Abstract
The standard deviation ([sigma] = [SIGMA]d 2/ (n [long dash] 1) ) of a single determination of the increase in weight of rats (bucks or does) in a 5 weeks'' growth test for vitamin A is not different from that of the normal increase in weight of rats for the same period and same weight. The deviation ([sigma]= 14.72 for bucks and 11.02 for does) of the experimental animals is the same whatever the mean increase in the test may be and whatever the average initial weight of the group. To obtain equal degrees of accuracy, about twice as many does as bucks must be used in a test.