however, and perhaps a majority of vermetids, it appears to rely also, to a smaller extent, upon mucus trap feeding. Novastoa lamellosa differs from Serpulorbis zelandicus in its position in the littoral, occurring much higher on the shore, immediately below the sessile cirripede formation on Poor Knights either Elminius plicatis or Chamaesipho brunnea. Possessing a close-fitting operculum when the foot is withdrawn into the shell, the animal is much more tolerant than Serpulorbis of exposure between tides and, like barnacles, to considerable extremes of temperature. The lower margin of the Novastoa zone is approximately at low-water neap tide—it is thus ecologically equivalent to the zone of serpulid worms—an association which seems to be unrepresented at Poor Knights Islands; it is the dominant species in Cranwell and Moore's Novastoa-encrusting coralline association. Serpulorbis zelandicus by contrast apparently never reaches the status of a dominant sessile organism—the animal is much less protected, having no operculum and relying only on deep retreat into the shell tube; it occurs at much lower tide level, singly or in patches, under stones in clean water at extreme low spring tide mark, in the zone of emergent brown algae (Carpophyllum) and encrusting pink corallines, or well below low tide on laminarian holdfasts. The third encrusting vermetid represented in the Cookian province is Stephopoma roseum, which may become locally dominant under stones, and just below fringes of emergent Corallina. Its tidal position is slightly higher than that of Serpulorbis, and it is a good deal more tolerant of sediment. Where the two forms occur together—as in the Noises Group (Morton, 1951)—Stephopoma forms a continuous strip just above low spring-tide mark, and Serpulorbis is found in patches with pink Melobesia, immediately below.

The leading ecological features of Novastoa lamellosa are summarized by Cranwell and Moore: “Still water is antagonistic to this species, which seems to march parallel with the other communities all round the islands on steep or moderately inclined slopes, but is lacking or poorly developed in tide pools or on shelving rocks where the water is likely to lie.” A preference for disturbed water is shared by the Australian “Vermetus” novae-hollandiae (Yonge, 1932), which occurs as a ciliary feeder on exposed portions of the Barrier Reef. Novastoa, however, would not appear to tolerate positions of maximum exposure to surf—where—on Poor Knights—Chamaesipho extends right down to D'Urvillaea or Xiphophora and the vermetid zone is cut out.

The general structure of the head, foot and pallial cavity of Novastoa (Plate 8, Fig. 1) resembles that of “Vermetus” novae-hollandiae. When the animal is extended during feeding, the opercular disc is raised, the head and foot widely protruded, and the margin of the mantle widely everted to lie over the rim of the shell tube. The animal is handsomely and distinctively coloured. The head (cph), terminating in the short cleft proboscis, is jet black behind, encircled in front by a broad scarlet band, widest laterally—just behind the tentacles, with a lighter yellow patch below at the base of either tentacle. The cephalic tentacles and the lobes bordering the mouth at the sides are black. The upper and lower borders of the mouth and the entrance to the buccal cavity are orange. The plug-like foot is whitish behind, orange-brown dorsally along the ciliated tracts (a.c.tr., l.c.tr.) with two

large black patches laterally—without cilia. The pedal tentacles (p.tn.) and membrane between them are black, the lower lip of the pedal gland aperture light yellow, while in front of the proboscis and pedal tentacles the foot is occupied by a semicircular orange-red pad (gl.ft.), strongly ciliated, representing the original sole region of the foot. The mantle margin is encircled by a narrow black line, followed immediately by a band of canary yellow within, and a broader band of scarlet, prominent when the edge of the mantle is everted, passing further back into translucent white. The foot differs from that of Serpulorbis in being surmounted by a large circular operculum (op.) frequently loaded by a hemispherical mass of attached calcareous alga. When the outer surface is clean, it is seen to be thin and horny, shallowly concave, its upturned edge slightly overlapping the margin of the foot. The black pigmentation of the periphery of the foot appears as a dark ring encircling the attachment area of the operculum, while at the centre the disc is raised up by the flat top of a calcareous axial plug, coated thinly with chitin, terminating bluntly above, and projecting strongly below as a round-tipped mamilla deeply inserted into the muscular column of the foot. In Finlay's words, the operculum is “shaped like an everted mushroom.” The structure of the operculum of Novastoa finds its closest resemblances among vermetids in the genus Spiroglyphus. A figure is given for comparison of the operculum of S. megamastus (Plate 8, Fig. 4c) from the Pacific coast of North America. The chief differences from Novastoa are the much wider overlap of the chitinous disc at the sides of the foot, the deeper concavity and the slighter development of the axial plug and mamilla.

The apex of Novastoa lamellosa (Plate 9, Fig. 6) is also strongly reminiscent of Spiroglyphus: in both genera there are two nuclear whorls only, horny-brown in colour, the second much larger and somewhat inflated, covered with fine granulations. In Novastoa lamellosa the lip of the nuclear aperture forms a slight projecting rim around the commencement of the adult shell, which is opaque white in colour, sculptured with close-set, rather indistinct, transverse rugae. The orientation of the nucleus is quite characteristic (Fig. 6b)—its axis directly transverse to that of the adult shell.

Fig. 1 illustrates the structure of the head and foot, with the pallial cavity opened by incision along the dorsal mid-line. The mantle margin is entire in the male, slit backwards along the mid-line in the female for the attachment of egg capsules to the shell as in the serpulorbids. On the left side is a wide triangular notch (inh.) (better developed in some specimens than in others) for the passage of the inhalant pallial current. The osphradium (os.) is a long, straight ridge, the endostyle is absent as typically in Vermetidae. The ctenidium (ct.) is of relatively small size, not larger than in Serpulorbis zelandicus, and, to judge from Yonge's figure, of rather smaller extent than in “Vermetus” novae-hollandiae. As in Serpulorbis zelandicus, the filaments are unspecialized, simple triangular leaflets, showing no tendency to become linear. The ciliary currents, however, especially lateral and frontal, are well-developed, and a stream of water carrying suspended food particles is continuously impelled through the pallial cavity, and particles strained out by passage between the gill filaments are deposited on the ciliated and glandular floor of the pallial cavity

at the right side. Here they are carried forward to the head, by strong ciliary currents along a narrow, rather deeply incised food groove (f.gr.). Material is conveyed to the neighbourhood of the mouth in a long, coherent mucus thread. The transport area is rather better developed than in Serpulorbis zelandicus, where a thin sheet of mucous secretion with embedded particles is carried along a broad, flat food tract. Ciliary currents proceed forward not only in the food groove but also over the whole lateral epithelium of the foot, on both left and right sides converging towards the mid-line on the scarlet pigmented area of the sole. In addition, there are strong currents beating around the periphery of the foot beneath the opercular attachment towards the dorsal mid-line, and wide ciliated tracts beating dorsally at the sides of the foot behind the black lateral patches. It is probable that all of these currents possess some food-gathering function, being freely exposed to alighting of particles when the foot is extended and the pallial current drawn to animal. They no doubt also have a rejectory function, especially the lateral tracts of the foot leading out of the pallial cavity, which carry faecal pellets and introduced carborundum particles to the mid-line of the foot, where enclosed in sheet of mucus they are presently released just below insertion of the operculum (see Fig. 1, f.p.). On the scarlet pigmented disc in front of the pedal tentacles, the mucus from the pedal gland is carried rapidly forward and towards the mid-line. The pedal gland secretion seems to have little connection with the contents of the food groove or with the rejection of waste particles.

Does Novastoa feed by mucus traps, and if so, to what extent does this method supplement ciliary feeding? All members of the Vermetidae (s. str.) possess a well-developed pedal mucus gland. The secretion of the gland does not appear to be employed either for the compaction of food collected by ciliary means, or for the rejection of waste particles. In some vermetids such as Serpulorbis gigas (Boettger, 1930) and Aletes (MacGinitie and MacGinitie, 1949) the gland is very large, and here mucous traps form the exclusive means of feeding. In Serpulorbis zelandicus, considered by the writer (supr. cit.) to engage in both mucus and ciliary feeding, it is also of considerable size. In “Vermetus” novae-hollandiae on the other hand, the gland is much smaller and narrower; in Novastoa (Figs. 1, 2) it is hardly larger, though even here still a very prominent feature, secreting a copious supply of mucus. It is invariably difficult to induce the formation of mucus food traps in the laboratory, but we shall probably not go far amiss in supposing that the great majority—if not all—of the vermetids make some use of the mucous gland in feeding in the natural environment. In none of the vermetids have the pallial organs become highly adapted for the collection of food—even in the ciliary feeding groups the filaments of the gill remain primitively triangular, and an endostylar tract is not developed. It is difficult to regard the mucous gland of “Vermetus” novae-hollandiae, Petaloconchus and Novastoa as undergoing reduction or loss of function. These genera on the contrary show every indication of being the more primitive section of the family, in which the gland probably arose, to be better developed in the more specialized Aletes and Serpulorbis, which have reduced their reliance on the gill. Fig. 1 shows the mucus mass extruded by