Belg. J. Zool., 133 (1) : 77-84 January 2003

On the osteology and myology of the cephalic region and pectoral girdle ofLiobagrus reini Hilgendorf, 1878, with a discussion on the phylogenetic relationships of the (Teleostei: Siluriformes)

Rui Diogo, Michel Chardon and Pierre Vandewalle

Laboratory of Functional and Evolutionary Morphology, Bat. B6, University of Liège, B-4000 Sart-Tilman (Liège), Belgium

ABSTRACT. The cephalic and pectoral girdle structures of Liobagrus reini are described in detail and compared with those of Amblyceps mangois, as well as of several other , as the foundation for a discussion on the phylogenetic relationships of the Amblycipitidae. Our observations and comparisons support de Pinna’s (1996) phylogenetic hypothesis, according to which the Amblycipitidae, plus the Akysidae, , and the , form a monophyletic , the superfamily . In addition, our observations and compari­ sons pointed out a potentially new amblycipitid synapomorphy, namely: the hypobranchial foramen of the paruro- hyal is markedly enlarged, with the length of the foramen being superior to half of the length of the main body of the parurohyal.

KEY WORDS : Akysidae, Amblycipitidae, , cephalic region, comparative morphology, Liobagrus, myology, pectoral girdle, phylogeny, Siluriformes.

INTRODUCTION lies forming the superfamily Sisoroidea, a hypothesis pos­ teriorly reiterated by d e P in n a (1998). The Siluriformes is “one of the economically important It is worthy of note that, despite the large number of studies groups of fresh and brackish water fishes in the world: in concerning silurifonn morphology (e.g., R eg a n , 1911; G a u - many countries, member form a significant part of b a , 1962,1966,1968,1 9 6 9 ;M aha jan , 1963,1966ab, 1967ab; inland fisheries; several species have been introduced into Alex a n d er , 1965;C h ard o n , 1968;G o sline , 1975;G h iot , fish culture; numerous species are of interest to the aquar­ 1978;G hiot et al., 1984;A rra tia , 1987,1990,1992;A rra tia ium industry where they represent a substantial portion of & Sch ultze , 1990; Mo, 1991;H e, 1996;D io go et al., 1999, the world trade” (T e u g e l s , 1996). The Amblycipitidae, 2000,2001 ab; D io go & C h a r d o n , 2000abc; etc.) the only few, with only three genera (Liobagrus Hilgendorf, 1878Am­ somewhat detailed morphological descriptions of the ambly­ blyceps Blyth, 1858 andXiurenbagrus Chen & Lundberg, cipitid catfishes published so far are those of T ila k (1967), 1994 ) and 15 species is one of the smallest families of cat­ C h ard o n (1968 ),C h en & L u nd berg (1994) andde P inna fishes (C h e n & L u n d b e r g , 1994). The phylogeny and sys- (1996). Moreover, as these descriptions are almost exclusively tematics of the Amblycipitidae were recently revised by restricted to the osteology and external anatomy of the ambly- C h e n & L u n d b e r g (1994). According to these authors, the cipitids, some important aspects of the morphology of these Xiurenbagrus is the sister-group of a clade formed fishes, such as, for example, the configuration of both the mus­ by the genus Liobagrus and the genus Amblyceps, with cles and the ligaments of the cephalic region and the pectoral these three genera being, in turn, the sister-group of a clade girdle, or the configuration of the structures associated with containing the two akysid genera, Ak\>sis and Parakysis their mandibular barbels, are practically unknown. This not (although C h e n & L u n d b e r g referred to a monogeneric only complicates the study of the functional morphology of the family Akysidae and to a monogeneric family Parakysi­ amblycipitids, but also restricts considerably the data available dae, the genera Ak\>sis and Parakysis are nowadays includ­ to infer the synapomorphies and the phylogenetic relationships ed in a single family, the Akysidae: see F e r r a r is & d e of these catfishes ( see d e P in na , 1996 ). P in n a , 1999 ). Such a sister-group relationship between the Amblycipitidae and the Akysidae was, however, contra­ The aim of this work is to describe in detail the bones, dicted in a later study by d e P in n a (1996). In fact, accord­ cartilages, muscles and ligaments of the cephalic region ing to d e P in n a (1996), the Amblycipitidae is not the (branchial apparatus excluded) and pectoral girdle of the sister-group of the Akysidae, but instead the sister-group amblycipitid Liobagrus reini Hilgendorf, 1878, and to of a clade formed by this latter family, the Sisoridae, the compare these structures with those of a representative of Aspredinidae and the Erethistidae, with all the five faini- one of the two other amblycipitid genera, namely Amblyc­ eps mangois ( Hamilton, 1822 ), as well as of several other catfishes, as the foundation for a discussion on the phylo­ Corresponding author : R. Diogo, e-mail : [email protected] genetic relationships of the Amblycipitidae. 78 Rui Diogo, Michel Chardon and Pierre Vandewalle

MATERIAL AND METHODS tus (Pimelodidae): USNM 226140, 1 (ale). Ictalurus punctatus (Ictaluridae): USNM 244950, 2 (ale).Laides hexanema (Shilbi­ dae): USNM 316734, 1 (ale). Leptoglanis rotundiceps (Amphili­ The fishes studied are from the collection of our labora­ idae): MRAC P. 186591 -93, 3 (ale). Liobagrus reini tory (LFEM), from the Musée Royal de l’Afrique Centrale (Amblycipitidae): USNM 089370, 2 (ale). Loricaria cataphra­ of Tervuren (MRAC), from the Université Nationale du cta (Loricariidae): LFEM, 1 (ale). Microglanis cottoides (Pime­ Bénin (UNB), from the Muséum National D’Histoire Na­ lodidae): USNM 285838, 1 (ale). Mochokus niloticus turelle of Paris (MNHN), from the University of Gent (): MRAC P.119413, 1 (ale); MRAC P.119415, 1 ( UG) and from the National Museum of Natural History of (ale). Mystus gulio (Bagridae): LFEM, 1 (ale). Nematogenys iti- Washington (USNM). Anatomical descriptions are made ermis (Nematogenyidae): USNM 084346, 1 (ale).Notoglanid­ after dissection of alcohol fixed or trypsin-cleared and ali­ ium thomasi (Claroteidae): LFEM, 2 (ale). Parakysis anomalopteryx (Akysidae): USNM 230307, 2 (ale). Paramphil­ zarine-stained (following T a y l o r & V a n D y k e ’s 1985 ius trichomycteroides (Amphiliidae): LFEM, 2 (ale). Pangasius method) specimens. Dissections and morphological draw­ larnaudii (Pangasiidae): USNM 288673,1 (Ac). Pangasius sian- ings were made using a Wild M5 dissecting microscope ensis (Pangasiidae): USNM 316837, 2 (ale). Paraplotosus albi­ equipped with a camera lucida. The trypsine-cleared and labris (Plotosidae): USNM 173554, 2 (ale). Phractura alizarine-stained (c&s) or alcohol fixed (ale) condition of brevicauda (Amphiliidae): MRAC 90-057-P-5145, 2 (ale); the studied fishes is given in parentheses following the MRAC 92-125-P-386, 1 (c&s). Phractura intermedia (Amphili­ number of specimens dissected. A list of the specimens idae): MRAC 73-016-P-5888, 1 (ale).Pimelodus clarias (Pime­ dissected is given below. lodidae): LFEM, 2 (ale); USNM 076925, 1 (ale). Plotosus lineatus (Plotosidae): USNM 200226), 2 (ale). Pseudeutropius Acanthodoras cataphractus (): USNM 034433, 1 brachypopterus (Shilbidae): USNM 230301, 1 (ale). Pseu­ (ale). Ageneiosus vittatus (Auchenipteridae): USNM 257562, 1 domystus bicolor (Bagridae): LFEM, 1 (ale), LFEM, 1 (c&s). (ale). Ailia colia (Shilbidae): USNM 165080, 1 (Ac). Aky’sis leu­ Pseudopimelodus raninus (Pimelodidae): USNM 226136, 2 corhynchus (Akysidae): USNM 109636, 2 (ale). Amblyceps (ale). Pseudoplatystoma fasciatum (Pimelodidae): USNM mangois (Amblycipitidae): USNM 109634, 2 (ale). Amiurtis 284814, 1 (ale).Rhamdia guatemalensis (Pimelodidae): USNM nebolosus (Ictaluridae): USNM 246143, 1 (ale); USNM 73712, 114494, 1 (ale).Rita chrysea (Bagridae): USNM 114948, 1 (ale). 1 (Ac). Amphilius brevis (Amphiliidae): MRAC 89-043-P-403, 3 Schilbe intermedius (Shilbidae): MRAC P.58661, 1 (ale). Silu­ (ale); MRAC 89-043-P-2333, 1 (c&s). Anadoras weddelii (Do­ ranodon auritus (Shilbidae): USNM 061302, 2 (ale). Silurus radidae): USNM 317965, 1 (Ac). Andersonia leptura (Amphilii­ asotus (Siluridae): USNM 130504, 2 (ale). Synodontis clarias dae): MNHN 1961-0600, 2 (ale); Arius hertzbergii (Ariidae): (Mochokidae): USNM 229790, 1 (ale). Tandanus rendahli (Plo­ LFEM, 1 (Ac). Arius heudelotii (Ariidae): LFEM, 4 (Ac). Aspre­ tosidae): USNM 173554, 2 (ale). Trachyglanis ineac (Amphilii­ do (Aspredinidae): USNM 226072, 1 (ale). Aucheni­ dae): MRAC P.125552-125553,2 (ale). Uegitglanis zammaronoi pterus dentatus (Auchenipteridae): USNM 339222, 1 (ale). (Clariidae): MRAC P-15361, 1 (ale). Wallago attu (Siluridae): Auchenoglanis biscutatus (Claroteidae): MRAC 73-015-P-999, USNM 304884, 1 (ale). magdalenae (Aspredinidae): 2 (ale). Bagarius yarreli (Sisoridae): USNM 348830, 2 (ale); USNM 120224, 1 (ale). Zaireichthys zonatus (Amphiliidae): LFEM, 1 (c&s). Bagre marinus (Ariidae): LFEM, 1 (ale); MRAC 89-043-P-2243-2245, 3 (ale)." LFEM, 1 (c&s). Bagrichthys macropterus (Bagridae): USNM 230275, 1 (ale). Bagrus boyad (Bagridae): LFEM, 1 (ale); LFEM, 1 (c&s). Bagrus docmak (Bagridae): MRAC 86-07-P- RESULTS 512, 1 (ale); MRAC 86-07-P-516, 1 (c&s).Belonoglanis tenuis (Amphiliidae): MRAC P.60494, 2 (ale). knerii In the anatomical descriptions, the nomenclature for the (Aspredinidae): USNM 177206, 2 (ale). Calophysus macropter­ osteological structures of the cephalic region follows basi­ us (Pimelodidae): USNM 306962, 1 (ale). Centromochlus heche­ cally that of Ar r a t ia ( 1997 ). However, for the several rea­ ln (Auchenipteridae): USNM 261397, 1 (ale). Cetopsis sons explained in detail in our recent papers (D io g o et al., caecutiens (Cetopsidae): USNM 265628, 2 (ale). Chrysichthys auratus (Claroteidae): UNB, 2 (ale); UNB, 2 (c&s). Chrysich­ 2001a; D io g o & C h a r d o n , in press), with respect to the thys nigrodigitatus (Claroteidae): UNB, 2 (ale); UNB, 2 (c&s). skeletal components of the Suspensorium we follow D io ­ Clariallabes melas (Clariidae): LFEM, 2 (ale). Clarias gariepi- g o et al. (2001a). The myological nomenclature is based nus (Clariidae): MRAC 93-152-P-1356, 1 (ale), LFEM, 2 (ale). mainly on W interbottom ( 1974 ), but for the different ad­ Conta conta (Erethistidae): LFEM, 2 (ale). Cranoglanis ductor mandibulae sections, D io g o & C h a r d o n (2000a) is bouderius (Cranoglanididae): LFEM, 2 (ale). Diplomystes followed. In relation to the muscles associated with the chilensis (Diplomystidae): LFEM, 2 (Ac). Doras punctatus (Do­ mandibular barbels, which were not studied by W in t e r - radidae): USNM 284575, 1 (ale).Doumea typica (Amphiliidae): b o t t o m (1974),D io g o & C h a r d o n (2000b) is followed. MRAC 93-041-P-1335, 1 (ale). Erethistes pusillus (Erethisti­ Concerning the nomenclature of the pectoral girdle bones dae): USNM 044759, 2 (ale).Franciscodoras marmaratus (Do­ and muscles, D io g o et al. (2001b) is followed. radidae): USNM 196712, 2 (ale). Gagata cenia (Sisoridae): USNM 109610, 2 (ale). Genidens genidens (Ariidae): LFEM, 2 (ale). Glyptosternon reticulatum (Sisoridae): USNM 165114, 1 Liobagrus reini (ale). Glyptothorax fukiensis (Sisoridae): USNM 087613,2 (ale). Goeldiella eques (Pimelodidae): USNM 066180, 1 (ale). Hara Osteology filamentosa (Erethistidae): USNM 288437, 1 (ale). Helicopha­ gus leptorhynchus (Pangasiidae): USNM 355238, 1 (ale). Helo­ Os mesethmoideum. Situated on the antero-dorsal sur­ genes marmaratus (Cetopsidae): USNM 264030, 1 (ale). face of the neurocranium ( Figs 1,2 ), with each of its ante- Hemibagrus nemurus (Bagridae): USNM 317590, 1 (ale);Hemi- ro-ventro-lateral margins ligamentously connected to the cetopsis candim (Cetopsidae): USNM 167854, 1 (ale). premaxillary. Hepapterus mustelinus (Pimelodidae): USNM 287058, 2 (ale). Heterobranchus longifilis (Clariidae): LFEM, 2 (ale). Heterop­ Os lateroethmoideum. The lateral-ethmoid (Figs 1,2,3 ) neustes fossilis (Heteropneustidae): USNM 343564, 2 (ale); presents a well-developed, laterally-directed articulatory USNM 274063, 1 (ale); LFEM, 2 (ale). Hypophthalmus edenta­ facet for the autopalatine. Osteology and myology of Liobagrus reini 79

m -Al-ost 1-ap

m-ep o-meth o-post-scl o-meth

m-hyp

o-prmx o-leth

o-ang-art o-mx afo pec-sp »-q-sym m-pr-pec o-iop o-fr 2 mm o-sph Fig. 1 (fig. above). - Lateral view of the cephalic musculature of Liobagrus reini. All the muscles are exposed. m-Al-ost, m-A2, sections of musculus adductor mandibulae; m-ep, musculus epaxialis; m-ex-t-1, section of musculus extensor tentaculi; m-hyp, musculus hy- poaxialis; m-l-ap, musculus levator arcus palatini; m-pr-pec, musculus protractor pectora­ pfo lis; m-re-t, musculus retractor tentaculi; o-ang-art, os angulo-articulare; o-apal, os autopalatinum; o-cl, os cleithrum; o-den, os dentale; o-fr, os frontale; o-iop, os interoper- o-pa-soc culare; o-leth, os latero-ethmoideum; o-meth, os mesethmoideum; o-mx, os maxillare; o- op, os operculare; o-pop, os praeoperculare; o-post-scl, os posttemporo-supracleithrum; o- o-pt prmx, os praemaxillare; o-q-sym, os quadrato-symplecticum; pec-sp, pectoral spine. Fig. 2 (fig. to the right). - Dorsal view of the neurocranium of Liobagrus reini, afo, anterior fontanel, o^/r, os frontale; o-leth, os latero-ethmoideum; o-meth, os mesethmoideum; o-pa- o-post-scl soc, os parieto-supraoccipital; o-post-scl, os posttemporo-supracleithrum; o-pt, os pteroti- cum; o-sph, os sphenoticum; pfo, posterior fontanel.

5 mm

Os praevomerale. Well-developed (Fig. 3), T-shaped Os basioccipitale. Well-developed, unpaired bone, bone without a ventral tooth-plate. forming the posterior-most part of the floor of the neuroc­ Os orbitosphenoideum. Posterior to the lateral ethmoid, ranium. Its ventro-lateral surfaces are ligamentously con­ with the dorsal edge of its lateral wall being sutured with nected to the ventro-medial limbs of the posttemporo- the ventral surface of the frontal. supracleithra. Os pterosphenoideum. Posterior to the orbitosphenoid, Os parieto-supraoccipitale. Large bone constituting the covering, together with this bone, the gap between the postero-dorso-median surface of the cranial roof, which frontals and the parasphenoid. bears a small, triangular postero-median process (Fig. 2). Os parasphenoideum. The parasphenoid (Fig. 3) is the Os angulo-articulare. This bone (Fig. 1), plus the den- longest bone of the cranium. It bears a pair of ascending tary, coronomeckelian and Meckel’s cartilage, constitute flanges, which suture with the pterosphenoids and proot- the mandible. Postero-dorsally, the ángulo-articular has an ics. articulatory facet for the quadrate-symplectic. Postero- ventrally, it is ligamentously connected to both the intero- Os frontale. The frontals (Figs 1, 2) are large bones that percular (Figs 1, 4) and the posterior ceratohyal (Fig. 4). constitute a great part of the cranial roof. They are largely separated anteriorly and posteriorly by the anterior and the Os dentale. The toothed dentary (Fig. 1) recovers the posterior fontanels, respectively (Fig. 2). great majority of the lateral surface of the mandible. The postero-dorsal margin of the dentary forms, together with Os sphenoticum. Slightly smaller than the pterotic (Fig. the antero-dorsal margin of the angulo-articular, a some­ 2), constituting, together with this bone, an articulatory what developed dorsal process (processus coronoideus). facet for the hyomandibulo-metapterygoid. Os coronomeckelium. Small bone lodged in the medial Os pteroticum. Well-developed, irregularly-shaped surface of the mandible. Postero-dorsally it bears a crest bone situated posteriorly to the sphenotic (Fig. 2). for attachment of the adductor mandibulae. Os prooticum. Together with the pterosphenoid and the Os praemaxillare. The two premaxillaries form a U- parasphenoid, it borders the well-developed foramen of shaped structure with its lateral tips curved posteriorly and the trigemino-facial nerve complex. ligamentously connected, by means of a strong, long liga­ Os epioccipitale. Situated on the posterior surface of the ment, to the antero-lateral tip of the ento-ectopterygoid neurocranium. The extrascapulars are missing. (Fig. 3). Ventrally, the premaxillaries bear a well-devel­ Os exoccipitale. The well-developed exoccipitals are oped tooth-plate with numerous small teeth having their situated laterally to the basioccipital. tips slightly turned backward (Fig. 3). 80 Rui Diogo, Michel Chardon and Pierre Vandewalle

Os entopterygoide-ectopterygoideum. Poorly-devel­ oped, with its posterior surface being sutured with both the hyomandibulo-metapterygoid and the quadrate-symplec- tic (Fig. 3). Antero-laterally, it presents a well-defined thickening for ligamentous connection with the sesamoid bone 1 of the Suspensorium. Os quadrato-symplecticum. The quadrate-symplectic (Figs. 1, 3) presents a well-developed anterior articulatory surface to articulate with the postero-dorsal surface of the angulo-articular. Os praeoperculare. Long and thin bone (Figs 1, 3) firm­ ly sutured to both the hyomandibulo-metapterygoid and the quadrate-symplectic. Os operculare. The opercular (Figs 1, 3) is a well-devel­ oped, roughly triangular bone attached ventrally, by means of connective tissue, to the interopercular. It presents a Fig. 3. - Ventral view of the neurocranium of Liobagrus reini. The well-developed, dorso-ventrally elongated antero-dorsal Suspensorium, palatine-maxillary system and the muscles and lig­ articulatory surface for the hyomandibulo-metapterygoid. aments associated with these structures are also illustrated, m-ad- ap, musculus adductor arcus palatini; m-ex-t-3, section of muscu­ Os interoperculare. Its anterior surface is ligamentously lus extensor tentaculi; o-apal, os autopalatinum; o-ent, os ento-ec- connected to the postero-ventral margin of the mandible topterygoideum; o-hm-mp, os hyomandibulo-metapterygoideum; (Figs 1, 4). Medially, the interopercular articulates with o-leth, os lateroethmoideum; o-mx, os maxillare; o-para, os paras- the lateral surface of the posterior ceratohyal. phenoideum; o-pop, os praeoperculare; o-prmx, os praeomaxil- Os ceratohyale posterior. Well-developed, somewhat lare; o-pvm, os praevomerale; o-q-sym, os quadrato- symplecticum; o-ses-1, sesamoidal bone 1 of the Suspensorium. triangular bone connected, by means of two strong liga­ ments, to the postero-ventral edge of the mandible (Fig. 4) and to the medial surface of the Suspensorium (the interhy- Os maxillare. The well-developed, elongated maxillary al is missing), respectively. is connected to the premaxillary by means of a strong, Os ceratohyale anterior. Elongated bone that supports, short ligament (Figs 1, 3). As in most catfishes, the maxil­ together with the posterior ceratohyal, the branchiostegal lary barbels are supported by the maxillaries. rays. Os autopalatinum. The autopalatine (Figs 1, 3) is a rod­ Os hypohyale ventrale. The ventral hypohyals are liga­ like, antero-posteriorly elongated bone with its posterior mentously connected to the antero-lateral edges of the portion somewhat expanded dorso-ventrally. Its posterior parurohyal (Fig. 4). end is capped by a small cartilage (Fig. 3). Its anterior end Os parurohyale. The pamrohyal is an irregular bone pre­ is tipped by a well-developed cartilage with two antero­ senting two well-developed postero-lateral arms and a poor­ lateral concavities, which accept the two proximal heads ly-developed postero-median process (Fig. 4). Its of the maxillary (Fig. 3). Medially, the autopalatine artic­ hypobranchial foramen (see Ar r a tia & S c h u l tze , 1990) is ulates, by means of a small, circular articulatory surface, markedly enlarged, with the length of this foramen being su­ with the lateral ethmoid (Fig. 32). perior to half of the length of the main body of the pamrohyal. Os hyomandibulo-metapterygoideum. The homology, and, Os posttemporo-supracleithrum. This bone (Figs 1, 2), thus, the correct denomination, of this bone, as well as of the plus the cleithmm and the scapulo-coracoid, constitute the other Suspensorium elements of catfish, has been the subject pectoral girdle. Its dorso-medial limb is loosely attached to of endless controversies (M cM u r r ic h , 1884;D e B eer, 1937; the neurocranium and its ventro-medial limb is ligamen­ H o e d em a n , 1960ab; G o sl in e , 1975;A r r a tia et al., 1978; tously connected to the basiocccipital. Its postero-lateral Arr a t ia & M en u m a r q u e , 1981; 1984;H o w e s , 1983ab; margin is deeply forked, forming an articulating groove 1985;A r r a tia , 1987; 1990; 1992;H o w es & T eu g e l s , 1989; for the upper edge of the cleithmm (Fig. 1). D io g o et al., 2001a; D io g o & C h a r d o n , in press; etc.). As Os cleithrum. The cleithmm (Figs 1, 4) is a large, stout, mentioned before, for the several reasons explained in detail well-ossified stmcture forming a great part of the pectoral in our recent papers (D io g o et al., 2001 a; D io g o & C h a r d o n , girdle and the posterior boundary of the branchial chamber. in press), the nomenclature used here to describe these ele­ It bears a deep crescentic, medially faced groove that ac­ ments follows strictly that presented by D io g o et al. (2001a). commodates the dorsal condyle of the well-developed pec­ The hyomandibulo-metapterygoid (Fig. 3) is a large bone ar­ toral spine. The two cleithra are attached in the antero- ticulating dorsally with both the pterotic and the sphenotic medial line via connective tissue. Antero-ventro-mesially and posteriorly with the opercular. each cleithmm bears a somewhat well-developed, antero- Os sesamoideum 1.Well-developed, 7-shaped bone at­ mesially pointed, anterior projection, with the stmcture tached, by means of two thick ligaments, to the ento-ec- formed by the association of the antero-mesial projections topterygoid and to the prevomer, respectively (Fig. 3). Its of both cleithra having, thus, a triangular shape (see Fig. 4). antero-dorso-lateral surface is weakly attached, via con­ The poorly-developed humeral process of the cleithmm is nective tissue, to the postero-ventral surface of the au­ surrounded posteriorly by soft tissue, which, in turn, is asso­ topalatinum. The sesamoid bones 2 and 3 (see (D io g o et ciated posteriorly with a long, thick ligament attached to the al., 2001a) are absent. anterior surface of the parapophysis of the sixth vertebra. Osteology and myology of Liobagrus reini 81

Musculus levator op erculi. It originates on the lateral margin of the pterotic and inserts on the dorsal surface of the opercular. Musculus adductor operculi. Situated medially to the levator operculi, it originates on the ventral surface of the pterotic and inserts on the dorso-medial surface of the opercular. Musculus adductor hyomandibularis. Small muscle sit­ uated mesially to the levator operculi but laterally to the adductor operculi. It originates on the ventral surface of the pterotic and inserts on the postero-dorso-median sur­ face of the hyomandibulo-metapterygoid. Musculus dilatator operculi. Well-developed, originat­ ing on the pterosphenoid, frontal, sphenotic and also on the dorso-lateral surface of the hyomandibula and insert­ ing on the antero-dorsal margin of the opercular. Musculus extensor tentaculi. This muscle is divided into three bundles. The extensor tentaculi 1 (Fig. 1) mns from both the orbitosphenoid and the lateral ethmoid to the postero-dor- sal surface of the autopalatine. The extensor tentaculi 2 orig­ inates on the lateral ethmoid and inserts on the postero-medial surface of the autopalatine. The extensor tentaculi 3 (Fig. 3) Fig. 4. - Ventral view of the cephalic region and pectoral girdle of Liobagrus reini. The muscles associated with the pectoral gir­ mns from the lateral ethmoid and the orbitosphenoid to the dle are also illustrated, m-ab-sup-1, section 1 of musculus abduc­ postero-ventral margin of the autopalatine. tor superficialis; m-arr-d-vd, ventral division of musculus Musculus retractor tentaculi. Well-developed muscle arrector dorsalis; m-arr-v, musculus arrector ventralis; m-hyp, situated medially to the adductor mandibulae (Fig. 1). It musculus hypoaxialis; mnd, mandible; m-sh, musculus stemohy- originates on the lateral surface of the hyomandibulo- oideus; o-ch-a, os ceratohyale anterior; o-ch-p, os ceratohyale metapterygoid and inserts, by means of a thick tendon, on posterior; o-cl, os cleithrum; o-hh-v, os hypohyale ventrale; o- the maxillary. iop, os interoperculare; o-op, os operculare; o-puh, os parurohy- ale; pec-ra, pectoral rays; pec-sp, pectoral spine. Musculus protractor hyoidei. This muscle (Fig. 5) has three parts. The pars ventralis, in which are lodged the car­ Ox scapulo-coracoideum. Elongated, irregular bony tilages associated with the internal and external mandibular plate suturing with the cleithrum along its antero-lateral barbels, originates on both the anterior and posterior cerato- edge. Medially it joins its counterpart in an interdigitation hyals and inserts on the dentary, meeting its counterpart in a of several strong serrations. There is a well-developed well-developed median aponeurosis (Fig. 5). The pars later­ mesocoracoid arch. alis mns from the posterior ceratohyal to the ventro-medial face of the dentary (Fig. 5). The pars dorsalis mns from both the anterior and the posterior ceratohyals to the dentary. Myology

Musculus adductor mandibulae. The adductor man­ m -intm dibulae Al-ost (see D io g o & C h a r d o n , 2000a) originates m -pr-h-v on the pterotic, preopercular, hyomandibulo-metaptery­ in-m nd-b goid and quadrate-symplectic, and inserts on the dentary (Fig. 1). The A2 (Fig. 1), which lies dorso-mesially to the c-in-m nd-b Al-ost, attaches postero-dorsally on the parieto-supraoc- ex-m nd-b cipital, pterotic, sphenotic and frontal antero-ventrally on the dorso-medial surface of the dentary. The adductor c-ex-m nd-b mandibulae A3’ originates on the hyomandibulo-metap­ terygoid and the preoprecular, and inserts tendinously on the coronomeckelian bone. There is no A3” nor Aco. m -hh-inf Musculus levator arcus palatini. The levator arcus pal­ m nd atini (Fig. 1) is a well-developed muscle situated medially to the adductor mandibulae A3 ’. It originates on the sphe­ notic, frontal and lateral-ethmoid and inserts on the lateral 5 m m face of the hyomandibulo-metapterygoid. Fig. 5. - Ventral view of the cephalic musculature of Liobagrus rei­ Musculus adductor arcus palatini. This muscle (Fig. 3) ni. c-in-mnd-b, cartilago intemus mandibularis tentaculi; c-ex-mnd- b, cartilago externus mandibularis tentaculi; ex-mnd-b, in-mnd-b; runs from the parasphenoid, pterosphenoid, orbitosphe- external and internal mandibular barbels; m-hh-inf, musculus hyo- noid and lateral ethmoid to the hyomandibulo-metaptery­ hyoideus inferior; m-intm, musculus intermandibularis; mnd, man­ goid, the ento-ectopterygoid and the sesamoid bone 1 of dible; m-pr-h-l, m-pr-h-v, pars lateralis and ventralis of musculus the Suspensorium. protactor hyoideus. 82 Rui Diogo, Michel Chardon and Pierre Vandewalle

Musculus retractor externi mandibularis tentaculi. diais to the antero-ventral margin of the dorsal part of the Small muscle situated dorsally to the pars ventralis of the pectoral fin rays. protractor hyodei and running from the dentary to the car­ Musculus protractor pectoralis. Well-developed muscle tilage associated with the outer mandibular barbel, which ( Fig. 1 ) running from the ventral surfaces of both the pter­ is connected with the cartilage associated with the internal otic and the posttemporo-supracleithrum to the antero- mandibular barbel and is markedly bifurcated posteriorly dorsal surface of the cleithmm. (Fig. 5). Musculus retractor intend mandibularis tentaculi. Amblyceps mangois Small muscle dorsal to the pars ventralis of the protractor hyodei. It attaches anteriorly to the dentary and posteriorly The principal differences between the structures of the to the cartilage associated with the internal mandibular cephalic region and pectoral girdle of this species and barbel. those of Liobagrus reini are that in Amblyceps mangois: 1 ) Musculus intermandibularis. Small muscle joining the the cartilage associated with the inner mandibular barbel is two mandibles (Fig. 5). pierced by a well-developed, triangular foramen; 2) the ar­ ticulatory surface of the opecular for the hyomandibulo- Musculus hyohyoideus inferior. Thick muscle (Fig. 5) metapterygoid is not elongated dorso-ventrally; 3) the in- attaching medially on a median aponeurosis and laterally terhyal is present; 4 ) the maxillary bone is not elongated on the ventral surfaces of the ventral hypohyal, the anterior proximo-distally; 5) the posterior fontanel of the neurocra­ ceratohyal and the posterior ceratohyal. nium is completely surrounded by the parieto-supraoccip- Musculus hyohyoideus abductor. It runs from the first ital, and not by both this bone and the frontal; 6) the (medial) branchiostegal ray to a median aponeurosis, muscle abductor profundus is significantly more devel­ which is associated with two long, strong tendons, at­ oped than in Liobagrus reini, extending mesially to reach tached, respectively, to the two ventral hypohyals. the postero-mesial surface of the scapulo-coracoid. Musculus hyohyoideus adductor. Each hyohyoideus ad­ ductor connects the branchiostegal rays of the respective DISCUSSION side.

Musculus sternohyoideus. It runs from the posterior Our observations and comparisons support d e P in n a ’s portion of the pamrohyal to the anterior portion of the clei­ ( 1996 ) phylogenetic hypothesis in that the Amblycipitidae thrum (Fig. 5). plus the Akysidae, Sisoridae, Erethistidae and Aspredini­ Musculus arrector ventralis. It mns from the cleithmm dae form a monophyletic clade, the superfamily Sisoroi­ to the ventral condyle of the pectoral spine ( Fig. 5). dea sensu D e P in n a (1996). In fact, our study not only confirmed the synapomorphies given by D e P in n a to sup­ Musculus arrector dorsalis. This muscle, dorsal to the port the monophyly of the Sisoroidea (see D e P in n a , 1996: hypoaxialis, the arrector ventralis and the abductor super­ 59-60), but also pointed out an additional synapomorphy ficialis, is differentiated into two well-developed divi­ to support this superfamily: Presence o f a well-developed, sions. The ventral division (Fig. 5), situated on the ventral antero-mesially-pointed projection on the antero-ventro- surface of the pectoral girdle, originates on the ventral mesial surface of the cleithrum (Fig. 4). margin of both the cleithmm and the scapulo-coracoid and inserts on the antero-lateral edge of the pectoral spine. The Plesiomorphically in catfishes the cleithmm lacks major dorsal division, situated on the dorsal surface of the pecto­ projections or processes on its antero-ventro-mesial sur­ ral girdle, originates on the dorso-medial edge of the face. However, in the catfishes of the superfamily Sisoroi­ scapulo-coracoid and inserts on the anterior edge of the dea, with the exception of the Aspredinidae, each dorsal condyle of the pectoral spine. cleithmm bears antero-ventro-mesially a well-developed, antero-mesially pointed, anterior projection, with the Musculus abductor pro fundus. It originates on the pos­ stmcture formed by the association of the antero-mesial terior surface of the scapulo-coracoid and inserts on the projections of both cleithra having, thus, a triangular shape medial surface of the dorsal condyle of the pectoral spine. (see Fig. 4). The presence of such a projection clearly ap­ Musculus abductor superficialis. This muscle is differ­ pears, thus, to represent a Sisoroidea synapomorphy sec­ entiated into two sections. The larger section (Fig. 5: m- ondarily reversed in aspredinids, since this is by far the ab-sup-1 ) mns from the ventral margins of both the clei­ most parsimonious hypothesis supporting the phylogenet­ thmm and the scapulo-coracoid to the antero-ventral mar­ ic position of the Aspredinidae within the Sisoroidea ( see gin of the ventral part of the pectoral fin rays. The smaller d e P in n a , 1996; D io g o et al., 2001c, 2002). section, situated dorsally to the larger one, mns from the According to D e P in n a (1996), the Amblycipitidae oc­ lateral edge of the scapulo-coracoid to the antero-dorsal cupy the most plesiomorphic position within the super­ margin of the ventral part of the pectoral fin rays. family Sisoroidea, with the intra-relationships among Musculus adductor superficialis. This muscle is situated members of this superfamily being: ( Amblycipitidae + on the posterior margin of the pectoral girdle and is divid­ (Akysidae + (Sisoridae + (Aspredinidae + Erethistidae ) ) ) ). ed into two sections. The larger section originates on the However, as noted by D e P in n a (1996: 76) himself, con- posterior surfaces of both the cleithmm and the scapulo- trarily to the grouping of the Sisoridae, Aspredinidae and coracoid and inserts on the antero-dorsal margin of the Erethistidae, which is very well supported ( and was subse­ dorsal part of the pectoral fin rays. The smaller section quently also confirmed by D io g o et al., 2001c, 2002), the mns from both the postero-ventro-lateral edge of the proposal of a sister-group relationship between the Akysi­ scapulo-coracoid and the dorsal surface of the proximal ra­ dae and these three families is based on scarce evidence. Osteology and myology of Liobagrus reini 83

In fact, this proposal relied “on three synapomorphies, one Sisoridae. It is also important to refer to three other characters of these ( supratemporal fossae present) shows reversals, discussed by de P in n a (1996: 69 ), namely the “morphology and the other two (supracleithrum strongly attached to of the first proximal pectoral radial”, the “presence of a spur­ skull; posterior nuchal plate with anterior process facet for like process on the quadrate” and the “state of the humerover- articulation with anterior nuchal plate ) have a number of tebral ligament”. These refer to derived features present in the putatively homoplastic occurrences elsewhere in sil- amblycipitids and in Parakysis, but not m. Akysis, which, thus, uriforms” (D e P in n a , 1996: 76). The evidence presented could either be interpreted as independently acquired in the by d e P in n a (1996) to support a sister-group relationship amblycipitids and in Parakysis, or as acquired in the ambly­ between the Akysidae and the clade ( Sisoridae + ( Aspred­ cipitids plus akysids and subsequently reversed in Afo’sis. In inidae + Erethistidae)) was not significantly stronger than this latter case, they would support a sister-group relationship the evidence supporting the alternative, most traditional between the families Amblycipitidae and Akysidae. (see, e.g., R e a g a n , 1911; C h e n & L u n d b e r g , 1994) hy­ In summary, it can be said that although this study sup­ pothesis of a sister-group relationship between the Akysi­ ports, with an additional synapomorphy, D e P in n a ’s dae and the Amblycipitidae ( with the clade formed by (1996) grouping of the Akysidae, Sisoridae, Erethistidae these two families being, in turn, the sister-group of the and Aspredinidae, the phylogenetic position of the Ambly­ clade formed by the Sisoridae, Erethistidae and Aspredini­ cipitidae remains a problematic topic within the interrela­ dae). tionships of the superfamily Sisoroidea. Our observations and comparisons confirmed the three With respect to the synapomorphies of the Amblycipiti­ synapomorphies provided by d e P in n a (1996) to support dae, six characters were presented by d e P in n a ( see D e P in ­ the grouping of the Akysidae, Sisoridae, Aspredinidae and n a , 1996: 60). Our observations and comparisons pointed Erethistidae, and, in addition, pointed out an additional out a quite peculiar, derived character that is found in both synapomorphy to support the grouping of these four fam­ Liobagrus m á Amblyceps, and in no other catfish examined, ilies. This additional synapomorphy is described below: which, thus, could eventually constitute an additional am­ Coronoid process of the mandible exclusively formed blycipitid by synapomorphy: the hypobranchial foramen of the the dorsal margin o f the dentary. pamrohyal is markedly enlarged, with the length of the fo­ ramen being superior to half of the length of the main body Plesiomorphically in catfishes the coronoid process of of the pamrohyal ( Fig. 4 ). However, since it was not possi­ the mandible is constituted by the dorsal surfaces of both ble to dissect an exemplar of the other amblycipitid genera, the dentary and the ángulo-articular bones. However, in Xiurenbagrus (see Introduction), and since the pamrohyal the Akysidae, Erethistidae and Sisoridae the coronoid of Xiurenbagrus xiurensis, the only species of this genus, is process is exclusively formed by the dorsal margin of the not described in C h en & L u n d b e r g (1994 ), it was not pos­ dentary (see, e.g., D io g o et al., 2002: Fig. 3A). Therefore, sible to confirm if such a feature is, or not, present in this this character appears to represent a synapomorphy of the species. Therefore, only when a more detailed osteological clade constituted by the Akysidae, Sisoridae, Aspredinidae description of this species becomes available will it be pos­ and Erethistidae, which was secondarily reversed in the sible to confirm if this feature represents, or not, an unam­ aspredinids (see D io g o et al., 2001c). biguous synapomorphy of the family Amblycipitidae. Altough the present study supports, thus, D e P in n a ’s (1996 ) grouping of the Akysidae, Sisoridae, Erethistidae and ACKNOWLEDGEMENTS Aspredinidae, we consider, as did D e P in n a , that it is impor­ tant to point out that the evidence to support this grouping is We thank G.G. Teugels (MRAC), P. Laleyè (UNB), J. Wilhams scarce and that there are some conflicting characters with this and S. Jewett (USNM) and P. Duhamel (MNHN) for kindly provid­ hypothesis. One of these characters is the peculiar 7-shape of ing a large part of the specimens studied in this study. A great part of the sesamoid bone 1 of the Suspensorium ( Fig. 3). D e P in na this work was realised by R. Diogo at the Division of Fishes, USNM ( 1996: 70) mentioned this peculiar feature, mentioning that it (Washington DC). We are thus especially grateful for the support, was present in the amblycipitids and in Parakysis. However, assistance and advice received from R.P. Vari and S.H. Weitzman this peculiar feature is also present in the other akysid genus, during this period. We are also especially grateful to G. Arratia, who, Akysis, as it was clearly noted by C h en & L u n d b e r g (1994: through her valued close cooperation concerning the “Catfishes” 795) and confirmed in the present study. The presence of this book, significantly contributed, although indirectly, to the long stay peculiar, derived feature in the Amblycipitidae and Akysidae of R. Diogo at the USNM. We are also pleased to acknowledge the helpful criticism, advice and assistance of L. Taverne, M. Gayet, genera thus conflicts with the grouping of the Akysidae with B.G. Kapoor, F. Meunier, S. He, O. Otero, T.X. de Abreu, D. Adri­ the families Sisoridae, Aspredinidae and Erethistidae. The aens, F. Wagemans, C. Oliveira and E. Parmentier. This project re­ other conflicting character concerns the bifurcation of the ba­ ceived financial support from the following grant to R. Diogo: sal cartilages of the external mandibular barbels. Contrary to PRAXIS XXI/BD/19533/99 ("Fundaçâo para a Ciência e a Tecno­ the situation in most catfishes, in both the akysids and the am­ logía", Portuguese Federal Government). blycipitids ( see Fig. 5), as well as in the glyptostemin Sisori­ dae (see D io g o et al., 2002: Fig. 5), the cartilages of the REFERENCES outer mandibular barbels are bifurcated posteriorly. Attend­ ing to the well-supported grouping of the Sisoridae, Aspredi­ A l e x a n d e r , R.M. (1965). Structure and function in catfish. J. nidae and Erethistidae in a monophyletic clade (see D e Zool. (Lond.), 148: 88-152. P in n a , D io g o 1996; et al., 2001c, 2002), the taxonomic dis­ A r r a t ia , G. (1987). Description of the primitive family Diplo­ tribution of this character would support a sister-group rela­ mystidae (Siluriformes, Teleostei, Pisces): morphology, tax­ tionship between the Akysidae and the Amblycipitidae, with onomy and phylogenetic implications. Bonn. Zool. Mono gr., an independent, homoplastic acquisition in the glyptostemin 24: 1-120. 84 Rui Diogo, Michel Chardon and Pierre Vandewalle

Ar r a t ia , G. ( 1990). Development and diversity of the Suspenso­ D io g o , R., M. C h a r d o n & P. Vandewalle (2002). Osteology rium of trichomycterids and comparison with loricarioids and myology of the cephalic region and pectoral girdle of (Teleostei: Siluriformes)./. Morphol., 205: 193-218. Glyptothorax fukiensis (Rendalii, 1925), comparison with Ar r a t ia , G. ( 1 9 9 2 ). Development and variation of the Suspenso­ other sisorids, and comments on the synapomorphies of the rium of primitive catfishes (Teleostei: ) and their Sisoridae (Teleostei: Siluriformes). Belg. J. Zool., 132: 93- phylogenetic relationships. Bonn. Zool. Monogr., 32 : 1 -1 4 8 . 101 . Ar r a t ia , G. (1997). Basal and teleostean phylogeny. F e r r a r is , Cd. & M.C.C. d e P in n a (1999). Higher-level names Palaeo. Ichthyologica, 7: 5-168. for Catfishes (: Ostariophysi: Siluriformes). Ar r a t ia , G. & H-P. S c h u l t z e (1 9 9 0 ). The urohyal: develop­ Proc. Calif. Acad. Sei., 51: 1-17. ment and homology within osteichthyans. J. Morphol., 2 0 3 : G a u b a , R.K. (1962). The endoskeleton of Bagarius bagarius 2 4 7 -2 8 2 . (Ham.), part I - The skull. Agra Univ. J. Res., 11: 75-90. C h a r d o n , M. (1 9 6 8 ).Anatomie comparée de l’appareil de We­ G a u b a , R.K. (1966). Studies on the osteology of Indian sisorid ber et des structures connexes chez les Siluriformes. Ann. catfishes, II. The skull of Glyptothorax cavia. Copeia, 802- 4: Mus. R. Afr. Centn, 169: 1 -2 7 3 . 810. C h e n , L u n d b e r g X. & J.G. (1994).Xiurenbagrus , a new genus G a u b a , R.K. (1968). On the morphology of the skull of catfish of amblycipitid catfishes (Teleostei: Siluriformes), and phyl­ Pseudecheneis sulcatus. Zool. Anz., 181: 226-236. ogenetic relationships among the genera of Amblycipitidae. G a u b a , R.K. (1969). The head skeleton of Glyptosternum retic­ Copeta, 1994: 780-800. ulatum McClelland & Grifith. Mon. Zool. Ital., 3: 1-17. De P in n a , M.C.C. (1996). A phylogenetic analysis of the Asian G h io t , F. (1978). The barbel movements of three South Ameri­ catfish families Sisoridae, Akysidae and Amblycipitidae, can pimelodid catfishes. Zool. Anz., 200: 1-7. with a hypothesis on the relationships of the neotropical As- prenidae (Teleostei, Ostariophysi). Fieldi ana (Zool.), 84: 1- G h io t , F., P. V a n d e w a l l e & M. C h a r d o n (1984). Comparaison 82. anatomique et fonctionnelle des muscles et des ligaments en rapport avec les barbillons chez deux familles apparentées de D e P in n a , M.C.C. (1998). Phylogenetic relationships of neotrop­ ical Siluriformes: Historical Overview and Synthesis of Hy­ poissons Siluriformes Bagroidei. Ann. Soc. R. Zool. Belg., 114: 261-272. potheses. In: M a l a b a r n a L.R, R.E. R e is , R.P. V a r i , Z.M. L u c e n a & C.A.S. L u c e n a (eds), Phylogeny and classifica­ G o s l in e , W. A. ( 1975 ). The palatine-maxillary mechanism in cat­ tion o f neotropical fishes, Edipucrs, Porto Alegre: 279-330. fishes with comments on the evolution and zoogeography of D io g o , R. & M. C h a r d o n (2000a). Homologies between differ­ modem siluroids. Occ. Pap. Calif Acad. Sei., 120: 1-31. ent adductor mandibulae sections of teleostean fishes, with a H e , S. (1996). The phylogeny of the glyptostemoid fishes (Tele­ special regard to catfishes (Teleostei: Siluriformes). J. Mor­ ostei: Siluriformes, Sisoridae). Cybium, 20: 115-159. phol., 243: 193-208. M a h a j a n , C.F. (1963). Sound producing apparatus in an Indian D io g o , R. & M. C h a r d o n (2000b). The structures associated catfish Sisor rhabdophorus Hamilton. J. Linn. Soc. (Zool.), with catfish (Teleostei: Siluriformes) mandibular barbels: or­ 44: 721-724. igin, anatomy, function, taxonomic distribution, nomencla­ M a h a j a n , C.F. (1966a). Sensory canals of the head in Sisor ture and synonymy. Neth. J. Zool., 50: 455-478. rhabdophorus Hamilton. Trans. Am. Micr. Soc., 85: 548-555. D io g o , R. & M. C h a r d o n (2 0 0 0 c ). Anatomie et fonction des M a h a j a n , C.F. (1966b). Sisor rhabdophorus - A study in adap­ structures céphaliques associées á la prise de nourriture chez tation and natural relationship. I. The head skeleton. J. Zool. le gerne Chrysichthys (Teleostei: Siluriformes). Belg. J. Zo­ (Lond.), 149: 365-393. o l, 130: 21-37. M a h a j a n , C.F. (1967a).Sisor rhabdophorus - A study in adap­ D io g o , R. & M. C h a r d o n (in press). On the homologies and ev­ tation and natural relationship. II. The interrelationships of olutionary transformation of the skeletal elements of catfish the gas bladder, Weberian apparatus, and membranous laby­ (Teleostei: Siluriformes) Suspensorium: a morpho-functional rinth. J. Zool. (Lond.), 151: 417-432. hypothesis. In: V a l A.L. & B.G. Ka p o o r (eds), Fish adapta­ M a h a j a n , C. F. (1967b). Sisor rhabdophorus A study - in adap­ tions, Oxford & IBH Publishing and Science Publishers, New tation and natural relationship. III. The vertebral column, me­ Delhi and New Hampshire. dian fins and their musculature. J. Zool. (Lond.), 152: 297- D io g o , R., P. V a n d e w a l l e & M. C h a r d o n (1999). Morpholog­ 318. ical description of the cephalic region of Bagrus docmak, Mo, T. ( 1991 ). Anatomy, relationships and systematics of the Ba­ with a reflection on Bagridae (Teleostei: Siluriformes) au- gridae (Teleostei: Siluroidei) with a hypothesis of siluroid tapomorphies. Neth. J. Zool., 49: 207-232. phylogeny. Theses Zoologicae, 17: 1-216. D io g o , R., C. O l iv e ir a & M. C h a r d o n (2000). The origin and R e g a n , C.T. (1911). The classification of the teleostean fishes of transformation of catfish palatine-maxillary system: an ex­ ample of adaptive macroevolution. Neth. J. Zool., 50: 373- the Ostariophysi: 2. Siluroidea./w;. &Mag. Nat. Hist., 388. 8: 553-577. T a y l o r , W.R. & G.C. V a n D y k e (1985). Revised procedure for D io g o , R., C. O l iv e ir a & M. C h a r d o n (2001a). On the homol­ ogies of the skeletal components of catfish (Teleostei: Sil­ staining and clearing small fishes and other vertebrates for uriformes) Suspensorium. Belg. J. Zool., 131: 93-109. bone and cartilage study. Cybium, 2: 107-119. D io g o , R., C. O l iv e ir a & M. C h a r d o n (2001b). On the osteol­ T e u g e l s , G.G. (1996). , phylogeny and biogeography ogy and myology of catfish pectoral girdle, with a reflection of catfishes (Ostariophysi, Siluroidei): an overview. Aquat. on catfish (Teleostei: Siluriformes) plesiomorphies. J. Mor­ Living Résout:, 9-34.9: phol., 249: 100-125. T il a k , R. (1967). The osteocranium and the Weberian apparatus D io g o , R., M. C h a r d o n & P. V a n d e w a l l e (2 0 0 1 c ). Osteology of Amblyceps mangois (Hamilton) (Pisces: Siluroidei) in re­ and myology of the cephalic region and pectoral girdle of Bu­ lation to taxonomy. Zool. Anz., 178: 61-74. nocephalus bierii, and a discussion on the phylogenetic rela­ W interbottom , R. ( 1974). A descriptive synonymy of the striat­ tionships of the Aspredinidae (Teleostei: Siluriformes). Neth. ed muscles of the Teleostei. Proc. Acad, Nat. Sei. (Phil.), 125: J. Zool., 51: 457-481. 225-317.

Received: July 5, 2002 Accepted: December 4, 2002