This study presents the results of an investigation to determine the resources utilised by possums in a spatially heterogeneous landscape in the Canterbury high country. The study area comprised a mosaic of forest, scrub, shrubland, grassland and swamp at the University of Canterbury field station at Cass, inland Canterbury. The vegetation communities at Cass were originally divided into seven different categories based on species composition. These areas are referred to as the swamp, grassland, shrubland, scrub1, scrub 2, native forest, and exotic forest. Communities were classified using two multivariate techniques; TWINSPAN and Detrended correspondence analysis. A vegetation map of Cass was produced using ground survey and aerial maps, and displayed the extent and coverage of vegetation communities. These vegetation communities reflect the influence of burning and farming. The study area comprises c. 195 ha. The overall possum density was low, with approximately one possum per hectare. This result may be due to resources that possums need for survival, being deficient in the area. Possum movements were investigated by radio-collaring six female possums and six male possums at Cass. These possums were radio-tracked six times during a period of ten months, with each radio-tracking session undertaken for a period of three consecutive nights. The data collected were used to derive individual home ranges using Minimum Convex Polygon and Kernel home range estimates. There was no statistically significant difference between female and male home ranges at Cass. No seasonal difference in possum home range was detected, although other possum home range studies have found seasonal differences. There were too few den site location areas to enable accurate information concerning seasonal variation of den sites. Possums had relatively large home ranges, which probably result from a low possum density in the areas, as well as spatial heterogeneity of key resources. Nine possums showed some degree of home range overlap, but there did not appear to be any difference between females and males with respect to this. Sixty percent of the possums studied exhibited bimodal home ranging behaviour. Bimodal behaviour almost certainly results from the spatial heterogeneity of key resources, with possums having to travel long distances in order to use all the resources. Possum diet was assessed using a point sampling technique, which identifyied all ingested stomach material greater than 3mm, at four times through 2001. There was no statistical difference between female and male diet, or between juvenile and adult diet. Thirty-eight different food items were consumed by possums over the study period, with four dominant food items that comprised of 50% of their diet (Aristotelia fruticosa, fungi, Podocarpus nivalis, and Blechnum penna-marina). Due to the spatial heterogeneity of vegetation communities at the Cass study area, possums utilised the different habitats non-randomly, preferring the scrub and forest communities. Possum management is not currently needed at Cass, although in the future, if possum density increases than management options should be reassessed in order to minimise possum impact.