Multilocational evaluation of selected local and improved cowpea lines in Uganda
Six improved cowpea (Vigna unguiculata) genotypes initially selected from lnrernational lnstitute of Tropical Agriculture collection were evaluated for 3 seasons at 4 locations in Uganda. with the objective of comparing their yield performance and assessing their adaptability under Ugandan conditions. The cowpea genotypes were MU- 93, TVX 337-025 (B) IT90K-109 (B), KVU-12349, IT82D-516-2, and IT85 F-2841. Across locations and seasons, MU-93 produced highest grain yield ( 1586 kg ha-1) while KVU-12349 produced the lowest grain yield (505 kg ha-1). At different locations the genotypes' performance varied, whereby IT85F-284l (1378 kg ha-1, MU-93 ( 1208 kg/ha), MU-93 (1235 kg ha-1 and IT90K-109(B) (2373 kg/ha) produced the highest grain yields in Kabanyolo, Pallisa, Kumi, and Kabaramaido, respectively. The Additive Main effects and Multiplicative Interaction (AMMI) model was used in the genotype by environment (GXE) analysis to determine the yield and stability of the genotypes in the multi - environment trial. The effect of environments, genotypes and GXE were highly significant (P<O.OO I ). The biplot revealed that the varieties; IT85F-2841 (G I), IT90K-1 09(B) (G2), 1T82D· 516-2 (G3), ·TVX 337-025'(B), (G4) and MU-93 (G6) were generally high yielding since AMMI placed them on the right hand side of the mid point of the axis representing the grand mean, Ebelat (G5), Icirikukwai (G7)' and KVU-12349 (G8) were generally low yielding and were placed on the left hand side of the midpoint on biplot. IT90K-109 (B) (G2) and lcirikukwai (G7) had the highest principal component analysis (IPCA) scores ( 15 .4 and -22.7, respectively) therefore were considered as unstable genotypes. Genotypes IT85F-2841, MU-93, IT82D- 516-2, KVU- 12349 and Ebelat had moderate IPCA 1 scores; hence moderately stable. TVX337 -025 (B) had a low IPCA 1 score and thus was relatively a stable genotype. Within environments AMMI and unadjusted means ranked genotypes differently. However, at one location in 2003, 6 out of 8 genotypes were ranked similarly by both estimates. WhileAMMI alone ranked MU-93 as the best genotype in all environments, both estimates reveal that MU-93 is superior in 6 out of 12 environments. MU-93 has now been recommended for official release in Uganda.
Key Words: Adaptation, AMMI model, G X E interaction, Vigna unguiculata