Biometrical Letters vol. 47(1), 2010, pp. 15-32

The relationship between 14 genotypes of black poplars (Sect. Aigeiros Duby) was analyzed with regard to the variability of sixteen early rooting characteristics and cutting survival. Principal component analysis (PCA) and cluster analysis were used in order to reduce the amount of data and to enable the presentation of relationships among the genotypes. Besides normal standardization, two alternative methods of standardization were used: standardization by standard deviation within genotype, and by standard deviation within genotype followed by indexing with the correlation between characteristics and cutting survival. Alternative standardization methods were used in order to emphasize the contribution of genotype to the effect of differences among genotypes on total variation and relationship of rooting characteristic with cutting survival. To preserve the effects of standardization, the principal component analysis was based on a matrix of covariances. Contribution to total variation rose from the first to the third method of standardization for the first principal component (PC), while the trend was the opposite for the second one. The distances between genotypes were more related to the survival of their cuttings than after normal standardization. There were no significant differences in loadings with the first two PCs among the considered types of standardization. The characteristics of the shoot and those describing rooting in the upper half of the cutting displayed the highest correlation with the first PC, while the characteristics with high loadings with the second PC mainly describe rooting on the basal part of the cutting.

principal component analysis, data standardization, nursery and plantation establishment