488 JOHN W. GOWEN 



In later corn hybrids, the products of more stringently selected inbreds, 

 the emphasis was reversed. The specific combining ability was zero to five 

 times as important as the general. Improved utilization of hybrid vigor has 

 seemingly selected and fixed general combining ability in the approved in- 

 breds. Further progress is dependent on specific combining ability. One com- 

 parison weakens this evidence. On exactly the same ten inbred combinations 

 one set grown at Ames, the other at Davenport, the specific combining ability 

 was five times that of the general at Ames, while at Davenport the general 

 and specific were identical. A place X genotype interaction in the general- 

 specific combining ability similar to that observed above for Drosophila egg 

 yield is also important even in these highly selected lines. 



SIGNIFICANCE OF ENVIRONMENT-GENOTYPE 

 INTERACTION IN HETEROSIS 



An experiment by Dr. Loh evaluating the significance of early or late test- 

 ing furnishes data on the part played by environment in the stability of the 

 hybrid phenotypes (1949). Fifty full brother X sister lines were formed from 

 each of three wild stocks having different geographical or chronological ori- 

 gin. Each line was then doubled and mated full brother X sister for as long as 

 possible, or until 37 generations were reached. The average productivity of 

 the initial lines crossed to the same synthetic strain at the start of the in- 

 breeding was Ames 1947, 179 ± 2; Ames 1943, 176 ± 3; and Amherst 1947, 

 166 ± 3 eggs for the 5, 6, and 7th days after the hybrid females hatched. All 

 surviving inbred lines were crossed to the same synthetic stock, and the hy- 

 brid females tested at the 8, 9, 16, 23, and 30th generations. The results were 

 consistent for the three stocks. The egg productions of the hybrids declined 

 2.4 eggs per generation on the average. This result was surprising, but may 

 possibly be accounted for by the fact that the inbred lines surviving in the 

 three stocks were becoming more and more like the synthetic tester due to 

 the fact that they were cultured on the same media and in the same way. The 

 favorable gene differences between the crossed lines became less each genera- 

 tion and resulted in a progressive lowering of hybrid yield. As the generation 

 times were confounded with time of year, it was also possible that the egg 

 yields showed some effects of the progressive changes in season. 



The surviving inbred lines were tested for egg yield on the 21, 26, and 31st 

 generations. They showed an average decrease in egg yield of 4.3 eggs per 

 generation. This decrease was greater than that observed for the inbreds X 

 synthetic cited above. This was not entirely unexpected, although it did in- 

 dicate continued and persistent heterozygosity in the inbred lines to a much 

 greater extent than was sometimes realized. The inbred lines produced 20 to 

 40 per cent less than the hybrids. The differences became greater as the in- 

 breeding advanced. 



The inbreds of the 15, 24, and 34th generations were crossed in all possible 



