Recently W.D. Hamilton and colleagues proposed a provocative new theory to explain the adaptive significance of autumnal leaf colours. They suggested that these colours were signals produced by the trees to warn potential insect herbivores of their defensive ability and tested this theory by an analysis of data on aphid species richness on different tree species. Here we argue that the principal assumptions of their theory do not match current knowledge of plant pigment biochemistry and aphid ecology. We therefore present further adaptive explanations for autumn leaf colours and suggest alternative reasons for the reported relationship between tree leaf colour and aphid species richness.