Evaluation of an Italian germplasm collection of Festuca arundinacea Schreb. through a multivariate analysis

Abstract

Forty-eight accessions of tall fescue (Festuca arundinacea Schreb.), collected from Northern to Southern Italy, were analyzed using both univariate and multivariate analysis applied to a set of 12 quantitative traits. Four principal components were found to explain 77% of the total variation in the dependence structure. Productivity characters together with heading time and dimensions of flag leaf appeared to be the major sources of diversity among tall fescue populations. On the basis of the 4 principal components similar populations were clustered according to minimal distance analysis. Seven clusters were identified. The results of cluster analysis confirmed the presence of a remarkable diversity within the germplasm collection and explained why results of a univariate analysis of variance did not reveal significant differences among groups of accessions collected in Northern, Central and Southern Italy. The multivariate approach seemed to point out a problem of genetic erosion of the local germplasm in the Po Valley (Northern Italy) and, on the whole, appeared to be a valid system for tall fescue germplasm evaluation.