and serrated by a row of fine denticles, much less pronounced than in Perissodonta. The marginals in Tylospira are rather short and blunt, and they have no fine denticulations. In Pelicaria the denticulations on the marginals are much reduced, but on most teeth still easily detected. In examining struthiolariid radulae it is important that teeth be selected which have come to lie as nearly flat as possible so that the cusp may be observed at its greatest length. Some of the older figures, such as that of Troschel (1856) contain examples of marginal

Fig. 1.—Portions of a single row of unworn radular teeth of Aporrhais pes-pelicani (a), Perissodonta mirabilis (b) and Tylospira scutulata (c). The lateral and marginals of one side are in each case omitted or interrupted, and in Tylospira scutulata a single central tooth, from further forward, with the cusp worn, is shown for comparison with the intact tooth. The scale in each case represents 0.2 mm.

Text-Fig. 2. — Diagrammatic views of the undersurfaces of the opercula of (a) Aporihais pes-pelicani, (b) Perissodonta mirabilis and (c) Tylospira scutulata. The thickened chitinised rim, produced in the Struthiolariidae to form the spine, is shown stippled.

rule, a clumsy, large apex points to a non-pelagic development, while a narrowly twisted apex, often with delicate sculpture, points to a pelagic development.”

In none of the perissodontid shells that the writer has examined was the apex free from erosion, but in one specimen of Perissodonta georgiana the protoconch was reasonably intact and was found to be small and closely coiled, as in Struthiolaria s. str. The four available animals of P. georgiana were unfortunately all males, and no information could be gained as to the presence or absence of an incubatory pouch. In was, however, previously assumed (Morton, 1950) on the evidence of the protoconch of P. georgiana that the larval life was in this genus unabbreviated, resembling that of Aporrhais (Lebour, 1938) or that—at least—as in Struthiolaria s. str. there remained a fairly lengthy free-swimming stage in the life history. The specimen of P. mirabilis lent by Dr. Rees was a female, and the structure of its reproductive organs greatly strengthens the view that Perissodonta closely resembles Aporrhais in its reproductive habits. The edge of the mantle (Text-Fig. 3a) has no incubatory pouch, in striking contrast to Struthiolaria and Pelicaria as described by Morton (1950). The ciliated oviducal groove, by which the egg path is prolonged forward from the opening of the capsule gland, terminates not—as in the modern struthiolariids—at the end of the food groove and opposite the incubatory pouch, but runs forward to the edge of the sole (Text-fig. 3a, Ov G A). By this means, in the Aporrhaidae and in some at least of the Strombidae, the eggs are carried forward to the level of the substratum where—as shown by Lebour (1938)—they are attached separately to sand grains. This is the basal condition of the reproductive organs in

the more primitive Strombacea, and for those Struthiolariids that retain it, we may reasonably predict a fairly long larval life, the eggs having comparatively little yolk and being deposited early. This supposition in turn helps to explain the surprisingly wide distribution of the genus Perissodonta at the present day—or more properly, the wide separation of its two remaining species, a fact greatly at variance with the very localised distribution of the three remaining struthiolariid genera.

In Tylospira scutulata, the female genital system is scarcely distinguishable from that of Struthiolaria and Pelicaria. The ciliated oviducal groove, leading forward from the capsule gland, terminates at the end of the food groove, while on the right side of the inner surface of the mantle opens a large, ovoid incubatory pouch (Text-Fig. 3b Br). The aperture of the incubation pouch is able, when the head is retracted into the mantle cavity, to be held closely against the termination of the ciliated oviducal groove, as in Struthiolaria (Text-Fig. 3b). The eggs are at this stage able to be transferred to the pouch, while at the hatching and escape of the young, the pouch is able, as in Struthiolaria, to be extended freely across the lip of the shell, by the expansion of the mantle. No eggs were found in the incubatory pouch of Tylospira, but from the appearance of the capsule gland, which is long and narrow and “segmented” into a succession of pockets as in Struthiolaria, it seems likely that the egg capsules are rather small and attached end to end in a string, in contrast to Pelicaria, where there is a small number of larger capsules, each of which in turn fills the whole capsule gland. Whether or not Tylospira has lost the free-swimming stage must remain uncertain, though the evidence of the bulbous scaphelloid apex would strongly suggest an abbreviated life history. The apical bulb is, however, a little smaller than in Pelicaria, and this fact, together with the resemblance of the capsule gland to that of Struthiolaria, would suggest that the amount of egg yolk in Tylospira is smaller than in Pelicaria. It is probable—on the whole—that in Tylospira, the young are retained in the incubatory pouch for the full duration of their development. The restricted range of Tylospira would suggest this to be so, and it is worthy of remark how closely the extent of the geographical range in the four existing genera of the Struthiolariidae reflects the stage of evolution reached in their reproductive habits.

The reproductive habits of the living genera of the Struthiolariidae provide an interesting exception to the general rule propounded for the prosobranchs by Thorson (1950). He states (p. 33) “all high arctic prosobranchs hitherto examined proved to have a non-pelagie development, and not a single larval prosobranch is known from arctic or antarctic plankton. As, in contrast to this, in the tropical marine area round the Bermudas up to 85% of the prosobranchs reproduce by pelagic larvae … this group of animals seems to be a very fine ‘barometer’ of ecological conditions.” The Struthiolariidae, alone, it seems, among the prosobranch families of which we have record, appear to reverse this rule: the antarctic and subantarctic genus Perissodonta is the only group with an unabbreviated life history, and among the genera of warmer temperate seas, two—namely, Tylospira and Pelicaria have a non-pelagic development. Thorson's rule—while undoubtedly holding good generally—requires modification in some points. As well as ecological conditions, there can be no doubt that phylogeny also plays its part in determining the character of the apex. Neglecting at first the influence of sea temperature on the type of larval life, it must be remembered that pelagic development with a small multispiral apex is likely to be the primitive condition in a given family, as it is among the prosobranches as a whole, and that abbreviated development, with a large, bulbous apex, is a derived condition which may, or may not, be developed in later forms. Finlay (1931), without laying stress on the influence of ecological conditions, makes clear his belief that the “scaphelloid” apex is derived from the multispiral type, and never the reverse. We may thus assume that evolution within a prosobranch genus or family moves from a closely coiled to a bulbous condition of the apex; and that ecological conditions come into play in determining the rate at which such a change will take place and the extent of its occurrence within a given group. Finlay's view that these two types of apex will never occur in different species of the same genus would be open to objection as taking a rather too restrictive view of the evolutionary range possible within a generic group. The English species of Littorina provide an example of differences in mode of development and form of apex existing within a single genus; so does the case discussed by Thorson (1950) in which two species of Natica are cited with pelagic development, two with non-pelagic. In the Struthiolariidae, if the living species Struthiolaria papulosa and Pelicaria vermis were considered alone, without knowledge of the long separation of their Tertiary ancestors, it would be difficult indeed to justify their location in separate genera.

In the present case, it may safely be held, following the work of Marwick (1923) and, especially, Finlay and Marwick (1937) that the Struthiolariidae are a southern derivative of the Aporrhaidae that originated on the shores of an antarctic land mass and achieved their present distribution by radial spread. It is thus the two most southern living species, Perissodonta georgiana and P. mirabilis, that must be considered closest to the primitive stock of the family; and it is in these species that the multispiral apex, with unabbreviated development has survived as a primitive condition, notwithstanding the apparent unfavourableness of the environment for such a life history. Conversely, and in contrast to Thorson's rule, the latter genera of the family, Struthiolaria, Pelicaria and Tylospira, and in spite of their warmer habitat, have each progressed, wholly or partially, towards that suppression of larval life characteristic of derived genera.