Landscapes are inherently heterogeneous. However, some portions of a landscape are more heterogeneous than others and are therefore not equally suitable for resource extraction by elephants. Elephants have large energy demands to meet and should spend the majority of their time in areas where they are able to forage optimally. Identifying the determinants of home range location and area may therefore provide insight into aspects of landscape utilization by elephants. Using vegetation structure as a surrogate, I investigated whether landscape heterogeneity explains the variability home range size and location of elephants occurring in the mesic savannas of Zambia and Malawi. I developed a landscape map for each of five study areas. Using these maps, I applied four FRAGSTATS metrics to quantify different aspects of landscape heterogeneity within the study areas, as well as elephant home ranges and randomly located ranges delineated using a 95% Kernel estimation. I placed similar study areas into comparable groups for each of the landscape heterogeneity metrics. Elephant home range size was not a function of landscape heterogeneity metrics and may therefore be explained by other factors. Landscape complexity and diversity of elephant home ranges varied within groups of similar study areas, suggesting that these metrics were important descriptors of home range location. Within study areas, with the exception of patch density, landscape heterogeneity metrics supported the expectation that wet season ranges would be more heterogeneous than those of the dry season. In addition, female ranges were more heterogeneous than those of males during the wet seasons with respect to both patch density and landscape diversity. In most cases, greater landscape heterogeneity within home ranges was only shown during the wet season and this suggests that water requirements preclude selection for more heterogeneous landscapes during the dry season. However, elephants of the Zambian study areas, besides Kafue, selected for metrics indicative of landscape complexity and diversity during both dry and wet seasons. I therefore concluded that elephants favoured complex landscapes with more vegetation types in irregularly arranged patches and landscape heterogeneity therefore determines the location of elephant home ranges. At a regional scale, a landscape comprises habitats of varying suitability to elephants. In a metapopulation framework, such areas may form sources or sinks and therefore contribute to driving elephant movements. The ability to identify areas of importance to elephant range utilization is therefore an essential tool to apply within the megaparks for metapopulations conservation framework. / Dissertation (MSc (Zoology))--University of Pretoria, 2007. / Zoology and Entomology / unrestricted