182 THE VITAMINES 



when the substance was recrystallized twice did these reactions dis- 

 appear. The purified substance was compared with that obtained 

 before, and mixed melting points were taken, whereby the new 

 product melted four degrees lower. Mixed in equal quantities, both 

 substances melted at 229C. (corrected). They were absolutely 

 identical as regards solubility, reactions and crystalline form. The 

 substance gave a white precipitate with Millon's reagent, but the 

 color reactions were always negative. In a 0.74 per cent solution, 

 a definite rotation in the polarimeter could not be shown. On 

 drying in vacuum at 100 C. and analyzing, the following figures 

 were obtained: 



3.267 mgm. yielded 6.63 mgm. CO 2 and 1.05 mgm. H 2 O; 55.35 per cent C, 3.60 



per cent H. 

 4.224 mgm. yielded 8.425 mgm. CO 2 and 1.36 mgm. H 2 O; 55.72 per cent C, 3.69 



per cent H. 

 4.256 mgm. yielded 8.66 mgm. CO 2 and 1.36 mgm. H 2 O; 55.50 per cent C, 3.58 



per cent H. 



3.048 mgm. yielded 0.373 cc. N. (710 mm. 15.5C.) 13.53 per cent N. 

 3.627 mgm. yielded 0.430 cc. N (712 mm.; 14C.) 13.21 per cent N. 

 3.286 mgm. yielded 0.401 cc. N (705 mm.; 18C.) 13.28 per cent N. 



These figures correspond to the formula C24Hi 9 O 9 NB (521.24), 

 which gives 55.25 per cent C, 3.68 per cent H and 13.44 per cent N. 

 Since a definite crystalline substance is noted on the tube when drying 

 at 100 C. in vacuum, analyses were made after the substance was 

 dried in vacuum over sulphuric acid; the following figures were 

 obtained : 



3.733 mgm. gave 7.775 mgm. CO 2 and 1.185 mgm. H 2 O; 56.80 per cent C, 3.55 



per cent H. 

 3.538 mgm. gave 7.35 mgm. C0 2 and 1.195 mgm. H 2 O; 56.66 per cent C, 3.78 



per cent H. 

 3.174 mgm. gave 0.372 cc. N (702 mm.; 17C.) 12.74 per cent N. 



These figures correspond to the formula C 2 6H 2 i09N 5 , which gives 

 57.01 per cent C, 31.87 per cent H and 12.81 per cent N. From this, 

 it appears that on drying at higher temperatures a substance rich in 

 carbon is lost. This characteristic and the great difference in 

 elementary composition excludes the possibility that this substance 

 was nicotmic acid. We have recently attempted to secure further 

 evidence in this connection by the use of the yeast method. If the 

 nitrogen in this substance is estimated by the method of 



