214 PHYSIOLOGIC GENETICS 



INBREEDING AND CROSSING 



The effects of inbreeding or of crossing lines on the mean expression of a quantita- 

 tive character provide some information about the dominance of the genes that affect 

 the character, and may perhaps allow us to draw tentative conclusions about the role of 

 the character as a determinant of natural fitness. The effects of inbreeding and of 

 crossing are the same but opposite in direction and they are both the outcome of the 

 same properties of the genes, namely dominance. If inbreeding produces a change in 

 one direction, then the genes on the average should show dominance of one allele over 

 the other and the recessive alleles should affect the character in the direction of the 

 change on inbreeding. To be specific, if the character declines on inbreeding, then 

 the alleles that reduce the character tend to be recessive to their alleles that increase 

 it. If the character does not change on inbreeding it cannot, however, be concluded 

 that the genes do not have dominance. It can only be concluded that they do not 

 have directional dominance; there could be dominance at every locus if some genes 

 were dominant in one direction and some in the other. The presence or absence of 

 directional dominance is the only thing to be learned about the genetic properties of the 

 character from the presence or absence of inbreeding depression or heterosis. 



An empiric conclusion based on much evidence is that characters obviously 

 connected with natural fitness exhibit inbreeding depression and heterosis. (By 

 characters obviously connected with natural fitness is meant characters reflecting some 

 aspect of fertility or general viability.) This gives some grounds, therefore, for suppos- 

 ing that if a character shows inbreeding depression and heterosis, then it probably 

 has a close connection with fitness, and natural selection in the past has favored 

 individuals with high rather than low or intermediate values. The presence of heterosis 

 is to be judged, in this connection, as any deviation of the F x from the mid-parental 

 value. 



THE NUMBER OF GENES 



The question of how many genes are concerned in the genetic determination of a 

 quantitative character is one of great interest. But unfortunately the answers that 

 can be obtained have not much meaning, and the question itself is not very meaningful. 

 The number of genes cannot be stated meaningfully without specifying the magnitude 

 of their effects. It is not impossible that all genes segregating in a strain affect every 

 character in some slight degree. However, if the suppositions are made that genes 

 either affect the character or do not and that all those that affect it do so by roughly 

 the same amount, then an estimate of the number that may have some meaning can be 

 obtained. 



The procedure for estimating the number of genes is to cross two strains that differ 

 in the mean value of the character and to measure the variance in the F 2 generation 

 and in the F 2 generation. The difference of variance between the F 1 and F 2 gives 



