54 CROSS-FERTILISATION IN TIIK INDIAN COTTONS. 



(2) In 1906 the seed of sixty plants was gathered from a Hold 

 crop of G. negledum Tod. and G. roseum Tod. and sown in separate 

 lots in the following year. These included the types (5 — 9) with both 

 "broad-" and "narrow-" lobed leaves and with yellow and 

 white flowers. The observations on the flower colour of this series 

 are given in Table IV. Owing to the fact that in these types yellow 

 is completely dominant to white it is impossible to detect crosses 

 between yellow and white-flowered plants in the progeny of the yel- 

 low forms. Here it is only possible to conclude that the thirty-four 

 plants selected were pure yellows. The white-flowered plants have, 

 however, given seven yellow offspring, and these seven are undoub- 

 tedly crosses. Accurate information concerning the leaf form is 

 not available since no determinations of leaf factor were made in this 

 series. A record has, however, been kept in which three groups, 

 broad-lobed, narrow-lobed, and intermediate, have been distinguished. 

 These are tabulated in the second section of Table IV. From this 

 it will be seen that thirty-eight plants out of the sixty gave offspring 

 without impurity ; that twelve were themselves piu'e, but that among 

 two hundred and seventy-five offspring occurred twenty-eight plants 

 which showed clear indications of being the result of a cross, 

 and that the remaining ten plants were themselves impure. Of 

 these last it is only necessary to remark on the paucity of intermedi- 

 ates. At the time these observations were made the leaf factor had 

 not been identified and no means of accurately recognising the limits 

 of the ' ' broad- ' ' and ' ' narrow- ' ' lobed types had been found. 

 There is little doubt, therefore, that many of the forms which, though 

 intermediate in reality, approached one or other extreme, were 

 included among the extreme forms and consequently it is impossible 

 to draw any conclusion from the marked divergence from the normal 

 in the numerical proportion of the groups. 



(3) In the cotton season of 1909 considerable damage resulted 

 from the ravages of the boll worm, numbers of the self-fertilised bolls 

 were attacked, and it was feared that loss of promising types might 

 result owing to the reduction of the quantity of seed harvested. 

 A further supply of seed was, therefore, gathered from the same 



