Prisoners in Their Habitat? Generalist Dispersal by Habitat Specialists: A Case Study in Southern Water Vole (Arvicola sapidus)

Godoy JA (2011) Prisoners in Their Habitat? Generalist Dispersal by Habitat Specialists: A Case Study in Southern Water Vole (Arvicola sapidus). PLoS ONE 6(9): e24613. doi:10.1371/journal.pone.0024613 Prisoners in Their Habitat? Generalist Dispersal by Habitat Specialists: A Case Study in Southern Water Vole (Arvicola sapidus ) Alejandro Centeno-Cuadros 0 Jacinto Roma n 0 Miguel Delibes 0 Jose Antonio Godoy 0 Paul Sunnucks, Monash University, Australia 0 1 Department of Conservation Biology, Estacio n Biol o gica de Don ana, Consejo Superior de Investigaciones Cient ficas , Sevilla , Spain , 2 Department of Integrative Ecology, Estaci o n Biolo gica de Don ana, Consejo Superior de Investigaciones Cient ficas , Sevilla , Spain Habitat specialists inhabiting scarce and scattered habitat patches pose interesting questions related to dispersal such as how specialized terrestrial mammals do to colonize distant patches crossing hostile matrices. We assess dispersal patterns of the southern water vole (Arvicola sapidus), a habitat specialist whose habitat patches are distributed through less than 2% of 2 the study area (overall 600 km ) and whose populations form a dynamic metapopulational network. We predict that individuals will require a high ability to move through the inhospitable matrix in order to avoid genetic and demographic isolations. Genotypes (N = 142) for 10 microsatellites and sequences of the whole mitochondrial Control Region (N = 47) from seven localities revealed a weak but significant genetic structure partially explained by geographic distance. None of the landscape models had a significant effect on genetic structure over that of the Euclidean distance alone and no evidence for efficient barriers to dispersal was found. Contemporary gene flow was not severely limited for A. sapidus as shown by high migration rates estimates (.10%) between non-neighbouring areas. Sex-biased dispersal tests did not support differences in dispersal rates, as shown by similar average axial parent-offspring distances, in close agreement with capture-mark-recapture estimates. As predicted, our results do not support any preferences of the species for specific landscape attributes on their dispersal pathways. Here, we combine field and molecular data to illustrate how a habitat specialist mammal might disperse like a habitat generalist, acquiring specific long-distance dispersal strategies as an adaptation to patchy, naturally fragmented, heterogeneous and unstable habitats. - Animal dispersal is commonly defined as the movement of individuals away from their home ranges with no subsequent return (at least, temporally) [1]. Although the decision of how, when and where to disperse is taken by individuals, its consequences extend to population and species levels. Individuals disperse as an effective strategy for the avoidance of inbreeding, resource competition, and kin competition [2], and this initiates important ecological and genetic feedbacks in spatially structured populations [3]. It has been classically debated whether patchy distributions of species result from pure distance effects (i.e. individuals mostly recruiting near their parents) [4], species-specific environmental responses [5] or the interaction of these two, which might depend on the scale at which the study is conducted [5,6]. In naturally or anthropogenically fragmented landscapes, the degree of fragmentation and the spatial configuration of the network of patches will influence dispersal routes and probabilities and, consequently, will affect the rates of colonization of empty patches and the distribution of genetic diversity [7]. These consequences make of dispersal a keystone process in ecological and evolutionary studies. In this sense, dispersal may be seen as the glue that holds populations connected, but also as the glue that connects different scales and disciplines [8]. Gene flow is one of the important consequences of effective dispersal (i.e. when it is followed by breeding success) and is expected to homogenize the genetic variation among populations and counteract the structuring effects of drift. Therefore, species might show strong genetic structure when gene flow among populations is reduced, either because the geographic distance exceeds average dispersal distance or because effective barriers (or filters) to dispersal separate the populations. Genetic structure will thus be greater for low mobility than for highly mobile species at a particular geographical scale. Classical analyses of patterns of gene flow have usually addressed their extent and distance components, often revealing a monotonic decrease of gene flow with distance (isolation-by-distance), where geographic distance is calculated as the Euclidean distance separating individuals or populations. This approach implicitly assumes that dispersing individuals travel in a straight line across a homogeneous or irrelevant landscape matrix. A more recent approach has highlighted the relative importance of the landscape matrix heterogeneity on the dispersal behavior of species, by showing a better correlation of gene flow with landscapemodified distances than with purely Euclidean distances [9]. Species are often classified into habitat generalist or specialists based on habitat requirements: while the former can exploit multiple habitat types or food sources, the latter are restricted to only one or few habitats. Like habitat generalists, specialists in large and continuous habitats can move rather freely across space, rendering populations with reduced spatial and genetic structure. Quite often, however, habitat specialists are restricted to more or less scarce and scattered patches of suitable habitat embedded in an unsuitable habitat matrix. Given that small and isolated populations have increased risks of extinction, highly specialized species inhabiting patchy habitats require a high ability to move through the matrix in order to avoid genetic and demographic isolation [10]: paradoxically, habitat specialists must behave as dispersal generalists. Generalist dispersal patterns have been described in plants [11,12] and invertebrates [13] occupying scarce and patchy habitats, although this possibility has not been yet assessed in mammals. We set out to test this prediction using southern water vole (Arvicola sapidus) as a case study of a species tightly associated to naturally fragmented habitats embedded in heterogeneous but largely hostile habitat matrices (see below). We first estimate gene flow among the populations of this rodent in the study area through indirect and direct approaches based on neutral autosomal microsatellite genotypes and mitochondrial control region sequences. We will then evaluate the relative role of the landscape matrix in shaping gene flow patterns through several landscape genetic approaches. According to our prediction, a high ability of southern water voles to disperse must be reflected on (...truncated)