322 SOW CROPS GROW. 



ized and cross-fertilized plants in the 1st, 3d, and 9th 

 generations were respectively as 64 to 100; 35 to 100, 

 and 26 to 100. 



In other cases, but, so far as observed, mnch less com- 

 monly, self-fertilization gives the best results both as 

 regards numbers and vigor of offspring. In Darwin's ex- 

 periments a variety of Mimulus luteus originated, of 

 which the self-fertilized progeny surpassed the cross-fer- 

 tilized, during several generations. In the seventh gen- 

 eration the ratio of superiority of the self-fertilized, as 

 regards numbers of fruit, was as 137 to 100, and in respect 

 to size of plants as 126 to 100. 



Continued self-fertilization, is thus limited by its ten- 

 dency, as statistically determined, to reduce both the 

 vegetative and reproductive vigor of the plant. On the 

 other hand, cross-fertilization is possible or practicable 

 only within very narrow bounds, and the increased pro- 

 ductiveness that follows it soon reaches a limit, as is 

 shown by the history of vegetable hybrids. 



That neither mode of fertilization is exclusively or speci- 

 ally adapted to the highest development of plants in gen- 

 eral, or of particular kinds of plants, is shown by the fact 

 that in the course of Darwin's researches on the Ipomea 

 purpurea, just referred to, in the sixth generation a self- 

 fertilized plant (variety) appeared, which was superior to 

 its crossed collateral, and was able to transmit its vigor 

 and fertility to its descendants. 



It is evident, therefore, that the causes which lead to 

 higher development co-operate most fully, sometimes in 

 the one, sometimes in the other, mode of impregnation 

 and do not necessarily belong to either. We must be- 

 lieve that excellence in offspring is the result of excel- 

 lence in the parents, no matter what lines their heredity 

 may have followed, except as these lines have influenced 

 their individual excellence. That crossing commonly 

 gives better offspring than in-and-in breeding is due to 



