286 
Journal of Agricultural Research 
Vol. XXV, No. 7 
pronounced. Similar facts have been established for cottonseed by 
Bidwell, 6 who has shown that cottonseed from the southwestern United 
States, on the average, is low in oil and high in protein, while seed from 
the Atlantic and Pacific coasts has a higher oil and lower protein con¬ 
tent in the order named. While Bidwell’s generalization is undoubtedly 
justified, exceptions must be anticipated, for over such large territories 
climatic and soil conditions must be exceptional here and there. This 
being so, exceptions will be encountered, if any reliance can be placed 
upon the work of Bain and Anders as reported by Cook (5), who writes 
as follows: 
The fluctuations induced by conditions of growth or associated with various degrees 
of maturity attained by the seeds were so large as to conceal inherent differences in 
individual plants or progenies. 
Hence the finding by Rast (9) of great variations in composition of 
Georgia cottonseed is not astonishing nor to be regarded as invalidating 
Bidwell’s generalization. 
If it be true that cottonseed varies greatly in composition and toxicity 
in different parts of the country and perhaps also in different crop years, 
then the toxic factor in cottonseed, whatever it may be, might well vary 
correspondingly. If Withers and Carruth ( 12 , 13 ) are right in attribut¬ 
ing the toxicity of cottonseed to the presence in it of a phenolic substance 
named by Carruth “gossypol,” then the gossypol content of different 
samples of cottonseed should vary as the toxicity varies. Apparently 
Carruth ( 3 ), recognizing that gossypol occurs in the so-called “gland 
dots” or “resin glands” in cottonseed, the distribution of which has 
been studied by Stanford and Viehoever (jo), does not believe the gossy¬ 
pol content of cottonseed to vary greatly. He states that “since all 
varieties of seed seem to have approximately the same number of glands, 
it would appear that gossypol does not vary to a greater extent than the 
oil or protein content.” This statement, however, is not based on experi¬ 
mental evidence. Hence it is just as reasonable to assume that, even 
if the number of glands in different cottonseeds were the same, the glands 
might vary both in size and in gossypol content. Carruth ( 3 ) found gos¬ 
sypol in cotton-root bark, where, according to the histological studies 
of Stanford and Viehoever (jo), there are also internal glands. Car¬ 
ruth ( 3 ) also attributed the differences of toxicity to variations in the 
method of manufacture of the press cake. However, this suggestion, 
while it may apply in some cases, does not explain the infrequency of 
poisoning in certain regions where the method of manufacture of the 
meal is generally the same as that in regions where poisoning is prevalent. 
There is, then, very definite evidence that the composition of cotton¬ 
seed varies widely and no evidence inconsistent with the possibility that 
the gossypol content may vary correspondingly. To determine whether 
or not the gossypol content of cottonseed varies and to determine whether 
there is any correlation between the gossypol, the protein, and the oil 
content is the purpose of the present paper. A study of the oil and 
protein content was included in the investigation, so that three criteria 
instead of one might be available for checking the results. The protein 
and oil analyses could be used to determine whether or not the sample 
was representative of the region from which the seed came and also to 
rule out “sports” or atypical seed. Consideration of the oil and protein 
content made it unnecessary to consider in detail the various factors 
affecting the seed. 
• Personal co mm u n ication from G. I,. Bidwell. Bureau of Chemistry. 
