Multi-species fruit and seed removal in a tropical deciduous forest in Mexico

We determined patterns of post-dispersal fruit and seed removal for 11 common native plants, as well as sunflower seeds (Helianthus annuus L.), within a tropical deciduous forest in Jalisco, Mexico. Removal values were high in Delonyx regia (Bojer) Raf. (90%), Crescentia alata Kunth. (87%), H. annus (81%), Pithecellobium dulces (Roxb.) Benth (81%), Albizia occidentalis Brandegee (80%), Coccoloba barbadensis Jacq. (80%), Recchia mexicana DC. (80%), Caesalpinia pulcherrima (L.) Sw. (79%), Enterolobium cyclocarpum (Jacq.) Griseb. (73%), moderate in Guazuma ulmifolia Lam. (42%) and Celtis iguanaeus Sarg. (44%), and low in Amphipterygium adstringens (Schlechtend.) Schiede ex Standl. (17%). Low removal rates (29%) in experimental patches open only to arthropods suggest that arthropods have a minor role in removing fruits and seeds of these species. Removal values were high in experimental patches open to all potential agents (77%) and semipermeable patches open to forest-dwelling rodents (76%), suggesting that the latter group, the most abundant terrestrial mammal in this forest, was an important agent for removing the fruits and seeds of study plants. Removal values were higher in experimental patches located in tropical deciduous forest than in tropical semideciduous forest, as well as higher in high density (patches of 30 fruits or seeds) than in low density (patches of five fruits or seeds) experimental patches for most study plants. Mice appear to selectively remove and hoard fruits and seeds according to their energy and nutritional content and the presence of secondary metabolites, and from from high-density food patches and preferred habitats. Nonindependent effects of species, habitat, and density suggest that a complex interplay of factors determines fruit and seed removal for the plants examined from the Chamela tropical forest.