252 THE BIOCHEMISTRY OF B VITAMINS 



a number of these factors mentioned had not been considered in the 

 experiment. 



With rats, mice, chickens, and most other smaller animals, growth has 

 frequently been taken as a suitable criterion for assessing the fulfillment 

 of the B vitamin requirement. In humans this is not the case, supposedly 

 because of the difficulties in the interpretation of such data. Indeed, studies 

 involving vitamin supplementation of adequate diets have been frequently 

 subjected to criticism for adopting "growth" as a criterion without any 

 further consideration of other factors involved. Moreover, from a physi- 

 ological point of view, "increased" and "improved" growth may be very 

 different things, there being no guarantee that the most rapid growth rate 

 is the most desirable. 27 



Melnick et al. 28 have developed B vitamins bioassay methods which 

 employ human subjects. These methods (p. 283) depend upon the study 

 of the urinary excretion of the B vitamins, which is presumably a function 

 of the intake. While not bearing directly upon the problem at hand, the 

 work of these investigators deserves mention at this time in that it is 

 subject to the variables previously mentioned, yet illustrates the value 

 that is attached to carefully controlled experimentation with human 

 subjects. 



Dietary surveys. Somewhat akin to controlled diet studies are those 

 on healthy and avitaminotic populations. By a careful consideration cf 

 numerous dietary surveys, it is sometimes possible to estimate the level 

 of nutrition which will bring about a deficiency of one or more of the 

 B vitamins. Excellent examples of this approach have resulted from 

 studies of prison camps and of circumscribed populations during the 

 recent war, since the diet was frequently rigidly controlled and permitted 

 unusually accurate assessment of the vitamin intake. One of the classical 

 studies of this nature, however, is an earlier one dealing with thiamine- 

 deficient diets in the Orient. 



Cowgill 12 studied some 180 human diets in regard to their thiamine and 

 calorific content and association with beriberi. Williams and Spies 14 later 

 reassessed these data and arrived at a more accurate estimate of the 

 minimal human requirement necessary to prevent beriberi. A modified list 

 of the diets studied in increasing order of the ratio of thiamine to Calories 

 in the diet is given in Table 2. A summary of the results is given in 

 Table 3. 



The validity of these data and the conclusions derived from them are 

 borne out adequately by a study of a large number of American dietaries 

 (Table 4) . 29 The thiamine-to-calorie ratio seems generally to be above 

 that associated with clinical beriberi, other dietary factors (i.e., fat; page 



