Phenotypic character gradient variation of Melia azedarach

Abstract

Canonical Correlation Analysis (CCA) was applied on the research data of five geographical-climatic factors and 18 phenotypic characters of 729 trees of 24 populations of Melia azedarach distributed in China. The eigenvalue of the first canonical variable is 0.997 9 (significant at 0.01 level), accounting for 78% of all eigenvalues. A study on the principal component analysis (PCA) was done, taking the first canonical variable coordinate values as the phenotypic character gradient axes (PCGA). The isogram of the PCGA was drawn out with 0.2 contours, which showed a geographical model with a northeast-southwest variation trend of the phenotypic characters of M. azedarach. Meanwhile, the path analysis results show the direct and indirect effects of phenotypic characters with phenotypic character gradient values, which prove that the propagative organs, are steadily changing.