Additive mean effects and multiplicative interaction analysis of maize yield trials in South Africa

Abstract

Thirty five maize hybrids were evaluated over 98 environments for genotype x environment interaction (GEI) and yield stability under dry land conditions during the 2003 to 2006 seasons in the eastern and western maize production regions of South Africa. The Additive Mean Effects and Multiplicative Interaction (AMMI) statistical model was used to describe the GEI, yield stability and adaptation to certain environments. This model has shown to be effective in understanding complex GEI’s in the national cultivar trials. The AMMI 2 combined analysis of variance (ANOVA) indicated highly significant differences among hybrids and environments as main effects. The environment and GEI contributed more towards the total variation. This demonstrated a high influence of the environment on the yield performance of maize hybrids in both regions. CRN 3604 and PAN 6611 were the most stable hybrids in the eastern region while PAN 6966 and PAN 6146 were the most stable hybrids in the western region. DKC 80–12B and DKC 80–10 were the best adapted hybrids for high potential environments in the eastern region while DKC 78–15B and DKC 80–12B were the best adapted hybrids for high potential environments in the western region. Caracal and to lesser extend Saffier were the best adapted hybrids for low potential environments in both the eastern and western regions.