Cultivar mixtures can stabilize yield and reduce pathogen spread in plant populations. A field experiment was performed to determine (a) whether a large difference between the cultivars in the mixture (e.g., plant height or earliness) would have an impact on mixture performance, and (b) whether such differences would modify the classical rules for mixture design. Mixtures were constituted from wheat cultivars with diversity for many traits, including plant height, flowering date, disease resistance, and yield potential. The field experiment was conducted over 3 years, testing each year 72–90 mixtures of two, four, or eight cultivars, and their corresponding pure stands. Disease severity and yield of cultivar mixtures were strongly related to the mean values of the component cultivars in pure stands. Despite the considerable diversity of the mixtures tested, the classic rules (e.g., proportion of susceptible cultivars) already tested in mixtures with similar height and earliness were effective for decreasing disease severity. Agronomic heterogeneity for traits such as plant height, yield potential, or earliness of the cultivars in mixtures did not have a negative impact on disease severity and yield relative to pure stands. Increasing the number of cultivars in the mixture from two to eight had no impact on the mean disease severity and yield of the mixtures, but reduced the variability of disease severity and yield in the mixture relative to pure stands. These results suggest that it may be possible to increase within‐field wheat diversity by combining more contrasting cultivars in mixtures than was previously thought.