Journal of Fish Biology (1997) 50, 1221–1257

The ontogeny of the chondrocranium in Clarias gariepinus:trends in siluroids

D. A W. V

University of Ghent, Institute of Zoology, K.L. Ledeganckstraat 35, B-9000 Ghent,Belgium

(Received 11 October 1996, Accepted 4 January 1997)

The ontogeny of the chondrocranium of 31 different stages of the African catfish Clariasgariepinus (Siluroidei: Clariidae) was studied, both from cleared and stained, and sectionedmaterial. The fish ranged from 4·1 (1 day post-hatching) to 127·0 mm SL (100 dayspost-hatching). The chondrocranium of C. gariepinus seemed to correspond to the generaladaptive trends in siluroids, especially in relation to the reduction of eye size and thedorso-ventral flattening of the skull. The platybasic neurocranium involved several modifica-tions related to the trabecular bars, the hypophyseal fenestra, the ethmoid region and even theolfactory nerves. Certain reductions were present, which have been observed in all siluroids(e.g. absence of the pila lateralis, the commissura lateralis, the myodomes) or are part of avariable trend within siluroids (e.g. reduction of the taenia marginalis anterior and the tectumsynoticum). Compared with some other siluroid species, the neurocranium of C. gariepinus iswell developed, for example in the otic region. The same was observed in the splanchnocraniumwhere some general siluroid trends persist (e.g. isolation of palatine from pterygoquadrate,presence of ‘ hyo-symplectic-pterygoquadrate ’ plate). Some trends, as observed in othersiluroids, were present also (e.g. interhyal continuous with suspensorium and ceratohyal,Meckel’s cartilage initially continuous with the suspensorium). The branchial basket is welldeveloped as all expected elements are present (basibranchials I–IV, hypobranchials I–IV,ceratobranchials I–V, epibranchials I–IV). Based on the observed ontogeny of C. gariepinusand data from the literature, a hypothesis was formulated which indicated the presence of ageneral reductional trend within siluroids. In C. gariepinus, all four (I–IV) infrapharyngo-branchials develop, although the anterior two are much reduced and fused with each other.

? 1997 The Fisheries Society of the British Isles

Key words: ontogeny; chondrocranium; Siluroidei; Clarias gariepinus.

INTRODUCTION

The study of the skull of vertebrates, and especially of osteichthyans, is restrictedin most cases to the adult bony part. Those studies dealing with the ontogeny ofthe skull, frequently focus on the development of the osteological componentsof the fish skull, with little or no attention paid to the relation with thechondrocranium (e.g. Kobayakawa, 1992; Vandewalle et al., 1995). A compi-lation of studies, made on the ontogeny of the cartilaginous skull of fishes, hasbeen given by de Beer (1937) and Daget (1964). Considerable attention has beenpaid to the chondrocranium of siluroids: Ictaluridae (Kindred, 1919; Arratia,1992), Callichthyidae (Hoedeman, 1960; Howes & Teugels, 1989), Ariidae(Bamford, 1948; Srinivasachar, 1958a), Plotosidae (Srinivasachar, 1958a),Schilbeidae (Srinivasachar, 1957a), Bagridae (Srinivasachar, 1957b); Clariidae(Surlemont et al., 1989; Surlemont & Vandewalle, 1991; Adriaens & Verraes,1994), Heteropneustidae (Srinivasachar, 1959).

Tel.: +(32) 9/264 52 18; fax: +(32) 9/264 53 44; email: dominique.adriaens@rug.ac.be

1221

0022–1112/97/061221+37 $25.00/0/jb970399 ? 1997 The Fisheries Society of the British Isles

1222 . .

Some important tranformations have been observed in the general mor-phology of the siluroid chondrocranium. The major part of these changes can berelated in some way to the benthic and nocturnal lifestyle of many catfishes.Most striking are the consequences of the reduction in eye size. The conse-quently decreased spatial competition allows the contra-lateral trabecular bars toremain separated along the greater part of their length, thus forming a largehypophyseal fenestra. This type of skull construction belongs to the platybasic(=platytrabecular) type, in contrast to the tropibasic (=tropidotrabecular,tropitrabecular, tropidobasic) type which is found in most teleosts (Daget, 1964).A reduction of the eye diameter also implicates a reduction in the bulk andlength of the extrinsic eye muscles, and no, or rudimentary, myodomes to housethem are needed (Daget, 1964; Alexander, 1965). The platybasic skull ofsiluroids generally becomes dorso-ventrally depressed, which is believed to be anadaptation for stability when lying on the bottom, as well as to reduce the dragin currents (Alexander, 1965). Another adaptation for this benthic behaviourmay be the loss of the interhyal during ontogeny in different catfishes (Adriaens& Verraes, 1994). A reduction in visual input may be compensated in catfishesby the presence of a set of oral barbels, equipped with taste and touch sensoryorgans (Alexander, 1965; Ghiot & Bouchez, 1980). One pair of these barbels, themaxillary barbels, is connected to a palatine–maxillary mechanism which enablesa controlled mobility. The development of this mechanism is enabled partiallyby the isolation of both skeletal and muscular elements of the palatine part of thepalatoquadrate [Gosline, 1975; Fink & Fink, 1981 (character 24); Ghiot et al.,1984; Arratia, 1992 (character 2)].In the present paper, a detailed description is given of the ontogeny of the

chondrocranium in the African sharptooth catfish Clarias gariepinus (Burchell,1822) [synonymized with C. lazera and C. mossambicus (Teugels, 1982)]. Severalstages are used to describe the morphological transformations that occur,resulting in the cartilaginous template for the bony skull. Seven of them aredescribed in this paper. The description of the developing osteocranium isbeyond the scope of this paper, but will be dealt with in a separate one. Someattention is paid to the development of the orbito-nasal region, as well as to theformation of the foramina of the otic region. Special attention is paid to thedevelopment of the infrapharyngobranchials, where an attempt is made todistinguish a trend within the Siluroidei. Based on these data and data fromliterature, some considerations are made concerning the comparative mor-phology of siluroid chondrocrania and some trends during its ontogeny. Theontogeny of the postcranial Weberian apparatus and vertebrae is not described(but refer to Radermaker et al., 1989).

MATERIALS AND METHODS

Several specimens of Clarias gariepinus of different ontogenetic stages were used toexamine the morphology of the chondrocranium. In total, 31 larvae of different lengthsand ages were studied, ranging from 4·1 mm SL (standard length) (1 day post-hatching)to 127·0 mm SL (100 days post hatching) (Table I). Eggs were obtained from theLaboratory of Aquaculture and Ecology (Catholic University of Leuven), and raised ata temperature of 25) C. At different time intervals, larvae were sedated in MS-222 and

1223

fixed in a 4% buffered formaldehyde solution. Specimens were used for serial sectioningand in toto clearing and staining. Clearing and staining procedures followed those ofHanken & Wassersug (1981), in which trypsin is replaced by a 1% KOH solution. Thecleared specimens were studied using a stereoscopic microscope (WILD M5). Six of theexamined specimens were used for serial sectioning. Fixation was in paraformaldehyde-glutaraldehyde or buffered formaldehyde solutions, whereas embedding was doneusing EPON or Paraplast. The 2-ìm sections were stained with toluidine. Sectionswere studied using a Leitz Diaplan light microscope. Drawings were made using acamera lucida, and 3D reconstructions were made by means of a commercial softwarepackage.Some of the examined specimens came from different egg batches, whereas 19 of them

belonged to a single, gynogenetical batch (Volckaert et al., 1994) (Table I).

T I. Specimens of Clarias gariepinus used in the present study

No. SL(mm)

TL(mm)

PAL(mm)

Age(days

post-hatch)Method Staining Used for

1 4·1 4·4 2·3 1·0 Clearing AB+ARS Drawing2 5·6 5·9 2·9 1 Serial sections T 3D reconstruction3 5·8 6·8 3·5 2 Clearing AB+ARS Drawing4 5·9 6·4 3·3 2 Serial sections T 3D reconstruction5 6·0 6·4 2·7 1 Clearing AB+ARS Drawing6 6·0 6·8 3·5 2 Clearing AB+ARS Drawing7 6·6 7·0 3·2 2 Clearing AB+ARS Drawing8* 6·8 7·4 3·1 3 Clearing AB+ARS Observation9 6·9 7·5 3·4 3 Clearing AB+ARS Drawing10* 6·9 7·5 3·6 7 Clearing AB+ARS Observation11 7·0 7·5 3·2 3 Clearing AB+ARS Observation12* 7·1 8·0 3·5 10 Clearing AB+ARS Drawing13 7·1 8·7 3·7 9 Clearing AB+ARS Drawing14 7·2 7·8 3·8 7 Serial sections T Observation15* 7·4 8·1 3·6 13 Clearing AB+ARS Observation16* 7·7 8·2 7·1 17 Clearing AB+ARS Observation17* 7·8 8·4 4·1 17 Clearing AB+ARS Observation18* 8·1 9·1 4·2 13 Serial sections T Observation19* 8·2 8·6 4·1 17 Clearing AB+ARS Observation20* 8·4 9·1 4·2 13 Serial sections T Observation21* 9·3 10·7 4·9 21 Clearing AB+ARS Observation22* 9·5 10·1 4·8 21 Clearing AB+ARS Drawing23* 10·0 11·3 5·0 21 Clearing AB+ARS Drawing24* 11·6 13·0 5·0 26 Clearing AB+ARS Observation25* 12·0 13·5 6·2 26 Clearing AB+ARS Observation26* 12·7 15·0 6·1 26 Clearing AB+ARS Observation27* 14·8 16·4 8·1 31 Clearing ARS Drawing28* 15·2 16·9 8·0 26 Serial sections T Observation29* 19·0 21·1 11·1 31 Clearing AB+ARS Drawing30* 21·5 24·7 11·6 31 Clearing AB+ARS Observation31 127·0 143·6 65·5 100 Clearing AB+ARS Observation

*Gynogenetic specimens.AB, alcian blue; ARS, alizarine red S; PAL, preanal length; SL, standard length; T, toluidine; TL, total

length.

1224 . .

RESULTS

The applied nomenclature for the anatomical structures follows the nomen-clature used by Daget (1964). Arguments for some divergent nomenclature aregiven in the discussion.

4·1-MM SL STAGE (Fig. 1)NeurocraniumOne day after hatching, the major part of the skull base is formed. The

chondrification seems to be concentrated in close contact with the notochord.The anterior part of the notochord is enclosed by two lateral cartilaginous strips,which are interconnected anteriorly. Based on their position, the lateral elementsmust correspond with the parachordal cartilages. The cartilage abutting againstthe anterior margin of the notochord then represents the acrochordal cartilage.Antero-laterally, an anterior basicapsular commissure connects the parachordalsto the anterior otic cartilage. At one side, in this specimen, the posterior oticcartilage can be distinguished already, bearing a small ventral extension. Noasymmetry, however, could be observed in any other stage. This part covers thepostero-lateral part of the otic vesicle. In general, the basicapsular commissures

tr(a) c-pol cm-bc-a

c-Meck

ch

c-ac c-ot-a c-ot-p

lm-bot

fs-mot

c-pc

0.2 mm

prc-opfs-mot

hsfr-tr-hm c-ot-p

cm-bc-a

lm-botc-pc

cb-I

hs

tr-cr

ch

c-Meck

p-q

cb-I0.2 mm

ih

0.2 mm

c-ot-a c-ot-p

ch

c-Meck

bb-I

prc-opp-q ih

(b)

(c)

F. 1. Chondrocranium of Clarias gariepinus (4·1 mm SL): (a) dorsal view; (b) lateral oblique view; (c)ventral view. bb-I, Basibranchiale I; c-ac, cartilago acrochordalis; c-Meck, cartilago Meckeli;c-ot-a, cartilago oticalis anterior; c-ot-p, cartilago oticalis posterior; c-pc, cartilago parachordalis;c-pol, cartilago polaris; cb-I, ceratobranchiale I; ch, ceratohyale; cm-bc-a, commissurabasicapsularis anterior; fr-tr-hm, foramen truncus hyomandibularis nervus facialis; fs-mot, fissurametotica; hs, hyosymplecticum; ih, interhyale; lm-bot, lamina basiotica; p-q, pars quadrata of thepalatoquadratum; prc-op, processus opercularis; tr, trabecula; tr-cr, trabecula cranii.

1225

connect the otic cartilages to the parachordal plate (or basal plate), the lattercorresponds to the fused parachordal cartilage and the basiotic lamina. How-ever, this lamina can hardly be distinguished as a separate element inC. gariepinus. The fusion with the parachordal cartilages may have occurredalready at the level of blastemes or the basiotic lamina may be lost. The partiallyenclosed, large foramen, surrounded by the otic cartilages, the anteriorbasicapsular commissure and the parachordal plates, will be penetrated by boththe glossopharyngeus (IX) and vagus (X) nerves. Consequently, this foramencorresponds to the metotic fissure [Fig. 9(a)].The interorbital region is supported already, as the paired cartilaginous

trabecular bars (trabeculae cranii) are extended from the antero-lateral margin ofthe acrochordal cartilage. Close to this connection, each trabecular bar is curvedoutwards slightly, whereas the larger rostral part runs parallel to the median planof the skull. Most probably, the caudal small part corresponds to the polarcartilage (cartilago polaris), whereas the rostral part corresponds to the trabecula(s.s.). The position of the fissure, through which passes the internal carotidartery, and which is observed clearly in later stages, supports this hypothesis (seeDiscussion). Anteriorly, the trabecular bars are still separated from each other,thereby bordering a relatively large interorbital space, as is to be expected in aplatybasic skull.

SplanchnocraniumThe cartilaginous dorsal and ventral elements of the mandibular and hyoid

arch are present already, whereas the premandibular palatine is not. Allmandibular and hyoid parts are fused to each other from the moment theyarise, thus forming a single cartilaginous piece. The distinction between thepterygoquadrate (dorsal part of the mandibular arch) and the hyosymplectic(dorsal part of the hyoid arch) cannot be made. This part of the fusedcomplex has been referred to as the ‘ hyo-symplectic-pterygoquadrate plate ’(Arratia, 1992). The hyosymplectic bears a distinct posterior opercular pro-cess, for the articulation with the opercular bone, which has started todifferentiate already at this stage. The foramen for the passage of thehyomandibular nerve trunk (VII) is formed already. Dorsally, the hyosym-plectic articulates with the neurocranium at the level of the anterior oticcartilage.Surprisingly, Meckel’s cartilage is fused to the pars quadrata of the pterygo-

quadrate, whereas the certohyal is fused to the hyosymplectic. The latterfusion occurs through a small cartilaginous bar which corresponds to theinterhyal, as can be derived from later stages. Left and right lower jaws havefused rostrally. The ceratohyals are fused together also, via a cartilaginousmass interconnecting their rostral tips. The latter cartilaginous mass has theposition of the basihyal although presumably does not correspond to it (seeDiscussion). Posterior to and continuous with this alleged basihyal, lies aforked chondrification. The anterior part corresponds to the first basi-branchial cartilage, whereas the two branches correspond to the future hypo-branchials I and ceratobranchials I. Both elements, however, cannot yet bedistinguished at this stage. No epi- or infrapharyngobranchials nor otherbranchial arches could be observed.

1226 . .

5·7-MM SL STAGE (Fig. 2)

NeurocraniumAt this stage, both posterior and anterior otic cartilages form at both sides.

Caudally, the posterior otic cartilages become extended, as they grow in thedirection of a lateral process of the parachordal plates. Serial sections reveal thetrue nature of this process, as the path of the glossopharyngeus (IX) and vagus(X) nerves could be followed. The glossopharyngeus nerve passes through theskull floor, anterior to this process, whereas the vagus nerve passes it posteriorly.Thus, this process corresponds to the posterior basicapsular commissure and notto the occipital arch, which lies posterior to the nervus vagus. Consequently, thelarge foramen is bordered by: (1) the anterior basicapsular commissure rostrally;(2) the parachordal plates medially; (3) the otic cartilages laterally; and (4) theposterior basicapsular commissure caudally. As the vagus nerve no longerpasses through the skull floor through this foramen, it no longer corresponds tothe fissura metotica but to the basicapsular fenestra. Antero-dorsally, theanterior otic cartilage bears a small process, which corresponds most probably tothe primordia of the taenia marginalis posterior. Anteriorly, the trabecular barshave become elongated and start to curve medially, but still remain separated

cb-III

(a)

0.1 mm

tn-m-ptr-cr

hs

cb-IIIcb-IIcb-I

c-ot-a c-ot-p c-ac

cm-bc-apal

sol-n

p-qc-Meckch

bb-I 0.1 mm

pal

prc-op

hs

sol-np-q

c-Meck

cd fn-bc

cm-bc-a

pc-pl

ih

ch

cb-I 0.1 mm

c-ot-pc-ot-a

cm-bc-p

cb-IIbb-I

(b)

(c)

F. 2. Graphical 3D reconstructions of the chondrocranium of Clarias gariepinus (5·6 mm SL): (a)frontal view of the chondrocranium situated in the head; (b) frontal view; (c) caudal view. bb-I,Basibranchiale I; c-ac, cartilago acrochordalis; c-Meck, cartilago Meckeli; c-ot-a, cartilago oticalisanterior; c-ot-p, cartilago oticalis posterior; cb-I, ceratobranchiale I; cb-II, ceratobranchiale II;cb-III, ceratobranchiale III; cd, chorda dorsalis; ch, ceratohyale; cm-bc-a, commissura basi-capsularis anterior; cm-bc-p, commissura basicapsularis posterior; fn-bc, fenestra basicapsularis;hs, hyosymplecticum; ih, interhyale; p-q, pars quadrata of the palatoquadratum; pal, palatinum;pc-pl, parachordal plate; prc-op, processus opercularis; sol-n, solun nasi; tn-m-p, taenia marginalisposterior; tr-cr, trabecula cranii.

1227

from each other in the ethmoid region, and the trabecular bars have formed alateral plate-like extension, the solum nasi.

Splanchnocranium

For the first time, the palatine is discernible, already articulating with thesolum nasi of the ethmoid plate. From this level on, the cartilaginous connec-tion between Meckel’s cartilage and the pars quadrata of the pterygoquadrateis lost, and an articulation is formed. The serial sections showed that anarticulation was present but not yet pronounced, which indicates that thearticulation may just have been formed at a SL of about 5·6 mm. Rostrally,both mandibular bars are still fused together. The ceratohyals are still continu-ous with the hyosymplectic, through the interhyal, as well as being connectedto each other rostrally. The cartilage interconnecting the hyoid bars is elongatedposteriorly, as it is fused to the basibranchial element of the first branchial arch.The hyosymplectic itself has undergone no major transformations, exceptbecoming more substantial. Rudiments from the first three branchial archescan be distinguished now. The anterior one consists of the ceratobranchialswhich are continuous with the basibranchials. The differentiation of thehypobranchial I, which arises from the cartilage, anterior to the cerato-branchial, could not be made yet. The following two arches are represented byrudimentary ceratobranchials only: the primordium of the ceratobranchial IIIbeing shorter than that of ceratobranchial II. Still no epi- or infrapharyngo-branchials could be discerned. Sections showed that these three branchial archeswere already functional, as hemibranchial structures were present at their caudalside.

6·0-MM SL STAGE (Fig. 3)

Neurocranium

At this level, the ethmoid region becomes differentiated. The trabecular barshave reached each other anteriorly, where they have fused. Based on the shapeof the ethmoid plate (cartilago ethmoideum) so formed, it seems that the fusionhas occurred through a medial extension of the trabecular bars and is not onlythe result of the mediad curvature of the rostral tips. Laterally, the ethmoidplate has small processes corresponding to pre-ethmoid cornua. This trabecularfusion occurs earlier as it is observed already in a 5·8-mm SL larva, thus forminga large hypophyseal fenestra. Rostrally, the ethmoid plate bears a dorsalprocess, which will fuse with the skull roof later and which will contributeeventually to the formation of the precerebral lamina. In most fishes a medialand single process is formed, referred to as the septum internasale. InC. gariepinus, however, two lateral processes are formed. To avoid confusionand misinterpretation, these processes in C. gariepinus will be referred to as theprecerebral process. At the lateral face of the ethmoid plate, the solum nasi hasbecome more pronounced and has formed a dorsal process: the orbito-nasallamina, sensu strictu. This lamina is situated at the level of the articulationbetween the solum nasi and the palatine. The taeniae marginales posterioresnow have become extended anteriorly, where they become branched at the levelof the orbito-nasal lamina sensu strictu. Three branches are formed: (1) medially,

1228 . .

the rudiments of the epiphysial bridge (pons epiphysialis) arise; (2) rostrally, aspheno-septal commissure reaches up to the precerebral process of the ethmoidplate; and (3) laterally, a spheno-ethmoidal commissure is directed to theorbito-nasal lamina, sensu strictu. The fact that the lamina orbito-nasalis, sensustrictu, is situated at the same level as the epiphysial bridge, suggests that theanterior taenia marginalis is reduced strongly in C. gariepinus. Prior to thefusion between the spheno-ethmoidal commissure and the orbito-nasal lamina,sensu strictu, a foramen for the ophthalmic ramus of the trigeminal nerve isformed [Fig. 8(a)]. Serial sections of a 5·9-mm SL stage revealed, however, thatthis foramen arises by the secondary formation of a cartilaginous bridge betweenthe spheno-ethmoidal commissure and the spheno-septal one, ventrally to theophthalmic ramus (Fig. 4). The fused spheno-ethmoidal commissure andorbito-nasal lamina, sensu strictu give rise to the first reinforcement between theskull roof and the skull floor, i.e. the orbito-nasal lamina, sensu latu. The

fs-sph(a)

tr-cr a-sc-c

pns-ep

cn-peth

pal tn-m-p prc-ptorb

fr-X

ot-cap

cd

0.2 mm

v-sc-c p-sc-c

cm-sphsep

fn-hyp

c-eth

sol-n

fs-car-i

pl-oc

fs-sph(b)

tr-cr hs

pns-ep

c-sbmx

pal c-Meck prc-ra

prc-op

fr-tr-hm

eb-IV

0.2 mm

tn-m-p ot-cap

cm-spheth

prc-pc

lm-on(s)

sol-n

ch

cb-IV

ih

p-q

prc-pt

cm-sphsep

prc-ra(c)

p-q prc-op

ch

pal

hb-II cb-II

cb-V

ot-cap

fr-X

0.2 mm

c-ac

bb-III

hh

c-Meck

c-sbmx

eb-II

ipb-IV

bb-II

bb-I

F. 3. Chondrocranium of Clarias gariepinus (6·0 mm SL): (a) dorsal view; (b) lateral view; (c) ventralview. a-sc-c, Anterior semicircular canal; bb-I, basibranchiale I; bb-II, basibranchiale II; bb-III,basibranchiale III; c-ac, cartilago acrochordalis; c-eth, cartilago ethmoideum; c-Meck, cartilagoMeckeli; c-sbmx, cartilago submaxillaris; cb-II, ceratobranchiale II; cb-IV, ceratobranchiale IV;cb-V, ceratobranchiale V; cd, chorda dorsalis; ch, ceratohyale; cm-spheth, commissura spheno-ethmoidalis; cm-sphsep, commissura spheno-septalis; cn-peth, cornua prae-ethmoidea; eb-II,epibranchiale II; eb-IV, epibranchiale IV; fn-hyp, fenestra hypophysea, fr-tr-hm, foramen truncushyomandibularis nervus facialis; fr-X, foramen nervus vagus; fs-car-i, fissura arteria carotisinterna; fs-sph, fissura sphenoidea; hb-II, hypobranchiale II; hh, hypohyale; hs, hyosymplecticum;ih, interhyale; ipb-IV, infrapharyngobranchiale IV; lm-on(s), lamina orbitonasalis, sensu strictu;ot-cap, otic capsule; p-q, pars quadrata of the palatoquadratum; p-sc-c, posterior semicircularcanal; pal, palatinum; pl-oc, pila occipitalis; pns-ep, pons epiphysialis; prc-op, processusopercularis; prc-pc, processus praecerebralis; prc-pt, processus pterygoideus of the palato-quadratum; prc-ptorb, processus postorbitalis; prc-ra, processus retroarticularis; sol-n, solum nasi;tn-m-p, taenia marginalis posterior; tr-cr, trabecula cranii; v-sc-c, vertical semicircular canal.

1229

now partially enclosed sphenoid fissure, surrounded by the taenia marginalisposterior, the trabecular bar and the anterior basicapsular commissure, is thenpenetrated by several cranial nerves: the olfactory nerve (I) (fila olfactoria), theoptic nerve (II) (fasciculus opticus), the oculomotorius nerve (III), the trochlearisnerve (IV), the trigeminus nerve (V), the abducens nerve (VI) and the facialisnerve (VII). At the level of their fusion with the otic capsules, the taeniaemarginales posteriores have become broader. This differentiation is referred to asthe postorbital process, which bears, at its ventro-lateral face, the anterior part ofthe hyosymplectic articulation.The trabecular bar bears a dorsal process, medial to the orbito-nasal lamina,

which will give rise to the preorbital base (preorbital root) [Fig. 8(b)]. In general,however, the preorbital base arises as a ventral outgrowth of the taeniamarginalis anterior, and not as a dorsal extension of the trabecular bars (Daget,1964). Posteriorly, the trabecular bars show a small, medial excavation for thehousing of the internal carotid artery. In the otic region, the formation of acartilaginous roof covering the brain has been initiated, as the otic cartilagesstart to grow medially. At this stage, both otic cartilages are connected throughcartilaginous bars to the parachordal plate, leaving two foramina. From theposition of the foramina and the penetration of certain cranial nerves, as well asobservations of further stages, it was made possible to reveal the possible truenature of these structures (Fig. 9). The basicapsular fenestra of the previousstage (5·6 mm SL) has now become subdivided into two small foramina by theformation of the basivestibular commissure. The anterior basicapsular fenestrais enclosed by: (1) the anterior basicapsular commissure anteriorly; (2) the oticcartilages laterally; (3) the parachordal plates medially; and (4) the basivestibularcommissure posteriorly. Presumably, no nerve or bloodvessel passes through

(a)

tn-m-p

0.1 mm

rm-mx-V

rm-mnd-V

pns-ep

cm-sphseprm-ophth-V

cm-spheth

n-II

tr-cr

pal rm-mx-V

rm-mnd-V0.1 mm

(b)

F. 4. Graphical 3D reconstructions of the orbital region of the chondrocranium (left side) of Clariasgariepinus, showing the path of the ramus ophthalmicus of the nervus trigeminus (5·9 mm SL): (a)caudal view; (b) frontal view (small circles indicate cartilage transections, shaded areas indicatenerve transections. cm-speth, Commissura spheno-ethmoidalis; cm-sphesep, commissura spheno-septalis; n-II, fasciculus opticus; pal, palatinum; pns-ep, pons epiphysialis; rm-mnd-V, ramusmandibularis nervus trigeminus; rm-mx-V, ramus maxillaris nervus trigeminus; rm-ophth-V, ramusophthalmicus nervus trigeminus; tn-m-p, taenia marginalis posterior; tr-cr, trabecular cranii.

1230 . .

this fenestra as it disappears later during ontogeny. The other foramen, theposterior basicapsular fenestra is then surrounded by: (1) the basivestibularcommissure anteriorly; (2) the posterior otic cartilage laterally; (3) theparachordal plate medially; and (4) the posterior basicapsular commissureposteriorly. This fenestra corresponds to the foramen of the glossopharyngealnerve since that nerve runs through it. Caudally to the posterior basicapsularcommissure, the occipital arch is formed, which is already fused with theparachordal plates. These paired pilae occipitales grow dorsally, as they borderthe notochord, but have a latero-dorsal extension as well, toward of the oticcapsules.

SplanchnocraniumThe major changes that occurred in this stage are related to the further

differentiations of the gill arches, whereas several bones have started theirdevelopment. In relation to the ossification of the maxillary bone, a submaxil-lary cartilage has formed at the rostral tip of each palatine. This secondarycartilage facilitates the articulation between the palatine and the maxillary bone,as part of the palatine–maxillary mechanism. At the posterior tip of Meckel’scartilage a more pronounced retroarticular process is formed, whereas a distinctarticulation with the quadrate is now present. At the antero-dorsal border ofthe quadrate, a small pterygoid process is present. The ceratohyals are ratherflattened structures, directed in an oblique vertical plane. They are still inter-connected rostrally, but the connection with the basibranchial is lost. At theirrostral tips, the hyoid bars now show a small incision at the anterior face, wherethe hyoid artery passes. Thus the first signs of the differentiation of thehypohyals have appeared. No sign of a basihyal is observed. The hyosymplecticarticulates with the ventro-lateral wall of the neurocranium, in front andpartially below the anterior semicircular canal.The ceratobranchials II and III have become extended now, apparently both

ventrally and dorsally. The ceratobranchials IV and V have developed as wellat this stage. The medial tips of ceratobranchials I and II are broadened,which corresponds to hypobranchials I and II. The hypobranchials have notyet become separated from the ceratobranchials. Between the left and righthypobranchials I and II, a single cartilaginous copula is present. At this stage,it consists of the fused basibranchials I, II and III, corresponding to theanterior copula. All epibranchials (I–IV), as observed in adults, are presentnow, as well as all infrapharyngobranchials (I–IV). No uncinate process couldbe observed on epibranchial III. All epibranchials are rod-like, although thelast one, epibranchial IV, seems to be more substantial than the other ones.The infrapharyngobranchials are built differently: I and II are strongly reducedand continuous with the epibranchials and have fused to each other and alsoto the third, whereas III and IV are separate, solid elements articulating withthe epibranchials. The latter two lie perpendicular to the correspondingepibranchials. The difference in shape is related to the fact that theinfrapharyngobranchials III and IV play a supportive role for the upperpharyngeal jaws, whereas the anterior two do not. These pharyngeal jaws,which are dermal bone plates, including dentition, were already observed inthis stage.

1231

6·6-MM SL STAGE (Fig. 5)

Neurocranium

As most structures were formed in the previous stage, further development isfor the greater part determined by fusion and enlargement of those existingstructures. In the ethmoid region, two additional reinforcements between skullroof and skull floor have occurred: (1) the precerebral process fuses with thespheno-septal commissure; and (2) the preorbital base fuses with the taeniamarginalis posterior. As a result, the sphenoid fissure of the previous stage nowbecomes completely closed and divided. The foramen, anterior to the preorbitalbase, is penetrated by the fila olfactoria. The foramen, posterior to this base, isthen penetrated by the other nerves which previously penetrated the sphenoidfissure (II–VII). The latter foramen is referred to as the sphenoid fenestra. Thefila olfactoria only penetrate the olfactory foramen and do not run through theorbito-nasal foramen. The latter foramen, bordered by the orbito-nasal lamina

fn-sph(a)

fs-car-i fn-hyp

sol-n

pal

tn-m-p tr-cr

cd

0.2 mm

ot-cap prc-ptot

cm-sphsep

lm-on

c-eth

pns-ep

c-ac

pl-oc

fn-sph(b)

lm-on pns-ep

porb-b

c-sbmx

ch ihprc-ra

prc-op

fr-tr-hm

eb-IV

0.2 mm

tn-m-p hs

fr-I

tr-cr

c-Meck

pal

p-q

cb-IV

prc-pt

cm-sphsep

prc-co(c)

cb-II eb-I

cb-I

cp-a

hb-IIIpal

cb-V

ibb-III

fr-X

0.2 mm

ot-cap

hb-I

c-Meck

ch

bb-IV

ipb-IV

prc-ptot

pl-oc

prc-op

hh

F. 5. Chondrocranium of Clarias gariepinus (6·6 mm SL): (a) dorsal view; (b) lateral view; (c) ventralview. bb-IV, Basibranchiale IV; c-ac, cartilago acrochordalis; c-eth, cartilago ethmoideum;c-Meck, cartilago Meckeli; c-sbmx, cartilago submaxillaris; cb-I, ceratobranchiale I; cb-II,ceratobranchiale II; cb-IV, ceratobranchiale IV; cb-V, ceratobranchiale V; cd, chorda dorsalis; ch,ceratohyale; cm-sphsep, commissura spheno-septalis; cp-a, copula anterior; eb-I, epibranchiale I;eb-IV, epibranchiale IV; fn-hyp, fenestra hypophysea; fn-sph, fenestra sphenoidea; fr-I, foramenfila olfactoria; fr-tr-hm, foramen truncus hyomandibularis nervus facialis; fr-X, foramen nervusvagus; fs-car-i, fissura arteria carotis interna; hb-I, hypobranchiale I; hb-III, hypobranchiale III;hh, hypohyale, hs, hyosymplecticum; ih, interhyale; ipb-III, infrapharyngobranchiale III; ipb-IV,infrapharyngobranchiale IV; lm-on, lamina orbitonasalis, sensu latu; ot-cap, otic capsule; p-q, parsquadrata of the palatoquadratum; pal, palatinum; pl-oc, pila occipitalis; pns-ep, pons epiphysialis;porb-b, preorbital base; prc-co, processus coronoideus; prc-op, processus opercularis; prc-pt,processus pterygoideus of the palatoquadratum; prc-ptot, processus postoticalis; prc-ra, processusretroarticularis; sol-n, solum nasi; tn-m-p, taenia marginalis posterior; tr-cr, trabecula cranii.

1232 . .

(dorsally and laterally), the solum nasi (ventrally) and the preorbital base(medially), is penetrated by the orbito-nasal artery. The medial processes of thetaeniae marginales posteriores have become extended, but have not reached eachother yet. The otic cartilages have become extended medially as well, whereasthey fused with the latero-dorsal part of the pilae occipitales. Consequently, theforamen for the vagus nerve is bordered completely. Left and right pilae remainseparated from each other. Postero-laterally, the otic capsule bears a distinctprocess, here referred to as the postotic process.

Splanchnocranium

The palatine has now developed a double-headed rostral tip, for the articula-tion with the maxillary bone (Adriaens & Verraes, 1997). This double-headedrostral tip does not show a double articulation with the maxillary bone, which isthe case in Diplomystidae (Arratia, 1987), but forms a slit into which fits andarticulates the latter bone. Meckel’s cartilage has hardly changed, whereas thepterygoid process has enlarged. The quadrate has become extended, comparedto the previous stage. This trend is noted in the later stages as well. Theceratohyal is still continuous with the hyosymplectic through the interhyal.Rostrally, left and right hypohyals have become separated from each other. Theincision for the hyoid artery can be distinguished clearly now, as well as aposterior process.The branchial basket is now almost completely formed. Only the fifth

basibranchial element seems to be missing. The fourth basibranchial lies in frontof the base of the fourth branchial arch. The hypobranchial cartilage ofbranchial arches I to III have become isolated now from the ceratobranchials,thus articulating with them from this stage on. Hypobranchial IV cannot bedistinguished yet, but as can be derived from ossification pattern, it is likely thatits cartilage does not become separated from that of the ceratobranchial IVduring ontogeny. No sign of hypobranchial V is apparent.

7·1-MM SL STAGE (Fig. 6)

Neurocranium

The cartilage between the precerebral processes becomes elevated, thus form-ing a transverse precerebral lamina. The olfactory foramen becomes smaller asthe preorbital base becomes expanded in an antero-posterior direction [Fig. 8(c)].The previously cylindrical orbito-nasal lamina has become plate-like now,transversely positioned on the solum nasi. It is pierced by the ophthalmicforamen, dorsally, and the orbito-nasal foramen, ventrally. Between theseforamina, the nasal barbel originates. The epiphysial bridge is now completed,as the left and right processes have fused. Consequently this narrow bridgedivides the large opening in the skull roof into prepineal and postpinealfontanella. The medial incision of the trabecular bar for the internal carotidartery becomes extended laterally, creating a rather narrow connection in theskull floor. Few changes have occurred in the shape of the otic capsules, exceptfor the occipital arches which have become extended. The left and right pilaeoccipitales have now fused above the neural tube, forming the tectum posteriusand thus enclosing the foramen magnum. The posterior position of this left and

1233

right fusion suggests that it is the pilae occipitales which have fused, and not theotic capsules, which would result in a tectum synoticum. During furtherontogeny, the otic capsules do not fuse with each other, when they becomereinforced by several ossifications. Posterior to the tectum posterius, thesupraneurals situated in front of the second and third vertebrae have fused,forming a solid cartilaginous plate overlying the basidorsals 2 and 3(Radermaker et al., 1989). The foramina of the glossopharyngeus and vagusnerves have decreased their diameter, whereas the anterior basicapsular fenestrahas disappeared. This must occur somewhere between 6·9 and 7·1 mm SL [Fig.9(c) and (d)].

fn-sph(a) fs-car-i fn-hyp

sol-n

lm-on

porb-b prc-ptorb

tt-p

0.2 mm

ot-cap

cm-sphsep

lm-pc

c-eth

pns-ep

c-ac

pl-oc

fr-ophth

tr-cr

sn

astsaglap

cb-V

fr-mcp-p

ipb-IV

fr-X

cm-sphsepfr-I

paltr-cr

porb-b pns-ep sntn-m-p fn-sph hs

0.2 mm

c-Meck ch prc-pt p-q ih

0.2 mmprc-unihhb-III

c-Meck

hh

hb-I

cb-I

cp-ac-sbmx

tr-cr

bd-v2

prc-op

eb-IVcb-IV

tn-m-p fn-sph prc-ptot

p-sc-c

v-sc-c

fr-mfr-X

fr-IX

a-sc-c0.2 mm

fs-car-ipalsol-n

c-eth

cn-peth

fn-hyp

c-sbmx

0.2 mmp-q prc-op

eb-IV

prc-un

ipb-IV

cp-p

ipb-IIipb-I

c-Meck

hh

hb-I

cb-I

cp-a

ch

prc-pt iph-IIIhsprc-op

(d) (e)

(b)

(c)

prc-ptotprc-opch eb-I

F. 6. Chondrocranium of Clarias gariepinus (7·1 mm SL): (a) dorsal view; (b) ventral view; (c) lateralview; (d) dorsal view of splanchnocranium; (e) ventral view of neurocranium. a-sc-c, Anteriorsemicircular canal; ast, asteriscus; bd-v2, basidorsale of vertebra 2; c-ac, cartilago acrochordalis;c-eth, cartilago ethmoideum; c-Meck, cartilago Meckeli; c-sbmx, cartilago submaxillaris; cb-I,ceratobranchiale I; cb-IV, ceratobranchiale IV; cb-V, ceratobranchiale V; ch, ceratohyale; cm-sphsep, commissura spheno-septalis; cn-peth, cornua prae-ethmoidea; cp-a, copula anterior; cp-p,copula posterior; eb-I, epibranchiale I; eb-IV, epibranchiale IV; fn-hyp, fenestra hypophysea;fn-sph, fenestra sphenoidea; fr-I, foramen fila olfactoria; fr-IX, foramen nervus glossopharyngeus(fenestra basicapsularis posterior); fr-m, foramen magnum; fr-ophth, foramen ramus ophthalmicusnervus trigeminus; fr-X, foramen nervus vagus; fs-car-i, fissura arteria carotis interna; hb-I,hypobranchiale I; hb-III, hypobranchiale III; hh, hypohyale; hs, hyosymplecticum; ih, interhyale;ipb-I, infrapharyngobranchiale I; ipb-II, infrapharyngobranchiale II; ipb-III, infrapharyngo-branchiale III; ipb-IV, infrapharyngobranchiale IV; lap, lapillus (utricular otolith); lm-on, laminaorbitonasalis, sensu latu; lm-pc, lamina praecerebralis; ot-cap, otic capsule; p-q, pars quadrata ofthe palatoquadratum; p-sc-c, posterior semicircular canal; pal, palatinum; pl-oc, pila occipitalis;pns-ep, pons epiphysialis; porb-b, preorbital base; prc-co, processus coronoideus of Meckel’scartilage; prc-op, processus opercularis; prc-pt, processus pterygoideus of the palatoquadratum;prc-ptorb, processus postorbitalis; prc-ptot, processus postoticalis; prc-un, processus uncinatusof the epibranchiale III; sag, sagitta (saccular otolith); sn, supraneurale; sol-n, solum nasi;tn-m-p, taenia marginalis posterior; tr-cr, trabecula cranii; tt-p, tectum posterius; v-sc-c, verticalsemicircular canal.

1234 . .

Splanchnocranium

Meckel’s cartilage, which has now developed a dorso-lateral coronoid process,is no longer fused to the contralateral one. The retroarticular process hasbecome even more elongated posteriorly. The pterygoid process is well devel-oped, as it grows in the direction of the palatine; however, neither contact oneanother. The quadrate has a well-developed articulatory facet for the articu-lation with the mandible. Still no distinction can be made between thepterygoquadrate and the hyosymplectic. The hypohyals can be distinguishedeven more from the ceratohyals as the rostral incision and the caudal process aremore pronounced.The anterior copula, which is the result of fusion between the anterior three

basibranchials, now fits with its anterior tip between the caudal processes ofthe hypohyals. The fourth basibranchial, which was observed in the previousstage, is now apparently elongated, its lateral margin abutting against the baseof the fourth branchial arch, its anterior margin lying in front of the fifthceratobranchials. Apparently the fifth basibranchial has formed and is fusedto the fourth, thus forming the copula posterior. The hypobranchials of thefirst and second branchial arches are situated laterally to the copula, articulat-ing with it. The third hypobranchials articulate with the posterior tip of theanterior copula, and also contact each other. The first signs of the fourthhypobranchials are evident, as the medial tip of the ceratobranchial isbroadened and bears a small incision. The third epibranchial bears a caudallydirected uncinate process. The fourth epibranchial is still more stout than theother ones, possibly acting as the support of the fourth infrapharyngo-branchial, which is also more substantial.

10·0-MM SL STAGE (Fig. 7)

Neurocranium

At this stage, some regions in the cartilaginous skull have become more solid.The ethmoid plate has become extended posteriorly, whereas the olfactoryforamen and the sphenoid fenestra have become smaller, due to the expansion ofthe preorbital base. This base now runs from well in advance of the orbito-nasallamina up to about halfway to the eye. No preorbital foramen was observedpenetrating the preorbital base. The spheno-septal commissures are more solid,as well as the precerebral lamina interconnecting the left and right commissures.The epiphysial bridge has also thickened, whereas a median posterior wideningis noted, corresponding to the taenia tecti medialis posterior. In the skull floor,the trabecular bars have become even narrower at the level of the fissure for theinternal carotid artery. This weakening is, however, compensated by thebroadening of the skull roof at that level, as well as osteological reinforcements.The posterior taeniae marginales have broadened as well, as they become moreconfluent with the otic capsules. The broadening of the taeniae is coupled to theformation of a foramen, penetrated by the ramus oticus of the nervus facialis(Surlemont, 1983). Different stages showed that this foramen arises from theinitial formation of a medial incision in the taenia marginalis (7·1 mm SL; Fig.6), which later gets closed off medially by a minute cartilaginous strip [between7·4 and 7·7 mm SL; Fig. 8(d)]. No commissura lateralis, nor a pila lateralis, is

1235

formed. The otic capsules still have not fused with each other, whereas thetectum posterius now seems to have fused with the previously mentionedsupraneurals.

Splanchnocranium

No significant changes have occurred, compared to the previous stage. Theforamen for the truncus hyomandibularis has enlarged, whereas the opercularprocess is more pronounced. The interhyal is still continuous with both theceratohyal and the hyosymplectic, although a constriction can be seen at thefusion with the ceratohyal. Later during ontogeny, in the 46·8-mm SL stage, this

pns-ep(a) tr-cr tn-m-p

porb-b

lm-on

fn-sph

tt-p

0.2 mm

ot-cap

cm-sphsep

lm-pc

c-eth

sol-n

fr-ot

c-ac

pal

fs-car-i

sncp-p

ipb-IV

fr-IX

lm-on

c-sbmx

snfn-sph fr-ot

0.2 mm

prc-co prc-pt p-q ih

0.2 mm

ot-capcp-a

c-Meck

hh

pal

ch

c-ethcn-peth

porb-b

prc-op

ipb-IV

fn-sph v-sc-ca-sc-cp-sc-c

cd

fr-m

prc-ptot

pl-oc

fr-IX0.2 mm

fn-hyppal tr-cr

c-eth

cn-peth

sol-n

c-sbmx

0.2 mmfr-tr-hm

cb-V

prc-un

ipb-IV

ipb-III

c-Meck

hh

hb-Icb-I

cp-a

ch

prc-pt hsp-q prc-op

(d) (e)

(b)

(c)

cb-I eb-Ifr-IX

ipb-Ieb-I

prc-co

hb-IV

cb-II hb-III eb-III cb-V

fs-car-i

tn-m-p

pns-ep

tn-t-m-p

hb-I

F. 7. Chondrocranium of Clarias gariepinus (10·0 mm SL): (a) dorsal view; (b) ventral view; (c) lateralview; (d) dorsal view of splanchnocranium; (e) ventral view of neurocranium. a-sc-c, Anteriorsemicircular canal; c-ac, cartilago acrochordalis; c-eth, cartilago ethmoideum; c-Meck, cartilagoMeckeli; c-sbmx, cartilago submaxillaris; cb-I, ceratobranchiale I; cb-II, ceratobranchiale II; cb-V,ceratobranchiale V; cd, chorda dorsalis; ch, ceratohyale; cm-sphsep, commissura spheno-septalis;cn-peth, cornua prae-ethmoidea; cp-a, copula anterior; cp-p, copula posterior; eb-I, epibranchialeI; eb-III, epibranchiale III; fn-hyp, fenestra hypophysea; fn-sph, fenestra sphenoidea; fr-tr-hm,foramen truncus hyomandibularis nervus facialis; fr-IX, foramen nervus glossopharyngeus(fenestra basicapsularis posterior); fr-m, foramen magnum; fr-ot, foramen ramus oticus nervusfacialis; fs-car-i, fissura arteria carotis interna; hb-I, hypobranchiale I; hb-III, hypobranchiale III;hb-IV, hypobranchiale IV; hh, hypohyale; hs, hyosymplecticum; ih, interhyale; ipb-I,infrapharyngobranchiale; ipb-III, infrapharyngobranchiale III; ipb-IV, infrapharyngobranchialeIV; lm-on, lamina orbitonasalis, sensu latu; lm-pc, lamina praecerebralis; ot-cap, otic capsule; p-q,pars quadrata of the palatoquadratum; p-sc-c, posterior semicircular canal; pal, palatinum; pl-oc,pila occipitalis; pns-ep, pons epiphysialis; porb-b, preorbital base; prc-co, processus coronoideusof Meckel’s cartilage; prc-op, processus opercularis; prc-pt, processus pterygoideus of thepalatoquadratum; prc-ptot, processus postoticalis; prc-un, processus uncinatus of the epibranchialeIII; sn, supraneurale; sol-n, solum nasi; tn-m-p, taenia marginalis posterior; tn-t-m-p, taenia tectimedialis posterior; tr-cr, trabecula cranii; tt-p, tectum posterius; v-sc-c, vertical semicircular canal.

1236 . .

pns-ep(a)

tn-m-p

lm-on(s)

cm-spheth

mx-b c-eth

sol-npal

fr-ophth

prc-pc

cm-sphsep

cm-bc-a

pns-ep

fs-sph

c-eth

sol-npal

ns-b

prc-pc

cm-sphsepcm-bc-a

tn-m-pc-ot-a

fs-car-i

pc-pl

mx-b

(b)

mx-b porb-b

pal

fr-ophth

prc-pccm-sphsep

fs-sph

sol-n

lm-on

ns-b

prc-pc

cm-sphsep

tr-cr

tn-m-p

c-ot-a

fr-on

sol-n

pal

fr-ophth

(c)

lm-on

mx-b

c-sbmx

pal

fr-ophth

cm-sphsep

fs-sph fr-on

sol-n

lm-pc

ns-b fr-I

cm-sphsep

tn-m-p

fr-on

sol-n

fr-ophthc-ot-a

pns-ep

tr-cr

pns-eptn-m-p

c-sbmx

lm-onporb-btr-cr

(d)

lm-on

fr-on

pal

0.1 mm

ot-cap

lm-on

sol-nfr-I

fr-I

cm-sphsep

cm-bc-afn-sph

fr-ot

fr-ot

sol-n

fr-ophth

c-ot-a

0.1 mm

pns-eptn-m-p

c-sbmx

porb-btr-crfn-sph

pns-ep

mx-b

tn-m-p fr-ophth

pal

0.1 mm 0.1 mm

0.1 mm0.1 mm

0.1 mm0.1 mm

lm-on

F. 8. Ontogeny of ethmo-orbital region (right side) of Clarias gariepinus (left: frontal view, right: lateralview): (a) (5·8 mm SL); (b) (6·0 mm SL); (c) (7·1 mm SL); (d) (9·5 mm SL). c-eth, Cartilagoethmoideum; c-ot-a, cartilago oticalis anterior; c-sbmx, cartilago submaxillaris; cm-bc-a, commis-sura basicapsularis anterior; cm-spheth, commissura spheno-ethmoidalis; cm-sphsep, commissuraspheno-septalis; fn-sph, fenestra sphenoidea; fr-I, foramen fila olfactoria; fr-on, foramen orbito-nasalis; fr-ophth, foramen ramus ophthalmicus nervus trigeminus; fr-ot, foramen ramus oticusnervus facialis; fs-car-i, fissura arteria carotis interna; fs-sph, fissura sphenoidea; lm-on, laminaorbitonasalis, sensu latu; lm-on(s), lamina orbitonasalis, sensu strictu; lm-pc, lamina praecerebralis;ot-cap, otic capsule; mx-b, maxillary barbel; ns-b, nasal barbel; pal, palatinum; pc-pl, parachordalplate; pns-ep, pons epiphysialis; porb-b, preorbital base; prc-cp, processus praecerebralis; sol-n,solum nasi; tn-m-p, taenia marginalis posterior; tr-cr, trabecula cranii.

1237

(c)

tn-m-p

fr-IX

cm-bc-p

porb-b

tt-p

fr-X

fs-car-i

fn-sph

tr-cr

cm-bc-a

cm-bv

(d)

sn0.2 mm

0.2 mm

tn-m-p

fr-IX

cm-bc-p

fr-X

fn-bc-a

fn-sph

tr-cr

cm-bc-a

cm-bv

(a)

0.2 mm

fs-mot

tr-cr

cm-bc-a

tn-m-p

fr-IX

cm-bc-p

pl-oc

fr-X

fn-sph

tr-cr

cm-bc-a

cm-bv

fn-bc-a

(b)

0.2 mm

F. 9. Ontogeny of the otico-occipital region (left side, ventral view) of Clarias gariepinus: (a) 4·1 mm SL;(b) 6·0 mm SL; (c) 6·9 mm SL; (d) 9·5 mm SL. cm-bc-a, Commissura basicapsularis anterior;cm-bc-p, commissura basicapsularis posterior; cm-bv, commissura basivestibularis; fn-bc-a,fenestra basicapsularis anterior; fn-sph, fenestra sphenoidea, fr-IX, foramen nervus glos-sopharyngeus (fenestra basicapsularis posterior); fr-X, foramen nervus vagus; fs-car-i, fissuraarteria carotis interna; fs-mot, fissura metotica; fs-sph, fissura sphenoidea; pl-oc, pila occipitalis;porb-b, preorbital base; sn, supraneurale; tn-m-p, taenia marginalis posterior; tr-cr, trabeculacranii; tt-p, tectum posterius.

1238 . .

fusion is lost, whereas the former is still continuous with the hyosymplectic(Adriaens & Verraes, 1994).At its median tip, the fourth ceratobranchial element has become differenti-

ated, as it has become broader and has developed a small caudal process. Thispart most probably corresponds to the fourth hypobranchial, which remainscontinuous with the cartilage of the ceratobranchial.

19·0-MM SL STAGE (Fig. 10)This stage is mentioned only to give the relation between the fully developed

chondrocranium and the initial osteocranium. However, some changes haveoccurred in the chrondrocranium, compared to the previous situation. Theethmoid plate has become extended even further posteriorly. The foramen forthe internal carotid artery now has cut its way completely through the trabecularbars, resulting in the confluence of the hypophyseal fenestra and the sphenoidfenestra. At this stage, this loss in reinforcement is compensated by theossifications of the skull floor and skull roof. Each of the cartilaginous parts ofthe neurocranium has become enclosed, covered or partially replaced by bones,with exceptions of some smaller regions between the bones.

DISCUSSION

Generally published data suggest that the chondrification of the neurocraniumstarts simultaneously with that of the splanchnocranium (de Beer, 1937).However, asynchronies have been observed where the neurocranial skeletonarises prior to the visceral one [e.g. Catostomus commersoni (Lacépède)(Cypriniformes, Catostomidae) (McElman & Balon, 1980); and Heteropneustesfossilis (Bloch) (Siluroidei, Heteropneustidae) (Srinivasachar, 1959), or later as inHeterobranchus longifilis Valenciennes] (Siluroidei, Clariidae) (Vandewalle et al.,1997). In the latter species, primordia of the mandibular and hyoid arch areformed prior to neurocranial elements. In Clarias gariepinus both neurocraniumand splanchnocranium were observed in the first stage, although the situationmay be comparable to that in the closely related H. longifilis.It has to be noted that when comparing with data from the ontogenetic study

of C. gariepinus made by Surlemont et al. (1989) and Surlemont & Vandewalle,1991), some ontogenetical evidence suggests that the 5·2-mm TL stage ofSurlemont et al. (1989) is more developed than the presently observed 5·6-mm SL(5·9-mm TL) specimen.

NEUROCRANIUMPlatybasic skullAs already mentioned, the adaptation of most catfishes to a benthic and

nocturnal lifestyle is reflected in several structural transformations. However,these are not restricted to the fully developed, bony skull, but arise early duringontogeny. When, for example, compared to the closely related Characiformes[according to the phylogeny proposed by Fink & Fink (1981)], the general trendin overall siluroid head morphology seems to involve a reduction of eye size, aswell as the dorso-ventral flattening of the skull. To a certain degree, this trendcan be expressed as the ratio of the eye diameter to the interorbital distance.

1239

In Characiformes this ratio may range from 0·3 (in Citharinops distichodus(Pellegrin), Distichodontidae) (Roman, 1966) up to 2·2 [in Hepsetus odoe(Bloch)] (Daget, 1962), whereas the average value is about 1·0 (based onbiometrical data from literature of 29 species of four families) (Boulenger,1909–1916; Giltay, 1929; Poll, 1945, 1954, 1967a, b; Daget, 1954; Poll & Gosse,1963; Roman, 1966; Poll & Daget, 1968; Mahnert & Géry, 1987; Teugels & Thysvan den Audenaerde, 1990). In Siluroids the ratio may range from 0·1 (in

(a)

tn-m-p

fr-IX

lm-on

prc-ptorb

sol-n

1 mm

ot-cap

cm-sphsep

c-eth

lm-pc

pns-ep

c-eth

cn-peth

porb-b

tt-p

fr-X

fn-hyp tn-m-p ot-cap

tr-cr fn-sph tr-cr

(b)

1 mm

F. 10. Chondrocranium of Clarias gariepinus in relation to the osteocranium (19·0 mm SL): (a) dorsalview; (b) ventral view (hatched areas indicate foramina and fenestra, lightly shaded areas indicatechondrocranium covered with bone, darkly shaded areas indicate superficially exposedchondrocranium. c-eth, Cartilago ethmoideum; cm-sphsep, commissura spheno-septalis; cn-peth,cornua prae-ethmoidea; fn-hyp, fenestra hypophysea; fn-sph, fenestra sphenoidea; fr-IX, foramennervus glossopharyngeus (fenestra basicapsularis posterior); fr-X, foramen nervus vagus; lm-on,lamina orbitonasalis, sensu latu; lm-pc, lamina praecerebralis; ot-cap, otic capsule; pns-ep, ponsepiphysialis; porb-b, preorbital base; prc-ptorb, processus postorbitalis; sol-n, solum nasi; tn-m-p,taenia marginalis posterior; tr-cr, trabecula cranii; tt-p, tectum posterius.

1240 . .

Dinotopterus cunningtoni Boulenger, Clariidae) (Teugels, 1986) up to 1·2 (inCheirocerus eques Eigenmann, 1917, Pimelodidae) (Stewart & Pavlik, 1985) withan average value of approximately 0·4 (based on biometrical data from literatureof 58 species of 17 families) (Steindachner, 1914; Boulenger, 1915; Pellegrin,1928; Schultz, 1942; Crass, 1960; Blache, 1964; Thys van den Audenaerde, 1964;Myers & Weitzman, 1954, 1966; Alfred, 1966; Roman, 1966; Poll et al., 1972;Glodek, 1976; Thys van den Audenaerde & De Vos, 1982; De Vos & Leveque,1983; Nijssen & Isbrücker, 1983, 1987, 1990; Stewart, 1985, 1986; Stewart &Pavlik, 1985; Teugels, 1986; Vari & Ortega, 1986; Risch, 1987; Reis & Schaefer,1992; Lucena et al., 1992). Within the siluroids, this trend is even more extremein Clariidae (Teugels, 1986). As can be expected, the spatial impact of the eyesize on the surrounding structures will be of more importance during the earlydevelopment than in the adult situation. It is observed that in Clarias gariepinusthe eye diameter of a 7·2-mm TL specimen is approximately 14·3% of the headlength (Surlemont, 1983) whereas in adults of 685 mm TL (600 mm SL) itreaches only 6·1% (Skelton & Teugels, 1992). As has been demonstrated byCorsin (1961), the development of the trabecula communis in tropibasic skulls isinfluenced by the spatial interactions with the developing eyes. When the eyesfail to develop properly, it is noted that in salmons the trabecular bars do notfuse with each other, and they also influence the growth of the taenia marginalis(Verraes, 1974). The reduced eye size in siluroids thus explains the absence ofa pronounced trabecular communis, and consequently the formation of aplatybasic skull.

Skull floor

The neurocranial ontogenetic sequence generally starts with the parachordalcartilages on either side of the notochord, followed by the chondrification of thetrabecular bars. Both fuse early in ontogeny, thus forming the initialneurocranial floor from the interorbital region up to the otic and occipital region.In general, the distinction can be made between the trabecula s.s. and the polarcartilages* (Srinivasachar, 1959). However, the position of the incision, intowhich passes the internal carotid artery, may indicate the anterior margin of thecartilago polaris (Goodrich, 1958; Bertmar, 1959). In relation to the platybasicskull configuration, the contralateral trabecular bars (trabecular cranii) remainseparated from each other by a broad fenestra hypophysea. This requires abroad ethmoid plate or medially curved trabecular bars. In most catfishes theethmoid plate is broad, although the trabecular bars seem to be slightly curved inas well, e.g. Clarias gariepinus (Figs 2 and 3). In Schilbeidae, which havelaterally compressed skulls, the ethmoid is narrower, with the trabecular barscurved medially (Srinivasachar, 1957a). In general the trabecular bars fuse withthe caudal border of the ethmoid plate, although in callichthyids it may fuse withits dorsal face (Hoedeman, 1960). The trabecular bars become broadened as wellas in both a lateral (for the support of the nasal sac), and a medial direction (for*Concerning the anatomical nomenclature of the ‘ polar cartilage ’, no consistent scientific name has

been introduced yet. In general these structures are referred to in a native language, as for example ‘ polarcartilage ’ (de Beer, 1937; Goodrich, 1958; Jarvik, 1980), ‘ cartilages polaires ’ (Devillers, 1958; Daget,1964) or ‘ Polknorpel ’ (Marinelli, 1936). In order to introduce a single, consistent and scientificterminology to avoid confusion, the term ‘ cartilago polaris ’ is used.

1241

the support of the brain). The lateral expansion results in a solum nasi, whichmay be very much pronounced (e.g. Silonia, Schilbeidae) (Srinivasachar, 1957a).It is well developed in clariids as well (Fig. 7) (Srinivasachar, 1959). The medialexpansion results in a fusion between the anterior part of the left and righttrabecular bars. Consequently, the platybasic siluroids do have, although onlyslightly, a trabecular communis (Srinivasachar, 1958a). In many siluroids,rostro-lateral processes of the ethmoid plate are formed. These preethmoidcornua are distinct in Ariidae (Srinivasachar, 1958a), whereas Plotosidae(Srinivasachar, 1958a) and Schilbeidae (Srinivasachar, 1957a) do not have them.In Clariidae and the closely related Heteropneustidae, as well as in Pangasiidae,the preethmoid cornua are rudimentary (Fig. 7) (Srinivasachar, 1957a, 1959).As a consequence of the broad ethmoid plate, a transversal precerebral lamina

is formed instead of a longitudinal septum internasale [Fig. 7(a)]. This, togetherwith the posterior position of the lamina orbito-nasalis, enables the fila olfactoriato lie in a transverse direction of the olfactory lobes, instead of an antero-posterior one. This implies that they can be rather short as the lobi olfactoriicome to lie medial to the nasal sacs, instead of postero-medially (Surlemont,1983). In Ictalurus nebulosus (Lesueur) it is noted that ‘ the olfactory lobeprotrudes through this foramen ’, the latter referring to the olfactory foramen(Kindred, 1919). In some cases, the precerebral lamina becomes extendedposteriorly in a ‘ nasal septum ’ (e.g. Arius jella Day and Ailia coila) (Hamilton,1822) (Srinivasachar, 1957a, 1958a). According to Daget (1964), this precerebrallamina and the internasal septum are homologous.

Skull roofGenerally, the chondrocranial roof in fishes is formed by the taenia marginalis

(frequently referred to as the ‘ orbital cartilages ’). They arise as separatecartilaginous elements which become elongated in an anterior and posteriordirection (de Beer, 1937; Daget, 1964). The taenia marginalis can be subdividedinto a taenia marginalis anterior (anterior to the epiphysial bridge) and a taeniamarginalis posterior (posterior to the epiphysial bridge). The anterior onenormally bifurcates at its rostral tip: an anteriorly directed commissura spheno-septalis and a laterally directed commissura spheno-ethmoidalis are consequentlyformed (Bertmar, 1959; Daget, 1964). In siluroids, however, the taeniamarginalis posterior seems to arise as a rostral extension of the anterior oticcartilage, at the processus postorbitalis, instead of being isolated from it[Fig. 2(a)] (399: Srinivasachar, 1959; 76: Hoedeman, 1960). The taeniamarginalis anterior becomes reduced or almost completely absent as in mostsiluroids the epiphysial bridge lies at the same level as the lamina orbito-nasalis.A distinct taenia marginalis anterior is present in Arius (Ariidae) and Plotosus(Plotosidae) (Srinivasachar, 1958a) but is short in Silonia and Ailia (Schilbeidae)(Srinivasachar, 1957a). In species like Heteropneustes (Heteropneustidae)(Srinivasachar, 1959), Pangasius (Pangasiidae) (Srinivasachar, 1957a), Ictalurus(Ictaluridae) (Kindred, 1919), Callichthys (Callichthyidae) and Clarias (Fig. 5)(Srinivasachar, 1959) the taenia marginalis anterior is completely reduced,whereas its bifurcation is situated at the level of the epiphysial bridge [Fig. 8(a)].In siluroids the spheno-septal and the spheno-ethmoidal commissures fuse withtwo dorsal processes of the chondrocranial floor: (1) the precerebral process of

1242 . .

the ethmoid plate; and (2) the lamina orbito-nasalis, s.s., of the solum nasi,respectively (Fig. 8). The latter fusion generally precedes the former one duringontogeny. The spheno-septal commissure then becomes elongated rostrally andfuses with a small precerebral process. Surprisingly, in H. longifilis this commis-sure fails to develop whereas it is the precerebral process that becomes elongatedin a caudal direction, until it fuses with the taenia marginalis (Vandewalle et al.,1997).A pons epiphysialis is present in most siluroids, although in Heteropneustes

fossilis it fails to develop (Srinivasachar, 1959). A small taenia tecti medialisposterior is present in Clarias gariepinus (Fig. 7), Arius jella, Plotosus caniusHamilton (Srinivasachar, 1958a), Ailia coila, Silonia silondia (Hamilton, 1822)and Pangasius pangasius (Hamilton, 1822) (Srinivasachar, 1957a), whereas it isnot distinct in Ictalurus nebulosus (Kindred, 1919), Callichthys callichthys L.(Hoedeman, 1960) and H. longifilis (Vandewalle et al., 1997).

Skull wallAt the same level of the lamina orbito-nasalis, a preorbital base is generally

present in siluroids. This structure arises as a ventral process of the taeniamarginalis which eventually fuses with the underlying trabecula cranii. InClarias gariepinus, however, a small process is found on the trabecula craniiwhich suggests that the preorbital base arises on the trabecula cranii in thisprocess [Fig. 8(b)]. In siluroids this base becomes broad, thus separating theolfactory and sphenoid foramenina, and may be perforated by a preorbitalforamen (‘ preoptic foramen ’) [e.g. Pangasiidae (Srinivasachar, 1957a), Ariidae,Plotosidae (Srinivasachar, 1958a), Ictaluridae (Kindred, 1919)]. Such a foramenis absent in Clariidae [Fig. 8(d)] (Srinivasachar, 1959; Vandewalle et al., 1997),Heteropneustidae (Srinivasachar, 1959) and Schilbeidae (Srinivasachar, 1957a).

Otic capsulesAlthough some changes in position and shape of the structures involved are

observed, the otic region is comparable in most siluroids. According to Daget(1964) the otic capsule arises as two consecutive elements (cartilago oticalisanterior and posterior). The anterior one becomes connected to the cartilagoparachordalis through a commissura basicapsularis anterior, whereas the pos-terior one is attached through a commissura basivestibularis (rostrally) and acommissura basicapsularis posterior (caudally). The identification of thesestructures is based on their relation with the nervus glossopharyngeus and thenervus vagus. In Clarias gariepinus, and in siluroids in general, the primordia ofthe otic region comprises a cartilago oticalis anterior, connected to the cartilagoparachordalis through a commissura basicapsularis anterior [Fig. 9(a)](Srinivasachar, 1957a, 1959). Posterior to this commissure the glossopharyngeusand vagus nerves pass through a fissura metotica. At the 6·0-mm SL stage twoforamina are present, penetrating the floor of the otic capsules [Fig. 9(b)]. Serialsections (of a 5·9-mm SL stage) revealed that the glossopharyngeus nerve passesthrough the posterior one. The vagus nerve, on the other hand, lies caudally ina fissure bordered by the pila occipitalis (medially) and the cartilago oticalisposterior (laterally). Eventually, this fissure becomes closed as the pila occipitalisand the otic capsule fuse dorsally. The anterior foramen becomes reduced and

1243

eventually disappears [Fig. 9(c) and (d)]. In a generalized teleostean situation,the foramina of these two cranial nerves are separated from each other by thecommissura basicapsularis posterior, whereas the glossopharyngeus nerve passesthrough the fenestra basicapsularis posterior. Consequently, in the case ofC. gariepinus the cartilaginous bar between the two foramina in the otic floor,corresponding to the fenestra basicapsularis anterior and posterior, representsthe commissura basivestibularis [Fig. 9(b)]. In the fully developed chondro-cranium, the posterior basicapsular fenestra (foramen nervus glossopharyngeus)and the foramen for the vagus nerve remain present [Fig. 9(d)]. A comparablesituation can be found in Heteropneustes fossilis (Srinivasachar, 1959) andPlotosus canius (Srinivasachar, 1958a). In Arius jella, the fenestra basicapsularisanterior remains present and is rather large, whereas the glossopharyngeus andvagus nerves still pass through the fissura metotica (Srinivasachar, 1958a). Thesame construction is found in the schilbeid Ailia coila, but foramina are absent inSilonia silondia. It seems more plausible that in the latter species all threecommissures have fused instead of that ‘ there are no basicapsular fenestrae andcommissures ’ (Srinivasachar, 1957a).In the occipital region, a fusion between the left and right part of the brain-

case is present in siluroids. This fusion can be spread over the otic capsules andthe occipital arches. However, in several catfishes only the pilae occipitales seemto fuse, thus forming a tectum posterius. The fusion between the otic cartilages,resulting in a tectum synoticum is absent (e.g. Clariidae, Callichthyidae)[Fig. 7(a)] (Hoedeman, 1960) or greatly reduced (e.g. Ariidae, Plotosidae,Pangasiidae) (Srinivasachar, 1957a, 1958a). In Ailia coila it is rather pronounced(Srinivasachar, 1957a). In many cases, however, the distinction between the twotecta is rather difficult to determine. An additional fusion observed in the nuchalarea of C. gariepinus occurs between this tectum posterius and the firstsupraneural cartilages [Figs 6(a) and 7(a)] (Radermaker et al., 1989). A closecontact between the third neural arch and the tectum posterius is observed inIctalurus nebulosus (Ictaluridae) (Kindred, 1919). Based on the relation betweenthe supraneurals and the basidorsals in C. gariepinus, it seems that supraneuralsof vertebrae two and three are involved [Fig. 6(c)]. This might explain theabsence of the anterior three supraneurals in adult Characiphysi, which isconsidered a synapomorphic feature by Fink & Fink (1981; characters 58, 59,61). Apparently, only the anterior one is absent (synapomorphy of Ostariophysi)(Fink & Fink, 1981; character 58), whereas supraneural two and three have fusedwith the neurocranium.

SPLANCHNOCRANIUMPremandibular archA synapomorphic feature of Siluroidei is the discontinuity of the palatine with

the pterygoquadrate (Fink & Fink, 1981; Arratia & Schultze, 1990; Arratia,1992). This decoupling has enabled the development of a highly specializedpalatine–maxillary mechanism for the controlled movements of the maxillarybarbel (Alexander, 1965; Gosline, 1975; Adriaens & Verraes, 1997). Exception-ally, a secondary fusion between the palatine and the pterygoquadrate is presentin Ariidae (380: Bamford, 1948; 993: Srinivasachar, 1958a). Srinivasachar(1957a, 1958a, 1959) rather confusingly, referred to the palatine as ‘ pterygoid

1244 . .

process ’, which he distinguished from the ‘ processus pterygoideus ’ of thepterygoquadrate. The palatine, which is considered to be homologous with theepipremandibular element (Daget, 1964; Jarvik, 1980) articulates with the laminaorbito-nasalis by means of an ‘ ethmopalatine process ’ [Fig. 5(B)] (Srinvasachar,1958a). Presumably this articulation corresponds to the ethmopalatine articula-tion of other teleosts. The question could be raised whether this isolation of thepalatine in siluroids is a secondary separation between the premandibular andmandibular arch, which would support the hypothesis that the palatine really isderived from the premandibular arch. As a consequence of the mechanical stressgenerated in the palatine–maxillary mechanism, more specifically in the articu-lation of the maxillary bone with the palatine, a secondary cartilage formation isformed, i.e. the cartilago submaxillaris [Fig. 7(c)].

Mandibular arch

In Siluroids, in general, a single cartilaginous ‘ hyo-symplectic–pterygoquadrate plate ’ is formed, which even may be continuous with Meckel’scartilage (Arratia, 1990). In some siluroids, however, this fusion is a secondaryprocess. In Galeichthys felis Valenciennes (Ariidae) four separate elementsare formed (pterygoquadrate, hyosymplectic, interhyal and ceratohyal withhypohyal) whereas the pterygoquadrate eventually fuses with the hyosymplectic(Bamford, 1948). In other catfishes a trend is observed for additional fusions ofthe ‘ hyo-symplectic–pterygoquadrate plate ’ with the interhyal [e.g. Ictaluridae(Kindred, 1919; Arratia, 1992), Schilbeidae, Pangasiidae (Srinivasachar, 1957a),Heteropneustidae (Srinivasachar, 1959), Callichthyidae (Hoedeman, 1960;Howes & Teugels, 1989), Trichomycteridae (Arratia, 1990)] and even with theneurocranium [e.g. Callichthyidae (Hoedeman, 1960; Howes & Teugels, 1989),Trichomycteridae (Arratia, 1990)]. This trend (i.e. the fusion with the interhyal)is persistent in the Clariidae as well, as is shown in the present study [Fig. 7(c)]as well as in literature (Srinivasachar, 1959; Surlemont et al., 1989; Surlemont &Vandewalle, 1991; Adriaens & Verraes, 1994; Vandewalle et al., 1997). In someof these siluroids the interhyal is continuous with the ceratohyal as well, whereasin others an articulation is found. An interhyal, which is fused initially to thehyosymplectic has been observed in other ostariophysans as well, for exampleChanos chanos (Forsskål) (Gonorhynchiformes, Chanidae) (Arratia, 1990) andHepsetus odoe (Characiformes, Hepsetidae) (Bertmar, 1959).In most siluroids the Meckel’s cartilages of both sides are fused to each

other (Ariidae, Plotosidae, Ictaluridae, Heteropneustidae and Clariidae)(Kindred, 1919; Srinivasachar, 1958a, 1959), which is not the case in Callich-thyidae (Hoedeman, 1960). A processus coronoideus is present, althoughgenerally it is not pronounced. However, in later stages of C. gariepinus, oncecartilage resorption and ossifications have started, the coronoid processbecomes more distinct (Fig. 11). The continuity of Meckel’s cartilage with thepars quadrata in the early stages is believed to play an important role for apassive mouth opening (Surlemont et al., 1989). The retroarticular process ispresent but small in C. gariepinus, which is the general trend in siluroids.However, in some species it may be reduced almost completely (Srinivasachar,1958a).

1245

Hyoid archThe hyosymplectic generally bears a foramen for the passage of the truncus

hyomandibularis of the nervus facialis. This foramen is generally observedin siluroids, although in several species the nerve passes in front of thehyosymplectic, through an anterior slit (Srinivasachar, 1957a, 1958a). Appar-ently the general ontogenic sequence involves a passage of the truncus in frontof the hyosymplectic which becomes enclosed by the anteriorly extendinghyosymplectic cartilage, until a complete foramen is formed. The absence of atrue foramen and the presence of a slit in some species can be considered as aneotenic feature. In Galeichthys felis (Ariidae) the nerve passes initially througha slit at the posterior margin of the hyosymplectic, where it seems to cut its waythrough until it passes in front of the latter (Bamford, 1948). Although thesymplecticum is considered to be absent in siluroids, some evidence suggests thepresence of a rudimentary, non-ossified symplectic region in the suspensorialcartilage. In Diplomystes camposensis Arratia, 1987 (Diplomystidae), whichrepresents the primitive catfish configuration, a rudimentary symplectic bone is

(a)

0.25 mm

(b)

(c)

0.25 mm

0.25 mm

F. 11. Differentiation of the processus coronoideus (arrows) in the mandibula of Clarias gariepinus: (a)12·0 mm SL; (b) 14·8 mm SL; (c) 19·0 mm SL.

1246 . .

present (Arratia, 1992: Fig. 14). The absence of the symplectic bone in themajority of siluroids may be a consequence of the cartilaginous link betweenthe hyosymplectic and the pterygoquadrate (210: Arratia, 1990). In a 21-mm TLstage of Clarias gariepinus, Howes & Teugels (1989) observed a gap where themargin between the pars quadrata and the symplecticum could be expected.Such a gap, however, was not observed in the specimens examined in the presentstudy.The hyoid bars of siluroids are well developed. As already mentioned, the

ceratohyals are continuous with the interhyal in some siluroids, whereas they arecontinuous with the hypohyals in all species. In general the latter bear ananterior groove for the passage of the arteria mandibularis, although a trueforamen is present in Arius jella (Srinivasachar, 1958a). In C. gariepinus thehypohyals are fused to each other initially as well as to the median cartilage of thebranchial basket [Figs 1(c) and 2(c)]. The branchial cartilage becomes separatedfrom the hypohyals, prior to the separation of the contralateral hypohyals [Figs3(c) and 5(c)]. Suggestions have been made that the anterior part of the medianbranchial cartilage, which is in close relation with the hypohyals, would be afusion between the basihyal and the first basibranchial (382: Srinivasachar,1959). However, the position and number of basibranchials indicate that thisanterior part is more likely to correspond to the first basibranchial only (seebelow). On the other hand, a basihyal would be expected anterior to thehypohyals, instead of posterior to them (Nelson, 1969). It is therefore generallyaccepted that the basihyal is absent in siluroids (Arratia & Schultze, 1990).

Branchial archesIn the branchial basket of siluroids, the following structures are generally

distinguished: (1) an anterior and a posterior copula; (2) the hypobranchialsI–III; (3) the ceratobranchials I–V; (4) the epibranchials I–IV. Concerningthe infraphrayngobranchials, some variation is observed (see below). Theontogenetic sequence, observed in siluroids, involves a differentiation fromanterior to posterior. For each arch, the sequence starts with the chondrificationof the ceratobranchials, followed by the hypobranchials and basibranchials,and eventually the epi- and infrapharyngobranchials (Srinivasachar, 1959;Vandewalle et al., 1997).Based on both chondrocranial and osteocranial evidence, it appears that in

siluroids the anterior copula comprises the anterior three basibranchials, whereasonly basibranchial II and III will ossify (Nawar, 1954; Srinivasachar, 1958b;Lundberg, 1982; De Pinna & Vari, 1995). Some authors considered these as theossa basibranchialia I and II, instead of II and III (Teugels et al., 1991; SrinivasaRao & Lakshmi, 1984). However, based on the fact that ‘ for teleostomes,established usage dictates that a basibranchial be given the name or numberof the paired arch-element behind it ’ (480: Nelson, 1969) suggests that thebasibranchial bones represent those of the second and third branchial arch. InDiplomystidae, the first basibranchial element becomes reduced, whereas theossified second and third ones become separated from each other (Arratia, 1987:Fig. 17). In a 127·0-mm SL stage of Clarias gariepinus, however, basibranchialsI–III remain fused to each other, and basibranchial IV remains fused to the fifthone, thus forming an anterior and a posterior copula [Fig. 7(b)]. These copulae

1247

correspond to the mistakenly named basibranchial I and II of H. longifilis[Vandewalle et al., 1997: Fig. 8(b)].The siluroid hypobranchials of the first three branchial arches become

separated from the ceratobranchial elements during ontogeny [Fig. 5(c)](Srinivasachar, 1959; Vandewalle et al., 1997). In most siluroids, the hyobranchi-als I and II articulate with the lateral margin of the anterior copula, whereas thethird one lies between the anterior and posterior copulae. Occasionally, the leftand right hypobranchials III fuse (e.g. Pangasius pangasius, Plotosus canius)(Srinivasachar, 1957a, 1958a). In some other species, they come into closecontact with each other (e.g. C. gariepinus) [Fig. 7(d)] or remain separated as thetwo copulae overlap in that region (e.g. Arius jella) (Srinivasachar, 1958a).Hypobranchial IV has never been observed as a separate element. However, thedifferentiation of the medial tip of ceratobranchial IV and the reduced ossifica-tion of that part suggests that hypobranchial IV is present but remains fused tothe ceratobranchial IV [Fig. 7(b)].Concerning the epibranchial configuration within siluroids, a consistent trend

is present. Epibranchials I–IV are always developed, whereas the fifth one,which frequently occurs in the closely related Characiformes, does not develop(Bertmar, 1959; Daget, 1964). An uncinate process of epibranchial III is presentin C. gariepinus [Fig. 7(d)], although it was not observed in several other siluroidspecies (Srinivasachar, 1957a, 1958a, 1959). Although the form and numberof epibranchials is rather constant in siluroids, the way they are connected tothe infrapharyngobranchials is very variable, as a result of the differentinfrapharyngobranchial constructions.The developmental status of the infrapharyngobranchials in siluroids, on the

other hand, has led to varied nomenclature (Table II, left part). According to theliterature, the number of infrapharyngobranchials may range between two andfour. Processes of reductions and fusions lie at the base of the high degreeof assumed variation. In the primitive diplomystid situation, all four infra-pharyngobranchials are present and even become separately ossified (Nelson,1969: Fig. 5(B); Arratia, 1987: Fig. 17). The anterior two infrapharyngobranchi-als are slender and lie parallel to the epibranchials, whereas the third and fourthones are more solid and lie transversely against the epibranchials. An importantfeature is the presence of a cartilaginous mass interconnecting the medial bases ofthe infrapharyngobranchials I and II with epibranchials I and II, as well as withthe anterior tip of infrapharyngobranchial III. Based on the interpretation ofthis cartilaginous mass, for siluroids two possible interpretations can be pro-posed (Fig. 12): (1) the cartilage is an additional cartilage which becomes reducedin most catfishes but remains present in C. gariepinus; or (2) no additionalcartilage is present in diplomystids, but the cartilaginous mass in questionrepresents the fused bases of the anterior two, partially reduced, infra-pharyngobranchials with the epibranchials, which may even or may not be fusedwith the proximal tip of the third infrapharyngobranchial. A more detailedstudy of the ontogeny of the epi- and infrapharyngobranchial elements inC. gariepinus, however, reveals that the cartilaginous mass, at the medial tip ofthe anterior two epibranchials, is confluent with them, as well as with the thirdinfrapharyngobranchial [Fig. 13(a): 9·1-mm SL stage]. This configurationremains established until the connection with the latter one is lost [Fig. 13(d):

T

II.StatusoftheinfrapharyngobranchialsinsomeOtophysi,basedon

datafrom

literature

Ordo-Subordo

Familia

Species

According

toliterature

According

tohypothesis

Reference

No.of

IPB

Reduced

Fusion

os-IPB

No.of

IPB

Reduced

Fusion

os-IPB

Cypriniformes

Cyprinidae

Rasboradaniconius

(Hamilton-Buchanan)

4—

1–2–3–4

?4

—1–2–3–4

?Tewari,1971

Notropisbifrenatus(Cope)

4—

2–3

1–(2–3)

4—

2–3

1–(2–3)Harrington,1955

CyprinuscarpioL.

4—

1–2–3

2–3

4—

1–2–3

2–3

Nelson,1969

Hom

alopteridae

Protomyzongriswoldi

4—

1–2–3

1–2–3–4*

4—

1–2–3

1–2–3–4Nelson,1969

Homalopteracaldwelli

4—

1–2–3–4

2–3

4—

1–2–3–4

2–3

Nelson,1969

Catostomidae

Catostomusmacrocheilus

Girard

34

—1–2–3

31†

—1–2–3

Weisel,1960

Characiformes

Hepsetidae

Hepsetusodoe

4—

—1–2–3

4—

—1–2–3

Fink&Fink,1981

4—

—?

Bertmar,1959

Hemiodontidae

Argonecteslongiceps

4—

?1–2–3–4

4—

?1–2–3–4Roberts,1974

HemiodussemitaeniatusKner

34

?2–3

34

?2–3

Roberts,1974

Bivibranchiaprotractila

Eigenman

34

?1–2–3

34

?1–2–3

Roberts,1974

AnodusmelanopogonCope

34

?1–2–3

34

?1–2–3

Roberts,1974

Parodontidae

SaccodonwagneriKner&

Steindachner

34

?1–2–3

34

?1–2–3

Roberts,1974

Characidae

BryconmeekiEigenman

&Hildebrand

34

?1–2–3

34

?1–2–3

Weitzman,1962

Siluroidei

Diplomystidae

Diplomysteschilensis(Molina)

4—

?1–2–3–4

4—

1–2–3

1–2–3–4Arratia,1987

34†

1–2–3

1–2–3

Nelson,1969

Ictaluridae

Trogloglanispattersoni

2(?)

1–2(?)

3–4

3–4

4—

1–2–3–4

3–4

Lundberg,1982

Pangasiidae

Pangasiuspangasius

4—

3–4

?4

—3–4

?Srinivasachar,1957a

Plotosidae

Plotsuscanius

4—

2–3–4

?4

—2–3–4

1–3

Srinivasachar,1958a

23–4

—1–2

Tilak,1963

Schilbeidae

Siloniasilondia

4—

3–4

?4

—3–4

?Srinivasachar,1957a

Ailiacoila

4—

3–4

?4

—3–4

?Srinivasachar,1957a

Clupisomagarua

4—

—1–2–3–4

4—

1–2

?Rastogi,1963

Schilbesp.

22–4‡

—1–3

4—

1–2

1–3–4

Nelson,1969

Eutropiichthysvacha

4—

3–4

1–2–(3–4)

4—

1–2/3–4

1–3–4

Tilak,1961

T

II.Continuedfrom

previouspage

Ordo-Subordo

Familia

Species

According

toliterature

According

tohypothesis

Reference

No.of

IPB

Reduced

Fusion

os-IPB

No.of

IPB

Reduced

Fusion

os-IPB

Ariidae

Ariusjella

4—

3–4

?4

—3–4

?Srinivasachar,1958a

Osteogeniosusmilitaris(L.)

4—

1–2–3–41–2–(3–4)

4—

1–2–3–4

1–3–4

Tilak,1965

AriustenuispinisDay

31

—3–4

4—

1–2

3–4

SrinivasaRao

&Lakshmi,1984

Siluridae

Wallagoattu(Bloch&

Schneider)

31

3–4

2–(3–4)

4—

3–4

2–3–4

Tilak,1963

Callichrousbimaculatus=

Ompokbimaculatus(Bloch)

31

3–4

2–(3–4)

4—

3–4

2–3–4

Tilak,1963

Amblycipitidae

Amblycepsmangois(Hamilton)

31

3–4

2–(3–4)

4—

1–2/3–4

3–4

Tilak,1967

Claroteidae

Parauchenoglanisguttatus

(Lonnberg)

31

—3–4

4—

1–2

3–4

Fink&Fink,1981

Cetopsidae

HelogenesmarmoratusGünther

3(?)

1(?)

—3–4

4—

1–2

3–4

dePinna

&Vari,1995

Paracetospsisbleekeri

Guichenot

3(?)

1(?)

—3–4

4—

1–2

3–4

dePinna

&Vari,1995

Clariidae

Clariasgariepinus

21–2

3–4

3–4

4—

1–2/3–4

3–4

Nawar,1954

Heterobranchuslongifilis

21–2

3–4

3–4

4—

1–2/3–4

3–4

Vandewalleetal.,1997

HeteropneustidaeHeteropneustesfossilis

21–2

3–4

?4

—1–2/3–4

3–4

Srinivasachar,1959

21–2

?3–4

Srinivasachar,1958b

Gymnotoidei

Sternopygidae

Eigenmannia

4—

—1–2–3

4—

—1–2–3

Fink&Fink,1981

Sternopygusmacrurus

Bloch&Schneider

4—

—1–2–3

4—

—1–2–3

Nelson,1969

Gymnotidae

GymnotuscarapoL.

4—

—2–3

4—

—2–3

Nelson,1969

Electrophoridae

Electrophoruselectricus(L.)

4—

3–4

1–2–3

4—

3–4

1–2–3

Nelson,1969

*InProtomyzontwosmallossificationsarepresent,oneinfrontofthesecond

andonebehind

thethirdinfrapharyngobranchials.

†Basedon

Fig.7ofWeisel(1960)andthedescriptioninthetext,itseemsmorelikelythatinfrapharyngobranchialIisabsent.

‡According

toNelson(1969)thefourthinfrapharyngobranchialIV

isconsideredasan

upperpharyngealjaw,although

itisaperichondralbone.

Reduced,serialnumberoftheinfrapharyngobranchialsthatarereduced;fusion,serialnumberoftheinfrapharyngobranchialsthathavefusedwitheach

other;

os-IPB,serialnumberoftheinfrapharyngobranchialsthatareossified(taxonom

yaccordingtoFink&Fink,1981).

1250 . .

21·5-mm SL stage]. The position and the relation of this cartilaginous mass tothe epibranchials during ontogeny suggests that indeed this mass represents theremains of infrapharyngobranchials I and II, instead of being an addition. If thishypothesis be applied for the configuration of other taxa described, a clear trendwithin siluroids can be proposed (Table II, right part). In descriptions of thesiluroid chondrocranium four infrapharyngobranchials are recognized except inthe Clariidae and closely related Heteropneustidae. Presumably, as the remnantsare very indistinct (in C. gariepinus), they may have become lost completely or

Partial reduction

eb-III

eb-IV

eb-I

eb-II

eb-III

eb-IV

eb-I

eb-II

Clarias gariepinus

Complete reduction

Clarias gariepinus

F. 12. Hypothesized strategies of infrapharyngobranchial configuration in Clarias gariepinus from theprimitive situation (Diplomystes nahuelbutaensis, Diplomystidae—after Arratia, 1987). eb-I,Epibranchiale; eb-II, epibranchiale II; eb-III, epibranchiale III; eb-IV, epibranchiale IV; ipb-I,infrapharyngobranchiale I ( ); ipb-II, infrapharyngobranchiale II ( ); ipb-III,infrapharyngobranchiale III ( ); ipb-IV, infrapharyngobranchiale IV (/); , additional cartilage.

1251

they may have been overlooked. In the descriptions of the osteocranium, all fourinfrapharyngobranchials are ossified, and consequently observed, in only a fewspecies. In some cases, a fusion is proposed between the bony third and fourthinfrapharyngobranchial (Tilak, 1961, 1963, 1965, 1967). However, when com-paring this fused complex with the configuration of C. gariepinus, it is more likelyto correspond with the fourth infrapharyngobranchial bone only, whereasTilak’s second one then corresponds with the third one in C. gariepinus. Theontogenetic evidence in C. gariepinus reveals no fusion between two ossificationcentra, which suggests that the complex infrapharyngobranchial might be amisinterpretation. Due to the fact that the strongly reduced first and secondinfrapharyngobranchials remain cartilaginous, they are consequently not recog-nized as such in osteological descriptions, which explains the assumed reductions(Nawar, 1954; Lundberg, 1982). The fact that the anterior two infrapharyngo-branchials become reduced in siluroids, may have led to the interpretation of thesingle unit-remnant of them, as the second infrapharyngobranchial only, sinceit is positioned in front of the third (Tilak, 1963, 1967; Fink & Fink, 1981;Srinivasa Rao & Lakshmi, 1984). Apart from the fusion between infrapharyngo-branchials I and II, a general trend in siluroids seems to involve a fusion betweeninfrapharyngobranchials III and IV (Table II). It is, however, important to

(a) ipb-I

ipb-II

ipb-III

ipb-IV

eb-I

eb-VI0.2 mm

(b) ipb-I

ipb-II

ipb-III

ipb-IV

0.2 mm

eb-I

eb-II

eb-III

eb-IV

(c) ipb-I

ipb-II

ipb-III

ipb-IV

0.2 mm

eb-I

eb-IIeb-III

eb-IV

(d)

ipb-I

ipb-II

ipb-III

ipb-IV

eb-I

1 mm

eb-II

eb-III

eb-IV

(e) ipb-I

ipb-II

ipb-III

ipb-IV

eb-I

1 mm

eb-II

eb-III

eb-IV

F. 13. Relation between the epibranchials and infrapharyngobranchials during ontogeny in Clariasgariepinus (dorsal view, pharyngeal tooth plates not drawn, light gray indicates cartilage, dark grayindicates perichondral ossifications): (a) 9·1 mm SL; (b) 10·0 mm SL; (c) 11·6 mm SL; (d) 21·5 mmSL; (e) 127·0 mm SL. eb-I, Epibranchiale I; eb-II, epibranchiale II; eb-III, epibranchiale III, eb-IV,epibranchiale IV; ipb-I, infrapharyngobranchiale I; ipb-II, infrapharyngobranchiale II; ipb-III,infrapharyngobranchiale III; ipb-IV, infrapharyngobranchiale IV.

1252 . .

mention that an extensive study of the ontogeny of the infrapharyngobranchialelements of different catfish species is required to confirm this hypothesis.

CONCLUSIONS

The overall ontogeny of Clarias gariepinus is comparable to the general trendobserved in siluroids. Structures like the pila lateralis and commissura lateralisare absent, as well as a myodome. The platybasic skull type, which is present inall presently known siluroids, involves a large fenestra hypophysea and aprecerebral lamina instead of an internasal septum. The notochord does notreach the latter fenestra, as the cartilago acrochordalis is present. The epiphysialbridge is well developed, as well as the preorbital base and the laminaorbito-nasalis. No foramen is observed in the preorbital base. The taeniamarginalis anterior seems to be reduced completely in Clarias, as is the case forsome other siluroids. The otic capsule is completely developed, except for thetectum synoticum. The tectum posterius becomes fused to the supraneuralcartilages of the second and third vertebrae. The fenestra basicapsularis anteriorarises initially, but disappears later during ontogeny. The glossopharyngeus andvagus nerves pass through separate foramina.The splanchnocranium corresponds to the general siluroid situation. The

palatine is separated from the pterygoquadrate from the moment it is formedand so secondary fusion occurs. The pterygoquadrate is fused to thehyosymplectic, thus forming a ‘ hyo-symplectic-pterygoquadrate plate ’, pene-trated by a foramen truncus hyomandibularis. This plate initially is continuouswith Meckel’s cartilage, at the level of the pars quadrata, and the hyoid archthrough the interhyal. The former connection is lost prior to the latter one. Awell-defined opercular process is present, as well as a distinct pterygoid process.The ceratohyals are fused to the hypohyals, whereas left and right hypohyalsinitially fuse in the middle. All basibranchials (I–V) develop, although theycould not always be distinguished as separate elements. Hypobranchials I–IIIbecome isolated from the ceratobranchials, which is not the case for the fourthone. All four epibranchials are present, as well as all four infrapharyngobranchi-als. Ontogenetic evidence indicates that the developmental status of the latterstructures may have been misinterpreted by several authors. A hypothesis,which suggests that the anterior two infrapharyngobranchials in many siluroidsmay be reduced to a single cartilaginous mass, may demonstrate after allthat a general trend in infrapharyngobranchial morphology is present withinsiluroids.

We thank F. Ollevier, F. Volckaert and W. Fleure for providing the specimens ofC. gariepinus; and G. De Wever and D. Vandenbroeck for making the serial sections ofC. gariepinus. Research was funded by the Institute for Science and Technology (DA)and the National Fund for Scientific Research (WV).

References

Adriaens, D. & Verraes, W. (1994). On the functional significance of the loss of theinterhyal during ontogeny in Clarias gariepinus Burchell, 1822 (Teleostei:Siluroidei). Belgian Journal of Zoology 124, 139–155.

1253

Adriaens, D. & Verraes, W. (1997). Ontogeny of the maxillary barbel muscles in Clariasgariepinus (Siluroidei: Clariidae), with some notes on the palatine-maxillarymechanism. Journal of Zoology (London), 241, 117–133.

Alexander, R. McN. (1965). Structure and function in catfish. Journal of Zoology(London) 148, 88–152.

Alfred, E. R. (1966). A new catfish of the genus Akysis from Malaya. Copeia 1966,467–470.

Arratia, G. (1987). Description of the primitive family Diplomystidae (Siluriformes,Teleostei, Pisces): morphology, taxonomy and phylogenetic implications. BonnerZoologische Monographien 24, 1–120.

Arratia, G. (1990). Development and diversity of the suspensorium of trichomycteridsand comparison with loricarioids (Teleostei: Siluriformes). Journal of Morphology205, 193–218.

Arratia, G. (1992). Development and variation of the suspensorium of primitive catfishes(Teleostei: Ostariophysi) and their phylogenetic relationships. Bonner ZoologischeMonographien 32, 1–148.

Arratia, G. & Schultze, H.-P. (1990). The urohyal: development and homology withinosteichthyans. Journal of Morphology 203, 247–282.

Bamford, B. T. W. (1948). Cranial development of Galeichthys felis (Ariidae). Proceed-ings of the Zoological Society (London) 118, 364–391.

Bertmar, G. (1959). On the ontogeny of the chondral skull in Characidae, with adiscussion on the chondrocranial base and the visceral chondrocranium in fishes.Acta Zoologica 40, 203–364.

Blache, J. (1964). Les Poissons du Bassin du Tchad et du Bassin adjacent du Mayo Kebbi:Étude Systématique et Biologique. Paris: O.R.S.T.O.M.

Boulenger, G. A. (1909–1916). Catalogue of Fresh-water Fishes of Africa in the BritishMuseum (Natural History). London: British Museum.

Boulenger, G. A. (1915). Diagnoses de poissons nouveaux. II. Mormyrides, Kneriides,Characinides, Cyprinides, Silurides. Revue Zoolique Africaine 4, 162–171.

Burchell, W. (1822). Travels in the Interior of Southern Africa. Volume I. London:Batchworth Press.

Corsin, J. (1961). Étude de quelques corrélations morphogénétiques dans le développe-ment du chondrocrane de Salmo. Bulletin de la Société Zoologique, France 86,772–785.

Crass, R. S. (1960). Notes on the freshwater fishes of Natal with descriptions of four newspecies. Annals of the Natal Museum 14, 405–458.

Daget, J. (1954). Les poissons du Niger Supérieur. Memoires Institut Francais d’AfriqueNoir 36, 1–391.

Daget, J. (1962). Les poissons du Fouta Dialon et de la basse Guinée. Memoires InstitutFrancais d’Afrique Noir 65, 1–210.

Daget, J. (1964). Le crane des Téléostéens. Mémoires du Musée National d’HistoireNaturelle Serie A 31, 163–341.

de Beer, G. R. (1937). The Development of the Vertebrate Skull. Oxford: Clarendon Press.de Pinna, M. C. C. & Vari, R. P. (1995). Monophyly and phylogenetic diagnosis of the

family Cetopsidae, with synonymization of the Helogeneidae (Teleostei:Siluriformes). Smithsonian Contributions to Zoology 571, 1–26.

Devillers, C. (1958). Le crane des poissons. In Aguathes et Poissons, anatomic, ithologic,systematique. Traite de Zoologic, (Grasse, P., Ed.) Paris: Masson et Cie.

De Vos, L. & Leveque, C. (1983). Étude systématique et morphologique du genreEutropius en Afrique de l’Ouest (Pisces, Schilbeidae). Revue Zoolique Africaine 97,469–532.

Fink, S. V. & Fink, W. L. (1981). Interrelationships of the ostariophysan fishes(Teleostei). Zoological Journal of the Linnean Society 72, 297–353.

Ghiot, F. & Bouchez, N. (1980). The central rod of the barbels of a South Americancatfish, Pimelodus clarias. Copeia 1980, 908–909.

Ghiot, F., Vandewalle, P. & Chardon, M. (1984). Comparaison anatomique et fonction-nelle des muscles et des ligaments en rapport avec les barbillons chez deux familles

1254 . .

apparentées de poissons Siluriformes Bagroidei. Annals du Société RoyalZoologique de Belgique 114, 261–272.

Giltay, L. (1929). Une characide nouveau du Congo Belge (Belonophago hutsebouti n. g.,n. sp. de la sous-famille des Ichthyoborinae). Revue de Zoologie et BotanieAfricaine 18, 271–276.

Glodek, G. S. (1976). Rhynchodoras woodsi, a new catfish from Eastern Ecuador(Siluriformes: Doradidae) with a redefinition of Rhynchodoras. Copeia 1970,43–46.

Goodrich, E. S. (1958). Studies on the Structure and Development of Vertebrates. VolumeI. New York: Dover Publications.

Gosline, W. A. (1975). The palatine-maxillary mechanism in catfishes with comments onthe evolution and zoogeography of modern siluroids. Occasional Papers of theCalifornian Academy of Sciences 120, 1–31.

Hanken, J. & Wassersug, R. (1981). The visible skeleton. A new double-stain techniquereveals the nature of the ‘‘ hard ’’ tissues. Functional Photography 16, 22–26.

Harrington, Jr, R. W. (1955). The osteocranium of the American cyprinid fish, Notropisbifrenatus, with an annotated synonymy of teleost skull bones. Copeia 1955,267–290.

Hoedeman, J. J. (1960). Studies on callichthyid fishes: 4. Development of the skull inCallichthys and Hoplosternum (1) (Pisces: Siluriformes). Bulletin of AquaticBiology 1, 73–84.

Howes, G. J. & Teugels, G. G. (1989). Observations and homology of the pterygoidbones in Corydoras paleatus and some other catfishes. Journal of Zoology(London) 219, 441–456.

Jarvik, E. (1980). Basic Structure and Evolution of Vertebrates. Volume I. London:Academic Press.

Kindred, J. (1919). The skull of Amiurus. Illinois Biological Monograph 5, 1–120.Kobayakawa, M. (1992). Comparative morphology and development of bony elements

in the head region in three species of Japanese catfishes (Silurus: Siluridae;Siluriformes). Japanese Journal of Ichthyology 39, 25–36.

Lucena, C. A. S., Malabarba, L. R. & Reis, R. E. (1992). Resurrection of the neo-tropical pimelodid catfish Parapimelodus nigribarbis (Boulenger), with aphylogenetic diagnosis of the genus Pimelodus (Teleostei: Siluriformes). Copeia1992, 138–146.

Lundberg, J. G. (1982). The comparative anatomy of the toothless blindcat, Trogloglanispattersoni Eigenmann, with a phylogenetic analysis of the ictalurid catfishes.Miscellanous Publications of the Museum of Zoology, University of Michigan 163,1–85.

Mahnert, V. & Géry, J. (1987). Deux nouvelles espèces du genre Hyphessobrycon (Pisces,Ostariophysi, Characidae) du Paraguay: H. guarani n. sp. et H. procerus n. sp.Bonner zoologische Beitragen 38, 307–314.

Marinelli, W. (1936). Kranium und Visceralsbelett: Algermeine probleme. In Handbuchder Vergleichende Anatomie der Wisbeltiere (Bolk, Goppert, Kallius and Lubosch,eds). Berlin: Urban & Schwarzenberg.

McElman, J. F. & Balon, E. K. (1980). Early ontogeny of white sucker, Catostomuscommersoni, with steps of saltatory development. Environmental Biology of Fishes5, 191–224.

Myers, G. S. & Weitzman, S. H. (1954). Another new Corydoras from Brazil. AquariumJournal 25, 93–94.

Myers, G. S. & Weitzman, S. H. (1966). Two remarkable new trichomycterid catfishesfrom the Amazon basin in Brazil and Colombia. Journal of Zoology (London)149, 277–287.

Nawar, G. (1954). On the anatomy of Clarias lazera: I. Osteology. Journal ofMorphology 94, 551–585.

Nelson, G. J. (1969). Gill arches and the phylogeny of fishes, with notes on theclassification of vertebrates. Bulletin of the American Museum for Natural History141, 479–552.

1255

Nijssen, H. & Isbrücker, I. J. H. (1983). Review of the genus Corydoras from Colombia,with descriptions of two new species (Pisces, Siluriformes, Callichthyidae).Beaufortia 33, 53–71.

Nijssen, H. & Isbrücker, I. J. H. (1987). Spectracanthus murinus, nouveau genre et espècede Poisson-chat cuirassé du Rio Tapajós, Est. Pará, Brésil, avec des remarques surd’autres genres de loricariidés (Pisces, Siluriformes, Loricariidae). Revue françaisdu Aquariologie 13, 93–98.

Nijssen, H. & Isbrücker, I. J. H. (1990). Lithoxus stocki, a species new to scienceof ancistrin loricariid catfish from the Maroni River drainage, with acomparison of the primary type specimens of the six species of Lithoxus (syn.:Paralithoxus) (Pisces, Siluriformes, Loricarriidae). Bijdragen tot de Dierkunde60, 327–333.

Pellegrin, J. (1928). Poisson du Kasai (Congo Belge), description d’un genre nouveauet de quatre espèces nouvelles. Bulletin de la Société de Zoologie, France 53,103–113.

Poll, M. (1945). Descriptions de Mormyridae et de Characidae nouveaux du CongoBelge, avec une étude du genre Stomatorhinus et des genres de Characidae nainsafricains. Revue de Zoologie et Botanie Africaine 39, 36–77.

Poll, M. (1954). Poissons de forét des environs de Yangambi (Stanleyville). Annals duMusée du Congo Belge 4, 56–68.

Poll, M. (1967a). Revision des Characidae nain africains. Annals du Muséum Royal del’Afrique Central de Sciences Zoologiques IN-8) 162, 1–158.

Poll, M. (1967b). Contribution à la faune ichthyologique de l’Angola. Publicacosculturais da Companhia de Diamantes de Angola 75, 1–381.

Poll, M. & Gosse, J. P. (1963). Révision des genres Nannaethiops Gunther 1871 etNeolebias Steindachner 1894, et description de trois espèces nouvelles (Pisces,Citharinidae). Annals du Muséum Royal de l’Afrique Central de SciencesZoologiques IN-8) 116, 7–41.

Poll, M. & Daget, J. (1968). Descriptions d’Hemistichodus lootensi (Pisces, Citharinidae).Bulletin du Muséum National d’Histoire Naturelle Serie 2 36, 1060–1065.

Poll, M., Lanza, B. & Sassi, A. R. (1972). Genre nouveau extraordinaire de Bagridae dufleuve Juba: Pardiglanis tarabinii gen. n. sp. n. (Pisces, Siluriformes). MonitorZoologica Italiano 15, 327–345.

Radermaker, F., Surlemont, C., Sanna, P., Chardon, M. & Vandewalle, P. (1989).Ontogeny of the Weberian apparatus of Clarias gariepinus (Pisces, Siluriformes).Canadian Journal of Zoology 67, 2090–2097.

Rastogi, M. (1963). The head skeleton of Indian schilbeid, Clupisoma garua (Ham.).Journal of Morphology 113, 205–214.

Reis, R. E. & Schaefer, S. A. (1992). Eurycheilus pantherinus (Siluroidei: Loricariidae), anew genus and species of Hypoptopomatinae from Southern Brazil. Copeia 1992,215–223.

Risch, L. (1987). Description of four new bagrid catfishes from Africa (Siluriformes:Bagridae). Cybium 11, 21–38.

Roberts, T. R. (1974). Osteology and classification of the neotropical characoid fishes ofthe families Hemiodontidae (including Anodontinae) and Parodontidae. Bulletinof the Museum of Comparative Zoology 146, 411–472.

Roman, B. (1966). Les poissons des hauts-bassins de la Volta. Annals du Muséum royalde l’Afrique Central de Sciences Zoologiques IN-8) 150, 1–191.

Schultz, L. P. (1942). The fresh-water fishes of Liberia. Proceedings of the U.S. NaturalMuseum 92, 301–348.

Skelton, P. H. & Teugels, G. G. (1992). Neotype description for the African catfishClarias gariepinus (Burchell, 1822) (Pisces: Siluroidei: Clariidae). IchthyologicalBulletin of the J.L.B. Smith Institute for Ichthyology, Grahamstown, S. Africa 56,1–8.

Srinivasa Rao, K. & Lakshmi, K. (1984). Head skeleton of the marine catfish Ariustenuispinis Day (Osteichthyes: Siluriformes, Ariidae). Journal of Morphology 181,221–238.

1256 . .

Srinivasachar, H. R. (1957a). Development of the skull in catfishes: part I: developmentof chondrocranium in Silonia, Pangasius and Ailia (Schilbeidae). Proceedings ofthe National Institute of Sciences of India 22B, 335–356.

Srinivasachar, H. R. (1957b). Development of the skull in catfishes: part II: developmentof chondrocranium inMystys and Rita (Bagridae). Morphologisches Jahrbuch 98,244–261.

Srinivasachar, H. R. (1958a). Development of the skull in catfishes: part IV: develop-ment of chondrocranium in Arius jella Day (Ariidae) and Plotosus caniusHam. (Plotosidae) with an account of their interrelationships. GegenbauersMorphologisches Jahrbuch 99, 986–1016.

Srinivasachar, H. R. (1958b). Development of the skull in catfishes: part V: developmentof skull in Heteropneustes fossilis (Bloch). Proceedings of the National Institute ofSciences of India 24B, 165–190.

Srinivasachar, H. R. (1959). Development of the skull in catfishes: part III: thedevelopment of the chondrocranium in Heteropneustes fossilis (Bloch)(Heteropneustidae) and Clarias batrachus (Linn.) (Clariidae). MorphologischesJahrbuch 101, 373–405.

Steindachner, F. (1914). Zur Fischfauna des Dscha, eines sekundaren Nebenfluâes desKongo, im Bezirke Molundu, Kamerun. Denkschrift von der KaiserlichesAkademie der Wissenschaften (Mathematische-Naturwissenschafftliche Klasse) 89,1–64.

Stewart, D. J. (1985). A new species of Cetopsorhamdia (Pisces: Pimelodidae) from theRio Napo Basin of Eastern Ecuador. Copeia 1985, 339–344.

Stewart, D. J. (1986). Revision of Pimelodina and description of a new genus and speciesfrom the Peruvian Amazon (Pisces: Pimelodidae). Copeia 1986, 653–672.

Stewart, D. J. & Pavlik, M. J. (1985). Revision of Cheirocerus (Pisces: Pimelodidae) fromtropical freshwaters of South America. Copeia 1985, 356–367.

Surlemont, C. (1983). Recherche sur les transformations postembryonnaires de la régioncéphalique d’un poisson-chat, Clarias gariepinus (Burchell, 1822). Ph.D. Thesis,Université de Liège.

Surlemont, C. & Vandewalle, P. (1991). Développement postembryonnaire du squeletteet de la musculature de la tête de Clarias gariepinus (Pisces, Siluriformes) depuisl’éclosion jusqu’à 6,8 mm. Canadian Journal of Zoology 69, 1094–1103.

Surlemont, C., Chardon, M. & Vandewalle, P. (1989). Skeleton, muscles and movementsof the head of a 5·2 mm fry of Clarias gariepinus (Burchell) (Pisces: Siluriformes).In Fortschritte der Zoologie, Trends in Vertebrate Morphology (Splechtna &Hilgers, eds), pp. 459–462. Stuttgart: Gustav Fischer Verlag.

Teugels, G. G. (1982). A systematic outline of the African species of the genus Clarias(Pisces; Clariidae), with an annotated bibliography. Annals du Muséum Royal del’Afrique Central de Sciences Zoologiques 236, 1–249.

Teugels, G. G. (1986). A systematic revision of the African species of the genus Clarias(Pisces; Clariidae). Annals du Muséum Royal de l’Afrique Central de SciencesZoologiques 247, 1–199.

Teugels, G. G. & Thys van den Audenaerde, D. F. E. (1990). Description of a newspecies of Bryconaethiops (Teleostei: Characidae) from Nigeria and Cameroon.Ichthyological Exploration of Freshwaters 1, 207–212.

Teugels, G. G., Risch, L., De Vos, L. & Thys van den Audenaerde, D. F. E. (1991).Generic review of the African bagrid catfish genera Auchenoglanis en Parauchenog-lanis with description of a new genus. Journal of Natural History 25, 499–517.

Tewari, S. K. (1971). The development of the chondrocranium of Rasbora daniconius(Ham. Buch). Gegenbauers Morphologisches Jahrbuch 116, 491–502.

Thys van den Audenaerde, D. F. E. (1964). Revision of the genus Eutropiellus Nichols &Lamonte (Pisces, Schilbeidae) with description of a new species from LowerNigeria, Eutropiellus vandeweyeri sp. nov. Revue de Zoologie et Botanie Africaine69, 214–230.

Thys van den Audenaerde, D. F. E. & De Vos, L. D. G. (1982). Description of Eutropiusdjeremi sp. n. Revue Zoolique Africaine 96, 179–184.

1257

Tilak, R. (1961). The osteocranium and Weberian apparatus of Eutropiichthys vacha(Ham.) and E. murius (Ham.): a study of interrelationships. Zoologisches Anzeiger167, 413–430.

Tilak, R. (1963). The osteocranium and Weberian apparatus of a few representatives ofthe family Siluridae and Plotosidae (Siluroidea): a study of interrelationships.Zoologisches Anzeiger 171, 424–439.

Tilak, R. (1965). The comparative osteology of the osteocranium and the Weberianapparatus of the Tachysuridae (Pisces: Siluroidei). Journal of Zoology (London)146, 150–174.

Tilak, R. (1967). The osteocranium and Weberian apparatus of Amblyceps mangois(Hamilton) (Pisces: Siluroidei) in relation to taxonomy. Zoologisches Anzeiger178, 61–74.

Vandewalle, P., Laleye, P. & Focant, B. (1995). Early development of cephalic bonyelements in Chrysichthys auratus (Geoffroy Saint-Hilaire, 1808) (Pisces,Siluriformes, Claroteidae). Belgian Journal of Zoology 125, 329–348.

Vandewalle, P., Gluckmann, I., Baras, E., Huriaux, F. & Focant, B. (1997).Postembryonic development of the cephalic region in Heterobranchus longifilis.Journal of Fish Biology 50, 227–253.

Vari, R. P. & Ortega, H. (1986). The catfishes of the neotropical family Helogenidae(Ostariophysi: Siluroidei). Smithsonian Contributions to Zoology 442, 1–20.

Verraes, W. (1974). Discussion on the shape of the eye and the influence of its size onshape and position of surrounding structures in normal and abnormal conditionsduring postembryonic development in Salmo gairdneri Richardson, 1836 (Pisces,Salmonidae). Forma et Functio 7, 125–138.

Volckaert, F. A. M., Galbusera, P. H. A., Hellemans, B. A. S., Van den Haute, C.,Vanstaen, D. & Ollevier, F. (1994). Gynogenesis in the African catfish (Clariasgariepinus). I. Induction of meiogenesis with thermal and pressure stocks.Aquaculture 128, 221–233.

Weisel, G. F. (1960). The osteocranium of the catostomid fish, Catostomus macrocheilus:a study in adaptation and natural relationship. Journal of Morphology 106,109–129.

Weitzman, S. H. (1962). The osteology of Brycon meeki, a generalized characid fish, withan osteological definition of the family. Stanfords Icthyological Bulletin 8, 1–77.