184 
Journal of Agricultural Research 
Vol. XXVIII ,No. 2 
gossypol, are presented. The rats fed cottonseed kernels grew on the average 
slightly better than the rats fed the estimated comparable quantities of gossypol. 
In the case of the experiments with the cottonseed kernel diet estimated to con¬ 
tain 0.135 per cent of gossypol, the diet was withdrawn early and replaced by the 
control diet of peanut meal. The sudden change in the rate of growth is very 
marked, especially in the case of the males. On the sixtieth day the maximum 
difference between the body weight of the largest and smallest male rats was 70 
gm. as compared with 30 and 35 gm. for normal rats on control diets. 
It is also shown that the rate of growth on Lone Star seed was decidedly poorer 
than that with Egyptian and Durango seed. Seed of the last two varieties, how* 
ever, were found by intraperitoneal injection of the ether extract to be as toxic 
as the analysis indicated, whereas the Lone Star seed was slightly less toxic than 
indicated. Such observations as these illustrate the necessity for a much 
larger number of experiments when exact estimates of toxicity are to be made in 
order to rule out experimental error due to variation in animals. 
Fig 11.—Effect of diet containing cottonseed kernels estimated to contain approximately 0.135 per cent 
crude gossypol. “C” indicates change to control peanut meal diet. “Average recovery” indicates 
growth when all rats were receiving control diet 
The cottonseed kernels fed in these experiments appeared to be slightly less 
toxic than was indicated by chemical analysis. The differences between the 
predicted and the observed toxicity, however, are not great. 
Feeding tests with a diet containing 26.2 per cent of Durango cottonseed, esti¬ 
mated to contain 0.257 per cent of gossypol, are similar to those in which the diet 
contains 0.225 per cent of gossypol. Although in the Durango diet the estimated 
quantity of gossypol was a tenth greater than it was intended to be, this excess is 
within the limits of- experimental error. The Durango seed, therefore, appears to 
be about as toxic as the gossypol analyses would indicate. The curves are given 
in figure 12. 
, The curves representing the rate of growth upon a diet containing 19.2 per cent 
of Egyptian cottonseed kernels, estimated to contain 0.227 per cent of gossypol, 
also presented in figure 12, are in practical agreement with the results of the 
tests in which the diet contained 0.225 per cent of added gossypol. The only 
difficulty in the way of stating that they are absolutely identical is the fact that 
one of the five rats on the Egyptian diet lived about twice as long as any of the 
