92 STUDIES IN GUJARAT COTTONS 
A third test may be quoted. In seeking to develop pure strains of Goghart 
cotton, we commenced by selecting in the field, with great care, plants which 
completely conformed to the Goghari type. The seeds from these plants were 
sown, and the plants from their produce were examined. The following 
table shows the percentage of these which conformed to the Goghars type 
in 1915-16 :-— 
Strain per cent, 
1a <= 4079 
Bo ine -- 40°3 
Chinere oe Gayl 
Allowing that a limited amount (not more than 3 per cent.) of cross-ferti- 
lization may have taken place in the field, this would be totally insufficient 
to account for such a large proportion of non-Goghart plants in the offspring 
of those selected. 
These results, therefore, seem to show conclusively in the herbacewm 
cottons of Gujarat, or at any rate in the Broach deshi and Goghari types, the 
following points :— 
(1) That cross-fertilization of the different types of herbaceum cottons 
does take place. 
(2) That this cross-fertihzation takes place normally to the BSD! of 
1 to 2°85 per cent. and perhaps more. 
(5) That the cotton plants in cultivation are very largely composed of 
hybrids between the various types described, and, hence, 
ordinary selection on the basis of the characters of the produce 
of a particular plant will not yield a constant type either of 
plant or produce. 
5. What characters are hereditary in Goghari cotton ? 
In considering which of the characters associated with Goghari cotton are 
hereditary, we have studied the following. We will first state what we have 
found with regard to each of them and then describe the data on which these 
conclusions are based. 
Boll characters. 
(1) The shape of the boll (spherical, tapering, ete. ) is hereditary. 
(2) The size of the boll, at any rate so far as the spherical type of boll 
is concerned, is hereditary. 
(3) The quantity of kapas per boll is hereditary. 
Plant Characters. 
(1) The proportion of bolls borne on primary fruiting branches is 
hereditary, 
