The effect of selection for combining ability within segregating lines of corn

It was possible to modify greatly the combining ability of S1 Krug lines in the subsequent selfed generations (S2 to S4) by selection from among a series of sublines within each family. Selection was based upon the top-cross performance of sublines after each generation of inbreeding. Selection of divergent high- and low-combining sublines within each family in and subsequent to S1 resulted in lines which in the S4 differed in combining ability by an average of 15 bushels per acre as measured in top-crosses. S1 lines classed as high or low in combining ability gave equal opportunity for modifying combining ability in subsequent selfed generations, though the highest yielding S4 lines were derived from the former. The mean yield of three high S1 lines in topcrosses was 98.6 bushels per acre while that of their divergent high- and low-combining S4 sublines was 107.8 and 93.3 bushels per acre in the same test. Similarly, the mean topcross yield of the three low S1 lines was 85.9 bushels compared with 94.1 and 77.6 bushels per acre for their divergent S4 high- and low-combining sublines, respectively. The situation relative to early testing as a method in corn breeding procedure may be summarized as follows: 1. Early testing of S1 lines (or S0 plants if the frequency of genes governing desirable agronomic characters is sufficiently high) will provide a better sample of material in which to inbreed than will a more random sample from the same population (8). 2. Although it has been demonstrated that the combining ability for a group of lines or families remains relatively stable from the S1 through subsequent selfed generations, it has not been stated directly nor implied, as indicated by some, that segregation for yield in the early generations of inbreeding is nonexistent. On the contrary, all published data show that the lines are segregating for factors determining yield as well as for other characters. However, it is doubtful whether the better combining segregates within families can be consistently identified without proper aids to selection for combining ability during the early segregating generations. 3. Since essentially comparable changes in combining ability may be expected in all lines during the inbreeding period, beginning with high-combining S1 lines will, in general, result in lines at fixation which are higher in combining ability than those obtained from low-combining S1 lines. In view of this fact maximum efficiency in the breeding program seems dependent upon the use of early testing, thereby permitting more intensive selection for other agronomic characters in families of known high combining ability.