Asking the right questions about evolution of development, larval morphology, and life history requires knowledge of ancestral state. Two hypotheses dominate current opinion about the ancestral life cycle of bilaterians: the "larva-first" and the "intercalation" hypotheses. Until recently, the larva-first hypothesis was preeminent. This proposes that the original indirect life cycle of bilaterians included a planktotrophic larva followed by a benthic adult. Phylogenetic evidence suggests that a planktotrophic larva is plesiomorphic for echinoderms. A preponderance of developmental studies on echinoderms may have fostered a tendency to extrapolate conclusions about echinoderm development to other clades, particularly the concept that larval and juvenile/adult bodies are mostly separate entities. However, some of the recent reconstructions of bilaterian phylogeny suggest that nonfeeding larvae may have been ancestral for bilaterians, and these may have been intercalated into a life cycle that was originally direct. I review comparative data on molluscan development that suggests the trochophore-like stage is little more than a gastrula with transient structures (prototroch and apical sensory organ) to allow a temporary planktonic phase during development. Most lineage founder cells of molluscan embryos generate progeny that develop through the veliger stage into structures of the juvenile, which becomes benthic when the prototroch and apical sensory organ are lost. In light of this, the model of separate larval and juvenile bodies with the latter developing from nests of multipotent cells within the larva is inappropriate for molluscs. The intercalation hypothesis may be a better model for interpreting development of molluscs and other lophotrochozoans.