Published online by Cambridge University Press: 02 September 2010
1. The data relating to 16 multi-location trials involving several varieties of egg-laying chickens per trial have been analysed to obtain estimates of variety × location (genotype-environment) interactions for rearing and laying mortality, hen-month and hen-housed egg production and 32-week egg weight. Significant interactions were found in about a third of the trials for all five traits though in only two cases did the interaction variance account for more than 20% of the total variance.
2. Genetic correlations for varieties between locations have been calculated as 0·49 for rearing mortality, 0·50 for laying mortality, 0·72 for hen-month production, 0·76 for hen-housed production and 0·91 for egg weight. These are in close agreement with previous published estimates.
3. The between-and within-variety components of variance have been calculated for each trait on each location separately. The regressions within trials and pooled values of the components on the mean level of all varieties on the location have been calculated. For hen-month production the pooled regres-son of the within-variety component on mean was negative, whilst for all other traits the pooled regressions were positive. The pooled regressions of the between-variety components on mean were all positive. It is suggested that this information indicates that comparison of varieties for single traits should be carried out on locations favourable for maximum expression of the trait. The difficulties which variety comparisons, for several traits having different optimum management regimes, would cause and the need for a more intensive investigation of the conditions conducive to genotype-environment interactions are stressed.