Assessing the impact of habitat fragmentation has become a central issue in conservation biology. Ecologists are urged to document and predict the ability of species to maintain viable populations in fragmented areas. Rarity in fragments is commonly interpreted as an indication of sensitivity to habitat fragmentation because fragmentation is thought to act as a ‘‘filter’’ impeding the movement or threatening the sustainability of sensitive species. By focusing on two common frugivorous bats in rainforests of French Guiana, we show that sensitivity as measured by abundance may not be consistent with sensitivity as measured by modifications in population structure (sex-ratio) or in physiological state (hematocrit level). Other mechanisms, such as competitive exclusion, may interfere with the filter model. In particular, highly mobile animals like bats and birds have the potential to rapidly modify their distribution as a function of competition. Special attention should be paid to the development of alternative methods to better assess species’ fragmentation-sensitivity