Squaliformes: Centrophoridae: Centrophorus) in North Atlantic Waters - Current Status and Taxonomic Issues

Home , Centrophoridae, Centrophorus, Centroscymnus, Dalatiidae, Deania, Gulper shark

W&M ScholarWorks

VIMS Articles Virginia Institute of Marine Science

2014

Species diversity of the deep-water gulper sharks (Squaliformes: Centrophoridae: Centrophorus) in North Atlantic waters - current status and taxonomic issues

A Verissimo Virginia Institute of Marine Science

CF Cotton Virginia Institute of Marine Science

RH Buch

J Guallart

GH Burgess

Follow this and additional works at: https://scholarworks.wm.edu/vimsarticles

Part of the Aquaculture and Fisheries Commons

Recommended Citation Verissimo, A; Cotton, CF; Buch, RH; Guallart, J; and Burgess, GH, Species diversity of the deep-water gulper sharks (Squaliformes: Centrophoridae: Centrophorus) in North Atlantic waters - current status and taxonomic issues (2014). Zoological Journal Of The Linnean Society, 172(4), 803-830. 10.1111/zoj.12194

This Article is brought to you for free and open access by the Virginia Institute of Marine Science at W&M ScholarWorks. It has been accepted for inclusion in VIMS Articles by an authorized administrator of W&M ScholarWorks. For more information, please contact [email protected]. bs_bs_banner

Zoological Journal of the Linnean Society, 2014, 172, 803–830. With 12 ﬁgures Downloaded from https://academic.oup.com/zoolinnean/article-abstract/172/4/803/2433424 by VIRGINIA INSTITUTE OF Marine Science user on 30 October 2018 Species diversity of the deep-water gulper sharks (Squaliformes: Centrophoridae: Centrophorus) in North Atlantic waters – current status and taxonomic issues

ANA VERÍSSIMO1,2*, CHARLES F. COTTON2,3,4, ROBERT H. BUCH5,6, JAVIER GUALLART7 and GEORGE H. BURGESS5

1CIBIO-UP – Center for Research in Biodiversity and Genetic Resources, Rua Padre Armando Quintas, Campus Agrário de Vairão, 4485-661 Vairão, Portugal 2Department of Fisheries Science, Virginia Institute of Marine Science, College of William and Mary, P.O. Box 1346, Gloucester Point, VA 23062, USA 3Department of Marine and Environmental Sciences, Savannah State University, 3219 College Street, Box 20467, Savannah, GA 31404, USA 4Florida State University Coastal and Marine Lab, 3618 Highway 98, St. Teresa, FL 32358, USA 5Florida Program for Shark Research, Florida Museum of Natural History, University of Florida, Gainesville, FL 32611, USA 6NOAA Fisheries Service, Northeast Regional Office (NERO), Gloucester, MA 01930, USA 7Laboratorio de Biología Marina, Departamento de Zoología, Universitat de València. E-46100 Burjassot (Valencia), Spain

Received 18 December 2013; revised 15 July 2014; accepted for publication 28 July 2014

The gulper sharks (genus Centrophorus) are a group of deep-water benthopelagic sharks with a worldwide distribution. The alpha taxonomy of the group has historically been problematic and the number of species included in the genus has varied considerably over the years and is still under debate. Gulper sharks are routinely caught in mid- and deep-water fisheries worldwide and some have shown a considerable decline in abundance in the last few decades. Clear and consistent species discrimination of Centrophorus is essential for an efficient and sustain- able management of these fisheries resources. Our study used molecular cytochrome oxidase subunit I (COI) and 16S ribosomal RNA gene sequences and morphometric data to re-evaluate the diversity of Centrophorus in North Atlantic waters, including the Gulf of Mexico, the Caribbean, and the Mediterranean Seas. Molecular data separated North Atlantic Centrophorus into five well-supported groups whereas morphometric data separated these same five groups and suggested three additional groups for which no molecular data were available. Four of the five groups identified in the North Atlantic also occur in the Indian and/or Pacific Oceans, thus extending the reported range of some species considerably. A species identification key for North Atlantic Centrophorus is provided based on our findings.

ADDITIONAL KEYWORDS: alpha taxonomy – dichotomous key – dogﬁshes – molecular genetics – morphometrics – squaloid.

INTRODUCTION of small- to medium-sized species (< 200 cm total length, TL) commonly known as gulper sharks. These are The genus Centrophorus Müller & Henle, 1837 (order benthopelagic sharks often found along the outer con- Squaliformes: family Centrophoridae) comprises a group tinental shelves and upper continental and insular slopes at depths between 50 and 2350 m throughout the world’s *Corresponding author. E-mail: [email protected] oceans (Compagno, 1984; McEachran & Branstetter,

1984). The ﬁrst description of a gulper shark dates back original species types and problematic specimens cannot to the late 18th century, i.e. Centrophorus (Squalus) be made or are of limited usefulness because precise

squamosus (Bonnaterre, 1788), although the genus morphometric data are almost impossible to obtain from Downloaded from https://academic.oup.com/zoolinnean/article-abstract/172/4/803/2433424 by VIRGINIA INSTITUTE OF Marine Science user on 30 October 2018 Centrophorus was proposed only in the mid 19th century the type specimens. Another complicating factor stems by Müller & Henle (1837a, b). The genus was based from the unreliability of some diagnostic morphologi- on the description of Squalus granulosus by Bloch & cal characters previously used for species identifica- Schneider (1801), which then became the type species, tion. Some often-used diagnostic characters vary with i.e. Centrophorus granulosus (Müller & Henle, 1841). ontogeny, such as shape and size of pectoral fin inner Since then, over 30 nominal species have been re- margins, and shape of dermal denticles, teeth, or of ferred to the genus Centrophorus (Eschmeyer, 2013) dorsal spines (Cadenat, 1959a, c; Maurin, 1968; Bass although some of them were subsequently assigned to et al., 1976; Ledoux, 1970; Cadenat & Blache, 1981; other genera (e.g. Centroscymnus Bocage & Capello, Tabit, 1993; Guallart, 1998; Guallart, García-Salinas 1864 or Deania Jordan & Snyder, 1902). & Catalán, 2013; White et al., 2013). As a result, dis- Although some of the species were described more tinct ontogenetic stages of the same species (ju- than two centuries ago, the global diversity of gulper veniles vs. adults) have been considered as distinct taxa sharks is not yet fully described (Castro, 2011; Naylor (e.g. White et al., 2013 determined Centrophorus acus et al., 2012; present study). Many authors have made Garman, 1906, described from a juvenile, to be a junior regional or global revisions of Centrophorus alpha tax- synonym of C. granulosus). onomy over the last 50 years, but they have also noted Ambiguous and inconsistent species identification plus the provisional character of their conclusions (Bigelow the continued debate amongst some systematists over & Schroeder, 1957; Bass, D’Aubrey & Kistnasamy, 1976; few, cosmopolitan species vs. multiple, less wide- Cadenat & Blache, 1981; Compagno, 1984; McEachran ranging species have generated much confusion in the & Branstetter, 1984; Bass, Compagno & Heemstra, 1986; literature, and a worldwide revision of the genus Last & Stevens, 1994). Other authors have described Centrophorus is much needed. This is particularly rel- previously unidentified morphotypes or have suggest- evant for certain species of gulper sharks that are com- ed the necessity to describe some new species (Cadenat monly targeted or bycaught in deep-water fisheries. & Blache, 1981; Castro, 2011; Naylor et al., 2012; present Studies have already shown signs of localized deple- study). As such, the number of species of Centrophorus tion for some Centrophorus species (Graham, Andrew is still not well defined (e.g. 13 recognized species in & Hodgson, 2001; ICES, 2006) and six species of Compagno, Dando & Fowler, 2005; 13 species in Ebert Centrophorus have been listed as Vulnerable, Near & Stehmann, 2012; 15 species in Eschmeyer, 2013). Threatened, or Endangered on the IUCN Red List Species discrimination within Centrophorus has been (iucnredlist.org accessed on 18 November 2013). Ad- problematic almost since its origin and still remains equate management and conservation efforts require problematic today. There are several issues contrib- reliable species-specific catch and mortality time series, uting to the confusion in the alpha taxonomy of as well as species-specific life history and ecological Centrophorus. One of the main problems is the high data. These data can be biased if species identifica- degree of morphological similarity between nominal taxa, tion is ambiguous or inconsistent, thus compromis- making species determination extremely challenging ing the effectiveness and adequacy of any management even for those experienced with the genus. Moreover, and conservation plan (Iglésias, Toulhoat & Sellos, 2010). many original species descriptions often fail to de- The eastern North Atlantic is one of the regions where scribe informative characters required to clearly dis- catches of deep-water sharks have been declining over tinguish nominal species from their congeners, a problem the past decades and total allowed catches (TAC) have that is particularly prevalent amongst the earliest de- been set to zero since 2010 for several species, includ- scribed taxa like in Centrophorus granulosus (Bloch ing C. granulosus and C. squamosus (ICES 2012). Un- & Schneider, 1801) and Centrophorus lusitanicus (Bocage fortunately, the taxonomic status of North Atlantic & Capello, 1864). Compounding the situation, holotypes Centrophorus remains unsettled and the ambiguity in are non-existent for some species [most notably species identification facilitates misreporting and casts C. granulosus and Centrophorus uyato (Rafinesque, doubt on the validity of species-specific catch records. 1810)] or cannot be located (Centrophorus niaukang For instance, no catches of C. lusitanicus were report- Teng, 1959), whereas others are incomplete (the ed prior to 2009 in the eastern North Atlantic, but the Centrophorus squamosus holotype is a dried head) and/ reported landings of C. lusitanicus have been over 271 or in poor condition (the Centrophorus atromarginatus metric tonnes since a zero TAC for C. squamosus and Garman, 1913 holotype is dried and deformed, and other C. granulosus has been enacted (ICES 2012). types are gutted and preserved in ‘U’ shapes and other Several studies have reported on the diversity of non-natural positions). In the absence of reference ma- Centrophorus in North Atlantic waters, including the terial of reasonable quality, comparisons between the Gulf of Mexico, the Caribbean, and the Mediterranean

Seas. Using morphological data, Bigelow, Schroeder & in North Atlantic waters, including the Gulf of Mexico Springer (1953) listed the presence of three nominal and the Mediterranean Sea. As the baseline for our

species, i.e. C. granulosus, C. uyato, and C. squamosus, comparisons, we considered the number of Centrophorus Downloaded from https://academic.oup.com/zoolinnean/article-abstract/172/4/803/2433424 by VIRGINIA INSTITUTE OF Marine Science user on 30 October 2018 in the North Atlantic and the Mediterranean Sea. Cadenat species accepted by Eschmeyer (2013) as the list of & Blache (1981) looked at both juvenile and adult species currently considered valid for the genus. In our Centrophorus from the eastern Atlantic, and indicated study, molecular data were obtained from specimens the existence of three morphotypes (‘formes’ in the origi- sampled throughout the North Atlantic as well as from nal work), rather than species, tentatively identified several locations around the globe, and compared under as C. ‘forme’ granulosus, C. ‘forme’ lusitanicus, C. ‘forme’ a phylogenetic framework to provide an independent uyato/machiquensis, in addition to C. squamosus and perspective of the Centrophorus species diversity at re- an undescribed Centrophorus. Compagno (1984) also gional and global levels. This approach had a three- relied on morphometrics in listing C. acus, C. granulosus, fold purpose: (1) allow for a direct comparison of C. lusitanicus, C. squamosus, and C. uyato, later re- Centrophorus species diversity between the North At- vising the list to include C. granulosus, Centrophorus lantic and the rest of the world; (2) identify the most harrissoni McCulloch, 1915 (pending confirmation), problematic taxa in terms of species discrimination glob- C. lusitanicus, C. niaukang, C. squamosus, and ally; and (3) provide an initial framework for species Centrophorus tesselatus Garman 1906 (Compagno et al., delimitation within the genus. Concurrently, 2005). Muñoz-Chapuli & Ramos (1989) attempted to morphometric and morphological data from North At- provide a comprehensive review of the alpha taxono- lantic Centrophorus specimens were collected and ana- my of Centrophorus for the north-eastern Atlantic and lysed using univariate and multivariate statistics. These Mediterranean Sea and discussed several of the taxo- data were obtained from individuals included in each nomic issues surrounding the genus. Also using mor- of the resulting genetic clades, as well as from addi- phological data, they concluded that only four species tional specimens representing distinct morphotypes but occur in north-eastern Atlantic waters, namely for which no genetic data were available. The C. granulosus, C. lusitanicus, C. niaukang, and morphometric traits exhibiting maximum clade dis- C. squamosus, and did not recognize C. uyato as a valid crimination were identified and used in the construc- species. They further confirmed the morphotypes pro- tion of a species identification key for North Atlantic posed by Cadenat & Blache (1981), although they used Centrophorus. Some biological data (e.g. fecundity, size- C. niaukang to designate C. ‘forme’ granulosus, and at-maturity) collected for different Centrophorus species C. granulosus to designate C. ‘forme’ uyato-machiquensis addressed in our study were also compiled and are pre- (Muñoz-Chapuli & Ramos, 1989). Guallart (1998) in- sented here, providing life history distinctions that vestigated the diversity of Centrophorus in the Medi- underscore genetic and morphological patterns. terranean Sea, specifically addressing long-stated records of both C. granulosus and C. uyato in the region, and concluded that only a single species occupies the Medi- MATERIAL AND METHODS terranean Sea. His conclusions were in line with those SAMPLING suggested by many previous authors but never objec- We aimed to maximize the geographical coverage of our tively tested (e.g. Maurin, 1968; Tortonese, 1969; Maurin samples throughout the North Atlantic sensu lato (i.e. & Bonnet, 1970; Capapé, 1985; Muñoz-Chapuli & Ramos, Atlantic waters north of the equator, including the Gulf 1989). Nevertheless, the two species were still includ- of Mexico, Caribbean, and Mediterranean Seas), irre- ed in the list of Mediterranean sharks by Serena (2005). spective of the original species designation or the type In his recent book, Castro (2011) used morphometric of data collected (i.e. molecular or morphological data). characters to distinguish seven species of Centrophorus Samples included tissue samples (fin clips or muscle tissue) in the north-west Atlantic off the United States east for downstream DNA extraction and analysis, as well as coast, Gulf of Mexico, and Caribbean Sea: Centrophorus whole-body specimens for morphological characteriza- isodon (Chu, Meng & Liu, 1981), C. niaukang, tion. Molecular and morphological data were collected C. squamosus, C. tesselatus, two undescribed species from the same individual whenever possible. Biological (Centrophorus sp. A and Centrophorus sp. B), and C. uyato. data regarding maturity stage and fecundity were also More recently, Ebert & Stehmann (2012) reported only collected from whole-body specimens whenever possible. four species of Centrophorus in the North Atlantic, C. granulosus, C. lusitanicus, C. niaukang, and C. squamosus, but indicated that more may be present. MOLECULAR DATA COLLECTION AND ANALYSIS Given the historical and current ambiguity and in- Tissue samples were obtained from fresh Centrophorus consistency in the alpha taxonomy of Centrophorus, specimens caught opportunistically between 2005 and the main objective of this paper was to provide a com- 2012, during deep-water scientific research surveys (Dyb prehensive reassessment of the diversity of gulper sharks & Bergstad, 2004; Menezes et al., 2004; Cotton, 2010)

© 2014 The Linnean Society of London, Zoological Journal of the Linnean Society, 2014, 172, 803–830 806 A. VERÍSSIMO ET AL. Downloaded from https://academic.oup.com/zoolinnean/article-abstract/172/4/803/2433424 by VIRGINIA INSTITUTE OF Marine Science user on 30 October 2018

Figure 1. Sampling locations of North Atlantic Centrophorus specimens. Empty squares, specimens with molecular and morphological data; yellow circles, specimens with molecular data only; red triangles, specimens with morphological data only. (Colour version of figure available online.) and from commercial longliners and trawling vessels op- 5 mg bovine serum albumin, 0.025 units Taq polymer- erating in the North Atlantic sensu lato (Fig. 1). The cor- ase, 1× Taq buffer with 1.5 mM MgCl2 (Qiagen) and responding whole specimen vouchers were retained autoclaved ultra-pure water. PCR conditions for the whenever possible and stored in a museum collection for COI fragment consisted of an initial denaturation morphological examination. Alternatively, a photo voucher of 5 min at 94 °C, followed by 30 cycles of 1 min at 94 °C, was obtained whenever possible if whole specimens were 30 s at 55 °C, 1 min at 72 °C, and a final extension not kept. All fresh specimens were tentatively identified step of 5 min at 72 °C. PCR conditions for the 16S upon capture by the various collectors (including but not fragment consisted of an initial denaturation of 10 min restricted to the authors). Additional tissue samples were at 93 °C, followed by 40 cycles of 30 s at 93 °C, 1 min at obtained from the Kansas University Museum of Natural 50 °C, 2 min at 72 °C, and a final extension step of 10 min History Tissue Collection as well as from collaborators at 72 °C. All amplicons were cleaned with a QIAquick throughout the world (complete list in the Acknowledge- PCR Purification Kit (Qiagen) according to the man- ments section), and the corresponding voucher speci- ufacturer’s protocol. Individual gene fragments were se- mens were examined when available. In this case, the quenced using an ABI Big Dye Terminator Cycle species designations previously assigned to the corre- Sequencing Kit (Applied Biosystems, Warrington, UK), sponding specimens were retained. All tissue samples and the reactions were run on an ABI Prism 3130xl genetic were preserved in 95% ethanol or in 20% dimethyl analyzer (Applied Biosystems). The resulting DNA se- sulphoxide buffer saturated with NaCl (Seutin, White & quences were imported into SEQUENCHER v. 4.8 Boag, 1991). The genomic DNA (gDNA) was extracted (Gene Codes Corp., Ann Harbor, MI, USA) and checked using a Qiagen DNeasy Tissue kit (Qiagen, Valencia, CA, for quality and accuracy in nucleotide base assignment. USA) according to the manufacturer’s instructions, except Additional nucleotide sequence data from Centrophorus for a final elution with 50 μL of Milli-Q autoclaved water. specimens collected within the North Atlantic The molecular data included nucleotide sequences of sensu lato and elsewhere were obtained from two mitochondrial gene regions, namely a 560-bp frag- GenBank (N = 49; accession numbers: AY147884; ment of cytochrome oxidase subunit I (COI; primers FishF2 DQ108227−108232; DQ108240−108242; EU003893 5′ TCGACTAATCATAAAGATATCGGCAC 3′; FishR1 5′ −003897; EU398647−398668; GU130627−130628; TAGACTTCTGGGTGGCCAAAGAATCA 3′; Ward et al., GU130701; HM239654−139655; JN596815−596821). The 2005) and a 507-bp fragment of 16S ribosomal RNA (16S; species names associated with each accession number primers 16SarL: 5′-CGCCTGTTTATCAAAAACAT-3′; were retained. Sequence data from a single specimen 16SbrH: 5′-CCGGTCTGAACTCAGATCACGT-3′; Palumbi of Deania calcea (Lowe, 1839) (family Centrophoridae) et al., 1991). The use of two gene regions was chosen to collected off Portugal, in the eastern North Atlantic, assess congruence in clade delimitation. The fragments was used as the outgroup taxon (GenBank Accession were amplified separately for each individual fish via No. KM281931 for the COI sequence, and KM281899 PCR in 25 μL reactions containing 10–20 ng gDNA, 1 mM for the 16S sequence). No specimen voucher was re- of each primer, 200 mM each deoxynucleotide triphosphate, tained for this specimen.

All sequences from each mitochondrial gene region neighbour-joining tree, and from additional morphotypes were aligned in GENEIOUS PRO v. 5.4.6 (Biomatters not represented in the molecular matrix. The initial

Ltd) using the ClustalW multiple alignment algo- morphological data matrix included a total of 110 Downloaded from https://academic.oup.com/zoolinnean/article-abstract/172/4/803/2433424 by VIRGINIA INSTITUTE OF Marine Science user on 30 October 2018 rithm (Thompson, Higgins & Gibson, 1994). The cor- morphometric measurements taken on both fresh and responding ﬁnal alignments were imported into MEGA preserved specimens (totalling 65 and 35% of meas- v. 5.0 (Tamura et al., 2011) in which neighbour- ured specimens, respectively). After a preliminary in- joining trees were constructed for each gene region using spection of the distribution of each morphometric genetic p-distances amongst haplotypes. Support for measurement amongst the different genetic clades, a individual clades was evaluated with 10 000 boot- subset of 21 measurements showing variable distri- strap replicates. The resulting topologies were not meant butions amongst clades was deemed potentially useful to infer the phylogenetic relationships amongst taxa in species discrimination and used for further analy- but only to provide information on clade delineation sis (Table 1). Measurements > 10 cm were performed based on sequence differences amongst specimens. For using a metric tape with a precision of ± 1 mm, whereas this purpose, we used genetic p-distances that calcu- measurements < 10 cm were carried out with calli- late the absolute number of nucleotide positions that pers with a precision of ± 0.1 mm. All measurements differ between each pair of sequences, thus empha- were taken by authors A.V. and C.C., and a subset of sizing differences amongst very similar sequences. By specimens was measured by both authors together to using strictly mitochondrial genes, it is not possible compare measuring methodologies and minimize to detect potential hybridization between species as the interobserver error. Biological data included matur- mitochondrial genome is strictly maternally inherit- ity stage (i.e. immature, maturing, and adult), based ed. Thus, in the case of hybrid individuals (if they exist), upon inspection of internal sexual organs and clasper only the maternal haplotypes were determined and no development and calciﬁcation (sensu Stehmann, 2002), information was obtained from the paternal species. and fecundity (i.e. number of enlarged/yolked oocytes or embryos in the uteri). To reliably and unambiguously discriminate species MORPHOLOGICAL ANALYSIS based on morphology, the set of diagnostic morpho- Morphological and biological data were obtained logical characters should ideally be independent of size from specimens representing each of the resulting (i.e. isometric growth) and sex. As such, the morpho- North Atlantic Centrophorus clades present in the logical data collected for each clade were taken from

Table 1. Morphometric measurements adapted from Compagno (1984). See Figure S1 for details

Abbreviation Description

TL Distance from snout tip to posterior tip of caudal fin when aligned with body axis PN Distance from snout tip to anterior margin of nostrils POR Distance from snout tip to anterior mouth opening POB Distance from snout tip to anterior margin of eyes HDL Distance from snout tip to fifth gill slit PP2 Distance from snout tip to pelvic fin origin, taken ventrally PD1 Distance from snout tip to first dorsal fin origin. PD2 Distance from snout tip to second dorsal fin origin. IDS Distance from first dorsal fin insertion to second dorsal fin insertion DCS Distance from second dorsal fin insertion to upper caudal fin origin PCA Distance from pelvic insertion to lower caudal origin INO Distance between anterior border of eyes MOW Mouth width measured between mouth corners P1A Length of pectoral fin anterior margin P1I Distance from pectoral fin insertion to posterior tip of pectoral fin inner margin D1B Distance from first dorsal fin origin to first dorsal fin insertion D1H Length from first dorsal fin base to apex of fin D2B Distance from second dorsal fin origin to first dorsal fin insertion D2H Length from second dorsal fin base to apex of fin CDM Distance from upper caudal origin to upper tip CPV Distance from lower caudal origin to ventral tip

© 2014 The Linnean Society of London, Zoological Journal of the Linnean Society, 2014, 172, 803–830 808 A. VERÍSSIMO ET AL. a wide range of sizes and roughly equal numbers of sulting discriminant functions. Additional assump- males and females (whenever possible) without cor- tions of nonmulticolinearity were tested prior to

recting for size or gender effects. Thus, significant dif- conducting the DFA, and prior probabilities of group Downloaded from https://academic.oup.com/zoolinnean/article-abstract/172/4/803/2433424 by VIRGINIA INSTITUTE OF Marine Science user on 30 October 2018 ferences amongst the groups are inclusive of ontogenetic membership were estimated based on the empirical and sexual dimorphism variation within and amongst group sample sizes. The statistical significance of the species. It should be noted that a detailed descrip- discriminant functions was tested with a MANOVA tion of the ontogenetic variation for each of the clades using Wilks’ lambda test. The reduced set of reported here was beyond the scope of this paper; morphometric ratios obtained with the DFA was used however, this topic should be explored in future in the construction of a species identification key for studies. North Atlantic Centrophorus. We also present the origi- Raw body morphometric data were log-transformed nal raw body measurements expressed as percentage and principal component analysis (PCA) was used to explore of total length to facilitate comparisons with pub- morphological differences amongst groups (i.e. clades and/ lished and future studies of gulper sharks world- or morphotypes), as well as to identify the principal set wide. All computations were performed in R (R of morphometric measurements contributing most for Development Core Team, 2012). amongst-group differences. In cases for which data were In addition to the morphometric analyses, general missing for a given body measurement (1.3% of total), descriptions of dermal denticles and teeth morphol- the corresponding value was estimated via the clade- ogy are provided for each morphotypic clade. These de- specific regression equation of body measurement vs. total scriptions refer exclusively to the adult stages and thus, length. Loadings in principal component (PC) 1 were given the known ontogenetic variation in these char- roughly equal for all variables and of the same sign, re- acters (see Introduction for details), should only be con- flecting the overall correlation of morphometric meas- sidered for comparisons with other adult specimens. urements with body size and thus were not informative in terms of amongst-group differences. The subset of measurements showing higher loadings on PC2 and PC3 were RESULTS combined into morphometric ratios focusing on different aspects of body, head, and fins. MOLECULAR DIVERSITY WITHIN CENTROPHORUS All morphometric ratios were initially inspected for Nucleotide sequences of the COI and 16S mitochondrial their within- and amongst-group variability, using box- gene regions were obtained from a total of 272 (67% plots. The subset of ratios exhibiting differential amongst- of North Atlantic samples) and 87 individuals (60% of group distributions were chosen for further analysis aiming North Atlantic samples), respectively. For either gene at determining their usefulness as potential diagnostic region, the data set included specimens originally as- characters. The selected morphometric ratios were log- signed to C. acus, C. atromarginatus, C. granulosus, transformed and tested for significant differences amongst C. harrissoni, C. isodon, Centrophorus moluccensis groups with a multivariate analysis of variance (MANOVA). McCulloch, 1915, C. niaukang, C. squamosus, C. uyato, Prior to the MANOVA, the dependent variables (i.e. and Centrophorus zeehaani White, Ebert & Compagno, morphometric ratios) were inspected for the presence of 2008, as well as several unidentified Centrophorus (i.e. outlier observations using box-plots (univariate outli- Centrophorus sp.). The 16S data set also included a ers) and Mahalanobis squared distances (multivariate single sequence from a specimen originally identified outliers). Multivariate normality was tested using the as C. lusitanicus. The putative Centrophorus taxa not Shapiro−Wilks’ test and deviations to multivariate nor- represented in our samples were Centrophorus mality were visually inspected with a Chi-square Q-Q seychellorum, Centrophorus robustus, C. tesselatus, and plot. Outlier observations were removed from the analy- Centrophorus westraliensis (but see Discussion). sis. Homogeneity of variances amongst species groups The alignment of homologous fragments of the was tested for each dependent variable, with alpha levels COI gene region yielded 35 unique haplotypes (GenBank adjusted for multiple comparisons via strict Bonferroni Accession no.s KM281900–KM281930 for newly correction. Because the data did not show multivariate recorted COI haplotypes; see Supporting Information normality, homoscedasticity was tested using nonparametric Table S1) and 67 variable positions, of which 58 were Fligner−Killeen tests. parsimoniously informative. A total of 18 haplotypes The morphometric ratios showing significant differ- was identified for the 16S gene region (GenBank ences amongst species were subsequently used in a dis- Accession no.s KM281885–KM281898 for newly rec- criminant function analysis (DFA). This analysis was orded 16S haplotypes; see Table S1), differing in 23 vari- aimed at identifying a reduced set of diagnostic char- able positions of which 15 were parsimoniously acters producing maximum discrimination amongst informative. The best alignment of 16S sequences groups, followed by testing predictions of group (species) implied the opening of several gaps (N = 10); thus, es- membership of individual specimens based on the re- timation of genetic p-distances and construction of the

© 2014 The Linnean Society of London, Zoological Journal of the Linnean Society, 2014, 172, 803–830 DIVERSITY OF NORTH ATLANTIC CENTROPHORUS 809 neighbour-joining trees were performed using two a sister clade to that of C. isodon/C. harrissoni in both methods: pairwisedeletion of gaps and complete de- the COI and 16S neighbour-joining trees, with high

letion of gaps. The clade structure of the resulting 16S and moderate bootstrap support values, respectively. Downloaded from https://academic.oup.com/zoolinnean/article-abstract/172/4/803/2433424 by VIRGINIA INSTITUTE OF Marine Science user on 30 October 2018 trees was congruent between methods but the pairwise Clade C specimens included individuals sampled in deletion option produced higher bootstrap support values. the North Atlantic off Cape Verde (designated as Thus, all results shown for the 16S data reflect those C. uyato) and in the Gulf of Mexico (designated as C. cf. obtained under the pairwise deletion option. The overall isodon), as well as several other specimens collected mean genetic p-distance amongst haplotypes (± SD) was off Taiwan and Indonesia originally designated as 0.036 ± 0.004 for the COI and 0.012 ± 0.003 for the 16S C. isodon. Assuming that the C. isodon specimen clus- gene regions, respectively. tering with C. moluccensis in the COI tree is a misi- The neighbour-joining trees of genetic p-distances for dentification, the relationship between C. isodon (and the COI gene region produced a total of eight clades thus Clade C) and C. harrissoni differed between gene of moderate to high bootstrap support values (73 to regions. Centrophorus harrissoni and C. isodon clus- 100%; Fig. 2A), whereas the 16S gene region recov- tered in the same clade in the 16S tree (100% simi- ered nine clades with comparatively lower bootstrap larity in their nucleotide sequences) whereas support values (< 93%; Fig. 2B). The number of clades C. harrissoni clustered in a separate clade from those and the clade composition were generally concordant identified as C. isodon in the COI tree (Fig. 2). However, between the two mitochondrial markers, although the in the latter case, Clade C was not monophyletic. relationships amongst some of the clades differed (Fig. 2). Clade D included specimens designated either as Each clade was generally composed of one widely dis- C. acus, C. granulosus,orC. niaukang. In the North tributed, high frequency haplotype and several locally Atlantic, specimens were collected from Portugal to Cape distributed, low frequency ones (one to three individ- Verde in the east, and from off Virginia (USA) south uals; see Table S1). to the Bahamas in the west. Several specimens were There was considerable heterogeneity in species des- also obtained from the Gulf of Mexico. Generally, the ignations within most of the genetic clades: speci- name C. granulosus was used to designate large specimens clustering in Clade A were originally identified mens caught in the eastern margin whereas those as C. granulosus, C. uyato, and C. zeehaani. In some caught on the western margin were identified as cases, specimens bearing the same species designa- C. niaukang. Other specimens clustering in this clade tion were included in two distinct clades: C. granulosus were sampled from the Indian and Pacific Oceans. All was used to designate specimens clustering in Clades C. acus specimens clustering in Clade D were collect- A and D. By contrast, some taxon names were found ed off Japan. exclusively in a single clade (C. moluccensis, C. niaukang, Finally, Clade E clustered all individuals originally C. uyato, and C. zeehaani) although that clade also in- identified as C. squamosus and sampled in all ocean cluded specimens bearing other species designations basins. In the North Atlantic, samples were collected (C. isodon and C. granulosus). Clades showing con- along the European continental slope from off Ireland sistent species identification refer only to those of south to Portugal, and on the mid-Atlantic ridge off C. atromarginatus and C. squamosus. the Azores. One unidentified specimen caught off Algeria, Both mitochondrial markers indicated the exist- on the western Mediterranean Sea, clustered in Clade ence of five distinct Centrophorus clades in the North E. However, no photo voucher or morphological ex- Atlantic, herein designated as clades A to E (Fig. 2). amination was obtained from this specimen. Clade A was composed of specimens sampled in all major Clades obtained in the current genetic analysis ocean basins, i.e. Atlantic, Indian, and Pacific Oceans. but not detected in the North Atlantic include those In the North Atlantic, specimens originally identified comprised of several specimens designated as as C. granulosus or as C. uyato were collected from the C. atromarginatus (COI and 16S trees), C. moluccensis Canary Islands to Cape Verde, as well as throughout (COI and 16S trees), and C. lusitanicus (16S tree only) the Mediterranean Sea (i.e. Greece, Algeria, and the (Fig. 2). There was also an additional clade present Balearic Sea) and in the Gulf of Mexico. Additional in the 16S tree represented by an unidentified samples collected outside the North Atlantic include Centrophorus from the Indian Ocean (Fig. 2B). specimens of C. zeehaani from Australia and a single unidentified sample from South Africa. Clade B included specimens caught exclusively in MORPHOLOGICAL ANALYSIS OF NORTH the Gulf of Mexico, the Bahamas, and Jamaica. No ATLANTIC CENTROPHORUS species identification was provided for most of the The PCA was conducted on a total of 75 specimens sampled specimens clustering in this clade except for including representatives from all genetic clades iden- two specimens tentatively identified as C. cf. uyato (sensu tified above (see Table S2 for details), namely A (N = 19), McLaughlin & Morrissey, 2005). Clade B clustered as B(N = 13), C (N =3),D(N = 29), and E (N = 11). The

Distribution Pacific Indian Atlantic

A) COI C. granulosus Downloaded from https://academic.oup.com/zoolinnean/article-abstract/172/4/803/2433424 by VIRGINIA INSTITUTE OF Marine Science user on 30 October 2018 88 C. uyato X X Clade A 98 C. zeehani 99 C. atromarginatus XX

99 C. moluccensis XX C. isodon

73 Centrophorus sp. X Clade B 99 C. isodon 98 89 C. granulosus XX X Clade C C. uyato 73 C. harrissoni XX

C. acus C. granulosus XX X Clade D 99 C. niaukang 93

C. squamosus XX X Clade E

Deania calcea

0.03 0.02 0.01 0.00 Distribution Pacific Indian Atlantic C. granulosus 93 C. uyato B) 16S C. zeehani XX X Clade A Centrophorus sp.

C. atromarginatus X X

Centrophorus sp. X

89 C. moluccensis X X C. cf. granulosus

93 C. granulosus X Clade D C. niaukang

C. lusitanicus XX , 80 C. isodon C. harrisoni X X Clade C C. granulosus X Centrophorus sp. Clade B

72 C. squamosus XX X Clade E

Deania calcea

0.012 0.010 0.008 0.006 0.004 0.002 0.000

Figure 2. Neighbour-joining tree of genetic p-distances amongst A, cytochrome oxidase I (COI) haplotypes and B, 16S ribosomal RNA (16S) haplotypes of Centrophorus. Support values for each clade, based on 10 000 bootstrap replicates, are indicated on top of each branch (if > 70%). The original species identification of specimens included in each clade is indicated in italics. The provenance of specimens in each clade (i.e. Atlantic, Indian, and Pacific Oceans) is marked with X, and the clades retrieved in the North Atlantic are identified from A–E.

Table 2. Loadings of morphometric measurements on the indicated lower contribution to clade discrimination ﬁrst three principal components (PC1, PC2, and PC3). Empty (F = 41.1 vs. 113.5 and 91.85, respectively).

cells refer to values close to zero Given the similarly low although significant contri- Downloaded from https://academic.oup.com/zoolinnean/article-abstract/172/4/803/2433424 by VIRGINIA INSTITUTE OF Marine Science user on 30 October 2018 bution of PN/POR (F = 13.2) and POB/HDL (F = 13.3) PC1 PC2 PC3 to clade discrimination, we conducted three independent DFAs: DFA1 included five morphometric ratios (i.e. Variance explained (%) 86 6.5 1.6 PN/POR, POB/HDL, P1A/P1I, D1H/D2H, and DCS/ TL −0.233 0.110 CDM); DFA2 excluded PN/POR; and DFA3 excluded PN −0.222 −0.380 POB/HDL. The purpose was to evaluate the rel- POR −0.226 −0.334 evance of including two ratios of snout proportions POB −0.227 −0.296 instead of one and, thus, to select the most parsimo- HDL −0.221 −0.138 nious set of morphometric ratios resulting in maximum PP2 −0.231 PCA −0.199 0.274 0.501 clade differentiation. PD1 −0.215 0.215 −0.176 All linear discriminant functions (LDs) were statis- PD2 −0.231 0.147 tically significant and had a percentage separation IDS −0.228 0.254 ranging between 76 and 80% for LD1, 17 and 20% for DCS −0.188 0.438 0.185 LD2, and 3.2 and 3.7% for LD3, accounting for c. 97.0 INO −0.226 −0.102 −0.115 to 97.5% of the total variance amongst clades (Table 3). MOW −0.224 −0.188 LD1 was most heavily loaded by the proportions of pec- P1A −0.229 toral fin margins (P1A/P1I) and dorsal fin heights (D1H/ P1I −0.169 0.525 −0.372 D2H), whereas LD2 was also influenced by snout D1B −0.209 −0.337 0.164 proportions (PN/POR and/or POB/HDL) (Table 4). LD3 D1H −0.210 0.273 0.161 was most affected by the relative caudal peduncle length D2B −0.210 −0.309 0.110 (DCS/CDM; Table 4). The classification precision was D2H −0.224 −0.130 89% for Clade A, 97% for Clade D, and 100% for Clade CDM −0.226 E (Table 5). Clade B had higher classification preci- CPV −0.223 −0.148 0.110 sion when using only four variables and excluding POB/ HDL, (92 vs. 85%, respectively; Table 5). See Table 1 for abbreviations. Considering the similarity in total variance explained by the three DFAs conducted, and the improved clas- first three PCs explained 94% of the total variance in sification precision of DFA3 using a smaller set of the data and the variables most strongly contribut- morphometric ratios (Table 4), we chose to use the cor- ing to amongst-group variance (i.e. had higher load- responding four morphometric ratios (PN/POR, P1A/ ings in PC2 and PC3) were related to snout shape, P1I, D1H/D2H, and DCS/CDM) as diagnostic traits to pectoral and dorsal fin shape, and caudal peduncle separate the four Centrophorus morphotypes included length (Table 2). After a preliminary inspection of the in the DFA analysis. LD1 and LD2 provided maximum distribution of several candidate morphometric ratios separation amongst Clade A + Clade B, Clade D, and Clade amongst the different genetic clades, six ratios exhib- E (Fig. 4A). This separation reflects the taller first dorsal iting unequal amongst-clade distributions were re- fins in Clades A and B compared with equal dorsal fin tained for downstream analysis: PN/POR, POB/HDL, heights in Clades D and E, as well as a much shorter P1A/P1I, D1H/D2H, D1H/P1I, DCS/CDM (see Table 1 pectoral inner margin in Clade E compared with the other for abbreviations; Fig. 3). clades (Fig. 3). LD2 and LD3 provided a good separation Downstream analysis included only the specimens between Clades A and B (Fig. 4B), mostly driven by the representative of clades A (N = 19), B (N = 13), D shorter snout length and caudal peduncle in Clade A (N = 29), and E (N = 11) owing to the small number compared with Clade B (Fig. 3). The four Centrophorus of specimens examined for Clade C (N = 3). The morphotypes corresponding to Clades A, B, D, and E are MANOVA showed significant differences in the shown in Figures 5–8. morphometric ratios amongst genetic clades (Wilks’ The small sample size of Clade C precluded any λ = 0.024,P< 0.001), and the corresponding one-way robust statistical comparisons of morphometric data ANOVAs conducted for each ratio independently were with the remaining four genetic clades, but it appears also significantly different amongst clades (P < 0.001), to be most similar to Clades A and B with regard to even after strict Bonferroni correction for multiple tests. the ratios of pectoral (P1A/P1I) and dorsal fins (D1H/ From the initial set of six candidate morphometric ratios, D2H), as well as caudal peduncle length (DCS/CDM) D1H/P1I showed strong colinearity with P1A/P1I and (Figs 3, 9). However, two ratios may be potentially useful D1H/D2H and was removed from further analyses in discriminating specimens in Clade C from those in because of its correspondingly smaller F-value, which Clades A and B (Fig. 3), namely a combination of short

© 2014 The Linnean Society of London, Zoological Journal of the Linnean Society, 2014, 172, 803–830 812 .VERÍSSIMO A. 1.6 04TeLnenSceyo London, of Society Linnean The 2014 © 0.50 0.52 TAL. ET 1.4 0.48 0.46 1.2 P1A/P1I PN/POR 0.44 POB/HDL 0.40 0.42 0.24 0.26 0.28 0.30 0.32 0.34 0.36 0.38 0.8 1.0

ABCDEFGH ABCDEFGH ABCDEFGH 2014, Society , Linnean the of Journal Zoological 0.45 1.4 1.2 D1H/P1I D1H/D2H DCS/CDM 0.35 0.40 0.30 0.8 1.0 0.4 0.5 0.6 0.7 0.8 ABCDEFGH ABCDEFGH ABCDEFGH

Figure 3. Box-plots of the six candidate morphometric ratios. Whiskers’ range: minimum to maximum; box range: first to third quartile; horizontal line: median. Box width is proportional to the sample size of each clade. Clades identified in the North Atlantic are identified from A−H (see text for details). See Table 1 for abbreviations.

803–830 172, Downloaded from https://academic.oup.com/zoolinnean/article-abstract/172/4/803/2433424 by VIRGINIA INSTITUTE OF Marine Science user on 30 October 2018 October 30 on user Science Marine OF INSTITUTE VIRGINIA by https://academic.oup.com/zoolinnean/article-abstract/172/4/803/2433424 from Downloaded DIVERSITY OF NORTH ATLANTIC CENTROPHORUS 813

Table 3. Discriminant function analyses (DFAs) using all ﬁve morphometric ratios (DFA1), and using only four morphometric ratios (DFA2 without PN/POR; DFA3 without POB/HDL). Amount of variance in data not explained by each DFA is in- λ

dicated by Wilks’ . Percentage of group separation provided by each linear discriminant function (LD1−LD3) is also Downloaded from https://academic.oup.com/zoolinnean/article-abstract/172/4/803/2433424 by VIRGINIA INSTITUTE OF Marine Science user on 30 October 2018 indicated. All values were statistically signiﬁcant (P < 0.001)

Wilks’ λχ2 LD1 % F statistics LD2 % F statistics LD3 % F statistics

DFA 1 0.0249 249.3 76.4 185.21 19.8 48.06 3.8 9.17 DFA 2 0.0292 238.4 79.5 184.53 16.8 38.97 3.7 8.52 DFA 3 0.0292 238.5 78.8 182.90 17.9 41.47 3.3 7.61

HDL, distance from snout tip to ﬁfth gill slit; PN, distance from snout tip to anterior margin of nostrils; POB, distance from snout tip to anterior margin of eyes; POR, distance from snout tip to anterior mouth opening.

Table 4. Loadings of morphometric ratios (Ratios) on each Table 5. Classiﬁcation precision of specimens in each clade of the linear discriminant functions (LD1−LD3) for each using the three discriminant function analyses (DFAs) of the three discriminant function analyses (DFA1: using all ﬁve morphometric ratios; DFA2 without PN/POR; DFA3 Clade Clade Clade Clade without POB/HDL) A B D E

Ratios LD1 LD2 LD3 DFA 1 Clade A 16 2 DFA 1 PN/POR −0.072 −0.503 0.269 Clade B 1 10 POB/HDL −0.124 −0.404 0.403 Clade D 1 29 P1A/P1I 0.759 0.564 0.477 Clade E 13 D1H/D2H −0.601 0.671 0.341 Classification 89 83 100 100 DCS/CDM −0.199 −0.176 0.594 precision (%) DFA 2 POB/HDL −0.142 0.613 −0.603 DFA 2 P1A/P1I 0.734 −0.502 −0.494 Clade A 16 2 D1H/D2H −0.612 −0.697 −0.276 Clade B 1 10 DCS/CDM −0.192 0.141 −0.621 Clade D 1 29 DFA 3 PN/POR −0.107 0.685 −0.603 Clade E 13 P1A/P1I 0.755 −0.666 −0.355 Classification 89 83 100 100 D1H/D2H −0.615 −0.693 −0.267 precision (%) DCS/CDM −0.190 0.154 −0.672 DFA 3 Clade A 17 1 DFA, discriminant function analysis. Clade B 11 See Table 1 for other abbreviations. Clade D 1 29 Clade E 13 Classification 94 92 100 100 snout (PN/POR similar to Clade A) with long pre- precision (%) orbital length (POB/HDL similar to Clade B). Additional putative Centrophorus morphotypes (Morphotypes F, G, H) were apparent amongst our samples Morphotype F was distinguishable from the other North for which no genetic data were available. These Atlantic morphotypes by a combination of taller first dorsal morphotypes exhibited a distinct combination of mor- fin (D1H/D2H > 1.1) and long first dorsal fin base (D1B/ phological traits that did not match those observed in TL > 0.17) (Fig. 10; Table S3). Indeed, the other morphotypes the above-described five clades (Fig. 3). One of these pu- exhibiting taller first dorsal fins (i.e. Clades A–C) had tative morphotypes (Morphotype F; Fig. 10) corresponds comparatively shorter first dorsal fin bases (D1B/ to a group of four specimens including the syntype of TL < 0.14; Table S3). C. lusitanicus from off Portugal (an immature male of One maturing male (79.8 cm TL, USNM 205781) and 73.0 cm TL; BMNH 1867.7.23.2); a nominal C. lusitanicus two adult males (84.2 and 90.2 cm TL, USNM 206031 specimen collected off Togo (an immature female of 57.0 cm and UF 30161, respectively) collected in the Carib- TL; MNHN 1969-0225), a nominal C. lusitanicus speci- bean Sea off Panama may comprise an additional pu- men collected off Cameroon (an immature male of 39.8 cm tative Centrophorus morphotype (Morphotype G). These TL; MNHN 1969-0276), and a North Atlantic specimen specimens are distinguishable by the combination of of unknown locality catalogued as C. granulosus (a ma- a taller first dorsal fin (D1H/D2H > 1.1) and long turing female of 79.5 cm TL; MNHN 1997-0477). interdorsal distance (IDS/TL > 0.34) (Fig. 11; Table S3).

A) B) Downloaded from https://academic.oup.com/zoolinnean/article-abstract/172/4/803/2433424 by VIRGINIA INSTITUTE OF Marine Science user on 30 October 2018 LD3 (2.8%) LD2 (17.7%) -3 -2 -1 0 1 2 -4-3-2-10123

-4 -2 0 2 4 6 -1.0 -0.5 0.0 0.5 1.0 1.5 2.0 2.5

LD1 (79.5%) LD2 (17.7%)

Figure 4. Plot of discriminant function scores of A, linear discriminant function 1 (LD1) vs. LD2 for each of four clades of North Atlantic Centrophorus (i.e. Clades A, B, D, and E), and B, of LD2 vs. LD3 for Clades A and B only. Black circles, Clade A; white squares, Clade B; black triangles, Clade D; white triangles, Clade E.

Figure 5. Clade A Centrophorus uyato. A, whole body perspective and B, teeth morphology in an adult male of 91.6 cm total length (TL) from the Gulf of Mexico (VIMS13355); and C, dermal denticles of an adult male of 82.7 cm TL from off western Ireland (MNHN2003-0532).

© 2014 The Linnean Society of London, Zoological Journal of the Linnean Society, 2014, 172, 803–830 DIVERSITY OF NORTH ATLANTIC CENTROPHORUS 815 Downloaded from https://academic.oup.com/zoolinnean/article-abstract/172/4/803/2433424 by VIRGINIA INSTITUTE OF Marine Science user on 30 October 2018

Figure 6. Clade B Centrophorus sp. 1. A, whole body perspective from a maturing male of 89.1 cm total length (TL) from the Gulf of Mexico (VIMS13346); and B, teeth morphology and C, dermal denticles of an adult male of 86.3 cm TL from the Bahamas (VIMS13518).

Finally, a third putative morphotype (Morphotype medial ridge and two or more lateral ones as exhib- H) corresponds to a single adult male of 95.2 cm TL, ited by Clade E specimens (Fig. 8C). collected in the Gulf of Mexico (UF 42592; Fig. 12). This In terms of teeth morphology, the upper and lower specimen exhibits striking differences in several head teeth showed dissimilar morphologies in specimens of measurements compared with the other Centrophorus all morphotypic clades. The upper teeth were smaller, morphotypes of similar sizes described above, most triangular-shaped, and with straight cusps at the centre notably PG1TL (0.13 vs. > 0.15, respectively), result- of the jaw but becoming progressively inclined towards ing in shorter overall horizontal body distances such the mouth corners. The one exception to this general as SVL/TL (0.50 vs. > 0.57, respectively) and PD2/TL pattern is from males in Clade G, in which the upper (0.56 vs. > 0.58, respectively) (Fig. 12; Table S3). teeth were also triangular but the cusps were always Regarding the morphology of the dermal denticles straight (i.e. not pointed to mouth corners; Fig. 11B). of the adult specimens in each of the morphotypic clades In general, the lower teeth were larger than the upper described above, there are three main types of denti- ones, with oblique to horizontal cutting edges in all cle morphology. One type refers to block-like, non- morphotypic clades. In some males, the tips of the lower overlapping denticles that are regularly arranged, with teeth cusps were turned upwards, as observed in sessile crowns and three or more low dorsal ridges, Morphotypes G and H (Figs 11B, 12B), and also in one as exhibited by Clades A−C and F−H (Figs 5C, 6C, 9C, adult male from Clade A (VIMS 13354) and another 10C, 11C, 12C). The number of dorsal ridges in this from Clade E (Fig. 8B). denticle type was highly variable within individuals, regardless of clade. A second type of denticle morphol- DISCUSSION ogy refers to granular, tear-drop shaped denticles with a rounded anterior edge and an extended posterior cusp, The alpha taxonomy within Centrophorus has histori- regularly arranged and slightly overlapping, as exhib- cally been problematic and our results highlight the ited by Clade D (Fig. 7C). A third type refers to leaf- difficulties in identifying species, as well as charac- shaped denticles, raised on pedicels, with a strong terizing species diversity on both regional and global

© 2014 The Linnean Society of London, Zoological Journal of the Linnean Society, 2014, 172, 803–830 816 A. VERÍSSIMO ET AL. Downloaded from https://academic.oup.com/zoolinnean/article-abstract/172/4/803/2433424 by VIRGINIA INSTITUTE OF Marine Science user on 30 October 2018

Figure 7. Clade D Centrophorus granulosus. A, whole body perspective and B, teeth morphology of a juvenile male of 92.7 cm total length (TL) from the Gulf of Mexico (VIMS13309); and C, dermal denticles of an adult male of 106.0 cm TL from off the Seychelles (MNHN2005-1010). scales. We found multiple instances where a single lar and morphological approach. The specimens of species name was used to identify multiple taxa, as Centrophorus sampled in the North Atlantic and ex- well as cases of individuals belonging to the same taxon amined here were consistently placed into five dis- being identified as separate nominal species. Our results tinct genetic clades by the COI and 16S molecular indicate that species diversity within the genus as well markers (Fig. 2), which corresponded to five distin- as the alpha diversity in many regions of the world guishable morphotypes (A–E). In addition to these, three may be falsely inflated because of the use of multiple putative morphotypes (lacking genetic data) were im- species names to designate the same taxonomic entity, plicit in our morphological data (Clades F, G, and H). as already noted in previous studies (Muñoz-Chapuli The correspondence of each morphotype to previous- & Ramos, 1989; White et al., 2013). In the present study, ly described species is discussed below. specimens originally identified as ten different Centrophorus species yielded only eight distinct genetic Clade A – Centrophorus cf. uyato clades when analysing the mitochondrial COI gene Clade A refers to medium-sized gulper sharks of about region. Concordant results were obtained using the 16S 120 cm TL maximum size, characterized by a short gene region, i.e. specimens labelled under 11 differ- snout (PN/POR < 0.45 and POB/HDL ≤ 0.33), first dorsal ent taxonomic names clustered into only nine genetic fin taller than second (D1H/D2H > 1.0), pectoral inner clades, two of which were composed of unidentified margin extended posteriorly in adults (P1A/P1I < 1.14), Centrophorus specimens. short caudal peduncle (PCA/TL < 0.16), and flat dermal denticles in adult specimens with five to six longitu- dinal ridges (Fig. 5). This is a globally distributed species CENTROPHORUS SPECIES DIVERSITY IN THE that historically has been referred to as C. granulosus NORTH ATLANTIC REGION or C. uyato in the Atlantic and Indian Oceans, as well In line with several recent efforts to clarify the tax- as in the Mediterranean Sea, and called C. uyato or onomy of the genus Centrophorus worldwide, our work C. zeehaani in the Pacific Ocean off Australia (Table S1). provides a reassessment of the species diversity in the The comparison of life history parameters estimat- North Atlantic sensu lato using a combined molecu- ed for this species under the different taxon names

© 2014 The Linnean Society of London, Zoological Journal of the Linnean Society, 2014, 172, 803–830 DIVERSITY OF NORTH ATLANTIC CENTROPHORUS 817 Downloaded from https://academic.oup.com/zoolinnean/article-abstract/172/4/803/2433424 by VIRGINIA INSTITUTE OF Marine Science user on 30 October 2018

Figure 8. Clade E Centrophorus squamosus. A, whole body perspective and B, teeth morphology of an adult male of 110.4 cm total length (TL) from off Portugal (MNHNC MB06-2448); and C, dermal denticles of an adult male of 99.3 cm TL from off France (Atlantic coast; MNHN2011-0883). shows great similarity amongst the geographical regions characteristic and unique leaf-shaped dermal denticles where it has been reported (Table 6). Sexual matura- of the latter species compared with the much smaller, tion is reached around 80 cm TL in males and about tear-drop shaped denticles in the former. A subse- 90 cm TL in females, and fecundity is one embryo per quent redescription of C. granulosus by Müller & Henle gestation cycle – one of the lowest amongst Centrophorus (1841) obfuscated the issue by depicting a c. 80 cm (see Table 6 for references) and also for elasmobranchs TL specimen (from off Sicily in the Mediterranean in general. Sea) representing a distinctly different morphotype, This species has generated much confusion in that corresponds well with our Clade A specimens. the literature owing mainly to the lack of holo- Indeed, a recent work by White et al. (2013) demon- types for C. granulosus and C. uyato, as well as to the strated that the taxon name C. granulosus refers to ambiguity in the original species descriptions and a morphotype distinct from that described by Müller the poorly chosen diagnostic characters. A full revi- & Henle (1841). sion of this long-standing taxonomic issue is beyond The taxon name C. uyato derives from earlier de- the scope of the present paper. Below we present a brief scriptions of Mediterranean squaloid sharks taking summary of the main nomenclatural problems sur- as a nomenclatural reference Rafinesque’s (1810) rounding this taxon with the purpose of providing a Squalus uyato but probably based on the subsequent general background to the reader, and to justify our and much more detailed description of Bonaparte’s proposed and tentative designation of Clade A as C. cf. (1834) Spinax uyatus. Rafinesque’s (1810) descrip- uyato. tion has been historically contentious because it Centrophorus granulosus (Bloch & Schneider, 1801) does not present features diagnostic for Centrophorus, refers to a large Centrophorus species reaching 150 cm and instead includes features of both Centrophorus TL as stated in the original species description. It is and Squalus, namely anterior dorsal fin spines joined clearly distinct from a similarly large species report- to the fin for a third of their length, large, oblong ed by Bloch & Schneider (1801) as Squalus squamosus eyes located above a small mouth, teeth small and Bonnaterre, 1788 (now C. squamosus) owing to the sharp, gill slits narrow with the fifth slit being the

© 2014 The Linnean Society of London, Zoological Journal of the Linnean Society, 2014, 172, 803–830 818 A. VERÍSSIMO ET AL. Downloaded from https://academic.oup.com/zoolinnean/article-abstract/172/4/803/2433424 by VIRGINIA INSTITUTE OF Marine Science user on 30 October 2018

Figure 9. Clade C Centrophorus isodon. A, whole body perspective; B, teeth morphology; and C, dermal denticles of a maturing male of 87.1 cm total length from the Gulf of Mexico (VIMS13308). longest (Rafinesque, 1810). Bonaparte (1834) de- ance, and biological parameters provided in White et al. scribed Sp. uyatus in great detail based on a 2-ft (2008) for C. zeehaani match those presented here (c. 61 cm) specimen, referring to it as the same taxon for Clade A, as well as previously published data previously described as S. uyatus Rafinesque, 1810, for Atlantic C. granulosus/C. uyato (see Table 6 for and clearly distinguished it from Squalus by noting references). the difference in the upper and lower teeth morphol- Our Clade A specimens also correspond to the Medi- ogy, and in the caudal fin shape. terranean Centrophorus described by Bonaparte (1834), The descriptions of Sp. uyatus by Bonaparte (1834) and to the recently described C. zeehaani from Aus- and of C. granulosus by Müller & Henle (1841), both tralia (White et al., 2008); thus, a single species name based on Mediterranean specimens, correspond well is needed to reduce this ambiguity. Centrophorus with our Clade A specimens, whereas there are strong granulosus refers to a distinct morphotype as de- doubts about the identity of Squalus uyato Rafinesque scribed in White et al. (2013), and corresponds to our (1810) as also noted by several other authors (Maurin, Clade D, and thus should not be used to designate speci- 1968; Tortonese, 1969; Cadenat & Blache, 1981; Bass mens bearing the Clade A morphotype. Until the no- et al., 1986). Notwithstanding the confused nomen- menclatural problems associated with Clade A are clatural history described above, another nominal species resolved following the recommendations of the Inter- has been found to correspond to our Clade A, i.e. national Code of Zoological Nomenclature, we retain C. zeehaani. Our molecular data clearly show that for this clade the name C. cf. uyato. C. zeehaani, originally described as an Australian endemic (White et al., 2008), is the same species des- Clade B – Centrophorus Species 1 ignated as C. uyato and/or C. granulosus in the At- Clade B refers to another medium-sized gulper shark lantic Ocean. Our data include a COI sequence from that reaches at least 107 cm TL, and to date has been one of the specimens used in the original species de- captured only in the subtropical western Atlantic, scription (CSIRO H6309-01, GenBank accession no. namely in the Gulf of Mexico, the Bahamas, and EU398660), which clusters with our Clade A speci- Jamaica. This species is morphologically similar to our mens from the North Atlantic and elsewhere. The Clade A specimens (Fig. 6), i.e. C. cf. uyato, a fact that comparison of body measurements, general appear- has undoubtedly contributed to its possibly undescribed

© 2014 The Linnean Society of London, Zoological Journal of the Linnean Society, 2014, 172, 803–830 DIVERSITY OF NORTH ATLANTIC CENTROPHORUS 819 Downloaded from https://academic.oup.com/zoolinnean/article-abstract/172/4/803/2433424 by VIRGINIA INSTITUTE OF Marine Science user on 30 October 2018

Figure 10. Clade F Centrophorus lusitanicus. A, whole body perspective and B, teeth morphology an immature male of 73 cm total length (TL) from off Portugal (syntype of C. lusitanicus BMNH 1867.7.23.2), and C, dermal denticles of a maturing male 79.5 cm TL from the North Atlantic (MNHN1997-0477). status or misidentification. Indeed, some previous studies specimens were also included amongst our samples, have referred to this species as C. cf. uyato, reflect- and fell within our Clade B. Our taxon coverage is ing its taxonomic uncertainty (McLaughlin & Morrissey, similar to that of Naylor et al. (2012) and both studies 2005). Our data indicate that Clade B may be distin- support the distinctiveness of Clade B from the re- guished from Clade A mainly by its slightly longer snout mainder of the Centrophorus taxa examined in the (PN/POR > 0.40 but more often ≥ 0.45, vs. < 0.45 in present study. However, both the present study and other clades). Some of our Clade B specimens (N =6; Naylor et al. (2012) used a genetic data matrix without MML, uncatalogued) were examined by Castro (2011), representatives from some currently recognized species who recognized these as a distinct, undescribed species of Centrophorus, i.e. C. seychellorum Baranes, 2003, of Centrophorus (namely Centrophorus sp. B, or slender C. robustus Deng, Xiong & Zhan, 1985, C. tesselatus, gulper). The limited biological data collected thus far and C. westraliensis. Nevertheless, specimens from most indicate that the species reaches sexual maturity over of the above species exhibit distinct morphological dif- 80 cm TL in males and over 90 cm TL in females, and ferences relative to Clade B specimens. For instance, fecundity ranges from one to two embryos per gesta- according to the original description of C. seychellorum tion cycle (McLaughlin & Morrissey, 2005; this study). (Baranes, 2003), the teeth morphology in this species These biological parameters match those presented by is very distinct from that of Clade B: both males and Sang (1991) for a taxon tentatively identified as females of C. seychellorum have similar-shaped teeth C. granulosus from off Puerto Rico, which thus might on both jaws, whereas in Clade B specimens the teeth correspond to our Clade B taxon. in the upper and lower jaws are distinct (Fig. 5). In Comparison with material from other Centrophorus addition, C. seychellorum has a longer snout and head specimens collected globally showed Clade B to be ge- dimensions than Clade B specimens of similar size netically distinct from all others included in the present (POR/TL: 0.12 vs. 0.10; HDL/TL: 0.25 vs. 0.21, respec- analysis. Naylor et al. (2012) also found a distinct clade tively). Centrophorus robustus has recently been of Caribbean Centrophorus (Centrophorus sp. 1) cor- synonymized to C. granulosus by White et al. (2013) responding to two specimens sampled off Jamaica and, as such, Clade B can be distinguished from this (GN1965 and GN1966). Tissue samples from these two species by its unequal dorsal fin heights (vs. roughly

© 2014 The Linnean Society of London, Zoological Journal of the Linnean Society, 2014, 172, 803–830 820 A. VERÍSSIMO ET AL. Downloaded from https://academic.oup.com/zoolinnean/article-abstract/172/4/803/2433424 by VIRGINIA INSTITUTE OF Marine Science user on 30 October 2018

Figure 11. Clade G Centrophorus sp. 2. A, whole body perspective; B, teeth morphology; and C, dermal denticles of an adult male 90.2 cm total length from off Panama (UF 30161).

equal heights in C. granulosus). The head morphometrics Clade C – Centrophorus isodon (Chu et al., 1981) of C. westraliensis indicate very distinct differences Clade C is derived from only a few specimens collect- between this species and Clade B specimens, with ed on both sides of the North Atlantic for which limited the former showing longer snout and head propor- morphological and biological data were available. These tions compared with the latter (e.g. POR: 10.9–12.4 specimens clustered either with C. isodon (COI) from vs. 8.3–9.5% TL; PSP: 14.1–14.8 vs. 11.1–13.2% TL; the Indo-Pacific, or with both C. isodon and C. harrissoni HDL: 23.8–24.7 vs. 16.7–22.1% TL, respectively; data (16S) from the Indo-Pacific and Australia, respective- taken from White et al., 2008). The comparison of Clade ly. The few morphometric data available for the Clade B specimens with the C. tesselatus holotype (MCZ 1031) C specimens examined here generally match those from off Japan shows great similarity in several body provided in White et al. (2008) for C. isodon (account- proportions between the two species, such as IDS/ ing for slight between-study differences in measure- TL, DCS/TL, and D1B/TL, amongst others (see Table S3 ment protocols). Thus, we consider our Clade C for more details). Thus, in the absence of genetic data specimens as C. isodon, pending further taxonomic work for specimens of C. tesselatus, we cannot discard the from the Atlantic and from the Indo-Pacific region hypothesis that these might be the same species (type locality of the species: South China Sea; Chu as our Clade B. However, caution should be exer- et al., 1981). cised when considering this hypothesis given the great Owing to the small number of measured speci- morphological similarity also seen between Clade B and mens, it is not possible at this point to indicate un- Clade A. ambiguous morphometric differences between Clade Further taxonomic work should be conducted to better C and two other North Atlantic morphotypes corre- ascertain the status of the Clade B lineage. More- sponding to Clades A and B. The combination of a over, although this species has been reported from a short snout (PN/POR: 0.41–0.43) and long pre-orbital very restricted region of the western Atlantic, it is likely distance (POB/HDL: 0.30–0.33) with a taller first dorsal that future surveys will extend its distribution range fin (D1H/D2H > 1.1) are thus merely indicative at as other Centrophorus species of similar size have wide present, and refer exclusively to immature and ma- geographical distributions (e.g. C. cf. uyato). turing specimens. Additionally, Clade C specimens have

© 2014 The Linnean Society of London, Zoological Journal of the Linnean Society, 2014, 172, 803–830 DIVERSITY OF NORTH ATLANTIC CENTROPHORUS 821 Downloaded from https://academic.oup.com/zoolinnean/article-abstract/172/4/803/2433424 by VIRGINIA INSTITUTE OF Marine Science user on 30 October 2018

Figure 12. Clade H Centrophorus sp. 3. A, whole body perspective, B, teeth morphology; and C, dermal denticles of an adult male 95.2 cm total length from the Gulf of Mexico (UF 42592).

small upper jaw teeth, turned towards the mouth Clade D – Centrophorus granulosus (Bloch & corners in immature specimens (Fig. 9), similar to those Schneider, 1801) described for juvenile specimens of C. uyato (Garman, Clade D refers to a large-sized gulper shark reaching 1906, 1913; Bigelow et al., 1953; Ledoux, 1970; Guallart, well over 150 cm TL (e.g. 176 cm TL; Cotton, 2010) and 1998; Guallart et al., 2013). with a wide geographical distribution including the Castro (2011) reported the presence of C. isodon Atlantic, Indian, and Pacific Oceans. This species is off the Bahamas, and our study shows its distribu- characterized by roughly equal dorsal fin heights tion to include both the western and eastern (D1H ≈ D2H), moderately extended pectoral inner Atlantic Ocean. These data also suggest that C. isodon margins, and teardrop-shaped dermal denticles in the has a wider geographical distribution than previous- adults (Fig. 7). This dermal denticle morphology gives ly reported, including (at least) the Indo-west the skin a granular feel, as reflected in its species name Pacific (South China Sea off the Philippines, and (from the Latin ‘granulosa’). Records of C. lusitanicus Indonesia, Chu et al., 1981; White et al., 2006), and from off South Africa and Mozambique made by Bass both sides of the North Atlantic (present study). et al. 1976, 1986) are in fact of C. granulosus, given Thus, it is plausible that this species also occurs the matching morphometric characters (‘first dorsal fin in the South Atlantic and west-central Indian Ocean. lower than the second dorsal’, ‘inner corner [of pecto- Cadenat & Blache (1981) reported an uniden- rals] being elongated to a point behind the exposed tified Centrophorus off Senegal very similar to their portion of the first dorsal spine’, Bass et al. 1976: 28) C. uyato-machiquensis (our Clade A), characterized and large maximum sizes (up to 160 cm TL, Bass et al., by a short first dorsal fin base and first dorsal fin 1976). In addition, Gubanov (1986) reported the pres- taller than second but distinguishable from it by its ence of C. tesselatus from the Indian Ocean on the Mada- longer snout. This suite of characters fits well with gascar Ridge although the description, morphometrics, our Clade C morphotype and may refer to the same and biological information collected on the specimens species. are also in line with those described for Clade D.

04TeLnenSceyo London, of Society Linnean The 2014 © Table 6. Summary of biological data available for all Centrophorus taxa studied so far (not all species described to date are included) TAL. ET Minimum size at Maximum maturity (mm) L50 (mm) Size at Original species size Uterine birth Revised species designation designation Region (mm) MF MFfecundity (mm) References

Centrophorus atromarginatus C. atromarginatus Indo-Pacific 987 580 735 n.d. n.d. 1 ∼361 White & Dharmadi, 2010 Centrophorus lusitanicus C. cf. lusitanicus Indo-Pacific 930 679 873 n.d. n.d. 1 n.d. White & Dharmadi, 2010 C. lusitanicus C. lusitanicus Western North Atlantic 1140 715 850 n.d. n.d. 1 > 305 Cadenat & Blache, 1981 Centrophorus granulosus Centrophorus niaukang Western North Atlantic 1690 1410 1500 n.d. n.d. ≤ 7 400–415 Cotton, 2010 C. granulosus C. granulosus Eastern North Atlantic 1660 1180 1380 n.d. 1470 ≤ 6 350–470 Bañón, Piñeiro & Casas, 2008 C. granulosus Centrophorus tesselatus Western Indian 1730 n.d. 1610 n.d. n.d. ≤ 7 n.d. Gubanov, 1986 Centrophorus harrissoni C. harrissoni Western South Pacific 1120 810 960 840 990 ≤ 2 ∼380 Graham & Daley, 2011 Centrophorus isodon C. isodon Indo-Pacific 1082 808 999 n.d. n.d. 2 ∼315 White & Dharmadi, 2010

2014, Society , Linnean the of Journal Zoological C. cf. isodon C. cf. isodon Gulf of Mexico 871 n.d. n.d. n.d. n.d. n.d. n.d. This study Centrophorus sp. 1 Centrophorus sp. Gulf of Mexico 1071 854 976 n.d. n.d. n.d. n.d. This study Centrophorus sp. 1 Centrophorus cf. uyato Caribbean Sea 1023 812 915 n.d. n.d. ≤ 2 > 350 McLaughlin & Morrissey, 2005 Centrophorus moluccensis C. moluccensis Indo-Pacific 937 767 909 n.d. n.d. 1 ∼330 White & Dharmadi, 2010 C. moluccensis C. moluccensis Western South Pacific 980 670 850 698 880 ≤ 2 ∼350 Graham & Daley, 2011 Centrophorus squamosus C. squamosus Indo-Pacific 1640 900 n.d. n.d. n.d. n.d. n.d. White & Dharmadi, 2010 C. squamosus C. squamosus Eastern North Atlantic 1440 n.d. n.d. 991 1263 ≤ 6 ∼440 Figueiredo et al., 2008 C. squamosus C. squamosus Eastern North Atlantic 1450 n.d. n.d. 864 1050 n.d. n.d. Clarke, Connolly & Bracken, 2001 C. squamosus C. squamosus Eastern North Atlantic 1460 960 1161 n.d. n.d. ≤ 10 > 380 Severino et al., 2009 Centrophorus uyato C. granulosus Mediterranean 1125 790 893 799 935 1 343–461 Guallart, 1998 C. uyato C. granulosus Mediterranean 1280 800 940 800 890 1 330–350 Capapé, 1985 C. uyato C. granulosus Mediterranean 950 745 850 n.d. n.d. 1 n.d. Megalofonou & Chatzispyrou, 2006 C. uyato C. granulosus Gulf of Mexico 1015 825 1015 n.d. n.d. 1 n.d. This study C. uyato Centrophorus zeehaani Western South Pacific 1110 790 960 785 957 1 ∼450 Graham & Daley, 2011

F, female; L50, size at 50% maturity; M, male; n.d., not determined;

This species reaches sexual maturity at a larger size Mate, 2002). In our study, a single tissue sample without and has a higher fecundity per gestation cycle than additional specimen data or a photo voucher is the only

most Centrophorus species, exceeding all but the simi- evidence suggesting the presence of C. squamosus in Downloaded from https://academic.oup.com/zoolinnean/article-abstract/172/4/803/2433424 by VIRGINIA INSTITUTE OF Marine Science user on 30 October 2018 larly large C. squamosus (Table 6). Males approach ma- this area. Historically intensive fishing and many ex- turity over 100 cm TL whereas females are sexually ploratory research surveys of deep waters have been mature over ∼140 cm TL (see Table 4 for references). carried out in the Mediterranean, consistently record- Centrophorus granulosus was originally described by ing the absence of C. squamosus. Thus, this species Bloch & Schneider (1801) as Squalus granulosus based may only occasionally enter the Mediterranean on a specimen measuring 5 ft, i.e. close to 150 cm. Of Sea. all Centrophorus species described to date, only four Centrophorus squamosus is one of the most studied have been reported to reach ∼150 cm in length: C. acus species of gulper sharks given its importance as a fish- Garman, 1906, C. granulosus (Bloch & Schneider, 1801), eries resource in parts of its distribution, most notably C. niaukang Teng, 1959, and C. squamosus (Bonnaterre, in the eastern North Atlantic (e.g. see ICES, 2012). 1788) (Bocage & Capello, 1866; Yano & Kugai, 1993; As in C. granulosus, sexual maturation of C. squamosus Compagno et al., 2005; Castro, 2011; Ebert & Stehmann, occurs at a large size of around 90 cm TL in males 2012). Centrophorus squamosus has large, overlap- and 116 cm TL in females, and fecundity can be up ping, leaf-shaped dermal denticles that are unique to ten pups per litter (see Table 6 for references). amongst Centrophorus and that make this species easily distinguishable from all other gulper sharks regard- Clade F – Centrophorus lusitanicus Bocage & less of size. Using molecular and morphological data, Capello, 1864 White et al. (2013) demonstrated that C. acus, Clade F is derived from three immature and one ma- C. granulosus, and C. niaukang are conspecific. Our data turing specimens characterized by having the first dorsal are in agreement with that study and support the use fin taller than the second (D1H > D2H), a markedly of the earliest described species name – C. granulosus long first dorsal fin base (DIB/TL > 0.17), and a short (Bloch & Schneider, 1801) – for this clade. caudal peduncle (PCA/TL: 0.12). Specimens in this clade were collected at three eastern North Atlantic loca- Clade E – Centrophorus squamosus tions and include the syntype of C. lusitanicus (Bocage (Bonnaterre, 1788) & Capello, 1864; Fig. 10). This nominal species has been Clade E specimens belong to another large-sized reported most commonly from off the western African Centrophorus reaching at least 164 cm TL (White & coast (Morocco south to Senegal: Cadenat, 1959b; Dharmadi, 2010), most commonly reported in the Cadenat & Blache, 1981; Muñoz-Chapuli & Ramos, eastern Atlantic and along the mid-Atlantic ridge, but 1989), with additional records from off Madagascar in also found in the Indian and Pacific Oceans (Compagno the western Indian Ocean (Naylor et al., 2012) and off et al., 2005). This species also occurs in the western Indonesia in the Indo-Pacific region (White et al., 2006; North Atlantic (Castro, 2011; Naylor et al., 2012; C. White & Dharmadi, 2010). As noted above, the pres- Cotton, pers. observ.), suggesting a wider Atlantic dis- ence of this nominal species in South African waters tribution than previously reported (Compagno, 1984; reported by Bass et al. (1976, 1986) refers in fact to Compagno et al., 2005). Centrophorus squamosus is well our Clade D, C. granulosus. characterized by roughly equal dorsal fin heights Bocage & Capello (1864) described C. lusitanicus from (D1H ≈ D2H), short pectoral fins (P1A/P1I > 1.3) with specimens captured off Portugal. However, two years a rounded posterior margin, a short caudal peduncle later (Bocage & Capello, 1866), they revised their pre- (DCS/CDM < 0.37), and large, overlapping, leaf- vious work and declared the species invalid, stating shaped dermal denticles in the adults (Fig. 8). The latter that the C. lusitanicus type specimens that they ex- character is unique amongst Centrophorus and diag- amined were in fact C. granulosus of different sizes and nostic for this species, which may explain the consist- collected in different seasons. Their type specimens were ently accurate identification throughout its worldwide originally deposited in the National Museum of Natural distribution. History and Science (MUHNAC) in Portugal (Bocage The presence of C. squamosus in the Mediterra- & Capello, 1866), with one specimen being sent to the nean Sea suggested by a single specimen included in Museum of Natural History, London (syntype our study has also been reported by previous authors BMNH1867.7.23.2). However, those remaining in Por- (Lozano-Rey, 1928; Lozano-Cabo, 1963; Lloris et al., 1984; tugal were lost in a fire in 1978. No other C. lusitanicus Muñoz-Chapuli, 1990; Barrull & Mate, 1996). However, specimens remained in the Portuguese collection, and none of these historical records were based on direct no other specimens have been deposited at MUHNAC observation of specimens but were rather the result since 1978 even though the type locality of C. lusitanicus of misspellings of species names, or to incorrect local- is Sesimbra, Portugal. Likewise, no C. lusitanicus speci- ity data of true C. squamosus specimens (Barrull & mens have been observed during intensive fish market

© 2014 The Linnean Society of London, Zoological Journal of the Linnean Society, 2014, 172, 803–830 824 A. VERÍSSIMO ET AL. sampling campaigns in Sesimbra ﬁshing port, the most its strikingly different body proportions from all other important regional port recording deep-water shark land- specimens examined in the present study (Fig. 12). The

ings (Figueiredo, Machado & Gordo, 2005) between 1998 head length is smaller (15.8% TL) and narrower (INO/ Downloaded from https://academic.oup.com/zoolinnean/article-abstract/172/4/803/2433424 by VIRGINIA INSTITUTE OF Marine Science user on 30 October 2018 and 2001 (A. Veríssimo, pers. observ.). It is likely TL: 0.06, MOW/TL: 0.06), leading to several other body that the only remaining original material from measurements being smaller (SVL, PD1, PD2; see Bocage & Capello’s (1864) description of C. lusitanicus Table 1 for abbreviations) when specimens of similar is deposited at the British Natural History Museum, size were compared. More specimens and further study which includes the syntype examined in the present are required to clarify the validity of this morphotype. study (BMNH1867.7.23.2). Our results demonstrate higher diversity of Bocage & Capello’s (1866) doubt about the validity Centrophorus than previously has been documented in of C. lusitanicus did not eliminate the usage of this the North Atlantic, including the existence of three po- species name. Most authors providing data about tentially undescribed species. We detected all four species specimens of this nominal species agree that the main previously reported from the north-eastern Atlantic: distinguishing feature is the very long first dorsal C. granulosus (also recorded as C. niaukang), fin base (> 17% TL), in combination with a taller first C. lusitanicus, C. squamosus, and C. uyato (also re- dorsal fin (D1H > D2H) (Cadenat & Blache, 1981; ported as C. granulosus) (Krefft & Tortonese, 1973; Muñoz-Chapuli & Ramos, 1989). Indeed, Muñoz-Chapuli Cadenat & Blache, 1981; Muñoz-Chapuli & Ramos, & Ramos (1989) provided not only biometric but also 1989; Compagno et al., 2005; Ebert & Stehmann, 2012), meristic characters (e.g. number of dorsal radials) to plus C. isodon (from off Cape Verde). All but the latter discriminate it from C. granulosus (our Clade A), were noted by Cadenat & Blache (1981) as occurring C. niaukang (our Clade D), and C. squamosus. in the eastern Atlantic and the unidentified species that In a recent study by Naylor et al. (2012), genetic they reported from off Senegal appears to correspond analysis of specimens collected off Madagascar, exhib- to our C. isodon. iting our Clade F morphotype and identified as C. cf. Our sampling efforts and data confirm the pres- lusitanicus, clustered together in a clade separate from ence of a single Centrophorus species bearing flat dermal their other Centrophorus samples collected world- denticles in the Mediterranean Sea, as noted by many wide, thus supporting its validity as a distinct previous authors (Maurin, 1968; Tortonese, 1969; Maurin Centrophorus lineage. Our own 16S tree shows our & Bonnet, 1970; Capapé, 1985; Muñoz-Chapuli & single C. lusitanicus sequence as a separate branch from Ramos, 1989; Guallart, 1998; Guallart et al., 2006). We the remainder of our Centrophorus clades. Thus, we provisionally refer this morph, our Clade A, to C. cf. refer Clade F specimens to C. lusitanicus pending further uyato pending future nomenclatural decisions. The pres- taxonomic revision. ence of C. squamosus in Mediterranean waters remains to be confirmed by further sampling but the evidence Clade G – Centrophorus Species 2 so far suggests this species as an occasional visitor to Clade G is based on a distinct morphotype exhibited the region. by three male specimens collected in the Caribbean Species diversity appears to be higher in the north- Sea off Panama. These are distinguishable from our western Atlantic than in the north-eastern Atlantic, other Centrophorus morphotypes in having a first dorsal with six or seven species historically reported from the fin taller than second (D1H > D2H) and a long former region. Bigelow et al. (1953) reported the pres- interdorsal distance (IDS/TL > 0.34). This clade (Fig. 11) ence of C. uyato in the Gulf of Mexico, whereas was identified as C. tesselatus by Castro (2011), who Compagno et al. (2005) listed both C. uyato (as examined one of the specimens included in our Clade C. granulosus) and C. tesselatus, and suggested the pres- G (UF 30161, erroneously labelled as UF 3161 in Castro, ence of a long-snouted Centrophorus in that region. A 2011). However, the comparison of body proportions previous report of the presence of C. granulosus in the between Clade G and the holotype of C. tesselatus (MCZ western North Atlantic by Moore et al. (2003; as 1031), most notably IDS/TL > 0.34 vs. 0.29, respec- C. niaukang) was confirmed in our study as we ex- tively, does not support this designation (see Table S3 amined a large number of specimens from off the US for more details). The validity of this clade needs to east coast, north of Florida. More recently, Castro (2011) be further examined with a larger sample size, using indicated the presence of seven species in the western both molecular and morphological data, in order to North Atlantic: C. isodon, C. niaukang (Clade D in perform robust comparisons with other Centrophorus the present study, C. granulosus), C. squamosus, taxa. C. tesselatus, C. uyato, and two unidentified Centrophorus species, i.e. Centrophorus spp. A and B, the latter cor- Clade H – Centrophorus Species 3 responding to Clade B in the present study. This clade is based on a single specimen (UF 42592) Centrophorus sp. A (referred to as ‘Minigulper’; Castro, taken in the Gulf of Mexico off Florida, which exhib- 2011) was not examined by us and its status remains

© 2014 The Linnean Society of London, Zoological Journal of the Linnean Society, 2014, 172, 803–830 DIVERSITY OF NORTH ATLANTIC CENTROPHORUS 825 uncertain. Based on our study, we recognize seven often worn or broken/damaged in both fresh and pre- species of north-west Atlantic gulper sharks, six of which served specimens of other species. Similarly, the ﬁrst

we examined here: C. granulosus, C. isodon, C. cf. uyato, dorsal fin base relative to total length has frequently Downloaded from https://academic.oup.com/zoolinnean/article-abstract/172/4/803/2433424 by VIRGINIA INSTITUTE OF Marine Science user on 30 October 2018 and three additional Centrophorus morphotypes (Clades been used to distinguish C. granulosus and C. lusitanicus B, G, and H from the Gulf of Mexico, the Bahamas, from other Centrophorus species, given its typically and/or the Caribbean Sea). The seventh species, longer base length. However, this character is highly C. squamosus, was not represented amongst our samples variable within C. granulosus (D1B/TL: 0.12–0.18) and from the north-west Atlantic but is readily identifi- may overlap with the range observed in other able and has been reported from the western North Centrophorus (e.g. C. uyato or Centrophorus sp. 1; Atlantic (Castro, 2011; C. Cotton, pers. observ.). Table S4). Much of this variability in fin base measurement probably stems from the fact that the first dorsal fin origin is difficult to distinguish in CAVEATS AND LIMITATIONS Centrophorus because of a moderately long skin fold The scarcity of available specimens historically has hin- anterior to the first dorsal spine, with highly vari- dered taxonomic investigations of many deep-water elas- able length within and amongst species. The same is mobranch taxa. The lack of adequate comparative true for the second dorsal fin, although the phenom- material of Centrophorus specimens precludes a more enon is not as pronounced. We suggest that the lengths thorough examination of the genus in the North At- of dorsal fin bases of Centrophorus (and of similar genera lantic and on a worldwide basis. Our examination of such as Deania) should be measured from the pos- recently collected gulper shark specimens has sug- terior margin of the dorsal spine, where it emerges from gested considerable ontogenetic variation within species the fin, to the dorsal fin insertion (D1B’; Fig. S1E). Using in many morphological characters, e.g. dermal denti- such a conspicuous landmark will probably reduce the cle morphology, pectoral fin shape, and dorsal fin spine variability in this measurement and may increase the morphology, supporting the findings of previous studies value of D1B/TL as a diagnostic character. (Cadenat, 1959a, c; Maurin, 1968; Ledoux, 1970; Bass et al., 1976; Cadenat & Blache, 1981; Muñoz-Chapuli & Blasco, 1984; Tabit, 1993; Guallart, 1998; Guallart CONSERVATION AND MANAGEMENT IMPLICATIONS et al., 2013; White et al. 2013). Sexual dimorphism, apart Our study of North Atlantic Centrophorus has docu- from the presence of male claspers in all members of mented the occurrence of species with widespread geo- the group, has been documented for some chondrichthyans graphical distributions encompassing entire ocean basins (Kajiura & Tricas, 1996; Orlov & Cotton, 2011), in- and/or multiple oceans. This underscores the need for cluding some squaloids (G. H. Burgess, pers. observ.) a global comparative approach to new species descrip- but remains largely unexplored within Centrophorus. tions and, ultimately, a global taxonomic revision of Recent projects [e.g. Tree of Life (tolweb.org), Deep-C the genus. That said, owing to the current paucity of (deep-c.org), Census of Marine Life (www.coml.org)] have museum specimens and the difficulty encountered with greatly promoted deep-water scientific sampling and the lack of and/or poor condition of types, regional facilitated the examination of more specimens, result- studies such as those of Teng (1959), Cadenat & Blache ing in a relative flood of new information for many species. (1981), Muñoz-Chapuli & Ramos (1989), Last & Stevens Although taxonomic studies have benefitted from this (1994, 2009), Baranes (2003), White et al. (2008), and recent increase in specimens collected, a wider size/ the present study provide critical data that can be in- sex range of specimens and broader distribution in geo- tegrated to resolve the global taxonomy of gulper sharks. graphical sampling are still needed to capture the full Based on our results, regional endemicity in the genus range of ontogenetic and sexually dimorphic charac- appears to be the exception rather than the rule. A ters in gulper sharks. recent population genetic study on a globally distrib- In addition to ontogenetically variable and sexual- uted gulper shark, C. squamosus, supports the concept ly dimorphic characters, some often used ‘diagnostic’ of long-distance movements along continuous conti- characters in gulper sharks have shown considerable nental shelf habitats as well as of movements across variation at the intra- and interspecific level, reduc- ocean basins within a generation time frame (Veríssimo, ing their value for clear and unambiguous species dis- McDowell & Graves, 2012). Direct information about crimination. For example, dorsal spine height relative movement patterns of different species of Centrophorus to dorsal fin height may be useful to identify some is still limited but indicates their ability for long- gulper sharks, e.g. adult C. granulosus has very short, distance movements (i.e. hundreds of kilometres in several exposed spines compared with any of the other weeks) along slope waters (Yano & Tanaka, 1986; Centrophorus, but this character not only varies with Rodríguez-Cabello & Sánchez, 2014; Daley et al., in press). ontogeny in this species (A. Veríssimo & C. Cotton, pers. Although this should be objectively confirmed for other observ.) but is often of limited use as fin spines are congeners, the wide distributions of most gulper sharks

CENTROPHORUS SPECIES KEY FOR THE NORTH ATLANTIC 1.a. Dorsal fins of about the same height (D1H/D2H < 1.2); adults can reach sizes > 120 cm TL...... 2 Downloaded from https://academic.oup.com/zoolinnean/article-abstract/172/4/803/2433424 by VIRGINIA INSTITUTE OF Marine Science user on 30 October 2018 1.b. First dorsal fin taller than second (D1H/D2H > 1.2): adults reach sizes < 120 cm TL...... 3 2.a. Pectoral inner margin slightly elongated (P1A/P1B < 1.3) and posterior margin concave; dermal denticles teardrop- shaped in adult specimens (Fig. 7)...... Centrophorus granulosus 2.b. Pectoral inner margin short (P1A/P1B > 1.3) and posterior margin rounded; skin very rough, with large, overlapping, leaf-shaped dermal denticles raised on pedicels (Fig. 8) ...... Centrophorus squamosus 3.a. First dorsal fin base (D1B) > 17% TL (Fig. 10)...... Centrophorus lusitanicus 3.b. First dorsal fin base (D1B) < 15% TL...... 4 4.a. Interdorsal distance (IDS) > 34% TL (Fig. 11) ...... Centrophorus Species 2 4.b. Interdorsal distance (IDS) < 34% TL...... 5 5.a. Distance from snout tip to first gill slit (PG1) < 14% TL; distance from snout tip to anterior edge of vent (SVL) < 55% TL (Fig. 12)...... Centrophorus Species 3 5.b. Distance from snout tip to first gill slit (PG1) > 14% TL; distance from snout tip to anterior edge of vent (SVL) > 55% TL...... 6 6.a. Distance from snout tip to nostrils (PN) always > 0.40 of distance from snout tip to anterior edge of mouth (POR), but more often ≥ 0.45 (Fig. 6) ...... Centrophorus Species 1 6.b. Distance from snout tip to nostrils (PN) always < 0.45 of distance from snout tip to anterior edge of mouth (POR)...... 7 7.a. Distance from snout tip to anterior edge of eye (POB) ≥ 0.30 of head length (HDL); distance from pelvic insertion to lower caudal origin (PCA) > 15.5% TL (Fig. 9)...... Centrophorus isodon 7.b. Distance from snout tip to anterior edge of eye (POB) ≤ 0.33 of head length (HDL), but most often < 0.30; distance from pelvic insertion to lower caudal origin (PCA) < 16% TL (Fig. 5)...... Centrophorus uyato

and the potential for long-distance dispersal between PENDING TAXONOMIC ISSUES REGARDING NORTH distant localities should be taken into account in future ATLANTIC CENTROPHORUS conservation assessments, as these are important factors to consider for the recovery of depleted populations. In Centrophorus Species 1 in the present study is dis- the absence of rigorous, comprehensive data for species tinct from other gulper sharks reported from the North distribution, the precautionary approach to manage- Atlantic, but further taxonomic work is needed to ment and conservation should be applied. compare this with other Centrophorus occurring else- One immediate consequence of the recent studies of where, thus assessing its validity as a new species. Centrophorus diversity (White et al. 2013; present study) In doing so, all currently recognized species of is the need to re-evaluate the conservation status of all Centrophorus worldwide need to be included in the com- gulper sharks. For instance, C. acus and C. niaukang, parison to avoid species duplication. Other pending taxo- now regarded as synonyms of C. granulosus (White et al. nomic issues include determining the status of Clades 2013), are listed as ‘Near-Threatened’, whereas G and H identiﬁed in our study. Larger sample sizes C. granulosus is listed as ‘Vulnerable’ according to the are needed for a more thorough taxonomic analysis of latest assessment by the International Union for the these morphotypes to determine whether they corre- Conservation of Nature Red List of Threatened Species spond to described species that have not been docu- (iucnredlist.org, accessed on 12 April 2013). Similarly, mented from the North Atlantic, or whether these are our study suggests that C. uyato and C. zeehaani, both truly undescribed species. When identifying or poten- of which have occasionally been referred to C. granulosus, tially describing new species, particular care should represent a single valid species, C. cf. uyato. This species be taken to compare morphotypes with specimens of has a particularly conservative life history with females similar sizes and maturity stages as considerable bearing but a single pup per gestation cycle (see Table 6 ontogenetic morphological changes have been shown for details); yet, it has experienced high ﬁshing pres- to occur within Centrophorus (see Caveats and Limi- sure in parts of its range (western Mediterranean, tations). Lastly, the status of C. cf. uyato is currently Guallart, 1998; off Australia, Graham et al., 2001). There under review by the authors and other collaborators. is, therefore, some urgency in a thorough reassess- Updated information on taxonomic synonyms in use, ment of the vulnerability and conservation status of important biological parameters, geographic distribu- gulper sharks worldwide. tion, and geographic variation in usage of the species

© 2014 The Linnean Society of London, Zoological Journal of the Linnean Society, 2014, 172, 803–830 DIVERSITY OF NORTH ATLANTIC CENTROPHORUS 827 name will be presented in a forthcoming manuscript Barrull J, Mate I. 1996. Els taurons dels paisos catalans.Bar- (W. T. White , D. A. Ebert, G. P. Naylor, A. Veríssimo, celona: Editorial Portic. Barrull J, Mate I. 2002. Tiburones del Mediterráneo. Arenys J. Guallart, R. H. Buch, C. F. Cotton, G. H. Burgess, Downloaded from https://academic.oup.com/zoolinnean/article-abstract/172/4/803/2433424 by VIRGINIA INSTITUTE OF Marine Science user on 30 October 2018 S. P. Iglésias unpubl. data). All of these efforts will help de Mar: Llibreria El Set-ciències. resolve the problematic taxonomy within the genus Bass AJ, Compagno LJV, Heemstra PC. 1986. Family n°5: Centrophorus and contribute to a future global revi- Squalidae. In: Smith MM, Heemstra PC, eds. Smith’s sea sion of the gulper sharks. fishes. Berlin: Springer-Verlag, 49–62. Bass AJ, D’Aubrey J, Kistnasamy N. 1976. Sharks of the east coast of southern Africa. VI. The families Oxynotidae, ACKNOWLEDGEMENTS Squalidae, Dalatiidae and Echinorhindae. Investigational Report No. 45, Oceanographic Research Institute. Durban: We are indebted to the people (and their respective South African Association for Marine Biological Research. institutions) that contributed with samples, feedback, Bigelow HR, Schroeder WC. 1957. A study of the sharks and enthusiasm to this study: D. A. Ebert; D. Grubbs; of the suborder Squaloidea. Bulletin of the Museum of Com- D. Kerstetter; E. Brooks; G. Menezes; J. Castro; J. Kiszka; parative Zoology 117: 1–150. J. Morrissey; R. Daley; S. Tanaka; T. Daly-Engel; Capt. Bigelow HR, Schroeder WC, Springer S. 1953. New and J. Ruhle; Capt. D. Ward; Capt. A. Ros; crew of R/V little known sharks from the Atlantic and the Gulf of Mexico. Weatherbird II, R/V Suncoaster, and of Cap Prim Segón Bulletin of the Museum of Comparative Zoology 109: 211– (Xàbia). We are also grateful to the following museum 276. staff for their assistance, support, and enthusiasm: E. Bloch ME, Schneider JG. 1801. Systema ichthyologiae iconibus Hilton, Z. Nicholls, P. McGrath, S. Huber, and P. ex illustratum. Vol. 2. Berlin. Konstantinidis (VIMS Fish Collection); K. Hartel, A. Bocage JVB, Capello FB. 1864. Sur quelque espèces inédites Holmes, and A. Williston (MCZ Fish Collection); B. Séret, de Squalidae de la tribu Acanthiana, Gray, qui fréquentent Z. Gabsi, and P. Pruvost (MNHN Paris); R. Robins les côtes du Portugal. Proceedings of the Zoological Society (FLMNH); D. Catania (CAS); J. Williams (USNM); J. of London 2: 260–263. Sparks, S. Schaffer, and R. Arrundell (AMNH); M. Sabaj- Bocage JVB, Capello FB. 1866. Apontamentos para a Parez (ANSP); M. J. Alves, N. Mesquita and F. Ribeiro Icthyologia de Portugal – Peixes Plagiostomos, Primeira Parte. (MUHNAC); O. Crimmon and J. Maclaine (BMNH). Lisboa: Tipografia da Academia Real de Sciencias. We would also like to thank G. Naylor, J. Musick, N. Bonaparte CL. 1834. Iconografia della fauna italica per le Straube, and W. White for feedback and stimulating quattro classi degli animali vertebrati. Tomo III – Pesci. Roma: discussions. A special thank you is extended to A. Bentley Tipografia Salviucci. and the Tissue Collection of the Kentucky Biodiversity Bonnaterre JP. 1788. Tableau encyclopédique et méthodique Institute. This work was generously supported by the des trois règnes de la nature – Ichthyologie. Paris. following funding awards and institutions: AES 2012 Cadenat J. 1959a. Notes d’ichtyologie ouest-africaine. XXI. Le genre Atractophorus Gilchrist 1922, stade juvénile de Symposium Fund; AMNH Collection Study Grant (A. V.); Centrophorus Müller and Henle, 1837 (Sélacien Squalidae). Deep-C (http://www.deep-c.org; C. F. C.); the Guy Harvey Bulletin de l’Institut Français d’Afrique Noire (A) 21: 735– Ocean Foundation (G. H. B.); MCZ Ernst Mayr Grant 738. (A. V.); MUNHAC; the National Shark Consortium; SYN- Cadenat J. 1959b. Notes d’ichtyologie ouest-africaine. XXII. THESIS Grant (A. V); Title VII (C. F. C.); and VIMS Centrophorus lusitanicus Bocage et Capello 1864 (Sélacien travel grants, Mini-Grants, and Student Research Grants Squalidae), espèce valable différent de C. granulosus. (A. V.). A. V. was funded by the Fulbright Commission Bulletin de l’Institut Français d’Afrique Noire (A) 21: 743– PhD scholarship 2005/2006 and by Fundação para a 747. Ciência e Tecnologia (SFRH/BD/40326/2007; SFRH/ Cadenat J. 1959c. Notes d’ichtyologie ouest-africaine. XXIII. BPD/77487/2011). Sur la valeur relative de la morphologie des spicules dans la systématique du genre Centrophorus. Bulletin de l’Institut REFERENCES Français d’Afrique Noire (A) 21: 748–756. Cadenat J, Blache J. 1981. Requins de Méditerranée et Bañón R, Piñeiro C, Casas M. 2008. Biological observa- d’Atlantique (plus particulièrement de la Côte Occidentale tions on the gulper shark Centrophorus granulosus d’Afrique) – Collection Faune Tropicale, no. XXI. Paris: Edi- (Chondrichthyes: Centrophoridae) off the coast of Galicia tions de l’Office de la Recherche Scientifique et Technique (north-western Spain, eastern Atlantic). Journal of the Outre-Mer. Marine Biological Association of the United Kingdom 88: 411– Capapé C. 1985. Nouvelle description de Centrophorus 414. granulosus (Schneider, 1801) (Pisces, Squalidae). Données sur Baranes A. 2003. Sharks from the Amirantes Islands, Sey- la biologie de la reproduction et le régime alimentaire des chelles, with a description of two new species of squaloids spécimens des côtes tunisiennes. Bulletin de l’Institut Na- from the deep sea. Israel Journal of Zoology 49: 33– tional Scientifique et Technique d’Océanographie et de Pêche 65. de Salommbô 12: 97–141.

Castro JI. 2011. The sharks of North America. New York: Oxford el Mar Balear (Mediterráneo occidental). Unpublished D. Phil. University Press. Thesis, University of Valencia. Chu Y-T, Meng C-W, Liu J-X. 1981. Description of a new Guallart J, García-Salinas P, Catalán A. 2013. Morphometric Downloaded from https://academic.oup.com/zoolinnean/article-abstract/172/4/803/2433424 by VIRGINIA INSTITUTE OF Marine Science user on 30 October 2018 genus and a new species of Squalidae of China. Acta ontogenetic changes in the little gulper shark Centrophorus Zootaxonomica Sinica 6: 100–103. granulosus (Squaliformes, Centrophoridae) from the Medi- Clarke MW, Connolly PL, Bracken JJ. 2001. Aspects of re- terranean Sea: a critical review of anatomic characters used production of the deep-water sharks Centroscymnus coelolepis for species discrimination within the genus. Book of Ab- and Centrophorus squamosus from west of Ireland and Scot- stracts, 17th European Elasmobranch Association Confer- land. Journal of the Marine Biological Association of the United ence 2013, Plymouth University, UK, 74. Kingdom 81: 1019–1029. Guallart J, Serena F, Mancusi C, Casper BM, Burgess Compagno LJV. 1984. FAO species catalogue. Vol. 4. Sharks GH, Ebert DA, Clarke M, Stenberg C. 2006. Centrophorus of the world. An annotated and illustrated catalogue of shark granulosus. In: IUCN 2013 – IUCN red list of threatened species known to date. Part 1. Hexanchiformes to Lamniformes. species (version 2013.1). Available at: http://www.iucnredlist.org Rome: FAO. Gubanov EP. 1986. New data on the dogfish, Centrophorus Compagno LJV, Dando M, Fowler S. 2005. A field guide tesselatus (Squalidae), from the Indian Ocean. Journal of Ich- to sharks of the world. London: Harper Collins Publishing thyology 26: 155–157. Ltd. ICES. 2006. Report of the working group on elasmobranch fishes Cotton CF. 2010. Age, growth, and reproductive biology of deep- (WGEF), 14–21 June 2006, ICES Headquarters (Copenha- water chondrichthyans. Unpublished D. Phil. Thesis, College gen, Denmark). ICES CM 2006/ACFM:31. 291 pp. of William and Mary. ICES. 2012. Report of the working group on elasmobranch fishes Daley RK, Williams A, Green M, Barker B, Brodie P. In (WGEF), 19–26 June 2012, Lisbon, Portugal. ICES CM 2012/ press. Can marine reserves conserve vulnerable sharks in ACOM:19. 547 pp. the deep sea? A case study of Centrophorus zeehaani Iglésias SP, Toulhoat L, Sellos DY. 2010. Taxonomic con- (Centrophoridae), examined with acoustic telemetry. Deep Sea fusion and market mislabelling of threatened skates: im- Research Part II Topical Studies in Oceanography. portant consequences for their conservation status. Aquatic Dyb JE, Bergstad OA. 2004. Mar-Eco – the cruise with M/S Conservation: Marine and Freshwater Ecosystems 20: 319– Loran, Summer 2004. 333. Ebert DAE, Stehmann MFW. 2012. Sharks, batoids, and chi- Kajiura SM, Tricas TC. 1996. Seasonal dynamics of dental maeras of the North Atlantic. FAO Species Catalogue for sexual dimorphism in the Atlantic stingray Dasyatis sabina. Fishery Purposes, Vol. 7. Rome: FAO. Journal of Experimental Biology 199: 2297–2306. Eschmeyer WN, ed. 2013. Catalog of fishes. California Academy Krefft G, Tortonese E. 1973. Squalidae. In: Hureau J-C, Monod of Sciences Available at: http://research.calacademy.org/ T, eds. Check list of the fishes of the North Atlantic and of research/ichthyology/catalog/fishcatmain.asp. Accessed 11 April the Mediterranean, Vol. 1. Paris: UNESCO, 37–50. 2013. Last PR, Stevens JD. 1994. Sharks and rays of Australia. Figueiredo I, Machado PB, Gordo LS. 2005. Deep-water Melbourne, Australia: CSIRO. sharks fisheries off the Portuguese continental coast. Journal Last PR, Stevens JD. 2009. Sharks and rays of Australia, of the Northwest Atlantic Fisheries Science 35: 291–298. 2nd edn. Canberra, Australia: CSIRO. Figueiredo I, Moura T, Neves A, Gordo LS. 2008. Repro- Ledoux JC. 1970. Les dents des Squalidés de la Méditerranée ductive strategy of leafscale gulper shark Centrophorus occidentale et de l’Atlantique Nord-ouest africain. Vie et Milieu squamosus and the Portuguese dogfish Centroscymnus 21: 309–362. coelolepis on the Portuguese continental slope. Journal of Fish Lloris D, Rucabado J, Del Cerro L, Portas F, Demestre Biology 73: 206–225. M, Roig A. 1984. Tots els peixos del Mara Català. I: Llistat Garman S. 1906. New Plagiostomia. Bulletin of the Museum de cites i referències. Treballs de la Societat Catalana de of Comparative Zoology 46: 203–208. Ictiología i Herpetología 1: 1–208. Garman S. 1913. The Plagiostomia (sharks, skates, and rays) Lozano-Cabo F. 1963. Nomenclatura Ictiológica. Nombres – memoirs of the Museum of Comparative Zoology 36. Harvard: científicos y vulgares de los peces españoles. Instituto Español Harvard College. Oceanografía 31: 1–271. Graham KJ, Andrew NL, Hodgson KE. 2001. Changes in Lozano-Rey L. 1928. Fauna ibérica. Peces (Generalidades, relative abundance of sharks and rays on Australian South Ciclóstomos y Elasmobranquios). Vol. 1. Madrid: Museo East Fishery trawl grounds after twenty years of fishing. Nacional de Ciencias Naturales. Marine and Freshwater Research 52: 549–561. Maurin C. 1968. Ecologie ichthyologique des fonds chalutables Graham KJ, Daley RK. 2011. Distribution, reproduction and Atlantiques (de la baie ibéro-marocaine à la Mauritanie) et population structure of three gulper sharks (Centrophorus, de la Méditerranée occidentale. Revue des Travaux de l’Institut Centrophoridae) in south-east Australian waters. Marine and des Peches Maritimes 32: 1–136. Freshwater Research 62: 583–595. Maurin C, Bonnet M. 1970. Poissons des côtes nord-ouest Guallart J. 1998. Contribución al conocimiento de la biología africaines (campagnes de la ‘Thalassa’ 1962 et 1968). Revue y la taxonomía del tiburón batial Centrophorus granulosus des Travaux de l’Institut des Peches Maritimes 34: 125– (Bloch & Schneider, 1801) (Elasmobranchii, Squalidae) en 170.

McEachran JD, Branstetter S. 1984. Squalidae. In: White- Rodríguez-Cabello C, Sánchez F. 2014. Is Centrophorus head PJP, Bauchot M-L, Hureau J-C, Nielsen J, Tortonese squamosus a highly migratory deep-water shark? Deep E, eds. Fishes of the North-eastern Atlantic and the Medi- Sea Research Part I: Oceanographic Research Papers 92: Downloaded from https://academic.oup.com/zoolinnean/article-abstract/172/4/803/2433424 by VIRGINIA INSTITUTE OF Marine Science user on 30 October 2018 terranean. Vol. I. Paris: UNESCO, 128–147. 1–10. McLaughlin DM, Morrissey JF. 2005. Reproductive biology Sang L. 1991. Aspects on the reproduction and embryonic de- of Cenrophorus cf. uyato from the Cayman Trench, Jamaica. velopment of the gulper shark, Centrophorus granulosus Journal of the Marine Biological Association of the United (Elasmobranchii, Squalidae), in Puerto Rico, with notes on Kingdom 85: 1185–1192. its biology. MSc Thesis, University of Puerto Rico. Megalofonou P, Chatzispyrou A. 2006. Sexual maturity and Serena F. 2005. FAO species identification guide for fishery feeding of the gulper shark, Centrophorus granulosus, from purposes – field identification guide to the sharks and rays the eastern Mediterranean Sea. Cybium 30: 67–74. of the Mediterranean and Black Sea. Rome: FAO. Menezes GM, Tariche O, Pinho MR, Duarte PN, Fernandes Seutin G, White BN, Boag PT. 1991. Preservation of avian A, Aboim MA. 2004. Annotated list of fishes caught blood and tissue samples for DNA analyses. Canadian Journal by the R/V ARQUIPÉLAGO off the Cape Verde archipela- of Zoology 69: 82–90. go. Arquipélago – Life and Marine Sciences 21A: Severino RB, Afonso-Dias I, Delgado J, Afonso-Dias M. 57–71. 2009. Aspects of the biology of the leafscale gulper shark Moore JA, Hartel KE, Craddock JE, Galbraith JK. 2003. Centrophorus squamosus (Bonnaterre, 1788) off Madeira archi- An annotated list of deepwater fishes from off the New pelago. Arquipélago – Life and Marine Sciences 26: 57–61. England region, with new area records. Northeastern Natu- Stehmann MFW. 2002. Proposal of a maturity stages scale ralist 10: 159–248. for oviparous and viviparous cartilaginous fishes (Pisces, Müller J, Henle FGJ. 1837a. Gattungen der Haifische und Chondrichthyes). Archive of Fishery and Marine Research 50: Rochen, nach ihrer Arbeit: Über die Naturgeschichte der 23–48. Knorpelfische. Preussische Akademie der Wissenschaften zu Tabit CR. 1993. Interspecific and intraspecific variability in Berlin 111–118. placoid scale morphology in relation to body form variabil- Müller J, Henle FGJ. 1837b. Über die Gattungen der ity in Squaliformes. D. Phil. Thesis, College of William and Plagiostomen. Archiv für Naturgeschichte 3: 394–401, 434. Mary. Müller J, Henle FGJ. 1841. Systematische Beschreibung der Tamura K, Peterson D, Peterson N, Stecher G, Nei M, Plagiostomen. Berlin: Verlag Von Veit und Comp. Kumar S. 2011. MEGA5: molecular evolutionary genetics Muñoz-Chapuli R. 1990. Els condrictis elasmobranquis. In: analysis using maximum likelihood, evolutionary distance, Carreras J, Folch R, Gonsàlbez J, Llimona X, Terradas J, and maximum parsimony methods. Molecular Biology and eds. Història Natural dels Països Catalans. Vol. 11. Barce- Evolution 28: 2731–2739. lona: Enciclopèdia Catalana, 125–153. Teng HT. 1959. Studies on the elasmobranch fishes from Muñoz-Chapuli R, Blasco M. 1984. Tendencias generales del Formosa. Pt. 6. A new species of deep-sea shark (Centrophorus crecimiento relativo en escualos. Investigación Pesquera 48: niaukang) from Formosa. Report – Laboratory of Fishery 303–317. Biology, Taiwan Fisheries Research Institute 9: 1–6. Muñoz-Chapuli R, Ramos F. 1989. Review of the Centrophorus Thompson JD, Higgins DG, Gibson TJ. 1994. CLUSTAL sharks (Elasmobranchii, Squalidae) of the eastern Atlantic. W: improving the sensitivity of progressive multiple se- Cybium 13: 65–81. quence alignment through sequence weighting, position- Naylor GJP, Caira JN, Jensen K, Rosana KAM, White specific gap penalties and weight matrix choice. Nucleic Acids WT, Last PR. 2012. A DNA sequence-based approach to the Research 22: 4673–4680. identification of shark and ray species and its implications Tortonese E. 1969. The squaliforms on the Ligurian Sea: for global elasmobranch diversity and parasitology. a revised list with notes. Israel Journal of Zoology 18: 233– Bulletin of the American Natural History Museum 367: 236. 1–263. Veríssimo A, McDowell JR, Graves JE. 2012. Genetic popu- Orlov AM, Cotton CF. 2011. Sexually dimorphic morpho- lation structure and connectivity in a commercially exploit- logical characters in five North Atlantic deepwater skates ed and wide-ranging deepwater shark, the leafscale gulper (Chondrichthyes : Rajiformes). Journal of Marine Biology 2011: (Centrophorus squamosus). Marine and Freshwater Re- Article ID 842821. search 63: 505–512. Palumbi SR, Romano A, McMillan WO, Stice L, Grabowski Ward RB, Zemlak TS, Innes BH, Last PR, Hebert PDN. G. 1991. The simple fool’s guide to PCR. Honolulu: Univer- 2005. DNA barcoding Australia’s fish species. Philosophical sity of Hawaii. Transactions of the Royal Society B 360: 1847–1857. R Development Core Team. 2012. R: a language and envi- White WT, Dharmadi. 2010. Aspects of maturation and re- ronment for statistical computing, v. 3.0.2. Vienna, Austria: production in hexanchiform and squaliform sharks. Journal R Foundation for Statistical Computing. Available at: http:// of Fish Biology 76: 1362–1378. www.R-project.org/ White WT, Ebert DA, Compagno LJV. 2008. Description Rafinesque CS. 1810. Caratteri di alcuni nuovi generi e nuove of two new species of gulper sharks, genus Centrophorus specie di animali e piante della Sicilia, con varie osservazioni (Chondrichthyes: Squaliformes: Centrophoridae) from sopra i medisimi. Palermo. Australia. In: Last PR, White WT, Pogonoski JJ, eds.

Descriptions of new Australian chondrichthyans. CSIRO Marine Yano K, Kugai K. 1993. Taiwan gulper shark, Centro- and Atmospheric Research Paper 022. CSIRO: 1–21. phorus niaukang, from the Okinawa Islands, Japan. Bul- White WT, Ebert DA, Naylor GJP, Ho H-C, Clerkin P, letin of the Seikai National Fisheries Research Institute 71: Downloaded from https://academic.oup.com/zoolinnean/article-abstract/172/4/803/2433424 by VIRGINIA INSTITUTE OF Marine Science user on 30 October 2018 Veríssimo A, Cotton C. 2013. Revision of the genus 41–49. Centrophorus (Squaliformes: Centrophoridae): part 1 – Yano K, Tanaka S. 1986. A telemetric study on the Redescription of Centrophorus granulosus (Bloch & Schneider), movements of the deep sea squaloid shark, Centrophorus a senior synonym of C. acus Garman and C. niaukang Teng. acus. In: Uyeno T, Arai R, Taniuchi T, Matsuura K, eds. Zootaxa 3752: 35–72. Proceedings of the Second International Conference on White WT, Last PR, Stevens JD, Yearsley GK, Fahmi, Indo-Paciﬁc Fishes. Tokyo National Museum, Ueno Dharmadi. 2006. Economically important sharks and rays Park, Tokyo: Ichthyological Society of Japan, 372– of Indonesia. Canberra: ACIAR Publishing. 380.

SUPPORTING INFORMATION Additional Supporting Information may be found in the online version of this article at the publisher’s web-site: Figure S1. Schematic representation of the morphometric measurements recorded in this study. A, lateral view; B, ventral view; C, dorsal head view; D, ventral head view; E, dorsal fin; F, pectoral fin; G, pelvic fin. Please note that total length (TL) is not represented in the figure as this measurement was taken when the caudal fin was aligned with the body axis, and not in its the natural position as depicted in the figure. Abbreviations follow those listed in Table 1. Table S1. Haplotype list per gene region with corresponding GenBank accession nos; number of specimens per haplotype; original species identifications revised species designation; ocean basins; and geographical locations where present. Table S2. Specimens used in the morphological data analysis: museum collection, catalogue reference, original species designation, location, geographical coordinates, genetic clade, sex (F, female; M, male), maturity stage, condition when measured (P, preserved; F, fresh), total length (TL), and morphometric ratio values (see Table 1 for abbreviations). Table S3. Ranges of body measurements as proportion of total length (TL) for the eight morphotypes identified in the North Atlantic sensu lato, in addition to Centrophorus tesselatus holotype (MCZ 1031); n, number of specimens measured.