School of Biological Sciencesii Frontispiece: A family group of eastern grey kangaroos at the Wilsons Promontory National Park in February, 2010, consisting of (left to right) adult female #6 (carrying a 3-month-old pouch young daughter that became #223), adopted young-at-foot daughter #116 (still suckling at approximately 15 months) and sub-adult daughter #7 of approximately 26 months. Photo by Wendy J. King.iii Abstract Animal grouping patterns are generally shaped by ecological constraints such as resource distribution or predation risk. Population density often increases competition, so that life-history consequences of association patterns should be particularly evident at increased densities. This thesis seeks to quantify how kinship affects association patterns in a species with unstable social bonds at high density. I also aimed to measure individual variation in sociability and evaluate its fitness consequences. Eastern grey kangaroos (Macropus giganteus) are gregarious herbivores, with a fission-fusion social system. I thus investigated how kinship and sex-age class affected behavioural and spatial associations in a high-density population of kangaroos at Wilsons Promontory National Park, Victoria.In Chapter 2, I examined relationships between individuals of different sex-age classes using halfweight indices (HWIs, which measure pairwise association strengths) and lagged association rates, and calculated six individual-based network measures. I found little social structure in the population. Mean HWIs were similar within and between different sex-age combinations (mean HWIs = 0.01 to 0.02). Individuals of all sex-age classes had preferred and avoided associates overall. Adult females showed a weak relationship (r = 0.05) between relatedness and HWIs, but this effect disappeared when controlling for geographic distance or spatial overlap between dyads.Lagged association rates decreased exponentially for adult males with adult females accompanied by young-at-foot in the breeding season. Most other lagged association rates decreased slightly in a linear fashion and few were constant. Although females without young-at-foot appeared to be more social than females with young-at foot, all associations among individual non-juvenile kangaroos were weak (mean HWI = 0.01). Sample size had strong positive effects on mean HWI and all six network measures (r = 0.40 to 0.66), which is more likely to occur in large populations where association strengths are generally low and individuals have many weak associations.I then combined observational data with genetic information to investigate sex-biased dispersal in Chapter 3. I found weak fine-scale genetic structure among adult females but none among adult males. Mean relatedness among adults was low (r = -0.01 to -0.03). Immature males moved away from their mothers at a younger age than did immature females. Although median distances of detected dispersals were 2-2.5 km for both sexes, only 6% of sub-adult females dispersed compared to 34% of sub-adult males. Adult fe...