In the breeding of rice, there is a need for an effective selection against pre-harvest sprouting. Two selection methods were tested with a hybrid population between non-dormant variety (P
1=Hokuriku 93) and dormant one (P
2=Koshihikari) in a greenhouse. during the rapid genaration advance of the breeding materials. First, F
3 seeds were classified by the germination speed into three groups, slow, moder-ate and fast. In F
4 seeds, the slowest germinating seeds were selected from the slow grop, moderately germinating seeds from the moderate group and the fastest germinating seeds Lrom the fast group. The same procedure as in F
4 Was repeated in F
5 seeds. The means of degree of pre-harvest sprouting of the three groups fairly corresponded to those of P
2, mid-parent and P
1 (Fig. 1). Second, F
3, F
4 and F
5 seeds were treated with high temperature at 70'C for 4 or 5 days, and the seeds survived were sown to raise next generation. There is a significant correlation between dormancy and to high temperature treatment. P
1 is a weakly dormant and low tolerant type and P
2 is a strongly dormant and highly tolerant type. This method was not so effective es the first one but showed a significant deviation toward the dormant parent (Fig. 1). Any adverse effect of the selection against pre-harvest sprouting was not observed on the other agronomic traits (Table 1). Correlations between pre-harvest sprouting and tolerance to high temperature treatment were observed in the F
6 Populations with P
1 and P
2 on both extreme sides (Table 2). It is inferred that the successful selection against pre-harvest sprouting was due to its high heritability. The heritability of the trait estimated by F
6 individuls and means of F
7 Iines was as high as that of culm length (Table 3). The importance and the merits of the selection against pre-harvest sprouting in early generations were discussed.
View full abstract