102 THE VITAMINS 



The doubt introduced by the work of Wilhams, Waterman and 

 Gurin (1929) as to the adequacy of destruction of vitamin B and 

 conservation of vitamin G in the autoclaving of yeast, has apparently 

 been cleared up by Bisbey (1930) who shows that the treatment used 

 and recommended by Chase is entirely satisfactory as applied to the 

 type of yeast ("bakers' ") which she (Chase) used, whereas other 

 types of yeast may require a different treatment or may be less suitable 

 for this purpose. Each investigator should, therefore, assure himself 

 by experimentation upon the yeast he proposes to use, that it is, after 

 autoclaving, sufficiently freed from vitamin B and a sufficiently good 

 source of vitamin G to be satisfactory for this purpose. Or the gen- 

 eral method here described may be used with something else than 

 autoclaved yeast as the source of vitamin G. (See also the work of 

 Chick and Roscoe, 1929.) 



Under the conditions described, the rate of growth appears to be 

 satisfactorily uniform (when averages of sufficiently large numbers 

 are considered), and proportional to the vitamin B intake, but not 

 necessarily arithmetically proportional. Quantitative comparisons are, 

 therefore, most accurate when made at the same levels of gain in 

 weight. 



It is suggested that quantitative comparisons of vitamin B values 

 be based on the determination of the amount of test material which 

 when fed daily under the conditions just described permits a gain of 

 3 grams per week during an experimental period of from 4 to 8 weeks. 

 This amount of vitamin B may be taken as a "unit" for use in those 

 comparisons or discussions in which such a unit (necessarily some- 

 what arbitrarily chosen) is considered a convenience. 



The unit here suggested is no more arbitrary than any such unit 

 must necessarily be in the present state of knowledge, and has the 

 advantage of direct relationship to the experimental operation upon 

 which it is based. It is also believed that a gain of weight of about 3 

 grams per week is favorable to a combination of delicacy and reliability 

 in quantitative comparisons, in that it ensures sufficiently healthy test 

 animals for reasonable constancy of results, yet is so far below normal 

 growth that there is ample room for response to an increased intake 

 of the vitamin. 



Tentatively, then, a unit of vitamin B may be defined as that amount 

 which, when fed as a daily allowance, results in a rate ,of gain of 3 

 grams per week in a properly standardized and controlled rat during 

 an experimental period of 4 to 8 weeks. The direct results of the 

 feeding test show what weight of the food under test contains one 



