eye. It must, however, be distinctly borne in mind that there is a fundamental difference between the eye of Vertebrates and of all other groups in the fact that in the Vertebrata the retinal body is itself a part of the central nervous system, and not a separate modification of the epidermis—myelonic as opposed to epidermic. The structure of the reputed eyes of several of the above-named genera has not been carefully examined. In Pecten and Spondylus, however, they have been fully studied (see fig. 21, and explanation). Rudimentary cephalic eyes occur in the Mytilidae and in Avicula at the base of the first filament of the inner gill, each consisting of a pigmented epithelial fossa containing a cuticular lens. In the Arcidae the pallial eyes are compound or faceted somewhat like those of Arthropods.
Fig. 24.—Embryo of Pisidium pusillum in the diblastula stage, surface view (after Lankester). The embryo has increased in size by accumulation of liquid between the outer and the invaginated cells. The blastopore has closed.
Generative Organs.—The gonads of Anodonta are placed in distinct male and female individuals. In some Lamellibranchs—for instance, the European Oyster and the Pisidium pusillum—the sexes are united in the same individual; but here, as in most hermaphrodite animals, the two sexual elements are not ripe in the same individual at the same moment. It has been conclusively shown that the Ostrea edulis does not fertilize itself. The American Oyster (O. virginiana) and the Portuguese Oyster (O. angulata) have the sexes separate, and fertilization is effected in the open water after the discharge of the ova and the spermatozoa from the females and males respectively. In the Ostrea edulis fertilization of the eggs is effected at the moment of their escape from the uro-genital groove, or even before, by means of spermatozoa drawn into the sub-pallial chamber by the incurrent ciliary stream, and the embryos pass through the early stages of development whilst entangled between the gill-lamellae of the female parent (fig. 23). In Anodonta the eggs pass into the space between the two lamellae of the outer gill-plate, and are there fertilized, and advance whilst still in this position to the glochidium phase of development (fig. 22). They may be found here in thousands in the summer and autumn months. The gonads themselves are extremely simple arborescent glands which open to the exterior by two simple ducts, one right and one left, continuous with the tubular branches of the gonads. In the most primitive Lamellibranchs there is no separate generative aperture but the gonads discharge into the renal cavity, as in Patella among Gastropods. This is the case in the Protobranchia, e.g. Solenomya, in which the gonad opens into the reno-pericardial duct. But the generative products do not pass through the whole length of the renal tube: there is a direct opening from the pericardial end of the tube to the distal end, and the ova or sperms pass through this. In Arca, in Anomiidae and in Pectinidae the gonad opens into the external part of the renal tube. The next stage of modification is seen in Ostraea, Cyclas and some Lucinidae, in which the generative and renal ducts open into a cloacal slit on the surface of the body. In Mytilus the two apertures are on a common papilla, in other cases the two apertures are as in Anodonta. The Anatinacea and Poromya among the Septibranchia are, however, peculiar in having two genital apertures on each side, one male and one female. These forms are hermaphrodite, with an ovary and testis completely separate from each other on each side of the body, each having its own duct and aperture.
The development of Anodonta is remarkable for the curious larval form known as glochidium (fig. 22). The glochidium quits the gill-pouch of its parent and swims by alternate opening and shutting of the valves of its shell, as do adult Pecten and Lima, trailing at the same time a long byssus thread. This byssus is not homologous with