多變量 解釋法에 依한 벼의 品種群 分類 및 이들 品種群間의 組合能力

Varietal Classification on the Basis of Multivariate Analysis and Combining Ability between Each Varietal Groups in Paddy Rice

An analysis of genetic divergence as measured by Mahalanobis’s D² statistic in 4005 possible comparisons among ninety varieties of paddy rice for eight characters, namely days to heading, culm length, panicles per hill, spikelets per panicle, a thousand grain weight, fertility ratio and grain weight per hill, and a single link cluster analysis based on the D²s were attempted. And the heterosis and combining ability analysis were performed on the F₁ hybrids of the two diallel cross sets of inter-and intra-varietal groups classified by the Mahalanobis’s D² and cluster analysis. The results are summarized as follows; 1. The varietal groups based on eight characters were most distinctly discriminated in camparison with the classifications based on three, four, and six characters using in order of eight characters mentioned above, respectively. However, the pattern of varietal classification based on six characters was nearly similar to that of eight characters. 2. Ten varietal groups which have minimum four varieties were classifed by the single link cluster analysis based on the D² computed in eight characters. The averages of inter-group D²s were greater than those of intra-group D²s. 3. Most of the varietal groups were separated according to the breeding origins and classified distinctly into early and late groups, long and short groups, and Indica and Japonica groups. 4. Days to heading and culm length among the eight characters were the most contributors to the D² in the inter- and intra-groups. 5. Dendrogram including all of the parents and F₁ genotypes from the inter- and intra-group diallel cross sets revealed that the parents and F₁ genotypes of infra-group were classified into the same group, and RD4 and it’s F₁s were widely distributed. 6. The heterosis over higher parents in the most of F₁ hybrids were significant for the days to heading, culm length, spikelets per panicle and a thousand grain weight in the inter-group diallel cross, and spikelets per panicle and grain weights per hill in the intra-group diallel cross. 7. The mean squares of general combining ability(GCA) and specific combining ability(SCA) for all characters except SCA of panicles per hill were significant in the inter- and intra-group diallel cross. The mean squares of GCA were greater than those of SCA for all characters except days to heading and fertility ratio in inter-group diallel cross and days to heading and fertility ratio in inter-group diallel cross and day to heading and grain weight per hill in intra-group diallel cross. 8. GCA and SCA effects computed in inter-group diallel cross were greater than those of intra group diallel cross. GCA effects of RD4 and Bulu & Gungil were positive for the most character, however, those of IR2061 were negative. 9. In the analysis of variance of Wr─Wr, the mean squars of array for days for days to heading, panicle length, a thousand grain weight, fertility ratio and grain weight per hill were significant, which suggest that varieties invoived in the inter-gruop diallel set may have non-allelic interactions for above five characters. 10. In the inter-group diallel cross analysis, the culm length and the spikelets per panicle were nearly complete dominant, however, the panicles per hill was predominated by the additive effects. taichung 164, CI9805 and Bulu & Gundil appeared to have many dominant genes for culm length, and Bulu & Gundil and CI9805 for the spikelets per pancile according to (Vr, Wr) graph.