Dispersal and population structure at different spatial scales in the subterranean rodent Ctenomys australis

Abstract

Background: The population genetic structure of subterranean rodent species is strongly affected by demographic (e.g. rates of dispersal and social structure) and stochastic factors (e.g. random genetic drift among subpopulations and habitat fragmentation). In particular, gene flow estimates at different spatial scales are essential to understand genetic differentiation among populations of a species living in a highly fragmented landscape. Ctenomys australis (the sand dune tuco-tuco) is a territorial subterranean rodent that inhabits a relatively secure, permanently sealed burrow system, occurring in sand dune habitats on the coastal landscape in the south-east of Buenos Aires province, Argentina. Currently, this habitat is threatened by urban development and forestry and, therefore, the survival of this endemic species is at risk. Here, we assess population genetic structure and patterns of dispersal among individuals of this species at different spatial scales using 8 polymorphic microsatellite loci. Furthermore, we evaluate the relative importance of sex and habitat configuration in modulating the dispersal patterns at these geographical scales.

Results: Our results show that dispersal in C. australis is not restricted at regional spatial scales (approximately 4 km). Assignment tests revealed significant population substructure within the study area, providing support for the presence of two subpopulations from three original sampling sites. Finally, male-biased dispersal was found in the Western side of our study area, but in the Eastern side no apparent philopatric pattern was found, suggesting that in a more continuous habitat males might move longer distances than females.

Conclusions: Overall, the assignment-based approaches were able to detect population substructure at fine geographical scales. Additionally, the maintenance of a significant genetic structure at regional (approximately 4 km) and small (less than 1 km) spatial scales despite apparently moderate to high levels of gene flow between local sampling sites could not be explained simply by the linear distance among them. On the whole, our results support the hypothesis that males disperse more frequently than females; however they do not provide support for strict philopatry within females.

Figure 1

Geographic distribution of Ctenomys australis sampling sites along the study area on the coast. Squares over the map show the surface of each of the three sample sites (Western, Central and Eastern) and their individual points of capture within them (black circles: individuals pertaining to the Western genetic cluster; white circles: individuals pertaining to the Eastern genetic cluster; lines: extra-cluster immigrants; triangles: ambiguous individuals).

Figure 2

Posterior assignment probabilities. Assignment probabilities (Q) to Western (dark grey) and Eastern (white) genetic clusters (K = 2), derived from the STRUCTURE analysis. Each individual is represented by a vertical bar.

Figure 3

Inference of genetic clusters (K) from Ctenomys australis sampling sites of Necochea (Argentina). The log-likelihood values (points) based on the STRUCTURE algorithm, for each 6 independent runs to the whole data set (n = 112) are also shown. The black arrow shows the best clustering solution.

Figure 4

Spatial autocorrelation analyses. Spatial genetic structure autocorrelograms for females and males from the Western (A and B, respectively) and females and males from the Eastern sampling sites (C and D, respectively) showing the genetic correlation coefficient (r) as a function of geographical distance (only four distance classes of 80 meters each one are represented). The 95% confidence intervals for the autocorrelation coefficients (r) are also shown (broken line).