THE ANTIBERIBERI VITAMINE 175 



clear from this work, that Barger (482) was correct in his statement 

 that the substance described by us in 1913, to which we then gave the 

 formula C26H2oN 4 O 9 , was in reality nicotinic acid. However, the 

 view expressed in the Report of the Medical Research Committee 

 (I.e. 333), that the substance isolated was nicotinic acid contami- 

 nated with vitamine is erroneous, since the analysis indicated, at the 

 time, pure nicotinic acid for which no curative action was claimed. 

 Summarizing our work with rice polishings, we were able to differ- 

 entiate the curative substance of 1911, but only when we undertook 

 the extraction of the rice polishings ourselves. When this was done in 

 the factory no curative substance was obtained, and hence no pub- 

 lication was made, thus explaining the non-appearance of the 

 protocols of the animal experiments. Therefore, it is not justifiable 

 to apply the conclusions drawn from our negative results with rice 

 polishings to our positive yeast findings, which we shall describe in 

 the next chapter. 



In conclusion we wish to call attention to an investigation by 

 Hofmeister and Tanaka (483). The impression was given that the 

 active vitamine had been isolated in the pure state from rice polish- 

 ings. It will be well to describe Hofmeister's method, different 

 from any previously used. He began his work with the notion that 

 our vitamine is, in reality, nicotinic acid, but this is not the case, 

 since in every case where nicotinic acid was isolated there was no 

 curative action. 



Hofmeister shook rice meal three times with double the volume of 

 80 per cent alcohol on the shaking machine. The filtered solution 

 was evaporated in the presence of a stream of air, in vacuum. Then 

 the residue was acidified up to 3 per cent with hydrochloric acid, the 

 fatty acids extracted with ether, the ether removed, the solution 

 concentrated to a syrup in vacuum at a low temperature, and again 

 taken up with 80 per cent alcohol to free it from colloidal impurities. 

 The clear filtrate was made faintly alkaline with sodium carbonate, 

 taking care to prevent an excess of alkali, and precipitated by bismuth 

 potassium iodide (prepared according to Kraut) with constant 

 stirring. It is necessary here to avoid a strongly acid reaction, else 

 the active substance precipitates too. The dirty grayish-yellow 

 precipitate of the choline fraction was filtered after standing for 5 

 hours; to the filtrate, was added one-tenth of its volume of 20 per 

 cent hydrochloric acid, and the active substance precipitated out 



