Serrata simplex, Boyer, 2008

Boyer, Franck, 2008, The genus Serrata Jousseaume, 1875 (Caenogastropoda: Marginellidae) in New Caledonia, Mémoires du Muséum national d'Histoire naturelle 196, pp. 389-436 : 425-436

Serrata simplex n. sp.

Fig. 52

TYPE MATERIAL. — Holotype (dd) MNHN 20630 .

TYPE LOCALITY. — Southern New Caledonia, 23°03’S, 166°58’E, 397-400 m [BATHUS 2: stn DW 730] .

MATERIAL EXAMINED. — Southern New Caledonia. BATHUS 2: stn DW 730, 23°03’S, 166°58’E, 397-400 m, 1 dd (holotype,

Fig. 52).

DISTRIBUTION. — Southern New Caledonia, one shell in 397-400 m.

DESCRIPTION. — Shell fusiform, thin, opaque. Protoconch paucispiral, low. Spire high, massive, whorls faintly convex. Aperture moderate, subvertical, base wide. Outer lip thin, vertical, weakly arched, shoulder sloping, low, outer margin wide, flat, weakly stepped, inner edge smooth. Four thin columellar plaits, decreasing a little in size from the lowest to the uppermost one.

Colour dirty white (faded).

Dimensions: 4.30 x 2.10 mm.

Radula unknown.

REMARKS. — Serrata simplex differs from S. robusta (Figs 44, 48) in its much smaller size, its less oval, more pyriform outline, its wider aperture with a dilated siphonal canal, its thin, smooth, non-reflected labrum, the outer margin without a bevelled dorsum, and the lack of an upper columellar pleat. Despite its juvenile appearance resulting from the thin, smooth labrum, the holotype of S. simplex seems to be an adult specimen with a perfectly mature outer margin.

Serrata simplex differs from S. tenuis principally in its slightly more inflated outline and its wider aperture. It differs from the biconical-fusiform form of S. magna (Figs 46-47) principally in its smaller size, its less produced spire and its subvertical straight labrum.

In its rather wide aperture, its thin, smooth and rather straight labrum and its produced spire, S. simplex resembles the S. hians group. But in its rounded base, the shape of its first columellar plait, its tiny size and its upper bathyal distribution, S. simplex resembles the S. boucheti group better.

ETYMOLOGY. — Latin simplex (adj., not complicated), referring to the simple morphology of the shell.

DISCUSSION

SPECIFIC DIVERSITY AND GEOGRAPHICAL DISTRIBUTION OF SERRATA

DIVERSITY AND DISTRIBUTION IN NEW CALEDONIA

The genus Serrata is now known from New Caledonia by 35 endemic species, and seems to be restricted to bathyal depths in this area. Twenty seven (77%) of the 35 species have been collected at only 1 or 2 stations, and 18 of them (51%) at only 1 station. Only 5 species (14%) are represented at more than 3 stations. Seventeen of the 35 species (49%) are known by only 1 to 3 specimens, and 11 of them (31%) by only 1 specimen. Therefore the high diversity is certainly still underestimated, despite intensive sampling during the last 20 years by successive French oceanographic expeditions in the New Caledonian Exclusive Economic Zone, including intensive workshops devoted to sampling reef formations.

Understanding of the interspecific affinities of the Serrata fauna of New Caledonia remains very sketchy and, despite the evident occurrence of several species groups, the formal subdivision of this genus is not possible at present. The S. boucheti species group is provisionally considered to include 16 species occurring in the upper bathyal zone (11 of them in the 200-480 m zone, the other 5 extending down to 500-775 m). The S. hians species group is provisionally considered to include 6 species distributed mainly in mid-bathyal depths (most in 675-1070 m, except 2 stations at 360 m and 570 m for S. hians ). Besides these two broad groups of species, several individual species seem to be morphologically isolated, such as S. arcuata (despite its tiny size and subcylindrical outline, which link it to the S. boucheti group), S. sinuosa (despite a radula resembling that in the S. boucheti group), S. stylaster (thought to be related to an Australian species), S. gradata (which has one of the most distinctive shell morphologies), and S. simplex (which seems to be intermediate between the S. boucheti group and the S. hians group). Serrata coriolis and S. exquisita possibly belong to a species group distributed on the Loyalty Ridge and off northern New Caledonia, and S. beatrix , S. granum and S. inflata are possibly western representatives of the S. translata group. Because of the similar pattern of their radular plates with unequal cusps (“saw blade” pattern), S. tuii , S. magna and S. robusta are provisionally proposed to belong in a more scattered S. tuii species group. None of these provisional species groups exhibit close similarities to the type species S. serrata , which seems to be an isolated form.

The geographical distribution of the Serrata species in the New Caledonian area is very asymmetrical. Only 4 species ( S. exquisita , S. beatrix , S. tuii , S. gradata ) are described from off the northwestern mainland (however, the single tiny shell of another Serrata species from this zone was broken during this study and is not described), one ( S. inflata ) is known from off the northeastern mainland, one ( S. orientalis ) from off the southeastern mainland and the Loyalty Ridge (although this latter population may, however, belong to a sibling species), and two ( S. lifouana , S. coriolis ) from the central Loyalty Ridge. All the other 27 species occur off southern New Caledonia, the majority being restricted north of 23°S and especially to the northern Norfolk Ridge, where the greatest diversity of species occurs, together with the highest rate of apparent micro-endemism. However, the micro-endemism exhibited by many species may be more apparent than real, considering that a few species occur at stations separated by noticeable distances in latitude (for S. ovata : 24’, for S. bathusi : 37’, for S. perlucida : 47’). Four species (12%) are known to range over more than 30’ of latitude (the widest distribution being that of S. robusta with 2°03’), and only 11 species (31%) range over more than 10’.

Although the intensity of sampling at bathyal depths was higher off southern New Caledonia and on the northern Norfolk Ridge than in other areas around New Caledonia, it seems clear that the genus Serrata is poorly represented off the far northern mainland, where the bathyal bottoms are often easy to sample and have been sampled during several expeditions. In contrast, off western and eastern New Caledonia the sharp relief and canyons are difficult to sample. The Loyalty Ridge remains poorly sampled and it is very probable that further expeditions in this zone will bring new Serrata species to light.

Thus, the data at hand allow the recognition of a pattern of increasing diversity of Serrata from northern to southern New Caledonia, the highest diversity occurring on the Norfolk Ridge itself. Further sampling south of 24°S (and overall south of 25°S) will verify whether this high specific diversity is specially focused on the northern Norfolk Ridge or extends further southwards.

DIVERSITY AND DISTRIBUTION IN OTHER PARTS OF THE PACIFIC AND IN THE INDIAN OCEANS

To the east of the Loyalty Ridge, the genus Serrata seems to be represented only by a small series of sibling species from shallow water in French Polynesia ( S. translata and 3 species recently described by Wakefield & McCleery 2002). No Serrata species were collected from bathyal depths during the IRD/MNHN expeditions to Tonga, Wallis, Fiji and Vanuatu, where marginellids are very poorly represented except for the genus Hydroginella , which is apparently able to disperse as a result of its behaviour of ectoparasitism on different species of sleeping fishes ( Boyer, Wakefield & McCleery 2003). Boyer (2002: 87) mistakenly mentioned Serrata instead of Hydroginella as the dominant genus in the bathyal of the Fijian archipelago.

The genus Serrata is not represented in collections from shallow water made recently in the southwest Pacific (Samoan Islands: E. Schwabe pers. comm.; Tongan and Fijian Islands: T. McCleery pers. comm.), or in collections from lower reef formations and bathyal depths made by French expeditions to the Chesterfield-Bellona Plateau (mainly MUSORSTOM 5, 1986 and CORAIL 2, 1988) and to the Solomon Islands (SALOMON 1, 2001). The genus also does not seem to be represented in the material from shallow water from India to South-East Asia, Indonesia and the Philippines currently deposited in public and private collections.

In contrast, the genus Serrata is well represented all around Australia by many species from shallow water, some ascribed to other genera and a good number undescribed (FB, unpublished). Serrata fasciata and probably the sympatric Marginella maoriana Powell, 1932 seem to be the only two representatives of Serrata off northern New Zealand.

Serrata is not recorded from bathyal depths off New Zealand and Australia. However, such depths remain poorly explored off Australia and a Serrata fauna may well occur there. A series of 4 undescribed Serrata species (among only 8 lots of marginellids: FB, unpublished) was collected off the Tanimbar and Kai Islands (far eastern Sunda Islands, northwest Arafura Sea) at bathyal depths by the French KARUBAR Expedition, 1991, making the occurrence of Serrata in the bathyal zone off northwestern Australia likely.

Together with the type species, S. serrata , which occurs in shallow water around the Mascarene Islands and ranges as far as the Glorieuses Islands (FB, unpublished), Marginella spryi Clover, 1974 from shallow water of northern Mozambique, Tanzania and southern Kenya, Serrataginella isabelae Bozzetti, 2005 from deep water off northern Mozambique and southern Tanzania (320-360 m) and Serrata brianoi Bozzetti, 1994 from deep water off Somalia (150-200 m) constitute a limited Serrata assemblage in the western Indian Ocean. However the bathyal levels are very poorly surveyed in the Indian Ocean, as in the Indonesian-Philippines archipelagos, so the occurrence of a diversified Serrata fauna is plausible at deep levels on the continental slopes of these areas.

Given the present state of marginellid taxonomy, none of the various species resembling Volvarina / Hydroginella occurring off South Africa can be transferred to Serrata , except for Marginella meta Thiele, 1925 , trawled off Cape Agulhas. This species has a reflected labrum with an excavated anterior area, and strong subequal denticles all along its length that are typical of Serrata . However several other species from deep levels in this area seem to intergrade between Serrata , Volvarina , Dentimargo and the Austroginella group.

Serrataginella spryi (Clover, 1974) , type species of the monospecific genus Serrataginella Coovert & Coovert, 1995 , is here transferred to Serrata , because of its similar shell morphology and radula pattern (see the discussion about the “ Serrata Group”, below). No Serrata species were collected in bathyal depths off Mayotte and the Glorieuses Islands, neither off Réunion, despite extensive sampling by the French expeditions BENTHEDI (1977) and Marion Dufresne Cruise/MD 32 (1982).

CONCLUSION

Although the bathyal fauna of the Indian Ocean and of the Australasian region remains very poorly investigated outside the New Caledonia Exclusive Economic Zone, the Recent geographical distribution of Serrata seems to consist of a “centre of diversity” or “diversity hot spot” ranging from the Australian platform to the north-south oriented oceanic ridges extending from New Caledonia to New Zealand, and of 2 peripheral zones with a few species in the western Indian Ocean and in the central South Pacific. Whether these patchy areas consist of pioneer lineages or relict species remains undetermined. The relative isolation of the New Caledonian occurrences and both the greater range of shell forms and the higher diversity displayed south of the New Caledonian mainland suggest that the principal dispersal events of Serrata in the southwest Pacific took place in subtropical latitudes between southeastern Australia, New Zealand and the Norfolk Ridge.

FOSSIL OCCURRENCES

EOCENE OCCURRENCES Several denticulate or non-denticulate species from the Middle Eocene of the Paris Basin and of western France, placed by Nieulande (1981) and Le Renard & Nieulande (1985) in Stazzania (species resembling Dentimargo ) or in Volvarina , seem to intergrade between Serrata and Recent forms of Volvarina , Dentimargo and several genera belonging to the Austroginella

group (mainly Protoginella Laseron, 1957 and Alaginella Laseron, 1957 ). These “intergrading” species possibly belong in the early history of the genus Serrata and of the whole Volvarina-Dentimargo complex. However, the question needs to be investigated on the basis of a more accurate study of the morphological series in the fossil material from various zones and periods, with the help of comparable characters found in Recent lineages. On the other hand, Euryentome anatina (Lea, 1833) from the Eocene of Alabama, the type species of Euryentome Cossmann, 1899 (see below), closely matches Recent species of Serrata , especially in the gently reflected upper part and clearly excavated lower part of the inner labrum, bearing numerous small subequal denticles all along its length ( Cossmann 1899: pl. 4, fig. 9; Palmer 1937: pl. 67, figs 4, 9; pl. 89, fig. 10). Also, the size and thickness of the columellar plaits increase from the anterior towards the posterior, as seen in several of the Recent Serrata species from New Caledonia. In many respects, E. anatina looks like a stout form of Serrata orientalis (Fig. 15) from southeastern New Caledonia. The most prominent differences between E. anatina and Recent Serrata species are the presence of a very short anterior columellar plait and of one to several horizontal lirations inserted high up on the parietal wall (columellar callus) in E. anatina . Although not common, an upper plait separated widely from the lower, closely spaced ones occurs in some Recent Serrata species as well, for instance in S. tuii (Figs 41- 43) and S. coriolis (Fig. 26). Three tiny species from the Eocene of France, with a keyhole-shaped anal canal and showing close affinities with the Austroginella group, were erroneously referred to Euryentome by Le Renard & Nieulande (1985: figs 37-39), on the basis of the incorrect concept of Euryentome used by Dockery (1980: pl. 40, fig. 4).

Euryentome Cossmann, 1899 is synonymized here with Serrata Jousseaume, 1875 . Palmer (1937: 416-418, pls 67, 89) introduced Marginella silabra as a replacement name for M. crassilabra Conrad, 1833 (non M. crassilabra Bory de St Vincent, 1827 ), type species of Euryentome . She designated a lectotype for both M. crassilabra Conrad, 1833 (as M. silabra ) and M. anatina Lea, 1833 as a specimen from the same lot that was used concurrently by Conrad and Lea in ANSP for the description of their respective taxa, and thereby made M. crassilabra Conrad, 1833 and M. anatina Lea, 1833 objective synonyms. Furthermore, Palmer (1937) listed M. anatina in her synonymy of M. crassilabra . Palmer (1937) did not discuss any morphological distinction between the two taxa, and based on her pictures, both species concepts are overlapping.

The replacement name introduced by Palmer was unnecessary, considering that the name M. anatina was available for this species. Palmer’s mistake derived apparently from the fact that she gave the original designation of the type species of Euryentome as “ M. crassilabra Conrad = M. silabra Palmer non M. crassilabra Lea ”, instead of “ non M. crassilabra Bory de Saint Vincent ” ( Palmer 1937: 416).

The few citations of the name E. silabra in the last 50 years do not justify its substitution, i.e. precedence inversion, for the older name M. anatina according to the Code (ICZN 1999: Article 23.9). Therefore, M. anatina Lea, 1833 , not E. silabra Palmer, 1937 , should remain as the valid name for the type species of Euryentome Cossmann, 1899 .

MIOCENE AND PLIOCENE OCCURRENCES

Marginella dongoensis Cox, 1927 from the Lower Miocene of Pemba Island ( Tanzania) appears to be very close to the few Recent species from the Western Indian Ocean attributable to Serrata . Cox (1927) considered M. dongoensis to be very close to M. nanggulanensis Martin, 1916 from the Javan Lower Miocene. The generic position of this latter species has not been determined.

Three slender fossil species from the Miocene of South Australia, illustrated by Cotton (1949), Marginella propinqua Tate, 1878 , M. winteri Tate, 1878 and M. woodsi Tate, 1878 , appear to be very similar to the type species of Serrata , S. serrata , and to some Recent Serrata species from New Caledonia. Marginella winteri was designated by Laseron (1957) as the type species of his genus Exiginella and M. propinqua was considered conspecific. Laseron (1957) noted that Cossmann (1899) placed M. winteri in the genus Serrata , but did not provide arguments for the new generic position. Exiginella was considered by Coovert & Coovert (1995) to be a junior synonym of Serrata .

Despite their very short spire and their long conoidal outline, Marginella inermis Tate, 1878 from the Miocene of South Australia, type species of the genus Conuginella Laseron, 1957 (see the discussion about the “ Serrata Group”, below), and Marginella strombiformis Tenison Woods, 1877 , from the Early Miocene of Fossil Bluff, Tasmania, also seem to belong to the genus Serrata on the basis of their subequal labial denticles, as also do the more inflated species Marginella wentworthi Tenison Woods, 1877 , also from the Early Miocene of Fossil Bluff, Tasmania, and M. kalimnae Chapman & Crespin, 1933 from the Lower Pliocene of eastern Victoria. Conuginella Laseron, 1957 is considered here as a junior synonym of Serrata . Numerous other fossil marginellid species illustrated by Cotton (1949) are compatible with the range of morphological variation displayed by Recent Serrata species off New Caledonia. Several other fossil species from the Australian Tertiary seem to intergrade between Serrata , Volvarina , Dentimargo , Hydroginella and the Austroginella group.

As there is no Cenozoic fossil record in New Caledonia, Cenozoic Serrata species remain unknown there. However, the Cenozoic fossil record of New Zealand, which lies on the same oceanic ridge, is well known and includes several Serrata species. Besides fossils of “ S. mustelina ” [which Marshall (2004) has shown is S. fasciata ] and S. maoriana, Beu & Maxwell (1990) listed 4 further fossil species from New Zealand as belonging to Serrata or Haloginella , and illustrated S. kirki ( Beu & Maxwell, 1990: pl. 42j) from the Late Pliocene of the southeastern North Island. This species appears to match S. amphora closely in general outline, S. tuii in its fifth columellar plait, and S. serrata in its labial denticles overlapping the ventral side of the outer lip.

CONCLUSION

Despite its scanty occurrence as distinct forms in the Middle Eocene, it seems that the genus Serrata was in the process of differentiating from the Volvarina / Dentimargo morphological group during this period, that is, precisely when the marginelliform gastropods bloomed. Serrata has become more clearly differentiated since the Miocene, and restricted to the Indo-Pacific Province, with the highest diversity apparently maintained until Recent time in the Australian/New Caledonian area. The huge morphological range of shell types represented by Serrata in New Caledonia and the display in this Recent fauna of most of the morphological combinations represented in the Cenozoic fauna both suggest that Serrata arrived there very early, and that a representative series of the Serrata radiation has been conserved there.

MORPHOLOGICAL DISSIMILARITY AND GENERIC UNITY

RADULA PATTERNS

The radulae of S. serrata (Fig. 53) and of S. mustelina (Fig. 54), respectively type species of Serrata and Haloginella , are here compared with the radulae of S. fasciata (Fig. 55) and S. translata (Fig. 56), previously documented by Coovert & Coovert (1995).

These radulae are united morphologically by the comb-like pattern of plates bearing a large number of subequal or unequal “saw blade” cusps (a few larger cusps alternating with a greater number of smaller cusps). This small series of radulae also displays significant disparity in the outline of the cusped border of the plates (from subhorizontal to arched or faintly depressed centrally). The high number of cusps on the comb-like plate of a uniserial modified rachiglossan radula, as found in Serrata / Haloginella , must be compared with the numbers prevailing in the marginellid subfamily Marginelloninae , suggested by Harasewych & Kantor (1991) to be the most primitive group in the Marginellidae : Marginellona gigas (Martens, 1904) has from 58-59 to 83-85 radular cusps ( Harasewych & Kantor 1991) and Afrivoluta pringlei (Tomlin, 1947) has 70-80 cusps ( Coovert & Coovert 1995). The high number of radular cusps is deduced to be a plesiomorphic character ( Boyer 2001). However, the number of radular cusps in S. serrata and S. translata is lower than the number in several species of the Volvarina-Prunum group (cf. Coovert & Coovert 1990), and similar patterns of unequal or subequal cusps are found in this group. Therefore, the radular patterns found in the few Serrata species so far studied can not be distinguished from those occurring in the Volvarina-Prunum group.

Overall radular disparity is relatively limited between the Serrata species from New Caledonia, despite the variation observed in details such as the width of the plates (53-208 Μm), the number of cusps (31-48), and the outline of the cusp series (straight to barbed). In most cases the number of cusps ranges from 31 to 38-42, as is seen in S. serrata and S. translata . Only S. perlucida , occurring in very deep waters, has radular plates with 48 cusps, as in S. mustelina and S. fasciata (49-59 subequal cusps along a straight external border in the latter species). Two principal radular patterns of comb-like plates (a “regular” or “straight” pattern of subequal cusps and an irregular “saw blade” or “barbed” pattern of unequal cusps) seem to occur in New Caledonia, not clearly correlated with the shell morphology of the species groups in which the radula is documented.

In the S. boucheti group the radulae of S. boucheti (Fig. 63) and of S. veneria (Fig. 58) are known, both of which have a straight anterior plate edge bearing subequal cusps. Despite the differences in its shell morphology, S. sinuosa (Fig. 59) has the same kind of radula as these two species. In the S. hians group only the radula of S. perlucida (Fig. 64) is known. It has a centrally depressed anterior plate edge bearing a high number of cusps. The shape of this radula is similar to that found in the Australian S. mustelina (Fig. 54), the New Zealand S. fasciata (Fig. 55) and the Polynesian S. translata (Fig. 56). The radula of the New Caledonian species attributed to the S. translata group are unknown, as is the case for the 2 species in the S. coriolis group. The radulae are not known in S. arcuata , S. stylaster , S. gradata and S. simplex , all of which are considered provisionally as not part of a species group. The series constituted by S. tuii (Fig. 57), S. magna (Figs 61, 62) and S. robusta (Fig. 60) is united on the basis of a similar radula with “saw blade” or “barbed” plates, comparable to that in the Indian Ocean type species S. serrata (Fig. 53). However, because of the heterogenous shell morphologies exhibited in this series of species, it cannot be considered at present as a “species group” in the sense adopted here, and its taxonomic status deserves further investigation.

The study of the radula in 7 further Serrata species from New Caledonia ( S. tuii , S. veneria , S. sinuosa , S. robusta , S. magna , S. boucheti , S. perlucida ) does not add new evidence of a distinctive radular pattern distinguishing Serrata . Despite a tendancy in these species for a higher average number of radular cusps, the radular patterns essentially still cannot be distinguished from that in the Volvarina-Prunum group. Furthermore, the comb-like radula with many cusps cannot be used as diagnostic by itself of a “ Serrata clade”, because it is assumed to be a plesiomorphic character state, which may have persisted or not within various parallel lineages, depending on reduction of animal size, change of feeding habits, etc. So in the same way that the comb-like radula has been lost in some, or most, of the Dentimargo lineages and conserved at least in one ( Boyer 2005), the comb-like radula with many cusps seems to have been conserved in some lineages or radiations of the Volvarina-Serrata complex and to have been replaced by radulae with less numerous cusps in others.

FIGS 53-64. Radulae. 53, Serrata serrata (Gaskoin, 1849) , La Réunion, FBC, H. 8.00 mm, scale bars 10 Μm. 54, Serrata mustelina (Angas, 1871) , Sydney Harbour, South East Australia,AMS, H 6.00 mm, scale bars 5 Μm. 55, Serrata fasciata (Sowerby, 1846) , Whangarus, North Island, New Zealand, H unknown (Coovert 1987), scale bars 15 Μm. 56, Serrata translata (Redfield, 1870) , Raiatea, French Polynesia, H 4.30 mm ( Coovert & Coovert 1995), scale bars 10 Μm. 57, Serrata tuii ( Cossignani, 2001) , northern New Caledonia, BATHUS 4 stn DW 929, H 7.20 mm, scale bars 20 Μm. 58, Serrata veneria n. sp., paratype MNHN 20596 BIOCAL stn DW 46, H 4.20 mm, scale bars 10 Μm. 59, Serrata sinuosa n. sp., southern New Caledonia,BIOCAL stn DW 51, H 6.30 mm, scale bars 20 Μm. 60, Serrata robusta n. sp., southern New Caledonia, SMIB 8 stn DW 193-196, H unknown, scale bars 50 Μm. 61, 62, Serrata magna n. sp.; 61, southern New Caledonia, BIOCAL stn DW 51, H 7.10 mm, scale bars 50 Μm; 62, BIOCAL stn DW 70, H 8.60 mm, scale bars 50 Μm. 63, Serrata boucheti ( Boyer, 2001) , SMIB 8 stn DW 167, H unknown ( Boyer 2001), scale bars 5 Μm. 64, Serrata perlucida n. sp., Norfolk Ridge, BIOCAL stn CP 61, H 7.20 mm, scale bars 50 Μm.

DIVERSITY, MORPHOLOGICAL DISPARITY AND GENERIC DIAGNOSIS

Since only 9 Recent species had been placed in the genus Serrata previously, the 35 Serrata species now recognized from New Caledonia increase five-fold the Recent species diversity of the genus. Correspondingly, the range in shell morphology of Serrata observed in New Caledonia is considerably broader than that exhibited in the genus previously. A redefinition of the diagnosis of Serrata is therefore required.

On the grounds of the combined study of the morphological disparity of the shells and radulae of the species here attributed to Serrata , the species belonging to this genus might be defined restrictively as Volvarina -shaped species from the Indo-Pacific with a medium to high number (> 28) of subequal to slightly more variable radular cusps, combined with 4 to 5 well-developed columellar plaits occupying the whole anterior third of the columellar border. The labrum may be reflected or not, excavated anteriorly or not, and provided or not with a series of numerous subequal denticles. The labial denticles, considered initially as the main diagnostic feature of the genus (Latin “ serrata = “saw”), are in fact noticeably variable within several species (from a distinctly denticulated labrum to a smooth labrum) and all kinds of intermediate species morphologies are represented.

The set of diagnostic features is, however, not satisfactory. The disparity of morphological shell types is so wide in Serrata that several species attributed to this genus can hardly be distinguished objectively from Volvarina (or even from Dentimargo , Hydroginella or the Austroginella series).

Even if not perfectly homogeneous in the Recent, we may consider the whole Volvarina-Dentimargo-Serrata series as a more inclusive clade, in which each of the 3 genera is not coherent on the basis of the data available. Both Volvarina mitrella (Risso, 1826) , type species of Volvarina Hinds, 1844 , and Serrata serrata (Gaskoin, 1849) , type species of Serrata Jousseaume, 1875 , have the same average number of 30 cusps on their radula. The cusps in V. mitrella are much more unequal than in S. serrata , but overall both species show a deeply different shell morphology (rather biconical with smooth inner labrum in V. mitrella , versus subcylindrical with many strong subequal labial denticles in S. serrata ) and occur in distinct biogeographic regions ( V. mitrella in the Mediterranean and S. serrata in the southwest Indian Ocean), at least in the Recent era. The well-defined V. mitrella species group (= Volvarina sensu stricto?) is profoundly different from the S. serrata species group (= S. serrata alone) as far as shell morphology is concerned. Besides its nominate species, the V. mitrella species group ( Boyer 2000) includes many Recent Atlantic species ranging from off western Morocco to off eastern Brasil, for instance V. exilis (Gmelin, 1791) , V. cf. cernita (Locard, 1897) , V. serrei (Bavay, 1913) , V. roberti (Bavay, 1917) and V. tunicata Boyer, 2000 , as well as V. eumorpha (Melvill, 1906) , which is restricted to the Gulf of Oman and is assumed to be a relict species descending from an old lineage isolated from the rest of the species group by the Tethys closure.

If we adopt a restrictive point of view based on the type species, Serrata is clearly distinct from Volvarina both in its shell morphology and its main geographic distribution, suggesting a phylogenetic gap between the two species groups. But if we adopt a more inclusive point of view, these species groups can be linked by various forms, in the fossil as well as in the Recent fauna, and may be united within a wider Volvarina-Dentimargo-Serrata clade. The issue of the real phylogenetic status and the correct taxonomic placement of this clade will have to be investigated more thoroughly, on the basis of further observations of variation in fossil and Recent faunas. In particular, DNA studies may allow determination of whether Serrata is the stem lineage from which Volvarina and Dentimargo originated, a sister lineage of Volvarina and Dentimargo , or if it cannot be separated from the Volvarina / Dentimargo series.

The biogeographic criterion seems to work correctly for the provisional discrimination of a less inclusive “ Serrata clade” in the Indo-Pacific. This is illustrated well in the New Caledonia area, where the species attributed to Serrata exhibit a wide range of shell morphology (with or without subequal labial denticles, with or without excavated lower labrum), forming a series with intermediates exhibiting the full range of possible variation. The unity of this “ Serrata clade” in New Caledonia derives mostly from this continuity or affinity of shell morphology and from their biogeographic coherence, but it is also characterized by the occurrence of a comb-like radula with many cusps (in the species whose radulae have been examined so far). This situation could be interpreted as a local radiation, i.e. a possible “ Serrata clade” resulting from long-term isolation.

Besides this local Serrata radiation, only one Volvarina species is known from the New Caledonia region, namely V. armonica Cossignani, 1997 , which occurs in 100-300 m and belongs to the same species group of slender shelled species from shallow water as V. philippinarum (Redfield, 1848) from the Philippines and V. rex ( Laseron, 1957) from northwestern Australia. Volvarina and Serrata appear also as well-differentiated groups in the western Indian Ocean where both genera can be syntopic (for instance in Mauritius: FB, unpublished), in the northern Indian Ocean and the Indonesia-Philippines region, where Volvarina is restricted to reef levels and where Serrata , when occuring, seems to be restricted to bathyal levels. No local intergrade in shell pattern (smooth lip versus labial denticles) nor in radular pattern (few cusps versus numerous cusps) is known from these areas. The most widespread Volvarina species group, the V. micans (Petit, 1851) species group, has an average of only 17-22 radular cusps (Coovert & Coovert 1990; FB, unpublished) as opposed to more numerous cusps in the Serrata species.

Two genera allied to Volvarina display a distinctive shell pattern in the Indo-West Pacific: Closia from the western Indian Ocean characterized by oval shells and Cryptospira from the Arabian Sea to the Philippines characterized by strong columellar plaits with a square section. Closia and Cryptospira (Coovert & Coovert 1990; FB, unpublished) cannot be distinguished from the local species referred to Volvarina sensu stricto on the grounds of their radulae with a low to moderate number of cusps, but their genuine and homogeneous shell features justify their generic status, reflecting their phylogenetic distinction from the rest of the poorly differentiated Volvarina / Prunum series.

The situation seems to be different off Southern Africa and Australia, where possible intergrades between Serrata and the Volvarina / Dentimargo / Austroginella series may occur.

In the Caribbean and Panamic provinces, the radulae of various species attributable to the Volvarina / Prunum series range from few cusped to many cusped (Coovert & Coovert 1990) but none of these species displays shell features commonly exhibited in the Indo-Pacific “ Serrata ” species group (subequal labial denticles and/or excavated lower labrum). The “ Volvarina ” fauna from Caribbean occurs mainly at the infralittoral level, very few Volvarina -like species being found at bathyal depths.

This situation can be interpreted as the persistence in the Indo-Pacific province of the descendents of a poorly differentiated Eocene stock belonging to the Volvarina / Dentimargo / Serrata complex, with the present “diversity hot spot” occurring in the New Caledonia area. According to this view, the clade would have become scarce outside the New Caledonia / Australia area and disappeared from the other biogeographic provinces. Future research will allow verification of whether the Indo-Pacific “ Serrata clade” deserves to be considered as a generic taxon or dismissed as a polyphyletic assemblage.

SUPRAGENERIC RELATIONSHIPS

CRITICISM OF THE MARGINELLINAE TRIBES

The division of the subfamily Marginellinae Fleming, 1828 by Coovert & Coovert (1995) into 3 tribes ( Austroginellini Coovert & Coovert, 1995, Prunini Coovert & Coovert, 1995 and Marginellini Fleming, 1828) seems to be poorly supported, principally because it does not take account of the close relationships seen in fossil faunas as well as Recent ones between the genera Serrata , Volvarina and Dentimargo , each attributed by Coovert & Coovert (1995) to a different tribe. In the Recent species, the radular patterns are not so clearly separated, as several kinds of intergrades exist, for instance between the comb-like radula with subequal cusps and assumed derived patterns, for example the narrow Austroginella radula with a chevron-shaped external border and a large central cusp. As demonstrated above regarding the variability of shell morphology and radular pattern in Serrata , separation from Volvarina is based on slight differences and both genera must be placed in the same tribe.

Furthermore, the apparent absence of a radula in Dentimargo has been shown not to be a diagnostic character, since the discovery ( Boyer 2005) of a comb-like radula with few cusps in a Recent Dentimargo species from the northern Caribbean that matches the type species of the genus in its principal shell features, including a sharp upper labial denticle. Very few of the living Dentimargo species have been checked for the presence of a radula, but this simple

discovery shows that the radula status in Dentimargo is eclectic and that Dentimargo (if constituting a homogeneous clade) may be derived from a radulate genus like Volvarina (or the Volvarina / Serrata complex). The loss of the radula in a uniserial modified rachiglossan group probably arises easily, and according to the proposition made by Boyer (2001), some lineages in Dentimargo may have conserved a radula while others may have lost it.

CRITICISM OF THE “ SERRATA GROUP”

Coovert & Coovert (1995) recognized in their tribe Austroginellini a “ Serrata Group” composed of 5 genera ( Serrata , Conuginella , Serrataginella , Stromboginella and Hydroginella ), as well as an “ Austroginella Group” composed of species assumed to have a narrow radula with plates showing a chevron-shaped external border. I revise this “ Serrata Group” in the following way.

Considering the shell morphology of its type species ( Marginella inermis Tate, 1878 from the Miocene of southern Australia), which is compatible with the range of morphological variation exhibited in the genus Serrata in New Caledonia, the fossil genus Conuginella Laseron, 1957 is regarded as a junior synonym of Serrata Jousseaume, 1875 .

S errataginella Coovert & Coovert, 1995 (type species: “ Marginella ” spryi Clover, 1974 from off East Africa, Recent) is also regarded here as a junior synonym of Serrata Jousseaume, 1875 , considering the subequal labial denticles of its type species, and its radular plates with 70 unequal cusps, arranged in a regular series of 9 to 11 large, pointed, strong cusps flanked on each side by 4 to 7 finer cusps increasingly set back, making a succession of subequal fringed chevrons ( Coovert & Coovert 1995: fig. 29). The morphology of this radular plate resembles the irregular “saw blade” or “barbed” pattern found in the type species S. serrata and in the 3 species grouped in the S. tuii series. The shell morphology in S. spryi is rather similar to that of S. tuii and S. robusta . From the point of view of its radular morphology (a very high number of cusps associated with a “saw blade” pattern), S. spryi appears to be intermediate between Serrata and the Marginellona / Afrivoluta group. Furthermore, the data at hand strongly suggest that the radula of M. spryi belongs to an ancestral state in Serrata . Whatever might be the historical position of M. spryi within the evolution of Serrata , there is no particular reason to classify it separately from this genus.

Because of its very narrow aperture and tiny columellar plaits that are very closely packed on the anterior part of the columellar border, the fossil genus Strombiginella Laseron, 1957 (type species: Marginella crassidens Chapman & Crespin, 1928 , Pleistocene of Australia) is considered to be a junior synonym of Hydroginella Laseron, 1957 . The shell outline exhibited by M. crassidens is perfectly compatible with the wide disparity of shell types observed in Recent Hydroginella species (cf. Boyer et al. 2003).

At present, Serrata and Hydroginella cannot be considered as deserving special association as “sister genera” in a common “ Serrata Group”. The distinctive shell characters of Hydroginella , characterized in most species by a very narrow aperture, a very light shell, and very small, closely packed, anteriorly situated columellar plaits, are not more similar to those of Serrata than to those of Volvarina . The occurrence of subequal labial denticles in some Hydroginella species can be considered a character shared with Serrata , but it possibly represents the homologous conservation of an ancestral character instead of a derived character. The radula of Hydroginella , with very narrow, square plates with resorbed cusps, associated with its behaviour of ectoparasitism on sleeping fishes (Bouchet 1989), is noticeably divergent from the comb-like patterns in Serrata and may be derived from a few-cusped Volvarina / Prunum radula pattern.

The placement of Serrata and Hydroginella in the tribe Austroginellini is not supported by concrete evidence, because most of the genera placed in the “ Austroginella Group” have very distinctive chevron-shaped narrow radular plates, very distinct from the comb-like patterns found in Serrata and the square patterns found in Hydroginella . Serrata and Hydroginella are not demonstrably more closely related to each other than either is to Volvarina / Prunum or to the Austroginella and Dentimargo groups, and the “ Serrata Group” sensu Coovert & Coovert (1995), composed of Serrata and Hydroginella , is therefore not accepted.

CONCLUSION

The tribes recognized by Coovert & Coovert (1995) in the subfamily Marginellinae Fleming, 1828 ( Austroginellini , Prunini and Marginellini) appear to have an eclectic composition. They do not reflect the real affinities among the Recent species included in the genera Volvarina , Dentimargo , Austroginella , Serrata and Hydroginella , in which some species seem to be weakly differentiated, or not differentiated at all, from a common stem. The use of the notion of a “ Volvarina / Dentimargo complex” at the suprageneric level is proposed as being more appropriate in the present state of knowledge.

The great disparity in the specific shell morphologies of Serrata , the frequent combination of features considered as typical in Volvarina and Dentimargo in the Recent, the occurrence of many morphological intergrades between these genera since the Mid-Eocene of the western Tethys Sea, and the widespread occurrence of the plesiomorphic character of a radula with numerous cusps, together suggest that the genus Serrata may be situated near the stem from which most of the Recent groups of the Volvarina - Dentimargo complex have differentiated.

ACKNOWLEDGEMENTS

I am very grateful to the following people for their kind help in support of this study: Philippe Bouchet (MNHN), Kathie Way (BMNH), Ian Loch and Holly Barlow (AMS) for the loan of material, Jacques Le Renard and Pierre Lozouet (MNHN) for help with fossil references, Andrew Wakefield (Buckhurst Hill, UK) and Bruce Marshall (NMNZ) for providing literature, Emilio Rolán (Vigo, Spain) for extraction and illustration of radulae, Jesús Mendez (CACTI, Vigo University, Spain) for access to SEM equipment, Philippe Maestrati (MNHN) for taking most of the photos and arranging the digital plates, Alan Beu ( New Zealand Institute of Geological & Nuclear Sciences, Lower Hutt) for correcting the English and for comments on the initial manuscript, Robert and Nicole Hasselot (Jouques, France) for typing it, and Virginie Héros (MNHN) for editorial supervision. I am also much indebted to Serge Gofas (Málaga University, Spain), who corrected the final manuscript and contributed to improving it.

REFERENCES

BEU A. G. & MAXWELL P. A. 1990. — Cenozoic Mollusca of New Zealand. COOVERT G. A. 1987. — Proper usage of the generic names Haloginella, New Zealand Geologic Survey, Paleontological Bulletin 58: 1-518. Hyalina, and Volvarina with a diagnosis of the genus Haloginella. Marginella BOUCHET P. 1989. — A marginellid gastropod parasitizes on sleeping fishes. Marginalia 3 (1): 1-7. Bulletin of Marine Science 45: 76-84. COOVERT G. A. 1999. — Revision of the nomenclature of the Kaicher Card BOUCHET P., HÉROS V., LOZOUET P. & MAESTRATI P. 2008. — A quater- Packs (Cystiscidae, Marginellidae). Marginella Marginalia 10 (3-6): 1-39. century of deep-sea malacological exploration in the South and West COOVERT G. A. & COOVERT H. K. 1990. — A study of marginellid radulae. Pacific:Where do we stand? How far to go?, in HÉROS V., COWIE R. H. & Part I: Type 6 radula, “ Prunum / Volvarina ” type. Marginella Marginalia

8 (1-6) & 9 (1-6): 1-68.

BOUCHET P. (eds), Tropical Deep-Sea Benthos 25. Mémoires du Muséum COOVERT G. A. & COOVERT H. K. 1995. — Revision of the supraspecific

national d’Histoire naturelle 196: 9-40. classification of marginelliform gastropods. The Nautilus 109 (2-3): 43-

BOYER F. 2000. — The genus Volvarina (Volutacea: Marginellidae) in Brazil. 110.

Part 1: revision of the species described by A. Bavay, and closely related COSSIGNANI T. 2001. — Descrizione di sei nuove marginelle (Gastropoda:

species. Novapex 1 (2): 35-55. Prosobranchia, Marginellidae e Cystiscidae ) della Nuova Caledonia.

BOYER F. 2001. — Espèces nouvelles de Marginellidae du niveau bathyal de Malacologia Mostra Mondiale (Cupra Marittima, Italy) 35: 12-17.

la Nouvelle-Calédonie. Novapex 2 (4): 157-169. COSSMANN M. 1899. — Essais de Paléoconchologie comparée 3. Paris

BOYER F. 2002. — Description of five new marginellids from bathyal levels of (Cossmann): 1 -201, pls 1-8.

southern New Caledonia. Novapex 3 (2- 3): 87-96. COTTON B.C. 1949.— Australian Recent and Tertiary Mollusca,Family Marginellidae.

BOYER F. 2005. — The discovery of a radula in a Dentimargo species and its Records of the South Australian Museum 9 (2): 197-224,pls 17-20.

taxonomic implications. Iberus 23 (2): 47-52. COX L. R. 1927. — Neogene and Quaternary Mollusca from the Zanzibar

BOYER F. & ROLAN E. 2004. — About a series of cylindrical shelled Granulina Protectorate. Report on the Palaeontology of the Zanzibar Protectorate.

( Marginellidae ) from north east Atlantic waters and the taxonomic Government of Zanzibar, pp. 13-102, pls 3-19.

organisation of the Granulininae. Iberus 22 (1): 155-165. DOCKERY D. T. III 1980. — The invertebrate macropaleontology of the

BOYER F., WAKEFIELD A. & McCLEERY T. 2003. — The genus Hydroginella Clarke County, Mississippi, Area. Mississippi Department of Natural

(Caenogastropoda: Marginellidae ) at bathyal levels from the Fiji Islands. Resources, Bureau of Geology, Jackson, Mississippi, Bulletin 122:

Novapex 4 (2-3): 67-77. 9-387, pls 1-82.

GOFAS S. 1989.— Le genre Volvarina (Marginellidae) dans la Méditerranée et MARSHALL B. A. 2004. — New names for four common Marginellidae l’Atlantique du nord-est. Bollettino Malacologico 25 (5- 8): 159-182. (Mollusca, Gastropoda) from northern New Zealand. Molluscan Research

HARASEWYCH M. G. & KANTOR Y. 1991. — Rediscovery of Marginellona 24 (1): 7-20.

gigas (Martens,1904),with notes on the anatomy and systematic position NIEULANDE F.A.D. van 1981.— Descriptions of Eocene Marginellidae (Mollusca, of the subfamily Marginelloninae (Gastropoda: Marginellidae). Nemouria Gastropoda) from the Paris Basin. Mededelingen Werkgroep voor Tertiare 37: 1-19. en Kwartaire Geologie 18 (2): 39-83, 4 figs, 6 pls.

JOUSSEAUME F.P.1875.— Coquilles de la famille des marginelles.Monographie. PALMER K. V. W. 1937. — The Clairbornian Scaphopoda, Gastropoda and Revue et Magasin de Zoologie 3 (3):164- 271, 429-435,pls 7-8 [also issued dibranchiate Cephalopoda of the southern United States. Bulletins of separately, pp. 1-115, pls 7-8]. American Paleontology 7 (32): part I, 1-548; part II, 551-730, pl. 1-90.

LASERON C. F. 1957. — A new Classification of the Australian Marginellidae PONDER W. F. 1970. — Some aspects of the morphology of four species of (Mollusca),with a review of species from the Solanderian and Dampierian the Neogastropod family Marginellidae with a discussion on the evolution zoogeographical provinces. Australian Journal of Marine and Freshwater of the toxoglossan poison gland. Journal of the Malacological Society of Research 8 (3): 274-311. Australia 2 (1): 55-81.

LE RENARD J. & NIEULANDE F.A.D.VAN 1985.— Les Marginellidae (Gastropoda, WAKEFIELD A. & McCLEERY T. 2002. — A review of the genus Serrata Volutacea) de l’Eocène moyen du Bassin de Nantes et du Cotentin (France). Jousseaume, 1875 (Gastropoda: Marginellidae) from French Polynesia. Mededelingen Werkgroep voor Tertiare en Kwartaire Geologie 22 (1): 3-69. Novapex 3 (4): 119-126.