We examine whether different guilds of foragers remove seeds differentially according to seed quality (seed size and insect infestation) and seed location (habitat and microhabitat) in a mixed oak-beech forest. Video recordings indicated that the wood mouse (Apodemus sylvaticus) was first to encounter seeds. Foragers preferred acorns to beechnuts, large to small size and sound to infested. Nevertheless, infested seeds were removed by rodents even when sound seeds were present. Seeds that were not preferred by scatter-hoarding rodents remained longer on the ground and were more vulnerable to predation and desiccation (4% moisture loss per day). However, seeds that were removed by scatter-hoarders were moved away from their mother trees (96%) and cached individually (32%), increasing their moisture content (3% per day). Buried seeds, simulating scatter-hoarding behavior, experienced only a 17% removal after 4 months. Seed removal differences among habitats were not due to habitat attributes but to the spatial distribution of rodent-preferred microhabitats. Thus, a significant lower seed removal was observed under the tree canopy with no shrubs. However, seed removal in forest gaps with deadwood cover was not significantly different from the preferred microhabitat (under shrub cover). In pure beech forests, seed removal by rodents only occurred under Ilex aquifolium (the only perennial cover) and under woody debris. This study concludes that seed quality and seed location determine the contribution of different removers (predators vs. dispersers), their seed selection and their removal speed, leading to different seed fates which will eventually affect tree regeneration.