Phenotypic variation in chesnut (Castanea sativa Mill.) natural populations in Hyrcanian forest (north Iran), revealed by leaf morphometrics

Sweet chestnut (Castanea sativa Mill.) is marginally distributed – as a rare species, in Hyrcanian forest, located in the north of Iran. In Iran, this species is economically important for timber and nut production, therefore its conservation is very necessary. However, no information exists on its variability in the Iranian chestnut populations. The aim of this study is to survey the variability in leaf morphology of three chestnut natural populations. Twenty trees per a population and forty leaves per a tree were sampled, data of nine characteristics (lamina length, lamina width, petiole length, distance from leaf base to the leaf maximum width, tooth width, tooth length, tooth distance, vein (count variable), teeth (counted variable) and four characteristic ratios (leaf length/leaf width, leaf length/petiole length, leaf length/distance from leaf base to the widest point, distance from leaf base to the leaf widest point/petiole length), were recorded. Principal components analysis (PCA) was used to separate inter-relationships into statistically independent basic components. Most of the variation (85%) was explained by the first four components, and leaf size emerged as the most important variable in the corresponding eigenvectors. We used one-way ANOVA on the scores of the factors extracted in the PCA. These analyses revealed significant between-population differences with regard to most of factors. The results of discriminant analysis showed a high percentage of correctly classified cases in all actual populations (in total 93%). The patterns of leaf plasticity exhibited low values for all parameters. We concluded that leaf parameters are suitable variables for detecting levels of phenotypic variability among chestnut natural populations. The high diversity observed in the populations is very important for the conservation of the species genetic resources