68 THE VITAMINS 



characteristic appearance of the gold double salt obtained by Jansen 

 and Donath. When the solution had been concentrated to about 1.5 

 cubic centimeters the precipitate was washed with absolute alcohol, and 

 both the aqueous suspension of the precipitate and the filtrate were 

 decomposed with hydrogen sulfide. The resulting solutions were evapo- 

 rated to dryness in vacuo over soda lime. In doses corresponding to 

 36 grams of embryo, the insoluble gold salt fraction produced but 

 subnormal growth, while the soluble gold fraction was inactive. 

 Administering these two fractions in combination, however, revealed 

 a supplementary action both in rat-growth and pigeon-curative 

 tests. This would suggest that the action of vitamin Bi is probably 

 to be ascribed to more than one factor. The gold precipitate would 

 appear to contain most of one factor and a little of the other, whereas 

 the soluble fraction would appear to contain the second factor almost 

 exclusively. 



The active picrolonic filtrate gave a very strong Pauly reaction 

 comparable with that of histidine. Heated with 40 per cent sodium 

 hydroxide this fraction emitted a strong smell suggestive of alkylamines, 

 but no smell of skatole or indole (Cf. Levene, 1928). No sulfur reaction 

 with lead acetate was obtained. Millon's, xanthoproteic, and purine re- 

 actions were negative, which would appear to exclude the possibility 

 of the vitamin being a hydroxyphenyl or purine derivative. The plati- 

 num fraction and the preparation from the gold precipitate also pro- 

 duced a very strong Pauly reaction, whereas the gold filtrate gave a 

 very faint one. The gold precipitate produced a very faint coloration 

 with Folin and Denis' reagent. Using the preparation obtained by 

 Jansen and Donath, Eijkman (1927) had observed a blue coloration 

 with this reagent, which, however, was stated to be weaker than that 

 produced by uric acid of the same concentration. These authors indicate 

 the importance of this observation in view of statements in the litera- 

 ture claiming Folin and Denis' reagent a valuable guide in the detection 

 of vitamin B. 



In a discussion of their observations, Guha and Drummond say: 

 "As regards the properties of vitamin Bi, its behavior appears to be 

 determined to an extraordinary extent by the presence of other sub- 

 stances and by the previous treatment of a given preparation, a fact 

 pointed out by Kinnersley and Peters (1928). These investigators 

 observed that the solubility of their curative preparations from yeast 

 in alcohol varied with their activity and with the hydrogen-ion concen- 

 tration of the solutions treated with alcohol. Similar variability has been 

 observed with fractionation by silver nitrate. In our first method of 



