PHENOTYPIC VARIATION AND POPULATION RELATIONSHIPS OF CHESTNUT (CASTANEA SATIVA) IN GREECE, REVEALED BY MULTIVARIATE ANALYSIS OF LEAF MORPHOMETRICS

Data regarding four leaf size parameters (leaf length, leaf width, petiole length, distance from leaf base to the leaf widest point), and four shape parameters (leaf length/leaf width, leaf length/petiole length, leaf length/distance from leaf base to the leaf widest point, distance from leaf base to the leaf widest point/petiole length), were recorded in six chestnut populations (three managed coppice, two old-growth natural and one fruit orchard population) originating from three different geographic localities.
Ten leaves per tree and 27 trees per population were sampled.
Univariate and multivariate analysis were carried out on a by population basis.
Analysis of variance revealed significant population differences, which were extended in most parameters as found via t-tests.
Multivariate methods involved both ordination and classification approaches (principal component analysis, multivariate analysis of variance and multiple discriminant analysis). Leaf size parameters emerged as most important variables in the corresponding eigenvectors.
Most of the variation (>85%) was resolved in low multidimensional space; nevertheless separation of populations in the first three components was poor.
Redistribution results to original classes using the linear discriminant function presented a high misclassification of return (34%). It is concluded that leaf parameters may be suitable variables in order to detect levels of phenotypic variability among populations, nevertheless do not possess adequate discriminative capabilities to distinguish populations and population types.

III International Chestnut Congress

F.A. Aravanopoulos

leaf morphometrics, discriminant analysis, PCA

https://doi.org/10.17660/ActaHortic.2005.693.28