The model level significance was α = 0.05. Unless
otherwise stated, values are provided as mean ± se. Colonies (n = 10) comprised 4–17 adult individuals (8.36 ± 1.45) and varied between 2–10 females and 2–7 males. Social interactions aboveground were very rare. In 612 h of observations (summer and winter), only 31 interactions were observed between colony members. Amicable interactions were observed on five occasions, involving two adults (n = 2) and mother and offspring (n = 3) sun basking together; allogrooming was observed in one adult pair. In contrast, agonistic interactions were observed 26 times (3% of the total observation time), significantly more than amicable interactions (z = 3.47, n = SB203580 mouse 10, P < 0.00; sign test), and always (n = 26) consisted of two individuals boxing briefly and one individual chasing the other for up to 12 m. No damaging fights were observed. Individuals of a colony always foraged alone. Season was a significant predictor of the aboveground home-range size (95% MCP), with home range being significantly greater in summer (367.39 ± 30.73 m2) than winter (164.27 ± 24.91 m2; F1, 36 = 20.58, P < 0.001). Sex was a significant predictor of home-range size (F1, 36 = 7.17, P = 0.011), being greater in females (327.01 ± 23.00 m2) than males (210.65 ± 25.88 m2). Season × sex was not a
significant predictor of home-range size (F1, 36 = 2.01, P = 0.165). Colony members exhibited a large degree of spatial overlap (percentage GDC-0199 research buy spatial overlap of 95% MCP; Fig. 1). Season was a significant predictor of home-range overlap (F1, 36 = 8.27, P = 0.001), with overlap being greater in winter (49.144 ± 2.47%) than summer (42.03 ± 2.93%). Sex (F1, 36 = 3.04, P = 0.143) and the interaction between season and sex (F1, 36 = 2.83, P = 0.101) did not predict home-range overlap in summer (females: 43.63 ± 2.46%; males: 42.46 ± 2.38%) and winter (females: 52.43 ± 2.09%; males: 43.31 ± 2.94%). The empty cage was approached in both seasons but cage biting occurred only
once (Fig. 2). Sitting in close proximity to conspecifics (i.e. tolerance) occurred infrequently (36–378 s) and was directed only at female stimulus subjects at Succinyl-CoA their capture site (i.e. non-displaced). Due to low incidences of tolerance, these data were not included in further statistical analyses. In contrast, agonistic behaviour was common (Fig. 2), mainly in the member and stranger treatments. Neither sex (Wald χ21 = 0.01; P = 0.995; GLZ) of the stimulus subject nor season (Wald χ21 = 0.00; P = 0.997) influenced aggression levels in any of the three treatments. However, there was a marked treatment effect (Wald χ21 = 95.99; P < 0.001). Post hoc tests revealed two groupings: low aggression for non-displaced stimulus subjects (β = 8.