Field-based mesocosm studies may be used as conceptual experiments to examine theoretical questions using the generated empirical data. A field-based Drosophilidae-nectarine mesocosm, comprising sun and shaded microclimate treatments arranged in a checkerboard design, was used in this thesis to examine four theoretical objectives. First, the efficacy of spatial analysis for detecting empirical pattern was examined. Natural variation weakened spatial structuring. However, hypothesis generation was shown not to be affected by this inclusion. Second, enhancing the hypothesis generating capacity of spatial methods was assessed, and confirmed to be possible, through the use of a sensitivity analysis method developed here. The method ( distinguished between mechanism categories (intrinsic, extrinsic and natural variation) and assessed the relative strength of each category. Next, an empirical test of the He and Gaston (2000a) parameterisation method and model to predict abundance from occupancy was conducted. Abundance estimates derived using the parameterisation method were underestimated because individuals were highly aggregated within fruit. This model and method require further exploration at fine scales for highly aggregated species. The incorporation of spatially explicit information may improve abundance predictions. Finally, the influence of spatial variation in temperature on adult body size in Drosophila simulans Sturtevant was investigated. The simple developmental effects of temperature differences, or the simple effects of stressful temperatures on thorax length, were overridden by interactive effects between temperature and larval density. As a result, flies attained the same final sizes in the shade and sun. Under natural conditions both mortality and non-lethal effects of temperature and/or crowding are likely to playa role in the evolution of body size. The results of this thesis provide i) an improved understanding of the influence of natural variation on spatial pattern, ii) an additional tool for spatial hypothesis generation, iii) an empirical test of an abundance-prediction model and iv) an understanding of interactive and non-lethal effects on body size under field conditions. The thesis therefore provides empirical support for the usefulness of field-based mesocosms to examine theoretical objectives.
Thesis (PhD (Entomology))--University of Pretoria, 2007.