Analysis of line x environment interactions for yield in navy beans. 1. Components of variance

Abstract

Multi-environment yield trials of navy bean (Phaseolus vulgaris L.) lines were grown over a diverse range of locations for the years 1983–1989 in Queensland, in an unbalanced set of line × location × year combinations. This is the first in a series of 3 papers reporting different perspectives on the genotype × environment (G×E) interactions in this series of experiments. In this paper, restricted maximum likelihood (REML) estimates of G×E components of variation were derived using trial means in both standard and extended models where concomitant genotype and location factors such as maturity, disease resistance, and experimental management regimes were investigated. Prior to estimating the variance components the heterogeneity of trial error variances was modelled. Several alternative trialing systems were compared using acceptance probabilities derived from the variance components. The interaction of genotype maturity with location considerably reduced the line × location × year variance component and further examination of the maturity × environment interaction suggested an advantage in stratifying the breeding program on maturity. There is no redundancy in the current trialing system and an increase in sampling locations can be justified.