sculpture of finely granulated spiral striae, is identical in all specimens examined, well-marked in the embryo, though sometimes eroded at the apices of mature shells. The operculum is highly distinctive: it is elevated or dome-shaped, composed of a spirally rolled chitinous band, of some 8 to 10 coils, edged with a single row of short, stiff setae, and enclosing a core of triangular cells, spirally arranged around a central axis, six cells to a whorl. The radula of a neozelanic specimen is illustrated in Figure 22. The central tooth has a single equilateral-triangular cusp, without any other denticles; the laterals have a broad cusp towards the mesial edge, and a single smaller denticle laterally. The marginals are narrow, curved, and without separate serrations.

The second New Zealand siliquariid, S. maoria (Powell, 1940) belongs to the same group as the more heavily built Australian ponderosa; its type locality is off Three Kings, and a second record may now be noted, from a specimen in the Dominion Museum, from Happy Valley, Wellington.

Pyxipoma weldii was obtained alive after an easterly gale at Milford, cast ashore from the sublittoral zone attached to the holdfasts of Ecklonio, where it is typically found embedded in a massive yellowish white sponge. It is invariably an offshore species, frequently turning up in the trawl from a depth of several fathoms, and never found in the littoral zone. A single sponge may be thickly studded with several hundred shells, with the dome-shaped opercula loaded with debris projecting above the surface. The animals dart back quickly within the shell tube, which is completely sealed by the operculum. Unfortunately there was no opportunity to examine living material for ciliary currents; animals were placed in fixative immediately upon collecting.

The structural features of the pallial cavity and alimentary canal (Fig. 2) are noteworthy in placing the animal very close to Stephopoma, in contrast with the vermetids hitherto described. Similar resemblances in mode of life may be inferred from the pallial organs. Pyxipoma is without doubt a ciliary feeder, with long, cirriform gill filaments, shortly attached to the mantle wall, and agreeing in histology with Stephopoma. The adaptation of the gill to form a sweeping fringe as in Stephopoma could not be observed in Pyxipoma, though from the similarity of the filaments may be expected to be present. The pallial cavity is longer than in Stephopoma, and the gill extends through a complete spiral turn occupying the extended last whorl of the shell. The mantle margin is simple, not papillose as in Stephopoma, but finely plicate around its inner edge. Corresponding to the shell fissure, the mantle is slit along the whole length of the right side, between the rectum and the pallial genital duct. It is not easy in Pyxipoma weldii, in which the shell is wholly embedded in sponge, to give any satisfactory explanation of the functions of the shell fissure and pallial slit. In siliquariids in general the slit may reasonably be supposed to have arisen as an adaptation for the rapid expulsion of waste in a ciliary feeder dependent upon a pallial water current. Detritus is probably compacted with mucus from the hypo-branchial gland which borders the slit, and cast off from the gill filaments directly through the slit. The same adaptation may also serve

as a means for the rapid expulsion of water when the head and trunk are sharply retracted into the pallial cavity. In contrast with Stephopoma the operculum is close-fitting, so as to prevent outward flow of water from the aperture of the closed shell. It should be noted that the pallial slit in the siliquariids in no way corresponds to the pallial fissure in the female Serpulorbis, Bivonia and other genera. In these vermetids the shell is never slit, and the pallial fissure is median, to the left of the rectum, never on the right side as in both sexes of siliquariids. In the latter the slit appears to serve the same function as that of the more primitive mesogastropoda; it is not, however, to be looked upon as a survival of a primitive character, but rather as an interesting example of independent acquisition in a highly specialised group.

The food groove in Pyxipoma weldii forms a deep incision (Fig. 2, f.g.) much narrower than in Stephopoma, and bounded along either side by a tall, straight ridge. It can thus be more effectively closed off from the pallial cavity than in Stephopoma, which may help to separate the food groove contents from the detrital particles passed outwards through the pallial slit. The groove extends well forward in front of the right tentacle, and its outer margin curves round sharply to terminate immediately in front of the mouth.

Text Fig. 3—Pyxipoma weldii.
Diagram to show relations of
head, food groove, and glandular
portion of the foot. AF, anterior
marginal flap of foot; BP, opening
of incubatory pouch; F [ unclear: ] ,
terminal portion of food groove;
GF, reduced glandular sole of
foot; MO, mouth; PF, pigmented
lateral region of foot; RO,
proboscis.

Text Fig. 3 illustrates the structure of the head and foot. The proboscis is cream-white in colour, finely rugose and deeply bifid, with a long, vertical mouth slit. The two whitish cephalic tentacles, with eyes at the bases, are longer and better developed than those of Stephopoma. The sole of the foot (gf.) consists of an oval, white, rather transversely furrowed, glandular area, covered with ciliated epithelium. This corresponds to the glandular pad in Stephopoma, and is produced anteriorly, below the mouth, into a broad, rectangular flap (af.) formed by the anterior margin of the sole, and representing