EFFECTS OF INBEEEDING AND CEOSSBEEEDHSTG. 37 



the segregation in the second generation proved that each strain 

 furnished a dominant factor lacking in the other. In this case the 

 factors were visibly different in effect, in that one increased the num- 

 ber of internodes, the other their length. 



The two hypotheses are not easy to distinguish. Under both of 

 them the departure from the conditicrn of the pure inbred strains is in 

 direct proportion to the amount of heterozygosis in at least certain 

 factors. Under the stimulation hypothesis, however, the differences 

 in vigor may be due to factors which of themselves have nothing to 

 do with vigor: i. e., there may be no difference between AA and aa in 

 vigor, while Aa may be superior to both. Under the dominance 

 hypothesis it is merely to be supposed that Aa is equal to the more 

 vigorous of the two homozygous types. Dominance indeed need 

 not be perfect. 



Under the dominance hypothesis it should be possible to isolate 

 inbred strains, homozygous in all factors conducive to vigor and hence 

 equal or superior to crossbreds. This should not be the case with the 

 stimulation hypothesis. Again, under the dominance hypothesis, 

 the crossing of two inbred strains, followed by random breeding, 

 should result in a population with a skew distribution as regards 

 measurable characters, according to the expansion of (f + i)" where 

 n is the number of factors. The distribution would be symmetrical 

 under the stimulation hypothesis. 



The invariable degeneration following self-fertilization which East 

 observed in corn, and the failure to find skew distributions as the 

 rule in the second generation of crosses, were urged by him in favor 

 of the stimulation hypothesis. These objections, however, were met 

 by Jones, who continued East's experiments with corn. He pointed 

 out that owing to the phenomenon of genetic linkage, which was 

 known to occur in corn as well as in several other plants and animals, 

 the consequences of the dominance hypothesis of hybrid vigor would 

 really be much closer to those of the stimulation hypothesis in the 

 above respects than had been recognized. It should frequently 

 happen that detrimental recessive factors would be linked with 

 favorable dominant ones. In these cases the homozygotes would be 

 of the types AbAb or aBaB, while the heterozygotes would be largely 

 AbaB. The heterozygotes, containing both dominant factors, would 

 be superior to both homozygous types, giving a situation almost 

 indistinguishable from that which follows the stimulation hypothesis. 

 In this suggestion Jones did not add a new hypothesis to the theory 

 of inbreeding; he merely pointed out the logical consequences of a 

 phenomenon, linkage, which had already been demonstrated. The 

 dominance hypothesis was thus greatly strengthened. Collins (1921) 

 has recently shown that even if linkage is disregarded, the objec- 

 tions to the dominance hypothesis are not serious in cases in which 



