The main objectives of this thesis were to analyse the relative roles of food availability and predation on the distribution, population dynamics and habitat selection of small mammals in Mediterranean areas of the Iberian Peninsula. Main results and conclusions Composition and abundance of small mammals communities can be completely ascertained by using two indirect methods of small mammals sampling, genet scats and barn owl pellets. Both methods reported all the small mammals species known to be present in the study area. Genet scats were used for the first time to study the composition of small mammal communitites with a biogeographical application. Small mammals species richness was negatively related to a gradient of productivity, which decreased with elevation. The increase of species density along elevation was observed at two spatial scales (plot and elevation belt). Small mammals mean abundance decreased with elevation, and this pattern was may be related to the decreased availability of resources and harsh environmental conditions with elevation. The interaction between negative feedback (density-dependence) and natural environmental disturbances (seasonal and interannual changes in cumulative rainfall) seem to be relevant forces driving the dynamics of wood mouse populations in Mediterranean forests. The role of intraspecific competition and climate were equally responsible of the strong population oscillations observed during three consecutive years. All small mammal species studied showed significant associations with gradients of short vegetation cover, either shrubs or herbs, in natural and man-disturbed habitats. Overall, species richness and abundance also showed positive associations with gradients of short vegetation cover. These patterns highlighted the dependence of small mammals on vegetation which provides food and shelter.Early successional stages supported larger small mammal abundances than unburnt forests after controlling for structural differences between plots. I hypothesised that early successional stages would be favourable habitats to small mammals due to a combination of reduced predation risk (high vegetation cover) and reduced predation pressure (less predators), and unburnt areas would be unfavourable habitats due to a combination of increased predation risk (low vegetation cover) and increased predation pressure (more predators). Fire-related fragmentation could have promoted the isolation of forest predators (owls and carnivores) in unburned forest patches, a fact that could have produced a higher predation pressure for small mammals. Behavioural responses of small mammals to predation were also surmised, with positive associations to vegetation cover in unburnt forests and negative associations in early burned habitats. These changes in small mammal-microhabitat relationships suggest differences in perceived predation risk that may have promoted the observed changes in microhabitat use. Foraging activity of mice shifted spatially and temporally in response to the presence of genets at relevant spatial scales, and decissions about how often and how long to forage were affected by the foraging activity of genets. Summarising, our findings demonstrate that predators have a prominent role in determining the patterns of distribution of mice along the post-fire succession, and that this role is mediated, at least in part, by indirect, behavioural responses of foraging mice to increased predator pressure along such succession. Effects of grazing on small mammals were mainly due to decreased availability of food and refuges for surviving harsh climatic conditions during winter, but not to food quality or increased predation risk. Taken together, our results indicate that small mammal communities in Mediterranean montane grasslands are bottom-up controlled by interspecific competition for food with cattle and by the limiting role of the availability of winter refuges, whereas top-down control due to generalists predators does not appear to be at work in these systems.