298 
Journal of Agricultural Research 
Vol. II, No. 4 
pollinated with the same type which furnished one of the parents of the 
original hybrid, it is conceivable that finally only the characters of this 
parent would continue to be expressed. A hybrid of this kind, although 
very “dilute,” might be expected to be in unstable equilibrium and, 
hence, to offer the proper conditions for the appearance of mutants. 1 
If we accept the hypothesis that mutability is a consequence of hybridi¬ 
zation, it is not difficult to account for the tendency of Egyptian cotton 
to produce mutants. As was pointed out on preceding pages, the type 
as a whole is believed by some authorities to have originated from crosses 
between distinct species of Gossypium. But even if this theory be 
rejected, the possibility that hybridization has been a factor in the 
mutability of this cotton is not removed, since (1) the different varieties 
which, although genetically related, are distinct in their characters, are 
in Egypt often grown in close proximity and their seeds frequently become 
mixed at the gins, so that a great deal of crossing takes place among them; 
(2) other types of cotton which readily hybridize with the Egyptian, such 
as the American Upland and Sea Island, have been repeatedly intro¬ 
duced into Egypt and precautions to keep them isolated have rarely if 
ever been taken; (3) practically every cotton field in Egypt contains 
numerous plants of the very distinct Hindi type, which crosses with the 
Egyptian, producing more or less fertile offspring. 2 
The mixed condition of the principal Egyptian varieties has been 
strikingly exemplified in the Arizona plantings grown from imported 
seed. These always contain a greater or less number of Hindi plants and 
of first generation hybrids between these and the Egyptian, together with 
a multitude of individuals which give evidence of earlier crossing by their 
less pronounced expression of Hindi characters. Even the Sakellaridis 
variety, which is only about eight generations removed from the parent 
individual, has shown itself to be badly contaminated. 
The theory that the mutability of Egyptian cotton is an after effect 
of hybridization with such distinctly different types as Hindi, Upland, or 
Sea Island, might be challenged on the ground that the recent mutants 
show no characters which can definitely be attributed to a non-Egyptian 
parent. This certainly appears to be the case with the Arizona varieties, 
which are purely Egyptian in the characters expressed. The objection 
may be met by assuming that the immediate ancestors of the mutating 
individuals were “diluted '* hybrids, which, while expressing only 
Egyptian characters, were in a condition of unstable equilibrium, favor¬ 
able to mutation. This assumption would imply that the remote non- 
Egyptian ancestor has made no direct contribution to the characters 
expressed in the mutating descendant and that the only remaining 
1 The possibility of inducing the production of desirable mutants by specially planned hybridization 
has been suggested by Mr. O. F. Cook (1913b, p. 86-87). 
2 According to Prof. Balls (1907, p. 57), “all varieties of Egyptian cotton, new or old, contain at least 
50 per cent of plants with hybrid constitutions.” In another work the same author (1912, p. 5) states: 
“The nominal varieties are more or less heterogeneous complexes of heterozygotes.” 
