Finally in Rudistes and the sessile Cirrho pods, Dr. Carpenter has pointed out the ex istence of a peculiar cancellated structure " like that of Pinna on a large scale" only that the segments of the prisms are hollow instead of solid. These hollow prisms are covered externally and internally by a struc tureless layer.
To complete this view of the different varieties of shell structure, it may now be interesting to consider the mode in which they are combined in the shells of the vari ous classes of the Mollusca. In the Bra ehiopoda, the calcareous shell is composed entirely of membranous laminw, which are superimposed at a very acute angle with the surface of the shell, and are further remark able for being thrown into sharp folds 20'00 tO of an inch apart, perpendicular to their planes. In the great majority of the recent species again, all the layers of the shell but the outermost are perforated by canals -0006 to -0024 of an inch in diameter, each of which contains a ccecal process of the mantle, corresponding with those processes which we have seen into the cellulose tunic of the Ascidians; the shells of Lingula and Orbieula are composed of horny laminw perforated by oblique tubuli like those of dentine (Car penter, /. c.). The shells of those families of Lamellibranchs, in which the lobes of the mantle are more or less united, are similarly composed almost entirely of laminated mem branous shell substance, e. g. Mytilus, Mo diolus, Tridaene, Isocardia, Conehacese, Nym phaceee.
The tubular structure is rnet with in the Areaexce, in Lithodomus, in Cardium, and has generally a marked relation with the costa tions or sculpturing of the outer surface; the membranous and prismatic structures are combined in the Myacew and Solenacece, and in those genera which have the lobes of the mantle disunited, as Ostrea, Unio, Pinna.
In the Gasteropoda the shell substance is invariably membranous, but the laminaa of which the shell is composed, usually' three in number, are marked by parallel lines into rhomboidal bodies, which are described by Dr. Gray as crystals, by Messrs. Bowerbank and Carpenter as elongated, mutually adhe rent cells. I believe that neither of these expressions is exactly correct, but that these bodies have the same origin as the prisnis of the lamellibranchiate shell ; a conviction in which I am strengthened by finding concen trically' laminated bodies, like those of the Lamellibranchiates, upon the inner surface of the shell of Helix (fig. 313. D).
In Patella the middle layer is composed of perpendicular prisms, like those of Pinna. Chiton resembles it in this respect, but the outer layer is here composed of fibres parallel to the surface, and is pierced by short canals. In Haliotis, calcified plaited laminw alternate with structureless horny layers, in immediate contact with which, says Dr. Carpenter, "is a thin layer of large cells of a very pecu liar aspect." Dr. Carpenter considers that the plaited laminw are cellular in this shell also.
Among the external shells of the Cephalo poda that of Nautilus has an external " cellu lar " layer as in Mya, and an internal nacreous layer like that of Haliotis.
The sh ells of all Lamellibranchiata, Brachio poda, and of the majority of Gasteropod Ce phalophora are external, being from their very origin never included in any involution of the mantle. It is different, however, with certain Cephalopoda and pulmonate Cephalophora, in which the shell commences its developtnent as an internal organ covered over by the outermost layer of the mantle, and may either remain so enclosed during life (e.g. Sepia, Limax), or ulti mately become naked as in Spirula and Clan silia. Although, however, these shells are truly internal (a distinction which, as I have endeavoured to show, carries with it soine important conclusions),* yet the careful ob servations upon their development in Sepia by Kolliker, and in Clausilia by Gegenbaur, appear to furnish abundant evidence that they are still truly ecderonic structures, and that they bear the same relation to ordinary shell as a nail bears to a horny epidermis among the higher animals. We know, in fact, that the nail, though to all intents and purposes mere cornified epidermis, is at first an internal structure, being covered over by the outer layers of the fcetal epiderm. A nail remaining so covered would correspond with the shell of Limax or Sepia, while an ordinary nail represents that of Clausilia. Gegenbaur, in fact, has shown that the shell of the latter mollusk commences at first like that of Limax by the deposition of a layer of calcareous par ticles in the midst of the cellular ecderon of the mantle beneath its outer layer of cells. The shell of Limax goes no further than this stage, while in Clausilia (and probably in Helix, &c.) it gradually increases by addition to its under surface, and finally bursts through the cellular investment which takes no share in its formation. It is the same with Sepia. Here the internal shell, or sepiostaire, is com posed of two layers, a dorsal and a ventral ; the former, according to Kolliker, is a thin membrane composed of slightly wavy, parallel, somewhat dark fibrils 0-001-2 " broad, which frequently appear to be composed of still niore delicate fibrillaa. So far as this membrane cor responds with the ventral layer, it is covered on both surfaces by a thin structureless la mina of carbonate of lime, which has a pearly aspect on the ventral surface where it is not covered by the ventral layer ; while it is gra nular on the dorsal surface, and on the ventral, where it is covered by the proper ventral layer, presents ridges to which the plates of the latter are attached. The thick ventral layer of the scpiostaire is composed of larnellw set at a very oblique angle to the dorsal layer, and united together by close-set partitions at right angles to their surface. Acted upon by acid, this portion of the shell leaves behind it a membranous skeleton of exactly the same form, but presenting no further structure.