HETEROSIS IN A NEW POPULATION 



447 



of the So or S3 lines in each case, however, was not based on test crosses for 

 specific combining ability with the other line involved. The advanced lines 

 used were picked from among their sister lines largely on the basis of desir- 

 able agronomic characters. It may be seen in Table 27.15 that, although 

 in many instances the differences were small, the yield of the crosses of S2 

 or S3 lines exceeded that of the corresponding cross with Si lines in every 



TABLE 27.15 



YIELDS OF SINGLE CROSSES BETWEEN 

 Si LINES COMPARED TO YIELDS OF 

 SINGLE CROSSES BETWEEN TWO 

 LINES* 



N = 14. 



Mean difference = 517.6 + 147 kilos (or 



9.2 per cent). 



* Derived from the respective Si's after two or three 

 generations of inbreeding and selection for agronomic char- 

 acters only. (Kilos per hectare.) 



case except two. The average difference between the fourteen paired crosses 

 was 9.2 per cent. It is highly probable that a greater increase would have been 

 obtained had the various S2 or S3 lines also been picked on the basis of tests 

 for specific combining ability. However, since selection can make a choice 

 only between the alleles present in a particular Si, a point of diminishing re- 

 turns may be rapidly reached upon straight selling. Experiments are under 

 way to determine in what generation of selfing maximum gains may be 

 reached. 



The data in Table 27.14 are of further interest from the standpoint of rela- 

 tionship between the combining ability of Si lines and the advanced genera- 



