HETEROSIS IN A NEW POPULATION 441 



ference between the various advanced generations, the 1^2, F3, and F4 yield- 

 ing 94.3, 95.4, and 95.1 per cent as much as the Fi respectively. Kiesselbach 

 (1933) compared the yield of the F2 and F;; generations of 21 single crosses. 

 The yield of F2 and Fg was approximately the same, being 38.4 and 37.8 

 bushels per acre respectively. Wellhausen and Roberts (unpublished) com- 

 pared the average yields of the Fi, F2, F3, and F4 generations in 18 topcrosses. 

 The average yield was 10.8, 9.9, 9.6, 9.8 kilos per plot for the Fi, F2, F3, 

 and F4 generations respectively. 



In the attempt to obtain still greater yield over Celaya, approximately 

 1000 So and S3 lines were crossed with the single cross C 67 X C 90. Both lines 

 in this single cross were first generation selfs from Celaya, and the majority 

 of the lines crossed with it were from varieties of this race. It is of interest 

 to note the kind of lines that gave the highest yields in combination with 

 C 67 X C 90. Among the ten that were finally selected as the best combiners 

 with C 67 X C 90, three were from the variety Jal. 35, one from a variety 

 from the State of Coahuila in northern Mexico, and the rest were from 

 Celaya. The total number of lines included from Jal. 35 and from other va- 

 rieties not classified as belonging to the race Celaya were relatively small 

 compared to the total number of Celaya lines involved in the test. However, 

 four of the ten best combiners with respect to yield came from varieties out- 

 side the race Celaya. This is in line with the belief that the possibility of ob- 

 taining high yielding hybrids is greater in the combination of lines from dif- 

 ferent varieties than in the combination of lines from the same or closely re- 

 lated varieties. Nevertheless, hybrids obtained from a recombination of 

 Celaya lines were satisfactory in yield and generally more disease resistant 

 and more acceptable from an agronomic standpoint than hybrids between 

 Celaya and non-Celaya lines. 



In the yield test results of all possible single crosses between the ten 

 selected good combiners with the tester C 67 X C 90, the two lines C 110-3 

 and C 126-5 from Celaya (the same variety from which the tester lines were 

 obtained) were of considerable interest. These two lines were not only good 

 combiners with the single cross tester C 67 X C 90, but also combined well 

 with each other. The single cross C 110-3 X C 126-5 was among the highest 

 of all the 45 possible single crosses among the ten selected lines. 



The tester single cross which was made up of two average Si Celaya lines 

 would tend to isolate genotypes which contribute the greatest number of 

 additional yield factors to its own genotype. These genotypes could be very 

 much alike or greatly different. Apparently the two genotypes represented 

 by the lines C 110-3 and C 126-5 were greatly different both genotypically 

 and phenotypically. In ear type they seemed to be opposite extremes in the 

 range of segregation among Celaya lines. As shown in Figure 27.5, C 110-3 is 

 a line with a fairly long 8-rowed ear, and phenotypically appears to be a 

 segregant in the direction of Tabloncillo which is one of the probable pro- 



