An account is given of the anatomy of a series of opisthobranch molluscs principally to assess the change in importance and functioning of the mantle cavity and columellar muscle throughout the transition from prosobranch to opisthobranch organization. Intermediate steps are represented by living tectibranchs, of which Philine and Scaphander are investigated in detail, Acteon, Bulla, Haminoea, Akera, Aglaja and Gastropteron more briefly. Though an opisthobranch, Acteon has an organization typical of a monotocardian prosobranch; the remainder show trends affecting the shell and visceral mass, mantle cavity and head-foot, which resulted finally in the production of nudibranch types. It is confirmed that the adaptations exhibited by primitive tectibranchs relate to the assumption of a burrowing mode of life. Initial changes were the reduction of the nuchal area and sealing of the mantle cavity anteriorly so that it opened on the right, where it became restricted, the first perhaps prompting the sealing. A broadening and an anterior elongation of the head-foot produced a wedge to facilitate burrowing. Change in disposition of the mantle edge, incurred by differential growth, produced an involute shell with a large body whorl, alignment changing from erect to horizontal. The resultant streamlining eased infaunal progression; no vertical insinking of the viscera was involved. Subsequently the shell became reduced and finally lost. A section of the mantle edge enlarged to produce a posterior mantle lobe upon which sit both the shell and viscera, and which later became redundant as posterior elongation of the head-foot produced a slug-like form, the viscera being incorporated within the head-foot. As the nuchal area became reduced, mechanical needs prompted alteration to both the form and attachment of the columellar muscle. In Acteon the muscle is like that of a prosobranch, but the proximal region has broadened, a change of proportion required by primitive tectibranchs in order to support the floor of the mantle cavity formed from the section of mantle skirt which in prosobranchs lies on the right. This was followed by reduction and re-alignment of the entire muscle along an anteroposterior axis as emphasis changed from the muscle effecting retraction into a shell to producing contorsions of the head-foot. The shell, similarly reduced, instead of providing anchorage, became itself anchored by additional anterior and posterior attachment zones with, in more advanced forms, dorsoventral muscles of the body wall rather than longitudinal muscles fastening to the former. Importance was placed on the mutual stabilization of constituent parts of the posterior body region. Re-alignment of the muscle induced breaking up of the longitudinal muscle sheet of the head-foot to produce muscle tracts, best exhibited in those tectibranchs which swim; they are derived from both the columellar muscle and intrinsic body wall muscles. In advanced opisthobranchs, the importance of the columellar muscle progressively diminishes and it is finally lost in the adult. The mantle cavity shallowed, partially due to lack of space on the right where the mantle abuts against the viscera, but principally to avoid instability of its walls. Without support the walls will, especially in larger animals, tend to collapse owing to the restricted inhalant flow of water caused by the absence of an effective siphon and the adverse infaunal conditions. The floor may tend perhaps to be pushed laterally by increases in pressure within underlying haemocoelic spaces. Tensor muscles arose to stabilize the floor, for this became distinct from the thickened mantle edge represented by the posterior mantle lobe, and viscera were interpolated between the inner surfaces of the two regions of this section of the mantle skirt. The separation of surfaces was a consequence of the creation of space posteriorly by reduction of the nuchal area, shell and proximal columellar muscle, all adaptations to produce a slug-like form; the first was the most important at an early stage in evolution, the latter two at a later stage. There is no evidence that any tensor muscle is derived from the columellar muscle. It is suggested that the first opisthobranchs were small, a feature which almost certainly favoured colonization of the infaunal niche, and lacked a gill, water flow being produced by ciliated bands as in various small gastropods. Upon a subsequent increase in size, a gill of different pattern to the prosobranch ctenidium evolved which is not important in producing water flow. The pallial caecum is a further respiratory innovation to offset functional inefficiencies which might otherwise have been incurred upon the increase in size which was undertaken under conditions of poor ventilation. Respiratory exchange was also facilitated by fusion of the pallial caecum to the visceral mass (Philine, Aglaja, Akera), which also enabled tensor muscles to attach to and stabilize its floor. In Philine, the roof also is stabilized by areas which adhere to the shell thereby ensuring that this caecum is always fully open. Discussion of both the mantle complex and columellar muscle indicates a high incidence of parallelism. It is suggested that the term detorsion be discarded. No rotation of the mantle skirt took place, but differential growth followed by folding to which the term posterior migration has been applied. Discussion of developmental studies indicates that torsion in opisthobranchs is halted at a stage which approximately corresponds to the position of the mantle complex in the adult, and in more advanced forms torsion is essentially abolished. The final changes leading to the assumption of the nudibranch condition, and the phylogenetic interrelations of the animals investigated are briefly discussed. It is concluded that the general pattern of opisthobranch evolution was one of initial assumption of infaunal life, followed, after varying intervals of time, by return to the surface; only a few groups, of which the Philinidae are a good example, have fully exploited the infaunal niche.