conclusively that the 8 cc. dose is the best one to use, three 
lots of six rats each were given diet 94 with the addition of 
6 cc., 7 cc., and 8 cc. of skimmed milk respectively per rat per 
day. The animals receiving 6 cc. of the milk gained an aver- 
age of 34 grams in eight weeks, those receiving 7 cc. averaged 
35 grams gain and the rats fed 8 cc. averaged 56 grams. 
Eight cc. of skimmed milk per rat per day is probably 
the best level at which to feed in order to detect any change 
in the vitamin B content of the milk, since it is at this level 
that we get the largest difference in the total gain in weight 
for eight weeks, as compared with the next lower amount fed. 
This does not furnish the optimum amount of vitamin B, since 
further addition of milk gave increased growth in many cases 
and in the general average. 
In another experiment the amount of milk given was 
varied in accordance with the weight of the rat. Each animal 
received 0.05, 0.075, 0.10 or 0.125 grams of skimmed milk 
powder (0.50, 0.75, 1.00 or 1.25 cc. reconstructed milk) per 
10 grams of rat per day. The amount of milk fed was calcu- 
lated on the basis of the weekly weighings, the milk fed each 
day for a week being determined from the weight of the rat 
at the beginning of the week. It was expected that this 
method would yield more consistent results and that the varia- 
tion among individuals receiving the same amount of vitamin 
B would be lessened by this method of feeding. The results, 
however, do not bear out this assumption. A comparison of 
the weight curves of the individual rats which were fed the 
same amount of milk throughout the experimental period re- 
gardless of body weight, with those of the rats receiving the 
graduated doses, shows that the former are as uniform as 
those in the other series. Table VII summarizes the results 
of the experiments in which the rats received the graduated 
doses. 0.1 gram of milk per day per 10 grams of rat is ap- 
parently the best level at which to feed the milk by this 
method, since for the same difference in amount of milk given 
a greater difference in weight gained is observed between this 
and the next lower amount, than between any other two suc- 
cessive doses. The method was not adopted since it gave no 
14 
