(go to NO.28 Contents)



In Table 1 the influence of this selection for ear length variance is shown on other characteristics of the selected M2-plants. For the three main components of yield, the variances were also higher in Group I. Accordingly, many other characteristics as well as ear length, could have been chosen to determine the induced variability. In addition, it can be seen from Table 1 that the means are generally lower in Group I, indicating the well-known tendency of any induced variability towards the negative side.

In the M3-generation, the high general variability of Group I was recovered and the selection, therefore, was successful. However, the difference between the two selection groups in respect to the variance of the ear length was lost in the M3. As another reduction of the total size of the experiment was necessary, it appeared inopportune to select again for ear length variance. Instead, from both groups as well as from the untreated control, a random sample of 16 families with 10 lines each was taken for the M4-field trial (split plot with 3 replicates in 1 x 2 lots). This procedure was considered satisfactory as the results of this trial indicated that the families giving lines with superior yield were scattered over the whole range of the M3-population.

In the M4-generation, the overall means for kernel yield within the two selection groups were still below the corresponding mean of the untreated control, though not significantly. The yield tests in M4, however, revealed several lines in each group with a significantly higher yield than the control mean. Table 2 shows that not only the number of such superior lines was higher in Group I, but also the best lines were found in this group. This demonstrates the success of our procedure of early selection on the basis of M2-variances.


<-- Back | --> Next    

(go to NO.28 Contents)