Habitat loss and fragmentation are key drivers of global species loss. In fragmented landscapes species must persist in small, isolated and often degraded habitat patches where they can be subject to high risk of extinction due to deterministic and stochastic forces. Species respond to habitat fragmentation according to species-specific life-history traits, with habitat generalist, edge or mobile species being less impacted compared to specialists and less mobile species. The impact of habitat fragmentation on species and their consequent probability of persistence depends on a series of key, concatenated events occurring at different biological and spatial scales. The response of single individuals to landscape change can translate into effects at the level of populations; coexisting species can reciprocally influence their responses through the alteration of interspecific relationships; inter-population dynamics can also occur, involving the movement of individuals between populations in different habitat fragments and affecting the persistence of entire systems of populations. Given the complexity of factors involved, including direct and interacting responses, it is extremely difficult to understand the actual effects triggered by habitat fragmentation without a thorough knowledge of the underlying ecological mechanisms. The aim of this PhD project was to contribute to understanding the mechanisms underlying the response of species to habitat fragmentation. By following a holistic approach, I used a set of mechanistic field studies on four rodent species specifically designed to investigate the series of key events involved in the persistence of species in fragmented landscapes: 1) Population and individual scale responses of small mammals to patch size, isolation and quality. The aim of this section was to determine the relative effects of landscape structure (habitat amount and configuration) and patch quality (here measured as abundance of shrub resources) on individuals (survival and litter size) and populations (density and colonization/extinction dynamics). A large-scale demographic field study was conducted, encompassing 30 woodland sites nested within three landscapes and surveyed monthly for three years by means of a capture-mark-recapture protocol. Model species was an arboreal rodent, the hazel dormouse (Muscardinus avellanarius), known to be sensitive to habitat loss and fragmentation. Habitat quality influenced populations at different biological scales by concatenated effects: it enhanced individual survival, increased the chances of colonizing vacant patches and sustained higher population densities. It was therefore related to the performance of single populations and systems of populations through re-colonization dynamics. Habitat quality, however, did not influence local extinction probability, which was ultimately related to the extent of available habitat, likely due to the absolute size of populations: a high absolute number of individuals reduces the chances of population extinction. 2) The role of interspecific interactions in shaping small mammal communities in fragmented landscapes. The aim of this section was to evaluate the strength of interspecific interactions as a shaping force of animal communities in fragmented landscapes. A large-scale demographic field study was conducted to measure the degree of competitive interference between species. Model system was constituted by the community of forest-dwelling ground rodents of central Italy, including the species Apodemus sylvaticus, Apodemus flavicollis and Myodes glareolus. Populations, inhabiting 29 wood patches in a fragmented landscape, were surveyed for two years by means of a capture-mark-recapture protocol. I modeled species' distribution as a function of landscape (habitat cover and connectivity provided by hedgerows) and habitat variables (vegetation structure and food resources) to look for evidences of competitive spatial segregation. Then I tested for each species the effect of competitors on several biological parameters: survival, recruitment, reproduction, body mass, population density. Even though populations' relative distribution was consistent with a mechanism of competitive spatial segregation, with habitat specialists being favored by high-quality, well-connected fragments and generalists exploiting more isolated and degraded patches, results on demographic parameters did not fully confirm this result. The strongest competitive effects were exerted by A. sylvaticus on A. flavicollis, whereas a little degree of interference was found between Apodemus spp. and M. glareolus. Nevertheless, competitive effects were weak, acting on a few biological parameters and not translating into strong effects at the level of populations (density of individuals). These results suggest that populations were mainly distributed according to their ecological requirements; competitive exclusion of specialists from isolated and degraded fragments was actually acting but was likely to play a minor role in determining the observed pattern of distribution. 3) Perceptual range and movement ability of small mammals in fragmented landscapes. The aim of this section was to broaden our understanding of animal orientation and movements in the agricultural matrix, with a special attention on the use of plantation rows as navigation cues. Experiments consisted in releasing individuals of forest-dwelling small mammals (species A. flavicollis, A. sylvaticus, M. glareolus) in fields characterized by different types of matrices: a bare field, a grass field with random pattern of vegetation, and a wheat field at three different stages of growth. Animals (N=119) were marked with fluorescent powder and released at progressive distances from target wood fragments; in this type of experiments individuals are assumed to go directly toward the wood as soon as they perceive it. Animal tracks were then analyzed to determine perceptual ranges and movement abilities. Perceptual ranges were species-specific, with habitat specialists perceiving woods at smaller distances compared to generalists. The presence of vegetation in the fields (either grass or wheat) strongly reduced perceptual ranges of all species by obstructing individuals' view. Furthermore, wheat plantation rows drastically influenced animal movements, possibly facilitating or hampering the reaching of a wood. Individuals of all species, in fact, followed the direction of wheat rows at any stage of growth, even if they were not directed toward the target wood. This study is one of the few examples investigating in detail the demographic mechanisms of response of species to habitat fragmentation. The holistic approach allowed me to provide an overview on the process by which factors such as landscape features, habitat characteristics, and co-occurring species affect the performance of populations in fragmented landscapes. Interspecific interactions play a minor role in shaping the community of small mammals in the studied system. A major role, instead, is played by landscape characteristics (such as habitat cover, connectivity, matrix properties) and local features (such as food resources and habitat structure), in both cases depending on species-specific life-history traits. Increased individual performance (e.g. due to habitat quality) can help to increase the viability of systems of populations; at the same time animals are constrained by the physical structure of the landscape where they live, and individual-scale effects are not necessarily transferred to the level of population. Results suggest that in order to increase the viability of animal systems in fragmented landscapes there is the need to manage the quality of habitat, which proves to be a major determinant of animal populations' performance. Nevertheless, findings also strongly suggest not to ignore the overall landscape context where populations are embedded. In landscapes that have been extensively cleared, restoration aimed to increase the amount of habitat and management of outside-patch landscape elements (hedgerows, agricultural fields) might also be a critical step to ensure the persistence of animal communities.