The unknown bathyal of the Canaries: new species and new records of deep-sea Mollusca

ABSTRACT A set of seven dredge hauls, between 195-215 m and 655-660 m deep on the NW slope of Gran Canaria (Canary Islands, Spain), recovered 15 000 specimens belonging to 295 species of molluscs. Of these, 254 are identified at the species level. Only 47 species, totalling 867 specimens, were live collected, which amounts to 84% of the species and 94% of the specimens represented only by shells. The dredges DW 133 (shallowest) and DW 130 (deepest) hold the highest number of species and abundance, representing about 90% of the material. Fifty-one species are new records for Canarian waters; of these, 23 are new for Spanish waters overall and three are the first reference in eastern Atlantic waters. Another 13 species, in the genera Mikro Warén, 1996, Discaclis Moolenbeek & Warén, 1987, Mucronalia A. Adams, 1860, Marginella Lamarck, 1799, Dentimargo Cossmann, 1899, Prunum Herrmannsen, 1852, Microvoluta Angas, 1877, Spirotropis G.O. Sars, 1878, Gymnobela Verrill, 1884, Mitromorpha Carpenter, 1865, Orbitestella Iredale, 1917 and Liostomia G.O. Sars, 1878, are described as new, most of them from the deepest haul at 655-660 m. Anatoma richardi (Dautzenberg & Fischer, 1896) is restored as valid species and the identification of Canarian specimens as Anatoma tenuis (Jeffreys, 1877) is disputed. Ancistrobasis lavaleyei Hoffman & Freiwald, 2017 is synonymized with A. reticulata (Philippi, 1844). Pleurotomella megalembryon (Dautzenberg & Fischer, 1896), type species of Azorilla Nordsieck, 1968, is assigned to Teretia Norman, 1888; Teretia strongyla (Dall, 1927) is synonymized with T. megalembryon, and Azorilla with Teretia. Pleurotoma teres Reeve, 1844 is selected under ICZN Art. 70.3.2 as type species of Teres Bucquoy, Dautzenberg & Dollfus 1883 and of its substitute name Teretia, discarding therefore Pleurotoma anceps Eichwald, 1830.


INTRODUCTION
The waters surrounding the Canary Islands are believed to be the most promising of all Spanish waters in terms of new discoveries in the benthic fauna (Gofas et al. 2017), taking into account that recent exploration of the benthos still yields new species or new records. Of the 127 species of molluscs currently listed as endemic of the archipelago, nearly half were described in this century and only six had been described prior to 1979.
This may be still more so in the bathyal zone, where exploration is technically demanding and requires the intervention of properly equipped research vessels. Here we report on the Mollusca from a suite of samples on the north-western slope of Gran Canaria, collected at the beginning of the SEA-MOUNT 2 campaign, aimed to the Central North Atlantic seamounts south of the Azores.

A BRIEF HISTORy OF THE ExPLORATION OF THE DEEP CIRCUMCANARIAN BENTHOS
The first insight into the deep-water benthos around the Canaries was an isolated station of the H.M.S. "Challenger" in 1873 (sta. 85, west of La Palma in 1125 fathoms [= 2057 m] in which only 20 species were collected (Smith 1885; Watson 1886). Six of them were described as new and seven remained unidentified. The expedition of the French vessel "Talisman" in 1883, aimed to a systematic sampling from the Bay of Biscay to Morocco and Mauritania, made several dredgings in Canarian waters ( Fig. 1) and provided the first significant contribution. All the zoological groups were studied and published monographs (for molluscs, Locard 1897-1898) were for more than a century the basic reference for the deep benthos of this region. In 1968, the British R.R.S. "Discovery" carried out a campaign in North West Africa to study the upwelling area (National Institute of Oceanography, 1968), and carried out several dredging operations in Canarian waters under the supervision of malacologist John Allen. However, little of the material has been systematically studied (for example, cumaceans by Corbera et al. 2001), and for molluscs, there are only isolated records in reports with a broader scope (Bouchet & Warén 1980, 1985, 1986, 1993Allen & Morgan 1981). At the end of the 20th century, the largest sampling effort carried out in the area was undoubtedly the Dutch CANCAP project, with several campaigns of the vessel "Tydeman", with some 1360 operations carried out deeper than 200 meters (van der Land 1987). The published results cover hydrozoans (Medel & Vervoort 1998Ansín Agís et al. 2001;Vervoort 2006), brachiopods (Logan 1983;1988), crustaceans (Holthuis 1984) and a few families of molluscs (De Boer 1985;Moolenbeek & Warén 1987;van der Linden 1995van der Linden , 1998Goud & Neefs 1996;Hoenselaar & Goud 1998;van Aartsen et al. 1998;Dijkstra & Goud 2002;Verhecken 2007). In addition to this, there were numerous campaigns, mainly in deep water, made in the 1990s by the research vessel "Taliarte" of the Canarian Institute of Marine Sciences (lCCM) of the Cabildo Insular of Gran Canaria. From this material, there is a comprehensive report on the decapod crustaceans (González Pérez 1995), but for molluscs we could only trace the isolated description of a marginellid species (Pérez-Dionis et al. 2009).
In this century, the INDEMARES project, intended to document offshore prospective marine protected areas for the EU's Natura 2000 network, targeted two deep-water areas, one to the southeast of Fuerteventura and Lanzarote and the other on Banco de la Concepción seamount (Almón et al. 2014a, b) but so far virtually nothing has filtered of the results and only those general accounts were published, recording only three molluscan species.
Those cumulated efforts have so far resulted in documenting 195 molluscan species living normally below 500 m (the "DeepSea" context in the World Register of Marine Species, Glover et al. 2017), of which ten are spurious and seven are not known outside Canarian jurisdictional waters (Gofas et al. 2017, supplementary material). Still, the SEAMOUNT 2 sampling is the only source for Gran Canaria and the islands of Tenerife and Gomera remain virtually unsampled in the deep sea.

MATERIAL AND METHODS
The SEAMOUNT 2 Oceanographic Campaign was held in January-February 1993 aboard R/V Le Suroît, with the second author as chief scientist and with the main objective of collecting information on larval dispersal capacity for colonization and speciation in the main underwater mountains of the Middle North Atlantic. During the campaign a total of 129 benthic samples were collected, of which 94 were obtained by Warén dredge (DW). The samples from off NW of Gran Canaria ( Fig. 1; Table 1), representing a total of 10 dredges in a depth range between 195 m and 680 m, are studied herein.
These samples were sieved on board in sea water on a column of 10, 5, 3 and 0.5 mm, sorted down to the higher zoological groups (phylum or class), and preserved for later identification and taxonomic work. Unsorted sediment samples were also preserved for subsequent separation in the laboratory of the finer fractions between 5 and 0.5 mm. The resulting mollusc material, including the types of new species, is deposited in Muséum national d'Histoire naturelle (MNHN).
The results of individual hauls were very uneven in terms of abundance and quality of the material, with the Warén dredge giving better results on rough bottoms than the epibenthic sled (which got lost in DE125). Dredge hauls that collected adequate material in terms of abundance are DW126 (in 345 m), DW128 (470-485 m), DW130 (655-660 m) and .
The samples used in this work were mostly preserved dry, except for some specimens fixed in ethanol when separated on board. Most of the material consists of shells, the proportion of live specimens being very low. The samples were fully separated at the species level, individually labelled and preserved in non-acidic glass tubes. Empty shells (hereafter, sh.) and live-taken specimens (hereafter, spm.) are listed separately.
The validity of the names used and their currently accepted familial placement was contrasted with the World Register of Marine Species (WoRMS Editorial Board 2018) using the "Match Taxa" tool therein.
The type of larval development was assessed from the characteristics of the protoconch, as detailed by Jablonski & Lutz (1980). A multispiral protoconch with differentiated protoconch I and protoconch II is taken as indicating planktotrophic development and this also indicates a considerable potential for dispersal at the larval stage. Conversely, a comparatively large, paucispiral protoconch without a differentiated protoconch II is taken as indicative of a non-planktotrophic larval stage, with a limited potential for dispersal.

SySTEMATIC PART
A total of about 15 000 specimens belonging to 295 species have been identified. Of these, 254 species, totalling 14 052 individuals, have been either identified at the species level or recognized as undescribed. Unidentified specimens at the species level include juveniles or specimens in poor condition which do not allow to observe diagnostic characters, and some which belong to genera or families whose taxonomy is problematic (Cocculiniformia Haszprunar, 1987, Eulimidae Philippi, 1853 for example).
Only 47 species, totalling 867 specimens, are represented by at least one specimen collected alive, which amounts to 84% of the species and 94% of the specimens represented only by shells.
The dredges DW 133 (shallowest) and DW 130 (deepest) are the ones with the highest number of species and abundance, representing about 90% of the total material.
Confronting the results of this study with the Spanish List of Marine Species recorded in Canary Island waters (Gofas et al. 2017) revealed 51 species as new records for Canarian waters. Of these, 23 are new records for Spanish waters overall and three are first references in eastern Atlantic waters. This count of new records do not include species already known from the Canary Islands under another name, for which we provide revised identifications. Surprisingly, only 11 of the species found in our material species (3.7% of the total) are listed as Canarian endemics in the checklist.
Another thirteen species seem new to science, most of them from the deepest dredge haul in 660 meters.
Hereafter we describe the species which are considered as new, provide taxonomic notes on some previously recorded species which we believe to have been misidentified or misplaced generically, and list all the newly recorded species. Our specimens have the suture of the last whorl either attached to the abapical keel of the selenizone, or dropping at some distance below this keel (leaving exposed a part of the previous whorl below the keel of the selenizone called "sutsel" by Geiger), with all transitional constructions. This questions the taxonomic value of this character, given by Geiger as a diagnostic of A. tenuis vs A. tenuisculpta (Seguenza, 1880). The latter species, originally described as a fossil from the Pliocene of southern Italy, is briefly compared to A. tenuis by Geiger (2012Geiger ( : 1106Geiger ( , 1116, and stated to differ, in addition to the sutsel in having a "longer teleoconch I (0.75 vs 0.5 whorl)" and "a wider umbilicus". Nonetheless in the description provided for A. tenuisculpta, teleoconch I is reported to be 0.5-0.85 whorl so that this difference from A. tenuis is not diagnostic. Our own Canarian specimens have about 0.75 whorl of teleoconch I (Fig. 3E). A further feature of the early teleoconch of both A. tenuisculpta as understood by Geiger (2012) and our specimens, is the presence of a single spiral cord on the outer third of teleoconch I, abutting on the beginning of the selenizone (Fig. 3D). This feature is found also on other species (e.g. A. aspera) and is stable on all the specimens we examined, but the lectotype of Scissurella tenuis shows only strong, arched axial ribs fading before reaching the suture, no spiral cord (Fig. 3E) and the teleoconch 1 extends hardly more than 0.5 whorl.
Our specimens are most likely conspecific with those reported by Geiger around the Canaries, Azores and Cape Verde Islands in the upper bathyal (310-1340 m), and the same species is also known to us from Galicia Bank off the NW Iberian Peninsula, but these may not be correctly identified as Anatoma tenuis, which has an abyssal type locality south of Greenland and has a discrepant morphology of the early teleoconch.
The fossil neotype of A. tenuisculpta has the last whorl considerably narrower and dropping more distinctly down from the suture, than in any of the Recent specimens here discussed. The specimens from Norway illustrated by Geiger as A. tenuis culpta have a much flatter spire and do not show essential difference with the Canarian specimens, apart from size and a somewhat less expanded last whorl. Conversely, Scissurella richardi Dautzen berg & H. Fischer, 1896, described from 1300-1396 m off the Azores is certainly the same. For the time being, we find premature to assume the synonymy of Anatoma richardi with either A. tenuis or A. tenuisculpta and propose to use the former name for the Recent Macaronesian species.  description Shell minute, globular in shape, with a moderately high, somewhat cyrtoconoid spire and a distinct umbilicus. Protoconch proportionally very large, egg shaped with no sign of coiling, with a maximum diameter of 270 µm, separated from the teleoconch by a distinct scar; protoconch surface with a microscopically pitted texture seen only at high magnification under the SEM. Teleoconch of about 1 3/4 whorl. Spire whorls with a weak spiral cordlet running along the suture, a very distinct adapical shoulder delimited by a sharp keel, channelled between the suture and the keel and slightly convex below it. Surface mostly smooth except for growth lines, a cluster of 4-6 spiral cordlets below the periphery of the last whorl, and another cluster of three stronger spiral cords surrounding the umbilicus. Aperture with a somewhat rhomboidal shape, with the adapical part definitely angled. Outer lip rather thick but simple, orthocline; parietal area without any callosity, columellar edge of the aperture increasing in thickness abapically. Umbilicus open, deep and relatively broad, inside with axial growth lines of a very rough texture, and with an additional spiral cord situated deep inside and abutting into the abapical part of the columella. Colour opaque white, with a somewhat pearly aspect, the protoconch slightly yellowish. Maximum diameter up to 0.8 mm (holotype 0.84 mm height × 0.80 mm diameter).

reMArks
The generic assignment of this species is tentative; shared characters with the type species Mikro globulus Warén, 1996 are the minute size, the protoconch with a rough but not distinctly sculptured surface, the presence of a distinct adapical keel at a short distance from the suture, and of a distinct spiral ridge inside the umbilicus. Lopheliella Hoffmann, van Heugten & Lavaleye, 2008 is similar in shape but has a distinct honeycomb sculpture on the protoconch which is not seen here; Lopheliella also either lacks an internal ridge inside the umbilicus or has it very close to its edge, not far inside as in Mikro globulus. Lopheliella species are also considerably larger, with an adult size of 2 to 3 mm.
Mikro oviceps n. sp. is unique in having a protoconch which shows no sign of coiling, contrary to M. globulus and M. hat tonensis Hoffman, van Heugten & Lavaleye, 2010 which were also collected in the same dredge haul.
Mikro globulus (Fig. 4I-L) is similar in size but has a higher spire, not cyrtoconoid. It lacks the clusters of spirals on the periphery and the abapical part of the last whorl; the subsutural shoulder disappears on the later whorls, and the umbilicus is narrower with only one periumbilical ridge, and no spirals inside. Mikro hattonensis ( Fig. 4M-P) is more similar to M. oviceps n. sp. in outline, but also lacks the spirals on the last whorl and is more broadly umbilicate. Both M. globulus and M. hattonensis have a protoconch as usual in Vetigastropoda i.e. coiled with hardly more than half a whorl ( Fig. 4L, P). Both also have been found in the same sample as M. oviceps n. sp. and are new records for Spanish waters (see below). Warén, 1996 ( reMArks New to Spanish waters. Compared to the paratypes from off Iceland illustrated in Warén (1996), our specimens are larger (1.5 mm vs 1 mm), with flatter whorls. Nevertheless the subsutural keel, very conspicuous at the beginning of the teleoconch and fading out later, and the narrow umbilicus bound by a deeply set spiral ridge, are similar. Considering that many other species (e.g. Rugulina fragilis (G. O. Sars, 1878), Ancistrobasis reticulata (Philippi, 1844), Cantrainea spp., Larsenia scalaroides Warén, 1989, among others) are shared with the assemblage described by Warén (1996), we tentatively consider them conspecific. reMArks New to Spanish waters. The figured specimens conform quite exactly with the original description of this species, from a depth of 796 m at 58°N in the northern Atlantic, including the shagreened microsculpture in the subsutural ramp and the umbilicus, bound by a narrow cord and also with a single cord inside. Another of our specimens has two cords inside the umbilicus.    Warén & Bouchet (1990) from Galicia Bank, in 985-100 m, and Warén (1991) from south of Iceland, in 970 m. It is actually quite well represented in our own material from Galicia Bank and from the Meteor and Atlantis seamounts, but the record here represents the southernmost known occurrence in the eastern Atlantic. The Western Atlantic A. costulata (Watson, 1879) described from off Culebra I., 390 fathoms, may be the same species but this has not been assessed.

Mikro hattonensis
Hoffman & Freiwald (2017) described Ancistrobasis lavaleyei based on a juvenile shell (1.43 × 1.89 mm) of A. reticulata collected on Coral Patch Seamount (34°58.00'N, 11°57.30'W, 1050 m) and included as a paratype an eroded shell from Galicia Bank. They mentioned that A. reticulata "is much larger, up to 10 mm; it has more than 10 spiral ribs that form a reticulated structure with nodes; and it lacks the angular outline on the first whorl." Ancistrobasis reticulata does have an angular first whorl of teleoconch (Fig. 6C herein and Warén 1991: fig. 1A) and the number of spirals gradually increases from one on the first whorl to about ten on specimens approaching full size about 10 mm. Based on the Canarian specimens and on additional material collected by the SEAMOUNT 2 expedition on the Meteor group of seamounts, we cannot see any grounds for distinguishing more than one species of recent Ancistrobasis in the North Atlantic, and treat A. lavaleyei as a synonym of A. reticulata.  reMArks This is the first report of this species after its original description from off NW Morocco in 529 m. It was diagnosed by the occurrence of a spiral sculpture whereas the interspaces between ribs are smooth in the similar Margarites laminarum ( Palazzia ausonia (Palazzi, 1988) MAteriAl exAMined. -1 sh., DW130.
reMArk Originally described from the Tyrrhenian Sea in 500 m depth, subsequently reported from off Iceland and Norway (Warén, 1991); new record for the Canaries.   van Aartsen, Bogi & Giusti, 1989 ( reMArk This species is readily recognized by its protoconch with an unusual reticulate sculpture. Originally described from 320-440 m in the Tyrrhenian Sea, new to Spanish waters. Family vAnikoridAe Gray, 1840

Larsenia scalaroides
MAteriAl exAMined. -100 sh., DW130. reMArks The occurrence of this species, originally described from off SW Iceland in 260-300 m, is surprising, but we could not see any substantial difference in the specimens collected off NW Gran Canaria. There are other shared species such as   description Shell small, spindle-shaped, translucent white with a brownish, styliform protoconch of two whorls. Teleoconch of a little more than three whorls, the first one markedly enlarged with respect to the protoconch and ascending over it so as to conceal partly the last protoconch whorl, the second one inflated, the third one elongate, forming more than 70% of the total shell height, slightly convex. Suture closely appressed, distinct, the suture internally visible by transparency but fuzzy.  (1986), 5.1 × 1.6 mm according to scale bar) and has a definitely more slender last teleoconch whorl. Mucronalia suava is also larger, and the spire angle is much less increased at the transition from protoconch to teleoconch. Our conclusion is that three different species are involved. The holotype of M. pinguicula n. sp. was live-collected and is presumed to be full-grown as deduced from the sturdy outer lip with a well rounded, not thinning, edge. We agree with Bouchet & Warén (1986) that the resemblance of the abovementioned species with the type species of Mucronalia, the Japanese M. bicincta A. Adams, 1860, based on the abrupt change of spire angle from protoconch to teleoconch, may be superficial but that introducing a new genus is not advisable at the present state of knowledge. Aclis sarsi differs from both in having only one protoconch whorl instead of more than two in the other two species Bouchet & Warén further reported a single shell from the Canaries as probably representing a new species, but refrained from describing it as the specimen was not in good enough condition. It was stated as having "the same kind of apical whorls as A. sarsi but the shape is similar to A. attenuans". The SEAMOUNT 2 material contains additional shells of the same Canarian species but this (including the specimen figured by Bouchet & Warén) has a little more than two protoconch whorls and is therefore similar to Aclis attenuans and does not have the paucispiral protoconch of Aclis sarsi. Admittedly, the Canarian specimens are slightly larger (3.7 mm instead of usually less than 3 mm for Mediterranean specimens) and the protoconch is also slightly larger in proportion, but are here taken as a range extension of Aclis attenuans and new record to Canary Islands waters.  description Shell very small, with a low spire and a blunt, protruding apex and with a deep and rather wide umbilicus. Larval shell globular, smooth and glossy, consisting of about 1 1/4 whorls with diameter 320 to 330 µm, demarcated from teleoconch by a distinct scar (difficult to see in optical microscopy due to transparency, conspicuous under SEM). Teleoconch of 1 1/4 whorl in specimens about 0.9 mm in diameter, up to 2 1/4 whorls in the largest specimens, with a narrow subsutural keel running very close to the suture and giving it a channelled aspect, with another blunt keel delimiting the abapical area around the umbilicus. Sculpture of numerous indistinct growth lines and, on the subsutural keel, of microscopic spiral striae. Within the umbilicus, 7-8 raised lamellae corresponding to earlier positions of outer lip. Aperture distinctly opisthocline, strongly expanded and flaring, especially in lower part prolonging the columellar edge. Outer lip not thickened. Maximum diameter (including the flaring lip) 0.9 to 1.3 mm (holotype 0.75 mm height × 1.0 mm diameter).
reMArks Discaclis lamellata n. sp. differs from the only species of the genus known so far, D. canariensis (Fig. 11A-C), in being twice as high with the same diameter, in having a considerably narrower umbilicus, and in having raised lamellae inside the umbilicus, which represent former stages of growth of the apertural lip. Discaclis canariensis is a shallow-water species with a type locality at 125 m and a record as shallow as 12 m in a submarine cave (Martínez et al. 2005 fig. 2a, b) and again by Bouchet & Warén (1986: figs 1135, 1154 for the number and pattern of spiral cords, with two prominent cords on the middle part of the whorl and an indication of a third one on the subsutural ramp. Admittedly, a large specimen     (Sykes, 1925), shell from DW130 (4.6 mm height); new to Spanish waters; J, protoconch of the same specimen; K, scanning electron micrograph of the protoconch of a juvenile, live taken specimen, same locality; the arrow points to the protoconch-teleoconch limit. Scale bars: 100 µm. description Shell medium-sized, consisting of 4 1/2 whorls, greyish white, very solid, smooth and glossy, with a moderately elevated spire. Apex blunt and rounded, protoconch about 2 mm in diameter, consisting of approximately 1 3/4 whorl, not delimited from the teleoconch. Spire whorls very slightly convex, with a thin suture. Last whorl representing more than 80% of total height. Aperture elongate, narrow, parallel-sided except at the adapical end where it tapers. Outer lip smooth inside, bordered externally by a broad, well delimited, evenly thickened labial varix, very slightly receding at its adapical end, forming there a moderately pronounced shoulder against the penultimate whorl. There are four columellar plaits, stout with a flattened crest, decreasing in size towards the abapical part of the columella, occupying two-thirds of the aperture length. No columellar callus. Animal (Fig. 15) colourless, with a broad and short foot extending flat when crawling, reaching about the same length as the shell longitudinally and about 120% of the shell breadth transversally; the propodium with a broad transverse flap and, behind this, an opaque white glandular area visible by transparency. Head bifid as usual in marginellids, with long, slender and tapering tentacles, lacking eyes but presenting a distinct bulge on each side, at the base of the tentacles where eyes should normally be situated. Siphon large, extending anteriorly about one-third of shell length. Dimensions of holotype: 16.7 height × 8.2 mm diameter; of paratypes, 14.1 × 7.6 mm, 14.2 × 7.5mm, 14.3 × 7.5 mm.   Warén (1985: 274-275), of relatively large size (20-30 mm). Marginella carmenae n. sp. (Fig. 14 A-F) is clearly differentiated by the shape of the lip of uniform thickness (swollen in the central part and thinning at the ends in M. marocana) and by the elongated, parallel-sided aperture (very wide, almost oval in M. subturrita). A further difference is that in M. marocana the abapical columellar plait is continued as a ridge to the abapical end of the shell whereas it fades out at a short distance in Marginella carmenae n. sp. Marginella gustavoi, the only species so far with a Canarian deep-water type locality, has a still more globose apex, a shorter and stouter spire and considerably thinner outer lip and columellar plait, with a very broad aperture.
From the platform of NW Africa, comparable species would be M. adamkusi Bozzetti, 1994, which differs by being smaller (10 mm), with a prominent colour pattern and with a swelling in the upper third of the lip that could justify its assignment to Dentimargo. Marginella belcheri Hinds, 1844 (Fig. 14G, H) is larger (18-20 mm) with a different colour pattern, has a much more pointed apex and the outer lip is thinning out at the adapical end (Goud & Neefs 1996). In M. carmenae n. sp., no definite limit could be seen between protoconch and teleoconch, neither on the live taken specimens nor on more or less worn shells. The approximate limit is taken where growth lines start to be apparent on the shell surface. description Shell very small, consisting of 3 1/2 whorls, whitish, translucent, smooth and glossy, with a moderately elevated spire. Apex blunt and rounded, protoconch 0.7 mm in diameter, consisting of 1 1/2 whorl, delimited from teleoconch by a tenuous line. Spire whorls very slightly convex, with a thin suture and an internal false suture visible by transparency.
Last whorl representing about 85% of total height. Aperture elongate, rather broad, parallel-sided except in the adapical part where it narrows very slightly. Outer lip smooth inside, bordered externally by thin labial varix, very slightly receding at its adapical end, forming there a hardly pronounced shoulder against the penultimate whorl. Abapical end bluntly pointed at the termination of the columella; there are four columellar plaits, equivalent in size, occupying slightly more than half of the aperture length. No columellar callus. Size up to 3.3 mm high (holotype 3.2 mm high × 1.8 mm diameter).
reMArks This species bears a certain resemblance, regarding profile and shape of the aperture, to Marginella colomborum (Bozzetti, 1995) described from Josephine bank, an isolated seamount located NW of Madeira, but M. pulex n. sp. is much smaller (2-3 mm compared to 10 mm) with a still thinner varix. Taking into account the large number of specimens, the thin outer lip is regarded as a diagnostic character state, not as an indication that the specimens are juveniles.  et al. 2014). The first was described from deep water off SW Spain and is the most similar, but it is smaller (7-8 mm), stouter, with a distinctly broader aperture, columellar plaits with a more acute crest, and the outer lip is not so thickened. The figure in Sykes (1905) shows a more pointed anterior end and a still less prominent labial denticle than the syntype pictured in Bouchet & Warén (1985), but in any case Sykes described the aperture as "broad". Dentimargo bojadorensis was described from a depth of 146 m off the mainland African coast. It is smaller, more fusiform without a distinct shoulder at the insertion of the outer lip, which narrows considerably at its adapical termination. The inner labial denticle is also different, situated towards the upper 1/3 of the aperture and more elongate, plait-like. description Shell small, consisting of nearly 4 whorls, yellowish to greyish white and opaque in dead shells, smooth, with a rather low spire. Apex blunt and rounded, protoconch about 1 mm in diameter, consisting of 1 1/4 whorl, delimited from teleoconch by a tenuous line. Spire whorls hardly convex, with a thin suture.
Last whorl representing about 90% of total height. Aperture elongate, narrow and parallel-sided. Outer lip smooth inside, bordered externally by thin labial varix, slightly receding at its adapical end, forming there a broadly rounded shoulder against the penultimate whorl. There are four columellar plaits occupying less than half of the aperture length, the adapical one relatively thin, the other three stout with a flattened crest. A broad and thin columellar callus covering the part of the last whorl adjacent to the aperture. Dimensions of holotype: 6.9 mm height × 4.15 mm diameter, of figured paratype 7.3 × 4.5 mm.
reMArks Only worn shells were collected, so that this species may live shallower and shells have been transported down the steep slope. This species, with a very embracing last whorl, is tentatively assigned to Prunum, as it shares with other (mostly Caribbean: see Cossignani 2006: 196-206) species the stout, ovoid outline with a conspicuous and tapering last whorl, and the smooth inner part of the outer lip. Admittedly, the separation between the genera Prunum and Volvarina Hinds, 1844 is not clear-cut and currently follows a rather arbitrary criterion where the large species with a strong callus are placed in Prunum, the slender species with a thin callus in Volvarina, leaving in between many ambiguously placed species like the one considered here.   description Shell whitish, consisting of 4.5 quite convex whorls, with a well marked, almost channelled suture; protoconch of one globular smooth whorl; teleoconch with oblique axial ribs of which 24 on the last whorl; spiral sculpture of rather deep grooves, rather obsolete on the first whorl and becoming more accentuated on the last whorl; grooves regularly spaced and forming granules at their intersections with the axial ribs. Aperture rather wide, somewhat arched. Three columellar folds, bulging into the aperture, occupying more that 3/4 of the apertural height, the remaining area in the adapical part of the columella somewhat concave; the adapical columellar fold is most developed and the abapical one is very slight; outer lip thin, siphonal canal short. Dimensions of holotype: 6.7 mm height × 2.8 mm diameter.
reMArks This species resembles in many respects Microvoluta superstes Bouchet & Warén, 1985, described from Gorringe Bank, off SW Portugal, but is less slender, more fusiform with the whorls not definitely shouldered as in M. superstes. Spiral sculpture is evenly developed over all the surface of the teleoconch, contrary to M. superstes which has smooth ribs in the median part of the whorls in most specimens. All shells are empty. A preliminary sorting of a large material from Ampère seamount, Josephine seamount and the Meteor group of seamounts collected during the SEAMOUNT 1 and SEAMOUNT 2 cruises revealed several undescribed congeneric species which are currently under study, with a differentiation between banks which may parallel that of Trituba Jousseaume, 1884, as described by Gofas (2003). Because no animal is available, the generic placement is only tentative. description Shell medium-sized, biconical with a tall and turreted spire, and the last whorl a little more than half of the total height. Protoconch of 1.5 whorls, globose, with the last half-whorl ornamented by a few flexuous incremental lines, separated from the teleoconch by an indistinct line. Teleoconch of about 7 regularly increasing whorls; teleoconch whorls distinctly shouldered, with a distinct subsutural rim, then a concave subsutural ramp between this and the shoulder. Shoulder bearing a series of axially elongate knobs (about 15 on penultimate whorl, becoming attenuated on the last whorl where approaching the outer lip). Labial notch well developed and deep, situated at the termination of the subsutural ramp, as is usual in the genus. Siphonal canal short and wide. Colour opaque white, without gloss. Dimensions of holotype, 16.2 mm height × 6.4 mm diameter, of paratypes, 10.7 × 4.3 to 20.1 × 7.6 mm. reMArks Spirotropis guancha n. sp. most resembles S. centimata (Dall, 1889), originally described from SE United States and subsequently reported from the Azores and from off western Morocco (Bouchet & Warén 1980). Spirotropis guancha n. sp. has a globose and smooth, whitish, paucispiral protoconch denoting non-planktotrophic development whereas S. centi mata has a conical, yellowish protoconch of c. 2.5 whorls, with a suprasutural keel, denoting a planktotrophic development.
The thick subsutural rim of S. guancha n. sp. is not seen in S. centimata, which also has much more prominent knobs along the shoulder. The aperture of S. guancha n. sp. is wider, the siphonal canal is shorter and broader than in S. centimata. We figure for comparison ( Fig. 20A-C) a specimen of the latter species collected off SW Morocco by R/V Talisman. description Shell small, whitish, consisting of a little more than three convex, definitely shouldered whorls, with a well marked suture; protoconch (visible part 550 µm high, 400 µm diameter) protruding and somewhat pointed apically, formed of a little more than one whorl, nearly smooth on the initial part but mostly covered by minute, irregular granules which are loosely aligned spirally. Protoconch/teleoconch limit very definite along an orthocline line. Teleoconch with elevated spiral cords, of which 4-5 on the spire whorls and 19-21 on the last whorl; cords narrower than the interspaces, one along the suture, the strongest one on the shoulder and the 2-3 abapical ones decreasing slightly in size; on the last half-whorl, an additional cord between the suture and the shoulder. The entire teleoconch covered by axial wrinkles which are much finer than the cords and override them. Aperture oval-elongate, somewhat truncated adapically at the shoulder of last whorl; outer lip simple, siphonal canal short and very broad. Dimensions of the holotype: 2.2 mm height × 1.15 mm diameter, paratypes up to 2.7 × 1.5 mm. reMArks "Gymnobela" multilirata n. sp., does not resemble closely any Atlantic species known to us, and even the generic and familial placement is tentative. It is guided by the overall resemblance with species like e.g. Gymnobela subaraneosa (Dautzenberg & H. Fischer, 1896) or G. lamyi (Dautzenberg, 1925) (see Bouchet & Warén 1980: 52) which also have a definite shoulder and a sculpture of widely separated spiral cords. Another superficially similar species is Nepotilla amoena (G. O. Sars, 1878) (see Bouchet & Warén 1980: 75, 76) but the latter has fewer and much stronger spirals and a much narrower canal and abapical part. The paucispiral protoconch of G. multilirata n. sp. is typical of a species with non-planktotrophic larval development but does not give a definite clue, although the profile of the protoconch and its definite, orthocline edge resembles the cancellariid Pseu dobabylonella minima (Reeve, 1856) (see Verhecken 2007: 351). However, teleoconch characters including the lack of folds on the columella and the elongate, tapering last whorl (usually blunt and short in cancellariids) do not support a placement in the latter family. Despite the uncertainty on its position, we consider this species as distinctive and abundant so as to be named.
Gymnobela abyssorum (Locard, 1897) (Fig. 22A (Dall, 1927) and called attention on the peculiar microsculpture of granules which is diagnostic of this genus. Pleurotomella megalembryon, the type species of Azorilla Nordsieck, 1968 (Fig. 23A-F), has this microsculpture and has a protoconch very similar to that of the type species of Teretia (Fig. 23G, H). Pleurotomella megalembryon is perceived as congeneric with Teretia strongyla and consequently we consider Azorilla a subjective junior synonym of Teretia. Bouchet & Warén (1980) already noted that "P. megalem bryon has a sculpture which recalls that of Terelia teres". The details of the protoconch sculpture, in both cases with spiral rows of cross-shaped nodules on protoconch 1 and a typical raphitomine sculpture on protoconch 2, are also shared (see Fig. 23  description Shell small, biconical, white; protoconch globose, smooth, of one and a half whorls. Teleoconch of 4 1/2 very slightly convex whorls. Sculpture of low axial ribs (about 15 on the penultimate and last whorl), slightly opisthocline, and of very flat spiral cords, much broader than their interspaces (three on first teleoconch whorl, five on penultimate, 17 on last whorl). Under high magnification, whole surface bearing axial wrinkles parallel to the growth lines. Suture shallow. Aperture narrow, outer lip with 4 denticles inside, the adapical one strongest, thickened outside; inner lip plain, with two strong folds near its middle, the adapical one the largest. Dimension of the holotype: 6.4 mm height × 2.8 mm diameter.  reMArks There are two other described species from the Atlantic with a similar cancellate sculpture. Mitromorpha biplicata Dall, 1889 has one more teleoconch whorl (5-6 instead of about 4 1/2) at roughly the same size (7 mm). The sculpture is also different, being raised spiral cords separated by broad interspaces, in which additional cordlets appear at the end of the last whorl, whereas in M. alabaster n. sp. the spiral cords are flat and much broader than the interspaces. description Shell minute, depressed, with a flat spire and wide umbilicus. Protoconch of a little less than 1 whorl. Teleoconch with three strong spiral ridges, of which one is adapical, the strongest one is on the periphery, and one is abapical. Microsculpture of irregularly arranged granules, loosely aligned axially, which give the surface a shagreened aspect and are attenuated on the ridges; the last 1/4 whorl marked in addition by conspicuous growth lines. Umbilicus broad and wide, delimited by an additional internal ridge. Aperture simple, rounded on the parietal and columellar side, polygonal along the outer lip, reflecting the termination of the ridges. Columellar edge of the aperture reflected outwards and strongly receding where approaching the parietal edge, abutting there on the abapical ridge of the preceding whorl. Dimensions of the holotype: 0.35 mm height × 0.8 mm diameter.

reMArks
The size, profile, and the characteristic recession of the abapical part of the aperture support the placement of this species in Orbitestella. The protoconch is damaged on the only available specimen and could not be observed properly. This minute species does not resemble closely any other gastropod known to us from the eastern Atlantic. The most similar is Orbitestella hinemoa Mestayer, 1919, from off New Zealand, which differs only in having a more even development of the three keels on the last whorl, the peripheral one not so much protruding as in Orbitestella pruinosa n. sp. The only other species of Orbitestella described from the eastern Atlantic and Mediterranean area is O. dariae (Liuzzi & Zucchi Stolfa, 1979), recorded from Madeira by Segers et al. (2009). It has a similar outline but an evenly convex abapical part with several flat spiral cords, and a spiral series of prominent knobs on the adapical side; furthermore it lacks the shagreened microsculpture of O. pruinosa n. sp. The specimen from a deep-water thanatocenosis off Sardinia, Western Mediterranean, figured as O. dariae by Bonfitto et al. (1994) is not that species and is so similar to O. pruinosa n. sp. that it is possibly the same; their specimen differs from the Canarian one in having definite axial riblets between the spiral ridges, instead of loosely aligned granules.
Family ciMidAe Warén, 1993 Genus description Shell minute, globose with a low, markedly stepped spire, and the last whorl occupying more than 4/5 of the total height. Protoconch c. 280 µm in diameter, with its apex obliquely immersed in the first teleoconch whorl (type B of van Aartsen, 1987), smooth and glossy. Teleoconch of two whorls, shouldered with a very strong keel separated from the suture by a broad channel, smooth except for growth lines which are particularly conspicuous within the subsutural channel. Last whorl with maximum convexity near the periphery, flat between keel and periphery and only slightly convex in the abapical part. Umbilicus broad and conspicuous, delimited by a blunt keel. No columellar tooth. Dimensions of the holotype: 1.0 mm height × 0.95 mm diameter.
reMArks This species does not resemble any known gastropod from the North Atlantic, but is clearly identified as a member of the Pyramidellidae by the heterostrophic protoconch. Among all the pyramidellid genera reported in the area, Liostomia was considered because its type species, although quite different in aspect, is also umbilicate, has the protoconch nucleus similarly immersed, lacks a columellar tooth, and has an incipient subsutural keel on the first teleoconch whorl, admittedly not so conspicuous (Høisaeter 2014). Liostomia mamoi Mifsud, 1993 was described from the Mediterranean and has a height/diameter ratio close to that of L. canaliculata n. sp., although lacking the spiral keels.

DISCUSSION
The specific richness in this transect (295 species Rolán & Hernández, 2009) are represented by some specimens collected alive. According to the master list, 127 species among the 1262 cited from the Canary Islands are endemic (10%). In the present list, the 11 species (3.7% of the total) recognized as Canarian endemics represent a much lower proportion than expected. On the other hand, if the 13 newly described species are taken into account, the proportion of endemics would be 8% of the total, a figure more in line with the expectations. This is likely to be verified as all the newly described species have paucispiral protoconchs and are therefore prone to have restricted ranges.
The increase provided by the new records (51 species, 4% of the total cited for Canarian waters) is considerable, exceeding, for example, figures for the Alborán platform in Peñas et al. (2006) where eight of 655 species were new to science and 39 were new to the Spanish Mediterranean coasts. Adding new records and the new species, over one-fifth of the species collected in this material brought new knowledge regarding large-scale distributions, which means that the inventory of deep waters around the Canaries is not even approaching completion. It is well known that the tropical deep-sea still holds a large fraction of biota to be discovered. Admittedly, these numbers do not match those found in the world's "golden triangle" of biodiversity (e.g. Bouchet et al. 2008 recorded 1028 species of molluscs from the New Caledonia Exclusive Economic zone from depths below 100 m, and 601 of these (58.4%) were new to science) but they are still remarkable for being so close to the supposedly well-explored European seas.
Most of the new species and new records in our material are concentrated in the deepest dredge haul DW130 in 655-660 m, and this must be placed in context with hydrological setting around the Canaries (Hernández Guerra et al. 2001;Almón et al. 2014a, b). The superficial waters (< 600 m depth) correspond to the North Atlantic Central Water (NACW) flowing south following the Canary current, at the eastern boundary of the North Atlantic Subtropical Gyre. The DW130 sample stands out as coming from below 600 m, where it should be situated within a northwards flowing tongue of Antarctic Intermediate Water (AAIW), formed where superficial waters of Antarctic origin sink as they meet warmer waters of the tropical zone (Machín & Pelegrí 2009). This flow, with a core at around 800 m depth and temperature around 8.3°C, certainly does not explain shared species with the western Atlantic and with Iceland, but two additional elements must be borne in mind. Firstly, the planktotrophic larvae of deepwater species reach superficial layers during their development (Bouchet 1976;Scheltema 1994, Bouchet & Warén 1994 and can therefore enter the North Atlantic Subtropical Gyre. Secondly, the deeper parts of the basin receive a thick layer of the North Atlantic Deep Water (NADW), which is formed in the Norwegian and Greenland Seas and passes into the North Atlantic in the deep channels that separate Scotland, Iceland, and Greenland. NADW is found throughout the Atlantic Ocean below 1500 m, although the flow around the Canaries is reported to be weak. Much has yet to be learned about the fauna affected by those deep and cold water masses, in an area where the deep-sea exploration still relies mostly on the xIxth century "Talisman" expedition.