Sociality (i.e., group-living) is a multi-dimensional aspect of behavior that occurs in many vertebrate species. Because living in spatially and behaviorally cohesive groups provides the foundation for most forms of complex, cooperative interactions, understanding the reasons for group-living is a fundamental goal for behavioral biologists. In mammals, ecological factors are hypothesized to play a major role in the formation of social groups; the ecological correlates of sociality in individual mammalian species, however, are often poorly understood. The torch-tail spiny rat (Trinomys yonenagae) is a South American hystricognath rodent endemic to semiarid sand dunes in northeastern Brazil. T. yonenagae is divergent from its congeners in that it is group-living, semi-fossorial, and desert-dwelling; other Trinomys species inhabit forests where individuals live aboveground and are solitary. To explore the adaptive bases for these distinctive attributes of T. yonenagae, I combined field studies with molecular genetic analyses to (1) characterize the social organization and kin structure of torch-tail spiny rats and (2) identify the primary ecological factors influencing sociality in this species. Most (76.2%) burrow systems monitored were occupied by more than one adult, including same-sex pairs, male-female pairs, and multiple adults of both sexes. Spatial overlap among burrow mates was extensive (72.0 ± 27.0%) and included the use of the same nest site. Kinship among adults decreased as the distance between the burrow systems in which individuals were resident increased. Burrow mates - particularly females - were typically close kin, although unrelated individuals (apparent immigrants from other burrow systems) were also detected within groups. Among adults captured in two successive field seasons, nearly half remained in the same burrow system; among the remaining animals, dispersal was male-biased. Individuals that dispersed to new burrow systems were more related to opposite-sex burrow mates than were individuals that remained in the same burrow system in consecutive years. At the same time, relatedness between dispersers and opposite-sex adults was lower in the new (as compared to the original) burrow system, suggesting that dispersal is related to inbreeding avoidance. Data regarding the distribution of vegetation on the study site revealed that proximity of food resources to a burrow system was significantly associated with group size. Protective vegetation and number of burrow openings, however, were better predictors of burrow sharing, suggesting that predation is the primary factor shaping social structure in this species. Comparisons of these findings with data from both other echimyids and other desert-dwelling rodent species yield intriguing new insights into the factors favoring sociality among burrow-dwelling rodents.