The yam bean (Pachyrizhus spp) was recently introduced as a root crop with high-yield potential, considerable protein and micro-nutrient concentration to investigate its potential for food production in Rwanda. Except for Chuin types (Pachyrizhus tuberosus) which have high storage root dry matter (RDM) (26 to 36%), most accessions are consumed raw and are reported to have low RDM. The present study aimed to evaluate and identify adapted high yielding yam bean accessions in major agro-ecological zones of Rwanda. Field experiments with 22 accessions were conducted in 2012 at three research sites representing the major agro-ecologies of Rwanda. Strict reproductive pruning was followed to enhance fresh storage root yields. Across locations, ANOVA indicated highly significant differences (p < 0.01) for genotypes (G), locations (L), seasons (S) and G x L effects for storage root yield, vine yield and harvest index and accounted for 21.88%, 43.41%, 1.43% and 13.25% of the treatment sum of squares, respectively. The GGE bi-plot revealed that EC209018 is high yielding but unstable. However, genotypes, AC209034, AC209035 and EC209046, were outstanding in terms of adaptation and relative stability across the 3 locations, suggesting consistent root yields irrespective of location and environmental conditions. The GGE scatter plot showed that all genotypes formed one mega-environment for storage root yield (Karama, Musanze and Rubona) and two mega-environments for biomass yield (Karama and Rubona as one mega-environment and Musanze the second one). This study revealed that Karama is the most suitable environment for evaluation and selection of yam bean for yield components in Rwanda.
Keywords: GGE bi-plot; GxE; Harvest index; High storage roots yield; Pachyrizhus; Rwanda.