An analysis of the population changes of Glossina palpalis palpalis (R.-D.) during and after two insecticide trials in Ivory Coast revealed that the effects of insecticide application depend very strongly on the characteristics of the population ecology of the fly, which showed considerable local variation. At sites with high fly populations and low natural mortality rates, the populations declined rapidly during the spray period as the insecticide killed a large proportion of the local, mainly resident, population. The high resilience of the populations, however, permitted rapid recovery, largely through reinvasion from neighbouring areas, so the reduction in fly numbers lasted for only 3–4 months at a time of year when, in the absence of insecticide, the population showed a significant seasonal decline lasting for up to three months. At other sites, with low populations and high natural mortality rates, the local populations were supported chiefly by immigrants and therefore insecticides only affected a very small proportion of the total population and thus did not cause an immediate decline in fly numbers. The low resilience of these populations, however, means that any repeated low additional mortality (such as that caused by removal trapping or deterioration of environmental conditions) would cause the populations to decline. Thus at both types of site, but for rather different reasons, single or short-term insecticide applications had a limited effect on the populations, and longterm, repeated control procedures are considered to be the only ones likely to reduce the populations significantly; through economic necessity these must rely on low-technology devices such as insecticide-impregnated screens or traps.