Galatheavalvidae Knudsen, 1970

Galatheavalvidae Knudsen, 1970 View in CoL

The family was diagnosed by having an internal shell, completely covered by the middle mantle fold; the presence of a single (inner) demibranch; a ventrally displaced anterior adductor muscle; two permanent mantle margin openings (inhalant– pedal and exhalant); and a well developed foot, with byssus. Additional characters reported for Galathea holothuriae Knudsen, 1970 , the only species thus far known of this genus, include the absence of teeth in the hinge plate and tentacles along the posterior part of the ventral margin, and the presence of a peculiar ‘dorsal brood pouch’, connected with the mantle cavity and extending beyond the shell. Although the superfamilial placement of this family was not determined by Knudsen (1970) at the time of erecting it, Galatheavalvidae currently appears listed under Cyamioidea by Bieler et al. (2010), a placement that appears improbable considering the above-mentioned morphological and anatomical characteristics. Alternatively, Bieler & Mikkelsen (2006) regarded Galatheavalvidae as Galeommatoidea . Huber (2010, 2015) followed this superfamilial placement, but considered Galatheavalvinae to be a subfamily of Galeommatidae .

SPORTELLIDAE : A FAMILY WITH UNCERTAIN AFFINITIES

From Thiele’s (1934) classification to the most recent classifications of bivalves (e.g. Ponder & de Keyzer, 1998; Bieler et al., 2010; Huber, 2010), the Sportellidae were always considered as cyamioideans. However, the concept of Sportellidae has varied greatly throughout time. In this regard, and as mentioned before, the genus Basterotia Hörnes, 1859 was included in Sportellidae for a long time, although it is currently considered as a different family of Galeommatoidea . Knowledge about the species that remain at present regarded as Sportellidae appears mostly to be restricted to shell morphology ( Coan, 1999). Sportella dubia ( Deshayes, 1824) , the type species of the genus (and the type genus of the family), is a fossil. Consequently, the diagnostic (reproductive, molecular and anatomical) characters used in the present study to define Cyamioidea cannot be studied in that species. The hinge of the specimen described and figured by Deshayes (1824) does not agree with those of Cyamiidae and Gaimardiidae . Ponder & de Keyzer (1998) figured a living specimen of ‘ Sportella sp. ’ as having pedal, inhalant and exhalant apertures, the pedal completely papillate, the exhalant projected in a siphon and with ‘well developed tentacles around the posterior inhalant and exhalant apertures’. Furthermore, this figured specimen shows the presence of a long foot, for which Ponder & de Keyzer (1998) described the presence of functional byssus gland(s?) and byssus groove in adults. The presence of a completely papillate pedal aperture and such a long foot are characters not observed in any other cyamioid genus. Some other living species were attributed by different authors to Sportella , although most of them were subsequently transferred to other genera, such as Ensitellops , Fabella , Neaeromya , Paramya and Pseudopythina (all of them currently regarded as Galeommatoidea ). The reproductive characteristics of other living species still grouped in Sportella have never been investigated. The information available at present does not allow us to draw a conclusion on whether Sportellidae corresponds (or not) to Cyamioidea .

FOCUSING ON CYAMIIDAE : A CASE OF CONSERVED LINAGES OR A FAMILY LESS DIVERSIFIED THAN PREVIOUSLY THOUGHT?

Ponder (1971) pointed out that anatomy is the most useful character for determining the generic relationships of Cyamioidea . To date, anatomical information is known for a relatively reduced group of cyamiid species, as follows: Cyamiocardium chuanisinense , Cyamiocardium crassilabrum , Cyamiocardium dahli , Cyamiocardium denticulatum , Cyamiocardium domaneschii , Cyamiocardium namuncurense , Cyamiocardium rotundatum , Cyamiocardium yeskumaala ( Soot-Ryen, 1951, 1957, 1959; Passos & Machado, 2014; Urcola & Zelaya, 2018; present study), Cyamiomactra problematica ( Ponder, 1971) , Cyamium antarcticum ( Ponder, 1971; present study), H. laminifera (present study), J. foveolata (present study), Perrierina taxodonta ( Ponder, 1971) , Ps. cardiformis ( Pelseneer, 1903; present study), Ps. franki ( Zelaya & Ituarte, 2009; present study), Ps. inexpectata (present study), Pt. georgiana (present study) and Reloncavia chilenica ( Soot-Ryen, 1957, 1959). Considering these species, no major anatomical differences appear among members of Cyamiocardium , Cyamiomactra , Perrierina , Cyamium , Heteromactra , Pseudokellya , Ptychocardia and Reloncavia to justify their generic separation. Instead, the characters previously regarded as distinctive at the generic level failed to separate groups when considering the intrageneric variability (such as, for instance, the variability observed in the relative size of the demibranchs or the degree of development of the posterior series of tentacles among different species of Cyamiocardium ), whereas other ‘differences’ seem to have originated in the state of preservation of the studied material; for instance, the byssus groove of Cyamiocardium , which, contrary to the ‘absence’ mentioned by Soot-Ryen (1951; for Cyamiocardium denticulatum ) is present in that genus ( Urcola & Zelaya, 2018; present study). Consequently, anatomy does not help to separate most of the genera of Cyamiidae considered in the present study.

Thiele (1934) and Chavan (1969) differentiated the genera of Cyamiidae based on the general shell outline and shell sculpture, as follows: Cyamiomactra subtrigonal and smooth; Cyamium transversely elongated and smooth; Kingiella ovate (longer than high) and with strongly radially sculptured; Perrierina , Legrandina , Cyamiocardium and Pseudokellya roundish to ovate, smooth or radially ribbed; and Ptychocardia oblong (higher than long), sculptured with numerous radial folds, some of them strong and producing undulations in the ventral margin. Undoubtedly, Engl (2012) followed this distinction when he placed Pseudokellya georgiana in Ptychocardia again. The above-mentioned shell characters seem to be inadequate for a distinction of these genera, because some Cyamiocardium species are subquadrate or rhomboidal ( Cyamiocardium chuanisinense , Cyamiocardium denticulatum and Cyamiocardium namuncurense ), and some species currently regarded under Cyamiomactra are not subtrigonal or smooth (e.g. Cyamiomactra chilensis Ramorino, 1968 and Cyamiomactra falklandica Dell, 1964 ). Considering shell outline and shell sculpture, Heteromactra should be regarded as a synonym of Cyamium and not of Cymiomactra, as currently considered (e.g. Lamy, 1906, 1910, 1911, 1917; Chavan, 1969). In a similar manner, Zelaya & Ituarte (2009) interpreted the existence of a continuum in the shell outline and sculpture of Pseudokellya species, ranging from the rounded and weakly radially sculptured Ps. cardiformis to the rhomboidal and strongly radially sculptured Pt. georgiana , and also including the ovate and smooth Ps. franki .

The number, morphology and arrangement of hinge teeth, ligaments and their supports have been (and still remain at present) used as (some of) the main morphological characters for delimiting genera of bivalves. However, and surprisingly, only two different hinge conformations are recognized when regarding the Cyamiidae considered in the present study. One conformation is present in Cyamium (as redescribed herein), Cyamiocardium , Cyamiomactra , Heteromactra , Reloncavia , Perrierina and Legrandina , the last two taxa with additional ‘crests’ along the dorsal margin, although these were interpretated as secondary acquisitions by Ponder (1971); in fact, these may be poorly developed in some species, such as Legrandina harrisonae Powell, 1935 and Perrierina matai Fleming, 1948 . The second conformation is present in the hinge of Pseudokellya and Ptychocardia . This low degree of variability in the overall architecture of the hinge exhibited by cyamiids could reflect: (1) that the general hinge conformation (and the anatomy of the soft part) have remained unchanged (‘conserved’) among different cyamiid genera; (2) that the hinge conformation (and anatomy) experienced phenomena of parallelism/convergence among different cyamiid genera; or, more probably, (3) that the generic diversity of Cyamiidae is considerably lower than previously thought, with several of the genera that are currently regarded as distinct being synonyms. The previous proposals by Lamy (1917) to consider Cyamiocardium as a synonym of Cyamiomactra and by Ponder (1971) to regard Cyamiomactra as a subgenus of Cyamium support the last hypothesis. However, the currently available molecular information is not sufficient to be conclusive about this issue; further taxa and markers need to be studied.

DISCREPANCIES OF THIS STUDY WITH PREVIOUSLY PROPOSED SCHEMES

The present study reveals that Cyamiidae and Gaimardiidae are closely related families. This grouping contrasts with the most traditional point of view, which considered them as belonging to two different superfamilies: Gaimardioidea and Cyamioidea ( Thiele, 1934; Chavan, 1969; Vokes, 1980). Although Ponder (1971) had previously suggested that these taxa were closely related (he considered them as two subfamilies of Cyamiidae ), this proposal did not receive acceptance by subsequent authors. In fact, in most recent studies both taxa were regarded as belonging to different superfamilies (e.g. Bieler et al., 2010; Lemer et al., 2018).

The difference of the scheme proposed in the present study from that of Ponder (1971) is not restricted to the taxonomic rank assigned to Cyamiidae /Cyamiinae and Gaimardiidae /Gaimardiinae, but also concerns the placement of Kidderia . According to Ponder (1971), Kidderia is closer to Cyamium than to Gaimardia (consequently, he regarded it as a Cyamiinae), whereas according to the present study, Kidderia is more similar to Gaimardia than to Cyamium , consequently being included in Gaimardiidae and not in Cyamiidae . The family placement for Kidderia proposed in the present study is in agreement with the previous usages by Odhler (1924), Thiele (1934) and Dell (1964); the placement in all these studies was based upon the characteristics present in the type species of Kidderia and the closely similar K. bicolor . However, the possibility that some other species currently placed under Kidderia could correspond to Cyamiidae cannot discarded, becuse Kidderia is presently considered as ‘possibly polyphyletic’ ( Ponder, 1971).

Another point of discrepancy with the previous literature concerns the placement of Pseudokellya . This genus was largely regarded as belonging to Lasaeidae / Kelliidae (Galeommatoidea) (e.g. Chavan, 1969; Vokes, 1980; Engl, 2012). The information obtained in the present study (namely, the presence of a follicular epithelium surrounding the oocytes and embryos and the presence of tentacles at the posterior end of the mantle border) allows us to confirm that Pseudokellya is a Cyamioidea , as previously reported by Thiele (1934) and Zelaya & Ituarte (2009).

CONCLUDING REMARKS AND FUTURE PERSPECTIVES

The present study reveals Cyamioidea as a monophyletic superfamily of bivalves. Members of this superfamily vary greatly in shell morphology and anatomy. However, all of them share a reproductive character: the presence of a follicular epithelium, which appears first at the early stages of oogenesis and persists until larvae are fully developed, brooded within the female gills. The follicular epithelium is unusual among bivalves and, in fact, it is not known for any other bivalve family; thus, it is valuable to define Cyamioidea . Curiously, and for the first time among marine bivalves, a reproductive character proves to be more informative than shell morphology and anatomy for the definition of a superfamily.

Despite the restricted significance of morphological and anatomical characters for defining Cyamioidea , they appear to be useful for delimiting families within this superfamily: Cyamiidae and Gaimardiidae , with the latter previously regarded as belonging to a different superfamily (Gaimardioidea). Neoleptonidae , previously regarded as cyamioideans, are definitively removed from this superfamily and reallocated into Veneroidea based on the molecular evidence obtained in the present study. The placement of Sportellidae in Cyamioidea requires confirmation.

As part of the present study, we confirm that Gaimardia and Kidderia can be reunited in Gaimardiidae and that Cyamiocardium , Cyamium , Heteromactra , Pseudokellya , Ptychocardia and Reloncavia belong to Cyamiidae . Within the latter group, we also tentatively place Cyamiomactra and Jukesena , although some anatomical and molecular evidence suggests that they might, in fact, correspond to a different family of Cyamioidea . Likewise, the relationship of Legrandina and Perrierina to Cyamiidae requires further study.

The synonymy or, alternatively, the eventual validity and relationship among different (sub)genera of Cyamiidae and Gaimardiidae also need to be investigated in more detail, particularly the affinities of Eugaimardia and Progaimardia with Gaimardia and of Costokidderia and Cyamionema with Kidderia . In contrast, cyamiids now appear to be defined mainly upon shell characters. The hinge conformation, usually considered as a key character for defining genera of bivalves, does not provide a clear distinction among Cyamiocardium , Cyamiomactra , Cyamium , Heteromactra , Legrandina , Perrierina and Reloncavia or between Pseudokellya and Ptychocardia . The general shell outline and shell sculpture, currently used to define these genera, prove to be more variable than previously thought, and the same characteristics appear in different taxa. Furthermore, the molecular studies performed in the present study, although providing full support for the entire superfamily, do not resolve the relationships among (sub)genera. All these facts imply either that several of these names are currently being used ambiguously or that some of them correspond to synonyms. The addition of more anatomical and molecular information for other species (in particular, molecular studies involving faster-evolving genes and further taxonomic sampling) appears crucial for an adequate re-evaluation of the validity of the different genera. We prefer to refrain from introducing nomenclatural changes or including new synonymy for these taxa until further study has taken place.