Optimizing breeding zones: genetic flexibility or maximum value?

When the performance of genotypes vary across an environmental gradient in a non-linear manner the matching of genotypes and sites is a complex problem. To overcome this breeding zones: aTe usually allocated so that sets of genotypes are used across different parts of the range to maximise production across all anticipated planting sites. Two different approaches to determining the number and constitution of these populations and the demarcation of their planting zone boundaries are contrasted. Both approaches use non-linear models to define the relationship between relative performance of a genotype and an environmental gradient. The approaches differ however, in the criterion used for choosing the populations. One approach will lead to a guaranteed minimum yield at each site whilst the other approach ma:lllmlSeS the expected yield over the total range of planting sites.