The pallial organs and the currents within the mantle cavity have been studied in many genera of aspidobranch Gastropoda including members of the Zeugobranchia, Fissurellidae, Patellacea (Docoglossa), Trochacea, Neritacea and Valvatacea. Relevant data on allied genera, not available alive, have been considered. The aspidobranch condition has been retained in a diverse variety of Prosobranchia; its possession does not indicate close relationship. The mantle cavity of the Mollusca is a respiratory chamber unique in its primitively posterior position and the opening into it of the alimentary, renal and reproductive systems. The respiratory current is created by lateral cilia on filaments alternately arranged on either side of the axes of the paired ctenidia. The current passes between the filaments invariably from the efferent to the afferent surface, i.e. in the opposite direction to the flow of blood. To withstand the water pressure the filaments are strengthened by skeletal rods under the zone of lateral cilia near the efferent surface. Frontal, abfrontal and terminal cilia (the last probably not primitive) are concerned with cleansing. Sediment is consolidated in mucus from hypobranchial glands situated on the roof of the mantle cavity where this is carried by the current. The osphradia always lie where the water current first impinges on the surface of the cavity; they are regarded as tactile organs for the estimation of sediment. Ctenidia, hypobranchial glands and osphradia form a functional unit. The anus and other openings discharge into the exhalant chamber dorsal to the ctenidia. There are correlated mechanisms in the gut for the consolidation of faeces. The primitive Mollusca probably crawled on a hard substratum as do the modern Loricata and all aspidobranch Gastropoda, except the Valvatacea where the pallial organs have migrated forward in the mantle cavity. The more highly specialized pectinibranchs, together with the Scaphopoda and Lamellibranchia, invaded soft substrata. Other pectinibranchs, like many modern Cephalopoda, took to pelagic life. In the evolution of the Gastropoda exogastric coiling of the shell probably preceded torsion. Of the theories concerning the origin of torsion, that of Garstang, which postulates larval mutation, is the most probable. The process still occurs in the development of modern Prosobranchia. The immediate advantage to the larva explains the selection of the original mutation and the absence of intermediate forms. So far from being of immediate advantage to the adult, the occurrence of torsion presented a series of problems only successfully solved by the pectinibranchs and the equally specialized Patellacea. The immediate problem, of 'sanitation', raised by the discharge of faeces and excrement into the anterior mantle cavity, was met by the appearance of a marginal slit (or aperture), with consequent withdrawal of the anus, which enabled the exhalant current carrying faeces and excretia to pass out clear of the head. The slit disappeared with the loss of the right ctenidium, when a left-right respiratory current was established and the rectum again extended to the margin of the cavity, and is now confined to the few surviving zygobranchiate genera. The asymmetry of the pallial organs was caused by the asymmetrical coiling of the shell which probably followed torsion. The organs of the right side were reduced or lost except where, as in the Fissurellidae, secondary shell symmetry was attained before the complete disappearance of these organs. The reproductive and renal organs are asymmetrical. There is no left gonoduct and possibly no left gonad (this loss may have preceded torsion); the right gonoduct opens into the reno-pericardial canal. In all Zeugobranchia (including Fissurellidae), Trochacea and Patellacea, the right kidney is retained and the left reduced; in the Neritacea, Valvatacea, Cocculinacea and all pectinibranchs, the right kidney disappears. Its loss is associated with the assumption of the monotocardiate condition and may be due to loss of the right ctenidium and so of the auricle. The retention of the right kidney in the other aspidobranchs cannot be correlated with changes in the pallial complex. Aspidobranchs with the right kidney have been faced with fundamental reproductive limitations; the impossibility of internal fertilization has prevented them from invading fresh waters or the land (unlike Neritacea, Valvatacea and pectinibranchs) or the abyssal seas (unlike Cocculinacea). In aspidobranchs four conditions have resulted from the initial asymmetrical coiling of the shell: (i) Asymmetrical shell with two asymmetrical ctenidia (Zeugobranchia); (ii) Secondarily symmetrical shell with two symmetrical ctenidia (Fissurellidae); (iii) Asymmetrical shell with loss of one ctenidium (Neritacea, Valvatacea, Trochacea); (iv) Secondarily symmetrical shell with loss of one or both ctenidia (Patellacea, Cocculinacea, some Neritacea, e.g. Septaria). The secondarily symmetrical limpet form has been acquired independently four times by these aspidobranchs as well as several times by other Gastropoda. In the Patellacea a functional series, Patelloida-Lottia-Patina-Patella, indicates the manner in which the monobranchs may have evolved into types with a ring of pallial gills and no ctenidium. With the loss of the left row of filaments in the remaining (left) ctenidium, the pectinibranch condition was assumed and the ideal solution found to the problem of respiratory circulation originally raised by torsion. Success is indicated by the wide range of adaptive radiation exhibited by the pectinibranchs. The primitive molluscan ctenidium was probably essentially similar to that of a modern asymmetrical Zeugobranch, e.g. Haliotis. The ctenidium in the aspidobranchs has been variously specialized; in the zygobranchiate Fissurellidae in association with the retention of the marginal slit, with increased afferent attachment in the monobranchiate Trochacea, and with greater freedom in the monobranchiate Patellacea, Valvatacea and Neritacea where skeletal rods are also absent. In the pectinibranchs, apart from specialization for food collection, the ctenidium varies little. In the Loricata, apart from multiplication and loss of skeletal rods, the ctenidia remain fundamentally unchanged. In the Scaphopoda they are lost. In the Lamellibranchia attachment is exclusively afferent. The evolution of the complex food-collecting ctenidia, representing a change in the function of the frontal cilia, is traced by way of those of the Nuculidae. The elongated filaments of the Filibranchia and Eulamellibranchia probably evolved not by downward growth of the tip of the filaments and later reflexion, but by ventral extension of the middle of the originally horizontal filament. Only then could the food groove function during development (and evolution). Pumping ctenidia have been independently evolved by the Nuculanidae and the Septibranchia. In the Cephalopoda respiratory currents are produced by contractions of the muscles of the funnel (Nautilus) and mantle (Dibranchia). The initial flow of water through the ctenidia is probably reversed, the exhalant chamber with anus and other openings being on the efferent side while skeletal rods appear in membranes on the afferent side of the filaments; but there may be some backflow of water in the efferent-afferent direction. The evolution of this type of respiratory current made possible the swimming movements of the Cephalopoda. The tetrabranchiate condition found in Nautilus is not regarded as primitive. There is no reason for assuming that primitive nautiloids were necessarily tetrabranchiate. The increased respiratory needs of the evolving Cephalopoda may have been met (i) by maintenance of the primitive circulatory system but duplication of the ctenidia, the current being produced by pulsations of the funnel only (e.g. Nautilus), or (ii) by retention of one pair of ctenidia, but increased efficiency by the acquisition of capillary circulation with branchial hearts and a more powerful respiratory current produced by the mantle musculature following the reduction and overgrowth of the shell (modern Dibranchia). The relation between respiratory surface and body weight in Gastropoda, Loricata and Nautilus is approximately similar. With the elongation of the filaments in ciliary-feeding Lamellibranchia and Gastropoda the current is increased beyond the respiratory needs of the animal. Hypobranchial glands occur in association with the ctenidia in all aquatic Gastropoda where sediment needs to be consolidated (i.e. not in Patellacea or Valvatacea); they disappear with the ctenidia, e.g. in Caecum and all terrestrial Gastropoda except the Neritacea. Similar glands occur in the Nuculidae and Solenomyidae (Protobranchia) and in Monia (Filibranchia); glands with similar functions occur in the pallial grooves of Patina and its allies (Patellacea) and in the Loricata. Osphradia are associated with ctenidia in all aquatic Gastropoda, persisting after the loss of these in the Patellacea and in Caecum, and reappearing in aquatic Pulmonata. Their presence is nowhere correlated with existence in water of variable character or with the type of food. There is some correlation between their size and the degree of sediment normally encountered. In certain Patellacea and in the Loricata subpallial sensory streaks in the pallial grooves may have the same function. Similar organs, of dubious function, occur in the exhalant cavity in the Lamellibranchia, but in the Cephalopoda osphradia are confined to Nautilus where alone sediment will accumulate in the mantle cavity. In all but the Lamellibranchia sediment carried in by the respiratory current is carried over the osphradial surface. Study of the bipectinate osphradium of Neptunea indicates that it is admirably adapted for the estimation of large quantities of sediment which passes slowly over its extensive sensory surface and is then rapidly removed by cilia and consolidated in mucus.