Hidden in plain sight: two co-occurring cryptic species of Supplanaxis in the Caribbean (Cerithioidea, Planaxidae)

Abstract The cerithioid Supplanaxisnucleus (Bruguière, 1789) is widespread in the Caribbean, where it lives in often dense aggregates on hard surfaces in the middle-high intertidal. Molecular evidence shows that it comprises two species that are in fact morphologically diagnosable. We fix the nomenclature of Supplanaxisnucleus by designating a sequenced neotype from Bruguière’s historical locality of Barbados, and identify the second, cryptic species as S.nancyae (Petuch, 2013). The two live syntopically across the Caribbean and form a closely related species group with the Panamic S.planicostatus (G.B. Sowerby I, 1825). Planaxisnucleola Mörch, 1876, described from St Croix, in the Virgin Islands, never again recorded in the literature but listed as a synonym of S.nucleus in taxonomic authority lists, is recognized as a valid species of Hinea Gray, 1847. Proplanaxis Thiele, 1929 and Supplanaxis Thiele, 1929, are synonyms and the latter is given precedence over the former.


Introduction
Supplanaxis nucleus (Bruguière, 1789) (Cerithioidea, Planaxidae) is a small, gregarious gastropod that lives in moderate to large population densities in intertidal, well-oxygenated habitats throughout the Caribbean. It occurs on hard substrates, from small pebbles and cobbles to large boulders or massive bedrock (Vermeij 1973;Bandel 1976;Houbrick 1987), from which it presumably grazes the biofilm. Individuals remain concealed from the sun under spray-moistened rocks, but emerge to feed, migrating with the tide (Bandel 1976). Rather little has been published on its biology, despite its ubiquity in the Caribbean. Similar to other planaxids, the species broods its young in a subhemocoelic brood pouch (Simone 2001;Strong et al. 2011), which are released from a pore on the side of the neck at the veliger stage (Thorson 1940;Bandel 1976;Houbrick 1987). Bandel (1976) observed freshly collected females to release veligers throughout the year. Troschel (1858) and Bandel (1984) provided descriptions of the radula of the species, and Houbrick (1987) and Simone (2001) described the radula and anatomy.
Following the serendipitous discovery that syntopic specimens from Guadeloupe clustered in two molecular groups, freshly collected material from Curaçao and Barbados confirmed the existence of two molecular clades within Supplanaxis nucleus. Museum material was then examined to evaluate the global distribution of the two clades.
In the present paper, we re-assess the taxonomy of Supplanaxis nucleus; we stabilize its nomenclature through the fixation of a neotype and review the nominal planaxid species currently treated as synonyms; and we attribute the second, molecularly distinct taxon to the little-known S. nancyae (Petuch, 2013). Finally, we remove Planaxis nucleola Mörch, 1876 from the synonymy of S. nucleus, and revalidate it as a species of Hinea Gray, 1847.

Materials and methods
Specimens for molecular and morphological study were collected intertidally from three sites in the Lesser Antilles (Guadeloupe, Curaçao, Barbados); tissues were separated from the shells following flash-boiling or microwaving (Fukuda et al. 2008;Galindo et al. 2014) and preserved in 95% EtOH.
Radulae were tissue digested overnight in 100 µl of ATL lysis buffer (Qiagen, Inc.) containing ~ 50 µg of Proteinase-K, sonicated and rinsed in de-ionized water (Holznagel 1997). Cleaned radulae were mounted on aluminum stubs using carbon adhesive tabs, then coated with 25-30 nm gold/palladium (60/40) and photographed using an Apreo scanning electron microscope (FEI Company) at the National Museum of Natural History. Shells were photographed using a Canon EOS 50D camera with a Canon EF 100 mm f/2.8 macro lens and Canon MT-24EX macro twin light flash (Canon USA, Inc.).
Whole genomic DNA was extracted from a ~ 1 mm 3 tissue clip of the foot using an Autogenprep965 (Autogen, Holliston, MA) automated phenol:chloroform extraction with a final elution volume of 50 µL. A 691 base pair (bp) fragment of cytochrome c oxi-dase subunit I (COI) was amplified using the jgLCOI primer (Geller et al. 2013) in combination with Cerithioid_COIR (Strong and Whelan 2019); a 509-511 bp fragment of 16S ribosomal DNA was amplified with the universal 16SAR/BR primers (Palumbi et al. 1991). PCR reactions were performed with 1 µL of undiluted DNA template in 20 µL reactions. Reaction volumes for COI consisted of 10 µL of Promega Go-Taq Hotstart Master Mix, 0.15 µM each primer, 0.25 µg/µL BSA, 1.25% DMSO and an amplification regime of an initial denaturation at 95 °C for 7 min, followed by 45 cycles of denaturation at 95 °C for 45 s, annealing at 42 °C for 45 s, extension at 72 °C for 1 min and a final extension at 72 °C for 3 min. Reaction volumes for 16S were 1x Biolase (Bioline, Taunton, MA) reaction buffer, 500 µM dNTPs, 3 mM MgCl 2 , 0.15 µM each primer, 0.25 µg/µL BSA, 1 unit Biolase DNA polymerase and an amplification regime of initial denaturation at 95 °C for 5 min, followed by 35 cycles of denaturation at 95 °C for 30 s, annealing at 48 °C for 30 s and extension at 72 °C for 45 s, followed by a final extension at 72 °C for 5 min. PCR products were purified using the Exo-SAP-IT protocol (GE healthcare, Piscataway, NJ). BigDye 3.1 (ABI, Foster City, CA) sequencing reactions and sequencing on an ABI 3730XL DNA analyzer capillary array were done following manufacturer's instructions.
Genes were sequenced in both directions to ensure accuracy. Chromatograms were visually inspected and corrected as necessary in Geneious Prime 2019 (Biomatters). COI alignments were translated into amino acids to check for stop codons and frameshift mutations, then trimmed to 658 bp representing the standard invertebrate barcoding region (Folmer et al. 1994). 16S sequences were aligned with MUSCLE (Edgar 2004) using default parameters as implemented in Geneious Prime. The final aligned length for the 16S dataset was 512 bp. All newly generated sequences have been deposited in GenBank (Table 1).
Sequences of additional planaxids [Supplanaxis niger (Quoy and Gaimard, 1833), Planaxis planicostatus G.B. Sowerby I, 1825] were downloaded from GenBank, including Planaxis sulcatus (Born, 1778), which was used to root the tree. Phylogenetic reconstruction was conducted using Maximum Likelihood (ML) in IQ-TREE ver. 1.6.12 (Nguyen et al. 2015) as implemented on the IQ-TREE web server (Trifinopoulos et al. 2016). The best-fit partitioning scheme and the most appropriate substitution model for each partition were estimated using ModelFinder (Kalyaanamoorthy et al. 2017) and partition models (Chernomor et al. 2016). Nodal support was estimated with 1,000 ultra-fast bootstrap replicates (Hoang et al. 2018).

Results
Thirty-two individuals from three sites in Guadeloupe, Curaçao and Barbados were sequenced for portions of the COI and 16S mitochondrial genes ( Table 1). The concatenated dataset was 1170 bp in length. The best-fit partitioning scheme used distinct models for each locus, with the best-fit model being MGK+F3X4 and HKY+F+G4 for COI and 16S, respectively. The ML tree constructed from the concatenated dataset ( Fig. 1) resolved two strongly supported clades among what has been traditionally recognized as a single species, Supplanaxis nucleus. Individuals collected in syntopy from all three sites occurred in both clades, which differed by 11.6-12.2% uncorrected pairwise sequence divergence in COI, with no geographic structuring among sites. Examination of the shells and radula revealed diagnostic morphological features consistent with the recognition of two species.  Remarks. Thiele (1929: 203) established Supplanaxis as a subgenus of Planaxis, for planaxids with rather small, dark colored, largely smooth shells, with spiral grooves mostly on the base of the last whorl and under the suture; and with a radula characterized by a broad rachidian with two to four denticles on either side, the long lateral extensions of the lateral teeth, and the outer marginal with a very broad, finely toothed cutting edge. Based on the morphological differences in shell, radula, soft anatomy, and embryonic development, Houbrick (1987) elevated Supplanaxis to full genus. Thiele (1929: 203) also established Proplanaxis as a new subgenus of Planaxis, with Planaxis planicostatus G.B. Sowerby I, 1825, as type species by monotypy. Houbrick (1987: 4) treated Proplanaxis as a synonym of Planaxis, but our molecular tree shows P. planicostatus to be nested within the Supplanaxis clade (see below), and not within Planaxis s.s., a taxonomic position in agreement with Laidre and Vermeij (2012), who had already used the combination Supplanaxis planicostatus. The names Supplanaxis and Proplanaxis having been published simultaneously, we act as First Revisers and, under Art. 24.2 of the ICZN Code, give precedence to the name Supplanaxis over Proplanaxis. (i) The shell from Barbados was accompanied by a non-binominal legend Buccinum B.r.paruum nigrum, ex toto laeve and the locality Barb., which was rendered by Bruguière as Buccinum brevirostrum parvum nigrum extoto laeve Barbadense. Some of the Lister collection was acquired by Sloane, and the Sloane collection was one of the founding collections of the British Museum (Natural History) [now The Natural History Museum], in London. However, there is no material in NHMUK corresponding to Lister's illustration (Wilkins 1953;and A. Salvador, pers. comm.), and this specimen is to be considered lost. (ii) Bruguière was a member of the second expedition (1773-1774)   as Banks, Forster and Solander. It is therefore possible that Broussonet could have acquired Cook artifacts from this circle and brought them back with him to Paris. He took part in the French Revolution but, as a member of the "Girondins", had to leave Paris in 1793 and his belongings were seized. However, Broussonet's artifacts are not nowadays traceable in any French museum where they would have been deposited by the Revolutionary powers. Dr Adrienne Kaeppler, curator for the Pacific Islands in the Department of Anthropology at the Smithsonian's National Museum of Natural History, has advised us that Maori cloaks from Cook's voyage to New Zealand are not known to be decorated with shells, and a Tonga or Hawaii provenance for Broussonet's artifacts would have been more likely.

Supplanaxis nucleus
There are thus no specimens left that could be considered a syntype of Buccinum nucleus. The specimens from Madagascar or the Pacific would not have been conspecific with the shell from the Caribbean illustrated by Lister; they might have been Supplanaxis niger (Quoy and Gaimard, 1833) ( Fig. 3D), which bears a strong resemblance to S. nucleus-especially to the eye of an 18 th century conchologist.
Because the description of Buccinum nucleus referred to more than one species, and because the Caribbean species designated under that name is a complex of two cryptic species, it is desirable to stabilize the nomenclature by the fixation of a neotype, which we designate herein.
Planaxis Description. Shell. Shell (Figs 3A, 4A-G, 7A-L at right) large for the genus, solid, littoriniform, medium high-spired, consisting of 5+ (apex generally broken or corroded in adults) moderately convex whorls separated by impressed suture, last whorl occupying ca 75-82% of total shell height. Shell surface usually covered with 15-17 even, deeply incised spiral grooves extending over all shell height, especially well marked adapically and on base behind outer lip. Aperture ovoid, expanding abaxially at mid-height, occupying ca 50% of total shell height, with 7-10 sharp internal lirae, columellar pillar extending almost to the abapical point of the aperture, delimiting a narrow siphonal notch, callus adpressed to parietal and columellar areas, parietal tooth strong, blunt. Color uniformly reddish brown to dark violet, parietal tooth and columellar callus orange violet. Height 11-18 mm.
Neotype (Figs 3A, 4D, 7L at right) reddish brown, with weak parietal tooth, height 16.9 mm. Radula. Radula taenioglossate (Fig. 5A). Rachidian pentagonal, with broad basal plate, long basolateral extensions, and short, rounded, median basal projection (Fig. 5B, D). Upper lateral part of basal plate with narrowly rectangular basal denticle. Rachidian cutting edge broad, comprising two-thirds width of tooth. Cutting edge shallowly and smoothly concave with large, squarish, spatulate median cusp, flanked by three to four smaller pointed denticles on each side. Lateral tooth with broad, high basal plate with central supporting ridge, rounded margins, and long lateral basal extension (Fig. 5C). Cutting edge broad with large, squarish, central cusp flanked by three to four inner and two to three sharp outer denticles. Marginal teeth elongate with curved paw-like tips (Fig. 5A, F). Inner marginal tooth with narrow flange along length of shaft outer edge, and with concave, rake-like tip with nine to ten rounded denticles. Outer marginal tooth with broad, membranous flange along distal outer edge, and with broad, bilobed tip bearing approximately 28 to 30 small, rounded denticles.
Distribution and ecology. Throughout the Caribbean in high energy, intertidal environments, on hard substrates, from large boulders to small cobbles and pebbles, in populations of moderate to large size (Vermeij 1973;Bandel 1976;Houbrick 1987). Its range extends from Palm Beach Inlet, Florida, in the north, to the northern coast of South America, from Veracruz, Mexico in the west (Tunnell 1974) and as far east as Trinidad and Tobago off Venezuela, including the Gulf of Mexico, Caribbean Sea and the Antillean Arc (Fig. 7). It is "rare" in Bermuda (Sterrer and Schoepfer-Sterrer 1986: 413), which may indicate that it only forms pseudopopulations there, and is absent from the Guyanas and Brazil.
Remarks. It is evident, from the morphology of the rachidian, that the radular descriptions of Troschel (1858: pl. 12, fig. 9A-D), Houbrick (1987: fig. 18A-E) and Simone (2001: fig. 88) were based on S. nucleus. Houbrick (1987: figs 19A, B, 20A-I, 21A-G, 22A-F) described the external anatomy and gross morphology of the mantle cavity, alimentary tract, and excretory, reproductive and nervous systems based on specimens from the same lot that yielded the radula (USNM 809780). Examination of the shells from this lot confirmed them all to be S. nucleus, although other lots collected from the same site at subsequent occasions comprised mixtures of the two species. Simone (2001: figs 56, 74, 88, 189-206) also provided a detailed description of the anatomy based on material obtained from several sites in Venezuela. The figured shells (figs 17-19) are S. nucleus, but both species occur in Venezuela and it is possible the anatomical descriptions are composite.
Despite its abundance in modern-day Caribbean faunas, Supplanaxis nucleus is surprisingly recorded as a fossil only in the Upper Pleistocene of Venezuela (Weisbord 1962: 168, pl. 14, figs 17, 18;and B. Landau, pers. comm.). It is not known from the well-preserved horizons of Florida. Description. Shell. Shell (Figs 3C, 4H-O, 7A-L at left) medium-sized for the genus, solid, littoriniform, high-spired, consisting of 5+ (apex generally broken or corroded in adults) weakly convex whorls separated by impressed suture, last whorl occupying ca 75-80% of total shell height. Shell surface rarely smooth, usually covered with 15-17 even, incised spiral grooves extending over all shell height, especially well marked on base. Aperture ovoid, occupying ca 50% of total shell height, with or without sharp internal lirae, 7-11 in number, outer lip regularly convex, columellar pillar truncated above broad siphonal notch, callus adpressed to parietal and columellar areas, particularly expanded adapically, its outer edge thickened and slightly raised, parietal tooth strong, blunt. Color uniformly reddish brown to dark violet, parietal tooth and columellar callus often lighter. Height 11-13.5 mm.
Radula. Radula taenioglossate (Fig. 6A). Rachidian pentagonal, with broad basal plate, long basolateral extensions, and short, rounded, median basal projection (Fig. 6B,  D). Upper lateral part of basal plate with robust, rounded basal denticle. Rachidian cutting edge rather narrow, comprising one-half width of tooth, but thick and heavily buttressed. Cutting edge sharply concave at midline, with narrowly rectangular, pointed median cusp, flanked by two to four smaller, robust, pointed denticles on each side. Lateral tooth with broad, high basal plate with central supporting ridge, rounded margins, and long lateral basal extension (Fig. 6C). Cutting edge broad with large, squarish, central cusp flanked by three to four inner and two to three sharp outer denticles. Marginal teeth elongate with curved paw-like tips (Fig. 6A, F). Inner marginal tooth with narrow flange along length of shaft outer edge, and with concave, rake-like tip with nine to ten rounded denticles. Outer marginal tooth with broad, membranous flange along distal outer edge, and with broad, bilobed tip bearing approximately 32 to 34 small, rounded denticles.
Distribution and ecology. The range of Supplanaxis nancyae extends from Miami Beach, Florida, in the north, to the northern coast of South America, from Veracruz, Mexico in the west (Tunnell 1974) and as far east as Trinidad and Tobago off Venezuela, including the Gulf of Mexico, Caribbean Sea and the Antillean Arc (Fig. 7). This species can occur in syntopy with Supplanaxis nucleus (Fig. 8) and museum lots across its range may comprise mixtures of the two species; ~ 30% of museum lots in the USNM with more than a single specimen included both species. Guadeloupe specimens of the two species are illustrated by Lamy and Pointier (2017: pl. 40, fig. 6A, B [nucleus], 6C-E [nancyae], both as S. nucleus). It is unknown whether this species also occurs in Bermuda.
Remarks. Petit (2013: 9) questioned the precision of the type locality and presented circumstantial evidence that Petuch's original material consisted of shells occupied by hermit crabs.
Planaxis nancyae was described based on two specimens, both unusually smooth and reddish for the species. In the absence of sequenced topotypic material, we are confident of the identity of the holotype based on our examination of more than 2,100 specimens of Supplanaxis from all over the Caribbean, and particularly material from Jamaica and the Virgin Islands that conforms with both our sequenced material from the Lesser Antilles and the Haiti holotype. The shell of S. nancyae differs from that of S. nucleus in its generally smaller size and lighter color, the regularly convex shape of the aperture, and its adapically expanded callus.
The protoconch figured by Houbrick (1987: fig. 17H) for S. nucleus may be that of S. nancyae, as all the adult shells in that lot (USNM 714035) represent the latter spe- cies. The radula of S. nancyae was described and figured as S. nucleus by Bandel (1984: fig. 59; pl. 2 figs 6, 8), as is evident from the morphology of the rachidian. Indeed, Bandel (1984) mistook the comparatively short lateral extensions of the rachidian described by Troschel (1858) as erroneous, but the latter were based on S. nucleus which has a broader cutting edge and comparatively shorter lateral extensions. The radular morphology of S. nancyae differs from that of S. nucleus primarily in its narrower and thicker cutting edge of the rachidian, the robustness and shape of the rachidian basal denticle, the narrower and more pointed rachidian central cusp, and the slightly greater number of denticles on the outer marginal tooth. There may be fewer flanking denticles on the rachidian in S. nancyae caused by fusion of the innermost denticles with the central cusp, as seen in the specimens examined herein from Guadeloupe (Fig. 6B, D), but the range of variation overlaps in the two species. These differences in tooth morphology strongly suggest the two species differ in trophic ecology, but this deserves further research. Regardless, in addition to features of the aperture of the shell, the characteristic rachidian provides diagnostic characters sufficient to separate S. nancyae from its co-occurring congener.
We note a strong resemblance with the fossil Planaxis ame Woodring, 1928 (342, pl. 25, fig. 16), known from the Upper Pliocene of Jamaica (type locality) and from the Upper Miocene Cercado Formation of the Dominican Republic (Landau, pers. comm.). Woodring compared Planaxis ame with P. nucleus, then the only known Recent Caribbean planaxid species, but the smooth forms of Supplanaxis nancyae are a better match, and we do not rule out that P. ame might turn out to be a senior synonym of S. nancyae. Description. Shell. Shell small for the genus, solid, littoriniform, high-spired, consisting of 5+ (apex broken or corroded in the two known specimens) weakly convex whorls separated by impressed suture, last whorl occupying ca 80% of total shell height. Shell surface smooth, except incised spiral grooves, 0-2 adapically below suture and 2-8 on shell base. Aperture ovoid, occupying ca 40% of total shell height, with 0-4 low internal lirae, with distinct siphonal notch, narrow callus adpressed to parietal and columellar areas, parietal tooth strong, blunt. Color dark brown olive in holotype, to light orange brown with white parietal tooth and columellar callus. Height 7.8-8.4 mm.

Genus
Remarks. Planaxis nucleola was described based on a single specimen, and the "probable holotype" (Fig. 3E)  Planaxis nucleola was explicitly separated from P. nucleus by Mörch (1876) by its much less acutely pointed shell and spiral grooves restricted to the base. It had never been illustrated, and we have not traced any citation of that species since Mörch (1876). It is currently indexed in MolluscaBase in the synonymy of S. nucleus based on Rosenberg (2009).
Discussion. The sympatry and syntopy of Supplanaxis nucleus and S. nancyae raise the question of the evolutionary mechanism that might have led to the emergence of two species. Unexpectedly, the phylogenetic tree did not resolve S. nucleus and S. nancyae as sister taxa. A specimen of the Panamic Supplanaxis planicostatus (G.B. Sowerby I, 1825) from Panama City [type locality: Galapagos Is; Fig. 3B], sequenced only for the 16S gene, is placed as sister to S. nancyae, albeit with no support. A further eastern Pacific Supplanaxis species is S. obsoletus (Menke, 1851) [type locality: Mazátlan, Pacific coast of Mexico]. The Caribbean and the Panamic species of Supplanaxis have obviously shared a long part of their evolutionary history, and more phylogenetic work needs to be done to properly assess their relationship.
funding from Fonds Européen de Développement Régional (FEDER) and Port Autonome de la Guadeloupe, and subsequently in 2017 by Dominique Lamy; in Curaçao as part of the Bivalve Assembling the Tree-of-Life project and supported by the U.S. National Science Foundation (NSF) Assembling the Tree of Life (AToL) program (award