Elephants damage and kill trees. This behaviour often appears to be excessive because their immediate, subsequent consumption of edible parts of these trees is often negligible. Some trees later resprout after snapping and toppling by elephants and thus produce forage at the animals' preferred feeding-height. We argue that this and other 'farming' hypotheses are group-selectionist and are thus not evolutionarily stable strategies. We suggest that excessive damage to trees is more likely to be due to social or sexual factors. More behavioural analyses and experiments are needed to understand this phenomenon and its implications for conservation.