Abstract
The objectives of this study, variation analysis of leaf characters among the natural populations of KaIopanax septemIobus Koidz. could be used for the conservation of gene resource and provide information to mass selection. The following results were obtained; For the study of the variation of leaf characters among natural populations of K. septemhbus in Korea, 10 populations were selected. ANOVA tests showed that there were statistically significant differences in all the leaf characters among populations as well as among individual trees within populations. Correlation analysis revealed that lobe width characters have highly positive relationship with Leaf blade length(LBL), Petiole length(PL), and Upper lobe angle(ULA) had negative related with LBL, lobe width and petiole characters. The result of principal component analysis(PCA) for leaf characters showed that the first for principal components(PC's) explained about 79% of the total variation. The first PC was correlated with those characters that mainly related with Maximum lobe width(MAW), Minimum lobe width(MIW), LBL, PL, Lower lobe width(LLW). The second PC was correlated with the ratio of LLW to LBL, No. of leaf lobe(NLL) and the third PC was correlated with the ratio of Middle lobe width(MLW) to LBL, the ratio of Upper lobe width(ULW) to LBL. The fourth PC was correlated with the ratio of PL to LBL. Therefore, these characters were important to analysis of the variation of leaf characters among natural populations of K. septemtobus in Korea. Cluster analysis using average linkage method based on leaf characters showed that natural 10 populations of K. septemlobus in Korea could be clustered into five groups. Group I consist of Mt. Sori, and Group II comprises Mt. Kaji and Mt. Jiri. Group III contains Muan and Mt. Halla, Group IV consists of Mt. Balwang, Suwon, and Mt. Worak, and GroupV comprises Mt. Chuwang and Mt. Kyeryong. Especially, population Mt. Sori was distinct from other populations. This result corresponded well with that of principal component analysis.