Solemya notialis, Abstract, 2009

Simone, Luiz Ricardo L., 2009, Comparative morphology among representatives of main taxa of Scaphopoda and basal protobranch Bivalvia (Mollusca), Papéis Avulsos de Zoologia 49 (32), pp. 405-457 : 423-450

Solemya notialis new species

( Figs. 39-46 , 111-117 )

Solemya patagonica : Rios, 1975: 186 (pl. 60, fig. 912-left) (part) (non E.A. Smith, 1885).

Solemya occidentalis : Rios, 1985: 207 (pl. 75, fig. 1058); 1994: 224 (pl. 78, fig. 1105) (non Deshayes, 1857).

Types: Holotype MZUSP 88440 . Paratypes: BRAZIL. Rio de Janeiro. Cabo Frio , MZUSP 35257 , 1 specimen (Paulo Gonçalves col.; v/2002) ; Off Maricá , 23°08’S 42°47’W, IBUFRJ 2172 . 1 specimen (NOAS sta. CF-VII-6165) ; off Saquarema , 22°59’S 42°19’W, IBUFRJ 1335 , 1 specimen (Geocosta Rio II sta. B3, 20/iii/1986) ; Angra dos Reis , 31 m depth, MZUSP 20396 , 2 specimens (R.V. Emilia sta. 53; 29/vi/1966) , 27 m depth, MZUSP 20397 , 1 specimen (R.V. Emilia sta. 54; 29/vi/1966) . São Paulo. Off Ubatuba , 33.5 m depth, MZUSP 20395 , 1 specimen (R.V. Emilia sta. 12; 17/xii/1965) .

Type locality: BRAZIL. São Paulo. Off Ubatuba , 23°25’S 44°43’W, 33.5 m depth (R.V. Emilia sta. 12, 17/xii/1965) .

Description

Shell ( Figs. 39-46 ): Thin, translucent, fragile, slightly rectangular; about 2.7 times longer than tall. Anterior and posterior ends rounded, similar-sized (anterior end slightly wider). Periostracum thin, translucent, glossy, beige-brown with radial, pale-beige bands, slightly more concentrated anteriorly ( Figs. 39, 40, 45 ); clearer close to umbones; extending about 1/4 beyond calcareous portion of shell ( Figs. 42, 45, 46 ), attached to ventral connection between both lobes of mantle ( Fig. 114 : pe). Umbones located preceding anterior quarter of shell length, not-protruded, flattened. Pair of oblique, shallow and narrow furrows, running gradually anteriorly to umbones towards posterior and ventral ( Figs. 41, 43 ). Ligament simple, restricted to posterior region of mantle, being abruptly wider posteriorly, reaching about 1/6 of dorso-ventral shell height ( Figs. 43, 44 : li, 46). Hinge edentulous. Inner surface white, bearing transverse, fine striation ( Fig. 44 ), gradually becoming oblique in anterior and posterior ends. Scar of adductor muscles approximately equal-sized; each one equivalent to 15% of calcareous portion of each valve ( Figs. 44, 46 ); anterior adductor scar located close to dorsal edge, just posterior to anterior quarter of valve’s length; posterior adductor scar located at short distance from posterior end ( Figs. 46 , 111 ). Pallial line simple, thick, located between middle and ventral thirds of shell height.

Main muscle system ( Figs. 111, 113 , 116 ): Anterior adductor muscle elliptical (with longer axis parallel to adjacent shell border) in section, attached in area equivalent to 15% of each valve area. Posterior adductor muscle similar in characters to anterior muscle, section slightly more rounded (for position of adductor muscles, see shell description). Pedal protractor muscle of foot ( Fig. 116 : fp) broad and thin; originating surrounding ventral edge of anterior adductor muscle on both sides; running immersed in ventro-anterior wall of visceral mass, disappearing along foot dorsal base. Pair of anterior pedal retractor muscles ( Figs. 111 , 116 : fa) narrow and long; originating just dorsal and posterior to anterior adductor muscle, in area equivalent to 1/15 of adductor muscle, antero-posteriorly elongated; running ventral and posteriorly with uniform width along its length; inserting splaying along wall between visceral mass and antero-dorsal foot base. Pair of median-anterior foot retractor muscles ( Fig. 116 : fb) very narrow and long, filiform; originating at small points approximately in middle of distance between umbones and anterior adductor muscles in area equivalent to 1/5 that of origin of anterior pedal retractor muscle; running towards ventral; inserting along ventro-anterior region of visceral mass, close to median line. Pair of auxiliary middle pedal retractor muscles ( Figs. 111 , 116 : fr) broad and thick; originating in umbonal cavity close to median line, antero-posteriorly long (about 4 times longer than wide); running towards ventral, narrowing gradually; inserting just internally to insertion of anterior pedal retractor muscles. Middle posterior foot retractor muscle ( Figs. 116, 117 : fv) a single muscular layer covering posterior wall of visceral mass anterior to pericardium; originating as wide, transverse horseshoe (concavity anterior) located between anterior and middle thirds of distance between umbones and posterior end, anterior region thicker (almost as thick as auxiliary middle pedal muscles), directed internally, remaining regions about half thinned, performing an arc; running ventrally separating reno-pericardial structures from gonad; inserting along postero-ventral walls of foot. Pair of posterior pedal retractor muscles ( Figs. 111 , 116 : fm), long and thick; originating just posterior and dorsal to posterior adductor muscle in approximately same area as adductors, slightly antero-posteriorly long (about twice longer than wide); running towards ventral and anterior almost half shell length, with uniform width along their length; inserting covering insertion of middle posterior foot retractor muscle in postero-ventral region of foot base. Large transverse muscle ( Fig. 117 : tm) located just ventral to pericardium, connecting both sides of middle posterior foot retractor muscle, approximately in its central region; size about 1/8 that of anterior adductor muscle. Pallial muscles see below (mantle).

Foot and byssus ( Figs. 111 , 116 ): Foot with about 1/3 of inner shell volume, position directed anteriorly and ventrally. Foot cylindrical, with half to a third shell length in retracted condition. Foot distal tip umbrella-like, edges thinner, undulating and with short, stubby papillae; expanding externally about double the foot base width. No byssus.

Mantle ( Figs. 111, 113-115 ): Mantle lobes about 25% larger than calcified portion of valves. Border of mantle lobes widely fused and thick, muscular. Pedal aperture anterior length about 1/3 that of shell. Excurrent aperture ( Figs. 113, 115 : se) located at middle level of posterior border about 8 times smaller than pedal aperture; edges simple; a series of papillae flanking excurrent aperture located at midway between inner edges and periostracum insertion ( Fig. 115 : ma); a larger dorsal papilla with approximately same length of excurrent aperture and width about 1/3 of adjacent mantle width, separated from excurrent aperture a distance equivalent to half its length; 2 pairs of papillae positioned between dorsal papilla and dorsal edge of excurrent siphon, outer pair with about half of dorsal papilla size, median pair with a quarter dorsal papilla size; 9 papillae surrounding ventral half of excurrent aperture, each one similarly sized to each other and with about 1/3 of dorsal papilla size. Pallial muscles mostly obliquely disposed ( Figs. 113, 114 : pm), with about double shell thickness; inserted in thick pallial line positioned between middle and ventral thirds of calcareous part of shell ( Figs. 44, 46 , 111, 114 ); about 2/3 of portion of mantle containing pallial muscles covered only by periostracum ( Fig. 114 ). Pallial longitudinal muscle running along ventral pallial fusion ( Fig. 114 : lm) from pedal to excurrent apertures; slightly wider than shell’s calcareous thickness; immersed dorsally by oblique muscles and ventrally by pallial tissue. Mantle edge located beyond calcareous portion of shell about 1/4 of its width ( Fig. 114 ), possessing two folds, outer folds with tip rounded, about as tall and calcareous shell thickness, periostracum attached to their inner surface; inner fold thin (about 1/4 of outer fold), tall and fused to its pair.

Pallial cavity ( Fig. 111 ): Occupying about 75% of each valve surface and more than half of shell volume. Pair of palps narrow and long (pp), about 5 times longer than tall, length approximately half of shell height; located about midway between anterior adductor muscle and gills’ anterior end. Palps inner surface smooth. Gills bipectinate, with about half shell length and height; located in posterior half of shell. Dorsal gill filaments slightly larger, more obliquely positioned and possessing more rounded tip than ventral filaments; ventral filaments with pointed tip and positioned vertically. Gill filaments with skeletal rod lying at external edge, occupying about 1/3 of filament width ( Fig. 112 ). Gill central, horizontal axis with efferent gill vessel externally and afferent gill vessel internally; gill longitudinal muscle lying along outer region of efferent gill vessel ( Fig. 112 : gm). Suprabranchial chamber small, mostly compressed by posterior foot retractor and adductor muscles and by kidneys ( Figs. 116-117 ).

Visceral mass ( Figs. 116, 117 ): Volume about 1/4 that of shell, located in central region of dorsal half. Mostly compressed by pedal and adductor musculature, divided almost completely in two halves by middle posterior foot retractor muscle (fv) (described above). Anterior half bent on dorsal surface of foot. Region anterior to middle posterior foot retractor muscle mostly filled by pale-cream gonad; region posterior to that muscle mostly filled by reno-pericardial structures, being pericardium dorsal and kidneys ventral, somewhat equivalently sized.

Circulatory and excretory systems ( Figs. 116, 117 ): Located in triangular space between middle posterior foot retractor muscle and posterior pedal retractor muscle; occupying about 1/4 of visceral volume. Pericardium oval, dorso-ventrally flattened; with about 90% of shell width. Pair of auricles lateral, positioned horizontally and transversally at middle pericardial region; connecting posterior end of gill efferent vessel with antero-lateral regions of ventricle. Ventricle oval, tapering posteriorly, with equivalent size of auricles and with about 1/5 of pericardial volume; surrounding 80% of intestinal loop crossing through pericardium. Kidneys solid, white, fulfilling ventral region of pericardium and dorsal surface of posterior pedal retractor muscle.

Digestive system ( Figs. 116, 117 ): Pair of palps described above (pallial cavity) ( Fig. 111 ). Mouth very small, located at central region between both palps, compressed between foot and anterior portion of visceral mass. Esophagus filiform, running between both auxiliary middle pedal retractor muscles up to their middle level; suddenly running posteriorly, gradually expanding and forming stomach; lacking clear limit between esophagus and stomach. Stomach with about 1/10 of shell length, about three times longer than wide; inner surface smooth, simple; pair of ducts to digestive diverticula in anterior third of gastric lateral walls. Digestive diverticula small, mixed with gonad. Intestine also filiform, possessing single loop anterior to middle posterior foot retractor muscle, along median plane; crossing dorsal edge of this muscle, running along median line almost straight, edging hinge, passing between origins of both posterior pedal retractor muscles and along dorsal and posterior surface of posterior adductor muscle. Anus simple, sessile, located in middle level of posterior surface of posterior adductor muscle.

Genital system: Gonad described above (visceral mass). No detected genital duct or pore.

Central nervous system ( Fig. 116 ): Cerebral ganglion a single, curved mass anterior to middle portion of esophagus, surrounding middle level of auxiliary middle pedal retractor muscle (fr); about 1/10 thickness of anterior adductor muscle in wider (anterior) region. Cerebral ganglion running along both sides horizontally towards posterior, through gonad, gradually narrowing and becoming connectives with visceral ganglion; no clear separation between ganglion and connective. Cerebro-visceral connective crossing laterally in dorsal region of middle posterior foot retractor muscle (fv) ( Fig. 117 ); crossing through dorsal muscular posterior wall of visceral mass ( Fig. 117 : mv), after this, running between gills base and dorsal region of posterior pedal retractor muscles (fm). Visceral ganglion a single transverse mass, slightly wider than preceding connective and with about 1/5 of cerebral ganglion size; located in middle level of dorsal third of posterior pedal retractor muscles. Two pairs of conspicuous lateral nerves running from each ganglion, running straight towards anterior in case of cerebral ganglion and on opposite side in visceral ganglion. No detectable pedal ganglion.

Measurements (in mm): Holotype: 10.9 by 4.1 by 2.3; paratypes 20395: 10.3 by 3.4 by 2.2.

Distribution: Brazil, from Rio de Janeiro to Northern São Paulo.

Habitat: Unconsolidated substrate, about 30 m depth.

Material examined: Types.

Discussion

Solemya notialis has a shell very similar to S. occidentalis , and has thus far been identified as such. Nevertheless, important anatomical differences reveal the specific separation. S. notialis has much longer palps than S. occidentalis ( Mikkelsen & Bieler, 2008) , which has a comparable palp size to S. cf. australis ( Reid, 1998, fig. 5-8E). On the other hand, there are some species lacking palps and entire digestive tubes ( Reid, 1980). Those species that still bear digestive systems, even slightly reduced, still maintain suspension feeding ( Krueger et al., 1992). S. notialis differs from the western American species S. reidi Bernard, 1980 , in having shallower pre-umbonal concavity, and from S. valvulus (Carpenter, 1864) in having a wider anterior region. S. notialis still differs from S. patagonica E.A. Smith, 1885, which has northern register in south of Rio Grande do Sul coast, in having shorter shell length (length/height tax of S. notialis is about 2.8; while that of S. patagonica is about 3.2), clearer color and less developed radial sculpture.

Discussion of Characters

Shell

1. Portion of shell weakly calcified between two calcified portions (ligament): 0 = absent; 1 = present ( Bivalvia) (CI = 100; RI = 100).

The ligament of the bivalve shell is one of the conspicuous synapomorphies of the class. The structure, however, is part of the shell and is produced by a portion of the dorsal fold between both mantle lobes ( Fig. 35 : hf). The ligament, different from the remaining shell regions, is only or mainly constituted by conchiolin and is weakly calcified ( Trueman, 1952; Owen, 1959), providing the abduction movement of the valves, antagonistic of the adductor muscles.

2. Hinge: 0 = absent; 1 = present ( Bivalvia); 2 = taxodont ( Propeleda carpentieri , Ennucula puelcha , Barbatia cancellaria ) (CI = 66; RI = 85).

The hinge is the interlocking region between both valves of the bivalve shell, which is connected by the ligament. This region of the shell is produced by the lateral surfaces of the dorsal fold between both mantle lobes ( Fig. 35 : hf), and as such fulfill the space between a tooth and respective socket. In the case of the taxodont type of hinge, a series of somewhat uniform teeth are present. These appear in the basal bivalve branches, and raise the other types of hinges.

3. Slit in posterior shell aperture: 0 = absent; 1 = present ( Coccodentalium carduus , Paradentalium disparile ) (CI = 100; RI = 100).

Although a slit is present in the posterior mantle border of all examined scaphopods, a slit at the shell’s posterior aperture is only found in the above mentioned species. Besides this character, no others are hereby applied to scaphopods because of polarization problems; however, further exploration of shell characters can be found in previous comparative analyses (e.g., Emerson, 1952; Scarabino, 1995; Steiner, 1996).

4. Periostracum: 0 = ending with or short beyond calcified shell region; 1 = extending long beyond it ( Solemya spp ) (CI = 100; RI = 100).

The normal fashion of the periostracum is to extend little beyond the calcified portion of the shell. This free periostracum surrounds the calcified edge and encases in the inner surface of the outer mantle edge fold ( Bottjer & Carter, 1980). In the case of solemyids, however, this free portion of the periostracum is much longer, extending about a quarter of the calcified portion beyond its edges ( Figs. 45, 46 ).

5. Umbos: 0 = absent; 1 = opisthogyre ( Solemya spp , Propeleda carpentieri , Ennucula puelcha ); 2 = prosogyre ( Barbatia cancellaria , Serratina capsoides ) (CI = 100; RI = 100).

A portion of the shell called umbos is the part which is oldest. This portion, exclusive of the bivalves, is a pair, i.e., an umbo in each valve. In the case of the protobranchs, the structures are opisthogyre, meaning they are turned backwards ( Fig. 37 ). This feature is particularly clearer in the nuculids, as their umbos are protruded. In the remaining protobranchs, however, the umbos are flat and difficultly individualized ( Figs. 39, 40, 45 ). On the other hand, the other bivalves have prosogyre umbos, i.e., turned forwards, being a notorious lamellibranch synapomorphy. Despite the fact that the states of this character are not considered additive, the result shows that they possibly are. Analyzing the cladogram, the opisthogyre umbos appear to be a bivalve synapomorphy, becoming prosogyre at node 9, which represents the lamellibranchs.

Several other shell features were researched, but were not utilized here due to an autapomorphic nature in the present sample, or because of difficulty in polarizing. One of them is the nacre, inner aragonite layer of the shell. This layer is present in basal taxa of most conchiferan classes, notoriously in Gastropoda, Cephalopoda and Bivalvia; and mostly lost in higher taxa ( Watabe, 1988). However, the nacreous layer is absent in Scaphopoda, in bivalve Solemyidae and in several lamellibranchs, including both sampled here. considering nacre in this analysis, it assists in support of node A (Conchifera minus monophacophorans), with remarkable reversions in nodes 2, 6 and 9. On the other hand, nacreous structure was considered apomorphic in bivalves (Giribet & Wheeler, 2002).

Another example is the shell sculpture.The feature is very difficult to polarize, as all kinds of sculptures are found both in ingroups and outgroups. Longitudinal striae are, however, found as basal stock, having smooth and annulated forms as derived states in scaphopod analyses (e.g., Reynolds, 1997; Steiner, 1998; Reynolds & Okusu, 1999). Similar difficulty in polarizing exists for other scaphopod shell attributes, such as curvature, placement of wider region, etc., which certainly are worthy of a longer analysis of the class.

Mantle

6. Mantle: 0 = conic; 1 = divided into two lobes ( Bivalvia, Scaphopoda) (CI = 100; RI = 100).

The conic state is found in those classes that possess a shell practically restricted to the dorsal region of the body; the cone can be flattened, as in monoplacophores, or deep (a bind-sac) as in gastropods and cephalopods. The lobed state is something like a horse saddle that mostly covers the dorsal and lateral regions of the animal’s body. In the case of bivalves, this bilobed condition persists from the larval phase to adulthood, and each lobe secretes each shell valve. A hinge fold of the mantle ( Fig. 35 : hf) lies on the intersection of both lobes, building the hinge and the ligament. While in the case of scaphopods, the bilobed condition is restricted to the larval phase. The ventral edges of these lobes fuse ventrally, producing a tubular conformation, which secretes the tubular shell.

Lateral mantle lobes have been used to base Diasoma upon some previous phylogenetic analyses (Steiner, 1996; as Loboconcha; Giribet & Wheeler, 2002, characters 50, 51).

7. Lobes free ventrally in adult phase: 0 = absent; 1 = present ( Bivalvia except Solemya ) (CI = 100; RI = 100).

As explained in the previous character, the characteristic state of scaphopods is to have a fusion of the mantle lobes during metamorphosis. However, a similar process happens in solemyids, which possess mantle lobes that are totally fused ventrally ( Figs. 42 , 113-115 ) ( Beedham & Owen, 1965). In fact, the solemyid mantle is fashioned in a very similar way to those of scaphopods. Analyzing the allocation of the character on the cladogram, the fusion of mantle lobes in metamorphosis appears to be a basal Diasoma feature; freedom of the ventral edge of the mantle lobes is only based on node 7.

Different degrees of mantle fusion are, on the other hand, observed in lamellibranchs (e.g., hiatellids – Simone & Penchaszadeh, 2008). However, those cases are convergences or, in the light of the present study, possibly reversions.

8. Mantle edge: 0 = mono- or bifolded; 1 = trifolded ( Ennucula puelcha , Barbatia cancellaria , Serratina capsoides ) (CI = 100; RI = 100).

The mantle edge of shelled mollusks is particularly an important structure, as it secretes most of the shell and bears some receptors to interact with the environment. The mantle edge has one or two lobes in most molluscan classes, including the basal taxa of the ingroup ( Figs. 47, 49 , 67 , 114 ). Protobranch nuculids ( Fig. 96 ) and lamellibranchs are those which have a trifolded mantle edge condition, revealing to be a conspicuous node 8 synapomorphy. In the embryology, the middle fold in higher bivalves appears later ( Morton et al., 1998, fig. 4.8A), which also indicate via ontogeny the apomorphic state of the trifolded mantle edge.

Additional features can be found in the epithelium of scaphopod anterior mantle edge, which can differentiate both orders ( Steiner, 1992a, 1996: 331).

9. Pallial cavity: 0 = opened ventrally; 1 = opening antero-posteriorly ( Bivalvia, Scaphopoda); 2 = a blind-sac ( Propilidium curumim , Nautilus pompilius ) (CI = 100; RI = 100).

The pallial cavity is one of the outstanding features of Mollusca, protecting important and delicate structures such as gills, osphradia, etc. The organization and the placement of this cavity are, however, differently structured in the molluscan classes. In the presently considered outgroups monoplacophores and polyplacophores, the pallial cavity is merely a furrow which surrounds the foot. This condition also appears in very basal, Cambrian mollusk-like forms [e.g., Odontogriphus omalus Morris, 1976 ( Caron et al., 2006); Wiwaxia corrugata (Matthew, 1899) ( Morris, 1985; Eibye-Jacobsen, 2004)]. In Conchifera, this condition is modified, except for monoplacophores. in the case of Cyrtosoma (node B), the pallial cavity is deeper, resembling a profound blind-sac. In the case of the Diasoma, the pallial cavity is arranged in an antero-posterior manner, flanking laterally the visceral sac, with a flow of water from anterior or antero-ventral to posterior. This condition (state 1) is an important synapomorphy of node 1.

Head and Foot

10. Head and appendages: 0 = absent; 1 = present ( Propilidium curumim , Nautilus pompilius ) (CI = 100; RI = 100).

This character was inserted in order to organize the pair of cyrtosome outgroups. It is important, however, to emphasize that no further research on this issue was performed, as Cyrtosoma is beyond the scope of this paper.

11. Foot position: 0 = ventral; 1 = mostly turned forwards (anterior) ( Bivalvia, Scaphopoda) (CI = 100; RI = 100).

A crawling foot is the normal fashion of Mollusca, lying along the ventral surface of the animal’s body. This attribute is found in all classes except those above mentioned, in which the foot instead is directed forward or, at least, in an antero-ventral manner ( Figs. 119, 120 ). The foot, in these cases, and as another indicative of homology, separates the mouth and respective appendages (palps or captacula) from the remaining pallial cavity.

The infaunal mode of life is not the only explanation for this feature, as with several other digging molluscs, such as some aplacophorans and gastropods (e.g., olivids and volutids), do not possess the same form.

As the foot of the Cephalopoda is highly modified, the state is not clear. however, the structures possibly derived from the foot, such as the siphon and ventral part of the arms, are ventrally positioned ( Shigeno et al., 2007).

12. Foot retractor muscles: 0 = 7-8 pairs; 1 = 4 pairs ( Bivalvia, Scaphopoda) 2 = 2 pairs ( Propilidium curumim , Nautilus pompilius ) (CI = 100; RI = 100).

Several pairs of foot retractors are the normal state in basal mollusk branches, including the conchiferan Monoplacophora ( Lemche & Wingstrand, 1959). A reduction of the number of pedal retractors is one of the synapomorphies that substantiates a branch of the Conchifera that excludes Monoplacophora, and groups Cyrtosoma and Diasoma. This taxon (node A) is still not named.

Despite the fact that this character appears divided into two apomorphic states that support Diasoma and Cyrtosoma respectively ( Fig. 122 , nodes 1 and B), it most probably is part of a single evolutionary tendency towards the simplification of the number of foot retractor muscles.

13. Foot main form of operation: 0 = peristaltic contractions; 1 = hydraulic inflation ( Bivalvia , Scaphopoda ) (CI = 100; RI = 100).

14. Foot function: 0 = crawling; 1 = digging ( Bivalvia , Scaphopoda ) (CI = 100; RI = 100).

The characters 13 and 14 refer to adaptations for digging. In these cases, there is a normal mode of working: the foot is introduced in the soft sediment, it inflates the tip, and, by means of foot retractor muscles, it pulls the remaining of the animal’s body inside the sediment. The constituents of the other classes, even those that are active diggers and have an infaunal mode of life, do not possess such modifications, and peristalsis is still present.

The typical Diasoma foot is, then, a somewhat hollow cylinder, in which fluid content contributes to its inflation. The inflation is wider in the foot’s tip.

Of course the basic form of the foot in Bivalvia and Scaphopoda was modified several times along their evolution. In the former, this structure even disappeared in several taxa. However, in these cases, the modifications appear to be derived from the basic plan described above.

A “Burrowing foot with anterior enlargement” was used as a character uniting scaphopods and bivalves in a morphological study by Giribet & Wheeler (2001: 303, character 109).

15. Foot distal region: 0 = simple; 1 = umbrella-like ( Coccodentalium carduus , Paradentalium disparile , Solemya spp , Propeleda carpentieri , Ennucula puelcha ) (CI = 33; RI = 60).

The tip of the foot of the species above listed has an additional modification for digging. There is an expansion surrounding the tip that assists the animal in anchoring into sediment ( Drew, 1900). This expansion is hereby called “umbrella-like” because of a similarity in appearance to an opened umbrella. This feature is found in an animal that is actively exploring a new environment. Analyzing the result, it is possible to infer that the umbrella-like foot appeared at the base of Diasoma (node 1), suffering a simplification in scaphopod node 4 and in lamellibranchs (node 9). This last taxon is also designated Pelecypoda, i.e., axe-like foot.

The lateral expansions of the foot in scaphopods have been named crenulated fringe and anchoring organ ( Steiner, 1992a, b, 1998). They are defined as a pair of epipodial flaps with rippled surfaces and crenellated edges, forming a dorsally interrupted collar ( Steiner, 1992b: 186).

It is equally parsimonious to consider state 1 as a Diasoma synapomorphy (node 1), with a reversion in Gadilidae (node 4), or a convergence between Dentalliidae and Bivalvia (nodes 3 and 5). The first optimization is shown in Fig. 122 , as the gadilids still remain a terminal disk when the foot is totally extended ( Steiner, 1992b, figs. 17, 18). As the homology between the gadilid terminal disk and the dentaliid umbrella-like foot tip is unclear, the gadilid state is coded as absent here. Despite this conservative approach, this concept is subject to change with further studies. In this case, the umbrella-like foot is maintained as a Diasoma synapomorphy, with a single reversion in lamellibranchs (node 9).

Pedal features are one of more clear similarities between bivalves and scaphopods, and were explored in previous phylogenetic proposals. Steiner (1996), for example, stated “contractile, burrowing foot, with longitudinal muscles associated with pedal wall, transverse muscles present, lateral epipodial lobes as anchoring organ” as synapomorphies of Diasoma (as Loboconcha).

16. Proboscis form of foot: 0 = absent; 1 = present ( Gadila braziliensis , Polyschides noronhensis ) (CI = 100; RI = 100).

The inversible foot of Gadilida representatives (node 4) resembles the proboscis of some gastropods in its organization ( Figs. 79 , 80 , 88 ). There is a pair of retractor muscles (fm) that can be used to retract and enfold the distal half of the foot. These disappear inside the more basal portion of the foot-visceral mass axis ( Steiner, 1992b; this study); only a small apical orifice remains in the fully retracted position ( Figs. 78 , 80 ). This is a notorious synapomorphy of the taxon and is responsible for the name Siphodentaliidae (a siphon-bearing tusk-shell, an allusion to the strange foot), a synonym of Gadilidae . The “proboscis” form is based on structural similarities between the gadilid foot and the gastropod proboscis. However, a comparison with nemertean proboscis is also found in the literature ( Steiner, 1992 b, Palmer & Steiner, 1998). A More detailed study on the pedal musculature has been used for phylogeny overtures (Steiner, 1996, 1998; Reynolds & Okusu, 1999). In such, a correlation between the internal layer of the longitudinal pedal muscles of Dentaliida and the Gadilida pedal inner retractor muscles is demonstrated. Additionally, transverse muscles appear to be reduced in Gadilida , amplifying the pedal sinus.

17. Retraction of postero-dorsal foot base: 0 = muscular integument; 1 = pair of conspicuous retractor muscles ( Bivalvia, Scaphopoda) (CI = 100; RI = 100).

The posterior region of the foot of most molluscan classes is retracted normally by a muscular integument, rather than a developed pair of muscles. This condition is not the case with bivalves and scaphopods. A conspicuous pair of posterior retractor muscles is present in both taxa which supposedly appear to be homologues. In the case of scaphopods, this pair of muscles surrounds the visceral mass and is positioned dorso-laterally as longitudinal muscles ( Figs. 48 , 53 , 65 , 76 , 88 : lm), while in bivalves, the pair is positioned more ventrally ( Figs. 99 , 116 : fm).

Suggestions on the homology of some muscular structures are represented in Figs. 118-120 , in which directly related structures are indicated, and some others are suggested with grey marks. Two of them, the scaphopod longitudinal muscles (lm) and the bivalve posterior pedal retractor muscles (fm), are mentioned above. Along the same point of view, the posterior origin of the longitudinal muscle (po) may be homologous to that of the bivalve retractor muscle. The scaphopod visceral muscle (vm) appears to be homologous to the bivalve middle posterior foot retractor muscle (fv). The difference in position is due to the more dorsal position of the scaphopod longitudinal muscles and the more ventral position of the bivalve retractor muscle. Both can be derived from laterally positioned muscles in rostroconchs ( Fig. 118 ), by opposite migration (dotted arrows). Further information on the similarities of pedal musculature between scaphopods and bivalves is found in Steiner (1992b, 1996) and in scaphopod inner branches in Steiner (1998).

18. Byssal gland: 0 = absent; 1 = present ( Barbatia cancellaria , Serratina capsoides ) (CI = 100; RI = 100).

A gland that secretes byssus is not present in protobranch bivalves. However, it being a remarkable lamellibranch synapomorphy. Most possess that gland only in early development ( Yonge, 1962), as is the case of Serratina . On the other hand, some pedal glands in nuculids have been referred to as possible byssal glands ( Rhind & Allen, 1992), an issue that merits further evaluation. This character was divided in three ordered states by Giribet & Wheeler (2001, character 105): absent, present in larvae and adults, and lost in adults.

19. Adductor muscles: 0 = absent; 1 = present ( Bivalvia); 2 = with two components ( Ennucula puelcha , Barbatia cancellaria , Serratina capsoides ) (CI = 100; RI = 100).

A possible homologous muscle to the bivalve’s anterior adductor muscle is the buccal muscle ring of scaphopod larva ( Wanninger & Haszprunar, 2002, fig. 2D-E: bm), which occupies a similar position in front of the esophagus and bears transverse muscle fibers. Despite the fact that the states of this character are not considered additive, the present result demonstrates that they can be regarded as such. The simple pair of adductor muscles is one of the Bivalvia synapomorphy, but a clear division between quick and slow components of each muscle ( Villarroel & Stuardo, 1998) is only detected in node 8.

20. Pallial component of main retractor muscle of foot: 0 = absent; 1 = present ( Scaphopoda) (CI = 100; RI = 100).

The pallial component originated from each longitudinal muscle is notable in the examined scaphopods ( Figs. 48 , 67 , 119 : pm). Although this pair of muscles is very thin in Gadilida .

Visceral Mass

21. Transverse muscles through surrounding visceral sac: 0 = absent; 1 = present ( Bivalvia , Scaphopoda ) (CI = 100; RI = 100).

22. Visceral sac: 0 = filling dorsal region of shell; 1 = exposed in pallial cavity ( Bivalvia , Scaphopoda ) (CI = 100; RI = 100).

In mollusks, the visceral mass, or visceral sac or hump, is normally a dorsal continuation of the foot. Rarely there is a clear separation between foot and viscera. Yet characters 21 and 22 explore special features of the Diasoma classes, which are absent in the remaining ones. In both scaphopods and bivalves, the visceral sac protrudes into the pallial cavity, forming a bulged sac at the ventral base of the foot. Additionally, transverse muscles are present, crossing from one side to another, through visceral glands (gonad and digestive) and the intestinal loops. These transverse muscles are possibly an adaptation to improve the hydrostatic high pressure necessary to project the foot forward.

In cephalopods, which have a visceral sac somewhat similar to the above mentioned features, important differences are present ( Simone, 1997 c). The visceral organs are not protected by a muscular integument, and neither transverse muscles are present.

Pallial Organs

23. Pair of gills: 0 = present; 1 = absent ( Scaphopoda) (CI = 100; RI = 100).

Total loss of the gills is a long known synapomorphy of Scaphopoda. Some pallial structures have been described as additional respiratory structures, as secondary gills (e.g., Boss, 1982). Nothing like that was found in the examined species, except for some series of low folds ( Figs. 47 , 67 : pf; 78: pl), which look more like glands than pallial folds. These folds are interesting differences amongst the species and appear to hold possible importance for scaphopod taxonomy.

The gonad of the Gadilida is intensely folded, and is separated from the posterior region of the pallial cavity by a very thin integument ( Figs. 78 , 81 , 87 : go). This special conformation certainly assists in blood oxygenation, as also do the gastric folds (digestive diverticula) of the Dentaliida ( Figs. 47 , 67 : dg).

24. Gills modified for filtering: 0 = absent; 1 = present ( Barbatia cancellaria , Serratina capsoides ) (CI = 100; RI = 100).

The formal taxon Lamellibranchia is named in order to appraise this important structure for nonprotobranch (and non-septibranch) bivalves (node 9). The increase in pairs of gills, each one forming a pair of demibranchs, is one of the necessary modifications for filter feeding, which is so characteristic of most bivalves.

25. Gills modified to support symbiotic bacteria: 0 = absent; 1 = present ( Solemya spp ) (CI = 100; RI = 100).

The thick gills filaments of solemyids are one of the more interesting features of this protobranch family. The large sized gills possibly raised some confusion in early literature, as protobranchs supposedly should possess only small, merely respiratory gills. However, increased solemyid gills appear to be a convergence, which, instead of being used for filtration, are used for symbiotic purposes ( Cavanaugh, 1983; Conway & Capuzzo, 1991; Coan et al., 2000). A parallel condition is also found in lamellibranch Lucinidae ( Reid & Brand, 1986; Glover & Taylor, 2001).

26. Gills positioned separating infrabranchial from suprabranchial chambers: 0 = absent; 1 = present ( Ennucula puelcha , Barbatia cancellaria , Serratina capsoides ) (CI = 100; RI = 100).

In most protobranch bivalves, gills serve merely a respiratory purpose, as in other mollusks. For this propose, gills are relatively free in the pallial cavity. In nuculids and in lamellibranchs, however, each gill is arranged in order to separate an infra- from a suprabranchial chamber. Ciliary connections attach the gills to neighboring structures, such as mantle lobes and visceral sac ( Figs. 95 , 101 ), which, in some eulamellibranchs, are substituted by tissue connections. In this arrangement, the flow of water necessarily must pass through gill filaments to be exteriorized. Some nuculids, additionally, have also facultative capacity to collect suspended particulate material, passing it anteriorly to the palp ( Stasek, 1961; Reid, 1998: 237), showing that the arrangement of gills in nuculids is a step towards the lamellibranch condition.

27. Gill suspensory stalk connected to ventral surface of posterior adductor muscle: 0 = absent; 1 = present ( Ennucula puelcha , Barbatia cancellaria , Serratina capsoides ) (CI = 100; RI = 100).

The gill suspensory stalk is a reinforced rod located between both demibranchs that support the posterior region of the gills. Its origin is on ventral surface of the posterior adductor muscle and is usually reduced, membrane-like in some eulamellibranchs, as the gills are connected to adjacent structures. However, it is muscular and retractile in nuculids and in filibranchs.

Reno-Pericardial Structures

28. Pericardium: 0 = surrounding heart; 1 = empty ( Coccodentalium carduus , Paradentalium disparile ); 2 = absent ( Gadila braziliensis , Polyschides noronhensis ) (CI = 100; RI = 100; non additive).

The loss of the heart is a well known scaphopod feature ( Reynolds, 1990a, b, 2002). However, dentaliids still retain a pericardial chamber ( Figs. 50, 51 : pc). This is totally absent in Gadilida . Although states were treated in a non additive optimization, an arrangement of the vestigial pericardium as an intermediary step towards its total loss is intuitive, and is represented in Fig. 122.

29. Auricle connection to ctenidial vein: 0 = terminal; 1 = subterminal ( Barbatia cancellaria , Serratina capsoides ) (CI = 100; RI = 100) (? in Scaphopoda).

This character, exclusive of the lamellibranchs, is related to the increase of gills. A portion of the gill augmented forward and the connection to auricles emmerge approximately at mid-level. The anterior portion of gills has an antero-posterior flow, while the posterior one has a contrary course. 30. Kidney: 0 = narrow; 1 = thick-glandular ( Bivalvia, Scaphopoda, Propilidium curumim , Nautilus pompilius ) (CI = 100; RI = 100).

Most basal mollusks, including the monoplacophorans, has narrow kidneys somewhat splayed along the haemocoel. The remaining Conchifera have kidneys concentrated in the sub-pericardial region forming a solid triangular chamber, which appears as a notable synapomorphy. For additional discussion on the molluscan kidney see Morse & Reynolds (1996), Ruthensteiner et al. (2001).

Digestive System

31. Mouth sphincter: 0 = developed; 1 = weak or absent ( Bivalvia, Scaphopoda) (CI = 100; RI = 100).

The occurance of a mouth sphincter is the rule in most molluscan classes. However, in scaphopods and bivalves this muscle is reduced or absent. The mouth, then, remains as a permanently opened orifice.

As explained in character 36, the scaphopod mouth is in the base of the conic oral tube ( Figs. 53-55 , 68 , 81 : mo), instead of in its distal tip.

32. Pair of lateral folds of mouth for food capture: 0 = absent; 1 = present ( Bivalvia , Scaphopoda ) (CI = 100; RI = 100).

33. Captacula: 0 = absent; 1 = present ( Scaphopoda ) (CI = 100; RI = 100).

34. Pair of palps: 0 = absent; 1 = present ( Bivalvia ) (CI = 100; RI = 100).

The homology between bivalve palps and the scaphopod captacula has been suggestively proposed in previous phylogenetic analyses (Steiner, 1996) as large labial appendages, as it is here corroborated ( Fig. 121 ). This proposition is based on several factors explained as follows:

The captacula clearly appear as lateral expansions of the mouth during early development of the scaphopods ( Wanninger & Haszprunar, 2002, fig. 1C). Afterwards the slender expansions grow. This is exactly the way that the bivalve palps appear and remain during the course of a lifetime. Additionally, the captacula are certainly separated from the foot since early development ( Shimek & Steiner, 1997) and, as in palps, do not appear to constitute a pedal structure.

Considering that the scaphopod mouth is at the base of the oral tube (see character 36), both captacula and palps are lateral expansions of this structure, being located in the same region.

The captacula originate from a double folded base that surrounds the mouth dorsally and laterally ( Figs. 52 , 54 ). A double folded flap also raises the palps ( Fig. 121 ).

Originally, both structures are designed to work in the same way. They are inserted into the surrounding sediment in search of food, which is then captured and conducted to the mouth. In filter-feeding lamellibranchs, however, the palps are modified to collect food from the proper pair of gills.

Another piece of evidence for the homology between palps and captacula is their innervations. In both structures, anterior nerves from the cerebral ganglia are responsible for sensorial and locomotive features ( Steiner, 1991, 1992a).

Supposed homologous structures are indicated schematically in Fig. 121 (in grey), which represents the above discussion.

Scaphopod captacula have been proposed to be homologous to cephalopod arms. However, dorsal arms of cephalopods, which occupy an equivalent location to the captacula base, are much more complexly organized as expansions of the muscular head-foot portion. It therefore more closely resembles gastropod cephalic tentacles rather than the thin, muscle-less scaphopod structure.

In relation to basal bivalves, nuculid palps have another character in common with those of lamellibranchs in that they are suspension-feeding organs ( Reid & Brand, 1986; Reid, 1998).

35. Proboscis of palps: 0 = absent; 1 = present ( Propeleda carpentieri , Ennucula puelcha ) (CI = 50; RI = 0).

The proboscis component of labial palps is a remarkable feature of most protobranchs ( Stasek, 1965). It is absent in the palps of the solemyids. However, the structure is almost useless, as most nutrients comes from the symbiotic mode of life in gills. Palps are almost atrophied in Solemyidae and even absent in some species.

It is equally parsimonious to consider that state 1 is a synapomorphy of node 7 with a reversion in eulamellibranchs (node 9), or a mere convergence between both above species. The former optimization is shown in Fig. 122 , as the proboscis is widely present in most protobranchs.

Regarding palps, Giribet & Wheeler (2002) considered two characters that resulted in relation to nuculids and nuculanids: a hypertrophied palp and palp appendages (= proboscis) (characters 80, 81 of that paper).

36. Conic oral tube: 0 = absent; 1 = present ( Scaphopoda) (CI = 100; RI = 100).

As suggested in the characters above related to the mouth, the oral tube appears to be a structure located in front of the mouth, homologous to the remaining mollusks, and is located at the base of the oral tube ( Figs. 53-55 , 68 , 81 , 121 : mo). The oral tube is a flexible and muscular cone projected forwards, normally full of prey, such as foraminifers and more rarely, small mollusks and other organisms.

The homology of the mouth at the base of the oral tube, instead of its tip, is mainly based on the position of the nerve ring and odontophore. These structures are always located close to the mouth in Mollusca. In the case of the nerve ring, those of the Gadilida are considered ( Figs. 80, 81 , 88 ). As in Dentaliida , the structure is decentralized.

The oral tube is also called proboscis, and the mouth is considered to be its tip ( Steiner, 1992a, Palmer & Steiner, 1998). This concept is explained above, but is not applied here.

37. Oral tube arrangement: 0 = absent; 1 = three internal chambers ( Coccodentalium carduus , Paradentalium disparile ); 2 = four equidistant projections at edge ( Polyschides noronhensis , Gadila braziliensis ) (CI = 100; RI = 100).

The important structure for scaphopods additionally includes the above mentioned differences that, so far, have upheld the two traditional orders (nodes 3 and 4). 38. Odontophore: 0 = as part of buccal mass; 1 = as appendix of esophagus ( Scaphopoda ); 2 = absent ( Bivalvia ) (CI = 100; RI = 100).

39. Radula : 0 = present; 1 = absent ( Bivalvia ) (CI = 100; RI = 100).

40. Horizontal muscle (m6): 0 = wide; 1 = narrow ( Polyschides noronhensis , Gadila braziliensis ) (CI = 100; RI = 100).

41. Pair of ventral tensor muscles of radula (m11): 0 = present; 1 = absent ( Scaphopoda ) (CI = 100; RI = 100;? in Bivalvia ).

42. Transverse muscle of odontophore (m3): 0 = absent; 1 = present ( Scaphopoda ) (CI = 100; RI = 100;? in Bivalvia ).

The odontophore and radular features explored here (characters 38-42) demonstrate some peculiarities of a highly modified structure. All bivalves lack any trace of odontophore, and such a structure is not indicated in embryological studies (e.g., Reverol et al., 2004; Costa et al., 2008). The odontophore loss is, then, a remarkable Bivalvia synapomorphy. On the other hand, nothing with respect to this structure can be inferred, and for this reason, a question mark represents the above character in the given matrix ( Table 1).

The odontophore of Scaphopoda is different from the normal pattern of the structure in other mollusks. It is positioned away from the esophageal axis, outside the buccal cavity. A short odontophoral tube ( Figs. 27, 28 : on) separates it from the remaining digestive tube. There is no clear ventral tensor muscles of the radula, although the pair of dorsal tensor muscles is very thick ( Figs. 59-62 , 71, 72 , 82-85 , 90-92 : m4). Therefore, the antagonistic muscle of the dorsal tensors is absent, and the radular return action, usually done by the ventral tensor, is a mystery. In the Dentaliida , there are two muscles that connect both sides of both odontophore cartilages, forming a strong ring ( Figs. 59, 60 , 71 : m3 and m6). One of them, however, is very narrow in the Gadilida ( Figs. 84, 85 , 91, 92 : m6). Observing the muscular and radular arrangement, it is possible to deduce that the scaphopod radula is not used to scratch the food, but to crush it. As the oral tube is full of foraminifer testa, which are not found after the odontophore level, i.e., in the stomach and intestinal loops, it is possible to infer that each prey is brought inside the odontophore to be squeezed.

Further comparison amongst the scaphopod radulae is found in the literature ( Morton, 1959; Steiner, 1996, 1998; Reynolds & Okusu, 1999). However, no applicable character was found to be considered here. The more important difference is in the rachidian, which is rectangular and smooth in the Dentaliida ( Figs. 8, 9, 14, 15 ), while the Gadilida is thin, narrow and somewhat irregular ( Figs. 21, 22, 26 ). This feature was explored phylogenetically at the ordinal level (Steiner, 1996: 330).

43. Stomach: 0 = wide; 1 = inconspicuous (a simple curve) ( Polyschides noronhensis , Gadila braziliensis , Solemya spp ) (CI = 50; RI = 66).

Scaphopods above mentioned, and possibly every Gadilida , lack any clear gastric region. A simple, several-looped intestine follows the esophagus ( Figs. 80, 81 , 88 ). This feature is convergently found in solemyids, which possess a reduced digestive tube because of the symbiotic mode of life; some species have even lost their digestive tubes ( Reid, 1980).

44. Stomach sorting areas: 0 = absent; 1 = present ( Propeleda carpentieri , Ennucula puelcha , Barbatia cancellaria , Serratina capsoides ) (CI = 100; RI = 100).

45. Style sac and crystalline style: 0 = absent; 1 = present ( Ennucula puelcha , Barbatia cancellaria , Serratina capsoides ) (CI = 100; RI = 100).

A large and complex stomach is a possible consequence of the simplification of the buccal mass in bivalves, and as such some characters are explored above (44, 45). Internal gastric folds and sub-chambers are found in bivalves after the node 7. A clear style sac, which forms a crystalline style, is found after node 8 ( Graham, 1949; Purchon, 1956; Halton & Owen, 1968; Villarroel & Stuardo, 1998; personal obs.).

46. Structure annexed to stomach: 0 = digestive gland; 1 = hollow digestive diverticula ( Bivalvia, Scaphopoda) (CI = 100; RI = 100).

Different from the solid digestive gland of most mollusks, the above taxa possess hollow diverticula directly connected to the stomach. This is a further modification for nutrient absorption that possibly characterizes Diasoma. However, in some classes that were not included in this study, Solenogastres and Caudofoveata, a single and large hollow chamber appears to work as a digestive gland ( Scheltema, 1993).

Certainly, the digestive gland and diverticula merit further studies with respect at least to their homology. Few studies have so far being realized related to protobranchs and scaphopods. An example reveals extracellular digestion postulated for nuculids ( Owen, 1959, 1972), and dentaliids ( Taib, 1981; Palmer & Steiner, 1998).

Central Nervous System

47. Central nervous system: 0 = surrounding esophagus/mouth; 1 = ganglia except cerebral removed from esophagus ( Coccodentalium carduus , Paradentalium disparile , Propeleda carpentieri , Solemya spp , Ennucula puelcha , Barbatia cancellaria , Serratina capsoides ) (CI = 50; RI = 80).

There is a clear tendency for the position of the ganglia of the central nervous system to be distant from the esophagus in most Diasoma, although some basal taxa still maintain a nerve ring surrounding it ( Figs. 80, 81, 86 , 93, 94 , 116 ).

It is equally parsimonious to consider state 1 as a Diasoma synapomorphy (node 1), with a reversion in Gadilidae (node 4), or a convergence between Dentaliidae (node 3) and Bivalvia (node 5). The first optimization was applied in Fig. 122.

48. Well-differentiated ganglia: 0 = absent; 1 = present ( Coccodentalium carduus , Paradentalium disparile , Ennucula puelcha , Barbatia cancellaria , Serratina capsoides ) (CI = 50; RI = 75).

49. Cerebral ganglia location: 0 = central, far from anterior adductor muscle; 1 = lateral, close to anterior adductor muscle ( Ennucula puelcha , Barbatia cancellaria , Serratina capsoides ) (CI = 100; RI = 100).

50. Visceral ganglia: 0 = inconspicuous; 1 = large ( Bivalvia ) (CI = 100; RI = 100).

51. Position of visceral ganglia: 0 = between posterior adductor muscle and posterior pair of foot retractors; 1 = ventral to posterior adductor muscle ( Barbatia cancellaria , Serratina capsoides ) (CI = 100; RI = 100).

In both Diasoma classes a clear tendency for ganglia to become distant from each other appears. In scaphopods, Gadilida still remain circum-buccal ganglia, which is also found in some protobranchs such as nuculanids and partially in solemyids. In the remaining taxa of both classes, the ganglia are positioned far, such as in Dentaliida ( Plate, 1892; Palmer & Steiner, 1998: fig. 10.11), and in bivalves after node 8 ( Burne, 1904; this study). The epiatroid condition of the nervous system, i.e., the cerebral and pleural ganglia closed or fused, was used to support Diasoma in previous phylogenetic analyses (Steiner, 1996).

Discussion of the Cladogram

General Aspects and Paleontology

In this section systematics and characters are discussed under light of the obtained cladogram ( Figs. 122 , 123 ), taking also into consideration the results of previous analyses of other authors. As stated above, comparative molecular studies have produced low resolution cladograms ( Steiner & Dreyer, 2003). However, the scaphopod relationship with bivalves is common ( Dreyer & Steiner, 2004). In paleontologic and evolutionary scenarios, inferences are also found in literature; evolution of the class Bivalvia has been debated, and even ancestors or early forms have been proposed (e.g., Morton et al., 1998, fig. 4.5).

Taking into consideration Paleontology, the Cambrian genus Heraultia has been indicated as the stem group of Diasoma, raising scaphopods, rostroconchs and bivalves ( Pojeta et al., 1972; Runnegar & Pojeta, 1974). Other less precise scaphopod stem groups have also been erected, such as Paleozoic Xenoconchia ( Starobogatov, 1974; Chistikov, 1979), a shell-less ancestor ( Yochelson (1978), and monoplacophoran stocks ( Edlinger, 1991; Carter et al., 2000). Also, Rostroconchia has been indicated as stem group of both classes ( Runnegar & Pojeta, 1974; Salvini-Plawén, 1980; Steiner, 1992a).

Despite some controversies with respect to the early fossil records of scaphopods ( Engeser & Riedel, 1996; Yochelson, 1999; Yochelson & Goodison, 1999; Palmer, 2001), the oldest recognizable scaphopod is Ordovician Rhytiodentalium kentuckyesnis Pojeta & Runnegar, 1979 . On the other hand, recognizable bivalves have been found since the early Cambrian ( Runnegar & Bentley, 1983; Pojeta & Runnegar, 1985; Bengston et al., 1990; Runnegar & Pojeta, 1992).

Naming Rostroconchia as stem group of Diasoma ( Scaphopoda plus Bivalvia) presents a dilemma in consideration of the following: 1) the univalve, pseudobivalved shell, and 2) the early appearance of the Bivalvia. Paleo-larval studies of rostroconchs ( Pojeta & Runnegar, 1985) revealed that the larva is univalved, becoming bivalve-like during metamorphosis. This shell is compressed laterally and possesses a pair of inseparable valves, united by continuous shell layers in dorsal margin. Inferred internal anatomy shows similarities with early bivalves ( Pojeta & Runnegar, 1985), except that rostroconchs bear a radular apparatus. The life history of these groups shows that rostroconchs are highly diverse in the lower Paleozoic, mainly in the Ordovician, being gradually replaced by bivalves ( Pojeta, 1978). Both occupied similar niches, an infaunal position anchored by foot in the sediment. Possibly the transition from rostroconchs to bivalves may be preceded by a phase lacking a ligament, but possessing adductor muscles (Runnegar & Pojeta, 1985). The dorsal shell layer between valves could be flexible enough to permit some adduction and abduction. Parallel evolution to closure of the shell aperture only by means of shell flexibility is found in some other groups of molluscs, such as the gastropod sacoglossans Ascobulla Marcus, 1972 and Cylindrobulla Fisher, 1857 ( Volvatellidae ). In these taxa an adductor-like muscle works in the shell aperture/closure, pressing the outer lip towards the inner lip ( Marcus, 1972; Mikkelsen, 1998), functioning only on the calcified shell wall. Even in the bivalve-like crustacean Conchostraca, in which the rostroconchs were previously classified (e.g., Nicholson & Etheridge, 1880; Kobayashi, 1933), a calcified dorsal layer functions withing the valve movement ( Eriksen & Brown, 1980). The same occurs in the bivalved crustacean Ostracoda ( Kornicker & Sohn, 1976). Both crustacean groups have a pair of adductor muscles. On the other hand, the fact that in early development the scaphopod shell is univalve has been used as an argument for contradicting the monophyly of the Diasoma ( Wanninger & Haszprunar, 2001).

Observing only the mantle, bivalves still retain a single piece that expands on both sides, forming the mantle lobes, each one producing a valve. The mantle is still continuous dorsally, showing no clear clue of valve separation except for a fold that builds the hinge ( Fig. 35 : hf). This fold stays introduced between each hinge tooth and its respective socket at the opposed valve. Its distal edge secretes the ligament. The mantle of bivalves, if the shell is extracted, is identical to the conformation of the mantle of scaphopods in early development ( Wanninger & Haszprunar, 2002, figs. 1A-B: pro).

The Ordovician ribeiroid rostroconch Pinnocaris Etheridge, 1878 have been erected as the stem group of scaphopods ( Runnegar & Pojeta, 1974; Pojeta & Runnegar, 1976, 1979). However, based on a different kind of protoconch, this possibility has been rejected ( Peel, 2004). A conocarioid rostroconchs has been, then, proposed to originate the class ( Yochelson, 1978; Peel, 2004). On the other hand, the theory of Pinnocaris , or Pinocaris -like basic stock for scaphopods does not seem strange, despite the argumentation regarding the protoconch position ( Peel, 2004) and body axis (Steiner, 1992). A scaphopod-like creature can be extracted from Pinnocaris , by means of a greater and curved development mainly from the anterior region. This modification could explain the posteriorization of protoconch ( Sasaki, 2007) and the rather oblique longitudinal axis in Scaphopoda.

Discussion of Each Branch of the Cladogram

( Fig. 123 )

Node 1: This branch represents a monophyletic relationship between Bivalvia (node 5) and Scaphopoda (node 2), which can be called Diasoma as originally introduced ( Runnegar & Pojeta, 1974), Ancrypoda ( Hennig, 1979), or Loboconcha ( Salvini-Plawén, 1980). This branch is the main focus of the present paper and is supported by 14 morphological synapomorphies as follows: the mantle divided into two lobes (at least in a phase of development) (character 6); an antero-posteriorly opened pallial cavity (9); the foot turned forward (11), with four pairs of retractor muscles (12), working mainly by hydraulic inflation for digging (13, 14), with an umbrella-like distal tip (15); the presence of transverse muscles in visceral mass (21); visceral sac exposed in the pallial cavity (22); mouth lacking sphincter (31); a pair of lateral expansions for food capture (32); the hollow digestive diverticula (46); and a tendency for ventral ganglia of nerve ring to be positioned far from esophagus (47).

On the other hand, molecular studies have sometimes revealed support of a Scaphopoda-Cephalopoda relationship, such as 18S rDNA ( Steiner & Dreyer, 2003). However, according to this same method, Bivalvia and Gastropoda are not supported ( Steiner & Müller, 1996; Steiner & Hammer, 2000). In a total-evidence approach, monophyly of Scaphopoda + Bivalvia is supported by morphological data, with Bremer index of 3 (Giribet & Wheeler, 2002, fig. 2); however, this relationship disappears with the addition of molecular data, transferring scaphopods to a place closer to cephalopods (Giribet & Wheeler, 2002, fig. 6).

The common morphological characters between Bivalvia and Scaphopoda have been considered mere convergences induced by the burrowing characters, such as enclosure of the body by the mantle and shell, and the burrowing foot innervated by concentrated pedal ganglia ( Steiner & Dreyer, 2003: 352). However, as commented above, such kind of modifications are not found in other burrowing mollusks, as, e.g., neogastropods Olividae and Volutidae , vetigastropod Umbraculinae, stromboidean Aporrhaidae , and many others. This leads to an interpretation that infaunal habits appear as insufficient explanations to justify the common organization between bivalves and scaphopods.

Another interesting argument in favor of the inclusion of Scaphopoda in Cyrtosoma, i.e., closer to Cephalopoda and Gastropoda instead of Bivalvia, is the appearance of cephalic retractors ( Wanninger & Haszprunar, 2002). However, such a character looks rather like a plesiomorphy. Bivalves have a totally reduced head and buccal apparatus. The absence of muscles that retract an atrophied region comprises a synapomorphy, i.e., the loss of cephalic retractors in bivalves that are present in scaphopods and cyrtosomes, and even in other mollusks.

In Summary, in the literature (e.g., Engeser & Riedel, 1996) there are propositions that Scaphopoda is considered a branch of Rostroconchia. This is hereby corroborated. However, Bivalvia must be included as well. Consequently, in the point of view of the present result, Rostroconchia is not an isolated extinct class, and two rostroconch branches are still living today – Bivalvia and Scaphopoda. In other words, bivalves and scaphopod are existent, modified, special rostroconchs. A parallel concept is applied in relation to birds, which are considered to be modified, existent dinosaurs ( Currie et al., 2004).

Node 2: This branch represents the class Scaphopoda, being well-supported by 8 synapomorphies, despite this not being the main goal of the present study. Among them and worthy of remark are: a branch of the longitudinal muscle coming from the mantle (character 20); loss of gills and heart (23, 28); the captacula (33); a conic oral tube (36); the odontophore as an appendix of the esophageal axis (38); loss of the ventral tensor muscle of the radula (m11) (41); and a great development of an auxiliary, single muscle from horizontal muscle, the m3 (42). The class is also well-supported by molecular approaches ( Steiner & Dreyer, 2003).

Nodes 3-4: The division of scaphopods obtained here agrees with the current classification, which divides the class into two orders: Dentaliida (node 3) and Gadilida (node 4) (e.g., Steiner & Kabat, 2001; Steiner & Dreyer, 2003). Dentaliida (node 3) is supported by 3 synapomorphies: the slit at posterior shell aperture (character 3); the tree internal divisions of the oral tube (37); and the differentiation of ganglia of the central nervous system (48). Gadilida (node 4) is supported by 7 synapomorphies, of which the more important are: the proboscis-fashion of the foot (character 16); the total reduction of the pericardium (28); the four equidistant projections at oral tube edge (37); the narrow odontophore horizontal muscle (m6) (40); and reduction of the stomach (43).

Node 5: This branch represents the class Bivalvia, supported by seven synapomorphies despite the fact that the investigation of the monophyly of this taxon is not the main goal of this study. From the synapomorphies, the interesting ones are: the shell having ligament, hinge and umbos (characters 1, 2, 5); the pair of adductor muscles (19); the loss of buccal mass structures (38, 39); and enlargement of the pair of visceral ganglia (50).

The first three branches of bivalves on the cladogram ( Figs. 122 , 123 , nodes 6-9) demonstrate that the taxon Protobranchia is a paraphyletic arrangement of basal Bivalvia. However, the monophyly has been postulated, based on shared features of endoderm reorganization and organogenesis during and after metamorphosis ( Gustafson & Reid, 1986, 1988; Reid, 1998). Some classifications still divide the Protobranchia into two orders: Nuculoida and Solemyoida (e.g., Sanders & Allen, 1973), based mainly on the contrast between a taxodont (former) and a cryptodont (last) hinge. Besides, even independent origin of protobranchs and lamellibranchs has been advocated in the literature ( Reid, 1998), based on different ligaments ( Waller, 1990; Cope, 2000), particularities of the stomach ( Graham, 1949; Purchon, 1956, 1987), and dissimilar form of larva – pericalymma versus veliger ( Gustafson & Reid, 1986, 1988; Cragg, 1996). On the other hand, paraphyly of protobranchs is the result of other phylogenetic analyses (e.g., Giribet & Wheeler, 2002).

Solemyidae , the first bivalve branch (node 6) appears to be quite ancient, with records since the Silurian, such as Janeia silurica Liljedahl, 1984 . However, it has been postulated that the family had originated from nuculoids ( Allen & Sanders, 1969; Kuznetzov & Sileiko, 1984; Reid, 1998; Carter et al., 2000). Additionally, some members of solemyids from Ordovician to the Carboniferous possess a taxodont hinge ( Reid, 1998; Cope, 2000), suggesting that the edentulous hinge of modern forms may be derived from it. This is an indicator that the taxodont kind of hinge (character 2, state 2) may be at the base of Bivalvia, i.e., in node 5 ( Fig. 123 ). Its appearance in node 7 is actually an artifact due to further modification of the modern examined solemyids.

Node 6: The node represents the family Solemyidae , supported by three synapomorphies such as: the extension of the periostracum longer beyond the calcified shell edge (character 4); the modifications of the gill to support symbiotic bacteria (25); and the reduction of the stomach (46). Solemyids have been, neverrtheless, considered at the base of a monophyletic Protobranchia in some phylogenetic approaches (e.g., Purchon, 1987, Cope, 1996, 1997). However, they are sometimes considered to be derived taxon ( Waller, 1990, 1998; Salvini-Plawén & Steiner, 1996). This approach is possibly based on a set of highly derived characters that modern solemyids have, which, when analyzed alone, suggest they could not be placed at the base of protobranchs. Those characters are: the lack of nacre, the edentulous hinge, the large gill (especially for a protobranch), the high degree of fusion of the mantle lobes, and the excurrent siphon. Conversely, under light of the present result, those derivations can be interpreted as remarkable convergences with other higher branches of Bivalvia, and that modern solemyids are highly modified animals. This is expressed, for example, by the well-developed foot of Solemya that even helps the animal to swim ( Reid, 1980; 1998: fig. 5.9).

Node 7: This branch reunites bivalves except Solemyidae , and is supported by four synapomorphies. They are: the taxodont hinge (character 2, but see observation above, in last paragraph of the discussion on node 5); the freedom of mantle lobes (7); the proboscis on palps (35 – which can also be transferred to the base of Bivalvia with further studies on the symbiotic solemyids); and the presence of sorting areas in stomach (44).

Node 8: This branch groups nuculids and lamellibranchs, supported by seven synapomorphies. The more important are: the trifolded mantle edge (character 8); the presence of two (quick and slow) components in the adductor muscles (19); the anatomical position of the gill separating an infra- from a suprabranchial chamber (26); the development and position of gill suspensory stalks (27); the gastric style sac (45); and differentiation of the central nervous system in differentiated ganglia placed away from each other (48, 49).

Some phylogenetic approaches have shown the Nuculoida as stem group of the lamellibranchs (e.g., Cope, 2000; as Autolamellibranchiata Grobben, 1894). However, the molecular data reveals Nuculanidae at the base of Autolamellibranchiata (Giribet & Wheeler, 2002, fig. 11). Another feature that could support this node is pedal feeding in early juveniles; this feature is common to nuculids and most lamellibranchs ( Reid et al., 1992). Nuculidae appeared in the Ordovician ( Keen, 1969; Reid, 1998).

The lamellibranchs are represented in node 9 by a filibranch and an eulamellibranch. Although, as mentioned above, no exhaustive research for lamellibranch synapomorphies was performed herein, the main characters from the seven that support this branch are: prosogyre umbos (character 5); loss of the umbrella-like foot distal end (15); the byssal gland (18); further gill modification for filter feeding (24), which is a character that could be dismembered in several others, such as the presence of demibranchs, food groves, etc.; the subterminal connection of auricles in the gills (29); loss of proboscis of palps (35); and the position of the visceral ganglia ventral to posterior adductor muscle (51).

Beyond these morphological characters, the absence of haemocyanin of the lamellibranch blood, a common feature in protobranchs and other mollusks (Morse et al., 1986), can be another synapomorphy.

The argumentation on independent origin between protobranchs and lamellibranchs (beginning of this section), can be used to further base the lamellibranch set of synapomorphies. Those characters erected as differences between both groups can perfectly be considered lamellibranch modifications.

Most phylogenies based mainly on morphology have shown monophyly of lamellibranchs, although normally protobranchs are not always properly tested ( Starobogatov, 1992; Cope, 1996; Morton, 1996; Waller, 1998; Steiner & Hammer, 2000). The age of appearance of the filibranch branch is late Cambrian or early Ordovician ( Cope, 1996, 2000), showing a very ancient path of evolution. The Cardiolarioidea appears to be the basic stock of the taxon ( Cope, 2000).

Node B: This branch is part of the outgroups, which, as explained above, do not have an exhaustive search for synapomorphies. Two synapomorphies support this node, although many others can be evoked. One of them is the derived position of the visceral connectives median to the dorso-ventral muscles ( Haszprunar & Wanninger, 2000).

Figures

FIGURE 121: Schematic representation of structures annexed to mouth of a protobranch bivalve (above) and a scaphopod (below), frontal-slightly right view.Captacula (cp) only partially represented. Oral tube (ot) represented as if it were transparent with mouth at base instead of its tip. Not to scale. Suggestions of homology of mouth position are based on topology and on location of nerve ring, just posterior to it in more basal taxa of both classes.

FIGURE 122: Most parsimonious cladogram with synapomorphies of each node represented by symbols (black square: non-homoplastic synapomorphy; white square: reversion; circle: convergence). The number above indicates the character, the number bElow indicates its state. Four outgroups also shown in the bottom. Length: 65; CI: 89; RI: 95.

FIGURE 123: Most parsimonious cladogram with each node numbered (ingroup) or named by letters (outgroups at right). Some suprafamiliar taxa already acknowledged indicated by dotted arrows. Length: 65; CI: 89; RI: 95.

FIGURES 39‑46: Shell aspects of Solemya notialis n. sp: 39) Holotype, left view, length = 10.6 mm; 40) Same, right view; 41) Same, dorsal view; 42) Same, ventral view; 43) Same, detail of peri-umbonal region, left-slightly dorsal view, scale = 1 mm; 44) Same, right valve, inner view, detail of hinge and anterior region, scale = 1 mm; 45-46) Paratype MZUSP 20397, right valve with extended periostracum, outer and inner views respectively, length = 12.9 mm.

FIGURES 111‑115: Solemya notialis anatomy: 111) Whole right view, part of right mantle lobe removed, some structures show by artificial transparency; 112) Gill, transverse section in its middle region; 113) Mantle border, right-slightly dorsal view, remaining portion of mantle lobes removed, topology of some adjacent structures also shown; 114) Same, transverse section in middle level of ventral region; 115) Mantle border in region of excurrent siphon, posterior-outer view. Scales = 1 mm.

FIGURES 116‑117: Solemya notialis anatomy: 116) Digestive tubes, main musculature and central nervous system in situ, right view, semi-diagrammatic representation artificially transparent, topology of some adjacent structures also shown; 117) Posterior region of visceral mass, dorsal view, dorsal wall of pericardium and adjacent mantle removed. Scales = 1 mm.

FIGURES 30‑38: Shell and anatomical aspects of Ennucula puelcha (MZUSP 19101; length = 13.0 mm): 30) Right valve, inner view; 31) Same, left valve; 32) Dorsal-slightly right view; 33) Left valve, detail of hinge, inner view; 34) Same, right valve; 35) Dorsal region with left valve somewhat still in situ, dorsal-slightly right partial view; 36) Whole shell, ventral view, valves somewhat deflected; 37) Whole left view, left mantle lobe partially removed; 38) Anterior-slightly ventral view.

FIGURES 47‑52: Coccodentalium caduum anatomy: 47) Whole ventral view, mantle opened longitudinally along median line and deflected; 48) Whole left view; 49) Detail of posterior region, dorsal-slightly right view; 50) Detail of middle and posterior region of pallial cavity roof, ventral view; 51) Anterior region of visceral mass, just posterior to haemocoel, transverse, semi-schematic section; 52) Foot and anterior region of haemocoel, left view. Scales = 2 mm.

FIGURES 65‑74: Paradentalium disparile anatomy: 65) Whole right view; 66) Detail of posterior region, dorsal view; 67) Whole ventral view, mantle opened longitudinally along median line and deflected; 68) Anterior region, mainly showing digestive tubes, right view, foregut opened longitudinally, topology of some adjacent structures also shown, trajectory of intestine indicated in indicated schematic representation at left, a transverse section of esophagus in indicated level also shown; 69) Odontophore, ventral view; 70) Same, dorsal view; 71) Odontophore, posterior view, radula completely removed, right m4 deflected; 72) Same, m3 sectioned transversally, both cartilages and left m4 deflected; 73) Pedal ganglia and statocysts, ventral view; 74) Cerebral ganglia, ventral view. Scales = 1 mm.

FIGURES 95‑100: Ennucula puelcha anatomy: 95) Whole right view, right valve and part of mantle right lobe removed; 96) Mantle edge, transverse section in middle portion of its ventral region; 97) Gill, transverse section in its middle region along 5 filaments; 98) Posterior half of visceral mass, dorsal-slightly right view, right mantle lobe and dorsal portion of pericardium removed; 99) Digestive tubes, main musculature and central nervous system in situ, right view, semi-diagrammatic representation artificially transparent; 100) Region of posterior adductor muscle, right-slightly posterior view, pallial lobes fully removed. Scales = 1 mm.

FIGURES 118‑120: Schematic representation of main retractor muscles of foot with indication of possible homologous structures. The grey marks are suggestions of homologies; 118) scheme of a transverse section in middle level of foot-visceral axis in a rostroconch, with putative intermediary lateral position of pair of longitudinal muscles, and generic basal scaphopod and bivalve, in which that pair of muscles migrated respectively towards dorsal and ventral (dotted arrows), originating from the longitudinal muscles and the posterior retractor pedal muscles (lm, fm). 119) A generic scaphopod; 120) A protobranch bivalve. Not to scales. Some structures seen by transparency.

FIGURES 75‑79: Polyschides noronhensis anatomy: 75) Whole left view; 76) Whole dorsal view; 77) Detail of posterior region, dorsal view; 78) Whole ventral view, mantle opened longitudinally along median line; 79) Haemocoel, dorsal view, mainly showing pedal musculature, foot and visceral sac opened longitudinally. Scales = 0.5 mm.

FIGURES 80‑86: Polyschides noronhensis anatomy: 80) Haemocoel, dorsal view, mainly showing pedal musculature, digestive tubes and part of nerve ring, foot and visceral sac opened longitudinally; 81) Digestive tubes, dorsal view, topology of some adjacent structures also shown; 82) Odontophore, ventral view, superficial layer of membranes removed; 83) Same, whole ventral view; 84) Odontophore, dorsal view, radula removed (only radular sac topology indicated), right m4 deflected; 85) Same, m3 sectioned transversally, both cartilages and m4 deflected; 86) Nerve ring, anterior view. Scales = 0.5 mm.

FIGURES 87‑94: Gadila brasiliensis anatomy: 87) Whole ventral view, anterior half of pallial cavity opened longitudinally along median line; 88) Foot and haemocoel, right view, more superficial structures, mantle and integument extracted, a portion of left retractor muscle of foot extracted to show intestine; 89) Odontophore, ventral view, superficial layer of membranes removed; 90) Same, posterior view; 91) Odontophore, radular ribbon removed, ventral view; 92) Same, both cartilages deflected, m3 transversally sectioned, right m4 deflected; 93) Central nervous system, left view, topology of adjacent esophagus also shown; 94) Same, anterior view. Scales = 0.5 mm.

FIGURES 53‑56: Coccodentalium caduum anatomy: 53) Whole ventral-slightly right view, haemocoel opened longitudinally exposing internal organs, stomach partially deflected, topology of some other structures also shown; 54) Detail of foot and anterior region of haemocoel, left view, captacula mostly extracted, except their base; 55) Digestive tubes, left view, foregut opened longitudinally, topology of some adjacent structures also shown, trajectory of intestine indicated in schematic representation at left; 56) Middle region of visceral mass, dorsal view, outer layer of structures, main mantle and longitudinal muscles, removed. Scales = 2 mm.

FIGURES 101‑103: Ennucula puelcha anatomy: 101) Whole ventral-slightly right view, right hemipalps deflected, pallial lobes removed; 102) Midgut and adjacent structures in situ, right view; 103) Stomach and adjacent digestive tubes,right view, mostly opened longitudinally. Scales = 1 mm.

FIGURES 16‑29: Shell and radular aspects of Gadilidae: 16-22) Polyschides noronhensis; 16) Holotype shell, left view, length = 8.6 mm; 17) same, paratype MZUSP 32011, length = 7.8 mm; 18) Paratype MZUSP 32011#1, left view, detail of anterior aperture in profile with some captacula, scale = 0.5 mm; 19-20) Paratype MZUSP 46736, detail of posterior aperture in profile, right and left views respectively, scale = 0.25 mm; 21-22) Radulae of 2 specimens in SEM, scale = 50 µm; 23-29) Gadila braziliensis MZUSP 46736; 23) #3, right view, length = 7.5 mm; 24) #2, dorsal view, length = 7.9 mm; 25) #4, young specimen, length = 3.4 mm; 26) Radula in SEM, scale = 50 µm; 27) Same in light microscopy; 28) Odontophore, radular ribbon removed, ventral view, maximum diameter = 0.7 mm; 29) #2, left view, detail of posterior aperture in profile, scale = 0.5 mm.

FIGURES 57‑64: Coccodentalium caduum anatomy: 57) Odontophore, dorsal view; 58) Same, ventral view; 59) Odontophore, anterior view, superficial layer of muscles and membranes removed, part of radula also removed (except portion inside radular sac); 60) Same, posterior view, radula completely removed, left portion of radular sac shown in situ, right m4 deflected; 61) Same, dorsal view, m3 transversally sectioned, both cartilages deflected; 62) Same, both m4 deflected downwards; 63) Cerebral ganglia, ventral view; 64) Pedal ganglia and statocysts, dorsal view. Scales = 0.5 mm.

FIGURES 1‑15: Shell, radular and anatomical aspects of Dentaliidae: 1-9) Coccodentalium caduum; 1) Shell, whole right view, MZUSP 32977, length = 74.0 mm; 2) Detail of anterior aperture, profile, right view; 3) Detail of sculpture of middle region. 4) Anterior aperture, anterior view, animal still inside; 5) detail of posterior aperture, dorsal view, scale = 2 mm; 6) Radula in situ, anterior view, scale = 1 mm; 7) Radula removed from odontophore, ventral view, scale = 1 mm; 8-9) Radulae of 2 specimens in SEM, scale = 200 µm; 10-15) Paradentalium disparile; 10) Shell, whole left view, MZUSP 25589#1, length = 35.1 mm. 11) Same, detail of posterior aperture; 12) Shell, whole left view, MZUSP 25589#2, length = 31.5 mm; 13) Detail of dissected anterior region, central view, captacula deflected to show foot, scale = 2 mm. 14-15) Radulae of 2 specimens in SEM, scales = 100 µm.