EVALUATION OF SEED YIELD CHARACTERISTICS IN 29 ACCESSIONS OF COCKSFOOT (DACTYLIS GLOMERATA) THROUGH A MULTIVARIATE ANALYSIS
In order to study the genetic variation for seed yield and its components among 29 accessions of cocksfoot (Dactylis glomerata L.), an experiment was conducted under spaced plant condition over two consecutive years in Karaj, Iran. Principal component analysis and cluster analysis was used to group the accessions. Analysis of variance showed significant differences among accessions for heading date, plant height, fertile tiller number, seed yield, panicle length, seed yield panicle-1, seed number panicle-1, thousand grain weight, harvest index and forage dry matter yield.
Using principal component analysis, the most important variables for seed yield characters were identified. First three independent components accounted for 72% of total variation. The first principal component indicates that seed yield, harvest Index and seed number panicle-1 were important characters for classification. Plant height and forage dry matter yield were important characters in the second component. Based on ward cluster analysis, 29 entries were divided into 5 groups. Accessions in clusters 5, and 3 averaged well above the overall mean for forage dry matter yield and seed yield, respectively. Majority of accessions in cluster 1 and 2 were originated from different parts of Iran, whereas, most of accessions in other clusters were originated from USA and European countries. Accessions from a defined geographic area tended to cluster, but some accessions from particular area spread among several cluster.
Keywords: Dactylis glomerata, seed yield, seed components, principal component and cluster analysis.