56 " A PURE-LINE METHOD IN CORN BREEDING 



together should not give in certain combinations considerably greater 

 yields than those produced by the combination of Strains A and B. 

 At any rate the result is sufficiently striking to suggest that the method 

 of separating and recombining definite pure-lines may perhaps give re- 

 sults quite worth striving for. 



This suggestion will be more readily appreciated perhaps if I discuss 

 briefly the theoretical aspect of this method of pure-line breeding as 

 compared with the method now in use among the most careful corn- 

 breeders. In the light of my results, the constant precautions that are 

 taken in the method now in use to prevent in-breeding, have for their 

 real object the retention of the most efficient degree of heterozygosis 

 or hybridity, and it is obvious that the selection of the most vigorous 

 individuals for seed, really picks out those individuals which have this 

 most efficient degree of hybridity. While I have not investigated the 

 inheritance of the various characteristics of the pure lines of maize 

 and am not in position to say that they all follow Mendel's law, many 

 investigations of particular characteristics in corn have show r n that 

 those characteristics are Mendelian. Even if some of the differentiating 

 characteristics of corn should not prove to be Mendelian, it seems not 

 improper to discuss the two methods on the Mendelian basis. 



In the method which selects for seed the most heterozygous indi- 

 viduals, the characteristic splitting and recombination of unit-characters 

 must produce an offspring of quite various degrees of heterozygosis. 

 Some individuals will be as complex as the selected parents, others will 

 have many of the same units in the homozygous condition, and thus be 

 less complex and consequently less vigorous. According to the laws of 

 chance a few individuals in the field may be expected to be almost or 

 quite completely homozygous, and as a result will be very inferior in 

 vigor and will produce but little grain. The result of such a process 

 must always be to give a crop of lower average yield than the average 

 of the selected seed. Moreover, these different combinations of unit- 

 characters and different degrees of hybridity in the offspring of a 

 complex hybrid must introduce a certain amount of heterogeneity into 

 the crop which will have the effect to also lower the average quality 

 with respect to any other desirable points which have been used as 

 guides in the selection of the seed-corn, and efforts at the attainment of 

 homogeneity by the method now in use tend to lessen physiological 

 vigor, and therefore lessen the yield, owing to the fact that such homo- 

 geneity in the offspring of hybrids is to be attained only through 

 homozygosis in respect to all those characteristics which affect the 



