114 THE COTTON PLANT IN EGYPT chap. 



the number of such cases in proportion to the total 

 number of plants, or bolls, originally chosen for testing. 

 This number was 5 in 75 in the original trials, but it must 

 be remembered that not all the vicinistic pollen parents 

 wUl have characters dominant over those of the mother 

 plant ; in any one character the chances will be even for 

 or against such recognition. These chances can be reduced 

 by taking more than one character, but when the two 

 intercrossed characters are identical, or nearly so, the 

 observational errors will prevent detection. Thus, we 

 shall not be far wrong if we double the number of cases 

 observed when working with two pairs of characters, so 

 that our figure becomes 10 vicinists in 75, or 13 "3 

 per cent. 



In the other case we find that some offspring bear 

 characters which are recessive to those of the parent. A 

 lint of 40 mm. in this case gives rise to some plants with 

 20 mm. lint. The presumption here is that the parent 

 was a heterozygote, produced by natural crossing in some 

 year antecedent to the one in which the seed was taken 

 from the field, having been simply self-fertilised in that 

 particular year. The number of such F^. plants in the 

 original total is nevertheless dependent on the amount of 

 crossing which takes place every year, and should serve as 

 a check upon the direct observation of the former case. If 

 now we assume simple Mendelian segregation for the 

 characters we employ, or prove it by growing further 

 generations from these splitting forms, we see that the 

 number of such F^ heterozygote plants depends not only 

 on the renewal of that number each year by fresh crossing 

 but also on its reduction each year by segregation of 

 homozygotes from them. 



If crossing to the extent of 10 per cent, per annum 

 were naccompanied by segregation, we should find the 

 number of vicinists increasing along a logarithmic curve, 

 since some of the crossed strains would be re-crossed in 



