Comparative Morphology of Dorsanum miran and Bullia granulosa from Morocco (Mollusca: Caenogastropoda: Nassariidae)

ABSTRACT The anatomy and taxonomy of two western African nassariids are explored, based on samples collected in Morocco. The species are Dorsanum miran, the type species of the genus, and Bullia granulosa, a characteristic member of Bullia. Both possess the typical morphological and anatomical features of the family, including a pair of metapodial tentacles, a well-developed proboscis, elongated odontophore with fusion of cartilages, and highly concentrated central nervous system. Both species have in common the socket-like heads, bifid columellar muscles, and reduction of the gland of Leiblein. D. miran has well-developed eyes, cement gland, and preputial protection at the penis tip. B. granulosa lacks eyes, has multiplicity of some buccal mass muscles (transverse muscles and main dorsal tensor muscles — m2), and a thick-walled and broad anterior oesophagus. The characters are discussed in the light of present knowledge concerning caenogastropod taxonomy.

this paper is part of a project reviewing the South American species usually included in the region's nassariid genus Buccinanops d'Orbigny, 1841. Since the original description, the species of Buccinanops have been included in several different genera (Pastorino 1993). two of the most used, Dorsanum Gray, 1847 and Bullia Gray, 1834 (in Griffith & Pidgeon 1834), both originally described from Africa, have some taxonomic problems associated with them and they are poorly defined at the anatomical level. The most recent and comprehensive revisions of these genera were published more than two decades ago. Adam and Knudsen (1984) regarded Dorsanum as a junior synonym of Bullia. All mon (1990) considered Dorsanum to be valid, but restricted to the type species only, i.e., D. miran (Bruguière, 1789). In that paper, he introduced the subfamily Bulliinae to include Bullia, with Buccinanops as a subgenus, and Dorsanum. Pastorino (1993) re-established the full generic status of Buccinanops for all South American species, but a more complete discussion on the genera Bullia and Dorsanum is still wanting. Only a few papers (e.g. Simone 1996Simone , 2011 describe the anatomy of soft parts of these taxa and all of those descriptions have been restricted so far to South American species. these three genera were included by Brown (1982) in what he called the "Bullia group". Later, Allmon (1990) extended this informal group, as Bulliinae, to fossil species, mostly from North America. He defined three subfamilies in Nassariidae: Nassariinae, Dorsaninae and Bulliinae. He included in Bulliinae a set of nassariids of relatively large size, having thin-walled shells, and confined to temperate and subtropical waters in the mid and south Atlantic and Indian Oceans. Dorsaninae is restricted to Dorsanum, while the genera Bullia (from Africa) and Buccinanops (from South America) comprise the Bulliinae. Despite these definitions, the differences amongst the subfamilies remain somewhat unclear. It is quite possible that additional information on their representatives, in terms of anatomy, could bring new insights, as the shell characters so far invoked have been insufficient to define the subfamilies and are not applicable to all included species.
As regards this scenario, the present paper deals with two species from the western coast of Africa: Dorsanum miran, type species of Dorsanum, and Bullia granulosa (Lamarck, 1822). Both are formally redescribed, including consideration of previously neglected anatomical aspects. Provision of the new information explored here has the target of laying a better foundation for future discussions and analysis concerning monophyly, and the phylogenetic and taxonomic allocation of the "Bullia group" (Brown 1982) and the Bulliinae/Dorsaninae (Allmon 1990), as well as the included genera Bullia, Dorsanum and Buccinanops. However, the present paper is almost purely descriptive, with discussion limited to data published so far.

MATERIAL AND METHoDS
Most of the material is from the Muséum national d'Histoire naturelle, Paris (MNHN); and some other comparative specimens are housed at the Muséum d'histoire naturelle de la Ville de Genève (MHNG), Academy of Natural Sciences of Drexel University, Philadelphia (ANSP) and the United States National Museum of Natural History, Smithsonian Institution, Washington, DC (USNM). The specimens had been fixed in 70 % ethanol. One female of D. miran was examined after the shell had been broken in order to complete the extraction of soft parts. In the remaining specimens, extraction was carried out by pulling tissue with forceps to obtain only the soft parts from the last whorl. Study of the visceral mass was precluded. the specimens were dissected by standard techniques under a stereo-microscope while immersed in ethanol. All drawings were made with the aid of a camera lucida. Scanning electron microscopy under a Phillips XL 30 was used to study the radulae, with the usual coating, at the Laboratório de Microscopia Eletrônica, Museu de Zoologia da USP, and Museo Argentino de Ciencias Naturales "Bernardino Rivadavia" (MACN). For the main terminology, including that relating to odontophore muscles, see Simone (2011).
Outer lip simple, sharp, rounded (Fig. 2). Inner lip simple, callus narrow, covering right surface of canal. Columella possessing strong spiral folds, located between middle and inferior thirds of each whorl (Fig. 4), ending as right edge of left canal wall, producing small protuberance located between middle and inferior thirds of inner lip (Fig. 5). as wide as dorsal surface of foot; attachment with operculum occupying ~ 70 % of its area ( Fig. 24: oa). Columellar muscle thick, ~1.5 whorl; distal end bifid, with right element wide, left element narrow, encased in furrow formed by columellar fold (Figs 21,24: cm). Female with small orifice of cement gland located on median line of anterior sole region ( Fig. 22: gc); inner space wider than narrow duct ( Fig. 23: gc). Operculum (Figs 1, 6, 7). Elliptical, horny, pale brown. Nucleus terminal, inferior. Outer sur face with normal concentric growth lines, forming undulations. Scar oval, occupying about ⅔ of inner surface, somewhat dislocated closer to inner edge.
Mantle organs . Mantle edge simple, thick. Siphon long, extending beyond mantle edge for 3-4× its base width ( Fig. 28: si). Low, broad fold of siphon's right base separates anterior end of osphradium and gill. Osphradium ~ ¼ width and length of pallial cavity; anterior end pointed, strongly curved towards left; posterior end rounded; osphradium filaments symmetrical, long, bluntly pointed, basal edge rein forced by wide rod (Fig. 26: os). Area between osphradium and gill very narrow. Ctenidial vein (cv) narrow, width uniform along its length. Gill (gi) ~ 80 % of pallial cavity length and ~ ½ its width; anterior end broadly pointed, located at some distance from mantle edge. Afferent gill vessel very narrow, lying at a short distance from right edge of gill ( Fig. 26: af). Gill separated from right edge of pallial cavity by an area equivalent to its width. Hypobranchial gland thin, beige, covering most of area between gill and rectum, including left and ventral surfaces of rectum ( Fig. 26: hg). Rectum narrow, with thick walls, running along right edge of pallial cavity ( Fig. 26: rt). Anus simple, siphoned, located in front of anterior quarter of pallial cavity (Figs 28, 36: an). Pallial gonoducts located between rectum and right pallial edge, described below. Visceral mass (Figs 28,40). Anterior quarter of whorl mostly occupied by kidney (ki) and pericardium (pc). Digestive gland beige, located along inferior region of each visceral whorl, covering middle digestive tubes and also two whorls posterior to stomach. Gonad pale beige, lying along superior and columellar surfaces of visceral whorls posterior to stomach. Stomach small, located half a whorl in front of pallial cavity ( running posteriorly, inserting in lateral side of anterior third of odontophore ( Fig. 29: m1); m1t, pair of narrow, lateral, transverse protractor muscles of odontophore, originating in lateral region of proboscis, running posteriorly and penetrating membrane surrounding odontophore, inserting in odontophore cartilage in region just posterior to m6 (Figs 32, 33); ma, peribuccal muscles and protractor of odontophore bearing longitudinal fibres, origin thin within dorsal wall of oral cavity, running along odontophore tube and becoming thicker, inserting into outer surface of cartilages externally to m6 and medially to m4 ( narrow, originating in ventral surface of haemocoel in region just posterior to proboscis, running dorsally, with median fibres running through nerve ring, inserting into posterior end of cartilages (Figs 29,30,(32)(33)(34)(35); m2a, auxiliary of m2, being single and running bet ween both m2 muscles, inserting on radular nucleus together with small branch of aorta (Figs 32-34); m2b, broad and thick auxiliary pair of m2 and dorsal tensor muscles of radula, originating as m2, running more medially, inserting along radular sac in its mid-region (Figs 32-34); m2c, ventral single, thin muscle auxiliary of m2, originating on ventral medial fibres of m2a, detaching from it in region just posterior to m6, runs covering medial region of m6, inserting on median-posterior region of ma (Fig. 33); m4, strong pair of dorsal tensor radular muscles, originating in odontophore cartilages on posterior-ventral surface, running towards dorsal surrounding lateral surface of cartilages, inserting laterally along radular sac (Figs 33-35); m5, pair of secondary dorsal tensor muscles of radula, originating in posterior-dorsal surface of cartilages, running dorsally and medially, joining with m4, inserting into radular sac alongside and medial to m4 insertion (Figs 33-35); m6, thin horizontal muscle, uniting both odontophore cartilages just posterior to anterior fusion for about 70 % of cartilage length, inserting along ventral and internal edge of cartilages, gradually becoming broader posteriorly (Figs 33, 35); m8, a pair of thick cartilages shorten muscles, which are located along dorsal edge of anterior region of cartilages for almost half their length, being thicker anteriorly and gradually narrowing posteriorly (Figs 34, 35); m9, one or two pairs of very narrow protractor muscles of radula, originating on posterior half of dorsal edge of cartilages, running forwards medially, inserting along m2a anterior fibres (Figs 34,  35); m9a, pair similar to m9, originating in anterior region of ventral edge of cartilages, just centrally to posterior region of m8, running medially and dorsally, inserting together with m2b fibres at level anterior to radular nucleus (Figs 34, 35); m11, pair of ventral tensor muscles of radula, thin, somewhat broad, originating in posterior-ventral end of cartilages, running anteriorly and medially covering m6, inserting into ventral edge of radula and subradular cartilage, and some inner portions in front of this (Figs 32-35). Subradular cartilage expands in exposed region of radula into buccal cavity, covering neighbouring surface of radula; oc, odontophore cartilages, elongate, furrow-like, flat, ~ 20× longer than wide; fusion between both cartilages in anterior-medial end, along ~ 5 % of their length (Fig. 35) their ducts very narrow, except for short proximal region that runs completely attached to anterior oesophagus wall and, more anteriorly, inside dorsal folds of buccal cavity ( Fig. 31: sd); opening is a very small pore ( Fig. 31: sa), into anterior-middle region of dorsal folds of buccal cavity. Anterior oesophagus with somewhat thick walls, a little longer than proboscis; inner surface with 12-15 narrow, uniform longitudinal folds.
Inner lip with small protuberance in base of canal (Fig. 12).
Mantle organs (Figs 45, 46). Similar to those of preceding species. Remarks and differen ces: Low, broad fold of siphon's right base separates anterior end of osphradium from gill and is somewhat taller (Fig. 46: sf). Osphradium (os) with filaments shorter, but with longer lateral projections ( Fig. 45: os). Ctenidial vein (cv) relatively broader, weakly expanded at mid-level. Gill with much shorter filaments, apex approximately at mid-level ( Fig. 45: gi). Anus shortly siphoned, located in front of anterior of pallial cavity ( Fig. 46: an). Visceral mass. Not seen in detail. Circulatory and excretory systems (Fig. 46). Similar to preceding species, except that the dorsal side of the renal lobe is more developed (kl).
Radula similar to that of preceding species (Figs 18-20): rachidian with ~ 21 cusps and thinner; base of the rachidian less curved; lateral tooth also similar, with inner cusp rising from base being more pronounced (Fig. 19). Salivary glands with similar features, Figs 49-53. Bullia granulosa foregut anatomy: (49) buccal mass, right view, odontophore partially removed and deflected to the right, buccal cavity and oesophagus partially sectioned longitudinally, topology of transverse muscles also shown; (50) odontophore, ventral view, superficial layer of membrane and muscles removed and shown partially deflected to the right, radula still in situ; (51) region of mid-oesophagus sectioned longitudinally to show inner surface; (52) odontophore, dorsal view, outer layer of membrane and muscles removed, cartilages also deflected, some muscles on right side with portions artificially sectioned; (53) odontophore cartilages and some adjacent portions of intrinsic muscles, dorsal view. Scale bars = 1 mm. except for their ducts, with clear expansion in region in front of their aperture, situated within dorsal folds of buccal cavity (df), and salivary aperture more laterally positioned ( Fig. 49: sa). Anterior oesophagus much broader, about as broad as odontophore (Figs 48,49: ea). Valve of Leiblein also broader ( Fig. 48: vl), with well-developed inner cilia ( Fig. 51: vc). Middle and posterior oesophagus with similar characters (Figs 48,51: em,ep). Gland of Leiblein very narrow, elongated, filiform (Figs 48,51: gl); ~ 3× longer than middle oesophagus and ~10× narrower than it; twisted between anterior and middle thirds, just in region where aorta passes (Figs 48: aa). Duct of gland of Leiblein almost undetectable, aperture simple ( Fig. 51: ga). Stomach and intestine not examined in detail. Rectum and anus described above (pallial cavity). Genital system. Male. No male was available for examination. Female (Fig. 46). General features similar to those of preceding species; except for relatively shorter albumen gland (ag) and female pore (fp) being situated further away from anus. No cement gland detectable.

DISCUSSION
Although there is no doubt about the validity of both species, the shells of Dorsanum miran and Bullia granulosa are so similar that they are commonly found together in the same lot in collections, as if the shells are those of a single species. Only after more detailed study of both shells and anatomy did the specific and even generic distinctions become apparent. the main shell differences include the apex, which is more sharply pointed in D. miran than in B. granulosa (compare Figs 1-3 with 9-11). the sculpture is also different in that D. miran has some broad nodulation in the first teleoconch whorl (Figs 1-3), whereas B. granulosa has a delicate pair of subsutural nodes extending all along the whorls (Figs 9-11). the sculpture of the region to the left of the aperture is also dissimilar, as D. miran has only a single pair of broad folds (Fig. 5), and B. granulosa has 7-8 delicate ones (Fig. 12). the protuberance in the canal's left base, almost a tooth, is well-developed in D. miran (Fig. 5) but weaker in B. granulosa (Fig. 12); it is at the end of the columellar fold (Fig. 4) in the former, a feature that was not confirmed for the latter because the shell could not be broken. the main anatomical differences and similarities were incorporated into the distinctive description of B. granulosa. the absence of eyes in B. granulosa (Figs 43,44) clearly separates this species from D. miran, where the eyes are very characteristic (Figs 24,25). the fold of the siphonal base separating the anterior end of the osphradium from the gill in B. granulosa (Fig. 46: sf) is taller than in D. miran (Fig. 28). the gill and osphradium filaments are much larger in D. miran (Fig. 26) than in B. granulosa (Fig. 45). the transverse musculature of the haemocoel is much more developed in B. granulosa than in D. miran in the ventral rhynchodeal wall (Fig. 47: mf) and between the oesophagus and odontophore (Fig. 49: tm). As regards odontophore muscles, the multiplicity of m2b that creates the pair m2d (Fig. 52) in B. granulosa is the main distinctive feature, this muscle being much simpler in D. miran. that aside, the pair m9a of B. granulosa are enlarged (Fig. 52). All these features indicate that B. granulosa has stronger odontophore musculature than D. miran. the extent of fusion between both odontophore cartilages is greater in B. granulosa (Fig. 53) than in D. miran (Fig.  35). Moreover, the anterior oesophagus and valve of Leiblein of B. granulosa are proportionally broader in B. granulosa (Fig. 48) than in D. miran (Fig. 29). Both species have in common a somewhat reduced gland of Leiblein. However, that of B. granulosa is elongated and filiform ( Fig. 48: gl) as compared with the "usual" form exhibited by D. miran (Fig. 29). Related to the nerve ring, B. granulosa has the cerebral and buccal commissures longer than those of D. miran (Figs 42,55), but both are very concentrated as in all buccinoid nerve rings (Bailey 1966).
the marked conchological similarity between D. miran and B. granulosa has been pointed out in the literature. Allmon (1990: 28, 29) invoked anatomical studies to resol ve the taxonomy. In fact, several characters allow the specific and even generic distinctions to be made, as reflected above. However, knowledge based on additional species is necessary for a more complete taxonomical evaluation. Despite B. granulosa having been used here as representative of the genus Bullia, it is important to emphasize that the type species, B. callosa, still needs to be studied. Generic inferences can then be made in the light of such an assessment. Allmon (1990) erected the subfamily Bulliinae in his revision of the so called "Bullia group", for which he studied mostly fossil representatives from North America. He mentioned the uncertain generic position of B. granulosa and several other genera. In his diagnosis of the subfamily, he included mainly shell characters, comparing Bulliinae with Dorsaninae and Nassariinae. Whereas those characters are clear when representatives of the Nassariinae are compared, they are not easy to see in Dorsaninae. He mentioned as main shell differences in Bulliinae the lack of a recurved siphonal channel with a carina on the dorsal side of the fasciole, and the reduced ornamentation.
the diagnostic features are the large size for a nassariid, considering that some species attain a shell length of 100 mm, such as Buccinanops cochlidium (Dillwyn, 1817), while the typical nassariid is approximately10 mm. the absence of a well-developed callus is another exclusivity, this structure being particularly robust in most nassariids. the callus and a more developed sculpture of the typical nassariids generate a thicker shell wall. the lack of or reduction in these features produces the thinner shell typical of the Bullia group. In general, the shells of the species so far included in the Bullia group resemble those of the Buccinidae more closely than shells of nassariids.
In regard to anatomy, the tendency for the gland of Leiblein to be reduced and the simplification of the female pallial oviduct are noteworthy attributes in comparison with the situation in other buccinoideans (Fretter 1942;Haasl 2000;Kantor & Harasewych 2008;Kosyan et al. 2009). the nassariid nature of the Bullia group species comes from the socket-like head and the metapodial tentacles, which are paired in both examined species, but single in Buccinanops (Simone 1996). the term "socket-like" for the head was coined by Marcus and Marcus (1962), and means a head that protrudes in a form that resembles an electric socket, in which the tentacles are the pins.
the nerve rings of both species are closely similar (Figs 41,42,54,55), highly concentrat ed in such a way that the pleural and cerebral ganglia cannot be distinguished from each other. this is an accepted feature for a neogastropod (Fretter 1942). However, the cerebral commissure is somewhat long in both species studied. the concentration of ganglia and the closure of the buccal ganglia has been referred to for another allied species, Bullia digitalis (Dillwyn, 1817) (Brown 1982, figs 16A, B), as well as the protuberances on opposite sides of the buccal ganglia (Figs 42, 54), which have been termed the sub and supra-intestinal ganglion, respectively (Brown 1982). they are not so distinct in the presently studied species, however. Conversely, the elongated suboesophageal ganglion close to the nerve ring, present in both species (Figs 41,54: su), is not shown by Brown (1982) for B. digitalis.
An interesting feature is the absence of eyes, which is typical for Buccinanops. this trait is shared with B. granulosa, but not with D. miran, which has well-developed eyes (Fig. 25). on the other hand, the reduced gland of Leiblein of the anatomically known Buccinanops resembles more closely that of D. miran.
the elongated odontophore of both examined species is typical for buccinoideans (Wilsmann 1942;Simone 1996Simone , 2011, including the main features of the muscles and cartilages. the degree of cartilage fusion and the length of the horizontal muscles (m6) are, by contrast, smaller in both examined species (Figs 35, 53). the pair or pairs of pro trac tor muscles of the radula, here called m9, is so far another exclusivity, but a more detailed assessment of these and other odontophore characters will be needed in the future, in comparison with Buccinanops.
B. granulosa has previously been assigned to the genus Dorsanum, more precisely the subgenus Fluviodorsum Boettger, 1885 (Cernohorsky 1984). However, since no charac ter fully agrees with that classification, we have kept the species in the genus Bullia, following most previous authors.
Although the original intention was to improve upon the definition of nassariid subfamilies, as set out by Allmon (1990), by providing anatomical details, is it still unclear whether the differences outlined above have validity as distinguishing characters at the specific, generic or subfamilial level. Discussion in this regard is accordingly post poned to a time when the anatomy of more species has become better known. A phylogenetic analysis will then be feasible; and this will be the next step in this ongoing investigation.