VII NATURAL CROSSING 125 



positive direction. Hence, in mixed pollination of Fj upon 

 parent, we should find a higher percentage of vicinists. 



The steps of the preceding argument were taken long 

 before the experimental results were available. These 

 results were as follows : — 



Mixed pollination of Fj upon Egyptian . . . 20/100 or 



20 per cent. 

 Mixed pollination of Fj upon Upland . . . 25/65 or 28 



per cent. 



Even allowing for all possible errors, and for the scanty 

 numbers of ovules involved, the diflference between 3 per 

 cent, and 24 per cent, as values for the " prepotency " of 

 mother-parent and of Fi pollen, under the same experi- 

 mental conditions, cannot be without significance. 



Hence, we conclude that some part of the Fj pollen from 

 an inter-specific cross is prepotent over all other pollen ; 

 that self pollen is prepotent over pollen from some other 

 species ; and that pollen from other species only makes its 

 way to the ovule with difficulty. 



Gametic differentiation. — A conclusion which may 

 be immediately drawn from these differences in prepotency 

 of pollen is, that the error from, natural crossing is likely 

 to be less in the field — where it is difficult to avoid — than 

 on the breeding plot, where it can be avoided by appro- 

 priate, though expensive methods. 



Again, it is within the bounds of possibility that 

 immune strains might be discovered or developed. Such 

 discovery is scarcely probable, but it may as well be 

 borne in mind. 



The chief speculation in which we may legitimately 

 indulge relates to gametic differentiation. If our postulate 

 as to correlation between gametic composition of the 

 microspore nucleus and growth-processes of the pollen 

 tube can be substantiated, it follows that those tubes 

 which succeed during mixed-pollination must possess 



