aes ae 
ra me : 
ron oe 
ee eaarend 
pA 

an cond 

—— 5 
be nd a. 
vas wire 


= — - n 
- = —— enna 
nese ~~ 


—— = 
Waietin~devans 
mh pad 


aaa aura 


= 
rime 


e 


i 


ea tg 


4 ‘ 
it i? 


—————— 


HANDBOUND 
AT THE 


& 


UNIVERSITY OF 
TORONTO peree 


Digitized by the Internet Archive 
in 2009 with funding from 
Ontario Council of University Libraries 


http://www.archive.org/details/biochemicaljourn16biocuoft 


oF 


AS eee 


THE 


BIOCHEMICAL 
- JOURNAL 


CAMBRIDGE UNIVERSITY PRESS 
C. F. CLAY, Manacer 
LONDON: FETTER LANE, E.C. 4 


Mines 


H. K. LEwIs & cO., LTD., 136, GOWER STREET, LONDON, W.C. I 
WHELDON & WESLEY, LTD., 2—4 ARTHUR ST,, NEW OXFORD ST., LONDON, W.C. 2 
CHICAGO: THE UNIVERSITY OF CHICAGO PRESS 
(AGENT FOR THE UNITED STATES AND CANADA) 

BOMBAY, CALCUTTA, MADRAS: MACMILLAN & CO., LTD. 
TOKYO: THE MARUZEN-KABUSHIKI-KAISHA 


All rights reserved 


THE 


BIOCHEMICAL 
JOURNAL 


EDITED FOR THE BIOCHEMICAL SOCIETY 


BY 


Sir W. M. BAYLISS, F.R.S. 


AND 


ARTHUR HARDEN, F.R:.S. 


EDITORIAL COMMITTEE 


Pror. G. BARGER Pror. F. G. HOPKINS 
Pror. V. H. BLACKMAN Sir F. KEEBLE 
Mr J. A. GARDNER Pror. W. RAMSDEN 


Sir E. J. RUSSELL 


VOLUME XVI 1922 


CAMBRIDGE 
AT THE UNIVERSITY PRESS 
1922 


\ 


CONTENTS 


No. 1 


I. Note on Tannase. By D. Ruryp and F. E. Smita 


II. The Action of Hypophysin, Ergamine and Adrenaline upon the 
Secretion of the Mammary Gland. By E. Roruir, R. H. A. PLimmMerR 
and A. D. HusBanp : ‘ ; ‘ ‘ 


Ill. The Rearing of Chickens on the Intensive System. Part I. The 
Vitamin Requirements. Preliminary Experiments. By R. H. A. PLIMMER 
and J. L. Rosepate. With the assistance of A. CricuTon and R. B. 
TOPPING 


IV. The Rearing of Chickens on the Intensive System. Part II. The 
Effect of ‘‘Good’’ Protein. By R. H. A. Pummmer and J. L. Rosepae. 
With the assistance of A. CricuTon and R. B. Toprrna 


V. Distribution of Enzymes in the Alimentary Canal of the Chicken. 
By R. H. A. Piimer and J. L. Ros—pALE 


VI. The Amino-Acids of Flesh. The Di-Amino-Acid Content of 
Rabbit, Chicken, Ox, Horse, Sheep, and Pig Muscle. By J. L. RosepaLe 


VII. Xylenol Blue and its proposed use as a new and improved in- 
dicator in Chemical and Biochemical work. By A. ConEN 


VIII. The Examination of some Indian Foodstuffs for their Vitamin 
Content. By 8. N. GuosE 


‘IX. Conditions of Inactivation of the Accessory Food Factors. By 
8. 8S. Zmva : : ; 


X. An Improvised Electric Thermostat Constant to 0-02°. By 8. C. 
BrapFrorp. (With Plate I) : } . ’ 


XI. Tethelin—the Alleged Growth-controlling Substance of the 
Anterior Lobe of the cpr Gland. By J. C. DrumMonp and R. K. 
CANNAN . ; ‘ ; : 


XII. The Estimation of Pectin as Calcium Pectate and the Applica- 
tion of this Method to the Determination of the Soluble Pectin in Apples. 
By M.H. Carré and D. Haynes 


XIII. On the Occurrence of Manganese in the Tube and Tissues of 
Mesochaetopterus Taylori, Potts, and in the Tube of a Vario- 
pedatus, Renier. By C. BERKELEY . ‘ 


‘e\ 


PAGE 


11 


53 


60 


70 


vi CONTENTS 


XIV. Mammary Secretion. II. 1. The Quality and Quantity of 
Dietary Protein. 2. The Relation of Protein to other Dietary Constituents. 
By G. A. Hartwe i. (With seven figures) ; : 


XV. Effect of Severe Muscular Work on Composition of the Urine. 
By J. A. CampBeLt and T. A. WEBSTER ; : : 


XVI. The Value of Gelatin in Relation to the Nitrogen Requirements 
of Man. By R. Rosison. (With one figure) ; ; 3 


XVII. Distribution of the Nitrogenous Constituents of the Urine on 
Low Nitrogen Diets. By R. Ropison 


XVIII. The Estimation of Total Sulphur in Urine. By R. RoBison 
XIX. The Action of Yeast-growth Stimulant. By O. K. Wricut 


XX. The Function of Phosphates in the Oxidation of Glucose by 
Hydrogen Peroxide. By A. HarpEN and F. R. HENLEY. ‘ 


XXI. Vitamin Requirements of Drosophila. I Vitamins B and C. 
By A. W. Bacor and A. HarpEN ~ : hs pe tars 


No. 2 


XXII. Studies on Carbonic Acid Compounds and Hydrogen Ion 
Activities in Blood and Salt Solutions. A Contribution to the Theory of 
the Equation of Lawrence J. Henderson and K. A. Hasselbach. By Errk 
JOHAN WARBURG A . ; ; j : ‘ 


No. 3 


XXIII. The Effect of Cold Storage on the Carnosine Content of 
Muscle. By W. M. Crrrrorp. (With one figure) ; ; 


XXIV. Investigations on the Nitrogenous Metabolism of the Higher 
Plants. Part LI. The Distribution of Nitrogen in the Leaves of the Runner 
Bean. By A C. Curpnai. (With one figure) ; , 


XXYV. Milk as a Source of Water-Soluble Vitamin. Il. By T. B. 
Ossorne and L. B. Menpen. With the Cooperation of H. C. Cannon. 
(With one figure) 


XXVI. The Estimation of Non-Protein Nitrogen in Blood, By 
K. Ponper ; ’ ; : , ; 


XXVIII. The Conditions Influencing the Formation of Fat by the 
Yeast Cell. By | Smeptey Mac ean 


PAGE 


78 
106° 
111 


131. 
134 
137 


143 


148 


153 


341 


344 


CONTENTS 


XXVIII. The Action of Whole Blood upon Acids. By E. L. Kenna- 
way and J. MoInrosu. (With four figures) i : 


XXIX. Salivary Secretion in Infants. By C. Nicory 

XXX. The Relation of Salivary to Gastric Secretion. By 'T. Naka- 
GAWA ; . : : : ; ; : : : 
XXXI. The Relation of the Fat Soluble Factor to Rickets and Growth 


in Pigs. I. By J. Gotpine, 8. 8. Ziva, J. C. DRumMonp and K. H. 
Cowarp. (With Plates II and III and one figure) . 3 


XXXII. The Estimation of Calcium in Blood. By A. R. pane and 
J. H. Busniiy. (With one figure) ; ‘ : } 


XXXIIT. The Minimum Nitrogen Expenditure of Man and the 
Biological value of various Proteins for Human Nutrition. By C. J. 
Martin and R. Rosison. (With eight figures) ; 


No. 4 


XXXIV. On the Change of the Osmotic Pressure of Solutions of 
certain Colloids under the Influence of Salt Solutions. By D. Ogara. 
(With one figure) ; : ; 


XXXV. Further Observations on the Nature of the Reducing Sub- 
stance in Human Blood. By E. A. Coopmr and H. WALKER . . 


XXXVI. Blood Enzymes. II. The Influence of Temperature on the 
Action of the Maltase of Dog’s Serum. By A. Compron. (With two figures) 


XXXVII. The Mode of Oxidation of Fatty Acids with Branched 
Chains. IT. The Fate in the Body of Hydratropic, Tropic, Atrolactic and 
Atropic Acids together with Phenylacetaldehyde. By H. D. Kay and 
H. 8S. Raper 2 ; ; ; ; ; : R : 


XXXVIITI. Structures in Elastic Gels caused by the Formation of 
Semipermeable Membranes. By E. HarscueKx. (With Plate IV) 


XXXIX. A Modification of Basal Diet for Rat Feeding mee aie 
By M. Bonb. (With one figure) : A g 


XL. Synthesis of Vitamin A by a Marine Diatom (Nitzschia closterium 
W Sm.) growing in Pure Culture. By H. L. Jamuson, J. C. Drummond 
and K H.Cowarp (With one figure) ; : : 


XLI. The Phospholipin of the Blood and Liver in Experimental 
Rickets in Dogs. By J. S. SHarpy ‘ ‘ ; ; : ; 


Vii 
PAGE 


380 
387 


390 


394 


403 


407 


449 


455 


460 


465 


475 


479 


482 


486 


Vili CONTENTS 
XLII. The Constituents of the Flowering Tops of Artemisia afra, Jacq. 
By J. A. Goopson . : i ; ; . g : 


XLII. A Method ae the Estimation of Small Quantities of Calcium. 

By P. P. Latpiaw and W. W. Payne F 
XLIV. Production of Hydrogen Peroxide by Bacteria. By J. W. 

McLeop and J. GorDON : ; : 5 ‘ ; ; ; 


XLV. Note on Urinary Tides and Excretory gsi By J. A. 
CaMPBELL and T. A. WEBSTER ; : é : j 


XLVI. Note on a New Tannase from Aspergillus Luchuensis, Inui. 
By M. NIERENSTEIN 


XLVII. A Qualitative Tannin Test. By E. Arkinson and E. O. 
HAZLETON ws at Se Rage CUES eR De Pe ss Ae 


XLVIII. The Origin of the Vitamin A in Fish Oils and Fish Liver Oils. 
By J. C. DrummMonp and § 8. Zuva, with the co-operation of K. H. 
COWARD : A ‘ : 


~ XLIX. The Respiratory Exchange in Fresh Water Fish. III. Gold- 
fish. By J. A. GAnpNER, G. Kine and E. B. Powrrs ; ; 


L. Notes on some Properties of Dialysed Gelatin. By D. Jorpan 
Luoyp. (With one figure) : § : ‘ : ; : 


No. 5 


LI. The Mechanism of the Reversal in Reaction of a Medium which 
takes place during growth of B. diphtheriae. By C. G. L. Woir 


LIL. The Properties of Dibenzoyleystine. By C. G. L. Woxr and 
E. K. Ripea. (With one figure) . ; : , 


LIL. The Nitrogen-Distribution in Bence-Jones’ Protein, with a 
Note upon a New Colorimetric Method for kdb Estimation in 
Protein. By E. LUscurr . ; A ‘ ; 


LIV. The Autolysis of Beef and Mutton. By W. R. Fearon and 
D. L. Foster. (With six figures) ‘ 4 q 


LV. Note on the Oxidation of Carbohydrates with Nitrie Acid. By 
P. Haas and B, Russett-WeLis 


LVI. Opsonins and Diets deficient in Vitamins. By G. M. Frnpiay 
and R. MACKENZIE 


LVIL. On Carrageen (Chondrus crispus). U1, The Constitution of the 
Cell Wall, By B. Russett-WeE.is . 


523 


530 


541 


548 


574 


578 


CONTENTS 


LVIIT. Studies of the Coagulation of the Blood. Part IT. Thrombin 
and Antithrombins. By J. W. Pickrerine and J. A. Hewrrr : 


LIX. Investigations on the Nitrogenous Metabolism of the Higher 
Plants. Part III. The Effect of Low-Temperature Drying on the Distribu- 
tion of Nitrogen in the Leaves of the Runner Bean. By A. C. CurpnaLn 


LX. Investigations on the Nitrogenous Metabolism of the Higher 
Plants. Part IV. Distribution of Nitrogen in the Dead Leaves of the 
Runner Bean. By A. C. CaTBNALL : : : i : 

LXI. Note on the Non-Protein Nitrogen in Goat’s Milk. By W. 
TAYLOR . 5 ‘ j , : ; 

LXII. Smell. By E. R. Watson 


LXIII. Studies on the Pituitary. I. The Melanophore Stimulant in 
Posterior Lobe Extracts. By L.'T. HoGBEen and F. R. Wrn'ron 


LXIV. On the Significance of Vitamin A in the Nutrition of Fish. By 
K. H. Cowarp and J. C. DrumMonp. (With six figures) ; 


LXV. Note on Knoop’s Test for Histidine. By G. HunTER 


LXVI. The Estimation of Carnosine in Muscle Extract—A Critical 
-Study. By G. HunTER , 


LXVII.. Identification of Inulin by a My ape Method. By A. 
CasTeELLaNt and FF. E. TayLor : 


LXVIII. Feeding Experiments in connection with Vitamins A and B. 
Ill. Milk and the Growth-Promoting Vitamin. IV. The Vitamin A 
Content of refined Cod-Liver Oil. By A. D. Stammers. (With three 
figures) : : ; , : : ; 

LXIX. On the Cardiac, Haemolytic and Nervous Effects of Digitonin. 
By F. Ransom. (With three figures) ; 

LXX. The Heat-Coagulation of Proteins. By W. W. LePrescHKIn . 


LXXI. The Synthesis of ee from papa ties By A. R. Line 
and D. R. NANgr : : ; 


LXXII. An Investigation of the Changes which occur in the Pectic 
Constituents of Stored Fruit. By M. H. Carr&. (With three figures) 


ix 
PAGE 


587 


599 


x CONTENTS 
No. 6 


LXXIII. The Presence of the Antineuritic and Antiscorbutic Vita- 
mins in Urine. By N. VAN DER WALLE. (With five figures) 


LXXIV. The Suction Pressure of the Plant Cell. A Note on Nomen- 
clature. By W. Stites d j i 


LXXV. Respiratory Exchange in Fresh Water Fish: Part IV. Further 
Comparison of Gold-fish and Trout. By J. A. GARDNER and G. Kina 


LXXVI. Respiratory Exchange in Fresh Water Fish. Part V. On Eels. 
By J. A. GARDNER and G. Kine : d 


LXXVII. On a Series of Metallo-Cystein Derivatives. I. By L. J. 


LXXVIII. The Influence of Fat and soa bein on the Sepik 
Distribution in the Urine. By E. P. Catucarr ; 


LXXIX. On the Influence of the Spleen upon Red Blood-Corpuscles. 
I. By N. A. Borr and P. A. Hemres. (With two figures) 


LXXX. The Enzymes of the Latex of the Indian ats Sti 
Somniferum). By H. E. ANNETT i ‘ 


LXXXI. The Blood of Equines. By C. P. Naser. (With five Graphs) 


LXXXII. The Food Value of Mangolds and the Effects of Deficiency 
of Vitamin A on Guinea-pigs. eo EK. Boook and J. TRevan. (With Ten 
Charts) ; : : ‘ , ; ‘ ; 


LXXXIII. The Catalytic Destruction of Carnosine in Vitro. By 
W. M. Cirrrorp. (With four figures) ; ; 


LXXXIV. On the Vitamin D. By T. B. Haron. (With one figure) 


LXXXV. A New Phosphoric Ester Produced y the Action of Yeast 
Juice on Hexoses. By R. Ropison : ; . 


LXXXVI. Mammary Secretion. IV. The Relation of Protein to 
other Dietary Constituents. By G. A. Harrwetu. (With five figures) 


INDEX 


PAGE 


713 


727 


729 


736 


739 


747 


754 


765 


770 


780 


792 
800 


809 


825 


838 


I. NOTE ON TANNASE. 


By DONALD RHIND anp FRANCIS EDWARD SMITH. 


From the Biochemical Laboratory, Chemical Department, 
University of Bristol. 


(Received December 14th, 1921.) 


Since the middle of last century, this enzyme has been the subject of many 
investigations, but it is notable that no method has been described whereby 
its hydrolysing power may be estimated. The investigation described in this 
paper was undertaken at the suggestion of Dr Nierenstein, since it was thought 
that some such method might be of assistance to other work at present in 
progress in this laboratory. 

This enzyme was prepared from Aspergillus Luchuensis Inui.' by a slight 
modification of Freudenberg’s method [1913]. It was used without further 
purification, a 1 °% solution in distilled water being employed in each experi- 
ment. 

The gallotannin used was Schering’s leviss. puriss. which had been further 
purified according to Fischer’s method [1912]. An approximately 0-3 % 
solution in distilled water was prepared, and the exact amount of gallotannin 
present estimated by Nierenstein’s caseinogen method [1911] as modified by 
Spiers [1914]. 

The hydrolyses were carried out in 250 cc. Erlenmeyer flasks, 100 cc. of 
gallotannin solution, to which various quantities of tannase solution were 
added, being employed in each experiment. During the reaction the flasks 
were plugged with cotton wool, and toluene was added to the reaction mixture 
to prevent the growth of fungi or bacteria, the whole being kept in a dark 
incubator at 23°. The amount of gallotannin present in the reaction mixture 
was estimated immediately after the tannase solution had been added and 
again after the hydrolysis had proceeded for a certain time. 

The details of the estimation are as follows. A solution of approximately 
0-1 % KMn0O, was prepared and standardised against pure ammonium oxalate 
of known strength; 20 ce. of a solution of 0-5 % of indigo carmine in dilute 
sulphuric acid (A-R reagent, British Drug Houses), being used as indicator. 
As the indigo carmine reacts with potassium permanganate, it is at first 
necessary to determine the quantity of the latter used by the indicator 
solution. To this end 20 cc. of this solution were placed in a large porcelain 
basin, 750 ce. of distilled water added, and titrated with the KMn0, solution. 
The end-point is reached when a faint pink colour can be seen round the edge 
of the solution. This titration gives the amount of KMn0Q, solution taken up 
by the indicator, which must be allowed for in all calculations. The gallo- 

1 We are indebted to Mr F. W. Mason of the Bureau of Biotechnology at Leeds for a pure 
culture of this fungus. 

Bioch. xvi 1 


2 D. RHIND AND F. E. SMITH 


tannin present is estimated as follows: 25 cc. of the reaction mixture are 
withdrawn and 4 ce. of this solution placed in a porcelain dish together with 
20 ce. of indigo carmine solution and 750 ce. of distilled water. The whole 
is then titrated with permanganate solution until the end-point as previously 
described is obtained. This titration represents both the tannin and non- 
tannins present. The remainder of the solution is shaken for 15 minutes with 
1 g. of fat-free caseinogen and then passed through a barium sulphate filter. 
This process is repeated twice. When all gallotannin is removed [ef. Nieren- 
stein, 1911], 4 cc. of this filtrate are then titrated as before. The difference 
between the two titration readings represents the amount of gallotannin 
present. The actual quantity may be calculated by means of Spier’s ammonium 
oxalate gallotannin ratio, viz. 1g. of ammonium oxalate is equivalent to 
0-4648 g. of gallotannin. 

In estimating the amount of gallotannin present after hydrolysis, the 
procedure is the same as that just described except that it is necessary first 
to filter off the precipitate of gallic acid formed. 

The following table gives the results obtained. 

In each case three titrations were made, the mean of these being recorded; 
also as a check each worker carried out an independent investigation. 


Ce. of 1 % tannase 
solution added to Duration of Percentage of 


Age in 100 ce. of gallo- hydrolysis allotannin 

Sample days tannin solution in hours ydrolysed 
1 112 10-0 23-75 14-70 
14-40 
2 2 5-0 19-50 10-52 
11-02 
2 2 5-0 45-00 16-60 
15:70 
3 2 5-0 19-50 13-48 
13-25 
4 2 7-5 19-50 13-00 
12:17 
4 2 75 45-00 18-25 
17-30 
5 2 10-0 19-50 15-00 
15-35 
5 2 10-0 45-00 20-00 
20:51 


From the results given in the above table it is evident that (1) this method 
represents a moderately accurate procedure for estimating the hydrolysing 
power of any sample of tannase; (2) the bulk of the hydrolysis takes place 
during the first 24 hours. 

It is proposed by Dr Nierenstein to continue this work at a later date, 
but since the authors are unable to pursue this research at present, it was 
thought advisable to publish this preliminary note. 


REFERENCES, 
Fischer (1912). Ber. 45, 923, 
Freudenberg (1913). Ber, 62, 177. 
Nierenatein (1911), Chem. Zeitg. 36, 31. 
Spiera (1914), J. Agric. Sct, 6, 77. 


Il. THE ACTION OF HYPOPHYSIN, ERGAMINE 
AND ADRENALINE UPON THE SECRETION 
OF THE MAMMARY GLAND. 


By ERNEST ROTHLIN, ROBERT HENRY ADERS PLIMMER 
AnD ALFRED DENNIS HUSBAND. 


From the Biochemical Department of the Rowett Research Institute of Animal 
Nutrition, University of Aberdeen and the North of Scotland College of Agri- 
culture, and the Physiological Institute, University of Zurich. 


(Received December 20th, 1921.) 


THE action of extracts of various organs, especially that of the pituitary gland 
(hypophysin) upon milk secretion has been frequently investigated. The 
results of these investigations have been summarised by Hammond [1913], 
who made an extensive study of the action of hypophysin upon the milk flow 
in the goat. He found that there was an immediate action and that the effect 
soon passed off, and he concluded that the action was not upon the muscular 
tissue of the gland but rather upon the glandular epithelium. The action 
depended upon the state of nutrition of the animal and upon the stage of 
lactation. The composition of the milk was not appreciably altered, though 
that produced by the action of the drug had a higher percentage of fat than 
other samples. The milk analyses were limited to estimations of fat and total 
solids and did not extend over long periods. 

Hammond and Hawk [1917] added some further details respecting the 
action in different states of nutrition and tested adrenaline, which was found 
to have no effect. 

As two goats under normal conditions, on a constant diet, but in different 
stages of lactation were available, there arose the opportunity of once more 
testing the action of hypophysin upon the stage of lactation and of extending 
Hammond’s results. Complete analyses of the milk of these goats had been 
previously made, and by continuing this procedure and including extra samples 
some further details could be ascertained. At the same time it was of interest 
to compare its action with that of adrenaline and of ergamine or histamine. 
The latter substances, as is well known, have an action corresponding with 
that of excitation of the sympathetic or para-sympathetic nervous systems. 
Their action, if any, would give probably some information upon the innerv- 
ation of the mammary gland with nerve fibres of the involuntary nervous 
system. Hammond’s results with hypophysin have been confirmed and ex- 
tended ; ergamine and adrenaline, as might have been expected, had no action. 


1—2 


q E. ROTHLIN, R. H. A. PLIMMER AND A. D. HUSBAND 


EXPERIMENTAL. 


The hypophysin, ergamine as acid phosphate, and adrenaline used for 
injection were the commercial preparations supplied by Burroughs, Wellcome 
and Co. The injections were always subcutaneous in the lumbar region of 
the body. They were made at different times of the day, usually from one 
to two hours before milking and generally in both goats about the same time. 
Before the injection of the drugs an injection of 0-9 % saline was made and 
once again later; it had no effect and thus the effect of the actual injection 
was excluded. 

The experiments were carried out in two periods; during the first period 
there was an interval of only one or two days between the injections; during 
the second period there was an interval of about seven days between the 
injections of the three drugs in order to ascertain if the total volume of the 
milk differed in the period of injection from that before and after. The in- 
jections were then made more frequently, sometimes twice in the day. 

The two goats used in these experiments were in different periods of 
lactation. Goat A had two kids on June 14th, goat N one kid on Feb. 2nd, 
1920. Goat A was in a good state of lactation, goat N in a poor state. 

Goat A during the first ten days was outdoors at grass by day and indoors 
by night; subsequently she was kept indoors permanently, fastened by a 
chain in a ground floor room of the building. Goat N was always kept in 
the building in the same room, but in a cage in which she could freely move 
and turn about. Both goats were given 112g. of oatmeal at about 8 a.m., 
noon and 5 p.m. This was eaten rapidly. Except for the first ten days, when 
goat A had grass during the day, after the oatmeal both had 454 g. of hay, 
which they were allowed to eat at their leisure. They were thus under fairly 
normal conditions and upon a constant diet. Their weights varied very little 
during the course of the experiment which lasted about three months, and as 
far as could be noticed they were in excellent health. 

The milk was not drawn by suction, but always by hand by the same 
experienced milker, who had drawn the milk previously over a long period. 
During this time the goats were milked twice daily, morning and evening; 
the two samples were mixed and analysed. To observe the effect of the drugs 
it became necessary to draw the milk more frequently and it was taken at 
9.30 a.m., 11.30 a.m. and 5 p.m. and occasionally at 6.30 or 7 p.m., depending 
upon the time of injection. 

Complete daily analyses of the milk of these goats had been previously 
made. Owing to the number of samples now taken (six to eight) it was not 
possible to analyse each separately and two or more samples were combined. 
The combination of two samples was made in such a way that the particular 
sample after injection was kept separate. The analyses are shown in the 
table, numbered 1, 2 or 3 according to the milking time of 9.30 a.m:, 11.30 a.m. 
and 5 p.m. 


SECRETION OF MAMMARY GLAND 5 


The methods of analysis were those usually employed in milk analysis. 
Fat was estimated by Soxhlet’s method, total protein by Kjeldahl’s method: 
caseinogen by this method after its precipitation from the sample by dilute 
acetic acid and washing; albumin in the filtrate from the caseinogen by pre- 
cipitation with tannic acid, washing and Kjeldahl’s method. The nitrogen 
figures by Kjeldahl’s method were multiplied by the factor of 6-38 to give 
the protein. Lactose was estimated in a sample, precipitated by “dialysed 
iron,” by Fehling’s method. Ash was determined by incineration in a small 
crucible. The data are expressed in grams per 100 cc. milk. 

The figures in the following table show the volumes at each milking, and 
the fat analyses, and the calculations to give the figures for the output of the 
whole day, the times of injection, etc. The data are so numerous that they 
are greatly abbreviated, and only those showing the essential features are 
given. The amounts of lactose, total protein, caseinogen and albumin except 
on the first two days are omitted and none of the figures of the second period 
is shown. 


RESULTS. 


(a) Changes of Volume. 


1. Hypophysin. The injection of hypophysin generally about two hours 
before the milking time produced a flow of milk only in goat A, which was 
in an early stage of lactation. The effect is distinctly marked in the volume 
of the second sample collected after an interval of two hours from the first 
sample. The volumes of 250 and 260 cc. were obtained in comparison with 
normal volumes varying from 35 to 145 cc. The volume of the next milking 
was then smaller than the normal, 230 and 150 cc. against 255 and 325 ce. 
An increase was not definitely noticed in the volume of the third sample 
collected after an interval of about six hours from the previous one. A volume 
of 350 cc. was obtained between volumes of 325 and 410 cc., but the average 
volume for this milking time was between 200 and 300 cc. No increase in 
volume was observed in the first sample collected at 9.30 a.m., 7.e. about 
16 hours after the sample at 5 p.m. The total daily volume of milk was not 
appreciably altered. A change over the total period could not be observed 
as the goat A gave a gradually diminishing volume of milk. 

No change in volume of the samples was observed in goat N after an 
injection of hypophysin. This goat gave a more regular daily volume and 
the effect was to make a slight general increase of the total volume; previous 
to the injection the average volume was 423 cc.; during the period it was 
448 cc, and after 468 cc. A similar slight increase was again noted later, the 
volumes being 395, 424 and 509 cc. The increase may however be due to 
other circumstances, as the goat N had at the time of this injection been in 


lactation for eight to nine months. The injection in goat A at the later period 


did not produce a rapid flow as at the former time, but there was a very 
slight increase of total volume; the averages of the periods were 710, 724, 730. 


6 E. ROTHLIN, R. H. A. PLIMMER AND A. D. HUSBAND 


PROTOCOLS. 

Goat A Goat NV 
Soon VoL Ash Fat Lact. T. prot. Cas. Alb. Vol. Ash Fat Lact. T. prot. Cas. Alb. 
7 1100 0:83 622 4-06 3-64 3-00 0-47 450 0-91 675 455 357 2-57 0-75 
8 1200 0-80 780 4:23 3:59 2-95 0-48 390 0:90 5-07 4:64 3-73 2-75 0-75 
*9 1200 0-76 7-02 4:26 361 2-95 0-51 430 0:97 550 460 3:52 2:53 0-77 
410 1 900 0:85 420 436 3-77 2:90 0-51 315 0-91 531 4:60 363 2-53 0-75 
2 250 080 11-21 4-00 361 2-75 0-46 10 — 12:07 4:37 368 2:50 0-80 
3 230 0-82 912 411 385 290 0-47 140 0:90 625 464 369 2:59 0-70 
per total 1380 0-83 629 4:26 3-75 287 0-49 465 0-91 847 461 3-65 2-55 0-74 
fil - ois p0°87 575 4:31 400 3-07 0-53 90 f0°9 522 4:70 3-77 264 0-75 
2 55 0-80 11:35 4-08 3-78 3-02 0-47 12° — 867 4:38 3-73 2:26 0-52 
1170 0-87 601 4:30 3:99 3-07 0-53 410 0:91 528 466 3-77 263 0-74 


* Both goats injected with 1 cc. 0-9 % NaCl after first milking. 
+ Both goats injected with 0-5 cc. hypophysin after first milking; milk collected one hour after injection 


(10.25 a.m.). t No injection. 
Goat A Goat NV 
Date 
Sept. Vol. Ash Fat Vol. Ash Fat Remarks 
12-1320 — — 310 — — No injection. 
2 100 — _— 35 _ _— 
3 25 — — 85 — 
1275 = 0-84 780 430 090 4-99 
13 1 840 0-83 500 300 0-91 4:89 Both injected with 1 cc. hypophysin 
2 260 0-74 14-09 37 =: 0-82 9-29 after first milking; milk collected 
3 150 0-90 7:37 83 0-94 4:77 one hour after injection. 
1250 0-82 709 420 8 0-91 5-25 
14 1 940 . : 335 y x No injection. 
3 3957 0°85 413 jJigr 0:96 4-31 
2 95 0-74 10°35 42 0-90 10-74 
1360 =-0-84 457 495 0-95 4-86 
1 1 680 086 465 325 0-92 4:50 Goat A injected with lec. 09% 
2 70 086 10-62 10 _ 9-65 NaCl at 3 p.m. Goat N injected 
3 350 0-82 900 130 8 0-90 6:29 with 1 ce. hypophysin at 3 p.m. 
1100-00-85 641 465 0-91 5-11 
16 Lt 520 O84 $400 280 096 410 No injection. 
2 95 0-80 6-30 40 380-90 7-52 
3 410 0-84 639 115 096 4-69 
1025 0-84 537 435 095 4-57 
17 1 800 —_ _ 305 = 0-94 3°89 Both goats injected with 1 ec. hypo- 
2 4 — _— 45 — 5-98 physin at 8.15 a.m.; milk collected 
3 265 —_ _— 95 0-92 4-00 one hour later. 
1210 = 0-83 554 445 O93 412 
6.3 ae — _ 325 O95 412 No injection. 
2 35 — _ 40 = 0-88 7-40 
3 2100 — — 110 O88 4-69 ‘ 
995 O81 551 475 0-93 4-53 
19 1 680 0-80 515 315 a No injection. 
2 50 0-80 9-30 35 — -- 
3 230 = 0-80 840 3110 —_ 
960 0-80 615 460 0-95 415 
4 30) «0-83 9-40 35 0-04 5-53 
2 1 660 O85 550 300 — —- Goat A injected with 1 cc. hypo- 
2 65 0-82 810 é — -— physin at 8.15 a.m.; milk collected 
é 175 = 0-88 652 105 _ a one hour later, 
030 =OBS 600 470 0-04 439 
2 64S ; y 305) 0-93 394 Goat N injected with 1 mgm. erga- 
‘ 145 O67 =—8S 45S 075 «B81 ~—_—>mine phosphate at 0.45 a.m.; milk 


2 70 O82 10-25 100 0-04 407 collected one hour later, 
835 O87 O19 450 Os 4-60 


Date 
Sept. 


22 


23 


26 


27 


28 


Oct. 


won wrdoe wwe 


wne 


one —- Whe 


one -F Whe wnoe 


wne 


woe 


SECRETION OF MAMMARY GLAND 7 


Goat A 
Vol. Ash 
570 30°85 
120 0-88 
110 0-74 
800 0-84 
730 0-90 
140 =0-82 
160 0-91 
1030 = 0-89 
660 0:85 
80 0-83 
200 80°86 
940 0-86 
640 0-90 
50 =: 0-80 
140 = 0:87 
830 80-89 
570 — 
70 a 
205 — 
845 0-89 
33. _ lost 
622 0-88 
20 ae 
225 0-87 
900 §=0-88 
600 0-87 
ry es 
130 =: 0-86 
780 0-87 
208 
230 y 0-86 
880 0:87 
540 0:87 
45 seal 
110 = 0-87 
695 0-87 
620 0-90 
Te a 
210 0-86 
870 0-89 
590 =0-81 
30 =e 
200 30-86 
820 0-84 
24 —_— 
536 30-87 
40 ee 
140 0-87 
740 0-87 
650 = 0:86 
30 os 
170 «0-88 
850 0-86 
480 0-91 
75 — 
190 0-89 
745 0-90 


s 
S 


— 


SASS APSA SaASa Seon 
S£8 S885 SSSu Sess 


re Ee 
& Se 5 


Goat N 
Ss a— 
Vol. Ash 
250 
35 = =©0-90 
385 0-94 
320 = 0-94 
18 — 
142 0-90 
480 0-93 
315 0-92 
30 = 0-86 
105 =: 0-93 
450 0-92 
300 = 0-92 
20 pe 
100s: O-91 
420 0-92 
295 — 
15 — 
120 — 
430 0-98 
33 0-93 
282 0-94 
10 — 
130 0-96 
455 0-94 
300 = 0-92 
15 — 
125 0-93 
440 0-92 
295 380-95 
10 
435 0-94 
295 0-93 
20 peas 
100 =: 0-93 
415 0-93 
285 0-97 
35 ol 
110 —s 0-94 
430 §=0-96 
300 = 0-93 
15 -— 
100s 0-91 
415 0-92 
20 come 
265 0-94 
35 BES 
100 0-94 
420 0-94 
315 = 0-91 
10 —_— 
110 = 0-92 
435 0-91 
260 3=0-95 
10 —_ 
115 = 0-94 
385 0-95 


4-56 


bp 
Sam 


ee a SFT | 
SS 2-48 


4-67 


Remarks 
Goat A injected with 1 mgm. erga- 
mine phosphate at 9.45 a.m.; milk 
collected one hour later. 


Both goats injected with 1 mgm. 
ergamine phosphate at 3.45 p.m.; 
milk collected at 5 p.m. 


No injection. 


Goat A injected with 0-5 mgm. erga- 
mine phosphate at8.15a.m. GoatN 
injected with 1-0 mgm. ergamine 
peers at 8.15a.m. Milk of 

th goats collected one hour later. 


Both goats injected with 1 mgm. 
ergamine phosphate at 3.30 p.m. 
Milk collected at 5 p.m. 


Both goats injected with 1 mgm. 
ergamine phosphate at 6.15 p.m. 
Milk collected at 9.15a.m. next 
day. 


No injection. 


Both goats injected with 1 mgm. 
adrenaline at 1.15 p.m.; milk col- 
lected at 5 p.m. 


Both goats injected with 1 cc. 0-9 %, 
NaCl at 1.30 p.m.; milk collected 
at 5 p.m. 


Both goats injected with 1 mgm. 
adrenaline at 9.45 a.m.; milk col- 
lected at 11 a.m. 


No injection. 


Both goats injected with 1 mgm. 
adrenaline at 1.15 p.m.; milk col- 
lected at 5 p.m. 


No injection. 


8 E. ROTHLIN, R. H. A. PLIMMER AND A. D. HUSBAND 


2. Ergamine. The injection of ergamine had no pronounced effect on the 
separate volumes of the samples of either goat and the average volume of 
the total milk per day was about the same as when no injection was made. 
A general effect over the whole period was not noticeable in the case of 
goat A which was giving a diminishing flow of milk, but a slight diminution 
can be made out in the case of goat N. During the second period of injection | 
the flow of milk in goat A was more constant, the average daily volumes 
being 645, 656, 657 cc. for the pre-, actual, and post times; the average figures 
of goat N at this time were 532, 493, 576. In both cases the changes are not 
essentially different from the ordinary daily volumes. 

3. Adrenaline. During the first period no effect could be seen in the case 
of either goat, both the separate volumes and the average daily volumes re- 
maining unchanged. During the second period there was an apparent increase 
in goat A, but the increase occurred at a time when the daily flow was 
lessening; the reverse effect can be made out in goat N. The general effect 
is not sufficient to indicate any definite action of adrenaline. 


(b) Composition of the Milk. 


No real change in the chemical composition of the total milk per day 
occurred in either goat after the injections of hypophysin, ergamine and 
adrenaline. If any change did occur, it was not greater than the normal 
variations in composition. These normal daily variations are most marked 
in the fat content. Possibly the fat content was lowered over the total 
period during which hypophysin was injected and raised during the ergamine 
period, but again the alterations were not more than those which occur daily. 

The three separate samples of milk taken at different times of the day 
after different intervals had normally a distinctly different composition. The 
fat content was always highest in the second sample taken two hours after 
the first, and the early morning sample had the lowest amount of fat; it was 
taken after an interval of 16 hours. The higher fat content was observed in 
both goats and is thus independent of the period of lactation. The second 
sample had less protein and lactose than the other samples. On the average 
the first sample at 9.30 a.m. had more protein and lactose than the third 
sample, but the differences are not distinctly marked off. 

If allowance be made for the ordinary daily variations, the separate 
samples do not show any appreciable difference from the normal. 


DIscussION. 


Since a marked increase in the volume of milk was only observed in goat A 
at the time of the second milking, and since it was followed by a diminish ed 
volume at the third milking, it appears that the action of hypophysin is 
powerful, rapid and of short duration. The phase of hyper-secretion after 
injection is followed by a compensatory phase of hypo-secretion. This effect 


ee eee er ee 
oi a Pn : 


SECRETION OF MAMMARY GLAND -9 


explains the absence of an increase in volume at the time of the first milking 
and third milking. The interval between the milkings was here 16 hours; an 
increase followed by a decrease would not be noticed. The interval between 
the first and second milkings was two hours. The gland secretes normally 
(Sept. 12) about 50 cc. every hour; after hypophysin the secretion was 125 ce. 
(Sept. 10) and 130 cc. (Sept. 13) per hour. Hypophysin had no effect in goat N 
and at the later period in goat A. The action of hypophysin is therefore 
probably not directly upon the gland. If the smooth muscle were stimulated, 
a flow of milk should follow at either period of lactation. If the secreting cells 
were affected a flow of milk should also follow in any case. It is most probable 
that it may act indirectly through the reproductive organs, which have been 
proved to contain substances acting upon the secretion of the mammary 
gland. At the earlier period of lactation the reproductive organs will be in 
a state of activity, whereas in the later stage their state of activity will have 
disappeared. Hypophysin thus may act upon an active organ which produces 
the galactogogue; further work will be required to determine whether this 
active organ be the ovary, corpus luteum, uterus, or placenta. 

In practice, the injection of hypophysin will only be of value at early 
stages of lactation and it may be able to bring a gland into activity when it 
is not already in that state. It must be remembered that the total volume 
produced by the gland is not appreciably increased, as shown above and 
previously described by other investigators. 

The composition of the total milk per day is not altered by the injection 
of hypophysin. It has been previously considered that hypophysin causes a 
flow of milk with a higher fat content, but, as the data show, it is followed 
by a milk with a lower fat content. The high fat content of a sample taken 
at a short interval between milkings is normal and is not really due to the 
action of hypophysin. 

As milk contains more fat if collected at a short interval after a previous 
longer interval, it appears that the secreting mechanism of the mammary 
gland is of a two-fold character; a mechanism producing fat and a mechanism 
producing protein and lactose. The fat-producing mechanism begins to act 
first and a milk of high fat content results; the protein and lactose mechanism 
acts later and dilutes the fat content to the normal value. The fat mechanism 
is probably more easily influenced by other conditions, since the amount of 
fat in milk is very variable, whilst those of protein and lactose are fairly 
constant. 

Former investigators have observed that adrenaline has no action upon 
the mammary gland; these results confirm their work. 

Ergamine also has no action. If these substances have a positive or negative 
action, it is rapidly followed by compensation. 

The inaction of adrenaline points to the absence of sympathetic nerve 
fibres from the mammary gland, if the general law of the relationship between 
the action of adrenaline and the presence of sympathetic fibres is here correct. 


10 E. ROTHLIN, R. H. A. PLIMMER AND A, D. HUSBAND 


It has been suggested from analogy to other glands (stomach) that ergamine 
perhaps acts on para-sympathetic fibres. The inaction points to the absence of 
such fibres from the mammary gland. 


SUMMARY. 


Hypophysin produces a flow of milk only in the early stages of lactation. 
Its action is rapid and powerful, but of short duration. The flow (hyper- 
secretion) is followed by a smaller quantity than normal (hypo-secretion). 
The total volume of milk per day is not altered. 

Since hypophysin does not act at later stages of lactation, it is probable 
that its action is indirect through the organs of reproduction. 

The quality of milk secreted after the action of hypophysin is not different 
from the normal. Normal milk has a high fat content, if it be collected at a 
short interval after the last milking. 

The secretory activity of the mammary gland is not influenced by the 
subcutaneous injection of adrenaline or ergamine. The gland probably does 
not contain sympathetic or para-sympathetic nerve fibres for the secretory 
mechanism. If they be present, the inaction of adrenaline and ergamine is 
exceptional. 


REFERENCES, 


Hammond, J. (1913). Quart. J. Eup. Physiol. 6, 311. 
Hammond, J. and Hawk, J. C. (1917). J. Agric. Sci. 8, 147. 


Ill. THE REARING OF CHICKENS ON THE IN- 
TENSIVE SYSTEM. PART I. THE VITAMIN 
REQUIREMENTS. 


PRELIMINARY EXPERIMENTS. 


By ROBERT HENRY ADERS PLIMMER & JOHN LEWIS ROSEDALE 
WITH THE ASSISTANCE OF 


ARTHUR CRICHTON & ROBERT BAYNE TOPPING. 


From the Biochemical Department, Rowett Research Institute, University of 
Aberdeen and North of Scotland College of Agriculture. 


(Received December 20th, 1921.) 


Tue attempts of previous workers to rear chickens in confinement under 
laboratory conditions have not been altogether successful. Drummond [1916] 
considered that there were other factors than an adequate diet which must 
be supplied to the growing chick. Osborne and Mendel [1918] commented 
upon various statements by Drummond and remarked “if the conditions 
under which chickens continue to grow normally in confinement can be 
learned, it will be possible to obtain much information of practical use in 
poultry husbandry.” Further they say “There is a widespread belief among 
poultry raisers that young chickens cannot be reared under the artificial con- 
ditions of housing and diet which many other experimental animals tolerate 
without detriment. The current ideas are expressed in the statement that 
the birds must be kept ‘on the ground,’ that they must have exercise, outdoor 
life, and green food.” 

In their previous experiment [1916] with four chicks from 3 to 4 weeks 
old, two died and two were observed to the age of 77 days. Two other birds 
81 days old at this time were actually kept for a total of 309 days in which 
time they reached maturity in apparent full vigour. The birds at the ages 
of 77 and 81 days had very ruffled feathers; this appearance was attributed 
to frequent handling. Osborne and Mendel believed the success was largely 
due to the supply of blotting paper as roughage. Out of another group of 
ten birds, two only reached maturity; the others failed to grow and showed 
leg weakness. We may record the better success of Buckner, Nollau and 
Kastle [1915] and of Palmer and Kempster[1919]. Hart, Halpin and Steenbock 
[1920] have emphasised the frequent occurrence of leg weakness in chickens 
reared in confinement and, like Osborne and Mendel, attribute the cause to 


12 R. H. A. PLIMMER AND J. L. ROSEDALE 


the absence of roughage from the diet. Drummond’s birds were fed upon ordinary 
commercial chicken food; Osborne and Mendel fed their birds upon the foods 
which they supplied to rats, containing butter fat and protein-free milk as 
sources of the A and B accessory food factors, a diet quite satisfactory for 
their growth. Buckner, Nollau and Kastle used grain mixtures and supplied 
some “green” food. Hart, Halpin and Steenbock used a diet similar to that 
of Osborne and Mendel. 

It is now known that, though rats and a few other animals can be reared 
on a diet containing only A and B factors, most animals, including man, require 
the inclusion of C factor for the maintenance of health. It was thus possible 
that the poor success was due to the absence of this third factor, although it 
was added to the diets fed by Hart, Halpin and Steenbock. With reference 
to this third factor, the work of Chick and her colleagues [1919] insists upon 
the inclusion of a definite daily quantity which must be above a certain 
minimal amount. The necessity for a daily amount of the other vitamins has 
not been so clearly established, but at any rate it is indicated by the experi- 
ments of Cooper [1913] and of Chick and Hume [1917]. 

In these experiments the birds have been supplied with the usual protein, 
carbohydrate and fat; salts other than those contained in the foodstuffs were 
not added. Accessory food factors have been given in the form of cod liver 
oil for A, autolysed yeast or marmite for B, and lemon or orange juice for C. 
The choice of these three substances was made on account of the general 
experience that cod liver oil is the most concentrated in A factor, marmite 
in B, and lemon or orange juice in C, as well as the fact that these substances 
are very fairly constant in their respective vitamin content. The diet was 
composed essentially of oatmeal and milk and no account was taken of the 
presence of vitamins in these foods. Water was always freely provided. 

Arbitrary amounts of the vitamin foods had to be chosen as the require- 
ments for chickens are unknown. Later work is designed to determine the 
minimal daily amount of each vitamin, and subsequently these amounts will 
have to be translated into the terms of natural foodstuffs as supplied by the 
poultry keeper. It must be pointed out that the natural foods are very 
unequal in their vitamin content and that at present, in the absence of chemical 
knowledge of vitamins, only comparable data are capable of investigation. 
Also, it must be clearly stated that the diet now supplied is not one that 
would be given under ordinary conditions to poultry. 

With the food supplied 21 chickens out of 24 have been reared to maturity; 
two were lost from illness apparently due to insufficient vitamin B in the 
diet and one was lost accidentally ; in other terms 91 °%, were successfully reared. 
In Part II [1922] it will be seen that nine birds out of nine, or 100 % were 
brought to maturity. 


— ae ee ge ee A, 
=; en = 


VITAMIN REQUIREMENTS OF CHICKENS 13 


EXPERIMENTAL. 


Housing. The chicks up to the age of about four weeks were housed in an 
ordinary type of foster mother, heated by a paraffin lamp’. After this age 
they were kept in wooden poultry houses measuring 6 x 4? x 64 feet high 
with an entrance door at one end and an opening into a run at the other end. 
Ventilation hatches were provided at each end at the top. The run of wire 
netting on wooden supports on a wood floor had the size of 6 x 4? x 3 feet 
high. Altogether there were four identical houses, two on the roof of the 
building, two in a room measuring 30 x 20 x 15 feet high on the ground floor 
of the Institute. The foster mother was kept in the ground floor room. 

The wooden floors of the foster mother, poultry houses and runs were 
covered with broken peat moss litter, except that during wet weather, sand 
and coal ashes were used in the houses on the roof. The peat moss litter was 
scraped about but, as far as could be ascertained, very little was eaten. Grit 
was supplied in the form of coal ashes from the age of about two months. 
The birds were weighed separately every two days. 

Diet. There is very little information about the daily quantity of food 
consumed by little chicks; it is stated [Board of Agriculture and Fisheries, 
1907] that chicks eat about 3 lbs. of dry food during the first eight weeks of 
their life. Consequently definite amounts were prepared every day; portions 
were given at intervals of 2-3 hours to the little chicks and 2—4 times daily 
as they grew older. In this way a record could be kept of the amount eaten 
at the different ages. It was found that the total consumption corresponded 
to that stated. 

Kilned oatmeal of pin head size was selected as the basal diet, mainly on 
account of its method of manufacture which would exclude the presence of 
vitamins, but their entire absence could not be expected. A small quantity of 
milk was also given to provide extra protein; the amount was at first rather 
variable, but later was always measured. Towards the end caseinogen was 
substituted for milk in the case of half the birds and at the same time B factor 
was excluded, since a comparison between caseinogen and marmite and at 
the same time an estimate of the vitamin content of the caseinogen, were 
wanted. 

The accessory food factors were added to the daily quantity of oatmeal. 
According to the health of the birds the original amounts required alteration. 
The course of the experiment is most conveniently divided into periods from 
which the variations in food consumption and vitamin needs are more easily 
ascertained. 

Twenty-four chicks were comprised in the group, 12 white and 12 black 
Leghorn. They were hatched in an incubator and were from Miss Fraser’s 
strains at the North of Scotland College of Agriculture. At the time of their 
arrival their age was approximately three days; several were suffering from 
diarrhoea and were very weak and some mortality was expected. 

1 Kindly lent by Miss Fraser of the North of Scotland College of Agriculture. 


14 R. H. A. PLIMMER AND J. L. ROSEDALE 


Period 1 from July 13th to Aug. 10th, 1920. The birds were housed in the 
foster mother and supplied daily with 120g. oatmeal, 5 cc. cod liver oil, 
0-5 g. marmite, 30 cc. lemon juice. 

Extra food, as required, after the above mixture had been eaten, was given 
in the form of dry oatmeal, or a mixture of dry oatmeal and biscuit meal; at 

first 30 g. was eaten daily, but it gradually increased until at the end of the 
- period 200g. was eaten. Including the basal mixture, the total food con- 
sumption per bird increased from 6 to 13 g. per day. 

In addition the birds were given 200 to 250 cc. of milk every day; it was 
mostly spilt and thus no record of its consumption could be made. On two 
occasions the yolk of an egg was given and lemon skins were put into the run 
for the birds to peck at and for their amusement; as far as could be seen, 
none was eaten. 

On July 21st, 2.e. on the ninth day, one of the weak birds was found dead 
in the outer part of the foster mother. ~ 

The other chicks recovered from the chicken diarrhoea and no signs of 
illness were noticed in any of the other birds. 

The average weight of the birds on July 27 was 78 g.; on Aug. 20 it was 
150 g. The increase averaged 5-1 g. per day. 

Period 2 from Aug. 11th to Sept. 3rd, 1920. The 23 birds were now divided 
into two groups. Group I had six black and six white; group II had five 
white and six black chicks. The selection was made so that the average 
weight of the two groups was approximately the same: group I weighed 154, 
group II 156 g. per bird. The two groups were housed in the poultry houses 
in the ground floor room. 

Each group received as basal diet 120 g. of oatmeal. The vitamins were 
added as before, but the amounts were divided between the two groups, 
except in the case of C factor. Each group received 2-5 ce. of cod liver oil 
and 0-25 g. of marmite per day. The difference in the amount of C factor 
was made on account of the stress that poultry keepers lay on green food for 
birds, and because, presumably, C factor is the most abundant in these foods. 
Group I was given 15 cc., group II 20 ec. of lemon juice per day; in the case 
of group I it was thus the same as before. 

For the first three days extra food was given as oatmeal and biscuit meal; 
each group consumed about 44 g. per day. On Aug. 14th, the extra food was 
given as a mixture of oatmeal and milk, 10 ec. of milk being mixed with 
30 g. of oatmeal. The proportion of milk was raised to 20 cc. on Aug. 26th, 
since the protein ratio was very low on the former mixture. The protein 
ratio of the total thus became | : 6 approximately. No account of the presence 
of vitamins in the milk was taken as it is now proved that the vitamin content 
of milk is variable and the quantity so supplied was expected to make little 
difference to that supplied directly. 

Five days after the separation into two groups, on Aug. L5th, illness was 
observed amongst the birds of group I. There was no particular characteristic 


VITAMIN REQUIREMENTS OF CHICKENS 15 


sign; the birds were less active and drooped. The cause was attributed to 
insufficiency of C factor. The quantities were therefore increased. Group I 
was given 35 cc. and group II 30 cc. per day, 7.e. in the reverse order to that 
previously. A little improvement was noticed, but the birds still remained 
unwell. Consideration was then given to the amount of B factor. The quantity 
of 0-25 g. to each group was very small and the amount was raised to 0-5 g. 
per day to each group on Aug. 19th. 

After this increase some birds of group II showed similar signs of illness 
to those of group I, and the quantity of lemon juice was again increased to 
40 and 50 cc. per day respectively to groups I and II. 

Group H. One bird of group I was so ill on Aug. 24th that it was removed 
and housed in the foster mother (not heated). The next day two other birds 
_ from group I and one bird from group I were also put into the foster mother. 
This so-called hospital group H was given the basal diet of oatmeal and 
vitamins supplied to group II, but with the cod liver oil increased to 5 cc.; 
so far no alteration had been made in the quantity of A factor and the illness 
was possibly due to its insufficiency. The increase had no effect; the birds 
got no better and on Aug. 27th the first bird of group I died. Its death was 
diagnosed by Dr W. Taylor as due to chicken cholera. 

Some of the birds now showed distinct signs of leg weakness, drooping 
wings and ruffled feathers. These symptoms have been described in the case 
of birds suffering from polyneuritis gallinarum in its early stages. The cause 
of the illness was thus indicated to be due to insufficiency of B factor. Extra 
marmite was consequently given, but owing to the poor appetite of the birds 
it could not be given in the food; on Sept. Ist, it was therefore supplied in 
the form of a 2 % marmite drink. This was greatly relished and on the third 
day unmistakable signs of improvement in health were observed. Just as 
this improvement was noticed two more birds from group II had to be ad- 
mitted to the hospital group (Sept. 3rd). With the marmite drink they also 
got better in three days. The drink was stopped on the 9th Sept. and as the 
birds were now quite well they were returned to their groups on Sept. 11th. 

Period 3 from Sept. 4th to 10th, 1°”). The birds remaining in groups I and I 
were supplied from Sept. Ist,to 5rd with the same daily diet consisting of 300 g. 
oatmeal, 120 cc. milk, 2-5 ce. cod liver oil, 0-5 g. marmite, 50 cc. lemon juice. 
Since there was no object in continuing the same diet to both groups, a change 
was made on Sept. 4th in the quantity of A factor. Group II was given 
10 cc. Loss of appetite was noti¢ed in both groups and several more birds 
in group II were ill. A drink of lemon juice bad no efiect. The remarkable 
effect of the marmite drink had just been obser ed in group H and it appeared 
clear that too little B factor was being supplied. The quantity of marmite 
_in the basal diet of both groups was therefore doubled on Sept. 6th. Group I 
relished the increase by showing an improved appetite, but group II still 
appeared ill. On Sept. 8th group II was given a drink of 2% marmite; its 
effect was again striking. It thus appeared clear that too little B factor had 


16 R. H. A. PLIMMER AND J. L. ROSEDALE 


been given and that chickens are very susceptible to an insufficiency; it was 
very noticeable that the leg weakness disappeared with the increase. Guinea- 
pigs show a similar behaviour, but their susceptibility is to C factor. The 
daily amount of marmite in the diet was doubled for the last two days of this 
period. 

Period 4 from Sept. 11th to 16th, 1920. On Sept. 11th the birds of the 
hospital group were returned to their respective groups. On account of the 
death each group had now eleven birds. 

The effect of the cod liver oil on group II had had the noticeable effect of 
producing a loss of appetite and signs of ill-health; another period of a similar 
diet was therefore tried. Both groups were given more marmite, 2-5 g. per | 
day and the same amount of lemon juice 30 cc. gradually reduced from 50 ce. 
Oatmeal remained as before. Group I had 5 ce. of cod liver oil; group II | 
12-5cce. The birds of group I (except one which was given 0-5 g. marmite, 
dissolved in water, by hand on Sept. 14th, after which it recovered) appeared 
well and had a good appetite. The birds of group II showed marked loss of 
appetite, and the quantity of marmite was doubled on Sept. 14th; by Sept. 
17th their appetite had again become normal. The extra cod liver oil was 
thus not itself the cause, but rather the extra fat required B factor for its 
assimilation. Braddon and Cooper [1914] have made similar observations in 
cases of beri-beri, and have suggested that the amount of B factor ran parallel 
to the caloric value of the diet. 

During this period a bird of group I swallowed a nail and died. 

Period 5 from Sept. 19th to Nov. 10th, 1920. The above observations 
indicating that the amount of B factor needed bore a constant ratio to the 
amount of carbohydrate and fat in the diet led to a general rearrangement 
of the diet. For every 30 g. of oatmeal a quantity of 0-5 g. of marmite was 
given; cod liver oil and lemon juice were given in a constant quantity per day. 
The only difference in the two groups was the extra cod liver oil for group IT 
and its balancing by 2g. of marmite. The daily food consumption for the 
first part of this period was the following: 


Group I Group It Extra for groups Tand II 

Cod rt Cod gi a 

liver Lemon liver Lemon 

Oatmeal oil Marmite juice ©» | oil Marmite juice Oatmeal Marmite Milk 

g. ce, é. ce. 4 ce. g. ce. g. g. ce. 

Sept. 19-21 180 5 a 30 180 12°5 5 30 120 2 80 
» 22-Oct. 17 180 5 3 30 L180 12'5 5 30 240 qd 160 
Oct. 18-21 180 5 . w 180) s-« 12-5 5 80 360 6 240 


The food was supplied in ty.o portions, and in consequence of the greater 
size of the birds the milk poytion was rapidly increasing above the original 
quantity. It was therefore de. ded to reduce the milk to 100 cc. per day and 


for greater convenience all te items were made into a single mixture: 
Por group I For group Il 


Ood 
liver 
oll 


~ a ™ 


Lemon liver Lemon 
Oates Milk Marmite, juice Oatmeal Milk oll Marmite juice 
“ ce oe, ‘ ce x. 06, oo, &. ce, 


Oct, 22-Nov. 10 600 100 a) 1! 30 660 100 125 18 30 


ao 
1 ¢ 


VITAMIN REQUIREMENTS OF CHICKENS 17 


These quantities were eaten daily by the birds of group II, but on two 
occasions they were not eaten by group I, due to the fact that this group had 
10 birds against 11 in group II. 

There were no signs of illness throughout this period of 54 days; the birds 
grew steadily, increasing on the average 25 g. every two days. Group I in- 
creased from 406 to 1019 or 613 g., group II from 329 to 1O11 or 682 g. It 
would appear that the extra cod liver oil had a stimulating action, but as 
two birds of group II were small at the start and two were cockerels of a 
large strain, it was not a real effect. 

An alteration in the housing was made on Oct. 12th, both groups being 
moved to the roof. The move made no difference to the regular weight increase 
nor to the daily food consumption, so that housing is of minor importance 
provided diet and cleanliness are satisfactory. During this period the cocks 
began to crow, the first one crowing on Oct. 13th. 

Period 6 from Nov. 11th to Jan. 13th, 1921. The quantities of the several 
accessory factors supplied in the previous period were evidently sufficient to 
keep both groups in perfect health. Group III (see Part Il) was receiving a 
diet containing caseinogen, secwa and 0-25 g. marmite per 30 g. oatmeal per 
day, 7.e. half the amount of marmite. The caseinogen and secwa would thus 


_ together contain 0-25 g. of B factor reckoned as marmite. It was possible that 


the caseinogen alone contained B factor equivalent to that in the marmite. 
Caseinogen was therefore substituted for marmite in group I. It was also of 
interest to compare the protein of caseinogen with that contained in marmite. 
The diets thus became: 


For group I For group IT 
Cod 
: liver Lemon liver Lemon 
Oatmeal Milk CiSeinogen oil juice Oatmeal Milk Marmi oil juice 
g. ee. ms ee. ce. g. ec. g. ce. ce. 
Nov. 11l—Dee. 6 660 100 - 13 5 30 660 100 13 5 30 
Dec. 7-24 990 150 19 7-5 45 990 =150 19 7-5 45 


The increase at the later date \-as an extra half ration on account of the 
greater daily consumption; as befoa», growp | with ten birds did not require 
food on two occasions. 

Both groups increased in weighty group | from 1018 to 1489 or 471 g., 
group II from 1045 to 1515 or 4708. per bird. No signs of ill-health were 
noticed. It was necessary to separate the cocks and hens at the last date. 
Two main groups were made: group | a white cock and four white hens, 
group IT a black cock and six black hens. The extra cock birds were kept in 
separate houses. The same diet was continued to the end of the period. The 
hens then began to lay and another rearrangement of groups was necessary. 
The white birds (group IV) were compared with group If (see Part IT) and 
the blacks were divided into two groups (I and [!) consisting of three hens 
and a cock, the extra cock being taken from the © xtra set. 

Period 7 from Jan 14th to Feb. 28th, 1921. Numerous experiments on the 
effect of vitamins on egg laying are conceivable. Since it had been found that 


Bioch. xv1 x 


18 R. H. A. PLIMMER AND J. L. ROSEDALE 


birds are most susceptible to insufficiency of B factor, the effect of extra 
marmite on egg laying was first tried. 

The birds (groups I and II) were given an identical diet except that 
group II was given double the amount of B factor. To compensate for any 
extra protein in the marmite the two groups were supplied with the same 
total quantity, but in the case of group I half was heated in the autoclave 
at 120° C. for one hour to diminish the B factor. The daily diets were: 


Group I Group IT 
oe es 
% Heated Cod liver Lemon Cod liver Lemon 
Oatmeal Milk Marmite Marmite oil juice Oatmeal Milk Marmite oil juice 
g. ce, g. g. ce. ce. g. ce. g. ce. ce. 
360 40 6 6 3-5 15 360 40 12 3°5 15 


No signs of ill-health were noticed in either group and the average weights 
remained constant at about 1609 g. for group I and 1638 g. for group II. 
There was a slight difference in the number of eggs: 


Group I Group IT 
Jan. 14-31 10 weighing 45-5 g. on average 4 weighing 47 g- on average 
Feb. 1-28 -y Fears 51-5 g. ¥ 36 »  obdg. Pe 


There were two double eggs weighing 84 and 73 g. 

The egg laying period was too short to yield a definite result, but if any 
recognition can be given to the actual birds in the groups, those in group I 
were better and started laying earlier than those in group II, and yet more 
eggs were laid by group II. 

The experiment had to be discontinued at this period. 


CONCLUSIONS. 

Chickens can be reared in confinement on a diet of oatmeal and milk if 
the three vitamins are all contained in the diet. 

Chickens are very susceptible to an insufficiency of B factor in the diet. 
The amount of B factor required in the diet seems to run parallel to the 
amount of carbohydrate and fat. It may be represented by 0-5 g. of marmite 
per 30 g. of oatmeal and per 5 cc. of cod livvr oil for 11 birds. 

A daily amount of A factor as conta’aed ‘» eod liver oil of Bec. and a 
daily amount of C factor as present in len on juice of 30 ce. suffice for 11 birds. 

The above figures are not to be regarded as minimal figures. 

It is probable that the necessary “green” food for fowls supplies extra 
B factor rather than the expected A and C factors. 

It would appear that the leg weakness of chicks is due to were of 


B factor in the diet. 
REFERENCES, 


Board of Agriculture and Fisheries (1907). Leaflet No. 114. 
Braddon and Cooper (1914). /. gprs 14, 331. 

Buckner, Nollau and Kastle (1015). Amer. J. Physiol. 39. 
Chick and colleagues (1919). Quote din Med. Riresrah Committee, Report No, 38 on Vitamines. 
Chick and Hume (1917), J. Roy. Army Med, Corps, August. 
Cooper (1013). J. Hygrene, 18, 436, 

Drummond (1916). Biochem. J. 10, 77. 

Hart, Halpin and Steenbock (1920). J. Biol. Chem. 48, 421. 
Osborne and Mendel (1916), J. Biol. Chem. 26, 293. 
Osborne and Mendel (1918), J. Biol. Chem. 38, 433. 
Palmer and Kempster (1919). J. Biol. Chem. 39, 299. 
Plimmer and Rosedale (1922). Biochem, J. 16, 19. 


IV. THE REARING OF CHICKENS ON THE IN- 
TENSIVE SYSTEM. PART II. THE EFFECT 
OF “GOOD” PROTEIN. 


By ROBERT HENRY ADERS PLIMMER & JOHN LEWIS ROSEDALE 
WITH THE ASSISTANCE OF 
ARTHUR CRICHTON & ROBERT BAYNE TOPPING. 


From the Biochemical Department, Rowett Research Institute, University of 
Aberdeen and North of Scotland College of Agriculture. 


(Received December 20th, 1921.) 


OsBorNE and MENDEL [1912,/1914] have shown that the growth of rats 
depends upon the presence of the amino acid, lysine, in the protein of the 
diet and that normal or average growth follows if 2 or 3 °% of this compound 
be present in the protein. They have shown further [1916, 1] that proteins 
vary greatly in their capacity for promoting growth. Lactalbumin was 
superior to caseinogen and caseinogen to edestin; cereal proteins were not of 
“good” quality’. Buckner, Nollau and Kastle [1915] found that a mixed 
grain diet of high lysine content produced more rapid growth than one of 
low lysine content in the case of chicks and this was confirmed by Osborne 
and Mendel [1916, 2]. Buckner, Peter, Wilkins and Hooper [1919] repeated 
the experiments with a larger number of birds kept under better conditions 
at a poultry farm and again noted the effect of a high lysine content. In all 
these experiments there was a considerable mortality amongst the birds, 
probably to be accounted for by an insufficient supply of vitamins. In Part I 
[Plimmer and Rosedale, 1922] it was shown that chickens could be reared 
in confinement if an adequate quantity of each of the three vitamins were 
added to the diet. It seemed that the B factor must be present in proportion 
to the amount of carbohydrate and fat in the diet. 

The present experiment was designed partly to confirm the previous one 
and partly to ascertain the effect of “good” protein upon the rate of growth. 
A mixture of lactalbumin and caseinogen was chosen; both these proteins 
contain a high amount of lysine; the main deficiency is that of cystine in 
caseinogen. Vitamins, additional to those in the commercial foods, were added 
and oatmeal formed the basis of the diet. The effect of the “good” protein 
produced the expected result of rapid growth. The cocks reached the average 
weight of 1828 g. in 122 days; after this time they increased more slowly 
reaching the average weight of 2002 g. in 144 days; they were really suitable 
for marketing at the former age. The hens started laying at the age of 139 days. 

1 A summary on “Quality of Protein in Nutrition” is given by Plimmer [1921]. 
2—2 


20 R. H. A. PLIMMER AND J. L. ROSEDALE 


EXPERIMENTAL. 

The group (No. III) consisted of nine white Leghorn chicks of the same 
strains as those in the previous experiment hatched under hens on Aug. 21st, 
1920. They were housed for the first three weeks in the foster mother and 
then in one of the poultry houses in the ground floor room of the Institute, 
except for a period of 21 days from Sept. 23rd to Oct. 13th, during which time 
they were kept in a house on the roof of the building. This change made no 
difference to their rate of growth, which was rapid throughout the experiment. 

The diet, except for the first five days, was composed of a mixture of 
equal parts of caseinogen! and secwa? or dried whey, mixed with oatmeal. 
This choice was made so as to give some comparison with the milk used 
previously. No allowance was made for vitamins in these foods: B factor 
would be present in the lactalbumin preparation, and both A and B factors 
in the caseinogen. To ensure an adequate supply they were added in the form 
of cod liver oil, marmite and lemon or orange juice. The cod liver oil and 
lemon juice were given, except for a short time at the beginning and again 
later, in constant daily amounts, but the marmite was increased with the 
increase of oatmeal required as the birds grew bigger. It was always added 
in the proportion of 0-25 g. to 30 g. of oatmeal. The food was prepared daily 
in the same way: the required amounts of oatmeal, caseinogen and secwa were 
weighed out; marmite was measured from a freshly prepared 25 % solution 
and mixed with the lemon or orange juice and cod liver oil; some extra water 
was added and the three ingredients shaken together and well stirred into the 
food mixture. A wet mash was thus the only food provided; dry food only 
consisted of the dried surface of the mixture. No extra salts were given and 
no grit, but on one day the birds were given by mistake some broken pieces 
of porcelain. The examination of the gizzards of the surplus cocks, after they 
were killed, revealed the presence of small black stones; they were presumably 
picked up from the peat moss litter used to cover the floor of the cage. 

The quantity of food eaten, protein ratio, average weight and average 
daily increase are shown in the following table: 


Cod Average 
No. of Casein- liver Lemon Protein Average daily 
days Oatmeal ogen Secwa oil Marmite juice ratio weight Increase 
&. g. &. co. g. ce, &. g 
Aug. 21-25 5 30 0 30 25 0°25 16: 248 39-69 6 
» 26-29 4 30 15 15 25 0:25 16 «41:25 69-80 3 
»  380-Sept. 3 5 60 =30 30 50 05 30 «1:25 80-111 6 
Sept. 4-11 8 90 45 45 75 0-75 45 1:25 111-181 9 
» 12-21 10 120 45 45 75 10 45 , 1:25. 181-299 12 
» 22-30 9 180 = 60 60 75 = 15 45 1:22 299-322 3 
Oct. 1-5 6 210 60 60 75 1:75 45 1:25 322-443 20 
ms 6-15 10 240 60 60 75 20 45 1:24 443-714 © 27 
» 16-26 11 30060 60 75 25 45 2:3 714-904 17 
» 27-Nov. 18 23 180 36 36 75 15 45 1:3 904-1233 14 
Nov. 19—Dee. 6 18 240 = 48 48 75 20 68" O38 1233-1445 12 
Dec, 7-26 20 360 72 72 10) 30 67 1:3 1445-1668 11 
» 2i-Jan. 13 18 360 36 36 75 30 45 1:4 1668-1767) 65 


The diminution in the daily food supply on Oct. 27th was due to the fact 
that the group consisted of five cocks and four hens and that it was necessary 


* Caseinogen: moisture 9-9, ash 6-5, fat 0-8, protein 82-7 %,. 
* Seowa: moisture 1-2, ash 9-8, fat 0-3, lactose 745, sol. lactalbumin 14-2 %,. 


Ss 


REARING OF CHICKENS 21 


to separate off the surplus four cocks (groups III A and B, below). The 
figures from this date thus refer to one cock and four hens. 

The birds in three weeks reached the weight of the birds in groups I and II 
[Part I] at the age of six weeks. The cocks in this group began to crow about 
Oct. 9th at the age of 49 days. The hens began to lay at the age of 139 days 
on Jan. 7th, 1921. It is recorded that Leghorns have started to lay at 
122 days, but these birds were hatched in the spring and had a free existence. 

Throughout the period the health of the birds was excellent and they 
moulted rapidly during October. As might be expected from the work of 
Palmer and Kempster [1919], the legs and beaks were not pigmented. These 
investigators showed that the yellow pigmentation depended on the presence 
of yellow pigments in the food. The food here supplied was almost colourless 
and the result is a confirmation of their work. 

Groups III A and III B. The four cocks after removal from the main 


_ group were kept in four separate small cages and were arranged in two groups _ 


of two birds as regards food. Both groups were supplied with the same food 
items as the main group, but with a higher proportion of caseinogen and 
secwa; this proportion was reduced on Dec. 3rd, so that the protein at first 
1 : 1:8 became 1 : 1-3. They were also given more cod liver oil and orange 
juice per bird than the main group, group III B more than HII A. The 
marmite was always kept at 0-25 g. per 30g. of oatmeal. The food con- 
sumption, average weight and daily average increase were as follows: 


Group III A, each bird. 


Cod Average 
No. of Casein- liver Lemon Protein Average daily 
days Oatmeal ogen Secwa oil Marmite juice ratio weight Increase 
g. g. g. ce. g. ce, g. g. 


Oct. 27—Dec. 3 39 60 30 30 2-5 0-5 15 1:18 1053-1730 17 
Dec. 4-Jan. 11 39 60 15 30 _ 25 0-5 15. 1:29 1730-2202 15 


Group III B, each bird. 


Oct 27—Dec. 3 39 60 30 30.50 0-5 30 1:18 1015-1587 15 
Dec. 4-Jan. 11 39 60 15 30. 5-0 0-5 30 1:29 1587-1804 6 


Sometimes the cocks would eat an extra daily allowance during the course 
of one or more days. 

The housing in the smaller cages made no appreciable difference to the 
daily weight increase, though it was slightly higher than that of the main 
group. In the first part of the period the greater amount of cod liver oil given 
to group III B had a lessening rather than the augmenting effect which might 
have been expected. This effect was very marked in the second part. It would 
support the suggestion made in Part I that the amount of B factor required 
in the food bears a definite ratio to the consumption of carbohydrate and fat. 
Cod liver oil was. not given for the last few days as the birds were to be killed 
on Jan. 11th for the purpose of examining the flesh. The flesh of the breast 
was pale, that of the legs about the colour usual with Leghorns. The flesh 
was also lacking in flavour, evidently due to the “tasteless” food. 

Group III, egg laying period. The hens began to lay on Jan. 7th, 1921; 
seven eggs weighing 286-5 g. or 40-9 g. per egg were produced up to Jan. 13th. 


22 R. H. A. PLIMMER AND J. L. ROSEDALE 


On Jan. 14th a comparison of egg laying on this diet was made with the 
same number of hens and a cock (group IV) from the birds of groups I and II. 
Their diet contained blood meal as protein in exactly the same protein ratio 
of 1 : 4. Group IV was given less lemon juice, and marmite was omitted from 
their diet, since it was possible that sufficient of this B vitamin was present 
in the blood meal. The following tables show the diets, weights etc.: 


Group III. 
Cod liver Lemon Average 
Oatmeal Caseinogen Secwa oil Marmite juice weight Increase 
g. g- g. ee. g. ce. g. g- 
360 36 36 7-5 3-0 45 1767-1815 48 
Group IV. 
Cod liver Lemon Average 
Oatmeal Blood meal oil juice weight Decrease 
g. g. ce. ce. g. g. 
360 87 7-5 45 1537-1456 81 


A hen of group III suddenly died on Feb. 4th apparently from choking. 
No signs of ill-health were noticed beforehand nor could any reason be found 
at post mortem examination. The other birds were perfectly well throughout. 

Group IV did not relish the blood meal; on several occasions it was not 
necessary to prepare the daily food for them. 

Grit was not supplied to either group; the eggs had always a sound shell. 


Group III gave 6 eggs during January weighing 253-5 g. or 42-5 g. per egg. 
PSY eae ee 5 6) = February ms 702-5 ¢. , 413g. 5, 

OBS | eT | Ps January an 316-5 ¢. ,, 452g. ,, 

” ” 3 ” February ” 134-4 &- »» 44-8g. 4, 


The very great difference in the number of eggs from the two groups may 
have been due to the blood meal in the food of group IV, but from the results 
recorded in Part I, it is more probably due to absence of B factor. Un- 
fortunately it was not possible to test this supposition, as the experiment 
had to be discontinued. 


CONCLUSIONS. 


On a diet containing sufficient of each of the three vitamins, chickens can 
be raised in confinement. This confirms the results of Part I. 

Rapid growth results on a diet having an albuminoid ratio of 1 : 3 and 
containing “good” proteins, caseinogen and lactalbumin. Cock birds began 
to crow in 49 days and hens to lay in 139 days; the cocks had an average 
weight of 1853 g. and hens of 1815 g. 


REFERENCES. 


Buckner, Nollau and Kastle (1915). Amer. J. Physiol. 39, 162. 

Buckner, Peter, Wilkins and Hooper (1919), Kentucky Agric. Rap, Stat, Bull. No, 220, 
Osborne and Mendel (1912). J. Biol, Chem. 12, 473. 

Oaborne and Mendel (1914). J. Biol. Chem. 17, 325. 

Osborne and Mendel (1916, 1). J. Biol. Chem, 26, 1. 

Osborne and Mendel (1916, 2), J. Biol. Chem, 26, 293, 

Palmer and Kempster (1919), J. Biol, Chem. 39, 299, 313, 331. 

Plimmer (1921). Proce, Roy. Inat.; Nature, 107, 664; J. Soc. Chem. Ind. 40, 227. 
Plimmer and Rosedale (1922), Biochem, J. 16, 11. 


V. DISTRIBUTION OF ENZYMES IN THE ALI- 
MENTARY CANAL OF THE CHICKEN. 


By ROBERT HENRY ADERS PLIMMER 
anp JOHN LEWIS ROSEDALE. 


From the Biochemical Department, Rowett Research Institute for Animal 
Nutrition, University of Aberdeen and North of Scotland College of Agri- 
culture. 

(Received December 20th, 1921.) 


THE presence of lactase in the intestines of animals and the non-adaptation 
of the pancreas and intestine to lactase by feeding with lactose was investi- 
gated by Plimmer [1906]. Lactase was always found to be absent from the 
intestine of chickens. A diet containing lactose had been used by us [1921] 
in feeding chickens from birth for a period of over three months. Examination 
of the birds’ excreta showed that reducing sugar was absent therefrom, a 
fact which indicated that the sugar was assimilated. Assimilation of disac- 
charides is usually preceded by hydrolysis to monosaccharides, which would 
imply the presence of lactase in the alimentary canal, either in the intestine 
by adaptation or in some other part. The intestines of the cockerels in this 
group of birds were therefore examined, after they were killed, for the presence 
of lactase: it was not found to be present, and the non-adaptation of this organ 
was verified. If hydrolysis of lactose previous to assimilation occur, it must 
take place in some other part of the gut. The crop, pancreas and proventri- 
culus were tested and lactase in small amount was detected in the crop. The 
investigation was then extended to the presence of other enzymes, as no 
information could be found in the literature about their occurrence in the 
alimentary canal of birds. The enquiry did not extend to the detection of all 
known enzymes, but was limited to those concerned in the digestion of the 
common foodstuffs. 


EXPERIMENTAL. 


The methods of preparing the enzyme solutions and detecting the presence 
of enzymes were in general in accordance with those usually adopted; in many 
cases a longer time of action (up to seven or ten days) was allowed, and in the 
case of the sucroclastic enzymes, proteins etc. were removed before testing 
for the reducing sugar formed by their action. 

The various parts of the alimentary canal were always taken foaun chickens 
killed the same day, or not later than the day previously; on account of the 


24 R. H. A. PLIMMER AND J. L. ROSEDALE 


small size of the crop, proventriculus and pancreas, the organs from four to 
eight birds were collected and examined together. A single small intestine 
provided sufficient material, but in most experiments several were combined 
as the whole series of sucro- or proteo-clastic enzymes were tested for simul- 
taneously. Separate tests were made for lactase. At least two experiments 
were made with each part, except the caeca. 


Preparation of enzyme solutions. 


The pancreas, on removal, was cut up into small pieces and ground with 
sand in a mortar; the ground mass was put into glycerol in which it was kept 
for several days in the presence of a few cc. of toluene. The solution was then 
prepared by diluting with rather more than an equal volume of water and 
filtering from sand, etc. 

The other parts of the alimentary canal were cut open and washed with 
running water to remove the contents. The mucous membrane was scraped 
off, ground up with sand and water and extracted for 24—48 hours with water 
in the presence of a little toluene to prevent putrefaction. The aqueous 
portion was strained off through cloth to remove sand and larger pieces and 
used for testing for enzymes. 

It was not possible to scrape off mucous membrane from the inside of the 
proventriculus. The organ is glandular, covered with numerous small teats, 
which, on pressing with a scalpel, emit a yellowish, viscous, distinctly acid 
secretion. This secretion was the material actually used after grinding with 
sand and mixing with water. Nothing could be scraped off the gizzard, the 
interior surface of which resembled parchment. 

Detection of enzymes. 

(a) Diastase and invertase. As substrates 100 cc. of 1% starch solution 
and 50 ec. of 3 % cane sugar solution were used. Two portions were measured 
out with a pipette in separate flasks; a known volume of enzyme solution was 
added to one, and the same volume of boiled enzyme solution, after cooling, 
to the other; 2 or 3 cc. of toluene were added to each, the flasks corked and 
put into an incubator at 37° for one or more days. A test for starch by the 
iodine reaction was made from time to time with a drop removed from the 
mixture. At the end of the reaction time, the mixtures were washed into a 
250 cc. measuring flask, a slight excess of colloidal ferric hydroxide added, 
any excess of the latter removed by a few crystals of magnesium sulphate, 
the volumes made up to the mark, the solutions filtered and reducing sugar 
tested for by the complete reduction of 10 cc. of Fehling’s solution. The 
control solutions*containing boiled enzyme did not reduce, or only gave a 
slight reduction due to sugar present in the extract. 

(b) Lactase. The detection of lactase was carried out in a similar way to 
that of diastase and invertase, using 50 ec. of 4% lactose solution as substrate, 
The enzyme and control mixtures were put directly into 250 ec, measuring flasks 


DIGESTIVE ENZYMES OF THE CHICKEN 25 


and made up to volume after clearing with colloidal ferric hydroxide and 
magnesium sulphate. The reducing sugar was estimated by the reduction of 
-10cc. of Fehling’s solution. The observed difference in reading indicated 
whether hydrolysis had or had not occurred. No difference in reading was 
observed in the case of the intestine or proventriculus, but a small though 
distinct difference was always noticed in the case of the crop extract; it varied 
from 0-2 to 0-5 cc. in a total of 10 or 10-1 cc. This slight difference indicated 
an hydrolysis of 10-20 % of the lactose. 

(c) Lipase. This enzyme was not looked for except in the case of the 
pancreas. Two exactly equal portions of oil in separate test tubes were made ~ 
just alkaline to phenolphthalein with 0-1 N caustic soda. Enzyme and boiled 
enzyme solution were added. On keeping at 37° and occasionally shaking, 
the pink colour of the tube containing enzyme solution disappeared and it 
was restored by adding a few drops of the soda. This could be repeated several 
times and altogether from 1-2 cc. of alkali were added; the control tube did 
not change colour. 

(d) Proteoclastic enzymes. Proteoclastic enzymes were detected by their 
action on Congo-red fibrin in neutral, acid and alkaline media. In the first 
case, a definite volume of enzyme solution and the same volume of boiled 
enzyme solution were put into separate flasks; in the other cases the same 
volumes of enzyme and boiled enzyme solutions were mixed with an equal 
volume of 0-2 N hydrochloric acid or 0-2 N sodium carbonate solution in 
separate flasks; 1 g. of Congo-red fibrin and 2 cc. of toluene were added to 
each and the several flasks were put in an incubator at 37° for one to seven 
days. Solution of Congo-red fibrin, which, in the case of hydrolysis, generally 
occurred in one or two days, was taken as indication of the presence of proteo- 
clastic enzyme; solution did not occur in those flasks with boiled enzyme 
solution. No investigation was made of the products of the hydrolytic action. 


RESULTS. 


The presence or absence of enzymes in the various parts of the alimentary 
canal is most easily seen from the following table: 


Proven- Intestine Duo- 
Crop triculus’ Pancreas whole denum TIeum _  Caeca 
Invertase 0 0 : + : - 0 
Diastase + 0 + + : ; + 
Lactase + 0 0 : 
Lipase - : + ; , ‘ > 
Proteoclastic in neutral 0 0 +slight 0 0 0 0 
o” acid +slight + +less rapid + + + 0 
* alkaline 0 0 +rapid +slight + +slight 0 


media 


The distribution of the sucroclastic enzymes corresponds in most parti- 
culars with that in the animal; most animals have invertase in the intestine, 
lactase is present in some, absent in others: diastase and lipase are generally 


26 R. H. A. PLIMMER AND J. L. ROSEDALE - 


present in the pancreas of animals. The proteoclastic enzymes show a differ- 
ence: the animal has trypsin acting in alkaline media; the chicken in both 
alkaline and acid media. The intestine of the chicken has an enzyme acting most 
rapidly in acid medium, less rapidly in alkali. The proteoclastic enzyme of 
the proventriculus acts only in acid medium; the organ corresponds to the 
stomach of animals. The caeca, as expected, had no enzyme of this group, 
but contained diastase. 


We wish tofthank Prof. J. A. MacWilliam, F.R.S., for kindly allowing us 
to carry out these experiments in his laboratory. 


REFERENCES. 


Plimmer (1906). J. Physiol. 34, 93; 35, 20. 
Plimmer and Rosedale (1921). J. Agric. Sci. 


VI. THE AMINO-ACIDS OF FLESH. 


THE DI-AMINO-ACID CONTENT OF RABBIT, CHICKEN, 
OX, HORSE, SHEEP AND PIG MUSCLE. 


By JOHN LEWIS ROSEDALE. 


From the Biochemical Department, Rowett Research Institute for Animal 
Nutrition, University of Aberdeen and North of Scotland College of Agri- 


culture. 
(Received December 20th, 1921.) 


A Lone series of food analyses has recently been made by Plimmer [1921, 1], 
who points out that by the ordinary routine method of analysis, in which 
the amount of protein is estimated by multiplying the nitrogen content by 
6-25, no discrimination is made between the flesh of different animals. The 
protein of one animal is regarded as being the same as that of another. The 
work of Emil Fischer and Kossel and their pupils has definitely proved that 
the various proteins differ very widely in their composition as regards the 
amino-acids, and this difference is emphasised by the experiments on the food 
value of the individual amino-acids by Hopkins in conjunction with Willcock 
and Ackroyd, by Osborne and Mendel and other American investigators!. 
These chemical and biological differences are sufficient evidence that quality 
of protein in nutrition must be taken into consideration. 

Complete analyses of the protein of the muscle of the ox, chicken, halibut 
and scallop have been made by Osborne and Heyl [1908] and Osborne and 
Jones [1909], and Drummond [1916] has made some analyses of muscular 
tissue by Van Slyke’s method. Both the more complete analyses by Osborne 
and co-workers and those by Drummond do not show any marked difference 
in the amino-acid content of the various muscle proteins. The flesh of various 
animals shows such distinct appearances, different both to the eye and palate, 
that it seems probable that greater differences may exist, and that there may 
be smaller differences in the flesh from various parts of the same animal’s 
carcase, such as back and leg. Some further amino-acid analyses have there- 
fore been made. 

The methods of protein analysis are far from perfect: Fischer’s ester 
method for the mono-amino-acids, as he pointed out, is not quantitative: 
Kossel and Patten’s method for the di-amino-acids, in spite of the numerous 
manipulations, is generally considered to be fairly accurate, but it has been 
largely superseded by Van Slyke’s method which gives higher values for these 
amino-acids. Van Slyke’s method also possesses the advantage of requiring 
only small amounts of protein and is more rapidly carried out. This method 
of protein analysis has been used in these experiments, since it was chiefly 

1 See summary by Plimmer [1921, 2]. 


28 J. L. ROSEDALE 


desired to compare the muscle protein of several animals with a view to more 
complete data at a later time. A comparison of these results with those by 
Kossel’s method has been made in a few cases. The results indicate that 
differences exist in the amino-acid content of the various muscle proteins. 
Duplicate analyses were always carried out; frequently these analyses were 
not so concordant as was expected. This inconsistency of the results was 
under investigation by Plimmer [1916] who tested the arginine determination; 
other details of the method are now being studied. 


EXPERIMENTAL. 


In the case of the smaller animals (rabbit, chicken) opportunity was taken 
of comparing the flesh of different parts of the body of the same animal. In 
other cases the flesh was taken from the thigh. The mode of operation was 
the same throughout. The flesh (about 350g.) was freed from inside fat, 
minced and put into about 2 litres of boiling water containing 0-1 % acetic 
acid and heated for about ten minutes so as to coagulate the protein and 
remove the extractives. The liquid was poured off and the coagulated protein 
squeezed dry in a cloth. This procedure was repeated twice. The coagulated 
protein (about 200g.) was then digested with 1g. pepsin in 2 litres of 
0-1N HCl, so as to separate nucleins, indigestible matter, etc. After digestion, 
which usually occupied about ten days at 37° the liquid was filtered off and 
the total nitrogen estimated. A portion containing about 6 g. of protein was 
then hydrolysed by boiling with hydrochloric acid added to the liquid so as 
to make a concentration of 20 %. The hydrolysis was carried on for 36 hours. 
The hydrolysed solution was evaporated to dryness in vacuo, made up to 
250 cc. and two samples of 100 cc. were analysed by Van Slyke’s method. 
This was performed as described except for the arginine estimation which 
was effected by Plimmer’s modification [1916]. In the earlier experiments it 
was impossible to make determinations of amide N owing to the facilities for 
vacuum distillation not being adequate. The analyses were made in duplicate 
and the percentage has been calculated from the average. 


Table I. Nitrogen percentages. 


Di-amino-acids Mono-amino-acids 
A — ‘rotal N 


ms Non- Argi- Histi- ci Non- of hydro- 
Humin Total Amino amino nine dine Lysine Total Amino amino lysed 
N N N N N N N N N N 


Amide solution 

Rabbit, back a — 457 215 240 15 19 115 49 — -- 94-7 
wi fore limb oe — 441 #179 27-7 88 30-9 65 507 — — 94-8 

a hind limb — — 448 184 266 13 25 58 563  — — 1011 
Chicken, breast 69 3 27 9 18 10 13 2 61-1 49-7 11-4 98-0 
wa legs as) 13 265 16 10-5 8 7 ll 68:5 66-6 19 1008 
Beef 63 056 285 15 13-5 «13:3 5 11-2 55 26-8 28-2 90:3 
Horse 2-9 09 37-1 #%I188 183 149 105 116 £70 58 119 110-9 
Mutton 65 06 383 223 156 15 18 43 54 52 2 993 
Pork 6-4 12 282 133 15 14 7 7 57 53 4 92:8 


Relatively little difference can be observed from the figures for the different 
meats. The amide N is almost similar, in each case averaging about 6-0 % 
of the total N, 


AMINO-ACIDS OF FLESH 29 


Table II. Percentages of amino-acids. Giving the amount of 
amino-acids in 100 g. of protein. 
Arginine Histidine Lysine Total di-amino N 
Rabbit, back 


8 10 13 31 

» fore limb 5 19 5 29 
hind limb 7 15 5 27 
Chicken, breast 6 8 1 15 
» legs q 4 10 18 

f 7 3 10 20 
Horse fj 6 9 22 
Mutton 7 ll 4 22 
Pork v} 4 6 17 


Humin N shows a difference. It is, if anything, higher in the white meats, 
e.g. breast of chicken 3%, legs 1-3 °%, pork 1-2 %, than in the red meats, 
where the average is 0-5 %, except in the horse, where 0-9 °% was found. The 
explanation of this slightly higher value may be that the animal was not 
properly bled on slaughter. 

Lysine figures are, with the exception of mutton, higher for the red meats, 
averaging about 11 %, while, of the white meats, rabbit limbs show only 5-5 %, 
chicken breast 2 °% and pork 7 %. 

Gortner and Holm [1920, 1], working with mixtures of pure amino-acids, 
have shown that tryptophan, and in the presence of aldehyde also tyrosine, 
and their analogues are the only known amino-acids which go to form humin. 
There is therefore no connection between the humin content and the lysine 
content of the meats; this is exemplified especially in the chicken, where the 
humin is high, and the lysine is low in the breast; and humin is low and 
lysine high in the legs. It may perhaps be mentioned that in the preparation 
of the di-amino-acids by the method of Kossel and Patten a distinct yellowish 
colouring adheres to the lysine portion. 

At the same time too much reliance must not be placed on the humin 
as an estimation of tryptophan and tyrosine. Gortner and Holm [1920, 2] 
and Thomas [1921] have shown that tyrosine and tryptophan which go to 
form humin are not necessarily the only substances giving a reaction with the 
phenol reagent of Folin and Denis. Estimations of substances giving the blue 
colour with this reagent were made during the progress of this work, both 
before the removal of the humin and afterwards. In the case of chicken 
breast, a white meat, the readings before removal of the humin represented 
4% “tyrosine” whereas after its removal the readings represented 3-5 %. 
In the case of beef however—a red meat—the difference was greater, the 
former reading being 3-5 °% and the latter 2-1 %, yet the humin N was much 
lower in the case of beef. 

The arginine figures are more constant at about 14 or 15 % except in 
rabbit fore limb and chicken legs where the average is 8 %. 

The histidine figures are less satisfactory, and exhibit perhaps a weak 
point in the method. In this connection it is of interest to point out that in 
the cases of abnormally high histidine the figures for the non-amino N are 
lower than normal and vice versa, e.g. beef 5 % histidine, 28 % non-amino N, 


30 _ J. L. ROSEDALE 


mutton 18 % histidine, 2 °% non-amino N, while in other cases this observa- 
tion cannot be made. This may be due, either to incomplete precipitation of 
the histidine by the phosphotungstic acid, or to washing. Work in this con- 
nection is in progress. 

It is not possible to draw any conclusions from the figures of the mono- 
amino fraction, which account for about 55 to 60 % of the total N. 

The average percentage of the di-amino N is 35. 

Comparison with former work on the hydrolysis of meat is difficult, because, 
with the exception of Drummond [1916] on chicken meat, the other figures 
relate to the method of Kossel, which generally gives lower results than the 
Van Slyke method. 

The above figures for chicken breast agree in the main with those of 
Drummond, his total hexone bases N 27-26 being the same as that above. 
The arginine figures are within 1 % and he records having used the same 
modification of that process as mentioned above. The figures for histidine 
and lysine are discordant, Drummond finding 8-45 and 9-81 respectively, 
while the total N of the mono-amino fraction is 4% Sie ase than that found 
by Drummond. 

In order to compare the figures of Osborne and co-workers with the above 
it is necessary to refer to the percentages not of total N but of actual arginine, 
histidine and lysine. The figures for arginine are generally constant within 
1 %, those for histidine are higher than Osborne’s, while the lysine figures, 
owing to the calculation in Van Slyke’s method, are dependent on the histidine 
values. Apart from the arginine values, only the beef of the present sets has 
given results comparable with those of Osborne, who found 7-5 % arginine, 
1-8 %, histidine and 7-6 °% lysine against 6-8 °% arginine, 2-6 % histidine and 
9-6 % lysine in this experiment. 


SuMMARY. 


1. Determinations have been made of the di-amino-acids of the protein of 
the flesh muscle of rabbit, chicken, ox, horse, sheep, pig by Van Slyke’s method. 
2. The red meats show a higher lysine content than the white meats. 


I wish to take this opportunity of expressing my gratitude to Dr Plimmer, 
who suggested this work, for his kindness and guidance throughout the time 
I was under him, and also to Professor J. A. MacWilliam, F.R.S., for so kindly 
placing his laboratory at my disposal. 


REFERENCES. 
Drummond (1916), Biochem. J. 10, 473. 
Gortner-Holm (1920, 1). J. Amer. Chem. Soc. 42, 821. 
Gortner-Holm (1920, 2). J. Amer. Chem. Soc, 42, 1682. 
Osborne and Heyl (1908), J, Biol. Chem, 22, 433; 23, 81. 
Osborne and Jones (1909). J. Biol. Chem, 24, 161, 437. 
Plimmer (1916). Biochem, J. 10, 115. 
Plimmer (1921, 1). Analyses and Energy Values of Foods (Stationery Office), 
Plimmer (1921, 2). Proc. Roy. Inat.; Wekits, 107, 664; J. Soc. Chem. Ind. 40, 227. 
Thomas (1921). Bul. Soc, Chim. Biol. 8, 197. 


VII. XYLENOL BLUE 


AND ITS PROPOSED USE AS A NEW AND IMPROVED 
INDICATOR IN CHEMICAL AND BIOCHEMICAL WORK. 


By ABRAHAM COHEN. 
From the Cooper Laboratory, Watford. 


(Received December 24th, 1921.) 


Durine the course of some studies on the most important of the sulphone- 
phthaleins synthesised by W. M. Clark and Lubs [1917], the writer was led 
to synthesise a new sulphonephthalein which has practically the same pro- 
perties as thymol blue but with the important advantage of being twice as 
intense and therefore more economical. 

Lubs and Acree [1917] in their earlier quests for evidence on the quinone- 
phenolate theory, now practically firmly established, noticed various rela- 
tionships between the chemical constitution of the sulphonephthaleins and 
the corresponding working ranges of utility. Thus the introduction of negative 
(bromine) groups shifts the working ranges to a more acid region, e.g. cresol 
red with a working range from P,, 7-2 to P;, 8-8 on bromination yields brom 
cresol purple with a working range from P,, 5-2 to P,, 6-8, a shift of two units 
for either limit. The same shift occurs on brominating thymol blue to produce 
brom thymol blue. On brominating phenol red to produce brom phenol blue, 
a tetra-brominated instead of a dibrominated sulphonephthalein results, and 
the shift is four units for either limit of the working range. 

Whilst then it would appear that negative group-substituents induce fairly 
regular changes of P,, ranges, another interesting aspect of the matter is 
presented by the shifting of the ranges towards the more alkaline region, by 
the introduction into the phenol residue of phenol red of positive groups like 
methyl and iso-propyl as in cresol red and thymol blue, which is a 1-methyl- 
4-iso-propyl-3-phenolsulphonephthalein. Here, however, the well defined acid 
range of thymol blue complicates matters. 

It occurred to me to compare the case of diethyl red and dipropyl red. 
These are azo indicators and homologues of the well-known methyl red, or 
more strictly dimethyl red, dimethylaminoazobenzene-o-carboxylic acid. 
The two homologues show the same P,, range, viz. P;, 4-5-6-5, in spite of 
the increased positive character of the propyl groups. The following show 
similar anomalous behaviour. : 


32 A. COHEN 


Table I. 


. Methyl Orange: Sodium p-dimethylaminoazobenzenesulphonate \ Py F144 

. Mono-ethyl Orange: Sodium p-monoethylaminoazobenzenesulphonate 

. Sérensen’s: a-naphthylaminoazobenzene | Px 3-7-5-0 
a-naphthylaminoazo-p-toluene. | 

4. Sodium p-benzylanilineazobenzenesulphonate \ Px 1-9-3-3 

5. Sodium p-aminoazobenzenesulphonate J 


One would suspect therefore the possibility of a similar anomalous un- 
shifted P,, range when the thymol residue in thymol blue is replaced by 
p-xylenol; the difference here is in the replacement of the iso-propyl group 
in thymol by another methyl radical. 

In view of this, it occurred to me, that owing to the smaller molecular 
weight, the resulting new sulphonephthalein should be used like phenol and 
cresol reds in 0-02 °% solution to produce the same intensity as 0-04 9% thymol 
blue or in strength between the two percentages. We should then have a 
more symmetrical list in which only the brominated dyes are used in the 
higher concentration of 0-04 % (or 1-2 % concentrated), alcoholic or aqueous 
monosodium salt, whenever phenol and cresol reds are chosen in half these 
strengths. These relative concentrations among the sulphonephthaleins have 
been recommended largely on the basis of equality of depth of colour, as far 
as can be judged by the eye with the reds and blues, though the reds and 
blues do not appear of the same quality, owing mainly to difference in degree 
of dichroism. 

By condensing the chloride or anhydride of o-sulphobenzoic acid with 
p-xylenol, a product was ultimately obtained having the same double Pj, 
ranges as thymol blue and similar colour changes. . 


a 


EXPERIMENTAL. 


The method of preparing 1-4-dimethyl-5-hydroxybenzenesulphonephtha- 
lein or xylenol blue is as follows [Chemicals & Bye-products, Ltd., and 
A. Cohen, 1921}: . 

A mixture of ten parts of o-sulphobenzoic dichloride, or of the acid anhy- 
dride; ten parts of fused zine chloride; and fifteen parts p-xylenol, M.p. 74:5°, 
p.P. 211-5°, is heated in a brine bath for six hours. The melt is heated with 
forty parts water till disintegrated. It is then filtered hot, washed with hot 
water, and afterwards with a little alcohol. Next, it is dissolved in excess 
caustic soda and precipitated with hydrochloric acid whilst being stirred. The 
mixture is filtered and crystallised from alcohol in the form of a brown solid. 

The reactions are: 

80,C1 (1) (1) OH 


CE + 2CH CHS 
, Ga ine en Ne eee 
(4) ee 


Sa YD Se Vinh f . 


XYLENOL |. 33 


Samples of the new indicator were titrated against 0-01 N NaOH bang 


a micro-burette up to 5 cc. and reading to 0-01 ce. 


‘Sample 1. 0-0205 g. in 150 cc. water (yellowish orange) required 4-90 cc. 
instead of the theoretical 5-00 cc. or 98 °% molecular equivalent alkali. 

Two other samples gave identical results. 

Carbon dioxide must of course be rigidly excluded in such titrations in- 
volving high dilutions. 

As in the cases of phenol and cresol reds, a 0-6 %% aqueous solution of the 
monosodium salt can be made by gently boiling, not merely warming, for 
3 minutes, 0-6 g. of the solid powdered dye in about 80 cc. water containing 
1-47 cc. of normal alkali, and then diluting when cold to 100 cc. 

If preferred, a strong alcoholic solution can be made by boiling 0-2 g. of 
the solid in 100 cc. absolute alcohol. 

A 0-02 % solution was prepared by a thirty-fold dilution of the 0-6 % 
aqueous monosodium salt solution, and added to cordite tubes (or test tubes 
of hard white glass, unflanged and of uniform diameter) each containing 
10 cc. of a standard buffer solution. Using a concentration of 0-05 cc. or 
1 drop indicator per cc. buffer solution, the double series P, 1-2-2-8 and 
P,, 8-0-9-6 (each in consecutive steps of 0-2 P,,) with xylenol blue, were 
similar to those with thymol blue added 4s a 0-04 % solution in the same 
concentration of 0-05 cc. per cc. The tints were, in fact, rather more distinct 
with xylenol blue in the alkaline range, appearing a purer blue in the day- 
light. Further, unlike dipropyl red in the anomalous case mentioned above, 
xylenol blue does not tend to precipitate out from buffer solutions even after 
three months’ standing. On the contrary, xylenol blue stands in this respect 
to thymol blue, as the preferred diethyl red stands to dipropyl red. 

Like thymol blue, it can be used for the estimation of the P,, of gastric 
juice and ammoniacal fluids; and in the differential acidimetric and alkali- 
metric titrations as described by A. B. Clark and Lubs [1918]. 

In Table II some results are given, obtained with mixtures of hydro- 
chloric and benzoic acids, A 5 cc. micro-burette graduated in hundredths of 
a cubic centimetre was used, and I have found it convenient to use a special 
easily constructed titration flask in which a cordite tube is fused horizontally 
to the side of the flask to facilitate colorimetric comparison with the standard 
coloured benzoic acid. The cordite tube attachment also serves to stir the 
liquid when the flask is tilted. Such a flask can be heated without any cracks 
developing in the joint, provided the heating is done at the side remote from 
the joint. I wish here to thank one of my colleagues Mr G. H. Wallis for 
his kindness in constructing this flask. 

0:5 cc. of 0-02 % aqueous xylenol blue was added to 50 cc. of an unknown 
benzoic acid solution, and 10 cc. of this standard kept in a cordite tube. The 
0-5 ce. indicator is arbitrary and others may find it better to use more or less, 
but the less indicator is used the better; the differentiating power of the eye 
being an inadequately known function of subjective colour differences. 


Bioch, xv1 3 


34 A. COHEN 


The mixtures titrated had, of course, the same concentration of indicator 
as the standard. 


Table Il. Titration of mixtures of hydrochloric and benzoic acids. 


0-993 VN NaOH G. of benzoic 
Benzoic acid 1N HCl r te + N HClfound acid per ce. 
ec. used ec. used to standard _to blue ce, _ found 
1 10 nil nil 0:39 nil 0-:0047 
2 10 3-03 3-06 3-46 3-04 0-0049 
3 25 4-04 4-05 5-00 4-02 _ 0-0046 


Experiments are in progress on the conditions for brominating xylenol blue. 


SUMMARY. 


p-Xylenolsulphonephthalein or xylenol blue, a new indicator having two 
working ranges of utility from P,, 1:2 (red) to 2-8 (yellow) and from P,, 8-0 
(yellow) to 9-6 (blue), can be successfully employed in all work for which 
thymol blue has heretofore been recommended. 

The fact, that p-xylenol is easily prepared from diazotised p-xylidene and 
that only half as much xylenol blue as thymol blue is required, should render 
its use eminently preferable. 


REFERENCES. 


Clark, W. M. and Lubs (1917). J. Bact. 21, 109, 137. 

Clark, A. B. and Lubs, H. A. (1918). J. Amer. Chem. Soc. 40, 1443. 

Lubs and Acree (1917). J. Amer. Chem. Soc. 88, 2772. 

Chemicals & Bye-products, Ltd. and A. Cohen (1921). British Specification, 30119. 


VIII. THE EXAMINATION OF SOME INDIAN 
FOODSTUFFS FOR THEIR VITAMIN — 
' CONTENT. 


By SUDHINDRA NATH GHOSE. 
From the Institute of Physiology, University College, London. 


(Received January 3rd, 1922.) 


THE subject of this paper was taken up at the suggestion of Dr J.C. Drummond, 
with a view to ascertaining whether the principal foodstuffs consumed by the 
mass of people in Bengal contain adequate amounts of different vitamins 
or not. 

Among the previous workers in this field mention should be made of 
Chick and her collaborators [1917, 1919], who tested certain cereals and flours 
and samples of dried fruits for the presence of vitamin B, etc. and Greig 
[1917, 1918], who examined the anti-scorbutic value of Indian grains and 
lentils. Shorten and Roy [1921] also examined some vegetables for vitamins 
Band C. Chick had found that pure white flour is deficient in vitamin B. 
Chick, Greig and Shorten found that vegetables and lentils under certain 
conditions contained good amounts of anti-scorbutic vitamin. 

The problem of supplying foodstuffs which should contain liberal amounts 
of different vitamins is very important in Bengal. Beri-beri occurs to some 
extent, and scurvy and rickets are encountered but are not widespread. From 
this but more from the low resistance of the people to various infectious 
diseases, and the presence of certain forms of eye-diseases, it may be suggested 
that the mass of the industrial classes may be living at or below the danger 
level of vitamin deficiency. 

The table given below—quoted from MacCay [1912]—shows typical diets 
ordinarily in use among the different classes in Bengal. 


Table I. 
Middle class Middle class 
(not above (above indi- 
Cultivators indigence) gence) Well-to-do 
ozs. OZs. ozs. ozs. 
Rice 20 16 6 f 
Dal l 1 1 4 
Vegetables 4 4 4 4 
1 — 1 
Fish 1 1 2 4 
Wheat flour — — 6 6 
Milk — — 8 16 
Butter a ae 1 4 
Calories (about) 2390 2310 2350 a 


(Meat and eggs (Meat and eggs 
also added) are taken, also 
sweets freely) 


3—2 


36 S. N. GHOSE 


From this table it will be seen that so far as calorific values, etc. are con- 
cerned, the diets of the people are fairly in compliance with accepted standards. 
Low resistance to disease, therefore, is probably due to vitamin deficiency — 
vitamin A specially—although MacCay is of opinion that this low resistance is 
due to want of sufficient protein in food. 

The above table may be regarded as roughly accurate, but MacCay omits 
to emphasise that all classes of the people eat large quantities of fruit during | 
the spring and summer months; also that the amount of lentils consumed 
by the poorer classes is much higher than that in the table. 

The principal “fatty foods” consumed are “ghee,” various preparations 
of butter, mustard oil, and, to a limited extent, coconut oil. These are mainly 
used in cooking. 


EXPERIMENTAL. 


This enquiry has been confined to an examination of different samples of 
“ghee,”’ mustard oil and coconut oil for the presence of vitamin A; “dals,” 
i.e. lentils, and some samples of Indian flours, for vitamin B. 

“Ghee” is made either from cows’ milk or buffalo milk. As a rule, the 
raw milk is first warmed over a low flame for about half-an-hour, and then 
cooled and inoculated with sour milk. After about 12 hours, when curdling 
is complete, water is removed and the curd churned until the fats separate 
out. The fat is removed, clarified, and then cooled into a semi-solid mass of 
buttery consistency. In some cases this product 1 is allowed to turn rancid 
and then heated and rendered. 

Fresh “ghee” remains good for months if well stored in the dark. 

The following table gives the analysis of different “ghees.” The results 
are similar to those quoted by Bolton and Revis [1910]. 


Table IT. 
Reich.-M. Free fatty 
value Sapon. V. Iodine N. acids 
Pure cows’ “ghee” (yellow) 28-7 230 31 2-2 
Pure cows’ “ghee” (remelted) eight 26 229 29 2-6 
months old sample (yellow) 

Pure buffalo “ghee” (white) 38:8 224 27-2 — 
Adulterated “ghee” (white) 14-4 213 29-3 -- 


“Ghee” is frequently adulterated with lard, mutton fat, coconut oil, and 
almond oil. 

Rats about one month old, which were over 50 g. in weight, were selected 
and put on a diet deficient in vitamin A [Drummond and Coward, 1920]. 
Generally it took about three weeks for the rats to become steady in weight 
and fit for the feeding experiment. Then each rat was given 0-2 g. of “ghee” 
every morning in addition to its usual basal diet. 

In the majority of cases the rats made steady increase in weight for about 
ten days (sometimes varying from 10 to 21 days) when they again became 


VITAMIN CONTENT OF INDIAN FOODSTUFFS 37 


stationary. The amount of “ghee” was then raised to 0-4 g. a day, when 


the rats again resumed growth. 


It is to be noted that the remelted sample did not induce growth, but 
some of the adulterated samples were fairly active when given as 5% in 


basal diet. 


It has been shown by Drummond, Coward and Watson [1921] that 
0-2-0-4 g. per day of average samples of butter is sufficient to maintain 


growth for rats. 


From the examination of the table of weighings it will be seen that both 


pure cows’ “ghee” and buffalo “ghee” 


Table ITT. 


(Showing the growth of rats on A-deficient 
diet and 0-2 g. of pure “ghee.” The weigh- 
ings of the rats during the preparatory period 
are omitted.) g. 


may be as good as European butters. 


Table IV. 


(Rats on A-deficient diet +0-2g. of pure 
“ghee.” Sample 2.) g. 


931, male 835, male 832, male 834, male 940, male 
First weighing 91 58 80 89 56 
After 10 days 102 74 85 lll YW j 
SB | Sa 104 83 90 120 76* 
Se hain vs 124 100 120 150 105 
ie | | Monee 138 110 129 162 112 
* Amount of “ghee” doubled 
Table V. Table VI. 


(Rats on remelted cows’ “ghee.’’) g. 


_ (Rats on pure buffalo “ghee.”) g. 


833, male 836, female 828, male 933, male 1176, male 932, female 
After 10 days 90 55 80 104 76 95 
After 1 ys 94 60 66 120 105 109 
” os 92 52 60 pa a eae) 
” ” 95 _— —- 136 150 136 
» 40 4, 151 — 148 
Table VII. Table VIII. 


(Rats on adulterated “ghee.” Sample I.) g. 


(Rats on adulterated “‘ghee.’’ Sample II, 
given as 5 % in basal diet.) g. 


1510, female 1177, male 1176, male _ 976, female 1300, male 
First weighing 72 65 68 80 60 
After 10 days 78 68 73 — a 
aay ns 75 72 72 110 95 
eet |) ae 80 76 74 125 109 
Table IX. Table X, 


(Rats on coconut oil and A-deficient diet. The oil was 


given in the basal diet 10-20 %.) g. 


1457, male 1456, male 1463, female 


(Rats on mustard oil: 10-20 % 
in the basal diet.) g. 


1508, male 1483, female 1511, male 


First weig 82 6+ 76 77 57 6+ 
After 10 My ae 95 69 94 84 62 76 
» 20 5 118 77 105 88 64 85 
> 130 91 102 98 72 92 


Some of the adulterated samples gave good results, but as this depends 
on the adulterants used, the different samples gave widely varying results. 


38 8S. N. GHOSE 


Daniels and Loughlin [1920] and Drummond, Golding, Zilva and Coward 
[1920] have shown that lard may contain a certain amount of vitamin 4, 
- the latter authors having shown that this occurs when the pigs have been 
grass-fed, and it should be borne in mind that lard is a most widely used 
adulterant in “ghees.” 

Contrary to the results of previous workers [cf. Jansen, 1920], the samples 
of coconut oil examined by me did produce fair growth in rats. It should be 
mentioned here that coconut oil was not administered in the same way as 
“ shees,”’ but was incorporated in the basal diet in 10 % proportion and replaced 
hardened fat. 

The results obtained with pure mustard oil are also interesting. The oil 
was administered in large quantities (10-20 % in diet) to the rats, and in 
some cases actual growth was obtained contrary to the results of previous 
workers. Both mustard oil and coconut oil were those prepared in India and 
most probably not so refined as in Europe or America. 

The buffalo “ghee” is pure white, and the presence of the growth-pro- 
moting vitamin A in it confirms Drummond and Coward’s work on vitamin A 
and colour [1920]. 

That renovation under certain conditions does not affect the vitamin 


content is also in accord with the results found by Drummond, Coward and 
Watson [1921]. 


Examination of different samples of lentils. 


The procedure for examining the samples of lentils was similar to that 
adopted in the case of “ghees,” except that in these experiments the rats had 
been previously fed on diet deficient in water-soluble B. Each rat, after it 
had begun to lose weight on this diet (B-deficient diet), was given 1 g. of the 
lentil to be tested, daily. 

The B-deficient diet was similar to the Drummond-Coward diet (A-defi- 
cient) except that no yeast extract was given and that butter replaced hard- 
ened fat. 

The weighed amount of lentil (1 g.) was kneaded with a small portion of 
the diet and given in the morning before the rat had the main bulk of its food. 

The different lentils thus examined included: 

“But” (yellow variety)—Cicer arientinum Linn. 

“Sona Moong” (golden-yellow in colour)—Phaseolus Mungo Linn. 
“Arhar” (small size, yellow-ochre variety)—Cejanus Indicus Spreng. 
“Lal Moong”’ (red variety)—Phaseolus radiatus Linn. 

“Mash Kalai” (greenish variety)—Phaseolus Mungo Linn. 

“Bara Lal But” (red variety)—Cicer arientinum Linn. 


It should be mentioned here that as far back as 1900 Grijns [1901] working 
in Java found “Katjang Hidjoe”—a variety of Phaseolus radiatus—a good 
cure for both human and avian beri-beri. Hulsoff [1917] also worked with the 


PSP Pr 


VITAMIN CONTENT OF INDIAN FOODSTUFFS 39 


anti-neuritic property of Katjong (Phaseolus radiatus). A similar result has 
also been found by Pol [1917]. 


Table XI. Composition of the various “ Dals”’ (lentils). 


Soluble 
Ether carbo- Woody 
H,O extract Protein hydrates fibre 
Name % % % % % 
5. Phaseolus Mungo Linn. 8-5 0-9 19-5 59-3 4:5 
2. » Yellow 10-99 0-83 20-5 59-0 4:8 
4. Phaseolus radiatus. Red 10-0 0-93 22-9 58-9 3:38 
1. Cicer arientinum. Yellow 11-4 4:9 18-6 56-3 6-1 
Go a Red 8-6 53 15-5 60-1 7-2 
3. Cejanus Indicus Spreng. 14-3 1-9 15-8 57:2 5:8 


Examination of the tables showing the growth of the rats on the lentils; 
shows that all the above varieties of lentils contain appreciable amounts of 
vitamin B. 


Table XII. Table XIII. 
Rats on B-deficient diet (1 g. of lentil a day). (Phaseolus Rats on Phaseolus Mungo—golden- 
Mungo—greenish variety.) g. yellow variety. g. 
1262, female 1267 a, female 1267,female 1282, male 1284, male 1285, male 
First weighing 90 102 75 92 100 80 
After 10 days 117 122 95 109 130 100 
n° 20 5 133 138 113 122 151 115 
Table XIV. Table XV. 
Rats on Phaseolus radiatus. g. Rats on Cejanus Indicus, Spreng. g. 
1258, male 1283, male 1259, female 1260, female 1261, male 
First Todays 72 100 63 86 95 
After 10 82 140 85 100 131 
” 86 162 104 115 146 
Table XVI. Table XVII. 

Rats on Cicer arientinum. Yellow variety. g. Rats on Cicer arientinum. Red variety. g. 
1264 1265 1266 1274 1277 1276 1275 
male female female male male male male: 

First weighing = 76, 76 90 82 96 80 95 
After 10 days 68 94 115 98 130 101 115 

en Oe 78 110 128 105 146 112 127 

Table XVIII. Table XIX. 
Rats on B-deficient diet and 1g. of pure 
unbleached Indian flour. g. Rats on crude “Attah,”” g. 
1278, male 1279, female 1266a, female 1268a,female 1269a, male 
First weighing 106 88 83 75 65 
After 10 days 120 115 101 105 98 
oe 206 45 135 120 111 117 117 
Table XX. 


Rats on bleached Indian flour. g. 
1292a, male 129la, female 1293a, female 
First weighing 81 69 76 
After 10 days 80 65 73 
79 Dead Dead 


” 20 ” 


40 8. N. GHOSE 


Examination of different samples of Indian flour. 


Here the samples examined were only those that are “gers used in Bengal. 
They were: 

1. Crude “attah.” (“Attah” is a mixture of various flours, e.g. wheat, 
pea and maize, etc., and contains the husks of the cereals in it.) 

2. Pure Indian flour. This flour has a finer texture than the “attah.” 
It is machine-ground, but it contains the ground-up husks of the cereals. 
This flour is not bleached. 

3.. White flour. This is machine-ground like the pure Indian flour, bat 
it does not contain the husks, and moreover it is bleached perfectly white. 
Sometimes chlorine is added in the bleaching process. 

1 g. of each of the above samples was given daily to the different sets of 
rats, in the same way as in the case of lentils. 

The composition of the different flours is given below. 


Table XXI. 


Hither <.:Cerho:. Debian ~ Orade 
H,0 extract hydrate Nx5-7 fibre 


1. Crude “ Attah” 14-6 2-9 67-1 11-5 3-9 
2. Pure unbleached flour 15 2 71-2 ll 0-8 
3. Bleached white 12 2 71:3 13-5 0-6 


The results found are in accord with those of Chick as well as with the 
facts announced by Col. Hehir [1919, 1, 2] in the Mesopotamia Commission 
Report on the presence of deficiency diseases in the besieged garrison in Kut. 
Willcox [1917] has also discussed “attah” as a protective against beri-beri. 

Johns and Finks [1920] found that a mixture of pea-flour and wheat- 
flour produces better growth in rats than simple wheat-flour, and these facts 
are also corroborated since “attah”’ is found to be better in promoting growth 
than pure (unbleached) flour. From the weighings of the rats, fed on the 
different samples of the flours, it will be seen that with the exception of 
bleached flour they contain appreciable amounts of vitamin B. 


CONCLUSION AND SUMMARY. 


Samples of pure Indian “ghee” proved to be as good a source of vitamin A 
as pure butter. 

Samples of remelted “ghee” (“makkum galana”’) failed to restore growth 
in rats whose weights had become stationary on a shortage of fat-soluble A. 

Two samples of “ghee” whose history is unknown failed to promote growth. 

Some samples of adulterated “ghee” only gave growth in rats when ad- 
ministered in large quantities (e.g. 5 % of the basal diet). 

Certain edible vegetable oils, e.g. (“narikel”’) coconut oil, and pure mustard- 
seed oil were examined, and these were found to give growth when ad- 
ministered in 10-20 °%, in the basal diet. 


VITAMIN CONTENT OF INDIAN FOODSTUFFS 41 


Various samples of lentils were examined and these showed good content 
of vitamin B. 

Bleached (pure white) Indian flour is found to be deficient in vitamin B, 
but both crude “attah” and unbleached Indian flour have considerable 
quantities of water-soluble vitamin B. 

In conclusion I take the opportunity of expressing my thanks to Dr J. C. 
Drummond and Miss Katharine H. Coward, for kind help and advice. 


My thanks are also due to Mr Suvaga for kindly giving me some samples 
of remelted “ghee,” and to my brother, Mr B. N. Ghose, for various samples 
of lentils, flours, oils and “ghees.” 


REFERENCES. 


Bolton and Revis (1910). Quoted in Analyst, 35, 343; from “Fatty Foods” (Churchill). 
Chick and others (1917). Proc. Roy. Soc. B. 90, 44. 

—— —— (1919). Biochem. J. 13, 199. 

Daniels and Loughlin (1920). J. Biol. Chem. 42, 359. 

Drummond and Coward (1920). Biochem. J. 14, 661. 

Drummond, Coward and Watson (1921). Biochem. J. 15, 540. 

Drummond, Golding, Zilva and Coward (1920). Biochem. J. 14, 742. 

Greig and others (1917). Indian Journ. Med. Res. 4, 818. 

—— —— (1918). Indian Journ. Med. Res. 6, 56. 

Grijns, quoted by Hulsoff (1901). Genees. Tijds. v. Ned. Ind. 41, 3. 

Hehir (1919, 1). Mesopotamia Commission Report. Appendix. Proc. Asiatic Soc. Bengal, 15, 212. 
—— (1919, 2). Indian Journ. Med. Res. Indian Science Congress, pp. 44-59 and 79-82. 
Hulsoff (1917). J. Physiol. 51, 432. 

Johns and Finks (1920). J. Biol. Chem. 42, 491. 

Jansen (1920). Genees. Tijd. v. Ned. Ind. 58, 173. 

MacCay (1912). “The Protein Element in Nutrition” (Arnold), p. 78. 

Pol, D. H. (1917). Ned. Tijd. v. Genees. 11, 806. 

Shorten and Ray (1921). Indian J. Med. Res. Annual Science Congress. Biochem. J. 15, 274. 
Willcox (1917). Lancet, ii, 77. 


IX. CONDITIONS OF INACTIVATION OF THE 
ACCESSORY FOOD FACTORS. | 


By SYLVESTER SOLOMON ZILVA. 
From the Biochemical Department, Lister Institute. 


(Received January Hh, 1922.) 


In some preliminary communications [Zilva, 1920, 1921] attention was drawn 
to the fact that the fat-soluble and the anti-scorbutic factors were inactivated 
on exposure to ozone in the case of the former and ozone and air in the case 
of the latter. It was further pointed out that the anti-scorbutic factor could 
withstand the temperature of 100° for two hours when air was excluded. These 
observations have found confirmation in the results obtained independently 
_ by other workers. Hopkins [1920] has shown that when air is bubbled through 
butter for four hours at 120° the fat-soluble factor becomes inactivated. 
Drummond and Coward [1920] have also demonstrated that the destruction 
of this factor takes place in butter-fat in presence of air rapidly at high 
temperatures and more slowly at lower temperatures. Hess [1920] and Hess 
and Unger [1921] found that milk incubated with hydrogen peroxide lost its 
anti-scorbutic properties, whilst orange juice, milk, and tomato juice treated 
with oxygen lost some of their anti-scorbutic potency. Dutcher, Harshaw, 
and Hall [1921] have since confirmed Hess and Unger’s observations. con- 
cerning the destruction of the anti-scorbutic factor by hydrogen peroxide 
and the writer’s observation on the stability of this principle at high tempera- 
tures in the absence of air. 

The object of this communication is to describe a series of experiments 
on the action of ozone and air on the three accessory factors, observations 
having been made at various temperatures. These experiments owe their 
origin to the fact that the inactivation of the fat-soluble factor in butter by 
exposure to a mercury-quartz lamp [Zilva, 1919] was found by the author 
to be due to the action of the ozone generated by the lamp and not to the 
action of the ultra-violet rays. 


THE INFLUENCE OF OZONE AND AIR ON THE FAT-SOLUBLE FACTOR. 


In the preliminary communication [Zilva, 1920] it was pointed out that 
when cod liver oil was exposed to the mercury-quartz lamp for 16 hours in 
the absence of air the activity of the oil was not destroyed by the action of 
the light. On the other hand six hours of exposure of the same oil to ozone 
in the dark entirely inactivated it. 

In testing the above oils for the fat-soluble factor 2 g. of the oil were added 
to the daily basal diet of 20 g. when the rats had ceased to grow owing to the 
deficiency in the diet. The following are the results obtained: 


- INACTIVATION OF ACCESSORY FOOD FACTORS 43 


With the addition of oil exposed to the mercury-quartz lamp 
in the absence of air. 
Rat I 2 gained 30 g. in 17 days. 
Rat II 2 gained 24 g. in 17 days. 
Rat IIT 2 gained 39 g. in 17 days. 


With the addition of oil exposed to ozone for six hours in the dark. 


Rat I 3. No gain in weight in four weeks. On addition at the end of 
that time of a similar dose of the unexposed oil the animal resumed growth 
and gained 37 g. in three weeks. 

Rat II 3. No gain in weight in four weeks. On addition at the end of 
that time of a similar dose of unexposed oil the animal resumed growth and 
gained 23 g. in 16 days. 

Rat III 3. Declined in weight and died 26 days after the commencement 
of the administration of the oil. Developed keratomalacia during the period 
of treatment with the oil. 

Rat IV 9. No gain in weight in four weeks. On addition at the end of 
that time of a similar dose of unexposed oil the animal resumed growth and 
gained 29 g. in three weeks. 

From the above experiments it is quite evident that ultra-violet light as 
such has no deleterious action on the fat-soluble factor. On the other hand 
ozone does inactivate it. Moreover the action of ozone must be very drastic, 
when one bears in mind that cod-liver oil is very potent. The oil employed 
in the above experiments has been found capable of inducing growth in rats 
in daily doses of 1-7 mgm. The experimental rats on the exposed oil consumed 
60-70 %, of their diet, in other words 1-2—1-4 g. of the oil or about 750 times 
as much as the minimum dose necessary to induce growth, without showing 
any tendency to grow. Possibly a much shorter exposure would have also 
produced inactivation of this factor. 

In the earlier part of this work whale oil was used instead of cod-liver oil. 
The former is much less active than the latter and was given up for that 
reason. At the time the experiments were carried out it was not definitely 
settled that the fat-soluble factor was associated with the unsaponifiable 
fraction of the oil and it was therefore considered advisable to study the 
change in the iodine value of the oil during exposure to the mercury-quartz 
lamp. The differences observed were of a low magnitude. The following are 
the figures of a representative experiment: 


Iodine value of whale oil before exposure __... 119 
Iodine value of whale oil exposed to the lamp for sic 
hours in the absence of air ... : id coe | BERG 
Iodine value of whale oil exposed for iatst eae in the 
presence of the generated ozone... a sew) LAG 


The change is not significant and the results are analogous to those ob- 
tained by Hopkins [1920] with butter heated in a stream of air at 120°. No 


44 S. 8S. ZILVA 


change was recorded in the iodine value of the butter after the treatment. 
It is very doubtful whether a chemical change such as the iodine value of 
the oil can have any bearing on the accessory factor. 

Some experiments in connection with the inactivation of the fat-soluble 
were carried out in connection with another inquiry. It was necessary to 
inactivate a large batch of very potent cod-liver oil and Hopkins’ method 
was adopted for that purpose. This method has an advantage over the em- 
ployment of ozone because it does not change the consistency of the oil to 
such a great extent. Air was aspirated through the oil heated at 120° for 
18 hours. By this means a very active oil was inactivated to the extent that 
a daily dose of about 0-75 g. did not induce any growth in rats. 


THE INFLUENCE. OF OZONE ON THE WATER-SOLUBLE FACTOR. 


Autolysed yeast juice was employed as a source for the water-soluble 
factor in this experiment. The juice was exposed to the ozone with constant 
shaking in the same apparatus and for the same time as the cod-liver oil. 
After the exposure the juice was very much altered in taste, smell and colour. 
It was incorporated in the basal diet free from the water-soluble factor after 
the rats had been on the restricted diet between three and four weeks. Three 
animals received 1, 2 and 3 cc. of autolysed yeast juice in their daily basal 
diet of 20g. All the three animals responded by resuming growth, gaining 
19, 24 and 15g. respectively in 17 days. As 1 cc. of autolysed yeast juice 
is not much above the minimum dose capable of inducing growth there could 
not have been much destruction of the water-soluble factor in the juice by 
this very drastic method and we may conclude that this factor is decidedly 
more resistant to oxidation than the other two accessory factors. This re- 
sistance explains why on exposing autolysed yeast juice in a Petri dish to 
a mercury-quartz lamp no inactivation of the water-soluble factor takes place 
[Zilva, 1919]. Whether drastic oxidation inactivates the water-soluble factor 
at higher temperatures still remains to be ascertained. 


THE INFLUENCE OF OZONE AND AIR ON THE ANTI-SCORBUTIC FACTOR. 


The influence of ozone in the dark on the anti-scorbutic factor was next 
investigated. Decitrated lemon juice, namely lemon juice from which the 
organic acids are removed by neutralisation with calcium carbonate and 
which retains almost the entire anti-scorbutic potency of the original juice, 
was exposed to ozone in precisely the same way as the active oils and the 
autolysed yeast juice. Fresh batches were prepared every two or three days 
so that no deterioration could take place on storage. The treated juices, as 
well as the control untreated juices of the same batch, were administered to 
guinea-pigs which had been kept on a scorbutic diet of oats and bran and 
autoclaved milk (40 ce.) for 14 days. Daily doses of 3, 5 and 7 cc. were ad- 
ministered with the following results: 


INACTIVATION OF ACCESSORY FOOD FACTORS 45 


Guinea-pig I. Dose 3.cc. Commenced losing weight after three weeks. 
Declined gradually and died 44 days after the commencement of the experi- 
ment. At the post-mortem examination evidence of scurvy was found. 

Guinea-pig II. Dose 5 cc. Declined and died 35 days after the commence- 
ment of the experiment. At the post-mortem examination evidence of scurvy 
was found. ; 

Guinea-pig III. Dose 7 cc. Maintained its weight for 53 days. Chloro- 
formed at the end of this period. At the post-mortem examination some 
evidence of scurvy was found. 

A control dose of 3 cc. of the original decitrated juice was found to be 
adequate to prevent scurvy for 56 days; the animal was still growing when 
the experiment was discontinued. 

One may conclude from the above experiments that the ozone inactivated 
the juice to a great extent. The writer has previously found [Zilva, 1919] 
that when decitrated lemon juice was exposed to the mercury-quartz lamp 
its anti-scorbutic potency was not destroyed or even impaired in spite of 
the generated ozone. This can be explained by the fact that the juice was 
exposed in a quartz tube and, unlike the butter in that experiment, was 
shielded almost entirely from the generated ozone. The anti-scorbutic, like 
the fat-soluble factor, is, therefore, inactivated by ozone but is not affected 
by ultra-violet rays. 

In view of the results obtained in the above experiments it became of 
interest to ascertain whether less drastic oxidative treatment such as exposure 
to air would also inactivate the anti-scorbutic factor. Air was therefore aspi- 
rated through decitrated lemon juice for 12 hours. Batches of juice prepared 
every two or three days were employed. The following is the protocol of the 
experiment: 

Guinea-pig I. Dose 3 cc. Just maintained its weight for 51 days gradually 
declining in health and showing symptoms of scurvy during this period. The 
animal: was chloroformed owing to its distressed condition. At the post- 
mortem, examination evidence of advanced scurvy was found. 

Guinea-pig II. Dose 5 cc. Commenced declining in weight after 17 days. 
Chloroformed 48 days after the commencement of the experiment. At the 
post-mortem examination evidence of scurvy was found. 

Guinea-pig III. Dose 7 cc. Gained 57 g. in 51 days. Chloroformed 51 days 
after the experiment commenced. Evidence of mild scurvy was found at the 
post-mortem examination in this case. 

A control dose of 3 cc. of the original juice was found quite adequate to 
prevent scurvy for 51 days. The control animals were still growing when the 
experiment was discontinued. 

This experiment affords evidence that the exposure of decitrated lemon 
juice to air for 12 hours diminishes its anti-scorbutic potency very appreciably. 

The above observations suggested the possibility that the inactivation of 
the anti-scorbutic factor by heat may be due in part if not entirely to the 


46 8. 8. ZILVA 


presence of air during the heating. Such a possibility was conjectured by 
Delf [1920] who found that orange juice and swede juice retained their anti- 
scorbutic potency to a marked extent after being heated for a considerable 
time at temperatures higher than 100°. In order to test this theory the following 
experiments were devised. Two lots of decitrated lemon juice were heated 
at 100° under a reflux condenser. One was heated for one hour during which 
time air was bubbled through. The second was heated for two hours, air 
being all the time rigorously kept out by bubbling carbon dioxide gas through 
the solution. In this case also fresh batches of decitrated lemon juice were 
employed every two or three days. The following results were recorded: 


Decitrated lemon juice heated in air for one hour. 


Guinea-pig I. Dose 1-5 cc. Commenced declining in weight after a month. 
Chloroformed 43 days after the commencement of the experiment. Evidence 
of scurvy was found at the post-mortem examination. 

Guinea-pig II. Dose 4.cc. Commenced to decline after three weeks and 
died 41 days after the commencement of the experiment. Evidence of acute 
scurvy was found at the post-mortem examination. 

Guinea-pig III. Dose 6 cc. Maintained weight for 36 days and then 
suddenly declined. Died 43 days after the commencement of the experiment. 
Evidence of scurvy was found at the post-mortem examination. 


The same juice heated in an atmosphere of carbon dioxide for two hours. 


Guinea-pig I. Dose 1-5 ce. Gained 122 g. in 74 days. Chloroformed. No 
evidence of scurvy was found at the post-mortem examination. 

Guinea-pig II. Dose 4c. Gained 48 g. in 74 days. Chloroformed. No 
evidence of scurvy was found at the post-mortem examination. 

Guinea-pig III. Dose 6 ec. Gained 138 g. in 74 days. Chloroformed. No 
evidence of scurvy was found at the post-mortem examination. 

This experiment affords proof that the anti-scorbutic factor is quite stable 
when heated in the absence of air. 

Decitrated lemon juice contains a considerable quantity of reducing sugar. 
On hydrolysis, however, the reducing capacity of the juice increases slightly, 
a reducing substance, most probably a sugar, being liberated. An experiment 
was devised to ascertain whether the anti-scorbutic potency is destroyed 
when the juice is hydrolysed under anaerobic conditions. Decitrated lemon 
juice was therefore hydrolysed on a water-bath under a reflux condenser for 
five hours with 2 N hydrochloric acid. The air was displaced during 
hydrolysis by carbon dioxide. After hydrolysis the solution was neutralised 
with sodium carbonate and tested out for its anti-scorbutic potency on guinea- 
pigs. A fresh preparation of lemon juice was hydrolysed every day for 
feeding purposes. The following results were obtained: 

Guinea-pig I. Dose 1-5 cc, Declined after three weeks and died 32 days 


INACTIVATION OF ACCESSORY FOOD FACTORS 47 


after the commencement of the experiment. Evidence of Scurvy, at the post- 
mortem examination. 

- Guinea-pig II. Dose 1:5 cc. Declined after 23 days. Died 30 days after 
the commencement of the experiment. Evidence of scurvy at the post- 
mortem examination. 

Guinea-pig III. Dose 1-5 cc. Declined after three weeks and died 26 days 
after the commencement of the experiment. Evidence of scurvy at the post- 
mortem examination. 

Guinea-pig IV. Dose 3 cc. Gained 30g. in 37 days. Chloroformed. At 
the post-mortem examination the animal was found in excellent ngs te 
but showed evidence of extremely mild scurvy. 

Guinea-pig V. Dose 3 cc. Gained 28 g. in 37 days. Chloroformed. At the 
post-mortem examination the animal was found in excellent condition showing 
no evidence of scurvy. 

Guinea-pig VI. Dose 3cc. Gained 16g. in 37 days. Chloroformed. At 
the post-mortem examination the animal was found in excellent condition 
showing no evidence of scurvy. 

Guinea-pig VII. Dose 5 cc. Gained 18 g. in 37 days. Chloroformed. At 
the. post-mortem examination the animal was found in excellent condition 
showing no evidence of scurvy. 

Guinea-pig VIII. Dose 5 cc. Gained 71 g. in 37 days. Chloroformed. At 
the post-mortem examination the animal was found in excellent condition 
showing no evidence of scurvy. 

Guinea-pig IX. Dose 5 ce. Gained 30 g. in 37 days. Chloroformed. At 
the post-mortem examination the animal was found in excellent condition 
showing no evidence of scurvy. 

No control was tested out with this experiment owing to some mishap, 
but numerous experiments performed by the writer with decitrated lemon 
juice show that the minimum dose capable of preventing scurvy in guinea- 
pigs is about 1:5cc. The conclusion one can draw from this experiment is 
that the hydrolysis diminishes the activity of the juice, but in spite of the 
very drastic treatment very significant activity is retained. This further 
confirms the view that the anti-scorbutic factor is stable towards heat pro- 
viding anaerobic conditions are observed. 


SUMMARY. 


1. The fat-soluble factor was destroyed in cod-liver oil by exposing it to 
ozone in the dark at ordinary temperature and by bubbling air through it 
at 120°. 

2. The exposure of autolysed yeast juice to ozone in the dark did not 
destroy the activity of its water-soluble factor to any appreciable extent. 

3. The exposure of decitrated lemon juice to ozone destroyed its anti- 
scorbutic activity. This was also destroyed by passing air through the solution 
at ordinary temperature. 


48 S. 8. ZILVA 


4. Ultra-violet rays had no deleterious action on the accessory food 
factors in the absence of air. 

5. Two hours’ boiling did not destroy the anti-scorbutic potency of deci- 
trated lemon juice in an atmosphere of carbon dioxide, whereas one hour’s 
boiling in the presence of air destroyed its potency almost entirely. 

6. Hydrolysis of decitrated lemon juice with hydrochloric acid for five 
hours impaired the anti-scorbutic activity of the juice. The juice however 
retained its activity to a very considerable extent. 


In conclusion I wish to express my best thanks to Professor Sensho Hata 
and Miss E. M. Low for help in some of the experiments in this inquiry. My 
thanks are also due to Dr J. 8. Edkins for having kindly permitted me to use 
his ozone generator. 


The expenses of this research were defrayed from a grant made by the 
Medical Research Council, to whom my thanks are due. 


REFERENCES. 


Delf (1920). Biochem. J. 14, 211. 

Drummond and Coward (1920). Biochem. J. 14, 734. 
Dutcher, Harshaw and Hall (1921). J. Biol. Chem. 47, 483. 
Hess (1920). Brit. Med. J. 147. 

Hess and Unger (1921). Proc. Soc. Exp. Biol. Med. 18, 143. 
Hopkins (1920). Biochem. J. 14, 725. 

Zilva (1919). Biochem. J. 18, 164, 

Zilva (1920). Biochem. J. 14, 740. 

Zilva (1921). Lancet, i, 478. 


X. AN IMPROVISED ELECTRIC THERMOSTAT 
CONSTANT TO 0:02’. 


By SAMUEL CLEMENT BRADFORD. 


From the Science Museum. 
(Received January 17th, 1922.) . 


In these days of elaborate physical apparatus it may be of interest to describe 
a simple thermostat which can be made easily from material at hand and is 
sufficiently constant for most purposes. The apparatus, shown in the photo- 
graph, was required for an investigation expected to extend over twelve 
months or more and needing only moderate constancy. 

The principle of the stirrer, on which the success of the thermostat depends, 
was suggested by Mr 8. Bradshaw, to whom the author is indebted, also, for 
much kind assistance in making the apparatus (Plate I). The stirrer comprises 
a solenoid, A, with an iron plunger, B, mounted on a glass tube carrying an 
ordinary cork bung, shown in dotted outline, at its lower extremity. The 
- upper part of the glass shaft bears two rubber corks, C and D, that strike the 
left arm of the brass lever, E, whose right arm carries a cylindrical plumbago 
contact. As the stirrer reaches the lower end of its stroke, the cork, C, pulls 
over the lever, causing the plumbago contact to engage with a similar con- 
tact, F, above. The current from the 110 volt supply now flows through the 
solenoid, raising the plunger until the cork, D, throws the lever back again, 
breaking the circuit and allowing the plunger to fall. To render the lever 
stable in either position, it is made top-heavy by means of the weight, G, 
carried on a stout wire soldered to the lever near its pivot. 

For temperatures up to 25°, a 16 c.p. carbon lamp is sufficient as a heater. 
For higher temperatures a more powerful lamp may be employed, or an 
auxiliary one continually alight. The brass parts of the lamp are protected 
from the water by a short piece of 1 inch rubber tubing. 

As pointed out by Lewis [1922], the heater should be placed near the 
thermo-regulator, M, which is similar to an ordinary toluene thermo-regulator 
for gas. This was made from a stout glass test-tube of about 50 cc. capacity 
when sealed to about 3 inches of 4mm. glass tubing. To this was joined 
about 6 inches of 1 mm. bore capillary tube, having a short wider tube at 
its open end. About 2 inches from the upper end of the capillary, a short 
piece of 4 mm. tube, N, was sealed at right angles. The wider tube, joining 
the capillary and bulb, was then bent through 180° near its junction with the 
bulb. The bulb contains toluene and some mercury which also fills the 


Bioch. xvr 4 


50 8. C. BRADFORD 


capillary tube. The height of the mercury in the latter is adjusted by sliding, 
within the side tube, a short piece of capillary tube bearing a platinum contact 
at its sealed end.. The joint between the side tube and its piston is closed by 
a slieve of rubber tubing. The other platinum contact is carried at the end 
of a piece of glass tube, O, drawn out so as to pass vertically within the wider 
end of the capillary tube and rest on the constriction, with the contact pro- 
jecting into the capillary tube. 

These two contacts are connected, by mercury seals, in series with a 
4 volt accumulator and the magnetic cut-out for the heater. For the cut-out, 
a discarded high-resistance telegraph relay, P, was fitted with iron wire dippers 
and two mercury cups in circuit with the heater. However, almost any high- 
resistance electro-magnet and armature could be adapted. By using an 
accumulator to actuate the cut-out, there can be no fouling of the mercury 
in the thermo-regulator, and it is unnecessary to provide for an up-and-down 
motion of the upper platinum contact. The capacity of the accumulator is 
immaterial, as it can be charged while in use through a single lamp, the 
potential difference across its terminals remaining at 4 volts. Actually a 
40 ampere-hour accumulator will last a month without recharging. The 
capacity of the bath is about 5 litres. A constant level is maintained by the 
Marriott’s bottle shown on the left. 

Had the temperature remained constant within 0-1°, expectations would 
have been fulfilled. Consequently, it was astonishing to find that the varia- 
tions were limited to 0-02°. The observations recorded below extended over 
a period of nearly 24 hours. The maximum variation occurs within a short 
period corresponding to, but lagging behind, the heating cycle. By bringing 
the heater right up to the thermo-regulator, probably even this small variation 
would be diminished. The disadvantage of doing so, when working for very 
long periods, lies in the chance of the mercury in the cut-out being fouled 
on account of the greater number of breaks. During the period under observa- 
tion only one reading was made differing by more than 0-01° from the mean, 
and this was only 0-05° greater. 

The solenoid, 44 inches long, with a pasteboard tube, allowing ;y inch 
clear space round the iron core, is wound with about eight layers of No. 27 
single cotton covered wire insulated with shellac varnish. The core is made 
from gas pipe, ? inch diameter, the same length as the solenoid, turned down 
to about sy inch thickness. Or it could be made from sheet iron. The glass 
shaft passes through two rubber corks in the ends of the core and is sup- 
ported by guides, Q and R, constructed of sheet brass. It is convenient to 
make the glass shaft in two pieces joined by a cork, shown at the level of the 
heating lamp. 

The contact lever, E, is the only part of the stirrer needing careful con- 
struction. It is fashioned from 18 gauge sheet brass. The pivot bearing is a 
piece of brass tube, 4 inch long, sweated through the lever at its centre of mass, 
and ground with a little emery and oil to fit smoothly over the shank of a 


AN ELECTRIC THERMOSTAT 51 


round-headed brass screw forming the pivot. The mass of the weight, G, is 
about 25g. One end of the bearing abuts against a brass plate screwed to 
the base-board, the other against the flange of the screw, which is adjusted 
so that the lever just swings easily, without looseness, between the two stops, 
K and L, jacketed with rubber tube. The thin steel springs, H and J, are 
carefully bent so that their pressures increase or diminish very gradually 
and evenly throughout the stroke. They engage the under side of the lever 
and the lower one should project upwards to the level of the end of the upper 
one. The pressures of the springs are adjusted by means of the screws seen 
near their centres, and must be only sufficient just to absorb the energy of 
the lever as it is brought to rest against the stops. To prevent sparking at 
the pivot, with consequent roughening of the bearing, the current is conveyed 
to the lever through a short piece of flexible wire soldered to it, as shown in 
the photograph. 


(a) Thermometer remote from 


thermo-regulator (b) Thermometer near thermo-regulator 
0 fa 0 Y 4 
h. m. 8. h.m. s 
oe —__—_—— 
23-52° 3.5 Op.m. 23-515° 12 30 Op.m 
“51 7 0 505 30 30 
50 7 30 505 31 0 
Bl 8 0 515 31 30 
+52 9 0 52 32 0 
515 32 10 
23-50 317 0 51 32 30 
51 33 (0 
23-50 345 0 
23-51 110 0 
23-50 40 0 515 ll O 
51 11 20 
23-51 415 0 51 12 0 
505 12 20 
23-52 10 30 Oa.m “51 13 0 
23-52 10 50 0 23-515 2 12 30 
“52 53 0 “bl 12 50 
53 54 0 “50 13 10 
-50 13 30 
23-515 ll 5 0 505 14 0 
51 14 10 
23-52 11 39 0 515 14 30 
51 39 30 51 14 50 
505 40 0 505 15 0 
“51 40 10 
+52 40 20 
525 41 0 
52 41 20 
23-52 12 26 Op.m 
“51 26 30 
“50 27 0 
51 27 20 
“52 27 30 


The induction in the solenoid causes a powerful spark at each break and 
a great deal of trouble was experienced with the contacts. When first set up, 
the apparatus worked for several hours, but the metal contacts used soon 
became corroded. Mercury contacts in oil lasted only about 48 hours. Finally, 


4—2 


52 8S. C. BRADFORD 


cylindrical plumbago contacts, 8 mm. in diameter, as used for magneto brushes 
were adopted. The lower one fits firmly in a clip soldered to the lever. The 
upper contact slides tightly in a thin split metal cylinder, which fits loosely in 
a short vertical piece of glass tube, clamped to the base-board, and has a pro- 
jection at the top to prevent the contact holder dropping through the glass 
tube. The flexible wire bringing the current is soldered to this projection. 
The position of the upper plumbago rod in its holder is adjusted so that, when 
the lever is in the contact position, the lower rod has raised the upper one 
about 7; inch above its resting position. Thus, at each make and break, the 
lower contact pushes the upper one up and down, and upon this motion the 
efficiency of the apparatus appears to depend. Probably carbon contacts 
would serve nearly as well as plumbago ones. 

’ Even with this arrangement, however, the apparatus could hardly have 
been regarded as satisfactory, owing to excessive sparking. Ultimately the 
device was adopted of allowing a small current, just insufficient to raise the 
plunger, to flow through the coil constantly. Only the extra current, necessary 
to actuate the stirrer, is passed through the contacts. With this arrangement, 
the coil is always on a closed circuit, and the smaller induced current, due to 
breaking the extra current, discharges through the closed circuit, almost 
entirely avoiding the spark. The constant current flows through an 8 c.P. 
carbon lamp, and the size of the bung is then adjusted so that its buoyancy is 
sufficient just not to float the plunger. The extra current passes through an 
8 c.p. lamp or preferably, a 16 c.P. lamp with a resistance (which may be made 
of iron wire), about 150 ohms, adjusted so that the plunger moves regularly 
with the required power. The position of the corks, C and D, is adapted to 
the length of stroke, about 24 inches. 

The only attentions the apparatus requires are to give a drop of oil to the 
pivot, the brass springs and the sliding contact holder, and to sand-paper the 
surfaces of the plumbago contacts about every 48 hours. Occasionally the 
pivot screw, or those of the springs, may need a fraction of a turn. The 
guides of the shaft are best’ lubricated with vaseline. At the time of writing 
the thermostat has been running smoothly for over a month. 


REFERENCE. 
Lewis (1922). Trans. Farad. Soc. 17, 1. 


BIOCHEMICAL JOURNAL, VOL. XVI. NO. 1 PLATE 


XI. TETHELIN—THE ALLEGED GROWTH-CON- 
TROLLING SUBSTANCE OF THE ANTERIOR 
LOBE OF THE PITUITARY GLAND. 


By JACK CECIL DRUMMOND 
AND 
ROBERT KEITH CANNAN (Beit Memorial Research Fellow). 


From the Biochemcial Laboratories, The Institute of Physiology, 
-Unwersity College, London. 


(Received January 17th, 1922.) 


CERTAIN conditions of disordered development, such as acromegaly and 
gigantism, have long been associated with abnormal conditions of the pituitary 
body and in consequence the function of this gland in the growth of the 
animal has been the subject of much investigation. The suggestion that the 
pituitary body supplies a hormone which regulates the growth of the body, 
particularly of the skeleton, arose directly from the pathology of the diseases 
in question and has directed research along two definite lines. On the one 
hand studies of the effect of extirpation of the gland have been made, on the 
other, light has been sought in experiments based on the administration of 
the gland or its extracts to a normal animal. 

The extirpation of one or other of the lobes of the pituitary body without 
injury to the remainder or to surrounding tissues is attended with extreme 
difficulty owing to the intimate association of the parts and, in consequence, 
results of such experiments have been conflicting. It is generally agreed that 
removal of the posterior lobe gives rise to no marked symptoms, whereas 
the balance of opinion points to the conclusion that with extirpation of the 
anterior lobe in young animals growth is checked, fatty deposits increase, 
metabolism is reduced and persistent infantilism results. 

The results of experiments on the administration of the gland to normal 
animals are, again, far from unanimous. Cerletti [1907] injected pituitary 
emulsion into young animals and reported retardation of bone growth and 
Cushing [1909] by repeated injection of anterior lobe extracts obtained a fall 
in weight of young animals. The feeding of pituitary by Sandri [1909] to 
young guinea-pigs arrested growth whilst Schafer [1912] found that oral 
administration of anterior lobe was without appreciable effect on the growth 
and metabolism of young rats. Aldrich [1921] from experiments on puppies 
came to the same conclusion. 


54 J. C. DRUMMOND AND R. K. CANNAN 


The general opinion at this period was, therefore, that oral administration 
of anterior lobe did not appreciably accentuate any regulating influence which 
the gland might normally have upon growth. The value of much of this 
earlier work is however depreciated by the fact that various forms of ad- 
ministration and different preparations of the gland were adopted by the 
different workers. 

The whole question was reopened by a series of papers by Brailsford 
Robertson in 1916. He claimed to have demonstrated that the feeding of 
the fresh anterior lobe of the ox to white mice leads to a “marked retardation 
of growth in the beginning of the third growth cycle (6-10 weeks) followed 
by a pronounced acceleration from the 20th to 60th weeks” [1916, 1]. Further, 
he isolated from the anterior lobe of the ox a substance to which he gave the 
name fethelin, and which he regards as the active growth-controlling principle 
[1916, 2]. 

We were, at first, impressed by the care with which Robertson’s feeding 
experiments had been carried out and it did appear that, by the employment 
of large groups of animals and by a statistical control of his data, he had 
guarded against all sources of error common to such experiments. A closer 
inspection of his work, however, revealed many inconsistencies which in our 
opinion seriously detract from the validity of his conclusions. In particular, 
it appeared that nothing but a substance of very impure character could 
possibly be obtained by the method he described for the isolation of tethelin 
[1916, 3]. 

The communications referred to were followed by others in which the 
chemical and physiological properties of tethelin were described and its effects 
in increasing the rate of tumour growth [1916, 4], in accelerating the healing 
of wounds and recovery after inanition [1916, 5] were reported. Meanwhile 
other workers were obtaining varying results. Goetsch [1916] and Marinus 
[1919] independently observed accelerated growth in rats and Wulzen [1916] 
accelerated division of planarian worms whilst Pearl [1916] obtained retarded 
growth of chickens fed upon anterior lobe. 

The conflicting nature of all these results is not dispelled by Robertson’s 
later papers in which, from a consideration of his accumulated data, he 
modifies his former views and comes to the conclusion that the influence of 
tethelin upon growth consists, as far as the whole animal is concerned, in 
retardation. His latest explanation of his former conclusions and the de- 
velopment of these views into a theory of growth catalysers is ingenious but 
quite unconvincing [1919, 1921]. We were led by the conflicting nature of 
all the evidence to undertake the re-investigation the results of which are 
reported in this paper. For the opportunity to carry out the work we are 
indebted to Mr T. O. Kent, who most kindly supplied us a Nth with 
quantities of fresh anterior lobe of the pituitary of the ox. 


TETHELIN 55 


Isolation of tethelin. 


The method for the isolation of tethelin described by Robertson [1916, 3] 
is briefly as follows.. The mixed anterior lobes dried by intimate mixture 
‘with an anhydrous inorganic salt are extracted for 48 hours with boiling 
absolute alcohol. The extract is concentrated im vacuo until separation of a 
precipitate begins. After cooling, one and one half volumes of dry ether are 
added and a voluminous white precipitate separates, which, after three hours, 
is filtered, washed with alcohol-ether mixture and dried in an incubator at 
30-35°. Robertson appears to claim chiefly from constancy of nitrogen and 
phosphorus content that the substance is a chemical individual. To anyone 
familiar with such methods for the extraction of tissues it must be obvious 
that only a crude product, consisting largely of lipoid substances would be 
yielded by his technique and, further, that no such product would preserve 
its characteristics unless carefully protected from oxidation throughout the 
preparation and drying. We have made several attempts to prepare this 
substance and to obtain a white precipitate such as Robertson describes by 
following the detail of his method. Our yield varied from 0-4 to 0-8 % of the 
moist gland, but the products obtained did not show constancy of nitrogen 
and phosphorus content as reported by Robertson. The analyses of two 
samples which agree in their content of nitrogen are given in Table I. In all 
cases our preparation was carried through in an atmosphere of carbon dioxide. 


Table I. Analyses. 


Our preparations 
o m1 Robertson Protagon 
Nit 2-88 2-80 2-58 2-39 
Phosphorus 1-76 0-81 1-41 1-07 


Our preparations resembled in many ways the mixture of lipoid substances 
which one can prepare from animal tissues by following a method essentially 
that of Robertson, and the analytical figures as well as the appearance and 
properties of the products show a resemblance to admittedly impure mixtures 
of this type such as protagon. Robertson described a numberof properties 
of tethelin which are typical of such lipoid mixtures but we were unable to 
confirm the colour reactions upon which he is rash enough to suggest the 
presence in the molecule of inositol and of an iminazolyl radicle. 

On the assumption that tethelin represents an impure mixture of sub- 
stances of the lipoid class, perhaps contaminated by other cell constituents— 
no purification of his product appears to have been attempted by the author— 
we submitted our preparations to a simple fractionation. By extraction with 
hot alcohol they could be divided into a fraction readily soluble in that solvent 
and an insoluble fraction. 

The soluble fraction deposited, on standing, a white pseudo-crystalline 
product of which 1-68 g. were obtained (7',). The mother liquor of 7’, yielded 


56 J. C. DRUMMOND AND R. K. CANNAN 


on evaporation a sticky brown residue, T3, whilst the fraction insoluble in 
alcohol was designated 7. On analysis these products gave the figures: 


T, Ts T, 
Nitrogen 1-63 3-43 3°23 
Phosphorus 0-88 1-54 3°94 


We endeavoured to carry the fractionation further but were led to abandon 
it owing to the small amount of material at our disposal and to our conviction 
that the matter was not worthy of further consideration. 

We did, however, examine the fraction 7, which appeared to have the 
properties of an impure mixture of ‘slctanidless 

A determination of the melting point as described by Rosenheim [1914, 1] - 
showed the following: 


Phrenosin (Rosenheim) ! be 
Softens 130-140° — 
Shrinks 170-190° 190° 
Darkens — 209° 
Melts 205-215° 214° 


In order to determine if the substance were mainly galactosides we hydro- - 
lysed 0-5 g. of T, by the method described by Rosenheim [1913]. 0-34 g. of 
impure esters was obtained and this upon hydrolysis yielded 0-23 g. of fatty 
acids, having a melting point of 58°. Recrystallisation from alcohol gave a 
crop of white needles melting at 70° which after a second recrystallisation 
melted at 76°. 

A few oily drops were separated from the products of hydrolysis of T, 
and may have been the sulphates of the bases, but there was insufficient for 
examination. 

The hydrolysate, freed from esters, was examined in the polarimeter. The 
rotation, if any, was slight and laevo in direction. A minute amount of a 
methylphenylosazone which melted at 182° was obtained (methylphenyl- 
galactosazone M.P. 190°). A further amount of 7’, was extracted with pyridine 
following the method of Rosenheim [1914, 2] and yielded fractions of varying 
nitrogen and phosphorus content. There was insufficient of these for further 
examination. 

There appeared, from all the data, to be little doubt that the product 
obtained by following the method described by Robertson for the isolation 
of tethelin is a very impure mixture of substances of the lipoid class and that 
this product is capable of separation by simple means into fractions of variable 
composition. 


Feeding experiments. 

Robertson appears to be well aware of the high variability of the growth 
curves of white mice but it is his claim that, by the employment of large 
groups of animals, a composite growth curve can be constructed which will 
to a high degree of probability reproduce the characteristics of growth of the 


individual. On this statistical basis he feels entitled to draw definite and far 
reaching conclusions from small differences in such composite curves for 


TETHELIN 57 


normal and tethelin-fed groups. An examination of his methods and data 
led us to query the justice of this argument. 

For example, it is significant that, in his experiments on the effect of 
feeding egg-lecithin [1916, 6], two batches of normal mice chosen at random at 
the age of four weeks showed a difference of 1-5 g. in mean weight, 7.e. 13 % of 
body weight—a difference of the same order as the defection of weight of 
tethelin-fed mice at the maximum divergence of this group from the normal. 
Yet this, in the former case, is explained away as being due merely to the 
high variability at that age, whilst in the latter case it is reported as being 
a marked defection due to the effect of tethelin. Many other examples of 
such biassed interpretation occur throughout this series of papers but one 
further example will suffice. In support of his contention that mice fed upon 
pituitary are, size for size, heavier than those normally fed. Robertson re- 
produces a photograph of two mice of the same age and linear dimensions, 
one a normal mouse weighing 30 g. the other a pituitary-fed mouse weighing 


_ 37g. [1916, 1]. When it is pointed out that the weight of the normal was 


-about the mean for his batch whereas the other weighed 27 % more than 
‘the mean for his batch—+.e. was an extravagantly abnormal member—the 
fallacy of the argument is apparent. 

On the basis of such criticisms we submitted the author’s papers to 
Professor Karl Pearson under whose direction, Mr H. Soper, of the Depart- 
ment of Applied Statistics, University College, London, undertook the ex- 
amination of Robertson’s statistical arguments, and who has kindly permitted 
us to publish extracts from his report. gHe says “if some mice are apt to 
be 3 days in advance of or behind the others in stages of growth as seems 
_ indicated, then the curve of mean growth will not reproduce the character- 
istics of individual growths. The addition of curves at different epochs or 
phases will tend to distort or obliterate the detail.” It is significant that just 
such a possible variation in stage of growth was introduced into the experi- 
ments by the empirical grouping together of animals for weighing. Thus 
animals weighed between the 25th and 31st days inclusive were regarded as 
having been weighed upon the 28th day. A similar bracketing of ages was 
employed throughout. Mr Soper sums up in these words: “The conclusion 
to be drawn is that on the figures presented and statements made there is 
a significant defect in weight in all the treated groups from the normal group. 
The normal group was not changed and was not large in numbers and it has 
been assumed that it was a random sample of the whole community under 
observation. The figures given are not so complete as they might have been 
and statements are made and procedures taken which could only be supported 
by data not detailed.” 

It is significant that the normal group was not strictly a “random sample 
of the whole community under observation” as their growth curve antedates 
that of the pituitary mice by three weeks and that of the tethelin mice by 
three months. The stages of growth of the contrasted batches were not con- 


58 J. C. DRUMMOND AND R. K. CANNAN 


temporary and there is no evidence that undetected variations in environ- 
ment incident to the two groups at different ages were not responsible for the 
differences in growth. Robertson himself [1919] has acknowledged that the 
normal growth curve varies year by year. 

Any experiment, claiming acceptance upon statistical considerations, 
which in the methods employed introduces a possibility for variability from 
that statistical basis must be open to criticism. Finally an inherent source 
of error vitiates all the conclusions. The animals were in possession of their 
own functioning pituitary glands, the activity of which could be neither 
measured nor controlled. 

EXPERIMENTAL. 


It was not possible with the material at our disposal, nor in view of our 
previous findings was it considered worth while, to carry out feeding experi- 
ments on the scale of those of Robertson, but all the other precautions em- 
ployed by him in respect to weighing, elimination of sickly mice and general 
care were taken and the errors which it is suggested existed in his méthods 
have as far as possible been avoided. 

The mice were placed as soon as weaned in groups of six of the same sex 
in special wooden boxes and fed on a basal diet of mixed grains and dry bread 
with meat and vegetables once a week and water ad libitum. To the treated 
animals dried anterior lobe was administered in a dose of 50 mg. per mouse 
per day (equal to about 3 mg. of tethelin). This was given in a paste of 
dried milk powder and starch in equal parts and was greedily consumed before 
the day’s ration of ordinary food was given. The same quantity of the paste, 
without pituitary, was given to the normal animals. 

In the first experiment mice were taken at an age between 3-5 weeks and . 
divided into batches of 25 normal and 25 pituitary-fed mice of each sex. The 
individual growth curves were plotted until maturity. In view of the un- 
certainty of age no composite growth curves were constructed but it was 
thought that a careful inspection of such groups of individual curves would 
reveal the general characteristics of the two rates of growth and the distortion 
of the normal curve found in Robertson’s tethelin curve would be detectable. 
No general tendency towards a differentiation of the shapes of the two batches 
of curves was to be observed, a close inspection merely emphasising the high 
variability of growth. 

In a second experiment mice were observed from birth so that composite 
curves might be drawn. The mothers of the treated mice were fed with 
pituitary until the young were weaned, when the latter were removed and 
treated as in the first experiment. The experiment was abandoned at 13 weeks 
owing to pituitary supply having become exhausted but as Robertson reports 
marked effects from 7 weeks onwards a comparison was possible. 

These growth curves show no characteristic differences. Admittedly the 
curves are built up from a statistically too limited number of mice but, in 
conjunction with the other work reported their negative result is significant. 


TETHELIN 59 


Male mice (ten in each group) Female mice (ten in each group) 
Mean weight. g. Mean weight. g. 
Age Normals _— Pituitary-fed Normals _ Pituitary-fed 

Birth 1-60 1-44 1-53 1-50 

1 week 3°92 4-20 4-00 4-10 

2 weeks 5-55 5-90 5-80 5-75 

rn 7-44 8-00 8-02 8-30 
a 9-70 10-0 10-1 10-1 
a 12-0 12-1 12-4 12-0 
ae 14:3 14-0 14-3 13-4 
ie 15-1 15-6 15-4 14-8 
oe 16-0 16-9 16-4 16-0 
oe 17-2 18-0 17-6 17-6 
Seem , 1&7 19-5 19-0 18-4 
ses 19-8 20-4 19-6 19-0 
B's 21-1 21-4 20°3 19-8 

CoNCLUSIONS. 


1. The product obtained by the method described by Robertson for the 
isolation from the anterior lobe of the pituitary gland of the substance he 
has called tethelin is a very impure mixture of substances of the lipoid class. 

2. Criticisms are advanced of the deductions made by Robertson as to 
the effect of tethelin and of anterior lobe of the pituitary gland upon growth 
and experiments are reported which fail to point to any influence upon the 
growth of mice of oral administration of anterior lobe of the pituitary gland. 


REFERENCES. 


Aldrich, T. B. (1912). Amer. J. Physiol. 30, 320 and 437. 
Cerletti, U. (1907). Arch. Ital. Biol. 47, 123. 
Cushing, H. (1909). J. Amer. Med. Assoc. 53, 251. 
Goetsch (1916). Bull. John Hopkins, 27, 29. 
Marinus, C. J. (1919). Amer. J. Physiol. 49 238. 
Pearl, R. (1916) J. Biol. Chem. 24 123. 
Robertson, T. B. (1916, 1). J. Biol. Chem. 24, 385. 
—— (1916, 2). J. Biol. Chem. 24, 397. 

—— (1916, 3). J. Biol. Chem. 24, 409. 

—— (1916, 4). J. Exp. Med. 23, 631. 

—— (1916, 5). J. Amer. Med. Assoc. 66, 1009. 
—— (1916, 6). J. Biol. Chem. 25, 647. 

—— (1919). J. Biol. Chem. 37, 377. 

—— (1921). Biochemistry. 

Rosenheim, O. (1913). Biochem. J. 8, 604. 

—— (1914, 1). Biochem. J. 7, 121. 

—— (1914, 2). Biochem. J. 8, 110. 

Sandri, 0. (1909). Arch. Ital. Biol. 51, 337. 
Schafer, E. A. (1912). Quart. J. Exp. Phys. 5, 203. 
Wulzen, R. (1916). J. Biol. Chem. 25, 625. 


XII. THE ESTIMATION OF PECTIN AS CALCIUM 
PECTATE AND THE APPLICATION OF THIS > 
METHOD TO THE DETERMINATION OF THE 
SOLUBLE PECTIN IN APPLES. 


By MARJORY HARRIOTTE CARRE anp DOROTHY HAYNES. 


Department of Plant Physiology and Pathology, Imperial College 
of Science and Technology. 


(Received January 18th, 1922.) 


In order to follow the changes which take place in the pectic constituents of 
stored apples, a number of preliminary investigations have been found to be 
necessary. In the first place a method was required by which pectin could 
be estimated accurately in dilute solution. This having been established, it 
has become possible to proceed to an investigation of the means by which 
the various pectic constituents of the fruit can be separately extracted. The 
present communication, in so far as the extractions are concerned, is confined 
to an examination of methods for extracting the soluble pectin of fruits. The 
difficulties which are encountered in the course of this extraction are almost | 
entirely mechanical—difficulties inseparable from the process of washing out 
this soluble colloidal material from the insoluble colloids of the cell wall with 
which it is intimately associated. The clearing up of these mechanical diffi- 
culties is a necessary preliminary to any further investigation of methods of 
extraction, for unless suitable methods of washing out are adopted much 
soluble pectin may remain behind to form a mixture with the pectin dissolved 
out by subsequent treatment. 

It has been the general practice hitherto to estimate the pectin content 
of solutions by precipitating the pectin with alcohol. This method at its best 
is inconvenient and lacking in accuracy. The pectin obtained is necessarily 
of variable composition, since pectin probably exists in a number of forms 
intermediate between neutral pectin and pectic acid, Moreover, as dilution 
increases precipitation becomes increasingly difficult, until at very low con- 
centrations no precipitate is obtained, even after prolonged standing with a 
large excess of alcohol. By the method now adopted the difficulty of precipi- 
tation is avoided, for calcium pectate can be made to flocculate from solutions 
of very low concentration, and a product of definite chemical composition is 
obtained. A comparison of the results of precipitating pectin with alcohol 
and as calcium pectate is given in the sequel. 


ESTIMATION OF PECTIN 61 


THE PRECIPITATION OF CALCIUM PECTATE. 


Since calcium pectate is a colloid, its state of aggregation varies greatly 
with the conditions under which it is precipitated, and in order to obtain a 
product which can be easily manipulated it is requisite to adjust these con- 
ditions carefully. Great difficulties were experienced in the early stages of 
this investigation by the occurrence of a sticky condition in which the pre- 
cipitate could only be washed and filtered with extreme difficulty. This has 
been found to be the result of precipitating in alkaline solution, in which case 
calcium hydroxide is absorbed by the gel [Haynes, 1914]; if the period of 
hydrolysis with sodium hydroxide is too prolonged, or if the alkali is too 
concentrated, a similar result tends to be produced. On the other hand, too 
small an excess of alkali above the theoretical amount produces unsatisfactory 
coagulation, and as the concentration of pectin is increased the excess of 
alkali required increases also. In the opinion of the writers, both effects are 
probably a consequence of the tendency which alkalies exhibit to replace the 
loosely combined water molecules of the pectin sol. 

To avoid these difficulties the precipitation of calcium pectate is carried 
out in the following stages: 


(1) Hydrolysis with sodium hydroxide. 
(2) Acidification with acetic acid. 
(3) Addition of calcium chloride. 


Since no independent method exists for checking the results of pectin esti- 
mations, it has been necessary to carry out series of experiments and to 
ascertain how the weight of the calcium pectate varies when variations are 
made from a standard procedure arbitrarily adopted. In most cases deter- 
minations of the ash content of the precipitate calculated as calcium were 
also carried out. Since the number of interdependent factors is considerable, 
this has involved a large amount of work. No useful purpose would be served 
_ by giving a complete account of the details of this preliminary investigation, 
but some remarks on the successive stages of the process are required, and the 
results of a few experiments will be quoted in support of the statements made. 
The results of ash determinations require more detailed discussion, and will 
be treated separately below. The pectin solutions used for the investigation 
were in almost every case obtained from apple. Many determinations were 
_ made directly on the liquid obtained by expressing and washing out apple 
pulp, previously killed by freezing; but confirmatory series were also carried 
out on pectin from a similar source which had been precipitated by alcohol, 
re-dissolved and allowed to stand under toluene until suspended impurities 
were deposited. The juice in each case was heated immediately after pressing, 
to destroy pectase. It was found convenient to work with a quantity of 
solution giving from 0-02-0-03 g. of calcium pectate, and to dilute to a con- 
centration of approximately 0-01 % before precipitation. Larger quantities 


62 M. H. CARRE AND D. HAYNES 


give so great a bulk of precipitate that it is difficult either to wash thoroughly 
or to dry satisfactorily. 

Hydrolysis. Fellenberg [1918] has stated that pectin hydrolyses very 
rapidly in the presence of a small excess of alkali. The following series show 
the general results obtained in the course of the present investigation. 


Weight of calcium pectate obtained 
A 


‘Serres I SERIES ie 
(Concentration of NaOH) (Concentration of NaOH). 
N/100 N/50 


Time — A A 


10 mins. Very little 0-029 0-026 
S055 0-031 0-0285 0-0315 0-032 
| 0-028 0-0285 0-0305 0-0305 
+ hrs. 0-028 0-028 — oe: 
24 0-028 0-028 0-0295 0-030 
48 o°5 0-028 0-0275 — — 


It-will be observed that the rate of hydrolysis increases very rapidly with 
the concentration of alkali. As it is advisable to keep this concentration low, 
it is necessary to allow the mixture to stand at least an hour. The best results 
have been obtained by leaving it overnight, in which case the precipitate 
flocculates readily and is easy to wash and filter. It is noticeable that rather 
high results were obtained in both series for the half-hour period. This is a 
very frequent result of incomplete coagulation which renders effective washing. 
extremely difficult. 

Acidification and addition of calcium salt. Calcium pectate can be boiled” 
with water without change, but with dilute acids it tends to pass into colloidal 
solution. The conditions of this change have not been fully investigated, but 
it appears that prolonged boiling reverses the peptising effect of acid. This 
is shown by the results of the following series of experiments carried out on 
equal quantities of pectin precipitated as calcium pectate. 


Weight of calcium pectate 
Treatment “ 


g ~ 
Boiled with water 0-0275 0-028 
Boiled with N/2 acetic acid 5 mins. 0-0205 0-0225 

” ” ” 0 ” 0-0215 0-024 
” ” ” 60 ” 0-026 0-0265 
Boiled with N/10 acetic acid 5 ,, 00245 0-027 
” ” ” Je 0-0285 0-028 
” ” ” 60 ” 0-026 0-023 


The precipitates which had been boiled with acid were found to contain 
a low percentage of ash. The action of acid must therefore be primarily 
attributed to the decomposition of calcium pectate; this being so, it is not 
surprising to find that an excess of calcium salts tends to reverse the effect 
and that calcium pectate may be boiled with dilute solutions of weak acids 
in the presence of a sufficient excess of calcium chloride without change of 
weight. The amount of calcium salt required increases very rapidly with the 
amount of pectin present. This is shown by a comparison of the two following 
series of precipitations: 


ESTIMATION OF PECTIN 63 


Precipitation carried out with: 100 ce. V/10 NaOH; 100 ce. N/1 acetic acid; x cc. M/1 CaCl,: 


Serres IT 
Serres | 400 ce. apple pectin as 
200 ce. apple pectin used in Series I 
Ce. sit CaCl, - A > cr A ~ 
ded (1) (2) (1) (2) 
2 0-0300 0-0300 0-0565 0-0510 
5 0-0300 0-0290 0-0560 0-0575 
10 0-0295 0-0300 0-0600 0-0555 
25 0-0305 0-0310 0-0605 0-0600 
50 0-0300 0-0300 0-0605 0-0600 
100 — — 0-0605 0-0615 


Effect of acid concentration. 
Precipitation carried out with: 100 cc. N/10 NaOH; « ce. acetic acid; 50 cc. M/1 CaCl,: 
II 


Series I SERIES 
200 ce. dilute apple 400 cc. dilute apple 
pectin pectin as in t 
Ce. acetic z A re A = 
N/10 (1) (2) (1) (2) 
2 0-1042 0-1270 — — 
10 0-0375 0-0365 — — 
20 0-0300 0-0305 — — 
50 0-0305 0-0290 — _— 
100 0-0300 0-0295 — 
N/l 
15 0-0310 0-0290 0-0645 0-0660 
30 0-0300 0-0295 0-0615 0-0620 
50 0-0300 0-0300 0-0610 0-0615 
100 0-0300 0-0300 0-0605 0-0610 


It will be seen that in the more concentrated series II the lowest limit of 
acid is only reached at 50 cc. of N/1 acetic acid, whereas in series I, 20 ce. 
-N/10 acetic acid are seen to be sufficient. The higher values in both series 
are due to the carrying down of calcium carbonate and occluded impurities 
which are only removed by excess of acid. 

It appears from these results that for ordinary concentrations of pectin 
the optimum concentration of calcium chloride is approximately M/1, and 
of acid N/5. Where the quantity of pectin is very small it has been found 
advisable to reduce the acid concentration to N/10, since the larger amount 
tends to produce stickiness. In this case the quantity of calcium chloride 
should be correspondingly reduced. 


METHOD OF ESTIMATING PECTIN AS CALCIUM PECTATE. 


As a result of these investigations the following method of estimating 
pectin has been adopted: 

A quantity of pectin is taken which will yield from 0-02 to 0-03 g. of 
calcium pectate; this is neutralised and then diluted to a volume such that 
after addition of all reagents the total volume measures about 500 cc. 100 ce. 
of N/10 NaOH are then added and the mixture is allowed to stand at least 
an hour, but preferably overnight. 50 cc. of N/1 acetic acid are then added, 
and after five minutes 50 cc. of M/1 calcium chloride. The mixture is then 
allowed to stand for an hour, after which it is boiled for a few minutes and 
filtered through a large fluted filter. If the precipitation has been properly 
carried out, filtration should take place very rapidly and subsequent washing 


64 M. H. CARRE AND D. HAYNES 


should be easy. The washing is continued with boiling water until the filtrate 
is free from chloride, after which the precipitate is washed back into the 
beaker, boiled, and filtered again. It is then again tested for chloride, and 
these processes are repeated until the filtrate from the boiled precipitate gives 
no indication of chloride with silver nitrate. It is then filtered into a small 
fluted filter, from which it can be transferred to a dish and finally to a Gooch 
crucible which has been previously dried at 100°. The precipitate is dried to 
constant weight at 100° which has been found to require about 12 hours. 

If the quantity of pectin is increased, the quantities of soda and calcium 
chloride should be correspondingly increased. If very small quantities of 
pectin are dealt with, the acid and calcium chloride should be reduced. 

Example of estimations using dilute apple pectin. 


Precipitations carried out with: 50 cc. pectin solution; 50 cc. N/10 soda; 100 cc. N/10 acetic 
acid; 100 cc. M/1 CaCl,: 


Series I Sertzs IT 
Purified pectin solution Unpurified pectin solution 
0-0236 . 00237 0-0280 0-0285 
0-0237 0-0236 0:0287 0-0300 
0-0240 0-0231 0-0265 0-0275 
0-0240 0-0230 0-0280 0-0280 
0-0229 0-0233 0:0280 0:0285 
Mean =0-0235 Mean =0-0282 


The quantity of acid used for these estimations is less, and the quantity of CaCl, greater, 
than prescribed in the text. It has been found that the alteration of these amounts facilitates 
the washing of the precipitate. 


If precipitated according to the method given above, calcium pectate is 
so insoluble that quantitative precipitation can be carried out in solutions of 
extreme dilution, the limiting factors being rather those governing the drying 
and weighing of very small quantities of colloidal material. It is thus possible 
to carry out estimations by the calcium pectate method over a large range of 
concentrations at which precipitation by alcohol is impossible. The lower 
limit of concentration at which pectin can be even partially precipitated by 
alcohol has been found to be 0-06 %. 


Example of method applied to very dilute solutions. 


Precipitations carried out with: 50 cc. dilute pectin solution (apple extract); 50 cc. N/10 
NaOH; 50 cc. M/1 CaCl,; 50 cc. N/1 acetic acid: 


Weight of calcium pectate 0-0053 00055 
0-0050 0-0055 
0-0050 00053 
00055 00050 
Mean =0-00525 


A few examples of estimation by precipitation with alcohol are given 
below for comparison. They are all carried out on similar quantities (10 ce.) 
of apple juice to which a small quantity of water extract has been added. 

Weight of precipitate 
Sens aa 


(1) ~ (2) 


25 cc, of alcohol 0-019 0-021 
50 ” ” (1) 0-018 0-015 
(2) 0-012 0-015 


100 ,, » 0-018 0-020 


ESTIMATION OF PECTIN 65 


The weights of calcium pectate obtained from 10 cc. of this extract were: 
(1) 0-023; (2) 0-023. It is probable that all the alcohol precipitations are too 
low. The solubility of pectin in alcohol is shown clearly by the following 
series of results, carried out on the same material as above: 

(1) The weight of precipitate obtained by adding 50 cc. of alcohol to 
10 cc. of extract diluted with 25 cc. of water was found to be (1) 0-019; 
(2) 0-020. 

(2) With 50 ce. of water no precipitate was obtained in two days in either 
sample. After this a small precipitate was obtained from one, but not from 
the other. 

(3) With 100 cc. of water no precipitate was obtained after standing for 
a week or even after doubling the amount of alcohol. 

These experiments have not been repeated on purified pectin owing to 
lack of material. Such experiments would probably give more regular results, 
since in ordinary juice sugar acids and other substances are present which 
may themselves tend to favour coagulation. 

In the foregoing account it has been assumed that the method of pectin 
estimation there described is dependent upon the production of a definite 
chemical compound—calcium pectate. This assumption obviously needs 
justification, more especially since pectin is a colloidal substance which is 
likely to form adsorption compounds and is known to form solid solutions 
with alkalies and akaline earths [Haynes, 1914]. 

Pectic acid was among the first of the pectic compounds to be described ; 
indeed the formation of pectic acid by hydrolysis still serves to define the 
class of substances known as pectins. The chemical nature of pectic acid is 
still obscure, but it is universally recognised as possessing definite acidic 
properties, and it has been found to possess the same chemical composition 
expressed by the empirical formula (,,H,,0,, [Schryver and Haynes, 1916]? 
when derived from very different sources (strawberry, turnip, and rhubarb 
stalks). 

The pectic acid from apples alone showed a slight variation, and the 
present work indicates that this may very possibly have been due to the 
presence of impurities. 

Fellenberg [1918] assigns the formula C,.Hyg0;. (CO,CHg)s to pectic acid 
on somewhat arbitrary grounds, and his own analyses of barium salts agree 
more nearly with the empirical formula given above. 

The number of decomposition products obtained by Ehrlich and others 
from pectin indicates, however, that the molecule of pectin is large, and it is 
therefore hardly likely that pectic acid can be represented by a simple multiple 
of C,,H,,0,,. It probably differs from a simple multiple by some small 


1 In this paper the name pectin was used as the equivalent of pectic acid. There is however 
an increasing tendency to reserve the term pectin for soluble products, especially those occurring 
naturally in fruits, and to denote by. pectic acid the last term in a series of pectic compounds 
obtained by hydrolysis of these soluble substances, This usage is advocated by Fellenberg, and 
is followed in the present paper. 


Bioch. xv1 5 


66 M. H. CARRE AND D. HAYNES 


quantity but the true formula cannot be ascertained with any certainty until 
the constitution of pectic acid is established; in the meantime it seems ad- 
visable to adopt with reservation the formula given above, and it is to this 
formula, assumed to be that of a dibasic acid that the results of the analyses 
given below are referred. 


THE COMPOSITION OF CALCIUM PECTATE OBTAINED FROM APPLES. 


Two methods of analysis have been employed: 
(1) Ignition to CaO. 
(2) Conversion of this to sulphate. 

The following results were obtained by the ignition of calcium pectate 
obtained from a purified solution of apple pectin which contained at most 
0-01-0-03 % of ash when precipitated by alcohol. When filtered carefully 
before precipitation the amount of ash was negligible. | 


Weight of Ca pectate Weight of CaO Percentage Ca 
0-0160 0-0017 7-59 
0-0226 0-0025 7-90 
0-0270 0-0029 ~~ 7-66 
0-0266 0-0028 7-51 
0-0746 0-0080 7-65 
0-0308 0-0033 7-64 
Mean =7-658 


In order to afford a comparison of the results of the two methods of ash 
determination, the following series are given. These were carried out on 
material similar to that used for the previous series. The calcium pectate was 
obtained by exactly similar methods of estimation in each case. 


As CaO As CaSO, 

Weight of Ca Percentage “Ca Percentage 
pectate CaO Ca pectate CaSO, Ca 
0-0746 0-0080 7-65 0-0240 0-0062 7-60 
0-0266 0-0028 7-52 0-0250 0-0065 7-65 
0-0270 0-0029 760° °° 0-0236 0-0062 7-70 
0-0260 0-0017 7:59 0-0234 0-0062 7:65 
0-0226 00025 7:70 00240 0-0064 7:67 

Mean =7-612 Mean = 7-654 


Mean of the two series = 7:63. 


It is evident that these results agree very closely with the theoretical per- 
centage, 7-66, and the same percentage has been obtained repeatedly from 
calcium pectate precipitated under widely different conditions, it may there- 
fore be regarded as established that calcium pectate obtained by this method 
is a definite chemical compound, the empirical formula of which approximates 
closely to C,,H90,,Ca. 

Since estimations of pectin have usually to be made in a water extract 
containing sugars, salts and other compounds, it has been nécessary to 
ascertain how nearly the composition of the precipitate from this unpurified 
material corresponds with theory. It has been found that the ash is usually 


ESTIMATION OF PECTIN 67 


rather high, but that if sufficient precautions are taken the error can be 
reduced to an amount which probably seldom exceeds 1 °% of the total ash. 
The following are representative examples, each set being estimated on 
different material : a 
. Ca pectate CaSO, Percentage Ca 


I 0-0243 0-0062 7-50 

0-0293 0-0076 7-63 

I 0-0553 0-0144 7-66 

0-0577 0-0173 8-30 

0-0294 - 0-0077 7-70 

Ill 0-0514 0-0138 7-90 

O-O151 0-0040 7-79 

2 0-0503 0-0129 7:54 
; 0-0425 0-O111 7-68 


In the second set, which gives the highest ash, much iron was found to 
be present. It must be borne in mind that an increase in the weight of ash in 
the calcium pectate precipitate may be due either to the carrying down of 
suspended mineral matter or to the occlusion of organic substances such as 
the salts of organic acids. These latter if present may introduce considerable 
error, but if the juice is carefully filtered and if the precipitation is carried 
out in the presence of sufficient acid, it has been found that the difference © 
can be reduced to a very small amount, which as is shown below may almost 
certainly be regarded as mainly of mineral origin. 

It is significant to find that the impurities associated with pectin vary 
with different stages of pressing. This is illustrated by the following estima- 
tions: 


Ash content of calcium pectate obtained from successive 
extractions of apple. 


Number of Proportion of total Ca = Ash (calculated as 
3 pressings pectate present percentage Ca) 
1-10 79-5 7-62 
10-20 14-0 9-20 
20-30 6-5 9-70 


In other cases final pressings were found to contain as much as 10-12 % 
of ash calculated as calcium. There seems much reason to think that this 
excess of ash is due to colloidal mineral matter, possibly associated with the 
protoplasm, which can only be pressed out when the cell is completely dis- 
integrated. This, together with a variable amount of colloidal iron, appears 
to constitute the principal source of the impurities carried down with the 
calcium pectate precipitated from apple juice. The mineral nature of this . 
impurity cannot be completely demonstrated without elementary analysis, 
but there seems to be very little doubt that it is largely, if not entirely, com- 
posed of inorganic material, in which case the error introduced into the 
weight of the calcium pectate precipitate will be very small. 

It may accordingly be stated as a general rule that pectin should be 
estimated on material obtained by exhaustive extraction, but that if ash 


5—2 


68 M. H. CARRE AND D. HAYNES 


determinations are made in order to ascertain whether or not the material 
affords calcium pectate in a state of comparative aie the later pressings 
should be rejected. 

A few words may also be added here as to esa precautions which it is 
advisable to take if satisfactory ash determinations are to be carried out. It 
must be remembered that pectin absorbs and carries down other substances 
very readily, and that these—once occluded—are removed only with ha 
great difficulty. 

Thus if calcium carbonate is contained in the precipitated gel, it is practi- 
cally impossible to remove it by subsequent treatment with acid. Pectin 
should therefore be precipitated in dilute solution and in the presence of 
considerable excess of acid. The excess of CaCl, should not be greater than 
is sufficient for satisfactory coagulation, and at each stage of the process 
sufficient time should be allowed for the reagents to penetrate the large 
particles of the pectin sol. 


THE EXTRACTION OF SOLUBLE PECTIN FROM APPLES. 


A considerable amount of investigation has been necessary in order to 
work out a practicable method for the complete extraction of the soluble 
pectin. After the first few extractions have been carried out, further extraction 
gives a very dilute solution, and it is easy to imagine that the whole of the 
pectin has been extracted at this stage. It has been found, however, that a 
large proportion—sometimes more than 50°,—comes out at so great a dilution 
that no precipitate is obtained from the solution even after prolonged standing 
with a large excess of alcohol. 

The process is of necessity laborious. The following has been found the 
best method of procedure: 50g. of finely minced apple are frozen in an 
efficient freezing mixture, and maintained at a low temperature for a con- 
siderable number of hours to ensure that the pulp is completely killed. The 
pulp is then warmed to the temperature of the air, and pressed through a 
cloth in a small hand press. After this the residue, which should be as dry as 
possible, is ground with fine purified sand in a mortar, and repeatedly washed 
with cold water and pressed. The residue should be well mixed after each 
addition of water. Sixty to eighty extractions are usually required, and the 
bulk of extract obtained is about two litres. Warm water should not be used, 
as this has been found to increase the amount of pectin obtained—presumably 
by promoting hydrolysis. 

After extraction the liquid is boiled to destroy the action of pectase, and 
before estimations are carried out it is filtered, first through muslin and finally 
through a fluted filter paper. An aliquot part, containing a suitable quantity 
of pectin, is then taken for estimation. The amount varies from about one- 
half in the case of rather immature apples to one-tenth in the case of apples 
juicy or over-ripe. 

The following estimations have been carried out on equal quantities of 


ESTIMATION OF PECTIN . 69 


apple previously finely minced and mixed to as uniform a condition as possible. 
They show the order of accuracy with which the process of extraction can be 
carried out. 


Weight of calcium pectate 
Inl /10 of extract of 50 g. From 100 g. of 

of apples apples 

0-0257 0-514 

0-0246 0-492 

0-0271 0-542 

0:0252 0-504 

0-0245 0-490 

Mean =0-0256 Mean =0-508 


[Note added Feb. 17, 1922.—Other series of results have been obtained 
which show similar differences. These differences appear to be due rather 
to the difficulty of mixing the material uniformly than to incomplete ex- 
traction. Thus in the following series, in which pressings were collected in 
successive fractions, there is no indication that later fractions tend to contain 
more pectin where the first fraction is relatively low. 

Weight of Ca pectate from 50 g. of apples 


“Main portion 2nd fraction 3rd fraction Total 
0-339 0-0050 25. 0-344 
345 0100 0-0060 361 
305 0085 0030 3165 
328 0070 i 3345 
325 0040 ea 329] 
SuMMARY. 


The precipitation of pectin as calcium pectate is described. It is shown 
that by a careful adjustment of conditions this can be used as an accurate 
method of analysis, and that the precipitate corresponds in composition 
closely to the empirical formula C,,H,.0,,Ca. A method is given for the 
extraction of soluble pectin from apples, and the results of a number of similar 
estimations are compared. 


REFERENCES. 


Fellenberg (1918). Biochem. Zeitsch. 85, 118. 
Haynes (1914). Biochem. J. 8, 553. 
Schryver and Haynes (1916). Biochem. J. 10, 539. 


XIII. ON THE OCCURRENCE OF MANGANESE IN 
THE TUBE AND TISSUES OF MESOCHAETO- 
PTERUS TAYLORI, POTTS, AND IN THE TUBE 
OF CHAETOPTERUS VARIOPEDATUS, RENIER. 


By CYRIL BERKELEY. 
From the Marine Hetogieat Station, Nanaimo, B.C, 


(Received January 18th, 1922.) 


A NuMBER of marine annelids excrete a mucous substance which hardens to 
form a tube in which they live. “Many of these tubes are partially soluble in 
cold dilute alkaline solutions. ~ 

In the course of an attempt to determine the nature of the material thus 
dissolved from the tube of an annelid which occurs commonly on the British 
Columbia coast, Mesochaetopterus Taylori, it was found that the substance 
precipitated on acidifying the alkaline extract yielded a bright’ blue ash on 
burning. Qualitative tests showed that this was due to manganese and led 
to the examination of the tube material and the tissues of the animal for that 
element. Preliminary quantitative tests indicated that it was present in both 
cases in unexpectedly large amounts and more careful determinations seined 
warranted. 

Traces of manganese are known to occur very peace in animal tines 
and Bradley [1910] has shown that the element is normally present in con- 
siderable amounts in the tissues of fresh-water mussels. It was also observed 
by that author that the nacre of the shells of both living and fossil Unionidae 
is rich in manganese and this, so far as the writer is aware, is the only case 
of its occurrence in quantity in an animal secretion at all comparable with 
the tube of a worm. 

All the more recent methods for the determination of manganese in organic 
material depend upon the extraction of the manganese salt from the ash and 
its oxidation to permanganic acid, which is estimated colorimetrically or by 
titration with arsenious acid or an arsenite. Reiman and Minot [1920] give 
an excellent summary of the literature of the subject and describe a method 
for determining the very small amounts of manganese occurring in human 
blood and tissues. In the present case the amount of manganese present was 
comparatively so large that the somewhat elaborate method of ashing the 
material and leaching the ash described by these authors. was unnecessary. 
In working with the tube material it was, indeed, not found necessary to use 


MANGANESE IN MARINE ANNELIDS 71 


any oxidising agent to complete the destruction of the organic matter as is 
recommended by Bradley. Direct burning at a moderate temperature pro- 
duced an ash, which, excepting some sand, was completely soluble in 35 % 
nitric acid, to which a few drops of hydrogen peroxide were added. Subse- 
quent fusion of the undissolved residue with potassium nitrate rendered no 
more manganese soluble. In the case of the tissues the addition of a small 
quantity of potassium nitrate was found necessary. With this modification 
and substitution of a solution of sodium arsenite for that of arsenious acid 
for titration of the permanganic acid, the method used has been substantially 
that of Bradley. The details of procedure have been as follows: 

0-2 to 1-0 g. of material is weighed into a procelain crucible. The bulk of 
the organic matter is burnt off at a low temperature. Potassium nitrate is 
‘then cautiously added if necessary. The crucible is then gradually brought 
to a bright red heat and maintained there for about half-an-hour. When 
cool 2 cc. of 35 % nitric acid and 5 drops of hydrogen peroxide are added to 
the contents of the crucible which is cautiously heated over a small flame. 
The acid is then diluted and decanted from the small amount of sand re- 
maining undissolved, and the latter is repeatedly leached with small quantities 
of hot water until the combined leachings amount to about 50 cc. 0-5 ce. of 
10 % silver nitrate is added to the solution which is then brought to the 
boiling point. The vessel is removed from the flame and 10 cc. of 20% 
ammonium persulphate cautiously added. The permanganic acid colour 
develops immediately. The solution is then boiled gently for five minutes 
and cooled in running water. It is then titrated with a very dilute solution 
(not stronger than 0-1 g. per litre) of sodium arsenite which has been pre- 
viously standardised against a solution of pure potassium permanganate 
reduced by sulphurous acid. An aliquot of this reduced permanganate solution 
is put through the foregoing oxidation procedure immediately before the 
sodium arsenite solution is standardised. 

In application to aliquots of the standard permanganate solution evapo- 
rated to dryness this method gave results of a high degree of accuracy for 
quantities of manganese as small as 0-00001 g. It is important to boil the 
solution sufficiently long after addition of the ammonium persulphate to de- 
compose the excess, otherwise too high readings are obtained. The presence 
of chlorides is stated by some authors to.affect the accuracy of determinations 
of manganese depending upon the principle involved in this method. This 
has not proved to be the case in the work to be described. Traces of chloride . 
were invariably indicated by the precipitation of silver chloride on adding 
the silver nitrate to the solution under analysis, but this dissolved on treating 
with ammonium persulphate and did not affect the determination. No appre- 
ciable difference in result was obtained whether the ash were heated with 
sulphuric acid, to drive off hydrochloric acid, before analysis or not. The 
figures recorded in Table I illustrate this point and indicate the degree of 
accuracy attained in duplicate samples. 


72 C. BERKELEY 
Table I. M. Taylori. 


Material Weight taken Manganesefound’-. Treatment 
g- % 
Tube 0-1592 0-0276 Ash not treated with sulphuric acid 
PA 0-3426 0-0305 me vk 
0-2000 0-0300 Ash treated with sulphuric acid 
+p 0-2176 0-0253 “fs % 
Posterior region 1-0690 0-0015 Ash not treated with sulphuric acid 
ms s 0-7098 0-0015 Ash treated with sulphuric acid 


Closer agreement than is indicated could not be expected owing to the 
difficulty of obtaining uniform samples. The material of the tube does not 
lend itself to very fine grinding and, as is subsequently shown (Table IT1), 
the different regions of the tube differ considerably in manganese content, so 
that such fluctuations as appear may easily be due to the regions being repre- 
sented in varying proportion in the samples. 


Determinations in individual tubes of Mesochaetopterus Taylori. 


It is extremely difficult to dig up the tube of M. Taylori entire. It usually 
runs down more or less vertically through the fine sand for one to two feet 
and then takes abrupt turns in a more horizontal direction through the coarser 
material underneath. Almost invariably, therefore, unless very special care 
is taken, the tube is cut off short of the closed end. The tubes of which the 
analyses are given in Table II were the longest which could be selected from 
a large number dug, but only in the case of No. 1 was the terminal portion 
present. They were freed from adherent sand as far as possible, but varying 
quantities remained imbedded in the tube material. This was determined by 
weighing the residue after the ash had been completely extracted with nitric 
acid and water. 


Table II. Tube of M. Taylori. 


Serial Weight of Manganese Manganese in Manganese in 
number material Sand found material material — sand 
g. % mg. % % 
1 08402 45:9 0-297 0-0353 0-0654 
2 0-3209 18-4 0-0715 0-0223 0:0273 
3 0-2600 13-9 0-0132 0-0508 0:0589 
4 02248 15:7 0-0385 0-0171 0-0203 
5 0:1427 22-0 0-0495 0-0347 0:0445 


The results show a considerable variation. It is noteworthy that although 
No. 3 gives the highest result on the entire material, this is in part due to its 
low content of sand and that, if calculation be made on the material minus 
sand, No. 1, in which the largest proportion of the basal end of the tube was 
present, gives the highest figure. This suggested that the basal end of the tube 
might be richer in manganese than the rest and that the variation in results 
was to some extent due to the varying proportion of this end of the tube 
represented in the material analysed, 


MANGANESE IN MARINE ANNELIDS 73 


Determinations in various regions of the tube of M. Taylori. 


The tube can be easily divided transversely into three distinct regions 
corresponding roughly to the anterior, median and posterior portions of the 
animal. The uppermost, most of which projects above the surface of the 
sand bed, is composed of very thin papery material and is almost white. In 
the median region the wall is thicker and pale brown, whilst that of the 
posterior region is comparatively thick and frequently two or three layered. 
It is much darker in colour than the rest of the tube and often contains 
thickenings where the tube has been broken and repaired. Material from each 
of these regions was separated from a number of tubes and duplicate deter- 
minations made from the bulked samples. The results are given in Table III. 


Table III. Regions of tube of M. Taylor. 


Weight of Manganese Manganese in Manganese in 
Region material Ash found material material — ash 
g- % mg. % % 

Anterior 0-2520 49-2 0-011 0-0043 0-0086 
. 0-2126 46-4 0-011 0-0051 0-0096 
Median 0-2746 39-9 0-0265 0-0096 0-0160 
i. 0-3584 36-7 0-0385 0-0107 0-0169 
Posterior 03426 26-7 0-1199 0-035 0-0475 
¥ 0-4250 27-2 0-1375 0-0326 0-0444 


The duplicate samples agree as well as could be expected having regard 
to the difficulty of drawing uniform samples of material. The results show a 
marked difference in the three regions, the posterior being the highest and the 
anterior the lowest. The ash content varies in the inverse sense. The greater 
part of the ash was sand, but separate determinations of sand were not made 
in these cases. Calculated to the material minus ash there is still a very large 
difference between the manganese content of the three regions. There is no 
doubt this accounts in some measure for the irregularities shown in the de- 
terminations in individual tubes (Table IT) since, as has been pointed out, 
these were not complete and it was in the proportion of material drawn from 
the posterior portion of the tube that they chiefly differed. 


Determinations in tissues of M. Taylori. 


The question whether the tissues of the worm itself contain manganese in 
notable quantity arises naturally out of its occurrence in the tube. The body 
of the worm is divided transversely into three distinct regions like the tube. 
For analysis these three regions were treated separately. They were removed 
from a large number of worms and kept in alcohol until hardened. The 
material was then dried, ground as finely as possible and duplicate samples 
from each region taken for analysis. The results are given in Table IV. 


74 C. BERKELEY 


Table IV. Tissues of M. Taylorv. 


Weight of Manganese Manganese in 
Region material found material 
g- mg. % 
Anterior 0-7067 0-044 0-0062 
= 0-7653 0-04.95 0:0065 
Median 0-4985 0-077 0-015 
iS 0-4493 0:0385 0-:0086* 
Posterior 1-069 0-016 0-0015 
= 0-7098 0-011 00016 


* This result is certainly too low. No potassium nitrate was used in ashing and combustion 
was therefore not complete. 


Manganese is present throughout the body of the worm, but in smaller 
proportion than in the tube. It is interesting to note that the median portion 
of the body contains most manganese since the chief glands which secrete the 
mucous material from which the tube is constructed are situated in this region. 


Determinations in tube of Chaetopterus variopedatus. 


C. variopedatus is the best known member of the family Chaetopteridae, 
but does not occur on our coast. A few tubes of this worm were obtained 
from Woods Hole, Massachusetts, for comparison with those of M. Taylori in 
respect of manganese content. The tube is a great deal wider in relation to 
the size of the inhabitant than in the case of the latter species and has both 
ends open. These both protrude above the surface of the muddy sand in which 
the animal lives. With the exception of these small and narrowed ends the 
material of the tube is very similar to the posterior portion of that of 
M. Taylori. It was found possible to divide the tube material longitudinally 
into two layers and it seemed of interest to determine manganese in the two 
layers separately. The inner one was smooth and clean and of a somewhat 
darker colour than the outer which was rough and heavily coated with sandy 
mud. This was cleaned as carefully as possible and the two lots of material 
dried separately. The results are given in Table V. 


Table V. Tube of C. variopedatus. 


Weight of Manganese Manganese in Manganese in 
material Ash found material material — ash 
g. % mg. % % 
Inner layer 0-3278 24-95 1-501 0-457 0-610 
‘ s 0-3042 28-27 1-314 0-431 0-602 
Outer layer 02806 49-14 0-66 0-235 0-462 
= 3 0°3021 45°81 0-676 0-224 0-410 


The tube of C. variopedatus is thus more than ten times as rich in man- 
ganese as even the posterior portion of that of M. Taylori. The inner layer 
contains about twice as much as the outer one, but this is in part due to there 
being more sand in the latter. Expressed on ash-free material the ratio is 
about 3:2. This, coupled with the observation that manganese occurs in 
comparable quantity in the tissues of the allied species M. Taylori, suggests 
very strongly that the manganese in the tube is secreted by the worm and is 


MANGANESE IN MARINE ANNELIDS 75 


not due to the direct action of sand or sea water on it. It is probable, on the 
contrary, that these agencies tend to remove manganese from the outer 
layers of the tube since the sand lining the cavity from which a tube has 
been dug is invariably stained dark brown and looks precisely as if a rusty 
iron bar had been withdrawn. This rustiness is almost certainly due to oxides 
of iron and manganese resulting from the decomposition and wearing of the 
outside of the tube by sand and water. 

In every case in which the test has been made iron has been found accom- 
panying manganese in the ashes of the materials dealt with in this paper. 


Examination of tubes of other worms for manganese. 


- The tubes of three other worms which occur commonly in this neighbour- 
hood were examined for manganese; those of two Sabellids and a Spiochaeto- 
pterus. It was present in just sufficient quantity to be determined in the 
material from one of the sabellid tubes. No trace could be detected in that of 
the other sabellid or of the spiochaetopterus using quantities of material up 
to 0:75 g. The last case is particularly interesting because, not only is the 
animal nearly allied genetically to the chaetopterids in the tubes of which 
manganese occurs, but also it is always found in the sand beds living im- 
mediately alongside M. Taylori. We seem to have here a very marked instance 
of a physiological divergence in species taxonomically and ecologically similar. 


Discussion. 


As an outcome of the observations recorded the question of the source 
and function of manganese in the tubes of worms naturally arises. It is fairly 
clear, for reasons which have already been advanced, that the manganese 
arrives in the tube from the worm outwards and not from the sea water or 
sand inwards. Whilst it is conceivable that the worms concentrate manganese 
from the sea water direct it is not probable, because the quantity of the 
element in sea water is relatively very small indeed and the worms do not 
pass large quantities of water through their bodies, but strain their food 
material out of it. Enders [1909], who has carried out a detailed study of the 
life history and habits of C. variopedatus, holds that, in the case of that species, 
any sand which is suspended in the water taken by the worm into its tube is 
also discarded and that only diatoms and small animals are taken into the 
digestive tract. This does not, however, seem to be the case with M. Taylors, 
since a considerable quantity of sand was found in the posterior region of the 
animal. A sample of the diatoms which are normally to be found in the water 
covering the sand beds in which this worm lives, and which probably constitute 
a large part of its food, was obtained by dragging a fine net over beds of 
eel-grass growing in the immediate neighbourhood. The diatoms were sepa- 
rated by fractional sedimentation from the small animals and debris accom- 
panying them, dried at 100° and ash and manganese determined. 60-8 % of 


76 C. BERKELEY 


ash containing 0-007 % of manganese was found. The manganese was all 
readily extractable from the ash by 35 % nitric acid containing hydrogen 
peroxide. No further manganese could be detected after fusing the residual 
silica with sodium carbonate, so that it appears to be contained in the organic 
portion of the diatom. Here, therefore, we have an immediate source of 
manganese readily available to the worm. It is, however, not impossible 
that it is also derived from the sand which, in the case of M. Taylori, un- 
doubtedly passes through the digestive tract in considerable quantity. 

Manganese was determined in the sand in which the worms were living 
by sodium carbonate fusion followed by application of the method previously 
used, and 0-158 °%% found present. About a fifth of this was directly soluble 
in 35 % nitrie acid and was therefore not in silicate combination. The sand 
passed through the animal therefore provides an ample source of manganese 
and its utilisation does not necessitate the assumption of any power of silicate 
digestion. 

The presence of manganese in organisms has generally been associated 
with respiration. No evidence is at hand to show that it is to be connected 
with this activity in worms. It is difficult to see how its presence in the tube 
can have any connection with respiration, and, in the absence of any evidence 
to the contrary, it seems most rational to assume that its presence there is 
of no physiological significance, but is due to the worm excreting its super- 
fluous manganese with the tube-building material, a theory which is sup- 
ported by the observation of manganese in larger quantity in the lining than 
in the outer layers of the tube of C. variopedatus, and in the mucous secreting 
region of the body of M. Taylori, more than in the other regions. The presence 
of more manganese in the posterior than in the other portions of the tube of 
the latter species is also explained on this assumption since this is the oldest 
part of the tube and has been most frequently relined and repaired. Bradley 
[1910] has shown that the excreta of fresh-water mussels living on their 
normal manganiferous diet contain the element in about the same proportion 
as it occurs in their food, so that a state of manganese balance is maintained— 
only if they are deprived of manganiferous food is the metal definitely retained 
by the animal. It is likely that a similar state of affairs exists in the case of 
worms having manganiferous tubes, but that some of the superfluous man- 
ganese passes into the tube material. The deposition of manganese in the 
nacre of the shell of the Unionidae observed by Bradley is no doubt a parallel 
case. 


SuMMARY. 


1. Manganese occurs throughout the tube of M. Taylori, but in much 
greater quantity in the posterior portion than in the other parts. 

2. Manganese is present throughout the body of M. Taylori, but in smaller 
quantity than in the tube. The median region of the body, in which the chief 
mucous secreting glands are situated is richest in manganese. 


MANGANESE IN MARINE ANNELIDS 77 


3. Manganese occurs in far larger quantity in the tubes of C. variopedatus 
than in those of M. Taylori. The lining of the tube contains much more man- 
ganese than the outer layer. 

4. The tubes of local species of Sabellid and Spiochaetopterus contain no 
manganese. 

5. Manganese is probably present in worm tubes as a waste material 
excreted by the worm with its tube building substance. It may be derived 
either from diatomaceous food or from sand passing through the digestive 
tract with the food, or from both. 


REFERENCES. 


Bradley (1910). J. Biol. Chem. 18, 237. 
Enders (1909). J. Morphology, 20, 479. 
Reiman and Minot (1920). J. Biol. Chem. 42, 329. 


XIV. MAMMARY SECRETION. III. 


1. THE QUALITY AND QUANTITY OF DIETARY PROTEIN. 
2. THE RELATION OF PROTEIN TO OTHER DIETARY CONSTITUENTS. 


By GLADYS ANNIE HARTWELL. 


From the Physiological Laboratory, Household and Social Science Diparteaci 
King’s College for Women, Kensington, London. 


Thesis approved for the Degree of Doctor of Science in the 
University of London. 


(Received December 21th, 1921.) 


INTRODUCTORY AND HISTORICAL. 


In a previous paper [ Hartwell, 1921, 2], it was shown that large quantities 
of protein in the diet of a lactating rat were detrimental to the young. The 
diet consisted of bread and protein in the proportion 15-0 g. bread to 5-0 g. 
protein. An obvious criticism was that such a diet is deficient in fat and 
vitamins, but it was pointed out that it was unlikely the bad effects were 
due to deficiency since the rat could bring up a healthy litter on a diet of 
white bread alone, a diet equally deficient. 

It is generally held that the diet of a nursing mother should be rich in 
protein, because young growing animals need large amounts of protein in 
their diet. It is probable that if large quantities of protein are ingested, some 
other factor must be present in the diet to safeguard the young from such 
harmful effects as those previously noted. The object of the present investi- 
gation was, therefore: 

1. To find out what proportion of protein can be fed to the mother 
without danger to the offspring. 

2. To determine what other factor or factors should be included in the 
diet in order that the large quantity of protein may be effectively metabolised. 

It is probable that not only should the diet be physiologically complete, 
but that a greater quantity of one or more constituents might be necessary 
when the intake of protein is high. For instance, when large quantities of 
fat are ingested, it is essential that carbohydrate should be included in the 
diet; otherwise the fat is not properly metabolised and acetone bodies are 
excreted, 


MAMMARY SECRETION IN RATS 79 


It is also probable that the large amount of protein deemed necessary has 
been somewhat overestimated, and that less amounts would give equally 
good results provided that. the diet was satisfactory in other respects. 

' That excess of protein in the mother’s diet has detrimental effects on the 
young is evident from the experiments of other workers, but such results 
have been explained differently. 

The experiments of McCollum and Simmonds [1918] are on almost identical 
lines with those described in this paper. These observers, however, were 
working with diets which were inadequate for growth of the young, and the 
main object of their research appears to be to show how far “the mother can 
serve as a protective agent in producing milk suitable for the nutrition of 
the young,” when she herself is getting a deficient diet. A great difference 
between the work of McCollum and Simmonds and that of the research now 
being described is the duration of the experiments. From the former’s curves 
it appears that the mother was left with the young for 40 days or more, 
which seems rather long, considering that a normal baby rat is perfectly capable 
of living alone from the 21st day. In one or two of their experiments the 
symptoms of the litter closely resemble those described in this series of experi- 
ments. In McCollum’s experiment 738 the mother rat was fed on rolled oats 
alone, and certain symptoms were noted in the young. “On about the 20th 
day a few of the young would throw themselves about the cage, and scream 
during this performance. Several of the young went into coma and died.” 

In some experiments where the mother appeared to have a high tolerance 
for protein, the babies behaved in the way McCollum describes about the 
20th-24th days; they had shown no symptoms at the earlier stage. Again 
in experiment 948 he says that “the young all died at a very early date. The — 
intestines were filled with gas. We are unable to account for the high mortality 
of these young.” The diet used in this case was rolled oats 82-0 and caseinogen 
18-0, 7.e. @ greater proportion of protein than when the rats had rolled oats 
only. The young died at intervals from the 3rd to the 23rd day. Seven 
experiments were made. 

The condition of the young certainly suggests a similar state to that found 
in this work [1921, 2], but only a few of the typical symptoms are recorded 
by McCollum and Simmonds. In their paper they state that the babies were 
weighed “frequently” and the curves are plotted at intervals of eight days. 
Unless daily weighings were made, and the babies carefully examined, it is 
highly probable that the spasms would be overlooked, since this stage is fre- 
quently of short duration, sometimes lasting only a few hours. The screaming 
fits exhibited by the older babies are much more obvious, the noise being 
distinctly heard in the next room. 

It is possible that these results are due to a deficiency of some dietary 
factor, but this deficiency is only evident when the diet is too rich in protein 
and therefore it may be that they are directly due to the excess of protein. 
_A similar diet, but with less protein, would not be deficient. 


80 G. A. HARTWELL 


GENERAL CONDITION OF THE ANIMALS. 


Only normal healthy animals were used in these experiments. If any rat 
became ill the results were discarded. There was no trouble with scab, 
pneumonia, or epidemic diseases. McCollum [1920] says that the “ palatability 
of the ration is the most important factor in animal nutrition,” and it is a 
well-known fact that many animals will starve rather than eat food which 
is distasteful to them. The rats used in this work readily ate any of the 
mixtures offered to them, and in some cases consumed enormous quantities. 
For instance, many rats were taking as much as 40:0g. bread (+ added 
protein and other constituents) on the ninth day of lactation. The initial 
ration of 15-0 g. bread etc. was usually adequate for the first two days, after 
which the animals’ appetites increased rapidly. The cages were examined for 
any hidden food, the “food-hoarding” instinct being especially marked in rats. 

The males and females were kept together until a few days before the 
birth of the litters, when the mothers were removed and placed in separate 
cages. After lactation the females were kept from the males for about two 
weeks to avoid over-strain of the maternal organism, and to give the mother 
a chance of recovery, should the special diet have upset her in any way. This 
method of procedure was adopted merely as a safeguard, because even when 
the litters died, the mothers appeared fit and well. Except during the actual 
period of lactation (and with one exception pointed out later) the males and 
females were fed on a diet of kitchen scraps, supplemented with small quan- 
tities of whole milk and white bread. This mixture proved a very good one 
and, as is shown later, the growth curves of the young animals fed on such a 
diet, are better than those given by Donaldson [1906] as normal for the 
white rat. 

Some white rats, and some white and black were used. It is possible that 
the latter should be somewhat heavier, since Donaldson states that black 
rats are heavier than white ones, yet there appeared practically no difference 
in size and weight. 


METHODS. 


The technique employed was the same as that previously described 
{ Hartwell, 1921, 1]. 

The litters, as before, were reduced to six. In a few cases less than six 
were born, and this is noted in the text and on the curves. The mothers and 
babies were weighed daily and the weights recorded graphically. In the 
“foster” experiments to be described later, the number in the litter varies, 
but is always stated. 

The foodstuffs used. (a) Pure caseinogen was not obtainable when these 
experiments were in progress, and a commercial form, “food casein,”! was 
therefore used. Experiments showed that the results obtained with “food 
casein’’ were exactly similar to those previously found with pure caseinogen. 


' I am indebted to Casein, Ltd., for supplying me with the composition of “food casein.” 


MAMMARY SECRETION IN RATS 81 


The “food casein” was 83 % pure, and the amounts given were calculated 
in terms of pure protein. 

(b) Pure egg albumin (as supplied by Baird and Tatlock) was used. It 
was very finely ground, in order to obtain an even mixture of the foodstuffs fed. 

(c) A commercial preparation of edestin was used. It was tested at the 
beginning of the experiments and found to contain only slight amounts of 
impurities. A new sample was tested lately and found to contain cane sugar. 
Every sample was not analysed, and as that obtained lately was less pure, it 
is possible that the results obtained with 1 g. and 2 g. edestin are due to the 
fact that smaller amounts of protein were actually fed. 

To find what amount of protein is really excess, it would be becal doko to 
work with absolutely pure foodstufis. 

(d) The lactose was pure, as supplied by Baird and Tatlock. 

(e) The salt mixture was that recommended by Mottram, and is a modifi- 
cation of the original one used by Hopkins. 


Sodium chloride NaCl . . 46°25. 
Magnesium sulphate feehind: ) Mgs0, a8 71-20 
Acid sod. phosphate NaH,PO,, H,0O .. ee 92-68 
Di. pot. phosphate K,HPO, ... 2 ... 254-60 
Calcium tetrahydrogen phos. Call (PO, )xH,0 144-20 
Calcium lactate Ca (C;H;0,).5H,O . ... 347-00 
Ferric citrate Fe (C,H,O,) 14. H,O ... ‘s 31-52 
Sodium fluoride NaF ... es os “D5 
Manganese sulphate MnSO, 5H, | eee és 2-00 
Potassium iodide KI... Pt oi me 10-00 


Total... 1000-00 


Water, ad lib, was always given unless the rats were having large quan- 
tities of milk. 


1. THE QUALITY AND QUANTITY OF DIETARY PROTEIN. 


In some recent experiments [Hartwell, 1921, 2] the initial diet consisted 
of 15-0 g. bread to 5-0 g. protein, both being increased proportionately. Only 
a very few babies survived and those which lived were not normal, therefore 
in these experiments less protein was given to the mother, in order to see if 
the litter would then survive. Small quantities of commercial yeast extract, 
marmite (about 0-2 g. per day) were added to the diet to make it more 
palatable. It has been found that small amounts make no difference to the 
final results and the animals ate such a mixture more readily. 

Three experiments were made with each protein.’ This involved 9 mothers 
and 54 young. The diet was not physiologically complete except in two of 
the experiments, when 0-5 g. butter, 0-7 cc. lemon juice and 0-7 g. salt were 
added to the mixture; this made no difference to the final results. 


Bioch. xv1 6 


82 G. A. HARTWELL 


(i) Proportion of 15 g. bread to 1 g. protein. 
(a) Caseinogen. Fig. 1 a. 
Exps. 160 and 173. 15-0 g. bread, 1-0 g. caseinogen, marmite. 
Exp. 175. 15-0 g. bread, 1-0 g. caseinogen, marmite, butter, lemon juice and salt mixture. 
_ Litter 160 were perfectly normal in every respect. One was weakly and 
died three days after weaning, but no typical symptoms (previously described 
[1921, 2]) were exhibited. 


160[- .173 
150 * oy 9160 
rs SiG 
140 A: 
oy * 
130 of 
Ps 
120+ ah 
Fs 
110} pe 
o * 
100 
un Case & 
8 oo 
as Fg a8 
80-- af we * Tah tee ao x 
70h at &, 
60+ aX 
KY ©) 
te an % 
50 XF Y oitt ‘ 
40 A Wis Weights of Litters 
ldead x79 ee be 
' 0 w 
ae \ 
Ba, Ae yi 
20 
10+ a 
D 
Pm Dew) ackad We eed ees Ves oes We ech eb aes Mie Gar cates Sea 
S 3 4 +8 6 7 8 6 10 11 19 18 14 15 16 17 10 18 20 21 


Fig. 1 a. Caseinogen. 

Litter 173 exhibited slight symptoms. However they recovered rapidly, 
the mother was removed on the 21st day, when all the babies appeared 
normal. They all survived weaning. On the 21st day, the average weight 
of babies 160 was 30-6 g. and that of litter 173 was 26-0 g. From the normal 
curve according to Donaldson [1906] the weight at 21 days is about 21-0 g. 
Hewer [1914] states that she used for subsequent feeding experiments any 
young rats which weighed at least 23-0 g. at weaning, so it would appear that 
both litters described above were well above average weight, in spite of the 
fact that slight symptoms developed in 173. It is possible that “breed’’ and 
number in the litter are factors influencing the actual weight of the babies. 
This is discussed later in the paper. 

Litter 175 showed bad symptoms and were all dead by the 19th day. 

The mothers all remained well. They were not full grown at the beginning 
of the experiment and put on a considerable amount of weight. 

(b) Edeastin. 


Exps. 141 and 195, 15-0 g. bread, 5-0 g. edestin, marmite. 
Exp. 163, 16-0. bread, 5-0 g. edestin, marmite, butter, lemon juice and salt mixture. 


MAMMARY SECRETION IN RATS 83 


All litters were normal in every respect, none of the typical symptoms 
were noticed in any one baby, not even excitability. The rate of growth in 
the middle part of lactation was extraordinarily good and apparently maximal 
(i.e. as good as when the mother is fed on bread and milk, a diet previously 
shown to give the best results). 


240;- 


230h- ak Ns 


Grams 
o 
° 
t 
* 


a~_ 
oi Nas 
. o— e 


170f- OYA cae, that Nox0 


pia 


160}- : Me 5206 


130L Weights of Mothers ee 4 REE foes 3 207 


4190 
160 a 


x 
150}- 3 A) 4 0206 


Grams 


Weights of Litters 


yale tub SSF 23 ae a a a Dd ae ee ce 


| 
oe. se ene ee se We te i) te 616. 17) (18 18 20° 21 
Fig. 16. Egg albumin. 


The mothers, however, showed a slight loss of weight during the first two 
weeks and this loss became very marked during the last week of lactation. 


6—2 


84 G. A. HARTWELL 


This seems rather extraordinary because during the last part the young were 
able to eat for themselves, and were not entirely dependent upon the mother. 


(c) Egg albumin. Fig. 1 b. 
Exps. 190, 206, 207. 15-0 g. bread, 1 g. egg albumin, marmite, 


In these experiments no litter was quite normal, but the mothers were 
removed on the 21st day and all the babies were successfully weaned. 

Litter 190 showed only slight spasms, and recovered rapidly. Two days 
after the mother was removed they were normal, but small and somewhat 
thin. 

Litter 206 were very similar to 190, but their condition was a little worse 
and they were not normal till a week after weaning. It will be seen from 
Fig. 1 6 that this litter lost weight for three days before they began to eat 
for themselves. 

Litter 207 showed worse symptoms than 190 and 206 but eventually all 
the babies recovered. 

At first the mothers maintained their weights or showed a slight loss, but 
from the 11th or 12th day they lost weight considerably. In ten days two of 
them lost approximately one-quarter of their body-weight and the third lost 
about one-fifth body-weight. The significance of this is discussed later (p. 86). 
Thus when 1-0 g. protein to 15:0 g. bread was fed to the mothers, slight 
symptoms developed in the litters, except with edestin. This is probably 
due to the fact that commercial edestin was used and, therefore, the mothers 
were receiving less amounts than with the other proteins. 


(ii) Proportion of 15-0 g. bread to 2-0 g. protein. 
(a) Caseinogen. Fig. 2. 
Exps. 161 and 198. 15-0 g. bread, 2-0 g. caseinogen, marmite. 
Exp. 176. 15-0 g. bread, 2-0 g. caseinogen, marmite, butter, lemon juice and salt mixture. 

Litter 198 were normal in all respects except that they were somewhat 
excitable and, as will be seen from the curves, gained only small amounts the 
two days before they were able to eat for themselves. 

Litter 161 showed the typical “excess protein” symptoms, but two lived 
till the 21st day and survived weaning, although they were not normal till 
two weeks after. 

Litter 176 were in a worse state than 161 and were all dead by the 18th day. 

The mothers 161 and 176 put on weight, whilst 198 lost weight. In this 
connection it is interesting to note that 198 litter were normal, but 161 and 
176 suffered badly. This is in agreement with a suggestion previously made 
that there is some correlation between the weights of the mother and babies. 

(b) Edeatin. 

Exps. 143 and 215. 15-0 g. bread, 2-0 g. edestin, marmite, 

Exp. 197. 150g. bread, 2-0 g. edestin, marmite, butter, lemon juice, salt mixture. 

Litter 215 were normal except that they developed slight “toe walking” 
towards the end of lactation. They recovered rapidly and were successfully 
weaned, ‘T'wo of the babies died the 2nd day and, therefore, the mother had 


MAMMARY SECRETION IN RATS 85 


only four to suckle. Since the average weight of these (see table on p. 87) 
is so much better than that of the two other litters, it is possible that it can 
be accounted for in this way. The fact that this rat may have had a greater 
tolerance for protein is also a factor to be considered. 

Litter 197 exhibited all the usual symptoms to some extent, but were 
never very bad. They soon recovered and were successfully weaned. 

Litter 143 developed very bad spasms accompanied by the usual symptoms. 
One died the 19th day and the others were all dead two days after the mother 
was removed. 


4198 
a 
eo 
Ff 
J 
oF 
TZ ® 
ae. 
Fan 
=" 
a” a—al6l 
2 survived 
Weights of Litters : 
a eet | si REN Bae bas a: Ae Yael | 


i | n 
7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 
Fig. 2. Caseinogen. 


The mothers’ curves are somewhat irregular, but on the whole show a 
maintenance of weight. 


(c) Egg albumin. 
Exps. 189, 208, 209. 15-0 g. bread, 2-0 g. egg albumin, marmite. 


All the litters survived and none showed any spasms. 

On the other hand none of the babies was quite normal. Litter 189 did 
best, but were somewhat excitable. Litter 208 were very excitable, and showed 
“toe walking.” 

Litter 209 had similar symptoms to a greater degree and were cold and 
very weak. 

_ The mothers appeared well, but, as in the case of those which received 
1-0 g. egg albumin, they lost weight rapidly. 


S 


86 3 G. A. HARTWELL 


Discussion oF RESULTS. 


The average weights of the young rats at 21 days are given below: 


Caseinogen Edestin Egg albumin 

1-0 g. protein: 

30-6 31-16 27:33 

26-0 31-4 24-66 

All died 31-16 18-16 
2-0 g. protein: 

30-33 33-0 (after 2nd day only four) 29-16 

22-0 (only two survived) 24-33 23-5 

All died 20-40 19-33 


(i) Irregular growth of the young. 


It has been previously pointed out [Hartwell, 1921, 1] that on any given 
diet for the mothers, similar growth curves could be obtained for the litters, 
and that such curves could be regarded as characteristic for that specific diet. 
This statement now needs some modification. 

The experiments described in this paper show that on any given diet the 
growth curves of the sucklings are by no means identical. It seems probable 
that when sufficient protein is given to produce bad symptoms in the young, 
then the above statement no longer holds good. This is well illustrated by 
the figures above. The average weights of a 21 day rat when the mother 
received 1-0 g. edestin are practically constant. These litters were normal in 
every way, so it appears that this amount of protein was not excessive. It 
is possible that less than 1 g. protein was given in these experiments, since 
recent samples of the commercial protein have been shown to contain cane 
sugar as impurity. 

In all the other experiments, varying degrees of abnormality were pro- 
duced and the weights are by no means constant. 

This point is again illustrated by experiments to be described later on in 
this paper. : 

Thus with sufficient protein in the mother’s diet to produce bad symptoms 
in the young, the individual metabolism of the mother rat seems to play an 
important part. 

The mother’s initial weight is not a factor to be considered, since on any 
given diet the babies of a heavy rat may die while those of a lighter rat may 
survive, or vice versa. 


(ii) Correlation between weights of mothers and offspring. 


It was previously suggested that there was some correlation between the 
weight of the mother and condition of the offspring, and that if the mother 
lost weight she might be able to spare the babies some of the symptoms. 
This suggestion is confirmed by these experiments. 

In Fig. 2 litter 198 have a fair growth curve, and showed no symptoms. 
The mother lost weight. Litters 176 and 161, on the other hand, exhibited 


MAMMARY SECRETION IN RATS 87 


typical symptoms and only two of the 12 survived. These mothers put on 
weight. 

In both sets of experiments (7.e. 1-0 g. and 2-0 g. protein) it is very obvious 
that the mothers fed on egg albumin lost weight considerably. 

Since albumin is a less good protein than either edestin or caseinogen, it 
may be that the mother is supplying the necessary amino-acid constituents 
from her own tissues. 

Even 5-0 g. egg albumin to 15-0 g. bread results in some loss in weight 
of the mother, in a deficient diet [Hartwell, 1921, 2], but practically no loss 
is found when that proportion is given in a physiologically complete diet 
(experiments described later). 

From the following figures it can be seen that, in some cases, the mother 
rat lost as much as one-third of her body weight; moreover this loss was 
very rapid and occupied only ten days. 


= at beginning Weight at Loss 
Exps. of lactation. g. end. g. g. 
1-0 g. egg albumin: 
190 221 164 57 
206 228 159 69 
207 168 142 26 
2-0 g. egg albumin: 
189 247 165 82 
208 214 174 40 
209 149 128 21 


Rat 207 lost less than 190 and 206 and her litter were in a worse condition 
and gained less weight than those of 206 and 190. 
Again, in the other series, rat 209 lost less and her litter were the worst. 


(iii) The amount of protein constituting excess. 


It is not possible to give a definite statement as to the proportion of 
protein which should be fed to a lactating animal, since with different proteins, 
a different amount will constitute excess. For instance, good growth curves 
were obtained when the mother was fed with 15-0 g. bread and 1-0 g. edestin, 
but in the three litters whose mothers had 1 g. egg albumin, there were no 
normal babies; with caseinogen, one litter was normal, one had slight symptoms 
and the babies of the third litter all developed bad symptoms and died. As 
previously explained (p. 81) there is a possibility of impurity in the edestin 
and hence the animals might be getting less than 1-0 g. of protein, which 
might account for the slightly better results obtained with the edestin. 

It is probable that the amount of protein constituting excess will vary 
with the presence or absence of other dietary constituents, and that such 
constituents must be considered quantitatively in relation to the protein. This 
point will be dealt with later in the paper. With caseinogen and edestin, a 
diet which is complete physiologically from a qualitative point of view gives 
just as bad results as one not physiologically complete. For example rat 176 
received butter, lemon juice and salt mixture in addition to bread, caseinogen 
and marmite, thus completing the diet qualitatively, yet her litter were worse 


88 G. A. HARTWELL 


than those of 198 and 161, whose mother had only bread, caseinogen and 
marmite. 


(iv) The quality of the protein. 


Certain typical symptoms, already described, were characteristic of the 
litters, irrespective of the protein fed to the mothers. With egg albumin, 
however, some extra symptoms were noticed. The weakness appeared to 
come on at an earlier stage, the babies were thinner and were very cold to 
the touch. This coldness was observed at an early age and was frequently 
the first indication that something was wrong. It was evident even when the 
symptoms were slight and the babies recovered. It is possible that the thin- 
ness of the babies is due to less good milk being produced by the mother. It 
is suggested that egg albumin is deficient or poor in certain amino-acids which 
are essential for the production of milk of adequate nutritive value. To prove 
such a point it would be necessary to feed with pure amino-acids. 


2. THE RELATION OF PROTEIN TO OTHER DIETARY CONSTITUENTS. 


(i) The effect of large quantities of protein in a 
physiologically complete diet. 

Hartwell [1921, 2] showed that an initial diet of 15-0 g. bread and 5-0 g. 
protein fed to nursing rats was detrimental to the young. It was, however, 
pointed out that such a diet was deficient physiologically, yet it was unlikely 
the bad symptoms were due to such deficiencies, since the rat can bring up 
a healthy litter when fed on white bread alone [Hartwell, 1921, 1], a diet also 
lacking in fat, vitamins and salts. 

In these experiments the proportion of protein to bread was kept the same 
as that used in the previous investigation (7.e. 15-0 g. bread to 5-0 g. protein), 
but 0-5 g. butter (fat and vitamin A), 0-2 g. marmite (vitamin B), 0-7 cc. 
lemon juice (vitamin C) and 0-7 g. salt mixture (composition given on P. 81) 
were added to complete the diet. 

Three proteins were tried, caseinogen, edestin, and egg albumin and in 
each case two series of experiments were done. 

1. The diet started as soon as possible after the birth of the litter; never 
more than 24 hours. 

2. Diet started before the birth of the litter, and at different stages of 
pregnancy. 

The bread and protein were increased proportionately as required, but the 
other constituents remained constant. 

(a) Caseinogen. Fig. 3. 
Exps. 113, 194, 171. Diet started at birth of litter, 
Exp. 126, Diet started 10 days before birth of litter. 
pe te “i 13 < +s 


os aoe ” 16 ” ” 
op 206, ” 27 ” ” 


MAMMARY SECRETION IN RATS 89 


The gestation period in the rat is given by some observers as 21 and by 
others as 23 days, so in the second series of experiments the rats were getting 
the excess protein diet for at least half the gestation period. Rat 166 received 
this diet during the whole of gestation. In this animal, either the gestation 
period was longer than normal, or else fertilisation was delayed, because she 
was separated from the males when the special feeding was started and the 
litter were not born for 27 days. | 


Se x 113 
200-7 
abet 
— * 
oa 
x 
180}- if 
x 
170+ Fs 
x 
160}— eats 
150 * (J 
x” 
140+. if Started at (X—~%* 
eee birth of litter pstagtsn 

130}-- 4 s—eu 

x Started before |a—a 
120. iar birth of litter | -—- 


90 oe fe 
ff 
“ 5 a 
80/- * Aiea Us 
fi 3 
70} 
Vee g~ ee" 
a * , ae a 
To, ge 4— 4-4 oe 
a “ Pw Pot nas e 
30 mee ned \ 
Path 
Pa A— Omer of 3 : ‘ 
20 ap F8 Li" Weights of Litters 
10 
eee ge at a gs Pg os 
5 RS aes Bae ha Oe 10s 1h Ne 18 te ID 16 17 18 18 20 at 


Fig. 3. Caseinogen. 
A. Feeding started after the birth of the litter. 

Litter 113 became excitable on the 9th day and from that time onwards 
showed signs of being hungry. 

No bad symptoms became evident until the 19th day when four of the 
litter had very bad fits, screamed, rushed round their cage, bit each other 
and hung on to the cage with firmly fixed jaws. These fits lasted from one to 
five minutes. Afterwards the babies were very exhausted and would sprawl 
on the floor of the cage. At this stage the mother frequently removed them 
to the nest and covered them up. The mother was taken away on the 21st 
day and all the litter survived, in spite of the fact that bad screaming fits 
were again noticed on the 22nd day. 


90 G. A. HARTWELL 


Litters 194 and 171 gave the typical “excess protein” curves and all died. 
The condition of the alimentary canals of the babies was exactly the 
same as previously described [Hartwell, 1921, 2]. 


B. Feeding started before the birth of the litter. 


The actual duration of the feeding before birth seems to be an unimportant 
factor, since all four curves are similar and the babies were all dead either 
the 14th, 15th, or 16th day, irrespective of when the feeding was started. 
In some experiments to be described later (p. 92) the mothers were fed on 
the caseinogen diet from the time they themselves were weaned and had, 
therefore, been eating this mixture for 70 days before the birth of the litter. 


(6) Edestin. 
Exps. 177, 112. Feeding started at birth of litter. 
Exp. 124. Feeding started 2 days before birth of litter. 
gn. 2 Ge a 11 ay “s 
rapa: t 35 21 i * 


A. Feeding started at birth of litter. 


Litter 177 showed quite good growth up to the 11th day, after which they 
only gained small amounts. They exhibited all the typical symptoms, though 
not so markedly as usual. They ate for themselves on the 16th day and were 
all successfully weaned. They were not normal for about ten days. 

Litter 112 grew much more rapidly than did 177, but they developed bad 
symptoms earlier and were much worse in every way. Only one survived 
and it was weakly for a month after being taken from its mother. 


B. Feeding started after the birth of the litters. 


In these experiments the time of feeding before birth bears a definite and 
inverse relation to the time the babies lived, but this does not apply to the 
litters whose mothers were fed with caseinogen and egg albumin. 

There is no reason to describe the experiments separately, or in detail, 
as the results are similar to those already described for caseinogen. 


(c) Egg albumin. 
Exps. 111, 193, 221. Feeding started at birth of litter. 
Exp. 125. Feeding started 4 days before birth of litter. 


” 129. ” 11 99 ” 
” 145. ” 9 ” ” 
” 130. ” 15 ” 9” 


A. Feeding started at birth. 


Litter 111 had the best growth curve in this series, but bad spasms were 
seen on the 13th day and after this the usual bad symptoms were marked. 

Only one of the litter survived and it was very weak and backward in 
every way. It did not eat for itself till the 21st day and therefore the mother 
was not removed until the 24th day. After this the young rat was weakly: 
for some time, but survived. 


MAMMARY SECRETION IN RATS 91 


Litter 193 were all dead the 17th day, having suffered from all typical 
symptoms, though not to a marked degree. It has been noticed that if 
extreme weakness comes on at an early stage, the acute spasms are frequently 
absent. 

Litter 221 suffered badly. They showed spasms and all typical symptoms. 
They were all dead the 16th day. 


B. Feeding started before birth of litter. 


These results again agree with those obtained with edestin and caseinogen. 

With all three proteins the mothers were very fit and in most cases put 
on weight. 

From these experiments one would conclude that feeding the mother rat 
with protein in the proportion 15-0 g. bread to 5-0 g. protein, even in a physio- 
logically complete diet is harmful to the litter. If the feeding is started before 
the birth of the litter, then no babies survive; also it appears to make very 
little difference if the feeding is started a few days before the birth or main- 
tained during the greater part of gestation. 

Failure in rearing the babies is of much more frequent occurrence when 
the mother is given an excess of protein in the diet (a¢ birth of litter) than 
when she is fed. on other mixtures. It is very easy to see if the young rats 
have been fed, and with an excess protein diet it was not infrequently found 
that the whole litter died in the first two or three days, and had hardly had 
any food. . 

The symptoms develop at approximately the same time and this seems 
to suggest a correlation with some stage of development of the young. This 
possibility is supported by the fact that in a foster litter the symptoms also 
develop about the same time (experiment to be described later). 

If the feeding is started after the birth of the litter, some babies may 
survive, but they are never normal; the majority die. 

The results with egg albumin are less good than those with caseinogen 
and edestin, but this as already suggested is probably due to the different 
composition of the proteins. Other differences will be commented on later 
(p. 98). 

The fact that some litters do better than others (e.g. litter 113 show a 
better growth curve than any other on the same diet) is possibly due to the 
greater tolerance of certain rats for high proportions of protein. In this con- 
nection it is also probable that the loss of weight of the mother is a factor to 
be considered; rat 113 lost weight, while other rats on that diet gained and 
their litters died. 

Thus the bad effects on the litters are obtained even when the diet is 
physiologically complete. It should be pointed out that the amounts of butter 
(fat and vitamin A), marmite (vitamin B), lemon juice (vitamin C) and salt 
mixture (inorganic constituents of diet) used were adequate for a normal diet 


92 G. A. HARTWELL 


(see experiment described below, when weanlings were fed on a similar 
mixture). 

The mothers remained well during the experiments and when the diet 
was started before the birth of the litter, the period of feeding lasted as long 
as five weeks. 


the” 
Ni 
\ 
tig 
or 
Peg 


4 days 
YS I Calls RE PS OY eS aes a EE RY PP ee Tes pa ee De TE DEA 


Fig. 4. 1, 4, 6 caseinogen; 2, 3 control. 


Growth of young rats on a diet similar to that fed to the mother. Fig. 4. 


One criticism which might be offered is that although the mother rats 
appeared well, they might not have remained so had the feeding been con- 
tinued longer. To test the adequacy of this diet, young rats were given a 
similar mixture. Two normal healthy litters were chosen and fed on bread 
and milk for two days after weaning. Then half of each litter was put on 
this diet, the protein used was caseinogen, and the other halves of the litters 
were used as controls, being fed on kitchen scraps supplemented with small 
quantities of bread and milk. 

The young rats were weighed separately every day at first and later every 
four days. The increase in weight of the males is shown in Fig. 4. 

The caseinogen-fed males did better than the controls, but the females 
grew equally on both diets. 

The experiment lasted 90 days, which should prove a suitable length of 
time for testing the diet. 

The protein-fed animals grew rapidly and were very fit, their coats were 
thick and silky and decidedly better than that of the average animal. The 
females became pregnant and the litters were born at about 13 weeks, which 


MAMMARY SECRETION IN RATS 3 93 


is recognised as the normal time for the appearance of the first litter. In the 
controls the first litter appeared about two weeks later. The three caseinogen- 
fed rats produced healthy litters, which compared very favourably both in 
size and weight with the average first litter. 

The litters were not reared. Their growth curves were typically those 
described previously when the mothers received excess protein in the diet. 

The weights of the animals at 70 and 90 days are compared with those of 
Donaldson [1906] and Osborne and Mendel [1911]. 

The average weights given by Osborne and Mendel are slightly less than 
those of Donaldson. The authors suggest that this is due to differences of 
“breed.” 

The rats used in these experiments were black and white and as before 
suggested they may be of a bigger breed. In other experiments both white, 
and black and white animals were used, and practically no differences in 
weight were noticed. The only difference was that the white rats do not make 
such good mothers and, therefore, for the later experiments, they were rarely 


used. 


Weight ofratat70days Weight at 


Average weight 90 days 
—. Average weight 
Male Female Males 
g- g- g: 
Donaldson 10 100 150 
Osborne and Mendel 120 100 135 
198 114 231 
Caseinogen-fed (3);201 (3), 143 (3) 202 Females not given 
l216 | 165 258 because they were 
Controls. Fed on kitchen scraps and ) 9 (178 103 2) {204 pregnant 
small amount bread and milk - J (2) | 182 (3) 158 ( | 223 
162 


This experiment proves the diet to be adequate in all respects except during 
lactation. It is extraordinary that a diet, so obviously good for a growing 
animal, suitable for fertilisation and production of offspring, should cause 
such disastrous effects on the suckling babes. 


Experiments with foster-mothers. 


It was suggested on p. 91 that the bad symptoms in the babies were in 
some way connected with development, since the symptoms usually came on 
at about the same time. It does not seem at all feasible to suppose that the 
milk contained toxic substances at a definite time, and then lost these pro- 
perties again. In support of this view it was found that if a rat could be 
induced to suckle more babies after her own had all died and the excess 
protein feeding continued, then the new litter gained for a time though not 
to any great extent, but finally died. The growth curve for the young is very 
similar to that obtained with the rat’s own litter, except that the increase of 
weight of the babies is less than with the original litter. (This fact applies 
also to a good diet, see experiment 107.) 

Several “foster” experiments were made. Rat 152 was fed on bread 


94 G. A. HARTWELL 


edestin, butter, marmite, lemon juice and salt mixture (the same diet as 
previously used) while nursing her own babies (experiment 112), and the diet 
continued with the foster litter, five in number. The foster babies gained 
weight until the 9th day, then ceased to gain, finally lost weight and died. 
No spasms were noticed, but weakness was a prominent feature. It was 
frequently found that slower-growing babies developed weakness earlier and 
showed only slight spasms, or none at all; however, this was not always the 
case. The mother, 152, lost weight considerably, but in experiment 112 she 
had gained a good deal. At the beginning of experiment 112 she weighed 212 g. 
and at the end of experiment 152 she was 205 g. In this instance loss of weight 
of the mother did not spare the litter, but the mother’s loss is hardly com- 
parable to that of other experiments, because in this case the rat had just 
put on what she now lost. . 

Rat 107 brought up foster babies when she was having a diet of bread 
and milk. Previously she was fed on a diet of egg albumin and bread, and all 
her own litter had died, with typical symptoms. These two experiments are 
representative of others, and it is inferred that after her own babies have died 
as a result of too much protein in her diet, the rat is capable of bringing up 
another litter. If a good diet is given in the second experiment, the babies 
are normal but small. They do not develop so rapidly as a rat’s own babies, 
and are not able to eat for themselves until about the 21st day. If the excess 
protein diet is continued, the babies gain small amounts for a time, but 
eventually die, having shown the usual symptoms, except that they do not 
suffer from spasms. 

In the experiments in which the mothers received excess protein, the foster 
babies did not develop symptoms until a few days later than the original 
babies. It has been pointed out that on a good diet the foster babies do not 
eat till later. This is in agreement with the suggestion that the development 
of the young is a factor to be considered. 

Quite a number of rats were unable to bring up a foster litter after their 
own had died as a result of too much protein in the diet. The rats took to the 
babies and mothered them, but were apparently unable to feed them. The 
babies lived from two to four days, during which time they lost weight. They 
did not appear to have been fed, although the rat took care of them in the 
nest. These babies were rarely eaten by the foster mothers. 

It was previously suggested [ Hartwell, 1921, 1] that excess protein caused 
an alteration in composition of the milk, and finally a cessation of the flow. 
From these experiments it would appear that in some rats the secretion could 
be re-started, but not in others. 

It is possible in some cases to alter the diet, either by adding milk, omitting 
the excess protein, or entirely changing the food to bread and milk, and thus 
to obtain recovery of the babies. 


MAMMARY SECRETION IN RATS 95 


Experiments on changing litters. 


That the milk of the rat on excess protein contains some toxic substance 
is further demonstrated by changing the litters. 

Rat 136 received bread and milk. Rat 135 had a diet with excess protein. 
The litters were born on the same day. As soon as the babies 135 developed 
spasms (the 12th day) the litters were changed. Litter 135 recovered rapidly 
when fed by rat 136, which was haying bread and milk, and were successfully 
weaned at the normal time. Two of litter 136 developed screaming fits and 
bad spasms the 20th day, but recovered. In another exchange the normal 
babies given to a mother, whose own litter were showing bad symptoms, lost 
weight for three days, after which they ate for themselves. In this experiment 
no typical symptoms were noticed. 

Thus it is evident that with large amounts of protein in the diet, which 
is complete physiologically, the milk is affected, and the babies show abnormal 
and typical symptoms. It is suggested that 


(i) The quality of the milk is altered, in that it possesses some toxic pro- 
perties responsible for the spasms. This may be due to excess of amino-acids, 
or possibly to some deficiency in the milk. 


(ii) The quantity of the milk is inadequate, since the babies lose weight. , 


Since large quantities of protein are usually recommended for a lactating 
animal, it is probable that definite and large amounts of one or more other 
constituents should be included in the diet. 

With a view to throwing some light on this, the following experiments 
were made. 


(ii) Effect of adding whole milk to a diet containing excess protein. 


The bread and protein were fed in the same proportion as in the last series 
of experiments. The milk was heated to 70° C. In these experiments the rats 
had no water to drink. (In some earlier experiments, when 100 cc. whole 
milk were given, it was noticed that the animals drank no water, and usually 
upset the containing vessel as soon as it was given to them.) 

The control experiment D was done at the same time of the year as the 
other experiments, though not actually the same year. So far the growth 
curves of the young when the mother is fed on bread and milk are practically 
constant, irrespective of the time of year, so long as the animals are kept in 
a well-heated room. Other observers find different growth curves according 
to the time of year. It is possible that this difference is minimised when the 
diet is exceptionally good, e.g. bread and whole milk. 

In these experiments no abnormal symptoms of any kind were shown, but 
in all cases the growth was depressed, as compared with the standard (bread 
and milk). 


96 G. A. HARTWELL 
This is well shown by the following figures: 


Average weight of young rat at twenty-one days. 


Exp. Diet Weight in g. 
D Bread and milk 39-2 
41 : 31:3 
108 a +  edestin 29-6 
a 
201 30-4 
202 fa * gelatin é 26-0 
203 : ' (29-0 
204 » 9 egg albumin 34-3 (litter of three only) 


In these experiments the rats were eating even larger quantities of protein, 
since milk contains from 2-5 %-4-0 % protein [Matthews, 1921]. The exact 
amount of milk required to prevent the babies from developing the typical 
symptoms has not been determined, but other experiments showed 50 ce. to 
be inadequate. In all probability the individuality of the rat would be a 
factor conditioning the amount required. Later in this paper (p. 101) it is 
demonstrated that the whey from 100 cc. milk is equally efficient in pre- 
venting bad results with caseinogen and edestin, and that the growth is equal 
to that obtained with bread and milk. 

Thus milk contains some dietary constituents which render it possible for 
the rat to ingest even larger quantities of protein and yet the litter is not 
affected by bad symptoms. The only detrimental effect is that the babies 
do not grow at a maximal rate. 

The next experiments were started with a view to finding the nature of 
such substance or substances. 

Two experiments were made with each protein and the average weight 
of the two litters does not correspond very closely. In spite of the fact that 
no symptoms developed in the babies, the mothers were ingesting very large 
amounts of protein and itis probable (as before suggested) that the individual 
metabolism of the mother is a partial explanation of these discrepancies. 


(iii) Effect of adding other constituents to a diet containing excess protein. 


In this series the proportion of protein to bread was kept constant as in 
the previous experiments, 7.e. 5 : 15. 


(a) Salts. Calcium given as lactate, 1-0 g. per diem (Fig. 5). 


Exps. 114, 172, 156. Bread, caseinogen, marmite, calcium lactate. 
Exp. 138. Bread, egg albumin, marmite, calcium lactate. 
» 140. Bread, edestin, marmite, calcium lactate. 


Litter 114 grew well the whole time; they were excitable and exhibited 
toe-walking, but otherwise no typical symptoms. It is probable that the 
mother had a greater tolerance for protein, as before suggested. Apart from 
this litter, the results were in agreement and the same for each protein tried, 


MAMMARY SECRETION IN RATS 97 


1.¢. caseinogen, edestin, and egg albumin. All the litters showed typical and 
very bad symptoms, but the babies lived longer than usual. 

Thus the addition of calcium lactate to a diet containing excess protein 
does not alleviate the symptoms but causes a slight prolongation of the life 
of the sucklings. 

(b) Salt mixture. Bread, caseinogen, marmite, salt mixture. No improve- 
ment in the litter was obtained when 1-0 g. salt mixture was added to the 
daily ration. The symptoms were bad, and the young were all dead on the 
15th day. 


210,- — 
200 , fs il4 
a 
190} . ov 
180}- 4 
170 oy 
ry 
160} Pd 
Pd 
150}-- 4 
140] ae 
} 4 
130}- A 
120-- oe 
s—., 
110 A ME Se 
Z A# 5, 
3 100}- WA ° . 
So “73 
Z a— a 
‘ s—2— 
a” —e-—* % 
ae ve 
=~ 
go" —A ° 
Z AG 
& 
B\ a> 
16a \Y 


am a=” aa 
Weights of Litters 


Ee Ee ae | vee 


Da 
ee ee ee ee 
9 ote. ee) I, wae ae ee ace. ee 2 ae BS ee 


Fig. 5. Calcium Lactate. 


(c) Milk ash’. The milk was evaporated to dryness, the residue scraped 
up and strongly heated in a crucible until a constant weight was obtained; 
0-7 g. were given per day (7.e. the amount in 100 cc. milk). 

These results were very similar to those obtained on adding calcium 
lactate to the excess protein diet. The symptoms were in no respect alleviated, 
but the life of the young was slightly prolonged. 

(d) Fat and vitamin A. No beneficial effects were obtained on adding 
3-5 g. butter per day to the excess protein diet. Edestin was used. 


1 T am indebted to Miss W. M. Clifford for the preparation of the milk ash. 
Bioch. Xv1 7 


98 G. A. HARTWELL 


(e) Lactose. 4-0 g. lactose were added to a diet containing excess edestin, 
but the spasms and bad effects were in no way improved. 

(f) Fat, vitamin A, and lactose. Casemogen was used and the diet was 
physiologically complete in all respects, fat and lactose being added in ap- 
proximately the same amounts as found in 100 cc. milk. 

The litter suffered very badly, two eventually survived, but they were 
weakly for some time and were not in a fit condition to leave the mother 
until the 26th day. 

Vitamin C. The diet consisted of bread and egg albumin in the usual 
proportions and 0-7 cc. lemon juice per day. Great difficulty was experienced 
in getting a rat to bring up a litter on this diet. In many experiments the 
babies died within the first three days. The mothers were well, the babies 
were average weight, and apparently all right at birth. This may be pure 
coincidence. No explanation is offered to account for this failure. 

Finally two experiments were successful. The litter all exhibited typical 
symptoms, and were dead by the 13th day. 


(iv) Effect of adding large quantities of yeast extract (commercial preparation, 
marmite) to a diet containing excess protein (Fig. 6). 

Two experiments were made with each protein (caseinogen, egg albumin 
and edestin). 

1. The diet consisted of 15-0 g. bread, 5-0 g. protein, and 3-0 g. marmite, 
all constituents increased proportionately as necessary. 

2. The bread, protein, and marmite were the same as above, but butter, 
lemon juice, and salt mixture were added to complete the diet. 

These curves are very similar to those obtained when 100 ce. whole milk 
were added to the excess protein diet. To bring out this point curve D (bread 
and milk standard) is included in the figure. With edestin and caseinogen a 
depression of the growth curve was obtained, very similar to that obtained 
when the mother had these proteins + 100 cc. whole milk. The babies were 
practically normal, but slight toe-walking and excitability were noticed; these 
features, however, were very quickly recovered from. With egg albumin, on 
the other hand, the results were not so good. No spasms were noticed, but 
the babies were far from normal. They were all right at first, but on the 
14th day they felt cold, were weak, inclined to roll over, thin and lethargic. 
They walked very badly and their movements were jerky and incoordinate. 
Two of the litter 180 died but the rest survived, as did all litter 182. The 
survivors were not normal for two weeks after weaning. 

The average weights of the young rats are given below. 

Average weight at 21 days. g. 


Exp. D Bread and milk (standard) 39-2 
es »  edestin, marmite 34-7 
» 165. RS ae Be butter, ete. 25:7 
jae by) 3 »  Caseinogen, marmite 34-0 
oo) nen a sb ea butter, ete. 30-7 
a... a » egg albumin, marmite 22-0 


» 182. “3 “a os butter, ete, 26:3 


MAMMARY SECRETION IN RATS 99 


- From these experiments it is clear that “something” in marmite is capable 
of obviating the bad results in the litters usually obtained with excess of 
caseinogen or edestin in the mother’s diet. Also the actual weights of the 
babies are better when the mother’s diet is lacking in fat, vitamin A, salt 
mixture and the antiscorbutic vitamin. This may be merely coincidence; a 
far greater number of experiments would be required before any conclusions 
could be drawn. It is interesting to note that similar results were obtained 


24or Poget, 
230} f 
220} 
210K aiid 
ae 
ey 2 Pa // 
190} 5 Pt ay eae Rs 
or - 
a Pea! 
ie a 
170} i re bs 
. a * 
160}— ff A Ae 182 
. a a 
£150/- if ef 2 veg 165 
g Leto a / a 
140} Vi pi ee 
130+ Ps a 1 A 
Mine. ve A 
o— 
PEGE 
yer a) Ome 
a7 a “e_ * 
ft ea 
re 
a \ oe 
Ce 
* 
Weights of Litters 
20/- 
10 an 
(TIES SSCS Sno SVE NES” Paley ize aber Mel cosy Cases a eines WR ce 
27s Soe 47; 8 o8 16, 1. 1¢ 18.14 16 16.17 18 18 90° 2 


Fig. 6. 


when less amounts of protein were given. Since the results are as good when 
the mother’s diet is deficient in fat, vitamins A and C and salt mixture, it 
appears that the “something” in marmite is the primary factor. It may be 
of interest that the marmite contained only traces of calcium salts. 

With egg albumin the improvement was not nearly so good. The litters 
certainly did better than when small amounts of marmite were fed. to the 


7—2 


100 G. A. HARTWELL 


mother, but they could not compare at all favourably with the caseinogen 
and edestin babies. These differences are discussed later (p. 102). 

The mothers maintained or increased their weights, except 180 (fed on 
egg albumin in a deficient diet) which lost appreciably. 


(v) The importance of protein. (Fig. 7). 


Exp. D. Control. Bread and milk (same as in Fig. 6). 
Exps. 118, 120. 15-0 g. bread, 5-0 g. edestin 
Exp. 167. 55 » 30 i 4 
i 168. - Co 8® ahaa os +whey from 100 cc. milk. 
Exps. 159, 162. ce Fa 0-0 protein 
Exp. L. Bread alone. [Taken from Hartwell, 1921 ,1]}. 
240 .D 
230}- J 
118 
220+ ps 
210+ be : 
: 120 
200}- “7 
a 
190+ | K-48 $5 
Pep : 
180}- ; ” Zw 
7 Fa 4168 
170 fed 
160 V 4 
a 
1s0l- 4 7 167 
sor V4 ‘ e 
130} / * 
Ronee 
120 ihe 
a —o162 
110} Ur rere ee yD epics 
a 
of Pe ae 
Op Se ae Paths 
90}-- rR ce o” ‘geen ys 
AL}. Leo ore am & 
rs |: 7 aon BAn 
ae 4 Z0%--*" a8 
70 _ . 4 hep ang a” 
pe Lae 
fe 4H a” 
WA a a a 
oh ra Fi ae A or 5 Weights of Litters 
40h A nich o% 
, So S ro 
42 QZ" 
0a 
20}-- 
10 
MG We ld PER: Waa WS er Wa RCE ve WOON Do GAPE Wr eSATA YO er tr: 
2 8 4 6 6 1 Oo SB. 10" Ts ARS ABYSS Gel ae, ne 2 
Fig. 7. 


Preparation of the whey. 100 cc. milk were warmed to 40° C., and a few 
drops of rennet added. ‘When the junket was firm, it was hung up in a muslin 
bag and the whey separated by dripping through. Each rat was given daily 
as much as could be obtained from 100 cc. milk. The protein and bread were 
increased proportionately as before. . 


MAMMARY SECRETION IN RATS 101 


It has already been mentioned that the whey was as effective as milk in 
preventing the bad symptoms in the litters. Also when the proportion of 
bread to protein was 15 : 5 the growth of the litter was practically equivalent 
to that obtained with a bread and milk diet, which up to the present time is 
regarded as maximal. It is extraordinary that such good growth should be 
obtained when the mother’s diet was so poor in fat and vitamin A!. 

In this connection it is interesting to note that all the mothers that were 
killed had large deposits of subcutaneous and peritoneal fat. It may be that 
with large reserves of body fat, the amount in the diet is immaterial from the 
point of view of milk production. 

In experiments 159 and 162 in which the mothers were fed on bread and 
whey, the litters did very little better than in experiment L when the mother 
received bread alone (taken from Hartwell [1921, 1]). Comparison of the 
protein rations in these experiments shows that the bread and whey contains 
only slightly more protein than bread alone. When more protein was added 
to the diet (experiments 167 and 168) a considerable improvement in growth 
of the young was obtained, though all the other constituents were the same 
as before. A further addition of protein (experiments 118 and 120) gave a still 
greater improvement in the growth of the litters. In this way it is seen that 
adding protein to the diet can raise the growth curve from almost the minimal 
to the maximal. Thus protein itself does play a very important part in the 
production of milk of good nutritive value, yet from other experiments de- 
scribed in this paper, it is obvious that the diet must be rich in other con- 
stituents as well or else harmful results occur in the litter. 

One of litter 118 was killed accidentally on the 7th day and the average 
weight of these babies is rather greater than that of 120, whose mother received 
a similar diet. It is quite possible that the actual number in the litter is a 
factor of importance. Experiments to investigate this point are now in 
progress. 


Discussion OF RESULTS. 


(i) Prevention of the bad symptoms in the young. 


It has been conclusively proved that large quantities of protein are by 
no means harmful to the rat itself, since the young and adult animals can 
live normally on a diet containing excess protein for considerable periods. 
The proportion of 15-0 g. bread to 5-0 g. protein proved an excellent diet for 
growing rats, but much less protein, even as little as 15-0 g. bread to 1-0 g. 
protein produced bad symptoms in the suckling litter. Therefore, it appears 
that the rat can metabolise the protein effectively from the point of view of 
its own growth, but is unable to produce normal milk, unless definite quan- 
tities of some other constituent (or constituents) are supplied. Our knowledge 


1 Since the caseinogen and edestim used were not absolutely pure, it is possible that traces of 
vitamin A may have been present. 


102 G. A. HARTWELL 


of the metabolism of individual amino-acids is very limited, but the experi- 
ments described in this paper show that there is some quantitative relation 
between the amino-acids and other dietary constituents. What these con- 
stituents are is a very difficult problem. The fact that whole milk can prevent 
the bad symptoms throws very little light on the subject. Owing to the 
complicated nature of milk, it merely proves that large quantities of protein 
in the mother’s diet produce no harmful effects in the litter, as long as 
“something” else is included. It is even possible that the necessary factor 
will vary with the nature of the protein. 

It has been pointed out that 3-0 g. marmite to 5-0 g. protein is adequate 
in preventing spasms, etc., in the case of caseinogen and edestin, but with 
egg albumin, although the symptoms are somewhat reduced, the litter still 
suffers badly. Also with egg albumin much better results were found when 
small quantities of butter, lemon and salt mixture were added to the diet, 
but with caseinogen and edestin, it was immaterial if the diet was complete 
or not as long as large quantities of marmite were fed. Since caseinogen and 
edestin are somewhat similar in constitution, and egg albumin is different 
there seems some evidence for suggesting that the factor will vary with the 
constitution of the protein. 

Some experiments with egg albumin now in progress show that the litter 
exhibit still greater improvements (but are not normal) when more butter is 
added to the ration, the diet being physiologically complete. 

It is quite obvious that some other constituents are vitally essential for 
the production of normal milk, if large quantities of protein are ingested. 

These are all present in milk, and in the case of caseinogen and edestin, 
present also in commercial yeast extract (“marmite’’). In the light of present- 
day investigations the most obvious explanation is that vitamin B is the 
factor, but this conclusion is by no means justified from the experiments 
described here. Again “marmite” alone was not wholly effective with egg 
albumin. It might be that insufficient was given, but this is hardly likely, 
because 3-0 g. to 5-0 g. protein is a large proportion. A more probable ex- 
planation is that some other constituent is also needed for the proper meta- 
bolism of this protein in a lactating animal. 

The fact that large quantities of milk and marmite were necessary suggests 
an actual chemical relation between the protein and some other factor, and 
not a catalytic or hormone action. It is possible that less marmite would be 
required if the right proportion of other constituents could be arrived at. 
Experiments have shown 1-0 g. marmite per day to be ineffective, yet the 
amount necessary to supply vitamin B to a rat is estimated as 0-2 g. per day. 

It is possible that the animal needs a greater amount of vitamin B for the 
effective metabolism of large amounts of protein; hence it is used by the 
mother and she has none to spare for her milk. On the other hand the mother 
is usually regarded as being a safeguard for the young and it is thought that 
she can sacrifice her own tissues for the benefit of her offspring. Also it has 


MAMMARY SECRETION IN RATS 103 


¢ 


been proved that in a non-lactating animal quite small amounts of vitamin B 
are adequate even when the intake of protein is high. 

The experiments of Karr suggest that vitamin B is an essential constituent 
of a dog’s diet, otherwise the animal refuses its food. Since the dog is a 
carnivor, it might be assumed that this vitamin plays some part in connection 
with protein metabolism. Karr, however, states that the intermediary meta- 
bolism of nitrogen is unaltered by adding vitamin B to the diet, but he appears 
to deduce such a conclusion from analysis of excreta and food ingested. It is 
possible that the nitrogen excretion might remain constant and yet the 
intermediary metabolism be altered. 

Thus there seems a certain amount of superficial evidence for associating 
vitamin B with protein metabolism. In the experiments described in this 
paper, all other recognised dietary constituents have been eliminated except 
vitamin B. This is only a negative proof and therefore of little value. The 
commercial preparation marmite is a somewhat crude product and contains 
many other substances besides vitamin B. It is even possible that extractives 
are of use in obviating the bad symptoms resulting from excessive amounts 
of protein in the mother’s diet. 


(ii) The amount of protein. 


At present it is impossible to state what amount of protein is harmful and 
what amount is beneficial, because it appears to depend largely on the other 
dietary constituents. The experiments described in this paper show that too 
much protein depresses the growth curve, even if no bad symptoms are shown. 
For example when the mother is fed on bread and protein (15 : 5) and whole 
milk, the litter show a less good growth curve than when the mother is fed 
on bread and milk only, although the babies are normal in each case. In the 
former experiment the rat was taking at least 8-0 g. protein per day, but no 
bad symptoms occurred, yet in a less complete diet 1-0 g. protein can produce 
a very bad condition in the young. 

As the results differ with different proteins it is probably certain amino- 
acids which are primarily responsible for the bad effects on the babies. If 
this is so, a diet of mixed proteins (such as taken by man) should prove less 
harmful, unless the protein is taken in very great excess. The actual com- 
position of foods themselves is a safeguard, because it is rare to find only one 
protein. 

(iii) Development of the young. 

It has been pointed out that the time of onset of the symptoms is re- 
markably constant and it is therefore suggested that the stage of development 
of the young rat is a factor to be considered. In foster litters, the babies 
develop a little more slowly and their eyes open about two days later than 
those of normal litters. They do not eat for themselves until a few days later 
than the original babies and therefore cannot be weaned at the normal time. 


104 G. A. HARTWELL 


(It has been found advisable to leave the mother with the young for at least 
three days after they can eat for themselves.) In general they seem more 
backward and less developed than normal babies of the same age. In foster 
litters, when the mother is fed on an excess protein diet, the symptoms 
usually develop about 2-3 days later, than in a litter left with its own mother. 
Also the symptoms come on at the same time, even if the mother has been 
fed with excess protein during gestation, or if she has first fed her own litter 
when she has had the excess protein diet. If the composition of the milk 
changed towards the middle and end of lactation, then the symptoms in the 
foster-litter should come on at an earlier stage, provided that the excess 
protein diet were continued throughout. But this is not so. | 


(iv) The value of milk in the diet of a lactating animal. 

“Milk in liberal amounts should always be included in the diet of the 
lactating mother.” This statement is made by McCollum [1920] and is 
probably a very true one. It was shown [Hartwell, 1921, 1] that bread and 
milk fed to the mother produced maximal growth of the young. Various 
other diets have been tried but as yet no curves have been obtained which 
are as good (except 15-0 g. bread to 5-0 g. protein + whey from 100 ce. milk— 
and this is a milk derivative). It has been proved in this paper that large 
quantities of milk entirely obviate the bad symptoms produced by excess’ 
protein in the diet. It is, therefore, possible that by taking large amounts of 
milk, the mother safeguards herself from any errors in diet. 

The value of milk is, no doubt, due to the fact that it contains all the 
necessary constituents of a good diet, and not excessive amounts of any. It 
has been suggested by some physiologists that milk contains a specific galacto- 
gogue. Some experiments now in progress show that there are other diets 
which will give as good results as bread and milk (at any rate with rats) 
provided that the right proportion of the various constituents can be ascer- 
tained. 

Another advantage of milk is that the proteins are adequate and it is 
quite obvious that the quality of the protein fed to a lactating animal is of 
primary importance. 

_SumMARY. 


1. In a lactating rat, the amount of dietary protein constituting excess 
varies with the type of protein and with the individuality of the rat. The 
proportion 15-0 g. bread to 1-0 g. protein (egg albumin or caseinogen) fed to 
the mother results in abnormal symptoms in the litters. 

2. On a diet of bread and egg albumin (either 1-0 g. or 2-0 g. egg albumin 
to 15 g. bread) the mother loses a considerable amount, in one case as much 
as one-third of her body weight. 

3. When excess protein is fed to the mother, the growth curves of the 
litters are not constant. The variations obtained are probably due to the 
individual metabolism of the different rats. 


MAMMARY SECRETION IN RATS 105 


4. Large quantities of protein fed to a nursing rat are detrimental to the 
young, even when the diet contains all the essential constituents. 

5. If such a diet is started at birth, some babies may survive, but they 
are not normal; the majority die. 

6. If the diet is begun before or during gestation, none of the young 
survive. 

7. The litters die at approximately the same time, irrespective of the 
length of time the mother has been fed on the excess protein diet. 

8. Foster babies suffer in exactly the same way as the opel family if 
the same diet is continued. 

9. The typical symptoms can be induced in a healthy litter by giving them 
to a mother whose own litter has developed this condition. 

10. It is suggested that the toxic substances in the milk affect the baby 
rat at some special stage of development. 

11. A diet containing large amounts of protein (15-0 g. bread to 5-0 g. 
protein) in a physiologically complete mixture is adequate for growth, fertili- 
sation and reproduction in the rat, and is unsuitable only during lactation. 

12. The bad effects are entirely obviated by adding 100 cc. whole milk to 
the mother’s diet, but the growth of the litter is not maximal. This applies 
to caseinogen, edestin, egg albumin and gelatin. With caseinogen and edestin, 
100 cc. whey are equally effective. 

13. With caseinogen, edestin and egg albumin the addition of calcium 
lactate to the diet prolongs the life of the young, but the symptoms are just 
as severe. Milk ash has a similar effect when added to a diet containing excess 
egg albumin or edestin. 

14. The addition of butter and lactose causes no improvement in the 
young when the mother is eating large quantities of edestin or caseinogen. 

15. Yeast extract (marmite) in large amounts added to the excess protein 
diet prevents all bad symptoms in the case of caseinogen and edestin; with 
egg albumin the bad condition is improved, but not entirely cured. 


I wish to express my thanks to Prof. V. H. Mottram for his interest in 
this work, the expenses of which were defrayed by a grant from the Medical 
Research Council. 

REFERENCES. 


Donaldson (1906). Boas Anniv. Vol., 5. 

Hartwell (1921, 1). Biochem. J. 15, 140. 
- Hartwell (1921, 2). Biochem. J. 15, 563. 

Hewer (1914). J. Physiol. 47, 480. 

McCollum and Simmonds (1918). Amer. J. Physiol. 46, 275. 
’ McCollum (1920). Newer Knowledge of Nutrition, 129. 

Matthews (1921). Physiological Chemistry, 306. 

Osborne and Mendel (1911). Feeding experiments. 


XV. EFFECT OF SEVERE MUSCULAR WORK 
ON COMPOSITION OF THE URINE. 


By JAMES ARGYLL CAMPBELL 
AND THOMAS ARTHUR WEBSTER. 


From the Department of Applied Physiology, National Institute 
for Medical Research, Hampstead. 


(Received January 23rd, 1922.) 


A FAIRLY complete series of observations was undertaken to determine the 
effect of severe muscular work on the composition of the urine, using recent 
standard methods. These observations form part of a research connected with 
industrial fatigue, the subject being employed at the same time for obser- 
vations on respiratory exchange. 

In a previous paper [1921] we recorded the results of observations on day 
and night urine under various routines—five days’ complete rest, five days’ 
ordinary laboratory work, and five days’ light muscular work, 67,500 kilogram- 
meters (kgm.) per day. In the present paper we give results obtained from 
the same subject, under similar conditions as regards diet, time of meals, etc., 
during five days’ severe muscular work. The diet was controlled as regards 
quality. About 4 litre more fluid per diem was swallowed during the severe 
muscular work than during the other routines. 

Our subject was accustomed to do 13,500 kgm. per hour, on a bicycle 
ergometer, working three hours before dinner—10 a.m. to 1 p.m.—and two 
hours in the afternoon—2 p.m. to 4 p.m.—the total, 67,500 kgm. producing 
only a slight degree of fatigue. For severe muscular work he attempted to 
do 100,000 kgm. per day at the rate of 20,000 kgm. per hour with the same 
interval for dinner. 100,000 kgm. represents a full day’s work for a vigorous 
labourer, and, as might be expected, our subject—a laboratory assistant— 
found this amount of work a severe task. Table I shows the details of the 
work actually performed. On the first day the subject had to give up the 
attempt at the end of the third hour, the first hour’s work of 20,000 kgm. 
affecting him very much. On the second day he was able to do more and 
on the third, fourth and fifth days he completed the full amount in the 
five hours. It was obvious from the subjective and objective symptoms 
that the experiment produced severe strain. The average composition of 
the 24 hours’ urine for the five days is shown in Table II; the significance 
of any of the five days’ results was similar. For comparison we have given 
the figures for the 24 hours’ urine obtained in our previous research [1921]. 
[t will be observed that the undetermined nitrogen, creatinine, neutral 


URINE DURING SEVERE MUSCULAR WORK 107 


sulphur—and, therefore, the total sulphur—and lactic acid were distinctly 
increased during the severe muscular work. Acetone bodies were present 
during the third and fourth days and the third night, probably indicating 
complete consumption of available carbohydrate. Purine nitrogen was also 
increased, but we had not complete figures for comparison. Other observers 
have obtained somewhat similar results. Cathcart [1921] reviews the most 
important of the recent and older researches in his monograph on protein 
metabolism. 


Table I. 
Working 
Work done __ period in 
Date in kgm. hours Remarks 
10. x. 21 44,482 3 Felt very hot, perspired freely; very much fatigued; 
severe headache; sore all over 
ll. x. 21 72,440 5 Felt very hot, perspired freely; very much fatigued; 
felt stiff all over 
12. .x. 21 104,760 5 Felt very hot, perspired freely; very much fatigued; 
‘felt stiff all over; acetone bodies in urine 
13. x. 21 100,000 5 Not so hot; not so fatigued; not so stiff as on previous 
days; acetone bodies in urine 
14, x. 21 100,000 5 Not so hot; not so fatigued; not so stiff as on previous 
days 


Osterberg and Wolf [1907] showed that the undetermined nitrogen was 
increased during activity. Shaffer [1908] found that muscular activity— 
walking ten miles—did not increase the excretion of creatinine. Pekelharing 
and Harkink [1911] obtained a similar result, but observed that after pro- 
longed tonic contraction the output rose. Our subject exhibited muscular 
stiffness and this may have been a cause for the increased excretion of 
creatinine. Hoogenhuyze and Verploegh [1908] found that in fevers and 
other pathological conditions in which there is an increased breaking down 
of tissue, the creatinine excretion is increased. In our subject there was the 
possibility that some muscle fibres were injured. The brain contains a fair 
proportion of creatine [Janney and Blatherwick, 1915], and was a possible 
seat of production of creatinine during the severe fatigue produced. Weinberg 
[1921] considers that an important influence on the excretion of creatinine in 
the urine should be assigned to the mind, high excretion being connected with 
emotion. Scott and Hastings [1920], Garrat [1898] and others noted that 
there is a rise in output of inorganic sulphate during activity. Although we 
observed a similar increase of sulphate during the light muscular work it was 
the neutral sulphur that was most increased during the severe work. This 
may have been due to the want of oxygen, so that the sulphur was not so 
completely oxidised during the severe work. With regard to lactic acid, many 
observers have found that this acid is increased by severe muscular exercise. 
Fletcher and Hopkins [1907] clearly demonstrated that lactic acid is produced 
in excised muscle only when the muscular contraction occurs in a deficiency 
of O,. When it occurs in an adequate supply of O,, CO, instead of lactic acid 
is produced. Hill and Flack [1909] showed that in the fatigue of athletes 
oxygen inhalation increases the lasting power and decreases fatigue, probably 


108 J. A. CAMPBELL AND T. A. WEBSTER 


by maintaining or restoring the vigour of the heart. This leads to the question 
whether creatinine, neutral sulphur and undetermined nitrogen were increased 
because of insufficient oxidation. This was a possibility, but in one respect 
these substances differed greatly from lactic acid; lactic acid was excreted in 
much greater amount during the day (Table II) than at night, whereas 
creatinine, neutral sulphur and undetermined nitrogen were more evenly dis- 
tributed between day and night. With regard to the latter substances we did 
not notice in our previous researches [1921] that there was any. parallelism 
in these excretions; but, nevertheless, there may be the same explanation for 


Table IT. 
Average 24 hours results 
r A — Average hourly day and 
Light Severe night results during severe 
muscular work muscular work muscular work 
Complete Laboratory (67,500kgm. (100,000 kgm. ~ 
rest work in five hours) in five hours) Day Night 
Amount ce. 1112 1116 971 686 32-4 25-9 
Acidity % 49-5 53-7 58-0 65-0 60-0 70-0 
Titratable acidity 309 312 332 342 14-0 14-4 
(Folin) ec. N/10 
Total acidity cc. N/10 703 673 723 [Oy a 29-0 30-7 
Total N g. 8-010 (100)* — 9-015 (100) 9-465 (100) — 10-250 (100) -446 (100)* -414 (100)* 
Urea N g. 5-990 (74-80) 7-131 (79-10) 7-647 (80-80) _ 7-680 (74-93) 336 (75:3) +308 (74-5) 
Ammonia N (A) g. -439 (5-48) *385 (4-27) 442 (4-67) 412 (4-02) 015 (3-4) “019 (4-6) 
Ammonia N (B) g. -548 -505 +552 +533 ‘021 023 
Amino-acid N g. 109 (1-36) “120 (1-33) “110 (1-16) 121 (1-18) 006 (1-4) 004 (1-0) 
Creatinine N g. 741 (9-25) 736 (8-17) 695 (7-34) 925 (9-02) 036 (8-2) 040 (9-6) 
Uric acid N g. 154 (1-92) -141 (1-56) 133 (1-40) 134 (1-31) 006 (1-4) 005 (1-2) 
Undetermined N g. -578 (7-19) 502 (5-57) “438 (4-63) ‘978 (9-54) 047 (10-3) — -038 (9-1) 
Chloride (NaCl) g. 9-410 7-207 8-801 6-736 394 200 
Phosphate (P,0;) g. 1-550 1-703 1-711 1-811 070 080 
Total S (SO,) g. 1-594 (100) 1-749 (100) 1-770 (100) 2-069 (100) 084 (100) — -087 (100) 
Inorganic S (SO,) g. 1-148 (72-0) 1-400 (80-0) 1-432 (80-9) 1-422 (68-7) ‘057 (67-9) —_-061 (70-1) 
Ethereal 8 (SO,) g. 212 (13-3) 132 (7-5) -140 (7-9) 209 (10-1) 007 (8-3) “009 (10-4) 
Neutral 8 (SO,) g. 234 (14-7) -227 (12:5) +198 (11-2) -438 (21-2) 020 (23-8) — -017 (19-5) 
Lactic acid g. trace trace trace “061 0040 “0014 
Calcium (CaO) g. os 281 +252 — — — 
Magnesium (MgO) g. — 173 156 — — _ 
Purine N g. 035 — — 058 0023 +0025 


(A) Van Slyke’s method. (B) Malfatti’s method. 
* Figures in brackets are percentages. Total acidity is Ammonia (B) + titratable acidity. 


the increase of each during severe work. It is well known that creatinine and 
neutral sulphur resemble one another in that they are of endogenous origin, 
and that they are practically constant in amount for each individual, being 
evidently connected with the mass of active living protoplasm. In our subject 
both creatinine and neutral sulphur were higher on the first day than on the 
succeeding days of severe work. We consider that the probable explanation 
of the increased excretion of creatinine, neutral sulphur and undetermined 
nitrogen was damage to protoplasm, both nervous and muscular by the ex- 
cessive fatigue and strain. 

Cathcart [1921] states that no one has been able to demonstrate clearly 
that muscular work affects greatly the total nitrogen excretion, provided the 


URINE DURING SEVERE MUSCULAR WORK 109 


supply of food, particularly of carbohydrate and of oxygen be sufficient. 
Under our conditions of experiment we found a slight and steady increase 
in total nitrogen from complete rest to severe work (Table II). We also found 
that during the five days’ severe work the total nitrogen rose from 8 g. on 
the first day to 10 g. on the second day, to 12 g. on the third day and fell to 
10 g. for the fourth and fifth days, whereas the increments in undetermined 
nitrogen, neutral sulphur, creatinine were marked from and including the 
first day. Garratt [1898] found that a rise of sulphate preceded a rise in total 
nitrogen as a result of exercise. Whilst the figures for total nitrogen are of 
interest and of some significance and show that the excretion of total nitrogen 
was increased, we cannot conclude that there was an increased katabolism of 
nitrogen in our subject as the result of exercise, since we did not estimate the 
total nitrogen in the food and in the faeces. Under ordinary conditions our 
subject’s total nitrogen averaged only 9 g. per diem during six months and 
was never as high as 12 g. The large amount of nitrogen, 12 g. on the third 
day of severe work, was probably due to lack of carbohydrate as indicated 
by the presence of acetone bodies on that day. Many observers have found 
an increase of nitrogen when there is an inadequate supply of nitrogen-free 
food available. 

Because it was necessary to increase greatly the amount of work for our 
subject before any definite change in composition of urine was detected we 
consider that under equally good dietary and atmospheric conditions a healthy 
industrial worker would not show any such change in urinary composition 
unless undertaking work far more severe than his customary daily task. We 
have done a few experiments with some subjects sitting at rest under hot 
(36° C.) and close atmospheric conditions and found that the day urine showed 
higher figures for titratable acidity and ammonia than when under com- 
fortable atmospheric conditions. We hope to extend these observations. One 
of us [J. A. C.] has noted high excretion of ammonia in a hot climate [1919, 
1920). 

Table II also gives the average hourly day and night results during the 
five days’ severe work. As in our previous research [1921] we found that the 
acidity, ammonia and phosphate were higher at night than during the day; 
also that the sulphur was evenly distributed between day and night, whilst 
the total nitrogen was higher during the day than at night. We consider that 
during the night there is an excretion of fixed acids which have been formed 
in the cells during the day and that these fixed acids are concerned in the 
production of fatigue and sleep. For all routines examined we observed that 
the differences between day and night urine were most marked during rest 
in bed and less marked as the activity increased, activity probably hastening 
the processes concerned. 

We are indebted to Mr C. Pergande, who was the subject in this research. 


110 J. A. CAMPBELL AND T. A. WEBSTER 


SUMMARY. 


A healthy subject, who was accustomed to do 67,500 kgm. of work, on a 
bicycle ergometer, in five hours, developed symptoms of muscular strain on 
attempting to do 100,000 kgm. of work in five hours and showed pathological 
changes in urinary composition. Creatinine, undetermined nitrogen, neutral 
sulphur and lactic acid were much increased, whilst acetone bodies were 
present during part of the experiment. 

Our previous results regarding composition of day and night urine were 
confirmed. 


REFERENCES. 


Campbell (1919). Biochem. J. 13, 239. 

—— (1920). Biochem. J. 14, 603. 

Campbell and Webster (1921). Biochem. J. 15, 660. 

Cathcart (1921). The Physiology of Protein Metabolism, 129. 
Fletcher and Hopkins (1907). J. Physiol. 35, 247. 

Garratt (1898). J. Physiol. 23, 150. 

Hill and Flack (1909). J. Physiol. 38, Proc. xxvm1. 

Hoogenhuyze and Verploegh (1908). Zeitsch. physiol. Chem. 57, 161. 
Janney and Blatherwick (1915). J. Biol. Chem. 21, 567. 

Osterberg and Wolf (1907). J. Biol. Chem. 3, 169. 

Pekelharing and Harkink (1911). Zectsch. physiol. Chem. 75, 207. 
Scott and Hastings (1920). Quoted in J. Indust. Hyg. Feb. 1921, 203. 
Shaffer (1908). Amer. J. Physiol. 22, 445. 

Weinberg (1921). Biochem, J. 15, 306. 


XVI. THE VALUE OF GELATIN IN RELATION 
TO THE NITROGEN REQUIREMENTS 
OF MAN. 


By ROBERT ROBISON. 
From the Lister Institute. 


(Recewed January 23rd, 1922.) 


THE story of the earliest attempts to discover the value of gelatin as a food- 
stuff has been told by Carl Voit [1872] in the introduction to his paper on this 
subject. It commences in 1682 when Dionys Papin prepared gelatin from 
bones by means of his digestor, and from the gelatin made soup with which 
he fed the poor. Such attempts were zealously renewed during the French 
Revolution by Cadet de Vaux, d’Arcet and others, and were supported by 
the Government, who issued official instructions extolling the nourishing pro- 
perties of gelatin soup above those of beef tea. The approbation of the 
Institute of France and of the Academy of Medicine was also forthcoming, 
but in spite of d’Arcet’s attempts to improve the flavour of the soup with 
spices, it did not meet with very great approval from the poor, who were 
expected to consume it. 

Gannal, a manufacturer of gelatin, fed himself and his family on gelatin, 
with and without bread, for some weeks until compelled to desist owing to 
the unsupportable nausea caused by the diet. The effects on the health of 
these people led him to conclude that gelatin is not only valueless as a food 
but actually harmful. 

Magendie’s Report in 1841 to the Paris Academy on the results of the 
investigations of the second Gelatin Commission was scarcely more favourable. 
Gelatin was considered to have no food value by itself and to reduce the value 
of other foodstuffs when fed in combination with them. A similar opinion 
was expressed by the Academy of Medicine in 1850 but less extreme views 
were held by some physiologists among whom were Boussingault [1846] and 
Frerichs [1845]. The latter ascribed to gelatin the same significance as that 
of the “Luxus” protein, 7.e. the excess protein in the diet over the require- 
ments of the body as represented by the protein decomposition during starva- 
tion. Though unable to replace the body protein gelatin could, he held, be 
utilised in the same way as the nitrogen-free foodstuffs (“ Respirationsmitteln’’). 

Somewhat similar views were held by Bischoff [1853] while Donders [1853] 
considered that gelatin might reduce the body’s needs for protein s since these 
are not restricted merely to the replacement of tissues. 


112 R. ROBISON 


Voit’s own experiments and those carried out in conjunction with Bischoff 
on dogs form the first systematic study of the nitrogen balances on diets con- 
taining varying quantities of meat, gelatin and fat. As the result of a large 
number of experiments he concluded that gelatin always spares protein and 
in a greater degree than fat or carbohydrate, but that gelatin plus fat reduces 
the protein decomposition more than does gelatin alone. On the other hand, 
however much gelatin and fat are given, body protein will be lost. 

The energy requirements of the animal were not sufficiently considered 
in these investigations and a large part of the so-called sparing action of 
gelatin can be ascribed to its ability to furnish energy and so to reduce the 
use of body protein for this purpose. I do not suggest that the whole effect 
is to be explained in this way, but to what extent gelatin can satisfy any 
portion of the dog’s specific nitrogen requirements cannot be ascertained from 
Voit’s results. 

During the next thirty years experiments upon dogs by feeding with gelatin 
were also carried out by Oerum[1879], Pollitzer[1885], Munk [1894], Kirchmann 
[1900] and Krummacher [1901]. In Oerum’s experiments the dog received a 
basal diet of starch, butter and meat extract equivalent to over 80 calories 
per kilo body weight. During successive periods of from four to eight days 
this diet was supplemented by meat or the equivalent amount of gelatin. 
Unfortunately only the urea nitrogen was determined (by Liebig’s titration 
method) and the faeces were not analysed, so that a nitrogen balance sheet 
cannot be made out, but the results appear to indicate that gelatin can save 
about half the amount of nitrogen excreted by a dog when receiving a carbo- 
hydrate diet of sufficient calorie value. 

Pollitzer also gave his dog abundant carbohydrate to which, during suc- 
cessive periods, were added equivalent quantities of meat, digestion products 
of meat (peptone, etc.) and gelatin. Positive nitrogen Beggs were obtained 
with all except gelatin. 

Kirchmann determined the amount of body protein spared by different 
amounts of gelatin, no other food except water being given. Taking the 
nitrogen output during starvation as 100 he found a saving of 25% when 
the gelatin given was sufficient to satisfy only 7-5 % of the energy require- 
ments of the animal, while eight times this amount was required to save 35 %,. 
He estimated that a maximum saving of 39% might be expected if the 
amount of gelatin could be increased to meet these energy requirements in full. 

Krummacher continued these experiments with still greater quantities of 
gelatin, and obtained a result closely agreeing with Kirchmann’s calculated 
figures. It is clear however that unless we also know the minimum nitrogen 
output when the energy requirements are fully met by nitrogen-free food- 
stuffs, the above relationships offer no evidence as to the capacity of gelatin 
to satisfy any of the specific nitrogen requirements of the animal. This value 
was not determined by either Kirchmann or Krummacher. 

Munk’s experiments were on a different plan from those mentioned above. 


GELATIN AS A FOODSTUFF 113 


He gave a dog a diet of rice, fat and meat equal to 58 calories per kilo and 
containing 0-6 g. nitrogen per kilo body weight, which was more than twice 
the starvation output. He was then able to replace five-sixths of this protein 
by the equivalent amount of gelatin and still keep the animal in nitrogen 
equilibrium. 

Kauffmann [1905] carried out a series of experiments on a similar plan 
but with the precaution of reducing the nitrogen intake of the standard diet 
to a much smaller amount than that given by Munk. This standard diet 
consisted of milk, rice, caseinogen (plasmon) and fat and was given in amount 
equal to 0-32-0-39 g. N and 63-72 calories per kilo body weight. Not more 
than one-fifth of the nitrogen of this diet could be replaced by gelatin nitrogen 
without an increase in the nitrogen output occurring. Kauffmann also in- 
vestigated the possibility of improving the value of gelatin by supplementing 
it with tyrosine, tryptophan and cystine and concluded from his experiments 
on dogs and on himself that with these additions ‘gelatin becomes of equal 
value with caseinogen. 

At a much earlier date Escher [1876] had fed dogs and pigs with gelatin 
supplemented by tyrosine and found that their body weight was maintained, 
but Lehmann [1885] was unable to obtain this result in experiments on rats. 

Rona and Miiller [1906] carried out a series of very careful experiments 
with dogs on the same plan as those of Kauffmann but were unable to confirm 
the latter’s conclusions. Their standard diet gave 0-2 g. N and 91 calories 
per kilo body weight. With this the animal was in nitrogen equilibrium, but 
when a portion of the milk was replaced by gelatin plus tyrosine and trypto- 
phan a negative balance was found. 

The value of gelatin fed in conjunction with other proteins has also been 
investigated by Murlin [1907, 1] both by experiments on dogs and on himself. 
With dogs on a diet containing one-fourth more than the fasting requirement 
of nitrogen, half of this being in the form of cracker meal and half in the form 
of caseinogen, it was not possible to. replace the caseinogen nitrogen by gelatin 
nitrogen without increased loss of body protein. With other diets however, 
in which the protein was in the form of meat, up to 58 % could be replaced 
without loss of body protein. The fuel value of all diets was greater than the 
energy requirements of the animal but Murlin attributes the high replace- 
ment value obtained in some diets largely to the greater proportion of calories 
supplied by carbohydrate in place of fat. 

It is possible that this factor may have influenced the result, though 
according to Zeller [1914] the nitrogen requirements are not affected by the 

_ proportion of fat to carbohydrate in the diet so long as this does not become 
greater than about 4 : 1. 

There seems however to be insufficient reason for assuming that the amount 
of meat given in some of these diets was the minimum required for nitrogen 
equilibrium, and unless this were so the fact that a part could be replaced by 
gelatin without affecting the balance would prove nothing. 


Bioch. xvr 8 


114 R. ROBISON 


On the other hand, on the cracker meal diets a negative balance was 
always obtained, which was recognised by Murlin as evidence of the lower 
availability of this form of protein. On these diets it was not possible to 
replace any part of the protein by gelatin without increasing the relative loss 
of body nitrogen. . 

The criticism that the protein in the diet after part of it had been replaced 
by gelatin may have still been in excess of the minimum required, applies 
even more forcibly to the experiment on himself, in which the basal diet 
contained 14-25 g. N, z.e. about 10% more than his nitrogen output during 
starvation. When two-thirds of this had been replaced by gelatin nitrogen 
he was still receiving 5-33 g. N (0-076 g. per kilo) derived from eggs, cream, 
butter and cereals. During the two days on which this diet was taken a 
positive balance was obtained, but this cannot be accepted as convincing 
evidence of the value of gelatin nitrogen. 

In a later paper Murlin [1907, 2] brought forward satisfactory proof that in 
a dog the reduction (about 30 %) of the fasting nitrogen output produced by 
small amounts of gelatin, was much greater than could possibly be accounted 
for by the dextrose which might be synthesised in the body from this 
gelatin!. 

From the investigations so far considered it may be taken as definitely 
established that: . 

1. Gelatin when given as the sole source of nitrogen is unable to maintain 
the animal body in nitrogen equilibrium. 

2. With dogs gelatin is able to reduce the loss of body nitrogen considerably 
below that occurring during starvation, and this effect is not proportional to 
the amount of potential energy thus supplied and cannot therefore be simply 
explained on these grounds. 

3. Some of the experiments indicate that when gelatin is mixed with 
other proteins, they may complement one another so that a proportion of 
the nitrogen of gelatin is utilisable. 

A critical examination of the results of these experiments does not enable 
us to form any definite conclusions as to the capacity of gelatin alone to 
satisfy any part of the specific nitrogen needs of the body in man, although 
some of the results obtained with dogs indicate a limited capacity in this 
direction if the nitrogen output on an abundant nitrogen-free diet is taken 
as representing these specific requirements. There is however a difficulty in 
accepting this since the results obtained with dogs do not fall into line with 
those obtained with man and some other animals, and suggest that the 
nitrogen metabolism of the carnivora varies from that of the omnivora and 
herbivora in some details. 

The fasting output of a man is equal to about 0-2 g. N per kilo body 


' A brief account of other researches by Ganz, Gerlach (1891), who investigated the value 
of gelatin peptones, Gregor (1901), who used gelatin for feeding infants, and by Brat (1902) and 
Mancini (1905), who fed it to convalescents, will be found in Murlin’s paper [1907, 1]. . 


GELATIN AS A FOODSTUFF 115 


weight, that of a large dog is of the same order. On an abundant carbo- 
hydrate diet the nitrogen output of man can be reduced to one-quarter of 
this amount, i.e. 0-05 g. N per kilo whereas according to most observations 
under the same circumstances the nitrogen output of a dog is only reduced 
by 10 % to 20 %. 

The difference in detail between the nitrogen metabolism of man and dog 
also emerges on comparison of the ratios of the total nitrogen to that excreted 
in the form of creatinine during starvation and on abundant nitrogen-free diets 
(see Table I). The constancy of the creatinine output and its probable re- 
lationship to the endogenous metabolism has been noted by Folin [1905], 
McCollum [1911], Zeller [1914] and others. 


Table I 
cae sieages Oneatiuitie 
Weight erkilo ‘per kilo Total urine N 
i Observer Animal Kg. Diet y weight y weight Creatinine N 
Cathcart [1907] Man V.B. 62-0 Fasting, 4th day 0-221 0-0056 39 
5 Bs FF ns 60-0 = 8th ,, 0-159 0-0053 30 
Benedict and Dog 39 76 a 3rd_,, =: 0360 0-0099 36 
Osterberg [1914] 
bs re 53° 188 12-7 ba 3rd", 0-280 0-0112 25 
Towles and mn 3 9-0 Bs 2nd 5, 0-294 0-0129 23 
Voegtlin [1912] 
Murlin [1907, 2] ” C 13-0 ” 4th ” 0-257 0-0080 32 
Folin [1905] Man H.B.H. 85-7 {mu cream, 0-0420 0-0070 6-0 
lg. 
Graham and Pie C5 t 62-4 Starch, cream, 0-0445 0-0093 4:8 
Poulton [1912] 912g. N 
re re PES Ale 72-4 Starch, cream, 0-0468 0-0107 4-4 
1-23 g. N 
af Klercker [1907] ,,  a.K. 88-0 Low N 0-0319 0-0079 4: 
Robison [1922] »  C.J.M. 60-5 ga 0-0352 0-0072 4:9 
at, -3 g. 
> it eae. S:: o 58-0 Bo 0-0355 0-0084 4:2 
McCollum [1911] Pig 10:9 Carbohydrate 0-0495 0-0095 5-2 
s a “ 68-4 ie 0-:0387 0-0069 5-6 
Mendel and Rose Rabbit 1-74 . 0-126 0-0172 7-3 
[1911] 
Murlin [1907, 2] Dog C 11-3 me 0-158 0-0104 15 


There is a close parallelism between the figures for men and dogs during 
starvation and between those for men, pigs and rabbits on abundant nitrogen- 
free diets. The creatinine excretion for Murlin’s dog C on such a diet is also in 
good agreement with the corresponding figures for men and pigs but the ratio 


Total urine N i. about three times as high as the same ratio for other animals. 
Creatinine N 


It is of course not possible to state on such evidence alone that the real endo- 
genous metabolism of this dog should be represented by a nitrogen output. 
of one-third the observed amount, but it is clear that the nitrogen metabolism 
of dogs differs in some way from that of man, and that caution must be used 
in applying conclusions from experiments with these carnivora to other 
animals and man. 

These criticisms however do not apply to the experiments of McCollum 
[1911] on pigs, for in these the constancy of the proportion of the endogenous 

8—2 


116 R. ROBISON 


metabolism represented by creatinine nitrogen was recognised and was used 
as a criterion for judging when the minimum nitrogen excretion of the animals 
had been reached. 

The pigs were fed on a basal nitrogen-free diet of ample fuel value con- 
sisting of starch, a salt mixture and water, until the nitrogen output had 
reached the minimum, whereupon an amount of the protein under examina- 
tion equivalent to this minimum (urine nitrogen only) was added to the diet 
during a further period, after which the basal diet alone was fed until the 
output had again fallen to the minimum, the nitrogen exereted during this 
last period being also included in the calculation. In the experiment recorded 
by McCollum 2-62 g. of gelatin nitrogen was given daily during eight days, 
z.e. 20-96 g. in all. The total output during these eight and the following 
four days on which no nitrogen was given, amounted to 41-71 g. in the urine 
and 12-48 g. in the faeces, 7.e. 54-19 g. in all, making a negative balance of 
33°23 g. 

Had the pig received no nitrogen at all its total output during these twelve 
days would have amounted to 31-44 g. in the urine and 12-48 g. in the faeces, 
making 43-92 g. in all, so that a saving of 10-69 g. nitrogen has been effected 
by 20-96 g. of gelatin nitrogen. This implies a utilisation of 50% of the 
nitrogen given in this form, which was confirmed by five other similar experi- 
ments the details of which are not given. If the result is stated in terms of 
body protein saved, this amounts to 1-34 g. per day (if reckoned on eight 
days), i.e. 37 %, of the minimum output in urine and faeces or 51 % of that 
in the urine only, which is taken by McCollum as representing the essential 
tissue metabolism of the animal. 

Boruttau [1919] has recently attempted to determine the biological value 
of gelatin by two experiments on dogs, using the method and formulae 
adopted by Karl Thomas [1909] and has obtained the figures 49-1 % caleu- 
lated by formula I and 67-3 % calculated by formula II. These values would 
agree much more closely had Boruttau not made an error in his use of 
formula I by taking the total food nitrogen as denominator in place of this 
amount less the nitrogen of the faeces, as intended by Thomas. In any case 
however such figures have no real significance in the case of gelatin since they 
will necessarily vary with the amount of the intake, and moreover the experi- 
ments were of too short duration to possess much value. 

Apart then from the experiments of McCollum no very satisfactory evidence 
has been produced regarding the value of gelatin alone to satisfy any of the 
nitrogen requirements of the animal body. Most of the investigations have 
in fact been concerned with its value when fed in conjunction with other 
proteins and this introduces the possibility of complementary effect, about 
which very: little is definitely known. That such effect is possible is shown by 
the experiments of Osborne and Mendel [1912] on rats. With gelatin as the 
sole protein the animals rapidly declined in weight but recovered when half 
of the gelatin was replaced by gliadin, a protein incapable of inducing more 


GELATIN AS A FOODSTUFF 117 


than a very slight growth when fed as the sole protein constituent of the diet. 
Further, almost all the previous work, including that of McCollum, has been 
carried out on animals, and the results might not necessarily apply to man. 
The whole question is of very great theoretical and practical interest because 
of its bearing on protein metabolism in general and the nature of the body’s 
requirements for particular compounds of nitrogen. 


EXPERIMENTAL. 


The investigation about to be described was an attempt to obtain more 
light on the problem by direct experiments on man. 

The subject of the experiment was myself, age 37 years, medium build, 
weight 59 kilo, height 173-5cem. My minimum nitrogen output had been 
determined by previous experiments which will be discussed in another paper. 

In the second of these experiments, in which a diet containing about 
0-3 g. N and equivalent to 2600-3000 calories (45-52 cals. per kilo) was taken 
for a period of seven days, the nitrogen output in the urine fell to a fairly 
constant level of 2-06 g., while the average amount of nitrogen excreted in 
the faeces was 1-13 g. per day. 

In the present investigation the basal diet supplemented by different 
quantities of gelatin was taken for periods of ten days, the nitrogen intake 
being kept absolutely constant during each period. 

Profiting by the experience of the previous experiments the basal diet 
was somewhat altered, the original attempt to introduce some variety and 
palatability being given up in favour of greater simplicity and uniformity of 
the food intake. The proportion of calories supplied by fat and the total 
nitrogen in the diet were both reduced. In the later experiments the process 
of simplification was carried to its furthest extent, the diet consisting of corn 
starch, lactose, sucrose and a salt mixture. Minimal quantities of lemon 
juice and cod liver oil were added to supply the antiscorbutic and fat soluble A 
accessory factors and agar-agar was taken to increase the bulk of the faeces 
and prevent constipation. The corn starch, lactose, salt mixture and agar 
for each day’s ration were weighed out and mixed together before the experi- 
ment began. The mixture was taken in the form of a cream made with cold 
or warm (but not boiling) water and washed down with more water. The 
uncooked starch grains were very well absorbed, extremely few being found 
in the faeces. Usually a third of the day’s ration was taken at 8 a.m., I p.m. 
and 7 p.m., but sometimes it was found necessary to increase the number of 
meals in order to consume the prescribed amount. The gelatin was dissolved 
in warm water and taken either by itself or mixed with some of the starch 
and lactose. The lemon juice, sweetened with cane sugar, was taken as a 
drink and a little weak tea with lemon was also permitted. The very small ' 
amount of nitrogen in the tea was assumed to be due to caffeine and to*he 
excreted unchanged in the urine. It was therefore always subtracted from 
the total nitrogen intake and from the output. 


118 R. ROBISON 


The salt mixture had the following composition: 


Calcium diacid phosphate CaH,(PO,),,H,O  20\ Na 25% 
Calcium lactate (C;H;03),Ca, 5H,O 30; K 13-5 
Potassium hydrogen phosphate K,HPO, 30| Needy 
Sodium dihydrogen phosphate NaH,PO,, H,O 15; P_ 13-6 
Magnesium carbonate MgCO, 3| Mg’ -85 
‘Tron carbonate” 2p BO, 27 


It was intended that 10 g. of this mixture with the addition of 5 g. sodium 
chloride should be taken daily. The amounts of calcium and phosphorus 
would then correspond with those recommended by Sherman as 50 % above 
the minimum requirements of the body for these elements [Sherman, Wheeler, 
and Yates, 1918; Sherman, 1920]. It was found necessary however to reduce 
these quantities to 6g. aad 4g. respectively on account of the diarrhoea 
caused by the diet. The ash of the above salt mixture is markedly alkaline, 
a point of importance in view of the observations by McCollum and. Hoagland 
[1913] on the increased nitrogen output caused by diets having an acid ash. 

The urine was collected from 8 a.m. to 8 a.m. and stored under toluene. 
The faeces were collected over the whole period and mixed with dilute 
sulphuric acid, those passed during the morning being considered as belonging 
to the previous day. Owing to the fluid consistency of the faeces the use of 
markers was found to be impracticable, but in view of the regular evacuation 
of the intestines and the length of the experiment, no serious errors can have 
been introduced in this way. Estimations of nitrogen in urine, faeces and in 
all components of the diet were carried out by the Kjeldahl method in dupli- 
cate. Creatinine was estimated by the method of Folin. 

The percentages of nitrogen found in the constituents of the diet and their. 
fuel values are given in Table II. 


Table IT. 
Nitrogen per Calories per 
100 g. 100 g. 

Gelatin (Coignet’s “‘ Extra.” Gold Label) 14-16 324 
Corn starch (a) 0-039 360 

7 ike) 0:027 360 
Dextrin 0-065 360 
Agar-agar 0-242 _ 
Lactose 0-013 370 
Sucrose aa 395 
Butter 0-080 775 
Cod liver oil Not determined 930 
Lemon juice (per 100 ce.) 0-067 40 
Vermouth a 0-005 140 
Tea infusion* re 0-008 oo 


* The strength of the tea infusion was kept as nearly constant as possible but the total 
nitrogen intake from this source was checked by removing an aliquot portion of all tea drunk 
during an experiment and estimating the nitrogen in the whole quantity. 


GELATIN AS A FOODSTUFF 119 
Concerning the purity of the gelatin used in the experiments. 


The source from which commercial gelatin is obtained and the methods 
employed in its manufacture are not likely to produce a pure product. One 
_ would expect to find it contaminated with traces of other animal proteins or 
their decomposition products. Such included impurities being colloids could 
not be removed by washing, and might conceivably possess a high value for 
the replacement of body nitrogen. 

Kirchmann drew attention to the fact that the best French gelatin gave 
a slight precipitate with Millon’s reagent and with potassium ferrocyanide 
and acetic acid, and claimed to have succeeded in removing the impurities 
to which these reactions were due. The unoxidised sulphur was also reduced 
from 0-387 % to 0-263 °%. He considered that the difference between his 
own results and those of previous workers was to be attributed largely to the 
presence of this protein in their gelatin. One of his methods consisted in 
soaking the gelatin first in water, then in 10% sodium chloride solution, 
again in water and finally in alcohol. Murlin [1907, 1], using the same methods, 
was unable to detect any improvement in the purity of the product. 

The gelatin used in the present investigation also gave a slight positive 
reaction with Millon’s reagent and with potassium ferrocyanide and acetic 
acid, and an attempt was therefore made to purify it by soaking it for 24 hours 
in N/20 HCl followed by N/20 NaOH, then for some days in running water. 
No appreciable reduction in the intensity of the colour produced with Millon’s 
reagent was observed after such treatment. 

Folin and Denis [1912] have recorded finding a trace of tyrosine in gelatin, 
using Folin’s colorimetric method, and Dakin [1920] has recently obtained a 
similar result using a gravimetric method. He estimates the amount of 
tyrosine at about 0-01 % and considers that it cannot be an integral part of 
the gelatin molecule. 

I attempted to estimate the amount of tyrosine present by means of 
Folin’s method, using relatively large quantities of gelatin. The tyrosine in 
a sample of dried ox muscle was also estimated by the same method. The 
results are shown in Table III. Millon’s reaction is not well adapted for 
colorimetric measurement but under suitable conditions it was found possible 
to make approximate determinations by comparing the colour with that 
developed by different amounts of pure tyrosine, and the results agree reason- 
ably well with those obtained by Folin’s method. 


Table ITI. 
Tyrosine estimated by Tyrosine estimated with 
Folin’s method Millon’s reagent 
Gelatin (Coignet’s extra) 0-57 % 0:6 % to 0-7 % 
» after purification 0-45 —_ 


47 bast 


120 R. ROBISON 


The accuracy of Folin’s method has been called in question by Abderhalden 
[1913, 1, 2] who has suggested that other amino-acids, tryptophan, hydroxy- 
tryptophan and hydroxyproline give the same colour reaction. Of these the 
first two are not present in gelatin but Dakin estimated the amount of hydroxy- 
proline as 14-1 %. Through the kindness of Prof. Leathes, F.R.S., who supplied 
me with a specimen of this amino-acid, I was able to test its behaviour with 
Folin’s reagent and found that a slight colour developed under the conditions 
laid down by Folin and Denis for the estimation of tyrosine, but that the 
intensity was about ;+,; of that produced by the latter. compound. 

A slight colour, similar to that given by hydroxyproline, was also obtained 
from a specimen of phenylalanine. 

In the face of the results given by gravimetric methods it would be rash 
to assert that the gelatin actually contained 0-57 % of tyrosine, but the colour 
produced with Folin’s reagent does not appear to be due to any of the other 
amino-acids known to be present. It is also probable that the same compound, 
tyrosine or other amino-acid, is the cause of the colour produced with Millon’s 
reagent. 

If the percentage of tyrosine is correct and if it is present as a constituent 
of another protein similar to ox muscle, the proportion of the latter in the 
gelatin would be about 10 %. This calculation however is based on too many 
assumptions to be of more than speculative interest. 

Up to the present neither cystine: nor any other dinisemiad containing 
sulphur has been isolated from gelatin though the presence of such unidentified 
compounds has been noted by Dakin [1920]. 

The gelatin used in these experiments after purification in the manner 
described above, contained 0-24 % of total sulphur, calculated on the dry 
substance. Krummacher [1903] after purifying gelatin by Kirchmann’s 
method found 0-28 % 8 (of which 0-02 % was in the form of SO, and SQ,). 
The original commercial product used by Krummacher contained 0-62 % 
total 8, of which 0-4 °% was present as SO, and SO,. Such an amount (0-24 %) . 
of unoxidised sulphur would correspond with 0-9 °% of cystine (or other com- 
pound containing a like proportion of 8) and this can hardly be ascribed to 
impurities in the gelatin. 


RESULTS OF THE EXPERIMENTS ON GELATIN DIETS, 


It was proposed to carry out three diet experiments in which low, medium 
and high amounts of gelatin nitrogen should be given in addition to the basal 
diet, in order to determine 

(1) whether the minimum nitrogen loss on abundant nitrogen-free diet 
can be still further reduced by gelatin, and if so, 

(2) what relation the amount of this reduction bears to the amount of 
gelatin ingested, 

Two experiments were completed during the early part of 1921, but the 


GELATIN. AS A FOODSTUFF 121 


third had to be broken off through illness shortly after it was begun. It was 
repeated in September 1921 the conditions being somewhat modified on 
account of certain results that had in the meantime been obtained from other 
experiments carried out with Prof. C. J. Martin. These appeared to indicate 
that the amount of certain proteins required for nitrogen equilibrium could 
be greatly reduced if a carbohydrate diet very much in excess of the energy 
requirements was taken. In this last experiment, therefore, I increased the 
fuel value of the diet to the maximum that could be tolerated, so that the 
body weight was maintained and even increased during the first half of the 
period (see Fig. 1, Curve C). Synthesis of fat from the carbohydrate of the 
food was also indicated by the high respiratory quotient. 

The diets for the three experiments are given together in Table IV. The 
diet taken during the experiment in which my minimum requirements were 
determined is also included for purposes of comparison. 


Table IV. 
Nitrogen minimum Gelatin I Gelatin II Gelatin LIT 
29/11/20—5/12/20 28/1/21—6/2/21 18/4/21—27/4/21 8/9/21—-17/9/21 


Total calories=2605 Total calories=2525 Total calories=2757 Total calories=3256 
Calories per kilo=:45" Calories per kilo=44 Calories per kilo=47 Calories per kilo=54 
masts (hae as Calories nd a as Calories supplied as Calories sup ag as 


Fat= Fat=6°5°/, Fat=0 
a —_—— ———_—_— — 
Food wt N Wt N wt N wt N 

g- g g. g. g- g. g. g. 
Gelatin _- a 66-672 12-000 27-142 4-885 41-892 7-540 
Corn starch 280 0-118 350 0-137 340 0-143 500 0-135 
Dextrin 50 0-033 16 0-010 — — — _ 
Butterand margarine 105 0-074 20 0-016 20 0-002 Cod oil3 — 
Honey 55 0-013 — -- ao — _ —— 
Sucrose 25 —_ — — 70 — 30 — 
Lactose 65 0-008 180 0-023 250 0-032 300 0-039 
Lemon juice 30 ce. 0-020 25 0-017 25 0-017 20 0-013 
Vermouth 25ce. 0-001 50 0-002 — — — _— 
Tea 1200 ec. 0-072 350 0-028 250 0-020 600 0-043 
Agar-agar 15 0-036 13 0-031 10 0-024 10 0-024 
Salt mixture 5 ae 108 -- 108 — 108 inne 

—~ 0-375 — 12-264 — 5-123 — 7-794 
Total fluid — — 2200-2500 — 2000 — 2200 —_ 


1 During the first five days of the experiment the fuel value of the diet was equal to about 
52 cals. per kilo. 

2 Weight calculated as gelatin. 

8 Includes 4 g. sodium chloride. 

No purification was attempted for experiment I. For II and III the gelatin, after purifi- 
cation by the method described in the text, was dissolved in hot water. Weighed amounts of 
this solution were transferred to bottles and sterilised in the autoclave. The figures for the 
nitrogen intake are based on a number of analyses of samples from different bottles. 


Gelatin I (28th Jan. to 6th Feb. 1921). 


During the three days previous to the experiment a mixed diet containing 
about 12 g. N was taken. . 
The experimental diet is given in Table IV. It included: 


N in the form of gelatin ax 12-00 g. 
N in accessories! (excluding tea) 0-23 
Total N ... Sak & ty! 12-23 


1 The nitrogen in the constituents of the basal diet. 


R. ROBISON 


122 


13. 9% 9% v2 Co 2 BW OF B BB UW oOo Gh if a zw ¢ sequiajdag 
: T T T T T T T iat T T T T T T es 
qydiom Kpog 6s 
9 ia oo 
os 
ig 
soply 
qa Apog 
s 
? 
1 9 
/ 
¢ L 
- aU Jo WYFIOM ooo! 
\ 8 
/ 
6 
q OL 00st 
OULAN UT NT [RIOT iu 
<4) 
(2) 00% 
Ni 
o0zt- NX (3) 
oUHf) J° 7M 
01% |- | 
aqnurut tad *9'9 Pat ° 
0% - peumnsuos uaSixg ot ee ut oor 
| 
ner y Joost 
over i 00st 
I 
09%} } oot 
SAY PZ Jad |sano[yy ; 
08 /- wistjoqe}ay | rst oost 
‘ura dod ‘o'a ; ob66 es 4, ‘a MY FG 
pownsuoogzz aDEeE \ | “diay Apog | l 1 i ! l i 1 rege | i l l i aad step 
uakkxQ eS OS OR. ee eee ee a ee ee ee reac et ae 


WIG [eUON —_poltad sayy (Ir =I UNH —- gat] euION’ poured arog 


GELATIN AS A FOODSTUFF 123 


A few nitrogen-free biscuits made from starch, dextrin and agar were 
chewed at each meal to increase the flow of saliva. The rest of the starch ete. 
was taken in the raw state as already described. 

The results of the experiments are shown in Table V. It will be seen that 
the nitrogen output in the urine remained almost stationary for three days, 
then. rose and a series of oscillations set in. 


Table V. 


y Weight Sp. gr. Creatin- TotalN Nin Total N 
Date exp. weight ofurine ofurine ine N _ inurinet faeces (U+F) Balance 


g- g- g. g- g. g. 
Jan, 28 1 58-75 1447 1-0175 0-44 12:37 (1-38)? (18-75) (-—1-50) 
“29 2 — 1790 1-014 0-53 12-42 1:38 13-80 — 1-57 
30 3 — 714 1-023 0-46 12-38 1-38 13-76 — 1-53 
31 4 — 1236 1-015 0:48 14:05 1:38 15-43 — 3-20 
Feb. 1 5 — 1550 1-013 0-52 12-84 1:38 14-22 — 1-99 
2 6 = 1496 1-015 0-53 14-42. 1-38 15-80 — 3-57 
3 7 57-80 1200 1-016 0-51 _ 13-13 1-38 14-51 — 2-28 
4 8 — 1300 1-015 0-53 14-20 1-38 15-58 — 3:35 
5 9 — 1601 1-014 0-50 12-27 1-38 13-65 — 1-42 
6 10 57-55 1914 1-014 0-54 14-44 1:38 15-82 — 3-59 
Average for whole period Aas eae Sea 0-50 13-25 1-38 14-63 — 2-40 
Average for the last six days OF sey 0-52 13-55 1-38 14-93 —2-70 


1 The “caffeine N” =0-03 g. (from tea) has been subtracted from the total urinary nitrogen. 
2 The faeces were only collected for the last nine days. 


Gelatin II (15th-27th April 1921). 


Durie the two days previous to this experiment the diet consisted of 
eggs, milk, bread, potatoes, butter and apples, and contained 11-2 g. basset ay 
The experimental period was divided into two parts. 

During the first three days the basal diet was supplemented by 250-2. of 
egg-white while during the last ten days this was replaced by gelatin. The 
two diets contained: 


Three days Ten days 
15th-17th Rpril 18th-27th "April 

Nin pe of see ara 5-0 

gelatin — 4-88 
Nin accessories sk arama tea) “21 +22 
Total N cae aes 5-21 5-10 
Fuel value... ex .-. 2580 cals=44 cals per kilo 2757 cals=47 cals per kilo 

(6-5 %, supplied by fat) 


The results are shown in Table VI. 

A pronounced negative balance occurred on the egg-white diet and the 
nitrogen output during the first few days of the gelatin period is slightly 
above the average of the last six days. 


Gelatin IIT (8th-17th Sept. 1921). 


During the preceding four days a mixed diet containing about 12 g. N 
and equal to about 2400 calories was taken. On the 7th Sept. 100 g. lactose 
and 100 g. starch were consumed in addition to the above, making the total 
calories 3130. 


124 R. ROBISON 


Table VI. 
Day of Body Weight Sp. gr. Creatin- Total N Nin Total N 
Date exp. weight of urine of urine ine N inurine! faeces (U+F) Balanpe 
; ; k. g. g. g. g. g. g. 
First period : 

Aprill5 #1 59-3 1351 1-017 — 7-30 = = — 

16 EF2 — 1155 1-019 — 5-91 — oe mae 

17.. £3 — 962 1019 © — 6-09 — — —_ 

Second period: 

April 18 1 58-85 1517 1-013 0-52 7-42 1-28 8:70 — 3-60 
19 2 — 1779 1-012 0-56 6:32 1-28 7-60 — 2-50 
20 3 _ 904 1-018 0:55 6-85 1:28 8:13 — 3-03 
21 4 — 1125 1-015 0:53 6-73 1:28 8-01 —2-91 
22 5 — 833 1-019 0°55 5-68 1-28 6-96 — 1-86 
23 6 — 1148 1-017 0-54 7-12 1:28 8-40 — 3:30 
24 7 — 1523 1-013 0-54 6-18 1-28 7-46 — 2:36 
25 8 58-35 817 1-021 0-53 6-12 1-28 7:40 — 2-30 
26 9 — 1209 1-015 0-52 6:77 1-28 8-05 — 2-95 
27 10 1020 1-017 0-53 6:56 1:28 7-84 —2-74 

28 _— 58-15 — — — — aes — 
Average of whole period _... ois -. 0°54 6-57 1-28 7:85 — 2-75 
Average of last six days’... Sn ~ 053 6-41 1-28 7-69 — 2-59 


1 The “caffeine N” =-02 g. (from tea) Seu been subtracted from the total urinary nitrogen. 
The experimental diet is given in Table IV. It included: 


N in the form of gelatin... ir, ee 7:54 g. 
N in accessories sees ay Sos 5% 21 
Total N aie of ne 15 


From Sept. 14th the fuel vali was duagduned by 100 calories taken in the 
form of starch biscuits and honey. The increased nitrogen is negligible. 

The daily exercise consisted of a walk of from five to seven miles. 

The results are shown in Table VII. 

Here as in Gelatin I the nitrogen output during the first four days is below 
the average for the whole period. 


Table VII. 
Day of Bod Weight Sp. gr. Creatin- TotalN Nin Total N 
Date exp. weight of urine of urine ine N inurine’ faeces (U+F) Balance 
k. g. g. g. g. g. g. 
Sept. 8 1 59-8 850 1-0225 0:59 8-77 1-54 10-31 — 2°56 ° 
7) 2 60-1 1038 1-0175 0:59 8-12 1-54 9-66 —1-91 
10 3 60-5 1764 1-011 0-60 711 1-54. 8-65 — 0-90 
ll 4 60:25 1680 1-013 0:58 8-43 1-54 9-97 — 2-22 
12 5 60°5 1735 1-013 0:60 9-49 1-54 11-03 —3-28 
13 6 60°35 1320 10165 0-59 8-87 1-54 10-41 — 2-66 
14 7 60-0 1347 1-0155 0-57 10:37 1-54 11-91 — 4:16 
15 8 59-93 1439 1-015 0-59 9-68 1-54 11-22 - 3:47 
16 9 59-65 859 1-0215 0-59 8-20 1-54 9°74 — 1-99 
17 10 59-78 1555 1-0145 0:60 9°84 1-54 11-38 — 3-63 
Average for the whole period 0-59 8:89 1-54 10:43 — 2:68 
Average for the last six days 0:59 9-41 1-54 10:95 ~ 3:20 


1 The “caffeine nitrogen” =-04 g. thecnn tea) has been subtracted from the total urinary 
nitrogen. 


BASAL METABOLISM. 


My basal metabolism was determined each day while on the experimental 
diet and during the periods immediately before and after. 


GELATIN AS A FOODSTUFF 125 


The method adopted was that of the Douglas bag, the expired air being 
analysed in Haldane’s gas analysis apparatus. From the 4th to the 7th of 
September the estimation was made at 9 a.m., fasting, after a walk of one 
mile followed by a resting period of at least 30 minutes. All the remaining 
determinations were made between 7.30 a.m. and 8 a.m. on waking. The 
results are shown in Table VIII. 

A slight decrease in the basal metabolism occurred at the commencement 
of the experimental diet but the normal level was regained by about the fourth 
day. This decrease coincides with the lower nitrogen output, as may be seen 
in Fig. 1 A and B, but it is not possible to say whether the two are in any 
way connected. No corresponding decrease occurred in the creatinire output 
which was constant throughout the period. 

Ia calculating the basal metabolism in calories per 24 hours the protein 
oxidation has been ignored but the error thus introduced is less than + 1 %. 
For those days in which the R.Q. is greater than 1 a correction has been made 
for the nett heat production due to synthesis of fat from carbohydrate by 
adding 0-3 of the calories equivalent to the excess ot the CO, output over 
the oxygen consumed. I am aware that this is an arbitrary estimate, but the 
possible error involved is only slight. 


Table VIII. 
Oxygen con- 
Day of Diet during previous 24 hours sumed per Calories per 
Date exp. (calories per kilo) minute (cc) R.Q. 24 hours 
Sept. 4 Fil Normal (40) 223-0 0-831 1575 
5 F 2 ” ” 222°5 0-830 1571 
6 F3 ss BR 222-4 0-803 1561 
7 Fa : +9 222-3 0-866 1580 
8 G1 Normal + 200 g. starch (54) 221:2 0-904 1582 
lactose mixture 

9 G2 Carbohydrate + gelatin (54) 210-1 1-012 1537 
10 G 3 ” ” 204:8 1-043 1510 
ll Ga4 Be : 199-0 1-093 1492 
12 G5 = as 215:3 1-065 1600 
13 G6 Pe ” 221-6 1-051 1633 
14 G7 » ” 222°8 0-967 1614 
15 G8 3 9 218-0 1-040 1604 
16 G9 99 ” 219-8 0-967 1592 
17 G10 ¥s ax 212-9 0-968 1542 
18 Al és 218-7 0-979 1587 
19 A2 Normal diet 239-8 0-823 1691 
20 A3 i: 269-3! 0-805 1893 
24 AT Bs 2243 0-801 1574 
25 A8 235-5 0-809 1660 
26 AQ e, 226-3 0-823 1596 
27 Al0 Py: 220-9 0-830 1559 


1 The body temperature was 99-4° when this determination was made.. Several attacks of 
vomiting had occurred during the previous night. The condition became worse and necessitated. 
some days’ rest. 

The normal basal metabolism calculated from Harris and Benedict’s 
formula for a man of age 37, weight 59-8 kilo, height 173-5 cm. would be 
66-4730 + 13-7516 x 59-8 + 5-0033 x 173-5 — 6-755 x 37 = 1506 calories. 


126 R. ROBISON 


GENERAL CONSIDERATIONS. 


My usual mode of life was followed throughout these experiments, eight 
or nine hours of each day being occupied with laboratory work. The only 
form of exercise was a walk of from two to seven miles. No great difficulty 
was found in consuming the diet although a feeling of nausea was frequently 
experienced. This was greatly intensified by the excessive quantity of food 
in the third experiment; the tongue became furred and more or less headache 
was common. Slight diarrhoea occurred in all three periods, faeces of fluid 
consistency being passed two or three times a day. The diet was however 
very well assimilated scarcely any starch being found in the faeces. Traces 
of reducing sugar were regularly present in the urine during the third en 
ment but only occasionally during the first two. 

It was intended that a period on nitrogen-free diet should follow im- 
mediately on the third gelatin diet in order to determine my minimum re- 
quirements once more. This was not possible, and indeed much difficulty 
was experienced in completing the ten days proposed for the above experiment. 

In the two experiments (GI, G IIT) which followed a normal mixed diet 
the average nitrogen output during the first four days was lower than that 
for the remainder of the period while the reverse of this was observed when 
the previous diet had been insufficient to satisfy the protein requirements 
(G II) and a considerable negative balance had occurred. These differences 
are probably due to the influence of the protein of the preceding diet, a 
diminishing store of which, perhaps in the form of amino-acids, remains in 
the body for some days. The first four days have therefore been excluded 
in considering the results. Another disturbing factor is the variation in the 
urine nitrogen from day to day. This frequently amounted to more than 
20% of the average output and showed no relationship whatever to the 
volume of the urine. Consequently it cannot be explained by diuresis. 


SUMMARY AND Discussion or RESULTS. 


The results obtained in the three experiments are summarised in Table LX, 
the average figures being given for the last six days of each ten day period. 
The average amounts of the nitrogen intake and output on the last three 
days of the earlier experiment on “nitrogen-free” diet are also included. 


Table [X 
maw A rus yake Nitrogen intake N itrogen ot output 
w of ; 
Date (kilos) cals per kilo Gelatin ZB tem cris Urine Faeces Total Creatinine Balance 
4 g a & g. g. j 
3. xii. 20-5. xii. 20 58-0 52-44 _— *B0 2:06 1:18 3°19 0:49 — 2-89 
1. ii. 21-6. ii. 21 57-8 Ad 12-00 23 13°55 138 14:93 0-52 — 2-70 


21. iv, 21-27. iv. 21 584 47 4°88 22 641 1:28 7:69 0-53 ~-2-59 
11. ix. 21-17. ix. 21 60:2 54 7-54 21 941 154 10-95 0:59 —3-20 


In attempting to calculate the amount of body protein spared by the 
gelatin from these results we are met by two difficulties, namely what is to 
be done with that part of the intake due to the nitrogen of the accessories, 
and with that part of the nitrogen output due to the faeces, 


GELATIN AS A FOODSTUFF 7 127 


Nitrogen of the accessories. The nitrogen from the tea does not appear in 
‘ the table, having been subtracted both from intake and output. The assump- 
tion that this is all “caffeine N”’ is not strictly correct but as the total amount 
is very small, usually under 0-05 g. any error involved in this mode of treat- 
ment must be negligible. The greater part of the accessory nitrogen comes 
from the corn starch. In McCollum’s account of his experiments on pigs, no 
mention is made of any nitrogen arising from this source although large 
quantities (1700 g.?) of corn starch were given. If this starch contained as 
much nitrogen as the samples used by me the nitrogen intake from this source 
may easily have been -0-5 g. or more. We do not know in what form this 
nitrogen is present nor its value in the human body and cannot therefore 
estimate its effect on the nitrogen output. If the amount and nature of such 
accessories are the same during the determination of the nitrogen minimum 
and experiments with gelatin, the following argument might be applied. If 
the value of this nitrogen in the accessories is zero then the real nitrogen 
requirements will be less than the observed output by the full amount of such 
nitrogen intake since the latter must be excreted in addition to the nitrogen 
resulting from the protein metabolism of the body. If the value of this 
nitrogen for the replacement of body nitrogen is 100 % then it will spare an 
equal amount of the latter and the observed output on the so-called “ nitrogen- 
free” diet will represent the actual minimum requirements. But in this case 
an equal amount of gelatin will also be spared when gelatin is taken, and the 
apparent sparing effect of the gelatin will thus be increased by the same 
amount. 

In either case the real saving of body nitrogen due to the gelatin will be 
less than the apparent saving, ¢.e. the difference between the negative balance 
on the gelatin and the minimum nitrogen output on the “nitrogen-free” diet, 
by an amount equal to the nitrogen of the accessories. This will also hold for 
all values of the latter between 0 and 100 °/,. Unfortunately the proviso that 
the accessories should be the same on both diets does not strictly hold in the 
above experiments but if the butter nitrogen be subtracted from the total 
intake on the “ nitrogen-free ’’ diet the remainder is practically the same as the 
accessory nitrogen on the gelatin diets. The butter nitrogen would probably have 
a high value and I have therefore assumed that it does not appreciably increase 
the nitrogen output. The rest of the accessory nitrogen (-22 g. average) has 
been deducted in calculating the amount of body nitrogen saved by the gelatin. 
This is not strictly accurate but is probably the best that can be done with 
the figures. The above argument however ignores the possibility of comple- 
mentary action of the accessories and the gelatin. It has been already pointed 
out that such action occurs when gelatin is fed with certain cereal proteins. 
McCollum, Simmonds and Pitz [1917] have shown that a mixture of oat 
protein and gelatin has a higher value than either alone or than oat protein 
plus caseinogen. It is impossible to say whether the results in my experiments 
were affected by such complementary action but if this did occur the real 
value of the gelatin alone is still less than the calculations appear to show. ‘| 


128 ~ R. ROBISON 


‘Faecal nitrogen. The problem of how to treat the nitrogen excreted in the 
faeces is even more difficult. In McCollum’s experiments the urinary nitrogen 
is alone considered, that in the faeces being estimated merely as a check on 
the complete absorption of the food protein. He considers that the nitrogen 
of the urine represents the essential tissue metabolism, while that of the faeces 
represents losses which may be termed accidental in character. 

It has been shown by Rubner [1919] that the increased amount of nitrogen 
in the faeces of men, when fed on various diets, above that found on a carbo- 
hydrate diet cannot be taken as entirely due to undigested food protein, but 
that a considerable proportion of the increase comes from the body. In my 
experiments somewhat large variations were observed in the nitrogen of the 
faeces and these may have been due to the slight diarrhoea which occurred. 
Probably most of the nitrogen came from the body but the possibility that a 
small amount of gelatin escaped absorption must not be overlooked. We 
know very little about the relationship between loss of body nitrogen through 
the intestines and that excreted in the urine, but there seems to be no evidence 
that an increase in the former is accompanied by a decrease in the latter. 
The reverse of this is perhaps more probable. 

This question remains at present the limiting factor for the accuracy of 
such experiments. I have attempted to define the limits between which the 
true conclusion from my results is to be found, by calculating the amount of 
body nitrogen saved by the gelatin in two ways. In the first (A) I have 
assumed that my minimum nitrogen requirements are represented by the 
sum of the nitrogen in the urine and that in the faeces on the nitrogen-free 
diet, and that the difference between the latter amount and the corresponding 
excretion on the gelatin diets represents unabsorbed gelatin nitrogen. 


Table X. 
A B 
+3 pss oe ye YisPa 
N Intake ) Body Nitrogen saved Body Nitrogen saved 
pa Sie Ne N ee N oo 
Gelatin Accessories Balance minimum em, i! minimum % of 
Experiment g. g. g. g. g. minimum g. g. minimum 
R.R. II 4:88 0-22 — 2-59 3-19 0:38 11-9 3:34 0-53 15-9 
R.R. U1 7-54 0-21 -3-20° 3-19 0 0 3-60 0-19 5:3 
R.R. I 12-00 0:23 — 2-70 3-19 0:26 8-1 3-44 0-51 14-7 
McCollum’s pig) 
average of last ; 2-62 ? — 2°35 3-68 1-33 36:1 3:66 1-31 36:3 
six day f 


In the second (B) I have assumed that the gelatin is completely absorbed 
and that the minimum requirements for such periods are represented by the out- 
put in the urine on “nitrogen-free” diet plus the nitrogen in the faeces on the 
gelatin diet under consideration. The truth probably lies somewhere between 
these two extremes. The results of these calculations are given in Table X. 
McCollum’s figures are alsoincluded for comparison. The accessory nitrogen has 
been in each case deducted from the apparent amount of body nitrogen saved. 

The maximum saving in terms of the nitrogen minimum is thus 11-9 % 
if the first method of calculation is employed and 15-9 % if the second is used. 


GELATIN AS A FOODSTUFF 129 


The fact that this was obtained with the lowest amount of gelatin would 
appear to prove that the effect is not due to any impurity, in which case there 
should be a proportionality between the amount saved and the gelatin intake. 
In this connection however the possibility that the increased protein in the 
diet entails an increased loss of body nitrogen must not be overlooked, although 
the creatinine excretion does not lend any support to such an hypothesis. 

The results of the first and second experiments agree well between them- 
selves, but in the third a much lower value was apparently indicated. It will 
be noticed however that the creatinine excretion in the experiment was higher 
than the normal and the body weight was also higher. This nitrogen minimum 
may therefore have been higher than the amount shown (perhaps owing to 
the excessive amount of food taken) in which case the calculated result would 
be too low. 

All these values are much lower than those found by McCollum and this 
discrepancy cannot be explained except by assuming a difference in the 
metabolism of man and pig. If the values are calculated in terms of the 
urinary output alone they are all proportionately increased—the value for 
my experiment @ IT then becoming 25-7 °% of the minimum, but the difference 
between the man and the pig still persists. 

The creatinine output has been shown to bear some close relationship with 
the minimum nitrogen output. I therefore attempted to.compare my results 
with McCollum’s in terms of the amounts of creatinine excreted in the several 
experiments. I ignored the faeces and calculated the negative balance by 
subtracting the nitrogen intake from the output in the urine. The ratio of 
this balance to the average amount of creatinine nitrogen excreted during 
the same period is shown in the last column of Table XI. The last six days 
of each period have been alone considered. 


Table XI. 
Nitrogen Nitrogen in Nitrogen Nitrogen balance 
Experiment intake urine balance Creatinine N Creatinine nitrogen 
g- g: g- & 
R.R. IL 5-10 6-41 -131 0-5 2-47 
R.R. IIT 7:75 9-41 — 1-66 0-59 2-81 
R.R. I 12-23 13-55 — 1-32 0-52 2-54 
McCollum’s pig 2-62 3:89 — 1-27 0-48 2-65 


There is obviously no discrepancy between our results when they are con- 
sidered in this way, though what this agreement implies is not easy to state. 
It has been shown in many papers by Grafe (1912-1914), Abderhalden [1915], 
Underhill and Goldschmidt [1913] and others, that many nitrogen compounds 
other than amino-acids, namely organic ammonium salts, urea etc., have the 
capacity to spare a certain proportion of the loss of body nitrogen occurring 
on a carbohydrate diet. The results obtained by these workers do not agree 
in all points but the amount of nitrogen thus spared appears to be of the 
same order as that spared with gelatin in my experiments. It may well be 
that the action of the gelatin is of the same nature as that of these simpler 


Bioch. xv: 9 


130 R. ROBISON 


compounds and consists essentially in the reduction of the waste of amino- 
acids derived from body protein through deaminisation and subsequent oxida- 
tion in the body. The amino-acids of the gelatin and the ammonia and urea 
produced from them can play a part in the reversible reactions that are 
constantly proceeding in the body and thus influence the resulting equilibrium. 
Amino-acid 7? Non-nitrogenous compound + NH; 7 urea. 
(Ketonic or hydroxy acid) 

If this is true the loss of body nitrogen when both carbohydrate and gelatin 
are fed may represent a “N-minimum” that corresponds more closely with 
the specific nitrogen requirements of the body than does the output on 
carbohydrate diet alone. This may perhaps be the explanation of the close 
agreement between the ratios of such loss to the creatinine nitrogen shown by 
my experiments and those of McCollum. 


In conclusion I would express my very sincere thanks to Prof. C. J. Martin 
for his constant encouragement and advice throughout this investigation. 


REFERENCES. 


Abderhalden (1913, 1). Zeitsch. physiol. Chem. 85, 91. 
(1913, 2). J. Biol. Chem. 15, 357. 

—— (1915). Zeitsch. physiol. Chem. 96, 1. 
Benedict and Osterberg (1914). J. Biol. Chem. 18, 195. 
Bischoff (1853). Der Harnstoff als Maass des Stoffwechsels, 70. 
Boruttau (1919). Biochem. Zeitsch. 94, 194. 
Boussingault (1846). Ann. chim. phys. 18, 444. 
Cathcart (1907). Biochem. Zeitsch. 6, 109. 
Dakin (1920). J. Biol. Chem. 44, 499. 
Donders (1853). Die Nahrun, rungstofte, 72. 

Escher (1876). Vierteljahrsschr. nat. Ges. Zurich, 36. _ 
Folin (1905). Amer. J. Physiol. 13, 117. 
Folin and Denis (1912). J. Biol. Chem. 12, 245. 
Frerichs (1845). Handwérterbuch Physiol. 3 (i), 683. 
Graham and Poulton (1912). Quart. J. Med. 6, 82. 
Kauffmann (1905). Pfliger’s Archiv. 109, 440. 
Kirchmann (1900). Zeitsch. Biol. 40, 54. 
af Klercker (1907). Biochem. Zeitsch. 3, 45. 
Krummacher (1901). Zeitsch. Biol. 42, 242. 
(1903). Zeitsch. Biol. 45, 310. 
Lehmann (1885). Sitzungsber Miinchener morph-phys-gesell. 
McCollum (1911). Amer. J. Physiol. 29, 210. 
McCollum and Hoagland (1913). J. Biol. Chem. 16, 299. 
McCollum, Simmonds and Pitz (1917). J. Biol. Chem. 29, 341. 
Mendel and Rose (1911). J. Biol. Chem. 10, 475. 
Munk (1894). Prfliger’s Archiv. 58, 309. 
Murlin (1907, 1). Amer. J. Physiol. 19, 285. 

—— (1907, 2). Amer. J. Physiol. 20, 234. 
Oerum (1879). Nordiskt medicinskt Arkiv. 11. 
Osborne and Mendel (1912). J. Biol. Chem. 18, 233. 
Pollitzer (1885). P/liger’s Archiv. 87, 301. 
Robison (1922). Biochem. J. 16, 131. 
Rona and Miiller (1 Bayete Zeitsch, physiol. Chem. 50, 263. 
Rubner (1919). Arch. 
Sherman (1920) J. Biok Chem. ‘44, 173. 
Sherman, Wheeler and Yates (1918). J. Biol. Chem, 34, 383. 
Thomas (1909). Arch. Physiol. 219. 
Towles and Voegtlin (1912). J. Biol. Chem. 10, 479. 
Underhill and Goldschmidt (1913). J. Biol. Chem. 15, 341. 
Voit, C, (1872). Zeitsch. Biol. 8, 297. 
Zeller (1914). Arch, Physiol, 213. 


XVII. DISTRIBUTION OF THE NITROGENOUS 
CONSTITUENTS OF THE URINE ON LOW 
NITROGEN DIETS. 


- By ROBERT ROBISON. 
From the Lister Institute. 


(Received January 17th, 1922.) 


Durtne the course of experiments carried out in conjunction with Prof. C. J. 

Martin on ourselves, with the object of determining our minimum nitrogen 

requirements and the biological value of certain proteins, the opportunity was 
taken of investigating the distribution of the nitrogenous constituents in the 

urine when the total nitrogen output had reached a very low level. 

In the particular experiments referred to in this paper, a diet consisting 
of carbohydrate and fat (corn starch, sucrose, lactose, honey, butter) together 
with a little lemon juice, inorganic salts and agar-agar was taken during 
seven days. The fuel value of this diet was equal to 44-52 calories per kilo 
body weight and the nitrogen content was about 0-3 g. A little tea and coffee 
was also taken but the nitrogen in these beverages was assumed to be “ caffeine 
nitrogen” and to be excreted as such in the urine during the same 24 hours. 
It was therefore subtracted from the total nitrogen before calculating the 
- percentage amounts of the other constituents. The above assumption is of 
course not strictly accurate, but in the one experiment (R.R.) the total amount 
of such nitrogen is very small, less than 0-1 g., so that any error thus intro- 
duced may be considered negligible. 

The period on this low nitrogen diet was immediately followed by another 
of five days during which the same basal diet with the addition of a little 
milk was taken, the nitrogen intake being thus raised to about 3 g. 

The total nitrogen in the urine and in all constituents of the diet was 
estimated by Kjeldahl’s method, urea by van Slyke’s urease method, ammonia 
by Folin’s aeration method, amino-acids by formal titration after removal of 
the ammonia, creatinine by the method of Folin, and uric acid by that of 
Hopkins as modified by Folin and Schaffer. 

The results are set out in Table I. In Table II the distribution of nitrogen 
on the days for which the total nitrogen output reached its lowest values has 
been compared with corresponding figures obtained by other investigators. 

It will be seen that these observations are in complete agreement with 
Folin’s [1905, 2] generalisations respecting the variation in the distribution of 
the urinary nitrogen. The creatinine nitrogen is practically constant and is 

: 9—2 


Ce ee ee a ee ne eee es 


ce NE 99 


Pel «88 , 88S (OOS CLS HEL OSS Ue lo LO RE aes REE (1 ie er Sa a a a “ 
| 60° 
FI eg TSS LE 8-7¢ 9-LT GLE 63° L0- og: F0- cE: PL: 266° { "re ¢ ele “WU “ 
ost 8 30 €3 909 66 LO 6% Ts eo OT OEE | ee te ae ees waneacis 
; ae (eter) pe 
92 — 619 SoSt 9F Gr B10 ge — se MOL GS 16 6 F9 ‘DD pave ureyes 
me oe em 92) 0D; AL. 8, an cnc re Sl sss wl Ww 6L9 — — [prerlsz 
ae. OF | CI r19 8h 86S ae 1 Oe = Le: IZ 9 Tks > aay, ; 
or Of OF = LOL O0T 209 0 rt — 8° Lt 88 Se ee eee 
e+) GZ CM O-Eh, eIt Lt |= Le 60° 09: = cr: SS OME | na Leg “H'aH [tr ‘coer) weg 
peur prow out poe vmUowUy vluoUUIy verQ ‘ON ‘SN ‘ON ‘ON ‘ON ‘ON ‘BN ‘ON ‘dxe yystea qoefqne TeAIssqQ 
~1yopunl ony) -uveL) ae +8019, peurut proe our plow euowMmy ereaQ eyo, eyequy jo Aeq Spog 
—  -1eJepuy «OO, -UNJearg -ouTWIY 
N 16904 JO sesequsoleg ; 
IT ®(4*.L 
‘9 ‘00q UO Weg IOP ALN ¢ - c'N ouLoyeo,, Se UayL}—aYoo pue vay WOIS NC 
C101 sI- 
Bf OR eRe OO RE OO She nee DIES ORS HOS RS OE Ree TE Atay ht gee iene ee 
: FIO er- 
Sar te SE ee OO POP AE 08 RS te ee OE OR OEY eee pee 
eit et OWE. et. Veh ene OF oe 2 Ge ge 50 OG. OS eee SCRE oe Lage org 
I 
wae OE WR LO OBR BD. RRS OE AOE ee TE OO: — ORR pee eee ge 
Sere ORS: OT RO ORT ROO Sm ee Om a OE Oe OR go eee gt 
Sat Seat ede COR IR RS BRR SU ee pe ee ieee A gee aS 
eR Obs OG BR CIR A. FBT BOR ee OE 0 Oe eA ge ee og Foe 
na ee eat Pat iaeeeg 0) fea 17 eal, ouaehe ae ees OS See ley 2 ane Sad Ro RE re 
“SO[EY L+LG—UIG *2ACT SO[LY 9-8S—YISS “AON INSIOM “AA :Wolqns 
PPG ors | tet) aie Ces, Me, Bee 80) (8 ae hoet ate, eet tere ee ee 
cet OR OST. BR. IR 08 SRO rR Oe gee A ae > 2 SR ace SO OTR Oe a) eas et ete 
I . 
gor 6 8GC«<C SCC BO 06 TS 6& 90 8h 80 SSL Ie eT eee cee gs 
ete Oe oe OC ROTC BIRO in: FS ee a a) Se OT oe gee gt 
ere ae eet? a ane ee ee, eee ett een Cae ace us eS Camas ee Sa 
gre) ceg a MeL Cpe ees Vor eee Se bec in ge Se a Oe ORE ee Pe 
peurm poe out poe  eruoulue eruoUTy Berl) ‘ON ‘ON ‘BN ‘BN ‘SN ‘SN ‘SN ‘BN eunm ‘BN ‘dxa eq 
srjopun, 9A -UnvaID -ouRY + "OLA pour = poe oT PR MMOMTUTY arf] oUFEyRD EIT, joud"ds eyruy yoke ORET 
‘ J Y s1ajepuQ oNQ -uNjeery -ourmy S89] "20 QUE ae 
(Nt outages sse[) N e909 F jo osequeored uy N [eqor yo oumyo, 


SOTY G-09—TIG “90 SSO[TY 6-TO—TIRS “AON FUSEAY “WWE sqoeLque 


T 18D 


URINARY NITROGEN ON LOW NITROGEN DIETS 133 


equal to 7-3 mg. (C.J.M.) and 8-2 mg. (R.R.) per kilo body weight. These 
amounts like those of the other constituents are very similar to those found 
by Folin and others. The values for the total nitrogen are amongst the lowest 
on record and correspond with still further reductions in the percentage of 
urea nitrogen, this falling in one instance to 37 % of the total. The last figure 
was however accompanied by a somewhat high percentage of ammonia, but 
the sum of the urea and ammonia nitrogen was only 54-6 % of the whole amount. 
It would have been interesting to discover whether this sum (urea -++ ammonia) 
could have been further reduced by the ingestion of alkalies or whether any 
decrease in the ammonia nitrogen would have been accompanied by an 
increase in the urea. 

The question as to whether any part of the urea nitrogen represents what 
Folin has termed the endogenous metabolism remains open. 


I wish to express my indebtedness to Prof. C. J. Martin for his interest 
and help in this work. 


REFERENCES. 


Folin (1905, 1). Amer. J. Physiol. 18, 66. 

(1905, 2). Amer. J. Physiol. 18, 117. 
Graham and Poulton (1912). Quart. J. Med. 6, 82. 
Zeller (1914). Arch. Physiol. 213. 


XVIII. THE ESTIMATION OF TOTAL 
SULPHUR IN URINE. 


By ROBERT ROBISON. 


From the Laster Institute. 
(Received January 27th, 1922.) 


THE methods for the estimation of total sulphur in urine are less satisfactory 
than those available for determining many of the other urinary constituents. 
A high degree of accuracy is specially important when we are concerned with 
the amount of unoxidised or “neutral” sulphur in the urine. This can only 
be determined by subtracting the amount of sulphates from that of the total 
sulphur and any error in the latter value falls on the relatively small difference. 

Benedict's [1909] method in which the oxidation is effected with a mixture 
of copper nitrate and potassium chlorate has been criticised on the ground 
that considerable losSes frequently occur through spattering during the initial 
stages of the reaction. Denis [1910] has stated that out of forty analyses 
mechanical loss resulted in every case. She has proposed to substitute a 
solution of copper nitrate, ammonium nitrate and sodium chloride in place 
of Benedict’s oxidising reagent, thereby reducing the vigour of the reaction. 

In my hands Denis’ modification has not proved so reliable as the original 
method. Decrepitation is certainly less troublesome but the results are fre- 
quently too high, the errors amounting at times to more than 10mg. The 
following series taken from nearly a hundred analyses are fairly representative: 

25 ec. urine (1) gave 0-0711, 0-0719, 0-0783, 0-0826 g. BaSQ,, 
‘s » (2) ,, 0-0864, 0-0901, 0-0865, 0-0899, 0-0877 g. BaSQ,. 

The residue, even after prolonged ignition always contained appreciable 
amounts of nitrate, this being probably due to sodium nitrate formed from 
the sodium chloride and copper nitrate since the latter is readily decomposed 
at temperatures lower than those used. Whether the presence of this nitrate 
would alone account for the high results I am not able to say. Considerable 
errors in this direction were observed by Kolthoff and Vogelensang [1919] 
using higher concentrations of both sulphate and nitrate. 

Consistent results were usually obtained by Benedict’s original method, 
using an electric hot plate as suggested by Givens [1917] although in spite 
of this precaution vigorous spattering sometimes occurred, Only slight traces 
of nitrate were found in the residues after ignition and it is probable that 
these arose from the small amounts of alkali salts in the reagent and in the 
urine itself. The necessity of igniting the residue at a red heat was however 


ESTIMATION OF TOTAL SULPHUR IN URINE 135 


a frequent cause of spoiled analyses through the cracking of the porcelain 
basins. Probably those used by Benedict were of better quality than are at 
present obtainable. 


Errors due to the use of coal gas for igniting the residue. 


Neither Benedict nor Denis appears to have recognised the possibility of 
errors from the presence of sulphur in the coal gas used for the ignition of the 
residue. Folin [1903] advised spirit burners for his original method and 
according to Liebesny [1920] errors of 8-10 °% in the amount of neutral 
sulphur may arise from the use of gas. My own experience confirms this. 
In determining the blank of the oxidising solution referred to below, 2+5 cc. 
was found to yield 0-7 mg. BaSQ,, using a spirit burner, and 2-4 mg. using 
an argand burner with coal gas. If this error were constant in amount it 
would be cancelled in subtracting the blank but such was not found to be the 
case. Probably the absorption of oxides of sulphur depends on the surface 
area of the residue as well as on the total quantity present in the gas. 

The use of a spirit burner is therefore essential if accurate results are to 
be obtained. 

Since copper nitrate, which forms the basis of Benedict’s reagent, is de- 
composed at comparatively low temperatures it seemed possible that the 
necessity of igniting the residue at a red heat might be avoided if neither 
chlorate nor alkali salts were introduced. Other substances were therefore 
tried as a diluent for the copper nitrate and of these copper chloride was 
found to be the most satisfactory. In the presence of a suitable proportion 
of this salt spattering never occurs, while the nitrate can be completely de- 
composed by heating the residue over a small spirit stove of the simplest 
kind. The temperature of a good argand burner is sufficiently high but gas 
cannot be used. 

The oxidising reagent finally adopted has the following composition: 

Copper nitrate (cryst.) 40g. 
Copper chloride (cryst.) 15g. 
Water to 100 ce. 


2-5 cc. of this solution are added to 10 cc. of the urine in a 4-inch porcelain 
basin and evaporated to dryness on a water-bath or electric hot plate. The 
oxidation can be started on the hot plate, or over a very small spirit flame. 
It proceeds rapidly but smoothly, leaving a coherent residue which frequently 
swells up. The dish is then heated over a broad spirit flame for 20 minutes. 
A spirit stove of the common kind is suitable but a sound tin, half filled with 
methylated spirit answers very well. A better flame is obtained if a number 
of holes are punched about half-way up the tin. The residue is dissolved in 
10 cc. of 2N HCl and diluted with 300 cc. distilled water. The sulphate is 
precipitated in the boiling solution with 10 cc. of a 5% solution of barium 
chloride, dropped in very slowly by means of a dropping tube. The precipitate 
_ is allowed to stand overnight before being filtered. 


136 R. ROBISON 


In the following analyses the precipitates were weighed in platinum 
crucibles. 

For 25 ce. urine, 6-25 ec. of the oxidising solution, 20 cc. 2N HCl, 400 cc. 
H,0 and 15 cc.—20 ce. of 5 % barium chloride solution were used. A blank 
lntarminalhe must be carried out with the reagents. 

Two series, each of eight estimations, of the total sulphur in a normal urine 
yielded the following results. 


BaSO, from 10 ce. urine BaSO, from 25 ce. urine 
1. 0-0451 1. 0-1140 
2. 0-0454 2. 0-1143 
3. 0-0460 3. 0-1135 
4, 0-0454 +t. 0-1141 
5. 0-0456 5. 0-1138 
6. 0-0457 6. 0-1133 
7. 0-0461 hs 0-1141 
8. 0-0454 8. 0-1148 
Average 0-0456 g. Average 0-1140 g. 


equivalent to 0-0456 g. for 10 ce. 

It is doubtful whether any gain in accuracy is achieved by the use of 
25 ec. instead of 10 cc. of urine. 

In all the above estimations the filtrates were examined for nitrates by 
the very sensitive diphenylamine test, slight traces being found in one or two 
cases only. 

The accuracy of the method was checked by estimating the total sulphur 
in two solutions of cystine containing also 2 % of urea and a definite volume 
of N/10 H,SO, neutralised by an equal amount of V/10 NaOH. 

The sample of cystine, for which I am indebted to the kindness of Dr J.C. 
Drummond, was analysed by the Carius method: 

0-1878 g. gave 0:3679 BaSO,; calculated 0-3681 g. BaSQ,. 
BaSO, from-10 ce. solution A BaSO, from 10 ce. solution B 


0-0671 0-0662 
0-0670 0-0659 
0-0675 0-0661 
0-0675 0-0659 
wan 00659 
= —.0-0663 
0-0673 g. 0-0661 g. 
Calculated 0-0678 of which Calculated 0-0668 of which 
0-0212 is from cystine. 0-0202 is from cystine. 


This method has been employed for some time by myself and others in 
connection with metabolism experiments in this laboratory and has proved 
satisfactory. 

I wish to thank Miss M. Tazelaar and Miss M. H. Carr for much valuable 
help in carrying out the numerous analyses, only a very small proportion of 
which have been recorded above. 


REFERENCES. 


Benedict (1909), J. Biol, Chem. 6, 363. 

Denis (1910), J, Biol. Chem. 8, 401. 

Folin (1903). Amer. J. Physiol. 9, 265. 

Givens (1917). J. Biol, Chem. 29, 15. 

Kolthoff and Vogelensang (1919). Pharm. Weekblad, 56, 122. 
Liebeany (1920), Biochem. Zeitsch, 105, 43. 


XIX. THE ACTION OF YEAST-GROWTH 
STIMULANT. 


By OSWALD KENTISH WRIGHT. 


From the Biochemical Department, Lister Institute, and the Bacterrological 
Laboratory, Household and Social Science Department, King’s College 
for Women. 


(Received January 30th, 1922.) 


WituiAMs [1919] and Bachman [1919] have confirmed the observation of 
Wildiers [1901] that certain yeasts are only able to grow at the expense of 
ammonium salts provided that a heavy inoculation is employed or a small 
amount of organic material (termed “bios” by Wildiers) is added to the 
medium, After a study of the ability of various substances to produce this 
effect, they suggest that the essential substance, which we will continue to 
call “bios” for the sake of convenience, is identical with the water-soluble B 
or anti-beri-beri vitamin. Williams [1920] has gone further, and has elabo- 
rated a method for measuring quantitatively the amount of anti-beri-beri 
vitamin present in any substance by observing its effect on the growth of yeast. 

This suggestion has given rise to much discussion [see Eddy, 1921] and at 
present it is generally considered that the case for the identity of the two 
principles is not proven. 

Lemon juice freed from citric acid by the method of Harden and Zilva 
[1918] when added to a mineral nutrient solution in small quantities enables 
a yeast to grow which could not grow in its absence. As animal feeding 
experiments show that the amount of water soluble B vitamin in lemon 
juice is very small relative to that in yeast extract [see Osborne and Mendel, 
1920], it was thought that an investigation of its effect on the growth of 
yeast in mineral nutrient solutions might throw some light on the way in 
_ which the stimulation is brought about. 

Preliminary experiments were performed to ascertain the smallest amount 
of lemon juice capable of producing growth in a nutrient solution made up as 
follows (Williams): 


Saccharose FF 20 g. 
(NH,),SO, dl 3 g. 
KH,PQ, ... a: 2 g. 
3S eee 4 0-25 g. 
MgSO, ... 3 0-25 g. 
Distilled water ... 1000 cc. 


A series of tubes was prepared containing this solution with the addition 
of increasing percentages of lemon juice. The yeast employed was a pure 


138 O. K. WRIGHT 


culture of a baker’s yeast isolated by Dr Harden. This was grown for 24 hours 
in yeast water glucose and the growth was washed three times with sterile 
distilled water by means of the centrifuge. A loopful of a suspension of this 
washed yeast was put into each tube of the series, the amount being judged 
so that a fair proportion of drops made, according to Lindner’s method, with 
a mapping pen on a cover slip contained one, two, or three cells each. The 
coverslip was sealed over a moist chamber with vaseline and, after noting the 
number of cells in each drop, was incubated for 24 hours at 25° and again 
examined for growth. 

By this method it was found that no growth took place in the mineral 
nutrient solution unless 5 % or more of lemon juice was added. It was also 
noticed however that some of the tubes containing smaller amounts of lemon 
juice than 5%, which had been standing in the laboratory, contained an 
amount of yeast appreciably larger than the amount originally inoculated. 
This suggested that growth did not take place as readily in the case of one or 
two cells in a minute drop as in a larger amount of medium. Subsequent 
investigation confirmed this, and the drop method was abandoned in favour 
of estimating the number of cells per cc. by means of a counting chamber. 

In order to avoid adding any appreciable amount of “bios” with the 
seed yeast, it is necessary to inoculate each tube with as small a quantity of 
yeast as possible. Attempts to inoculate tubes by picking up a drop of dis- 
tilled water containing one, two, or three cells from a coverslip by means of 
sterile filter paper proved unsuccessful. The method adopted was to make a 
light emulsion in sterile distilled water of yeast cells prepared and washed as 
described above. A slide was marked with several circles with a grease pencil 
and a drop of the emulsion transferred to the centre of each by means of a 
platinum loop. The drops were dried and stained and the number of cells 
in each counted, and by this means an emulsion was prepared which con- 
tained somewhere between 50 and 500 cells in a loopful. One loopful was 
then used to inoculate each tube. | ; 

A series of ten tubes was prepared with a mineral nutrient solution con- 
taining (NH,),SO, with increasing percentages of lemon juice, A similar series 
was prepared without the (NH,).SO,. 

Each tube contained 7 cc. of the following solution (Williams’ solution 
without the (NH,),S80,, see p. 137): 


Saccharose bn 20 g. 
KH,PO, ... es 2 g. 
ah 5; me 025 g. 
MgSO, ... oss 0:25 g. 
Distilled water ... 1000 ce. 


The tubes containing (NH,),SO, received 1-5 ce. of a solution containing 
2g. of the salt in 100 cc, The saccharose and the (NH,),SO, were purified | 
by several recrystallisations. Repeated attempts to grow the yeast in tubes 
of this solution with (NH,),8O0, inoculated with from 50 to 500 cells in the 


ACTION OF YEAST-GROWTH STIMULANT 139 


manner described above and without the addition of “bios” failed, no growth 
being observed even after a lapse of several weeks. 

The lemon juice was prepared by the method of Harden and Zilva [1918], 
namely as follows: calcium carbonate was stirred into fresh lemon juice in 
excess of the amount required to neutralise all free acid. Absolute alcohol 
to twice the volume of the original lemon juice was added, and the whole 
filtered. The filtrate was evaporated down almost to dryness im vacuo at 35°. 
and the residue made up to the original volume of the lemon juice with dis- 
tilled water. The sample used contained 0-0868 mg. of nitrogen per cc. of 
which 0-0112 mg. was ammoniacal nitrogen. The desired amount of this 
lemon juice was added to each tube of the series which was then made up to 
10 ce. with distilled water. The tubes were sterilised by steaming for 30 minutes 
on each of three successive days. Each was then inoculated with from 50 to 
500 cells as described above, the tubes were aerated by blowing air into them 
through a sterile glass tube plugged with wool and having its end drawn out 
to a fine point, and were incubated at 25°. 

The tubes were examined each day and the number of cells counted by 
means of a Thoma haemacytometer. At least four drops were counted from 
each tube and if any considerable discrepancy was found between them the 
counting was continued until the margin of error was reduced below 5 %. 
Owing to the number of tubes it was not possible to count all of them every 
day, but the different rates at which the yeast grew in the various tubes 
rendered this unnecessary. The results are set out in Table I. 


Table I. 
Percentage Days: growth in millions of cells per cc. 
Number of lemon x A : 
of tube juice (NH,),SO, 1 2 3 4 5 6 9 
1 0-1 0 8.2. 8.2. 8.2. 8.2. -- 8.g. 1-6 
il 0-1 + - nd . _ ae 15 29-0} 
2 0-25 0 ms a ms Lt _ 1-8 1-0 
12 0-25 + ‘s a - ne os 4:3* 32-67 
3 0-5 0 ‘ia ma is 1-3 -—~ 3-5 1-0 
13 0-5 + “ < is 1-7 2-1* 34-2} 
+ 0-75 0 # ns 5 0-9 — 2-5 1-8 
14 0-75 + te - pi 1-0 4-9* 26-77 
5 1-0 0 Ar en 2:8 3-1 — 2-7 6-1 
15 1-0 + 5, Md Ll 1-4 a 4:6* 22-6 
6 2-5 0 a 2-8 4:3 -- 6:3 — ~~ 
16 2-5 + “ 3-1 55 — 21-2 — — 
at: i 5-0 0 i 4-9 71 _ — 
17 5-0 + € 4:8 27-4 — -- — 
8 75 0 es 6-7 9-2 — — — — 
18 7-5 + es 21-1 19-4 -- —_— — 
9 10-0 0 x 12-0 12-1 — — — — 
19 10-0 + ” 26-5 29-4 _ — — — 
10 15-0 0 m 15-9 19-6 _ — — 
20 15-0 + nb 24-3 40-6 — — 


S.g. means some growth but too slight to count. 

* Involution forms present. 
: { Films formed. Accurate counting impossible owing to clumping of cells. Many involution 
orms. 


140 O. K. WRIGHT 


It is well known that the growth of yeast is susceptible to slight variations 
of conditions which cannot be controlled absolutely, such as the degree of 
aeration and the formation of products of metabolism and fermentation. 
Further, the accuracy of the counting is affected by the smallness of the drop 
examined in the low counts and by the tendency of the cells to form clumps 
in the higher counts or in the later stages of growth, so that below 3,000,000 
and above 15,000,000 cells per cc. the counts of different drops of the same 
sample are found to be noticeably less uniform than between those figures, 
and the counts made after more than six days of growth can only be regarded 
as approximate. In considering the results obtained regard must be paid to 
these facts, and account taken only of large differences in the numbers. The 
table shows, however, that, until the yeast reaches a concentration of some- 
where in the neighbourhood of five or six million cells per cc., the rate of 
growth is independent of the presence or absence of (NH,),SO, and depends 
on the concentration of the lemon juice. 

It does not appear unreasonable that it should take longer for an individual 
cell to collect sufficient nutriment to enable it to multiply when the nutriment 
is dilute than when it is concentrated, and possibly the fact that cells were 
unable to multiply in a minute drop in dilutions that enabled growth to take 
place in larger quantities of medium is explained by the fact that the whole 
drop did not contain sufficient nutriment. 

After the yeast reaches a concentration of five or six million cells per cc. 
it is able to continue growing freely in the (NH,),SO, tubes. Further it was 
observed that after six days involution forms, as described by Will [1895] in 
his investigation of film formation, begin to appear in all tubes containing 
(NH,).SO, where the growth had not reached the critical point of five or six 
million cells per cc., and by the ninth day there was a heavy growth of film 
yeast on all these tubes. All the remaining tubes after a lapse of three weeks 
showed only very rare involution forms and no film formation, but most of 
the cells were highly refractile with many large vacuoles and granules and a 
thick membrane, similar in appearance to the permanent cells described by 
Will. It would appear, then, that after an interval of six or seven days the 
cells in smaller concentrations than five or six million per cc. are able to 
adapt themselves to the use of (NH,),SO, and that film formation results. 

A similar series of tubes was prepared with increasing percentages of 
aqueous yeast extract instead of lemon juice. This was prepared by Osborne 
and Wakeman’s method [1919]; about 250 g. of wet brewer’s yeast was washed 
and boiled for two minutes with 1 litre of distilled water containing 0-01 % 
of acetic acid and the liquid was separated off in the centrifuge. The solid 
residue was again heated with 500 ce. of 0-01 % acetic acid and the extracts 
united and concentrated to 500 cc, This extract contained 2-856 mg. of 
nitrogen per cc., of which 0:2273 was ammoniacal nitrogen. As this is rather 
more than 30 times the amount contained in the lemon juice, the amounts 
added to the various tubes of the yeast extract series were much smaller than 
in the lemon juice series. The results are set out in Table IT. 


ACTION OF YEAST-GROWTH STIMULANT 141 


Table IT. 
Percentage Days: growth in millions of cells per cc. 
Number of yeast — A ene 
of tube extract (NH,),SO, 1 2 3 + 5 6 11 
21 0-01 0 8.2. 8.2. 8.2. 8.2. = = 3-1 
31 0-01 + ” ” ” oe bag —— 10-6* 
22 0-02 0 ” ” ” ” = 3-0 
32 0-02 + - aa 7 be — 10-5* 
23 0-05 0 ” ” ” ” ere os 4:0 
33 0-05 + rd Pry s ” ood sera e 15-0* 
24 0-1 0 r ‘s o4 —_ 1-6 10-0* 
34 0-1 4 as sy 3 = — 15-1* 20-0* 
25 0-15 0 7 wy 2 6-1* — 10-0* —_— 
35 0-15 + “- os ~ 3-6 ~ 11-5* — 
26 0-2 0 a * 0-5 3-6 5-1 — 
36 0-2 + ¥ bs 0-5 3-7 15-5 — obacs 
27 0-5 0 ” 4:9 7-6 ound — —_ — 
37 0-5 + ” 4-4 23-7 — _ — 
28 1-0 0 rp 5:3 14-5 “= — — 
38 1-0 + Go 11-8 39-5 —_— ml a 
29 2-0 0 si ' 136 27-2 — te 
39 2-0 + te 19-9 43-0 one nie 
vt <4 . " } Very heavy growth. Not counted. 


oe means some growth but too slight to count. 
* Involution forms present. Very slight film formation. 

On comparing the results obtained with the yeast extract with those 
obtained with the lemon juice it will be seen that their general trend is similar, 
and that the yeast extract is approximately ten times as effective as the 
lemon juice although its nitrogen content is more than 30 times as great. The 
results obtained are somewhat less uniform than those obtained with the 
lemon juice, and tubes Nos, 24 and 25 fall out of line on the 11th and 4th days 
respectively. Film formation does not occur so readily, only very slight 
traces being present in tubes Nos. 31, 32, 33, 34 and 35 by the 11th day. 

These differences may be due to differences in the nature of the nitrogen 
in the two substances under examination or to the presence of toxic substances 
in the yeast extract which are absent from the lemon juice. The yeast experi- 
ment, however, affords confirmation of the fact that the rate of growth is at 
first independent of the presence or absence of (NH,),SO, and depends on 
the concentration of the “bios” until the yeast has reached a concentration 
of somewhere in the neighbourhood of five or six million cells per cc., after 
which it proceeds further in the presence of the (NH,),SO,. 


GENERAL CONCLUSIONS. 


The suggestion that “bios” may be of the same nature as the vitamins, 
which are widely held to be necessary for the nutrition of higher animals, 
and the further suggestion that it may be identical with one of the recognised 
vitamins, open up possibilities for investigating the subject of vitamins which 
are particularly attractive, owing to the ease with which a biological process 


142 O. K. WRIGHT 


can be studied in a yeast as compared with a similar study in the case of a 
more complex organism. Before we can proceed to investigate the general 
question of vitamins by such methods it is necessary, firstly, to be certain 
that “bios” is actually a vitamin, 7.e. a substance whose presence in the 
food in relatively small quantities is necessary to enable the remainder of the 
foodstuffs to be properly assimilated and utilised by the organism, and 
secondly that the substance in question is identical with one of the recognised 
vitamins. The present investigation points to the fact that “bios” does not | 
enable the yeast to assimilate (NH,),SO, simply by its presence or by being 
consumed at the same time, but that the yeast grows solely at the expense 
of the “bios” until it reaches a certain degree of concentration, and after 
that it is able to use the (NH,),SO,. No explanation is offered of this pheno- 
menon which requires further investigation. 


I wish to thank Prof. A. Harden, who suggested the investigation, for his 
advice and assistance in carrying it out. 


REFERENCES. 


Bachman (1919). J. Biol. Chem. 39, 235. 

Eddy (1921). The Vitamine Manual, Chap. rv (Baltimore). 
Harden and Zilva (1918). Biochem. J. 12, 259. 

Osborne and Wakeman (1919). J. Biol. Chem. 40, 383. 
Osborne and Mendel (1920). Jbid. 42, 465. 

Wildiers (1901). La Cellule, 18, 313. 

Will (1895). Zeitsch. ges. Brauwesen, 18, 1. 

Williams (1919). J. Biol. Chem. 38, 465. 

(1920). J. Biol. Chem. 42, 259. 


XX. THE FUNCTION OF PHOSPHATES IN THE 
OXIDATION OF GLUCOSE BY HYDROGEN 
PEROXIDE. 


By ARTHUR HARDEN anp FRANCIS ROBERT HENLEY. 
From the Biochemical Department, Laster Institute. 


(Received January 25th, 1922.) 


THE conditions under which the oxidation of glucose by H,O, can take place 
in aqueous solution have been studied by W. Léb and his co-workers [1910, 
1911, 1912, 1915]. He believes that the oxidation depends on the hydroxyl 
ion concentration of the solution, and further that phosphate ions exercise 
a specific accelerating effect on the reaction. He was unable to produce any 
oxidation of glucose by H,O, in presence of water alone or when other sub- 
stances were substituted for phosphates. 

For instance neither Sdérensen’s mixtures of glycocol and NaOH nor of 
borate and HCl could be used to replace the phosphate mixture. Witzemann 
[1920] confirmed Léb’s observations and agreed that phosphate mixtures 
exercise a specific effect. He was unable to produce the same effect with 
mixtures of NaHCO, and Na,CO, as with phosphate mixtures. 

Witzeman [1920] suggested that the phosphates might produce an inter- 
mediate compound with glucose of the same nature as the hexosephosphate 
which has been shown by Harden and Young [1910] to be produced during 
alcoholic fermentation, and which appears from the work of Embden and 
others to be connected with the production of lactic acid in muscle. 

Witzemann was unable to detect the formation of any compound of 
glucose and phosphate in absence of H,O,. 

As more exact knowledge of the mechanism of this reaction is desirable 
in view of the part played by phosphate in alcoholic fermentation, further 
experiments were made to try and ascertain whether an intermediate com- 
pound of the nature of a phosphoric ester of glucose was formed during this 
oxidation. For this purpose free phosphate was estimated at intervals in a 
mixture similar to one of those used by Witzemann for oxidation of glucose 
by H,O, at 37°, both by the method of Schmitz [1906] and by the ordinary 
magnesium citrate method. No change in the amount of free phosphate was 
observed. . 

Besides the possibility of the formation of an intermediate compound 
between phosphate and glucose there are at least two other possible explana- 
tions of the effect produced by phosphate in this reaction: 


144 A. HARDEN AND F. R. HENLEY 


(1) Phosphate may act as a peroxidase, thus rendering possible the oxida- 
tion of glucose by H,0,. If this were so phosphate might act in the same way 
towards mixtures of H,O, and benzidine or guaiacum. No evidence could be 
obtained of such an action, but it does not follow from this negative result 
that no such action occurs with H,O, and glucose. 

(2) The phosphate mixtures may simply act as buffer substances, the 
maintenance of the P,, within certain points being essential. If this hypothesis 
is correct it should be possible to replace the phosphate mixtures by other 
substances, provided the P;, be maintained within the same limits. 

We have carried out experiments on the oxidation of glucose by H,0, 
with solutions the P,, of which was never over 7-3 and find that glucose is 
oxidised by H,O, in presence of 0-125 M NaHCO, saturated with CO,; or 
0-25 M sodium arsenate saturated with CO,; or 0:25 M sodium acetate. 

The phosphate ion cannot therefore be regarded as playing a specific part 
in the reaction. It appears more likely that the buffer action of the salts 
employed is the important factor, whether as providing and maintaining the 
most suitable conditions for oxidation or as protecting the H,O, from too 
rapid decomposition. In this connection the experiment of Witzemann [1920, 
p. 12, sec. 7] with a mixture of sodium carbonate and bicarbonate should be 
referred to. This shows that in presence of these salts, at a P,, slightly over 
9-3, H,O, is rapidly decomposed and little oxidation takes place, whereas 
in our experiment in presence of 0-125 M NaHCO, saturated with CO, (Py, 6-8) 
a considerable amount of oxidation is produced. 

Experiments were accordingly made roughly to test the stability of H,O, 
in solutions the P,, of which was maintained at various levels by the use of 
buffer solutions. The rate of decomposition was found to increase with rise 
of alkalinity. In those cases in which the buffer solutions consisted of phos- 
phate mixtures the rate of decomposition of the H,O, was not so rapid as in 
the corresponding experiments with other buffer solutions; so that in this 
sense the phosphate ion may be said to have a specific action. 

In the experiments of both Léb and Witzemann the P,, of the solutions 
was only carefully measured in those experiments in which Sérensen mixtures 
were used. In most if not all of the experiments in which the effects of other 
substances were compared to that of phosphate mixtures the reaction of the 
solution was tested with litmus paper only. The glycocol + NaOH and 
borate + HCl mixtures used are only efficient buffers between P,, 8-0 and 
10-2, and at these higher degrees of alkalinity the H,O, is rapidly decomposed. 
The same applies to the NaHCO, + Na,CO, mixture used by Witzemann 
which must have been much more alkaline than the phosphate mixture. 

The experiments in which he compared the effect of Na,CO, and NaOH 
were not carried out at the same P;,. The reaction of the solution in the latter 
case was suitable for the oxidation, whereas the alkalinity of the solution used 
in the former case was too high. 


OXIDATION OF GLUCOSE BY H.,0, 145 


EXPERIMENTAL. 


I. Experiments to determine if mixtures of Na,HPO, and NaH,PO, will act as 
a peroxidase towards H,O, and benzidine or guaiacum. 

Phosphate mixtures of P,, 9-0, 7-4, 6-9, 6-6, 6-2, 4-8 were tested with H,O, 
and benzidine or guaiacum. No action was observed in any of the experi- 
ments. 

To test whether the optimum P,, for the action of a peroxidase lies within 
the above range, similar experiments were made with the same phosphate 
mixtures, H,O, and guaiacum with the addition of potato juice. The maximum 
effect was produced at room temperature from P,, 7-4 to 6-9. A similar result 
was obtained when benzidine was substituted for guaiacum. 


II. Oxidation of glucose by H,O, in presence of various buffer mixtures. 
Seven flasks were made up as follows: a 


Na,HAsO, NaC,H,0,. 5-7 % sol. 
Glucose NaHCO, 12H,0° 34,0 K,HPO, KH,PO, H,0, 

g. g. g. g. g g.- cc. 
Ll 875 26 Gi i oft as 62 
2 $765 ees 25-1 te ee we 62 
3&5 3 as 8-5 wi eS 62 ‘ Es 
4 875 — ~~ ee OR ee eet ecte 
56 1125 26 pas 24 ah ca ) 
6 11-25 es 25:1 x me em om 
7 11-25 ie oo 8-5 be fat a 


In Nos. 1, 2, 5, 6 the solution was saturated with CO,. 

The P,, of each solution was estimated colorimetrically, and the glucose 
estimated by Bertrand’s method and by the polarimeter. 

All the flasks were then kept in the incubator at 37° for 24 hours and the 
P,, and glucose again estimated in each. The results are shown below: 


At start of experiment 
glucose g. per 100 cc. Apparent loss of glucose 
7 A —— After 24 hours in g. per 100 ce. 
Pu Bertrand Polarimeter _— A — a 
A B P A B by A by B 
1 6-8 4:16 3-50 6-7-6-8 3-39 2-99 0-77 0-51 
2 6-8 3-96 3-55 6-7-6-8 2-96 2-38 1-00 1:17 
3. 68-70 3-64 3-54 5-2 3-01 2-50 0-63 0-53 
+ 7-0 3-65 3-20 56-58 2-24 1-85 1-41 1-35 
5 6-8 _ 4-55 7-6 — 4-55 -- 0 
6 6:8 — 4-50 7-0 — 4-50 —_— 0 
7 7:3 oa 4 7 — 4-25 -- 0 


Note. The sample of glucose employed for these experiments showed a 
5 % higher content of glucose as estimated by Bertrand’s method than when 
the polarimeter and [a], 52:5° were used. This, and the fact that the Bertrand 
estimations of glucose at the start of the experiment were carried out before 
the polarimeter readings were made, may account for the discrepancies be- 
tween the two sets of figures shown above. 

Only traces of H,O, remained in 1, 2 and 4 after 24 hours in the incubator. 

No. 3 still contained a considerable amount of H,O, and the action was 
allowed to proceed for a second period of 24 hours. After this the Py was 


Bioch. xvr 10 


146 - A, HARDEN AND F. R. HENLEY 


found to be 4-6-4-8 and glucose estimated polarimetrically amounted to 
1-76 g. per 100 cc.: a total apparent loss of glucose of 1-78 g. per 100 cc. 

These experiments show that in absence of H,O, no loss of glucose has 
apparently taken place (Nos. 5, 6, 7). Whereas in presence of H,O, (Nos. 1, 2, 
3, 4) an apparent loss of glucose has occurred in every case, greatest in the 
two solutions containing phosphates and arsenates, but considerable in the 
other two solutions. 


Ill. Effect of change of Py on the stability of H,O.. 
All solutions contained 5-0 cc. of 5:7 % H,O, per 100 cc. solution. The 
H,0, was estimated at the start and again after 24 hours and 48 hours at 37° 


by titration with standard permanganate. 
°/ H,0, lost 


i" O Fe Paes! 
after 24 hours after 48 hours 


No. Mixture Pu 
1 HO)? Sis a: ph ate 1:0 —_ 1 
2 Sérensen’s glycocol+ HCl ... 1-9 1 2-1 
3 Na acetate +acetic acid 4-6 2 2°1 
t Phosphates... 6:8 2 4-2 
5 = a 7-1* 4:9 7:5 (43 hours) 
6 Se 7-8* 6-1 7-4 (43 hours) 
- e ae aK ois 8-0* 7:5 13-1 
8 Palitzsch’s borax + boric acid 6-9 20-2 36-1 
9 * = Zo 71 20-2 37-2 
10 uM in ee 7-7 22:3 39:3 
ll aA a 79 37:2 59-5 * 
12 = G BS 8-2 55:3 74:4 
13 es = ue 8-4 65-9 84-0 
14 ” ” ” 8-6 65-9 84-0 
15 f as at 8-9 67-0 85-1 
16 Sérensen’s borate + HCl cae 8-8 — 78-5 
17 ” +NaOH ... 9-2 — 99-0 
18 Ee a cae 12-8 100 — 


* In these cases the concentration of the hydrogen peroxide used was 5 % instead of 5-7 %. 


IV. The P,, of some of the solutions tested by Witzemann. 


A solution was made up as described by Witzemann [1920, p. 12, sec. 7]: 
2-43 g, Na,CO,.10H,O; 0-72 g. NaHCO,; 35 cc. H,O; 20 ce. (0-2 g.) glucose 
solution and 20 cc. of 3 % H,0,. 

The Py, estimated colorimetrically, was over 9-3, At this degree of alkalinity 
H,0, is rapidly decomposed and it is therefore not surprising that little 
oxidation took place. 

To compare the effect of NaOH and Na,CO, Witzemann [1920, p. 14, 
sec, 9 (2) and sec, 10 (2)] prepared solutions as follows: 

P. 14, sec. 9: 32-0 cc, 0-33 M Na,HPO,; 3-0 cc. HO; 20-0 ee. 1 % glucose; 

20-0 ce, 3% H,O,; 0-61 g. NagCO,.10H,0; 
and sec. 10(3) 25-6 ec. 0-33 M Na,HPO,; 6:4 cc. 0-33 M NaH,PO,; 20-0 cc. 
1 % glucose; 20-0 ec. 3 % H,O,; 3 cc. H,O; 0-1714 g. NaOH. 

The P,, of the former was 9-3-9-6 and of the latter 7-1. The amount of 
oxidation was greater in the second than in the first case and Witzemann 
argues that NaOH exercises a less harmful effect on the reaction than does 
Na,CO,, 


OXIDATION OF GLUCOSE BY H,0, 147 


It does not seem that this inference can fairly be drawn unless the experi- 
ments are carried out under the same conditions of P,,. The acidity of the 
H,0, solution used by Witzemann is not stated in his paper. It is therefore 
not possible to say what was the exact P,, in his experiments. But in any 
case the alkalinity of the solutions used in experiments 7 and 9 (2) must 
have been much higher than in experiment 10 (3). _ 


SUMMARY. 


1. The oxidation of glucose by H,O, takes place in presence of the following 
buffer substances: NaHCO, + CO,; Na,HAsO, + NaH,AsO,; NaC,H,0,; 
K,HPO, + KH,PO,, provided the P,, does not rise much above 7-3. 

2. The stability of H,O, in aqueous solution is increased by the presence 
of phosphates which do not however show a specific action in other respects. 


REFERENCES. 


Harden and Young (1910). Proc. Roy. Soc. B. 82, 321. 

Lob, W. (1911). Biochem. Zeitsch. 32, 43. 

Léb, W. and Beysel (1915). Biochem. Zeitsch. 68, 368. 

Lob, W. and Gutman (1912). Biochem. Zeitsch. 46, 288. 

Lob, W. and Pulvermacher (1910). Biochem. Zeitsch. 29, 316. 
Schmitz (1906). Zentr. anal. Chem. 45, 513. 

Witzemann (1920). J. Biol. Chem. 45, 1. 


XXI. VITAMIN REQUIREMENTS OF DROSO- 
PHILA. 1. VITAMINS B AND C. 


By ARTHUR WILLIAM BACOT ann ARTHUR HARDEN. 
From the Departments of Entomology and Biochemastry, Lister Institute. 


(Received January 30th, 1922.) 


Lors [1915] ascertained the fact that flies of the genus Drosophila could be 
reared on a culture medium containing as mineral constituents only K, Mg, 
PO, and SO, ions. In these experiments micro-organisms were not excluded, 
and the growing larvae undoubtedly lived at the expense of these. Loeb also 
found that when Na was substituted for K no growth took place and con- 
cluded that K was essential for the development of the insect. It is however 
possible that the absence of K stopped the growth of the micro-organisms 
and thus deprived the larvae of nitrogen in an available form. Loeb’s experi- 
ments therefore only prove that the mineral constituents named above are 
adequate, but not that they are essential. 

Subsequently Loeb and Northrop [1916] succeeded in sterilising the eggs 
of Drosophila by exposure for 6-7 minutes to 0-1 % aqueous HgCl, or a satu- 
rated alcoholic solution of the same salt and found that the larvae hatched out 
from the few eggs which survived the disinfection were sterile. They ascer- 
tained that these sterile larvae could not be reared on sterile media consisting 
only of purified proteins, carbohydrates, fats and salts, but that when yeast 
killed by heat was added, normal development occurred. Alcoholic extracts 
of yeast were however inactive. Northrop [1917] pursued the subject further 
and found that the addition of caseinogen and similar substances along with 
cane sugar to yeast greatly increased the number of flies which could be 
reared at the expense of a definite amount of yeast and concluded that they 
acted as foodstuffs when properly supplemented by yeast. Similar results 
were obtained by the addition of kidney, liver and pancreas from the dog, 
but many other organs of the dog were inactive. The author was inclined to 
interpret these results in the sense that yeast supplies one or more accessory 
substances, necessary for the growth of the organism. 

The question was again investigated by Baumberger [1919] who found 
that the pupae and eggs of flies reared in presence of a pure culture of yeast 
could readily be sterilised by immersion for 20 minutes in 85 % alcohol. The 
sterile larvae obtained from such eggs could easily be reared in sterile culture 
media provided the latter contained killed yeast or nucleo-protein prepared 
from yeast. Baumberger came to the conclusion that the essential requisite 


VITAMIN REQUIREMENTS OF DROSOPHILA 149 


was a sufficient concentration of yeast nucleo-protein and that accessory 
food factors were not concerned. 

The following experiments were instituted with the object of ascertaining 
more definitely whether any of the recognised vitamins are essential for the 
nutrition of these flies. The subject is of general interest as regards the extent 
to which the need for accessory food factors is distributed in the animal 
kingdom. Special interest attaches to the possibility of introducing a sim- 
plified technique by which results could be obtained more rapidly and econo- 
mically than by the tedious and costly feeding experiments which are at 
present necessary. 


TECHNIQUE OF THE EXPERIMENTS. 


It was soon found that the eggs could be sterilised much more readily 
and certainly than the pupae and this method was always employed. 

Following Baumberger’s instructions the flies were reared in banana-agar 
(approximately equal parts of mashed banana and 1-5 % water-agar) which 
had been autoclaved and subsequently inoculated with yeast (S. cerevisiae). 
After a few sub-cultures on this medium the flies were found to be free from 
bacteria, both anaerobic and aerobic, and the pure stock was readily maintained 
by occasionally passing a few flies into a fresh banana-agar-yeast tube. 

When eggs were desired for experimental purposes, freshly emerged flies 
were transferred to a special apparatus. This consisted of an inverted wide 
necked bottle, standing in a Petri dish and having a strip of moistened filter 
paper on the wall. A small circular glass dish with vertical walls was also 
placed on the Petri dish, so as to be covered by the inverted bottle. The 
walls of this dish were lined with hardened filter paper moistened with water. 
The bottle was in turn covered by an inverted cylindrical tinned iron can and 
the whole apparatus was autoclaved at 120° for half-an-hour. A few cc. of 
a 24-48 hour culture of yeast in yeast extract containing glucose were then 
placed in the small dish, the flies introduced into the inverted bottle and 
the whole apparatus placed in an incubator at 30°. Under these conditions 
the eggs were laid within 24-48 hours almost exclusively on the hardened 
filter paper lining the small circular glass dish containing the yeast culture. 
The strips of paper were removed, placed on a sterile slide under a dissecting 
microscope and the eggs picked off singly by means of sterilised dissecting 
needles and placed in 85 % alcohol as directed by Baumberger. They were 
then divided up into collections of the number required for each tube by means 
of a sterile pipette provided with a rubber teat, and were finally, after half- 
an-hour’s exposure to the alcohol, sucked up into the pipette and delivered 
into the appropriate tube of medium. It was found most convenient to place 
20 eggs in each test tube (6x ?’’) containing 5-5-5 cc. of medium in the 
form of an agar slope. 


The agar jelly in the medium tube must be so soft that the larvae can 
readily penetrate it in their search for food. As a rule 3 cc. of 13 % agar 


150 A. W. BACOT AND A. HARDEN 


were added to 2 cc. of medium and the whole then steamed for 20 minutes 
on three successive days, and cooled in the form of slopes. | 

The inoculated tubes were incubated at 30° and examined daily, note being 
made of the number hatched, the general progress and the time of appearance 
of pupae and flies. 

At the close of each experiment the tubes were tested for sterility by 
making cultures on beef broth-peptone-agar and incubating at 37°. Con- 
tamination by yeast usually showed itself in the experimental tube, but was 
when necessary tested for by culture on beer wort agar at 25°. 


__ Diet. 
A basal diet consisting of 
Caseinogen (purified) ... ae ... 015g. per tube 
Starch and salts! (94 ie starch and 6 ¥, salts) 0-05 g. a 
Cane sugar a 3 i ae agree Oe me ey: 


was employed, to which various additions of butter-fat (vitamin A) yeast- 
extract (vitamin B) and lemon juice freed from citric acid (vitamin C) were 
made as required, 2 cc. of liquid in all being added, and then 3 ce. of 15% 
agar. The caseinogen had been extracted with alcohol and light petroleum 
or in some cases heated at 120° for many hours and repeatedly stirred. 


RESULTS. 


The experiments were directed in the first instance to ascertaining which, 
if any, of the three recognised accessory food factors are necessary for the 
growth and metamorphoses of these insects. 

The results, which are summarised below, showed definitely that: 

(1) in the presence of butter-fat and yeast extract (vitamins A and B) 
growth occurred irrespective of the presence or absence of 1 cc. of lemon 
juice (vitamin C); 

(2) in the presence of butter-fat (vitamin A) growth occurred in presence 
of 1 ce. of yeast extract (vitamin B), but not in its absence, this result being 
independent of the presence or absence of lemon juice; 

(3) in the presence of 1 cc. of yeast extract (vitamin B), with or without 
1 cc. of lemon juice (vitamin C), growth occurred in the presence of two drops 
of butter-fat, but not in its absence. 

In the following summary only results of experiments in which the tubes 
remained sterile are quoted. 

Additions to basal diet 


Butter- Yeast Lemon Tubes Eggs Pupae Flies 


fat extract juice inoculated hatched formed emerged 
! - + + 10 100 0 0 
2 + + - 1h 115 92 63 
3 + - + 10 57 0 0 
4 + + + 10 68 48 42 
5 - + - 2 19 0 0 
6 - - - 2 16 0 0 
1 


As used in this laboratory for rats, 


VITAMIN REQUIREMENTS OF DROSOPHILA 151 


Control experiments were carried out to ascertain whether the yeast 
extract alone was capable of supplying sufficient nitrogen for the growth of 
the larvae, but with entirely negative results. This was done by omitting the 
caseinogen from mixtures 2 and 4, which gave positive results in its presence. 
Even the addition of 2 cc. of the yeast extract gave entirely negative results 
in the absence of the caseinogen. 


Other sources of vitamin B. 


As flies of this group are specially associated with yeast, which probably 
forms the chief food of the larvae in their natural condition, it was thought 
that in order to prove the necessity for vitamin B, it would be essential to . 
show that it could be supplied from other sources. 

The first experiments were made with milk, but without success. Recourse 
was then had to wheat germ, the alcoholic (80%) extract of which is known 
to contain a considerable amount of vitamin B. 

50g. of wheat germ were heated on the water-bath for one hour with 
200 ec. of 80 % alcohol, pressed out and the clear liquid evaporated to dryness 
at 50°. The residue was shaken up with 50 ce. of water and filtered, the clear 
filtrate being used. This contained 10-47 g. total solids, 0-21 g. mineral matter, 
and 0-196 g. N per 100 ce. 

Tubes were made up with 2, 1 and 0-5 ce, of this extract in addition to the 
usual basal diet of caseinogen, starch, salts, cane sugar and butter-fat. Some 
difficulty was experienced in adjusting the stiffness of the agar jelly, but 
finally 0-5 ce. HO and 3 ce. of 1-25 % agar were added to each tube, so that the 
total volume was 5-5 cc. as shown below. 


Caseinogen, Cane Butter- Wheat germ 1-25 % 

starch and salts sugar fat extract H,O agar 
g. g. ce. ce. ce, 
7 0-2 0-1 2 drops 2 0-5 3 
8 0-2 O-1 & l 1-5 3 
9 0-2 0-1 aS 0-5 2-0 3 

Starch and salts 

10 0-05 0-1 ” 2 0-5 3 


The following are the summarised results of three quite independent and 
consistent experiments: 
Wheat germ Eggs known to 


extract have hatched Pupae Flies 
ll 2 34 18 9 
12 1 56 24 13 
eee 0-5 39 0 0 
14 2 ce. as sole 9 0 0 
source of N 


The experiments were somewhat hindered by the hot weather prevailing 
at the time, which caused unduly rapid drying of the tubes, but they show 
quite conclusively that, like yeast extract, wheat germ extract is capable of 
supplying some factor necessary for the growth and metamorphosis of the fly. 

It is at the same time obvious that the wheat germ extract (1 ce. of which 
corresponds with 1 g. of wheat germ) is much less efficacious than the yeast 


152 A. W. BACOT AND A. HARDEN 


extract (1 ce. of which corresponds with 0-5 g. yeast) and this is in accordance 
with what is known of the vitamin B content of these two materials, 1 g. of 
yeast being approximately equivalent to 3-6 of wheat germ. 

Taken together the experiments with yeast extract and wheat germ extract 
leave little doubt that these insects require vitamin B for their development, 
thus confirming Northrop’s view. Experiments are in progress on the amount 
of yeast extract required and the effect of varying concentrations on the 
duration of the period of development. 


THe Fat-soLUBLE Factor (VITAMIN A). 


The striking fact that in presence of 1 cc. yeast extract growth occurred 
in the presence of two drops of clarified butter-fat, but not in its absence, may 
be taken to show either that a fat of some kind is necessary or that the fat- 
soluble factor is required or that both are required. To decide between these 
possibilities it is necessary to ascertain whether or not growth occurs in 
presence of a fat devoid of vitamin A. Attempts of this kind are being made, 
but no definite result has yet been obtained. The preliminary results however 
indicate that the fly is certainly able to develop in the presence of very small 
amounts of the vitamin if not in its entire absence. 


SUMMARY. 


The complete development of Drosophila requires the presence of vitamin B — 
but not of vitamin C. 


REFERENCES. 


Baumberger (1919). J. Hap. Zoology, 28, No. 1. 
Loeb (1915). J. Biol. Chem. 28, 431. 

Loeb and Northrop (1916). J. Biol. Chem. 27, 309. 
Northrop (1917). J. Biol. Chem. 30, 181. 


XXII. STUDIES ON CARBONIC ACID COMPOUNDS 
AND HYDROGEN ION ACTIVITIES IN BLOOD 
AND SALT SOLUTIONS. 


A CONTRIBUTION TO THE THEORY OF THE EQUATION 
OF LAWRENCE J. HENDERSON AND K. A. HASSELBALCH. 


By ERIK JOHAN WARBURG. 
From the Laboratory of the Finsen Institute (Copenhagen). 


(Received March 3rd, 1922.) 


CONTENTS PAGE 
Introduction t ° . : ‘: > ‘ : ‘ ; ‘ ; 153 
CHAP. 
I. Some general physico-chemical Rin ene nes to the new caucaratee dis- 
sociation theory of Bjerrum 155 
II. The hydration of CO, and the Aleanclaiitin - satiade iad The detivailens a 
Henderson’s and Hasselbalch’ 8 ent and the modification of these in 
accordance with Bjerrum’s theory . 169 
III. The development of a modified Henderson- Sassattadch equation ‘tae use in the 
case of heterogeneous solutions ; 177 


IV. Methods for the electrometric determination of the ap even ediengial ion sailstier 
in solutions containing CO, with the hs g of the platinum hydrogen Sey. 


and the technique employed. - 180 

V. The direct empirical determination of hen i in ‘the modified a Hasethaleh 
equation in its application to blood . 193 
VI. The determination of the pA’) variation. : é ‘ ‘ ; : ; 202 
VII. The reaction of the blood corpuscles. ‘ 222 

VIII. The determination of the first dissociation constant ia cushoadn anid ren the 
deviation coefficients of the bicarbonate ion =; 233 

IX. Preliminary considerations eroneres the mode of Acuabhandbees of balbcais etid 
inthe blood ‘ 261 

X. A brief historical review of the older unin of ‘the sciesibindlen of C0, i in the 
blood 266 

XI. Further saiatilenaticgil a e iments on the pareve of the sastbhesiod of CO, 
in the blood, illustrated by the “carbonic acid combination” curve . 275 


XII. The factors which determine the partition of permeating ions between the blood 
corpuscles and the serum, the volume of the blood corpuscles is pape’ ss 
the reaction, as well as the potential of their surfaces ; 307 


INTRODUCTION 


K. A. Hasse.Batcu in 1916 introduced a formula with the help of which it 
is possible to determine the hydrogen ion concentration of the blood from the 
amount of its combined and_ dissolved carbonic acid, basing his equation upon 
the work of L. J. Henderson carried out from 1908-1910. Henderson em- 
ployed the then generally known equation for the first dissociation of carbonic 
acid, namely, 
CrCHco’s _ K 
= 


Y CH2COs 
Bioch. xv1 11 


154 E. J. WARBURG 


Applied to the determination of the Cy: of the blood he got 
. — CH2C0s , Ki, 
CHco’s 

By substituting the best known values at the time for combined and dis- 
solved CO,, for the first dissociation constant of carbonic acid and for the 
dissociation of bicarbonate, he found a value for Cy: which corresponded 
fairly well to the reaction of the blood as it was thought to be at that time. 
Henderson [1909-1910] concluded from his calculations that the major portion 
of combined carbonic acid in the blood was in the form of bicarbonate. 

Hasselbalch [1916] determined in the same sample of blood (and serum) 
Cy, Cyco’, and Cy,co,, and found that Cy was the same as calculated by 


substituting in the equation the value of “ for a solution of pure sodium 


bicarbonate. He thought he could infer from this that all the combined 
carbonic acid of the blood was in the form of bicarbonate and that the Cy: 
of the blood could be estimated by means of the equation when Cy,co, and 
Cyco’, Were known. 

As investigations on kindred subjects progressed it became clear the theory 
could only be recognised as a first approximation—a view which a knowledge 
of 8. P. L. Sérensen’s Studies on Proteins [1915-1917] strongly supported— 
and the question immediately arose, how is it possible the calculated and 
experimental values of Cy: can coincide when the suppositions of the calcu- 
lation are inadequate? 

It further became evident from experience gradually accumulated re- 
garding the Cy" measurements by the Hasselbalch-Michaelis principle (shaking 
method combined with minimal immersion), that experimental difficulties 
were present which had not been entirely overcome, since measurements with 
various electrodes gave constant differences. And as the Henderson-Hassel- 
balch equation seemed to be destined to play an important rdéle in physiology 
and pathology it was necessary to subject both the measurements and the 
theories which formed the basis of the equation to a further test. 

The present work therefore has a double object, viz.: 


(1) to investigate whether the Henderson-Hasselbalch equation gives 
correct results when the measurements are carried out by another method 
than that of Hasselbalch; 

(2) to develop a theory connecting the amount of carbonic acid in the 

blood with its reaction. 
In connection with this a number of points have naturally arisen concerning 
the distribution of ions between blood cells and serum, and a series of experi- 
ments on the reactions in simple watery bicarbonate solutions in equilibrium 
with different tensions of CO, have been made. 


THEORY OF THE HENDERSON-HASSELBALCH EQUATION 155 


CHAPTER I 


SOME GENERAL PHYSICO-CHEMICAL CONCEPTIONS RELATING TO THE 
NEW ELECTROLYTIC DISSOCIATION THEORY OF BJERRUM. 


Osmolar Concentration. 


By the molecular or molar concentration of a solute in a solution we under- 
stand the number of gram-molecules of solute in a litre of the solution. We 
also speak of the molar concentration of the ions and molecular aggregates, 
looking upon the particles in question as if they were molecules when deter- 
mining the gram-molar weight. In physiology a distinction has often been 
made between molecular and molar concentrations (cf. Hamburger [1902], 
Hedin [1915], Ege [1920]), by imagining all solutes to exist in molecular form 
(thus Na’Cl’ as NaCl) when reckoning the total molecular concentration—a 
conception which is also made use of in this work in stating the molecular 
concentration—while in estimating the molar concentration the solute is con- 
sidered to be in its actual state of aggregation. This distinction has been of 
especial significance in physiology for questions of osmotic pressure as it is 
the total number of dissolved particles that is the determining factor (corrected 
for secondary effects), but as this terminology is not recognised in physical 
chemistry it must be abandoned. In place of the molar concentration of 
some physiologists I have therefore employed the expression osmolar concen- 
tration. 

Activity of a Solute. 

In 1908 G. N. Lewis formed the conception activity. As regards the 
thermodynamical definition of this conception I shall confine myself to referring 
to the article of Lewis [1908], however I shall here quote slightly modified 
the very perspicuous explanation which Lewis gives in the paper mentioned. 

The activity of an ideal gas is equal to the concentration of the gas. - 

The activity of a solute which forms an ideal solution is equal to its 
concentration +. 

If a substance has the same activity in two phases, the substance will 
not by itself go from the one phase to the other. 

If the activity of a substance is greater in one phase than in the other, the 
substance will go from the first phase to the second as soon as it is possible. 

By the activity coefficient of a solute we mean the factor the concen- 
tration of the solute must be multiplied by to give its activity. For 
uncharged mols in aqueous solution the coefficient will nearly always be 
greater than unity as it is then merely an expression for what we call depression 
of solubility, salting out effects, action of salt, etc. For charged mols, “ions,” 
in weak aqueous solutions the coefficient will be less than unity, in strong 
solutions often greater than unity, as will shortly be shown in discussing 
electrolytic dissociation. Following Bjerrum [1916, 1918, 1919], we distinguish 

1 Lewis writes ...is by constant temperature and pressure proportional to the concentration. 
11—2 


156 E. J. WARBURG 


between the apparent and true activity coefficients, the apparent being the 
relation between the activity of the desolvated solute and the concentration 
of the solvated solute, while the true coefficient eliminates solvatation effects. 
In so far as we do not know whether a solute is solvated we will use the 
expression apparent activity coefficient. 


Solubility and Activity of Gases. 

In the case of a solution of the ordinary gases: at room temperature (and 
low pressure), the activity of the gas can be estimated by determining its 
tension in a gaseous phase in equilibrium with the solution. This will be 
understood from what is said later. The gases mentioned in this work, oxygen, 
hydrogen, carbon dioxide and nitrogen, for the ranges of pressure and tem- 
perature used, can in practically every respect be regarded as ideal’. The 
tension of such a gas will thus be a direct measure of its activity. For instance © 
if Henry’s law applies, the amount of gas which is dissolved in a solution in 
a state of equilibrium will be proportional to its pressure. The factor which 
expresses the amount of gas dissolved at a given pressure is called the absorp- 
tion coefficient. Bunsen’s absorption coefficient is used here which is called a, 
and which gives the number of cc. of gas reduced to 0° C. and 760 mm. Hg, 
that are dissolved in 1 cc. of solution at a pressure of 760 mm. of the gas 
itself. At a pressure of P mm. Hg, a cc. of gas will then be dissolved in 1 cc. 


of the solution 
Fi fb Gc a ae a cn eee (1) 


760 
That we use Bunsen’s absorption coefficient and not Ostwald’s, is because 
we want to estimate the concentration of the gas in the solution and the 
former way of expressing it is most suitable for this purpose. A gram-molecule 
of an ideal gas at 0° C. and 760mm. Hg occupies a volume of 22393 ce. 
A solution containing a cc. of gas (0° C. and 760 mm. Hg) in 1 ce. will contain 
1000 a ce. per litre and the gas will therefore have a concentration of 


rapg = 01044656 x @ = Chacal gas. -sssesseseeseseereees (2) 
Expressing the dissolved gas in volumes per cent. (Vols. %) this will be equal 
to 100 a and the concentration will be 

000044656 x (100 a) = Cideat GAB. vvesswvcvssvvcvsesece (3) 
The constant 22393 is valid for an ideal gas; for CO, the corresponding 
constant is 22263*. The difference between these two values is a little over 
} % (the deviation of CO, from Avogadro’s law at 0° C. and 760 mm. Hg). The 
last constant is of course used for the calculation of the concentration of CO,. 

For carbon diowide therefore we have the corresponding equations 


COMMIT 3 bss Og. oases cea aaa (4) 
and 0-00044917 (100 a) = Coo, : eo eeeeeoeeverveeooens (5) 


* An ideal gas obeys the “gas laws.” 
* These constants are caleulated according to F. W. Kiister, Logarithmische Rechentafeln fir 
Chemiker u.s.w, 19th edition, Leipzig, p. 40, 1918. 


THEORY OF THE HENDERSON-HASSELBALCH EQUATION 157 


Tf we desire to express the concentration in terms of the absorption 
coefficient and the tension we have, for an ideal gas, 


Pat x 0000058756 = Cacat gas sss+s+eseteeessesesesees (6) 


and for CO, 
Paz X 0-000059101 = Cog. eeceececesseseceseeeeee (7) 


Activity Coefficients and Depression of Solubility of Gases. 
We will now consider the value @ a little more closely. Equation (1) 
Pa _ q ig valid without exception for pure solvents. In this work we shall 


760 = 
assume the same is the case for all solutions if the correct absorption coefti- 
cient is used but it must be mentioned there may be small deviations in 


heterogeneous solutions. If a gas dissolves in a pure solvent, 
Cgas= Ages (Activity of the gas). ............00c0000. (8) 
If other solutes are already present in the solvent a correction must be intro- 


duced. In such solutions the absorption coefficient is 
100-Y 
“corrected ~ Top (9) 


and the concentration of the gas will then be 


| Pr C5 Ba Cea, oicoecnsceesessecesnsenes (10) 
while the activity of the gas is 

Br Ee ADAM, pica iensvevevencixeccucecess (11) 

and Cgas x Wey ME MMMM co be eddie wis vevicscsesss (12) 


Y is called the relative depression of solubility’, 
- is called the relative absorption coefficient, and 


a y is the apparent activity coefficient of the gas. 


‘Tt will be noticed that the apparent activity coefficient is the reciprocal of the 
relative absorption coefficient. The apparent activity coefficient for a gas is 
called F, (gas), the value @ is reserved for the absorption coefficient. in the 
pure solvent. 

For watery solutions the relative absorption coefficient has been determined 
in a large number of experiments, in different ways, for numerous gases, and it 
has been shown that its value in a given solution is about the same for all gases. 

The relative absorption coefficient has also been determined for various 
solid substances which we have reason to think do not unite with other 
solutes, by estimating the partition of the substance between the solution 
and a liquid which does not mix with the latter and in which the solute under 
investigation may be assumed to form a “true” solution. It has been found 
the same laws apply as for gases. 


1 This applies to homogeneous solutions, whilst the symbol © is used for this quantity in 
a heterogeneous solution. 


158 E. J. WARBURG 


Lastly it has been discovered that the relative absorption coefficient, within 
the experimental error, is independent of temperature. The solutes that most 
markedly depress solubility in water are the electrolytes, but non-electrolytes 
can also have a large effect, e.g. cane sugar. There is an extensive literature 
on the solubility of gases in water which is exhaustively dealt with in 
Rothmund’s monograph [1907] and cited in Landolt and Boérnstein’s tables, 
to which the reader is referred. 

Bohr and Bock’s [1891] values for the absorption coefficients have been 
used. . : 


Table I. The Absorption Coefficients of CO, (Chr. Bohr and J. C. Bock). 


Temp. a Temp. a 
15 1-019 20 0-878 
16 0-985 21 0-854 
17 0-956 22 0-829 
18 0-928 37 0-567 
19 0-902 38 0-555 


Reduction to 0° C. and. 760 mm. Hg has been done in the usual way 
according to Boyle’s and Gay Lussac’s laws. 


Mass Action Law. 


We have now arrived at the point where it will be advisable to discuss the 
‘question of the chemical equilibrium which takes place between substances 
which react with one another in solution. 
When the mass action law is expressed as a formula and the apparent 
activity of the mol B is designated by az the following holds good: 
The reaction! bB + cC + dD= mM + nN + 00 proceeds from left to right 
with the rate 


er Be pak Sarr tptenN SHORE CON AE (13) 
and from right to left with the rate 


m n 0 

Ay - An -A9-kyy, eee ee eee eee ee ee ee eee eee (14) 
b ce d mm on o 

therefore Ap Ag. Ap. ky = Ay. Ay Ag hy vevesercoceecseaee (15) 

Pg a” ae 
MN § OO , 

and ta alas "i hrs abr (16) 

Gp Ag. 4p 


The expression for a reaction like the following (a binary dissociation) is 
especially simple, B = M + N, namely . 


* That is to say b mols of B react with ¢ mols of C and d mols of D, giving m mols of M +n mols 
of N +o mols of O. But the reaction can also be expressed thus—m mols of M react with x mols 
of N and o mols of O giving b mols of B+e mola of C +d mols of D. 


THEORY OF THE HENDERSON-HASSELBALCH EQUATION 159 


If we wish to express the law for the reacting (solvated) mols in terms of 
their concentrations we get 


Cu Fan: Cn Fan _ 
Cy-F ip f 


in which the activity of a mol is defined as the concentration multiplied by 
the apparent activity coefficient. If the apparent activity coefficient is 1, or if 


Fao Fim) _ | 
mae Ta ee at 3 


PB) 
Cu-Cy 
Cp 

When the reaction is in equilibrium and when it is dissociated so much 
that xCg:7r) mols of M and N are formed, (19) then becomes 


ed 
;_ 7. Car) = K, Peer ewer reer eeeeeerereseserese (20) 


l-2- 


(18) reduces to Me ik cell ckek G1 ibbokeed aaab cue’ (19) 


where Cg,7) is the concentration B would have if the reaction proceeded 
completely from right to left. 


Osmotic Pressure. 


In watery solutions which only contain non-electrolytes the “ideal” con- 
ditions for the osmotic pressure are very nearly fulfilled, particularly at higher 
temperatures, but here also certain corrections must be introduced for hydra-_ 
tion (see Findlay [1913]) which will be more pronounced if several solutes 
are present simultaneously in the solution some of which appropriate con- 
siderable quantities of the solvent. 

In solutions of electrolytes the conditions will be still more complicated 
when we consider the osmotic activity of the charged particles (ions). It has 
until recently been generally accepted, though not without hesitation, that 
the osmotic activity of the ions was the same function of their concentration 
as in the case of non-electrolytes, but latterly it has been appreciated that 
the electrical forces between the ions must diminish their osmotic activity. 

Milner (quoted from Bjerrum [1916, 1918, 1919]) has attempted to calcu- 
late the magnitude of this effect, which Bjerrum calls the Milner effect, without, 
according to Bjerrum, having actually succeeded in doing so, but as regards 
the questions dealt with here this is of no importance because we accept the 
empirical formula given by Bjerrum which will be discussed later on. But 
first we must recall some of the eeEy points in Arrhenius’ theory of 
electrolytic dissociation. 


Electrolytic Dissociation. 
In the following sketch of the fundamental conceptions of the theory of 


electrolytic dissociation aqueous solutions are exclusively referred to. 
Water dissociates to a slight extent into hydrogen and hydroxy] ions, 


Meg = (Ry ah ae a (21) 


160 E. J. WARBURG 


According to the law of mass action, 


ay X Gon)’ ee 
a Ke on's vcs teks niahd esta Dea (22) 
In weak solutions the following equation is very nearly true, 
44,0 Rae Ce Ey CET tae an et Me ar ete Re er (23) 
From (22) and (23) we get 
ay’ X Aon) = | oR Terererrerre errr roo (24) 


where K,, is of the order of 10-14. [See also Lewis, 1916, 1, p. 325, and 
8. P. L. Sérensen, 1912, 12, p. 400.] . 
When an electrolyte is dissolved in water it will be split up—dissociated— 
and positively and negatively charged ions will be formed. 
The ions carry one, two or more electrostatic units and the dissociation 
can therefore take place according to the following schemes: 


My Ay MY A ae ee (25) 
(electrolyte) *" (cation) (anion) 
M, M, Ay M5 + a ee (26) 
MM, M, Ay, OM + My Ae (27) 
or M, MM, Ans Mt MGA ee (28) 
M, My A Ay Ase (29) 
or My, Ay'Aype Mg A A (30) 
My A FEO ae ahi ae (31) 


and so on. 


M’; is a monovalent cation, M"*;; a divalent cation, etc.; A’; and A’’ yy 
respectively monovalent and divalent anions. There is reason to distinguish 
between three kinds of electrolytes. 

Electrolytes can be divided into: 


Alkali salts and similar salts 
Totally dissociated electrolytes { Strong acids 


» bases 
Dissociation varying {Weak acids 
Electrolytes the dissociation of which | with the reaction 5, bases 
varies with the concentration Ampholytes 


Complex salts 


The acids dissociate according to the following schemes: 
For a monovalent acid 


HAS a Als tcisetsnsonpaneae (32) 

and for a divalent acid 
H H Ay, — H’ + H A’; (partial dissociation), ............ (33) 
FF A EE A yi ies hacemssiptsdniericsantiooeen koe (34) 


THEORY OF THE HENDERSON-HASSELBALCH EQUATION 161 


The bases dissociate as follows: 
PR FE OT) 5 eocscctcescnendicscevasiaveee (35) 
My, (OH) (OH) — M™;; + (OH)’ + (OHY’. ......... (36) 
For an ampholyte which can split off one hydrogen ion and one hydroxyl 
ion per molecule the following hold: 
for the hydrogen ion dissociation 


Peter ee EI COMR) fice teases tess. (37) 
for the hydroxyl] ion dissociation 
BOM NOR ich ieiiccit eke: (38) 
and for the simultaneous dissociation of H” and (OH)’ 
BO) oo (OT) FTE oi avicescenseseescn. (39) 


Although the dissociation of weak acids and bases, and ampholytes follows 
the mass action law, difficulties are encountered with the strong acids and 
bases which compel us to believe that they are completely dissociated (¢.e. 
the reaction progresses completely from left to right), as will now be more 
fully explained in speaking of salts. 


“Strong” Electrolytes. 


The question which has caused most trouble in connection with the 
electrolytic dissociation theory is whether the “strong” electrolytes follow 
the “mass action law.” 

We will not discuss further the extremely wide development the subject 
has undergone; in all the larger textbooks of physical chemistry it is dealt 
with, and in Arrhenius’ [1912] Theories of Solution a whole chapter is re- 
served for it to which the reader is referred. Suffice it to say that in whatever 
way the dissociation is estimated—whether by depression of the freezing point 
or by conductivity methods, the mass action law is not fulfilled. In using 
the method involving the depression of the freezing point which is propor- 
tional to the osmotic pressure it was assumed the depression caused by ions 
and non-electrolytes was the same function of their concentration. As an 
example of such an estimation the dissociation (25) 

M, A; si M’, + A’; 
may be considered. The solution has a freezing point depression which is 7+ 
times greater than a non-electrolyte of the concentration of My; A, (if the 
reaction proceeded completely from right to left). If we regard the freezing 
point depression as proportional to the osmolar concentration without correc- 
tions for the Milner effect, the following equations will apply: 


_ (Cy, 4, (7) 18 the concentration of the salt when the reaction proceeds com- 
pletely from right to left.) 
Cap = (= 1) Cur ayer) = Cay verter teens (40) 
and Bie RR MY Ogg nyt) boistas nis. Siedisi ebb (41) 


1 <4 is van ’t Hoft’s coefficient. 


162 E. J. WARBURG 


substituting in (19) we get 

(¢-1) Pa 

Q=5 . Cy, 4, (T) art K. See ee Eee ee eee eee eee (42) 

In calculations made from conductivity experiments it was assumed the 
velocity of the ions was independent of the other ions present in the solution. 
It was thought therefore that 
the equivalent conductivity at the concentration Cy,4,(7T) wu, 
the equivalent conductivity at infinite dilution a ace fi ws 


was equal to the value z in (20) and it was subsequently substituted in the 
equation, but corrections for altered viscosity were often introduced. 

As it appeared to be impossible to obtain any concordance between the . 
experiments and the requirements of the mass action law—whether the dis- 
sociation was estimated by the methods described or in any other way— 
either the assumptions which formed the basis of the calculation were 
erroneous, or it must be concluded the strong electrolytes do not follow the 
mass action law, a conclusion which under these conditions would be de- 
structive to the law. 

It has for many years been known that the dissociation determined by 
conductivity methods (see for example Jahn [1900]) was not quite accurate 
while that determined by depression of the freezing point has as a rule been 
regarded with greater confidence, at any rate by physiologists, although there 
has been some suspicion about it also. 

The problem of elucidating the nature of the complications mentioned 
seems first to have been mastered by the Dane, Niels Bjerrum [1909, 1916, 
1918, 1919]. 


Dissociation Theory of Bjerrum. 


Bjerrum’s theory can aptly be called the theory of the complete dissociation 
of the strong electrolytes. 

From a number of experiments he carried out in 1906 on the chromium 
salts Bjerrum came to the conclusion the strong electrolytes were completely 
dissociated. It was found that they possessed the colour of the ions both in 
weak and in strong solutions if complex combinations were not formed 
(Bjerrum)!. The fact that the colour of a strong electrolyte was independent 
of its concentration led him in 1909 to formulate the hypothesis that the 
strong electrolytes must be taken to be completely dissociated into their ions 
in solution provided complex compounds were not produced. 

In 1919 Bjerrum developed this theory further and showed that a large 
number of the difficulties of the electrolytic dissociation theory disappeared . 
on his supposition. 

It can thus be said that Bjerrum showed in some very characteristic cases 


1 I shall follow extremely closely Bjerrum’s exposition which to a large extent is quoted 
word for word. 


THEORY OF THE HENDERSON-HASSELBALCH EQUATION 163 


that it was not necessary to attribute any catalytic effect to the undissociated 
acid (in systems where hydrogen ions were the catalyst), which was the case 
when accounting for the dissociation determined by conductivity. 

Bjerrum and Gjaldbaek [1919] succeeded in getting agreement between 
the calculated and experimental reaction constants for calcium carbonate 
and their investigations will be briefly discussed later in this work. 

They also succeeded in calculating the potential for the concentration 
cell Hg/HgCl, 0-1n KCl/0-01ln KCl, HgCl/Hg; after the elimination of the 
diffusion potential the potential was 


e = 0-0548 volt. 
Reckoned with the Nernst formula 
€ = 0-059] log? volt, .......c.ccsccessrcessesee (43) 


1 
by actual concentration « = 0-0591 volt, 


(corrected) by the degree of conductivity 

e = 0-0569 volt, 
(corrected) by the activity coefficient 

e = 0-0553 volt. 


The following is a translation of a part of Bjerrum’s paper in the reports 
of the Nobel Institute, those sections only being included which are necessary 
for the understanding of the theory, as it will be indispensable for those 
wishing to work at the subject to consult the original. 


“Tn the year 1887 Arrhenius enounced his famous hypothesis according 
to which the ions in a solution of electrolytes are present in the solution in 
the free state. During the 26 years passed since then this hypothesis has 
become an indispensable part of physics and chemistry, and even to-day it 
acts as a fruitful working hypothesis thus showing the powerful richness and 
force of the original conception. In the course of time, however, several 
difficulties have attended this hypothesis especially in explaining the 
properties of strong electrolytes. But after the work of later years these 
difficulties appear to vanish when considering the effects that must be produced 
on the properties of the solutions of electrolytes by the electric forces between 
the ions. The greatness of these effects may be indicated by means of coeffi- 
cients expressing the relation between the real value of the property in question 
and the value the property should have held supposing the electric forces did 
not act between the ions. Thus the activity-coefficient f, indicates the effect of 
the interionic forces on the activity of the ions; the conductivity-coefficient f, 
indicates the influence of the interionic forces on the conductivity; the osmotic 
coefficient f, indicates the influence of the interionic forces on the osmotic 
pressure etc. 

If we suppose that the strong electrolytes are completely dissociated into 
ions we may determine the value of the osmotic coefficient by means of 
freezing-point determinations. From Noyes’s and Falk’s excellent exposition 


164 E. J. WARBURG 


it thus results that the osmotic coefficient of all electrolytes with monovalent 
ions may, with fair approximation, be rendered by the formula 


Lb fg = BAC): ck sncdixs Speed auoeeu sees (44) 
in which k varies between 0-146 and 0-225 the mean being 0-17. 

This formula, no doubt, does not agree with Milner’s values, but within 
the limits of concentration where both experiments and Milner’s calculation 
are fairly reliable, viz. 0-01m—0-1m, the agreement is a rather fair one. This 
fact must be looked on as a support of the assumption that the strong electrolytes 
are completely dissociated or, at any rate, are much more dissociated than 
originally assumed. 

When the formula of the osmotic coefficient is known, the formula of the 
activity-coefficient may be deduced by means of the following thermodynamic 
relation: 


fot oot = Up se aterer a (45) 
If this be combined with (44) the following expression is obtained: | 
Inf = — 4 WO sa aaaisccioesdices fa ensnaee (46) 


When trying to determine experimentally the value of the activity-co- 
efficient f, for an ion, various difficulties are thrown in one’s way. 

First, it is difficult to make sure that the electrolyte is really completely 
ionised or, if that is not the case, to determine its degree of dissociation. If 
the molar conductivity of the electrolyte increases with decreasing concen- 
tration according to Ostwald-Walden’s valence rule it will however be natural, 
as I have developed formerly, to presume complete dissociation. In many 
cases this point may be investigated by means of spectro-photometric measure- 
ments the colour of the ions being, to a rather high degree, independent of the 
interionic forces and, on the other hand, greatly changing when the ions combine 
chemically together. 

Secondly, it is not possible, even if the ion concentration is known, to 
state with certainty what the activity of the ion would have been if the inter- 
ionic forces did not exist. This difficulty particularly manifests itself in 
concentrated solutions. As the measure of the activity one should, most 
likely, take neither the mol number per litre nor the mol number per 1000 g. 
of water but what is known as the mol fraction, 


where n is the mol number of the solute and »’ the mol number of the solvent. 
We shall designate this mol fraction as z-concentration. As to the justification 
for putting the activity equal to the #-concentration we may for instance refer 
to the excellent exposition of ideal solutions in Alex. Findlay’s book: The 
osmotic pressure. 

Thirdly, the water content of the ions in the solution must be considered. 
When the ions form hydrates in solution the consequence is that the «#-con- 


' Table II gives the corresponding values of the constant factors in equations (44) and (46). 


Table I, 
1-fq Infa 
Ne 2-3026 A/c 
0-146 025 
O17 0:30 


THEORY OF THE HENDERSON-HASSELBALCH EQUATION 165 


centrations of the ions will be smaller than they would have been without the 
hydration. The effect of hydration is well known and frequently discussed. 
There is, however, another effect of hydration which has not as yet been 
taken into account; although it is rather of greater importance than the 
former effect. If we want to determine the activity of an ion by means of 
potential measurements, say with metal or hydrogen electrodes or with 
mercurous chloride electrodes, we obtain for a hydrated ion by means of the 
ordinary Nernst formula not a measure of the activity of the ion itself but only 
of the activity of the ion without water. If the ion contains m H,O we should to 
obtain the measure of the activity of the ion-hydrate multiply the activity of 
the ion without water, which we may call A, by the activity of the water to the 
mth power. We have a good measure of the activity of the water in the vapour 
pressure of the solution p divided by the vapour pressure of pure water pp. 
So the activity of the actual ion may be put at 


a=A (2). (Sn ie WAR Dae a a (48) 


Strange to say no account has till now been taken of this correction in spite 
of the fact that in the electrometric determination of the hydrogen-ion- 
concentration, so frequently used of late, it is of considerable importance, 
the hydrogen-ion being very strongly hydrated. The measurement of the 
potential of a hydrogen-electrode in a solution does not directly give us either 
the hydrogen-ion-concentration or the hydrogen-ion-activity, but only a calcu- 
lative quantity which may be designated as the activity of the dehydrated 
hydrogen-ion. It is only for want of due consideration of this fact that the 
conclusion that the hydrogen-ion-activity increases by plentiful addition of 
neutral salts to hydrochloric and other strong acids has been arrived at 
(comp. for instance Harned, Journ. Amer. Chem. Soc. 37, 2460 (1915), 38, 
1986 (1916)). This correction should also be made in the case of activity 
determinations founded on the solubility of salts. 

Fourthly, the exact determination of ion-activities in concentrated solu- 
tions is rendered difficult by the association of the water, which is of im- 
portance in calculating the z-concentration, and which is not properly known.” 


By making calculations based on the above Bjerrum has succeeded in 
estimating the amount of hydrate water of the ions. 

F, is the apparent activity coefficient and f, the true activity coefficient. 

— log F, is the apparent activity exponent. 

In Fig. 1 which is reproduced here Bjerrum shows the relation existing 
between F, and ¢ for different degrees of hydration of the ions. 

While it is the hydrogen ion concentration, as Bjerrum [1916, 1918] has 
shown for the catalytic action of hydrogen ions (esterification of organic acids), 
which is the determining factor, it is the apparent hydrogen ion activity 
which is the important factor in the mass action law relating to those chemical 
actions in which the hydrogen ions take part. The apparent hydrogen ion 
activity can be determined by the electrical method in the same way as the 
“hydrogen ion concentration” has hitherto been estimated. 

In measuring the potential difference between a hydrogen-platinum elec- 


1 It may be noted that Bjerrum is using the symbols a and A with a significance differing 
from that which we have accepted in these articles. 


166 E. J. WARBURG 


trode and a solution, and a constant electrode we have the following form of 


Nernst’s equation, 
EB -E, 
Pr =) log ay = 0-0577 +0-0002 (¢-18°) PIP eeeseccecccsive se (49) 


E being the measured potential and Hy, a constant dependent upon the 
electrode used for comparison. If a 0-1n KCl—HgCl electrode (0-1n calomel 
electrode) is used, 8. P. L. Sérensen on the basis of conductivity experiments 


at 18° obtained E, = 03377 volt (SGresisen). |<. 35 haccpcs vesekeees (50) 


1-4 T 


Fe 0 S ae: 
1°2 

10, 

fs pcs 


0-6 — a 
3H,0 


0-4 Se re Soh 


02 | mr 


a | 
0 05 1°0 15 0 2°5 3-0 
Fig. 1. (After Bjerrum.) 


Bjerrum and Gjaldbaek [1919] from calculations based on the activity coeffi- 


cient obtained E, = 0-348 (Bjerrum) .....sessseccseesesenees (51) 
from which EF, (Sérensen) — 0-0029 = Hy (Bjerrum)'. ............ (52) 


Both results are valid for a hydrogen pressure of 760 — fmm, Hg where f 
is the pressure of saturated aqueous vapour at the temperature in question. 
If the hydrogen pressure is different, e.g. Py,, the following value (e) must be 


added to E, 


__ 0-0001983 poe Pu, 


—3z — (278 + ¢) log a ay EEE (53) 
where ¢° is the temperature on the centigrade scale, Py, is the Hy pressure 
in the electrode, and f is the pressure of saturated aqueous vapour at 0°. 

* At 38° 2, (Bjerrum) =0-3313. 


THEORY OF THE HENDERSON-HASSELBALCH EQUATION 167 


It will be observed that Bjerrum’s Ey is 2-9 millivolts less. than Sérensen’s. 
The corresponding difference in py: from 5-00 to 8-00 at room temperature 
or body temperature can with sufficient accuracy be expressed by 


pu’ (Bjerrum) = py (S6rensen) + 0-048. ............... (54) 


It is thus easy to transfer the values from one scale to another. 
As will be seen from Bjerrum’s reasoning, 


RU rN a ed asSoeesj soectassevavesin (55) 


In weak solutions, TP (LON) oe tO Big (LOD) io voce cnseseancenvennss (56) 


and the concentration of an ion can be calculated when its activity is known. 
For hydrogen ions according to Bjerrum [1916, 1918] the following equa- 
tion holds, 
eat Te (BEY eS A/G essitcies inne sacsenecnees (57) 
For the bicarbonate ion & is not known but it will be shown later that & for 


sodium bicarbonate is about 0-46. 
For an n-valent ion, according to Bjerrum and Gjaldbaek [1919, p. 79], 


ME WU WUE AIG Sec des iota ssessibiivisseecue sos (58) 


where k has the same value as for a monovalent ion in the same solution, and 
where it varies within the same limits as in the case of monovalent ions. 
Table III, according to Bjerrum, gives the values of the deviation coefficients 
for KCl. 5 


Table IIT. 
f, a 
Molec. cone. to i. ster 
0-001 0-985 0-943 0-979 
0-01 0-963 0-882 0-941 
0-1 0-932 0-762 0-861 
1-0 0-854 0-552 0-755 


Since fg (KCl) = Vf, (Cl’) f, (K’). 


It is a matter of fact that all salts do not follow the same law of dis- 
sociation. Thus while the salts of the alkali metals and presumably also the 
majority of those of the alkaline earths are completely dissociated (Bjerrum 
and Gjaldbaek [1919], Gjaldbaek [1921]), this is not the case for many of the 
salts of the heavy metals. The latter form complex compounds, that is to 
say they are only partially dissociated—often only very slightly dissociated. 
The equilibrium between the ions and the undissociated salts is determined 
in a similar manner to that between hydrogen or hydroxyl ions and the 
electrolytes the dissociation of which varies with the reaction, and the remarks 
anent these can almost be directly applied to the complex salts (electrolytes 
the dissociation of which varies with the concentration) by substituting metal 
ions for hydrogen ions. 

As it is by no means a matter of course that a salt is completely dissociated 
it will be one of the objects of the present work to investigate whether protein 
salts are complex or not. The working hypothesis that alkali albuminates and 


168 E. J. WARBURG 


protein salts with Cl’ and HCO’; are completely dissociated has been adopted 
throughout, and it has been investigated whether any facts following from this 
contradicted the hypothesis. 


Complications arising from Heterogeneity. 


The following equation holds for a heterogeneous solution, 

Cs, (100 — Cs, Di . 
Se bee ee (59) 
where @ is the volume of the internal phase expressed as a percentage of the 
total volume of the solution; Cg, is the true concentration of the solute in 
the external phase, the possibility of an aggregation being neglected, and Cg, 
is the concentration in the inner phase; D is a proportionality factor deter- 
mined by Os, 


and Cy, is the concentration of the solute in the whole system provided it is 
evenly distributed over the whole volume. We shall call Cg, the mean con- 
centration. 

In biological chemistry only scant attention has been paid to the con- 
ditions expressed by (59) but there are a number of investigators who have 
been cognizant of them. Thus Hamburger [1902] was the first to recognise 
the importance of the volume of the disperse phase in experiments on the 
osmotic conditions of the blood corpuscles. 

Rona [1910, 1913, 1, 2] and his collaborators also corrected for the 
volume of the disperse phase in dialysis of serum but they did not discuss 
any more than Hamburger the equation of the solubility of the solute in the 
continuous phase. 

8. P. L. Sérensen in his monumental work Studies on Proteins was the 
first thoroughly to appreciate these conditions and the author of this essay 
willingly admits he has been greatly influenced by the above work. 

After the appearance of Studies on Proteins Rich. Ege [1920, 1, 2, 3, 4; 
1921, 1,2, 3, 4] published an excellent piece of work on the partition of glucose 
between the blood cells and serum in which he pays the proper attention to 
the characteristic qualities of colloid solutions just mentioned. 

In Henderson’s and Hasselbalch’s equations the heterogeneity of the solu- 
tion is not allowed for and it was this that originally prompted me to investi- 
gate the theory of these equations. 

T. R. Parsons in 1917 called attention to the difficulties caused by the 
different distribution of bicarbonate between the blood corpuscles and serum. 
L. 8. Fridericia [1920] as well as Joffe and Poulton [1920] have recently 
determined the conditions of distribution but none of these authors regarded 
the serum and the fluids of the blood corpuscles as heterogeneous systems 
separate from one another. The work of all these authors is extremely valuable 
in relation to the questions before us and it will be fully dealt with later in 
this work. 


THEORY OF THE HENDERSON-HASSELBALCH EQUATION 169 


CHAPTER II 


THE HYDRATION OF CARBON DIOXIDE AND THE DISSOCIATION 
OF CARBONIC ACID. 


The Derwation of Henderson's and Hasselbalch’s Equations and the Modifica- 
tion of these in accordance with Bjerrum’s Theory. The Limits within which the 
Derived Equations are Valid. 


In 1 ce. of water in equilibrium with a gas mixture in which the tension 
of CO, is Poo,mm. Hg, there will be dissolved according to Henry’s law 
(equation (1)), 


= 5008 ec. CO, (at 0° and 760 mm. Hg). 
Some of the dissolved carbon dioxide is hydrated, forming carbonic acid, 
Eerie Seat ee AEs arapicesscndiececiovanestei (61) 
and the mass action law gives for this reaction 


ACO: “H20 Le 
Cee agains c cobev cae kanturs (62) 
Since a@,9 can be regarded as constant under the experimental conditions, 
ed 8 ae eee I est lege ae a (63) 
and therefore from (7) and (11), 
Ay,co, = Poo, Ky 9-00005910.  ........-. ee eee es (64) 


This equation in accordance with what was said on p. 157 applies to watery 
solutions also, with the proviso that Ky is the same for pure water and for 
watery solutions. How far this is the case cannot be estimated at present, 
but as from (62) and (63) 


and we regard ay, as constant, Ky will only vary from one solution to another 
if ky varies. 

Such a variation under the conditions investigated is improbable, but if 
it should happen it would be without importance for the following argument 
and would be included in the total constants. Only in the investigation of 
the extent to which the activity of the bicarbonate ion conforms to Bjerrum’s 
[1916, 1918, 1919] equation for the calculation of the activity coefficient, 
would such a condition be of real importance and invalidate the proof. 
Assuming however that Bjerrum’s theory is correct the above-mentioned 
investigation can be used as a proof that K, does not vary appreciably with 
the ionic concentration within the limits of concentration investigated. 

In equation (63) 

AH2COs __ 
Ati © 
@y,co, 18 the apparent activity of carbonic acid, and in so far as it is not 
Bioch. xvi 12 


170 E. J. WARBURG 


hydrated (i.e. it is not H,CO;, H,O for example), we have ay,cy, = Aun,co,- 
The error involved by putting 
H2COs CH2COs 
Aco: ©coz 66) 

will in all cases be small, as the activity coefficient of non-electrolytes is an ex- 
pression of the salting out effect which ought to be about the same for all non- 
electrolytes. Ky indicates therefore how much of the dissolved CO, is hydrated. 
Experiment has shown that Ky is small. Thus Thiel and Strohecker [1914] 
have calculated from the rapidity with which CO, neutralises bases that only 
4 % of the dissolved CO, is hydrated and therefore our Ky is roughly 0-005. 

Walker and Cormack [1900] make some remarks in their experiments on 
conductivity in CO, solutions about the value of Ky. They show that such 
solutions of carbonic acid fulfil the requirements of the mass action law when 
the total amount of dissolved CO, is substituted for the carbonic acid in the 
equation. If only a small part of the dissolved CO, was hydrated the mass 
action law would not be satisfied in the form employed and they think there- 
fore they are right in concluding about half the dissolved CO, is hydrated 
(which means that our Ky should be 0-5). Walker and Cormack’s determina- 
tions however certainly do not justify them in drawing these conclusions. 

Carbonic acid dissociates in the following manner: 


00. 2. H + HCO’, (1st dissociation), ...... (67) 
(Carbonic acid) (Hydrogen ion) (Bicarbonate ion) 


HOE? = H “+ CO”, (2nd dissociation). ...... (68) 
(Bicarbonate ion) (Hydrogen ion) (Monocarbonate ion) 
From the mass action law 


ee eee eee eee ee ee) 


“HBO's = k, (1st dissociation of carbonic acid), ...... (69) 
“5 oO" = K, (2nd dissociation of carbonic acid). ...... (70) 
Multiplying equations (63) and (69) we get 
Ary: &@ vad 

ay Poet = kK, eee ee eee eee eee eee eee eee ee 2 (71) 

and therefore “Bp HOO's ek CURT TT ee reyes (72) 
Since, almost without exception before the appearance of Bjerrum’s theory, 

er PEROT» KAMA Ls RN OE: (73) 

40, a Coo,» Cee emcees eee e Deere ee eeeseevereeeebeseoece (74) 

Ayco’, = Cuoo’,¥ = (Croo’,S.)s ssseeeeseensens (75) 


where y was the “conductivity dissociation” of the bicarbonate in the con- 
centration under investigation, (72) became 


On CHOO’sY _ 
Oco. — K,, TEPER EUEEERTE ETE EEE (76) 
which was transformed into 
gee K, Coos 
Cy = y Oncota’ Tt titteseeeeeeesseens (77) 


This is Henderson’s equation. 


THEORY OF THE HENDERSON-HASSELBALCH EQUATION 171 


Equations like (76) and (77) have repeatedly been used in physical 
chemistry, for instance in W. Nernst’s experiments on the dissociation con- 
stants of weak monovalent acids (cited from Datta and Dhar [1915]), in 
Datta and Dhar’s [1915] determinations of the second dissociation constant 
of weak divalent acids, and in investigations on the second dissociation 
constant of carbonic acid by McCoy and Smith [1911] and McCoy and Test 
[1911], Auerbach and Pick [1912], and Seyler and Lloyd [1917, 1, 2], and lastly 
in work upon the solubility of the carbonates of the alkaline earths, for 
example, by Johnston [1916]. 

An equation analogous to (72) is one derived by Bjerrum and Gjaldbaek 
[1919, p. 82, equation (24) ]. 

In logarithmic form (77) becomes 


Pu = pK, + log Cyco’, — log Coo, MOE Ves 2 escucnncwnes (78) 

As will be shown later in considering the limits between which an equa- 
tion similar to (77) is valid, Cyco,, in the cases treated by Hasselbalch and 
Henderson (as always in these investigations) can be assumed to have origi- 
nated from the combined carbonic acid, as the amount of HCO’, derived from 
the dissociation of the dissolved carbonic acid is negligible in comparison 
with it. If we know the total quantity of CO, in a solution and its pressure 
in a gaseous phase in equilibrium with the solution, the temperature and the 
relative absorption coefficient of the solution, it is easy to calculate the 
quantity of combined carbonic acid by subtracting the amount dissolved 
from the total amount. If now it is assumed all the combined carbonic acid is 
in the form of bicarbonate, and the CO, is expressed in volumes per cent., (77) 
compared with (5) becomes 


Cy: — Ei Vol. % dissolved CO, 0-0004492 (79) 
H “ y vol. % combined CO, 0-0004492’? "TTT 
os 
icici dciseicimesss. (80) 


7 val % combined CO, 
This is Henderson’s equation in a simplified form, expressed in the same way 
as that of Hasselbalch. 

Hasselbalch [1916, 2] regarded carbonic acid as a divalent acid and put the 
concentrations in (77) in terms of normality, assuming a solution which con- 
tained a gram-molecule of CO, in a litre to be twice normal as regards carbonic 
acid, while he regarded a solution which contained a gram-molecule of CO, 
bound as bicarbonate in a litre, as normal with respect to bicarbonate!. 

Equation (77) under these circumstances becomes 

__ K, vol. % dissolved CO, . 


. ae 2 
Cae are meeeA CG, De oct teneeeteceeen (81) 


And since Hasselbalch substitutes K, for = his equation becomes 


o3 vol. % dissolved CO, 
Cy — Ky vol. % combined co, 2. eee eee eee 


1 Parsons [1919] and L. Michaelis [1920] have recently pointed out that Hasselbalch uses 
this unusual convention and Michaelis has protested against it. 


12—2 


172 E. J. WARBURG 


If we now compare (80) with (82) it will be seen that 


= Menor) me DK, (Hasselbaleh). och aw eves (83) 


Hasselbalch’s equation in logarithmic form becomes 


Pu’ = Px, + log vol. % combined CO, — log vol. % dissolved CO, — 0-3010, 
and when (83) is put in logarithmic form we get 


pk, (Henderson) + log y = pK, (Hasselbalch) — 0°3010....... (84) 


Although the above appears to be correct it should be noted that Hassel- 
balch determined his total constant by the potentiometrical method. If this is 
recognised (83) becomes 

K, (Henderson) _ 9, (Hasselbalch) (85) 
F, (HCO’,) T (SEASSOLDAICH) ones ct sos sean enas 
and (84) is transformed accordingly, but at the same time a correction ought 
to be introduced into all Hasselbalch’s calculations so that px, always be- 
comes 0-048 times larger (given by the difference between Sdrensen’s and 
Bjerrum’s £,). 

If we abide by what is the usual custom in the literature of physical 
chemistry there is no doubt we ought to use molar concentration and not 
normality in such equations, and as a similar convention also has priority in 
physiological literature, Hasselbalch’s mode of expression should be given up. 

As the assumptions underlying (73), (74) and (75) are no longer tenable 
the equation should be modified in accordance with Bjerrum’s requirements. 
When the mols in the equation with the exception of ay: are expressed by 
their concentrations, (72) becomes 

ap CHCO’s Pg (HCO’s) Bi 


feo FAG) oe Pg. joes cehanbpeanicnaaiontey (86) — 
It will be seen from (54) that 
1-117 age = Cy (Sdrensen), 00.0... sccsenersees (87) 


and we thus obtain 


Cy: (Sérensen) CyCo’s __ Fa(CO,) 1-117 
CCO2 > Kk, F, (HCO’,) See eee (88) 


Substituting Oe a ee (89) 
we get from (88) and (85) 


, __ 1-793 Ky (Hasselbalch) 
K tor i. (CO,) ee Seer ee eee eee eee eee ee ere (90) 


and from (83) and (90) 


, 0-896 K, (Henderson) 
K', = vf, (GO,) a Paka Leva een Cok cee (91) 
If the relative absorption coefficient for blood is taken as 0-92 (Bohr [1905}), 
F,, (CO,) becomes 1-087 and we get for blood 


Blood K’, = 1-65 K; (Hasselbalch), — ....c.s.ssssoeeessens (92) 


pK’, = pK, (Hasselbalch) — 0-218. ns. sss esas (93) 


Now that we have developed the relations between the various constants 


THEORY OF THE HENDERSON-HASSELBALCH EQUATION 173 


we will revert to (72) and express it in the same way as Hasselbalch’s equation 


by using (89) ne l. % dissolved CO. 
| age = Ky By (COs) Fe cerimeg Con vee eeeseentiese (94) 
, Pco.« 
or ny = K * 7-60 vol. % combined co, © Pewee eereseereececere (95) 


Putting (94) in logarithmic form we get 


Pu (Bjerrum) = pK’, + log vol. % combined CO, — log vol. % dissolved CO, 
— log F, (CO,). ......... (96) 


We have now worked out the relations between Henderson’s, Hasselbalch’s 
and our own equations as expressed in (90) and (91), but a considerable qualifi- 
cation of their significance must be made. 

Henderson’s and Hasselbalch’s equations were evolved, as will be remem- 
bered, on the assumption that the degree of dissociation measured by con- 
ductivity gave the degree of activity of the bicarbonate, but since this is 
incorrect as Bjerrum later has shown, (73), (74), (75), (76), (77), (78), (79), 
(80), (81), (82), (83), (84), and (91) are only approximations to the true 
equations and have only mathematical significance, while (88) and (90) are 
correct actually as well as mathematically. 

It is perhaps also worth pointing out that (90) only holds good if Ky, 
(Hasselbalch) and K’, are calculated from the same experiments or from 
experiments which give the same results with the same method of calculation, 
and that Sérensen’s Hy (50) should be used for the calculation of Ay; (Hassel- 
balch), while #, (Bjerrum) (51) should be used for the calculation of K’;. 
If (90) is not satisfied by a correct calculation from two series of experiments 
it is proof that the experimental results do not agree. 

In the equations evolved up to the present we have assumed without 
hesitation that all the combined carbonic acid is present as bicarbonate. We 
will now inquire how far the equations are valid if the second dissociation of 
carbonic acid (68) takes place to any considerable extent. We will first find out, 
however, the reaction at which this happens. 

: Ay: Aco” 
Equation (70) a = K, 


: Oco’s _ Kz Fa(HCO’,) 
can be transformed into Tacos Gar FCO") © eee (97) 


The value of the right side of the equation is still not accurately known in 
spite of a great deal of work on the subject, since McCoy’s [1911], Auerbach and 
Pick’s [1912], Shield’s [1893], and Seyler and Lloyd’s [1917, 1] results must be 
recalculated according to the new ideas. Bjerrum and Gjaldbaek [1919] give 
K, = 10-1°”2 from which we may expect to get a good approximation by 
putting — K, Fa(HOO’,) 

F,(CO”s) 
It should however be noted that the value may vary somewhat with the ionic 
concentration. 


i es Br 5 ee (98) 


174 


Pr 


E. J. WARBURG 


If with this value ao is calculated from (97) the table below is obtained. 


Table LV. 
OC0"* vocm temp. CCO’’s room temp. 
ay pu CHCO’s ay PH CHCO’s 

1x10"? 7-00 0-001 0-25 x 1078 8-60 0-040 
0-5 7:30 0-002 0-2 8-70 : 0-050 
0-3 7-52 0-003 0-17 8-77 0-059 
0-2 7-70 0-005 0-15 8-82 0-067 
0-15 7-82 0-007 0-13 8-89 0-077 
1 x 10-8 8-00 0-010 0-11 8-96 0-091 
0-5 8-30 0-020 1x10~-° 9-00 0-100 
0-3 8-52 0-033 


It will be seen from the table that at py: 7-00 the ratio is only ae at 
8-00, a and at 9-00 it is 6: It may be concluded from this that in 


reactions of physiological importance all the carbonate may be considered to 


be 


in the form of bicarbonate—a result which numerous workers have pre- 


viously arrived at in a similar way (e.g. L. Meyer [1857], Heidenhain and 


L. 
[19 


Meyer [1863], Zuntz [1868], Chr. Bohr [1909], van Slyke and Cullen 
17], Parsons [1919], Bayliss [1915] and others). We shall revert briefly to 


this subject in discussing the variations of volume of the blood corpuscles. 


are 


It will now be shown that with a little alteration the equations evolved 
valid even if the reaction is so alkaline that a large amount of monocar- 


bonate ions are present in the solution. Instead of the concentration, the 
corresponding amounts of CQ, are substituted expressed in volumes per cent., 
and the amount 


corresponding to Coo, is called So, 


ry) Cyco’, »? S,, 
2 Coo”, ” S.. 
(94) then becomes gst ae By FOO en ae (99) 
And (97) combined with (98) 
tg Bigs ss ciastasephaveane (100) 
Calling the total amount of combined CO,, B, we get 
8. See Be eRe (101) 
and from (100) and (101) 8, = cr ms pe BLT Mayet (102) 


When this is substituted in (99) we get a quadratic equation 


ay’ = 


SoK’sFa(CO,) ROASTS (CO,))®+45,K’,F, (00,2 Ky B 
2B +“ +B . 


The numerator under the square root sign is equal to 


(S,K’,F, (CO,) + 2K’, B)® — 4K’,? BY. 


4K’,* B® can be neglected as K’,® is very small, and as the negative sign is 
without physical significance, we get 


K’, F, (CO,) So 
B od | Bas Ie 


ay’ = {- K’,. ond eed.ove beseneeeeous (103) 


THEORY OF THE HENDERSON-HASSELBALCH EQUATION 175 


The equation (103) thus developed is valid generally (independent of hydro- 
lysis of the ions), so long as the concentration of the combined carbonic acid is 
large compared with the concentration of the dissolved CO,, and so long as all 
the combined carbonic acid is present as monocarbonate and bicarbonate. 

It will be noticed that it is identical with the earlier equation (94) (the 
modified Henderson-Hasselbalch equation) when K’, is small compared with 


3 vol. % dissolved CO, 
K Pf a (CO,) vol. % combined CO, ° 


In the following short table the first column contains values of 
, vol. % dissolved CO, _ 
KPa (CO,) vol. or combined CO, 
the second column contains the corresponding values of 
4 vol, % dissolved CO . 
K'P a (COs) yo, 0% combined Co, + K's» 
K’, being again 10-19; in the last column the corresponding py: values are 
given. 


Table V. 
K’ F Sy , S, ’ 
14 (CO,) S K’, F4(CO,) S, + K’, Pu 
10-8 1-010 x 10-8 7-996 
10-° 1-100 x 10-° 8-959 
10-% 2-000 x 107" 9-699 


It will be seen from the table that at reactions more acid than 10-8 the 
correction is negligible and the equation becomes identical with the modified 
Henderson-Hasselbalch equation. 

We have still to investigate how small the quantity of combined carbonic 
acid can be without being small in comparison with the amount of bicar- 
bonate arising from the dissolved CO,. 

The hydrogen ions in a solution containing dissolved CO, may either 

(1) exclusively be derived from the dissociated carbonic acid (except for 
the minimal amount due to the dissociation of water), or 

(2) partly come from the carbonic acid and partly from another acid, or 

(3) the reaction may be so acid that the carbonic acid is undissociated, 
in which case there can be no carbonate present in the solution and the 
equation does not apply. 

We will only consider the case when carbonic acid is the sole source of 
hydrogen ions, because the field is more restricted than in the other cases. 

Let us call the concentration of bicarbonate ions which is equivalent to 
the amount of available base, C,,, the bicarbonate ions which are derived 
from the dissociation of carbonic acid being, as already mentioned, equal to 
Cy. Equation (86) taken in conjunction with (89) and (85) then gives 


a a) RICCO), che eee (104) 
from which Cy = — a + af ons ae eye Ccieireasaauet (105) 


176 E. J. WARBURG 


in which the negative sign is without actual meaning. If C,, is equal to 0, (105) 
reduces to 


i = VK’, F, (CO,) ee (H’) Coo, - obesvecceseves (106) 
(106) applies to reactions in solutions of CO, where there is no combined car- 
bonic acid. 

A hard and fast rule for using (105) instead of the modified Henderson- 
Hasselbalch equation is difficult to give, but a good approximate rule may be 
arrived at in the following way. 

(104) is compared with (86) and (89) in the form 


Cc * Fa(H') Cm _ A ; 
Gry Bi Fa (C02). vssereerierrerreeen (107) 


The difference between (104) and (107) is due to the difference between 
the factors C,, and (Cy: + C,,). If Cy: is small in comparison with C,,, the 
equations become the same. Making the approximate assumption that Cy: 
is equal to ay: we will evaluate ae ay is calculated from Te which only 
differs from (105) in form, and an estimation of the value of @ -is made. If 
this does not exceed 0-01 the use of equation (94) is Kacenae 


Table VI. 
Vol. 4 
an: On combined CO, mm. Hg, CO, 
10-° 10~* 0-22 0-563 
10-5 10-* 2-24 53-2 
10~* 10? 22:39 4710 


In the table the values which C,,, must have at different reactions so that 
the above requirements for the applicability of (94) may be fulfilled, are given. 
In the third column C,, is expressed in volumes per cent. of CO, and in the 
fourth column the corresponding tension of CO, calculated with the help of 
the constants determined for sodium bicarbonate in Chapter VIII. 


Résumé. 


From the above results it may be said that in homogeneous solutions: 
I. Henderson’s equation is in agreement with the equation evolved on 


Fa(COs) is substituted 


the basis of Bjerrum’s new dissociation theory when K, 5 (HCO’,) 
@ 3 


for af in the former. 


II. Hasselbalch’s K, = K,** TnGO 


III. The value found then becomes ay: (Bjerrum), not Cy: (Sérensen), 
and instead of py: (Sérensen) we have py: (Bjerrum). 

IV. The limits for the modified Henderson-Hasselbalch equation are given. 

V. The complete equations for the relation between the hydrogen ion 
activity, the amount of dissolved CO, and the combined carbonic acid are 
evolved. 


THEORY OF THE HENDERSON-HASSELBALCH EQUATION 177 


CHAPTER III 


THE DEVELOPMENT OF A MODIFIED HENDERSON-HASSELBALCH EQUATION 
FOR USE IN THE CASE OF HETEROGENEOUS SOLUTIONS. 


As already hinted at several times the conditions in heterogeneous solu- 
tions are complicated by the fact that the concentration of the solute in the 
individual phases is not directly given by its mean concentration, as the 
solute cannot be assumed to be evenly distributed over the whole system. 
Further we cannot regard the activity coefficients in the single phases as 
already known, and lastly there is the possibility that the solute may be con- 
centrated by adsorption in the interfaces between the different phases. 

The influence of the first condition on the determination of reaction by 
Henderson’s and Hasselbalch’s equations has latterly been recognised by 
Fridericia [1920], Parsons [1917] and by Joffe and Poulton [1920], but they 
have not realised the full significance of it, only regarding serum-blood cells 
as a heterogeneous system, and failing to appreciate that serum (plasma) and 
the fluid of the blood cells are likewise heterogeneous systems. 

Fridericia’s and Parsons’ contributions appeared while the experimental 
part of this research was in progress, but Joffe and Poulton’s work only came 
to my notice after I had embarked upon the literary section. These investi- 
gations are of the greatest importance for the questions here considered and 
they will be fully dealt with in a later chapter of this work. 

We will proceed with our calculations wnder the assumption that all the 
combined carbonic acid is present as bicarbonate (that no CO, is bound by 
complex combinations or adsorbed), and we will investigate how the ex- 
perimental results accord with this assumption. 


Let - Vol. % combined carbonic acid = f, 


B being the mean concentration in the heterogeneous solution. Let us first 
consider a system of two watery phases in mutual equilibrium, and in equili- 
brium with a gaseous phase containing CO,. 

100 parts of the system consist of q parts of phase (2) and 100 — q parts 
of phase (1). 

From this we get 


‘vol. of phase (2) a 
100 x vol. of phase (1) + vol. of phase (2) = q: oes eecsevecesees (108) 
Kiko Vol. % combined CO, in phase 5 d (1 09) 


Vol. % combined CO, in phase (1) — -— ee eee eee eee eee eee 


Seeking now the amount of combined CO, in the single phases expressed by f, 
d and q, 


Viogd . Vay (100- 
us ri aw (Cw ie Fe aeek salsa ci saad eaae ks (110) 
100 : . 
Vay = CErers (vol. % combined CO, in phase (1)), ...(111) 
and Vig = eH (vol. % combined CO, in phase 2)). ...(112) 


178 E. J. WARBURG 


From (111) in conjunction with (95) we get 
, 100 -—g(1-d Poo.a 
an" (1) = K 1() cl x ioe Sivd wale sss eoueua ale (113) 
where ayy"(,) is the apparent hydrogen ion activity in phase (1) and K’,() is K, 
divided by the apparent activity coefficient of HCO’; in phase (1). 
Analogous to this we have 
. = Rt 100=q(k-d) PB 
Sern =H sca) angg tee ee (114) 
K’y(y and K’; (9; are not different from K’, in a homogeneous solution and the 
indices only signify that Ff, (HCO’,) refers to a particular solution. 
Comparing (95) with (113) and (114) it will be seen that the last two can 
be expressed in the same form as (95), viz. (113) in the form 


a a 
ana) = Xb Gere. Hovde qe tandee (115) 
and (114) in the form Oar tay = Aa GOES + nssvesscteansessoned «ses(116) 
< , 100 -—qg(1-d 
Since * Aa) =i K 1() en) Seen meee eee eereeeenens (117) 
, 100 -—q(l-d 
Xe) => K 1(2) — 20) a Terre rey ah eee ee (118) 


Ay) and A(z) occur in (115) and (116). These are only constants as long as the 
factors in them do not vary or the variations neutralise one another. 

RK’ yy and K’; (2) will only vary with the reaction if the activity coefficient 
of the bicarbonate ion varies with it. 

L. Michaelis [1920] has lately drawn attention to this condition, regarding 
it however from the classical standpoint of Arrhenius, and therefore assuming 
that K’, varies with the degree of dissociation of the bicarbonate. His views 
appear in the form of a criticism on the theory which H. Straub and Klothilde 
Meier [1918-1920] put forward for the carbonic acid combining power of the 
blood on the strength of a series of experiments carried out by them. Straub 
and Meier’s experiments and theories are however subject to an elementary 
error, and as will later be shown they are quite fallacious, but Michaelis’ 
comments naturally do not lose their interest on this account. 

According to Bjerrum’s! theory there is reason to expect that the activity 
coefficient of an ion will decrease in a solution containing ions with an opposite 
charge and this will be especially the case if these ions are polyvalent or have 
a large molecular volume. 

If there are already rather many ions present in the solution the effect 
of the new-comers will be small, and possibly in special cases it will not be 
noticed at all, but it is not right, without due consideration, to regard K’,(,) 
and K’,(. a8 constants, as it is a fact that the ionic concentrations in phases 
(1) and (2) vary with the reaction. 


The factor 1” = ~) in (117) and the corresponding factor in (118) 


' I have repeatedly had the opportunity of discussing this condition with Prof, Bjerrum 
about whieh he has up to the present only published a rough sketch. 


THEORY OF THE HENDERSON-HASSELBALCH EQUATION 179 


will vary provided that qg and d vary. From 8. P. L. Sérensen’s [1915-1917] 
experiments on egg albumin there is every reason to expect they will vary 
with the reaction, but what influence the variations will have on the “total 
constants” experimental investigations alone can inform us at present. 

As will be shown in one of the subsequent chapters 4,,) in serum and the 
liquid phase of the blood corpuscles is constant within the limits of the 
reactions investigated (experimental errors are not very small in experiments 
with the fluids of blood cells), and we will therefore continue our calculations 
under this assumption. 


Let @y* in the liquid phase of serum = ay"(s)> 
and ay in the liquid phase of blood cells = ay‘), 
Aq in serum =X», 


Ai) in blood cells = A,,). 


The volume of the blood cells expressed as a percentage of the total volume 
of the blood is Q. The partition of HCO’, between blood cells and serum is D, 


so that 
CHCO’s in blood cells D, 
Cyco’s im serum 


and lastly the mean concentration of combined carbonic acid in the blood 
expressed in volumes % = B, and 


~ vol. of blood 
From (115) and (116) we then get 
Bhetah Nay ee OOS, cvnenscrdensvanecs (119) 
PI, TO a “ou Fe a asncseapianaraenye (120) 
(119) and (120) can be put in a form analogous with (95), viz. 
We Nig Gag ae ssahoisisponesntgcesnic (121) 
Wer tah Wicah He Y ehesscdaiseassiacasseseeesnndse (122) 
Bay ins casi susescscevencissearts (123) 
Neo = Xe) 10-0.-D), An Eee ate, (124) 


It follows from the above that in A;,) and A,,) four variables take part, of 
quite a different kind from those which determine K’, in (95), and it will thus 
be understood that we cannot immediately draw conclusions regarding the 
combination of carbonic acid in the blood from a numerical agreement be- 
tween A,,) and K’, in any bicarbonate solution. 

Fridericia [1920], Henderson [1908-1910], Hasselbalch [1916, 2], Parsons 
[1917-1920], Joffe and Poulton [1920], Michaelis [1920] and many others 
have not realised the full consequences of this condition (which in an almost 


180 E. J. WARBURG 


analogous way can be developed by the classical dissociation theory), and we 
are therefore compelled to investigate anew the questions mentioned in the 
introduction: 


I. Can the hydrogen ion concentration of the blood be calculated by an 
equation similar to Henderson’s and Hasselbalch’s? 
II. Is all the combined carbonic acid of the blood ionised ? 


I have attacked the first problem from a purely empirical standpoint. For 
practical work it is convenient to employ (121) in logarithmic form 


Pa \s) = PAG) + log B — log 00x f ocauywaaheeekssws (125) 


and transform it thus 
Pa’ = PA + log B — log fg, © tsscveessecssees (126) 


where f, is the mean concentration of dissolved CO, in the blood expressed 
in volumes %, and 


PA’, = PA, — log De (COg), «vse crees tonthenenes (127) 


while ®, (CO,) is the reciprocal of the relative absorption coefficient in blood. 
The relative absorption coefficient in serum is 0-975 according to Bohr 
[1905] and 0-81 in blood corpuscles, from which ®, (CO,) can be calculated 
by the following equation: 
®, (CO,) = a Ri a Rept cE RA SS (128) 


_19Q 2:5 (100 -Q)" 
1p 100 100 


Résumés. 


I. The modified Henderson-Hasselbalch equation is adapted for use with 
heterogeneous solutions. 

II. It appears from the resulting equation that nothing can be concluded 
from the size of the total constant about the way the carbonic acid is combined, 
without taking the heterogeneity into account. ; 


CHAPTER IV 


METHODS FOR THE ELECTROMETRIC DETERMINATION OF THE 
APPARENT HYDROGEN ION ACTIVITY IN SOLUTIONS CONTAINING 
CO, WITH THE HELP OF THE PLATINUM HYDROGEN 
ELECTRODE, AND THE TECHNIQUE EMPLOYED. 


The general principles and methods of estimation by the potentiometrical 
method have been often described and are to be found in every handbook of 
physical chemistry. The method used hereis that describedin§. P. L. Sérensen’s 
Etudes enzymatiques, 11 [1909-1910] and those interested are referred to this 
work for further information, Within the limits of reaction obtaining in the 
following experiments the reaction of a solution will be a function of the CO, 
tension with which it is in equilibrium, What effect a change of the CO, tension 


THEORY OF THE HENDERSON-HASSELBALCH EQUATION 181 


will have on the apparent hydrogen ion activity cannot be expressed in a 
general way, because the buffer action of the solution will determine it. 

If it is desired to determine the reaction in a solution in equilibrium with 
a certain CO, tension, care must be taken that the CO, content does not vary 
during the measurement. There are two ways of avoiding this, either by 
saturating the solution with a hydrogen-CO, mixture (H6ber’s principle, 1903), 
or by bringing a very limited amount of hydrogen in equilibrium with the 
solution and taking care by suitable means that the CO, tension in the hydrogen 
only very slightly departs from that existing originally in the solution 
(Hasselbalch’s principle, 1910), the potential of a platinum plate in contact 
with the solution being then measured. 

For didactic reasons it is convenient to first describe Hasselbalch’s prin- 
ciple. It is a development of methods with “stationary hydrogen”’ first used 
by Farkas in 1903 and later modified in different ways, see for example 
Botazzi in Neuberger’s Die Harn. 

Farkas allowed the liquid, the reaction of which he wished to estimate, 
to come into equilibrium with a quantity of confined hydrogen by diffusion 
through its surface, and he then measured the potential of a platinum plate 
dipping into the liquid. K. A. Hasselbalch showed however in 1910 that 
equilibrium is not reached in the time occupied by the experiment by this 
method, and as it was impossible to increase the duration because changes 
would occur in the blood he modified the method in such a way that the 
hydrogen was quickly brought into equilibrium with the liquid by rocking 
the electrode vessel containing the hydrogen and liquid a certain number of 
times. The liquid was then replaced by a fresh supply and the process re- 
peated. Hasselbalch combined his method with the device of Michaelis and 
Rona [1909] which consisted in minimal immersion of the platinum in the 
liquid. Michaelis and Rona had demonstrated that by this refinement the 
depolarising action of the oxygen could be avoided or at any rate reduced 
to a minimum. 

Konikoff [1913], who also worked at the “oxygen error,” realised that the 
rocking must be continued until all the oxygen is reduced by the hydrogen 
(with the help of the platinised platinum). This is however hardly practicable 
in blood on account of the slow speed of the reduction (both electrode and blood 
become changed during the experiment). But there is no great difficulty in 
removing the free oxygen from inorganic salt solutions, (human) urine and 
similar liquids if a sufficiently large platinum plate is used, but a platinum 
wire of the size usually employed (Michaelis [1914], Hasselbalch [1916, 2] 
electrode at the bottom of Fig. 3) cannot remove the oxygen from solutions. 
Hober [1914], Peters [1914] and T. R. Parsons [1917] have also pointed out 
the possibility of depolarisation by O, in Hasselbalch’s measurements and 
they recommended estimations on Héber’s principle. 

It is easy to convince oneself of the truth of the above by measuring the 
reaction of a phosphate mixture by Hasselbalch’s method. In a vessel con- 


182 E. J. WARBURG 


taining a platinum plate electrode the liquid can be renewed several times 
without the potential practically varying at all, as Hasselbalch [1910 and 
1911] showed a long time ago. But if the estimation is carried out in a vessel 
containing a platinum wire as electrode the potential! will fall on renewal 
(in the electrode vessel illustrated in Fig. 3 the potential fell about 2 millivolts 
for each renewal of liquid). This is due to the fact that besides H, and CO, 
the gas becomes contaminated with O, which depolarises the H,—Pt electrode. 
No information is known to me in the literature about the activity of small 
quantities of O, in this connection. In measurements with an electrode vessel 
on Hdber’s principle, to be described later, I have obtained a rough idea of 
the magnitude of this factor. O, below 2mm. Hg does not give an error of 
1 millivolt, 3 mm. gives about 14 millivolt, and 8 mm. gives a large error if 
the potential otherwise remains constant. I can also confirm the frequent 
experience that no conclusion from the way in which the potential becomes 
established can be drawn about the presence of traces of O, in the electrode 
hydrogen. 

Until 1916 Hasselbalch employed a small platinum plate as an electrode, 
while Michaelis used a platinum wire. In 1916 Hasselbalch adopted a hook- 
shaped wire for blood estimations, the point of which was allowed to touch 
the side wall of the vessel just over the surface of the liquid, and in this way 
he obtained minimal immersion. 

According to experience at the Finsen Institute therefore wire electrodes 
should be used for liquids containing haemoglobin, and platinum plates for 
other liquids. In estimations of blood there are four points in particular to 
be noted: 

(1) Contact between platinum and liquid should be minimal. 

(2) The electrode must not be moved while waiting for the potential to 
become constant. 

(3) The three-way tap must not become fixed. 

(4) There must be no air bubbles in the connecting tube to the KCl 
solution. 

We will now quite shortly consider a few points of importance in making 
measurements by Hasselbalch’s method, neglecting the depolarising action of 
the oxygen, but remembering that if we did not do this the potential would 
be lower (and py smaller). We will assume also that the volume and pressure 
of the gas do not vary during the determination. 

If a limited quantity of liquid, saturated with a gas at a given tension, 
which forms no combinations that split off the gas under the conditions of 
experiment, e.g. bicarbonate, is allowed to come into equilibrium with a 
limited space occupied by a gas, it can easily be shown? that at a given tem- 


* By the potential is understood, here and henceforth when not otherwise stated, the total 
potential of the chain of concentration cells. 

* See for example D. D. van Slyke and Cullen's [1917] equation for the caleulation of the 
amounts of CO, in the apparatus of van Slyke. 


THEORY OF THE HENDERSON-HASSELBALCH EQUATION § 183 


perature the same proportion of the dissolved gas always goes over into the 
space, so that its partial pressure in the space will always be the same fraction 
of the pressure with which the liquid was originally in equilibrium. If the 
space is now allowed to come into equilibrium with a fresh amount of liquid, 
the partial pressure of the gas will rise further, but it will again be a definite 
fraction of the original pressure irrespective of its absolute magnitude. 

It will be seen from equation (95) that ay: is proportional to the CO, 
pressure, from which it follows that the error, occasioned by the CO, tension 
in Hasselbalch’s electrode being too low, will be independent of the absolute 
CO, tension and will only be dependent on the relation between the volume 
of liquid and gas and on the absorption coefficient of the investigated liquid 
for CO,. 

The statement of Hasselbalch [1910, 1911] that more renewals of the liquid 
should be made in the case of higher CO, tensions than with lower to obtain 
a constant potential, is due to the fact that a slight alteration of the H, 
tension has hardly any effect on the potential, while a large change greatly 
affects it (see equation (53)). 

Neither Hasselbalch, Michaelis nor any other person who has worked with 
the method, corrected for the decreased H, tension (the correction is usually 
small), but Hasselbalch came to the conclusion that the best results are 
obtained by omitting this correction. It follows from what has gone before 
that we ought to be very cautious about admitting measurements by Hassel- 
balch’s method as standard ones, but they can legitimately be used for com- 
parison with one another if they are done with the same electrode and the 
CO, tensions do not differ very much from one another. 

Similar considerations apply to blood, as the CO, split off does not greatly 
disturb the conditions but rather tends in favour of equilibrium being attained 
more easily than in bicarbonate solution. It appears that Hasselbalch and 
his collaborators found different values for py: at different times. This is 
perhaps most evident from the fact that Hasselbalch and also Michaelis until 
1916 held that blood with 40 mm. CO, had about 0-20 py: greater at 18° than 
at 38°, but in 1916 when he changed the electrode he found the same py: 
value at 18° as at 38°. 

It is also worth while recounting a few experiences we have had at the 
Finsen Institute with regard to measurements of blood in the Hasselbalch 
electrode, which have also been noticed in numerous other laboratories. When 
the change of potential is followed after rocking it will be noted that it rises 
to a maximum and then slowly falls a little (the destruction of the electrode 
by the proteins). We have always taken the maximum reading as the correct 
potential. 

When the electrode vessel is ready for an estimation the blood corpuscles 
sink, leaving the platinum in contact with serum alone. This is unquestionably 
of no small importance in the estimations, a view which is shared by many 
others. McClendon [1917] has even devised an electrode in which the blood 


184 : E. J. WARBURG 


corpuscles and serum are separated by centrifuging before the estimation is 
made; and T. R. Parsons [1917] has saturated the blood with a mixture of 
H, and CO, and then centrifuged at the temperature of saturation (38°), 
keeping the saturation constant, and he showed that the potential of the 
separated serum was the same as that of the blood. 

The other technical points are the same as those met with in measurements 
with Héber’s electrode, to which the reader is referred. 

Hober’s principle, as already mentioned, depends on first bringing the 
liquid into equilibrium with a mixture of hydrogen and CO, of known con- 
stitution, and it is tacitly assumed that the oxygen is expelled by the satura- 
tion. 

Numerous electrode cells have been constructed on this principle (e.g. by 
Bjerrum and Gjaldbaek [1919], Héber [1903], Peters [1914], McClendon 
[1917], Milroy [1917], Walpole [1913] and others). There can be no doubt that 
the measurements made on this principle should be corrected for diminished 
hydrogen tension, but there are no difficulties arising from the depolarising 
action of the oxygen (see however later measurements of the blood cell fluids). 


GENERAL TECHNIQUE. 


The gases used were stored in the ordinary commercial metal cylinders. 
The hydrogen was prepared electrolytically and the cylinders contained as a 
rule about 99 % H, and about 0-5 % O,. Once it was strongly contaminated 
with atmospheric air. 

The hydrogen was led through palladium-asbestos kept red-hot in an 
electric oven. On account of the war I had only about 2 g. of 25 % palladium- 
asbestos at my disposal. It was packed in a combustion tube between asbestos 
corks. With this device I could get rid of the oxygen in 1 litre of hydrogen 
in a minute. The hydrogen was first washed with 5 % sulphuric acid and 
then with concentrated aqueous sublimate solution. Hydrogen prepared in 
this way contained about 99-5 % H 

The gases were mixed in a spirometer. For the hydrogen mixtures a 
20-litre spirometer with water sealing was used, and care was taken that 
oxygen never gained access to it, a precaution it is absolutely necessary to 
observe. If the spirometer was not used for several days it had to be washed 
with O,-free air several times. In addition to this spirometer two others of 
125 litres were employed. In using a spirometer with water sealing it is im- 
portant to recognise that gas mixtures containing CO, slowly change because . 
CO, diffuses out through the water. The best way to avoid error is to take 
care that the CO, content of the gas mixture only differs slightly from the 
mixture that was previously in the spirometer, so that in investigating the 
effect of different tensions of CO, on a solution we start with the weakest or 
the strongest mixture and let the next nearest mixture follow on. 

I have convinced myself a large number of times that the same gas mixture 


THEORY OF THE HENDERSON-HASSELBALCH EQUATION 185 


is obtained from the spirometer (20 litres) if 9 litres are removed at the end 
of half-an-hour or 18 litres at the end of an hour. Only very occasionally 
have I found slight differences. In working with the large spirometers I have 
always taken the precaution to prepare 1001. of the mixture and only with- 
draw 60 |. for the experiment. In this way the same mixture will be obtained 
after the lapse of one hour as after the lapse of four hours. Occasionally 
however I have noticed a slight fall in the CO, tension. Mixtures which 
greatly differed from the one which was last in the spirometer have usually 
been prepared on the evening before they were to be used. The mixing in 
the spirometer took place simply by moving the bell up and down. With 
the 20-litre spirometer, 25 times was sufficient; with the large one, 50 times 
was necessary. The gas mixture was conducted from the spirometer through 
a wash bottle which often contained 0-9 % NaCl when I was working with 
blood and strong salt solutions, otherwise pure water. I have however not 
strictly adhered to this but no appreciable error can have arisen from this 
cause. Before reaching the wash bottle in the thermostat the gas passed 
through a lead tube 1-5 metres long and about 0-3 cm. diameter so that it 
should attain the correct teniperature. A Roux thermostat was employed 
(see Fig. 2). As the temperature in the upper part was often several degrees 
higher than in the lower, the perforated copper plates were replaced by wide- 
meshed galvanised iron netting and the netting covering the floor was raised 
somewhat. Under this a stirrer was installed which was driven by a small 
electric motor. 

The stirrer! is illustrated in detail in Fig. 5 (p. 193). Its springs were 
similar to those of a dentist’s drill (Claudius Ash and Co.), When it revolved 
at the proper speed and had the right inclination the temperature could be 
kept constant in the thermostat to within 0-5°. The conducting wires to the 
electrodes were led through the wall of the thermostat in paraffined glass 
tubes which were fixed in paraftined corks. 

For the sake of economy in space it was necessary to place the thermo- 
regulator somewhat lower. In addition to the stirrer, the wires and the 
thermo-regulator an axle pierced the wall of the upper part to rotate the 
saturator and another the wall of the lower part to rock the Hasselbalch 
electrode. Both axles were worked with the help of a small electric motor. 

The saturators used for small amounts of liquid were composed of 250 cc. 
bottles like those Hasselbalch [1916, 2] has used, but the inlet tube was not 
fitted with the piece of rubber tube which acts as a valve. In such an instru- 
ment 4-12 ce. of liquid were rotated about 60 times a minute for half-an-hour 
and about 1 litre of gas blown through every three minutes, at the same time. 
Under these conditions equilibrium was always established except in a very 
few cases with high CO, tension and very viscous liquids and when preliminary 


1 This and various other items of the apparatus used were constructed by the mechanic of 
the Finsen Institute, the late Mr Eriksen, after consultation with me and my best thanks are due 
to him. . 

Bioch, xvr 13 


186 E. J. WARBURG 


— p. 
b- 
LC 
ii: —— 
L 
—_——£ 
U AA 
#. 
( 
ig 
3 +-O. 


Vig. 2. The thermostat employed. In the upper department a small saturator electrode is 
rotating; in the lower, a similar one is y for the estimation. a, small saturator electrode 
vessel; b, metal holder for the same during rotation; c, hollow wooden block; d, thermo- 
meter; ¢, wash bottle; f, calomel electrode; g, stand for calomel electrode; h, Petri dish 
filled with sealing-wax; i, veasel containing 3-5n KCI solution; &, fan for ae the air in 
the thermostat; /, paraffined cork; m, platinum electrode; o, stand for Hasselbalch electrode; 
p, ground joint for tap, 


THEORY OF THE HENDERSON-HASSELBALCH EQUATION 187 


treatment with pure CO, had been undertaken, but the difference was only 
slight. The experiment below will show the maximum error in the first case. 
Preliminary treatment with CO, in the small saturator was only employed 
in experiments where absolute accuracy was unimportant for the conclusions 
drawn, while it should be remembered that some of the CO, combinations 
may be a few vols. per cent. too low with tensions above 350 mm. CQ,. 


Fig. 3. Above, a large saturator; below, the Hasselbalch electrode E VI used as a Héber electrode. 


The tap is in position III during saturation and in position I when the liquid is transferred 
from the saturator to the electrode. 


Ox blood containing 22-7 vols. % combined O, (concentrated by centri- 
fuging) was treated with pure CO,; half-an-hour after there were 178-8 vols. % 
total CO, in the blood; after a further treatment with CO, for 25 hours 183-4 
vols, % CO, were present. 

13—2 


188 E. J WARBURG 


For larger amounts of liquid a saturator of the form illustrated in Fig. 3 
was used. The capacity was a litre. A curved tube for extracting samples 
and a conducting tube for the gas were led through the cork in a similar 
manner to the small saturator. In a number of the experiments a stirrer was 
installed in the vessel, rotating with it. The frame of the stirrer was of heavily 
gilt brass and its blades of wood. It became evident however that the stirrer 
was superfluous. As equilibrium was always established in half-an-hour with 
70 ce. in the saturator without the stirrer I dispensed with it afterwards, but 
rotated for an hour at the same time blowing about a litre of gas through 
every three minutes, total amount about 20 litres. 

The saturator was rotated by being mounted in a holder of practically the 
same type as that illustrated in Fig. 2 (without wooden blocks), and fixed 
in it by means of straps. 

The electrodes used were on Hober’s principle and five kinds were 
employed. : 

I. Small saturator electrode!. This electrode which rather resembles a 
form described by McClendon [1917] was constructed as a development of 
the above saturators. The details of its structure will be understood from 
Fig. 2 in which an electrode is seen which is rotated in the upper part of the 
thermostat while gas is blown through it, and in the lower part an electrode 
ready for an estimation in a little wire basket fastened in a Petri dish with 
sealing-wax. 

The electrode consists of a vessel with a rounded end. Its capacity is 
about 250 cc. In the middle a small funnel is blown which is continued into 
a thin glass tube with a small glass tap having a single hole. On the end of 
the tube a piece of rubber tube is placed which partly serves for taking 
samples of the liquid and partly for making a connection with the KCl solu- 
tion. In the plane of the glass tube at right angles to the long axis of the 
vessel the outer part of a ground joint is blown in the side of the vessel and 
it is prolonged about a cm. through the wall. The axis of the ground joint 
and that of the glass tube form an angle of about 150° with one another and 
the axis of the ground joint cuts off a sector of the cross section (not looking 
towards the axis of the vessel). In the ground joint there is an inner part 
which is continued as a glass tube through the end of which a 0-5 mm. platinum 
wire passes. The wire carries a 0-5 x 1 cm. platinum plate of about 0-5 mm. 
thickness. Connection with the sealed platinum wire is made by means of 
mercury or by welding a copper wire on to it. The neck of the vessel also 
consists of a ground joint the internal half of which inside the vessel is pro- 
longed into a tube and outside bears a single bored tap. 

The electrode vessel in the upper portion is fixed in a hollow block of wood 
in which three 5 mm. strong rubber coated iron wires are fastened. The block 
of wood is held against an iron plate by means of straps which go round the 
neck of the vessel and fasten on to brass knobs attached to the iron plate. 


' The electrode vessels were made at Jacob’s technical glass works, Copenhagen. 


THEORY OF THE HENDERSON-HASSELBALCH EQUATION 


: 


” 


ee eee 
t 


“i 


“+ 
an at a, 
' 


——— 
. 
N 
‘ 
1 
ry 


4 oe ee 


224 led 

|e al ee 

ted Tee 
ty ct v, é 
rE 
ote 
4 
‘ 
ia 


fe 


Fig. 4. On the extreme left is the large saturator LII in its stand after the potential has been 
measured. Then comes the small receiver filled with mercury for taking samples of the liquid. 
Following this is seen an “Ente” and in connection with its tap there is an ordinary gas 


receiver. 


189 


190 E. J. WARBURG 


On the opposite side to the wooden block the iron plate carries a box which 
fits into a screw thread in the axis of the rotator. 

The ground joints are lubricated with vaseline. The exit taps are not 
greased. In the electrode vessel 10 cc. of liquid are placed and rotated for 
half-an-hour while blowing through 8-10 litres of the H,-CO, mixture. 
Samples of the liquid are taken while the introduction of the gas mixture 
continues, in a little mercury filled receiver which is seen in Fig. 4 (p. 189). 
The exit tap is then closed and afterwards the inlet tap, and the outer part 
of the ground joint is put in its place in the neck. The connecting tube is 
removed and the electrode placed ready for the estimation in the lower portion 
of the thermostat, the tube dipping into a vessel of 3-5n KCl and a connection 
made with the copper wire by a binding screw. 

In constructing such an electrode again it would be an Sitges to 
replace the small straight-bored exit tap by a tap with tail boring, as the 
conductivity through it is better. 

In cases where the carbonic acid binding capacity of the ,Tiquid was known 
no samples were taken, and only 3-5 cc. were therefore introduced into the 
electrode vessel. In working with this saturator electrode and the saturator 
described above, samples of the gas blown through were taken immediately 
after the saturation was completed. 

Now that the small saturator electrode has been described in detail it will 
only be necessary to give a short account of the other electrodes employed. 

II. Large Saturator electrodes. (L I and L ILI.) 

The electrode vessel was, on the whole, of the same type as the previously 
described smaller electrode but it was much longer and had a capacity of 
1050 cc. The exit tap was an ungreased 3-way tap with tail boring (the 
straight boring being filled with paraffin). The inlet tap was a 3-way tap 
similar to the one fitted to the evacuation receiver (see Fig. 5, p. 193, upper 
row of taps) and where the neck of the vessel (the ground joint) joined the 
body a thin tube was sealed in at an angle of 45° with the axis of the vessel. 
This tube which was used for taking samples of the liquid had a small greased 
single-bore tap. The electrode vessel was mounted in a similar stand to that 
shown in Fig. 4, which carries the electrode vessel shown there. This stand 
had a box with a screw clamp and could be fixed to a rotating axle. 

In measuring haemoglobin solutions (haemolysed blood) the following 
procedure was adopted. 20 cc. of the liquid was put in the electrode vessel 
and rotation started. Then either a strong stream of CO, was sent through 
it for a quarter to half-an-hour, or this preliminary treatment was dispensed 
with and 17 to 18 litres of H,-CO, mixture passed through the electrode 
for half-an-hour. Then the ground joint was put in position and the inlet 
tap closed, and the electrode was rotated for half-an-hour with closed tap. 
A similar gas mixture was again passed through the electrode while rotating 
and finally the electrode was again rotated for half-an-hour with closed tap. 

The stand was then removed from the rotating axle and hung up by silk 


THEORY OF THE HENDERSON-HASSELBALCH EQUATION 191 


threads, so much liquid being sucked out that the exit tube was filled, and then 
the rubber tube to the tap was filled with KCl solution making a connection 
for the measurement in the usual way (the potential was not affected beyond 
the experimental error (1 millivolt) by further rotation). After the estimation 
was completed the electrode was putin such a position that its long axis formed 
an angle of about 45° with the horizon and the inlet tap was lowest. The 
latter was now put in connection with a receiver for taking samples of gas 
and its two inlet tubes filled from the receiver with mercury. Through the 
small glass tube a sample of the liquid was taken; and then a sample of gas 
for analysis, through the inlet tap. In this way there was a control on the 
equilibrium by comparing the analysis of the gas mixture used for the satura- 
tion with that in the electrode. 

The same electrode vessel (LI) was used for an investigation of the liquid 
in equilibrium with low tensions of CO,. The electrode was filled with pure H, 
and then the H, + CO, saturated liquid was introduced while H, was blown 
through the vessel, and the ground joint was then placed in position and it was 
rotated for one hour. As it seemed desirable not to allow contact between the 
haemoglobin solution and the platinum plate to take place before the measure- 
ment was actually begun, the electrode vessel shown in Fig. 4 (L IT) was con- 
structed. Its capacity was about 1050 cc. and it consisted of a glass tube with 
a ground joint in each end. The inner portion of the small ground joint carried 
the platinum electrodes; on the prolongation of the outer portion a side 
tube was joined with an ungreased 3-way tap with tail boring, with the 
help of which communication with the KCl solution and the removal of 
samples of the liquid could take place. A tube for taking samples of gas 
was sealed in the wall of the vessel and prolonged almost to the opposite side. 
This tube had a straight-bored tap. By means of the contrivance which is 
to be seen on the drawing the tube and the 3-way tap were rinsed with 
about 10 cc. gas from the electrode before the sample was admitted to the 
receiver. 

It is useful to let the platinum wire in the electrode project 2 cm. on 
either side of the point where it is sealed into the glass. The other details 
will be appreciated with sufficient clearness from the diagrams. 

The technique of saturation was the same as with the electrode L I. 

III. Electrode vessels of the Hasselbalch type were likewise used but with 
H,-CO, mixtures. One was the electrode shown by Hasselbalch and Gammel- 
toft [1915] and was called E III. It was only used for salt solutions. A little 
' liquid was introduced and an H,—CO, mixture bubbled through. Connection 
with the KCl solution was made by filling the 3-way tap and exit tube with 
KCl solution. The gas mixture was passed through until the potential became 
constant, the ground joint was turned and the 3-way tap adjusted in such a 
way that the bore of the tap was in line with that of the electrode vessel after 
which the potential was measured again. . 

The other electrode vessel of the Hasselbalch type called E VI was that 


192 : E. J. WARBURG 


illustrated in Fig. 3. It will be seen that there isa CaCl, tube with a ground joint 
from which the handle is removed, a platinum wire through the top of. the 
joint being substituted. The wire is coiled with a single turn in the interior 
of the ground joint and is bent in the form of a hook against its wall. The 
electrode vessel is washed out before the estimation with the H,-CO, gas 
mixture and the gas is then led through the saturator as shown in Fig. 3. 
As soon as saturation is complete the stream of gas is turned off and the gas 
is conducted, by turning a T-tap to the position marked in the illustration, 
directly out into the saturator. The electrode vessel is then filled with liquid 
to the side opening in the ground joint avoiding the introduction of air into 
it and the measurement made in the usual way. 

In using all these electrodes it must be remembered to wash out the tube 
and taps with the gas mixture in order to displace any air present. 

As connecting liquid 3-5n KCl was used. 

As standard electrode a n/10 KCl calomel electrode was prepared according 
to Ostwald and Luther’s directions. For a large number of the estimations 
a calomel] electrode and KCI solution was used which with long standing had 
become too concentrated giving an error of 2 millivolts, but this has been 
allowed for. Later four new electrodes were made which agreed very well 
with one another and numerous measurements of 8, P. L. Sérensen’s [1909- 
1910] phosphate solutions gave results which did not deviate 1 millivolt from 
Sérensen’s own value. It is however probable that the stated value of py: 
is 0-008 too low which corresponds to an error of $ millivolt. ‘There is possibly 
thus a systematic error running through all the results but I have not corrected 
for it as it falls within the errors due to the method of regulating the tem- 
perature which I employed. A Weston cell was used as a standard cell and 
was tested twice during the investigations. 

The gas analysis was carried out partly with the analysis apparatus of 
Haldane modified by A. Krogh and partly with a Petterson, Bohr, Tobiesen 
apparatus. CO, was absorbed in 50% KOH, and O, in strongly alkaline 
pyrogallic acid solution. The maximum accuracy with Haldane’s apparatus 
was 0-01 %; with Petterson’s 0-02 %,. 

The gas content of the liquid was determined by exhausting with a Tépler- 
Hagen-Bohr mercury gas pump and analysing in Petterson’s apparatus. 

The liquids were collected in the little receiver shown in Fig, 4, the 3-way 
tap being rinsed with the liquid from which the sample was taken by suction, 
and the volume of the liquid being determined by weighing the mercury that 
had run out, 

The evacuation receiver had a 3-way tap, like those shown in the upper- 
most row in Fig. 5. The sample was transferred to it by filling the dead space 
with distilled water and turning the taps as shown in the diagram (from left 
to right). It was necessary, as indicated, to have the tap closed with water 
and mercury during the pumping. 

The extraction of gas from blood and amino-acids took place with a 


THEORY OF THE HENDERSON-HASSELBALCH EQUATION § 193 


saturated solution of boric acid in the receiver while in the case of inorganic 
salts 5 °% sulphuric acid was used as a rule. The receiver was usually heated 
during the process in a water bath at 38°-45°. 


Fig. 5. Below is the small (liquid) receiver in connection with the inlet tap of the evacuation 
receiver. In the first position (from the left) the dead s is filled with distilled water, 
in the fourth position the liquid is led into the pump and in the last position the tap of the 
evacuation receiver is cl and the small receiver can be removed and pumping begun. 
Above is the stirrrer for the thermostat. 


CHAPTER V 


THE DIRECT EMPIRICAL DETERMINATION OF A,,. IN THE MODIFIED 
HENDERSON-HASSELBALCH EQUATION IN ITS APPLICATION TO BLOOD. 


The values obtained in the present chapter are calculated from experi- 
ments in which the apparent hydrogen ion activity (H,—Pt electrode) and 
the combined carbonic acid were estimated on the same sample of blood, 
and in which the CO, tension with which the blood was in equilibrium was 
known. Inthe literature similar estimations by K. A. Hasselbalch [1916, 2], T. R. 
Parsons [1917] and J. F. Donegan and T. R. Parsons [1919] will be found. 

- Hasselbalch’s measurements were made with a small hook-shaped wire 
electrode. The volume of the hydrogen was about half that of the liquid in 
the electrode vessel used. Hasselbalch reckoned his constants from these ex- 
periments but did not correct for depression of hydrogen pressure. The con- 
stants have been recalculated by means of the equations given in chapter ITI, 
and the values obtained by correcting for decreased hydrogen pressure are given 
on the assumption that the CO, tension in the electrode vessel was the same 
as that in the saturation mixture and that CO, and H, alone were present in 
the electrode vessel. This proviso as we have seen in a previous chapter is 


194 E. J. WARBURG 


only approximately true as the liquid is usually not renewed in Hasselbalch’s 
measurements in this electrode but it will be observed from the table that 
the correction is only small. Parsons’ and Donegan and Parsons’ measure- 
ments were done on Hober’s principle with Walpole’s electrode vessel and are 
already corrected for diminished H, tension. 

There is reason to believe that the electrical measurements of Parsons and 
his collaborators are the best in the literature because they replatinised their 
electrodes each day and only let them come in contact with the blood when 
the measurement was about to commence, saturation being accomplished in 
a special saturator. Parsons made use of a water thermostat and possibly 
obtained a more constant temperature than we did at the Finsen Institute. 
But it seems likely that his estimations of combined CO, are a little uncertain. 
This assertion is supported by an examination of his experiments by plotting 
them as curves, and the view is also shared by E. Jarlév [1919]. The electrical 
measurements recorded in Table VIII were carried out on serum which was 
centrifuged from blood saturated with CO,, at the saturation temperature, 
saturation being maintained throughout. In experiments with reduced blood 
the saturation was carried out with H,-CO, mixture and the serum was 
resaturated with the same mixture. In experiments with oxygenated blood 
saturation with CO,-air mixture was performed and the serum separated by 
the centrifuge was resaturated with H,—CO, mixture with the same content 
of CO, as in the mixture used for the blood. The measurements given in 
Table IX were made directly on H,—-CO, saturated blood. All Parsons’ 
figures for combined CQ, refer to blood. 

I have myself performed a number of similar experiments most of which 
were carried out with the small saturation electrode, but some with low CO, 
tension with the large saturation electrode LI. The platinum electrodes were 
platinised before each measurement, which was found to be necessary (cf. 
Hasselbalch and Gammeltoft [1915]) otherwise the potential was lower than 
with freshly platinised ones. It was found unnecessary after a few experiments 
to heat the electrode to redness before platinising, but later on measuring 
strong haemoglobin solutions this was found to be essential, for which reason 
it would be advisable in future to avoid all trouble by always heating the 
electrode to redness first. 

The potential always became constant a few minutes after the electrode 
was ready for an estimation (except at 38° when the air in the thermostat got 
cool on account of the unavoidable manipulations connected with the taking 
of samples etc.), and this was taken as one of the indications that everything 
was in order, 

The relative absorption coefficient in Hasselbalch’s, Parsons’, Donegan 
and Parsons’ and the author’s blood experiments was always taken as 0-92, 
while in measurements of serum-rich and serum-poor mixtures of blood. I 
have used 0-95 and 0-855 respectively, 

When not otherwise stated the blood used in the present experiments 


THEORY OF THE HENDERSON-HASSELBALCH EQUATION 195 


came from the cattle market and was not used before the day after the animal 
was slaughtered (so that “Ist day” signifies the first day after slaughter). 
All the experiments with blood at room temperature are marked series A 
(and also a series with serum). They were all carried out in the autumn of 
1918 and I have reason to believe all the potentials were too low without 
however being able definitely to state the cause. The mutual agreement 
within a series of experiments was however sufficiently good and therefore I 
have considered the experiments should be published; the course of the curve 
for the combination of carbonic acid is also of some interest. 

The experiments marked B are subject to no such uncertainty. 

Certain points about the value of A;,) can be surmised in advance. From 
the very considerable experience accumulated in different quarters on the 
reaction of the blood (serum) with various CO, tensions it can be concluded that 
the activity of the bicarbonate ion in serum cannot be greatly different from 
its mean concentration in the corresponding blood, from which it follows that 
Ais) must be of the same order of magnitude as K,. Hasselbalch considered 
that Ky, (Hasselbalch) varied with the mean concentration of bicarbonate in 
the same way as it varied in sodium bicarbonate solutions while Milroy [1917], 
Parsons [1919] and Michaelis [1920] on theoretical grounds thought the value 
ought not to vary in blood. I can support the opinion of these authors in so 
far as I have grounds for believing A,,) does not vary but one cannot predict 
to what extent the variations of the volume of the blood corpuscles (Q in 
the equation (123)) react upon A;,) or whether the partition of the bicarbonate 
ion (D) between serum and blood corpuscles is variable. These conditions 
will be thoroughly investigated in a subsequent chapter. Haggard and 
Y. Henderson in 1919 calculated this constant from a number of Hasselbalch’s 
and Parsons’ experiments. The results differ somewhat from those about to 
be given chiefly because Haggard and Henderson have dealt with “hydrogen 
ion concentration” while I have dealt with “apparent activity” (Bjerrum), but 
furthermore these two authors have not realised that a number of Hassel- 
balch’s experiments which they used in their calculations are really not valid 
for this purpose because the “hydrogen ion concentration” was not determined 
electrically, but reckoned from Henderson and Hasselbalch’s equation. 

In calculating the mean figures of the experiments I have felt justified 
in neglecting the values in brackets. This applies to the experiments of Parsons 
[1917] and Donegan and Parsons [1919] which refer to very low tensions of 
CO, because a small error in the determination of the combined CO, will 
produce a large error in pA’;,) in these cases. Further I have excluded all the 
exepriments of Donegan and Parsons where pA’;,) is under 6:10, since as a 
matter of experience too low values are much commoner than too high ones ° 
with the electrode cells employed, and the low values will therefore not be 
balanced in the mean value without the adjustment alluded to. 

The correction for the temperature in my experiments at room tem- 
perature is effected in such a way that 0-005 was subtracted from pA‘) at the 


196 E. J. WARBURG 


temperature of the experiment for each degree below 18° and added for each 
degree above 18°. 
The mean of Hasselbalch’s 24 determinations at 38° was 
pA'(s) = 6-139 + 0-00391. 
From Parsons’ and Donegan and Parsons’ 48 determinations at 37° 
pA'(s) = 6-178 + 0-0039. 
From my own 15 determinations at 38° 
pA'(s) = 6-147 + 0-0079. 


Parsons [1917] and Donegan and Parsons [1919] have remarked that there 
is a difference between their py: values and those calculated according to 
Hasselbalch’s method. They write: 

“Tt will be seen that the general direction of the curve (py—Pco, curve) 
is the same but that Hasselbalch’s blood appears to be more acid than mine 
at each CO, pressure. The reason for this may be to a certain extent in an 
individual variation, and without more data of this kind for the blood of a 
number of individuals the extent to which this factor operates must remain 
undecided. But the divergence may possibly be partly explained by differences 
in our experimental procedure. It is a significant fact that practically all 
errors which are likely to occur in electrometric determinations on blood (with 
the exception of loss of CO,, which is out of the question in the experiments 
here described) tend to produce a reduction in the value of the E.M.F. with 
a consequent shift of the results to the acid side. The particular points in 
which Hasselbalch’s procedure differs from ours is that he runs oxygenated 
blood into the electrode vessel, and so his results are liable to be affected y 
an error due to depolarisation of the electrode.” 

Parsons has repeatedly emphasised that blood is a heterogeneous system 
and he also realises that Hasselbalch’s constant is a total constant, but in 
the article written in conjunction with Donegan they draw some conclusions 
which will not bear criticism. They write: 

“There is no doubt of the applicability of Hasselbalch’s calculations in 
homogeneous systems containing CO, and sodium bicarbonate in equilibrium. 
But while aqueous solutions of sodium bicarbonate and also blood plasma? 
represent such systems, the same is not true of whole blood. Here we have a 
two-phase system, composed of corpuscles and plasma in which the differences 
of composition of the two phases lead to marked differences of their combining 
powers for CO,. Therefore the relation between CO,-tension and CO,-content 
in the whole blood will be different from that in plasma, Now it has been 
shown previously that it is the relation between free and combined CQ, in 
the plasma,-which determines the value obtained in an electrometric deter- 
mination of py of the whole blood, while it is from the different relation 


' The “mean error” of the mean determination. 
* It follows from what was said in chapter IIT, that Tam unable to support Donegan and 
Parsons on this point, 


THEORY OF THE HENDERSON-HASSELBALCH EQUATION | 197 


between free and combined CO, in the whole blood we shall obtain the caleu- 
lated value of its py:. In other words the difference between the observed 
and the calculated py:’s of blood is simply an expression of the difference in 
behaviour of corpuscles and plasma towards CQ,.” 

Although I am in agreement with Parsons [1917] that the oxygen error 
may have caused Hasselbalch’s measurements to be too low, he overlooks 
the fact that the CO, tension has not been properly appreciated in Hassel- 
balch’s electrode (no renewal) which will militate against the O, error so that 
it is impossible to predict which factor will be paramount. 

Donegan and Parsons also overlook the fact that Hasselbalch’s constant 
was determined by measurements on whole blood so that the constant contains 
a correction for the different bicarbonate contents of serum and blood cor- 
puscles, and that the calculated and measured py: must be the same if 
Hasselbalch’s and Parsons’ techniques give concordant results. 

That the pA’;,) value of Parsons and Donegan and Parsons is 0-023! greater 
than mine can only be due to lack of uniformity in the technique of the 
electrical measurements and as they are all carried out on the Héber principle 
- it seems to me that provided the difference is a real one there can only be two 
possibilities : 

(1) either our calomel electrodes are different, 

(2) or my lower value depends upon my not having glowed my platinum 

electrode before each platinising. 


The difference between our constants is really very small, although it 
seems to be real judging from the mean error, and it would mean a consider- 
able number of new experiments to determine what the true value ought to be. 

If the tables are examined more carefully it will appear that pA‘;,) from 
Donegan and Parsons’ measurements seems to increase greatly with reactions 
more alkaline than py: 8-0. This is however doubtful and my own measure- 
ments with a marked alkaline reaction though few in number do not show this. 

On the face of it there appears to be no certain variation of pA‘, caused 
by change of reaction but a more careful investigation of the question might 
be useful. 

If we divide Parsons’ and Donegan and Parsons’ measurements in such 
a way that all the experiments in which pg: was equal to or over 7-46 are 
placed in one group, and all those in which py: was equal to or below 7-45 
are placed in a second group, we get 21 experiments 


Py = 746 pA, = 6-188 + 0-0063 
with a standard deviation of 0-029; and 27 experiments 
745 = py pA'c — 6-170 + 0-0049 
with a standard deviation of 0-023. The difference between the constants is 
0-018 + 0-008. 
1 Corrected for the influence of the temperature. 


198 E. J. WARBURG 


It seems from the above as if pA’;,) increases a little with the py value. 
The standard deviation in the two series is of the same order so that we are 
justified in assuming there is a real deviation of the “constant,” and the 
results group themselves about the mean in the same way in the two groups 
so that we can*compare the mean values. 

Dealing with my own experiments at 38° in a similar way, we have 8 ex- 
periments pa = 7:39 pA’ = 6-161 + 0-0099 
with a standard deviation of 0-028; and 9 experiments 

7:38 5 po pA) = 6-133 + 0-012 
with a standard deviation of 0-036. The difference between the constants is 
0-028 + 0-0155; pA’;,) seems therefore to increase a little with the py-. It is 
clear however from the calculations that the change in pA‘, is small in all cases 
and a large number of experiments would still be necessary to accurately 
determine its magnitude. 

At 38° the following is the mean of Parsons’ [1917], Donegan and Parsons’ 
[1919] and my own measurements: 

pA‘) = 6-159, from which pA) = 6-193 


and A = 6-41 x 1077. 
At 18° I found pA‘, = 6-24, from which pA(,) = 6-27 
and Avy = 54 x 10-%. 

Résums, 


I. The constant A;,) has been calculated at body temperature and room 
temperature from all the suitable experiments to be found in the literature. 

II. A considerable number of new experiments for the determination of 
the constant have been performed. 

III. A small discrepancy has been shown to exist between the constant 
reckoned from Parsons’ and Donegan and Parsons’ experiments and my own. 
and the possible causes of the disagreement are discussed. 

IV. It has been shown to be probable that the constant increases a little 
with increasing hydrogen ion activity. 


THEORY OF THE HENDERSON-HASSELBALCH EQUATION 199 


Table VII. Calculated after K. A. Hasselbalch, Biochem. Zeitsch. 78, p. 123. 


Corrected 
mm. Hg Vols. % co, a 
CO, (combined) Pus) pA'(s) Pu: pA(sy 
38°. Ox blood I 39-3 42-6 7-31 6-11 7-32 6-12 
28-1 37-2 7-43 6-14 7-44 6-15 
63-1 47-9 7-19 6-14 7-21 6-15 
95-6 62-7 7-09 6-10 7-12 6-13 
38°. Ox blood IT 10-7 31-6 7:77 6-13 7-77 6-13 
33-6 47-8 7-47 6-15 7-48 6-16 
96-7 65-0 711 6-11 7-14 6-14 
61-2 58-7 7-27 6-10 7-29 6-12 
20-1 39-8 7-62 6-15 7-63 6-16 
43-7 51:8 7-36 6-11 7:37 6-12 
74-0 64-7 7-22 611 7:24 6-13 
38°. Ox blood IIT 41-1 48-8 7:37 6-12 7:38 6-13 
41-0 48-5 7-37 6-13 7-38 6-14 
38°. Ox blood V 42-7 44:3 7-31 6-13 7:32 6-14 
39-3 44-5 7-33 6-10 7-34 611 
40-9 42-2 7:29 6-12 7:30 6-13 
38°. Ox blood VI 36-3 45-2 7:39 6-12 7-40 6-13 
54-4 52-0 7-32 6-17 7-34 6-19 
Mean 6-14 
38°. Human blood (K.A.H.) 50-8 57-1 7-31 6-09 7-33 6-11 
45-7 52-9 7-37 6-13 7-39 6-15 
32-7 45-1 7-43 6-12 7-44 6-14 
18-5 35-1 7-60 6-15 7-61 6-16 
22-4 36-7 7-54 6-15 7-55 6-16 
80°7 61:3 717 6-12 7-20 6-15 
Mean 6-145 
18°. Human blood (K.A.H.) 44-8 54:3 7°23 6-20 7-24 6-21 
22:3 44:3 7-47 6-22 7-48 6-23 
Mean 6-22 
Table VIII. T. R. Parsons, J. Physiol. 51, p. 448. 
Vols. % CO, 
(combined) Pu pa “(s) 37° 
Per meneame ———— 
mm. Hg Oxygenated Reduced Oxygenated Reduced Oxygenated Reduced 
No. . lood blood blood blood blood blood 
l 37-4 43-4 47-3 7:37 7-45 6-13 6-18 
2 0-79 — 12-4 wa 8-55 =P (6-19) 
3 19-6 34-4 39-9 7-62 7-69 6-20 6-22 
4 72-1 59-9 64-4 7-25 7-28 6-17 6-17 
5 10-1 30-0 31-5 7-79 7-86 6-16 6-21 
6 8-1 _ 27-4 —_ 7-91 _ 6-22 
7 55:3 — 55-1 = 7:35 — 6-19 
8 33-4 41:5 47-6 7-48 7-53 6-13 6-21 
9 5:7 21:3 23-8 7-96 8-01 6-22 6-22 
Mean 6-17 6-20 
-- 


6-19 


200 E. J. WARBURG 
Table [X. Calculated after Donegan and Parsons, J. Physiol. 52, pp. 317-318. 


mm. Hg Vols. % CO, mm. Hg Vols. % CO, 
co, (combined) py: pA‘(s) 37° CO, (combined) py pA‘(s) 37° 
0-7 4-5 8-29 (6:32) 1-0 12-4 8-28 (6-01) 
9-7 33-1 7-78 (6-08) 21-3 38-5 7-58 6-16 
29-6 44-3 7-51 6-17 40-5 46-8 7:37 6-14 
44-0 53-1 7:39 6-14 60-0 50-8 7-26. 6-17 
65-7 60-7 7:30 6-17 2-0 8-9 8-15 (6-34) 
1:3 6-4 8-26 (6-30) 14-2 29:3 - 7-67 6-19 
12-7 30-5 7-75 6-20 30:8 38-9 7-43 6-17 
37-0 45-0 7-45 6-20 45-9 45-0 7:34 6-18 
67-6 52-6 7-24 6-18 2-0 8-5 8-23 (6-32) 
21 8-4 8-26 (6-29) 19-1 33-5 7-61 6-20 
18-8 36-3 7-67 6-22 ~ 40°8 44-6 7:39 6-19 
31-5 48-2 7-49 6-14 60-7 54-9 7:27 6-15 
2-2 10-4 8-30 (6-36) 10-9 26-0 7-74 6-20 
- 18-1 39-4 7-68 6-16 29-3 37-8 7-47 6-20 
38-9 43-8 7-44 6-22 45:3 46-2 7-32 6-20 
59-3 53-9 7°30 6-18 0-7 8-8 8-02 (6°75) 
13 7-7 8-19 (6-25) 11-6 27-4 7-58 (6-03) 
20-0 35°5 7-59 6-18 31:3 37-0 7:37 6-13 
42-7 45-1 7-36 6-17 46-1 41-3 7-27 6-15 
65-9 53-0 7-24 6-17 12:5 23-0 7-61 6-18 
1-3 10-0 8-27 (6-22) 25-6 30°3 7-42 6-18 
15-9 35°6 7-66 6-15 40-9 36-9 7-28 6-16 
38-2 46-9 7-41 6-16 eer 
B15 507 7-31 6-14 Moan 6-18 


Table X. Determination of pA’; in blood at 38° with the small saturator 
electrode, Series B. 
mm. Hg Vols. % combined 
ol aoe 


co, Oz CO, O, pu = pA'(s) 
13-6 0-4 27-7 0-2 765 = 6:18 
74-6 0-5 26-5 0-6 -— — Human blood (E.J.W. 31. i. 19). Placed 
39-5 0-4 42-8 0-6 7-39 . 6:20 on ice immediately 
109-2 0-5 64-7 0-2 710 866-15 
544-5 0-5 99-5 0-1 6-60 6-17 In the ice safe overnight 
Ordinary air 24-8 Mean 6-18 
501-9 04 842 00 653 613 
83:5 0-2 557 0-1 7:07 608 
36-0 05 441 03 7:34 610 Ox blood 
8-4 0-3 242 oo TB OE 
108-7 02 60-5 04 %707 615 
39-6 + ord. air 39-6 20-1 Mean 6-10 
33-4 03 443 02 743 614 Rabbit blood from four animals (quite 
120-9 0-4 65-4 0-2 704 6-14 fresh, placed on ice immediately) 
39-4 0-3 45:5 0-7 7:38 6-15 
378 140-0 41-2 13-2 Mean 6-14 
518-1 O-7 994 0-0 6-55 6-10 
39-6 0-3 49-8 0-7 745 6-18 Horse blood 
97-3 05 65°7 05 716 «619 
319 OS 47-0 0-6 751 617 In the ice safe overnight 
341 137-5 40-0 20-0 Mean 6°16 
26-0 10 36:3 OL 748 616 Human blood, fresh (chronic nephritis) 
19-9 0-2 — 01 759 — 
96-6 O4 _ 710 oo 


364 1368 34-7 208 — “ 
Mean of all the determinations 6-15. 


THEORY OF THE HENDERSON-HASSELBALCH EQUATION 201 


Table XI. Determination of pA‘,,) in blood at 18° with the small and 
large (LI) saturator electrodes, Series A. 


mm. Hg Vols. % combined 
cmmmmatieeaamme | 
CO, O, Co, O, 
250-4 0-8 112-1 0-1 
91-8 0-6 89-2 0-4 
24-1 0:8 59-4 — 
26°3 0-8 56-5 — 
37-7 0:3 — — 
16-6 0:8 48-9 0:2 
73-1 0-7 78-6 0-5 
58-0 143-8 65-6 22-7 
58:0 =: 143-8 65:9 23-1 
4:2 0:8 26-3 0-8 
147-1 0:8 74:0 0-2 
80-2 0-6 65°3 0-4 
58-8 0:8 59-8 0:5 
58-8 0:8 59-6 0-0 
4196 1-1 85:8 0-4 
51-7 1-2 58-4 0-0 
16-2 0-6 42-3 0-0 
22-3 1-0 48-4 0-2 
5:3 0-6 29-1 0-0 
Air from water blower 14-9 
” ” 15-1 
77-1 0:8 68-5 — 
29:3 0-3 53-5 —_ 
56-4 0:3 62-4 0-4 
127-5 0-2 76-0 0-2 
337-6 0-5 88-8 0-2 
2-2 0:7 21-8 0-4 
Air from water blower 19-2 
” ” 19-2 
41-5 0:3 63-0 0-4 
24-2 0:3 53-1 0-0 
83-1 0-5 76-5 0-2 
368-0 0-2 107-0 0-4 
44-2 0-5 64:1 0-4 
16-0 0:3 48-6 0-5 
81-7 0-4 73:8 0-1 
91-5 0-6 83-5 0-7 
22-0 0:3 56-2 0-7 
48-7 0:3 69-9 0-7 
3-0 0-4 25-2 0-4 
39-6 0-2 66-3 0-5 
484-7 0-2 112-2 0-9 
80-2 0-2 78-2 — 
Air from water blower 22-3 
” ”, 22:9 
128-6 0-6 82-0 0-6 
58-5 0-6 71-9 0-5 
38-8 0:3 66-6 0-4 
269-8 0-4 90-3 — 
126-9 0-4 90-3 — 
42-6 0-4 59-6 — 
72-1 0-2 74:3 Ate 


Bioch. xv1 


3) IIIS 
So 
bs 


el aeerem 


e gl! 
: 


7:26 
7-57? 


E 


AIetena 
eeecee | 


PAs) 
6-23 
6-23 
6-28 
6-32 


6-25 


6-27 


(6-22)? 


6-23 
6°24 


6-24 


Human blood (E.J.W. 3. x. 18), 
fresh, placed on ice 


2nd day 


3rd day. Large saturator elect. 
(LI) 


Ox blood 


CO, preliminary treatment 
Next day. CO, preliminary 
treatment 


” ” 


3rd day. "Large saturator elect. 
(LI) 


Dog blood (0-1 % oxalate) 


Large saturator elect. (L 1) 


60’ after blood Rabbit blood 


90’ was taken. (on ice) 
150° ” 

195’ ue 
240’ “4 Vol. % Oz 
300’ * (combined) 
390’ S 16-6 


lst day. Horse blood 


LI 


2nd day 
3rd day 


2nd day. Equal parts of horse 
blood and serum. V=5 


3rd day 


3rd day. Horse blood cell sus- 
pension 


14 


202 b E. J. WARBURG 
Table XI (continued) 


mm. Hg »° Vols. % combined 
“a —A—— 
Temp. CO, O, co, Q; Pu pA'(s) 
195 536-4 0-4 102-2 —_ 6-45 6-22 2nd day Ox blood 
21 91-9 0-3 755 _— 7-13 6-25 3rd day 
19-5 24-3 0-2 53-3 — 7-52 6-23 4th day. 
— 156-0 — 18-96 Mean 6-23 
_— 156-0 —_ 18-97 
19 510-5 0-5 89-6 — 6-44 6-25 2nd day Equal parts of ox 
21 91-9 0-3 72-5 _— 7-08 6-23 3rd day blood and serum. 
20 34:3 0-4 60-7 _ 7-45 6-25 v=5 
20 24-3 02 S78 — 7568 693 4th day 
— — 156-6 — 8-5 Mean 6-24 
19 508-0 — 115-8 — 6-59 6°24 Ist day Ox blood cell sus- 
20 65-7 0-3 70°3 — 7:24 6-23 3rd day pension. ¥=14-5 
20 34:3 0-4 57-1 — 7:42 6-26 SY 
20 2-8 1-1 34-4 — 7-80 6:25 5thday LI 
— —_ 156-6 — 33°8 Mean 6-25 
18 65:3 0-5 73-0 0-1 —_— _ Ist day Ox blood 
19 51-0 1-2 66-6 0-2 7:29 6-21 2nd day 
19 6-0 _— 36-5 0-3 7-96 6-22 3rd day LI 
18 58-0 140-7 —_ 16-5 Mean 6-215 
18 58-0 =-140-7 — 16-6 
19 38-6 0-8 64-0 0-5 7-42 6-28 Ist day Blood and serum, 
19 51-0 1-2 66-7 0-4 7-29 6-23 2nd day equal parts. Y=5 
18 144-5 0-6 81-6 _ 6-91 6-22 3rd day 
— a 146-3 — 6-9 Mean 6-24 
18 53-7 10 710 O09 733 623 2ndday Ox blood cell sus- 
18 64-9 1:3 76-7 0-1 7-24 6-19 3rd day pension 
18 64-9 13 75:4 0-6 7-24 6-20 
—— a 146-3 — 32:3 Mean 6-20 
CHAPTER VI 


THE DETERMINATION OF THE paA‘,, VARIATION. 


From equation (127) and (123) in logarithmic form we get 


, 100-Q(1-D 
pA’) = PX — log “O=2E—9) _ tog ©, (C04). «..-+--»-(129) 


If we now calculate py.) for blood by using the following equation 
Pu'(s) (uncorrected) = pA») + log B — log fy = pXs) + log D, (COg) 

’ As 

+ log B — log “ag s+++++(180) 
instead of (126) Pu = PA + log B — log fo, 


THEORY OF THE HENDERSON-HASSELBALCH EQUATION 203 


where B is as usual the volume % of combined CO, in the blood and Jy is the 
dissolved CO, in blood, the py: value is not correct because 


. 100-Q(1-D 
+ (— log ®, (CO,) — log WE") 


must be added to it. 

As will be seen from (129) this gives in addition the pA‘(,) variation pro- 
vided pA) is constant, which a priori is very probable as the volume of the 
protein phase in serum is relatively small and the amount of salt large in 
comparison with the quantity of electrolytes which vary with the reaction. 

®, (CO,) has been determined from (128) on the assumption that 100 ce. 
blood corpuscles combine with 49 cc. O,. This is undoubtedly incorrect but 
the error involved in the above approximation is but trifling compared with 
the uncertainty of the relative absorption coefficient in blood corpuscles 
(0-81) [Bohr, 1905}. 

When these investigations were begun there were only a few rather un- 
certain experiments which allowed of the determination of D. 

( Be e.c. combined CO, in 100 c.c. blood erpune 


: c.c. combined CO, in 100 c.c. serum 
Since then some experiments by L. 8. Fridericia [1920] and by J. Joffe and 
E. P. Poulton [1920] have appeared from which this factor may be deter- 
mined. The authors have themselves made the calculation which in many 
ways is similar to the one I am about to put forward, and they have also 
drawn attention to the difficulties with the Henderson-Hasselbalch equation 
which arise in connection with it. Joffe and Poulton have moreover shown 
how the difficulties can be evaded. 

I have however made a recalculation of these authors’ experiments because 
I believe the method of calculation I have employed has considerable ad- 
vantages over that of Joffe and Poulton and it also better admits of a com- 
parison of the various series of experiments than would otherwise be possible. 
A. Schmidt [1867] and N. Zuntz [1867] showed almost simultaneously in 1867 
that the blood corpuscles contain considerable amounts of combined COy. 
They further proved that this amount rose rapidly with the CO, tension 
(although Schmidt on the basis of some poorer experiments from a technical 
standpoint, with dog’s blood, considered that the CO, combination passed 
through a minimum with increasing CO, tension), and lastly they showed 
that the combination of CO, with serum which was obtained from blood 
saturated at high CO, tensions was considerably greater than with serum of 
the same CO, tension provided the serum was got from (the same) blood at 
lower tensions. They explained this phenomenon—which I shall repeatedly 
revert to and which I have called the Schmidt-Zuntz phenomenon—by 
assuming a transference of sodium from the blood corpuscles to the serum 
(plasma) under the conditions cited. The experiments do not permit D to 
be calculated with reasonable certainty. L. Fredericq [1878] has shown that 
in horse blood corpuscles a little over half as much CO, is combined as in the 


14—2 


204 E. J. WARBURG 


serum at physiological CO, tensions, while at 745 mm. about nine-tenths are 
combined in the blood corpuscles as compared with an equal volume of serum. 

From experiments carried out by Fr. Kraus [1898]—experiments which 
seem to me to be much too little known—it is clear that D in ox blood at 
physiological CO, tensions is about }, while at a couple of hundred mm. 
CO, it is about i" 

‘Quite recently W. Falta and Richter-Quittner [1920] have asserted that 
the partition of Cl, glucose and non-protein nitrogen between blood corpuscles 
and serum is absolutely different from that between blood corpuscles and 
plasma because they thought they were able to show that the blood corpuscles, 
before coagulation sets in, do not contain appreciable amounts of these sub- 
stances, while after coagulation has taken place they contain considerable 
quantities—as is well known. They further claim to have proved that different 
treatment of the blood such as the prolonged action of cold, addition of 
oxalate, etc. has the same effect as defibrination. If Falta and Richter- 
Quittner’s experimentsare really sound they will be of epoch-makingimportance 
in the subject of the osmotic conditions of the blood corpuscles and the per- 
meability of their surfaces, and pursued to their ultimate conclusion they will , 
occasion a complete change of the current view concerning the conditions of 
equilibrium between blood corpuscles and plasma, and thereby influence our 
ideas on the production of lymph and kindred subjects. 

L. Fredericq’s [1878] experiments show that the partition of combined 
CO, between blood cells and serum is the same as between blood cells and 
plasma. Fredericq hindered the coagulation of the blood (horse blood) by 
preserving it in an excised jugular vein. 

As early as 1893 H. J. Hamburger showed that the partition of chloride 
between blood corpuscles and plasma (obtained under oil) was the same as 
that between blood corpuscles and serum when the blood was defibrinated 
without admitting the air. 

K. L. Gad Andresen [1920] has raised objection to the experiments of 
Falta and Richter-Quittner, as he has demonstrated that the partition of urea 
is independent of coagulation (hirudin-plasma). Rich. Ege [1920] has shown 
the same for chlorine and glucose. A. Norgaard and H. C. Gram [1921] have 
shown that the Cl content of the blood is a simple function of that of the 
blood corpuscles (or amount of haemoglobin) when it is assumed the corpuscles 
contain about half the amount of Cl present in the serum (per unit volume). 
They also found this ratio in blood rendered incoagulable by the addition of 
isotonic sodium citrate solution. 

H. C. Hagedorn [1920] has shown in numerous experiments that human 
blood corpuscles obtained from blood with the addition of hirudin contain 
considerable amounts of glucose. In a short paper I have myself [Warburg, 
1920] shown that chlorine and combined CO, are distributed in the same 
manner between blood corpuscles and hirudin plasma as between blood 
corpuscles and serum even though the CO, tension is varied greatly. 


THEORY OF THE HENDERSON-HASSELBALCH EQUATION 205 


Falta and Richter-Quittner [1921] have maintained their position in oppo- 
sition to these experiments of Gad Andresen, Ege, Hagedorn and Warburg, 
or at any rate have only modified it to a very slight extent as regards the 
main problem. It is however impossible to bring the two views into agree- 
ment with one another. There are two possibilities—either Falta and Richter- 
Quittner are unable to analyse the contents of the blood corpuscles with 
regard to the substances in question, or Fredericq, Hamburger, Gad Andresen, 
Ege, Norgaard and Gram, Hagedorn! and the author are unable to do it. 

It may further be of interest to discuss an erroneous assumption which 
seems to be widespread. 

N. Zuntz [1867, 1868] showed in the paper previously referred to, that the 
alkali in blood which was estimated by titration decreased on coagulation 
and concluded from this that acid was formed in the process. This fact has 
subsequently been confirmed several times, amongst others by Loewy and 
Zuntz [1894] and by A. Jaquet [1892], but it has been curious that the amount 
of acid formed varied comparatively greatly from one experiment to another. 
Quite recently Howard Haggard and Yandell Henderson [1920, 1] have again 
demonstrated the presence of such an acid, having found, at high CO, tensions, 
several volumes °% less CO, combined in defibrinated blood than in oxalated 
blood (a difference which is however small compared with what Zuntz [1868] 
originally found). J. Joffe and E. P. Poulton [1920] have carried out numerous 
CO, determinations of Joffe’s blood, defibrinated and oxalated, and they found 
no difference in the combined CO,. A. Krogh and G. Liljestrand [1921] have 
also found the same amount of CO, combined in the defibrinated and oxalated 
blood of Liljestrand. Dr Chr. Lundsgaard and Dr E. Méller have kindly 
reported to me that in a number of experiments not yet published they have 
found no difference in the CO, combined in defibrinated and oxalated blood 
(man). 

In the previously mentioned experiments of the author no difference in 
the combined CO, in defibrinated and “hirudin” blood (horse) was found. 
That some investigators discovered a decreased titrimetric alkalinity or CO, 
combining power after coagulation can undoubtedly be ascribed to the fact 
that a quantity of blood corpuscles may be removed by the defibrination, the 
bearing of which on the present problem has not hitherto been appreciated, 
as these at higher CO, tensions or more acid reactions function as a base 
(see chap. XI) and combine with acid. 

L. 8. Fridericia [1920] has reported three experiments with 0-1 % oxalated 
ox blood which permit of the calculation of D at three different reactions. 
The reactions were calculated from the plasma figures, the value for pX,) 
which will be found below being used. 

Hasselbalch and Warburg [1918] have reported experiments (see TableXIV) 
from which a similar calculation can be made, as the carbonic acid combination 


1 EB. Wichmann [Pfliiger’s Arch. 1921, 189, p. 108] has likewise shown that Falta and Richter- 
Quittner’s view regarding the difference of hirudin blood is incorrect. 


206 E. J. WARBURG 


curve of the blood was determined (which had not been previously reported) 
simultaneously with the experiment-with the ox blood which combined with 
13-4 vols. % O,. The values obtained in this experiment are very uncertain 
because the blood contained few blood corpuscles and there were also local 
difficulties with the centrifuging. The other experiment of Hasselbalch and 
Warburg which allows of the calculation being made should give much better 
results. I have also carried out a few experiments with defibrinated ox blood 
which will be found in Table XVII. 

In calculating D from ox blood experiments it is always assumed that 
100 vols. blood corpuscles combine with 45 vols. Og. 

I have performed a number of experiments with defibrinated horse blood 
which appear in Tables XVIII and XIX. The details of the technique were as 
follows: 50-70 ce. blood were saturated for one hour in the saturator shown 
in Fig. 3. Instead of the electrode shown in the diagram, a glass cylinder 
with a rubber cork bored with two holes was substituted through which a 
long and a short glass tube were introduced. The long tube could be pushed 
up and down through the cork, the movement being airtight. The gas was 
conducted by the rubber tube through the glass cylinder before reaching the 
saturator. 

After the completion of saturation a sample of blood was first taken for 
testing (By), whereupon the T tap was turned to position I and the exit rubber 
tube was removed from its glass tube. The cylinder was then filled with the 
blood through the glass tube by siphon action avoiding contact with the air. 
The position of the tap was finally as shown in position II. 

When the blood corpuscles had settled a sample was taken through the 
long glass tube with the help of a rubber tube for pumping out of the serum 
(Sz) and then another of the blood corpuscle emulsion (Cz). A portion of the 
serum (S;;) was then saturated again and also a portion of blood corpuscles 
(Cy) in a small saturator (Fig. 5) and their gas content estimated in the usual 
way. Lastly a sample of blood was saturated afresh which before the be- 
ginning of the main experiment was saturated with the gas mixture and had 
been standing in the thermostat (at room temperature), closed in a small 
saturator, and the sample was pumped out and analysed (Byy). 

At higher CO, tensions the other gas in the spirometer was O,. 

Haematocrite estimations were performed on B;, Cy and By; as a rule, 
usually two at a time, and centrifuging was continued until the column of 
corpuscles was transparent. In those cases where haematocrite estimations 
were not done the volume of the blood corpuscles was calculated from the 
amount of oxygen, which is duly noted in the tables. 

In the experiments in which solutions rich in blood corpuscles were employed 
these were the only ones used for the calculation of D, other experiments being 
used as controls. The values for the oxygen combining power of the corpuscles 
given in Fig. 9 are all derived from experiments with blood cell emulsions. 

Most of the horse blood came from the cattle market. The blood for 


THEORY OF THE HENDERSON-HASSELBALCH EQUATION 207 


experiments 1-3 was fresh and obtained from the serum laboratory of 
“ Landbohdéjskolen.” 

A complete experiment lasted 4-5 hours. Saturation was usually finished 
after 34-4 hours. ' 

In experiments with ox blood it was passed from the saturator into a 
centrifuge tube, the rubber tube being put right down to the bottom. The 
blood then rose slowly in the tube which was closed immediately with a cork, 
and melted paraffin poured over it. After centrifuging the serum and blood 
cell emulsion were resaturated separately. 

Hasselbalch [1916, 2] has carried out the determinations! of pd‘(,) given in 
Table XII. They were done with the technique of Hasselbalch and the remarks 
previously made about this technique refer also to these experiments. It 
should be stated however that the oxygen error is much less in these experi- 
ments than with blood. 

In Table XIII a number of determinations of pdA’;,) are given which were 
carried out with the small saturator electrode in the usual way. The oxygen 
in the spirometer gas was always estimated but on no occasion did it exceed 
1 mm.; as a rule it was about 0-5 mm. 

From the mean of Hasselbalch’s four experiments at 38° 

px'(3) = 6-13 and consequently pr,,) = 6-14. 
ean the 23 estimations given in the table I get at 38° 
pr (5) = 6-144 + 0-0042, 
The standard error was 0-018, From the mean of 19 estimations at 18° 
pX'(s) = 6-283 + 0-0042. 
The standard error was 0-019. From these we get at 38° 
pA) = 6155 + 0-0042, 
and at 18° PAs) = 6-294 + 0-0042. 


The tables show that there is no variation of the constants determined by 
the reaction, which is in harmony with the small standard deviation. 
If the estimations at room temperature and at 38° are combined we get 


PAt? = 6-294 + (18° — #°) 0-007. ..eseeeeeeeee. (131) 


In Fig. 6 the values of D are plotted as ordinates and the apparent hydrogen 
ion activity exponents as abscissae. The individual determinations, as will be 
seen, are well distributed about the curve and it will be noted D varies con- 
siderably in the range of reaction dealt with. The cause of this variation will 
be the subject of a thorough inquiry in the last chapter. 

Some points on the figure are marked with a circle. They refer to the 
determinations in Table XIX. It will be observed that the addition of a 
solution isotonic with the blood, 5 cc.—100 cc. of 0-9 % NaCl solution (0-154), 
does not change D, while a similar quantity of a hypertonic salt solution, 
0-5n NaHCO, and 1-0n NaCl, makes D smaller. This last fact is of par- 

1 pd’() has been formed analogically with pA‘,, in equation (127). 


208 E. J. WARBURG 


0-4 

0-3-- = 

0-2-- =] 

i eS 
Pry 

0-0 Pate Sen ieee BSE ie fey ee Mis 


80 79 718 77 #%T6 75 74% TS I 1 170° 69 68 67 766 G5 
Fig. 6. Horse blood (defibrinated), room temperature. 


ec. combined CO, in 100 ce. blood corpuscles _ 


ee. combined CO, in 100 cc. serum = 


ticular interest and is in harmony with what would be expected from theoretical — 
considerations. 

In the experiment with the addition of physiological salt solution the 
blood combines with 62-5 vols. % CO, at a CO, tension of 37-3 mm., with a 
serum of py: 7:50, while in the corresponding serum 73-5 vols. % CO, (com- 
bined) are found. In the experiment with hypertonic salt solution the blood 
combines with the same amount of CO, (62-7 vols. °%%) at practically the same 
tension (38-5 mm.) and the same py: (7:50), but the serum combines with 
rather more CO, (76-0 vols. %). 

The CO, combined by the serum is 4 % greater in the last experiment 
than in the first. At the same time it can be demonstrated that the volume 
of the blood corpuscles has diminished by 12 %. 


Table XII. pA‘;,) calculated from K. A. Hasselbalch’s experiments, 
Biochem. Zeitsch. 78, p. 123. 


Corrected 
mm. Hg Vols. % CO, r A ae 

CO, combined pr(s) PN(s) Px’ (s) pr(s) 

38°. Ox serum III 41:8 62-0 7:44 6-12 7-45 6-13 
14-5 54:3 7:83 611 7:83 6-11 

38°. Ox serum IV 40-6 60-9 745 6-13 7:46 6-14 
59-8 59-8 745 612 7:46 6-13 

Mean 6:13 

18°, 41-0 64-6 7-36 6:24 7:37 6-25 


THEORY OF THE HENDERSON-HASSELBALCH EQUATION 


Temp. 


18-5 
18-5 


Table XIII. Determinations of pA‘) in serum carried out with 
the small saturator electrode. 


2 
512-5 
119-6 


on 
I bo ——s 
J82 sacse 
QOIH ae 


eres 
PSSERS 
HK AOCo & 


SECs xses 


Wr BODO LOT bo 


— 
oR 


mm. Hg Vols. % CO, 
C 


Mean 


combined pris) 
71-1 6-31 
64-4 a 
58-2 — 
62-6 7-36 
55-8 7-82 
75-9 6-32 
69-4 6-95 
64-7 7-36 
58-5 771 
69-1 6-43 
71-4 6-28 

—_— 7-19 

65-3 7-12 
50°7 7-89 
55-6 7-85 
58-1 7-50 
61-1 7-49 
75-5 — 
65-1 6-37 
68-1 7-24 
76-1 7-16 
75-4 -- 
71-9 -- 
65:3 7-46 
61-8 7-80 
66-0 7-40 
68-5 7-22 
795 6-38 
74:4 7-12 
63-0 7-86 
65-6 6-75 
67-6 7-32 
67-5 7-48 
74:9 7-12 
63-5 7-82 
61-8 8-03 
63-4 7-70 
63-8 7-72 
67-5 7:45 
71:8 7-14 
77-3 6-56 
71-2 6-64 
61-9 7:24 
74-9 6-43 
69-7 7-02 
63-3 7-54 
58-1 8-01 
69-2 711 
80-1 
74:7 715 
62-6 7-74 
69-0 7-34 
78-5 a 
67-8 7-40 
78-7 6-44 
68-7 7-23 
57-2 -- 
60-7 7-81 


P's) prs) 
18° ° 
6-24 Be 
6-24 =e 
6-24 ae 
6-21 — 
6-26 — 
6-24 ran 
6-22 — 
623 
— 6-18 
6-24 os 
6-28 —_— 
— 6-14 
6-28 a 
fon 6-10 
6-29 sass 
6-30 = 
ne 6-16 
6-28 a 
6-31 ae 
6-28 vik 
6-25 ate 
6-29 pi 
6-30 oe 
6-30 = 
6-28 _ 
6-25 
6-29 
6-29 doves 
6-30 te 
6-28 al 
6:28 — 
6-283 ‘ 
Pu 6-13 
a 6-14 
ab 6-15 
nok 6-16 
—_ 6-15 
6-15 
a 6-15 
=n 6-14 
— 6-14 
bs 6-13 
— 6-12 
— 6-15 
iad 6-13 
— 6-16 
— 6-14 
6-16 
a 6°15 
= 6-18 
— 6-14 


Mean 6-144 


Series A 
Ist day 


Horse blood 
2nd day 


Series A 
Ox serum 


Series B 
Ist day 


Horse serum 
2nd day 


Series B 
Ox serum 


Series B 
Ist day 
Ox serum 
3rd day 


Series B 
Horse serum 


Series B 
Horse serum 


Series B 
Horse serum 


Series B 
Horse serum 


Series B 

Ox serum 
Series B 

Ist day 
Horse serum 
3rd day 


Series B 
Ox serum 


Series B 
Ox serum 
Series B 


209 


Human serum. (Patient 
with chronic nephritis) 


210 E. J. WARBURG 


Table XIV. The value D is calculated from Hasselbalch and Warburg’s 
experiments 4 and 6. Biochem. Zeitsch. 86, p. 417. Defibrinated ox 
blood. 

In experiment 4 the volume of the blood cells is estimated at 27-5 % as the blood combined 
with 12-4 vols. % Oy. 
In experiment 6 the volume of the blood cells is estimated at 68 % in the blood cell sus- 

pension B; the suspension combined with 31-3 vols. % O,. 


The values for the corresponding carbonic acid combinations in serum and blood cells are 
determined by interpolation. 


Experiment 4. 
Blood ‘ Serum 
7 “be a ec “gh cae 
mm. Hg CO, 
mm. Hg Vols.%CO, mm.Hg Vols.%CO, during 
O, combined CO, combined centrifuging 
12-5 43-3 110-4 62-3 12-7 
193-9 64-4 
12-7 44-5 
46-5 61-1 
58-8 65-7 25-0 61-8 58-8 
94-1 69-7 96-8 69-9 
181-9 80-0 173-9 74-0 
185-4 79-7 +: 
29-0 © 67:3 89-7 
377-5 90-2 101-1 75-4 
182-9 76:3 : 
23-4 68-5 181-9 
134-9 79-3 
401-4 87-7 
430-5 88-0 17-9 75-4 430-5 
160-0 88-2 
737-0 91-6 
737-5 99-3 26-4 80-6 737-0 
161-1 92-6 
428-3 95-0 
Vols. % combined CO, 

mm. Hg , A — pu’ in 
co, in serum in blood cells D serum 
12-7 52 23 0:35 7-78 
588 66 66 1-00 7:22 
89-7 74 52 0-70 7-09 

181-9 82 65 0-80 6-83 
430-5 90 75 0:87 6-49 
737-0 96 107 1-11 6-29 
Experiment 6. 
Serum B, CO, tension during 
centrifuging, 65 mm, Blood cell suspension B 
mm. Hg ‘Vols. % CO,  mm.Hg Vols. % CO, 
CO, combined CO, combined 
19-5 61-6 298 46-1 
798 69-6 86:8 67:7 
182-0 73-0 159-8 81-2 
378-3 787 387°7 100-8 


from which at 65-0mm, Hg CO, in serum 68 vols, %, in blood cells 55:9, D =0-82, py: in serum =7°23. 


THEORY OF THE HENDERSON-HASSELBALCH EQUATION 211 


Table XV. The value D calculated from J. Joffe and E. P. Poulton’s experi- 
ments with human blood at 38°. J. Physiol. 54, pp. 148-149, Tables IT, 


III and V. 
Oxygenated blood Reduced blood 
Vols. % combined CO, Vols. % combined CO, 
(ee, ———_’—_—~. 
mm. Hg in blood Pa calc. in blood PH: cale. 
co, inserum cells OD inserum in serum ce D inserum 
10 29-2 14-5 050 7-77 31:3 21-5 0-69 7-81 
20 38-0 23-3 0-61 7:59 42-1 30-2 0-72 7:64 J.J.’s defibrinated 
30 44-8 29-6 0-66 7-49 48-7 36-7 0-75 7-52 blood 
40 50-1 35-1 0-70 = =7-41 54:4 41-6 0-77 7-45 
55 57:3 40-7 0-71 = =67-31 61-4 48-1 0:78 7:36 Blood cell volume 
70 63-1 45-4 0-72 7-27 67:3 54-2 0-81 7-30 50-93 . 
90 68-0 49-5 0-73 7-19 73-4 61-4 0:84 7-23 
156-0 77:2 58-3 0:76 7-01 — — — — Blood cell volume 
180-0 = 81-5 57-2 0-70 6-97 — — = 53-5 
110°3. 73-8 54-2 0-74 7-14 -—— —- — —- 
157-0 76-4 65-1 0-85 7-00 — _ _ 
376-3 97-5 83-6 0-85 6-72 os -- os -— 
610-0 108-9 92-9 0-85 6-56 — — — _ 
278-0 97-2 73-9 0-76 = 6-86 — -— — — 
477-°0 100-5 91-8 0-91 6-63 — -- _ -- 
40 55:3 25-5 0-46 7:45 60-8 36-0 0:59 7:49 J.J.’s oxalated blood 
55 65-3 25-7 0-39 7:38 67:8 42-5 0-63 7:40 Blood cell vol. 44-5 
15 36-6 (22-7 0-62 7-70 40-8 20-6 0-51 7-75 W.R.’s oxalated 
25 45:8 243 0:53 7:58 49-9 31-8 0-64 7-6) blood 
35 52:0 29-6 0-57 7:48 55-9 40-7 0-73 7-51 Blood cell vol. 39-77 
45 58-2 33-0 0-57 7-42 59-8 44-5 0-74 7-44 
60 62-2 40-6 0:65 7:32 63-8 45-4 0-71 7:34 
75 63°38 42-4 0-66 7-23 65-0 47-9 0-74 7-25 
Table XVI. The value of D calculated from L. 8. Fridericia’s experiment 7. 
J. Biol. Chem. 42, p. 254. Ox blood with 0-1 % sodium oxalate. 38 vols. % 
blood cells at 17°. 
Vols. % CO, Vols. % CO, Gram NaCl Gram NaCl 
combined combined py (cale.) in 100ce. in 100 ce. 

mm. Hg in plasma _ in blood cells D(HCO’,) in plasma plasma _ blood cells D(C’) 
0-08 23-4 16-6 0-71 9-46 0-578 0-260 0-45 
6-1 42-6 27-6 0-65 8-04 0-533 0-334 0-63 

39-1 67-6 62-0 0-92 7-44 0-511 0-369 0-72 
Table XVII. The value of D in ox blood. 100 ce. ox blood cells 
combined with 45 cc. Oy. 
Vols. % Vols. % 
mm. Hg combined (px‘) combined 
Temp. CO, O, (cale.) O, 
18 35-5 64-0 7-46 -- Serum 
58-6 _- 16-0 Blood haematocrite determination 35 
50-7 — 35:8 Blood cell suspension 
from serum and blood ane -.. D=0°79 
from serum and blood cell suspension D=0-74 
18 43:3 74:4 7-44 os Serum 
69-1 a 10-9 Blood 
61-6 —_ 29-6 Blood cell suspension 
from serum and blood ss -.. D=0-79 
from serum and blood cell suspension D=0-74 
18 21-9 63-5 7-67 a Serum 
60-2 —_ 10:9 Blood 
48-9 _ 33-5 Blood cell suspension 
from serum and blood aa ... D=0-78 
from serum and blood cell suspension D=0-69 
18 9-8 56:8 7:97 oa Serum 
50-2 — 10-9 Blood 
from serum and blood cen ... D=0-52 
18 22-5 62-5 7-65 — Serum 
56-9 — 15:7 Blood 
49-2 ~- 24-5 Blood cell suspension 
from serum and blood vee oo. D=0-73 


from serum and blood cell suspension D—0-75 


219 E. J. WARBURG 


Table XVIII. The value of D in defibrinated horse blood 
at room temperature. 


Vols. % Vols. % 
Vols.% Mean Vols. % combined combined 
com- ofpH' com- Haemato- in CO, in 
mm. Hg bined calc. bined crite blood blood 
Temp. CO, CO, inserum OQ, number _ cells cells D 
19 347 59:3 203 41:5 48-9 By from By, Sz and haematocrite By 44-4 
69-9 0-2 a= — ws 
51-1 36-0 — — Oo y Oe a Pr By 44-4 
68-7 oo — ee 
7-51 0-64 
17 15-4 46-2 20:2 400 505 By ,, By, St * inet 30-0 
57-0 0-6 ~- — ws nae 
40-4 35-1 — — Cy HES cy Ee By 33-1 
56-1 06 — eget 
7-96 0:57 
18 82-2 76-1 20:0 . 410 489 B, ,,. By, Sy ” By 61-7 
87-1 1:3 — — ws 
69-2 34-7 — ae Og eu SE ” By 61-1 
69-4 33-2 66:5 49:9 Cyy 
87-2 Sa — &s&& » O77, Sy >» Cyy 608 
76-9 19-9 As Pusey Br ” Br, Sir ” By 62: 
7-23 0-70 
18 38:4 59-3 19S | 404. 400- Bes Bg ey ” By 37:2 
74-1 0-5 — — ws 
55-1 35:8 — — Cy ” Cy, Sy ” By 48-1 
53-5 34:1 66:0 51:8 Cyy 
71:8 Fs Ie — Sy » Cy, Sy i Cyr 45:2 
60-1 192 37:5 51:1 Bry ,, By, Str ” Byy 40-7 
7:48 0-64 
17 17-3. 47-4 19-7 403 492 By, ,, By, Sy ” By 32:3 
57-7 0-4 a aa I 
39-7 37-9 _ — Cy ” Ct, Sy ” By 34:3 
39-1 36-2 65:8 (55-0) Cy 
56-4 OS — S8&y » Cy, Sry Pa Br 29-9 
47-0 188 39:5 476 Bry ,, By, Sry ” Byy 32:7 
7:72 0-56 
17 86-2 78-4 19-8 41-4 48-5 By ” By, Sy ” By 63:8 
88-9 1-1 — — ss 
71-7 35-5 _— — CT ” Cy, Sy ” By 65-4 
89-0 1-1 a — Sir 
74:8 31:3 ~~ om Oy. »° Op an ” By 65-8 
78:1 18-6 36-5 51-0 Bry ” Bit, Str ” Bir 59-2 
7-21 0-74 
19 1783 90-4 184 413 446 By, ,, By, Sy ” By 78-0 
99-1 0:8 a= — ws 
81-9 35-7 -= — Oy 4 Oy Sy ” By 77-6 
82-2 350 755 466 Oj; 
98-7 SARIN as — Sy » Cry, Sty .» Cry 76-9 
89-1 179 396 45:3 By ,, Bry, Str ” By = 745 
6-95 0:78 
18 1827 95-1 23-8 489 487 By; ,, By, Sy ” By 80-4 
109-3 1-4 — I 
88-5 36-6 _ pase CT ” C7, Sy ” Cry 81:8 
88-1 366 774 473 Oj 
109-6 0-8 ae? Fah Sir 9 Ci, Sir ” Orr 82:4 
96-4 227 488 465 Bry ,, Bry, Sir ” By = 826 
6-98 0:75 
18 760 T51 230 475 484 By, ,, By, Sy ” By 57-9 
90-7 10 — — 
68-6 36-0 -- — O; » Oy, & ” Oxy 61-2 
67-1 366 763 480 Ory 
89-6 0-5 cae: iy Sir » Or Str ” OTL 60:2 
76-1 227 470 483 By , Bry, Sry ” By = 587 


7-28 0-67 


THEORY OF THE HENDERSON-HASSELBALCH EQUATION 213 
Table XVIII (continued) 


Vols. Vols. % 
ea 


Vols.% Mean Vols.% combin combined 
com- ofpH' com- Haemato- O,in CO, in 
mm. Hg bined cale, bined crite blood blood 
Temp. CO, CO, inserum Oy, number cells cells D 
18 35:5 66-7 _ 39-0 — By from By, Sy and haematocrite By 46-6 
79-5 0-5 — — ss 
55-5 37-6 a ete pps Ote OE Be Cir 48-4 
53-8 38:8 79-7 487 Oj7 
78-7 0-4 — Sy » Cr, Si A Cyr 47-7 
66-7 Rae 38°3 — It » Bri, Sir ” By 47-3 
7.55 0-61 
18 461-2 115-7 19-0 — — By , By, Sy » Cir 982 
128-2 0-4 _ — ws 
113-8 34-7 _ — Qty , Cy, 8 “ Cy, 1093 
110-1 353 775 45:6 Cyy 
126-5 0-7 — — Sy », Cry Su ‘ Cyy 1053 
117-6 18:7 41-8 8 By , Bry Su oy Bir 105-2 
6-65 0-84 
xg 12-6 60-5 0-4 — — Sy Oy, By 32 Cyy 329 
40-1 36-9 _ — Cf 
35°8 39-4 «791 «= 49-8 Cy Cg:)«=~C ry, St ” Cyy 29-4 
60-0 0-4 == —_ II 
49-0 185 374- 495 By , By, Sry ra By 305 
7-87 0-52 
18 443-1 99-5 13-4 304 441 By , By Sy ” By 78-6 
97-6 33-2 — — Cc 
- 108-7 0-2 — — S& , Cy, Sy s Cjy 93:8 
90-9 ~326 730 447 Cyy 
104-7 —o — Sy » Cy, Si ” Cyy 85:8 
98-5 13-6 310 43:9 By , By Sry Mas Byy 84-4 
6-59 : 0-84 
19 8215 52-7 13-7 279 401 By, ,, By, Sy é By 32-4 
60-7 0-2 — — wf 
40-6 38-5 reas aa Cy ” CL, Sy ” Cy 34-9 
38-4 38:8 783 49-6 Cyy 
59-8 0-7 _ — Sy » Cry, Su ” Cy 325 
52-6 131 265 494 By , By, Sty * By 328 
7-66 0-56 
18 413-9 108-1 18-2 41-3 44-2 By ” Bi, S] ” By 98-4 
114-9 0-4 — — ws 
103-3 33-9 _ — Qt , Cy, Sy ” By 99-8 
101-2 33:2 750 443 Cyy 
111-8 0-7 — — Sy » Cry, Sty ek Cyr 976 
6-64 0-87 
18 143 49-2 189 378 SOl By , By & ie By 26-8 
63-2 0-6 — — ws 
38-7 38-7 oo — Cy » Cy Sy = By 30-0 
37-7 38:2 760 650-2 Cy 
58-9 0-5 — — Sy » Cr, Si Re Cyy 297 
48-7 175 «368 «6475 0«|—ByySCy,)=~ Buy, Str ” By 3h 
7-83 0-49 
18 2173 99-4 188 408 461 By ,, By, Sy ” By 81-1 
111-9 0-5 -- — ws 
94-2 35:8 —_— — Cy ” Cy, Sy ” Cry 89-1 
92-5 353 765 461 Cyy 
110-1 0-5 — — Sy » Cy, Siu ¥s Cyy 8885 
99-3 185 39:9 479 By , By, Sry ae By = 833 
6-91 0-80 
17 194 51-4 -208 415 503 By , By, St ” By 39-6 
59-8 0-9 — — ws 
40-0 38-9 — met MO oy Oy, Oy x Cyt 32:7 
41-8 38:1 755 60-4 Cy ‘ 
59-3 0-5 — — Sy 5 Cry Str a Cy 31:3 
50-2 199 40:0 49:7 By ,,° By, Sry * Byy 36-0 


7-68 0-54 


Temp. 
17 


19 


19 


Temp. 


18 


19 


19 


214 


com- 
mm. Hg bined 


CO, 
45-1 


42-3 


172-4 


37:3 


77-6 


38-5 


com- 
mm. Hg bined 
CO. 


CO. 


SeSSSE 
OAWR AGA 


~1-+1 0 


Seas 
@enw-aro 


_— ; 
‘calc. 


in serum 


7-44 


7-38 


E. J. WARBURG 


Table XVIII (continued) 


Vols. % 
com- Haemato- 
bined crite 
O, number 
20-6 41‘8 
. 0-6 oe 
37-7 = 
35-9 72:9 
0-9 oo 
19-9 as 
216 45-0 
0-3 — 
37-7 — 
0-7 — 
36-0 75-0 
20-7 43-4 
0-7 — 
35-8 7 
35°3 77-0 
1-0 — 


Vols. % 
combined 

Oz in 
blood 

cells 
49-4 
49-3 


By trom By, Sy and haematocrite By 


ST 
CT oy (Gg, Sy ” Cry 
CTr 
Sir» Crp Siu ” Ci 
By» By Si . Cir 
By ” By, ST ” By 
ST 
OT ” CL St} ” Cir 
Sir 
Cy » Cry Siu 9 Cy 
By» By Si B By 
St} ” CI, ST ” Cir 
CT 
Cir ” Cit, Si ” CTL 
Sir 


Table XIX. The value of D in defibrinated horse blood 
with various additions. 


Vols.% Mean 


CO. 


IAAIS 
Seles 
wm =1 GO GO Or C1 


118-5 
138°5 


of pa’ 
calc. 


bined 
inserum Oy, 


number 


100 cc. blood +5 cc. 0-9 % NaCl sol. (isotonic with blood) 


7-60 


19- 


6 


40:3 


758 
38°8 


48-6 


49-2 
47-4 


ps from By, Sy and haematocrite By 
I 
Cy ” 


C Cy, Sy ” Cir 
II 

Sty » Crp Sir » Cy 
By » By, Str » By 


100 cc. blood +5 ec. n/2 NaHCO, (hypertonic to blood) 


7-46 


7-60 


By from By, Sy and haematocrite By 


ST 

CT ” Cl, Sy ” Ci 
OIL ; 

Sir» Crp Str ” Orn 
By ” Bu, Sir ” Bry 


100 oc. blood +5 ec. n/1 NaCl (more hypertonic) 


186 36:0 51-7 
OS ne ee 
7 ee ee 
375 724518 
0-7 cas es 
189 360 525 
191 340 56-2 
0-2 a _— 
35-4 nS a 
306 545 © 56-1 
4 —-— — 


es from By, Sy and haematocrite By 


I 
CTY w Oy, Sy ” Oy 
Cn 
Sir ow» Cry Sir » Oy 


0-78 


0-63 


0-59 
36:8 
44-7 


40°7 
0:56 


The curve with a continuous line in Fig. 7 is the same as that in Fig. 6 
(horse blood at room temperature); the “recumbent” crosses represent the 
values of D in oxygenated blood calculated from the experiments of Joffe 


THEORY OF THE HENDERSON-HASSELBALCH EQUATION 215 


and Poulton [1920] with defibrinated human blood at 38°; the “erect” crosses 
represent the values of D in almost completely reduced blood; the marks 
enclosed in squares refer to oxalated blood. 

aay UGTA IEG FAS sam eit ce ice) Oe i 


0-30}— od 
0-20}— o 
0-10}— ¥ 
oool_| | | ee, | bts eee | Prey 


80 79 78 77 16 75 74 73 72 T1 70 69 68 67 66 65 
Fig. 7. Joffe and Poulton’s results compared with experiments with horse blood. 


Oxygenated horse blood at room temperature, defibrinated. 
—— Reduced human blood at 38°, defibrinated. 


x  Defibrinated human blood, oxygenated 
+ reduced 

+: ” 38°. Th : 
® Oxalated ro al eveew of D 
fA ” ” reduced 


Joffe and Poulton’s experiments with defibrinated human blood at 38° 
give values for D which, allowing for experimental error, are identical with 
those for horse blood cited above, while reduced blood gives values which are 
fairly accurately represented by a curve situated 0-125 higher in the diagram 
and which runs parallel with the curve for horse blood. The experiments 
with oxalated blood all give smaller values for D than corresponding ones 
with defibrinated blood but the values are more irregular. Apart from this 
the experiments with oxalated blood agree with those with defibrinated blood 
as regards all the other factors. 

The addition of oxalate therefore causes a change in the distribution of 
carbonic acid in the blood and it is quite analogous to the alteration which 
takes place in horse blood on the addition of hypertonic salt solution; sodium 
oxalate in point of fact introduces fresh cations into the plasma which makes 
the serum hypertonic and causes the blood corpuscles to shrink. 

Referring to J. Joffe’s experiments on human blood! we find at a pressure 
of 40 mm. CO,, 45-0 vols. % CO, (total) in defibrinated oxygenated blood 
and 44-5 vols. % in oxalated blood, while in the corresponding serum there 
are 52-8 vols. % and in plasma 58-0. As regards reduced blood defibrinated 
contains 50-4 vols. % total CO, and oxalated 52-3, while there are 57-1 vols. % 
in serum against 63-5 vols. % in plasma. At 55mm. CO, exactly similar 
conditions prevail. It is seen therefore that the CO, content of plasma is 
t J. Joffe and E. P. Poulton [1920, p. 148, Table IT]. 


216 E: J. WARBURG 


- 10 % higher than that of serum and the blood corpuscles have shrunk about 
13 %. 

Similar conditions would presumably be found in osmotically hypertonic 
blood (nephritis and sweating baths, diarrhoea, etc.). 

D. D. van Slyke and G. E. Cullen [1917] some years ago introduced a 
method for determining the amount of combined bicarbonate in the blood 
which is in many ways extremely useful and quite accurate. This method 
appears to give more variable values for the combined CO, than might have 
been expected from the experiments of J. Christiansen, Douglas and Haldane 
[1914], Hasselbalch [1916, 2], Donegan and Parsons [1919] and Jarlév [1919] 
on the CO, combining capacity of whole blood when it is assumed the partition 
between plasma and blood corpuscles is constant at the same reaction. 

In the previously mentioned paper Hasselbalch and Warburg [1918] have 
cast suspicion on methods based upon the determination of the combined CO, 
in the serum (or plasma) when it is desired to draw conclusions about the 
“alkali reserve” (van Slyke) or the “reduced hydrogen number” [Hassel- 
balch, 1916, 1], having pointed out that the Schmidt-Zuntz phenomenon may 
play some part, provided the blood is not centrifuged at the alveolar CO, 
tension. Recently Joffe and Poulton [1920] in their much cited work have dealt 
with the same question and put forward a similar explanation of the degree 
of the dispersion in the experiments of van Slyke and his collaborators. 

It has for long seemed strange to the author that the “fortuitous” varia- _ 
tions which are associated with van Slyke’s method appear to be considerably 
greater than those we should expect to arise from the Schmidt-Zuntz pheno- 
menon. I think I have already made it clear that the most important cause of 
the variations is to be sought in the addition of varying amounts of oxalate 
to the blood before centrifuging it. Van Slyke and Cullen [1917] have shown 
that oxalate mixed with serum does not alter its combining power with CO, 
but they have overlooked the fact, as far as I am aware, that the oxalate 
could change the distribution of CO, in the blood. Van Slyke and Cullen 
recommend the addition of 0-5 % oxalate, which is a rather large amount. 
I have little doubt it would be an advantage to the method if the amount of 
oxalate employed was always the same and was as small as possible because 
there will always be a slight want of uniformity on account of the variation 
in the quantity of blood corpuscles in different experiments. 

In the final chapter it will be shown that the Cl’ partition between blood 
corpuscles and serum follows exactly similar laws to that of HCO’;. It is 
thus important in determining the amount of Cl’ in plasma to pay attention 
to the amount of oxalate added (the same applies also to fluoride and citrate), 
just as we should naturally work with a known CO, tension as has been par- 
ticularly pointed out by H. J. Hamburger [1902] and L. 8. Fridericia [1920]. 

After this digression let us revert to the curves. In Fig. 8 the value of D 
will be found in oxygenated ox blood at room temperature calculated from 
Fridericia’s [1920], Hasselbalch and Warburg’s [1918] and my later experi- 


esi te nk nt etn leita 


—-" 


THEORY OF THE HENDERSON-HASSELBALCH EQUATION 217 


ments. The values emanating from the previously mentioned unreliable 
experiments of Hasselbalch and Warburg are put in parentheses. The other 
experiments fall evenly on a curve similar to that for reduced human blood 
according to Joffe and Poulton’s [1920] experiments. The experiments are 
too few to bear further investigation. 


© 9 op ad astra ta npr ae, Pe ees 9 FG OO A ea a 


1-00}- @ 
0-90}- as 
0-80 x pe oii 

0°70 }— x pin (x) 4 
0-60 }- + 


0-40 

x) 
0-30 }—- ad 
0-20 = — 
0°10} ~<d 


oi: 
aay Pea Silt Toe aes Tha | ike. Wee Hi 
80 79 78 77 176 15 74 73 72 71 70 69 68 67 66 65 64 63 


Fig. 8. D in ox blood at room temperature. 


I | | | | beet et | : Fy ll 
56 |— 2 100 
u 
55;- + 99 
54}— —+ 98 
53t- + 97 
52 x6 7 96 
51}- 95 
ae i 
50+— x . x n + 94 
49+ onan | 93 
48 x— 92 
ait ae | o1 
46 }— x — 90 
x 
45;-— x —} 89 
x 
44 88 
a 
> 
43}- : — 87 
42/- ; Pee a 
(oe Sees | ANE | | | 


ped 
80 79 78 7:7 76 75 74 73 72 71 +70 69 68 67 66 65 


Fig. 9. Curve I. cc. of O, in 100cc. horse blood corpuscles. 
» Il. The volume of blood corpuscles as a % of their volume at py: 6-50. 


Bioch. xv1 15 


218 E. J. WARBURG 


Fig. 9, curve I, and Table XX show how the oxygen in 100 cc. of horse 
blood corpuscles varies. In the earlier literature we find the statements of 
H. Nasse [1878] and Hamburger [1902] on the volume of the blood corpuscles 
at low CO, tensions and at tensions about 1 atmosphere. These statements, 
however important they may have been for the development of our know- — 
ledge of the physical chemistry of the blood, are obviously not accurate. 
Jofie and Poulton have not been able to demonstrate with certainty the change 
of volume determined by the reaction with greatly varying CO, tensions. The 
variations shown in the diagram are well grouped about the curve with the 
exception of two points, so that it will have considerable value on account of 
its novelty. The variations in volume will in chapter XII be the subject of a 
theoretical investigation. Curve II gives the volume of the blood corpuscles 
as a percentage of the volume at py: 6-50. 

At py 7-40 as will be seen horse blood cells combine with 49 ec. O, per 
100 ce. blood cells. A. Norgaard and H. C. Gram [1921] found values diverging 
from mine in an examination of human blood at physiological reaction; as 
a mean of three determinations I found 48 vols. %. The difference is pre- 
sumably in the haemoglobin determinations (see chapter XI, p. 285 footnote). 
With regard to the volume of ox blood corpuscles I have only a very few 
determinations available, the mean of which gives 45 cc. O, in 100 cc. blood 
corpuscles at “physiological” py-, but the determination can very well be a 
couple of vols. % out. 


Table XX. 
Vol. of blood Vol. of blood 
Vols. % cells as a % Vols. % cells as a % 
combined O, of vol. at combined O, of vol. at 
PE in blood cells pr 6-50 PY in blood cells pr 6-50 
6-50 43-7 100-0 7:30 48-6 89-9 
6-60 44-5 98-2 7-40 49-0 89-1 
6-70 45-2 96-7 7-50 49-4 88-4 
6-80 45-9 95-2 7-60 49-8 87-8 
6-90 46-6 93-9 ° 7-70 50-1 87-2 
7-00 47-2 92-7 7-80 50-4 86-7 
7-10 47-7 91-7 7-90 50°7 86-2 
7-20 48-2 90°7 


From the curves in Figs. 6 and 9 we can now calculate the pA’(,) variations 
for horse blood at room temperature. This is done in Table XXI and the 
results are graphically represented in Fig. 10. It will be noted the curves also 
give a measure of the error committed, if py: of the blood is estimated by the 
Henderson-Hasselbalch equation and pA’;,) used in the calculation. The error 
would then always be 0-011 greater than the correction in the table. It will 
also be observed that it is very easy by choosing a suitable constant to get 
a good result in calculating py from the Henderson-Hasselbalch equation, 
and that a larger (logarithmic) constant should be used in an alkaline reaction 
than in a more acid reaction. This is in agreement with what we found in 
the preceding chapters, The rather steep rise of the curves indicates that 


THEORY OF THE HENDERSON-HASSELBALCH EQUATION 219 


Table XXI. 
log 100-2 (1 =P) 
100-Q(1-D) _,,,100-QUL-D) ~!°8 yop 
pH D Q 100 8 100 —log &, (CO,) 
10 vols. % O Vv =5-9 , (CO,) =1-06 log ©, (CO,) =0-027 
7-90 0-50 19-7 0-901 0-045 0-018 
7-60 0-60 20-1 0-920 0-036 0-009 
7-40 0-66 20-4 0-931 0-031 0-004 
7-27 0-70 20-7 0-938 0-028 0-001 
6-88 0-80 21-5 0-957 0-019 —0-008 
6-50 0-88 22-9 0-973 0-011 ~ 0-016 
15 vols. % Oz Vv =8-0 , (CO,) =1-09 log ®, (CO,) =0-036 
7-90 0-50 29-6 0-852 0-069 0-033 
7-60 0-60 30-2 0879 0-056 0-020 
7-40 0-66 30-6 0:896 0-048 0-012 
7-27 0-70 31-1 0-907 0-042 0-006 
6-88 0-80 32:3 0-935 0-029 ~0-007 
6-50 0-88 34-4 0-959 0-018 ~0-018 
20 vols. % O, v=9-9 $,(CO,)=110 log 4 (CO,) = 0-042 
7:90 0:50 39-4 0-803 0-095 0-053 
7:60 0-60 40-2 0-839 0-076 0-034 
7-40 0-66 40-8 0-862 0-064 0-022 
7-27 0-70 41-4 0-876 0-057 0-015 
6-88 0-80 43-0 0-914 0-039 ~ 0-003 
6-50 0-88 45-8 0-945 0-025 ~0-017 
25 vols. % O, v=11 ,(CO,)=1-12 log ®, (CO,) =0-049 
7:90 0-50 49:3 0-753 0-123 0-074 
7-60 0-60 50-9 0-796 0-099 0-050 
7-40 0-66 51-0 0-827 0-082 0-033 
7-27 0-70 51:8 0-845 0-073 0-024 
6-88 0-80 53-8 0-893 0-049 0-000 
6-50 0-88 57-3 0-931 0-031 -0-018 
30 vols. % O, v=12 , (CO,) =114 log , (CO,) =0-057 
7-90 050 59-1 0-705 0-152 0-095 
7-60 0-60 60:3 0-759 0-120 0-063 
7-40 0-66 61-2 0-792 0-101 0-044 
7:27 0:70 62:1 0-814 0-089 0-032 
6-88 0:80 64-5 0-871 0-060 0-003 
6-50 0:88 68-7 0-916 0-038 —0-019 


Donegan and Parsons’ estimations in marked alkaline reactions are mostly 
correct. 

As already mentioned the curves are so constructed that the correction 
for defibrinated horse blood can be directly read when the py: is calculated 
with (130). 

In the calculation of the py-(,, of human defibrinated oxygenated blood at 
38° the curves are similarly directly applicable. (The condition for this should be 
that D as well as Q is the same in oxygenated human blood at 38° as in oxygen- 
ated horse blood at 18°, but the error due to Q being slightly different will be 
negligible.) In using the correction curves for calculating the py'(,) value of 
oxygenated defibrinated ox blood at room temperature and of reduced de- 
fibrinated human blood at 38°, one proceeds in such a way that having deter- 
mined the py: value by means of (130), the py.) value is looked for on the 
continuous line curve of Fig. 7, which has the same D value as the calculated 
Pus) On the interrupted line curve, and then the correction corresponding to 


15—2 


220 EK. J. WARBURG 


the new py: is read off Fig. 10. This manoeuvre is most easily performed by 
subtracting 0-40 from the value of py: calculated with (130) and then using 
the resulting number for seeking the correction in Fig. 10. The error com- 
mitted by this method will hardly be appreciable in the second decimal place. 

It will be noticed from the course of the correction curves that pA’;,) 
increases with the pg: and with increasing amount of haemoglobin in the 
blood. That this can only just be demonstrated with the potentiometer in 


To de ed ee ie ciwee ot ae 


0-01 


0-00 
3 

-0-01+ 3% ea 
£2 

-0-02/- SS i 


Py: Uncorrected 
BEA EGE. | | | Bae | EE, i eT | 
80 79.78 77 76 75 74 73 72 71 70 69 68 67 66 65 
Fig. 10. Correction for use in the calculation of pyy-;s) for horse blood 

at room temperature and human blood. at 38°. 


the case of the variation determined by the reaction and not at all in the case 
of that determined by the concentration is due to the fact that potential 
measurements, even in the excellent arrangement employed by Parsons whose 
method can hardly be improved upon at present, are not sufficiently reliable 
in their application to blood. It follows therefore that the py: value ought 
to be determined according to the principles given here if the best possible 
results are desired. 

It will also be observed that pA’,), when calculated as described in this 


THEORY OF THE HENDERSON-HASSELBALCH EQUATION 221 


chapter, is larger than the constant found in an earlier one. The reason is 
presumably that the blood corpuscles in spite of all precautions are to a certain 
extent “deleterious” to the platinum electrode, even in Parsons’ and Donegan 
and Parsons’ experiments where saturation and measurement of potential 
were undertaken in different vessels. 

Lastly it is important to note that the calculations of Hasselbalch using 
his pg, curve gave much more certain results than by a measurement of 
potential, and that the results obtained can easily be rectified with the help 
of the constants given here. 

It has often been attempted to estimate the reaction of blood (and serum) 
colorimetrically and in later years such methods—partly combined with 
diffusion processes—have been described by Levy, Rowntree and Marriot 
[1915], W. M. Bayliss [1919] and quite recently in a very refined form by 
H. H. Dale and C. Lovatt Evans [1920]. 

J. Lindhard [1921] has devised a micro-method on Dale and Evans’ 
principle which allows of a determination of the py value with a relative 
accuracy of 0-02. 

With the method of calculating py given here, which is entirely based on 
a refinement of the Henderson-Hasselbalch principle, the apparent hydrogen 
ion activity exponent by using the proper method of CO, determination can 
be reckoned with a relative accuracy of 0-005 in any two determinations in 
human blood at 38° and horse blood at room temperature, and with an error 
which does not exceed the absolute value by more than 0-015. In human blood 
and ox blood at room temperature two determinations can be carried out 
which do not differ by more than about 0-01 relative to each other and with 
an error which does not exceed the absolute value by more than 0-03. 

In conjunction with the above a few examples of the calculation of py.) 
with Henderson’s and Hasselbalch’s equation may be given. Let us examine 
a sample of human blood with 20 vols. % combined O, at 38°. The relative 
absorption coefficient is 0-91. At 10 mm. CO, we find 25 vols. % total CO,. 

The amount absorbed is 

Oe! = 0-7 vol. %. 
The amount combined is thus 24-3 vols. %: 
log 24-3 — log 0-7 = 1-563. 
Converting the values from Hasselbalch’s 1916 curve to py: (Bjerrum) we get 


Pu'(s) = 1°727. 
Using Parsons’ and Donegan and Parsons’ pA’;,) we get 
Pu'(s) = 7-734. 


With the author’s constant = py) = 7-710, 
and by calculating with (130) and correcting with Fig. 10 
Pu'(s) = 7751. 


222 E. J. WARBURG 


At 40 mm. we find in a similar blood 50 vols. % total CO, and conse- 
quently 47-3 vols. % combined CO,, from which, in a similar manner to the 
previous example, we get 


Warburg 
Hasselbalch Parsons and Donegan (126) (130) and correction 
Pu'(s) = 7-374 7-421 7-397 7-418 


and for blood at 300 mm. CO, with 90 vols. % total CO, and 70-1 vols. % 
combined CO, 


Warburg 
Hasselbalch Parsons and Donegan (126) (130) and correction 
Pao) = 6719 6-787 6-763 6-750 


CHAPTER VII 
THE REACTION OF THE BLOOD CORPUSCLES. 


Although the majority of workers who have investigated the reaction of 
the blood have realised that what is generally called “the hydrogen ion con- 
centration of the blood” is really only “the hydrogen ion concentration of 
the serum,” few have attempted to get some knowledge of the reaction in 
the interior of the blood corpuscles, a question however which is of the greatest 
interest because analogies with other cells may be drawn from it if similar 
factors determine the difference in reaction between serum and blood cells, 
and serum and tissue cells. Hasselbalch and Lundsgaard [1912] and later 
J. M. de Corral y Garcia [1914] claim to have shown that blood cells at 
physiological CO, tensions are more acid than the corresponding serum. The 
evidence produced is however faulty, as the phenomenon discovered by 
Hasselbalch and Lundsgaard is only an expression of the Schmidt-Zuntz 
effect as it appears in A. Schmidt’s [1867] and N. Zuntz’s [1867, 1868] and 
many other old experiments, and as is very clearly seen in Hasselbalch and 
Warburg’s [1918] experiments and expositions. Hasselbalch and Lundsgaard 
showed that the reaction at constant CO, tension was more acid in blood 
than in the serum centrifuged from it. The experiments of the above men- 
tioned authors show that the content of the serum in bicarbonate and therefore 
the reaction of the serum at a given CO, tension is a function of the CO, 
tension in the blood at the time of centrifuging, so that we cannot conclude 
anything about the reaction in the interior of the blood cells from a difference 
of reaction between serum in blood and separated serum. 

The Schmidt-Zuntz phenomenon, which has been the cause of Hasselbalch | 
and Lundsgaard’s error, indicates that some kind of equilibrium prevails 
between the activity of the ions in blood cells and serum but it is not possible 
at the moment to say how this equilibrium is maintained. In the fina] chapter 
an attempt will be made to give a theoretical and experimentally workable 
solution of the problem on a relatively broad basis. 


THEORY OF THE HENDERSON-HASSELBALCH EQUATION 223 


If the reaction of blood is determined electrically and then haemolysis 
produced (e.g. by freezing) without letting the blood come in contact with 
any new gas mixture, the reaction after haemolysis will show whether there 
is any difference of reaction between blood corpuscles and serum. The assump- 
tion involved in this reasoning—as for numerous other instances later in the 
chapter—is that the dissociation of the electrolytes determined by the reaction 
is the same at the same reaction before and after the haemolysis. 

Konikoff [1913] has reported experiments in which he estimated the reac- 
tion of blood electrometrically using Hasselbalch’s method. He employed a 
special electrode vessel with a relatively large platinum plate and assumed 
he avoided the oxygen error in this way (demonstration of the oxygen error 
was the real object of his work). Having determined the reaction of the blood 
he haemolysed it by freezing and found that the reaction had become much 
more acid. 

Milroy [1917] determined the reaction in haemolysed blood at a CO, 
tension of about 40 mm. and found a py of roughly 6-60, that is about ten 
times as large a hydrogen ion activity as in blood serum at the same CO, 
tension. 

Although Konikoff devised a special technique to avoid the O, error (he 
did not however make use of minimal immersion) and Milroy employed 
Héber’s principle and was aware of the existence of the O, error, I do not 
hesitate to say that the results of both these investigators are misleading as 
they undoubtedly had considerable quantities of O, in their haemoglobin 
solutions. Neither of them attempted to prove that the oxygen was actually 
dissipated during the measurements and it is beyond all question that too 
low potentials will be obtained if the estimations are carried out with deeply 
immersed platinum electrodes in strongly oxygenated haemoglobin solutions. 

Parsons [1917] has published some electrical determinations in haemolysed 
blood in which the reaction was almost the same as that usually met with in 
serum with a similar CO, tension. 

L. E. Walbum [1914, p. 231] has shown that the reaction in a solution 
of blood corpuscles (10 blood + 90 physiological NaCl) is the same before and 
after haemolysis. Although the quantity of blood corpuscles in the experi- 
ments was rather small this is counterbalanced by the liquid containing them 
(serum + NaCl solution) being relatively poor in buffer substances. L. S. 
Fridericia [1920],and J. Joffe and E. P. Poulton[1920], in the papers extensively 
referred to in a previous chaper, calculated the reaction in blood corpuscles 
and serum at the same CO, tension in a manner which in essence is identical 
with that employed in the experiments about to be described, but they made 
the assumption that pA.) was the same as pz, in bicarbonate solutions of the 
same carbonic acid binding power, an assumption which is to a certain extent 
supported by Hasselbalch’s [1916, 2] rather scanty estimations in dialysed 
haemoglobin solutions. 

It is therefore hardly possible from experiments in the literature to con- 


224 E. J. WARBURG 


clude anything with certainty about the reaction in the blood corpuscles but 
I believe that Joffe and Poulton’s contribution must be looked upon as the 
most important on this subject even though it is open to objection as pA) 
was not experimentally determined. 
100-Q(1-D) Poco 
If (120) Ax () = Ae) 100D x eae 


is divided by (119) — agyg) = Ae) td aoe =?) a 


eo _ fo 132 
we get Fd 1) Oe (132) 


which in logarithmic form becomes _ 
Pu'(s) — Pa’) = PX) ir Pre) ag log PIS bats sx ceten (133) 


Since we have previously determined pA.) and D, we only require the value 
pc for estimating the difference in reaction between blood corpuscles and 
serum, which we will now attempt to determine. The determination of pA; 
was associated with much greater difficulties than I originally expected. One 
of the most important was to get rid of the oxygen at the reactions and 
temperatures dealt with, but this was overcome to a large extent by the 
technique described in chapter IV. It was easy to haemolyse ox blood by 
repeatedly freezing so that it became completely transparent and only a 
trifling amount of blood corpuscles was left, but it was practically impossible 
by freezing alone to haemolyse horse blood so thoroughly. Even after freezing 
and thawing three times numerous intact blood corpuscles are present and 
many amorphous fragments are seen with the microscope. If the volume of 
the disperse phase is determined by the haematocrite—which can easily be 
done—it will never be found to be over 5 % of the whole system even when 
very concentrated blood cell suspensions are used. If a liberal amount of 
saponin is added to horse blood it will become completely transparent and 
only a few formed constituents (about 1 % in the haematocrite) can be seen 
with the microscope. This difficulty of haemolysing horse blood by freezing 
led me to work with blood haemolysed by saponin as it was found that the 
combined CO, was the same whichever of the two methods was employed as 
the following experiment shows. 

Defibrinated horse blood was concentrated by centrifuging. Haematocrite 
reading 59-5. 

mm, Hg CO, Vols. % combined CO, 


24:3 45-2 Saponin 
24-0 45-0 Freezing 
79-2 72-7 ie 

79-9 72-9 Saponin 


In using a concentrated solution of horse blood haemolysed by saponin 
a new difficulty arose. When it is treated for a long time with high tensions 
of CO, it becomes very viscous and shortly afterwards a large quantity of 
haemoglobin crystals separate out so that the experiment has to be abandoned, 
It has been found that this precipitation never takes place in the first quarter 


| 


THEORY OF THE HENDERSON-HASSELBALCH EQUATION 225 


of an hour so that a preliminary treatment of the haemoglobin solution with 
CO, may be undertaken for this space of time and the experiment continued 
with lower CO, tensions (lower ay"). This crystallising out of horse haemo- 
globin will be reverted to in chapter XI. 

In horse blood haemolysed by freezing (concentrated in the centrifuge) 
I have only once seen a similar crystallisation, and the haemoglobin solution 
was in this instance cooled to 0°. 

That haemoglobin very readily crystallises out at high CO, tensions has 
been known a long time and is mentioned for example by Preyer [1871] 
without any particular comment. I. Setschenow [1879, p. 48] reported similar 
observations with strong concentrated frozen horse blood at room temperature 
(CO, and H,SO, addition). 

As already repeatedly mentioned the potential in an electrical determina- 
tion of reaction falls when the platinum electrode has been in contact with 
protein solutions for some time. The drop is not large and in the course of 
2-3 hours an almost constant potential seems to be reached (within $ millivolt), 
but I tried nevertheless to avoid any possible error from this cause (“‘ deteriora- 
tion” of the electrode) by developing the technique employed with L II 
described earlier in this work. In using this electrode vessel, in which it will 
be remembered the platinum electrode does not come in contact with the 
haemoglobin before the potentiometry is started, the potential was found to 
rise quickly about 10-20 millivolts in the first quarter of an hour after contact 
(total immersion) was established but quite irregularly. Then it became 
constant for a time and afterwards slowly declined. When the platinum was 
heated to redness before platinising the rise was much less, but as a rule a 
few millivolts. This must be what Parsons [1917] referred to when he wrote 
that it is essential to heat the electrode red hot before every determination in 
haemolysed blood. 

In Table XXII pA») is calculated from Hasselbalch’s experiments with 
dialysed haemoglobin in weak sodium bicarbonate solution the conversion 
being carried out in the same way as in the preceding chapters. 

PAm) 18 in agreement with pA,) and pA) and therefore in hase 
blood we have | 


PXm) = Pu? + log 700 eh. ds ees ry (134) 
where f is the mean concentration of sc atbicsd CO, (expressed in vols. % 
CO,) in the haemolysed blood. 

In Table XXIII+ a number of determinations of pA;,,) in haemolysed ox 
blood are given. They were done with the small saturation electrode and within 
the same period as the experiments with blood designated series A in chapter V. 


1 The temperature corrections here and in what follows are made by adding 0-0075 to the 
value found for each degree over 18°, and subtracting the same amount for each degree under 18°. 
In the calculations from experiments in chapters V and VI 0-005 was used as the correction, but 
the difference is so small that I have not found it necessary to recalculate these earlier experi- 
ments with the correction employed in this chapter. 


226 KE. J. WARBURG 


It is extremely probable that the apparent hydrogen ion exponent at this 
time was 0-06 too low, according to which pAim) in haemolysed ox blood with 
about 33 vols. % combined O, should be about 6-27. 

The determinations in Table XXIV of haemolysed ox blood with 29 vols. 
°%, Og at 20° and 38° give respectively values of 6-26 and 6-15. They belong 
to series B and are carried out with the small saturation electrode. 

The experiments in Table XXV were made in the large saturation elec- 
trode LI. In almost all cases a preliminary treatment with CO, for }-} an 
hour was undertaken. Analyses of two gas mixtures are given in the table 
corresponding to one measurement of potential, the first relating to the last 
gas in the spirometer and the second to the gas in the electrode vessel. 

Pm) at 18° referring here to horse blood cells haemolysed by saponin with 
about 38 vols. % combined O, is found to be 6-32; for ox blood haemolysed 
by freezing with 35 vols. % Oy, it is 6-26. 

In Table XXVI measurements are given carried out with L II without 
heating the platinum electrode red hot before platinising. The results are 
quite in agreement with those reported in Table XXV, as pA) in horse 
serum haemolysed blood cells with about 40 vols. % combined O, is also 
6-32, while in haemolysed ox blood with about 32 vols. % it is 6-27. It is 
worth noticing that the final potential is not appreciably different although 
the platinum plate with one technique was 2-2} hours in contact with the 
protein and with the other technique only }-} hour. I should however expect 
that with a sufficiently extensive series of measurements a deviation of 1-2 
millivolts might be demonstrated. 

In Table XXVII are given measurements carried out in L II with platinum 
plates freshly heated to redness. The results are as will be seen a little different 
from those just obtained, pA;,) in haemolysed horse blood and ox blood of 
similar constitution being respectively 6-35 and 6-28 at 18°. These values 
differ but slightly from those of the first series and the difference is hardly 
greater than the experimental error, a calculation of the mean error as in the 
case of the experiments with serum being hardly feasible. 

It will be remembered that pA.) is 6-29 (see chapter VI) and since prim) 
in a mixture of one part serum and three parts haemolysed blood corpuscles is 
6-35 we shall not make a large error by assuming pA) — pi.) is 0-07, but as 
this difference is rather uncertain calculations have been made using (133) 
with values ranging from 0-05 to 0-09. The results are to be found in Table 
XXVIII. In the last column the values for (132) are given, the difference 
here being 0-07. 


Table XXII. pA) in dialysed haemoglobin with NaHCO, 0-025n 
calculated from K. A. Hasselbalch’s experiment at 38°. 


mm, Hg Vols. % combined PH 
CO, CO, corrected Pm) 
20-2 42-2 7-59 7:13 
94-2 55-4 7-06 716 


70 32-9 7:95 714 


THEORY OF THE HENDERSON-HASSELBALCH EQUATION 227 


Table XXIII. pAgn) in mixtures of serum and blood cell fluids (freezing) 
determined by the small saturator electrode. Series A. 


mm. Hg Vols. % combined 
Ree rege Basia aya 


Temp CO, O, CO, O, PH Pum) 18° 
19-5 459-9 05 117-3 0-6 6-51 6-19 Ist day 
19-0 77-7 0-5 70-5 0-7 7-09 6-21 
19-0 + 73-7 0-4 69-8 0-7 7-12 6-23 Ox blood 
18-0 50-2 0-9 60-6 — 7-21 6-21 2nd day 
18-0 50-2 0-9 63-0 0-8 7-19 6-18 Preliminary treatment with CO, 
18-0 28-6 0-5 50-2 1-8 7-36 6-18 
19-5 28-6 0-5 49-5 0-6 7-35 6-20 - ” 
Colorimetric 32 
19-0 37-4 0-5 61-1 0-7 7-33 6-20 Ist day 
19-5 57-1 0-5 68-6 1-4 7:23 6-23 
19-5 57-1 0-5 70-2 0-5 7:19 6-19 Preliminary treatment with CO, 
19-0 = 131-9 0-5 90-3 1-6 6-97 6-22 Ox blood 
18-0 = 575-7 05 136-2 0:3 6-49 6-20 2nd day 
19-0 82-6 0-5 79-1 0-8 7-12 6-22 
18-5 82-6 0-5 82:7 0-7 7:13 6-22 Preliminary treatment with CO, 
18-0 14:3 0-5 40:7 13 7:58 6-21 
Mean 6-21 
20-0 38-6 =: 147-0 51-9 32-2 _— — 3rd day 
20-0 3144 3208 103-2 34-0 _— — 
19-0 =: 103-5 0-5 77-4 1-1 7-08 6-29 Ist day 
19-0 40:3 0-2 — — 7:33 — 
19-0 46-7 0-3 64-5 — 7:28 6-22 Horse blood 
19-0 = 481-5 0-0 128-2 0-9 6-58 6-24 2nd day 


Colorimetric circ. 32 


Table XXIV. Washed ox blood cells haemolysed by freezing circ. 225 cc. 
+ 6 cc. n/3 Na,CO,. Determinations of pA.) in the small saturator 
electrode. Series B. 


mm. Hg Vols. % combined 
Creer PH Pm) Pm) PH 
Temp. co, O, co, O, (measured) 20° 38° = (calculated) 

27. ii. 19 
38 589-2 0-9 80-8 0-4 -- a -— 6-42 
20 615-8 1-0 101-5 0-7 — —_ —_ 6-41 
38 114:3 0-3 47-2 0-6 6-90 — 6-15 6-90 
20 119-5 0-3 67-2 0-5 6-95 6-26 -- 6-95 
38 64-7 0-5 37-3 0-2 7-05 —_ 6-15 7-05 
20 64-6 0-5 55-4 0-7 7-15 6-27 — - 7-14 

28. ii. 19 
38 121-7 0-6 49-1 0-5 6-88 -- 6-13 6-90 
21 127-2 0-6 69-2 0-4 6-93 6-25 -- 6-93 
38 12-1 0-5 13-7 0-7 7-32 —_ 6-14 7-33 
20 12-1 0-5 27-4 0-4 7-53 6-24 — 7-55 
38 41-0 0-4 29-8 —_ 7-09 _ 6-09 7-15 
.20 47-4 0-4 49-3 0-8 7-21 6-26 — 7-21 

Mean 6-26 6-15 

1. iii. 19 
38 36-2 140-0 22-8 28-4 -- _ 7:10 
20 37°8 146-5 38-6 29-3 _ _ —_ 7-20 
38 112-6 565-3 42-2 29-0 — —_ — 6-87 
20 117-6 590-6 62-8 —_— —_ — — 6-92 
38 501-3 199-1 78-4 27-2 — — _— 6-39 


228 


E. J. WARBURG 


Table XXV. pA,p) in mixtures of serum and blood cell fluids estimated 
with LI. Series B. 


Temp. 


19-5 
20-5 
19-0 
20-0 
19-0 
18-0 
19-0 
18-0 
18-0 


18-0 


mm. Hg Vols. % combined 

prs 

co, O, co, O, 
46-9 0-5 — — 
46-5 0-6 64-4 0-3 
106-2 0-7 — — 
105-4 2-3 865: 1-9 
29-4 0-8 —_ — 
29-4 0-9 50-4 0-9 
Colorimetric 32 

40-2 0-5 — = 
40-4 1-2 64-7 1-2 
16-6 0-2 — — 
15-3 1-2 40-7 2:1 
12-3 0-5 — nae 
12-9 1-1 38-4 1:3 
Air from blower 37-6 

48-4 0-2 _— en 

47-7 0-7 66-7 ? 
586-5 0-2 — — 
586-5 0-4 144-9 0-2 
60-8 0-5 — —_— 
66-4 0-5 85-3 0-2 
27-2 0-2 — — 
26-7 0-7 64-1 0-2 


Air from blower 35:1 - 


er Pie 
7:39 633 
7-13 630 
7-48 633 
"7-48 635 
765 631 
7-67 6-28 

Mean 6-32 
740 «6-34 
6-57 627 
7:28 6-26 
755 6-26 
~ Mean 6-26 


lst day 
2nd day Horse blood haemo- 


lysed with saponin 
3rd day 


lst day Horse blood haemo- 
lysed with saponin 
2nd day 


3rd day 


Human blood, frozen 


Ist day Concentrated by 
centrifuging 
Ox blood, frozen 


2nd day 


Table XXVI. Determinations of pA;,) in mixtures of serum and blood cell 
fluids made with LIT. The electrode was not heated to redness before each 
measurement. Series B. 


Temp. 


21-0 


20-0 


mm. Hg Vols. % combined 

SS 

co, O, Co, O, 
470 Or a a 
47-0 2-9 60-7 1-4 
258°3 0-1 — —_— 
288 O02 M30 — 
74-2 0-1 _ _ 
73-7 1-6 75-9 0-5 
45-6 0-1 — — 
45-2 1-2 62:1 0-5 
334-7 0-3 — — 
335-2 0:3 122-1 0-1 
Air from blower 39-7 
24-6 58 43-4 3:3 
82-2 0-3 — _ 
31-6 38 48-8 21 
101-3 0-6 _— — 
95-0 0-8 746 0-4 
305-2 O1 _ _ 
305-2 7 104-4 0-3 
Air from blower 31-8 


Haematocrite number 72-8 


Pm) 

PH 18° 
7:35 6-31 
689 6-33 
7:23 6-30 
7:40 6°33 
678 631 
Mean 6-32 
747 +630 
738 627 
706 6:28 
oo en 
Mean 6:27 


Ist day 
Horse blood, frozen 
2nd day 
Ist day 
2nd day Ox blood, frozen 


3rd day: 


THEORY OF THE HENDERSON-HASSELBALCH EQUATION 229 


Table XXVII. Determinations of pA;,,) in mixtures of serum and blood cell 
fluids made with LII. Freshly burnt out platinum electrode. 


Temp. 
19-0 
19-0 
19-0 
19-0 
16-0 
16-0 
18-5 
18-5 
17-0 
16-0 
16-0 
16-0 


mm. Hg Vols. % combined 

Tog moet man 

Co, O, CO, O, 
33°7 0-0 — — 
33°3 1-9 53-0 1-5 
3753 Ol st aa 
375-0 0-1 117-6 0-1 
20-1 0-1 a — 
23-6 0-6 54-5 0-0 
60°3 0-0 — — 
61-6 0-0 75-1 0-0 
Air from blower 36-3 

29-6 0-1 —_ _— 
29-7 0-1 57-5 0-0 
Air from blower 41-6 

18-9 0:3 — — 
18-9 0-7 44:3 1-4 
343-1 1-0 — — 
342-1 2-4 120-7 1-2 
89-9 0-2 — _— 
90-5 1-7 80-8 0-0 
Air from blower 38-2 

388-9 0-1 — — 
388-4 13 110-8 0-6 
22-8 0-2 _ — 
22-6 2-4 51:3 1-2 
78-4 0-4 —_ — 
78-9 0:2 76:3 0-1 
215-1 0-2 — — 
214-5 0:6 99-8 0-1 


Air from blower 27:8 


oe 
7-50 6-38 
6-73 6-32 
7-63 6-33 
736 6-37 
750 631 
757 6-30 
685 639 
725 6-39 

Mean 6-35 
6-70 6-34 
751 6-25 
717-628 
682 6-25 

Mean 6-28 


Ist day 
3rd day 


4th day 


Horse blood, frozen 


Horse blood, frozen 


Horse blood, frozen 


Horse blood, frozen 


Ox blood, frozen 


Table XXVIII. The difference of reaction between blood cells and serum 
(horse blood at room temperature). 


PAs) — PXe) 
PH'(s) —log D 
6-50 +0:055 
6-88 +0-097 
7:27 +0°155 
7-40 +0-180 
7:60 +0-222 
7:90 +0-301 


—0-05 — 0-06 — 0-07 — 0-08 —0-09 — 0-07 
PH (s) ~— PH (ey CH) 
tS Fees et AF (s) 
+0-005 — 0-005 —0-015 ~- 0025 -—0-035 0-966 
+0-047 +0-037 +0-027 +0-017 +0-007 1-06 
+0:105 +0-095 +0-085 +0-075 +0-065 1-22 
+0-130 +0-120 +0-°110 +0-100 +0-090 1-29 
+0°172 +0-162 +0°152 +0-142 +0-130 1-42 
+0:251 +0:241 +0:231 +0-221 +0211 1-70 


A dotted line is drawn through the table which indicates the reactions 
at which blood cells and serum have the same reaction. It will be seen this 
is the case between 6-50 and 6-88. Fig. 11 is a graphic representation of the 
table, the apparent hydrogen ion exponents of serum being the abscissae and 
the differences between the exponents of serum and blood cells the ordinates. 
For clearness only the curves relating to pA.) — p,) 0-05, 0-07 and 0-09 are 


given. 


230 ‘BE. J. WARBURG 


It will be further seen from the tables and curves that the apparent hydrogen 
ion activity is larger in blood corpuscles than in serum at serum reactions 
more alkaline than py: = 6-9 (ag: = 1:26 x 10-7) and that the difference in- 
creases with the hydrogen ion exponent. 


PH'(s) — PH'(e) 
iad ony ers ei es aa ARE 
0-24 : 
0:23 é ve 
0-22 F- ae] 
021-- Sad 
0-20 
0-19-— ra 


\ 
0-07} . a. 
0-06 |— a 
0-05 


: ‘ 
0:03 }— ss = 
0-02 }— aid 
001-- Ae ~ 
Il 1 
0-00 


1 1 
‘ Dir. 
ee eat WS Sac {hoe aay ee 


80.79 76 77 76 75 74 73 72 71 70 69 68 67 66 6S 64 
Fig. 11. pyr) — PH 

As D in human blood at 38° is the same as D in horse blood at 18° the 
above considerations will also apply to this species of blood if pAq)— pA) is 
the same. In the tables there is only one measurement of human blood at room 
temperature to be found but this fits in quite well in the horse blood series. 
The question can however only finally be settled by many estimations. 

According to Fr, Kraus’ [1898] experiments D for ox blood appears to be 
1-00 at py about 6-8. When this is compared with the determinations of 
PX») in Ox blood the same reaction should exist in serum and blood cells 
at this reaction because pA.) and pA) are very nearly identical in such blood. 


THEORY OF THE HENDERSON-HASSELBALCH EQUATION 231 


It appears also from the few determinations of D for ox blood which are 
reported in a preceding chapter that the difference of reaction between blood 
cells and serum increases with the py. 

If we assume that the dissociation of electrolytes which varies with the 
reaction is the same at a similar reaction before and after haemolysis we can 
draw conclusions from the combination of CO, at the same CO, tension, before 
and after haemolysis, about a possible difference of reaction between blood 
cells and serum. The combination of CO, increases with the apparent hydrogen 
ion activity in serum and in the fluids of the blood cells but it increases most 
in the latter case. If the same py: persists after haemolysis as there was in 
serum and blood cells before haemolysis (e.g. in horse blood at py 7-60) then 
the CO, combination is not altered. If however there was a higher ay: in blood 
cells than in serum before haemolysis, a,° will be midway between the original 
reaction of the serum and blood cells under the given conditions after haemo- 
lysis and less CO, will be combined with the electrolytes varying with the 
reaction in blood cells but more with those in serum. The result will be that, 
altogether, less CO, will be combined after haemolysis than before if the 
volumes of blood cells and serum are equal. 

I have only made a few experiments (about ten) with ox blood with this 
object in view but they all go to show that less CO, is combined after haemo- 
lysis than before, at alkaline reactions. At reactions about 6-30 the combined 
CO, was almost or actually the same in blood before and after haemolysis. 
In some experiments in which the osmolar concentration of the blood was a 
little diminished the difference was smaller than in ordinary blood. In an 
isolated experiment where the osmolar concentration of the blood was rather 
increased the opposite was the case. These effects of the changes in the 
osmolar concentration are in accordance with the theory, as D varies with the 
variations in volume of the blood corpuscles. 

As will be noticed there is a disagreement, although a small one, between 
the results attained by the first and last mentioned principles for the deter- 
mination of the difference of reaction in ox blood. From the first principle 
we concluded that the reaction was identical in serum and corpuscles at 
pg’ 6:8 while the last pointed to the fact that there was no difference as far 
as py 6-3. Further experiments are needed to clear up the matter. 

In Table X XIX some of the experiments! mentioned are given. Haemo- 


1 The experiment at 38° was particularly interesting as the relation between oxy-haemoglobin 
and reduced haemoglobin is not altered by haemolysis. This result was supported by several 
ee with saponin haemolysed blood which I pave. Wier to have the opportunity of 
publishing. The phenomenon itself is not without interest because it indicates that the quantity 
and kind of salt does not play so great a part in determining the form of the O, combination 
curve of haemoglobin as Barcroft [1914] and his collaborators imagined. That the dilution of 
the haemoglobin due to the haemolysis plays no great part in the relation between oxy-haemo- 
globin and reduced haemoglobin was only to be expected because they are both diluted to the 
same degree and there should thus be no change in the extent of oxygenation of the haemoglobin 
either according to the interpretation of the process expressed by G. Hiiffner [1901] in his later 
papers or by A. V. Hill [1910, 1913, 1921}. 

If we could estimate the degree of oxygenation with sufficient accuracy we might however 
expect to find a little greater oxygenation after haemolysis than before at reactions more alkaline 
than py 6:8 because the extent of oxygenation is a function of the reaction; cf. Chr. Bohr, 
K. A. Hasselbalch and A. Krogh [1904], R. A. Peters [1914, 2], K. A. Hasselbalch [1916, 2] and 
L. J. Henderson [1920]. 


232 
mm. Hg 
eS 
co, O, 
14-9 151-6 
14-9 151-6 
53-0 143-3 
53-0 143-3 
144:3 163-8 
144-3 163-8 
164-5 313-2 
164-5 313-2 
467-6 256-2 
467-6 256-2 
731-5 trace 
731-5 e 
10-7 153-0 
10-7 153-0 
34-1 148-1 
34-1 148-1 
67:3 141-3 
67-3 141-3 
135-3 127-0 
135-3 127-0 
133-0 126-4 
133-0 126-4 
405-9 69-9 
405-9 69-9 
736-0 trace 
736-0 > 
17-4 152-1 
17-4 152-1 
98-4 135-1 
96-1 135-6 
430-4 64:8 
429-2 65°3 
37-2 150-4 
37-2 150-4 
37-2 150-4 
37-2 150-4 
153 144-5 
15-3 144-5 
42-3 138-6 
42-3 138-6 
114-6 124-2 
114-6 124-2 
213-7 100-1 
213-7 100-1 
4556 48-0 


E. J. WARBURG 
Table X XIX. 


Vols. % combined CO, Vols. % combined O, 
A A 


SRR pe ee ee, | 
in blood inhaemoglobin in blood inhaemoglobin calculated 
Blood and water-haemolysed blood at 19° 


24-1 = 15-0 oe 7:25 
— 21-6 _— 14-6 — 
— 35:8 — 14-4 — 

37-5 —_ — —_ 6-98 
—_ 47-3 — 14-4 — 
48-6 — 14:5 — 6-73 
50-5 = 14-6 — 6-29 
— 48-5. — 14-7 — 
— 60-5 — 14:5 oa 
61-6 -- 14:3 -— 6-32 
— 64-9 — -- — 
67-0 — 1:3 1-4 6-15 

Blood and water-haemolysed blood at 19° 
25-9 _- 8-0 — 7-60 
obi 24-1 — 8-1 — 
35-4 _ 8-0 _ 7-22 
a 32-9 ~- 7-9 = 
40-0 _— — — 6-99 
a 38-7 — 8-2 — 
46-1 — 77 — 6-75 
— 44:8 — 7-4 -- 
45:8 — — — 6-76 
ei) 44-6 a Pa gine 
54:3 _— — _— 6-33 
a 54-3 ike eS tea 
59-4 —_ 3:3 — 6-13 
anes 60-0 od 16 — 

Blood and blood haemolysed by freezing at 18° 

47-1 _- 16-0 —_ 7-65 
-— 45-1 —_— 15-6 — 
ads 72:8 mee wim les 

72-1 — 15:8 — 7-07 
— 96-5 _- 12-9 _ 

96-3 -- 14-0 — 6-50 

Blood and water-haemolysed blood at 18° 
oe 28-0 — 12:9 _ 

30-1 ~- 12-9 — 711 
— 28-7 — 13-0 _— 

30-2 ~- 12-8 — 7112 

1 CO, preliminary treatment for half an hour. 
Blood and water-haemolysed blood at 38° 
i 12-9 -— 81 — 

13-4 — 7:8 — 7:25 
— 19:5 _ 77 -— 

20-2 _ 7:4 _ 6-98 
— 27-0 -— 71 — 

27-5 — 74 — 6-62 

32-6 _ 62 — 6-46 
Sins 32-0 — 6-2 —., 

401 — 2:3 _- 6°23 
= 413 _ 2:2 — 


THEORY OF THE HENDERSON-HASSELBALCH EQUATION 233 


lysis was brought about either by freezing or by the addition of water. When 
haemolysis by water was completed sufficient NaCl was added to make the 
salt content up to 0-9 % again. 
I. Setschenow [1879, p. 44] reported the following experiment at 37°-37°-5. 

Emulsion of dog blood corpuscles: 

514-6 mm. CO,, total CO, in 50-18 cc. = 50-31 ce. (0°, 1 mtr.). 
After freezing: 

513-3 mm. CO,, total CO, in 50-18 cc. = 49-98 ce. (0°, 1 mtr.). 


From this py is 6-22 if ‘Y= 15 and pA, = 6-20, at which reaction there 
seems to be the same reaction in dog blood cells and serum. 


RésumMs. 


The apparent hydrogen ion activity in horse blood corpuscles has been 
determined. 


CHAPTER VIII 


THE DETERMINATION OF THE FIRST DISSOCIATION CONSTANT 
OF CARBONIC ACID AND THE DEVIATION COEFFICIENTS 
OF THE BICARBONATE ION. 


In the foregoing chapters we have determined the value of pA;,) and pA,,), 
and it will now be interesting to inquire into the factors which control these 
constants rather more closely. 

It will be remembered that in chapter III it was shown that the apparent 
activity coefficient of the continuous phase of serum and of the blood cell 
fluid participates in the constants and a rather large number of determinations 
have therefore been performed of the apparent activity coefficients in salt 
solutions. I have at several points pursued the investigations further than 
was absolutely necessary for the problem being dealt with, because it may 
be of particular interest from a purely physico-chemical standpoint, especially 
since the appearance of Bjerrum’s theory. 

The first dissociation constant of carbonic acid has been determined by 
the conductivity method by J. Walker and W. Cormack [1900] and by 
J. Kendall [1916] on the basis of experiments carried out by himself, by 
Pfeiffer’, by Knox! and by Walker and Cormack. The determinations of 
Walker and Cormack and Kendall are better than the earlier ones (Pfeiffer’s 
and Knox’s) and therefore the values calculated from them are the most 
valuable. The molecular conductivity of the bicarbonate ion at “infinite 
dilution” comes into the calculation. This value is obtained by extrapolation 
from conductivity determinations of sodium bicarbonate solutions (and also 
calcium bicarbonate solutions), but it seems to the author that Walker and 
Cormack and Kendall have not executed this extrapolation in a satisfactory 

! Cited from Kendall [1916]. 
Bioch. xvi ‘16 


234 E. J. WARBURG 


manner. A slight error in the extrapolation is of hardly any consequence in 
the determination of the first dissociation constant of carbonic acid because 
the molecular conductivity of the bicarbonate ion has to be added to the much 
greater molecular conductivity of the hydrogen ion and it is the sum of these 
quantities which is used as a factor in the calculation (see chapter I). But if 
it is required to determine the conductivity coefficient an error in p,, will 
have a great effect particularly in weak solutions. The question as to how 
#,, Should be obtained by extrapolation from conductivity determinations 
of salt solutions has up to the present been the subject of much controversy 
but I am unacquainted with any well-grounded theoretical method of per- 
forming the extrapolation. The best way must therefore be to obtain a 
relation between a function of the salt concentration and a function of the 
molecular conductivity which is the equation for a straight line. We can then 
determine y,, either graphically or by the method of least squares. . 

It will as a rule be much the most convenient method to perform the 
extrapolation graphically but the method of least squares has the advantage 
of giving an estimation of the accuracy of the determinations. 

That I only occasionally estimate the error in this chapter is because the 
value obtained is only of real significance if a systematic error can be excluded 
and a determination of the error can only be used in comparing results obtained 
by the same experimentalist with the same technique in a similar process. 
In the majority of cases the graphic method carried out in’a suitable way, 
coupled with a good idea of the accuracy of the technique, will give more 
valuable information than the more arduous determination of the error. 

Kohlrausch (cited from Lehfeldt [1908, p. 61]) recommends that the cube 
roots of the concentrations be plotted as abscissae and the molecular con- 
ductivities as ordinates. The curve will] then be a straight line for many salt 
solutions. This kind of extrapolation has proved very suitable in the case 
of Walker and Cormack’s determinations but particularly so for Kendall’s 
measurements of conductivity in sodium bicarbonate solutions. 

In Table XXX Walker and Cormack’s and also Kendall’s results are given. 
In the first column the number of litres in which a gram equivalent is present 
are recorded; in the second column, the concentration in terms ofnormality ; 
in the third the cube root of this amount; in the fourth, the equivalent con- 
ductivity. In Fig. 12 these values are plotted, </c as abscissae and p, ex- 
pressed in reciprocal ohms as ordinates, The ordinates which refer to sodium 
bicarbonate are on the left, the numbers referring to the calcium bicarbonate 
series are on the right. It will be seen from the figure that the curve which 
represents Kendall’s measurements of sodium bicarbonate solutions corre- 
sponds extremely well to a straight line. Extrapolation can therefore be 
undertaken with very considerable certainty and it gives p= 992 re- 
ciprocal ohms (25°), while Kendall himself made the extrapolation to be 
97-5, From this—using the same corrections as Kendall did—j,, at 18° is 
86°3 reciprocal ohms, while Walker and Cormack’s determinations on extra- 


THEORY OF THE HENDERSON-HASSELBALCH EQUATION 235 


polation give 88-1 reciprocal ohms. The figure shows that the extrapolation 
of Walker and Cormack’s measurements is rather uncertain, the two estima- 
tions in the most dilute solutions being a little high. The want of agreement 
between Walker and Cormack’s and Kendall’s results is presumably due to 
this cause and I shall regard Kendall’s as correct. 


NaHCO; 
Be } Ca (HCO,)s 
co Oo 22 GR ph SSA Py 5 ia SS OR A eR: ey een Tet ae am Ma em a pa Cm 
<a — 
2% bain el Ke 
94 —4109 
rN Pins 
92}~ age 107 
90}— NY {105 
88 +. 
A ig 1 
86/— + — 101 
84}— epee — 9 
82)—- ee ALL 
oo Ng NS 
78h—- +4 93 
tg - 
ag Ny Sista 
m4 — 89 
4 Xn 
72 87 
70}- sie — %& 
66;— 3, | 81 
‘eh Ley Wa 1 Aa) Pe | BL ME eT b ieee | 172, l 
46 


ha 4 
0:00 -02 04 -06 08 10 12 14 16 +18 -20 -22 -24 -26 -28 -30 -32 34 -36 38 40 “42 -44 « 
Fig. 12. Conductivity of sodium bicarbonate solutions. 


Upper curve from Kendall’s measurements; lower curve from Walker 
and Cormack’s measurements, 
4 $ calcium bicarbonate. 


The curve of the calcium bicarbonate experiments (triangles) demonstrates 
that only the three experiments with the strongest solutions lie on a straight 
line—the experiments with the two weakest solutions lie considerably above. 
Extrapolation, having regard to these two points, gives quite an unreasonable 
value for u,,, and I therefore think it cannot be performed with these experi- 
ments. 

If we calculate the molecular conductivity of carbonic acid from Kendall’s 
experiments with sodium bicarbonate (carbonic acid being regarded as mono- 
basic) we get at 25° 395-5, 


,, 18° 355:8, 
, 0° 265:6. 


Using these figures, values for the first dissociation constant of carbonic 
acid are obtained which diverge in the third decimal place from those calcu- 
lated by Kendall [1916]. In Table XX XI Walker and Cormack’s [1900] and 
Kendall’s results are recalculated. Pfeiffer’s and Knox’s experiments are not 
given and the reader is referred to Kendall’s paper for these. The small 


16—2 


236 E. J. WARBURG 


difference between Kendall’s value for p,, for carbonic acid and mine is 
without real significance in the calculations from these experiments. 

The mean of Walker and Cormack’s and Kendall’s determinations at 18° 
of the first dissociation constant of carbonic acid is 3-1 x 10-?, from which 
Px, = 6509. At 25° we obtain from Kendall’s experiments K, = 3:47 x 10~’, 
Px, = 6-460, and at 0°, A, = 2-21 x 10-7, pg, = 6-656; pg, therefore de- 
creases 0-20 between 0° and 25° which is equivalent to 0-008 per degree. 

By thermodynamic methods it is possible to determine the change in px, 
with temperature when the heat of ionisation of carbonic acid is known. 
Julius Thomsen has estimated it at 2800 calories, from which the change in 
Px, can be calculated by van ’t Hoff’s equation (cf. Henderson [1909], Hassel- 
balch [1916, 2] and Kendall [1916]), and turns out to be 0-0065 per degree, 
which is in good agreement with Kendall’s direct determination. Kendall 
has pointed out that the heat of hydration of CO, participates in Thomsen’s 
determinations and that the calculation of the change of the dissociation 
exponent with temperature according to van ’t Hoff’s equation is only correct 
if the degree of hydration is independent of the temperature. He thinks 
however he is justified in concluding, from the good agreement between the 
calculated and experimental values, that the calculation is permissible—that 
is to say that the change in the degree of hydration is negligible in this con- 
nection. 

From Fig. 12 it appears that the relation between the molecular conduc- 
tivity of sodium bicarbonate and the concentration of the salt is expressed 
by the following equation: 


HOR NES O88 ei Ra ad ae (135) 
from which we get f= gp = 1 — B08 Ho. escasioersexeinaves (136) 


No general relation is known -between the conductivity coefficient of a 
salt and its activity coefficient but Hasselbalch, in his frequently cited paper 
(1916, 2}, has carried out determinations which permit of the estimation of 
the apparent activity coefficient of the bicarbonate ion. 

The experiments referred to were electrical estimations of hydrogen ion 
activity in sodium bicarbonate solutions of known concentration with simul- 
taneous estimation of the CO, tension with which the solutions were in 
equilibrium. From these experiments Hasselbalch determined pg, (Hassel- 
balch) with the help of the equations developed in chapter III, In Table XXXII 
these experiments will be found recalculated to give our px’, and, as before, 
the values corrected for depressed hydrogen tension as well as the uncorrected 
are given. Hasselbalch’s experiments were performed by his own method 
with a plate electrode, For the sake of clearness we will postpone the further 
examination of these experiments, 

Hasselbalch’s experiments were carried out, as repeatedly stated, on the 
principle he evolved himself. I began therefore by repeating these experi- 
mente with the earlier deseribed technique on Hober’s principle, widening 


THEORY OF THE HENDERSON-HASSELBALCH EQUATION 237 


the scope of the investigation by including sodium bicarbonate solutions 
which in addition contained NaCl or KCl in the concentrations 0-0718n, 
0-145n, and 0-291n. These experiments are given in Tables XXXIII- 
XXXVIII and belong to series A but were done early in this series. I do not 
think they are subject to any systematic error but I cannot express myself 
with absolute certainty on this point. 

In the first column of the tables the constitution of the solutions is given. 
The bicarbonate content of the solutions was estimated with the exhaust 
pump, at least two samples being used and usually several more. The error 
of this method is negligible compared with that of the electrical method. 
NaCl and KCl were added to the solutions in measured quantities, the salts 
being washed from a watch-glass into a measuring flask containing the bi- 
carbonate solution. In a number of cases NaOH was used instead of NaHCO, 
and the solutions were therefore treated with CO, for a quarter of an hour 
before the actual saturation. It was impossible to demonstrate any difference 
between solutions prepared with the bicarbonate and the hydroxide. 

In calculating the combined CO, (B), dissociation of the bicarbonate into 
the monocarbonate is taken into account. 

In accordance with what was said in chapter II we have 

Ges: 
Si= gaz B, 
where S, is the amount of combined CO, corresponding to the bicarbonate 
expressed in vols, %, and B is the total amount of combined CO, expressed 
inthe same way. _ 

Since S, is the amount of combined CO, corresponding to the mono- 
carbonate, and 7 is the amount of CO, in vols. °%% which would be combined 
if the combined CO, was in the form of bicarbonate, we have 


S,+28,=T, Terrrrrrerrerrerrerrrrriry (137) 
from which with the aid of (102) we get 
WH 4) 
Bree Fe aici hcecscenagsss (138) 
e+? 


When B and ay: are known, therefore, 7 can be calculated and inversely 
B can be calculated from T and ay. In the pumping out experiments for 
the determination of Cyco’,, B was put equal to 7 when py: was below 7-70; 
if it was above, (138) was used in the calculation, K’, being taken as 10-". 
In calculating pg’,, 7 was put equal to B when py: was less than 7-90; when 
greater, (138) was used. In these last experiments equation (103) was also 
used instead of (95) (in logarithmic form) for calculating px’. 

In Table XX XIX the results of this series of measurements are compared. 
They show that px’, undoubtedly decreases with increasing salt concentration 
and that the effect of the sodium ion on the constant is more pronounced than 
that of the potassium ion. The measurements in pure sodium bicarbonate 


238 E. J. WARBURG 


solutions in the case of the two weakest concentrations (7' = 8-88 and T = 17-96) 
are uncertain because the conductivity of these solutions was so slight that 
the resistance in the exit tap was very great and the oscillations of the capillary 
electrometer were too small. 

In Table XX XVIII there are a number of determinations of pg’, at room 
temperature in sodium bicarbonate solutions with the addition of equivalent 
amounts of NaCl or KCl. The experiments with the two series of salt solutions 
were performed on the same day and with the same electrode and they are 
therefore as comparable as possible, with the technique employed. 

In Tables XL and XLIV experiments. belonging to series B are given. 
Px’, is here determined for sodium bicarbonate solutions partly with the addi- 
tion of salt, at room temperature and at 38°. 

In Table XLI measurements are given, carried out with the Hasselbalch 
electrodes EIII and E VI, employed as Héber electrodes in the manner de- 
scribed in chapter IV, at room temperature. I consider these last measure- 
ments to be the best of all fromm a technical point of view, and I shall examine 
them a little more in detail. In the first place it will be observed that px’, 
for pure sodium bicarbonate solutions decreases with the concentration. 


n/100 NaHCO, gives px’, 6-400 


n/40 s: > 6898 
n/20 i aria Ss 
n/10 : ol OS BEG 


In potassium bicarbonate solutions for 
n/100 KHCO,, px’, = 6-413 
n/10 53 » =6317 
If we compare the solutions with salt concentrations of n/10 we get 
n/10 NaHCOs, px’, = 6:279 
n/10 KHCOg, pg’, = 6°317 
n/20 NaCl + n/20 NaHCOsg, px’, = 6-297 
n/20 KCl + n/20 KHCOg,, px’, = 6-312 
n/10 NaCl + n/1000 NaHCOg, px’, = 6-313 
It is seen therefore that px’, for n/10 sodium salt solutions is rather less than 
the constant for n/10 potassium salt solutions. This is still more noticeable 
in the determinations in n/5 salt solutions, and from the latter it also appears 
that px’, for mixtures of Na and K salts lies between those for the pure salts. 
n/10 NaCl + n/10 NaHCO, gives px’, = 6-246 
n/10 KCl + n/l0 KHCO, © ,, », = 6-260 
n/10 KCl + n/10 NaHCO, _,, »» = 6-282 
Turning again to the n/10 solutions it will be noticed that px’, seems to 


be a little greater for solutions containing NaCl than for those containing 
equivalent amounts of NaHCO,. The difference in py’, for solutions containing 


THEORY OF THE HENDERSON-HASSELBALCH EQUATION 239 


pure NaHCO, and for those containing equal parts of NaCl and NaHCO, is 
however so small that it certainly falls within the experimental error, but the 
difference between n/10 NaHCO, and n/10 NaCl + n/1000 NaHCO, appears to 
be real. I think that in this case the method of least squares ought to be used. 
From equation (89) in logarithmic form we get 
Px’, = Px, + log Fy (HCO’s). 
Sree, to Bjerrum (cf. equation (58)) 
log fg (HCO’;) = — k Ve. 


If we put F, (HCO’;) = fa (HCO’s), 
which is permissible at n/10 concentration, we get 
Di i ESOS iy ios ir nets choses cesses: (139) 


where c is the molar concentration of the cations and k is the constant in 
Bjerrum’s above cited equation, while pg, is the negative logarithm of the 
first dissociation constant of carbonic acid. 

If we have a series of equations of the form 


ra en pe ee eh eee es (140) 
‘ we can calculate the constants which fulfil the conditions best with the 
following equations: Sal- Yar! 
= aaa (aj eter solcy Serugevées (141) 
D1 2a? - Tal = 
Y aaa ocr eeceteneeeseensenes (142) 


If we employ the above equations in connection with all the 26 experi- 
ments which only contain NaHCO, in the series recorded we obtain 
PK hae 6-512, 
k = 0-475. ; 
An estimate of the mean error of px’, can be obtained by putting the 
values found for it and & in (139), and ee the deviations from the 


above value, thus 
M= og ee x A = OO1TS. 


For a series of nine experiments we may ahh expect a mean error of 
0-0058. Thus for such a series of measurements with n/10 NaHCO, 

Px’, =.6:291 + 0-0058. 
In the experiments with the solution of n/10 NaCl + »/1000 NaHCO, 

Px’, = 6-313 + 0-0045. 
The difference is therefore 0-022 + 0-0073 and in all probability it is a real one. 

It is an obvious consequence of Bjerrum’s activity theory, as it has been 
expounded in chapter I and the present chapter, that the activity coefficient 
in a solution of a binary salt consisting of two monovalent radicles shall not 
vary! if the solution is diluted with another which contains an equivalent 
amount of a similar salt having no ions in common with the first. 
1 Apart from secondary effects represented by the difference in & in equation (46). 


240 E. J. WARBURG 


According to a view widely held by physiologists, which for example is put 
forward by Hamburger [1902], Hedin [1915] and Ege [1920], the dissociation 
of a monovalent salt will increase if the salt solution is diluted with another 
monovalent salt having no ions in common with the first, even if the two 
solutions are equivalent. This view depends however on an incorrect appli- 
cation of Arrhenius’ [1888] theory of “isohydric” solutions, but I was myself 
involved in it until Prof. Bjerrum kindly pointed out the fallacy to me. In 
view of the general acceptance it has received I will enter a little more deeply 
into the question regarding for the time being Arrhenius’ classical theory of 
the incomplete dissociation of salts as correct. 

Let us consider a /1 solution of NaCl and a n/1 solution of KI. 

Then according to Arrhenius’ theory extended to isohydric dissociation 


Cna'+Cor _ k 
Cnal? 
Sr abs als ee 
and Oa k. 


The equilibria in a n/1 KCl solution and a n/1 Nal solution will be given by 
nl yK" nfl yl _ 1 

m/l KCl(1 —7) ; 

nj/lyNanj/lyl _ k 

nj Nal(l-y) 

y being the (common) degree of dissociation of the salts. 

If we mix some NaCl solution with some KI solution, e.g. equal parts, an 
equilibrium will be established which will satisfy all the four equations. This 
is only possible at a quite definite degree of dissociation of the different salts 
as the determining equations can only have one solution. 

Assuming that the dissociation does not change on mixing the solutions, 
we have 


and 


Cnet = Cor’ = Cg: = Cy = n/2y, 
and Cyac = Cxr = n/2 (1 — y). 

Now NaCl and KI will react with each other and equal quantities of Nal 
and KCl will be formed if the same laws hold for all the undissociated salts 
in question. Thus we have the following equilibria: 

Cyaci = Car = Cnat = Cxoi = n/4 (1 — y), 
n/2y Na’ n/2yCV _ k 
n/4Nal(l-y) ~~ ™” 
n[2yK°n/2yV _ k 
naKl-y 
n/2yNa'n/2yV — k 
n/4 Nal (1 —-+) i. 
n/2yK'n/2 yor _ k 
njaKCl(l-y ~~ ™ 
which are correct in accordance with the four original equations. As only 
one state of equilibrium is possible the above mathematical reasoning is proof 
that no change in the dissociation of the salts takes place on mixing, 


THEORY OF THE HENDERSON-HASSELBALCH EQUATION 241 


Finally it may be stated that as far as the proof is concerned it is im- 
material whether the dissociation equations are really equal to a constant or 
only to a function of the total salt concentration, provided that all the 
equations are equal to the same quantity. 

Equation (139) is a straight line and corresponds to one in a rectangular 
coordinate system in which the abscissae are +/e and the ordinates px’,, the 
straight line cutting the ordinate axis at pg,, while the tangent of the angle 
made by the straight line with the abscissa axis measured in the second or 
- fourth quadrant is equal to x. In Fig. 13 ¥/c is plotted as abscissae and px’, 
as ordinates, the total cation concentration in the solutions which contain 
either sodium salts or potassium salts alone being calculated. Only the ex- 
periments for which 4/c is given in the tables are plotted in the figure. 

Hasselbalch’s experiments (the two lowest dotted curves) are seen to be 
about 0-09 lower in the coordinate system than mine, so that Hasselbalch’s 
Px, (18°) is 6-412, pg, (38°) is 6-302, pg, (18°) — px, (38°) = 0-11, while & 
is 0-52. 

The continuous line refers to the series of pure sodium bicarbonate solu- 
tions fully discussed above, while the uppermost dotted line is the one that 
best represents, as far as one can judge, all the experiments with sodium salts 
(the experiments with n/1000 NaHCO, are however omitted). This line gives 


Pr, = 6-514, 
k = 0-46. 


_ There is only slight uncertainty in drawing the line and without doubt only 
a small real error. It is worth noting that the equation 


— log F, (HCO’s) = 0°46 Ve wns se eeeeeeeeeees (143) 


appears to hold up to 0-4n with good approximation. 

The lowest dotted curve is parallel with the uppermost but 0-11 lower. 
It represents the determinations at 38° tolerably well. It will be observed 
that the change in the constant per degree is 0-0055 in Hasselbalch’s experi- 
ments and also in mine, therefore considerably less than the change found by 
Kendall [1916], namely 0-008, and rather less than that obtained by thermo- 
dynamic methods with Thomsen’s value for the heat of reaction, namely 
0-0065. The uppermost dotted curve represents the determinations of pg’, 
for pure potassium salt solutions. It gives pg, = 6-497 and k = 0:38. The 
value found for px, is presumably a little too low because it is hardly possible 
there can be any real difference between px, in sodium and potassium solu- 
tions, but there are too few determinations in low concentrations in the series 
to enable us to attribute the same importance to the constants for potassium 
as to those for sodium salts. 

From their experiments on the reaction of CaCl, solutions saturated with 
Ca (HCO), Bjerrum and Gjaldbaek [1919] calculated the “reaction constant” 
of calcium carbonate as 

log K = — 5-02 at 18°. 


E. J. WARBURG 


242 


96-0 080 920° 220 BHO 
| 


2t-0 68-0 f £e-0 pe | $2-0 02-0 | a te 


100-0 


1 l l it ve i 
tL Gb OL 89 95 $9 co 09 8S 9% $5 cc OS 8h SF th ch Ob BE GE be ce Of 8c 9c be ca Oc SI OF FI cl OF 8 SO tO a Jd 
fa Soe Sees Se eae a he Geer Coane iat fe Me ll: SRS DSA Wipes PP a Ean eed ols Wor a ed win Be eo car Soar icy a ae 9 lla ar 
bak nes 
ao yi 409 
— . 909 
— A - < 30-9 
are ee 
>. >. 
ie 
x 5 eS 
Bes ; ig 
se x 
+— Rowe 
or 
— x_™ 
~ gy monm ey me) 
= hee x 
Ps, Vv Stee 
We oo — — 
~. 
= i By Ns 
™. 
4 
ag v 
— ‘ . a 
i a7 "Sieg IN 8H 0= 4 
= @-~t Vv Sag: 
Sate v~, 


Vv 
+ — ae v 
~~ SG 
ra . 
-- awe 
en ba 
Se 
~ 
~ 


s 


THEORY OF THE HENDERSON-HASSELBALCH EQUATION 243 


This experimental determination they compared with a calculation of the 
reaction constant by means of the following equation, Bjerrum and Gjaldbaek’s 


equation (28), log K = log K, + log Kp —4log BK’, ........eee.. (144) 

where K, is the first dissociation constant of carbonic acid (log K, is therefore 

equal to — px,), Kp is the “solubility” constant of CO, which is estimated as 
log Kp = — 1:38, 

and K’ is a constant, which we shall shortly consider a little more closely, 

the value used by Bjerrum and Gjaldbaek being 


log K’ = — 5-68. 
Bjerrum and Gjaldbaek employed the mean of Walker and Cormack’s and 


Kendall’s determinations, namely — 6-51, as the value for log K,, from which 
they obtained log K = — 5-05, 


a figure which therefore was only 0-03 less than that determined directly. 

Now it will be seen that equation (144) reversed can be used for the 
calculation of log K, if log K’ is known, and that by substituting the value 
found, namely log K’ = — 5-02, we obtain 

log K, = — 6-48. 
This value agrees extremely well with that previously obtained from Walker 
and Cormack’s and Kendall’s experiments, and with those I myself deter- 
mined, but it will be shown the agreement can be made still better. 

The constant K’ determines the following equilibrium, Bjerrum and 
Gjaldbaek’s equation (15), 
: MME eT claaluaicitetds ate cccoescs (145) 
Apropos of this they write: 

“The ionic activity (a) can be calculated by multiplying the ionic con- 
centration (c) by the ionic activity coefficient (F,). For the monovalent 
bicarbonate ion the activity coefficient is approximately given by 

log F, = — 03 Veron ? 
and for the divalent calcium cation it is given by 
log F, = — 2 x 0:3 4/e,,,; 
Cro, Standing for the ionic normality of the solution. (145) now assumes the 
following form: 


Coa: .C 's : A 
“Co Ho" per Oe ttn es (146) 


If we call the equivalent concentration of the calcium carbonate which 
Schlésing? found dissolved C, we have 


Coa": = $C, Cuco’s => C; Cton = C. 
+ Bjerrum and Gjaldbaek substitute poo, for my Pgg., and naturally their numbering of 


equations is different from that given here. 
2 Cited from Bjerrum and Gjaldbaek [1919]. 


244 E. J. WARBURG 


By substitution and taking logarithms (146) becomes 

3.log C — log Poo, — 1:24/e — log 2 = log K’. ......... (147)” 
Bjerrum and Gjaldbaek calculate the values of log K’ with the help of (147) 
from Schlésing’s! experiments on the solubility of calcium carbonate at 
different CO, tensions, and their results are recorded in Table XLV in the 
first three columns. It will be noted they have calculated the activity of the 
bicarbonate ion from the equation 


log F (H0O',) + 0804/0,5) a4. ie ok (148) 


but as we have previously found the value 0-3 is too low their tables have 
been recalculated with the equation 
3 log C — log Poo, ~ 1:52 Ve — log 2 = log K’, ......... (149) 

having assumed that the constant in the equation for calculating the activity 
coefficient of the bicarbonate ion in calcium carbonate solutions is the same 
as in sodium salt solutions (see above). The values obtained have been 
put in the fourth column of Table XLV. Log K’ becomes — 5-709 and it 
will be observed that the agreement between the individual experiments is 
better calculated in this way than by Bjerrum and Gjaldbaek’s method. 
Schlésing’s! experiments were carried out at 16° and Bjerrum and Gjaldbaek 
therefore convert the values to 18° using a temperature coefficient derived 
from experiments of R. C. Wells!. Correcting the constant in the same way 
we obtain log K’ = — 5-749 at 18°. 


Now it appears that Bjerrum and Gjaldbaek’s “reaction constant” for 
calcium carbonate can be used without conversion if we confine ourselves to 
the experiments they signify as best suited for the calculation (0-1 and 0-02n 
CaCl,), because a recalculation using 0-46 instead of 0-30 gives no difference in 
the second decimal place. Calculated by (144) px, becomes 

Px, = 6515 at 18°, 3 
which is in surprisingly good agreement with Walker and Cormack’s, Kendall’s 
and my own determinations. 

At 18° the first dissociation exponent is therefore according to 


Knox 6-426 Kendall 6-507 
Walker and Cormack 6-512 Bjerrum and Gjaldbaek 6-515 
Hasselbalch 6-420 Warburg? 6-514 


Of these values Knox’s was obtained with a relatively indifferent tech- 
nique. In view of what was said in chapter IV Hasselbalch’s experiments 
may very well have a systematic error; all the other determinations give 
values which approximate closely to 6-51, Therefore 

K, = 31 x 10-7 (18°). 

T. H. Milroy [1917] has carried out electrical determinations at 37-5° with 
pure 0-2n sodium bicarbonate solution and the same diluted with 0-2n sodium 
chloride solution from which pg’, at the concentration 0-2n can be calculated. 


' Cited from Bjerrum and Gjaldbaek [1919]. ® Sodium salts, 


THEORY OF THE HENDERSON-HASSELBALCH EQUATION 245 


I have gone over Milroy’s experiments and the technique is hardly as 
good as could be wished. The dispersion is rather large and px’, has a lower. 
value than would be expected, being not as much as 6-00 on an average. 
J. F. McClendon, A. Shedlov and W. Thomson [1917] have also made measure- 
ments in salt solutions containing bicarbonate but they have only published 
their results graphically so that it is impossible to adjudicate upon the accuracy 
of the determinations. 

Lastly L. Michaelis and P. Rona [1914] have made some measurements in 
sodium bicarbonate solutions but as Hasselbalch has shown they are subject 
to a technical error which makes them very doubtful. 

The osmotic coefficient of the bicarbonate can be calculated with equa- 
tion (44), 1—fy= kv/e, 
where k is determined by (46), 

log, fa = — 4k Ve. 
By substituting the value 2-303 x 0-46 for 4k we get 
PAA Fe FSO WN Boe iis iocssetviccies (150) 
According to (136) —f,, (NaHCO ) = 1 — 0-51 Ve, 
and from (143) we have 
log f, (HCO’,) = (log F, (HCO’,)) = — 0-46 Ve. 

With the help of these equations the deviation coefficients for the bicar- 
bonate ion have been calculated and the results recorded in Table XLVI 
and graphically displayed in Fig. 14. 

When one compares the deviation coefficients given here with those valid 
for KCl which are recorded in chapter I, Table III one cannot help thinking 
the marked depression of the activity of the bicarbonate ion and the relatively 
low conductivity coefficient are an indication that the salt is not completely 
dissociated as we previously supposed. 

It would be of great interest to compare the deviation coefficients of a con- 
siderable number of salts with one another and with the dissociation constant 
of the corresponding acid, but such an investigation lies outside the scope of 
this work and it is to be hoped it will soon be undertaken by someone more 
expert in that branch of the subject. It may however be put forward that 
the fact that the activity of the bicarbonate ion can be determined with 
(143) over a large range of concentration does not at all well fit in with the 
idea of an incomplete dissociation of sodium bicarbonate. If we assume that 
equation (143) only represents the relation between the bicarbonate concen- 
tration and the activity of its ions, while the relation between the ionic 
concentration and activity of bicarbonate and sodium is given by 


pet ME gs ee PAD 4/0. nn scene ceescapesnsns snantoass (I) 
we can construct the following equation: 
oy 10-925 Hey = ¢ 10-946 Ye oe eee eeeees (II) 


and therefore log y = Vc (0-46 — O-25 V/y).  ceececeeecceescceneee (IIT) 


246 E. J. WARBURG 


ss | ae a we ts ret: BERS Bs | Ea aes Saeeye | ETS He ey | 
0-38}— = 
0-96 fA Bes! 
0-94 ‘hn 
o-92 \ — 
0-90 
ce |} f, (C09) a 
0-86 |} Pe Finge sn 
0-84 — 
0-82 -~ 
ot ; ae 
seo a: f, (NaHCOs) s 
0-76 }— it 
0-74;— Ped tl 
0-72}— =i 
0-70 - 
pate ine 2 
0-66 |— e Be 
0-64+-— aa 
os2|- f,(HCOs) “ 

ieee 


0-56 }— 7 4 
0-54 >-—- 


52} pe 
Pr) Baad SA Bia we eal AEN od OSE ee ea Sel DG ier Cs | 
0:00 001 02 03 “04 05 06 07 08 09 10 “11 “12 43 14 15 16 17 18 19 -20 


Fig. 14. Deviation coefficients of sodium bicarbonate solutions. 


According to the mass action law the following holds: 


Ona‘ Fa (Na*) . Cyco’s Fa(HCO’s) 
awad Waiedoes MT pee Bs gecan Cyeth cannabis (IV) 
If we put F, (NaHCO,) = 1 we get 
10-95 Vey 
aoe rg eee ee ee ee (V) 


If now equation (IIT) is solved by successive approximations and the values 
found are substituted in (V) we obtain 


c ¥ K 
0-001 0-952 1-7 x 10-? 
0-01 0-897 6-1 x 10-7 
O-1 0-790 1:8 x 10-3 


The example given shows therefore that “the constant” in (V) must increase 
with the salt concentration and that the mass action law cannot therefore 
be satisfied if the salt is incompletely dissociated, provided that an equation 
for the relation between the concentration and ionic activity of the form of 
(143) holds good. 

In Table XLVII are given determinations of px’, in solutions which 
contain small amounts of phosphate in addition to NaC] and NaHCO,. It 


THEORY OF THE HENDERSON-HASSELBALCH EQUATION 247 


is probable that px’, allowing for experimental error is identical with the 
constant one might expect if phosphate was substituted by chloride. 

In Table XLVIII are recorded two series of estimations carried out in 
about 0-01n sodium bicarbonate solutions which were at the same time 0-501m 
as regards cane sugar. 

The solutions were prepared by making stock solutions double the strength 
of those used in the experiments. These were kept on ice. Equal parts of 
sodium bicarbonate and cane sugar solutions were measured with Geissler 
pipettes and mixed half an hour before saturation was commenced during 
which time they were in the laboratory at room temperature. 

The potential established itself—in contrast to the pure salt solutions— 
not immediately, but only when the electrode had been rocked 10-25 times. 
The first series of measurements belongs to series A and was performed late 
in this series so that px’, is too low. It shows that the relative absorption 
coefficient is about 0-835 and y is therefore 16-5. The second series of measure- 
ments was carried out with E III and E VI simultaneously with the experiments 
previously noted as technically the best. It gives pg’, = 6-461 (18°). As px’, 
in the corresponding pure NaHCO, solution is 6-413, it appears that the 
addition of cane sugar increases the constant. If now we regard the depression 
of solubility of CO, as an expression of the amount of water the cane sugar 
has appropriated (we assume that CO, and NaHCO, are insoluble in hydrated 
cane sugar), and correct the calculations for this we obtain 

— log F, (CO) = 6-383, 
and at the same time the solution may be regarded as 0-012n with respect 
to bicarbonate, which corresponds to 
Px’, = 6-402. 
The difference between 6-383 and 6-402 does not lie outside the experimental 
error for certain, so much the more so because of the peculiarity in the estab- 
lishment of the potential in cane sugar solutions, just referred to. 

The result of this small series of experiments is that the activity of the 
bicarbonate ion is not affected with any certainty by the cane sugar molecules 
when a correction for the hydration is made. 


Résume, ' 

I. The first dissociation constant of carbonic acid has been recalculated 
from the best experiments in the literature, and it has been determined afresh 
by the electrical potential method at 18° and 38°. 

II. The apparent activity coefficient of the bicarbonate ion has been 
determined in sodium and potassium solutions up to 0-4n. 

III. The apparent osmotic coefficient of the bicarbonate ion has been 
calculated from the apparent activity coefficient. 

IV. The apparent conductivity coefficient of the bicarbonate ion has been 
recalculated. 

V. It has been rendered probable that the alkaline bicarbonates are 
completely dissociated. 


248 | E. J. WARBURG 


ADDENDUM 


After this chapter was completed a paper by C. Lovatt Evans appeared 
(J. Physiol. 1921, 54, pp. 353-366) which necessitates a certain amount 
of criticism. The author himself is rather cautious, as he writes: “Although 
perhaps some doubt still remains as to the finality of the conclusions which 
will be presented here etc....” 

The object of the paper was to prove that electrical determinations of 
the reaction of bicarbonate solutions give too acid values (too low py:) and 
that correct results can be obtained by the colorimetric method. The explana- 
tion of the reaction being found too acid by potential measurements is given 
by the author as a consequence of the formation of formic acid in sufficient 
quantity to set up a reduction potential by the catalytic action of the platinum 
black on the electrode, according to the equations 


CO, + H, HCOOH 
and HCO’; + H, = HCOO’ + H,0. 


(1) It can very easily be shown that the combined CO, in an alkali bi- 
carbonate solution does not decrease when the solution is treated for half an — 
hour (or one hour) with hydrogen in the presence of platinum black. I have 
myself carried out over 50 such experiments. However it would be very 
probable that the combined amount of CO, would decrease if a reduction 
potential was set up as Evans suggests. 

(2) If such a process took place px’; would be dependent upon the CO, 
tension and HCO’, concentration and not, as has been shown, practically 
only upon the cation concentration. 

(3) The calculation of px, from Bjerrum and Gjaldbaek’s and from my 
own experiments would give lower values than from conductivity determina- 
tions if the formation of formic acid took place. 

(4) The agreement between the calculated amount of combined CO, and 
that found experimentally in the phosphate experiments in chapter XI 
would be bad if Evans was right in his conclusions. 

The most important of Evans’ experiments are the following (px’, caleu- 
lated by the author): 


Neutral red Phenol red Electrically 
mm. Hg — A —, - A —, —_—__——_—_ 
co, PH PKs Pr PK’ PH PK’ 
0-02n NaHCO, 20° 
8-93 8-23 (6°59) 8-04 6-40 7:96 (6°33) 
21-8 7-70 6-45 7-42 6-17 7-55 6:30 
33°3 7-57 6-51 7-28 6-22 7:34 6-28 
46-0 7:37 6-45 717 6-25 7:16 6-24 
68-9 7:19 6°34 6-97 6-12 7:07 6-22 
Mean 6-44 Mean 6:19 Mean 6:26 
0-02n NaHCO, + 0-18 NaCl 20° 

6-5 8-13 (6°36) = Sh om die 

11-5 777 6°25 —_— — _ —_— 
13-9 7-66 6-22 7:67 6-23 7:60 6-16 
18-8 — -_- — _ 7:34 6-06 

26-0 7-48 6-31 — — — _ 
30-5 7-33 6-23 734 6-24 715 6-05 

42-8 7-23 6-28 —_— _ mae —_ 

44-0 7-21 6:27 — —_— _ _ 
46-1 - -- _— — 7-08 6-16 

57-2 7A 6-28 — —_ _ _— 
58-5 — _— — —_— 6-88 6-10 

65-4 7: 627 ~~ —_— _— —_ 


Mean 6°26 Mean 6-236 Mean 611 


THEORY OF THE HENDERSON-HASSELBALCH EQUATION 249 


From the determinations made in the preceding chapter px’; in the first 
series is 6-38 and in the second series 6-23 at 20°. 

The reason Evans did not obtain agreement between the electrical and 
colorimetric results is probably the fact that his electrode vessel is of an 
unsuitable type. As far as I can judge it must be almost impossible with 
his electrode to avoid the O, error. Evans himself writes: “ Definitive potentials 
were obtained almost at once provided the gas was free from oxygen; when 
this was not the case, there was first a somewhat lower potential than that 
finally obtained after one or two hundred inversions.” 

The colorimetric determinations of the reaction in salt solutions seem 
from the above to be rather good (agreement between px’, in each series), 
but the curves of the reaction of the blood indicate too alkaline a reaction 
possibly because the blood was saturated at body temperature, the dialysis 
and colorimetry however being performed at room temperature. 

The reactions of salt solutions which are “saturated” with smaller CO, 
tensions than 10 mm. are more alkaline than would be expected as Evans 
himself has noted. He ascribes this to hydrolysis which is incorrect as hydro- 
lysis at the reactions of the experiments is negligible in this connection (it 
is of the same order as doy’). The strongly alkaline reactions probably arise 
from equilibrium not being reached during the “saturation” at the low CO, 
tensions (cf. chapter XI). 

I shall not enter further into a discussion of the theory of the “degree 
of dissociation” of the bicarbonate here but refer the reader to the preceding 
chapters. 


Table XXX. Conductivity determinations in bicarbonate solutions. 


V c 8/e My 
12-1 0-0826 0-436 77:3 Kendall 
24-2 0-0413 0-346 81-7 NaHCO, 25° 
48-4 0:0207 0-275 85-3 
96-8 0-0103 0-217 88-2 
193-6 0-:00516 0-173 ae 
387-2 0-00258 0-137 2-4 
774-4 0-:00129 0-109 93-8 
1548-8 0-000646 0-0864 94-8 
3097-6 0-000323 0-0686 95-5 
_ 0 0 99-2 
32-0 00313 0-315 69-8 Walker and Cormack 
64-0 0-0156 0-250 72-9 NaHCO, 18° 
128-0 0-00782 0-198 75-6 
256-0 0-00391 0-158 78-6 
518-0 0-00195 0-125 80-8 
were 0 0 88-1 
64-0 0-0156 0-250 83-3 Kendall 
128-0 0-00782 0-198 88-8 1/2 CaCO, 25° 
256-0 0-00391 0-158 93-4 
512-0 0-00195 0-125 96-7 
1024-0 0-000977 0-0992 102-0 
2048-0 0-000488 0-0787 107-7 
_ 0 0 ? 


» Bioch. xv1 17 


250 E. J. WARBURG 


. 


Table XXXI. Determinations of the first dissociation constant of carbonic 
acid calculated from Walker and Cormack’s and from Kendall’s experi- 


ments. 
y: My ¥ OEE UU ed 
31-25 1-104 0-00310 3-09 Walker and Cormack 
62-5 1-570 0-00441 3-13 18° CO, from marble 
93-7 1-916 0-00539 3-11 ; 
125-0 2-218 0-00623 3-12 
roa) 355°8 Mean 3-11 
PK, 6-507 
27-5 1-033 0-00291 3:07 Walker and Cormack 
55-0 1-454 000409 — 3-05 18° CO, from carbonic acid snow 
82-5 1-754 0-00502 3:05 
110-0 2-052 0:00577 3-04 
ra) 355°8 Mean 3-05 
PK, 6-516 
25-4 0-631 0:00237 2-23 Kendall 
38-3 0-770 0-00290 2-20 0° 
50-0 0-880 0-00331 2-20 ‘ 
76:3 1-081 0-00407 2-18 
99-8 1-242 0-00468 2-21 
152-6 1-548 0-00584 2-24 
o 265-6 Mean 2-21 
PK, 6-656 
30-9 1-100 0-00309 3-11 Kendall 
42-0 1-281 0-00360 3-11 18° 
61-2 1-550 , 0:00437 3-14 
83-4 1-792 0-00504 3-06 
fs) 355°8 Mean 3-11 
PK, 6-507 
36-4 1-403 0-00355 3:47 Kendall 
51-3 1-659 0-00320 3°44 25° 
72:8 1-977 0-00500 3:45 
102-4 2-341 0-00592 3:44 
145-5 2-820 0-00713 3°54 
ra) 395-5 “Mean 3:47 
PK, 6-460 


Table XXXII. Calculation of px’, in pure sodium bicarbonate solutions from 
K. A. Hasselbalch’s experiments, Biochem. Zeitsch. T8, p. 119. 


PHY PK’, 

Pe PE. ee v - 

Corrected for decreased 
Concentration mm, CO, Uncorrected H, pressure 

0-05n 47-2 747 6-19 7:48 6-20 

&/c =0°368 403-2 6-53 6-18 6:68 (6°33)! 
18° 74-0 7:28 6-20 7:30 6-22 
Mark 4 147-7 6-97 6-19 701 6-23 
Mean 6°19 Mean 6:22 
0-0ln 54:3 6-82 6-29 6°83 6:30 
&e=0-215 134-3 642 6-28 6:46 6°32 
18° 101-8 651 6°25 6-54 6-29 
Mark 4 84-5 6-59 6°25 661 627 
68:8 6-73 6°30 6°75 6°32 
92-4 659 629 6-62 6:32 
Mean 628 Mean 6:30 


1 Not included in the mean value. 


THEORY OF THE HENDERSON-HASSELBALCH EQUATION 251 


Table XX XII (continued) 


PH PK’, 
PH PK’ Se v 4 
: Corrected for decreased 
Concentration mm. CO, Uncorrected H, pressure 

0-05n 168-9 7-02 6-06 7-08 6-12 

&/e=0-368 247°8 6:87 6-08 6-97 (6-18)? 
38° 52-6 7:55 6-09 7:57 6-11 

Mark 9 717-0 6-40 6-07 ? ? 

Mean 6-08 Mean 6-12 
0-03 40-6 7:48 6-12 7-49 6-13 
e=0-311 80-7 7-15 6-09 7-18 6-12 
38° 135-0 6-92 6-09 6-96 6-13 
Mark 9? 96-9 7-08 6-10 7-11 6-13 
Mean 6-10 Mean 6:13 
0-02n 123-3 6-83 6-13 6-87 6-17 
3/e =0-271 36-7 7-33 611 7-34 6-12 
> 62-8 7-13 6-14 715 6-16 
Mark? Mean 6-13 Mean 6-15 
0-01n 85-0 6-70 6-14 6-73 6-17 
&/ce=0-215 218-6 6-33 6-18 6-40 6-25 
137-4 6-51 6-16 6°55 6-20 
Mark 9? 189-2 6-40 6-19 6-46 6-23 
119-8 6-57 6-16 6-61 6-20 
149-0 6-44 6-13 6:48 6-17 
Mean 6:16 Mean 6:20 
0-005n 23-5 7-02 6-21 7-03 6-22 
3/e = 0-171 41°35 6°75 6-18 6-76 6-19 
Mark 9 78 6-53 6-20 6-55 6-22 
Mean 6-20 Mean 6-21 


1 Not included in the mean value. 


Table XX XIII. Determinations of px’, for 18° in sodium bicarbonate solutions 
carried out with the small saturator electrode. Series A. Mark A. 


mm. Hg 
(ny 
Concentration Temp. co, O, PH’ PK’, 18° 
T' =8-88 16-0 240-2 0-6 5-84 6-37 
0-00399n HCO’, 16-5 118-6 0-2 6-19 6-42 
0-00399n Na’ 16-5 41-6 0-2 6-56 6-33 
16-5 15-2 0-2 7-04 6-37 
16-5 58 0-2 7-45 6-37 
Mean 6-37 
T =17-52 19-5 543-4 0-3 5-80 (6-37) 
0-00786n HCO’, 19-5 141-0 0-5 6-38 6-37 
0-00786n Na* 19-5 55-2 0-5 6-82 6-40 
19-5 18-0 0-5 7:29 6-38 
20-0 Tl 0-3 7:70 6:39 
20-0 4-2 0-3 8-05 (6-51) 
Mean 6-38 
T =55-5 19-5 486-5 0-7 6-35 6-37 
0-0249n HCO’, 19-0 124-4 0-5 6-95 6-38 
0-0249n Na* 19-0 51-0 0-5 7:33 6-37 8/e=0-273 
19-0 30-9 0:3 7-51 6-35 
19-0 11:3 0-3 8-01 6-40 
Mean 6°37 


17—2 


252 E. J. WARBURG 
Table XX XIII (continued) 


mm. Hg. 
aor Bae ET 
Concentration Temp. co, O, Pr pK’, 18° 
T =110-0 20-5 538-0 0-6 6-60 (6-37) 
0-0494n HCO’; 20-0 267-0 0-5 6-91 6-37 
0-0494n Na® 20-0 67-2 0-5 7-50 6-36 8/e =0-367 
20-0 25-2 0-2 7-91 6:35 
20-0 9-2 0-2 8-36 6°37 
Mean 6-36 
T =221-9 19-5 126-3 0-5 7-44 6-28 
0-0997n HCO’; 19-5 37-5 0-5 8-00 6-32 
0-0997n Na’ 19-5 14-7 - 0:3 8-39 6°31 
19-0 58-8 0:3 7-80 6-31 
17-0 193-4 0-2 7-28 6-31 &/c =0-464 
17-0 85-3 0-9 7-63 6-31 
17-0 25-1 1-2 8-08 6-33 
17-0 50-1 0-7 7:88 6-32 
17-0 130-1 0-5 7-41 6-27 
17-0 56-6 03 7-81 6-32 
Mean 6-31 


Table XXXIV. Determinations of px’, for 18° in sodium bicarbonate solu- 
tions with the addition of 0-42 °% NaCl (0-0718n). Carried out with the 
small saturator electrode. Series A. Mark A. 


mm. Hg 
gS. Cheese 
Concentration Temp. co, us OS PEL pK’, 18° 
T =8-88 18-5 245:8 0-6 5-78 6:31 
0-00399n HCO’, 19-0 29-5 0-6 6-70 6-31 
0-0718n Cl’ 19-0 40-7 0-6 6-56 6:30 &/c=0-421 
0-07479n Na* 19-0 63-8 0-4 6-23 6-28 
Mean 6-30 
T =17-52 19-0 395-5 1-0 5:86 6-31 
0-00786n HCO’, 19-0 108-9 0-4 6-42 6:28 
0-0718n Cl’ 19-5 34:5 0-2 6-91 6-29 /e=0-429 
0-07966n Na* 19-5 6-0 0-2 7-68 6-30 
19-5 14-4 0-2 731 (6°31) 
Mean 6-29 
T =55°5 20-0 36-7 6-6 7:34 (6-24) 
0-0249n HCO’, 20-0 119-2 12° 6-86 6:27 
0-0718n CI’ 20-0 106-2 0-7 6-93 6:29 
0-0967n Na* 20-0 7:3 0-5 8-10 6:30 &/e=0-459 
20-0 512-0 0-6 6:22 (6-25) 
20-0 126-7 0:3 6-85 6-29 
20:0 36-1 0-2 7:40 6-29 
20-0 31 0-2 8-46 (6°31) 
Mean 6-29 
T =110-0 21-0 144-2 0-2 7-10 6:29 
0-0494n HCO’, 21-0 48:3 0-2 7-57 6:29 
0-0718n Cl’ 21-0 34-7 0-2 7-72 6:29 YVe=0-495 
0-1212n Na’ 205 77 0-2 8°36 6:29 
20-5 38 0-2 8:77 (6-43) 
Mean 6:29 
1 = 221-9 17-0 476-4 Lb 6-76 (6:18) 
0-099Tn HCO’, 17°5 122-5 0-8 TAl 6-23 
0-07 18n Cl’ 17-0 58-2 0-3 777 6:28 Ve = 0-556 
0-1715n Na’ 17-0 23:3 03 815 6:27 
17-5 22:2 0:3 817 6:27 


Mean 6:26 


THEORY OF THE HENDERSON-HASSELBALCH EQUATION 253 


Table XXXV. Determinations of px’, for 18° in sodium bicarbonate solutions 
with the addition of 0-85 % sodium chloride (0-145n). Carried out with 
the small saturator electrode. Series A. Mark A. 


mm. Hg 
Ser etres, 
Concentration Temp. C04 O, PH px’, 18° 
T =8-88 19-0 238-4 08 5-78 6-30 
0-00399n HCO’, , 19-0 111-6 0-0 6-10 6-29 
0-145n Cl’ 19-0 30-1 0-3 6-64 6-26 X/e =0-530 
0-1490n Na* 19-0 12-2 0-3 7-03 6-26 
19-0 63-7 0-3 6-33 6-27 
19-0 12-6 0-3 7-02 6-27, 
Mean 6-28 
T =17-52 20-0 587-6 0-7 5-60 (6-20) 
0-00786n HCO’, 19-5 149-5 0:3 6-24 6-26 
0-145n Cl’ ye. “1G 43-9 0-5 6-76 6-24 8/e =0-535 
0:1529n Na* 19-0 14:3 0-4 7-23 6-22 
19-0 6:3 0-4 7-61 6-24 
19-0 36 0-4 7:87 (6-26) 
Mean 6-24 
T =55-5 21-0 152-0 0-4 6-72 6-22 
0-0249n HCO’, 20-5 47-6 0-4 7:24 6-24 
0-145n Cl’ 21-0 14-6 1-6 777 6-26 &/c =0-554 
0:1699n Na’ 20-5 2-7 0-4 8-55 (6-25) 
19-5 598-0 1-1 © 613 (6-24) 
19-5 411-7 5 6-27 6-22 
Mean 6-24 
T =110-0 19-5 60-9 02 7-44 6-27 
0-0494n HCO’, 19-5 26-5 © 0-2 781 628 
0-145n Cl’ 19-5 77 0-2 8-34 6-29 
0-1844n Na* 20-5 565-0 1-4 6-42 (6-25) &/e =0-569 
20:5 142-0 0-6 7-06 6-23 
20-5 27-7 0-6 7-79 6-28 
Mean 6-27 
T =221:9 17-0 460-5 0-6 6-73 (6-14) 
~  0:0997n HCO’, 17-0 161-2 0-6 7-24 6-19 
0-0997n Cl’ 17-0 49-1 0-2 7-79 6-23 
0:2447n Na* 17-0 15-5 0-2 8-28 6-23 &/e =0-625 
17-0 67-9 0-2 8-64 6-25 
19-0 45-0 0-6 7:83 6-22 
19-0. 90-9 0-7 7-56 6-26 
19-0 32:8 0-6 8-00 6-25 
Mean 6-23 


Table XXXVI. Determinations of px’, for 18° in sodium bicarbonate solutions 
with the addition of 1-7 % NaCl (0-291). Carried out with the small 
saturator electrode. Series A. Mark A. 


mm. Hg 
eS RE aaa 
Concentration Temp. co, O, PE pK’, 18° 
T =8-88 18-5 243-4 0-4 5-69 6-22 
0-00399n HCO’, 19-0 115-5 0-4 6-00 6-20 
0-291n Cl’ 19-0 36-1 0-4 6-52 6-22 o/c =0-666 
0-2950n Na* 19-5 11-7 0-4 | 7-00 6-21 
19-5 16-1 0-4 6-85 6-19 

e Mean 6-21 
T =17-52 18-5 529-3 1-1 5-57 (6-13) 
0-00786n HCO’; 18-5 149-5 0-8 6-17 6-19 
0-291n CY’ 18-5 46-2 0-2 6-67 6-18 2/c =0-669 
0-2989n Na* 19-0 15-3 0-2 717 6-19 


Mean 6-19 


254 E. J. WARBURG 
Table XXXVI (continued) 


mm. Hg 
"yc aaep aE ey E 9 
Concentration Temp. co, 0, PH PK’, 18° 
T =55-5 19-5 602-7 1-1 6-08 (6-20) 
0-0249n HCO’, 19-5 416-2 0-5 6-27 6-23 
0-291n Cl’ 19-5 131-1 0-6 6-73 6-19 &/c=0-681 
0-3159n Na* 19-0 62-0 0-5 7-09 6-23 
19-0 17-2 0-5 7-61 6-18 
19-0 6-0 0-5 8-08 6-19 
19-0 2:8 0-5 8-48 (6-19) 
Mean 6-20 
T =110-0 20-0 518-9 0-4 6-50 (6-25) 
0-0499n HCO’, 20-0 139-1 0-4 7-01 6-19 
0-291n CY’ 20-5 48-4 0-4 7:50 6-21 Se =0-699 
0-3409n Na’ 20-0 16-2 0-4 7-94 6-19 
Mean 6-19 
T =221-9 17-0 442-1 0-6 6-76 615 
0-0997n HCO’, 17-0 40-0 0-4 784° 6-19 
0-291n Cl’ 17-0 20:7 0-4 8-09 6-15 
0-3907n Na’ 17-0 40-1 0-4 7:83 6-18 
17-0 14-5 0-4 8-23 614  gce=0-731 
18-0 95-2 0-6 7:50 wien. 
18-0 106-4 0-6 7:39 6-15 
18-0 365-2 0-7 6-86 6-17 
18-0 26-1 0-4 8-04 6-19 
Mean 6-17 


Table XXXVI. Determinations of px’, for 18° in sodium bicarbonate solu- 
tions with the addition of varying quantities of KCl. Carried out with 
the small saturator electrode. Series A. 


mm. Hg 
Fn ns 
Concentration Temp. CO, O, PE PK’, 18° 

T =44:3 20-0 80-4 0:3 6-98 6-32 
0-0199n HCO’, 20-0 132-0 0:3 6-79 6°34 
0-54 % KC1(0-0718n) 20-0 15-4 0-6 7-71 6°33 
Mean 6-33 
T =110-0 18-0 58-6 0-5 7:48 6:28 
0-0494n HCO’, 18-0 98-7 1-7 7-26 6-29 
0-54 % KC1(0-0718n) 18-0 21-2 0-5 7-92 6-29 
Mean 6-29 
T =443 19-0 1303 | 0-6 6-72 6-27 
0-0199n HCO’, 19-0 37-1 0-8 7:27 628 
1-08 % KCL(0-145n) = 19-5 13-8 0-6 7-69 6-26 
19-5 56-2 0-6 7-09 6-27 
Mean 6:27 
T =44:3 19-0 64-3 0-5 6-98 6-22 
0-01992 HCO’, 20-0 95:3 0-7 6°84 6:24 
216% KC1(0-291n) =. 20-0 30-1 0-7 7:34 6°25 
19-5 39:3 0-7 7:20 6-23 
Mean 6:24 

T =55°5 17-0 132-2 0-5 6-76 620. * 
0-0249n HOO’, 17-0 55-0 O-4 7:14 6-23 
2-16 %, KC1(0-291n) 17-0 18-1 0-6 7:61 6:22 
Mean 6:22 
T = 110 18-0 58-6 0-6 7-40 621 
0-0494n HOO’, 18-0 98:7 1-7 TAT 621 
216% KCI(0-291n) 18-0 21-2 0-5 7-85 6-22 


THEORY OF THE HENDERSON-HASSELBALCH EQUATION 255 


Table XX XVIII. Determinations of px’, for 18° in sodium bicarbonate solu- 
tions with the addition of 0-145” NaCl or KCl. Carried out with the small 


saturator electrode. Series A. 


mm. Hg NaCl KCl 

7a pe Wale Ser EP | i meni, "an WSS GE 

Concentration Temp. CO, O, PE PK’, Pr PK’, 18° 
T =56-25 17-0 162-2 0-2 6-66 6-21 6-69 6-24 
0-0253n HCO’, 17-5 59-1 0-7 T1l 6-23 7-16 6-28 
Mean 6-22 Mean 6:26 
T=111-0 17-0 125-3 0-4 7-07 6-21 7:10 6-25 
0-0499n HCO’, 17-5 34-2 1-7 7-68 6-23 7:70 6-28 
17-5 45:3 0-4 7-54 6-24 7-57 6-28 
Mean 6:23 Mean 6:27 


Table XXXIX. Determinations of px’, for 18° in sodium bicarbonate 
solutions with varying additions of salt. Series A. 


0-0718n 0-0145n 0-291n 
No salt So my 
T added — NaCl KCl NaCl KCl NaCl KCl 
8-88 (6:37) 6-30 re 6-28 ee 6-21 a 
17-5 (6-38) 6-29 cs 6-24 oe 6-19 aes 
44:3 tse me 6-33 ent 6-27 ae 6-24 
55:5 6-37 6-29 a 6-24 es 6-20 6-22 
110-0 6:36 6-29 6-29 6-27 6-27 6-19 ja 
111-0 sae ne tics ae 6-23 ae 6-21 
221-9 6-31 6-26 io 6-23 es 6-17 ae 


Table XL, Determinations of px’, for 18° in sodium bicarbonate solutions 
with the addition of salt. Carried out with the small saturator electrode. 
Series B. Mark +. 


mm. Hg 
Concentration Temp. CO, rr PK’, 18° 

T =45-2 21-0 50-1 7-18 6-30 
0-02038n HCO’; 21-0 8-7 7-90 6-26 g/e=0-452 
0-42 % NaCl (0-0718n) 21-0 27:7 7-46 6-32 
0-0718n CY’ M 6-29 
0-0921n Na* copii 
T =56-5 15-0 579-4 6-18 6-32 
0:0254n HCO’, 15-5 139-0 6-79 6-31 
0-42 % NaCl (0-0718n) 15-5 30-1 7-50 6-35 B/e=0-458 
0-0718n Cl’ 19-5 61-8 7-20 6-34 
0-0962n Na* 19-5 52-0 7:27 6-32 

19-5 41-6 7:39 6°35 

20-0 41-6 7:38 6-33 

Mean 6-33 

T =45-2 21-0 51-2 7:16 6-30 
0-0203n HCO’, 21-0 36-0 7:30 6-28 8/c=0-452 
0-42 % NaCl (0-0718n) es 
0-0718n CV Mean 6:29 
0-0921n Na’ 
T=55:7 20-0 36-5 7-42 6-30 
0-0250n HCO’, 20-0 23-2 7-55 6-24 
0:85 % NaCl (0-145n) 20-0 22-2 7-60 6-27 8/c=0-554 
0-145n Cl’ 21-0 38-4 7:36 6-26 
0-170” Na* 19-0 142-6 7-76 6-24 

19-0 53-2 7-22 6-27 

Mean 6-26 


256 E. J. WARBURG 
Table XL (continued) 


mm. Hg 
Concentration Temp. CO, PH PK’, 18° 

T =56-5 21-5 579-4." 6-13 (6-22) 
0-0254n HCO’; 21-0 111-3 6-92 6-29 
0-85 Joa aes (0- 145n) 21-0 48-5 7-26 6-27 &/c=0-558 
0-145 21-0 31:8 7:46 6-29 
0-174n ote 21-0 49-0 7:26 6:26 

21-0 19-7 7-68 6:30 

Mean 6-28 

T =56-5 19-5 199-8 6-55 6:19 
0-0254n HCO’, 19-0 81:3 6:97 6-21 
1-7 % NaCl (0° 291n) 19-0 36:8 7:29 6-18 ~/ce=0-681 
0-291n Cl’ 19-0 ~. 20-0 7-58 6-21 
0-:3164n Na’ 19-0 8-4 7:96 6:22 

19-0 17-1 7-65 6-21 

Mean 6-20 

T=56-5 19-5 61-8 7-18 6:30. 
0-025n HCO’, 19-5 52-5 7:25 6-31 8/e=0-554 
1-08 % KCl (0: 145n) 19-5 41-6 7:34 6-30 
0-145n Cl’ 20-0 41-6 7:36 6-30 
0:145n K* Aah 
0-025n Na’ Mean 6-30 


Table XLI. Determinations of px’, in salt solutions for 18°. Carried out by 
Hober technique in E VI and EIII. Series B. Mark O. 


pK’, 18° 
mm. Hg a A — 
Concentration Temp. CO, Pr EVI EMMI 
T =224-4 19-0 96-2 7-533 6:247 —_ n/10 NaHCO, 
0-1008n HCO’, 19-0 65-9 7-737 6-287 — 
0-1008n Na* 19-0 40-7 7-969 6-299 -—— 
19-0 31-6 | 8-051 6-281 — Ye =0-464 
19-0 31-6 8-061 6-292 a 
19-0 30-7 8-070 6-288 os 
19-0 32-7 8-026 6-261 — 
Mean 6-279 
7 =112-2 18-0 86-9 7-364 6-340 — n/20 NaHCO, 
0-05041n HCO’, 18-0 86-9 7:364 —_ 6-340 
0-05041n Na’ 18-0 61-5 7-506 6-321 — 
18-0 61-5 7-497 — 6-323 8/e = 0-369 
18-0 50-7 7-604 6-346 _— 
18-0 50-7 7-597 — 6-349 
Mean 6°335 6:337 
= 6°335 
T =56-49 18-5 52-7 7-324 6384 — n/40 NaHCO, 
0-02538n HCO’, 20-0 24:8 7-680 6-400 _ 
0-02538n Na* 20-0 14-7 7-890 6:383 —_ &/c =0-294 
20-0 32-0 7-571 6-402 a 
19-0 47:8 7-382 6391 — 
Mean 6393 
T' = 22-06 17-0 104-1 6-644 6-392 — n/100 NaHCO, 
0-01031n HCO’, 17-0 104-1 6-636 _ 6°385 
0-010381n Na’ 16-5 43-4 7-038 6-416 _ 
16-5 43-4 7-035 — 6413 Se= 0-217 
19-5 82-2 6-742 6-382 — ; 
19-5 35-9 7-133 6-413 _ 
19-5 35-9 7-133 — 6-418 
19-5 37°3 7-090 6-386 - 
Mean 6°398 6-408 


” 6-400 


THEORY OF THE HENDERSON-HASSELBALCH EQUATION 257 
Table XLI (continued) 


PK’, 18° 
mm. Hg — A — 
Concentration Temp. CO, Pr EVI EIit 
T=111+4 21-0 28-2 7-803 6-291 — n/20 KHCO, 
0-05004n HCO’, 20-5 33-3 7-751 6-308 — 
0-0497n Cl’ 20-5 33°3 7-742 — 6-299 n/20 KCl 
0-1000n K* 20:5 49-8 7-600 6-332 — 
20-5 49-8 7-593 — 6-323 3/e=0-464 
22-0 26-2 7-876 6-319 — 
Mean 6-313 6-312 Mark ® 
” 5 6-312 
T =224-4 20-0 83-3 7-621 6-268 — n/10 NaHCO, 
0-1008n HCO’, 20-0 83:3 7-618 — 6-265 
0-1012n CY 20-0 41-1 7911 6-250 — n/10 KCl 
0-1008n Na’ 20-0 41-1 7911 — 6-260 
0-1012n K* &/e =0-585 
0:09965n Cl’ 22-0 91-7 7-586 6-269 — 
0:09965n K* 22-0 38-3 7-938 6-242 — 
22-0 38:3 7-931 —- 6-235 
T =222-7 21:5 85-1 7611 6-271 — n/10 KHCO, 
0-1000n HCO’, 21-5 85-1 7-604 _ 6-264 
0-1012n Cl’ 21-5 44-5 7-896 6-275 — __n/10 NaCl 
01000” K* 985 : 
0:1012n Na: Mean 6-262 6 256 
“ 6-260 
T =2:244 19-5 35-2 6-029 6-310 — 
0-001008” HCO’, 19-5 35-2 6-026 — 6-310 n/1000 NaHCO, 
0-0994n Cl’ 20-0 47-0 5-904 6-303 —_ 
0:1004n Na’ 20-0 38-6 5-991 6-298 — n/10 NaCl 
20-0 38-6 5-991 — 6-298 
19-0 37-4 6-034 6-334 — 
19-0 37-4 6-031 — 6-333 
19-5 39-5 5-987 6-318 — 
19-5 39-5 5-984 _- 6-315 
Mean 6-313 6-314 
ts 6-313 
T =221-45 19-5 44-1 7-929 6-322 — n/10 KHCO, 
0:09948n HCO’, 19-5 44-1 7-918 — 6-315 
0-09948n K* — 19-5 29-2 8-101 6-317 — 
19-5 29-2 8-094 — 6-310 /ce=0-464 
19-5 60-9 7-799 6-324 — 
19-5 60-9 7-789 — 6-314 
Mean 6-321 6-313 Mark © 
” 6-317 
T =22-14 19-5 38-9 7-092 6-422 — 
0-009948n HCO’, 19-5 38-9 7-095 — 6-425 n/100 KHCO, 
0-009948n K* 19-5 48-4 6-979 6-404 — 
19-5 48-4 6-979 — . 6-404 8/e=0-215 
19-5 50-2 6-969 6-410 — 
Mean 6-412 6415 Mark @ 
Ae 6-413 
T =224-4 20-0 57-8 7-737 6-226 —~ n/10 NaHCO, 
0-1008n HCO’, 20-0 57:8 7-737 — 6-226 
0-1002n CY 19-5 41-5 7-906 6-267 — n/10 NaCl 
0-2010n Na‘ 19-5 41-5 7-904 — 6-265 
22-0 81:3 7-623 6-254 — 
22-0 81:3 7-621 — 6-252 3/c=0-585 
22-0 29-5 8-048 6-228 — 
22-0 29-5 8-058 —- 6-238 Mark + 
22:5 78:2 7-642 6-249 — 
22:5 718-2 7-639 — 6-256 
Mean 6-245 6-247 


”? 6-246 


258 E. J. WARBURG 
Table XLI (continued) 


px’, 18° 
mm. Hg. c A — 
Concentration Temp. co, PH EVI El 
T =112-2 21-0 75-6 7-411 6-314 — n/20 NaHCO, 
0-05041n HCO’, 21-0 75:6 7-408 — 6-311 
0-0501n Cl’ 22-0 24:3 7-889 6-295 — n/20 NaCl 
0-1005n Na* 22-0 24:3 7-871 — 6-277 
20-0 31-6 7-773 6-300 — &/e=0-464 
20-0 31-6 7-761 6-288 
Mean 6303 6-292 Mark + 
oa 6-297 
T =222-7 21-5 80-9 ~~ 7-647 6-285 — n/10 KHCO, 
0-1000n HCO’, ~ 21-5 80-9 7-637 —- 6-275 n/10 KCl 
0-0995n Cl’ 21-5 29-2 - 8-070 6-276 _— 
0-1995n K* 21-5 29-2 8-068 — 6-274. 8/c=0-585 
21-5 30-4 8-063 6-286 — 
21:5 30-4 8-063 ae 6-296 
Mean 6-282 6-282 Mark @ 
* 6-282 


Table XLII. Determinations of px’, at 38°; salt solutions. Carried out by 
Hober technique in EVI. Series B. Mark o. 


mm. Hg 
[ances fe ret ate PK - 
Concentration Co, O, PE’ 38h" 
T =56°3 465-7 0-2 6-25 (6-02) /c=0-681 
0-0253n HCO’, 108-7 0-2 7:20 6-15 
1-7 % NaCl (0-291n) 44-3 0-2 7:35 6-11 
0-291n CY’ 17-9 0-3 "7-22 6-09 
0-3163n Na’ 40-1 0-3 7:38 6-09 
Mean 6-11 
T =56-5 587-6 0-0 6-21 (6:09) 8/c=0-681 
0:0254n HCO’, 338-4 0-1 6°45 6-10 
1-7 %, NaCl (0-291n) 127-7 0-0 6-89 6-11 
0-291n Cl’ 45:8 0-0 7-44 6-11 
0-3164n Na’ 8-7 0-6 7-04 6-08 
Mean 6-10 


Table XLII. Determinations of px’, at 38° in sodium bicarbonate solutions 
without the addition of salt. Carried out by Héber technique in E VI. 
Series B. Mark o-. 


mm. Hg 
: - PR’, 
Concentration CO, Oy Pr 38° 
T =56°5 54-1 — 7:42 6-27 3/c=0-287 
0-0254n HCO’, 41-4 — 7-56 6-29 
0-0254n Na’ 43-4 — 7:55 6°30 
95-7 a 7-19 6-28 
32-1 ~ 7:66 6:27 
Mean 6:28 
T =224-4 74-0 -— 7-82 620 Se=0-466 
0-101n HCO’, 50-4 ~~ 7-90 6-18 
0-101n Na’ 90°7 - 775 6-22 
5O5 oa 7:87 615 
112-0 — 7-58 614 
50-2 — 7-92 614 


Mean 617 


THEORY OF THE HENDERSON-HASSELBALCH EQUATION 259 


Table XLIV. Determinations of px’, at 38° in sodium bicarbonate solutions 
with the addition of salt. Carried out with the small saturator electrode. 
Series B. Mark x. 


mm. Hg 
: PRY 
Concentration CO, O, Pr 38° 

T =45-2 48°8 —- 7-31 6-21 Re =0-452 
0-0203n HCO’, 8-7 — 8-01 6-17 
0-42 % — (0-0718n) 26-5 — 7-56 6-18 
0-0718n CY pees 
0-0921n Na’ Mean 6:18 

T =56°5 34-6 0-1 7-55 6-19 g/ce=0-458 
0:0254n HCO’, 114-0 0-6 7-02 6-19 
0-42 % NaCl (0-0718) 52-3 0:3 7:34 6-17 
0-0718n Cl’ 39-3 0-0 7-51 6-21 
0-0962n Na Mean 6-19 

T =45-2 543-5 6-15 (6-09) /e=0-549 
0-0203n HCO’, 62-7 — 710 6-11 
0-85 % NaCl (0-145n) 49-2 0-8 7-26 6-17 
0-145 Cl’ 36-2 0-5 7-41 6-17 
0:1653n Na’* 34-4 0-1 7-43 6-18 
19-1 0-5 7-66 61 
Mean 6-16 

T =56°3 529-7 0-2 6-28 6-12 Sfe=0-554 
0:02538n HCO’, 107°6 0-2 7-00 6-15 
0-85 % NaCl (0-145n) 37-4 0-2 7-44 6-12 
0-145 Cl’ 35°3 0-2 TAT 6-13 
0-1703n Na Mean 6-13 

T=56°5 567°3 0-0 6-27 614 3e=0-554 
0-0254n HCO’, 142-0 0-1 6-89 6-16 
0-85 % NaCl (0-145) 25-6 0-2 7-61 6-13 
0-145 Cl’ 34:3 0-0 7-53 6-18 
0-1704n Na* 34-0 0-6 7-54 6-18 
56-6 0-6 7-30 6-17 
39-9 0-6 7-45 6-16 
Mean 6:16 


Table XLV. Conversion of Bjerrum and Gjaldbaek’s Table VI. 


log K’ 
Poo. c 108 log K’ converted 
0-000504 1-492 — 5-619 — 5-666 
0-000808 1-700 — 5-660 — 5-698 
0-00333 2-744 . — 5-677 —5:°754 
0-01387 4-462 — 5-688 —5-741 
0-0282 5-930 — 5-649 — 5-707 
0-0501 7-200 — 5-661 — 5-723 
0-1422 10-66 — 5-635 — 5-705 
0-2538 13-27 — 5-621 — 5-697 
0-4167 15-75 — 5-630 —5-710 
0-5533 17-71 — 5-613 — 5-697 
0-7297 19-44 — 5-620 — 5-705 
0-9841 21-72 — 5-616 — 5-704 


Mean —5-641 Mean —5-709 


260 E. J. WARBURG 


Table XLVI. Deviation coefficients of the bicarbonate ion 
in sodium salt solutions. 


Cc Ye Sa fo Su (NaHCO,) 
0-001 0-100 0-900 0-974. 0-949 
0-005 0-171 0-834 0-955 0-913 
0-01 0-215 0-796 0-944 0-890 
0-02 0-271 0-751 0-929 0-862 
0-03 0-311 0-719 0-919 0-841 
0-04 0-342 0-696 0-911 0-827 
0-05 0-368 0-677 0-904 0-812 
0-06 0-391 0-661 0-898 0-801 
0-07 0-412 0-646 0-893 0-790 
0-08 0-431 0-634 0-888 - 0-780 
0-09 0-448 0-622 0-884 0-778 
0-10 0-464 ~~ 0-612 0-879 0-764 
0-12 0-493 0-593 0-872 0-749 
0-14 0-519 0-577 0-865 0-735 
0-16 0-543 0-563 - 0-859 0-723 
0-18 0-565 0-550 0-853 0-712 
0-20 © 0-585 0-538 0-848 0-702 
0-22 0-604 0-528 0-843 0-692 
0-24 0-621 0-518 ' 0-839 0-683 
0-25 + 0-630 0-513 0-836 0-679 


Table XLVII. Determinations of px’, in salt solutions containing phosphates, 
for 18°. Carried out with the small saturator electrode. Series A. 


mm. Hg Vols. % CO, PK’, 
Concentration Temp. CO, combined Pr 18° 
0-85 % NaCl 19-0 49-3 35°8 7-03 626 Y=6 
70 ce. NaHCO, (7' =43-0) 19-0 156-9 45-9 6-63 6-24 Phosphate 
6 ce. KH, PO, (0-066m) 19-0 48-2 — 34:8 6-45 6-27 
24 cc. Na, HPO, (0-066m) 19-0 16-3 28-9 6-41 6-24 
0-02m phosphate 19-0 38-3 33-6 6-44 6-25 
0-145n Na’ : 7 on 
0-004n K* Mean 6:25 
0-42 % NaCl 18-0 51-5 38-5 7:06 6-28 - Yes 
70 ce. NaHCO, (7 = 43-0) 18-5 28-9 33-4 7-25 6-28 Phosphate 
6 ce. KH, PO, (0-066m) 18-5 53-6 37:5 7:05 6-29 
24 cc. Na,HPO, (0-066m 19-0 156-9 48-2 6-66 6-25 
0-02m phosphate A.08 
0-0718n CY Mean 6-28 
0-101n Na’ 
0-004n K* 


Table XLVIII. Determinations of px’, for 18° in cane sugar solutions 
(0:501m) with sodium bicarbonate (circ. m/100). Carried out with E VI 
and EIII. 


mm. Hg Vols. % CO, 
Concentration Temp. CO, Pr PKs combined 

0-501m saccharose 18-0 519-9 6-03 6°36 276 8 Y=16-5 
T' =27-7 18-0 1173 6-66 6:38 28:1 EVI 
0-0125n HCO’, 18-5 28-6 7:29 6:39 27:6 Series A 
0-0125n Na’ 19-0 16-9 747 6°34 27:6 

19-5 523-4 6-03 6:40 27:7 

18-5 34-7 —- (6°31) 27-9 

Mean 6:38 Mean 27:7 

0-501m saccharose 20-5 43-9 7-098 6-471 Se Y=165 
T' = 22°44 20-5 43-9 7-092 6-465 a EVI and EMI 
0-01008% HCO’, 20-5 41-4 7-102 6-450 —- Series B 
0-01008n Na’ 20-5 41-4 7102 6-450 _ 

205 43-6 7-095 6-465 — 

20-5 43-6 7-005 6-465 


Mean 6-461 


THEORY OF THE HENDERSON-HASSELBALCH EQUATION — 261 


CHAPTER IX 


PRELIMINARY CONSIDERATIONS CONCERNING THE MODE 
OF COMBINATION OF CARBONIC ACID IN THE BLOOD 


In the first seven chapters of this work the relations existing between the 
apparent hydrogen ion activity, the CO, tension and the combined carbonic 
acid of the blood have been developed. These relations depended on the 
assumption that all the combined carbonic acid in the various phases of the 
blood is present as bicarbonate?, but on further scrutiny it will be obvious 
this proviso is by no means proved by the fact that the apparent hydrogen 
ion activity can be calculated from the modified Henderson-Hasselbalch 
equations. On the contrary it will be perceived that in the empirically deter- 
mined “constants,” corrections for combined carbonic acid not in the form 
of bicarbonate can very easily be hidden. 

When I started these investigations I thought that the corroboration of 
Henderson’s and Hasselbalch’s “proof” that all the combined carbonic acid 
in the blood was present as bicarbonate, would not involve great theoretical 
difficulties so that by taking the heterogeneity of the solution into considera- 
tion it would be feasible to determine whether the evidence was really sound. 
But the more I have pondered over the theoretical conditions, the greater the 
difficulties have seemed to be. I have nevertheless, as far as I have been able, 
fulfilled my original plan, inasmuch as, it seems to me, questions of con- 
siderable interest have arisen but I have furthermore endeavoured to establish 
the proof in another way in order to avoid the most important difficulties. 
The evidence will be found in chapter XI but for the present it will be shown 
how far we can go with the original idea. The value that gives the greatest 
trouble is the apparent activity coefficient of the bicarbonate ion in the liquid 
phase of the blood corpuscles, and uncertainty in evaluating this will produce 
a significant error in the calculations in this chapter and chapter XII. I do 
not think however the conclusions arrived at can be disputed from a quali- 
tative point of view but I quite realise that quantitatively, especially in this 
chapter, they may rather diverge from the results that may eventually be 

obtained. 
In chapter III an equation was evolved for the calculation of the apparent 
hydrogen ion activity in the water phase of a two-phase system. With this 
equation (115) we can estimate the amount of bicarbonate which will be 
found in the serum proteins and haemoglobin if the combined carbonic acid 
is present as bicarbonate: 


Poo. a 
(115) ap =Aww FEpg 


1 No account is taken here or henceforth of the small quantities of monocarbonate found in 
the blood (cf. chapter IT). 


262 E. J. WARBURG 


the condition for which is 


, 1LO0O- g(1-d) 
(117) Aw = K's 99 — 
from which we obtain d= es) eed eS ateyhtacivaatls saraneths (151) 


in which the significance of the constants is discussed in chapters II and III. 
Let us first consider serum. We found in chapter V 
Nay = Ag = 10 5-20 10: 
K’, can be determined from the equation 
(143) — log F, (HCO’;) = 0-46 ¥/c. 
For the determination of the apparent activity coefficient we only lack the 
value of c. 

Numerous and thorough investigations by Hamburger [1902], Hedin 
[1915], Ege [1920] and many others have shown that serum is almost isotonic 
(isosmotic) with a 0-9 °% sodium chloride solution. For further information 
respecting this I would refer the reader to Rich. Ege’s thesis. It would be 
erroneous however immediately to conclude that c was the same as in 0-9 % 
NaCl solution, that is 0-154, because NaCl mols are not the only osmotically 
active ones. Now the mols which are not electrolytes are only present in such 
a small concentration in serum that for this purpose we can neglect them 
entirely. But at the reactions here in question there are also some protein 
anions present in serum which neutralise corresponding amounts of cations. 
The quantity of these never exceeds 0-02n and presumably is usually about 
0-0ln, so we are justified in estimating c at between 0-16 and 0-18n. 

The volume of the disperse phase in serum is certainly very nearly 8 %_ 
and putting this in equation (151) we obtain 


c=016 K’, = 10-62 — 5:50 x 10-7 d = 0-32; 
c=0-18 KK’, = 10-%351— 5-61 x 10-7 d= 0-07, 
The calculation is rather uncertain because a slight error in K’, or in X) 


has a great effect, but it seems as though some bicarbonate is dissolved in 
the serum proteins. 


Table XLIX. Calculation of the value of d in the blood corpuscles 
with equation (151). 


d 
A 
epee ei | 
c R’, x 1077 q=40 q=35 
0-20 5°30 0-51 0-44 
0-225 5-41 0-47 0-40 
0-250 552 0-42 0:35 


The concentration of cations in the water phase of the blood corpuscles 
is not certainly known. It has been estimated here as between 0-2 and 0-25, 
The reason for this will be discussed later in chapter XII. 

Ay) = Ae) in horse blood cell fluid, according to the measurements reported 
in chapter VII, is about 6-37. The volume of the disperse phase in blood cells 


THEORY OF THE HENDERSON-HASSELBALCH EQUATION 263 


according to Ege’s determinations is between 35 % and 40 %. If d is caleu- 
lated with these values we obtain the results given in Table XLIX, using 
equation (58) with the constant 0-40, as a large part of the cations of blood 
cells consists of potassium. 

From the table it will be seen that the value of d, under the given assump- 
tions, lies between 0-35 and 0-51. 

As emphasised above it is quite possible that the calculation may be rather 
uncertain quantitatively but there is hardly any doubt that d is really a 
positive quantity. If d= 0 and ¢ and %) at the same time have the values 
employed in the calculation above, we can estimate K’, and ¢ since in that 
case the following holds: ae ans Xi 
and therefore for g = 40 we get K’, = 7-12 x 10-7 and c = 0-74n, and for 
q = 35, K’, = 6-57 x 10-7 and ¢ = 0-55n. These values for c are undoubtedly 
too high. 

It may therefore be taken as settled that a part of the bicarbonate is 
dissolved in the haemoglobin provided that all the combined carbonic acid 
of the blood is really present as bicarbonate, and the calculations also show 
that the bicarbonate is about half as soluble in haemoglobin as in a roughly 
n/4 KCI solution (the dispersion medium). 

This result is by no means antagonistic to what we already know about 
the solubility of salts in proteins. Thus 8. P. L. Sérensen [1915-1917] found 
that ammonium sulphate was freely soluble in egg albumin but, on the other 
hand, it compels us to inquire whether a part of the combined carbonic acid 
is not present in some other form than the bicarbonate ion (or undissociated 
bicarbonate in the haemoglobin phase). It would be interesting to estimate 
first what amount of the combined carbonic acid of the blood may be expected 
to be dissolved in the protein phases according to the above calculations. 

As an example let us take horse blood at 18° with a py*;,) value of 7-40 and 
at a CO, tension of 40 mm., and further we will assume the volume of the 
blood cells is 40 % of the total volume of blood, which according to Fig. 10 
corresponds to 19-6 vols. °% combined O, (Haldane 106). We now calculate 
with (130) the combined carbonic acid (B) making the necessary correction 
with the help of’Fig. 10, and we obtain 58-4 vols. %. From Fig. 7 we read 
that the partition ratio of combined CO, at this reaction (D) is 0-66. 

In 100 cc. blood therefore there are 58-4 cc. combined CO, (0° and 760 mm.), 
of which 40-6 cc. are present in the 60 cc. serum, equivalent to a content of 
67-6 ce. in 100 cc. serum; and in 40 cc. blood cells there are consequently 
17-8 cc. combined CO, corresponding to 44-6 cc. in 100 cc. blood cells. Of the 
60 cc. serum the water phase is 55-2 cc. and the protein phase (8 %) 4-8 cc., 
and if we assume d is 0-35, we obtain 39-4 in the water phase corresponding 
to 71-3 cc. in 100 cc. and 1-2 ce. in the protein phase which is equivalent to 
25-0 ce. in 100 ce. protein. If we carry out a similar calculation for the blood 
corpuscles and put g = 40 and d = 0-45, the water phase becomes 24 cc. and 


264 EK. J. WARBURG 


the protein phase 16 cc. There are then 13-7 cc. in the water phase, equivalent 
to 57-2 ce. in 100 cc. and 4-1 cc. in the protein phase, equivalent to 25-7 cc. 
in 100 cc. In the two disperse phases there are altogether 4-1 + 1-2 = 5:3 ce. 
combined CO, which is equivalent to about 9% of the total quantity of 
combined CO,. We have found that 71-3 vols. % combined CO, are present 
in the water phase of serum, while there are only 57-2 vols. % in the water 
phase of the blood cells. Now the apparent activity coefficient of the bicar- 
bonate ion is very nearly identical in the water phase of serum and blood cells 
because the greater cation concentration in the blood cells is almost balanced 
by the smaller Milner effect of the potassium ions when it is compared with 
the sodium ions, and the relation between the amounts of combined CO, in 
the two water phases will thus directly afford a measure of the relation be- 
tween the activity of the bicarbonate ion in them, so that 


4@yCO, in the water phase of the blood cells _ 0-8 
@HCoO, in the water phase of the serum , 


This result is by no means at variance with what we should expect, as it 
will be shown in chapter XII that on the basis of Ostwald’s [1890] and 
Donnan’s [1911] partition law such a distribution is to be looked for. 

The above results are condensed in Table L. 


Table L. 
Vols. % combined Cc. combined 

2 2 

100 ce. blood at py: 7°40 and 40 mm. ... oes 58-4 58-4 
In the serum of the above (60cc.) ~ ... ay 67-6 40-6 
», water phase of the serum (55-2 cc.) ahs 71:3 39-4 

z Pheri phase of the serum (48 ce.) ... 25-0 1-2 

»» blood cells (40 cc.) 44-6 17:8 

»» water phase of the blood ‘cells (24 ce. ys 57-2 13-7 

», protein phase of the blood cells (16 ce. y 25-7 4-1 


The amount of CO, combined in a form other than bicarbonate would thus 
appear to be small and, according to the above calculations, at most 9 % 
of the total. We will now try to get an idea of what form of combination 
this might theoretically be. 

The combined carbonic acid is determined as the difference between the 
total dissociable combined OO, and the dissolved amount calculated by 
Henry’s law. The relative absorption coefficient of blood and serum has been 
determined indirectly by Bohr [1905], the assumption being that it is the same 
forall gases. This is of course only approximately true but the divergence is not 
so large that a grave error can arise if the CO, tension is not too high. It is how- 
ever possible that CO, could be adsorbed at the interface between protein and 
the dispersion medium, That such an adsorption actually takes place in blood 
is claimed by Findlay for various gases as will be referred to in the next 
chapter but he undoubtedly overestimates the significance of his experiments. 
It must also be noted that A;,) could not be independent of the reaction if such 
an adsorption took place for reasons identical with those that will be advanced 
at the end of this chapter against the existence of a protein-carbamino-acid. 


THEORY OF THE HENDERSON-HASSELBALCH EQUATION 265 


The possibility may also be raised that some of the bicarbonate may be 
present as undissociated protein bicarbonate (complex salt). This assumption 
is however not quite correct because in alkaline reactions only very small or 
negligible amounts of protein cations are present, since proteins function 
almost entirely as acids and so there can be no question here of undissociated 
protein bicarbonate. At reactions in the region of, or more acid than, 10~7*° 
some haemoglobin cations are in all probability formed so that there is the 
possibility of the existence of undissociated haemoglobin bicarbonate and in 
such a case p,.) might decrease with increasing hydrogen ion activity, but 
in chapter VII we saw this was not so. We can therefore conclude that un- 
dissociated bicarbonate does not exist in sufficient quantity to compromise 
our theories and calculations. Further it is possible that the bicarbonate ion 
might be adsorbed at the interface between protein and salt solution. As a 
matter of experience however such an adsorption only takes place when the 
disperse phase has a charge of opposite sign to that of the adsorbed ions, and 
so the theory of the adsorption of the bicarbonate ion by proteins is untenable 
for the same reason as applies to undissociated protein bicarbonate. Further- 
more the problem of the adsorption of the ions is presumably identical with 
the problem of the large Milner effect of the colloid ions, according to the 
views of which Prof. Bjerrum has kindly given me the benefit. There remains 
the possibility that some of the CO, combines with the protein in some other 
form than protein bicarbonate. Such a form of combination has been stated 
by Siegfried to be protein carbamino-acid. 

It will be clear to the reader that the combined carbonic acid referred to 
in this work is only the dissociable combined carbonic acid, that is to say the 
carbonic acid which can be liberated by the addition of acid or by diminishing 
the CO, tension; that which can only be split off from the organic molecules 
by such vigorous processes that the latter are destroyed is naturally not 
included, it can moreover be shown that this dissociation of the carbonic 
acid combination of the blood is completely reversible. If protein carbamino- 
acid took part in the reversible carbonic acid combination ,) (or A,,,)) could 
not be constant in the fluid of the blood cells because the non-ionised com- 
bined H,CO, would then vary with the CO, tension. This evidence that all 
the combined carbonic acid is really ionised in the blood is well supported by 
the determinations of pA;,,) in chapter VII, but is weakened to some extent 
by the determinations being so extremely difficult to carry out. But it also 
involves the assumption that q and d in equation (117) do not vary appre- 
ciably with the hydrogen ion activity. From a close examination of (117) 
however it appears the variations must be fairly large to make themselves 
felt in the determinations in question. 

As it would certainly involve very great experimental difficulties to proceed 
further with the questions raised in this chapter there is every inducement to 
examine the literature for experiments which can throw light on these 
problems. 


Bioch. xv1 . 18 


266 K. J. WARBURG — 


In conclusion it may be remarked that the experiments and views put 
forward give no information as to whether the bicarbonate dissolved in the 
protein phase is present exclusively as ions, or as undissociated salt as well. 


R&suME. ; 
It has been shown that a small part of the combined carbonic acid is not 
present in the dispersion medium of the serum and blood cells, and the 


manner of combination of this amount of combined acid is discussed. It is 
probable that it is dissolved in the protein phases. 


CHAPTER X 


A BRIEF HISTORICAL REVIEW OF THE OLDER THEORIES OF 
THE COMBINATION OF CO, IN THE BLOOD. 


The first to extract gas out of blood was John Mayow of Oxford (quoted 
from P. Bert), who reported his results ina paper in 1674. He thought the gas 
obtained was what we now call oxygen. 

In 1783 Pietro Moscati of Milan stated he had extracted a gas out of blood 
and he believed he had proved it to be carbonic acid (fixed air), but his 
evidence was not good as by allowing blood to stand for 24 hours with lime 
water he obtained a precipitate [1784]. 

In 1799 Humphrey Davy proved that CO, could be driven out of blood 
by heating it, and in 1814 Vogel showed that CO, could be pumped out of 
blood. Up to 1837 when Magnus finally established that CO, could be pumped 
out of blood, the question was one of considerable interest as it was closely 
bound up with the current discussion on where oxidation took place in the 
organism. 

Lavoisier had originally discussed whether oxidation occurred in the peri- 
pheral circulation or in the lungs. In his later works he claimed to have 
proved it took place in the lungs (according to W. Edwards and Ch. J. B. 
Williams). He believed that the carbon in a state of fine division was carried 
by the blood to the lungs and burned up there in association with oxygen. 
It became therefore of prime importance at that time to ascertain whether 
carbonic acid (CO,) was free in the blood or only in a state of combination. 
In those days it was evidently not considered sufficient that carbonic acid 
could be driven out of the blood by heat, or by treatment with another gas 
such as hydrogen or atmospheric air, because it was supposed these gases 
replaced carbonic acid in loose combinations although it was partly realised 
similar processes took place in the lungs. 


THEORY OF THE HENDERSON-HASSELBALCH EQUATION 267 


The first to have pumped CO, out of blood seems to have been, as already 
mentioned, Vogel [1814] and, apparently almost simultaneously, Vaquelin. 
This last investigator's experiments were never published by himself but 
were recorded by W. Edwards [1824]. 

Between 1814 and 1837, Scudamore, Brande and others succeeded in 
pumping CO, out of blood, while Darwint, Davy [1803], Stromeyer [1832], 
Misterlich, Gmelin and Tiedemann [1834] failed in an attempt to do so. 

Stevens’ [1832, 1835] experiments and theories are especially interesting. 
He was a Scotsman who lived the greater part of his life in the then Danish 
West Indies, and he seems to be almost forgotten in scientific literature, which 
is partly due to the immense range of his works, but he appears to have 
introduced a treatment for cholera which although based on faulty experi- 
mental material, constituted a real step forward at the time (cf. Observations 
on the Nature and the Treatment of the Asiatic Cholera, London, 1853). 

In his book of 1832 Stevens put forward the theory, on pp. 22-26, that 
the carbonic acid of the blood is present as alkaline carbonates and as free 
CO,, comparing the change of reaction of serum on exposure to the air with 
the change which takes place in the alkaline mineral water highly charged 
with CO, from a spring at Saratoga in New York State, when it is allowed 
to stand in the air. 

In 1835 he communicated a series of experiments to the Philosophical 
Transactions which showed that he was unable to pump CO, out of venous 
blood immediately after it was taken but he succeeded after the blood had 
stood for some time in contact with hydrogen or air. It had been repeatedly 
shown that blood could absorb considerable amounts of CO,—a fact which 
Stevens himself demonstrated experimentally—but he could not get the CO, 
out again without letting the blood stand some time in the air or hydrogen. 
Stevens thought therefore that these gases liberated CO, allowing it to be 
pumped out. It is perfectly clear to us now that the disagreement between 
the different observers was due to their gas pumps being of different degrees 
of efficiency and those who had the worst pumps could not pump CO, out of 
freshly drawn blood. The one who finally proved with certainty that the 
blood contained CO, which could be expelled by blowing hydrogen through 
it or by the exhaust pump was Magnus. His article in Poggendorft’s Annalen 
of 1837 is still worth reading not only on account of its historical interest but 
also in view of the great insight of the author, and on perusing it, it is striking, 
as is the case with so many other works of the first half of last century, how 
little physiological problems have changed their character and how often 
the same phenomena are rediscovered. 

A sentence in an article in 1834 by R. Hermann of Moscow is of interest. 

On the 5th September, 1833, Drs Stevens, Markus, Wyllie, Janchen and 
R. Hermann met in Moscow for an investigation.—‘‘Nachdem man sich 
einstimmig dariiber ausgesprochen hatte, dass das Serum als eine Fliissigkeit 

1 Quoted from: Williams [1835]. 
18—2 


268 K. J. WARBURG 


zu betrachten sei, die neben doppelt-kohlensaurem Natron noch freie Kohlen- 
saure erhalte (on account of the reaction of the serum), ging man zu Unter- 
suchung des Blutes iiber.”’ 

It was not all however who realised that blood contained free carbonic 
acid. Misterlich, Gmelin and Tiedemann [1834], and also Davy [1803] and 
others, thought that blood only contained combined CO, while one gets the 
impression that Magnus [1837] regarded all the CO, as free. Even in the middle 
of the sixties W. Preyer [1867] could proclaim with success that no free CO, 
was to be found in the blood because it could not be present in an “alkaline” 
liquid. When Zuntz [1868] pointed out that free CO, might be found in blood 
when it was in equilibrium with an atmosphere containing CO, the question 
was finally decided. P. Bert however in a paper in 1878 opposed this view. 

Magnus, whose technique was a considerable advance on that of earlier 
workers, contributed analyses which even in those early days. gave fairly 
correctly the quantity of CO, in the blood. These experiments were the best 
until 1857 when Fernet in France and Lothar Meyer in Germany published 
extensive investigations on the subject. Fernet’s technique—which Meyer 
also used—consisted in driving the gas out of blood by means of a vacuum 
produced with the help of water vapour and afterwards analysing the gas 
driven out. Fernet and Meyer thought the combined CO, in blood could be 
divided into two kinds, namely that which could be liberated with a vacuum 
and heat alone, and that which required the addition of acid. About this time 
Bunsen’s first absorption coefficients appeared and it was then possible to 
calculate the amount of combined CO, in the blood. The values found by 
Fernet and Meyer were however too low—in fact in the case of CO, they were 
no better than those of Magnus—on account of the relatively bad vacuum. 
Fernet proposed the theory that the CO, of the blood was combined with 
phosphates as a double salt but Heidenhain and Meyer [1863] showed a few 
years later that Fernet’s analyses with respect to the double salt were not 
correct, and Sertoli proved in 1868 that the amount of free phosphates in the 
blood was so small that it could be entirely neglected as regards its capacity 
for combining with CO, in view of the relatively large amount of combined 
CO, which had been found. Later investigations have fully supported Sertoli’s 
findings and according to the latest. experiments which have been collected 
amongst others by Poul Iversen [1920], blood and serum contain about 
m/ 1000 dissociated phosphate. The combination of CO, in phosphate solutions 
has nevertheless had no small influence on our knowledge of CO, compounds 
in blood, as it is a rather simple example of a CO, combination in an electrolyte 
the dissociation of which varies with the reaction. Heidenhain and Meyer 
[1863], Setschenow [1879], Henderson and Black [1908] and K. A. Hasselbalch 
[1910] have all used phosphates as a solution comparable with blood, and the 
author will also report some similar experiments later on in this work. 

Ludwig [1865] and his collaborators Schéffer [ 1861, 1868] and Setschenow 
[1859], in the latter half of the “fifties” and in the “sixties,” carried out a 


THEORY OF THE HENDERSON-HASSELBALCH EQUATION 269 


number of experiments upon the CO, compounds in blood, and these investi- 
gators also considered that all the CO, could not be extracted out of the blood 
(with Ludwig’s mercury air-pump) without the addition of acid. It was not 
until Pfliiger [1864] in conjunction with Geissler constructed a new pump 
that the extraction was so complete that the yield was not increased by adding 
acid. Pfliiger also showed that all the CO, could not be pumped out of serum 
(when it was separated from the blood corpuscles) without adding acid. Pfliiger 
finally showed that in addition to the CO, which blood yielded to a vacuum 
a further supply could be liberated from sodium carbonate. Pfliiger fully 
realised that pumping out in the presence of sodium carbonate was accom- 
panied by a transformation of monocarbonate into bicarbonate and it was 
owing to this that the CO, was driven out (cf. H. Rose [1835] and Marchand 
[1846]). He considered the acids which drove out the CO, were haemoglobin 
and the serum proteins. Hoppe-Seyler thought that haemoglobin only acted 
as an acid under the influence of the exhaust pump, being destroyed during 
the pumping, in spite of the fact that such an assumption is unnecessary for 
explaining the phenomenon, and, so far as I am aware, a proof has never 
been furnished; the hypothetical acid could only drive out a very small 
quantity of CO,, because Gaule [1878] and Zuntz [1882] have shown that 
after pumping out, blood combines with almost as much acid as before. 

Pfliiger [1864] thought the difference between blood and serum consisted 
in the fact that less acid was formed in serum on pumping out than in the 
blood corpuscles. 

Sertoli [1868] and Setschenow [1879] both showed that globulin is an acid 
which can expel CO, from sodium carbonate, and taking everything into 
consideration there is now hardly any doubt that the proteins really possess 
an acid character. As it is also an aecepted fact that they can combine with 
acid their ampholytic nature is proved, It is an observation over 100 years 
old that the proteins function both as bases and acids. About the beginning 
of the present century they were classified as ampholytes, at which time the 
peculiar properties of this group of electrolytes became known, and their 
physico-chemical characteristics were thoroughly worked out by Bredig, Walker, 
Hardy and Pauli, and later particularly by Sérensen and Michaelis and their 
numerous collaborators, while Henderson and Spiro and Hasselbalch have 
rather elaborately dealt with the consequences of the ampholytic character 

in relation to CO, combinations in the blood. 

In 1882 Zuntz, in Hermann’s handbook, collected the investigations 
relating to the combination of carbonic acid in the blood in a deservedly 
celebrated monograph. He was able, largely on the basis of his own and 
Setschenow’s experiments, to enunciate the theory for the combination of 
carbonic acid, that CO, was combined with the alkali of the blood in such a 
way that carbonic acid and the blood proteins divided the alkali between 
them, and also the haemoglobin at high CO, tensions acted as a base and thus 
combined with CO, itself. 


270 E. J. WARBURG 


Some years before, Setschenow [1879] had come to a very similar conclusion 
about the combination of CO,, the only difference being that he interpreted 
CO, combinations at high CO, tensions not as a production of haemoglobin 
bicarbonate but as a more specific reaction for haemoglobin. 

The view which is held to-day regarding the CO, compounds in the blood 
only slightly differs from Zuntz’s, the most important advance being the 
knowledge that the proteins belong to the group of ampholytes. The premises 
upon which Zuntz built his theory were pre-eminently the following: 

(1) The well established fact that proteins combine with both acids and 
bases. 

(2) Pfliiger’s observation that haemoglobin can drive CO, out of sodium 
carbonate, in conjunction with Sertoli’s and Setschenow’s demonstration of 
the same thing for globulin and paraglobulin. 

(3) Setschenow’s demonstration that the blood can combine with more 
CO, than corresponds to the difference between the cations and anions the 
dissociation of which varies with the reaction (expressed in modern terminology). 

Of these premises only Pfliiger’s has later become open to doubt as 
Buckmaster tried to show in 1917, but the latter author is certainly wrong. 
His results can only be explained by his having added too much alkali so that 
the amount of bicarbonate formed is so small in relation to the amount of 
monocarbonate that there is not a measurable CO, tension in the system and 
thus CO, could not be pumped out. Quite recently Adolph [1920] has also 
shown Buckmaster was in error, but the latter’s experiments have raised so 
much discussion that I think it worth while reporting one of two experiments 
I have undertaken which were in good agreement and which well support 
Pfliiger’s experiments. 

Experiment. 

Defibrinated ox blood kept in the ice cupboard from 19. viii. till the be- 
ginning of the experiment. 

The blood corpuscles were separated by centrifuging. 

On 20, viii. 6-98 ec. of the blood corpuscle emulsion in equilibrium with 
40-5 mm. CO, were pumped out into an acid-free evacuation receiver. It con- 
tained 39-1 vols. % CO, and 38-2 vols. % O, (0° and 760 mm.). 

On 21. viii. 7-12 ee. blood cell emulsion in equilibrium with 40-9 mm. CO, 
and 146-1 mm, O, at 18° contained 38-5 vols. % CO, and 32:3 vols. % O, 
on pumping out in the presence of boric acid. 

In an acid-free evacuation receiver were put 6-64 cc, NaOH in equilibrium 
with 40-9 mm. CO,. This NaOH, according to previous estimations, combined 
with 219-7 vols. %%, and there was physically absorbed 5-01 vols. %, that is 
altogether 224-7 vols. °% (0° and 760 mm.). 

Further there was put in the receiver 10-55 ec. of the blood cell emulsion. 


1898 45). » 


0° 760 mm, 
In 6-64 cc, NaHCO, ... «» 1492 ec, CO, 
Therefore 10-56 cc. of the above blood cell emulsion in equilibrium 
with 40-9 mm, CO, and 1461 mm, O, contained . eh ww 4:06 


THEORY OF THE HENDERSON-HASSELBALCH EQUATION 271 


0° 760 mm. 
In one hour there was pumped out... Wen ... 18-35 ec. CO, 
After the addition of boric acid for 7 mins. a a further saa Pep ear ei yee 


Each pumping out took place at 38°. 19-00 ,, 


Pfliiger’s [1864] phenomenon that all the combined CO, of the blood can 
be pumped out and further CO, can be driven out by addition of sodium 
carbonate shows beyond doubt that substances are present in the blood 
which act as acids at reactions more alkaline than the neutral point, but does 
not prove that all the combined CO, in the blood is in the form of bicarbonate 
because every compound of CO, with another substance will be dissociated 
as soon as the CO, tension is diminished. 

It is justifiable to assume that the electrolytes of the ‘iinod the dissociation 
of which varies with the reaction are mainly proteins, particularly paraglobulin 
(Setschenow [1879]) and haemoglobin (Zuntz [1868], Mathieux and Urbain, 
Setschenow [1879], Bohr [1891, 1898, 1904, 1909], Hasselbalch [1916, 2], 
Campbell and Poulton [1920] and many others). But Setschenow [1879] has 
shown that lecithin also acts as a weak acid, and it is known that phosphates 
act in the same way although as already mentioned they are only found in 
quite small concentrations. 

There are some theories in the literature which differ from the Zuntz- 
Setschenow theory. The first is Bohr’s [1904]. He believed the combination 
of CO, took place in the following way: 


CO, + 4,HbA, = CO, (A,Hb) + A, 


(the symbols being Bohr’s own). In other words he imagined haemoglobin 
contained an acidic and a basic group and that at the same moment in which 
it combined with CO, it split off a new acid. Bohr has worked out this relation 
on the assumption that there is no alkali in the system and that the process 
is independent of the reaction of the solution. 

So far as I can see the only proof which Bohr has tried to give that the 
process really proceeds is that the CO, compounds in a haemoglobin solution 
and in the blood can be calculated by an equation derived under the above 
assumptions, but it is generally accepted a proof of this kind is always very 
weak, Although I cannot assert that a reaction of the kind in question does 
not take place particularly when developed according to Siegfried’s [1905, 1, 2; 
1908, 1, 2; 1910] theory, as will be explained later on, it is probable at any 
rate that such a process only proceeds to a very slight extent. Bohr has reported 
numerous experiments on CO, compounds in pure haemoglobin and claims 
to have shown that various modifications of haemoglobin exist which can 
pass into one another but he is unable to show how. This complicates Bohr’s 
view still more and it seems to me it would be useless to express an opinion 
on the CO, compounds in pure haemoglobin solutions without carrying out 
fresh experiments. 

Siegfried and his collaborators have shown that by treating amino-acids 


272 E. J. WARBURG 


with CO, in the strongly alkaline barium hydroxide or calcium hydroxide 
solution at a temperature about freezing point a complex compound of car- 
bonic acid and amino-acids is formed and he has further shown that the same 
compound is also formed even if no Ba or Ca is present in the solution. When 
a solution in which a complex CO,-amino-acid compound is formed is heated 
and Ba or Ca is present, a precipitate of the carbonates of these elements 
will separate out. 

Siegfried believes that complex compounds of the amino group and CO, 
are formed according to the following scheme: 


WO H 

Pe 
—R—N—H+CO, Core 
COOH COOH 


and that they are strong acids which form salts as follows: 
H 
Pst 
—R—N—COO 


obo Ba 


(Siegfried seems to imply by this formula that the barium salt is not dis- 
sociated; this is however unimportant in the present connection.) 

In my opinion Siegfried has only shown that CO, is combined in a complex 
manner with the amino group in the cold and in a very alkaline reaction, 
and he appears to have failed to supply any proof that the compounds formed 
are acids, let alone strong acids. 

Siegfried [ 1908, 2] claims to have shown that proteins form with CO, com- 
plex compounds similar to the amino-acids. The experiments he considers 
most telling in this respect are certainly not convincing and they are probably 
quite wrongly interpreted. Siegfried passes CO, through a protein solution 
and shows that the conductivity increases more than can be accounted for 
by the CO, alone. He concludes from this that dissociated carbamino-acids 
are formed but it seems to me more natural to ascribe the effect to the forma- 
tion of dissociated protein bicarbonates. Siegfried has asswmed that protein 
carbamino-acids are formed at the temperature and reaction of the body and 
that it is a reversible reaction and they would be formed in greater or less 
amounts according to the prevailing CO, tension, and further that these 
compounds participate in the transport of CO, from the tissues to the lungs. 
This view has gained some acceptance (cf. for example Hammarsten [1914] 
and Wilstitter and Stoll[1918]) but in my opinion there is no evidence that 
the compounds exist or are dissociable at physiological temperatures, reactions 
and CO, tensions. Per contra it has naturally not been proved such bodies 
do not exist, 

The third theory relating toCO, compounds in the blood is mainly supported 
by experiments of Findlay and his collaborators [1908, 1910, 1911, 1912, 1918, 
1915). The theory itself has not been actually propounded but Findlay’s 


THEORY OF THE HENDERSON-HASSELBALCH EQUATION 273 


experiments and his whole method of reasoning in his papers certainly suggest 
the enunciation of such a theory. Findlay’s idea is that gases can be adsorbed 
by colloids and that it is possible such an adsorption takes place in the blood. 
The experiments of Findlay and his collaborators have from a physiological 
standpoint had an unfortunate fate. In 1908 Findlay formulated the working 
hypothesis referred to and investigated in the following years the adsorption 
of various gases in several colloid solutions. By this great pioneer work 
Findlay has brought to light a considerable number of previously unknown 
conditions which still await further investigation. From his experiments he 
could conclude that inactive gases are frequently adsorbed in colloid solutions, 
and that in some cases CO, is adsorbed and in others combined (as bicarbonate). 

In 1911 Findlay and Creighton reported some experiments on the absorp- 
tion of oxygen, nitrogen, nitrous oxide, carbon monoxide, and carbon dioxide 
in blood and serum but these experiments, from which he concluded that 
oxygen, nitrogen, nitrous oxide and carbon monoxide were adsorbed by the 
proteins, are from a technical point of view so indifferent that the results are 
far from convincing. Findlay and Creighton have not taken sufficient notice 
of the fact that there are large quantities of combined O, and CO, in the blood 
which are dissociable and that it is necessary to make sure the compounds of 
these gases have not changed if one wishes to determine the absorption con- 
ditions of another gas by Findlay’s technique; but this has certainly not been 
the case in the experiments referred to. In physiological circles this has been 
generally known since the days of Fernet and Meyer, and it is only the fact 
that Findlay has been trained as a chemist that makes it intelligible he should 
overlook the point. 

I will quite briefly show from a single series that Findlay and Creighton’s 
experiments are much too uncertain to allow of any conclusion being drawn. 
Findlay and Creighton have determined the absorption coefficient (Ostwald’s) 
at a number of different tensions. It is easy with the help of the equation 
first employed by Fernet [1857] and Meyer [1857] to estimate the absorption 
coefficient between two tensions. I have performed this and then calculated 
this value as the relative absorption coefficient at the temperature of the 
experiment, in pure water. 

A few of the experiments will be found recalculated in the following table: 


CO in blood which had been exhausted 


before the saturation* CO in blood not exhausted 
Relative absorp- Relative absorp- 
mm, Hg CO, tion coefficient mm. Hg CO, _ tion coefficient 
751-871 0-32 751-944 1-12 
871-1056 1-61 944-1144 1-52 
1056-1140 1-08 1144-1371 0-95 
1140-1274 1-07 1371-1434 2-75 
1274-1543 0-90 1434-1528 1-80 


It appears further from other experiments that the absolute values found 
for the absorption coefficients are on an average much lower than they have 


1 With a water exhaust pump. 


274 E. J. WARBURG 


been found before, in fact inexplicably low. There is no alternative but to con- 
clude that Findlay’s experiments with blood are subject to such a large technical 
error that they cannot be utilised. The issue itself which Findlay has raised 
is very interesting. Jolyet and Sigalas [1892] claim to have shown that blood 
absorbs more nitrogen than it should if the absorption coefficient were the 
same as for pure water. Bohr [1909] and later Buckmaster have been able to 
confirm this although Buckmaster [1911-1912] only finds about 10% too much. 
There is yet another fact which might be brought forward in support of the 
contention. Bohr [1905] has found the relative absorption coefficient for 
hydrogen in blood is 0-92 and Fahr [1912] found, approximately the same 
values. 

In serum Bohr estimated 0-975 as the coefficient for hydrogen. 

In serum Fahr estimated 0-94 to 0-95 as the coefficient for hydrogen. 

In blood cells Bohr estimated 0-81 as the coefficient for hydrogen. 

In blood cells Fahr estimated 0-78 to 0-88 as the coefficient for hydrogen. 

R. Siebeck [1909] has found values for nitrous oxide in blood, blood cell 
solutions and haemoglobin solutions which are a little over 1 (for blood 1-03), 
while Lindhard [1914] found about 0-97 for blood; and for serum Siebeck 
found the same value with nitrous oxide as Bohr found with the hydrogen?. 

The disagreement between the depression of solubility for hydrogen and 
nitrous oxide is pronounced. Should this be supported by experiments it 
points strongly to complicating factors in the solubility of the gases in protein 
solutions. 

Thorup [1887] states he has demonstrated the existence of a CO,-haemo- 
globin compound spectroscopically. He has shown that reduced haemoglobin 
when treated with high tensions of CO, develops a different absorption 
spectrum. From a modern physico-chemical point of view we cannot judge 
by this method whether complex combinations are formed or whether the 
haemoglobin becomes ionised, because electrolytes with characteristic spectra 
usually change their absorption bands on ionisation. 

In conclusion I will discuss a view put forward by Jaquet [1892]. He has 
asserted that the CO, combination curve of the blood, when plotted with the’ 
CO, tensions along one axis and the combined CO, along the other, is shifted 
parallel with itself on addition of acid or base to the blood. If Jaquet’s state- 
ment were right it would prove that the CO, combination with haemoglobin 
was independent of the reaction and thus was not a bicarbonate compound 
as it appears to be from the reasoning in the next chapter. But Jaquet’s 
experiments are a long way from proving his contention because in the first 
place the curves are not truly parallel and secondly they are produced by 
points corresponding to experiments with different samples of blood. 


' D. D. van Slyke and W. C. Stadie (J, Biol. Chem. 1921, 54, 1) in accordance with earlier 
investigators have found greater quantities of N, in the blood than is to be expected from Henry’s 
law. A similar observation was made by Smith, Dawson and Cohen (Proc, Soc, Lap, Biol, and 
Med. 1919-20, 17, 211), quoted from van Slyke and Stadie. At the Finsen Institute I have myself 
frequently had a similar experience. 


THEORY OF THE HENDERSON-HASSELBALCH EQUATION = 275 


Résume. 


A search of the old literature does not disclose any facts which would lead 
us to abandon Zuntz’s theory modified in accordance with modern physico- 
chemical requirements. 


CHAPTER XI 


FURTHER CONSIDERATIONS AND EXPERIMENTS ON THE NATURE 
OF THE COMBINATION OF CO, IN THE BLOOD, ILLUSTRATED BY 
THE “CARBONIC ACID COMBINATION” CURVE}. 


In a solution in equilibrium with air containing CO,, a quantity of CO, 
is found to be dissolved according to Henry’s law, as repeatedly stated. This 
CO, can be extracted from the solution by diminishing the CO, tension in 
the gas phase over the solution, and further by this process more CO, which 
must have been reversibly “combined” in the solution can often be liberated, 

It has frequently been stated in earlier chapters that the combined CO, 
may be present 

(1) as adsorbed CO, if the eigen: is heterogeneous, 

(2) as carbonate ions (monocarbonate and bicarbonate), 

(3) as undissociated carbonate (complex salt), 

(4) as adsorbed carbonate which will presumably be identical with (3), 
since the molecular groups lying on the surface in a high molecular solute 
may be expected under certain circumstances to form compounds with the 
carbonate, 

(5) as a dissociable combination in which CO, takes part after assumption 
into the molecule, e.g. carbamino-acids. 

At reactions more acid than ay: = 10-* we can regard the carbonate ions 
present solely as bicarbonate (cf. chapter II). 

We will first examine systems in which adsorption combinations and 
complex salts are not found. In general, the following equation for the 
equilibrium of the positive and negative charges is valid for any solution 
whatever: 

Cy + Wy + BC yyy + se NC yt, «.. n Charges 
= C4, + 2C,’ pag = mf +. NC 4... m Charges, ...... (152) 
where Cy", stands for the concentration of monovalent cations and C 4’, for 
the concentration of monovalent anions and so on. As hydrogen and hydroxy] 
ions at reactions not far removed from the neutral point only occur in very 
small concentrations we can neglect them as terms in the above equation. 


1D. D. van Slyke [1921, 1] has recently reviewed the question of the combination of CO, in 
blood from the theoretical as well as from the clinical standpoint. Those interested are referred 
to the original articles which only became known to me after the following had been written. 


276 E. J. WARBURG 


In a solution in which only strong acids and their salts are present combined 

CO, cannot exist as there are no cations to counterbalance the bicarbonate 

ions. But in a solution containing a free base the hydroxy] ions will react thus: 
CO, + OH’ = HCO’. 

Under the given conditions this reaction practically proceeds quantitatively. 

In such a solution the combination of CO, will, as is well known, be inde- 

pendent of the pressure (if monocarbonate is not formed). 

If a weak base or the salt of a weak acid is present in a solution, CO, will 
be combined in amounts varying with the CO, pressure if the reaction is 
suitable. : 

For a weak monovalent acid, as shown in chapter I, the following equation 
is valid: 


Ay x G4’ 
Be Rn de ey gerry (153) 
On saturation the following reaction 
CO, + H,O +.A’3 = HCO's + HAy uses. ecceeeee. (154) 


proceeds in such a way that a part of A’; is substituted by HCO’, until the 
equilibrium as4 


CO's 
1aqAat COs HAt eee ee) (155) 
is reached. 
If a weak monovalent base is present in a solution the reaction 


CO, -+ My OH = My + HCO’, ooeeeeecsesessss (156) 


proceeds until the following equilibrium is reached : 
am: K ACO! 
warn * Kaan = Rig ee ee (157) 

The base is thus dissociated with the production of cations and bicarbonate 
ions. When equilibrium is attained an amount of bicarbonate ions will be 
present in the solution equivalent either to the cations the dissociation of 
which does not vary with the reaction (which before equilibrium were balanced 
by other anions the dissociation of which varies with the reaction) or to newly 
formed cations the dissociation of which varies with the reaction. The amount 
of combined CO, therefore becomes a measure of the available cation concen- 
tration. D. D. van Slyke and his collaborators have used the term “alkali 
reserve” for this amount (in serum) and E. Jarliv has employed the almost 
analogous expression “available alkali.”’” The reason I consider these ex- 
pressions should be given up is because of the possibility of misunderstanding 
them. Thus Parsons, Davies, Haldane and Kennaway have erroneously 
assumed that in the “maximum” amount of CO, combined we have a measure 
of the difference between the cations and anions the dissociation of which 
does not vary with the reaction, 
In order to illustrate the above I have performed experiments with sodium 
hydroxide solutions to which NaCl and phosphate solutions were added. The 
experiments will be found in Table LI and graphically in Fig. 15. 


: 


THEORY OF THE HENDERSON-HASSELBALCH EQUATION 277 


Table LI. 
With equation (171) With equation (172) With equation (180) 
a, —_. _eooeooor~ 
Vols % Vols. % Vols. % Vols. % 
mm. Hg CO, pH’ aH’ x10-7 CO, Diff. pH’ CO, Diff. © ff. 
CO, combined calc. cale. combined Vols. % cale. combined Vols.% combined Vols. % 
1 2 3 4 5 6 i 8 9 10 ll 
s {2—5) (2—8) (2—10) 
I. 30 cc. m/15 KH,PO, +50 ce. 0-:04941n NaOH + H,0 to 100 cc. + 0-42 g. NaCl. 20°. 

7-0 16-1 7-585 0-260 17:0 —0-9 7-608 17-6 -1:5 15-9 +0-2 
127 194 7-407 0385 199 -05 7418 202 -08 197 -03 
23:4 235 «7-225 «(0-506 = 238-7 2-0-2 7-290 888-0388 - 0-1 
46-4 28:0 7:004 0990 287 -O7 7015 29-1 -09 283. . -03 
76-5 31-9 6-844 1:43 32-7 -0:8 6-854 33-2 -1:3 32-4 -0°5 

1723. 384 6571 2-68 393 -09 6582 308  -14 376 +08 
745-0 487 6038 9-15 488 -01 6039 492 -05 491 -0-4 


Il. 6 cc. m/15 Na,HPO,+24 cc. m/15 KH,PO,+50 cc. 0-04941n NaOH +H,0 to 100 cc. +0:42 g. 
NaCl. 19°. 


86 49-2 7954 O-111 49-4 -0-2 17-956 49-4 —0-2 43-1 +61 
144 51-0 7-746 0-180 51-2 -0-2  17:747 51:3 —0°3 47-8 +3-2 
20-2 52-1 7608 0247 52-9 -0-8 7-614 52:8 —0-7 50-7 +14 
26:7 53-5 7-498 0-318 54-3 -08 7-504 54-6 -1-l 53-2 +0°3 
47-3. 57-7 7-283 ©0521 58-4 -0-7 7-288 58-6 -0-9 57-9 —0-2 
82-5 62-0 7-073 0-846. 63-0 -10 7-080 63-0 -1-0 62-7 —0-7 

436-4 77:5 6-446 3-58 78:8 -13 6-453 78-1 —0-6 76-4 +11 
750-2 80-6 6-228 5-92 82-9 -23 6-240 82-1 —15 81-4 —0-8 


IIT. 30 ce. m/15 Na,HPO, +50 ce. 0-04941n NaOH + H,0 to 100 cc. +0-42 g. NaCl. 20°. 


52 6563 8240 00576 567  -04 8242 568 ~-05 483 +80 
125 588 7878 0132 588 00 7878 588 00 541 +47 
196 60-4 7-694 0-202 60:7 -03 7696 608  -04 58-1 +23 
25:2 624 7599 0-252 618 +06 7595 619 +05 60:2 +22 
435 65-4 «= 7-383. 0-415 «4 00 7:383 65-4 00 649 +05 
810 701 7142 O721 73 j@2 17144 704 £-03 70-1 0-0 
186-2 77:4 6824 1:50 78-1 -0-7 6828 781 -07- 770  +0-4 
4144 842 6513 3-07 855 -13 6520 © 85-7 -15 838 +04 
731-5 875 6283 5-21 90:1 -26 6296 902 -27 888 -1:3 


IV. 24 ce. m/15 Na,HPO,+6 cc. m/15 KH,PO, +0-0243n Na,CO, to 100 cc. +0-85 g. NaCl. The 
two first experiments at 18°, the others at 19°. 


69 32:8 7835 0-146 33:3 -05 7-841 33-4 0-6 27-6 +5-2 
104 341 7-673 0-212 34:9 -08 7-683 35-0 -0-9 31-4 +2°7 
21-1 37-6 7-416 0-384 38-6 -10 17-427 38-7 -1-1 37-3 +0°3 
42-7 42-4 7-162 0-689 43-6 —-12 7174 43-9 -1-5 43-1 —0-7 
85:3 48-6 6-921 1-20 49-3 -0-7 . 6-927 49-5 -0-9 48-7 -O-1 

2218 57-1 6-575 2-66 58-0 -0-9 6-582 58-3 —1-2 56-7 +0-4 
4140 63:0 6-347 4:50 63-0 0-0 6:347 63-0 0-0 62-0 +1-0 
747-3. 67-0 6117 7-63 66-8 +02 6-117 66-6 +0-4 67-3 —0:3 

7-68 66-8 -02 6114 66-8 —0-2 67:3 —0-7 


7473 66-6 6-115 


V. 24 cc. m/15 Na,HPO,+6 cc. m/15 KH,PO,+50 ce, 0-04941n NaOH +H,O to 100 ce. + 0-85 g. 
NaCl. 19°. 


100 499 7859 0138 50:1 -O02 7-861 50-1 -02 474 +265 
22:7 530 7529 0296 537 4-07 7534 539 £4=-09 538 -08 
477 «578 «7-244 «2057089 20-11 7-252 -23B-G 
85:2 623 7-023 0945 639 -16 7036 641 -18 636 -13 
178-6 685 6-744 1:80 10-9 -24 6759 712 -27 69:0 -065 
3752 769 6473 3-37 716. OT. O4636. 97-2 3082 148. 487 
743-2 810 6198 6-34 $26 -16 6207 826  -16 795 +15 


278 EK. J. WARBURG 


In the first three solutions 0-42 9% NaCl was present and 50 cc. 0-04941n 
NaOH in 100 cc. solution. In the last two solutions 0-85 % NaCl was present 
and inthe first of this series 70 cc. 0-0243n Na,CO, in 100 cc. In the last, 
NaOH was of the same concentration as in the first three. In all the solutions 
there was also 0-02m phosphate (30 cc. of 8. P, L. Sérensen’s phosphate 
mixtures to 100 cc. solution), the alkali content of the solutions being varied 
by adding different amounts of primary and secondary phosphate. The ex- 
periments were performed as usual (with the small saturator). In the first 
column of the table the CO, tension is given; in the second column the 
combined CO, in vols. %; in the third column py: calculated in the usual way, 
the values for K’, given later being used; the fourth column contains ag:. 


cack RS ae WS ae ee ae ee Pie ben eat aga 
a fleet 
gol- >. eee. ee Sa cuans 
< ee | ++ 
> 
soa FE ee ea nent a 
Opti yp oe er ponent 
es vlaeeetrns: ape 
A ee aes pe eg a) 
30;- matt — 
a Ee gage 
20 cat 
10-—- = 
a é Py. 
es ee oma! Ears be | cecka (A Eat l MES fee 


80 79 78 77 76 75 74 73 72 71 70 69 68 67 66 65 64 63 62 61 60 
Fig. 15. The combination of CO, in 0-02m phosphate solution. 


We will let C; stand for the difference between the total cations and total 
anions the dissociation of which does not depend upon the reaction. The 


value for K’, employed is calculated by equations (139) and (143). The first 
dissociation of phosphoric acid 


HPO, = + BPO eee (158) 


progresses, under the conditions of the experiment, from left to right com- 
pletely, the dissociation constant according to Abbott and Bray [1909] being 
10-*, The third dissociation of phosphoric acid under the same conditions 
does not take place at all, that is to say the reaction 


APO", = 4 POO OP ee ee (159) 


proceeds completely from right to left, the dissociation constant according to 
Abbott and Bray being 10-!*, The only significant reaction therefore is 


HPO!) B+ HPO eee (160) 


THEORY OF THE HENDERSON-HASSELBALCH EQUATION 279 


At a given CO, tension the following equilibrium holds for (160): 


nee LORS” CH2P0's Fg (H,PO";) 
ay’ = 1°7-60B 2 CHRO”, F, (HPO” iy eeccesesecce 


(An almost analogous equation was evolved by L. J. Henderson as early 
as 1908.) 


Pq (H,PO"%,) 


If we now put hg ar F, (HPO”,) = OP EE ey yo ee ee (162) 
» Poo.a 1, ¢ P 
we get Ce Ee) el 2, pee (163) 
The total phosphate concentration is expressed by 
Con = Caypoy + Crpory: seereteeeceeeeeeeeees (164) 
The anv expressed in normalities corresponding to the phosphates then 
becomes uCp, = Cy.Po’ 4 + 2Cypo”, eee eee eee eee eee eee (165) 
From (163), (164) and (165) we get 
15-20Bk’, + K’,Poo. « 
| uU= 760 Bk, + EPoo. Se sreapsedpovansebnedeasevuey sa (166) 
2k’, + ayy" 
or | any “ere re (167) 
Further, C; is defined by 
Cy —= Cn00's + Cy.po, ++ 2CyPo”, ove ceneeeceseseses (168) 
and therefore from (168) and (165) 
= (Cy — uCp,) 2226. Cee eeerereeseerecesccece (169) 
From (169) and (166) we obtain 
K’, Poo, a + 2226 x 7-60k', (2Cpp -Cy) , K’, 2226 Poo, a (CPh-C ‘pa 
B?+ B+ 7-60, V+ 7-608’, “= 0. ...(170) 


Substituting in this the constants given in Table Ll a we get, for the first 
series B = 5-2 — 0-217 Poo, + V0-0471 P¥oo, + 21-5 Poo, + 27-0; 
for the second series 

B = 23-0 — 0:230 Poo, + V0-0529 P¥¢9, + 31°6 Poo, + 529; 
for the third series 

B = 27-5 — 0-226 Poo, + V0-0511 P%o, + 32°6 Poo, + 756; 
for the fourth series at 18° 

B = 14-5 + 0-252 Poo, + V0-0635 P%o, + 29°8 Poo, + 210; 
at 19° B= 14-5 + 0:248 Poo, + V0-0615 P¥o9, + 29:3 Poo, + 210; 
for the fifth series 

B = 23-0 + 0-250 Poo, + ¥0-0626 Poo, + 33°9 Poo, + 529. 


een ease (171) 
Table LI a. Constants used in the calculation of (171). 
Total cation 
Series concentration Cy R’, CPh ¥; Temp. 

I 0-1165 00447 10~%-286—5-18 x 10-7 002 10-*9*=1:38x 1077 © 19 
I 0-1325 0:0607  10-*-478 = 5-36 x 10-” 0:02. 10°*-**=1-38 x 10-7 =—-:20 
Il 0-1365 0-0647 10—-?68 —5-40 x 1077 0:02 10-§8=1-38x10-7. 19 
TV (18°) 0-1940 0:0530 10—*:244—5-70 x 107? 0:02: 10-$"§= 1:38 x 1077.18 
IV (19°) 00-1940 00530 10—*-239—§-77 x 1077 0-02 °10-**6=1-38x10-7 19 
V 0-2027 0-0607 106-285 5-82 x 10-7 0:02 107**=1-38x10-?7 19 


280 E. J. WARBURG 


k’, is taken as 10-*-86, which has been very accurately determined by 8. P. L. 
Sérensen [1909- 1910] in a mixture of equal parts of m/15 KH,PO, and 
m/15 Na,HPO,. It is probable that the constant in our case only differs 
slightly from this value. In solutions of 0-42 % NaCl, Y is taken as 6, in 
other solutions, as 8. The results of the calculations will be found in Table LI 
in the fifth column; in the sixth column the differences between the values 
found and calculated are given. If we put aside the experiments with the 
highest CO, tensions (which are subject to the largest experimental error), 
it will be seen the agreement is as good as could be expected on the whole, 
but it should be noted the calculated value is always a little greater than that 
found by experiment. The cause of this is not quite clear but it may be 
pointed out that the agreement would have been very nearly ideal if k’, was 
taken as 10-89 

Now that we have effected these calculations let us turn again to the 
equations (171). It willimmediately be obvious that it is difficult to form a 
clear conception of the course of the five carbonic acid combination curves 
if we only have these equations to go by, but it will be simple if from (167) 
and (169) we evolve the equation 


B= (Oy — Op, Ge) 2226. oo seccseeeccsseee (172) 


ky +a 


This equation indicates that all the CO, combination curves in solutions 
which contain the same phosphate concentration run parallel with one another 
if they are drawn in a right angled coordinate system where the apparent 
hydrogen ion activity, or a value which contains no other variable than this, 
is placed along one axis and the combined CO, along the other. If the com- 
bined CO, is expressed in vols. % the distance between the two curves pro- — 
jected on the vols. % axis will be the difference between the Cy x 2226 
(vols. °%) of the two solutions. 

The above reactions are of general interest in that, in any similar solution 


whsterer:6 Tare w= (Oy — Oe f (ace )) 22aGes cos ace (173) 


where Cy is the concentration of the electrolytes of which the dissociation 
varies with the reaction, and where C; and Cy (aq°) can have positive or 
negative values. Even though we do not know Cy, Cy, orf (ay*) we can, under 
the given assumptions, be certain of the shift in the CO, combination curve 
if Cy varies on the addition of alkali (or acid), the volume of the system being 
constant. If for instance C; is increased to Cy + 6 we get 


B; = (Cy + 6 — Oy ff (@yq")) 2226. ....... cee ee eee ees (174) 
Subtracting (173) from this we obtain 
By — B a 22360, vcs ivsecrotuisdvamvivesies (175) 


which is the algebraic expression of the fact that any CO, combination curve 
will be shifted parallel with itself in an ay*—vols. % CO, graph provided 
adsorption combinations of CO,, or CO, compounds in the form of complex 


THEORY OF THE HENDERSON-HASSELBALCH EQUATION — 281 


bodies (e.g. carbamino-acids), are not formed. The distance between the 
curves is then 2226) vols. %. 

If the CO, in a solution can be adsorbed or combined in the complex manner 
indicated above and if these combinations are reversible, the amount combined 
will be a function of the CO, pressure. The general expression for the com- 
bination will be 

B = (C, — Cy f (@u") — fi (Poo,)) 2226, «.. eee (176) 


and the difference in CO, combined in a solution before and after the addition 
of alkali at the same reaction will therefore, from analogy with (175), be 


B, — B = 2226b — 2226 (fy (Poo,) — fr (Pco,,))- +++ (177) 


In this equation the amount 


— 2226 (f, (Pco,) — fi: (Pco,,)) 
will always be positive because the combination of CO, must increase with 
the CO, tension in the given conditions and in order to produce the same 
reaction a higher CO, tension will be necessary after the addition of alkali 
than before (cf. equation (95)). 

It is thus possible to state in general that a CO, combination curve, in 
which the combined CO, is set off along the one axis and an expression for 
the active reaction along the other axis, will be shifted on the addition of 
alkali in such a way: 

(1) that the distance between the original curve and the new curve is 
always equal to the combination of CO, corresponding to the added amount 
of alkali provided that the variable combination in each curve is only a 
function of the reaction; 

(2) that the distance between the original and new curves will increase 
with the apparent hydrogen ion activity provided that the variable com- 
bination is also a function of the CO, tension; 

(3) that the distance between the two curves can never be less than the 
combined CO, corresponding to the added alkali. 


We will now revert to the phosphate solutions again. In the eighth column 
of the table the combined CO, expressed in vols. % will be found, calculated 
by (172) with the help of the apparent hydrogen ion activity in column 4 
reckoned directly from the experiments. The agreement is, as expected, good. 
In Fig. 15 the combined CO, is plotted as ordinates and 


— log ay" = pr 
as abscissae, however on a hundred times greater scale. In such a coordinate 
system, as explained, the course of the curves will be parallel. Now it appears 
from the figure that the CO, combination curves are straight lines for a long 
portion of their courses, and we will therefore, as an example of what is to 
follow, calculate these sections of the curves by the method of least squares, 
using the figures in columns 2 and 3, and include all experiments in which 
Bioch. xv1 19 


282 E. J. WARBURG 


Pu is less than 7-59. Equations (141) and (142) are employed in the calcu- 
lation and the mean error of the constants is estimated by the equation 


M 
My = My = Fase ee (178) 


where M = af. = a SUE IS Maa boaus ob DAP O ee eee ee (179) 
Xa being the sum of the squares of the deviations of the found and calculated 
values and » the number of experiments. The equation of the straight line 
referred to above is © log ty by = Be OA aialciov taba (180) 
where z is the tangent to the angle made by the line with the axis of the 


abscissa, measured in the first or third quadrant, and y is the part of the 
ordinate axis cut off by the line. The calculation gives 


No. of 
Series x y M experiments 
21-460 +.0-057 178-65 +.0-057 0-51 7 
i 22-139 +-0-17 219-15 +0-17 0-91 5 
Il 21-720 40-16 225-24 +.0-16 0-87 5 
IV 23-076 +-0-077 208-41 +-0-077 0-68 7 
Vv 19-301 4.0-27 199-13 +-0-27 1-9 6 


It will be seen that a determination of the error indicates that x is different 
n the various series, but as it has been proved already by calculations that 
the different curves are really parallel, allowing for experimental error, we 
cannot attribute any real significance to the mean error calculated by this 
method in the present and similar cases. There are several reasons for this 
rather unsatisfactory result. In the first place the number of experiments in 
each series, within the range of reaction employed, is far too few to permit of 
an accurate determination of the error; in the second place the points deter- 
mined are not regularly distributed over the various sections of the curve, 
in fact they do not all cover exactly the same range of reaction, which is of 
importance as it will be shown the portion of the curve in question is not in 
reality a straight line. For the appreciation of what is to follow it will be 
necessary to examine this a little more closely. 

By equation (172) it is easy to estimate the true slope of the curve. In 
the second column of the following table the combinations of CO, in a 0-02n 
phosphate solution of py varying from 7-50 to 6-20 are given, 2226 Cy = 52:3 
vols, %, the combined CO, at py: 7-50 being here put equal to 0. In the third 
column will be found the tangent to the straight lines (angle of inclination 
which is x) which join the nearest points together two by two. If the point 
referring to py’ 7:50 is joined with that referring to py 6-20 the tangent to 
the angle of inclination of the line is 21-6. In the fourth column will be found 
the amount of combined CO, expressed in vols. %, which, according to a 
calculation based on the straight line smoothing formulae (141) and (142) from 
the figures in columns 1 and 3, should be found combined at the given py: 
when the best constants are used. These constants are 

a = 22-719 + 0-021, y= 169-77 4 0-021. 
In the fifth column the differences between the CO, combinations calculated 
by (172) and by the smoothing are given. The standard deviation is calculated 
to 045 vols. %. 


THEORY OF THE HENDERSON-HASSELBALCH EQUATION 283 


Table LI db. 
Vols. % CO, Vols. % CO, 
combined. combined. 
Calculated Algebraic Difference 
Pr by (172) x smoothing vols. % 
1 2 3 4 5 

(2-4) 

7:50 0-0 -0-6 +0°6 
17 

7:40 1-7 1-7 0-0 
19 

7:30 3-6 3-9 —0°3 
21 

7:20. 57 6-2 -0°5 
23 

7:10 8-0 8-5 -—0°5 
24 

7-00 10-4 10-7 —0°3 
25 

6-90 12-9 13-0 —O1 
26 

6-80 155 15:3 +02 
25 

6-70 18-0 17-6 +0:4 
24 

6-60 20-4 19-8 +06 
22 

6-50 22-6 22-1 +0°5 
21 

6:40 24-7 24-4 +0°3 
‘ 18 

6-30 26-5 26-6 —O1 
16 

6-20 28-1 28-9 —0-8 


If column 5 is examined it will be observed that the differences change 
their sign three times, that is to say the straight line cuts the true curve 
(calculated by (172)) three times. If it should happen, in the determination 
of a similar CO, combination curve, that points were obtained at these places 
of intersection, it would naturally follow that even an ideal method of analysis 
would be unable to show that the curve departed from a straight line. It will be 
realised that a very considerable number of experimental points are required to 
determine the true course of the curve and that in using the smoothing method 
of calculation one must be exceedingly cautious about one’s conclusions. 

As mentioned the constant x was determined by the smoothing method 
to be 22-719 + 0-021. If this is compared with the constants found by calcula- 
tion from the experiments it will be seen that only that from series II appears 
to be of the same magnitude, which is another plea for the contention that 
the mean error of the constants is really too sharply defined. 

There is still another point in relation to the smoothing method to be 
touched upon. It is not possible by this method in the form which is used 
here with the aid of the mean error of the constants to determine the mean 
error from a single point on the curve. This is due to the choice of the zero 
point of the coordinate system, because this lies, as will be seen, far outside 
the experimentally determined points. The reason I have given the numbers 
for the constants correct to the first decimal place in spite of the objections 


19—2 


284 EK. J. WARBURG 


put forward relating to the mean error, is in order to show them with all the 
figures used in the calculations and the mean error is therefore only employed 
roughly to demonstrate the uncertainty of the constants. 

It is accepted that a large part of the reversibly combined CO, of the blood 
is present as the bicarbonate ion, the haemoglobin and other protein sub- 
stances acting as ampholytes. The objections against this view which have 
been put forward in the last few years have been to some extent negatived 
in previous chapters and will also be questioned at the end of the present one. 
It can therefore be concluded that no general condition for the combination 
of CO, in the blood can be formulated which cannot be traced back to an 
equation of the form of (173) or (176). Provided no CO, is adsorbed or com- 
bined in a complex manner (173) will be the correct equation. 

In 1920 T. R. Parsons put forward a mathematical treatment of the CO, 
combination in the blood but it seems to me he has made a mistake on this 
point. Firstly Parsons assumed that haemoglobin is a monobasic acid, each 
haemoglobin molecule however encompassing several such groups. Secondly 
he assumed that in blood especially there were so many more cations than 
anions the dissociation of which did not vary with the reaction, that they 
were able just to neutralise the haemoglobin acid. Thirdly he believed that in 
the “maximum” of combined CO, we have a measure of the concentration 
of the haemoglobin acid. 

The first of these assumptions is quite uncertain; the second is undoubtedly 
erroneous, because the blood can expel CO, from alkaline carbonate as Pfliiger 
[1864] was the first to show. The third assumption is incorrect as there is no 
maximal CO, combination in blood at atmospheric CO, tensions. Parsons’ 
final relation evolved as it is from faulty assumptions has no real significance 
and it is only one of many instances that nearly every equation with a suffi- 
cient number of constants can be brought into agreement with an observed 
fact when the constants chosen are the best possible. 

At present there seems to be no possibility—contrary to what was the case 
with the phosphate experiments—of establishing an a priori relation between 
ay and the combination of CO, with the blood proteins. There is no other course 
open therefore than to find the simplest empirical relation between these values. 

In the previously mentioned paper of Hasselbalch and Warburg some CO, 
combination curves for blood and serum are given in a py'—vols. % curve. 
They are straight lines. This fact appears to have been quite overlooked, but I 
believe the appreciation of this phenomenon will be of far-reaching importance 
in the physiology of respiration. In this work I shall content myself with more 
thoroughly demonstrating the relation and employ it for elucidating the question 
of the existence of complex CO, combinations as well as some other problems. 

In tables LII—LVII (pp. 287-290) the combination of CO, at various CO, 
tensions in a number of blood samples from different people and animals is 
given. In the last column but one the results of the algebraic smoothing are 
recorded, the CO, combination curve in the py:—vols. % combined CO, chart 
being taken as a straight line. 


THEORY OF THE HENDERSON-HASSELBALCH EQUATION 285 


It is therefore assumed that 
@ log Gy + Y= Be oececececesseccseeseee ..+..(181) 
or ee Me MSE RRs sci nls cnc cascartaundeusccss (182) 


~ That these relations, within the experimental error, are really true seems 
to follow from the curves and tables, but I have further investigated some 
rather extensive unpublished material dealing with the CO, combination 
curves in ox blood which has been elaborated of late years for another purpose 
in the laboratory of the Finsen Institute, without being able to find any sign 
of a departure from the relations. After what was said about the phosphate 
curves it will readily be understood I do not wish to urge that the sections 
of the curves being dealt with are actually straight lines, but it is asserted 
the deviations from a straight line are small compared with the variations 
in the combination of CO, at different apparent hydrogen ion activities. 

In the calculation of py;,) (130) was used. In the case of oxygenated 
human blood at 38° and oxygenated horse blood at room temperature the 
corrections given in Fig. 10 were added to the values found. In the case of 
reduced human blood at 38° and reduced horse blood at room temperature 
the correction corresponding to py: (uncorrected — 0-40) was added. In the 
case of Hasselbalch’s [1916] “half reduced” blood 0-20 was subtracted from 
Pu (uncorrected), while 0-40 was subtracted from oxygenated ox blood at 
room temperature. Equation (130) was used without correction in the calcu- 
lation of the other blood samples. 

The correction curve corresponding to 18-5 vols. °%% combined CO, was used 
‘in the calculations relating to Haldane’s blood, from Christiansen, Douglas 
and Haldane [1914], the 15 vols. % curve for Parsons’ [1917] blood,. the 
20 vols. % curve for Davies’ [1920] and Hasselbalch’s [1916, 2] blood and the 
25 vols. % curve for Joffe’s [1920] and Warburg’s blood!. In the calculations 


1 T have not attempted to discover how much combined O, was actually present in Davies’, 
Haldane’s, Joffe’s and Parsons’ blood but I have contented myself with estimating their py-—CO, 
combination curves by means of the corrections mentioned because an error of 5 vols. % com- 
bined O, only causes a difference in the constant x of about 1. The reason of this is that I have 
not been able to obtain agreement between the O, determinations got by pumping out and the 
values obtained by Haldane’s ferricyanide method. I originally became aware of this by 
comparing the relation between the O, combined by the blood and the volume of the blood 
corpuscles of the same blood. A. Norgaard and H. C. Gram and later H. C. Gram [1921] found 
about 38-5 vols. % combined O, in 100 cc. human blood corpuscles but V. Bie and P. Mdller 
[1913] found about 45 vols. %, a similar quantity to that I have myself found. Later on, it 
appears, W. B. Cannon, J. Fraser and A. N. Hooper [1919], in oxygenated blood from patients 
suffering from shock, found about 43 vols. %. While Norgaard and Gram’s haemoglobinometer 
was corrected by the ferricyanide method, Bie and Mdller’s apparatus was corrected by Fridericia 
by pumping out, just as my own results were obtained. Dr Marie Krogh was kind enough to 
determine the O, in some blood samples from man and horse with a Haldane haemoglobinometer 
which was very accurately calibrated by the ferricyanide method. She constantly found lower 
O, values than I did with the exhaust pump. I have no doubt whatever that too low readings 
are obtained under special circumstances with Haldane’s ferricyanide method as shown lately by 
van Slyke and Stadie [1921] and I hope soon to take up this question again. F. Miiller [1904] 
previously found the same for dog blood but considered it only applied to quite fresh blood and 
under unfavourable circumstances. 


286 E. J. WARBURG 


relating to ox and horse blood the curve which most nearly corresponded to 
the O, determinations was employed. 

The value used in the calculations for pX.) + log ®, (CO,) with human, 
ox, and horse blood at 38° was 6-190; at 37°, 6-185; at 17°, 6-334; at 18°, 6-327; 
at 19°, 6-320. In the case of pig and pigeon blood at 38°, 6-20 was used; with 
dog and rabbit blood at 19° and 18-5°, 6-35. With haemolysed blood equation 
(134) was employed with the following constants: at 17°, 6-325 was employed; 
at 18°, 6-320; at 19°, 6-315. With serum equation (121) was used and as 
constants at 38°, 6-165 was used; at 19°, 6-286; at 20°, 6-281; at 21°, 6-274. 
The results of these calculations and of the smoothing on the curves for some 
of the experiments carried out in the laboratory of the Finsen Institute will 
be found in Tables LIT, LITI and LVIII and in Figs. 16, 17, 19, 20and21. The 
combination of CO, in human, pigeon, horse, dog, rabbit, pig and ox blood and 
in some sera can therefore be expressed by an equation of the form of (181) 
or (182). 

It will be observed, as already mentioned, that all the curves determined 
in this laboratory are straight lines, allowing for experimental error. In the 
case of a number of the serum curves however the point corresponding to the 
highest CO, tension is too low. This seems to me only to be an indication that 
equilibrium had not been quite reached in these experiments as at the time 
I undertook them I was not aware of the deficiency of the small saturator. 
I have omitted these experiments in the calculus of smoothing, and in the 
tables this is denoted by a line drawn through them. It will be easily seen 
without a special note which experiments are included and which rejected. 
Those portions of the curves which correspond to the rejected experiments 
are drawn with a dotted line. In one of the tables an experiment has been 
excluded because it seems to be subject to an unusually large error. This 
experiment is put in parentheses. 

The experiments with very low tensions were carried out with the large 
saturator electrodes using the technique described in chapter IV. 

In Table LIV I have performed similar calculations for Haldane’s blood 
(taken from Christiansen, Douglas and Haldane’s [1914] paper) and in Table 
LVI for Davies’ blood (taken from Davies, Haldane and Kennaway’s [1920] 
paper). The curves are given in Fig. 18. 

It will be noticed the CO, combinations corresponding to the most alkaline 
values are too high both in oxygenated and reduced blood—in Davies’ blood. 
the most alkaline values are too high and the most acid too low. I believe 
I am right in saying these deviations from a straight line are due to the fact 
that the technique of saturation employed is not applicable at high and low 
CO, tensions. The experiments just recorded show this and I myself had a 
similar failure before I became aware of the relatively great difficulties of 
saturation. Davies, Haldane and Kennaway’s mistake in thinking the com- 
bination of CO, in the blood is maximal at the highest tensions is only due to 
equilibrium not being established, Calculations relating to J. Joffe’s blood, 
Table LVI. and Fig. 19(from Joffe and Poulton’s[ 1920] paper) also demonstrate 


THEORY OF THE HENDERSON-HASSELBALCH EQUATION 287 


Table LII. 


Pig blood (O,-combination 21-1 vols. %). Half saturated with O,, 38°. Calculated 
after K. A. Hasselbalch. pd) +log ,=6-20 in every case. 


So Vols. ve we 

mm. Hg Vols. % CO . combin 

Co, combined . caleaiahed calculated Difference 

8-4 21-9 7-75 20:8 +11 

(11-4 22-7 7-64 26-5 — 3-8) 
15-7 28-6 7:60 28-6 0-0 
30-4 36:8 7-42 37-9 -11 
30-8 36-4 7-41 38-4 —2-0 
53-1 46-3 7:28 45-2 +11 
76-1 51:8 7:17 51-0 +08 


*=51-865+0-5. y=422-764+05. M=1-4, 


Pigeon blood (O,-combination 23-7 vols. %). Half saturated with O,, 38°. 
alculated after K. A. Hasselbalch. pd, + log &,=6-20 in every case. 


5:8 21-0 7-90 20-9 +01 
18-3 _ 35-0 7-62 35-1 —O-1 
38-5 45-3 7-41 45-7 —0-4 
75-0 56-2 7-21 55-9 +03 


x =50-774+0-08. y=421-96+0-08. M=0-4. 
Dog blood (O,-combination 19-2 vols. %) totally reduced, 19°. pd,,) + log &, = 6-35 in every case. 


2-2 21:8 8-27 22-1 — 0:3 
29-3 53-5 7-54 53-6 —O-1 
56-4 62-4 7-32 63-1 —0-7 
717-1 68-5 7-22 67-5 +10 

127-5 76-0 7:05 74:8 +1-2 
337-6 88-8 6-70 89-9 —-1-] 


x =43-230+0-1. y=379-58 +01. M=1-0. 
Rabbit blood totally reduced, 18-5°. pd;.) + log &, = 6-35 in every case. 


16-0 48-6 7:76 44-4 +42 
24-2 53-1 7-62 52-9 +0-2 
41-5 63-0 7:46 62-6 +0-4 
44-2 64-1 7:44 63-9 +0-2 
81-7 73:8 7:23 76-6 —2-8 
83-1 76-5 7-24 76-0 +0-5 
368-0 107-0 6-74 106-4 +0-6 


x=60-838+0-4. y=516-49+0-4. M=1-5, 
Reduced horse blood (vols. % combined O,, 22-6), 19°. 


3-0 25-2 8-24 26-1 —0-9 
22-0 56-2 7-68 55:7 +05 
39-6 66-3 7-48 66-3 0-0 
48-7 69-9 7-41 70-0 -—0-1 
80-2 78-2 7:24 79-0 —0:8 
91-5 83-5 7-20 81-1 +2-4 

484-7 112-2 6-59 113-4 -—1-2 
x=52-908+.0-2. y=462-04+0-2. M=1-3. 
Oxygenated horse blood (vols. % O,, 25:4; 25, 1; 25, 2), 20°. 
32:7 51-5 7-50 51-4 +01 
75-2 69-4 7:24 69-4 0-0 
194-5 91-1 6-92 91-4 -—0°3 
364-0 106-1 6-71 105-9 +0-2 
558-4 116-2 6-56 116-2 0-0 
«=68-897 +0-06. y=568-17+0-06. M=0-2. 
Oxygenated ox blood, 38°. Hasselbalch and Warburg. 
PAs) + log &, =6-20 in every case. 
43-5 41-2 7-31 40-7 +0°5 
93-6 54:3 7-10 54-9 - 0-6 
150-9 63-9 6-96 64-5 —0°6 
203-9 70-1 6-87 70-4 —0:3 
269-4 76-5 6-79 75:8 +0°7 
361-8 83-4 6-70 81-9 $25 
692-9 101-9" 6-50 95:3 + 6-6 


1 O, deficit. 
x =67-458+0-03. y=533-81+0-03. M=0-7. 


288 E. J. WARBURG 


Table LIII. 


K. A. Hasselbalch’s half reduced blood, 38°. K. A. Hasselbalch [1916, 2]?. 
Vols. % CO, 


mm. Hg Vols. % CO, PH is) combined 
CO, combined calculated calculated Difference 

18-5 35-1 7-63 36-6 -1-5 
22-4 36-7 7-56 40-6 -3-9 
32-7 45-1 7-48 45-2 —0-1 
40-9 48-2 7-41 49-3 -1-1 
45-7 52-9 7-40 49-9 +3-0 
50-8 57-1 7-39 50-4 +6°7 
67-0 57-9 7-27 57-4 +0-5 
80-7 61:3 7-21 60-9 +0-4 
102-2 62-5 711 66-6 -4-1 

x =57-858-+40-9. y=477-:98+40-9. M=3-6, 

E. J. Warburg’s reduced blood, 38°. 

13-6 27:7 7-66 25-9 18 
39-5 42:8 7:36 45-9 -3-1 
74-6 56-5 7-20 56-5 0-0 
109-2 64-7 7-09 63-8 +0-9 
544-5 99-5 6-56 99-1 +0-4 


a =66-567 +.0-5. y =535-79+0-5, M=2-1. 


1 The experiments were performed with blood taken on three different days and were not 
specially designed for the determination of the CO, combination curve. 


Table LIV. 


J. 8. Haldane’s oxygenated blood, 37°. 
Vols. % CO, 
mm. Hg Vols. % CO, PH'(s) combined 
CO, combined _ calculated _ calculated Difference 
2-6 14-9 8-15 3-8 +11-1 
8-7 25-7 7-84 23-0 + 2-7 
18-9 35°8 7:64 35:5 - 03. 
37-7 48-6 7-46 46-6 + 2-0 
44-1 49-4 7-40 50-4 - 10 
41-7 48-2 7-42 49-1 - 09 
56-1 54:5 7:33 54:7 - 0-2 
73-3 58-1 7°25 59-7 - 16 
78-5 60-3 7°23 60-9 - 06 
105-2 68-1 7:15 65:8 + 23 
«=62-01640°5. y=509-32+.0-5. M=1-6. 
J. 8. Haldane’s reduced blood, 37°. 
2-15 153 8-24? 4:2 +111 
7°15 30-3 7:94 23-6 + 6-7 
8-1 30-0 7-92 24-9 + 51 
17-9 40-4 7:69 39-8 + 06 
37-9 53-5 7:48 53-4 + O1 
58-2 60-5 7:34 62-5 - 2-0 
74-2 65:3 7:27 67-0 - 17 
78:7 69-9 71:27 67-0 + 29 
Lill 754 715 74:8 + 06 


2 =64-73140-7, y=537-60+40-7, M=18, 


THEORY OF THE HENDERSON-HASSELBALCH EQUATION 289 


Table LY. 


T. R. Parsons’ oxygenated blood, 37°. 
Vols, % CO, 


mm. Hg Vols. % CO. PH combined 

CO, combined » calculated calculated Difference 
5-7 21:3 7-93 21-4 —O-1 

10-1 30-0 7-83 26-2 +3°8 

19-6 34-4 7-59 37-6 —3-2 

33-4 41-5 7-43 45-2 -3-7 

37-4 43-4 7-40 46-6 —3-2 

72-1 59-9 7-25 53-7 +6-2 


*%=47-44942-4. y=397-7042-4. M=5-4. 


T. R. Parsons’ reduced blood, 37°, ° 


5:7 23-8 7-96 23-0 +0°8 

8-1 27:4 7:87 27-7 -0:3 
10-1 31-5 7:83 29-6 +19 
19-6 39-9 7-64 39-8 +01 
33-4 47-8 7-48 48-1 -0:3 
37:3 47-3 7-42 51:3 -40 
55:3 55-1 7-31 57-0 -1-9 
72:1 64-4 7:26 59-6 +48 


x=52-319+0-7. y=439-48+0-7. M=3-0. 


Table LVI. 
H. W. Davies’ blood, 38°. 
Vols. % CO, 
mm. Hg Vols, %, CO, (2) combined 
Co, combined culated calculated Difference 
8:3 26-3 7-88 6-0 +20-3 
9-02 29-6 7-90 4:3 +253 
11-2 28-7 7-78 14-5 +14-2 
19-3 35-5 7-63 27-2 + 83 
29-8 41-4 7-50 38-2 + 3-2 
— «38-9 46-4 7:44 43:3 + 31 
39-0 45-5 7-42 44-9 + 06 
39-1 43-1 7-39 47-5 - 44 
46-6 47-3 7:35 50-9 - 36 
51:3 53-2 7:36 50-0 + 3-2 
64:9 57-4 7:29 55-9 + 15 
68-5 60-0 7:28 56-8 + 32 
76:7 62-0 7-25 59-3 + 2-7 
89-9 65:7 7:20 63-5 + 2-2 
102-9 66-5 7-15 67-8 - 13 
202-0 81-8 6-93 86:3 - 45 
215-0 82-6 6-91 88-1 ‘— 55 
411-0 ‘97-3 6-69 106-7 — 9-4 
455-0 97-6 6-65 110-1 ~125 
497-0 95:7 6-60 114-4 -18-7 


2 =84-657-40-4, y=673-0940-4. M=4:8, 


290. E. J. WARBURG 


Table LVII. 


J. Joffe’s defibrinated oxygenated blood, 4. viii—28. xi. 1919. 
Joffe and Poulton, Tables V and VII. 


Vols. % CO, 
mm. Hg Vols. % CO, PH) combined 
co, combined calculated calculated Difference 

4-0 13-7 7:93 5-4 + 83 

5-0 16-8 7:92 6-1 +10-7 
12-9 23-8 7-65 24-0 — 0-2 
19-7 29-4 7-54 31:3 - 19 
24-9 36-0 7-53 31-9 + 4-1 
34-6 40-0 7-42 39-3 + 0-7 
44-4 44:8 7:36 43-2 + 1:6 
50-2 44-] _ 7:30 47-2 - 31 
58-0 47-6 7-26 49-9 — 2:3 
65:3 52-6 7-25 50-5 + 2-1 
67-8 53-9 7-25 50-5 + 3-4 
82-2 56-0 7:17 55:8 + 0-2 

110-3 63-3 7:10 60-5 + 2:8 
149-0 62-6 6-96 69-8 — 7-2 
156-0 67-1 6:97 69-1 — 20 
157-0 70:3 6-98 68:4 + 19 
180-0 68-5 6-91 73-1 — 46 
376°3 90-0 6:69 87-7 + 23 
§10-0 100-3 6-53 98-3 + 2-0 
«=66-345+0-16. y=531-52+0-16. M=3-2. 
J. Joffe’s defibrinated reduced blood, 4. viii.-9. x. 1920. 
Joffe and Poulton, Table VII. 

9-9 27:9 7-81 20-1 +7:8 
15-0 34-4 7-72 26:3 +81 
19-4 35:3 7-61 34-0 +1:3 
27-6 40-1 : 7:50 41-6 -1:5 
30-9 44-1 7-50 41-6 +2°5 
36-8 47-9 7:45 45:1 ' +28 
41-7 46-0 7:37 50-6 —4-6 
47-0 50-4 7:36 51:3 -0:9 
50-5 49-8 7°33 53-4 -—3-6 
55-7 52:9 7-31 54:8 -1:9 
70-1 55-6 7-23 60-4 —4:8 
76-8 62-1 7-23 60-4 +1:7 
77:7 64-0 7-24 59-7 +43 
92-7 67-6 7:19 63:1 +45 


2 =69-475+40-7. y=562-65+0-7. M=3-6. 


that the most alkaline values are too high but the acid values are correct. 
Parsons’ [1917] blood gives values scattered irregularly about a straight line | 
(Table LV and Fig. 19). 

The constant x, which is the tangent to the angle the straight line makes 
with the axis of the abscissa measured in the first or third quadrant, is for 


Davies’ oxygenated blood ww. — 84°7 
Haldane’s 4 NERY SY «» 62:0 
i reduced Aes Pe oe 6497 
Hasselbalch’s half oxygenated blood 57-9 
Joffe’s oxygenated blood - 663 
3s reduced me w» 69-5 
Parsons’ oxygenated ,, ve AT*4 
reduced m » 523 


Warburg’s zs sv p00. OOD 


THEORY OF THE HENDERSON-HASSELBALCH EQUATION 291 


fe SE ie i ot Seg i i a Oe Pe Pee ete ees 
ok —4110 
x + 
zr —100 
ob Seen]. : ee 
0 }— ates eo blood 20° # 
70}— —— (concentrated) Bs 
60 }-— ee — 60 
50}— -—{ 50 
Dog blood 19° A é7] 
30;-—- : —| 30 
20} + 20 
10}-— Py ‘a —j 10 
Rid fue Cee 1S | | Se) ROA easy ali | al l 
83 82 81 80 79 78 77 76 75 74 73 72 71 «70 69 68 74°73 72 %71 70 69 GB 67 66 6 ‘5 
60;— — 60 
50-—- et —| 50 
40|- wf 
Pig blood 38° Pe blood 38° ~| 
a “nt (K. A. H.) pe ale (K.A.H.) 4% 
20 |— 4 7] ot —-| 20 
10}— Pu... —{ 10 
l | bas, TS TS ae | ee (ee ee ee eS ree a ae 


82 81 80 79 78 77 76 175 74 73 72 TL 70 69 68 67 


Fig. 16. 


79 78 «77 «176 «+15 «+74 


i, 4 


i 


eee 


Horse blood 19° 


+ (concentrated ) 


ll Sec died os ee Pe st 


dT a 


Py, 


Re AE tia CNY Reed RE DS 


aad 


he 


81 80 79 78 77 76 75 74 73 72 71 70 69 68 67 66 65 66 


+ 


Rabbit blood 18°5° 


+7 (K.A.H. and E. J. W.) 
7“ 


110 


— 4100 
A 


et 


| 


a 


+ 
Ox blood 38° 


is) 


Py 
EET a CARRE IA Sa 


! 
78 77 76 75 74 73 72 71 70 69 68 


Fig. 17. 


73°72 71 70 69 68 67 66 65 


73° 72 «T1 70 


292 


E. J. WARBURG 


be Ps FOR MRR SPS Scala Re a Ey OS i ES 
70}—- aa 
60 = — 
se: i J. S. Haldane’s blood 
4 37-5° es 4 
40-- = x , 4 
S ee 
30}— Br = a pa: 
a 
20+— 4 fo poe 
+ ee pe. 
eal 1G 7 ries. Px.) 2 
ee ace RB DORs SS, Gabe CoS VE RES ad | 
82 81 80 79 78 77 76 75 74 73 72 71 70 69 r- 
110}— — 
100 }— a 
90}— 4 
so|— ad 
oY Be H. W. Davies’ blood 37°5° + 
ol Ss a 
of 2 z 
40|— +7 * J 
‘ { 
30--—- +, + J ~ 
20/— 7 ml 
A 
st rie : Pity Ty 
Sa Pac crs a os FES, Waal ERY ASH ae Nl 


83 82 81 80 79 78 77 76 75 74 73 .7:2 


Fig. 18. 


71 70 69 68 67 66 65 64 63 


mae As pny as aoe Mila Sie Ws Be Bay is - 
et 70 
Para i {60 
Ae ss 
°o 
pe o 
ae. SB had 
K. A. Hasselbaleh’s blood 38° > _] 49 
| 20 
Pir — 10 
a Dell eae RTE Bae) Bae is Bee ! | esi Ae AE Sa | Se | 
79 78 77 76 «75 74 «73 «72 «71 «70 69 6B 67 66 76 75 74 73°72 «714, 70 69 6S 67 
110)— —{ 110 
10u}— 
Wh 
i) 
i ae 
w}- 60 
wr * —{ 50 
wl 4 +E. J. Warburg's blood 38° _] go 
x s+ : 
i) oe x 4 —{ 30 
75 20 
Pr D 
Tt in p 10 
- Pn 
QBS eid tee, Fey Us ce ay A ee | A ee ce Poe Pee i 
60 79 7% 77 714 «16 74 734 «72«71«~«:70 60 68 67 OW 76 76 74 73 72 71 70 60 OB 67 


Pig. 19. 


THEORY OF THE HENDERSON-HASSELBALCH EQUATION 293 


It will be observed in the first place that the constant varies from blood 
to blood. The greater part of the difference is undoubtedly due to variations 
in the amount of haemoglobin as Hasselbalch [1916, 2] has pointed out and 
as is evident from the experiments of Schmidt [1867], Zuntz [1867, 1868], 
~ Setschenow [1879], Jaquet [1892], Bohr [1905] and many others, but the 
examination of numbers of curves has shown that this cannot be the sole 
determining factor, as Peters, Jr. and Barr [1921] have also indicated from 
a rather different point of view (Peters and Barr’s curves are constructed as 
Poo,—vols. % total CO, curves). It will also be seen that z is a little greater 
in reduced blood than in the corresponding oxygenated blood. The difference is 
certainly a real one and is of almost the same magnitude in Haldane’s, Joffe’s 
and Parsons’ blood. L. J. Henderson [1920, 1] has recently demonstrated the 
same thing graphically and he has discussed the cause in a paper dealing with 
the influence of oxygen on the combined CO, of the blood. This paper seems 
to be of very considerable interest although I do not agree with the author 
in all his views but as I have not yet succeeded in quite clearing up several 
of the problems I shall only refer those interested to the original article. 

By plotting a py-—vols. % combined CO, curve it is possible to form a 
good idea of the reliability of the curve. I have done this for a number of 
Haggard and Henderson’s experiments and for some of Peters, Barr and 
Rule’s also. It appears especially in the case of Haggard and Henderson’s, ~ 
that the technique is not so good as the authors believe and it seems to me 
problematical whether the accuracy is sufficient to draw any conclusion about 
the tension with which the blood is in equilibrium from the total amount of 
CO, in the blood, as these authors do. Haggard and Henderson’s [1920, 1] 
discovery that there is no difference between the combined CO, in oxygenated 
and reduced oxalate blood, and that the combined CO, of the blood decreases 
irreversibly by blowing air through it vigorously, is presumably due to ex- 
perimental error. 

I must now discuss some experiments published in the last few years by 
H. Straub and Klothilde Meier [1918-1920]. They have unfortunately gained 
some recognition, e.g. by L. Michaelis! and by Parsons. But their results are 
so surprising, that Joffe and Poulton [1920], Peters, Barr and Rule [1921] 
for example have announced that they are sceptical. Straub and Meier’s 
chief claim is that they have shown the combined CO, in the blood does not 
increase before a reaction of about py 6-70 is reached when suddenly (in an 
interval of py: 0-01) an amount of CO, as great as the oxygen capacity is 
combined by the blood and then all further combination ceases. In haemo- 

* Michaelis and Airila have recently [1921] shown by cataphoresis experiments that the 
change of haemoglobin varies uniformly with the reaction over a very large range of reaction 
and is thus greatly at variance with Straub and Meier’s assertions about the ionisation of 
haemoglobin. * The conclusion may be drawn from these interesting experiments that the com- 
bination of CO, in haemoglobin solutions can be represented by a curve which, if not a straight 


line, is a very close approximation thereto, in a py—vols. % diagram over a very considerable 
range of reaction. 


294 ---B J. WARBURG 


lysed blood the sudden change is at py- 7-00 and is of the same extent. They 
believe serum is quite unable to combine with CO, in quantities varying with 
the pressure. Straub and Meier claim to have shown that various substances 
shift this sudden change backwards and forwards, and they have formed a 
theory which for extravagance is only surpassed by the obvious imperfection 
of their technique. With regard to these authors I cannot suppress my 
surprise that they really seem to be quite ignorant of the older, and as far 
as the present question is concerned, of the more recent literature on the 
combined CO, of the blood and it is to me completely incomprehensible that 
they have not paid more attention to their technique. 

As an illustration of Straub and Meier’s technique I will give one of their 
experiments [1920, 2, p. 250]. 

Experiment 9. 38°. 


mm. Hg Vols. % CO, 

. total 

16-8 15-1 
50-2 19-6 
72-6 30-6 
153-3 38-7 
41-4 35-9 
64-0 34-9 


The only possible explanation is, as the authors themselves have realised, 
_ that equilibrium has not been reached again at the low CO, tension. Instead 
of examining their technique they employ the experiment in propounding 
one of their extraordinary theories, and they fail to see that py: cannot be: 
calculated by Henderson and Hasselbalch’s equation if equilibrium in the 
system has not been attained. The following similar experiment shows that 
equilibrium can easily be attained. 
Ox blood slightly concentrated by centrifuging. Combined O,, 22-8 vols. %. 

70 cc. blood in the large saturator. 


mm. Hg 

21°. co, Vols. % CO, total 
First 29-0 47-9 after 30 mins. 

Then 740 186°3 = 53°°.80 2a; 

” ” 184°3 ” 65 rr) 

Afterwards 29-0 49-8 ,, 30 ',, 

” ” 47:8 ” 60 9 

” ” 47-7 ” 120 ” 


In Table LVIII and in Figs. 20 and 21 experiments and algebraic smoothing 
will be found referring to different sera. These curves also appear to be straight _ 
lines although a number of points at high CO, tensions fall a little below the 
curve. It should however be noted that deviations from a straight line course 
are more difficult to detect than in the case of the blood curves on account 
of the relatively small value of zx, 

That serum and plasma really form a dissociable combination with CO, 
is raised beyond all doubt and has already been demonstrated by Setschenow 
[1879] and by Jaquet [1892], and later by Hasselbalch [1916, 2] and Hassel- 
balch and Warburg [1918] among others. It is also quite certain that varying 


THEORY OF THE HENDERSON-HASSELBALCH EQUATION 295 


carbonic acid is combined as bicarbonate, as is shown by Loewy and Zuntz 
[1894, 1] and by Giirber’s [1895, 2] old diffusion experiments. Giirber even 
claimed to have proved that CO, quantitatively was present as bicarbonate, 
a conclusion which is strongly supported by the determinations of pA.) in 
chapter VI of this work. 


Table LVIII. 
Ox serum 38°. Hasselbalech and Warburg. I, 
Vols. % CO, 
mm. Hg Vols. % CO, PH (s) combined 
CO, combined calculated calculated Difference 
41:3 54-4 7:42 54-1 +03 
74:5 58-5 7-20 58-6 -—0-1 
140-5 62-2 6-95 63-7 —1:5 
203-0 67-5 * 6-82 66-4 +1-1 
271°5 68-9 6-71 68-6 +03 
710-0 76:3 6-33 76-4 -O-1 
x=20-507+40-2. y=206-23+0-2. M=1-0. 
Ox serum 19° (serum B). Hasselbalch and Warburg. II. 
19-5 61-6 7-73 Unsuitable for calculation 
79-8 69-6 717 (see Fig. 20) 
182-0 73-0 6-82 
378°3 73-7 6-52 
Horse serum 38°. III, 
23:5 63-8 7:73 63-9 —O1 
24-6 63-4 771 64-1 —0-7 
45:6 67-5 7-47 67-0 +05 
104-6 71:8 7-14 71-0 +08 
430°3 17:3 6-57 77:8 —0°5 
#=12-01640-2. y=156-76+.0-2. M=0-7. 
Horse serum 38°. IV. 
10-4 58-1 8-05 579 +0-2 
35:2 63-3 7-56 63-6 —0:3 
106-7 69-2 711 68-9 +03 
130-7 69-7 7-03 69-8 —O-1 
535-6 749 6-42 76-9 —2-0 
2=11-64440-1. y=151-6740-1. M=0°3. 
Ox serum 38°. V. 
22:7 62-6 7:74 64-2 —16 
59-5 69-0 7:37 69-4 —0-4 
103-7 74-7 7-16 72-4 +2:3 
510-7 80-1 6-50 81-6 -155 
x=14-:012+0-6. y=172-6940-6. M=2-3, 
Human serum 38°. VI. 
15-2 57-2 7:88 58-7 -1:5 
18:3 60-7 7:82 59-5 +1:2 
84:5 68-7 7-21 68-2 +05 
562-1 78-7 6-45 79-0 —0:3 
x =14-237+0°3. y=170-83+0-3. M=1-4. 
Horse serum 19°. VII. : 
12-7 55:8 7-86 56-5 -0-7 
18-7 58-2 7-71 57-9 +03 
41-2 62-6 7-40 60-9 +1-7 
119-6 ; 64-4 6-95 65-2 -—0:°8 
512-5 71-1 6:36 70-9 +0- 
x=9-5901+40-2. y=131-87+0-2. M=1-2. 


296 


E. J. WARBURG 


Table LVIII (continued) 
Horse serum (19-5°) 20°, VIII. 


Vols. % CO, 
mm. Hg Vols. % CO, PH '(s) combined 
O, combined calculated calculated Difference 
13-1 55-6 7-86 56-7 - -1-1 
34-1 61-1 7-48 60-5 +0°6 
83-1 65:3 7:13 64-0 +1:3 
573-7 71-4 6-32 72-2 —0:8 
x=10-025+0°3. y=135-5140°3. M=1-4. 
Ox serum 19°. IX. 
16-0 58-5 7:78 59-3 —0°8 
43-1 64-7 7-40 63-9 +0°8 
118-6 69-4 6-99 68-9 +0:5 
545-9 75-9 6:37. 76-4 —0:5 
x=12-127+40-2. y=153-65+0-2. M=0°9. 
Ox serum 21°. X. 
16-1 61-8 781 61-4 +0:4 
40-5 66-0 7:44 66-0 0:0 
40-6 65:3 7-44 66-0 —0:7 
70-2 68-5 7-22 68-8 —0:3 
122-9 71:9 7-00 71:5 +0°4 
595-0 75-4 6-33 79-8 —4-4 
x=12-412+40-2. y=158-38+0-:2. M=0-7. 
Vol. % CO, * 
sot T] | | | ] | | 1 | ! ! | | | ek a | a 
60 ss es eo ene a * 
50-—- =a 
80}— PN oe els — 
y 
60}- “é a 
80}— au 
er anand ‘eRe cae 
70;— pare Tit! AV “4 
A faye sree > 16S epee each F J 
-+- 
50} as 
70}— + mW <i 
60}— Bee 
80 ha ao oan ae 
aT 
70}~ geen ee = 
Py —+ hn Sd Th a 
80 }— 9 fAgs- 
10\—~ 5 aaa a 
; I 
60\— bb & camel ae 7 
eta p 
50 |. Be 
| | i | ee st | l | eae: Saeed | 


2 al Ce i 
60 79 76 77 76 75 74 +73 ~=72 


71 70 69 68 67 66 65 64 63 62 


Fig. 20. The combination of CO, by sera, 


re oe aera ee 
me? an 

aS . 

~ 


ms 


THEORY OF THE HENDERSON-HASSELBALCH EQUATION — 297 


In a paper on the neutrality of the blood W. M. Bayliss [1919] has recently 
advanced the opinion that the proteins of serum and plasma could not function 
as ampholytes at the reactions obtaining in my experiments and in the above 
mentioned papers. I do not agree with Bayliss on this point and as the subject 
is of prime importance for the problems we are discussing I shall attempt 
briefly to dispute his statements. 


Vol.% CO, 
eee eet te tT 
ae eee, + 
70|— Seer . 
6o}- ——* 4 
+ 
60|— i a 
50-- = 
80/- ry 
710 i ioe cad =“ 
s+ vill 
—+ 
50}- en 
80}— at 
SL cea VII 
Py 
ere hk ype hte ef hy oy 


80 79 78 77 76 75 74 73 72 T1 70 69 BS 67 66 65 64 63 62 


Fig. 21. The combination of CO, by sera. 


Bayliss? asserts that serum does not combine with CO, in a reversible 
manner at the usual CO, tensions, supporting this by an experiment without 
doubt wrongly interpreted. It is obvious from the experiment that the variable 
combination—which Bayliss considers unimportant—is of the same order as 
that found by other authors. Bayliss further maintains that proteins do not 
act as buffers at reactions in the neighbourhood of the neutral point. This 
result is absolutely opposed to the finding of L. J. Henderson [1909-1910] and 
T. B. Robertson [1908-1910, 1912] with serum globulin and to the above 
mentioned experiments. The reason he has come to this erroneous conclusion 
must be that he has overrated the accuracy with which the reaction of plasma 
can be determined colorimetrically. In the same paper W. M. Bayliss has 
concluded by analogy that haemoglobin cannot act as a buffer. This analogy 
is however unjustifiable, quite apart from whether the experiments with 
plasma are correct or not, because the ampholytic character of the various 


* Mukai (J. Physiol. 1921, 55, 356) has quite recently, in Bayliss’ laboratory, found that 
serum combines with variable amounts of CO, of a similar order to those found by earlier authors, 


Bioch. xvr 20 


298 E. J. WARBURG 


proteins is very different as is undoubtedly shown by the work of Robertson 
[1912] and his collaborators, and by the studies of Pauli [1920] and his co- 
workers. 

In Table LIX and Fig. 22 the calculations from Campbell and Poulton’s 
[1920] experiments with dialysed haemoglobin solution at 38° will be found. 
The corresponding curves are here also straight lines allowing for experimental 
error. As 2226C] was 91-5 vols. % in the experiments the isoelectric point is 
7-14. This is, as Campbell and Poulton [1920] themselves have remarked, 
considerably more alkaline than Michaelis and Takahashi [1910] found at 
room temperature the isoelectric point according to these observers being 
py (Bjerrum) 6-79. Now it can be shown from a glance at the CO, combina- 
tion curves—as Hasselbalch [1916, 2] has pointed out—that the isoelectric 
point for haemoglobin must be more acid at body temperature than at room 
temperature (the difference is about 0-30), so that the difference between 
Campbell and Poulton’s and Michaelis’ results is real}. 


Table LIX. Dialysed solution of haemoglobin (Haldane 52) with 0-041n 
NaHCOsg, 38°; pA(m) = 6-23. Calculated after J. M. H. Campbell and E. P. 


Poulton. 
Vols. % CO, 


mm. Hg Vols. % CO, . pr combined 
co, combined calculated calculated Difference 
4:8 52:3 8-40 36-2 +161 
4-9 52-1 8-39 36-6 +155 
11-0 57-5 8-08 50:3 + 7:2 
21-9 63-3 7:83 61-2 + 21 
40-7 72-9 7-62 70-4 + 25 
43-6 76:5 7-61 70-9 + 56 
50-7 75:5 7-54 70-0 + 55 
52-5 70:2 7:47 77-0 —- 68 
78:1 74:1 7:34 82:8 - 87 
85:5 79-8 7:34 82:8 - 30 
90-0 77:1 7:30 84:5 — 7:4 
110-5 92°3 7-29 85-0 + 73 
119-8 86:3 7°22 88-0 - 17 
151-4 93-5 7:16 90-7 + 2:8 
221-9 99-1 7-02 96:8 + 2:3 
226-7 97-8 7:00 97-7 + O1 
310-4 101-9 6°88 103-0 -. 1] 
338-0 107-0 6:87 103-4 + 36 
435°3 110-0 6°77 107:8 + 2:2 
596-4 112-2 6-64 113-5 - 13 


%=43-95940-2. y=405-41402, M=4:5. 


Some experiments I have performed give an indication of the position 
of the isoelectric point. According to Hardy the agglutination optimum of 


' T. R. Parsons and Winifred Parsons (Biochem. Zeitschr. 1921, 126, 108) have recently 
pointed out that Campbell and Poulton’s haemoglobin solutions must have contained inorganic 
salte because when they incinerated the haemoglobin they found more ash than could have come 
from the iron alone. Calculating the amount of HCO’, which would be balanced by metallic ions 
on the assumption the excess of ash was K,CO, we get 30 vols, % and the isoelectric point is 
therefore about 6-5, which agrees with Michaelis’ determination when we take the difference of 
temperature into account, 


THEORY OF THE HENDERSON-HASSELBALCH EQUATION 299 


a colloid is at the iseelectric point. The crystallising out of haemoglobin 
at high CO, tensions (after saponin haemolysis) mentioned in chapter IV 
never began at reactions more alkaline than py: 7-10 at room temperature 
under the conditions of the experiment, but the ease with which crystallisation 
was set going increased until py 6-80 was reached which supports Michaelis’ 
result. Should further experiments confirm the difference between the iso- 
electric point estimated from the combined CO, and from the other method, 
it will indicate that some CO, is adsorbed or combined in a complex manner 
with the haemoglobin. 


110}- E i nl 


60 a 
50tt- > a lu Sd 
- 
a 

bi 

30}—- fel 
20 + 
~ 10 Py 4 

Teo | | | i l UO lie Sa je | | l | | 


84 83 82 81 80 79 78 77 76 75 74 73 72 71 70 69 68 67 66 65 64 


Fig. 22. CO, combined by dialysed haemoglobin (with the addition 
of 0-041n NaHCO,). (J. M. H. C. & E. P. P.) 


With the help of the relations developed in this chapter, which found their 
simplest expression in (173) and (176), it will be possible to investigate whether 
all the reversibly combined CO, is present as bicarbonate. Before reporting 
the experiments I have made in this connection I would draw attention to 
the fact that an error in pA,) or pAym) Or in the correction taken from Fig. 10 
is without significance here, because it will influence the estimation of the 
hydrogen ion exponent in the original and in the curve displaced by the 
addition of alkali to the same extent and thus the relative positions of the 
curves will not be altered. 

The experiments were carried out as follows: blood or saponin haemolysate 
(4 % saponin) which was slightly concentrated by pipetting off serum was 
sloped in a 100 cc. measuring flask so that it filled about half of it. With a 
pipette either 5 cc. n/2 NaCl solution or 5 cc. n/2 NaHCO, solution was then 
added during shaking, the titre of which was determined beforehand by the 
exhaust pump. Two drops of octyl alcohol were next added to prevent 
frothing after which the flask was filled up with blood (or haemolysate). 

20—2 


300 E. J. WARBURG 


With the mixtures thus prepared CO, combination curves were constructed 
with the help of the large saturator. The calculations were made as before, 
and the results are given in Table LX and in Figs. 23, 24 and 25. 


T T ] | ] | | | | 
170 2 + 4 
160 + + — 
150 ea) 
140}- é: _ 
130|— oe eS ts 
120 % 
ar a 
: a 
110} esi 
+ os 
100 
Pa 
es ira 
Ps ¢ Horse blood A 
80}- Pe He 
ce) 
> 
oe 
60;- = 
50 
20 ee 
10}—- ~~ 
Pr.) 
| | i | | | | l | 
78 677 «76 =#«75 «14 «73 «72 Tt 70 69 68 67 66 65 6-4 
Fig. 23, 
Let us first compare the constant 2. 
Lower curve Upper curve Difference 
Horse blood A 66-9 4-0-3 69-2 4-03 23-404 
Horse blood haemolysate B 87-0 4.0°6 89-2 +-0°5 2:2+0°7 
Ox F ne C 57-9 4-1-1 57-8 4-0-9 ~O-1+1:4 
Horse ,, pe D 813401 83:3 40-3 2-0-40:3 
” ” ” E 64-0 4-0-1 70:3 40:3 6:3-+.0:3 


THEORY OF THE HENDERSON-HASSELBALCH EQUATION 301 
Table LX. 7 
Defibrinated horse blood A. Curve L Y=11. 
mm. Hg Vols. % combined Vols. % CO, 
E———_, —“——7 (8) combined Difference 
Temp. CO, 0, CO, O, culated calculated vols. %, 
17 43-5 147-0 65-6 26-5 7-45 65-6 0-0 
6 99-8 135-5 83-8 26-4 718 83-7 + O01 
m 104-6 _ 85-4 26-4 717 84-3 + Ll 
ms 358-8 — 112-7 _— 6-72 114-4 - 17 
i. 537-8 — 123-5 — 6-59 123-1 + 0-4 
Mean 26-4 
x =66-883-+0-3. y=563'87+0-3. M=1-2. Theoretical distance 57-0. 
Curve II. 
17 40-4 147-6 105-6 26-9 7-71 104-9 + 0-7 
s 59-9 143-7 114-5 26-5 7-56 115-3 - 08 
re 168-0 120-5 140-9 26-4 7-18 141-6 - 0-7 
” 303-8 — 158-1 — 6-96 156-8 + 13 
PP 522-4 — 171-5 cane | 6-74 172-0 — 05 
Mean 26-6 
x =69-17240-3. y=638-24+0-3. M=1-1, 
Saponin haemolysed horse blood B. Curve lL. ¥=11. 
18 14-4 _ 33-6 26-2 7-60 31:3 + 23 
ik 45-2 _ 54-2 25-9 7-31 56-5 - 23 
* 84:8 _ 69-3 — 715 70-4 - Il 
ts 359-3 — 107-4 — 6-71 108-7 - 13 
‘a 536-8 — 121-4 — 6-59 119-2 + 2-2 
am 137-4 —_ 75-0 _ 6-97 86-1 -1)-1! 
BS 359-3 — 92-6 _ 6-64 114-8 — 22-2! 
1 Crystallisation. 
#=87-033 40-6. y=692-70+0-6. M=2-5, Theoretical distance 53-0. 
Curve II. 
18 20°3 — 69-4 _ 7-77 68-3 + ll 
& 45-6 ih 89-0 ie 7-52 90-6 =~ 16 
= 70-6 —_ 102-5 _ 7-40 101-3 + 1-2 
a 141-6 — 120-8 _ 7-16 122-7 -— 19 
ie 360-3 _ 150-7 — 6-86 149-5 + 1-2 
«x =89-176+40-5. y=761-2140-5. M=1-8. 
Saponin haemolysed ox blood C. Curve lL. ¥=8. 
19 19-4 _— 38-7 20-6 7-54 37-7 + 10 
be 60-2 — 56-7 —_ 7-21 56-8 - 01 
rs 394-1 . — 93-6 _— 6-62 91-0 + 2-6 
2 166-9 _ 73-6 — 6-86 77-1 - 35 
©=57-899+1-1. y=474-27+1-1. M=3-2. Theoretical distance 54-5. 
Curve IL. 
19 44-4 _ 91-1 20-3 7-55 91-6 - 05 
398-5 oa 135-4 il 6-77 136-6 - 12 
ms 164-9 — 120-1 _ 7-10 117-6 + 2-5 
F 82-3 _ 102-9 _— 7-34 103-7 -~ 08 
#=57:77240-9. y=527-7540-9. M=2-1. 
Saponin haemolysed horse blood D. Curve I. Y=11. 
18 18:3 133-9 36-7 25-0 7-53 36-4 + 03 
% 18-3 133-9 36-5 24-4 7:53 36-4 + 0-1 
17 52-9 — 57:8 — 7-26 58-3 - 05 
a9 91-9 _ 70-3 _— 711 70-5 — 0-2 
18 531-8 — 114-6 —_— 6-57 114-4 + 0-2 
17 153-4 —_ 81-5 — 6-95 83-5 — 2-0} 
e 348-0 re 97-4 sy 6-67 106-3 ~ 39 
1 Crystallisation. 


x=81-260+40-1. y=648-25+0-1. M=0-5. Theoretical distance 54-5. 


302 Kk. J. WARBURG 
Table LX (continued) 


mm. Hg Vols. % combined Vols. % CO, 
o—_——_ 7 —*-- PH (s) combined Difference 
Temp. Co, O, CO, 0, calculated calculated Vols. % 
Curve II. 
18 532-4 — 157-0 — 6-70 157-3 — 03 
17 16-8 —_ 66-9 — 7-82 64-0 + 29 
s 52-6 — 91-9 —- 7-47 93-2 - 13 
72-8 _ 100-0 — 7-36 102-3 — 23 
~ 147-5 120-4 —- 7-14 120-7 — 03 
ss 515-9 — 158-3 — 6-71 156-5 + 18 
a 320-0 — 141-8 _- 6-87 143-2 - 14 


1 Crystallisation. 
x =83-32240°3. y=715-5740-3. M=1-6. 


Saponin haemolysed horse blood E. Curve I. ¥=11. 


18 19-8 _ 41-7 25-7 7-55 41-7 0-0 
~ 61-7 — 61-6 — 7-23 62-2 - 06 
17 80-0 —_ 68-0 —_ 7-15 67-3 + 0-7 
= 155-6 —_— 80-9 —_ 6-94 80-8 + 0-1 
x 342-3 — 97-1 — 6-68 97-4 -— 03 
‘e. 545-9 —_ 107-8 —_— 6-52 107-7 + Ol 


x =64:052+0-1. y=525-29+0-1. M=0-5. Theoretical distance 54:5. 


Curve IT. 
18 22-1 —_ 78-6 26-0 7°78 q7-1 + 15 
2 101-9 — 111-3 —_ 7:27 113-0 - 17 
© 185-9 — 126-1 — 7-06 127:8 -—- 17 
17 301°3 — 140-0 — 6:89 139-8 + 0-2 
me 514-8 — 153-3 — 6-70 153-1 + 0-2 
ue 512-3 — 154-7 26-1 6-70 153-1 + 16 


«=70-335+40°3. y=624-:37+40-3. M=1-6. 


Apart from ox blood haemolysate C, which appears to be subject to 
a far greater fortuitous error than the other members of the series, the upper 
curves all exhibit a rather steeper ascent than the lower ones. Although we 
should be very cautious, as previously explained, in drawing definite conclusions 
from the size of the mean error about an actual discrepancy between the 
constants it is extremely probable the upper curve really has a somewhat 
steeper course than the lower. In accordance with this the equation for the 
combination of CO, in the blood and haemolysate should not be 


(173) B= (Cy — Cyf (ag:)) 2226, 
but (176) B= (Cy — Cyf (an’) — fi (Poo,)) 2226, 


and therefore some CO, should be combined in blood in some other form than 
bicarbonate. There is however an alternative explanation which seems to 
me more probable. As will be remembered it was explained on p. 280 that 
the difference between the curves at no place could be less than 2226b, which 
is the value entered in the tables as the “theoretical distance.” We will now 
investigate how far this is correct. 


THEORY OF THE HENDERSON-HASSELBALCH EQUATION — 303 


Horse blood A. 

” »» haemolysate B. 
Ox es pa C. 
Horse ,, As D 

¥ E. 


B Difference 
>. op a | ¥- A ee | 
Atpy: Atpy Atpg- At pq Theoretical 
7:50 6-70 7-50 6-70 distance 
Upper curve 119-5 1748 : F i 
wer ,, 630 1143 565 60-5 57-0 
Upper _,, 92-4 163-8 ; f 3 
Lower ee 40-0 108-6 52-4 55-2 53-0 
pper ,, 94:3 140-7 : : i 
Lower 40-0 sg4 4 54:3 54-5 
. Upper ,, 90-7 157-3 : . z 
Lower .. 388 1038 59 53-5 54-5 
Upper ,, 96-9 153-1 t é 
Lower 44-9 1. 2° = 8a 545 


There seems therefore to be an inclination to obtain too small differences 
in the most alkaline reactions. In the case of the haemolysates D and E this 
seems to exceed the experimental error as the latter may be estimated as 
below 2 vols. % of the difference between the curves, determined by five 
points or more. (The standard deviation for all the determinations with the 


160 


| 1 | | | | i] a4 | | 
150 
140 _—— ~ 
130 * mac 
120}- i < 
ik pia ie 
st + 
S Pi 
100 }—-3¢ 
s 
90}. > F . af 
80 |. “ 
x 
70} e 4 
Horse blood, haemolysed 
ry by saponin 3) 
B 
50 Y 
40 }- os] 
+ 
16 a 
ae 
| | 1 J | i | | | | 
era ee ee ee eee ee ek PO SR | EB. UG) 6G. BS 


304. E. J. WARBURG 


exception of- the horse blood haemolysate D is 1-2 vols. %.) A possible 
explanation of the phenomenon is that on the addition of bicarbonate the 
blood was very little changed so that before the bicarbonate was completely 
mixed with the blood a slight decomposition of the proteins took place with 
the formation of new acid radicles. 

The result of these experiments is therefore that it is saehcbes that CO, 
is combined in the blood only as carbonic acid, that is to say as bicarbonate 
ion, but that the possibility cannot be entirely excluded that small quantities 
may be combined in other ways. 

In the haemolysates B and D haemoglobin crystallised out in the acid 
reactions. It will be noticed that the combination of CO, decreases when 
crystallisation takes place but it must be remembered we cannot be sure of 
getting a homogeneous sample of the whole system when the crystal phase is in 
process of formation. Setschenow [1879] has observed a similar decrease of 
the combined CO, of horse blood when the haemoglobin crystallised out. 


160 
| i i] | | | | | | | 
ij f Ww 
150 P a 
we 
140 F- a 
130 }- 4 


+ 
Horse blood, haemolysed by saponin _ 


60 4 ? 
oS r 
50 


+ 
40}- Y - 
20 me — 
20 = —_ 
| ae | | Saw Be aa ES Pa ie | 
80 70 %78 77 #14 76 #974 73 #72 T4 70 69 68 67 66 65 G4 


Fig. 25. 


In conclusion I will report some experiments carried out in a similar 
manner with histidine hydrochloride with which Prof. Henriques kindly 
supplied me, which make it probable the CO, combination curve of histidine 


ee 


THEORY OF THE HENDERSON-HASSELBALCH EQUATION — 305 


can be shifted in a py'—vols. % CO, diagram in a similar way to that of blood. 
The preparation was impure as it only contained 17-1 % N, while the theo- 
retical amount for histidine mono-hydrochloride is 26-5 °% and for histidine di- 
hydrochloride 21-5 %. A 5 % solution was 0-439n as regards chlorine which 
corresponds to a mixture of mono- and di-hydrochloride. It can however 
be concluded from the course of the CO, combination curve that chloride 
must also have been present in some other form than histidine hydrochloride. 
The solutions were prepared in such a way that for the lower curve a 5% 
solution of histidine hydrochloride was mixed with an equal part of a sodium 
bicarbonate solution which combined with 995-9 vols. % just before each 
experiment, while for the upper curve a bicarbonate solution which combined 


with 1106-5 vols. °% was used. The theoretical distance between the curves 


should be 55-3 vols. %. The experiments willbe found in Table LXI and in 
Fig. 26. It will be observed the curves are very nearly parallel with an inter- 
vening distance of 53-5 to 57-5 vols. % which may be regarded as sufficiently 
accurate for curves of this type. It should be remarked however that the 
distance between the curves increases a little with increasing apparent hydrogen 
ion activity and that the distance in the most alkaline reactions is about 
2 vols. % less than the theoretically smallest possible distance so that there 
must either be small experimental errors or the theory proposed must be 
incomplete in one way or another. 


Table LXI. 2-5 % histidine hydrochloride pg’, = 6-300 at 18° 
and 6-305 at 17°. Y =8. 


Curve I. Curve IT. 
mm. Hg Vols.%CO, py mm. Hg Vols. % CO, , 
Temp. CO, combined calculated Temp. co, combined calculated 

At 22-5 35-7 7-41 18° 23-5 76-1 7-72 
ee 40-5 44-3 7:24 e 48-8 86-8 7-46 
a 83-4 57-7 7-05 ‘ 113-2 101-6 7:17 

ss 188-5 78-7 6-83 me 207-7 119-0 6-97 

* 392-3 103-0 6-63 ia 390-4 141-7 6-77 
731-0 127-5. 6-45 & 730-0 167-5 6-50 


Theoretical distance 55-3 vols. %. 


In the above discussion an assumption has been made which is not really 
tenable, namely that a complex combination between the CO, and proteins 
or histidine does not drive CO, out of the bicarbonate. If such a combination, 
for example a carbamino-acid, is in itself an acid it will in virtue of its cations 
be able to drive out CO,. Equation (176) should therefore—if this possibility 
be allowed—have the form 


= (Cy — Cy f (ax) — fi (Poo,) — fz (Poo, » %a")) 2226, 
and the above experiment would be explained without the assumption of 
an experimental error. Further experiment can alone determine whether this 
explanation accords with fact but the above experiment proves at any rate 
that only very small amounts of carbamino-acid can be formed at the reactions 
investigated. 


306 E. J. WARBURG 


As the final result of the investigations reported in chapters VIII, IX, X 
and XI it is clear that the CO, of the blood is exclusively or almost exclusively 
combined as bicarbonate and that the variable amount combined is due to . 
the presence in the blood of electrolytes the dissociation of which varies with 
the reaction. Such electrolytes are chiefly proteins. On account of the rela- 
tively small concentration of the serum proteins their effect will be subordinate 
to that of the proteins of the blood corpuscles. Haemoglobin is the most active 
of the blood corpuscle proteins judging from experiments of Setschenow [1879], 
Bohr [1905], Hasselbalch [1916, 2] and Campbell and Poulton [1920]. It is 
impossible to estimate the activity of haemoglobin in this respect but it is 
not improbable that the other proteins and lecithin (Setschenow) are also 
active. To a limited extent the phosphates will also function in a similar 
manner. 


150 


Vol. % CO, 


53} 


10} + 


i 


L 


| 


1 


| 


| 


| 


| 


Hes 


80 


79 


78 


Fig. 26. Histidine hydrochloride, Theoretical distance 55-3 vols. %. 


717 


76 


7% 


i4 


13 


V2 


he 


70 


69 


68 


67 


66 


65 


64 


63 


THEORY OF THE HENDERSON-HASSELBALCH EQUATION — 307 


Résumé. 


I. The theory of the combination of CO, in a solution which contains — 
both electrolytes the dissociation of which does and does not vary with the 
reaction is elaborated. 

II. This theory has been tested on phosphate solutions and found to be 
satisfactory. 

III. An empirical relation on a rational basis has been proposed for the 
combination of CO, in serum, haemoglobin solutions and blood. 

IV. It has been shown it is only possible for a trifling amount of CO, to 
be adsorbed or bound in a complex manner (as carbamino-acid) in blood or 
haemolysate. 

V. A similar proof has been produced for the combination of CO, in a 
histidine solution. 


CHAPTER XII 


THE FACTORS WHICH DETERMINE THE PARTITION OF PERMEATING IONS 

BETWEEN THE BLOOD CORPUSCLES AND THE SERUM, THE VOLUME OF 

THE BLOOD CORPUSCLES DEPENDENT UPON THE REACTION, AS WELL 
AS THE POTENTIAL ON THEIR SURFACES 


As repeatedly stated A. Schmidt [1867] and N. Zuntz [1867] discovered 
independently of one another that the amount of alkali in serum increased 
when the blood was treated with high tensions of CO,. Zuntz explained this 
phenomenon by assuming that the effect of CO, was to split off sodium from 
the sodium-protein compounds of the blood cells and that some of the sodium 
diffused out of the blood cells in the form of sodium bicarbonate until the 
sodium bicarbonate concentration was the same in blood cells and serum. 
When we remember the state of physical, and especially physiological, 
chemistry at that time the absolute genius displayed in Zuntz’s hypothesis 
excites the greatest admiration. 

The converse of this bicarbonate diffusion was discovered in 1874 without 
its importance in this connection being noticed. Hermann Nasse?, who at that 
time was an old investigator and who had gained a considerable reputa- 
tion in haematology, discovered in 1874 that the chlorine in serum decreased 
when blood was treated with high CO, tensions. At the same time he found 
that the blood cells swelled under CO, treatment, taking up water from the 
serum. Although he again reported his discovery in Pfliiger’s Archiv in 1878 
his name has quite disappeared from textbooks on physico-chemical biology 
because the honour of discovering the phenomena in question was falsely 
ascribed to Hamburger [1892, 1902] and v. Limbeck [1894]. In the literature 


1 Do not confuse with the son O. Nasse, the discoverer of isotonic salt solution, or with the 
father, the celebrated clinician of Bonn, Christian Friedrich Nasse. 


308 E. J. WARBURG 


dealing especially with ionic equilibrium in the blood I have only found 
H. Nasse mentioned by Giirber [1895, 1]. 

When it was decisively shown by the work of Gryns, Hamburger [1902], 
Hedin [1915], Overton, Giirber [1895], Koeppe [1897] and many others about 
the beginning of the century that the membrane of the blood cells was im- 
permeable for cations—Arrhenius’ ionic theory was at that time generally. 
accepted by physiologists—it became necessary to seek a new explanation 
of the diffusion of ions between blood cells and serum. From excellent but 
wrongly interpreted experiments Giirber [1895] formed a theory that CO, 
split up NaCl with the formation of free HC] and sodium bicarbonate in serum 
and that the HCl subsequently diffused into the blood cells. 

Koeppe [1897] advanced the theory that by CO, treatment monocarbonate 
ions are formed in the blood cells and that these are exchanged with chlorine 
ions from the serum—a theory which is only a special application of the 
principle enunciated by W. Ostwald in 1890 for ionic diffusion through a — 
semi-permeable membrane. In his handbook Hamburger [1902] has thoroughly 
dealt with these theories but considers it is impossible to judge which is right. 
E. Petry [1903] has partly identified himself with Giirber’s theory. But Giirber’s 
theory is undoubtedly fallacious because firstly carbonic acid could not split 
up an appreciable amount of NaCl, even if the latter was not completely 
dissociated, and secondly, the same CO, tension must be present in the blood 
cells and serum, so that the CO, must have the same effect on the NaCl on 
both sides of the membrane. Koeppe’s theory was falsified by the fact that 
he assumed monocarbonate ions could be formed in appreciable quantities by 
CO, treatment, but on account of the hydrogen ion activity this is impossible 
as shown by L. Meyer [1857], Setschenow [1879], Bohr [1905] and others. 
But if we substitute bicarbonate ions for monocarbonate ions Ostwald and 
Koeppe’s theory is sound. Koeppe was not clear why more carbonate ions 
should be formed in the blood cells than in serum when the blood is treated 
with high CO, tensions but we can fall back upon Zuntz’s [1882] old theory 
for the explanation, and it is of little importance in this connection whether 
we imagine the sodium to be originally combined in a complex manner with 
the proteins of the blood cells or whether we believe it to be present as ions 
partially balanced by electrolytes the dissociation of which varies with the 
reaction. 

Loewy and Zuntz [1894] have imitated the ionic exchange process by 
enclosing serum or blood cell fluid in a parchment dialysing bag according 
to Kiihne and dialysing against salt solutions, and they have shown that by 
treating the inner liquid with CO, alkali bicarbonate passed into the outer 
liquid. Loewy and Zuntz knew that alkali bicarbonate was liberated in this 
way both in the plasma and in the blood cells, but that by far the greatest 
amount appeared in the blood cells, and they have drawn attention to the 
connection between this phenomenon and the great capacity of the blood 
cells for combining with acid, Giirber [1895, 2] and Spiro and Henderson 


ee ee eee ee ae ee 
“| - et oe 


THEORY OF THE HENDERSON-HASSELBALCH EQUATION — 309 


[1909] have similarly imitated the exchange of ions with protein and salt 
solutions separated by parchment membranes with the same results. Spiro 
and Henderson have put the question very concisely by combining Koeppe’s 
hypothesis (with the revision mentioned) with Loewy and Zuntz’s views con- 
cerning the different combining properties of blood cells and serum. Hassel- 
balch and Warburg [1918]; without at the time being cognisant of Spiro and 
Henderson’s work, advanced a similar interpretation of the phenomenon, but 
taking into account the modern view of proteins as ampholytes, which however 
is not of great moment for the theory. Loewy and Zuntz, Giirber, Rona and 
Gyorgy [1913, 2] have made it very probable by dialysis experiments that 
the greater part of the combined CO, in serum is present as bicarbonate, but 
the explanation of the experiments involves many difficulties. Giirber, Loewy 
and Zuntz, C. Lehmann [1894] and Petry have shown that the alkali metals 
do not diffuse on treating the blood with CO,; the small differences found in 
the experiments are undoubtedly partly due to slight inaccuracies in the 
analyses and partly to the fact that the authors, with the exception of Petry, 
have disregarded volume changes of the blood cells. Hamburger and Bubanovic 
[1910] alone found a wandering of cations in some briefly reported experi- 
ments but this result is in opposition to what is known otherwise of the per- 
meability of the membrane of the blood cells. Reference should also be made 
to the handbooks and to Ege’s important work of 1920. 

L. J. Henderson [1920] has quite recently pointed out that O, and CO, 
combinations in blood, and Cl and bicarbonate in serum, are so related that 
we can conclude from the phase rule that there is a common cause for their 
mutual equilibria, but he did not explain if in detail. A question of con- 
siderable interest with regard to the considerations which are to follow is, how 
many inorganic cations there are in blood cells and serum, but the question 
cannot be solved at present as the published analyses do not agree with one 
another. The distribution of inorganic cations between the external and internal 
phases of the blood cells is unknown but we shall not be far wrong in assuming 
the concentration in the external phase to be about 0-15m. We can briefly 
express the Zuntz theory of the distribution of ions between blood cells and 
serum, taking into account the advance since Zuntz propounded his theory 
and remembering the work of Koeppe, Zuntz and Loewy, Rona and Gyérgy, 
and Spiro and Henderson, as follows. 

The blood cells are surrounded by a membrane impermeable to proteins 
and cations. Both blood cells and serum (plasma) contain electrolytes the 
dissociation of which varies with the reaction, but in the neighbourhood of 
the neutral point the content of the blood cells in these substances is greatly 
in the ascendant. Thus more bicarbonate ions are formed in the blood cells 
than in the serum (and fewer hydrogen ions in the blood cells than in the 
serum), when the blood is brought into equilibrium with an atmosphere in 
which the CO, tension is higher than that with which the blood was originally 


in equilibrium. As the equilibrium is disturbed in this way bicarbonate ions 


310 EK. J. WARBURG 


from the blood cells are exchanged with other anions (chiefly chlorine ions) 
from the serum till equilibrium is again established—until the apparent 
bicarbonate ion activity in the blood cells and serum is the same. It is of no 
importance whether the bicarbonate ions are balanced by inorganic or by 
protein cations. The above form of Zuntz’s theory is subject to a considerable 
defect which was foreign to the original theory. While Zuntz thought a part 
of the alkali of the blood cells was combined undissociated to the blood cell 
proteims and that these compounds were split up on treatment with CO, 
allowing sodium bicarbonate to diffuse from the blood cells into the serum 
until the concentration was the same in both, we must assume, since Gryns, 
Koeppe, Hamburger [1902], Hedin [1915], Giirber [1895 1, 2] and others have 
demonstrated the impermeability of the membrane of the blood cells to cations, 
that bicarbonate ions are exchanged for other anions (chiefly Cl’) until the 
concentration is the same in blood cells and serum. This opens up a new 
problem which has never been adequately discussed. If there was equilibrium 
before the CO, treatment of the blood between the chlorine ions in the blood 
cells and serum it requires explaining how it is possible there can also be an 
equilibrium after chlorine ions have wandered from the serum into the blood 
cells. In the Zuntz theory as modified by Loewy and Zuntz, Koeppe, Spiro 
and Henderson, and Hasselbalch and Warburg the assumption has however 
been made that the bicarbonate ion concentration or as it ought to be ex- 
pressed, the apparent bicarbonate ion activity, is the same in the blood cells 
and the serum independent of the reaction (and CO, tension). If on the other 
hand we assumed that the exchange of ions only proceeded until the relative 
decrease in tension of each kind of ion between blood cells and serum was the 
same we should have a rational explanation. 

It will now be shown thermodynamically with the help of Donnan’s [1911] 
distribution law that this explanation is a necessary consequence provided 
the assumptions relating to the permeability of the membrane of the blood 
cells and the electrolytes the dissociation of which varies with the reaction 
are correct. 

As early as 1890 W. Ostwald showed that diffusible ions must distribute 
themselves in a characteristic manner between two solutions separated by a 
membrane when ions were present which could not pass through it. Ostwald’s 
demonstration of this fact has however not been very fruitful to physiology 
and its application to the problem of the unequal distribution of ions referred 
to above was on the whole very difficult to grasp until Donnan published his 
work. As I fully share Donnan’s conviction that a knowledge of his distri- 
bution law will be of extreme importance for the appreciation of ionic equili- 
brium in the organism and a number of other physiological problems, e.g. in 
muscle physiology, I will shortly review the fundamental facts of the theory. 
A complete exposition is given in W. ©. McC, Lewis’ A System of Physical 
Chemistry [1916] as well as in Donnan’s own paper, to which the reader is re- 
ferred. 


THEORY OF THE HENDERSON-HASSELBALCH EQUATION 311 


Let us imagine two solutions separated by a membrane which is im- 
permeable to an anion R’, while the ions Na’ and Cl’ diffuse freely. There are 
originally only Na’ and RF’ in the one solution and only Na’ and Cl’ in the 
other according to the scheme 


Na’ Na’ 
R’ GU deny assess. (183) 
(1) (2) 
When equilibrium is established Na* and Cl’ have diffused from (2) to (1) 
so that we now have R’ Na‘ 
Na’ 
hs ee (184) 
(1) (2) 


Diffusion goes on until no work can be done by the simultaneous isothermous 
transference of a differential mol of Na’ and Cl’, so that we can obtain no 
energy by simultaneously carrying out the two isothermous differential pro- 


cesses dn mol Na’ (2) — (1), 
8n mol Cl’ (2) — (1). 


We therefore have the equilibrium equation 


CNa: (2) Cor (2) 
dn RT log, a) + 8m RT log, Gayrg{y = 0+ veeeeeeeeees (185) 
Ona‘ (1) Cor (2) 
Therefore a (®) = Gay (1) eteeseeeeeetesteeeseeeeee (186) 


on the assumption the solutions are ideal. 
Similarly for other ions in general we get 


Craw Cy ya Cae _ [un (1) [eum ri eb o[ea%s (2) ote 
= = = = Cl —_—_—__— *s 
Creve Cui Cara Cun Cur a1 (2) Cana 


where M’; is any monovalent cation and A’; any monovalent anion and so on. 
Donnan gives the following example: two solutions in equilibrium have 
originally the constitution 
a ee 
CG, @G 
(1) (2) 

where C, and C, are the molar ion concentrations in the solutions. Assuming 
complete activity and equal volumes of the solutions on either side of the 

membrane during the whole process, we get, when equilibrium is reached, 
Na’ ee 
Cita 0, x 

(1) (2) 

From this it follows that G, 100 is the quantity of NaCl which diffuses from 


312 E. J. WARBURG 


(2) to (1) expressed as a percentage of the original amount of NaCl in (2), and 


C,-=x 


is the distribution ratio of NaCl between (2) and (1) corresponding to 
the equilibrium. 
The equation in this case is 


C2 
sel OA (I) 
from which ar ee oer Ss (II) 
C.-2 0,40, 
eer ee Co (III) 
If C, is small compared with C, we-can put 
= poe es 
Gy Gp) cere eteeeeeeeseteeseseneeenens .-(IV) 
Ui<eo-gh 
and won Sats Tasacupes nates tee eptaear ees en (V) 


2 / 
If we take for example C, = 1 C,, then a = 0° that is to say only ragth 
2 


of the original amount of NaCl diffuses from (2) to (1). If on the other hand 
C, is small compared with C, 


and 


In the following table Donnan gives some calculations made by equations 
(I), (IL) and (IIT). 


Table LXII. 
Distribution 
Original Original Original % NaCl ratio of NaCl 
concentration concentration ratio of NaR diffused from between 
of Na R in (1) of NaCl in (2) to NaCl (2) to (1) (2) and (1) 
: C; x C,-x 
CO, 0, Os C,10 = 
0-01 1 0-01 49-7 1-01 
0-1 1 0-1 47-6 11 
l 1 l 33 2 
1 0-1 10 3 ll 
l 0-01 100 1 101 


It is clearly seen from the equations and tables that the ions for which 
the membrane is freely permeable will become very unevenly distributed 
between the two solutions if the concentration (activity) of the ions for which 
the membrane is impermeable is large in comparison with the former. 

In Fig. 27 we have the conception of Donnan in schematic form. The 
figure shows the cross-section of a cylinder in which a piston can move. The 
piston represents the semi-permeable membrane. Solution (1) is on the left 
of the piston, solution (2) on the right. The piston is situated in the middle 
of the cylinder and is fixed there, so that the volumes of the two solutions 
are equal. The ions for which the piston is impermeable will attempt to force 
the piston to the right (as they are only present in solution (1)), and since 
the ions of opposite charge (in this case Na’) are held back in the same amount 


THEORY OF THE HENDERSON-HASSELBALCH EQUATION 313 
by the electrostatic force, the osmotic pressure (the pressure on a square 
centimetre of the piston) will be 


pet 5 i Sentinal ene a ey ee (VII) 


provided that (2) contains only pure water. But as in the present example 
(2) also contains Na’ and Cl’, and as these ions are unevenly distributed in 
(1) and (2), the sodium chloride will exert a counter-pressure on the piston 
when equilibrium is established which will be 


P=2(C,—2) RT —2cRT = 2 (C, — 22) RT. ......(VII) 


The force per unit surface which will attempt to drive the piston to the 
right when equilibrium is established will then be 


P, =P, — P=2RT(C,-—C, + 22). Ce eeeeseeeeeees (IX) 
Therefore combining with (I) BA é 3 rs x noida gugdbaah ds daaesesuclastess (X) 


If C, is small compared with C,, P, = }P,; if C, is small compared with 
C,, P, = Py as would be expected. 
The following table (after Donnan) shows the relations: 


C, P, 
0, P, 
0-1 0-92 
1 0-67 
2 0-60 
10 0-52 


It will be observed therefore that if the concentration of the permeating 
ions is large compared with that of the non-permeating ions the counter- 
pressure of Donnan will have a very important influence on the magnitude 
of the osmotic pressure. 


(1) (2) 


SASSASS SASL 


bo 
~I 


Fig. 


Donnan and Allmand [1914] have investigated the distribution of KCl 
between two solutions separated by a copper ferrocyanide membrane, potassium 
ferrocyanide being also present in one of the solutions, and they have found 
the theory is supported on the whole. Bayliss [1909, 1911] has demonstrated 
the effect directly in experiments with congo red, H. R. Procter [1914] and 
‘Procter and J. A. Wilson [1916] have studied the gelatin-acid systems, and 


Bioch. xv1 21 


314 E. J. WARBURG 


S. P. L. Sérensen has very thoroughly investigated to what extent such an 
effect might be expected to influence his measurements of the osmotic pressure 
of egg albumin. 

The unequal distribution is of great importance in the purification of non- 
permeating substances by dialysis (cf. Donnan [1911] and Sorensen [1915- 
1917]), and may sometimes be important in experiments to determine the 
active concentrations (apparent activities) of permeating substances by com- 
pensation dialysis and ultrafiltration, as has been shown particularly by Rona 
and Gyérgy [1913] and H. J. Hamburger [1918]. 

Donnan has further shown by mathematical treatment analogous to the 
above that the amount of undissociated NaCl must be the same in the two 
solutions, which leads to the equation 

CNa: Cor 

CNaCl 

but this relation we know to be wrong (cf. chapter I). Donnan is unable to 
give an adequate explanation of this but it will be seen the difficulty disappears 
when we assume with Bjerrum that undissociated NaCl does not exist in watery 
solutions in the range of concentration in question. This point also brings out 
the superiority of Bjerrum’s dissociation theory over the classical theory of 
Arrhenius. : 

In the above example of Donnan it is assumed there are no non-permeating 
ions in solution (2) and that the volumes of the solutions do not change (the 
piston in Fig. 27 being fixed). If we assume on the other hand that NaR is 
present in both solutions and that water can be freely interchanged between 
them it will be obvious without any special mathematical treatment that the 
solutions will be of the same constitution so that there will be no uneven 
distribution of Na’ and Cl’. In this case therefore we may imagine that the 
piston is freely movable without obstruction and that water can freely pass 
through it. 

If there are non-permeating mols without a charge in one or both of the 
solutions but in a different ratio to that which the non-permeating (in this 
case negative) ions bear to one another, different concentrations of the non- 
permeating ions will be found in the two solutions when equilibrium is estab- 
lished and the permeating ions will consequently also be unequally distributed. 
A point of importance is the possibility that some force or other opposes the 
change in volume of the solution. In the illustration we can imagine this as 
happening by the piston doing work against spiral springs. In such a case 
the volumes, and therefore the concentrations, will not be dependent alone 
upon the degree of the permeating forces of the mols, and thus unequal 
distribution of the ions can ensue and conditions dependent thereon. If some 
of the non-permeating ions have a different valency to the permeating ones, 
a different number of permeating ions and non-permeating ions will be in 


= constant, ....... wevsessseeseeses-(190) 


* Jaques Loeb has published a series of papers in the Journal of General Physiol. 1920-1921 
on the influence of the Donnan effect on numerous factors concerned with gelatin solutions. 


THEORY OF THE HENDERSON-HASSELBALCH EQUATION 315 


electrostatic equilibrium mutually. If the volumes of the solutions are only 
determined by the non-permeating ions and the osmotic activity of the 
permeating ions electrostatically balanced by the former, unequal distribution — 
of the ions can arise. 

The last complicating factor of great significance is that the solutions in 
question are not ideal. The values in equation (187) are therefore not the 
concentrations but the apparent osmotic activities so that it becomes 


ONar y * Fa (Na*) ay) _ OM * Fa (M's) cy _ CA * Fa (AD ek ig, (191) 
ONar (2) X Pa(Na’) (2) = OM*1 (2) * Fa( Mt) (2) CA’) * Fa (AD) 

Let us briefly collect the results of the above. Donnan’s theory shows that 
the relative decrease in osmotic tension between solutions separated by 
semi-permeable membranes is the same for all monovalent ions, and that the 
relative osmotic tension of cations is the reciprocal of that of anions, and also 
that the relative decrease in osmotic tension of an n-valent ion is the nth 
root of that of a monovalent ion. If the concentrations of those permeating 
positive and negative ions which are balanced by non-permeating mols are 
different in solutions which are in equilibrium through a semi-permeable 
membrane, the osmotic activity of a permeating ion will be different in the 
solutions as first discovered by Ostwald [1890]. 

Let us consider a system of two homogeneous watery solutions separated 
by a semi-permeable membrane. The solutions contain both permeating and 
non-permeating ions. The non-permeating ions are distinguished by their 
symbols being put in parentheses, e.g. (M',(,)), that is, the non-permeating 
monovalent cation in solution (1). As the following exposition is only de- 
veloped as far as the present limited problems require and as experiments 
have not been carried out to make a general expression possible we will assume 
that in the solutions of permeating ions only chlorme and bicarbonate ions 
occur in such concentrations as are of significance for the exchange of ions, 
since the relations of the other ions are analogous with these two. 

The following non-permeating ions are also present in the system: 


(M";), (M1), (A’y) and (Ay). 
According to the law of equal positive and negative charges we have 
(Cura) + (2C ay) — (Ca) — (20a) — Ca’ Cuco’,@ = 9, (192) 
and (Cyg*,(2)) + (2C g4°*14(2») — (C_a’y¢a)) — (2C 4x12) — Con’ (a) — Cxc0’, (2) = 9, (193) 
therefore Cy‘ + Cyco’,) = (Cary) +(2€u" pw) — (C4) —(2C 47), (194) 
and Cy’ (2) + Cuco’,¢2) = (Cary) + (2C (2) — (C.a‘a)) — (20.4 y¢2))- (195) 


Col.) _ CHCO’s (2) _ CCl’ (2) + CHCO’s (2) 196 
From (187) we get & de Oa atCnoa eee (196) 


and therefore 


Corie) _ CHCO aia) _ (CM*1) + (2C M11 @) - (CA’t() - (20 An @) sie § (197) 
Cora, Cyco’s@ (CM 1a) +(2C Mn) -(CA1@) +264") 


316 E. J. WARBURG 


If the solutions are not ideal we obtain from (191) 


Cra) Fa (CV) (a) _ CHCO’s (2) Fa (HCO’s) (a) _ COV (a) Fa (Cl) (2) + CHCO’s (a) Pa (HCO'S) (2) 
COV Fa (Cl) ay ~~ CHCO’s w Fa (HCO’s) (1) CoV (1) Fa (Cl) a) + CHCO’s (a) Fa (HCO’s) (1) * 


For the further treatment of the present problem it will be convenient to 
rearrange equation (198) thus: 


CCl’ 2 ; 

Cora __ Fa(HCO’s)i» Fa (Cl) (199) 
CHCO’s (3) F, (HCO’s) aR Fi (CT) ‘ah © eeccvccessccccccoe 
CHCO’s «) 


If solutions (1) and (2) are heterogeneous, as is the case with serum and blood 
cells, (199) will be further complicated as the concentrations in the solutions 
are then mean concentrations and we therefore have to introduce a correction 
for this in equation (199) (cf. chapter III). Adding the subscript (c) to the 
symbols relating to blood cells and (s) to those relating to serum we get 


CCl’) 
Pq (HCO’S) (e) Pa (Cl’) ¢s CCl’ «s 
D' = fie AO a bi Se SeyPERRESS AO BORA 200 
#q (HCO's) (s) Pa (Cl) (ey CHCO’s te) ( 
CHCO’s («) 
The symbol D’ is introduced for the sake of convenience. 


Pa (Cl) oy . eee 
The value %, (HC0",)o C22 be regarded with very close approximation as 


being constant on account of the relatively small variations serum undergoes 
on changing the reaction. This is in the last resort due to the relatively small 


disperse phase in serum (circ. 8 %). Be 


, F, (HCO’ ) (ec) 7 
The value %(HO0%)@ jg equal to ete et PCL ,-..(201) 
Py (Cl die F (Cl’) 100 
oe £100 = qe (= 4 (CY) 


where d (Cl’) and d(HOO’,) are the relations between the concentrations of 
Cl’ and HCO’; respectively in the protein phase and external phase of the 
blood cells, while the activity coefficients refer to the external phase. In 


accordance with the reasoning in chapter VIII we may undoubtedly regard 


a (HCO’s) ¢e ae : 
Scan as constant—within the experimental error—for subsequent in- 


vestigation. The value 100 
100 — q& (1 —d (HCO’s)) 
ohm Se N cesdeaeets ene eaeees (202) 
100 — qe) (1 —d (Cl’)) 

must also be nearly constant as will be shown. 

4c) 18 the volume of the disperse phase expressed in vols. %. Ege [1920, 
1921] has shown this is between 35 and 40 %, of the blood cells. If we assume 
that 4%) in horse blood cells at py: 6-50 is 40 it will be seen from Fig. 9 that 
Yc) 4t Pye 7°90 is 40,02 = 34:5. d(Cl’)«) and d(HCO’;),) are the relations 
between the solubility of the chlorine and bicarbonate ions in the protein 


and water phases of the blood cells. In chapter IX for the bicarbonate 
ion we found d(HCO’,),. equal to about 4. Provided that the chlorine ions 


THEORY OF THE HENDERSON-HASSELBALCH EQUATION — 317 


are not soluble in the protein phase and d (Cl’).) therefore is 0 the variation 
of the conditions with the reaction will be a maximum. Substituting these 
values in (202) we obtain at py: 6-50 the value 75, and at py: 7-90, 79. 

It will therefore be appropriate to investigate the value of D’ at different 
reactions with the knowledge that we may expect it to be fairly constant. 
It is however not improbable it increases a little with decreasing apparent 
hydrogen ion activity. 

In chapter VI there are numerous determinations of pace 
fibrinated horse blood at room temperature and the results are given dia- 
grammatically in Fig. 6. I have therefore undertaken some determinations 
of the distribution of Cl between blood cells and serum, and they are to be 
found in Table LXIII. The technique employed was as follows: Defibrinated 
horse blood was saturated in the large saturator with gas containing CO, (by 


in de- 


Table LXIII. Horse blood. 
, Milligram 

uivalents of 

mm. Hg Vols. % combined Haema- lin 1 litre 


a7 - PH is tocrite (ene ceremony, 
Temp. CO, O, CO, O, cale. number blood cells D’ 


21 345 = 146-4 56-5 22:72 7-40 38-5 a 83-8 58-6 0-46 
106 
7445 C 683 50-8 0-48 
21 616-6 24:2 117-4 9-3 6-46 465 B 838 71-0 0-75 
S 95-0 
750 C 755 69-1 0-73 
21 18-9 152-4 46-8 22-2 7-69 370 B 838 35-7 0-32 
8 112 
705 C 68-7 50-5 0-45 
20-5 718 137-0 71-1 22-6 7-27 440 B 838 59-6 0-58 
S 103 
05 C712 57-9 0-56 
20" 197-5 = 112-5 91-6 22-0 6-92 455 B- 838 67-5 0-69 
S 915 
TU8.: EC T6 64-0 0-66 
20" 375-1 75:6 96-7 21:1 6-65 478 B 838 71-6 0-75 
S 95-1 
818 C 741 69-4 0-73 
20 =6150 155-9 48-2 21:3 7-81 42-0 rs 81-9 47-4 0-44 
107 
750 C 663 52-8 0-49 
20 741-5 0-7 122-0 0-6 6-43 475 B_ 81-9 70-6 0-77 
S 921 
725 C 126 65-2 0-71 
19 96-4 136-6 84-0 20-9 7-20 440 B- 81-9 59-6 0-60 
S 90-5 
1-0: (CO: 102 59-6 0-60 
19 2373 «102-9 = 107-6 21-9 6-90 460 B_ slg 66-6 0-70 
S 95-0 
763 CG 745° 68-2 0-72 
19 30-4 = 150-3 58-2 21-6 7-57 42-5 ri 81-9 53-4 0-52 
103 
222 19-4 155-4 48-6 20-6 425 B 819 46-8 0-43 
19-4 155-4 59-4 0-6 7-73 S 108 
194 155-4 37-9 39-1 7170 C 63-4 50-1 0-46 
222 744-9 0-0 122-4 0-4 463 B_ 819 70-7 0-77 
744-9 0-0 127-4 0-0 6-47 8 91-6 
744-9 0-0 8 116-4 — 780 C 746 69-9 0-76 


The blood was five days old and had a slightly sour smell. ® Hirudin blood. 


318 E. J. WARBURG 


the method described in chapter IV) the constitution of which was determined 
by analysis of the spirometer gas at the conclusion of the saturation. Samples 
of the blood were taken to determine the amount of combined CO,, and it 
was transferred to a glass cylinder, the CO, saturation being maintained, where 
it was allowed to sediment in contact with the saturating gas. The chlorine 
in the serum and blood cell suspension (and blood) was then estimated by 
means of a slightly modified titration according to Mohr and I. Bang’s [1916] 
method. The volume of the blood cells was estimated in the usual way with 
the haematocrite. In estimating the chlorine, serum, blood or blood cell sus- 
pension was transferred with a 1 ce. pipette to a 50 cc. measuring flask and 
the pipette then washed out three times with distilled water. The flask was | 
then filled about up to the mark with absolute alcohol with continuous shaking 


1-0 


T ! | | ! | I I if l | ] | | Beas 


0-9 + vad 


— 

08 See a hil 
%xX pa 
+ 


O7 ee at = eee 


ey 

06 r- - asi 
x, 

+ 


05 + 4 


0-4 _— — 
03 “= 
02 : ; pa 


Ol “i 


80 79 78 %17 «76 #75 .94 #%73 72 #94 +90 69 68 G7 66. 65 64 
Fig. 28. 


which produced a very fine precipitate. After 24 hours’ standing, during which 
the flask was usually shaken several times, the extraction was complete and 
it was filled right up to the mark with absolute alcohol, whereupon the flask 
was rotated several times. The solution was then filtered into small Bang- 
flasks the funnel being covered with a watch glass to prevent evaporation 
and the tube passed through a cork which fitted loosely into the neck of 
the flask. 10 cc. of the filtrate was transferred to three flasks and titrated 
with 7/100 silver nitrate with potassium chromate as indicator as described 
by Bang [1916]. 

In explanation of Table LXIII it is only necessary to state that B stands 
for blood, S for serum and C for blood cell suspension, 

In Fig. 28 all the results are given, those that are obtained from blood 
and serum are designated by “recumbent” crosses and those from blood cell 


THEORY OF THE HENDERSON-HASSELBALCH EQUATION 319 


suspensions and serum by “erect” crosses. A curve through the points was 
drawn freehand. This curve is marked D (Cl’). The curve from Fig. 6 (the 
curve showing the distribution of combined CQ,) is also given in the figure 
and is marked D(HCO’,). D’ is caleulated from the two curves and the 
result plotted as an independent curve. It will be seen that D’ is about 0-85. 
The experiments do not justify an opinion as to whether the small variation 
found is a real one or due to experimental error. The result of these experi- 
ments very strongly supports the hypothesis advanced for the distribution 
of ions between blood cells and serum—an hypothesis which may briefly be 
recapitulated thus: The permeating monovalent anions distribute themselves 
between blood cells and serum in such a way that the relative decrease in 
osmotic tension is the same for them all. The relative concentration (mean 
concentration) of the different anions will be a function of the relative osmotic 
tension, of the apparent activity coefficients of the ions in the external phase, 
and the solubility of the ions in the internal phase. 

It would have been of great help in the further elucidation of these 
problems had it been possible to measure the apparent (potentiometric) 
activity of the chlorine ions in blood cells and serum, but it is not. I shall 
take this opportunity of explaining why such an estimation must fail because 
it appears largely to have been overlooked and a knowledge of the facts is 
of fundamental importance in judging the value of many of the studies of 
Pauli and his collaborators—as is very evident from Pauli’s book which ap- 
peared in 1920. 

Bugarsky and Liebermann as early as 1898 attempted to measure the 
chlorine*ion concentration in egg albumin by means of a calomel electrode. 
Later Manube and Matula [1913] and Blasel and Matula [1914] amongst 
others used the same electrode for a similar determination in protein solutions. 
Oryng and Pauli [1915] noticed that serum albumin combined with mercurous 
ions but they claim to have shown these mercurous ions were split off again 
at more acid reactions. I cannot see however that the authors have succeeded 
in demonstrating this or that it has been proved the mercurous combination is 
independent of the salt concentration. Pauli and his collaborators have them- 
selves made interesting contributions to the problem of the combination by 
the proteins of various heavy metal ions, a combination which takes place 
beyond all doubt. 

If the chlorine ion concentration (activity) of a solution is to be determined 
with the help of an electrode of the second class, be it a mercurous or silver 
electrode, two conditions must be fulfilled. 

I. The mercurous or silver ion concentration in chloride-containing solu- 
tions must be determined solely by the activity ofthe chlorine ions, that is 
to say calomel or silver chloride must be precipitated as soon as more - 
than a trace of mercurous or silver nitrate is added. If the mercurous or 
silver ions are combined in a complex manner to any solute thereby diminishing 
the activity of these ions to an amount smaller than necessary for the pre- 


320 | E. J. WARBURG 


cipitation of the calomel or silver chloride, the estimation of the chlorine ion 
activity is not possible. 

II. None of the solutes in the solution (apart from a negligible amount of 
chlorine ions) must be affected by the presence of mercurous or silver ions in 
the solution. Several investigators including Pauli and his co-workers have 
shown that the proteins form complex compounds with mercurous and silver 
ions which have an extraordinarily low dissociation tension. I have myself 
some experience from unpublished experiments relating to the effect of the 
addition of silver nitrate to blood and the following results may be stated: 


(1) Considerably more silver nitrate than corresponds to n/100 can be 
added to the blood without silver chloride being precipitated. 


(2) Haemolysis gradually takes place which becomes total if sufficient 
silver nitrate is added. 


(3) At body temperature the haemoglobin is relatively slowly destroyed, 
and quicker in acid than in alkaline reactions. 


(4) The combination of CO, is rather less than in the original blood at 
the same reaction. 


(5) The O, combination curve at body temperature is much steeper than 
usual. 


It will be seen that neither of the two conditions for the determination of 
the chlorine ion activity is fulfilled. It is appropriate to draw attention here 
to a fact which in a large degree obscures many of the experiments of Pauli 
and his collaborators and undoubtedly compromises many of their results. 
Pauli and his collaborators, as was generally accepted at the time, always 
took electrically determined apparent activity to be equivalent to the corre- 
sponding concentration, which gives rise to an error in the solutions that were 
poor in salt and moderately rich in protein, which in some cases may be very 
significant. 

The experiments dealing with the distribution of chloride between blood 
cells and serum are further interesting as they help to throw light upon the 
conditions of the exchange of bicarbonate and chlorine ions. D, D, van Slyke 
and G. E. Cullen [1917] have tried to show that only about half the bicar- 
bonate produced by treating serum with CO, (measured at roughly constant 
reaction) is due to exchange with chlorine ions, while L. 8. Fridericia [1920] 
has found that the chlorine diffusion almost completely accounts for the 
production of bicarbonate. Van Slyke and Cullen as well as Fridericia have 
however disregarded the fact that the water yield of the serum to the blood 
cells affects their calculations. Van Slyke and Cullen found that serum from 
a sample of freshly-drawn blood contained 

0-1050n Cl’ and 0-031n CO,. 
After saturation with CO, at atmospheric pressure the serum yielded 
O-111n Cl’ and aequired 0-0206n HCO’. 
If we assume the serum was concentrated 10 °% the chlorine loss would be 


THEORY OF THE HENDERSON-HASSELBALCH EQUATION 321 


0-0216n, and the bicarbonate formation 0-0266n. So that the difference is 
really much less than van Slyke and Cullen calculated. 
Fridericia found that serum separated from blood at 0-08 mm. CO, con- 


tained 0-0992n Cl’ and 0-0250n HCO’,, at 163 mm. CO,. 
Serum separated from blood at 162 mm. CO, contained 

0-0814n Cl’ and 0-0463n HCO’,, at 163 mm. CO,. 
If we assume the serum was concentrated 8 % we obtain 

Cl’ HCO’; 
Difference — 0-0256n + 0-0183n 

so that more chlorine has passed over than corresponds to the bicarbonate 
which presumably is due to experimental error. 


In an experiment (Table LXIII) I found that serum separated from blood 
at 19-4 mm. CO, had a chlorine content of 0-108n, and that the bicarbonate 


content was 0‘027n at py 7°73. 


The volume of the serum was 57-5 % of the blood. 
Serum separated from blood at 744-9 mm, CO, had a chlorine content of 
0-0916n, and a bicarbonate content of 


0-0572n at py: 6-47. 


The volume of the serum was 53-7 % of the blood. 

In accordance with the determinations in a previous chapter horse serum 
will combine with about 12-5 vols. % CO, in changing from py: 7-73 to 
py 6:47. Subtracting this (0-006n) and correcting for the loss of water which 
is 9-3 % of the original volume of the serum we get 


Cl’ HCO’; 
Difference — 0-026n + 0-022n 


which result like that of Fridericia’s experiment gives too large a diffusion of 
chlorine and is probably likewise due to experimental error. 

F. C. McLean, H. A. Murray and L. J. Henderson [1920] have reported 
experiments according to which only two-thirds of the diffused bicarbonate is 
accounted for by the chlorine but these experiments are only known to me 
through a short reference in Medical Science Abstracts and Reviews, 5, p. 88, 
and it is not clear from this whether the authors corrected for the water loss 
of the serum?. 


1 In an article which recently appeared in the J. Biol. Chem. (1921, 47, 377) E. A. Doisy 
and E. P. Eaton have shown by accurate direct analyses that sodium and potassium do not 
diffuse on changing the CO, tension. They further showed that the diffusion of Cl’ on varying 
the CO, tension between 35 mm. and 100 mm. accurately corresponded to the diffusion of HCO’, 
(the change of volume was allowed for); on passing to very high CO, tensions in agreement with 
Fridericia and myself they found a too great diffusion of Cl’. Apart from possible experimental 
errors I am unable, as I have mentioned, to give an explanation of the latter phenomenon. 
Mukai (J. Physiol. 1921, 55, 356) has likewise shown by direct analysis of the salts that the 
cations do not diffuse through the surface of the blood cells on changing the CO, tension. 


322 E. J. WARBURG 


Before proceeding further with the investigation of equation (198) it will 
be well to examine the conditions which determine the change in volume 
of the blood cells at different CO, tensions. It must however be remarked 
that this investigation deals with the fundamental rather than the quanti- 
tative relations of this process because the experimental material in my 
possession was not performed with the present purpose in view, but is a 
by-product of the experiments in chapter VI, and the constants used are not 
very accurately ascertained. 

The discovery that the volume of the blood cells is increased by CO, 
treatment of the blood is credited to_H. Nasse, 1874, as already stated, but 
he did not attempt any explanation of the phenomenon. Later it was con- 
firmed by v. Limbeck [1894] and by Hamburger [1902]. Hamburger essayed 
to give a theoretical explanation of the processes on which it is based by 
accepting Koeppe’s theory of ionic exchange between blood cells and serum. 
Hamburger’s reasoning is as follows: 

When blood is treated with CO, monocarbonate ions are formed in the 
blood cells. One of these is exchanged for two chlorine ions from the serum 
(plasma). In this way the osmotically active mols in the blood cells are 
increased and they therefore swell up, water from the serum entering them 
until the osmotic pressure in blood cells and serum is the same. Hamburger’s 
view is erroneous—as is of course generally known—because monocarbonate 
ions are not present in sufficient quantities in the blood cells at the reactions 
employed and they do not increase in numbers on treating the blood with 
CO,; on the contrary they decrease. 

Koeppe’s [1897] theory which Hamburger at any rate originally supported 
is in its earliest form incapable of explaining the problem as Koeppe had no 
idea as to how the CO, (according to Koeppe the monocarbonate ions) was 
combined. 

In Spiro and Henderson’s work the old theories of Zuntz [1882], Koeppe 
[1897] and Loewy and Zuntz [1894] were revived in a modern form. Spiro 
and Henderson’s [1909] views were expressed in a rather different manner in 
conformity with the theory of ampholytes of Hasselbalech and Warburg 
[1918], which has been alluded to in more detail above. It is possible from the 
interpretation given by Loewy and Zuntz, Spiro and Henderson and Hassel- 
balch and Warburg of the distribution of ions in the blood to explain the 
change in volume of the blood cells which is determined by the reaction as 
Spiro and Henderson have briefly indicated. 

Rich. Ege [1920], who has made the most important contribution in 
physiological literature since the beginning of the century to the question of 
the volume of the blood cells and who from nearly every point of view has 
decisively shown that it is a simple function of the osmotic pressure of the 
liquid they are suspended in, failed when it came to explaining the change of 
volume caused by varying reaction. He has however reported an experiment 
in which he added HCl to a blood cell suspension in physiological salt solution. 


THEORY OF THE HENDERSON-HASSELBALCH EQUATION 323 


This caused the apparent hydrogen ion activity to rise to py 3-7 and the blood 
cells swelled 16 %. A freezing point determination of the external liquid and 
blood cell liquids gave the same osmotic pressure and Ege could therefore 
conclude the conditions obtaining in the volume change consequent on change 
of reaction did not differ fundamentally from those which determine the other 
kinds of volume change of the blood cells'. 

The mols which can exert an osmotic pressure on the membrane of the 
blood cells must either be incapable of passing through it, that is, be non- 
permeating, or they must be retained on one side of the membrane by electro- 
static forces. The non-permeating mols of any importance in this respect 
are the cations which are all considered to be non-permeating, and protein 
mols (both charged and uncharged). The permeating mols which are important 
are consequently the anions. There is over 30 °% of haemoglobin in the blood 
cells and about 8 % of other stroma proteins. If we assume with A. V. Hill that 
the haemoglobin molecules are aggregated with a mean aggregation number 
of 2-5, the molecular weight of the aggregated haemoglobin is about 40,000. 
We shall therefore not incur a large error in assuming the osmolar concen- 
tration of the blood cell proteins to be m/100, provided they ionise without 
destruction of the molecule which is now accepted by nearly everyone (cf. 
Pauli’s [1920] criticism of T. B. Robertson’s hypothesis of protein ionisation) ; 
and if the protein aggregates do not further aggregate or split up perceptibly 
under the given conditions the osmotic pressure exerted by the protein mols 
will not be altered by varying the reaction. Even if the assumptions con- 
cerning the aggregation should not turn out to be absolutely correct it will 
be of no consequence on account of the small osmotic activity of the proteins. 
On ionisation of the proteins however different amounts of permeating ions 
can be combined by the blood cells and it is this condition which determines 
the change in volume of the blood cells at different reactions. We will dwell 
upon this condition a little and examine it systematically on account of its 
great theoretical interest. 

If we consider the conditions of an apparent hydrogen ion activity where 
the number of positive and negative charges associated with the blood cell 
proteins are equal, the amount of permeating anions retained in the blood 
‘cells will be equal to the amount of inorganic cations present, which in this 
connection may be taken to be the same as the potassium and sodium con- 
centrations. If we now examine the conditions in a more alkaline reaction, 
protein anions will be formed in the blood cells in greater quantities and 
protein cations will have disappeared (cf. chapter XI); at the same moment 
there will be liberated an equivalent amount of bicarbonate ions in the 
form of CO,, as protein ions now are balancing the inorganic cations corre- 
sponding to the expelled CO,. In this way the osmolar concentration is 


1 Mukai (J. Physiol. 1921, 55, 356), by determining the osmotic pressure of the blood cell 
fluids and serum by Barger’s method, has shown that it is the same in both phases whether we 
,are dealing with defibrinated blood treated with CO, or with untreated defibrinated blood. 


324 K. J. WARBURG 


decreased and the blood cells will shrink as water is transferred from the blood 
cells to the serum until the osmotic pressure is equal on either side of the blood 
cell membrane. 

If we now consider the conditions when the reaction is more acid than 
before protein cations will be formed in excess of protein anions, the latter 
being simultaneously decreased in amount, and therefore bicarbonate ions 
will be held in equilibrium with the protein cations, whereby the osmolar con- 
centration of the blood cells is increased and they will increase in size. It 
does not matter whether the bicarbonate ions on reaching equilibrium are 
exchanged with the same number of chlorine ions as the osmolar concentration 
is not altered in this way. The conditions are actually rather more involved 
than indicated here, three complications particularly being active: 

I. In serum as well as in blood cells there are proteins (ampholytes), the 
dissociation of which depends on the reaction, which will oppose the change 
of volume but their action is only about 75th of the blood cells. 

II. During the ionic equilibrium an osmotic counter-pressure arises from 
the unequal distribution of the permeating ions as previously explained in 
accordance with Donnan’s views. It may be surmised however that this 
pressure must be relatively slight. 

III. The apparent osmotic coefficients should change a little on ionisation 
of the proteins. 

I have made an attempt by a rough approximation to find out the 
mathematical relations of the volume change determined by the reaction and 
I have compared the result with the experimental one shown in Fig. 9. It 
should however be mentioned that the experiments in the first part of 
Table LXIII are in poor agreement with my other experiments in this con- 
nection but I can give no explanation of the fact. In the following no account 
is taken of the Donnan counter-pressure nor of any changes in the osmotic 
coefficient which might occur, but it is obvious that the Donnan counter- 
pressure will cause the volume change in the alkaline reactions to be smaller 
than that calculated. 

The osmotic pressure on the membrane of the blood cells exerted by their 
fluids is made up of three components, namely the pressures due to the 
proteins, the cations, and the anions. If further we assume that approxi- 
mately no exchange takes place between the mols of the protein phase and 
the external phase of the blood cells, we get this expression for the osmotic 
pressure in terms of normality exerted by the fluids of the blood cells, 


(K*Q) + C*¥ 4) Fo (A); 
and for that exerted by the serum, 
(K*.) 7 O* 4/0) Fy (A’G)), 


where O*,’ only refers to the anions the concentration of which varies with 
the reaction, the phosphates being excluded, and where an asterisk signifies 


— eS ee eee 


a ee 


THEORY OF THE HENDERSON-HASSELBALCH EQUATION 325 


that the symbol refers to an external phase. Further we assume with Ege that 
the osmotic pressure in the blood cells is the same as that in serum, which is 
approximately true. 

On this assumption we will now consider a sample of horse blood in which 
the volume of the blood cells at py: 6-50 is 40 vols. % of the whole blood, 


therefore at pa 650 Q = 40, 


and the disperse phase of the blood cells is 35 vols. % of the blood cells, 
therefore at py: 6-50 q(.) = 35. Moreover we put at py: 6-50 q,,) = 9. 


8 y oo 
H g , External phase E35 
1 Fy External phase of Serum : of Er 2 
33 4 Blood Corpuscles S38 
Qo, Qu 


Fig. 29. 


We can represent this blood diagrammatically at py: 6-50 by Fig. 29. 
The contents of the cylinder are 100 cc. of blood, the piston is the semi- 
permeable membrane which separates serum from the blood cells, dividing 
the cylinder into one compartment of 60 cc. which corresponds to the serum 
and another of 40 cc. which corresponds to the blood cells. From the serum 
space a portion 5-4 cc. is cut off which is the volume of the protein phase, 
and from the blood cell space a portion 14 cc. which is the volume of the 
blood cell proteins. The piston is assumed to be freely movable without 
resistance and is kept in place by the osmotic forces. We further assume that 
the osmotic pressure at py* 6-50 both in serum and blood cells corresponds 
to an osmotic activity of 0-3 mol. We can therefore write 

K*,.) = K*,,) = 0°30 at Pu 6-50. 

If during a change of reaction in the water phase of the blood cells a osmotic 
activities per litre disappear and in the water phase of the serum b osmotic 
activities, the osmotic activity of the blood cells will be 0-30 — a, and of the 
serum 0-30 — b. 

If there is always to be the same osmotic activity in the blood cells and 
in the serum, water must be exchanged between them. We will call the 


quantity of water, expressed in vols. % of the whole blood, which must pass 
from blood cells to serum and therefore 


q - Ma 100- 9-24. 
(0:3 how a) =e = (0-3 = b) 100 -@ 9 ...(203) 
= 100 - Q- 100 V(s) +X 


from which it appears that the volume change of the blood cells is a function 
of the volume of the blood cells in the blood, which is easy to understand when 
we remember that in the limiting case where no serum is present the blood 
cells are quite unable to alter their volume by the appropriation of water. 


326 E. J. WARBURG 


In an earlier chapter we found that the combination of CO, in blood cells 

and serum can be expressed by 
B= y — =p. 

For haemolysed horse blood with a combined O, content of 24-7 vols. % 
we found x = 81, and for the other haemolysate with 26-0 vols. % combined 
O,, x = 87. In the latter haemolysate there must have been about 50 vols. % 
serum and since the constant « for horse serum at room temperature is about 10 
we obtain x = 80 (cire.) for the share of the blood cell proteins. In pure blood 
cell fluids we may therefore expect a constant of about 160. The newly formed 
ions distribute themselves between the protein phase and the external phase 
but at the same time their osmotic activity is somewhat depreciated by the 
Milner effect. I have therefore calculated that the change in the osmotic 
activity of the bicarbonate ions on passing from py" (1) to py" (y—and conse- | 
quently after exchange of bicarbonate and chlorine ions—can be expressed 
in terms of activity by 200 (pEr w) — PHT «@)) 

. 2226 
In the case of serum I have estimated half of the compound the dissociation ~ 
of which varies with the reaction comes from the phosphates, and that the 
change in the osmotic activity of the bicarbonate ions (and the chlorine ions 
exchanged with them) can be expressed by 


5 (PE a) — PH) . 
2226 


As the variation in the combined CO, is a direct measure of the altered ionisa- 
tion of the proteins we have, in the case under consideration, 


200 5 
@ = Song (Pu — 6°50) and b = 5555 (Pu — 6°50). 


Substituting these values in (203) we get 


125 (py: — 6-50) 
oa ton eae ne ee (204) 


and expressing the volume of the blood cells as a percentage of the volume 


t Du 6°50 Fe 1250 (py — 6-50) 
at py 6 @y = 100 — grat ap DOBO)’ Cette tnettees (205) 


In the first column of the following table the py: values are given; in the 
second, the volume expressed as a percentage of the volume at py: 6-50 read 
directly from curve II, Fig. 9; and in the last column the same amount 
calculated with (205). 


Table LXTV. 


Volume of blood cells in % of their 
volume at py 6:50 


ie aciply ' 
Pr Found Calculated with (205) 
6:50 100 100 
6°80 95-2 96 
7-00 92-7 93 
7°20 90-7 91 
740 89-1 88 
7-60 87-8 85 
7:80 86-3 81 


THEORY OF THE- HENDERSON-HASSELBALCH EQUATION 327 


It will be seen that the calculation indicates more shrinking of the blood 
cells than is actually found. In the calculation however the assumption has 
been made that py: is the same in blood cells and serum but as shown in 
chapter VII this is not true. Taking this into account and using for the reac- 
tion of the blood cells the value found from Fig. 11, curve II the agreement 
between the found and calculated results becomes surprisingly good. 

By substitution in, and modification of, equation (203) in accordance with 


the above it becomes 
x, = 100 14200 (0-000 (pir) — 50) 0-002 (Prig—O50)) ( 
97 — (9-2 (py"(e) — 6°50) + 0°49 (pH (s) — 6°50) 
Table LXV below gives the results of calculation with this revised equation 
compared with the experimentally obtained values. 


Table LXV. 


Volume of blood cells in % of their 
volume at py 6-50 
pS 


METLR | 
PH) PH) Found Calculated with (206) 
6-50 6-49 100 100 
6-80 6-78 95-2 97 
7:00 6-96 92-7 94 
7:20 713 90-7 91 
7:40 7-29 89-1 90 
7:60 7:45 87-8 88 
7-80 7-60 86-3 86 


As stated the agreement is very good, better indeed than we had a right 
to expect. The result is greatly in favour of the theory being the right one and 
that complicating factors of much importance do not play a part. 

Let us revert to equation (198) 


CHc0’s (a), Fa (HCO's) (2) _ COV'(2) Fa (Cl’) (2) + CHCO’s (2) Fa (HCO’s) (2) 
Cyco’s a) “* fa (HCO’s)q) CON Fa(Cl’) a) +CHco’sa Fa (HCO’s) (1)? 


from which we approximately assume, at the concentrations employed, 
@C0's (2) CCI2) + CHCO’s (2) 
Be eee petrnetsoneevaseactances (207) 
which in conjunction with (194) and (195) gives 


{CO's (2) _ (CM°1(2)) +(2C- Morr (2)) —(C.A’1(2)) — (2C-A’t1 (2) (208) 
%C0's) (OC M*1 ay) + (20 Meir ay) — (C44) — (20 Ara) © 


Indicating the concentrations relating to the external phases of blood cells 
and serum with an asterisk we get 
@* 00's _ (O*¥ M1) + (20* M11) — (C* 411) — (20* A") . ...(209) 

@*HC0's (8) (C*M*1(a)) + (20* Me-qx(2)) — (C*A1(0)) — (20% A" t18)) * 

The numerator in the fraction on the right side of the equation stands for 
the difference between the non-permeating cations and anions in the external 
phase of the blood cells, and the denominator for the same difference in the 
external phase of the serum. At a reaction in the neighbourhood of py: 6-50 
the ionic activity in serum and blood cells is the same as regards the per- 
meating ions. Consequently the difference referred to in the external phase 
of the blood cells and serum must also be of the same magnitude if (209) is 


328 E. J. WARBURG ° 


to be satisfied. And since the osmotic pressure of serum is almost identical 
with that of a 0-15m NaCl solution we can put the difference at this reaction 
without incurring a great error, at 0-15. If we also assume that the variation 
_ In the difference can be expressed in the same way as in the case of the volume 
of the blood cells we obtain 
200 
reotee (o-15 - Sy4 (PH © - 6-50) ) (173 - 0-732) 
* 5 ‘ 
HCO (0-15 — S55 (our to ~ 6°50) ) (I-54, ~54) 


where 2, is determined by (206). The results are given in Table LXVI. In 


the third column the most probable values for pre will be found, being 


the reciprocals of the ratio between the apparent hydrogen ion ADULTE of 
blood cells and serum. 


Table LXVI. 
@* CO's) 3 
a* s 
HCO 3(s) Poona 
= a. “aioess Seer st TES 
PH'(s) PH (ec) Found Calculated Calculated 
6-49 6-50 (1-02) 1-00 (1-02) 
6-80 6-78 0-96 0-90 1-07 
7-00 6-96 0-91 0-84 1-08 
7-20 7-13 0-85 0-77 1-10 
7-40 7-29 0-78 0-66 1-15 
7-70 7:45 0-71 0-59 1-20 
7-80 7-60 0-63 0-49 1-29 


According to calculation the difference in activity should be considerably 
greater than it is found to be, but it must be remembered the approximation 
we made in evolving (207) is not very good. 

On the whole the results obtained, including the last part of the problem, 
support the theories advanced although it is quite evident several of the 
assumptions made are only rough approximations and are undoubtedly sus- 
ceptible of considerable refinement. 

Donnan [1911] has developed a simple relation between the ionic equili- 
brium of solutions separated by a semi-permeable membrane and the electrical 
potential difference between the two sides. The following is taken from 
Donnan and W. C. McC. Lewis. 

Let us return again to the system (183) in the equilibrium (184) 


R’ Na’ 
Na’ 
Cl’ - Cl’ 


Let 7, and 7, be the potentials of the positive electricity in solutions (1) 
and (2). If we transfer isothermally the very small amount of positive electricity 
Fn from (2) to (1) we have the two following relations: 

(a) The increase in free electric energy = Fn (a, — 7). 

(b) pén mols of the Na’ ions are transferred from (2) to (1) and simul- 
taneously q5n mols Cl’ ions from (1) to (2), where p + q = 1 (p and q represent 


— ee 


en 


re 


-THEORY OF THE HENDERSON-HASSELBALCH EQUATION — 329 


the fraction of the total current which the respective ions have carried, or 
in other words they are the transference numbers of the ions). The maximal 
work for (6) is in an ideal solution given by 


ponRT log Ae xe (2) fe qanRT log 5 ao: of 


’ (9)? 
and as the electric forces balance fe osmotic we ak 


F8n (7) — 7) = pSnR T log ee ~——®. gn RT log eee em taht (211) 


It has been shown above that the failowites spate holds 


CNa‘ (2) _ COV) 
CNa‘a) CCl‘2 


and since Pe Ss a Set (212) 
RT CNa’ tans RT Col, ) 
we get 71) — M2) = Fr log GN haat “er log G a: Ratouesades (213) 


As it has already been shown that the permeable ions distribute them- 
selves with variable activity between blood cells and serum, a potential 
difference will exist between the two sides of the membrane of the blood 
cells. This can easily be calculated by 


RT ON te a* i 
Ts) — Mc) = I "Pasa 2.5 = ‘log cage ee seeeecceeses (214) 
At 18° we therefore have 
Ts) — Te) = 0-058 (Pao a Dio) eter eee ceeeccescess (215) 


If we now calculate the numerical values of (215) with the aid of the most 
probable corresponding values of py.) and pyr(,) We obtain 


Table LX VII. 


T(sy—T(e) 
PH (s) PH volt 

6-49 6-50 — 0-006 
6-80 6-78 +0-0012 
7-00 6-96 +0-0023 
7-20 7-13 +0-0041 
7-40 7:29 +0-0064 
7:60 7-45 +0-0087 
7:80 7:60 +0-0116 


The table shows that serum is positively charged against the blood 
cells, in other words the blood cells wander to the anode. At a reaction in the 
neighbourhood of py 6-50 the blood cells change the sign of their charge. 
As far as one can judge, this result is in harmony with the facts but the 
experimental investigations into the charge of the blood cells will hardly bear 
a searching criticism. Although it is far from being my purpose to deal 
thoroughly with the views put forward concerning membrane potentials in 
the physiological literature (muscle and nerve physiology is particularly rich 
in them), it will be worth while adding a few remarks about the above calcu- 
lations. Those especially interested are referred to Zangger’s [1908] and 
Beutner’s [1920] monographs and to the textbooks of physico-chemical 
biology, particularly Héber’s [1914] and McClendon’s [1917]. 


Bioch. xvr 22 


330 E. J. WARBURG 


The above calculations are distinguished from all earlier calculations and 
views touching upon the membrane potentials at the surfaces of cells in that 
they are based on Donnan’s theory. The new fact is made use of that it does 
not matter which permeating ion we credit with the production of the potential 
as all the permeating ions give rise to the same potential. A condition of the 
views advanced is here—as elsewhere in this work—that equilibrium is reached 
in the various systems, or that the equilibria at all events are so nearly 
approached that any deviations which occur take place very slowly. From 
this point of view they differ fundamentally from Beutner’s ideas as he in- 
vestigated the potential difference for systems of two aqueous electrolyte 
solutions separated by non-aqueous phases in which the cations and anions 
diffused with very different but nevertheless appreciable speed, so that the 
development of the potential difference between the watery solutions was 
determined by the different rate of diffusion in the intermediate liquid. Con- 
sequently the potential will not have been developed in Beutner’s systems 
when equilibrium is reached. 

As early as 1904 Héber [1904, 2; 1914] had experimentally examined 
the potentials on the surface of the blood cells. He used a chamber the height 
of which was only 100u. It seems probable that in such a shallow chamber_ 
Héber could not ensure that the movement of the blood cells he observed under 
the influence of a weak current did not arise from the motion of the water 
along the surface of the glass since, according to Burton [1916], the stream 
of water is present for a distance of about 25u. Héber conjectures from his 
experiments that the blood cell membrane is originally impermeable for 
cations and anions, and only becomes permeable for anions after CO, treat- 
ment. Some of H. Nasse’s [1878] and Hamburger’s [1902] experiments on 
the diffusion of chlorine in blood on blowing gas through it had however 
already made Hoéber’s view improbable. Fridericia [1920] has also recently 
shown that it must be fallacious. I shall not further discuss Héber’s con- 
tentions as I feel convinced Héber himself could not wish to retain them in 
their original form after the appearance of Donnan’s work. | 

A few years ago Fahraeus [1918] attributed the tendency of the blood cells 
to aggregate (and consequent rapid sinking) to a diminished potential on their 
surface. The experimental technique of Fahraeus however leaves so much 
to be desired that the hypothesis advanced stands on insecure foundations. 
In a recently published very exhaustive investigation on “The Suspension- 
stability of the Blood,” F&ahraeus [1921] seems partly to have relinquished 
his original view as he has shown that the substances which favour the 
aggregation of blood cells are the fibrinogen and the globulins of the plasma 
without however having elicited any definite connection between the amount 
of fibrinogen and the potential. H. C. Gram [1921] independently of Fahraeus 
has come to the same conclusion regarding the relation between the 
fibrinogen of the plasma and the rate of sinking of the blood cells by the 
examination of a very large number of samples of human blood. Linzenmeyer’s 


THEORY OF THE HENDERSON-HASSELBALCH EQUATION © 331 


[1920, 1921] work also supports to some extent this view, the relation of 
which to the haematology of x Bypone days is so excellently portrayed by 
Fahraeus. 

I have attempted to carry out new determinations of the charge on the 
blood cells but I encountered great experimental difficulties. Such determina- 
tions would be of considerable interest for the problems before us as we could 
test the accuracy of (215) with their help. 

In concluding this work I would again draw attention to the very great 
difficulties involved in bringing the blood into equilibrium with a new CO, 
tension, a fact which we time and again encounter in the literature and 
which vitiates the significance of a large portion of the experiments reported 
there. It is generally assumed it is the mixing that is the sluggish process 
in the saturation of the blood cells with CO,, and that the diffusion of CO, 
through the membrane of the blood cells, the interchanging of the ions and 
the ionisation of the proteins take place rapidly in comparison with it. This 
has however not been experimentally proved but the question of the cause 
of the difficulties of saturation gains physiological importance if we try to 
argue from the combined CO, at equilibrium in the blood in vitro to the 


conditions in vivo. 
RésuMeE. 


I. Donnan’s theory of the distribution of permeating ions in two-phase 
systems containing non-permeating ions is discussed and developed particu- 
larly with regard to its application to blood. 

II. The distribution of chloride between the blood cells and serum of 
horse blood has been determined. 

IIT. It has been proved that the distribution of bicarbonate and chlorine 
ions between blood cells and serum is in accord with Donnan’s theory. 

IV. The relation between chlorine ion and bicarbonate ion exchange has 
been discussed on the basis of experiments by van Slyke and Cullen, Fridericia, 
and the author. 

V. The view that the volume change of the blood cells which is deter- 
mined by the reaction is an osmotic phenomenon has been strengthened. 

VI. The electric potential of the surface of blood cells has been calculated 
with Donnan’s equation. 

VII. It has been shown that it is theoretically impossible directly to 
determine the apparent chlorine ion activity of the blood with an electrode 
of the second class. 


I am greatly indebted to all those who in various ways have helped me 
while this work was in progress. My thanks are especially due to my teacher, 
Dr K. A. Hasselbalch, late director of the laboratory of the Finsen Institute 
in Copenhagen, and to the present director, Dr C. Sonne, for the excellent 
facilities for working afforded me as well as for their kindness and assistance. 


332 E. J. WARBURG 


Prof. Niels Bjerrum has on numerous occasions discussed various physico- 
chemical problems with me and he has kindly read through chapters I, II, 
III and VIII. Without his aid it would hardly have been possible to write 
these sections, and I must express my deep gratitude to him. 


Lastly I am indebted to Dr E. E. Atkin of the Lister Institute, London, 
for the great care with which he has translated the work from the Danish. 


BIBLIOGRAPHY. 


The bibliography has been elaborated essentially with regard to the problem 
of the nature of the combination of carbon dioxide in the blood. 

The older literature dealing with allied subjects is collected’in H. J. 
Hamburger’s Osmotischer Druck und Ionenlehre etc., 1, Wiesbaden, 1902. 

The handbooks and larger monographs in the list are marked with an 
asterisk. Reference will be found in them to the more recent work on physico- 
chemical biology. The collection of the literature was practically completed 
in the first half of 1921. ; 


Abbott, G. A. and Bray, W. C. (1909). The Ionisation Relations of Ortho- and Pyro-phosphoric 
Acids and their Sodium Salts. J. Amer. Chem. Soc. 31, 729. 
Amsiph, E. F, (1920). The Liberation of CO, from Carbonate by Blood and Serum. J. Physiol. 54. 
roc. XXXiv. 


Arrhenius, 8. (1888). Theorie der isohydrischen Liésungen. Zeitsch. physikal. Chem. 2, 284. 
(1912). Theories of Solution. New Haven. 
(1915). Lehrbuch der Elektrochemie, 3 Aufl. Leipzig. 
Auerbach, F. and Pick, H. (1912). Die Alkalitit wasseriger Lésungen kohlensaurer Salze. 
Arbeiten aus dem K. Gesundheitsamte, 38, 243. ; 
*Bang, I. (1916). Methoden zur Mikrobestimmung einiger Blutbestandteile. Wiesbaden. 
Barcroft, J. (1913). The Combination of Haemoglobin with Oxygen and with Carbon Monoxide. 
Biochem. J.7, 481. 


Barr, D. P. and Peters, Jr, J. P. (1920-1921). The carbon dioxide absorption curve and carbon 
dioxide tensions of the blood in severe anaemia. J. Biol. Chem. 45, 571. 
Bayliss, W. M. (1909). The Osmotic Pressure of Congo-red, and some other Dyes. Proc. Roy. 
Soc. B, 81, 269. 
ae ae The Osmotic Pressure of electrolytically dissociated Colloids. Proc. Roy. Soc. B, 
. London. 
(1919). The Neutrality of the Blood, J. Physiol. 53, 162. 
*Beért, P. (1870). Legons sur la physiologie comparée de la respiration. Paris, 1-30, 66-152. 
(1878). La pression barométrique. Paris. 
ser gs Sur |’état dans lequel se trouve l’acide carbonique du sang et des tissues. Compt. 
‘end, 87. 
Beutner, R. (1920). Die Hntstehung elektrischer Strome in lebenden Geweben u.s.w. Stuttgart. 
Bie, V. and Méller, P. (1913), Undersogelser af normale Menneskers Blod, Ugeskrift for Lager. 
a N. (1909). A new Form for the electrolytic Dissociation Theory. 7th Internat. Congr. 
Appl. Chem. Sect. x, 58. 


* —— (1914). Die Theorie der alkalimetrischen und azidimetrischen Titrierungen. Stuttgart. 
(Samml. Ahrens.) 


——~ (1916), De sterke Elektrolyters Dissociation, Forh, ved 16 skand, Naturforskermede 1916, 
———~ (1918). Die Dissoziation der starken Elektrolyten. Zeitsch. Llektrochem. 24, 321. 
——~ (1919). On the Activity-coefficients for Ions. Medd. fran Nobelinst. 5. 


Se a Se ee ee eee a 


THEORY OF THE HENDERSON-HASSELBALCH EQUATION 333 


Bjerrum, N. and Gjaldbaek, J. K. (1919). Undersggelser over de Faktorer, som bestemmer Jord- 
bundens Egenskaber. I. Om Bestemmelser af en Jords sure eller basiske Egenskaber. 
Il. Om Reaktionen af Vesker som er mattede med Calciumcarbonat. Landboh¢gjskolens 
Aarsskr. (Deutscher Zusammenfassung: I. Uber die Bestimmung der sauren und basischen 
Figenschaften des Bodens. II. Uber die Reaktion von Fliissigkeiten die mit Calcium- 
karbonat gesittigt sind.) 

Blasel, L. and Matula, J. (1914). Unters. iiber physik. Zustandsinderungen der Kolloide. 
XVI. Versuche mit Desaminoglutin. Biochem. Zeitsch. 58, 417. 

Bohr, ©. (1886). Uber die Verbindung des Himoglobins mit Kohlensiure. Festschrift fiir 
Indwig. Leipzig, 164. : 

—— (1891). Beitrige zur Lehre von den Kohlensiureverbindungen des Blutes. Skand. Arch. 
Physiol. 3, 47. 

—— (1898). Uber die Verbindungen von Methimoglobin mit Kohlensiure. Skand. Arch. 
Physiol. 8, 363. 

——- (1904). Theoretische Behandlung der qualitativen Verhiltnisse der Kohlensiure- 
verbindung des Hiimoglobins, Centr. Physiol. H. 23. 

—— (1905). Absorptionscoefficienten des Blutes und des Blutplasmas fiir Gase. Skand. Arch. 
Physiol. 17, 104. 

—— (1909). Blutgase und respiratorischer Gaswechsel. Nagel’s Handb. d. Physiol. 1, 54-129. 
Braunschweig. 

—— and Bock, J. C. (1891). Bestimmung der Absorption einiger Gase in Wasser bei Tem- 
peraturen zwischen 0° und 100°. Wiedemann’s Ann. 44. 

—— Hasselbalch, K. A. and Krogh, A. (1904). Uber den Einfluss der Kohlensiureverbindung 
auf die Sauerstoffaufnahme im Blute. Centr. Physiol. H. 23. 

Kiger H. (1910). Blut und Lymfe. Nagel’s Handb. d. Physiol. Ergiinzungsbd. 1-78. Braun- 

schweig. 

Botazzi, F. (1911). Physikalisch-chemische Untersuchung des Harns und der anderen Kérper- 
fliissigkeiten. C. Neuberg: Der Harn u.s.w. LU. Berlin. 

Boyle, R., cit. Loewy (1870). Die Gase des Kérpers. 

Buckmaster, G. A. (1917, 1). The Relation of Carbon Dioxide in the Blood. J. Physiol. 51, 105. 

—— (1917, 2). On the Capacity of Blood and Haemoglobin to unite with Carbon Dioxide. 
J. Physiol. 51, 164. 
—— and Gardner, J. A. (1911-1912). The Nitrogen-content of Blood. J. Physiol. 48, 401. 
vias ao ag 8. and Liebermann, L. (1898). Uber das Bindungsvermégen eiweisshaltiger Kérper 
fiir Salzsiiure, Natriumhydroxyd und Kochsalz. Pfliiger’s Arch. 72, 51. 
*Burton, E. F. (1916). The physical Properties of colloidal Solutions. London. 

Campbell, J. M. H. and Poulton, E. P. (1920). The Relation of Oxyhaemoglobin to the CO 
of the Blood. J. Physiol. 54, 152. 

Cannon, W. B., Fraser, J. and age oe A. N. (1917). Some Alterations in the Distribution and 
Character of the Blood in Shock. Medical Research Committee. Wound Shock and 
Haemorrhage. London, 1919, 49. 

Christiansen, J., Seeger C. G. and Haldane, J. 8. (1914). The Absorption and Dissociation 
of Carbon Dioxide by human blood. J. Physiol. 48, 244. 

Clark, W. M. (1915). A Hydrogen Electrode Vessel. J. Biol. Chem. 23, 475. 

* — (1920). T'he determination of hydrogen ions. Baltimore, reprinted 1921. 
—— and Lubs, H. A. (1916). Hydrogen Electrode Potentials of Phthalate, Phosphate and 
Buffer Mixtures. J. Biol. Chem. 25, 479. 

Dale, H. H. and Lovatt Evans, C. (1920). Colorimetric Determination of Reaction of Blood by 
Dialysis. J. Physiol. 54, 168. 

Datta, A. K. and Dhar, N. (1915). A New Method of finding the Second Dissociation 
Constants of Dibasic Acids. J. Chem. Soc. 107, 824. 

Davies, H. W., Haldane, J. B. 8. and Kennaway, E. L. (1920). Experiments on the Regulation 
of the Blood’s Alkalinity. I. J. Physiol. 54, 32. 

Davy, H. (1803). (Edinburgh Med. and Surg. J. 29), cit. Henry, W. Gilbert's Ann. 591. 

ges ger J. F. and Parsons, T. R. (1919). Some Further Observations on Blood Reaction. 

. Physiol. 52. 

Donnan, F. G. (1911). Theorie der Membrangleichgewichte und Membranpotentiale bei Vor- 
handensein von nicht dialysierenden Elektrolyten. Ein Beitrag zur physikalisch-chemische 
Physiologie. Zeitsch. Elektrochem. 17, 572. 

—— and Allmand, A. J. (1914). Ionic Equilibrium across Semi-permeable Membranes. 
J. Chem. Soc. 105, 1941. 
*Edwards, W. F. (1824). De linfluence des agens physiques sur la vie. Paris. 


22—3 


334 E. J. WARBURG 


*Ege, R. (1919). Studier over Glukosens Fordeling mellem Plasmaet og de rgde Blodlegemer o.s.v. 
Kobenhavn. 


—— (1920, 1). Etudes sur la distribution du glucose entre le plasma et les globules rouges 
du sang et sur quelques problémes qui s’y rapportent. Compt. Rend. Soc. Biol. 88, 697. 

—— (1920, 2). Uher die Restreduktion des Blutes. Biochem. Zeitsch. 107, 229. 

—— (1920, 3). Zur Frage der Permeabilitét des Blutkérperchen gegeniiber Glucose und 
Anelektrolyten. Biochem. Zeitsch. 107, 246. 

—— (1920, 4). Uber die Bestimmung des Blutkérperchenvolumen. Biochem. Zeitsch. 109, 
242. 

—— (1921, 1). Die Verteilung der Glucose zwischen Plasma und roten Blutkérper. Biochem. 
Zeitsch. 111, 189. 

—— (1921, 2). Wie ist die Verteilung der Glucose zwischen den roten Blutkérperchen und 
den aiusseren Fliissigkeit zu erkliren? Biochem. Zeitsch. 114, 88. 


(1921, 3). Studien iiber das osmotische Verhaltniss der Blutkérperchen. I. Unter 
suchungen iiber das Volumen der Blutkérperchen in gegenseitig osmotischen Lésungen. 
Biochem. Zeitsch. 115, 109. 


—— (1921, 4). II. Der osmotischen Druck in Blutkérperchen und Plasma. Biochem. Zeitsch. 
115, 175. 


*Eichwald, E. and Fodor, A. (1919). Die physikalisch-kemischen Grundlagen der Biologie. Berlin. 
Enderlin, C. (1844). Physiologisch-chemische Untersuchungen. Ann. d. Chem. u. Pharm. 49, 
317; 50, 53. 
Evans, C. Lovatt (1921, 1). On a probable error in determinations by means of the hydrogen 
electrode. J. Physiol. 54, 353. 
(1921, 2). The regulation of the reaction of the blood. J. Physiol. 55, 159. 
Fahr, G. (1912). New Determinations of certain Absorption Coefficients in Blood-gas Investi- 
gations. J. Physiol. 48, 417. 
Fahraeus, R. (1918). Uber die Ursachen der verminderten Suspensionsstabilitit der Blut- 
kérperchen waihrend der Schwangerschaft. Biochem. Zeitsch. 89, 355. 


* (1921). The Suspension-stability of the Blood. I-IV. Acta medica scandinavica, 55, 
1—228. 


Falta, W. and Richter-Quittner, M. (1920). Uber die Verteilung des Zuckers, der Chloride und 


der Reststickstoffkérper auf Plasma und Kérperchen im strémenden Blute. Biochem. 
Zeitsch. 100, 148. 


—— (1921). Uber die chemische Zusammensetzung der Blutkérperchen. Biochem. Zeitsch. 
114, 144. 


Farkas (1903). Uber die Konzentration der Hydroxylionen im Blutserum. Pfliiger’s Arch. 98, 551. 


Fernet, E. (1855). Note sur la solubilité des gaz dans les dissolutions salines, pour servir & la 
théorie de la respiration. Compt. Rend. 41, 1237. 


—— (1857). Du réle des principaux éléments du sang dans l’absorption ou le dégagement 
des gaz de la respiration. Ann. Sc. Nat. (Zool.) 8, 125. 


Findlay, A. (1908). Einfluss von Kolloiden auf die Absorption von Gasen, insbesondere von 
Kohlendioxyd in Wasser. Kolloid. Zeitsch. 3, 169. 


(1913). Osmotic Pressure. London. 

. (1915). The Phase Rule and its Applications, London. 

and Creighton, H. J. (1909). The Influence of Colloids and Colloid Suspensions on the 
Solubility of Gases in Water. 7th Internat. Congr. Appl. Chem. Sect. x. 


(1910). The Influence of Colloids and Fine Suspensions on the Solubility of Gases 
in Water. I. Solubility of Carbon Dioxide and Nitrous Oxide. J. Chem. Soc. 97, 536. 


——~ (1911). Some Experiments on the Solubility of Gases in Ox Blood and Ox Serum. 
Biochem, J. 5, 294. 


——— and Shen, B. (1912). Il. The Solubility of Carbon Dioxide and Hydrogen, J. Chem. 
Soe, 101, 1459. 


—— and Wililams, T. (1913). III. Solubility of Carbon Dioxide at Pressures lower than atmo- 
spheric. J, Chem. Soc, 108, 637. 


and Howell, O. R. (1915). IV. The Solubility of Carbon Dioxide in Water in the Presence 
of Starch. J. Chem. Soc. 107, 282. 


*Frédérieq, L. (1878). Recherches sur la constitution du plasma sanguin, Gand. 
*Freundlich, H. (1909). Kapillarchemie. Leipzig. 
Fridericia, L. 8. (1920). Exchange of chloride Ions and of Carbon Dioxide between Blood 
Corpuscles and Blood Plasma, J. Biol. Chem. 42, 245. 
*Forster, F. (1915). Llektrochemie wasseriger Lésungen, 2. Aufl, Leipzig. 


Gad Andresen, K, L, (1920), Uber die Verteilung der Reststickstoffkérper auf Plasma und 
Kérperchen im strémenden Blute, Biochem, Zeitsch. 107, 250, 


= 


THEORY OF THE HENDERSON-HASSELBALCH EQUATION = 335 


eee? se _ Corral y (1914). La reaccién actual de la sangre y su determinacion electrometrica. 
: 

Gaule, J. (1878). Die Kohlensiiurespannung im Blut, im Serum und in der Lymphe. Arch. 
(Anat.) Physiol. 469. 

Gjaldbaek, J. K. (1921). Undersogelser over de Faktorer, som bestemmer Jordbundens Reaktion. 
III. Magniumkarbonatets forskellige Former, og om Reaktionen af Vesker,°som er 
mettede dermed. Landbohgjskolens Aarsskr. (Deutscher Zusammenfassung iiber die 
verschiedenen Formen des Magnesium-Karbonats und iiber die Reaktion von Flissigkeiten, 
die damit gesiittigt sind.) 

Gram, H. C. (1921). Studier over Fibrinmengden i Menneskets Blod og Plasma. With an English 
Summary. (Studies on the percentage of fibrin in human blood and plasma.) Kobenhavn. 

Gréhant, N. Les gaz du sang. Paris. 

Giirber, A. (1895, 1). Die Salze des Blutes. Verhl. physikal.-chem. Ges. Marburg, N.F. 28, 129. 

—— (1895, 2). Uber den Einfluss der Kohlensiure auf die Verteilung von Basen und Saéuren 
zwischen Serum und Blutkérperchen. Sitzwngsber. physikal.-med. Ges. Wiirzburg, 28. 

Hagedorn, H. C. (1920). Einige Bemerkungen iiber die Verteilung der Glucose zwischen Blut- 
kérperchen und Plasma. Biochem. Zeitsch. 107, 248. 

Haggard, H. W. and Henderson, Y. (1919). Hemato-respiratory Functions. I. The CO, Diagram 
of the Blood. II. Laws of Respiration. ITT. Heapisatory Decompensation and Acidosis. 
IV. The Cy? Scale and the Dissociation Characteristic. J. Biol. Chem. 39, 163. 

—— —— (1920,1). VII. The reversible Alteration of the H,CO, : NaHCO, Equilibrium in 
Blood and Plasma under Variations in the CO, Tension and their Mechanism. J. Biol. 
Chem. 45, 189. 

—— —— (1920, 2). VIII. The degree of Saturation of Corpuscles with HCl as a condition 
underlying the amount of Alkali called into use in the Plasma. J. Biol. Chem. 45, 199. 

—— —— (1920, 3). IX. An irreversible Alteration of the H,CO, : NaHCO, Equilibrium of 
Blood, induced by temporary Exposure to a low Tension of CO,. J. Biol. Chem. 45, 209. 

(1920, 4). X. The Variability of the reciprocal Action of Oxygen and CO, in Blood. 
J. Biol. Chem, 45, 215. 
Hamburger, H. J. (1892). Uber den Einfluss der Athmung auf die Permeabilitat der roten 
Blutkérperchen. Zeitsch, Biol. 28, 405. 
—— (1893). La constitution du sang veineux et du sang artériel. Arch. physiol. norm. et 
pathol. 8. v, 5, 1. 
(1902). Osmotischer Druck und Ionenlehre u.s.w. I. Wiesbaden. 

—— (1918). Anionenwanderungen in Serum und Blut unter dem Einfluss von CO,, Saure 
und Alkali. Biochem. Zeitsch, 86, 309. 

—— and Bubanovie, Fr. (1910). La perméabilité des globules rouges, spécialement vis-d-vis 
des cations. Arch. internat. Physiol. 10, 1. 

*Hammarsten, 0. (1914). Lehrbuch der physiol. Chemie. 8. Aufl., 805. Wiesbaden. 

Hardy, W. B. (1900, 1). Uber den Mechanismus der Erstarrung in umkehrbaren Kolloid- 
systeme. Zeitsch. physikal. Chem. 33, 326. 

—— (1900, 2). Eine vorliufige Untersuchung der Bedingungen welche die Stabilitét von 
nicht umkehrbaren Hydrosolen bestimmen. Zeitsch. physikal. Chem. 33, 385. 

Hasselbalch, K. A. (1910). Elektrometrische Reaktionsbestimmungen kohlensiure-haltiger 
Fliissigkeiten. Biochem. Zeitsch. 30, 317. 

—— (1911, 1). Elektrometrisk Reaktionbestemmelse af kulsyreholdige Vesker. Medd. fra 
Carlsberg Lab. 10, 64. 

—— (1911, 2). Détermination électrométrique de la réaction des liquides renfermant de l’acide 
carbonique. Compt. Rend. Carlsberg, 10, 69. 

yeah 1 poy 1). Die reduzierte und die regulierte Wasserstoffzahl des Blutes. Biochem. Zeitsch. 


——- (1916, 2). Die Berechnung der Wasserstoffzahl des Blutes aus der freien und gebundenen 
Kohlensiure desselben, und die Sauerstoffbindung des Blutes als Funktion der Wasser- 
stofizahl. Biochem. Zeitsch. '78, 112. 

—— and Lundsgaard, C. (1912). Elektrometrische Reaktionsbestimmungen des Blutes bei 
K6rpertemperaturen. Biochem. Zeitsch. 38, 77. 

—— and Gammeltoft, 8S. A. (1915). Die Neutralititsregulation des graviden Organismus. 
Biochem. Zeitsch. 68, 206. 

——- and Warburg, E. J. (1918). Ist die Kohlensiureverbindung des Blutserums als Mass fiir 
die Blutreaktion verwendbar? Biochem. Zeitsch. 86, 410. 

*Hatschek, E. (1916). An Introduction to the Physics and Chemistry of Colloids. London. 
Reeag= sae - (1915). Grundziige der physikalischen Chemie in ihrer Beziehung zur Biologie. 
iesbaden. 


336 E. J. WARBURG 


Heidenhain, R. and Meyer, L. (1863). Uber das Verhalten der Kohlensiure gegen Lésungen 
von phosphorsaurem Natron. Studien des physiol. Inst. zu Breslau, 2, 103. 

Henderson, L. J. (1908). A Note on the Union of the Proteins of Serum with Alkali. Amer. 
J. Physiol. 21, 168. 

—— (1908). Concerning the Relationship between the Strength of Acids and their same 
to preserve Neutrality. Amer. J. Physiol. 21, 172. 


—— (1908). The Theory of Neutrality Regulation in the animal Organism. Amer. J. Physiol: 
21, 427. 


(1909). Das Gleichgewicht zwischen Basen und Sauren im tierischen Organismus, Hrgebn. 
Physiol. 254. 


—— (1910). On the Neutrality Equilibrium in Blood and Protoplasm. J. Biol. Chem. 7, 29. 


—— (1920, 1). The Equilibrium between Oxygen and Carbonic Acid in the Blood. J. Biol. 
Chem. 44, 401. 


—— (1920, 2). Blood as a physico-chemical System. J. Bioch. Chem. 45, 411. 


—— and Black, O. F. (1908). Study of the Equilibrium between Carbonic Acid, Sodium 
Bicarbonate, Mono-sodium Phosphate and Di-sodium Phosphate at Body Temperature. 
Amer. J. Physiol. 21, 420. 


—— and Spiro, K. (1909). Zur Kenntniss des Ionengleichgewichts im Organismus. I. Biochem. 
Zeitsch. 15, 106 


Hermann, R. (1834). Uber die saure Beschaffenheit des venésen Menschenbluts und iiber den 
Unterschied zwischen arteriellen und venésen Blute. Poggendorff’s Ann. 1 (2), 311. 


Hill, A. V. (1910). The possible Effects of the Aggregation of the Molecules of Haemoglobin 
on its Dissociation Curves. J. Physiol. 40. Proc. iv. 


—— (1913). The Combination of Haemoglobin with Oxygen and with Carbon Monoxide. 
Biochem. J. 7, 471. 


——- (1921). The Combination of Haemoglobin with Oxygen and Carbon Monoxide, and the 
Effects of Acids and Carbon Dioxide. Biochem. J. 15, 577. 
Héber, R. (1900). Uber die Hydroxylionen des Blutes. I. Pfliiger’s Arch. 81, 522. 
(1903). IL. -Pfliiger’s Arch. 99, 572. 
—— (1904, 1). Resorption und Kataphorese. Pfliiger’s Arch. 101, 607, 627. 


—— (1904, 2). Weitere Mitteilungen iiber Ionenpermeabilitat bei Blutkérperchen. Pfliiger’s 
Arch, 102, ae 


* 


. OC. Oppenheimer’s Handb. der 
Bioch. II. 2. “Hilfte, 1. Leipeig 


‘abe G. (1901). Neue Versuche iiber die Dissociation das Mabe, Arch. (Anat.) 
Physiol. 187. 

Iversen, P. (1920). Untersuchungen iiber den “siurelislichen Phosphor” in Blut und Plasma 
bei verschiedenen Tieren, sowie einige Studien tiber die Toxikologie der verschiedenen 
Phosphate. Biochem. Zeitsch. 109, 211. 

Jahn, H. (1900). Uber den Dissociationsgrad und das Dissociationsgleichgewicht stark dis- 
socierter Elektrolyte. I. Zeitsch. physikal. Chem, 38, 544. 


Jaquet, A. (1892). Uber die Wirkung miissiger Siurezufuhr auf Kohlensiiuremenge, Kohlen- 
siurespannung und Alkalescenz des Blutes. Ein Beitrag zur Theorie der Respiration. 
Arch. exp. Path, Pharm, 30, 311. 

Jarlév, E. (1919). Om Syre-Baseligevagten i det menneskelige Blod serlig ved Sygdomme. 
Kobenhavn. 


Joffe, J. and Poulton, E. P. (1920). The Partition of CO, between Plasma and Corpuscles in 
oxygenated and reduced Blood. J. Physiol. 54, 129. 
Johnston, J. (1916). Determination of carbonic acid, combined and free, in solution, particu- 
larly in natural waters. J. Amer. Chem. Soc. 38, 947. 
Jolyet and Sigalas (1892). Sur l’azote du sang. Compt. Rend. 114, 686. 
Kendall, J. (1912). The Velocity of the Hydrogen Ion, and a general Dissociation Formula for 
Acids, J. Chem, Soc, 101, 1275. 
—— (1916). The specific Conductivity of pure Water in Equilibrium with atmospheric Carbon 
Dioxide. J. Amer, Chem. Soc. 38, 1480. 
Koeppe, H. (1897). Der osmotische Druck als Ursache des Stoffaustausches zwischen roten 
Blutkérperchen und Salzlésungen, Pfliiger’s Arch, 67, 189. 
—— (1903). Uber das Lackfarbenwerden der roten Blutscheiben, Pfliiger’s Arch. 99, 33. 
Konikoff, A. P. (1913), Ober die Bestimmung der wahren Blutreaktion mittels der elektrischen 
Methode. Biochem. Zeitech. 61, 200, 
*Kordnyi, A. v. and Richter, P. F. (1907-1908). Physikalische Chemie und Medizin, Leipzig. 


THEORY OF THE HENDERSON-HASSELBALCH EQUATION — 337 


*Kraus, F. (1898). Uber die Verteilung der Kohlensiure im Blute. Festsch. Univers. Graz. Graz. 
Krogh, A. and Liljestrand, G. (1921). Eine Mikromethode zur Bestimmung der Kohlensaure 
des Blutes. Biochem. Zeitsch. 104, 300. 
*Lehfeldt, R. A. (1908). Hlectro-chemistry. 1. London. 
Lehmann, ©. (1894). Untersuchung iiber die Alkalescenz des Blutes und speciel die Einwirkung 
der Kohlensiure daran. Pfliiger’s Arch. 58, 428. 
Levy, R. L., Rowntree, L. G. and Marriot, W. McK. (1915). A simple method for determining 
the Hydrogen-ion Concentration of the Blood. Arch. internal. Med. 16, 389. 
Lewis, G. N. (1908). Umriss eines neuen Systems der chemischen Thermodynamik. Zeitsch. 
physikal. Chem. 61, 129. 
*Lewis, W. C. McC. (1916). A System of Physical Chemistry. I-II. London. 
Limbeck, R. v. (1894). Uber den Einfluss des respiratorischen Gaswechsel auf die roten Blut- 
kérperchen. Arch. exp. Path. Pharm. 35, 309. 
Lindhard, J. (1914). Undersggelser over Hjcertets Minutvolumen i Hvile og under Muskelarbejde. 
Kobenhavn. 
—— (1921). Colorimetric determination of the concentration of hydrogen ions in very small 
quantities of blood by dialysis. Compt. Rend. Carlsberg, 14, No. 13. 
Linzenmeyer, G. (1920). Untersuchungen tiber die Senkungsgeschwindigkeit der roten Blut- 
kérperchen. I. Pfliiger’s Arch. 181, 169. 
—— (1921). Il. Pfliiger’s Arch. 186, 272. 
Loeb, J. (1921, 1). Ton series and the properties of proteins, I. J. General Physiol. 3, 85. 
—— (1921, 2). Il. J. General Physiol. 3, 247. 
—— (1921, 3). Ill. J. General Physiol. 3, 391. 
—— (1921, 4). Donnan equilibrium and the physical properties of proteins. I. Membrane 
potentials. J. General Physiol. 3, 667. 
—— (1921, 5). Il. Osmotic Pressure. J. General Physiol. 3, 667. 
—— (1921, 6). ILI. Viscosity. J. General Physiol. 3, 827. 
—— (1922). The origin of electrical charges of colloidal particles of living tissues. J. General 
Physiol. 4, 351. . 
Loewy, A. (1894). Untersuchungen zur Alkalescenz des Blutes. Pfliiger’s Arch. 58, 462. 
Spit og Die Gase des Kérpers. C. Oppenheimer’s Handb. d. Bioch. tv, 1. Halfte, 1-79. 
ipzig. 
—— and Zuntz, N. (1891). Uber die Bindung der Alkalien in Serum und Blutkérperchen. 
Phliiger’s Arch, 58, 511. 
—— —— (1892). Einige Beobachtungen iiber die Alkalescenzinderungen des frisch entleerten 
Blutes. Pfliiger’s Arch, 58, 507. 
Ludwig, C. (1865). Zusammenstellung der Untersuchungen iiber Blutgase. Separat-Abdruck 
aus den med. Jahrbiichern. Zeitsch. Ges. Aerzte Wien. i 
Beagrie H. (1908). Affinitdtsmessungen an schwachen Sduren und Basen. Stuttgart. (Samml. 
ns.) 


* 


Magnus, G. (1837). Uber die im Blute enthaltenen Gase, Sauerstoff, Stickstoff und Kohlensaure. 
Poggendorff’s Ann. 40, 583. 
‘Manube, K. and Matula, J. (1913). Untersuchung iiber physik. Zustandsinderungen d. Kolloide. 
XV. Elektrochem. Unters. am Siureeiweiss. Biochem. Zeitsch. 52, 369. 
Marchand, R. F, (1846). Ueber die Anwesenheit der kohlensauren Salze in dem Blute. J. pr. 
Chem. 37, 321. 
*McClendon, J. F. (1917). Physical Chemistry of Vital Phenomena, Princeton. 
—— Shedlov, A. and Thomson, W. (1917). Tables for finding the alkaline Reserve of Blood 
Serum in Health and Acidosis, from the total CO, or the alveolar CO, or the py- at known 
CO, Tension. J. Biol. Chem. 31, 519. 
McCoy, H. N. and Smith, H. J. (1911). Equilibrium between Alkali-earth Carbonates, Carbon 
Dioxide and Water. J. Amer. Chem. Soc. 33, 468. 


—— and Test, C. D. (1911). Equilibrium between Sodium Carbonate, Sodium Bicarbonate 
and Water. J. Amer. Chem. Soc. 38, 473. 


’ McLean, F. C., Murray, H. A. and Henderson, L. J. (1920). The variable Acidity of the Haemo- 
globin and the Distribution of Chloride in the Blood. Proc. Soc. Exper. Biol. and Med. 
17, 180. Quoted from Medical Sc. 5, 88, 1921. : 
Meyer, L. (1857). Die Gase des Blutes. Diss. Gottingen. 
Michaelis, L. (1914). Die Wasserstoffionenkonzentration. Berlin. 


—— (1920). Uber die Analyse der Kohlensiuregleichgewichte im Blute nach H. Straub und 
K. Meier. Biochem. Zeitsch. 103, 53. 


—— and Airila, Y. (1921). Die elektrische Ladung des Himoglobins. Biochem. Zeitsch. 118, 144. 


338 E. J. WARBURG 


Michaelis, L. and Bien, Z. (1914). Der isoelektrische Punkt des Kohlenoxydhimoglobins und 
des reduzierten Himoglobins. Biochem. Zeitsch. 67, 198. 
—— and Davidoff, W. (1912). Methodisches und sachliches zur elektrometrischen Bestim- 
mung der Blutalkalescenz. Biochem. Zeitsch. 46, 131. 


—— and Davidsohn, H. (1911). Der isoelektrische Punkt des genuinen und denaturierten 
Serumalbumins. Biochem. Zeitsch. 33, 456. 


—— and Mostynski, B. (1910). Der isoelektrische Punkt des Serumalbumins. Biochem 
Zeitsch. 24, 79. 

—— and Rona, P. (1909). Elektrochemische Alkalinitaéitsmessungen an Blut und Serum. 
Biochem. Zeitsch. 18, 317. 

— ress (1914). Die Dissoziationskonstante der Kohlensiure. Biochem. Zeitsch. 67, 
182. 


—— and Takahashi, D. (1910). Die isoelektrischen Konstanten der Blutkérperchenbestand- 
teile und ihre Beziehungen zur Siiurehimolyse. Biochem. Zeitsch. 29, 439. 

Milroy, T, H. (1917). The Reaction Regulator Mechanism of the Blood before and after _ 
Haemorrhage. J. Physiol. 51, 259. 

Misterlich, E., Gmelin, L. and Tiedemann, F, (1834). Versuche iiber das Blut. Poggendorff’s 
Ann. 1 (2), 289. : ’ 

Moscati, P. (1784). Osservazione ed esperienze sul sangue, fluido e rappreso; sopra l’azione 
dell’ arterie; e sui liquori, che bollono poco riscaldati nella machine pneumatica, 1783. 
Quoted from Crell’s Chem. Ann. 1, 91. 

Miiller, F. (1904). Uber die Ferricyanidmethode zur Bestimmung des Sauerstoffs im Blute ohne 
Blutgaspumpe. Pfliiger’s Arch. 103, 541. 

Nasse, H. (1878). Untersuchungen iiber den Austritt und Eintritt von Stoffen (Transudation und 
Diffusion) durch die Wand der Haargefiisse. Pfliiger’s Arch. 16, 604. 

*Nernst, W. (1913). Theoretische Chemie. 7. Aufl. Stuttgart. 

Norgaard, A. and Gram, H. C. (1921). Relation between the Chloride Content of the Blood and 
its percentic Volume of Cells.. J. Biol. Chem. 49, 264. 

Oryng, T. and Pauli, W. (1915). Untersuchungen iiber physikal. Zustandsiinderungen der 
olloide. XIX. Uber Neutralsalzeiweissverbindungen. Biochem. Zeitsch. 70, 368. 
Ostwald, W. (1890). Elektrische Eigenschaften halbdurchlissiger Scheidewande. Zeitsch. 

physikal. Chem. 6, 71. 
* (1911). Grundriss der Kolloidchemie, t. Dresden. 


*Ostwald and Luther. Hand- und Hiilfsbuch zur Ausfiihrung physiko-chemischer Messungen. 
3. Aufl. bei R. Luther und K. Drucker. Leipzig. 


Parsons, T. R. (1917). On the Reaction of the Blood in the Body. J. Physiol. 51, 440. 


—— (1919). The Reaction and Carbon Dioxide carrying Power of Blood—a mathematical 
Treatment. L J. Physiol. 53, 42. 


—— (1920). IL. J. Physiol. 53, 340. 


—— and Parsons, W. (1919-1920). Hydrogen Ion Measurements on Blood in the neighbour- 
hood of the Isoelectric Point of Haemoglobin. J. Physiol. 53. Proc. c. 

—— (1922). The relations of Carbon Dioxide in Acidified Blood. J. Physiol. 56, 1. 

*Pauli, W. (1920). Kolloidchemie der Eiweisskérper. I. Dresden and Leipzig. 

and Matula, J. (1917). Untersuchungen iiber physikalische Zustandsinderungen der 
Kolloide. XXI. Uber Silbersalzproteine. Biochem. Zeitsch. 80, 187. 

Peters, R. A. (1914, 1). A combined Ionometer and Electrode Cell for measuring the H ion con- 
centration of reduced blood at a given tension of CO,. J. Physiol. 48. Proc. vii. 

(1914, 2). Quoted from Bareroft. The respiratory Function of the Blood. Appendix ITT. 


Peters, Jr, J. P. and Barr, D. P. (1921). The Carbon Dioxide Absorption Curve and Carbon 
Dioxide Tension of the Blood in Cardiac Dyspnoea, J. Biol. Chem. 45, 537. 

and Rule, F. D. (1921). The Carbon Dioxide Absorption Curve and Carbon Dioxide 
Tension of the Blood of normal resting Individuals, I, J. Biol. Chem. 45, 489. 

Petry, E. (1903). Uber die Verteilung der Kohlensiiure im Blute. Hoffmeister’s Beitriige, 3, 1. 

Pfliiger, E. F. W. (1864), Uber die Kohlensture des Blutes, Bonn, 


Philip, J. C. and Bramley, A. (1915). Influence of Sucrose and Alkali Chlorides on the solvent 
Pani of Water. J. Chem. Soc, 107, 377. 


Preyer, W. (1866). Uber die Kohlensiiure und den Sauerstoff im Blute, Centr. med. Wiss. 321. 
———. (1867), Beitriige zur Kenntniss des Blutfarbstoffs. Centr, med. Wiss, 273. 
* _... (1871). Die Blutkrystalle. Jena. 
Procter, H, R. (1914). The Equilibrium of dilute Hydrochloric Acids and Gelatin. J. Chem. Soc, 
105, 313. 
—— and Wilson, J. A. (1916), The Acid-Gelatin Equilibrium, J. Chem. Soc, 109, 307. 


THEORY OF THE HENDERSON-HASSELBALCH EQUATION 339 


Robertson, T. B. (1909). On the Nature of the Chemical Mechanism which maintains the 
Neutrality of Tissue-fluids. J. Biol. Chem. 6, 313. 
—— (1910). Concerning the Relative Magnitude of the Parts played by the Proteins 
and by the Bicarbonate in the Maintenance of the Neutrality of the Blood. J. Biol. Chem. 
7, 351. 
(1908, 1909, 1910). Note on the Applicability of the Laws of amphoteric Electrolytes to 
Serum Globulin. J. Biol. Chem. 5, 155; 6, 313; 7, 351. 
* —_ (1912). Die physikalische Chemie der Proteine. Dresden. 
Rona, P. (1910). Uber das Verhalten des Chlors im Serum. Biochem. Zeitsch. 29, 501. 
and Gyérgy, P. (1913, 1). Uber das Natrium und das Carbonation im Serum. Beitrag 
zur Frage des nicht diffusiblen Alkalis im Serum. Biochem. Zeitsch. 48, 278. 
—— (1913, 2). Beitrag zur Frage der Ionenverteilung im Blutserum. Biochem. Zeitsch. 


ee Paar ocd Uber die Verbindungen der Alkalien mit Kohlensiure. Poggendorff’s Ann. 
*Rothmund, V. (1907). Léslichkeit und Léslichkeitsbeeinflussung. Leipzig. 
Schmidt, A. (1867). Uber die Kohlensiure in der Blutkérperchen. Ber. k. Sachs. Ges. Wiss. 
Math.-physikal. Cl. I. 19, 30. 
Schéffer, A. (1861). Die Kohlensiiure des Blutes und ihre Ausscheidung mittels der Lunge. 
Sitzungsber. k. Akad. Wiss. Wien, 41, 589. 
—— (1868). Die Kohlensiure des Blutes. Centr. med. Wiss. 657. 
Scudamore, C. (1826). Versuche mit dem Blut, 84-90. Wiirzburg. Quoted from Zuntz, Nagel’s 
Handbuch, . 
Sertoli, E. (1868). Uber die Bindung der Kohlensiure im Blute und ihre Ausscheidung in der 
Lunge. Hoppe-Seyler’s med.-chem. Unters. 3, 350. 
Setschenow, I. (1859). Beitrige zur Pneumatologie des*Blutes. Sitzwngsber. k. Akad. Wiss. 
Wien, 36, 293. 
—— (1879). Uber die Konstitution der Salzlésungen auf Grund ihres Verhalten zur Kohlen- 
siure. Nouv. mém. Soc. Imp. sc. nat. de Moscou, 15, 199. 
Seyler, C. A. and Lloyd, P. V. (1917, 1). Studies of Carbonates. II. Hydrolysis of Sodium 
elie a and Bicarbonate and the Ionisation Constants of Carbonic Acid. J. Chem. Soc. 
—— —— (1917, 2). Studies in Carbonates. III. Lithium, Calcium and Magnesium Car- 
bonates. J. Chem. Soc. 111, 994, 
Shields, J. (1893). Uber Hydrolyse in wiisserigen Salzlésungen. Zeitsch. physikal. Chem. 12, 166. 
Siebeck, R. (1909). Uber die Aufnahme von Stickoxydul im Blut. Skand. Arch. Physiol. 21, 368. 
Siegfried, M. (1905, 1). Uber die Bindung von Kohlensiure durch amphotere Amidokérper. 
I. Zeitsch. physiol. Chem. 44, 85. 
—— (1905, 2). Il. Zettsch. physiol. Chem. 46, 401. 
—— and Neumann, C. (1908). III. Zettsch. physiol. Chem. 54, 423. 
—— and Libermann, H. (1908). IV. Zeitsch. physiol. Chem. 54, 436. 
* —_ (1910). Der Carbamino- und Hydroxylkohlensiure-reaktion. Ergebnisse Physiol. 9, 334. 
Slyke, D. D. van (1921, 1): The normal and abnormal variations of the acid-base balance of 
the blood. J. Biol. Chem. 48, 153. 
—— (1921, 2). The Carbon Dioxide Carriers of the Blood. Physiol. Rev. 1, 141. 
— and Cullen, G. E. (1917). Studies of Acidosis. I. The Bicarbonate Concentration in the 
Blood Plasma, and its Determination as a Measure of Acidosis. J. Biol. Chem. 30, 289. 
Smith, L. W., Means, J. H. and Woodwell, M. N. (1920-1921). Studies on the distribution of 
carbon dioxide between cells and plasma. J. Biol. Chem. 45, 245. 
Sérensen, S. P. L. (1909-1910). Etudes enzymatiques. II. Sur la mesure et l’impotence de la 
concentration des ions hydrogénes dans les réactions enzymatiques. Compt. Rend. Carlsberg, 8. 
(1912). Uber die Messung und Bedeutung der Wasserstoffionenconcentration bei biolo- 
gischen Prozessen. Ergebnisse Physiol. 12, 393. 
—— (1915-1917). Studies on Proteins. Compt. Rend. Carlsberg, 15. 
Spiro, K. and Henderson, L. J. (1909). Zur Kenntniss des Ionengleichgewichts im Organismus. 
I, Einfluss der Kohlensaiure auf die Verteilung von Elektrofyten zwischen Blutkérperchen 
und Plasma. Biochem. Zeitsch. 15, 114. 
Stevens, W. (1832). Observations on the healthy and diseased Properties of the Blood. London. 
—— (1835). Observations on the Theory of Respiration. Communicated by W. T. Brande. 
Phil. Trans. 345. 
Stieglitz, J. (1909). The Relation of Equilibrium between the Carbon Dioxide of the Atmosphere 
and the Calcium Sulphate, Calcium Carbonate and Calcium Bicarbonate in contact with it. 
Carnegie Inst. Publ. 107, 233. 


* 


340 E. J. WARBURG 


Straub, H. and Meier, K. (1918, 1). Blutgasanalysen. I. Qualitativer und quantitativer 
Nachweis von Sauren in kleinen Blutmengen durch Bestimmung von Verteilungsgleichge- 
wichten. Biochem. Zeitsch. 89, 156. 

(1918, 2). II. Hamoglobin als Indicator. Ein Beitrag zur Theorie der Indicatoren. 

Biochem. Zeitsch. 90, 305. 

(1920, 1). III. Die Chlorionenpermeabilitaét menschlicher Erytrocyten. Biochem. 

Zeitsch. 98, 205. 

(1920, 2). IV. Der Einfluss der Alkalikationen auf Himoglobin und Zellmembran. 

Biochem. Zeitsch. 98, 228. 

(1920, 3). V. Der Einfluss der Erdalkalien auf Himoglobin und Zellkolloide. 
Biochem. Zeitsch. 109, 47. 

Stromeyer, E. C. F. (1832). Ist wirklich freie Siure im Blute? Schweiger’s J. tv, 4, 95. 

Thiel, A. and Strohecker, R. (1914). Uber die wahre Stiirke der Kohlensiure. Ber. 47 (1), 945. 

Thorup, 8. (1887). Om Blodets CO,-Binding. Kobenhavn. 

Tobiesen, F. (1895, 1). Bidrag til Leren om Blodets Rolle ved Vavenes respiratoriske Stofskifte. 
Kobenhavn. 

(1895, 2). Uber den spezifischen Sauerstoffgehalt des Blutes. Skand. Arch. Physiol. 6. 

Vogel, A. (1814). Versuche mit Urin. I. Uber die Existenz der Kohlensiure im Urin und im 
Blute. Schweiger’s J. 9, 399. ; 


Walbum, L. E. (1914). Die Bedeutung der Wasserstoffionenkonzentration fiir die Hamolyse. 
Biochem. Zeitsch. 63, 221. 


Walker, J. and Cormack, W. (1900). The Dissociation of very weak Acids. J. Chem. Soc. 77, 5. 
Walpole, G. 8. (1913). Gas-electrode for general use. Biochem. J. 7, 410. 
Warburg, E. J. (1920). Einige Bemerkungen iiber die Verteilung von Anionen zwischen Blut- 
kérperchen und Plasma. Biochem. Zeitsch. 107, 252. 
*Weimarn, P. P. von (1911). Grundziige der Dispersoidchemie. Dresden. 
*Williams, C. J. B. (1834-1835). Observations on the Changes produced in the Blood in the 
Course of its Circulation. With Experiments. Read before the R. Med. Soc. of Edinburgh 
1823, ete. London Med. Gazette, 16, 718, 724, 745-751, 783-788, 807-813, 842-848, 871-876. 


*Willstatter, R. and Stoll, A. (1918). Untersuchungen iiber die Assimilation der Kohlenséure, 
182-185. Berlin. 


*Zangger, H. (1908). Uber Membranen und Membranenfunktionen. Ergebnisse Physiol. 7, 99. 
*Zigmondy, R. (1918). Kolloidchemie. 2. Aufl. Leipzig. 


Zuntz, N. (1867). Uber den Einfluss des Partialdrucks der Koblensiure auf die Verteilung dieses 
Gases im Blute. Centr. med. Wiss. 529. 


—— (1868). Beitrdge zur Physiologie des Blutes. Diss. Bonn. 


(1882). Physiologie der Blutgase und des respiratorischen Gaswechsels. Hermann’s 
Handb. Physiol. Leipzig. 


* 


XXIII. THE EFFECT OF COLD STORAGE ON 
- THE CARNOSINE CONTENT OF MUSCLE. 


By WINIFRED MARY CLIFFORD. 


From the Physiology Department, Household and Social Science 
Department, Kensington. 


(Received February Ist, 1922.) 


TuE flavour of meat has long been believed to be associated with the presence 
of extractives, although whenever isolated in a pure condition such substances 
are tasteless. However one extractive—f-alanyl-histidine or carnosine—does 
disappear in cold storage and this may account for the inferior flavour of 
imported meat when compared with home-killed. 

The idea of such a disappearance was suggested by differences found when 
estimating carnosine in various samples of beef, by a colorimetric method 
previously described in this journal [Clifford, 1921]. Research on this ex- 
tractive was begun just after the war when the beef commonly available was 
imported. Such samples gave yields of 0-35-0-37 °% carnosine. On one occa- 
sion some English steak was obtained and gave the unexpectedly large yield 
of 1-1 %. It has been shown that for a given species the carnosine content 
of muscle is constant [Clifford, 1922] and the great difference between English 
and imported beef could not be explained unless cold storage had effected 
a reduction of the base. In order to test this various samples of English and 
imported meats were analysed. 


CARNOSINE CONTENT OF ENGLISH AND IMPORTED MEAT. 


Beef. 

Twelve separate samples of English beef have been examined over a period 
of two years with the following results: 1-1, 1-0, 1-1, 0-98, 0-96, 1-0, 0-97, 
0-98, 1-1, 0-98, 0-98, 0-97 %. Average 1:0 %. 

Five samples of imported beef gave: 0-37, 0-34, 0-36, 0-35, 0-36 %. Average 
0356 %. 

Similar experiments have been carried out on veal, mutton and lamb and 
in every case the carnosine content of imported meat has been very much 
lower than that of English meat. Actual figures obtained were: 


English Imported 
Beef 0-96-1-1 % (12 samples) 0-34-0-37 % (5 samples) 
Veal 1-05-1-12 (Oy 3.) 0364036 (4 ,, .) 
Mutton 0-37-0-38 Perish Se 1G. (2. 3) 
Lamb 0-40-0-42 eee ye O16 (2 yy |) 


Tn all these experiments about one-third of the carnosine value was found 
in the imported meats as compared with English killed. 


Bioch. xvr 23 


342 W. M. CLIFFORD 


EXPERIMENTS ON THE EFFECT OF COLD STORAGE ON MEAT. 


Since all the samples of imported meat were perfectly fresh and in an 
edible condition, the loss of carnosine could not be a putrefactive change, 
but might have been due to the length of time the meat had been in cold 
storage. 

It was possible to test this idea owing to the kindness of Dr H. H. Dale 
who allowed a sample of English meat to be placed in a cold room in the 
Research Institute at Hampstead. Portions were removed at various times 
for analysis. When first put in, carnosine was present to the normal value 
of 0-99 %. Samples were taken at-intervals of about one month. A fall of 
carnosine at first great, then slight, followed by a second steep fall was shown 
(Fig. 1). 


0-10g—2 
0-091-0-9 
0-08}-0:8 

0-07+-0-7 00:07 
0-06}-0-6 


0-05+0-5 C6 


0-04+-0-4 


003}-0°3 


0:02+-0-2 


0-01-01 


The actual figures were: 


Original 0-99 % carnosine 4 months 0-50 % carnosine 
lmonth 0-90 rp 6 e 0-48 fe 
2 months 0-55 rm BE gs 0-40 a 


5) ” 0-50 ” 9} ” 0°35 ” 


CARNOSINE IN MUSCLE DURING COLD STORAGE — 343 


EFFECT OF COLD STORAGE ON RAT MUSCLE. 


The loss of carnosine in beef might have originated in some process which 
took place before the meat was put into the cold room. Therefore seven rats 
were killed with coal gas and put into cold storage immediately after death. 
The normal value of 0-11 % carnosine was given by one animal, estimated 
directly, and the others were removed at intervals. The curve of carnosine 
loss ran parallel with that for beef for a period of four months (Fig. 1). This 
was followed by a steep fall in the fifth month and by six and a half months 
all trace of the base had disappeared. 

Results were: 


Original 0-11 % carnosine 64 months Absent 
l month 0-11 “ (i ease 99 

2 months 0-07 = 8} ,, as 

4  ,,. 0-052 is 


The above experiment shows that loss of carnosine occurs in muscle put 
into a cold room directly after the death of the animal in a similar manner to 
that shown by bought English meat kept at the same temperature. 

The temperature of the cold room was just below 0° C. Spicules of ice 
were found in every sample analysed and the extractions were made before 
thawing out had completed. 


DISCUSSION OF RESULTS. 


From experimental results it may be stated that measurement of the depth 
of red colour produced on diazotising carnosine in a watery muscle extract 
gives an easy and rapid test for distinguishing between English fresh killed 
and cold storage meats. A sample of beef or veal with a carnosine percentage 
below 0-8 % or of mutton or lamb below 0-3 °% would not come from a freshly 
killed animal, but probably from a carcase which had been chilled in order 
to preserve it. If the percentage were as low as 0-3 % in the case of beef or 
0-15 % with mutton the sample would be 9-12 months old. Further experi- 
ments are in progress with the object of obtaining a curve by which the age 
of cold storage meat may be read off after determining the carnosine content. 

The mechanism causing the loss of carnosine is unknown, but experiments are 
being carried out to investigate the problem. Since the reaction takes place at 
freezing point it is improbable that bacterial or enzyme action can account 
for the change, which is more likely to be initiated by a simpler type of catalyst. 


Thanks are due to Dr H. H. Dale for allowing the meat to be deposited 
in the cold room at the Research Institute, Hampstead, and to Prof. V. H. 
Mottram for suggestions during the course of the work. 

The expenses of this research were defrayed by the Medical Research 


Council. 
REFERENCES. 


Clifford (1921). Biochem. J. 15, 400. 
—— (1921). Biochem. J. 15, 725. 


23—2 


XXIV. INVESTIGATIONS ON THE NITROGENOUS 
METABOLISM OF THE HIGHER PLANTS. 


PART II. 


THE DISTRIBUTION OF NITROGEN IN THE LEAVES 
OF THE RUNNER BEAN’, 


By ALBERT CHARLES CHIBNALL. 


From the Biochemical Department, Imperial College 
of Science and Technology. 


(Received February Ist, 1922.) 


In the first part of this series [Chibnall and Schryver, 1921] a method of 
isolating part of the protein complex of fresh leaves was described. Subse- 
quent work carried on with the same object, mentioned in that paper, in view, 
namely that of throwing light on the protein metabolism in the leaf, showed 
that it was necessary to investigate two problems in detail. Firstly the nature 
of the nitrogen retained in the residue of cellular matter after applying the 
above method remained to be determined, and secondly it was necessary to 
find out how far it is justifiable to compare leaves picked at different, and even 
unknown, periods of growth. 

The present research on the leaves of the runner bean was primarily under- 
taken to solve these two problems. A method has been evolved whereby the 
whole of the nitrogenous material in the leaf is obtained in a state suitable 
for subsequent analysis. The water-soluble products are freed from all protein 
substances other than proteoses, about three-fourths of the protein N of 
the leaf is isolated as an amorphous powder, whilst the residue of the cell 
matter, containing 5-15 % of the total leaf N, is shown to contain only 
protein N. This makes it possible to estimate the total protein and water- 
soluble N in the leaf. 

By applying this method at frequent intervals during the life history of 
the runner bean leaf, an attempt has been made to solve the second problem 
enumerated above. Not only has the ratio protein N to water-soluble N been 
found, but by suitable analysis either before or after hydrolysis, the distri- 
bution of the protein N in five groups, and of the water-soluble N in seven 
groups, has been determined. 

A study of the protein groups indicates that the protein undergoes little, 
if any, change during the life history of the leaf, a fact that should be of 
assistance in future research into the chemistry of the leaf proteins. The 
water-soluble groups however are continually varying, indicating that com- 
parisons of the composition of leaves, unless regard is paid to exact conditions 

' Thesis approved for the Degree of Doctor of Philosophy in the University of London. 


‘tad 
‘ta: *. 


NITROGEN OF THE LEAVES OF THE RUNNER BEAN 345 


of age and growth, is almost impossible [compare Jodidi, Kellogg and True, 
1920}. 

As the research progressed it was found possible to extend the scope of 
enquiry to the diurnal variations, and the effects of starvation, the results of 
which throw considerable light on the protein metabolism in the leaf. The 
synthesis of protein in the leaf from nitrates appears to take place through 
the amino acids, whilst the products of protein degradation, in striking analogy 
to what is found in the animal kingdom, pass into what appear to be urea 
derivatives, 

A certain amount of research has been performed in the past to throw 
light on most of the problems discussed in this paper [see Czapek, 1920], but 
it was nearly all carried out before the development of modern methods for 
estimating the nitrogenous bases and monoamino acids, so that comparison 
of results is of very little value. 


I. GENERAL EXPERIMENTAL METHODS. 


Materials used. The plant used was Phaseolus vulgaris v. multifloris 
(Scarlet Champion). The seeds were planted in unmanured ground following 
potatoes on the 28th April, 1921, at the Physic Gardens, Chelsea. 

The seedlings appeared above the ground about 20 days later, and all ages 
of plants stated henceforth are calculated from the 18th May. The growth 
was normal, flowers appearing when the plant was five to six weeks old. At 
this time a few aphides appeared, but attention prevented them spreading, 
and all evidence of the disease had vanished three weeks later. By the middle 
of July it was found that although the inflorescence was normal, very few of 
the flowers had set, with the result that the number of pods formed was small, 
not averaging more than three or four per plant. This failure must be attri- 
buted to the abnormal dryness and high temperature of the atmosphere and 
not to the drought experienced during July, August and September, since the 
plants received abundance of water daily. 

Methods of sampling. The procedure adopted was to pick the leaves from 
the plants selected, go lightly over the surface with a brush, remove all stems 
and take the total weight. It is from this weight, obtained before any appre- 
ciable evaporation from the surface of the leaves could have taken place, that 
the “N per fresh weight of the leaves,” and hence all the figures given later 
in Tables XI and XIII, are calculated. The leaves were then counted, sorted 
out into six pans according to size, the content of each pan weighed, and the 
sample for grinding (200-250 g.) taken proportionately from each. By this 
means an average sample, giving a mean weight per leaf similar to that of 
the whole batch, was obtained. From several of the batches duplicates were 
taken to test the accuracy of the sampling and experimental methods used. 
The remainder of the batch was air-dried in an oven at 37° through which 
a slow stream of air was drawn. | 

Each batch picked was given a series number, which is retained throughout 


346 A. C. CHIBNALL 


in all subsequent operations. Details of the pickings, with the weather con- 
ditions, are given in Tables I and II. On account of the exceptionally fine 
weather conditions throughout the experiment the leaves were picked in a 
normal state without excess of moisture. Furthermore, the absence of high 
wind and storms kept the surface of the leaves clean and free from soil material. 
Only in one case, that of series 4, did rain fall in the preceding two or three 
days. In this rain fell during the night, but when the sample was taken in the 
morning no excess of moisture, as shown by lightly pressing between filter- 
papers, was found. Fresh weight determinations then can be considered as 
exceptionally good. 3 

Dry weight and total nitrogen of the leaves. The air-dried remainder was 
used for this purpose. The dry weight was taken from a small sample placed 
for 24 hours in an oven heated to 108°. | 

The total N was determined by the modified Kjeldahl-Gunning method, 
whereby the nitrate N is first converted into amino-phenol by salicylic acid 
and sodium hyposulphite. The practical details followed were those given by 
Moore [1920]. 

General methods of manipulation (separation of protein, soluble products and 
cellular matter). Two methods which aim at this separation have so far been 
described, that of Osborne and Wakeman [1920] and the one communicated 
in the first paper on these investigations [Chibnall and Schryver, 1921]. In 
the former the leaves were passed through a Nixtamel mill three times; the 
pulp obtained mixed with water and returned twice more to the mill. The 
result was a green solution containing protein in colloidal suspension, from 
which the cellular matter was removed by centrifuging. The protein was 
afterwards flocculated by the addition of alcohol. 

Briefly the second method consists of treating the minced leaves with 
ether-water and pressing out through thick muslin in a tincture press, whereby 
a green solution, containing most if not all of the water-soluble matter and 
part of the protein in colloidal suspension, was obtained. The protein was 
afterwards flocculated by warming. For the purpose of the present research 
this method, although easy to manipulate and rapid, was of little use unless 
the nitrogenous matter of the residue, about one-third of the total, could be 
separated from the cell matter for further investigation. 

Osborne’s method was certainly more thorough, the residue in this case 
containing only 18 % of the total N. It was inconvenient for the present 
purpose however, as it would have required the continuous use of a centrifuge. 
It was therefore decided to see if the ether-water method could not be improved. 

Some cabbage leaves were accordingly treated by this method and the 
colloidal solution and residue examined microscopically, using a }th objective. 

The solution appeared colourless, with two distinct sets of particles. The 
first, deep green in colour and irregular in shape, were obviously fragments 
of disintegrated chloroplasts, and were large enough to be free from move- 
ment. The second were very numerous, showed no trace of green colour, and 


NITROGEN OF THE LEAVES OF THE RUNNER BEAN 347 


were small enough to be in rapid Brownian movement. On warming to coagulate 
the protein the particles were all removed, and the coagulate appeared as a 
greyish granular mass in which the fragments of the green chloroplasts could 
be distinctly seen. 

The residue to the naked eye showed large disintegrated lumps of leaf 
matter. One of the smaller of these was examined under the microscope. The 
outer layer of cells had their walls broken and the major part of the content 
expelled; those inside were unruptured but flattened. In addition the vacuole 
content of the latter had been largely expressed, even though the layer of 
protoplasm was unbroken. 

It seemed clear from this examination that the rupturing of some of the 
cells was due to the action of the cutter in the mincing machine, and that 
the function of the ether-water was merely to kill the semi-permeable mem- 
brane of protoplasm in the remainder, so that the fluid content could be 
squeezed out by the press. The large quantity of N remaining in the residue 
therefore was due to the protoplasmic membrane remaining in the unruptured 
cells. Clearly then to reduce the N content of the residue it was necessary to 
increase the number of cells ruptured. 

No access could be had to a mill of the type used by Osborne. The only 
one available was that known technically as an End-runner Mill, which is 
nothing more than a mechanical pestle and mortar. It was found that if 
200 g. of cabbage leaves were broken up and fed slowly into this, then 100 cc. 
of water added, the whole was reduced to a very fine green pulp in about 
20 minutes. On pressing through muslin a solution, deeper coloured than 
before, was obtained. The grinding and pressing was repeated on the residue, 
and the N in the two colloidal extracts and final residue determined. The 
first extract contained 74-2 %, the second 14:8 %, and the residue 11 % of 
the total N. , 

Under the microscope the solution appeared much as before, except that 
the fragments of green chloroplasts were much more numerous and irregular 
in size, ranging from quite coarse lumps, which appeared however free from 
cell matter, down to small, but distinctly green, particles in Brownian move- 
ment. : 

The residue appeared different. The large masses of leafy matter had 
disappeared. Lumps of cell matter however could still be seen under the 
microscope, though the number of unruptured cells had greatly diminished. 
In addition numerous shreds of broken cells, containing only infrequent 
portions of protoplasm, were to be seen. 

It is on this property—that the cellular matter is not ground down to a 
powder (but only ruptured and torn so as to set free the cell content)—that 
the success of the method, which was used throughout the present research, 
depends. The thick muslin and the tincture press were found to be quite 
unnecessary. All that is required is a glass funnel opening to a diameter of 
about 8 inches and a piece of fine cotton lawn 18 inches square. The whole 


348 A. C. CHIBNALL 


of the liquid, with its heavy charge of colloidal matter, is expressed by very 
slight hand pressure in under a minute, leaving behind a pasty mass that 
can be easily removed from the lawn and returned to the mill for further 
grinding. The fact that this fine lawn can be used renders the method appli- 
cable for quantitative work, since none of the cell matter or protein can be 
retained in its pores, and the amount of water-soluble N absorbed is negligible. 

Reference to Table IV will show how complete a separation was obtained 
in the present research. Six grindings of the samples (200-250 g.) were given 
and the colloidal solution, as a safeguard against the pores of the lawn being 
enlarged by the passage of the liquid under slight pressure, was run through 
a second piece spread out over another funnel. 

In several of the intermediate batches ether-water was used in place of 
distilled water, but no benefit was experienced and it was later discarded. 
In case some of the cells should have escaped disruption, and thereby still be 
retaining their water-soluble contents, the residue from the sixth grinding 
was thrown into 200 cc. of boiling water. The cellular matter at once swelled 
up, imbibing the water, and on cooling this was squeezed through the same 
lawn as before. The operation was then repeated. In the earlier workings, 
when the total N in the residue was low, very little N passed into these two 
extracts. In the later however, when the residual N rose to 12 % and over, 
sufficient to make an appreciable difference was found in the first, though a 
negligible quantity was always found in the second. The protein was floccu- 
lated in the first six series by warming to 45°, but the temperature was after- 
wards raised to 60°, since it was found that the precipitate was then rendered 
almost granular, and could be readily filtered in a few minutes, whereas before 
it had taken about an hour. The filtering was accomplished by means of a 
Soxhlet extractor thimble standing upright in a glass funnel. After all the 
liquid had drained through, the retained protein was washed with two extracts 
from the boiled residue described above. After draining a second time the 
thimble was then transferred to an extractor described by Schryver [1908] 
and washed for three or four hours with 95 % alcohol. As first filtered off 
the protein is a pasty mass, part of which clings to the sides of the thimble, 
but the majority collects in a large button at the bottom. As the alcohol 
washes it the surface becomes dehydrated and shrinks, thereby coming away 
from the wall of the thimble. The whole can then be shaken out and broken, 
so as to allow the alcohol access to a large surface and thus hasten the 
washing. It was returned to the thimble and after a further three or four 
hours in this extractor the thimble was removed and stood overnight in 
absolute alcohol. In the morning it was placed in a Soxhlet apparatus and 
extracted with ether until all colouring matter was removed. The protein 
was then shaken out of the thimble, ground up in a small mortar, returned 
to the thimble and left 24 hours to dry, The weight of the thimble was deter- 
mined before the operation, and the increase when it contained the air-dried 
protein was taken as the weight of the latter. 


NITROGEN OF THE LEAVES OF THE RUNNER BEAN 349 


In the previous communication by Chibnall and Schryver [1921] it was 
found that after the colloidal protein of cabbage was flocculated by warming, 
a secondary flocculation could be obtained by raising the temperature to 
about the boiling point. The same was found to occur in the present case, : 
but the amount was small, and diminished when the temperature of the first 
flocculation was raised to 60°. Experiment on a moderate quantity of this 
second flocculent substance prepared from the cabbage showed that it was 
formed from some water-soluble protein by coagulation, since it was insoluble 
in all the solvents tried, and the percentage of N in it was about 12. In the 
present case it was separated by filtering on a Buchner funnel, and the filter 
paper, with the retained coagulate, transferred to a Kjeldahl flask and the 
total N estimated. 

The aqueous extract, after boiling to remove all flocculent matter, was 
cooled, its volume taken, and stored in a stoppered bottle with a little 
chloroform and toluene for future use. As a precaution against any unknown 
reaction or separation on standing, the distribution of N in this was per- 
formed without delay. The alcohol and ether washings of the protein were 
evaporated down, transferred to a Kjeldahl flask and the total N estimated. 
The residue was air-dried at 37° in the same oven used for the main bulk of 
the leaves, as stated above. Total N was estimated by Kjeldahl’s method. 

It will not-be out of place here to give some idea as to the time occupied 
in the separation. The grinding (10 mins. first, 6-7 mins. the others) and 
squeezing out take between 1 and 1} hours; raising the temperature to 60°, 
5 mins., filtering through the Soxhlet thimble 15 mins., boiling the filtrate 
10 mins., and the separation of the coagulated protein about 15 mins. About 
two hours, therefore, after the leaf cells are killed the colloidal protein is in 
alcohol, the clear aqueous extract has been boiled, cooled and shaken up with 
antiseptics, whilst the residue, containing none of the water-soluble N, is 
drying at 37°. 


DISTRIBUTION OF N IN THE WATER-SOLUBLE PRODUCTS. 


The total N in solution. Determined by the modified Kjeldahl-Gunning 
method mentioned above. 

Nitric nitrogen. Reduction by Devada alloy in alkaline solution in place 
of the more usual Schulze-Tiemann method was employed, as the same appa- 
ratus as that required for free ammonia and amide N could be utilised.. This 
consisted essentially of a two-necked Claisen flask without side tube, connected 
by means of an ordinary Kjeldahl steam trap to a double-surface condenser 
leading into the distillate flask of 500 cc. capacity containing N/100 sulphuric 
acid. The apparatus was set up in duplicate, the two larger flasks being 
immersed in the same water-bath. Both were also connected to the same 
vacuum pump (Geryk), so that similar conditions as to 7’ and P were main- 
tained in each. 

The procedure was as follows: 50 cc. of the extract was run into each 


350 ' A. C. CHIBNALL 


flask late in the afternoon, then 10 cc. of 30 94 NaOH, and into one only a 
small quantity of powdered Devada alloy. In this latter the evolution of 
hydrogen proceeded steadily during the night, with reduction of the nitric N 
to ammonia. Sufficient quantity of V/100 acid had been run into the receiver 
flasks to cover the tube leading from the condenser and the apparatus 
closed. In the morning 400 cc. of ammonia-free distilled water was run in 
by means of a tap funnel, the water-bath warmed to about 45° and the 
duplicate apparatus evacuated. Distillation was allowed to proceed for about 
two hours, 200 to 300 ce. of liquid coming over. The difference between the 
ammonia found in the two receivers was taken as that due to reduction of 
the nitric N. How necessary is the blank distillation with soda will be seen 
from the fact that the ammonia coming over was about twice that obtained 
from the same volume of extract using magnesia (as below), indicating a very 
slight decomposition by the strong alkali of some nitrogenous bodies (pro- 
teoses?) in the solution. Duplicate readings of nitric N by this method, 
without any regard being paid to the exact 7 or P, agreed to 0-2 cc., an 
accuracy in the solutions containing the lowest concentration of nitric N of 
1 : 100, sufficient for the present research. 

Ammonia nitrogen. 50 cc. or, if sufficient solution was available, 100 cc. 
of the extract were run into the Claisen flasks, then 400 cc. of ammonia-free 
distilled water followed by sufficient cream of magnesia to make the solution 
distinctly alkaline. Duplicate determinations were made at the same time. 
The amount of ammonia measured was equivalent to 1-5-3-5 ec. of N/100 
acid, a quantity sufficiently small to be affected by errors due to the action 
of the MgO on the proteoses (see Nash and Benedict [1921] for discussion on 
methods of estimating free NH, content of blood). The figures given in the 
tables therefore only indicate the very low concentration of ammonium salts. 

Amide nitrogen (asparagine and glutamine N). Sachsse’s [1873] method 
was employed, whereby the extract was boiled with 4 % HCl for two hours. 
Sufficient extract to contain one-fourth to one-third of the total N in solution 
was used for this purpose. After this mild hydrolysis the solution was concen- 
trated im vacuo almost to dryness to remove HCl, transferred with about 
400 ce. of ammonia-free water to the Claisen flask described above, then cream 
of magnesia added to make the solution distinctly alkaline. The total ammonia 
present was found by distillation im vacuo, and from this the free ammonia 
originally present in the solution, as estimated above, was deducted. When 
the details of this research were planned, it was not realised that any quantity 
of proteoses would be present. It is possible that these might give amide N 
under the conditions stated above [see Osborne and Nolan, 1920], but the 
fact that no amide N at all can be detected in the samples picked at night 
shows that the error under this heading is not appreciable. 

Humin nitrogen and nitrogen precipitated by phosphotungstic acid, After 
the distillation for amide nitrogen the procedure follows that of Hausmann for 
the hydrolysis products of protein, as modified by Osborne and Harris [1903]. 


NITROGEN OF THE LEAVES OF THE RUNNER BEAN 351 


Monoamino nitrogen. The filtrate and washings from the phosphotungstic 
acid precipitation were made just alkaline to litmus with 30 % NaOH, then 
just acid by a few drops of glacial acetic acid. They were then concentrated 
in vacuo to about 100 cc. and made up to a standard volume. Monoamino N 
was determined by Van Slyke’s method. 

“Other nitrogen.’ Calculated by difference; the value is therefore subject 
to the sum of the individual errors in the six determinations mentioned above. 

Proteose nitrogen. Estimated by saturation with zinc sulphate in acid 
solution. The precipitate obtained was filtered off, washed with a saturated 
solution of zinc sulphate until the washings were colourless, and redissolved 
in water. The solution was then again acidified, saturated with zinc sulphate, 
the precipitate washed as before, re-dissolved in water and made up to a 
standard volume. Total N was then estimated in an aliquot part by Kjeldahl’s 
method. 

Nitrous nitrogen. This was tested for by both the Griess and meta- 
phenylenediamine reagents, but none was detected in any of the extracts 
(before or after heating) prepared in the present research. 


DISTRIBUTION OF N IN PROTEINS. 


Colloidal protein. Osborne and Harris’ modification of Hausmann’s method, 
referred to above, was used, monoamino N being determined by the method 
of Van Slyke. 

Dried residue. When the total N in these (given later in Table IV) was 
determined they had not been washed in alcohol and ether to remove fats, 
lecithins, etc., since the extractors were being continuously used for work on 
the colloidal proteins. Series 7 B and 98, which were to be further investi- 
gated, were first washed with these solvents, when the N extracted was found 
to be respectively 0-60 % and 0-78 % of the total leaf N. These two ex- 
tracted residues were then boiled up with 500 cc. of 1 % HCl for six hours 
under a reflux condenser, cooled, the undissolved portions filtered off, washed 
with water, and their N content determined by Kjeldahl’s method. In terms 
of the total leaf N they contained 3-2 % and 4-13 % respectively. About 
three-fourths of the N left in the dried residue had therefore passed into 
solution. These extracts were then evaporated to a small bulk in vacuo, made 
up to 100 cc. so that they contained 20 % of HCl, hydrolysed tor 16 hours, 
and the Hausmann numbers determined in a similar way to the colloidal 
protein above. They are given in Table VI later, and show that the N ex- 
tracted was of protein origin. It is probable that the 3-4 % of N in the final 
residues mentioned above was also protein N that had escaped extraction by 
the dilute acid [compare Osborne and Wakeman, 1920]. 

Total protein and non-protein N. Previous work on the colloidal leaf 
proteins [Osborne and Wakeman, 1920; Osborne, Wakeman and Leaven- 
worth, 1921; Chibnall and Schryver, 1921] indicates the possibility of a small 


352 A. C. CHIBNALL 


quantity of the N being of non-protein origin. In all the analyses therein 
mentioned, however, either hot or cold alkali was used for extraction, and 
since the resultant proteins have the properties of alkali-albumins, it is con- 
sidered here that this “non-protein” N has been set free from the original 
protein complex by hydrolytic action. The constancy of the distribution of N 
in the colloidal proteins of the present research (Table V), especially in the 
starvation experiment to be described later, supports this view. In Table IV 
therefore the estimated total protein N is taken as the sum of the N in the 
colloidal precipitate, the dried residue and the protein coagulated by boiling. 

There is evidence that more than-one protein exists, amongst which may 
be cited the fact that the Hausmann numbers of the protein in the dried 
residues differ from those in the colloidal precipitate and also that a small 
part of the protein, as already mentioned, cannot be flocculated at 60°, 
but only precipitates when the liquid containing it is boiled. Furthermore, 
proteoses are present, but separate analysis to determine these could only be 
performed in the later series. 

An analysis of the leaf N (series 9) in terms of protein N, proteose N and 
non-protein N is given in Table 1X, whilst the distribution into seven groups 
after complete hydrolysis is given in Table X. 


Il. DESCRIPTION OF EXPERIMENTS. 


Three main series of experiments have been carried out to determine: 
(a) Seasonal variations. 
(6) Diurnal variations. 
(c) Results of starvation. 


(a) The seasonal variations. 


These were studied by a complete analysis of eleven batches of leaves 
picked at frequent intervals during the life of the plant. Except for the change 
of temperature at which the colloidal protein was flocculated from 45° to 60°, 
as stated earlier, the conditions of the individual analyses remained standard 
throughout. 

(b) Diurnal variations (series 5-6 and 9-8). 

These were studied at the end of the seventh and eleventh weeks. The day 
reading was obtained by picking the batch about half-an-hour before sunset, 
when the leaves had been subjected to 14 hours’ light; the night reading by 
picking half-an-hour before sunrise, when the leaves had been in the dark 
about six hours. 

From a climatic point of view the conditions under which these series 
were taken were ideal. In each case no rain had fallen for several days pre- 
viously, nor was there any dew during the night, the relative humidity at 
the time of the night pickings being 85 °/, and 75 °% respectively. This being 
so, the large decrease in the dry weight of the leaves during the night (10 % 
in both cases) can be taken as due entirely to translocation away from the leaf. 


NITROGEN OF THE LEAVES OF THE RUNNER BEAN 353 


(c) Starvation experiments (series 10). 

These were designed so to alter the conditions in the leaf that light would 
be thrown on the mechanism of either protein synthesis or degradation. 

At the time that the batch for series 10 was picked two samples, 10 © 
and 10 p, were cut with long stems, placed in cups containing ordinary tap 
water so that their stems only were immersed, and left in a subdued light, 
10 c for 100 hours and 10 p for 114 hours. In this way the leaves were kept 
quite fresh, but starved of nitrate and of sugar (other than that due to 
photosynthesis). 


Table I. Giving details of samples picked. 
Age above Time of da 


Series ground when icked, Date when Condition of Condition of 
number in days G.M.T. picked weather plants 
*] 14 8 A.M. June 14 Sunny Height 6-8 ins, 
*2 21 Ps » 21 Dull, no sun Not very different from above 
3 26 Hi » 13 Sunny Average height 12 ins. Just 
starting to climb 
4 33 o » 20 Dull, rainy Starting to flower 
night 
oO eH 46 8.30 P.M July 3 Very fine day Height 6-8 ft., flowers just 
developing 
6 47 2.15 A.M » 4 Fine night As above 
7 60 3 P.M. » 17 Very fine day An appreciable increase of new 
shoots and leaves 
+7N 66 2 A.M. » 23 Fine night Pods just starting to form 
8 77 ae Aug. 3 Heh i Fully grown. Fertilisation had 
not taken place on account 
of the heat. A few pods still 
forming 
9 81 3 P.M. » 7 Very fine day Ditto 
10 97 9 A.M. » 23 Fine day Only a few pods fully formed. 
Numerous new shoots and 
leaves still forming 
ll 143 m Oct. 8 ~~ Most of the leaves old and be- 
ginning to show signs of 
chlorophyll degeneration. A 
few new shoots and leaves 
* Planted 14 days later than the rest of the plants. 
+ No analyses were made with this batch. 
Table II. Giving further details of samples picked. 
Number of Average number Average weight 
Series plants of leaves per of leaves per Average weight 
number used plant plant in g. of leaf in g. 
1 68 57 53 0-935 
2 50 6 5-4 0-715 
3 25 15-3 15-6 1-027 
13 30-5 29-7 0-975 
5) day 12 56-3 50-9 0-904 
6) night 12 53-6 55-6 1-086 
7 12 109-0 90-0 0-827 
7N 12 a 126-0 — 
8) night 12 218-0 172-0 0-791 
9\ day 12 259-0 202-0 0-781 
1 6 372-0 245-0 0-683 


354 © A. C. CHIBNALL 


Table III. (Compare Fig. 1.) Showing the dry weight and total nitrogen. 
Series Fresh weight of | Percentage Percentage of N Percentage of N 


number sample in g. dry weight per fresh weight per dry weight 
1 160 11-43 0-623 5-44 
2 88 13-06 0-671 5:13 
3 177 13-43 0-658 4:90 
4 191 13-16 0-588 4-48 
5) day 193 15°85 0-678 4-28 
6) night 236 14-46 0-653 4-51 
7 590 15-90 0-762 4-76 
8) night 1365 14-42 0-656 4-53 
9\ day 1925 15-51 0-675 4-35 

10 921 16-31 0-724 4-44 

ll 164 16-13 0-622 3:86 


Table IV. Distribution of protein and water-soluble nitrogen in the leaves. 
(In percentages of total leaf N.) 


N of N of N of alcohol and 
colloidal protein Estimated ether washings N remaining 
Series protein coagulated N of protein N of colloidal in solution 
number isolated by boiling residue ‘3 protein after boiling 

1 59-55 2-84 6-88 69-27 2-53 28-20 

2 62-56 1:76 7-15 71-47 2-78 25-75 

3 64-90 3-30 5-44 73-64 2-81 23-55 

4 60-18 5-29 7-86 73°33 2-93 23-74 

5A 56-32 3-61 13-34 73°27 2-44 24-29 

5B 50-70 3°88 19-00 73-58 2-52 23-90 

6A 56-02 2-97 13-37 72-36 2°75 24-89 

6B 48-52 4-36 19-62 72-50 2-80 24-70 

7A 59-61 1-34 12-25 73-20 2-46 24-34 

7B 58-87 0-74 14-11 73-72 2-30 23-98 

8A 61-28 0-68 14-92 76-88 2-02 21-10 

8B 59-37 0-91 16-62 76-90 2-06 21-04 

a 58-11 0-68 16-81 75-60 2°72 21-68 

9B 59-70 0-76 15-09 75-55 2-88 21-57 
10a 56-82 0-68 18-17 75-67 2:17 22-16 
108 52-47 0-82 21-99 75-28 1-07 22-75 
11 58-43 1-43 15-83 75-69 1-90 22-41 

Starvation experiments: 

100 49-55 1-79 12-30 63-64 2-22 34-14 
10D 51-85 0-72 8-75 61-32 2-50 36-18 


* The sum of the three columns on the immediate left. 


Table V. Percentage distribution of nitrogen in the colloidal protein. 


Series Amide Humin Basic Monoamino “Other” 
number N 
1 7:44 3°52 21-65 58-90 8-49 
2 6°56 3°50 22:40 60°25 7°29 
3 6-68 3°30 21-56 58-60 9°86 
4 6-59 3°63 21-08 58:20 10:50 
ost day 6-68 418 20-55 59-67 8-78 
68) night 6°36 3-64 19-82 59-54 10-64 
78 6°37 3°57 20°55 59°10 10-41 
88) night 614 3-55 19-33 59-25 + 1173 
9n\ day 6°34 3°64 22-00 58°88 914 
108 6-50 3°82 21-78 59-22 8-68 
*l0c 6-60 3°56 21-63 58-50 9-71 
ll 6-40 3-82 21-32 58-50 9:96 


* Starvation experiment. 


- NITROGEN OF THE LEAVES OF THE RUNNER BEAN 355 


Table VI. Showing the distribution of N in the products extracted from the 
residues by 1 %, HCl (after complete hydrolysis). 


Series Amide Humin © Basic Monoamino “Other” 
number N N N N N 
7B 8-38 5-73 25-18 48-20 12-53 
9B 8-19 5-82 24-66 48-78 12-64 


Table VII. Distribution of the water-soluble nitrogen (in percentages of total 
leaf nitrogen). After hydrolysis with 4% HCl for two hours. 


Proteose Mono- 
Series Ammonia Asparagine Nitric Humin + basic amino “Other” 
N 


number N N N N N N 
1 0-67 0-79 7:78 0-78 7-63 6-43 4-64 
2 0-22 0-62 7°35 0-79 6-77 6-41 3°59 
3 0:27 0-94 3-93 0°72 6-29 5-54 5°87 
4 0:27 0-51 2-45 0-74 6-67 5-52 7-58 
BAL aa 0:47 0-49 2-71 1-32 6-81 9:57 2-92 
5By C8Y 0:36 0:43 2-96 0-66 6-25 9-25 3-99 
Gis 0-45 nil 3-89 1-19 6-60 10-37 2-39 
ht 
6nj 8 0-34 nil 3-66 1-07 6-79 10-16 2-77 
7A 0-39 0-21 3-77 0-83 7-90 9-33 1-91 
1B 0-32 0-24 4:25 1-06 7-49 8-60 2-02 
84) Vi oht 0-34 trace 3-38 0-94 6-87 7-08 2-49 
spi MS 0-38 trace 3-74 1-07 6-62 7-05 2-18 
9A) aa 0-41 0-48 3-04 1-03 6-67 7-72 2-33 
9p\ C8Y 0-44 0:53 2-75 0-99 6°35 7-99 2-45 
10B 0:37 0-47 3°53 1-06 689 6-81 3-62 
ll 0-61 0-92 1-28 0-96 6-97 6-61 5-05 


Table VIII. Showing the percentage of nitrogen found in the 


colloidal precipitates. 

Series Percentage of Series Percentage of Series Percentage of 
number nitrogen number nitrogen number nitrogen 

1 11-42 ae 11-84 9A 11-66 

2 11-44 6B 10-75 9B 11-71 

3 11-76 oI 11-75 10a 11-54 

4 9-73 7B 11-81 10B 11-36 

pat 10-88 at 11-84 10c) x 11-35 

5B 10-68 8B 11-75 10D 11-42 


* Starvation experiment. 


Table IX. Showing the amounts of protein, proteose and non-protein nitrogen 
in the leaves of series 9 (age 14 weeks). 


Percentage of 
total N 
Colloidal protein ... 59-70 
Protein in residue extracted d by 1 1% HCl 10-24 
Undetermined in residue .. 4-13 
Coagulated by boiling... A ae 0-76 
Total protein N tsp “yk si 74-83 
Proteose N ve 4:35 
Non-protein N, soluble in water.. rE 17-22 
Total... =o rye bas 96-40 


Soluble in alcohol and shi a va 3-60 


356 


A. C. 


CHIBNALL 


Table X. Showing the final analysis of the leaf nitrogen in series 9B 
(in terms of total leaf N). After complete hydrolysis. 


Mono- 
Ammonia Nitric Amide Humin Basic amino “Other” 
N N N N N N 

Colloidal protein — — 3-78 2-17 13-13 35°16 5:46 

Residue — — 1-18 0-82 3°54 7-01 1-82 

Proteoses, water- 0-44* 2-73* 1-26 1:08 2-78 9-01 4-25 

soluble N 

ate — — — — — 0:76 

Soluble in aloohol’! as =e ve sa =f Total N } 3.60 
and ether ; 

Total 0-44 2-73 6:22 4:07 19-45 51-18 15:89 


* Determined before hydrolysis. 


Table XI. (Compare Fig. 1.) Showing the amounts of protein N, water-soluble N and the 
components of the water-soluble N in percentages of the fresh weight of the leaves. 


Water- Aspara- 


Proteose Mono- 


Series Protein soluble Ammonia Nitric sine Humin  -+hasic amino “Other” 
number N N N N N N N N 

l 0-4315 0-1757 0-0042 0-0485 0-0049 0-0048 0:0475 0-0401 0-0289 

2 0-4795 0-1728 0-0045 0-0493 0-0041 0-0053 0-0454 0:0430 0-1241 

3 0-4846 01550 0-0017 0-0259 .0-0062 0-0047 0-0414 0-0358 0-0386 

4 0-4322 0-1396 0-0016 0-0144 0-0030 0-0043 0-0392 0:0325 0-0446 

5A 0-4967 0+1647 : : 

5 4 day 0-4989 01620 0-0027 0-0192 0-0031 0-0067 0-0444 0:0637 0-0237 

Ca); 0-4725 0-1626) , a. : : : } : ; The last seven 

os night 0-4732 0-1613; 9°0026 00-0247 0-0 0:0072 0-0438 0:0670 0-0168 cohitting aaa 

7A 0-5575 0-1855) 2. , ; : Anat , , the mean of 
mt day 0-5620 0-1828 9 0026 0-:0291 0:0016 0-0072 0-0567 0-0683 0-0149 the respective 
8a] _. 0-5045 0-1382) .. : : : ‘ : , A and B values 

ra night 05045 0-1380\ 9 0024 0-0234 0-0 00066 0:0443 0:0463 00-0152 

94 05102 0-1463 i : : 

pd day 0-5100 0-1458\ 2°0028 00196 0-0034 0:0067 0:0439 0:0523 0-0172 
soot 05477 = 0-1605 — — — — — _ — 
10B 05450 0-1647 0-0027 0-0256 0-0034 0:0077 0-0499 0-0493 0-0262 
11 0-4708 01394 0:0038 0-0080 0:0058 0:0059 0:0423 0-0411 0-0314 

Table XII. Showing variations introduced in the distribution of the 
water-soluble N of series 10 by starvation. 
(Figures given are in percentages of the total leaf N) 
seers 
Aspara- Amide N Mono- 

Series Strength ime Nitric Ammonia ne otherthan  Humin  Proteose Basic amino “Other” 
number 1 in hours N N N asparagine N WN N N N N 
1l0A& B 4%, 2 3°53 0°37 0:47 = 1-06 4:59 2°30 6-81 3-62 
100 (1) 4 2 3°42 0:46 0-49 os 0-88 (3-60)! — 0°67 695 17:66 
10c (2) 20 16 342 0-46 0-49 5-69 1-71 — 5-03 8:88 8-46 
10 (1) — _ a= ~ _ _ 3°52 — “= 
10p (2) 20 16 (3-42)* (0-46)* (0-49) 6°57 1-72 — 5-61 9-64 8-27 

9p (1)* 4 2 2-73 0-44 0-53 — 0-99 4:35 2-00 7:76 2:77 

9p (2)* 20 16 2-73 0-44 0-53 0-73 1-08 — 2-78 9-01 4-25 


! Estimated by comparison with 10p (1). 


® Not determined; assumed, by comparison with 10c, to be unchanged. 
* Extract 98 was used as none of 10a or 108 was available. 


we ae ee 


<ul i 


Series 
number 


NITROGEN OF THE LEAVES OF THE RUNNER BEAN 357 


Table XIII. Showing the diurnal variation in the N of series 5-6 and 9-8, in 
percentages of the fresh weight of the leaves. 


Water- Aspara- Proteose Mono- 
Total Protein soluble Ammonia Nitric ne Humin -+basic amino “ 
ihe | N N N N N N N N 


Diurnal difference —0-0250 —0-0250 —0-0014 —0-0001 +0-0055 —0-0031 +0-0004 —0-0006 +0-0033 - 


8B 


day. Mean 06750 0-5101 0-1460 0-0028 0-0196 0-0034 0-0067 0-0439 0-0523 
night. Mean 0-6560 0-5045 0-1392 0-0024 0-0234 0-0 0-0066 0-0443 0-0463 


Other” 
N 


day. Mean 0-6780 0:4980 0-1634 0-0027 0-0192 0-0031 0-0068 0-0444 0-0637 0-0237 
night. Mean 0-6530 0-4730 0-1620 0-0026 0-0247 0-0 00072 0-0438 0-0670 0-0168 


0-0069 
0-0172 
0-0152 


Diurnal difference —0-0190 —0-0056 —0-0068 —0-0004 +0-0038 —0-0034 —0-0001 +0-0004 —0-0060 -—0-0020 


+ Indicates that there is a diurnal increase. — Indicates that there is a diurnal decrease. 


Table XIV. Showing the difference between certain duplicate readings in 


Tables IV and VII. (Expressed as a percentage of the mean value of the 
two readings.) 


Series Protein Water-soluble Nitric Monoamino “Other” 
number N N N N N 
ad 0-44 1-69 8-44 3-37 25-0 
* 0-15 0-83 5-91 2-05 14-0 
in 0-75 1-48 11-30 8-49 5-5 
0-0 0-15 9-60 0-43 12-0 
Ps 0-0 0-50 9-54 3-50 6-0 
“jm 0-50 0-77 Es se SS 


The results of the above experiments are given in Tables III to XI. All 
the figures in the hydrolyses are the means of duplicates. Since the fresh 
weight of the sample taken for grinding might be too low on account of 
evaporation from the surface of the leaves whilst sorting, the distribution 
of N was calculated in terms of the total N. To convert this distribution into 
one in terms of the fresh weight of the leaves, as is given in Tables XI and XIII, 
the appropriate value for “N per fresh weight” given in Table III was used. 
In connection with these tables the following definitions must be noted: 

(1) “Amide N” in Tables V and VI. This is used in the standard way 
to denote the ammonia set free on the complete hydrolysis of proteins, ete. 

(2) “Asparagine N.” Ammonia obtained by hydrolysis with 4% HCl 
for two hours (after deduction of free ammonia nitrogen). (See reservation 
re Osborne and Nolan given on p. 350.) 

(3) “Amide N other than asparagine N.”” In Table XII this is the ammonia 
set free on complete hydrolysis minus that estimated as asparagine in (2), 
and as free ammonia. 

(4) “Monoamino N” refers only to the amino N that is estimated by 
Van Slyke’s method. 

Bioch. xv1 ‘ 24 


358 A. C. CHIBNALL 


A few 8 Chlorophyll 
Young plant Climbing Flowering forming Fully grown degeneration 
q “74 


O55 


nD 
oO 
> 
3 
x 
—s 
°o 
= 
<= 47h 
Miu 
oO 
= 464 
a 
£ 045+ 
~ 
444 
2, 
Ss 043t 
2 
= 018- 
= ~ 
~~ 174 
=) xer-solub 
? e 
S 464 S N. 
£ 
3 154 
Zod 
O7Fr 
O6r 
O54 aa 
—_ te, 
04 “ 
oar ‘Other N’ 
\ Vy, 
02+ “hry, 
' Jc 
Ol} 
i | i i i a, i i i i i i i i i i i j 
0.1 .=68:°8 456.600 To Oise 0 10... ea A Cece oa 


SABE] JO 4YSIOM Ysory tod uoZ0I4IU Jo odeyUDI1Eg 


NITROGEN OF THE LEAVES OF THE RUNNER BEAN 359 


(5) “Other N” is the N.not estimated in any of the other groups given. 
Calculated by difference. 

(6) “Protein N” is the sum of the N in the colloidal precipitate, dried 
residue and coagulated protein found by boiling the extract (see p. 351). 

Fig. 1 illustrates Table XI. The curves are drawn to show a value for 
about 9 a.m. to 11 a.m. (after 6-8 hours daylight). Whenever only sunrise 
and sundown values have been determined, the value shown on the curves 
(Fig. 1) is the mean of the two. 

Sampling and experimental errors in the distribution of N*. In no case was 
it possible to take more than duplicate samples of the same batch of leaves, 
so that probable errors, based on Peter’s formula, cannot be obtained. 
Table XIV however shows the differences between duplicate samples of the 
more important groups given in Tables IV and VII calculated as a percentage 
of the mean reading. The protein N shows a maximum difference of 0-75 % 
only, so that all the changes indicated in Fig. 1 can be taken as significant. 
The water-soluble N shows a maximum difference of 1-69 %. The nitric N 
shows a maximum difference of 11-3 % and a minimum of 5-91%. This is 
much greater than the experimental error, and must be due to variation of 
the concentration in different leaves. As none of the changes indicated by 
Fig. 1 is less than 20 % they can be regarded as significant. 

Monoamino N_ shows a maximum difference of 8-49 °%, although that in 
the other four readings does not exceed 3-5 °%%. The experimental error here 
is undoubtedly high, due to the small volume of N actually measured, but a 
variation of 12 % or over should be fairly significant. 

“Other N” is calculated by difference, and shows a maximum of 25 %. 
Considering that it is subject to all the errors in the rest of the N groups, a 
variation of 50 % is required to be at all significant. 


II. DISCUSSION OF RESULTS. 


The main results, as indicated in the tables and figure, may be summarised 
as follows: 
(a) Seasonal variations. (Tables I, If, II, V; Fig. 1.) 


(1) Total N and protein N, calculated as percentages of fresh weight, vary 
with growth, decreasing when this is very rapid or pods are forming. 

(2) The distribution of N in the colloidal protein remains practically un- 
changed throughout the series. 

(3) Asparagine N and free ammonia N remain low throughout the series. 

(4) Nitric N and monoamino N vary directly with the protein N, indi- 
cating that they may be connected with protein synthesis. Similarly the 
“other N” varies inversely as the protein N, indicating that it may be con- 
nected with protein degradation. 

1 Sampling errors for dry weight and total N were not obtained in the present research. 
Earlier work on the leaves of the broad bean showed, for a batch of 250 g., the following probable 
errors calculated by Peter’s formula: dry weight, 0-48 %; N per dry weight, 0-63 9; N per fresh 
weight, 0-91 %,. 

24—2 


360 A. C. CHIBNALL 


(5) The water-soluble products show considerable variation throughout 
the series, indicating that comparison of leaves from plants of different age 
is impossible [see Jodidi and colleagues, 1920]. 


(b) Diurnal variations. (Table XIII.) 


(1) There is a diminution of total solids at night. 

(2) This is accompanied by a diminution of total N and protein N. 

(3) There is a distinct rise in the nitric N at night. 

(4) The asparagine N disappears at night. 

(5) The other water-soluble products remain more or less unchanged in 
amount during the night. 

(6) The most significant diminution of nitrogenous products at night must 
be assigned to the proteins, indicating translocation of unchanged protein or 
its decomposition products. 

The results for series 9-8 are not so reliable, since three days elapsed be- 
tween the pickings. The conclusions to be drawn from them are similar to 
those stated above, except that there is a loss of both water-soluble N and 
monoamino N. This difference may be connected with the fact that pods 
were forming at the time. 


(c) Starvation experiments (series 10). Tables IV, V and XII. 


(1) There is a considerable decrease of protein N, with increase of water- 
soluble N. 

(2) The distribution of N in the colloidal protein remains unaltered. 

(3) Nitric N and ammonia N remain unaltered. 

After mild hydrolysis: 

(4) Asparagine N and monoamino N remain unchanged. 

(5) Proteose N and basic N have decreased. 

(6) There is a large increase of “other N,” equivalent to the decrease in 
the protein N. 

After complete hydrolysis: 

(7) The “other N” is greatly reduced, and allowing for proteoses in 
solution, half of the loss reappears as ammonia (amide N other than aspara- 
gine N in Table XII) and one-third as bases. 

Perhaps the most interesting of these results is that the protein disappears, 
and is replaced, not as one might expect by amide N and amino N, which 
remain more or less constant, but by N products of undetermined composition, 
given under the heading “other N.” Light is thrown on the nature of these 
products by the different results of mild and complete hydrolysis. It would 
appear that they consist, in large part, of substances that fulfil the following 
conditions : 

After mild hydrolysis: 

(1) Yield no free ammonia, i.e. contain no acid-amide linkage R—CO—N Hy. 

(2) Give no precipitate with phosphotungstic acid. 

(3) Give no increase of free amino groups capable of reacting with 
nitrous acid. 


NITROGEN OF THE LEAVES OF THE RUNNER BEAN 361 


After complete hydrolysis: 

(4) Break down giving about one-half their N as ammonia and one-third 
as bases. 

This great increase of free ammonia on complete hydrolysis suggests a 
urea derivative in which the amino groups are coupled to form compounds 
capable of resisting mild hydrolysis. The ureides are substances of this type; 
further they are, as a class, strongly acidic and capable of forming salts. 
As an example take barbituric acid, 


NH—CO 
Ps .,* 
co CH,. 
NH Co” 


This substance contains both replaceable imino and methylene hydrogen atoms, 
and only on prolonged hydrolysis breaks down giving ammonia from the urea 
nucleus. A condensation product of such an acid with one or more of the 
nitrogenous bases might possibly resist disruption on mild hydrolysis, and so 
give no precipitate with phosphotungstic acid due to liberated bases. In the 
absence of further chemical evidence this view can only be regarded as a 
speculation, but it is worth noting that these products appear to be present 
in small quantities in the unstarved leaf extract. A complete hydrolysis (see 
9B(1) and (2), Table XII) of this extract gives an increase of ammonia 
equivalent to 17 % of the proteose N, a value three times greater than one 
would expect from published data concerning the amide N content of pro- 
teoses. Furthermore, Fosse [1912, 1913, 1914] has demonstrated the presence 
of traces of urea in the leaves of higher plants. 

Summarising the results under (a), (b) and (c) above, it would appear that 
the nitrogenous metabolism of the leaf runs on the following lines: 

Protein 


monoamino N “other N” —-> translocated 


nitric N 

This will explain the following points brought out in the above discussions: 

(1) The graphs of the nitric and monoamino N in Fig. 1 vary with that 
of the protein N. 

(2) In spite of rapid protein degradation at night, there is very little 
change in most of the constituents of the water-soluble N. 

(3) There is an accumulation of “other N” during the fourth and fifth 
weeks when protein degradation appears to be rapid, also later in the season 
when the plant is becoming aged, and translocation less active. 

(4) In the starvation experiment, though protein degradation is rapid, 
there is no increase in amide N or monoamino N, and, since translocation 
away from the leaf has stopped, there is a large increase in “other N.” 

Protein metabolism, however, does not depend on N alone, and due regard 
must be paid to the supply of carbohydrate and other substances when inter- 
preting some of the results outlined above. Thus in the starvation experiments, 


362 A. C. CHIBNALL 


since nitric N is present, it might be asserted that protein synthesis would go 
on until this was used up. But in this case the leaves were starved of sugar, 
and since they were placed in a subdued light photosynthesis would be weak. 
In all probability then, protein synthesis had stopped for lack of carbohydrate 
[compare Suzuki, 1898]. Similarly the accumulation of nitric N at night is 
probably due to the fact that starch, stored during the day, has been partially 
or entirely used up, so that the supply of carbohydrate for protein synthesis 
is reduced or is stopped. 

The low concentration of free ammonia and asparagine N observed 
throughout the series confirms the work of Emmerling [1887] on the broad 
bean, and of Kosutany [1897] on the vine. The researches of Wasilieff [1901] 
and Prianischnikoff [1899] have shown that the growth of seedlings depends 
in part on asparagine derived from the decomposition of the reserve proteins 
in the cotyledons. This may indicate that the function of asparagine in the 
general metabolism of the plant is connected with growth, so that the total 
disappearance at night, an observation confirmed by Kosutany, may be due 
to translocation away from the leaf to the growing points or roots. The 
starvation experiments, where the asparagine N remains unchanged, support 
this view, and show that this substance is not derived from the decomposition 
of the leaf proteins. No indication as to its origin can be deduced from the 
results of the present research, but they lend support to the theory of 
Butkewitsch [1909] that its formation is to remove excess of free ammonia. 


This research was carried out at the instigation of Prof. 8. B. Schryver, 
whom the author thanks, together with Prof. V. H. Blackman, F.R.S., for 
continued advice and criticism. His thanks are also due to Mr Hales, of the 
Chelsea Physic Gardens, for the care bestowed on the plants used. 


This research was made possible by a grant from the Department of 
Scientific and Industrial Research. 


REFERENCES. 


Butkewitsch (1909). Biochem. Zeitsch. 16, 411. 
Chibnall and Schryver (1921). Biochem. J. 15,60. 
Czapek (1920). Biochemie der Pflanzen (Jena), 11, pp. 291-307. 
Emmerling (1887). Land. Vers. Stat. 34, 1. 
Fosse (1912). Compt. Rend. 155, 851. 
(1913). Compt. Rend. 156, 1934. 

———- (1914). Compt. Rend. 158, 1374; 159, 253. 
Jodidi, Kellogg aad? True (1920). J. Amer. Chem. Soc. 32, 1061. 
Kosutany (1897). Land. Vers. Stat. 48, 13. 
Moore (1920), J. Ind. Bng. Chem. 12, 669. 
Nash and Benedict (1921). J. Biol. Chem. 48, 2. 
Osborne and Harris (1903). J. Amer. Chem. Soc, 25, 323. 

———< and Nolan (1920). J. Biol. Chem. 48, 2. 

—— and Wakeman (1920). J. Biol. Chem. 42, 1. 

~———<- Wakeman and Leavenworth (1921). J. Biol. Chem. 49, 1. 
Prianischnikoff (1899), Land, Vers. Stat, 52, 137, 347. 
Sachase (1873). J. pr. Chem. n. F. 6, 118. 

Schryver (1908), J. Physiol. 87. Proc. xxiii. 

Suzuki (1898). Bull. Coll. Agr. Tokyo, 2, 409. 

Waesilieff (1901), Land. Vera. Stat. 55, 45. 


XXV. MILK AS A SOURCE OF WATER-SOLUBLE 
VITAMIN. III’. 


By THOMAS BURR OSBORNE anp LAFAYETTE BENEDICT MENDEL 
WITH THE COOPERATION OF 


HELEN C. CANNON. 


From the Laboratory of the Connecticut Agricultural Experiment Station and the 
Sheffield Laboratory of Physiological Chemistry, Yale University, New Haven. 


(Received February 14th, 1922.) 


THE question of the vitamin content of milk has, very properly, attained 
considerable prominence not only in its scientific aspects in relation to nutrition 
but also from the standpoint of public health. The classic experiments of 
F. G. Hopkins [1912] indicating an unexpectedly favourable effect upon health 
and growth secured by the addition of as little as 2 cc. of cow’s milk per day 
to a “synthetic diet” fed to young rats have given the occasion for the current 
emphasis upon the richness of milk in that nutritive factor now commonly 
designated as vitamin B. Our own early experiments [1911] published in 1911 
had indicated that what we originally termed “protein-free milk”? furnishes 
something without which rats cannot grow satisfactorily when they are kept 
upon “synthetic” diets consisting of mixtures of more or less isolated food 
substances. However, the quantities of “protein-free milk” which we have 
found necessary to supply the essential food factor to rats in numerous trials 
at various times have been decidedly larger than would correspond to 2 cc. 
of milk. This might have been due to a variety of obvious possibilities in- 
cluding the loss of vitamin in the process of preparation of the milk product. 
Consequently we [1918] undertook a re-investigation of the subject, using 
milk as the added source of vitamin B in the ration. The experiments indicated 
the necessity of feeding as much as 16 cc. of fresh cow’s milk to secure the 
desired growth. Again [1920] we made further attempts to demonstrate a 
possible greater vitamin B potency by feeding unpasteurised milk of known 
origin. In some of these trials the diets used by Hopkins were imitated as 
closely as the materials at our disposal would permit. Again “the outcome 


1 The expenses of this investigation were shared by the Connecticut Agricultural Experiment 
Station and the Carnegie Institution of Washington, Washington, D.C. 

® This product consists of the dried solids of skim milk from which the casein has been re- 
moved by precipitation with acid and the coagulable proteins removed by heat. 


364 T. B. OSBORNE AND L. B. MENDEL 


was in harmony with all our experience in showing that even additions of 
10 ce. of fresh milk per day were insufficient to effect a food intake adequate 
for growth at a normal rate” [1920]. 

Neither seasonal variations, differences in the rations fed to the lactating 
cows, nor manipulations of the milk prior to marketing appeared to offer a 
satisfactory explanation of the differences between the results recorded by 
Hopkins and ourselves. He [1920]! has lately reported the outcome of a new 
series of tests of 2 cc. of milk per day as a source of vitamin B for rats. His 
experiences are related as follows: 

“On receipt of Osborne and Mendel’s private communication, I made a 
few experiments in the winter of 1919. At this time the results were frankly 
disappointing. When the animals receiving the 2 cc. of milk in addition to 
the synthetic diet were compared with others which had the latter alone, 
the favourable effect upon health and upon the survival periods of the animals 
was unmistakable. Growth, however, was very slow, and the death rate of 
the rats was higher than that of animals normally fed. 

“In April of this year fresh experiments were begun and continued during 
the summer. The results now became such as to confirm entirely my earlier 
experiments. Out of 20 animals, each receiving daily 2 cc. of fresh cows’ 
milk as an addition to a highly purified synthetic diet, not one failed to grow 
almost normally throughout the period of experiment. The observations were 
not extended beyond 60 days, as they were meant only for comparison with 
those described in my original paper, and not to determine the amount of 
milk necessary for continued growth. The technique was exactly that originally 
used by me. In each experiment control animals were fed upon the synthetic 
diet alone and the contrast between them and a corresponding set receiving 
2 cc. of milk was just as marked as in my original experiments. There has 
been no selection of results.” 

It will be noted in most of Hopkins’ experiments that he began the feedings 
when the rats were still quite young, often weighing not more than 40 g. 
We concluded to repeat our tests upon rats somewhat younger than those 
with which our previous experiments were begun. The standard diet con- 
sisted of 


Casein ... A id tf 8 Butterfat ... 9% 
Salt mixture? ... 4 TAT 2 Oe, ty | 
Starch ... o (O04 


Each day 2 ce. of milk, supplied in a separate dish, were promptly consumed 
by the rats. The tests began October 19, 1921. The milk was obtained un- 
pasteurised from a nearby supply. A bacterial count made in November on 
samples from the same source showed 230,000 bacteria per cc. 


* In this paper a reference (p. 724) to previous experiments by Osborne and Mendel is 
erroneously cited as Biochem, J, 41, 515, It should be J. Biol, Chem, 41, 515. 
* (Osborne and Mendel, 1919.] 


— Ss 


MILK AS A SOURCE OF VITAMIN B 365 


The changes in body weight are shown graphically in the appended chart, 
the weekly food intakes (exclusive of milk solids) being indicated, where 
available, on the curve of growth. During the periods represented by the 
broken lines (- - - - - - ) 40 mg. per day of a yeast fraction prepared by Osborne 
and Wakeman’s method [1919] replaced the milk in the case of rats 7462, 
7485, 7488, 7489. Rat 7506 received 0-1 g. dried brewery yeast per day. 
Rat 7460 died with symptoms characteristic of lack of vitamin B. 


MILK AS A SOURCE OF 


WATER-SOLUBLE VITAMIN 


Initial Body Weight 


40 


41 


41 


2 ec. milk 


Ns 
Se K 20 days > 
44 


K20 grams? 


Fig. 1. 


If variations in the vitamin B content of cow’s milk from different sources 
actually reach the magnitude represented by the discrepancies between the 
English reports and our own they can scarcely be due to variations in the diet 
of the cows. The differences seem too large to be accounted for by such a 
probability. Hopkins himself states: 


366 T. B. OSBORNE AND L. B. MENDEL 


“The incomplete observation just mentioned, and the experience of 
Osborne and Mendel, suggest, at any rate, that the apparent seasonal varia- 
tion in the results of my experiments was not likely to be due to differences 
in the milk. That there is a seasonal factor in the growth energy of rats is 
I think sure, but it is doubtful if it could account for the large difference in 
the experimental results now described. I am endeavouring to obtain further 
light on the matter.” 

In harmony with our own experiences alike with fresh cow’s milk and 
“protein-free milk”’ prepared therefrom are reports from other investigators. 
Thus Johnson [1921], who studied the “growth-promoting properties” of 
milk and dried milk preparations used in the form of “reconstructed” milk 
in the Hygienic Laboratory of the U.S. Public Health Service, found that his 
animals required, as an adequate source of vitamin B, an amount of the milk 
equivalent to that reported by ourselves. Thus he states that “a food mixture 
consisting of purified foodstufis plus milk as the sole source of water-soluble 
vitamine must contain at least 2} parts of milk to 1 part of the basal ration 
in order to produce normal growth. Since a full-grown male rat weighs 
about 280-300 grams, and a full-grown female about 180 grams, a rat upon 
such a diet consumes just about 16-18 cubic centimeters of milk daily. These 
figures agree with those of Osborne and Mendel. It is also seen that the rate 
of growth of rats receiving less than 2} parts of milk in their diet was 
accelerated after increasing the amount of milk in the diet.” 

According to Johnson and Hooper [1921] “Pigeons fed upon mixtures of 
spray process skim milk powder with polished rice require 30 per cent. of the 
food in skim milk powder in order to get full protection from polyneuritis. 
This corresponds to about 75 cc. daily liquid milk.” 

Dutcher [1921], Dutcher, Kennedy and Eckles [1920], Dutcher and 
colleagues [1920, 2], have found spring milk obtained after the cows were 
placed on green grass to be superior to winter milk in “antiscorbutic and 
nutritive properties.” However, when the milk was at its best it required 
10 ce. to furnish “sufficient water-soluble and fat-soluble vitamine for normal 
growth in the albino rat.” Even the greatest reported increases in anti- 
scorbutic potency in the milk of cows fed upon especially suitable rations 
[ Hart, Steenbock and Ellis 1920; Dutcher and colleagues, 1920, 1; Hess, Unger 
and Supplee, 1920] do not parallel the increment which the differences 
represented by 2 cc. (Hopkins) and 10 cc. respectively per day denote. If 
the variations in the vitamin-content of the diet of cattle in different places 
will account for the marked differences in the vitamin B content of cow’s 
milk they must be far more pronounced than has been suspected hitherto. 

Through the cooperation of the Department of Obstetrics in the School 
of Medicine, Yale University, we have been enabled to test mixed samples 
of human milk for its content of vitamin B in the same way as has been 
reported for cow’s milk. The product, which represented the secretion from 
several persons, was not essentially richer than we have found cow’s milk 


MILK AS A SOURCE OF VITAMIN B 367 


to be. 5cc. of human milk per day added to the standard diet devoid of 
vitamin B was insufficient to secure continued increments of body weight. 
When the milk was desiccated and fed in the form of tablets the equivalent 
of 10 cc. sufficed, in addition to the standard food eaten, to secure growth 
at a fairly good rate. In no case was the human milk equivalent in its potency 
as a source of vitamin B to the cow’s milk used in Hopkins’ most successful 
tests. 


REFERENCES. 


Dutcher (1921). J. Ind. and Eng. Chem. 18, 1102. 
Dutcher, Eckles, Dahle, Mead and Schaefer (1920, 1). J. Biol. Chem. 45, 119. 
Dutcher, Eckles, Dahle, Mead and Schaefer (1920, 2). Science, 52, 589. 
Dutcher, Kennedy and Eckles (1920). Science, 52, 588. 
Hart, Steenbock and Ellis (1920). J. Biol. Chem. 42, 383. 
Hess, Unger and Supplee (1920). J. Biol. Chem. 45, 229. 
Hopkins (1912). J. Physiol. 44, 425. 
—— (1920). Biochem. J. 14, 721. 
Johnson (1921). Public Health Reports, 36, 2044; Chem. Abstracts, 15, 3664. 
Johnson and Hooper (1921). Public Health Reports, 36, 2037; Chem. Abstracts, 15, 3664. 
Osborne and Mendel (1911). Carnegie Institution of Washington, Publication 156, pt. ii, 83. 
—— —— (1918). J. Biol. Chem. 34, 537. 
—— —— (1919). J. Biol. Chem. 87, 572. 
—— —— (1920). J. Biol. Chem. 41, 515. 
Osborne and Wakeman (1919). J. Biol. Chem, 40, 383. 


XXXVI. THE ESTIMATION OF NON-PROTEIN 
NITROGEN IN BLOOD. 


By ERIC PONDER. 
From the Department of Physiology, Edinburgh University. 


(Received February 16th, 1922.) 


THE object of this paper is to describe a rapid micro-method for the esti- 
mation of non-protein nitrogen in blood. Several methods have hitherto been 
described, but, for various reasons, are unsatisfactory. The method described 
by Cole [1919] is very lengthy, and requires 5 cc. of blood—an amount which 
cannot be obtained from small animals, and only with some inconvenience 
from man. The method of Folin and Denis [1916] is even more unsatisfactory. 
Several cc. of blood are required. The digestion mixture does not appear to 
be a suitable one, the contents of the incineration tube turning solid before 
the incineration is complete, and being afterwards quite insoluble in water: 
this defect is apparently due to an excess of phosphoric acid. Even if this 
occurrence be avoided, Nesslerisation is very difficult: Cole has observed this, 
and notes that a cloud rapidly appears. Frequently a brick-red precipitate 
appears also. 

The following method, which may be applied to small quantities of blood, 
obviates these difficulties, and may be used with success when estimating 
the non-protein nitrogen in the blood of small animals. 

Special apparatus etc. required: 

1. The solutions required for the preparation of blood filtrates, as described 
by Folin [1919]. ; 

2. A digestion mixture. To 50 ce. of a 5 % copper sulphate solution add 
100 ce. of 85 % phosphoric acid, and 300 ce. of pure concentrated sulphuric 
acid. ) 
3. A boiling tube, 10 x 1 em. The tube need not be of special glass. It is 
graduated at 3-5 cc. It is provided with a small loose stopper, shaped like 
a specific gravity bulb, which is introduced neck downwards into the boiling 
tube, so as to close the orifice. 

4. A micro-filtration apparatus, such as is described in the paper dealing 
with the author’s method for the estimation of blood-sugar [Ponder and 
Howie, 1921]. This is essentially a small filter which filters under slight suction. 

5. Small calibrated pipettes, to deliver 0-2, 0-15, 0-5 and 1:5 ce. 

6. Standard ammonium sulphate solution, containing 0-4716 g. of the 
purified salt per litre; Nessler’s reagent, as described by Folin or Cole. 

The estimation is performed as follows: 

1. Blood is drawn from the finger, or from an animal’s vein, into a 0-2 ce. 
pipette. The contents of the pipette are added to 1 cc. of distilled water: 


NON-PROTEIN NITROGEN IN BLOOD 369 


the pipette is twice filled with water, and these volumes, carrying with them 
all traces of blood from the walls of the pipette, are added to the tube con- 
taining the blood and distilled water. 0-2 cc. of sodium tungstate and 0-2 cc. 
of 2/3 N sulphuric acid are added: the tube is allowed to stand for five minutes 
at least, its contents being occasionally shaken. 

2. A filtrate is produced from the contents of the tube by about two 
minutes filtration in the micro-filter. 

3. 0-5 ce. of filtrate is placed in the boiling tube, and to it is added 0-2 ce. 
of the digestion mixture diluted 1 in 4. The contents are boiled over a micro- 
burner, very gently. As soon as boiling occurs, the stopper is placed in the 
mouth of the tube. The boiling is continued for two minutes. The flame 
should be very low; bumping is thus prevented. At the end of the incineration, 
the flame is removed, and a few drops of water are added. The tube is allowed 
to cool, and is then filled to the mark with distilled water. 

4. A standard is prepared. Place 3-15 cc. of distilled water in a tube, 
add 0-15 cc. of standard ammonium sulphate solution, and v. 2cc. of the 
diluted digestion mixture. 

5. Nesslerise the contents of the boiling tube and the standard simul- 
taneously, by adding 1-5 cc. of Nessler’s reagent. There is no difficulty in 
obtaining a clear yellow solution. 

6. The solutions are matched in the colorimeter in the usual way. If the 
standard be set at 15, the non-protein nitrogen in 100 cc. of blood is 15 
divided by the reading, and multiplied by 30. 

It has not been possible, for the reasons given above, to compare this 
method with that of Folin and Denis. The method has however been applied 
to nitrogenous substances, of known nitrogen content, which require incinera- 
tion before producing a colour with Nessler’s reagent, and has been found 
accurate and satisfactory, the results obtained by the method and the calcu- 
lated results agreeing closely. 

Some results obtained in the examination of blood of various animals are 
given in the following table: 


Animal mg. in 100 ce. Animal mg. in 100 cc. 
Man 36 Rabbit 43 
a 25 Frog 93 
Cat 67 Lizard 76 


In the case of the frog and lizard, the blood was obtained from the heart. 

It is claimed for this method that (1) a very small quantity of blood is 
required, (2) that the technique is simple and rapid, (3) that the final Nessleri- 
sation presents no difficulty, and (4) that the results given are accurate. 


REFERENCES. 


Cole, S. W. (1919). Practical Physiological Chemistry, 260. 

Folin, O. (1919). A Laboratory Manual of Biological Chemistry, 179. 
Folin and Denis (1916). J. Biol. Chem. 26, 491. 

Ponder and Howie (1921). Biochem. J. 15, 171. 


XXVIII. THE CONDITIONS INFLUENCING THE 
FORMATION OF FAT BY THE YEAST CELL. 


By IDA SMEDLEY MACLEAN. 
(Received February 22nd, 1922.) 


Very little is known of the story of fat metabolism in the lower organisms 
although a considerable amount of work has been carried out on the con- 
ditions attending the formation of fat in yeast. Yeast offers a particularly 
favourable field of study as it can readily be grown in large quantity and 
considerable variations can be effected in the conditions of its growth. In 
spite of the work that has been done but little progress has been made since 
the work of Naegeli and Loew [1878, 1879] carried out more than forty 
years ago. 

The conclusions arrived at by Naegeli and Loew were briefly as follows: 

(1) The fatty acid of the yeast cell consisted chiefly of oleic acid. 

(2) The amount of fat obtainable from yeast was about 1 to 2 % if the 
dried yeast was directly extracted by ether; this figure might be raised to 
about 5 °% by first evaporating the yeast with concentrated HCl several times 
on a water-bath. The acid destroyed the cell wall and the ether then was no 
longer prevented from extracting the fat by the impermeability of the cell wall. 

(3) The more vigorous the growth of the yeast cell, the greater was the 
amount of fat formed. Both the total amount of dry substance formed and 
the percentage of fat it contained were raised. 

(4) Other conditions being similar, the percentage of fat formed increased 
with the supply of oxygen. They found that the fat content of a yeast 
grown in a solution of sugar containing ammonium tartrate (2 9%) which was 
aerated throughout the time of growth was 12-5 %, whereas a yeast grown 
on peptone and sugar at a low temperature with scanty respiration contained 
only 5 % of fat. 

(5) Naegeli discussed the question of the origin of the fat formed and 
concluded that under different conditions both carbohydrate and protein 
might act as sources of fat. He drew attention to the rape seed which, before 
maturity, is filled with starch grains and from which, when ripe, the oil is 
pressed and to the case of fungi in which, when put into water, the plasma 
diminishes with the appearance of fat, the cellulose membrane also increasing 
during fat formation. 

(6) Naegeli recognised also that the fat contained a sterol which he termed 
cholesterol and further that the conditions which led to an increase of fat 
led also to an increase of sterol. 


FORMATION OF FAT BY THE YEAST CELL 371 


The nature of yeast fat. 


This has since been elucidated by the work of Hinsberg and Roos [1903, 
1904], Neville [1913], Gérard [1895], and Smedley MacLean and Thomas [1920]. 
As the result of these investigations it has been shown that the fatty acids 
present are palmitic, oleic and linoleic with a small quantity of lauric. The 
fat is chiefly remarkable for the large proportion of sterol which it contains, 
the sterol being apparently identical with ergosterol, a characteristic con- 
stituent of the fats of all the lower plant world. 


Lhe amount of fat present in yeast. 


In the normal yeast cell the percentage of fat present has been described 
as from 1 to 5 % of the dried yeast. Naegeli pointed out that the extraction 
of the air-dried yeast with ether only removed part of the fat, and that if the 
cells were first treated with concentrated HCl from two to three times as 
much fat could be extracted, this being however hydrolysed to the free fatty 
acids. Naegeli’s results have been criticised by later observers who consider 
that prolonged treatment with the strong acid may give rise to ether-soluble 
decomposition products of the yeast cell and that the higher figures obtained 
do not represent the true percentage of fat. 

Hinsberg and Roos [1903] and Bokorny [1916, 3] therefore retain the 
ether extraction method in determining the fat percentage of dried yeast. 
Bokorny [1916, 2] indeed did not confirm Naegeli’s results; he treated the 
air-dried yeast for 24 hours with concentrated HCl and found that the per- 
centage of fat was only 0-66 % of the weight of dried yeast compared with 
2-66 % of fat obtained by ether extraction without the preliminary treatment 
with acid: the sticky mess obtained by treating the yeast with acid, he found 
unsuitable for ether extraction. 

In a series of experiments carried out with the object of determining the 
proportion of fat to carbohydrate in the yeast cell under different conditions, 
I adopted the method of hydrolysing the yeast by boiling with N HCl for 
two hours, filtering and washing the residue with water until the washings 
no longer reduced Fehling’s solution; the residue was then air-dried overnight 
at the laboratory temperature and extracted with ether in a Soxhlet apparatus 
and the filtrate and washings used for the estimation of carbohydrate. After 
evaporating off the ether, the residual fat was taken up with dry ether and 
dried to constant weight in a vacuum desiccator at the laboratory temperature. 

I found that the amount of fat found in this way might be several times 
as great as the amount obtained by the direct extraction of the dried yeast 
with ether. In the latter method the yeast was dried by treating it with a 
large volume of absolute alcohol, the alcoholic extract was evaporated and 
the residue added to the dried yeast before extracting it with ether. 

A comparison of the fat obtained by the two methods showed that the 
two specimens were similar in appearance and from both sterol separated on 


372 I. SMEDLEY MACLEAN | 


- standing; the iodine values of both varied considerably in different experi- 
ments but no consistent difference could be detected: the difference between 
the Wijs and Hubl numbers which may be taken as an indication of the amount 
of sterol present [Smedley MacLean and Thomas, 1921] also showed no con- 
stant variation between the two series of experiments. 


Table I. Showing the amount of fat extracted by ether before and after 
hydrolysis of the yeast. 


Amount of fat 
A. 


Weight of (a) Without hydrolysis (b) After hydrolysis 
dried=. as paBE I Carr 
yeast Weight %ondry Weight % ondry 


Sample of yeast g. g. yeast g. yeast 
Pressed yeast (12-5 g.) 2-99 0-1118 3°74 = — 
3:07 _ 0-1948 6-35 


3°05 — —_ 0-1892 6-20 
12-5 g. above sample incubated 48 eed 
hours at 26° in glucose solution 


Oxygenated 4:17 0-0960 2-29 _ — 
Re 4-80 — — 05402 11-24 
Not oxygenated 4-18 O-1114 2-67 = = 
” ” 4-80 _- — 0-2434 5-07 
A pure culture of yeast grown on . 
wort containing lactic acid (N/10) 
at 30° 
Oxygenated 8-05 0-1076 1-34 _ _— 
* 8-05 — — 01975 2-45 
Not oxygenated 6-49 0-0996 1-53 —_ _- 
- " 6-49 _— —_ 0-1524 2-30 


It must be admitted therefore that ether extraction of the yeast dried 
either in air or by means of alcohol, does not remove all the fat from the cells. 
The criticisms brought forward against Naegeli and Loew’s method of re- 
peated evaporation of the yeast with concentrated HC] cannot be urged 
against the much less drastic treatment of boiling for two hours with 3-6 % 
or even with 1-8 °% HCl, a method by which the fat is not hydrolysed, its 
acid value being barely affected. 


The state in which the fat occurs in the yeast cell. 


Naegeli and Loew apparently regarded the hydrochloric acid as acting by 
impairing the cell membrane and thus permitting the entrance of the solvent 
into the cell containing the fat. They considered it probable that continued 
treatment with alcohol or ether would completely remove the fat, an expecta- 
tion which does not appear to be realised even when the extraction is continued 
for a very long time. Two views present themselves: (1) a proportion of the fat 
may exist in the free state in the cell, being probably formed as a decomposition 
product of the cell-plasma. The remainder of the fat may be in combination 
in the plasma of the cell and may only be liberated on hydrolysis when some 
complex substance in the plasma is itself decomposed, (2) The sub-microscopic 
fat particles may be retained in a protein meshwork, only the larger fat par- 
ticles being extracted by the ether. The smaller fat particles would then only 


FORMATION OF FAT BY THE YEAST CELL 373 


be liberated by the breaking down of the protein when they would be extracted 
by the ether. 

All the known facts as to the extraction of fat from yeast are in agreement 
with the hypothesis that the fat is closely associated with the sterol and protein 
and possibly with the carbohydrate of the cell; this association may be of 
the nature of a chemical combination. 

The evidence upon which this view is based, is as follows: 


(a) Extraction with alcohol and ether removed readily from 1 to 3% 
of yeast fat, calculated on the dried yeast, after which only traces of fat were 
obtained by long continued extraction. 


(b) As stated above about twice as much fat may be removed from the 
yeast cell after boiling with normal or semi-normal acid as is obtained by 
direct ether extraction of the dried yeast. The greater part of the fat which 
is obtainable from dried yeast by direct extraction with ether is removed in 
a comparatively short time; further prolonged extraction only results in the 
separation of traces of fat. 

Thus in an experiment carried out by Miss D. Hoffert, 12-5 g. yeast were 
soaked overnight in alcohol, the alcohol evaporated, the residue added to the 
yeast, the whole dried overnight and extracted with ether in 14 hours; the 
amounts of fat extracted varied from 0-0822 to 0-1060 g. 12-5 g. of the same 
sample of yeast were hydrolysed with N HCl for two hours, the solid residue 
dried overnight and extracted for 14 hours with ether; 0-211] g. fat was 
obtained. ; 

The amounts of fat obtained per hour by direct extraction with ether are 
shown in the following table. 


Table IT. 
7th to Total in 


Time in hours... Ist 2nd 3rd 4th 5th 6th 7th 14th 14 hours 
(1) Weight of fat from 
12:5 g. yeast 0-0334 0-0216 0-0146 0-0090 0-0082 0-0079 0-0055 0-0058 0-1060 
(2) is is 0:0652 0-0077 0-0023 0-0013 0-0014 0-0013 0-0011 0-0019 0-0822 
(3) ” » Se a ee ee 00888 
(c) Old yeast cells or cells grown under unfavourable conditions, e.g. an 
abnormally low or high temperature, give considerably higher fat percentages 
than normal cells when extracted directly with alcohol and ether. Such cells 
when examined microscopically show small globules of fat staining with 
osmic acid. In experiments where the period of incubation is long, partial 
autolysis of the yeast probably takes place—the amount of yeast formed is 
much reduced and the proportion of fat extracted by ether is greater. The 
total amount of fat obtained is decreased; but since the total amount of yeast 
is proportionately still less, the percentage of fat is raised. 
Tubes containing 10 cc. wort were inoculated with a pure culture of 
brewer’s yeast and after 48 hours the contents added to 1500 cc. sterilised 
wort and incubated. The yeast was centrifuged, filtered and dried with 


Bioch. xvi 25 


374 I. SMEDLEY MACLEAN 


alcohol; the residue from the alcohol was added to the dried yeast and the 
product extracted with ether in a Soxhlet apparatus for 14 hours. From the 
figures given below it will be noted that (1) the quantity of yeast formed is 
in inverse ratio to the fat percentage, (2) the percentage of fat tends to be 
higher when the time of incubation is long and (3) incubation in the presence 
of 1% lactic acid at 35° is particularly unfavourable and leads to the pro- 
duction of the smallest amount of yeast and the highest percentage of fat. 


Table III. Showing that a higher percentage of fat is extracted by ether from 
yeast which has been grown under unfavourable conditions. 


Temperature 25—26° 
(a) Reaction neutral (b) 1% lactic acid added 
No. Weight Weight dry Fat No. Weight Weight dry Fat 
days fat. g. yeast. g. % days fat. g. yeast. g. % 
2 0-105 7-4 1-4 2 0-131 6-5 2-0 
2 0-205 8-9 2:3 2 0-117 5:3 2-2 
3 0-143 17 1-8 2 0-110 5-0 2-2 
3 0-152 7-4 2-0 3 0-168 58 2-8 
3 0-155 6-5 2-4 3 0-130 58 2:2 
6 0-155 4-6 3:3 5 0-149 4-2 3-6 
6 0-153 3°7 3:3 
Mean... 0-153 71 2-2 0-134 5-2 2-6 
Temperature 35-36° 
4 0-127 4-1 3-1 4 0-148 3-1 4:6 
4 0-101 2-1 4:7 1 - 0-063 2-6 2-4 
4 0-105 2-9 3-5 1 0-076 13 5-2 
6 0-083 2-4 34 8 0-087 0-96 9-0 
8 0-051 2-9 1-7 12 0-032 0-49 6-5 
12 0-058 15 3:8 12 0-028 0-69 4-0 
Mean... 0-087 2-65 3-4 0-072 1:52 53 


(d) Bokorny [1916, 3] found that when yeast is submitted to the action 
of protoplasmic poisons the amount of fat obtained by extraction was very 
considerably increased. He soaked pressed yeast for some hours in such 
solutions as phenol (5 %), formaldehyde, mercuric chloride, etc. While it 
seems unlikely that living processes would continue to be carried on by yeast 
soaked in 5 % phenol solution, it is certainly conceivable that such treatment 
might decompose the complex substance in the plasma and liberate fat from it, 
if fat be indeed one of its constituents. 

Bokorny’s experiments were carried out on small amounts of yeast and 
the quantities of fat weighed were small. It is known that protein readily 
absorbs phenol [Cooper, 1912] and it is possible that when the phenol-treated 
yeast was extracted with ether, the small amount of fat present may have 
been augmented by traces of phenol which contributed to the 12 % of fat 
obtained. In repeating this work it was found very difficult completely to 
remove the phenol, but I think there is no doubt that after the soaking with 
phenol the proportion of fat extracted by ether is appreciably increased. 
Bokorny regarded the increase as being caused by an abnormal secretion of 
fat deposited as a protection against unfavourable conditions of growth. More 
probably it is to be regarded as fat liberated from combination in the cell 


FORMATION OF FAT BY THE YEAST CELL 375 


contents by the action of the poison. It is interesting to note that a German 
patent (D.R.P. 309,266) recommends the auto-digestion of the yeast to ensure 
the liberation of the fat globules from the cells before extracting with solvents. 


Conditions affecting the amount of fat in the cell. 


During the Great War the question of using the lower organisms as sources 
of fat became one of practical importance especially in Germany where a 
good deal of work was carried out on these lines. Lindner’s [1916, 1919] 
“mineral yeast” (Endomyces vernalis) was cultivated as a source of both fat 
and protein and in this organism a fat percentage of 18 % was claimed. The 
work of Bokorny and other observers was directed to producing a similar 
result with yeast; Bokorny [1916, 1] found that by using peptone as his nitro- 
genous food and repeatedly adding sugar, the fat content of yeast could be 
raised. In one German patent (D.R.P. 320,560) it is claimed that by applying 
the methods used by Lindner to increase the fat content of mineral yeast, 
the fat content of beer yeast and of pressed yeast may be raised to from 
20 to 50 % viz. by growing a surface culture on a non-nitrogenous medium. 
The method of estimating the fat is not given and only the microscopic 
appearance is described. Another patent (D.R.P. 307,789) describes the 
application of hydrogen peroxide and of violent aeration of the glucose 
solution to increase the fat content of yeasts which do not form surface 
cultures. Here the increase of fat is not stated nor is the method of extraction 
indicated. 

In the first series of experiments I carried out, the air-dried yeast was 
extracted with alcohol and ether. Pure cultures of yeast were grown on wort 
with and without aeration of the medium; pressed yeast was added to glucose 
solutions and to wort and incubated with and without aeration of the medium. 
No marked differences were produced in the amount of fat extracted and the 
oxygenation of the medium appeared to be without result. If however the 
yeast was first hydrolysed with normal HCl in the manner already described 
and the solid residue extracted with ether, very marked variations were 
observed and the fat content appeared to be largely raised by the aeration 
of the medium. 

Pressed yeast incubated for 44 hours in glucose solution without oxygena- 
tion showed a marked increase in the total weight of the dried yeast, and the 
total weight of fat present was increased although the percentage of fat 
calculated on the dry weight was decreased. But in the oxygenated glucose 
solution, not only was the total dry weight of the yeast increased, but the 
percentage of fat was sometimes more than doubled. Thus in one experiment 
the fat percentage rose from 6-0 to 11-6 %; but while 11-6 % of fat was 
extracted after hydrolysis, part of the same material treated with alcohol 
and then extracted with ether showed only 3 % of fat. The increased amount 
of fat formed when yeast is grown in a glucose solution which is oxygenated 
throughout the experiment appears to be held in combination in the cell 


25—2 


376 I. SMEDLEY MACLEAN 


plasma and is not extracted by ether until by hydrolysis it is set free from the 
cell complex. The ether-soluble material which was weighed as fat contained 
both fat and sterol. 

Yeast incubated in a solution of glucose to which nitrogenous material 
has not been added is characterised by a high percentage of carbohydrate. 
If a given quantity of yeast be incubated for 44 hours in (1) a solution of 
glucose and in (2) wort containing the same percentage of carbohydrate, the 
total amount of yeast formed in the wort is from two to three times as much 
as in the carbohydrate solution. The total amounts of carbohydrate contained 
in the yeasts after incubation in (1) and in (2) respectively are approximately 
the same, though, since much more yeast is formed during the incubation in 
the wort, the percentage of carbohydrate in the latter specimen is much less. 
The total amount of fat as well as the percentage are considerably higher i in 
the yeast incubated in the glucose solution. 

It is not clear whether the decreased amount of fat in the yeast incubated 
in the medium rich in nitrogen is due to a lessened synthesis of fat or to an 
increased breaking down of the fat after it has been formed. The amount of 
fat is however markedly greater in the yeast from the oxygenated wort than 
in that from the wort which has not been aerated. 

The part played by the oxygen requires further elucidation and is at 
present being further studied; it is uncertain whether, as Slator[ 1921] suggested 
in his study of the conditions affecting the growth of yeast, it acts by removing 
the detrimental influence of the carbon dioxide or by some specific action of 
its own. In all the experiments which I have so far carried out an increase 
in the total amount of fat has been associated with a high percentage of 
carbohydrate in the cell. 

The observation of Naegeli that the more vigorous the growth of the yeast 
cell (as is the case in the aerated medium) the greater the amount of fat formed 
is therefore confirmed. It will be remembered that Naegeli extracted the fat 
after previously warming the yeast with concentrated HCl. Bokorny’s ob- 
servation that in strongly growing yeast cells there is a diminished fat content 
probably depends on the fact that he extracted the dried yeast directly with 
ether, for when this method of extraction is used his results are in agreement 
with those quoted above [Bokorny, 1916, 2]. 

In the following experiments 12-5 g. of pressed yeast were incubated for 
44 hours at 26° in 1500 ce. of the sterilised medium, in some of the experiments 
a current of oxygen being passed through the medium during the experiment. 
The yeast was then filtered and the total reducing substance determined after 
hydrolysis by Bertrand’s method, the result being calculated as glucose. The 
figures given for fat refer to the total amount of material soluble in dry ether. 


The nature of the substance from which fat is formed. 


Naegeli and Loew [1879] appear to have recognised clearly that fat is 
formed in moulds and yeasts from substances existing in the plasma of the 


FORMATION OF FAT BY THE YEAST CELL 377 


cell; they noticed that as the fat globules appeared the plasma diminished, 
and argued therefore that the fat could not be derived from the nitrogen-free 
carbon compounds since these were only present in very small quantity in 
the cell contents. In this case therefore they claimed that the fat must have 
been formed from protein. When peptone was used as the source of nitrogen, 
they regarded the sugar or tartaric acid in the medium as the source of fat. 


Table IV. Showing the effect of oxygenation. 


Carbohydrate 
Yeast dry Fat Fat —_-_os* Nitrogen 
Medium Oxygen weight. g. weight. g. % weight. g. % % 
He 3-17 0-1643 5-2 0-9 28-2 Se 
Original yeast 3-19 0-1812 57 oi uk Es 
In water - 2-64 0-1734 6-2 0-55 20-8 — 
Ae + 2-65 0-1343 51 0-55 20-6 oles 
»» glucose pS 3-72 0-1892 51 2-0 53-8 nES 
Wee ye 4:17 0-4532 10-8 1:85 44-1 nee 
he 3-40 0-2277 6-7 0-92 27-2 ae 
Original yeast 3-48 02158 6-4 0-88 26-0 sm 
In water na 2-87 0-1542 5-4 0-66 23-0 ps 
bee + 3-03 0-1812 6-0 0-64 21-2 ea 
»» glucose - 4:98 0-2824 5-7 2-50 50-1 — 
sas 4 5-69 0-6623 11-6 2-64 46-3 e3 
& 3-13 0-1892 6-05 0-49 15-7 7-87 
Original yeast 3-07 01948 6-37 0-55 17-9 7-67 
Tn glucose - 4-80 0-2440 5-07 2-40 50-0 5-10 
ae ‘fe 4-80 0-54.02 11-24 2-21 46-1 4-73 
ee - 2-98 0-2183 7-32 0-51 17-2 8-78 
Original yeast 2-90 02193 7-56 0-52 18-1 8-69 
In glucose + 4-65 0-6267 12-43 2-00 23-3 4-77 
Original yeast 3-93 0-2481 6-30 0-84 21-5 7-96 
In wort > 11-26 0:1637 1-45 2-42 21-5 8-55 
pai ss 12-27 0-2153 1-76 2-56 20-9 8-72 
oat + 11-73 0-2846 2-43 2-63 21-1 8-51 
Original yeast 
(1) Inwort - 6-44 0:1487 2-31 1-72 26-7 8-70 
yee <n 3-29 0-1169 3-51 1-05 28-4 8-97 
(1) » x + 10-04 0-4165 4-15 2-61 26-0 8-67 
yaaa + 9-27 0°3535 3-65 2-30 23-72 is 


(1) Instead of 12-5 g., 2-5g. of yeast containing 0-0505 g. fat was added to the solution. 
(2) The inoculating dose was here 0-25 g. containing 0-005 g. fat. 


They thought it probable that all materials for fat formation were first 
built into the protoplasm protein and in this way utilised for the formation 
of fat, the fat being subsequently split off and stored as reserve material. 

The figures already quoted (Table IV) show conclusively that when yeast 
is grown in oxygenated glucose solution, where no external supply of nitrogen 
is available, the loss in protein even if the protein were wholly converted to 
fat could account only for a small portion of the fat formed. As it is extremely 
unlikely that all the amino-groups present would be converted into fatty 
chains, it is clear that some other source must be found for the manufacture 
of the fat. When yeast is grown in glucose solution there is a large accumu- 
lation of carbohydrate in the cell and since the protein which has disappeared 
is insufficient to account for the increase in fat, either the carbohydrate of the 


378 I. SMEDLEY MACLEAN 


external medium or that of the yeast cell itself must have acted as the starting 
material for the formation of fat. The following figures will make this 
apparent. 


Dry Carbo- 
yeast hydrate Fat Nitrogen Protein 
g. g. g. g. g. 
1. 12-5 g. of the original sample of 
yeast contained 3-1 0-52 0-192 0-2409 1-50 
12-5 g. yeast incubated 44 hours 
in oxygenated glucose solution 
at 26° and filtered. The residue 
contained Be 4-80 2-21 0-540 0-2270 1-42 
Gain or loss “te. oo het _ +169 +0348 — 0-08 
2. 12-5 g. of the original ‘etnias of 
yeast contained 2-94 0-52 0-219 0-257 1-61 
After incubating 44 hours at 26° 
in glucose solution and filtering. 
The residue contained as 465 2-00 0-627 0-222 1:39 
Gain or loss a oe. Tie ae ef! +1-48 +0-408 “— 0-22 


I endeavoured to find out whether the carbohydrate stored in the yeast 
cell was converted in the presence of a free supply of oxygen into fat; a sample 
of yeast which had been grown in glucose solution for 44 hours and contained 
a high carbohydrate percentage (46-7 °4) was then transferred to water and 
a rapid current of oxygen passed through for about 20 hours. The carbohydrate 
and fat were determined in the yeast before and after it was submitted to the 
action of the oxygen. 

3°84 g. dry yeast contained before being submitted to the action of oxygen 
0-1932 g. fat and 1-79 g. carbohydrate; afterwards 0-2146 g. fat and 0-96 g. 
carbohydrate were detected. In another experiment 3-32 g. dry yeast con- 
tained 0-1477 g. fat and 1-446 g. carbohydrate, after growing in the glucose 
solution. After oxygenation in a solution containing 1 % of potassium and 
magnesium phosphates 0-1736 g. fat and 0-88 g. carbohydrate were found. 
Though the differences in the amounts of fat are not very large the increase 
suggests that the fat is formed from carbohydrate material stored inside the 
yeast cell. 


CONCLUSIONS. 


Free fat exists in the normal yeast cell in small amount; the percentage 
of fat is increased in old and degenerating cells, in cells grown under un- 
favourable conditions and in cells exposed to the action of protoplasmic 
poisons. A large part of the fat normally present in yeast and of the sterol 
associated with it, is in some form of combination in the plasma of the cell 
and is not extracted from it by treatment with alcohol and ether. This 
complex is decomposed by boiling with dilute mineral acids and the fat can 
then be readily extracted from it by ether. It is possible that part of the 
protein and carbohydrate which are hydrolysed by the dilute acid are also 
combined with the sterol and fat. 

A free supply of oxygen and a non-nitrogenous medium rich in carbo- 
hydrate are conditions producing an increased amount of fat in yeast but 


FORMATION OF FAT BY THE YEAST CELL 379 


the fat formed in this way appears to be entirely held in combination in the 
yeast cell, and is only set free by hydrolysis. 

The fat and sterol formed in this way are derived from carbohydrate. 

This investigation has been carried out for the Food Investigation Board 
of the Department of Scientific and Industrial Research, and I desire to 
express my thanks for the grants which have been received for this purpose. 
I should like also to thank Miss Florence Banks for the assistance she has 
given in the experimental part of the work. 


REFERENCES, 


Bokorny (1916, 1). Allgemein. Brauer- und Hopfen-Zeitung, 56, 603. 
—— (1916, 2). Allgemein. Brauer- und Hopfen-Zeitung, 56, 1479. 
—— (1916, 3). Biochem. Zeitsch. 75, 346. 

Cooper (1912). Biochem. J. 6, 362. 

Gérard (1895). Compt. Rend. 126, 909. 

Hinsberg and Roos (1903). Zeitsch. physiol. Chem. 38, 1. 

—— —— (1904). Zeitsch. physiol. Chem. 42, 189. 

Lindner (1916). Wochsch. Brauerei, 33, 193. 

—— (1919). Zeitsch. techn. Biol. 7, 79. 

Naegeli and Loew (1878). Annalen, 193, 322. 

—— —— (1879). Miinchen Akad. Sitzber. 289. 

Neville (1913). Biochem. J. 7, 347. 

Slator (1921). J. Chem. Soc. 119, 115. 

Smedley MacLean and Thomas (1920). Biochem. J. 14, 483. 

—— —— (1921). Biochem. J. 15, 319. 


XXVIII. THE ACTION OF WHOLE BLOOD 
UPON ACIDS. 


By ERNEST LAURENCE KENNAWAY anp JAMES McINTOSH. 


From the Bland-Sutton Institute of Pathology, Middlesex Hospital, 
and the Cancer Hospital Research Institute. 


(Received February 23rd, 1922.) 


ALTHOUGH very many investigations have been made upon the alkali reserve 
of serum and plasma, and upon the movements of ions which take place when 
red corpuscles are exposed to different tensions of CO, it seems that no 
quantitative study of the acid-neutralising power of whole blood has been 
made. It is commonly assumed and stated that whole blood is more perfectly 
buffered than plasma. There is no difficulty in investigating this matter if 
the blood is placed in an isotonic medium. 


MeEtTuHop. 


Whole blood was shaken with 0-01N H,SO, containing 0-9 % NaCl. The 
mixture was then centrifuged and the supernatant fluid titrated to ascertain 
how much of the acid had been neutralised, and examined by the electrode 
for H-ion concentration. The number of volumes of 0-01N acid added to 
one volume of blood varied from 2-28 to 9-66; the actual amounts used are 
given in the tables. The mixture of blood and acid was in all cases made up 
with 0-9 % NaCl to 8-25 ce. as this volume gave a convenient amount for 
duplicate titrations and H-ion determination. It was then shaken vigorously 
for five minutes in a wide centrifuge tube in a strong current of CO,-free air 
to remove CQ,, and centrifuged in corked tubes. Of the. supernatant fluid 
duplicate portions of 2-5 cc. were taken, mixed in test tubes with 5 cc. freshly 
boiled distilled water and nine drops of 0-1 °% phenol red in alcohol, and 
titrated with 0-01N NaOH in a comparator, a phosphate solution of py 7:4 
being used as colour standard. The remainder of the supernatant fluid was 
used for H-ion determination. Plasma was treated in exactly the same way 
as the whole blood except that the centrifuge was not used. Two series of 
observations were made, namely upon human blood derived from one healthy 
subject over a period of about three months, and upon guinea-pig’s blood 
(Tables I and II, and Figs. 1 and 2). 

It is at once evident in Fig. 1 that the behaviour of whole blood is quite 
different from that of plasma. However much acid is added to the human 


THE ACTION OF WHOLE BLOOD UPON ACIDS 381 


plasma, roughly the same amount is neutralised; in fact plasma behaves, 
though with much wider variations, like a solution of caustic soda. With 
whole blood on the other hand the amount of acid neutralised increases with 
the amount added in such a way that round about 80 °% of the acid is neutral- 
ised in every case. The exact percentages can be seen in Table I; they range 
from 89 to 78 %, and tend to diminish as the amount of acid increases. 


6 ee ae CRO ON re ce ee eS 


lot 


Fig. 1. NEUTRALISATION OF ACID BY HUMAN WHOLE BLOOD AND PLASMA. The points of inter- 
section of the diagonal line with the ordinates represent amounts of acid added to 1 cc. blood 
or plasma; the points marked @ and ( on these ordinates represent the fractions of these 
amounts which are neutralised by the blood (@) or plasma (0). A second oblique line has been 
drawn at a height which is 80 % of the height of the diagonal. 


The results obtained with the whole blood of guinea-pigs (Fig. 2) are 
confirmatory; the amounts neutralised keep rather closer to the 80 % line. 
With the plasma, however, extremely wide variations were found; in some 
cases the plasma had scarcely any neutralising power. This is probably due 
to the disturbances of circulation and respiration brought about in the taking 
of the blood, which are quite absent when blood is drawn from the human 
subject. It is remarkable that there is no indication in the results obtained 
from whole blood of these variations in the plasma. The maximum amount 
of 0-01N acid neutralised by 1 ce. plasma is about the same (2-3 to 2-4 cc.) 
in both cases (man and guinea-pig) and indicates that the “alkali reserve” 
of the plasma is equivalent to that of a 0-024N solution. 

The addition of increasing amounts of acid to whole blood cannot of course 
be carried on indefinitely, because haemolysis supervenes. This occurs in 
human blood when more than about 11 ce. 0-01N acid is added to 1 ce. blood, 
whereas | cc. guinea-pig’s blood is lysed by 7 or 8 cc. 0-01N acid. The amounts 


382 E. L,. KENNAWAY AND J. McINTOSH 


of acid used in these experiments are not so large as to be altogether beyond 
the range of what may occur in the body; thus in diabetic coma the concen- 
tration of compounds of aceto-acetic and f-hydroxybutyric acids in the blood 
may be that of a 0-03N solution [Kennaway, 1918]. This would be in some 
degree comparable with an experiment in which 3 cc. 0-01N acid was added 
to 1 cc. blood (see Figs. 1 and 2). No methhaemoglobin was detected with 
the spectroscope in these mixtures even when more than enough acid to 
produce lysis was added. 


10 + 
fe) 


8 


Fig. 2, NEUTRALISATION OF ACID BY GUINEA-PIG’S WHOLE BLOOD AND PLASMA. 
For description, see Fig. 1. 


rl j oe i i i j i eG i i n n 
2 4 6 8 10 12 14 
2.c. 0:01 N. Accd added to1ce.Blood or Plasma. 


Fig. 3. pyy OF MIXTURES OF ACID AND HUMAN BLOOD (@) OR PLASMA (1). 


THE ACTION OF WHOLE BLOOD UPON ACIDS 383 


Table I (see Figs. 1 and 2). Neutralisation of acid by whole blood. 
All mixtures of blood and acid were made up to 8-25 cc. with 0-9 % NaCl. 


Mixture Human blood Guinea-pig’s blood 
cc. — 
~~ Ratio 1 ce. blood 1 ce. blood 
Blood + Acid cc.0°01N neutralised Fesgentage neutralised Percentage 
— — acid to cc. 0°01N of aci cc, 0-01N of acid 
001N 0°005N 1ce. blood acid neutralised Final py acid neutralised Final p,, 
1-75 4-0 — 2-286 1-83 80- 7-05 1-906 83-4 — 
” ” — ” 1-83 80-2 = nd rong he 
* a — ye 1-94 85-0 71 — _ — 
5 i — ‘5 2-03 88-8 6-6 — — — 
1-0 — 5-0 2-5 2-20 88-0 -— as -- _ 
1-50 4-0 — 2-66 2-22 83-4 — 2-20 82-7 -- 
1-50 5-0 — 3°33 2-90 87:1 6-5 2-76 82-9 6-15 
s ss — a2 2°84 85:3 —- 2-84 85-4 — 
Fr pe — As 2-87 86-0 6-35 — _ — 
0:75 — 5-0 Ws 2-83 85-0 — — — — 
1-0 — 7:25 36 3-04 84-4 — — — — 
0-75 — 6-0 4-0 3°39 84-8 — — — — 
0°75 — 7:25 8648 4-05 84-3 _ — — — 
1-0 5-0 —_ 5-0 4-15 83-0 5-2 4-01 80-2 53 
” ” INES ” 4-08 81-6 5-65 4:13 82-6 5:8 
0°75 — 75 Ze 4:25 85-0 — — — _ 
0-8 5-0 — 6-25 — _— — 5-14 82-2 5-55 
0°75 5-0 — 6-66 5:35 3 5-4 5-52 82-9 5-05 
_ ty — Be 5-53 83-0 5:3 5:34 80-2 5:8 
1- 7-25 — 7-25 5-93 81-8 5-05 lysis — — 
¥ ‘i _ * 6-01 82-5 4:8 — — _ 
0-5 —~ 7-75 7°16 6-16 79:5 — — nae ia 
0°75 6-0 — 8-0 6-28 78-5 5-65 — — — 
0-75 6-5 — 8-66 6-79 78-4 4:95 — — _ 
0°75 7-0 — 9-33 7-48 80-1 — — — — 
0:75 7:25 — 9-66 7-55 78-2 4-5 — — — 
” ” — ” 7-58 78-5 arene — ea — 
0: 5:5 — 11-0 lysis -- = — — — 
Table II (see Figs. 1 and 2). Neutralisation of acid by plasma. 
Acid and plasma were mixed in the same amounts as those given in the first 
column of Table I for acid and whole blood. 
Human plasma Guinea-pig’s plasma 
iy a > t “ —, 
Ratio 1 ce. plasma 1 ce. plasma 
ce. 0-01N neutralised Percentage neutralised Percentage 
acid to ec. 0-01N of acid ec. 0-01N of acid 
1 ce. plasma acid neutralised Final p,, acid neutralised Final Py 
2-286 — - os 0-579 25-3 5-05 
re — — — 1-797 78-6 6-4 
4 — — _ 1-825 80-0 6-7 
2-66 1-93 72:5 6-5 1-308 50-0 5-55 
ad 2-25 84-6 — 1-59 60-0 6-6 
3°33 2-41 72-4 6-15 1-595 47-7 58 
5-0 2-38 47-6 — 0-347 7-0 4-2 
ri 2-40 48-0 == 2-18 43-6 4-5 
nS — — a 2-31 46-2 4:35 
6-66 1-915 28-8 4-0 0-84 12-6 4-0 
2-22 33°3 4-25 2-18 32-7 3-95 
= a — — 1-10 16-5 4-1 
8-0 2-06 25-8 — — — pees 
8-33 1-88 22-6 — 0-96 11-6 — 
9-66 2-09 21-6 3-65 0-73 7-6 3-45 
as 2-06 21:3 — 1-86 19-3 3-05 
2-11 21-8 = 2-16 22-4 3-4 


384 E. L. KENNAWAY AND J. McINTOSH 


The measurements of H-ion concentration (Fig. 3 and Tables I and II) 
show that the whole blood does not bring this concentration to any constant 
level when different amounts of acid are added. With the addition of in- 
creasing amounts the H-ion concentration increases at a diminishing rate, 
roughly from py 7 to 4. In striking contrast to their effects upon titratable 
acidity, the whole blood and the plasma have very much the same effect upon 
py; the plasma shows a rather greater acidity, as one would expect. The 
rather wide variations in py are no doubt due to the errors of pipette measure- 
ment, and they would be wider still if the mixtures were not buffered; thus an 
error of only 0-01 cc. in the pipetting of 0-01N acid in these experiments 
would, in an unbuffered solution, be nearly sufficient to produce the difference 
between py 5 and 6. 

This quite unexpected behaviour of whole blood towards acids requires 
further investigation. The word “neutralisation” has been used above, for 
want of any less committal term, to denote simply a disappearance of titratable 
acid, which is all that is actually observed; this disappearance might be due 
to combination with inorganic or organic bases, or to some form of adsorption, — 
or to both of these. The inorganic bases of the red corpuscles are insufficient 
to account for the neutralisation of these amounts of acid. 1 cc. of blood can 
neutralise 7-5 cc. 0-O01N acid, of which 1-2 éc. is to be assigned to the plasma 
and 6-3 cc. to the corpuscles (see Tables I and II). The amounts of K and Na 
present, according to Schmidt’s analyses, in 0-5 cc. of red corpuscles, are 
equivalent to about 5 cc. 0-01N. Hence these bases would be unable to neu- 
tralise such an amount of acid even if they were wholly combined with carbonic 
acid, whereas the analyses show that they must be combined chiefly as chlorides, 
phosphates, and sulphates. 

The following points suggested themselves for investigation: 

(a) Is the amount of acid neutralised dependent upon its dilution? 

Fig. 4 and Table III show the amounts of acid neutralised when constant 
amounts of human blood (0-75 cc.) from the same subject and of 0-01N acid 
(5 cc.) were allowed to react in the presence of varying amounts of saline, 
the acid being diluted before the blood was added, and the total volume 
ranging from 6-25 to 45-75 ce.; the concentration of acid thus varied seven-fold. 
The differences observed in the case of any one of the four samples of blood 
examined thus are mostly within the range of experimental error. Hence, 
under the conditions studied, dilution does not have any distinct influence 
upon this neutralisation of acid. 

(b) Is the acid taken up by the corpuscles liberated if they are re-suspended 
in a fresh neutral medium? This was found not to be the case. 1 cc. blood was 
shaken with from 5 to 7-5 ce. 0-01N acid in saline in the manner described 
under “Method” above, and centrifuged, and the supernatant fluid drawn 
off as completely as possible; the corpuscles were then re-suspended in neutral 
saline, shaken for five minutes, let stand for a time, the mixture then centri- 
fuged and the supernatant fluid titrated. Only traces of acid (0-05 cc. 0-01N) 


THE ACTION OF WHOLE BLOOD UPON ACIDS 385 


were recovered; such an amount might well have remained enclosed between 

the packed-down corpuscles when the first supernatant fluid was removed. 
(c) Is the neutralisation of acid by whole blood dependent upon the 

structure of the red corpuscles, or does lysed blood act in the same way? 


5 
SS —. 
4- 1 2 
ee SE 
0-01N. 
aL 
ib 
L Paes | L L | lL | a I 
Total Volume 5 15 25 35 45 


Fig. 4. N&UTRALISATION BY WHOLE BLOOD OF ACID IN DIFFERENT CONCENTRATIONS. 
0-75 cc. blood was mixed with 5 cc. 0-OLN acid diluted with saline. Points plotted 
represent the amount of acid neutralised. The numbers 1, 2, 3, 4 refer to four 
different samples of blood. 


Table III (see Fig. 4). Neutralisation by whole blood of acid 
in different concentrations. 
0-75 ce. blood +5 ce. 0-OLN acid. 


Sample of Total volume cc. 0-01N acid Sample of Total volume ce. 0-01N acid 
blood ' ce. neutralised blood ce. neutralised 

1 6-25 4:17 3 11-25 4:37 
25-75 4-31 16-25 4-40 

2 15-75 4-32 20-75 4:37 
30-75 4:24 4 6-75 4-41 

35-75 4-25 18-25 4-43 

45-75 4-05 30-75 4:37 

45-75 4:27 


The difficulty in settling this question is due to the very deep colour of 
lysed blood; this necessitates considerable dilution before titration can be 
carried out. The small amount of acid to be titrated is in very dilute form 
and it is difficult to prevent the presence of appreciable quantities of CO, in 
the large volume of water required. It might be possible to employ dialysis, 
but it seemed that this might introduce various other sources of error. It 
may be best to describe one of the experiments in detail. The lysed blood 
was prepared by freezing and thawing four times. The following mixtures 


were made: 


Blood + 0-01LN H,SO, in , Normal Total Burette reading 
normal saline saline volume cc. 0-01N NaOH 


: 0-5 4 4 8-5 0-20, 0-21 
Unlysed ios 0 -: 8 ‘ O-1l, 0-12 
0-5 4 4 i 0-50, 0°50 

Lysed tO 0 8 . 0-115, 0-11 


The mixtures were shaken for five minutes and centrifuged, in the ordinary 
way; the two with 8 cc. saline only and no acid served as “blanks.” Duplicate 
portions of 1 cc. of each supernatant fluid were taken, diluted with 36 cc. 


386 E. L. KENNAWAY AND J. McINTOSH 


freshly boiled water and 2 cc. 0-1 % phenol red and titrated in the comparator; 
the colour in the lysed samples was very slight at this dilution (1 of blood 
in 663) but was matched by diluted blood placed in front of the colour standard. 


The results of three such experiments were as follows: 
oe) 


neutralised 
I 1 cc. unlysed blood neutralises 6-21 cc. out of 8-0 ce. 0-01N 77-7 
* (Lec. lysed = e 1:54 a 55 19-2 
ll \lcc. unlysed ,, a 6:28 a = 78-5 
* jl ce. lysed Red . 1:37 ie e 17-1 
III lec. unlysed ,, <a 6-93 * <3 86-9 
* lee. lysed o re 2°31 aA ms 29-0 


As has been explained above, this form of experiment is not adapted to 
a high degree of accuracy; but it does show beyond doubt that lysed blood 
has only from one-quarter to one-third of the acid-neutralising power of un- 
lysed blood. The actual difference measured in this comparison of lysed and 
unlysed blood is that between two burette readings; with the burette employed 
the difference of level was about 24 inches, a sufficiently obvious amount. 
The acid-neutralising power of lysed blood is not greater than that of plasma 
(see Table II). The remarkable property of whole blood which is illustrated 
in Figs. 1 and 2 seems then to be due to the structure or physical condition 
of the red corpuscles. Probably there are many such forms of adsorption or 
pseudo-adsorption which have not yet been described. 


SUMMARY. 


The action of whole blood upon acids is quite different from that of plasma. 
If a series of amounts of sulphuric acid is added to plasma, and to whole 
blood, so as to produce py between 7 and 4, it is found that the amount of 
titratable acid removed by whole blood increases with the amount added in 
such a way that a roughly constant percentage, namely about 80 %, of the 
acid disappears; whereas the amount neutralised by plasma does not vary 
with the amount added. 

It was not found possible to wash out the acid taken up by the red 
corpuscles. Within the range investigated, the amount of acid removed was 
independent of the dilution. The inorganic bases of the corpuscles are in- 
sufficient to neutralise the acid. Since lysed blood does not take up acid in 
this way it seems that the action of whole blood upon acids must be due to 
some form of adsorption which is dependent upon the structure or physical 
condition of the red corpuscles. 


_ REFERENCE. 
Kennaway (1918). Biochem. J, 12, 120, 


XXIX. SALIVARY SECRETION IN INFANTS. 


By CLEMENT NICORY. 
(Received February 28th, 1922.) 


THERE seems to be great difference of opinion as to the age of onset of the 
secretion of ptyalin in the saliva and its relation to the food of the child. 

Schiffer [1891] says that the ptyalogenous function is present in the new- 
born child in the submaxillary gland, but is not established in the parotid 
under the age of two months. 

Moll [1905] obtained saliva containing ptyalin from a parotid fistula in 
a seven months foetus. 

Ibrahim [1908] confirmed the work of Moll, and found that ptyalin appeared 
earlier in the parotid than in the submaxillary gland. This seems natural 
when one remembers that the parotid is a purely serous gland, and most of 
the ptyalin in the mixed saliva of the adult comes from it as shown by Pavlov. 

Berger [1899] examined the dry residue of the saliva in new born babies, 
and found that the foetal and adult salivary secretions do not differ very 
considerably in their consistency. The dry residue in new-born babies being 
0-82 % and in adults about 0-79 %,. 

According to most investigators the secretion of saliva is very slight until 
after the sixth week of life when the amount increases considerably. Korowin 
succeeded in collecting as much as 1} ce. in five minutes in a child of four 
months. . 

Finzi [Rev. Hyg.| found that up to the age of six months the amylolytic 
power of the saliva was the same, whether the food contained starch or not, 
but after that age the power increased if the food contained starch. 

Histologically the infant’s salivary glands are relatively rich in connective 
tissue and in blood vessels, and the mucous cells are few in number. At two 
years of age however it is impossible to distinguish microscopically an infant’s 
salivary gland from that of an adult. 

In an extensive investigation of the subject I have recently obtained the 
following results at St Thomas’s Hospital and the Evelina Hospital for 
Children, London. 

The saliva was collected by giving infants plugs of gauze weighing 500 mg. 
to suck, the amount of saliva being estimated by the increase in weight of 
the plugs. The latter were extracted with 5 cc. of distilled water for one hour 
and then two series of tests were done. 


388 C. NICORY 


(a) Equal parts of salivary extract and 1 % starch solution were mixed. 
(b) To 1 cc. of the salivary extract 2 cc. of 1 % starch solution were added. 
The test tubes were kept at a temperature of 40° C. 

The following table shows some of the results obtained: 


Equal parts of Two parts starch 
starch solution solution to one 


Condition of Nature of and salivary part salivary 
Age infant ° diet extract extract 
34 weeks (premature) — Whey 98 minutes 130 minutes 
36  ] 2? ane ” 30 ” 104 ” 
36—Ci«y, a Not as good se No change after two hours 
as others ' 

5 hours (after birth) Fair Breast 23 minutes 30 minutes 
7, ee Good 16 = 30 ye 
y! ee Poor se 29 ‘Fr 41 iS: 
"| ee Good ne 17 Er 29 Re 
AS Very poor nS : No change after two hours 
12 days Poor Cows’ milk and 17 minutes 65 minutes 

(one of twins) water 

9 weeks Fair Breast 8 Ss ll - 
10 ” Good ” 4 ” 5 ” 
ll ” Fair ” 7 ” 8 ” 

3 months Poor Cows’ milk 6 oe 8 2 

' pee Very poor Breast and bread 9 = 13 fe 

+: eaters Fair Breast 1 minute 3k 

1 year Good —_—- 42 seconds 1 minute 


The time indicated in the table was taken at the moment when the mixture 
of enzyme and starch solution had reached the achromic point as indicated 
by the iodine test. At this point the starch has reached the stage of convertion 
into achroodextrin and some maltose. The presence of reducing sugar was 
demonstrated with Fehling’s solution. 

An analysis of the results obtained from the saliva of 80 different infants 
supported the following conclusions: 

(1) Ptyalin is present in the saliva of infants at least 1} months before 
term although only in small quantities. 

(2) The amount of enzyme increases gradually to the age of one year, 
when the composition of the child’s saliva becomes identical with that of the 
adult. 

(3) At all ages the quantity of ptyalin present varies with the general 
condition of the child, being larger in strong than in weak infants. 

Inquiries were made among the mothers attending the out-patient de- 
partments as to the nature of the diet administered to their infants, and one 
was surprised to find that a large number of them fed their babies on bread 
soaked in milk and ground rice at the early age of 11 weeks and even younger, 
and barley water was a diluent used by most of them. Quite a number of 
these children were in fair condition and suffering from no digestive trouble. 

According to Halliburton [1913] and others no diastatic enzyme is present 
in the pancreatic juice of infants. 

In view of the fact that an amylolytic enzyme is present in the saliva of 
babies in quite appreciable quantities, and that a number of infants fed on 
milk plus starchy foods manage to utilise a large quantity of the starch, 


SALIVARY SECRETION IN INFANTS 389 


there seems every reason for assuming that diastase is present in the pancreatic 
secretion of infants. 

At the commencement of my investigations it appeared as if children who 
had taken starch in their diet secreted more ptyalin than those who had had 
none, indicating a capacity of adaptation as postulated by Pavlov, but after 
examining a larger number it seemed that the presence of starch in the diet 
makes no appreciable difference in the quantity of ptyalin present in the 
saliva. 


REFERENCES. 


Berger (1899). Inaug. Dissert. St Petersburg. 

Finzi. Rev. Hyg. et Méd. Inf. 8, No. 3, 224. 

Halliburton (1913). Handbook of Physiology, p. 520. 

Ibrahim (1908). Naturforsch. Vers. Kéln. 

Moll (1905). Monatschr. f. Kinderheilk. 

Schiffer, Korowin and Sweifel (1891). Cited in Halliburton’s Chemical Physiology. 


Bioch. xvi 26 


XXX. THE RELATION OF SALIVARY TO 
GASTRIC SECRETION. 


By TOMOICHI NAKAGAWA (Osaka). 
From the Institute of Physiology, University College, London. 


(Received March 1st, 1922.) 


Sativa, besides its main functions in assisting deglutition and exercising some 
amylolytic action, has other functions of secondary importance which also 
play a certain réle in digestive processes. Maxwell [1915] found that boiled 
starch has a marked inhibitory action on peptic digestion, but that the 
products of amylolytic hydrolysis have not this effect. Thus the amylolytic 
ferment of the saliva acquires a new importance by removing the inhibitory 
factor and thus indirectly assisting in the gastric digestion. Maxwell, in all 
his experiments, used commercial pepsin. The object of the experiments 
described in this communication was to repeat Maxwell’s experiments with 
natural gastric juice and to extend them to the rennet action of gastric juice. 


A. SALIVA AND PEPTIC DIGESTION. 


Pure gastric juice was collected by “sham feeding” from a dog which had 
a gastric fistula and an oesophagostomy. Fresh samples were used for each 
experiment. Mett’s method was found to be sufficiently accurate for the 
purpose of these experiments and was used throughout the research. The 
Mett’s tubes were left to digest for ten hours at 40°. All the experiments were 
repeated a considerable number of times, and those given in this communica- 
tion serve only as examples. 

Exp. 1. The seven small flasks containing the solutions named in Table | 
were kept at 40° for one hour; at the end of this time 5 ce. of gastric juice 
were added to each flask, and after another half hour the Mett’s tubes were 
placed i in each sample. In ten hours the readings of the digested tubes given 
in Table I were observed. 


Table I. 


Digestion of Mett’s 
tubes in mm, 


1. 3ce. of 2% boiled potato starch +1 cc. fresh human saliva 53 
2. ” ” ” ” ” qa boiled ” ” 39 
3. ” ” raw ” ” i: op ” ” ” hed 
4. > » soluble starch + 55 Ab * rs 5:3 
5. - »  erythrodextrin + gy a bs ey 5:2 
6. ” ” maltose lpg ” ” ” 54 
7. ” ” glucose TY os ” ” ” 55 


The figure for flask 2 is the only one that shows a deviation from the rest. 
This flask was also the only one that contained the starch in a colloidal form. 
The hydrolysis of the starch in flask 1 was carried to the achromic point. 


GASTRIC AND SALIVARY SECRETION 391 


Exp. 2. In this experiment the actions of hydrolysed and unhydrolysed 
starches of different origin were compared. The different pairs given in 
Table II must not be compared with each other, because, although the hydro- 
lysis was carried on for the same length of time, it reached a different stage 
with each starch as determined by the iodine reaction. 


Table II. 


Digestion of Mett’s 
tubes in mm. 


1. Potato starch hydrolysed 5-3 
”» »,  unhydrolysed 3°9 
2. Maize » hydrolysed 3-0 
= »  unhydrolysed 2-2 
3. Wheat ,, hydrolysed 3:3 
” »  unhydrolysed 2-6 
4. Rice » hydrolysed 5-2 
*9 » unhydrolysed 4-2 


In each case 1 cc. of saliva, boiled or unboiled, was added to 3 cc. of the 
2% solution, kept in the thermostat for one hour, and then mixed with 
5 ce. of gastric juice. The readings were taken at the end of ten hours. 

Maxwell’s observations made with commercial pepsin are extended in 
these experiments to normal gastric juice. 

The theory of adsorption of the pepsin by the colloidal starch advanced 
by Maxwell seems to be the most probable explanation of the experiments 
described. There is one factor of general interest which should be mentioned. 
As has been described by several observers, the gastric juice secreted on a 
carbohydrate food is always the richest in pepsin. On the other hand, carbo- 
hydrates, when in a colloidal form, hinder peptic digestion. This seems to 
be another case of a delicate adaptation of the organism in providing more 
ferment when its action meets with difficulties. 


B. SALIVA AND MILK CLOTTING. 


We have seen that the carbohydrates in a colloidal form behave towards 
normal gastric juice and the commerical preparation of pepsin in the same 
manner. We now proceeded to determine whether the carbohydrates behave 
in-the same way towards the rennet of gastric juice. 

The experiments with milk clotting were performed in the same way as 
those with peptic digestion, but instead of the Mett’s tubes, 5 cc. of cow’s 
milk were added to each test tube and the time required for clotting noted. 

The action of the boiled starch on milk clotting was found to be the 
opposite to its action on peptic digestion. In each experiment the addition 
of starch in a colloidal form produced a considerable acceleration of the 
clotting. Exp. 3 serves as an example. 

Exp. 3. The seven test tubes containing the solutions were placed in a 
thermostat at 40°. After one hour 0-3 cc. of gastric juice was added to each 
test tube and after a further 30 minutes 5 cc. of cow’s milk. The time of 
clotting of each sample is given in Table ITI. 


26—2 


392 ; T. NAKAGAWA 


Table III. Time of clotting 
in mins. 
l. 3 ccc. boiled potato starch +1 cc. fresh human saliva 15 
2. 2? 29 >” ” 5 > boiled > 9° 3 
3. 29 Taw > = o> 9 3? 9° 15 
4. ,, soluble starch +S m rey * 14 
5. ,, erythrodextrin + '55 ne PA % 14 
6. ,, maltose = eae}. % % ” 15 
ie ” glucose Slee ” ” ” 16 


As in the first experiment, the only figure-which deviates from the others 
is the one corresponding to the test tube containing the starch in a colloidal 
form. But in this experiment the difference is in the opposite direction, showing 
a marked acceleration. 

Exp. 4. Some other starches (beans, sago, barley, oats, etc.) were tried, 
and always with the result that the starch in a colloidal form produced an 
acceleration of milk clotting. Table IV gives some of the figures obtained. 


Table IV. 
Time of 
1 ce. Gastric Milk clotting 
saliva juice ce. mins. 

3 cc. boiled potato starch fresh 0:3 - 5 9 
9? 9? 3? 9? boiled > ” 3 
» 9 bean » fresh » ” 8 
9? 9 ”? ”? boiled > > 3 
» 99 +~=«Fice » fresh ” ” 8 
9? ” 9 3? boiled ” > 5 
5 os Maize ...,, fresh es %5 6 
° > ” ”? boiled 9° ”> + 


In all the above experiments human filtered saliva was used. The filtered 
saliva was quite watery and contained hardly any mucin as determined by 
precipitation with acetic acid. The result of the experiments with milk 
clotting indicates that the amylolytic action of saliva indirectly slows the 
clotting of the milk in the presence of colloidal starch. The experiments 
provide one more example of the difference in behaviour of pepsin and rennin. 

Under normal conditions it is rather exceptional for saliva to have this 
effect on milk clotting, and it certainly never occurs in animals deprived of 
the salivary amylase. It is of some interest, however, that even in such 
animals (dog) a fair amount of saliva is secreted on milk although it is of no 
use for its deglutition and has no direct relation to its digestion. In the dog 
the submaxillary saliva secreted on milk is richer in mucin than the saliva 
secreted on any other food. In sucklings there is a profuse secretion of viscid 
saliva. These facts suggest that saliva itself, besides its indirect action de- 
scribed above, has some importance in the digestion of milk. Borisov [1914] 
has shown that in dogs with gastric fistula milk gives finer clots when mixed 
with saliva, and ascribes it to the physical action of the viscous saliva, Allaria 
[1912] found occasionally a slight slowing in milk clotting when mixed saliva 
from an infant was added. 

The following experiments were performed in order to determine whether 
saliva has any direct effect on milk clotting. 

Parotid and submaxillary saliva were collected from a dog having salivary 


a ae 


GASTRIC AND SALIVARY SECRETION 393 


fistulae of the two glands. The saliva was diluted ten times with thrice dis- 
tilled water. To each 5 cc. of the diluted saliva 5 cc. of milk were added and 
then mixed with 0-5 ce. of gastric juice. The time required for milk clotting 
in the thermostat at 40° is given below. 


Table V. Time of clotting 
in mins. 
5 cc. tap water +5 ce. milk 
ry distilled water +5e ce. milk 
ie nade saliva +5 cc. milk as 
»» submaxillary saliva+5cc. milk ... 


submaxillary saliva (10 times Se! + Be ce. milk 
», tap water +5 cc. milk pS 


3. »» pure submaxillary saliva (undiluted) ‘ba ce. adie 
» tap water +5 cc. milk 


The sample of parotid saliva used in iia eeatitistint did not contain any 
mucin, which is often present in dog’s parotid secretion. 

Brennemann [1917] showed that in rapid clotting larger curds are formed 
than when the clotting of milk is slow. As the presence of the viscid saliva 
delays the clotting of the milk, this explains the observations of Borisov 
quoted above. It’ is certain that this effect of the submaxillary saliva does 
not depend on a change in the reaction. The addition of colloidal starch as 
used in the above experiments to acid solutions did not change their py as 
determined colorimetrically. Parotid saliva, being not less alkaline than sub- 
maxillary, has no such delaying effect on milk clotting. The action of the 
viscid saliva seems to be better explained by ascribing a protective action 
to the mucin. 


ms wa Qwtyr 


CONCLUSIONS. 


1. Maxwell’s experiments on the inhibitory effect of colloidal starch on 
peptic digestion are confirmed and extended to natural gastric juice. 

2. The amylase of saliva hydrolysing the starch removes its inhibitory 
effect on gastric digestion. 

3. Pepsin and rennin behave towards colloidal starch differently—the 
action of the pepsin is inhibited by colloidal starch and the action of the rennin 
accelerated. 

4. The amylase of saliva hydrolysing the starch removes its accelerating 
action on rennin. 

5. The saliva has a delaying action of its own on milk clotting; this action 
was found to be due to the protective action of mucin. 

6. The secretion of saliva in sucklings is important because it delays the 
milk clotting and thus produces finer and more easily digested curds. 


I wish to express my thanks to Dr G. V. Anrep for his kind help and 
criticism during my experiments. The gastric juice and saliva were collected 
from dogs operated on by Dr Anrep. 


REFERENCES. 


Allaria (1912). Zentr. Biochem. Biophys. 18, 536. 

Borisov (1914). Quoted after Babkin’s Die dussere Sekretion der Verdauungsdriisen, 8, 19. 
Brennemann (1917). Arch. Pediatrics, 34, 81. 

Maxwell.(1915). Biochem. J. 9, 323. 


XXXI. THE RELATION OF THE FAT SOLUBLE 
FACTOR TO RICKETS AND GROWTH IN PIGS. IL. 


By JOHN GOLDING, SYLVESTER SOLOMON ZILVA, JACK CECIL 
DRUMMOND anp KATHARINE HOPE COWARD (Beit Memorial 
Research Fellow). 


From the National Institute for Research in Dairying, Reading, the Biochemical 
Department, Laster Institute, the Institute of Physiology, University College, 
London. 


(Received March 14th, 1922.) 


THE experiments described in this communication form a part of an inquiry 
now in progress, the main purpose of which is to study the part played by the 
accessory factors in the nutrition of agricultural stock and in the etiological 
factors concerned in the causation of rickets in pigs. In a previous communi- 
cation [Zilva, Golding, Drummond and Coward, 1921] in which the etiology 
of rickets in pigs was investigated we have shown that a rigorous elimination 
of vitamin A from the diet of young pigs did not conduce to the production 
of well-defined rickets. We were, however, able to demonstrate in those 
experiments that such a dietetic deficiency has a very marked effect on the 
development of these animals. Indications were also obtained that the de- 
privation of this dietetic principle possibly has some bearing on the production 
of healthy young. 

Having been unsuccessful in producing rickets in pigs experimentally by 
depriving them of vitamin A alone we next attempted to ascertain whether 
a dietetic deprivation of calcium and vitamin A would lead to the production 
of the disease. The results obtained by us form the subject of this communi- 
cation. 


EXPERIMENTAL. 


Our animals were divided into four groups. Group I received a diet de- 
ficient in vitamin A and in caleium (— A — Ca), group IT received a diet 
deficient in vitamin A (— A + Ca), group III received a diet deficient in 
calcium only (+- A — Ca), group IV received a diet containing calcium and 
vitamin A (+ A + Ca). Two pigs were placed in each group. Group I con- 
sisted of two boars, group II of one boar and one sow, group III of two sows 
and group LV of two boars, The young pigs, which belonged to the same 


RICKETS AND GROWTH IN PIGS 395 


litter, were started on their special diets when they were 53 days old. They 
were young Berkshires, born of a sow 14 months old. The sow was kept for 
some time on a diet of toppings and whey deficient in vitamin A and mani- 
fested this deprivation by retarded growth [Drummond, Golding, Zilva and 
Coward, 1920]. On supplementing the above diet with lucerne the animal 
resumed growth and continued an apparently normal existence. After 74 days 
of correct feeding she was served by a pedigree boar and soon after was 
placed again on a diet of toppings, whey and swedes. This diet which was 
poor in vitamin A was purposely planned as we did not desire the young pigs 
to be born with a store of the vitamin. After 116 days nine pigs were born 
and at the age of 65 hours were divided into two sections. The chart below 
shows the history and subsequent allocation of the pigs. One lot received 
its supply of vitamin A entirely from the mother, the other received additional 
vitamin A in the form of cream (} oz. per day) and eventually in the form 
of cod liver oil (4 rising to 1 oz. per day). Two of the animals in the former 
section received an equal supply of additional oil in the form of inactive olive 
oil, made into an artificial cream at first and later given to balance in nutrients 
the cod liver oil given to Section II. 

The two remaining pigs in the first section received only the sow’s milk; 
the sow being fed all the time on a diet shown to be poor in the vitamin A. 


Sow on a diet deficient in vitamin A from weaning to 7 months old 
Normal diet given for 74 days during which growth was restored 
Served by boar at 94 months old. Diet again restricted in vitamin A 
Nine es born after 116 days. Eight lived 


itamin deficient Vitamin A 
groups groups 
No addition to Olive oil iateiand later 
sow’s milk cod liver oil given 
y 
After 53 days After 53 roads After 53 days 
= 
j 1 
Y { Y 
Group I Group IT Group III Group IV 
—vit. A —Ca —vit. A+Ca +vit. A —Ca +vit. A +Ca 


As soon as it was convenient the young pigs were given additional toppings 
which at the end of this preliminary period of 53 days (period I) reached a 
ration as high as 3 lbs. per day for the litter of eight pigs. The actual intake 
of food and the corresponding increase in weight for the entire experiment 
is summarised in Table I. The growth of the animals is graphically represented 
in Fig. 1. The average daily gains in pounds for period I were for pigs in 
section I receiving no oil 0-533 and 0-514, for those receiving olive oil 0-377 
and 0-5, while the daily gains of the pigs in section II receiving cod liver oil 
were 0-481, 0-453, 0-344 and 0-509. 

The agreement between the rates of growth of the pigs in the two sections 
is made more evident by employing the formula advised by R. A. Fisher, 


LEZ PLE 
LET =< 
Ves MOG 
0 GE-€ 9 G2-§ 
CLF CLF 
8-LFT 9FI 
Zerl €F-1 66-0 Itt 
SZ-0€ IE CZ-81 G-EZ 
SZ-6E1 GOIL &6-€6 &Z-801 
ZOE 0-F 
are ah 
VoL MOG 
09-¢ 09-¢ 
269 19 
8o-LST ect 
66-0 T@1 16-0 $8-0 
6Z F:6E 0G 9-12 
66 cs GL GL‘F8 
GG { 0-€ 
V 98-9 AN > 
O-% ae 
9 90-9 9 90-9 
Oo Lg OX Lg 
OLT OLI 
LoL 08-1 FET L¥-T 
OF CZ9-FE 90°8Z SZ9-FE 
OL GZ9-G¢ 9¢-F¢ SZI-€9 
\ ee Nae oe eX a vi) 
oyeydsoyd pus uIntoyeo + ‘ UENTOTeO MO'T 
VY uroreyrA + 
JOATT POD IOAT] pO—) AAAI, pog JOATT pod 
$e €8-E oh oG-F 
609-0 PFE-O ESF-0 I8t-0 
LZ GZ-81 O-F% GCF 
€ CLS GZ O-€ 
0€ 1G g-9Z G8Z 
‘8 ‘s *sq[ “Sq| *Sq| “sq| 
aoe sSrd 1og iwog MOS stid g 10g ~=MO0g 
, SEY a P| \- a) 
AT dnory TIT dnoay 


€8°3 6E- 
ii = 
vé M CLT 
0 86-E 0 Go-1 
6°€ 1 
GF 69-66 
89-0 99-0 8L-0 
£6 S66 SZ-0L 
g18 GZ-89 Se-e9 
oF 9-F1 
are red 
VoL} MELT 
09-¢ O8<Z 
H L-G H G8-Z 
III SSS 
IL-0 IL-0 800 Bd Treg ZO 
GL-ST &1 ST co-F 
GL°IL 62 = 486-GE 0-¢¢ 
8-€ FF 
V 98-9 baid 
OG ne 
0 90-9 0 90-9 
HL HL-¢ 
OLT OLI 
etl ee-1 86-0 66-0 
G96 GSS 1g 88-06 
99 oF 1¢ 8&-0¢ 
, ee “7 y | |= ~y ae 
eyeydsoyd puv untoyeo + UWINIO[BO AO'T 
ee a 
V uroreztA — 
euON eATTO ouON eATlO 
16°F FL 0°-F I¢-F 
€ES-0 LLE-0 FIS-0 ¢-0 
GB-8E 0-06 SG-LE £93 
SBS Qs CLs Og 
9-08 iret 0& £63 
‘Sq ‘ ‘sql sq] ‘sq sq 
reog = said is ray a aiid ~ iq swog 
Nica _ J Neen v ani 
Tr dnory T dnory 


T orne@eyT 


5 ee ae 


ony, ‘uedourese) 
iayyeu Arp ‘ssurddoy, 
OBI YRMOIS OAYRIOY 
sXep ZZ Ul ures) 
puzz “ydeg uo yy F104 
ystems 
ot UT sevolour “q[ 
ioy3eu1 Arp spunog 
4TeGO 
a 
uaSourese 
103yeur Axp ‘sdurddoy, 
9781 Y}AMOIT OAL Y 
sXep cg Ur ures) 
STE “ny wo FA 
ysIem 
OAY] UI osvalOUI “q] | 
zed 19}3vu1 Arp spunog 
Teooregy 
34) 
30 tO 
uasourese) 
zoyyeur Arp ‘sfurddoy, 
O7UI YIAOIT OAHVOY 
sep #¢ Ur UIeD 
W32z Ajur uo Wyse 


WoAls TO 

QIU YPAOIT OAHVPEY 

Aep sod “sqj ur ures) 

sep ¢¢ Ul UTED 
Windy = * 

Pag GUNS UO BqTIOA\ 


pag ounr 

WIT Edy 

AOS OU} UITAL 
[ poueg 


‘dn pouvofo sXvaye Jou orem YOryA YLeyYO pure [wOoTeYS Jo ssoy 4ysureSe yos ore oyrULIeUT JO Sosop [eUOIsv000 9Y} pure oornl UOMO, 
jo osop A[rep [[vuts oy, ‘punod v jo speutoop pue spunod ur possoidxe pue ueals syuorperZur oy} UO pozefNoyeo st pouINsuOO 10}4Ur Arp [8404 ey, “ALON 


“Teorey 
PooM=A “TwoorwyH [BUY =V poUDYD “TIO PANO PATJoBUT=O “TIO JOAYT poQ=o "NO “poywoH=H ‘poywoquQ=n wahowssyg *yX¢g ‘Buy uo [ dnoss ‘Brg jo WYFIEM = » 


VysIom 
OAT] UT OSBEIOUT “q] | 
SE 6-F LF 6-8 red 1033e01 Arp spunog 
1 os 1 mIe4O 
VOG MOG V9G M ¢-T yeooreyy 
es 9 Ta Dae 0 
2 6E 2 6-€ H 6-€ H C6-T é uos0uTesey) 
LOL Ost Lg 3-93 qoygvur Arp s,Burddoy, 
69-0 OLT 48-0 9LL-0 96-0 19-0 02-0 OBI YYMOIS OATFRIOY 
G1Z ‘ ce SB-81 g-61 o9 96-6 SSL sXep 0% Ul ures uoZ0 
LOL Z9L = SB o-98I £6 GL-08 G3-9L PUZZ “390 UO FYSIEM -UTosBO poyeoH] (2) 
qysIom 
OAT] UL OS¥BOTOUL “I T 
I&€ €-8 69-1 9z-T _ ded 10yge01 Arp spunog 
81-0 — 8L-0 ee yreqD 
Vv LE-0 M LE-0 V LE-0 MZ0 yeooreyy 
9 8F-0 9 8F-0 0 8F-0 O 2-0 TO 
O 9¢-0 9¢-0 a 9¢-0 2 8z-0 ueZouresey) 
9-93 GE-LS 96-L *9-F royyeur Arp ‘sBurddoy, 
18-0 1€-1 0¢-0 1¢-0 SLI L-0 IL OVI YAMOIS OAIZL[OY, 
GE g oT GL-1 oF oT CS-F sep g Ul urexy) wueZo0UTseO 
G-CFl Ost 101 9Il S-98 GIL 69 PUZ “990 UO 4YDIOM peyeoyuy) (9) 
qySIoM OAT] UE 
esvarout “q] [ ted pooy 
G6 COP SL-L a ur 1oyyeur Arp spunog 
£F-0 re cF-0 boot: AT®4GO 
V 18-0 M 18-0 Vv L8-0 M FF-0 Teoorey) 
9ST 9 SLT CCL T 9 Z¢-0 TO 
aT QE1 HE€-T HZ9-0 uesouleseyR ; 
IF-8¢ 80-L¢ GL-E1 Z-01 roygeut Arp ‘sdurddoy, woo 
FET 0-1 36-0 LL-O 60-0 98-0 880'T OWI YMOIS OATVV[OY -UlosvO poywoFT (”) 
CL-ZI og 93-9 0-9 c-0 CLT ¢-0 - sXep p Ur ures) TI pus J sdnorgy 
GPt Sel S66 Se-FIL és OL SL-¥9 496% “dog uo 4yFI0 A potieg 


J. GOLDING AND OTHERS 


398 


Period V 


me a ew a a 3 oe few ee me ee ele rn en 


ee 


Group I 


€---30lis—- — — > 


i " A oi 


ant 
a Pe ce oo 


Period II 


Period I 


Oct.8 18 


Is 


Sept.8 


a8 


24) 


‘ 
4 


30 Aug.9 19 


30 July 10 20 


SL Juneld 20 


21 


21 May! 


Aphl 


Fig. 1. 


RICKETS AND GROWTH IN PIGS 399 


Chief Statistician, Rothamsted Experimental Station, for calculating the re- 
lative growth rate per day per cent. using natural logarithms, viz.: 
log, M, —log, M, 100 


K, i Ky 
where M, = the weight at the end of the period, 
M, = a x commencement of the period, 


K, — K, = duration of period in days. 

The following figures are obtained for Section I (— A) without oil 4-91 
and 4-50, with olive oil 4-14 and 4:31, for Section II (+ A) 4:25, 4:45, 3-83 and 
4-34. 

It is evident from the above figures that a sow, even when fed on a diet 
deficient in the fat-soluble factor and having undergone a previous depriva- 
tion in this factor, is capable of rearing her young satisfactorily. This confirms 
further our earlier observation that the requirements of the pig for the fat- 
soluble factor are not of a high order. 

At the end of this period, namely 53 days after birth, the animals were 
weaned and placed on their respective experimental diets. Each group was 
placed in a separate sty. The sties faced south and were divided by wooden 
partitions, the floor being partly wood and partly concrete. The bedding con- 
sisted of wood shavings or sawdust. _ 

The basal diet for all groups consisted of toppings or wheat offal having 
the following average composition: 


Moisture 11-51 % Mucilage 59-60 % 
Oil 4-29 Woody fibre 5-86 
Protein 14-81 Ash 3°93 


The dry matter of the above contained 0-338 °% calcium. Besides the basal 
diet the pigs received supplementary protein in the form of caseinogen, which 
in the case of the vitamin-free diets was previously inactivated by being 
heated for 24 hours at 120°C. This inactivation was carried out for us by 
Dr R. T. Colgate in Messrs. Huntley and Palmer’s Laboratory, for which we 
wish to express our thanks. 

The food was weighed out three times a day at 8a.m., 12 noon and 
4.30 p.m.; it was mixed with cold water just before feeding and given in the 
form of a thin cream. The food was given on a live weight basis, being 
regulated to the quantity which the pigs would clean up. 

The other two accessory factors, namely the antiscorbutic and the anti- 
neuritic factors, were supplied in the form of freshly prepared lemon juice 
and marmite (Commerical Yeast extract), about 7 cc. of the former being the 
daily dose, whilst $ oz. of the latter was given at intervals. Groups III 
(+ A — Ca) and IV (+ A + Ca) received a daily dose of cod liver oil as a 
source of the fat-soluble factor whilst the other two groups received an 
equivalent dose of inactive olive oil. Groups II and IV also received a daily 
dose of 1 oz. each of precipitated chalk and 1 oz. animal charcoal containing 


400 J. GOLDING AND OTHERS 


67-3 per cent. of calcium phosphate as a source of additional calcium and 
phosphate; the other two groups received only the small amounts of calcium 
from the basal diet. 

During the following 54 days (period II) the intake in all the four groups 
was the same. The increase in weight in the animals of the various groups 
however showed a marked disparity. The two animals in group I gained 
41-88 lbs., in group IT 49 lbs., in group III 62-7 Ibs., and group IV 74-6 lbs. 
The relative growth rates, as calculated from the formula given above, were 
group I 0-99 and 0-98, group IT 1-32 and 1-13, group III 1-47 and 1-34, and group 
IV 1-80 and 1-57. Theaverage dry matterin the food consumed per pig for each 
pound of gain in live weight was: group I 4-4 Ibs., group II 3-8 lbs., group III 
3-0 lbs. and group IV, 2-5 lbs. By this time very marked differences in the 
general appearance of the animals in the respective groups could be discerned. 
The animals in group I developed a scurfy skin and saddle back, weak legs and 
joints painful to pressure. They were easily tired and were not playful. Those 
in group II also showed lack of vitality and a saddle back, those in group III 
were in good condition possessing glossy coat, whilst the animals in group IV 
were decidedly in the best condition. One animal in group I received an 
injury from a fall and died 87 days after the commencement of the experi- 
‘ment. At the post mortem examination it was found that the vertebral 
column was broken. 

During period III, 7.e. 35 days following period II, the increase in weight 
wasas follows: group 14-62 Ibs. (one animal), group II 28-75 lbs., group III 42 lbs., 
and group IV 58-4 lbs., and the average dry matter in the food consumed 
per pig for each pound of gain in live weight was: group I (one animal) 
14-6 lbs., group IT 4-5 Ibs., group III 4-0 Ibs. and group IV 3-02 lbs. : 

Owing to the low condition of the remaining animal in group I it was 
decided to administer a small amount of the fat-soluble factor in order to 
save the pig. This was done by introducing caseinogen, which had not been 
previously inactivated by heating, in the diet during period IV. Thisaddendum 
had its desired effect and the animal responded after about five days by 
resuming growth. 

During this period of 22 days (period IV) the animals gained in weight as 
follows: group I (one animal) 10-25 lbs., group II 19 lbs., group III 41-7 lbs., 
and group IV 61-7 Ibs. In the last period (period V) the inactivated caseinogen 
was alternated with crude caseinogen in order to keep the weight of the 
animals in group I and group II in check. 

The experiment was terminated 145 days after its commencement. The 
animals were slaughtered with the exception of one sow in each of the groups 
If and ILI. The condition of the animals before slaughter was as follows: 
groups I and II wrinkled skin, ears carried forward, down on hind legs; 
group III skin rather rough, lack of size and bloom, flesh not quite firm, 
otherwise normal; group IV skin healthy and animals perfectly normal. 

The photographs of the pigs before slaughtering are shown in Plate I, 


RICKETS AND GROWTH IN PIGS 401 


two photographs having been taken of the pig in group I. The photographs 
were taken to scale. 

The post mortem examination revealed no abnormalities of the organs 
beyond that the ribs of the animals in group I were easily fractured. It is 
also to be reported that the fat of the pigs of groups I and II was, in the 
butcher’s opinion, softer and not of such good quality as the fat of the other 
pigs. This was confirmed by the estimation of the melting and solidifying 
points of the fats. — 


Melting points Solidifying points 
Group I 21° 22-9° 19° 
ms pS yf hag 23-8° 19-3° 
Oa PEE: BB 25-20° 22° 
oe ee ts RD 32-8° 22-5° 


Table II gives the size, weight, breaking points, and calcium content of 
the bones. 


Table II. 

Weight of Distance of Breaking Same CaO 

humerus bearing points weight corrected to percentage of 

in g. in ine in tons 34” length dry matter 

Group I 103 3 0-210 0-210 29-52 
ve Ir 127 3 0-345 0-345 29-05 
Peete? i! 117 4 0-292 0:336 33-52 
sg) ot Oh 152 4 0-466 0-532 38-08 


No abnormal flavour or taste was reported by a number of people who 
consumed the joints derived from the animals fed on the cod liver oil. 

The fifth, sixth and seventh ribs of one animal in each group were examined 
for us histologically by Prof. V. Korenchevsky to whom we are also indebted 
for the interpretation of the results. The following is a summary of the 
observations made: 

(1) Group IV (+ A + Ca) showed a somewhat abnormal picture with very 
slight osteoporosis and a belated deposition of lime salts in the newly formed 
bone. 

(2) In all cases the bone marrow especially in the region of the secondary 
spongiosa consisted of a fine fibrous reticulum with but few bone marrow cells. 

(3) The histological condition in animals belonging to groups II (— A+ Ca) 
and ITI (+ A — Ca) was not very different from that of group IV (+ A + Ca). 
Only a higher degree of osteoporosis resulting from a diminished activity of 
osteoblasts could be seen. 

(4) In group I (— A — Ca) the condition of osteoporosis was more pro- 
nounced. Moreover a more frequent incursion of the proliferating cartilage 
into the bone marrow was in evidence. In these places was also noticed 
defective calcification in the zone of provisional calcification. 

It is quite evident that in spite of the very marked changes which have 
been effected by our restricted diets, no rickets in the pathological sense of 
the word has been induced. The animals in groups I and II have on various 


402 J. GOLDING, AND OTHERS 


occasions during the experiment displayed a condition which would have 
been described by the practical man as the pigs being “off their feet.” No 
doubt such a condition has been before now loosely referred to as “rickets.”’ 
Although defective calcification was found in the zone of provisional calcifica- 
tion in the case of group I (— A — Ca) no increase in the amount of osteoid 
tissue could be established and therefore no faulty deposition of calcium in 
the newly formed bone in the sense of rickets can be asserted. ; 

A remarkable feature in our experiments is that even in the case of group IV 
(+ A + Ca) which acted as our control, a normal histological picture was not © 
obtained. This requires further investigation. Possibly the restricted diet of 
the mother may be responsible for this. Another point to be considered is 
that the animals consuming the calcium-deficient diet received 0-338 °% of 
calcium in their food. With animals, such as pigs, which consume large bulks 
of food it was difficult from a technical point of view to reduce this calcium 
intake. However the calcium deficiency was definite enough to impair the 
growth and diminish the calcium content in the bones of the animals in 
group III (— Ca + A) and it is very doubtful whether pigs which develop 
rickets under farm conditions, receive a diet much poorer in this element. We 
refrain from reviewing the literature which has appeared during the last few 
monthsin connection with the etiology of rickets. Most of the experimental work 
was done on rats and the results and conclusions of the various investigations 
are conspicuously contradictory. We cannot however conclude without briefly 
referring to the results obtained by Korenchevsky [1921]. This investigator, 
working with rats, obtained a definite condition of rickets on a diet free from 
the fat-soluble factor and calcium. Whether our apparently different results 
were due to the higher content of calcium in our experimental diets, or whether 
it was due to the different character of the experimental animal will most 
probably be decided by future investigation. We are continuing our experi- 
ments and although the dietetic hypothesis of the etiology of rickets forms 
our main line of work, we are not excluding such a hypothetical factor as light, 
especially in view of the latest work of Hess and Unger [1921], of Powers, 
Park, Shipley, McCollum and Simmonds [1922]. 

The expenses of this research were defrayed from a grant made by the 
Medical Research Council, to whom our thanks are due. 


REFERENCES. 
Drummond, Golding, Zilva and Coward (1920). Biochem. J. 14, 742. 
Hess and Unger (1921). Soc. Lxp. Biol. Med. 18, 298. 
Korenchevaky (1921). Brit. Med. J. 547. 
Powers, Park, Shipley, McCollum and Simmonds (1922), J. Amer. Med. Ass. 78, 159, 
Zilva, Golding, Drummond and Coward (1921). Biochem. J. 15, 427. 


BIOCHEMICAL JOURNAL, VOL. XVI. NO. 38 PLATE Il 
PHOTOGRAPHS OF PIGS TAKEN ON OCT. 1l0ru, 1921 


Group I — Vitamin A. 


OE | } 
Ace rehearse tn peel 4 RON Wa ge 


j 

1 ’ 
} | | 
; 


Group IL — Vitamin A +Calcium and Phosphate. 


| 
| 
| 


Group IV + Vitamin A +Calcium and Phosphate. 


/ | | 
} | , 


} 


BIOCHEMICAL JOURNAL, VOL. XVI NO.8 PLATE Ill 


\< 


Pigs from Groups [V, III, I, I. 


Cross sections of pigs from Groups IV, III, I, I. 
Group I —Vitamin A. 
Group II — Vitamin A + Additional Calcium and Phosphate. 
Group III +Vitamin A. 
Group IV +Vitamin A + Additional Calcium and Phosphate. 


XXXII THE ESTIMATION OF CALCIUM 
IN BLOOD’. 


By ARTHUR ROBERT LING ann JOHN HERBERT BUSHILL. 


From the Department of Biochemistry of Fermentation, 
University of Birmingham. 


(Received March 17th, 1922.) 


INTRODUCTORY. — 


Tue estimation of calcium in blood, serum etc., is a matter of extreme delicacy 
seeing that, in the case of human blood at least, it is not often easy to obtain 
more than 1 to 5cc. of the sample. Working with such small quantities 
therefore it is necessary to employ a method capable of measuring calcium 
to a limit of accuracy of at least 0-006 mgm. It is obvious also that in these 
circumstances the reagents employed must.be of the highest purity, especially 
so far as their freedom from calcium is concerned. 

Among published methods are some based on the use of the haemacyto- 
meter (counting the calcium oxalate crystals) and on principles, such as 
(1) the determination of the concentration of calcium ions necessary to produce 
clotting of the blood, (2) mixing the blood with ammonium oxalate solutions 
_made isotonic with sodium chloride, and calculating the calcium from that 
concentration of oxalate which just prevents clotting, (3) mixing the blood 
with sufficient ammonium oxalate to prevent coagulation and subsequently 
determining the quantity of calcium chloride necessary to produce clotting. 
We agree with de Waard [1919] that these give comparative rather than 
absolute results. Recourse must therefore be had in our opinion to chemical 
methods. : 

McCrudden [1909] suggests a gravimetric method. The calcium is precipi- 
tated as oxalate in a solution in which the hydrogen ion concentration is so 
adjusted that the iron and phosphates remain in solution. 

Halverson and Bergeim [1917] propose to estimate calcium in blood 
which has not been incinerated, but from which the proteins have been pre- 
cipitated by picric acid. The precipitation of the calcium is carried out under 
McCrudden’s conditions, the hydrogen ion concentration being adjusted to 
the point at which alizarin gives a violet coloration. The calcium oxalate is 
washed by centrifuging and after dissolving it in dilute sulphuric acid the 
solution is titrated with N/100 potassium permanganate. This method, ac- 
cording to the examples shown in the paper, gives results agreeing closely, 

1 This investigation was undertaken at the request of my colleague, Prof. Carlier. (A. R. L.) 


404 A. R. LING AND J. H. BUSHILL 


i.e. within the limit of experimental error, with similar estimations made on 
the ash of the blood, ete. 

Marriott and Howland [1917] have devised an ingenious method for the 
estimation of calcium and magnesium in blood serum. It depends on the fact 
that solutions of ferric thiocyanate are decolorised by oxalates and by phos- 
phates. Calcium is precipitated as oxalate and magnesium as ammonium 
magnesium phosphate. The precipitates are dissolved in acid, and added to 
solutions of ferric thiocyanate, the degree of decolorisation being determined 
by comparison in small Nessler tubes. It is said to be possible by this method 
to work on 2 cc. samples of serum with a maximum error of less than 5 %. 

Jansen [1918] describes a gravimetric method of estimating calcium. He 
incinerates the blood; and he considers the removal of iron and phosphoric 
acid necessary before precipitating the calcium as oxalate. . 

De Waard [1919] has devised a method for the estimation of calcium in 
organic substances for which he claims an accuracy of 4%. The substance 
is incinerated, the ash dissolved in hydrochloric acid in a special tube, 
drawn out to a capillary end, which is subsequently used for centrifuging 
the liquid when the calcium has been precipitated. The tube containing the 
hydrochloric acid solution of the ash is placed in a bath of boiling water, 
0-5 cc. of saturated ammonium oxalate solution added, then a slight excess 
of ammonia, and finally a slight excess of acetic acid. In these circumstances 
he finds that the iron, magnesium and phosphates remain dissolved, whilst 
the calcium is precipitated quantitatively. 


EXPERIMENTAL. 


The method we have adopted is based on the methods of McCrudden, 
Halverson and Bergeim, and de Waard. : 

The blood is incinerated in a platinum dish, the ash dissolved in concen- 
trated hydrochloric acid, and the solution washed into a centrifuge tube of 
special shape. In this tube the calcium is precipitated as oxalate as described 
later. When the liquid is centrifuged, the precipitate collects in the narrow 
tube leaving the supernatant liquid clear. This liquid can then be removed 
by a tube drawn out to a capillary without disturbing the precipitate. The 
calcium oxalate is washed twice in this way, dissolved in dilute sulphuric 
acid and the solution titrated with N/100 potassium permanganate. 

The shape of the centrifuge tube is shown in Fig. 1. It has the following 
dimensions: diameter ‘+5 em., total length 12 em., length of narrow tube 
1-3.cm., bore of capillary tube 0-5 cm. The tube is graduated for 2 ce. and 
for 25 ec. A glass rod capable of entering the narrow tube (as shown) is 
used for stirring the precipitate after the addition of sulphuric acid before 
titrating. The advantage of this centrifuge tube is that it is of such dimen- 
sions that all the operations can be performed in it. It is large enough for 


_ spongy mass.of carhonaceans mattaniatlate 


ESTIMATION OF CALCIUM IN BLOOD 405 


the precipitation to be carried out, the narrow tube at the base enables the 

precipitate to be washed easily without 

being disturbed, and finally the end point <--tdcms-—> 

can be seen easily when titrating as the ages 

tube does not taper gradually to a point. 

The solutions may be titrated to 0-Olcc. <A f | 
The procedure is as follows: 
Incineration. About 2-5 cc. of blood is | 

| 
1 


measured into a platinum dish. This is 
heated on a piece of wire gauze with a 
coating of asbestos, by a very small flame 


H— 25 C8. 
until all the water has evaporated and a 


ERRATUM 
Tue Estimation of Catctum 1x Broop (Lina AnD BusHIt.) 


The diameter of the centrifuge tube should be 2°6 cm. and not 1°5 as stated in the text 


and diagram on p. 405. 


mr une esulmation OT caiclum in blood. 


McUrudden’s method as modified by Hal- 
verson and Bergeim. The solution in the centrifuge tube is neutralised 
with 0-880 ammonia using alizarin as indicator. The neutral point is overshot 
and then titrated back with N/2 HCl until the colour just commences to 
change. Then 2-5 cc. of N/2 HCl and 2-5 ce. of 2:5 % oxalic acid solution are 
added and water to make to the 25 cc. mark. The tube is then placed in a 
water-bath and brought to the boiling point. To the hot solution 5 cc. of 
3 % ammonium oxalate is added and the tube kept at the boiling point for 
15 minutes when it is removed, placed in ice water and 5 cc. of 20 % sodium 
acetate solution added (or until the colour commences to change). When 
the calcium oxalate has precipitated which takes about 6-8 hours (it is usually 
allowed to remain over night), the solution is centrifuged. The clear super- 
natant liquid is removed by means of a tube (drawn out to a capillary) leading 
to a bottle connected with a filter pump. By this means the liquid is removed 
to the 2 cc. mark on the tube. The outside of the capillary tube is washed 
with water to prevent the loss of any precipitate which may adhere to it. 
Bioch. xv1 27 


494; 


404 A. R. LING AND J. H. BUSHILL 


z.e. within the limit of experimental error, with similar estimations made on 
the ash of the blood, etc. 

Marriott and Howland [1917] have devised an ingenious method for the 
estimation of calcium and magnesium in blood serum. It depends on the fact 
that solutions of ferric thiocyanate are decolorised by oxalates and by phos- 
phates. Calcium is precipitated as oxalate and magnesium as ammonium 
magnesium phosphate. The precipitates are dissolved in acid, and added to 
solutions of ferric thiocyanate, the degree of decolorisation being determined 
by comparison in small Nessler tubes. It is said to be possible by this method 
to work on 2 cc. samples of serum witha maximum error of less than 5 %. 

Jansen [1918] describes a gravimetric method of estimating calcium. He 
incinerates the blood; and he considers the removal of iron and phosphoric 
acid necessary before precipitating the calcium as ovalota 


---vae wee aruspoun, aud ae wWaard. 

The blood is incinerated in a platinum dish, the ash dissolved in concen- 
trated hydrochloric acid, and the solution washed into a centrifuge tube of 
special shape. In this tube the calcium is precipitated as oxalate as described 
later. When the liquid is centrifuged, the precipitate collects in the narrow 
tube leaving the supernatant liquid clear. This liquid can then be removed 
by a tube drawn out to a capillary without disturbing the precipitate. The 
calcium oxalate is washed twice in this way, dissolved in dilute sulphuric 
acid and the solution titrated with N/100 potassium permanganate. 

The shape of the centrifuge tube is shown in Fig. 1. It has the following 
dimensions: diameter ‘t5 em., total length 12 cm., length of narrow tube 
1-3.cm., bore of capillary tube 0-5cem, The tube is graduated for 2 cc. and 
for 25 ce. A glass rod capable of entering the narrow tube (as shown) is 
used for stirring the precipitate after the addition of sulphuric acid before 
titrating. The advantage of this centrifuge tube is that it is of such dimen- 
sions that all the operations can be performed in it. It is large enough for 


ESTIMATION OF CALCIUM IN BLOOD 405 


the precipitation to be carried out, the narrow tube at the base enables the 

precipitate to be washed easily without 

being disturbed, and finally the end point <—-iscmer =? 

can be seen easily when titrating as the 

tube does not taper gradually to a point. 

The solutions may be titrated to 0-01 cc. 
The procedure is as follows: 


Incineration. About 2-5 ce. of blood is 
measured into a platinum dish. This is 
heated on a piece of wire gauze with a 
coating of asbestos, by a very small flame 
until all the water has evaporated and a 
spongy mass of carbonaceous matter is left, J 
when it is heated more strongly. During , FH --I's ol 
the final heating the dish is placed on a 
pipe-clay triangle and brought to redness. 
To prevent extraneous matter from gaining 
access to the dish during the first part of - 
the incineration, an inverted funnel is fixed 
over it by a clamp. The residue is treated 
with 1-5 cc. concentrated hydrochloric acid _ | 
0-5 ce. at a time—each time the solution | 
is gently warmed and washed into the | 
centrifuge tube. The dish is then washed 
three times with water. Care must be taken | 
in the operation to keep the volume to a Page 
minimum. OScms. — 

Precipitation. This was carried out by Ms: began nome, —e Ss 
McCrudden’s method as modified by Hal- 
verson and Bergeim. The solution in the centrifuge tube is neutralised 
with 0-880 ammonia using alizarin as indicator. The neutral point is overshot 
and then titrated back with N/2 HCl until the colour just commences to 
change. Then 2-5 cc. of N/2 HCl and 2-5 cc. of 2-5 % oxalic acid solution are 
added and water to make to the 25 cc. mark. The tube is then placed in a 
water-bath and brought to the boiling point. To the hot solution 5 ce. of 
3 % ammonium oxalate is added and the tube kept at the boiling point for 
15 minutes when it is removed, placed in ice water and 5 cc. of 20 % sodium 
acetate solution added (or until the colour commences to change). When 
the calcium oxalate has precipitated which takes about 6-8 hours (it is usually 
allowed to remain over night), the solution is centrifuged. The clear super- 
natant liquid is removed by means of a tube (drawn out to a capillary) leading 
to a bottle connected with a filter pump. By this means the liquid is removed 
to the 2 cc. mark on the tube. The outside of the capillary tube is washed 
with water to prevent the loss of any precipitate which may adhere to it. 

Bioch. xv1 27 


o 


a8 25 ccs. 


——--—--—---> 


12-0 cms 


20S. 


406 A. R. LING AND J. H. BUSHILL 


Water is then added to the 25 ce. mark, the centrifuging repeated, and the 
supernatant liquid removed as before to the 2 cc. mark, etc. Itis washed twice 
in this way, when 4 cc. of 5 % sulphuric acid are added and the precipitate 
stirred with the glass rod. The tube is then placed in a water-bath at 65° C. 
for 1-2 minutes, when the oxalic acid is titrated with N/100 potassium per- 
manganate by means of a burette capable of reading to 0-01 cc. lcc. of 
N/100 potassium permanganate is equivalent to 0-020 mgm. of calcium. 

A “blank” experiment must be performed on the materials. Coagulation 
of the blood was prevented by the addition of sodium citrate to the extent 
of 2-5 %. 

Solutions used. All materials used were recrystallised twice from water 
which had been redistilled twice from a Jena glass flask, and such water was 
used throughout the operations. It was found necessary to have the end of 
the burette drawn out to a fine point and vaseline was placed on the outside 
so that it was possible to reduce the volume of one drop to 0-01 ce. 

The permanganate solution. About 0-5 g. of potassium permanganate 
crystals is dissolved in 1 litre of redistilled water in a litre flask. The flask is 
placed on the water-bath for 2-3 days, a funnel and watch glass being placed 
in the neck to prevent extraneous matter from entering. The solution is then 
carefully filtered through an ashless filter paper on a Buchner funnel to 
remove the deposit of manganese dioxide. The solution should be kept in 
the dark. It is standardised against an oxalic acid solution prepared by 
dissolving about 0-25 g. of recrystallised oxalic acid in-250 ce. of water. When 
standardising allowance must be made for the titre of the sulphuric acid used. 
It is as well to standardise the permanganate about once every fortnight. 

Alizarin solution, An aqueous solution of 0:2 % strength was used. 


The following are some of the results obtained: 
Calcium per 


100 ce. g. 
1. Standard calcium solution (containing 0-0076 g. per 100 ec.) 00077 
2. ” ” ” ” : ” 0-0077 
3. ” ” ” 9? ” 0-0077 
4. ” ” ” ” ” 0-0077 
1. Human blood _... des tee avs ee as oes 0-0082 
2. ” ” + eee tee oer see eee tee 0-0080 
1. Ox blood ... see ans we ie aye as v4 0-0070 
2. ” ” ose one oe + eee tee see eee 0-0070 
lL. 5 9 (elotted) ry eve abe TY eae avi 00055 
2. ” ” ” tee eee eee vee tee tee 0:0055 
3. ” ” ” +e wee wee see eee tee 00056 
In all cases 5 cc, were used for the determination with the exception of 


the human blood when 3 cc, and 2 cc. respectively were used. 
Duplicate experiments show an agreement equal to two drops of NV/100 
permanganate or in other words to 0-006 mgm. of calcium, 


REFERENCES, 


Halverson and Bergeim (1917). J. Biol. Chem. 32, 159. 
Jansen (1918). Zeitech. physiol. Chem. 101, 176. 
MeCrudden (1909), J. Biol. Chem, 7, 83. 

Marriott and Howland (1917). J. Biol. Chem. 32, 223. 
De Waard (1919). Biochem, Zeitsch. 97, 176. 


XXXIII. THE MINIMUM NITROGEN EXPENDI- 
TURE OF MAN AND THE BIOLOGICAL VALUE OF 
VARIOUS PROTEINS FOR HUMAN NUTRITION. 


By CHARLES JAMES MARTIN anp ROBERT ROBISON. 
From the Laster Institute, London. 


(Received March 22nd, 1922.) 
HISTORICAL. 


Unit comparatively recently, the search for the minimum protein require- 
ments of the human body has been made on the assumption that protein is 
an entity, little regard being paid to whether the proteins were derived from 
meat, milk, cereals, etc. 

The more important of the numerous investigations undertaken with this 
object are those of Hirschfeld [1887], Kumagawa [1889], Klemperer [1889], 
Peschel [1891], Lapicque and Marrette [1894], Sivén [1900] and Albu [1901]?. 

In all of them the protein fed was derived from more than one source and 
often from several. The observations, which vary in precision, were made 
under conditions which were not uniform, particularly in regard to the total 
calories taken. Nevertheless, each experimenter succeeded in establishing that 
nitrogenous equilibrium could be maintained over short periods with one-third 
to two-thirds the standard laid down by Voit of 118 g. protein, equal to 
0-39 g. N per kilo. 

The minimum arrived at by the above experimenters varied between 
0-08 g. and 0-18 g. N per kilo, most of the results being round about 0-1 g. 
The lowest value is that of Sivén, who considers that he ultimately attained 
nitrogenous equilibrium on a mixed diet containing 0-08 g. N per kilo, only 
0-03 g. of which he regards as true protein, but as the only evidence that this 
small amount was sufficient is a positive balance of -04 g. N on the last day 
of a four days experiment, decided negative balances occurring on the first 
three days, this conclusion would appear questionable. In another series in 
which the nitrogen intake was 37 % higher the evidence of nitrogenous 
equilibrium is satisfactory. 

1 The earlier observations have been collected by Atwater and Langsworthy in their “ Digest 
of Metabolism Experiments,” Bulletin 45, U.S. Dept. of Agriculture, 1898. An excellent review 
of the work previous to his paper is given by Sivén. Most of the literature on the subject to 
date is referred to in Mendel’s “Theorien des Eiweissstoffwechsels,”’ Ergebnisse der Physiologie, 
11 Jahrgang, 1911; Caspari’s article “ EKiweissstoffwechsel” in Oppenheimer’s Handbuch der 


Biochemie, 1911, and Catheart’s Physiology of Protein Metabolism (1921). All of these contain 
good bibliographies. 


27—2 


408 C. J. MARTIN AND R. ROBISON 


The investigations of Neumann [1902] and Chittenden [1904] were under- 
taken with a somewhat different object, namely, to ascertain whether health 
and activity could be maintained over prolonged periods on a mixed diet of 
low content in protein. Neumann made three experiments on himself, each 
lasting four to ten months. The total calories of the diet amounted to 30-40 
per kilo. Chittenden’s observations were made upon 26 individuals and the 
duration varied between six and nine months in different cases. The total 
calories varied between 35 and 45 per kilo body weight. Nitrogenous equili- 
brium was obtained by Neumann on an intake of 0-15 g. N per kilo and in 
Chittenden’s experiments with 0-1 g. to 0-17 g. per kilo in the different indi- 
viduals. There is no reason to suppose that these figures represent minima. 
The conclusion drawn is that a nitrogen intake of one- to two-thirds of 
the amount laid down by Voit is sufficient to maintain health and efficiency 
over the periods during which the observations endured. 

During the last decade of the 19th century, physiologists were becoming 
increasingly alive to the possible significance of differences discovered in the 
elementary composition and chemical properties of the proteins, and the 
uniform value hitherto attributed to them in nutrition was coming under 
suspicion. Rubner [1897] appears to have been the first to formulate the 
view that proteins had different biological values. He used this conception 
to interpret some early experiments of his on the utilisation of various food- 
stuffs [1879] in which he had observed that less nitrogen was excreted in the 
urine on potato diets than on bread diets although the adverse N-balance was 
smaller on the former. In the same article Rubner expressed the opinion that 
the old search for a protein minimum must be fruitless since there will be, not 
one, but many minima according to the nature of the foodstuff used. He does 
not appear to have attributed such variation to any difference in chemical 
constitution although doubtless this possibility was in his mind. At this date 
the chemical constitution of the proteins was obscure although a number of 
amino acids had been isolated from the decomposition products of proteins 
and differences in the amounts of these had been observed. A little later Kossel 
and Kutscher [1900] determined the histidine, arginine, lysine and ammonia 
derived from the hydrolysis of a number of proteins and Kossel [1901] as a 
result of his own and others’ work came to the conclusion that the habit 
of regarding protein as a physiological unit was unsound and that, since 
proteins possess different chemical compositions, they will also have different 
values for the organism. 

About the same time discoveries were being made in another direction. 
The researches of Cohnheim [1901, 1906], Kutscher and Seemann [1901], 
Abderhalden and his co-workers and others, proved that a much more ex- 
haustive break up of the protein molecule than had previously been supposed 
takes place in the small intestine prior to absorption, and that what the body 
really receives is a mixture of amino acids and simple polypeptides. Loewi 
[1902], and later, Abderhalden and Rona [1904] showed by means of feeding 


BIOLOGICAL VALUE OF PROTEINS 409 


experiments with previously digested proteins that the abiuret products of 
such digestion were capable of maintaining animals in nitrogen equilibrium. 
Abderhalden believed at first that such amino acids were at once utilised 
for building up of blood and tissue proteins and in conjunction with Samuely 
[1905] attempted to ascertain whether the composition of the body proteins 
varied with the character of food proteins. A horse was fed for three days 
on gliadin containing 36-5 % of glutamic acid, but no increase in the very 
low content of this amino acid in the serum proteins could be detected. From 
these results Abderhalden and Rona [1906] drew the conclusion that in the 
renewal of body proteins a proportion of the amino acids arising from the 
food will be left over unless the body is capable of synthesising one amino 
acid from another. Since this proportion will depend on the relative com- 
position of the food and body proteins the protein minimum must also be 
variable. 

These discoveries provided a theoretical basis for Rubner’s empirical con- 
clusions and a stimulus for the further investigation of the subject. 

Meanwhile the réle play »d by protein in satisfying the energy requirements 
of the body, and the effect on the protein minimum of insufficient as against 
abundant provision for these needs from non-protein sources, was becoming 
more clearly realised. 

The discovery of the variable composition of proteins and of the fact that 
certain of them are almost entirely deficient in one or more of the amino 
acids was followed by interesting researches to determine to what extent the 
animal body could, by the practice of economy or synthesis, dispense with 
the missing complexes. The deficiency of gelatin in tyrosine was ascertained 
early and the absence from it of cystine and tryptophan was discovered when 
these amino acids became known as protein constituents. The inability of 
gelatin to preserve the body in nitrogenous equilibrium has been shown by 
many investigations, but as this aspect of the subject has recently been dealt 
with by one of us in this journal [ Robison, 1922, 1] it is unnecessary to review 
it here. 

The discovery of tryptophan by Hopkins and Cole [1901] and of the de- 
ficiency of this amino acid in zein by Osborne and Harris [1903] was followed 
by an experimental enquiry by Willeock and Hopkins [1907] to ascertain 
‘whether zein would serve as an exclusive source of nitrogen for mice and if 
not whether the addition of tryptophan would enhance the nutritive value 
of this protein. Zein alone failed to maintain the animals but this was achieved 
by supplementing with tryptophan. 

Following this pioneer work Osborne and Mendel [1911] planned a lengthy 
investigation of the biological value of different proteins in the light of the 
new knowledge of their chemical structure. Their earlier work was carried 
out before the importance of accessory factors was recognised but they became 
aware from their experiments with individual proteins that some factor other 
than the supply of protein, salts and energy was complicating their observa- 


410 C. J. MARTIN AND R. ROBISON 


tions. In the continuation of their researches, the results of which have 
appeared in some 30 papers in the Journal of Biological Chemistry from 1912 
up to the present, the error due to absence of vitamins was obviated. Osborne 
and Mendel [1912-1920] confirmed and extended the observations of Willcock 
and Hopkins and showed that the addition of 3 % tryptophan to the zein 
given was sufficient to maintain rats over a period of 182 days, but that they 
failed to grow. When 2 % of lysine, in which zein is also deficient, was added, 
growth occurred. The problem of maintainance is therefore distinct from that 
of growth. The observations of Osborne and Mendel are some of the most 
important contributions to our knowledge of nutrition. The choice of rats 
enabled great numbers of experiments to be carried out and the extension of 
the periods of observation to cover a large, fraction of the normal life of the 
animal. They prove that rats cannot supply some of the missing amino acids 
and that the minimum requirements and relative nutritive value of any 
particular protein depend upon the proportion of essential amino acids it 
contains. Their work also shows how a knowledge of the composition of 
particular proteins may be used for the eqpnomical adjustment of the nitro- 
genous portion of an animal’s dietary by arranging that one protein shall 
compensate for the deficiencies of another. 

The supplementary value of proteins from different sources has also been 
investigated by McCollum, Simmonds and Pitz, and McCollum, Simmonds and 
Parsons [1917 to 1921], whose observations, like those of Osborne and Mendel, 
were carried out upon rats. They found that cereal proteins could be satis- 
factorily supplemented by the proteins of milk, meat, kidney and casein 
and gelatin. Proteins of various leguminous seeds also usefully supplemented 
cereal proteins, e.g. wheat together with navy beans or peas. 

The ultimate test of the nutritive adequacy of a protein is its capacity 
' to nourish a young animal and provide for its complete growth and develop- 
ment and this, as far as the rat is concerned, is the criterion of the American 
investigators to which we have briefly referred. It may be surmised that, 
broadly speaking, conclusions arrived at from experiments on rats will be 
applicable in general to human nutrition. On the other hand the human 
mechanism may differ in detail. It will, for obvious reasons, be long before 
information as to the complete adequacy of individual proteins and quanti- 
tative data as to their biological values is forthcoming for human nutrition. 
In the meantime the findings of Osborne and Mendel, and McCollum and his 
co-workers have been applied with advantage to the feeding of stock. 

From this more general survey of the subject we will now return to con- 
sider observations upon the minimum requirements for equilibrium when 
nitrogen is supplied in different forms. We have already referred to the 
observations of Rubner which led him to the conclusion that a different 
nitrogen minimum would be discovered for different proteins, This surmise 
was subsequently investigated in his laboratory by Karl Thomas [1909] who 
introduced the term “ biological value.” The expression “ physiological value” 


BIOLOGICAL VALUE OF PROTEINS 411 


had been previously suggested by Voit and Korkunoff [1895] for a similar 
conception. Karl Thomas defined biological value as the number of parts 
of body nitrogen replaceable by 100 parts of the nitrogen of the foodstuff. 
Thomas’s definition is not concerned with the relative digestibility of the 
protein. The replacement of the “ Wear and Tear” quota was recognised as 
the only proper basis for comparison and in order to determine this value he 
fed himself on a carbohydrate diet (starch, sucrose, lactose) of high calorie 
value for periods of several days, during which the daily output of nitrogen 
in faeces and urine was determined. The figure to which this output fell was 
taken as his minimum requirements for the time being. During succeeding 
periods varying from one to four days a similar carbohydrate diet supplemented 
by a certain amount of the foodstuff under examination was taken and the 
N-intake and output determined as before. The N-intake was not as a rule 
kept constant and sometimes varied considerably on the different days. In 
most cases a negative N-balance was obtained. From the results Thomas 
calculated his biological value by three formulae based on the above definition, 
but differing from one another agcording to the way in which the nitrogen 
of the faeces is dealt with. 

Some of Thomas’s experiments lasted four to five weeks though no indi- 
vidual foodstuff was taken for longer than four days at a time. A period of 
one or two days on nitrogen-free diet was usually interposed between the 
experiments. Sixteen foodstuffs were investigated and their biological values 
recorded. These varied from 100 % in the case of milk to 30 % in the case 
of maize. We shall have occasion to discuss some of his results after dealing 
with our own experiments. 

Shortly before this work of Thomas appeared the results of experiments 
upon dogs with a similar object were published by Michaud [1909]. The 
output of nitrogen on diets of dog-flesh, sugar and fat was compared with 
that on diets of horse-flesh, caseinogen, gliadin, and edestin and on the 
carbohydrate and fat alone. Nitrogenous equilibrium was attained’ with an 
amount of nitrogen in the form of dog-flesh equivalent to the nitrogen output 
on the nitrogen-free diet. Negative balances were obtained with the other 
proteins, the greatest being in the case of gliadin and edestin. 

Zisterer [1910] found differences between caseinogen, flesh and gluten. 
These were however, in his opinion, too small to have practical significance. 

Observations upon pigs were made by McCollum [1911]. These animals 
lend themselves to metabolism experiments of this kind as they will consume 
sufficient of a diet free from nitrogen to obtain the necessary calories over a 
considerable period. Their minimum nitrogen expenditure can therefore be 
determined with reasonable accuracy. 

_ After a period of a week upon a diet of starch alone, the animals were fed 
for several days with the same ration to which was added a small amount 
of gelatin, zein, caseinogen or other protein. This was followed by the starch 
ration for a further period of some days. An amount of nitrogen in the 


412 C. J. MARTIN AND R. ROBISON 


form of gelatin equal to that of the urine upon the starch diet was found to 
cover 39 % and in the form of zein 73 % of the animal’s expenditure. The 
same amount of nitrogen in the form of cereal protein did not cause any rise 
in the nitrogen of the urine and with caseinogen the rise was small. McCollum’s 
experiments seem to avoid all the obvious pitfalls and his results indicate a 
much higher biological value for cereal proteins when fed to the pig than those 
arrived at by Thomas’s experiments upon himself. 

Hindhede [1913, 2] concludes that nitrogen equilibrium may be attained 
on a diet of potatoes and margarine containing only 20 g. of digestible protein. 
The figure is, however, arrived at by deducting the nitrogen of the faeces from 
the intake. This method of calculation is not in accordance with our know- 
ledge of the origin of a considerable portion of the faecal nitrogen and will 
furnish a too favourable balance sheet. . 

Hindhede [1914] vigorously contests the findings of Rubner and Thomas 
and claims to have attained nitrogenous equilibrium on as small an amount 
of protein in the form of bread as of potatoes. He declares as a result of his 
lengthy experiments that the proteins of potatoes, bread and meat can replace 
those of the body gram for gram. With Hindhede’s criticisms of some of 
Thomas’ experiments and treatment of his data, we are, for the most part, 
in agreement but must at the same time admit the justice of a great part of 
Rubner’s equally severe criticisms of Hindhede’s evidence, in particular, as 
regards the justification for assuming that all the nitrogen of the faeces 
represents undigested food proteins. 

Abderhalden, Fodor and Rose [1915] carried out some experiments to 
determine the minimum requirement of nitrogen in the form of different kinds 
of bread and potatoes. The subject of the experiments was Hofrat Rése who 
possessed some peculiarly advantageous characteristics. Rése was accustomed 
to a monotonous diet, neither smoked nor drank alcohol and was in the habit 
of chewing his bolus 120 times before swallowing it. Experiments of three to 
eight days’ duration were made on diets of potatoes, white wheaten bread, 
Swedish bread and kommiss brot, the last two being made from rye and con- 
taining bran. The experimental facts seem to us to warrant the conclusion 
that a gross intake of 4-5 g. of potato nitrogen, equal to 0-074 g. N per kilo, 
were adequate in the case of this Hofrat who chewed so long and so well. 
9g. N in the form of white wheaten bread was not quite sufficient and 
10-8 g. N as supplied in the rye bread was only just enough to reach equili- 
brium. This is not, however, the interpretation placed upon the results by 
Abderhalden and his co-workers, who conclude that bread nitrogen is as good 
as potato nitrogen and that for both of them the minimum nitrogen re- 
quirement is round about 4 g. for a man of 60 kilos. 

Rubner [1919] in the course of some studies of the capacity of certain 
vegetable nutriments to satisfy nitrogen needs, undertaken during the war, 
investigated different sorts of bread and the effect of milling to varying extent 
on the value of the product as a source of nitrogen, The paper covers a good 


BIOLOGICAL VALUE OF PROTEINS 413 


deal of ground and contains some particularly useful experiments with white 
wheat bread which can be compared with our own upon whole wheat. The 
bread was made, in one series, from white flour, 30 % milled, in the other 
of the same flour mixed with rye-bran to the extent of 30%, so-called 
“Finkler brot.” 10g. of the N as contained in the fine flour and between 
10 and 11g. of that in the Finkler bread were adequate to maintain 
equilibrium. 

Recently Sherman and his co-workers [1918, 1 and 2, 1919, 1920] have 
obtained results which are difficult to harmonise with those of Thomas. In 
experiments upon men and women, nitrogenous equilibrium was attained with 
an intake of 0-08 g. N per kilo, nine-tenths of this being supplied by cereal 
proteins and one-tenth by those of milk or apple. Wheat, maize and oats were 
found of equal value as a source of nitrogen and the view is taken that these 
cereal proteins possess a higher biological value than Thomas found. The 
effect of the supplementary action of the small quantity of milk may, in the 
light of the observations of Osborne and Mendel [1917] and of McCollum, 
Simmonds and Parsons [1921] be considerable. 

Boruttau [1915] believes that the low value of cereals as a source of 
nitrogen is greatly improved when these are consumed without the removal 
of the bran, etc. The biological value of 145 °% he obtained for the nitrogen of 
bran, is, in our opinion, an instance of the misuse of arithmetical formulae. 

R. O. Neumann [1919] made an excellent experiment upon himself in 1917 
in which he lived exclusively on rye bread, cane sugar and water for 40 days. 
Nitrogenous equilibrium was attained with 1000 g. bread and 300g. cane 
sugar (= 9-9g. N). The total calories of this diet amounted to 3630 or 63-8 
per kilo. On raising the calorie value of the intake to 4434 (or 73 per kilo) 
the nitrogen excreted steadily fell to 7-3. This indicates that Neumann in a 
long continued experiment could maintain nitrogenous equilibrium on less 
than 7 g. of nitrogen in the form of rye-proteins if excess of calories were 
furnished by sugar. The experiment is also interesting as indicating the 
sensitiveness of the nitrogen balance to the addition of carbohydrate. This 
aspect of the experiment will be discussed later. 

From a survey of the literature it is clear that certain of the proteins 
possess very different biological values both for growth and maintainance, 
There is, however, much uncertainty as to the degree to which the admixed 
proteins occurring in individual foodstuffs, where one protein to some extent 
complements the deficiencies of another, vary in value as a source of nitrogen. 

The divergence of opinion is most marked when it is based upon metabolism 
experiments on man over limited periods. 


414 C. J. MARTIN AND R. ROBISON 


OUR OWN OBSERVATIONS. 


INTRODUCTORY. 


We commenced our investigation lightheartedly with the comparatively 
modest object of re-determining the relative values of certain cereal proteins 
in human nutrition, in particular that of maize, in view of the significance 
given by Goldberger and others [1915, 1920] and.Wilson [1921] to the low 
biological value of maize in the causation of pellagra. The difficulties in 
arriving at values which could justifiably be compared were soon, however, 
apparent and it became essential to investigate thoroughly the conditions 
under which valid results might be obtained. In so far as the problem can 
be solved by metabolism experiments on adult animals the one unexceptional 
way to determine the relative biological values of proteins would be to 
ascertain the minimum intake on which nitrogen equilibrium can be maintained 
in each case. This sounds simple but unfortunately a positive balance only 
tells one that the intake is sufficient but not how much it is in excess and a 
number of experiments have to be performed to ascertain the minimum 
quantity. 

We were ourselves the subjects of the experiments. This is inconvenient 
but advantageous, for the experiments are exacting and necessitate constant 
supervision of one’s actions if sources of error are to be avoided. The partial 
abandonment of the joys of life is to some extent compensated by interest in 
the results. 

Nevertheless, the unnecessary multiplication of irksome experiments on 
one’s self, each extending over many days, is a thing to be avoided and it 
would be very desirable if a couple of observations could be made and the 
minimum requirements calculated from these with sufficient accuracy. This 
is what Thomas attempted to do. But in adopting such a method an assump- 
tion is made, the truth of which is by no means self-evident, namely, that 
the value of any protein for biological purposes remains uniform whatever 
the amount taken. The assumption would be justified if the nitrogen were 
utilised in the first instance to form some complex, such as ‘‘ Vorratseiweiss.” 

In this case the biological value, as pointed out by Abderhalden, would 
be determined by the ratio of the percentages, in food protein and body 
complex respectively, of that amino acid for which this ratio has the lowest 
value, unless the body has the capacity to synthesise that particular amino 
acid from others. 

It might also be true,if the nitrogen requirements are of varying nature 
so long as they are also indivisible, that is that no single requirement can be 
satisfied unless at the same time all the others are satisfied. 

The former of these two conceptions would appear to have been accepted 
without question by Rubner and Thomas though the case of gelatin obviously 
could not be treated in this way. Gelatin was considered to be capable of 
sparing body protein to the extent of 30-40 % when fed in relatively small 


BIOLOGICAL VALUE OF PROTEINS 415 


amounts but unable to do more than this however much was taken. Its 
biological value, if calculated by any of Thomas’s formulae would therefore 
appear quite appreciable when the intake was small but almost zero if the 
intake was very large. Yet, Boruttau [1919] has actually made use of these 
formulae to calculate the biological value of gelatin and has obtained a result 
of 58-2 %. 

Another possible disturbing factor (which we have reason to suppose 
occurs) is the varying economy with which the body deals with the amino 
acids supplied to it, according to their abundance. 

The various possibilities stated above may be made clearer by a diagram 
in which abscissae represent real nitrogen intake and ordinates the real 
nitrogen output. 


O pw 4 X 
Fig. 1. 

Let OM (=m) be the output on a N-free diet of adequate fuel value. 
Then m is equal to the nitrogen minimum. 

Suppose that an ideal protein (B.V. = 100) is fed in gradually increasing 
amounts and is utilised without waste. So long as the intake remains lower 
than m the output will remain constant and equal to m since the food protein 
saves an equal amount of body protein. The graph of intake and output will 
therefore be a line parallel to the #-axis and at E where ME = MO the body 
will be in nitrogen equilibrium. 

_ If now the intake be further increased, equilibrium will again result (unless 
the body is in a growing condition or has been previously starved of N) and 
the graph will now follow the line HE, at an angle of 45° to the axis. 

In the case of a protein of value less than that of the ideal protein just 
considered, equilibrium will not be attained on an intake equal to m but on 


416 C. J. MARTIN AND R. ROBISON 


some greater amount e, (at the point £,). On all amounts less than this, the 
output will exceed the intake and the graph will follow some line joining M/E,. 
Whether this line is straight or curved will depend on the conditions set out 
above, viz.: 

(1) indivisibility of the nitrogen requirements of the body; 

(2) uniform economy with varying nitrogen intake. 

If these conditions are obtained the line ME, will be straight and its 
equation will be y= m+ @ tan @ where y is a real output corresponding 
with any real intake z less than e,. 

For higher values of x the graph will follow the line E£,£,,. 

Thomas’s formulae can be very simply expressed in terms of 6; thus 
formula B 


B.V. = 100 Urine N in N-free diet + faeces N + balance 
See N-intake 


becomes B.V. = 100 miie-y) 


— 100 @+=- —(m+<2 tan @) 
x 


= 100 (1 — tan @). 

If the above conditions do not obtain, e.g. if a number of different 
amino acids are required for specific purposes, which are distinct and can be 
separately satisfied, the graph of a protein rich in certain of these acids but 
poor in others would be a curved line such as the dotted line joining M and EF, 
in the diagram. This curvature would express the fact that a certain fraction 
of the body’s needs could be satisfied by a smaller amount of this protein 
than would correspond with the amount required to obtain equilibrium. The 
angle 6 and the biological value would then vary for different values of x. 

The graph of a protein, unable by itself to satisfy any portion of the body’s 
nitrogen requirements, would be a straight line MK parallel to OEF,, since 
the nitrogen output would always be equal to the intake + m. For this line 
6= 45° and the equation y= m-+atan@ becomes y= m-+ a while the 
biological value = 100 (1 — tan 45°) = 0 

In our opinion there was very little reason for assuming that these graphs 
would necessarily prove to be straight lines. It is true that Thomas calculates 
his values from individual daily balances and takes the average of the results, 
but such daily balances are too variable and are subject to too great experi-- 
mental errors to offer any satisfactory proof of the uniformity of the value. 
We therefore set out to obtain evidence on this question by determining as 
accurately as possible a number of points for the same protein but for different 
values of «. Our results will be considered later but we may here state that 
in the case of bread proteins the points do lie on, or close to, a straight line. 
In the experiments with nitrogen in the form of milk results were at first 
obtained indicating pronounced curvature of the line, and nitrogen equili- 
brium was not obtained until more than 11g. of milk nitrogen was taken 
per day. By increasing the amount of carbohydrate however, so that the 


BIOLOGICAL VALUE OF PROTEINS 417 


fuel value of the diet was greatly in excess of requirements and the respiratory 
quotient greater than 1, equilibrium was finally reached with half this amount, 
and bearing in mind that when z, the intake, is very small, the physiological 
errors of experiment become relatively great, the observations could now 
perhaps be expressed by a straight line. As long as any doubt exists of the 
rectilinear character of the line ME it will obviously be prudent to place 
reliance only upon observations in which x and y are as large as possible 
short of equality. 


FACTORS WHICH MUST BE CONSIDERED IF VALID RESULTS 
ARE TO BE OBTAINED. 


1. The time required to reach a uniform N-output on a constant intake. 


The effect of the previous diet upon the N-output and the length of time 

required to reach a constant output on a constant intake which is either 
greater or less than that of the preceding period, was clearly demonstrated 
by the old experiments of C. Voit [1866, 1867]. The N-output of a dog during 
the first days of starvation varied with the amount of protein in the previous 
diet but fell gradually until a relatively constant figure was reached on the 
fifth or sixth day. When- the dog was given a constant meat diet for some 
days and then a considerably greater (or less) amount daily during a further 
period a similar gradual increase (or decrease) in the N-output was observed 
during five or six days before equilibrium again set in at the new level. These 
observations have been repeatedly confirmed by Grubner [1901], Landergren 
[1903], Kinberg [1911] and many others. 
The rapidity with which the nitrogen excretion diminishes obviously de- 
_ pends on the difference between the N-intake during the experimental and 
the preceding periods and will be greatest when a period of nitrogen starvation 
follows one of high protein intake or vice versa. There is no reason to suppose, 
however, that this gradual change is ever replaced by an immediate jump 
to the new level even when the difference between the two planes of N-intake 
is but small, though naturally the absolute amounts of the variations will be 
correspondingly less. 

Whether the N-output is also influenced by the nature as well as the 
amount of the protein taken in the foregoing period is more difficult to decide. 
If part of the nitrogen of the previous diet is stored up in any form that can 
be utilised by the body (e.g. amino acids) and not merely in the form of 
unexcreted end products, we should expect the N-output during the first few 
days of the succeeding period to be influenced thereby—unless during both 
periods the body is in N-equilibrium. For example if the diet during the first 
period contains 10 g. of N from caseinogen, and during the second period a 
negative balance occurs on 10 g. of N from zein, the amount of this negative 
balance might very well be less during the first few days of the zein diet than 
on the latter days of the same period owing to the supplementary action of 
amino acids stored up during the caseinogen diet. The results of experiments 
[Robison, 1922] in which a diet containing gelatin as the sole protein followed 


418 C. J. MARTIN AND R. ROBISON 


a diet of mixed proteins suggest that this does occur, and that therefore when 
the diet is changed in any way—either in amount or nature of the protein— 
the N-output cannot be considered to represent that of the second ~— until 
some days have elapsed. 

It follows that metabolism experiments are subject to error from these 
causes and that the error diminishes as the duration of the experiment in- 
creases. As a compromise, we have exlcuded from calculation the figures for 
the first three or four days in arriving at the N-output. It is also advisable 
that the N-intake should not vary during the experiment. Most of Thomas’s 
experiments are subject to both these sources of error. 


2. The time required for the elimination of errors due to the 
fluctuation in the N-output. 


The very considerable fluctuations that may occur in the amount of 
nitrogen excreted in the urine by men receiving an absolutely constant diet 
have been noted by Bornstein [1898], Atwater and Benediet [1902] and by 
Falta [1906], by all of whom the cause was considered to be psychical. 
Atwater and Benedict also noted the increased diuresis which often accom- 
panied a high N-output and thought it probable that the diuresis was the 
direct result of the psychical stimulus and the cause of the increased N-elimina- 
tion. Falta observed variations of 4—5 g. in the N-output on individual days 
although in equilibrium over the period as a whole. 

Neumann [1899] studied the influence of variations of the urinary flow 
upon the daily output of nitrogen, the intake remaining constant. When 
diuresis was produced by increasing the water drunk from one to three litres, 
the N-output increased from 10-5 to 14-3 g. and did not reach the original 
level until the third day. 

In almost all our experiments such fluctuations, amounting frequently to 
25 % of the mean output, have occurred and not least in those experi- 
ments in which the most rigorous attention has been paid to constancy of 
diet and fluid intake, and to regularity in the mode of life. 

At first we also were disposed to attribute these fluctuations to increased 
diuresis but the frequent lack of any correlation between the two compelled 
us to modify our opinion. Some factors (e.g. mental strain or excitement) 
may possibly affect both diuresis and N-output, but the latter does not always 
coincide with the increased volume of urine and may even vary in the opposite 
sense. In one experiment, for example, the minimum N-output corresponded 
with the maximum volume of urine. 

Judging from the experiments of Voit and other workers on dogs and of 
McCollum on pigs, it would appear that in the case of these animals the 
fluctuations in N-output on a constant diet are usually less considerable than 
with man. It is clear that calculations based upon the nitrogen balance sheet 
for single days, a procedure frequently resorted to, are subject to very large 
errors and that these can only be eliminated by taking the average results 
over a number of days. 


BIOLOGICAL VALUE OF PROTEINS 419 


3. The necessity for abundant energy supply. 


It has been universally recognised that proteins will be used as fuel unless 
an adequate supply of fat and carbohydrate is provided, but there have been 
different opinions as to what constitutes adequacy in this respect. In their 
search for the protein minimum some of the earlier workers considered it 
necessary to supply a diet of fuel value very greatly in excess of the energy 
requirements of the organism. Thus in Klemperer’s [1889] experiments on 
two young men, a diet of 5020 calories, equal to about 75 calories per kilo was 
given. Sivén, however, was of the opinion that such excess was unnecessary 
and that the minimum could be reached without increasing the calorie value 
of the diet above the normal. The fuel value of Sivén’s diets was equal to 
about 40 calories per kilo body weight. 

Hindhede’s [1913, 2] attitude is somewhat difficult to understand. He con- 
siders that an abundant calorie intake is necessary if the protein minimum is 
to be attained, but that this minimum will vary with the calorie value of the 
diet. He does not believe that a quiet old man, for whom a diet of 1500-2000 
calories is sufficient, can have the same minimum as an active young man 
who requires a diet of 3000-5000 calories. From the results of his experiments, 
he calculates by simple proportion, the minimum for a standard diet of 3000 
calories. As on a diet of 3900 calories F. Madsen’s minimum was equal to 
25 g. of digestible protein so, according to Hindhede, for 3000 calories the 
minimum would be 19 g. of digestible protein. 

Rubner [1919] has criticised this procedure, and in our opinion justly, on 
the ground that it is unwarranted by the facts, and considers that the values 
so calculated to 3000 calories possess no scientific basis. 

Rubner’s own conclusions are that the N-minimum may sometimes be 
reached when no more than a third of the total energy requirements are 
satisfied, in other cases only when they are fully met, while in others a diet 
considerably in excess of these requirements will be necessary. These differ- 
ences he considers are due to the varying nutritional condition of the body 
cells. The minimum is, however, most easily reached on an abundant carbo- 
hydrate diet. 

In our own experiments on a diet nearly N-free we appeared to reach our 
N-minimum with an intake of 45-50 calories per kilo body weight of which 
about one-third was taken in the form of fat, whereas on a diet of milk (with 
additional carbohydrate), equilibrium was not readily obtained until the fuel 
value was increased to about 55 calories per kilo of which only 10 % was in ~ 
the form of fat. 

The effect of diets containing varying proportions of fat and carbohydrate 
on the protein minimum has been studied by Zeller [1914], who found that 
the mixed diet was just as efficacious in reducing the consumption of body 
protein as one of carbohydrate alone, so long as the eee of carbohydrate 
to fat did not fall below 1 : 4. 


420 C. J. MARTIN AND R. ROBISON 


Neumann’s [1919] experiment upon himself, with a diet composed of bread 
and sugar, affords a striking demonstration of the effect of excess of carbo- 
hydrate in lowering the protein minimum. It is obvious that Prof. Neumann 
readily stores fat. Otherwise, he could not consume 73 calories per kilo over 
a period of three weeks, unless doing hard work. In his case, presumably, 
a greater excess of carbohydrate would be required to maintain the blood 
sugar at a high level than in our own, owing to the greater greed of his 
connective tissue cells. 

It would seem, therefore, that the most certain way of determining the 
protein minimum would be to take a diet consisting mainly of carbohydrate 
and so much in excess of the energy requirements that the blood sugar is 
maintained high, the liver and muscles are kept well stocked with glycogen 
and the surplus is being stored as fat, as indicated by a respiratory quotient 
above unity. This was accomplished in the latter part of our experiment on 
milk. It is by no means easy for one of the meagre habit of the subject of 
the experiment (C.J.M.) as it increases the distaste for the sufficiently un- 
appetising ration of starch and lactose and if persisted in too enthusiastically 
it produces unpleasant symptoms. 


4. Reduction of the nitrogen in the basal diet to a minimum, 


The carbohydrate (starch, lactose, etc.) and fat, etc. which form the basal 
diet for these experiments are nominally but not absolutely nitrogen-free. 
The amount of nitrogen taken in this form can of course be estimated but its 
biological value is unknown and this complicates the results. It is therefore 
important to reduce the nitrogen in the basa] diet to a minimum by careful 
selection of the most suitable forms of such foods. 

Most observers have neglected to take account of the nitrogen in the starch, 
etc. fed. As large quantities of such basal ration are consumed it may not 
be negligible in the case of experiments in which small quantities of some 
protein are being given. 


5. Accessory food factors and inorgame salts. 


The duration of metabolism experiments is limited by the difficulty of 
providing an adequate supply of the accessory food factors. The ill effects 
of long continued low-protein diets observed in some of the older animal 
experiments was no doubt sometithes due to the deficiency of one or other 
of these factors. 

Fat soluble A can be introduced in the form of rendered butter fat or 
cod liver oil, and water soluble C in the form of lemon juice, the nitrogen 
content of which is very low, but we have found no method of introducing 
the water soluble B without an undue amount of possibly very valuable 
nitrogen. 


ph Soe, aoe 


BIOLOGICAL VALUE OF PROTEINS 421 


When the diet is deficient in inorganic salts or is such as to afford an 
acid ash, adjustment by suitable amounts of a salt mixture is essential. 
McCollum and Hoagland [1913] found that the endogenous metabolism of 
the pig reached its lowest level when the animal was given an abundant 
carbohydrate diet together with a salt mixture of an alkaline character. 
When an acid salt mixture was given the urinary output rose, the increase 
occurring in the amount. of ammonia. They concluded that this animal 
is not able to use the nitrogen of the urea fraction for the neutralisation 
of acid, oe, 


6. The apportioning of the nitrogen in the faeces. 


The difficulty of correctly apportioning the nitrogen in the faeces to un- 
absorbed food nitrogen and excretion from the alimentary tract respectively, 
is the limiting factor in most experiments in which the total intake of nitrogen 
is small, and the way in which different observers treat the faecal nitrogen 
has given rise to much controversy and recrimination. 

The daily nitrogen output in the faeces on a protein-free diet usually 
amounts to about 1 g. On other diets the amount may be considerably greater 
than this and the question arises—what is the significance of this excess? 
Does it represent unabsorbed food residues or increased loss of body nitrogen? 
This difficulty was recognised and discussed by Karl Thomas, whose three 
formulae for the calculation of the biological value of protein differ only in 
the assumptions that are made with regard to this point. In formula A the 
whole of the nitrogen of the faeces is assumed to represent unabsorbed food; in 
formula B it is assumed to arise entirely from the body, while in formula C 
1 g. N (which is taken as the average output on protein-free diet) is assumed 
to be body nitrogen, any excess over this amount being ascribed to unabsorbed 
food. 

Rubner [1915] has investigated this problem in connection with his re- 
searches on the digestibility of various foodstuffs. He has devised a method 
for estimating the amount of nitrogen in the faeces present in the form of 
undigested food residues (vegetable cell membranes) by making use of the 
insolubility of the latter in acid alcohol and in a concentrated solution of 
chloral hydrate in which he states bacteria, epithelial cells, etc. are dissolved. He 
considers that the rest of the nitrogen comes from the body and represents 
metabolic products, and he concludes that such body nitrogen forms a very 
considerable proportion of the increase in the total nitrogen of the faeces that 
commonly occurs when the diet consists largely of whole cereals, vegetables 
or fruit. 

Rubner found that on a vegetable diet nearly half the total nitrogen of 
the faeces was soluble in acid alcohol whilst on a diet consisting chiefly of 
animal foodstuffs the proportion of soluble nitrogen was still greater. We have 
obtained similar results with faeces resulting from a mixed diet, but do not 

Bioch. xv1 28 


422 C. J. MARTIN AND R. ROBISON 


know in what form the whole of this soluble nitrogen is present, and are there- 
fore not able to draw definite conclusions as to its origin. 

So long as this uncertainty remains, the nitrogen of the faeces will be the 
limiting factor for the accuracy of such experiments as those here described. 
We have therefore followed Thomas’s plan and have calculated our results 
in two ways; the first, assuming that the whole of the nitrogen of the faeces 
comes from the body and that this amount plus the urine N on N-free diet 
represents the body’s minimum requirements under the conditions of the 
experiment; the second, assuming that the amount by which the nitrogen of 
the faeces exceeds the average amount excreted on a N-free diet represents 
unabsorbed food and is therefore to be subtracted from the total intake in 
order to arrive at the true nett intake, 7.e. the absorbed nitrogen. The truth 
will probably lie somewhere between these two extremes. baal 


EXPERIMENTAL. 


Firstly as to general procedure. The N of ell foodstuffs and beverages used 
was determined. The food was weighed in the same condition as that in which 
the N was determined and the total intake of N recorded in the tables can 
be relied upon to plus or minus 0-01 g. The urinary excretion of each 24 hours 
was collected in the presence of toluene, weighed, and duplicate determinations 
of the N content made. Care was exercised to see that invalid sampling from 
the deposition of either urates or ammonium magnesium phosphate did not 
occur. The faeces, after collection each day, were mixed with some H,SQ, to 
prevent decomposition and possible loss of NH, and great care was taken at 
the end of each period to ensure the validity of the samples taken for analysis. 
N determinations were made upon about 10 g. in duplicate, only closely con- 
cordant results being accepted. | 

The evacuation of the intestines is not usually so regular and complete 
that any great value can be placed on the figures for the nitrogen excretion 
on individual days. The best that can be done is to determine the nitrogen 
in the faeces for the whole experimental period and from this to calculate the 
average daily output. The latter may vary within rather wide limits even 
when the diet contains little or no nitrogen, and appears to depend to some 
extent on the bulk and character of the faeces. In order that these might 
be kept as uniform as possible agar-agar was taken when the diet consisted — 
wholly or largely of completely digestible foodstuffs. 

Food was taken in approximately equal amounts three times a day at 
the customary hours, and we led our usual life. In the experiments upon 
milk we abandoned all attempts to make our basal ration of fat and carbo- 
hydrate resemble a repast and drank a suspension of uncooked corn-starch 
in a saturated solution of lactose. This was followed by an alkaline salt 
mixture and 2g. of agar-agar. In this way 600 calories were contained in 
about a tumbler full. Microscopical examination of the faeces showed that the 


BIOLOGICAL VALUE OF PROTEINS 423 


starch was completely digested. When the starch-lactose mixture exceeded 
250 g. at one meal some glycosuria occurred temporarily. On the N-minimum 
experiment the faeces were olive green as the bile pigment was unreduced. 
Little gas was formed. 

The following are the essential data regarding the subjects of the experi- 
ments, ourselves: 


C.J.M. Age 56. Weight 61 kilo. Height 183c. Very thin, stores fat with 
difficulty. Mode of life: laboratory work. Exercise: lawn tennis before . 
breakfast for three-quarters of an hour and about three miles walk 
during the day. 

R.R. Age 37. Weight 59 kilo. Height 173-5. Spare, does not store fat 
readily. Usual mode of life consists chiefly in laboratory work. Very 
little regular exercise beyond daily walk of two to four miles. 


Minimum nitrogen expenditure on carbohydrate-fat diet. 


Our minimum nitrogen expenditure was determined in two experiments, 
during which our diet was as nearly as possible nitrogen-free. In the vain 
attempt to make this diet appetising much labour was expended in en- 
deavouring to prepare the food in a varied and attractive manner. Biscuits 
made of corn starch with 20 % of fat proved quite palatable when taken in 
small quantities but nauseating in bulk. A biscuit, whose chief defect was 
its hardness, made from starch, dextrin and a little fat was finally adopted 
and formed, with butter and honey, the chief article of diet. A starch mould 
flavoured with lemon juice was also taken. Weak tea with lemon and some- 
times black coffee and a little vermouth was drunk. After the first few days 
of this diet no desire was felt either for this or for any other food, nor did the 
sight of our first normal meal at the close of the experiment arouse any 
appetite. The drinking of a glass of hot milk, however, excited in a few 
minutes a very keen appetite and desire for food. The quantities of the 
individual constituents of the diet varied somewhat from day to day but 
were always accurately measured and noted. Those for a single typical day 
are set out in Table I while the total nitrogen intake and fuel value of the 
diet for each day are shown in Table II in which are also set out the daily 
output of nitrogen in the urine and faeces and the nitrogen balance. In this 
and in all other experiments the nitrogen in the tea and coffee has been 
assumed to consist chiefly of caffeine and to be excreted unchanged in the 
urine. It has therefore been subtracted in all cases from the total intake 
and from the urinary nitrogen output. Any error involved in this method of 
treatment must be of negligible dimensions. 

The nitrogen in the urine fell steadily until the last day of the experiment 
when a rather considerable rise occurred in the case of both C.J.M. and R.R. 
This is probably to be explained by the fact that, owing to the difficulty of 
consuming the food when all appetite was in abeyance, the fuel value of the 
diet was reduced during the last day or two. 


28—2 


424 C. J. MARTIN AND R. ROBISON 
Table I. 
N-minimum 2. Diet R.R. 4. xii. 20 
N per Calories per — ie — 
Foodstuff 100 g. 100 g. Weight g Ng. Calories 
Corn starch 0-042 360 280 0-118 1008 
Dextrin 0-065 360 50 0-033 180 
Butter 0-080 775 90 0-072 698 
Margarine (rendered) 0-010 900 15 0-002 135° 
Honey 0-023 327 55 0-013 180 
Sucrose — 395 25 — 99 
Lactose 0-013 370 65 0-008 . 241 
Lemon juice 0-067 40 30 ce 0-020 12 
Vermouth 0-005 140 25-5 0-001 35 
Tea infusion 0-006 — 1200 ,, 0-072 — 
Agar-agar 0-242 - 15g 0-036 — 
Salt mixture — —_ Bos5 — _— 
; Total 0-375 2588 
Excluding tea N 0-303 45 per kilo 
Table II. : 
N-minimum 1. Subject: R.R. Fluid Intake: 2000-2500 ce. 
Daily Intake 
hs — Output 
N N - Xx *% 
Body (excluding tea and Calories N N N Balance 
Date weight teaand coffee per urine faeces total N 
1920 k coffee) g. g. kilo g. g. g. g. 
Nov. 15 58-5 0-30 0-16 56 8-64 — = — 
» 16 0-28 0-06 + 5-31 1-05 6°36 — 6-08 
eet 0-30 0-15 55 _ 3-63 ss 4-68 —4:38 
wi. 18 0-25 0-10 51 2-66 eF 3-71 — 3-46 
are | 0-22 0-06 43 2-25 = 3°30 — 3-08 
ee 57-7 0-23 0-15 49 2-82 - 3:87 —3-64 
(21. xi) 
N-minimum 2. Subject: R.R. Fluid Intake: 4000 ce. 
» 2 58-6 0-27 0-07 49 8-79 1-13 9-92 — 9-65 
» 30 (23. xi) 0-30 0-07 50 4-71 7 5-84 — 5-54 
Dec. 1 0-30 0-08 51 3-46 s 4-59 —4:29 
” 2 0-35 0-08 51 2-71 Ms 3°84 -— 3-49 
*” 3 0-32 0-09 52 1-99 A 3-12 — 2-80 
” 4 0-30 0-07 45 2°17 Ps 3°30 — 3-00 
Pa 5 57-7 0-27 0-06, dA 2-01 =" 3:14 — 2-87 
Average of last 3 days 0-30 0-07 -- 2-06 1-13 3:19 — 2-89 
N-minimum 1. Subject: C.J.M. Fluid Intake: 2000 ce. 
Nov. 14 61-7 0°33 0-20 56 8-61 —_ — _ 
» 415 0-30 0-20 47 4-89 1:24 6-13 ~ 583 
os 16 0°30 0-20 47 3-28 es 4-52 — 4:22 
nc ae 0-38 0-21 52 3-28 pi 4:52 ~4-14 
ie ae 0-32 0-21 46 2-48 ‘- 3°72 ~ 3-40 
» 19 0-30 0-21 45 2-25 ” 3-49 -319 
» 20 60-9 0-21 0-18 35 2-49 os 3°73 -— 3-52 
(21. xi) 
N-minimum 2. Subject: C.J.M. Fluid Intake: 4000 ce. 
» 2 61-9 0-34 0-31 59 8-89 ~~ _ _ 
” 30 0-34 0:27 55 4-78 — ite 7 
Dec, 1 0-34 0-33 55 3-40 117 4:57 - 4:23 
” 2 0°33 0-32 as 2-41 i 3-58 — 3°25 
” 3 0-35 0-32 52 251 ” 3-68 — 3°33 
” 4 O35 0-29 52 2-40 ‘ 357 — 3°22 
” 5 60-5 0-33 0°26 48 213 ” 3:30 -2-97 ° 
Average of last 3 days 0-34 0-29 o- 2-34 117 3-51 =3:17 


BIOLOGICAL VALUE OF PROTEINS 425 


On plotting the amounts of urine nitrogen as ordinates against the time 
in days as abscissae it became apparent that all the points except the last 
would lie on, or close to a curve for which a simple logarithmic expression 
was found (Figs. 2 and 3). Among several possible interpretations of this 
curve is the simple one that the falling nitrogen output represents washing 
out of some metabolic products from the tissues, the amount washed out each 
day being proportional to that still present. 

A second similar experiment was therefore carried out in order to confirm 
this result and also to discover whether the steepness of the curve could be 
altered by drinking large quantities of water and thus greatly increasing the 
volume of the urine. Apart from the volume of total fluid, which was doubled, 
the diet did not differ from that taken in the first experiment. The results 
are set out in Table IT. 

The regularity in the fall of the nitrogen output was again observed but 
the change in the rate of this fall was very small. 

The average outputs in urine and faeces on the last three days of this 
experiment have been taken as representing the minimum nitrogen expen- 
diture on such a diet. Whether this amount also represents the minimum 
expenditure on a diet that is absolutely nitrogen-free will depend on the 
biological value of the small amount of nitrogen in the food consumed during 

the above experiment. If this has a value of 100 %%, 7.e. if it can replace and 
therefore spare an equal amount of body nitrogen, the output thus deter- 
mined will be equal to the real minimum expenditure. If the value of the 
food nitrogen is zero, 7.e. if it is unable to satisfy any fraction of the body’s 
requirements, it must be excreted in addition to the amount representing the 
latter and the minimum expenditure will therefore be equal to the observed 
output less the full amount of the nitrogen intake. Probably this nitrogen, 
which was present chiefly in the corn starch and butter, has a value inter- 
mediate between 100 and zero.. The minimum nitrogen expenditure will 
therefore be some amount between those shown in the last two columns below. 


Minimum nitrogen expenditure. 


Urine Faeces Total output Total output —intake 
Subject g. g. g. g. 
C.J.M. 2-34 1-17 3°51 3-17 
R.R. 2-06 1-13 3°19 2-89 


Tue BIOLOGICAL VALUE OF THE PROTEINS OF WHOLE WHEAT. 


A series of experiments was carried out with whole wheat, flour with two 
objects in view: 

(1) to determine the Biological Value of the wheat proteins from the 
minimum amount with which nitrogen equilibrium can be attained; 

(2) to discover whether the Biological Value is uniform for varying amounts 
of wheat nitrogen. 

The first of these has been investigated upon man by other workers, whose 


426 C. J. MARTIN AND R. ROBISON 


results will be considered with our own. The great practical importance of 
this question and the astonishing divergence between the results of previous 
investigations were sufficient reasons for further study. 


Method of experiment. 


The large variations in the percentage of water, and consequently of 
nitrogen, in different parts of a loaf of bread and the difficulty of obtaining 
a satisfactory sample, render it impossible to estimate the total nitrogen 
content of the loaf with sufficient accuracy for these experiments. We decided 
therefore, to base our calculations on the flour and to bake the bread ourselves. 
By suitable manipulation it was found possible to prepare loaves from 500 g. 
of flour with a maximum loss of less than 0-1 %. The other materials used 
were butter (5 %), salt, baking powder (prepared from tartaric acid, sodium 
bicarbonate and corn starch) and water. The loaves were baked in the labo- — 
ratory for about one hour at a temperature of 240°-250°, and were only very 
slightly browned so that no appreciable loss of nitrogen can have occurred — 
during the baking. For the first ‘period, a somewhat coarsely ground whole 
wheat flour containing 1-85 % N was used but on increasing the daily ration 
from 300 g. to 450 g., considerable discomfort was experienced from the large 
particles of bran, and the bread was poorly absorbed. For the second and 
remaining periods a very finely ground flour prepared from a mixture of 
English and foreign whole wheat was employed. We found it very palatable 
and well absorbed, as the figures for the nitrogen in the faeces indicate. Only 
when the daily consumption had been raised to 550 g. and the total calories 
to 63 per kilo did we experience any discomfort. 

The experiment was carried out in duplicate on ourselves and commenced 
with a total nitrogen intake of nearly twice the amount of our minimum re- 
quirements. On this diet a considerable negative balance occurred and the 
amount of wheat nitrogen was therefore increased during successive periods 
until equilibrium was finally attained. 

The daily ration of flour was kept constant during each separate period 
of the experiment, but some latitude was permitted in the amounts of the 
remaining constituents of the diet. The actual quantities taken were, however, 
measured and the variations in any one period were not such as to affect 
appreciably the total nitrogen intake or greatly alter the fuel value of the 
diet. The diet set out in Table III shows the average amounts consumed by 
R.R. from the 24th to the 3lst of October but except for the quantity of flour 
it would with slight variations serve for the whole experiment, 

The experimental results are set out in Tables IV and VY. 

As in the previous experiments it has been assumed that the nitrogen 
consumed in tea and coffee would be excreted unchanged and the amount 
has therefore been subtracted from both intake and output (urine N), 

During certain periods indicated by the letter (A) in the first column of 


BIOLOGICAL VALUE OF PROTEINS 427 
Table III. 
R.R. Diet during period 5, 
Whole wheat. 24-31 Oct. 1920 
N per Calories per _ A — 
Foodstuff 100 g. 100 g. Weight g N g. Calories 
Flour, whole wheat 2-162 360 450 9-730 1620 
Corn starch 0-042 360 36 0-015 130 
Butter 0-080 775 56 0-045 434 
Dripping 0-016 885 80 0-013 708 
Honey 0-023 327 40 0-009 131 
Marmalade 0-052 341 20 0-010 68 
Cane sugar —_ 395 56 — 221 
Lemon juice 0-067 40 20 ce 0-013 8 
Tea 0-006 — 960 ,, 0-058 — 
Coffee 0-041 — 100 ,, 0-041 — 
Vermouth 0-005 140 25 ,, 0-001 35 
Total 9-935 3355 
Excluding tea and coffee N 9-836 58 per kilo 
Table IV. 
Diet: whole wheat. Subject: C.J.M. 
Daily Intake Daily Output 
A. Sw 
Bod N N Calories N N N Balance 
Date weight bread total oy urine faeces total N 
Period 1920 k g. g. ilo g. g. g. g. 
1. (A) Sept. 30-Oct.2 61-80 5-55 58 46 — — — — 
(26. ix) 
(O) Oct. 3-5 61-50 “a 5-70 5-62 2-14 7-76 — 2-06 
(3. x) 
2. (A) » 10-14 60-45 8-32 8-63 — — — — 
tere 8. (10. x) a pe es 8-26 2:17 10-43 — 1-80 
” 16 ” ” ” 7-62 ” 9-79 —1-16 
RS 61-44 > Pe ss 6-98 a: 915 —0-52 
Average 8-32 8-63 56 7-62 2-17 9-79 —1-16 
3. (A) a 48 8-65 8-97 57 7-86 2-35 10-21 — 1-24 
” 19 ” ” ” 7-25 ” 9-60 — 0-63 
» 20 ” ” ” 715 ” 9-50 — 0-53 
Average 8-65 8-97 57 7-42 2-35 9-77 -0-80 
4. (A) sa ad 60-65 9-73 9-98 53 7-25 1-74 8-99 +0-99 
3 Oe MA me ie 7-92 Pa 966 +0-°32 
Eas ae 61-03 i. e “2 7-30 nt 9:04 +0-94 
Average 9-73 9-98 53 7-49 1-74 9-23 +0-75 
5. (QO) » 24 60-95 9-73 9-85 60 8-65 1-94 10-59 —-—0-74 
Raw nA mr no 9-23 es 11-17 — 1-32 
Ras a os a 8-77 Es 10-71 — 0-86 
” 27 “99 ” ” 7-84 ” 9-78 +0-07 
ae > — a . 8-93 53 10-87 — 1-02 
a foe Pe me Ps 8-63 2 10-57 —0-73 
” 30 ” ” ” 7-68 ” 9-62 +0-23 
Pe 61-60 an ne a 7:67 ‘5 9-61 +0-24 
Average 9-73 9-85 60 8-43 1-94 10-36 —0-51 
6. (O) Nov. 1 11-89 12-00 61 9-13 2-53 11-66 +0-34 
” 2 ” ” ” 8-33 ” 10-86 +1-14 
Reincab | re 4 Re 9-83 oe 12-36 — 0-36 
/ ae | 61-40 a eA Ps 9-63 ob 12:16 -0-16 
(5. xi) 
Average 11:89 = 12-00 61 9-23 2-53 11-76 +02 


428 C. J. MARTIN AND R. ROBISON 


the table, a quantity of stewed apples (225 g. raw fruit containing 0-11 g. N) 
was included in the diet. The letter (O) indicates that this fruit was omitted. 

For several days prior to Sept. 30, 1920 (period 1) a bread diet containing 
about 7-3 g. N had been taken in order to eliminate the disturbing effect of 
the previous high protein dietary. For the same reason the first three days 
of this period and the first five days of period 2 have been excluded in calcu- 
lating the average nitrogen balance. 


Table V. 
Diet: whole wheat bread. Subject: R.R. 
Daily Intake Daily Output 
c si = ne + oe 
Body N N Calories N N N Balance 
Date weight bread total a urine faeces total N 
Period 1920 - . ge g. ilo g. g. oy oh 
1. (A) Sept.30-Oct.2 58-65 555 582 48 Sis ua fe on 
(27. ix) 
(O) Oct. 3-5 57-70 - 5-71 49 5-62 1-47 7-09 — 1-38 
(6. ix) 
2. (A) » 10-14 5820 832 857 654 a et ra eas 
(13. x) 
”° 15 ” °” ” 7-21 1-98 9-19 — 0-62 
” 16 ” ” ” 8-30 ” 10-28 -1-71 
ee | | 58-13 ap re PR 6-41 me 8-39 +0-18 
Average 8-32 8-57 54 7:31 1-98 10-29 - 0-72 
3. (A) ee 865 891 656 806 - 125 931 -040 
7 we ” ” ” 8-70 ” 9-95 — 1-04 
is) 20 ” ” ” 7:79 ” 9-04 —0:13 
Average 8-65 8-91 56 8-18 1-25 9-43 -0-52 
4. (A) » 21-23 57-89 9-73 10-00 58 8-32 1-70 10-02 — 0-02 
(2i.'x) 
5. (0) » 24 973 984 58 841 181 10-22. -0-38 
ag J ” ” ” 8-25 ” 10-06 — 0-22 
” 26 ” ” ” 10-51 ” 12-32 — 2-48 
o) meee 57-75 3 “ aS 8-83 ve 10-64 — 0-80 
” 28 ” ” ” 7-67 ” 9-48 +0:36 
” 29 ” ” ” 9-74 ” 11-55 —-1-71 
” 30 ” ” ” 9-50 ” 11-31 —1:47 
ee) 57-44 3 ms os 8-54 $y 10:35 —-0-51 


Average 9-73 9-84 58 8-93 1-81 10:74 —0-90 


6. (0) Nov. 1 1189 1198 63 930 268 1108 0 
” 2 ” ” ” 9°75 ” 12-43 — 0-45 
” 3 ” ” ” 8:86 ” 11-54 +044 
NS Wer eee = # 9-46 » Se aS 


(5. xi) 
Average 11:89 ~—:11-98 63 9-34 2-68 12:02 -0-04 
A marked rise in the nitrogen output occurred with both C.J.M. and R.R. 
at the beginning of period 5 and the coincidence of the omission of the apples 
that had been included in the diet during the preceding period led us to 
consider whether there was here any relation of cause and effect. Two possi- 
bilities suggest themselves: (1) the small amount of apple protein (0-11 g. N) 
might possess high value to supplement the wheat protein; (2) the alkaline 


BIOLOGICAL VALUE OF PROTEINS 429 


ash of the fruit would partially neutralise the acid ash of the bread and this 
might affect the nitrogen expenditure. Neither of the above appears adequate 
to explain the facts and further, the results obtained during period 4 do not 
agree any better with those for periods 2 and 3, in which apples were included 
in the diet, than with those for period 5 in which they were omitted. The 
increased. output must therefore, like the variations which occur from day to 
day, be left for the present unexplained. 

From the results of these experiments we have calculated the biological 
value of the wheat proteins by two formulae, which differ only in the assump- 
tion made with regard to the nitrogen of the faeces. Both are based on 
Thomas’s definition of this value as the number of parts of body nitrogen 
spared by 100 parts of the nitrogen of the food, and, when reduced to their 
simplest form, can be thus expressed 


Balance [P]— Balance [M] 
B.V. = 100 Food N absorbed — 100 Intake [P]-— Intake [M] 


where P signifies the experiment with the protein under investigation and 
M the nitrogen minimum experiment. 

This correction for the small amount of nitrogen in the diet of the N- 
minimum experiment is strictly accurate only if certain provisos hold, viz. 
(1) that the same amount of nitrogen in the same form, or of the same bio- 
logical value, enters also into the second diet (P); (2) that the value of this 
nitrogen is not increased by supplementary action with the other proteins of 
the second diet. 

In our experiments with wheat and milk proteins the first proviso is partly 
but not entirely satisfied. Whether or not the second is also satisfied cannot 
be decided. The method, however, certainly involves a less error than if the 
N-intake in the N-minimum experiment is altogether ignored. 

When the different assumptions as to the faeces are made the two formulae 


become: 
Balance [P] — {Intake [M] — (Urine N [M] + Faeces N [Pt 
I. B.V. = 100 Intake [P] — Intake [/] 


Balance [P] — Balance [M 
TE BV, cry dcees WLP = Poses Be) Totals La 

In I the whole of the food nitrogen is assumed to have been absorbed so 
that the total intake is also the real intake. In calculating the minimum 
expenditure corresponding with the period in question the faeces N for this 
period is added to the urine N [M] and the intake [M] subtracted from the 
sum. This formula corresponds with Thomas’s formula B. 

In II when the nitrogen of the faeces is in excess of that occurring in the 
N-minimum experiment this excess has been assumed to represent unabsorbed 
food and has been subtracted from the total intake to obtain the real intake. 
The minimum expenditure has been taken as the actual balance on N- 
minimum diet, ¢.e. Urine N [M] + Faeces N [M] — Intake [M]. This formula 
corresponds with Thomas’s formula C except that in the latter an average 


430 C. J. MARTIN AND R. ROBISON 


figure of 1 g. N, has been taken to represent the faeces N on a N-free diet. 
Both procedures have obvious disadvantages but there is very little difference 
in the results whichever is adopted. 

A summary of the results for the separate periods with the biological values 
calculated from both the above formulae is given in Table VI. 


Table VI. 
Diet: whole wheat bread. Subject: C.J.M. 
Intake 

7 A — Balance Biological value 

Total N Absorbed N Calories per N oN 
Period g. g. kilo g. (1) (2) 
1. (O) 5-70 4-73 44 — 2-06 38-8 25:3 
2. (A) 8-63 7-63 56 —1-16 36-3 27-6 
3. (A) 8-97 7-79 57 — 0-80 41-1 31-8 
4. (A) 9-98 9-41 53 +0°75 41-0 43-2 
5. (O) 9-85 9-08 60 -0°51 36-0 30-4 
6. (O) 12-00 10-64 61 +0-24 40-9 33-1 
Average 39-0 31-9 

Diet: whole wheat bread. Subject: R.R. 

1. (O) 5-71 5:37 49 — 1-38 34-2 29-8 
2. (A) 8-57 7-72 54 -0-72 36-8 29-3 
3. (A) 8-91 8-79 56 — 0-52 28-7 27-9 
4, (A) 10-00 9-43 58 — 0-02 35:5 31-4 
5. (O) 9-84 9-16 58 —0-90 28-0 22-5 
6. (O) 11-98 10-40 63 . —0-04 37-7 28-1 
Average 33-5 28-2 


Tue BIOLOGICAL VALUE OF THE PROTEINS OF Cow’s MILK. 


The first experiment with nitrogen in the form of milk protein followed 
immediately after the second period on low nitrogen diet and was carried out 
in duplicate on C.J.M. and R.R. The total nitrogen intake was approximately 
equal to the minimum nitrogen expenditure and though equilibrium was not 
obtained the negative balance was not very large, from which we concluded that 
milk proteins would be found to possess a relatively high value. When the 
experiments were repeated with larger amounts of milk we were surprised to find 
the nitrogen balance still negative and equilibrium was only reached with diets 
containing over 11 g. of milk nitrogen. This result appeared so extraordinary 
as to lead us to suspect that it might be due to a deficiency in the energy 
value of the diet, although this was amply sufficient to cover our normal 
requirements. The unusually rapid loss in weight which occurred in some of 
these experiments pointed in the same direction, as did also the coincidence 
of the fall in the nitrogen output occurring in one period (C.J.M. 6-R.R. 3) 
with a pronounced rise in the temperature of the air. 

A further series of experiments was therefore carried out on one of us 
(C.J.M. periods 7, 8, 9), the fuel value of the diet being increased to the 
maximum that could be consumed, In the first of these, nitrogen equilibrium 


BIOLOGICAL VALUE OF PROTEINS 431 


was practically attained with a diet containing 6-84g. N and furnishing 
57 calories per kilo. After an interval of one day, on which the basal ration 
together with a very little milk (3-0 g. N) was consumed, period 8 was begun. 
The diet contained 5-28 g. N and furnished 55 calories per kilo, but the effects 
of the continued excessive diet made themselves unpleasantly obvious in the 
form of a bilious attack, which threatened to terminate the experiment. By 
reducing the amount of carbohydrate consumed, so that the fuel value fell 
to 36 calories per kilo it was however just possible to carry on and after two 
days the condition was so far improved that the full diet was resumed. The 
nitrogen intake was not altered at all but the effect of the reduced diet was 
very marked in the increased nitrogen output, which persisted for several 
days after the calories were again increased. The period was extended for 
six days after the output had reached a fairly constant level and only the 
results for these days have been considered in calculating the average output. 
A decided though small negative balance occurred. 


Table VII. 
Milk. C.J.M. 
Diet during period 9 
20-25. viii. 21 
Nitrogen Calories — A oy 
Foodstuff per 100g. perl00g. Weightg. Nitrogen g. Calories 
Milk 0-508 65 830 4-216 539 
Starch 0-027 360 280 0-076 1008 
Lactose 0-013 370 210 0-027 777 
Honey 0-023 327 250 0-057 818 
Margarine (rendered) 0-010 900 18 0-002 162 
Sucrose “= 395 10 — 40 
Tea 0-008 — 750 ce. 0-060 — 
Agar-agar 0-242 — 6 0-014 — 
Salts — + a — 
Total 4-452 3344 
Excluding tea N 4-392 54 per kilo 


In period 9 the quantity of milk was again reduced but the high calorie 
value of the diet was maintained. The usual daily game of tennis was dis- 
continued and no exercise was taken so that the excess of energy supplied 
was even greater than before. The weather also was very hot. The average 
of the last six days of this period showed a considerable negative balance 
and the biological value calculated from this agrees fairly well with that 
obtained from the results of periods 7 and 8 and also of period 2 (C.J.M.). 
This point is interesting because the calorie value of the diet in period 2 was 
lower than in any other, only 44 per kilo. The composition of the diet during 
period 9 (C.J.M.) is given in Table VII and the results of the experiments are 
set out in Tables VIII (R.R.) and IX (C.J.M.), while Table X is a summary 
of these showing the biological values calculated from the two formulae. It 
is obvious from the amounts of nitrogen in the faeces that the milk proteins 
were very completely absorbed so that formula 1 probably gives the closest 
approximation to the truth in this case. 


432 C. J. MARTIN AND R. ROBISON 


Table VIII. 
Diet: milk. Subject: R.R. 
Daily Intake Daily Output 
Body N N Calories N N N Balance 
Date weight milk total Fee urine faeces total N 
Period 1920 k aed g. ilo g. g. g. g. 
“N-free” Dec. 5 57-75 0 0°27 44 2-01 1-1 3-14 — 2-87 
lL. ae 2-73 2-98 48 2-49 0-66 3°15 -0-17 
pode: 5 2-69 2-94 a 2-95 a3 3°61 — 0-67 
oe aes 2-76 3-09 es 3°22 a 3°88 —0-79 
9 2-59 2-91 a 2-86 . 3°52 —0-61 
ee i | 57-50 2-67 2-97 is 2-83 oo 3-49 — 0-52 
(11. xii) : 
Average of last 4 days 2-68 2-98 48 2-97 0-66 3°63 — 0-65 
1921 
2: May 28 59-65 4-10 4-30 40 7-12 — — — 
i 6-14 6:29 48 5:97 1:18 7-15 — 0-86 
30 ” ” ” 591 ” 7-09 ed 0-80 
Pry 31 °° ” ” 6°34 > 7-52. , 1-23 
June 1 ~ sf a 7-09 er 8-27 — 1-98 
” 2 ” ” ” 6-55 ” 7°73 — 1-44 
ian 57-90 os . is 6-21 = 7:39 —1:10 
(4. vi) 
Average of last 4 days 6-14 6-29 48 6-55 1-18 7:73 — 1-44 
June 16 Mixed diet 12-29 — — —_ 
3. PEE i 59-60 13-61 13-69 47 13-48 1:37 14-85 — 1:16 
Prey ” ” ” 13-13 ” 14-50 —0-81 
” 19 ” ” ” 12-29 ” 13-66 +0-03 
” 20 ” ” ” 13-76 ” 15-13 — 1-44 
” 21 ” ” ” 12-39 ” 13-76 —0:07 
” 22 ” ” 29. 12-60 ” 13:97 Me 0-28 
tt oe 58-50 = et Sy. 11-10 ES 12°47 + 1-22 
oo ee “i a z 11-19 me 12-56 +1:13 
Average of last 5 days 13-61 13-69 47 12-21 137 «613-58 = =6+ 0-11 
Table IX. 
Diet: milk. Subject: C.J.M. 
Daily Intake Daily Output 
PRE A tet ’ & a es | 
Bod N N Calories N N N Balance 
Date weight mi total = urine faeces total N 
Period 1920 k g. g. ilo g. g. g. g. 
“N-free” Dec. 5 60-45 0 0°33 48 2-13 1-1 3:3 — 2:97 
1. esp 3-15 3°37 49 2-42 1-29 3:71 — 0°34 
Basen | = 3°36 Pe 2-57 a 3°86 - 0-50 
aS * 3°42 sy 3-01 = 4°30 —0-88 
” 9 ” 3-40 ” 3°12 ” 4-41 a 1-01 
of AD 59-30 A 3-45 = 3°22 Es 4-51 — 1-06 
(11. xii) < 
Average of last 3 days 3-15 3-42 49 3-12 1-29 4-41 - 0-99 
1921 
2. April 30 62-60 4-02 4°17 44 4:74 1-22 5:97 — 1:80 
ay 1 ” ” Pa 4:58 Fe 580 -1-63 
” 2 ” ” ” 4-09 ” 4-32 = 1-15 
” 3 ” ” ” 4°25 ” 547 - 1:30 
ae 62-20 a 3 + 4°53 a 5:75 ~ 1-58 
Average of last 4 days 4-02 417 At 4:36 1-22 5-58 ~141 
May 10 Mixed diet 8-00 _ _ — 
3. pro: | | 62°20 6-20 6-35 48 6-13 1-18 7-31 — 0:96 
” 12 ” ” ” 571 ” 6°89 — 0:54 
ce ” ” ” 526 ” 6-44 ~ 0-09 
” 14 ” ” ” 5:80 ” 6-98 -~ 0:63 
” 15 ” ” ” 6-08 ” 7:26 = 0-91 
oo» 16 ” ” ” 6-07 ” 7:25 - 0:90 
io’. oe 62-15 ak as pe. 6-21 Hh 7-39 — 1:04 
(18. v) 
Average of last 4 days 6-20 6°35 48 6-04 118 7:22 8-087 


BIOLOGICAL VALUE OF PROTEINS 


Daily Output 
_. A esse | 
N N N 
urine’ faeces total 
&- g- g- 
8-38 — — 
7-05 0-96 8-01 
711 = 8-07 
7-55 = 8-51 
7-48 0-98 8-46 
8-58 Pm 9-56 
8-41 ~ 9-39 
8-67 ‘ 9-65 
8-55 0-98 9-53 
10-12 — — 
9-71 1-45 11-16 
10-46 BS 11-91 
10-21 - 11-66 
10-19 11-64 
10-82 ” 12-27 
8-09 - 9-54 
9-60 7 11-05 
9-71 a 11-16 
10-28 oe 11-73 
9-78 1-45 11-23 
6-33 — — 
5-66 1-28 6-94 
5-55 a 6-83 
5-35 os 6-63 
5-70 Ae 6-98 
6°35 és 7-63 
5-04 es 6-32 
5-61 1-28 6-89 
5-16 — —- 
4-54 1-09 5-63 
3:99 ne 5-08 
5-57 2. 6-66 
5:28 i 6-37 
5-29 o 6:38 
4-71 BS 5-80 
4-63 mt 5-72 
4-83 Be 5-92 
4-06 Pe 5:15 
4:65 Pr 5-74 
4:24 a 5:33 
4-52 1-09 5-61 
15-79 — —_ 
10-78 — _— 
6-46 — —_ 
5:83 —_ _ 
4-69 1-00 5-69 
4-00 ss 5-00 
4:10 - 5-10 
5-04 59 6-04 
4:78 a 5-78 
4-72 ve 5°72 
4:56 1-00 5-56 


Table IX (continued) 
Diet: milk. Subject: C.J.M. 
Daily Intake 
Bod N N Calories 
Date weight milk total = 
Period . 1921 k g. g. ilo 
“*N-free” May 27 — — -- 
4. see 62-40 7-63 7-78 47 
” 29 ” ” ”” 
” 30 ” ” ” 
5. Pen: 5 | 8-62 8-77 47 
June 1 ” Ps * 
” 2 61-30 ” ” ” 
” 3 ” ” ” 
Average of last 3 days 8-62 8°77 47 
June 16 Mixed diet 
6. a ST 61-80 11-32 11-44 46 
” 18 ” ” ” 
” 19 ” ” ” 
” 20 ”? ”? ” 
” 21 ” ” ” 
” 22 ” »” ” 
”? 23 ” ” ”? 
” 24 ” ” ” 
” 25 61-70 » ” ” 
(26. vi) 
Average of last 6 days 11-32. 11-44 46 
July 2 Mixed diet 
eS 61-10 6-65 6-84 57 
” + ” ” ” 
” 5 ” ” ”? 
” 6 ” ” ” 
” 7 ” ” ” 
” 8 61-80. ” ” ” 
Average of last 4 days 6-65 6-84 57 
July 9 62-10 — 3-00 54 
fe 5-16 5-28 55 
” 1 1 ” ” 36 
” 12 61-00 ” ” 36 
2» 8 60-80 ” ” 53 
» 14 61-10 x Fi 57 
seeks? BEBO ma a 55 
» 16 61-60 et ae 54 
eee |y | 61-90 ma fe 53 
ee SPER jy a“ 52 
» 19 61-85 5-04 5-18 52 
eter) 62-00 fe 2 52 
Average of last 6 days 5-12 5°25 53 
9 Aug. 16 ~~ 61-60 4:24 4-58 53 
” 17 61-85 ” ” ”? 
> 18 61-85 ” ” ” 
” 19 61-90 ” ” ” 
” 20 ” 4-39 ” 
” 21 61-80 ” ” ”? 
” 22 61-80 ” ” ” 
” 23 61-60 ” ” ”? 
” 24 61-60 ” ” 58 
pee 61-75 5 a 55 
Average of last 6 days 4-2 4-39 55 


433 


Balance 


434 C. J. MARTIN AND R. ROBISON 


Table X. 
Diet: milk. Subject: C.J.M. 
Intake 

Z A —, Balance Biological value 

Total N Absorbed N Calories per N —_—_— 

Period g. g. kilo g. (1) (2) 
1, 3-42 3°30 49 — 0-99 74-7 73-6 
2. 4-17 ~ 4:12 44 -1-41 47-3 46-6 
3. 6-35 6°34 48 _ — 0-87 38-4 38:3 
5. 8-77 8-77 47 — 0-76 26-3 28-6 
6 11-44 11-16 46 +0-21 33-0 31-2 
7. 6-84 6-73 57 — 0-05 49-7 48-8 
8. 5-25 5-25 53 — 0°36 55-6 57:2 
9 4:39 4-39 55 -1:17 45-2 _ 49-4 
Average for last 3 periods 50-2 ee 

Diet: milk. Subject: R.R. =u 

1. 2-98 2-98 48 — 0-65 66-0 83-6 
2. 6-29 6-24 48 — 1-44 - 25-0 24-4 
3. 13-69 13-45 47 +0-11 24-2 22-9 


Table XI. Basal metabolism of C.J.M. on normal diet and on carbohydrate 
and fat (= 55 cals. per kilo) with 4-4 g. milk N. 


Diet during Oxygen consumed Calories per 

Date previous 24 hrs. per min. ce. R.Q. 24 hrs. 
Aug. 10 Normal 225-6 : 0-711 1557 
io ms 236-3 0-928 1699 
» 3 iat 227-1 0-684 1566 
> ae e 220-9 0-820 1557 
Average 1595 
nae Milk 211-5 0-974 1529 
» 21 (439 g. N, 55 cals. 178-0 1-159 1359 
9» 22 per kilo) 197-3 1-093 1478 
oe ‘ 185-7 1-152 1415 
ae we 195-0 1-044 1440 
i ae “ 196-1 1-055 1453 
» ae os 210-4 1-017 1540 
Average 1459 
ow ae Normal S819; 0-915 1590 
ae » 216-8 0-889 1550 
iene | » 228-0 0-811 1604 


Average 1581 


During period 9 the subject’s basal metabolism was determined on waking 
and the results were compared with similar determinations carried out during 
previous and succeeding periods when the diet was normal. These results are 
set out in Table XI and show a decrease of about 8 °% in the basal metabolism 
on the milk diet. 


BIOLOGICAL VALUE OF PROTEINS 435 


DISCUSSION OF RESULTS. 


Tue Mintmum NItTRoGEN EXPENDITURE. 


During the second experiment to determine our nitrogen minimum the 
output of nitrogen in the urine fell to an amount equal to 0-035 g. per kilo 
body weight in the case of C.J.M. and 0-034 g. per kilo for R.R. The average 
amounts for the last three days of this period were slightly above the minima, 
being 0-038 g. and 0-035 g. respectively. If the nitrogen of the faeces is in- 
cluded the average output for the same period was 0-057 g. per kilo for C.J.M., 
and 0-055 g. for R.R. The nitrogen intake amounted to 0-005 g. per kilo. 

These figures are somewhat lower than most of those recorded by other 
workers [Landergren 1903; Folin, 1905; Kinberg, 1911; Graham and Poulton, 
1912], but this may be explained by the fact that their diets unavoidably 
‘contained more nitrogen than ours. Karl Thomas [1910] determined his 
minimum expenditure on a purely carbohydrate diet in seven experiments 
carried out over a period of two and a half years and found that this minimum 
fell from experiment to experiment, the final amount for the output in the 
urine alone being 2-2 g. or 0-029 g. per kilo, while that for urine and faeces 
combined was 2:9 g. or 0-039 g. per kilo. At this period Thomas weighed 
75 kilos and had put on a good deal of body fat. In some of his earlier experi- 
ments, however, it seems probable that the minimum output was never 
reached as the diet was continued for too short a period, 

In McCollum’s [1911] experiments on pigs, which were fed on a diet of 
starch, a salt mixture and water, the minimum nitrogen output fell to a level 
corresponding closely with that reached by us. Thus a pig weighing 68-4 kilos 
excreted 0-039 g. N per kilo in the urine. For smaller animals the output per 
kilo was somewhat greater. 

The determination of the nitrogen in urine and faeces does not, of course, 
give a complete account of the loss of nitrogen from the body. To these must 
be added the loss through hair, beard and nails, through loss of epidermis and 
in sweat. Except for the last named these losses are all very small in amount 
but the loss through the sweat may be considerable. Benedict [1906] has 
shown that a resting man may excrete 0-071 g. N per day in this way while 
with moderate work the loss may amount to 0-13 g. N per hour. McCollum 
considers that the nitrogen of the faeces should also be classed with these as 
representing losses that may be termed accidental in character and that the 
nitrogen of the urine alone is to be taken as representing the essential tissue 
metabolism of the body. 

The amount of nitrogen excreted in the urine on the successive days of the 
experiments in which the diet was nearly free from nitrogen, is plotted in 
Figs. 2, 3, and 5. The amount diminishes in a fairly regular manner, at first 
quickly and then more slowly until the minimum is reached. The points lie 
on or near the graph of a simple logarithmic equation 


log(y—A)=a-—ke * 


C. J. MARTIN AND R. ROBISON 


Days 
4 


N. minimum 1 N. minimum 2 


(C.5.M.) (C.I.M) 


| } | | J ! | 


I | I | i} | | 


N. minimum 1 N. minimum 2 
(R.R.) (R.R.) 


© 
log (y-2°0) «1:13-0:33 Lx log (y= 2°0) «1198 - 0°366x 


= 


1 2 3 4 5 6 7 8 


BIOLOGICAL VALUE OF PROTEINS 437 


in which z is the number of days, y is the nitrogen output in the urine, and 
A the minimum value of y. A closer agreement is obtained if A is given a 
value slightly lower than the minimum actually reached. 

Thus in Fig. 3 the curve R.R. 1 is drawn from the equation 


log (y — 2:0) = 1-153 — 0:331a 


Urine Nitrogen g. 
1 


\— 


log (y- 1°04) = 1:08 - "182 


| i ae | } | J 

1 2 3 4 5 6 7 8 9 10 
Days 

Fig. 4, E. V. McCollum’s experiment on a pig. Daily output of nitrogen in urine 

on starch diet after ingestion of zein. 


and fits the points reasonably well, but a still closer approximation is obtained 
with the equation —_ Jog (y — 1-73) = 1-1209 — 0-28322 


the agreement between the observed and calculated values of y being extra- 
ordinarily good for all points except the last. 


Bioch, xv1 29 


438 C. J. MARTIN AND R. ROBISON 


Day of y calculated from equation 
experiment log (y — 1-73) =1-1209-0-28322 —_y obtained 
1 8-61 8-64 
2 531 5-31 
3 3°60 3-63 
4 2-70 2-66 
5 2-24 2-25 
6 1-99 2-82 
20 1-73 — 


A question is thus raised: Does this amount 1-73 g. represent the real 
minimum expenditure, which, from some cause was not realised in either 
experiment? At present this must be left unanswered, but further experi- 
ments may give some information on the point. 


Urine N. g. 


8r 


ol log (y-2°66) = 1:210-0'8332, 


l 


5 


Nor 


Days 
Fig. 5. E. V. MeCollum’s experiment on a pig. Daily output of nitrogen in urine 
on starch diet after ingestion of urea, 


The difference between experiments 1 and 2 lies only in the amount of 
total fluid taken which was about 2000 cc. daily in the first and 4000 cc. in 
the second. This makes but slight difference in the value of k, that is in the 
rate at which the nitrogen output falls. Kinberg [1911] has previously drawn 
attention to this regularity but has not attempted to deduce any mathematical 


BIOLOGICAL VALUE OF PROTEINS 439 


expression from his results. Thomas [1910] also recognised that on a protein- 
free diet the nitrogen of the “ Vorratseiweiss” leaves the body with varying 
rapidity according to the amount present, but was unable to find any exact 
relationship either in his own results or in those of Landergren. He calculated 
the amount of “ Vorratseiweiss” excreted, by subtracting the minimum output 
(“ Abnutzungsquota’”’) from the urine nitrogen, but the value of this minimum 
was taken from experiments of four days’ duration and was probably too high. 
Landergren’s results show a fair agreement with the graph of an equation of 
the type given above if A is taken as 2-5 instead of 3-0. Thomas’s results are 
more irregular. In considering the agreement of the results of such experiments 


N N. Intake g 

ait 

15 kK SSS 

. SYS N_ output 
in urine 


1ob Y 


YZ 


e @ 
(oh Deeg ao) EY Ig BIRR! Set Ree SEN Me Cee en | 
August 16 17 18 19 20 21 22 23 24 25 26 27 28 29 
Fig. 6. 


with those calculated from these equations, the tendency of the nitrogen 
output to oscillate even when the nitrogen intake is constant must be borne 
in mind: Such oscillations are present in most of the experiments considered, 
but do not alter the general character of the curve. In McCollum’s experi- 
ments on pigs a similar regularity also appears as may be seen from Figs. 4 
and 5. Fig. 4 shows the nitrogen output on a starch diet after a diet containing 
zein and Fig. 5 shows the output after the ingestion of urea. The curves are 
of the same type but “k,” 7.e. the rate at which the output falls, is very much 


A 


: 


440 C. J. MARTIN AND R. ROBISON 


greater after urea than after zein. This point is of interest as it seems to 
indicate that the store of nitrogen which exists in the body after a protein 
diet, and which is rapidly given up on a diet free from nitrogen is not present 
entirely in the form of urea. As to the form in which this nitrogen occurs, 
very little can be definitely stated. It may be as resynthesised protein (Vor- 
ratseiweiss), or as amino acids adsorbed in the tissues, or compounds of inter- 
mediate complexity. It is certainly present partly as urea and other end 
products of metabolism. If our conclusions as to the rate at which this storage 
nitrogen is excreted are correct the interpretation is, that the amount removed 
from the body on any day is proportional to the amount still present. This 
might hold whether the reaction involved was the hydrolysis of protein, the 
deaminisation of amino acids or simply the washing out from the tissues of 
the end products of nitrogen metabolism. 

On reversing the process, and after a minimal N-intake for ten days, 
suddenly increasing the nitrogenous food consumed, nitrogen at first remains 
in the body and equilibrium between intake and output does not occur for 
several days. Fig. 6 is a graph of the results of an experiment designed to 
show this. The broken line represents intake of N, the solid line output. The 
shaded area on the left hand represents the stored N gradually removed on 
dropping the intake from 17 g. to 4-4 g. and the right-hand area the amount 
again stored on resuming a diet containing 16 g. N. The curves are reciprocal 
and the two shaded areas approximately correspond. 


Tue NATURE OF THE MINIMUM NITROGEN REQUIREMENTS OF THE Bopy. 


The low level to which the nitrogen output falls on a protein free diet is 
evidence of the smallness of the body’s daily requirements in this respect. 
We know from Folin’s [1905] researches that the reduction in the nitrogen on 
such a diet occurs mainly at the expense of the urea fraction, the ammonia 
and uric acid being reduced to a relatively much smaller extent while the 
creatinine remains constant. These facts led Folin to conclude that protein 
metabolism is of two types, (1) “tissue” or “endogenous,” which tends to be 
constant, and is represented largely by such products as creatinine, neutral 
sulphur, and to a less extent by uric acid, (2) “exogenous” which varies with 
the amount of protein consumed and is represented chiefly by urea. The 
nitrogen required for processes of the first type is alone essential, but Folin 
recognised that equilibrium at this low level may not be possible since a certain 
amount of protein may always fall prey to the exogenous metabolism. The 
distribution of the nitrogenous constituents of the urine was determined during 
our second minimum experiment and the milk diet immediately succeeding this. 

The results | Robison, 1922, 2] are similar to those of Folin. The minimum 
nitrogen output was lower than any recorded by him, and the percentage of 
urea was correspondingly reduced, the minimum figure being 37-2 %, while 
the sum of the urea and ammonia amounted to 54-8 °%, of the total nitrogen. 


BIOLOGIGAL VALUE OF PROTEINS . 441 


The multifarious transactions involved in endogenous metabolism are not 
likely to be conducted with perfect economy. When much protein is hydrolysed 
and the products mobilised and used for the synthesis of proteins of another 
composition such as those of the blood or for the manufacture of thyroxin or 
adrenaline it is unlikely that the whole balance of unwanted amino acids 
escapes deaminisation and conversion. Leakage of this kind may account 
for no inconsiderable fraction of the minimum nitrogen expenditure and it 
is perhaps in this direction that, by adaptation, the body may effect some 
saving. The experiments of Thomas and Hindhede would appear to show that 
this does in fact occur. 

The amount of carboliydrate eaten may also influence the degree of this 
waste since the process of deaminisation is reversible and will be affected by 
the rapidity with which the non-nitrogenous products, hydroxy or ketonic 
acids, are removed by oxidation or conversion into carbohydrate. It is pro- 
bable that an excess of carbohydrate in the blood would retard either action 
and in consequence deaminisation. 


THE QUESTION OF UNIFORMITY OF THE BIOLOGICAL VALUE. 


We have seen that the validity of the method adopted by Thomas for the 
determination of the biological values of proteins depends in the first place 
on the uniformity of this value when varying amounts of the same protein 
are consumed. The investigation of this question was one of the objects of 
our experiments and the results must now be considered from this point of 
view. The three proteins so far studied were chosen on account of the wide 
difference in the values attributed to them. The biological values of milk 
and wheat proteins as given by Thomas are 99-71 and 39-56 respectively, 

‘while gelatin has long been known to be deficient in several essential amino 
acids, so that no amount, however great, can completely satisfy the body’s 
nitrogen requirements. 

The results of the experiments described in this paper together with those 
on gelatin, previously recorded by one of us [Robison, 1922, 1] are plotted in 
Figs. 7 and 8, in which the values of the nitrogen intake for different periods 
are the abscissae and the corresponding amounts for the total output in urine 
and faeces are.the ordinates. For this purpose the nitrogen of the faeces in 
excess of the amount on the “protein-free” diet has been taken as repre- 
senting unabsorbed food and has been subtracted from both intake and output. 

In the case of whole wheat proteins, the results are strikingly similar 
with both individuals and the points obviously lie on or close to a straight 
line passing through the points representing the minimum expenditure on 
the low nitrogen diet. The agreement is remarkably good considering the 
errors inseparable from such experiments. The results for periods 4 and 5 show 
some discrepancy but these periods were continuous and should probably be 
considered as a whole, the only difference in the diets being the inclusion of 


26—3 


C, J. MARTIN AND R. ROBISON 


442 


‘S'S 
B eyequy uatouyin 
Z 


‘S eyequy uatosjiy 


Z 


"L “BI 


9 


S 


v € 4 


ESS 
Sie aii Pees ied Recee Se3 ae 


unejan O 
way 914M @ 


a: Oe 


| 


t 
_—joO 
N 


an O 
FeO AA FOUN co 


Wro 


—4 


BIOLOGICAL VALUE OF PROTEINS 443 


a small quantity of apples in the former and its omission from the latter. 
The average for the two periods gives a point lying close to the line. The 
uniformity of the biological value is therefore satisfactorily proved for the 
proteins of whole wheat. 

In the case of gelatin also the points fit a straight line reasonably well, 
but if this line passes through the point representing the nitrogen minimum 
(as drawn in the figure), it will at some distant point intersect the “equili- 
brium line” (at 45° to the axis), 7.e. nitrogenous equilibrium will be attained 
at this point, which we believe to be impossible. On the other hand if it is 
drawn parallel to the “equilibrium line” it will cut the axis of y at a point 
slightly below the observed minimum. We must conclude, therefore, either 
that the line is slightly curved near the minimum or that the observed value 
of the latter is somewhat higher than the real one. If the capacity of the 
gelatin is limited to the reduction of “leakage” these two alternatives have 
practically the same significance. 

Milk presents a more difficult case. Only the results for those periods in 
which a diet of. abundant fuel value was taken (C.J.M. 1, 7, 8, 9) have been 
plotted but even these do not agree well with any straight line or other 
regular curve. The most that can be said is that in view of the possible errors 
in such experiments where the intake is small, the results are not inconsistent 
with the uniformity of the biological value of this protein. 

From the consideration of these three cases we may conclude that the 
general assumption of this principle made by Thomas, occasioned no serious 
errors in his conclusions. Caution must be used in extending this principle 
to all cases without investigation and reliance should not be placed upon 
results of experiments in which the negative balance is large. 


THE BroLoGIcAL VALUES OF THE PROTEINS OF WHOLE WHEAT AND MILK. 


Whole Wheat. 


The biological values calculated from the different periods of our experi-. 
ments with whole wheat proteins agree very well amongst themselves and give 
an average of 35 for those obtained on C.J.M. and 31 for those obtained on 
R.R. Thomas’s value is somewhat higher, 39-56. The latter figure was based 
on the result of two experiments each of three days’ duration and one of four 
days. During none of these experiments was the nitrogen intake constant; 
in one, the amounts for the separate days were 4-0 g., 7-3 g., 9-0 g. respectively. 
The minimum requirements corresponding with these three periods were taken 

_as 4-63 g., 3-991 g. and 3-316 g. N (urine only), these amounts being determined 
in experiments of four and three days’ duration. In the first of these the nitrogen 
output on the four days was 18-32 g., 10-17 g., 7-39 g., 4-63 g. but there is no 
evidence that the last figure represented Thomas’s minimum requirements. 
The biological value of wheat proteins was calculated from each separate 


444 C. J. MARTIN AND R. ROBISON 


day’s balance and out of these widely varying results those for certain days 
were selected in a somewhat arbitrary fashion. The experimental results 
of his third period, in which the intake was nearly constant, indicate a similar 
biological value for wheat proteins to that found by ourselves. 

The result Hindhede [1913, 1] obtained upon F. Madsen with white bread, 
namely a positive balance of about 1 g. per day over a 28 day experiment 
on 13-73 g. N was certainly not a minimum quantity and the amount necessary 
for N equilibrium may be considerably less. His further experiments [1914] 
were made with rye bread (“Schwarzbrot”) and the diets contained con- 
siderable amounts of fruit or vegetables; which accounted for 0-5 g. to 1-9 g. N. 
These experiments were carried out in duplicate on F. and H. Madsen and 
were continued for four months so that they possess great value. In the final 
period of six days the fruit was omitted and positive balances were obtained 
on 13-52 g. (F.M.) and 10-68 g. (H.M.) bread nitrogen respectively: By sub- 
tracting the whole of the nitrogen in the faeces Hindhede calculates that the 
amounts absorbed were 8-49 g. and 6-28 g. respectively. We do not agree 
with this method of treating the faecal nitrogen and cannot accept his con- 
clusion that bread proteins possess equal value with those of potatoes, of 
meat and of the body; but there is no doubt that in his experiments, equili- 
brium was obtained on a lower intake of nitrogen in the form of rye bread 
(and still lower when supplemented by fruit) than our minimum for whole 
wheat. This is confirmed by the experiments of Neumann [1919], who, by 
long continued diet of very high calorie value, was ultimately able to retain 
nearly 3g. N daily with an intake of 9-9 g. N in the form of rye proteins. 
Neumann’s experiment was upon himself, and was in every way unexcep- 
tionable, but it may perhaps be significant that it was preceded by a prolonged 
period of semistarvation during which he was investigating the German 
civilian ration of 1916-17. This is a further indication that, apart from the 
influence of loss of body weight, the organism can gradually accommodate 
itself to a lower ration of nitrogen, perhaps by the exercise of greater economy. 

Abderhalden’s [1915] observations on Rése, although as pointed out 
earlier in this paper, not susceptible of the interpretation he places upon them, 
do indicate that the latter could get into equilibrium with about a gram less N 
in the form of white bread than we could with bread made from the whole 
grain. Rése’s nitrogen expenditure on a nitrogen-free diet was not ascertained 
80 we cannot estimate the biological value for this diet. 

The recent observations of Rubner [1919] upon the proteins of white flour 
are interesting in relation to our own, because they show, we think, that from 
the point of view of biological value, the proteins of the endosperm are equal, 
if not superior to, those of the whole seed. 

In most of Sherman’s experiments on the value of cereal proteins these 
are supplemented by a certain amount of milk and the results are not directly 
comparable with ours, but in some experiments on white bread [1920] the 
diet consumed would not appear to be greatly different from that taken by 


BIOLOGICAL VALUE OF PROTEINS 445 


us and we are therefore the less able to explain the difference between his 
results and ours. Sherman’s subject, a man weighing 80 kilos, attained equili- 
brium on a diet containing 6-0 g. N over 95 % of which was derived from 
white bread and the remainder from apples and butter. The energy value 
was only 34 calories per kilo. The bread was purchased from a bakery and 
probably contained a small amount of milk but how much was not known. 


Milk. 


The results of our experiments with milk proteins do not agree so closely 
as those for the whole wheat. The very low values calculated from the results 
of periods 3, 5, 6 (C.J.M.) and 2, 3, (R.R.) are not easy to explain. That they 
are in some way due to the lower calorie value of the diet seems clear but this 
was in no case below that of our normal diet and amply covered our energy 
requirements. The high values obtained with both subjects in period 1 are 
perhaps connected with the previous nitrogen starvation and a consequent 
increase in the economy with which the body may deal with the protein 
supplied to it. 

If we consider only periods 7, 8, 9 (C.J.M.) in which the conditions were 
the same and the energy supply abundant the biological value for milk proteins 
is equal to 51 %. This value is only half of that found by Thomas, but the 
criticisms we have made in discussing his experiments with bread apply with 
still greater force in this case. His value (100%) was calculated from the 
nitrogen balance on a single day of an experiment lasting only two days 
on which the intake was 6-24 g. and 7-28 g. respectively. The minimum 
requirements were taken as 3-99 g. which is probably much too high. If the 
value of milk protein were as high as Thomas makes out it is difficult to see 
how he could explain the large negative balances occurring in his experiment 
with “Frauenmilch” (cow’s milk with extra sugar and cream). During the 
first two days of this experiment the fuel value of his diet was obviously too 
low, but in the last three days it was equal to 40-45 calories per kilo, the 
N-intake being 15-3 g.—17-3 g. yet the negative balance was never less than 
1-0 g.. The results of this experiment appear to agree with our experience, . 
and to suggest that a high intake of milk nitrogen tends to result in increased 
expenditure of body nitrogen, unless the fuel value of the diet is raised very 
much above the normal. So far as we are aware the value of milk protein 
has not been the subject of any other investigation on man. 


CONCLUSIONS. 


From our observations upon ourselves we conclude: 

(1) That our minimum nitrogen expenditure by the urine is somewhat 
less than 0-038 g. and 0-035 g. per kilo in C.J.M. and R.R. respectively. 

(2) That on taking a diet of carbohydrate and fat of adequate calorie- 
value the nitrogen excreted in the urine falls in a regular and orderly manner, 


446 ©. J. MARTIN AND R. ROBISON 


capable of simple mathematical expression, approaching a minimum in five 
to seven days. On resuming an ordinary nitrogenous diet the reciprocal 
phenomenon occurs. 


(3) Bearing in mind the considerable experimental errors, the ratio 
Body N saved 
Food N absorbed 
taken in the form of whole wheat bread, until equilibrium is reached. 


(4) In the case of milk the experimental errors are proportionately greater 
and the most we can say is that this ratio may remain constant. 

(5) In the case of gelatin the ratio certainly does not remain constant 
and there is no indication that the amount of body nitrogen saved increases 
- beyond that effected by the smallest quantity of gelatin fed. 

(6) The application of Thomas’s method of determining biological values 
is justified in the case of bread, doubtful with milk and impossible with 
gelatin. 

(7) Until Thomas’s procedure has been ascertained to be justifiable for 

Body N saved 
Food N absorbe 
close to, but below, the point of equilibrium. 

(8) The mean biological value of the nitrogen contained in the whole 
wheat grain as determined by six experiments on each of two adults was 
35 % (C.J.M.) and 31 % (R.R.). 

(9) The mean biological value of the nitrogen in cow’s milk, derived from 
three experiments upon C.J.M. in which an excess of calories (55 per kilo) 
was taken, was 51 %. 

(10) Biological values arrived at from experiments of comparatively short 
duration, however well justified, have a limited significance. 


appears to remain constant. whatever amount of nitrogen is 


the particular proteins concerned, the ratio q Should be determined 


REFERENCES, 


Abderhalden and Rona (1904). Zeitsch. physiol. Chem. 42, 528. 
(1906). Zeitsch. physiol. Chem. 47, 397. 
Abderhalden and Samuely (1905). Zeitsch. physiol. Chem. 46, 193. 
Abderhalden, Fodor and Rése (1915). Pfliiger’s Archiv. 160, 511. 
Albu (1901). Zeitsch. klin. Med. 48, 75. 
Atwater and Benedict (1902). Mem. Nat. Acad. Science, 8, 233. 
Benedict (1906). J. Biol. Chem. 1, 263. 
Bornstein (1898). Berl. klin. Woch. Nr. 36. 
Boruttau (1915). Biochem. Zeitsch. 69, 225. 
—— (1919). Biochem. Zeitach. 94, 194. 
Chittenden (1904). Physiological Economy in Nutrition, New York. , 
Cohnheim (1901), Zeitsch. physiol. Chem. 38, 451. 
—— (1906). Zeitsch. physiol. Chem. 49, 64. 
Falta (1906), Arch. klin. Med. 86, 517. 
Folin (1905). Amer. J. Physiol. 18, 66 and 117, 
Goldberger, Waring and Willets (1915). U.S., P.H. Reports. Reprint’ 307, 5. 
Goldberger, Wheeler and Sydenstricker (1920). U.S., P.H. Service Reports, 35, 648, 
Graham and Poulton (1912). Quart. J. Med, 6, 82. 
Grubner (1901). Zeitech. Biol. 42, 407, 


BIOLOGICAL VALUE OF PROTEINS 447 


~ Hindhede (1913, 1). Skand. Arch. Physiol. 28, 165. 
—— (1913, 2). Skand. Arch. Physiol. 30, 97. 
—— (1914). Skand. Arch. Physiol. 31, 259. 
Hirschfeld (1887). P/fliiger’s Archiv. 41, 533. 
—— (1889). Virch. Archiv. 114, 301. 
Hopkins and Cole (1901). J. Physiol. 27, 418. 
Kinberg (1911). Skand. Arch. Physiol. 25, 291. 
Klemperer (1889). Zeitsch. klin. Med. 16, 550. 
Kossel (1901). Ber. 34, 3214. 
Kossel and Kutscher (1900). Zeitsch. physiol. Chem. 31, 165. 
Kumagawa (1889). Virch. Archiv. 116, 370. 
Kutscher and Seemann (1901). Zeitsch. physiol. Chem. 34, 527 
Landergren (1903). Skand. Arch. Physiol. 14, 112. 
Lapicque (1894). Arch. Physiol. Norm. et Path, 26, 596. 
Lapicque and Marrette (1894). C.R. Soc. Biol. 10 ser. 1, 274. 
Loewi (1902). Arch. Exp. Path. Pharm. 48, 303. 
McCollum (1911). Amer. J. Physiol. 29, 210. 
McCollum and Hoagland (1913). J. Biol. Chem, 16, 317. 
McCollum, Simmonds and Pitz (1917). J. Biol. Chem. 29, 341. 
McCollum, Simmonds and Parsons (1921). J. Biol. Chem. 47, 111, 139, 175, 207, 235. 
Michaud (1909). Zeitsch. physiol. Chem. 59, 405. 
Neumann (1899). Arch. Hyg. 36, 248. 
—— (1902). Arch. Hyg. 45, 1. 
—— (1919). Vierteljahrschrift gericht. Med. 52. 
Osborne and Harris (1903). J. Amer. Chem. Soc. 25, 853. 
Osborne and Mendel (1911). Pub. no. 156. Carnegie Inst. Washington. 
—— —— (1912-1920). J. Biol. Chem. 12 to 45. 
. —— —— (1917). J. Biol. Chem. 29, 69. 
Peschel (1891). Inaug. Dissert. Berlin. 
Robison (1922, 1). Biochem. J. 16, 111. 
—— (1922, 2). Biochem. J. 16, 131. 
Rubner (1879). Zeitsch. Biol. 15, 115. 
—— (1897). V. Leyden “Handbuch der Ernaihrungstherapie.” 
—— (1915). Arch. Physiol. 145. 
—— (1919). Arch. Physiol. 81. 
Sherman (1920). J. Biol. Chem. 44, 97. 
Sherman, Wheeler and Yates (1918, 1). J. Biol. Chem. 34, 383. 
Sherman and Winters (1918, 2). J. Biol. Chem. 35, 301. 
Sherman, Winters and Phillips (1919). J. Biol. Chem. 39, 53. 
Sivén (1900). Skand. Arch. Physiol. 10, 91. 
Thomas (1909). Arch. Physiol. 219. 
—— (1910). Arch. Physiol. Suppl. 249. 
Voit, C. (1866). Zeitsch. Biol. 2, 307. 
—— (1867). Zeitsch. Biol. 3, 1. 
Voit, E. and Korkunoff (1895). Zeitsch. Biol. 32, 58. 
Willcock and Hopkins (1907). J. Physiol. 35, 88. 
Wilson (1921). J. Hygiene, 20, 1. 
Zeller (1914). Arch. Physiol. 213. 
Zisterer (1910). Zeitsch. Biol. 58, 157. 


XXXIV. ON THE CHANGE OF THE OSMOTIC 
PRESSURE OF SOLUTIONS OF CERTAIN 
COLLOIDS UNDER THE INFLUENCE OF 

SALT SOLUTIONS. 


By DAIZO OGATA (Fukuoka). 
From the Department of Biochemistry, Oxford University. 


(Received March 17th, 1922.) 


In spite of certain controversies among investigators on the osmotic pressure 
exerted by colloidal solutions the positive proof that a definite, though small, 
osmotic pressure is set up was given by Starling [1896], Moore and Parker 
[1902], Lillie [1907], Moore and Roaf [1907] and others. 

This osmotic pressure is affected by several factors, such as the mode of 
the formation of the colloidal solution [ Lillie, 1907], the change of temperature 
[Moore and Roaf, 1907], the acidity or alkalinity of solution [Adamson and 
Roaf, 1908]. Further that the added salts also increase or diminish the size 
of the colloidal particles before any visible precipitation takes place, 1.e. 
lower or raise the osmotic pressure, has been proved by Biltz and Vegesack 
[1909], Lillie [1907], Moore and Roaf [1907] and Loeb [1918]. 

The present research was started, by suggestion of Prof. Moore and under 
his direction, to prove by means of osmotic pressure measurements whether, 
first, salt solutions affect the “solution aggregate” of different colloidal solu- 
tions always in the same way, and, secondly, what is the influence, if any, 
of certain anaesthetics upon colloidal solution. While definite and regular 
results were obtained in the former experiments, those in the latter were not 
at all decisive. 


METHODS. 


As colloidal solutions gum arabic (1%), egg albumin (5 %) and sheep 
serum (half diluted) were used. The membrane (dialyser) was of collodion 
and made on the outside of a large test tube. This was fixed to a well-fitting 
rubber cork with a strong string. The cork had two holes, one for a glass 
manometer tube and the other for an inlet and outlet tube. The length of 
the membrane was so adjusted as to limit its content to approximately 15 cc. 

In the earlier part of the experiments the colloidal solution only was 
introduced into the dialyser, ¢.e. without mixing previously with any salt 
solution, thus letting the latter penetrate gradually from the outside. But 


Bioch. xv1 30 


4 


450 D. OGATA 


the procedure became more elaborate with the progress of the experiments, and 
the final method consisted of mixing the colloidal solution with the salt solution 
just before setting up the apparatus and then introducing into the membrane. 
The membrane was immersed up to the upper surface of the cork in a definite 
volume (120 cc. as a rule) of salt solution contained in a beaker. The mano- 
meter tube, together with a millimetre scale, was clamped in a vertical position 
during the experiments. The salts employed were copper sulphate, calcium 
chloride and alkaline sodium phosphate respectively. 

The experiments were performed in series, e.g., 1/10 N, 1/100 N, 1/1000 N 
solution and solvent (distilled water) only, the last as a control. The same 
salt solution of the same concentration was always placed inside and outside 
the membrane. 

Usually the osmotic pressure was not very high, therefore the ponaiferstion 
of the distension of the membrane was of little importance. 

As the salt solution could easily pass through the membrane the column 
of solution in the manometer increased in length, at first, very rapidly. After a 
certain time it began to fall, hardly staying at the highest point any measurable 
length of time. Then the column fell very slowly to a certain level at which 
it remained for a considerable time. When almost the same level was kept 
for at least two days, it was understood, in my experiments, that the expected 
equilibrium of osmotic pressure was attained. In most cases when this 
equilibrium was reached the outside liquids were replaced with distilled water, 
four or five times in succession with an interval of one day or more. The 
reversibility of pressure was fairly satisfactory. 


RESULTS OF EXPERIMENTS. 


The following are the examples of the abridged records of experiments 
carried out during five months. Similar results were always obtained when 
due precautions were taken. 


Experiments with gum arabic solution. (See Fig. 1.) 


Inside the membrane: 1 % gum arabic solution, 15 cc. 
Outside the membrane: salt solutions of different concentrations, 120 ce. 


Table I. 80,Cu. 


Osmotic pressure 
Day Hour Temp. c A — 
( ‘ 
Sept 1 atncon wot wp) {SAO loo. oon dai 
ge 4.00 p.m. 189° C, 2-5 mm. 13-2 mm. 48-0 mm. 163-3 mm. 
> es 2.30 16-9 75 14-0 45-1 150°5 
ie 4.00 18-9 8-5 12:8 42-0 135-5 
6 3.20 18-3 9-0 13-0 42-2 130-0 
Pa "of the outside solution, about 9-0 7-2 65 65 


The reversibility was tested: 
Sept. 13 10.00 a.m, 17:3 92-0 100-0 106-0 120-0 


The curves in Fig. 1 indicate the rate at which equilibrium was reached both in direct and 
“reversed” experiments, 


OSMOTIC PRESSURE OF COLLOID SOLUTIONS 451 
Table II. Cl,Ca. 


Osmotic pressure 
Day Hour Temp. c A —~ 
ee 1/10 N 1/100 N 1/1000 N distilled 
Sept. 2 about noon (set up) | Ca Cl,Ca Cl,Ca och 
oe 3.30 p.m. 17-5° C. 12-3 mm. 17-0 mm. 50-0 mm. 108-5 mm. 
pu of the outside solution, about 6-0 6-5 6-0 6-5 
The reversibility was tested: 
Sept. 13 10.00 a.m, 16-9 93-0 94-0 113-0 117-5 
mm. 


200 


Osmotie Pressure. 
= = 
8 ra 


a 
oO 


I i 


Hour™ ioj224 
Diag % ¥o 40 59 §o fe Yo ¥9 '% 1Y9 149 Yo 

Fig. 1. Exp. with gum arabic solution. Effect of SO,Cu. At the : line on the 

abscissa, the outside liquid was replaced by dist. water in succession. 


Table III. PO,HNa,. 


Osmotic pressure 
Day Hour Temp. l ‘ = Cmca 
: 1/10 V 1/100 NV 1/1000 istille 

Sept. 8. morning (eet up) { PO,HNa, PO,HNa, PO,HNa, water 

rama: 12.00 17-2° C, 15-2 mm. 270mm. 110-5 mm. 149-0 mm. 
pu of the outside solution, about 9-0 7-2 6-5 6-5 

The reversibility was tested: 
Sept. 13 10.00 a.m. 17-2 251-0 220-0 129-0 130-0 
Experiments with egg albumin. 


Inside the membrane: 5 %, egg albumin solution, 15 cc. 
Outside the membrane: salt solutions of different concentrations, 120 cc. 


Table IV. SO,Cu. 


Osmotic pressure 
A 


Day Hour Temp. — — 
é {1/2000 N—_1/20,000 N_ _ 1/200,000 N_—s distilled 
Rept; 18° morning {set 'up) 1 ‘S0,Cu S0,Cu S0,Cu obra 
>» 20 12.00 16°1° C. 12-0 mm. 53-9 mm. 55-1 mm. 52-2 mm. 
pu of the outside solution, about 6-0 6-0 6-0 6-0 
The reversibility was tested: 
Sept. 27 4.00 p.m, 16-0 36-0 64-0 62-0 59-0 


30—~2 


452 D. OGATA 
Table V. Cl,Ca. 


Osmotic pressure 
A 


Day Hour Temp. “asatill a 
Sept. 16 afternoon (set up) pms Me 1/ Chca ae 
» 20 12.00 15-7° C. 14-0 mm. 17-0 mm. 36-0 mm. 50-0 mm. 
pu of the outside solution, about 6-75 6-0 6-0 6-0 
The reversibility was tested: 
Sept. 27 4.00 p.m. 15-5 40-0 41-0 49-0 58-0 


Table VI. PO,HNag,. 


Osmotic pressure 


Day Hour Temp. = me Pane, 
{ 1/20N 1/200 N 1/2000 istille 
Sept. 16 afternoon (set up) | PO,HNa, PO,HNa, PO,HNa, water 
oe 2.00 p.m. 16-5° C. 22-6 mm. 30-0 mm. 44-0 mm. 47-0 mm. 
pu of the outside solution, about 9-0 7-2 6-75 6-5 
. The reversibility was tested: 
Sept. 27 4.00 p.m. 15-9 40-0 48-0 51-0 ~ 59-0 


Experiments with blood serum. 


Inside the membrane: sheep serum half diluted, 15 cc. 
Outside the membrane: salt solutions of different concentrations, 120 ce. 


Table VIL. S0,Cu. 


Osmotic pressure 
A. 


ens reas —e N 1/200,000 N distilled 
1/2000 N___1/20,000 1/200, istille 
Sept. 30 afternoon (set up) ea S0,Cu * —80,Cu S0,Cu ‘eratiiet 
Oct. 5 12-00 16-6° C, 161-0mm. 189:0mm. 192-0 mm. 184-0 mm. 
pu of the outside solution, about 6°75 7-2 7-2 7-2 
Sulphate in 100 ce. of the outside 
solution as SO, _... ye 5:36 mg. 2-86 mg. 0-06 mg. 0-00 mg. 
The reversibility was ented 

Oct.17 10.00 a.m. 13-4 115-0 168-0 163-0 170-0 


Table VIII. Cl,Ca. 


Osmotic pressure 
A. 


a api tek 1/20 N 1/200N 1/2000 N_ distilled 
A , istille 
Sept. 30 afternoon (set up) C1,Ca Cl,Ca C1,Ca whiteg 
Oct. 5 12.00 16-0° C, 105-0mm. 120-0mm. 169-0 mm. 176-0 mm. 
pu of the outside solution, about 7:2 7-2 72 72 
Chloride in 100 ce. of the outside 
solution as ClNa . ode eae 37-7 mg. 11-5 mg. 9-25 mg. 8:0 mg. 


The reversibility was tested: 
Oct. 17 10.00 a.m. 13-1° C. 178-0 mm. 1685 mm. 1550 mm. 163-0 mm. 


Table IX. PO,HNay. 


Osmotic pressure 
A. 


Day Hour Temp. 


af ars | 
: , { 1/20 N 1/200 N 1/2000 NV distilled 
Oct, l morning (set up) | PO,HNa, PO,HNay PO,HNa, water 
a 5 12.00 165° C, 133-5 mm. 163-0mm. 166-0 mm, 176-0 mm, 
pu of the outside solution, about 9-0 72 6:75 6-75 
Phosphate in 100 ce, of the ee 
solution aa P,Os ... bie «+ 204-0 mg. 27:2 mg. 8-0 mg. 3-0 mg. 


The reversibility was tested: 
Ort. 17 10.00 a.m, 13-2° ©, 690mm. 101-0mm.  137:0 mm. 158-0 mm, 


OSMOTIC PRESSURE OF COLLOID SOLUTIONS 453 


Unless specially mentioned the protein did not appear in the outside 
‘solution through the membrane (acetic acid and potassium ferrocyanide test). 

As clearly seen in the above tables the salts used have all a more or less 
prominent influence upon the osmotic pressure exerted by the colloidal solu- 
tions such as gum arabic, egg albumin or blood serum. The salt solutions 
always reduced the osmotic pressure of the colloidal solution, and the more 
dilute the salt solution, the less decrease was brought about, thus gradually 
approaching the control with distilled water only. This fact suggests, at once, 
the important influence of electrolytes, because within the limits of concen- 
tration of the salt solutions employed here it may safely. be concluded that 
the more concentrated the solution, the greater will be the number of elec- 
trically dissociated molecules. The trouble is the appearance of precipitation 
in the long run of experiments of this kind, and, therefore, all the statements 
here are based necessarily on the comparison with the control otherwise 
under the same conditions. Of copper sulphate an extremely dilute solution 
was made use of simply to avoid the unnecessary precipitation as far as 
possible. 

The temperature, no doubt, affects the height of the osmotic pressure, 
but this effect is rather slight compared with that due to the electrolytes 
themselves. Notwithstanding the fluctuation of osmotic pressure parallel to 
that of temperature, the main course of change of osmotic pressure cannot 
be ignored. 

The hydrogen-ion is neither the sole nor the principal cause in these cases, 
since the osmotic pressure does not always keep pace with its concentration. 
To quote a few examples, the hydrogen-ion concentration was about the 
same through the series in Tables II, IV, VII and VIII respectively. Never- 
theless the pressure difference is very notable. 

The more or less prominent reversibility of the pressure leads us to seek 
the probable cause in the physical reaction rather than the chemical change, 
as was propounded by several previous investigators. Mayer, Schaeffer and 
Terroine [1907] observed under the microscope that the submicroscopical 
granules of colloidal solutions (metal colloid, starch, albumin and others) dis- 
appeared on alkalisation and appeared on acidification. This is one of the 
convincing proofs that colloid aggregation can change at different reactions. 
It, moreover, strongly suggests that the decrease of the osmotic pressure of 
colloidal solutions under the influence of electrolytes may be most conveniently 
ascribed to a similar phenomenon. 

So far as the results of our experiments were concerned the effects of three 
salts on the three colloidal solutions showed no qualitative difference among 
themselves. — 

Briefly speaking I am inclined to think, with previous investigators, that 
the diminished osmotic pressure under the influence of salt solutions depends 
upon an increased aggregation of colloidal particles. 

The effects of anaesthetics (chloroform, ether, alcohol) were tested in the 


454 D. OGATA 


same way. The only difference of the method lay in the use of a flask, instead 
of a beaker, to prevent the evaporation of these anaesthetics, connection 
with the manometer tube being made by a side tube. The solution inside 
the membrane was 15 cc., outside about 500 cc. The concentrations of 
anaesthetics were not always the same. Saturated solutions of chloroform in 
distilled water, 10 % ether, and 5 % alcohol were the most concentrated 
solutions in the series of experiments in each case. Roaf and Alderson [1907] 
found that anaesthetics such as chloroform, ether and carbon dioxide detached 
no salts from blood serum, which differed thus from other tissues (red blood 
corpuscles, brain, liver, muscle, kidney). Moore and Roaf [1907] observed 
that the presence of anaesthetics (chloroform) or organic solvents (benzene) 
did not alter the osmotic pressure of the serum protein (pig’s serum). In their 
experiments the serum saturated with chloroform was placed against saline 
- (0-175 %) also saturated with chloroform. 

Despite the imperfections of the present osmotic pressure determinations 
the results of experiments with anaesthetics seem to favour somewhat positive 
effects though they are not very decisive. 


SuMMARY. 


1. Salt solutions (SO,Cu, Cl,Ca, PO,HNa,) added to colloidal solutions 
(gum arabic, egg albumin, blood serum) decreased the osmotic pressure 
exerted by the latter. No qualitative difference was confirmed among the 
salts or among the colloidal solutions within the concentrations here employed. 
The remarkable reversibility of the osmotic pressure suggests, at once, that 
it is a physical phenomenon, i.e. increased aggregation of colloidal pe 
by the electrolytes. 

2, The effects of anaesthetics (chloroform, ether, alcohol) on the osmotic 
pressure of colloidal solutions were studied. The results, although not quite 
definite, seem to show that there is some effect. 


I wish to express my great indebtedness for Prof. B. Moore’s constant 
direction in this work. 


REFERENCES, 


Adamson and Roaf (1908). Biochem. J. 8, 422. 

Biltz and Vegesack (1909). Zeitsch. physikal. Chem. 68, 357. 
Lillie, R. 8. (1907). Amer. Ji Physiol. 20, 121. 

Loeb, J. (1918). J. Biol. Chem. 35, 497. 

Mayer, Schaeffer and Terroine (1907). Compt. rend. 145, 91° 
Moore and Parker (1902). Amer, J. Physiol. 7, 261. 

Moore and Roaf (1907). Biochem. J, 2, 34. 
Roaf and Alderson (1907), Biochem. J. 2, 412. 

Starling (1896). J. Physiol. 19, 312. 


XXXV. FURTHER OBSERVATIONS ON THE 
NATURE OF THE REDUCING SUBSTANCE 
IN HUMAN BLOOD. 


By EVELYN ASHLEY COOPER anp HILDA WALKER. 
From the University of Birmingham. 


(Received April 3rd, 1922.) 


In the previous communication [1921] it was shown that the reducing power 
of human blood was sometimes increased by acid hydrolysis. This pointed 
to the presence in addition to glucose of a more complex substance, which 
was shown not to be glycogen, and thus appeared to be a disaccharide. The 
quantitative determination of this accessory substance was found to be 
complicated by two factors: 

(1) the destructive effect of the HCl used for hydrolysis upon the reducing 

substance; : 

(2) the inhibitory effect of NaCl (formed by neutralisation after hydrolysis) 

upon the reduction of the copper carbonate by sugar. 
Further work on the subject has since been carried out, and the results are 
recorded in the present paper. | 

First of all, it was necessary to select an acid, which did not itself destroy 
sugar, and the salts of which had no effect on the reducing process. 

In the previous work sodium sulphate had been found to have no in- 
hibitory effect, but the use of sulphuric acid for hydrolysis is inadmissible, 
because sodium sulphate is used for removing the proteins from the blood, 
and its presence would diminish the ionisation of the acid. 

Further experiments showed that potassium chloride, like sodium chloride, 
had a retarding action on the reducing process, while the sulphate had no 
effect. The use of other chlorides for comparative purposes was not possible 
for various reasons. For example, ammonium chloride is decomposed by the 
alkali in the copper solution, and the liberated ammonia destroys the glucose, 
while the salts of several metals, e.g. Ca, Ba, Mg, give precipitates with the 
copper carbonate. The results, however, show that the chloride ion has a 
specific retarding effect. 

Experiments with varying concentrations of sodium and potassium 
chloride showed that the retarding action first appeared when the concen- 
tration was approximately N/1. This inhibitory effect was found to be 
irreversible, i.e. when a solution of glucose containing NaCl was diluted with 


456 E. A. COOPER AND H. WALKER 


a salt-free sugar solution, the reducing action of the sugar upon the copper 
solution was still (proportionately) retarded. The reduction was also retarded 
when the salt was added to the boiling mixture of sugar and copper solutions. 
This suggests that the compounds of glucose with the chlorides are much 
more stable than is commonly supposed. aghast ca and iodides of the alkali 
metals also inhibited the reducing process. 

Sodium citrate was moreover also found to exert an inhibitory action 
when present in concentration equivalent to 2 % citric acid, which is em- 
ployed for hydrolysis in sugar analysis. 

Di-sodium phosphate, on the other hand, had no inhibitory effect, and it 
thus seemed that phosphoric acid might be a suitable acid for hydrolysing 
the blood-sugar extracts. Separate portions of an extract (prepared by 
MacLean’s method, as in the previous communication) were hydrolysed with 
‘N/10 HCl and N/10 H,PO,, and it was found that the reducing substance 
was destroyed to the extent of about 10 % by both acids. 

Pure glucose was next dissolved in sodium sulphate solution, or in blood- 
extract, and the mixture hydrolysed with N/10 or N/100 HCl. The results 
obtained were similar to those obtained with blood-extract itself [Cooper and 
Walker, 1921]; sometimes there was no destruction of reducing substance at 
all, but occasionally a destruction occurred amounting to about 10 %. 

These results afford an explanation of our previous work. Evidently the 
hydrolysable substance, which may be a disaccharide, is only occasionally 
present in blood, so that on hydrolysis there is often no increase in the re- 
ducing power of the blood-extract, or there may actually be a loss in reducing 
power, owing to destruction of some of the glucose. In fact, out of 18 samples 
of normal human blood examined, only six cases showed an increased reducing 
power after acid hydrolysis, and as a general rule the chief reducing substance 
present is glucose. 

Two blood-extracts were subjected to dialysis. In one case the hydrolysable 
substance dialysed completely. In the other case, however, it was only par- 
tially dialysable, and this suggests that occasionally substances of intermediate 
complexity may be present in blood, possibly related to malto-dextrins. 


THE ESTIMATION OF BLOOD-SUGAR. 


Up to the present time nearly all the methods for estimating sugar in 
physiological work are modifications of Fehling’s method. Although fairly 
reliable, this is an empirical method, and the interaction of sugar with alkali 
copper solution is of great complexity. Recently, the iodometric method for 
estimating sugars has come to the forefront [see Baker and Hulton, 1920). 
This is a simple method, consisting in the quantitative oxidation of aldoses 
to the corresponding carboxylic acid by means of iodine in alkaline solution, 
and it can be carried out in a few minutes at room temperatures, We have 
therefore attempted to adapt it to physiological work. 


THE REDUCING SUBSTANCE IN HUMAN BLOOD 457 


We have found, however, that in estimating pure glucose in concentration 
approximating to the average content of the blood-sugar extracts by the 
iodine method, slightly low results are obtained. Varying the proportion of 
alkali, and extending the period of reaction did not affect the results. In 
estimating the sugar in blood by this process, however, the results were twice 
as high as those obtained by MacLean’s method. The alkali copper reagent 
is thus more selective than iodine in its action, and the iodometric method 
does not appear to be suitable for physiological work. 


INFLUENCE OF FATIGUE ON THE BLOOD-SUGAR CONCENTRATION. 


Estimations of the blood-sugar by MacLean’s method were carried out on 
normal persons immediately before, and directly after, exercise. The persons 
examined were students, both men and women, of ages ranging from about 
18 years to 26. Estimations were made on one man before and after half-an- 
hour of strenuous boxing, the other estimations on men were made before 
and after Rugby football, and on the women, before and after hockey. Of 
twelve experiments, there was a rise in the blood-sugar concentration in nine 
cases. The increase, as shown in the appended table, varied from 10 to 50 %, 
but in one case the amount was nearly trebled after 1} hours’ play. 


% before % after 
Boxing. 4 hour. A 0-04 0-05 
Rugby. hr. 10 mins. B (1) 0-052 0-077 
(2) 0-072 0-130 
C 0-097 0-090 
D 0-093 0-10 
E 0-115 0-10 
F 0-087 0-094 
G 0-118 0-094 
Hockey. 1 hr. 15 mins. H 0-087 0-253 
I 0-065 0-075 
J 0-118 0-150 
K 0-083 0-092 


The player B is a first class forward, and in good training; he is of a very 
nervous and imaginative temperament, and before the match is always in 
a state of suppressed excitement. C and D are also fast players; C is of a 
quiet type, not easily roused to excitement. H, F and G were not in dee 
and did not take the game so seriously as the others. 

Of the hockey players, H was playing back; she is of a highly nervous 
temperament, but well controlled. Z and K were half-backs, and J played 
forward. 

It will be seen that the players most easily roused to excitement over the 
game (B and H) are those whose blood-sugar has increased the most, although 
H, playing back, probably underwent less physical fatigue than most of the 
others. Psychic influences, therefore, probably play a considerable part. 


458 E. A. COOPER AND H. WALKER 


CoNDITION OF THE SUGAR IN THE BLOOD. 


We next considered the question of the structural condition of the sugar 
in the blood, and its bearing on physiological and pathological problems. It 
is well known that sugars do not merely exist as aldoses and ketoses, but may 
also pass into cyclic forms known as oxides. The ethylene oxide form: 


H.OH 
oC 
CH 


id (OH)], ° 
H,OH 


is characterised by being extremely chemically reactive, and it decolorises 
permanganate rapidly. 

Hewitt and Pryde [1920] showed that this active form of sugar could be 
actually formed by contact of an aqueous solution of glucose with the in- 
testinal wall in vivo. 

It is possible that in the animal organism sugar is normally metabolised 
in this active condition, and that an enzyme exists for transforming ordinary 
sugar as ingested into this form. Now it is known that fructose is more 
readily converted into the ethylene oxide structure than glucose, and that in 
diabetes the organism may still be able to metabolise fructose, although it 
has lost the power to deal with glucose. This suggests that the causation of 
diabetes is associated with some disturbance in the enzyme mechanism, which 
normally converts inactive sugar into the reactive ethylene oxide form. 

' We therefore proceeded to ascertain whether normal human blood can 
cause ordinary sugar to pass into this reactive state. A few cc, of fresh blood 
were placed in a small dialyser, immersed in }—1 % solutions of glucose and 
fructose. At varying intervals of time, samples of the dialysate were with- 
drawn and examined, either polarimetrically or with a solution of perman- 
ganate. No evidence, however, was obtained that blood, under the above 
conditions, can produce the ethylene oxide form from ordinary glucose or 
fructose. 

Since these experiments were carried out, Hewitt and Souza [1922] have 
found that even im vivo normal blood is unable to induce this structural 
change. 


SuMMARY. 


1. Chlorides, bromides, iodides and citrates inhibit the reduction of copper 
carbonate by glucose, as carried out by MacLean’s method, Sulphates and 
phosphates have no effect. 

2. Glucose is slightly destroyed by boiling with V/10 HCl and V/10 H,PO,. 
The reducing substance present in blood is also destroyed by acid to about 
the same extent. This affects the determination of the hydrolysable reducing 
substances in blood. 


THE REDUCING SUBSTANCE IN HUMAN BLOOD 459 


3. Glucose is the chief sugar occurring in human blood, and reducing 
substances of a more complex nature are only occasionally present. 


4. The estimation of blood-sugar by the iodine method gives resulta much 
higher than those obtained by MacLean’s process, and the iodine method does 
not seem suitable for physiological work. 


5. The blood-sugar concentration may rise considerably as the result of 
muscular exertion. 


6. There is no evidence that human blood can transform ordinary glucose 
or fructose into the reactive ethylene oxide form. 


7. A theory as to the causation of diabetes is put forward. 


REFERENCES. 


Baker and Hulton (1920). Biochem. J. 14, 756. 
Cooper and Walker (1921). Biochem. J. 15, 415. 
Hewitt and Pryde (1920). Biochem. J. 14, 395. 
Hewitt and Souza (1922). Biochem. J. 15, 667. 


XXXVI. BLOOD ENZYMES. II. THE INFLUENCE 
OF TEMPERATURE ON THE ACTION OF 
THE MALTASE OF DOG’S SERUM. 


By ARTHUR COMPTON. 
From Laboratoire de Chimie Biologique, Institut Pasteur, Paris. 


(Received April 4th, 1922.) 


In a recent communication [Compton, 1921, 2], we have called attention to 
the fact that it is possible to divide mammals into two groups, depending 
upon whether the ‘enzyme maltase is present, or not, in their blood-stream. 
For convenience, these groups have been named the “dog group” and the 
“rabbit group.” The former group is represented in our experiments by such 
mammals as the dog, the pig, the ox, the horse, the goat, the sheep and the 
rat; while the latter group is represented by the rabbit, the guinea-pig, the 
cat and man. 

In view of the general interest of the question, and the fact that it offers 
a starting-point for various researches, we have set out in the present in- 
vestigation to determine, under certain well-defined conditions, the effect of 
heat on the enzyme as met with in the blood of the dog—an animal typical 
of the maltase group. 

EXPERIMENTAL. 


For this investigation three dogs were utilised (2-P, 3-P, 4-P). Blood was 
withdrawn aseptically by venous puncture from the jugular, the animals 
having been in general in the fasting state for 24 hours previously’. Dog 2-P 
was examined three times, an interval of three weeks elapsing between the 
examinations (1) and (2), and one week between the examinations (2) and (3); 
dog 3-P twice, 18 days elapsing between the examinations; while blood was 
only once examined from dog 4-P. 

As soon as withdrawn, the blood was transferred to a sterile centrifuge 
tube, centrifuged, and the supernatant serum pipetted off from the clot. The 
serum thus collected was conserved in presence of a few drops of toluene in 
a closed test-tube, and utilised as required in the studies which follow. 


I, Activity in maltase of various specimens of dog’s serum. 
The activity in maltase of the different specimens of serum was determined 
at the temperature of 47°, in an action of 16 hours’ duration and concentration 


+ I desire to express here my indebtedness and thanks to Monsieur A. Frouin, for kindly 
supplying me with specimens of dog’s blood for these experiments. 


THE MALTASE OF DOG’S SERUM 461 


in maltose M/20: that temperature having previously been found under like 
circumstances to be the optimum temperature of a specimen of vegetable 
maltase prepared from Aspergillus oryzae [Compton, 1914]. 

For the determination, 90 mg. of pure hydrated maltose were weighed out 
into each of four clean test-tubes, to which were added 2 cc. of pure water 
(redistilled) to make a solution; then doses of the serum under investigation 
were added to the tubes as follows: 0-1, 0-2, 0-4, 0-8 ec. The contents of the 
tubes were then completed rapidly to 5 cc. with a further addition of water 
respectively as follows: 2-9, 2-8, 2-6 and 2-2 cc. Next, three drops of toluene 
were added to each tube, the latter shaken, closed with clean corks, and 
plunged into a water-thermostat regulated at or about 47°. After 16 hours’ 
incubation, the tubes were withdrawn, the corks removed and the tubes 
heated for seven minutes in boiling water to stop the enzyme action. When 
cold, the contents of each tube were diluted to 50 cc., and 20 ce. of the diluted 
mixture employed to determine, by Bertrand’s method, the proportion of 
maltose hydrolysed. The numbers obtained are set out in Table F. 


Table I. 


Dose of serum Temperature of Percentage Information as to the 
Lab. No. of employed experiment of maltose age of serum at the 
the animal ce *C. hydrolysed time of testing 


2-P (1) 47-0-46-8 25-4 Fresh 


> 
S $ 


2-P (2) 47-4-46°8 34:3 Two days 


2-P (3) 47:5-46-0 31-2 Fresh 
3-P (1) 47-0-46-9 18-3 Two days 
3-P (2) 46-8 26-8 Fresh 


4-P 47-2 47-9 One day 


POSSE SS2S SESE FESSP SFESP SELLE 
RK OREN DR DENK DRY DRI 
~ 
— 

«1 


i) 
wo 
S 
w 


On plotting the percentage of maltose hydrolysed against the quantities 
of serum in action, these numbers give the curves indicated in Fig. 1. 

An outstanding feature of this figure is the close relationship existing 
between the individual examinations of a given animal. Thus, we find dog 2-P 
giving a series of curves: 2-P (1), 2-P (2) and 2-P (3), crowded together in 
the same region of the figure. Indeed, the curves 2-P (2) and 2-P (3) are 
practically identical. Again, the same is true, although to a lesser degree, of 


462 A. COMPTON 


the dog 3-P: the curves 3-P (1) and 3-P (2) presenting a very similar allure, 
and being situated lowest in the diagram. The animal 4-P, on the other hand, 
represented by only a single examination, gives a curve which stands well 
apart, in a quite different region from the curves of the other dogs, being 
situated highest. 


100° 


90F 


a 
o 


S 
=) 


MALTOSE HYDROLYSED (4)——~> 
& D 
=) =) 


ia) 
° 


Oo 


02. O46. 86. 08 210 
QUANTITY OF SERUM iN CM°—> 
Fig. 1. 

The conclusion therefore, would seem justified, that while the content of 
the blood serum of the dog in the enzyme maltase may vary considerably 
from animal to animal, still for the individual it is remarkably constant, varying 
only between comparatively narrow limits. 


II. Optimum temperature of the maltase of dog’s serum. 

For reasons relating to another investigation, the quantity of serum 
utilised for the determinations of optimum temperature described in this 
paper is 0-09 ce. To employ conveniently this dose we operated as follows: 
1-35 ec. of serum were measured out into a small 15 cc, measuring flask and 
the content made up to the 15 ce. mark with pure water, After shaking, 


THE MALTASE OF DOG’S SERUM 463 


the mixture was distributed in portions of 1 cc. to a series of seven or 
eight test-tubes which had been prepared in advance, containing 90 mg. of 
maltose dissolved in 4 cc. of water and five drops of toluene, each tube being 
already placed in a water-thermostat regulated to a constant temperature. 
The object of this latter procedure was that the substrate solution in the 
tubes should already be in temperature equilibrium with the thermostat at 
the moment of adding the enzyme solution. The tubes were closed with clean 
sterile corks and incubated for 16 hours, after which the corks were removed, 
and the enzyme action stopped as previously. The proportion of maltose 
hydrolysed in each tube was determined as already described. The numbers 
obtained are set forth in Table IT. 


Table IT. 


Temperature at Maltose hydrolysed % 


beginning and 
end of each Dog 2-P Dog 3-P Dog 


experiment 
i (1) 


) 
hid 


~ 
a ~] 


. 
J 


a 
s 
| | ws 
@ -_— 
Besa 
ie 2) 
ne 
wm bo 
On 


| 
||) |8 


_ 
| © 
<I 


— 
a 
or 
bho 
a 
o 


bo 
for} 
o 
~ 
oa 
BIEL LILI 


ow 
~ 
ow 
or 


bo 

_ 

~ 
Pere 


> 
oon 
o & to 
or 
te 
or 


atte 
SILI Iii El tii 


SdH ASS 


z 
oO 
SPSL Lib Sri Sr iit 


oo 
bo 
ios) 


Deeee eee eee cee eeeee: 


SrtA ett 


eee SSS 


> 
& 
n 


qaenhe 

SLE It SEibSee 
A 
S11 S| | | 


SESE 
me 
oan 


: 


S2esy 
e2eee 
Sane 

Silt 
1181118 

SI l dt 

PILI I Str iid 
Papeeeee 

DeeRanee 


37-5 


On plotting the percentage of maltose hydrolysed against the mean tem- 
perature of the experiment, these numbers give the series of optimum 
temperature curves indicated in Fig. 2. 

Fig. 2 reveals for the three animals studied the same optimum temperature: 
55°, in an action of 16 hours’ duration and reaction of medium that of the 


464 A. COMPTON 


serum diluted in pure water. It is of interest that this same optimum tem- 
perature is found, whether successive studies are made at different times of 
the same animal, or whether different animals are studied. 


70- 


r°7) 
=] 
’ 


: 


MALTOSE HYDROLYSED (%)——> 


rm 
i=) 


10} TREES 
Oe eh ake 
30.40 250 60 eae 
TEMPERATURE ——> 
Fig. 2. 
CONCLUSION. 


The enzyme maltase, met with in the blood of the dog, exhibits from 
animal to animal a constant optimum temperature. This stability of optimum 
temperature constitutes an experimental fact which cannot but prove to be 
of considerable practical value as a reference point, in connection with any 
subsequent study of the enzyme. It is in contrast with the variability which 
characterises the actual amount of enzyme present at any time in the blood 
of different animals. : 

From what we know of the remarkable sensitiveness of the optimum 
temperature of an enzyme to variations of py of the medium [Compton, 1915; 
1921, 1], there can be little doubt that the theoretical interpretation of the 
above effect lies in a certain constancy of the chemical reaction, or hydrogen-ion 
concentration, of the animal’s blood serum. The same delicate mechanism which 
maintains constancy of the one, will obviously maintain constancy of the other. 

From the point of view of general physiology, this stability of optimum 
temperature of a blood enzyme raises problems of even wider interest. For 
instance, one may ask how far it may be possible to define a species by 
humoral reactions of enzymic nature, That is a question to which we hope 
to return in the course of subsequent studies in this series. 


REFERENCES. 


Compton (1914). Proc. Roy. Soc, B, 88, 258. 
(1915). Proe. Roy. Soc. B, 88, 408. 
——— (1921, 1). Proc. Roy. Soc. B, 92, 1. 
—— (1921, 2). Biochem J. 15, 681. 


XXXVII. THE MODE OF OXIDATION OF FATTY 
ACIDS WITH BRANCHED CHAINS. II. THE FATE 
IN THE BODY OF HYDRATROPIC, TROPIC, ATRO- 
LACTIC AND ATROPIC ACIDS TOGETHER WITH 
PHENYLACETALDEHYDE. 


By HERBERT DAVENPORT KAY anp HENRY STANLEY RAPER. 


From the Department of Physiology and Biochemistry, the 
University of Leeds. 


(Received April 11th, 1922.) 


In a previous communication, Raper [1914] advanced an hypothesis to explain 
the mode of oxidation in the body of fatty acids substituted at the a-carbon 
atom by a methyl group. The experimental evidence available at that time 
indicated that these acids underwent oxidation in the same way as the straight 
chain fatty acids from which they were derived. It was suggested that this 
behaviour might be accounted for if oxidation of the a-methyl group took 
place first, giving rise to a derivative of the semi-aldehyde of malonic acid, 
which being very unstable would lose carbon dioxide to produce a normal 
fatty aldehyde. This aldehyde, in turn, would give rise to its corresponding 
fatty acid by further oxidation. The series of reactions would thus be as 
follows: 
CH, CHO 
p-CH_COOH—-+2-CH—COOH—->R—-CH, CHO +CO,—--R—CH,—COOH. 


In this way an a-methylated acid would give rise to a fatty acid with a straight 
chain structure but containing one carbon atom less. In support of this hypo- 
thesis it was shown that isobutyric acid and a-methylbutyric acid, on oxidation 
with hydrogen peroxide, yielded propaldehyde and normal butaldehyde 
respectively. 

Owing to the relative scarcity of data concerning the fate of branched 
chain fatty 4cids in the animal organism it was decided to investigate the fate 
of phenyl derivatives of the a-methylated fatty acids. The present com- 
munication is concerned with the simplest member of this series, namely, 
hydratropic acid (a-phenylpropionic acid), and some of its derivatives. Ex- 
periments with other members of the a-substituted series are in progress and 
will be communicated in due course. The results have been disappointing so 
far as the isolation of intermediate products of oxidation is concerned but the 
fate of the various acids investigated, (I) to (IV), has enabled conclusions to 


Bioch. xvi 31 


466 H. D. KAY AND H. 8. RAPER 


be drawn which indicate the probable manner in which hydratropic acid is 
oxidised. 


CH, CH,OH 
(I) o,H,—H_co0H, (II) cH, HCOOH, 
CH, CH, 
(III) 0H, (0H)—COOH, (IV) 0,H,C_COOH. 


Hydratropic acid (I) administered to dogs produced slight toxic effects in 
doses of 0-25 g. per kilo and was oxidised to the extent of about two-thirds. 
The remaining third was excreted in the urine, partly combined with glycine, 
and was dextrorotatory. A resolution of hydratropic acid, not yet complete, 
has shown that this was not the pure dextro form but a mixture of the dextro 
and laevo modifications in which the dextro component was in excess. No 
intermediate products of oxidation were detected. Inactive tropic acid (II) 
was found to be very resistant to oxidation, over 90 % of the acid given 
being recovered from the urine. The acid excreted, as expected, was optically 
inactive. Inactive atrolactic acid (III) was also very resistant to oxidation 
and over 80 % of this acid was excreted unchanged. It also was optically 
inactive. Atropic acid (IV), on the other hand, was found to be the most 
toxic of the four acids investigated, but in doses of 0-13 g. per kilo, which 
were tolerated, it was completely oxidised. No intermediate products were 
detected in the urine but a small but definite trace of succinic acid was 
excreted on two occasions and was not found on careful examination of the 
normal urine of the same dog. It is difficult however to connect this directly 
with the oxidation of atropic acid. 

It is of great general interest to find that tropic acid is much more 
resistant to oxidation in the body than hydratropic acid, a result which the 
greater in vitro oxidisability of the hydroxylated acids (when acted on by the 
usual oxidising agents) would hardly lead one to expect. Dakin [1909] has 
observed the same phenomenon with f-phenylpropionic acid and f-phenyl- 
B-hydroxypropionic acid. The former is much more easily oxidised than the 
latter. These results show that the introduction of a hydroxyl group in the 
f-position in phenyl derivatives of propionic acid does not facilitate their 
oxidation in the body but rather hinders it. The fact that tropic and atro- 
lactic acids are much less easily oxidised than hydratropic acid excludes them 
as possible intermediate products in its oxidation. It is also very improbable 
that hydratropic acid undergoes a-oxidation since either atrolactic acid, 
phenylglyoxylic acid, mandelic acid or benzoic acid would be expected as 
intermediate products. A careful search failed to reveal any of these, so 
that it may be assumed that hydratropic acid is oxidised in the f-position. 
Further, since the methyl group containing the B-carbon atom is unsubstituted 
both in atrolactic and hydratropic acids, it would be expected, that the former 
would be oxidised to a greater extent than was found to be the case. This 
makes it probable that some other factor besides the non-substitution of the 


OXIDATION OF BRANCHED FATTY ACIDS 467 


methyl group is important in facilitating the oxidation of hydratropic acid. 
We believe this factor to be the ability to form an unsaturated linkage 
between the a- and f-carbon atoms by the loss of two atoms of hydrogen. 
This is not possible in atrolactic acid but it is in hydratropic acid and would 
result in the formation of atropic acid. Atropic acid is not resistant to oxida- 
tion and although it may be urged that it is more toxic than hydratropic 
acid and therefore is not a likely intermediate product in the oxidation of 
the latter, it has to be borne in mind that the toxicity of a substance when 
produced slowly by metabolic processes in the tissues may be much less 
marked than when a comparatively large dose is introduced quickly by sub- 
cutaneous injection. Hydratropic acid, also, is distinctly more toxic than 
atrolactic and tropic acids. In the first case, doses of 0-25 g. per kilo give 
toxic effects whereas | g. per kilo of either of the other two acids gives no 
toxic symptoms. For these reasons therefore we are led to believe that 
atropic acid is the most likely primary oxidation product of hydratropic acid. 
It follows from this that the change from hydratropic acid to atropic acid is 
brought about by the direct oxidative removal of two hydrogen atoms and 
not by the loss of water from either-tropic or atrolactic acids, since these are 
apparently not capable of conversion to atropic acid in the body. If they 
were they would be oxidised to a greater extent than was found to be the case. 
It is also possible that atropic acid may be produced from hydratropic acid 
by a process of dehydrogenation as pictured by Wieland [1912]. 

As to the further stages in the oxidation of hydratropic and atropic acids 
we can only speculate. From analogy with cinnamic acid B-oxidation might 
be expected with the formation of formylphenylacetic acid (V): 

i CH (OH) CHO 


C,H,—C—COOH——>C,H,—C—COOH—-> (V) C,H,—CH—COOH. 


This would be unstable and lose carbon dioxide to give phenylacetaldehyde, 
or by further oxidation it might give phenylmalonic acid. The fate of the 
latter substance in the body is not known, but Dakin [1909] has described 
experiments in which phenylacetaldehyde in dilute alcoholic solution was 
given subcutaneously to dogs. About 84 % was completely oxidised and the 
remaining 16 °% converted into phenylacetic acid which was excreted in the 
urine as phenaceturic acid. We have repeated and confirmed Dakin’s experi- 
ment and found that by oral administration also a considerable part of the 
phenylacetaldehyde administered undergoes complete oxidation. Since neither 
hydratropic nor atropic acid gave rise to the excretion of phenylacetic acid 
it is not certain that phenylacetaldehyde represents an intermediate stage in 
their oxidation, although it would be expected to be such if the original 
hypothesis put forward previously [ Raper, 1914] were correct. It is however 
not improbable that phenylacetaldehyde produced in vivo may undergo com- 
plete oxidation in spite of the fact that when introduced orally or subcutan- 
eously it is partially oxidised to phenylacetic acid. 


31—2 


468 H. D. KAY AND H. 8S. RAPER 


In a recent communication, Hanke and Koessler [1922] have called atten- 
tion to the phenomena of the oxidation of fatty acids in the body in relation 
to their electronic structure. The arguments put forward in the case of 
butyric acid, for which the following electronic structure is developed, will 
serve as an example: 

“Gs +C+-C- ries 
- - - +4 
ecg ek. 
In this formula the a- and y-carbon atoms are represented as quadruply 
negative and the carboxylic carbon atom as quadruply positive, whereas the 
B-carbon atom is doubly positive. Evidence is brought forward by Hanke 
and Koessler which indicates that quadruply positive or negative carbon, and 
especially the former, is resistant to oxidation. This being so, it would be 
expected that butyric acid when oxidised would be attacked most readily at 
the B-carbon atom which is only doubly positive. It is therefore suggested 
that in this way the predominance of f-oxidation of fatty acids in the body 
may be explained. On our part, we feel that there is, as yet, insufficient 
knowledge, either qualitative or quantitative, and especially the latter, of 
the mechanism of oxidation processes on which an explanation in relation to 
their electronic structure can be based. The idea that B-oxidation is pre- 
dominant is supported essentially by qualitative evidence for there is little 
of a quantitative nature. There are no animal experiments on record which 
enable one to postulate that butyric acid is oxidised only or chiefly in the 
B-position. Further, Cahen and Hurtley [1917] have shown that when butyric 
acid is oxidised with hydrogen peroxide, a process which simulates to some 
extent oxidation in the body, the main product is succinic acid and not 
acetone, so that oxidation in this case is most marked at the y-carbon atom. 
There is also evidence which indicates that in certain cases, B-oxidation, when 
it occurs, may be preceded by “desaturation” between the a@- and B-carbon 
atoms. The oxidation of hydratropie acid to produce atropic acid, as sug- 
gested above, is a case in point, but others are: the production of cinnamic 
acid from phenylpropionic acid [Dakin, 1909] and the formation of furfuryl- 
acrylic acid from furfurylpropionic acid [Sasaki, 1910; Friedmann, 1911). 
Again, the electronic formulae of B-phenyl-8-hydroxypropionic acid and f- 
phenyl-propionic acid, developed by Hanke and Koessler, show that in both 
compounds the B-carbon atom is doubly positive and the a- and carboxylic- 
carbon atoms quadruply negative and quadruply positive respectively. If 
the ease of oxidation in the body of a particular carbon atom of a substance 
is sufficiently explained by the electronic formula then there would be no 
reason to expect that these two substances should not undergo oxidation 
equally readily. As a matter of fact the non-hydroxylated acid is much more 
readily oxidised. The experiments with hydratropic and tropic acids, de- 
scribed in the present communication, show equally marked differences and 
do not seem to be capable of adequate explanation on an electronic basis 


OXIDATION OF BRANCHED FATTY ACIDS 469 


alone. It would seem that until we have a more definite knowledge of the 
mechanism of the oxidation process, its explanation, based solely on views 
of polarity in the substance oxidised, or on its electronic structure, cannot be 
entirely successful. These views, developed largely by a study of the phenomena 
of addition and substitution in organic compounds, have given a satisfactory 
general explanation of previously inexplicable phenomena; but with the 
meagre quantitative data at our disposal and our lack of knowledge as to 
how oxidation of very simple substances takes place either inside or outside 
the body, the polarity alone of the atoms in the substance oxidised is in- 
sufficient as a basis of explanation. As an instance of the probable com- 
plexity of an apparently simple process such as the oxidation of carbon 
monoxide to carbon dioxide when the former is burnt in moist air, the work 
of von Wartenberg and Sieg [1920] may be referred to. It is suggested that 
in this oxidation four intermediate reactions are concerned. Firstly the pro- 
duction of formic acid by the addition of water, secondly its dehydrogenation 
to produce carbon dioxide and hydrogen, thirdly, oxidation of the hydrogen 
to produce hydrogen peroxide and lastly decomposition of the peroxide with 
the formation of water and oxygen. If the oxidation of carbon monoxide is 
as complicated as this scheme represents, then it is at least probable that 
the processes of oxidation of fatty acids are not less complex. 


EXPERIMENTAL. 


Methods. Unless stated otherwise, the acids used in this investigation 
were administered hypodermically as sodium salts in aqueous solution. The 
urine was collected from the time of the first dose until 36 hours after the 
last dose was given. Urines were preserved on ice. Volatile products were 
sought for by distilling all, or part of the urine in steam and examining the 
distillate. The urine was then concentrated under reduced pressure to about 
one-fifth of its original volume and acidified strongly with syrupy phosphoric 
acid. Subsequently it was extracted with ether in a continuous extractor for 
30 to 48 hours. The ether extract was then examined for unchanged acid or 
products of oxidation. 

Hydratropic acid. The acid was obtained in two ways. Firstly, by the 
reduction of atropic acid with sodium amalgam and secondly, by the oxidation 
of hydratropic aldehyde with silver oxide in the presence of a slight excess 
of sodium hydroxide. The hydratropic aldehyde was prepared by the method 
of Tiffeneau [1907]. The former method was found to be the more expeditious. 
The acid was purified by distillation under reduced pressure and had B.P. 
148-9° (14 mm.). In doses of 0-25 g. per kilo a slight lassitude was developed 
but no other ill effects were noted. Dogs, only, were used in this experiment 
and larger doses were not tried. No volatile products of oxidation were de- 
tected when the urine was distilled with steam. The ether extract from the 
acidified urine was obtained as a light brown syrup which refused to crystallise. 
Extraction with warm light petroleum removed some unchanged hydratropie 


470 H. D. KAY AND H. 8. RAPER 


acid and traces of fatty acids. The latter were removed as insoluble barium 
salts and the hydratropic acid recovered was weighed. In this way, in two 
experiments, 3 and 22 % of the acid originally given was recovered. Its 
equivalent was 151 (calculated 150), and it was dextrorotatory. In absolute 
alcohol (concentration 3 %), it had [a] + 21-2° to + 26-9°. The ether extract, 
after removal of the hydratropic acid as just outlined, was treated in many 
ways to induce it to crystallise but all were fruitless. It was dextrorotatory, 
reduced Fehling’s solution on boiling and gave the orcinol reaction, indicating 
the presence of a glycuronic acid derivative. The original urine also gave 
these reactions. In addition, the ether extract contained some nitrogenous 
substance. Having failed to obtain any crystalline compound from this ex- 
tract, and suspecting the presence of a glycine derivative because of the marked 
nitrogen content, it was hydrolysed by boiling with concentrated hydrochloric 
acid for an hour. After hydrolysis, the solution was diluted and extracted 
with light petroleum and then with ether. The latter yielded only a trace of 
oily substance which was neglected, but the former was found to contain 
dextrorotatory hydratropic acid (equivalent 150-8). In 3 % solution in abso- 
lute alcohol it gave [a]? + 27-2°. The hydrolysate after ether extraction was 
evaporated to small bulk under reduced pressure and esterified with absolute 
alcohol saturated with hydrochloric acid. It was seeded with a crystal of 
glycine ester hydrochloride and placed on ice for two days, whereby a mass 
of crystals of glycine ester hydrochloride was obtained. The identity of the 
glycine was confirmed by benzoylation, when hippuric acid, M.P. 186—7°, was 
obtained. Careful search failed to reveal any other definite products of hydro- 
lysis, phenylacetic, tropic and atropic acids being specially sought for. The 
chief substance present in the ether extract of the urine thus appeared to be 
hydratropylglycine. A sample of this substance was made by the Schotten- 
Baumann method from hydratropyl chloride and glycine and was found to 
be more difficult to crystallise than the commoner acyl derivatives of glycine 
such as phenaceturic and hippuric acids, and its contamination with some 
oily impurity, possibly a glycuronate of hydratropic acid, with about the 
same solubilities, explains its non-crystallisation from the ether extract. In 
a later experiment it was found possible to obtain crystalline hydratropyl- 
glycine from this extract with considerable loss, by the following method. 
The acids extracted from the urine by ether, after removal of unchanged 
hydratropic acid, were converted in turn into sodium salts and ferric salts, 
The acids obtained from the latter by decomposition with dilute sulphuric 
acid and extraction with ether were mixed with half their equivalent of 
sodium hydroxide and the free acids present in the solution shaken out 
with ether. The remaining sodium salts were now decomposed with dilute 
sulphuric acid and the liberated acid extracted with ether. This extract 
crystallised readily on inoculation with hydratropyl glycine and after two 
recrystallisations from water had m.p. 103°, which agreed with that of the 
synthetic product made from the dextrorotatory hydratropic acid recovered 


OXIDATION OF BRANCHED FATTY ACIDS 471 


from the urine, and glycine. The pure inactive glycine derivative has m.P. 105°. 
In one experiment carried out as quantitatively as possible for the estimation 
of the hydratropic acid excreted, either free or combined, three doses, each 
of 1-5 g., of the acid were given on three successive days. It was found that 
0-156 g. (= 3-5 %) of the acid administered was present in the free state and 
1-225 g. (= 27%) in combination with glycine, or possibly partly with 
glycuronic acid. In order to determine whether any derivative was present 
in the urine as a glycuronate the concentrated urine after the usual ether 
extraction was freed from phosphoric acid by baryta and the reducing sub- 
stance precipitated by basic lead acetate. This precipitate was decomposed 
with hydrogen sulphide and to the resulting solution, after removal of 
hydrogen sulphide in a current of air, dilute sulphuric acid was added to make 
the concentration about 5 %. It was boiled for eight hours and the resulting 
solution, after clarification with charcoal, extracted continuously with ether 
for eight hours. The ether extract contained only a trace of resinous substance. 
The reducing substance was probably, therefore, free glycuronic acid which 
had originally been combined with hydratropic acid and which had been split 
off during the ether extraction of the concentrated and strongly acidified 
urine. It is well known that many glycuronates are easily hydrolysed under 
such conditions. 

Tropic acid. Inactive tropic acid was used in these experiments and its 
fate in both cats and dogs was determined. Doses of | g. per kilo caused no 
noticeable ill effects. The urine on steam distillation yielded no volatile oxida- 
tion products. The ether extract of the concentrated urine, on removal of 
the ether, gave a brown mass which crystallised at once on cooling and after 
one recrystallisation gave pure tropic acid, m.p. 118-119°. The acid was 
optically inactive. In two preliminary experiments, one on a cat and the 
other on a dog, the urine was extracted with ether for only 12 hours. In the 
first, 2-8 g. of the acid were given and 2-25 g. (80 %) of the pure acid were 
recovered from the urine. In the second, 5-9 g. of the acid yielded 4-4 g. 
(74 %) in the urine. A third experiment was carried out and the extraction 
with ether continued for 30 hours, when all the acid had been extracted. 
Of 2-94 g. of the acid administered, 2-77 g. were recovered, i.e. 94 %. Ex- 
amination of the mother liquors after the separation of the tropic acid failed 
to reveal the presence of any oxidation products. 

Atrolactic acid. The inactive acid was prepared from acetophenone by 
Spiegel’s method using the modification described by McKenzie and Clough 
[1912]. Both dogs and cats were used for the experiments and doses of 0-9 g. 
per kilo were well tolerated, no ill effects being noted. No volatile oxidation 
products were detected in the steam distillate. The residue from the ether 
extract crystallised on cooling and was extracted with hot benzene to remove 
unchanged atrolactic acid. Most of the residue was soluble in benzene but a 
small amount of brown tarry substance remained after the extraction. This 
was dissolved in water and re-extracted with ether. The ether residue was 


472 H. D. KAY AND H. S. RAPER 


then repeatedly extracted with moist light petroleum whereby a further 
quantity of atrolactic acid was obtained. A small amount of brown oily 
substance remained which was insoluble in cold water, benzene or light 
petroleum and would not crystallise. The only definite substance obtained 
from the urine, therefore, was atrolactic acid. It was optically inactive and 
melted at 92-3° after one recrystallisation from water. In one experiment 
on a dog, 5-8 g. of the acid were administered and 4-9 g. (84 °%) recovered from 
the urine. In another case two cats were used and of 4-5 ¢. of the acid 
administered, 3-42 g. (76 %%) were recovered from the urine. 

Atropic acid. The acid used in these experiments was prepared from tropic 
acid by boiling with 10 °% aqueous baryta for 18 hours. It was obtained quite 
pure after one crystallisation from alcohol and had m.p. 106-7°. The yield 
was 67 % of the theoretical. Doses of 0-4 g. per kilo were invariably fatal 
_ both in cats and dogs within 30 hours and no oxidation products or unchanged 
acid were found in the urine. A dose of 0-2 g. per kilo, given to a dog, pro- 
duced only slight lassitude, but when repeated after 24 hours caused death 
at the end of 60 hours. Doses of 0-13 g. per kilo were well tolerated. In one 
instance a dose of 0-18 g. per kilo, given to a dog, caused haematuria but 
the animal recovered. The urine on steam distillation yielded no detectable 
oxidation products. The residue of the ether extract was very small in amount. 
It was dissolved in hot water, clarified with charcoal, and on standing de- 
posited a small amount of a colourless, crystalline substance. In one experi- 
ment in which 1-9g. of the acid were given, 0-04 g. of this crystalline 
substance was obtained, and in another, 2:8 g. yielded 0:05 g. It was 
recrystallised from a mixture of nine parts benzene and one part alcohol and 
had m.p. 178-9°. On heating above its melting point it gave a crystalline 
sublimate. It was soluble in water and alcohol, slightly in ether and almost 
insoluble in chloroform and benzene. It gave insoluble ferric and lead salts 
but a soluble calcium salt. When recrystallised from dilute nitric acid it 
melted at 183°. The ammonium salt heated with zine dust gave the pyrrole 
reaction. These properties agree with those of succinic acid but enough 
material was not obtained for analytical identification. No other substance 
was isolated from the urine. A control experiment with a 48 hours sample 
of urine from the same dog to which atropic acid had been given yielded none 
of the crystalline acid on ether extraction. 

Phenylacetaldehyde. In a mixture of four parts alcohol to six parts water 
(by volume), 0°93 g. of phenylacetaldehyde was dissolved and given subcu- 
taneously to a cat weighing 2 kilos. The cat appeared to be affected by the 
alcohol and was drowsy for two days, but recovered. The urine was collected 
for four days and on steam distillation yielded no detectable oxidation pro- 
ducts. The ether extract was dissolved in water, distilled with steam to 
remove acids, again acidified with phosphoric acid and extracted with ether. 
The ether extract yielded a small amount of oily substance which crystallised 
after standing for two days. The mass was rubbed up with a little water and 


OXIDATION OF BRANCHED FATTY ACIDS 473 


filtered. In this way 0-26 g. of a crystalline acid was obtained. After re- 
crystallisation from water it gave a very unsharp M.P., the main part melting 
at 135-140° and the rest at 170—-180°. It appeared therefore to be phenaceturic 
acid mixed with some higher melting substance. By boiling with 20% 
sulphuric acid for a short time the phenaceturic acid was hydrolysed and the 
resulting phenylacetic acid was extracted with benzene. It weighed 0-102 g. 
and was not quite pure. The higher melting substance was extracted by 
ether from the acid liquid and proved to be hippuric acid, which is more 
resistant to hydrolysis with hot dilute acids than phenaceturic acid. In this 
experiment therefore, not more than 10 % of the phenylacetaldehyde given 
appeared as phenaceturic acid. A second experiment was carried-out on a 
dog to test the effect of oral administration, the use of alcohol as a solvent 
for subcutaneous injection being objectionable. 1-9 g. of phenylacetaldehyde 
in its own volume of alcohol were administered to a dog by the mouth in 
gelatin capsules. No abnormal symptoms were observed. The urine was 
collected for the ensuing 48 hours and was treated in the same way as that 
from the cat. A fairly pure crop of crystals of phenaceturic acid was obtained 
(m.p. 141-3°). It weighed 0-85 g. The mother liquors on hydrolysis yielded 
0-1 g. of impure phenylacetic acid and again a little hippuric acid was isolated. 
The phenaceturic and phenylacetic acids thus obtained correspond to 0-62 g. 
of phenylacetaldehyde, which represents 32 °% of that administered. Nearly 
70 % had therefore been completely oxidised. These experiments thus confirm 
those of Dakin [1909]. 


SuMMARY AND CONCLUSIONS. 


1. The experiments described in this paper were carried out to determine 
the mode of oxidation in the body of phenyl derivatives of fatty acids with a 
branched chain structure. 

2. The acids investigated were, hydratropic, tropic, atrolactic and atropic 
acids. In addition previous observations by Dakin on the fate of phenylacet- 
aldehyde were confirmed. 

3. Hydratropic acid is moderately well oxidised and undergoes a partial 
resolution in the body, the unoxidised acid found in the urine being dextro- 
rotatory. Atropic acid is also well oxidised and is more toxic than the 
other three acids. Tropic and atrolactic acids are comparatively resistant to 
oxidation. 

4. From the observations made, it is concluded that hydratropic acid on 
oxidation is converted directly into atropic acid (though this was not directly 
proved), and that this then undergoes complete oxidation. 

5. Phenylacetaldehyde largely undergoes complete oxidation in the body 
and may therefore be an intermediate product in the oxidation of hydratropic 
and tropic acids. 


474 H. D. KAY AND H. S. RAPER 


We desire to express our thanks to Messrs Boots Pure Drug Co. Ltd., for 
a gift of residues from the saponification of atropine. From these the tropic, 
atropic, and part of the hydratropic acids used in our experiments were 
obtained. 


REFERENCES. 


Cahen and Hurtley (1917). Biochem. J. 11, 164. 
Dakin (1909). J. Biol. Chem. 7, 203, 242. 

Friedmann (1911). Biochem. Zeitsch. 35, 40. 

Hanke and Koessler (1922). J. Biol. Chem. 50, 193. 
McKenzie and Clough (1912). J. Chem. Soc. 101, 393. 
Raper (1914). Biochem. J. 8, 320. 

Sasaki (1910). Biochem. Zeitsch, 25, 272. 

Tiffeneau (1907). Ann. Chim. Phys. (8) 10, 176. 

von Wartenberg and Sieg (1920). Ber. 53, 2192. 
Wieland (1912). Ber. 45, 484, 679, 2606. 


XXXVI. STRUCTURES IN ELASTIC GELS 
CAUSED BY THE FORMATION OF 
SEMIPERMEABLE MEMBRANES. 


By EMIL HATSCHEK. 


(Received April 20th, 1922.) 


WHEN a reaction is produced in a gel, the latter may undergo slight changes 
which, however, are not striking or even obvious, and have received little 
attention compared with that devoted to the characteristics of insoluble pre- 
cipitates obtained in the gel. These changes are, generally speaking, altera- 
tions in the water content of the gel, such as dehydration by contact with 
very concentrated salt solutions, e.g. the 20 or 25 % silver nitrate solution 
used in the classical Liesegang experiment; or increased hydration caused by 
the diffusion of acid or of one of the strongly lyotropic salts such as iodide 
or thiocyanate. If the experiments are carried out in test-tubes, these changes 
show themselves most readily in alterations of the gel meniscus. 

Much more complicated conditions arise, and the gel itself develops a 
complicated structure, when the reaction leads to the formation of a semi- 
permeable membrane. The most convenient and striking instance is the 
formation of copper ferrocyanide in gelatin gel, and the most suitable arrange- 
ment is as follows: clean test-tubes are filled to about half their height with 
a gelatin sol containing 10 g. of air dry gelatin in 100 ce. of a 2 % solution © 
of crystallised potassium ferrocyanide (K,Fe (CN), .3H,O). After setting com- 
pletely—which is somewhat retarded by the ferrocyanide—the gel is covered 
with a 5% solution of crystallised cupric chloride or cupric sulphate 
(CuCl,.2H,O or CuSO,.5H,O respectively), or, as will be explained below, 
with a mixed solution of cupric and sodium sulphate. The molar concentra- 
tions are accordingly M/21 K,Fe(CN), in the gel, and M/3-4 CuCl, or 
M/5 CuSO, in aqueous solution. The molar concentration of the latter is thus 
considerably greater than the molar concentration of the salt in the gel, the 
necessary condition when a reaction producing an insoluble compound is to 
proceed into the gel. 

A few minutes after putting on the copper salt solution a pale brown and 
perfectly transparent membrane of copper ferrocyanide can be noticed. The 
next change to be noticed is a gradual sinking of the surface of the gel. In 
the first stages there is no great alteration in the curvature of the meniscus, 
but it appears to travel down bodily, leaving a thin film of gel adhering to 
the glass. The constant formation or exposure of new surface entails fresh 


476 E. HATSCHEK 


formation of copper ferrocyanide, and the layer of gel on the glass as well 
as the meniscus turns dark brown and opaque. If the preparation is observed 
superficially only and at long intervals, this shifting of the meniscus and the 
formation of a dark brown zone on the glass seem to be all that happens. 
Careful observation however shows the process to be discontinuous: the 
curvature of the meniscus increases and finally the membrane composing it 
bursts somewhere near the centre, the copper salt solution penetrates through 
the crack and a fresh membrane is formed. 

What happens can be readily deduced a priori from known properties of 
the semipermeable membrane and of elastic gels. As soon as the membrane 
is formed, water flows through it from ferrocyanide solution in the gel to 
the strongly hypertonic copper salt solution above it: in other words, the gel 
“dries.” In doing so it behaves exactly as if drying proceeded in the usual 
way by evaporation into the atmosphere, 7.e. the meniscus gradually sinks 
and becomes more and more curved. The surface film now consists of very 
concentrated gel and—provided the gelatin adheres well to the glass—is in 
tension. At the same time, owing to the withdrawal of water and the low 
diffusion velocity of the potassium ferrocyanide, with a molecular weight of 
368-5, the concentration of the latter in the zone immediately below the mem- 
brane increases markedly; this can easily be seen by the colour, which is a 
clear yellow, although the rest of the gel, with the initial concentration of 2 %, 
is barely coloured. The ferrocyanide has, like the other cyanides, a marked 
softening effect on gelatin, which shows itself in concentrations of about 10 %, 
so that the layer below the tough surface film is considerably softer than the 
body of the gel. The surface film accordingly is torn off when the tension has 
increased sufficiently, and is at the same time ruptured by the atmospheric 
pressure, the copper salt solution penetrating under it. The whole process 
then repeats itself, and this continues until the copper salt solution, which 
is continually diluted, becomes approximately isotonic with the potassium 
ferrocyanide solution in the gel. 

If the gelatin does not adhere to the glass wall sufficiently, it begins to 
contract below the reaction zone and copper salt solution flows into the space 
thus formed, covering the exposed gelatin surface with a semipermeable 
membrane. This leads to further shrinkage and finally a thin cylinder of very 
tough gel only may be left. 

The analysis of the process, set forth above, was confirmed by a detailed 
examination of a large number of preparations. In the first instance the test- 
tubes were cut through about 1 em. below the bottom of the reaction zone. 
Immediately the cylinder of gel is cut across, it contracts sharply and de- 
velops a deep constriction below the last membrane. The dehydrated gel 
containing the precipitate of copper ferrocyanide no longer adheres to the 
glass, and the reaction zone can, with a little care, be removed intact from 
the tube and cut through longitudinally. Two halves of such preparations, 
made in large boiling tubes 35 mm. in diameter, are shown in Fig, 1 (Plate IV). 


SEMIPERMEABLE MEMBRANES IN ELASTIC GELS = 477 


An intact test-tube, in which the reaction has proceeded for some days, is 
illustrated in Fig. 2. The rupture in the last membrane but one is clearly 
visible, and the membrane in course of formation, which is still almost trans- 
parent, can be seen faintly. Fig. 3 shows a preparation in which the gel has 
detached itself from the glass and has become covered completely with 
copper ferrocyanide; the consequent shrinkage is very marked. 

If the reaction proceeds throughout the gel, without the latter detaching 
itself, the final aspect of the preparation is extremely remarkable. It then 
consists of a thin tube of tough gel with a large number of septa, convex 
towards the lower end of the cylinder and always ruptured somewhere near 
the centre. A cross section through a specimen of this description is shown 
in Fig. 4. It may be mentioned that this preparation was obtained by using 
a mixed solution of copper and sodium sulphate, to which reference has been 
made, instead of copper sulphate alone. The function of the copper salt is 
a twofold one: it provides material for the formation of the semipermeable 
membrane, and it maintains the osmotic pressure so that dehydration of the 
gel takes place. The latter function can of course be performed equally well 
by a salt which does not enter into the reaction, and sodium sulphate has the 
advantage of not tanning the gelatin to the same extent as copper and other 
heavy metal sulphates. It is the latter action which seems to lead, or to 
contribute largely, to the detaching of the gel from the glass, which makes it 
difficult to obtain specimens showing more than two or three septa with 
copper salt alone. Incidentally the condition of the glass surface is of con- 
siderable importance: test-tubes cleaned carefully with bichromate-sulphuric 
acid mixture ete: consistently gave bad results, while new test-tubes simply 
dusted with a dry cloth gave the best. 

If the explanation of the formation of septa, which has been given above, 
is correct, it follows that they should be formed in the course of any reaction 
which satisfies the following three conditions: (1) a semipermeable membrane 
is produced; (2) the solution above the gel is, and remains, hypertonic, and 
(3) the gel contains a lyotropic agent which, in moderately high concentration, 
reduces its modulus and tensile strength. Only one such combination—apart 
from the copper ferrocyanide reaction—readily suggests itself: gelatin gel 
containing potassium iodide as lyotropic agent, and on it a concentrated sugar 
solution containing tannin. The general results were the same as with copper 
ferrocyanide: depression of the meniscus, followed by the detachment and 
rupture of membranes. The latter were very irregular, probably because one 
set only of concentrations, chosen at random, was tried; a detailed investi- 
gation appeared unnecessary in view of the somewhat artificial nature of the 
arrangement. 

The results described show that the possibilities of periodic structures in 
gels, as the result of continuous and uncontrolled diffusion, are by no means 
exhausted by the normal Liesegang phenomenon. In the latter the gel, while 
of course influencing the character of the results very considerably, primarily 


478 E. HATSCHEK 


serves as the support for periodic deposits of insoluble reaction product, and 
does not undergo any profound changes. If a semipermeable membrane is 
formed by the reaction, the gel itself becomes segmented and the final structure 
is one which it would be extremely difficult to explain a posteriori—quite as 
difficult as it was to explain other periodic structures before the study of the 
Liesegang rings provided a clue. While the author does not know any natural 
structure resembling those described and while he is, generally speaking, by 
no means inclined to overrate the cogency of the conclusions drawn from 
artificial imitations of natural forms, he ventures to think that the highly 
complicated consequences which follow the formation of a semipermeable 
membrane in a suitable elastic gel may have some biological or histological 
interest. 


BIOCHEMICAL JOURNAL, VOL. XVI, No. 4 PLATE IV R 
47 ¢ 


Fig. 2 Fig. 3 


XXXIX. A MODIFICATION OF BASAL DIET FOR 
RAT FEEDING EXPERIMENTS. 


By MURIEL BOND. 


From the Physvologwcal Laboratory, London (Royal Free Hospital) 
School of Medicine for Women. 


(Received April 21st, 1922.) 


CASEINOGEN, the usual protein constituent of a basal diet for rat feeding 
experiments, possesses several disadvantages. Owing to its natural association 
with milk fat it is difficult to render it free of fat-soluble A—if the purification 
be done by alcohol and ether extractions the cost of the material is greatly 
increased, and if the caseinogen is treated by the Drummond method [Coward 
and Drummond, 1921], careful supervision is required to ensure a proper 
oxygenation and a constant temperature. In addition it often seems impossible 
when using caseinogen to obtain a deficiency curve, although apparently no 
fat-soluble A is present in the diet. The rats continue to grow steadily and 
seem apparently healthy. 

In view of these difficulties which were experienced while investigating 
the fat-soluble A content of certain foods, some rat feeding experiments were 
started, in which for the caseinogen dried egg white was substituted. Dried 
egg white, unlike caseinogen, contains very little associated fat and if it is 
an adequate protein when used in a percentage similar to that normally used 
for caseinogen, would be considerably cheaper. 

The first series of experiments was done with a diet containing 30 % 
protein—instead of the usual 20 °4—as analyses of egg white have shown it 
to be somewhat deficient in tyrosine. 

When the diet was first used the dried egg white was crushed and the dry 
protein mixed with the other ingredients of the diet. It was found, however, 
that on this diet many of the rats developed diarrhoea and little or no growth 
was obtained. A diet was then tried in which the dried egg white was soaked 
overnight in tap water and the solution gently heated to coagulate the pro- 
teins. While heating the solution it was constantly stirred and the broken 
clot was then mixed with the other constituents. A fairly solid mass was thus 
obtained, which was readily taken by the rats and was satisfactorily digested!. 

The experiments were carried out on young healthy albino rats, weighing 
about 40-60 g. at the beginning of the experiments (see Fig. 1). Usually the 


1 Subsequently I read Bateman’s paper [1916] on the digestibility of egg proteins and find 
that his results and mine are completely in accord. 


480 M. BOND 


young rats were fed on bread and milk for about a week previously and any 
animal not showing a decided gain in weight during that time was discarded. 
The rats were fed twice daily—usually at 9 a.m. and 3 p.m.— and were weighed 
twice a week. The amount of food given was so arranged that there was still 
a little left uneaten when a fresh quantity was due. 

Each litter was divided into two groups. One group received diet 4 and 
the other group diet B. 


Soaked and cooked egg white 30 g. 30 g. 
_ Salts (Hopkins’ mince) vps 5g. 5g. 
Marmite om ‘iss 5 g. 5 g. 
Strained lemon j juice 65 ida 5 ec. 5 ec. 
Potato starch . ahs ss, BOR, 40 g. 
Fat f butter... 15 g. — 
hardened cotton seed oil — 15 g. 
Water to drink os -- ad lib. ad lib. 


The rats on the B diet showed a Bice | curve, which in a few weeks 
began to flatten out. These rats also frequently suffered from eye inflamma- 
tions. The rats on the A diet grew well, appeared to be healthy and when 
they were mated reproduction was obtained (Fig. 1). In this series of ex- 
periments 20 rats were used. 


170¢ ; 
S 
160 fe) 


a VE Ee: 


120} a WA fo, 
~ od Na i ¢ 
110k myd are 
a-yf if a a: B65 
+ ~e.--* 4 ewe eeeay 


¢ ene 


80P WA fe . 
ih + Ws ¢ Mf diet A 
~ “ rs oY On ee ee ee diet B 


70PF Ss &, ‘y 
- Py S 
oF 
50 i i A A A A 4 A‘ 4 i‘ L a aie =| A A 


weeks 
Fig. 1. Litter of 6. Feeding extending over 8 weeks, A and B diets. 

The results of feeding the lactating rat and her growing young on the egg 
white diet are not very conclusive. One litter has been successfully reared 
on this diet, but as a rule it was found that the young died about the 20th-30th 
day after birth, and the mother became ill unless the diet was changed to one 
of bread and milk as soon as the young rats began to move about the cage. 


BASAL DIET FOR RAT FEEDING EXPERIMENTS 481 


This is probably due to the large proportion of protein in the diet [ Hartwell, 
1921, 1922]. Hartwell has shown 30-40 % protein to be excessive during 
lactation, although the experiments described by her are somewhat different 
from those described here. 

1. In the experiments described in this paper the diet used for the 
lactating rats always contained all the necessary food factors and had been 
given to the mother rats during growth and gestation. 

2. The egg white used has been soaked and cooked, whereas in the cases 
quoted by Hartwell, the protein was used crushed, dry and uncooked. This 
condition of the egg white in itself frequently leads to diarrhoea and death 
in non-lactating and non-sucking rats. 

_ 3. The unhealthy symptoms were not observed as early as in her experi- 
- ments—the young rats always having begun to move about the cage and get 
food for themselves before they began to show signs of weakness. The charac- 
teristic spasms and screaming fits observed by Hartwell were never noticed, 


' but the young rats became less active, were cold and died after a period of 


coma. Weighing of the young rats was not carried out so that it is not possible 
to say when growth ceased. 

A second series of experiments is now in progress, in which the rats are 
receiving a diet containing only 20 % egg white. This is a diet comparable 
to that normally used, but with the substitution of egg white for caseinogen. 
The growth of rats on the 20 % diet is being compared with that of rats 
on the 30 % diet and so far no signs of inadequacy can be detected among 
the rats on the lower protein diet. There has not yet been time to ascertain 
whether this 20 °% diet is adequate for reproduction nor whether it is more 
satisfactory than the 30 % diet during lactation. 

A point in favour of egg white as a source of protein seems to be the 
constancy with which, when this material is used, the characteristic flattening 
of the growth curve is obtained if the diet be deficient in fat soluble A—a 
constancy not always obtained when caseinogen is used. 


REFERENCES. 


Bateman (1916). J. Biol. Chem, 26, 263. 
Coward and Drummond (1921). Biochem. J, 15, 531. 
‘Hartwell (1921). Biochem. J. 15, 140, 563. 

—— (1922). Biochem. J. 16, 78. . 


Bioch. xv1 32 


XL. SYNTHESIS OF VITAMIN A BY A MARINE 
DIATOM (WITZSCHIA CLOSTERIUM W.Sm.) 
GROWING IN PURE CULTURE. 


By HENRY LYSTER JAMESON, JACK CECIL DRUMMOND anp 
KATHARINE HOPE COWARD (Beit Memorial Research Fellow). 


From the Biochemical Laboratories, Institute of Physiology, 
University of London. 


(Received April 28th, 1922.) 


Durtine the early stages of an investigation of the nutrition of certain species 
of marine molluscs it became necessary to gain some information concerning 
the food value of a number of marine plants upon which the shell-fish directly 
or indirectly feed}, 

Amongst other dietary factors to be studied that Shown as vitamin A 
claimed our early attention. By now it is generally understood that the 
higher land animals draw their supplies of this indispensable factor directly 
or otherwise from green plants, since their own tissues do not appear to possess 
the power to synthesise this organic complex. That a parallel condition would 
exist in the sea appeared to us to be highly probable from the fact that certain 
marine algae had been found to share with land plants the power to synthesise 
vitamin A [Coward and Drummond, 1921]. The algae examined in the former 
work were Fucus, Ulva, Cladophora, Polysiphonia and Enteromorpha, but 
these are not truly representative of the ultimate food supply of the species 
of molluscs with which we were working. We therefore decided to test whether 
a typical unicellular marine alga can synthesise vitamin A. 

To obtain the necessary material it was decided to grow the organisms in 
pure culture following the technique described by Allen and Nelson [1910]. 
Pure cultures of three common marine diatoms, Nitzschia closterium W. Sm. 
forma minutissima, Thalassiosira gravida, and Skeletonema were very kindly 
placed at our disposal by Dr E. J. Allen, F.R.S., Director of the Marine 
Biological Laboratory, Plymouth. 

To grow the relatively large quantities of the organisms required for the 
feeding tests was not a simple matter, and in the case of two of the three 
species named our attempts failed owing to their slow rate of growth. With 
Nitzschia, however, we found it a relatively simple matter to prepare ample 


* Unfortunately owing to the sudden and untimely death of our colleague Dr H. Lyster 
Jameson the main line of this work has been discontinued, The present communication records 
the chief result which we had obtained up to the time of his death, by which science and par- 
ticularly marine biology suffers a grievous loss. J. C, D, 


SYNTHESIS OF VITAMIN A BY A MARINE DIATOM 483 


quantities of the diatom. The cultures were prepared by carefully isolating 
a few of the organisms from the original Plymouth sample and allowing them 
to grow in small flasks containing Miquel’s culture fluid or sterilised sea water 
(see reference to Allen and Nelson’s work). At the end of a month the bottom 
of the flask was covered with a thick brown growth of the diatom, but, as Allen 
and Nelson also observed, the cultures do not thrive well after more than a 
month so that at that stage the cultures were transferred to a number of large 
Kilner jars of 4-litre capacity containing the sterile culture fluid. The covered 
jars were allowed to stand in diffuse daylight in front of a window facing 
north, and the growth was accompanied by quite an appreciable evolution 
of oxygen on the brighter days. Frequent microscopical examination of the 
cultures demonstrated their freedom from contamination. By repeated sub- 
culturing in this manner large amounts of the organism were prepared without 
much trouble. 

We were at first doubtful whether the administration of the fresh diatoms 
to rats in the usual manner of testing for the presence of vitamin A would be 
of any value but decided to test this method before going to the trouble of 
preparing by chemical means a special fraction. Somewhat to our surprise 
the organisms were digested during their passage through the alimentary 
canal of the rats as was indicated by the recovery of growth and by the 
presence of empty frustules in the faeces. The usual method of testing was 
followed [Drummond and Coward, 1920], the supplement being administered 
before the daily ration of basal diet was given. To prepare the organisms for 
administration the fluid was decanted off from a month-old culture and the 
thick flocculent growth of diatoms on the bottom filtered on to a thin layer 
of starch at the bottom of a Gooch crucible. After washing with a very small 
quantity of distilled water the almost dry mixture of starch and organisms 
was thoroughly mixed and an aliquot part given to the animals who invariably 
consumed it with readiness. 

The organisms although possessing a deep brown colour whilst in culture 
assumed a dark green tint on drying. The total fat content (ether extract) 
was 5-25 % of the dry weight, and in addition to the characteristic pigment 
of the brown algae, fucoxanthin, there were also present chlorophyll and rela- 
tively large amounts of the lipochrome pigments carotene and xanthophyll. 

Although it was impossible accurately to measure the dose of Nitzschia 
given daily an approximation was made on the basis of the average dry weight 
of a month-old culture which was 0-32 g. 

The curves shown in Fig. 1 show that this diatom may serve as a source 
of vitamin A, the resumption of growth being especially marked at the higher 
dosage of approximately 0-08 g. (dry weight) per day. 

The extraordinary potency of these diatoms as a source of the vitamin A 
may be judged from the fact that this ration of diatom represents approxi- 
mately 4 mg. of fat; which recalls the observations of Zilva and Miura on the 
potency of cod liver oils [1921]. 


32—2 


484 H. L. JAMESON, J. C. DRUMMOND AND K. H. COWARD 


It is significant that this organism promoted growth very much more 
effectively than did the common seaweeds we previously tested. This differ- 
ence may possibly be accounted for by taking into account the much greater 
relative surface of the diatoms. 

The enormous stores, relatively speaking, of vitamin A which are found 
in the tissues of certain marine animals lead one to conclude that they are 
derived from some highly potent source of that factor such as is represented 
by the diatom we have been studying. 


140F 
130F 
120} 
110 


— 
56 8 
Sri oe. ae 


Body weight in grams 


~“ 
oO 


eof - / a 
50 i i i A i 7 i‘ a i i | i i i 


Fig. 1. Daily ration of Nitzschia (dry weight) shown in each period. Preparatory 
period shown by dotted line. 


<—4 weeks — > 
i i‘ 7 ‘? a 


—— 


We essayed to prove the transference of the vitamin from the diatoms to 
the molluscs we were investigating but owing to the sad death of our colleague 
the experiments were carried no further than the demonstration that many 
species of molluscs are valuable sources of the A factor. The production of 
vitamin A in green land plants capable of carbon assimilation [Coward and 
Drummond, 1921] is therefore paralleled by a similar synthesis in marine plants 
containing photocatalytic pigments. Further, the fundamental dependence 
of terrestrial animal life on the fresh green leaf for its supplies of vitamin A 
would appear to be paralleled by a similar dependence of marine animals on 
the marine flora, particularly the microscopic plants. 


SUMMARY. 


1. Pure cultures of a common marine diatom, Nitzschia closterium, grown 
in Miquel’s solution or sterilised sea water synthesise large amounts of 
vitamin A. . 


2. A parallel is drawn between the dependence of land animals on fresh 
green leaves and that of marine animals on the synthetic activity of the marine 
flora for their supplies of vitamin A. 


3. A number of molluses were found to contain considerable amounts of 
vitamin A. 


SYNTHESIS OF VITAMIN A BY A MARINE DIATOM —§ 485 


The greater part of the expense incurred in this research was defrayed out 
of a grant made by the Medical Research Council to whom our thanks are due. 


We also wish to express our appreciation of the assistance extended to us by 
the Ministry of Agriculture and Fisheries. 


REFERENCES. 


Allen and Nelson (1910). Quart. J. Micros. Sci. 55, 361. 
Coward and Drummond (1921). Biochem. J. 15, 530. 
Drummond and Coward (1920). Biochem. J. 14, 661. 
Zilva and Miura (1921). Biochem. J. 15. 


XLI. THE PHOSPHOLIPIN OF THE BLOOD AND 
LIVER IN EXPERIMENTAL RICKETS IN DOGS. 


By JOHN SMITH SHARPE. 
From the Institute of Physiology, University of Glasgow. 


(Received May Ist, 1922.) 


In a paper in which he advances a theory of the metabolism in rickets, 
Prof. Noél Paton [1922] points out that recent work seems to indicate that 
' the supply of phosphorus rather than of calcium may be the limiting factor 
in ossification and develops the idea that a disturbance in the metabolism of 
the phospholipins may be involved. 

Since the phospholipins and especially lecithin play so important a part 
in the metabolism of phosphorus an attempt has been made to discover 
whether there is any modification in the amount in the blood and liver in 
rickets. 

Unfortunately, the material available has not been of such a nature as 
to enable a conclusive answer to be given, for in each of the experiments in 
which the blood and liver were preserved for analysis, either all the pups 
remained free of rickets or all developed the condition. The results obtained 
are however sufficient to justify their being recorded and to allow a tentative 
conclusion to be drawn. 

A considerable amount of work has been recorded upon the phospholipin 
of the blood but little of it has a bearing upon the present investigation. 
Bloor [1915] showed that in feeding with fat the lecithin content of the blood 
is raised and [1918] that this is due to an increase of phospholipin in the 
corpuscles. 

MeTHOp. 


The blood and the livers were taken at the time of death by chloroform 
and bleeding, from experiments 5, 6, 7 and 8 of the series described by Noél 
Paton and Watson [1921]. The details of experiment 12 have not yet been 
published, but all the pups remained free from rickets. Unfortunately the 
non-phospholipin phosphoric acid was determined only in lots 7 and 8. 


Metruops or ANALYSIS. 


Livers. The livers were preserved in formalin. A weighed amount was 
repeatedly extracted with alcohol-ether (10-1) by rubbing up in a mortar. 
The extract was dried in a steam-bath and in a vacuum desiccator and weighed 
and from this total fats and alcohol-ether soluble P,O, were determined. The 
alcohol-ether-insoluble P,O, was estimated in the residue. Hach was ignited 


&-€ 3 e-€ GZ 0-6 eZ ra LZ cE a 9% L3 a rar 
#800 610-0 001-0 20-0 #800 9200 0600 +200 8900 1200 GL00 80-0 910-0 220-0 
83 080TH €8-Z TT. OG © SEU AS) OT EIS" 280 
IOATT pool, IOATT pool TOATT pool TOATT pool, TOAVT pool IOATT poold TOATT poolg 
(a °A 
pue gq ‘A 
uaeMjoq WOl4Ip 
SPOYOUY SJOYOI YS soMory syoyory SOY poylVpL = WOO) SoH == SOON 4 TITS 
cg . cel 191 911 3 99 9g 
YOO WIC Joye 
‘3 CF [ROUTyRG YOoM Y9G Joye YooM RE 199;e 
09 OCT 0} ‘3 QF [voulzeO ‘3 CP [RoulyVg 
peonpar “09 (¢Z ‘B 0 (pep) *00 OT 0} peonper YOom WIC Joqyze “B oF Yoo WAC 1ozye “3 CF 
yyeca poyeredeg yyraa sfoO"UM *00 OZ Yrur spoy, M [TeoulzVO “00 OGZ Arar poyeredeg yTeourzeg ‘00 O¢Z Yr afPouM, 
’ mes “ ~—" , ee a San, 
€d ‘IA ca ‘A 92 TA td ‘IA td A Id ‘A Gq ‘IA 
“plo syxooM ET 48 TA fo esoqy “pjo SYOOM JT 48 peT[oT A jo sdng 
*SLUMOIY 
3 ror = iar Lt 98 BE SESE 8% SyBy 1230} Jo % sv 
mon 0-0 — 720-0 0280 800 981-0 690-0 602-0 290-0 
= 06-1 — Sit L9 St 82 Of 9 981 
TOATT poo[g TOATT pool JOAVT poolg TOATT pool IOATT poold 
o¢ 00z €8 OZI 002 
*B OG 07 posvoroUr “3 Og prorg 
“3 PE 0} posvoroUl “3 ¢-Zz (polap) Arr rear [INT 
¢ t € z I 
ROR NB ¢ 
"PIO syoom GT 98 porey sdng “Ty 
6 81 + 61 6-F raat c-9 TZ 0-8 13 oF LT SyBy [%I04 JO 
60Z-0 990-0 LLT-O S200 $970 6900 6FT-0 8900 e170 £900 F980 160-0 eee a ta 
CIL0 ZOO 601-0 S800 L9TO 9200 I910 L200 9600 800 8600 660-0 : : 
FEO O10 982-0 ITO SeFO 280-0 O1€0 800 6080 8600 Z9F0 OFI-0 : % 
6% FES «SS «= OS rE Cis GS WI ZI = ODT GLGCtCHG| ey 
JOATT = poojg «= OAT «= poor’: SssOATY «= pOOTCsCOAT © poof =ATT «= poof: «OAT — pOOTT 
998 00€ 009 OFZ 08z 0z9 PY aa as 
‘3 OE prog 
‘3 ¢-Z% (perp) ‘BF privy “5 Og peag ‘B OF pearg 
YF ureoro [NT ‘8 ¢-2e (partp) ATrur poyeredog ‘8 ¢-2 (petsp) sITar poyeredag eA 
: an wa” a pes | 
¢qd ‘IIA td ‘IIIA td ‘IIA td ‘IIA @q ‘IIIA L@ ‘IIA 


‘PIO SY9OM ST 4B TILA JO esoyy “Pfo syoom QT 9 por] ITA Jo sdug *T 
“TVNUON 


eee 878} [B40} jo % 
se utdyoydsoyd ur *O%g 

% urdroydsoyd ut *o*q 
eee oe % 8}RF 1879], 


ba uolyIpuoyD 
% FBIM Ul UTeX) 


worp sod yor 


syuouted xy 


sutdryoydsoyd ur *o%q 
surdyoydsoyd ur *O*g 
oF “et Ye S¥8F [CIOL 


% PqsIOM. Ur UTES 


torp 10d yor 


% se suidyoydsoyd ut *O%q 
% Joyooye ur eyqnjosut °O*q 


soyd ut *Q%q 
% *0%d T24OL 
ve -%, 8yeF [BIO 


% surdrjoy 


% FSI ur ures 


aorp rod qoIq 
syuouts0d x 7] 


488 J. S. SHARPE 


with an excess of phosphorus-free calcium carbonate, the residue dissolved 
in hydrochloric acid with the addition of a little nitric acid and the phosphorus 
estimated by the molybdate method in the usual way. 

Blood. The blood was drawn into a measured quantity. of methylated 
spirit, well shaken and allowed to stand several days. This was filtered into 
a weighed flask and the residue extracted with aleohol-ether (10-1) many 
times, the washings being added to the filtrate. This was evaporated and the 
total fat and the P,O; determined as in the liver. 


SUMMARY OF TABLES. 


P,O; in Phospholipin. 
Non-rachitic Rachitic 
aE 2 See, fee es ee 
Maximum Average Minimum Maximum Average Minimum 
Blood 0-059 0-043 0-026 0-033 0-025 0-019 
Liver 0-320 0-162 0-096 0-100 0-082 0-068 
CONCLUSIONS. 


The tables show: 


1. That the percentage of phospholipin in the blood and in the liver bears 
no relation to the gain in weight, to the amount of fat in the diet, to the total 
fat in blood and liver or to the age of the animal. 

2. That in the rachitic pups the phospholipins of the blood and of the 
liver are lower than in the non-rachitic. . 

3. That the amount of phospholipins of the liver is smaller in proportion 
to that of the blood in rachitic than in non-rachitic animals. 


I have to thank Prof. Noél Paton for the help and advice he gave me 
during the investigation. The expenses for the work were defrayed by a 
grant from the Medical Research Council to which body I tender my thanks. 


REFERENCES. 


Bloor (1915). J. Biol. Chem. 28, 317. 
—— (1918). J. Biol. Chem. 36, 33. 
Paton, D. Noél (1922). Brit, Med. J. i, 379. 
Paton, D. Noél and Watson, A. (1921). Brit. J. Rup. Path, 2, 75. 


XLII. THE CONSTITUENTS OF THE FLOWERING 
TOPS OF ARTEMISIA AFRA, JACQ. 


By JOHN AUGUSTUS GOODSON. 
From the Wellcome Chemical Research Laboratories. 


(Received May 2nd, 1922.) 


Tue genus Artemisia comprises some 350 species, of which only about 30 have 
been chemically examined. In most cases the investigation has been confined 
to the essential oil, but a few of the more important species such as wormwood, 
Artemisia Absinthium Linn., used in the preparation of absinthe, and worm- 
seed, A. maritima var. Stechmanniana Bess. which grows in Turkestan and is 
the sole source of the anthelmintic, santonin, have been more thoroughly 
examined. In consequence of the difficulty of obtaining santonin since 1914, 
other sources of this indispensable drug have been sought and in this con- 
nection attention has been given to other species of Artemisia, of which the 
author has examined four. A. brevifolia Wall. (which according to the Index 
Kewensis is a form of A. maritima Linn.) from India was found to contain 
santonin, as previously recorded by Greenish and Pearson [1921] and 
Simonsen [1921]; specimens of A. afra Jacq. from South Africa, A. mexicana 
Willd. from Mexico, and A. monosperma Delile (A. deliliana Bess.) from Egypt 
were found to contain no santonin. The closely related Egyptian plant 
Santolina chamaecyparissa was also found to be free from santonin. 

Judging from these results and others previously recorded it would seem 
that the occurrence of santonin is restricted to species of Artemisia indigenous 
to East Europe and Asia, the only exception so far recorded being A. gallica 
Willd., which occurs in France, and according to Heckel and Schlagdenhauffen 
[1884] contains santonin. These authors, however, give no evidence for this 
statement and their observation has not been confirmed}. 

Of the four species examined by the author only one, A. afra, was available 
in quantity and this was fully investigated to see if it contained anything that 
could be regarded as a precursor or a derivative of santonin. The results given 
below show that this plant contains camphor, a wax-ester probably ceryl 
cerotate, triacontane, scopoletin and quebrachitol, none of which can be 
regarded as connected with santonin. 

l-Camphor has been recorded in several species of Artemisia, e.g. in 
A. Herba-alba Ass. [Grimal, 1904], A. cana Pursh [Whittelsey, 1909] and 
A. annua Linn. [ Yoshitomi, 1917]. In A. afra the camphor is dextrorotatory, 


1 Since this paper was submitted for publication a preliminary notice of a communication 
by Viehoever and Capen [1922] has appeared, in which it is stated that santonin has been 
isolated from A. mexicana and A. neo-mexicana. J.A.G. 


490 J. A. GOODSON 


but the rotation is much lower than that of normal d-camphor. Two specimens 
were found to have [a]*? + 9-7° and [a]? + 9-3° instead of [a]?? + 42-4°. 
As it is usually considered that only two optically active isomerides can exist 
notwithstanding the fact that the camphor molecule contains two asym- 
metric carbon atoms, and as camphor is not easily racemised, it would appear 
that both d- and /-camphor are present in A. afra. It may -be noted that 
inactive camphor has been found in Chrysanthemum sinense var. japonicum 
[Keimatsu, 1909]. 

Scopoletin and quebrachitol have not been recorded previously in the 
genus Artemisia or even in the natural order Compositae. The former has 
been found in Atropa Belladonna Linn. | Kunz, 1885], Fabiana imbricata Ruiz 
et Pav. [Kunz-Krause, 1899] and Scopolia japonica Maxim. [EKykman, 1884] 
of the natural order Solanaceae; Ipomoea purga Hayne (Convolvulaceae) 
' [Power and Rogerson, 1910]; Gelsemiuwm sempervirens Ait. (Loganiaceae) 
[ Moore, 1910] and Prunus serotina Khrh. (Rosaceae) [Power and Moore, 1909]. 
Quebrachitol has only been found in four other plants, viz. Aspidosperma 
quebracho Schlect (Apocyanaceae) [Tanret, 1889]; Grevillea robusta A, Cunn 
(Proteaceae) [Bourquelot and Fichtenholz, 1912]; Heterodendrum oleaefolium 
Desf. (Sapindaceae) [ Petrie, 1918] and Hevea brasiliensis Muell. (Euphorbia- 
ceae) [Pickles and Whitfield, 1911], all of which belong to different natural 
orders. The other components ceryl cerotate and triacontane are fairly widely 
distributed in plants. 


EXPERIMENTAL. 


For the material used in this investigation the author is indebted to Mr 
I. B. Pole Evans, Chief of the Division of Botany, Union of South Africa. 
It consisted of the flowering tops of the plant including stems, leaves and 
florets. 

When extracted with Prollius’s fluid it yielded a mere trace of alkaloid 
and furnished the following percentages of extract on exhaustion in a Soxhlet 
apparatus with solvents in the order named: petroleum (b.p. 35-60°), 7-0; 
ether, 9-5; chloroform, 2-1; ethyl acetate, 3-4; alcohol, 15-5, Special search 
was made for santonin with negative results. For the purpose of a more 
complete examination a quantity (11-1 kilograms) was extracted in succession 
with hot solvents, petroleum (b.p. 35-60°), ether and alcohol. The alcoholic 
extract was further fractionated by drying on a quantity of the flowering tops 
previously exhausted with petroleum and ether and re-extracting hot with 
chloroform, ethyl acetate and alcohol in succession. The petroleum extract 
on distillation in steam and extraction of the distillate with ether yielded 
55-2 g. of essential oil; a further 28-6 g. was subsequently obtained from the. 
ether extract, equivalent in all to 0-75 % by weight of the flowering tops. 

The material left behind in the distillation flask on boiling with petroleum 
(b.p. 35 to 60°), deposited on cooling about 81 g. of crude wax-ester melting 
at 74-76°. The residue left on removal of the petroleum was boiled with ether 


CONSTITUENTS OF ARTEMISIA AFRA 491 


and this solution on cooling gave 17 g. of crude hydrocarbon, melting at 62 
to 66°. The ethereal solution was then extracted with dilute hydrochloric 
acid, but yielded no alkaloid; it still contained some free and combined fatty 
acids, which were not investigated, and some unsaponifiable matter, which 
appeared to contain a sterol giving a red coloration with acetic anhydride and 
sulphuric acid. 

Examination of the Essential Oil. The oil Oeaiaa the odour of camphor, 
had a specific gravity 0-9453 at 15°/15° and specific rotation [a])? + 5-8°. 
On washing with solutions of sodium carbonate, sodium hydroxide and sodium 
bisulphite (saturated), it lost to each quantities of material too small to be 
investigated in detail. The acids extracted by sodium carbonate were frac- 
tionally precipitated as silver salts, the fractions containing 24-7, 34-9, 38-6, 
39-9 and 41-3 % of silver respectively. Silver pelargonate, CsH,,COOAg 
requires Ag = 40-4. The saponification value of the oil before acetylation 
was 33-9 and after acetylation 73-9. 

On distillation the following fractions per cent were obtained (1) below 
180°/760 mm., 31-8; (2) below 100°/25 mm., 27-5; (3) at 100—-120°/25 mm., 
12-1; (4) at 120-140°/25 mm., 12-3; (5) at 140-180°/25 mm., 8-0; (6) and above 
180°/25 mm., 7-7. 

Isolation of Camphor. Fractions (2) and (3) on standing deposited camphor, 
a further quantity of which separated on redistilling the fractions and freezing 
the distillates with solid carbon dioxide, the total quantity obtamed amounting 
to 13-5 % of the crude oil, but this does not represent the total amount of 
camphor in the oil as much still remained in the various fractions. The crude 
camphor was recrystallised from dilute alcohol until its melting point was 
constant at 178° (corr.) and showed no depression of melting point on ad- 
mixture with natural d-camphor melting at 180° (corr.). The specific rotation 
was low, being only [a]?? + 9-7° in 95 % alcohol (a7) = + 0-98°, 1 = 1 dem. 
C = 10-148), that of natural d-camphor used for comparison being [a]? + 42-4° 
(a2 — + 4-35°, 1 = 1 dem., C = 10-261). Found C = 78-9; H = 10-5. Calcu- 
lated for camphor, C,9H,,0, C = 78-9; H = 10-6. 

_ The oxime was prepared and after recrystallisation melted at 118-119° 
(corr.), the melting point remaining unchanged on admixture with camphor- 
oxime prepared from natural d-camphor. It was laevo-rotatory [a], — 3-4° 
(@p = — 0-16°, 1 = 0-5 dem., C = 9-296). The semicarbazone after recrystalli- 
sation from alcohol melted with decomposition at 241° (246° corr.) and when 
mixed with camphor-semicarbazone prepared from natural d-camphor, m.p. 
246° (corr.) with decomposition, it melted at 242° (247° corr.). The melting 
point of the semicarbazone of d-camphor is usually given in the literature as 
236-238°, but this is undoubtedly too low. (Found C = 63-4; H = 9-4. Calcu- 
lated for camphor-semicarbazone, C,,H,,N,0, C = 63-1; H = 9-2.) . 

Examination of waz-ester. The wax-ester was recrystallised many times 

from ethyl acetate or petroleum and then melted constantly at 79°, although 


492 J. A. GOODSON 


it appeared still to contain a small quantity of hydrocarbon. From the analyses 
of the ester and its hydrolytic products, it would appear to be ceryl cerotate 
or a closely related ester. (Found C = 82-5, 82-5; H = 13-5, 13-7. Calculated 
for ceryl cerotate, C,,H,;COOC,,H;,, C = 82-0; H = 13-8.) 

The alcohol obtained on hydrolysis melted at 74°, and gave an acetyl 
derivative melting at 64-66°. Ceryl alcohol melts at 81°, and ceryl acetate at 
64-5°. (Found C = 81-6, 81-5; H = 14-2, 13-9. Calculated for ceryl alcohol, 
Cy,H;,0, C = 81-6; H = 14-2.) 

(Found C = 79-0; H = 13-3. Calculated for ceryl acetate, CH. s000C Hs, 
C = 79-2; H = 13-3.) 

The ree acid produced on hydrolysis melted at 76°. (Found C = 78-7, 
78-8; H = 13-3, 13-5. Calculated for cerotic acid, C.,H;.0.,C = 78-7; H = 13-2; 
m.p. 82°.) 

Isolation of Triacontane. The crude hydrocarbon was distilled under re- 
duced pressure and recrystallised several times from ethyl acetate and 
petroleum until it melted constantly at 66°. (Found C = 84-9; H = 14-9. 
Calculated for triacontane, C5)7Hg, C = 85:2; H = 14-8; m.p. 65-5°.) 

After removal of the essential oil, wax-ester and hydrocarbon from the 
ether extract, the latter was extracted with dilute hydrochloric acid, which 
removed no alkaloid or other basic material and then with sodium carbonate 
solution, followed by potassium hydroxide solution.. 

Isolation of Scopoletin. The sodium carbonate extract was acidified with 
hydrochloric acid, and extracted with ether. The ethereal solution on con- 
centration deposited a crystalline substance, a further quantity of which was 
obtained by extracting the dilute hydrochloric acid extract of the original 
ether extract with ether. The substance was purified by recrystallisation from 
ethyl acetate, yielding 2 g. of pure material, which formed pale yellow needles 
and dissolved in sodium carbonate solution giving a solution possessing a 
striking blue fluorescence. It melted at 203° (208° corr.) the melting point 
being uninfluenced on admixture with scopoletin. (Found C = 62-4, 62-4; 
H = 4-2, 4-5. Calculated for scopoletin (4-hydroxy-5-methoxycoumarin), 
C,9H,0,, C = 62-5, H = 4-2.) 

The chloroform extract yielded no crystalline substance and dilute hydro- 
chloric acid removed from it only a trace of material giving alkaloid re- 
actions. 

Isolation of Quebrachitol (Methyl l-inositol). The residues of the ethyl acetate, 
and alcoholic extracts, after concentration, deposited a mixture of resinous 
matter and crystals, which was extracted with water, the solution was boiled 
with animal charcoal, filtered, and concentrated and finally alcohol was added 
when a quantity of crude methyl-l-inositol crystallised out corresponding to 
0-43 %, of the flowering tops used. After several recrystallisations it melted 
constantly at 194° (corr.), and had a specific rotation [a], — 81-6° (ap) = — 
816°, 1 = 1 dem., C = 10-0), Found C = 43:2, 43-4; H = 7-0, 7-3. Calculated 
for quebrachitol, CH,OC,H, (OH),, C = 43:3; H = 7:3. 


CONSTITUENTS OF ARTEMISIA AFRA 493 


The acetyl derivative, prepared by the action of acetic anhydride in 
presence of pyridine, was recrystallised from ethyl acetate and obtained in 
rosettes of needles, melting at 94-95°. (Found C = 50-8; H=6-1. Calculated 
for penta-acetylmethylinositol, CH,OC,H, (OOC.CH,);, C = 50-5; H = 5-9.) 

There is no published record of previous chemical work on Artemisia afra 
but the author desires to state that Messrs H. W. B. Clewer, and R. R. Baxter 
have made preliminary examinations of the plant in these laboratories, and 
their records, which include the isolation of camphor and of the crystalline 
substance now shown to be quebrachitol, have been available to him. 

In conclusion, the author desires to express his warmest thanks to Dr 
T. A. Henry for his advice and criticism throughout the course of the work. 


REFERENCES. 


Bourquelot and Fichtenholz (1912). J. Pharm. Chim. 6, 346. 
Eykman (1884). Rec. trav. chim. Pays-Bas, 3, 171. 

Greenish and Pearson (1921). Pharm. J. (4), 52, 2. 

Grimal (1904). Bull. Soc. Chim. 31, 694. 

Heckel and Schlagdenhauffen (1884). Compt. Rend. 100, 804. 
Keimatsu (1909). J. Pharm. Soc. Japan, 326, 1. 

Kunz (1885). Arch. Pharm. 223, 701. 

Kunz-Krause (1899). Arch. Pharm. 237, 1. 

Moore (1910). J. Chem. Soc. 97, 2223. 

Petrie (1918). Proc. Linnean Soc. N.S.W. 43, Part 4. 

Pickles and Whitfield (1911). Proc. Chem. Soc. 54. 

Power and Moore (1909). J. Chem, Soc, 95, 243. 

Power and Rogerson (1910). J. Amer. Chem. Soc. 32, 95. 
Simonsen (1921). J. Indian Industries and Labour, Nov. 539. 
Tanret (1889). Compt. Rend. 109, 908. 

Viehoever and Capen (1922). J. Amer. Pharm. Assoc. 11, 393. 
Whittelsey (1909). Wallach Festschrift, 668. 

Yoshitomi (1917). J. Pharm. Soc. Japan, 424, 1. 


XLII A METHOD FOR THE ESTIMATION OF 
SMALL QUANTITIES OF CALCIUM. 


By PATRICK PLAYFAIR LAIDLAW anp . 
WILFRED WALTER PAYNE. 


From the Pathological Department, Guy's Hospital. 
(Received May Sth, 1922.) 


Our knowledge of calcium metabolism is limited, and probably hampered, by 
the lack of a good and accurate method of estimating small quantities of 
calcium. In quite a number of biological problems the estimation of amounts 
of calcium of the order of 0-1 mg. is frequently desired. The method about 
to be described is applicable to even smaller quantities than this; and is, we 
believe, remarkably accurate if used, with reasonable care, and superior to 
those in use at the present time. — 

The modern methods are of two types (a) biological, (b) chemical. The 
biological methods depend on the fact that calcium is necessary for blood 
coagulation. Wright [1912] and later Vines [1921] have developed methods 
with this fundamental fact as basis. Inasmuch as the nature of the coagula- 
tion process is only imperfectly known these methods suffer from being built 
on an insecure foundation and are liable to break down for unknown reasons. 
They are, however, sensitive and within limits not clearly defined useful. If 
a method equally sensitive with a secure basis were available they would 
rarely be employed. The more recent chemical methods such as those of 
Howland and Kramer [1920], Kramer and Tisdall [1921] are difficult to use, 
and consisting as they do of microtitrations of oxalate by N/100 potassium 
permanganate are liable to error. Thus small amounts of organic matter such 
as carbon from imperfect incineration of a biological product, dust, or even 
the organic material in poor distilled water, may each and all cause significant 
error. The weak solution does not keep well, and the end point is not sharp. 
With elaborate precautions and in the hands of the select few they give good 
results; but we obtained very disappointing results on trying them out. 

The principle of the method. The caleium is separated in the first place as 
oxalate and the oxalate is converted into calcium alizarinate under defined 
conditions. The crystalline calcium alizarinate is collected and washed on a 
Gooch filter and decomposed by oxalic acid. The alizarin is dissolved in alcohol 
made up to known volume with a relatively large volume of dilute ammonia, 
and the resultant solution of ammonium alizarinate is compared in a Dubose 
colorimeter with a standard dilution of ammonium alizarinate. The primary 


ESTIMATION OF SMALL QUANTITIES OF CALCIUM = 495 


separation as oxalate is necessary since in spite of many trials no clean 
separation of magnesium and calcium alizarinates proved feasible. Since the 
calcium in the alizarinate is only one-sixth the weight of the alizarin, the 
initial weight of calcium to be estimated is multiplied by six; and since the 
colorimetric comparison is carried out against a 0-001 % scladien of alizarin, 
the method is very sensitive. Given cleanliness, reagents free from calcium, 
pure alizarin, and avoiding undue hurry, 0-1 mg. calcium may be estimated, 
to 0-002 or 0-005 mg. without much trouble. 

Precvpitation of calcium oxalate, This is done along the usual lines and 
may be carried out on the ash of a biological product, or sometimes, as in 
the case of blood serum, directly. After careful incineration the ash is dis- 
solved in 0-5 cc. of N HCl and transferred with about 2 cc. wash water to a 
centrifuge tube. One cc. of a saturated solution of ammonium oxalate is added, 
followed by 2 cc, of a saturated solution of sodium acetate, to which some 
ammonium oxalate has previously been added to exclude calcium contamina- 
tion. The whole is mixed and allowed to stand for three hours, or if more 
convenient till next day. In our experience three hours’ standing, especially 
in the case of direct estimations in blood serum, gives more uniform results 
than shorter periods, though most of the oxalate separates as a rule in a much 
shorter time. The oxalate precipitate is centrifugalised to the bottom of the 
tube (two minutes at 4000 per min. is sufficient) and the supernatant fluid 
sucked off as completely as possible without disturbing the oxalate precipitate 
at the bottom of the tube. Three cc. of a 0-1 °% solution of ammonium oxalate 
are then added, the precipitate stirred up and the centrifugalisation repeated 
at once. The supernatant fluid is once more sucked off as completely as 
possible, In this way all the calcium is obtained as oxalate free from all but 
traces of magnesium, and with only a little excess of oxalate. 

Conversion of oxalate into alizarinate. The oxalate is dissolved in 0-5 ce 
of N HCl, and transferred to a clean test-tube. The centrifuge tube is washed 
out four times with water. The resultant dilute solution in the test-tube 
should measure 8 to 10 cc. A small volume is disadvantageous, as oxalate 
may separate at a later stage if the volume is kept too small. Excess of an 
alcoholic solution of alizarin is then added (1 cc. of a 1 in 1000 solution is 
sufficient for quantities of calcium of about 0-1 mg.). The test-tube is then 
warmed on a water bath to about 80° C., and five drops of strong ammonia 
solution added. The contents of the tube turn deep purple on mixing. It is 
best to add at this stage a minute amount of crystalline calcium alizarinate 
though it is not essential. The step ensures the full and ready separation of 
the calcium alizarinate in a crystalline form which facilitates filtration later. 
The tube is kept warm for about one hour then set aside to cool and stand till 
the following day. The whole of the calcium should then have separated in 
clumps of blue black microcrystalline needles at the bottom of the tube and 
the supernatant fluid should be pale in colour. ; 

The calcium alizarinate is filtered off on a Gooch crucible witha fine asbestos 


496 P. P. LAIDLAW AND W. W. PAYNE 


layer at a water pump, and the precipitate and filter washed with dilute | 
ammonia. At this point care must be taken not to suck much air through the 
filter since acids, even the carbonic acid of the air, will decompose the calcium 
salt. Paper filters are not so good as asbestos. The ordinary papers contain 
significant amounts of salts which interfere; and extracted papers have their 
own. difficulties. 

The delivery from the Gooch crucible is now arranged to return fluid to 
the test-tube in which crystallisation of the alizarinate took place, and 1 ce. 
of a strong solution of oxalic acid in 50 % alcohol allowed to run all over the 
asbestos and precipitate. The calcium alizarinate decomposes at once and the 
blue black colour becomes orange. The alizarin set free is washed through 
with warm 95 % alcohol along with excess of oxalic acid into the test-tube. 
The asbestos should now be pink or very pale purple. The nature of this 
- colour is unknown, but it does not appear to be due to any calcium compound 
and is usually greater when the oxalate is precipitated directly from blood 
serum, and less in the case of ash estimations. The alcoholic solution of alizarin 
is transferred to a 50 cc. standard flask, the test-tube washed out with dilute 
ammonia, the contents of the flask made just alkaline with ammonia, and 
the volume made up to the mark with water. 

The standard solution for comparison is made up as follows. Two cubic 
centimetres of an accurate M/1000 solution of alizarin in alcohol are run into 
a 50 cc. standard flask, followed by 1 cc. of the oxalic acid solution in 50 % 
alcohol, and a volume of alcohol approximately equal to that used in ex- 
tracting the alizarin from the Gooch crucible when dealing with the unknown. 
Water is added, the contents of the flask are made just alkaline with ammonia, 
and the volume made up to the mark. In this way a perfect colour match is 
readily obtained as the reactions of the test and standard solutions are nearly 
the same. A large excess of ammonia in one flask yields a purple tint as com- 
pared with a red one. A small difference in ammonia content is insignificant 
owing to the buffer effect of the oxalate. 

The unknown and the standard solutions are then compared in a Dubosc 
colorimeter. 20 to 25 mm. depth of the standard solution yields a suitable 
tint for colour match. 

The alizarin in the unknown solution is calculated from the depth of the 
column of fluid which gives a perfect match with the standard, and this 
figure divided by six gives the weight of the calcium. 

The method is simple and though time must be spent over the whole 
process owing to the waits for complete precipitation at two stages, the actual 
time spent on manipulation is not great, and if a series of estimations is being 
done the time spent on any one is still smaller. It may be stated that there 
are many indications that if still smaller amounts than 0-1 mg. are in question, 
the use of small scale apparatus would enable these to be estimated with fair 


accuracy. 


ESTIMATION OF SMALL QUANTITIES OF CALCIUM = 497 


EXPERIMENTAL, 


The composition of the calcium alizarinate as prepared from thé oxalate, 
under the defined conditions of excess alizarin, in the presence of ammonia, 
and in weak alcoholic solutions, was determined by incineration of three 
separate specimens of the crystalline solid; conversion of the calcium carbonate 
so obtained to sulphate; and weighing after heating and cooling: 


Alizarinate 
taken Sulphate Calcium Theory 
0-2424 g 0-1023 0-0301 0-0307 
0-4855 g. 0-2053 0-0604 0-0614 
0-2535 g. O-1111 . 00327 0-0321 
0-2535 g. calcium alizarinate dried over sulphuric acid heated to 
110° C. lost 0-0158 g. 1H,0 =0-0144 g. 
120°C. ,, 0-0213 g. 
160°C. ,, 0-0276 g. 2H,0 =0-0288 g. 


190° C. began to decompose. 

The weighing of the heated alizarinate was difficult since there was gain 
in weight during the weighings. The data show that under the conditions of 
experiment a salt of the composition Ca,Aliz, + 2H,O separates. We also 
have evidence of the formation of another calcium salt of a different com- 
position when there is not excess of alizarin present. 

The method was tried out on an artificial salt mixture, made up to imitate 
the ash of blood serum but with the magnesium in excess, of the following 


composition: 
NaCl ... “s .. 46340 
KH,PO, eg .. 00395 
MgSO, (dry) ... in “OTST 
CaCO, ee ... 02410 
Oe. ut: ae 8 cc. Water to 500 ce. 


A trace of iron was present probably due to impurity in the magnesium 
sulphate. This was no disadvantage since this metal may be present in blood 


serum ash from haemoglobin. 
The following table shows some of the results we obtained when estimating 


various amounts of the salt mixture: 


Solution taken Ca found Calculated Difference 

ce. mg. mg. mg. 
0-9 0-171 0-1737 —0-0027 
0-7 0-140 0-135 +0-005 
0-6 0-117 : 0-115 +0-002 
0-5 0-102 0-096 + 0-006 
0-4 0-0815 0-077 +0-0045 
0:3 0-056 0-058 — 0-002 


Repeated estimations of quantities of about the amount normally present 
in blood serum gave the following figures: 


Calcium taken Found Difference 
mg. mg. mg. 
0-116 0-113 — 0-003 
0-116 0-118 +0-002 
0-116 0-1196 +0-0036 
0-096 0-098 +0-002 
0-096 0-098 +0-002 
0-096 0-096 0-000 


Bioch. xv1 33 


498 P. P. LAIDLAW AND W. W. PAYNE 


Estimations on mixed samples of human blood serum are given below. 
It will be observed that the figures for the ash of the serum and the direct 
determinations are practically the same. Other figures show that the agree- 
ment is even closer if 24 hours be allowed for the separation of the oxalate 
in the case of the direct estimation. Other experiments on heated serum 
indicate that after heating some of the calcium becomes bound to the serum 
proteins and cannot be estimated directly: 


Ca in ash, Ca directly without 


mg. ashing, mg. 
0-098 . 
Sample 1 {0-100 - 0-008 
0-097 “091 
Sample 2 10-004 rox e 
SuMMaRY. 


A method for estimating small quantities of calcium is described. The 
method is colorimetric and depends on the insolubility of calcium alizarinate. 
It is accurate on 0-1 mg. to 0-002 mg. 


REFERENCES. 


Howland and Kramer (1920). J. Biol. Chem. 48, 35. 

Kramer and Tisdall (1921). J. Biol. Chem. 48, 223. Bull. Johns Hopkins Hospital, 32, 44. 
Vines (1921). J. Physiol. 55, 86. : 

Wright (1912). The Technique of the Teat and Capillary Tube, 85. 


XLIV. PRODUCTION OF HYDROGEN PEROXIDE 
BY BACTERIA. 


By JAMES WALTER McLEOD ann JOHN GORDON. 
_ From the Department of Pathology, University of Leeds. 


(Received May Sth, 1922.) 


A stupy of phenomena of auto-inhibition in bacterial cultures was published 
by McLeod and Govenlock [1921] in 1921 showing that a series of investiga- 
tions on this subject had led to the demonstration of such phenomena in 
connection with the growth of many bacteria. C. Eijkman’s results [1904 
and 1906, 1, 2], which have been more questioned than accepted, were fully 
confirmed. ; 

It was shown further that the Pneumococcus, a bacterium, which did not 
appear to have been investigated previously in this respect, produced in the 
course of growth a substance inhibitory to its own growth and to that of 
most bacteria. This inhibitory effect was more powerful than that observed 
in connection with the growth of any other bacterium examined and the 
inhibitory substance was found to be thermolabile (85°). The labile character 
of the substance produced suggested an analogy with toxins and ferments 
and for this reason the name “bactericidins” was suggested for such sub- 
stances. Subsequent work has shown however that the substance produced 
by the Pneumococcus, at all events, is simpler in character and is in all 
probability hydrogen peroxide: but the question whether other bacteria 
produce specific inhibitory substances analogous to ferments remains an 
open one. 

A short summary of the work done to show that the Pneumococcus pro- 
duces hydrogen peroxide in culture has already been published as an abstract 
of a communication to the Pathological Society [McLeod and Gordon, 1922]. 

It is the purpose of this paper to give in some detail the experimental 
evidence for this conclusion together with some more recent observations on 
the subject. 


Conditions necessary for the production of pneumococcal inhibitory 
substance in fluid media, 

The original observations establishing the existence of a thermolabile 
bactericidal product in pneumococcal cultures had been made in solid media- 
serum agar—and such were obviously unsuitable for attempting to concen- 
trate the bactericidal substance. 

? 33—2 


500 J. W. McLEOD AND J. GORDON 


In trying to obtain evidence of the production of these substances by the 
Pneumococcus in fluid media, the Staphylococcus was used as a test micro- 
organism because it grows freely and with great constancy in all ordinary 
bacteriological media and because it had proved to be particularly sensitive 
to the inhibitory body produced by the Pneumococcus. 

The experiment consisted in preparing a series of dilutions in peptone 
water of a 10 % serum bouillon culture of Pneumococcus, which had been 
heated for 30 minutes at 60° to kill off the Pneumococcus, and inoculating 
these with Staphylococcus. The inhibitory potency of the pneumococcal 
culture was judged by the extent to-which it had to be diluted i in order to 
permit of growth of the Staphylococcus. 

By using such methods it was soon determined that aeration of the culture 
was of importance; thus little or no inhibitory effect would be produced by 

‘growing a Pneumococcus in 10 cc. of 10 % serum broth in a narrow test-tube 
and the Pneumococcus was always alive at the end of 48 hours, whereas 
the same organism grown in 10 cc. of the same medium spread out over the 
base of an Erlenmeyer flask gave rise to a markedly inhibitory culture in 
which the Pneumococcus had usually died at the end of 48 hours. It was 
found further that inhibitory cultures could be obtained more constantly if 
a current of air was kept bubbling through the culture during incubation, by 
connecting the flask containing the medium with a water suction pump. 
Control observations in which air was passed through sterile media kept for 
similar periods of time in the incubator gave negative results—the medium 
did not develop any inhibitory quality. 

All the earlier observations were made with a bouillon medium prepared 
from meat extract and Parke Davis peptone, 1 %, to which 10 % of horse serum 
was added. The horse serum had been heated for an hour at 56° and was several 
months old. It contained very little catalase and subsequent observations 
have shown that it is important for constant production of the inhibitory 
body that the medium should contain little or no catalase. Some of our 
observations also tended to show that the amount of inhibitory substance 
produced was influenced by the kind of peptone used but the point has not 
yet been investigated carefully enough to permit of a definite statement. 


Methods of concentrating the inhibitory substance. 


Under the impression that the substance under investigation was probably 
a labile protein similar to toxin an attempt was made to obtain concentration 
by precipitating out the proteins of the culture with absolute alcohol at 0°, 
and redissolving the precipitate in a small quantity of water. The solutions 
so obtained showed little or no power of inhibiting bacterial growth. Further 
experiment showed, however, that the bactericidal substance dissolved in 
alcohol and passed into the filtrate. By concentrating such filtrates in vacuo 
at 35° 45° fluids of considerably enhanced bactericidal potency were obtained. 


PRODUCTION OF HYDROGEN PEROXIDE BY BACTERIA 501 


The average culture before concentration had an antiseptic potency suffi- 
cient to prevent development of Staphylococcus aureus in peptone water con- 
taining from | part in 3 to | part in 8 of pneumococcal culture, and by con- 
centrating this culture to one-sixth or one-eighth of its original volume in the 
manner outlined above a residue was obtained at least four times as active as 
the original culture, 7.e. capable of completely inhibiting growth of Staphylo- 
coccus in peptone water if present in the proportion of | part to 20 of peptone 
water. That is to say the fluid obtained was about equal to 4 % carbolic acid 
as regards its antiseptic effect on Staphylococcus aureus: it differed from the 
latter however in the marked instability of the antiseptic substance contained. 

By reprecipitating the first concentrate with an excess of alcohol and 
reducing to a considerably smaller bulk it was possible to obtain fluids of 
higher antiseptic potency. ~ 


Chemical evidence of the presence of peroxide in the inhibitory 
cultures and concentrates. 


The antiseptic concentrates from pneumococcal cultures had not been 
long under investigation before the observation was made that the addition 
of such fluids to emulsions of blood was followed by evolution of gas. We are 
indebted to Mr H. D. Kay of the Physiology Department, Leeds University, 
for suggesting that the effervescence was probably due to the presence of 
peroxide. This appears to be the correct explanation. Both cultures and 
concentrates give a strong blue colour when mixed with starch iodide paste 
containing traces of FeSO, (Schénbein’s reagent). Wolff [1912] has pointed 
out that under certain conditions nitrites may give this reaction and cause 
confusion in regard to presence of peroxides in plant juices; but the absence 
of nitrites from the cultures is proved by failure to obtain any colour reaction 
on adding diphenylamine. Cultures so concentrated also give quite definitely 
a yellow or orange colour with titanium sulphate solution. Lastly a compound 
blue in colour and soluble in ether but disappearing rapidly if not extracted 
with ether immediately was obtained on adding 10 % chromic acid to a 
concentrate. 

The chief evidences from the chemical standpoint for considering this 
substance to be hydrogen peroxide rather than some organic peroxide are: 

(1) Its sensitiveness to decomposition by the catalase of blood and serum 
or by special catalase preparations from the liver [Morgulis, 1921]; Novy and 
Freer [1902] state that the various organic peroxides with which they worked 
differed from H,O, in being relatively insensitive to the action of catalase. 

(2) The fact that it tends to decompose on heating, but only does so very 
rapidly when the temperature is raised to 85°, in which respect it closely 
resembles H,O,, which distils at 84° under reduced pressure when's latively © 
pure and concentrated but is rapidly destroyed about the same temperature 
in dilute and impure solutions [ Baur, 1908]. 


502 J. W. McLEOD AND J. GORDON 


(3) The gas evolved on adding catalase to a concentrate was proved 
to be oxygen since it was not absorbed by KOH solution but was absorbed 
to the extent of 90-95 % by alkaline pyrogallol solution and was cap- 
able of rekindling a glowing splint placed in it. Further the oxygen evolved 
corresponded very nearly to the amount calculated as available when the 
concentrate was titrated with potassium iodide, sodium thiosulphate and 
starch with a view to determining its H,O, content. 


Bacteriological evidence of identity of inhibitory substance in 
pneumococcal cultures with H,O,. 


Two lines of investigation have been pursued. 

First of all a number of bacteria have been compared as regards their 
_ varying sensitiveness to the inhibitory effect of (a) small concentrations of 
H,0, in agar plates (b) the substances diffusing from deep plate cultures of 
Pneumococcus to superimposed layers of agar [McLeod and Govenlock, 1921}. 
It has been found that if bacteria are graded according to the greatest concen- 
tration of H,O, in an agar plate which is compatible with their growth the 
order of their resistance to the H,O, is as follows: B. subtilis, B. prodigiosus, 
some Streptococci, B. coli, other Streptococci, Staphylococcus aureus, Anthrax, 
Cholera, Shiga, Typhoid. 

Among the bacteria which were observed often enough on inhibitory 
pneumococcal plates to give a reliable average result the order of resistances 
was: Streptococcus, B. coli, Staphylococci and Shiga. Although complicating 
factors such as favouring effect of other products of pneumococcal growth on 
certain bacteria may be present, the results are remarkably alike. 

The second line was to determine how far the antiseptic effect of the con- 
centrate of a pneumococcal culture corresponded to that of a dilute solution 
of H,O, of similar strength, i.e. one which would give a similar figure for 
available oxygen when that was determined by titration with potassium 
iodide and thiosulphate solution in presence of starch. Four experiments of 
this kind were performed and the results are given in Table I. 


Table I. 


Antiseptic potencies expressed as highest dilution in peptone water 
which inhibited growth of Staphylococcus aureus completely 


11,0, of concentrated 3 {?) Solution of () Concentrated 3 
culture calculated 2 in saline. culture inactivated (d) Concentrated 
in “ volumes” by trength as by heat: +-H,O culture 
KILN m80. (a) Concentrated calculated for of same strength inactivated 

Experiment + starch titration culture concentrate as in (b) by heat 
1 O05 1-34 1-34 1-24 — 
2 0-8 1-24 1-50 1-50 _ 
3 0-4 1-12 1-32 1-18 —- 
4 0-38 1-25 1-50 1-25 nil 


It will be seen that the figures for antiseptic potency are nearly identical 
for concentrates and for heat inactivated concentrates which had received 


PRODUCTION OF HYDROGEN PEROXIDE BY BACTERIA 503 


addition of H,O, sufficient to restore them to the same titration figure for 
available oxygen as that of the concentrates. 

Further in the fourth experiment, where figures for concentrate and heat 
inactivated concentrate + calculated amount of H,O, exactly correspond, a 
determination of volumes of gas liberated by catalase from 2 cc. of H,O, of 
calculated strength was 0-7 cc. and that from 2 cc. of concentrate 0-67 cc. 

That the antiseptic potencies of the concentrates should be usually less 
than those of solutions in 0-85 % saline of H,O, of similar strengths is not 
surprising since in the strong solution of amino-acids etc. and suspended fatty 
bodies which constituted the concentrate the tendency to decomposition of 
the H,O, is likely to be greater. 


Changes in blood pigment effected by growth of Pneumococcus. 


It has long been recognised that certain streptococci produce a green 
colour when grown on a blood agar plate, hence the name “viridans” 
[Schottmiiller, 1903]. The Pneumococcus resembles this micro-organism in 
the appearance of its growth on blood agar plates. It has also been recognised 
for some time that pneumococci and certain streptococci produce methaemo- 
globin [Stadie, 1921]. 

Methaemoglobin is not however a green pigment and so far as we know 
no full investigation of the pigmentary changes associated with the develop- 
ment of the green colour has been made. 

In recent years a medium called “chocolate agar” [Crowe, 1915, 1921; 
Neurin and Gurley, 1921], consisting of agar and heated blood mixed in 
varying proportions, has been much used. The blood here is no longer 
present in the form of haemoglobin or methaemoglobin and on such media 
the pigmentary changes produced by pneumococci etc. are much more 
striking than on unheated blood media. All gradations in colour between 
dark olive green and light yellow may be seen around pneumococcal colonies. 
If, however, small drops of dilute solutions of H,O,—+} to 1 “vol.’”—are re- 
peatedly applied to the surface of such media a very similar range of colours 
is obtained. If the application of H,O, is sufficiently powerful the medium is 
completely bleached. It would seem therefore that the varying colour changes 
which develop around pneumococcal colonies on such media are due to H,0, 
formation in varying degrees of intensity. Also that the heated blood agar 
plate may reasonably be used to detect bacteria capable of forming H,0Q,. 


Production of H,O, by bacteria other than the Pnewmococcus. 


The following conclusions have been reached by using the “chocolate agar” 
_ plate as a primary indicator of H,O, formation and then proceeding to confirm 
the fact that peroxide is produced by the bacteria which form green colonies, 
by growing them in bouillon or serum bouillon media in presence of abundant 
oxygen supply and testing their cultures with hydrogen peroxide reagents 


504 J. W. McLEOD AND J. GORDON 


such as Schénbein’s or titanium sulphate solution. No bacteria have been 
met in the course of routine work or on going through a large series of stock 
cultures which produce a green colour on “chocolate agar,’ with the exception 
of various forms of cocci. Amongst the streptococci green-forming strains 
predominate but both amongst the haemolytic and the non-haemolytic forms 
strains are met which do not produce green on chocolate agar, and all such 
strains which have been examined have also failed to produce fluid cultures 
giving H,O, reactions. 

As a general rule, although with some distinct exceptions, the streptococci 
are differentiated from the pneumococci by producing a green colour later 
and with lesser intensity, similarly the presence of HO, is usually detected 
with the appropriate reagents at a later period in streptococcal than in pneu- 
mococcal culture, 7.e. after 36 hours rather than after 18 hours. 

In addition to the streptococci, bacteria from the urethra of sarcinal type 
and a coarse coccus unclassified have been found to produce a green colora- 
tion on “chocolate agar” and also to give H,O, reactions in fluid media. 

_ Amongst the many bacteria which do not produce green coloration on 
“chocolate agar”’ plates three were also tested for production of H,O, in flasks 
of fluid media in which an air current was maintained. These were B. coli, 
Paratyphoid B. and Staphylococcus. All gave negative results and a number 
of different bacteria occurring from time to time as contaminations in such 
experiments have also given negative results. 


Importance of catalase in connection with these phenomena. 


It is an old finding in bacteriological work that most bacteria produce 
catalase [Gottstein, 1893]. 

If the production of H,O, by bacteria as described above is a fact, certain 
observations may be reasonably expected to follow; amongst these are 

(1) that bacteria producing H,O, will not produce catalase; 

(2) that such bacteria will grow much better in media containing catalase ; 

(3) that a medium containing some substance producing the catalase 
effect will be the best for maintaining stock cultures of such bacteria. 

These observations have all been made; no evolution of oxygen over 
Pneumococcus colonies has ever been observed on pouring dilute solutions of 
H,O, over cultures of these bacteria. The same holds for most streptococci. 
A more marked turbidity is developed in a serum bouillon culture of Pneumo- 
coccus than in a parallel culture where the serum bouillon has been heated 
to 65° for 30 minutes so as to destroy the catalase; that is if the cultures are 
incubated under conditions allowing of moderate access of oxygen. In fact 
it is very probable that the “vitamin” effects of fresh tissue fluids in pro- 
moting growth of Pneumococecus which Kligler [1919] describes are partly 
due to catalase, 

Lastly it has been found that peptone bouillon containing 10 °% of washed 
blood, a medium which retains to a considerable extent the power of de- 


PRODUCTION OF HYDROGEN PEROXIDE BY BACTERIA 505 


composing H,O, even after heating for 20 minutes at 115°, is the most con- 
venient for maintaining stock cultures of the Pneumococcus; no difficulty 
has been met in preserving various strains over long periods when subcultures 
have been made at 3-4 week intervals. 


Possible importance of H,0, formation by bacteria. 

The most interesting possibility that arises in this connection is that H,0,- 
forming bacteria may under certain circumstances tend to kill themselves in 
the body by peroxide formation, just as they do in the test-tube. The most 
concrete case in which this possibility arises is the crisis in pneumonia. In 
the explanation of this there is admittedly a good deal of difficulty at the 
present moment, since the observed phenomena of phagocytosis, development 
of bactericidal quality in the serum, antitoxin formation etc. are inadequate. 
The chief difficulty in accepting such a theory of the crisis in pneumonia is 
that the pneumonic exudate is rich in catalase and it is most unlikely that 
any concentration of H,O, similar to that which kills off the pneumococcus 
in culture can occur. There are two possible ways out of this difficulty. One 
is to suppose that something occurs to paralyse the catalase of the pneumonic 
exudate. Changes of reaction and enzyme reactions have been demonstrated 
in the pneumonic exudate which do not occur in the healthy tissues of the 
body [Lord, 1919, 1,2; Lord and Nye, 1921, 1,2]. Another is to suppose 
that a much smaller concentration of H,O, may be effective in killing off the 
Pneumococcus in the lung than is required in cultural experiment on account 
of the oxidising ferments present in the exudate. Both however are subjects 
for extended experimental investigation. 


Explanation of the formation of H,O, by bacteria. 

Wieland’s [1912, 1,2; 1913] theory of oxidation supposes that the essential 
phenomenon is the liberation of hydrogen and that the réle of oxygen is that 
of an acceptor for the hydrogen liberated, while Wartenberg and Sieg [1920] 
show that, under certain conditions at all events, the first stage in the union 
of Hand Ois H,O,. Such a sequence of events would fit in well enough with the 
observed phenomena in pneumococcal cultures in which H,O, is only formed 
as a bye-product where there is sufficient access of oxygen and little or no 
catalase. 

The reason why H,O, does not appear in the cultures of other bacteria 
may be either that they are catalase formers—the great majority of known 
bacteria; or that although not catalase formers they are too sensitive to the 
antiseptic action of low concentrations of H,O, to grow sufficiently in the 
presence of oxygen to produce any recognisable traces of that substance— 
the anaerobes. When however certain strains of streptococci are considered 
which we have found to be incapable of producing catalase and relatively 
insensitive to H,O,, but which do not produce H,0, in their cultures, it is ob- 
vious that other factors which await investigation must enter into the problem. 


506 J. W. McLEOD AND J. GORDON 


CoNCLUSIONS AND SUMMARY. 


1. The inhibitory substance developed in pneumococcal cultures to which 
there is abundant access of oxygen is H,Q,. 

2. This is proved both by chemical reactions and by the comparison of 
the antiseptic effects of the pneumococcal cultures and those of dilute solutions 
of H,O, reckoned by titration etc. to contain a similar amount of available 
oxygen. 

3. The early death of Pneumococcus in culture is usually due to accumu- 
lation of excess of H,Q,. 

4. The green or yellow colorations produced on heated blood media by 
certain bacteria are due to H,O, formation. 

5. In addition to pneumococci the only bacteria which have been shown 
to produce H,O, are many streptococci, both haemolytic and non-haemolytic, 
and a few other coccal forms. 

6. These findings can be utilised practically in putting bacteriological 
technique on a more definitely scientific basis in several respects. 

7. The substance in fresh tissue fluids which specially promotes the growth 
of the Pneumococcus and which has been supposed to be of the nature of 
vitamin is most probably catalase. 


In conclusion we have pleasure in expressing our indebtedness to Prof. 
M. J. Stewart in whose department this work was carried out and to Prof. 
H. 8. Raper of Leeds University for valuable advice. 


REFERENCES 


Baur (1908). Abegg’s Handb. Anorg. Chem. 2, 1te Abth. 87. 

Crowe (1915). Lancet, ii, 1127. 

(1921). J. Path. Bact. 24, 361. 

Eijkman (1904). Centralbl. Bakt. Par. O.1. 87, 436. 
—— (1906, 1). Centralbl. Bakt. Par. O.1. 44, 367. 
—— (1906, 2). Centralbl. Bakt. Par. O.1. 44, 471. 

Gottstein (1893). Virchow’s Archiv, 188, 295. 

Kligler (1919). J. Hap. Med. 30, 31. 

Lord (1919, 1). J. Amer. Med. Assoc. 72, 1364. 
—— (1919, 2). J. Amer. Med, Assoc, 78, 1420. 

Lord and Nye (1921, 1). J. Hap. Med. 34, 199. 
—— (1921, 2). J. Hap. Med. 34, 201. 

McLeod and Govenlock (1921). Lancet, i, 900. 

McLeod and Gordon (1922). J. Path. Bact. 25, 139. 

Morgulis (1921). J. Biol. Chem, 47, 341. 

Neurin and Gurley (1921). J. Immunology, 6, 5. 

Novy and Freer (1902). Amer. Chem. J. 27, 161. 

Schottmiiller (1903), Miinchener Med. Woch. 50, 849. 

Stadie (1921). J. Lap. Med. 38, 627. 

Wartenberg and Sieg (1920). Ber. 58, 2192. 

Wieland (1912, 1). Ber, 45, 484. 

(1912, 2). Ber. 45, 2606. 

(1913). Ber. 46, 3327. 

Wolff (1912). Ann. Inst, Past, 27, 501. 


XLV. NOTE ON URINARY TIDES 
AND EXCRETORY RHYTHM. 


By JAMES ARGYLL CAMPBELL anp 
THOMAS ARTHUR WEBSTER. 


From the Department of Applied Physiology, National Institute 
for Medical Research. 


(Received May 10th, 1922.) 


INTRODUCTION. 


In previous papers [1921, 1922] we gave the results of observations on the 
composition of the day and night urine of the same subject under four different 
routines, each of five days duration. We showed that the differences between 
day and night urine were constant under a routine of complete rest in bed, a 
laboratory routine, a light muscular work routine and a severe muscular work 
routine. We found, in each case, that the total N, urea N, water and chloride 
were excreted in much greater amounts during the day than at night; that 
the amino-acids and uric acid were excreted in slightly greater amounts during 
the day than at night; that the acid, ammonia N and phosphate were 
excreted in much greater amounts during the night than during the day; 
and that creatinine and sulphate were mort or less evenly distributed between 
day and night. In other words, there was an excretory rhythm which was 
not altered by any of the conditions of the four routines examined. In this 
rhythm there were definite tides during the day of water, chloride, total N 
and urea N, whilst at night there were phosphate, ammonia N and acid 
tides. It was considered that the greater excretion of acid, ammonia N and 
phosphate at night was due to delayed excretion of “certain fixed acids” 
formed in the cells during the day and that the phosphate tide at night was 
connected with the greater acidity at night. 

To investigate further these various tides and this excretory rhythm we 
employed the same subject under further changes of routine. In the present 
paper we refer to differences between day and night urine, (1) during a routine 
of day starvation for 48 hours with food only at night, (2) during a routine 
of complete starvation for 24 hours and (3) during a reversed routine, for 
96 hours, in which work was performed and food was taken during the night 
whilst the subject slept during the day. 


508 J. A. CAMPBELL AND T. A. WEBSTER 


METHODS. 


We have kept the methods constant throughout all routines. 

Our subject took his usual diet but it was controlled to exclude articles 
which are known to influence greatly the composition of the urine and also 
to keep the average daily diet for each routine as similar in substance as 
possible. We.employed his usual diet as we were undertaking a prolonged 
research in which the subject would at times be exposed to somewhat strenuous 
conditions. By making each routine cover several—and the same—consecutive 
days and by keeping the quality of the average daily diet constant, we thought 
that errors would be lessened. Moreover, preliminary observations under his 
normal daily diet showed that the differences between day and night urines 
were constant. 

We collected the urine between 7 a.m. and 5 p.m. for the day and between 

5 p.m. and 7 a.m. for the night. It was not considered that these periods were 
the best. They were chosen as being most convenient for our research. For 
a large part of this research the subject lived at the laboratory. 

There is no accurate way to decide whether a waste product formed in a cell 
is excreted as soon as it is formed by the cell, or for how long the cell retains _ 
it. All periods for collecting urine have their advantages and disadvantages. 
The best results will probably be attained by a comparison of figures obtained 
during various periods of collection from each of many subjects under various 
conditions. 

In estimating the acidity of the urine, we employed four methods—py, 
Folin’s titration method, titration to py 7-4, and Leathes’ acidity percentage 
method [1919]. As other observers have pointed out we found that the 
agreement between these methods as regards detail was not marked, but 
nevertheless the broad significance of the conclusion as regards acidity was 
similar with each method. Thus we were able to demonstrate the presence of 
the so-called alkaline tide in our subjects and to determine the difference in 
- acidity between the day and night urines by any of these methods. 


Starvation Routines. 


Our subject starved all day taking food only between 5 p.m. and 10 p.m., 
the object of this experiment being to determine whether taking food at night 
would alter any of the tides, particularly the acidity and total N tides. For 
two days this routine was followed, and on the next day the subject starved 
for 24 hours. Table I shows the results for both these routines, namely, 
“day starvation” and “complete starvation.’’ On comparing these results 
with our previous work [1921, 1922] it will be noticed that practically no 
change from our previous figures occurred as regards the differences between 
day and night urine. That is total N, urea N, water and chloride were still 
excreted in greater quantities during the day than at night, and acidity, 
phosphate and ammonia N were still higher at night than during the day. 


URINARY TIDES 509 


Lusk [1917] records that the total N excretion during starvation was higher 
during the day than at night. From our results during starvation, it was 
obvious that the urine, under normal conditions, was not more acid during 
the night than during the day because taking of food during the day produced 
so-called alkaline tides. We have carried out many observations on our 
subject with regard to the so-called alkaline tide and found that any normal 
meal was followed by such a tide, in many cases the tide being well marked. 

During the “day starvation” routine, our subject carried out sedentary 
laboratory work, whilst during the “complete starvation” routine he rested in 
bed reading most of the day. 

We found about 20 mg. of creatinine N per hour in our subject, as seen 
in Table I. In our previous papers [1921, 1922] owing to an error we gave 
about 30 mg. per hour as the figure for this subject; all the figures in these 
previous papers for creatinine N should be reduced by one-third and the 
figures for undetermined N correspondingly increased. The general conclu- 
sions, however, are not disturbed by this correction, as the error was constant. 


Reversed Routine. 


In the “reversed routine” our subject worked and had most meals and 
fluid at night whilst he slept during the day from about 10.30 a.m. to 5.30 p.m. 
Meals were taken at 8 p.m., 2a.m., 5 a.m. and 8 a.m. instead of the usual 
times 7.30 a.m., 1 p.m., 4 p.m. and 7.30 p.m. respectively. Four consecutive 
days of this “reversed routine” were spent by our subject C.P. and one of us 
(J.A.C.). Neither subject had become accustomed to the new routine by the 
fourth day; although both subjects had slept well during the day they felt 
very sleepy at night. Table I shows C.P.’s average results for the last 72 hours 
- of this routine and Table IIa gives J.A.C.’s average results. Average figures 
are given because the results for any one day were very similar to these. 
Subject C.P.’s results show that no change from the normal rhythm was 
obtained except that the phosphate for the day was much increased, the 
night phosphate remaining about the usual figure. 

In the case of J.A.C. a similar increase of phosphate during the day was 
noted but also there were other changes; now the acidity and ammonia N 
were higher during the day when sleep was taken, so that the reversal of routine 
produced some reversal in composition in J.A.C.’s urine (see Table II a). 

It is interesting to note that after four days of complete reversal of routine 
and meals (including fluid) the day and night partition of water and nitrogen 
and of practically all other excretions remained unchanged. This points to- 
wards a fixed physiological rhythm of the cells which cannot be altered easily. 
It is not likely that the kidney was responsible for this rhythm. All the body 
cells were probably responsible. The rhythm may be due to following closely 
a fixed daily routine—of work, meals and sleep—for a long period. Allowance 
for this phenomenon must, therefore, be made where necessary when carrying 
out experiments of this nature. 


510 
7 
v4 
48 hours’ day starvation (food 
between 5 p.m. and 10 p.m.) 
Day Nicut 
(Sedentary (Food and 
Fess no food) sleep) 
Amount 08. ot - 56-0 25-0 
Acidity % . 46-4 78-6 
Titratab le acidity (Folin) 
ec. N/10 6-0 15-7 
Total acidity, ce. NAO. 15-0 30-7 
Total N g. $3 -379 (100)* = -307 (100)* 
Urea N g. -315 (83-00) +248 (80-80) 
Ammonia N (A). g. -010 (2-64) -019 (6-19) 
Ammonia N (B) g. -012 -021 
Amino-acid N a 002 (0-52) -002 (0-65) 
Creatinine N g. 020 (5-28) -019 (6-19) 
Uric acid Ng. ... 006 (1-58) 004 (1-30) 
Undetermined N g. -026 (6-98) 015 (4-87) 
Chloride (NaCl) g. 472 213 
chi apes (P,0;) g. -068 082 
Total S (SO,) g. 055 (100) -068 (100) 
Inorganic § (SO,) g. 039 (70-9) +052 (76-5) 
Ethereal 8 (SO,) g. 006 (10-9) 009 (13-2) 
Neutral § (SO,) g. -010 (18-2) 008 (10-3) 
0026 


Purine N g- oS ee 
Calcium (CaO) g. abe _ 


J. A. CAMPBELL AND T. A. WEBSTER 
Table I. Average hourly results. Sulyect C.P. 


24 hours’ complete 


88-0 
33-0 


6-0 
14:0 


starvation 
Day Nicut 
(Rest in 
bed) (Sleep) 

17-9 

58:3 

10-0 

24:3 
+392 (100)* —-195 (100)* 
+305 (77°80) -142 (72-80) 
‘010 (2-55) ~—--018 (9-23) 
‘O11 -020 
‘001 (0-25)  -002 (1-02) 
-019 (4:85) -020 (10-25) 
006 (1-53) = -003 (1-54) 
-050 (13-02) -010 (5-16) 
-497 -134 
“068 -078 
-056 (100) -041 (100) 
‘044 (78-6)  -035 (85-4) 
005 (8-9) “O01 (2-4) 
‘007 (12:5) = -005 (12-2) 


— 


72 hours’ reversed 


routine 
Day NIGHT 
~ (Work and 
(Sleep) food) 
71:3 34:5 
42-4 70-4 
8-0 12-4 
17:3 29°5 
-373 (100)* +337 (100)* 
+313 (84:00) -250 (74-20) 
‘010 (2:68) —-018 (5:34) 
‘013 024 
-003 (0-80) —-006 (1-78) 
-021 (5-63) +022 (6-53) 
-006 (1-61) ~ -007 (2-07) 
020 (5°28) — -036 (10-08) 
+335 261 
“093 081 
-063 (100) -072 (100) 
*054 (85-7) 057 (79-2) 
-002 (3-2) -006 (8-3) 
-007 (11-1) = -009 (12-5) 
‘O11 “014 


(A) Van Slyke’s method. (B) Malfatti’s method. Total acidity is ammonia (B) + titratable acidity. 
* Figures in brackets are percentages. 


Table Il a. Hourly averages. Sulyject J.A.C. 


Reversed routine Ordinary routine 
’ cme a ee | ee: * My 
Night Day Day Night 
Work and food Sleep Work and food Sleep 
Amount ce. 43-1 76-0 60-7 31-5 
Acidity % ese 53-8 57:3 42:8 72:3 
Titratable > acidity, Folin cc, N flo = 11-2 13-6 8-2 10:7 
Total acidity cc. NV ly dee « 868 35°3 26-5 30:6 
Total N g. -408 596 +533 451 
Ammonia N (B) g. 025 ‘030 026 ‘028 
Phosphate (P,0;) g. 074 097 “055 075 


Table Il 8. Hourly averages. Subject C.P. 


Titratable Total 
acidit: acidit 


Experi- Acidity Phosphate Ammonia Total N 
ment no. %y (Py0,) Bg. cc,N/l co. N/lO N(A)g. g. Time Sleep Routine 
1 20°6 045 54 18:8 ‘O14 386 day - : 
2 702-080-182 367-022-206 night | Rest in bea 
3 448 065 10-4 54 O15 428 ay ~ ' 
4 646 080 168 330 019 +332 night + Ordinary 
5 50-0 066 11-6 28-1 ‘O11 ‘447 = day - Light muscular 
6 64-0 O75 15-4 31-6 O19 ‘357 ~—s night + work 
7 60-0 ‘070 14-0 20-0 O15 446 = day - Severe muscular 
8 pk 080 14-4 30-7 019 414 — + work 
9 ‘4 068 6-0 15-0 O10 ‘379 ay - : 
10 «786 = 082—ss«s1G-7S—s80-7—=—«—19—s—‘— OT~Ssinight + Day starvation 
il 33-3 068 6-0 140 010 B92 day “a Complete 
12 58-3 O78 10-0 24:3 ‘018 195 ~—s night + starvation 
13 42-4 093 8-0 17-3 ‘010 373 day ay Reversed 
14 704 081 124 20°5 018 ‘837s night - 


URINARY TIDES 511 


In both subjects the phosphate for the day and therefore for the whole 
24 hours was distinctly increased. This increase was probably due to some 
special nervous metabolism connected either with sleep itself or with the 
desire to sleep. In Table II 8, we have drawn up some results from all our 
experiments including those previously published [1921, 1922]. These show 
that a high phosphate excretion always accompanied sleep and that the higher 
of the day and night figures for phosphate excretion in each routine could be 
separated from every other factor in our experiments but sleep (see Table II B). 


SLEEP AND AcIpITY oF URINE. 


In our previous papers we found that sleep was accompanied by a definite 
increase of ammonia, acidity and phosphate and a relative increase of sul- 
phate. We suggested that there was a delayed excretion of “certain fixed 
acids” formed in the cells during the day, and that when formed in certain 
quantities, they were responsible for sleep and fatigue. In our present paper 
the higher figures for phosphate whether found by day or by night always 
accompanied sleep and sleep was independent of every other factor including 
acidity (see Table IIB). Therefore, although these “certain fixed acids” 
might have been responsible for sleep and although higher acidity usually 
accompanied sleep it was not a necessary accompaniment of sleep (see experi- 
ment 13, Table Il B). This is contrary to the finding of Leathes [1919], who 
considered that the CO, tension of alveolar air was increased at night above 
that for the day, as a result of depression of the respiratory centre during sleep. 
Collip [1920] also considered that an increase of Cy of the blood at night was 
due to sleep. 

We see no reason to abandon the suggestion that sleep might be due to 
“certain fixed acids” since there was in the case of subject C.P. in the 
“reversed routine” an increased acidity of the urine previous to sleep although 
not accompanying it. He was obviously sleepy during the time of excretion 
of the larger amount of acid, but kept awake purposely. If the high acidity 
was due to “certain fixed acids” formed by the activity of the cells and capable 
of causing sleep, the cells would only recover after sleep itself and not after 
excretion of the “certain fixed acids.” However, it was obvious that the 
subject had not become accustomed to the “reversed routine,” so that our 
results really belonged to a transitional period and it is at present difficult to 
draw conclusions. We have carried out some observations on a subject who 
was somewhat more accustomed to work and food at night, with sleep during 
the day. He worked on a night shift every third week. The results we have 
obtained from him are in complete agreement with the results here described. 
Results from subjects accustomed to long periods of night work are required. 


PHOSPHATE EXCRETION. 


We conclude from all our results regarding the higher figure for day and 
night phosphate excretion in each routine that it was intimately related to 


512 J. A. CAMPBELL AND T. A. WEBSTER 


metabolism either connected with sleep or occurring at the same time as sleep. 
We were able to separate the higher phosphate excretion in each routine from 
every factor, but sleep (see Table IIB). Although, as a rule, the higher 
phosphate excretion accompanied the higher acid and the higher ammonia N 
excretion there was a marked exception to this rule (see experiment 13, 
Table II 8). 

Broadhurst and Leathes [1920] have suggested that the phosphate tide 
at night may be connected with some special muscular or nervous metabolism. 
Our previous results indicated that there was no connection with muscular 
metabolism and less clearly that there was none with nervous metabolism. 
Broadhurst and Leathes also found that the phosphate tide was not dependent 
upon food. Our results during the starvation routines confirm this finding. 
Fiske [1921], who considered that the phosphate tide was due in part to 
‘retention of phosphorus,” also showed that the phosphate curve was not 
affected by the taking of sodium bicarbonate so that the phosphate curve 
was not connected with excretion of acid. This may be the case, but there is 
much evidence from other observers that phosphate may have, as one of its 
functions, the removal of acid. Haldane [1921] found that acid was removed 
by phosphate if there was any phosphate available. In our subject C.P. a high 
acidity, as interpreted by a high titratable acidity and a high ammonia N 
excretion, was always accompanied by a high phosphate excretion. 

We estimated the calcium in the urine and found that it was not connected 
with the increase of phosphate, the calcium being low when the phosphate 
was high and vice versa. When in the “‘reversed routine” the phosphate was 
absolutely increased, no such marked change was noted in the amount of 
calcium excreted. 

According to text books [ Hawk, 1919], some investigators hold that during 
extensive decomposition of nervous tissue the phosphate is increased. Mendel 
found that phosphate was increased after sleep produced by potassium 
bromide or chloral hydrate, so that it is possible that special metabolism 
connected with sleep was the main factor for the higher of the day and night 
phosphate excretions in our subject. Nervous metabolism may have been 
concerned. 

It is interesting to note that in children we have found no such correla- 
tion; in them the phosphate followed closely the total N. In children, the 
condition is rendered more complex by the presence of growth metabolism 
together with the maintenance metabolism. Children did not show the same 
rhythm, probably because of different conditions with regard to sleep. 


SUMMARY. 


1. Observations on day and night urine during a routine of day starva- 
tion, a routine of complete starvation and a reversed routine—that is with 
work and meals at night and with sleep during the day—are recorded. — 


URINARY TIDES 513 


2. Evidence was obtained that the urine was more acid at night, neither 
because the taking of food during the day produced alkaline tides, nor because 
the respiratory centre was depressed during sleep. 

3. After four days of complete reversal of habit most of the differences 
between day and night urine still remained as before, so that there must be 
a fixed physiological rhythm connected with excretion by the body cells. 
Thus, although in the “reversed routine,’ most of the fluid was swallowed at 
night, the greater amount of urine was still excreted during the day which 
now included the sleep period. Also, although most of the food taken was 
eaten at night, the total N excretion still remained higher during the day than 
at night. ; 

4. In the excretory rhythm referred to, the total N, urea N, water and 
chloride were excreted in greater amount during the day whilst the ammonia N, 
acidity and phosphate were higher during the night than during the day. 

5. The higher figures for phosphate whether found by day or by night 
always accompanied sleep and could be separated from every other factor 
except sleep; and, with one exception, the higher phosphate excretion ac- 
companied both the higher acidity and the higher ammonia N excretion in 
all routines examined. An absolute increase in phosphate occurred in the 
‘reversed routine.” 7 : 


REFERENCES. 


Broadhurst and Leathes (1920). J. Physiol. 54. Proc. xxviii. 
Campbell and Webster (1921). Biochem. J. 15, 660. 
—— (1922). Biochem. J. 16, 106. 
Collip (1920). J. Biol. Chem. 41, 473. 
Fiske (1921). J. Biol. Chem. 49, 171. 
Haldane (1921). J. Physiol. 55, 272. 
Hawk (1919). Practical Physiological Chemistry, p. 435. 
Leathes (1919). Brit. Med. J. ii, 165. 
Lusk (1917). The Science of Nutrition, p. 110. 


Bioch. xvi 34 


XLVI. NOTE ON A NEW TANNASE FROM 
ASPERGILLUS LUCHUENSIS, INUI. 


By MAXIMILIAN NIERENSTEIN. 


From the Biochemical Laboratory, Chemical Department, 
University of Bristol. 


(Received May 2nd, 1922.) 


PavuLuinia tannin’ is hydrolysed by emulsin, dextrose and f-gambier- 
- catechincarboxylic acid being produced [Nierenstein, 1922]. It was therefore 
surprising to find that tannase, which had been prepared by growing Asper- 
gillus Luchuensis, Inui in a solution of gallotannin and which is known to 
hydrolyse gallotannin into dextrose and gallic acid [Rhind and Smith, 1922], 
had no effect on paullinia tannin. This suggested the possibility that the . 
hydrolysing properties of tannase depended on the medium in which it had 
been produced by the fungus and that a catechutannin, such as paullinia tannin, 
required a tannase, which had been formed in a solution of a catechu- and not 
gallotannin. This has actually been found to be the case, since Aspergillus 
Luchuensis yields in a medium in which gallotannin had been replaced by 
catechutannin from cube-gambier (Owrowporia Gambier, Baill), a new tannase, 
which hydrolyses paullinia tannin, but which has no action on gallotannin. 
We have therefore to distinguish between these two kinds of tannase and it 
is proposed to refer to them as gallotannase and catechutannase respectively. 
In this connection reference must be made to what may in time prove a 
dangerous practice, which has recently been introduced by Freudenberg and 
Vollbrecht [1921]. These chemists prepare tannase by growing an unknown 
species of Aspergillus in the aqueous extract of ground myrobalams, the fruit 
of Terminalia chebula, Retz. Such a medium must obviously affect the general 
properties of the tannase and ought to be avoided. 

Freudenberg’s method [1920], in which gallotannin is replaced by an 
aqueous solution of the residues of cube-gambier, from which the catechin is 
removed by extraction with ethyl acetate, is used for the preparation of 
catechutannase. The catechutannase thus obtained resembles gallotannase in 


* This catechutannin is present in the seeds of Paullinia Cupana, H.B. and K., which is a 
synonym of P. sorbilis, Mart. These seeds are used for the preparation of the paste known as 
“pasta guarand” and the tannin is consequently also referred to as guarana tannin [compare, 
for example, Perkin and Everest, 1918}. Harvey [1921] and also Procter [1922] have wrongly 
assigned the name P. sorbilis to “ guara,”’ which contains a gallotannin and not a catechutannin, 
as“ guara” signifies Cupania americana, L. For this information I am indebted to the authorities 
of Kew Gardens, who also inform me that in Cuba the names “guarand” and “guara” are 
apparently used for Cupania macrophylla, A, Rich, and Guarea trichiliodes, L, respectively, 


A NEW TANNASE 515 


appearance and produces when added to a solution of paullinia tannin (0-5 g. 
catechutannase are used for 2 g. paullinia tannin in 150 cc. of water to which 
5 ec. of chloroform are added and the mixture kept in a dark incubator at 23° 
for ten days), a bulky precipitate, which when filtered off crystallises from 
water in small, pointed needles, which melt at 252-253°, carbon dioxide being 
evolved. These are in every respect identical with the inactive B-gambier- 
catechincarboxylic acid, previously obtained by the action of emulsin. 
(Found: C = 57:2; H = 4-6. Calculated: C = 57-5; H = 4-2 %.) For further 
identification a small quantity of the acid was methylated with diazomethane, 
when the corresponding pentamethoxy methyl ester was obtained. It crystal- 
lises from light petroleum in long needles, which melt at 74°. This melting 
point is not depressed on admixture with the same substance (m.p. 74°) 
previously obtained by the action of diazomethane on the acid. 

The filtrate from the crude B-gambier-catechincarboxylic acid, freed from 
unchanged paullinia tannin with the aid of lead acetate and hydrogen sulphide, 
contains dextrose, which is identified as the dextrosazone (m.p. and mixed 
m.p. 200—203°). 

In conclusion it is interesting to note that catechutannase produces the 
identical B-gambier-catechincarboxylic acid when added to the aqueous extract 
of fat-free cocoa-beans or to a solution of catechutannin from cube-gambier. 
This indicates a very close relationship between these tannins which has so 
far not been suspected. 

REFERENCES, 


Freudenberg (1920). Ber. 53, 958. 

Freudenberg and Vollbrecht (1921). Zeitsch. physiol. Chem. 116, 277. 
Harvey (1921). Tanning Materials, p. 23. 

Nierenstein (1922). J. Chem. Soc. 121, 23. 

Perkin and Everest (1918). The Natural Organic Colouring Matters, p. 442. 
Procter (1922). The Principles of Leather Manufacture, p. 325. 

Rhind and Smith (1922). Biochem. J. 16, 1. 


34—2 


XLVII. A QUALITATIVE TANNIN TEST. 


By ETHEL ATKINSON anp EDITH OLIVE HAZLETON. 


From the Biochemical Laboratory, Chemical Department, 
University of Bristol. 


(Received May 15th, 1922.) 


TANNING consists in the fixation of the tannins by animal fibre. A reliable 
tannin test must therefore demonstrate this specific property and it is re- 
markable that no such test has so far been devised. The reactions which are 
generally used for the identification of the tannins, such as the colorations 
produced by iron salts or the precipitations which are given by potassium 
dichromate and gelatin [compare, for example, Onslow, 1920], obviously only 
indicate some incidental properties of the tannins. These reactions are also 
shared by other organic substances, thus phenols and hydroxybenzoic acids 
are known to give colorations with iron salts, whilst both gallic acid and 
gallotannin are precipitated by potassium dichromate [Drabble and Nieren- 
stein, 1907]. Similarly gelatin is not only precipitated by tannins, but also 
by gum arabic [ Pelletier, 1813], starch [Tollens, 1914], inulin [Tollens, 1914], 
methyl gallate [Nierenstein, 1905, 1912; Fischer, 1919; Freudenberg, 1920] 
and other substances. The want of a specific tannin test was therefore felt 
for many years in this laboratory, with the result that the following method 
has been devised by us at the suggestion of Dr Nierenstein. 

A small piece of gold-beater’s skin’, about 4 inch long and ? inch wide, is 
pinned on a flat surface of paraffin-wax, which is prepared by pouring melted 
paraffin-wax into a watch-glass. The skin is covered with a few cc. of water, 
in which it is left soaking for five minutes so as to make it better permeable 
to the tannins. The water is then poured off and the skin tanned by covering 
it with 1 ec. of a tannin solution, which is prepared by extracting for half-an- 
hour on a boiling-water bath 1 g. of the material to be tested with 50 cc. of 
water. After many experiments we have come to the conclusion that half-an 
hour’s tanning suffices even for the weakest tannin solution, but that solutions 
of usual strength (about 1-2 °%) tan in a few minutes. It is, however, 
advisable to tan not less than 15 minutes. The tanned skin is washed for two 
minutes at a constant drip of two drops per second and then stained for five 


* Gold-beater’s skin is the outside membrane of the large intestine of the ox. It has many 
advantages over other animal tissues, It is easily obtainable (the material used by us was 
supplied by the British Drug Houses) and it is very thin (1 yard x 6 inches weighs only about 
3g.). It consequently tans rapidly and requires only little tannin, which should make this test 
of use not only for chemical, but also for botanical work, 


A QUALITATIVE TANNIN TEST 517 


minutes with | cc. of a 1 % solution of ferric chloride. It is then again washed, 
dried and mounted for reference. 

We find that only tanned gold-beater’s skin is stained by ferric chloride. By 
this method we have been able to demonstrate the presence of tannins in the 
following plant products!, all of which are known to contain tannins: 

Aleppo galls, Chinese galls, Chinese plum galls, blue Basra galls, white 
Basra galls, Knopper galls, Sumac, Mangrove bark, Mimosa bark, Myrobalams, 
Valonia, Oak bark, Quebracho wood, Hemlock bark, Divi-Divi, Algarobilla, 
Canaigre, Pistacia Lentiscus, Golden Wattle, Acacia Arabica, Larch bark, Tea, 
Coffee, and Cocoa. 

Gallotannin behaves exactly in the same manner, whereas gallic acid and 
pyrogallol, both of which are known not to possess tanning properties, are not 
fixed by gold-beater’s skin, which is consequently not stained by ferric chloride. 
Identical results were also given by the following substances, all of which are 
known not to possess tanning properties: 

Phenol, catechol, quinol, resorcinol, phloroglucinol, salicylic acid, proto- 
catechuic acid and f-resorcylic acid. 

The combination between the tannins and the gold-beater’s skin is of a 
very permanent character. This is evident from the fact that the tanned and 
subsequently stained gold-beater’s skin may be decolorised with dilute hydro- 
chloric acid and again re-stained with ferric chloride. This process of de- 
colorising and re-staining may be repeated several times without the slightest 
effect on the colour of the re-stained material. 


In conclusion we would like to suggest that this test be referred to as the 
“‘gold-beater’s skin test for tannins.” We also wish to thank Prof. McCandlish 
of the University of Leeds and Mr M. C. Lamb of the Leathersellers’ Technical 
College at London for the different tanning materials used in this investigation. 
Our thanks are also due to the Research Fund Committee of the Chemical 
Society for a grant allocated to Dr Nierenstein for his work on gallotannin, 
from which the expenses of the present investigation were defrayed. 


REFERENCES. 


Dekker (1913). Die Gerbstoffe. 

Drabble and Nierenstein (1907). Biochem. J. 2, 96. 

Fischer (1919). Ber. 52, 821. 

Freudenberg (1920). Ber. 53, 236. 

Nierenstein (1905). Collegium, p. 307. 

(1912). Ber, 45, 837. 

Onslow (1920). Practical Plant Biochemistry, p. 91. 

Pelletier (1813). Ann. Chim. 87, 106. 

Tollens (1914). Kurzes Handbuch der Kohlenhydrate, pp. 525 and 551. 


1 We refrain from giving the Latin names of the plants we have used, since they are well- 
known tanning materials, which can be found in any book which deals with them [compare for 
example, Dekker, 1913]. 


XLVIII. THE ORIGIN OF THE VITAMIN 4 
IN FISH OILS AND FISH LIVER OILS. 


By JACK CECIL DRUMMOND anv SYLVESTER SOLOMON ZILVA, 
WITH THE CO-OPERATION OF 
KATHARINE HOPE COWARD (Beit Memorial Research Fellow). 


From the Biochemical Laboratories Institute of Physiology, University College, 
London, and Biochemical Department, Laster Institute, London, 


(Received May 18th, 1922.) 


Durine the last two years the authors have been collaborating in an ex- 
haustive investigation of the nutritive value of the edible oils and fats with 
particular reference to the dietary unit referred to as vitamin A. Of the 
time spent in this work a very large proportion has been devoted to the study 
of the oils of marine origin and in particular the fish liver oils, chiefly because 
it has been shown by Zilva and Miura [1921] that these oils represent by far 
the most valuable sources of the vitamin of all the groups into which the 
edible oils and fats are divided. Our study of the liver oils has been naturally 
divided into several phases each concerned with a distinct aspect of the subject 
and the present communication deals with the question of the origin of the 
relatively enormous stores of vitamin A which are found in the livers of many 
species of fish. The results set out in this paper supplement some recorded 
by Hjort who at our suggestion has studied a number of marine organisms 
in connection with this enquiry [1922]. 

It is now generally agreed that the higher land animals derive their 
supplies of vitamin A either directly or indirectly from green plants and the 
probability of a similar relationship existing in the sea has been suggested 
by Jameson, Drummond and Coward [1922]. The fundamental dependence 
of all marine animal life on the marine flora, particularly the microscopic 
flora, has been recognised for many years past, and the studies of Jameson, 
Drummond and Coward have shown that a typical marine diatom grown in 
pure culture can synthesise relatively large quantities of vitamin A, whilst 
Coward and Drummond [1921] showed earlier that higher forms of marine 
algae, whilst far less potent sources of this factor than diatoms, may contain 
a concentration approximately as high as that of typical land plants such as 
cabbage. 

There appear to be no adequate grounds for doubting that the ultimate 
origin of the vitamin A in marine oils and liver oils is represented by the 


ORIGIN OF VITAMIN A IN FISH LIVER OILS 519 


marine plants possessing photocatalytic pigments, particularly those which 
are included in the term plankton. Our studies of fish eggs show us that the 
young fish normally begins its life with a considerable store of vitamin A 
derived from the yolk-sac. We have reasons for believing that there may be 
a transference of vitamin from the liver or tissues of the spawning fish to the 
eggs although as yet our experiments on this subject are incomplete. 

Apart from our experimental work the existence of such a transference 
would appear probable from our knowledge concerning the transport of 
vitamin from tissues to the eggs or milk of-certain terrestrial species. 

Probably the vitamin-A content of the yolk-sacs is sufficient for the re- 
quirements of the developing larvae for some time, but it is well known that 
the period between the stage at which the contents of the yolk-sac are nearly 
absorbed and that at which the young fish are entirely dependent on external 
food is a very critical one. 

The careful studies of Lebour [1918, 1919, 1, 2; 1920, 1921] and others 
have shown us that post larval fish feed very largely on copepods. Her work 
may for our purpose be summarised by quoting two paragraphs from one of 
her most valuable papers [1920]. 

“One finds that certain copepods and other entomostraca constitute by 
far the larger part of the food of nearly all the very young fish, and that 
usually each species of fish selects its own favourite food to which it keeps, 
indiscriminate feeding seldom or never taking place, and one can usually 
assign to each fish its own particular food.” “Few fish are vegetarians, and 
it is unusual for any but the youngest fish to eat diatoms.” “Very young 
herring or sprat and a few others often contain green remains which probably 
belong to some algae, and occasionally diatoms can be recognised in this but 
even before the yolk sac is absorbed the gut may contain larval molluscs and 
small crustaceans.” 

- But whilst there is little or no direct feeding of fish on diatoms these 
and other microscopic organisms form the diet of the copepods and larval 
molluses which play so important a réle in the nutrition of young fish. The 
relationship between these two groups of organisms has been traced by a 
number of investigators with some care, but it is only necessary here to call 
attention to the very rapid multiplication of the plant life of the sea which 
takes place in the spring and which is intimately connected, as Moore has 
shown [1921], with the increase in light intensity during that period. Some 
short time after the rapid rise in diatoms and other plant organisms there is 
an associated rise in the numbers of microscopic animals, particularly cope- 
pods, and the occurrence of these two related phenomena has a very important 
bearing on the survival and development of the young fish hatched out at that 
season [Hjort, 1914]. z 

By the kindness of Dr O. Borley of the Fisheries Laboratory, Lowestoft, 
we were enabled to study the vitamin-A content of a number of plankton 
samples collected during the spring of 1922. 


520 J. C. DRUMMOND AND §&. 8S. ZILVA 


The samples after being dredged, sifted through muslin and pressed, were 
preserved in an equal bulk of absolute alcohol and sent directly to us in 
London. On arrival they were evaporated to dryness in vacuo at 40° and 
ground well to mix the organisms thoroughly with the material extracted by the 
alcohol. The dry material was administered to rats in the usual manner for 
testing for vitamin A, and all four samples were found to be highly potent as 
sources of that factor when administered in daily supplements of 0-1 g. to 
the rats. . 

The four samples studied were representative of the staple food of a large 


majority of the small fish. The analysis of the samples for which we are 
indebted to Dr Borley is given below. 


Table I. Plankton samples. 


; 1 2 3 3 4 
Locality 52-2N.x3-8E. Sandetti Lightship Pas de Calais Pas de Calais 
51-15 N. x 1-52 E. 50-58 N. x 1-30 E. 50-58 N. x 1-30 E. 
Principal contents . 
(1) Copepods Temora Temora Temora 
longicornis* Pseudocalanus Pseudocalanus 
Pseudocalanus Calanus Calanus 
elongatus 
Calanus 
finmarchicus Acartia 
Acartia Centrophages 
longiremus typicus 
(2) Mysids Schistomysis Macropsis 
spiritus Gastrosaccus 
Gastrosaccus 
spinifer 
(3) Amphipods Paratylus 
swammerdami 
(4) Larvae Decapod Decapod Decapod Decapod 
(5) Miscellaneous Sagitta Oikiopleura Sagitta 
hexaptera Polynoid 
Pleurobrachia polychaetes 


(6) Post-larval fish Herring 


* The species in italics were the most numerous in each sample. 


Samples 1 and 2 were particularly interesting in that they consisted almost 
entirely of the very widespread genera of copepods Temora, Pseudocalanus © 
and Calanus. These organisms directly or indirectly represent a very im- 
portant stage in the food supply of the cod and thrive on diatoms such as 
Nitzschia closterium {Allen and Nelson, 1910], which has been shown to be a 
very rich source of vitamin A [cf. Jameson, Drummond and Coward, 1922]. 

Small fish. Having shown that the food of the majority of the small fish 
and certain other marine organisms is relatively rich in vitamin A, the next 
step was to study whether this factor is transferred to the tissues of the 
species which feed upon them. For this purpose we have taken one or two 
representative species from different groups without attempting to study the 
matter exhaustively, and have where possible selected those which most 


ORIGIN OF VITAMIN A IN FISH LIVER OILS 521 


frequently form the food of the cod and the related species which yield large 
quantities of marketable medicinal oils. 

The food of the cod is not at all times the same. In Norway the early 
spring rise in diatoms leads to the vast multiplication of copepods, which 
form the food of enormous shoals of small fish which come in to the northern 
coast (Finmarken) to spawn. These shoals are chiefly composed of the caplin 
or capelan (Mallotus vilosus) and they are followed by great numbers of cod, 
coal-fish, haddock and other species which devour them in enormous quan- 
tities [Hjort, 1914]. Other species of young fish are frequently consumed by 
these larger species, but their diet also includes certain shell fish, larger 
crustacea, salps, and especially in the case of cod on the Newfoundland banks, 
squid. 

The results of our few tests are given in Table II, but they only represent 
the testing of a few miscellaneous products we had on hand. No attempt has 
been made to examine these and similar products exhaustively. 


Table IT. 
Approximate daily 

Species dose to rats Results 

Caplin, Mallotus vilosus 0-1 g. Rapid growth 
Sprat, Clupea sprattus 1-0 g. » 
oung herring, Clupea harengus 0-4 g. a ‘ 
Mussel, Mytilus edulis 4-0 g. Good ,, 
Periwinkle, Littorina littorea 1-0 g. ” » 


Shrimp, Crangon allmanni 0-01 g.* < ae 
* Of unsaponifiable matter. 

Of these species the periwinkle is interesting as being an example of a 
direct transference of the vitamin from the algae on which it feeds (Entero- 
morpha, etc.). The fat extracted from these animals after feeding is deeply 
coloured with chlorophyll and other pigments derived from these green weeds. 

It would be an almost insuperable task to demonstrate an actual trans- 
ference of vitamin A right through from a diatom such as Nitzschia to its 
final location in the liver of the cod, but it would appear that we are justified. 
in assuming that such transference does indeed take place by relying on the 
fact that as yet synthesis of this dietary factor by an animal organism has 
not been demonstrated, whereas such synthesis is readily carried out by 
certain plants both marine and terrestrial. The dependence of marine animals 
on certain products of the synthetic powers of marine plants may indeed be 
as fundamental as that which exists between these two great groups of 
organisms on land, for many examples may be recalled such as that given by 
Fabre-Domergue and Bietrix [1900] who showed that a supply of microscopic 
plant life would save the lives of larval fish (soles) which were failing from 
malnutrition in spite of the fact that their yolk was not completely absorbed. 

We have included in this paper a few of the unpublished results obtained 
with molluscs by the late Dr H. Lyster Jameson as well as the examination 
of certain material he had collected for other purposes since they have a certain 
bearing on the main subject. 


522 J. C. DRUMMOND AND §&. S. ZILVA 


SUMMARY. 


1. The ultimate origin of the vitamin A found in the oils derived from 
fish, and particularly the fish liver oils, would appear to be chiefly the uni- 
cellular marine plants. Except very occasionally these organisms are not 
consumed directly by the fish. 

2. The extraordinary rise in the number of marine plants which begins 
as soon as the intensity and duration of sunlight increase early in the year is 
followed by a rapid rise in the organisms, largely copepods, and larval 
decapods and molluscs, whose growth and development are dependent on 
their food supply which consists of minute plants. These minute animals, 
which form a large proportion of plankton, contain relatively large quantities 
of vitamin A presumably derived from the diatoms on which they have 
thriven. 

3. The plankton forms the staple food of innumerable species of marine 
animals from small fish to some whales. This no doubt accounts for the 
presence of vitamin A in the tissues or fat depots of these animals. 

4. The fish which yield the large bulk of the liver oils, cod, haddock, 
coal-fish, etc., feed on many species. At one season in northern Norway and 
Newfoundland they feed extensively on a small fish, the capelan or caplin, 
which has been found to be very rich in vitamin A, doubtless derived from 
its food. | ; 

5. The origin of the vitamin A in fish oils and fish liver oils has therefore 
been traced back to the synthetic powers of the marine algae which form the 
fundamenta] food supply of all marine animals. 


It is a pleasure to acknowledge the valuable advice and assistance given 
to us by Dr Allen and Dr Lebour of the Marine Biological Laboratory, 
Plymouth, Dr Borley of the Fisheries Laboratory, Lowestoft, Professor Hjort 
and our late colleague Dr H. Lyster Jameson. 

We also beg gratefully to acknowledge a financial grant from the Medical 
Research Council which defrayed the cost of this investigation. 


REFERENCES. 


Allen and Nelson (1910). Quart. J. Micros. Sci. 55, 361. 
Coward and Drummond (1921). Biochem. J. 15, 530. 
Fabre-Domergue and Bietrix (1900). Bull. de la marine marchande, Paris. 
—_ (1914). Rapports, Conseil permanent international pour 0 Rxploration de la Mer, Copen- 
agen, 20 
(1922). Proce. Roy. Soc. B. 93, 440 
Jameson, Drummond and Coward (1922). Biochem. J. 16. 
Lebour (1918). J. Mar. Biol, Asa, 11, 433. 
—« (1919, 1). J. Mar. Biol. Ass. 12, 9. 
—— (1919, 2). J. Mar. Biol. Asa, 12, 21. 
—— (1920). J. Mar. Biol, Aas, 12, 262, 
(1921). J. Mar. Biol, Asa, 12, 458, 
Moore (1921). J. Chem. Soc. 149, 1555, 
Zilva and Miura (1921), Laneet, i, 323. 


XLIX. THE RESPIRATORY EXCHANGE IN 
FRESH WATER FISH. III. GOLD FISH. 


By JOHN ADDYMAN GARDNER, GEORGE KING 
AND EDWIN BOOTH POWERS. 


From the Physiological Laboratory, University of London, 
South Kensington. 


(Received May 22nd, 1922.) 


In part I of this series of papers [Gardner and Leatham, 1914, 1] a detailed 
account was given of apparatus for measuring the respiratory exchange in 
fish, and measurements were recorded showing the influence of temperature, 
size of the animal, etc. on the respiratory exchange in the case of brown trout. 
In this paper we give an account of experiments by the same method on 
gold fish. 
METHOD. 


Preliminary experiments showed that gold fish live at a much lower plane 
of metabolism than trout, and it was found necessary to modify slightly the 
procedure adopted with the latter animals. 

In the case of trout the duration of the experiments varied from four to 
six hours. Gold fish, however, both at low and at ordinary temperatures use 
much smaller volumes of oxygen per hour, and it was found necessary to 
continue the experiments over periods of 24 hours in order to obtain accurate 
measurements. At higher temperatures more oxygen was used and satis- 
factory results could be obtained in experiments of six hours’ duration. 

In most of the 24 hour experiments it was not found practicable to continue 
the pumping of the air through the water during the whole period, owing to 
irregularities during the night of the electric power used in working the pumps, 
though in one or two experiments the pumping was continuous. Usually the 
experiment was commenced about 11 a.m. and the pumping of the air through 
the water bottle was continued until as late as convenient in the evening. 
'The pump was started again early in the morning and continued until the 
end of the experiment. Experience showed that the absorption of oxygen 
was so slow that, with the large bulk of water used, the tension of the dis- 
solved oxygen was not reduced sufficiently during the time the pumps were 
out of action to affect the fish. Oxygen, equivalent to that used during the 
24 hours, was gradually added to the air above the water in the experimental 
bottle during the period before the cessation of pumping. 


524 J. A. GARDNER, G. KING AND E. B. POWERS 


To obtain accurate results by this method, it is essential to take the 
greatest care in measuring the temperature and pressure of the air above the 
water at the beginning and end of the experiment. . 

Some minor improvements were made in the thermostatic arrangements 
of the apparatus used for the trout, and the pressure gauges were fitted with 
spirit levels. Care was also taken to paint over all joints, and particularly 
the line of contact of the rubber bung and the neck of the bottle with a 
rapidly drying acid proof enamel paint. The gold fish used were about five 
to six inches long and averaged about 50 g. in weight. They were kept in a 
large tank outside in. slowly running water. The tank had a gravel bottom 
and contained suitable water plants. The fish were fed in the usual way and 
remained perfectly healthy during the long period—more than a year—over 
which the experiments extended. 


EXPERIMENTAL RESULTS. 


A detailed example of the method of calculation was given in the paper 
referred to above [1914, 1, pp. 379-380]. 

In the following sauteobls we give only details of the number and weight 
of fish used, the temperatures, and the fotal initial and final volumes of free 
and combined carbon dioxide of oxygen and of nitrogen in cc. reduced to 0° and 
760. The volume of nitrogen should theoretically remain constant and the 
small positive and negative differences actually found afford a ready indica- 
tion of the experimental errors. 


Experiments at low temperature. 


(1) Seven fish; weight 362 g.; individual variations 45-60 g. The fish 
prior to experiment had lived outside in cold weather. Initial temperature 
of water in experimental bottle 5-8°, final temperature 6-1°. Duration of 
experiment 22-43 hours. 


CO, Initial 1352 ce. Final 1459-4 ce. Difference + 107-4 ce. 
Oxygen re 5646-8 » 8454-4 tS — 192-4 
Nitrogen »  20356-4 » 20349-3 a - Fl 


Nitrogen error —0-034 %. 
(2) Seven fish; 348 g.; variations 42-59 g. LT. 7-4°, F.T. 6-2°. These fish 
had been kept at from 3° to 6° for several days prior to the experiment. 
Duration 23-17 hours. 


co, Initial 1571-8 ce, Final 1689-1 ce. Difference + 117-3 ce. 
Oxygen os 5162-2 » 6069-1 o — 93-3% 
Nitrogen ms 1868-8 » 18705-4 oi + 186 


Nitrogen error + 0-099 %. 
(3) Twelve fish; 601 g.; variations 38-65 g. These fish had had a long 


spell of mild winter weather prior to experiment. I.T. 6-15°, F.T. 1:4°. 
Duration 23-75 hours. 


CO, Initial 1475 ce, Final 1667 ce. Difference +192 cc 
Oxygen ae 5261°7 » 60488 , ‘i - 212:9 
Nitrogen » 189919 »» 19036°7 a + 44:8 


Nitrogen error +023 %. 


Fe 


RESPIRATORY EXCHANGE IN GOLD FISH 525 


smal] | Seven fish; 382 g.; variations 40-65 g. L.T. 6-2°, F.T. 6-3°. The fish 
had mild winter temperature prior to experiment. Duration 23-78 hours. 


CO, Initial 2167 cc. Final 2230-3 ce. Difference + 63-3 ce. 
Oxygen » 4321-9 » 4181-8 a — 140-1 
Nitrogen »  15469-4 » 15426-4 ae — 43:0 


Nitrogen error —0-28 %. 


(5) Seven fish; 368 g.; variations 40-60 g. These fish had been in water 
covered with a film of ice for the previous three days. I.T. 5-7°, F.T. 6-05°. 
Duration 23 hours. 


CO, Initial 1648-6 ce. Final 1781-5 ce. Difference + 132-9 ce. 
Oxygen 7 5216-4 » 8091-2 re — 125-2 
Nitrogen »  18894-4 » 188841 Pe —- 103 


Nitrogen error —0-054 %. 


At medium temperatures. 


(6) Seven fish; 355 g.; variations 35-60 g. I.T. 13-55°, F.T. 16-7°. Duration 
24-25 hours. ae 


CO, Initial 2066-9 cc. Final 2459-2 cc. Difference +392-3 ce. 
Oxygen Pe 4427-2 » 4020-7 ms — 406-5 
Nitrogen »  16044-3 » 16017-6 a — 26-7 


Nitrogen error -0-17 %. 


(7) Seven fish; 355 g.; variations 35-60 g. I.T. 13-55°, F.T. 15-9°. Dura- 
tion 23-25 hours. 3 


CO, Initial 2123-9 ce. Final 2392-0 cc. Difference + 268-1 ce. 
Oxygen a 4413-4 »  93976:3 a — 437-1 
Nitrogen »  15966-4 », 15931-4 a ~ 35-0 


Nitrogen error —0-22 %. 


At higher temperatures. 


(8) Seven fish; 370 g.; variations 45-62 g. I.T. 21-4°, F.T. 20-6°. Duration 
23-18 hours. In this experiment the valves of the pumping apparatus were 
out of order and the air could not be pumped through the water. At the end 
the water in the experimental bottle had a strong faecal smell. 


co, Initial 1518-4cc. | Final 2256-5cce. Difference +7381 cc. 
Oxygen = 5098-7 53> 4521-2 Ses -—577°5 
Nitrogen »  18551-2 ys 185191 ‘i - 321 


Nitrogen error 0-17 %. 


(9) Seven fish; weight 361 g.; variations 45-63 g. I.T. 20-7°, F.T. 20-4°. 
Duration seven hours. At the end no objectionable smell was noted. 


co, Initial 1649-1 ce. Final 1855-2 ce. Difference + 206-1 cc. 
Oxygen sh 4640:8 » 4354-0 a — 286-8 
Nitrogen +  16805-0 » 16819-9 os + 14:9 


Nitrogen error +0-088 %. 


The results of these experiments are given in Table I. 


é 


526 J. A. GARDNER, G. KING AND E. B. POWERS : 


Table I. a the 
Average Volume of Volume of Volume of Volume of : 
weight oxygen con- CO, produced oxygen con- CO, produced 
of fish sumed per fish per fish per sumed per kilo of _ per kilo of fish 
No. of usedin perhourince.at hourince. at fish per hourin per hour in cc. Respiratory 
exp. g- 0° and 760 0° and 760 cc. at 0° and 760 at 0° and 760. quotient 
At low temperature: 
(I) 51-7, 1-225 0-684 23-70 13-23 0-56 
(2) 49-7 0-574 0-723 — 11-56 14-55 1-26 
(3) 50-0 0-747 0-674 14-91 13-45 0-98 
(4) 54:5 0-842 0-380 15-42 6-97 0-45 
(5) 52-5 0-778 0-825 14-79 15-70 1-06 
. Mean 0-833 0-657 16-07 12-78 
At medium temperature: © ~ , 
(6) 50-7 2-395 2-309 47-22 45-54 0-97 
(7) 50-7 2-686 1-647 52-96 32-48 0-61 
Mean 2-541 1-978 . 50-09 39-01 
At higher temperature: me 
(8) 52-8 3-560 4-550 67-35 86-08 1-28 
(9) 51-6 5-853 4-207 113-50 81-55 0:72 
Mean 4-706 4:379 90-43 83-81 


INFLUENCE OF TEMPERATURE. 


It will be seen from Table I that the volume of oxygen consumed per fish, 
or per kilo of fish, increases with the temperature and very roughly in pro- 
portion to the temperature. 

We should scarcely expect to find any exact proportion in experiments 
on the living animal, more especially as, owing to the long period over which 
the work extended, different sets of fish were used, and some latitude must 
be allowed for individual idiosyncrasy. Further in most of the experiments 
it was necessary to keep the animals in the experimental bottle for about 
24 hours and during some hours the pumping of air was discontinued, so that 
during some part of the experiment the oxygen tension in the water would 
be below saturation. Experiments 8 and 9 at higher temperatures need some 
comment, as they show marked difference in the oxygen absorption, though 
most of the fish used were the same in each. 

In experiment 8, which was of 24 hours’ duration, the valves of the 
apparatus went wrong at the beginning, so that it was impossible to pump 
the air of the bottle through the water. Consequently the oxygen tension in 
the water progressively decreased, and at the end the partial pressure of the 
oxygen in the water was only 0-583 °% of an atmosphere. At the end it was 
also noticed that the water had an objectionable faecal smell, a condition not 
found in any of the other experiments. It seemed probable therefore that the 
oxygen consumption figure was below normal. The experiment was therefore 
repeated (No. 9) and the duration was reduced to seven hours. Pumping was 
continuous, The oxygen figure is markedly higher than in No 8, and the 
respiratory quotient approximately normal for an unfed animal. | 

In the experiments at low and at normal temperatures, owing to the 
relatively slow oxygen absorption of the fish, it did not seem likely that the 


} 


RESPIRATORY EXCHANGE IN GOLD FISH 527 


small reduction of the oxygen tension in the water during the non-pumping 
stage of the experiment would have much effect. 

To test this, however, we made the following experiment at a moderate 
temperature keeping the oxygen tension below normal during the whole time. 

(10) Seven fish; weight 329 g.; variations 22-65 g. I.T. of water 16-8°, 
F.T. 16°. The water had been previously partially denuded of oxygen and 
the gas over the water was air and nitrogen. Duration six hours. The partial 
pressure of the dissolved oxygen was found to be at the beginning 6-17 % 
and at the end 2-69 % of an atmosphere. 


CO, Initial 436-0cc. | Final 510-lcc. ‘Difference + 74:1 ce. 
Oxygen » 2301-9 » 2168-2 5 — 133-7 
- Nitrogen » —-:23818-5 » 23847°6 i + 29-1 


Nitrogen error +0-12 %. 


Oxygen absorbed per fish per hour 3-114 ce. or per kilo of fish per hour 
66-26 cc. Carbon dioxide evolved per fish per hour 1-726 cc., or per kilo per 
hour 36-72 cc. The respiratory quotient is 0-55. 

At the temperature used the oxygen absorption should be a little higher 
than the mean of experiments 6 and 7. This is evidently the case, and we may 
conclude that the non-pumping interval in the experiments at low and at 
medium temperatures did not appreciably affect the result. 


COMPARISON OF THE OXYGEN REQUIREMENTS OF TROUT AND GOLD FISH. 


In Table II the oxygen requirements of trout and gold fish, at various 
temperatures, are compared. The gold fish were 5 to 6 inches in length, and 
the sets of trout 4 inches and 8 inches respectively, but comparisons in terms 
of length are somewhat imperfect, owing to the difference in shape of the 
two kinds of fish. We were unable to obtain gold fish of the same average 
weight as the larger trout, in sufficient number. A glance at the table will 


Table II. Comparison of oxygen consumption of gold fish and trout. 


4 inch trout, 
average weight 23 g. 8 inch trout, 5 to 6 inch gold fish, 
ooo average weight 102 g. average weight 51 g. 
Per fish Per kilo ——e——, _ A — 
per hour of fish Per fish Per kilo Per fish Per kilo 
°-7° 2-05 89-04 10-48 102-53 0-83 16-07 
13°-16° 2-82 117-40 17-84 192-36 2-54 50-09 
°-22° 4-63 213-80 19-57 204-24 4-71 90-42 


25° 5-64 _ 258-87 _ _— _ _ 
show that the gold fish live at a much lower plane of metabolism than trout, 
- and require much less oxygen. They also appear to react to temperature over 
a wider range than trout. At temperatures round 16° gold fish use about 
three times as much oxygen as at 4° to 7°, whereas the 8 inch trout use about 
double. A further rise in temperature from about 16° to between 20° and 22° 
again doubled the oxygen used by the gold fish, but caused only a small 
increase in the case of trout. It would appear at first sight that in the case 
of the large trout the organism ceases to react to ris€ of temperature about 


aN 


528 J. A. GARDNER, G. KING AND E. B. POWERS 


the limit of 16° to 17°, but that in the case of gold fish this limit is higher, 
for the ratio of oxygen consumption of gold fish to trout at 4° to 7° is about, 
1:6, at about 16° 1:4, and at 20° to 23° nearly 1:2. Itis probable, however, 
that this difference in reactivity to temperature is more apparent than real, 
for it has been shown [Gardner and Leatham, 1914, 2] that trout can live and 
keep well for long periods at a lower level of oxygen partial pressure than that 
corresponding to full saturation, and it is well known that gold fish flourish 
at very much lower levels of saturation; in addition the large trout were some- 
what restricted in their movements in the experimental bottle compared with 
the smaller animals, and were consequently less favourably situated for assisting 
the pumping of water through the gills by swimming movements. If we 
make the assumption that a fish absorbs the whole of the oxygen from the 
water passing through its gills, then calculating on the basis of Winkler’s 
figures [1905] for the oxygen content of water saturated with air at various 
temperatures, a 4 inch trout at 7° to 8° would need to pump about 4 cc. per 
minute through its gills, while at 21° he would need about 12-4 cc. per minute 
and at 25°, 16 cc. per minute. On the same basis an 8 inch trout would need | 
to pump at 6°, 20 cc. per minute, and 16°, 44 to 45 cc. The above 
assumption is perhaps scarcely warrantable, so that these figures must be 
regarded as minimum values. It is therefore perhaps not surprising that the 
large fish should apparently cease to react to increases of temperature at a | 
lower temperature than the small fish. 

Another factor which appears to bear on the difference between trout and 
gold fish in this respect is that gold fish are able to bear with impunity higher 
temperatures than trout. As was shown in the paper referred to above 
[1914, 1] temperatures round about 25° appear to be dangerous or fatal to 
trout. Measurements of respiratory exchange were successfully carried out 
on 4 inch trout at 25°, but on attempting to determine the oxygen absorption 
for 8 inch trout at this temperature, of the three fish used in the experiment, 
one died in about two minutes, though the water was fully oxygenated, and 
the other two turned over on their backs in about ten minutes, and would 
no doubt have died, but on adding cold water they recovered completely and 
rapidly, and appeared to be quite well the next day. This result was not due 
to any sudden change of temperature, as the fish were carefully and gradually 
warmed up beforehand during the course of one hour. In another experiment 
four 8 inch fish were accidentally plunged into well-oxygenated water at 33°, 
The fish gave a few violent leaps and collapsed as one might perhaps imagine 
a warm-blooded animal doing on falling into boiling water, and were all quite 
dead in under one minute. The water was afterwards found to be quite 
frothy, as though it contained some saponaceous material. Some experiments 
were made with gold fish to ascertain the maximum temperature compatible 
with life. A strong healthy gold fish (about 50 g.) was placed in a large earthen- 
ware basin of water at 20°. After 15 minutes warm water was carefully added, 
with good stirring, until the temperature rose to 25°, At intervals of 10-15 


RESPIRATORY EXCHANGE IN GOLD FISH 529 


minutes more hot water was gradually added, and the temperature very 
gradually increased. At 30° the fish was very active, swimming about more 
rapidly and continuously than at lower temperatures, but beyond this showed. 
no apparent sign of distress, though the breathing, as one would expect, was 
rather rapid. At 35° the fish apparently died without any struggle, remaining 
on his side at the surface, the mouth being closed. No sign of life being ob- 
served after ten minutes at this temperature, the water was then gradually 
cooled down to 17°. After about one hour the fish was observed to be breathing 
feebly. It then began to recover more rapidly; it was put into the store tank 
outside, and the next day appeared to have completely recovered. 

It would seem from these experiments that trout can live in water up to 
a temperature of from 20-25°, but gold fish can exist without apparent harm 
up to at least 30° or even higher. This result is, we believe, in accordance 
with the geographical distribution of these fish. As far as we are aware trout 
are never found in tropical rivers, except such as have a snow mountain 
source, 

RESPIRATORY QUOTIENT. 


We are not yet able to explain the variations noted in the respiratory 
quotient, particularly the very low values observed with trout at low tem- 
peratures, but a careful examination of our experimental results has con- 
vinced us that the variations cannot be attributed to experimental error. 

Experiments are in progress which we hope will throw light on this point. 


We take this opportunity of expressing our thanks to the Government 
Grant Committee of the Royal Society for aid in carrying out this work. 


REFERENCES. 


Gardner and Leatham (1914, 1). Biochem. J. 8, 374. 
—— —— (1914, 2). Biochem. J. 8, 591. 
Winkler (1905). Landolt Bornstein Meyerhoffer, Tabellen, 3 Auflage, 184. 


Bioch. xv1 35 


L. NOTES ON SOME PROPERTIES OF 
DIALYSED GELATIN. 


By DOROTHY JORDAN LLOYD. 


From the Biochemical Laboratory, Cambridge, and the Laboratory of the 
British Leather Manufacturer's Association. 


(Received May 29th, 1922.) 


I. Material. 


THE gelatin on which the following observations was made was purified by 
dialysis at the iso-electric point (py = 4:6) and subsequent precipitation in 
strong alcohol. Details have been published in a previous paper [Jordan 
Lloyd, 1920]. The dry gelatin contained 0-06 to 0-00 % of ash. It was obtained . 
as a snow white, fibrous and brittle solid, with a low solubility. In one specimen 
the maximum solubility was 1-5 % in boiling distilled water. In others 2-2-5 % 
could be dissolved under the same conditions, but in no case was it possible 
to dissolve more than 2-5 % of the pure fibrous gelatin in pure boiling water. 
This agrees with the observations of Dheré [1910] and Dheré and Gorgolewski 
[1913] that prolonged dialysis is accompanied by diminished ash content and 
decreased solubility. The hot water solutions were clear and colourless. On 
cooling to 15°, if the concentration were above 0-9 % the sol set to a turbid 
white gel in 24 hours, below 0-9 % but above 0-6 %, gelation took from 
2-5 days. At a concentration of 0-6 % or less, the clear hot sols cooled to 
form turbid white sols at 15°, the turbidity decreasing with decreasing gelatin 
content. The gelatin in these sols is not precipitated by ferric, mercuric or 
lead salts, nor by ferrocyanides, dichromates or tannic acid. Gels formed 
under these conditions are not stable, the gelatin separating from the water 
[see Jordan Lloyd, 1920; Smith, 1921]. The solubility of dialysed gelatin is, 
as would be anticipated, immediately increased by the addition of acid or 
alkali. Gels formed in the presence of free acid or base are glassy clear, 
colourless and stable. 


Il. The influence of hydrochloric acid, sodium hydroxide and sodium 
chloride on the gelling power. 


“Gelling power” is a difficult phrase to define precisely, and the property 
which it purports to describe is equally difficult to measure accurately. The 
following notes do not claim to set out any satisfactory measurement or 
definition of gelling power, the differences described between the experimental 
fluids being purely comparative, and observed under conditions kept as 


PROPERTIES OF DIALYSED GELATIN 531 


uniform as possible. The solutions of gelatin used for this comparative work 
were made by dissolving dialysed gelatin in hot water, raising to boiling point, 
and then allowing to stand for 48 hours at 15°. The minimum concentration 
of gelatin required to produce a gel under these conditions was assumed to 
give an inverse measure of the gelling power. A fluid was taken as gelled 
when it remained in its place in an inverted test-tube. The influence of hydro- 
chloric acid and sodium hydroxide was examined in the absence of sodium 
chloride and in the presence of 0-1 N, 0:4 N and N sodium chloride. The 
reactions of the 1 % gelatin solutions were measured electrometrically and 
also by adding a few drops of indicator at the time of mixing, and comparing 
the colours with suitable standards at the end of the experiment. It is well 
known that the addition of sodium chloride to a solution containing hydro- 
chloric acid increases the concentration of hydrogen ion but Harned [1915] has- 
shown that the increase is not great. In the present case it will be well within 
the experimental error. The measurements of reaction are accurate to + *2 py. 
The results of the experiment are summarised in Tables I, III, IV and V, 
and a diagram, in which the minimum concentration that will cause gelling 
is plotted against reaction, is given in Fig. 1. Areas above the curves are sols, 
those below are gels. 

Conc. % 


0. 


AREY ee ee ave eee ee 0 ke Me Me 
Iso-electric point of gelatin 


Fig. 1. 
Abscissae = pq. 
Ordinates = Minimum Pigerys sor of gd cent. at which gel f forms. 15° and 48 hours. 
x : =0: AN NaCl * x prt ON NaCl 


Pure gelatin under the given conditions (48 hours’ standing at 15°) requires 
a minimum concentration of 0-8 % to form a gel, and will only form gels at 
1% concentration from py = 12-3, and again rather surprisingly between 
Pu 2 and 0-7. The disappearance of the power of gelation in acid solutions and 
its temporary reappearance in stronger acids is a phenomenon which has not 
previously been described. The conditions necessary to observe it are rather 
restricted, temperature being an important factor as is shown in Table I. 


532 D. JORDAN LLOYD 


Table I. 
1 % gelatin after 48 hours 
Concentration of — A — 
pa. HClor NaOH 0° C. 15°.C. 20° C. 
0-7 -25 N HCl — - Clear fluid | — 
0-8 -20 Transparent fluid » very soft gel Clear fluid 
0-9 15 ” ” soft gel ” ” 
1-1 10 ” ”? ” gel ” 9 
1-5 05 ” soft gel ” 2 ” ” 
1-6 “04 —_ » soft gel _- 
ad 03 mae ” ” Ta 
2-0 -02 -- » fluid — 
2-8 ‘O01 Transparent fluid = a3 Clear fluid 
3-8 -005 5 soft_gel » soft gel ye teae 
002 Gel with ice crystals —_ » gel 
0025 — Clear gel — 
“0015 Very faintly turbid gel oo Clear gel 
-0010 Faintly turbid gel os Faintly turbid gel 
0005 Turbid gel; hysteresis — Turbid gel 
5-1 Distilled water Opaque white gel Opaque white gel Be Bult 
0005 N NaOH = Faintly turbid gel -— Very faintly turbid gel 
-0010 Very faintly turbid gel _— Clear gel 
0015 ” ” : bani > ” 
002 ” ” Rare ” ” 
10-4 005 Clear gel Clear gel » fluid 
11-7 ‘O01 ;, viscous fluid on re i 
12-15 -02 — » very soft gel — 
12-4 03 — » viscous fluid —- 
12-5 04 — » fluid — 
12-6 05 Clear fluid “ zs oe 
12-9 10 oF % 3g 3 Clear fluid 
13-1 20 ” ” ” ” ” ” 


At concentrations of 0-1 N, and 0-4 N, sodium chloride lowers the minimum 
concentration of gelatin required to produce gelation, except at reactions 
more alkaline than py = 12-2 where in no instance was a gel obtained. At 
normal concentration sodium chloride may either increase or decrease the 
gelling power, according to the value of the hydrogen ion concentration of 
the solution under investigation. Its influence is most marked from py = 5 
to py = 1-5, v.e. at the range where pure gelatin gels least readily. It is 
interesting that this range is immediately on the acid side of the iso-electric 
point. The relationship between sodium chloride content and gelling power 
is obviously not a simple one and requires further study. One clear point 
however emerges from the study of Fig. 1, ¢.e. the greater the concentration 
of neutral salt, the less the influence of the hydrogen ion. 

If gelation is regarded as a function of viscosity, the influence of sodium 
chloride as found by the experiments described here, does not agree with 
previous work by Mérner [1889] and Freundlich [1909]. . 

The inhibition of setting in strongly acid solutions both in presence and 
absence of salt can be shown to be due to the hydrogen ion since on neutralisa- 
tion with sodium hydroxide the gelling power is restored. Neutralisation does 
not restore gelling power to the strongly alkaline sols, The gelling power in 
this case is permanently destroyed. 


PROPERTIES OF DIALYSED GELATIN 533 


Ill. The influence of hydrochloric acid, sodium hidroxide and sodium 
chloride on the turbidity of sols and gels. 


The observations on turbidity were made in all cases on solutions of 
gelatin of 1 % or weaker, and refer only to their condition at 15°. 1% gels 
of gelatin in water are turbid under these conditions; in 0-1 N sodium chloride 
they are less turbid, in 0-4 N sodium chloride they are again slightly less 
turbid than in 0-1 N salt and in normal sodium chloride they are only faintly 
opalescent. In acid solutions the influence of sodium chloride is reversed. 
In the absence of sodium chloride the 1 % gelatin solutions in hydrochloric 
acid are generally clear. A slight turbidity develops on standing if the acid 
concentration does not exceed 0-005 N [Jordan Lloyd, 1920]. The addition 
of sodium chloride in concentrations of 0-1 N, 0-4 N and 1-0 N causes the 
clear gels to become turbid. In the last case the gels containing 0-01-0-008 N 
HCl are actually cloudier than those with salt alone. In dilute alkaline solu- 
tions (0-002 N or less) the addition of sodium chloride causes the clear gels 
to become turbid. In stronger alkaline solutions (0-05 N or more) the addition 
is without effect, but evidence from other sources shows that the gelatin is 
actually destroyed at this reaction. 

The influence of acid, alkali and salt is summarised in Tables II, III, IV 
and V. Miss Laing has very kindly examined the turbid water gels for me 
by means of the ultra-microscope and finds that they contain long fibres 
similar to the fibres in soap curds. 

The relations of the turbidity or “degree of dispersion” of the gelatin to 
acid, alkali and salt form a close parallel to those for globulin described by 
Hardy [1905]. 

Table II. 


48 hours at 15°. Salt concentration =0-00. 
Concentration of gelatin 
A. 


— : — 
Reaction 10% 09% 08% 7% 0-6 % 
:25 N HCl po =0-7_ Clear fluid — ies BE 
‘20 N 0-8 » soft gel Clear ‘gela- — — ae 
tinous fluid 
15 N 0-9 ” ” ” ” 4 — — 
“10 N Te] » gel is a = = — 
05 N 15 ” ” ” ” are wae rar 
“04 N 1-6 bot SOR 3, Ee — Se lees 
03 NV LT ” ” > ” _—; ——e == 
02 N 2-0 » fluid Clear fluid — os aes 
‘01 N 2-5 ” ” ” ” = —s = 
005 NV 3:8 » softgel ,, — a Cee 
0025 N 4-4 Hy. me Clear gela- a Eas Nas 
tinous fluid 
Water 5-0 Firm opaque yop soft Turbid soft Turbid gela- Turbid fluid 
gel gel gel tinous fluid 
005 N NaOH 10-4 Clear gel Clear gel Clear gela- Clear fluid _ 
tinous fluid 
‘O01 N 11-7 ” ” ”» ” “oo ” ” ” ates 
02 N 12-1 Clear very » fluid ier a aay 
soft gel 

03 NV 12-4 Clear fluid — — <a ei 
04 N 12:5 

‘05 N 120] nis — — — — 
‘10 N 12-9 


Dp. JORDAN LLOYD 
Table ILL. 


48 hours at 15°. Salt concentration =0-10 WN. 


Concentration of gelatin 
A 


pu = 13-0 B at 


Reaction 1-0 % 0-9 % 0-8 % 0-7 % 0-6 %, 0-5 %, 
Hydrochloric acid 
pu= 08  Feaintly ~ —_ — mee — 
turbid gela- 
tinous fluid | A 
Ppu= 1-1 Faintly Faintly Clear fluid _— — — 
turbid gel _—_ turbid soft 
gel 
Pu 1-5 oo ” ” ” 2 ” s — eid 
pPu= 2-7 ae aris Faintly Faintly Clear gel Clear gela- Clear fluid 
turbid gel _—_ turbid soft tinous fluid 
gel 
Pu= 3°8 ” ” ” ” Faintly ” ” Clear soft ” ” 
turbid gel gel 
Water 
Pu= 5-1 Turbid gel ‘Turbid gel aera Faintly Faintly Faintly 
turbid gel turbid soft turbid 
gel fluid 
Sodium hydroxide 
pu=10-4 =‘ Faintly ee teh Sir ds Faintly Faintly — 
turbid gel turbid soft turbid fluid 
gel 
Pu= 11-8 ” ” ” ” ” ~< Faintly ” ” — 
turbid gel 
pu=12-6 — Clear fluid ae _ — — a 
Ppu= 13-0 ” ” —— bea Nanas ge) - 3 
Table IV. 
48 hours at 15°. Salt concentration =0-40 N. 
Concentration of gelatin 
Reaction 1-0 % 0-9 %, 08 % 0-7 %, 0-6 % 0-5 %, 
Hydrochloric acid 
pu= 08 ~~ Faintly — — — — — 
turbid fluid 
pu= 11 Turbid gel Turbid soft Faintly Clear fluid — a 
gel turbid soft 
gel 
pu= V5 id se és 7” e Faintly Faintly = 
turbid gela- turbid fluid 
tinous fluid 
Pu= 2°7 As »  Turbid gel  Faintly = a Faintly Clear fluid 
turbid gel turbid gela- 
tinous fluid 
pu= 38 ” ” ” ” ” ” ” ’ ” ” ” 98 
Water 
Pus il ” ” ” ” ” ” Faintly ” ‘ Faintly 
turbid gel turbid fluid 
Sodium hydroxide 
pu=104 = Turbid gel ve - » * yor soft ns » Clear fluid 
ge ° 
pu =i1bs ‘fe a ” »  Faintly Clear gela- Clear fluid — 
turbid soft — tinous fluid 
gel 
yu2l26 Clear fluid — — — —— 


PROPERTIES OF DIALYSED GELATIN 535 
Table V. 
48 hours at 15°. Salt concentration =1-0 NV. 
Concentration of gelatin 
Reaction 10 % 0-9 % 08% 
Hydrochloric acid 
pu= 08 Faintly turbid fluid — _ 
pu= 09 Faintly turbid Faintly turbid fluid oa 
gelatinous fluid 
pu= Il Faintly turbid soft gel ns os _- 
pu= 15 Faintly turbid gel Faintly turbid Faintly turbid fluid 
gelatinous fluid 
pu= 2-5 Turbid gel Faintly turbid gel Faintly turbid 
gelatinous fluid 

pu= 38 ” ” “ ” ” » ” 

Water pu= 5-1 Faintly turbid gel = rs " v 
Sodium hydroxide 

pu=l0-4 Clear gel Clear gel wes » 

pu=1b7 ” ” ” fluid oF 

pu=121 » fluid a ee 

Pu 12-5 ” ” — te 


IV. Optical rotation. 


The rotatory power of a solution of gelatin is not a constant one but varies 
with temperature. This property has been examined by Trunkel [1910] and 
Smith [1919]. | 

Smith considers that above 35° gelatin is converted to a form which he 
calls gelatin A which has not the property of forming gels. Below 15° Smith 
considers gelatin exists in a form which he calls gelatin B, which is the sub- 
stance capable of gel formation. Between these two temperatures, a solution 
of gelatin is considered to be a mixture of the two isomers. The conversion 
of A to B is reversible. Smith, using the sodium line, gives the following 
values: 

[a], at 35° and above = — 141°, 


[a], at 15° and below = — 313°. 


The rotatory power of the purified gelatin used throughout these experi- 
ments was examined in solutions of distilled water, dilute acid, or dilute alkali. 
The reaction of the solution in distilled water is approximately py = 5. It 
was found that all three solutions showed mutarotation, and that the rotatory 
power differed in acid, alkaline and distilled water solution. The mercury green 
line was used as the source of light in all experiments. [¢]i¢ green 18 the rota- 
tion caused by a column of fluid 10 cm. in length when each cc. of solution 
contained 1 g. of pure gelatin. Constant values were obtained for the higher 
temperature by keeping the tubes for three hours at 37°: at the lower tem- 
perature by keeping the tubes for 48 hours at 10° before taking a reading. 
Intermediate values were not taken. All three solutions set to gels at the 
lower temperature, the acid and alkaline gels are clear, the water gel cloudy. 


536 D. JORDAN LLOYD 


For this reason a short tube had to be used in the last case. The following 
experimental figures were obtained: 


Solvent 0-017 NHCI Water po =5 0:017 N NaOH 
Concentration of gelatin 1-67 % 2-00 % 1-67 % 
Length of polarimeter tube 18-9 cm. 9-9 cm. | 20 em. 
Observed rotation at 37° — 5-03° —2-81° -- 497° 
Observed rotation at 10° —9-55° — 6-08° —9-34° 

[a}Hg at 37° — 160° — 142° — 150° 

{alte at 10° — 302° — 307° — 280° 


These values for [@],;, at 10° and 37° are reversible. After several heatings 
and coolings, the value of the alkaline, and to a less extent of the acid tube, 
tends to sink. The difference in the values of the acid, alkaline and neutral 
solutions at 37° is a considerable one. A further investigation of the influence 
of temperature and reaction on the value of [@]iz¢ green 18 being carried out. 


V. Note on the combination of gelatin with hydrochloric acid and sodium 
hydroxide and on molecular structure in sols and gels. 


If the gelling power of a gelatin solution be plotted against the reaction, 
in a system free from inorganic salts, as in Fig. 1, it can be seen that the 
former is reduced over a range of reaction immediately on the acid side of 
the iso-electric point, Michaelis and Grineff’s [1912] figure being taken for the 
last point. Over this same range, the gelatin is combining rapidly with hydro- 
chloric acid to form soluble hydrochlorides [ Procter, 1914; Loeb, 1919; Lloyd 
and Mayes, 1922]. On the theory that gels are formed of a framework of 
uncombined gelatin, penetrated with a solution of soluble gelatin salts, it is 
possible to suppose that the loss of gelling power at these reactions is due to 
the reduction of the uncombined gelatin to below the minimum necessary for 
the formation of a rigid framework. The temporary reappearance of the gel 
state at higher reactions is probably due to the reduction of the active mass 
of the water owing to the hydration of the protein ions. At higher py values 
the gel state is again inhibited. Immediately on the alkaline side of the iso- 
electric point, the presence of free hydroxyl ions has very little effect. Yet it 
is known that the combination of gelatin with sodium hydroxide begins 
immediately on the alkaline side of py = 4:6, though the rate of combination 
is slower by about a third for bases up to a concentration of 0-01 N of free 
hydrogen or hydroxyl ions. At this point on the curve of “base fixed” the 
curve inflects and this point corresponds to a py value of 12 beyond which 
gelation cannot exist [see Lloyd and Mayes, 1922, Fig. 5]. Smith [1921] finds 
that the rise of osmotic pressure to a maximum is more rapid in alkaline 
solutions than in acid. Evidence has been given elsewhere [Lloyd and Mayes, 
1922] that acids and bases combine with gelatin at different positions in the 
molecule and this is supported by the observations given above on the gelling 
power, It even seems possible that the sodium salts of gelatin may be able 
to build themselves into the jelly framework in the undissociated state. The 
loss of gelling power at py = 12 is to be attributed to actual destruction of 


PROPERTIES OF DIALYSED GELATIN 537 


the gelatin molecule in the presence of the strong alkali. Patten and Johnson 
[1919] also found that liquefaction accompanied high alkalinity (py = 8, 9 
or 10 in their experiments) but as their observations were all made in the 
presence of buffer salts they are not comparable with those given in this paper. 
The effect of sodium chloride in diminishing the minimum concentration of 
gelatin required to form a gel is obscure, but a possibility suggests itself that 
it may be due to withdrawal of water from the system by the hydration of 
its ions. This hypothesis is consistent with the fact that sodium chloride 
increases the turbidity of gelatin in acid or alkaline solution but is difficult 
to harmonise with the observation that sodium chloride decreases the turbidity 
of gels near the iso-electric point, and when in sufficient concentration also 
decreases the setting power. (It may be remarked in parenthesis that Dr 
Holker has examined the influence of sodium chloride on some of the gelatin 
used in this work, at a temperature of 40°. At this temperature, according 
to Smith, the gelatin would be entirely transformed into the non-gelling form. 
Dr Holker finds that at this temperature the curve of turbidity against sodium 
chloride content shows the characteristic sine form [Holker, 1921]. This fine 
variation of turbidity at 40° is however different from the coarser variations 
described in this paper which are more of the nature of an actual separating 
out of gelatin from its solutions.) 

In a recent paper [Jordan Lloyd, 1920] the theory was put forward that 
gelatin gels consist structurally of a framework of solid iso-electric or neutral 
gelatin together with water, through the open meshes of which lies an aqueous 
solution of ionised gelatin salts, either acid or basic. The solid phase was 
assumed to be the seat of the elastic forces of the gels, and the liquid phase 
of the osmotic forces. The necessity for gelatin in both the uncombined, un- 
charged, and the ionised, charged condition for the formation of a stable gel, 
was postulated. This theory has been criticised adversely by both Procter 
[1920] and Laing and McBain [1920] on the grounds that the structure postu- 
lated was supposed to be of microscopic dimensions. This, however, is a 
misinterpretation of the position taken up by the writer. There is no need 
to assume that the framework, even if solid, is necessarily of an order of 
magnitude that would be visible under a microscope, and the existence of 
glass-like gels of gelatin, showing that they are free from particles of an order 
of magnitude greater than 10~® cm. in diameter, is against this. Adam [1921] 
has shown that films of palmitic acid on water only one molecule in thickness 
may either be in the solid or the liquid state. This proves that the solid state 
is not incompatible with structure of molecular dimensions. The writer does 
not consider that there is any fundamental difference between the “ molecular 
network” of Procter, the “exceedingly fine filamentous structure” of Laing 
and McBain, or the “solid framework” postulated by herself, though originally 
suggested by Hardy. (It is noteworthy that though Hardy made his gelatin 
frameworks of microscopic dimensions by the arbitrary use of alcohol, he 
explicitly states in his original paper [1900] that the droplets may be any size 


538 D. JORDAN LLOYD 


from macroscopic to sub-microscopic according to the conditions under which 
they separate.) The existence of structure in both gels and sols would seem 
to have been proved by work on the viscosity of melted gels [see P. van 
Schroeder, 1903; Garrett, 1903; Levites, 1908; Gokum, 1908; Schorr, 1911; 
and more recently Hess, 1922], and further the anisotropic properties de- 
monstrated by Hatschek [1921] are evidence that a part at any rate of a gel 
exhibits the characteristics of a solid. Bearing in mind the work of Adam 
already referred to, it does not seem impossible that the liquefaction and 
solidification of gels may be associated with a liquefaction and solidification 
of the framework similar to the liquefication and solidification of a film of 
palmitic acid one molecule in thickness. 

Procter considers gelatin gels to be of the nature of solid solutions, presumably 
similar to glass. Laing and McBain state that “a gel is identical with a sol 
_ except for its mechanical properties.” It is perhaps helpful to consider the 
matter in the light of the kinetic theory. According to this theory it is possible 
to visualise the solid state as one in which the molecules composing it are 
able to vibrate only round a mean position which is fixed, and the fluid state 
as one in which the molecules (or ions) composing it are free to migrate in 
any direction. If the molecules of solid are displaced from their mean position 
by the application of an external force, they will return towards it on the 
removal of the force, i.e, provided that the body has not been strained beyond 
its elastic limit. Molecules in a liquid displaced by an external force will not 
return to their previous position on the removal of the force. Gelatin gels are 
elastic solids with a modulus that has been determined [Leick, 1904], but the 
classical researches of Graham [1850, 1851] and Dumanski [1907] show that 
ions can migrate through them as rapidly as through water provided the gel 
be not too concentrated; 7.e. experiments on diffusion show that part of the 
gel at least is fluid and that in gels not too concentrated the fluid phase is 
continuous. Hardy’s observations on the reversal of continuity of phases in 
the alcohol-water-gelatin system, with increasing concentration of gelatin, 
have a bearing on this point. Laing and McBain have shown the similarity 
of gels and sols for other physical properties. The simplest theory harmonising 
all these facts does seem to be that in sols there is a meshwork of protein in 
the fluid state through which is an aqueous solution non-miscible with it; 
in gels the framework is not fluid but solid. The justification for this is to be 
found in the work of Adam on palmitic acid films, The extension of the theory 
to suppose that the framework was iso-electric gelatin, and the solution gelatin 
salts, led to the deduction that perfectly pure gelatin would be unable to form 
stable gels. This supposition has since been confirmed by the observations of 
the writer [Jordan Lloyd, 1920]; Field [1921] and Smith [1921]. 


PROPERTIES OF DIALYSED GELATIN 539 


SuMMARY. 


1. Gelatin purified by dialysis and precipitation in alcohol is only soluble 
with difficulty in boiling distilled water. On cooling, if the concentration be 
0:8 % or above, the solution sets to an opaque white gel; if the concentration 
be 0-7 % or less, the solutions on cooling become turbid fluids. The gels are 
not stable but show syneresis. The addition of sodium chloride in small 
quantities (0-4 N) decreases turbidity and increases gelling power, 0-6 % 
gelatin setting to a soft almost clear gel at 15° in 48 hours in the presence of 
0-1 N sodium chloride: at 1-0 N concentration sodium chloride decreases 
turbidity and decreases gelling power. 

2. In the presence of hydrochloric acid, gelatin becomes much more 
readily soluble. Sols and gels are transparent to the naked eye. Hydrochloric 
acid diminishes the gelling power, the degree of diminution passing through 
a maximum between the reactions py = 2 and py = 3, and again beyond | 
py = 0-7 at 15°. The addition of sodium chloride in small quantities (0-1 to 
1-0 N) increases turbidity and increases gelling power up to a py of 0-9. In 
stronger acid solutions the effect of salt is to decrease gelling power. The 
presence of the sodium chloride does not appreciably affect the reaction of 
the system. 

3. In the presence of caustic soda, gelatin is more soluble than in water. 
Sols and gels are transparent to the naked eye. Caustic soda slightly decreases 
the gelling power between py = 10 and py = 12 and completely prevents 
gelation at concentrations above py = 12. At concentrations py = 12 and 
above slow hydrolysis occurs. The addition of sodium chloride in small 
quantities (0-4 NV) increases turbidity and increases gelling power. In greater 
concentrations (1-0 NV) gelling power is decreased. The presence of the sodium 
chloride is without appreciable effect on the reaction of the system. 

4. Gelatin therefore can exist in the forms clear sol, clear gel, turbid sol, 
turbid gel, at the same temperature and the same concentration. 

5. Aqueous solutions of pure gelatin are not precipitated by the salts of 
the heavy metals. 

6. The rotatory power of gelatin is not a constant but varies (a) with 
temperature, (b) in acid, neutral or alkaline solutions. 

7. The bearing of the above properties on the theory of gelation is 
discussed. 


REFERENCES. 


Adam (1921). Proc. Roy. Soc. A, 99, 336. 

Dheré (1910). J. Physiol. Path. Gen. 18, 158, 167. 

Dheré and Gorgolewski (1913). Compt. Rend. 150, 934. J. Physiol. Path. Gen. 12, 646. 
Dumanski (1907). Zeittsch. physikal. Chem. 60, 553. 

Field (1921). J. Amer. Chem. Soc. 43, 667. 

Freundlich (1909). Kapillarchemie (Leipzig), 418. 

Garrett (1903). Phil. Mag. (6), 6, 374. 

Gokum (1908). Koll. Zeitsch. 3, 84. 


540 D. JORDAN LLOYD 


Graham (1850). Phil. Trans. 140, 1, 805. 

—— (1851). Phil. Trans. 144, 483. 
Hardy (1900). Proc. Roy. Soc. 66, 95. 

—— (1905). J. Physiol. 33, 251. 

Harned (1915). J. Amer. Chem. Soc. 37, 2460. 
Hatschek (1921). Trans. Faraday Soc. 16, Appendix I, 35. 
Hess (1922). Koll. Zeitsch. 27, 154. 

Holker (1921). Biochem. J. 15, 216. 

Leick (1904). Ann. Phys. 14, 139. 
Levites (1908). Koll. Zeitsch. 2, 210, 239. 
Laing and McBain (1920). J. Chem. Soc. 117, 1506. 
Lloyd (1920). Biochem. J. 14, 147, 584. 
Lloyd and Mayes (1922). Proc. Roy. Soc. B, 93, 69. 
Loeb (1919). J. Gen. Physiol. 1, 379, 487. 
Michaelis and Grineff (1912). Biochem. Zeitsch. 41, 373. 
Mérner (1889). Zeitsch. physiol Chem. 28, 471. 
Patten and Johnson (1919). J. Biol. Chem. 38, 178. 
Procter (1914). J. Chem. Soc. 105, 313. 
(1920). J. Soc. Leather Trades Chem. 4, 187. 
Schorr (1911). Biochem. Zeitsch. 37, 1911. 
Schroeder, P. v. (1903). Zeitsch. physikal. Chem. 45, 75. 
Smith (1919). J. Amer. Chem. Soc. 41, 135. 

—— (1921). J. Amer. Chem. Soc. 48, 1350. 
Trunkel (1910). Biochem. Zeitsch. 26, 493. 


LI. THE MECHANISM OF THE REVERSAL IN 
REACTION OF A MEDIUM WHICH TAKES PLACE 
DURING GROWTH OF B&B. DIPHTHERIAE. 


By CHARLES GEORGE LEWIS WOLF. 


From the John Bonnett Memorial Laboratory, Addenbrooke's Hospital, 
Cambridge. 


(Received June Ist, 1922.) 


A Report to the Medical Research Council. 


In the preparation of diphtheria toxin for immunising purposes it is known 
that the initial and final reactions of the medium are factors of prime import- 
ance in harvesting a toxin of high potency. The more recent investigations of 
L. Davis [1918], Bunker [1918, 1919] and Hartley [1922] have placed this 
beyond a doubt. Bunker, indeed, goes so far as to say that, outside certain 
well-defined final hydrogen ion concentrations, potent toxins are never 
harvested, and for this reason has suggested that a determination of the 
hydrogen ion concentration by modern methods is a criterion for ascertaining 
when the organism has elaborated the greatest amount of toxin. Hartley 
questions this and has been able to show that while Bunker’s statement is 
approximately correct, there are cases in which a very high grade antigen is 
produced outside the limits set by Bunker. 

For the purposes of the present paper it is only necessary to refer to the 
course which the medium takes when reaching this final reaction. 

If one starts with a bouillon peptone medium of py = 7-8, during the 
earlier period of growth the reaction becomes more acid, often exceeding the 
neutral point of py = 7-0. A sudden change in reaction occurs when this acid 
point is reached and the medium becomes alkaline. The alkalinity may reach 
the high value of py = 8-6 or even 8-8. It is during the period of alkalinity 
that the toxin is formed. 

The chemical reactions which occasion the change to the acid side of 
neutrality and then to the alkaline side with this bacillus have been little 
studied and not much is known about them. The acid production early re- 
ceived the attention of Theobald Smith [1895, 1899] who believed that it was 
closely connected with the presence of glucose in the bouillon. Later on he 
seems to have modified his view and to have attached more importance to 
the presence of inositol, the so-called muscle sugar, as a producer of acid, and 
in certain experiments on toxin production he sought to remove all fermentable 
carbohydrates by the preliminary treatment of the medium with B. coli or 


Bioch. xvr 36 


542 C. G. L. WOLF | 


yeast. It is noteworthy that he suggests that after this treatment a small 
amount of glucose should be added to the medium for the more efficient 
production of toxin. Smith’s experiments have had a very considerable in- 
fluence on the technique of toxin production. Even as late as 1921, Dernby 
and David [1921] have recommended the preliminary fermentation of media 
by yeast before proceeding to the inoculation with the diphtheria bacillus. 

Smith’s statements regarding glucose in the bouillon were based on amounts 
of gas obtained in a fermentation tube after sowing the bouillon with B. coli, 
and there does not seem to be any real information available as to the amount 
of glucose in meats as ordinarily purchased. In a subsequent paper I hope to 
be able to give some data on this point. For the present one may say that, 
assuming the amount of glucose to be 0-06 %, this would yield, if completely 
converted into carbon dioxide by fermentation, 0-088 °%% by weight of this gas. 
This expressed as gas per 100 g. of meat at room temperature and normal 
pressure would be about 47 cc. It is improbable that such a reaction occurs. 
One might also assume that a complete transformation of glucose to 
lactic acid took place. In this event, the amount of lactic acid produced 
would be equivalent to 0-66 cc. of normal lactic acid per 100 cc. of nutrient 
broth. Wolf and Harris (1917, 2) have shown that by adding this amount 
of lactic acid to broth the reaction shifted to the acid side to about the © 
same extent as is seen when B. diphtheriae acts upon peptone bouillon. As 
the amount of meat corresponding to the bouillon is at least half the quantity 
reckoned as above, it seems inevitable to conclude that the acids produced are 
not derived solely from the fermentation of the sugar contained in the 
medium, 

The only statement regarding the amount of inositol to be found in meat 
is an early one of O. Jacobsen [1871] who estimated that fresh muscle con- 
tained 0-035 %. In any case, so far as I am aware, no controls have ever been 
made to ascertain the effect of inositol added to a medium which has been 
fermented with yeast or with B. coli. 

That there are other views on the origin of the acid appears from the dis- 
cussion which took place between Madsen [1897] and Lubenau [1908]. The 
former author contended that the reversal of the medium was determined by 
the initial reaction. Lubenau asserted that with the absence of fermentable 
carbohydrates there was never any acid reaction, provided that the growth 
took place in the presence of air. Jacobsen [1911] attempted to reconcile these 
two points of view and tried to show that peptone also has a capacity for 
forming acids. Petruschky [1891] also’ believed that peptone furnishes material 
for the production of acid substances. 

The probability is that the truth lay rather with Madsen. With a high 
initial alkalinity and enough efficient buffer substances in the medium, despite 
the formation of acids, the reaction of the medium would never reach an acid 
point as determined with litmus. 

When it came to the question of reversal, the matter was on less certain 


REVERSAL IN REACTION OF CULTURE MEDIA 543 


ground. The obvious product in fermentation which can lead to an alkaline 
reaction is ammonia, and to this was attributed the change to an alkaline 
reaction. 

There were certain indications that this could not be the case. For example, 
Teruuchi and Hida [1912] examined the daily production of ammonia in 2 °%, 
bouillon peptone, with and without the addition of small quantities of glucose. 
While they obtained a small increase of ammonia, commencing on the fourth 
day, they did not find that the alkalinity bore any relation to the amount of 
ammonia produced. 

The more recent work of Ayers and Rupp [1918] gives a strong clue to the 
mechanism whereby the medium becomes alkaline. In their paper on the 
well-known reversion produced by organisms of the colon-aerogenes group it 
is shown that, with B. coli, acid and alkaline fermentations proceed simul- 
taneously. Glucose is converted to organic acids and these, in turn, are 
transformed to carbonates. They suggest that the reversion of reaction is 
probably due to bicarbonates. A similar simultaneous acid and alkaline fer- 
mentation has been analysed by Harris and the writer [1917, 1] in the case of 
strongly proteolytic anaerobic organisms. Here the stabilisation of the reaction, 
in spite of the production of large amounts of organic acids, is due in great 
_ part to the production of ammonia. It seemed, therefore, of some importance 
to ascertain to what the reversal in reaction with B. diphtheriae was due. 

The first experiment was designed to follow out the changes ih hydrogen 
ion concentration, and volatile acid and ammonia production. For this purpose 
portions of 100 ce. of a standard bouillon-peptone medium were sterilised in 
450 cc. rectangular medicine bottles. These were incubated lying on their sides, 
so that the surface of medium exposed to the air was about 115 sq.cm. At 
stated times a bottle was taken from the incubator and the hydrogen ion 
concentration, volatile acids and ammonia determined. Table I gives the 
results of the experiment. 


Table I. 


Mgm. ammonia Volatile acids, 
nitrogen per ce. N/10 alkali 


Hours pu 100 ce. per 100 ce. 
Control 75 6-13 4-4 
24 7-0 7:85 21-2 
48 71 8-40 24-8 
72 7-9 8-98 8-8 
96 8-4 8-52 7-2 
120 8-5 8-52 6-8 
144 8-5 8-42 7-2 
168 8-5 8-98 78 
192 8-4 9-25 8-0 
216 8-5 11-20 8-0 
240 8-5 10-07 8-0 
264 8-5 8-12 6-0 


Certain quite clear results emerge from the analyses. Starting with an 
initial hydrogen ion concentration of py = 7-5 the reaction follows the usual 
course, becoming as acid as py = 6-9, When reversion took place, a final 


36—2 


544 C. G. L. WOLF 


value of pq = 8-5 was reached which persisted as long as the experiment was 
continued. The amount of ammonia produced is definite but very small. 
Starting from an initial value of 6-13 mgm. ammonia N per 100 cc. of medium 
it does, on one occasion, reach a concentration of 11-2 mgm., but the average 
is somewhat less than this. This experiment confirms what Teruuchi and Hida 
found, viz. that the amount of ammonia formed during the metabolism of the 
bacillus is very small indeed. At the same time a very remarkable change 
takes place in the volatile acid production. Starting with an amount of acid 
equivalent to 4-4 cc. of N/10 NaOH per 100 cc. of medium, in the first 24 hours 
the acid had risen to 21-2 cc. with-a further rise to 24-8 cc. at the end of 
48 hours. With the sudden reversal of reaction from py = 7-1 to 7-9 a large 
proportion of the acid disappears and in the fourth sample only 8-8 cc. are 
found. From this time on the amount of acid rises only slightly si a final 
alkalinity of the-medium of py = 8-5. 

It is thus obvious that a large part of the volatile organic acids formed in 
the first stages of fermentation are utilised in the sense of Ayers and Rupp to 
form basic substances. 

A second experiment was undertaken to ascertain to what extent the 
fixed acids took part in the changes above referred to. The results are embodied 
in Table II. 


Table II. 


' Volatile acids Fixed acids 
CO,, ce. per N/10 alkali N/10 alkali 


Hours Pu 100 ce per 100 ce per 100 ce. 
Control 7:8 3-0 3-6 27:2 
24 75 — 4-4 20-5 
48 7:3 _— 14-8 6-6 
72 7-4 — 15-4 lll 
96 75 15:8 74 
120 8-1 18-0 7:2 4:5 
144 8-2 _ 6:8 4-1 
192 8-5 27-0 7:2 2-9 
240 8-5 27-0 7:4 3°5 


The initial alkalinity of the medium was greater than in Exp. | and 
possibly for this reason the medium did not become so acid. The final reaction 
was, however, the same as in the previous test. 

As the amount of medium was not sufficient, no analyses for ammonia were 
performed. The general trend of the volatile acid curve is the same as in the 
preceding experiment; one finds a sudden increase in the concentration of 
these compounds corresponding with a decrease in py. The increase in alkalinity 
is coincident with the sudden lowering of the volatile acids. The course of the 
fixed acid curve is less easy to explain. From the time of inoculation there is 
a steady fall, only one value—that obtained at 72 hours—being outside the 
curve. 

The main non-volatile organic acid in bouillon is undoubtedly lactic acid, 
and this appears to have been metabolised, apparently during the earlier hours 
of the fermentation, to substances still having an acid reaction and it may well 


REVERSAL IN REACTION OF CULTURE MEDIA 545 


be that some of these are represented in the increase to be found in the volatile 
acids. Finally, however, the fixed acids reach a level represented by about 
4-0 mgm. of N/10 alkali per 100 cc. of medium. 

During the course of this experiment I was led to ascertain whether the 
available carbon dioxide in an old culture was different from that in a medium 
which had not been inoculated. For this work I employed the van Slyke 
apparatus, using 1-0 cc. of medium for a determination. It appeared at once 
that the amount in the former was over five times that of the latter. 

Using the present series of cultures it was found that at the 120th hour 
an amount of carbon dioxide was disengaged by dilute sulphuric acid equivalent 
to 20-0 cc. of gas per 100 cc. of medium. This was at a time when the reaction 
of the medium was py = 8-1. When the fluid reached an alkalinity of py = 8-5 
the amount of carbon dioxide disengaged was 29-0 cc. per 100 cc. of medium. 
The amount of carbon dioxide disengaged from the control medium was only 
3-0 ce., and of this 2-0 cc. were due to gases dissolved in the reagents used for 
the determination. 

It would thus appear that the reversal of reaction was closely connected 
with the formation of carbonates, confirming what Ayers and Rupp had found 
with B. aerogenes. The matter was therefore studied from the commencement 
of fermentation. 

At the same time it was decided to obtain some information regarding the 
action of B. diphtheriae on certain well-known organic salts. Preliminary 
experiments performed some time previously made it appear that certain 
organic acids added to the medium would be available as a source of energy, 
and a detailed examination of this point was made. Not a great deal of 
information is available on the latter point. 


Table ITT. 

Control Sodium succinate Sodium tartrate 
Hours _ Pellicle pu Co, Pellicle pu CO, Pellicle pu co, 
0 78 1 78 1 78 1 
48 +++ 76 10 +++ 8-0 23 +++ 7:3 9 
72 +++ 8-1 9 +++ 8-1 21 +++ 8-0 17 
120 +++ 8-3 23 +++ 8-6 57 +++ 8-2 19 
168 8:3 34 8-6 65 7:9 12 
240 7:3? 5? +++ 8-5 71 +++ 8-0 ll 
336 8:3 27 8-6 75 8-2 22 
456 8-5 27 8-8 90 8-5 28 
624 8:4 21 8-6 56 8-3 13 

Sodium malate Sodium citrate Sodium formate 
Hours __—iPellicle pu CO, Pellicle pu co, Pellicle pu CO, 
0 7:8 1 7:8 1 7:8 1 
48 +++ 77 29 0 7:3 4 +++ 7-5 10 
72 +++ 8-3 50 0 7-2 5 +++ 75 16 
120 +++ 8-6 75 0 7-3 7 +++ 7-9 20 
168 8-6 66 7-5 7 8-3 Wi 
240 +++ 8:1? 20? 0 75 6 +++ 8-3 20 
336 8-8 90 8-3 21 8-4 26 
456 8-8 92 8-3 23 8-4 26 
624. 8-6 71 7-4 3 8:3 19 


546 Cc. G. L. WOLF 


Maassen [1896] many years ago in an elaborate investigation of the action 
of bacteria on organic salts found in a series of 21 organic acids that malic 
acid was the only one attacked by B. diphtheriae and that very feebly. 

Altobelli [1914] in a somewhat recent paper also states that this acid is 
the only one attacked by B. diphtheriae. 

In the present experiments the fermentations were carried out in test- 
tubes, the quantity of medium used being about 10 ce. 

The salts were added in a concentration of about 1-0 %. After the addition 
of the salts the medium was adjusted to py = 7-8 and sterilised at 120° for 
30 minutes. The tubes were all inoculated from the same culture and kept at 
37°. The culture used (Park and Williams, No. 8) had been transferred at 
24-hour intervals in bouillon peptone and was growing very actively (Table III). 

It is clear that various organic acids are utilised in quite a special way by 
B. diphtheriae. The two which are most strongly attacked are.malic and 
succinic acids. Judged by the amount of carbonate formed, succinic acid is 
nearly as completely used as malic acid and it is difficult to understand how 
Maassen overlooked its effect on metabolism. Tartaric acid and formic acid do 
not seem to have been utilised at all, for the amounts of carbonates formed are 
very close to those given by the control medium. Citric acid appears to have 
had an inhibiting effect on the growth of the bacillus. Reversal was very 
much delayed and the alkalinity did not reach the height obtained with other 
cultures. The carbonate fermentation was below that of the control medium. 

A special interest attaches to formic acid, as this compound has been shown 
to be one of the steps through which the higher acids are finally oxidised to 
carbon dioxide. In the katabolism of glucose by B. coli the intermediary 
between this compound and carbon dioxide is said to be formic acid [Harden 
and Penfold, 1912]. This does not appear to be the case with the diphtheria 
bacillus, for the addition of formic acid to the medium is not associated with 
an increase in carbon dioxide above that found in the control medium. 


SUMMARY. 


The reversal of reaction in carbohydrate-free media caused by the growth 
of B. diphtheriae is due, in its acid phase, in part to the production of volatile 
acids. When the medium reverses to a more alkaline reaction, these acids are 
converted into carbonates. 

Certain organic acids, such as malic and succinic acids, are also utilised to 
produce carbonates. 

The formation of carbonates from the acids examined does not appear to 
take place with formic acid as an intermediary product, 


REVERSAL IN REACTION OF CULTURE MEDIA 547 


REFERENCES. 


Altobelli (1914). Ati Soc. Toscan. d’Igiene, quoted from Centr. Bakt. 1916, 64, 188. 
Ayers and Rupp (1918). J. Inf. Dis. 23, 188. 
Bunker (1918). Abstr. Bacteriol. 2, 10. 
—— (1919). J. Bacteriol. 4, 379. 
Davis, L. (1918). J. Lab. Clin. Med. 3, 358. 
Dernby and David (1921). J. Pathol. Bact. 24, 180. 
Harden and Penfold (1912). Proc. Roy. Soc. B, 85, 417. 
Hartley, P. (1922). In press. 
Jacobsen, O. (1871). Annalen, 157, 227. 
Jacobsen, K. A. (1911). Centr. Bakt., Orig. I, 57, 17. 
Lubenau (1908). Arch. Hyg. 66, 305. 
Maassen, A. (1896). Arb. kaiserl. Gesundhmt. 12, 340. 
Madsen (1897). Zeitsch. Hyg. 26, 157. 
Petruschky, J. (1891). Bawmgarten’s Jahresb. 6, 488 (footnote). 
Smith (1895). Centr. Bakt. 18, 1. 
—— (1899). J. Expt. Med. 4, 373. 
Teruuchi and Hida (1912). Saikingashi, Nos. 205 and 768, quoted from Biochem, Zentr. 1913, 14, 
748, 749. 
Wolf and Harris (1917, 1). J. Pathol. Bacteriol. 21, 386. 
Wolf and Harris (1917, 2). Biochem. J. 11, 218. 


LII. THE PROPERTIES OF DIBENZOYLCYSTINE. 


By CHARLES GEORGE LEWIS WOLF! 
anp ERIC KEIGHTLEY RIDEAL. 


From the John Bonnett Memorial Laboratory, Addenbrooke's Hospital, 
Cambridge, and the Chemical Laboratory, Cambridge University. 


(Received June 20th, 1922.) 


NearLy all the organic substances exhibiting gel formation in water are 
compounds of complex structure and high molecular weight. 

In order to elucidate the mechanism of gel formation, in recent years 
search has been made for substances of relatively low molecular weight and 
simple constitution, and the dye-stuffs, benzo-purpurin, chrysophenone, 
thionin blue, the soaps, the organic sulphonic acids and camphorylphenylthio- 
semicarbazide have been used. As Gortner and Hoffman have recently 
pointed out [1921], benzoylcystine is the simplest known type of organic 
substance which exhibits gelation with water and it therefore appeared to 
present a suitable material with which to study some of the elementary 
properties of gels in greater detail. 

The preparation of dibenzoylcystine used in this work was essentially that 
of Brenzinger [1892] and Gortner and Hoffman [1921], except that the sub- 
stance, after recrystallisation from dilute alcohol, was extracted for several 
hours with benzene in a Soxhlet apparatus to remove completely all traces of 
benzoic acid and benzoyl chloride. The melting-point of our product was 189° 
(uncorrected) which is some eight degrees higher than that given by the older 
authors. 

While the method of dispersion in water from an alcoholic solution is well 
known, yet for many purposes the presence of a second component is un- 
desirable, and accordingly attempts were made to obtain direct dispersion of 
the material in water. It was found that aquagels may be prepared by 
triturating the solid in small quantities at a time with warm water and com- 
pleting peptonisation by stirring the solution for long periods at a temperature 
of 80°-90°. The rigidity of the gels obtained by this method was essentially 
of the same order as that of those described by Gortner and Hoffman using 
alcohol for dispersion. 

The sol-gel transformation was found to be perfectly thermally reversible, 
thus differing from that of silicic acid, and decomposition does not appear to 
occur as in the case of gelatin. Dilute gels are perfectly transparent; but at 
higher concentrations, 0-2 ° and over, an opalescence makes its appearance, 
increasing with the amount present in the liquid, Dilute gels, on standing, 


* Working under a grant from the Medical Research Council, 


THE PROPERTIES OF DIBENZOYLCYSTINE 549 


become somewhat cloudy. On prolonged standing syneresis occurs with con- 
traction of the gel to a core and the exudation of water. Simultaneously small 
nuclei make their appearance in the gel, which continue to grow to form 
spherites. Microscopic examination of the spherites shows them to consist of 
minute elongated radiating crystals. The appearance of these crystalline 
spherites in the gel, differing from the shot-like spheroids formed in the gelation 
of gelatin [ Bradford, 1921], lends additional support to the two-phase theory 
of gel structure. The thread or fibrillar nature of the spherite group, in addition, 
supports the assumption that the natural crystal growth of dibenzoylcystine 
in water is in the form of poly-molecular crystalline rods, the gel consisting 
of a number of intersecting fibrils built up of these rods in which the water is 
enclosed. The fibrils in the gel are evidently relatively coarse, since the gel is 
both moist and elastic and exhibits syneresis. 

The examination of the spherites thus classifies dibenzoylcystine as be- 
longing to the fibrillary crystalline gels, which include nearly all the gels 
which have been examined, with the possible exception of dilute gelatin, agar 
and silica. These latter, although generally regarded as a class apart, may be 
brought into the same category, if one assumes, in the disperse phase, the 
existence of highly elastic or mobile fibrils undergoing a process analogous to 
crystal twinning. 

Dibenzoyleystine is a relatively strong acid, a gel containing 2-66 g. per 
litre showing a py of 3-05 at 20°, the value being obtained both by indicators 
and the hydrogen electrode. This value, if the acid be considered as mono- 
basic, corresponds to a dissociation constant of 0-3 x 10-%. The degree of 
dissociation was also determined by the conductivity method. The molecular 
conductivities of a 0-1 % solution and solutions more dilute were determined 
at 25° in a conductivity cell of the usual type, with the aid of a potentiometer, 
telephone and coil. The value of A oc was found to be 250, a limit to which the 
molecular conductivity had almost attained at a dilution of 0-00014 molar. 
The value of K, the dissociation constant, varied from 1-4 x 10-8 to 1-5 x 10-8 
with a mean value of 1:48 x 10-8. The acid is thus comparable to mono- - 
chloroacetic acid and monobromosuccinic acid. No discontinuity in con- 
ductivity in the 0-1 % solution, which exhibits a feeble tendency to set, was 
observed, whether the liquid was agitated or kept at rest prior to the determina- 
tion. This is a phenomenon which has been observed by McBain [1919] and 
his co-workers in their studies of soap solutions. The discrepancy between the 
dissociation constant value as calculated from the hydrogen ion concentration 
and determined by the conductivity method is probably to be attributed to a 
liquid junction potential in the former due to the non-diffusibility of the anion 
in the dibenzoyleystine. This discrepancy is of the order of 10 millivolts. 

On the alteration of the hydrogen ion concentration of the medium by the 
addition of sodium hydroxide or hydrochloric acid several interesting points 
were noted. On the addition of increasing amounts of alkali the gel formation 
gradually became less, until the contents of the tube were entirely liquid. 


550 . ¢ G. L. WOLF AND E. K. RIDEAL 


On the addition of acid, the rate of setting increased and the gels were dis- 
tinctly more opalescent in character, while on the addition of relatively large 
quantities of acid the gels became distinctly granular and less resistant to 
penetration. The gel possessed the maximum stability at the isoelectric point. 
The marked change in the physical characteristics on the alteration of the py 
of the solution was made evident by the following experiments. 

Water Retention. One gram of dibenzoyleystine was dissolved in 6-0 cc. of 
hot alcohol and poured into 250 ce. of boiling water as quickly as possible with 
constant stirring. From this stock solution 20 cc. were run into boiling tubes 
which contained 10 cc. of solutions of alkali and acid of varying concentration 
as given in Table I. 20 cc. of the mixtures were then pipetted into Petri dishes 
for determinations of their setting power. The remaining 10 cc. were used for 
hydrogen determinations. At the end of 72 hours the jellies contained in the 

Petri dishes were removed and allowed to drain under their own grayity on 
filter papers in funnels. The liquid running through was measured. The results 
of these determinations are given in Table I. 


Table I. 0-30 % dibenzoyleystine. 


Water drained 
in 2 hours 
from 20 ce. 
Dibenzoyl- of mixture 
N/20 N/10 stine after setting Observations made on 
No. H,SO, NaOH Water solution p,, 20°C. for 3 days mixing and cooling 
ce. ce. ce ce. ce. 
ee 2 8 20 3-65 16-5 Quite fluid 
2 0 1 8 20 3-48 16-5 ” ” 
3 «(0 0-5 9-5 20 3-36 15-5 Moderately set, not very opalescent 
4 0 0-25 9-75 20 3-26 13-0 Not quite stiff, pours 
5 0 0 10 20 3-05 13-0 Stiff, does not pour 
6 05 0 9-5 20 12-65 13-0 ir stiff 
’ epee BY 0 9-0 20 2-59 14:0 of air bubbles, set at once 
8 15 0 8:5 20 | 2-52 13:0 is * * 
. 20 0 8-0 20 2-41 13-0 Opalescent, set at once 
10 30 0 7-0 20 2-29 11-0 a mh 
a8 4) 6-0 20 2-19 11-5 Very opalescent 
12 80 0 2-0 20 1-93 11-5 Solid had separated out 


It is evident as the hydrogen ion concentration decreases below py 3-0 the 
water-retaining power of the gel is very seriously diminished. In the more acid ° 
solutions a slight alteration takes place but the curve of drainage is not so 
regular. 

In the determination of the py of the alkaline solutions it was noticed that 
the internal resistance of the cell became extremely high, possibly owing to the 
formation of a pellicle at the interface of the solution and the saturated 
potassium chloride junction fluid. It was not, however, sufficient to interfere 
with the readings. 

Effect of Electrolytes. In common with other gels, the water-retaining power 
of the fibrils is much affected by the presence of neutral electrolytes. The effect 
of alteration of the osmotic pressure and the reduction in the free water content 
and of ion hydration naturally varies with the nature of the added electrolyte. 
Quantitative investigations on the effect of the addition of ammonium thio- 
cyanate to a typical emulsoid albumin to gelatin and to silicon hydroxide gels 
are to be found in the literature, 


THE PROPERTIES OF DIBENZOYLCYSTINE 551 


Ammonium thiocyanate exerts a similar lyotropic effect on the dibenzoyl- 
cystine gels. Syneresis is more pronounced in the presence of small quantities 
of this salt and relative instability of the gel structure may be obtained in 
fairly dilute solutions as is indicated by the following draining experiments 
performed in a manner similar to those previously described. 


Table II. Effect of Ammonium Thiocyanate on the setting of Dibenzoyleystine. 


- 0-4 % dispersion of dibenzoylcystine. 
5-0 % solution of ammonium thiocyanate. 
Ammonium Dibenzoyleystine 
thiocyanate Water solution Filtrate cc. in 
Number ce. ce. ce, 1 hour 

l 1 9 20 14-5 
2 2 8 20 17-5 
3 3 7 20 18-0 
4 4 6 20 19-0 
5 6 4 20 21-0 
6 8 2 20 21-0 
7 10 0 20 21-0 
8 0 10 20 14-0 


Nos. 1 and 8 went solid in about 10 minutes. The others some time afterwards. 


Resistance to Penetration. In order to obtain some information on the 
tenacity of the gels as affected by change in the py of the dispersion medium 
a simple penetration test was devised. This consisted essentially in measuring 
the rate of fall of a hemispherically ended plunger through a column of the 
gel maintained at room temperature. It was noted that a uniform velocity 
was attained after a fall of two or three centimetres. The time in seconds for 
a penetration of 1 cm. is plotted in the following curve. 

100 


BOfr- 


60;- 


40 


Penetration coefficient seconds per cm. 


20fF- 


to 


oF - -—--—--—--_--—-_ wa 


0 


1 


> 
cee. N NaOH Pure gel. ce. N HCl per 10 ce. gel. 
per 10 ce. gel, 


552 C. G. L. WOLF AND E. K. RIDEAL 


It will be noticed that the influence of the alkali is extremely marked, 
whilst a distinct reduction in the internal viscosity is produced in relatively 
acid solutions. The conclusion that qne may draw from these experiments is 
that on the addition of alkali the number of fibrils decreases, presumably 
because the sodium salt has no gelatinising qualities and is entirely soluble, 
while, on the addition of acid, the fibrils become relatively much coarser and 
consequently fewer in number. 

Protective Action. An attempt was made to determine the protective power 
of dibenzoylcystine, using red colloidal gold prepared according to Zsigmondy’s 
method. This gold gave correct values when tested against gelatin. Owing to 
the sparing solubility of the compound, it was impossible to carry out the test 
in exactly the way which Zsigmondy [1920] prescribes. 

With a 0-2 % solution of the compound partial protection of 10 cc. of gold 

_suspension occurred with 4-0 cc. of the gel. Complete protection was afforded 
by 6-0 ce. The gold number should therefore be about 10, which is close to 
that of dextrin and is nearly a negligible quantity. 

Diffusion Experiments. In dilute concentrations gels such as gelatin and 
agar offer no abnormal resistance to the diffusion of electrolytes. With more 
concentrated gels the coefficient of diffusion is somewhat reduced. Attempts 
to estimate the fibril mesh of such gels by observations on the rate of diffusion 
of electrolytes, semi-colloids and colloids of varying particle diameter have 
frequently been made. The results, however, have been extremely variable. 
The same abnormalities have been noticed with dibenzoylceystine which being 
of much simpler chemical constitution permitted partial interpretation of the 
results obtained. 

The experiments were made by taking a gel of 0-2 % and allowing it to 
set in tubes of uniform size. As soon as the gel was fairly stiff, solutions of 
dyes were carefully pipetted on top of it and the level marked. At intervals 
the entrance of the dye into the gels was estimated with calipers. All the dyes 
used were of Griibler’s make, The results are given in Table ITI. 

It will be noted that the gel offers very slight resistance to an electrolyte 
such as potassium dichromate, a distance of 100 mm. being traversed in a few 
hours. 

Dibenzoyleystine is a negative colloid, being itself a relatively strong acid, 
and although the ion, as evidenced by the highly dissociated sodium salt, 
possesses feeble if any gelatinising properties, yet there is no doubt that the 
fibrils are partially ionised. In the presence of positive crystalloid or colloid 
dyes adsorption or mutual precipitation is to be anticipated. For such basic 
dyes, diffusion far into the interior is not to be looked for, and the fibrils 
should be markedly stained, the partition coefficient between fibril and sur- 
rounding liquid being very high. The low apparent diffusibility is for these 
dyes to be ascribed to adsorption and not to a small fibrillary mesh. The zone 
of precipitation naturally varies with the colloidal character of the dye and 
ite diffusibility. 


‘THE PROPERTIES OF DIBENZOYLCYSTINE 553 


Table IIT. 
Diffusion rate 
Diffusion of Nature Adsorption Precipitation mm. in 24 hrs. 
Potassium dichromate Crystalloid - _ 135 
Basic dyes, cationic ; ; 
Night blue Colloid x x x x 0 
Neutral red Semi-colloid ees x 0-2 
Methyl violet 58 ye Pe he x x 1-0 
Bismarck brown Crystalloid se - 1-6 
Toluidine blue Colloid x 2 x 2-0 
Methylene blue mn x x 2-0 
Malachite green x - 2-0 
Thionin x x - 2-4 
Safranin Crystalloid x x - 3-0 
Fuchsin Semi-colloid - - 3-0 
Chemically reacting dyes: 
Eosin Colloid x x 336 0-2 
Rose bengale a x x x x 1-0 
Acid dyes, anionic : 
Azo blue Pes x= - 1-1 
Congo red ” x= - 1-6 
Benzopurpurin P x= x 1-6 
Alkali blue Pe - - 2-0 
Nigrosin - - 2-4 
Carminic acid - - 3:3 
Lewd blue zs - - 4-2 
ight n F.S. - - 4-4 
ree fuchsin Semi-colloid - - 5-0 


POSSIBLE STRUCTURE OF THE GEL FIBRIL. 


Evidence is collecting to show that the union of molecules to form molecular 
complexes is brought about by forces identical with, though generally more 
feeble in character than, those uniting atom to atom within the molecule; 
further that these residual force fields, the product of these internal molecular 
asymmetries, are localised within certain definite areas in the molecule itself. 
These forces are electrical in character, and for convenience the various residual 
affinities may be designated as positive or negative relative to other such fields. 
The active areas on the molecules are associated with the presence of specific 
chemical groups which are electrically unsaturated. 

The results on the diffusion of the dyes indicate that mutual precipitations 
of the cationic dyes and the anionic gel do actually take place, and that with 
the more colloidal dyes the precipitation zone is more limited and the colour 
more intense, confirming the work of Biltz and Teague and Buxton. 

The halide-containing dyes, eosin and rose bengale, appear to react 
chemically with the dibenzoylcystine. But feeble, if any, adsorption was 
observed in the case of the anionic dyes. Those that were distinctly colloidal 
in character were feebly adsorbed. Benzopurpurin, a dye very similar to 
electrolytes, was apparently relatively easily precipitated. In general, the 
order of diffusibility observed with dyes and benzoylcystine was very similar 
to that observed by Bechold and others [1906, 1919]. 

Some idea of the nature of the linkages involved in the gel fibril may be 
obtained from a survey of the effect of substitution in dibenzoylcystine of 


various groups on the gel formation. 


554 C. G. L. WOLF AND E. K. RIDEAL 


These may be: 

1. Replacement of the carboxylic hydrogen by a more electropositive 
element. For example, Na destroys the gel-forming properties. 

2. Replacement of the electronegative benzoyl group attached to the 
amino nitrogen by an electropositive group destroys the gel-forming power of 
the substance. Attempts to prepare the cinnamy] derivative, which should be 
more electronegative than the hydrogen or the acetyl, have not succeeded. 
The electronegative di-m-nitrobenzoylcystine has been prepared by Dr F. W. 
Dootson and exhibits the characteristic gelatinising qualities of the simple 
benzoyl compound. : 

3. Replacement of the —S—S— linkage by groups such as —CH,—CH,— 
or —CH = CH— results in a loss of gelatinising properties. Gel formation 
thus appears to be dependent on the presence of an electronegative group 

_attached to the amino nitrogen, which must not be too polar in character, on 
the presence of a relatively negative carboxyl group and on the presence of 
the electropositive —S—S— grouping. Evidence for the unsaturated and 
electropositive character of the —S—S-— group is furnished by the interaction 
of cystine with halide aromatic substances in the body, e.g. the formation of 
bromophenylmercapturic acid from bromobenzene. Linkages of this double 
character would result in the formation of molecular aggregates arranged in 
echelon permitting the growth of needle-like or fibrous crystals, as shown in 
the following suggested formula. 

Formula. 


OH ah OH 
LeJ rege he 5 be id reget exe | 
is ie eae Antec bering a ghee ate avai v= 
ieee fe ae { 
1 pee ee ge cca ey ae H 
| ‘,H,CO (-) | (-) C.H,CO 
‘H, SH, OH 
4 { | 
rae Bie tes ee 


§ t+**(-) 8,00 


| 4  9ter 65) (ee Ve 0 C “eee eee C,H,CO.HN 1 


HF ceccceee Pa eae een Men Ciera er Eye Wi Sm re 
a x i H 
( ‘ 
“y [ 2 | 2 
HC-N HONHC,H,CO ete. 
Ont d.u,co OH -C 


| 
) 0 


THE PROPERTIES OF DIBEN ZOYLCYSTINE 555 


SUMMARY. 


Gels of dibenzoyleystine may be prepared by peptisation with hot water. 
The substance is a relatively strong acid with a dissociation constant of 
1-49 x 10-8. The gel structure appears to be fibrillary and relatively coarse. 
The sodium salt of the acid exhibits no gelatinising properties. The presence 
of acids greatly reduces the solubility of the compound. This was confirmed 
by observations on the water-retaining capacity and internal viscosity of the 
gel. Lyotropic salts, such as ammonium thiocyanate, reduce the water- 
retaining power of the gel, producing ultimate liquefaction. Cationic dyes are 
adsorbed and precipitated by the gel. Anionic dyes diffuse in a normal manner, 
whilst halogen dyes apparently react chemically with the sulphur group. Di- 
m-nitrobenzoyleystine has similar properties to the simple compound. A 
possible structure for the gel fibril based upon the effect of chemical substitu- 
tion on the gel formation is suggested. 


REFERENCES. 


Bechold (1919). Colloids in Biology and Medicine, 427. 

Bechold and Ziegler (1906). Zeitsch. physikal. Chem. 56, 105. 
Bradford (1921). Koll. Zeitsch. 28, 214 and Biochem. J. 15, 553. 
Brenzinger (1892). Zeitsch. physiol. Chem. 16, 573. 

Gortner and Hoffman (1921). J. Amer. Chem. Soc. 43, 2199. 
McBain, Laing and Titley (1919). J. Chem. Soc. 115, 1280. 
Zsigmondy (1920). Kolloidchemie, 174. 


LIII. THE NITROGEN-DISTRIBUTION IN BENCE- 

JONES’ PROTEIN, WITH A NOTE UPON A NEW 

COLORIMETRIC METHOD FOR TRYPTOPHAN- 
ESTIMATION IN PROTEIN. 


By ERY LUSCHER. 
From the Biochemical Laboratory, Cambridge. — 
(Received June 6th, 1922.) 


THE physical properties of Bence-Jones’ protein have been more or less fully 
described in all investigated cases. On the other hand, there is but little 
reference to the exact chemical constitution and especially to the amount of 
individual amino-acids. Magnus-Levy [1900] determined the N-balance by 
Haussmann’s method [1899]. Abderhalden and Rostoski [1905] estimated the 
monoamino-acids by esterification and Hopkins and Savory [1911] all the 
amino-acids by a similar method. This last investigation was carried out in two 
different cases and it is to one of these that the protein belongs that I have 
used for the analysis described in this paper. I determined the nitrogen 
distribution by van Slyke’s method and the tryptophan content by the colori- 
metric estimation of von Fiirth and Lieben [1920]. The details of the analysis 
are described later. For the clinical symptoms of this case and the physical 
properties of the protein the original paper by Hopkins and Savory [1911] 
must be consulted in which the case is described as Case C. 

Table I gives the results of the different analyses carried out on Bence- 
Jones’ protein. 


Table I. 
The figures are % of total nitrogen. 
Hopkins 
Magnus- Abder- — A = Present 
Levy halden Case A Case C analysis 
Amide-N 8-6 ‘ 8-02 8-00 9-43 
Melanin-N ‘ - : ; : 
Total-N of bases 24-7 ‘ ‘ ; 23-11 
Cystine-N ° 3 0-41 1:25 
Arginine-N + 11-9 12-0 9:27 
Histidine-N 4 1-44 1:33 4:54 
Lysine-N : + ‘ 4:34 8:04 
Total-N of filtrate 64-2 ° ‘ ‘ 66-84 
Amino-N of filtrate ; ; : . 61-69 
Non-amino-N of filtrate ; ‘ . . 515 
Glycine-N ‘ 1-95 4 + : 
Alanine-N , 4:37 4 + 
Valine-N ‘ ; ‘ 4:22 
Leucine-N ; 6-98 ‘ 4:29 
Aspartic acid-N ° 2-89 134 1-41 
Glutamic acid-N . 3:52 4-42 4:73 
Phenylalanine-N ° 0-78 2-47 2:57 
Tyrosine-N ‘ O81 1-97 2-02 
Proline-N ; 1-40 1-93 2-00 
Oxyproline-N ‘ - - - : 
Tryptophan-N 0-69 ° 2-44 


NITROGEN-DISTRIBUTION IN BENCE-JONES’ PROTEIN 557 


In the papers of Abderhalden [1905] and Hopkins [1911] the amounts of 
amino-acids are given in percentage of the protein. Assuming the total 
nitrogen of the protein to be 16-21 %, a value found by Hopkins, it is possible 
to calculate the nitrogen of each amino-acid in percentage of the total nitrogen. 
The above table is expressed according to the latter. 

It is a matter of course that only those estimations are directly comparable 
which have been performed by the same method. This holds good in the two 
cases examined by Hopkins [1911], which show almost identical values. More- 
over Hausmann’s estimation of the nitrogen distribution and that by van 
Slyke are based on the same principle, so that those cases may also be com- 
pared. They also show a close agreement. This indicates that in the two cases 
examined by Hopkins [1911] and in the case of Magnus-Levy [1900] the protein 
shows the same relative composition of mono- and diamino-acids. The assump- 
tion that these three cases have so far the same chemical constitution by mere 
chance is not probable. It seems more likely that in all cases of Bence-Jones’ 
protein-uria there appears one and the same protein or mixture of proteins in 
the urine. Abderhalden [1905] found in his case somewhat different values 
from those of Hopkins in spite of a similar method. But esterification is a very 
complicated method and does not give quantitative results. 

Table II shows the relations between Bence-Jones’ protein and those of 
serum and tissue. 


Table IT. 
The figures are % of total nitrogen. 
Bence-Jones’ Serum- Serum- Normal 

protein globulin? albumin* tissues* Tumors? 
Amide-N 9-43 7:95 6-65 5-82 4:73 
Melanin-N 0-90 2-30 0-95 3-09 2-42 
Total-N of bases 23-11 25-25 34-75 30-71 30-54 
Cystine-N 1-25 1-65 3-10 0-93 1-12 

inine-N 9-27 8-00 9-90 11-13 10-09 

Histidine-N 4-54 4:80 5-85 7:86 7-09 
Lysine-N 8-04 10-80 15-90 9-83 12-04 
Total-N of filtrate 66-84 64-80 58-15 60-82 62-14 
Amino-N of filtrate 61-69 62-65 56-55 55-88 54-24 
Non-amino-N of filtrate 515 2-15 1-60 4:94 7:90 
N-recovered 100-28 100-50 100-30 100-4 99-8 
Tryptophan-N 2-44 . 2-5-3-55 _— —_ — 


1 and 2 carried out by Hartley [1914]. 
3 and ‘ carried out by Drummond [1916]. 
5 carried out by von Fiirth and Lieben [1920]. 


Hartley’s [1914] analysis and that described in the present paper show the 
same nitrogen distribution for mono- and diamino-acids in Bence-Jones’ pro- 
tein and serum-globulin. The serum-albumin is characterised by a higher 
proportion of diamino-acids. Moreover, the tryptophan content is almost 
identical in serum-globulin and Bence-Jones’ protein. On the other hand, the 
latter as also the serum-albumin yields less melanin. But this fraction depends 
much on the hydrolysis and is therefore the most uncertain. The only real 
difference between serum-globulin and Bence-Jones’ protein has been found 
in the non-amino-nitrogen of the filtrate, i.e. the proline and oxyproline 


Bioch, xm 37 


558 E. LUSCHER 


fraction. It is much greater in Bence-Jones’ protein than in serum-globulin 
and approaches the amount in tissue proteins. But these, on the other hand, 
contain the diamino-acids in a greater proportion than Bence-Jones’ protein. 
Analyses of the tumor-proteins in the bone-marrow have not yet been carried 
out. Those registered in the above table were prepared from breast-cancers. 
They do not show any close resemblance ip their amino-acid balance with 
Bence-Jones’ protein. But it must be understood that Drummond coagulated 
the proteins in tissues and tumors by heat in acid solution, which also pre- 
cipitates the nucleo-proteins. It is not known in which fraction the nitrogen 
of the purine and pyrimidine-ring appears and in what manner the relations 
are changed. This may make the comparison between his analyses and those 
in which no nucleo-proteins are present somewhat uncertain. On the whole, 
Bence-Jones’ protein seems to be a substance, not only characterised by its 
- physical behaviour, but also by its distribution of amino-acids, which differs 
from all the proteins analysed up to the present time. 


EXPERIMENTAL DETAILS. 


A. Preparation of Bence-Jones’ protein: it was precipitated by heat not 
exceeding 60°, washed with water, alcohol and ether, and dried at 110° till of 
constant weight. [See Hopkins and Savory, 1911.] 

B. Analysis by van Slyke: 3-4 g. of protein were hydrolysed with a volume 
of 20 % HCl 35 times its weight for 48 hours at 100°. For each of the two 
analyses the hydrolysis was carried out separately. 

The following modifications mentioned by Crowther and Raistrick [1916] 
were applied: 

(a) To the solution of N/10 sulphuric acid were added some crystals of 
KI and 1-2 ce. of saturated solution of KIO,. The iodine produced was then 
titrated with sodium thiosulphate using soluble starch as indicator. 

(b) The melanin-precipitate was filtered off under slight suction. 

(c) The amino-nitrogen was determined by the micro-apparatus. Triplicates 
done on 2-5 ce. varied within 2 %. 

(d) I found it easier to estimate the total nitrogen of the bases in a different 
portion from arginine as Plimmer [1916] has pointed out. I carried out the 
arginine estimation itself in the original way described by van Slyke. Bumping 
was almost entirely avoided by using a rose-burner. 

After concentrating the hydrolysed protein the solution was made up to 
100 ec. and I proceeded in the following manner: 

1. Total-N in 10 cc,: duplicate. 

2. Ammonia-N in’ 75 ce. 

3. Melanin-N in 75 ce. 

The solution of the bases was made up to 100 ce, In it were estimated: 

1. Total-N in 10 ce. 

2. Cystine-N in 20 ce, 

3. Arginine-N in 50 ce, 


NITROGEN-DISTRIBUTION IN BENCE-JONES’ PROTEIN 559 


4, Amino-N in 2-5 cc.: triplicate. 

The filtrate of the bases was made up to 200 cc. and in this were estimated: 

1. Total-N in 50 cc.: duplicate. 

2. Amino-N: 10 cc. were diluted to 25 cc. and 2-5 cc. taken for estimation 
of amino-N: triplicate. 

Table III gives the results in percentage of total nitrogen of the two 
analyses carried out, the values corrected for solubility of the bases and 
according to blank tests. 


Table III. 
I Il Average 
Amide-N 9-36 9-50 9-43 
Melanin-N 0-92 0-82 0-90 
Total-N of bases 23-05 23-18 23-11 
Cystine-N 1-26 1-24 1-25 
Arginine-N 9-15 9-40 9-27 
Histidine-N 4-82 4:27 4:54 
Lysine-N 7:82 8-27 8-04 
Total-N of filtrate 66-88 66-81 66-84 
Amino-N of filtrate 61-98 61-40 61-69 
Non-amino-N of filtrate 4:90 5-41 5-15 
Total-N recovered 100-26 100-31 100-28 


C. Colorimetric estimation of tryptophan by von Fiirth and Lieben. For 
details the original paper must be consulted [1920]. 

The ordinary cuvettes of the Dubosque colorimeter can be used, putting 
a rubber ring between the bottom and the upper part and screwing the latter 
tightly down. Experiments show that the fluid does not come in contact with 
the metal and the colour remains unchanged. 

The colour reaction is better produced in two measuring flasks of 25 ce. 
than in two test-tubes. Thus the volume of the standard solution and that of 
the unknown solution can be made equal with greater accuracy. 

Each time to 2 cc. of a standard solution of 0-1 % pure tryptophan (pre- 
pared in the Biochemical Laboratory of Cambridge) and to 2 cc. of a solution 
of Bence-Jones’ protein of about 5 °% were added 1 drop of formaldehyde 
24 % and 20 cc. of saturated hydrochloric acid solution. Ten minutes after- 
wards sodium nitrite was added, the amount varying from 10-50 drops of 
a 0-05 solution in order to get the maximum depth of colour. 

The following estimations were carried out: 

1. 4-80 % solution of Bence-Jones’ protein in 0-6 % NaOH. A 20 minutes’ 
heating at 100° dissolved the protein. 20 drops of sodium nitrite produced the 
maximum colour: triplicate. 

2. 6:29 % solution of Bence-Jones’ protein in 15 % NaOH. Dissolved in 
20 minutes at 100°. 50 drops of sodium nitrite produced the maximum colour. 

3. 4:57 % solution of Bence-Jones’ protein in 20 % NaOH. Dissolved in 
15 minutes at 100°. 50 drops of sodium nitrite produced the maximum colour. 

Table IV gives the calculated tryptophan content of Bence-Jones’ protein 
in percentage: 

37—2 


560 E. LUSCHER 


Table IV. 
1 
c % aa | 
a b c 2 3 Average 
2-74 2-88 2-95 3-14 2-75 2-89 


The difference between the extreme values is 14 % of the highest. The 
difference in the amount of alkali used in dissolving the protein and therefore 
the difference in the acidity of the colour solution has no influence on the 
colour depth. This fact ought to be noticed since von Fiirth and Lieben [1920] 
pointed out that the colour depth in pure tryptophan solutions varies with 
the acidity. - 

In the course of the experiments I found two difficulties: 

1. The standard solution often changes the character of the colour in a 
short time and matching becomes uncertain even with a colour-screen. This 

‘is probably due to the beginning of clouding. On the other hand, the protein 
solution does not cloud but reaches its maximum colour more slowly than 
the pure tryptophan. 

2. The colour in pure tryptophan is blue, that in the protein is more or 
less red, so that matching without a colour-screen is impossible. It is not 
difficult to find a screen making the two colours identical and a solution of 
CuSO, proved to be best. But the following must be considered. Fearon [1920] 
has shown that formaldehyde and glyoxylic acid give with tryptophan two 
coloured compounds, which he calls tryptophan-red and tryptophan-blue 
respectively. The colour obtained under ordinary conditions is a mixture of 
both and in pure tryptophan the blue prevails, in protein the red. Since both 
colours are due to tryptophan it must be considered what influence a colour- 
screen may have on the estimation. If the screen takes away the red rays, 
then one part of the tryptophan molecules is eliminated and it is uncertain if 
the calculation only based on the remaining blue would give right values for 
the whole amount of tryptophan. As Fearon remarks, a reaction producing 
two different colours is not suitable for a colorimetric method. 

The blue colour of pure tryptophan turns more red by cooling, but at 10° 
the colour of pure tryptophan and that of protein are-not yet matchable. 

There is one aldehyde which gives nearly the same blue colour in pure 
tryptophan and in protein solutions, namely benzaldehyde. Salicylaldehyde 
gives also red and blue. As to colour, benzaldehyde would therefore be the 
most suitable reagent for a colorimetric method. The reaction is specific for 
tryptophan, no other known amino-acid producing it. Pure gelatin does not 
give any blue colour. 

I therefore tried with benzaldehyde and used the following solutions: 

1. Solution of pure tryptophan of 0-05 °% in distilled water. 

2. Solution of pure benzaldehyde of 1-9 % in pure hydrochloric acid of 
about 38 %,. 

3. Solution of sodium nitrite of 0-05 °%, in distilled water. 

4. Pure hydrochloric acid of about 38 %. It is most essential to use only 


NITROGEN-DISTRIBUTION IN BENCE-JONES’ PROTEIN 561 


very pure hydrochloric acid without or only with a very slight yellow colour. 
Otherwise the colour reaction does not reach its maximum. On the other 
hand, I found it unnecessary to saturate with hydrochloric acid gas if the 
acid is 38 °% or more. 

After a good many trials the following technique proved to be best: I 
mixed 2 ce. of the tryptophan solution, | cc. of the benzaldehyde solution and 
10 ce. of hydrochloric acid in a measuring flask of 25 cc. After 5 minutes 
I added 5 drops of the sodium nitrite solution and let it stand together for 
15-60 minutes. Then I filled up to the mark with hydrochloric acid. 

The experiments revealed several advantages in using benzaldehyde in- 
stead of formaldehyde. 

1. The colour solution does not cloud and even after several days no 
precipitate is produced. There is a slight diminution of the colour depth in 
24 hours, but it does not change appreciably in the course of | or 2 hours. 

2. The maximum colour reached with benzaldehyde is about double as 
deep as with formaldehyde, using in both cases the same concentration of 
tryptophan. This is, of course, welcome in the case of more or less insoluble 
proteins. The solution of 0-05 % tryptophan gives a satisfactory colour depth. 
The reaction goes on as quickly or even more quickly with benzaldehyde. 

3. Both pure tryptophan and protein give blue colours. There is a slight 
difference in shade and sometimes they are not matchable without a colour- 
screen. On page 560 I discussed the error possibly arising from this source by 
using formaldehyde and this holds good for benzaldehyde too. But in the latter 
case the two solutions are nearly the same, whereas in the former they are 
very different. Therefore one would assume that the error with formaldehyde 
- might be much greater than with benzaldehyde. The estimation is uncertain 
in both cases with coloured protein solutions or very impure proteins. Traces 
of iron, for instance, hinder the colour reaction. 

In order to decide if the colour depth varies proportionally to the con- 
centration, I estimated solutions of varying concentrations of tryptophan by 
comparing them with a standard solution and obtained the following results: 


Table V. 
Calculated Estimated 
Number tryptophan % tryptophan % Error % 
1 0-1092 — — 
2 0-0546 0-0510 — 6-6 
3 0-0546 0-0574 +5-0 
4 0-0546 0-0510 — 6-6 
5 0-0273 0-0256 - 6-0 
6 0-0546 0-0549 +0°5 
7 0-0546 0-0535 -1:7 
8 0-0273 0-0273 +0-0 
9 0-0546 0-0549 +0°5 
10 0-0546 0-0560 +25 
11 0-0273 0-0278 +2-0 
12 0-0546 0-0546 +0-0 
13 0-0546 0-0568 +40 
14 0-0273 0-0282 +5-0 


Average —_ = -~0-2 


562 E. LUSCHER 


There is a maximum error of 6-6 %. Von Fiirth had one of 20 %, therefore 
benzaldehyde seems to give more accurate results. I can not give any explana- 
tion of the error. The colour reaction is capricious and an unknown influence 
may sometimes come into play. 

I tried the same with caseinogen and got a maximum error of 5 %. 

In all cases I used the same amount of benzaldehyde and of sodium nitrite. 
Within the above limits it is not necessary to vary them proportionally to the 
amount of tryptophan, since the maximum colour is reached without such a 
complication of the method. In comparing protein and pure tryptophan the 
concentration of tryptophan in both solutions ought to be about the same. 

In the case of water-insoluble proteins von Fiirth dissolved them in more or 
less alkaline solutions. Since I did the same in my estimations I had to deter- 
mine the influence of alkali on the colour reaction. Instead of a neutral 

‘solution of pure tryptophan different concentrations of NaOH were used as 
indicated in the following table: 


Table VI. 
Number % NaOH Calculated % Estimated % Error % 

1 0 0-1092 0-1092 =~ 

2 10 0-1092 0-0729 50 

3 5 0-1092 0-0926 -18 
4 2-5 0-1092 0-1051 -+ 

5 l 0-1092 0-1081 -1 

6 


0-5 0-1092 0-1050 -4 


Up to a concentration of about 2-5 % NaOH there is practically no in- 
fluence on the maximum colour. Higher concentrations diminish the colour 
depth. 

It is essential to know that things are different in protein solutions. The 
following table gives the tryptophan content of Witte-Pepton dissolved in 
different concentrations of NaOH: 


Table VII. 
Number % NaOH Tryptophan % Error % 
1 0-6 3-0 ; _ 
2 15 2-85 -5 
3 20 3-0 +0 
4 30 2:9 -3 


After filtering off the precipitate of NaCl the colour depth is the same in 
all cases. Therefore insoluble proteins may be dissolved in strong alkali of 
20-30 %, without changing the colour reaction. It is not sure that this holds 
good for all proteins and it ought to be proved in every special case by using 
different concentrations of alkali. 

The above data show that the colour reaction is not so easily disturbed in 
proteins as in pure tryptophan. Therefore in making comparisons one is apt 
to get too high values of tryptophan in proteins, 

The following table gives the tryptophan content of several proteins 
estimated by formaldehyde and benzaldehyde: 


NITROGEN-DISTRIBUTION IN BENCE-JONES’ PROTEIN 563 


Table VIII. 


Tryptophan-content % 
estimated by 


if ——, Difference 
Protein Benzaldehyde Formaldehyde or 
Caseinogen 1-1 1-39 27 
Serum-protein 1-65 2-2 33 
Ovalbumin 0-97 1-5 53 
Witte-Pepton 3-0 4:8 60 


In all cases formaldehyde gives values 30-60 °% higher than benzaldehyde. 
Most probably the estimation by benzaldehyde is the more correct, since it 
avoids several disadvantages of the formaldehyde method. No other method 
for exact tryptophan estimation is known, so that a direct test is impossible. 
Therefore the method described in this paper seems to be the best at present 
available for tryptophan estimation in proteins. 


SuMMARY. 


1. The nitrogen distribution of Bence-Jones’ protein is determined by van 
Slyke’s method. 

2. Bence-Jones’ protein is a substance, not only characterised by its 
physical behaviour, but also by the distribution of nitrogen, which differs 
from that of all the proteins analysed up to the present time. 

3. ‘There is some evidence that the same protein appears in the urine in 
all cases of Bence-Jones’ protein uria. 

4. The tryptophan content of Bence-Jones’ protein is estimated by von 
Fiirth’s colorimetric method. Objections to this method are discussed. 

5. The writer proposes to use benzaldehyde instead of formaldehyde. 
There are several advantages in doing so and the method becomes more correct. 
The values of tryptophan estimated by formaldehyde are probably 30-60 % 
too high. 


My very best thanks are due to Prof. F. G. Hopkins for his kind hospitality 
in his laboratory and his kind help and advice during the progress of this work. 


REFERENCES. 


Abderhalden and Rostoski (1905). Zeitsch. physiol. Chem. 46, 125. 
Crowther and Raistrick (1916). Biochem. J. 10, 434. 

Drummond (1916). Biochem. J. 10, 473. 

Fearon (1920). Biochem. J. 14, 548. 

von Fiirth and Lieben (1920). Biochem. Zeitsch. 109, 124, 153, 
Hartley (1914). Biochem. J. 8, 541. 

Hausmann (1899). Zeitsch. physiol. Chem. 29, 136. 

Hopkins and Savory (1911). J. Physiol. 42, 189. 

Magnus-Levy (1900). Zeitsch. physiol. Chem. 30, 200. 

Plimmer (1916). Biochem. J. 10, 115. 


LIV. THE AUTOLYSIS OF BEEF AND MUTTON. 
By WILLIAM ROBERT FEARON any DOROTHY LILIAN FOSTER. 


Report to the Food Investigation Board from the Biochemical Laboratory, 
Cambridge. 


(Received June 13th, 1922.) 


I. THE AUTOLYSIS OF BEEF AND MUTTON COMPARED. 


Tue study of the course of autolysis in beef and mutton described in the 
following experiments was undertaken in the hope that it would reveal some 
explanation of the inherent differences observed in the effect of cold storage 
on these two kinds of muscle. 

It is a fact of common knowledge that beef cannot be frozen and thawed 
again without marked changes taking place in the appearance, palatability, 
and general physical state of the meat, while nothing of the sort happens when 
mutton is similarly treated. Frozen beef on thawing becomes bluish in colour, - 
flabby to the touch, and, most serious of all, there is an exudation of juice 
which contains valuable nutrient material. “Frozen” beef must be carefully 
distinguished from ‘“‘chilled” beef. The latter is beef which has been cooled 
to a temperature low enough to prevent bacterial action, but not so low as to 
cause the tissue fluids to become solid. Meat so treated is not immune from 
attack from moulds, and can only be preserved for a limited length of time. 

The fundamental differences in the two kinds of flesh underlying the 
differences in their behaviour on freezing and subsequent thawing may be 
chemical or physical. Autolysis is a term used to describe a series of post- 
mortem chemical changes in tissues, and it was hoped that any radical 
difference between the chemical constitution of beef and that of mutton would 
reveal itself in the course of the autolysis of the two kinds of meat. 

Autodigestion was the term used by early workers on this subject, and it 
implies that, under certain conditions, tissues digest themselves. The digestion 
of an organ by its own self-contained enzymes must clearly be distinguished 
from that produced by external agencies, as, for instance, the bacteria of 
putrefaction. If an excised organ is kept aseptically in vitro, it undergoes 
characteristic autolytic changes. Since certain antiseptics inhibit bacterial 
action, while they are without effect on enzymes, autolysis pursues a normal, 
or nearly normal, course in vitro in the presence of antiseptics such as toluene, 
chloroform, or acriflavine, In some pathological conditions autolysis can occur 
in the living animal. 

As in the case of digestion by the ordinary enzymes of the alimentary canal, 
autodigestion is a series of processes by which the complex compounds of the 
tissues are broken down into simpler ones, It is certain that there are changes 
in the carbohydrates and fats, but the course of the protein autolysis is that 
which has received most study, Proteolysis implies a degradation of coagulable 


THE AUTOLYSIS OF BEEF AND MUTTON 565 


protein material into simpler substances, which are not coagulated by heat, 
nor precipitated by certain protein reagents, and hence autolysis is accompanied 
by a relative increase in the non-coagulable, or soluble, nitrogen. As this is 
capable of easy estimation, the course of autolysis is usually followed by 
making successive determinations of the soluble nitrogen present. 

All animal tissues are subject to post-mortem autolysis, but the process 
proceeds at very different rates in different organs, being most rapid in glandular 
tissue, such as the liver, and slowest in striated muscle. In order to develop 
the details of technique, a preliminary examination and comparison was made 
of the course of autolysis in sheep and ox liver, and essentially the same 
methods were later applied to the study of autolysis in muscular tissue. 


Experimental details. 


About one and a half pounds of the liver were transferred immediately 
after the death of the animal, and with as little handling as possible, to a 
sterile, covered dish. On arrival at the laboratory it was cut into small slices 
and scraped to a fine pulp with sharp scalpels on a glass plate. In this way 
the glandular material could be almost entirely freed from connective tissue. 
The pulp was well mixed, weighed in a glass dish, and made up into a 20 % 
suspension in toluene water. For example, in one case (Exp. 12), to exactly 
400 g. of well-mixed liver pulp, 50 cc. of toluene were added and sufficient 
distilled water to make a total volume of 2 litres. The suspension was poured 
into a sterile flask with a well-fitting cork which had previously been soaked 
in toluene. The mixture was incubated, generally at 37°, the flasks being shaken 
thoroughly twice a day. The course of the autolysis was followed by successive 
estimations of the soluble nitrogen in an aliquot part of the whole suspension. 
Fractions were obtained by withdrawing a known volume by means of a 
blunt-ended pipette. The tip was cut off an ordinary, rather wide, 20 cc. 
pipette, which was then found to deliver 19-4 cc., and this was used throughout 
all the experiments. Provided that the suspension had been carefully made up, 
no difficulty was found in obtaining representative samples by this means. 
Samples were withdrawn immediately after making up the suspension for 
determination of the initial total and soluble nitrogen, and then at regular 
intervals, generally once a day for the first four days, and later every two or 
three days. 

The total and soluble nitrogen were estimated by Kjeldahl’s method, and 
the amino-nitrogen by Sérensen’s formaldehyde method. The value for the 
total nitrogen was obtained by incinerating 20 cc. of the suspension, corre- 
sponding to a known weight of muscle, and that of the soluble nitrogen by 
incinerating an aliquot part of the filtrate after precipitating the coagulable 
protein. In all cases the ammonia was steam-distilled into the standard acid, 
which was N for the total, and N/10 for the soluble nitrogen. The caustic soda 
was used in equivalent strengths, and was CO, free. The indicator used was 
methyl red. 


566 W. R. FEARON AND D. L. FOSTER 


The coagulable protein was precipitated either by 25 94 metaphosphoric 
acid, or 2-5 % trichloroacetic acid, according as the filtrate was to be used 
for the determination of amino-nitrogen or not. It was found impossible to 
do a formaldehyde titration in the presence of so large an excess of phosphate, 
while trichloroacetic acid showed no such “buffer” action. 

The following is a detailed account of the procedure. 20 cc. of the suspension 
were run into a stoppered graduated cylinder, and 10 cc. of a 25 °% solution of 
glacial phosphoric acid (freshly made) added, and the volume made up to 
100 ec. After thorough shaking, the cylinder was allowed to stand over night. 
The precipitate was then removed by filtration through a dry, fluted paper, 
when a crystal clear filtrate was obtained. The nitrogen was estimated in 10 cc. 
of this filtrate. Duplicate estimations were always made. When trichloro- 
acetic acid was used, 50 cc. of a 2-5 % solution were added instead of the 
‘phosphoric acid, but the procedure was otherwise the same. 


Expression of results. 


In reviewing the literature on this subject, a great diversity in the method 
of expressing results has been found. The most general practice has been to 
give the number of cc. of acid neutralised in the Kjeldahl estimation of the 
nitrogen; sometimes the amount of alkali equivalent to the excess of acid. 
As each observer chooses his own standards, this method is far from satis- 
factory, and it is suggested that a uniform method of expressing results should 
be adopted. We have found that the clearest picture of the course of autolysis 
is presented by expressing the amount of non-coagulable nitrogen present as 
a percentage of the original total nitrogen. In this way the progressive de- 
gradation of the nitrogenous pep hima is most easily followed and graphically 
represented, 

THE AUTOLYSIS OF OX AND SHEEP LIVER. 


It was found that the autolysis of the livers of both these animals follows 
a normal course, and they are strictly comparable. There is nothing in the 
breakdown of the chemical complexes after death to suggest any fundamental 
difference during life. The curves representing the increase in soluble nitrogen 
follow the same course and are similar to those obtained by other workers. 
Estimations were not made sufficiently soon after death to show the latent 
period, but both show a rapid initial increase followed after above five days 
by equilibrium. The curves obtained from amino-acid titrations show the 
same characteristics. Curves from a typical experiment are appended (Figs. 1 
and 2), 

THE AUTOLYSIS OF SHEEP AND OX MUSCLE. 

In the main the methods used for the examination of the autolysis of liver 

were applicable to muscle tissue, but slight modifications were necessary in 


preparing the suspension. The meat was separated from fat and connective 
tissue, and then shredded on a glass plate. The resulting pulp was passed 


Sol. nitrogen as % of total nitrogen. 


Mgm. sol. nitrogen per g. tissue. 


THE AUTOLYSIS OF BEEF AND MUTTON 


567 


3 -o— 
° 
o—-— 0x 
X—*— Sheep 
] i ] | j | l 1 


1 2 3 4 5 6 7 8 9 
Time in days. 
Fig. 1. The autolysis of sheep and ox liver. 
4:00 ‘s P lee 3 | 
3:00F 
o—o— Ox 
X—*— Sheep 
2-00 n i 4 ; A : 


2 3 4 5 6 7 8 9 
Time in days. 


Fig. 2. The amino-nitrogen curve of liver autolysis, 


568 W. R. FEARON AND D. L. FOSTER 


twice through a fine mincer, and then rubbed to a paste in a mortar. A definite 
amount (e.g. 100 g.) was then weighed on a tared clock glass and transferred 
quantitatively to the mortar, and well mixed with water and toluene. The 
suspension was washed into a cylinder, and made up to a known volume; a 
fixed volume of water was then used to wash this into a sterile flask, so that 
the total volume of suspension was accurately known. The flask was closed 
by a cork soaked in toluene. In general, 20 % solutions in 5 % toluene were 
used, and 20 ce. fractions withdrawn as already described. The technique of 
precipitation etc. was exactly the same as for liver autolysis. 

As in the case of the liver, the study of post-mortem autolysis in these two 
species of muscle failed to reveal any intrinsic differences. It was found, 


c 


© 
© 


FP 


Sol. nitrogen as % of total nitrogen, 


x— Mutton 


o—e— Beef 


| WON mens oe TO eRe | 
2 3. 4 5 6 7 8 esos 19 ie 


Time in days. 
Fig. 3. The autolysis of beef and mutton at 37°. (Exp. 13.) 


however, that in mutton the soluble nitrogen forms a larger percentage of the 
total nitrogen, so that the curves are parallel, but not superposable. The 
course of both, however, is perfectly regular and typical, and there is nothing 
to point to any important differences in the chemical constitution of these 
two kinds of muscle sufficient to account for their different behaviour in the 
cold store. 

The progress of the autolysis is markedly slower in muscle than in glandular 
tissue, and equilibrium is not reached till about the eighth day. Moreover, 
the degradation stops short at a smaller percentage of soluble nitrogen. 

Curves are given of a typical autolysis at 37°, and also of one at a much 
lower temperature. It is to be noted that though autolysis is retarded at the 
lower temperature, the two curves remain parallel (Figs. 3, 4 and 5). 


Sol. nitrogen as % of total nitrogen. 


Sol. nitrogen as % of total nitrogen. 


THE AUTOLYSIS OF BEEF AND MUTTON 


569 


o——e——- Beef 
X—9—- Mutton 
18k 
16 
14 x apy 


om 
NO 


10 


De) 
~ 


to 
ey) 


to 
(ee) 


= 


© 


— 
~S 


ve 2 ee 
6 


2 3 4 5 7 8 S740. 


Time in days. 


Fig. 4. Attolysis of beef and mutton at 6°. (Exp. 13.) 


ae Mutton at 37° 


Mutton at O° 


4N 7X 


bp © Beef at 37° 


| } i i 


ee 
OV 


3 6 9 12 


Time in days. 


Fig. 5. The autolysis of beef and mutton at 37° and at 0°. (Exp. 14.) 


570 W. R. FEARON AND D. L. FOSTER. 


GENERAL CONCLUSIONS AND SUMMARY. 


An examination of the post-mortem autolysis of beef and mutton has 
thrown no light on the cause of their different behaviour after being frozen. 
The processes in both are exactly parallel, both at incubator temperature and 
at low temperatures. In the case ‘of mutton, equilibrium is reached at a higher 
percentage of soluble nitrogen than in beef; but the initial non-coagulable 
nitrogen is higher, so that the curves are comparable. It is probable, therefore, 
that the differences in the two kinds of muscle in this respect lie not in their 


chemical constitution, but in the structure and physical properties of the 
fibres. 


B.air frozen 


* D. air frozen 


> 
t 


A. contro] 
Choe unfrozen) 


—" 
Ww 
' 


v. 


. a oe 
C. brine frozen 


_ 
ie) 


Sol. nitrogen as % of total nitrogen. 


— 
— 


10 i } I j sll i l l 
8 12 16 20 24 28 32 
Time in days. 


Fig. 6. The autolysis of frozen and unfrozen beef. 


Il. THE AUTOLYSIS OF FROZEN BEEF. 


While working on the autolysis of beef it seemed of interest to determine 
whether previous freezing had any influence on the course of autolysis after 
thawing. It was found that not only was the course and degree of autolysis 
profoundly altered by such treatment, but that the method of freezing, that 
is the rate of freezing, is of great significance. This is made clear by the 
curves given in Fig. 6. It will be seen that previous freezing of the muscle 
results in a much greater degradation of the nitrogenous constituents during 
autolysis, and also that equilibrium is greatly delayed. The rate of freezing of 
carcasses is known commercially to be most important, and recent work has 
shown that, if frozen sufficiently rapidly, the meat on thawing approaches 


‘THE AUTOLYSIS OF BEEF AND MUTTON 571 


more nearly in appearance and palatability to fresh beef. On the large scale, 
rapid cooling is achieved by immersing the carcass in brine at about — 20° C. 
instead of freezing in air chambers. The beneficial effect of rapid freezing is 
borne out by the autolysis curve of beef frozen in brine. This curve approaches 
much more closely to that of the control from fresh beef, than that of air- 
frozen meat. 

The following are the details of one such experiment. The animal was 
killed at 2.30 p.m. on June 29th, and immediately, before skinning, about 
24 lbs. were cut off the shin, and placed in a sterile pan in ice. On arrival at 
the laboratory, the skin and fat were removed, and the muscle divided into 
four portions. 

“‘A’’ was used as control, and a 20 % suspension was made by the method 
described above, and placed in the incubator at once. 

“‘B” was suspended in a glass jar, and placed in the cold store at 18° F. at 
3.20 p.m. It was removed the next day, and allowed to thaw for six hours at 
room temperature, and then the autolysis suspension made up as “ A.” 

“©” was treated in the same way as “‘B” except that, in the cold store, it 
was sunk in saturated brine at 18° F. 

“TD” was kept at 0° for20 hours, and then placed in cold store on June 30th, 
after which it received the same treatment as ‘ B.” 

In each case the autolytic mixtures consisted of 100 g. muscle in 5 % 
toluene. Samples were removed at the times indicated on the curve, and the 
subsequent procedure was exactly as previously described. 

Curves are given for one experiment only, but similar relationships were 
found to hold in other experiments. 

It is very difficult to explain this effect on autolysis produced by freezing. 
Obviously the equilibria of the cell have been disturbed by the rupture of the 
membranes which occurs during slow freezing. As there is an absence of 
“drip” after brine freezing it would seem that by this method any extensive 
damage to the cell is avoided, and this is borne out by the autolytic changes 
in the brine-frozen beef. 


Our sincere thanks are due to Prof. Hopkins for his interest and help 
throughout this work. 


LV. NOTE ON THE OXIDATION OF CARBO- 
HYDRATES WITH NITRIC ACID. 


By PAUL HAAS anp BARBARA RUSSELL-WELLS, 
(Received June 16th, 1922.) 


Durine the course of an examination of the carbohydrate constituent of 
Chondrus crispus we had occasion to oxidise the material with nitric acid, 
with a view to determining the amount of mucic acid produced. On examining 
the mother liquor from the mucic acid it was found to have a powerful reducing 
action upon Fehling’s solution in the cold. On repeating the oxidation under 
rather more drastic conditions and evaporating the mixture to a syrup the 
same result was obtained. This seemed sufficiently remarkable to be worth 
investigating and samples of cane sugar, lactose, glucose and levulose were 
accordingly oxidised in the same way, and in every case it was found that the 
resulting product readily reduced Fehling’s solution in the cold. Experiments 
were thereupon started upon pure glucose with the intention of isolating the 
material responsible for this strong reducing action. While these experiments 
were in progress a paper appeared by Kiliani [1922] in which he described the 
production of certain new reducing acids formed by carefully oxidising various 
carbohydrates at room temperature in the absence of air. As this author 
expressed the wish to reserve this field for himself we felt bound to discontinue 
our investigation, but thought it desirable to put on record our observations 
so far as they go in order to draw attention to the fact that powerful reducing 
substances are produced, not only under the rather peculiar conditions 
described by Kiliani, but even by oxidation of carbohydrates with hot nitric 
acid and without special precautions, although the yields obtained may 
possibly be smaller than those obtained by Kiliani. 25g. of glucose were 
dissolved in 100 cc. of nitric acid (sp. gr. 1-15) and warmed over a boiling 
water-bath; after a short time a violent action set in which gradually subsided 
and the heating was continued until the solution had been reduced to one-third 
of its original volume. In order to avoid undue dilution the solution was 
neutralised by the addition of powdered crystallised sodium carbonate. The 
resulting liquid, which was of a light yellow colour, was found to darken slowly 
even when kept in a stoppered bottle protected from the light; it reduced 
Fehling’s solution at once in the cold and likewise ammoniacal silver oxide. 
The solution was treated with 20 °%, lead nitrate until no further precipitate 
was formed. This precipitate, A, was filtered off and when washed, suspended 
in water and decomposed with hydrogen sulphide yielded a solution which 
reduced ammoniacal silver nitrate but not Fehling’s solution. The filtrate from 
precipitate A, though containing an excess of lead nitrate, gave no further 
precipitate on standing; the fact that the solution still reduced Fehling’s 


OXIDATION OF CARBOHYDRATES WITH NITRIC ACID 573 


solution at once in the cold was taken to indicate that the reducing substance 
did not form an insoluble lead salt and was possibly not an acid, but on in- 
vestigation it was found that the solution had acquired a markedly acid 
reaction, and on addition of a little alkali until the reaction was just on the 
acid side of neutrality a further heavy precipitate B was formed, which, after 
washing and decomposing with hydrogen sulphide, yielded an acid which 
reduced ammoniacal silver nitrate and Fehling’s solution on heating but not 
in the cold. The filtrate from precipitate B which still reduced Fehling’s 
solution in the cold was next treated with basic lead acetate, when it yielded 
a third precipitate C; the latter, when decomposed, gave the unknown acid 
X for whose isolation these experiments were undertaken; it was found to 
reduce both ammoniacal silver nitrate and Fehling’s solution in the cold. The 
filtrate from precipitate C no longer had any reducing properties. 

Subsequently it was found that the same acid X could be more rapidly 
obtained by replacing lead nitrate by lead acetate, since in using this reagent 
there was no change in the reaction of the solution; the addition of an excess 
of lead acetate caused the simultaneous precipitation of both the precipitates 
A and B of the previous experiment, leaving a filtrate containing only the 
acid X which could then be precipitated by the addition of basic lead acetate; 
after adding an excess of this reagent the precipitate was filtered and washed 
repeatedly with hot water until the washings were free from nitrate, an opera- 
tion which resulted in considerable loss owing to the tendency of the precipitate 
to go through the filter paper. The washed precipitate decomposed with 
hydrogen sulphide yielded a powerfully reducing acid solution which,evaporated 
in vacuo, left a pale brown, viscous material which would not crystallise and 
which gradually darkened on keeping, especially if heated. This substance, 
dissolved in water and heated with concentrated hydrochloric acid, evolved 
a considerable quantity of furfural, thus showing it to contain at least five 
carbon atoms. The aqueous acid solution, treated with baryta solution, pre- 
cipitated a small quantity of a sparingly soluble salt, the filtrate from which, 
however, still contained barium and reduced Fehling’s solution in the cold. 
The neutral solution now also gave an immediate precipitate with lead acetate 
although this reagent would not precipitate it from the original oxidation 
mixture; this precipitate dissolved in caustic soda still reduced Fehling’s 
solution in the cold. 

Attempts to prepare a crystalline oxime, phenylhydrazone, p-bromo- 
phenylhydrazone and a cinchonine salt were unsuccessful. From the above 
facts it would appear that this substance, though in some respects similar to 
glycuronic acid, is not identical with it, but in view of the circumstances set 
forth above (with regard to the claim by Kiliani to reserve the field) and the 
possible identity of this substance with one of the compounds obtained by 
this author we have decided not to pursue the investigation any further. 

REFERENCE. 
Kiliani (1922). Ber. 55, B. 75. 
Bioch. xvi 38 


LVI. OPSONINS AND DIETS DEFICIENT 
IN VITAMINS. 


By GEORGE MARSHALL FINDLAY anp RONALD MACKENZIE. ~ 
From the Royal College of Physicians’ Laboratory, Edinburgh. 


(Received June 19th, 1922.) 


ALTHOUGH it is well known that animals fed on diets deficient in vitamins 
frequently die from intercurrent bacterial infections, the reason for this 
apparent failure in the protective mechanism of the body is at present un- 
known. Hektoen [1914] found that the serum reactions were quite normal in 
animals fed for considerable periods on artificial diets, while later, Zilva [1919] 
could not determine any decrease in the power to form agglutinins, comple- 
ment and immune body in animals fed on diets deficient in vitamins. 

In the present investigation the opsonic activities of the serum were studied 
in animals fed on diets deficient in vitamins. 


DIETS DEFICIENT IN VITAMIN A. 


Rats were employed for this experiment. Four partially grown rats varying 
from 106 to 120 g. were placed on a diet consisting of caseinogen, starch, cotton 
seed oil, marmite and an inorganic salt mixture. After ten weeks the animals 
were killed. One of the animals was suffering from keratomalacia at the time 
of its death. The sera from the four animals were collected in the usual manner 
and pooled before use. i 

Four normal animals fed on a diet of caseinogen, starch, cod-liver oil, 
marmite and inorganic salts were also killed and their sera collected to serve 
as a control. Leucocytes were obtained from a normal rat. The technique em- 
ployed was that recommended by Wright [1921]. 

The opsonic activity of the two groups of sera was determined for two 
types of organisms, Bacillus coli and Staphylococcus aureus. One hundred 
polymorphonuclear leucocytes were counted on each slide. Notes were taken 
of (1) the number of leucocytes containing bacteria, (2) the total number of 
bacteria phagocytosed, The average number of bacteria per polymorphonuclear 
leucocyte was termed the phagocytic index. The data are recorded in Table I. 


OPSONINS AND DIETS DEFICIENT IN VITAMINS 575 


Table I. 
No. of poly- Total 
morphonuclear No. of no. of 
Date of Condition leucocytes cells with organisms Phagocytic 
examination Organism of animal counted bacteria ingested index 
Staphylo- } 100 79 128 1:3 
coccus ~ Vitamin A 
After10 | B.coli } 100 40 46 0-5 
weeks on 
the diet Staphylo- 100 84 136 1-4 
coccus + Vitamin A 
B. coli 100 ~ 42 69 0-7 


It will be seen that there is no significant difference in the readings obtained 
with the two sera and thus little evidence of any decrease in the opsonic 
activity as the result of a diet deficient in vitamin A. 


DIETS DEFICIENT IN VITAMIN B. 


Rats were also used for this experiment. Six rats were fed on a diet of 
caseinogen, starch, cod-liver oil and an inorganic salt mixture. Three rats were 
killed after 21 days, the remainder after 6 weeks on the diet. Rats fed on a 
complete diet were used as controls. The technique employed was the same 
as in the first experiment. The data are given in Table II. 


Table II. 
No. of poly- Total 
morphonuclear No. of no. of 
Date of Condition leucocytes cells with organisms Phagocytic 
examination Organism of animal counted bacteria ingested index 
Staphylo- 100 73 96 1-0 
coccus ~ Vitamin B 
After 3 B. coli 100 74 98 1-0 
weeks on 
the diet Staphylo- 100 70 129 1-3 
coccus | + Vitamin B 
B. coli 100 82 126 1:3 
Staphylo- 100 77 135 1:3 
coccus — Vitamin B 
After 6 B. coli 100 59 79 0-8 
weeks on { 
the diet Staphylo- 100 67 141 1-4 
coccus + Vitamin B 
B. coli 100 76 88 0-9 


DIETS DEFICIENT IN VITAMIN C. 


Guinea-pigs were used for this experiment. As is well known, when fed 
on a diet containing only small quantities of vitamin C these animals develop 
chronic scurvy, a condition characterised by no very definite clinical symptoms 
except a failure of normal growth. Six guinea-pigs, varying in weight from 
250 to 300 g., were fed on an ad libitum diet of oats and bran, and 70 cc. of auto- 
claved milk a day, with the addition of 2 cc. of orange juice every third day. 
Six controls were placed on the same basal diet but with the addition of 10 ce. 
of orange juice every day. Experiments, as before, were carried out with two 
organisms, Staphylococcus aureus and Bacillus coli. Readings were made after 

; 38—2 


576 _.G. M. FINDLAY AND R. MACKENZIE 


four weeks and after eight weeks on the dietary, three scorbutic and three 
control animals being killed on each occasion. The technique was the same as 
in the experiments with rats except that the leucocytes in this instance were 
those of a normal guinea-pig. The results are summarised in Table III. 


Table IIT. 
No. of poly- Total 
morphonuclear No. of no. of 
Date of Condition leucocytes cells with organisms Phagocytic 
examination Organism of animal counted bacteria ingested index 
Staphylo- 100 94 315 3-1 
coccus —VitaminC _ 
After 4 B. coli 100 87 115 1-1 
weeks on 
the diet Staphylo- 100 | 91 330 3:3 
coccus + Vitamin C 
B. coli 100 84 148 1-5 
, Staphylo- 100 87 261 SF 88 
; coccus — Vitamin C 
After 8 B. coli 100 82 104 1-0 
weeks on 4 
the diet Staphylo- 100 92 282 2:8 
coccus + Vitamin C 
\ B. coli 100 76 95 0-9 


THE PHAGOCYTIC ACTIVITY OF THE POLYMORPHONUCLEAR LEUCOCYTES IN 
SCURVY. 


Since the opsonic activity of the serum was practically the same both in 
scorbutic and in control animals, an attempt was next made to determine 
whether the liability to infection in scurvy was in any way associated with a 
decreased phagocytic power of the polymorphonuclear leucocytes themselves. 
For this purpose two guinea-pigs of equal weight were selected, one being 
quite healthy, while the other had been fed on a diet deficient in vitamin C 
for the previous eight weeks. 1 cc. of a suspension of S. awreus heated for an 
hour at 55° and containing 200 million organisms was injected intraperitoneally 
into each animal. At intervals some of the peritoneal exudate was removed 
by means of a capillary tube; smears were made, stained with “Leishman” and 
examined microscopically. Two hundred polymorphonuclear leucocytes were 
counted on each film, a note being taken of the number of polymorphonuclear 
leucocytes containing 0, 1, 2, 3, and more than 3 staphylococci. The results 
are shown in Table IV. 


Table IV. 
Percentage no. of polymorphonuclear leucocytes containing 
hi No 1 2 3 More than 3 
Condition Staphylo- Staphylo- Staphylo- Staphylo- Staphylo- 
Time of animal cocci coccus cocci cocei cocci 
3 hours after {Scorbutic 50 27 13 8 2 
injection | Control 44 26 16 10 1 
6 hours after (Scorbutic 16 32 22 16 14 
injection | Control 33 18 25 12 12 
18 hours after (Scorbutic 36 24 26 6 8 
injection Control 36 18 24 16 6 
30 hours after (Scorbutic 48 16 21 9 6 
injection | Control 37 20 21 13 9 


OPSONINS AND DIETS DEFICIENT IN VITAMINS 577 


An examination of the pleural exudate 54 hours after the injection showed 
that mononuclear leucocytes were now the predominant type of cell in both 
animals while the polymorphonuclear leucocytes still present were undergoing 
degeneration. 

There is no evidence to suggest any decrease in the phagocytic activity of 
the polymorphonuclear leucocytes as the result of a diet deficient in vitamin C. 

In many ways the bacterial infections so frequently found in association 
with the deficiency diseases are analogous to the terminal infections met with 
in other chronic diseases. It is therefore of some interest to note that quite 
recently Cross [1921] has investigated the question of the opsonic index in 
relation to terminal infections associated with chronic bacterial infections. 
He was unable to determine any decrease in activity against any bacteria not 
concerned in the primary infection even in the very last stages of the disease. 

As Bordet [1909] was the first to point out, the phagocytic power of the 
body would appear to be a relatively stable function and one not easily 
influenced by conditions which profoundly affect other vital activities. 


CONCLUSIONS. 


1. Rats fed on diets deficient in vitamins A and B do not show any 
decrease in the opsonic activity of the serum. 

2. Guinea-pigs fed on a diet deficient in vitamin C do not show any 
decrease in the opsonic activity of the serum. 

3. The phagocytic activity of the polymorphonuclear leucocytes of guinea- 
pigs with chronic scurvy is not decreased. 


REFERENCES. 


_ Bordet (1909). Studies in Immunity. New York. 
Cross (1921). Johns Hopkins Hosp. Bull. 32, 350. 
Hektoen (1914). J. Infect. Dis. 15, 278. 
Wright (1921). The Technique of the Teat and Capillary Glass Tube. 2nd edition. London. 
Zilva (1919). Biochem. J. 18, 172. 


LVII. ON CARRAGEEN (CHONDRUS CRISPUS). Il. 
THE CONSTITUTION OF THE CELL WALL’. 
By BARBARA RUSSELL-WELLS. 
(Received June 21st, 1922.) 


HISTORICAL. 


Many workers have from time to time investigated Chondrus crispus, but of 
recent years the most exhaustive work has been done by Sebor. The presence 
of galactose had been established by Fliickiger and Obermayer [1868], and 
Bente [1875, 1876] had found among the products of the action of acid upon 
carrageen levulinic acid and a sugar, which showed little or no optical activity. 
Haedicke, Bauer and Tollens [1867] established in carrageen mucilage the 
presence of fructose and calculated that it contained 28 % of galactose. They 
suggested that these sugars might be present under the form of raffinose. Sebor 
[1900] determined to identify the remaining sugars. He confirmed the presence 
of galactose and fructose, found small quantities of pentosan and showed that 
raffinose is not present. He also stated that he had found evidence of glucose. 
He oxidised the carrageen mucilage, but, after filtering off the mucic acid, 
failed’ to obtain any acid potassium saccharate on treating with potassium 
carbonate and acetic acid. As an analysis of the silver salt gives the same 
result for both mucate and saccharate he tested the potassium solution for 
optical activity, but found it inactive, whereas the saccharate should be active. 
He therefore abandoned the oxidation products and tested for glucose in a 
hydrolysed solution of mucilage. He hydrolysed with 6N HCl to destroy the 
fructose, and when the solution no longer gave a Selivanoff reaction, added 
benzylphenylhydrazine. He separated off the benzylphenylhydrazone of 
galactose and found the remaining compound had a melting point of 163°, 
that of glucose benzylphenylhydrazone being 165°. He declared that the 
hydrazone was a mixture and his paper gives the impression that he was not 
satisfied with the agreement of melting points referred to above. This appears 
to be all the evidence he had in favour of the presence of glucose among the 
constituents of carrageen. Miither and Tollens [1904] state that glucose is 
among the constituents of carrageen, but all the evidence they offer is an 
estimation of the amount of silver present in the silver salt of an oxidation 
product; this is obviously useless since the amount of silver is the same in 
the salts of both saccharic and mucic acids, 

Abderhalden’s Biochemisches Handlexikon {1911| contains the statement 
that the pure mucilage of carrageen corresponds to the formula O,H,)0,, but 


* Thesis approved for the Degree of Doctor of Philosophy in the University of London. 


CARRAGEEN 579 


it is not stated on what evidence this formula is based; it must be borne in 
mind that the mucilage is known to contain both nitrogen and considerable 
quantities of ash. Stocks and White [1903], on the other hand, have shown 
by analysis that the mucilage’does not possess this formula. 


Found for ash Calculated for 


free substance C,H,,0; 
Carbon 37:94 % 44-44 % 
Hydrogen 5-92 6-17 
Oxygen 54-95 49-39 
Nitrogen 1-19 —_— 


These figures conclusively demonstrate that the formula C,H,,0, does not 
agree with the results of analysis. 

Since, as will appear later, it has now been proved that carrageen mucilage 

contains at least two substances which are carbohydrate esters of sulphuric 
acid, and various metallic radicles, any calculations of the above nature are 
futile. 
_ Previous investigators have commented on the high ash content of car- 
rageen and reported that it is largely composed of calcium sulphate, but most 
of them have ignored the significance of this and proceeded to examine the 
organic radicles obtained by various means from the plant. They also appear 
to have worked either on the seaweed itself or on a direct hot water extract. 
In 1921 Haas and T. G. Hill [1921] showed that this extract consists in reality 
of two distinct fractions, and in the same year Haas [1921] published a method 
for separating the two, together with a systematic investigation of the inor- 
ganic constituents of one of them. He found that while one was readily soluble 
in both hot and cold water, the other was readily soluble in hot, but only 
sparingly in cold water. The method of extraction and separation was based 
_on this difference of solubility, and was, with slight modification, the method 
adopted in the preparation of larger quantities of the two substances required 
for the present investigation. 

The distinctions between the hot extract (H.E.) and the cold extract (C.E.) 
described by Haas and Hill were confined to physical characteristics such as 
different solubility in cold water and different gelatinising properties. 

It was also shown by Haas that the H.E. contained 17-6 °% of ash which 
could not be reduced by dialysis and was found to be due to the presence 
of the calcium salt of an ethereal sulphate, in which the calcium is fully ionised 
and can be precipitated by the ordinary reagents, while the sulphate radicle 
is not ionised and can only be precipitated by barium chloride after hydrolysis 
of the compound. 

Haas and Hill also state that they found nitrogen in both fractions. They 
also got positive reactions for protein in the seaweed itself and in the water- 
soluble product obtained from it. 

From the foregoing it will be seen that carrageen can be separated into 
three constituent fractions, the cold water and hot water soluble portions and 
the residue. The physical characteristics of thetwo extracts have been described 


580 B. RUSSELL-WELLS 


by these previous workers, and the presence of an ethereal sulphate complex 
definitely established in the H.E. 


PRESENT INVESTIGATION. 


The present investigation was undertaken with a view to instituting a 
chemical comparison between the two extracts of carrageen, both as to 
their inorganic and organic constituents, to examine the residue of the weed, 
remaining after the extraction of both fractions, for cellulose and to discover 
if other examples could be found among the seaweeds of ethereal sulphate 
grouping. 2 . 

The hand picked carrageen was first washed free from dust by holding it 
for a few seconds in running water, and then soaked for about an hour in 
distilled water. The aqueous extract thus obtained was poured off, filtered 
and evaporated, the resulting scales constituting the “cold extract.’ This 
process was then repeated and thus a further quantity of C.E. obtained. 
After only two hours’ soaking much C.E. still remained in the weed but 
it was decided to sacrifice this quantity lest a more prolonged extraction 
should result in contamination with some of the H.E. which is also 
slightly soluble in cold water. The weed was therefore washed for several 
days by soaking in water until it was no longer slimy, all the wash water 
being rejected. The weed thus deprived of its C.E and some of its H.E. was 
then washed several times in distilled water and air-dried. In order to obtain 
the H.E. the dried carrageen was heated on a water-bath with fresh distilled 
water and the liquid from this yielded, on filtration and evaporation, the 
“hot extract.” 


EXAMINATION OF THE COLD EXTRACT. 


The crude C.E., just as it is taken from the evaporating pans, has 
a distinctly saline taste and on incineration was found to contain about 
32 °%, of ash which showed a marked tendency to fuse. After dialysis the salt 
taste had disappeared and the ash content had fallen to 21-86 °%. No amount 
of dialysis would reduce the ash further. The salts which dialysed out were, 
no doubt, mainly impurities arising from contact of the seaweed with sea- 
water. During dialysis it was observed that the water outside the dialyser 
turned slightly yellow as though some organic substance had come through 
the parchment. This water was therefore collected and evaporated down 
over a water-bath on tin-lined copper trays. A smal) quantity of a brown, 
sticky, amorphous material was obtained; this did not dry in scales like the 
hot and cold extracts, and although composed very largely of the salts 
which had dialysed out from the C.E., gave a strong reaction with Molisch’s 
reagent, thus showing the presence of a diffusible carbohydrate; as it did not 
reduce Fehling’s solution to any appreciable extent it was thought possible 
that it might be trehalose, which Kylin [1915] claims to have found in carra- 


CARRAGEEN ; 581 


geen. Accordingly a dilute solution was tested in a polarimeter, but not the 
slightest trace of rotation could be observed. 

A qualitative analysis of the ash of the C.E. showed the presence of 
sulphate, calcium, magnesium and small quantities of sodium and potassium 
with traces of iron. These were all found in the ash of the H.E. as well, but 
here the amount of alkali metals was so small that the ash showed no tendency 
to fuse during the process of incineration. 

The calcium in the C.E. as in the H.E. is fully ionised and can be 
precipitated quantitatively from a 0-35 % solution by means of ammonium 
oxalate. The actual figures obtained were as follow: 

1. Calcium by incineration 3-98 %. 

2. Calcium by direct precipitation 4-03 %. 

The sulphate however is not ionised, no precipitate being formed by the 
addition of barium chloride. But, after hydrolysis with hydrochloric acid, 
barium chloride gives a copious precipitate of barium sulphate, showing that 
hydrolysis liberates the sulphate radicle in an ionised condition, while 
previously it was bound to some organic radicle. The method of hydrolysis 
employed was as follows: 

About 0-5 g. of C.E. was dissolved in 200 cc. of water and 5 cc. of con- 
centrated hydrochloric acid added. This solution was boiled over a Bunsen 
flame for at least six hours to ensure complete hydrolysis. On estimation the 
percentage of sulphate obtained after hydrolysis was found to be rather more 
than twice that obtained on incineration, the actual figures being: 


I Il Mean 
1. SO, by incineration 13-75 % 14-16 % 13-95 %" 
2. SO, by hydrolysis 30-14 30°32 30-23 


As in the case of the H.E. it is extremely difficult to get rid of all 
moisture, the substance tending to char when heated in a steam oven owing 
to the liberation of sulphuric acid. 

_ If the C.E. contained only a calcium salt of the ethereal sulphate corre- 
sponding to the formula which has been given for H.E., viz. 


0.80,.0 


K ‘ve 
oh if: 
0.S80,.0 


the sulphate ratio would be exactly two to one. Actually, however, it is more 
than two to one. This discrepancy could be accounted for on the assumption 
that some of the ethereal sulphate was combined with ammonium instead of 
with calcium. Such a combination might be present in a mixed calcium and 
ammonium salt, or there might be slight traces of the pure ammonium salt 
mixed with the calcium salt. There is not, however, enough evidence to show 
which combination is the more probable. It is obvious that in a salt of this 


582 B. RUSSELL-WELLS 
nature the sulphate moiety combined with (NH,) would leave no ash on in- 
cineration, although such sulphate could be estimated after hydrolysis. 

In order to discover if any ammonium radicle were in fact present, some 
of the C.E. was distilled with magnesium oxide, when ammonia was given 
off. This ammonia was estimated and the results, calculated as N, were as 
follow: I Wl Mean 


N by magnesium oxide 0:24 % 0:22 % 0:23 % 


The quantity of nitrogen found would, in the form of ammonium, satisfy 
0-80 % of sulphate calculated as SO,. Thus assuming the nitrogen actually 
present is contained in such an ammonium salt as suggested above, the 
experimental value for incinerated sulphate is raised by 0-80 % from 13-95 % 
to 14-75 %, which is only 0:36 % below the value required by theory, a 
difference well within the limits of experimental error. In addition to the 
nitrogen which can be driven off by distillation with magnesium oxide, C.E. 
contains other nitrogen not liberated in this manner. 

I Il 
N by Kjeldahl 0-55 % 0:57 % 
As shown above, nearly half this nitrogen can be liberated by boiling with 
magnesium oxide and is therefore probably present in an ammonium radicle. 
The rest is almost certainly there as protein nitrogen. That protein is present 
in both H.E. and C.E. is indicated by the fact that both give a positive 
reaction with Millon’s reagent and with nitric acid. Whether this protein is 
combined or not is at present impossible to say, but it is constantly present 
though it may be removed by prolonged heating with alkali. 


Ammonium salt of extracts. 


An attempt was made to prepare the ammonium salt of the C.E. in the 
hope of removing all the calcium and magnesium and so obtaining an ash 
free material. For this purpose some C.E. was dissolved in water, ammonia 
and ammonium chloride were added and then enough ammonium phosphate 
to precipitate all the calcium and magnesium. The solution was left to cool 
and settle over night and was filtered next morning. The liquid was filtered 
with the aid of the filter pump first through a Chardin and then through a 
quantitative filter-paper on a Buchner funnel. The filtrate, which was much 
clearer than a solution of ordinary C.E., was dialysed for five days to remove 
excess of ammonium phosphate and other impurities. Since it charred when 
evaporated to dryness over a water-bath, the filtrate was evaporated in a 
vacuum-desiccator and clear, almost colourless scales were obtained. These 
were dried to constant weight in the desiccator, and it was then found that 
the ash content had fallen from 21-26 % to 587%. A qualitative analysis 
of the ash showed it to consist very largely of magnesia, from which it was 
concluded that the magnesium present in the C.K. is not in an ionised con- 
dition, or else it would have been precipitated by the phosphate. 


CARRAGEEN 583 


In addition to magnesium the ash of the ammonium salt of the C.E. 
contained potassium and sodium but no calcium. 

The same procedure was adopted with the H.E. The filtrate, after pre- 
cipitation with ammonium phosphate, was dialysed for six days, and then 
evaporated over a water-bath in a platinum basin. Like the C.E. it showed 
a tendency to char, and when evaporated to small bulk turned acid and 
reduced Fehling’s solution. It was therefore kept alkaline with ammonia all 
the time it was being evaporated, and when reduced to a small volume it was 
taken from the water-bath and set aside to cool. After cooling it set to a stiff 
jelly, proving that the loss of calcium does not destroy the gelatinising pro- 
perties of the H.E. On complete evaporation in a vacuum the substance 
was found to be very like the ammonium salt of the C.E. in appearance, 
and consisted of clear, nearly colourless scales which readily absorbed moisture 
from the atmosphere. On incineration a considerable amount of ash was left, 
which like that of the C.E. contained magnesium. So here again the mag- 
nesium is present in an un-ionised condition. 

Another sample of the ammonium salt of the H.E. was prepared by 
precipitating with ammonium oxalate instead of with ammonium phosphate 
and the ash in this was estimated and found to have been reduced from the 
original 17-6 % to 4:38 %. 


COMPARISON OF ORGANIC SUBSTANCES IN HOT AND COLD EXTRACT. 


The organic complexes in the two extracts of carrageen were also investi- 
gated with a view to discovering further differences of composition between 
the two fractions. 

Both H.E. and C.E. consist largely of carbohydrates. Neither of them - 
will reduce Fehling’s solution without previous hydrolysis, but both will do 
so after.such treatment. 

Since galactose has been found in carrageen mucilage by Fliickiger and 
Obermayer [1868] and in Chondrus elatus by Takahashi [1920] it was decided 
to test for it in the two extracts. They were therefore oxidised with nitric 
acid according to the Creydt modification of the Kent and Tollens method. 
Mucic acid was obtained from both, thus proving galactose to be present in 
both fractions. The amount of mucic acid obtained was then estimated, 
according to the method of van der Haar [1920], the average values obtained 
being 21-12 °% from the H.E. and 24-82 % from the C.E. These figures corre- 
spond to 29-47 % and 33-72 % of galactose respectively in the two extracts. 

The mother liquors from the mucic acid were examined qualitatively for 
saccharic acid, but in no case was any found; on the other hand, both tartaric 
and oxalic acid were shown to be present, the latter in larger quantity in the 
case of the H.E. 


584 B. RUSSELL-WELLS 


ESTIMATION OF PENTOSES. 


Sebor [1900] expressed the view that carrageen mucilage contained a 
small quantity of pentosan or methyl-pentosan. The H.E. and C.E. were 
accordingly tested by the Kréber-Tollens method to discover whether pentoses 
could be discovered in either or both of these fractions. The actual results 
obtained were: H.E. 1-89 % of pentosan (mean of two estimations); C.E. 
1-38 % of pentosan. 

Since concentrated solutions of earrageen mucilage will set to a jelly on 
cooling it was thought that this mucilage might contain pectic bodies. Accord- 
ingly a direct hot water extract of Chondrus was tested according to the 
method of von Fellenberg [1914], which consists in steam distilling an alkaline 
solution of the substance under consideration, and testing the distillate for 
methyl alcohol and acetone [Tutin, 1921]. Two grams of the extract of 
carrageen were therefore gently boiled for 6? hours under-a reflux condenser 
with 100 cc. of V/10 NaOH, and the alkaline solution, which had darkened 
slightly, was then steam distilled. The distillate was practically neutral and 
slightly yellow in colour. It was tested for alcohol by the acid and bichromate 
test and for acetone by the iodoform and nitroprusside tests, but the result 
was negative in each case. This was therefore taken to prove the absence of 
pectic bodies from the mucilage of carrageen. 


RESIDUE LEFT AFTER EXTRACTION BY COLD AND HOT WATER. 


The residue of the Chondrus left after the removal of the C.E. and H.E. 
showed a great reluctance to part with its last traces of water-soluble material. 
It was boiled for several hours on a water-bath, the water being changed 
periodically, but each time this water on being filtered off and tested with 
Molisch’s reagent gave a strong positive reaction. A fresh portion of residue 
was therefore heated over a Bunsen burner with three successive changes of 
water, but this treatment also failed to remove the last traces of water-soluble 
material. Recourse was then had first to a steriliser and finally to an auto- 
clave. After several hours in each of these, however, water-soluble material 
was still coming out, and was found to contain calcium. 

Kylin [1915] claims to have found an insoluble calcium compound in 
Chondrus crispus. He apparently cut sections of fresh weed, boiled them in 
water and then left them to soak over night, concluding that this treatment 
removed all water-soluble substances present. He then found calcium to be 
present in the sections and assumed it to be in an insoluble form. In view of 
the extreme difficulty experienced during the present investigation in getting 
rid of the last traces of water-soluble material from the residue of carrageen 
which had already been soaked for several days in water, it seems unlikely 
that all was removed by Kylin’s treatment. Hence the calcium-containing 
substance he found was in all probability merely the last traces of the H.K. 

Since it was found so difficult to extract the residue completely merely 
by heating with water, a portion was boiled under a reflux condenser with 


~ 


CARRAGEEN 585 


3% NaOH. After two days of this treatment the contents of the flask were 
filtered, and the residue on the filter-paper washed free from alkali. It was 
then soaked in dilute hydrochloric acid to remove the carbonate found to 
have been formed during the alkali decomposition. When washed free from 
acid the residue was found to be protein- and ash-free, and completely soluble 
in cuprammonia, showing it to be pure cellulose. Cellulose was also obtained 
from the residue by acid decomposition. This was effected by boiling for 
some hours with hydrochloric acid. The solution was filtered and the solid 
matter left on the filter-paper was washed and tested with cuprammonia. 
The total amount of cellulose in carrageen, estimated by alkali decomposition, 
was found to be 1:3 %. 


CERAMIUM RUBRUM. 


After it had been definitely established that the sulphate in both the 
C.E. and H.E. of carrageen is present as an ethereal sulphate linked to an 
organic radicle it seemed probable that this grouping might be found to occur 
in other seaweeds. Accordingly it was decided to examine another alga which 
was known to yield an ash with a high sulphate content, and for this purpose 
Ceramium rubrum was selected. 

Three grams of dried Ceramium were heated with distilled water on the 
water-bath for an hour. The seaweed was then removed and the liquid 
filtered and evaporated. Scales of a transparent, horny material were ob- 
tained. Some of this material was dissolved in distilled water and the solution 
was divided into two portions A and B, barium chloride being added to each. 
To solution A a small quantity of concentrated hydrochloric acid was added 
while solution B was left neutral; both were then boiled for one and a half 
hours. At the end of this time a precipitate of barium sulphate had been 
formed in the acid solution A only, thus demonstrating the presence of the 
ethereal sulphate grouping in Ceramium rubrum. The investigation is being 
continued. 


GENERAL. 


Microscopic examination of sections of the thallus of Chondrus show that 
the cell walls become very thick and swollen in distilled water. This is what 
one would expect if they were largely composed of H.E. and C.E., since on 
being placed in water dried scales of these materials absorb much moisture 
and swell up enormously. While only 1-3 % of the dry weight of carrageen 
consists of cellulose the cell walls appear under the microscope to constitute 
the greater part of the section, and over 80 % of its weight consists of H.E. 
and C.E. Since these contain complex polysaccharides they are allied to 
cellulose and are possibly degradation products of it. 

The exact nature of the organic radicle or radicles present in the ethereal 
sulphates has not been determined. There may be one bi-valent or two 
uni-valent groups in each molecule (see formula on p. 581). 


586 B. RUSSELL-WELLS 


The present investigation has shown that the two extracts, though differing 
chemically as well as physically, are nevertheless closely allied. It may be 
that each extract contains a mixture of various closely related bodies and that 
different ones predominate in each extract. The similarity of the oxidation 
products points to both extracts consisting of substances belonging to the 
same group of compounds, but there is not enough evidence to show whether 
the differences are due to different bodies or mixtures in different proportions 
of the same substances. On the whole the former of these alternatives appears 
the more likely. | 

SUMMARY. 


1. Dialysable organic matter can be separated from the C.E. of carrageen. 

2. The C.E. contains calcium and ammonium ethereal sulphates and its 
ash contains, besides sulphate and calcium, magnesium, sodium, potassium 
and traces of iron. The ash of H.E. also contains these radicles, but it has 
less sodium and potassium and more calcium than that of C.E. 

3. The ammonium salts of both extracts have been prepared by replacing 
the ionised calcium by the ammonium radicle. 

4. Un-ionised magnesium is present in both extracts. 

5. The main oxidation products of both extracts consist of mucic, oxalic 
and tartaric acids. More mucic acid, but less oxalic acid, is obtained from 
C.E. than from H.E. 

6. Pentose radicles are present in both extracts, but more in cold than 
in hot. 

7. Pectic bodies are absent from both extracts. 

8. Cellulose is present in the residue of carrageen left after extraction 
with both cold and hot water. 

9. There are indications that the ethereal sulphate grouping is present 
in Ceramium rubrum. 

In conclusion the writer wishes to express her thanks to Dr Haas for his 
help and advice during the progress of the work, and his kind interest 


throughout. 
REFERENCES. 


Abderhalden (1911). Biochemisches Handlexikon, 2, 75. 
Bente (1875). Ber. 8, 417. 

—— (1876). Ber. 9, 1158. 
von Fellenberg (1914). Chem. Zentr. 2, 942. 
Fliickiger and Obermayer (1868). Repert. Pharm. 380. 
van der Haar (1920). Monosaccharide u. Aldehydsiiuren 
Haas (1921). Biochem. J. 15, 469. 
Haas and Hill (1921). Ann. Appl. Biol. 7, 352. 
Haedicke, Bauer and Tollens (1867). Annalen, 238, 302. 
Kylin (1915). Zeitach. physiol. Chem, 94, 337. 
Miither and Tollens (1904). Ber, 37, 298. 
Sebor (1900). Ocsterr. Chem. Zeit. 8, 441. 
Stocks and White (1903). J. Soc. Chem. Ind. 22, 4. 
Takahashi (1920). J. Coll. Agr. Hokkaido Imp. Univ. Japan, 8, 183. 
Tutin (1921). Biochem, J. 15, 494. 


“~~ 


+ 


LVIII. STUDIES OF THE COAGULATION 
OF THE BLOOD. 


- PART Il. THROMBIN AND ANTITHROMBINS. 


By JOHN WILLIAM PICKERING anp JAMES ARTHUR HEWITT. 
From the Physiological Department, the University of London (King’s College). 


(Received June 26th, 1922.) 


In a recent paper in this Journal the present writers questioned the current 
view that antithrombin is secreted by the liver. It was suggested that the 
action of that organ in promoting the fluidity of the blood after the injection 
of “peptone” depends on variation of carbon dioxide content of the blood 
[ Pickering and Hewitt, 1921]. Experiments were devised to test this suggestion 
with the somewhat unexpected result that in pithed cats respiring air, typical 
inhibition of the coagulation of the blood followed the rapid injection of 
“peptone” into animals with the liver out of the circulation. Further, the 
anticoagulant action of “peptone” was annulled by partial asphyxia and 
was restored by administration of air. It was found also, when suitable 
precautions were taken to preserve the surface conditions of the blood, that 
the addition of “peptone” to blood in vitro inhibited clotting when the 
concentration of “peptone” was no greater than was required to produce a 
like effect in vivo. 

Immunity to the anticoagulant action of “peptone” was also found to 
follow the slow injection of that substance into cats with the liver out of 
circulation [Pickering and Hewitt, 1922]. 

Having thus demonstrated that the typical action of “peptone” on the 
blood of the cat can be obtained without invoking the aid of the liver, it 
became of interest to examine the action of thrombin on the circulating 
blood of animals under similar conditions, and also to re-examine the evidence 
for the extraction of antithrombin from the liver, blood and other tissues. 
Observations on these latter points are presented in this communication. 


Tue INTRAVASCULAR INJECTION OF THROMBIN INTO ANIMALS 
DEPRIVED OF HEPATIC CIRCULATION. 


The thrombin was dissolved in mammalian Ringer’s solution and was 
extremely active as a coagulant in vitro; thus the addition of 0-25 cc. of a 
solution, prepared by dissolving 0-0048 g. of solid thrombin in 100 cc. Ringer, 


588 J. W. PICKERING AND J. A. HEWITT 


to 0-75 ce. of cat’s blood caused coagulation to commence! in 45” on glass, 
completion! of this process being one minute later. On paraffined surfaces 
these figures were 45” and 2’ 30” respectively. 


Experiment 1. Black cat, 2060 g. Pithed. Artificial respiration was maintained for 20’ before 
the aorta and inferior vena cava were ligatured. The following table shows the coagulation times 
of blood shed on to glass before and after the injection of thrombin into the circulation. 


Table I. 
Commencement of Completion of 

Time Notes coagulation coagulation 

0’ 0” Intact animal Loithses ae 7’ 40” 12’ 0” 

io Animal a vessels ligatured — sin ae — — 
36’ 30” eth ¥ if 50” ne kd 48” 
36’ 45” 7-5 cc. of 0-0048 % thrombin in Ringer injected — — 
39’ “1 gs eer prise ries \ 0” if 10” 
44’ 0” ued e pee 1%, 10” 7 50”’ 
71’ 0” a -—- 2’ 50” 8" 307 
86’ 0” Pat: Me a pn a +? 45’ 8’ 50” 


Notes: (1) Between 44’ and 86’ several observations were made. The coagulation times given 
at 71’ show the greatest variations. 
(2) Post-mortem examination revealed no intravascular clots. 
(3) Other experiments of the same nature yielded similar results. 


Experiment 2. In this observation 0-01 g. of thrombin was injected into a 2} kilo. cat under 
precisely similar conditions to those of Exp. 1. Similar results were obtained. The subsequent 
injection of 7-5 cc. of 10 % calcium chloride in distilled water caused no material alteration in 
the coagulation times of shed blood. A further injection of 5 cc. however killed the animal. 
Post-mortem examination did not show any clots in those portions of the vascular system where 
blood had been circulating. 


These experiments show: (1) That relatively large doses of thrombin can 
be injected into cats respiring air and with the liver out of circulation without 
causing intravascular coagulation. (2) That subsequent injection of a lethal 
amount of calcium chloride in distilled water fails to induce thrombosis. 
(3) That blood shed after the injection of thrombin into animals deprived of 
hepatic activity coagulates much more rapidly than the normal blood of the 
same animal under otherwise similar conditions. 

It may be remarked that the present authors have found, in the course 
of a large number of observations on pithed cats, that if the air supply of the 
animals is regulated and kept constant, then constant times of commence- 
ment and of completion of clotting of shed blood can be maintained for 
considerable periods. 

The great rapidity of clotting of shed blood after the injection of thrombin 

may in part be explained by the fact that in the experiments recorded above 
no excretion of injected thrombin by the kidneys was possible; thus, directly 
the blood was shed, the surface conditions of the plasma were altered and 
the thrombin was able to exert its coagulant action. 


* “Commencement” of clotting indicates the first visible departure from normal fluidity : 
“completion’’ that coagulation had advanced so far that the vessel could be inverted without 
spilling. 


STUDIES OF THE COAGULATION OF THE BLOOD 589 


The somewhat delayed coagulation reported by Davis [1912] following 
the injection of thrombin into intact animals under ether anaesthesia and 
preceding the typical hypercoagulability has not been observed in this 
series of experiments. Unfortunately Davis records only the times of com- 
pletion of clotting of shed blood after injection of thrombin and does not give 
the coagulation times before, recording instead the “bleeding times,” esti- 
mated by the rate of decrease of haemorrhage from a small skin wound, such 
as in the lobe of the ear. Davis employed the method of Duke [1910]. It 
should be noted that Weil [1920] has shown that prolonged bleeding time 
may coexist with normal coagulation time. Moreover Mendenham [1915] 
found that coagulation is hastened by ether anaesthesia. Unless precautions 
were taken to maintain a constant depth of anaesthesia and a constant carbon 
dioxide and oxygen content in the blood slight variations of coagulability, 
like those observed by Davis, might well occur, These would be masked later 
by the coagulant action of the thrombin. 


THE BEARING OF THE FOREGOING EXPERIMENTS ON THE 
‘“THROMBIN THEORIES” OF COAGULATION. 


The experiments reported in this paper are directly opposed to a com- 
monly accepted view, advocated by Howell [1918], that the injection of 
thrombin stimulates the liver to secrete an excess of antithrombin, which 
latter neutralises the injected thrombin and so maintains the fluidity of the 
blood. They are also dissonant with the suggestions of Howell [1912] ‘that 
prothrombin or thrombin itself constitutes a hormone which excites the 
secretion of antithrombin,” and are concordant with the fact that the typical 
action of “peptone” on blood can be obtained in cats deprived of hepatic 
activity. Support is thus given to the conclusion | Pickering and Hewitt, 1922] 
that the retarded coagulability of shed blood after the rapid injection of 
“peptone” into cats is due neither to the secretion of antithrombin nor of 
excess of alkali by the liver. This latter view, advocated originally by Dastre 
and Flouresco [1897], and later by Mellanby [1909] is also dissonant with 
the recent work of Gratia [1921] who, using Marriott’s colorimetric method 
[Marriott, 1916], was unable to find any evidence of the slightest difference 
in the alkalinity of normal and “peptone” plasmas. 

In the hypotheses of Morawitz [1904, 1] and of Fuld and Spiro [1904], 
the formation of thrombin from thrombogen (prothrombin) is said to be the 
prelude of the actual coagulation of the blood. Thrombin, it is postulated, 
acts directly on the fibrinogen of the plasma and is the immediate cause of 
clotting. 

So long as it was reasonable to assume that the liver secreted sufficient 
antithrombin to neutralise the effect of massive doses of thrombin these 
theories remained tenable. As large amounts of thrombin can be injected 
without thrombosis into the circulation when such hepatic secretion is im- 
possible, the view that thrombin acts directly on the fibrinogen of unaltered 


Bioch. xvt 39 


590 J. W. PICKERING AND J. A. HEWITT 


plasmas must be discarded and the views of Morawitz and of Fuld and Spiro 
become untenable. The theory of Howell [1912] explains the fluidity of normal 
blood by the presence of antithrombin which prevents the calcium of the 
blood from activating prothrombin to thrombin. The coagulation of shed 
blood is attributed to the neutralisation of antithrombin by thromboplastin, 
derived from platelets and tissue cells. Thrombin, so soon as formed, is again 
regarded as the immediate cause of clotting, a conclusion which is opposed 
to the evidence brought forward in this communication. 

The most recent development of the thrombin theories, that propounded 
by Bordet [1920], accepts the belief-in the existence of antithrombin and 
other anticoagulants in normal blood, but dissents from the view of Howell — 
as to the réle played by cytozyme (thromboplastin) in coagulation, citing 
the work of Gratia [1920] on the neutralisation of hirudin by cytozyme in 
support of his contention. There is also another important difference between 
the views of Howell and of Bordet. The former author accords primacy to 
antithrombin in the maintenance of fluidity, the latter, together with his 
co-workers [Bordet and Gengou, 1903; Bordet and Delange, 1913], assigns 
equal importance to surface conditions. Bordet maintains that contact of 
the blood with foreign bodies causes the interaction of lipoidal cytozyme 
(thromboplastin), formed largely from platelets but also from leucocytes and 
other sources, with serozyme, a product of the plasma formed immediately 
prior to coagulation. Serozyme and cytozyme are then said to combine, in 
the presence of calcium, to yield thrombin which in turn acts on fibrinogen 
to give clots of fibrin. Thus apart from the preliminary processes involving 
the production of thrombin, the first act of coagulation is believed to be the 
action of thrombin on fibrinogen. The evidence in this paper shows that if 
the physical conditions of the blood are preserved by remaining in the living 
vessels, then thrombin, in amounts sufficient to cause coagulation in vitro, 
does not cause thrombosis im vivo. It thus appears that although Bordet is 
correct in assigning importance to the surface conditions of the blood in the 
maintenance of fluidity, it is not to the disintegration of formed elements 
yielding thrombin, under the influence of physical change, that the inception 
of coagulation is due, but rather to physical change destroying the stability 
of the clotting complexes of the plasma. It is only after such change that 
thrombin can act as a coagulant. In this view, thrombin would be regarded 
as an accelerator of coagulation rather than as an initiator of that process. 
The work of Vines [1921] indicates that the complex associated with the 
inauguration of clotting exhibits distinctive differences from the prothrombin 
of Howell. 

This view of the action of thrombin is supported by the observation of 
the present writers [Pickering and Hewitt, 1921] that in blood surrounded 
by oil, coagulation commences as a reversible gel. Further the work of Howell 
[1916] indicates that the change from a gel to a typical clot is a physical 
process due to ageing and condensation, 


STUDIES OF THE COAGULATION OF THE BLOOD 591 


The addition of thrombin to blood in the gel stage immediately transforms 
che gel into typical fibrin, the speed of coagulation being identical with the 
speed of clotting when thrombin, in the same concentration, is added to blood 
shed on to clean glass by means of a paraffined cannula. This latter observa- 
tion is concordant with the fact discovered by Gratia [1918], that the coagula- 
tion of blood induced by thrombin proceeds at the same rate on paraflined 
as on glass surfaces. 

Biirker [1904] found platelets to remain intact in blood on a paraffined 
surface, yet under these conditions the first stage in coagulation—the forma- 
tion of a reversible gel—sets in. This falls into line with the observations of 
Achard and Aynaud [1908] that if coagulation takes place in a moist vaselined 
chamber at 16~-18°, then the platelets remain independent of the fibrin fila- 
ments. These authors maintain that physical changes of platelets are the 
result, rather than the cause of coagulation. Morawitz [1904, 2] found that 
thrombokinase (cytozyme) could be obtained from platelets. Cramer and 
Pringle [1913] showed that thrombokinase is not present in oxalate plasma 
and concluded that recalcification “induces coagulation primarily by causing 
the breaking up of platelets and only secondarily by influencing the chemical 
reaction which takes place in the presence of the substance (thrombokinase) 
liberated by the platelets,” also the “‘filtered oxalate plasma. ..is readily 
coagulated by both filtered (through a Berkefeld filter) and unfiltered platelet 
extract in the presence of soluble calcium salts.’ On the other hand Gratia 
[1918] concluded that contact with foreign substances capable of being wetted 
by blood, does not act solely on platelets and other formed elements but also 
on the colloids of the plasma. The same writer also states that calcium 
chloride, if dissolved in distilled water, disintegrates platelets, liberating, pre- 
sumably, cytozyme (thromboplastin or thrombokinase). Attention is drawn 
to experiment 2 of this paper in which massive doses of calcium chloride in 
distilled water were injected into the circulation of an animal deprived of 
hepatic activity and subsequent to a massive dose of thrombin. In this 
experiment, if Gratia is correct, platelets were disintegrated; it appears there- 
fore that thrombin: plus cytozyme from platelets is unable to cause clotting 
when the surface conditions of the blood are maintained by contact with the 
living walls of the blood vessels. It also indicates that platelets in concentra- 
tions existing in circulating blood, play only a secondary part in provoking 
coagulation. 

It has been shown that the rapid addition of a tissue extract (a source of 
cytozyme) to bird’s blood in vitro has a coagulant action, while the slow 
addition of similar amounts has an anticoagulant action [Pickering and 
Hewitt, 1921]. Demonstration has also been given that the intravascular 
injection of either calcium chloride | Léwit, 1892] or of inactive nucleoprotein 
[Halliburton and Brodie, 1894], each of which causes leucolysis and therefore 
liberates cytozyme into the circulation, is not followed by thrombosis. 

These facts indicate that the relatively slow liberation of cytozyme by the 


39—2 


592 J. W. PICKERING AND J. A. HEWITT 


disintegration of formed elements in the blood is insufficient to induce 
thrombosis. 

Bloch [1920] suggested that in circulating blood the calcium ions exist in 
an “inactive and latent state” and that on issuing from the vessels, contact 
with the air, dust and so-called thromboplastic material may transform 
“inactive” calcium to an active electrolytic precipitant of the colloids of the 
plasma. The work of Chio [1917] has indicated that the tension of carbon 
dioxide in the blood may affect the equilibrium of the calcium salts which in 
turn regulates the physical state of the blood, while Vines [1921] has shown 
that a change from combined calcium to ionised calcium takes place during 
clotting. 

In Cramer and Pringle’s experiments [1913] blood was shed through the 
air directly into oxalate!. If Bloch’s view is correct change towards clotting 
had already taken place. Caution should, it is thought, be exercised in deducing 
any theory of coagulation from the behaviour of blood after exposure to the 
air, oxalation and replacement of ionised calcium. 

In the experiments of Cramer and Pringle referred to, platelets, leucocytes 
and some erythrocytes were removed from oxalate plasma, and from these 
was prepared a “platelet extract” whose coagulant action was tested in vitro. 

The work of Spring [1900], Héber and Gordon [1904], Paine [1911] and 
Galecki [1912] shows that the speed of addition of a coagulant electrolyte is 
an important factor in coagulation of inorganic sols, while the present writers 
have demonstrated that similar conditions determine the coagulation of bird’s 
blood by tissue extract in vitro [Pickering and Hewitt, 1921]. If experiments 
are to show that lysis of platelets is the initial factor in normal coagulation, 
it would be necessary to prove: (a) that the “platelet extract,” or the aggre- 
gates of platelets, was free from products extracted from leucocytes and 
erythrocytes; (b) that the addition of the “platelet extract’ was made at the 
same rate as the liberation of the coagulant material occurring normally in 
shed blood; (c) that the concentration of either “platelet extract” or of 
aggregates of platelets added, was not greater than the concentration of 
extract resulting from the disintegration of platelets in normal shed blood; 
(d) that the lysis of platelets added to oxalate plasma in the presence of 
ionised calcium took place at the same rate as the disintegration of platelets 
in normal shed blood. 

* In a private communication to one of the present authors (J. A. H.) Dr Cramer expressed 
the opinion that our remarks on p. 719 of Vol. xv of this Journal regarding the technique of 
Cramer and Pringle were liable to misinterpretation; he suggested that we might refer to it in 
our next communication on the subject in question. We stated “that blood was not withdrawn 
through a paraffined or similarly treated cannula.”” The method actually was [Cramer and Pringle, 


1913] “to cut the artery so that the blood flowed directly into the receiving vessel.” It will be seen 
that the statement of fact is strictly correct. 


STUDIES OF THE COAGULATION OF THE BLOOD 593 


THE SIGNIFICANCE OF THE METHODS OF PREPARING ANTITHROMBINS 
AND ALLIED SUBSTANCES. 


Doyon and his colleagues Morel, Policard, Dubrulle and Sarvonat [1910- 
1919] describe the following ways of preparing antithrombin. The liver of 
the dog is frozen and thawed three times over a period of 48 hours; during 
the intervals between freezing and thawing it is exposed to the air at. room 
temperature. The liver is then perfused with 0-9 % sodium chloride and 
dilute alkali carbonate. The perfusate has no anticoagulant action until 
heated to 100° when it yields antithrombin. Alternative methods involve 
such processes as autoclaving liver at 120° or exposing minced liver to chloro- 
form vapour or allowing intestine to undergo autolysis. 

By the use of one or other of these methods Doyon has prepared anti- 
thrombins from the majority of the organs of the body, from the muscles of 
the cray-fish and from macerated earthworms. Recently however Doyon 
[1921, 1, 2] finds that antithrombin can only be obtained from the liver of 
graminivorous birds by freezing and thawing, a process which involves breaking 
up the colloidal complexes of the cells. In this connection it is noteworthy 
that Howell ascribes the fluidity of shed bird’s blood on glass to excess of 
antithrombin secreted by the liver. Doyon also found that even freezing and 
thawing, in the case of the rabbit's liver, failed to produce an anticoagulant; 
yet Davis [1912] has shown that thrombin can be injected into the circulation 
of the rabbit, in large doses, without causing thrombosis. 

The method adopted by Howell [1914] to demonstrate the presence of 
antithrombin in the blood was by heating oxalate plasma or peptone plasma 
to 60°. The filtrate delays the action of thrombin on fibrinogen and is con- 
sequently said to contain antithrombin. This antithrombin, Howell states, is 
destroyed by heating at 80°-85°. Collingwood and Macmahon [1914] found. 
that serum, particularly when kept for two days, exhibits an inhibitory influence 
on thrombin; this antithrombin is destroyed by heating to 60 or 65°. The 
retarded coagulability of “‘peptone” plasma is commonly ascribed [ Howell, 
1918] to excess of antithrombin secreted by the liver. This antithrombin is 
completely destroyed only at 100° | Nolf, 1908]. 

Even allowing a wide margin for differences in stability due to the presence 
of various protective colloids in the respective solutions and also to individual 
experimental error, it is difficult to believe that liver antithrombin stable in 
solution at 100°, or even after autoclaving at 120°, is the same substance as 
the antithrombin prepared from plasma which is destroyed at 80-85°, or that 
which appears in disintegrating serum and is destroyed at 60°, or the anti- 
coagulant of “peptone” plasma which is active up to 100°. Yet if the current 
view is accepted that antithrombin obtained from the liver is actually secreted 
into the circulation, these anomalies must be ignored. 

The most recent method for the preparation of antithrombinisthe extraction 
of rapidly dried and hashed lungs by benzene; the extract is dissolved in 


594 J. W. PICKERING AND J. A. HEWITT 


0-9 % sodium chloride and purified by precipitation at its iso-electric point. 
The anticoagulant substance produced is stated by Mills, Raap and Jackson 
[1921] to be antithrombin. 

The earlier methods of manufacturing antithrombin demand drastic dis- 
integration of the colloidal complexes from which the anticoagulant is pre- 
pared. The term “manufacturing” is used as such methods cannot imply 
unaltered extraction. Moreover, Mills, whose technique is less drastic, admits 
his method involves removal of a phospholipin from a protein-phospholipin 
complex, so that again there is only a product of cellular break-down. Like- 
wise in serum, in self-digested viscera_and in the exudation of decaying liver 
exposed to chloroform, the anticoagulant may well be considered as a product 
of autolysis. 

When more physiological methods were employed negative results were 
obtained. Menten [1920] found that perfusion of the liver with normal saline 
failed to yield any anticoagulant substance. Minot [1916] applied Doyon’s 
chloroform method to solutions of Howell's antithrombin (prepared from 
oxalate plasma by heating to 60°), and also to plasma and serum both un- 
heated and heated to 60°. In each case no anticoagulant was found. Loeb 
[1904] was unable to obtain any antithrombin-like material by extraction of 
linings of blood vessels, a result which throws doubt on the view, recently 
revived, that antithrombin is secreted by the endothelium of the vascular 
wall. Rettger [1909] found no evidence of the production of antithrombin 
after the serial injection of minimal doses of thrombin into the dog and into 
the rabbit. 

Attempts to cause the liver to secrete antithrombin by means of bile, bile 
salts, secretin and by electrical stimuli gave negative results [Denny and 
Minot, 1915]. It is noteworthy that Bulger [1918] found an increased coagula- 
tion time in cats, rabbits and guinea-pigs suffering from anaphylactic shock, 
associated with a decrease of antithrombin estimated by Howell’s method. 
Thus the altered coagulability of anaphylactic shock, which closely resembles 
that of “peptone” shock, also affords no evidence of the secretion of anti- 
thrombin. 

In short, no direct evidence has been obtained of the existence of anti- 
thrombin in living animals. 

Two other organic anticoagulants merit notice. Schikele [1912] stated that 
the extract obtained from the uterus by pressure retarded coagulation. Bell 
[1914] repeated these experiments but failed to find any anticoagulant. 
Howell and Holt [1918] extracted dried powdered liver with ether. The 
solution thus obtained was precipitated by acetone, redissolved in ether and 
reprecipitated by absolute alcohol at 50°. This last process was repeated 
from 12 to 20 times and yielded a substance (or mixture) termed by them 
heparin, which is a powerful anticoagulant in vivo and in vitro; it differs 
from antithrombin in that it does not neutralise the action of thrombin, Here 
again the product obtained appears to be the result of more or less drastic 


STUDIES OF THE COAGULATION OF THE BLOOD 595 


disintegration of cells. Howell and Holt admit that there is no direct evidence 
of the existence of heparin in the blood. Nevertheless they believe it is an 
important factor in the maintenance of the fluidity of the blood, and suggest 
that it inhibits clotting by preventing the activation of prothrombin to 
thrombin and also by activating pro-antithrombin with the production of 
antithrombin. 


THE PREPARATION OF ANTITHROMBIN FROM VEGETABLE CELLS. 


In view of the evidence indicating that antithrombins are artificial products 
formed by the dissolution of complex substances of animal origin, it became 
of interest to enquire if the employment of Doyon’s technique would produce 
an anticoagulant from vegetable sources. 

The raw materials selected were baker’s yeast and a very pure crystalline 
specimen of edestin. It may at once be stated that from both these vegetable 
products active anticoagulants were prepared. 

In each case the materials were suspended in water and heated to 120° 
under pressure for 4 hours on three separate occasions; when not being heated 
they were freely exposed to the air. The product was centrifuged and to the 
clear supernatant liquid sodium carbonate was added until the solution was 
neutral to litmus. In the case of the yeast the neutralised solution was dried 
at room temperature at low pressure and was dissolved in water when 
required. 

The action of the products obtained from yeast will be evident from the 
following experiment. 

Blood was shed from a pithed cat respiring air and with the liver out of circulation, through 
a paraffined cannula into similarly treated vessels. Thrombin was dissolved in 0-9 % sodium 
chloride. The yeast extract was dissolved in distilled water. 5 cc. of blood were employed in 


each experiment. 
Blood plus Blood plus 


Blood plus 2cc. yeast 2 cc. yeast 
Blood on _ Blood plus 0-75 ce. Blood plus _—_ extract plus extract plus 
paraffined 2 cc. distil- 0-5 % 2 ce. yeast 0-5 ce. 0-75 ce. 
surface led water thrombin extract thrombin thrombin 
Commence- ) ( Unclotted 
mentof ;- 26'10” 4’ 45” 0’ 45” + 2 hours 24’ 10” 18’ 40” 
clotting j later 
Completion (Small 
of otting 29’ 50” 8’ 45” 2’ 45” clots next 43’ 36” 41’ 10” 
| morning 


Note: Similar results were obtained when the yeast extract was dissolved in 0-9 % sodium 
chloride. 

In other experiments of this nature in which non-paraffined vessels have been used, clotting 
has been observed to occur in fractions. In this, the first clot formed contracted rapidly, while 
a portion of the fluid remained uncoagulated. Later a second small clot appeared and con- 
tracted normally and so on. In clotting induced however by addition of thrombin the clot 
invariably assumed the typical gel form and later contracted in the normal manner. 


Similar results were obtained with rat’s blood and with human blood. In 
the latter case the shed blood had been in contact with cut tissues. With 


596 J. W. PICKERING AND J. A. HEWITT 


dog’s blood shed through a paraffined cannula into glass vessels, from animals 
in which the liver was not acting, the anticoagulant action, though evident, 
was not so pronounced. Employing the same materials as in the experiment 
just given, delays of 6 and 10} minutes were obtained in the times of com- 
mencement and completion of coagulation. 

The action of the yeast extract on the coagulation of recalcified oxalate 
plasma was investigated. 


Fresh oxalate plasma was centrifuged for 30’ at 4000 r.p.m. On recalcification with calcium 
chloride the times of clotting were: commencement 8’, completion 9’ 15’. 


To the same volume of plasma was added 0-192 g. of yeast extract; after solution and 
recalcification the times were 28’ and 31’ respectively. 


Employing the usual animal technique previously described, 0-1 g. of the 
desiccated material, dissolved in 0-9 % sodium chloride, was injected into the 
heart. Blood withdrawn 10 minutes after injection yielded a minute clot in 
8 minutes, but the great bulk of the blood was fluid after 22 hours. After 
26 hours however large gelatinous clots were present. The normal coagulation 
times of the blood of the animal in question were 8’ 15” and 9’ 50” for com- 
mencement and completion. Other experiments gave similar results. 

The crude material employed in the above experiments was extracted 
with boiling ether and the ether removed at room temperature. The resultant 
product was also found to possess well marked anticoagulant properties when 
tested on recalcified oxalate plasma and on unsalted blood; it did not however 
neutralise the coagulant action of thrombin added to shed blood. It thus 
resembles in its action the substance extracted by Howell and Holt from the 
liver and termed heparin. 

Treatment of edestin as mentioned above gave a product which behaved 
like that obtained from yeast except that the coagulant action of added 
thrombin was not neutralised. 

The preceding experiments show that extract of yeast exhibits the typical 
characteristics of the antithrombin prepared by Doyon. It not merely acts 
as an anticoagulant in vivo and in vitro but antagonises the action of thrombin; 
this and the preparation of a material resembling in its action heparin, when 
considered in relation to the methods of obtaining antithrombins from the 
colloids of cells and of serum, indicate that these substances are not secreted 
by organisms possessing blood streams but are post-mortem products of the 
break-down of colloidal complexes. 

Doyon [1921, 1, 2] has recently shown that the nucleic acids prepared by 
Neumann’s method from thymus and from ganglia are anticoagulants in vivo 
and in vitro; he assigns the former action to the secretion of a nucleo-protein 
by the liver. Experiments to be reported in a subsequent paper have demon- 
strated that in an animal deprived of hepatic activity the anticoagulant 
action is equally well marked, 


STUDIES OF THE COAGULATION OF THE BLOOD 597 


SUMMARY. 


1. The massive intravascular injection of thrombin into cats deprived of 
hepatic activity does not produce thrombosis but accelerates the coagulation 
of shed blood. A phase of retarded coagulability of shed blood preceding hyper- 
coagulability was not observed when the above conditions were adhered to. 

2. The view that the continuance of the fluidity of circulating blood after 
the injection of thrombin is due to the secretion by the liver of an excess of 
antithrombin is dissonant with the results of the foregoing experiments. 

3. In the theories of Morawitz, of Fuld and Spiro and of Bordet, apart 
from the preliminary processes involving the formation of thrombin, the first 
act of coagulation is held to be the action of thrombin on fibrinogen. The 
evidence in this paper indicates that the inception of coagulation is due to 
physical change destroying the stability of the colloidal complexes of the 
plasma. Thrombin appears to be an accelerator of clotting rather than an 
initiator of that process. 

4. The addition of thrombin to blood in the state of a reversible gel 
causes immediate coagulation. 

5.. Accepting the statement of Gratia that platelets are destroyed by 
calcium chloride, it is shown that detritus of platelets, in concentrations not 
greater than can be formed by lysis in circulating blood, plays only a secondary 
part in provoking coagulation. 

6. Attention is drawn to certain inherent difficulties in interpreting experi- 
mental results obtained by the use of “platelet extracts.” 

7. An analysis of the methods of obtaining antithrombins from liver, 
sundry tissues and from serum, indicates that the anticoagulant substances 
so formed are post-mortem products. 

8. The employment of Doyon’s technique for preparing antithrombin 
from liver, yielded from yeast an extract which exhibited the properties of 
antithrombin. Further extraction with ether yielded a material similar in its 
action to heparin. The hydrolysis of edestin also gave an anticoagulant. 

9. It is concluded that antithrombins are not phylogenetic products of 
the animal kingdom arising late in evolution as a protection against throm- 
hosis, but are products resulting from the hydrolysis of protein. 


The authors would express their thanks to Prof. M. Doyon and to 
Dr P. A. Levene for placing at their disposal preparations of nucleic acids, 
and to Prof. B. J. Collingwood for an exceedingly active sample of thrombin. 


598 J. W. PICKERING AND J. A. HEWITT 


REFERENCES. 


Achard and Avian (1908). Compt. Rend. Soc. Biol. 64, 716; 65, 459. 
Bell (1914). J. Path. Bact. 18, 462. 
Bloch (1920). Lancet, ii, 301. 
Bordet (1920). Ann. Inst. Pasteur, 34, 561. 
Bordet and Gengou (1903). Ann. Inst. Pasteur, 17, 822. 
Bordet and Delange (1913). Ann. Inst. Pasteur, 26, 657; 27, 341. 
Bulger (1918). J. Infect. Diseases, 23, 522. 
Biirker (1904). Pfliiger’s Arch. 102, 36. 
Chio (1917). Arch. Farm. Sperim. 23, 206. 
Collingwood and Macmahon (1914). J. Physiol. 47, 53. 
Cramer and Pringle (1913). Quart. J. Exp. Physiol. 6, 1. 
Dastre and Flouresco (1897). Arch. de Physiol. 5th series, 9, 216. 
Davis (1912). Amer. J. Physiol. 29, 161. 
Denny and Minot (1915). Amer. J. Physiol. 38, 246. 
Doyon (1910, 1). Compt. Rend. Soc. Biol. 69, 340. 
—— (1910, 2). Compt. Rend. 151, 1074, 
—— (1912, 1). Compt. Rend. Soc. Biol. 72, 26, 485, 727 and 766. 
—— (1912, 2). J. Physiol. et Pathol. 14, 229. 
—— (1919). Compt. Rend. Soc. Biol. 82, 570. 
—— (1921, 1). Arch. Internat. de Physiol. 16, 343. 
—— (1921, 2). Compt. Rend. 172, 1212; 173, 1120. 
Doyon and Dubrulle (1912). Compt. Rend. Soc. Biol. 73, 546. 
Doyon, Morel and Policard (1911). Compt. Rend. Soc. Biol. 70, 175, 232, 433 and 797. 
Doyon and Sarvonat (1913). Compt. Rend. Soc. Biol. 74, 312. 
Duke (1910). J. Amer. Med. Assoc. 55, 1185. 
Fuld and Spiro (1904). Beitrdge, 5, 174. 
Galecki (1912). Zeitsch. anorgan. Chem. 74, 179. 
Gratia (1918). J. Physiol. Pathol. Gén. 17, 772. 
—— (1920). Compt. Rend. Soc. Biol. 83, 313. 
—— (1921). Ann. Inst. Pasteur, 35, 529. 
Halliburton and Brodie (1894). J. Physiol. 17, 172. 
Hdber and Gordon (1904). Beitrdge, 5, 432. 
Howell (1912). Amer. J. Physiol. 29, 208 and 209. 
(1914). Arch. Int. Med. 13, 76. 
—— (1916). Amer. J. Physiol. 40, 526. 
—— (1918). A Text-Book of Physiology, 466, 469. 
Howell and Holt (1918). Amer. J. Physiol. 47, 328. 
Loeb (1904). Virchow’s Arch. 176, 10. 
Léwit (1892). Studien z. Physiol. u. Pathol. d. Blutes u. d. Lymphe. Jena. 
Marriott (1916). Arch. Int. Med. 17, 840. 
Mellanby (1909). J. Physiol. 38, 502. 
Mendenham (1915). Amer. J. Physiol. 38, 51. 
Menten (1920). J. Biol. Chem. 48, 383. 
Mills, Raap and Jackson (1921). J. Lab. Clin. Med. 6, 379 
Minot (1916). Amer. J. Physiol. 39, 135. 
Morawitz (1904, 1). Beitrdge, 5, 133. 
—— (1904, 2). Arch. klin. Med, 79, 215. 
Nolf (1908). Arch. internat. Physiol. 7, 42. 
Paine (1911), Proc. Camb. Phil. Soc. 16, 430. 
Pickering and Hewitt (1921). Biochem. J. 15, 710. 
———— (1922). Proc. Roy. Soc, B, 93, 367. 
Rettger (1909). Amer. J. Physiol, 24, 431. 
Schikele (1912), Biochem, Zeitsch, 88, 169. 
Spring (1900). Rec, trav. chim. Pays-bas (2), 4, 204. 
Vines (1921). J. Physiol. 55, 204. 
Weil (1920). Rev. Md. 37, 81. ‘ 


LIX. INVESTIGATIONS ON THE NITROGENOUS 
METABOLISM OF THE HIGHER PLANTS. 


PART III. THE EFFECT OF LOW-TEMPERATURE 
DRYING ON THE DISTRIBUTION OF NITROGEN 
IN THE LEAVES OF THE RUNNER BEAN. 


By ALBERT CHARLES CHIBNALL. 


From the Biochemical Department, Imperial College 
of Science and Technology. 


(Received June 27th, 1922.) 


In any research into the metabolic processes taking place in plants, where 
fresh material is used, the period during which the research can be carried 
out must necessarily be restricted to the appropriate season during which 
the plants can be grown. To overcome the difficulty, thus entailed, some 
workers have used leaves that had been previously dried, either at low 
temperatures or from 50—70°. 

As far as is known, however—and this statement applies especially to 
those who have investigated the N and protein metabolism in plants—no 
one has first investigated the possible changes taking place during drying. 
Both protein synthesis and degradation are continually going on in the leaf 
cell; consequently, proteolytic enzymes must be present, and it is strange 
that so many workers have assumed that these would remain inactive during 
the process of drying, often at temperatures of 30-50°, the most favourable 
for enzyme action. 

The present research shows that in as far as the leaves of the runner bean 
are concerned, considerable proteolysis takes place, with consequent increase 
in the simpler water-soluble nitrogenous products, chiefly ammonia (in the 
form of ammonium salts or as the amide N of asparagine) and monoamino 
acids, the former being increased from 1 % to 5 % or 6 % of the total leaf N, 
the amount of increase depending on the conditions of drying. 

On reviewing, therefore, the work done in the past to throw light on the 
N metabolism in the leaf, it would appear that the predominance given to 
asparagine as a final product of protein metabolism is based, to a certain 
extent, on an erroneous conception as to the amount of it actually present 
in the living leaf. 

Boussingault [1860] was one of the first systematically to study the 
asparagine content of leaves, and he emphasised that it could only be isolated 


600 _ A. C. CHIBNALL 


when the plants were young. His results were confirmed by Meunier [1880] 
and Miiller [1887], who extended their researches over a wide range of plants. 
Emmerling [1887], who studied the N content of the fresh leaves of Vicia 
Faba maj. throughout the life history of the plant, never found that either 
ammonia N or asparagine N exceeded 1-25 % of the total leaf N, a result the 
author of the present paper confirmed in an unpublished series of experiments 
in 1921. The latter also [Chibnall, 1922] found less than 1 % for each of these 
substances during the life of Phaseolus vulgaris v. multifloris. Kosutany [1897] 
found a similar result for Vitis vinifera. Yet side by side with these results 
others were being published which indicated that the asparagine content of 
leaves could be quite high. On investigation it would appear that in all these 
cases the leaves had been dried before analysis. Thus Suzuki [1897] gives 
5 % of asparagine in leaves of Phaseolus vulgaris dried in an air oven at an 
unstated temperature. Wassilieff [1901] found 21-46 % of the total N as 
asparagine in young leaves of Lupinus alba dried at 70°. Jodidi, Kellogg and 
True [1920] found in spinach leaves (age unstated) dried at atmospheric 
temperature for 7 days, 4-8 % of the total N as ammonia and about 12 % 
as “amide N of asparagine.” Yet the present author found in the fresh 
leaves from plants three weeks old only 0-6 % as ammonia and 2-05 % as 
“amide N of asparagine.” 

From the results given later, combined with those stated above, it seems 
fairly certain that these high values must be attributed, at any rate in part, 
to proteolysis during the drying. The same remark must apply to the results 
obtained by Prianischnikoff [1899], who investigated the seedlings of several 
legumes, though he analysed the plants as a whole, and not the leaves alone. 

Whilst proteolysis thus affects the water-soluble products of the leaf, it is 
found that the general character of the leaf proteins, as shown by Hausmann’s 
method of analysis, undergoes but little change, if any, during the drying. 


EXPERIMENTAL DETAILS, 


The plant used was Phaseolus vulgaris v. multifloris (Scarlet Champion). 
The samples of leaves investigated were from series 8, 9, and 10, of a former 
research [Chibnall, 1922]. In these the total weight of leaves picked was—in 
series 8, 2064 g. from 12 plants; in series 9, 2424 g. from 12 plants; in 
series 10, 1860 g. from 8 plants. For the distribution of N in the fresh leaves 
only about 500 g. were required, the remainder, in each case, being air-dried 
in a closed oven through which a slow stream of air was drawn. Series 8 was 
dried at 50° in approximately 100 hours, series 9 at 40° in approximately 
120 hours, and series 10 at 40° in approximately 50 hours. The dried leaves 
were afterwards ground to a moderately fine powder and stored in a stoppered 
bottle. About six months elapsed before the present research was commenced. 

Table I gives further details of the air-dried samples. As the drying took 
place in the dark, the leaves of series 8, which had been picked one hour 
before dawn, received no opportunity for photosynthetic action before death. 


NITROGENOUS METABOLISM OF THE HIGHER PLANTS 601 


Table I. Showing details of the air-dried samples used. 
Approximate age Time ofday Fresh weight Air-dried % of moisture 


Series of plant above when picked of sample weight of in air-dried 
number — ground in weeks (G. M. T.) in g. sample in g. sample 
8 11 2 a.m. 1365 213-0 7-65 
9 ll 3 p.m. 1925 335-1 10-92 
10 14 9 a.m. 921 161-8 7:17 


General methods of manipulation (separation of protein, soluble products and 
cellular matter). The method of grinding in an end-runner mill, as used with 
the fresh leaves [Chibnall, 1922], was found to be applicable in the case of 
the dried leaves. The ground up particles of cellular matter imbibe the water 
and swell up, so that only the very fine ones pass through the pores of the 
lawn when gentle pressure is applied. By using lawn of very close texture, as 
in the case of series 8, this can be reduced to a minimum, the percentage of 
N per dry-weight of the colloidal protein being but very little lower than that 
from the fresh leaves. Even when moderately coarse lawn is used, so as to 
allow the separation to be conducted in the minimum of time, the amount of 
carbohydrate mixed with the protein is not sufficient to interfere with the 
chemical analysis of the latter (vide series 9 in Table IV). 

In some preliminary experiments the ground leaves were first shaken up 
with a small quantity of water and allowed to stand for some hours, with the 
object of facilitating the extraction of the less soluble products, but subse- 
quent analysis of the extracts, as will be shown later, indicates the presence 
of proteolytic enzymes, which are activated by the addition of water. It is 
therefore necessary to carry out the extraction in the minimum of time. 

In the preparations from 50 g. of the dried leaves of series 8 and 9, upon 
which the subsequent analyses given herein are based, 6 extractions by grind- 
ing, followed by six extractions with boiling water were given. Extraction 
with boiling water without the preliminary grinding in the cold was also tried. 
The protein- and proteose-free water-soluble N agreed with that given by the 
other method, but only a part of the proteoses passed into solution and, of 
course, no colloidal protein was obtained. On account of lack of material, 
the analysis given for series 10 in Table IV was of an extract prepared by this 
method. Table II illustrates the difference between the extracts prepared by 
the two methods. } 


Table II. Showing the water-soluble N extracted from the dried 
leaves by cold and hot extraction. 
(In percentages of total leaf N.) 
Precipitated by ammonium 


sulphate 
Number of extractions c A a8 Protein- and 
A A ~ Total Not resoluble proteose-free 
Series By cold With hot water-sol. in water Proteose water-soluble 
number grinding water N (protein?) N N 
8 6 6 30-78 0-51 3-62 26-65 
‘3 10 27-65 0-48 1-45 25:71 


0 
9 6 6 33-68 0-39 3°74 29-55 
9 0 10 30-94 0-28 0-69 29-97 


602 A. C. CHIBNALL 


Analytical methods, The details given in the former paper were followed, 
with the following exceptions: 

Proteose N. This was estimated by ammonium sulphate instead of zinc 
sulphate in acid solution [Baumann and Bémer, 1898]. No difficulty had 
been experienced with the latter method when using the extracts from the 
fresh leaves, but in those from the dried leaves the zinc sulphate separates a 
quantity of non-nitrogenous material of a slimy nature, which renders the 
subsequent filtering slow, and the washing of the separated proteoses almost 
impossible. No difficulty was experienced with the ammonium sulphate, the 
proteoses filtering readily at the pump. They were then washed with a 
saturated solution of ammonium sulphate until the washings were colourless 
and redissolved in water. The operation was repeated, and the N of the 
aqueous solution of the proteoses, after expulsion of the free ammonia by 
distillation with magnesia in vacuo, was determined by Kjeldahl’s method. 
The N of the small insoluble residue on the filter paper, probably coagulated 
protein, was also determined, after thorough washing, by Kjeldahl’s method. 

Residue after extraction of the colloidal protein and water-soluble products. 
As this still contained about half the protein N of the leaf, three extractions 
with 700 cc. of 1 % HCl were given, the extracts being united and analysed 
as before. 

RESULTS. 


Effect of drying on the proteins. Table III gives a detailed account of the 
proteins extracted from the dried leaves of series 9, with the corresponding 
data for the fresh leaves. The amount of protein passing into colloidal solution 
is proportionately much smaller than with the fresh leaves, but could no 
doubt be increased by further grindings, since the 6th extract appeared as 
heavily charged with colloidal matter as the earlier ones. The bulk of the 
protein not passing into colloidal solution, however, is extracted by the HCl. 
Assuming, as in the case of the fresh leaves, that the small amount of N in 
the final residue of cellular matter is due to unextracted protein, it will be 
seen that the total protein N has fallen from 74-83 % to 61-15 % of the total 
leaf N. 


Table III. Comparison of the proteins and water-soluble N in the 
Fresh and dried leaves of series 9. 


(In percentages of total leaf N.) 


Fresh Dried 

leaves leaves 
Colloidal protein extracted bee see ose ai 59-70 29-44 
Protein extracted by 1% HCl ... ae re wis 10:24 ~~: 25-00 
Remaining in residue __... ive Ff ie oe 413 5:90 
Coagulated by boiling... 1-76 0-42 
Coagulated by saturation with ammonium sulphate () 0-00 0:39 
Total protein N... ie tbe ish * ai 74:83 6115 
Proteose N . 1 RAs 4:35 374 
Protein: and proteose- free water-soluble N_... ps 17-22 29°55 


Extracted by alcohol and ether . i als vos 3°60 2-96 


NITROGENOUS METABOLISM OF THE HIGHER PLANTS 603 


Table IV gives the distribution of N in the colloidal proteins from the 
fresh and dried leaves of series 8 and 9. It will be observed that no apparent 
change has taken place in the protein molecule. 


Table IV. Comparison of the distribution of N in the colloidal proteins 
from the fresh and dried leaves. 


(In percentages of total protein N.) 
% of N 


Series per dry Amide Humin Basic Monoamino Other 
number weight N N N N N 

8 fresh 11-75 6-14 3°55 19-33 59-25 11-73 
8 dry 10-11 6-26 4-63 21-20 57-93 9-98 
9 fresh 11-71 6-34 3°64 22-00 58-88 9-14 
9 dry 8-64 6-73 4-71 21-67 56-30 10-51 


Table V gives the distribution of N in the proteins extracted by the HCl 
from the fresh and dried leaves of series 9. In comparing them it must be 
remembered that they represent very different proportions of the total leaf 
protein. The HCl extract from the fresh residues represents only about } of 
the total protein, that from the dried leaves nearly 4. Since the leaf cell 
undoubtedly contains more than one protein, it follows that if the dried 
leaves, like the fresh, contain a small amount of protein rich in amide and 
basic N, these characteristics, on analysis, may be masked by the predomi- 
nance of unextracted protein, of the type represented by the colloidal protein. 
If allowance be made for soluble carbohydrates due to the mild hydrolytic 
action of the HCl on the pentosans etc. of the cell wall material, which will — 
increase the humin N at the expense of the monoamino N, the distribution 
of N in the HCl extract from the dried leaves does not differ from that of 
the corresponding colloidal protein, indicating that the major part of it is of 
this type. Interpreting the various distributions of N in terms of the amounts 
of the total leaf protein that they represent, it would appear that the leaf 
proteins do not undergo any appreciable change in character during the 
process of drying. 


Table V. Comparison of the distribution of N in the proteins 
extracted by 1 % HCl. 


(In percentages of total N in the extract.) 


Series Amide Humin Basic Monoamino Other 
number N N N 

9 fresh 8-19 5:82 24-66 48-78 12-64 
9 dry 6:96 6:16 21-57 54-01 11-30 


Effect of drying on the proteoses. Table VI shows that in the cold extracts 
prepared from the dried leaves the proteoses have diminished. Since Table IT 
shows that they are but slowly soluble in hot water, it may well be that after 
drying they are not so readily soluble in cold water, in which case the apparent 
loss may be due to incomplete extraction. 


604 A. C. CHIBNALL 


Table VI. Comparison of the distribution of N in the aqueous extracts 
from the fresh and dried leaves. 
(In percentages of total leaf N.) 
Total Protein- and 


water- proteose-free Am- Amide Mono- 

Series soluble water-soluble monia WN of Nitric Humin Basic amino Other 
num ber N N N Asparagine N N N N 
8 fresh 21-07 * 0:36 0-0 3-56 1-00 * 7:06 2-34 
8 dry 30-78 26-65 6-02 0-0 3-57 1-15 1-81 10-55 3:55 
Diff. 9-71 — 5-66 — — — — 3°49 1-21 
9 fresh 21-57 17-22 0-44 0-53 2-73 0:99 2-00 7:76 2:77 
9 dry 33-68 29-55 5-32 1-51 1:62 0-82 2-84 13-22 4-22 
Diff. 12-11 12-33 4:88 0-98 1-11 — 0-84 5:46 1-45 
10 fresh 22-75 18-16 0:37 0-47 3-53 1°06 2-30 6-81 3°62 
10 dry (1) 27-07 2-68 2-26 3-01 0-89 3-51 10-60 4-12 
Diff. — 8-91 231 1-79 0-52 _— 1-21 3-79 0-50 


* Proteoses not determined. 
(1) Determined from an extract made with hot water only, which gives a low value for the 
proteoses. ‘ 


Effect of drying on the protein- and proteose-free water-soluble products. 
Table VI shows that in all three cases this has increased, the amount being 
roughly proportional to the time of drying. The products of the protein 
autolysis fall chiefly into ammonia (as ammonium salts or “amide N of 
asparagine”) and monoamino acids. This is illustrated better by Table VII. 
Leaving for a moment the relation between the ammonium salts and aspara- 
gine, which is discussed later under the heading of enzymes, it will be seen 
that in the two samples picked after some hours’ daylight, series 9 and 10, 
the amount of ammonia N is nearly equal to the monoamino N, whereas in 
series 8, picked after 6 hours’ dark, the ammonia is in considerable excess. 
In the latter case this may possibly be due to the higher temperature of 
drying (50° instead of 40°) or, more probably, to the fact that the leaves, 
before drying, were deprived of most of their reserve carbohydrate by trans- 
location away from the leaf at night. 


Table VII. Showing the ammonia and amino acids formed 
by proteolysis during drying. 
(In percentages of total autolysed protein N.) 


Ammonia N 
Series Ammonia N+ amide Monoamino Total: {| Amide N of asparagine 
N 


number N of asparagine Monoamino N 
8 58:3 B50 94:2 
9 47-5 44°3 91-8 
10 46-0 42-5 88:5 


In series 9 and 10 again, there is a loss of nitric N, probably due to the 
necessary carbohydrate for the synthesis of organic nitrogenous products 
being present, since in series 8 there is no loss. The higher temperature 


NITROGENOUS METABOLISM OF THE HIGHER PLANTS 605 


of drying may have accounted for this, however, since Couperot [1909] 
has shown that leaves dried at room temperature lost 25-50 % of their 
nitric N as hydrocyanic acid whereas those dried rapidly at 60° suffered no 
such loss. 


Table VIII. Showing the enzyme action produced by allowing the 
dried leaves to stand with cold water. 


(In percentages of total leaf N.) 
Protein Protein- 
and 


d 
Hours Total proteose N gubtenns- 
standing water- removed by free water- Amide 

Series with soluble ammonium soluble Ammonia N of Nitric Humin Basic Monoamino Other 

No. water N sulphate N N asparagine N N N N N 
8 0 30:78 4-13 26-65 6-02 0-0 3°57 1-15 1:81 10-55 3°55 
8 16 35-90 3°16 31-74 2-30 2-96 3:59 0-89 3-76 13-56 5-68 
9 0 33-68 4:13 29°55 5:32 1-51 1-62 0-82 2-84 13-22 4-22 
9 16 39°35 3°39 35-94 1-69 4-58 1-67 0-90 5-63 15-80 5-67 
9 116 39-82 2-86 36-96 1:20 4-10 1:56 1-04 5-67 17-48 5-91 
10 0 * * 27-07 2-68 2-26 3-01 0-89 3°51 10-60 4-12 
10 16 31-60 3°21 28°38 1-19 3°57 3-04 1:02 3°33 11-14 5-09 


* Determined in a hot water extract only, which gives a low value for the proteoses. 


Enzymes present in the dried leaves. As was stated earlier, if the ground - 
leaves are allowed to remain suspended in water for some hours (under a 
layer of toluene), proteolytic enzymes become active, with consequent increase 
of the simpler nitrogenous products in solution. Table VIII shows the 
analyses that have been made to illustrate this. A study of the figures shows 
that the action is complex, but the following points stand out fairly clearly: 

(1) There is at first a rapid increase in the water-soluble N tending to 
reach a limit with lapse of time. 

(2) Monoamino N shows a continuous increase. 

(3) Basic N shows an increase that is not maintained with lapse of time. 

_ (4) Ammonia N shows at first a rapid decrease, tending to reach a limit. 
At the same time there is a nearly corresponding increase in the ‘‘amide N 
of asparagine.” 

(5) The sum ammonia N + ‘amide N of asparagine” shows a slow 
decrease, indicating re-synthesis of more complex bodies. 

From (2) and (3) it is clear that an enzyme of the nature of a pepsin is 
present (the extract is slightly acid). But the chief point of interest is the 
rapid conversion of the ammonia N into the amide N of asparagine. In the 
leaf the origin and relationship between these two substances have not yet been 
definitely established. Under normal conditions the evidence is that the 
concentration of both of them is low. Whether the ammonia is due only to 
direct translocation from the root, or to protein degradation in the leaf itself 
or to both of these is not yet clear. Undoubtedly an enzyme capable of 
breaking down the protein into ammonia and amino acids exists in the dried 
leaves, but this may not be so in the normal fresh leaf. In a previous paper 
[Chibnall, 1922] there was described an experiment in which the living leaf 

Bioch. xv1 40 


606 A. C. CHIBNALL 


was starved under conditions that precluded the translocation of the products 
of metabolism, and in spite of considerable protein degradation, no appreci- 
able increase in either ammonia or “amide N of asparagine” was noted!. 
Suzuki [1897], by starving moist leaves for 48 hours, claims to have increased 
the “asparagine N” from 5 % to 13 % of the total N, but his results must be 
queried since he dried the leaves before analysis. 

Asparagine itself cannot be a primary product of protein degradation, but 
must be a product of re-synthesis from the ammonia and amino acids. In 
very young plants direct translocation of it from the cotyledon reserve is 
possible, since Prianischnikoff [1896] showed in germinating seeds of Vicia 
sativa a steady increase of asparagine at the expense of the protein. Butke- 
witsch [1909], who experimented with seedlings of lupins and peas, considers, 
as does Suzuki [1894], that the formation of asparagine is to remove an 
injurious excess of free ammonia. This view, if modified to include an excess 
of monoamino-dicarboxylic acids also, is very likely, and the reaction 
“ammonia + aspartic acid == asparagine” is probably due to an enzyme 
governed by the H ion concentration of the medium in which it is working. 
As far as is known, an enzyme of this nature, an asparaginase, has not yet 
been indicated, and it is therefore of great interest to observe that one appears 
to be present in the dried leaves used in this research. Furthermore, under 
the conditions of the present experiment, it appears to have marked synthetic 
activity. 

It is hoped at an appropriate season to repeat these enzyme experiments 
with fresh leaves, and if an asparaginase should be indicated, to study the 
effect of change in the H ion concentration upon its action. 

In conclusion the author would like to thank Prof. 8. B. Schryver for his 
advice and encouragement throughout the research, which was made possible 
by a grant from the Department of Scientific and Industrial Research. 


SUMMARY. 


1, The effect of low-temperature drying on the nitrogenous bodies in the 
leaves of the runner bean has been investigated. 

2. Protein autolysis takes place, with increase in the simpler water- 
soluble N products. 

3. These products are chiefly ammonium salts, asparagine and amino acids, 

4. The leaf proteins, whilst they are diminished in amount, are not 
appreciably changed in character. 

5. Proteolytic enzymes are present in the dried leaves, and are activated 
by the addition of water. 

1 The season 1921 was abnormal, the bean plants bearing but few pods, At the time this 
experiment was carried out, 23 August 1921, no pods were in process of formation. Further 
starvation experiments are being carried out this season, and the results so far to hand from 
leaves picked in August 1922, when pod formation was active, indicate that considerable 


quantities of asparagine are formed by degradation of the leaf protein, Ammonia N, as before, 
is unchanged, 


NITROGENOUS METABOLISM OF THE HIGHER PLANTS 607 


6. The presence of an asparaginase, activated by the addition of water, 
is indicated. Under the conditions of the present research it possesses marked 
synthetic activity. 

7. The position of ammonia and asparagine in the N metabolism of the 
leaf is discussed. 


REFERENCES. 


Baumann and Boémer (1898). Zeitsch. Untersuch. Nahr. u. Genussm. 1, 106. ° 
Boussingault (1860). Agronomie, chimie agricole et physiologie, 2nd ed, 
Butkewitsch (1909). Biochem. Zeitsch. 16, 411. 
Chibnall (1922). Biochem. J. 16, 344. 
Couperot (1909). J. Pharm. Chim. 29, 100. 
Emmerling (1887). Land. Vers. Stat. 34, 1. 
Jodidi, Kellogg and True (1920). J. Amer. Chem. Soc. 32, 1061. 
Kosutany (1897). Land. Vers. Stat. 48, 13. 
Meunier (1880). Ann. Agron. 6, 275. 
Miiller (1887). Land. Vers. Stat. 33, 326. 
Prianischnikoff (1896). Land. Vers. Stat. 46, 459. 
—— (1899). Land. Vers. Stat. 52, 137, 347. 
Suzuki (1894). Bull. Coll. Agr. Tokyo, 2, 409. 
—— (1897). Bull. Coll. Agr. Tokyo, 3, 245. 
Wassilieff (1901). Land. Vers. Stat. 55, 45. 


40—2 


LX. INVESTIGATIONS ON THE NITROGENOUS 
METABOLISM OF THE HIGHER PLANTS. 


PART IV. DISTRIBUTION OF NITROGEN IN THE DEAD 
LEAVES OF THE RUNNER BEAN. 


By ALBERT CHARLES CHIBNALL. 


From the Biochemical Department, Imperial College 
of Science and Technology. 


(Received July 12th, 1922.) 


In a former paper [Chibnall, 1922, 1], the seasonal variations in the N content 
of the leaves of the runner bean, Phaseolus vulgaris v. multifloris (Scarlet 
Champion), were demonstrated and discussed. The period examined covered 
the first twenty weeks of the plant’s life, from the seedling stage, until the 
leaves were showing signs of chlorophyll degeneration. Eleven samples were 
analysed, numbered consecutively “series 1—-1f.”’ 

The plants died during the 24th week, following two nights’ hard frost, 
and the present communication deals with the distribution of N found in the 
dead leaves, which will be denoted henceforth as series 12. As picked from 
the dead plants they were shrivelled up and dry, whilst their green colour 
was masked by a brownish-red stain, probably due to iron salts, since it was 
easily washed away with cold water. 

The results given later for series 12 have in all cases been compared with 
the corresponding ones for series 9, 10 and 11, the four together covering the 
latter half of the plant’s life, from the 11th to the 24th weeks. During this 
period the plants were fully grown and pod formation was complete, so that 
the general metabolism would be, presumably, on the down grade. The tables 
given later show how remarkably stable is the general equilibrium between 
the nitrogenous substances in the leaf during this period, the changes, even 
at death, being of quite a small order. There is no evidence of any great with- 
drawal of N from the leaves before death, nor is there any accumulation of 
asparagine. Miyachi [1896] assumed that asparagine accumulates in the aged 
leaves of Paeonia albiflora since old leaves from dying plants, if placed under 
water for 14 days, become rich in this substance. In the present author’s 
opinion this experiment merely demonstrates the presence of active enzymes 
in the old leaves. (Compare Chibnall [1922, 2], where observations on dried 
leaves suspended in water are recorded.) 

The methods of manipulation and analysis were those used for dried leaves, 
as described by Chibnall [1922, 2). 


NITROGENOUS METABOLISM OF THE HIGHER PLANTS 609 


The results can be summarised under the following headings: (1) There is 
no great withdrawal of N from the leaves to the stems or roots when the 
plant becomes aged [vide Czapek, 1920]. (2) The protein- and proteose-free 
water-soluble N of series 12 shows a slight increase, due chiefly to ammonia 
(as ammonium salts or “amide N of asparagine”) and monoamino N. From 
the author’s experiments on the changes in the leaf N due to low-temperature 
drying [Chibnall, 1922, 2], it would appear that this increase has taken place 
during the period of dehydration of the leaf on the plant. Towards the latter 
end of the plant’s life the nitric N has fallen, as one would expect from the 
ageing of the root system, otherwise the equilibrium between the water- 
soluble nitrogenous products has remained very steady, indicating a period 
of low metabolic activity in the leaf. (3) The proteins in the leaf have under- 
gone slight amidisation, a change that must be ascribed to the ageing of the 
leaf, since it was not observed in the proteins of leaves dried at low tempera- 
tures. The ratio of protein to non-protein N in the leaf has suffered no appreci- 
able change with age or death. 


Table I. Showing the percentage dry weight and total N 


in the leaves of series 9-12. 
Age above 

Series ground in % dry % of N per % of N per 
number weeks weight fresh weight dry weight 

9 ll 15-51 0-675 4-35 

10 14 16-31 0-724 4:44 

ll 20 16-13 0-622 3°86 

12 24 91-6 _ 3-60 


Table Il. Showing the amounts of protein and non-protein N 
in the dead leaves (series 12). 


(In percentages of total leaf N.) 
Colloidal protein extracted... as nae 29-22 
Protein extracted by 1 % HCl as és 29-49 
N in residue (pro tein ?) 6-62 
*Protein coagulated by ammonium sulphate .. 0-77 
Total protein N es od as Ses 66-10 
Proteose .. és ee 6-28 
Proteose and protein- free water-soluble N_ ... 22-10 
Soluble in alcohol and ether ... maa 2-15 
Total N accounted for ... ve ska a 96-63 


* In estimating proteoses by saturation with ammonium sulphate, a small amount of N, 
probably due to traces of coagulated protein, is not resoluble in water. 


Table III. Showing the water-soluble N in series 9-12. 
(In percentages of total leaf N.) 


Series Total water- Proteose Coagulated Protein- and 
number soluble N N protein N* proteose-free N 
9 21-57 4-35 _ 17-22 
10 22-75 4-57 — 18-16 
ll 22-41 4:38 _ 18-03 
12 29-15 6-28 0-77 22-10 


* As in footnote for Table IT. 


610 4 A, C. CHIBNALL 


Table [V. Showing the distribution of the water-soluble N in series 9-12. 
(In percentages of total leaf N.) 


Total 
protein- and Amide N Mono- 
Series proteose- Ammonia of Nitric Humin Basic amino Other 
number free N N asparagine N N N N N 

9 17-22 0-44 0-53 2-73 0-99 2-00 7-76 2-77 

10 18-16 0-37 0-47 3-53 1-06 2-40 6-81 3-62 

ll 18-03 0-61 0-92 1-28 0-96 2-59 6-61 5-06 
12 22-10 0-97 1-94 1-61 1-32 2-22 7-74 6:30 


Table V. Percentage distribution of N in the colloidal proteins of serves 9-12. 


Series Amide Humin Basic Monoamino Other 
number N N N N N 
9 6-34 3-64 22-00 58-88 9-14 
10 6-50 3-82 21-78 : 59-22 8-68 — 
ll 6-40 3°82 21-32 58-50 9-96 ~ 
12 7-21 6-12 19-27 56-22 11-32 
12* 8-81 


6-67 19-27 51-90 13°35 
* HCl extract of series 12. 


REFERENCES. 


Chibnall (1922, 1). Biochem. J. 16, 344. 

—— (1922, 2). Biochem. J. 16, 595. 
Czapek (1920). Biochemie der Pflanzen (Jena), 2, pp. 293-295. 
Miyachi (1896). Bull. Coll. Agr. Tokyo, 2, 458. 


LXI. NOTE ON THE NON-PROTEIN NITROGEN 
IN GOAT’S MILK. 


By WILLIAM TAYLOR (Carnegie Research Scholar), 
The Rowett Institute, Aberdeen. 


(Received June 28th, 1922.) 


As a result of an investigation into the factors controlling the percentage 
composition of milk, during which a goat was fed on various diets each of 
which was abnormally high in some one of the three energy-yielding consti- 
tuents of the food, it was found that irrespective of the nature of the diet 
the percentages of protein, fat, lactose and ash all tended to vary with the 
daily volume of milk secreted, the percentages of protein, fat and ash tending 
to vary inversely and the percentage of lactose directly with the daily volume 
[Taylor and Husband, 1922]. So far as these constituents are concerned, 
therefore, the percentage composition of milk depends upon its rate of secre- 
tion rather than upon the percentage composition of the diet. 

In the case of non-protein nitrogen no relationship could be traced bétween 
the percentage and the daily volume of the milk. The results of the following 
experiment show that the percentage of this constituent depends upon the 
nature of the diet. ; 

EXPERIMENTAL. 


A goat was fed on various diets of constant composition for periods varying 
from 8 to 16 days, and the total protein, caseinogen and albumin plus globulin 
contained in the milk were estimated daily. The figure for non-protein 
nitrogen was obtained from the difference between the percentage of total 
protein and the sum of the percentages of caseinogen, albumin and globulin, 
The figure for non-protein nitrogen is expressed in terms of milk protein, as 
the non-protein nitrogen was included in the estimation of total nitrogen 
from which the total protein was calculated. 

The urine was collected and a daily estimation of the total excretion of 
urinary nitrogen was also carried out. 

The following table shows, for each dietary period, the average daily total 
quantity of nitrogen in the urine and the average percentage of non-protein 
nitrogenous substances in the milk. The periods are arranged in the table 
according to the level of nitrogen excreted in the urine. 

The table shows a definite correlation between the daily output of nitrogen 
in the urine and the percentage of non-protein nitrogen in the milk. 


612 W. TAYLOR 


Average percentage 
of non-protein 


Average daily nitrogenous Duration of 
excretion of substances in experimental 
urinary nitrogen the milk Nature of period 
g. per cent. diet Days 
26-8 0-39 High protein 16 
22-5 0-37 = ‘s 12 
9-6 0-22 Potatoes 8 
9-4 0-22 Grass 14 
4:3 0-20 High fat 14 
4-1 0-20 Spt ee 14 
2-3 0-16 High carbohydrate 13 


Discussion.oF RESULTS. 


It has been shown that the non-protein nitrogenous substances in milk 
consist of amino-acids, creatine, creatinine, uric acid, and urea, the last 
named being present in greatest amount [Denis, Fritz Talbot and Minot, 
1919]. The concentration of these in the blood seems to determine both the 
amount excreted in the urine and the percentage present in the milk. The 
mammary gland acts to some extent as an excretory organ, waste non-protein 
nitrogenous substances filtering through from the blood to the milk. The 
percentage in which these are found in milk seems to be determined by the 
degree of concentration in the blood of the end products of protein metabolism. 


CONCLUSIONS. 


1. In a lactating animal there is a correlation between the daily output 
of nitrogen in the urine and the percentage of non-protein nitrogen in the 
milk, both apparently being determined by the amount of protein in the food. 

2. The mammary gland acts as an excretory organ passing through from 
the blood to the milk end products of protein metabolism. 


REFERENCES. 


Denis, Fritz Talbot and Minot (1919). J. Biol. Chem. 39, 47. 
Taylor and Husband (1922). J. Agr. Sci. 12. 111. 


LXII. SMELL. 


By EDWIN ROY WATSON. 
From the Dacca College, Dacca, Eastern Bengal. 


(Received July 3rd, 1922.) 


Tue aim of this investigation was to find a method of accurately and scientifi- 
cally describing smells. Sounds and colours can be accurately described by 
the frequencies of the vibrations of which they are composed, but no such 
method has yet been discovered for describing smells, and the best that can 
be done at present is to say that there is some kind of resemblance between 
the smells of certain substances though at the same time it is generally felt 
that such a description is very unsatisfactory and incorrect. Not only is it 
impossible to describe any smell satisfactorily but there is at present no 
criterion as to whether a smell is simple or complex. It is true we can some- 
times detect more than one substance in a mixture by the smell but we are 
generally helped in such an analysis by the different volatility of the different 
constituents of the mixture, and in fact such analysis is very uncertain and 
liable to error. 

The problem is difficult, because, so far as is known at present, smell is 
not due to any kind of vibration such as the air vibrations causing sound or 
the ether vibrations causing light. The fact that most and probably all 
odorous substances are volatile and that no odorous substance can be detected 
by its smell when enclosed in an air-tight receptacle made of any ordinary 
material has led to the general belief that contact between the odorous sub- 
stance and the olfactory organ is necessary to cause the sensation of smell. 
At any rate temporarily the author has assumed that smell is due to chemical 
or physical reaction between the odorous substance and the olfactory organ. 
It is well. known that certain odorous substances react chemically with 
proteins, e.g. the halogens, nitrous fumes, hydrochloric acid, formaldehyde. 
And as the olfactory organ is undoubtedly composed chiefly of proteins it 
seemed worth while first to see whether any chemical reaction could be 
detected between proteins and odorous substances in general. The method 
used by Bugarsky and Liebermann [1898] to prove chemical reaction 
between non-coagulated white of egg and hydrochloric acid seemed specially 
suitable. These investigators showed that the addition of egg-white to 
aqueous hydrochloric acid caused an elevation of the freezing-point, pointing 
to a decrease of the number of hydrochloric acid molecules in the solution. 
A similar experiment was therefore tried by the present investigator with 
aqueous ethyl acetate solution and white of egg but no elevation of the 
freezing point was observed. An experiment was also tried by shaking solid 
egg-albumin (Merck’s) with aqueous ethyl acetate to see whether the amount 
of ethyl acetate was reduced. But the egg-albumin, although not readily 


614 E. R. WATSON 


soluble in water, was sufficiently soluble to make filtration very difficult. 
The experiment was therefore modified by using alcoholic solutions of various 
odorous substances, shaking them up with egg-albumin and ascertaining 
whether the amount of odorous substance in solution had decreased. At any 
rate it was expected that acids and bases would react with the egg-albumin. 
A decrease was observed in the case of hydrochloric acid but none in the case 
of ethyl acetate, acetic acid, pyridine or citronellal. It is obvious, therefore, 
that under these circumstances there is no chemical reaction between several 
typical odorous substances and the typical protein, egg-albumin. 

In looking round for a physical reaction between odorous substances and 
the olfactory organ which might be the cause of smell the author was struck 
by the parallel between intensity of smell and depression of the surface tension 
of the aqueous solutions of various organic substances investigated by Traube. 
Physiologists give much attention to the problem of how chemical substances 
can enter the living cell, for it is obvious that no reaction can take place with 
the cell-substance unless entry can be obtained into the cell. Surface tension 
and adsorption are likely to play an important part in bringing substances 
into contact with the living cell, and it was therefore very interesting to 
notice in Traube’s results that the strongest smelling substances produced 
the greatest depression of the surface tension in aqueous solutions, 

But before much progress could be made in detecting a parallel between 
smell and depression of surface tension it was necessary to find a method of 
measuring the intensity of smell, and this seemed likely to be a very difficult 
matter as common experience seems to indicate that different animals have 
very different sensibilities as regards smells. Dogs by their sense of smell 
can obviously detect things which are quite inodorous to human beings, and 
one human being seems to have a much keener sense of smell than another. 
Moreover it seemed likely that the intensity of smell of a substance would 
depend among other things on its vapour pressure. Practical psychologists 
describe an instrument in which air is inhaled by the nostrils over a definite 
area of filter-paper surface moistened by a solution of the substance to be 
tested. But before experimenting with this instrument the simpler plan was 
tried of making aqueous solutions of different strengths of the substance to be 
tested and ascertaining the minimum strength of solution in which the sub- 
stance could be detected by smelling the solution contained in an ordinary 
bottle at the ordinary temperature. Almost contrary to expectation it was 
found that consistent results could be obtained in this way, and that several 
persons obtained practically the same results. Details are given in the experi- 
mental part. 

These determinations of the intensity of the odour of the substances 
examined by Traube confirmed the conclusion previously suspected, viz. that 
those substances had the strongest smell which produced the greatest depression 
of the surface tension in aqueous solution, The intensity of odour and also. 
the surface tension of aqueous solutions of additional substances were deter- 


SMELL 615 


mined, especially of essential oils and other substances well known for their 
strong odours, and with a few exceptions the above rule was found to hold. 
The constituents of the well-known perfumes, e.g. citral and geraniol, produced 
a very great depression of the surface tension. Out of 22 substances examined 
it was found that 17 would be arranged in exactly the same order whether 
they were arranged according to increasing depression of surface tension or 
increasing intensity of odour. The list is as follows: formic acid, methyl 
alcohol, acetic acid, ethyl alcohol, propionic acid, methyl acetate, methyl- 
amine, phenol, ethyl acetate, butyric acid, iso-amyl alcohol, quinoline, 
cinnamic aldehyde, citral, allyl sulphide, geraniol. The most noteworthy 
exceptions were ammonia, pyridine and ethyl mercaptan which did not depress 
the surface tension to the extent anticipated from their powerful odours. 

The method used for the determination of surface tension was of the 
simplest character and it is well known that such determinations are liable 
to large experimental errors. Rather than spend more time on eliminating 
possible errors in surface tension determinations it was decided to make some 
determinations of adsorption. According to theory adsorption should run 
parallel with surface tension and it is obvious that from the point of view of 
contact of the odorous substance with the olfactory organ we are more 
interested in adsorption than in depression of surface tension which was 
merely used as an index of adsorption. 

Arranged according to adsorption from aqueous solution by animal char- 
coal two of the exceptions above noted disappear, viz. pyridine and ethyl 
mercaptan, which have adsorptions in keeping with the intensity of their 
odours. 


Arranged according to Arranged according to Arranged according to 
depression of surface tension adsorption by charcoal intensity of odour 


Ammonia Ammonia Formic acid 
Formic acid Formic acid {Asati acid 
Acetic acid Acetic acid Propionic acid 
Ethyl mercaptan Propionic acid Ammonia 
Pyridine Pyridine Butyric acid 
Propionic acid Butyric acid idine 
Butyric acid Ethyl mercaptan Ethyl mercaptan 
Citral Citral 


We may therefore enunciate the general rule that those substances have 
the strongest smell which are most readily adsorbed from aqueous solution. 

It was thought that by replacing animal charcoal by a protein we might 
obtain a closer approach to the olfactory organ. Some experiments on ad- 
sorption by wool from aqueous solutions were tried but did not lead to any 
result of interest. 

Although, apparently, no progress has been made in solving the problem 
outlined at the beginning of this paper, it is interesting to have found a 
connection between smell and physical properties capable of exact measure- 
ment, and it does not seem too much to hope that it may prove a first step 
in solving the problem which the author has set before him. 


616 E. R. WATSON 


EXPERIMENTAL. 


Determination of the minimum concentration of aqueous solution in 
which odorous substances can be detected by smell. 


The solutions were all placed in similar bottles which were stoppered and 
shaken just before they were smelt by bringing the nose close to the mouth 
of the bottle. In order to avoid fatigue the more dilute solutions were pre- 
sented to the observer before the more concentrated ones and in order to 
avoid the influence of suggestion he was not told what solutions were being 
presented to him. The results obtained by the author, two Indian colleagues 
(Dr A. C. Sirear and Babu S. C. Ganguli) and one Indian student (Babu J. K. 
Mazumdar) are recorded. It was more difficult to obtain consistent results 
with the fatty acids than with the other substances examined. 

The substances are arranged in order of increasing intensity of odour. 


Percentage strength in which detected by 
A 


" Substance ERW. 8.64. EM AOS. 
Formic acid a 0-1 1 1 — 
Acetic acid yaa 0-1 (0-3) 1 (0-1) 0-1 — 
Propionic acid af 1 (0-1) 1 (0-03) 0-3 a 
Aniline ae wel 0-1 — sine sae 
Methyl alcohol are 0-1 (0-012) 0-1 0-01 a 
Ethyl alcohol 0-1 0-1 0-1 — 
Methyl acetate 0-1 0-03 0-1 — 
Chloroform ... is -- 0-1 _ = 
Butyrie acid sae 0-01 0-1 (0-003) 0-1 (2) — 
Ammonia ... a 0-01 — = 0-01 
Methylamine hE 0-01 — — ce 
Ethyl acetate ore! 0-01 0-01 0-01 — 
Phenol Sas a5 — 0-01 —_ _— 
Pyridine was Bo — <0-001 an 
Quinoline ; <8 0-001 — nie oes 
Isoamy] alcohol te 0-001 0-001 0-001 -- 
Cinnamic aldehyde ... 0-001 -- — _— 
Nitrobenzene eis 0-001 0-001 — — 
Ethyl mercaptan... 0-0001 -- -- — 
Citral sar ais 0-0001 0-0001 — 0-0001 
Geraniol aan aie = 0-0001 — _ 
Allyl sulphide on 0-0001 — ie ats 
Phenyl isocyanide ... 0-00001 — — os 


Determination of surface tension of aqueous solutions of odorous substances. 


This was done by observing the rise of the solution in a capillary tube. 
In the following list the substances examined are arranged according to 
effect on surface tension, those having least effect being placed first: 


1. Ammonia 9. Methyl acetate 17. Isoamyl alcohol 
2. Formic acid 10,11 Methylamine 18, Quinoline 
3. Methyl alcohol and 12, ) Phenol 19, 20, (Ally! sulphide 
4. Acetic acid "| Aniline and 21. Citral 
6 and | Ethyl mercaptan 13. Nitrobenzene * (Cinnamie aldehyde 
6. | Ethyl aleoho 14. Ethyl acetate 22. Geraniol 
7. Pyridine 15. Butyric acid 


8. Propionic acid 16. Phenyl isocyanide 


SMELL : 617 


Adsorption from aqueous solution by animal charcoal. 


The determination was easy for acids and bases, which can readily be 
estimated by titration. An approximately normal solution (50 cc.) was shaken 
with 2 g. animal charcoal (Merck’s, washed and dried) and allowed to stand 
with occasional shaking over two nights, then filtered and titrated. Similarly 
100 ce. of an approximately decinormal solution were shaken with | g. animal 


charcoal. 
Titration value (in ce. N soln.) of 10 cc. 


Substance Original solution Final solution 
Formic acid 21-7 x 0-939 19-9 x 0-939 
17-7 x 01108 16-1 x 0-1108 
Acetic acid 16-8 x 0-939 14:9 x 0-939 
13-3. x 0-1108 11-55 « 0-1108 
Propionic acid 12-3. x 0-939 10-25 x 0-939 
10:3. x 01108 8-15 x 0-1108 
Butyric acid 8-7 x 0-939 6-6 x 0-939 
6-65 « 0-1108 4-35 x 0-1108 
Ammonia 12-4 x 1-036 11-4 x 1-036 
12:3. x 0-11036 11:35 x 0-11036 
Pyridine 9-7 x 1-036 76 x 1-036 
10-3. x 0-11036 73 & 011036 


Citral. It was found that citral in very dilute aqueous solution such as 
required for these experiments could be estimated by extracting with ether 
and titrating the ethereal extract with Hiibl’s solution. 

A solution of 0-3555 g. citral in 1 litre of water was shaken with 10 g. 
animal charcoal and allowed to stand for 24 hours. The smell of citral had 
entirely disappeared, so that concentration had been reduced to less than 
0-000001. 

A solution of 0-4 cc. citral dissolved in 1 litre was shaken with 2 g. charcoal 
and allowed to stand for 24 hours. The smell had almost entirely disappeared. 

An estimation gave the approximate results: 


Concentration in water = 0-00000256 
charcoal = 0-1782 


For the same concentration of butyric acid in charcoal we can calculate: 


Concentration in water = 0-00201 
a charcoal = 0-1782 


So that the adsorption of the citral is very much greater. 

Ethyl mercaptan. An attempt was made to estimate ethyl mercaptan in 
aqueous solution by precipitating with lead acetate, filtering off, washing 
and weighing the lead mercaptide. It was not very satisfactory owing to the 
solubility of lead mercaptide. 

Ethyl mercaptan (2 cc.) was dissolved in 250 cc. water. 

(a) 40 cc. of this solution precipitated with 1 g. lead acetate dissolved in 
15 cc. water; precipitate filtered off and washed with a measured quantity 
of water, 10 cc. at a time = 0-2800 g. (corresponds to 0-1055 mercaptan). 


618 EK. R. WATSON 


(b) 10 cc. of mercaptan solution diluted to 40 cc. and treated with 1 g. 
lead acetate solution, etc., exactly as above, using same quantity of wash 
water; lead mercaptide pbtindl 0-0472 g. 

(c) 50 cc. of mercaptan solution was shaken with 0-1 g. animal charcoal 
filtered and 40 cc. of filtrate treated exactly as above; lead mercaptide 
obtained = 0-0463 g. 

(d) 50 cc. of mercaptan solution was shaken with 0-5 g. animal charcoal, 
filtered and 40 cc. filtrate treated exactly as above; lead mercaptide obtained 
= 0-0122 g. 

Although the method of estimating mercaptan is unsatisfactory we can 
safely say that 40 cc. of the filtrate from (c) contained about the same quantity 
of mercaptan as 10 cc. of the original solution, so that 

Concentration in water = 0-00166 
- charcoal = 2-502 

From these data we can arrange the substances according to their adsorp- 
tion by animal charcoal, those which are most adsorbed being placed last in 
the list: ammonia, formic acid, acetic acid, propionic acid, pyridine and 
butyric acid (bracketed together), ethyl mercaptan and citral (bracketed 
together). 


The work described in the paper was carried out in the second half of 
1919. 


REFERENCE, 
Bugarsky and Liebermann (1898). Pfliiger’s Arch. 72, 51. 


LXIII. STUDIES ON THE PITUITARY. I. 


THE MELANOPHORE STIMULANT IN POSTERIOR 
LOBE EXTRACTS. 


By LANCELOT THOMAS HOGBEN ann FRANK ROBERT WINTON. 
(Received July 4th, 1922.) 


INTRODUCTION. 


To the manifold physiological responses evoked by extracts of the posterior 
lobe of the pituitary gland, it has lately become necessary to add that of 
inducing pigmental changes in those lower vertebrates (Fishes, Reptiles, 
Amphibia) which possess considerable capacity for temporary modification of 
bodily colour through the activity of a special type of effector organ, the 
chromatophore (melanophores, etc.). Recent work on the part played by 
endocrine organs in amphibian metamorphosis has in particular brought to 
light a number of interesting data relative to the factors which control pigment 
response in these organisms. One of the most characteristic consequences of 
pituitary removal in tadpoles noted by Allen [1917] and others was a condi- 
tion of extreme pallor resulting from general contraction of the dermal 
melanophores (black pigment cells). More recently Swingle [1921] has shown 
that implantation of the pituitary in pale tadpoles induces a darkening of 
the skin, a result shown to follow injection of posterior lobe extracts into 
adult frogs by the writers [1922]. The present communication attempts to 
explore the use of frog melanophores as indicators of pituitary extracts and 
to emphasise the remarkable specificity of this reaction. Few people are aware 
how marked is the capacity for pigmental change possessed by the common 
frog (R. temporaria), the same individual varying from a light yellow to a 
coal black or dark grey tint within comparatively short time limits: the 
difference is mainly due to the “expansion” and “contraction” of the dermal 
melanophores. In general between temperatures of 0-25° frogs placed in 
bright illumination are pale if kept dry whether on a light or dark background. 
They may become dark if placed on a dark background in water. For the 
study of pigmental darkening by pituitary injection, it is best to leave the 
animals singly in dry glass dipping jars for a few hours in bright light on a 
white background to ensure maximal contraction of the melanophores, 
Observations can be supplemented by microscopic preparations of skin, fixed 
in Bouin’s fluid, dehydrated, cleared and mounted without more elaborate 
treatment. In earlier experiments [Hogben and Winton, 1922] decerebrated 


620 L. T. HOGBEN AND F. R. WINTON 


animals were used. For injection Burroughs Wellcome’s all glass tuberculin 
syringe graduated in 0-1 cc. was found to be most convenient: unless otherwise 
stated the intraperitoneal method was adopted. 

In these experiments the frogs were stored in white porcelain tanks with 
glass tops, and were thus always ready for use. To facilitate observation 
individuals in which the skin is uniformly of a yellowish hue when pale, rather 
than the varieties with irregular markings, spotted or speckled aspect should 
be selected. 

QUANTITATIVE EXPERIMENTS. 

The following experiments were carried out to determine how far the 
melanophore response might be utilised to estimate the activity of posterior 
lobe extracts. The preparations employed were Burroughs Wellcome’s 0-5 ce. 
phials of liquid sterile posterior lobe (21 %) extract (“infundin”) stated 
to be standardised by action on the isolated mammalian uterus. Injection with 
0-5 ce. of 1/1000 solution of this preparation invariably invoked the character- 
istic darkening of the skin, while 0-5 cc. of a 1/10,000 solution failed to do so 
in every case. The following preliminary experiments in which the animals 
employed were not weighed may be quoted to indicate the range of sensitivity 
of the frog’s melanophores to pituitary extracts and the relative sensitivity 
with different methods of injection. The average temperature was 72° F. 

Sample A. Three series of six frogs received an intraperitoneal injection 
equivalent to 0-00031, 0-00025, 0-00019, 0-00012, 0-00009 and 0-00006 cc. of 
infundin respectively. In two series the animals which received 0-00012 ce. 
and upwards underwent darkening of the skin and in the other series those 
which received 0-00019 cc. and upwards gave a positive reaction. 

Sample B. Two series of six frogs were injected with 0-00025, 0-0002, 
0-00015, 0-0001, 0-000075, 0-00005 ce. of infundin. In one series (intra- 
peritoneal) the individuals which received 0-00015 ce. and upwards showed 
visible darkening of the skin accompanied, as seen on microscopic examination, 
with full expansion of the melanophores. The individual which received 
0-0001 cc. displayed the melanophores in a slightly expanded (stellate) con- 
dition, while in the remaining frogs the melanophores were fully contracted. 
In the remaining series (dorsal lymph sac injection) microscopic examination 
revealed a precisely similar seriation, but the frog which received 0-0001 ce. 
showed a slight visible darkening which was not noticeable in the corre- 
sponding animal of the preceding group. 

Sample C. Six frogs received an intraperitoneal and six an intravenous 
injection equivalent respectively to 0-0005, 0-0004, 0-0003, 0-00025, 0-0001 
and 0-00005 ce. of B.W. liquid sterile extract. In both series the animals 
which received 0-0005, 0-0004 and 0-0003 ec. doses reacted with full expansion 
of the melanophores; a slight darkening of the frogs injected with 0-00025 
and 0-0001 ce. was visible, but microscopic examination revealed the expan- 
sion of the melanophores in the former though not in the latter case. The 
animals injected with a dose of 0-00005 cc, did not react at all. 


STUDIES ON THE PITUITARY 621 


In these experiments summarised in Table I the frogs used were in no 
way selected, and it may be inferred that such methods are suitable for 
detecting the presence of posterior lobe extracts in quantities exceeding that 
equivalent to5 x 10->g. of fresh gland (ox) substance, 7.e. about 0-05 % of a 
clinical dose. In view of the relation between sensitivity and size indicated 
below it will be seen that by selecting sufficiently small individuals (or species 
of Amphibia), it should be possible to detect posterior lobe secretions in 
considerably smaller quantities than this and therefore in amounts that lie 
much beyond the limits of sensitivity of methods at present employed. We 
have found no evidence of a difference in the sensitivity of the reaction with 
the three methods of injection described. 


Table I. 
Sample Method of injection Minimal dose Subminimal dose 
ce. ce. 
A Intraperitoneal 0-00012 0-00009 
A Intraperitoneal 0-00012 0-00009 
A Intraperitoneal 0-00019 0-00012 
B Intraperitoneal 0-00015 0-0001 
B Dorsal lymph sac 0-00015 0-0001 
C Intraperitoneal 0-00025 0-0001 
Cc Intravenous 0-00025 0-0001 


Sample D. Two series, each of six frogs, were weighed and arranged in a 
heavy and a light series. Doses were injected so that pairs, one from each 
series, received the same dose per unit body weight—the results are shown 
in Table II, and demonstrate proportionality between body weight and the 
minimal effective dose. 


Table IT. 
Approximate 
ose per Heavy series Light series 
weight Wt. Dose Response Wt. Dose Response 
ce. g. ce. g. ce. 
0-00025 31 0-00037 Darkening 16 0-0002 Darkening 
0-0002 30 0-0003 - 15 0-00015 w 
0-00015 25 0-0002 Pe 15 0-00011 be 
0-00015 23 0-00018 re 14 0-0001 ” 
0-0001 23 0-00012 — 13 0-00007 _ 
0-00005 21 0-00005 -- 12 0-00003 — 


Sample E. (See Table III.) Eight frogs of 18-20 g. body weight received 
0-0002, 0-00016, 0-00014, 0-00012, 0-0001, 0-00008, 0-00006, 0-00004 cc. infundin 
(intraperitoneal). The first five showed darkening: the rest remained pale, 
i.e. the minimal dose to produce maximal expansion of the melanophores lay 
between 0-0001 and 0-00008 ce. Six other frogs weighing 15-16 g. were injected 
via the peritoneum with 0-0002, 0-00015, 0-0001, 0-000075, 0-00005, 0-000025 
cc. infundin. The first four only showed darkening of the skin, 7.e. the minimal 
dose lay between 0-000075 and 0-00005 cc. A third set of eight frogs weighing 
13-14 g. were tested from the same sample of infundin with the following 
doses (intraperitoneal): 0-0002, 0-00015, 0-0001, 0-00008, 0-00006, 0-00004, 

Bioch. xvi 41 


622 L. T. HOGBEN AND F. R. WINTON 


0-00002, 0-00001 cc. The melanophores were fully expanded in the first three 
and contracted:in the last three, while the frogs which received 0-00008 and 
0-00006 ce. were found to display the pigment cells in a condition of partial 
expansion. 

Sample F. (See Table III.) The following doses were administered to 
twelve individuals. A 1 (30 g.) and A 2 (25 g.), 0-00035 cc.; B1 (30 g.) and 
B2 (26 g.), 0-0003 cc.; C1 (30 g.) and C2 (26 g.), 0-00025 cc.; D1 (29 g.) 
and D 2 (27 g.), 0-0002 cc.; E 1 (29 g.) and E 2 (28 g.), 0-00015 cc.; and F 1 
(28 g.), and F 2 (28 g.), 0-0001 cc. infundin. The pairs A, B, C, D, uniformly 
displayed an intense darkening of the skin after half an hour had elapsed. 
The two pairs E and F on the other hand remained pale. This set of animals 
was deliberately chosen for large size; none of the females were carrying ripe 
eggs at the time. 

The results obtained from the three foregoing experiments, although 
obtained with three different samples of “infundin,”’ and therefore hetero- 
geneous, present sufficient uniformity to be instructive in connection with the 
problem of what order of error is introduced in the method. It will be seen 
that the interval between minimal and subminimal doses, representing the 
margin of uncertainty, can usually be reduced to 25 %-30 % of the minimum 
effective dose; occasionally however, as in the third series of experiment (£), 
unsuitability of frogs may extend this even to 60 %. 

That the simple correlation between body weight and minimal dose does 
not invariably obtain, is illustrated by a further experiment which was under- 
taken with a view to determining the sensitivity of the reaction to a dose 
injected about two hours subsequent to a previous dose in the neighbourhood 
of the minimal dose. 

Sample G. (See Table III.) A preliminary experiment with four frogs 
showed that this sample was somewhat more potent than some used pre- 
viously, and suggested the doses described. 32 frogs were weighed, and divided 
into four groups. 

Eight frogs (9-11 g. body weight) were injected respectively with 0-0002, 
0-00015, 0-000125, 0-0001, 0-000075, 0-00005, 0-000025, 0-00001 ce. infundin, 
Visible darkening of the skin occurred in the first five, the remaining three 
were unaffected. The minimal dose was in this case between 0-000075 and 
0-00005 ce, 

Kight frogs (12-13 g. body weight) were injected with 0-00025, 0-0002, 
0-00015, 0-000125, 0-0001, 0-000075, 0-00005, 0-000025 ec. infundin. The first 
three only showed visible darkening of the skin. The minimal dose was thus 
between 0-00015 and 0-000125 ce. 

Kight frogs (14-15 g. body weight) were injected with the same series of 
doses. The first four in this instance displayed visible pigmental change, the 
dose necessary being between 0-000125 and 0-0001 ce. 

Kight frogs (17-19 g. body weight) were injected with 0-00035, 0-0003, 
000025, 0-0002, 0-00015, 0-0001, 0-000075 and 0-00005 ec. The first three 


STUDIES ON THE PITUITARY 623 


only showed darkening of the skin, 7.e. the minimum dose for this series was 
between 0-00025 and 0-0002. 


Table III. 
Actual dose Dose per 20 g. body weight Error 
a A r A in 
Sample Wt. Min. Submin. Wt. Mean min. Min. Submin. min. 
g. ce. ce. g. ce. ce. ce. % 
D 23 0-00017 0-00012 23 é : : 
D 14 0-0001 0-00007 13 0-00015 0-00015 0-0001 0 
E 18 0-0001 0-00008 20 0-0001 0-00008 —17 
E 15 0-000075 0-00005 | 0-00012 40-0001 0-00007 —17 
E 13 0-0001 0-00004 14 0-00015 0-00006 + 25 
F 29 0-0002 0-00015 pid 0-000145 0-00014 0-0001 — 3 
F 27 0-0002 0-00015 28 (| 0-00015 0-00011 + 3 
G 19 0:00025 0-0002 17 0-00025 0-00023 +14 
G 15 0-000125 0-0001 14 0-00017 0-00015 — 21 
G 12 0-00015 0-000125 12 0-00025 0-00021 +15 
G 8 0-000075 0-00005 10} 0-000215 4 0-00019 0-0001 —1l 
G 8 0-000075 0-00005 10 0-00019 0-0001 —ll 
a, aod 
17 0-0002 0-00015 18 .0-00023 0-00017 + 7 


(repeat) 
Mean deviation 11-8 %. 


Effect of subsequent doses. After injection with concentrated extracts the 
condition of melanophore expansion may last 12 hours or more. After injection 
with minimal doses it passes off in about 1-24 hours. After 1? hours following 
the first injection, the last series injected with sample @ received corresponding 
doses of the same magnitude. The minimal dose was slightly lower, 0-0002— 
0-00015 ce. On injecting the first series two hours after the initial treatment 
the minimal dose was the same as before. It thus appears: first that given 
sufficient time for recovery the sensitiveness of the melanophores is not appre- 
ciably changed by previous treatment, second, that the effect of repeated doses 
is, in contrast with the action of the pressor principle [ Howell, 1898], identical. 

Like Dale and Laidlaw’s original method! for estimating the uterine 
stimulant, this method is appropriate to comparison rather than absolute 
standardisation of extracts. Several factors which may lead to anomalous 
results might be considered, such as possible seasonal variation of sensitivity?, 
and variations depending on the immediate previous history of the frogs with 
respect to moisture, temperature, etc. Again, different strains of frogs may 
respond with varying facility, and we have certainly noted that some are 
more and others less satisfactory in giving a sharply defined end point and 
well contrasted reaction. Furthermore, isolated frog’s skin placed in Ringer 
shows the characteristic response to posterior lobe extracts, and may react 
with more constant sensitivity than an intact animal; these and other relevant 
considerations we have at present no opportunity to examine. The specificity, 

1 In Dale and Laidlaw’s [1912] standard method for estimating the activity of pituitary extracts 
by their action on the isolated guinea pig’s uterus, the minimal dose of infundin which produces 
a maximal contraction varies between -003--005 cc. which is diluted to about 100 ce. 


2 Compare the striking variation in the electrical phenomena of frog’s skin which occurs 
during the breeding season—[ Bayliss and Bradford, 1886]. 


41—2 


624 L. T. HOGBEN AND F. R. WINTON 


certainty and delicacy of the melanophore response will recommend it for 
such biochemical and pharmacological purposes as do not require a degree of 
accuracy higher than that we have described, namely consistency of minimal 
dose, to the extent of a mean deviation of 12 %. In each of the foregoing 
experiments the actual number of frogs used was the minimal number requisite 
for demonstrating the end point in a series. It is evident that in testing 
samples on a large scale the degree of activity might be compared with greater 
confidence by employing a larger number of individuals. In any case a simple 
and direct method is provided for testing the retention of activity of pituitary 
products employed in the dispensary or hospital. 


SPECIFICITY OF THE MELANOPHORE STIMULANT. 


Concerning the mode of action of the melanophore stimulant there is no 
need to add anything further in this place to our previous observations except 
to emphasise the local response obtained on isolated strips of skin placed in 
Ringer’s extracts. It is our intention to publish in the near future a compre- 
hensive account of the reactions of frog melanophores to drugs; but it is of 
special interest to insert at this point a few observations under this heading 
to lay stress on the highly specific nature of the response evoked in amphibian 
melanophores by posterior lobe extracts. The physiological effects hitherto 
described as characteristic of posterior lobe extracts are equally characteristic 
of a group of other well-known drugs. In the case of the melanophore reaction 
described in this communication we have to deal with a property of the 
extract which is not precisely simulated by any of the more familiar pharmaco- 
logical reagents, as the accompanying table signifies. The case of digitalis 
merits special comment. The water-soluble digitalin (Parke Davis) was 
employed. In a dosage of 1 mg. death followed 30-45 minutes after injection 
via the dorsal lymph sac and no colour change resulted. 0-5 mg. was generally 
fatal after a few hours and accompanied in one case by a partial expansion of 
the melanophores. Doses of 0-3 mg. (i.e. one half the minimum lethal dose) 


Table IV. 
Pale frogs were injected with doses of the drugs undermentioned. 

(a) Pilocarpine ... oes 25 mg., 12-5 mg., 2-5 mg. bbs ote Remained pale 
(6) Atropine dis is 10 mg., 3 mg., 1 mg. wae vue » 
(c) Strychnine ... 1 mg., 1/25 mB» 1/100 mg. ats ae ” 
(d) Curare (decerebrated) Cura: rised she sen ~ 
(ec) Histamine... 3-6 mg., 36 mg. Ms 036 mg. Vea + < 
(f) Ergatoxin ... ons 6-5 mg., 65 mg., 065 mg. one ie ” 
(9) re 2 vee cas 20mg.,2mg.,"02 mg. ... aT bie *» 
(h) Veratrine ois ey: 1/4 mg., 1/20 mg., 1/100 mg... oa 
(i) Digitalis oP oes Numerous doses of varying er below 

M, L.D. (of. text) eee * ” 
() Barium chloride eve 1 mg., 1/5 mg: abe dak oie aed ” 
(k) Sodium nitrite éha 5 mg. * an vie ” 
(1) Caffeine bes nie 10 mg., 2 mg. ” 


2-5 mg.—5 mg. without motor aralysis 

10 mg. motor paralysis within si 0 mins. Slight darkening 

(n) Nicotine {1/10 mg. no motor paralysis —_... ses Remained pale 
wy . (1/5 mg.—1 mg. with motor paralysis... Partial darkening 


(m) Apocodeine 


STUDIES ON THE PITUITARY 625 


or less produced absolutely no effect on coloration. But no effect was obtained 
on the isolated skin. A partial expansion usually but not invariably occurs 
with nicotine and apocodeine only in dosage sufficient to produce motor paralysis 
without immediate lethal effects. In none of these cases is the reaction readily 
comparable with the pituitary reaction. 


Some CHEMICAL PROPERTIES OF THE MELANOPHORE STIMULANT. 


Experiments were carried out to test how far the chemical properties of 
the pituitary melanophore stimulant agree with those of the other active 
constituents of the posterior lobe extracts. 

Inactivation by boiling with dilute HCl. There appear to be at least two 
autacoid substances contained in extracts of the posterior lobe, since it is 
possible to destroy or reduce the pressor activity of “infundibular” extracts 
without diminishing in a corresponding manner their power to excite plain 
muscle. Dudley [1919] by extraction of dried and powdered infundibulum 
with acidified water, treatment of the solution with colloidal ferric hydroxide, 
and subsequent continuous extraction of the filtrate at reduced pressure with 
butyl alcohol, succeeded in separating a residue containing all the uterine 
stimulant (‘“‘oxytocic” principle) together with a portion of the pressor sub- 
stance. The latter has been shown by Abel and Nagayama [1920] and Dale 
and Dudley [1921] to be rapidly destroyed by boiling with 0-5 °% HCL The 
oxytocic principle on the other hand is only slowly destroyed by such treat- 
ment. In this respect it appears that the melanophore stimulant is like the 
uterine principle and is not identical with the pressor substance. 

The effect of continued boiling with dilute HCl was investigated as follows. 
A 0-5 % solution of the commercial extract was made up in 0-5 % HCl. This 
mixture was subjected to continuous boiling for five hours, a sample being 
removed at the end of thirty minutes. At the conclusion of the experiment a 
sample of the unboiled mixture, the portion which had been subjected to only 
half an hour’s hydrolysis, and the residue were respectively neutralised with 
soda and diluted to a concentration approximately isotonic with frog’s Ringer. 
From each of the three solutions, A (unboiled), B (boiled half an hour), 
C (boiled five hours), 0-5 ec. was injected into a pair of frogs whose pigment 
cells were fully contracted. The macroscopic and microscopic examination of 
the six animals at the conclusion of an hour revealed a marked contrast. The 
A and B pairs were dark and showed the typical pituitary reaction. The pair 
C remained pale. Microscopic preparations of the skin showed that in the 
C pair the melanophores were fully contracted, and the A and B pairs expanded. 
The result of the experiment indicates that pituitary extracts retain a con- 
siderable potency to induce melanophore response after half an hour’s boiling 
with 0-5 % HCl; hence the melanophore stimulant is not identical with the 
pressor substance. According to Abel and, Nagayama the oxytocic activity 
of the pituitary extracts after being reduced to about one-fifth by hydrolysis 


626 L. T. HOGBEN AND F. R. WINTON 


for 30 minutes resists further destruction. Dale and Dudley however did not 
find that the residue was stable to prolonged boiling with acid; and the 
behaviour of the melanophore stimulant in this experiment conforms to their 
experience of the uterine stimulant. 

The oxytocie potency of the samples taken after 30 minutes’ boiling with 
0-5 % HCl was found by Dale and Dudley to be between a quarter and a fifth 
of the previous activity. The potency of pituitary extract to induce melano- 
phore expansion after half an hour’s hydrolysis appears to be reduced in 
somewhat similar proportions. In a further experiment an approximation to 
the amount of melanophore stimulant remaining after half an hour’s boiling 
with 0-5 % HCl was investigated thus. Three portions A, B, C (15 cc.) were 
taken from a 2 % “‘infundin” solution. A was boiled for half an hour, after 
which 15 cc. 1 % HCl and sufficient 1 % soda to neutralise it were added. 
To B were added 15 cc. of 1 % HCl, which was neutralised before boiling 
for 30 minutes. C was acidified with 15 cc. of 1 % HCl, boiled for half an 
hour, and subsequently neutralised. Each solution thus contained originally 
0-3 ec. of sterile extract. A and B were made up to two litres so that an 
injection of 0-3 cc. would be equivalent to 0-00015 cc. infundin. C was made 
up to half a litre: one-half of this (C 1), representing a dilution one quarter of 
A and B, was used for injection, the other half was diluted one in four (C 2). 
Twenty-eight frogs were injected as in the following table from which it 
appears, on making the correction necessary for body weight, that the samples 
A and B (controls) were about four times as potent as C, the solution which 
had been subjected to boiling with 05% HCl ~~ 

The positive sign signifies darkening of the skin, the numbers in brackets 
refer to the body weight in grams, and the extreme left hand column gives 
the dosage in ce. injected. 


Table V. 
A B 01 C2 
0-8 ce. + (12). (12) + (12) + (11) 
0-7 ,, + (13) has) + (13) — (11) 
0-6 ,, + (14) + (13) + (14) — (11) 
0-5 ,, + (16) + (16) + (16) — (11) 
0-4 ,, + (17) + (18) + (17) 
03 ,, — (18) + (19) — (18) 
0-2 ,, — (20) — (20) + (20) 
O1,, — (23) — (23) — (23) 


Action of proteoclastic enzymes. Perhaps the most significant fact indicative 
of the type of chemical compounds to which the infundibular autacoids are 
allied is furnished by the action of proteoclastic ferments. Schafer and 
Herring [1906] stated that pepsin destroys the pressor activity of pituitary 
extracts leaving intact the diuretic principle. They denied further that 
trypsin affected either. Later work by Dale [1909] and Dudley [1919] indicates 
that trypsin rapidly destroys both the pressor and oxytocie principles. To 
test the action of proteoclastic ferments on the melanophore stimulant six 
solutions were made up: A (0-5 % infundibular extract in 0-2 % HCl and 


STUDIES ON THE PITUITARY . 627 


0-5 % pepsin); B (0-5 % infundin in 0-2 % HCl); C (0:2 % HCl and 0-5 % 
pepsin); D (0-5 % infundin in 0-5 % saline trypsin); H (0-5 % saline trypsin); 
F (0-5 % infundin in 0-5 % boiled trypsin). After two hours’ digestion A, B 
and C were neutralised, boiled and diluted till isotonic with Ringer. D, F 
were boiled. On injection into pairs of pale frogs, the pairs injected with A, 
B and F showed darkening of the skin with expansion of the melanophores. 
The rest remained pale with the melanophores contracted. Thus the melano- 
phore stimulant is rapidly destroyed by trypsin and is not rapidly destroyed 
by pepsin. In this connection evidence is provided that the melanophore 
stimulant is not identical, as Abel and Kubota [1919] once believed the 
uterine principle to be, with histamine, since the latter (Dudley) is ‘not 
destroyed by tryptic digestion. As a matter of fact histamine does not in any 
concentration cause melanophores to expand. 

Relation to other pituitary autacoids. In view of the possible identity of 
the diuretic and pressor substances, and the likelihood that the oxytocic 
stimulant is responsible for the galactogogue action of posterior lobe extracts, 
it is not necessary to postulate the existence of more than two distinct auta- 
coids secreted by the juxtaneural epithelium, It is possible that the melano- 
phore stimulant is not identical with either the pressor or the oxytocic 
autacoid, but there is strong evidence that it is not identical with the former, 
namely (a) its slow rate of inactivation by boiling with dilute hydrochloric 
acid; (b) no reversal of response after successive action; (c) failure of drugs 
which agree with pituitary extracts in their effect on the blood pressure to 
evoke the same reaction. For the present therefore we may provisionally 
attribute the melanophore response to the oxytocic substance. 


' SOURCE OF THE Prrurrary MELANOPHORE STIMULANT. 


In Swingle’s [1921] experiments upon the implantation of the pituitary 
gland in amphibian larvae, the pars intermedia was found to be responsible 
for the darkening of the skin. The results obtained from injections are con- 
sonant with the conclusion that the juxtaneural epithelium is the source of 
the secretion which induces melanophore expansion. 

The demarcation between the three parts of the gland in the case of ox 
pituitaries is very striking; and ox pituitaries secured within an hour of 
killing were therefore employed in the following experiment. The three 
portions of the gland were carefully separated: each portion was weighed, 
ground up with sand, and extracted with Ringer’s solution at 35° for two 
hours. A 0-1 % and a 0-02 % solution were made up from each extract so 
prepared. A pair of frogs were injected with each of the six solutions. At 
the end of half an hour the two pairs injected respectively with the strong 
and weak anterior lobe extracts remained pale. The pair injected with a 
weak extract (0-02 %) of pars nervosa extract also remained pale. While the 
pairs injected with both weak and strong extracts of pars intermedia as well 
as the pair injected with the stronger pars nervosa extract were conspicuously 


628 L. T. HOGBEN AND F. R. WINTON 


darkened through, as microscopic examination displayed, the general expan- 
sion of the dermal melanophores. In view of the non-glandular character of 
the pars nervosa, and the higher concentration of the melanophore stimulant 
in the pars intermedia, it seems likely that the melanophore stimulant is 
secreted by the latter portion of the gland, diffusing rapidly into the former. 


PHYLETIC DISTRIBUTION OF THE MELANOPHORE STIMULANT. 


The following observations indicate the occurrence of the same melano- 
phore stimulant in the pituitary gland of various classes of Vertebrata. 

Mammalia. An adult female rabbit was decapitated, and its pituitary 
gland together with pieces of muscle, brain, ovary, pancreas, liver, suprarenals 
and spleen removed instantly. The tissues mentioned were severally ground 
with sand and extracted with Ringer in the usual way, after having been 
first weighed. A pair of frogs was injected with 0-5 cc. per individual of a 
1 % solution of each extract. An additional pair received a corresponding 
amount of a putrid meat extract, while another pair received 0-5 cc. of a 
0-1 % solution of the same pituitary extract. The two pairs injected with 
1 % and 0-1 % pituitary extracts alone displayed a visible darkening of the 
skin, and, upon microscopic examination, the expansion of the melanophores. 

Birds. Pituitaries were removed from eight ducks, weighed, ground up 
with sand, extracted and made up to 2 %, 0-4 %, and 0-08 % solutions which 
were administered respectively to pairs of pale frogs (0-5 ec. per individual). 
The animals injected with the two stronger solutions displayed the character- 
istic pigmentary response: the pair injected with the 0-08 % extract however 
remained pale. 

Reptiles. Into each of a pair of frogs an injection of an extract of the 
pituitary of the lizard (Lacerta viridis) was made. The characteristic reaction 
was evoked. Two other frogs injected with an extract of brain tissue as 
controls, remained pale. 

Amphibia. The experiment with frog’s pituitary is of special interest as 
bearing on the réle of the juxtaneural epithelium in normal pigmentary 
responses. Six frogs’ pituitaries were removed and made up to 5 ce. in Ringer’s 
solution. An injection of 0-5 cc. into each of a pair of pale frogs evoked an 
intense darkening of the skin, The remaining solution was diluted 1-5 in 10; 
and a pair of frogs injected with 0-5 cc. of this solution also underwent 
darkening of the skin. The 1-5/10 solution was again diluted 1/5. A pair of 
frogs injected each with 0-5 cc. of the latter showed in one case a darkening, 
while the other individual remained pale. On further dilution no response 
was induced by injection. It appears from the foregoing data that the 
amount of pituitary melanophore stimulant in the gland of one frog is not 
only sufficient to account for the darkening of the skin in one animal of the 
same species but in as many as fifty-six individuals. 

‘ishes. Pituitaries were removed from six cods (Gadus) and were made up 
to 2%, 02 % and 0-04 % extracts. Two pairs of frogs injected with the 
usual amount of the two stronger solutions displayed the characteristic 


STUDIES ON THE PITUITARY 629 


darkening of the skin. Only one of two individuals injected with 0-5 cc. of 
the 0-04 % extract underwent pigmentary darkening: a pair injected with a 
5 % extract of cod’s brain remained pale. 

Tunicata. The widespread distribution of the pituitary melanophore 
stimulant in the Vertebrate series encouraged the search for a similar substance 
among the lower Chordata; and extracts of the so-called dorsal tubercle or 
subneural gland of the Tunicate, regarded by morphologists as the homologue 
of the Vertebrate hypophysis, were accordingly prepared with this end in 
view. Ascidiella, a solitary Ascidian about 14 inches in length, was selected 
for the purpose. The dorsal tubercle in this form is very large and quite 
conspicuous through the translucent mantle after the gelatinous test has been 
removed. The organ in question was removed from twenty-four individuals 
and a 10 cc. extract was made. Injection into pale frogs was not followed 
by any pigmentary change. 


REACTION OF CHROMATOPHORES TO PITUITARY EXTRACT. 


The observations above mentioned refer to the behaviour of the dermal 
melanophores of the frog, and in their reactions to pituitary extracts these 
resemble those of other Amphibia, of which we have examined Hyla, Bombinator 
(Anura) and Amblystoma (Urodela). The frog possesses in addition at least 
three other types of pigmentary effector organs: (a) epidermal melanophores, 
responding to pituitary extracts in a manner similar to the dermal melano- 
phores, though their sensitivity may be different; (b) the dermal xantho- 
leucophores (yellow pigment cells) which respond in the opposite sense, 1.e. 
contracting after pituitary injection—this property not being shared by Hyla 
in which the xantholeucophores appear to be non-contractile; (c) the retinal 
pigment cells. Fujita and Bigney [1918] have shown that these expand after 
treatment with adrenaline, an observation which we have confirmed. Pituitary 
injection hasno effect upon them either in the expanded or contracted condition. 

The fact that the frog itself possesses two types of chromatophores which 
differ diametrically in their mode of response to posterior lobe extracts is of 
interest in view of the fact that Spaeth [1917] describes the contraction of 
the melanophores of the fish (Fundulus) after action of the pituitary autacoid 
in certain circumstances. In view of the conformity of data derived from 
glandular extirpation and injection in the case of Amphibia, Spaeth’s obser- 
vations require in our belief reconsideration, though the above-mentioned 
fact does not permit one to regard his conclusions as necessarily improbable. 


SUMMARY. 

1. The pars intermedia of the mammalian pituitary secretes a specific 
stimulant inducing expansion of the dermal melanophores in Amphibia: this 
property is shared by extracts from the pituitary of birds, reptiles, amphibia 
and fishes: it was not found in the subneural (hypophysial) gland of Tunicates. 
Sufficient can be extracted from the gland of one frog to induce intense visible 
darkening in more than fifty other pale individuals of the same species. 


630 L. T, HOGBEN AND F. R. WINTON 


2. The action of pituitary extracts on melanophores is one of extreme 
delicacy, and can be used to detect with certainty the presence of posterior 
lobe secretions in quantities equivalent to 5 x 10-°g. fresh glandular sub- 
stance (7.e. about 1/2000th ordinary clinical dose) by their power to induce 
intense darkening of the skin in pale frogs. 

3. A method of approximate quantitative estimation of potency of 
posterior lobe extracts is indicated by experiments on the sensitivity of the 
melanophore response of frogs, which further show within what limits of error 
such a procedure may be successful. The delicacy of response does not depend 
on the mode of injection, nor is it affected by previous injection of minimal 
doses of the extract; it is however correlated with body weight. 

4, Whereas the physiological effects hitherto described as characteristic 
of posterior lobe extracts are equally characteristic of a group of other well- 
known drugs (e.g. digitalis, BaCl,, histamine), the melanophore response 
described above is a property which is not precisely simulated by any of the 
more familiar pharmacological reagents. An effect in the same sense is indeed 
produced only by nicotine and apocodeine in doses sufficient rapidly to produce 
motor paralysis, and even here the darkening is much less conspicuous. 

5. The melanophore stimulant in pituitary extracts is rapidly inactivated 
by tryptic but not by peptic digestion; its activity is diminished by boiling 
with 0-5 %% HCl at about the same rate as is that of the uterine stimulant, 
but much more slowly than that of the pressor substance. The probable 
identity of the uterine and melanophore stimulants may be inferred. 

6. In view of this probable identity of the uterine and melanophore 
stimulants, the accessibility of the materials, facility of manipulation, speci- 
ficity and shortness of time implied by the method involving this response, 
it would seem to be serviceable for testing the activity of posterior lobe 
extracts in use in the laboratory and hospital. 

The experiments were carried out in Prof. MacBride’s laboratory at the 
Imperial College of Science, and the writers were advised on several points 
by Dr H. H. Dale, F.R.S., to whom grateful acknowledgment is made. The 
expenses of the research were defrayed by a grant made by the Government 
Grants Committee of the Royal Society. 


REFERENCES, 


Abel and Kubota (1919). J. Pharm. Exp. Ther. 18, 243. 
Abel and gs ee (1920). J. Pharm. Rup, Ther. 15, 347. 
Allen (1917). Biol. Bull. 32, 117. 

Bayliss and Bradford (1886). J. Physiol. 7, 216. 

Dale (1909). Biochem. J. 4, 427. 

Dale and Dudley (1921). J. Pharm. Bap. Ther. 18, 27. 
Dale and Laidlaw (1912). J. Pharm. Rup. Ther. 4, 75. 
Dudley (1919). J. Pharm. Rup. Ther. 14, 295. 

Fujita and Bigney (1918). J. Bap. Zool. 27, 391. 
Hogben and Winton (1922). Proc. Roy. Soc. B. 98, 318, 
Howell (1898). J. Bap. Med. 8, 2. 

Schafer and Herring (1906). Proce. Roy. Soc. B. 77, 57). 
Spacth (1918). J. Pharm. Rap, Ther. 11, 209. 

Swingle (1921). J. Hap. Zool. 34, 119. 


LXIV. ON THE SIGNIFICANCE OF VITAMIN A 
IN THE NUTRITION OF FISH. 


By KATHARINE HOPE COWARD (Beit Memorial Research Fellow) 
anp JACK CECIL DRUMMOND. 


From the Biochemical Laboratories, Institute of Physiology, 
University College, London. 


(Received July Sth, 1922.) 


Ir has recently been shown that the relatively large amounts of vitamin A 
found in the liver of the gadoid fishes is probably derived from their food 
and ultimately from the marine algae which synthesise this dietary factor. 
[Coward and Drummond, 1921; Hjort, 1922; Jameson, Drummond and 
Coward, 1922; Drummond and Zilva, 1922.] That the vitamin A probably 
fulfils some réle in the development of the young fish is indicated by its 
presence in the gonads [McCollum and Davis, 1915; Hjort, 1922] and recently 
it has been shown that the amount stored in the reproductive cells of such a 
fish as the cod is extraordinarily large [Zilva and Drummond, 1922]. To study 
the fate of the vitamin in the fish ova it was decided to work with the brown 
trout which is easy to rear and maintain in the ordinary laboratory. 


VITAMIN A IN Trout Eaas. 


Feeding tests on rats by the usual procedure demonstrated that trout 
eggs are very rich in vitamin A. A dose of two eggs per day per rat resulted 
in a resumption of normal growth, and it is to be regretted that tests were 
not made with smaller doses of the material, but none had been reserved 
( Fig. 2). 

REARING OF YounG TROUT IN THE LABORATORY. 

A thousand fertilised eggs were obtained from the Trout Fisheries, Stirling, 
and were allowed to hatch out in a single trough which we are describing 
because it enables trout to be reared with success for considerable periods of 
time in an ordinary laboratory, and we have encountered many cases where 
workers have failed to achieve this. All the eggs were placed in a flat boat- 
shaped receptacle (Fig. 1) 18’ long and cross section 5’ x 5’’. The sides are 
of wood but the ends and bottom are made of narrow-mesh zine gauze. Such 
receptacles are placed in an ordinary porcelain sink of which the outlet has 
been adjusted so as to maintain a depth of half aninch of water whilst a constant 
stream of water is supplied by means of a pipe leading from the tap. The 


632 K. H. COWARD AND J. C. DRUMMOND 


incoming stream issuesfroma glass tube of half-inch diameter and is not directed 
centrally down the length of the “boat” but at an angle of about 20° to that 
axis. By this means the water stream does not pass directly out at the far 
end of the ‘‘boat” but tends to form a current round the sides. In this manner 
the unconsumed food, which must always be borne to the fish down stream, 
is carried round several times before being washed down the outflow pipe 
and wastage is thereby prevented. 

The sinks and “boats” are kept scrupulously clean and the water current 
maintained without interruption. Partial shade for the trout is provided by 
placing a board crosswise over part of the sink. If these precautions are 
followed it is possible to rear the young trout almost without loss, at any rate 
for several months. 


Fig. 1. 


VITAMIN A IN LARVAL TROUT. 


The ova hatched out with few exceptions and the larvae were allowed to 
develop without any extra food being given for four weeks. The depth of 
water was increased to about one inch. It is difficult to obtain information as 
to the feeding habits of young fresh-water fish, and as far as we are aware 
there are available no careful studies analogous to Dr Lebour’s investigations 
on marine fish [1918; 1919, 1, 2]. It is commonly believed that larval trout 
take no food during that stage of their development and that they are purely 
carnivorous when they begin to feed as soon as the yolk sac is nearly absorbed. 
Somespecimensof young brown trout about one month old from atrout hatchery 
were examined by us and were found to contain definite remains of partially 
digested green microscopic plants in their alimentary tracts. It is possible 
that these were taken in accidentally by the fish or that they were previously 
ingested by minute crustaceans which the fish had swallowed. The similar 
presence of marine plants in young fish has been described by Lebour. 

About half-way through the larval period our experimental trout were 
tested again for vitamin A and it was found that although the contents of 
the yolk sac were about half absorbed there yet remained considerable amounts 
of this dietary factor, At this stage one young fish per day was suflicient 


VITAMIN A IN NUTRITION OF FISH 633 


Gi OY *<~4weeks -> 


aoe 


Fig. 2. Fig. 3. 
Periods on — A diet only in dotted lines. 
Fig. 2. Resumption of growth of rats on feeding two trout eggs previous to basal diet each day. 
Fig. 3. Resumption of growth of rats on feeding one larval trout with yolk sac half absorbed. 


oy? <«-4 weeks ~ > 


. ‘ ‘ 
50) re m 4 caer re? ay 4. re r my i. + r re re Sy aS eee 4. + + i. 4 " Sa 


Periods on ~— A diet in dotted lines. 


Fig. 4. Very poor growth of rats on feeding one larval trout with yolk sac completely 
absorbed. 


Fig. 5. Very poor growth of rats on feeding one trout that had been fed on cod muscle from 
the time of the absorption of yolk sac. 


<-4 weeks- > 


Periods on ~—A diet only in dotted lines. 


Fig. 6. Resumption of growth of rats on feeding one trout that had been fed on liver and 
yolk of egg from the time of absorption of the yolk sac. Two short periods when the rats refused 
to eat the trout are seen in curves of rats 2566 and 2571. After a short time on —A only again 
the rats ate the trout and growth was rapid. 


634 K. H. COWARD AND J. C. DRUMMOND | 


to cause a good resumption of growth in the test rats (Fig. 3). After about 
four weeks the yolk sacs were practically absorbed in every case and another 
test was made. From the results it was apparent that the vitamin A content 
of the fish had fallen very greatly and that actual utilisation of that originally 
present in the yolk sac had occurred (Fig. 4). 


THE FEEDING OF PostT-LARVAL TROUT. 


Having found that the supplies of vitamin A present in the yolk sac of 
the newly hatched fish are apparently used up in the larval period, it was 
important to determine whether such fish are subsequently dependent on 
their food to maintain their supply of this factor. 

Accordingly the remaining fish were divided up into three groups of about 
250 each which were treated as follows: 

Group I. Diet rich in vitamin 4; fresh minced rats’ liver and ground up 
yolk of hard boiled egg. 

Group II. Diet deficient in vitamin A, freshly ground up cod muscle. 

Group III. No extra food. 

It was found necessary to feed Groups I and II several times a day by 
slowly dropping the finely minced food pulps into the stream of water in the 
“boats” so that the small particles were borne down to the fish on the current 
when the trout would immediately arrange themselves upstream and feed 
voraciously. 

The influence of the feeding on the trout themselves was marked. Those 
in Group I grew rapidly and were very active, plump and well. Except by 
an accident late in the experiment when a large number of deaths followed 
the stoppage of the current of water, there was hardly a death in this group. 

The fish on the vitamin A-deficient diet of cod muscle lived fairly well 
but did not grow much or show the vigour of those in Group I. Their weaker 
constitution was demonstrated by the increasing frequency of deaths as the 
experiment progressed. The group kept without food exhibited little or no 
growth but few died until about three weeks after the experiment had been 
in progress when deaths were suddenly very frequent. 


INFLUENCE OF Foop ON VITAMIN A RESERVES OF PostT-LARVAL TROUT. 


The feeding tests on rats demonstrated in a very striking manner the 
importance of the qualitative composition of the food supply for maintaining 
the vitamin A in the tissues of these fish. Very slight or no growth (Figs. 4 
and 5) was given by supplements of the fish of Group III or those fed on the 
white fish muscle known to be deficient in vitamin A [Drummond, 1918]. 

The fish from Group I, however, served as an excellent source of the 
vitamin for rats and led to a prompt resumption of growth (Tig. 6). 

Confirmation of this significance of the food supply was given by the 
survivors of Group II whose diet was later supplemented by an addition of 


VITAMIN A IN NUTRITION OF FISH 635 


egg-yolk pulp. They themselves immediately began to grow again and, when 
used as test material for the one rat surviving this test, showed that their 
tissues had once again become stored with the vitamin A. Similarly, those 
trout which had been previously starved were fed for ten days with egg; 
and the rats which had been used to test the starved trout were in the 
meantime fed only on the diet deficient in vitamin A. But when the trout 
now being fed were added to their diet, the rats immediately resumed growth. 

The demonstration of the importance of this qualitative aspect of the diet 
of young fish may be of some interest in connection with the well-known 
critical period which coincides with the absorption of the yolk sac. Trout 
hatchers have assured us that this period is one of great uncertainty in the 
rearing of young trout, and it is possible that our experiments may suggest to 
them some modification of the usual] existing methods of feeding which might 
lessen the loss which frequently occurs at that time. Some of our results, 
which are not fully recorded in this paper, tend to show us that mashed liver, 
the usual dietary of the fish in hatcheries, may not always supply sufficient 
of the vitamin A for their optimum development and well-being. The full 
absorption of the contents of the yolk sac is an equally critical stage in the 
development of marine fish. Fabre and Domergue [1900, 1905] found that 
certain young fish unable to obtain external food at this stage become anaemic 
and die, but that an addition of microscopical plant organisms to the water 
enables them to be reared successfully. The application of such observations 
to the practical question of sea fishery is well pointed out by Hjort [1914] 
who emphasises the importance of the immediate availability of suitable 
plankton food for the survival of larval marine fish. 


SuMMARY. 


1. The ova of the brown trout normally contain relatively large amounts 
of the vitamin A. 

2. During the subsequent development of the young larval fish this sub- 
stance is in some way utilised so that at or shortly after the stage at which 
the contents of the yolk sac are absorbed the supplies are almost exhausted, 
so far as our technique enables us to ascértain. 

3. It is probable that this fact is one of the reasons why this stage is so 
critical a period in the development of the young fish. 

4. If, in the post-larval period or even before the yolk sac is completely 
absorbed, the fish are given food rich in vitamin A their growth and develop- 
ment are satisfactory and they appear able to store that factor in their 
tissues. 

5. On a diet containing adequate protein but deficient in the factor A no 
such storage occurs and growth is subnormal. 

6. These experiments confirm the previous findings, referred to in the 
text, which show that the stores of vitamin A in the tissues of fish can be 
derived from the food. 


636 K. H. COWARD AND J. C. DRUMMOND 


We are indebted to Mr Peart, manager of the Chorley Wood Trout 
Hatcheries, for invaluable advice in the rearing of the trout, and also to the 


Medical Research Council for a financial grant which enabled the experiments 
to be made. 


REFERENCES. 


Coward and Drummond (1921). Biochem. J. 15, 530. 
Drummond (1918). J. Physiol. 52, 95. 
Drummond and Zilva (1922). Biochem. J. 16, 518. 
Fabre and Domergue (1900). Bull. de la marine marchande, Paris. 
—— (1905). Développement de la Sole, Paris. 
Hjort (1914). Rapports, Conseil International pour l exploration de la Mer, 20. 
(1922). Proc. Roy. Soc. B. 98, 440. 
Jameson, Drummond and Coward (1922). Biochem. J. 16, 482. 
Lebour (1918). J. Marine Biol. Ass. 11, 434. 
—— (1919, 1). J. Marine Biol. Ass. 12, 9. 
—— (1919, 2). J. Marine Biol. Ass. 12, 22. 
McCollum and Davis (1915). J. Biol. Chem. 20, 641. 
Zilva and Drummond (1922). Lancet, i. 1243. 


LXV. NOTE ON KNOOP’S TEST FOR 
HISTIDINE. 


By GEORGE HUNTER, Strang-Steel Scholar. 
Institute of Physiology, University of Glasgow. 


(Received July 17th, 1922.) 


In performing this test Knoop [1908] adds bromine water to the solution to 
be tested until there is just a very slight excess. On heating, the excess 
bromine disappears and in the presence of histidine a brownish red colour is 
developed; if the latter is in sufficient concentration a dark precipitate 
settles out.. The solution to be tested should be slightly acid. The only 
substance other than histidine found to give this test is iminazolylethylamine 
or histamine. The test, according to Knoop, is given with histidine at a 
dilution of 1 : 1000. 

When using this test the writer found that different intensities of colour 
were produced from solutions of the same concentration of histidine. This 
peculiar behaviour was accounted for by slight variations in the bromine 
excess. The colour produced is very soluble in excess of bromine and in 
dilute solutions of histidine an initial excess may prevent all colour develop- 
ment. The difficulty of adding just sufficient bromine especially to slightly 
coloured solutions has been overcome by adding a definite excess followed by 
washing the solution repeatedly with chloroform in a small separating funnel 
until the chloroform is no longer coloured. By this procedure even strongly 
coloured meat extracts and other fluids may be successfully tested by Knoop’s 
method. The excess of bromine appears to oxidise the colouring matters 
and after washing with chloroform such solutions are almost colourless, so 
that on heating the brown colour due to histidine is readily detected. Most 
uniform results have been obtained when the washed solution is transferred 
to a test tube and set in a boiling water-bath. 

That the chloroform does not extract any of the histidine compound was 
shown by adding to a series of test tubes equal amounts of 0-1 % histidine 
monohydrochloride and adding increasing small amounts of bromine water 
from a pipette until definite excess of bromine was present. By heating all 
the test tubes together on the water-bath it was found that there was first a 
rise, then a fall in colour values corresponding with insufficiency or excess of 
bromine. The maximum thus obtained was less than that obtained by the 
chloroform method. 

Bioch. xv 42 


638 G. HUNTER 


The maximum colour appears to be developed when the histidine has 
absorbed just three atoms of bromine per molecule. A dilute solution of bromine 
water was standardised against NV/100 sodium thiosulphate using a mixture 
of potassium iodide and starch as indicator on a tile. The bromine solution 
was 0-00785 N. 

In one burette was placed 0-1 °% solution of histidine monohydrochloride 
and in another the standard bromine water just before using it. At ordinary 
temperature it was found that 10 cc. of the histidine solution absorbed 18 cc. 
of the standard bromine water. When the solution was kept cold in ice the 
same quantity of histidine absorbed-16 cc. of the bromine solution, whilst at 
about 40° in a water-bath it absorbed 21-8 cc. of the bromine solution, (These 
amounts are not quite definite as on Jong standing more bromine is ab- 
sorbed.) 

The histidine-bromine solution was in each case transferred to a 100 ce. 
measuring flask which was filled up to the mark with water, so that the solu- 
tion represented 0-01 % original histidine. About 10 cc. of each were put 
into three dry test tubes and heated on the water-bath. The colour developed 
was maximum when the bromine was absorbed at ordinary temperature. 

It follows from this that one molecule of histidine monohydrochloride 
absorbs approximately three atoms of bromine at ordinary temperature. 

To narrow the limits still further five lots of 5 ec. of 0-1 % histidine solution 
were taken and to these respectively were added 8, 8-5, 9, 9:5 and 10 cc. of 
standard bromine solution. On diluting to 50 cc. and heating portions, the 
colour was greatest where 9 cc. had been used. This is the same as the above. 
The maximum colour developed here was not greater than that got from 
solutions of the same concentration washed with chloroform. 

By the modification above suggested Knoop’s test can be obtained 
with certainty in solutions of histidine at a dilution of 1 : 10,000; a faint 
colour in pure solutions is still observable at 1 : 20,000. 

The colour developed on heating is very variable according to the reaction 
of the liquid. In markedly acid solution it is a dull yellow brown. In approxi- 
mately neutral solution, the shade is dark brown. In solutions made faintly 
alkaline with sodium hydroxide—after bromine has been added and the 
excess removed with chloroform—a bright pink colour develops which is less 
stable than that in neutral or acid solutions. If made very faintly alkaline 
with ammonia instead of caustic soda, a deep purple colour rapidly develops 
on heating. The presence of ammonia in just the requisite amount gives a 
more intense colour than in the other.cases, but since it is difficult to regulate 
this amount, and further, since there is reason to believe that the specific 
value of the test may suffer—for carnosine gives a faint yellow by this 
treatment—the use of alkali cannot be recommended. 

The colour developed from tryptophan by addition of bromine is readily 
distinguished from the colour due to histidine. he former develops in the 
cold and is easily extractable with amyl alcohol. The coloured substance in 


NOTE ON KNOOP’S TEST FOR HISTIDINE 639 


Knoop’s test is not extracted by any of the ordinary solvents and is peculiarly 
unaffected by reagents other than bromine. 


SUMMARY. 


By a modification of Knoop’s test for histidine its delicacy is increased 
at least ten times and the test can be performed with greater certainty 
especially in coloured fluids. The proportion of three atoms of bromine per 
molecule of histidine is found to give the maximum colour on heating. 


REFERENCE. 
Knoop (1908). Beitrdge, 11, 356. 


42—-2 


LXVI. THE ESTIMATION OF CARNOSINE IN 
MUSCLE EXTRACT—A CRITICAL STUDY. 


By GEORGE HUNTER, Strang-Steel Scholar. 
Institute of Physiology, University of Glasgow. 


(Received July 17th, 1922.) 


In a recent communication [Hunter, 1921] a colorimetric method for the 
estimation of carnosine in muscle extract was generally outlined. This was 
based on a method of estimation of iminazoles devised by Koessler and Hanke 
[1919, 1]. This method depended on the coupling of carnosine with diazotised 
sulphanilic acid in alkaline solution to form an azo dye. The colour produced 
was measured against a standard solution of a mixture of Congo red and 
methyl-orange. Pure solutions of carnosine were found to react very satis- 
factorily, and in different extracts of the same muscles the colour values were 
consistent. The accuracy of the method could not be doubted, but the certainty 
that carnosine was responsible for the total colour production was called in 
question. 

The problem—of the extent to which carnosine is responsible for the 
production of the azo colour in muscle extracts—may be attacked in two 
ways: 1. By the elimination of all interfering substances. 2. By a direct 
method of confirming the carnosine content. 


I. Tue Evimtnation or INTERFERING SUBSTANCES. 


Potentially interfering substances. It is well known that aromatic amines © 
and phenols readily couple with a diazotised aromatic amine and are thus 
potentially interfering substances with the reagent here employed for the 
estimation of carnosine. 

The iminazoles form another well defined group—of which carnosine is 
only one member—giving the diazo reaction. 

The purines which contain the iminazole ring are generally described as 
substances giving a positive diazo test—this applies at least to the members 
adenine, hypoxanthine, and xanthine [v. Plimmer, 1918]. 

There are various other substances diversely referred to throughout bio- 
chemical literature as giving the diazo test. Of these may be mentioned 
thymine [Thierfelder, 1908], bilirubin [Neubauer-Huppert, 1913], urochro- 


THE ESTIMATION OF CARNOSINE IN MUSCLE EXTRACT 641 


mogen and urobilinogen (Neubauer-Huppert), besides the ‘ ‘neutral- sulphur” 
compounds of urine (Neubauer-Huppert). 

An exhaustive investigation into the behaviour of all substances—confined 
even to biochemical literature—towards the diazo reagent would probably 
be endless as well as fruitless. Attention will thus be confined only to those 
reagents ordinarily employed in the preparation of muscle aqueous extracts 
and to substances that may be present with a fair degree of probability in 
the extracts. 

A. Reagents. Soluble chlorides, sulphates, nitrates, phosphates and 
sodium acetate, have been tested in relatively high concentrations and found 
to have no effect on the diazo reagent. 

Ammonia and its salts not only interfere with the production of colour 
from other substances such as histidine [Koessler and Hanke, 1919, 2] but 
when added alone to the diazo reagent give a yellow colour. 

Soluble sulphides give a colour in low concentration. The presence of 
sulphur in cystine probably accounts for the similar behaviour of that sub- 
stance towards the diazo reagent. 

Ethyl alcohol was found to give no colour with the reagent though 
Koessler and Hanke find that it inhibits colour production. 

Hydrogen peroxide, formaldehyde and acetone all give yellow colours. 

Tannic acid in very low concentration gives a marked colour. This is 
accounted for by its phenolic constituents. 

All solutions in which carnosine is to be estimated by means of the diazo 
reagent should thus be free from ammonium salts, sulphides, and tannic acid. 
Formaldehyde or phenols, such as thymol, should not be used as preservatives 
in this connection. Muscle tissue may be preserved in alcohol and extracts 
by a layer of toluene. 

B. Muscle constituents. Certain normal muscle constituents which may 
possibly act as interfering substances in the extracts require a more detailed 
examination, as these are much less within the control of the worker than 
substances which may be used in the preparation of the extracts. 

The presence of bilirubin, urochromogen and urobilinogen in muscle 
extracts is too unlikely to claim further consideration for those substances. 
Thymine may also be dismissed on account of the difficulty with which it is 
liberated by hydrolysis from nucleic acid [v. Jones, 1920]. 

(1) Phenols and aromatic amines. The presence of these substances in 
protein-free extracts has not been shown by any of the tests used to detect 
them. Extracts from several types of muscle were shaken in acid solution 
with ether. The ether extracts were evaporated to dryness and the slight 
residues taken up in small quantities of water. These gave no diazo reaction 
nor has Millon’s test been found positive either in the extracts themselves or 
in the ether fractions. 

It is known that aromatic amines couple with the diazo reagent in acid 
solutions, whereas iminazoles and phenols require a weak alkaline medium before 


642 G. HUNTER 


a colour is produced. Thus if a little 1, 2. 4-diaminotoluidine is added to the 
reagent without the addition of sodium carbonate, a strong orange colour is 
developed. If an iminazole or phenol is added under the same conditions no 
colour is produced?. 

Various extracts have been tested with the acid reagent but all with 
negative results. It may thus be concluded that there are no aromatic amines 
contributing to the azo colour developed in meat extracts?. 

(2) Tyrosine. This substance, which is not extractable by ether, has 
received considerable attention from the point of view of the diazo reaction. 
Totani [1915] devised a method to distinguish the azo colour developed by 
histidine from that developed by tyrosine. The colours were first reduced 
with zine and hydrochloric acid, twice the volume of 25 % ammonia added 
and in both cases a golden yellow colour was obtained. In the case of 
histidine the addition of hydrogen peroxide changed the colour to lemon 
yellow whereas the colour was destroyed in the case of tyrosine. A dilution 
greater than 1 : 20,000 histidine was necessary to get the characteristic final 
colour. 

Among the various substances considered by Totani no mention is made 
of iminazoles other than histidine, nor of purines, nor of aromatic amines 
nor of phenols other than tyrosine. 

On repetition of Totani’s procedure with histidine alongside carnosine, 
the two substances went through approximately the same colour changes. 
The final colour was not destroyed and so the presence of tyrosine could not 
be detected by this method. 

Tyrosine gives a positive Millon’s test; histidine and carnosine are 
negative to Millon’s. Tyrosine gives a faint Millon’s test at a dilution of 
1 : 25,000. 

A still more delicate test for tyrosine—on which is also founded a method 
of estimation—has recently been devised by Hanke and Koessler [1922]. 
This depends on a further modification of the diazo reaction and is approxi- 
mately as delicate as that reaction is for carnosine. For quantitative purposes 
the procedure is the same as that adopted for iminazoles and phenols. The 
test cylinder is allowed to stand for exactly 5} minutes after the tyrosine 
solution has been added to the alkaline reagent. This gives rise to a primary 
yellow colour the intensity of which is not directly proportional to the amount 
of tyrosine used. 

2 cc, of 3N sodium hydroxide solution are now added and the contents 
of the cylinder mixed. This gives rise to a colour intensification with a change 
of tint towards the red. 

One minute after the addition of the sodium hydroxide 0:10 ce. of a 20 % 
solution of hydroxylamine hydrochloride is added and rapidly mixed, After 


* The aromatic amines also give a colour in alkaline solutions. 
* This test would also eliminate bilirubin which gives a blue colour in acid solution [Ehrlich, 
quoted by Neubauer- Huppert). 


THE ESTIMATION OF CARNOSINE-IN MUSCLE EXTRACT 643 


a latent period of 5 to 10 seconds an intense bluish red colour rapidly develops. 
This colour is stable and is directly proportional to the amount of tyrosine 
present. 

Hanke and Koessler note that this intensification is also given by sub- 
stances capable of a keto-enol tautomerism such as acetaldehyde, acetone 
and aceto-acetic acid. 

It was decided to test this new modification on muscle extracts. It was 
first observed that the addition of sodium hydroxide and hydroxylamine 
hydrochloride in the manner above described to the azo colour developed by 
histidine or carnosine had the effect merely of a proportional dilution. Thus 
a solution of carnosine gave by the ordinary method a reading of 22-5 mm.} 
by the new method a reading of 18-3 mm., 7.e. approximately the same as 
would be obtained by adding 2-1 ce. wate to the cylinder. The convene 
results were obtained on the extracts. 


Reading - Reading with diazo Calculated reading— 
with diazo reagent, NaOH and assuming proportional 


reagent NH,OH . HCl dilution Intensification 
mm. mm. mm. % 
Ox muscle 27-2 21-4 21-6 None 
Cat ,, 18-0 16-5 14:3 15-4 
Rabbit muscle 10-8 11:3 8-5 33-0 
Otter Pe 15-5 18-5 12-3 50-4 
Salmon , 13:7 21-5 10-8 100-0 


The percentage of intensification of these extracts is in the order of the 
yellowness of their azo colours. Thus ox muscle extracts give a colour which 
matches the carnosine colour standard. The colour developed from cat 
muscle extract is generally slightly more yellow; that from rabbit almost 
matches the histidine colour standard; that from otter is still more yellow 
and that from salmon is so yellow as to be almost unmatchable. 

The intensification colours were pink. They did not show the purplish 
tinge given by tyrosine. No Millon’s reaction was obtained in concentrated 
extracts of salmon muscle and it is concluded that tyrosine is not responsible 
for the intensification obtained in that tissue. 

The specific cause of this intensification has not been determined. 

(3) Purines. These are normally present in muscle extracts so that a 
consideration of the individual members is rendered necessary. Adenine and 
guanine were prepared from commercial plant nucleic acid. This was hydro- 
lysed by suspending in methyl alcohol and passing dry hydrochloric acid gas 
according to the method employed by Levene [1921] for animal nucleic acid. 
This process was found to work very satisfactorily. The precipitated chlorides 
of adenine and guanine were filtered off and separated according to methods 
described by Jones [1920]. 

Hypoxanthine nitrate and xanthine were prepared from parts of the 
adenine and guanine respectively by deaminisation and subsequent purifica- 
tion. 


644 G. HUNTER 


Cytosine and uracil were also prepared from the residues freed from methyl] 
alcohol by further hydrolysis in the autoclave at 160° with 25 % sulphuric 
acid for five hours according to Jones [1920]. 

Neither of these pyrimidines gave a colour with the diazo reagent. 

0-05 % solutions were made of adenine sulphate, of guanine chloride, of 
hypoxanthine nitrate and of xanthine—all in 1-1 % sodium carbonate 
solutions. 

1 ce. of each of these was tested in the ordinary way. Guanine and xanthine 
gave marked colours. The adenine solution showed only a slight reaction and 
there was no colour in the case of hypoxanthine. With a tenth of the above 
amounts guanine and xanthine were still strongly positive whilst adenine and 
hypoxanthine were entirely negative. The intensities of the colour were approxi- 
mately the same in like concentrations of guanine and xanthine. With 
0-025 mg. of guanine hydrochloride in the cylinder a reading of 8-5 mm. was 
obtained with the histidine colour standard. With 0-05 mg. the reading was 
13 mm. Xanthine in the same amounts gave the respective readings 8 mm. 
and 12-5mm.! The colour production is not directly proportional to the 
amounts of guanine and xanthine in the cylinder; nor do guanine and 
xanthine, mixed with a known amount of carnosine, give proportional colours. 
Thus 1 cc. of a mixture of 1 cc. 18-6 mm. per cc. carnosine solution with 1 ce. 
13-5 mm. per cc. guanine solution gave a reading of 13-2 mm. with the test 
cylinder set at 20mm. This is 2-8 mm. short of the calculated reading of 
16 mm. A similar result was found with xanthine. Adenine and hypoxanthine 
do not give a colour or inhibit colour production at this concentration, but 
they tend to make the colour due to carnosine too yellow. Thus with 1 ce. 
of a mixture of carnosine solution of the above concentration with | cc. 
0-05 mg. per cc. of adenine sulphate solution, the calculated reading of 9-3 mm. 
was obtained. Hypoxanthine behaves similarly. 

Guanine and xanthine are about half as sensitive towards the diazo reagent 
as carnosine. If present in appreciable amounts in muscle extract they must 
seriously affect the estimation of carnosine. 

Uric acid in excess gives a slight yellowness to the diazo reagent. 

To what extent are purines likely to interfere with the estimation of 
carnosine in muscle extracts? On the assumption that there is present in 
extract of meat free purine nitrogen to the extent of 0-045 % as quoted by 
Lusk [1921] it is unlikely that more than 0-020 % nitrogen represents purines 
that give a colour with the diazo reagent. If the nitrogen be taken as repre- 
senting 42 °%, of the purine molecule—an average figure from adenine, guanine, 
hypoxanthine, xanthine and uric acid—the purines affecting the diazo reagent 
amount to about 0-05 %. In muscles with a carnosine content of less than, 


say 0-1 % the presence of such a proportion of purines would make the 
results worthless, 


' The readings are only approximate as the colours are much more yellow than the 
standard. 


THE ESTIMATION OF CARNOSINE IN MUSCLE EXTRACT 645 


To determine the actual purine interference under the conditions of 
extraction previously recommended, it was then considered necessary to 
carry out some quantitative fractionations!. 

A quantity of ox muscle was extracted at 70° with small amounts of water 
until the filtrate no longer gave a positive diazo test. The proteins were 
precipitated with excess of lead acetate and the excess of lead was removed 
with disodium phosphate. The filtrate was neutralised with sodium hydroxide 
and its colour value measured with the carnosine colour standard. This was 
found to be 1050 mm. per cc. 

50 ec. of the extract were taken in a small beaker, made just acid with 
nitric acid, and silver nitrate was added till a drop of the solution on a tile 
gave a brown colour with baryta. The test drops were washed back into the 
beaker. . 

After settling, the contents of the beaker were filtered by suction through 
washed asbestos in a small Hirsch funnel. The precipitate was carefully 
washed with water and after sucking dry was transferred with the asbestos 
to a small beaker, stirred up with a little water and hydrogen sulphide passed 
to precipitate the silver. The contents were again filtered through asbestos 
in the same way and the hydrogen sulphide removed by prolonged aeration 
with the help of a pump. The filtrate was neutralised. This is the purine 
fraction. 

To the filtrate from the purine fraction excess of solid finely ground baryta 
was added. The precipitate was allowed to settle, filtered as above, and 
washed with saturated baryta water. The silver was removed with hydrogen 
sulphide and the barium with slight excess of sulphuric acid. The slightly 
acid solution was aerated to remove the hydrogen sulphide and the solu- 
tion was then neutralised with sodium hydroxide. This is the carnosine 
fraction. 

The filtrate from the carnosine fraction was freed from barium and hydrogen 
sulphide and neutralised as above. This is the final filtrate. 

The exact volume of each of the three fractions was noted. The colour 
value per cc. of each of the fractions was then measured and their total colour - 
values calculated. The sum total colour value was then obtained and com- 
pared with the original, 7.e. 50 x 1050 or 52,500 mm. 


1 All the carnosine can be extracted from muscle with much smaller proportions of water 
than recommended by the writer in his previous paper on this subject. A filtrate of about 
50 cc. per g. of muscle used was there recommended mainly because the proteins are easily 
precipitated by slight acidification with acetic acid and heat under those conditions. In con- 
centrated extracts this simple process is not effective and the use of lead acetate or other pre- 
cipitant is necessitated. The number of washings, however, should not be reduced. There is no 
need to reduce the proportion of water to muscle when the extract is to be tested only for colour 
value unless the carnosine content is less than 0-1%. With that amount in muscle a reading 
of 12mm. will still be obtained with the 50 cc. per 1g. proportion. But for precipitation 


purposes and other tests it is desirable to keep the filtrate small rather than have to evaporate 
it to a low volume. 


646 3 G. HUNTER 


The distribution of colour value is shown by the total colour values in 
mm. of the three fractions: 


Purine fraction ane 1,560 
Carnosine fraction me 47,587 
Final filtrate... we 2,200 
Sum total io as 51,347 


The unrecovered colour value is thus 1153 mm. or approximately 2 %. 
The purine fraction accounts for about 3 % and the final filtrate for about 
4-5 %. The carnosine fraction represents almost 91 % of the original colour 
value. 

The relatively small error due to purines by the gentle method of extrac- 
tion recommended is further seen in the table to follow. 

Three portions were taken from the same muscle and two extracted as 
above. The third was extracted at a temperature just under boiling point. 

Aliquot portions of the filtrates from the lead precipitates were taken for 
total nitrogen estimation and for further fractionation into purine and 
carnosine fractions. The final filtrates were rejected without evaluation. The 
total nitrogen was also estimated in the carnosine fractions. In the table 
carnosine, nitrogen and purine (calculated as carnosine) are shown as per- 
centages of the original muscle. The colour values of the purine fraction are 
at best only a rough estimate on account of the difficulty in comparing with 
a much redder standard, The error due to the purines in ox muscle is however 


small. 
Difference in 


Carnosine Total N Carnosine Total N Colour % carnosine ~ 
from lead _— from lead from from value in in lead acetate 
Meat acetate acetate carnosine carnosine purine filtrate and in 
No. filtrate fraction fraction fraction fraction carnosine fraction 
1 0-517 0-530 0-498 0-246 0-004 0-019 
2 0-512 0-546 0-495 0-230 0-012 0-017 
3 0-533 0653 0-487 0-258 0-014 0-046 


Before considering other substances likely to interfere with the estimation 
it may be noted that the more drastic extraction process of muscle No, 3 
results in a higher percentage colour value. The total nitrogen in this extract 
is also raised. In the carnosine fraction the percentage falls again into line 
with the other two samples. The loss is not accounted for by a higher purine 
value. Though the colour value of the lead acetate filtrate is raised by the 
more rigorous extraction, the actual percentage of carnosine is lowered. The 
lowering of colour value becomes obvious even in the lead acetate filtrate if 
boiling is prolonged. Thus two samples of muscle treated in the usual way 
showed an average of 0-639 % carnosine in the lead acetate filtrate, whilst 
three samples of the same muscle after boiling for about 30 minutes showed 
an average value of only 0-614 % or a fall of 0-025 % reckoned as carnosine. 

In muscles | and 2 approximately 96 % of the colour is recovered in the 
carnosine fraction, Calculated as carnosine this accounts for about 50 % of 


THE ESTIMATION OF CARNOSINE IN MUSCLE EXTRACT 647 


the nitrogen in the same fraction. There is no reason to doubt that carnosine 
is responsible for all the colour production in the carnosine fraction from ox 
muscle. In the case of rabbit muscle and notably salmon muscle, fraction- 
ation does not appreciably improve the match. Thus two portions of a rabbit 
muscle showed from the lead acetate filtrate colour values of 0-092 °%% and 
0-099 % reckoned as carnosine. The carnosine fraction from the same muscles 
showed each a value of 0-072 % as carnosine. In the latter case the azo colour 
developed was still very yellow. The colour from salmon muscle extract is 
slightly more red in the carnosine fraction than in the lead acetate filtrate, 
yet it is far from satisfactory. 

The precipitation process occupies much time and to obtain reliable 
quantitative results from it, great care must be exercised. With the additional 
possibility that the carnosine fraction may still contain other colour-giving 
substances, it would appear of little use to attempt the elimination of inter- 
fering substances by the silver-baryta precipitation except in very special cases. 

(4) Iminazoles. Of this group only histidine and carnosine are known 
with certainty to occur in animal tissue. Histamine has been found in the 
intestinal mucosa by Barger and Dale where its presence is attributed to 
bacterial action [Barger, 1914]. Its isolation from the pituitary gland by 
Abel is accounted for by Hanke and Koessler through Abel’s use of a com- 
mercial product. Hanke and Koessler [1920, 2] show that it is absent from 
fresh hypophysis cerebri. 

In protein-free extract from fresh muscle there is no evidence that histidine 
is present provided the extraction process has been gentle. The ease with 
which carnosine may be hydrolysed (v. later), along with other factors to be 
considered, renders it important that there should be some ready means of 
detecting histidine in the presence of carnosine. 

As far as the writer is aware there is no known method of separating these 
substances. They both precipitate in the carnosine fraction in the “arginine 
separation.” Though histidine is more readily precipitated with ammonium- 
silver than carnosine, yet the latter also precipitates in such relatively dilute 
solutions as 0-5 % carnosine. Nor could any test be found in the literature 
to distinguish the two substances. 

Various colour tests for histidine were performed on solutions of carnosine. 
The main differences were found in their behaviour towards the biuret and 
Knoop tests, 

The biuret reaction was found to be entirely negative even in about 30 % 
solution of carnosine and after standing at least one hour. This difference in 
behaviour is however of little value for discrimination purposes as relatively 
concentrated solutions of histidine must be used to get a positive biuret. 
Thus 1 % solution of histidine monohydrochloride gives the reaction in the 
cold only after standing for about 14 hours. 

The writer has found Knoop’s test to be the most useful way of detecting 
histidine in the presence of carnosine. Carnosine is quite negative to Knoop’s 


648 G. HUNTER 


test. By a modification of the test as explained in the preceding paper, the 
writer has been able to increase both its delicacy and certainty as a test for 
histidine. The presence of histidime may be detected with certainty at a 
dilution of 1 : 10,000. ; 

Though this reaction has not yet the delicacy that might be desired, its 
very specific nature makes it valuable for work of this kind. Knoop [1908] 
states that the reaction is also positive for histamine. No other substance, 
as far as the writer is aware, gives the reaction. 

Certain muscle extracts suspected of histidine, were then tested by the 
modified Knoop’s test. Extracts of-rabbit and salmon which give notably 
yellow colours were tested in sufficient concentration but without giving an 
indubitable Knoop’s test. It is thus concluded that histidine is not a factor 
in causing the yellow colour of those extracts. 

In various cases of purchased butcher’s meat a positive Knoop’s test has 
been obtained. If the meat is allowed to become just noticeably putrid a 
very marked reaction is given. Histamine may in part be responsible for the 
positive test in such cases. 

The test may perhaps be used to most advantage as an indirect test for 
carnosine where the presence of that substance is doubtful in any tissue or 
fluid. If a protein-free extract is at first negative to Knoop’s test, and after 
hydrolysis (v. later) is positive to the test, it appears necessarily to follow 
that the unhydrolysed liquid contains carnosine. 

The test can also be applied to elucidate some other problems that had 
arisen in the course of this work. The effect of heat on carnosine solutions 
will first be considered. 


The effect of heat on solutions of carnosine. 


In the writer’s preliminary communication on the estimation of carnosine, 
a slight fall was noted in the colour values of solutions of carnosine heated 
on the water-bath for one hour. Similar experiments have since been repeated 
but the period of heating has been extended to four hours. Thus a number 
of solutions of carnosine nitrate with an original colour value of 22 mm. per 
cc. showed after four hours on a boiling water-bath an average value of 19-4mm. 
per cc, or a fall of 2-6 mm. per ce. 

10 ce. of a 0-5 % solution of carnosine as nitrate was brought to dryness 
three times in an open basin on a steam-bath. The total period of heating 
was about four hours, It was finally taken up in about 2 cc. of water and 
Knoop’s test applied. A decided brown colour was developed indicating 
that in the process the carnosine had been partly hydrolysed, giving rise to 
histidine. This observation confirms the fall in colour value. 

Though four hours is a relatively long period of heating, the fact that carno- 
sine is thus destroyed in pure neutral solution indicates that care should be 
exercised in the process of extraction from muscle, . 

The fall in colour value as estimated in carnosine from 0-639 % to 0-614 % 


THE ESTIMATION OF CARNOSINE IN MUSCLE EXTRACT 649 


on prolonged extraction of the muscle, previously noted (p. 646), may be 
accounted for by the destruction of carnosine. The occurrence of histidine 
in muscle extracts will thus arise more likely from the hydrolysis of carnosine 
than from the hydrolysis of histidine-containing proteins. For if the latter 
process occurred to a marked degree the colour values would tend to rise 
rather than to fall. . 

It will thus be observed that any process of evaporation at atmospheric 
pressure must result in at least a partial destruction of the substance. 

Solutions of carnosine whether pure or in muscle extracts may be evapor- 
ated in vacuo to dryness with very little destruction. Thus 10 ce. of a neutral 
dilute meat extract were evaporated to dryness at 20 mm. pressure of mercury 
from a 200 cc. distilling flask on a water-bath about 60°. Exactly 10 cc. of 
water were then pipetted into the flask and rinsed round to dissolve the dried 
residue. The colour value of the original solution was 25-6 mm. per cc. whilst 
that of the dried substance taken up in 10 cc. of water was 25 mm. per ce. 
(The volume of solid in the original 10 cc. was regarded as negligible.) Repeti- 
tion of this experiment still showed a loss in colour value of about 2 %. 

This experiment at the same time goes to show that a very small propor- 
tion, if any, of the colour-producing substances in muscle extract is volatile. 
It serves to confirm the results obtained by ether extraction and Millon’s test. 

(5) Other substances. The effects of cystine, leucine and arginine on the 
diazo reagent have been measured by Hanke and Koessler [1920, 1]. These 
workers find that there is no interference so long as the ratio of cystine to 
histidine does not exceed 6 : 1, which is higher, they note, than has yet been 
found in any protein. Cystine may thus be dismissed as a probable interfering 
factor in muscle extracts. 

The effects of arginine and leucine are less marked than those of cystine. 

Creatine in large amounts gives a slight yellowness to the reagent. 

Creatinine, urea and lactic acid have been found to have no effect on the 
reagent. : 

Such a process of elimination as has here been attempted must yield 
mainly negative conclusions. None of the substances most likely to interfere 
with the colour values has any very marked effect. In the case of ox muscle 
it has been shown that most probably less than 5 % of the colour value is 
not due to carnosine. In cat and frog extracts, judging merely from the colour 
development, the non-carnosine colour value is likely to lie within the same 
limit, but of salmon little can be said except that it contains carnosine. Any 
elimination method must necessarily be unsatisfactory unless the substance 
can be obtained pure; and many factors militate against that possibility in 
the case of carnosine. Among those may be noted the high solubility of the 
substance in water, its facility for adsorption, the inadequacy of the methods 
for its fractional precipitation and the ease with which it is hydrolysed. It 
is upon this last property that the writer has sought a more direct method 
for confirming the values. 


650 G. HUNTER 


II. Hyprotysis EXPERIMENTS. 

(1) The stability of histidine. Histidine has been observed by Koessler 
and Hanke [1919, 2] to be remarkably stable towards hot concentrated 
hydrochloric acid. They observed that when 2 cc. of a 1% solution of 
histidine dihydrochloride were heated on a boiling water-bath for 10 hours 
with 25 ec. concentrated hydrochloric acid the histidine was quite unchanged, 
as shown by the colour values recovered. Over a shorter interval the writer 
has confirmed those findings as shown by the following experiment. 

Into each of five numbered test tubes there was introduced 1 cc. of a 
solution of histidine along with 3 cc. concentrated hydrochloric acid. These 
were then set on a boiling water-bath and one removed every 15 minutes. 
As each test tube was removed from the water-bath it was cooled, about 
1 sq. mm. of litmus paper introduced, and the contents neutralised with 
sodium hydroxide. The contents were again cooled and poured into a 25 cc. 
flask. The test tube was repeatedly washed into the measuring flask, then 
water was added up to the mark. 1 cc. of this was taken for measuring the 
colour value against the histidine colour standard. The original value of 
23 mm. was obtained by directly diluting 1 cc. of the original to 25 cc. and 
taking 1 cc. for the test. The colour values after periods of boiling lasting 
15, 30, 45, 60 and 75 minutes were found respectively to be 23-2, 23-1, 23, 
22-9 and 23 mm. High concentrations of sodium chloride affect neither the 
tint nor the intensity of the colour developed. 

(2) Hydrolysis of carnosine—theoretical considerations. It thus seemed 
very probable that if carnosine were hydrolysed with hydrochloric acid, the 
colour value would continue to fall until the reaction was complete. If the 
diazo value of the solution were measured before hydrolysis and again after 
hydrolysis, the ratio of the colour values thus obtained should be that of 
their molecular colour values. The molecular colour value of histidine was 
previously found to be 117 million mm. as compared with 114 million mm, 
determined by Koessler and Hanke. A revised determination of the molecular 
colour value of carnosine has led the writer—mainly from increased skill in 
matching the histidine and carnosine colour standards—to the conclusion 
that the molecular colour value of carnosine is somewhat greater than the 152 
million mm. previously published, viz. 161 million mm. This value is checked 
by the following considerations. 

The histidine colour standard is a mixture of 0-22 ce. methyl-orange with 
0-20 ec. Congo red in 100 cc. water. For convenience call this S;,. 

The carnosine colour standard is a mixture of 0-10 cc. methyl-orange with 
(25 ec, Congo red in 100 ce, water. Call this Sg. 

These amounts were taken from 0-1 % stock methyl-orange and from 
0-5 %, stock Congo red [v. Hunter, 1921}. 

On comparing in the colorimeter it is found that 24 mm. Sg = 28-5 mm. Sy. 

From the table previously published | Hunter, 1921] 0-040 mg. of carnosine 
in the test cylinder set at 20 mm. showed a reading of 24 mm. Sg. 


THE ESTIMATION OF CARNOSINE IN MUSCLE EXTRACT 651 


0-04 mg. of carnosine should thus give a reading of 28-5 mm. Sq. }..., 
The molecular colour value of carnosine is thus 


ae x 226 x 105, 


or approximately 161 million mm. Sy. 

Now, carnosine yields by theory on hydrolysis 68-5 % of histidine, 7.e. 
0-04 mg. of carnosine yields on hydrolysis 0-0274 mg. of histidine. 

0-04 mg. of histidine monohydrochloride gives a reading of 22:4 mm. Sy. 


[N,B. The molecular colour value of histidine is thus 
so x 209 x 105, 


i.e. approximately 117 million mm. Sq. } 


Histidine monohydrochloride contains 74 % histidine, 7.e. 0-04 mg. of 
histidine monohydrochloride contains 0-0296 mg. of histidine. 


0-01 mg. of histidine will thus give by calculation a reading of ae or 
7-57 mm. Sy, and 3 

0-0274 mg. of histidine will give 20-7 mm. Sy. 

That is, a solution of carnosine giving an original reading of 285 mm. 
Sy will give after hydrolysis a reading of 20-7 mm. Sy. 

The ratio of the final reading to the original reading should be the same 
as the ratio of the molecular colour value of histidine to the molecular colour 
value of carnosine. 


te 20-7 117 


28:5 161 

For convenience this figure will be termed the hydrolysis quotient. 

(3) Haperimental. The results obtained from the actual hydrolysis of 
carnosine solutions are less satisfactory than one might expect from the above 
considerations. Thus a series of carnosine solutions were treated in exactly 
the same way as that described for histidine. 0-1 °% solution of carnosine 
was used. The original value of 1 cc. from 25 cc, dilution was 28-5 mm. Sy. 
After hydrolysis for periods of 15, 30, 45, 60 and 75 minutes 1 cc. showed the 
respective colour values 20-4, 19, 18-2, 18-1 and 18 mm. Sy. After 15 minutes’ 
hydrolysis the colour was more yellow than Sq. 

Carnosine is thus very readily hydrolysed and under the above conditions 
part of the histidine is destroyed. After the initial sudden fall from 28-5 mm. 
to 20-4 mm. continued boiling lowers the value at a much diminished rate. 
After 30 minutes’ boiling the values remain almost constant but the colours 
are too yellow—indicating some destruction. It would appear that in the 
course of the hydrolysis the histidine passes through an unstable phase in 
which the iminazole ring is readily disrupted. The histidine that emerges 
from that hypothetical intermediate condition resists further boiling. 

Other hydrolytic agents, such as sodium hydroxide and acetic acid, besides 


= 0-73. 


652 G. HUNTER 


different conditions of temperature and concentration were tested, in the 
attempt to overcome the difficulty. Finally an approximately constant- 
boiling mixture of hydrochloric acid was used and the hydrolysis conducted 
at 90°. Under these conditions a strictly quantitative conversion is not yet 
attained, but a comparison of the results from various extracts under standard. 
conditions would appear to be of some weight. The extracts hydrolysed were 
of such a concentration that 1 cc. of the final dilution gave approximately 
the same values. To the amount of extract to be hydrolysed exactly half its 
volume of concentrated hydrochloric acid was added. The following are the 
results from a solution of pure carnosine and from muscle extracts of cat, of 
ox, and of salmon. This ox extract was slightly positive to Knoop’s test. 
All the readings were taken with the histidine standard. 


Cat muscle Ox muscle —- Salmon muscle 

Time Carnosine extract extract extract 
mins. mm. mm. mm. mm. 
Original 28:5 29-0 28-5 27-0 
15 25-4 24-3 26-0 26:5 
30 23-3 23-0 23-5 24-0 
45 21-5 21-8 22-7 22-5 
60 21-0 21-2 22-0 22-1 
75 20-9 20-7 21-7 21-4 

Hydrolysis quotients 

after 75 mins. 0-73 0-71 0-76 0-79 


The carnosine solution and the cat muscle extract agree very well in both 
their rate of fall and in their hydrolysis quotients. The ox muscle quotient 
is slightly high, probably owing to the presence of either histidine or histamine 
in the original. The salmon extract quotient is still higher, but it is remark- 
able in face of the very yellow azo colour given by the original that the rate 
of fall accords so well with the others. ; 

Although this method of hydrolysis is not sufficiently sensitive to show 
that a definite percentage of the colour is due to carnosine, it at least gives 
one a sense of assurance in the use of a very unspecific reagent for the esti- 
mation of carnosine in such a complex solution as muscle extract. It would 
further seem to indicate that carnosine is responsible for a very high percentage 
of the colour as measured from fresh muscle extracts treated only with lead 
acetate. 

The specific cause of the yellow colour produced in such cases as salmon 
has not been found. In the case of rabbit muscle, it has been observed that 
the colour is more red when the carnosine content of the muscle is high and 
yellow when low. Some cat and frog muscles with a low carnosine content 
also showed a yellow colour. This is not surprising as the chances of inter- 
ference are greatly increased when the test portions are less dilute. 

Though purines certainly give rise to a small error and exact quantitative 
results cannot be got when the colour is not of the right tint, yet the total 
error in the lead acetate filtrate is too small to eliminate. With a better 
hydrolytic agent it might yet be possible to evaluate the error. 


THE ESTIMATION OF CARNOSINE IN MUSCLE EXTRACT 653 


CARNOSINE CONTENT OF SOME MUSCLES. 


With the method of extraction and treatment of the extract as previously 
described, the following results have been obtained. Each result represents 
-the percentage colour value reckoned as carnosine in the fresh skeletal 
muscle. Further, each result represents the average of at least two results 
‘obtained from different pieces of the same muscle. | 

For convenience the amounts found are given from lowest to highest 
carnosine contents. The contents for four members of each species of animal 
are given. 


Rabbit muscles 0-026 0-064 0-090 0-101 % carnosine 
Frog i? 0-107 0-128 0-142 0-280 ,, as 
Cat me 0-123 0-203 0-336 0-380 ,, . 
Ox a3 0-340 0-400 0-515 0-640 ,, » 

The results show that the carnosine content of muscle varies not only 
with the species of animal but varies greatly in different animals of the same 
species. The highest values obtained in the one species are two to four times 
greater than the lowest values in the same species. This finding is at variance 
with that of Clifford [1921, 2]. This worker finds, for example, ox muscle to have 
an almost constant carnosine content of 1-1 %. But apart from the constancy 
of the results found, the writer is compelled to question their accuracy. In a 
previous paper [1921, 1] Clifford finds that 0-02 mg. of histidine gives a value 
of 10-75 mm. measured with the Koessler and Hanke histidine standard. 
With these data—and assuming the author means 0-02 mg. of histidine 
dihydrochloride—the molecular colour value of the histidine is 

aoe: x 228 x 10%, 
which is approximately 122 million mm. Sy. 

With the same colour standard Clifford records that 0-1 mg. of carnosine 
gives a reading of 30 mm. The molecular colour value of Clifford’s carnosine 
is thus 30 x 226 x 104 or approximately 68 million mm. As previously stated, 
the writer finds the molecular colour value of carnosine against this same 
colour standard to be 161 million mm. Assuming this figure represents 100 % 
pure carnosine, the carnosine employed by Clifford is thus only about 43 % 
pure. The carnosine content of the various muscles tested by Clifford should 
from this point of view be about 43 % of the values actually given. 


SuMMARY. 


1. Extracts to be tested for carnosine should be free from ammonium 
salts, sulphides, phenols and aldehydes. 

2. The degree of yellowness of the azo colours developed from muscle 
extracts is in the order of the percentages of the intensification of the azo 
colours as given by sodium hydroxide and hydroxylamine. 

3. In ox muscle purines are responsible for about 3 °% of the colour value 
reckoned as carnosine. In the same tissue there is probably about other 


Bioch. xv1 43 


654 G. HUNTER 


2 % of the colour not due to carnosine. The error in the method due to colour- 
producing substances other than carnosine is probably about 5 % in ordinary 
muscles. 

4, Histidine under certain circumstances may be present in muscle 
extracts. 

5. A test has been found to distinguish histidine from carnosine. 

6. A more certain means than the diazo reagent is given for the identifi- 
cation of carnosine in any tissue. 

7. The fall in colour value of hydrolysed meat extracts agrees with that 
of pure solutions of carnosine. 3 

8. The carnosine content of muscle varies with the species of animal and 
with different members of the same species. 


The writer is much indebted to Prof. Cathcart for his inspiring guidance 
throughout this work. 


REFERENCES. 


Barger (1914). Simple Natural Bases. 
Clifford (1921, 1). Biochem. J. 15, 400. 
Clifford (1921, 2). Biochem, J. 15, 725. 
Hanke and Koessler (1920, 1). J. Biol. Chem. 48, 527. 
—— (1920, 2). J. Biol. Chem. 48, 557. 
—— —— (1922). J. Biol. Chem. 50, 235. 
Hunter (1921). Biochem. J. 15, 689. 
Jones (1920). Nucleic Acids. 
Knoop (1908). Beitrdge, 11, 356. 
Koessler and Hanke (1919, 1). J. Biol. Chem. 39, 497. 
(1919, 2). J. Biol. Chem. 39, 521. 
Levene (1921). J. Biol. Chem. 48, 177. 
Lusk (1921). Science of Nutrition. 
Neubauer-Huppert (1913). Analyse des Harns. 
Plimmer (1918). Practical Organic and Biochemistry, 293-296. 
Thierfelder (1908). Chemische Analyse, 164. 
Totani (1915). Biochem. J. 9, 385, 


LXVII. IDENTIFICATION OF INULIN BY 
A MYCOLOGICAL METHOD. 


By ALDO CASTELLANI anp FRANK E. TAYLOR. 


From the London School of Tropical Medicine and King’s College, 
University of London. 


(Received June 1st, 1922.) 


In previous publications [1917, 1919, 1920] we have described a general 
mycological method, theoretically devised by one of us (C.) some years ago 
in Ceylon, which we have found useful in the identification of various carbo- 
hydrates and other carbon compounds. We propose in the present paper to 
describe briefly how this method can be applied to the determination of 
inulin. . 

It is generally stated that there is no organism which induces a complete 
fermentation of inulin, that is to say, fermentation with production of gas, 
but one of us (C.) has found a fungus which causes a complete fermentation 
of this carbohydrate with large production of gas. This fungus is Monilia 
macedoniensis Castellani and allied species, which ferment with production of 
gas in addition to inulin the following carbohydrates: glucose, levulose, 
galactose and saccharose. 

By means of, this fungus in conjunction with certain other fungi, it is 
possible to identify inulin, using a modification of the general mycological 
method we described some time ago for the identification of various sugars. 

Technique. Let us suppose we have a substance about which we want to 
decide whether it is inulin or not. A sterile 1 °% solution in sugar free peptone 
water is made and distributed into two tubes, No. 1 and No. 2, each con- 
taining a Durham’s fermentation tube or similar appliance. The following 
procedure is then used: 

(a) No. 1 tube is inoculated with Monilia macedomensis Cast., No. 2 
with Monilia tropicalis Cast. The two tubes are placed in an incubator at 
35-37° for 72 hours. If after that time, No. 1 tube contains gas and No. 2 
tube does not, we can come to the conclusion that the substance is inulin. 
This is easily understood by keeping in mind the fermentative reactions of 
the two monilias: Monilia macedoniensis ferments with production of gas, 
only the following carbon compounds: glucose, levulose, galactose, saccharose 
and inulin. Monilia tropicalis Cast. ferments with production of gas, only 
glucose, levulose, maltose, galactose and saccharose. 

Monilia macedoniensis Cast. + ieee 
Monilia tropicalis Cast. 12 ae : 
43—2 


656 A. CASTELLANI AND F. E. TAYLOR 


(6) No. 1 tube is inoculated with Monilia macedoniensis Cast.; No. 2 
with Monilia rhoi Cast. The two tubes are placed in an incubator at 35-37° 
for 72 hours. If after that time No. 1 tube contains gas and No. 2 does not 
we can come to the conclusion that the substance is inulin. This is easily 
understood remembering that Monilia macedoniensis ferments with produc- 
tion of gas, only glucose, levulose, galactose, saccharose and inulin, and 
Monilia rhoi ferments with production of gas, only glucose, levulose, galactose 
and saccharose. 

; Monilia macedoniensis Cast. +- 
Monilia rhoi Cast. 0 


(c) No. 1 tube is inoculated with Monilia macedoniensis; No. 2 with 
B. pseudocoli or B. neapolitanus, or any other strain of the communior group 
of B. coli (ferment saccharose). The tubes are incubated at 37° for four days. 
If then tube No. 1 contains gas and tube No. 2 does not, we can again come 
to the conclusion that the substance is inulin, since glucose, levulose, galactose 
or saccharose would have been fermented also by B. pseudocoli or B. neapoli- 
tanus or any other strain of the Coli communior group. 


= Inulin. 


Monilia macedoniensis Cast. B 3 
B. coli communior (B. pseudocoli Cast., = Inulin. 
B. neapolitanus Emmerich, etc.) 0 


(dq) No. 1 tube is inoculated with M. macedoniensis Cast., No. 2 tube 
with B. asiaticus Cast. The two tubes are placed in an incubator at 37° for 
four days. If after that time No. 1 tube contains gas and No. 2 does not, we 
can come to the conclusion that the substance according to all probabilities 
is inulin. This is easily understood by remembering the fermentative reactions 
of the two organisms. M. macedoniensis ferments only glucose, levulose, 
galactose, saccharose and inulin with production of gas; whilst glucose, 
levulose, galactose and saccharose are also fermented by B. asiaticus; it must 
therefore be inulin. 


Monilia macedoniensis Cast. + ratte 

B. asiaticus Cast. he ae : 

IDENTIFICATION OF INULIN WHEN PRESENT WITH SOME OF THE 
MORE COMMON FERMENTABLE SUBSTANCES. 


If we suspect that a liquid contains inulin mixed with some of the more 
usual fermentable substances such as glucose, levulose, maltose, etc., we can 
find out the presence of inulin in the following manner. The mixture is 
fermented with Monilia tropicalis Cast. If, after exhaustion with M. tropicalis, 
the liquid can still be fermented with M. macedoniensis with production of 
gas, the inference is that the liquid contained inulin. Of course, the precaution 
should be taken of selecting strains of M. tropicalis and M. macedoniensis 
with approximately equal fermentative power on glucose, levulose, galactose 
and saccharose, which carbohydrates they both ferment. 


IDENTIFICATION OF INULIN BY A MYCOLOGICAL METHOD 657 


ADDENDUM. 


For the reader’s convenience we annex a table containing the fermentative 
characters of the various fungi and bacteria we use in our method, and we 
give also a list of the principal mycological formulae which we have devised 
and employed in the identification of various sugars and other carbon com- 
pounds. It is essential to use strains with permanent biochemical reactions. 
Acid fermentation without production of gas is not taken into account. 


Table showing fermentation reactions of certain fungi and bacteria, 


Monilia baleanica Cast. 


- Krusei an 


M. gn Sehr Cast. 


M., tropicalis Cast. 
M. rhoi Cast. 


M, macedoniensis Cast. 
Bacillus coli Escherich 


B. pseudocoli Cast. 


B. paraty phosus B var. M ea 
B. parat _ovg uct sa enaeaaaal 


B. asiaticus Cast. 


B. pseudoasiaticus Cast. 
G = gas; 0 = absence of gas. Simple acid fermentation is not take: 


RRAQQOM MAMDMOE| Glucose 
RMAMMQQM MRAMQD|ARS Levulose 
RARMRMPMQM COMRBMRoSOS Maltose 


Myco.ocicaL FoRMULAE. 


QMQQQANM QMAQARecce Galactose 


Monitlia macedoniensis Cast. 
M. tropicalis Cast. ose 


M. macedoniensis Cast. 
M. rhoi Cast. 


M. macedoniensis Cast. 


Bacillus coli communior (B. ‘pseudocoli, B. ‘neapolitanus) 


M. macedoniensis Cast. 
B. asiaticus Cast. 


SSS S55 


SS 55 5555 


tropicalis Cast. 


macedoniensis Cast. 


metalondinensis Cast. ... 


macedoniensis Cast. 
pinoyt Cast. 

kruset Cast. 

pinoyi Cast. ... 
macedoniensis Cast. 


metalondinensis Cast. ... 


pinoyt Cast. 


metalondinensis Cast. ... 


kruset Cast. 
macedoniensis Cast. 


Agcy 
ERE 
3 8 3 
Rn ae 
g--6° 0. 0 
0.5396"? 6 
0. -.8:°-6° 70 
Oh 2-0 
G0 0 0 
GO 0)-:8 
G 0 0 0 
0GG4GG 
G GGG 
00GG 
0048 G 
Go GO 
GO0O0G+«G 


Inulin. 


Maltose. 


RRRARRRRWcoocccoe Dextrin 


Galactose. 


RROORR cococccoe Raffinose 


RPRRRPRRocococcoe Arabinose 


a ae 

€43 E 8 

ns) bs aa 

q4aan Dd 

OS OO  Or-8 

UAT Tt) Cet joa 

OO. 70) 6-0 

oe Se ae a I 

OO 3Os. sine 

O20" -6 105-8 

OAS On "Oe @ 

006464064 

00GO0G 

G65 G20". 

Of" E4: 20. 6 

00G0G 

00G4G06G 
nm into account, 

+) _ ‘ 

+) _ : 

| = Inulin 

So x ° 

) | = Inulin 

+) — Inui 

| = Inulin 

} | = Maltose 

0 

ao Go 

5 | = Maltose 

cial tse 

) | = Maltose 

+). 

5 | = Maltose 


SoOOoORCS coooccoce Inositol 


ROSOMR coocccoe AHalicine 


coosos oococco Amygdalin 


RRPARMAR coooccoe Isodulcitol 


eccooceso Scoosccco Erythrytol 


658 


A. CASTELLANI AND F. E. TAYLOR 


M. tropicalis Cast. 

M. bronchialis Cast. 
M. tropicalis Cast. 

M. macedoniensis Cast. 
M. krusei Cast... oe 
B. paratyphosus B Schottmiiller 


Saccharose. 


tropicalis Cast. 
M. metalondinensis Cast. . 


M. rhoi Cast. 


” M. pinoyi Cast. 


M. tropicalis Cast. Se: 

B. coli communis (sensu stricto) . 

M. tropicalis Cast. pa 

B. paratyphosus B Schottmiiller 

M. macedoniensis Cast. a 

B. coli communis ies stricto) ... 

B. coli communior .. is “ee 

M. macedoniensis Cast. bis re 

B. paratyphosus B Schottmiiller 

B. coli communior . ey aia 
B. coli communis Escherich pone ham 
B. neapolitanus Emmerich “s ts 
B. coli communis Escherich ee srt) 
B. asiaticus Kat 


poe 

M. krusei Cast. 

M. pinoyi Cast. , my 
Glucose. 

M. baleanica Cast. Ge We re eat 

M. krusei Cast. ae re = Bie 
Inositol. 


B. paratyphosus B var. M Schottmiiller 
B. paratyphosus A Schottmiiller a 


Galactose (continued). 


> } = Saccharose 

‘ } = Saccharose 

. | = Saccharose 
) 

> = Saccharose 

+ 

0 = Saccharose 

ob 

a 

0 = Saccharose 

+) 

af = Saccharose 

ia co charose 

; Levulose 


of = Glucose 


3} = Inositol 


CHEMICO-MYCOLOGICAL FORMULAE. 


Saccharose. 
Fehling 
M. ne THE. Cast. ave ave 
Lactose. 
Fehling 


B. poralyphosue B Schottmiiller Sis is 
B. coli communis Escherich ove my 


Pentose. 


Fehling a 

M. tropicalis Cast. Sid 
B. paratyphosus B Schottmiiller 
B. coli communis Escherich 


. | = Saccharose 


+ 

0 = Lactose 

+ 

0] 
= (generally 

7 arabinose) 


+ = gas; 0 = no gas; simple acid fermentation is not taken into account. 


REFERENCES, 
Castellani (1920). Lancet, i, 847 


Castellani and Taylor (1917), Brit. Med, J, ii, 855. 


—— —— (1919). Brit. Med. J. i, 183. 


LXVIII. FEEDING EXPERIMENTS IN CONNEC- 
TION WITH VITAMINS A AND B&B. 


III. MILK AND THE GROWTH-PROMOTING VITAMIN. 


IV. THE VITAMIN A CONTENT OF REFINED COD- 
LIVER OIL. 


By ARTHUR DIGHTON STAMMERS. 
Work carried out at the Research Laboratory, Port Sunlight. 


(Received July 10th, 1922.) 


III. 


Durine 1920 Osborne and Mendel [1920] published a paper entitled “Milk 
as a Source of Water-soluble Vitamin,’ in which they state that they are 
unable to produce results similar to those obtained by Hopkins in his classical 
experiments published in 1912 [ Hopkins, 1912}. 

The administration of 2 cc. of milk per day in these experiments was 
sufficient to provide the accessory factors for growing rats, but the American 
authors did not obtain anything approaching normal growth until at least 

16 ec. per day were supplied. 
Osborne and Mendel’s experiments included the use of winter milk (which 
was probably deficient in vitamins owing to stall-feeding of the cattle) and 
also milk from grass-fed animals. In each case comparable results were 
obtained. The authors also considered the possibility of an unsuspected 
inorganic substance in Hopkins’ salt mixture, but experiments showed that 
this was not the cause of the discrepancy. 

Subsequent to this, Hopkins [1920] repeated his experiments and con- 
firmed his previous findings, without, however, being able to throw any light 
on the discrepancigs of the American investigators. He suggests that a 
seasonal variation in the food value of milk may be a contributing cause, 
although an experiment upon goats’ milk, secreted on winter and summer 
diets, planned to elucidate this point, failed to yield much evidence. He also 
suggested that there was a seasonal variation in the growth energy of rats. 

In view of these conflicting results, the experiment now to be described 
may be of some interest, although the circumstances under which it was 
carried out were not comparable (as regards age of animals) to those in the 
experiments of Hopkins and of Osborne and Mendel referred to above. 

The chief point of difference was that the animals used in the present 
work were the survivors from a previous experiment which had already 


660 A. D. STAMMERS 


lasted for 101 days, the results of which showed conclusively that the materials 
tested were almost entirely deficient in vitamin A. The average weight of 
the animals at the commencement of this experimental period was 57 g. and 
at the termination 80 g.—an increase of only 23 g. as compared with the 
normal for this period, which is 142. 

There were ten of these survivors and their average age when selected for 
use in this new experiment was 150 days. Eight of them were showing 
marked signs of keratomalacia and general lack of condition and the other 
two were also out of condition but not to the same extent. 

They received throughout the experiment 2 cc. of cow’s milk per animal 
per day and the only other food supplied was entirely deficient in vitamins 
and composed as follows: 


Basal dietary Jés is ip ones 
Steam distilled tatietedcnal oil . 15% 


[For method of preparation see Stammers, oe No antineuritic or anti- 
scorbutic was given, the milk supplied being the only source of these factors 
available to the animals. 

The experiment was carried on for 111 days, and during this period the 
animals were weighed twice a week, first thing in the morning, before feeding. 

It should be stated that the experiment was carried out during the months 
of December, January, February and March, and the milk used was obtained 
from stall-fed cows. Butter-fat obtained from the same source during this 
period showed the effect of stall feeding in giving growth far below the normal 
(see graph) when used as a control in other experiments, so that the milk 
may be considered to have been of fairly poor quality as regards its vitamin 
content. 

RESULTS. 


As already mentioned, the average weight of ten animals at the commence- 
ment of the experiment was 80 g. and their average age was 150 days. In 
111 days the increase in average weight was 38 g., i.e. 80-118. 

The normal increase for this period and age is 43 g., ¢.e. 202-245; hence 
there is a deficit of only 5 g. in average weight over a period of nearly sixteen 
weeks, 

The animals improved out of all knowledge in general condition, the 
keratomalacia disappearing and the only difference from normal animals was 
in their weight and size. Otherwise, they appeared to be in excellent health. 

A graph is appended (Fig. 1) which shows 


A, growth from milk, 
a, normal for A, 
B, C, growth from winter butter-fat, 
b, c, normals for B, C. 


It will be noticed that the butter-fat curves are almost super-imposable. 


FEEDING EXPERIMENTS WITH VITAMINS A AND B_ 661 
250 ; 


240 


230 


220 
210 
200 
190 


180 


170 


D 
=) 


Weight in grams. 
a 
2) 


: i i s 
0 10 20 30 40 50 60 70 80 90 100 110 
Days of experiment. 
Fig. 1. 
a, b, c, Donaldson’s curves of normal growth. 
A, growth from milk. B,C, growth from butter-fat. 


662 A. D. STAMMERS 


CoNCLUSIONS. 


While it is not maintained that this experiment either definitely confirms 
the results obtained by Hopkins, or disproves the statements of Osborne and 
Mendel, it seems to throw some additional weight of evidence in favour of 
the former, although, for the reasons already mentioned, the conditions under 
which the various experiments were carried out are not comparable. 

It is considered, however, that the probabilities are that if an attempt 
had been made to duplicate Hopkins’s experiments by making the conditions 
similar, the results, even with the sample of winter milk used, would have 
confirmed the findings of this worker. 

If young animals had been used, the initial “growth momentum,” as it 
were, derived from the food consumed prior to weaning, plus the 2 cc. of milk 
administered daily during the experimental period, would probably have 
sufficed to give normal growth, having regard to the seasonal variation; in 
other words the growth from milk would have compared favourably, if 
nothing more, with that from winter butter-fat. 

It may be urged that an adult requires less vitamins to keep it in condition 
than does a young animal and an explanation of the results of this experi- 
ment may be found in this fact: on the other hand, this hypothesis is, in the 
writer’s opinion, only applicable in the case of an animal which has received 
a normal diet up to the age of maturity and not to animals such as those 
used in this work, which had been fed on a vitamin-deficient diet since they 
were seven weeks old. 

A further point of interest is now touched upon. When an animal has been 
fed upon a deficient diet and is suddenly placed upon one containing a greater 
or less quantity of the growth-promoting vitamin, a considerable stimulus is 
usually applied by the latter to the growth impulse and this can generally 
be well seen on the graphic record of the growth. It is quite common for 
animals in such a case to put on a large increase in weight in the first week, 
but if this occurs there is usually an almost corresponding drop within the 
next week or so. It will be noticed that, in this case, the effect of any stimulus 
applied by the administration of the milk was not reflected in a sudden jump 
in weight; in fact, during the first week there was a drop and if the curve is 
smoothed out, it will be seen that it remains almost parallel to the normal. 

A possible explanation is that the deficiency disease caused by the previous 
experiment had suppressed the growth impulse to such an extent that the 
latter could not immediately respond to the stimulus applied. It is also 
possible that the greater the age of the animal the greater must be the magni- 
tude of the stimulus to produce this effect, even though this may appear, 
under certain circumstances, to contradict the axiom that the older an 
animal is, the less vitamin it requires to keep it in health. 

A possible criticism may be applied (owing to the advanced age of the 
animals used) to the conclusions drawn and therefore attention is invited at 
this stage to the next paper (IV), 


bd 


FEEDING EXPERIMENTS WITH VITAMINS A AND B_ 663 


It will be observed that in four of the experiments there described, the 
animals were also of a mature age (varying from 130 to 160 days) and in 
three of the experiments had been previously fed upon a diet which had 
failed to maintain adequate growth, thereby approximating to the conditions 
under which the present experiments were carried out. 

Cod-liver oil is well known to be a rich source of vitamin A, but in spite 

of this fact its administration failed to raise the weights of the animals con- 
cerned to the normal, but the curves ran roughly parallel to that of normal 
growth (Donaldson) without any marked approach to the latter. 
_ It might have been expected that cod-liver oil would be sufficiently potent 
to overcome the stagnation of the growth impulse and raise the weights to 
somewhere near the normal, but this did not take place even in any individual 
rat. 

It therefore seems that age is the important factor and that once an 
animal has reached a certain age, either (a) the amount of vitamin which 
would have been adequate to give normal growth in a young animal, is 
insufficient, or (6) no amount of vitamin can fully restore the growth impulse. 

The writer’s opinion is that the latter is probably the correct explanation. 

A study of these results therefore would seem to indicate that the curve 
of normal growth remains constant, within limits, for age and not for weight, 
and, if for any reason the growth impulse has been inhibited, an attempt at 
restoration of the impulse by however powerful a stimulus cannot succeed 
to an extent greater than that which would be expected from the age of the 
animal. 


IV. 


Six experiments are described in this paper and two varieties of cod-liver 
oil were used. It has already been established that crude cod-liver oil is 
exceedingly rich in vitamin A [Zilva and Drummond, 1921]. Of the two 
varieties used in the experiments now described, one, A, was an ordinary 
sample procured from a pharmacist and of a type commonly supplied for 
medicinal purposes; the other, B, was more crude and a brand suitable for 
administration to cattle. Sample A was used in Exps. 1, 2, 4, 5, and 6, and 
sample B in Exp. 3. 

The animals in Exp. 1 were 160 days old at the commencement and 
had been fed (for the previous 110 days) upon butter-fat and used as a 
control in another set of experiments. Their average weight was 148 g. 
Those in Exp. 2 were also 160 days old and had previously been fed (also for 
110 days) upon a partially deficient diet. Their average weight was 130 g. 
In Exps. 3 and 4 the animals were also of mature age (135 days) and had also 
been fed (for the previous 90 days) upon a partially deficient diet. Their 
average weights were 100 and 90 g. respectively. 

Exps. 5 and 6 were confined to young animals. In Exp. 5 they were 43 
and in Exp. 6, 33 days old. Their average weights were 77 and 49 g. respec- 


664 A. D. STAMMERS 


tively. The experimental technique was that usually adopted in this laboratory 
and has already been described in this journal [Stammers, 1921]. 

Exps. 1 and 2 were carried on for 101 days, Exps. 3 and 4 for 73 and Exps. 
5 and 6 for 87 days. 


250 
240 
230 
220 
210 
200 
190 
180 
170 


160 


Weight in grams. 


150 


140F / 2 
130 / 2 n0t” 


120 SP ie oe 


if ‘7 
110F a 
rae 
4 di 
100K 
é 
Ys 
9OLe i 1 i i i i i " 1 i ) 
0 10 20 80 40 §0 60 70 80 90 100 110 
Days of experiment. 
Fig. 2. 


a, Normal for Expts. 1 and 2. 6, Normal for Expts. 3 and 4. 
Expt. | Expt, 2 --—--. Expt, 3 vvrreee Expt. 4 ---- 


FEEDING EXPERIMENTS WITH VITAMINS A AND B_ 665 
200r | by, 
190 
180 
170 
160 


150 


140 


© 
i=) 


Weight in grams 
S 
(2) 


j 4 t 4 1 j 
0 10) 20. SO * ae 60." 60°: 70. $0. 80-.100 
Days of experiment. 
Fig. 3. 


a, Normal for Exp. 5. 6, Normal for Exp. 6. 
<1 Expt. 5 (cod-liver oil). Expt. 6 (cod-liver oil). 


RESULTS. 


The results are given in tabular form and graphs are also appended (Figs. 
2 and 3) which show Donaldson’s normal curves in addition to the growth 
recorded, 


666 - A. =D, STAMMERS 


Normal 

Age at Weight at increase (for 

Previous commence- commence- Weightat Actual age at com- 

Exp. treatment ment ment conclusion increase mencement) 

days g. g. g. g. 

1s Butter-fat 160 148 180 32 37 
2 ~=— Partially deficient 160 130 191 - 61 37 
3 ——Partially deficient 135 100 133 33 41 
4 Partially deficient 135 91 145 54 41 
5 None 43 77 164 87 141 
6 None 33 48 156 108 144 


Taking the first four experiments, 7.e. those in which mature animals 
were used, it will be seen that in Exps. 1 and 3 the growth was subnormal, 
while in Exps. 2 and 4, although it was supernormal (for age) the growth 
shown was not nearly sufficient to cause the curves to approximate to that 
of normal growth—in other words, although cod-liver oil is probably the 
richest source of vitamin A available, its administration in these cases was 
not attended by any very marked or maintained increase in weight. The less 
refined sample, used in Exp. 3, which, ceteris paribus, would be expected 
to give the best results, actually proved the least effective. 

As regards Exps. 5 and 6 in which young animals just weaned were used, 
the growth was slightly subnormal in each case although the animals were 
considerably above normal weight at the commencement of the experiment: 
the curves in each case fell off gradually and crossed the normal between 
the seventh and eighth weeks. When the experiments were terminated, 
however, they were still on the up grade. 


SUMMARY AND CONCLUSIONS. 


Two samples of cod-liver oil, one more highly refined than the other, 
were tested upon rats of mature age, some of which had previously been fed 
on a diet partially deficient in vitamin A. The more highly refined oil was 
also tested on young animals. 

In the latter case (Exps. 5 and 6) the growth registered was slightly sub- 
normal and, in view of the potency of cod-liver oil as a source of vitamin A, 
expectations were hardly realised. The sample used was, however, highly 
refined and this may be responsible. 

As regards the first four experiments, No. 3 may be considered first. 
Although the sample used in this case was fairly crude, it failed to restore 
deficiency animals to normal weight even after a period of 73 days, moreover 
there was no sharp initial rise in the growth curve as might have been 
expected, 

In the other experiments an initial rise did take place, but after three 
weeks the curves flattened out and the growth recorded ran roughly parallel 
to the normal (for rats of that age). Exp. 2 showed the best results, but even 
after 101 days the animals averaged 54 g. below the normal. 

The explanation offered for these results is (a) that the growth impulses 
of the animals concerned in these experiments had, owing to their ages (135 . 
and 160 days), been deprived to a great extent of their power to react to the 


FEEDING EXPERIMENTS WITH VITAMINS A AND B_ 667 


stimuli applied even by a fat rich in vitamin A and consequently they failed 
to show substantially greater growth than the normal for their age, or (b) 
that the oil had in the process of refinement lost a large part of its original 
vitamin A, 

Point is given to the former theory by the fact that in Exp. 1 the animals 
had previously been fed on a diet containing butter-fat, which would, in 
itself, not be expected to impair the growth impulse. 


I am indebted to Messrs Lever Brothers, Limited, for permission to publish 
this research. 


7 


REFERENCES. 


Hopkins (1912). J. Physiol. 44, 425. 
— (1920). Biochem. J. 14, 721. 
Osborne and Mendel (1920). J. Biol. Chem. 44, 15. 
Stammers (1921). Biochem. J. 15, 491. 
Zilva and Drummond (1921). Lancet, ii, 753. 


LXIX. ON THE CARDIAC, HAEMOLYTIC AND 
NERVOUS EFFECTS OF DIGITONIN. 


By FRED RANSOM. 


From the Pharmacological Laboratory of the London (Royal Free Hospital) 
School of Medicine for Women (University of London), 


(Received July 19th, 1922.) 


DIGITONIN, a sapo-glucoside found in Digitalis purpurea, has certain well- 
known properties—it causes haemolysis, forms a combination with cholesterol 
and some other sterols and has a marked action upon cardiac muscle. 

The objects of the present investigations were (1) to compare the effects 
of certain allies of cholesterol upon the cardiac and haemolytic actions of 
digitonin and (2) to ascertain whether the activities of a digitonin solution 
are modified by shaking with an indifferent adsorbent. 

The particular digitonin used was prepared and carefully purified by 
Mr J. A. Gardner, whom I desire to thank for his kind help. 

In the first place it was necessary to measure both the cardiac and the 
haemolytic activity of the preparation. 

Haemolysis. Graduated amounts of digitonin, in each case dissolved in 
4 cc. of 0-95 % NaCl solution, were placed in a series of test tubes and 1 ce. 
of a 10 % dilution of defibrinated ox blood in physiological salt solution was 
added to each tube, so that all reactions took place in 5 ce. of fluid. The tubes 
were left standing in the laboratory for about 18 hours and then the condition 
of their contents was noted, i.e. whether haemolysed or intact. In Table I 
the results of such a test are set out. Repeated experiments showed that the 
minimum quantity of the digitonin which would, under the above conditions, 
completely lake 0-1 cc. of ox blood was from 0-0032—0-0036 mg., corresponding 
to a dilution of the digitonin of from 1 : 156,000 to 1 : 139,000, With cat’s 
blood the minimum complete laking dose was rather less than 0-004 mg. 


Table L. Haemolytic power of Gardner's digitonin. 
Digitonin NaCl Ox blood Dilutionof Amountof Condition after 20 hrs. 


0-004 %, 0-95 % 10%, digitonin —_ digitonin at room temp. 

0-4 ce. 3-6 co. 1 co, 1:312,500 0-0016 mg. Intact 

Ob 35 1 1: 250,000 0-0020 Very slight haemolysis 
Ob 3-4 1 1: 208,000 0-0024 Slight haemolysis 

7 +3 1 1; 178,577 0-0028 Nearly haemolysed 

0-8 3-2 l 1; 156,250 0-0032 Very nearly haemolysed 
09 +1 l 1: 139,000 00-0036 Completely haemolysed 
10 3-0 | 1 ; 125,000 0-004 ” ” 


PHYSIOLOGICAL ACTION OF DIGITONIN 669 


Evaporation to dryness on a water-bath of a digitonin solution, whether 
in water or 90 % alcohol, did not affect the haemolytic power (Table II) but 
the presence of NaCl in hypertonic solution increased the resistance of the 
red cells to haemolysis by digitonin (Table III), as was previously observed 
by Luger [1921] for saponin. 


Table II. Digitonin solution evaporated to dryness and redissolved. 


Digitonin NaCl Ox blood Condition after 
0-004 % 0-95 % 10% 18 hours 
0-4 ce. 3-6 ce. 1 co. Trace of haemolysis 
0-6 3-4 1 Partial > 
0-8 3:2 1 Complete oa 
0-9 3-1 1 ri Fe 


Table III. Protective action of hypertonic NaCl. 
Condition after 18 hours 


Digitonin NaCl 10 % blood at room temp. 
‘004 mg. 4 cc. -95 %, 1 ce. Complete haemolysis 
ao . ce. ar: Zo : i 3 

° cc. 6:0 % oO se 

008 4 cc. 5-0 o 1 Complete a 


i ve F (Mii \ i 
mie hl alts 


Fig. 1. Frog’s heart perfused with Ringer’s fluid. At arrow perfusion 
changed to Ringer + 0-004 % digitonin. 


Cardiac action. Solutions of the digitonin in oxygenated Ringer’s fluid 
were perfused through frogs’ hearts in situ from a cannula inserted into the 
inferior cava, the fluid passing at a constant and moderate pressure through 
the heart and escaping by the cut aorta. A number of such perfusions showed 
that 0-0001 % of digitonin had very little or no effect; 0-001 % produced 
increase in the beat followed by rise of tone, diminution of diastole and finally 
stoppage or nearly so: 0-004 % produced the same effects somewhat more 
quickly (Fig. 1). Evaporation of a watery solution to dryness on a water- 
bath and redissolving in Ringer’s fluid did not affect the activity. Perfusion 
of a frog’s heart with a Ringer’s fluid modified by omitting Ca quickly causes 
diminution of efficiency; if now, in the continued absence of Ca, 0-001 % 
digitonin is added to the perfusing fluid the normal efficiency is first partly 
restored and then the characteristic effect of digitonin is developed. 


Bioch, xvi 44 


670 F. RANSOM 


It is well known that digitonin treated with cholesterol loses its haemolytic 
power, it seemed therefore of interest to ascertain in how far bodies more or 
less closely allied to cholesterol such as phytosterol, 8-cholestanol, coprosterol 
and pseudo-coprosterol affect the action of digitonin upon the blood and 
upon the heart. 


EXPERIMENTS WITH SUBSTANCES ALLIED TO CHOLESTEROL. 


Method. A solution containing 0-004 % (or other percentage) of digitonin 
in 90 % alcohol was mixed with an equal quantity of a solution of the same 
strength of one of the cholesterol allies, also in 90 % alcohol. The mixture 
was evaporated to dryness on a water-bath and the dry residue taken up in 
so much Ringer’s fluid as would give a possible concentration of digitonin 
and the sterol equal to that of the original alcoholic solution. This solution 
of the dry residue was filtered and the filtrate tested for its haemolytic power 
and cardiac effect. The results obtained with phytosterol, 8-cholestanol, and 
coprosterol were identical—in each case both the haemolytic effect and the 
cardiac action were either cut out or greatly diminished (haemolytic effect, 
Table IV). On the other hand, the mixture digitonin + pseudo-coprosterol 
retained both its haemolytic and cardiac activity. 


Table IV. Effect of phytosterol, B-cholestanol, coprosterol and y-coprosterol. 


Possible concentration of digitonin 0-004 %. 
‘Solution NaCl Ox blood Dilution of Condition after 18 hrs. 


of residue 0-95 % 10% digitonin at room temp. 
Phytosterol or 1 ce. 3 ec. 1 ce. 1: 125,000 No haemolysis 
B-Cholestanol , 2 2 1 1: 62,500 os 
or Coprosterol 4 -- 1 -1:31,250 
y-Coprosterol 1 3 1 is 125, 000 Complete haemolysis 


Control for above: Digitonin 0-004 % in 90 % alcohol evaporated to dryness and taken 
up in NaCl, so.as to give a possible 0-004 % digitonin. 


Solution NaCl Ox blood Dilutionof Condition after 18 hrs. 


of residue 0-95 % 10% digitonin at room temp. 
0-6 ee, 3:4 ce. 1 ce, 1: 208,333 No haemolysis 
0-8 3-2 1 1: 156,250 ‘Partial haemolysis 
1-0 3 1 1: 125,000 Complete _,, 


When stronger solutions of digitonin were mixed with correspondingly 
increased amounts of phytosterol, B-cholestanol or coprosterol the protection 
against haemolysis and cardiac action was still complete (Table V). 


Table V. Protective action of phytosterol, B-cholestanol or coprosterol. 


Possible concentration of digitonin 0-1 %. 


Solution NaCl Ox blood Dilution of Condition after 18 hrs. 
of residue 0-95 %, 10% digitonin at room temp, 

05 ce. 3-5 ce, 1 ce, 1: 10,000 No haemolysis 

10 30 1 1 ; 5,000 ” 


40 — 1 1: 1,250 ” 


PHYSIOLOGICAL ACTION OF. DIGITONIN 671 


If the haemolytic test was carried out with mixtures containing a certainly 
haemolytic dose of digitonin and a gradually diminishing amount of phyto- 
sterol, B-cholestanol or coprosterol, the effect of these bodies in diminishing 
and finally extinguishing the haemolytic effect of the digitonin (Table VI) 
was perhaps still more strikingly in evidence. 


Table VI. Protective action of phytosterol. 


Digitonin Phytosterol Cat’s blood Condition after 20 hrs. 
0-01 % 0:01 % 10% at room temp. _ 
O5ec. + O5cc. } 1 ce. No haemolysis 
0-5 + 0-4 Evaporated 1 Pa 
0-5 + 03 to ess 1 3 
0-5 + 0-2 and 4 cc. NaCl 1 Pe 
0-5 + Ol added to 1 Some haemolysis 
0-5 + 0:05 each, then 1 Complete haemolysis 
0-5 — 1 


” 


In all the above experiments the solutions used were of equal percentage 
for the digitonin and the cholesterol allies; now the molecular weight of 
Gardner’s digitonin is 1202, that of cholesterol and its isomer phytosterol 
is 386, so that when equal percentages by weight are used 1 phytosterol 
corresponds molecularly to 3-1 of digitonin. It was found that 0-31 cc. of a 
0-01 % solution of digitonin (0-0031 mg.) will very nearly or sometimes com- 
pletely haemolyse 0-1 cc. of ox blood. If the relationship between digitonin 
and cholesterol in the digitonin-cholesteride is monomolecular [Windaus, 
1910], then 0-31 cc. of a 0-01 % solution of digitonin should be inactivated 
by 0-1 cc. of a 0-01 % solution of phytosterol or B-cholestanol or coprosterol 
(since the difference between the molecular weight of phytosterol, 386, and 
that of B-cholestanol or its isomer coprosterol, 388, is too small to be of im- 
portance in the conditions of the experiment); moreover multiples of these 
quantities should also be inactive. On the other hand, if the mixtures contain 
relatively an excess of digitonin more or less haemolysis should occur. In 
Table VII the results of an experiment carried out on these lines are shown. 
It will be seen that when molecular equivalents were mixed the neutralisation 
of the haemolytic action of the digitonin was complete, even when multiples 
of the minimum haemolysing dose were employed; when, however, the 
digitonin was relatively in excess haemolysis occurred and was most marked 
in the tubes containing the smallest relative proportion of the sterol. 

As regards the cardiac activity it was found that the effect upon the heart 
ran parallel with that upon the red cells, the non-haemolytic mixtures had 
no cardiac effect and when the haemolytic action was not completely cut 
out there was also some cardiac effect. The results recorded in Table VII 
are strongly reminiscent of what occurs when diphtheria or other toxins are 
mixed with their respective antitoxins, and might perhaps be explained as 
simple adsorption phenomena (see below) were it not that other evidence 
points to a definite chemical union between digitonin and cholesterol. More- 
over the fact that pseudo-coprosterol does not interfere with either the 


44—2 


672 F. RANSOM 


haemolytic or the cardiac action of digitonin strongly supports the chemical 
view. In view of the chemical relationship between cholesterol and cholalic 
acid a series of experiments was made with this substance on the same lines _ 
as those recorded above, using a 0-01 % solution of the acid with 0-01 % of 
digitonin, but no evidence was obtained which would indicate that cholalic 
acid has any influence either upon the haemolytic or the cardiac action of 
digitonin. 


Table VII. 
All reactions in 5 cc. of fluid. 
0-01 % sol. of 0-01 % sol. of Cat’s blood Condition after 18 hrs. 
coprosterol digitonin 10% at laboratory temp. 
0-1 cc. 0-31 cc. lee. No haemolysis 
0-2 0-62 1 Molecular e, 
0-4 1-24 1 equivalents %. 
1-0 3-1 1 ” 
0-05 0-31 1 No haemolysis 
0-1 0-62 1 Digitonin Trace of haemolysis 
0-2 1-24 1 in excess Partial = 
0-5 3-1 1 Complete _,, 
0-2 3-1 1 sy Digitonin in ( Completely haemolysed 
0-4 3-1 1 excess, constant, | am - 
0-6 3-1 1 coprosterol Nearly bs 
0-8 3-1 1 varied | No haemolysis 
— 0-1 1 No haemolysis 
_ 0-2 1 ome 
Controls for 2 ‘ 
— 0-3 1 eens Complete haemolysis 
wei 0-4 1 ae eaigiay s ‘ 
— 0-5 1 


”? ” 


ADSORPTION EXPERIMENTS. y 


The importance of adsorption in many pharmacological problems suggested 
experiments to ascertain if the solutions of digitonin are affected by indifferent 
adsorbents. To this end experiments were carried out with animal charcoal, 
kaolin and starch. 

Method. 50 ce. of 0-004 % solution of digitonin in distilled water were 
gently shaken with 2 g. of purified animal charcoal (washed and dried) or 
kaolin or starch for 30 mins., then filtered and to the filtrate salts to make 
Ringer’s fluid were added; this isotonic solution was then iter? as to its 
haemolytic and cardiac activities. 

Charcoal. The digitonin solution before treatment with charcoal, but with 
salts for Ringer’s solution added, caused haemolysis as usual (0-8 cc. com- 
pletely); after charcoal 5 cc. had no haemolytic effect. As to cardiac action 
the solution which had been treated with charcoal had no effect, but when 
the perfusion fluid was changed to the digitonin solution untreated with 
charcoal the usual digitonin effect appeared at once. The charcoal remainder 
was shaken with 25 ce. 90 %, alcohol for 30 mins., the alcohol evaporated 
and the residue taken up in 50 ee, Ringer’s fluid, this solution caused haemo- 
lysis and had a marked cardiac action. It is evident that animal charcoal 
very effectively adsorbs digitonin. If the period during which the charcoal 


PHYSIOLOGICAL ACTION OF DIGITONIN 673 


was applied or the amount of charcoal used was diminished the removal of 
the digitonin was not so complete. 
Starch. In this case the digitonin solution was 0-001 %. This solution 
perfused through a frog’s heart gave rise to increased systole and rise of tone, 
after treatment with starch the filtrate showed only a slight remnant of 
digitonin action. As to haemolysis, the results of the test are given in Table 
VIII and show that whereas before treatment with starch 1 cc. of the digitonin 
solution caused complete haemolysis, after starch 5 cc. did not do so. Digitonin 
then is adsorbed by starch and the filtrate from the starch is found to have 
lost both in haemolytic and cardiac activity. 


Table VIIT. 
Before starch 1 cc. of 0-001 % digitonin gave complete haemolysis 
After Hi | 3 = no haemolysis 
9 ” 2 ” > ” ” ” 
” » 3 ” ” ” trace only 
” » 45 % ” ‘. nearly complete haemolysis 


Kaolin. In this case also the digitonin solution used was 0-001 %. Before 
treatment with kaolin 1 cc. completely haemolysed 0-1 cc. of ox-blood, after 
kaolin 5 ce. caused only partial haemolysis (Table IX). As to cardiac action, 
before kaolin the solution gave a characteristic digitonin effect, after kaolin 
there was still some evidence of digitonin effect but much less marked. Clearly, 
then, kaolin adsorbs digitonin from watery solution. 


Table IX. 
Before kaolin’! cc. of digitonin sol. gave complete haemolysis 
After este: = ma Ze no haemolysis 
” ” 4 ” ” ” trace of haemolysis 
» » 45 aS cag ge incomplete haemolysis 


At first glance it would appear as though of the three adsorbents animal 
charcoal were the most effective but much stress cannot be laid upon this 
point, since no attempt was made to insure the same number and size of 
particles, 7.e. surface area, in the amounts of each of the adsorbents. Never- 
theless there are good grounds for believing that charcoal occupies a peculiarly 
favourable position among adsorbents [Michaelis and Rona, 1920]. 

Windaus founded his method of estimating cholesterol by means of 
digitonin upon his discovery that cholesterol forms with digitonin a stable 
compound which is not haemolytic and dissolves with difficulty in alcohol 
and upon the fact that cholesterol esters [Hausmann, 1905] do not form such 
a compound. He considered the digitonin-cholesteride to be an addition 
product and not an ester and gave its formula as C;;H,,O,. (digitonin) + 
Cy,H,,0 (cholesterol), regarding it as a combination of one molecule of chole- 
sterol with one of digitonin without loss of water. The results obtained in 
the present investigation are confirmatory of this view, since they show that 
when phytosterol, coprosterol or 6-cholestanol are mixed in monomolecular 
proportions with digitonin the latter loses completely its haemolytic power, 


674 F. RANSOM 


whereas, if the mixtures contain less than this proportion of the sterol, more 
or less haemolysis takes place according to the degree of deficiency. 

But now the very interesting question arises as to how the cardiac action 
of digitonin is affected. In the first place it may be noted that the effect of 
phytosterol, 8-cholestanol or coprosterol upon the cardiac action runs strictly 
parallel to their effect upon the haemolytic power of digitonin—with mono- 
molecular proportions both the haemolytic and the cardiac actions were cut 
out, whereas in mixtures containing a deficiency of sterol the more nearly 
the haemolytic action was neutralised the less was the cardiac effect and 
vice versa. Further, pseudo-coprosterol which does not prevent haemolysis, 
i.e. presumably does not combine with digitonin, did not affect the cardiac 
action either. These facts suggest that the action of digitonin upon the 
heart muscle depends upon the same factor as that which determines the 
haemolytic effect, viz. that there is present in cardiac muscle a body resembling 
or identical with cholesterol or coprosterol, etc., and that this sterol is at any 
rate in part free and not esterified. Moreover it has so important a function 
to perform in the muscle that its withdrawal as digitonin-sterol profoundly 
affects the physiological activity of the muscle cell. This supposition receives 
important confirmation from the fact that pseudo-coprosterol, which does 
not form a digitonide and does not affect haemolysis, has no effect Eee the 
cardiac action of digitonin. 

To show that the heart muscle probably does fix seanbotiee the following 
experiment was carried out; a kitten was killed by pithing the brain and 
bleeding, the blood being collected and defibrinated. The heart was removed, 
washed free from blood, and the coronary vessels washed out with Ringer. 
The excised heart (7 g.) wasspounded to as smooth a paste as possible, then 
well mixed with 50 cc. 0-002 °% digitonin in Ringer and the mixture placed 
in the ice-chest. After 20 hours the mixture was centrifuged, the liquid poured 
off and divided into two parts, one of which was used to test haemolysis 
(Table X); with the other a frog’s heart was perfused, with the result that 
both the haemolytic and the cardiac effects were markedly lessened. 


Table X. 
Heart emulsion Kitten’s blood 

+ digitonin 0-002 %, Ringer 10 % After 20 hrs. 
0-8 ce, 4-0 co, 0:2 cc. Intact 
10 3-8 0-2 > 
2-0 2-8 0-2 a 
5-0 _ 0-2 E 

Digitonin sol 
0-002 %, 

0°8 ce 4-0 ce, 0-2 ce, Partly haemolysed 
1-0 38 0-2 Nearly completely 
2-0 28 0-2 Completely 


It may be recalled that many of the saponins, including digitonin, are 
powerful fish poisons, acting by paralysing the C.N.S., but that the cholesterol- 
saponide is non-toxic as well as non-haemolytic, In like manner saponin, 


PHYSIOLOGICAL ACTION OF DIGITONIN 675 


e.g. agrostemma saponin, found in the corn cockle, when it causes poisoning 
in human beings does not kill by its action upon the blood but by its effect 
upon the C.N.S. and upon the heart. On the other hand, tetanus toxin, 
besides its characteristic action upon the C.N.S., has definite haemolytic 
powers, indeed tetanus toxin is usually regarded as a mixture of tetanospasmin 
and tetanolysin. Cholesterol neutralises tetanolysin without diminishing the 
specific action of tetanospasmin. So the question arises; is the central nervous 
effect of saponin due to the same factor as enables it to cause haemolysis and 
alteration in cardiac muscle or do the conditions resemble those in tetanus 
toxin? To throw light upon this point the following experiment was carried 
out. A young rabbit was killed by a blow on the neck, the carotids were then 
cut and the blood collected and defibrinated; brain removed (6-7 g.), rubbed 


Fig. 2. Frog’s heart perfused with Ringer’s fluid. At arrow perfusion changed to emulsion of 
rabbit’s brain in Ringer + 0-004 % digitonin. Compare with Fig. 1. 
to a smooth paste in a mortar and then thoroughly mixed with 50 cc. of a 
0-004 % solution of digitonin in Ringer’s fluid; the emulsion was -placed in 
the ice-chest for 20 hours; then well centrifuged, the supernatant fluid poured 
off and divided into two parts, one of which was used to perfuse a frog’s heart, 
the other to test the haemolytic power. Table XI shows the results of the 
haemolytic test; the perfusion tracing is shown in Fig. 2. It will be seen at 
once that the digitonin-brain emulsion has no haemolytic power, at least not 
up to 5 cc. which is very much more than the ordinary haemolytic dose, and 
further the heart tracing shows merely a trace of digitonin action. 


Table XI. Protective action of brain substance. 


Possible concentration of digitonin, 0-004 %. 


Brain Ten rabbits’ . 
+ digitonin Ringer blood After 18 hrs. 

0-6 ce. 4-2 cc. 0-2 ec. Intact 

0:8 4-0 0:2 As 

1-0 3:8 0-2 or 

2-0 2:8 0-2 a 

5-0 — 0-2 f 


Digitonin 0-004 % 
0-8 ce. 4-0 ce. 0-2 cc. Completely haemolysed 


676 F. RANSOM 


Considered in conjunction with the previous experimental evidence, these 
results appear to justify the conclusion that the action of saponin in causing 
narcosis in fishes and paralysis of the C.N.S. in man is an analogous pheno- 
menon to digitonin haemolysis and digitonin cardiac action, 7.e. there is free 
and unesterified in the brain a body allied to or identical with cholesterol, 
phytosterol, etc., which unites with digitonin. Moreover the function which 
this cholesterol-like body has to perform is of such importance that when it 
is inactivated by combination with digitonin (saponin) serious central nervous 
symptoms ensue. 


A 


a Hl 


; a as on il : nl 


ny bi ani isi a 


' 


ce 7 i a mi i | 
pic Mh a ii te nn M aut \ ii 


WENN Wana iti 


i 


Fig. 3. Two frogs’ hearts perfused with Ringer’s fluid. A, at arrow perfusion fluid changed to 
emulsion of cat’s brain in Ringer + 0-001 % strophanthin. B, at arrow perfusion fluid 
changed to Ringer + 0-001 % strophanthin. In each case the break in the tracing represents 
an interval of 15 mins. 


Kobert [1904] has shown that the toxic action of saponin as well as the 
cardiac effects is abolished by the combination of saponin with cholesterol, 
and further that the extreme toxicity of ‘sapotoxin solutions for fishes is 
entirely lost in the cholesterol-sapotoxin combination. 

The effect of digitonin is also neutralised by adding milk to the digitonin 
solution. 

[t is interesting to note that strophanthin, which is a glucoside but not a 
saponin, does not lose its cardiac action when brain emulsion is added to the 
strophanthin solution (Fig. 3). 


PHYSIOLOGICAL ACTION OF DIGITONIN 677 


SUMMARY. 


Monomolecular combinations of digitonin with phytostero], coprosterol, 
B-cholestanol are non-toxic for red blood cells and for the frog’s heart; any 
excess of digitonin beyond this proportion can be detected by the action of 
the solution upon red cells and upon the heart. Pseudo-coprosterol does not 
effect the activity of digitonin. Digitonin mixed with an emulsion of brain 
substance loses its toxicity for red cells and for the heart. 

It is suggested that the toxic action of digitonin upon red blood cells, 
cardiac muscle cells and cells of the central nervous system (fishes) is essenti- 
ally the same—it attacks a substance allied to or identical with cholesterol 
which is present free and not esterified in these various cells. This free sterol 
has in each case so important a function to perform that its inactivation by 
combination with digitonin leads to a profound modification of the physio- 
logical activities of the respective cells. 

Digitonin is readily adsorbed from solution in water (e.g. 1 : 25,000) by 
animal charcoal, kaolin or starch; the adsorption by charcoal appears to be 
particularly complete. 

Cholalic acid has no effect upon the action of digitonin. 


REFERENCES. 


Hausmann (1905). Beitrdge, 6, 567. 

Kobert (1904). Saponinsubstanzen, Stuttgart. 

Luger (1921). Biochem. Zeitsch. 117, 145. 

Michaelis and Rona (1920). Biochem. Zeitsch. 102, 268. 
Windaus (1910). Zeitsch. physiol. Chem. 65, 110. 


LXX. THE HEAT-COAGULATION OF PROTEINS. - 


By W. W. LEPESCHKIN. 
From the Botanical Laboratory, the University of Kasan. 


(Received July 21st, 1922.) 


Ir has been long pointed out that the heat-coagulation of proteins is not so 
simple a phenomenon as it at first may seem to be. Schmidt and his pupils, 
Aronstein [1874], Heynsius [1874], Kieseritzky [1882] and Rosenberg [1883], 
showed that albumin solutions from which the electrolytes had been 
removed by dialysis did not coagulate on heating; heat-coagulation could 
only be produced after the addition of electrolytes to such solutions; it took 
place also if a small amount of any salt was added to the cold albumin 
solutions which had been preliminarily heated. These authors therefore sug- 
gested the hypothesis that heating may transform the albumin into an 
unknown body, the solutions of which coagulate on the addition of a very 
small quantity of salts [Rosenberg, 1883, p. 34]. 

Hardy [1899, 1900, 1906] showed later that in solutions of protein altered 
by heating ultramicrons appear and that the protein can now be precipitated 
by adjusting the reaction so as to render the particles isoelectric with the 
solution. When the protein is made negative by adding alkali the valency of 
cations is of importance for the precipitation; on the contrary, the cations 
have no part in the precipitation if protein is made positive. 

Then Michaelis came to the conclusion [Michaelis] that the properties of 
albumin solutions may be changed by heating from those of an “emulsoid” 
to those of a “suspensoid,” whereas the high viscosity and the low surface- 
tension of solutions of albumin altered by heating point rather to their hydro- 
philic character. These solutions probably do not fall into either category of 
the colloidal state but form an intermediate group with some of the properties 
of both. 

Chick and Martin [1913, 1] who investigated the heat-coagulation of pro- 
teins and the influence of various factors upon its course confirm the conclusion 
that this phenomenon consists of two distinct processes: (1) “denaturation” 
and (2) ‘‘agglutination” of the denaturated protein by electrolytes. They 
regard the former as a reaction between protein and hot water without deter- 
mining this reaction more precisely. | 


THE HEAT-COAGULATION OF PROTEINS 679 


Contradicting the results of the above investigations and of his own earlier 
experiments Pauli [1899, 1908] affirms that he has observed that protein 
solutions free from electrolytes (““amphoteres Hiweiss”’) coagulate on heating 
and that the addition of electrolytes to such solutions increases their coagu- 
lation temperature. Criticism and detailed consideration of this new opinion 
may be postponed till some experiments related to heat-coagulation have 
been described. 

A number of scattered observations indicated that temperature is an 
accelerating factor in the process of heat-coagulation. Buglia [1909] was the 
first who precisely determined the dependence of the velocity of heat-coagula- 
tion upon the temperature, and found this dependence to be logarithmic. 

Chick and Martin [1910] found that the coagulation-rate increases 
1-3-1-9 fold per 1° temperature-rise. Such a high temperature coefficient 
bears only on the denaturation, while the agglutination-velocity does not 
increase so rapidly with the temperature as the denaturation-rate, its tem- 
perature coefficient being equal to 2-5 per temperature-rise of 10° [Chick 
and Martin, 1913, 2]. These authors showed moreover that, if precautions 
are taken to keep the concentration of hydroxyl or hydrogen-ions constant 
during the process, the denaturation may proceed as a reaction of the first 
order. 


(1) ExprrtMentaL Meruops. 


My experiments were made before I was acquainted with the papers by 
Chick and Martin, and the method used by me differs from theirs. I find 
however that it is more convenient for observing slight, and sometimes short, 
alterations of velocity in both the processes that are comprised in heat- 
coagulation. 

In order to estimate alterations in the velocity of heat-denaturation 
and in coagulation (“agglutination”) of the denaturated protein under 
influence of various factors, in my experiments, the time required to produce 
a slight turbidity in the protein solutions like that of a ground glass (chosen for 
all the experiments) was noted. The solution to be tested was usually intro- 
duced by means of a capillary pipette into a small vessel made of two cover- 
glasses (20 x 10 x 0-1 mm.) cemented with oil-lacquer (linseed-oil and 
minium) on both sides of a rectangular brass plate (0-7 mm. thick, 22 mm. 
long, 13 mm. broad) the middle of which was rectangularly cut out, so that 
the rim-breadth (at three sides) was equal to 3 mm. 

A drawing of this vessel is given in one of my previous papers devoted 
to the reactions between starch and water brought about by heating!, There 
are also given a drawing and a photograph of the thermostat used. 


1 Lepeschkin, “ Etude sur les réactions chimiques pendant le gonflement de l’amidon dans 
leau chaude.”’ Le manuscrit est renvoyé 4 M. le professeur Chodat (Généve, Suisse) pour la publi- 
cation dans l’ Archives des sciences naturelles. 


680 W. W. LEPESCHKIN 


The above vessel with the protein solution to be tested was immersed in 
the water of the thermostat, in which the temperature variation did not 
surpass 0-03°. Close by the vessel, a similar rectangular brass plate (cut 
out in the middle) was placed; instead of the cover-glasses (see above) 
a piece of ground glass of the same size was cemented on one side of this plate. 
Both (viz. the vessel and the plate with ground glass) were observed through 
the front glass-wall of the thermostat box by means of a horizontal micro- 
scope of which the eye-piece was removed and the objective was substituted 
by a black plate having two rectangular apertures corresponding respectively 
tothe front cover-glass of the vesseland to the ground glass. The protein solution 
and the ground glass were illuminated by an electric lamp (50 candles) through 
the back glass-wall of the thermostat, and the microscope-tube was situated 
obliquely to the front glass-wall (in order to observe most easily the turbidity 
produced in the protein solution by heating). 

The water of the thermostat was uninterruptedly stirred by means of an 
electric stirrer. 

The time-intervals were defined by means of a seconds pendulum. 

The preliminary experiments showed that the liquid contained in the 
above vessel takes up the temperature of the thermostat water in about 20 
seconds. This interval of time was always deducted from the time required to 
produce the above turbidity of the solution to be tested. 

The material used was filtered egg-white, egg-albumin Kahlbaum from 
which electrolytes had been removed by dialysis, serum-albumin Kahlbaum 
similarly treated or purest recrystallised egg-albumin. 

The first experiments are concerned with denaturation-velocity at various 
temperatures. 


(2) THE TEMPERATURE COEFFICIENT OF DENATURATION. 


Preliminary experiments showed that the coagulation-velocity of the 
denaturated proteins in solutions containing a sufficiently great quantity of 
electrolytes is extraordinarily great in comparison with the denaturation- 
velocity at the same temperature. It could thus be assumed that the coagula- 
tion of denaturated protein proceeds instantaneously at the temperatures 
the influence of which upon the denaturation was investigated in my experi- 
ments. The appearance of the turbidity in the protein solutions containing 
a sufficient quantity of electrolytes marked, therefore, the denaturation pro- 
duced by heating (or is it truer to say the degree of denaturation corresponding 
with the turbidity of the solution like that of the ground glass). The results 
were the following. 

Filtered egg-white. The temperature of water in the thermostat was in 
the experiments: 63-01°, 62-01°, 61-02°, 60-03°, 59-03°, 58-01°, 57-04° and 
56-00°. The average (from six experiments) times (in seconds) passed from the 
moment of immersing the vessel with egg-white into the water to the appear- 
ing of the required turbidity of the solution 7’ were found respectively: 


THE HEAT-COAGULATION OF PROTEINS 681 


40 secs., 97 s., 230s., 595 8., 1535 s., 3720 s., 9520 s., and 22,600s. The tempera- 
ture coefficients per rise of 1° thus were found to be: 2-4, 2-4, 2-5, 2-6, 2-4, 
2-6, 2-4 or on the average 2-5. 

3% solution of egg-albumin Kahlbaum. This had been dialysed during 
5 days and no longer coagulated on heating; in order to make it coagulable 
0-4 % ammonium sulphate was added. Temperatures of the thermostat: 
70-04°, 69-00°, 68-00°, 67-04°, 66-04°, 65-06°. The average (from five experi- 
ments) times T': 91 secs., 136 s., 207 s., 3068., 452s., 690 s. The average tempera- 
ture coefficient per rise of 1° = 1-5. 

2-5 % solution of serum-albumin Kahlbaum. This had been dialysed during 
20 hours and no longer coagulated on heating. 0-7 % potassium chlorate 
was added. The temperatures were: 75-00°, 72-00°, 70-00°, 68-00°. The 
average times 7’: 59 secs., 2058., 470s., 1298s. The average temperature 
coefficient = 1-6. 

1 % solution of recrystallised egg-albumin (not coagulating on heating). 
1-5 % ammonium sulphate was added. The temperatures were: 70-00°, 
69-00°, 68-00°, 67-00°. The average times 7’: 154 sees., 331 s., 784 s., 1860 s. 
The average temperature coefficient = 2-3. 

The coefficient of denaturation thus is equal to 1-5-2-5 per 1°, viz. 58-9540 
per 10°. A similar coefficient was found by me for the chemical reaction 
between starch and water during the swelling of starch in hot water. See 
footnote p. 679. 


(3) ON THE NATURE OF DENATURATION. THE INFLUENCE OF HYDROGEN- 
AND HYDROXYL-IONS UPON THE DENATURATION-RATE. 


Chick and Martin’s researches [1910] show that water is of great signifi- 
cance in the denaturation process of proteins, and reasoning from analogy with 
the above mentioned reaction between starch and water, I should incline, 
in accordance with Chick and Martin, to regard this process as a chemical 
reaction between protein and water. Further we shall see indeed, that the 
chemical properties of denaturated protein differ considerably from those of 
native protein. 

In my cited paper it was suggested that during the swelling of starch in 
hot water no hydrolysis takes place but the starch forms a compound with 
water. Indeed, the velocity of the reaction between starch and water increases 
not at all proportionally to the hydrogen-ion concentration but depends upon 
the nature of the acid added, e.g. sulphuric acid, in weak concentrations, 
even lowers the reaction-rate. So then the hypothesis, that starch pre- 
liminarily forms a compound with acid before it reacts with water, must be 
adopted. The preliminary formation of compounds with potassium and sodium 
hydroxide which react more easily with water than free starch is also certain. 

The fact that denaturation proceeds as a reaction of the first order (mono- 
molecular reaction) indicates that the protein reacts with the solvent, but 
leaves it uncertain whether the denaturation is a hydrolysis, like the inversion 


682 W. W. LEPESCHKIN 


of sugar, or is accompanied by the formation of a compound between protein 
and water. It is important in this connection to know the effect of acids 
and alkalies upon the denaturation-velocity. 

Chick and Martin [1913] who found the denaturation-rate to be increased 
with increasing concentration of hydrogen- or hydroxyl-ions in protein 
solutions suggest that the cause of this phenomenon lies in the fact that 
protein forms salts with acids and alkalies, and that these salts are in a more 
intimate association with water. 

The researches of Hofimann [1889, 1890], however, had already shown that 
slight quantities of acid which are added to a pure protein solution form no 
salts with protein, and this result is in accordance with the recently found 
weakly acid nature of proteins (about 0-00001 N hydrogen-ion concentration). 
Nevertheless my experiments showed that a very remarkable increase of 
denaturation-rate is observed even when albumin solutions contain 0-00016 
mol. nitric acid per litre. The increasing effect of such small concentrations of 
acid upon the denaturation-rate cannot be explained except by an acceleratory 
influence of hydrogen-ions like that observed in the case of hydrolysis of 
carbohydrates, viz. by their catalytic properties. The denaturation-velocity 
was however found in my experiments not to be proportional to the con- 
centration of hydrogen-ions present in the solution when strongly different 
concentrations of the same acid had been tested. It increased more slowly 
than the hydrogen-ion concentration when the latter was not greater than 
0-002 mol. per litre, but it began to/increase more rapidly when the hydrogen- 
ion concentration attained 0-01 mol. per litre. 

However, if various acids were tested at about the same molecular 
concentrations, the denaturation-velocity was found to depend upon the 
hydrogen-ion concentration, so that contrary to the case of the reaction 
between starch and water, the nature of the acid has here no influence upon 
the denaturation-velocity. 

The material used in my experiments was a solution of egg-albumin 
Kahlbaum which had been dialysed during many days. The solution was 
diluted partly with water, partly with nitric acid so that the albumin-content 
in all solutions was 0-66 %. In order to make the solutions obtained coagulable 
0-8 °% (NH,),80, was added. When the concentration of nitric acid was 0 N, 
0-00016 N mol., 0-00032 N, 0-002 N, 0-016 N, the average time of coagulation 
at 63° was found respectively to be 2440 secs., 1550 s., 1220s., 315s., 11s. 
In the case of different acids the average coagulation-time at 53° was found 
to be equal to 167 secs, in the presence of 0-016 N HNO, (viz. hydrogen-ion 
concentration = 0-015 N), 161 sees. in the presence of 0-015 N HCl (hydrogen- 
ion concentration = 0-014 N), 86 secs. in the presence of 0-018 N H,SO, 
(hydrogen-ion concentration = 0-024 N ) and 1275 secs. in the presence of 
0-016 N acetic acid (hydrogen-ion concentration = 0-0016 N). 

The absence of proportionality between the denaturation-velocity and 
the concentration of hydrogen-ions in the case of strongly different concen- 


‘ 


THE HEAT-COAGULATION OF PROTEINS 683 


trations of the same acid, is not unexpected. Indeed a strict proportionality 
between the hydrogen-ion concentration and the reaction-rate also does not 
exist in the case of the hydrolysis of sugar by acids, particularly when salts 
are present in the sugar solution. 

Arrhenius [1889] who investigated the influence of salts upon the inver- 
sion-velocity showed that the latter departs from its normal value almost 
proportionally to the salt-content of the solution; moreover, the weaker the 
acid concentration, the more considerable is the influence of salt; so for 
instance, the inversion-rate increased by 35-40% on adding KBr to the 
concentration of 0-4 N, when hydrobromic acid was 0-002 N, and it increased 
by 200-240 °% when the acid was 0-0005 N. 

At the same time, it was necessary, in my experiments, to add ammonium 
sulphate to the albumin solutions to bring about the coagulation of denatur- 
ated albumin. No doubt the salt increased the catalytic power of the hydro- 
gen-ions normally contained in the albumin solution, the presence .of which 
is due to the electrolytic dissociation of the protein molecules (hydrogen-ion 
concentration of about 0-00001 N); then, the salt induced thereby an extra- 
ordinarily strong increase of the denaturation-rate!. However, this increasing 
effect of salt upon the catalytic power of hydrogen-ions strongly diminished 
on the further increase of the concentration of these ions by adding nitric 
acid so that the denaturation-rate increased only by 1-6 fold, while the 
hydrogen-ion concentration increased by 15 fold. By virtue of the same cause 
the denaturation-velocity continued to increase more slowly than the con- 
centration of hydrogen-ions on the addition of acid. When, however, this 
concentration became more than 0-002 N the denaturation-rate began to 
increase more rapidly than the hydrogen-ion concentration (an increase of 
the latter of 8-3 fold was accompanied by an increase of the former of 28 fold). 
The same was found by Chick and Martin [1912]. 

The more rapid increase of the denaturation-rate by stronger concentra- 
tions of acid could not be explained except by the formation of compounds 
between albumin and acid which react more easily with water than the 
unaltered albumin, the acid concentration 0-016 N being evidently sufficient 
for the formation of such compounds. 

As the increasing effect of salts upon the catalytic power of hydrogen-ions 
varies only with a very considerable alteration of concentration of these ions 
(Arrhenius, see above), it is very comprehensible that the denaturation-rate 
remained in my experiments almost proportional to the hydrogen-ion con- 
centration in the case where denaturation proceeded in solutions containing 
similar concentrations of various acids. 

In view of the importance of hydrogen-ions in the reaction between protein 
and water, it seems to be very probable that the denaturation is a slight 
hydrolysis of protein; such hydrolysis cannot be discovered by chemical 


1 It is uncertain whether this increase of denaturation-rate is due to the formation of an 
active albumin, analogically to the active sugar of Arrhenius. 


684 W. W. LEPESCHKIN 


analysis, the protein molecule being too large, and it is not surprising that 
Chick and Martin could obtain no evidence of hydrolysis by denaturation of 
pure proteins in the chemical way. 

We go on now to consider the influence of alkali upon slavntoeaiion: 

From Chick and Martin’s experiments [1913] it follows that in the presence 
of very weak alkali concentration (hydroxyl-ion concentration equal to 
10-**§ V) the denaturation proceeds too slowly and only on addition of more 
alkali (hydroxyl-ion concentration 10-**7 V) does the denaturation-rate 
increase so as to be comparable with that in the original, slightly acid solution. 
These experiments confirm therefore the above hypothesis that the denatur- 
ation of albumin, free from electrolytes, is accelerated by its own hydrogen- 
ions present in the solution in concentration of about 10-5 N: after the 
neutralisation of albumin by a small amount of alkali the accelerating effect 
of these ions is abrogated and the denaturation-rate strongly diminished. 
The strong increase of the latter on the addition of more alkali could then be 
explained by the formation of albumin salts with alkali (alkali albuminate) 
which react more easily with water than the unaltered albumin. 

Unfortunately it was only possible for Chick and Martin to institute a 
cumpai:ton over a small portion of the curves expressing the dependence of 
the denaturation-rate upon the hydroxyl-ion concentration (namely, from a 
concentration equal to 10-*8 to 10-**7), Nevertheless it seemed to me to be 
desirable to investigate the effect of stronger concentrations of alkali upon 
the denaturation-velocity. My experiments made for this purpose showed 
‘vat an increase of the hydroxyl-ion concentration beginning from 10-?*? NV 
only influence’ the densturation-rate very little. Thus an increase of alkalinity 
of over sixty-fold was accompanied by a decrease in denaturation-rate of 
27 %. 

In my experiments a solution of serum-albumin Kahlbaum dialysed 
during 6 weeks and containing 0-49 % albumin was used. This solution was 
diluted with an equal volume of a solution of potassium hydroxide of known 
concentration and in order to keep it coagulable ammonium sulphate (con- 
centration of 0-2 N, or 0-4 N when the alkali concentration was great) was 
added to it. When the hydroxyl-ion concentrations were 0-006 N (=10-**?), 
0-012 N, 0-067 N, 0-150 N, 0-400 N, the average denaturation-times at 70° 
were found to be equal respectively to 270 secs., 285 s., 300 s., 320 s., 
370 8. 

The study of the influence of alkalies and acids upon the denaturation 
showed therefore that the above hypothesis concerning the nature of this 
reaction is true: the denaturation is a weak hydrolysis of protein. 

This hydrolysis can be accompanied also by the splitting off of any amino- 
acids containing sulphur from the protein molecule [Moll, 1904] and by an 
increase in the amount of carboxyl group, for, as will be shown below, the 
acid properties of albumin increase after the denaturation. 

Concerning the known increase of alkalinity of a solution of albumin, 


THE HEAT-COAGULATION OF PROTEINS 685 


after denaturation, Chick and Martin showed that the diminution of acidity 
during heat-coagulation takes place only in acid solutions, while in alkaline 
solutions a diminution of alkalinity is observed [Chick and Martin, 1913, 1}. 
It has already been pointed out that protein can unite with acids and alkalies 
before denaturation; the coagulation following the latter would then evidently 
remove not only protein from the solution but also a quantity of acid or 
alkali so that the acidity or the alkalinity of solution would be diminished 
during the heat-coagulation. 


(4) INFLUENCE OF SALTS UPON THE DENATURATION-VELOCITY OF PROTEIN, 


As mentioned above, salts increase the catalytic power of hydrogen-ions 
and produce thereby an increase of inversion-rate of sugar. Salts-probably 
also accelerate the denaturation of protein for they increase the catalytic 
power of hydrogen-ions whose presence in the electrolyte-free protein solution 
is due to the electrolytic dissociation of the protein molecules which are more 
acid than alkaline (see above, hydrogen-ion concentration of about 10-5 N). 
Moreover, Arrhenius found the accelerating effect of salts upon inversion to 
increase with the increasing salt concentration. On the other ha_., a8 +s 
known, salts diminish the electrolytic dissociation of electrolytes, and they 
may thus diminish that of protein; moreover the researches of Lillie [1907] 
showed that salts considerably diminish the osmotic pressure of albumin 
solutions; that is to say they lower their degree of dispersion and, therefore, 
the particle-surface of the protein; so that the reaction between protein ani* 
water may be retarded. 

We should therefore expect that an increase of salt concentration would 
accelerate the denaturation of albumin when the salt concentration is small, 
it would however diminish it when the salt concentration is great and it 
would leave it unaltered if the salt concentration is intermediate. 

My experiments showed that this suggestion is right. But, unfortunately, 
they could not render clear the effect of very small concentration of salts upon 
the denaturation-velocity, the method employed having required the pre- 
sence in the protein solution of an amount of salt sufficiently great to produce 
the coagulation of the denaturated albumin (see above). 

The material used in my experiments was egg-albumin Kahlbaum, which 
had been dialysed many days (content 0-66 °%). When the concentration of 
added ammonium sulphate was 0-03 N, 0-06 N, 0-12.N, 1-1N and 1-32 N, 
the average times of denaturation at 70° were found respectively to be 932 secs., 
185 s., 185s., 270s. and 605s. These results are in accordance with the 
results of Chick and Martin, who investigated only the concentrations from 
0-1 to 1 N and above. 

As is known, salts accelerate the inversion of sugar by acid in different 
degree, according to their lyotropic properties, which are chiefly due to the 
effect of anions: SO, < Cl< Br< SCN [Spohr, 1888]. It was therefore to be 


Bioch. xv1 45 


686 W. W. LEPESCHKIN 


expected that salts would accelerate the denaturation of acid solutions of 
protein in the same series of anions. Indeed, my experiments showed that 
the denaturation-velocity of protein in acid solutions is incomparably more 
increased by potassium thiocyanate than by potassium sulphate; potassium 
chlorate takes the middle position but it is, in its effect, nearer to potassium 
sulphate. 

The material used was egg-albumin Kahlbaum, the 2 % solution of which 
had been dialysed during 5 days. The solution obtained was diluted with 
water and 0-1 N HNOg, so that the content of albumin in all the solutions 
was 0-41 %, and the content of acid 0-01 mol. per litre. To the solutions such 
quantity of various salts was added as to render the salt-content of each 
solution equal to 0-083 mol. per litre. In the presence of KCNS, KCl, K,SO, 
the average time of denaturation at 50° was found respectively to ae 15 secs., 
3000 s. and 4650 s. 

If albumin solutions contain no free hydrogen-ions but are slightly alkaline, 
the presence of salts could not evidently accelerate the denaturation (there 
are no hydrogen-ions in the solution the hydrolytic power of which could 
be increased), and it is to be expected that the nature of the anion would have 
no influence upon the denaturation-velocity. 

Indeed, my experiments showed that the denaturation-rate in this case is 
almost equal in the solutions containing salts of various anions. A slightly 
alkaline solution of egg-albumin Kahlbaum had been dialysed until it no 
longer coagulated on heating. When 0-1 N K,SO,, KCl, KI and KOCOCH, 
were added, it coagulated at 75° respectively in 212, 220, 225 and 227 secs. 

Concerning the increasing influence of various cations upon the catalytic 
power of hydrogen-ions in the case of denaturation, it is almost the same for 
various alkali metals. The acceleration of denaturation by salts of alkaline 
earths and heavy metals is however so strong, as to make necessary the 
suggestion that the salts in question first enter into chemical unions with 
protein which are soluble in water and react with it more easily than the free 
protein, and therefore are more easily denaturated. The other suggestion 
that alkaline earths and heavy metals form insoluble compounds with native 
protein (which are therefore no longer capable of denaturation by heating), 
the formation of which is accelerated by high temperature as are all other 
chemical reactions, is less probable, for the temperature coefficients of the 
reaction were found in my experiments to be too high: 1:28-1:38 per rise of 1°. 

A solution of egg-albumin Kahlbaum, after dialysis during 6 days, was 
diluted with water and salt solutions so that all the solutions to be tested 
contained 0-3 %, albumin. In the presence of 0-1 N KCl the average time of 
denaturation was found, at 75°, to be 2095 secs., while in the presence of 
0-1 N BaCl, denaturation proceeded instantaneously at the same temperature ; 
in the presence of 0-003 N BaCl, it required 95 secs., 0-003 N MgCl, 120 secs., 
0-0001 N HCI 52 sees., while it proceeded instantaneously‘ in the presence of 
0-003 N HCl. 


“THE HEAT-COAGULATION OF PROTEINS 687 


The study of the influence of salts upon denaturation showed, therefore, 
that there was no contradiction with the foregoing hypothesis. 


(5) THE CAUSE AND THE TEMPERATURE COEFFICIENT OF THE 
COAGULATION OF DENATURATED PROTEIN. 


As mentioned above, it has long been known, that protein solutions made 
heat-incoagulable by dialysis acquire the property to coagulate on heating 
when some electrolyte is added to them. They coagulate also on the 
addition of electrolytes when preliminarily heated and cooled. The presence 
of electrolytes is therefore required to produce the second process comprised in 
heat-coagulation. In order to render this process clear it was first of all 
necessary to answer the questions, whether the coagulation of denaturated 
protein is a pure physical process or, at the same time, also a chemical reaction; 
whether this process is, in its nature, like the other processes of coagulation 
of suspension-colloids by electrolytes. 

In this connection it was interesting to estimate the temperature coefficient 
of coagulation of denaturated protein and to compare it with that of chemical 
reactions and of other coagulations. 

My experiments showed firstly that the coagulation of denaturated protein 
proceeds at a certain temperature considerably more rapidly than the de- 
naturation at the same temperature if the protein solution contains a suffi- 
ciently great amount of salt. 

Thus, in one of my experiments, the average time required to produce 
the necessary turbidity of 0-57 °% solution of dialysed egg-albumin Kahlbaum 
to which ammonium sulphate was added to the concentration of 0-05 N 
at 65° was found to be 2220 secs., while the average time required to produce 
the same turbidity of the same albumin solution which had been boiled and 
to which after cooling the same amount of ammonium sulphate was added 
(7.e. the time to produce the coagulation) was found to be 10 secs. at 65°. 

Further, in accordance with Chick and Martin, my experiments showed 
that the average temperature coefficient of coagulation of denaturated 
albumin is usually equal to 1-1 (it varies from 1-08 to 1-2) per temperature- 
rise of 1°. But sometimes it increases to 1:3. The results of four of my 
experimental series are the following: 

I. Solution of egg-albumin Kahlbaum made non-coagulable on heating 
(by dialysis) and containing 0-57 °% protein was boiled. After cooling (NH,),S0, 
was added to it to the concentration of 0-025 N. The average times of coagu- 
lation at 80°, 75°, 70°, 65°, and 60° were found to be equal respectively to 
165 secs., 405 s., 943 s., 1922 s., 2800 s. The temperature coefficient = 1-15. 

II. Solution of egg-albumin Kahlbaum. Content of albumin = 0-41 %. 
Content of ammonium sulphate added after boiling and cooling the solution 
= 0-05 N. The average times of coagulation: at 65° 10 secs., at 60° 40 secs., 
at 55° 138 secs., at 53° 251 secs., at 50° 687 secs., at 47° 1630°secs. The tempera- 
ture coefficient = 1-33. 


45—2 


688 W. W. LEPESCHKIN 


IIT. 0-14 % solution of recrystallised egg-albumin, made non-coagulable on 
heating by dialysis. Content of ammonium sulphate added after boiling and 
cooling to the solution= 0-1 N. The average times of coagulation were: 
at 45° 72 secs., at 40° 152 secs., at 35° 425 secs., at 30° 1200 secs. The average 
temperature coefficient = 1-2. 

IV. Egg-white had been dialysed till the absence of heat-coagulation, 
diluted with water fivefold, boiled and cooled. (NH,),SO, was then added to the 
concentration of 0-1 VN. The average times of coagulation were: at 60° 35 secs., 
at 50° 297 sees., at 45° 687 secs., at 40° 1325 secs. The average temperature 
coefficient = 1-18. : 

As known, the denaturated protein is regarded as a suspension-colloid 
(see above). It was therefore interesting to compare the temperature coeffi- 
cients found with those of coagulation of other suspension-colloids. But 
there are yet no investigations dealing with the dependence of coagulation 
upon temperature. Scattered up and down a number of observations only 
pointed out that colloidal solutions are generally less stable at higher tempera- 
tures [see, for example, Zsigmondy, 1912]. 

I therefore made some experiments with colloidal solutions of arsenic tri- 
sulphide and lecithin by means of the method above described, which, in the 
case of arsenic, was a little modified; the vessel for the colloidal solution was 
observed with the intact horizontal microscope (small magnification) and 
the time-intervals were noted which were required to produce a complete 
coagulation. This was marked by the appearance of orange particles of 
arsenic trisulphide, all through the liquid, on the pure white ground. 

A 0-5 % colloidal solution of arsenic trisulphide (obtained by passing 
sulphuretted hydrogen into water containing powdered As,O,), from which 
H,S had been removed by a current of air, was centrifuged until particles 
visible under the microscope were absent. Then ammonium sulphate was — 
added to the resulting colloidal solution to the concentration of 0-01 mol. per 
litre. 

The average times of coagulation were found to be at 85° 228 secs., at 
75° 401 secs., at 65° 601 sees., at 55° 1110 secs. The temperature coefficient 
= 1-05 (per rise of 1°). 

In another experimental series where 0-4 °% solution of arsenic trisulphide 
was used and the concentration of added salt (NaCl) was 0-04 mol. per litre, 
the average times of coagulation were: at 60° 19 mins., at 50° 29 mins., at 
40° 39 mins., at 30° 56 mins. and at 17-3° 86 mins. The temperature coefficient 
was therefore 1-04 (per rise of 1°). 

1 %, solution of lecithin in water was made by mixing an ethereal solution 
of lecithin with water and removing ether by blowing air through the solution. 
Twenty-four hours after the preparation ammonium sulphate was added to 
the solution (concentration 0-08 N), The average times of coagulation were 
at 90° 106 secs., at 80° 500 sees., at 70° 910 sees., at 60° 1960, secs. The average 
temperature coefficient was therefore 1:10. When the concentration of 


THE HEAT-COAGULATION OF PROTEINS 689 


lecithin was 0-5 %, the times of coagulation were: at 90° 376 secs., at 80° 
606 secs., at 70° 1356 secs., at 60° 3290 secs. The average temperature 
coefficient was therefore 1-07 (per rise of 1°). 

From the experiments cited it may be seen that the temperature coefficient 
of coagulation of denaturated albumin is greater than that of arsenic tri- 
sulphide or lecithin but it is nearer to that of lecithin, which is known to 
occupy an intermediate position between hydrophilic and hydrophobic 
colloids (resp. emulsoids and suspensoids). In the case of arsenic trisulphide 
the temperature coefficient of coagulation is like that of the diffusion of salts. 
The temperature-rise only increases the quantity of salt acting upon the 
colloidal particles. The temperature effect, in the case of lecithin coagulation, 
is however more complex. The temperature influences in this case not only 
the salt quantity acting upon the colloidal particles, but also the degree of 
dispersion and therefore the reacting surface of the colloidal particles, and 
the coefficient is greater. 

The most complex effect of temperature is that upon the coagulation of 
denaturated protein. The temperature influences probably, in this case, not 
only the dispersion of the colloid and the acting salt quantity, but also the 
rate of reaction between salt and protein. For, a formation of loose chemical 
compounds (“adsorption compounds”) of salts with denaturated protein 
during the coagulation is very probable (see below). 


(6) INFLUENCE OF ACIDS AND ALKALIES UPON THE COAGULATION 
OF DENATURATED PROTEIN. 


Denaturated proteins are known to be soluble in alkalies and in strong 
acids [Cohnheim, 1904], so that the formation of compounds of denaturated 
proteins with alkalies and acids is undoubted. From the solution of de- 
naturated albumin in weak potassium hydroxide the protein can be precipi- 
tated by ammonium sulphate. The precipitate is however easily soluble in a 
stronger solution of this alkali, but a greater quantity of salt precipitates 
protein anew. This precipitate is also soluble in more alkali, etc. The increase 
of alkali concentration produces therefore a continuous change of colloidal 
properties of denaturated albumin owing evidently to the formation of alkali 
compounds of protein, and the greater the alkali concentration of the solution 
the more alkali these compounds contain. 

We should therefore expect that alkalies and acids would have a consider- 
able influence upon the coagulation-velocity of denaturated albumin. My 
experiments showed, indeed, that, generally, acid strongly increases and 
alkali strongly diminishes the coagulation-rate of denaturated albumin. 
Moreover 0-00005 N HNO, increases this rate more strongly than 0-0001 and 
even 0-0002 N acid so that the increasing effect cannot be explained by the 
influence of hydrogen-ions; it is evidently due to the formation of acid com- 
pounds of denaturated protein the colloidal properties of which change with 


690 W. W. LEPESCHKIN 


the amount of acid contained in them. The same result was also found for 
the alkali effect. The decrease of coagulation-velocity is in this case not at 
all proportional to the hydroxyl-ion concentration: 0-001 N hydroxyl-ion 
concentration, in the case of ammonia, produces a similar diminution of 
velocity as 0-011 N hydroxyl-ion concentration in the case of potassium 
hydroxide. Conversely the like molecular concentrations of ammonia and 
potassium hydroxide bring about a like decrease of coagulation-velocity so 
that the decreasing effect of alkali cannot be explained except by the forma- 
tion of alkali compounds of denaturated albumin having other colloidal 
properties than the free albumin. ; 

The experiments were made with egg-albumin Kahlbaum, the solutions 
of which had been dialysed during many days and no longer coagulated on 
heating. 

I. A solution of albumin (containing 2-07 % protein) was boiled and, 
after cooling, diluted fivefold partly with water, partly with nitric acid. To 
the solutions obtained potassium sulphate was added to the concentration of 
0-083 mol. per litre. The average time of coagulation was found at 75°, in 
the absence of acid, to be 2700 secs., whilst it was found with 0-00005 N acid 
to be 102 secs., with 0-00010 WN acid 236 secs., 0-00020 N acid 178 secs. But 
in the presence of 0-01 WN acid the coagulation proceeded instantaneously. 

II. A solution of albumin (containing 3-2 % protein) was boiled and, 
after cooling, diluted fivefold partly with water, partly with solutions of 
potassium hydroxide or of ammonia of certain concentrations. Ammonium 
sulphate was added to the concentration of 0-06 N. The coagulation pro- 
ceeded at 70°, in the absence of alkali, instantaneously, while the average 
time of coagulation, with 0-0018 N ammonia (H-ion concentration = 0-00031 
N), was 370 secs., with 0-018 N ammonia (H-ion concentration = 0-00100 J), 
760 secs., of 0-2 N ammonia (H-ion concentration = 0-00400 NV), 2200 sees. 
In the presence of 0-018 N KOH (H-ion concentration = 0-01100 NV) it was 
770 secs. 

My experiments showed further, that the temperature coefficient of 
coagulation is in the presence of acid still less than that of coagulation of 
denaturated albumin in the presence of alkali and even than that of coagula- 
tion of arsenic trisulphide. So in a series of experiments, in which dialysed 
egg-albumin Kahlbaum in the presence of 0-0001 mol. nitric acid per litre 
was tested, the coagulation time was found to be equal to 237 secs. at 75°, 
to 400 secs. at 65°, to 720 secs. at 55°. So that the average temperature 
coefficient was 1-05. In another series, in which the acid concentration was 
00002 mol. per litre, the coagulation time was: 760 secs. at 75°, 1597 sees. at 
55° and 1720 sees. at 50°. The average temperature coefficient of coagulation 
was, therefore, 1-02. In the presence of 0-018 mol. ammonia per litre this 
coefficient was found to be 1-12 and in the presence of 0-18 mol. ammonia per 
litre 1-19. 


We might thus suggest, that the colloidal state of acid compounds of 


THE HEAT-COAGULATION OF PROTEINS 691 


denaturated albumin is nearer to that of typical suspension colloids, while 
that of alkali compounds is nearer to the state of emulsion colloids. The free 
denaturated albumin seems to occupy an intermediate position between the 
alkali compounds and the acid compounds. 

Finally, my experiments showed that the acid compounds of denaturated 
albumin are not identical with the denaturated acid compounds of native 
albumin, formed in the presence of the same concentration of acid. Its 
colloidal state is different. 

One portion of solution of dialysed egg-albumin containing 2-07 °% protein 
and not coagulating on heating was diluted two and a half-fold with water 
and boiled up. After cooling, to this protein solution a solution of potassium 
sulphate and diluted nitric acid were added, altogether in such quantities 
that the original albumin solution was diluted fivefold. The concentration 
of acid was 0-0002 mol. per litre, that of potassium sulphate 0-083 mol. The 
average time of coagulation at 65° was found to be 240 secs., at 55° 365 secs. 

The second portion of the same solution of egg-albumin was directly 
diluted fivefold with diluted nitric acid so that the acid concentration was 
also 0-0002 mol. per litre. Subsequently the solution was boiled up and cooled; 
afterwards potassium sulphate was added to the concentration of 0-083 mol. 
per litre. The average time of coagulation at 65° was found to be 320 secs., 
at 55° 660 secs. 


(7) INFLUENCE OF THE DEGREE OF DISPERSION OF DENATURATED ALBUMIN 
UPON ITS COAGULATION-VELOCITY. 


My experiments relating to the comparison of the time required to produce 
a certain degree of denaturation of an albumin solution with the time required 
to bring about the coagulation of the same albumin denaturated by boiling 
led to an interesting observation which is in accordance with the fact cited 
above (found by Lillie) that salts lower the degree of dispersion of albumin 
solutions. 

My experiments showed, namely, that the appearance of the definite 
turbidity of a 0-57 % solution of dialysed egg-albumin Kahlbaum, containing 
0-1 mol. potassium chlorate per litre, required on an average 220 secs. at 75°, 
and 1630 secs. at 71°. However, the same turbidity appeared in 1380 secs. 
at 75° and in 2020 sees. at 71°, when before adding potassium chlorate the 
same albumin solution had been boiled and cooled. In both cases the turbidity 
marked the coagulation of denaturated albumin, but in the first case this 
coagulation, following the denaturation, proceeded in a lapse of 220 and 
1630 secs. (or still more rapidly), while in the second case this coagulation 
required 1380 and 2020 secs. 

It is evident that the colloidal properties of denaturated albumin formed 
in the solution containing 0-1 mol. potassium chlorate were unlike those of 
denaturated albumin formed in the solution containing no potassium chlorate, 
for the former albumin required less time to be coagulated than the latter. 


692 W. W. LEPESCHKIN 


This phenomenon can only be explained by a greater degree of dispersion of 
the denaturated albumin formed in absence of potassium chlorate. Lillie 
showed that the degree of dispersion of albumin in the presence of salts 
is smaller than in the absence of it, even before the denaturation takes place. 
The denaturation of an albumin of a smaller degree of dispersion leads there- 
fore to the formation of a denaturated albumin which also has a smaller 
degree of dispersion. 

In agreement with the results obtained by Lillie, my experiments showed 
that the presence of potassium chlorate even in concentration of 0-01 mol. 
per litre affects the colloidal properties of albumin. So in one of my experi- 
ments a | % solution of dialysed egg-albumin Kahlbaum (not coagulating on 
heating) was boiled and, after cooling, diluted with an equal volume of 0-2 N 
KCl. This albumin solution coagulated at 75° on an average in 1380 secs. 
On the other hand, to the same solution of albumin Kahlbaum KCl was first 
added to the concentration of 0-01 mol. per litre and the solution was then 
boiled and after cooling diluted with an equal volume of 0-19 N KCl. In this 
case the protein solution coagulated at 75° on an average in 225 secs., though 
the concentration of KCl was in both cases the same, namely 0-1 mol. per litre. 

The coagulation-velocity of denaturated albumin formed in the presence 
of potassium chloride was found therefore to be almost sixfold as great as 
the coagulation-velocity of denaturated albumin formed in the absence of 
salt. The result obtained corresponds precisely with the value of diminution 


of osmotic pressure of albumin solution by potassium chloride found by 
Lillie. 


(8) INFLUENCE OF SALTS UPON THE COAGULATION 
OF DENATURATED ALBUMIN. 


In the preceding part of this paper it has several times been pointed out that 
protein solutions from which salts have been removed by dialysis do not 
coagulate on heating, but they become coagulable when salts are added to 
them. Conversely Pauli and his pupil Handovsky affirm that after a very 
long dialysis the serum (viz. serum-albumin) acquires the property to 
coagulate on heating without adding salts. Salts added to such serum 
bring about a rise of its ‘coagulation temperature,’ and thiocyanates and 
iodides make it even heat-uncoagulable [Pauli, 1908; Pauli and Handovsky, 
1908, 1909}. 

These authors call the protein contained in their serum “amphoteric 
protein,” pointing out that it conducts the electrical current not more than 
distilled water and shows no cataphoresis. It is however a pity that the 
authors did not define the salt content of their protein, as for instance, by 
determining the quantity of ash, for salts might not have been dissolved 
in the water of the serum but adsorbed on the protein particles, Then after 
the denaturation of albumin they might be liberated anew and produce the 
coagulation. 


THE HEAT-COAGULATION OF PROTEINS 693 


At all events, it is impossible to regard the small salt content of ‘‘ampho- 
teric protein” of Pauli, as the cause of the possibility of its heat-coagulation, 
and to suppose that the absence of heat-coagulation of dialysed protein 
solutions which has been observed by all other authors is due to the incom- 
pleteness of the dialysis. Indeed, the addition of salts to the “amphoteric 
protein” could cause only a rise of coagulation temperature, and only thio- 
eyanates and iodides (normally absent in natural liquids) made, in the 
experiments of Pauli, the heat-coagulation of serum impossible. 

In order to render the influence of salts upon the coagulation of denaturated 
protein clear, I made experiments with egg-albumin and serum-albumin 
Kahlbaum. We consider first the influence of salts upon the coagulation of 
egg-albumin. 

The solutions of egg-albumin (5-10 %) were usually centrifuged and 
filtered. They had a slight alkaline reaction (litmus) and were almost water- 
clear. After drying a part of the solutions in vacuum and afterwards in the 
drying-stove the dry albumin was found to contain about 3 % ash, of which 
the most part (2°) was water-soluble (KCl, NaCl, K,S0,, K,CO,). The 
solutions coagulated on heating. 

The dialysis was executed by means of the thinnest parchment paper 
which could be purchased. If the dialysis was prolonged an antiseptic was 
added to the solutions tested and the dialyser was placed in a covered vessel. . 
The solution layer in the dialyser never exceeded } cm. in depth. 

After 24 hours of dialysis (distilled water was changed each two hours) 
the protein solutions usually became neutral (litmus) in reaction and then lost 
_ the property of coagulating on heating (if the protein-content was relatively 
great, e.g. 10 %) or they lost this property only after two or three days of 
dialysis (if the protein-content was small). 

The dialysed solutions were usually centrifuged and filtered anew till they 
became water-clear. The content of protein and of mineral substances in the 
solutions (which did not coagulate on heating) is set out below. 


Content of 
Content soluble mineral 
of mineral substances in 
Number Original Concentration substances in g. per 100 ce. 
of tested concentration Duration of albumin dry albumin of solution 
solutions of albumin _ of dialysis after dialysis after dialysis after dialysis 
%o _ days % % 
1 10 1 75 1-8 0-05 
2 4-9 2 3-29 i! 0-02 
3 41 5 2-07 1-2 0-01 


. In order to answer the question whether solutions of egg-albumin which 
no longer coagulate on heating would acquire the property to coagulate 
anew by a still more prolonged dialysis, solutions 1 and 3 were employed for 
a further dialysis. 

To solution No. 1 an excess of chloroform was added as an antiseptic. 
After 10 days of dialysis at 20° this solution showed heat-coagulation anew. 


694 W. W. LEPESCHKIN 


But it acquired an unpleasant scent and, in spite of the chloroform, contained 
many bacteria. The heat-coagulation was evidently due in this case to 
organic acids or salts which had been formed by the bacteria. Indeed, after 
a dilution of the solution with water the heat-coagulation disappeared anew, 
but it reappeared if some ammonium sulphate was added to it. 

To solution No. 3 an emulsion of melted thymol in water was added 
as an antiseptic. The dialysis lasted three weeks. The solution contained 
after such dialysis 0-99 % albumin. The dry albumin (drying in vacuum) 
was found to contain 0-8 °%% ash of which about a half was soluble in water. 
The solution contained 0-008 % dissolved mineral substances and coagulated 
on heating. An unpleasant scent and bacteria were however absent. 

The albumin solution thus obtained showed the required degree of coagu- 
lation (turbidity like that of the ground glass, see above) at 75° in 27 secs. After 
dilution of the solution with water to double the volume, the same degree of 
coagulation was observed at 75° in 700 secs., at 80° in 370 secs., at 70° in 
1260 secs. The striking increase of coagulation-time after a relatively incon- 
siderable dilution with water indicates that the very small amount of salts 
contained in the albumin, which was sufficient to bring about the coagulation 
of protein in 27 secs. at 75°, after the dilution with water, could no longer 
produce the coagulation in this time. At the same time, the small temperature 
coefficient—1-12 and 1-14 per temperature-rise of 1°—showed that the 
coagulation-velocity, after the dilution with water, was smaller than the 
denaturation-velocity. 

When this albumin solution was diluted fivefold with water, it ceased to 
coagulate even on boiling, whereas it coagulated easily if some ammonium 
sulphate was first added to it. Similarly it coagulated if after boiling and 
cooling this salt was added to it. 

It seemed, therefore, that the heat-coagulation of denaturated egg-albumin 
has two optimal concentrations of salts present in the solution. Namely the 
coagulation on heating takes place when this concentration is more than 
0-1 % (viz. about 0-01 NV), it ceases to be observed when this concentration 
is between 0-01 and 0-05 % (viz. about 0-001 and 0-006 NV), it appears anew 
when this concentration diminishes to 0-008 % (viz. about 0-0008 NV), and, 
finally, it ceases again to appear when the salt-concentration becomes about 
0-001 % (viz. about 0-0001 N). 

Similarly, two most favourable salt-concentrations, viz. two coagulation- 
zones, were observed also on the other colloids, as for example, on lecithin 
and cholesterol [Porges and Neubauer, 1908]. Nevertheless, my experiments 
could not confirm the existence of any coagulation-zones in the case of heat- 
coagulation of protein. 

They showed that a gradual increase of salt-content produces, instead of 
a retardation or a cessation of heat-coagulation, an acceleration of this 
coagulation of protein. In my experiments, the above solution of albumin 
containing 0-008 °, dissolved mineral substances was diluted with an equal 


THE HEAT-COAGULATION OF PROTEINS 695 


volume of solution of potassium chlorate of various concentrations and the 
heat-coagulation time was determined by the previous method. 

When potassium chlorate was substituted by potassium thiocyanate, the 
phenomenon remained the same, and only very strong concentrations of this 
salt (2-5 mol. per litre) brought about a diminution of coagulation-velocity. 
A suppression of heat-coagulation by adding KCNS could never be observed. 
At a concentration of KCl of 0-003 N the temperature coefficient was found 
to be equal to 1-17; and only at the concentration of 0-01 N did this coefficient 
attain the value of the normal denaturation coefficients 1-35-1-40, and the 
denaturation began to proceed more slowly than the coagulation of the 
denaturated albumin. The same was also observed, when a sufficient amount 
of KCNS (0-1 mol. per litre) was added to the solution. On the other hand 
an excess of this salt (2-5 mol. per litre) retards the coagulation anew and the 
denaturation begins to proceed more rapidly than the coagulation (the 
temperature coefficient = 1-04). 

The albumin contained in the solution which had been dialysed during 
four weeks is therefore not identical with that contained in the solution 
which had been dialysed during 5-7 days. After denaturation the former 
albumin coagulates more easily under the influence of salts than the latter. 
The degree of dispersion of the former is probably smaller than that of the 
latter. At all events the very prolonged dialysis evidently alters the original 
albumin. Both albumins are however “amphoteric” (see Pauli), for neither the 
one nor the other showed in my experiments cataphoresis by an electrical 
current of 110 volts; but both migrated to the anode after denaturation 
(according to the increase of alkalinity of the solution). 

We pass now to the experiments with serum-albumin. 

A 5 % solution of serum-albumin Kahlbaum was centrifuged and filtered. 
The slightly alkaline solution obtained was very faintly opalescent, coagu- 
lated on heating and contained 3-07 % dry residue of which 0-25 % was 
mineral matter. 

The dry protein therefore contained 8-4 °, mineral substances. The 
soluble part of the latter consisted of sodium chloride, potassium chloride, 
sodium carbonate, etc., the insoluble part was also insoluble in acids 
(Si0,?). 

The albumin solution obtained was dialysed during 20 hours. After the 
dialysis it was centrifuged and filtered anew. It reacted neutral (litmus) 
and no longer coagulated on heating, but only became opalescent. It now 
contained 2-4 °% dry residue, of which 0-07 % was mineral and 0-05 % 
water-soluble. The dry albumin contained now about 3% mineral sub- 
stances. ) 

After adding salts to this albumin solution heat-coagulation could again 
be produced on heating. 

The high temperature coefficient of heat-coagulation (= 1-55) after adding 
potassium chloride showed that 0-1 mol. per litre of this salt was already 


696 W. W. LEPESCHKIN 


sufficient to make the coagulation of the denaturated albumin proceed more 
swiftly than the denaturation. 

In fact, heating the solution of dialysed serum-albumin after a preliminary 
boiling and cooling and the subsequent addition of potassium chloride to a 
concentration of 0-1 mol. per litre, showed in my. experiments that the 
coagulation of the denaturated albumin at 70° required on the average only 
35 secs., at 68° 50 secs., at 65° 90 secs., etc. On the other hand the same 
concentration of potassium thiocyanate only brought about a coagulation 
which proceeded slower than the denaturation, and the temperature coefficient 
was found in this case to be low. Only between 65° and 70° did the denatura- 
tion-velocity begin to be greater than the coagulation-velocity and the 
temperature coefficient increased, while the coagulation-time became equal to 
that observed in the presence of potassium chloride. 

Further dialysis renders the solution of serum-albumin heat-coagulable 
again, and even after 6 weeks of dialysis the solution coagulated on heating. 
At the same time even after 5 days of dialysis the filtered solution contained 
1-28 % dissolved solids, of which 0-02 °%% was mineral; the dry albumin there- 
fore now contained 1-4 % mineral substances (half of which was soluble in 
water). Further dialysis however altered this salt-content of the albumin 
solution very little. 

Similarly to the solutions of egg-albumin those of serum-albumin, when 
dialysed during a long time, had a low temperature coefficient of heat- 
coagulation, the denaturation proceeding evidently more rapidly than the 
coagulation of the denaturated albumin. The heat-coagulation at 70° 
required in my experiments on the average 840 secs.’and at 80° 510 secs. ; 
the temperature coefficient per rise of 1° was thus found to be equal to 
1-05. 

Further, a very small concentration of salts (KCl or KCNS) was found in 
my experiments to accelerate the coagulation of denaturated serum-albumin, 
and the increase of salt concentration accelerated it so strongly that its 
velocity became greater than the denaturation-velocity, and the temperature 
coefficient was found high again and like that observed for solutions of serum- 
albumin which had not been dialysed (1-69). 

The results obtained in my experiments on serum-albumin are therefore 
like those obtained on egg-albumin. The difference is only quantitative. 
The solution of serum-albumin loses its mineral substances considerably 
more readily by dialysis than the solution of egg-albumin. After two 
days of dialysis the solution of serum-albumin has lost the greater part of its 
mineral matter; and after 5 days only one-twelfth of this remains in the 
solution. 

Further, the disappearance of heat-coagulation of serum-albumin is 
observed when the salt-content of the solution is between 0-03 and 0-07 %, 
while the absence of heat-coagulation of egg-albumin takes place when this 
content is between 0-01 and 0-05 %. On the other hand the addition of salts 


THE HEAT-COAGULATION OF PROTEINS 697 


(even of potassium thiocyanate) to the solutions of both albumins, made 
coagulable on heating by a very prolonged dialysis, cannot render them 
incoagulable on heating anew. 

We have thus to conclude that in both cases albumin is altered by a very 
prolonged dialysis in such a manner that, after denaturation it shows a 
greater susceptibility to salts than before. In what however this alteration 
consists, ‘is still unknown. It is probable that the degree of dispersion of both 
albumins became smaller by lapse of time, or that a small part of the salts, 
contained in albumin solutions, are chemically united with the native albumin 
and the prolonged dialysis decomposes such salt compounds. Free albumin, 
after denaturation, is in this latter case more susceptible to salts than its 
unions with salts. 

It was formerly pointed out that Pauli and Handovsky affirm that they 
have found the addition of salts to ““amphoteric albumin” to bring about an 
increase of its “coagulation-point.” This observation contradicts the results 
of my experiments which showed, as mentioned, an increasing effect of salts 
upon the coagulation-velocity of albumins. But nevertheless, according to 
Pauli and Handovsky, my experiments showed that potassium thiocyanate 
and iodide in strong concentration (more than 1 mol. per litre) diminish the 
coagulation-velocity of denaturated albumin and in the case of serum-albumin 
can even make it incoagulable on heating when they are present in almost 
saturated solution. This phenomenon can be explained by the formation of 
chemical compounds of these salts with denaturated serum-albumin. This 
albumin is still soluble in a strong and boiling solution of potassium thio- 
cyanate [Pauli and Handovsky, 1908]?. 

Moreover, the results of my experiments concerning the influence of 
dilution with water upon the coagulation-velocity of denaturated albumin 
and particularly upon that of denaturated serum-albumin also contradict 
the result obtained by Pauli and Handovsky. These authors point out that 
an increasing concentration of protein produces a considerable rise of the 
coagulation-point. My experiments show however that dilution of heat- 
coagulable albumin solutions which had been very long dialysed diminishes 
the coagulation-velocity and a sufficiently strong dilution makes it so small 
as to render it inaccessible to observation. 

What is the cause of all these contradictions? Could it not lie in that 
conjuncture which had taken place in the experiments of Pauli and Handovsky, 
who, before determining the coagulation-point of serum dialysed 6 weeks, 
left it to settle during 3-5 months [1908, pp. 416-8]? It is possible that 
during such a long time the serum-albumin was altered under the influence 
of water and the oxygen of the air. In one of my experiments, at least, a 
solution of serum-albumin, having been dialysed during 6 weeks and coagu- 


1 The degree of dispersion of compounds of albumin with thiocyanates and iodides is prob- 
ably only greater than that of free denaturated albumin, a sufficient amount of ammonium 
sulphate making them coagulable anew. 


698 W. W. LEPESCHKIN 


lating on heating lost this property and became slightly alkaline after standing 
during one year (thymol as antiseptic). 

We now pass to the experiments on the influence of salts upon the coagu- 
lation-velocity of denaturated albumin dialysed during some days. In these 
experiments albumin solutions were dialysed until they no longer coagulated 
on heating without the addition of salts. 

We have seen above that increase of salt concentration accelerated the 
coagulation of denaturated albumin which had been very thoroughly dialysed 
and coagulated on heating. By virtue of low temperature coefficients we 
have, indirectly, concluded that in-this case of heat-coagulation we have 
been dealing with the coagulation of denaturated albumin. To investigate the - 
influence of salt concentration upon the coagulation of denaturated albumin 
not coagulating on heating, without the addition of salts, is evidently much 
simpler. The solutions of dialysed albumin were in my experiment first 
boiled up and, after cooling, mixed with salt solutions of a certain concen- 
tration; the albumin solutions obtained were then tested at certain tempera- 
tures in the thermostat. I will cite here an example of my experiments. 

A solution of egg-albumin Kahlbaum containing 2-85 % protein and not 
coagulating on heating (owing to dialysis during 5 days) was boiled up, and, 
after cooling, diluted fivefold with solutions of ammonium sulphate of various 
concentrations so that the albumin solutions obtained contained 0-025, 0-05,. 
0-1 and 0-15 mol. ammonium sulphate per litre. At 47° the times required to 
produce the standard turbidity were found to be respectively equal (on an 
average) to 10,600 secs., 1620 s., 40 s. and 17 s. 

The experiment cited shows that the dependence of the coagulation-rate 
upon the salt concentration of the solution is not like the dependence of the 
adsorption of salt upon the salt concentration. The coagulation is therefore 
not a simple adsorption of electrolytes whose ions electrically discharge the 
colloidal protein particles, 1 

According to the well-known adsorption-isotherm = = acn, the adsorbed 
quantity of dissolved substance increases more slowly than the concentration 
of this substance in the solution. In my experiment however the increase of 
the salt concentration of twofold was accompanied by an increase of coagula- 
tion-velocity of ten- or fourteen-fold. 

As mentioned above Hardy showed that the denaturated protein can be 
precipitated by adjusting the reaction so as to render the particles iso-electric 
with the solution; but when the protein is made negative the valency of the 
cation is of importance for its precipitation. It has already been pointed out 
that the colloidal particles in the solution of egg-albumin Kahlbaum, which 
has been dialysed for some days (4-6 days), are iso-electric with the solution 
(no cataphoresis was observed) and that they become negative after denatur- 
ation. According to Hardy we should therefore expect that the cation only 
would be important in the coagulation of denaturated protein of the albumin 
solution in question. My experiment showed however that not only cations 


THE HEAT-COAGULATION OF PROTEINS 699 


but also anions are of importance in the process. The results of experiments 
are set out below. 

I. The material used was egg-albumin Kahlbaum, a solution of which 
had been dialysed for 6 days (albumin-content 2-8 %). It was boiled and 
after cooling diluted with solutions of salts (fivefold). The average times of 
coagulation at 75° were found to be 2100 secs. in the presence of 0-1 N KCl, 
1450 secs. at 0-1 N NH,Cl, 1180 sees. at 0-1 N NaCl, 1057 secs. at 0-1 N LiCl, 
1287 secs. at 0-003 N BaCl,, 1320 secs. at 0-003 N MgCl,, 1080 secs. at 
0-1 NV FeCl,, 2540 secs. at 0-1 N AICI,. In the presence of 0-02-0-1 N BaCl, 
or MgCl, the coagulation proceeded instantaneously. In the presence of 
- 0-003 N FeCl, very slowly. 

II. The same solution of egg-albumin. The average times of coagulation 
at 75° were found to be 215 secs. in the presence of 0-1 N K,SO,, 826 secs. at 
0-1 N potassium tartrate, 1070 secs. at 0-1 N KNO,, 1363 at 0-1 N KCl, 
1406 secs. at 0-1 N KBr, 1826 secs. at 0-1 N KI and 3826 secs. at 0-1 N 
KCNS. 

Therefore, the cations act upon the coagulation-velocity of denaturated 
albumin not at all proportionally to their valency; the tervalent cations 
(Al, Fe) produce an acceleration of this velocity almost equal to that produced 
by the univalent ions. The bivalent cations (Ba, Mg) are however strikingly 
active, producing in the concentration of 1/33 of the univalent ions an almost 
equal acceleration of coagulation-velocity. In the single series of cations of 
equal valency the accelerating effects are as follows: Li> Na>NH,Cl>K; 
Ba > Mg and Fe > Al. In all cases the temperature coefficients are relatively 
great (1-1), and then the suggestion emerges that the different effects of cations 
are due to some chemical influence produced by these cations. The latter 
could, for example, form adsorption compounds with denaturated protein. 
The yellow colour of the precipitate brought about by ferric chloride directly 
confirms this suggestion. At all events, the observed effect of cations shows 
that the coagulation of denaturated albumin cannot simply be regarded as a 
process of electrical discharge of protein particles by the ions. The same results 
also from the observed effect of various anions. 

In spite of the slight negative charge carried by the particles of denaturated 
albumin the anions employed can be placed in the following series: 


SO, > tartr. > NO, > Cl > Br > I > Thiocyanate. 


Both the series of ions (that of univalent cations and that of anions) 
are the well-known lyotropic series of ions which were found for all precipi- 
tations of emulsion-colloids. 

The observed difference in effect of single anions becomes still sharper 
when the solution of denaturated albumin is made more alkaline by adding 
potassium hydroxide. In this case, as already mentioned, protein certainly 
forms compounds with alkali, which are definitely nearer to the emulsion- 
colloids than the free denaturated albumin (see above). In my experiments 


700 | W. W. LEPESCHKIN 


the coagulation of denaturated albumin proceeded at 75° instantaneously 
when the solution contained 0-001 mol. per litre potassium hydroxide and 
0-5-1 mol. K,SO,; whilst the average time of coagulation was 180 secs. in 
the presence of the same quantity of alkali and 1 mol. KCl, it was 1220 secs. 
with 1 mol. KNO,. In the presence of 1 mol. of KI or KCNS the coagulation 
did not appear at all. KCNS produces a coagulation only when it saturates 
the solution. 

As already mentioned, dialysed albumin shows, after denaturation, a 
cataphoresis to the anode, so that it is probable that the reaction of the 
solution becomes in this case slightly alkaline (see above, Hardy). But it 
was pointed out that even very slight concentrations of alkali can cause the 
formation of compounds with denaturated protein. It is therefore possible 
that the lyotropic series of ions found is observed only when the coagulation 
of alkali compounds of protein is investigated. In order to study the influence 
of ions upon the coagulation of alkali-free albumin, I added to the albumin 
solution, after denaturation, nitric acid to the concentration of 0-0002 mol. 
per litre. This concentration is not sufficient for the formation of acid-com- 
pounds of albumin (see above). Nevertheless the lyotropic series of anions 
was perceptible even in this case. 

If in my experiments the acid concentration increased to 0-004 mol. per 
litre and above, the denaturated albumin formed compounds with acid (see 
above) and the lyotropic series of anions became indefinite, but the valency of 
the anions came to the foreground. Potassium sulphate, for instance, was 
found to accelerate the coagulation considerably more strongly than potassium _ 
chloride; on the other hand, according to its lyotropic property (the series is 
in this case inverse) potassium thiocyanate had a still greater effect upon the 
coagulation than potassium sulphate. The results obtained by Hardy were 
confirmed only in part. 

The study of the influence of salts upon the coagulation-velocity showed 
generally that the process of coagulation of denaturated albumin is not simply 
a pure physical phenomenon of discharge of colloidal protein particles owing 
to an adsorption of electrolytes (ions), but that it is, at least partly, a chemical 
phenomenon, in which not only electrical properties but also chemical pro- 
perties of salts are significant. 


REFERENCES, 


Aronstein (1874). Pfliiger’s Arch. 8, 75. 
Arrhenius (1889). Zeitech. physikal. Chem. 4, 239. 
Buglia (1909). Kolloidzeitsch. 5, 291. 

Chick and Martin (1910). J. Physiol. 40, 413. 
—— (1912). J. Physiol. 48, 2. 

——ww me (1913, 1). J. Physiol. 45, 61. 

———- ee (1913, 2). J. Physiol. 45, 288, 
Cohnheim (1904). Chemie der Eiweisskérper, 132. 


THE HEAT-COAGULATION OF PROTEINS 701 


Hardy (1899). J. Physiol. 24, 158. 
(1900). Zeitsch. physikal. Chem. 33, 385. 
—— (1906). J. Physiol. 30, 251. 
Heynsius (1874). Pfliiger’s Arch. 9, 514. 
Hoffmann (1889). Zentr. klin. Med. 793. 
(1890). Zentr. klin. Med. 521. 
Kieseritzky (1882). Die Gerinnung d. Faserstoff. etc., Dissertation, Dorpat. 
Lillie (1907). Amer. J. Physiol. 20, 127. 
Michaelis. Phys. Chemie der Kolloide, Richter-Koranyi’s Handb. 2, 391 
Moll (1904). Beitrdge, 4, 563. 
Pauli (1899). Pfliiger’s Arch. 78, 315. 
(1908). Kolloidzeitsch. 3, 2. 
Pauli and Handovsky (1908). Beitrdge, 11, 415. 
(1909). Biochem. Zeitsch. 18, 340. 
Porges and Neubauer (1908). Biochem. Zeitsch. '7, 154. 
Rosenberg (1883). Vergleichende Untersuchungen betr. d. Alkalialbuminate, etc., Dissertation, 
Dorpat. 
Spohr (1888). Zeitsch. physikal. Chem. 4, 237. 
Zsigmondy (1912). Kolloidchemie, 66, 


Bioch. xvi 46 


LXXI. THE SYNTHESIS OF GLYCINE FROM 
FORMALDEHYDE. — 


By ARTHUR ROBERT LING anp DINSHAW RATTONJI NANJI. 


From the Department of Biochemistry of Fermentation, 
University of Birmingham. 


(Received July 25th, 1922.) 


For the purpose of some work on which we are engaged, it became necessary 
to secure a considerable quantity of glycine. The published methods for the 
preparation of this compound are unsatisfactory and difficult to carry out. 
Glycine was first obtained by Braconnot [1820] by the hydrolysis of gelatin 
with dilute sulphuric acid or baryta. Perkin and Duppa [1858] prepared it 
by treating bromoacetic acid with ammonia. It has also been prepared by 
treating chloroacetic acid with ammonia or ammonium carbonate. This 
method of preparation is dealt with in subsequent papers by Heintz [1862], 
Nencki [1883], Mauthner and Suida [1888, 1890]. When prepared by either 
of these methods, the glycine must be isolated by means of the copper salt, 
and the yield never exceeds 20 % of the theoretical. 

Gabriel and Kroseberg [1889] obtained an almost theoretical yield of 
glycine by hydrolysing ethylphthalylglycine with hydrochloric acid. This 
method from an economic standpoint is unsuitable for the preparation of 
large quantities of glycine. ; 

Eschweiler [1894] states that he obtained glycine in almost theoretical 
yield by treating methylene cyanohydrin with a large excess of ammonia. 
Here the difficulty is the preparation of the cyanohydrin. 

Attempts have been made by us to devise a direct method for the synthesis 
of glycine from formaldehyde. Methylene-aminoacetonitrile can be obtained 
in a yield of 60 % of the theoretical by the condensation of formaldehyde 
(2 mols.) with ammonium cyanide as shown by Klages [1903], thus: 


2H . CHO + NH,CN = CH, : N. CH, . CN + 2H,0. 


Klages’ method is carried out as follows. Finely powdered ammonium 
chloride (360 g.) is added to 40 % formaldehyde (1000 g.), in a wide-necked 
glass jar, cooled to 5° in a freezing mixture, the solution being stirred by means 
of a strong electric turbine. Potassium cyanide (440 g.), dissolved in water 
(600 cc.) is slowly run in during 3 hours, the temperature being kept during 
this time below 10°. It may be mentioned that commercial 96 % cyanide, 
which consists of a mixture of potassium and sodium cyanide, may be employed. 
When half the cyanide solution has been added, the ammonium chloride will 
have completely dissolved. The remainder of the cyanide and at the same 
time glacial acetic acid (250 ce.) is then dropped in, When the whole of the 


THE SYNTHESIS OF GLYCINE FROM FORMALDEHYDE 703 


acetic acid has been added methylene-aminoacetonitrile commences to separate 
in glistening white crystalline flocks. The solution is now stirred with a 
turbine for 2 hours and the crystalline mass collected on a Buchner funnel. 
After washing with cold water and drying on a porous plate, the yield is 
about 280 g. or 60 % of the theoretical. It melts at 129°. 

When the nitrile is hydrolysed with hydrochloric acid, the plan adopted 
by Klages, considerable difficulty is encountered in separating the glycine 
from the ammonium chloride produced simultaneously. It was found, how- 
ever, that the nitrile could be easily hydrolysed by boiling it with a concen- 
trated (40 %) solution of barium hydroxide. 

The nitrile (10 g.) is added in small portions at a time to a boiling 40 % 
solution of barium hydroxide (100 cc.) in an open beaker. The mixture is 
boiled until no more ammonia is evolved, the volume of the liquid being 
kept approximately constant by the addition of water from time to time. The 
total period of boiling is 3 hours. From the solution, which now contains the 
methylene derivative of glycine, the barium is precipitated with sulphuric 
~ acid, the filtrate and washings from the barium sulphate are acidified 
until 3 % of sulphuric acid is present, and the liquid is boiled in an open 
beaker until no more formaldehyde is given off, which usually requires about 
4 hours. The reaction proceeds as follows: 


CH,(NCH,)COOH + H,O = CH,(NH,)COOH + HCHO. 


The sulphuric acid is removed by the addition of barium hydroxide, and 
the filtrate and washings containing the glycine are decolorised by boiling 
with norit. Traces of barium are removed by the addition of the requisite 
quantity of standard sulphuric acid, The filtrate is concentrated on a water- 
bath and the glycine which separates is recrystallised from alcoho]. The yield 
is 90 % of the nitrile employed or taking the yield of the nitrile as 60 % of 
the theoretical, the yield of glycine from formaldehyde is 54% of the 
theoretical. 

This is the highest yield of glycine yet recorded as a result of its direct 
synthesis from formaldehyde. Our procedure has many advantages over the 
methods previously published, one being that the glycine is obtained un- 
contaminated with inorganic salts which are very troublesome to separate. 
Indeed methods of separating these salts have formed the subject matter of 
patents [Farbw. Meister, Lucius and Briining, 1903; Siegfried, 1907]. 


REFERENCES. 


Braconnot (1820). Ann. Chim. Phys. [2], 18, 114. 
Eschweiler (1894). Annalen, 278, 229. 

Farbw. Meister, Lucius and Briining (1903). Zeitsch. angew. Chem. 16, 527. D.R.P. 141,976, 
Gabriel and Kroseberg (1889). Ber. 22, 426. 

Heintz (1862). Annalen, 122, 257. 

Klages (1903). Ber. 36, 1506. 

Mauthner and Suida (1888). Monatsh. 9, 732. . 

(1890). Monatsh. 11, 373. 

Nencki (1883). Ber. 16, 2827. 

Perkin and Duppa (1858). Quart. J. Chem. Soc. 11, 22. 
Siegfried (1907). Chem. Zentr. 11, 1466. D.R.P. 188, 005. 


LXXII. AN INVESTIGATION OF THE CHANGES 
WHICH OCCUR IN THE PECTIC CONSTITUENTS 
OF STORED FRUIT. 


By MARJORIE HARRIOTTE CARRE. 


Department of Plant Physiology and Pathology, Imperial College 
of Science and Technology, London. 


(Received July 27th, 1922.) 


Aw accurate method of estimating pectin in dilute solution by precipitating 
as calcium pectate, has recently been established. The method has been 
successfully applied to the estimation of pectin in some fruit juices and by 
the application of this method it was proved that a complete extraction of 
the soluble pectin of apples can be effected by a process of continuous washing 
out with water. 

As a result of this preliminary work it was found possible to follow the 
changes which take place in the pectic constituents of apples kept in cold and 
ordinary storage and to determine if the improved keeping properties of the 
apples in cold storage is accompanied by any marked difference in pectin 
content. 

The soluble pectin probably develops from an insoluble pectic substance 
contained in the cell wall, which is left behind in the pulp residues after the 
aqueous extraction of the soluble form. This insoluble pectin corresponds to 
the protopectin of Fellenberg [1918], and to the pectose of earlier investigators. 

Estimations of the soluble pectin were systematically carried out at regular 
intervals throughout the period of storage. The apples chosen for the experi- 
ments were very different, Lane’s Prince Albert, a hard and acid cooking 
apple, and the soft, sweet Cox’s Orange Pippin. Equal quantities were stored 
in cupboards at the ordinary temperature, and in a refrigerator at 60° C. 

In view of the possibility that the time of harvesting might influence the 
keeping properties of the fruit, the estimations were carried out on three 
pickings, which were subsequently kept in cold and ordinary store. Batches 
were gathered three weeks earlier than the usual harvesting, at the normal 
time in early September, and the third picking was left on the trees till 
October. | | 

Estimation of the Soluble Pectin. 


The following is a summary of the process of extraction and estimation of 
the soluble pectin; a more detailed account of the method has recently been 
published [Carré and Haynes, 1921]. 

Samples of ten apples were cut up, freed from skin and core, and thoroughly 
mixed, Fifty grams were used for each estimation throughout the season. 


THE PECTIC CONSTITUENTS OF STORED FRUIT 705 


After killing the cells, by freezing the weighed portion of the material in an 
efficient freezing mixture, the extraction of the soluble pectin was carried 
out by washing with water and pressing out in a small hand press. These 
processes were repeated until all the pectin was extracted. The completeness 
of the extraction was verified in all cases by testing the last washings of the 
pulp for pectin by precipitation as calcium pectate. 

The dilute juice was filtered through fluted papers, to remove insoluble 
matter and disintegrated cell substance. It was then boiled to destroy enzyme 
action, and after cooling, the solution was made up to a known volume. 

Aliquot portions were taken for each estimation, the quantity used de- 
pending on the concentration of the pectin solution. It was found advisable 
to make a rough preliminary determination of the amount of pectin contained 
in the juice solution, since, as the fruit approached maturity, the development 
of the soluble pectin from an initial negligible quantity to as much as 0-8 % 
necessitated a careful adjustment of the conditions of estimation. 

The best results were obtained by sie a volume of solution yielding - 
0-02-0-03 g. of calcium pectate. 

For this weight of pectin, 100 cc. N/10 eas were used for the hydrolysis, 
and after 24 hours’ standing the pectin was precipitated as calcium pectate 
by acidification with 50 cc. of N/1 acetic acid, and subsequent addition of 
50 cc. M/1 calcium chloride. During the later period of storage in the case 
of both Cox and Lane apples, it was observed that, although the pectin 
content did not appreciably differ and the conditions of experiment were 
exactly as before, the coagulation of the calcium pectate became ill defined, 
and the precipitate was slimy and difficult to wash and filter. It was found 
however that a satisfactory coagulation could again be obtained by using 
double the amount of N/10 soda previously found adequate for hydrolysis. 
The larger amount of soda was found to be necessary as long as the supply 
of fruit was available, and it seems reasonable to conclude from these obser- 
vations that some change occurs in the state of aggregation of the pectin sol 
as the fruit develops, or possibly an alteration may take place in the consti- 
tution of the pectin. 


THe DEVELOPMENT OF SOLUBLE PECTIN. 


The accompanying graphs (Figs. 1 and 2) are representative of the results 
obtained from the estimations. Fig. 1 illustrates the development of soluble 
pectin in cold and ordinary storage, and Fig. 2 shows the very close similarity 
in the development of the soluble pectin in all three pickings of apples. 

' The experiments indicate that in the early stages of the maturation of the 
apple there is no soluble pectin, but it gradually develops as ripening proceeds 
till it attains a maximum amount when the fruit reaches its fully ripe con- 
dition. The Cox apples ripened earlier than the Lane, and in both cases the 
fully ripe state was found to be coincident with the greatest amount of pectin 
developed. The maximum pectin content was maintained in all cases for 

46—3 


706 M. H. CARRE 


about 4 weeks, during which the apples were in their prime condition. This 
period was followed by a condition of general softening and a sudden drop in 
the pectin content was found to accompany the change of condition in the 
fruit. This softening process is the concluding phase in the apple’s life, and 
it leads eventually to a state of physiological breakdown or of general decay 
brought about by fungal or bacterial action. During the whole softening 


tb Cold store 

Ss Ordinary store ==<=» 

Lol _ 

= 

© “7 = 

to 

[2] 

_ . “6 & 4 

ee g 

2H “5p 2f End End 

% : 

E ee 4 a Fy By) re 

= 2 sf 

2s s rey, 

= -3F S 4 
ng 

Seal 

o “Oh 4 

-_ 

a 

A Saitek | ' — 

7) 

= i 1 L 1 Ll l lL j j 1 Ores | i i 1 1 L j 1 lL i 

Sept. Oct. Nov. Dee. Jan Feb. — Mar. May June 


, Apr. 
19.27.41, 25:8) 629. 6 2OR AS IB as 8 eb bones iD 26 1656 94 eh 
Fig. 1. Development of soluble pectin in cold and ordinary storage (Lane—1Ist picking). 


5 Lave lo 
is | Lane I] ss. : 
S 8F Lane II]----- “ 
sl ; 
—_ 
7 4 
34° 
o's a 
& B, “5E 
os 
ES, 4+ et 
Bf 
3 | al 
iS) 
s 2 4 
> ‘IF : i 
2 ir 
i 1 n L i 1 i i l j L i L 1 1 i - + 1 1 1 
Sept. Oct. Nov. Dee, Jan. Feb. — Mar. Apr. May. June, 
19 27. 17 96. #8). -22: 6... 90:: 4) AB VIB 1 16" 20292 26 1G ea 7 a 


Fig. 2. Development of soluble pectin in cold stored Lane’s picked at different times. 


period the soluble pectin content showed considerable variations, In order 
to ascertain if these variations could be accounted for by differences in the 
condition of the fruit estimations of pectin were made on sets of apples in 
different states of softness, each set being carefully selected to be as uniform 
as possible. The results obtained show that there is a marked decrease in 
pectin as the fruit softens, The following experiments illustrate this point: 


Hard gample Soft sample 
Weight of soluble pectin from 100g... O51 g. 0-42 g. 


THE PECTIC CONSTITUENTS OF STORED FRUIT 707 


It may therefore be concluded that the fluctuating values obtained for the 
soluble pectin are to be attributed to the great variability which was observed 
in the state of maturity of the individual apples taken for each estimation, 
and it follows that in samples of ten taken for each separate determination 
a large sampling error might easily result. It is proposed during the next 
season to carry out estimations on each of the individual apples and to calcu- 
late the probable error due to sampling. 

The routine estimations showed that apples contained varying but appreci- 
able amounts of pectin as long as supplies of fruit were available, and in no 
case where sound apples were examined did the pectin wholly disappear. On 
the other hand when diseased apples were examined there was found in most 
cases to be a complete absence of soluble pectin. 


Errect oF CoLp STORAGE ON THE DEVELOPMENT OF SOLUBLE PECTIN. 


The curves showing the development of pectin in cold and ordinary stored 
fruit (Fig. 1) are very similar, the only effect of the low temperature being 
to prolong the period of ripening. The maximum point of soluble pectin 
content occurs six weeks to two months later in the cold stored fruit. 

After the maximum was passed and during the breakdown of the fruit the 
pectin content of the cold and ordinary stored fruit was found to be practically 
identical as long as samples of both were available for comparison. 

It was possible to carry out these comparisons up to the end of April on 
the cold and ordinary stored Lane’s, but the Cox’s Orange were so badly 
affected by Bitter Pit that the supply of ordinary stored fruit was exhausted 
by the end of December, soon after the maximum point had been reached. 
The greatly prolonged life of the cold stored Cox’s as compared with those 
kept in ordinary storage was due to the effect of cold in retarding the progress 
of disease in the fruit. 

If the results of the estimations of soluble pectin developed in cold and 
ordinary store during the normal life of the apple are added together, the 
quantity is found to agree very closely for apples of the same kind whether 
they have been picked at different stages of maturity or kept under different 
conditions of temperature. 


Total of (12) estimations during six months, 
Cold stored Ordinary stored 


Lane [* ise sit 4-05 g. 4-07 g. 
Lane II ee es 3-99 4-03 
Total of (7) estimations during three months. 
Cox T°... wai eee 3-01 g. 3-07 
Cox II se .F 2-77 2-82 
Picking 
l 2 3 
Lane a nea 4-29 4-11 4-45 
Cox Se ie 4:95 4-70 4-64 


1 — the numerals indicate the different pickings of the apples. 


708 M. H. CARRE 


This agreement is too close to be merely fortuitous, but it is impossible 
to attempt any complete interpretation at present. It may however be 
inferred from the foregoing results that a definite amount of soluble pectin is 
produced during the process of ripening and that it is subject to secondary 
change during the later stages of maturation. 

Estimations of protopectin which are described below, throw some further 
light on the subject. 


ESTIMATION OF PROTOPECTIN. 


With a view to examining the pectin changes in the apple more thoroughly 
it became necessary to devise a quantitative method of estimating the 
insoluble pectin constituent or protopectin!, as it will be called in the following 
account, and so to obtain some idea of the relations between the soluble and 
insoluble forms. 

The method adopted was to treat the residue after extraction of the soluble 
pectin with 100 cc. N/20 HCl in an autoclave for about an hour at a tempera- 
ture of 110°. The material became thoroughly disintegrated during the process 
and the protopectin was converted into the soluble form, presumably by a 
process of hydrolysis. It was easily washed out with very little pressing, 
and the extract was estimated by the calcium pectate method. That the 
treatment removes all the insoluble pectic substances capable of being trans- 
formed by the acid was proved by repeating the autoclaving with a similar 
amount of N/20 HCl in which case no soluble pectin was found in the second 
extract. Subsequent treatments with greater concentrations of acid NV/5 and 
N/1 also failed to produce any further trace of pectin. 

To test the accuracy of the method equal amounts of apple pulp were 
taken, and the soluble pectin in them was extracted and estimated. The 
residues were treated as described above in order to convert the protopectin 
into soluble pectin, and the extracts thus obtained were also estimated. The 
following results illustrate the measure of agreement in the amounts obtained 
from two samples of uniform material. Both protopectin and soluble pectin 
are estimated in terms of the calcium pectate obtained from them. 


Weight (g.) obtained from 100 g. of pulp. 
Total pectin (Protopectin 


Soluble pectin Protopectin and pectin) 
(a) 0-43 0°36 0:79 
(b) 0-385 0-45 0-835 


It has been found possible to estimate the total pectin directly from the 
apple pulp without preliminary washing out of the soluble form. 


* The adoption of this name for the hydrolysable pectin present in the cell wall implies that 
this is the precursor of soluble pectin, While there is much reason to believe that this is so, the 
fact cannot be regarded as finally established, and the present nomenclature must therefore be 
considered provisional, 


THE PECTIC CONSTITUENTS OF STORED FRUIT 709 


Apple Sample I from Lane. 
(Weight obtained from three similar samples of 100 g. of apple.) 
Picking 
1 2 3 
Soluble pectin ... oo _ 0-348 
Protopectin vie — — 0-094 
Total i 0-404 0-418 0-442 
Apple Sample II from Coz. 
Soluble pectin ... — — 0-52 
Protopectin Sas _- — 0-24 


Total pectin 0-71 0-71 0-76 


The amounts estimated in this way are in fairly close agreement with the 
total arrived at by adding the weights obtained by separate estimation of 
the soluble and insoluble pectin. The slightly lower values obtained in both 
these examples where the total pectin is extracted from the pulp may be 
entirely due to experimental error, but it is possible that some change may 
be brought about by the acids and salts contained in the juice. 


ESTIMATION OF PROTOPECTIN OF APPLES KEPT IN COLD 
AND ORDINARY STORE. 


During the latter half of the period of storage this method was systemati- 
cally used in a preliminary examination of the protopectin content of apples. 
Estimations were carried out on the residue of the weighed sample from which 
all the soluble pectin had been extracted. The residue was then autoclaved 
with acid as described above, and the resulting soluble pectin precipitated as 
calcium pectate and weighed. In this way the weight of protopectin and 
soluble pectin was determined in a given sample of apple. 

The results obtained are of necessity incomplete, and the following account 
must be regarded as in the nature of a preliminary survey. | 

The protopectin determinations were made after the maximum soluble 
pectin content had been passed and the protopectin shows the tendency to 
fluctuate which was observed in the case of the soluble pectin. 

The following estimations of the protopectin content of hard and soft 
apples show that there is less in the softer fruit. As the samples contained 
individuals differing greatly in their degree of maturity these fluctuations may 
be attributed, as in the case of the soluble pectin, to the variation of individual 
apples in each set taken for estimation. 


Weight of protopectin 


in 100 g. of samples Hard sample Soft sample 
A. Lane II ordinary stored 0-45 0-35 
B. Lane II cold stored 0-28 0-18 


A series of estimations of pectin and protopectins carried out at regular 
intervals on the same apples, showed that a very definite relationship exists 
between them and that the changes in the two constituents tend to be equal 


710 M. H. CARRE 


and opposite in amount. The accompanying graph (Fig. 3) is typical of the 
results obtained: 

The foregoing results show that after the maximum soluble pectin content 
is passed the quantity of soluble pectin tends to decrease and that the decrease 
is roughly proportional to a simultaneous increase in the insoluble pectin. On 
the other hand both the protopectin and soluble pectin have been shown to 
decrease markedly in soft over-ripe fruit as compared with harder less ripe 
fruit. 

These observations seem to indicate that there is a balance between the 
protopectin and soluble pectin during-the normal life of the apple, but that 
in the later stages of breakdown the total pectin tends to decrease. On the 
other hand there is some evidence to show that the total pectin content tends 
to increase as the apple ripens. The following estimations were carried out 


Soluble pectin —— 
Protopectin ~<-<- 


g. calcium pectate 
: - 3 
or 
LJ 
s 
‘ o~ 
= o 
>/2 
! _ 
' 
& 
| 


bes rOt ~ 

20 == 2 Peet; 
* 
tee ee Mere fe 
. *. ee. 
15 i oe “Ness ond 
bd "ae 

10 2 


l lL l l l | | 1 | l 
Mar 12 19 26 Apr.9 16 23 30 May 7 14 at. 


Fig. 3. Relationship of pectin and protopectin in Lane II ordinary store. 


on cold stored “Cox’s” when the apples were in the early stages of their 
development and during the period of fluctuation when the fruit was beginning 
to get over-ripe. 


Soluble pectin Protopectin Total 
Early unripe state (Sept. 20)... 0-05 0-48 0:53 
Fully ripe state (Feb. 16—Mar. 16)" 0-65 0:36 1-01 


* These figures represent the mean of the weights obtained during this period. 


These results need confirmation but appear to show that in the few weeks 
following the picking of the apples the total pectin increases rapidly owing to 
the marked development of soluble pectin. The change in the protopectin 
content is much less pronounced and is insufficient to account for the increase 
of the soluble form. It may therefore be inferred that the total increase in 
pectin is due to a third source of pectin, which either gives rise to soluble 
pectin directly, or is transformed into it through the intermediate state of 
protopectin, 


‘ 


THE PECTIC CONSTITUENTS OF STORED FRUIT 711 


THE FACTORS CONTROLLING THE DEVELOPMENT OF SOLUBLE PECTIN. 


The following investigations were made to obtain some information as to 
whether the production of soluble pectin is due to enzyme activity. 

Samples of apple which had been previously killed by freezing were 
crushed with sand and thoroughly washed and pressed free from soluble 
pectin. The residues were left with water and a little thymol added to prevent 
bacterial action. At the end of a week the residues were again pressed out 
and tested for soluble pectin. In all cases examined pectin was found to have 
developed. The samples were put back in water and the process repeated 
every week till no more pectin was found to have developed. The following 
examples carried out on ordinary stored “Cox’s”’ illustrate the results so 
obtained. 


WEIGHT OF SOLUBLE PECTIN DEVELOPED FROM 100 G. OF APPLE PULP. 


A B 
Cox III Cox IT 
’ Ordinary stored Ordinary stored 
Ist week ... mae 0-22 0-30 
2nd week ... ore 0-35 0-05 
3rd week ... nA 0-04 0:05 
4th week ... ta 0-00 0-06 
5th week ... Mi 0-00 0-02 
6th week .,. ae 0-00 0-00 


A series of experiments on similar lines was carried out on swedes previ- 
ously killed by ether and washed free from soluble pectin. As with the 
experiments on apple material a sample left standing with water showed a 
steady development of soluble pectin. Another portion was boiled with water 
for three hours and then pressed out and examined. Much soluble pectin was 
found in the extract. The residue was washed free and then again boiled for 
another three hours. No soluble pectin was obtained in this case. The residue 
was divided into two parts for different treatment. One portion was auto- 
claved with a solution of N/20 HCl at 110° and subsequent examination 
showed that a great deal of pectin had been produced in the process. The 
second portion was kept in water for one month but no soluble pectin was 
developed. It was then autoclaved with acid and a great deal of pectin was 
produced. 

The foregoing observations show that a certain amount of pectin appears, 
if a residue previously washed free from pectin is left with water for a time, 
but that a steady state is finally reached when no more pectin is produced 
however long the sample is kept. Similar results are obtained by boiling; a 
certain amount of pectin is brought into solution but subsequent boiling for 
long periods has no further effect and no more pectin is developed after pro- 
longed standing with water. On the other hand in both these cases large 
quantities of pectin are produced by autoclaving with acid. A great deal of 
further investigation is necessary before any definite conclusions can be made, 
but such facts as have already been ascertained support the theory that the 


712 _ M. H. CARRE 


production of soluble pectin is due to the action of an enzyme which must 
be supposed to be present in the cell walls in a form which it was not possible © 
to extract. 

SUMMARY. 


An account is given of the development of soluble pectin in different kinds 
of apples kept in cold storage and at ordinary temperature. It is shown that 
the pectin reaches a maximum during the process of ripening and then gradu- 
ally falls as the apple becomes over-ripe. The date of picking of the fruit has 
no effect on the development of the pectin in either cold or ordinary store. 

A method is described for the quantitative estimation of protopectin by 
hydrolysis with weak acid. Preliminary work carried out by the application 
of this method suggests that there is a definite relationship between the amounts 
of soluble and insoluble pectin constituents. The experiments indicate the 
possibility of a third source of pectin. 

Preliminary experiments tend to show that the development of the soluble 
pectin may be attributed to enzyme activity. 


This investigation has been carried out for the Food Investigation Board 
of the Scientific and Industrial Research Department. The author wishes to 
express her gratitude to Dr Haynes of this Department for her valuable 
criticism and advice. 


REFERENCES. 


Carré and Haynes (1921). Biochem. J. 15, 60. 
Fellenberg (1918). Biochem. Zeitsch. 85, 118. 


PHYSIOLOGICAL 
moe tf RAC TS 


Monthly, price 5s. net. Annual Subscription, post free, 42s. 


This Journal is issued by the PHYSIOLOGICAL 
SOCIETY, acting in co-operation with numerous 
physiological organisations in Great Britain, America, 


and other countries. The Editor, Professor HALLIBURTON, 
is assisted by a staff of competent abstractors both at 


home and abroad. 


The Journal aims at issuing promptly abstracts of 
the papers published throughout the world, in physio- 
logical and allied sciences (including Plant Physiology). 


A number is published at the beginning of each month. 
Volume VII. began with the April (1922) issue. 


Subscriptions are only taken for April to March of the 
next year, and are payable in advance. 

Back numbers, excepting some which are out of print, may be 
obtained from the Publishers, or any Bookseller. An extra charge is 
made for some numbers the stock of which is nearly exhausted. Prices 
on application. 


Binding Cases, price 2s. net, are supplied by the Pub- 
lishers, who can also arrange for binding; prices on application. 


Published by H. K. LEWIS & CO. LTD. 
136 GOWER STREET, LONDON, W.C. | 
Publishing Office, 28 GowER PLAcE, Lonpon, W.C. | 


Subscriptions may be paid through any Bookseller 


i 


Cambridge University Press 


Practical Plant Biochemistry. By M. WHELDALE ONSLOW, formerly 
fellow of Newnham College, Cambridge, and Research Student at the John Innes 
Horticultural Institution. Royal 8vo. 16s. e¢ 


Basic Slags and Rock Phosphates. By GEORGE SCOTT ROBERTSON, 
D.Sc. (Dunelm.), F.I.C., witha Preface by Sir E. J. RUSSELL, D.Sc. (Lond.), F.R.S. 
With 8 plates and a map. Small Royal 8vo. 14s. met. Cambridge Agricultural 
Monographs. 


A Course of Practical Chemistry for Agricultural Students. By 
H. A. D. NEVILLE, M.A., F.I.C., and-L. F. NEWMAN, M.A., F.I.C. Demy 8vo. 
Vol. I, 10s. 6a. mez; Vol. II, Part I, 55. met. 


A Course of Practical Work in Agricultural Chemistry for Senior 
Students. By T. B. WOOD, M.A., F.R.S., Drapers Professor of Agriculture in the 
University of Cambridge. Demy 8vo. Papercovers. 2s. met. 


A Course of Practical Physiology for Agricultural Students. By 
J. HAMMOND, M.A., and E. T. HALNAN, M.A. Demy 8vo. With blank pages 
interleaved. Cloth, 6s. 6d. net; paper covers, 45. 6d. net. 


Physiology of Farm Animals. By '. B. WOOD, C.B.E., M.A., F.R.S., and 
F. H. A. MARSHALL, Sc.D., F.R.S. Part I, General. By F. H. A. MARSHALL. 
Demy 8vo. 16s. net. 


Insect Pests and Fungus Diseases of Fruit and Hops. A Complete 
Manual for Growers. By P. J. FRYER, F.I.C., F.C.S. Crown 8vo. With 24 
plates in natural colours and 305 original photographs and diagrams. 455. met. 


Fungoid and Insect Pests of the Farm. By F. R. PETHERBRIDGE, M.A. 
With 54 illustrations, Crown 8vo. 55. 6d. me¢. Cambridge Farm Institute Series. 


Manuring for Higher Crop Production. By Sir E. J. RUSSELL, D.Sc., F.R.S. 
Second edition, revised and extended. Demy 8vo. With 17 text-figures, 55. 6d. med. 


The Fertility of the Soil. By Sir E. J. RUSSELL, D.Sc., F.R.S. With 9 illus- 


trations. 16mo. Cloth, 3s. ze¢; lambskin, 4s. wef. Cambridge Manuals Series. 


A Student’s Book on Soils and Manures. By Sir E. J. RUSSELL, D.Sc. 
Second edition. Revised and enlarged. Large crown 8vo. With 41 illustrations. 8s. xe¢. 


Inorganic Plant Poisons and Stimulants. By WINIFRED E. BRENCH- 
LEY, D.Sc., F.L.S., Fellow of University College, London. . Royal 8vo. With 
19 illustrations. 9s. et. 


Plants Poisonous to Live Stock. By HAROLD C. LONG, B.Sc. (Edin.), 


of the Board of Agriculture and Fisheries. Royal 8vo. With frontispiece. 8s. net. 


Brewing. By A.C. CHAPMAN, With 14 illustrations. Royal 16mo. 2s. 6d. net 
and 3s. #ef. Cambridge Manuals Series. 


Chemistry for Textile Students. By B. NORTH and N. BLAND. Demy 8vo. 
30s. net. Cambridge Technical Series, 
Cambridge University Press 
Fetter Lane, London, E.C, 4: C. F. Clay, Manager 


THE UNIVERSITY OF CHICAGO PRESS 


THE UNIVERSITY OF CHICAGO SCIENCE SERIES 


The Evolution of Sex in Plants. By JOHN MANNING COULTER. x+140 pages, I2mo, 
cloth; 7s. net. 

Individuality in Organisms. By CHARLES MANNING CHILD. x +122 pages, small 12mo, 
cloth; 8s. 3d. net. 

‘The Origin of the Earth. By Tuomas C. CHAMBERLIN. xii+272 pages, small 12mo, 
cloth; gs. 6d. net. 

Finite Collineation Groups. By Hans F. BLICHFELDT. xii+194 pages, small 12mo, 
cloth; 8s 3d. net. 

A Chemical Sign of Life. By Surro TasHIRO. x+142 pages, small 12mo, cloth; 
7s. net. 

The Biology of Twins. By Horatio Hackett NEWMAN. x+186 pages, small 12mo, 
cloth; 8s. 3d. net. 

Food Poisoning. By EDWIN OAKES JORDAN. viii+116 pages, small 12mo; 7s. net. 

The Electron: Its Isolation and Measurement, and the Determination of Some of 
its Properties. By Rosprert ANDREWS MILLIKAN. xii+268 pages, small 12mo, 
cloth; gs. 6d. net. 

The Living Cycads. By CHARLES JoSEPH CHAMBERLAIN. xiv+1I72 pages, smal] 12mo, 
cloth; 8s. 3d. net. 

Problems of Fertilization. By Frank R. LILLIE. xii+278 pages, small 12mo, cloth; 
gs. 6d. net. 


Volumes planned or in preparation 


Black Body Radiation. By CHARLES E. MENDENHALL. 

Mechanics of Delayed Germination in Seeds. By W. CRocKER. 

The Rigidity of the Earth and of Materials. By A A. MICHELSON. 

Linear Integral Equations in General Analysis. By Er1akim HastinGcs Moore. 


Sold in the British Empire (except Canada) by 
CAMBRIDGE UNIVERSITY PRESS, FETTER LANE, LONDON, E.C.4 


Cambridge University Press 


The Naturalisation of Animals and Plants in New 
Zealand. By Hon. Gro. M. Tomson, M.L.C., F.L.S., F.N.Z.Inst. 
Royal 8vo. 42s. net. 

Cattle and The Future of Beef-Production in England. 
By K. J. J. Mackenzie, M.A., with a preface and chapter by F. H. A. 
MarsHAaLL, Sc.D., F.R.S. Demy 8vo. 7s. 6d. net. 


Microscopic Analysis of Cattle=-Foods. By T. N. Morris, B.A. 
With 54 text-figures. Crown 8vo. 35. net. 

British Grasses and their Employment in Agriculture. 
By S.F.Armstronc,B.A.,F.L.S. With 175 illustrations. Demy 8vo. 10s.6d.net. 

The Chemical Examination of Water, Sewage, Foods and 
other Substances. By J.E. Purvis, M.A. and T. R. Hopcson, M.A. 
Second and enlarged edition. Demy 8vo. 20s. net. 

The Bacteriological Examination of Food and Water. 
By W.G. Savace, B.Sc., M.D.,D.P.H. Second edition. With 16 illustrations. 
Demy 8vo. tos. net. 


Food Poisoning and Food Infections. By W. G. Savacz, B.Sc., 
M.D., D.P.H. Demy 8vo. 15s. net. 


Cambridge University Press, Fetter Lane, London, E.C.4: C.F. Clay, Manager 


PHYSIOLOGICAL REVIEWS. 


PUBLISHED BY THE 


AMERICAN PHYSIOLOGICAL SOCIETY 


UNDER THE EDITORIAL DIRECTION OF 


W. H. HOWELL, BALTIMORE 
REID HUNT, Boston 
F.S. LEE, New York 


J. J. R. MACLEOD, Toronto 
LAFAYETTE B. MENDEL, New Haven 
H. GIDEON WELLS, Cuicaco 


D. R. HOOKER, Managing Editor, BALTIMORE 


Containing short but comprehensive articles dealing with the recent 
literature in Physiology, using this term in a broad sense to include 
Bio-Chemistry, Bio-Physics, Experimental Pharmacology and Ex- 


perimental Pathology. 


CONTENTS. VOLUME III, 1923 


JANUARY 
A. J. Cartson: The Gastric Secretion in Health 
and Disease 
J. A. Gunn: Cellular Immunity 


W.T. Lonccope: Hyper-sensitization in Man 
and its Relation to Disease 


E. A. Park: The Etiology of Rickets 


PeyTon Rous: Destruction of Red Blood Cor- 
puscles in Health and Disease 


APRIL 
H. H. Date: Chemotherapy 
W. McK. Marriotr: Anhydremia 


YANDELL HENDERSON: Volume Changes of the 
Heart 


F. H. Pixe: Functions of the Vestibular Ap- 
paratus 


A. T. SHout_: Mineral Metabolism 


PUBLISHED 
QUARTERLY 


Subscriptions should be sent to 


SUBSCRIPTION 
IN ADVANCE ONLY $6.50 FOREIGN 


JULY 
C. Lunpscaarp and C. A. L. Brncer: Lung 
Volume 


D. W. Witson: Neutrality Regulations in the 
Body 


F. H. A. MarsHa.t: Internal Secretions of the 
Reproductive Organs 


P. A. SHAFFER: Intermediary Metabolism of the 
Carbohydrates 


A. I. KenpA.L: Bacterial Metabolism 


OCTOBER 
C. S. SHERRINGTON: Functional Development 
of the Tracts of the Central Nervous System 
S. R. Benepicr: Purin Metabolism 
T. H. Morcan: The Mechanism of Heredity 


H. C. BrapLey: Autolysis in its Relation to 
Atrophy 
G. W. Corner: Ovulation and Menstruation 


$6.00 DOMESTIC 


Dr D, R. HOOKER, 
1222 ST PAUL STREET, 


BALTIMORE, Mb., U.S.A, 


LXXIII. THE PRESENCE OF THE ANTINEURITIC 
AND ANTISCORBUTIC VITAMINS IN URINE. 


By N. VAN DER WALLE. 
From C. Eujkman’s Laboratory of Hygiene, Utrecht, Holland. 


(Received July 29th, 1922.) 


Funk [1914], Serpexy [1921] and others applied chemical methods in order 
to determine the structure of the antineuritic vitamin. Their efforts to isolate 
these mysterious substances were not perfectly successful and therefore we 
do not wonder that nothing is known as yet of the metabolism of these factors. 
However, recently some experiments have been made to examine the excretion 
of antineuritic vitamins from the organism. 

Cooper [1914] succeeded in curing polyneuritis-pigeons with an alcoholic 
extract of faeces. 

Muckenfuss [1918] tried to find antineuritic vitamins in ox-bile, human 
urine and saliva by shaking up these substances with fuller’s earth. The 
fuller’s earth, treated in this way, purified with water and alcohol and then 
dried, proved to be efficient in curing pigeons with polyneuritis. 

Gaglio [1919] at Rome also investigated the result of administering urine 
to polyneuritis-pigeons. For this purpose he concentrated the urine and found 
that even 3-4 cc. were able to cure the birds. Urine of rabbits that had fasted 
for 15-20 days was likewise active, though less than that of normal rabbits 
which were fed on cabbage leaves and bran. The ash of urine proved to be 
without action. 

Gaglio [1919] as well as Curatolo [1920] are convinced that these cures are 
not to be attributed to the action of vitamins but to the influence of non- 
specific products of metabolism. On this supposition is founded the research 
of di Mattei [1920], who tried to cure pigeons with coffee. Indeed, 5 % of 
mocha proved to have a strong curative power, but to his astonishment he 
perceived that pure caffeine (in non-toxic doses) had only a slight effect. 

Funk [1912] had previously demonstrated that some purine and pyri- 
midine derivatives had a more or less favourable influence on polyneuritis. 

It has also been proved that a mixture of NaCl and KCl may have a good, 
though temporary effect on the disease [Hijkman, van Hoogenhuyze and 
Derks, 1922]. 

Still more remarkable is the observation made by Theiler who described the 
favourable result of subcutaneous injection of distilled water [1921]. 

Bioch, xv 47 


714 N. VAN DER WALLE 


It is difficult to explain this effect from a stimulation of vitamins, still 
present in the organism. Theiler attributed this action to “spontaneous re- 
covery.” 

Jansen [1920], in the Medical Laboratory at Weltevreden (Java), confirmed 
the observations of Eijkman and Theiler. 

Yet, the results described by Gaglio and Curatolo are so striking, that we 
doubt whether such a lasting and absolute recovery may be attributed to the 
influence of non-specific products of metabolism. 

Funk informs us that the recovery obtained by treatment with hydantoin 
lasted only nine days, whereas a pigeon treated with thymus-nucleic acid 
lived on for a fortnight. Allantoin, adenine and other pyrimidine derivatives 
were less active. From the administration of uric acid he did not obtain any 
effect. 

It occurs to us that much is to be ae in favour of the opinion of 
Muckenfuss who believes antineuritic vitamins to be present in urine. 

By a series of experiments on pigeons we tested the opinions of these 
investigators and tried to find the most probable explanation. 

First, 12 pigeons (Nos. 1-12) were fed on polished rice and water till they 
showed symptoms of polyneuritis. Then 5 cc. of urine that had been concen- 
trated in vacuo at 45° were administered per os. In six cases this had 
undoubtedly a favourable influence. 

In most cases the symptoms had much improved one aay after the ad- 
ministration; sometimes even the influence was evident after a few hours. 
The body weight, having diminished before the appearance of the clinical 
symptoms, went up again in all cases of improvement. Yet we did not succeed 
in regaining the initial weight. 

Eight pigeons (Nos. 13-20) were treated prophylactically with the ad- 
ministration of concentrated urine from the very first. The result was that 
the average incubation period increased from 22 to 42 days. 

To enhance, if possible, the curative power of the urine the latter was 
dried in large open Petri dishes in the incubator at 37°. After about 16 hours 
a thin layer of tough, semi-solid substance was found which was dried in an 
exsiccator at room temperature; 1g. of dried substance was found to be 
equivalent to 25 cc. of urine with a s.q. of 1-030. 

Seven polyneuritis-pigeons (Nos, 21-27) were treated with this dried urine, 
dissolved in autoclaved milk and administered in quantities of 1-5-2 g. per day. 
Four of these pigeons were cured rapidly and absolutely. After stopping the 
administration of dried urine, some days passed before the birds showed 
further symptoms of the disease. We succeeded in curing a second attack 
in two pigeons. 

Three pigeons (Nos. 28-30) were treated by subcutaneous injection of 
dried urine, dissolved in sterilised milk. These however died after a few days. 

It is well known that strong heating destroys the antineuritic vitamin. 
We therefore tried the action of urine which had been heated for one hour 


VITAMINES IN URINE 715 


at 130°. After sterilisation the urine was filtered, neutralised with hydro- 
chloric acid and then concentrated in vacuo to a specific gravity of 1-050. 

Ten polyneuritis-pigeons (Nos. 31-40) were treated with 15 cc. of this 
liquid daily. In one case the disease lasted only seven days, whereas the 
medicament had not the least effect on the other birds. This proves that by 
strong heating the curative power of the urine is decreased. 

We also investigated the action of urine ash. For this purpose dried urine 
was incinerated, by which process it lost about 75 % of its weight. Ten 
pigeons (Nos. 45-54) were treated per os with this ash, suspended in water, 
and it was found to be almost inactive. 

Chamberlain and Vedder [1911] observed that the antineuritic factor was 
removed by filtering through animal charcoal. We made use of this dis- 
covery and shook up urine with purified animal charcoal, 200 cc. of urine 
being shaken up for 45 minutes with 5 g. of animal charcoal; afterwards the 
urine was centrifuged and the charcoal shaken up anew with 200 cc. of fresh 
urine. This process was repeated three times. The charcoal was then washed 
once with distilled water and dried at 37°. 

Six polyneuritis-pigeons (Nos. 60-65) were treated with this “carbo cum 
urina,” 2g. of this substance suspended in a little water being administered 
per day. All the birds recovered within a short time and we kept them 
healthy for at least a fortnight. The second attack too was cured. The 
smallest active dose proved to be 1 g. 

Crude (non-purified) charcoal, shaken up with urine and treated in the 
same way, was far less active (Table I). 


Table I. Treatment of polyneuritis-pigeons with carbo cum urina. 


Weight (g.) 
Laboratory Incubation Period of Dose of - 
no. of period illness* Effect of Carbo cum Beginning of End of 
pigeon (days) (days) treatment wrina, g. Remarks experiment experiment 
Crude charcoal 
55 16 1 0 2 Died 240 200 
56 21 2 0 2 = 265 175 
57 45 1 0 2 » 240 175 
58 22 5 ck 2 Tempor. improved 350 225 
59 16 12 -E 2 aa ‘9 260 200 
Carbo anim. puriss. 
60 24 18 + 2 Much improved 295 175 
61 17 18 + 2 > $s 265 175 
62 35 —- + 2 i oe 280 220 
63 22 30 + 2 $3 +3 380 315 
64 14 30 + 2 a na 270 ? 
65 32 — + 2 = 3 265 ? 
66 14 1 + 1 s Me ? ? 
67 18 15 ck 1 Slightly improved 355 260 
68 17 19 + 0: 9 99 330 290. 


* The period of illness is counted from the moment the treatment begins until death or, in 
case of recovery, until the end of experiment. 


Another experiment on 12 pigeons (Nos. 69-80) showed that the non- 
treated animal charcoal had absolutely no effect. 
47—2 


716 N. VAN DER WALLE 


Normal urine, shaken up with a surplus of charcoal, then filtered and con- 
centrated in vacuo, had no effect (pigeons Nos. 81-89). 

Then eight polyneuritis-pigeons (Nos. 90-98) were treated with animal 
charcoal, which had been shaken up with urine, heated previously for three 
hours at 120°. We expected this charcoal to be ineffective. On the contrary, 
two pigeons recovered and walked normally for ten days: six other pigeons 
were not cured. Without a doubt the curative effect of this charcoal is far 
inferior to that of charcoal shaken up with fresh urine. However, it must be 
admitted that there are some substances present in urine, possessing a curative 
power, and capable of being adsorbed by charcoal, and yet not bearing the 
character of the so-called vitamins. We tried to find out whether these sub- 
stances have an organic or an inorganic character. For this purpose a mixture 
of inorganic salts was made, approximately in the same proportion as those 
found in human urine: 


NaCl 9-2 % NagHPO, 0% 
K,S0O, 2-7 MgCl, 15 
(NH,).SO0, 1:5 CaH,(PO,). 0:25 


KH,PO, 2:8 
This solution was shaken up with purified animal charcoal and ten pigeons 
(Nos. 99-108) were treated with this charcoal, which proved to have a very 
small effect as was expected beforehand, Rona and Michaélis [1919] having 
demonstrated that animal charcoal will only slightly adsorb inorganic salts, 
the cations and anions present in urine being least adsorbed. 

We may therefore safely assume that sterilised urine has a definite curative 
power owing to the presence of non-specific organic substances. 

Finally we made an investigation as to whether the bacilli present in urine 
had any influence. The urine used was fresh, but not sterile. We did not 
expect to have a positive result here as the quantity of bacilli as far as weight 
was concerned, was very small. 

Besides it has been already shown that B. coli does not contain any anti- 
neuritic vitamin even when cultivated in an extract of rice bran [| Hijkman, 
van Hoogenhuyze and Derks, 1922], whilst Damon [1921] demonstrated that 
the B-vitamin is absent from B. paratyphosus B., B. coli and B. subtilis. Fresh 
urine was kept at 37° for 24 hours. The bacilli out of this urine were culti- 
vated on agar, then suspended in distilled water and afterwards heated at 
100° for 30 minutes, This suspension of dead bacilli was then filtered through 
animal charcoal. After this process the latter was dried at 37°. As expected 
this charcoal proved to have scarcely any curative power (Table IT), 

Finally the fact that yeast is only active in a medium that contains 
antineuritic vitamin or its components was made use of [Kijkman, van 
Hoogenhuyze and Derks, 1922]. Baker’s yeast was cultivated at 28° for 
24 hours in fresh urine, the acid reaction of which had been diminished and to 
which had been added 5 % of glucose. This yeast was centrifuged, washed 
once with distilled water and dried at 37°. It proved to have a strong curative 


VITAMINES IN URINE “17 


action on polyneuritis of pigeons, 2 g. of yeast being sufficient to ensure the 
birds an absolute and lasting cure. This experiment makes it probable that 
antineuritic vitamins or their components are present in fresh, normal urine. 


Table IT. Treatment with bacilli-charcoal. 


Weight (g.) 
= 


Laboratory Incubation Period of Dose of 

no. of period illness Effect of _bacilli- Beginning End of 

pigeon (days) (days) treatment charcoal, g. Remarks of experiment experiment 
109 27 5 if 2 Slightly improved 330 230 
110 31 6 ck 2 ms i 360 350 
111 19 4 0 2 Died 300 190 
112 20 5 0 1 an 300 250 
113 28 4 0 0-5 vA 340 205 
114 28 10 + 0-25 Slightly improved 350 200 
115 15 7 a 0-25 es F 325 260 
116 19 ft 0 0:25 Died ? ? 
117 19 4 0 0-25 oF ? ? 
118 19 4 0 0-25 ve 305 240 


In order to trace whether diet has any influence on the activity of urine, 
the urine of a dog also was examined. Charcoal shaken up with this urine 
proved to possess as curative an action as human urine. Two g. per day were 
quite sufficient, but one pigeon even recovered from the use of 0-25 g. per day. 
As long as we were experimenting with its urine, the dog was fed on potatoes, 
polished rice and meat. This food contains a sufficient quantity of anti- 
neuritic vitamin. Then the diet was altered by autoclaving all the dog’s food 
(bread, rice, meat and potatoes) during three hours. Only three weeks after 
the dog had been taking the vitaminless diet, the urine was again examined 
every day and shaken up with charcoal. The urine proved to have lost all 
its curative power (Table ITI). 


Table IIT. Treatment with charcoal shaken up with urine of a dog. 


A. The dog was fed on normal diet. 


Weight (g.) 
Laboratory Incubation Period of 
no. of riod illness Effect of Dose of Beginning of End of 
pigeon (days) (days) treatment charcoal, g. Remarks experiment experiment 
123 13 — + 2 Much improved 280 230 
124 23 —_ + 15 99 ” 340 285 
125 28 — + 1 9 Fre 380 300 
126 25 _— +. 1 - as 380 250 
127 42 1 0 1 Died 300 195 
128 33 8 + 0-5 Slightly improved 310 290 
129 31 1 0 0-5 Died 320 190 
130 37 _ + 0-25 Improved 290 245 
131 35 8 0 0-25 Died ? ? 
B. The dog was fed on a vitamin-free diet. 

132 23 6 0 2 Died 320 260 
133 28 2 0 2 PP 300 220 
134 22 + 0 2 Pa 315 200 
135 21 2 0 2 99 360 ? 

136 26 5 0 2 PFs 330 ? 

137 22 6 ch 2 Slightly improved 350 250 
138 17 2 0 2 Died 300 240 
139 26 5 0 2 a 340 205 


718 N. VAN DER WALLE 


It became more and more obvious that the favourable result must at 
least partly be attributed to the presence of the antineuritic vitamin, yet we 
went on to treat some fowls with urine because up till now it has never been 
demonstrated that non-specific substances have any influence on the poly- 
neuritis of fowls. We found that fresh, concentrated urine (s.G. 1-050), as well 
as dried urine, when administered in large quantities is able to cure poly- 
neuritis-fowls. Still more efficient was the action of animal charcoal, shaken 
up with fresh human urine. This however has to be administered in doses of 
15 g. (equivalent to 2-5 litres of urine!) a day. Even then it required ten days 
to cure the fowl whilst we did not always succeed in curing the symptoms of 
paralysis absolutely. By a check experiment on six fowls it was shown that 
the charcoal itself did not exercise any curative action, and a fowl treated 
with charcoal that had been shaken up with urine, heated aetis! at 130° 
for one hour, did not recover. 

The result of the experiments on fowls strengthened our opinion as to the 
presence of a small quantity of antineuritic vitamin in urine. 


Antiscorbutic vitamin. 


We then examined whether the antiscorbutic vitamin was also present in 
the urine. 

For this purpose some guinea-pigs were put on a scorbutic diet receiving 
in addition fresh urine per os. The standard diet selected was a mixture of 
oats and bran, which together with water was given ad lib. Great care was 
taken to give the animals the best conditions possible. Following the example 
of Delf [Delf and Tozer, 1918] five guinea-pigs (Nos. 1-5) (Fig. 1) were given 
about 60 ce. of milk, previously autoclaved for an hour at 120°. Two other 
animals (Nos. 6 and 7), kept on the same diet, received daily 10 cc. of urine 
(which had been diluted to a specific gravity of 1-020). All these guinea-pigs 
developed scurvy and succumbed after some weeks. By post-mortem histo- 
logical examination of the costochondral junctions we found the abnormalities 
described by Tozer [Delf and Tozer, 1918] in the severe stage of scurvy: irregu- 
larity of the junction, disorder of the rows of cartilaginous growing cells, 
usually a well-developed reticular zone, haemorrhages etc. 

The average loss of weight of the guinea-pigs which were getting urine 
came to 29 % as against 19 % of the former group, the average lifetime being 
28 days against 27 of the first group. 

Evidently the administration of urine had no influence on the severity of 
the microscopical changes and on the average lifetime, whereas the loss of 
weight had increased, 

(Though we made a histological examination of the ribs of all the guinea- 
pigs we are conscious that we have to be cautious in judging the results. It 
is known that Tozer [1921] demonstrated that the lack of A-vitamin may 
bring about departures from the normal which closely resemble those which 
are present in chronic cases of scurvy.) 


719 


‘oul jo "00 OI+ “ ce se sé cy *9 ‘SON ‘soid-vourny) 
"9% PY (OCT 4% “Y 1) ¥[ur “Jooyne f10yeM pue ueIq ‘s}¥O :yaIq “C-] ‘SON ‘sB1d-voumny 
‘DT pue jy syuommodxg *] “Sq 


VITAMINES IN URINE 


007 
Kar ed 
ee Kf ot ees 
a ae 
a) tt ott + i+ a ie 
o ae x — oo XK Seas ane O0¢ 
| 
oO 
) xX 
4 Oo 
ge re | Om 
Bin ae I ge 
— 
~~. - e —_—ee 
Ege: —_ Oe ee dl neni ee i 
~~ —_ . - 
, 
. { 
: 
i + aes 00¢ 
| —_ - ~~ : bf 
x re —_—_ 
=. al > 
— _— SS 
—_—_-— x pee « Bs 
aes ae i pen 
th 
oy 
4 £ 2 £ Vv 


‘ouLIM JO ‘00 OT + * fs + a *LI-8 ‘SON ‘sd1d-eournsy 
“YIU *[oOyNR Jo ‘00 OT ‘10ywM “URIG ‘8}VO 24a7q7 “GT-ZI “SON ‘satd-eouny) 
pT] pue Ty syuounedxg °% “Sry 


00T 


Ves 


N. VAN DER WALLE 


720 


a | 
oe. 
arpa 8 A ee bal oe Pe 
fg— {=i , Eatin {— 
2 ves 
“ih 
oo 
¢ Zz / ° 9 


VITAMINES IN URINE 721 


It appeared to us that the guinea-pigs were loth to take the autoclaved 
milk and it was necessary to give an additional quantity with a syringe, which 
took a great deal of time. 

Experiment II. Four guinea-pigs (Nos. 12-15) were given with a syringe 
10 cc. of autoclaved milk besides the usual scorbutic diet. Just as had been 
demonstrated by Delf, it was found that the loss of weight had increased 
(about 34 %). The average lifetime was about the same (25 days), and the 
microscopical changes were very similar in both experiments (Fig. 2). 

Expervment IIa. Another series of animals (Nos. 8-11), fed on the same 
standard diet, were also given 10 cc. of fresh urine daily. The average 
loss of body weight was 31 % and the average lifetime was 26 days. The 
microscopical abnormalities were not less than those in the above experiments. 

Experiment II b. In this experiment the amount of urine was increased; 
20 cc. being administered instead of 10. The average loss of weight of this 
group of animals (Nos. 16-19) was 26 %, and the average lifetime was 23 days 
(Fig. 3). On histological examination we found marked disorganisation of the 
normal structure. In this experiment too the urine proved to be without 
antiscorbutic power. (Although the administration of a large quantity of 
autoclaved milk had a favourable influence on the weight and the general 
state of the guinea-pigs, we thought it advisable to administer only 10 cc. of 
autoclaved milk to the animals in the following experiments, by which method 
it was still possible to make a comparison with the former groups.) 

As we did not think it advisable to administer a greater dose of urine, we 
tried to find out in the following experiments whether urine was able to 
increase the antiscorbutic power of a small quantity of green cabbage. 

Experiment III, Four guinea-pigs were given besides the diet of oats and 
bran 1 g. of green cabbage. These animals (Nos. 20-23) died in about 33 days, 
the average loss of weight being 32%. The macroscopical as well as the 
microscopical changes were far less than in the former experiments (Fig. 4). 

Experiment III a. 1g. of green cabbage was evidently not- quite enough 
absolutely to prevent the onset of scurvy. We then gave four guinea-pigs 
(Nos. 24-27) in addition 15 cc. of fresh urine. It was found that the average 
lifetime of these last animals, as compared to that of those in Expt. III, had 
increased only two days, the average loss of weight being 24 %. The result 
of the microscopical examination was similar to that of Group III. (We think 
the decrease of the average loss of body-weight was only owing to guinea-pig 
No, 23, the effect of the green cabbage being extraordinarily favourable in 
this case.) Without a doubt this experiment shows that the antiscorbutic 
power of 1g. of green cabbage was great (a well-known fact), but that a dose 
of 15 ce. of urine was unable to exercise any curative effect. 

Experiment IV. A further group of animals (Nos. 28-32) was given, besides 
the standard diet, 2 g. of green cabbage. By this treatment the average loss 
of weight was only 17 % whereas the average lifetime increased to 69 days. 
The influence of this quantity of cabbage proved to be much greater than the 


‘euLIN JO °00 OZ + YI[TUr *[90yN*e Jo ‘00 OT + 10zvM “‘URIG ‘8}VO *7axT “GI-9T “SON ‘s81d-eoutns) 
"Q IT quowmodxg “g “31 


N. VAN DER WALLE 


722 


‘Cor 
~~ 
~ 
“—, 
— 
~ nig 
one on 
eee” ieee : ee Se ee 
iy PTA , ENE P ~~ sisi 
~™ — 
* ——— 5 ® 
re 
* x ee. vf wi ie oe 
* — ——— * ° 
2 aes 7s SS hee: 
* + 
a at 
nae 3 & 
Tt lp eee a* 
oF | 
vefels 
bog 
£ z é ° 


723 


‘ouLIn jo 99 cI+ ‘é “cc + “ce “ec Ae “é “ec “cc "LOPS "SON, ‘sSid-eoums 
*yyrut *Pooyne jo "90 OI+ ‘é “cc + ‘é “ “ "eZ ‘ON ‘Sid-eoumns) 
adeqq evo Users Jo “3 [+ 10}eM ‘UBIG ‘s}yVO 27917 “ZZ—-OZ “SON ‘Ss1d-voumny 

‘D TIT pur [IT syucupodxy “fF “S17 


VITAMINES IN URINE 


| ’ Oo 
} 
+: 
hyo roe 
TING 
= a a 7 = 
~ | if | 2 4% 
aid dite Se | | ¥ x 
» 4; <2 ¥ 
\Q ae 4 cs PN A ee he ees 
: pee ie Ce eee 
. SS 4. | o ° ale ~ = + * és 
. aw + SS er sz Pi SS Pd x 
“ST <a, ~ wll at? vog 
“—_ Satie x pa i . 
grr roe Wetman oe oe pales Hr = 
Sie or ok ee ge 
ag aa lt ke 
| 7 
O% 
{ 
= | 
| | 
ee 
th 
| j | | 1 70g 
2 9 Ss 4 . z ¢ 0 


N. VAN DER WALLE 


724 


*yprur *fooyne Jo *00 OT + 


ee 


oh 


“< 


‘Zé ‘ON “Sid-eourny) 
‘eSevqqeo UeeI3 Jo “3 T+ 107VM ‘UBIG ‘S}VO 291g “*[E-8Z “SON ‘st1d-eoumny 
"AT quoupredxg *¢ “Sry 


go 


) | 
\ 
+~ ° 
=| 
a xe _ So! ad: | —_— |? 
a es pe “oe res Meee = 
—_ 
oe gz * ie cat 
-_- -_ % + Dh 
tE ae | 
yz a Pee Suk 
4 i 5 > F 
* } 
. ‘ 
N“ Kee 
= | ‘\ 
X T 
Rais | 
i 2 
i \ | er | “ 
| | [ 
| I | | 
) 
! 
| I 
$s hh vy ba “ % 6 ry ) 9 Ss 4 ¢ ? ‘ 2 


‘00 


00} 


On 


oo¢ 


VITAMINES IN URINE 725 


effect of giving a combination of 1 g. of cabbage plus 15 cc. of urine (Fig. 5). 
The microscopical changes found post-mortem were approximately those 
described by Tozer in the incipient stage. 

In none of these experiments were we able to detect any antiscorbutic 
power of the urine. Of course it is possible that the urine might prove to 
be effective if given in greater quantities. This was practically impossible. 
Theoretically one might try to concentrate the urine just as Harden and 
Robison [1920] concentrated orange-juice and Bassett-Smith lemon-juice. We 
did not think it advisable to do this with urine because of the unknown 
influence of the other substances in urine on the labile antiscorbutic vitamins. 
Harden and Zilva [1918] demonstrated that charcoal does not adsorb the 
antiscorbutic vitamin and so we could not use this method of procedure, which 
proved so successful in the case of the antineuritic vitamin. 


CONCLUSIONS. 


1. It was shown that concentrated, normal urine may exercise a favourable 
influence on polyneuritis of pigeons. 

2. Urine, dried at 37° for 16 hours, was likewise active. 

3. By heating the urine at 130° for one hour, the curative power was 
destroyed. 

4. The ash of urine was inactive. 

5. Carbo anum. puriss., shaken up with urine, centrifuged, washed with 
distilled water and then dried, had a strong curative power. 

6. Urine, treated with a surplus of charcoal and then filtered, had lost its 
curative power. 

7. Charcoal, shaken up with urine, which had been heated previously for 
three hours at 120°, was far less active. 

8. Charcoal not shaken up with urine proved to be quite inactive. 

9. Charcoal, shaken up with a solution of inorganic salts, containing the 
cations and anions present in urine, had only a slight effect. 

10. The curative power exercised by urine may not be attributed to the 
presence of bacilli. 

11. Yeast, cultivated in urine, to which had been added 5 % of glucose, 
exercised a strong curative action. 

12. The urine of a dog had as strongly curative an action as that of man, 
but when the dog was fed on a vitamin-free diet, the urine lost its curative 
power. 

13. It was shown that in some cases fresh, normal urine or dried urine, 
when given in large quantities, may have a good influence on polyneuritis of 
fowls. 

14. Charcoal shaken up with normal urine was likewise active. 

15. The charcoal itself had no curative influence on the polyneuritis 
of fowls. 


726 N. VAN DER WALLE 


16. Charcoal treated with urine heated previously for one hour at 130° 
had no effect. 

17. It was pointed out that a small quantity of antineuritic vitamin is 
present in the urine, as proved by the above experiments. 

18. We were unable to detect the presence of the antiscorbutic vitamin 
in urine. 


LITERATURE. 


Chamberlain and Vedder (1911). Philippine J. Sci. 6, 395. 

Cooper (1914). J. Hygiene, 14, 20. 

Curatolo (1920). Il Policlinico. : 

Damon (1921). J. Biol. Chem. 48, 379. 

Delf and Tozer (1918). Biochem. J. 12, 416. 

Eijkman and van Hoogenhuyze (1913). Archiv Schiffs- und Tropenhygiene, 17, 328. 
Eijkman, van Hoogenhuyze and Derks (1922). J. Biol. Chem. 50, 311. 

Funk (1912). J. Physiol. 45, 489. 

—— Die Vitamine (Bergmann, 1914). 

Gaglio (1919). Jl Policlinico. 

Harden and Robison (1920). Biochem. J. 14, 171. 

Harden and Zilva (1918). Biochem. J. 12, 260. 

Jansen (1920). Mededeelingen uit het geneesk. Laboratorium te Weltevreden, 3° serie, 23. 
di Mattei (1920). Il Policlinico. 

Muckenfuss (1918). J. Amer. Chem. Soc. 40, 1606. 

Rona und Michaélis (1919). Biochem. Zeitsch. 94, 240. 

Seidell (1921). J. Ind. Eng. Chem. 18, 1111. 

Theiler (1915). The third and fourth reports of the Director of Veterinary Research, Pretoria, Nov. 
Tozer (1921). J. Pathol. Bact. 24, 306. 


LXXIV. THE SUCTION PRESSURE OF THE 
PLANT CELL. 


A NOTE ON NOMENCLATURE. 


By WALTER STILES. 
From the Department of Botany, University College, Reading. 


(Recewed August 4th, 1922.) 


As a result of the work of Pfeffer and de Vries during the last thirty years of 
the nineteenth century, the vacuolated cell came to be regarded as a system 
consisting of a solution containing osmotically active substances (the cell sap) 
surrounded by a semi-permeable membrane (the protoplasm) permeable to 
water but impermeable to the solutes in the cell sap, the whole being enclosed 
by an elastic cellulose envelope (the cell wall), supposed to be permeable 
both to water and dissolved substances. How far such a view really represents 
the facts is not in question here. 

When such a system is immersed in water, it is to be expected that water 
will enter the cell on account of the osmotic pressure of the cell sap. This 
must result in the development of a hydrostatic pressure and consequent 
stretching of the wall, which, being elastic, exerts an inwardly directed pressure 
on the cell contents. At any time, therefore, there is a pressure, the osmotic 
pressure, tending to send water into the cell, and the inwardly directed 
component of the pressure exerted by the wall tending to compress the cell 
contents and so send water out. The hydrostatic pressure of the water in the 
cell is called the turgor pressure, to which the inwardly directed wall pressure 
must be equal and opposite. If this pressure is denoted by 7’ and the osmotic 
pressure by P, the net pressure sending water into the cell when this is sur- 
rounded by water is thus P — 7, which varies from P when the cell is 
completely deturgid to zero when it has absorbed so much water that T = P 
and the cell can be regarded as saturated with water. 

Just as it is convenient to have the term osmotic pressure to denote the 
pressure developed when a solution is separated from pure water by a semi- 
permeable membrane, so it is convenient to have a term to denote the actual 
net pressure sending water into the cell when this is immersed in pure water. 
For this quantity Ursprung and Blum [1916] have used the term “suction 
force” (Saugkraft) and there appears to be a tendency for other continental 
writers to adopt this. The chief objection to this term is that the quantity 
is a pressure and not a force, and is equated with a pressure (osmotic pressure). 
In this country Thoday [1918] has proposed the term “ water absorbing power” 


728 W. STILES 


of the cell for this quantity. This term appears particularly unfortunate, for 
not only is it perhaps a trifle clumsy, but if the quantity in question is not 
a force, still less is it a power, a rate of doing work. Nevertheless there is a 
tendency for the term to be accepted by English writers, one of whom indeed 
gives an equation in which a quantity defined as a surface tension (a force 
per unit length) is put equal to a quantity defined as a pressure (a force per 
unit area) less a quantity defined as a power (work per unit time). 

To avoid such anomalies in nomenclature, I propose for this quantity, 
already described as a force and a power, but which is in reality a pressure, 
the term “suction pressure.” This gives an accurate description of the 
quantity, is not clumsy, and practically embodies the term used by Ursprung 
and Blum, whose expression, other things being equal, is entitled to pre- 
cedence on the ground of priority. 


REFERENCES. 


Thoday (1918). New Phyt. 17, 108. 
Ursprung and Blum (1916). Ber. deut. bot. Ges. 34, U39. 


LXXV. RESPIRATORY EXCHANGE IN FRESH 
WATER FISH. 


PART IV. 
FURTHER COMPARISON OF GOLD-FISH AND TROUT. 


By JOHN ADDYMAN GARDNER anv GEORGE KING. 
From the Physiological Laboratory, University of London, South Kensington, 
(Recewed August 21st, 1922.) . 


In Part IT of this series of papers [1914, 2] the behaviour of trout at various 
temperatures under reduced oxygen tension was described, and measurements 
were recorded of the quantity of oxygen in the water at the asphyxial point 
for the various temperatures. It was found that the oxygen content of the 
water at the asphyxial point was much larger at higher temperatures than at 
lower, a result to be expected since, as was shown in Part I [1914, 1], the 
oxygen consumption of trout in fully saturated water increases with the rise 
in temperature—within the ordinary limits of temperature of natural water 
consumption doubles for rise of 10°—and consequently the fish would have 
to pump & much larger volume of water through their gills at a higher tem- 
- perature than at a lower to obtain the necessary amount of oxygen. 

Gold-fish as we showed in Part III [1922] require much less oxygen than 
trout and live at an altogether lower plane of metabolism, though their oxygen 
requirements are also proportional to temperature. It seemed of interest to 
measure the oxygen tension at the asphyxial point at various temperatures in 
the case of these animals. 


MeErTHop. 


The apparatus described in Part II proved unsuitable for gold-fish, as we 
found it impossible to reduce the oxygen in the water to the asphyxial tension 
within a reasonable time. We therefore used a much smaller apparatus based 
on the same principle, and filled with water which had previously been heated 
in order to get rid of the bulk of the dissolved gas and allowed to cool in a 
closed bottle. A six litre vacuum desiccator with slightly domed lid served 
for the water tank. This was filled to the edge of the ground flange with the 
water so that the gas space above the water was only that of the shallow 
dome of the lid. The tubulure of the lid was closed by a rubber stopper 
carrying (1) a tube leading to the bottom of the vessel and ending in a per- 
forated bulb through which nitrogen could be sprayed through the water, 
(2) an exit tube for nitrogen bent over so as to dip into a beaker of water 

Bioch, xv1 48 


730 J. A. GARDNER AND G. KING 


and (3) a capillary tube fitted with a tap through which samples of the water 
could be drawn off as required. The desiccator stood in a thermostat. The 
fish were placed in the tank and the air in the water was gradually displaced 
by a current of nitrogen, which also served to stir the water. The compressed 
commercial nitrogen used contained about 0-5 °% of oxygen; this quantity of 
oxygen was too much, and the gas was purified by passing through several 
wash bottles containing strong alkaline pyrogallol. In this way the percentage 
of oxygen in the nitrogen could easily be reduced to 0-14. A sample of the 
water was withdrawn for analysis at the asphyxial point, which was taken 
‘ as that point at which the fish rolled over on their sides and ceased to breathe. 
The sample was withdrawn by displacement of mais and may by the 
method fully described in previous papers. 

Between temperatures of, say, 10° and 27° the oxygen content of the water 
could fairly easily be reduced to the asphyxial tension by means of a current 
of nitrogen, but at low temperatures it was not found possible to reduce the 
oxygen sufficiently by this method. In experiments at low temperatures the 
oxygen was displaced as far as possible by a current of nitrogen, and the 
apparatus was then closed and the fish made to use up the pan eaay oxygen 
themselves. 


EXPERIMENTAL. 


Experiment at high temperature. 


Two fish, A weighing 53g. and B weighing 58 g., were placed in the 
partially boiled out water at a temperature of about 22° and the tank put in 
the thermostat at 28°; nitrogen was then bubbled through the water. At the 
beginning the fish were very active, A being particularly violent for a gold-fish, 
and attempting to jump out of the water. The respirations of both animals 
were about 117 per minute. As the oxygen tension decreased the respiration 
became more spasmodic, irregular and slow. The movements of the fish 
gradually became less vigorous, and towards the end of the experiment the 
animals took short intervals of rest with scarcely any perceptible movements 
of the gills, alternating with short rushes round the vessel mainly near the 
surface. After about 90 minutes A remained on his side without any obvious 
sign of life and B was nearly in the same condition. A sample of water was 
now withdrawn for analysis, the temperature being 27-2°. The fish were now 
removed from the experimental vessel and placed in a trough of tap water 
at laboratory temperature. After two minutes B recovered his normal 
position, while A began to breathe and showed slight movements after five 
minutes, but still remained on his side. The fish were then placed in the store 
tank in running water and after two to three hours A recovered. Next day 
both fish appeared to be perfectly normal and healthy. On analysis | litre 
of the water was found to contain the following quantities of gas measured 
at 0° and 760:—free and combined CO, 12-25 ce., oxygen 0-422 cc. and nitrogen 


RESPIRATORY EXCHANGE IN FRESH WATER FISH 731 


8-287 cc. Taking Winkler’s value 0-02727 as the absorption coefficient of oxygen 
at 27-2° (Landolt, Bornstein, Meyerhoffer’s tables) the partial pressure of the 
dissolved oxygen was therefore 1-547 % of one atmosphere. 


Experiment at temperature 11-12°. 


Two fish, C 51 g/ and D 40 g., were placed in six litres of partially boiled 
out water as before and purified nitrogen bubbled through the water. Both 
animals were fairly active at first and their respirations 80-90 per minute, 
the normal in fully aerated still water being between 40-50 per minute. During 
the first two hours of the experiment, as the oxygen gradually decreased, the 
fish became less continuously active, the number of respirations decreased to 
40-50 per minute, and the breathing, if we can use this term, became shallower. 
A noticeable feature was that under these low oxygen tensions the fish indulged 
in spells of backward swimming. After 135 minutes the nitrogen current was 
stopped and the fish reduced the oxygen content of the water naturally. The 
breathing became shallow and very spasmodic and the respiration rate was 
reduced to 20-30 per minute. At the end of the fourth hour it was noticed 
that the fish remained inert for periods of 1-14 minutes. This was followed 
by a few respirations then another period of inertness and so on. After six 
hours both fish were on their sides and seemed to have ceased to respire, a 
sample of water was therefore withdrawn for analysis. The temperature was 
11-4°. One litre of the water contained 10-732 cc. free and combined CO,, 
0-397 cc. oxygen and 15-362 cc. nitrogen measured at 0° and 760. Taking the 
coefficient of absorption for oxygen in water at 11-4° as 0-03686, the partial 
pressure of the oxygen was 1-077 % of one atmosphere. The fish gradually 
recovered on being restored to the store tank. 


Experiment at low temperature. 


It was not found possible at the low temperature of the ice chest to reduce 
the oxygen tension sufficiently, by means of a current of nitrogen, to in- 
commode, let alone asphyxiate, the fish. After reducing the tension as far as 
possible by this method we had to rely on the fish to use up the remaining 
oxygen, and this proved a very slow process. 

Two fish, # 52 g. and F 42 g., were placed in 54 litres of boiled out water 
and a rapid current of nitrogen bubbled through for two hours on March 27th. 
The nitrogen inlet and outlet taps were then closed and the vessel placed in 
the ice chest and left for a week. On the second day the fish appeared to be 
quite normal and during the week they remained very quiet and sluggish 
sometimes near the surface and sometimes at the bottom of the vessel. On 
April 5th they showed a rolling gait symptomatic of distress. On the morning 
of April 6th F was more or less on his side and in the afternoon appeared to 
be dead or nearly so. # was still breathing occasionally in a spasmodic manner 
and had a rolling gait as though finding difficulty in keeping upright. The 
temperature recorded by the thermometer in the upper part of the chest 


48—2 


a 


732 J, A. GARDNER AND G. KING 


varied from 3-5° during the week. A sample of water was withdrawn for 
analysis at 2.35 p.m. on April 6th; its temperature was 2-6°. The quantity of 
oxygen in the water was less than the limit of the experimental error. On 
analysis one litre of water was found to contain the following volumes of gas 
reduced to 0° and 760:—free and combined CO, 96-96 cc., oxygen nil, nitrogen 
22-97 ce. 


Comparison with trout. 


Trout normally use very much larger quantities of oxygen per kilo per 
minute than gold-fish, and at a diminishing oxygen tension behave quite 
differently. The behaviour of trout was described in detail in Part IT [1914, 2, 
p. 595]. As the oxygen diminished the respiratory movement gradually in- 
creased in rapidity and force and acquired a dyspnoeic character. The animals 
became more active, rushing about and jumping into the air space above the 
water. These periods of activity alternated with periods of rest at the bottom 
with the dyspnoeic respiration mentioned. The periods of rest gradually 
became longer than the periods of activity, and as the oxygen approached the 
asphyxial tension they remained nearly all the time at the bottom and ulti- 
mately quietly rolled over on their backs. In this condition they died in two 
or three minutes, but if oxygen were at once sprayed into the water they re- 
covered in the course of a few minutes and afterwards appeared to suffer no 
ill effects. The asphyxial condition in the case of trout takes place within 
quite small limits of oxygen tension; at or below this tension they very quickly 
die but slightly above they recover. Gold-fish on the other hand could stand 
much lower oxygen tensions for any given temperature and are much less 
sensitive, so that the precise asphyxial tension is difficult to ascertain. This 
difference is still more marked at low temperatures. In Table I we give a 
comparison of the measurements for trout and gold-fish. 


Table I. 
8” trout, weight 80-90 g. 5” to 6” gold-fish, weight 
Oxygen in water at 45-55 g. Oxygen in water 
asphyxial point* at asphyxial point Normal water 
we —- ——_—-—---r.. 
ec, of oxygen partial pressure cc, of oxygen partial pressure 
measured at of oxygen in measured at ofoxygenin ce, of oxygen rtial pressure 
0 and 760 per per cent. of 1 0° and 760 per percent.of1 at 0° and 760 in percent. of 1 
Temperature litre of water atmosphere litre of water atmosphere per litre atmosphere 
2-6° — — nil nil 9-49 
64° 0-79 1-89 — _ 8-60 
9-5 to 10° 0-81 2-19 a — = 8-00 
114° —_ — 0-397 1-077 7-9 
17° 1-37 417 —_ -- 6:75 app. 21 % 
18” 1-49 4-62 —_ —_ 6-61 
24° 1-97 6:83 _ —_ 5:89 
25° 2-40 849 — a 5:78 
27-2° - _ > 0-422 1-547 5-52) 


* Gardner and Leetham [1914, 2). 


3 


RESPIRATORY EXCHANGE IN FRESH WATER FISH 733 


GLYCOGEN CONTENT OF TROUT AND GOLD-FISH UNDER 
VARIOUS CONDITIONS. 


In Part I [1914, 1] of this series it was shown that at low temperature, 
v.e. below about 6°, trout appeared to be in a condition, for a time at any 
rate, akin to that of hibernating animals, and their respiratory quotients 
were very low. These low respiratory quotients could not be directly con- 
nected with nitrogen metabolism, and it was suggested that a possible ex- 
planation might be found in the fat. If at low temperature the animals were 
in a state of hibernation or starvation they might be living on their fat and 
partially converting it into glycogen and sugar. Some such process might 
result in a low respiratory quotient, as indicated in the following equation: 

2C'sH,(CysHg30z)s + 640, = 16C,H,,0, + 18CO, +8H,0 
CO,/0,=43 =0-281. 

In the hope of throwing light on this point comparative estimations of 
the glycogen content of starving trout at various temperatures were made. 
The glycogen was estimated by Pfliiger’s method. The precipitated glycogen, 
after solution and reprecipitation two or three times, was hydrolysed and the 
sugar estimated by Fehling. If in sufficient quantity the cuprous oxide was 
weighed, otherwise it was dissolved in ferric sulphate and the ferrous salt 
produced estimated by permanganate. The first experiments were done on 
four-inch trout which however proved too small for the purpose. 


EXPERIMENTAL. 
Experiments with four-inch trout. 

(1) Control. Four fish weighing 44-65 g. were kept in running water in a 
large tank with a gravel bottom. The average temperature of the water 
was 19°. The fish were not fed though they may have picked up some food. 
After treatment by Pfliiger’s method and hydrolysis of the glycogen 0-0013 g. 
Cu,0 was obtained. Percentage of glycogen was therefore 0-0052. 

(2) Four fish weighing 40-6 g. were placed in an open vessel in the ice 
chest and the water kept well oxygenated by a spray of oxygen. At the end 
of the first day the temperature of the water was 7°, at the end of the second 
day 6° and the third day below 5°. The fish were very inert and sluggish and 
scarcely moved when touched. On the afternoon of the third day they were 
killed and analysed. 0-0032 °% of glycogen was found. 

(3) Five fish weighing 57-5 g. were kept in the ice chest under the same 
conditions as in experiment 2 for six days. On the last day two died and as 
soon as this was observed the rest were killed and the whole analysed. The 
final temperature of the water was 3-5°. The percentage content was 0-0022. 

(4) Three fish weighing 27-8 g. were kept in the ice chest for two days at 
reduced oxygen tension. They were then analysed but no glycogen could be 
detected. 

As the fish were too small to give a suitable weight of material without 
using an inconvenient number of fish the experiments were repeated using 
eight-inch animals. 


734 J. A. GARDNER AND G. KING 


Experiments with eight-inch trout. 


(5) Control experiments. The fish were kept in a large outside tank in 
running water for*several days. They were not fed and the temperature of 
the water was about 16°. On analysis, fish A weighing 89-1 g. gave no trace 
of glycogen, fish B weighing 109 g. was found to contain 0-0012 %. 

(6) Two fish were placed in separate earthenware bowls in the ice chest 
and a spray of oxygen passed through the water by means of circular lead 
pipes punctured with small holes. The fish were kept at 2-6° for two days; 
they were very sluggish but otherwise appeared healthy. Fish C weighing 
114-3 g. contained no trace of glycogen nor did fish D weighing 134-3 g. 

(7) Two fish were kept under similar conditions to those in experiment (6) 
for five days. Final temperature of the water was 1-7°. On the fifth day 
one of the animals turned over and they were therefore cut up and analysed. 
Fish £ weighing 101-3 g. gave 0-1179 g. of Cu,O and contained 0-0469 % 
glycogen. Fish F weighing 108 g. gave 0-0798 g. Cu,O and hence contained 
0-0294 % glycogen. ? 

We do not know what is the limit of accuracy of Pfliiger’s method for 
estimating glycogen, but as the conditions—proportion of alkali, duration of 
hydrolysis, volume of alcohol—were kept as nearly as possible the same in 
experiments 5, 6 and 7 it is clear that the fish in experiment 7 contained a 
very much larger percentage of glycogen thanthe trols “*~ -~ 
to fit in with our hypothesis as to the cause of the :ow respira.vry quouents at 
low temperature. At the same time we cannot pretend that the experiments 
are altogether conclusive. 


GLYCOGEN CONTENT OF GOLD-FISH. 


It seemed of interest to ascertain how glycogen content of gold-fish, which 
did not show abnormal quotients at low temperatures, compared with that 
of trout. 

(8) Control experiments. Two gold-fish were taken from the running water 
tank in which they had been living many months. The tank contained plenty 
of food. Two fish weighing 73 g. analysed together were found to contain 
0-86 %, of glycogen. Temperature of the tank when the fish were caught 
13-14°. Two fish weighing 90-5 g. taken from water at about 16° contained 
1:57 %. Two fish weighing 61-2 g., temperature of tank 18°, contained 1-235 % 
glycogen. 

(9) Two fish weighing together 73-3 g. were kept in an open vessel in the 
ice chest for seven days; the water was aerated by a current of oxygen and 
the final temperature was 2~3°. The fish were not fed, and on analysis they 
contained 1-21 %. 

(10) In this experiment two fish weighing 81-8 g. were placed in the 
apparatus used for determining the asphyxial point and the oxygen tension 
considerably reduced by bubbling nitrogen through the water. The apparatus 


RESPIRATORY EXCHANGE IN FRESH WATER FISH = 735 


was then placed in the ice chest and kept at about 2° for seven days. Per- 
centage of glycogen was found to be 0:52. 

(11) Two fish total weight 61-6 g. were placed in a trough of ordinary 
tap water through which oxygen was bubbled occasionally and kept in the 
ice chest for eleven days; the fish were not fed. Percentage of glycogen was 
found to be 0-54. 

(12) Two fish weighing 60-45 g. were placed in six litres of partially boiled 
out tap water in the asphyxial point apparatus referred to above. Pure 
nitrogen was passed through the water for half an hour when the nitrogen 
tubes were closed and the apparatus placed in the ice chest and left for 
several days. The average temperature of the water was 3-7°. On the morning 
of the fifth day one fish was found on its side but the others were still alive 
and in normal position. They were killed and the whole analysed and found to 
contain 0-024 % glycogen. 

These results appear to be in conformity with the low level of metabolism 
of these fish compared with trout. 


We take this opportunity of expressing our thanks to the Government 
Grants Committee of the Royal Society for aid in carrying out this work; and 
also to Mr A. R. Peart of Hungerford who supplied the trout. 


Nant yit't bi REFERENCES. 


or - «= ./@ Gardner arld Leetham (1914, 1). Biochem. J. 8, 374. 
—— (1914, 2). Biochem. J. 8, 591. 
Gardner, King and Powers (1922). Biochem. J. 16, 523. 


LXXVI. RESPIRATORY EXCHANGE 
IN FRESH WATER FISH. 


PART V. ON EELS. 


By JOHN ADDYMAN GARDNER anp GEORGE KING. 


From the Physiological Laboratory, University of London, 
South Kensington. 


(Received August 25th, 1922.) 


EELS are said to be peculiarly averse to cold, and the fact that the tem- 
perature of the brackish water of estuaries is usually higher than that of 
unmixed salt or fresh water has been advanced as one of the reasons for their 
seaward migration on the approach of winter. During the cold of winter 
these fish lose their appetite and become torpid, large numbers of them con- 
gregating together for the sake of the additional warmth thus obtained, and 
burying themselves to a depth of many inches in places where the receding 
tide leaves them more or less dry. 

In the eel the bronchial openings are small and lead into a sac, from which 
another sac is given off. The gills are thus exposed only slightly to the drying 
influence of the atmosphere, and it is believed that it is owing to this and 
to the slimy condition of the skin, that eels are able to exist for a very con- 
siderable time, compared with other fish, out of water. They can evidently 
exist at a low plane of metabolism, but like most animals which pass the 
winter in a torpid condition, are relatively voracious during the summer 
months. 

For these reasons it seemed of interest to study the respiratory exchange 
of these animals under various conditions of temperature, and compare them 
with other fish, particularly trout, which, as we have shown, are also sluggish, 
for a time at any rate, on exposure to cold. 

The fish used in these experiments were the common Anguilla vulgaris and 
were 15 to 18 inches in length. They were kept for several days in a large 
tank with gravel bottom in running water before being used for experiment. 
The tank contained a number of small worms and the like, which the eels 
could have picked out of the gravel, but whether they did so we have no 
means of judging. The fish however were in good condition, 

The method adopted was that fully described in Part I [1914] and Part III 
[1922] of this series. 


RESPIRATORY EXCHANGE IN FRESH WATER FISH = -737 


EXPERIMENTAL. 


At high temperatures. 


(1) Six fish, weight 940 g., were -placed in a tank of water at_the tem- 
perature of the outside tank—about 15°—and gradually warmed up to 20° 
before being placed in the experimental bottle. The initial temperature of 
the water in the bottle was 21-1° and the final temperature 23-5°; the duration 
of the experiment was 5 hours 4 minutes, and during this period 275 ce. 
of commercial oxygen were added (measured at 16-5° and 760 mm.) to the 
air above the water. Analyses of the air and water at the beginning and end 
of the experiment gave the following results in cc. reduced to standard tem- 
perature and pressure. 


_ Initial vol. Final vol. Difference 
Free and combined CO, 1827-4 2023-0 + 195-6 
Oxygen 4102-3 3810-2 — 292-1 
Nitrogen. 14693-1 14721-4 + 28-3 


Nitrogen error +0-1990 %. 


At the end of the experiment the water (24619 cc.) had a strong faecal 
smell, and contained a good deal of dejecta. The eels at this high temperature 
were very active and difficult to catch by hand. 


At medium temperatures. 


(2) Five fish, weight 650g. Initial temperature of water 16-5°, final 
temperature 16-5°. Duration of experiment 5 hours 45 minutes. Oxygen was 
added as before. The fish were lively. 


Initial vol. Final vol. Difference 
Free and combined CO, 1821-0 1950-1 +129-1 
Oxygen 4147-5 3981-5 — 166-0 
Nitrogen 15982-2 15981-8 -— 0-4 


Nitrogen error —0-0025 %. 


(3) Five fish; weight 725 g. Initial temperature 11-7°, final temperature 
8:4°. Duration 3-44 p.m. to 12.10 p.m. = 20 hours 36 minutes. Oxygen was 
added, and pumping was continuous except for a few hours owing to a slipped 
belt. 

Initial vol. Final vol. Difference 


Free and combined CO, 1552-9 1828-8 +275-9 
Oxygen 4868-8 4548-4 — 320-4 
Nitrogen 17594-7 17607-8 + 131 


Nitrogen error +0-074 %. 


(4) Four fish, weight 680 g. This experiment was commenced at 11.37 a.m. 
and continued until 12 noon on the following day—24 hours 23 minutes. 
The initial temperature of the water in bottle was 16-6° and the air 16-7°. 
After about three hours the water in the thermostat was packed with ice, 
and the temperature of the thermometer in the air of the bottle was noted 
at intervals, the final temperature of the water was 6-4°. It was estimated 


738 J. A. GARDNER AND G. KING 


that the average temperature was between 9° and 10°. Oxygen was added 
during the experiment as before. During the night the pumping was inter- 
rupted for a few hours owing tothe slipping of a belt, probably in the early 
morning owing to.change over of the dynamos in the power station. 


Tnitial vol. Final vol. Difference 
Free and combined CO, 1650-1 1949-8 + 299-7 
Oxygen 4655-4 4371-7 — 283-7 
Nitrogen .16875-8 16903-4 + 27-6 


Nitrogen error + 0-16 %. 


At low temperatures. 


(5) For this purpose the eels were kept in well oxygenated water in the 
ice chest for 24 hours before the experiment, the average temperature being 
between 3° and 4°. The animals were very inert and sluggish and scarcely | 
moved when handled. Five fish, weight 725 g. Initial temperature of water 
6-5°, final temperature 5-3°. Duration of experiment 24 hours 8 minutes. 
Oxygen was added during the experiment and the pumping was continuous. 


Tnitial vol. Final vol. Difference 
Free and combined CO, 1343-5 1414-3 + 70-8 
Oxygen "5559-5 5397-2 — 162:3 
Nitrogen - 20211-6 20211-5 - Ol 


Nitrogen error —0-00049 %. , 
In all the above experiments except that at high temperatures the water 
remained clear and no unpleasant smell was noticed. The fish remained 
perfectly healthy after the experiments. The results are summarised in the 
following table. 


Oxygen Oxygen 


K aeet absor CO, i rch CO, cap hia 
weight of oe hour Bata or hott per Dose 
No. of Average fish used in ce. at at 0° and at 0° and at 0° and Respiratory 
experiment temperature in g. 0° and 760 760 760 760 quotient 
(1) 22-2° 157 9-60 61-29 6-43 41-04 0-67 
(2) 16-5° 130 5-77 44-41 4-49 34-54 0-78 
(3) 10° 145 3-11 21-45 2-68 18-47 0-86 
(4) about 9° 170 ae + ') | 17-11 3-07 18-08 1-06 
(5) 5-9° 145 1:35 9-28 059 405 0-44 


It will be noticed that the oxygen consumed per fish, and per kg. of 
fish per hour is proportional to the temperature and if the oxygen is plotted 
against temperature the values are approximately on a straight line. 

On comparison with the figures given in Part I [1914] for 8-inch brown 
trout it is evident that eels live at a much lower plane of metabolism than 
trout. At medium temperatures, about 16°, trout use about four times as much 
oxygen as eels, and at low temperatures 10 to 12 times as much. 

The respiratory quotients in the first four experiments are approximately 
normal but at low temperature, at which the animals are in a torpid con- 
dition, the value falls much below normal. In this respect their behaviour 
recalls that of trout. 

REFERENCES. 


Gardner and Leetham (1914). Biochem. J. 8, 374. 
Gardner, King and Powers (1922). Biochem. J. 16, 523. 


LXXVII. ON A SERIES OF METALLO-CYSTEIN 
DERIVATIVES. I. 


By LESLIE JULIUS HARRIS. 
From the Biochemical Laboratory, Cambridge. 


(Report to the Committee of Scientific and Industrial Research.) 
(Received September 14th, 1922.) 


WueN a stream of air is blown through an alkaline solution of cystein, the 
oxidation is generally accompanied by the production of a violet coloration; 
an observation that will be familiar to workers who have had occasion to 
prepare cystine in this manner. Although the phenomenon has been long 
known its cause has remained obscure, and the present investigation was 
undertaken primarily with the object of ascertaining its nature. 

The fact that cystein gives with ferric chloride and ammonia an intense 
coloration, which becomes darker on shaking, was first noticed by Andreasch 
[1884]. Arnold [1911], describing a number of tests by which cystein may be 
recognised, distinguishes between the colour produced with ferric chloride and 
that with ammonia. “Ifa solution of cystein is made weakly alkaline a violet 
colour is produced which is more permanent than that given by sodium 
nitroprusside....” There can be little doubt however that the coloration is 
in reality due not so much to the action of the alkali upon the cystein, as 
would appear from Arnold’s account, as to the presence of traces of a metallic 
(generally ferric) salt, which produce the typical coloration in alkaline solution. 
Taking care to ensure the complete absence of a metallic ion I have been 
unable to observe the production of the slightest colour. 

An important feature of the reaction is its impermanent character. Arnold 
observes that the colour fades after a while, but can be renewed by shaking. 
The nature of this curious alternation of fading and restoration of colour 
receives no explanation from Arnold. 


The Reversible System: Cystein—Iron—Ammonia. 


When cystine is isolated by the use of mercuric sulphate reagent, the 
Hg-cystine compound which is precipitated, yields, on decomposition with 
H,S, a solution of the reduced cystein. This can be re-oxidised, after removal 
of H,S, by making alkaline with ammonia and blowing a stream of air through 
the solution. Under these circumstances there is generally enough iron present 


740 L. J. HARRIS 


as impurity to produce a noticeable violet coloration in the solution as soon 
as the active oxidation is started. Often a state of delicate equilibrium is set 
up in which the colour disappears upon the air current being stopped, and 
immediately reappears upon its continuation. The necessary conditions may 
be reproduced artificially by adding to a test tube containing (say) 4 cc. of 
a 5 % eystein solution a single drop of weak FeCl, solution. Upon making 
alkaline with weak NH,OH, a violet colour appears, only to fade almost _ 
immediately. The colour reappears if the tube be shaken, or if air (or, better, _ 
oxygen) be otherwise brought into intimate contact with the solution. Bubbling 
hydrogen, carbon dioxide or nitrogen through the liquid has no effect. 

Variation in the concentrations has a visible effect on the course of the 
phenomena. The depth of colour produced is directly proportional to the 
quantities of both cystein and ferric chloride present together. Again, the greater 
the concentration of ferric chloride relative to that of cystein, the longer does 
the violet colour survive the cessation of the air current, or, to look at it in 
another way, increasing the proportion of cystein in the mixture, brings about 
a more rapid fading when the air current is stopped. The solution remains 
violet (or alternatively capable of yielding a violet colour on aeration) so long 
as unoxidised cystein is present; the presence of the latter can be demon- 
strated by the extremely delicate nitroprusside test?. When oxidation is 
complete the solution consists of cystine together with a minute quantity of 
Fe(OH),, derived from the added iron salt. The effect of increasing the 
alkalinity is to yield a solution which less readily decolorises in the absence 
of the air current, but in which complete and permanent loss of colour (or 
ability to produce it on addition of the iron salt) occurs more rapidly, owing 
apparently to the greater rapidity of the transformation cystein —~ cystine in 
the more alkaline solution. 

Numerous ferric salts, also suspended ferric hydroxide, have been investi- 
gated as substitutes for FeCl,. All were equally efficient. Potassium ferri- 
cyanide was without action. 


Explanation of the Phenomena. 


Several explanations might be adopted to account for the observations 
described above. At first sight it would seem likely that the colour is due to 
the formation of an unstable compound in a state of oxidation intermediate 
between cystein and cystine; its presence being made manifest as a result of 
its power to yield a coloration with ferric salts. 

Secondly, it may be assumed that direct oxidation of cystein to a body of 
the cystine peroxide type occurs, and that this hypothetical substance is 
reactive to FeCl,. The consecutive reaction is between cystein and the 
peroxide to yield the end product, cystine. There are theoretical grounds for 
postulating the possible existence of such a peroxide, _ 


* It is possible to fix approximately the completion of oxidation by the disappearance of the 
violet coloration, 


A SERIES OF METALLO-CYSTEIN DERIVATIVES 741 


In view, however, of similar reactions between cystein and salts of certain 
other metals, notably copper, the following explanation would appear to fit 
the facts more closely than either of these. 
Cystein is an acid, and it may be supposed that in the presence of Fe™ 
ions, and in alkaline solution, it forms a violet-coloured ferric derivative. 
The main characteristic of cystein is its reducing power: 


2COOH—CH(NH,)—CH,—S|H + 0 —-> (COOH—CH(NH,)—CH,—S—), + H,0. 


_ If excess is present and the supply of air be limited the cystein is oxidised 
to cystine at the expense of the ferric ion, which is thereby reduced to the 

ferrous state. The latter yields no coloured derivative with cystein, and the 
solution therefore fades. If now air be blown through the mixture, the Fe™ 
is rapidly re-oxidised to Fe", Fe’ as is well known being unstable in alkaline 
solution when available oxygen is present. The colour is therefore regenerated 
provided some unoxidised cystein remains to form the anion of the coloured 
body, and it will be retained until the process of oxidation is complete. In 
the presence of abundance of free oxygen the cystein will be oxidising con- 
tinuously, and the iron will be maintained in the ferric condition. 

_ This explanation accounts for the following experimental observations: 

(1) The intensity of colour on making alkaline with ammonia is propor- 
tional to the amounts of cystein and Fe salt present together (7.e. the amount 
of cystein iron derivative produced). 

(2) Excess of cystein causes a more transient coloration (since the Fe’ 
is more rapidly reduced in the presence of excess of the reducing agent). 

(3)) Excess of iron causes a more stable coloration. 

(4) Excess of alkali causes the coloration to be less transient (since’Fe™ 
is increasingly unstable in alkaline solution). 

Apparently the reduction potential of cystein does not develop with in- 
creasing alkalinity so rapidly as that of Fe(OH), —+ Fe(OH). 

(5) With increasing alkalinity the total duration of the colour phenomena 
is less, since the cystein disappears more rapidly. (Cystine becomes more 
readily autoxidisable with increase of py.) 

(6) It is found that the addition of free cystein to the coloured solution 
aids the (temporary) decolorisation. (Explanation as in (2).) 

(7) It is a well known experimental fact that iron acts as a catalyst in 
accelerating the oxidation of cystein to cystine. This is readily explained on 
the basis of the above theory: the iron acts as a carrier; in the ferrous state 
it is a more efficient oxygen acceptor than the cystein, and oxidised it 
functions as an oxygen donator. 

(8) The most convincing support to these views is afforded by the be- 
haviour of copper (vide infra). It is found moreover that reversible colorations 
analogous with that given by iron could be obtained only by those metals 
which exist in more than one state of oxidation, and in which change of 
valency from the one form to the other can occur with tolerable ease. 


742 L. J. HARRIS 


With regard to the catalytic action of iron (see (7) above) mention should 
be made of the work of Mathews and Walker [1909] on the velocity of spon- 
taneous oxidation of cystein in the presence of salt solutions. This paper was 
published prior to that of Arnold and I was unaware of the explanation therein 
advanced of the mechanism of the catalysis when the work described in the 
present paper was carried out. 

“An explanation is given” by Mathews and Walker “of the action of 
different metals based on their solution pressure.” These authors account for 
the catalytic action by assuming the momentary formation of an intermediate 
metallic compound 


“MeO + Cystein — Cystein . MeO. 
Cystein. MeO = Cystin + Me + H,0.” 


It would seem that this view fails to account for the experimental obser- 
vation that the coloured derivative is only instantly decomposed when excess 
of free cystein is present. The pure cupric cystein derivative can be isolated 
and shows no tendency for internal spontaneous scission such as is suggested. 

The action of cystein upon the cupric and other coloured metallic de- 
rivatives, can leave little doubt that the mechanism of the reaction consists 
in the reduction of the oxidised metallic atom of the derivative by the action 
of excess of cystein. 


The formation of various metallic derivatives of cystein. 


Fe’. The behaviour of iron is explained above. The deep coloration 
obtained when (excess of) ferric chloride is added to cystein was first described 
by Andreasch, and then by Arnold. The latter author distinguished between 
the colour given by FeCl, and by “ammonia.” With sufficiency of iron the 
colour remains until oxidation of cystein to cystine is complete, and cannot 
then be regenerated by aeration. 

Fe’. Under ordinary conditions Fe(OH), cannot be precipitated suffi- 
ciently free from Fe(OH), to prevent the production of the violet ferric com- 
pound. Investigations have been started under anaerobic conditions. 

Double salts (ferric alum, ete.) in which iron functions as Fe’ act like 
FeCl, or Fe(OH),. When iron functions in the complex anion as in K,FeCy, 
or K,FeCy, it is without action on the cystein anion. 

Mn” and Mn”. If a few drops of manganous sulphate be added to cystein 
(in HCl solution) and the liquid made alkaline with ammonia a fine green 
coloration (Mn-cystein cpd.) appears. It rapidly fades provided sufficient 
cystein be present. On blowing air through the faded solution the colour is 
regenerated, The general behaviour of manganese towards cystein resembles 
that of iron described at length above. 

It appears probable that the green body is the Mn” derivative and the 
decolorised solution contains Mn”. The addition of alkali to a manganous 
salt yields Mn(OH),, which however rapidly goes to Mn(OH), or MnO 3, H,O 


A SERIES OF METALLO-CYSTEIN DERIVATIVES 743 


unless air be excluded. When cystein is also present the green colour is given 
in air, while in absence of air the solution remains colourless (Mn"*). When 
oxidation of cystein has reached completion brown Mn,0O,, H,O remains. The 
addition of fresh cystein to this yields the green sernynn re and the sequence 
can then be repeated afresh. 

Thunberg [1913] has found that in the presence of MnCl, the autoxidation 
of various thio-compounds is accelerated, a fact which is readily accounted 
for by a cycle such as the above, in which the salt functions as an oxygen 
carrier. According to Mathews and Walker [1909] however MnCl, has no 
effect on the velocity of oxidation of cystein. 

Mn“, Dilute KMnO, added to cystein solution suffers decolorisation. On 
making alkaline with ammonia a rose-red colour appears, fading on standing 
and reappearing on aeration. The depth of colour is increased with increasing 
concentration of the two substances; excess of cystein producing more rapid 
fading on stoppage of the air current, coupled with a less readily regenerated 
colour on aeration, while increasing the amount of KMnQ, gives a more stable 
coloration. Here again the behaviour is parallel with that of iron salts. 

The rose colour may be due to sexvalent manganese, corresponding with 
MnO,, and the colourless reduced solution may contain quadrivalent Mn— 
(corresponding with MnO,) formed from the MnO, by the reducing action of 
excess of cystein. 

The first stage would be the production of MnO, 

sENAO, +08. S—OH(NH,), 000K = 2MnO, + 3(S—CH,—CH(NH,)—COOH), 
+2KOH +2H,0. 
In alkaline solution however Mnv¥! is stable, 
Cg. Mn,0, —~ 2Mn0,+0 
alkali 


3K,Mn™!0, +2H,O — ™ 2KMnv0, + MnO, +4KOH 
in alkaline solution in acid solution 


hence we may have Mn¥! in the alkaline aerated cystein mixture. In the 

reduced solution MnO, is almost certainly present. The action of KMn0Q, is 

being further investigated. 

- Cu” and Cu’. The existence of a copper derivative of cystein has been 

known for some time [Suter, 1895]. The addition of excess of a cupric salt 

to cystein in neutral solution produces a blue-black precipitate of a Cu”-_ 
‘cystein compound. The latter is almost completely insoluble in neutral or: 
slightly acid reaction, and may advantageously be employed for the separa- 

tion and estimation of cystein [Embden, 1901]. When the solution is somewhat 

acid it is better to employ moist freshly precipitated Cu(OH), in place of the 

copper salt, following the procedure used by Hopkins [1921] in the separation 

‘of glutathione. The blue-black precipitate dissolves on addition of ammonia 

forming an intense dark brown (sepia) solution. It is reprecipitated on re- 

ducing the py to about 7 and dissolves to an almost colourless solution in 

excess of acid. 


744 L. J. HARRIS 


- Interesting results have been obtained by the use of excess of cystein, 
The addition of cystein to the alkaline (brown) solution of the cystein-copper 
derivative causes rapid decolorisation. It will be shown that this is due to 
the reduction of the copper to the cuprous condition, and that the decolorised 
solution contains the cuprous-cystein derivative. If the decolorised solution 
be now shaken the dark brown colour reappears: the Cu’ has been re-oxidised 
by the air in the alkaline solution to Cu”, On standing, the solution fades 
once more; and these changes can be repeated so long as sufficient unoxidised 
cystein remains. By the prolonged action of oxygen or air on the brown 
solution the cystein is completely oxidised to cystine and a cuprammonium 
colour persists. a : 

Copper is singular among the metals so far examined in that both its 
eystein salts (cuprous and cupric) are insoluble in neutral solution. As men- 
tioned above an alkaline solution of the cuprous derivative is formed when 
the ammoniacal cupric—i.e. the brown—solution reduces spontaneously in 
the presence of excess of cystein. On neutralisation of the decolorised solution 
by the addition of dilute HCl the cuprous cystein compound is precipitated 
as a highly characteristic voluminous white precipitate. It may be obtained 
in a variety of ways. Add to a few cc. of a 5 % cystein solution one drop 
of CuSO,aq. A blue-black precipitate of the cupric derivative first appears 
but is rapidly transformed into the white cuprous derivative. The Cu” has 
become reduced to Cu’ by excess of cystein. The white precipitate is readily 
soluble in excess of acid or ammonia to a colourless solution. The alkaline 
solution readily darkens in presence of air owing to oxidation of Cu’. The ~ 
blackened alkaline solution will of course again yield the cupric cystein de- 
rivative on neutralisation, while the air-free clear alkaline solution will give 
the white precipitate. 

The cuprous derivative has also been obtained directly by the use of 
cuprous iodide. To a cystein solution a little well washed freshly precipitated 
cuprous iodide was added and the solution made alkaline with ammonia. An 
almost colourless solution resulted. This was neutralised and yielded the 
characteristic white precipitate. The strongly alkaline solution readily darkens 
in air—owing to oxidation of the cuprous ion—especially when the amount 
of copper is large. The colour can be seen extending downwards from the ex- 
posed surface of a large tube. 

Co. The addition of a drop of cobalt acetate to cystein solution followed 
by ammonia resulted in the production of a dark yellow colour. No appre- 
ciable intensification resulted on prolonged passage of air through the mixture. 
The intensity of colour was increased by addition of further Co solution. 

Ni. Cystein solution containing a little NiSO, assumed an orange colour 
on addition of ammonia, The colour deepened to a reddish brown (ma- 
hogany) appearance when some more cystein and nickel sulphate were added. 
The colour was retained so long as cystein remained in the solution, 

Cr. On adding cystein solution to a weak, acidified solution of K,Cr.0, 


A SERIES OF METALLO-CYSTEIN DERIVATIVES 745 


an intense chrome yellow colour appeared. On making alkaline with ammonia 
the usual greenish chromate colour was observed; it was not formed so readily 
however on the direct addition of ammoniacal K,Cr,0, to cystein. Potassium 
dichromate is peculiar in giving the coloration in acid solutions. Potassium 
chromate, which also contains the hexavalent chromium atom, behaves in 
the same way. 

Sni’, Sn. The formation of a tin derivative of cystein may be demon- 
strated in the following manner. Have two test tubes each containing 3 ce. 
of water, and to one add also about 30 mg. of cystein. Add a little SnCl, 
to both tubes. Now make both tubes alkaline by the addition of ammonia 
and then acid by means of sulphuric acid. The presence of a tin cystein de- 
rivative is shown by its ready solubility in acid and more especially in am- 
moniacal solution. The tube containing the cystein will be seen to become 
perfectly clear and colourless in presence of ammonia or larger excess of acid; 
while a bulky precipitate remains in the blank tube under the same circum- 
stances. 

It might be thought that the presence of SnCl, would favour oxidation of 
the cystein. There seems no doubt however that the production of the clear 
solution is due to unoxidised cystein, since its presence is shown by the nitro- 
prusside reaction. Indeed the presence of SnCl, seems to inhibit the oxidation. 
In one experiment a merest trace of cystein was added to a bottle containing 
a large excess of SnCl, and water, and the mixture was made alkaline with 
ammonia. It still showed a nitroprusside reaction after the lapse of five days. 

Tin has a particularly strong affinity for sulphur and no doubt it protects 
that atom in the cystein molecule from oxidation. 

With SnCl, the results observed were in every respect identical with those 
described above for SnCl,. 

As”, As’. Arsenic appears to be too electro-negative to afford any ready 
evidence of functioning as the basic part of a cystein derivative. As,S, and 
As,S3 were examined but without result. 

Pb. The addition of Pb(NO,), to cystein solutions failed to yield a colora- 
tion. 

Hg". If insufficiency of “HgSO, reagent” is added to cystein the white 
precipitate first formed disappears. The same result can be obtained by 
addition of excess of cystein to the mercury precipitate. The phenomenon 
appears to be due to the reduction of the mercury from the bivalent to the 
univalent state, Hg’, which is inefficient as a precipitant. Analogous results 
may be obtained with other mercuric derivatives. If cystein be added in 
excess to a suspension (1) of mercuric iodide, or (2) of the double ammonium 
compound which is precipitated when ammonia is added to acidified mercuric 
iodide, the precipitate will be seen to dissolve in both cases. 

When sufficient mercuric salt is added, however, precipitation occurs. 
Mercuric nitrate, chloride and acetate all yield insoluble compounds with 
cystein. These substances have no precipitant action on cystine. 


Bioch. xv1 5 : 49 


746 L. J. HARRIS 


Bi. Add cystein to an ammoniacal suspension of bismuth hydroxide 
(obtained from bismuth nitrate or subnitrate). The precipitate readily dis- 
solves with the production of a yellow solution, indicative of a Bi cystein 
compound. 

The action of BiO, and Bi,O; remains to be investigated. 

It will be seen that the observations recorded above open up considerable 
ground for further inquiry. A number of metals have still to be investigated. 
In view of the fact that considerable supplies of cystein will be required for 
continuation of the research and in view of the calls of other research work it 
has been thought advisable to give an account of the observations so far made. 

I should like to express my sincere gratitude to Professor Hopkins for the 
help and encouragement he has given me and the interest he has shown in 
this work. 

SumMMARY. 


(1) The so-called “ammonia” test for cystein is only effective in the 
presence of traces of a metallic compound. 

(2) The following metals—viz. Fe”, Mn"’, Mnvi@), Cu, Hg"—can form 
part of a system which is reversible so long as unoxidised ecystein is present; 
the controlling factors being the relative concentrations of cystein, metal salt 
and atmospheric oxygen. In alkaline solution and in presence of oxygen, the 
reduced metal is equivalent to an oxygen acceptor, and in absence of oxygen 
the oxidised metal is equivalent to a donator of oxygen, the cystein forming 
the oxidisable body. 

(3) The fading of alkaline solutions containing a metal plus cystein and 
the recoloration on aeration are explained on these grounds. 

(4) Coloured metallic derivatives of cystein have been formed from Fe’, 
Mn", Mnv), Cu’, Co, Ni, Cr, Bi. Cu’ gives a characteristic white derivative 
insoluble in neutral solutions. Sn and Hg derivatives are also discussed. 

(5) Cystein exhibits greater readiness to form metallic derivatives than i is 
the case with cystine. 

REFERENCES. 


Andreasch (1884). Maly’s Jahresber. 76. 

Arnold (1911). Zeitsch. physiol. Chem. 70, 317. 
Embden (1901). Zeitsch. physiol. Chem. 32, 94. 
Hopkins (1921). Biochem. J. 15, 290. 

Mathews and Walker (1909). J. Biol. Chem. 6, 306. 
Suter (1895). Zeitsch. physiol. Chem. 20, 575. 
Thunberg (1913), Skan. Arch. Physiol. 30, 293. 


LXXVIII. THE INFLUENCE OF FAT AND 
CARBOHYDRATE ON THE NITROGEN 
DISTRIBUTION IN THE URINE. 


By EDWARD PROVAN CATHCART. 
Institute of Physiology, University of Glasgow. 
(Received September 22nd, 1922.) 


IN previous papers it has been shown that the complete withdrawal of carbo- 
hydrate from the diet produces well marked changes in the metabolism. 
Landergren [1903] in his original paper showed clearly that (1) on an ex- 
clusively carbohydrate diet the output of total nitrogen steadily fell and 
(2) when the diet was changed to one composed exclusively of fat the output 
of total nitrogen rose. This finding I confirmed and extended [1909]. The 
series of experiments detailed in the present paper were intended to extend 
further our knowledge of the subject. They were all carried out previous to 
July 1914 on Mr R. Lang and it was intended to complete the series with the 
effects of such limited diets on the capacity of the individual to do muscular 
work, but unfortunately Mr Lang on his demobilisation was no longer available 
and so far no other subject, either willing or suitable to undergo the necessary 
privation, has been found. 


Meruops. 


Previous to the ingestion of the various experimental diets the subject on 
each occasion attempted for two or three days to consume roughly the same 
diet so that the base line would be approximately the same. Unfortunately 
owing to circumstances over which the subject had no control this was not 
always possible. The life led by the subject was uniform throughout. 

The oil used was the finest olive oil procurable. It was always emulsified 
before taking by shaking, after the addition of about 1 g. of potassium car- 
bonate dissolved in water. In this form the oil was fairly readily consumed 
and during the three days of the duration of the experiment it gave rise to 
no untoward digestive disturbance or diarrhoea. As a matter of fact in one 
only of the three day experiments did a single movement of the bowels take 
place. Mr Lang found it impossible to carry on for more than three days on 
account of the nausea induced by the mere sight and smell of the emulsified 
oil. The addition of even 150g. of glucose did not materially lessen this 
disgust although the subject always declared that a feeling of “tiredness” 
which was prominently associated with the ingestion of pure oil was distinctly 
reduced when sugar was added even in comparatively small amounts. The 


49—2 


748 E. P. CATHCART 


sugar used was chemically pure anhydrous dextrose. The calorie values of 
the various diets were kept approximately constant; they varied between 
3102 C. with pure oil and 2978 C. with oil plus 150 g. sugar. The analytical 
methods employed were: total nitrogen, Kjeldahl; ammonia, Folin; urea, Folin 
and urease; uric acid, Hopkins-Folin; creatinine and creatine, Folin. 


RESULTS. 


Output of Nitrogen. 


Ing. ~ In percent. of T.N. 
Day of Am- Total Uric Unde- ; Am- Total Uric Unde- - 
Diet experiment Total Urea monia creatinine acid termined Urea monia creatinine acid termined 

323g. oliveoil 1 10-00 7:31 0:32 0:50 0-094 1-76 73:1 33 50 0-94 17-6 
2 14:24 11-08 0:66 0-62 0-031 1-84 778 46 44 0-22 12:9 

3 [10-75] 7-75 1:07 0-63 0-033 1-27 721 99 58 031 11:8 

323g. oliveoil 1 10:95 7-88 0-29 0-49 0-123 2-16 720 27 45 112 19-7 
2 14:35 11:05 0:56 0-52 0-059 2-17 770 39 36 O41 15-1 

3 14:18 10:44 1:13 0-59 0-029 1-99 73:6 8-0 42 020 14-0 

310 g. olive oil 1 9-41 6-94 0-21 0-58 0-136 1-53 73:7° 2:2 62 1:44 16-2 
30 g. dextrose 2 11-72 748 044 057 0-087 3:15 63:8 _ 3:7 49 0:74 26-7 
3 10-23 6-60 0-58 0-60 0-066 2-38 64:5 5:7 59 0:64 23-2 

297 g. olive oil 1 9-78 7:36 0:26 0-52 0-142 1-50 -15:3 = 2-6 53. 145 153 
60 g. dextrose 2 9-31 7-07 O31 0:49 O-119 1-32 75:9 3:4 53 128 14-2 
3 8-60 586 0-38. 0-50 0-108 1-76 68:1 4-4 58 1:25 20-4 

279 g. olive oil 1 7-97 5:38 0-26 0-48 0-133 1-73 67-5 3-2 569 167 21-7 
100 g. dextrose 2 7°75 5:24 0-25 0-54 0118 1-60 67:6 3:3 70 1:52 206 
3 7-12 4:79 0-18 0-54 0-145 1:47 67:3. 2-5 76 2:03 20:6 

257 g. olive oil ] 10-18 766 0-37 0-56 0-121 1:47 75:2 36 55 118 144 
150 g. dextrose 2 9-95 728 0-24 0-56 0-152 1-68 73-2 24 56 152 168 
3 7:37 505 0-16 0-56 0-143 1:46 68-5 2:2 76 194 19-8 


Total nitrogen. 


The table clearly shows that with increasing amounts of carbohydrate there 
is a general tendency for the output of total nitrogen to decrease. On the 
pure oil, although in each instance there is a rise in the output of total nitrogen 
on the second day of the experiment, there is not the further rise noted on 
the third day of the experiment which both Landergren and I found in 
previous experiments. This may be due to the fact that in this series, in 
contradistinction to the earlier ones, olive oil emulsified by alkali was used in 
place of butter fat or cream. There is no doubt that the olive oil and alkali 
is much more readily tolerated by the intestine than either cream or butter. 
Instead of experiencing the acute diarrhoea which usually results from the 
ingestion of these products when taken alone, as above noted Mr Lang was 
constipated throughout the experiments. 

Zeller [1914] has also carried out a series of observations, both on dog and 
man, on the effect of varying the fat and carbohydrate values of practically 
protein-free diets. It is to be regretted that his experimental periods on the 
fat-rich diets were so short, two days with the.75 % fat diet and only a single 
day with 100 °% fat. Experiments of such very short duration do not permit 


FAT AND CARBOHYDRATE METABOLISM 749 


of the metabolism adjusting itself to the new conditions. The effect of the 
fat-rich diet had however the usual result, viz. a marked rise in the output 
of total nitrogen both in dog and man. 


Urea. 


The urea output resembles very closely that of total nitrogen. It is inter- 
esting to note that when the percentage outputs of urea and ammonia nitrogen 
are compared there is a very marked fall in the output of urea after the sugar, 
even with the smallest dose, although at the same time there is also a fall in 
the output of ammonia, whereas it might have been expected that with the 
decrease in the ammonia output there would have been a rise in the urea 
output. In the case of the experiments with oil alone there does seem to be 
a balance between the two outputs, a rise in the ammonia output being 
associated with a fall in the urea. 

When the outputs of both ammonia and urea are taken together it may 
be definitely stated that when oil is given alone there is a definite and steady 
rise in the excretion of these nitrogenous materials, whereas in the experiments 
in which sugar is added there seems to be just as general a tendency for the 
united output to fall during the three days of the experiment. 

Zeller in his series also found that the urea output followed very closely 
the curve for the output of total nitrogen. His percentage output of urea 
reached however a far lower level than any that I obtained. Thus when a 
pure carbohydrate diet was given immediately after a pure fat diet he found 
in one instance that urea only formed on the average 49-7 °/ of the total 
nitrogen and in the other it reached the very low level of 39-7 %. 


Ammonia. 


In each case with the pure oil diet a very definite steady rise both of the 
absolute and the percentage output of ammonia nitrogen occurs on the three 
days of the experiment. A slight rise is also noted on the third day when 
30 and 60g. of sugar are given although, even with these small amounts of 
sugar, there is a definite reduction in the total ammonia output. When the 
larger amounts of sugar are given there is a steady fall in the output of 
ammonia during the three days of the test. 

Zeller too in his experiments found, like other observers, that the giving 
of fat-rich diets led to a marked rise in the output of ammonia. In all of his 
experiments, just as he found a much lower percentage output in the case of 
urea, he obtained much higher outputs of ammonia, both with the carbo- 
hydrate-rich and carbohydrate-poor diets, than those given in the present 
paper. His maximum percentage output of ammonia was 15-4 % with the 
100 % fat diet. 

Urie acid. 

The variation in the excretion of uric acid as the result of the alterations 

in the composition of the diets is very striking and most interesting. As the 


750 piety EK, P. CATHCART 


table shows it may generally be stated that when carbohydrate is completely 
removed from the diet there is a well marked reduction in the output and that 
this output gradually and steadily rises as the amount of glucose added is 
increased. It was shown in a previous paper [1909] that the output of uric 
acid varied with the carbohydrate intake; that the output was high when the 
subject was on a carbohydrate-rich diet and low when on a fat-rich, carbo- 
hydrate-poor diet. Graham and Poulton [1913] who studied the influence of 
carbohydrate and fat on the output of endogenous purine also found that a 
diet in which fat predominated was associated with a low output and a diet 
in which carbohydrate predominated with a high output of uric acid. Later 
Umeda [1915] working in my laboratory confirmed, both in the case of man 
and the dog, the variation in uric acid output as the result of alteration of 
the diet. He found in man that, although there was a definite fall in the 
output of total purines, the output of purine bases on a fat diet was higher 
than on a carbohydrate one. Graham and Poulton had also noted that when 
the carbohydrate content of the diet was reduced there was a tendency for 
the purine base output to rise. In the case of the dog, Umeda found that the 
allantoin output behaved like uric acid in its relation to the nature of the 
diet. Zeller also investigated the influence of carbohydrate and fat diets both 
on the output of uric acid and the purine bases. He too found that there was 
a reduction in the output of uric acid on a 100 % fat diet but there was little 
or no influence on the output of the purine bases. As already noted Zeller’s 
experiments were of exceptionally short duration. 

In an interesting and suggestive paper Ackroyd and Hopkins [1916] 
showed very definitely that (1) when both arginine and histidine were re- 
moved from the diet the amount of allantoin excreted in the urine of the rat 
was much decreased; (2) it was somewhat diminished when only one of these 
amino acids was present and (3) when both were restored to the diet the 
allantoin excretion returned to normal. Their conclusion that arginine and 
histidine probably constitute the most readily available raw material for the 
synthesis of the purine ring in the animal body would seem to be justified. 
Still, although these two amino acids may be considered to supply the raw 
material for the synthesis, this does not dismiss the probability that the 
presence of carbohydrate is necessary as in the experiments of Ackroyd and 
Hopkins the carbohydrate supply was abundant. Further, the well known 
experiment of Knoop and Windaus [1905] in which they demonstrated the 
formation of the iminazole ring when a solution of dextrose was acted upon 
by ammonia in the presence of zinc and sunlight, shows, at least, that a syn- 
thesis from very simple compounds is possible. If the rest of the diet played 
but a minor part in the synthesis of the purine it would have been expected 
that the output of purines would have risen when a general increase in the 
breakdown of the protein molecule took place, as is the case when oil alone is 
the food material supplied, i.e. when there would be in all probability an 
increased amount of amino acids including arginine and histidine free in the 


FAT AND CARBOHYDRATE METABOLISM 751 


tissues. Further the diminution of the output of uric acid on the fat diet is 
not due to delayed oxidation of other purine bodies because reference to the 
output of undetermined nitrogen, which would include purine bases, shows 
that the excretion on the fat diet is actually lower than when sugar is given. 
It is probable then that, although arginine and histidine may be regarded as 
the actual nitrogenous source of the purine, before the synthesis can take place 
carbohydrate in some form or other must be present. 


Creatinine and Creatine. 


Inasmuch as many of the estimations of these substances were made 
without the precautions suggested by Graham and Poulton the total creatinine 
output only is given. It may be stated that in two other experiments when 
every precaution was taken the presence of creatine in the urine was definitely 
shown. The work of Underhill and Baumann [1916] showed very conclusively 
that acidosis alone cannot account for the presence of creatine in the urine. 


Undetermined nitrogen. 


Although the absolute amount of undetermined nitrogen daily excreted 
does not vary very greatly in any of the experiments, yet, when calculated 
on a percentage basis, it is found that there is a very definite rise when carbo- 
hydrate is added to the diet. Zeller in one of his pure fat experiments found 
the highest percentage output whereas in the other the output was low, if 
not the lowest of the series. Zeller in addition to the substances referred to 
in this paper also determined amino and peptide nitrogen. The output of 
these substances rose when the pure fat diet was given. 


DISCUSSION. 


The consideration of the question of isodynamic replacement and the 
inferences to be drawn from the consideration of experiments on complete 
replacement such as those described, irrespective of the question as to whether 
such experiments fall within the normal capacities of the tissues, open up very 
wide issues. The mere statement of the isodynamic law virtually upholds 
the thesis that the basis of nutrition is the exchange of energy and not the 
exchange of material or, at the very least, that Kraftwechsel predominates 
over Stoffwechsel. Rubner [1883] who enunciated the hypothesis that fat and 
carbohydrate are mutually replaceable in a diet in isodynamic amounts, 
although in a much more recent paper he definitely stated that it was im- 
possible to replace completely any of the proximate principles, definitely: 
selected the term isodynamic after the consideration of such less specific terms 
as gleichwertig. 

The work of Rubner is generally stated to be substantiated by the work 
of Atwater and Benedict [1903]. The experiments of these workers do un- 
doubtedly support it but the type of experiment they adopted could not be 


752 EK. P. CATHCART 


expected to determine the ultimate degree of replacement, as only variations 
in a mixed diet on the human subject under approximately normal conditions 
were employed in contradistinction to Rubner’s method of a rigorous adminis- 
tration of a single food stuff. It is undoubtedly true that within limits fat 
and carbohydrate may replace one another in the diet and it is obvious that 
in the average diet such an arrangement is automatically adopted. The state- 
ment of the case in the form of a general isodynamic law is simply untenable. 
At present the evidence is clear and convincing in support of the statement 
that it is impossible to replace carbohydrate completely by fat but so far the 
evidence available in support of the view that fat as such is a necessary con- 
stituent of a diet is scanty and unsatisfactory. 

No one will of course seriously maintain that nutrition can ultimately be 
reduced merely to the satisfying of the energy demands: the calorie factor 
may be regarded as strictly secondary to the supply of material. We do not 
live on calories, yet all our general estimates of food requirements are quite 
properly for the most part made in terms of calories. Calorie value is simply 
a very convenient physical standard for the assessment of diets, but merely 
because such a standard has proved of great utilitarian value there is no real 
justification for placing this standard as the foundation stone of hypotheses 
framed to offer an explanation of cellular activity. Many writers are obsessed 
with the idea of the calorie, forgetting that the organism is certainly not a 
heat engine. It is perfectly true that calories are a measure of heat, but it 
must not be forgotten that we do not consume actual heat units but only 
potential heat-giving substances which can eventually be degraded to the 
form of heat and be measured as such. The thermal aspect of nutrition is 
unduly stressed, for, while heat may be a necessary product of tissue activity, 
it is, after all, a by-product. 

The use of the term isodynamic in connection with problems of nutrition 
should be strictly limited. One can undoubtedly speak of isodynamic quan- 
tities of various substances but it does not follow that they are of equal, or 
indeed of any, value to the organism. When dealing with foodstuffs we ought 
to keep constantly in view that the material side is fully as important as the 
energy side. Therefore one ought not to stress so much the equality in energy 
as the equality in sparing or preventing tissue breakdown, the isoeconomic or, 
as I prefer to call it, the isotamieutic (Gk. tamieuo = to husband or to spare) 
value. Such a value is more physiological than isodynamic as it covers all 
phases of cellular activity. At present the data available do not suffice to 
permit of any adequate explanation of metabolic phenomena. Considerations 
such as those on which Carl Voit based his theory of metabolism, so actively 
rebutted by Pfliiger, are not dismissed by the more modern hypothesis of 
Folin. Folin simply dealt with entirely superficial results and within these 
limitations the hypothesis is admirable, He did not attempt to elucidate the 
causal factors which lie beneath the phenomena which he correlated in his 
papers. The old question discussed so energetically by Voit and Pfliiger as 


FAT AND CARBOHYDRATE METABOLISM 753 


to whether the newly ingested material becomes an integral part of the living 
molecule before utilisation is still unanswered. 


CONCLUSIONS. 


1. The output of total nitrogen, urea, and ammonia rises on a fat diet 
and falls on the addition of carbohydrate. 


2. The output of uric acid is low on the fat diet and increases on the 
addition of carbohydrate. 

3. The output of total creatinine is but little affected by the change of 
diet. Small amounts of creatine are excreted on a carbohydrate-free diet. 

4. The output of undetermined nitrogen is greater on diets containing 
carbohydrate than on those from which carbohydrate is absent. 


REFERENCES. 


Ackroyd and Hopkins (1916). Biochem. J. 10, 551. 

Atwater and Benedict (1903). U.S. Dept. Agr. Bull. 136. 
Cathcart (1909). J. Physiol. 39, 311. 

Graham and Poulton (1913). Quart. J. Med. 7, 13. 

Knoop and Windaus (1905). Beitrdge, 6, 392. 

Landergren (1903). Skan. Arch. Physiol, 14, 112. 

Rubner (1883). Zettsch. Biol. 19, 312. 

Umeda (1915). Biochem. J. 9, 421. 

Underhill and Baumann (1916). J. Biol. Chem. 27, 127 et seq. 
Zeller (1914). Arch. Physiol. 213. 


LXXIX. ON THE INFLUENCE OF THE SPLEEN 
UPON RED BLOOD-CORPUSCLES. I. 


By NICOLAAS ALBERT BOLT anp PIETER ANTON HEERES. 
From the Physiological Laboratory, University of Groningen, Holland. 


(Received October 3rd, 1922.) 


Srnce the researches of Hunter [1892] on the destruction of red blood-cor- 
puscles, an active réle in this action has been ascribed to the spleen. The 
appearance of phagocytosis and haemolysis in the spleen had already been 
discovered by Koelliker [1847]. From this discovery resulted the researches 
of Virchow, Quincke and others on the transformation of blood pigments, in 
which for the rest the spleen was only considered as an organ of accumulation. 

But it was Hunter who pointed out that the cause of the blood destruction 
partly lies in an action of the spleen itself. This view was based specially 
upon this author’s researches on toluylene-diamine poisoning. As is known 
the action of this poison is much stronger in normal than in splenectomised 
animals, so that the opinion is justified that the spleen plays an interfering 
part in toluylene-diamine poisoning. 

To determine the haemolytic power of the spleen the following four 
methods have been used: 

1. Microscopical examination of spleen pulp in normal, experimentally 
modified and pathological conditions. This morphological method is not well 
adapted for learning the mechanism of the process observed. 

2. Study of the consequences of splenectomy. We do not propose to 
discuss all the experiments made in this direction. The most generally accepted 
result of this operation is the increased resistance of the red blood-corpuscles 
of the animal against hypotonic salt solutions, a fact which has mainly been 
settled by the Dutch investigator Pel Jr. [1911]. 

3. Study of the action of splenic extracts. The results of this method, 
which appears to us as a very unreliable and inaccurate one, are moreover 
very contradictory. 

4, Comparative examination of blood from the splenic vein and splenic 
artery. The latter may of course be replaced by arterial blood in general. 

Destructive action of the spleen being admitted, it was obvious to inves- 
tigate traces of this action in the blood of the vena lienalis. There are three 
principal qualities of this blood, which may be examined: 

(a) the number of erythrocytes in a cubic millimetre; 

(b) the amount of decomposition products of red cells, especially haemo- 

globin and bilirubin; 

(c) the resistance of red blood-cells to haemolytic agents. 


SPLEEN AND RED BLOOD-CORPUSCLES 755 


The number of erythrocytes in a cubic millimetre of the blood of the splenic 
vein, as compared with blood from the splenic artery, is not a very exact 
measure of the function of the spleen. In the first place the difference which 
might be expected would be so small as to be within the limits of technical 
error. If the lifetime of a red cell is estimated at 20 days (Quincke, Bénard; 
Rubner consider it to be at least 70-90 days, Dekhuysen still longer) and if 
one supposes that the total quantity of blood of the body passes through the 
spleen about thirty times a day [Burton-Opitz, 1912], it is clear that of each 
volume of blood which passes through the organ only 1/20 x 1/30 = 1/600 is 
destroyed. So one might expect to find numbers like 6,000,000 and 5,990,000 
for splenic artery and vein. This difference is too small to be determined with 
sufficient accuracy. Larger differences might be found, if this action of the 
spleen were intermittent, but there is no evidence that this is so. 

Moreover the operation which is necessary for the puncture of the splenic 
vessels will cause a stagnation of the blood which has a great influence upon 
the number of erythrocytes. 

Of late this method has been used again by Frey [1920]; his values for the 
difference between the blood from an artery of the abdominal wall and the 
splenic vein are so large (on an average 800,000 in a cubic millimeter) that in 
our opinion they must be due to some error in technique. 

An increased quantity of bilirubin in the splenic vein, as compared with 
that in the blood from other vessels, might possibly be due to the formation 
of bilirubin in the spleen (Hymans van den Bergh and Snapper, Ernst and 
Szappanyos), it would not be conclusive proof of a haemolytic function of 
that organ. 

The best method in our opinion is the investigation of the resistance of 
the red cells against hypotonic salt solutions, the so-called osmotic resistance. 
Itis Eppinger[ 1920], who also stated that the osmotic resistance of the erythro- 
cytes was decreased in the spleen, asserting that this fact is “eigentlich die 
einzig fassbare Tatsache die im Sinne einer zerstérenden Tiatigkeit der Milz 
spricht, vorausgesetzt dass die Resistenzherabsetzung als ein einleitender 
Vorgang der Erythrozytenzerstérung aufgefasst werden kann.” 


French authors were the first to use the resistance determination in hypotonic salt solutions, 
originating from Hamburger [1904], to estimate the condition of the red blood-cells in several 
diseases. The value of this method however has been greatly increased by the modifications and 
extensions given by Brinkman [1922]. This author showed in the first place that the haemolysis 
in hypotonic NaCl solution is not only due to the difference of osmotic pressure, but also to the 
lyotropic effect of pure NaCl, the liquefaction of the cellular membrane-colloids by the Na-ions, 
Especially when the erythrocytes are washed several times with an isotonic NaCl solution, as is 
done by many authors, the cells are injured, as is shown by the diminished osmotic resistance 
(Snapper). This lyotropic effect may be abolished by the addition of Ca-ions to the solution in 
a very definite concentration. It would lead us too far to discuss the modified Ringer solution, 
which is.used by Brinkman; we give here only the composition: 


NaCl ae sec, HORE: Gage 
NaHCO, : 4.5). ha 0-8 2%: [H"]=0-45 x 10-7, 
KCl Frnt bac’ OE 
CaCl,6aq. ... segn O'OR Yat 


Hypotonic solutions are obtained by diminishing only the concentration of NaCl. 


756 N. A. BOLT AND P. A. HEERES 


If the red blood-cells are washed with this isotonic solution, the result is an increase of the 
osmotic resistance. This is due, as has been shown by Brinkman, to the removal of an auto- 
haemolytic substance from the surface of the erythrocytes, a phosphatide, perhaps lecithin’. 
This same substance, if you make an emulsion of it in modified Ringer solution in an appropriate 
concentration, added to red blood-cells, may dissolve them. 

In the surface of the blood-cells the haemolytic action of the phosphatide is balanced by the 
cholesterol, which is kept in solution by the first. The osmotic resistance has been shown to be 
dependent on the proportion cholesterol/phosphatide. 

It must still be mentioned, that the modified Ringer liquid may be substituted by a mixture 
of primary and secondary phosphate, containing 16-30 g. Na,.HPO,. aq. and 2-18 g. KH,PO, per 
litre. The freezing-point of this liquid in which the erythrocytes have the same volume as in their 
plasma, is —0-46°. Without any theoretical consideration, the equivalence of this solution with 
the modified Ringer solution has been proved empirically. Its advantages are the easiness of 
its preparation and its stability, combined with the physiological properties of the modified 
Ringer solution. It was necessary to mention this liquid, because we used it in our resistance- 
determinations. For further particulars reference must be made to the publications of Brinkman. 

After this digression we will discuss briefly the results obtained by others 
in the determination of the resistance of red blood-cells from the splenic vein 
in comparison with that of erythrocytes from arterial blood. _ 

Gabbi [1893] finds like Hammarsten an increased resistance in the splenic 
vein. 

Pugliese and Luzatti [1900] observed however a lowered resistance of the 
red blood-cells leaving the spleen. 

Chalier and Charlet [1913] once more found that both minimum and 
maximum resistance of the blood-cells from the splenic vein (7.e. respectively 
the resistance of the weakest and of the strongest erythrocytes in regard to 
haemolysis by hypotonic salt solution) were higher than those of erythrocytes 
from the splenic artery. From their observations they conclude that the spleen 
only lets the strongest erythrocytes pass and so exerts a haemolytic action 
upon the blood. This conclusion would be more justified, if only the minimum 
resistance had increased, the weakest erythrocytes entering the spleen having 
been destroyed. How is the increase of the maximum resistance, however, 
from this point of view to be explained, 7.e. the fact that the most resistant 
blood-corpuscles which enter the spleen, leave the organ still more resistant? 

The remaining investigators, Strisower and Goldschmidt [1914], Widal 
and Abrami [cited by Eppinger, 1920], Banti and Furno [cited by Eppinger, 
1920], Polak Daniéls and Hannema [1916], Eppinger [1920], all found a 
diminished resistance of the erythrocytes leaving the spleen. We have already 
quoted Eppinger’s opinion on the importance of this fact. 

We also have tried to study the haemolytic function of the spleen by 
investigating its influence upon the resistance of the red blood-cells passing 
through the organ. For practical reasons we took our materials from sheep 
at the abattoir. Immediately after the killing of the animal, some of the 
blood spurting out of the carotid artery was collected in a glass tube and 
defibrinated by shaking it with glass beads. Then as soon as possible the 


' It may be obtained by evaporating the washing fluid and extracting the phosphatide from 
the dry residue, or, by direct extraction of the erythrocytes with chloroform, after they have 
been extracted with light petroleum, in which cholesterol and neutral fats dissolve. 

* We were able to do so by the kindness of the authorities of the abattoir in Groningen. 


SPLEEN AND RED BLOOD-CORPUSCLES 757 


spleen was removed, care being taken not to injure the capsule and the 
splenic vessels. The organ was then rapidly conveyed to the physiological 
laboratory. By introducing into the splenic artery a glass canula, which was 
connected with a reservoir, containing the modified Ringer solution, placed 
about 175 cm. above the working table, it was possible to obtain from the 
splenic vein the blood which had remained in the organ, diluted with the salt ° 
solution. The suspension of erythrocytes obtained in this manner was centri- 
fuged immediately. These red corpuscles, which had been for some time in 
the spleen, longer than they are normally, but too short a time for autolysis 
to take place, seemed to us very well adapted for the investigation of an 
eventual resistance-decreasing power of the spleen. By the dilution of the 
blood from the spleen with modified Ringer solution, the resistance of the 
erythrocytes is practically not changed. The effect of washing red corpuscles 
with this solution, which we have already described, fails practically to be 
evident, if the total blood is “washed” with the liquid. To produce the same 
conditions, however, we have also diluted the arterial blood from the carotis 
and centrifuged it to get the corpuscles. 

Both kinds of the corpuscles were then put in a series of tubes containing 
the phosphate mixture in geometrically decreasing concentrations, viz.: 

9/10 isotonic (obtained by adding one vol. aq. dest. to 9 vols. of an iso- 
tonic solution). 

(9/10)? isotonic (obtained by adding again one vol. of aq. dest. to 9 vols. 
of the 9/10 isotonic solution). 

Ete. 

Ten minutes after the addition of the erythrocytes, the latter were centri- 
fuged and the degree of haemolysis read by Arrhenius’ method. In 20 deter- 
minations we always reached the same result, that is to say, the blood from 
the spleen was less resistant than that of the carotis. Both the beginning and 
the complete haemolysis of the erythrocytes of the spleen occurred in a salt 
solution of higher concentration than those of the blood-corpuscles of the 
carotis. In a hypotonic salt solution of a definite concentration the degree of 
haemolysis of the blood from the spleen was always larger than that of 
erythrocytes from the carotis. 

We do not think that in vivo the difference of the resistance of red blood- . 
corpuscles from art. and vena lienalis will be as large as in our determinations, 
because in the latter the erythrocytes had remained for some time in the spleen. 
Eppinger observed e.g. that if the blood stagnated for 10 minutes in the 
. Spleen the resistance of erythrocytes from the vena lienalis was distinctly 
lower than before the stagnation. There is, however, evidence that in vivo 
also stagnation in the organ is possible. 


758 -_-N. A. BOLT AND P. A. HEERES 


Table I contains the determinations: 


Table I. 


% haemolysis 
Salt concentrations 
A 


No. Kind of blood Isotonic 0- 


mG 
9 0-9? 0-98 0-94 0-9° 0-98 

1 Spleen 0 0 10 30 70 100 — 
Carotis 0 0 0 0 0 0 _ 

2 Spleen 0 0 0 20 50 90 — 
Carotis 0 0 0 0 - 0 0 — 

3 Spleen 0 20 40 90 100 100 _— 
Carotis 0 o- 0 0 30 75 _- 

4 Spleen 0 5 10 50 70 100 — 
Carotis 0 0 0 0 5 10 —_ 

5 Spleen 0 0 5 20 50 90 _ 
Carotis 0 0 0 0 5 — — 

6 Spleen 0 0 0 50 90 95 100 
Carotis = — — 0 5 60 90 

7 Spleen 0 0 5 50 100 100 100 
Carotis — — 0 DY 30 50 70 

8 Spleen — — 10 30 50 100 100 
Carotis a = 0 10 50 80 100 

9 Spleen — 5 5 20 50 80 100 
Carotis — 0 0 5 10 50 70 

10 Spleen — 0 5 10 40 60 90 
Carotis — 0 0 0 5 30 90 

ll Spleen — 5 20 40 60 100 100 
Carotis —- 0 5 10 30 50 70 

12 Spleen _ 0 0 10 50 70 100 
Carotis _— 0 0 0 10 - 50 70 

13 Spleen —_ 0 0 10 30 60 100 
Carotis — 0 0 0 5 20 50 

14 Spleen — 0 5 10 40 80 100 
Carotis a 0 0 5 30 50 70 

15 Spleen —- 0 5 20 50 100 100 
Carotis — 0 0 0 30 70 90 

16 Spleen -- 0 5 10 30 70 100 
Carotis — 0 0 0 30 70 90 

17 Spleen — 5 10 30 50 80 100 
- Carotis — | 0 0 10 40 80 
18 Spleen _ 5 15 30 80 100 100 
Carotis - 0 0 10 30 70 100 

19 Spleen = 5 10 40 70 90 100 
arotis 0 0 10 40 100 100 

20 Spleen a 5 20 50 100 100 100 
Carotis —- 0 0 10 60 100 100 


If one calculates for every concentration the average degree of haemolysis 
one obtains the following table: 


Table II. 


% haemolysis 


~~ amie Ta So 
Salt concentration Carotis Spleen 


0-9 x isotonic 0 0 
9 x # 0 5 
0-9 x a 0 30 
0-9" x a 20 60 
0-0 x a 50 85 
0-0 x a 75 100 


00 x: 100 100 


SPLEEN AND RED BLOOD-CORPUSCLES 759 


One may consider this as an accidental case. Fig. 1 shows the “‘resistance- 
curves” of this case (the left one of the corpuscles of the spleen, the other of 
the carotis corpuscles): 


100 


b0r 


Haemolysis in % 


0,9x isot 0.97 09° 9% O9° 09° a9’ 
Salt concentrations 
Fig. 1. 


To deduce from the decreased osmotic resistance of the erythrocytes 
leaving the spleen conclusions as to the haemolytic function of the organ, it 
is necessary to investigate how far the phenomenon is specific. Of course it 
is not necessary that only the erythrocytes from the vena lienalis should have 
a decreased resistance; it might be also the case with blood-corpuscles leaving 
the liver and perhaps other organs in so far as these are concerned with 
blood destruction. We have contented ourselves for the present with showing 
that an organ like the kidney to which no haemolytic function is assigned 
fails to show a resistance-diminishing power. Table III shows our deter- 
minations: 

Table IIT, 


% haemolysis 
Salt concentrations 
A 


No.. Kind of blood 0-9 isotonic 0-9? 0-98 0-94 0-98 0-98 
1 Kidney 0 0 5 50 75 100 
Carotis 0 0 5 60 75 100 

2 Kidney 0 5 10 50 80 100 
Carotis 0 5 10 30 50 70 

3 Kidney 0 0 10 50 80 _ 100 
Carotis 0 0 10 30 70 100 


We have already mentioned the opinion of Chalier and Charlet and of 
Gabbi, that the spleen would exert a destroying action upon the weakest 
blood-corpuscles, so that the result is an increased resistance of the erythro- 
cytes in the splenic vein, as was found by these authors, and we have observed 
that it is difficult to explain the increase of the maximum resistance from this 
point of view. From our researches, however, it would follow that the spleen 
exerts a haemolytic influence upon all blood-corpuscles, passing through the 


760 N. A. BOLT AND P. A. HEERES 


organ, as is proved by the diminished minimum and maximum resistance of 
erythrocytes from the vena lienalis. The parallel course of the curves in Fig. 1 
gives a clear image of this fact. If one makes e.g. the assumption that the 
spleen produces an internal secretion, which causes an alteration of all erythro- 
cytes passing through the organ—and perhaps also has some influence upon 
the bone marrow—the mode of action of the spleen becomes more intelligible. — 
We also point to the fact that after splenectomy a general increase of the re- 
sistance of the blood-corpuscles has been found,7.e. an increase of the minimum 
and maximum resistance whilst the “breadth of resistance” (the difference 
in concentration between the solutions in which beginning and complete 
haemolysis is observed) remains the same. Pel gives e.g. the following table 
as the average of a large number of experiments: 


Table IIT ed 
Normal dog Splencct. dog 
% % 
Average concentration of NaCl, in which beginning 0-42 0:35 
* haemolysis is observed 
Average concentration of NaCl, in which just com- 0-30 0-23 
plete haemolysis is observed 
Breadth of resistance 0-12 0-12 


This phenomenon would also be explained in the most simple way by the 
assumption that after removal of the spleen the production of an internal 
secretion ceases, that has normally a harmful action upon the erythrocytes, 
which is measured by the diminished resistance. 

We have tried to ascertain more particulars about the resistance-dimin- 
ishing power of the spleen. Assuming the action of an internal secretion, it 
was obvious that in the first place the surface layer of the erythrocytes would be 
altered by such an action. From the researches of Brinkman and Miss van Dam 
[1921] we know however the importance of the proportion cholesterol/phos- 
phatide in the surface layer for the osmotic resistance. A decrease of this 
proportion means a diminution in the osmotic resistance. To test the assump- 
tion that only or principally the surface layer of the erythrocytes is altered 
by the action of the spleen it is only necessary to investigate the resistance of 
the blood-corpuscles of the vena lienalis, when this surface layer has been 
removed by washing the erythrocytes with modified Ringer solution. We have 
made these determinations by washing the blood-corpuscles obtained in the 
manner described on p. 757 and also the erythrocytes from the art. carotis 
three times with the modified Ringer solution and then determining the re- 
sistance in the same way as described above. In all our determinations we 
found a considerable difference between the two kinds of corpuscles, before 
washing, which disappeared however almost entirely after this process. 
Table LV contains the values of these determinations: 


SPLEEN AND RED BLOOD-CORPUSCLES 761 


Table IV. 
Kind of blood. % haemolysis. 
Three times Salt concentrations 
washed with r A ~ 

No.! modified Ringer Isotonic 0-9 0-9? 0-98 0-94 0-9% 0-98 
1 Spleen 0 0 0 0 5 20 _— 
Carotis 0 0 0 0 0 5 — 
2 Spleen 0 0 0 0 20 50 — 
Carotis 0 0 0 0 0 0 5 
3 Spleen 0 0 0 0 5 20 a= 
Carotis 0 0 0 0 0 5 — 
4 Spleen = 0 5 10 30 50 100 
Carotis — 0 0 10 30 50 100 
5 Spleen ~- 0 0 0 5 10 40 
arotis — 0 0 0 5 10 40 
6 Spleen — 0 0 10 40 80 100 
Carotis — 0 0 10 40 80 100 
7 Spleen — 0 0 20 50 70 100 
Carotis _ 0 5 20 30 70 100 
8 Spleen 0 0 10 30 60 90 
Carotis — 0 0 20 30 50 90 
9 Spleen - 0 0 0 10 40 100 
Carotis — 0 0 0 10 30 70 
10 Spleen ae 0 0 10 40 80 100 
Carotis ~- 0 0 20 50 80 100 


Table V contains the average values which are represented also in Fig. 2. 


Table V. 


% haemolysis 


Washed spleen Washed carotis 


Salt concentration Corpuscles 
0-9 x isotonic 0 0 
eae 0 0 
0-98 x ‘3 5 6 
0-04. -,, 20 15 
OO se a5 ,: 45 35 
POF x <5, 90 85 
/00 
k 2 
. 
50F 


Haemolysis in % 


— 


= 


Ow ation 
0,9x isot 09° 09% 09% O9% 09° 
Salt concentrations 

Fig. 2. 


1 The resistance of the unwashed corpuscles of these cases may be seen respectively in 
Nos. 1, 2, 3, 4, 11, 13, 14, 15, 16, 17, 18 of Table I. 


Bioch. xv1 50 


762 N. A. BOLT AND P. A. HEERES 


The resistance of the arterial erythrocytes does not increase in the same 
degree as that of the blood-corpuscles of the spleen, so that after the washing 
the difference between the two kinds of blood-corpuscles has been reduced 
considerably, as is shown by Table V and Fig. 2. 

There remains a small difference; this is perhaps partly due to the de- 
fibrinating of the blood from the art. carotis (the resistance is increased hereby 
somewhat); moreover arterial erythrocytes are always somewhat more re- 
sistant than blood-corpuscles from venous blood (Hamburger). The same 
difference exists between the resistance of erythrocytes from the vena renalis 
and art. carotis. 

From our experiments we draw the conclusion that the influence exerted 
by the spleen upon the red blood-cells is confined to the adsorbed surface layer 
of the latter, 7.e. to the layer which may be removed by washing with a 
modified Ringer solution. As we have not found any conception in the 
literature about the mode of alteration of the erythrocytes themselves in the 
spleen, except the following of Jacobsthal [1921], it seemed worth while to 
quote it. “Ich habe mir die Vorstellung gebildet, dass die Abnahme der 
Blutkérperchenresistenz beim haemolytischen Ikterus sich so entwickelt, dass 
die die Milz passierenden Blutscheiben jedesmal nicht haemolysiert, sondern 
an threr Iipoidhiille sozusagen nur angenagt werden.” 

The three following alterations in the surface layer of the red blood- 
corpuscles in the spleen are possible: 

1. Amount of phosphatide the same; amount of cholesterol decreased. 

2. Amount of phosphatide increased; amount of cholesterol the same. 

3. Amount of phosphatide and cholesterol both increased, the former 
however in the highest degree. 

In all these cases the proportion cholesterol/phosphatide would diminish 
and this means a decrease of osmotic resistance. It seems that of these three 
possibilities the third one holds true. It has appeared to us that the erythro- 
cytes leaving the spleen contain more cholesterol than the erythrocytes from 
the art. carotis'. It is therefore necessary that the phosphatide should undergo 
a considerable augmentation. We want to point out here already that this 
increase may be as well of a qualitative as of a quantitative nature. The 
specific influence of the phosphatide and cholesterol of the surface of the 
erythrocyte upon its resistance is undoubtedly due to the physico-chemical 
properties of these substances. If one makes an emulsion of lecithin ex ovo 
in an isotonic modified Ringer solution, this emulsion has a low surface tension 
and a great haemolytic power. If, however, cholesterol is added in increasing 
small amounts, the surface tension and the viscosity of this emulsion rise 
gradually and at the same time the haemolytic power decreases. The haemo- 
lytic power of an emulsion of lecithin is dependent on the quantity of the 
substance per cc, of the solvent, but apparently small qualitative alterations 
of the lecithin also have an enormous influence upon its haemolytic power. 


* Publication of these investigations will follow. 


SPLEEN AND RED BLOOD-CORPUSCLES 763 


This has been shown by Delezenne and Fourneau [1914], who withdrew oleic 
acid from lecithin by means of cobra venom and in this way obtained a 
substance, called by them “désoléolécithine” or “lysocythine,” which had a 
very strong haemolytic power. It is possible that in vivo also analogous 
chemical changes play a part in haemolytic processes. Perhaps it will be 
possible to settle the nature of the actual quantitative or qualitative changes 
in the phosphatides in the erythrocytes in the spleen and elsewhere by further 
researches. 

To judge the value of our opinions, developed after investigations on the 
sheep’s spleen, for normal and pathological physiology (Eppinger’s Hyper- 
splenie) of the human spleen, it will be necessary to repeat the determinations 
of the resistance described by us in man. This is possible where splenectomy 
is performed. Till now only one case has been investigated in this respect. 
For these figures we are indebted to Dr H. H. de Zoo de Jong, curator of the 
clinic of Prof. L. Polak Daniéls, University Hospital, Groningen. Splenec- 
tomy was performed in the case of a haemolytic anaemia. Some blood was 
taken from the vena lienalis and the red blood cells compared with erythro- 
cytes from the peripheral blood in regard to their osmotic resistance before 
and after washing with a modified Ringer solution. Table VI contains the 
determinations : 


Table VI. 
a % haemolysis 
A ‘ 
Before washing ~ After washing — 
. a" — 
Salt concentration — Peripheral Spleen Peripheral Spleen 
3 isotonic 20 30 0 0 
0-9 x = 35 65 40 40 
0-9? x " 70 80 70 70 
0-98 x i 80 85 80 80 
0-94 x oa 95 95 90 95 
0-98 x af 100 100 100 100 


The difference between peripheral blood and blood from the vena lienalis would perhaps have 
been more evident if the resistance had also been determined in solutions stronger than } x iso- 
tonic. Then the beginning of haemolysis would also have been determined. 

For the rest it is clear that in this case the investigations on the sheep’s spleen are wholly 
confirmed. It will, however, be necessary to repeat these determinations in regard to man if 
possible. 


CONCLUSIONS. 


The spleen has the power of diminishing the osmotic resistance of the 
erythrocytes. These are prepared hereby for haemolysis, which partially 
takes place in the organ itself. 

Erythrocytes from the vena lienalis, which have been washed with an 
equilibrated salt solution, do not show this decreased osmotic resistance. This 
fact proves that the point of attack of the haemolytic power of the spleen 
lies in the removable surface layer of the erythrocytes. The alteration in this 
layer must consist in a decrease of the proportion cholesterol/phosphatide. 


50—2 


N. A. BOLT AND P. A. HEERES 


REFERENCES. 


Brinkman (1922). Résistance osmotique et phosphatides du sang, Groningen. 
Brinkman and Frl. E. v. Dam (1921). Biochemische Zeitsch. 108. 
Burton-Opitz (1912). Pfliiger’s Archiv, 146. 

Chalier and Charlet (1913). J. Phys. Path. Gén. 13. 

Delezenne and Fourneau (1914). Bull. Soc. Chim. (4), 15. 

Eppinger (1920). Die hepatolienalen Erkrankungen, Berlin. 

Frey (1920). Deutsch. Arch. klin. Med. 133. 

Gabbi (1893). Betirdge u. allgem. Path. 

Hamburger (1904). Osmotischer Druck und Ionenlehre, Wiesbaden. 

Hunter (1892). Lancet. 

Jacobsthal (1921). Verhandl. deutschen path. Gesell. 66. 

Koelliker (1854). Micr. Anatomie oder Gewebelehre des Menschen, Leipzig. 
Pel Jr. (1911). Onderzoekingen by miltlooze dieren, Amsterdam. 

Polak Daniéls and Hannema (1916). Folia microbiologica 4. 

Pugliese and Luzatti (1900). Arch. Ital. Biologie. 

Strisower and Goldschmidt (1914). Zeitsch. gesammt Med. 14. 


LXXX. THE ENZYMES OF THE LATEX OF THE 
INDIAN POPPY (PAPAVER SOMNIFERUM). 
By HAROLD EDWARD ANNETT. 
Agricultural College, Cawnpore. 


(Recewed October 3rd, 1922.) 


THIs paper must be considered in the nature of a preliminary note. The work 
done has simply been qualitative and pressure of other work has prevented 
a more detailed treatment hitherto. In view of certain interesting results 
obtained and as it is possible the work may have to be discontinued it is 
deemed worth while to put it on record. 

When poppy capsules are lanced for opium! the latex which immediately 
exudes varies considerably in colour from capsules in the same field and from 
the same pure race of poppy and even at times from capsules on the same 
plants. It may be pure white, smoky grey, light pink or deep pink in colour. 
It rapidly darkens to some shade of brown, either light chestnut or deep 
mahogany and at times it is almost black. When this opium is dried at air 
temperature and powdered the resulting powder may vary from a pale straw 
colour to almost black. 

Moreover in a field sown with a single pure race of poppy, samples of 
opium collected on the same day at the same time under apparently identical 
conditions, on being subsequently air dried and powdered vary considerably | 
in colour, 7.e. from pale straw colour to almost black. 

It would appear probable that the change is due to oxidation and one 
would expect to find powerful oxidising enzymes in the latex. 

That the change is due to oxidation seems proved by the following experi- 
ments. 

A. Some of the fresh latex diluted slightly with water to make it more 
liquid was drawn into the bulb of a Lunge nitrometer over mercury. The 
tap was turned off, thus leaving some of the latex exposed in the cup of the 
nitrometer. The remainder of the latex was effectually sealed up by the 
mercury out of contact with air. The latex exposed to air rapidly darkened. 
That sealed up after three months still shows no sign of darkening. 

B. Portions of fresh latex were placed in small distillation flasks which 
were exhausted with a Geryk pump and sealed off. After three months these 
‘samples still show no sign of darkening whereas latex in similarly sealed flasks 
which were afterwards cracked rapidly darkened. 

1 For details of method of lancing see Annett [1921]. 


766 H. E. ANNETT 


Latex kept in closed vessels over alkaline pyrogallol darkens because the 
complete absorption of oxygen by this reagent is a slow process unless the 
vessel is vigorously shaken and this is not feasible in the presence of the latex. 

These preliminary experiments would appear to show that the darkening 
of the poppy latex is due to an oxidation process. It therefore appeared of 
interest to examine the latex for the presence of oxidising enzymes. 

(1) Some fresh latex was shaken with water and divided into portions 
(a) and (6). The flask containing (a) was placed in boiling water: for two 
minutes. (a) and (b) were then both filtered. Next day the filtrate from (6) 
was much darker than the filtrate from (a), thus indicating that the boiling 
had destroyed the oxidising enzyme. 

(2) Latex mixed with water and fresh guaiacum tincture gave no blueing 
but when the test was repeated in presence of H,O, a strong blue colour was 
immediately produced. It would therefore appear that a peroxidase is present 
in the latex. Spence [1908] working with rubber latex obtained a similar 
result but on dialysis with water for 24 hours he obtained both oxidase an 
peroxidase. 

(3) Some fresh latex was immediately tested for oxidising enzymes with 
the following results: 

(a) Latex and water + guaiacum tincture gave an immediate green colour 
but when the test was repeated in presence of H,O, no reaction was obtained. 

(b) A similar test with benzidine gave a deep colour immediately in 
absence of H,O,, but when the test was repeated in presence of H,O, no 
reaction was obtained. 

(c) A similar test with tyrosine suspension in absence of H,O, gave 
darkening within 15 minutes and next day the liquid was very dark. On 
repeating the test in presence of H,O, no darkening was obtained. 

(d) Pyrogallol gave similar results, i.e. oxidation in absence of HO, but 
no reaction in its presence. 

This result is directly opposed to that obtained under (2) above. The latex 
used under (3) was, however, used within 20 minutes of the lancing of the 
capsules while that used under (2) was probably two hours old. Some of the 
same latex used for (a), (b), (c) and (d) above tested about an hour later gave 
similar results except in the case of benzidine with which reagent a strong 
reaction was obtained both with and without H,O,. 

In these cases controls were done with boiled liquids and all gave negative 
results. 

This would indicate that the relative freshness of the two lots of latex 
used had nothing to do with the different results under (2) and (3). 

(4) 10-15g. of latex, collected for experiments in (3) above, were rubbed 
up in a mortar with distilled water and poured into a cleaned goat’s stomach. 
This was allowed to diffuse in a vessel through which a slow stream of tap 
water was passed, The tap water was passed through thymol before reaching 
the diffusing vessel. Next day the walls of the goat’s stomach were quite 


ENZYMES OF POPPY LATEX 767 


black, indicating the presence of tyrosinase in the latex. 100 ce. of the dialysed 
liquor, the total volume of which was about 300 cc., were removed for enzyme 
tests. It was very faintly acid in reaction to litmus paper. 40 cc. of the liquid 
were boiled in a water-bath for five minutes. The table sets out the results 
of experiments then performed with the boiled and unboiled dialysed liquid. 
3 cc. of the liquid were taken for the test in each case: “—” indicates no 
reaction, “+” a reaction, and “‘+-+”’ a very powerful reaction. Where H,0, 
was used one drop of the 10 volume reagent was employed. 


Unboiled dialysed liquid Boiled dialysed liquid 
Reagent NoH,0, With H,0, NoH,0, With H,0, 
Guaiacum ++ - - - 
Benzidine ++ = = ~ 
f + } 
Pyrogallol 1 ++ next day( - ~ - 
; + 
Tyrosine ++ next day i y, 


After another two days more dialysed liquid was removed from the goat’s 
stomach. It was then quite neutral to litmus. The above tests were repeated 
with exactly similar results except that the unboiled liquid gave a very faint 
reaction with benzidine in presence of H,Q,. 

Another portion of the dialysed liquid was filtered and tested as above. 
Almost identical results were obtained. That is to say reactions were obtained 
with guaiacum, benzidine and tyrosine in absence of H,O,. In the case of 
guaiacum and tyrosine no such reaction was obtained in presence of H,O,, 
but in the case of benzidine a very faint reaction was obtained in the presence 
of H,O,. 

(5) Some latex was collected and tested within 10 minutes of lancing with 
guaiacum and a deep green colour was immediately obtained. The test when 
repeated in presence of H,O, gave an even more intense reaction. This is 
again a contradictory result. 

The rest of this latex was shaken with water and divided into two portions 
A and B. A was rendered faintly alkaline with NaHCO, and one-half of it, 
A’, was boiled in a water-bath for five minutes. One-half of B was boiled 
similarly and is called B’. A, A’, B and B’ were then tested with guaiacum, 
benzidine, and tyrosine with and without addition of H,O,. 

The boiled solutions gave no reaction in any case. The reaction with 
guaiacum in case of both A and B solutions was very doubtful probably owing 
to the latex having been too largely diluted. 

With benzidine both in the case of A and B a very strong reaction was 
obtained in absence of H,O,, the reaction being distinctly stronger in the 
alkaline solution. In the presence of H,O, a faint but positive reaction was 
obtained with both solutions A and B but the reaction was much weaker than 
in absence of H,O,. 

With tyrosine a practically similar result was observed namely a strong 
reaction in absence of H,O, in both solutions A and B,-but it was not more 


~ 


768 H. E. ANNETT 


marked in the alkaline solution. In presence of H,O, a very faint but positive 
reaction was obtained in the case of both A and B solutions. 

(6) Some of the latex collected for tests under section (3) was immediately 
after collection rubbed up with water and poured into a tall cylinder. Toluene 
was poured on the surface and the cylinder allowed to stand undisturbed. 
The insoluble portions of the latex settled out at the bottom of the cylinder. 
It will be remembered that when tested under section (3) a strong reaction 
was given with both guaiacum and benzidine in absence of H,O, but there 
was no reaction in presence of H,Q,. 

After the cylinder had stood for eight days some of the liquid was drawn 
up from the bottom with a pipette. It contained pieces of the insoluble matter 
of the latex. This liquid reacted with guaiacum both in presence and absence 
of H,O, the reaction being slightly stronger in the latter case. It reacted with 
benzidine in absence of H,O, but not when H,O, was present. 

(7) It appeared of interest to test old powdered opium for oxidising 
enzymes. 

Some opium powder two years old was rubbed up with water and dialysed. 
‘The dialysed liquid was then tested with guaiacum and benzidine in presence 
and absence of H,O,. Guaiacum gave no reaction in either case; benzidine 
reacted strongly in absence of H,O, but in presence of H,O, there was no 
reaction. 


Tests for other enzymes in the latex. 


We have been unable to find any reference to work on the enzyme content 
of poppy latex, with the exception of a statement that it contains only a small 
amount of protease [Czapek]. 

Fifty capsules were lanced and the latex which exuded was immediately 
scraped off. It was macerated with sand and 100 cc. water. After straining 
through coarse cloth the filtrate was divided into two portions A and B. 
The latter was boiled. The usual tests for enzymes were then carried out on 
the following substances: 1 % starch solution, 1 % cane sugar solution, 1 % 
a-methyl glucoside solution, 1 °% urea solution and white flour extract (10 %). 
The result of these tests indicated the complete absence of amylase, invertase, 
maltase and urease. The test for protease indicated the possible presence of 
protease in small quantity as a very faint positive result was obtained. 

Emulsin was tested for with amygdalin as in Armstrong and Horton’s 
test, and a negative result obtained. 

There was one interesting observation we made in all these tests, namely 
that the liquids in the tests with unboiled latex all darkened distinctly with 
reference to the control experiments carried out with boiled latex. This is 
another indication of the presence of oxidising enzymes. 


ENZYMES OF POPPY LATEX 769 


CONCLUSIONS. 


1. The darkening of the latex of the Indian Opium Poppy is a process of 
oxidation. 

2. The latex has a powerful oxidising action on guaiacum tincture, pyro- 
gallol, benzidine and tyrosine in the absence of H,O,. In almost all the 
experiments tried it was found that H,0, inhibited these reactions. 

Chaudat and Staub [1907] found that hydrogen peroxide retarded the 
action of tyrosinase. 

Bach [1906] and also von Fiirth and Jerusalem [1907] have found that 
though hydrogen peroxide in certain amounts retards the action of tyrosinase, 
yet minute quantities accelerate the action of the enzyme. 

3. The dialysed latex both before and after filtration also oxidises guaiacum 
tincture, pyrogallol, benzidine and tyrosine in absence of H,O, but the reaction 
is inhibited in the presence of this reagent. 

4. The actions on benzidine and tyrosine were particularly powerful. In 
the case of tyrosine the darkening always appeared as a surface film on the 
liquid and gradually diffused downwards. 

5. Opium powder stored for three years has been shown to possess an 
oxidising enzyme which acts on benzidine. 

The author has recently shown that dry opium powder loses a considerable 
proportion of its morphine on storage. This may be due to the action of 
oxidising enzymes. 

6. The following enzymes were tested for in the latex and not found, viz. 
amylase, invertase, maltase, emulsin and urease. There were indications of 
a weak proteolytic activity. 


REFERENCES. 


Annett (1921). Mem. Dept. Agric. India, 6, Nos. 1 and 2. 

Bach (1906). Ber. 39, 2126. 

Chaudat and Staub (1907). Arch. Sci. Phys. Nat. 24, 172-191 (vide “The Oxidases,” J. H. 
Kastle, p. 85). 

Czapek. Biochemie der Pflanzen, 3, 716. 

von Fiirth and Jerusalem (1907). Beitrdge, 10, 131. 

Spence (1908). Biochem. J. 3, 165, 351. 


LXXXI. THE BLOOD OF EQUINES. 


By CHRISTIAN PETRUS NESER. 


Preliminary Communication} from the Division of Veterinary Education 
and Research, Onderstepoort, Union of South Africa. 


(Received October 3rd, 1922.) 


Technique. Without offering details of technique it may be briefly stated 
that blood, drawn from the jugular vein except where otherwise mentioned, 
was collected in 10 cc. bottles containing a measured adequate amount of 
sodium citrate solution. From this were determined, (a) percentage volume 
of erythrocytes by centrifuging in specially prepared uniform tubes of 2 cc. 
capacity, blown out slightly at the sealed end to facilitate rapid sedimentation; 
(b) corpuscle count after diluting 1 in 200 with Hayem’s fluid for erythrocytes, 
and 1 in 10 with 0-5 % acetic acid for leucocytes, the Biirker chamber being 
used in both cases; (c) haemoglobin content by Sahli’s method. 

Smears were made by Craandyk’s [1918] method and also by a modifica- 
tion of Ehrlich’s method which is so simple and effective as to merit brief 
description. A small drop of blood is quickly transferred to a slide by means 
of a suitable platinum loop, and a cover slip, as broad as the slide, immediately 
lowered over it. The droplet at once spreads as a thin circular layer, and on 
drawing the slip lengthwise over the slide all the blood is left upon the latter 
as a thin film in which the distribution of leucocytes is remarkably uniform, 
and eminently suitable for differential count after Giemsa staining in the 
usual way. 

Limits of error. From numerous observations expressly designed to deter- 
mine the degree of accuracy of methods in conventional use, it was concluded 
that (a) the centrifuge gives very accurate and consistent data, (b) red counts 
are liable to an error of up to 10 % even with the improved Biirker chamber, 
(c) white counts are less liable to error, (d) the Sahli reading cannot be relied 
upon to within five scale units, (e) smears made by the described modification 
of Ehrlich’s method show a very regular distribution of leucocytes. 


THE RED CORPUSCLES. 


Influence of work. Comparative study of the blood of different horses 
showed a remarkable difference in percentage volume of erythrocytes from 
individual to individual, but a fairly constant figure for any given individual 
over short intervals of time. The outstanding difference between the blood 
of different horses was at first very puzzling. Since all were fed alike, diet 


1 The full paper, giving detailed protocols, will appear in the forthcoming “Report of the 
Director of Veterinary Education and Research” (P.O. Box 593, Pretoria). 


THE BLOOD OF EQUINES 771 


was not a factor. Age was excluded by statistical comparison of data. That 
sex played no part was evidenced by similar variation in mares and geldings. 
Since all the animals under observation had been purchased for the horse- 
sickness experiments of the institution, they were a mixed lot whose histories 
were generally unknown, and no mere inspection sufficed to explain the ob- 
served variations. 

The clue, however, was given by two animals (laboratory Nos. 11144 and 
11775) one of which was permanently stabled and showed a percentage volume 
of 23, while the other was used as a saddle horse and showed a figure of 
approximately 40. This remarkable difference at once suggested that work 
was the deciding factor in determining the percentage volume of erythrocytes 
in horse blood. Following up this clue, horses were grouped according to the 
work performed, as “fast working,” “permanently stabled,” and “other 
horses.”’ It was at once noted, without exception, that the blood of fast-working 
horses showed a high percentage volume of erythrocytes, or high red count, 
while in horses stabled for six months or more the figures were correspondingly 
low. With the third group the data were variable, but it is probable that if 
the previous histories of these miscellaneous animals had been known, many 
of them would have been classified in one of the other groups. At a later 
date it was found possible to procure the blood of a few race-horses in training, 
and their high volume of red corpuscles bore out the earlier conclusions in 
most striking fashion. Graph I represents a summary of the data from 
over 200 animals from all sources, Graph II the data from three race-horses 
in various stages of training, while Graph III illustrates the change actually 
occurring during hard training of three young race-horses. 

These data leave no doubt that hard fast work brings about a marked 
increase in the percentage volume of red blood corpuscles in the horse. The 
contrast shown in Graph II, for three race-horses from the same stable, is 
particularly striking, the figure for the fully trained animal showing that 
52 %, or more than half the total volume of blood, may be made up of erythro- 
cytes. This corresponds to a count of approximately 12 million as against 
an average count of about 8 million for the ordinary slow working horse, and 
less than 6 million for permanently stabled laboratory horses. 

These already conclusive observations are further clinched by reference to 
Graph III, which shows a steady increase in numbers of red corpuscles during 
the actual course of hard training. In five weeks the count increased from less 
than 7 million to over 9 million in one case, and from 6-3 million to 8-8 million 
in a second case. In the third case training was interrupted owing to injury 
about the third week, the rapidly rising count then remaining stationary. 
Of special interest is the further fact that although the high count follows so 
rapidly in the wake of hard training, the reverse effect, or diminution of red 
blood corpuscles with rest, is a slow process. Thus in two young horses a 
decrease of two millions occurred only after three months, while in several 
old animals the decrease after five months was hardly noticeable at all. 


772 


% volume of red corpuscles 


50 


40 


ie) 
\) 


20 


C. P. NESER 
Graph I. Graph II. 


50 


40 


WV 


SS 


YL, 


% volume of red corpuscles 


2 WMI WK 


Y 
WY U 


oS 


AW 


SSSA W 


beads 


/ Y é a 
Lh Wi ook ZA UZ7. 
- , Stabled horses. pa erat 
, Medium horses. , Untrained. 
C, Fast working horses. ; B, Partly trained. 
D, Race-horses in full training. C, Fully trained. 


Showing the influence of hard fast work upon the % volume of red corpuscles. 


Graph ITI. 
9 alec 
sae RE 
ee te 
aa a 
28 fo " 
| 7 7 wy 
2 Ving hs ‘ 
Be tea 
ye t / 
2 7 rsh ye Y 
“4 "4 sf? 
4 
6 


Weeks under observation. 


Showing the influence of time upon the red count in three 
young race-horses, during hard training. 


THE BLOOD OF EQUINES 773 


Relationship between volume, count, and measurement of erythrocytes. In 
numerous cases, the percentage volume as determined by the centrifuge, the 
actual count in the Biirker chamber, and the diameter of the cells as measured 
in smears, were compared for the same horses. 

The diameter of individual corpuscles varied between 4 and 8p: but, 
with one exception, the average of 200 cells remained fairly constant at 5-5. 
In respect, therefore, to average size of erythrocytes, different horses are very 
uniform. 

The ratio between percentage volume and numerical count of red blood 
corpuscles varied between 4k and 4-7k, with an average of 4:35k—k being a 
constant dependent upon the units of measurement used. From the relatively 
constant average ratio, the observed liability of the red count to an error of 
10 %, and the observed constancy of centrifugal readings, it may be con- 
cluded that the variations observed are mainly due to errors in counting and 
not to real variations in the average size of the corpuscles. These considera- 
tions led to the conclusion that a count calculated from percentage volume 
on an assumed ratio of 4:35k is even more correct than an actual count; at 
least for the blood of healthy horses. It is quite certain that the percentage 
volume and the red count give the same information, and that the observed 
increase of the former with work is due to a real increase in the number of 
normal red cells per unit volume of blood; normal also in respect to haemo- 
globin content as determined by the Sahli method. This increase in erythro- 

cytes is probably absolute, and a genuine response of the blood-forming 
tissues to the increased demands made upon the oxygen-carrying capacity of 
the blood. 

Diurnal variations. Quite remarkable differences in percentage volume of 
erythrocytes were observed from day to day, and even from hour to hour. 
Numerous experiments in which horses were worked, fasted and then fed, 
allowed to go thirsty and then watered, all failed to account for the variations 
observed in jugular blood. The data obtained indicated that any change in 
the concentration of the blood, arising from these factors, is at most very 
transient. The real explanation of the variations, however, was at once found 
on tapping at different points in the circulation. In comparing blood from 
the jugular vein with that from the ear, it was found, with very few exceptions, 
that the latter gave a higher red count than the former, and was also subject 
to far greater variation. The difference between ear blood and jugular blood 
was greatest in animals feeding at rest, and least in animals excited or at work; 
an increase in one count being generally associated with a decrease in the 
other. . 

These observations indicated the mechanical state of the circulation as 
the important factor in determining distribution of erythrocytes throughout 
the body, and the few simple experiments recorded in Graph IV provided 
direct evidence of its operation. 


774 C. P. NESER 


Graph IV. 
Fig. 1. Fig. 2. Fig. 3. 
10 
fO} 
if aes ax Z : is 
~o=t \ 
9 / t ei \ 
co) / \ 4 \ 
_ . / a 
5 f \ ae é 
» / ‘ : 
s 8 N. 4 X 
Ss «|! \ ev 
o 
3 3 | i if _* 
a. le W, ye- 
> ~ Beg, 4 eee Ss. far | jus 
> a = 3 a bP Fa 


0 1° 2 8 4.6. 6..0. 10 20.90.40.-0- 30 07am 
Hours after food. Horse excited at Minutes fed upon return Minutes taken from 
time of feeding. from work. food and excited. 


Showing the influence of the mechanical state of the circulation, upon the distribution of 
the red cells. 


Fig. 1 of Graph IV represents the changes in red corpuscle count in an 
animal excited just prior to feeding; Fig. 2 those of an animal rested and fed 
immediately upon return from work; Fig. 3 those of an animal taken from 
the manger and threatened for 30 minutes by exhibition of a familiar whip. 
In all these cases the counts on peripheral and systemic blood converge most 
closely when the circulation is most active, and disagree most widely when 
the circulation is more sluggish. A less active circulation thus results in a 
concentration of the red corpuscles in the peripheral capillaries, possibly owing 
to relative increased lymph formation and possibly owing to sedimentation. 

On many occasions it was also noticed that both ear and jugular red 
counts had increased or decreased together; observations which may be ex- 
plained by assuming concentration or dilution in some other part of the circu- 
lation. — 

The general results, however, show that in horses the jugular blood is far 
more representative of the average circulation than is blood taken by ear 
puncture; that factors beyond the control of the observer may give very mis- 
leading results when ear blood is studied; but that even jugular blood is not 
free from variation. 

Pregnant and nursing mares. The results obtained from a few animals of 
this class may be summarised by stating that no influence of gestation or 
lactation, as such, could be established. The red corpuscle counts varied 
greatly between individual mares, being always highest in those with a past 
history of active life, and such differences obscured possible variations cen- 
tring around sex, 


THE BLOOD OF EQUINES 775 


Foals. In foals the observed red count was very high just after birth, but 
in two months had decreased considerably. Two cases will serve to show that 
the most rapid fall occurs in stabled foals at rest: 


Percentage volume of red corpuscles 
A 


Foal At birth After 2 months 
(a) Permanently stabled 48 30 
(6) Running with mother 42 36 


Again the influence of physical activity is manifest. 

Donkeys and Mules. The red cell picture of donkey and mule blood differs 
in an interesting way from that of the horse. The average diameter of the 
erythrocytes is approximately 6-2, for both donkey and mule as against 
5-5 for the horse. For the donkey the ratio of percentage volume and 
numerical count was generally found greater than 5k, while for mules it 
approximated 4-35k, the figure for the horse. The inference is that the cor- 
puscles of the mule are thinner than those of donkey and horse. 


THE LEUCOCYTES. 


The classification of the leucocytes adopted in this work is that which is 
generally accepted to-day, as fully discussed by du Toit [1917] in his article 
on bovine blood. Equine leucocytes are not unlike those of bovines, as studied 
in stained smears, with the exception of the eosinophiles and basophiles. The 
granules of the latter cells are very large and form a characteristic feature of 
equine blood. 

The number of leucocytes in the blood of healthy horses is very variable, 
not only as between individuals, but also in the same animal from time to 
time. The general range is from 5 to 20 thousand per cubic millimetre. In 
any individual animal the hourly and daily variations are not great, although 
exceptions are not uncommon. 

Several experiments were undertaken to ascertain the influence of ordinary 
non-pathological factors upon the white count. Results may be summarised 
as follows: 


(a) Moderate variations in supply of food and water showed no definite 
effects, and the jugular blood certainly reflected no evidence of any “digestive 
leucocytosis.” 


(b) Exercise always increased the white count of jugular blood, in some 
cases by over three thousand per cmm. On cessation of exercise the numbers 
often decrease again very rapidly, even to a figure below the pre-exercise 
count. 


(c) Simultaneous observations upon ear and jugular blood yielded no 
conclusive results, except with exercise. In this case the results obtained 
indicate that the distribution of leucocytes also depends largely upon the 
mechanical state of the circulation. 


776 C. P. NESER 


To ascertain the extent to which technique of sampling influenced the 
white count, variations were made upon the usual procedure of first slapping 
the ear and collecting from a large puncture. A small puncture was made and 
blood obtained by squeezing at the base and at the apex. The following data 


represent one case: 
Leucocyte count 


Ear, ordinary technique 11-3 x 108 
Ear, squeezed at base 7-4 x 108 
Ear, squeezed at apex ‘14:3 x 108 
Jugular blood for comparison ——-12-4 x 108 


From these marked variations it may be concluded that the white cells 
cling to the walls of the capillariesand the wound, especially if the blood flow 
is slow. When the circulation is rapid, or when the blood is forced out under 
_ considerable pressure, the leucocytes become detached, or are prevented from 
clinging, with consequent increase in apparent numbers. This explains why 
the ear white count is usually lower than the jugular. 

Differential Counts. Differential leucocyte counts, obtained by the useful 
method (modified Ehrlich) already described, showed the following general 
figures: 


Differential count of equine Lympho- Mono- Neutro- Eosino- Baso- 
leucocytes cytes cytes philes philes philes 
Horses, average of 7 each: . 
(a) Stabled 36 4 54 5 1 
(6) Medium 39 4 52 + 1 
(c) Fast working 40 5 50 4 a 
(d) General average 38 f 53 4 1 
Normal variations 45-30 8-2 60-45 9-3 3-0 | 
Extreme variations in 200 cases 50-25 9-0 62-54 15-1 3-0 
Donkeys, average 53 4 34 8 1 
Mules: ; 
(a) Clinically healthy 41 4 49 6 ae 
(6) Slight injuries 32 3 61 4 _ 
(c) Inoc. earlier against anthrax 52 3 39 6 1 


The most interesting feature of this comparative table is the curious re- 
versal of the figures for lymphocytes and neutrophiles, as between horses and 
donkeys. The horse shows lymphocytes 38 and neutrophiles 53 as general 
average, while the donkey shows neutrophiles 34 and lymphocytes 53. The 
count for mules approximates that of horses, when clinically healthy animals 
are considered, The slightly higher neutrophile count in group (6) mules may 
perhaps be accounted for by the slight injuries from which they were suffering. 
The high lymphocyte count in group (c) mules may perhaps be due to the fact 
that these animals had been inoculated against anthrax some months earlier, 


Breeding mares and foals. 


To these data may be added a few observations upon mares and foals. 
Taken just before or after parturition, the mares showed a very high neutro- 
phile count, even over 70 %. An even higher neutrophile count, about 80 %, 
was characteristic of foals shortly after birth. After about two months this 
neutrophilia had disappeared, provided the animals remained healthy. 


THE BLOOD OF EQUINES 777 


Influence of other factors. 


Various minor deviations from the usual mode of life did not appear to 
influence the differential count in any regular way; except for the eosino- 
philes, which were generally increased by water-drinking after a period of 
thirst or fasting. Graph V offers a typical example of transient eosinophilia 
following large consumption of water after dry feeding. 


Graph V. 

12 9. 

1b. ae] 

10--S 3 / 4 

8 r%\ 

9-8 +6 \ 

st § et \ 
a 15 >! N, 
3g 7s ! .. 
5 6F > ! 
EVE: 
Pa aber 3 


if 
ec. 


—- YN @© 
| 


9 10 1 12 1 2 3 4. 5 
Hour of day 
Showing the influence of food and water upon the % of eosinophiles in one horse. 


GENERAL DISCUSSION. 


It is of interest to compare the findings recorded in this paper with the 
limited data already available in the same direction. References to healthy 
equine blood in the general haematological literature appear to be very scanty, 
and it is to be regretted that South African library facilities do not always 
allow of consultation of such references as can be tracked by abstract. 
Burnett quotes various authors as obtaining erythrocyte counts ranging from 
6-3 to 8-5 million, and leucocyte counts varying between 5-6 and 15 thousand, 
for adult horses. The white count is approximately that recorded in this 
paper, but the red counts quoted by Burnett suggest that other workers have 
contented themselves too readily with animals of one class. The data now 
recorded show a much wider range, considerably below 6-3 million for per- 
manently stabled horses and even up to 12 million for race-horses in hard 
training. 

Frei [1909], working in this laboratory, compared the centrifuge and the 
counting chamber, obtaining 47k for the ratio between percentage volume 
and numerical count, and preferring the volumetric method for determining 
the proportion of erythrocytes. His ratio is thus about 8 % higher than the 
one now offered on a more extensive series of observations. The horses 


Bioch. xv1 51 


778 C. P. NESER 


available to Frei belonged to the same miscellaneous class as our own original 
lot, and showed variations from 22 to 43 for corpuscular volume. The aston- 
ishing thing is that no serious attempt was made to elucidate the nature of 
these differences, and the extraordinary effect of regular work seems to have 
wholly escaped his notice. 

In regard to corpuscular dimensions, some authors give the average size 
of erythrocytes as 5-5, but others give figures as high as 5-8u. The former 
figure is the average now established upon several hundred horses. 

The differential counts quoted by Burnett for horses show higher neutro- 
phile and lower lymphocyte percentages than those recorded in this paper. 
Stephens and co-workers [1921] have called attention to the variation in 
distribution of leucocytes in different parts of smears made in the ordinary 
way, and since our own technique has overcome this difficulty, the data now 
recorded would seem the more reliable. , 

By far the most striking observations recorded by us concern the remark- 
able influence of regular exercise upon the proportion of erythrocytes in horse 
blood. The available recent works upon horse blood all deal with pathological 
conditions, and nowhere does any reference seem to be made to the funda- 
mental influence of normal work. In the various standard text-books upon 
Physiology and Haematology, the same omission is conspicuous. Dealing, as 
they do, largely with the human subject, in which normal variations are less 
noticeable, no stress is laid upon this factor, which plays so remarkable a 
role in the physiology of the horse. It is generally accepted that slight varia- 
tions occur with mode of life, but whether or not these are specifically related 
to exercise is hardly discussed, although factors such as “altitude” are almost 
unduly stressed—perhaps because immediately detectable. Some workers on 
human blood have found that “robust” people have a higher, and “obese”’ 
people a lower, red count. The numbers are also lower for women, and it has 
been recorded that school children show a higher count at the end of a long 
vacation, than at the beginning. It is true that the variations generally re- 
corded in healthy human individuals are registered in hundreds of thousands 
as against the tremendous variation of 4 million to 12 million between the 
stabled laboratory animal and the fully trained race-horse, but is it not 
possible that the minor differences in the human are due to differences in the 
amount of exercise rather than to other factors? 

For the horse, the facts now recorded are perhaps not so surprising when 
considered in the light of its evolutionary history. The ancestor of the modern 
horse depended upon speed and endurance for its very existence; two qualities 
conditioned by a highly efficient circulatory system, and the demand for a 
high oxygen-carrying capacity of the blood. Is it then astonishing that the 
modern horse is born with a potential capacity for high speed and great 
endurance, and that during hard training his store of haemoglobin should be 
increased as fast as his muscles develop, so that finally more than half the 
total blood volume of a “finished race-horse” consists of erythrocytes? 


THE BLOOD OF EQUINES 779 


In how far does this phenomenon of increase of red corpuscles with 
exercise, so strikingly shown here for the horse, apply to other animals? The 
problem at least seems a promising one for future investigation. 


REFERENCES. 


Burnett. Clinical Pathology of the Blood of Animals (Taylor and Carpenter, Ithaka). 

Craandyk (1918). Folia Haematologica, 23, Heft 2. 

Frei, W. (1909). Physical Chemical Investigations into South African diseases: Transvaal De- 
partment of Agriculture. Report of the Government Veterinary Bacteriologist for the year 1907-8. 

Stephens, J. W. W., Yorke, W., Blacklock, B., Macfie, J. W. S., Cooper, C. Forster, and Carter, 
Henry F. (1921). Annals of Trop. Med. and Parasit. 

du Toit, P. J. (1917). Archiv wissenschaft. prak. Tierheilkunde, 43, Heft 2 and 3. 


51—2 


LXXXII. THE FOOD VALUE OF MANGOLDS AND 
THE EFFECTS OF DEFICIENCY OF VITAMIN A 
ON GUINEA-PIGS. | 


By ELLEN BOOCK anp JOHN TREVAN. 
From the Wellcome Physiological Research Laboratories. 


(Recewed October 11th, 1922.) 


Our colleagues Glenny and Allen [1921] presented to the Pathological 
Section of the Royal Society of Medicine an account of an investigation into 
an epizootic amongst a stock of guinea-pigs. They presented conclusive evi- 
dence that the epizootic could be entirely controlled by alteration of the diet. 
The diet which resulted in the outbreak consisted of bran, oats, water and 
mangolds. Substitution of the mangolds by cabbage, grass or lucerne stopped 
the epizootic. When guinea-pigs in adjacent runs were fed on the two diets, 
those on mangolds were attacked, whilst those with grass remained healthy, 
although no precautions whatever were taken to prevent infection of the 
healthy animals. Isolation was not carried out, the runs were not disinfected, 
and the attendants handled both groups of animals indiscriminately. These 
results are of such importance that the following attempt was made to de- 
termine the factors or deficiencies in the mangold which rendered the guinea- 
pigs sensitive to the epizootic. A few experiments were done by Glenny in 
which an alcoholic extract of carrot was given to the guinea-pigs with a view 
to supplying a possible deficiency of vitamin A. These indicated that the 
supply of vitamin A, although it reduced the incidence of the disease, did not 
stop it, and the guinea-pigs did not put on weight. 

The following experiments were begun after the epizootic had been con- 
trolled by dietary means, and the effect of the diet was re-investigated. 

Chart I shows the weight curves of four guinea-pigs fed on unlimited bran, 
oats, mangolds and water, the animals eating about 40g. of mangold each 
per diem, and Chart II those of guinea-pigs kept on a control diet of bran, 
oats, autoclaved milk and greenstuff. These latter gained weight rather faster 
than Miss Hume’s [1921] “standard” guinea-pigs (43-7 % in 25 days, as 
against 34-2 % in Miss Hume’s experiments). 

It was found that all the guinea-pigs placed on the mangold diet even- 
tually succumbed within a period of about two months. In Chart I, the 
guinea-pigs were about 200 g. weight when started. In Chart III, they were 
younger—about 150 g.—and the younger guinea-pigs succumbed earlier than 
the older ones. 


GUINEA-PIGS AND VITAMIN A 781 
218 GMS. 


228 | 


20 
r 100GMS. 


+ 


20 DAYS. 


INTUSSUSCEPTION. KILLED. 


The charts are drawn to different scales indicated in each case. The oblique straight lines 
on each chart represent the greatest and least rate of growth shown in the experiment with a 
normal diet represented in Chart II. The abscissa for each curve is at a different level; the figure 
at the beginning of each curve shows the initial weight of the guinea-pig in grams. 
Chart I. Effect of basal diet of bran, oats, and mangolds. One of the animals had an intus- 
susception and was killed. We have met this condition fairly frequently in animals under 
200 g. in weight which have beer deprived of greenstuff. 


1006MS. 
221 
? 
20 DAYS. 


Chart II. “Normal” diet consisting of bran, oats, lucerne, grass and autoclaved milk. This diet 
was used as a control for some experiments on scurvy. We have obtained since writing this 
paper even better results than these, with the ‘‘ synthetic ”’ diet described later. 


782 E. BOOCK AND J. TREVAN 


100GMS. 
10 DAYS. 
176 GMS. ¥ 
INFECTED. 
URINE ACID. 


158 


SALT 1 iwescrep. 
: URINE ACID 


Chart III. Effect of basal diet of bran, oats, and mangolds on younger guinea-pigs. It will be 
seen that these animals died sooner than those in experiment 1. 


| Gipaa son URINE ACID. 


225 


SALT x * 
CASEIN x 


24 


1.0 CC, CODLIVER OIL. 


224 


PAPER ADDED AT ARROWS. 


Chart IV. Guinea-pigs on basal diet as in I and ITI, to which was added packing paper at the 
arrows. One of the animals died the same day, the others started to grow again, one growing 
for over two months before decline set in. As the result of experiments which had been 
going on in the meanwhile, salt mixture (calcium lactate and sodium chloride), caseinogen 
and cod-liver oil were added to the diet (see further charts). It will be seen that the limiting 
factor in this guinea-pig was the absence of something supplied by cod-liver oil—presumably 
vitamin A, 


It was noticed whilst weighing the animals on the deficient diet, that they 
eagerly nibbled at any pieces of paper or cardboard which they could reach; 
so clean white packing paper was placed in the cages. The guinea-pigs each 
consumed about’5 g. of paper every day. Chart IV indicates the effect on 


GUINEA-PIGS AND VITAMIN A 783 


the growth, and it will be seen that out of four guinea-pigs, one died on the 
first day that the paper was given, but the other three gained in weight as a 
result, and one of the three continued to gain for over two months before a 
decline set in. Suggestions are put forward below as to the explanation 
of this. 

Post-mortem examinations revealed several abnormalities in the guinea- 
pigs whilst on the mangold diet. 

As in Glenny’s experiments, there was constantly some sign of an infection 
present, either 

(a) Alkaline stomach contents. 

(b) Pneumonic consolidation or abscess of the lung. 

(c) Purulent pericarditis. 

(d) Enteritis. 

When the urine was tested during life and post-mortem, it was always 
found to be strongly acid, and therefore quite free from the normal precipitate 
of phosphates which renders the urine of the healthy guinea-pig turbid. This 
observation led to the suggestion that the mangold, bran and oats diet is de- 
ficient in certain basic constituents, and following the indication of McCollum’s 
feeding experiments on rats [1920] (from which he concludes that roots and 
seeds, besides other deficiencies, lack the three inorganic elements calcium, 
sodium and chlorine), it was decided to add a salt mixture to the bran, oats 
and mangold diet. A mixture of calcium lactate and sodium chloride was 
made and mixed as thoroughly as possible with the bran and oats, about 
1-5 g. of calcium lactate and 1-0 g. of sodium chloride per guinea-pig per diem 
being given. This resulted in a return of the urine to the normal alkalinity, 
when examined during life and post-mortem, and improved the growth curves. 
(First part of Chart V.) 

In connection with this, we made an examination of the ash, by drying 
mangold at 100°, and then incinerating in a muffle furnace. Magnesium, 
sodium, potassium and iron were found to be present, but no qualitative test 
for calcium could be obtained, and in attempting to estimate calcium by 
Cahen and Hurtley’s method [1916], no potassium permanganate was used up 
in the final titration. 

The beneficial effect of the salt mixture is therefore probably due to its 
supplying the extra calcium. The packing paper also helps in the same 
direction, for analyses of the ashed paper showed that there i is a certain amount 
of calcium present: 


» 


Percentage of ash in paper... ...._ :1:7% _—_ Ist determination. 
s ae a ne sah PG. 2nd ue 
,, calcium in paper ... 0126 % Ist a 


> > ”> ” 0- 190 yA 2nd > 


Using Sherman and Gettler’s tables 91st we have calculated that the 
calcium supplied by the 5 g. of paper eaten per guinea-pig per diem would 


784 E. BOOCK AND J. TREVAN 


be about equal to that supplied by 6 cc. of cow’s milk per diem, and is about 
equal to half the calcium in 40 g. of cabbage (Miss Hume’s standard amount). 

Besides supplying this small amount of calcium, it is possible that the 
paper may act as “roughage” in the diet, the alimentary tract of the guinea- 
pig being only suited to an extremely bulky diet, and requiring a compara- 
tively large amount of ballast in order to secure normal intestinal movements. 
It is, however, very difficult to believe that a guinea-pig devouring 40 g. of 
mangold per diem is in need of any further cellulose, and we are inclined to 
the opinion that the paper acts chiefly by partially supplying the calcium 
deficiency. Mangold, though deficient in calcium, contains about the same 
percentage of potassium as cabbage. 


GMS. r 4 


t URINE ALKALINE. 


Tt F 
100GMS. 
eon 
160 Z “a 
AD 
10 DAYS. 

tT 
155 
196 * 


0.5 CC. CODLIVER OIL. 
Chart V, Guinea-pigs on basal diet as in I and III, plus salt mixture and paper. 0-5 cc. cod- 
liver oil per guinea-pig per diem administered at the point marked by the arrows. One 
| oermesee ig only then approached a normal rate of growth, the deficiency of protein probably 


ing the limiting factor in the case of the other three guinea-pigs, one of which ultimately 
succumbed, 


An important point in connection with the first part of Chart V arises in 
consideration of the changes in the acidity of the urine. 

On the mangold diet, the urine is acid. On a greenstuff diet, the urine is 
alkaline, 40g. of cabbage, according to Miss Hume’s results, is sufficient for 
normal growth in guinea-pigs. If, however, a comparison is made between 
the ash of mangold and cabbage, it will be found that whereas 40g. of ashed 
mangold give sufficient base to neutralise 35-2 ce. N/10 acid (this is obtained 
from a direct titration of the ash—is probably therefore too high because of 
the loss of sulphur, but the error is only of the order of 2 or 3 cc.), 40 g. of 
ashed cabbage will neutralise 42-6 ec. N/10 acid (Sherman and Gettler, caleu- 


GUINEA-PIGS AND VITAMIN A 785 


lated), not a significant difference. There are two explanations that offer 
themselves: 

(1) The potassium salts of the mangold are not absorbed by the intestinal 
mucous membrane in the absence of calcium salts to balance their toxic 
effects. It is a commonplace that potassium poisoning is very difficult or 
impossible to produce by oral administration of potassium salts. As a result, 
the acid-base balance is disturbed—the animal’s only source of salts is the 
bran and oats, the ash of which is acid. 

(2) The deficiency of calcium may lead to a disturbance of the metabolism, 
possibly secondarily to an infective process, which leads to excessive break- 
down of body protein, and the consequent production of abnormal amounts 
of sulphuric and phosphoric acids in excess of the neutralising value of the 
food salts absorbed. 

The experiments so far done are in favour of the first. explanation, for 
animals that died as a result of other deficiencies in diet, but with a good 
supply of calcium lactate, mostly had an alkaline urine, wliereas animals fed 
on the diet without mangold, and with no other source of vitamin C, although 
supplied with calcium lactate, developed an acid urine. 

This is only one of many indications of the need for further investigation 
of the physiological nature of the disturbances of the animal economy by a 
deficient diet, and is one which we are further examining. 

In Chart V, four guinea-pigs were started on an initial diet of mangolds, 
bran and oats, salt mixture and paper, and it will bé seen that the salt mixture 
and paper did not sufficiently supplement the diet to give continued normal 
growth curves, and after about a month, all the guinea-pigs started to decline 
in weight. The similarity of the post-mortem lesions to those found in rats 
fed on diets deficient in vitamin A suggested the addition of some source of 
vitamin A. This had been attempted before with carrot extract prepared by 
Zilva’s method, but without much success, for other deficiencies of the mangold 
were not made up. We now tried cod-liver oil instead of the carrot extract 
and 0-5 cc. per guinea-pig per diem was administered by hand. This was 
found:to cause a great improvement, and the guinea-pigs, with only one ex- 
ception, resumed a fair rate of growth. 

It follows, therefore, that the mangold, bran and oats diet, besides being 
deficient in calcium and perhaps roughage, is also lacking in the fat-soluble 
vitamin A. Our impression is that the dose of 0:5 cc. of the oil used in this 
experiment for our guinea-pigs was marginal, which may account for the 
failure of one of the four guinea-pigs to survive on the oil-supplemented diet. 
We have frequently noted that rats kept for a prolonged period on a diet 
deficient in vitamin A seem to undergo some permanent change, which 
renders them unresponsive to any treatment with cod-liver oil. It is possible 
that guinea-pigs also develop a similar condition of unresponsiveness, when 
deprived of the vitamin for prolonged periods of time. 

But further feeding experiments with another set of guinea-pigs on a diet 


786 E. BOOCK AND J. TREVAN 


of bran and oats, mangolds, salt mixture, paper and oil, revealed a further 
deficiency in the diet, viz. protein. 

Chart VI shows the weights of two guinea-pigs which were started with 
a diet of bran, oats, mangolds, paper and salt mixture, and then when a 
decline in weight, due to deficiency of fat-soluble A, occurred, oil was given, 
with a result that growth was again resumed. After about a month of this 
supplemented diet, however, a fresh drop in the weights of both guinea-pigs 
occurred, and one of them succumbed, but the other guinea-pig quickly 
picked up weight again on adding caseinogen to the diet. This guinea-pig 
continued a normal rate of growth, and was still continuing to grow at the 
time of writing, so that it is concluded that a fully supplemented diet has 
been attained. (See also the latter part of Chart IV.) The question arises 
whether the caseinogen acts beneficially by: 

(a) Supplementing the phosphorus intake. 

(b) Supplementing deficient protein, either a deficiency in certain amino- 
groupings, or a deficiency in total amount. 


GMS, 
170” % PNEUMONIA. 


125 *& CASEIN. 


1006Ms. 
150 % cc. cODLIVER Ort. 


16 20 DAYS. 


Chart VI. Guinea-pigs on basal diet as in I and II, plus salt mixture and paper. Two succumbed 
before administration of oil was commenced (at arrows). Of the two survivors, one eventually 
succumbed to protein deficiency, but the other quickly resumed a normal rate of growth on 
the addition of caseinogen to the diet (at star). 


Osborne and Mendel [1918] have shown with rats, that a shortage of 
phosphorus leads to a considerable slowing in growth followed by a fall in 
weight, but on adding caseinogen to the phosphorus-free diet, an improve- 
ment was made, and still greater improvement when inorganic phosphorus 
was added, The addition of edestin (a phosphorus-free protein), however, only 
led to complete cessation of growth and a decline. 


GUINEA-PIGS AND VITAMIN A 787 


The 12-5 g. of bran and oats eaten in a day contains 0-05 g. of phosphorus 
[Sherman and Gettler, 1912], whereas 40 g. of cabbage only adds 0-013 g. of 
phosphorus, an amount which could be easily covered by a small extra con- 
sumption of bran and oats, such as often takes place. We are led to conclude 
that, since cabbage can fully supplement a bran and oats diet, the addition 
of phosphorus is probably not the influencing factor, but that both caseinogen 
and cabbage supply a certain amino-acid or acids not present in bran and 
oats or in mangolds. This is a rather remarkable conclusion, considering the 
small amount of nitrogen in greenstuff. It is paralleled by the observations 
of Thomas that the N of potato is especially effectual in supplying the N re- 
quirements of the body. 


KERATOMALACIA. 


oo 


GMS 
158 


1oocms. T 
1 CC. CODLIVER 
OIL 


168 20 DAYS. 


Chart VII, Guinea-pigs on diet of bran, oats, salt mixture, caseinogen and mangolds. After 
definite symptoms of vitamin A deficiency had become apparent, cod-liver oil was ad- 
ministered, at the arrows. Duration of the keratomalacia marked by straight line. 


Two guinea-pigs, Chart VII, were started on a diet of bran and oats, 
caseinogen, salt mixture, paper and mangolds, and it will be seen that this 
diet was complete with one exception, viz. fat-soluble A. After a good initial 
growth for over a month, both guinea-pigs dropped in weight considerably, 
and one of them developed keratomalacia. There was a clouding of the cornea 
of one eye, which gradually became opaque. No haemorrhagic discharge was 
seen, and no conjunctivitis. In the Report of the Medical Research Com- 
mittee [1919], it is stated that in rats it is the swelling of the eyelids and 
conjunctivitis which appears first, and this, if untreated, leads to thickening 
and clouding of the cornea, and ultimate blindness. With guinea-pigs, how- 
ever, we have found that the corneal clouding always appears first, and several 
times we have received guinea-pigs from stock with completely opaque corneas, 
which have gone on to panophthalmia. We have never observed any haemor- 


788 E. BOOCK AND J. TREVAN 


thagic or purulent discharge preceding the keratomalacia in guinea-pigs. In 
rats, we have sometimes observed the uncomplicated corneal change first, 
but more frequently conjunctivitis precedes the corneal cloudiness. 

When the guinea-pigs had each lost about 50-100 g., and the keratomalacia 
had developed unmistakably in the one, 1 cc. of cod-liver oil was given to 
each guinea-pig every day, which resulted in an almost immediate increase 
in weight, and a resumption of a normal rate of growth. There was a com- 
plete disappearance of the eye trouble after about one week’s feeding with 
the oil. 

We have confirmed the absence of fat-soluble A from mangold by feeding 
experiments with rats (Chart VIII). 


, p 
GMS. MANGOLD é 
25 
ee 
25 
ae t 
*CONJUNCTIVITIS. 
25 
10 DAYS. 


Chart VIII, Rats on purified synthetic diet deficient in vitamin A. Conjunctivitis developed at 
point marked by cross, and here mangold was added to the diet. No improvement in the 
condition and weight of the rats resulted, and they all eventually died. 


The fact that our guinea-pigs developed no symptoms of scurvy whilst on 
the mangold diet supplemented as above, for a period extending over more 
than two months, indicates that the mangold supplies an adequate amount 
of the anti-scorbutic vitamin C. Vitamin B is supplied in sufficient quantity 
by the bran and oats mixture, and is also present in the mangold. 

Our experiments therefore lead to the conclusion that mangold is deficient 
in the following: (1) calcium, (2) fat-soluble A, (3) protein, (4) ? roughage; 
but contains a sufficiency of the anti-scorbutic vitamin to keep guinea-pigs 
free from scurvy. This confirms the rule laid down by McCollum for seeds, 
roots and tubers generally. 

The only essential dietary substance added by the mangold to bran and 
oats is vitamin ©, although there is of course some energy value in the various 
other constituents, and by feeding experiments on rats we have shown that 
there is a fair quantity of vitamin B present. 

Once again, as other observers have noted, vitamin A deficiency accom- 
panies a low calcium content, so that it can almost be laid down as a general 
rule that where there is vitamin A deficiency, calcium deficiency probably 


GUINEA-PIGS AND VITAMIN A | 789 


runs parallel with it; and conversely, where there is a good supply of the fat- 
soluble vitamin, there is often a high percentage of calcium, as the following 
figures taken from Sherman and Gettler’s tables show: 

Vitamin A Calcium content % 


I. Egg yolk + 0-143 
Cow’s milk... + 0-124 
Beans (dried) ... + 0-165 
Nuts se + 0-270 (almonds) 
II. Egg white - 0-011 
Fish (white) - 0-022 
Potatoes = 0-006 
Rice - 0-006-0-01 


NO CASEIN. 


ot 


NO CASEIN. 


100 GMS. 


195 20 DAYS. 


Chart IX. Guinea-pigs on complete diet of bran and oats, salt mixture, orange juice, cod-liver 
oil and caseinogen. Flattening of the growth curves when caseinogen was withdrawn from 
the diet for a period marked by the horizontal line. 

We have made various efforts to feed guinea-pigs on a basal artificial diet 
similar to that used in feeding experiments on rats, but made up according to 
calculations based on analyses of bran and oats, milk and grass, with the 
idea of supplementing this basal diet with the three vitamins A, B and C in 
turn. All our experiments, however, have been brought to an untimely con- 
clusion by the fact that the guinea-pigs refuse to eat any of the artificial diet. 

Starting with bran and oats as a basal diet, and supplementing this with 
heated caseinogen, salt mixture as above, and paper, one can, however, study 
the effects of the vitamins A and C in the guinea-pig in a more or less un- 
complicated way. 

Two guinea-pigs were placed on this “synthetic” diet, as shown in Chart IX, 
bran and oats, heated caseinogen, salt mixture and paper being given, together 


790 E. BOOCK AND J. TREVAN 


with 3 cc. lemon juice and 0-3 cc. cod-liver oil per guinea-pig per diem. After 
a time, however, 5 cc. orange juice was substituted for the lemon, as it was 
found that the guinea-pigs were averse to the sour taste of the lemon, whereas 
they drank the orange juice eagerly. It will be seen from the curves that a 
very satisfactory rate of growth can be maintained on this diet. A flattening 
of the growth curves occurred when caseinogen was withdrawn from the diet 
for a short period, thus emphasising the need of supplementing the protein 
of the bran and oats. It is, of course, of great importance that the temperature 
of the animal house in which the guinea-pigs are kept should be kept as 
constant as possible. | 


: , 50 GMS. 
CALCIUM LACTATE. 
10 DAYS 
218 
Tica L URINE ACID. 


LACTATE. 


Chart X. Showing effect of calcium deficiency. Guinea-pigs on bran and oats, mangold and 
cod-liver oil. (Paper and caseinogen were omitted because of the possibility of their con- 
taining small quantities of calcium.) A rapid decline set in, from which both animals 
quickly recovered on the addition of calcium lactate to the diet. One guinea-pig later 
succumbed to protein deficiency. 


We are unable to confirm Miss Hume’s statement [1921] that guinea-pigs. 
have an intolerance for unemulsified fat, for we have performed a large number 
of experiments in which we have fed guinea-pigs with cod-liver oil, and in 
every case the oil has had a beneficial effect provided other essential food 
factors are not missing. We would suggest that Miss Hume’s failure to obtain 
growth on oil was probably due to the basal diet of bran and oats and orange 
juice being inadequate, so that when the test substance, e.g. butter-fat or oil, 
was added to the diet, the beneficial effect of the addition of vitamin A was 
masked by the protein and inorganic deficiencies. For example, in Chart X, 
guinea-pigs were put on a diet of mangolds, bran and oats and cod-liver oil, 
and very soon a decline in weight occurred, which might have been ascribed 


GUINEA-PIGS AND VITAMIN A 791 


to an intolerance for oil on the part of the guinea-pig. That this was not the 
case was shown by supplying the deficient calcium of the dietary, when a 
quick recovery of the weight of the guinea-pigs was obtained, and they con- 
tinued growing satisfactorily for some time, the oil still being continued, until 
finally, the deficiency of protein became apparent. 

It is obvious that the need of the guinea-pig for an adequate supply of 
calcium is an urgent one, and the effects of its absence show themselves 
earlier than those of the vitamins. Vitamin A deficiency takes about 40-50 
days to become apparent, and scurvy takes about three weeks to develop. On 
a diet containing inadequate calcium, however, guinea-pigs lose weight in a 
few days (see Chart X). 

These experiments fully confirm Miss Hume’s conclusion that guinea-pigs 
require a large amount of vitamin A, and they have the advantage that the 
addition of vitamin A to the diet was made in the form of cod-liver oil— 
much less admixed with other essential substances than in the case of the 
greenstuff and milk studied by her. The effect of deprivation of vitamin A is 
much more regularly obtained in guinea-pigs than in our own stock of rats, 
several. of which grow for long periods on a deficient diet and even continue 
to grow whilst developing obvious eye changes. The larger dose of oil necessary 
for the guinea-pigs also suggests that they may be more suitable than rats for 
the estimation of vitamin A, and we are making some experiments with that 
in view. 

CONCLUSIONS. 


(1) The deficiencies of a diet of mangold, bran and oats and water which 
was the controlling factor in an epidemic amongst guinea-pigs have been 
investigated. 

(2) Vitamin A and calcium salts have been shown to be deficient, and the 
protein to be deficient in quantity or composition or both. 

(3) Keratomalacia in the guinea-pig as a result of vitamin A deficiency 
has been observed and cured by the administration of cod-liver oil. 

(4) The administration of cod-liver oil to guinea-pigs has been shown to 
be an adequate means of administering vitamin A, and to be well tolerated 
by guinea-pigs if the diet is otherwise satisfactory. 

(5) Direct confirmation of the necessity for vitamin A in the diet of 
guinea-pigs, inferred by Miss Hume, has been obtained. 

(6) Another instance is provided of a foodstuff in which a deficiency of 
vitamin A is accompanied by a deficiency of calcium. 


REFERENCES. 
Cahen and Hurtley (1916). Biochem. J. 10, 308. 
Glenny and Allen (1921). Lancet, ii, 1109. 
Hume (1921). Biochem. J. 15, 47. 
McCollum (1920). The Newer Knowledge of Nutrition, 36, 49, 64. 
Osborne and Mendel (1918). J. Biol. Chem. 34, 131. 
Report of the Medical Research Committee (1919). Special reports, 38, 17. 
Sherman and Gettler (1912). J. Biol. Chem. 11, 328. 


LXXXIII. THE CATALYTIC DESTRUCTION 
OF CARNOSINE IN VITRO. 


By WINIFRED MARY CLIFFORD. 


Physiology Department, Household and Social Science Department, 
King’s College for Women, Kensington. 


(Received October 20th, 1922.) 


In a previous paper [Clifford, 1922] an account was given of the destruction 
of carnosine in beef and rat muscle kept in cold storage. Since the temperature 
at which the destruction occurred was at or below zero, it was improbable 
that it was due either to enzyme or bacterial action. The time taken however 
was prolonged—6-10 months—and therefore a slow change due to an enzyme 
or to bacteria might have been the cause. In order to exclude these possi- 
bilities experiments were carried out at 100°. A similar though far more rapid 
disappearance of carnosine took place at this temperature pointing to a simpler 
catalyst than an enzyme, while bacterial life is also inconsistent with a tem- 
perature range of 0°—100°. 


EXPERIMENTAL RESULTS. 


(1) Destruction of carnosine in beef at 100°. 


From 3-5 g. of minced lean beef were weighed into test-tubes and 10 cc. 
water added. The tubes were then plugged with cotton wool and placed in 
a water-bath kept at 100°. When necessary more water was put in the tubes 
to prevent drying of the meat. 

At intervals of 1-4 days a tube was removed, the contents made up to 
95 ce. with distilled water and 5 cc. of 20 % metaphosphoric acid added to 
precipitate proteins. The carnosine present in the filtrate was estimated 
colorimetrically by the method described in a previous paper [Clifford, 1921, 1]. 

The numerical results of four experiments were as follows: 


Days at 100° Ss 15 29 29.4 
Start 0-96 %, 0-99 % 1-05 % 1-05 % 
I ites 0-80 — — 
2 0°70 ae 0-90 0-89 
4 i a 0-80 0°74 
5 0-60 0-66 — —_— 
7 O-b4 0-60 0-68 0-65 
10 rs ea 0-60 0-64 
12 0-58 ~~ at —_ 
\4 8 0-60 0-55 0:57 
16 0°30 i “ ~ 
17 oe 0-40 0-43 0-40 


21 _ 0:33 -- — 


CATALYTIC DESTRUCTION OF CARNOSINE 793 


These results, Fig. 1, indicate the presence of a catalyst which removes 
carnosine from muscle. This removal is considerably hastened! by a rise of 
temperature and is therefore probably a chemical and not a physical change. 
On comparing the results with those previously published [Clifford, 1922] it 
is found that for the percentage of carnosine to fall to the same level 9-10 
months is necessary at 0° as against 21 days with a temperature of 100°. 


1:0 
0:9 


0:8 


0'7 ons 
a 
4. . 


§ 0-6 . pee 5 Sh 
: oh oe 
8 0-5; 
z 029 
© 0-4}— 294 
" ald 

0:3}—- bs 

0-2; 

0-1t- 

Days 
| ecg gfe a) CSS SR I a Be lH OP NE 


eee 
Js Se Oy Ct agee, O10 11) 12:38 14 16> 16 17 18> 162021 
Fig. 1. 3-5 grams meat in 10 cc. water. 


(2) Catalyst not removed by boiling. 


Experiments were performed to find whether the catalyst were present in 
an extract made by boiling beef with water. Minced lean beef was put into 
a flask with water and boiled in a water-bath for 30-60 minutes. It was then 
filtered, and the residue again boiled with water for a similar period. The 
filtration was repeated and the two filtrates united. An estimation of the 
carnosine percentage of the extract was made, and portions of 10 cc. put in 
test-tubes plugged with cotton wool. The tubes were then placed in a water- 
bath at 100°. 

In five experiments lasting 13-26 days no change took place in the carnosine 
content of the extract. Similar results were obtained in two other 13. day 
experiments, one kept at 60°, and the other at 37° after sterilisation by boiling. 

1 The coefficient is 1-5 for 10° rise whilst for physical actions it is generally 1:03-1-05. 

Bioch. xv1 52 


794 W. M. CLIFFORD 


Since the carnosine content of the extract in these experiments remained 
unchanged, the catalyst cannot be removed from beef by extraction with 
water for 1-2 hours at 100°. 


(3) Catalyst remains in beef after boiling. 

A series of test-tubes was taken containing 10 cc. of beef extract prepared 
as above together with 3 g. of boiled minced beef which had been washed in 
cold running water for 15 minutes to remove traces of extractives. The result 
as expected was identical with that obtained with fresh beef in water. 


(4) Catalyst removed by prolonged washing of boiled meat. 


Further experiments were carried out to find the result of prolonged 
washing of the boiled meat. 

Five experiments with series of tubes containing 10 cc. of extract and 3 g. 
of boiled meat washed 3, 18, 24, 48, 48 hours respectively were kept at ~ 
with the following results: 


21 20 19 30 9 
3 hours’ 18 hours’ 24 hours’ 48 hours’ 48 hours’ 
washing washing washing washing washing 


Days % % % % % 
Start 1-0 1-0 1-05 0-99 0-98 
3 0-9 1-0 1-0 0-99 0-98 
6 0-9 0-9 1:0 0-99 0-98 
8 0-45 0-6 0-77 0-99 0-98 
10 0-45 0-6 0-70 0-99 0-98 
13 0-40 0-55 0-65 0-99 0-98 
23 0-35 0-35 0-50 0-99 0-98 
27 0-30 0-33 0-48 0-99 0-98 


From these results it appears that the catalyst present in boiled beef 
which destroys carnosine can be removed by prolonged washing in running 
water, probably by separation of an adsorption compound from the muscle 
protein. 

In one experiment the preliminary boiling of the meat was omitted and 
fresh unboiled minced beef was washed in running water for 24 hours. Three g. 
of this washed meat was then added to each of a series of tubes containing 
10 ce. extract at 100°. The result was a disappearance of carnosine similar to 
that given by boiled beef washed for three hours. 


Start 10% 10 days 040% 
Bdays 0-95 % 13 ,, 040% 
6, 080% 23, 025% 
8 , 0-45 % 27, «025 % 


These results, Fig. 2, indicate that the catalytic destroyer of carnosine is 
present in some physical combination with the protein of beef muscle and 
may be removed by prolonged washing in cold running water. This removal 
is facilitated by boiling the meat before washing, though mere boiling for 
1-2 hours does not extract the catalyst from the muscle. 


CATALYTIC DESTRUCTION OF CARNOSINE 795 


The curve of disappearance of the base is peculiar and it is difficult to 
account for the delay in the start of the reaction with washed meat and 
extract as compared with fresh meat and water. In all the curves there is 
a sharp fall between the 6th and 8th days followed by a stationary period 
and then a slower steady fall. . 

This three-step curve is indicated in experiments on fresh meat and water 
(Fig. 1) and is strongly marked in experiments with cod muscle and with liver. 
A paper previously published [Clifford, 1922] shows a similar curve from beef 
kept in cold storage. 


Lee 
pe is 


0-9 


0:8}-— 


0-6}-— 


0:5;— 


0-4}— 


Percentage of carnosine 


Days 
error Cie beh eee he 


| 
23 45 6 7 8 9 10 11 12 13 1415 16 17 18 19 20 21 22 23 24 25 26 27 


Fig. 2. 10 cc. meat extract +3 grams washed meat. 
30, 9. Meat boiled 1-2 hours then washed 48 hours. 


19. ” %”? ”? 24 ” 
20. ” ” ” 18 ”? 
21. Bos sa 


22. Meat unboiled but waahed 24 hour, 


(5) Presence of catalyst in muscles of white fish. 


Carnosine has been shown [Clifford, 1921, 2] to be absent from the muscles 
of white fishes. A series of experiments was performed using 3g. of fresh 
minced cod muscle in beef extract in place of 3 g. of boiled and washed beef. 
Unexpectedly it was found that the carnosine disappeared giving a similar 
curve to that with washed beef in its. peculiar three-stepped nature, though 
the actual position of the stationary period varies (Fig. 3). 

Three experiments were carried out at 60° after preliminary sterilisation 
by boiling and three were kept constantly at 100°. 


52—2 


796 W. M. CLIFFORD 

P H 7 E 14 24 
60° 60° 60° 100° 100° 100° 

Days %_ % % % % % 
Start 0:96 0:99 0-99 0-99 104-- 40 
1 aoa aw sae ~~ 0-80 aos 

2 0-90 0-90 as 0-90 An _ 

3 ‘bes a es es as Ll 

4 = we 0-95 oe Be: vi 

5 0-85 ne aes iis 0-80 ee 

6 oe ae we hen ae 0-75 

7 0-58 0-80 0-72 0-90 0-55 = 

8 athe sa irs sais fe 0-60 
ee eae: as eS sae oe exes 
10 =e 0-80 ah 0:88 i 0-46 
ll os ay 0-60 at = — 
13 0-58 os — oh at 0-46 
14 soe 0-66 0-55 0-50 0-50 — 
16 ea 0-66 ped 0-44 oe si 
17 0-52 pate iss sik es eae 
18 ee 0-44 0-55 ore 0-45 oad 
20 ee: es as a ws ah 
21 a 0-40 0-50 0-40 0-35 ~ 
23 oa 0-35 as Be i a 
25 vast 0-30 es a si 0-30 
28 — _ 0-40 ses at ik 


Percentage of carnosine 


O1}-— 


Days 


be i ea Us Wl es ss is SA A aS ws SR Eat Sle) SA ah, 29 


1 2°S 2.56 O67 8 9 10 11121314 15 16 17 18 19 20 21 22 23 24 25 26 27 28 


Fig. 3. 10 cc. meat extract +5 grams cod muscle. 
Unbroken lines 100° C, copia: 
Broken lines 60° C. constantly after sterilisation by boiling. 


(6) Presence of catalyst in liver. 

An experiment was performed using tubes with 10 cc. of beef extract and 
3g. of calf liver kept constantly at 100°. Carnosine estimations again gave 
a stepped curve of disappearance of the base. 

A second experiment in which 1 g. of freshly killed rat’s liver was used 
in place of the 3 g. of calf liver, and in which the temperature was 60° in place 


CATALYTIC DESTRUCTION OF CARNOSINE 


797 


of 100°, also showed a loss. As was expected from the lower temperature 
and smaller amount of tissue taken, the rate of disappearance was slower 


than in the first experiment (Fig. 4). 


16 Q 
Days of 100° and 3 g. 60° and 1 g. 
experiment liver. % liver. % 
Start 1-0 0-96 
1 0-80 — 
2 —- 0-90 
5 0-55 0-77 
7 0-55 0-77 
13 0-64 
14 0-50 — 
17 0-42 0-60 
21 0-25 — 
1:0 
0:9 


ie 
08} 4 ee 
oT) ee Pare 


; a 
0-6 |— AS cee 


0-4|— 


Percentage of carnosine 


0:3;-— 


0-1-- 
Days 
Ora J0R8 GSS ae I AIR DER a 


4/6 


ot 


Fig. 4. 10 cc. meat extract + liver. 
Q 1 g. rat liver kept at 60°C. 
16 3 g. calf liver kept at 100° C. 


eet a We Bat | 
12 3 4 65 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 


A third experiment lasting 18 days, in which 3 g. of frozen calf liver was 
used as a source of the catalyst, showed no loss of the base. These findings 
point to the presence of the carnosine-destroying catalyst in liver as well as 
in muscle, whilst prolonged keeping in cold storage destroys the catalyst itself. 


798 W. M. CLIFFORD 
(7) Absence of catalyst from kidney. 


Since skeletal muscle and liver had both been shown to cause the dis- 
appearance of carnosine from muscle extract, it was possible that the catalyst 
was present in all cells. Therefore an experiment was carried out using | g. 
of freshly killed rat’s kidney, in place of liver or muscle, in 10 cc. of beef 
extract. The tubes were kept at 60°. No change in carnosine content took 
place till the 21st day, when a very slight fall took place. This fall however 
lay well within the limits of experimental error. 

Similar results were obtained using 3 g. of sheep’s kidney, 3 g. of bullock’s 
kidney and 3 g. of frozen bullock’s kidney, all kept at 100° for at least 20 days. 

There is, therefore, no carnosine-destroying catalyst in the kidney, and 
consequently the possibility of existence of this catalyst in all cells cannot be 
maintained. 


(8) Absence of catalyst from invertebrate muscle. 


A series of test-tubes containing 10 cc. of beef extract together with 3 g. 
of fresh lobster muscle was left in a bath at 100°. No change in the carnosine 
content of the tubes was observed up to the end of 46 days, and therefore 
the catalyst was not present in the lobster muscle. 

Oysters were next used as a further type of the invertebrata. Here from 
the first day onwards no pink colour could be obtained on diazotising the 
protein-free filtrate. However a deep yellow colour resulted and if estimated 
as carnosine showed no loss of base up to the 23rd day. 

In all other experiments, e.g. with liver and muscle, or in simple keeping 
of meat in cold storage, disappearance of the base coincided with a less depth 
of colour, and not an alteration of hue. Therefore the lack of red colour on 
diazotising the oyster-beef extract was probably due to the presence of an 
inhibitory substance similar to that shown to exist in salmon muscle [Clifford, 
1921, 2]. 

Unfortunately there was not enough of the beef extract with oyster muscle 
to use the precipitation method of Dietrich [1914] to show whether this were 
the case or not, but experiments are in progress to determine this. point. 

The definitely negative result with lobster muscle and the probably nega- 
tive findings with oyster point to a lack of carnosine-destroying catalyst in 
the invertebrata. 


Discussion OF RESULTS. 


The experiments described in this paper indicate the presence of a sub- 
stance which is capable of destroying carnosine in muscle extracts, and owing 
to its temperature range it must be a relatively simple catalyst. 

It is not known whether the absence or lessening of red colour on diazo- 
tising filtrates from meat kept several days in water at 100° is due to actual 
destruction of the iminazole ring, or to a synthesis which prevents coupling 
of this ring with the diazo-reagent. Experiments are in progress with the 


CATALYTIC DESTRUCTION OF CARNOSINE _ 799 


object of elucidating this problem. In either case, whether the loss of carnosine 
is due to a synthesis or to a breaking up of the base, it will be of interest to 
know the fate of the B-alanine portion of the molecule, as this is the only 
known f-amino acid in the animal organism. 

The stepped nature of the curves of loss of carnosine is peculiar and totally 
unlike any enzyme curve. It may indicate that not one but two agents are 
active in the change, and that they act in an analogous way to enzymes and 
co-enzymes. On the other hand the catalyst may bea single substance and the 
reaction take place in two stages separated by an inactive period. Neither 
of these explanations can account for the initial delay shown when using 
washed meat as a source of catalyst, since after this delay the three-stepped 
curve appears. 

The catalyst is present in ox, rat, and cod muscle and therefore is probably 
found in vertebrate skeletal muscle generally, but experiments with lobster 
and oyster muscle indicate its absence from invertebrates. The livers of the 
rat and ox also contain this catalyst, which however is absent from frozen 
ox-liver. This indicates another connection between hepatic and muscular 
activity already seen in glycogen metabolism and also with urea, creatine and 
creatinine. 

The catalyst is not present in all cells since experiments have proved it 
to be absent from the kidney of the rat, ox, and sheep. The existence of this 
catalyst may account for the non-appearance of ingested carnosine in the 
urine. 

Earlier experiments in which carnosine was shown to be absent from the 
striped muscles of the white fishes and the finches appeared to suggest that 
carnosine is a merely accidental substance, somewhat arbitrary in its presence 
or absence. But the experiments detailed above, particularly those that show 
the presence of a carnosine-removing catalyst in the muscle of white fish, 
suggest that it is an intermediary product of metabolism and that its appear- 
ance and its percentage in muscle are determined by the rate at which it is 
formed and the rate at which it is removed in the different types of muscle. 


The expenses of this research were defrayed by a grant from the Medical 
Research Council. 


Thanks are due to Professor V. H. Mottram for the interest shown by him 
in the work and for his helpful criticism. 


REFERENCES. 


Clifford (1921, 1). Biochem. J. 15, 400. 
—— (1921, 2). Biochem. J. 15, 725. 
(1922). Biochem. J. 16, 341. 
Dietrich (1914). Zeitsch. physiol. Chem. 92, 212. 


LXXXIV. ON THE VITAMIN D. 


By TREVOR BRABY HEATON. 
From the Department of Pharmacology, Oxford University. 


(Received October 25th, 1922.) 


THE well-known fact, first described by Wildiers, that yeast cells when in 
low concentration fail either to ferment sugar or to grow, but may be induced 
to do so by the addition of “bios,” is as yet imperfectly understood. The 
need of yeast cells for “bios” depends clearly on dilution; for yeast, when in 
adequate initial concentration, is able to synthesise this substance indefinitely 
from the ingredients of a simple medium. It is an organic substance, soluble 
in water or alcohol, dialysable, and thermostable; it appears therefore to 
belong to the group of the vitamins. It seems, however, to be more thermo- 
stable than the vitamin which promotes the growth of rats [Souza and 
McCollum, 1920]; it is more thermostable also than the co-enzyme of zymase 
[Tholin, 1921]. The capacity for activating yeast in this manner has been 
used nevertheless by Williams [1919] and others [see Sherman, 1921], as a 
measure of the richness of any material in the vitamin B; and there is no doubt 
that in many cases, in yeast itself for instance, this vitamin and “bios” are 
closely associated. 

Now the comparative value of the various organs of animals, for pre- 
venting the onset of polyneuritis in pigeons fed on polished rice, has been 
determined by Cooper [1914], who places them in the following order: liver, 
heart, cerebrum, cerebellum, muscle; 0-5 g. of yeast being equivalent for this 
purpose to 0-9 g. of liver, 1-2 g. of brain, or 5-0 g. of muscle (dried weights in 
each case). In respect of promoting the growth of young rats, Osborne and 
Mendel [1918] place the tissues in a similar order. The distribution of “bios” 
among the organs of animals, therefore, seemed of interest as a test of the 
closeness of association between it and this vitamin; and a comparison in 
this respect between the organs of normal animals and those of animals suffering 
from vitamin-B deficiency, as an indication how far this deficiency can be 
attributed to a shortage of “bios.’’ Such an investigation forms the subject 
of the experiments to be described. 


Meruop. 


Two methods have been adopted for measuring “bios,”’ or the substance 
which activates yeast; the one depends upon the rate of multiplication of 
yeast cells, the other upon the rate of fermentation, as indicated by the 
evolution of CO,. The majority of workers have preferred to study growth, 


ON THE VITAMIN D 801 


as being a less erratic phenomenon than that of CO, production. The number 
and complexity of the factors which influence growth, however, are likely to 
be no less great than those which influence fermentation; and the fermentation 
method has been used in the present experiments, since it has the advantages 
of convenience and speed, and allows therefore many estimations to be made 
and an average drawn. 

The medium employed has been Niageli’s solution, omitting ammonium 
nitrate. Ammonium nitrate of course is necessary for the growth of yeast, 
and leads also to more rapid fermentation; but where the rate of fermentation 
by a given quantity of yeast is alone in question, a factor which only promotes 
growth is a source of error; and results without it, while less sensitive, seemed 
likely to be more accurate than in its presence. The composition of the 
medium used, then, was: 


Cane sugar id ws nae A Be 
Potassium biphosphate «oe OB 
Magnesium sulphate ... ... 025g. 
Calcium phosphate... .. 005g. 
Distilled water to bss .«« 100 cc. 


To this medium is added 0-01 % of fresh brewer’s yeast (1 cc. of a 1 % sus- 
pension to 100 cc.), a quantity in itself insufficient to produce any fermenta- 
tion whatever in 24 hours: this constitutes the standard yeast-Nageli medium. 
To 10 cc. in each of a series of tubes is added a graduated quantity of a watery 
extract of the substance under investigation. The mixtures are transferred 
with a serum-syringe to a series of 3 cc. glass ampoules, such as are used for 
vaccines, two ampoules from each tube. The experiments are performed in 
duplicate, so that four determinations are made for each concentration of 
the activating substance. The ampoules are then inverted over a receptacle, 
and incubated at 33° for 24 hours. As fermentation occurs, fluid is displaced, 
and drips out of the neck of the ampoule into the receptacle. Each ampoule 
is weighed empty (A), full (B), and at the end of the experiment (C). The 
difference, B—C, represents the cc. of CO, produced in B—A cc. of solution; 
from which the percentage CO, formation may be calculated. 

Results found to be most reliable are obtained when the fermentation 
amounts to about 5-10 %; the yeast tends to adhere to the glass as the fluid 
recedes, vitiating results when larger amounts of fermentation have occurred. 
And the progressive emptying of the ampoule constantly diminishes the 
quantity of fluid in which the recorded fermentation is taking place. Other 
influences,.moreover, which may affect the rate of fermentation, such as the 
presence of amino-acid or of phosphate, cause an increasingly greater per- 
centage error. The suggestion has indeed been made that such influences 
completely vitiate the test. At these concentrations of yeast, however, the 
error that they produce is not large. Table I shows the fermentation produced 
in the medium, in the absence of a specific activating substance, by various 
influences of this kind. 


802 T. B. HEATON 


Table I. Showing absence of fermentation in standard yeast-N dgeli medium, 
in absence of specific activating substance. 


Added to 10 cc. yeast-Nagelimedium  ... -05 +1 -15 = :2ee. 
Percentage fermentation 


Sodium phosphate 2 % see He ae 0 1 } 
Commercial glucose 2 % 5" ass sta | 1 1 1 
Dried blood 2 % extract NE Se ae fet? | 1 1 2 


Table II shows, further, that when an extract of dried yeast, which is an 
efficient activating substance, is added to the medium, the further addition 
of phosphate makes very little difference. 


Table IT. Showing absence of effect of sodium phosphate on fermentation 
produced in activated yeast-N dgeli medium. 
10 cc. yeast-Nageli medium, with 0-075 ce. 


of 2 % extract of dried yeast Fermentation — 
Without sodium phosphate... as isp 4:5 
With sodium phosphate ee a nibs 4-25 


A more searching test, however, as to whether the activation of minimal 
quantities of yeast is or is not the effect of a single substance, and whether 
the fermentation thereby produced is or is not a satisfactory measure of such 
substance, is afforded by a comparison of the fermentation curves obtained 
when the activating substance is derived from various sources. The following 
have been used for this purpose: cow’s milk, a 10% solution of “Glaxo,” 
a 2 % extract of the desiccated spleen of the calf (Armour), and a 2 % extract 
of dried yeast. Table III represents the average of a large number of deter- 
minations of the fermentation produced in the standard medium, after addi- 
tion of these four substances in increasing quantity. 


Table III. Average fermentation obtained in standard yeast-Ndgeli medium, 
after addition of increasing amounts of various activating substances. 


Amount added to 10 ce. 
yeast-Nigeli medium... -05 -075 “1 125 +15 = =+175 2 225 «+253 ee. 


Substance added Percentage fermentation in 24 hours 
Cow’smik .. .. 25 46 86 118 165 169° 934 — — — 
10 % Glaxo eos . — 4 — 7 — 85 — 15 24 
2 % dried yeast o 16 88 77 116 162 242° — — — . = 
2 % desiccated spleen... — 25 44 65 12 1456 181 215 — — 


It is clear from this table that while the activating power of cow’s milk 
is almost exactly the same as that of 2% yeast extract, it is considerably 
greater than that of either 10 % Glaxo or 2 % spleen. If, however, the acti- 
vating powers be calculated from the above figures, of 17 % Glaxo and of 
2:6 % spleen respectively, and the results plotted on a curve, it is found 
that such fermentation curves, for the four activating substances used, almost 
exactly coincide (Fig. 1), This seems to indicate that the activating substance 
is a single and definite entity, and that the method is a fair one for estimating 
its quantity in any material. 


a 


ON THE VITAMIN D 803 


In the following experiments the curve shown in Fig. | has been taken 
as a standard, indicating the fermentation produced by a 2 % extract of dried 
yeast. The organs of various animals have been air-dried at 33°, and powdered, 
and their activating power determined in 2 % watery extract. By comparison 
with Fig. 1 (or Table III) it is possible to say how much dried yeast is the 
equivalent in activating power of unit weight of the material investigated. 
This, for convenience, I have called the “‘ yeast-equivalent” of the substance. 
From Table III, for instance, it will be seen that the yeast-equivalent of this 
preparation of desiccated spleen is 0-8. 


< oO 
®& 
20|— 
e 
o 
8 
15|- a 
e 
Be 
10-— 
qd + 
(e) 
FB 
i a 
*: ° 
Os 
oO 
| | | | | [ | 


oe trees) be 26 1 PFS 20 
Fig. 1. Curve of fermentation in yeast-Nageli medium, induced by addition of various activating 
substances. 


Abscissae: Percentage addition to the medium of the various substances. Ordinates: 
Percentage fermentation in 24 hours at 33°. Crosses: 17% Glaxo!. ‘Squares: Cow’s milk. 
Hollow circles: 2 % dried yeast. Solid circles: 2-6 % desiccated spleen. 


1 Calculated from observations with 10 % Glaxo and 2 % spleen respectively. 


804 (ZL B. HEATON © 


RESULTS. 


(1) Pigeons. The organs were taken and dried under the same conditions 
from a series of birds, being (1) normal pigeons, and (2) pigeons which on an 
exclusive diet of polished rice had developed avian polyneuritis, and had 
either died from this condition, or had been killed at the point of death. It 
was thought that if the substance which activates yeast were the same as that 
whose absence causes polyneuritis, the organs of polyneuritic birds should 
show a deficiency of this substance; and this deficiency should be most marked 
in the brain, where the symptoms are so nearly exclusively localised. Funk 
[1912], as is well known, has shown a deficiency in the N- and P-content of 
the brains of birds in this condition. 

Because the symptoms of avian polyneuritis are markedly coreliolize | in 
type, the cerebral hemispheres and cerebellum were dried and estimated 
separately. In order to illustrate the degree of consistency of the results, the 
figures obtained in the case of the brain (cerebral hemispheres) are given in 
detail in Table IV. Table V is a summary of results from all organs, the 
average alone being given in each case, eg ige! with the calculated ‘ Tea 
equivalent.” 


Table IV. Details of experiments showing percentage fermentation obtained in 
yeast-N dgeli medium by addition f 2 % extracts of dried cerebral hemispheres 


of pigeons. 
Pigeon 1 2 3 + Av 1 2 3 4 Av. 

I. Polyneuritic 
P 25. 6 256 76 549 ll ll 11-5 9 10-5 
Q 45 35 4 45 41 9 7 8 4-5 71 
O 3 3 1-5 15 2:2 5 5 8-5 6 6-1 
4 8 4 35 65 52 8 9 10 10 9-2 
Z 4 8 4 5 5-2 9 ll 10 12 10-5 
B 3 2 15. 35 25 6 9 10 6 77 
».¢ 35 3 45 4 3:8 8 5 75 4 6-1 
Average 41 42 31 49 40 8 8-1 9-4 74 8-2 

II. Normal 

D 5 3 55 4 4-4 8 115 11 6 9-1 
5 3 5 5 4-5 9 12 10 5 9-0 
A 4 5 6 4 4:7 ll 10 9 ll 10-2 
Average 47 37 65 43 465 93 112 10 7:3 9-5 


Tables IV and V show, firstly, that the power to activate yeast is approxi- 
mately the same in all the organs of pigeons that were examined, and that this 
power is not very much less than that of dried yeast itself: the liver, which is 
slightly more active than other organs, being scarcely at all inferior to yeast. 
This offers a striking contrast with the figures of Cooper, already referred to, 
showing the relative efficiency of the organs in preventing the development of 
polyneuritis. 

So marked a contrast seems to show quite clearly, not only that the yeast- 
activating substance is not identical with the antineuritic vitamin, but that 
the two are not associated together in constant relative proportions. 


ON THE VITAMIN D 805 


Table V. Showing average fermentation produced in standard yeast-Ndgeli medium by 
addition of 2% extracts of dried organs of pigeons, and “ yeast-equivalents’’ calculated 
therefrom. 

Organ ... 2%cerebrum 2 % cerebellum 2 % liver 2 % kidney 2 % heart 2 % muscle 

Ce, added to'10 ce. 


yeast-Nigeli 
medium vow tL 15 ‘1-15 ‘1 +15 1 +15 ‘1 +15 1 +15 
Yeast Yeast Yeast Yeast Yeast Yeast 
Pigeon % Ferm. equiv. %Ferm. equiv. % Ferm. equiv. % Ferm. equiv. % Ferm. equiv. % Ferm. equiv. 
I. Polyneuritic 
Pg 5 105 8. 15° 4 4 95145 10 7 12 2 8-12 95 5 11 8 
Q 37 T §& 12° 85 “45 19 8 4 #8 15 5 9° 15 5511 85 
O Bae RR «cae 75 85 14 10 45 10 a7-4. 28 i anes, es | “‘T 
¥ Sxee 8 6565 11 Soa 14 ‘9 — “85 — — _ — 
Z 5 105 8 See eo. OD aren atl Cen bow was WB 
B 25 T5 6 — — 8 12 9 6-105 65 6 8° °8 55 95.8 
Xx es 6 35 8 ‘T 65 85 8 _ — — os a _ 
Average 4 82 7 4° 9 1 17512 9 §4 10 8.9 BT ote oS! OO 5, 8 
II. Normal z 
D 45 9 8 5 10 8 85 25 Ltt 8 ‘75 65 11 85 °4- 5 “6 
H 45 9 Soa 8 ‘15 —_ — — — — — — _ 
A 45 10 8 -- oe Oe The 8s 14. TO 26°98 oe S::-6 “‘T 
Average 45 93 8 45 9 § 67163 9 6 1 85 5:5 10 8 45 65 +7 


The tables show, moreover, secondly, that the activating power of the 
organs of polyneuritic pigeons is only very slightly less than that of the normal 
organs. This fact serves to strengthen the conclusion that deficiency of the 
substance which activates yeast is not the deficiency which causes poly- 
neuritis. 7 

There might indeed be a possibility that the organs of polyneuritic animals, 
while retaining their percentage of this activating substance, suffered a loss 
in their total content, by diminution in their whole weight. However, as 

- Abderhalden has shown [1921], such a diminution in the weight of individual 

- organs in polyneuritis occurs only in the case of the liver and voluntary 
muscles, not in that of the brain. This observation was confirmed in the fore- 
going experiments, as is seen in Table VI. 


Table VI. Showing weights of normal and of polyneuritic pigeons, 


and of their organs. 
Total Weights of organs 
bod c 7 . 
Pigeon weight Cerebrum Cerebellum Liver 
I. Polyneuritic 
150 0-98 0-22 4-52 
Q 165 1-05 0-20 ‘7-78 
O 145 1-13 . 0-22 3°15 
x 150 1-12 0-23 5°22 
Z 165 1-01 0:26 6-10 
B 224 1-11 0-27 8-67 
x 199 0-94 0-27 5:85 
Average 171 1-05 0-24 5-90 
Il. Normal 
D 270 1-08 0-28 6-39 
ae es 1-03 0-25 — 
A 323 1-15 0-24 6-86 
- : Average 296 1:09 0-26 6-68 


806 T. B. HEATON 


(2) Rats. It follows from the above experiments, that the yeast-activating — 
substance cannot be identified with the antineuritic vitamin, its amount 
being very nearly as great in the organs of polyneuritic as in those of normal 
pigeons. Similar experiments were performed on rats, the organs of normal 
animals being compared with those of others, which had been either killed 
when moribund, or had died, as the result of a diet free from water-soluble 
vitamin; and with those of a third series which, as the result of such a diet, 
had ceased to grow for several weeks, but had heen killed before any other 
symptoms had manifested theinselves. The diet consisted. the usual mixture 
of purined caseinogen, starch, sugar, butter, and salts. Upon this diet growth 
ceased at once; but if the initial weight of the animals was not less than about 
70 g., they remained active and in good condition for many weeks; eventually — 
dying with or without irregular nervous symptoms. The course of this con- 

_ dition, as has often been observed, is in striking contrast with the regular and 
rapid onset and constant symptoms of polyneuritis in birds. Table VII gives 
the summary of the average r+ $sofexperiments on the’ Yontent of the 
organs of these rats in the substance which activates yeast. 

It will be seen that in the case of normal rats, the “ yeast-equivalents”’ 
approximate very closely to those found in pigeons; but that in the rats which 
had died as the result of deprivation of water-soluble vitamin, this “ equi- 
valent”? has diminished considerably. The rats which were killed after a 
shorter period (six weeks) of vitamin starvation occupy in this respect an 
intermediate position. — 


Table VII. Showing average percentage fermentation produced in standard yeast-Ndge 
medium by addition of 2% extracts of dried organs of rats, with “ yeast-equivalents 
calculated therefrom. 


Organ Fas 2% brain 2 % liver 2% kidney 2% heart 
Uc. added to 10 cc. _ = —_ —_— 
yeast-Nigeli medium +1 -15 +2 : “1. +36 “2 *] “15 2 “1 +15 +2 
9 Yeast r, t 


% Yeast Yeas Y 
Rat number fermentation equiv. fermentation equiv. fermentation equiv. fermbttation ve 


I. Normal rats 


il 6513 — 085 5517 — O98 14 2856 — 13 5.10 16 0:8 
6 56 11 195 08 6 dH — 09 71. = 00" BS 8 45 0-8 
7 8 20 — 10 35 6 — 06 5516 — O89 56 7 16 0-78 
Average 6 1 — O9 5 12 — 08 9 205 — 10 5. 9 16 (OF 
Il. Vitamin-free diet six weeks. No symptoms ; 
8 45 6 13 OF"? 6: 10) 1S 0-8 6 9 14 08 35 45 6 0-6 
9 2 48550 05 4 9% 18 075 75 9 16 0-85 15 25 6 0-4! 
pee peewee Po ES 05 25 4 8 0-5 6 ll 14 08 8 ct ae 0-6 
Average 3 4310 O08 4 8 12 OF 65 9715 O8 27 47 7 8 
IIl, Vitamin-free diet till death or moribund 
I 1 — 85 06 15— Ii 0-6 25 — 75 05 1 — 25 O8 
2 2 — 1056 0565 1 — 5 O4 1 — 5 04 2 — TH Od 
3 1s — 10 O66 15— 75 046 4 — II 055 1 — 6 Od 
4 5 — 65 0465 38 — 75 05 — 3 8 05 — 1 6 OF 
ies - 6 9 08 — 7 9 O55 — B95 165 08 - 3 45 O-4 
Average 156 6 9 0565 15— 8 0886 25 6 95 06 +165 2 53 Od 


ON THE VITAMIN D 807 


There is therefore a considerable difference between the condition of these 
rats and that of polyneuritic pigeons. This is due, no doubt, to the long 
duration of the diet; for pigeons on a diet of polished rice are invariably dead 
from polyneuritis within 28 days. But it shows at any rate, that rats suffering 
' from deficiency of water-soluble vitamin are depleted of a substance which 
polyneuritic pigeons still possess; since death occurs before this apqon de- 

_.ency is manifested. 

These findings, afford.sypport to the conclusions of Funk and Dubin 1921], 
. vattiu vuOie are tyig water-soluble vitamins present in yeast: it is the absence 
of one of these, vitamin B, which causes polyneuritis in birds; while the other, 
which these authors call vitamin D, is the “bios” of Wildiers, necessary to 
the activity of low concentrations of yeast cells; both of them are necessary 
for the proper growth of rats. The prevailing uncertainty as to the identity 
of the antineuritic and the growth-promoting vitamins is due, therefore, to 
the fact that the latter includes the antinenritic substance, but contains also 
another essc~tial facté’, the vitam‘n ™ 

So far as yeast is concerned, the “ bios” effect indicates that the vitamin D 
alone is necessary for the activation of low concentrations of yeast cells, and 
leads not only to fermentation but also to growth. This must not be taken 
to imply, however, that fermentation and growth are identical phenomena; 
nor that vitamin D is the only requisite for the manifestation of either or both 
of them. The fact that vitamin B is present in yeast cells is a proof to the 
contrary; and other vitamins, it may be, are responsible for various functions 
in the life-history of the yeast-plant. The “bios” effect signifies merely that 
vitamin D is one of the substances necessary to the life of the cell, and that 
it differs from most other such substances in undergoing dilution by diffusion 
into the surrounding medium. 


SuMMARY. 


(1) The substance which activates minimal concentrations of yeast is a 
definite chemical entity, though obtainable from many sources. It may be 
measured fairly satisfactorily by the fermentation to which it gives rise. 

(2) It is present, in approximately the same amount, in the following 
organs, both of pigeons and rats: cerebrum, cerebellum, liver, kidney, heart, 
and voluntary muscle; its amount in any of these is about 80 °% of its amount 
in dried yeast. This distribution is contrasted with that of the antineuritic 
vitamin. : 

(3) In the organs of pigeons rendered polyneuritic by a diet of polished 
rice, it is present in the same amount as in those of normal pigeons. 

(4) In the organs of rats fed on a diet purified of water-soluble vitamin, 
its amount progressively diminishes, and when death occurs is only about 
one-half the normal. 

(5) For these reasons this activating substance cannot be identified with 


808 T. B. HEATON 


the antineuritic vitamin B. The deprivation of water-soluble vitamin from 
rats is held to involve a double deficiency; a deficiency of vitamin B in the 
first place, and of the yeast-activating substance in the second. This is the 
substance which Wildiers called “bios,’’ and Funk the vitamin D. 


The author wishes to express his cordial thanks and obligation to Dr J. A. 
Gunn, Professor of Pharmacology in the University of Oxford, for facilities 
to carry out the above work in his laboratory. 


REFERENCES, 


Abderhalden (1921). Pfliiger’s Archiv, 198, 355. 
Cooper (1914). J. Hygiene, 14, 12. 

Funk (1912). J. Physiol. 44, 50. 

Funk and Dubin (1921). Proc. Soc. Exp. Biol. 19, 15. 
Osborne and Mendel (1918). J. Biol. Chem. 34, 17. 
Sherman (1921). Physiol. Reviews, i, 598. 

Souza and McCollum (1920). J. Biol. Chem. 44, 113. 
Tholin (1921). Zeitsch. physiol. Chem. 115, 235. 
Williams (1919). J. Biol. Chem. 38, 465. 


LXXXV. ANEW PHOSPHORIC ESTER PRODUCED 
BY THE ACTION OF YEAST JUICE 
ON HEXOSES. 


By ROBERT ROBISON. 
From the Biochemical Department of the Lister Institute. 


(Received October 30th, 1922.) 


THE effect of sodium phosphate in increasing the fermentative power of yeast 
juice was first observed by Wroblewski [1901] and again by Buchner [Buchner, 
K. and H. and Hahn 1903] by whom it was attributed to the alkalinity of the 
salt. That such was not the true explanation followed from the work of 
Harden and Young [1905; 1906, 1 and 2; 1908, 1 and 2; 1909; 1910, 1 and 2; 
1911] whose extended investigations revealed the important rédle played by 
phosphates in alcoholic fermentation. It was shown by these authors that 
when a suitable amount of a soluble phosphate is added to a fermenting 
mixture of yeast juice and glucose, fructose or mannose, the rate of fermen- 
tation rapidly increases but after a short period again falls to a constant rate, 
which is only slightly greater than that of the original yeast juice and sugar. 
During this period of increased fermentative activity the total evolution of 
carbon dioxide is increased by an amount which is equivalent, molecule per 
molecule, to the phosphate added, while the latter undergoes a transformation 
into a form that is no longer precipitable by magnesium citrate mixture. 

From such solutions Young [1907; 1909] succeeded in isolating an ester 
of phosphoric acid by precipitation in the form of its insoluble lead salt. He 
showed [1909; 1911] that this compound possessed the constitution of a 
hexosediphosphoric acid, C,H,)0,(PO,H,),, and that the same substance was 
formed from glucose, fructose or mannose when fermented with yeast juice 
or zymin in the presence of phosphate. The acid was slightly dextrorotatory 
([a],, = + 3-2°) and on hydrolysis yielded phosphoric acid and a laevorotatory 
reducing substance from which fructose was isolated, although Young was not 
satisfied that the latter was the only sugar produced. A number of salts and 
derivatives were described, including the phenylhydrazine salt of a phenyl- 
hydrazone in which both phosphoric acid groups were retained, and an osazone 
in the formation of which one molecule of phosphoric acid had been split off. 
The evidence brought forward by Young was, however, insufficient to decide 
fully the constitution of the ester. 


Bioch. xv1 : 53 


810 R. ROBISON 


About the same time this compound was also discovered by Ivanov [1905; 
1907; 1909], who attributed to it the constitution of a triosemonophosphoric 
ester, and at a somewhat later date was examined by Lebedev [1909; 1910] 
who at first supposed it to be a hexosemonophosphoric ester but afterwards 
[1911] came to the same conclusion as Young on this point. 

During the course of some experiments carried out in 1913 by the author 
in conjunction with Professor A. Harden indications were noted that hexose- 
diphosphoric acid is not the only compound of this type produced during the 
fermentation of sugar by yeast juice in presence of phosphate. It was ob- 
served that during the preparation of hexosediphosphoric acid from fructose 
and after the precipitation of its lead salt by the addition of lead acetate, the 
filtrate always contained very appreciable amounts of phosphorus in organic 
combination, and that a further precipitate containing most of this phosphorus 
was thrown down on the addition of basic lead acetate. On examining the 
acid solutions obtained by decomposing such basic lead precipitates with 
hydrogen sulphide it was found that the ratio of the reducing power, as deter- 
mined by Bertrand’s method, to the phosphorus content was much higher 
than that given by solutions of hexosediphosphoric acid, while polarimetric 
observations gave evidence of the presence of a substance much more strongly 
dextrorotatory than the latter. 

The barium salt of this new acid was prepared, and proved to be readily 
soluble in both hot and cold water, differing in this respect markedly from 
barium hexosediphosphate. The salt was amorphous and was obviously con- 
taminated with substances derived either from the yeast juice or formed 
during the fermentation of the fructose, but the results of analyses appeared 
to indicate that the compound was the barium salt of a hexosemonophos- 
phoric acid, C,H,,0;P0,Ba. These facts were published in a preliminary note | 
[Harden and Robison, 1914] and the investigation of the compound was con- 
tinued. Its purification, however, proved unexpectedly difficult and was not 
completed when the work was interrupted by the war. It was resumed in 
1919 but was again held up by the difficulty experienced in obtaining yeast 
juice of reasonable fermentative power. The variability in the activity of 
juice prepared from the same type of yeast and apparently under similar 
conditions has been frequently observed, but no explanation could be found 
for the long run of inactive or feebly active juices obtained during 1919 unless 
it were the disheartening effect of the war beer on the yeast. That difficulty 
no longer exists and improved methods for the isolation and purification of 
the new compound have been worked out. Sufficient evidence has now been 
obtained to confirm the opinion already stated as to its constitution. It is, 
however, not identical with the hexosemonophosphorie acid prepared by 
Neuberg [1918], by partial hydrolysis from hexosediphosphoric acid. A speci- 
men of barium hexosemonophosphate prepared according to the method 
described by Neuberg was found to be optically almost inactive but to yield 
a strongly laevorotatory product on further hydrolysis. The compound iso- 


=<. 


HEXOSEMONOPHOSPHORIC ACID 811 


lated from the products of fermentation is, on the other hand, strongly dextro- 
rotatory, and on hydrolysis yields a dextrorotatory reducing substance from 
which glucosazone has been obtained. The ratio of the rotatory power of this 
hydrolysis product to its reducing power as estimated by Bertrand’s method 
is, however, much lower than that given by solutions of pure glucose. In this 
and in other respects the behaviour of the ester suggests that it may possibly 
be a mixture of isomeric hexosemonophosphoric acids, while the substance 
described by Neuberg may represent one or more other isomers. The iso- 
merism may be due to the hexose molecule itself or to the position of the 
phosphoric acid radicle. 

So far, only one salt of hexosemonophosphoric acid, that of brucine, has 
been obtained in a crystalline condition and the hope that this salt might be 
used to effect the purification of the acid, and its possible resolution into two 
or more isomers was for long disappointed owing to the difficulty with which 
crystallisation was effected. Within the last few weeks, however, more success 
has been obtained and it is hoped that the question may yet be solved in 
this way. 

With phenylhydrazine, hexosemonophosphoric acid yields the phenyl- 
hydrazine salt of an osazone in which the phosphoric acid radicle is retained. 
This compound is not identical with the osazone of the same empirical formula 
prepared by Young, Ivanov and Lebedev from hexosediphosphoric acid. 

Hexosemonophosphoric acid is only very slowly hydrolysed by boiling 
with mineral acids and a considerable portion of the sugar produced is de- 
stroyed during the operation. It is also hydrolysed by emulsin, yielding free 
phosphoric acid and a dextrorotatory reducing substance. Emulsin likewise 
hydrolyses hexosediphosphoric acid but, as with acid hydrolysis, the reducing 
substance produced is strongly laevorotatory. It is therefore improbable that 
hexosemonophosphoric acid is formed during the fermentation by partial 
hydrolysis—enzymic or otherwise—of the diphosphoric ester. 

The behaviour of hexosemonophosphate towards yeast, yeast juice and 
zymin is being studied in conjunction with Dr Harden and is yielding inter- 
esting results that are, however, difficult to interpret. The alkali salts are 
readily fermented by yeast juice and by zymin and the initial rate is relatively 


high, even approximating to the “phosphate rate” for glucose, but rapidly 


falls to a lower level, which nevertheless is decidedly higher than that for 
hexosediphosphate. 

The same compound is apparently formed during the fermentation of 
either fructose or glucose in presence of phosphate. It is true that some 
differences were observed in the specific rotations of the various preparations, 
but such differences were not greater between preparations made from glucose 
and those from fructose than between different preparations made from the 
same sugar. This may perhaps be taken as further evidence that all specimens 
were mixtures of isomeric compounds in somewhat varying proportions. 

Owing to the many processes employed in the different attempts to obtain 


53—2 


812 ~ R. ROBISON 


the ester in a pure condition, and the fact that even the best of these methods 
involves some loss of the substance, the record of yields obtained does not at 
present give very trustworthy evidence as to the most favourable conditions 
for the production of the monophosphoric ester or for determining the relative 
amounts of this and the diphosphoric ester that are formed during fermen- 
tation. 

Calculations based on the values of the ratio 


Reducing power (as glucose) 
Phosphorus 


as determined for the acid solutions derived from both normal and basic lead 
precipitates would appear to indicate the presence of a much higher proportion 
of the monophosphoric ester than has ever been actually isolated. It is just 
possible that in spite of the thorough washing to which they were subjected, 
these precipitates still held some adsorbed sugar which would affect the above 
ratio. The possibility of other reducing substances being present must also be 
considered. 

The best yields of the hexosemonophosphoric acid actually obtained from 
fructose and from glucose correspond with only 15 % and 7 % respectively 
of the sodium phosphate added during the fermentation. These figures are 
based on the weight of barium salt of at least 90 9% purity; the total amounts 
formed would certainly be higher and might well correspond with double 
these percentages. 

The question as to what rdle, if any, is played by hexosemonophosphoric ° 
acid in alcoholic fermentation must be left unanswered until more information 
is gained on the various points indicated above. 


EXPERIMENTAL. 


A typical series of values of the ratios 
Reducing power (as glucose) Rotation in 4 dm. tube 
Phosphorus and Reducing power (as glucose) in 100 ce. 

as determined for the acid solutions derived from normal and basic lead pre- 
cipitates, is given below. They are taken from one of the first experiments, 
which led to the discovery of the monophosphoric ester, and are quoted be- 
cause of their bearing on the relative amounts of the two esters formed during 
the fermentation. 

The reducing power was estimated by Bertrand’s method and calculated 
as glucose; the phosphorus was estimated by Neumann’s method. 

Fermentation: 1220 ce. juice from 4 kilos of fresh pressed brewery yeast 
was allowed to ferment 122 g. fructose, 131 g. (0-4 Mol.) of Na,HPO,12H,0 
being added gradually, as indicated by the rate of evolution of CO,, during 
1} hours. Fermentation was stopped and protein coagulated by blowing steam 
through the solution. After neutralisation of the filtered solution with caustic 
soda and removal of free phosphate by precipitation with magnesium acetate, 
a slight excess of lead acetate was added to the filtrate. The precipitate was 


HEXOSEMONOPHOSPHORIC ACID 813 


removed by centrifuging and washed nine times with water. The filtrate and 
washings were treated with basic lead acetate and the precipitate again 
separated by centrifuging and thoroughly washed. These precipitates were 
decomposed by hydrogen sulphide and the excess of the latter removed by 
means of a current of air. The acid solutions were each neutralised to phenol- 
phthalein by the addition of caustic soda and were again treated successively 
with normal and with basic lead acetate, any traces of free phosphate being 
first removed. These operations were repeated as shewn below (Table I). 

The acid solutions derived from the various lead precipitates gave the 
following ratios: 


Table I. 
Description of precipitate Reducing power Rotation in 4 dm. tube {a} , of acid 
No. of from ig cokuti ion was as glucose Keducing power as calculated on 
sol. P glucose in 100 ce. P-content 


1 First Sean gn normal lead 
acetate (crude lea ee 


phate) ha 2-50 _ on 
2 = Solution 1 eicinlinted a were ins 
by normal lead acetate ... a 2-25 — — 


3 Solution 2 precipitated a third time 
by ved mo acetate. Precipita- 
tions carried out in two fractions: 
lst fraction gue 1-50 —_— — 
4a Solution 3 precipitated a isa sad 
fifth time with normal lead acetate, Calculated for 
the fifth time in two fractions: CgHyo0,(PO,H,), 
Ist fraction Lae aed 1-21 0:647 +3-56° 
4b 2nd ae Ske its Ay 1-67 -- _ 


5 _—_ Solution 4 a precipitated a sixth time 
with normal lead acetate es (pare 
lead hexosediphosphate) .. nS 1-16 0-623 + 3-26°* 


6 _ First precipitate by basic lead acetate 

after removal of first normal lead 

precipitate .. eee 5-50 — —- 
7a Solution 6 precipitated a a re 

with normal then with basic lead 

acetate: 

; Normal lead pre as _ as — 

7b Basic om 4:24 — _ 
8a _ Solution 76 pescipitated a third aaa 

with normal then with basic lead Calculated for 

acetate: C.H,,0,(P0,H,) 

Normal lead precipitate an 3-96 1-60 +19-0° 

8b Basic __i,, rE 4:36 2-40 +31-2° 
9a Solution 8b precipitated a fourth 

time with normal then with basic 

lead acetate: 

Normal lead precipitate aie 4-13 2-18 +25-8°F 

9b Basic lead precipitate 

Amount too small for determina- 

tion of ratios eis as -- — 


* [a], of ee aiub takin acid = +3-2° (Young). 
+ [a] # of hexosemonophosphoric acid = + 25-0° (p. 817). 


Solutions 9a and 96 were each treated with baryta until neutral to 
phenolphthalein and the barium salts precipitated by the addition of an equal 


814 R. ROBISON 


volume of alcohol. They were purified by redissolving in water and repre- 
cipitating with alcohol-and were then dried at 100° to constant weight and 
analysed : | 


Barium salt from solution 9a. Ba = 34-33%; P = 7-07 %. 
* ee »  9b.- Ba = 34-86%; P ='7-22-%. 
Calculated for C,H,,0;PO,Ba. Ba = 34-75 %; P = 7-85 %. 
It seemed probable therefore that the chief constituent of the basic lead pre- 
cipitates was the salt of a hexosemonophosphoric acid, but that this was con- 
taminated with salts of other acids containing less phosphorus or none at all. 
Whether the differences in the ratios for solutions Nos. 6 to 9 are due entirely 
to phosphorus-free compounds, or whether they indicate the presence of yet 
another phosphoric ester, e.g. of a disaccharide, is impossible at present to 
say. 

In the case of the solutions derived from the normal lead precipitates a 
steady decrease in the ratio Reducing power/P occurs as the purification pro- 
ceeds. If the final ratio of 1-161 be accepted as correct for pure hexosediphos- 
phoric acid, then it is obvious that solution 1 must have contained a con- 
siderable proportion of some other reducing substance. The presence of sugar 
would seem to be precluded by the thorough washing of the precipitate. That 
the precipitate contained some hexosemonophosphate is certain since, al- 
though this ester gives in concentrated neutral solutions no precipitate with 
normal lead acetate, a copious precipitate, probably of basic lead salt, is 
formed in dilute solutions. . 

If solution No. 1 contained no reducing substance or phosphorus com- 
pound other than the two hexosephosphoric acids, then the ratio 2-50 would 
indicate that nearly 50 % of the phosphorus was present in the form of the 
mono-ester, in addition to the amount remaining in the filtrate and pre- 
cipitated by basic lead acetate. As already stated, the yields of hexosemono- 
phosphoric acid actually isolated in the form of its barium salt have never 
reached this proportion and have usually been much lower. 

The purification of the ester proved a problem of considerable difficulty and 
much time was spent in solving it. Attempts to prepare salts in a crystalline 
condition were unsuccessful except in the case of the brucine salt, and frac- 
tional precipitation of the metallic salts by alcohol effected only slight im- 
provement in the analytical figures. Fractional precipitation of the syrupy 
acid from its alcoholic solution by ether, was also of little use. Repeated 
extraction of the aqueous solution of the acid with ether or ethyl acetate did 
remove a small proportion of crystalline substance which proved to be succinic 
acid, but the phosphorus content of the barium salt was thereby only slightly 
raised. Further fractional precipitation by lead acetate, normal and basic, 
gave a purer product but involved too rapid diminution in the amount of 

* The values for this ratio calculated from Young’s data lie between. 2-1 and 2:3. They are, 


however, not comparable with those shown above as the reducing power was determined by a 
different method{ Pavy). 


HEXOSEMONOPHOSPHORIC ACID 815 


substance to be practicable. The greatest improvement was effected by the 
use of mercuric acetate which precipitates a considerable proportion of the 
impurities and by combination of this with other processes, a product giving 
satisfactory analytical figures was at length obtained. 


Preparation and purification of the hexosephosphoric esters. 


The methods employed in the preparation of the esters have been con- 
siderably modified as experience suggested. Heating the fermentation mixture 
to destroy the enzymes and coagulate the protein was discontinued on account 
of the risk of causing partial hydrolysis of the diphosphoric acid. Preliminary 
separation of the two esters by means of their lead salts was given up owing 
to the large proportion of the monophosphoric acid brought down in such 
dilute solutions by lead acetate. The method now used is as follows: 

Juice prepared from fresh pressed English mild ale yeast is mixed with 
10 % of its weight of fructose or glucose and warmed to 26° in a water-bath. 
Fermentation having commenced a 20 % aqueous solution of Na,HPO,12H,O 
| is added from time to time in such quantity as to produce the maximum rate 
: of evolution of carbon dioxide. The addition of phosphate solution is best 
regulated by means of a guide experiment carried out with 25 ce. juice in the 
apparatus described by Harden, Thompson and Young [1910] (v. “ Alcoholic 
Fermentation” by A. Harden, 1923 edition, p. 28). When the total volume 
of phosphate solution added exceeds half the original volume of juice a further 
quantity of sugar should be dissolved in the fermentation mixture. When the 
further addition of phosphate no longer causes any considerable rise in the 
rate of evolution of carbon dioxide, solid barium acetate in amount equal to 
the total weight of crystalline sodium phosphate that has been added, is 
dissolved in the reaction mixture which is then rendered just alkaline to 
phenolphthalein with baryta, and treated with an equal volume of alcohol. 
The barium salts of the two hexosephosphoric acids and of any excess of 
free phosphoric acid are thus precipitated together with the protein of the 
yeast juice. The precipitate is filtered off on a large Buchner funnel, and 
thoroughly washed with 70 % alcohol. It is then treated with boiling absolute ~ 
alcohol and allowed to remain in contact with the alcohol over night. In this 
way the protein present is denaturated and rendered insoluble. The crude 
barium salts are then dried in an evacuated desiccator over sulphuric acid or 
more quickly by exposure to a current of warm air. They are next thoroughly 
ground up with 10 parts of cold water which dissolves the barium hexose- 
monophosphate, but scarcely any of the hexosediphosphate. The residue is 
washed twice with small quantities of water (being preferably removed from 
the Buchner funnel and again ground up) and is then extracted with 200 parts 
of water which dissolves the barium hexosediphosphate, leaving behind the 
insoluble barium phosphate etc. After removal of any traces of mineral 
phosphate by precipitation with magnesium acetate, the hexosediphosphoric 
acid is precipitated from the filtered solution in the form of its lead salt by 


gL ORE FE rr 


ma,’ 


816 R. ROBISON 


the addition of lead acetate. The precipitate is filtered or centrifuged off and 
washed. It is then suspended in water, decomposed by a current of sulphur- 
etted hydrogen, the clear filtrate is freed from sulphuretted hydrogen by a 
current of air and finally neutralised to phenolphthalein with caustic soda. 
The precipitation of the lead salt is repeated several times [v. Young, 1909]. 
The first aqueous extract of the crude barium salts contains the barium 
hexosemonophosphate together with a little hexosediphosphate and other | 
barium salts of acids derived from the yeast juice or formed during the fer- 
mentation. It is treated with basic lead acetate and the insoluble basic lead 
salt is filtered and washed. It is then decomposed with sulphuretted hydrogen 
as described above and the acid solution neutralised to phenolphthalein with 
a hot saturated solution of baryta. The solution is filtered and the precipita- 
tion with basic lead acetate repeated, after which the barium salt is again 
formed and is precipitated from its solution by alcohol, filtered off, washed 
with absolute alcohol and dried in an evacuated desiccator over sulphuric 
acid. It is dissolved in 10 parts of water to which 10 % of alcohol is finally 
added. The filtered solution is treated with mercuric acetate so long as this 
causes any precipitate to form, and is allowed to stand for some hours before 
filtering. The basic lead salt is then reprecipitated from the clear filtrate and 
once more converted into the barium salt, which is finally purified by re- 
peatedly dissolving it in 10 % alcohol, filtering and reprecipitating with an 
equal volume of alcohol. In presence of a small proportion of water the salt 
readily forms a syrup, and it should therefore be washed several times with 
absolute alcohol and rapidly dried in an evacuated desiccator. If the results 
of analysis are not satisfactory the purification by mercuric acetate etc. is 
repeated. The mercury precipitate and a sparingly soluble barium salt, which 
is obtained in very small amounts during the final stage of the purification, 
both contain phosphorus and are being submitted to further investigation. 


Hexosemonophosphoric acid and its salts. 


Barium hexosemonophosphate, prepared by the above method, is a colour- 
less, amorphous substance very readily soluble in hot and cold water, but 
insoluble in organic solvents. Dried over sulphuric acid it still retains about 
4-5 % moisture which is given up at 100°, the salt at the same time taking 
on a brownish tint. The percentage of water corresponds approximately with 
one molecule (calculated for C,H,,0;P0,BaH,0, H,O = 4:36 %) but is not 
constant. The anhydrous salt takes up about the same amount of moisture 
with great avidity at ordinary temperatures even when kept in a closed vessel 
over calcium chloride. 

For analysis the salt was dried to constant weight in an evacuated vessel 
(2 mm, pressure) at 78° over phosphorus pentoxide, only slight discoloration 
taking place under these conditions. Table II gives the results of analyses 
of specimens prepared from glucose and fructose, together with the specific 


HEXOSEMONOPHOSPHORIC ACID 817 


rotation of the barium salt and of the acid prepared from it by the addition 
of the exact amount of sulphuric acid. 

The results obtained with a specimen of barium hexosemonophosphate 
prepared by hydrolysis of hexosediphosphoric acid (Neuberg’s method) are 
also shown in the same table for purposes of comparison. 


Table II. 
Sugar from which salt 

No. was prepared Ba P C H [a] of salt La} of acid 

1 Fructose 35:35 = 7-61 — — — — 

2 ® 34:89 7-86 — — +12:5°(c=3-9%) +24:9°(e=2-1 %) 
3 Glucose 35:12 7-60 — — _- — 

4 rv 35:05 7-66 — — +14-4°(c=14 %) — 

5 a 34:82 7-45 — — +12:7°(c=17 %) — 

6 ” 34:83 7-73 — — +15-6° (c=11:2 %) +28-8° (c=2-7 %) 
ae 782 1897 209) aoeeaa7%) 4260" (0=8% 


Calculated for C,H,,0;P0,Ba 34°75 7:85 18-20 2-81 — — 
8 Hexosemonophosphoric acid 


obtained by partial hydrolysis 

of hexosediphosphoric acid 

(Neuberg’s method) 33:26 758 — — +0:38°(c=5:2%) +1:49° (c=3-3 %) 
Calculated for CgH,,0;PO,BaH,O 33-22 7-51 —_ — _— _ 

Molecular weight determination by the cryoscopic method yielded the 
following results: 

1. 0-0849 barium salt in 13-17 g. water gave A = 0-062°, whence apparent 
Mol. weight = 198. 

2. 03037 barium salt in 13-17 g. water gave A = 0-192°, whence apparent 
Mol. weight = 228. 

_C,H,,0;P0,Ba requires Mol. weight = 395-5; if 100 °% dissociated, apparent 
Mol. weight = 197-8. 

C,H» 909(PO,Ba),H,O requires Mol. weight = 791; if 100 % dissociated, 
apparent Mol. weight = 263-7. 

The experimental results are therefore consistent with the first formula 
but not with the second. 

The brucine salt was prepared by dissolving the calculated amount of the 
base in an aqueous solution of hexosemonophosphoric acid prepared from the 
barium salt. On evaporation of the solution at a low temperature a crystalline 
residue was obtained. This was very soluble in water, moderately soluble in 
warm methyl alcohol but sparingly soluble in ethyl alcohol. Recrystallisation 
was not readily effected but on treating the concentrated aqueous solution 
with acetone and cooling to 0°, a semi-crystalline precipitate was obtained. 
After twice repeating this operation the salt was dried to constant weight at 
78° and 2 mm. pressure over phosphorus pentoxide, no discoloration taking 
place. 

Phosphorus estimated by Neumann’s method: found P = 2-73 %. 

Calculated for C,H,,0;P0,H,(C.3H.gN,0,)2; P = 2:96 %. 

Specific rotation of salt in water (conc. = 6-6 %); [a]; = — 23-4°, 
The investigation of the brucine salt is being continued. 


818 R. ROBISON 


The normal lead and copper salts are also soluble in concentrated solutions 
but on diluting these, precipitates, presumably of the basic salts, are formed. 
Hexosemonophosphoric acid itself was obtained as a colourless, very viscous 
liquid by evaporating its aqueous solution 7 vacuo. It is moderately soluble 
in alcohol but very sparingly soluble in ether or acetone. Whether prepared 
from fructose or glucose it reduces Fehling’s solution and gives a red coloration 
when heated with resorcinol and hydrochloric acid (Selivanov’s reaction), 
though the colour is not so intense as that given by the equivalent amount of 
fructose. With water it yields a strongly acid-solution which can be titrated 
sharply with alkalis (two equivalents) using phenolphthalein as indicator. All 
attempts to obtain the acid in crystalline condition have failed. 


Osazone of hexosemonophosphoric acid. 


Attempts to prepare a phenylhydrazone of hexosemonophosphoric acid 
were not successful but on warming the solution with excess of the base, a 
crystalline derivative was obtained. 10 g. barium hexosemonophosphate, pre- 
pared from fructose, were treated with the exact amount of sulphuric acid and — 
the barium sulphate removed by filtration. The clear solution was mixed with | 
15 g. phenylhydrazine dissolved in acetic acid, and was heated on the water- 
bath during 30 minutes. It was then cooled to 0° and the yellow precipitate 
filtered off, washed with chloroform, and dried. 

Weight of product = 8g. M. pt. 135°-137°. 
Recrystallisation was effected by dissolving the compound in boiling ethyl. 
alcohol, adding an equal volume of hot chloroform and cooling to 0°. Short, 
pale yellow needles were formed. M. pt. 139° with decomposition. Moderately 
soluble in methyl alcohol, less soluble in ethyl alcohol, still less in chloroform 
and insoluble in light petroleum. Very sparingly soluble in water but readily. 
soluble in dilute sodium hydroxide. 

Analysis of substance dried at 66° in vacuo over phosphorus pentoxide: 
N = 15-33 %; P = 5-60 %. 

Calculated for C,H,;NHNH,H,PO,.C,H;(OH),C(N,HC,H,)CHN,HC,H, ; 

N = 15-38 %; P = 5-68 %. 
The compound therefore appeared to be the phenylhydrazine salt of the 
osazone of hexosemonophosphoric acid. It is, however, not identical with the 
compound of the same empirical formula obtained by Young from hexose- 
diphosphorie acid, with elimination of one molecule of phosphoric acid, since 
the latter melted at 151°. 

The chloroform extract from the crude osazone yielded, on evaporation, a 
small quantity of a crystalline compound readily soluble in ethyl alcohol, 
which, by reerystallisation from ethyl alcohol and chloroform, was obtained 
in yellow plates, melting with decomposition at 190°. The substance con- 
tained phosphorus and was readily soluble in dilute sodium hydroxide, though 
very sparingly soluble in water. Unfortunately the quantity was insufficient 
for analysis. 


es eee 


Oia Cath ei a aS 


a 


HEXOSEMONOPHOSPHORIC ACID 819 


Hydrolysis of hexosemonophosphoric acid by acids. 


Hexosemonophosphoric acid is very resistant to hydrolysis by acids at 
ordinary temperatures. 

A solution of the acid in N sulphuric acid, after being kept for four months 
at room temperature contained only a trace of free phosphoric acid. When 
the solution of the acid is boiled either alone or with the addition of mineral 
acid, free phosphoric acid is slowly liberated and a reducing substance formed. 
Table III gives the data of an experiment in which 5-4 g. of hexosemono- 
phosphoric acid prepared from glucose, was heated with 250 cc. of N/5 sul- 
phuric acid at 97°, the rotation and free phosphoric acid being estimated from 
time to time. The total phosphorus = 0-643 g. 


Table IIT. 
Time of P present as free 
heating, Rotation in P present as free PO, in percentage 
hours 1 dm. tube PO,, g. of total P 
0 +0-625° 0 0 
1 +0-568° _- — 
3 +0°544° — a 
163 +0-507° 0-170 26-4 
36 +0-427° 0-274 42-6 
80 +0-327° 0-374 58-2 


After 80 hours the solution had become dark brown in colour and the 
rotation was difficult to measure. The hydrolysis was accordingly stopped 
and the solution neutralised with baryta. The precipitate, which consisted 
entirely of barium phosphate and sulphate, was filtered off and the clear 
filtrate treated with an equal volume of alcohol, which precipitated the barium 
salt of the unchanged hexosemonophosphoric acid. The phosphorus content 
and specific rotation of this salt agreed with those of the original barium 
hexosemonophosphate. The alcoholic filtrate reduced Fehling’s solution and 
contained only a trace of phosphorus. After distilling off the alcohol the re- 
ducing power of the solution was estimated by Bertrand’s method and its 
rotation measured. 

Reducing power calculated as glucose = 0-083 g. per 100 cc. equivalent to 
0-98 for total original solution. 

Rotation in 4 dm. tube = + 0-097°. 


Rotation in 4 dm. tube 
ace Sauer sa gincces ta 100 oo. 1-17. Calculated for glucose = 2-11. 


The amount of sugar formed during the hydrolysis, calculated from the 
amount of phosphoric acid liberated is equal to 2-17 g. or more than twice 
the quantity actually estimated in the solution. It would seem therefore that 
a considerable proportion of the sugar is destroyed during the boiling. This 
solution yielded an osazone which, after being twice recrystallised from dilute 
pyridine, melted at 205°-206° and did not lower the melting point of glucos- 
azone when intimately mixed with the latter. 


820 R. ROBISON 


In another experiment, Table IV, a solution of hexosemonophosphoric 
acid prepared from laevulose was boiled under a reflux condenser, the re- 
ducing power and the rotation being measured from time to time. 


Table IV. 
Time in Rotation in Reducing power as Rotation 
hours 4 dm. tube glucose in 100 cc. Reducing power 
0 +1-339° 0-59 2-27 
2 + 1-064° 0-64 1-66 
4 +1-024° 0-64 1-60 
10 +0-896° 0-64 1-40 
16 +0-788° 0-65 1-21 


At the end of 16 hours’ boiling the free phosphate amounted to 0-0795 g. P 
per 100 ce. or 53 % of the total phosphorus (0-1499 g. per 100 cc.). After pre- 
cipitating the unchanged ester by basic lead acetate the filtrate was freed from 


lead by sulphuretted hydrogen and the reducing power etc. determined. The 


: Rotation in 4 dm. tube 
value found for the ratio Reduchin power aa glisale 1100. Oe equal to 0-90. 


Hydrolysis of hecosemonophosphoric acid by emulsin. 


Hexosemonophosphoric acid is hydrolysed by emulsin yielding also in this 
case, free phosphoric acid and a dextrorotatory substance from which glucos- 
azone can be obtained. 

A solution of 1-876 g. anhydrous barium hexosemonophosphate in 24 ce. 
water was treated with 2 cc. of an extract of emulsin! prepared from sweet 
almonds, and was incubated at 37° in presence of toluene. At intervals the 
rotation was measured, after filtering off the precipitated barium phosphate. 
Table V gives the results of these determinations. On the 10th day a further 
1 cc. emulsin extract was added but the rotation has been calculated for the 
original volume of the solution. 


Table V. 
Time of incubation, Rotation in 2dm. tube Percentage of salt hydrolysed 
days at 20° as measured by free PO, 
0 + 1-680° — 
1 +1-579° —_ 
4 +1-376° — 
10 + 1:334° oo 
13 + 1-339° 66 


On the 13th day the remaining solution was neutralised to phenolphthalein 
with baryta and the total free phosphate estimated. The clear filtrate was 
treated with 14 volumes of alcohol and the precipitate filtered off, dried at 
78° in vacuo, and examined, 

Weight — 0-370 g. equivalent to 0-53 g. for the total original solution. 

Found Ba = 32:8 %; P = 6-87 %; [a]? = + 113%. 


* This extract contained a slight trace of reducing substance. 


HEXOSEMONOPHOSPHORIC ACID 821 


The substance probably consisted chiefly of unchanged barium hexose- 
monophosphate. The filtrate and washings from this precipitate were made 
up to 100cc. They contained a reducing substance equivalent to 0-240 g. 
glucose as estimated by Bertrand’s method. This amount is equivalent to 
0-347 g. for the original solution. 

Rotation in 4 dm. tube = + 0-350°. 


Rotation in 4 dm. tube aa 
Reducing power as glucose in 100 cc. f 


On evaporation this solution yielded a syrup which could not be made to 
erystallise but with phenylhydrazine yielded an osazone. After recrystallisa- 
tion from dilute pyridine this melted at 204°-205°, and did not lower the 
melting point of glucosazone when intimately mixed with the latter.. 


Hydrolysis of barium hexosediphosphate by emulsin. 

The hydrolytic action of emulsin on hexosediphosphate was established by 
Harding [1912] who estimated the phosphoric acid set free but did not further 
examine the products. 

In view of the possibility that the optically active substances produced 
under these conditions might differ from those produced by acid hydrolysis 
it was thought advisable to investigate this point. 

1g. of barium hexosediphosphate was incubated with 1 cc. emulsin extract 
and 10 cc. water at 37°, in presence of toluene. The salt only dissolved to a 
slight extent in the water. After 14 days the solution was filtered from the 
solid matter, which consisted of unchanged barium hexosediphosphate and 
barium phosphate, and examined polarimetrically. 
a8 in 1 dm. tube = — 0-530°. 

Estimation of free and total phosphate in both precipitate and solution 
indicated that about 60% of the phosphoric acid had been liberated. On 
neutralising the solution with baryta and adding alcohol a small amount of 
soluble barium salt was precipitated while the solution contained the laevo- 
rotatory substance. 

It appears therefore that the hydrolysis of hexosediphosphoric acid by 
enzymes results in products similar to those obtained by acid hydrolysis. 
Further experiments are, however, being carried out. 


The hexosemonophosphoric acid produced by partial hydrolysis of 
hexosediphosphoric acid. 

A quantity of this ester was prepared by Neuberg’s [1918] method in 
order to compare it with the compound obtained by the fermentation of sugar. 

9-3 g. of crude barium salt was thus obtained from 19 g. barium hexose- 
diphosphate by boiling the latter with a solution of 9-5 g. oxalic acid in 
150 cc. water. After purification by repeatedly dissolving it in 10 % alcohol 
and precipitating the clear solution with alcohol, the product (5-3 g.) was 
dried at 78° in vacuo over phosphorus pentoxide and analysed. 


822 R. ROBISON 


Found Ba = 33-26%; P = 7-58 %; [a]}? in 5-2% aqueous solution 

= + 0-38°. 

Calculated for C,H,,0;PO,BaH,0O; Ba = 33-22 %; P = 7-51 %. 
According to Neuberg the barium salt holds one molecule of water which is 
not given up at 105° and the above analyses.are in agreement with this view. 

The specific rotation of the acid, prepared from the solution of the barium 
salt by adding the calculated amount of oxalic acid was also determined: | 


[a]*? in 3-3 % aqueous solution = + 1-49°. 
With such small polarimetric readings, however, the possible error in the 
above values is relatively large. The solution of the acid was boiled with such 


excess of oxalic acid as brought the total acidity equal to NV. 
The solution rapidly became laevorotatory, as is shown in Table VI. 


Table VI. 
Time of boiling — Rotation in Percentage of salt hydrolysed as 
(hours) 2 dm. tube measured by free phosphate 
0 +0-094 ‘ 0 
4 /  —0-176 = 
1} —0-481 — 
43 — 0-822 68 


The behaviour of fermentation hexosemonophosphoric acid on hydrolysis 
is thus very different from that of hexosediphosphoric acid, and of the mono- 
phosphoric acid obtained from it by partial hydrolysis, both of which yield 
laevorotatory products. Table VII gives the results of experiments with each 
of the above compounds for comparison, the figures for the diphosphoric 
ester (Nos. 7~9) being those of Young [1909]. 


Table VII. 
Rotation in 4 dm. tube 
Percentage ee, 
of phosphate Before After 
Substance Method of hydrolysis liberated hydrolysis hydrolysis 
1. Hexosemonophosphoric Solution of acid in N/5 H,SO, at 58-2 +2:50° — +1-31° 
acid from glucose 97° 80 hours 
2a. DO. Do. Solution of acid boiled 27 hours 73 +0°374°  +0-234° 
3. Do. from fructose ‘ re 1B 33 53 +1:339°  +0-788° 
4. Do. Do. ” ” 10) ,, 42 +0°816°  +0-532° 
5. Barium hexosemono- Emulsin; 13 days at 37° 66 +1-680°* +1-339°* 
phosphate from glucose 
6. Hexosemonophosphoric Solution of acid in N oxalic acid 68 +0-094° -0-822° 
acid from heximathads: boiled 44 hours 


yhoric acid by partial 
nydrolysis (Neuberg’s 
method) 
7. Hexosediphosphoric Solution of acid boiled 27 hours ot - —_ 
acid from glucose 


(Young) 

8. Do. Do, i * My _- +0°161° | —0-658° 

9. Do. from fructose re Ps js ee oe +0-416° —1-514° 
(Young) 

10. Barium hexosediphos- Emulsin; 14 days at 37° 60 (+0:30f calc.) —0°530t 
phate (salt not in 


solution) 
* 2 dm. tube. + 1 dm. tube, 


HEXOSEMONOPHOSPHORIC ACID 823 


It would seem improbable therefore that the hexosemonophosphoric acid 
found in the products of fermentation is formed from hexosediphosphoric acid 
by hydrolysis, enzymic or otherwise. It may conceivably be an intermediate 
stage in the formation of the diphosphoric ester and indeed its behaviour with 
yeast juice and zymin lends some weight to such a view. On the other hand 
it may be formed by the synthetic power of the yeast enzymes but’ have no 
part in the main cycle of reactions which constitutes alcoholic fermentation. 

‘The analytical data so far obtained are all in agreement with the formula 
assigned to the compound, that of a hexosemonophosphoric ester, but it is 
admitted that further criteria of purity and homogeneity are desirable before 
this formula can be considered as established. The formation of an osazone 
retaining the phosphoric acid radicle indicates that the latter is not attached 
either to the terminal aldehydic carbon atom or that adjacent to it, but its 
position in the molecule cannot be definitely stated at present. The nature 
of the hexose is also undecided. The reducing substance produced on hydrolysis 
is dextrorotatory but less so than glucose. The ester itself whether produced 
from glucose or fructose gives Selivanov’s reaction for fructose but. not so 
strongly as an equivalent quantity of the latter sugar. These facts and the 
formation of glucosazone from the products of hydrolysis may perhaps indicate 
that the compound is a mixture of the monophosphoric esters of glucose and 
fructose. More cannot be said until further work has been done. 


SUMMARY. 


1. When fructose or glucose is fermented by yeast juice in presence of 
suitable amounts of a soluble phosphate, an ester having the composition of 
a hexosemonophosphoric acid is formed in addition to the hexosediphosphoric 
acid described by Young, Ivanov and Lebedev. 

2. A method is described for the separation of this new phosphoric ester 
from the other products of the fermentation and for the purification of its 
barium salt. 


3. Hexosemonophosphoric acid is strongly dextrorotatory ; [a] = + 25-0° 


‘in aqueous solution. 


The metallic salts with the exception of the basic salts of the heavy metals, 
are all readily soluble in water and are amorphous. A crystalline brucine salt 
has been obtained. 

4. The phenylhydrazine salt of the osazone of hexosemonophosphoric 
acid has been obtained. It is: not identical with the compound of the same 
empirical formula prepared by Young and Lebedev from hexosediphosphoric 
acid. 

5. On hydrolysis by acids or by emulsin, hexosemonophosphoric acid 
yields free phosphoric acid and a dextrorotatory reducing substance from 
which glucosazone has been obtained. The rotatory power of this product is, 
however, less than that of pure glucose. 


824 R. ROBISON 


6. By its specific rotation and its behaviour on hydrolysis the compound 
is sharply distinguished from the hexosemonophosphoric acid prepared by 
Neuberg by partial hydrolysis of hexosediphosphoric acid. ; 

7. The alkali salts of hexosemonophosphoric acid are readily fermented 
by yeast juice and zymin. 

8. It is improbable that hexosemonophosphoric acid is produced by the 
hydrolysis of the diphosphoric ester during fermentation. It may perhaps be 
an intermediate stage in the formation of the latter compound. 


In conclusion I wish to thank Professor A. Harden for the interest which 
he has taken in this investigation. - 


REFERENCES. 


Buchner, E. and H. and Hahn (1903). Die Zymasegirung, Miinchen. 
Harden and Robison (1914). Proc. Chem. Soc. 30, 16. 
Harden, Thompson and Young (1910). Biochem. J. 5, 230. 
Harden and Young (1905). Proc. Chem. Soc. 21, 189. 
(1906, 1). Proc. Roy. Soc. B. 77, 405. 
—— (1906, 2). Proc. Roy. Soc. B. 78, 369. 
—— (1908, 1). Proc. Roy. Soc. B. 80, 299. 
—— (1908, 2). Proc. Chem. Soc. 24, 115. 
—— (1909). Proc. Roy. Soc. B. 81, 336. 
—— (1910, 1). Centr. Bakt. Par. Abt. 1, 26, 178. 
—— (1910, 2). Proc. Roy. Soc. B. 82, 321. 
—— (1911). Biochem. Zeitsch. 32, 173. 
Harding (1912). Proc. Roy. Soc. B. 85, 418. 
Ivanov (1905). S. Travaux de la Soc. des Naturalistes de St Petersburg, 34. 
—— (1907). Zeitsch. physiol. Chem. 50, 281. 
—— (1909). Centr. Bakt. Par. Abt. 11, 24,429. 
Lebedev (1909). Biochem. Zeitsch. 20, 114. 
—— (1910). Biochem. Zeitsch. 28, 213. 
—— (1911). Biochem. Zeitsch. 36, 248. 
Neuberg (1918). Biochem. Zeitsch. 88, 432. 
Wroblewski (1901). J. pr. Chem. (2), 64, 1. * 
Young (1907). Proc. Chem. Soc. 23, 65. 
(1909). Proc. Roy. Soc. B. 81, 528. 
—— (1911). Biochem. Zeitsch. 32, 178. 


LXXXVI. MAMMARY SECRETION. IV. 


THE RELATION OF PROTEIN TO OTHER 
DIETARY CONSTITUENTS. 


By GLADYS ANNIE HARTWELL. 


From the Physiological Laboratory, Household and Social Science Department, 
King’s College for Women, Kensington, London. 


(Received October 24th, 1922.) 


I. Iyrropuctory AND HIsToRICAL. 


Iv is generally held that the diet of a lactating animal should contain a 
large amount of protein, and it seems quite obvious that a maximal growth 
of the young cannot be obtained if the mother’s protein intake is low. On 
the other hand it has been demonstrated [ Hartwell, 1921, 2] that large amounts 
of protein fed to the mother, result in harmful effects to the young. It has, 
however, been found [Hartwell, 1922] that large quantities of whole milk, 
and large amounts of yeast extract added to the mother’s excess protein diet, 
prevent the bad symptoms in the litters. The obvious, though possibly not 
the correct explanation of such results, was that the important factor was 
vitamin B, since this is present in both milk and yeast. Also it was shown 
[Hartwell, 1922] that lactose, butter, milk-ash and calcium lactate were in- 
effective in improving the bad condition. Consequently the milk effect was 
not due to any obvious dietary factor. 

The experiments described in this paper were started in the hope of 
throwing some light on the relation of vitamin B to protein in the metabolism 
of the lactating animal. 

That there is a possible relation between vitamin B and protein is to be 
seen from the work of Karr [1920]. He found that the dog refused food 
unless vitamin B was included in the diet, and since this animal is a carnivore, 
it may be that the vitamin is in some way essential for protein metabolism. 

In the experiments to be detailed here the mother rats were fed on an 
excess protein diet, to which were added various fruit and vegetable juices 
and extracts made from other foods said to contain vitamin B. From the 
results described, it appears that the addition to the mother’s protein-rich 
diet of juices of, or extracts from, foods reported to contain vitamin B, is 
effective in rendering the nature of the milk more normal. 

Bioch. xvz 54 


826 ; G. A. HARTWELL 


At the present time, the published results as to presence or absence of — 
vitamins in foods show discrepancies. This is probably due to 

(1) The different methods used for testing the presence of the vitamins. 

(2) The lack of quantitative estimations. 

It appears to be generally accepted that vitamin B is widely distributed in 
the plant kingdom, although there are differences of opinion as to the amount 
present. 

Osborne and Mendel [1920, 2] state that the edible part of an orange 
contains vitamin B and that the potency is similar to that of comparable 
volumes of cow’s milk. Grape juice they consider to be less potent than 
orange; apples contain the vitamin, but are not rich in it. These observers 
[1920, 1] also find that tomatoes and carrots contain appreciable amounts of 
the water-soluble vitamin; potato is quite a good source, but not so good as 
tomato. Beetroot contains less of the substance. The occurrence of vitamin B 
in animal tissues is possibly less extensive, but it may be present in smaller 
quantities. Cooper [1913] prepared a curative solution of the antineuritic 
substance from horse flesh and from beef. 

According to Drummond [1919] animal tissues (with one or two exceptions) 
are deficient in the water-soluble antineuritic factor. This observer fed young 
rats on protein-free extracts of muscle from herring, cod and salmon and found 
that his solutions contained no vitamin B. When the extract was made from 
the whole herring, he obtained a slight growth and concludes that the source 
of the vitamin in this case was the ova and glands, shown to contain the 
antineuritic vitamin by Chick and Hume [1917]. Drummond does not explain 
how he prepared the protein-free solutions and it is possible that any vitamin 
present was adsorbed when the protein was precipitated. If the factor in- 
vestigated in the experiments below be vitamin B the results obtained con- 
tradict those of Drummond. 


II. Meruops. 


The technique employed has been fully discussed in a previous paper 
{ Hartwell, 1921, 1]. At least three mother rats were fed on the various diets 
and each attempted to rear six baby rats. One typical curve is shown as 
representative of the three. In some experiments where two or three members 
of one litter survived and another litter all died (e.g. cucumber juice, Fig. 1) 
two growth curves are given as showing the best and worst growth on the 
given diet. 

The basal diet was: 

(i) White bread, 

(ii) Butter for fat and vitamin A, 

(iii) Lemon juice for vitamin CO, 

(iv) Caseinogen / 

(v) Salt es nehe | [Hartwell 1922], 
to which was added one of the following as potential source of vitamin B; 


A SS 


“7 ay 


MAMMARY SECRETION . 827 


(a) Fruit or vegetable juices. These were made as follows: the soft fruits 
were placed in muslin and the juice squeezed out by hand. The harder 
vegetables (e.g. potato, carrot) were grated and then squashed in a hand 
press. The juice was filtered through muslin and allowed to stand in a cylinder 
so that the starch (if present) settled at the bottom of the vessel and the juice 
was decanted. All juice was freshly prepared, just before being fed to the 
animals. 


240 
526 Apple 
230 ¥ ‘ x PP 
220}— , ge 
x 
210} A pod Apple 
200}~ 4 * 
; *, Jpgee Orange 
180} galery Ao e Lee 
170 7 Fel. Beetroot 
160 7362, Carrot 
150 
~ 373. P. 
J 

130 Va F otato 
120 
110 
100 

90 

a 
80 Va pe 495 Cucumber 
; fe” eT 

70 ‘a 

60 

50 

#0 WEIGHTS OF LITTERS 

30 = 

20 

10 Days 


72 384 56 7 8 9 10111213 141516171819 20 21 


Fig. 1. 
Mothers’ diets 


bal x x 15g. bread, 5g. caseinogen, 0-5 g. butter, 0-7 cc. lemon juice, 50 cc. apple juice 
473 A—A = g - * 50 cc. tomato juice 
422 oO —-——O ” ” ” ” 50 ce. orange na 
437 e--e ie oe ee - 50 ce. artichoke ,, 
421 e ” % " ae 50 ce. beetroot ,, 
6s." O——O * nt Pe = 50 cc. carrot je 
373 a Sr 2 yaae E4 ” ” ” % 50 ce. potato » 
608 ) ——% » ” ” ‘s 50 cc. marrow ,, 
} 2 Vos cae? ” ” i ei 50 cc. cucumber ,, 
597 @ ~aeer, Te we ” ; ” ” ” 50 ce, grape sd 


54—2 


828 G. A. HARTWELL 


(6) Meat or fish extracts. The meat was freed from fat as far as possible 
and minced finely. The fish muscle was removed from the bones (in the 
herring the backbone was taken out, but probably a few small bones were 
left) and minced. No roe was included. 

The minced meat and fish were weighed, placed in a bottle, warm distilled 
water added (1 cc. water to 1 g. flesh) and the whole shaken on a mechanical 
shaker for 14 hours. The contents of the bottle were then transferred to a 
flask, which was placed in a boiling water-bath for some of the protein in the 
solution to coagulate. The whole was filtered through paper and the filtrate 
fed to the animals. 

(c) Wheat germ extract. The wheat germ was added to warm distilled 
water in the proportion of 3 g. wheat to 50 cc. water. This mixture was shaken 
for 14 hours and then filtered through paper. The volume of the filtrate was 
measured and the liquid boiled to coagulate some of the protein. When cold 
the volume was made up with distilled water to what it was before boiling and 
the protein filtered off. This filtrate was fed to the animals. 

(d) Extract rich in vitamin B made from crude “marmite.” 250 g. “ mar- 
mite’ were dissolved in 400 cc. distilled water and transferred to a Winchester 
quart bottle. 1600 cc. absolute alcohol were added and the whole shaken for 
3 hours. This liquid was decanted (filtered if necessary) and distilled in vacuo 
at 40°-50° until all the alcohol had been distilled off. The remainder was left 
to cool, filtered, and the volume made up to 250 cc. with distilled water. It 
was boiled for 4 hours to sterilise. 

(e) Egg yolk solution. The egg yolk was separated from the white as much 
as possible, 50 cc. warm distilled water added and the whole well mixed. 


The diet given. 

In all the experiments to be described in this paper the proportion of white 
bread to protein was the same as that previously used as an excess protein 
diet, 7.e. 15 g. bread; 5 g. protein (caseinogen). 

The initial basal ration consisted of: 15 g. bread, 5 g. protein (cassinogent), 
0-7 g. salt mixture, 0-5 g. butter, 0-7 cc. lemon juice. 

To this were added 50 cc. of fruit or vegetable juice, meat extract, or egg 
yolk solution ete. 

When wheat germ or soya bean was given as solids, 3 g. were added to 
the above ration and water was used to mix the dry constituents. 

All the constituents were increased proportionately when necessary. 


III. Resvutrs. 


(i) Vegetable and fruit juices. 
Jerusalem artichoke, carrot, tomato and potato gave excellent results. The 
baby rats were normal in all respects and were extraordinarily fit and healthy. 
The litters belonging to the mothers which were given tomato juice in their 


MAMMARY SECRETION 3 829 


diets, were, perhaps, just fitter than the others and had very thick coats. 
Representative growth curves are shown in Fig. 1, 437, 382, 473, 373. 

Apple, orange, beetroot and vegetable marrow juices were also effective in 
preventing very bad symptoms in the young, but the babies were not so 
healthy and fit as were those whose mothers received artichoke, carrot, tomato 
or potato juice, although the “apple babies” grew at a greater rate. 

In all these experiments slight spasms were seen in the young, but the 
condition of the babies was never very bad. The spasms were generally 
noticed for one day only, and the growth curve was not interfered with, 7.e. 
no loss of weight was observed. In the case of the “apple babies” the rate of 
growth was practically maximal. From the 18th-20th day screaming fits 
[ Hartwell, 1921, 2] were noticed, but in spite of this the babies seemed well 
and gained weight. All were successfully weaned at the 21st day (Fig. 1, 526, 


B41, 422, 421, 608). 


Cucumber juice when fed to the mother was of no effect in relieving the 
bad symptoms in the young. All three litters suffered badly, the spasms were 
severe and only 4 out of the 18 young survived. These were weakly for about 
two weeks after weaning (Fig. 1, 495, 501). 

Grape juice, gave even worse results than cucumber juice. All three litters 
suffered badly and none of the young survived (Fig. 1, 597). 


(ii) Wheat germ and soya bean. 

These two foods were tried because both are reported to be rich in 
vitamin B. | 

(a) Wheat germ 3-0 g. and wheat germ extract 50 cc. (preparation see p. 828) 
were most effective. The litters did well and showed none of the bad symptoms. 
In the first series of experiments, it is interesting to note that the mother was 
receiving still more protein, since wheat germ itself contains 24-3 °% of protein 
[Plimmer, 1921] (Fig. 2, 405, 480). 

(b) Soya bean. When 3g. were added to the mother’s basal diet the 
litters suffered severely; all the babies had bad spasms; 11 out of 16 
survived, but were weakly for some time after weaning. 

Soya bean contains 33-7 °% protein [ Plimmer, 1921] and therefore it seemed 
possible that the bad results of the above experiments might be explained by 
the increase of protein in the mother’s diet. Accordingly three more experi- 
ments were tried in which the soya bean was increased from 3 to 6g. in 
proportion to the 15g. bread, 5g. protein etc. In this case the babies all 
survived and were practically normal. Slight spasms were noticed, but only for 
one day. The results were comparable to those obtained when the mother 
received apple juice (Fig. 3). 


(iii) Fish and meat extracts. Egg yolk. Whey. 


Lean muscle (beef ) and herring extracts when added to the mother’s diet 
entirely protected the young from any harmful effects (Fig. 4, 390, 433). 


830 G. A. HARTWELL 


Cod extract was nearly as good, but not quite. No spasms were noticed, 
but the baby rats were not quite so lively as normal animals. However, three 
days after weaning (on a bread and whole milk diet) they had recovered en- 
tirely (Fig. 4, 447). 

200;-— 

190 

180 

170 

160 

150 

140 

130 

120 

110 

100 

90 
80 
70 
60 
50 
40 
30 
20 
1} 
ey om et ez 
9 10 1112 13 14 15 16 17°18 19 20 21 
Fig. 2. 
Mothers’ diets 


405 « e 15g. bread, 5 g. caseinogen, 0-5 g. butter, 0-7 g. salt mixture, 0-7 cc. lemon 
juice, 3-0g. wheat germ. 


WEIGHTS OF LITTERS 


* baby rats eating ; 


480 B® —— @ 15b., 5c., 0-5b., 0-7 salt, 0-7 lemon (see above), 50 cc. wheat germ extract. 

369 A — A fe ‘i m 3cc. extract from marmite. 

531 @--@ } 15b., 5c., 0-5 b., 0-7 salt, 0-7 lemon (see above), 50 cc. wheat extract previously 

627 0 --o0 boiled with charcoal. 

533 A——A 15b., 5c., 0-5b., 0-7 salt, 0-7 lemon (see above), 50 cc. marmite extract pre- 
viously boiled with charcoal. 


Egg yolk was extraordinarily good. The growth curve of the young was 
equal to that obtained on any diet (7.e. maximal) and the babies were abso- 
lutely fit in all respects. Their coats were especially thick and long. This is 
interesting because egg yolk itself contains 15% of protein (Fig. 4, 464). 

Whey. It has previously been shown that the whey from 100 cc. whole 
milk when added to the mother’s excess protein diet was adequate in safe- 
guarding the young. In these experiments 50 cc. were tried in order to 
compare these results with those obtained when the mothers received 50 cc. of 
fruit juice, ete. 


= S. - 


~~ 


MAMMARY SECRETION 831 


The babies suffered badly; 13 out of 24 survived, but these were miserable 
specimens. They were weakly and their coats were thin and dirty (a typical 
sign that an animal is not fit). They were put on a bread and milk diet on 


weaning, but it was over a fortnight before they could be considered normal 
(Fig. 4, 515). 


(iv) Ezatract rich in vitamin B made from crude “marmite.” 
(Preparation given on p. 827.) 
3cc. of the extract were added to the basal ration, and proved most 
effective. All three litters did well and were normal in all respects (Fig. 2, 369), 
The results of the above experiments are tabulated on p. 832. The dis- 
tribution of vitamin B as given in the report of the Medical Research Council 
[1919] and by Eddy [1921] is included. 


180 ;— Pati: 
170}— re 
160}— Seteadaee 
150}—- ont 
140} . 
130}- Lelie. oer 0432 
Po a" > ft 
120}— ie ie oo & 9 
i 
110 = z a yah Lae 
100}~ § di Pe & 387 
g0}-> vi nee a 
80}— er ee 5 on. ac 
a 
70}— "aed ae % 
Pe ee a Se 386 
60}— fg a ot a as 
A e ga Ks a 
50k ow PF P03 % baby rats eating Ag 
she er. 
o 1 gti, WEIGHTS OF LITTERS 
30f<a—a—a— 
20;+— 
10}— Days 
‘book cused a ED SAE Fe A  eem  ek, NN s 
eed. Sa oo G29 Oye 401) 12 18 14°10 16 ‘17:18. 19 20 21 29 2a 24 
Fig. 3. 
Mothers’ diets 
543 = 15 g. bread, 5 g. caseinogen, 0-5 g. butter, 0-7 g. salt mixture, 0-7 cc. lemon 
432 0 ——-oO juice, 6 g. soya bean. 


386 A —— nt 15 g. bread, 5 g. caseinogen, 0-5 g. butter, 0-7 g. salt mixture, 0-7 cc. lemon 
387 A —— A juice, 3g. soya bean. 


(v) Experiments with a diet physiologically incomplete. 
The rats were fed on a diet of bread and caseinogen, as before, in the 


proportion of 15:5 and to this was added tomato juice, wheat germ or 
marmite extract. Such diets were incomplete physiologically, since salt 


mixture and fat were omitted. Also vitamin A would be poorly represented, 
if present at all. As explained in a previous paper [Hartwell, 1922] the 


832 - G. A. HARTWELL 


caseinogen used was not pure and might therefore contain a small amount of 
vitamin A, as would also the wheat germ and tomato juice. 

Three mothers were fed on each diet and in all nine experiments they re- 
mained fit and well. The young had a good growth curve and no spasms were 
seen. The baby rats appeared perfectly fit and healthy, and were quite equal 


to those whose mothers were getting a more complete diet (Fig. 5, 396, 519, 
553). 


Fruit and vegetables. 
M.R.C. Eddy 
50 ce. juice on 3+system on 4+system Condition of litters 
Applet Not given ~ ++ Slight spasms. Growth good 
Artichoke BF Not given Nospasms. Growth good 
Beetroot? Pk + Spasms. Babies weakly, but all 18 
survived 
Carrot* ae +++ No spasms. Growth good 
Cucumber Not given Not given Bad spasms. 4 out of 18 young sur- 
vived 
Grapes* Pa + Bad spasms, all young died 
Orange? us +++ Slight spasms. Growth good 
Potato + +++ No spasms. Slight jerky type of 
walking. Growth good 
Soya bean (3 g.) ++ + + Bad spasms, 11 out of 16 young sur- 
vived, but were weakly 
as » (6g.) ++ +++ Very slight spasms. Babies normal 
and healthy 
Tomato? Not given +++ No spasms. Growth good. Babies 
coats very thick. Young very fit 
Vegetable marrow ” _ Notgiven Slight spasms, but quite good growth ~ 
Wheat germ® 3 g. +++ +++ No spasms. Very good growth 
Wheat germ extract — — *s bee 35 
Yeast 3 g.° ++ 4 t++++ A ne nN 
Yeast extract “marmite” +++ t++ va a is 
Meat, fish, eggs, milk. 
Cod® White fish very + No spasms. Good growth. Not quite fit 
slight if any 
Egg yolk® Eggs + + + Eggs+ + No eer Very good growth, extra- 
ordinarily fit babies 
Herring Very slight if any ++ No spasms. Quite good growth 
Meat extract® 1 g. l ce. Lean meat + 0 No spasms. Good gro 
water 
Whe Milk + +++ Bad spasms. 13 out of 24 young sur- 
rf viet: but all were weakly 
1 Osborne and Mendel [1920, 2]. 
8 {1919}. 
® Sugiura and Benedict [1918]. 
Denton and Kohman [1918]. 


* Osborne and Mendel [1920, 1]. (Commercial grape juice.) 
5 Chick and Hume [1917, 1, 2}. 
® Cooper [1913). 
From such a series of experiments it seems justifiable to conclude that 
protein and the substance supplied in tomato, marmite, wheat germ etc. 
(? vitamin B) are of primary importance in the diet of a lactating animal. 


(vi) Haperiments demonstrating adsorption by charcoal of the accessory factor. 
From the experiments just described it is clear that the protective sub- 
stance has a distribution similar to that of vitamin B. It has been shown by 
Chamberlain and Vedder [1911] that the active principle (prepared from rice 


MAMMARY SECRETION 833 


polishings) which prevents polyneuritis in fowls is adsorbed by filtering through 
animal charcoal. 

Cooper [1913] found that a curative solution prepared from horse flesh 
lost its activity to the extent of 30 % when filtered six times through a bed 
of animal charcoal. It was therefore decided to try if the protective substance 
shared this property of vitamin B. Accordingly the following experiments 
were made in which the mother rats were fed on marmite and wheat germ 
extracts which had previously been boiled with animal charcoal. 


250/— 
240|— 
230}— ) 
220|— Ws 
210} 
200}— */ 
190}— A 
180}— *. 
170}— / ee 
160}— ae foe 
150} ee “ “9 18 
2 he Z*447 
140}—- v4 ms & V4 Pai 
130}— 3 : 
120}- eR ae coe 


110;— Sy ese ° 
100}— DIG ites 
-90}- f Peeks 


80}— Aas 
70 a % baby rats eating 


60}— yy, 
50}— ge 


a464 


Grams 
be 
9 
Ne 
\ 
bs 
\ 
Dd 
¥ 


40 BAZ WEIGHTS OF LITTERS 


30 
20;- 


10}— Days 


“2s Ss dS a a a 
7 8 9 10 11 12 18 14 15 1617 18 19 20 21 


Fig. 4. 


Mothers’ diets 
464 9 —— 1 15g. bread, 5g. caseinogen, 0-5g. butter, 0-7 g. salt mixture, 0-7 cc. lemon 
juice, 50 cc. egg yolk solution. 
433 A —— A 15b., 5c., 0-5 b., 0-7 salt, 0-7 lemon (see above), 50 cc. herring extract. 
390 e —— e Ra a a a 50 cc. meat extract. 
515 wy Pi ” Pf +: 50 cc. whey. 
447 e--e aa se ¥3 - 50 cc. cod extract. 


54—5 


834 G. A. HARTWELL 


Preparation of extracts. 

(a) Wheat germ. The extract was prepared as previously déscribed (p. 828). 
Animal charcoal was added and the whole boiled. When boiling, a little of 
the liquid was taken in a spoon; if the charcoal settled rapidly and left a clear 
liquid on top, the whole was allowed to cool. If the charcoal remained sus- 
pended in the liquid, more charcoal was added and the boiling continued until 
the clear liquid settled out when tested in the spoon. When cool the volume 
was made up with distilled water to what it was before boiling, and the whole 
filtered. This clear filtrate was added to the diet. 


210;— 
200}- of 519 
190;— ° Bf 

180}— : iy 

1 0 — ey : +553 
160}— *Y oe 

150}+— Je */ 

140/- 4 ts or 

130}— eee, 


120}— oo 
110}— Z, e 
Vj 
100;— WZ 
90;+— Ve toms 
80+- Vite < 
70 jf * 
60K Ze : 


ME KEE 2 WEIGHTS OF LITTERS \ 
40 6 


30 wes \ 
= 


Gra ms 


20 


10 Days 


‘ \ 
Ele: aes Fs PA ste a igi SE WE tah SR BP 

6 7 8 9 10 11 12 18 14 15 16 17 18 19 20 21 
Fig. 5. 


1, me 


Mothers’ diets 
396 « « 15g. bread, 5 g. caseinogen, 3 cc. extract from marmite. 
519 O——O * 3 g. wheat germ. 
553 & A ” ” 50 ce, tomato juice. 

N e--e sa s 50 cc, water (taken from Hartwell (1921, 1)) for 

comparison, 

(b) “Marmite” (yeast extract), 6 4, marmite solution was made with hot 
water. This was boiled with charcoal (as described above), cooled, the volume 
made up and filtered. This process was repeated (usually three times) until 
a clear, colourless filtrate was obtained. This was used for the feeding. 


MAMMARY SECRETION | 835 


The diet was the same as that used in the preceding experiments, 7.¢. 
15g. bread, 5g. protein, 0-5 g. butter, 0-7 g. salt mixture, 0-7 cc. lemon 
Juice + 50 cc. of one of the above extracts. 

Three experiments were made with each diet. 

Results. When the mothers received the marmite extract which had 
been boiled with animal charcoal, the babies suffered badly and none survived. 
A typical curve is shown in Fig. 2 with the yeast “charcoal” extract, two 
litters died and one survived, though all had bad spasms. The young which 
survived were weakly and were not normal for some time after weaning. 

The slight difference in these results may be accounted for by the fact 
that the marmite extract was boiled three times with charcoal and the wheat 
extract, only once; so that in the former case there might be more complete 
adsorption (Fig. 2, 527, 531, 533). 


IV. CoMMENTs. 


Many natural foods or extracts made from them, contain a factor which 
protects the suckling rats from the harmful effects of excess of protein in the 
mother rat’s diet. 

The distribution of this factor is similar to that of vitamin B. The foods 
reported to be especially rich in this vitamin, e.g. egg yolk, wheat germ, when 
added to the mother’s diet not only prevent the bad symptoms, but provide 
for good growth in the young. Grape juice when fed to the mother was of no 
effect in relieving the symptoms in the baby rats; all the 18 young (3 litters) 
died. According to Eddy the vitamin B content of grapes is represented by +- 
(see Table, p. 832). 

There are, however, discrepancies between my findings, the work of other 
physiologists, and the tables published by the Medical Research Council and 
Eddy. 

Drummond [1919] finds that there are no appreciable amounts of the 
water-soluble factor in the muscles of cod and herring. Eddy gives the 
vitamin B content of cod as + and herring + +. According to the experi- 
ments described in this paper, the extracts from cod and herring muscle had 
appreciable protective properties, since no spasms were seen in the young, 
when such extracts were added to the mother rat’s diet. 

Again Eddy represents orange and tomato ds equal and containing + + + 
and apple as + +. My experiments show tomato to be very superior to orange 
and that apple and orange are about equal. When the mothers received a 
ration of 50 ce. orange juice per day, the litters developed spasms, while a 
ration of 25 cc. tomato juice added to the mother’s diet was sufficient to give 
complete protection to the young. 

Soya bean gave interesting results (Eddy + + +, M.R.C. Report + +). 
When 3g. were fed to the mother rats, the babies developed spasms and 
were in a very poor condition; but with 6g. instead of 3g., the litters 


836 G. A. HARTWELL 


were practically normal, only slight spasms were noticed and the growth was 
good. Soya bean contains 33-7 % of protein and hence adding 3g. soya 
bean to the mothers’ diet also increased the protein ration. With 6¢g., 
however, the vitamin content of the diet was nearly doubled, while the per- 
centage of protein was very little different from that of the diet with 3 g. 
protein. On the whole my results agree more closely with those of Eddy 
than with those published in the Medical Research Council’s Report. 

It is possible that the disagreement may be due to: 

1. Method of investigation—some observers find the curative dose, while 
others merely add a definite amount of the food under investigation to a diet 
deficient in vitamin B. 

2. The method of preparation of extracts used, since vitamins easily form 
adsorption compounds, ; 

Further the protective factor is similar to vitamin B in its solubility and 
its property of adhering to charcoal. Consequently there is presumptive 
evidence that the protective factor is vitamin B. This is interesting as 
vitamin B is thus definitely linked with protein metabolism; the one gland 
that synthesises protein in large amount functions badly in the absence of 
vitamin B when there is excess of amino acids in the blood. The lactating 
animal, when there is excess of protein in the diet, needs much more vitamin B 
than normal, not for herself, but to regulate the activity of the mammary 
glands. 


SUMMARY. 


1. When 50 ce. tomato, artichoke or carrot juices are added to the mother’s 
basal high protein diet, the litters do well and show no harmful effects, such 
as would otherwise be produced by such a diet. 

3g. wheat germ or 50 cc. wheat germ extract give equally good results. 
Soya bean, 3g., are ineffective, but 6g. added to the above diet greatly 
improve the condition of the litters. 


2. Apple and orange juices give good results, but not quite equal to those 
obtained with potato etc. The litters show good growth, but slight spasms 
are noticed. These do not last long and the growth curve is not impaired. 


3. Wheat germ extract and marmite extract after being boiled with 
animal charcoal have no beneficial effect. 


4. Beef muscle, herring, and cod extracts, when added to the mother’s 
diet have also the property of protecting the young from any bad effects. 
Egg yolk has a similar protective action and the growth curve of the young 
is exceptionally good. | 

5. The substance in foods which exercises this protective action resembles 
vitamin B in its distribution, its solubility and the ease with which it is 
adsorbed by charcoal. | 


MAMMARY SECRETION 837 


I wish to thank Prof. V. H. Mottram for his criticism and advice. Thanks 
are also due to the Medical Research Council for a grant which defrayed the 
expenses of this research. . 


REFERENCES. 


Chamberlain and Vedder (1911). Philippine Journal of Science, 6, 395. 
Chick and Hume (1917, 1). J. R.A.M.C. 29, 121. 
—— (1917, 2). Proc. Roy. Soc. B. 90, 44. 
Cooper (1913). Biochem. J. 7, 268. 
Denton and Kohman (1918). J. Biol. Chem. 36, 249. 
Drummond (1919). J. Physiol. 52, 95. 
Eddy (1921). The Vitamine Manual, 59. 
Hartwell (1921, 1). Biochem. J. 15, 140. 
—— (1921, 2). Biochem. J. 15, 563. 
—— (1922). Biochem. J. 16, 78. 
Karr (1920). J. Biol. Chem. 44, 255. 
Medical Research Committee (1919). Special Report Series, No. 38, 50. 
Osborne and Mendel (1919). J. Biol. Chem. 39, 29. 
—— —— (1920, 1). J. Biol. Chem. 41, 451. 
—— —— (1920, 2). J. Biol. Chem, 42, 465. 
Plimmer (1921). Analyses and Energy Values of Foods. 
Sugiura and Benedict (1918). J. Biol. Chem. 36, 171. 


INDEX 


Accessory food factors, conditions of inactiva- 
tion of (Zilva) 42 


Andrenaline, action of, on secretion of mam-. 


mary gland (Rothlin, Plimmer and Husband) 
3 


Amino-acids of flesh (Rosedale) © 27 
Annett, H. E. The enzymes of the latex of the 
Indian poppy (Papaver somniferum) 765 


Antineuritic vitamin, in urine (van der Walle) 


713 

Antiscorbutic vitamin in urine (van der Walle) 
713 

Antithrombins lao and Hewitt) 587 

Apples, soluble pectin of (Carré and Haynes) 
60 


Artemisia afra, constituents of flowering tops 
of (Goodson) 489 , 

Arxtnson, E. and Hazieron, E. O. A qualita- 
tive tannin test 516 

Atrolactic acid, fate of, in body (Kay and 
Raper) 465 

Atropic acid, fate of, in body (Kay and Raper) 
465 


Autolysis of beef and mutton (Fearon and 
Foster) 564 


Bacillus diphtheriae, mechanism of reversal in 
reaction of medium, caused by (Wolf) 541 

Bacor, A. W. and Harpen, A. Vitamin require- 
ments of Drosophila. I. Vitamins B and C 
148 

Bacteria, production of hydrogen peroxide by 
(McLeod and Gordon) 499 

Beef, autolysis of (Fearon and Foster) 564 

Bence-Jones’ Protein, nitrogen-distribution in 
(Liischer) 556 

Berkevey, C. On the occurrence of manganese 
in the tube and tissues of Mesoc terus 
Taylori, Potts, and in the tube of Chaeto- 

Variopedatus, Renier 70 
Biological value of proteins (Martin and Robison) 


407 

Blood, calcium in, estimation of (Ling and 
Bushill) 403 

Blood, coagulation of (Pickering and Hewitt) 
587 


Blood-corpuscles, red, influence of the spleen 
upon (Bolt and Heeres) 754 
Blood enzymes (Compton) 460 
Blood, human, nature of reducing substance 
in (Cooper and Walker) 455 
Blood, non-protein nitrogen in, estimation of 
(Ponder) 368 
Blood, of equines (Neser) 770 
Blood, wale, action of, upon acids (Kennaway 
and Mel 7 
Bour, N. A. and Heeres, P. A. On the in- 
fluence of the spleen upon red blood-cor- 
puscles I, 754 
Bonp, M. A modification of basal diet for rat 
feeding experiments 479 


Booox, E. and Trevan, J. The food value of 
mangolds and the effects of deficiency of 
vitamin A on guinea-pigs 780 

Braprorp, 8. C.. An improvised electric 
thermostat constant to 0-02° 49 

Buswiwy, J. H., see Lina, A. R. 


Calcium, estimation of, in blood (Ling and 
Bushill) 403 
Calcium, estimation of small quantities of 
(Laidlaw and Payne) 494 
CAMPBELL, J. A. and Wesstsr, T. A. Effect of 
severe muscular work on composition of the 
urine 106 wy, 
CampBety, J. A. and Wesster, T. A. Note on 
urinary tides and excretory rhythm 507 
Cannan, R. K., see DRumMonD, J.C. ~ 
Cannon, H. C., see OsBoRNE, T. B. 
Carbohydrate, influence of, on nitrogen distribu- 
tion in urine (Cathcart) 747 
Carbohydrates, oxidation of, with nitric acid 
(Haas and Russell-Wells) 572 
Carbonic acid in blood and salt. solutions 
(Warkurg) 153 
Carnosin& catalytic destruction of, in vitro 
(Clifford) 792 
Carnosine, content of muscle, effect of cold 
storage on (Clifford) 341 
Carnosine, estimation of, in muscle extract 
(Hunter) 640 
Carrageen, constitution of cell wall of (Russell- 
Wells) 578 
Carré, M. H. An investigation of the changes 
which occur in the pectic constituents of : 
stored fruit 704 
Carri, M. H. and Haynzs, D. The estimation 
of pectin as calcium pectate and the applica- 
tion of this method to the determination of 
the soluble pectin in apples 60 
CasTeLLAnt, A, and Taytor, F. E. Identifica- 
tion of inulin by a mycological method 655 
Carucart, E. P. The pay bee of fat and carbo- 
hydrate on the nitrogen distribution in the 
urine 747 
Chaetopterus variopedatus, occurrence of man- 
ganese in tube of (Berkeley) 70 
CutBnaLy, A. C. Investigations on the nitro- 
enous metabolism of the higher plants. 
art II. The distribution of nitrogen in the 
leaves of the runner bean 344 
Curpnaui, A, C. Investigations on the nitro- 
nous metabolism of the higher plants, 
art III, The effect of low-temperature 
drying on the distribution of nitrogen in 
the leaves of the runner bean 599 
Curenaty, A. C. Investigations on the nitro- 
nous metabolism of the higher plants. 
art [V. Distribution of nitrogen in the 
dead leaves of the runner bean 608 
Chickens, a o4 good protein on (Plimmer and 
© 


INDEX 


Chickens, enzymes in alimentary canal of 
(Plimmer and Rosedale) 23 

Chickens, vitamin requirements of (Plimmer 
and Rosedale) 11 

Chondrus crispus, constitution of cell wall of 
(Russell-Wells) 578 

CiirForD, W. M. The effect of cold storage on 
the carnosine content of muscle 341 

CiirForD, W. M. The catalytic destruction of 
carnosine in vitro 792 

Coagulation of blood (Pickering and Hewitt) 
587 

Cod liver oil, refined, vitamin A content of 
(Stammers) 659 

Conren, A. Xylenol blue and its proposed use 
as a new and improved.indicator in chemical 
and biochemical work 31 

Colloids, osmotic pressure of solutions of, in- 
fluence of salt solution on (Ogata). 449 

Compton, A. Blood enzymes. II. The influence 
of temperature on the action of the maltase 
of dog’s serum 460 

Cooper, E. A. and WaLKEr, H. Further obser- 
vations on the nature of the reducing sub- 
stance in human blood 455 

Cowarp, K. H. and Drummonp, J. C. On the 
significance of vitamin A in the nutrition 
of fish 631 

Cowarp, K. H., see also. DrumMmonp, J. C., 
Goxtpina, J., Jameson, H. L. 

Cricuton, A., see Primmer, R. H. A. 

Cystein, metallo-derivatives of (Harris) 739 


Di-amino-acid content of muscle (Rosedale) 27 

Diatom, marine, synthesis of vitamin A by 
(Jameson, Drummond and Coward) 482 

Dibenzoyleystine (Wolf and Rideal) 548 

Diet, basal, for rats, modification in (Bond) 
479 

gm physiological effects of (Ransom) 


Drosophila, vitamin requirements of (Bacot and 
Harden) 148 

Drummonp, J.C. and Cannan, R. K. Tethelin— 
the alleged growth-controlling substance of 
the anterior lobe of the reas 4 gland 53 

Drummonp, J. C. and Ziiva, 8. 8., with the 
co-operation of Cowarp, K. H. The origin 
of the vitamin A in fish oils and fish liver 
oils 518 

Drummonp, J. C., see also Cowarp, K. H., 
Goupina, J., Jameson, H. L. 


Eels, respiratory exchange of (Gardner and 
King) 736 

Enzymes, distribution of, in alimentary canal 
of chickens (Plimmer and Rosedale) 23 

Enzymes of blood (Compton) 460 

— of latex of Indian poppy (Annett) 
765, 

Equines, the blood of (Neser) 770 

Ergamine, action of, on secretion of mammary 
gland (Rothlin, Plimmer and Husband) 3 

Excretory rhythm (Campbell and Webster) 507 


Fat, formation of, by the yeast cell, conditions 
influencing (Smedley MacLean) 370 

Fat, influence of, on nitrogen distribution in 
urine (Cathcart) 747 


839 


Fat-soluble factor, relation of, to rickets and 
growth in pigs (Golding and colleagues) 394 

Fatty acids, with branched chains, mode of 
oxidation of (Kay and Raper) 465 

Fearon, W. R. and Foster, D. L. The autolysis 
of beef and mutton 564 

Frypiay, G. M. and Mackrnzir, R. Opsonins 
and diets deficient in vitamins 574 

Fish, fresh water, respiratory exchange in 
(Gardner, King and Powers) 523 

Fish liver oils, origin of vitamin A in (Drummond 
and Zilva) 518 

Fish oils, origin of vitamin A in (Drummond 
and Zilva) 518 

Fish, vitamin A in nutrition of (Coward and 
Drummond) 631 

Foster, D. L., see Fearon, W. R. 

Fruit, stored, changes which occur in pectic 
constituents of (Carré) 704 


Garpner, J. A. and Kine, G. Respiratory ex- 
change in fresh water fish. Part 1V. Further 
comparison of gold-fish and trout 729 

GarpDNER, J. A. and Kine, G. Respiratory 
exchange in freshwater fish. Part V. On 
eels 736 

GARDNER, J. A., Kina, G., and Powmrs, E. B. 
The respiratory exchange in fresh water 
fish. II]. Gold fish 523 

Gelatin, dialysed, properties of (Lloyd) 530 

Gelatin, value of, in relation to nitrogen require- 
ments of man (Robison) 111 

Gels, elastic, structures in, caused by formation 
of semipermeable membranes (Hatschek) 
475 

Guoskg, 8. N. The examination of some. Indian 
foodstuffs for their vitamin content 35 

Glycine, synthesis of, from formaldehyde (Ling 
and Nanji) . 702 


Gold-fish, respiratory exchange in (Gardner and 


King) 729 

Gold-fish, respiratory exchange in (Gardner, 
King and Powers) 523 

Go.pinG, J., Zirva, 8. 8., DRummMonp, J. C. and 
Cowarp, K. H. The relation of the fat- 
soluble factor to rickets and growth in 
pigs, IT 394 

Goopson, J. A. The constituents of the flowering 
tops of Artemisia afra, Jacq. 489 

Gorpon, J., see McLxop, J. W. 

Growth, in pigs, relation of fat-soluble factor 
to (Golding and colleagues) 394 

Guinea-pigs, effects of vitamin A deficiency 
upon (Boock and Trevan) 780 


Haas, P. and RussgeLtt-Wetis, B. Note on 
the oxidation of carbohydrates with nitric 
acid 572 

Harpen, A. and Henry, F. R. The function 
of phosphates in the oxidation of glucose 
by hydrogen peroxide 143 

Harpen, A., see also Bacor, A. W. 

Harris, L. J. On a series of metallo-cystein 
derivatives, | 739 

Hartwetyt, G. A. Mammary secretion. ITI. 
1. The quality and quantity of dietary pro- 

tein. 
2. The relation of protein to other dietary 
constituents 78 


840 


HartweE tt, G. A. Mammary secretion. IV. The 
relation of protein to other dietary con- 
stituents 825 

Hartscuek, E. Structures in elastic gels caused 
by the formation of semipermeable mem- 
branes 475 

Haynes, D., see Carré, M. H. 

Hazzeton, E. O., see ATKINSON, E. aS. 

Heat-coagulation of proteins (Lepeschkin) 678 

Heaton, T. B., On the vitamin. D 800 

Heerss, P. A., see Bout, N. A. 

Henderson and Hasselbach’s equation, con- 
tribution to theory of (Warburg) 153 

Hewirt, J. A., see PickERING, J. W. 

Hexoses, new phosphoric Ester produced by 
action of yeast juice on (Robison) 809 

Histidine, Knoop’s test for (Hunter) 637 

Hoesen, L. T. and Winton, F. R. Studies on 
the pituitary. I. The melanophore stimulant 
in posterior lobe extracts 619 

Hunter, G. Note on Knoop’s test for histidine 
637 

Hunter, G. The estimation of carnosine in 
muscle extract—a critical study 640 

Husspanp, A. D., see Roruit, E. 

Hydratropic acid, fate of, in body (Kay and 
Ra 


per) 4659 
Hydrogen ion activities in blood and salt solu- 
Tiydeogen vunictics yootienlich of er 
nm peroxide, uction of, acteria 
7 (McLeod and Gordon) 499 : 
Hypophysin, action of, on secretion of mam- 
mary gland (Rothlin, Plimmer and Hus- 
band) 3 


Inulin, identification of, by mycological method 
(Castellani and Taylor) 655 


Jameson, H. L., Drummonp, J.C. and Cowarp, 
K. H. Synthesis of vitamin A by a marine 
diatom (Nitzschia closteriwum W.Sm.) growing 
in pure culture 482 


Kay, H. D. and Rargr, H. 8. The mode of 
oxidation of fatty acids with branched 
chains. II. The fate in the body of hydra- 
tropic, tropic, atrolactic and atropic acids 
together with may pe seaenening 465 

Kenwnaway, E. L. and MoIyrosn, J. The action 
of whole blood upon acids 380 

Kuna, G., see Garpnir, J. A. 


Larwiaw, P. P. and Paynz, W. W. A method 
for the estimation of small quantities of 
calcium 494 

Lerescuxix, W. W. The heat-coagulation of 
proteins 678 

Luya, A. R. and Busuit, J. H. The estimation 
of calcium in blood 403 

Lana, A. R. and Nangt, D. R. The synthesis 
of glycine from formaldehyde 702 

Iaoyp, D. J. Notes on some properties of 
dialysed gelatin 530 

Liscurn, E. The nitrogen-distribution in Bence- 
Jones’ protein, with a note upon a new 
colorimetric method for tryptophan-estima.- 
tion in protein 556 


Mackenziz, R., sce Finpiay, G. M. 


INDEX 


Maltase, of dog’s serum, influence of tempera- 
ture on action of (Compton) 460 

Mammary gland, action of hypophysin, ergamine 
and adrenaline on secretion of (Rothlin, 
Plimmer and Husband) 3 

Mammary secretion (Hartwell), 78, 825 

Manganese, occurrence of, in marine annelids 
(Berkeley) 70 

Marigolds, food value of (Boock and Trevan) 780 

Martry, C. J. and Rostson, R. The minimum 
nitrogen expenditure of man and the bio- 
logical value of various proteins for human 
nutrition 407 

McIntosu, J., see Kennaway, E. L. 

McLzop, J. W. and Gorpon, J. Production 
of hydrogen peroxide by bacteria 499 
Melanophore stimulant in extract of posterior 

lobe of pituitary (Hogben and Winton) 619 
MENDEL, L. B., see OspornzE, T. B. 
Mesochaetopterus Taylori, occurrence of man- 

ganese in tube and tissues of (Berkeley) 70 
~— ae growth promoting vitamin (Stammers) 


Milk as a source of water-soluble vitamin 
(Osborne and Mendel) 363 

Milk, goat’s, non-protein nitrogen in (Taylor) 
611 


Muscle, carnosine content of, effect of cold 
storage on (Clifford) 341 

Muscle, diamino-acid content of (Rosedale) 27 

Muscle extract, estimation of carnosine in 
(Hunter) 640 

Mutton, autolysis of (Fearon and Foster) 564 


Naxacawa, T. The relation of salivary to 
gastric secretion 390 

Nana, D. R., see Lina, A. R. 

Neeser, C. P. The blood of equines, 770 

Nicory, C. Salivary secretion in infants, 387 

Nrerenstery, M. Note on a new tannase from 
Aspergillus Luchuensis, Inui. 514 

Nitrogen, distribution of, in dead leaves of 
runner bean (Chibnall) 608 

Nitrogen, distribution of, in leaves of the runner 
bean (Chibnall) 344 

Nitrogen distribution of, in leaves of the runner 
bean, effect of low temperature drying on 
(Chibnall) 599 

Nitrogen, minimum expenditure of, in man 
(Martin and Robison) 407 

Nitrogen, non-protein, estimation of, in blood 
(Ponder) 368 

Nitrogen, non-protein, in goat’s milk (Taylor) 
611 


Nitrogen requirements of man, value of gelatin 
in relation to (Robison) 111 

Nitrogenous metabolism of the higher plants 
(Chibnall), 344, 599, 608 

Nitzschia closterium synthesis of vitamin A by 
(Jameson, Drummond and Coward), 482 


Oaata, D, On the change of the osmotic 
pressure of solutions of certain colloids under 
the influence of salt solutions 449 

Opsonins and diets deficient in vitamins (Findlay 
and Mackenzie) 574 

Osnorn®, 'T. B. and Munper, L. B. (with the 
cooperation of Cannon, H, C.), Milk as a 
source of water-soluble vitamin, III 363 


INDEX 841 


Osmotic pressure of colloid solutions, influence 
of salt solutions on (Ogata) 449 

Oxidation of fatty acids with branched chains 
(Kay and Raper) 465 

Oxidation of glucose by hydrogen peroxide, 
function of phosphates in (Harden and 
Henley) 143 


Papaver somniferwm, enzymes of latex of 
(Annett) 765 

Payne, W. W., see Larpiaw, P. P. 

Pectic constituents of stored fruit; changes 
which occur in (Carré) .704 

Pectin, estimation of, as calcium pectate (Carré 
and Haynes) 60 

Pectin, soluble, in apples (Carré and Haynes) 


60 
Phenylacetaldehyde, fate of, in body (Kay and 
465 


—) 

Phosphates, function of in oxidation of glucose 
by hydrogen peroxide (Harden and Henley) 
143 


Phospholipin of blood and liver in experimental 
rickets in dogs (Sharpe) 486 

Phosphoric ester, a new, produced by the action 
of. yeast juice on hexoses (Robison) 809 

Pioxerinea, J. W. and Hewirrt, J. A. Studies 
of the coagulation of the blood. Part II. 
Thrombin and antithrombins 587 

Pigs, rickets and growth in, relation of fat- 
soluble factor to (Golding and colleagues) 
394 

Pituitary, melanaphore stimulant in extracts of 
posterior lobe of (Hogben and Winton) 619 

Plant cell, suction pressure of the (Stiles) 727 

Prrmmer, R. H. A. and Rosepare, J. L., with 
the assistance of Cricuton, A. and Toprrne, 
R. B. The rearing of chickens on the inten- 
sive system. Part I. The vitamin require- 
ments. Preliminary experiments 11 

R. H. A. and Rosepatg, J. L., with 

the assistance of Cricuton, A. and Toprrna, 
R. B. The rearing of chickens on the inten- 
sive system. Part II. The effect of “good” 
protein 19 

Purmmer, R. H. A. and Rosepare, J. L. 
Distribution of enzymes in the alimentary 
canal of the chicken 23 

PurmeERr, R. H. A., see also Roruiin, E. 

PonvEr, E. The estimation of non-protein 
nitrogen in blood 368 

Poppy, Indian, enzymes of latex of (Annett) 
765 


Powers, E. B., see GARDNER, J. A. 

Protein, dietary, effect of quality and quantity 
of, on mammary secretion (Hartwell) 78 

Protein, relation of, to other dietary con- 
stituents, effect of on mammary secretion 
(Hartwell) 78 

Protein, relation of, to other dietary constituents 
in effect on mammary secretion (Hartwell) 
825 

Proteins, biological value of (Martin and 
Robison) 407 

Proteins, heat-coagulation of (Lepeschkin) 678 


Ransom, F. On the cardiac, haemolytic and 
nervous effects of digitonin 668 
Raper, H. S., see Kay, H. D. 


Reaction, reversal of, on growth of B. diphtheriae 
(Wolf) 541 

Reducing substance in human blood, nature of 
(Cooper and Walker) 455 

Respiratory exchange in fresh water fish 
(Gardner and King) 729, 736 

Ratp, D. and Surru, F. E. Note on tannase 1 

Rickets, experimental in dogs, phospholipin of 
blood and liver in (Sharpe) 486 

Rickets, in pigs, relation of fat-soluble factor 
to (Golding and colleagues) 394 

Rivet, E. K., see Wor, C. J. L. 

Rostson, R. The value of gelatin in relation 
to the nitrogen requirements of man 111] 

Rosison, R. Distribution of the nitrogenous 
constituents of the urine on low nitrogen 
diets 131 

Rosison, R. The estimation of total sulphur 
in urine 134 

Rostson, R. A new phosphoric ester produced 
by the action of yeast juice on hexoses 809 

Rostson, R., see also Marttn, C. J. 

Rosepatz, J. L. The amino-acids of flesh. The 
di-amino content of rabbit, chicken, ox, 
horse, sheep, and pig muscle 27 

RosEpAtE, J. 1 a see also PurmmeEr, R. H. A. 

Rorati, E., Prrmmmer, R. H. A., and HusBanp, 
A. D. The action of hypophysin, ergamine 
and adrenaline upon the secretion of the 
mammary gland 3 

Runner bean, dead leaves of, distribution of 
nitrogen in (Chibnall) 608 

Runner sadly distribution of nitrogen in leaves 
of (Chibnall) 344 

Runner bean, effect of low temperature drying 
on distribution of nitrogen in leaves of 
(Chibnall) 599 

Russett-WELLs, B. On carrageen (Chondrus 
crispus). III. The constitution of the cell 
wall 578 

RussELL- WELLS, B., see also Haas, P. 


Salivary secretion in infants (Nicory) 387 

Salivary secretion, relation of, to gastric 
(N awa) 390 

Semipermeable membranes, structures in elastic 
gels caused by formation of (Hatschek) 475 

Suarpeg, J. 8. The phospholipin of the blood 
and liver in experimental rickets in dogs 486 

SmepiEy MacLean. I. The conditions in- 
fluencing the formation of fat by the yeast 
cell 370 

Smell (Watson) 613 

Smuntu, F. E., see Rump, D. 

Spleen, influence of, on red blood-corpuscles 
(Bolt and Heeres) 754 

Srammers, A. D. Feeding experiments in con- 
nection with vitamins A and B. IL. Milk 
and the growth-promoting vitamin. IV. The 
vitamin A content of refined cod-liver oil 


659 
Stites, W. The suction pressure of the plant 
cell. A note on nomenclature 727 
Sulphur, total estimation of, in urine (Robison) 
134 


Tannase, new, from Aspergillus Luchuensis 
(Nierenstein) 514 
Tannase, note on (Rhind and Smith) 1 


O 


842 


Tannin, test, a qualitative (Atkinson and 
Hazleton) 516 

Tay tor, F. E., see CASTELLANT, A. 

Taytor, W. Note on the non-protein nitrogen 
in goat’s milk. 611 : 

Tethelin (Drummond and Cannan) 53 

Thermostat, animprovisedelectric( Bradford) 49 

Thrombin (Pickering and Hewitt) 587 

Toprrne, R. B., see PLimmer, R. H. A. 

TREVAN, J., see Boock, E. 

Tropic acid, fate of, in body (Kay and Raper) 
465 


Trout, respiratory exchange of (Gardner and 
King) 729 

Tryptophan, colorimetric estimation of, in 
protein (Liischer) 556 


Urinary tides (Campbell and Webster) 507 

Urine, antineuritic and antiscorbutic vitamins 
in (van der Walle) 713 

Urine, effect of severe muscular work on com- 
position of (Campbell and Webster) 106 

Urine, estimation of total sulphur in (Robison) 
134 

Urine, nitrogen distribution in, influence of fat 
and carbohydrate on (Cathcart) 747 

Urine, nitrogenous constituents of, on low 
nitrogen diets (Robison) 131 ° 


Vitamin A content of refined cod liver oil 
(Stammers) 659 

Vitamin A deficiency, effects of, on guinea-pigs 
(Boock and Trevan) 780 

Vitamin A in nutrition of fish (Coward and 
Drummond) 631 

Vitamin A, origin of, in fish- and fish liver-oils 
(Drummond and Zilva) 518 

Vitamin A, synthesis of by a marine diatom 
(Jameson, Drummond and Coward) 482 

Vitamin B, feeding experiments in connection 
with (Stammers) 659 


INDEX 


Vitamin content of Indian foodstuffs (Ghose) 35 

Vitamin D (Heaton) 800 

Vitamin requirements of chickens (Plimmer and 
Rosedale) 11 

Vitamin requirements of Drosophila (Bacot and 
Harden) 148 

Vitamin, water-soluble, milk as a source of 
(Osborne and Mendel) 363 

Vitamins, diets deficient in, and opsonins 
(Findlay and Mackenzie) 574 


WALKER, H., see Coopmr, E. A. 

Wate, N. van per. The presence of anti- 
neuritic and antiscorbutic vitamins in urine 
713 

Warsura, E. J. Studies on carbonic acid com- 
pounds and hydrogen ion activities in blood 
and salt solutions. A contribution to the 
theory of the equation of Lawrence J. 
Henderson and K. A. Hasselbach 153 

Warson, E. R. Smell 613 

WessTErR, T. A., see CAMPBELL, J. A. 

Winton, F. R., see Hocsen, L. T. 

Wo tr, C. G. L. The mechanism of the reversal 
in reaction of a medium which takes place 
during growth of B. diphtheriae 541 

Worr, C. G. L. and Ripxgat, E. K. The properties 
of dibenzoyleystine 548 

Wricut, O. K. The action of yeast-growth 
stimulant 137 


Xylenol blue (Cohen) 31 


Yeast cell, formation of fat by, conditions in- 
fluencing the (Smedley MacLean) 370 
Yeast-growth stimulant, action of (Wright) 137 


Zitva, 8. 8. Conditions of inactivation of the 
accessory food factors 42 

Ziuva, 8. S., see also DrumMonp, J. C. and 
GoLpine, J. 


CAMBRIDGE: PRINTED #Y J. B. PEACE, M.A., AT THE UNIVERSITY PRESS 


C 


BINDING SECT. JUN 10 1964 


| 
| 


OP The Biochemical journal 
re [422 
v.16 
cop.4 

Biological 

. & Medical 

Serials 


PLEASE DO NOT REMOVE 
CARDS OR SLIPS FROM THIS POCKET 


UNIVERSITY OF TORONTO LIBRARY 


STOP AGE 


BR ITIE Jouente 


‘Yu/Y 


BS 


esd 
oleae 


Sees 


poets 


oe 
ot 


eye = 


See 


~— 


a 


i aoeearaane 
ean 


waist eee 


ae ee