This thesis assesses the interactions among tigers Panthera tigris, leopards Panthera pardus, their prey and their habitats in a subtropical forest in Bardia National Park in lowland Nepal. The first part concerns the influence of habitat on the densities, distributions and space use of the ungulate prey species. Then the population developments of the predators and their prey during 22 years of protected area management and their interactions were investigated. The last part of the thesis assesses the competitive relationships between tigers and leopards and their behaviour. Field study was conducted from 1998 to 2000. Results showed that habitat productivity strongly influenced ungulate densities. The most abundant species, axis deer Axis axis and hog deer Axis porcinus, were both associated with the most productive habitats, grasslands and early successional forests, resulting in considerably higher densities of ungulates in these areas. Habitat periodicity and predictability were identified as influential factors on the spacing patterns and social structures of hog deer and muntjac Muntiacus muntjak. The densities of axis deer, hog deer, muntjac and wild boar Sus scrofa were affected by different habitat parameters. Hence, habitat heterogeneity seemed to promote niche separation among these ungulates. Survey of predators by camera trapping in 1998-99 and 2000-01 showed that tiger numbers had increased to high densities of 13.3 plus or minus 2.08 and 19.3 plus or minus 2.18 animals per 100 km², respectively. Leopard density was lower (approximately 5 per 100 km²), indicating different numerical responses among the predators to the overall increase in ungulate densities. Analyses of tiger and leopard scats (n=460) showed that some low density prey species were preferred (wild boar) or killed according to availability (barasingha and nilgai antelope), whereas the two most abundant species, axis deer and hog deer, were killed less frequently than expected. None of the predators exhibited prey-switching at low prey densities. In terms of tiger and leopard behaviour, competition and coexistence, exploitative competition between the two predators was not a plausible cause of the low leopard density, mainly because the removal rates of the main and shared prey was relatively low, and also because food consumption of leopards was higher than recorded in other studies. The distribution patterns of scats and territorial markings showed that leopards stayed closer to the park border than tigers and inhabited the margins of the tiger home ranges. This pattern fits with the distribution expected from a predator influenced by interference competition. Tiger and leopard diets differed mainly in three aspects: (1) leopards hunted more small prey than tigers; (2) leopard hunted more domestic animals than tigers because of their distribution near human settlements; and (3) tigers consumed more hog deer than leopards, probably due to different availability of this prey and avoidance of risky habitat by leopards. Hence, different availability of prey due to avoidance behaviour among leopards seemed to be an important factor explaining the diet differences. The findings in this study contrast with the vision of large carnivores as maintainers of biodiversity, and emphasize the importance of taking into account potential cascading effects on other faunal groups when initiating conservation measures to increase the abundance of large predators.