Abstract:

Based on the 5 geographical-climate factors and 18 phenotypical characteristics data of 729 trees of 24 populations of Melia azedarach in China distribution area, canonical correlation analysis(CCA)was applied for. The eigenvalue of the first canonical variable is 0.997 9 (significantly at 0.01 level), accounting for 78% of all eigenvalue, and then further principle component analysis(PCA) was done, taking the first canonical variable coordinate values as phenotypic character gradient axes, namely PCGA.The isogram of the PCGA was drawn out with 0.2 contours which showed the geographical model of northeast-southwest variation trends of phenotypial characteristics of M. azedarach. Meanwhile the path analysis results showed the direct and indirect effects of phenotypical characteristics on phenotypic character gradient values, which proved the fact that the propagative organs are changing with steady steps.

Key words: Melia azedarach, canonical correlation analysis(CCA), principle component analysis(PCA), path analysis, phenotypical character gradient axes(PCGA)