Stability analysis of grain yield of tef (Eragrostis tef) using the mixed model approach

Abstract

Multi-environment trials are usually used to assess the grain yield stability of genotypes. Most of the common stability measures such as Finlay and Wilkinson (1963), Eberhart and Russell (1966) and Shukla (1972), can be accommodated by a linear mixed model where genotypes are a fixed effect and environments are a random effect. The objective of this paper was to examine the stability differences of tef genotypes using the mixed model approach. Yield stability of 13 tef genotypes was studied based on data collected over 12 environments where an environment constitutes a particular location and year combination. The combined analysis of variance from the 12 environments revealed significant differences among genotypes, environments and their interaction effects on the grain yield indicating the differential responses of genotypes to diverse environments. The likelihood ratio tests for Finlay-Wilkinson and Shukla models showed that the stability of tef genotypes was best assessed by an Eberhart-Russell model. The mixed model analysis showed that genotypes DZ-01–974, DZ-01-D12, DZ-01–58, DZ-01–795 and DZ-01–35012bs were the most stable genotypes with small Eberhart and Russell variance of deviations. Genotypes DZ-Cr-358, DZ-01–1465, DZ-01–1576, DZ-01–1809 and DZ-01-Dg were the unstable genotypes that had high mean yield.