g. deer mouse Peromyscus maniculatus; Blank, Nelson & Buchberger, 1988). Ice rats

do not hibernate or enter torpor, and their physiology (e.g. greater thermal conductance; Richter, Webb & Skinner, 1997) resembles that of their congeners inhabiting warmer, low altitude, climates (e.g. vlei rats Otomys irroratus). However, they have morphological adaptations, such as an elongated small intestine for increased energy uptake (Schwaibold & Pillay, 2003) and thick fur for insulation (Rymer, Kinahan & Pillay, 2007). Behaviourally, they time their activity to the warmest periods (Hinze & Pillay, 2006) and bask by withdrawing PI3K inhibitor the limbs, tucking in the head and orientating the back to capture Wnt inhibition the sun’s rays (Rymer et al., 2007). Ice rat colonies occupy an aboveground area of 98–1200 m2 (mean: 720 m2), depending on the number of individuals in a colony (i.e. colony size; Hinze, pers. obs.). Colony members (kinship unknown) collectively

construct an intricate interlinking tunnel system 280–357 mm belowground, consisting of ≥25 entrance holes and one to two nesting chambers, providing refuge at night and during adverse weather (Hinze, Pillay & Grab, 2006). Ice rats are vulnerable to thermoregulatory stress due to their high surface area to volume ratio, which creates greater heat exchange with the environment (McNab, 1983). In accordance with the social thermoregulatory hypothesis, ice rats huddle in their burrows (Hinze & Pillay, 2006), presumably reducing the per capita energy expenditure for thermoregulation

(Scantlebury et al., 2006). We expected that group living could also be explained by the benefits accrued from burrow sharing (burrow-sharing hypothesis), the prevailing ecology of the Maluti mountains (resource dispersion and food competition hypotheses) or a combination of these. We excluded predation risk as a determinant of group living because ice rats experience minimal predation (Schwaibold & Pillay, 2006) and spend much of their time openly sun basking. We also excluded the resource-defence hypothesis because preferred food of ice rats is not abundant throughout the year (Schwaibold & Pillay, 2010). For the burrow-sharing hypothesis, medchemexpress we predicted high overlap of home ranges between colony members because the nest is a shared resource (Schradin & Pillay, 2005). Moreover, we predicted that colony members would be: (1) amicable and/or show reduced aggression (social tolerance) to group members; and (2) aggressive to non-colony members, as observed in striped mice Rhabdomys pumilio (Schradin, 2004). We also expected that social behaviour would be predictable over time because groups are stable throughout the year (Hinze et al., 2006).