LEAF STOMATAL TRAITS VARIATION WITHIN AND AMONG FOURTEEN EUROPEAN BEECH (Fagus sylvatica L.) PROVENANCES

In the present study, variations of leaf stomatal traits for fourteen European beech provenances (Fagus sylvatica L.), originating from six countries (Bosnia and Herzegovina, Croatia, Germany, Hungary, Romania and Serbia), were evaluated at the inter- and intra- provenances levels. Ten stomatal traits (stomatal density (SD), guard cell length (LA) and width (WB), stomatal aperture length (La) and width (Wb), pore surface of guard cell (SPSLAWB), pore surface of a widely opened pore (SPSLaWb), coefficient (SSC), potential conductance index (PCI) and relative stomatal pore surface (RSPS)) were examined using methods of univariate (ANOVA and Tukey's HSD test) and multivariate statistical analyses (principal component analysis (PCA), discriminant analysis and cluster analysis). The results obtained provide significant insight into the genetic diversity and differentiation of European beech provenances and confirmed high genetic variability within and between examined European beech provenances, according to all examined parameters. Differences at the intra-provenance level had a bigger impact on most of the examined stomatal traits (with 70-80% of total variation), while differences at the interprovenances level accounted for only 20-30% to the total expected variance. The results of PCA and canonical analysis suggest ecotypic pattern of genetic variability related to seed origin, which influenced leaf stomatal traits in F. sylvatica provenances. Forward stepwise discriminative analysis showed that WB, SD, PCI, SPSLaWb and Wb traits accounted for the same level of discrimination, as all the examined stomatal parameters together. However, the analysis detected considerably small correct allocation of the model that included all examined traits (only 38.85%). In order to discriminate examined beech provenances more precisely, additional phenotypical traits (e.g. morphological, physiological, etc.) should be included in the model.