256 CHAMBERLAIN AND VEDDER. 



varies from 2 to 5 per cent so that a jjrief calculation will show that 

 the extract preiDared according to this method contains between 1/1^000 

 and 1/5,000 of the total amount of phosphorus contained in the original 

 polishings. In other words, at least 999 parts out of each 1,000 parts 

 of phosphorus are proven to be unimportant in the prevention of 

 polyneuritis of fowls. It is impossible to state positively that this 

 minute trace of phosphorus, remaining in the extract is not the neuritis- 

 preventing element, but we can at least say that it appears, utterly 

 incredible that it can be of the slightest importance. Therefore, the 

 theor}^, especially advocated by Schaumann, ( 2 ) that beriberi and poly- 

 neuritis gallinarmn are caused by a lack of phosphorus compoimds, is 

 strongly discredited to say the least. 



However, these findings do not necessarily conflict with the statement made 

 by Fraser and Stanton (3) and by others, including this Board, to the effect 

 that the proportion of phosphorus pentoxide present in a rice is an index of its 

 beriberi-preventing powers. Such a statement has never, so far as we are aware, 

 been intended to convey the impression that beriberi was due to a lack of phos- 

 phorvis compounds, but merely means that a rice containing phosphorus in 

 amounts above 0.4 per cent phosphorus pentoxide necessarily has a sufficient 

 quantity of pericarp adhering to the kernel; i. e., it is sufficiently undermilled. 

 As shown by a former publication of the board(5) the presence of a certain per- 

 centage of jDotassium appears to be an equally reliable index of safety, and so 

 in fact would be present, x in a rice of a definite percentage of any substance which 

 is found exclusively or almost exclusively in the pericarp and the aleurone layer. 



The 0.0406 per cent of nitrogen contained in this extract would indicate 

 a considerable quantity of proteid matter provided the nitrogen were 

 present in combination with proteids. This would not seem to be 

 the ease, however, since the extract failed to give the xanthoproteic 

 and biuret reactions and moreover we should not expect to find proteids 

 in a fluid obtained by extracting with alcohol, but in order to prove 

 this point we determined to test the extract by the method of dialysis. 



Experiment U. — The brown residue obtained from 2 kilograms of 

 rice polishings, macerated with cold alcohol according to the method 

 described above in experiment 2, was mixed with 300 cubic centimeters 

 of distilled water and filtered until clear. This small bulk of water 

 was used in order that the aqueous extract resulting might be quite 

 concentrated. This aqueous extract, which contained only those sub- 

 stances soluble in cold water and cold alcohol, was then placed in a 

 parchment bag and suspended in distilled water. This parchment was 

 previously tested with solutions of egg albumen and sodium chloride, 

 and it was found that the sodium chloride passed freely through the 

 membrane while the egg albumen did not pass through at all. During 

 the process of dialysis the apparatus was kept in the ice-box to prevent 

 decomposition of the extract. At the end of two days, when the fiuid 

 on the outside of the bag had become a deep yellow, the diffusate was 



