Gains from selection of broadly adapted and specifically adapted sugarcane families

Abstract

Gains from selecting sugarcane families exhibiting specific and broad adaptation were examined in a region where genotype×site interactions are known to be large. The families were firstly evaluated across a range of sites. Families exhibiting different types of adaptive response were selected, and random clones from each selected family and from random families were evaluated in a further sample of environments in the same region. Of particular interest was the determination of how well the average yields of the families across all sites predicted yield of progeny clones at different sites, compared with the predictive value of family yields at specific sites. General selection of families on the basis of broad adaptation (ie. based on mean yield across all sites) gave the best or equally best gains from selection in most environments. In considering clone performance at individual sites, in most cases, genetic correlations with (previously measured) average family yield across all sites were as large or larger than correlations with (previously measured) family yields at individual sites. Consistent with this result, families selected on the basis of broad adaptation also provided a greater proportion of the most elite clones in each progeny trial. Some possible reasons for the superiority of families selected on the basis of broad adaptation are suggested. Overall, the findings suggest that selecting families on the basis of broad adaptation would result in better gains than selecting for specific adaptation to individual sites in regional sugarcane selection programs targeting a similar range of environments as used in this study.