According to the coevolution theory of autumn colours, the bright colours of leaves in autumn are a warning signal to insects that lay their eggs on the trees in that season. If the colour is linked to the level of defensive commitment of the tree and the insects learn to avoid bright colours, this may lead to a coevolutionary process in which bright trees reduce their parasite load and choosy insects locate the most profitable hosts for the winter. We try to clarify what the theory actually says and to correct some misunderstandings that have been put forward. We also review current research on autumn colours and discuss what needs to be done to test the theory.