Improvement of <i>Vitis vinifera sativa</i> D. C. taxonomy

Abstract

A method of mathematical statistics for evaluation of the taxonomic usefulness of characters included in the Vitis vinifera sativa D. C. cultivar clustering has been developed. The method works as follows: The irregularity of the distribution of characters in taxons is calculated, and the probability of achieving such estimates is calculated when all taxons are assigned to the same general population after a certain character. The method does not depend on the type of the distribution of characters and is computer-programmed (a CM-4 computer). Using the method, it is possible to conclude that the higher the taxon hierarchy, the smaller is the amount of taxonomically useful characters the taxons differ in, i.e. in order to differentiate taxons of lower hierarchy a smaller amount of descriptive characters is necessary. Var. transcaucasica and var. mediiasica have been isolated in the taxon convar. orientalis subconvar. caspica NEGR. and in the taxon subconvar. antasiatica NEGFL, while the taxon convar. pontica NEGR has been supplemented with subconvar. meridionali-balcanica and georgica-caspica along with subconvar. balcanica and georgica NEGR.
Using discriminant analysis, assignment of many cultivars to proper taxons has been verified. The phenogenetic proximity of taxons has been established. Some problems of the theory of evolution and those of practical breeding are discussed.