For a mixture to be effective, it should incorporate functional diversity. One measure of functional diversity is divergent response to a series of environments. Here, environments can be defined in the usual way of a location within year, or in terms of the composition of a pathogen population with which a crop is confronted. Using the former definition, a variety's functional profile is determined by its responses to environments relative to a reference population. This is is GxE reaction. Using the latter definition, a varietie's functional profile is determined by the set of pathogen isolates to which it is resistant. In both cases, a functionally diverse mixture can be obtained by mixing varieties whose functional profiles have a low -- hopefully negative -- correlation. We illustrate this idea with using the resistance specificities of a set of oat cultivars to oat crown rust and by assessing the performance of oat mixtures assembled either to maximize or minimize the covariance of their GxE reactions.