410 H. H. KOEFOED-JOHNSEN AND T. MANN I954ram spermatozoa by 50%; 00008-00012M wasrequired for complete inhibition of fructolysis andmotility. Among similarly active substances weredodecyl sulphate, 2-phenoxyethanol, dodecylaminehydrochloride, sodium taurocholate, 'Lubrol' and'Teepol'.

3. The surface-active compounds induced achange in the permeability of spermatozoa as indi-cated by the diffusion ofcytochrome c from the cellsinto the surrounding fluid.

4. When succinate was added to detergent-treated immotile spermatozoa, the percentage in-crease of the oxygen uptake was much greater thanin intact motile sperm; thus, the stimulating effectof succinate on the respiration of spermatozoa wasan indication of a moribund cell population.

This investigation was carried out on behalf of theAgricultural Research Council. We wish to thank Dr J. R. G.Bradfield of the Cavendish Laboratory, Cambridge, andDr J. H. Schulman of the Colloid Science Department,Cambridge, for their help. One of us (H. H. K.-J.) wishes toacknowledge the receipt ofa travellinggrant awarded by theRoyal Veterinary and Agricultural College, Copenhagen.

REFERENCES

Hockenhull, D. (1948). Nature, Lond., 162, 850.Hugo, W. B. & Street. H. E. (1952). J. gen. Microbiol.

6,90.Humphrey, G. F. & Mann, T. (1949). Biochem. J. 44, 97.Keiln, D. & Hartree, E. F. (1945). Biochem. J. 39,

289.MacLeod, J. (1943). Annu. Rev. Phy8iol. 5, 399.Mann, T. (1945). Biochem. J. 39, 451.Mann, T. (1946). Biochem. J. 40, 481.Mann, T. (1948a). J. agric. Sci. 38, 323.Mann, T. (1948 b). Lancet, 254, 446.Mann, T. (1949). Advanc. Enzymol. 9, 329.Mann, T. (1951 a). Biochem. J. 48, 386.Mann, T. (1951b). Biochem. Soc. Symp. 7, 11.Pethica, B. A. & Schulman, J. H. (1953). Biochem. J. 53,

177.Roe, J. H. (1934). J. biol. Chem. 107, 15.Salton, M. R. J. (1951). J. gen. Microbiol. 5, 391.Schwartz, A. M. & Perry, J. W. (1949). Surface ActiveAgent& New York and London: Interscience Publishers.

Walton, A. (1946). Notes on Artificial Insemination ofSheep, Cattle and Hor8e8. London: Holborn SurgicalInstrument Co.

Two Methods for the Determination of Glycogen in Liver

BY J. VAN DER VIESDepartment of Pharmacological Re8earch, N. V. Organon, 088 (Holland)

(Received 16 December 1953)

Several methods for the determination of glycogenin tissues have been described (Good, Kramer &Somogyi, 1933; Morris, 1948; Boettiger, 1946;Seifter, Dayton, Novic & Muntwyler, 1950; vanWagtendonk, Simonsen & Hackett, 1946; van derKleij, 1951). As we needed a simple technique forthe study of the glycogenetic properties of adrenalsteroids a survey was made of several techniquesoften used (see Table 1). For serial determinationsthe methods of Seifter et al. and van der Kleij areattractive, in part because the time-consumingseparation of the glycogen by precipitation withalcohol is avoided. However, in Seifter's methodthis separation can only be omitted if the glycogencontent of the liver exceeds 1 %, while van derKleij's technique is not specific for glycogen, sincesubstances like glucose and glucose phosphates,which are known to occur in liver, are also estimated.The results submitted in this paper show that the

process of extracting the glycogen with trichloro-acetic acid as proposed by van der Kleij can be usedin combination with colorimetric techniques for theestimation of glycogen as described by van Wag-tendonk et al. (with iodine) and Morris (with

anthrone). In this waytwo methodswere developed,which can be used for the serial determination ofglycogen in liver. In particular the application ofaniodine reagent proved to be very convenient forroutine determinations of glycogen. In our hands,the use of this method instead of the more elaboratetechniques of Pfluger (1905) and others led to aconsiderable saving of time and materials. Satis-factory results have been obtained, despite somelimitations ofthe method, which will be discussed indetail.

REAGENTS AND METHODS

Extraction of glycogen from liver withtrichloroacetic acid

If the iodine reagent is used for the determination ofglycogen in the extracts the concentration of trichloroaceticacid (TCA) should not exceed the limits 4-9-5-1% (w/v).This is checked by titration with 0-1N alkali, and thesolution adjusted if necessary by adding water or a concen-trated solution of TCA.

Procedure. A 200 mg. sample of liver is weighed on atorsion balance and finely ground with 20 ml. of 5% TCA

DETERMINATION OF GLYCOGENin a mortar or preferably in a homogenizer (Potter &

Elvehjem, 1936). The precipitate of proteins is filtered offand the clear filtrate submitted to analysis.

E8tinmation of glycogen in the extract8

(a) With iodineIodine reagent: 16-5 ml. of Lugol's solution, prepared by

dissolving 1 g. of iodine and 2 g. ofKI in 20 ml. of water, isadded to 990 ml. of an aqueous solution, containing 25%(w/v) of KCI.

Procedure. In a colorimeter tube (1.2 cm. diameter) 2 ml.of a liver extract is added to 3 ml. of iodine reagent. Aftermixing, the optical density is read in a photometer at650 mu. against a blank, obtained by adding 2 ml. of 5%TCA to 3 ml. of reagent in the same way. The amount ofglycogen is read from a calibration curve, which must beprepared using glycogen of the same origin as the sample tobe estimated.

(b) With anthroneAnthrone reagent: 0*5 g. of anthrone is dissolved in

250 ml. of 95% (w/v) H2SO4. As this reagent deterioratesrapidly it should be freshly prepared before each series ofdeterminations.

Procedure. 2 ml. of the liver extract are pipetted intoa test tube calibrated at 20 ml., 2 ml. of ION-KOH are

added and the tube placed in boiling water for 1 hr. Aftercooling, 1 ml. of glacial acetic acid is added to neutralize theexcess of alkali, and the fluid brought up to the mark withwater: 2 ml. of the solution are added slowly to 4 ml. ofanthrone reagent in a test tube, which is placed in coldwater to prevent excessive heating. After mixing by lateralshaking, the tube is placed in boiling water for exactly10 min. for colour development and cooled with tap water.The optical density is read within 2 hr. in a photometer at650 m,u. against a blank, which is prepared by submitting2 ml. of5% TCA to the same procedure. For the calculationofthe amount ofglycogen, glucose standards are prepared ineach series as follows: 2 ml. ofTCA are heated to 1000 with2 ml. of 1ON-KOH and the solution is neutralized with 1 ml.of glacial acetic acid as described above. Then 4 ml. of a

solution containing glucose (0-15 and 0-60 mg.) are added.After dilution to 20 ml. with water the procedure is con-

tinued in the usual way.

Preparation of glycogenThe livers (or muscles) of 10-20 rats were quickly removed

and transferred as rapidly as possible to 300 ml. of 30%(w/v) NaOH at 1000. Heating was continued for 2 hr. andthe crude glycogen precipitated by adding twice the volumeof 96% (v/v) ethanol. After separation, the precipitate wasdissolved in a small volume of water and the solutionadjusted to pH 3 with dilute HCI. The glycogen was re-

precipitated by the addition of an equal volume of ethanol.This process of precipitation was repeated once more, andthe resulting substance washed with ethanol and ether andfinally dried in vacuo. The purity ofthe glycogen determinedafter conversion into glucose usually amounted to 90-95%.For these estimations glucose was determined with theHagedorn & Jensen method. In all experiments opticaldensities were measured in a Lumetron photometer (model402 E) provided with monochromatic filters.

. C

.,-~b -a

0

0

z

0 CQ; g E]

o"

0~~~

beD

6 0~~~

0

Q

el

0

.E0D

I'd

,0

eJ e c

ce=~U 0 s m -B

*^~ ~ ~ ~ ~~b

r ° 0 Q z { | P

04

e'

CS

aC)~0.0

0 ~~~ ~ ~ ~~~~~~0~~~

0Z -,5t. 1.

0

p

'd( 4N

0r-4 C)

04

P3 a

(1-~It

Ca

biDo00 0

4-

0 ~ ~ ~ ~

4Q 4.'Caa~p~

Vol. 57 411

J. VAN DER VIES

RESULTS

Estimation with iodine reagentFig. 1 shows the relationship between the opticaldensity of a glycogen-iodine complex and theglucose equivalent of glycogen at various wave-lengths. This diagram was constructed with aglycogen sample obtained from rat liver. As theserelationships can be represented by straight lines theoptical density can be read at different wavelengths.In practice we adopted 650 m,u. as the most suitableone for estimation of glycogen over a wide rangeof concentrations without necessitating dilution ofglycogen-rich liver extracts. The colour intensity ofthe glycogen-iodine complex decreases with in-crease in the concentration of TCA in the extract,but within the limits of 4 9-5 1 % TCA, variationsin optical density due to different acidity can beneglected.

Glycogen is very stable in the extracts. Thedifferences between estimations of thirty-sevensamples from various livers, performed immediatelyafter extraction of the tissues and 18 hr. later,respectively, averaged + 0-01 0/ + 0 03, the greatestdeviation being 0-08 %. The extracts can thereforebe left overnight, which adds to the usefulness oftheTCA method.The reproducibility of the technique was in-

vestigated by performing duplicate determinationson the same liver. In view of the irregular distribu-tion of glycogen in liver found by Nijs, Aubert & deDuve (1949), adjacent samples of tissue were takenfor analysis. In eighteen experiments a meandifference of 0-06 + 0-068 in percentage liver glyco-gen was found between duplicate samples. Similar

515 m,.0700 X 490 mj.

0-600 //

0.500 4A~ ~ ~ ~ / 5750 mi.

C 0-400 -

.Li 0-300OF ,C /5~~~~~~~6262 mA

00-200 640 m,u0Z100 ~~~~~~~0 ~660 m,u.0.100

000625 0-125 0*250

Glucose equivalent (mg./ml.)

Fig. 1. Relations between optical density of a glycogen-iodine complex and glucose equivalent of glycogen atvarious wavelengths. A homogenous sample of glycogenfrom the livers of rats was used. The glucose equivalentsindicated refer to 1 ml. of iodine-stained solution.

liver samples were analysed using the more elaboratePfluger method; in eleven experiments a meandifference of 0-33 + 0-225 was found, so with theiodine method more reproducible results are ob-tained than with Pfluger's technique._ The calibration curves relating optical density toglucose equivalent have a different slope withglycogen samples of different origin (Fig. 2); thisemphasizes the necessity for a standard curve,obtained with glycogen from the same source as theglycogen samples which are estimated. The magni-tude of this slope may be used to characterizeglycogen for comparative purposes and is termedhere the 'colour intensity', being related to theextinction coefficient. Using the data of Fig. 2,values of colour intensity for samples of glycogenfrom mouse liver, rat liver and rat muscle werecalculated to be 1- 1, 1 00 and 0.81, respectively. Insubsequent experiments the colour intensity ofa glycogen was determined by submitting a solutionto two determinations, one by iodine-staining andthe other by anthrone.

It was found that differences in colour intensityexist between samples of glycogen prepared fromthe livers of different individuals of one species(rats). The colour intensities were estimated atvarious wavelengths for five samples of glycogenprepared from different livers (Table 2). Deter-minations in triplicate were performed on one

/A0-700-

0-600

0-500

0-400

03000

0-200 -

.01004100X I ~~~~~~~~~~~~I I

0 0-1 0-2 0:3 0 4 0-5 0-6Glutcose equivalent (mg./ml.)

Fig. 2. Calibration curves for glycogen from differentsources. Optical densities were read at 650mIt. Theglucose equivalents indicated on the abscissa refer to1 ml. of iodine-stained solution and were calculated fromthe estimated purities of the glycogen preparations.A, mouse liver; regression coefficient of the line, 1l51.B, rat liver; regression coefficient, 1-00. C, rat muscle;regression coefficient, 0 81.

I954412

DETERMINATION OF GLYCOGENhomogeneous sample ofglycogen as well. The varia-tions in the values for glycogen from different livers(standard deviations in upper part of the table)exceed significantly those solely due to technicalerrors (standard deviations in lower part of thetable). This was proved statistically by computinga mean square for each part of the table by poolingthe standard deviations according to the formula(E82)/n. The ratio of these mean squares was foundto be 4*36 and indicates a significant difference withP <0O001. Hence, colour intensities of glycogendiffer from one liver to another.As the differences between the samples of

glycogen existed at all wavelengths, only colourintensities at 650 mp. were compared in thefollowing experiment. After preliminary starva-tion, rats were dosed with glucose and the colourintensities of liver glycogen compared with those ofnormally fed rats (Table 3). The results suggest thatfollowing a treatment which effects a rapid deposi-tion of glycogen in the liver the colour intensity ofthe glycogen increases, but independently of theactual amount of glycogen in the liver. Thissuggests that newly formed glycogen has thegreater colour intensity.

Estimation with anthrone reagentThe estimation of glycogen in liver extracts by

means of the anthrone reagent includes a simul-taneous hydrolysis of glycogen to glucose, so thisestimation is not influenced by the origin of theglycogen. Treatment ofliver extracts with hot alkaliresulted in an elimination of interfering glucose,without affecting the glycogen (Table 4); thisenhances the specifiaty of the method.A liver incubated at 370 in Krebs-Ringer phosphate

buffer is a suitable object to demonstrate the estimation ofglycogen in the presence of increased amounts of physio-logical degradation products, such as can occur in thesystem to be studied. Experiments were performed witheight livers, incubated at 370 for 15 min. The livers were

extracted by grinding 500 mg. of tissue with 30 ml. of TCA.In order to hydrolyse the glycogen, 10 ml. of extract were

heated with 0-5 ml. of conc. HCl for 3 hr. at 100°. Afterneutralization of the acid with alkali and readjusting thequantity of fluid to 10 ml. with water, 10 ml. of 10% TCAwere added, giving a final concentration of 5 %, as the TCAoriginally present was decomposed by boiling with hydro-chloric acid. A comparable solution was obtained by mixingthe components in a sequence which prevented the hydro-lysis of the glycogen. In the solutions, glycogen was esti-mated both including and excluding a treatment with alkali

Table 2. Colour inten8ities at various wavelengths for glycogen from the livers of individual rats

Wavelength (mP.)465 490 515 550 575 595 620 640 660

Colour intensities,' ~~~~~~~~~A

Meani S.D. of glycogen preparationsfrom the livers of five different rats

Mean+s.D. triplicate determinationson one glycogen preparation

2-12+0-232-63

+0-29

2-53+0-322-78

+007

2-46±0-24

2*85+0-06

2-24 1 61±0-24 ±0-21

2-30 1-64±0-02 +0-05

1*32+0-251-30

+0-01

1-06+0-15

1-07+0-01

0-74+0-110-73

±002

0-68+0090-63

±0t02

Table 3. Colour intensities at 650 mp. for glycogen from the livers of individual rats

Normally fed rats Glucose-treated rats*

Liver glycogen(% of fresh wt.)

1-762-425-761-345*05

Av. 3-27

Colourintensity

0-590-890.950-800-870-82

Liver glycogen(% of fresh wt.)

3501-361-251-68

1*95

Colourintensity

1-381-521-801-85

1-64* After fasting for one night each animal received 7 hourly doses of 0-5 g. of glucose subcutaneously. The mean

difference between the colour intensities is statistically significant (t7=5-21, P<0-01).

Table 4. Influence of treatment with alkali on gluCose and glycogen estimation with anthrone

Time of treatment with alkali at 1000 (min.)

0No. of R

observations ,Glucose (0-60 mg./ml.) 4 100Glycogen (glucose equivalent 0-32 mg./ml.) 2 100The experiments were performed as indicated in the Methods section.

15 30 45 60 90lecovery of glucose or glycogen (%)

4 3 4 3 3104 104 103 99 101

VoI. 57 413

J. VAN DER VIES

for 1 hr. at 1000 (Table 5). By omitting the treatment withalkali both glycogen and glucose are estimated. Afterprevious hydrolysis of the glycogen the same results areobtained. Hydrolysis of glycogen with subsequent alkalitreatment eliminates both glycogen and glucose and con-sequentlyno carbohydrate is found. Finally, treatment withalkali only destroys glucose, leaving the glycogen intact,which results in lower figures than without alkali treatment.

For routine determinations treatment withalkali for 1 hr. at 1000 was adopted.. After addingknown amounts of glycogen to liver extracts theglycogen was fully recovered (Table 6).

Compari8on ofmethod8for e8timation ofglycogen inliver and mu8cle. The results of the method werecompared with those of the technique described byGood et al. (1933). Two adjacent samples were takenfrom several livers, one sample extracted withTCA and the other one submitted to analysisaccording to Good et al. In the extracts glycogenwas estimated with the anthrone reagent. The sameexperiments were performed with muscles of rats.To extract the glycogen from muscle, a 500 mg.sample was ground finely in a mortar with about1 g. of purified sand and 10 ml. of 5% TCA. Thesand and proteins were filtered off and the clearfiltrate was used for analysis.The data in Table 7 indicate a satisfactory

agreement for liver but not for muscle, suggestingan incomplete extraction of glycogen from thelatter. This assumption was tested as follows:Two samples were taken from the same muscle;

one extracted with TCA and the other one heatedwith alkali until a homogenous solution was ob-tained. From this solution the glycogen was pre-cipitated by adding twice the volume ofethanol andthe precipitate separated. The glycogen was dis-solved in 10 ml. of TCA, yielding a solution com-parable with that obtained by direct extraction ofmuscle. In the two extracts glycogen was estimatedwith the anthrone reagent (Table 8). Lower figureswere obtained when the muscle was directly ex-tracted with TCA than after liberation of the glyco-gen by destruction of the tissue with alkali. Soapparently a considerable part of the glycogen inmuscle is not extracted with TCA. Subsequentextractions of the protein residue with fresh TCAdid not increase the yield ofglycogen. However, theunextracted glycogen could be recovered afterdestruction of the protein residue with alkali.

In a similar series of experiments both theextractable and non-extractable part ofthe glycogenwere estimated in the livers of rats, submitted tovarious treatments. The non-extractable glycogenleft in the protein residue of the tissue after extrac-tion with TCA was liberated by alkali destructionand separated by ethanol precipitation. In somecases the colour intensities of the glycogen-iodinecomplexes were determined as well. All results are

Table 5. Influence of treatment with alkali on gluco8eand glycogen in extracts of incubated liver8

Results* (as % glucose equivalent of fresh wt. of liver)

Acid hydrolysis previously

Including WithoutKOH KOH

treatment treatment0-01±d0022 1*01 ±0-424

No acid hydrolysis

IncludingKOH

treatment0-33+0-328

WithoutKOH

treatment0-99±0-471

* Each result is the average of eight experiments.

Table 6. Recovery of glycogen added to liver extract8

Estimation with anthrone reagent.

Addedglycogen(mg./ml.)0-1590-2200-2720-3160-3560*389

Recoveredglycogen(mg./ml.)0-1570-2210-2800-3130-3660*382

Recovery(%)98-8100-3102-999.1102-998-4

Av. 100-4

Table 7. Agreement between e8timation8 with theanthrone reagent and the method of Good et al. (1933)

Glycogen(Mean % fresh wt. tissue)

With Withanthrone Good'sreagent method

Liver (10) 2-72 (10) 2-51Muscle (12) 0 34 (12) 0-52

Meandifference

+0-21 ±0-227- 0-18±0-050

The figures in parentheses give the numbers of observa-tions. For the mean values of the glycogen amounts inliver and muscle no standard deviations have been calcu-lated as the latter are only indicative of the variabilitybetween individual animals. For the comparison of thetwo methods only the differences between the data ob-tained with both methods in the same organ are relevant.The relatively greatest mean difference between the esti-mations in muscle is highly significant (tll= 12-5, P < 0-001).

Table 8. Determinations of glycogen in muscle ofrats u8ing trichloroacetic acid extraction or alkali-dedtruction

Glycogen found(Mean % fresh wt.)

No. ofobservations

18

Glycogenextractedwith TCA

0-42

Glycogenliberated bydestructionof the tissuewith alkali

0-61

Meandifference0-19±0-126

Statistical significance of mean difference t47 = 6-37,P<0-001.

414 I954

DETERMINATION OF GLYCOGEN

Table 9. Extractable and non-extractable glycogen in the liver8 of rat8 and mice after various treatment8

Expt.1 Normal rats, not starved

2 Normal rats, starved for 24 hr.

3 Adrenalectomized rats, starvedfor 24 hr.

4 Adrenalectomized rats, starved for24 hr. and treated with cortisonet

5 Adrenalectomized mice, starved for24 hr. and treated with glucose:

6 Adrenalectomized mice, starved for24 hr. and treated with cortisoneand glucose$

Glycogen fractionExtractableNon-extractableExtractableNon-extractableExtractableNon-extractableExtractableNon-extractableExtractableNon-extractableExtractableNon-extractable

No. of Glycogenobservations (% of fresh wt.)

10 4-71±2-220-16±0-021

10 0-04±0-024- 0-03±0-0165 0-06±0-045

0-01±0-0075 0-63±0-240

0-08±0-0505 0-46±0-274

0-10±0-0495 1-98±0-541- 0-13±0-32

* Statistical significance of differences between colour intensities of extractable and non-extractable glycogen fractions:Expt. 1: t9 =7-72, P < 0-001; Expt. 2: t7=3-26, P<0-05; Expt. 3: t7=9-50, P<0-00l.

t After fasting for 16 hr. each rat received 5 doses of 0-18 mg. cortisone acetate in peanut oil suboutaneously.t Treatment with glucose and cortisone as described by Venning, Kazmin & Bell (1946). Total doses for each animal:

cortisone acetate 0-08 mg., glucose 35 mg.

~0

0-cE

-v

bo0oZ-,

bo0

t1-

0*

-1

-1 u +1-

log % glycogen (anthrone method)Fig. 3. Parallelism between the results of the anthrone and

the iodine method. The logarithms of the percentageglycogen found with the two methods are plotted againsteach other.

summarized in Table 9. In the livers of normalanimals the non-extractable fraction is only a small

part of the total glycogen. The colour intensitieswith iodine at 650 mi. of the non-extractableglycogen fractions are less than those ofthe extract-able parts. Finally it can be seen that alterations inthe glycogen content of liver, caused by starvationor treatment with cortisone acetate, are mainlypresent in the extractable fraction of the glycogen.

Comparison of anthrone and iodine methods. Inmany cases, the determination of the extractableglycogen will furnish sufficient information aboutthe glycogen contents of the liver, and then the

TCA extraction procedure may be adopted, withsubsequent estimation by the iodine or anthronemethod. A comparison between the anthrone andthe iodine method was made using extracts ob-tained during a glycogenetic test.A close parallelismwas found between both series of results as can beseen from Fig. 3 and the final results of the testproved to be nearly the same, irrespective of themethod used for the estimation of glycogen. It maybe noted that in the graph logarithms of the per-centage are plotted. It could be shown that thelogarithms of liver glycogen values in rats have anormal frequency distribution. So the logarithmrather than the actual percentage should be usedfor the statistical calculation of a test.

DISCUSSION

In the foregoing the usefulness of the iodine andanthrone methods has been demonstrated. Bothmethods are simple and much more appropriate forserial determinations than all other methodspreviously described. Nevertheless they have theirlimitations. The TCA extraction, which is used forboth methods, is very convenient but a considerablepart of the glycogen in the muscles of rats isnot extracted with TCA. Consequently the TCAextraction cannot be recommended for the estima-tion of glycogen in muscle.

In liver the non-extractable glycogen fractionproved to be only a minor part of the total glycogenand alterations in the amount of glycogen in thistissue were found to be mainly associated with theextractable glycogen. For many purposes thedetermination of the extractable glycogen will givesufficient information about the amount of glyco-gen in a liver. The suitability of the extraction

Colourintensity*0-93±0-0850-58±0-121

0-87±0-2240-50±0-1741-26±0-2630-64±0-131

I~~~~ *,

z --! I -I' I- I

I-- - -- -

-11 A -11

VoI. 57 415

416 J. VAN DER VIES I954procedure should be determined for other organs. Ifone wants to be sure that 100% of the glycogenpresent in any organ is determined, the usualhydrolysis of the tissue with strong alkali followedby precipitation ofthe glycogen remains the methodof choice. Using the iodine reagent only approxi-mate figures are obtained for the glycogen content ofa liver extract. Nevertheless, this very convenientmethod may be useful for investigations where someaccuracy may be sacrificed in order to make possibleexperiments with a large series of rats. For instancethis was the case in our studies of the glycogeneticproperties of adrenal steroids by estimation of liverglycogen, in which the experimental animals aretreated following a standard procedure, and theresults are always mutually compared. The resultsof these tests were hardly influenced by the variablecolour intensities of the glycogen from differentlivers. However, if approximate values cannot beaccepted the use of the anthrone method is in-dicated. Providing the lower saccharides arepreviously destroyed with alkali, more correctresults are obtained, but the method is moreelaborate than the estimation with iodine.

SUMMARY

1. Two methods are described for the determina-tion ofglycogen. In both procedures the glycogen isextracted from the tissue with trichloroacetic acid.

2. In these extracts glycogen can be estimatedapproximately by means of an iodine reagent. The

colour intensity of the glycogen-iodine complexvaries from one liver to another and is further in-fluenced by the treatment of the animals.

3. More correct results are obtained if theglycogen in the extract is determined using ananthrone reagent, after treating the liver extractswith alkali which eliminates interfering glucose.

4. The application of the methods is brieflydiscussed.

I wish to express appreciation and thanks to Dr G. A.Overbeek for his advice and encouragement given during thecourse of this work and to Miss C. W. N. N. Schellekens andMr J. J. de Koning for valuable technical assistance.

REFERENCES

Boettiger, E. G. (1946). J. cell. comp. Phy8iol. 27, 1.Good, C. A., Kramer, H. & Somogyi, M. (1933). J. biol.Chem. 100, 485.

Kleij, B. J. van der (1951). Biochim. biophy8. Acta, 7,481.

Morris, D. L. (1948). Science, 187, 254.Nijs, A., Aubert, X. & Duve, C. de (1949). Biochem. J. 45,

245.Pfluger, E. F. W. (1905). Da8 Glykogen, Bonn. p. 53.Potter, V. R. & Elvehjem, C. A. (1936). J. biol. Chem. 114,

495.Seifter, S., Dayton, S., Novic, B. & Muntwyler, E. (1950).

Arch. Biochem. 25, 191.Venning, E. H., Kazmin, V. E. & Bell, J. C. (1946). Endo-

crinology, 38, 79.Wagtendonk, W. J. van, Simonsen, D. H. & Hackett, P. L.

(1946). J. biol. Chem. 163, 301.

The Use of Substances Depressing Antithrombin Activityin the Assay of Prothrombin

BY P. FANTLBaker Medical Research In8titute, Alfred Hospital, Melbourne, Australia

(Received 12 September 1953)

Determination of the prothrombin activity of bodyfluids has become of clinical interest in connexionwith the diagnosis of haemorrhagic disorders, andduring treatment with anticoagulants. Mosttechniques for the determination of prothrombinare based on one or other oftwo principles. In both,prothrombin is converted into thrombin in thepresence of optimum amounts of the other clottingfactors. In one, the time interval required for clotformation in the specimen is taken as a measure ofprothrombin activity; this is a direct method, andis referred to in the literature as the one-stagetechnique (Quick, 1942). In the other, the thrombinpotency developed is estimated with an added

substrate containing fibrinogen; this is an indirectmethod often called the two-stage technique(Warner, Brinkhous & Smith, 1936).In the direct method a fibrin clot appears at a

certain time corresponding to the formation of acertain concentration of thrombin. The rate ofthrombin formation depends upon the concentra-tion ofprothrombin and the activity of acceleratingand inhibiting factors. The direct method is usefulin determining the haemostatic condition of apatient, but it does not indicate the actual pro-thrombin concentration of the specimen.

In the direct method, fibrin formation takes placeat a very low thrombin concentration, which is