Mixing different herbivore species is assumed to increase vegetation use because of the complementarity of their feeding choices and to reduce animal parasite burden through the dilution effect. Here, we compare the effects of mixed horse-cattle grazing and monospecific horse grazing (1.4 LU/ha) on animal foraging behaviour, sward characteristics and horse parasitism in a mesophile grassland of central France. In both treatments, animals were stocked alternately on two subplots with rotation lengths between 15 and 21 days according to season. The horses quickly acclimated to the cattle, and very few agonistic interactions were observed between them. All the horses selected short (<= 4 cm) and intermediate (5-8 cm) high-quality regrowths and avoided reproductive and dead herbage areas contaminated by their faeces as a consequence of their latrine behaviour. Cattle, which are more constrained by plant height, selected intermediate and tall vegetative swards. However, the alternate stocking of animals also provided them with high-quality regrowth on the shortest patches. Consequently, the cattle used these short patches proportionally to their availability while avoiding reproductive and dead herbage areas. We assume this limited their consumption of infective cyathostome larvae, which are concentrated close to horse dung. Moreover, co-grazing horses and cattle did not reduce sward structural heterogeneity and thus did not enhance overall herbage quality. We conclude that rather than considering mixed grazing as a turn-key solution, its management needs to be adapted to support the complementarity of horse and cattle dietary choices and thus provide the expected benefits of multi-species grazing.