Äîêóìåíò âçÿò èç êýøà ïîèñêîâîé ìàøèíû. Àäðåñ îðèãèíàëüíîãî äîêóìåíòà : http://www.sevin.ru/laboratories/Marine_Invertebrates/kantor/113_Puillandre_et_al_2010_Gemmuloborsonia.pdf Äàòà èçìåíåíèÿ: Fri Feb 12 15:16:38 2010 Äàòà èíäåêñèðîâàíèÿ: Mon Feb 4 05:14:17 2013 Êîäèðîâêà: Ïîèñêîâûå ñëîâà: aurora

CRYPTIC S PECIES IN GEMMULO BORSONIA (GASTROPODA: CO NOIDEA)
N. PUILLANDRE 1, C . CRUAUD 2 AND Y U. I. KANTOR
1

3

UMR 7138, Museum National d'Histoire Naturelle, Departement Systematique et Evolution, CP26, 57 rue Cuvier, 75231 Paris Cedex 05, France; 2 GENOSCOPE, Centre National de Se´quencage, 2 rue Gaston Cre´mieux, CP 5706, 91057 Evry Cedex, France; and ¸ 3 A.N. Severtsov Institute of Ecology and Evolution of Russian Academy of Sciences, Leninski Prosp. 33, Moscow 119071, Russia (Received 17 December 2008; accepted 19 May 2009)

ABSTRAC T
During a broad molecular taxonomic and phylogenetic survey of the gastropod superfamily Conoidea, 80 specimens of several species of the genus Gemmuloborsonia were sequenced for the cytochrome c oxidase subunit I gene. The genus, originally established for fossil species from the Plio-Pleistocene of the Philippines, now includes living species from bathyal depths of the Indo-Pacific Oceans. The molecular data demonstrated the presence of five separate entities, while only four `morphospecies' could be isolated by visual examination. The two largest groups, representing separate species from the molecular data, were impossible to distinguish with certainty using shell or anatomical characters. To examine shell morphology in more detail the shape of the last whorl was analysed by Fourier analysis, and the Fourier coordinates were used in canonical variate analysis. The majority of the specimens were separated into two groups, but 21.6% of the specimens were impossible to distinguish by morphological characters. One of these two forms was attributed to the known species Gemmuloborsonia moosai Sysoev & Bouchet, 1996, while the other is described as a new species Gemmuloborsonia clandestina. Bathytoma colorata Sysoev & Bouchet, 2001 is transferred to Gemmuloborsonia on the basis of molecular analysis and radular morphology. Another species, represented in our material by a single specimen, remains undescribed.

INTRODUC TION
The taxonomy of shell-bearing molluscs was, and continues to be, largely based on shell characters. For more than 250 years of scientific malacology shell characters have proved to be effective species-level identifiers, especially when the protoconch is also included. Similar characters have been employed for the exceptionally good fossil record of molluscs. Moreover, reliance on shell characters is also justified by the fact that many species have simply never been collected alive; for example, during intensive surveys of coral reef sites 28% of the species were not collected alive (Bouchet et al., 2002). For the overwhelming majority of described species of molluscs the primary namebearing types are shells, these usually being the only part of the animal preserved in collections. This leaves conchological characters as the major if not the only source of evidence for taxonomic decisions. The conventional approach to documenting molluscan diversity or revisionary taxonomy is to sort material to morphospecies on the basis of shell characters, with subsequent testing of taxonomic hypotheses with all available data, such as anatomy, biogeographic information and, more recently, with molecular analyses. The final stage of any taxonomic decision in malacology is the critical reevaluation of the shell characters in order to identify and formalize reliable discriminating features, and this requires the estimation and evaluation of the intraspecific variability of the shell. Convergence and homoplasy render shell characters much less reliable predictors of relationships at higher taxonomic levels ( family, genus) within the gastropod superfamily Conoidea ( ¼ Toxoglossa). Sometimes the incongruence between the shell and internal anatomy is startling and species with very similar shells may be very distantly related. For instance, shells of Toxicochlespira Sysoev & Kantor, 1990 (Conidae) strongly
Correspondence: N. Puillandre; e-mail: puillandre@mnhn.fr

resemble representatives of Cochlespira Conrad, 1865 (Turridae) (Sysoev & Kantor, 1990); shells of Strictispira McLean, 1971 (Strictispiridae), are hardly distinguishable from those of many species of Crassispira Swainson, 1840 (Turrridae, Crassispirinae) (Tippett, 2006); and the radula-less species Cenodagreutes aethus Smith, 1967 is said to be conchologically indistinguishable from the radulate Raphitoma leufroyi (Michaud, 1828) (both Conidae, Raphitominae) (Fretter & Graham, 1985). Cryptic, or sibling, species are ubiquitous among marine animals and molluscs are no exception (see reviews by Knowlton, 1993, 2000). In reality, most recently discovered cryptic species of Gastropoda are forms with superficially similar shells that can usually be reliably distinguished by anatomical characters. A recent example is the discovery of two conchologically very similar pairs of species that were initially placed in the genus Xenuroturris Iredale, 1929, but differ markedly in radular morphology (Kantor et al., 2008). Molecular techniques are now more frequently employed in taxonomic analysis and are revealing numerous cases of cryptic species in all groups, including molluscs (e.g. Williams & Reid, 2004; Collin, 2005; Reid et al., 2006; Duda et al., 2008; Malaquias & Reid, 2008). Molecular data are now routinely used in combination with shell and anatomical characters for taxonomic purposes and sometimes become the ultimate proof of the existence of separate species. Cryptic species in molluscs pose significant nomenclatural problems since unambiguous assignment of older type specimens, which in molluscs are nearly always represented by the empty shell (sometimes even severely `beach worn' and often without good locality data), to one of several forms may be extremely difficult or impossible. During the course of a broad-scale taxonomic and phylogenetic survey of the superfamily Conoidea, several cases were found where molecular data conflicted with hypotheses based on conventional shell and sometimes even anatomical characters (Puillandre et al., 2008). A remarkable example is the
doi:10.1093/mollus/eyp042

Journal of Molluscan Studies (2010) 76: 11 ­ 23. # The Author 2010. Published by Oxford University Press on behalf of The Malacological Society of London, all rights reserved


N. PUILLANDRE ET A L . genus Gemmuloborsonia Shuto, 1989. The genus was established for four fossil species of Turridae from the Plio-Pleistocene of the Philippines, and Late Miocene of Indonesia and Italy (Shuto, 1989). Subsequently, five living species were described from bathyal waters of the Indo-Pacific by Sysoev & Bouchet (1996). Gemmuloborsonia was initially assigned to the subfamily Borsoniinae ( ¼ Clathurellinae fide Taylor et al., 1993), but transferred on the basis of radular characters to the Turrinae (Sysoev & Bouchet, 1996). However, recent molecular analyses do not support this result (Puillandre et al., 2008). A molecular analysis of 80 specimens of Gemmuloborsonia from the central Indo-Pacific using the cytochrome c oxidase subunit I (COI) gene demonstrated the presence in our material of five separate entities, while only four `morphospecies' could be isolated by visual examination. The two largest groups, representing separate species based on the molecular data, were initially impossible to distinguish by shell characters and it is not surprising that these forms avoided recognition even in the latest taxonomic revision of the genus (Sysoev & Bouchet, 1996). These discoveries led us to make a more detailed examination of the Gemmuloborsonia species complex, combining the molecular results with studies of the radula and multivariate analysis of shell form. sequences were calculated using MEGA 3.1 (Kumar, Tamura & Nei, 2004). Phylogenetic reconstructions were conducted using Bayesian Analysis (BA), consisting of two Markov chains (10,000,000 generations each with a sampling frequency of one tree per 1,000 generations) run in four parallel analyses using MrBayes (Huelsenbeck, Ronquist & Hall, 2001). When the log-likelihood scores were found to stabilize, a consensus tree was calculated after omitting the first 25% trees as burn-in. Only the number of nucleotide substitution rates categories (6) was fixed, the other parameters of the substitution model being estimated during the BA.

Fourier analysis
Morphometric analyses were performed in an attempt to identify morphological differences between two groups recognized genetically but indistinguishable morphologically using traditional characters. The shape of the last whorl was analysed by Fourier analysis, as previously described by Puillandre et al. (2009). Shells were placed horizontally, aperture up, and digitized at the same magnification using a macro stand, to reduce possible optical distortions. As the outer apertural lip of some shells was broken, this part of the whorl was not included in the analysis, as shown in Figure 1. Five landmarks were defined, corresponding to adapical and abapical margins of the peripheral keel on both sides of the shell, and to the tip of the siphonal canal (Fig. 1). The five landmarks, as well as the outlines, were digitized using TpsDig (Rohlf, 1996). The same starting point, corresponding to the first landmark, was always used. All pictures and outlines were taken by the same operator (N.P.). Outlines were used as input for an EFA (Elliptic Fourier Analysis; Dommergues et al., 2003; Baylac & Friess, 2005). The five landmarks were used as control points to rotate the outlines into the same orientation. The images were then centred and normalized for size (using square roots of the surface). A visualization of Fourier reconstructions using different numbers of harmonics, compared to the original outline, was used to estimate that 40 harmonics were sufficient to reconstruct the outlines with high accuracy (Fig. 1). The obtained Fourier coordinates were used in canonical variate analyses (CVAs), using two different grouping variables: (1) the genetic groups as defined by the molecular analysis and (2) the cruise of collection (Table 1). Visualizations of the outline deformations along the canonical axes were made using the procedure described by Monti et al. (2001). Assignation of a specimen to one or another genetic group was tested by a `leave-one-out' cross-validation (1,000 bootstrap replicates). Several type specimens, not preserved in alcohol and thus not available for molecular analysis, were added to the CVAs and assigned successively to each of the genetic groups. All analyses were performed using specially devised MATLABv5.2 functions implemented by Michel Baylac. ´ Institutional abbreviations: MNHN, Museum National d'Histoire Naturelle, Paris; NM, Natal Museum, Pietermaritzburg, South Africa; NMNZ, National Museum of New Zealand, Wellington; PPPO-LIPI, Pusat Penelitian dan Pengembangan Oseanologi LIPI, Jakarta, Indionesia; ZMMSU, Zoological Museum of Moscow State University, Moscow.

MAT E RIAL AND M ETHODS Sampling
A total of 80 living specimens potentially belonging to the genus Gemmuloborsonia were collected between 2004 and 2007 in the Philippines, Solomon Islands and Coral Sea (Table 1). Specimens were preserved in 90% or 100% ethanol specifically for molecular analysis by clipping pieces of the head ­ foot from anaesthetized specimens, thus keeping the shell intact for morphological analyses. All material is deposited in the collections ´ of the Museum National d'Histoire Naturelle, Paris (MNHN). In order to test the monophyly of the genus Gemmuloborsonia, we used as outgroups several species of Turrinae (Lophiotoma albina Lamarck, 1822; Turris babylonia Linnaeus, 1758; Gemmula diomedea E.A. Smith, 1894; Lucerapex sp.), and several species belonging to other subfamilies of Conoidea (Clavus sp., Raphitoma sp., Conus orbignyi Kilburn, 1975). Additionally, one specimen included in our samples was identified as Bathytoma colorata Sysoev & Bouchet, 2001, but this species is thought to belong to the genus Gemmuloborsonia (see Discussion). To test this hypothesis, we also included three specimens of the genus Bathytoma Harris & Burrows, 1891. A species of Harpa (Neogastropoda: Harpidae) was used as a distant outgroup.

Extraction and sequencing
DNA was extracted from a piece of foot, using 6100 Nucleic Acid Prepstation system (Applied Biosystems). A fragment of 658 bp of the COI mitochondrial gene was amplified using universal primers LCO1490 and HCO2198 (Folmer et al., 1994). All PCRs were performed in a volume of 25 ml, containing 3 ng of DNA, 1 á reaction buffer, 2.5 mM MgCl2, 0.26 mM dNTP, 0.3 mM of each primer, 5% DMSO and 1.5 U of Q-Bio Taq (MPBiomedicals). Amplifications were performed according to Hebert et al. (2003). PCR products were purified and sequenced by a sequencing facility (Genoscope). In all cases, both directions were sequenced to confirm accuracy. For GenBank accession numbers see Table 1.

R E S U LT S
Eighty specimens were sequenced for COI, resulting in a 658-bp fragment. The Bayesian tree supports the monophyly of Gemmuloborsonia [ posterior probability (PP) ¼ 0.99], and shows five different groups, numbered from 1 to 5 in Figure 2. Each group includes from 1 to 49 specimens. Groups 3 ­ 5, each 12

Phylogenetic analysis
COI sequences were manually aligned, because no ambiguous indels were found. Genetic distances ( p-distances) between


CRYP TIC S PEC I ES IN CO NOIDEA
Table 1. Identification number (MNHN ID), cruise, station, species identification, percentage of assignation obtained with the CVA for the two groups identified as Gemmuloborsonia moosai ( percentage provided only when the specimens were assigned to the wrong group) and BOLD (Barcode Of Life Database) and GenBank numbers are given for each specimen.
ID 17849 41918 41919 41920 41921 41922 41923 41924 41925 41926 41927 41928 41929 41930 41931 41932 41933 41934 41935 41936 41937 41938 41939 41940 41941 41942 41943 41944 41945 41946 41947 41948 41949 41950 41951 41952 41953 41954 41955 41956 41957 41958 41959 41960 41961 41962 41963 41964 41965 41966 41967 41968 41969 41970 41971 BOLD CONO192-08 CONO841-08 CONO597-08 CONO598-08 CONO599-08 CONO758-08 CONO780-08 CONO822-08 CONO811-08 CONO798-08 CONO812-08 CONO844-08 CONO534-08 CONO560-08 CONO559-08 CONO555-08 CONO556-08 CONO557-08 CONO558-08 CONO547-08 CONO542-08 CONO548-08 CONO549-08 CONO550-08 CONO537-08 CONO546-08 CONO554-08 CONO540-08 CONO538-08 CONO536-08 CONO539-08 CONO553-08 CONO541-08 CONO535-08 CONO552-08 CONO545-08 CONO544-08 CONO813-08 CONO814-08 CONO815-08 CONO816-08 CONO817-08 CONO824-08 CONO825-08 CONO826-08 CONO829-08 CONO830-08 CONO831-08 CONO832-08 CONO833-08 CONO834-08 CONO835-08 CONO836-08 CONO837-08 CONO840-08 GenBank EU015658 FJ462616 FJ462589 FJ462588 FJ462587 FJ462590 FJ462593 FJ462602 FJ462595 FJ462594 FJ462596 FJ462619 FJ462557 FJ462583 FJ462582 FJ462578 FJ462579 FJ462580 FJ462581 FJ462570 FJ462565 FJ462571 FJ462572 FJ462573 FJ462560 FJ462569 FJ462577 FJ462563 FJ462561 FJ462559 FJ462562 FJ462576 FJ462564 FJ462558 FJ462575 FJ462568 FJ462567 FJ462597 FJ462598 FJ462599 FJ462600 FJ462601 FJ462603 FJ462604 FJ462605 FJ462606 FJ462607 FJ462608 FJ462609 FJ462610 FJ462611 FJ462612 FJ462613 FJ462614 FJ462615 Cruise EBISCO (Chesterfield Islands) Salomon 3 (Solomon Islands) Norfolk 2 (Norfolk ridge) Norfolk 2 (Norfolk ridge) Norfolk 2 (Norfolk ridge) Salomon 2 (Solomon Islands) Salomon 2 (Solomon Islands) Salomon 3 (Solomon Islands) Salomon 2 (Solomon Islands) Salomon 2 (Solomon Islands) Salomon 3 (Solomon Islands) Salomon 3 (Solomon Islands) Aurora 07 (Philippines) Aurora 07 (Philippines) Aurora 07 (Philippines) Aurora 07 (Philippines) Aurora 07 (Philippines) Aurora 07 (Philippines) Aurora 07 (Philippines) Aurora 07 (Philippines) Aurora 07 (Philippines) Aurora 07 (Philippines) Aurora 07 (Philippines) Aurora 07 (Philippines) Aurora 07 (Philippines) Aurora 07 (Philippines) Aurora 07 (Philippines) Aurora 07 (Philippines) Aurora 07 (Philippines) Aurora 07 (Philippines) Aurora 07 (Philippines) Aurora 07 (Philippines) Aurora 07 (Philippines) Aurora 07 (Philippines) Aurora 07 (Philippines) Aurora 07 (Philippines) Aurora 07 (Philippines) Salomon 3 (Solomon Islands) Salomon 3 (Solomon Islands) Salomon 3 (Solomon Islands) Salomon 3 (Solomon Islands) Salomon 3 (Solomon Islands) Salomon 3 (Solomon Islands) Salomon 3 (Solomon Islands) Salomon 3 (Solomon Islands) Salomon 3 (Solomon Islands) Salomon 3 (Solomon Islands) Salomon 3 (Solomon Islands) Salomon 3 (Solomon Islands) Salomon 3 (Solomon Islands) Salomon 3 (Solomon Islands) Salomon 3 (Solomon Islands) Salomon 3 (Solomon Islands) Salomon 3 (Solomon Islands) Salomon 3 (Solomon Islands) Species colorata sp. neocaledonica neocaledonica neocaledonica moosai moosai moosai moosai moosai moosai moosai clandestina clandestina clandestina clandestina clandestina clandestina clandestina clandestina clandestina clandestina clandestina clandestina clandestina clandestina clandestina clandestina clandestina clandestina clandestina clandestina clandestina clandestina clandestina clandestina clandestina moosai moosai moosai moosai moosai moosai moosai moosai moosai moosai moosai moosai moosai moosai moosai moosai moosai moosai 0.8906 0.9887 0.6045 0.5562 0.7968 0.9879 0.9565 0.6838 0.6995 0.7025 0.733 % wrong group assignation

Continued

13


N. PUILLANDRE ET A L . Table 1. Continued
ID 41972 41973 41974 41975 41976 41977 41978 41979 41980 41981 41982 41983 41984 41985 41986 41987 41988 41989 41990 41991 41992 41993 41994 41995 41996 17700 17754 17756 17865 17890 17902 17921 17929 40569 40813 42305 FRANZ270-08 CONO570-08 BOLD CONO842-08 CONO843-08 CONO846-08 CONO847-08 CONO848-08 CONO849-08 CONO850-08 CONO851-08 CONO853-08 CONO854-08 CONO855-08 CONO856-08 CONO857-08 CONO858-08 CONO859-08 CONO860-08 CONO861-08 CONO862-08 CONO762-08 CONO759-08 CONO551-08 CONO543-08 CONO571-08 CONO580-08 CONO581-08 CONO147-08 CONO226-08 CONO481-08 CONO242-08 CONO279-08 CONO225-08 CONO296-08 CONO363-08 GenBank FJ462617 FJ462618 FJ462620 FJ462621 FJ462622 FJ462623 FJ462624 FJ462625 FJ462626 FJ462627 FJ462628 FJ462629 FJ462630 FJ462631 FJ462632 FJ462633 FJ462634 FJ462635 FJ462592 FJ462591 FJ462574 FJ462566 FJ462584 FJ462585 FJ462586 EU015643 EU015677 EU127882 EU015687 EU015713 EU015676 EU015721 EU015742 EU685626 EU820609 FJ462636 Cruise Salomon 3 (Solomon Islands) Salomon 3 (Solomon Islands) Salomon 3 (Solomon Islands) Salomon 3 (Solomon Islands) Salomon 3 (Solomon Islands) Salomon 3 (Solomon Islands) Salomon 3 (Solomon Islands) Salomon 3 (Solomon Islands) Salomon 3 (Solomon Islands) Salomon 3 (Solomon Islands) Salomon 3 (Solomon Islands) Salomon 3 (Solomon Islands) Salomon 3 (Solomon Islands) Salomon 3 (Solomon Islands) Salomon 3 (Solomon Islands) Salomon 3 (Solomon Islands) Salomon 3 (Solomon Islands) Salomon 3 (Solomon Islands) Salomon 2 (Solomon Islands) Salomon 2 (Solomon Islands) Aurora 07 (Philippines) Aurora 07 (Philippines) EBISCO (Chesterfield Islands) EBISCO (Chesterfield Islands) EBISCO (Chesterfield Islands) Vanuatu Philippines Vanuatu Philippines Philippines Philippines Philippines Solomon Islands Vanuatu Philippines Chesterfield Islands Species moosai moosai moosai moosai moosai moosai moosai moosai moosai moosai moosai moosai moosai moosai moosai moosai moosai moosai moosai moosai moosai moosai neocaledonica moosai moosai Bathytoma sp. Turris babylonia Lophiotoma albina Bathytoma tippetti Raphitoma sp. Clavus sp. Conus orbignyi Bathytoma sp. Harpidae, Harpa sp. Gemmula diomedea Lucerapex sp. 0.9999 0.8149 0.9973 0.9068 0.9572 % wrong group assignation

including several specimens, are well supported (PP . 0.95). Mean genetic distance between groups ranges from 5.6% (between Groups 1 and 3) to 10% (between Groups 1 and 2). The high genetic distances found between the different groups within Gemmuloborsonia are generally interpreted as interspecific distances (see e.g. Hebert et al., 2003; Smith, Fisher & Hebert, 2005; more specifically for molluscs: Mikkelsen, Schander & Willassen, 2007; Malaquias & Reid, 2008; Puillandre et al., 2009). Furthermore, all the groups that include several specimens are reciprocally monophyletic. These two findings suggest that these five entities certainly correspond to different species (Samadi & Barberousse, 2006).

Group 4: This contains four specimens collected off New Caledonia and in all respects similar to Gemmuloborsonia neocaledonica Sysoev & Bouchet, 1996 (Fig. 3G). Groups 3 and 5: These are the most numerous, containing 25 and 49 specimens, respectively. In general shell shape and in multispiral protoconch both groups resemble Gemmuloborsonia moosai Sysoev & Bouchet, 1996. The similarity between the two forms is striking; nevertheless the molecular data suggested the presence of two separate species. Both groups are reciprocally monophyletic, and the average genetic distance between them is 6.2%. Moreover, both forms were sympatric in a single dredge haul (400 ­ 500 m) from the Philippines. Studies of the radulae (Fig. 4C ­ D and E ­ F) revealed no significant differences, nor did standard measurements of the shell. A Fourier analysis was performed on the 74 specimens included in genetic groups 3 and 5. Using CVA, the two genetic groups are separated along the axis, although not completely (Fig. 5). Among 74 specimens, 16 are not assigned to the correct genetic group (Table 1). Using the collection locality (cruise) as a discriminant variable (Fig. 6), specimens collected during the Salomon 2 cruise are separated from the 14

Taxonomic position of studied forms
Group 1: This group is represented by a single specimen identified as Gemmuloborsonia sp. (Fig. 3F). Group 2: Group 2 consists of a single adult specimen from New Caledonia, identified as Gemmuloborsonia colorata (Sysoev & Bouchet, 2001) comb. nov. (Fig. 3D ­ E).


CRYP TIC S PEC I ES IN CO NOIDEA

Figure 1. Outline reconstructions with increasing number of harmonics indicated within outlines. In black, original outline. The five landmarks are represented on the shell picture.

others on the first axis (representing 48.07% of the variance). On the second axis (39.44% of the variance), specimens collected during Aurora 07 (Philippines) and Salomon 3 (Solomon Islands) expeditions are discriminated, but the two groups overlap. Results of the CVA obtained both using genetic groups or collection locality as discriminant variable are similar, because genetic group 3 is exclusively from the Philippines, and genetic group 5 is exclusively from the Solomon Islands and Chesterfield Plateau, except for two specimens collected in the Philippines. The axes discriminating genetic groups 3 and 5 (Fig. 5) and specimens from the Solomon Islands and Philippines (Fig. 6) show a contrast between forms with stouter shells with lower last whorls and those forms with more elongated shells and taller last whorls. The presence of two forms poses the question as to the applicability of names. One form is broadly distributed from the Coral Sea to Solomon Islands and Philippines, while the other is found only in the Philippines. The multivariate analysis demonstrated that the holotype of Gemuloborsonia moosai falls within Group 5 (Table 2). Thus, the name moosai can be attributed to Group 5.

SYSTEM AT IC DESCRIPTIONS
Gemmuloborsonia sp. (Fig. 3F) Material examined: Solomon Islands (stn 2850), one live specimen, shell length 25.3 mm (MNHN 41918). Remarks: The single specimen examined was an immature female. The radula was studied (Fig. 4B) and is typical for the genus. It consists of only 42 rows of teeth, 1.75 mm in length [0.20 of aperture length (AL)]. The marginal teeth are duplex, about 110 mm long (1.25% of AL), with the central formation (Kantor, 2006) nearly rectangular, formed of fused central and lateral teeth. The cusp ( ¼ central tooth) is rather weak and narrow. The specimen lacks a protoconch and is characterized by very prominent, bulging subsutural fold and peripheral keel. 15

Figure 2. Bayesian tree obtained with the COI gene. Posterior probabilities (above 0.5) are given for each node. Groups are numbered from top downwards from 1 to 5. For each group and for each cruise within a group, one shell is illustrated (numbered from 1 to 7).

Nevertheless, in nearly all studied species of Gemmuloborsonia the degree of prominence of fold and keel decreases with age and therefore it is difficult to predict the definitive shell form. Therefore, despite the fact that it is rather distinct from other species we refrain from description of a new species until additional fully grown specimens become available. Gemmuloborsonia colorata (Sysoev & Bouchet, 2001) comb. nov. (Fig. 3A ­ E) Bathytoma colorata Sysoev & Bouchet, 2001: 294, 296, figs 97 ­ 98 (Vanuatu, NE of Tanna, 198220 S, 1698260 E, 408 ­ 410 m,


N. PUILLANDRE ET A L .

Figure 3. Examined species of Gemmuloborsonia. A­ E. G. colorata (Sysoev & Bouchet, 2001). A. Holotype, MNHN, shell length 42.5 mm. B. Specimen from New Caledonia, Norfolk Ridge, MNHN, shell length 30.0 mm. C. Specimen from French Polynesia, E off Rapa, MNHN, shell length 44.7 mm. D. Specimen from New Caledonia, Lansdowne Bank, MNHN 17849, shell length 45.5 mm. F. Undescribed species, Solomon Islands (stn 2850), MNHN 41918, shell length 25.8 mm. G. G. neocaledonica Sysoev & Bouchet, 1996, specimen from New Caledonia, Norfolk Ridge (stn 2097), MNHN 41921, shell length 25.2 mm. A­E, shells at the same scale; F, G, not to scale.

Expedition MUSORSTOM 8, stn CP982; holotype and paratype MNHN). Material examined: Type material; Lansdowne Bank, 208060 S, 1608230 E, New Caledonia, 490 ­ 550 m (Expedition EBISCO, stn DW 2619, 20 October 2005), one specimen sequenced (MNHN 17849). Remarks: This species was described in the genus Bathytoma (Conidae: Borsoniinae) on the basis of three empty shells from Vanuatu. Due to the lack of data on protoconch and radula the authors did not attribute it to any recognized subgenus of Bathytoma. Later, additional material was collected off New Caledonia (Norfolk Ridge), in the Coral Sea (Chesterfield Plateau) and French Polynesia. The specimen analysed (shell length 45.5 mm) (Fig. 3D, E) was collected alive from the Lansdowne Bank. Conchologically it is very similar to the holotype (Fig. 3A), but differs in the much paler coloration. Similarly coloured specimens have been found in French Polynesia (Fig. 3C), with intermediate ones from New Caledonia (Fig. 3B). The protoconch 16

appears typical of the subfamily Turrinae of Turridae. It is dark brown, multispiral, formed of 3.25 whorls, diameter 790 mm. It consists of c. 1.75 smooth whorls of protoconch I and 1.5 whorls of protoconch II covered with arcuate strongly prosocline ribs not reaching the suture below. In general shape, size and ornamentation it is very similar to the protoconchs illustrated and described for Gemmuloborsonia neocaledonica Sysoev & Bouchet, 1996 and G. moosai Sysoev & Bouchet, 1996. The radula of Gemmuloborsonia colorata (Fig. 4A) is typical for subfamily Turrinae and is extremely similar to that of other examined species of Gemmuloborsonia. It is formed of about 60 rows of teeth, 22 immature, 3.75 mm in length (0.23 of AL). Marginal teeth are duplex, about 210 mm long (0.013 of AL). The central formation is rather short, and strongly notched anteriorly. The cusp is strong and curved in profile. The molecular analysis groups the species unambiguously within Gemmuloborsonia (Fig. 2). The species is widely distributed in the Indo-Pacific from French Polynesia to Vanuatu, New Caledonia and westward to


CRYP TIC S PEC I ES IN CO NOIDEA

Figure 4. Radula of examined Gemmuloborsonia species. A. G. colorata (Sysoev & Bouchet, 2001), MNHN 17849, shell see Fig. 3D ­ E. B. Undescribed species, Solomon Islands (stn 2850), MNHN 41918, shell see Fig. 3F. C, D. G. moosai Sysoev & Bouchet, 1996, Philippines, Aurora 07, stn CP2658, MNHN 41993, shell see Fig. 7D. E, F. G. clandestina sp. nov. E. Paratype, MNHN 41952, shell see Fig. 8E. F. Paratype, MNHN 41943, shell length 18.7 mm. Scale bars 50 mm, for D ­ 10 mm.

17


N. PUILLANDRE ET A L .

Figure 5. CVA for the two groups identified as Gemmuloborsonia moosai, using genetic groups as grouping variable. A. Genetic group 1. B. Genetic group 2. C. Superimposed outlines for minimum ( grey line) and maximum (black line) projections onto the axis are represented.

Ta ble 2. Assignation o f the holotype and the two paratypes of Gemmuloborsonia moosai obtained with CVA.
Locality Holotype Paratype stn CP78 Paratype stn CP118 Indonesia Philippines Philippines G. moosai 0.982 0.0003 0.0823 G. clandestina 0.018 0.9997 0.9177

Material examined: Holotype and nine paratypes in MNHN; see also Table 1. Remarks: Our specimens match the types and were collected in close proximity to the type locality at similar depths. The species is distributed in New Caledonia, Loyalty Islands and the southern New Hebrides arc, at depths of 420 ­ 550 m. Gemmuloborsonia moosai Sysoev & Bouchet, 1996 (Fig. 7A ­ H) Gemmuloborsonia moosai Sysoev & Bouchet, 1996: 82, 84 ­ 85, figs 2C, E, 5A ­ G (E of Palau Jamdena I., Indonesia, 088200 S, 1328110 E, 405 ­ 399 m, Expedition KARUBAR, stn CP59; holotype and 48 paratypes MNHN, 2 paratypes PPPO-LIPI, 2 paratypes NM, 2 paratypes ZMMSU). Material examined: Holotype and 48 paratypes in MNHN; see also Table 1. Remarks: Sysoev & Bouchet (1996) remarked on the high variability of the shell characters, including sculpture, shell outline and relative height of the last whorl. The holotype is one of the most slender specimens examined. Some of the paratypes from Indonesia match well with our specimens collected around the Solomon Islands at similar depths. The species is extremely similar to Gemmuloborsonia clandestina sp. nov. and cannot be distinguished visually. The Fourier analysis of the shell outline showed that in general the specimens of Gemmuloborsonia moosai have slightly narrower and slightly taller last whorls. These differences are easily obscured by the shell sculpture. In addition, there is overlap in characters and not all of specimens can be distinguished by the morphometric analysis. 18

Figure 6. CVA for the two groups identified as Gemmuloborsonia moosai, using cruise of collection as grouping variable. Superimposed outlines for minimum (dotted line) and maximum (black line) projections onto the two principal axes are represented.

Madagascar and Reunion. It is illustrated as Lucerapex indagarotis (Finlay, 1927) on the website http://vieoceane.free.fr/, dedicated to the molluscs of Reunion Island. Gemmuloborsonia neocaledonica Sysoev & Bouchet, 1996 (Fig. 3G) Gemmuloborsonia neocaledonica Sysoev & Bouchet, 1996: 76 ­ 78, figs 1, 2A, D, 3A ­ D (Southern New Caledonia, 248400 S, 1688380 E, 650 m, Expedition CHALCAL 2, stn DW74; holotype and nine paratypes MNHN, one paratype ZMMSU, one paratype NM, one paratype NMNZ).


CRYP TIC S PEC I ES IN CO NOIDEA

Figure 7. A ­ H. Gemmuloborsonia moosai Sysoev & Bouchet, 1996. A­ C CP2658, MNHN 41993, shell length 25.2 mm. E. Coral Sea, EBISCO, stn CP2660, MNHN 41992, shell length 20.9 mm. G. Solomon Islands, Fig. 4C, D. H. Solomon Islands, Salomon 3, stn 2857, MNHN 41927, moosai) shell length 32.2 mm.

. Holotype, MNHN, shell length 32.4 mm. D. Philippines, Aurora 07, DW2546, MNHN 41996, shell length 20.1 mm. F. Philippines, Aurora Salomon 2, stn 2177, MNHN 41922, shell length 28.6 mm. See radula shell length 27.4 mm. I. Gemmuloborsonia clandestina sp. nov. ( paratype of

stn 07, on G.

The two species co-occur in the Philippines and were found in a single dredge haul (stn CP2658). In contrast to the usual phenomenon of character displacement in the zone of overlap between closely related species, these Gemmuloborsonia that co-occurred were very similar. Most of the specimens from this station, which were molecularly identified as G. clandestina new species, were placed in the wrong group (that is G. moosai) by CVA (Table 1). We examined the radulae of both species, including those of sympatric specimens (Fig. 4C ­ F). No significant differences of specific value could be found. The radula of G. moosai (Fig. 4C, D) is typical for the genus. The marginal teeth are duplex, about 145 mm long (0.0184 of AL), while the central formation is rather short and notched anteriorly. The cusp is strong and curved in profile. 19

Shell morphometry of the two paratypes of G. moosai from the Philippines revealed that they belong to Group 3 (G. clandestina) (Table 2). There is a possibility that the Indonesian population of G. moosai (i.e. from the type locality) represents yet another species, separate from Groups 3 and 5. In this case we are dealing with three species, but until material suitably preserved for molecular analysis becomes available from the type locality of G. moosai, we prefer to use this name for specimens in Group 5. Gemmuloborsonia moosai is distributed off the Tanimbar Islands (Banda and Arafura Seas), Indonesia, Philippines (our material ) and Solomon Islands (our material ). The species was also thought to inhabit the Mozambique Channel, although the specimens from this area differ significantly in


N. PUILLANDRE ET A L .

Figure 8. Gemmuloborsonia clandestina sp. nov. A­C. Holotype, MNHN 41937. D­ J. Paratypes, stn CP2658. D. Shell length 24.9 mm, MNHN 41949. E. Shell length 22.4 mm, MNHN 41952. F. Shell length 22.3 mm, MNHN 41953. G. Shell length 26.3 mm, MNHN 41950. H. Shell length 21.4 mm, MNHN 41948. I. Shell length 24.8 mm, MNHN 41947. J. Shell length 24.8 mm, MNHN 41946. H­ J were erroneously attributed to G. moosai with Fourier analysis (see text).

having a rather different shell outline, in particular the much taller last whorl. They are much more different from the specimens from Indonesia and Solomons than those of G. clandestina. We consider that the Mozambique Channel should be excluded from the distributional range area of this species, and that these specimens perhaps constitute a different taxon. Gemmuloborsonia clandestina new species (Fig. 8) Types: Holotype: MNHN 41937; 13 paratypes: MNHN 41941 ­ 41953. Type locality: E of Luzon Island, Philippines, Philippine Sea, 158580 N, 121849.10 E, 422 ­ 431 m (Expedition Aurora, stn CP2658). 20

Material examined: Type material; E of Luzon Island, Philippines, 16800.90 N, 121851.20 E, 342 ­ 348 m (Expedition Aurora, stn CP2657), one live (MNHN 41929); 15856.40 N, 121848.90 E, 460 ­ 480 m (Expedition Aurora, stn CP2659), one live (MNHN 41936); 15852.20 N, 121848.80 E, 506 ­ 542 m (Expedition Aurora, stn CP2660), three live (MNHN 41938 ­ 41940); 15801.40 N, 121844.80 E, 431 ­ 493 m (Expedition Aurora, stn CP2673), one live (MNHN 41932); 15804.10 N, 121841.10 E, 368 ­ 442 m (Expedition Aurora, stn CP2707), three live (MNHN 41933 ­ 41935); 158190 N, 121833.90 E, 300 ­ 318 m (Expedition Aurora, stn CP2727), one live (MNHN 41931); 15858.10 N, 121849.20 E, 418 ­ 456 m (Expedition Aurora, stn CP2744), one live (MNHN 41930); Philippines, 138490 N, 1208280 E, 441 ­ 550 m (Expedition MUSORSTOM 2, stn CP78), two dead ( paratypes of Gemmuloborsonia moosai Sysoev &


CRYP TIC S PEC I ES IN CO NOIDEA Bouchet, 1996); Philippines, 118580 N, 1218060 E, 448 ­ 466 m (Expedition MUSORSTOM 3, stn CP118), one dead ( paratype of Gemmuloborsonia moosai Sysoev & Bouchet, 1996). Etymology: clandestinus ­ Latin, hidden, concealed, with reference to the extreme similarity of the species to G. moosai. Description (holotype): Shell elongate-biconic, strong, mediumsized, slightly glossy, covered by thin light yellowish smooth periostracum. Protoconch eroded, brown, multispiral, of c. 2.6 whorls, diameter 670 mm. Transition from protoconch to teleoconch clearly marked by change in colour. Teleoconch of 9.75 low whorls separated by shallow-channelled suture. Whorls bear well-developed subsutural fold and peripheral keel. Subsutural fold appears on first teleoconch whorls and is narrower than keel, although its relative width progressively enlarges with shell growth. Subsutural fold covered by rounded blunt gemmules, which occupy whole fold on early whorls, but on last and most of penultimate whorl are confined to narrowing cord in middle of fold which becomes more flat on last whorl; there are 39 gemmules on last whorl and 28 on penultimate whorl. On seventh teleoconch whorl the cord at upper edge of subsutural fold appears and over the extent of one whorl this cord is split in two, which become more convex and well developed on last whorl. Peripheral keel bears longitudinally elongate gemmules that are arcuate on last and penultimate whorls; there are 36 gemmules on last whorl and 28 on penultimate whorl. Interspace between subsutural fold and peripheral keel is very narrow and smooth on upper 5 teleoconch whorls; later there appears an initially narrow cord, becoming progressively broader and more pronounced. On last whorl the interspace between fold and keel is broad with three narrow but distinct spiral cords. Body whorl occupies 0.61 of shell length. Periphery of whorl below keel, shell base and canal are covered by narrow granulated cords slightly differing in width. There are in total 22 such cords, with interspaces not exceeding cord width. Aperture is narrow and its width slowly decreases to a broad and obliquely truncated canal. Inner lip smooth and convex in its parietal part and nearly straight in columellar part; covered with thin off-white glossy callus. Weak columellar pleat encircles columellar obliquely. Outer lip projects strongly forward below anal sinus; sinus is deep, U-shaped, slightly adapically directed. Shell length 25.3 mm, body whorl length 15.6 mm, aperture length 8.5 mm, canal length 3.0 mm and shell diameter 8.5 mm. Remarks: Although there is clear conchological similarity among all specimens that are attributed to the new species on the basis of molecular data, G. clandestina is rather variable in terms of shell outline, especially in terms of shell slenderness. The type locality is remarkable in this respect, since the variability within the type series equals the maximal variability within the species (shell width/shell length ratio varies from 0.34 to 0.40, average 0.36 + 0.01, n ¼ 24). The radula was examined in three specimens, including two paratypes (Fig. 4E, F). The radula is formed of about 48 ( paratype MNHN 41943) to 60 rows of teeth (MNHN 41934), 12 ­ 20 nascent, 2.51 ­ 2.87 mm in length (0.32 ­ 0.34 of AL). The marginal teeth are duplex, about 105 ­ 137 mm long (1.75 ­ 1.37% of AL) and the central formation is rather short, with the anterior border indistinct and fused with the membrane. The cusp is strong and curved in profile. The species is extremely similar conchologically to G. moosai and some specimens cannot be distinguished even by morphometric analysis. For discussion and comparison see Remarks on the previous species. The holotype of the new species is distinguished from the holotype of G. moosai in having a smaller and broader shell. Both species can be readily distinguished by COI sequences. 21 The species has so far been found only in the Philippines, at depths of 342 ­ 542 m.

DISCUSSION
The conventional practice of distinguishing species is to find the gaps in the morphological continuum. Therefore, until discrete differences (at least in some of the parameters) are found, entities are usually not considered as separate species. Prior to use of molecular techniques the status of allopatric forms was, in reality, decided arbitrarily. In our analysis, two discrete entities, revealed by the DNA analysis, are not morphologically distinguishable. Before conducting time-consuming Fourier analysis we tried the more standard morphometric parameters that are operational for different groups of marine gastropods (Bouchet & Kantor, 2004; Kantor & Bouchet, 2007), but we were not able to delimit the species. Similarly, the radular morphology and gross anatomy did not reveal any significant differences. Finally, Fourier analysis allowed us to attribute most, but not all specimens to one of the two groups. However, COI sequences clearly suggested that two different species (Gemmuloborsonia moosai and Gemmuloborsonia clandestina) were included in what was considered before as single `G. moosai'. Genetic distances between G. moosai and G. clandestina are similar to those found between others unquestioned species included in Gemmuloborsonia (Fig. 2). Although our conclusion is based on a single gene, it is unlikely that there is gene flow between these two species, because they co-occur sympatrically in the Philippines. We also sequenced a nuclear gene (28S rRNA; results not shown) to test if the genetic differences found in the mitochondrial marker were supported, thus reflecting the species boundaries (see e.g. Nichols, 2001; Funk & Omland, 2003), but all the specimens shared the same 28S sequence. A more variable nuclear marker would test the hypothesis proposed from the results using the COI gene. The described situation with G. moosai and G. clandestina is presently very uncommon for molluscs. Although species are now commonly delimited using molecular data (e.g. Meyer & Paulay, 2005; Mikkelsen et al., 2007; Campbell et al., 2008), species limits are most of the time illustrated by morphological differences, known a priori or found a posteriori. Even when cryptic species are revealed by DNA surveys, morphological differences can usually be identified. In several cases, morphologically indistinguishable entities have been recognized as separate species in molluscs. One is the recognition of two sister pairs of species of Bulla (Bullidae) (Malaquias & Reid, 2008). The other is a recognition of a cryptic speciation of the genus Bostrycapulus Olsson & Harbison, 1953 (Calyptraeidae) (Collin, 2005). In both cases the specific status was proved on the basis of molecular data for allopatric forms. Using COI sequences Gittenberger & Gittenberger (2006) demonstrated the existence of several morphologically indistinguishable parasitic species of the genus Leptoconchus (Coralliophilidae). In this case, sympatric species inhabited different species of the hosts, scleractinian corals of the family Fungiidae. The authors refrained from formal introduction of 14 recognized new species. Our case is different from those mentioned above. Firstly, the unexpected diversity was found without any clue from conchology, anatomy or ecology. Secondly, comparison of syntopic specimens of G. moosai and G. clandestina, found in the same dredge haul, revealed that they are more similar to each other than the specimens of allopatric populations. Most of specimens from that station which molecularly were shown to belong to G. clandestina were placed in the wrong group (that is G. moosai) by CVA (Table 1). Usually the situation is the opposite and the


N. PUILLANDRE ET A L . sympatric specimens of different species are easier to recognize than those from allopatric populations. Although we were unable to demonstrate that these species are morphologically discrete, we feel that it is necessary formally to recognize two clades within what was previously considered to be G. moosai, as separate species, even if DNA constitutes the only tool to separate them. We agree with Collin (2005) that there is no theoretical reason to expect that mechanisms of speciation should always result in species that can be distinguished visually, especially for recent speciation events. For the practical purpose of the discrimination of G. moosai from G. clandestina, only the specimens from the Philippines constitute a problem (as far as our sampling revealed). All of the specimens collected in the Solomon Islands and Chesterfield Plateau belonged to G. moosai and 92.6% of the specimens collected in the Philippines were G. clandestina (only 2.7% of the specimens were not correctly assigned). Morphological characters failed to distinguish 21.6% of specimens. Although discriminating morphological (or anatomical ) characters were not found in Gemmuloborsonia species, this does not mean that some discrete differences might not be identified, most probably by detailed anatomical study. At the same time such an intensive search may not prove to be operational. Empty shells cannot be identified with certainty (at least from Philippines). At present, even the examination of the radulae with the scanning electron microscope (not to mention serial histological sectioning) is probably more costly and labour intensive than molecular sequencing. Molecular analysis is becoming more and more a standard procedure with rapidly decreasing costs that can be performed by personnel without taxonomic expertise. Current developments in malacology clearly demonstrate that the recognition of cryptic species, indistinguishable by shell characters and anatomy, is becoming a more common phenomenon. This highlights a general problem of traditional taxonomic malacology ­ the absence of a reliable link to the overwhelming majority of existing name-bearing types of molluscs, which are represented by material unsuitable for molecular sequencing. At present, when describing new species the preference when designating the types is given to wellpreserved adult shells. It is advisable that in future preference should be given to specimens for which either a sequence exists, or at least of which the samples are suitable for future molecular analysis. directed by Guillaume Lecointre. We are also pleased to thank Michel Baylac and Allowen Evin (Plateforme Morphometrie) for their help in the morphological analysis. Philippe Bouchet helped a lot with his taxonomic expertise at all stages of the project. The work was done during a visiting curatorship of Yu.I.K. at MNHN and he expresses his thanks to Virginie ´ Heros, Barbara Buge, Philippe Maestrati and Pierre Lozouet for assistance during his stay in Paris.

R E FE RE NC ES
BAYLAC, M. & FRIESS, M. 2005. Fourier descriptors, Procrustes superimpositions and data dimensionality: an example of cranial shape analysis in modern human populations. In: Modern morphometrics in physical anthropology (D.E. Slice ed.), pp. 145 ­ 165. Kluwer, Chicago. BOUCHET, P. & KANTOR, YU.I. 2004. New Caledonia: the major center of biodiversity for volutomitrid mollusks (Mollusca: Neogastropoda: Volutomitridae). Systematics and Biodiversity, 1: 467 ­ 502. BOUCHET, P., LOZOUET, P., MAESTRATI, P. & HEROS, V. 2002. Assessing the magnitude of species richness in tropical marine environments: exceptionally high numbers of molluscs at a New Caledonia site. Biological Journal of the Linnean Society, 75: 421 ­ 436. CAMPBELL, D.C., JOHNSON, P.D., WILLIAMS, J.D., RINDSBERG, A.K., SERB, J.M., SMALL, K.K. & LYDEARD, C. 2008. Identification of `extinct' freshwater mussel species using DNA barcoding. Molecular Ecology Resources, 8: 711 ­ 724. COLLIN, R. 2005. Development, phylogeny, and taxonomy of Bostrycapulus (Caenogastropoda: Calyptraeidae), an ancient cryptic radiation. Zoological Journal of the Linnean Society, 144: 75 ­ 101. DOMMERGUES, E., DOMMERGUES, J.-L., MAGNIEZ, F., NEIGE, P. & VERRECCHIA, E.P. 2003. Geometric measurement analysis versus Fourier series analysis for shape characterization using the gastropod shell (Trivia) as an example. Mathematical Geology, 35: 887 ­ 894. DUDA, T.F., BOLIN, M.B., MEYER, C.P. & KOHN, A.J. 2008. Hidden diversity in a hyperdiverse gastropod genus: discovery of previously unidentified members of a Conus species complex. Molecular Phylogenetics and Evolution, 49: 867 ­ 876. FOLMER, O., BLACK, M., HOEH, W., LUTZ, R. & VRIJENHOEK, R. 1994. DNA primers for amplification of mitochondrial cytochrome c oxidase subunit I from diverse metazoan invertebrates. Molecular Marine Biology and Biotechnology, 3: 294 ­ 299. FRETTER, V. & GRAHAM, A. 1985. The prosobranch molluscs of Britain and Denmark. Part 8 ­ Neogastropoda. Journal of Molluscan Studies, Suppl. 15: 435 ­ 556. FUNK, D.J. & OMLAND, K.E. 2003. Species-level paraphyly and polyphyly: frequency, causes, and consequences, with insights from animal mitochondrial DNA. Annual Review of Ecological Systems, 34: 397 ­ 423. GITTENBERGER, A. & GITTENBERGER, E. 2006. A largely cryptic, adaptive radiation of parasitic snails: sibling species in Leptoconchus (Gastropoda: Caenogastropoda: Coralliophilidae), associated with specific coral hosts (Scleractinia: Fungiidae). In: The evolutionary history of parasitic gastropods and their coral hosts in the Indo-Pacific (A. Gittenberger ed.). PhD dissertation, Leiden University, pp. 62 ­ 87. HEBERT, P.D.N., CYWINSKA, A., BALL, S.L. & DEWAARD, J.R. 2003. Biological identifications through DNA Barcodes. Proceedings of the Royal Society of London B, 270: 313 ­ 321. HUELSENBECK, J.P., RONQUIST, F. & HALL, B. 2001. MrBayes: bayesian inference of phylogeny. Bioinformatics, 17: 754 ­ 755. KANTOR, YU.I. 2006. On the morphology and homology of the `central tooth' in the radulae of Turrinae (Conoidea: Turridae). Ruthenica, 16: 47 ­ 52. KANTOR, YU.I. & BOUCHET, P. 2007. Out of Australia: Belloliva (Neogastropoda: Olividae) in the Coral Sea and New Caledonia. American Malacological Bulletin, 22: 22 ­ 73.

AC KNOW LE DG E M EN TS
The material was collected during several deep-sea cruises of the Tropical Deep Sea Benthos programme as follows. Philippines: the AURORA 2007 cruise on board M/V DABFAR associated with the National Museum of the Philippines (NMP, co-PI Marivene Manuel ), MNHN (co-PI Philippe Bouchet) and BFAR, made possible through a grant from the Lounsbery Foundation. Coral Sea, Norfolk Ridge and Solomon Islands: the EBISCO (PI Philippe Bouchet), SALOMON 1 (PI Bertrand Richer de Forges), SALOMON 2 (PI Philippe Bouchet), SALOMONBOA 3 (PI Sarah Samadi) and NORFOLK 2 (PI Sarah Samadi) cruises, on board R/V Alis deployed from Noumea by the Institut de Recherche pour le Developpement (IRD). Marie-Catherine Boisselier and Ellen Strong are thanked for their role in molecular sampling during these expeditions. This work was supported by the Consortium National de Recherche en Genomique and the Service de Systematique Moleculaire of the Museum National d'Histoire Naturelle (IFR 101). It is part of the agreement no. 2005/67 between the Genoscope and the Museum National d'Histoire Naturelle on the project `Macrophylogeny of Life' 22


CRYP TIC S PEC I ES IN CO NOIDEA
KANTOR, YU.I., PUILLANDRE, N., OLIVERA, B. M. & BOUCHET, P. 2008. Morphological proxies for taxonomic decision in turrids (Mollusca, Neogastropoda): a test of the value of shell and radula characters using molecular data. Zoological Science, 25: 1156 ­ 1170. KNOWLTON, N. 1993. Sibling species in the sea. Annual Review of Ecology and Systematics, 24: 189 ­ 216. KNOWLTON, N. 2000. Molecular genetic analyses of species boundaries in the sea. Hydrobiologia, 420: 73 ­ 90. KUMAR, S., TAMURA, K. & NEI, M. 2004. MEGA3: integrated software for molecular evolutionary analysis and sequence alignment. Briefings in Bioinformatics, 5: 150 ­ 163. MALAQUIAS, M.A.E. & REID, D.G. 2008. Systematic revision of the living species of Bullidae (Mollusca: Gastropoda: Cephalaspidea), with a molecular phylogenetic analysis. Zoological Journal of the Linnean Society, 153: 453 ­ 543. MEYER, P. C. & PAULAY, G. 2005. DNA barcoding: error rates based on comprehensive sampling. PLoS Biology, 3: 1 ­ 10. MIKKELSEN, N., SCHANDER, C. & WILLASSEN, E. 2007. Local scale DNA barcoding of bivalves (Mollusca): a case study. Zoologica Scripta, 36: 455 ­ 463. MONTI, L., BAYLAC, M. & LALANNE-CASSOU, B. 2001. Elliptic Fourier analysis of the form of genitalia in two Spodoptera species and their hybrids (Lepidoptera: Noctuidae). Biological Journal of the Linnean Society, 72: 391 ­ 400. NICHOLS, R. 2001. Gene trees and species trees are not the same. Trends in Ecology and Evolution, 16: 358 ­ 364. PUILLANDRE, N., SAMADI, S., BOISSELIER, M.-C., SYSOEV, A.V., KANTOR, Y.I., CRUAUD, C., COULOUX, A. & BOUCHET, P. 2008. Starting to unravel the toxoglossan knot: molecular phylogeny of the `turrids' (Neogastropoda: Conoidea). Molecular Phylogenetics and Evolution, 47: 1122 ­ 1134. PUILLANDRE, N., BAYLAC, M., BOISSELIER, M.-C., CRUAUD, C. & SAMADI, S. 2009. An integrative approach to species delimitation in Benthomangelia (Mollusca: Conoidea). Biological Journal of the Linnean Society, 96: 696 ­ 708. REID, D.G., LAL, K., MACKENZIE-DODDS, J., KALIGIS, F., LITTLEWOOD, D.T.J. & WILLIAMS, S.T. 2006. Comparative phylogeography and species boundaries in Echinolittorina snails in the central Indo-West Pacific. Journal of Biogeography, 33: 990 ­ 1006. ROHLF, F.J. 1996. TpsDig. State University of New-York at Stony Brook, http://life.bio.sunysb.edu/morph/. SAMADI, S. & BARBEROUSSE, A. 2006. The tree, the network, and the species. Biological Journal of the Linnean Society, 89: 509 ­ 521. SHUTO, T. 1989. Gemmuloborsonia, a new genus of the family Turridae (Gastropoda) from the Plio-Pleistocene Cabatuan Formation, northwest Luzon. Transactions and Proceedings of the Palaeontological Society of Japan, New Series, 153: 48 ­ 54. SMITH, A.M., FISHER, B.L. & HEBERT, P.D.N. 2005. DNA barcoding for effective biodiversity assessment of a hyperdiverse arthropod group: the ants of Madagascar. Philosophical Transactions of the Royal Society, Series B, 360: 1825 ­ 1834. SYSOEV, A.V. & BOUCHET, P. 1996. Taxonomic reevaluation of Gemmuloborsonia Shuto, 1989 (Gastropoda: Conoidea), with a description of new Recent deep-water species. Journal of Molluscan Studies, 62: 75 ­ 87. SYSOEV, A.V. & BOUCHET, P. 2001. New and uncommon turriform gastropods (Gastropoda: Conoidea) from the South-West Pacific. In: Tropical Deep-Sea Benthos. Vol. 22 (P. Bouchet & B.A. Marshall, eds), pp. 271 ­ 320. Me´moires du Muse´um National d'Histoire Naturelle, Vol. 185, 1 ­ 406. SYSOEV, A.V. & KANTOR, YU.I. 1990. A new genus and species of `Cochlespira-like' turrids (Gastropoda, Toxoglossa, Turridae). Apex, 5: 1 ­ 6. TIPPETT, D.L. 2006. The genus Strictispira in the western Atlantic (Gastropoda: Conoidea). Malacologia, 48: 43 ­ 64. WILLIAMS, S.T. & REID, D.G. 2004. Speciation and diversity on tropical rocky shores: a global phylogeny of snails of the genus Echinolittorina. Evolution, 58: 2227 ­ 2251.

23