In pursuit of effective control strategies against Leishmania chagasi and American Visceral Leishmaniasis, we investigated the ecology of Lutzomyia longipalpis in a series of laboratory and field experiments in Amazonian Brazil. In Chapter II, we show that bloodfeeding success in peridomestic animal pens was dependent on the density of females feeding at the host. As the density of biting flies increased, hosts became more agitated, and bloodfeeding was interrupted. However, flies did not appear to distribute themselves between the available peridomestic hosts to minimise these costs. In Chapter III, we infer from the results of mark-recapture experiments that pheromone-mediated attraction and arrest was the principal determinant of fly abundance in sheds. Males are also found to aggregate preferentially to the site of the previous night's activity. We use these results to explain the sub-optimal distribution of Chapter II. In Chapter IV, we find that residual insecticide spraying caused a dramatic decrease in fly abundance in animal pens only when neighbouring aggregation sites remained untreated. Bringing together evidence from changes in Lu.longipalpis sex ratio, abundance of the different female Lu.longipalpis gonostates and abundance of other phlebotomine species, we argue that by disrupting pheromone production, the principal effect of spraying in this study was to stimulate the formation of aggregations at untreated and previously under-exploited sites, such as dining-huts, rather than the required masskilling. Finally, in Chapter V we report on laboratory feeding experiments that provide some evidence that digestion-mediated killing of parasites in the sandfly gut may be common to many types of animal blood. In conclusion, we suggest that unless blanket spraying achieves close to 100% coverage, in the absence of a synthetic pheromone bait the best approach to disease control would be the selective treatment of susceptible host sites.