Abstract
In many cases for the cross-pollinating plants, estimating of heritability and genetic correlation is based on parent-offspring regression, assuming mate being completely random. But this assumption is questionable in the case of usual condition. Therefore, the actual changes of the three strains (E, L and R) were compared with the expected changcs. The mother plants of E and L strain were respectively selected such that the plants were randomly selected from phe earlier or later part of the 'original population. R strain were randomly selected over all population. The performances of the mother plants and their pedigrees are shown in Table 1 and 2, respectively. The expected values of change by selection were obtained according to I.M. Lerner. The expected value of change by selection is composed by two parts, direct response and correlated response, and they are shown in Table 4. The direct response is deviation for a character under selection and the correlated response is deviation for an unselected character correlated with a character under selection. The expected changes did not completely agree with the actual changes but both values were resemble . Accordingly, for estimatingr of change by selection, heritability and genetic correlation above mentioned are useful. In this experiment, the character under selection was date of lodging and for this character selection was effective. In the other hand, shape of bulb significantly changed by selection for date of lodging, that is, the early strain (E) became to be flatter and the late strain (L) giober. These performances were able to predict from the expected values, too.
