I began with an observation that there exists consistent variation among fruit-parasitic female flies (Rhagoletis pomonella) with regard to search time allocation. In essence, populations appear to be composed of flies that are "optimistic" and "pessimistic" about their chances of locating other higher quality patches. I posed the question "how can such variation be maintaned over time?". To answer this question I developed stochastic, dynamic, state-variable model that considered patch emigration decisions by individuals as a function of: (1) current patch quality, (2) average patch quality, (3) density of patches, (4) time of day, and, (5) egg load. The model was then altered to allow for optimistic and pessimistic estimates of patch availability. The optimal behaviour for such flies was then solved. Results obtained showed that optimists move more often within trees and pessimists move more frequently among trees. Further, calculation of daily reproductive output showed that both optimists and pessimists performed nearly as well as flies with errorless estimates of patch availability so long as over and under-estimates were moderate. This is because of the interaction between egg limitation, host availability and time limitation. When patch estimate errors were large (e.g. 90%), however, pessimists performed less well than optimists. These results allowed me to derive fitness curves for optimists and pessimists. These curves were then used to predict the distribution of search allocation by flies in the field. Predictions as to the shape of the distribution were consistent with field data (i.e. optimists are over-represented in samples).