The idea that the last common ancestor of bilaterian animals (Urbilateria) was segmented has been raised recently on evidence coming from comparative molecular embryology. Leaving aside the complex debate on the value of genetic evidence, the morphological and developmental evidence in favor of a segmented Urbilateria are discussed in the light of the emerging molecular phylogeny of metazoans. Applying a cladistic character optimization procedure to the question of segmentation is vastly complicated by the problem of defining without ambiguity what segmentation is and to what taxa this definition applies. An ancestral segmentation might have undergone many complex derivations in each different phylum, thus rendering the cladistics approaches problematic. Taking the most general definitions of coelom and segmentation however, some remarkably similar patterns are found across the bilaterian tree in the way segments are formed by the posterior addition of mesodermal segments or somites. Postulating that these striking similarities in mesodermal patterns are ancestral, a scenario for the diversification of bilaterians from a metameric ancestor is presented. Several types of evolutionary mechanisms (specialization, tagmosis, progenesis) operating on a segmented ancestral body plan would explain the rapid emergence of body plans during the Cambrian. We finally propose to test this hypothesis by comparing genes involved in mesodermal segmentation.