Taxonomy of some Galeommatoidea ( Mollusca , Bivalvia ) associated with deep-sea echinoids : A reassessment of the bivalve genera

The type species of Axinodon ellipticus Verrill & Bush, 1898 and Kellia symmetros Jeffreys, 1876 are re-described. It is concluded that the two species are not conspecific and that K. symmetros cannot be placed in the genus Axinodon. The family affinity of Axinodon is not resolved, although it is probable that this genus belongs to the Thyasiridae. Kellia symmetros is the type species of Kelliola and is placed in the Montacutidae. Kelliola symmetros is most probably associated with the echinoid Aeropsis rostrata and is not the species previously recorded from North Atlantic Pourtalesia echinoids under the name of Axinodon symmetros. This commensal associated with the North Atlantic Pourtalesia is here described as new and placed in the new genus as Syssitomya pourtalesiana gen. nov. sp. nov., Syssitomya gen. nov. differs from all other genera in the Montacutidae by having laminar gill filaments modified for harbouring symbiotic bacteria and it is thus assumed to be chemosymbiotic. A montacutid associated with the hadal Pourtalesia heptneri is described as Ptilomyax hadalis gen. nov. sp. nov.


Introduction
The bivalve currently known as Axinodon symmetros (Jeffreys, 1876), a commensal associated with the deep-sea echinoid Pourtalesia A. Agassiz, 1869, has been found to have a gill structure similar to that of the Thyasiridae Dall, 1900, and to harbour symbiotic bacteria suggesting that it may be chemosymbiotic (Oliver, Southward & Dando in press).In that study, some key taxonomic issues arose that now require resolution before publication of that study can proceed.
Firstly, there has been debate about the family placement of Axinodon Verrill & Bush, 1898.Originally, it was placed in the Thyasiridae (Verrill & Bush 1898) and this was followed by Aartsen (1996) and Coan et al. (2000).Ockelmann (1965) placed it in the superfamily Galeommatoidea Gray, 1840 and family Montacutidae Clark, 1855; this placement was maintained by Chavan (1969) and is followed in the current European checklist (CLEMAM 2012).The implications of this debate are twofold: • If Axinodon were a thyasirid then the chemosymbiosis might be expected, but the commensalism would be a first for that family.
• If Axinodon were a montacutid then Axinodon would be the first of that family and the first commensal bivalve shown to harbour symbiotic bacteria.
Secondly, it was also discovered that commensal bivalves were present on the echinoid Aeropsis rostrata (Wyville Thomson, 1877), but although similar to that on Pourtalesia, they were not identical.
Eve Southward (pers.comm.)noted that the type locality and sample number for Aeropsis rostrata is the same as that for Kellia symmetros Jeffreys, 1876, both from "Station 9" of the "Valorous Expedition" in the Davis Straits.This brought into question the identity of K. symmetros and the identity of the Pourtalesia commensal.This is further exacerbated by the conflicting descriptions of Axinodon, some stating that it is edentulous (Aartsen 1996) other showing a distinct tooth in the right valve (Bouchet & Warén 1979).
Bivalves belonging to the Galeommatoidea and the Montacutidae in particular have long been known to be commensal with species of sea-urchins and some shallow water species have been studied in considerable detail.There are numerous family and subfamily taxa proposed within the Galeommatoidea (Bieler & Mikkelsen 2006) but their distinction and relationships remain controversial (Mikkelsen & Bieler 2007) although the latter authors do recognise the Galeommatidae Gray, 1840 as distinct form all others.Mikkelsen & Bieler (2007) and Carter et al. (2011) refer all non-galeommatids to the Lasaeidae Gray, 1842, that being the earliest family name.Many other authors continue to follow the classification of Chavan (1969) and adopt Montacutidae Clark, 1855 for those taxa morphologically similar to Montacuta Turton, 1822 (Jespersen et al. 2004;Kamenev 2008;Gofas & Salas 2008).Jespersen et al. (2004) argue for recognition of Montacutidae on grounds of hinge and anatomical characters in that the Lasaeidae, Leptonidae and Kelliidae have both cardinal and lateral teeth, and the presence of dimorphic sperm is peculiar to species placed in the Montacutidae.Unfortunately, galeommatoids are poorly represented in molecular phylogenies (Taylor et al. 2007) and there is no resolution at the family level.Here, Montacutidae is preferred over Lasaeidae, as the species under consideration are morphologically most similar to Montacuta and other associated genera that have been grouped in the Montacutidae.Popham (1940) and Oldfield (1961) Jespersen, Lützen & Nielsen (2004) studied the fine anatomy of three species all associated with Brissus latecarinatus (Leske, 1778) and created two new genera Montacutella Jespersen, Lützen &Nielsen, 2004 andBrachiomya Jespersen, Lützen &Nielsen, 2004.Table 1 gives a compilation of galeommatids known to be commensal with echinoids.
Associations with deep-sea echinoids are known, but taxonomic recognition has been restricted to a single taxon associated with North Atlantic species of Pourtalesia.Currently this taxon is known as Axinodon symmetros (Jeffreys, 1876) and is associated with Pourtalesia jeffreysi Wyville Thomson, 1877 and P. miranda Agassiz, 1869 (Bouchet & Warén 1979;Gage et al. 1985).The synonymy of Kellia symmetros and Axinodon ellipticus Verrill & Bush, 1898 was first made by Ockelmann in Chavan (1969).Dall (1899) noting discrepancies in Jeffreys original description of K. symmetros placed it in a new genus Kelliola Dall, 1899; this was synonymised with Axinodon by Chavan (1969).
In the original descriptions of "Kellia" symmetros (Jeffreys 1876) and Axinodon ellipticus (Verrill & Bush 1898) there is no mention of an association with echinoids.The first mention of an association is by Ockelmann & Muus (1978) and is restated by Bouchet & Warén (1979), both giving Pourtalesia as the host.Resolution of the issues raised in this introduction requires examination of type materials but, although available, are restricted to holotypes consisting of shells only.For this paper, the holotypes of the species Kellia symmetros and Axinodon ellipticus are re-described and their supposed synonymy and family affinities are analysed.This is followed by descriptions of the commensals from Pourtalesia and Aeropsis Mortensen, 1907 and their relationship to previously described species is reviewed.Finally, the specimens figured by Mironov (1978) are formally described.

Material and Methods
The materials used in this study are all from museum collections derived from late 19 th century oceanographic explorations and from recent research studies by British and Norwegian deep water programmes.The modern collections were made by grab or sledge hauls and the bivalves were found attached to the spines of sea urchins.Details of each collection site are given for each specimen examined.
Specimens are of dry shells or whole animals previously fixed in formaldehyde and preserved in 70% ethanol.
Stereo microscope observations were made with Leica M8 or M10 instruments and macrophotography with the latter, attached to a digital camera system, employing enhancement via AutoMontage™ software.Scanning electron micrographs of type material held by the United States National Museum were supplied by that institution and all others were made in the National Museum of Wales using a Jeol Neoscope.Soft tissues were prepared by critical point drying and subsequent gold coating for SEM examination.
Right valve lacking any projecting teeth; left valve with a short, weak.marginal flange beneath the lunule.Isomyarian.Pallial line entire.

Redescription of holotype (Fig. 1)
Shell small, length 3.5 mm, height 3.3 mm.Thin, fragile.Equivalve.Inequilateral, beaks behind the midline.Umbos prominent, beaks prosogyrate.Outline obliquely subcircular, distinctly expanded anteriorly; anterior dorsal margin short, as a poorly defined lunule; posterior dorsal margin indistinct sloping into broadly rounded posterior, anterior broadly rounded more so than posterior, ventral margin broadly rounded no distinct junctions with lateral margins.Sculpture of dense, fine commarginal ridges most obvious on margins.Prodissoconch II distinct, 754 µm across, with weak commarginal lines (Fig. 1H).Hinge plate narrow, ligament deeply sunken on a groove running from under the beaks posteriorly for about one-third of the posterior dorsal slope.Right valve lacking any projecting teeth, hinge plate slightly and irregularly thickened below lunule.Left valve with a short, weak.marginal flange beneath the lunule; a very weak protuberance is visible below the beak.Adductor scars oval, roughly of equal size; pallial line entire.Interior with feeble radial striae.

Differential diagnosis
At the generic level Chavan (1969) lists Kelliola Dall, 1899 as congeneric with Axinodon and assigns them to the Galeommatoidea in the family Montacutidae.This is despite Verrill & Bush (1898) describing Axinodon as edentulous and Dall (1899) describing Kelliola with teeth.The hinge of A. ellipticus has been examined carefully to ascertain if it is edentulous as stated by Aartsen (1996) or if a cardinal peg was present but has been broken off.The SEM images presented here (Fig. 1A) show no indication of a broken tooth confirming the edentulous condition.Comparisons with figures in Chavan (1969) should not be made as these incorrectly show a shell that is expanded posteriorly and teeth that are not shown in the accurate figures made by Verrill & Bush (1898).From the description below, Kelliola is seen to have a cardinal peg in the right valve (Fig. 2A) and therefore Axinodon and Kelliola are not congeneric.In outline, Axinodon looks like Kelliopsis Verrill & Bush, 1898 but the latter has small, distinct teeth (Verrill & Bush 1898).Verrill & Bush (1898) placed Axinodon in the Thyasiridae as did Aartsen (1996) and Coan et al. (2000), and there are similarities with the thyasirid genus Mendicula Iredale, 1924.Without anatomical data, it is not possible to confirm the affinity with the Thyasiridae but it is, however, very doubtful that Axinodon is a galeommatid as the latter all display some degree of dentition in the right valve.

Diagnosis
Small to minute shells, Equivalve.Slightly inequilateral, beaks behind the midline.Outline subovate, anterior a little more expanded than posterior.Hinge with an anterior cardinal peg in the right valve, an anterior marginal flange in the left valve, posterior teeth lacking; ligament internal on a recessed resilifer beneath and posterior of the beaks.Sculpture weak primarily of commarginal lines, faint radial striations may be present along with microscopic notches in the ventral margins.Adductor muscles of approximately equal size; mantle edge with a large pedal gape, fused from mid point posteriorly with a small exhalant aperture; foot with an active byssus.Ctenidia of single, un-reflected, demibranchs.
Redescription of the holotype (Fig. 2) Shell minute, length 1.0 mm, height 0.77 mm.Thin, translucent.Equivalve.Weakly inequilateral, beaks just behind the midline.Umbos prominent, beaks orthogyrate.Outline subovate, slightly extended anteriorly; anterior dorsal margin sloping into broadly rounded anterior margin; posterior dorsal margin indistinct sloping into broadly rounded posterior margin, this slightly less expanded than anterior margin; ventral margin weakly curved.Sculpture weak, of indistinct commarginal lines most obvious on lateral margins.Prodissoconch II distinct, 373 μm across, with weak commarginal lines.Hinge plate weak, Ligament short, internal, attached to a shallow resilifer situated beneath and posterior of the beaks.Right valve with a single, projecting, cardinal peg, immediately anterior to this tooth a slight depression.Left valve with a short, weak, marginal flange in a posterior lateral position.Adductor scars oval, roughly of equal size; pallial line entire.

Description (based on material from Bay of Biscay)
Shell. (Fig. 3) Minute, largest of length 1.3 mm, height 1.0 mm.Thin, fragile.Equivalve.Inequilateral, beaks behind midline.Umbos weakly inflated, beaks orthogyrate.Outline subovate, longer than high, length to height ratio 1.3:1, slightly extended anteriorly; anterior dorsal margin sloping, rather straight merging smoothly with rounded anterior; posterior dorsal margin shorter and sloping more steeply than anterior, merging smoothly with rounded posterior margin; anterior slightly more expanded than posterior; ventral margin gently curved.Sculpture weak almost smooth, of fine commarginal lines; radial lines apparent under transmitted light (Fig. 3A) but these very faintly raised (Fig. 3J).Prodissoconch II distinct, 380 µm in diameter sculptured with commarginal lines (Fig. 3I); Prodissoconch I weakly demarcated, 140 µm in diameter with a punctate micro-sculpture (Fig. 3I).Ligament short, internal, attached to a shallow resilifer situated beneath and posterior of the beaks.Right valve with a single, projecting, cardinal peg, immediately anterior to this tooth a slight depression (Fig. 3D).Left valve with a short, weak, marginal flange in a posterior lateral position (Fig. 3E).Adductor scars oval, roughly of equal size; pallial line entire.Ventral margin dissected by minute transverse grooves (Fig. 3K).
AnAtomy. (Fig. 4A) Mantle margin free for most of its length, joined and attached to terminal of gill axis, anterior (pedal) aperture extensive, posterior aperture very small.Adductor muscles of approximately equal size.Foot with a large toe and small heel, byssus functional producing a mass of threads arising from a single stalk.Anterior pedal retractor inserted above the anterior adductor, posterior pedal retractor above the posterior adductor.Ctenidium of a single demibranch, with nine non-reflected filaments in the largest specimen.Filaments rod shaped, lacking abfrontal extension or harbouring symbiotic bacteria.Labial palps small but projecting.
ASSociAtion.Attached by byssus threads to the spines of the echinoid Aeropsis rostrata (Fig. 5).

Distribution
Kelliola symmetros is known only from the type locality and from the Bay of Biscay, at abyssal depths.The host echinoid is widely distributed in the North Atlantic (Echinoid Directory 2012).

Differential diagnosis
The hinges of K. symmetros and the Aeropsis commensal are almost identical, but K. symmetros has slightly more prominent umbos and lacks the marginal transverse grooves.Given that K. symmetros was taken in the same sample along with Aeropsis, but not attached to it, it is possible that the two are associated.This suggests an ecological affinity with the Aeropsis commensal described here.Despite the wide geographical separation of the samples considered here, Aeropsis rostrata is regarded as pan Atlantic and having an abyssal bathymetric range (WoRMS 2012).With so few specimens at hand and the poor condition of the holotype of K. symmetros, we have chosen to be conservative and regard the Aeropsis commensal from Biscay conspecific with Kelliola symmetros.Consequently, at the family level the hinge and anatomical characters of Kelliola are entirely in keeping with the Montacutidae, consisting of a single cardinal peg in the right valve, a marginal flange in the left valve and an internal ligament.The ligament is attached to an elongate shallow depression extending below the beaks and is therefore most similar to Montacuta sensu stricto (Fig. 6).Montacuta substriata,  the type species of Montacuta, has a longer cardinal tooth, has radial ridges and ovate in outline (Fig. 6).Anatomically Kelliola is similar to many montacutids, where the ctenidium is reduced to a single demibranch.However, in this genus, the filaments are very few and not reflected (Fig. 4A).This condition could be due to the small size of the specimens.Dall (1899) noted that Jeffreys' (1876) description of the hinge of K. symmetros was incorrect and, in re-describing it, created the new genus Kelliola for it.Dall did not make comparisons with other genera but noted that it was similar to Aligena Lea, 1846.This is difficult to understand as Aligena species have a cardinal peg in each valve (Harry 1969).Comparisons with other montacutid genera are currently complex due to a lack of compatibility in descriptions of characters and the widely varying use of generic names.Such difficulties were amply recognised by Gofas & Salas (2008) in their review of Mysella Angas, 1877 and consequent creation of the genus Kurtiella Gofas & Salas, 2008.For the purposes of this paper, comparisons are restricted to genera that have a single cardinal peg in the right valve only and left valve with varying degrees of pseudocardinal development.In hinge characters, Kelliola is most similar to Montacuta sensu stricto (Fig. 6) and Neaeromya Gabb, 1873 (Coan et al. 2000) in that the development of the posterior teeth is limited to a marginal flange in the left valve and the ligament is attached to a shallow depression beneath and posterior to the beaks.For N. rugifera (Carpenter, 1864) Narchi (1969) states that there is a tooth in each valve, although it may be reduced in the left valve (Paul Valentich-Scott pers comm); a further example of the contradictory descriptions found for the montacutid species.In Tellimya T. Brown, 1827 the resilifer is developed and the hinge plate thickened accordingly (see Ockelmann 1965, Fig. 2).In Montacutella the left valve flange is developed as a small projection (Jespersen et al. 2004) and approaches the condition seen in Aligena.The shell of Brachiomya is like that of Tellimya (Jespersen et al. 2004).
Kelliola is as different from Montacuta sensu stricto as are the other genera and a molecular study is required to evaluate the significance of the morphological characters.Kelliola is retained here until such a study is undertaken.
Species level comparisons are restricted to the few abyssal galeommatid species that have been described and none other than that described below under Syssitomya pourtalesiana sp.nov.have been found attached to echinoids.The shell of S. pourtalesiana sp.nov. is more expanded anteriorly, has a more depressed lunule and lacks marginal notches.The ctenidium is highly modified with laminar filaments whereas that of K. symmetros is not modified in this manner.Other described, Atlantic, deep-sea, galeommatids have been assigned to the genera Mysella (now Kurtiella) (Gofas & Salas 2008), Epilepton Dall, 1899 (Allen 2007) or Draculamya Oliver & Lützen, 2011(Oliver & Lützen 2011) none having a dentition identical to Kelliola or Montacuta.Among ten undescribed galeommatoids from the deep Atlantic, Allen (2008) lists two undescribed Montacuta species that may or may not be similar to K. symmetros.

Diagnosis
Shell small, thin.Equivalve.Inflated.Inequilateral, beaks behind the midline.Outline roundly subovate, distinctly expanded anteriorly, lunule depression distinct.Hinge with an anterior cardinal peg in the right valve, an anterior marginal flange in the left valve, posterior teeth lacking; ligament internal on a recessed resilifer beneath and posterior of the beaks.Sculpture weak primarily of commarginal lines.Ctenidium of single partly reflected demibranchs; filaments laminar (Fig. 4E), extended abfrontally, abfrontal surfaces lined with bacteriocyte cells densely packed with symbiotic bacteria.A detailed scanning electron and transmission electron microscopy study of the ctenidium is in preparation and is beyond inclusion in this taxonomic paper (Oliver, Southward & Dando in press).

Etymology
From the Greek, syssitos a messmate and mya a clam, referring to the commensal habit and bacterial symbiosis.

Etymology
After Pourtalesia, the host echinoid; and the Latin termination -iana, to denote belonging with.
AnAtomy.(Figs 4B, E; 8A 2 ) Mantle margin free for most of its length, joined and attached to terminal of gill axis, anterior (pedal) aperture extensive, posterior aperture very small (Fig. 8A 2 ).Anterior mantle edge thrown into folds (Fig. 4B).Adductor muscles of approximately equal size.Foot with a large toe and small heel, byssus functional producing a mass of threads arising from a single stalk.Anterior pedal retractor inserted above the anterior adductor, posterior pedal retractor above the posterior adductor.
Specimens from abyssal depths in the Rockall Trough do not differ significantly from those from the Norwegian Sea except for a large shell that shows greater anterior expansion.The ctenidia of this specimen show the same laminar filaments with abfrontal extension.Specimens collected in the month of February were gravid, the suprabranchial chamber holding hundreds of sub-triangular larvae, on average 112 µm in diameter.
There is no record of any association with an echinoid but these specimens originate from the same sampling programme reported upon by Gage et al. (1985) where they were attached to Pourtalesia miranda.
ASSociAtion.Attached by byssus threads to the spines of Pourtalesia jeffreysi and P. miranda (Fig. 9).

Distribution
Confirmed from the NE Atlantic, Norwegian Sea to Bay of Biscay at depths from 800-3617 m.Records (Allen 2008) from the NW and SW Atlantic have not been confirmed although one of the host species P. miranda has been recorded in these areas.
Pourtalesia jeffreysi has two recognised subspecies with both hosting Syssitomya gen.nov.: Pourtalesia jeffreysi gibbosa Mironov, 1995 has a bathyal range while the subspecies lata Mironov, 1995 is abyssal.The known geographic range for P. jeffreysi is the Norwegian Sea and Russian Arctic Ocean (WoRMS 2012).

Differential diagnosis
The shell characters of Syssitomya gen.nov.are montacutid in all respects, notably the anterior expansion, internal ligament and the hinge reduced to a single cardinal peg in the right valve and a marginal flange in the left valve.In shell character, similarities are greatest with Kelliola but Syssitomya gen.nov., uniquely within the Galeommatoidea, has highly modified ctenidia with abfrontally extended, laminar filaments.
At the species level, for comparison with Kelliola symmetros see above.The somewhat inflated, anteriorly expanded form of S. pourtalesiana sp.nov. is rather distinctive and not like the form of other deep-sea montacutids such as species of Kurtiella (Gofas & Salas 2008) or Epilepton (Allen 2007) that are more ovate, compressed and have different dentition patterns.From external appearances it more resembles some thyasirids notably Thyasira subovata (Jeffreys, 1881) (see Oliver et al. 2012) and if not found attached to its host could easily be mistaken for a thyasirid.

Diagnosis
Very small, juvenile shell equivalve, subovate, slightly expanded anteriorly.Adult shell inequivalve dorsal margin of left valve elevated and rolled over, lateral dorsal margins extended, alate.Ligament small, posterior, deeply sunken; dorsal margin with thickened periostracal "ligament".Hinge plate not examined.Sculpture almost smooth with faint commarginal lines.Ctenidium of single reflected demibranchs; byssus of multiple strands; pedal aperture large, exhalant aperture small.

Etymology
hadalis -referring to the hadal zone of the deep ocean (deeper than 6000 m).The material has been poorly preserved, thus the shells are considerably softened, and the tissues are rather contracted.No attempt has been made to examine the hinge, as the valves are so soft that they do not survive being separated.However, the distinctive external form of the shell and the anatomy are sufficient to describe it and place it at the family level.

Description
Shell. (Fig. 10A-C) Small, (largest specimens 2.5 mm L x 1.9 mm H).Inequivalve in adult, lateral dorsal margins of right valve elevated and rolled over projecting beyond left valve.Outline slightly inequilateral, beaks slightly prosogyrate, just behind the mid line, anterior more expanded than posterior; dorsal margins more or less straight, those in adult right valve projecting laterally, becoming alate; forming distinct lateral sinuses with adjacent anterior and posterior margins; lateral margins broadly rounded, anterior more so; ventral margin curved.Ligament small, posterior, deeply sunken; dorsal margin with thickened periostracal "ligament".Hinge plate not examined.Sculpture almost smooth with faint commarginal lines.Prodissoconch I, 105 µm; prodissoconch II, 265 µm, with fine commarginal lines.
AnAtomy. (Fig. 10D) Mantle edge mostly unfused with extensive anterior-ventral pedal aperture; fused with terminal of gill axis; exhalant aperture small.Anterior-ventral inner margin raised into series of prominent evenly spaced transverse ridges.Anterior and posterior adductor muscles of similar size, oval in section.Posterior pedal retractor and anterior pedal muscles present.Ctenidium of a single demibranch, ascending and descending lamellae of almost equal size.Labial palps small.Foot with well developed toe, sole flattened and apparently with tissue projections (these may be an artefact of preservation); heel distinct, byssus functional producing numerous thin filaments.

Differential diagnosis
The projecting and rolled over dorsal margins of the right valve are reminiscent of pteriids but Ptilomyax gen.nov. is clearly allied with the galeommatids as evidenced by the anatomy.Ptilomyax gen.nov. shares many common features with other montacutids associated with echinoids including: a large anterior pedal aperture, mantle fusion limited to junction with gill axis, ctenidium with a single demibranch, foot with prominent toe with a sole and with a functional byssus, ligament internal and weak hinge with poorly developed teeth (Oldfield 1961;Ockelmann 1965;Ponder 1968;Jespersen et al. 2004).The Galeommatoidea exhibit an extraordinary range of shell morphologies as can be seen in Chavan (1969) but none have projecting dorsal margins.This unique feature is sufficient to warrant the creation of a new genus and new species.

General discussion
This study has been necessary largely because of inadequate descriptions of small species based on shells alone and from few specimens.Family placement of these taxa cannot be made on shell characters alone and when anatomical data are lacking contemporary studies are compromised.Two key taxa based on unique and poorly preserved shells have been the centre of this paper with the taxonomic crux being their relationship to either of the commensals fully described here from the echinoids Aeropsis and Pourtalesia.This paper has excluded Axinodon but has not clarified the family affinity of that genus.The shell characters of the presence of a single cardinal peg in the right valve, marginal flange in the left valve and an internal ligament attached to a flat or grooved resilifer are shared by the commensals and K. symmetros.Although this paper concludes that the Aeropsis commensal is K. symmetros there was an option to regard K. symmetros as a nomen dubium because of the poor condition of the holotype and the lack of anatomical data for further comparisons.Despite the holotype of K. symmetros being very small and in poor condition, its almost symmetrical outline and lack of lunule distinguish it from the Pourtalesia commensal.As a consequence of this decision the Pourtalesia commensal was left without a species name and is named here as pourtalesiana.The shell characters suggest that it should be placed in Kelliola but the highly modified gill and bacterial symbiosis warrant further distinction and the new genus Syssitomya gen.nov. is erected for it.The use of anatomical characters to define genera and species within the Galeommatoidea is not novel, most recently Jespersen et al. (2004) used anatomical characters to define their montacutid genera Montacutella and Brachiomya and Oliver & Lützen (2011) did likewise for Draculamya.It is accepted that using anatomical characters makes systematic comparisons, with species known from shells alone, difficult but no more so than when applying molecular data to morphological systems.In this case, the adaptive radiation from suspension feeding to bacterial symbiosis represents a considerable change in both morphology and ecology.In my view, this adaptation is of considerably more significance than small changes of shell morphology that have traditionally been recognised by many bivalve taxonomists.

Fig. 1 .
Fig. 1.Holotype of Axinodon symmetros Verrill & Bush, 1898, USNM 35175.A-B.SEM of hinges of right and left valves.C-D.SEM of internal of right and left valves.E. Photo micrograph of internal of right valve.F-G.SEM of external of right and left valves.H. SEM of prodissoconch.

Fig. 2 .
Fig. 2. Holotype of Kellia symmetros Jeffreys, 1876, USNM 170626.A-B.SEM of hinges of right and left valves.C-D.SEM of internal of right and left valves.E-F SEM of external of right and left valves.G-H.photo micrographs of internal and external of right valve.

Fig. 3 .
Fig. 3. Kelliola symmetros (Jeffreys, 1876) from Biscay.A-C.Photo micrographs a external of left valve, internals of both valves.D-E.SEM of hinges of right and left valves.F-H.SEM of internal of both valves and external of left valve.I. SEM of prodissoconch.J. SEM of anterior area showing weak radial sculpture.K. SEM of margin showing transverse grooves.

Fig. 7 .
Fig. 7. Syssitomya pourtalesiana sp.nov.from Norwegian Sea.A-C.Photo micrographs a external of left valve, internals of both valves.D-E.SEM of hinges of right and left valves.F-G.SEM of internal of both valves.H. SEM of prodissoconch.I. SEM of internal of right valve from NORBI cruise, abyssal, Courtesy of Anders Warén.