INFLUENCE OF INBREEDING ON HYDATINA SENTA. 9 



mum must be the same 1 for every pure line. Inbreeding must, 

 on this view, always eventually reduce vigor if there is random 

 segregation and recombination. 



On Bruce's view, according to which vigor depends on the 

 number of different kinds of gene present, there ought to be some 

 cases in which the inbreeding of a heterozygous line or the selfing 

 of a heterozygous individual would result in F], or F 2 , or F 3 , etc., 

 having as many present genes as the parent. Thus a parent 

 having the constitution AaBbCcDd might, on selfing for one or 

 more generations, have a few progeny with the formula AABB- 

 CCDD. Such individuals should be as vigorous as the original 

 parent, or, if dominance is not complete, even more vigorous; 

 and their progeny, produced by self-fertilization, should never 

 show decrease of vigor. Any individuals that came to have the 

 constitution AABBCCdd would be a little less vigorous. Those 

 having the formulas AABBccdd and AAbbccdd would be still less 

 vigorous, and the constitution aabbccdd would represent the 

 minimum of vigor. Thus, if four genes were concerned with 

 vigor, it should be possible to isolate four pure lines, each with 

 its own degree of vigor, which would not thereafter decrease. 

 Each line reaches a minimum of vigor, but that minimum is not 

 the same for different lines. 



Either of these two views, as stated, fits the evidence so far 

 obtained from cases in which inbreeding reduces vigor, though 

 evidence could be obtained, at the expense of sufficient labor, 

 which would not fit both. East and Hayes's addition to the 

 heterozygosis view, which would probably be assented to by 

 Shull, namely, the postulation of an "inherent vigor' not 

 dependent on heterozygosis, would make it possible to produce 

 pure lines having different degrees of irreducible vigor. But in 

 no case could a pure line, derived by successive inbreeding from 

 an FI that was more vigorous than its parents, be as vigorous, on 

 the view of Shull, or of East and Hayes, as the original FI; this 

 would be possible, as explained above, on the view of Bruce. If 

 many genes were concerned with vigor, testing the correctness of 

 the two hypotheses on the basis of this distinction would probably 

 require a prohibitive amount of labor. 



1 Note, however, the effect of East and Hayes's addition to this theory, discussed 

 below. 



