April 15, 1941 ANAL.

YTICAL EDITION 225

constituents that are claimed to give the color by Kuhn and

LijW.
Swaminathan has also used the phenol reagent of Folin and
Ciocalteau and by the use of either of these two methods, he
has obtained the vitamin B6 content of various foodstuffs.
In Table I1 are listed his values as well as those obtained in
HO H 2 C f i I i
the authors’ laboratory using a new bioassay procedure for
vitamin Bg.
Literature Cited
-%mold, A., Schreffler, C. B., and Lipsius, S. T., IND.E m .
CHEM.,Anal. Ed., 13, 62 (1941).
Bandier, E., Biuchem. J., 33, 1130 (1939).
TABLE 11. VITAMINBs CONTENT
O F FOODSTUFFS Bandier. E.. and Hald. J.. Ibid.. 33, 264 (1939).
Henderson, Waisman,
Euler, H. von, Schlenk, F., Heiwinkel. H., and Hogberg, B.,
Foodstuff and Elvehjem Swaminathan 2. physiol. Chem., 256, 208 (1938).
Brewer’s yeast 40 54
Gibbs, H. D., J . Biol. Chem., 72, 649 (1927).
Autoclaved brewer’s yeast 40 53 Harris. L. J.. and Raymond, W. D., Chemistry Industry, 58, 652
Rice polishings ... 13.4 (1939).
Sheep liver 12.6 13.4 Karrer, P., and Keller, H., H e h . Chim. Acta, 21, 463 (1938);
Sheep muscle 11.7 4.5
Cow’s milk .4.8
.. 1.7 22, 1292 (1939).
Yellow corn 7.1 Kodicek, E., Biochem. J., 34, 712, 722 (1940).
Whole wheat ... 7.4 Konig, W., J . prakt. Chem., 69, 105 (1904) n. f .
Polished rice ... 3.0
Krinestad. H.. and Naess. T., 2. _vhwsiol.
_
Soybean ... 8.0 Chem.. 260. 108
Cabbaeo ... 3.1 (1939).
Beet root 1.3 Kiihn, R., and Low, I., Ber., 72, 1453 (1940).
...
..I

Pork liver 9.9

Pork ham 24.0 ... Melnick, D., and Field, H., J . B i d . Chem., 134, 1 (1940); 135,
Veal muscle
Pork heart
16.3
15.5
...
... 53 (1940).
Pearson, P., Ibid., 129, 491 (1939).
Perlzweig, W-.A., Levy, E. D., and Sarett, H., Ibid., 133, lxxiv
(1940).
Reitzenstein, F., J . prakt. Chem., 68, 251 (1903).
The liberation of vitamin Bs from naturally occurring ma- Ritsert, K., Klin. Wochschr., 18, 934 (1939).
terials depends on a pepsin digest followed by tungstate pre- Rubicon Co., Philadelphia, “Notes on Operation of Evelyn
cipitation for removal of proteins and silver nitrate precipita- Photoelectric Colorimeter”.
Scudi, J. V., Koones, H . F., and Keresztesy, J. C., Proc. SOC.
tion for the removal of purines, pyrimidines, and amidazole Ezptl. Bid. Med., 43, 118 (1940).
bases. The vitamin was absorbed, then eluted and further Shaw, G . E., and McDonald, 0. A., Quart. J . Pharm. P h a r m -
treated to remove interfering materials. The pH was ad- coZ., 11, 380 (1938).
justed to 7 and the color developed with the diazotized sul- Swaminathan, M., Indian J . Med. Science, 26, 427 (1938).
Swaminathan. M., Nature. 145, 780 (1940).
fanilic acid. This method is open to criticism since the reac- (22j Vongerichten, E., Ber., 32, 2571 (1899).
tion is not specific. However, the adsorption on an alkaline (23) Waisman, H. A., Henderson, L. M., and Elvehjem, C. A., un-
earth does minimize the carrying over of certain other cell published data.

Chemical Methods for Determination of

Vitamin C
C. G. KING
University of Pittsburgh, Pittsburgh, Penna.

N EARLY all the methods that have been proposed for

the estimation of vitamin C have been based upon the
reversible oxidation of ascorbic acid to dehydroascorbic acid.
gation concerning vitamin C, both in foodstuffs and in physio-
logical research, analyses for dehydroascorbic acid are nearly
as important as tests for ascorbic acid itself. This situation
Although several of the oxidation methods give reasonably arises primarily because dehydroascorbic acid is approxi-
satisfactory results, there are pitfalls to be avoided in all the mately as active antiscorbutically as ascorbic acid and
procedures that have been published; hence it is of distinct theories concerning the function of vitamin C in vivo are based
value to have available an independent approach, such as in part on the analyses for dehydroascorbic acid.
that developed by Roe (9), in which ascorbic acid is decom-
posed to furfural for colorimetric estimation. The reaction Indophenol Method
with 2,6-dichlorophenolindophenol in metaphosphoric acid
solution provides the most widely used, and in the author’s For most investigations direct visual titration of ascorbic
experience the best, general basis for ascorbic acid analysis. acid with the dye 2,6-dichlorophenolindophenolin a pH range
The polarigraphic and oxidation-reduction titrimeter methods of 1 to 3.5 gives a reasonably accurate and satisfactory meas-
are of interest as specialized oxidation procedures, but neither ure of the vitamin C content of tissues. The extracts must be
has been developed sufficiently to merit confidence for general properly prepared and close attention must be given to details
use. of the procedure (1, 2, 7 ) . The two most crucial points to be
Figure 1 will serve to outline the basic steps in the watched, apparently, are avoidance of oxidative changes pre-
methods that have been proposed. I n many types of investi- vious to titration, and awareness of the circumstances
226 INDUSTRIAL A N D ENGINEERING CHEMISTRY Vol. 13, No. 4

aA TH I/-\
0 0 slight fading of the dye is then not bothersome

H OH H C-C
OH
I
OH
I
(dye)
-2H
‘Ie and interference by sulfhydryl compounds is
markedly decreased.

A
H O - -C-
)!I
’ A’
\ /
L o
+2H
HO- -C-
,/”=” 0
The indophenol dye that has been used so
widely is not ideal, especially because of its
relatively high oxidation potential, but it is
0 doubtful whether any other oxidizing agent
Ascorbic acid (HzS) Dehydroascorbic acid that has been proposed up to the present

1 fHC1
Furfural COS+
(color with aniline)
IpH>5
Decomposition products-+. g.,
oxalic acid and 1-threonic acid
time possesses any practical advantage. A
few laboratories have reported satisfactory
results with methylene blue as an oxidant.
The reaction in this case is slow, however,
unless there is an accompanying exposure to
FIGURE 1 light. For testing urine samples there have
been encouraging reports based upon precipita-
tion of the vitamin by 2,4dinitrophenylhy-
under which other reducing substances will or will not inter- drazine, followed by quantitative reduction of the nitro groups
fere with the titration. or by hydrolysis, reduction, and conversion to furfural. The
method proposed by Roe (9),based upon furfural formation,
In relation t o the first point, aerobic oxidation through the has the advantage of completely avoiding interference by non-
agency of enzymes (chieffy copper- but to a lesser extent iron- ascorbic acid reducing material such aa sulfhydryl compounds
containing specific proteins) or added copper (picked up from and the 3- or 4-carbon sugar decomposition products, but the
reagents and equipment) can generally be avoided by the use of
2 to 3 per cent metaphosphoric acid as an extractant, if sufEcient procedure is much more time-consuming and necessitates
care is given to other details. InsufEcient grinding and extraction avoidance of possible interfering substances that yield furfural
of the sample have also been common sources of error. The on decomposition, such as the uronic acids and pentoses.
author has not observed evidence of the existence of si nificant I n regard to the use of such special instruments as the oxi-
amounts of combined, non-acid-extracted ascorbic acii as re-
ported by Reedman and McHenry (8) and a few other investi a- dation titrimeter or the polarigraph, the author’s experience
tors. Neither has he found sulfuric acid, as recommended%y has been very limited, but he has not found that they possess
Tressler and associates (6),t o be desirable as an extractant, even any significant advantage over the photoelectric colorimeter
though it affords some economy. Althou h acetic and trichloro- for general use. For meeting special difficulties these instru-
acetic acids were widely used in earlier wo,!r and give good results
with many products, they are not so satisfactsry as metaphos- ments have obvious advantages. The polarigraphic method
phoric acid. The protective action of acids is not simply a H holds promise for detecting and perhaps avoiding errors
phenomenon, because at the same pH, metaphosphoric acicf is that arise from interfering reducing substances that exhibit
distinctly su erior to sulfuric acid and the latter in turn is su- slightly higher or lower reduction potentials, such as charac-
perior to hy%ochloric acid.
In relation to the second point, sulfhydryl compounds and terize the sulfhydryl compounds.
carbohydrate decomposition products are especially likely to in-
terfere. The concentration of interfering substances in most Tests for Dehydroascorbic Acid
natural products is not sufficient to cause significant errors, I n studying ascorbic acid synthesis in tissue slices, Dr.
however. When making direct visual titrations it is very im-
portant to use an end point of about 5 seconds or less. With the Smythe in this laboratory observed that certain added sub-
hotoelectric colorimeter method, two or more readings should strates gave reactions with hydrogen sulfide that were com-
g e made at 15-second intervals, from which one can in part cor-
rect for the interference due t o more slowly reacting substances.
parable to the reaction generally used for estimating dehydro-
Thiosulfate, ferrous salts, and certain sulfhydryl compounds cause ascorbic acid. The study was extended to establish the fact
interference when present, as do a number of sugar decomposition that many organic compounds give such a reaction. It is
products having a reductone type of structure. reasonably clear that many of the published data relative to
the oxidized form of the vitamin have been misinterpreted.
I n some natural products such as strawberries and beets because of this interference by newly formed reducing mate-
the presence of coloring matter makes i t nearly impossible rial that results from hydrogen sulfide treatment. The risk of
to carry out direct titrations. In such cases the use of a photo- encountering such an error was pointed out in an earlier pub-
electric colorimeter is satisfactory, however, because one lication (4). The error that is introduced by hydrogen sulfide
measures only the decrease in color caused by ascorbic acid, treatment is also inherent in the modified methods in which
when the dye is added, and the natural pigment which masks mercuric salts are added to remove sulfhydryl compounds
the disappearance of the indophenol dye is not reduced during from tissue extracts, because these methods include a final
the test. treatment with hydrogen sulfide.
Table I lists some of the compounds that were tested and
Bessey has described the details of procedure to be followed in found to interfere seriously with the dehydroascorbic acid
using the instrument (1) and has cited man related publications. determination, together with a number of compounds that did
The papers by Mindlin and Butler (6) an8Farrner and Abt (9) not cause serious interference under the conditions studied.
are of special interest in relation to blood analyses. It is now I n several cases the interference was almost molefor-mole
generally agreed that cyanide should not be used and that as- equivalent to the reversibly oxidized vitamin. During a
corbic acid is fairly stable in whole blood, plasma, and serum
except when hemolysis has occurred. There may be a slight loss considerable part of the titration, the reaction of the newly
caused by oxidation arising from ferri forms of compounds and formed sulfhydryl compounds with the dye was apparently
loosely bound oxygen carried by hemoglobin. Bessey has pointed complete within 1 to 5 seconds. Toward the end of the titra-
out the value of using an extractant buf€ered at pH 3.5 for titra- tion there was generally a slower fading of the dye that r e
tions with the instrument, since in that range the reaction with the
d g is reasonably specific, the reading does not drift because of a sembled the typical glutathione reaction. Because of the
effect on the dye, and conditions are about optimum for de- rapidity of the reaction the interference could not be wholly
gydroascorbic acid analyses. The same author has pointed out avoided either by a correction curve with the photoelectric
the need for a check readin after decolorizing all of the dye, in colorimeter or by visual titration a t a pH of 1 or less. The
order t o correct for light a%sorption by suspended matter and
other material in the test preparation. In making direct visual best direct methods for detecting and avoiding such an error
titrations one can profitably use a lower pH (about 1) because the would seem to be the use of the furfural reaction, the use of a
April 15, 1941 ANALYTICAL EDITION 227

polarigraph, or to decompose the dehydroascorbic acid a t a hydes, ketones, and quinones give rise to an interfering
pH above 7 and note the resultant change in value after reaction when reduced by hydrogen sulfide. Among the in-
hydrogen sulfide reduction. terfering compounds are pyruvic acid, pyruvic aldehyde,
glyceric aldehyde, dihydroxyacetone, acetaldehyde, manno-
saccharic acid, 5-ketogluconic acid, 1,4-benzoquinone, 1,4-
TABLEI. DEHYDROASCORBIC
ACID TESTS naphthoquinone, and 2-methyl-1,4-naphthoquinone(thylo-
( H t S for 2 hours at pH 3.5; titration with 2,6-dichlorophenolindophenol) quinone). Three methods for detecting and avoiding such
Positive Negative interference are pointed out.
(Interference > 20%) (Little or No Interference)
Pyruvic acid Acetone Piperitone
P ruvic aldehyde Sorbose Phorone Acknowledgment
& ceric aldehyde
ydroxyacetone
Acetaldehyde
Levulose
Glucose
Xylose
Oxalic acid
Kojic acid
8-Ketobutyric acid The author is indebted to the Buhl Foundation for a re-
Mannosnccharic acid Cyclohexanone @-Ketoglutaricacid search grant in support of investigations in this field of study.
bKetogluconic acid Hexadienal Glucuronic acid
14-Benno uinone Butyraldehyde
1:4-Naph%oquinone
2-Methylquinone (vitamin IC) Literature Cited
Bessey, 0. A., J. Biol. Chem., 126, 771 (1938); J . Am. Med.
Assoc., 111, 1290 (1936).
Summary Bessey, 0 . A., and King, C. G., J. Biol. Chem., 103,687 (1933).
Farmer, C. J., and Abt, A. F., Proc. SOC.Exptl. Biol. Med., 34,
When adequate precautions are taken, the reaction of as- .----,-
146 (1936).
K i i g C. G.,Physiol. Reo., 16,238 (1936).
corbic acid with 2,6-dichlorophenolindophenolcan be used in Mack, G.L.and Treader, D. K., J. Biol. Chem., 118,735 (1937).
direct titrations or with the photoelectric colorimeter to give Mindlin, R. L.,and Butler, A. M., Zbid., 122,673 (1938).
relatively satisfactory quantitative analyses. Other methods Musulin, R. R., and King, C. G.,Zbid., 116,409 (1936).
of analysis may be preferable under special circumstances b e Reedman, E.J., and McHenry, E. W.,Biochem. J., 32,85 (1938).
cause of interference by other reducing materials. Roe, J. H.,Science, 80,561(1934);J.B i d . Chem., 116,609 (1936).
Methods for the measurement of dehydroascorbic acid, CONTRIBUTION418 from the Department of Chemiatry, University of
however, are subject to great interference, because many a l d e Pittsburgh.

A Spectroscopic Method for the Quantitative

Estimation of Vitamin D
NICHOLAS A. MILAS, ROBERT HEGGIEI, AND J. ALBERT RAYNOLDS'
Massachusetts Institute of Technology, Cambridge, Mass.

did not improve the spectroscopic values appreciably, since the

D URING the past 4 years three independent physico-
chemical methods (1, 6, 6, 16, 16) have been proposed
for the estimation of vitamin D.
presence in the oils of sterols and possibly carotenoids gives rise
to colors which seem to have an additional effect on the 500 mp
maximum of vitamin D. This made necessary the use of an addi-
tional correction for the presence of sterols. Although such a
The method originally used by Brockmann and Chen has been procedure ave satisfactory results for oils containing more than
more extensively studied and modified by several investigators
(8, 8, 18, IS, 14). This method is based on the discovery that
i.
10,000 U. P. vitamin D units er gram, it was found cumber-
some and not universally ap licaile, as it is necessary to have a
vitamin D forms with a saturated solution of antimony trichloride method of measuring the visihe spectrum accurately and rapidly.
in chloroform a pink color which exhibits a prominent band in
the region of 500 mp. Figure 1 shows the antimony trichloride In view of these objections the authors have adopted a
transmission spectra of vitamins Dg and Da and that of the gluco- modified procedure in which vitamin A, 7-dehydrocholesterol,
side of vitamin Dn. Irrespective of the minor differences in
structure, all transmit exactly in the same region with the head
of the characteristic band located a t 500 mp. In the course of 100
/
their work on the separation of vitamins A and D from fish liver
oils (10, I I ) , the authors followed the purification by measuring 90 ,
the extinction coefficients at 500 and at 620 mp, the two maxima
attributable to vitamins D and A, respectively. By the use of
the appropriate correction factors, it waa found ossible to esti-
mate vitamin D in the presence of considerabE quantities of
vitamin A with a reasonable de ree of accuracy. This method a
has now been extended and mogfied to make possible the esti-
mation of vitamin D in fish liver oils. 0
u)
In preliminary attempts, the authors used as a standard a GLUCOSIDE-CONG.0 0165 X
fish liver oil, the biological potency of which had been accurately O~~CONC~000391!4
determined a number of times by several laboratories and re-
ported to be 50,000 U. s. P. vitamin D units per gram. This 50 D3-CONC '0025EX
oil was kindly su plied by the Atlantic Coast Fisheries Corpora-
tion of New YorE. Usin this standard, the potencies of some
of the oils tested agreed girly well with the reported biological
40
400 500 600
J
700
potencies. However, a correction for the presence of vitamin A WAVE LENGTH, MILLIMICRONS

1 Present address, The American Chicle Company, New York, N. Y. FIG~JRE1. TRANSMISSIONSPECTRAOF ANTI-
1 Present address, The Atlantic Coast Fisheries Corporation, New York, MONYTRICHLORIDE COLOROF VITAMINSDPAND
s.Y. D, AND VITAMIND p G ~ ~ ~ ~ ~ ~