Abstract
The heading habit and seasonal growth of twenty six clones of tall fescue were investigated for three years at two locations, Nishigoshi and Aso. In the statistical analysis all effects, except reprications, were considred fixed. The analysis of variance revealed significant differences among genotypes for heading date and scores of vigor in every seasons. Also, the interactions of genotypes×locations and those of gentypes×locations×years were significant for heading date and vigor in every season (Table 3,7). Principal component analysis was applied to the genotype×environment interaction matrix. In the result of principal component analysis for heading date in six environments, first component scores mean the yearly fluctuation of genotypes in Nishigoshi, and second ones mean those in Aso (Table 5, 6). In the case of vigor in 20 environments, first component shows genotypes×locations interaction, second one shows the seasonal growth of genotypes, and third one appears to indicate genotypes×locations inteaction in the relation between vigor in initial growth stages and that of perennial growth stages (Table 9). The clones which showed good growth at both locations were originated in the populations of Kyushu or the cultivars bred in U.S.A. (Fig. 2). Two clones originated in Algerian populations showed good growth at Nishigoshi and poor growth at Aso (Fig. 1). We need to investigate adapatability and stability of tall fescue clones from the results of evaluation tests for several years at two or more locations. The clones which exhibited remarkable interaction of genotype×environment on heading habit might cause to drift on the genetic characters of synthetics presupposing random mating amang the clones. Whereas the clones which exhibited interaction on vigor might improve stability of synthetics.
