232 DONALD F. JONES 



were due solely to the interaction between the dominant and recessive mark- 

 ers, there would have been a rapid approach to the level of vigor now re- 

 maining. If it were due to a larger number of genes distributed rather evenly 

 along the chromosome, the reduction in heterosis would be gradual, as it 

 has proved to be. Small amounts of heterosis may persist for a long time un- 

 til all of the genes contributing to it are removed by crossing over. 



A recent experiment by Stringfield (1950) shows a difference in produc- 

 tiveness between an F2 selfed generation and a backcross having the same 

 parentage. The amount of heterozygosis as measured by the number of 

 allelic pairs is the same in both lots. In the backcross there are more indi- 

 viduals in the intermediate classes with respect to the number of dominant 



TABLE 14.2 



INCREASE IN HEIGHT OF PLANT IN SUCCES- 

 SIVE BACKCROSSED GENERATIONS HET- 

 EROZYGOUS FOR A DOMINANT GENE 

 MARKER 



genes. This indicates a complementary action of favorable dominant genes. 

 Gowen et al. (1946) compared the differences in egg yield in Drosophila 

 between random matings, 47 generations of sib mating, and homozygous 

 matings by outcrossing with marker genes. The differences are significant, 

 and indicate a large number of genes having dominant effects on the repro- 

 ductive rate. 



INTERACTION OF GENES AND CYTOPLASM 



The suggestion has been made many times that heterosis may result from 

 an interaction between genes and cytoplasm. Within the species, differences 

 in reciprocal crosses are rare. In commercial corn hybrids, reciprocal differ- 

 ences are so small that they can usually be ignored. Evidence is accumulating 

 that there are transmissible differences associated with the cytoplasm, and 

 that these must be considered in a study of heterosis. Small maternal effects 

 are difficult to distinguish from nutritional and other influences determined 

 by the genotype of the mother and carried over to the next generation. 



The cross of the two different flowering types of tobacco previously cited 

 shows a maternal effect. The first generation cross of the indeterminate or 



