A study of the ecological variability in yield of selected clones of Arabica coffee (Coffea arabica L.) from the coffee-breeding programme in Kenya

Cherry yield of eleven elite clones of Arabica coffee was evaluated over a five-year period at three locations in a continuing study to determine the extent and influence of Clone X Environment interactions on coffee breeding lines developed in Kenya. Clone mean yield comparisons indicated a better performance of all the improved clones compared with the Standard. In analysis of variance (ANOVA) significant main effects and non-significant Clone X Environment interactions (P < 0.05) were found. Interactions were however, further investigated using ecovalence values, the Eberhart and Russell (1966) regression model as well as the Additive Main effects and Multiplicative Interaction effects (AMMI) model. The clones were then clustered according to their response patterns using principal components and cluster analyses to obtain a delineation of ecological districts and more accurate predictions of clone performance. The main effects for clones reflected breeding advances with the standard variety having the smallest mean. Improved materials from the hybridization scheme appeared to do better than their pure-line counterparts with clones 8, 3, 1 and 2 being the most advanced of the 11 clones. Similarly, the main effects for the environment reflected the overall site/year quality with Kiambu having the best production conditions particularly due to the favourable weather in 1983. This indicates that the coffee growing upper zone may be favourable for the clones especially when they are raised by top-working old coffee. Makuyu environments had the least interaction effects.