It is believed that the resemblance of many hoverflies to stinging hymenopterans is a case of Batesian mimicry, though there is little experimental evidence that it is effective in protecting them from predators. In this study the effectiveness of hoverfly mimicry was investigated for humans by presenting groups of university students and schoolchildren with a questionnaire which included pictures of stinging hymenopterans, mimetic hoverflies and dipteran controls. More people thought that the mimics would sting than either of the control flies, though fewer than those who thought that the mimics' hymenopteran models would sting. This showed that the hoverflies' mimicry worked but was not 100% effective. More people thought that the good mimics would sting than poor mimics which were black and yellow, showing that the reaction was not just due to their warning coloration. Students' identification skills were poor; only 77%, 66% and 50% were able to correctly identify wasps, bumblebees and honeybees, respectively, but even knowledgeable students were confused by mimetic hoverflies. Significantly more of the students who had been stung thought that the Hymenoptera would sting and identified Hymenoptera correctly. Students who thought a hymenopteran would sting were in turn more likely to think that its mimic would sting. This suggests that the mimicry is partly mediated by experience. However, even students who had never been stung showed the same pattern of discrimination as those who had, suggesting that information is also passed on culturally. These results suggest that mimicry is effective and might help hoverflies avoid predation by birds but, as many of the subjects said they would kill a stinging insect, this would actually increase their chances of being killed by humans.