The mechanisms underlying the evolution of morphology are poorly understood. Distantly related taxa sometimes exhibit correlations between morphological differences and patterns of gene expression, but such comparisons cannot establish how mechanisms evolve to generate diverse morphologies. Answers to these questions require resolution of the nature of developmental evolution within and between closely related species. Here I show how the detailed regulation of the Hox gene Ultrabithorax patterns trichomes on the posterior femur of the second leg in Drosophila melanogaster, and that evolution of Ultrabithorax has contributed to divergence of this feature among closely related species. The cis-regulatory regions of Ultrabithorax, and not the protein itself, appear to have evolved. This study provides experimental evidence that cis-regulatory evolution is one way in which conserved proteins have promoted morphological diversity.

The horned soldier aphids of the Cerataphidini, unlike most social insects that reside in nests, live on the open surface of plants. The lack of a nest and other obvious ecological correlates makes it unclear why secondary-host soldiers might have evolved. Here I present a molecular phylogenetic analysis of 32 species of the Cerataphidini, including 10 species from the genera Ceratovacuna and Pseudoregma that produce horned soldiers. The phylogeny suggests that horned soldiers evolved once and were lost once or twice. Most horned soldiers are a morphologically specialized caste and two species that have unspecialized soldiers are independently derived from species with specialized castes. The genus Ceratovacuna appears to have undergone a relatively rapid radiation. Mapping secondary-host plants and geographic ranges onto the phylogeny suggests that bamboos were the ancestral secondary-host plants and that the Asian tropics and subtropics were the ancestral geographic regions for the genera Astegopteryx, Ceratoglyphina, Ceratovacuna Chaitoregma, and Pseudoregma and possibly for the entire tribe. There is evidence for vicariant events that separate the tropical and subtropical lineages in all of the major lineages of the tribe and for dispersal of some lineages. Based on these results, I present hypotheses for the causes and consequences of horned-soldier evolution.