On Austrosassia, Austroharpa, and Austrolithes, New Genera; with some Remarks on the Gasteropod Protoconch.

Transactions and Proceedings of the Royal Society of New Zealand, Volume 62, 1931-32, Page 7

On Austrosassia, Austroharpa, and Austrolithes, New Genera; with some Remarks on the Gasteropod Protoconch. By H. J. Finlay, D.Sc. [Issued separately. 23rd May, 1931.] Genus Austrosassia nov. Type: Septa parkinsonia Perry, 1811. I Propose to replace the use of Austrotriton Cossmann in New Zealand by the above name. Iredale (Trans. N.Z. Inst., vol. 47, p. 459; 1915) applied Austrotriton to the Recent parkinsonia Perry, stating that it “stands quite alone when compared with the other Recent species, so that I make use of the generic fossil name, basing its use upon Kesteven's studies.” Later (Proc. Linn. Soc. N.S.W., vol. 49, pt. 3, p. 253; 1924) he re-affirmed this location, and on the strength of this I have used Austrotriton for a number of New Zealand Tertiary species (Trans. N.Z. Inst., vol. 55, pp. 453–455; 1924). All this depends in the first instance on Kesteven's remark (Proc. Linn. Soc. N.S.W., vol. 27, p. 454; 1902), “L. parkinsonianum is the recent representative of L. radiale, abbotti, textile, woodsi, and tortirostris. This group is more distinct than any I have studied.” Later (l.c., vol. 37, p. 49; Pl. 1; 1912) he again discussed and figured parkinsonianum, tortirostris, and abbotti to show their intimate relationship. Undoubtedly the last three species are related, but I think Kesteven's earlier “parkinsonianum-group” was a heterogeneous assemblage. The shell habit in all is somewhat similar, but the protoconchs of, say, woodsi and tortirostris are in my opinion fundamentally distinct. I think the only satisfactory basis for the classification of the Cymatiidae is the protoconch, and would reject a species from any of the shell groups if it does not agree with the other members in apex. This is in direct opposition to Kesteven's opinion (l.c., vol. 27, p. 470), “If the group is to be split up into various genera it will be impossible to disregard the form of the apices, now that we know so much about them; and we shall have such dissimilar species as L. costatum Born, and L. cornutum Perry, in the same genus, and species so absolutely alike as L. tortirostris and L. radiale in different genera. Considered as varieties of one type, they may, for the convenience of the monographer, be disregarded.” Emphatically no!—that is fatally easy, but unscientific. Reduction of this dictum to its essentials would end in classifying all gasteropods as varieties of one type. Kesteven's own results sufficiently attest the uselessness of this method; using it, he ended up by classing all the Tritons in one genus!

I would distinguish at least three general types of protoconch in this Family:— (A) Tall, narrowly conical, pointed, of about 4 whorls, set slightly oblique to the axis of the shell, entirely coated with a persistent brown horny envelope, strikingly differentiated in colour from the rest of the shell, e.g., Lampusia pilearis (Linn.). This is the same kind of apex as possessed by Harpa, Neothais, Janthina, and, in a wider and lower form, by Tonna and Cassis, and is very common among tropical molluscs. It almost certainly represents a Sinusigera, free-swimming embryo, such as Iredale has commented on (Proc. Mal. Soc., vol. 9, pt. 5, pp. 319–323; 1911). I endorse his dictum that “no species should be allotted to a group of which the type has been shown to possess a Sinusigera apex, unless it also possesses such an apex.” (B) Similarly polygyrate, but turbinate, with globose rapidly increasing whorls, usually wider than high, the tip minute and planorbid; set almost symmetrically on shell, polished, white, and shining, with no trace of a horny envelope, differentiated from normal shell only by inception of adult sculpture; in the Cymatiidae most examples of this type of apex also have a few distant linear spiral ridges on its lower whorls, e.g., Fusitriton kampylum (Watson), Cymatiella quoyi (Reeve), Austrosassia parkinsonia (Perry), etc. This is the type seen in Pseudotoma in the Turridae, and in Voluta, Yetus, Aulica, etc., in the Volutidae. (C) Paucispiral, of about two whorls, the last normal and much as in B, but the initial whorl quite irregular in shape, roughened, and markedly differentiated in texture from the polished succeeding whorl, beginning in a lateral blob or point which may be quite erect and sharp; the whole appearance is that of a scar left by the loss of some integral part. The conspicuous differences between this style of apex and the previous one are well illustrated by Harris (Cat. Tert. Moll. B.M., pt. 1; Pl. 6, Figs. 6 and 7; 1897) in his figures of the embryos of radialis and annectans Tate. This is the Scaphelloid, or in some cases even Caricelloid apex so common in the Volutidae; it is absolutely the apex of Voluta junonia, the type of Scaphella, but is not the type seen in undulata, which has been often quoted as the genotype. The scar seems to denote the loss of a previous entirely horny part, which did not envelope the whole embryo, calcification proceeding before this was completed. This type seems to be absent in Recent Cymatiidae, but was characteristic of the Australian Balcombian group to which the name Austrotriton Cossmann should be applied. There is a fourth type of apex, which I have not observed in the Cymatiidae, but which is common in other Families:— (D) Paucispiral, of 1½–2 whorls, rounded and globose, generally quite small but occasionally large, everywhere smooth and polished, generally marked off from adult whorls by a distinct varix, but otherwise not differing anywhere in texture or material from the normal conch, the initial whorl rather bulbous and asymmetrically placed; nowhere horny. This is the simplest type of apex, the true Fusid

one as seen in Colus, Cominella, etc., and, in a modified taller form, in many genera of the Cancellariidae and Pyrenidae. Evolution must proceed in the apex as well as in the shell, and as a working basis I suggest that the course normally followed is from Paucispiral calcareous (D)—Polygyrate calcareous (B)—Paucispiral part calcareous and part horny (C)—Polygyrate horny (A). The horny parts may and mostly do have a limy stratum underneath, but this when exposed lacks the regularity and polish of the normal shelly and horny surfaces. Dall (Tert. Fauna Florida, pt. 1, p. 66; August, 1890) remarks as regards the Volutidae, “There are several forms of the shelly nucleus. It undoubtedly preceded the horny one in this group.” According to this theory, Lampusia is more advanced in development than Austrotriton, which in turn has gone further than Austrosassia, while the simplest type from which this evolved is not known to us (perhaps a Muricoid genus). The geological age of these genera seems to support this; Lampusia being a Recent development, probably not antedating the Pliocene, Austrotriton being an Oligocene-Lower Pliocene genus, and Austrosassia dating back at least to the Eocene, but like many primitive lines, surviving even yet. It is but just, however, to note that arguments based on the geological age of lineages are always liable to be upset by new discoveries, so that this remains a pure hypothesis. This apical classification is of course not laid down as a primary division, for it would be absurd to say that all species which agreed in type of apex were congeneric. But I would definitely say that species which do not agree in type of apex cannot be congeneric. There are the nervous system, radula, operculum, and shell formation also to be taken into account, and exactly the same remarks apply to these. Which one of these is of prime importance must depend on circumstances; no hard and fast rules seem definable. Thus we must inevitably get a large number of genera based on different combinations of these five main variables; as in mathematics, if any four of these remain constant and the fifth varies, no equation results, and a distinct group ought to be recognised. To return to Austrotriton: Cossmann proposed this name (Ess. Pal. Comp., livr. 5, p. 98; Dec., 1893) chiefly on variation from the normal in shell features: he figured the apex (type C) of cyphus Tate, but did not realise its importance, as he included in the genus abbotti Ten.-Woods, which has a type B apex. Kesteven figured the apices of nearly all the Australian Tertiary species, but also did not realise their evolutionary importance, and, regarding them “as varieties of one type,” lumped every species in one genus Lotorium. His figures, especially of the polygyrate apices, do not show well the essentially different appearance of type A from type B. Triton woodsi Tate has the most typical Caricelloid apex (type C), but is closely simulated by radialis and gibbum, while armatum, cyphum, protensum, and tumulosum, all of Tate, have a lateral beginning, not raised into a spike, but still absolutely of type C. On the other hand, tortirostre and crassicostata Tate, and abbotti T.-W. have a well-formed type B apex, and are, as Kesteven noted,

the real relatives of parkinsonia Perry. Annectans Tate I have not seen, but if the specimen Harris figured was correctly identified, it belongs to this latter group. All the New Zealand Tertiary species whose apices have been seen also have this type. Since Cossmann definitely named Triton radialis Tate as type of Austrotriton, there is no alternative but to apply it to the type C group. A new name is therefore needed for the New Zealand and other type B species, and I therefore propose Austrosassia as above. It is with some hesitation that I separate this group from Sassia Bellardi, 1871 (Genotype: Triton apenninicum Sassi; a Pliocene species), which apparently has the same apex. Cossmann (l.c.) records species of this genus from Palaeocene to Recent, and placed therein tortrostris, annectans, tumulosus, protensus, oligostirus, and gemmulatus, all of Tate, an apically heterogeneous series. Kesteven (l.c., p. 472) stated that apenninicum “might, judging by Hoernes and Auinger's figures, be included in my quoyi-group.” Cossmann's figure supports this suggestion, and he gave quoyi, eburneus, and verrucosus Reeve as the Recent representatives of the genus. These are all members of Iredale's genus Cymatiella (Proc. Linn. Soc. N.S.W., vol. 49, pt. 3, p. 254; Oct. 24, 1924), typified by quoyi Rve. Sassia does not quite agree in shell habit with the Austral Cymatiella series, or with Austrosassia, so the three names are best treated as distinct until more is known of the lineage in Europe. Austrosassia procera nom nov. I propose this to replace Triton minimus Hutton, 1873 (Cat. Tert. Moll., p. 5), non Triton minimus Giebel, 1847 (Fauna d. Vorwelt, vol. 1, pt. 2, p. 188). I have described and figured the unique holotype (Trans. N.Z. Inst., vol. 55, p. 454; Pl. 48, Fig. 5), from Broken River, Trelissick Basin, and named the common Awamoan and Hutchinsonian New Zealand form Austrotriton maorium; this is very close to parkinsonia. Insignitum, cyphoides, and decagonia Finlay also all have this style of apex and shell, and may be referred to Austrosassia. Austrosassia reticulata (Suter). 1917. Streptosiphon (Streptopelma) reticulatum Suter, N.Z. Geol. Surv. Pal. Bull. No. 5, p. 17; Pl. 11, Fig. 25. Waihao River. 1924. Cymatium marwicki Finlay, Trans. N.Z. Inst., vol. 55, p. 456; Pl. 51, Figs. 1a, 1b. McCulloughs Bridge. These two specimens were described from the same locality, and I have no doubt that they represent the same species. Suter's description of the columella as having “a distinct oblique fold at the deflection towards the canal, and two smaller folds higher up” at first sight does not correspond well with my description of the columella as “straight, with five plaits on lower part, and a few rugosities above these.” But I have since observed that my Cymatium decagonium (l.c., p. 460; Pl. 48, Fig. 4) and the specimen described below both show two obsolete pillar plaits far within the aperture. These are not strong as in Semitriton, but like the rudimentary plaits

seen in some Recent species such as lotorium and femorale. I do not doubt that these same plaits exist in marwicki, and were either overlooked or are covered with matrix; every other detail of figure and description corresponds so well that the synonymy seems certain. A. maorium has nearly always only two strong elongated tubercles at the base of the pillar, one of them marking the twist; rarely there are a few denticles above these, but none below and no internal plaits. A. reticulata, and decagonia have two ridge-like tubercles, three denticles decreasing in strength below these, a few rugosities due to the basal spirals above them, and two (or perhaps more) low internal plaits. But as these details are more or less distinctly shown by tortirostris Tate, and the shells are otherwise so similar, separation does not seem justified. The pillar plaits of Semitriton are quite different. A. reticulata is the direct descendant of decagonia Finlay, though this was not noticed by me at the time of description. Additional specimens of each have since been collected, and demonstrate this fact; a perfect shell from McCulloughs Bridge differs in a few respects from the type, being more nodular, and having a strong keel; it is here described by comparison with decagonia to show the close alliance between these two:— Shell directly descended from decagonia Finlay, but more slender. Apex of four convex whorls, early ones flat, tip planorbid and minute, later ones globose and inflated, with four linear distant spiral ridges, merging without varix into adult sculpture. Spirals same in number and arrangement as in decagonia, but coarser, wider, and blunter, the interstitial riblets almost filling up the spaces instead of linear and distant. Axials same in number, but much stronger and closer together, 1½–2 times their width apart instead of 3–4. (Additional specimens of decagonia show that the axials are usually closer together than in the type; there are normally 7 intervariceal ribs, but another weak one may occur just before a varix). Tubercles considerably larger and more prominent. Aperture exactly as in decagonia. Height, 27 mm.; width, 13 mm.; height of spire, 13.5 mm. Genus Austroharpa nov. Type: Harpa pulligera Tate, 1889. Tate (Trans. Roy. Soc. S.A., vol. 11, pp. 149–152; April, 1889) described eight Australian Tertiary species of Harpa; lamellifera, sulcosa, spirata, cassinoides, abbreviata, tenus, pulligera, and clathrata, and later (Proc. Roy. Soc. N.S.W., vol. 27, p. 173; 1894) named another one, pachycheila, noting in this case that it belonged to the Section Eocithara. In 1896, Verco described (Trans. Roy. Soc. S.A., vol. 20, p. 218) a new Recent species of Harpa from South Australia as H. punctata, noting that “the two other Australian forms, H. ventricosa Lamk. and H. minor Lamk., are inhabitants of the warmer regions of the North and North-West parts of the continent.” In April 1899, Cossmann (Ess. Pal. Comp., livr. 3, p. 76) referred sulcosa, lamellifera, pachycheila, spirata, and tenuis to Eocithara, mentioning no Australian forms under Harpa s. str. Finally, Verco

in 1913 (Trans. Roy. Soc. S.A., vol. 37, p. 446) compared his punctata with the Tertiary species, examined all the types of these, and referred every one to Eocithara, regarding punctata as the only survivor of this otherwise extinct section. He quoted the differential characters of the section given by Cossmann, whose final words in this respect were, “Finally the riblets are more completely folded upon the suture, and cover it, joining one another, though this last character is less visible in the South Australian Eocithara, which have besides a more globular protoconch.” This is the only hint given of a fundamental distinction from both Harpa and Eocithara shown by several of the Australian Balcombian species. True Harpa has the same apex as described under Austrosassia, type A; this again seems to be a comparatively recent evolution, restricted to warm seas. Cossmann quotes one species from the Oligocene, and one from the Miocene, but these are almost certainly not truc Harpa; his Miocene species is quite certainly not, for his description of the protoconch of the genus, based on it, is “lisse, petite, globuleuse, à nucléus planorbulaire.” At the same time (l.c., p. 75) he described and figured the apex of Eocithara mutica (Lamk.), the genotype, as “lisse, globuleuse, composée de trois tours, à nucléus déprimé et obtus”; this is a type B protoconch, and would alone separate Eocithara from Harpa. Of all the Australian forms only one, H. lamellifera Tate, has this apex; it is large and turbinate, of nearly three (not two, as given by Tate) smooth, slightly shining, very convex whorls, rapidly increasing in width, and separated by canaliculate sutures, the tip somewhat prominent and globular, slightly intuned. Harris (Cat. Tert. Moll. B.M., pt. 1, p. 79; Pl. 4, Figs. 3a, 3b; 1897) gives a good figure of it and remarks that “Compared with the protoconch of ….H. mutica Lk…. it is relatively much larger and is more depressed though of the same general character.” H. pachycheila, and perhaps cassinoides Tate (which I have not seen), would, if belonging to this group, have apices referable to a Variety of type D, of about two smooth and shining shelly whorls, the first very inflated, upturned, and somewhat immersed (compare this with the apex of Hypocassis Iredale, later described). But I suggest that these two species would be better referred to Family Cassididae, perhaps as Oniscidia. Nearly all the remaining species have an apex quite different in character from Harpa and Eocithara. It consists of a well defined last volution, of not quite one whorl in extent, distinctly but not prominently marked off from adult sculpture, narrow, very lightly convex, smooth and shining, sides almost perpendicular; surmounted by an initial whorl in the form of a more or less inflated dome-shaped blob, of indefinite coiling, roughened and pitted, appearing like the top of an egg, most of which is immersed in the succeeding whorl, from which it is distinctly marked off by a very oblique dividing line in good specimens. It is exactly like the apex of Pterospira Harris in the Voutidae, with the initial egg-whorl more immersed. Tate's original figure of pulligera (Trans. Roy. Soc. S.A., vol. 11; Pl. 6, Fig. 9; 1889) and Harris's good figurcs of the protoconchs of tenuis and

abbreviata (Cat. Tert. Moll. B.M., pt. 1; Pl. 4, Figs. 4, 5) give a clear idea of the formation of this peculiar embryo. It will at once be seen that this is a type C apex, quite comparable in general style to that of Austrotriton, though different in detail. Of the Balcombian species, pulligera, sulcosa, abbreviata, tenuis, and the Recent punctata certainly possess it, spirata probably has it, while clathrata I have not seen. My arrangement, then, of the Australian Harpidae would be as follows:— Genus Harpa Pallas, 1774. (Spic. Zool., vol. 10, p. 33). Genotype (by tautonymy): Buccinum harpa Linn. Harpa Humphrey, 1797; Mus. Calonn., p. 18. Harpa Bolten, 1798; Mus. Bolten, vol. 2, p. 149. Harpa Lamarck, 1799; Mem. Soc. H.N. Paris, p. 71. Harpalis Link, 1807; Beschr. Nat. Samml. Univ. Rostock, vol. 3, p. 114. Harparia Rafinesque, 1815; Analyse de la Nat., p. 145. 1. Buccinum harpa Linn., 1758; Syst. Nat., ed. 10, p. 738; ed. 12, p. 1201, 1767. (= major Bolten, 1798, = ventricosa Reeve, 1843). (Hedley, in his Queensland list, also included, probably on Melvill's authority, H. amouretta Bolten, 1798 = minor Lamk., 1822, but Iredale in Mem. Queensland Mus., vol. 9, pt. 3, p. 283, June 29, 1929, has discussed the matter, and there makes the “first authentic record” for North Australia, a juvenile of harpa Linné). Genus Eocithara Fischer, 1883. Man. Conch., p. 601). Genotype (by original designation): Harpa mutica Lamk. 3. Harpa lamellifera Tate, 1889; T.R.S.S.A., vol. 11, p. 149; Pl. 6, Fig. 2. Genus Austroharpa Finlay, n. gen. Genotype: Harpa pulligera Tate. 4. Harpa punctata Verco, 1896; T.R.S.S.A., vol. 20, p. 218; Pl. 6, Figs. 3, 3a, b. 5. Harpa sprata Tate, 1889; Id., vol. 11, p. 150; Pl. 6, Fig. 3. 6. Harpa abbreviata Tate, 1889; l.c., Pl. 6, Fig. 7. 7. Harpa sulcosa Tate, 1889; l.c., Pl. 6, Fig. 10. 8. Harpa tenuis Tate, 1889; l.c., p. 151; Pl. 6, Fig. 1. 9. Harpa pulligera Tate, 1889; l.c., Pl. 6, Fig. 9. 10. (?) Harpa clathrata Tate, 1889; l.c., Pl. 6, Fig. 8. 11. Austroharpa tatei Finlay, n. sp. (described below). The other genera of the Family belong to the Cryptochordinae, and do not occur in Australia. They are as follows: Cryptochorda Moerch, 1858 (genotype: Buccinum stromboides Herman.; Eocene), Silia Mayer-Eymar, 1877 (Umgeg. von Einsiedeln, p. 59, genotype:

Silia zitteli M.-E., n. sp.; Eocene—Dall says this is known only from an internal cast), Lyrianella Nelson, 1925 (Butt. Dept. Geol. Sci., Univ. Cal. Pub., vol. 15, No. 11, p. 432, genotype: Cryptochorda lyrata Nelson, n. sp.; Eocene), and Tritonoharpa Dall, 1908 (Bull. Mus. Comp. Zool., vol. 43, No. 6, p. 319, genotype: T. vexillata Dall, n. sp.; Recent). One may note in conclusion that Cossmann and Pissaro (Mem. Geol. Surv. India, Pal. Indica, N.S., vol. 3, Mem. 1, p. 22; 1909) described a Harpa morgani of which Vredenburg remarks (l.c., vol. 10, Mem. 4, p. 35; 1925), “In Cossmann and Pissarro's memoir this species has been compared also with Harpa jacksonensis G. Harris from the Upper Eocene of Mississippi, and with the numerous fossil species from the Australian Tertiary. The latter are stated to differ on account of their more numerous or more overlapping lamellae. This difference does not apply in the case of Harpa cassinoides Tate from the Murray Desert, which however is readily distinguished by its much more ventricose shape.” But as elsewhere on the same page Vredenburg remarks that morgani is hardly separable varietally from Harpa mutica Lk., and describes the same sort of protoconch for it, it is evident that this Indian Eocithara is not comparable with the Australian Austroharpa. An Oligocene species described later by Vredenburg (Mem. Geol. Surv. India, vol. 50, pt. 1, p. 122; 1925) as Harpa (Eocithara) narica is still an Eocithara, and Vredenburg says of it, “In several respects the Nari shell (narica) seems somewhat intermediate between Harpa morgani and the living Harpa conoidalis Lamk., of which it perhaps represents an ancestral form.” Austroharpa tatei n. sp. Extremely close to sulcosa Tate, and evidently derived from it. Somewhat smaller and more solid. Apex exactly same type but smaller. Shoulder of whorls flat, not concave as in sulcosa, the bordering angle not sharply ridged, but bluntly rounded. Three broad low spiral bands on spire whorls and 10 on body whorl, as against 4 and 14 respectively in sulcosa; interstices but little narrower than ribs instead of considerably narrower. Thirty-three narrow crisp axial lamellae over-riding spirals on body whorl, as against 45; not projecting as lamellose hollow points on shoulder keel. Fasciole with the S-shaped lamellae less prominent on the ridge, more prominent in the umbilical excavation. Aperture narrower. Height, 25.5 mm.; width, 17 mm. Locality—400–500 ft., Abbatoirs Bore, Adelaide, South Australia (“Older Pliocene”). Type in Finlay collection. Genus Austrolithes nov. Type: Fusus bulbodes Tate, 1888 (as restricted by Pritchard). Pritchard (Proc. Roy. Soc. Vict., vol. 17, pp. 320–324; Sept., 1904) showed that Tate confused two species when he described his Fusus bulbodes (Trans. Roy. Soc. S.A., vol. 10, p. 139; Pl.7, Fig. 8; 1888). He separated these as Clavella bulbodes (Tate), typical,

common at Balcomb Bay, and widespread in the Balcombian, and C. platystropha sp. nov. from Muddy Creek lower beds. The latter he remarked was “very much more of the type of C. longaevus than any other described Australian species.” Tate (l.c., p. 141) also remarked of Fusus tateanus T.-W. that it was “remotely related to F. longaevus of the European Eocene.” One other species, Fusus incompositus Tate (l.c., p. 137; Pl. 3, Fig. 9) is generally referred to Clavella or Clavilithes by Australian conchologists. The remarkable series of Eocene shells classed as Clavella s. 1, in older literature has been ably discussed by Grabau (Phylog. of Fusus, pp. 98–144, 1904). He has divided this series into four main groups, Clavilithes typical, Clavellofusus, Rhopalithes, and Cosmolithes. The last three have tiny protoconchs, and the latter two have plicate columellas. The Australian species show a superficial likeness to these “phylogerontic Fusidae,” but even in conch characters are not quite the same as the English and Paris Basin shells. The sutural shoulder, so strongly developed on the later whorls of the European forms is absent, the whorls clasping each other much more tightly; the basal angulation is more rounded and much less conspicuous; and the pillar is long and very regularly decreasing, without a strong twist at the beginning of the canal as in true Clavilithes. But apart altogether from these differences, the character of the protoconch renders separation imperative. The only genus we need consider in this respect is Clavilithes itself; Grabau says of it (l.c., p. 104), “The protoconch of this genus is very striking and is distinctive…. The first whorl is depressed and naticoid, with a minute apical portion. The whorl gradually enlarges, but after the first volution the proportional increase in size is much less, so that the whorls produce a nearly cylindrical protoconch. There are from two and a half to four whorls thus giving the protoconch a distinctively papillose appearance.” The British species have, according to Grabau, uniformly a slightly different apex from the Parisian ones; the apex of deformis Solander he describes thus (l.c., p. 120): “The protoconch is much larger and more robust than is even the case in the French specimens of the genus. Its median whorl has a diameter of nearly 4.5 mm., while the average diameter of the median whorls in the French species is less than 3 mm., seldom exceeding 2.5 mm. In one specimen from Barton the diameter of the median whorl was found to be 5 mm…. The number of volutions varies from three to nearly four, and they almost always show an irregularity in thickness. A characteristic feature not found in the French species is the flattening of the upper exposed portion of the protoconch, thus giving a sloping or trochiform character to the apex. The apex of the protoconch of the Parisian species is naticoid with the convexity of the whorl unimpared.” Compare with this the description of the apex of Fusus bulbodes; Tate says “terminating in a large ovoid summit,” while Pritchard notes the “remarkably large mammillate embryo.” This inadequately describes the most curious feature of this species; Tate's figure shows it moderately well, but all Pritchard's figured shells are decollate. The apex is really of about two whorls, the second quite narrow, with

straight sides, a trifle angulated near lower suture, somewhat shining, and towards its close with gradually more distinct spiral scratches which become the grooves of the normal whorls, merging imperceptibly into the adult conch; the first whorl relatively enormous, egg-shaped, well separated from the second whorl by a change in texture and irregularity in suture line, running in same spiral as second whorl for but a short distance, then suddenly rising erect and expanding into a large bulb, which curves over and has its nucleus immersed in its own volution and that of the second whorl, the whole surface coarse, chalky, and pitted; the bulb measures 5 mm. across and nearly 6 mm. high. This is exactly the embryo of the Volutid genus Pterospira Harris, except that it is smaller and has the nucleus totally immersed and hidden instead of lateral. Here again, then, is a kind of type C apex, but the initial horny-covered stage entirely dwarfs the remaining normally calcareous volution. Still, there is no apical difference further than relative size of the two whorls between Austrotriton, Austroharpa, Austrolithes, and Pterospira; a regular and most interesting gradation is shown in the comparative size of the “blob” in these four genera. The protoconch of Clavilithes is truly Scaphelloid, being quite like that of Amoria Gray in the English forms, or some species of Alcithoe in the case of the Parisian ones. It is, therefore, also a type C, but is so different in shape and formation from that of bulbodes, that close connection seems improbable. Here, then, is a case of two genera of a group having apices of the same type but of quite different formation, which is probably intimately connected with ovicapsular conditions. Perhaps there are two series in Australia; no apex of platyslrophus or tateanus has been recorded, and my specimens lack it; incomposita (in my only specimen from Aldinga) has a smaller embryo with more than a whorl of rough texture before the nucleus is immersed. It may be a true Clavilithes, but till more material is examined is best referred with the other species to Austrolithes. Grabau has explained that the correct name of the genotype of Clavilithes Swainson, 1840 (Treat. Mal., pp. 90, 304) (= Clavella Swainson, 1835; non Oken, 1815) is Fusus (Cyrtulus) parisiensis Mayer-Eymar, Fusus longaevus Lamk., 1803 (for which Swainson instituted the genus) being preoccupied by Solander in 1766. Family Struthiolariidae. This exhibits another case of very similar shells with fundamentally different embryos. Here, however, only two types have been observed. Monolaria Marwick, Struthiolaria Lamk., and Callusaria Finlay have a 2½-whorled regularly coiled shelly almost planorbid nucleus, the tip minute and symmetrical—this may be referred to type B. Pelicaria Gray, on the other hand, has the regular diminution of the apical adult whorls suddenly interrupted by an irregular roughened blob of shelly material, which is quite Scaphelloid in character, and was evidently deposited inside a horny coating. The Australian Tylospira Harris at first sight appears related to Struthio-

laria in its nodose ornament, but the apex (as seen in some juveniles of scutulata dredged in 6–12 fathoms, Twofold Bay, N.S.W) is exactly like that of Pelicaria, only considerably smaller. On inspection, it is also seen that the ornament of Tylospira is really more similar in style to Pelicaria than to Struthiolaria, so that a close alliance between the two former genera is indicated. The earliest Pelicaria so far known is Struthiolaria nana Marwick, 1926 (Trans. N.Z. Inst., vol. 56, p. 318, pl. 73, Fig. 4) from Tirangi Stream, a Taranakian horizon, while the oldest Pelicaria is probably the new species described below from the “Older Pliocene” of the Adelaide basin. These two deposits are probably approximately equivalent in age, though there is no resemblance between nana and marwicki. This is interesting and suggestive in view of the hypothesis advanced by me in another paper in this volume (“On Turbo postulatus Bartrum, etc.”). This Pelicaria -Tylospira apex is probably a variety of type C. After the initial rough blob, fine spiral sculpture starts immediately, and if there is another whorl to the protoconch, as in typical type C, it is ornamented precisely as the normal conch, grades into it, and cannot be detected. But as there is probably all gradation in apices of this kind from those entirely coated with chitin to those which have a subsequent shelly portion, I include under type C at present all those embryos generally known as Scaphella, Caricella, or Cymba nuclei. Three very tiny apical fragments of T. scutulata in quite fresh condition, from 6–12 fathoms Twofold Bay, N.S.W. presented exactly the same condition as New Zealand Recent and Tertiary species of Pelicaria, the blob being the same colour as the rest of the shell and showing no trace of horny matter; its rough surface, however, almost certainly indicates the previous presence of such a coating, lost while still in the ovicapsule, as in Aurinia and other Volutidae. Pelicaria marwicki n. sp. Shell closer to coronata Tate than to clathrata Tate, but, like these two, much shorter in the spire and with a more thickened aperture than the Recent scutulata Mart. It has the short spire of clathrata, but no lamelliform axials; whorls angled, and growth lines and outer lip far more sigmoid. Related to coronata in general habit of shell and callus deposit (which extends over whole of body whorl, and sometimes ¼ of next whorl; clathrata has only two-thirds of a whorl callused over) and the tendency to sulcation of the suture anteriorly, but this never goes to the same depth and extent, the callus shoulder remaining mostly as a flat or lightly concave platform. Shell more squat than coronata, spire being lower than aperture; angle of whorls much less developed and lower down, almost entirely hidden by callus on penultimate whorl. Peripheral knobs about the same in number, but very much weaker; small sharp tubercles instead of projecting nodules. Apertural callus very heavy, forming three marked “lips,” one projecting upwards and outwards at suture, another forming a pad on base, a third at end of the strongly marked fasciole. Columella very much bent (almost a semicircle) to the right.

Height, 38 mm.; diameter, 31.5 mm. Locality—400–500 ft., Abbatoirs Bore, Adelaide, South Australia (“Older Pliocene” of Tate) eight shells. Type in Finlay collection. This is probably ancestral to both clathrata and coronata, which are from the Gippsland coast Kalimnan beds, and, through coronata, to scutulata Martyn. The significance of these and other types of gasteropod protoconch remains to be demonstrated by biological research. Some of the questions that need investigation are:— (1.) What is the reason for the change from shelly to horny embryo, and can any artificial conditions bring it about? (2.) Why are the horny types apparently more advanced, and in general more characteristic of later geological epochs? (3.) Why are the type A apices practically confined to tropical and sub-tropical seas and so abundant there, while the much older type C embryos are more widely spread? Did the horny type of apex evolve on account of tropical conditions? (4.) Is the type A apex a further evolution from type C or is it very distinct? Why is the chitin so persistent in it and so easily and early lost in type C. (5.) What is the significance of the change from paucispiral type D to polygyrate type B in the shelly series, and from paucispiral type C to polygyrate type A in the horny series? Do these different embryos show any notable difference in swimming powers, and hence distribution? (6.) Is there any relation between the fecundity, stability, and adaptability of the species and its type of embryo? These are questions highly interesting from a systematic and evolutionary point of view. Dall has fairly fully discussed the appearance and types of horny apex in the Volutidae, but the matter has not gone much further. I now suggest that the Scaphelloid and pseudo-Scaphelloid apex is more common in the mollusca than was previously thought. In the present paper I have shown its existence in Austrotriton (Cymatiidae), Austroharpa (Harpidae), Clavi-lithes and Austrolithes (Colidae) and Pelicaria and Tylospira (Struthiolariidae), while elsewhere (Trans. N.Z. Inst., vol. 57, p. 414, 1926) I have noted its occurrence in Glaphyrina (Fasciolariidae). Finally, comparison may be made with the protoconchs of Columbarium von Martens and Hypocassis Iredale, where the apex has the same bulbous and irregularly globular first whorl, turned almost at right-angles to the plane of the next, the nucleus immersed; but the embryo is polished and smooth throughout as if there had been no horny envelope. This seems to be a form of the type D apex in which the initial whorl is of irregular growth. It may be remarked that Hypocassis Iredale (Rec. Austr. Mus., vol. 15, No. 5, p. 329, April 6, 1927), which was founded on shell characters and the existence of “a long lineage in the south of Australia” (genotype: Cassis bicarinata var. decresensis Hedley), is perfectly distinct from Cassis in its apex alone. The large wide

polygyrate horn-covered embryo of Cassis s.str. (a form of type A with the same shape as type B) is here replaced by a relatively small blunt paucispiral shelly apex, the first whorl distorted and immersed. (Apex described from the Balcombian Cassis exigua T.-W., which Iredale says is a fossil representative of bicarinata). This is almost identical with the apex of Columbarium as seen in the Balcombian acanthostephes (Tate). After a number of years careful examination of gasteropod apices, I am fully satisfied, in spite of what several authors have written, that the protoconch is one of the most valuable criteria for systematic classification. Not only have I never found it to vary from type in a homogeneous genus, but I have also found it so generally constant that in my opinion considerable importance must be placed on it in determining lineage relationships. To the palaeontologist it is as important as the radula is to the malacologist, and should be given just as much consideration.