56 BULLETIN 1134, U. S. DEPARTMENT OF AGRICULTURE. 



matured. Thus, in 1917, among 71 bolls which matured from a lot 

 of bagged flowers, 3 contained only 2 seeds, 6 contained only 3 seeds, 

 3 contained only 4 seeds, and 5 contained only 5 seeds. Hence, in 17 

 bolls (or 24 per cent of the number which reached maturity) fewer 

 than one-fourth of the mean number of ovules had been fertilized. 

 Counts were made of the number of seeds in 633 bolls which matured 

 from bagged flowers in 1920, and of these 2 contained only 1 seed, 7 

 contained only 2 seeds, and 14 contained only 3 seeds. Few of the 

 bolls which mature from flowers naturally open pollinated show 

 fertilization of less than half of the average number of ovules. Of 

 447 such bolls in 1920 only 22 (or 4.9 per cent) contained fewer than 

 10 seeds. On the other hand, of 633 bolls which matured from 

 bagged flowers during approximately the same period 201 (or 31.7 

 per cent) contained fewer than 10 seeds. 



It has been shown on preceding pages that the pollen of Pima cot- 

 ton is highly viable and perfectly self-compatible. Evidence has 

 been presented, also, that the number of pollen grains normally 

 reaching the stigmas vastly exceeds the number of ovules and that 

 fertilization of only a few of the ovules is necessary in order to in- 

 sure retention and maturation of the boll. It may be concluded, 

 therefore, that with this variety, under the conditions existing at 

 Sacaton, deficient pollination and fertilization are not important 

 factors in the shedding of bolls. 



INBREEDING IN RELATION TO FERTILITY. 



The effect upon cotton of continuous and strictly controlled self- 

 fertilization appears to have been little investigated. Leake and 

 Prasad (36, pp. 39-45), working in India with cottons of the 

 Asiatic type, obtained indications that partial sterility, as expressed 

 by the percentage of bolls retained, occurred in the first and later 

 inbred generations and that this tendency could be overcome by sub- 

 sequent cross-pollination. Different types apparently differed in 

 their tendency to sterility. One type snowed a marked tendency to 

 imperfect development of the anthers in the second inbred genera- 

 tion which, however, disappeared in the third generation. In one 

 strain self-pollination with pollen from another flower on the same 

 plant resulted in the retention of a much smaller percentage of bolls 

 than did cross-pollination with pollen from a sister plant; but in 

 another strain of the same type no such difference was observed. 

 The numbers of flowers dealt with in these experiments were too 

 small to afford conclusive results. 



A different conclusion was reached by Kottur (&£), who also 

 worked with the Asiatic type of cotton, in India. He found that 

 sterile anthers, containing no pollen, are of common occurrence. 

 In an unselected open-pollinated stock of the Kumpta variety, 

 of which 500 flowers were examined, sterile anthers were found 

 in all but 128, the proportion of empty anthers having been as high 

 as 43 per cent of the total number in some of the flowers. Con- 

 trolled self-fertilization during six successive generations did not 

 increase this form of sterility. In fact, it is shown by Kottur's data 

 that in the continuously self- fertilized strain the proportion of the 

 flowers having more than 10 sterile anthers was only 13.6 per cent, 

 as compared with 35.9 per cent in the open-pollinated stock. Kottur 

 also observed that the rate of boll shedding was lower and that the 



