Simulation of the effects of movement patterns and resource density on the egg distributions of Pieris rapae (Lepidoptera) at multiple spatial scales.

This thesis presents a spatially explicit, agent based simulation, used to explore the ovipositing behaviour of the Small Cabbage White butterfly, Pieris rapae (Lepidoptera). The study concerns the effects of host-plant (Cabbage, Brassica oleracae) density upon P. rapae egg distribution patterns, at multiple scales. A general review of the literature is provided which covers the ecology of animal movement, methods of quantifying movement, models of movement, ecological theory of herbivore responses to plant density (Resource Concentration Hypothesis) and the biology of P.rapae. The construction of the simulation is described in detail and the source code plus an executable version of the software are available as a companion CD. A number of simulation experiments are reported which demonstrate the basic behaviour of the simulation over a simplified resource layout. The framework is then used to explore more complex layouts which are compared to field experiments conducted as part of a separate PhD thesis (Hasenbank, in prep). A Correlated Random Walk simulated a negative relationship between forager egg distributions and resource densities, observed at all scales. Including a diffuse attraction to resources (e.g. olfaction), simulated a negative relationship between egg distributions and resource densities at smaller scales, and a positive relationship at larger scales. This work builds on a large body of previous simulation studies and attempts to produce a useful framework for subsequent researchers to explore the effects of animal movement through the use of random walks. It demonstrates the use of the framework with a specific example concerning the egg distributions of P. rapae and the effect of scale. It provides some useful insights into both the analysis of results from a complex spatial experimental layout, and potential responses which may be observed. It demonstrates that a simple model can, in the case of P rapae be used to obtain relatively realistic egg distributions