A study of carabids and cicindelids was carried out in eastern South Africa using the same methodology as has been used in the northern hemisphere to obtain a southern hemisphere perspective. The study used the macroecology approach to compare patterns and responses of these animals to anthropogenic disturbances in visually similar habitats (forests, grasslands). Although this is essentially a local component of a larger macroecological study, it is shown that even though species and identities differ between the north and southern hemispheres, the general patterns of community response to anthropogenic disturbances are surprisingly similar. Changes in carabid assemblages were assessed across eight sites or landscape elements experiencing a range of disturbance types, both regular and irregular (such as mowing, fire, irrigation, fertilizers, pesticides, forestry). Direct comparisons were made with similar studies in the Palaearctic. Species diversity, seasonal population variations, population phenology, spatial patterns and mean body size of species assemblages relative to the landscape elements are described. As in the north, intensively disturbed biotopes were impoverished, and natural patches of moist forest acted as habitat sources for these disturbed sinks. Roadside verges were species-rich analogues of natural habitats. The mean body size of carabid assemblages in forest and grassland sites decreased with increasing intensity of disturbance. One contrary comparison with the north was that a plantation, in this case macadamia, was exceptionally rich in carabid species and individuals. Classification and ordination methods identified and characterised the eight sites to six ecological meaningful biotopes for carabids and cicindelids. This also allowed inferences as to how the various landscape disturbances in natural forests, planted pine forests, macadamia plantation, recreational park, road verge grasslands and hayfields affect carabid and cicindelid species richness and abundance. Species assemblages that responded to these anthropogenic impacts were potential indicator groups that can assist in the planning and management of forest and grassland landscapes for conservation of biodiversity. Some management recommendations for these landscapes are given. Individual species-environment relations were investigated using both univariate and multivariate analyses. The solutions to these analyses were then used to describe how species are distributed along major environmental gradients. It was shown that soil characteristics (pH, moisture, twig and/or leaf litter) determine carabid and cicindelid assemblages. Land-use and management regimes influence these patterns. The effect of altitude is masked by the presence of soil characteristics in a multivariate analysis, and more so in the presence of pH and moist soil-sand gradients with changing altitude. In the absence of soil characteristics and in univariate analysis, altitude becomes very important. Altitude has therefore an indirect effect in that it determines climate, which, in turn, determines soil and vegetation type which then determines species presence and abundance. It is concluded that the macroecological approach has great potential for teasing apart local effects from global ones, and can contribute to the conservation of biodiversity at both small and large scales. / Thesis (Ph.D.)-University of Natal, Pietermaritzburg, 1998.