Additive Main Effects and Multiplicative Interaction Analysis of Arabica Coffee (Coffea arabica L.) for Yield

The southern region of Ethiopia is the second largest producer and supplier of Arabica coffee in the country and shares 46% of the national market. Coffee (Coffea arabica L.) grows in all parts of the region particularly in Gedeo, Sidama, Bench Maji, Shaka, Kembata-Tembaro, and Gamo Goffa administrative zones. A genotype by environment interaction experiment was conducted on 43 genotypes of Arabica coffee using RCB design with 4 replications, 2m by 2m spacing and 10 trees per plot, in two agro-ecologically contrasting locations (Dale and Yirgachefe districts) of southern Ethiopia region. The study evaluated genotype by environment interaction effect of coffee genotypes with the objective of selecting the best performing and adaptable genotype. Yield data was collected for 4 consecutive years in each location and analysis was computed in order to predict yield based on the selected model. AMMI analysis result showed a very high significant difference (p<0.0001) of the environment, genotype and GxE interaction components. The first two principal components explained 74% of the GxE interaction component. This indicated reasonable and parsimonious AMMI model with the first two interaction principal component axes in partitioning the treatment sum of squares effectively. The AMMI bi-plot of mean yield of the 43 genotypes showed that genotypes 1, 2, 3, 9, 32, 12, 25, 31, 10 and 29 were stable and genotypes 33, 4, 23, 34 and 27 were most stable. In general, genotypes 1, 9, 2, 3, 32, 12 and 25 had the best performance, and genotype 1 had the highest yield and was acceptably stable. The AMMI analysis was effective in identifying the best performing genotypes which the breeding program could advance.