locality has its own type of colour pattern; curiously, the Tasmanian shells, though near group 2 in shape, have patterns very like those of the Port Phillip Bay shells.

[The section below cannot be correctly rendered as it contains complex formatting. See the image of the page for a more accurate rendering.]

Following Iredale's (1924, p. 210) observations at the British Museum, the stout shells of group 1 can be taken as kingi and the more elongate ones of group 2 as subspecies lamarcki Gray. Gray's original description of Cytherea kingi reads: “Shell ovate, heart-shaped, white or pale brown with dark rays, each formed of several narrow lines, the umbones white, 1 inch long, 8/10 inch high.” This description suggests that it was written from juveniles of the Port Phillip Bay kind.

The writer's former description of Notocallista (Marwick, 1924, p. 592) was based on specimens of the New Zealand multistriata and parki, together with a specimen of the Tasmanian diemenensis wrongly identified as kingi. It is therefore quite misleading.

The surface of the adult kingi is almost smooth, being very irregularly and weakly concentrically striated, even on the juvenile such concentric striae as are developed are irregular. The periostracum on some specimens, e.g., from Twofold Bay, New South Wales, is well developed and has a high glaze, the posterior area in many shells bearing fine radial wrinkles like those of V. chione L. The periostracum of others, however, especially some from southern districts such as Port Phillip Bay is not firmly attached and peels off, leaving the shell with a dull chalky-looking surface very like that underlying the periostracum of V. chione.

The hinge of N. kingi appears to the writer to agree in every significant respect with that of V. chione. Take Jukes-Browne's criteria: there is no well-defined channel leading back from the anterior lateral pit; a shallow depression, about normal to the long axis of the pit, leads down to the inter-cardinal socket, the anterior cardinal being undercut to the same small extent in both species; the left posterior cardinal of kingi, as of chione, is relatively short, thin, and closely joined to the nymph.

Owing to its broader hinge-plate, the cardinal teeth of kingi (s. str.) are in general relatively longer than those of chione and their set differs (as measured by the more nearly vertical anterior cardinal). That the difference is unimportant is shown by the hinge of subspecies lamarcki, especially of the lightly built Gulf of St. Vincent shells. (See Pl. 11, figs. 2, 3.)

The left anterior lateral tooth of both kingi and chione continues up towards the umbo as a definite ridge, simulating a cardinal. The presence of a similar “buttressed” tooth in the West American Oligocene Pitar arnoldi Weaver was the outstanding feature used by N. M. Tegland (1929, pp. 276, 280) to differentiate Katherinella as a new subgenus of Pitar which was said to lack the buttress.

Grant and Gale's suggestion (1931, p. 347), that the anterior lateral of the Pitarinæ is really a cardinal tooth moved forward is rather attractive when one considers a hinge like that of Notocallista, in which the tooth in question extends well up under the umbo.

Ontogenetic evidence, however, is against the idea. On young multi-striata and parki of 2 mm. diameter the left anterior lateral is a small, bluntly conic tubercle without any ridge or buttress, apparently originating as an independent growth on the hinge-plate. The buttress develops later and becomes quite prominent (Pl. 10, figs. 1—3). Nor does phylogenetic evidence support the cardinal nature of the anterior lateral. Cretaceous species do not have the tooth buttressed any more than, if as much as, many later species do. The buttress has apparently developed as the anterior margin of the socket into which the right anterior cardinal fits.

The pallial sinus of both kingi and chione is obliquely truncated to form a rather sharp point though that of chione is less ascending, and bulges out along its lower boundary to be more nearly equilateral.

Finally, in both species the pedal-retractor scar is well separated from that of the anterior adductor.

In view of this remarkably close agreement in essential characters, one is tempted to conclude that the two species are quite closely related, differing chiefly in shape, and that Notocallista must take the place, without the seniority of Callista; which is not legally available (Stewart, 1930, p. 239).

The position, however, is not so simple, and when the species of the two new groups Striacallista and Fossacallista are considered, the European chione seems with equal if not greater probability to be a remarkably close parallelism to Notocallista. Fossacallista is an Oligocene and Miocene group that lived in Australian and New Zealand seas. Besides having a hinge closely similar to that of Chionella (Pl 10, figs 9, 10), it has a sunken ligament, confluent pedal retractor (Pl 14, fig. 10a), and ascending pallial sinus with a rounded end. The left posterior cardinal is not so well separated from the nymph as that of Chionella, but in some species a well-defined though shallow groove separates them. Fossacallista grades through such species as N. mollesta n.sp. into Striacallista, which has a high ligament, separate pedal retractor (Pl 12, fig. 5a) and truncated sinus; and from Striacallista the change to Notocallista, which also has these characters, is mostly one of increase in size and obsolescence of sculpture.

If the parallelism of kingi and chione is an actual fact, then Notocallista must be a recent development from Striacallista for it is not known as a fossil. On the other hand, if chione is really as closely related to kingi as it superficially appears to be, that is, if it is a Notocallista, then the European fossil occurrences take the group back to the Miocene, and the American probably to the Eocene.

The geological factor favours solving the problem by parallelism, of which many cases are known in the family.

To trace the origin of the European chione, account must be taken of the strongly sculptured Costacallista. The hinges of species of this group are remarkably constant and agree in every respect with that of chione. The groups also agree in size, shape, lunule, ligament, pedal retractor and pallial sinus. Costacallista has been traced back by Palmer to weakly sculptured species in the Eocene [C.