The tethered-flight technique used previously for studies of the factors regulating the flight performance of Spodoptera exem_p ýto moths was improved and developed to increase reliability and incorporate computerised datalogging permitting faster and more extensive analysis. Larvae of S. exempta show a density-dependent phase polyphenism and tee ect of larval phase on adult flight was examined. A consistently greater flight performance was observed in female moths reared as re aria larvae compared with those reared as solitaria, but no significant effect was found in males. The distribution of flight durations was examined and found to be log-normal. Using normalised data the heritability of flight duration was estimated in a number of strains. Significant heritabilities were obtained for flights beginning before, but generally not after, midnight when radar observations in the field have shown flights achieving substantial displacements to occur. It is suggested that prolonged flights in the laboratory which start before midnight and whose durations are under genetic regulation represent migratory flights in the field. Shorter flights, particularly those starting after midnight, achieve local redistribution of moths, some of them possibly representing the "pluming" behaviour observed using radar. Flight duration was shown to respond to selection for both short and long flight. The Fl generations of the lines of four strains selected for increased flight capacity revealed a bimodal distribution of flight durations indicating two types of moth. It is suggested that a major gene could account for this effect and that-the longer fliers represent potential migrants while the shorter fliers are capable only of flights achieving local dispersal. The effect of flight on the fecundity of females was examined. It was found that in moths subsequently fed water (necessary for oocyte development), a decrease in fecundity resulted which was related to the duration of flight. In moths fed sucrose fecundity was maintained at the level of unflown moths. These results are discussed in relation to the behaviour and ecology of S. exempta in East Africa and the prospects for the success of a regional approach to its control.