Evaluation of chickpea genotypes for resistance to Ascochyta blight (Ascochyta rabiei) disease in the dry highlands of Kenya

Abstract

Chickpea (Cicer arietinum) is an edible legume grown widely for its nutritious seed, which is rich in protein, minerals, vitamins and dietary fibre. It’s a new crop in Kenya whose potential has not been utilized fully due to abiotic and biotic stresses that limit its productivity. The crop is affected mainly by Ascochyta blight (AB) which is widespread in cool dry highlands causing up to 100% yield loss. The objective of this study was to evalu- ate the resistance of selected chickpea genotypes to AB in dry highlands of Kenya. The study was done in 2 sites (Egerton University-Njoro) and Agricultural Training centre-ATC-Koibatek) for one season during long rains of 2010/2011 growing season. Thirty six genotypes from reference sets and mini-core samples introduced from ICR- SAT were evaluated. There were significant (P<0.001) differences in AB responses and grain yield performance in test genotypes in both sites. AB was more severe at Egerton-Njoro (mean score 5.7) than ATC-Koibatek (mean score 4.25), with subsequent low grain yield. Genotypes ICC7052, ICC4463, ICC4363, ICC2884, ICC7150, ICC15294 and ICC11627 had both highest grain yield in decreasing order (mean range 1790-1053 Kg ha-1) and best resist- ance to AB. Further evaluation is needed in other multi-locations and their use in breeding program determined especially because of their undesirable black seed color. Commercial varieties (LDT068, LDT065, Chania desi 1, and Saina K1) were all susceptible to AB, but with grain yield >1200 Kg ha-1. The findings of the study showed that chickpea should be sown during the short rains (summer) in the dry highlands of Kenya when conditions are drier and warmer and less favorable for AB infection. However yield could be increased by shifting the sowing date from dry season to long rain (winter) thus avoiding terminal drought if AB resistant cultivars with acceptable agronomic traits could be identified.