Journal of Animal Ecology 2008, 77, 1250–1256 doi: 10.1111/j.1365-2656.2008.01448.x © 2008 The Authors. Journal compilation © 2008 British Ecological Society Blackwell Publishing Ltd Microhabitat heterogeneity influences offspring sex allocation and spatial kin structure in possums Sam C. Banks 1 *, Emma J. Knight 1 , Jean E. Dubach 2 and David B. Lindenmayer 1 1 The Fenner School of Environment and Society, The Australian National University, Canberra, ACT 0200, Australia; and 2 Department of Conservation Biology, Daniel F. and Ada L. Rice Conservation Biology and Research Center, Brookfield Zoo, Brookfield, IL 60513, USA Summary 1. Sex allocation theory predicts that where dispersal is sex biased, the fitness consequences of produc- ing male or female offspring are mediated by resource availability and maternal competitive ability. Females in poorer condition are expected to favour dispersing offspring to minimize resource competition with kin. Environmental heterogeneity may drive spatial variation in sex allocation through resource competition-related benefits to females and territory quality benefits to dispersing or philopatric offspring. 2. Here, we demonstrate that microhabitat heterogeneity can drive extremely fine-scale spatial heterogeneity in offspring sex allocation. Female bobucks (Trichosurus cunninghami) in temperate rainforest were more likely to produce male offspring than those in surrounding Eucalyptus forest. 3. A maternal physiological effect was identified, in that females of lower body mass were more likely to produce male offspring. This finding is consistent with resource competition predictions, in that smaller females are expected to have poorer competitive ability. 4. Genetic spatial autocorrelation analysis identified males as the more dispersing sex. Furthermore, overproduction of males by mothers in the rainforest habitat was geographically concordant with reduced philopatry, as inferred from spatial genetic analysis. This provides empirical validation of dispersal-related explanations of offspring sex allocation: that production of offspring of the dispersing sex minimizes the potential for resource competition with kin. 5. Spatial variation in dispersal via sex allocation responses to environmental heterogeneity can potentially contribute to spatial patterns in population dynamics. Key-words: genetic spatial autocorrelation, habitat heterogeneity, local resource competition, sex allocation, Trichosurus Journal of Animal Ecology(2007) >doi: 10.1111/j.1365-2656.2007.0@@@@.x Introduction Evolutionary hypotheses for offspring sex allocation by females have incorporated the trade-off between the relative maternal cost and potential fitness returns of rearing male and female offspring. Fitness returns can be influenced by variation in reproductive potential between the sexes (Trivers & Willard 1973), the probability of local settlement in relation to habitat quality and sex-biased dispersal (Julliard 2000), and heritability of fitness-related competitive abilities from male or female parents (Silk 1983; Burley 1986; Sheldon & West 2004). Maternal costs relate to the energy requirement to wean offspring (Trivers & Willard 1973) and the potential for local resource competition with philopatric offspring (Clark 1978; Silk 1983). While sex allocation findings typically have been inter- preted in regard to the major ‘competing’ Trivers–Willard (T– W) and local resource competition (LRC) hypotheses, there is an increasing theoretical and empirical appreciation of the potential for environmental effects (via resource availability) on the costs and benefits of producing male or female off- spring (Julliard 2000; Johnson et al. 2001; Byholm et al. 2002; Bradshaw, Harcourt & Davis 2003; Rubenstein 2007). Varia- tion in resource availability can influence maternal condition, thereby mediating T–W sex allocation predictions, under which females in good condition should favour sons due to the potentially greater fitness returns of high-quality males resulting from their reproductive potential (Trivers & Willard 1973; Bradshaw et al. 2003; Rubenstein 2007). *Correspondence author. E-mail: sam.banks@anu.edu.au