Analysis of phyllometric parameters efficiency in discrimination of Croatian native <i>V. vinifera</i> cultivars

Abstract

The aims of this study were to achieve correct cultivar classification of leaf samples by using phyllometric parameters to determine the performance of parameters applied as the discriminant criteria and to determine the minimal number of leaf parameters needed to accurately classify leaf samples within cultivars. Seventy-nine phyllometric parameters were measured/calculated on 360 leaf samples from eleven grapevine cultivars, gathered during several years, from different growing conditions and with some differences in the sampling methods applied. Stepwise discriminant analysis was used to rank the phyllometric parameters according to their efficiency in the discrimination of cultivars. A series of discriminant analyses was performed with successive introduction of new parameters as discriminatory elements until 100 % correct classification of leaf samples into the correct cultivar was achieved. This was achieved using only the seven highest ranking phyllometric parameters from the stepwise discriminant analysis. Additionally, canonical discriminant analysis was performed to evaluate the differences between cultivars in the same parameters. This study represents a model for analysis of the efficiency of different ampelometric parameters for discrimination of V. vinifera cultivars.