A dramatic morphogenetic movement (‘ontogenetic torsion’) during the development of gastropods has been proposed as a recapitulation of the original developmental departure that established the novel gastropod body plan. Nevertheless, speculative literature about ontogenetic torsion and its evolutionary significance has far outstripped empirical observations and recent results suggest that the developmental process may be somewhat different than the traditional description. I used scanning electron microscopy, immunohistochemistry, phalloidin labeling, and histological sections to monitor displacements of five components of the visceropallium with respect to axial coordinates of the cephalopodium in developing embryos of the caenogastropod, Trichotropis cancellata. Embryos of this species achieve a transient stage of anatomical organization that also arises during development of a vetigastropod (Haliotis kamtschatkana), although morphogenetic processes that generate this stage are different in these two species. At the stage of similarity, the embryonic shell has achieved its definitive orientation with respect to the cephalopodium, but the developing mantle cavity, sensory osphradium, and anus are confined to the right side. I also show that this stage of anatomical organization is recognizable during the development of other gastropods, which collectively represent three major gastropod clades. I propose that ontogenetic torsion should be viewed as a conserved stage of anatomical organization during development, rather than a conserved process of 180° rotation between the visceropallium and cephalopodium. The results lead to the suggestion that the mantle cavity of extant gastropods evolved by enlargement of the right side of the mantle cavity in a monoplacophoran-like ancestor. Under this interpretation, there is no need for a hypothetical pre-gastropod with a mantle cavity that was restricted to the posterior end. J Exp. Zool (Mol. Dev. Evol.) 297B:11–26, 2003. © 2003 Wiley-Liss, Inc.