Abstract

A set of twenty four cowpea (Vigna unguiculata L. Walp) genotypes were evaluated in field conditions for three consequtive seasons based on 13 quantitative characters. A combined analysis of variance revealed a highly significant difference among the genotypes for most of the traits (P <0.001). Introduced genotypes performed better in nine of the characters, revealing the advantage of introduced genotypes. Principal Component Analysis explained a total variation of 76.16%, where the first two PCs accounted for 51%. Most variation was largely dependent from rainfall use efficiency, days to maturity, seed yield per hectare and pod width. Cluster analysis dendrogram delineated the 24 genotypes into two major clusters, with a mixture of local and introduced in both groups which indicated their potential relationship. Exceptional local genotypes B137B, with highest seed yield (866 kg/ha), and the singleton B342 with a number of desirable agronomic traits must be included in future cowpea breeding programs in Botswana.

Key words: Cluster analysis, rainfall use efficiency, SPAD reading, Vigna unguiculata.