Weaver ant colonies in mango trees are reported to deter oviposition by fruit flies into the developing fruits and, with controlled ant colonies, this can be sufficiently effective to become incorporated into commercial practice. A widely-offered explanation for this deterrent effect is that patrolling ants disturb flies on the fruit before successful ovioposition. Volatile compounds that might be contained in deposits left on fruits by patrolling ants have also been proposed as a deterrent factor. This study investigates the effectiveness of weaver ant control in a community of smallholder mango growers in Tanzania, where typical practice is to harvestmature, but not ripe, fruits for local sale. Initial interviews with growers provide an estimate of 10-25% crop losses due to fruit fly infestation and no strong perception of any significant impact of weaver ant colonisation on these losses. However, direct observation shows a reduction in fruit fly landings on fruits in the presence of ants, confirming earlier observations, but analysis of volatile emissions from fruits using gas chromatography/mass spectroscopy has not identified compounds that consistently reflect the presence of weaver ants. During ripening the fruits emit ethyl crotonate, an attractant and ovipositing stimulant for fruit flies. These data suggest that, for the specific cultivation conditions and uncontrolled ant populations, visual deterrence has a limited, benefit and there is no contributory, deterrent effect of volatile compounds. Initial observations suggest that the major infestation problems may occur during postharvest transit and storage, where ants are not present.