Cultivar mixtures can be used to improve the sustainability of disease management within farming systems by growing cultivars that differ in their disease resistance level in the same field. The impact of canopy aerial architecture on rain-splash dispersal could amplify disease reduction within mixtures. We designed a controlled conditions experiment to study single splash-dispersal events and their consequences for disease. We quantified this impact through the spore interception capacities of the component cultivars of a mixture. Two wheat cultivars, differing in their aerial architecture (mainly leaf area density) and resistance to Septoria tritici blotch, were used to constitute pure stands and mixtures with 75% of resistant plants that accounted for 80% of the canopy leaf area. Canopies composed of 3 rows of plants were exposed to standardized spore fluxes produced by splashing calibrated rain drops on a linear source of inoculum. Disease propagation was measured through spore fluxes and several disease indicators. Leaf susceptibility was higher for upper than for lower leaves. Dense canopies intercepted more spores and mainly limited horizontal spore transfer to the first two rows. The presence of the resistant and dense cultivar made the mixed canopy denser than the susceptible pure stand. No disease symptoms were observed on susceptible plants of the second and third rows in the cultivar mixture, suggesting that the number of spores intercepted by these plants was too low to cause disease symptoms. Both lesion area and disease conditional severity were significantly reduced on susceptible plants within mixtures on the first row beside the inoculum source. Those reductions on one single-splash dispersal event, should be amplified after several cycle over the full epidemic season. Control of splash-dispersed diseases within mixtures could therefore be improved by a careful choice of cultivars taking into consideration both resistance and architecture.

Cultivar mixtures can stabilize yield and reduce pathogen spread in plant populations. A field experiment was performed to determine whether (i) a large difference between the cultivars in the mixture (e.g. plant height or earliness) would have an impact on mixture performance and whether (ii) such differences would modify the classical rules for mixture design. Mixtures were constituted from cultivars with diversity for many traits, including plant height, flowering date, disease resistance and yield potential. The field experiment was conducted in three years testing each year 72 to 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 contrasted cultivars in mixtures than was previously thought.