170 THE VTTAMINES 



precipitated with silver and baryta, instead of phosphotungstic acid. 

 We observed that this extract behaved quite differently from the 

 previous one; it appeared to contain less potassium and no free 

 choline. On decomposing the silver precipitate, a substance was 

 obtained (not the vitamine fraction), which melted at 231C. and 

 was identified as allantoine. Schaumann (464) described the isola- 

 tion of a substance containing nitrogen, which was perhaps allantoine 

 but was not investigated either chemically or therapeutically. 



We have therefore isolated from rice polishings two known 

 nitrogenous substances, allantoine and choline, and the vitamine 

 fraction. The method described in our first paper on rice polishings 

 seems to be peculiarly adapted for the isolation of the vitamine 

 fraction, although it has shown the problem to be more complicated 

 than was apparent at first sight. We met with particular difficulty 

 when the rice polishings were not extracted in the laboratory but in 

 a chemical factory, and in large quantities. The investigation was 

 begun with 380 kilos of rice polishings (465) using the method 

 described in our earlier work on the same material. For technical 

 reasons, however, it seemed necessary to remove the fat first. Ordi- 

 nary instead of hydrochloric acid alcohol was used, and the evaporated 

 extract was hydrolyzed for two hours with 5 per cent sulphuric acid. 

 In other respects the procedure was the same as above; the individual 

 fractions were not investigated but careful experiments were made 

 with the vitamine fraction. The other substances, of no direct 

 concern in our problem, were kept for later work. We thought at 

 first that we had isolated two different substances by means of frac- 

 tional crystallization of the vitamine fraction, although they seemed 

 to crystallize out together, especially when the concentration was too 

 great. When the first fraction was recrystallized, needle-like crystals 

 were obtained which melted at 233C. (uncorrected) and gave the 

 following figures on combustion according to the method of Pregl : 



4.796 mgm. substance yielded 10.34 mgm. CO 2 and 1.685 mgm. H 2 O corre- 

 sponding to 58.80 per cent C and 3.93 per cent H. 



4.212 mgm. substance yielded 9.095 mgm. COz and 1.465 mgm. H 2 O, corre- 

 sponding to 58.89 per cent C and 3.89 per cent H. 



4.367 mgm. yielded (at 713 mm. 18C.) 0.420 cc. N corresponding to 10.58 per 

 cent N. 



Only a small part of the nitrogen could be determined by Van 

 Slyke's method. Nitrogen could not be estimated by Kjeldahl's 



