Credit: RICHARD CUMMINS/CORBIS

Most of us are happy simply to marvel at the autumn colours of deciduous trees, as shown here. But the late William Hamilton, one of the most influential thinkers on evolution in the twentieth century, and Sam Brown felt compelled to ask what such a show is in aid of. They hypothesize that trees are sending the message 'pick on someone else' to their insect enemies.

Most researchers, when they have considered autumn displays at all, have assumed them to be a side effect of senescence. But Hamilton and Brown felt that trees would not, without good reason, make large amounts of potentially costly compounds in their leaves just before shedding them.

In autumn, many insects are looking for plants on which to overwinter and to feed and reproduce the next year. Hamilton and Brown concentrated on aphids, because aphid species tend to be choosy about what they eat and use colour cues to find their hosts. The amount of damage aphids can inflict also gives trees a strong incentive to deter them.

Using published data, Hamilton and Brown surveyed 262 tree species and found that the yellowness or redness of a tree's autumn leaves correlates with the number of aphid species that attack it (Proc. R. Soc. Lond. B 268, 1489–1493; 2001). Maples, for example, which put on some of the most spectacular displays, are some of the most heavily aphid-infested species, fitting in with the idea that tree species suffering greater insect damage should invest more in colour signalling. Costly autumn pigments would be an 'honest signal' — only trees that were truly committed to defence would make them.

The connection between colours and herbivores raises questions for exploring the hypothesis further. For instance, does the link hold good within species, as well as between them — that is, are brighter individuals left alone? And can the idea be broadened beyond aphids?