The supratemporal system and the pattern of ramification of cephalic sensory canals in Denticeps clupeoides (Denticipitoidei, Teleostei): additional evidence for monophyly of Clupeiformes and Clupeoidei

The supratemporal system and the pattern of ramification of cephalic sensory canals in Denticeps clupeoides (Denticipitoidei, Teleostei): additional evidence for monophyly of Clupeiformes and Clupeoidei

Fabio Di Dario^I,II; Mário C. C. de Pinna^I,III

^ISeção de Peixes, Museu de Zoologia, Universidade de São Paulo, Caixa Postal 42494-970, CEP 04218-970, São Paulo, SP, Brasil

^IIPost-doctoral fellow; [e-mail]

^IIIProfessor and Vice-Director; [e-mail]

ABSTRACT

The cephalic portions of the latero-sensory canal system in Denticeps clupeoides are described and analyzed. The species, a small herring-like fish from relictual West African streams, is the most primitive living clupeomorph and sole Recent representative of suborder Denticipitoidei. As sister group to over 360 species in Clupeoidei, Denticeps is a key taxon in understanding clupeomorph and lower teleostean relationships. Observations on recently-collected specimens of Denticeps clupeoides revealed comparatively-important and previously-unrecorded details of the cephalic latero-sensory canals which are relevant for understanding relationships at different levels in clupeomorph phylogeny. The infraorbital, supraorbital, preopercular, extrascapular and post-temporal canals of Denticeps have unbranched tubules in soft tissue, as in the hypothesized plesiomorphic condition for lower teleosts. Contrastingly, the presence of a complex network formed by a high order of branching of cephalic canals is hypothesized as a synapomorphy of the Clupeoidei. Denticeps and the Clupeoidei share an exclusive sensory branch that originates at the junction between the extrascapular bone and the recessus lateralis, here hypothesized as an additional synapomorphy of Clupeiformes. A supratemporal system is newly recorded in Denticeps, and the character is proposed as a synapomorphy of Clupeiformes, and not of Clupeoidei as previously thought. The hypothesis that the supratemporal system is homologous to the supraorbital cavern is refuted, and the latter is corroborated as an autapomorphy of Denticeps. Another autapomorphy of Denticeps (or Denticipitoidei) is the presence of the postorbital bulla, a hitherto unrecorded specialization of the infraorbital canal associated with infraorbitals 4 and 5. Homologies of other tubules of the cephalic sensory canals in Denticeps are also discussed, with emphasis on their bearing on the recognition of homologies of infraorbital bones in Denticeps and other lower teleosts. In general, data from the cephalic latero-sensory system corroborate Denticeps as the sister group to all other Recent clupeomorphs, and provide additional support for the monophyly of Clupeoidei and Clupeiformes.

Keywords: Clupeiformes, Denticeps clupeoides, latero-sensory system, lower teleosts, phylogenetic relationships, postorbital bulla, sensory biology.

RESUMO

As porções cefálicas do sistema de canais látero-sensoriais em Denticeps clupeoides são descritas e analisadas. A espécie, uma pequena sardinha com distribuição restrita a poucos riachos relictuais na África Ocidental, é o mais primitivo Clupeomorpha vivente, e único representante Recente da subordem Denticipitoidei. Como grupo irmão das mais de 360 espécies incluídas em Clupeoidei, Denticeps é um táxon-chave no entendimento das relações filogenéticas em Clupeomorpha e outros Teleostei inferiores. Observações com base em espécimes recentemente coletados de Denticeps clupeoides revelaram detalhes inéditos e comparativamente importantes dos canais látero-sensoriais cefálicos, relevantes para o entendimento de relações em diferentes níveis da filogenia de Clupeomorpha. Os canais infraorbital, supraorbital, preopercular, extrascapular e pós-temporal de Denticeps possuem túbulos simples nos tecidos moles, como na situação considerada plesiomórfica para teleósteos inferiores. Em contraste, a presença de uma rede complexa formada por ramificações múltiplas dos canais cefálicos é considerada uma sinapomorfia de Clupeoidei. Denticeps e Clupeoidei compartilham um ramo sensorial exclusivo que se origina na junção entre o osso extrascapular e o recessus lateralis, aqui proposto como uma sinapomorfia adicional de Clupeiformes. O sistema supratemporal é registrado pela primeira vez em Denticeps, e o caráter é consequentemente proposto como sinapomorfia de Clupeiformes, e não de Clupeoidei como se acreditava previamente. A hipótese de que o sistema supratemporal seria homólogo à caverna supraorbital é refutada, e a última é considerada autapomórfica para Denticeps. Uma outra autapomorfia de Denticeps (ou Denticipitoidei) é a presença da bula pós-orbital, uma especialização previamente desconhecida do canal infraorbital associada com os infraorbitais 4 e 5. As homologias de outros túbulos dos canais sensoriais cefálicos em Denticeps também são discutidas, com ênfase no reconhecimento de homologias dos ossos infraorbitais em Denticeps e em outros Teleostei inferiores. Em geral, os dados do sistema látero-sensorial cefálico corroboram a hipótese de Denticeps como grupo irmão de todos os outros clupeomorfos Recentes, e fornecem evidência adicional para o monofiletismo de Clupeoidei e Clupeiformes.

Palavras-chave: Clupeiformes, Denticeps clupeoides, sistema látero-sensorial, teleósteos inferiores, relações filogenéticas, bula pós-orbital, biologia sensorial.

INTRODUCTION

Denticeps clupeoides Clausen, 1959 (Fig. 1) is a peculiar little herring-like fish that inhabits a few isolated coastal streams in West Africa, from Eastern Benin to Cameroon (Teugels, 2003). The species is sole Recent representative of its genus, family Denticipitidae, and suborder Denticipitoidei. The Denticipitidae also includes the strikingly similar Paleodenticeps tanganikae from the Tertiary of Tanzania, East Africa (Greenwood, 1960). Denticipitoidei is sister-group to about 360 species of the fishes commonly known as sardines, herrings, and anchovies, grouped in the Clupeoidei (Greenwood et al., 1966; Greenwood, 1968a, 1968b; Grande, 1985; Nelson, 2006). The Denticipitoidei and the Clupeoidei together comprise the Clupeiformes, which includes all living species of the Clupeomorpha (Grande, 1985). Despite their abundance and importance, the anatomy and relationships of clupeomorphs are still poorly known. Comparative or phylogenetically-oriented works on the group or its subunits have been relatively few, both on the basis of morphological and molecular data. It seems to us that current disputes about the relationships among Osteoglossomorpha, Elopomorpha, Otocephala and basal euteleosts (e.g., Johnson & Patterson, 1996; Arratia, 1997, 1999; Ishiguro et al., 2003; Lavoué et al., 2005) would benefit from a deeper understanding of character variation in clupeomorphs in general. Denticeps clupeoides in particular, because of its basal position among clupeomorphs, is a key taxon for understanding relationships among lower teleosts.

The acusticolateralis system has a series of unique features in different clupeomorph subgroups, particularly in the head region (Tracy, 1920; Wohlfahrt, 1936, 1937; O'Connell, 1955). The recessus lateralis is a complex that probably acts as a center for integration of stimuli perceived by the lateral line. Its presence is hypothesized as a synapomorphy of the Clupeiformes (Grande, 1985; Di Dario, 2004). The recessus lateralis is an intracranial space in the otic region into which the main branches of cephalic sensory canals converge, namely the supraorbital, infraorbital, preopercular, and extrascapular/post-temporal (Grande, 1985; Di Dario, 2004; Figs. 2 and 3). The exact mechanism and function of the recessus lateralis are not yet understood, although it is probably related to detection and analysis of small vibrational pressures and displacements (Hoss & Blaxter, 1982).

In species of the Clupeoidei, there is an ontogenetic increase in the branching of the main cephalic canals that results in a highly complex network formed by superficial soft-tissue tubules (Stephens, 1985; Fig. 2). The branches lack neuromasts and always end in a pore (Greenwood, 1968a; Coombs et al., 1988). The pattern of ramification in representative species of the Clupeoidei has been accurately described and illustrated in a series of papers (Wohlfahrt, 1937; Bamford, 1941; Gunter & Demoran, 1961; Hoss & Blaxter, 1982; Nelson, 1983; Stephens, 1985; Blaxter, 1987). The network formed by the branches covers virtually the whole head of the fish (Fig. 2), and is likely responsible for an increase in the detection of mechanical waves (Stephens, 1985). Another distinct structure of the sensory apparatus of some clupeomorphs is the supratemporal system, originally described by Wohlfahrt (1937: "supratemporale System"). The supratemporal system is a complex structure formed by branched canals at the temporal region that originate from a sac-like expanded sensory canal that fills the temporal foramen (Patterson, 1970). A limited connection between the supratemporal system and the rest of the latero-sensory system persists via the supraorbital canal in most adult clupeomorphs (Bamford, 1941; Di Dario, 2004), although according to Wohlfahrt (1937) those systems are completely independent in Sardina pilchardus and possibly other clupeomorphs. The modified, sac-like, sensory canal or vesicle that fills the temporal foramen is the sinus temporalis (Wohlfahrt, 1937; Fig. 3). The temporal foramen is apparently exclusive to the Clupeiformes (Lauder & Liem, 1983). Typically, it consists of a roughly oval foramen dorsal to the recessus lateralis in the region of articulation between the frontal and parietal bones (Grande, 1985; Fig. 3). Bamford (1941) speculated that the supratemporal system is a sensory organ that detects differences of pressure around the head, caused by movements of the fish. Association between the supratemporal system, the recessus lateralis, and the highly branched head canal system have been proposed as one of the main factors responsible for the extraordinary ecological success of clupeomorphs, assisting in a variety of functions such as the precise perception of the position of the fish in a school, and in pressure balance during vertical migration (Blaxter, 1987; Webb, 1989).

The acusticolateralis system of Denticeps clupeoides is still poorly known. It is clear, nevertheless, that it is composed of an expected mosaic of autapomorphic and primitive features (Di Dario, 2004). One hypothesized autapomorphic structure of the acusticolateralis system of Denticeps is the supraorbital cavern (Figs. 4A and 5, cv), which is a sac-like canal that occupies a relatively large region between two bony bridges of the frontal, which are themselves also autapomorphic for Denticeps (Greenwood, 1968a; Fig. 4B, fbs). Other features of the cephalic sensory canals of Denticeps have been noted, although their phylogenetic significance remains nebulous. Clausen (1959) observed five to seven soft unbranched aborbital tubules crossing the opercle of Denticeps in an oblique posteroventral direction. These tubules originate "from the postemporal and extrascapular portions of the main trunk, and ( ) from the near junction between the infraorbital and preopercular branches." (Clausen, 1959:144). Similarity between those tubules and the branched sensory canals in the head of species of the Clupeoidei led Greenwood (1968a) to consider their presence as evidence that Denticeps is indeed a clupeomorph, at a time when characters from other complexes indicative of such a relationship were unknown or speculative.

Collecting efforts in the Republic of Benin recently yielded a large number of specimens of Denticeps clupeoides for study (Fig. 1). Observations on well-preserved specimens cleared and stained for bone, cartilage, and nerves, revealed details of the cephalic sensory canals unrecorded or not precisely described so far. The exact patterns and points of ramification of sensory tubules from main cephalic trunks could thus be identified and described for the first time. On the basis of new data, specific hypotheses of homologies between those tubules and sensory branches of members of the Clupeoidei are proposed. The aim of this paper is to report on these new findings and to interpret their phylogenetic significance.

MATERIAL AND METHODS

A total of 21 specimens of Denticeps clupeoides, between 26.1 and 40.7 mm SL, were prepared according to the clearing and counter-staining techniques of Taylor & van Dyke (1985) and Song & Parenti (1995), which allow for visualization of bone and cartilage. The nerve staining phase of Song & Parenti (1995) was successfully done in six of those specimens. In order to facilitate visualization of the cephalic sensory canals in cleared and stained specimens, India ink diluted in 50% glycerin was injected into that system, through a small opening in the usually large preopercular sensory canal. A similar procedure was employed for seven alcohol preserved specimens bleached with hydrogen peroxide, but in this case, India ink was diluted in 70% alcohol. Dissection of cleared and stained specimens followed the protocol described in Weitzman (1974), which exposes most of the relevant skeletal structures while minimizing damage to articulations. Drawings were made with a camera lucida attached to a stereomicroscope. Anatomical terminology follows Weitzman (1962), and Grande (1985). Cephalic canal terminology follows Nelson (1972) and Di Dario (2004).

Institutional abbreviations

ANSP, Academy of Natural Sciences, Philadelphia

FMNH, Field Museum, Chicago

INPA, Instituto Nacional de Pesquisas da Amazônia, Manaus

MCP, Museu de Ciências e Tecnologia da Pontifícia Universidade Católica, Porto Alegre

MZUSP, Museu de Zoologia da Universidade de São Paulo, São Paulo

SAIAB, South African Institute for Aquatic Biodiversity, Grahamstown

UMMZ, University of Michigan Museum of Zoology, Ann Arbor

USNM, National Museum of Natural History, Smithsonian Institution, Washington DC

Anatomical abbreviations

ao, antorbital

aor, anterior opening to the chamber of recessus lateralis

atc, accessory temporal sensory canal

atio₁, anterior tubule of infraorbital 1

atpc, anterior tubule of post-temporal sensory canal

atss, anterior tubule of supratemporal system

bs, basisphenoid

cv, supraorbital cavern

dfp, dorsal opening of the pre-epiotic fossa

dmpb, dorsal middle tubule of postorbital bulla

dsp, dermosphenotic

dtpb, dorsal tubule of postorbital bulla

dtpp, dorsal tubule of preopercular sensory canal

ec, extrascapular sensory canal

ehc, ethmoidal sensory canal

epo, epioccipital

exo, exoccipital

ext, extrascapular

fbs, parasagittal frontal bridge

fp, pre-epiotic fossa

fr, frontal

gr, groove on the lateral wing of the frontal

ic, infraorbital sensory canal

io₁₅, infraorbital bones 15

iop, interopercle

lopoc, ventral opening for main preopecular sensory canal

mc, mandibular sensory canal

mor, middle opening to the chamber of recessus lateralis

mx, maxilla

na, nasal

op, opercle

optt, middle opening of the post-temporal

os, orbitosphenoid

oscl, posterior opening of the supracleithrum

pa, parietal

pcl, postcleithrum

pio₄, posterodorsal process of laminar region of infraorbital 4

pmx, premaxilla

pop, preopercle

por, posterior opening to the chamber of recessus lateralis

posc, postorbital branch of supraorbital sensory canal

ppc, preopercular sensory canal

pro, prootic

ptc, post-temporal sensory canal

ptec, parietal tubule of extrascapular sensory canal

ptio₁, posterior tubule of infraorbital 1

pto, pterotic

ptpc, posterior tubule of post-temporal sensory canal

ptrc, posterior tubule of recessus lateralis

pts, pterosphenoid

ptss₁, posterior tubule of supratemporal system 1

ptss₂, posterior tubule of supratemporal system 2

ptt, post-temporal

pvor, posteroventral opening to the chamber of recessus lateralis

rec, chamber of the recessus lateralis

reco, common opening to the chamber of the recessus lateralis

sc, supraorbital sensory canal

scl, supracleithrum

so, supraorbital

soc, supraoccipital

sop, subopercle

sp, sphenotic

ss, supratemporal system

st, sinus temporalis

stec, supraoccipital tubule of extrascapular sensory canal

tcv, tubule of supraorbital cavern

tf, temporal foramen.

tfc, chamber of temporal foramen

tlopoc, tubule of the ventral opening for main preopercular sensory canal

tpvr, tubule of the posteroventral opening of recessus lateralis

tuopoc, tubule of the dorsal opening for main preopercular sensory canal

uopoc, dorsal opening for main preopecular sensory canal

vmpb, ventral middle tubule of postorbital bulla

vtpb, ventral tubule of postorbital bulla

Comparative material examined

Number of examined specimens follows catalog number, and refers to number of cleared and stained specimens examined (unless noted: alc, alcohol preserved, unstained specimen; skl, skeleton), not to total number in lot:

Clupeiformes

Denticipitidae: Denticeps clupeoides, MZUSP 62480, 1; MZUSP 84776, 20 + 7(alc).

Chirocentridae: Chirocentrus dorab, MZUSP 62467, 1; MZUSP 72930, 1(skl).

Clupeidae: Alosa caspia, USNM 143891, 1; A. pseudoharengus, MZUSP 62471, 1; Anodontostoma chacunda, MZUSP 62476, 1; Brevoortia aurea, MZUSP 11729, 1; B. pectinata, MCP 7722, 1; MCP 7725, 1; MCP 7727, 1; Clupea harengus, ANSP 157065, 1; Clupeichthys aesarnensis, MZUSP 62465, 3; C. bleekeri, MZUSP 63114, 1; C. perakensis, MZUSP 63104, 3; Dorosoma cepedianum, MZUSP 62481, 3; USNM 272899, 6; Dussumieria acuta, MZUSP 62468, 3; Etrumeus teres, MZUSP 62469, 3; USNM 188934, 3; Gilchristella aestuaria, MZUSP 64115, 3; Harengula clupeola, MZUSP 18672, 1; H. jaguana, MZUSP 11269, 2; MZUSP 10791, 1; Herklotsichthys dispilonotus, MZUSP 63115, 1; Lile piquitinga, MZUSP 11215, 1; Ophistonema oglinum, MZUSP 10844, 3; Pellonula afzeliusi, UMMZ 195029, 2; Platanichthys platana, MZUSP 10629, 3; MCP 19409, 3; Rhinosardina amazonica, MZUSP 11231, 2; MZUSP 11452, 1; Sardina pilchardus, MZUSP 37394, 3; MZUSP 12123, 1; Sardinella albella, MZUSP 63117, 2; S. aurita, USNM 272875, 2; S. brasiliensis, MZUSP 12124, 1; MZUSP 11418, 2; S. maderensis, MZUSP 37382, 1; Tenualosa ilisha, USNM 276407, 2; Spratelloides delicatulus, MZUSP 62470, 3.

Coiliidae: Coilia neglecta, USNM 357380, 6; C. rebentischii, MZUSP 62483, 1; Lycothryssa crocodilus, MZUSP 62482, 1; Setipinna melanochir, MZUSP 64119, 1; Thryssa baelama, ANSP 63228, 1; T. hamiltoni, USNM 364595, 1; T. mystax, ANSP 60578, 1.

Engraulidae: Amazonsprattus scintilla, MZUSP 28475, 2(alc); Anchoa filifera, MZUSP 18528, 2; A. lamprotaenia, MZUSP 11508, 2; A. lyolepis, MZUSP 11476, 2; A. spinifer, MZUSP 11454, 2; A. tricolor, MZUSP 11547, 2; Anchovia clupeoides, MZUSP 11559, 1; Anchoviella brevirostris, MZUSP 11578, 2; A. carrikeri, MZUSP 5728, 3; A. guianensis, MZUSP 5726, 1; A. lepidentostole, MZUSP 51826, 2; Cetengraulis edentulus, MZUSP 11634, 3; Encrasicholina heteroloba, MZUSP 63118, 2; Engraulis anchoita, MZUSP 18393, 10; Pterengraulis atherinoides, MZUSP 11723, 1; Stolephorus indicus, MZUSP 63112, 2; MZUSP 63113, 1; MZUSP 63106, 1.

Pristigasteridae: Chirocentrodon bleekerianus, MZUSP 11097, 3; Ilisha africana, MZUSP 62463, 4; I. amazonica, MZUSP 33266, 2; I. elongata, UMMZ 219537, 1; I. kampeni, MZUSP 62477, 2; I. megaloptera, MZUSP 62472, 2; I. melastoma, MZUSP 62473, 2; Neoopisthopterus tropicus, MZUSP 62478, 2; Odontognathus mucronatus, MZUSP 11264, 3; MZUSP 11267, 3; MZUSP 10835, 2; Opisthopterus dovii, MZUSP 62462, 3; O. equitorialis, MZUSP 62479, 2; O. tardoore, MZUSP 62475, 2; Pellona castelnaeana, INPA 4959, 1; MZUSP 5492, 1; P. ditchela, ANSP 63526, 1; USNM 189996, 1; P. flavipinnis, MZUSP 40063, 1; MZUSP 18728, 1; P. harroweri, MZUSP 11283, 3; MZUSP 11308, 1; MZUSP 11374, 3; Pliosteostoma lutipinnis, FMNH 2818, 2; Pristigaster cayana, MZUSP 30338, 3; P. whiteheadi, INPA 8555, 3; Raconda russeliana, MZUSP 62466, 4.

Gonorynchiformes

Chanidae: Chanos chanos, MZUSP 62601, 1; USNM 347536, 9.

Gonorynchidae: Gonorynchus sp., MZUSP 63663, 1.

Kneriidae: Kneria auriculata, MZUSP 63121, 4; USNM 290762, 2.

Cypriniformes

Catostomidae: Catostomus commersoni, USNM 238094, 2; USNM 238111, 2; USNM 340759, 2.

Cyprinidae: Opsariichthys uncirostris, USNM 87445, 2; Zacco platypus, MZUSP 62597, 1.

Gyrinocheilidae: Gyrinocheilus aymonieri, USNM 271692, 1.

Characiformes

Citharinidae: Citharinus latus, MZUSP 84480, 1.

Distichodontidae: Xenocharax spilurus, MZUSP 50358, 1.

Characidae: Hollandichthys multifasciatus, MZUSP uncataloged, 5.

Curimatidae: Curimata roseni, MZUSP 55740, 1.

Argentiniformes

Argentinidae: Argentina striata, USNM 188212, 2.

Alepocephalidae: Searsia koefoedi, USNM 206873, 1; Talismania aphos, USNM 215540, 1.

Salmoniformes

Galaxiidae: Galaxias auratus, USNM 344893, 3; USNM 344895, 1.

Osmeridae: Osmerus mordax, MZUSP 64116, 1.

Salmonidae: Salmo sp., MZUSP uncataloged.

Esociformes

Esocidae: Esox americanus, USNM 237257, 2; USNM 120051, 2.

Umbridae: Dalia pectoralis, USNM 034033, 1; Umbra pygmaea, USNM 333152, 3; USNM 345523, 1.

Elopomorpha

Albulidae: Albula vulpes, MZUSP 10625, 2.

Elopidae: Elops sp., MZUSP 60346, 1.

Megalopidae: Megalops cyprinoides, USNM 173580, 3.

Osteoglossomorpha

Hiodontidae: Hiodon tergisus, MZUSP 28450, 1; H. alosoides, USNM 350554, 1. Notopteridae: Chitala sp., MZUSP uncataloged, 1.

RESULTS

Supratemporal system

The temporal foramen of Denticeps clupeoides is overlain by a flange formed mostly by the frontal (Greenwood, 1968a; Patterson, 1970; Fig. 5). A small portion of the posterior region of that flange, and consequently of the temporal foramen, is of parietal origin. The flange is arched laterally, resulting in a somewhat bulged chamber with a ventrally directed aperture. As the temporal foramen of Denticeps is housed within that chamber, the ventral opening of the chamber may be regarded as the actual exit to the exterior of the sinus temporalis. That distinguishes Denticeps from other clupeiforms, where the opening of the temporal foramen is directed laterally (Fig. 3). In spite of major differences in the architecture of the temporal region, the relative sizes of the opening of the chamber of the temporal foramen in Denticeps and the temporal foramen of other clupeiforms are similar.

The sinus temporalis completely fills the chamber of the temporal foramen in Denticeps clupeoides (Figs. 4A and 5). Consequently, the structure is relatively larger in Denticeps than in species of the Clupeoidei. Additionally, a superficial connection between the sinus temporalis and the supraorbital canal, present in some members of the Clupeoidei (e.g., Fig. 3), is absent in Denticeps. The region between the recessus lateralis and the opening of the chamber of the temporal foramen is slightly concave, with a pronounced angle close to the middle of the opening. A small portion of the sinus temporalis projects from the ventral opening of the chamber and rests on that concavity.

A series of unbranched tubules with diameters approximately similar among themselves, but somewhat variable along their length, originate from the exposed portion of the sinus temporalis (Figs. 4A and 5). Each tubule crosses the temporal region in a posteroventral direction and end in a pore. Tubules of the supratemporal system in Denticeps are extremely delicate and are often damaged during preparation and dissection of specimens. Of 21 cleared and stained specimens examined here, only nine had intact tubules on both sides of the head. Of those, one specimen has four tubules on each side, whilst two other specimens have three tubules on each side. All other specimens show different numbers on the left and right sides, revealing rampant bilateral variation in supratemporal tubule number in Denticeps. Two of the same nine specimens have three right and four left tubules, and two other specimens have those values reversed. A single specimen has three right and five left tubules, but with the anteriormost of those five extremely short. Another single specimen has three right and two left tubules. In nine additional specimens, supratemporal tubules could be counted on one side only (usually the left one, due to normal dissection practice). Seven of those have three tubules and two have four tubules. Finally, three other specimens were extensively dissected in the head, and tubules could not be counted on either side, although their severed remains in the soft tissue definitely indicate their presence.

In spite of variation in numbers of tubules of the sinus temporalis, the length of tubules and the region of the cranium they cover are approximately constant. One or two of the anterior tubules are short and reach only to the anteroventral margin of the extrascapular (Figs. 4 and 5). The extrascapular bears sections of the extrascapular and post-temporal canals, and both canals converge to a single opening at the posterior region of the recessus lateralis (Di Dario, 2004). Remaining one to three posterior tubules of the supratemporal system are longer and overlie the extrascapular bone (and associated canal) and the anterior region of the post-temporal bone (Figs. 4 and 5). A tubule of the post-temporal canal projects from a somewhat oblique opening at the post-temporal bone (Fig. 4B, optt), which encloses that portion of the main trunk of the post-temporal canal. The oblique opening at the post-temporal bone marks the posterior limit reached by any tubule of the supratemporal system.

The region ventral to the opening of the chamber of the temporal foramen and portions of the dorsal regions of the extrascapular and post-temporal overlain by tubules of the supratemporal system are devoid of odontodes, which otherwise cover most bones of the head, the dorsal portion of the pectoral girdle, and associated scales (Clausen, 1959; Greenwood, 1968a; Sire et al., 1998).

Postorbital bulla, and ramification of cephalic sensory canals

The pattern of ramification of cephalic sensory canals is remarkably constant among specimens of Denticeps examined, with just minor variation. The overall aspect is much less complex than that in members of the Clupeoidei. The infraorbital, supraorbital, preopercular, and post-temporal canals have primary tubules, which typically do not ramify (Fig. 4A).

A hitherto unrecorded specialization of the infraorbital canal of Denticeps clupeoides adds to the plethora of outstanding structures of the acusticolateralis system of clupeomorphs. The proximal portion of the infraorbital canal of Denticeps is markedly dilated, forming an elongated and clearly recognizable bulla, here called postorbital bulla (Figs. 4 and ⁶). Hypertrophy of that portion of the infraorbital canal reflects on the structure of associated infraorbital bones. Infraorbital 5 encloses the proximal region of the postorbital bulla and is therefore unusually inflated (^{Fig. 6}), to such an extent that it is laterally the most prominent bone in the head. The mid to anterior portion of the postorbital bulla is partially covered by the anterior margin of infraorbital 4 that gradually tapers distally. The posterodorsal region of the laminar component of infraorbital 4 in most examined specimens of Denticeps bears a delicate process that shores the posterior margin of the median section of the postorbital bulla (^{Fig. 6}, pio₄). A major portion of the postorbital bulla, located between the projection of the anterior margin and the process at the posterodorsal region of the laminar component of infraorbital 4, is not enclosed by bone. The anterior end of the postorbital bulla is marked by an abrupt decrease in diameter of the infraorbital canal (^{Fig. 6}), which then remains roughly the same for the rest of its length (Fig. 4).

Four tubules of the latero-sensory system of the head of Denticeps originate from the postorbital bulla. These tubules are not enclosed by bone, are approximately parallel, have the same length and diameter, and project from the postorbital bulla in roughly posteroventral direction (Fig. 4A). During their course, they overlie the preopercle and associated canal. As with tubules of the supratemporal system and those of other main cephalic canals, tubules of the postorbital bulla end in a pore. The dorsal tubule of the postorbital bulla originates near the connection between the infraorbital canal and the recessus lateralis, and ends approximately on the middle of the opercle (Figs. 4 and ⁶, dtpb).

Two tubules project from the middle region of the postorbital bulla. The dorsal one originates at the articulation between infraorbitals 4 and 5 (Figs. 4 and ⁶, dmpb). It is separated from the ventral middle tubule by a small bony process (^{Fig. 6}, pio₄, described above). The two middle tubules of the postorbital bulla are closer to each other than each is to other tubules of the postorbital bulla. The dorsal middle tubule typically does not branch and ends in a region close to the middle of the posterior margin of the preopercle (Fig. 4). Variation in course and ramification was observed in the dorsal middle tubule of the postorbital bulla. In four examined specimens of Denticeps, the distal region of the dorsal middle tubule abruptly turns dorsally at a roughly 90º angle upon passing the posterior border of the preopercle and ends on the anterior third of the opercle. In one examined specimen, the dorsal middle tubule bifurcates. The dorsal branch of the bifurcation ends at a pore located approximately on the anterior third of the opercle, while the ventral one ends on the middle to ventral region of the posterior border of the preopercle. The ventral middle tubule never ramifies (Fig. 4A, vmpb), and ends in a pore at a region close to the dorsal opening for the main preopercular latero-sensory canal (Fig. 4B, uopoc).

The fourth tubule of the postorbital bulla originates ventrally at the anterior region of infraorbital 4 (Figs. 4A and 5, vtpb). It is oriented towards the angled posteroventral border of infraorbital 4, and ends at a region close to the middle of the ventral border of the preopercle.

Two short tubules of the infraorbital canal originate at the anterior region of the canal sector located in infraorbital 1. They are slightly curved posteriorly, and each ends in a pore at the ventral border of infraorbital 1 (Fig. 4A, atio₁ and ptio₁).

The main trunk of the preopercular canal is entirely enclosed in the somewhat inflate preopercle except for a relatively small region of connection with the recessus lateralis, where a posteroventrally-directed dorsal tubule of the preopercular canal originates (Figs. 4A and ⁶, dtpp). That tubule is dorsal to the dorsal tubule of the postorbital bulla. Both tubules are roughly parallel and have similar lengths and diameters, with the dorsal tubule of the preopercular canal ending in a pore located on the middle of the opercle (Fig. 4). Two additional tubules of the preopercular canal are located at the posteroventral region of the preopercle. The shape of the preopercle and the pattern of ramification of that portion of the preopercular canal were accurately described by Greenwood (1968a). He labeled the openings in the posteroventral region of the bone from which those tubules project as the upper and lower openings of the main preopercular sensory canal (Greenwood, 1968a:238239; Fig. 4B, uopoc and lopoc, respectively). The upper opening of the preopercular canal leads to a narrow groove that houses its tubule (Fig. 4A, tuopoc). The lower opening of the preopercular canal is a long slit (Greenwood, 1968a). Both tubules end in pores at the margin of the posterior tip of the preopercle, which is extremely sharp (Greenwood, 1968a).

A total of four tubules are associated to the recessus lateralis. The two ventral ones have been described above in association with the preopercular canal and the postorbital bulla. A long tubule originates from the posteroventral opening of the recessus lateralis, located between its middle and posterior openings (Fig. 4B, pvor). Di Dario (2004) hypothesized that opening of the recessus as autapomorphic for Denticeps clupeoides. The tubule of the posteroventral opening of the recessus lateralis crosses in a ventral oblique direction the dorsal third of the opercle and ends in a pore at the posterior border of the bone (Fig. 4A, tpvr). The dorsal and last tubule associated to the recessus lateralis (Fig. 4A, ptrc) originates at the connection between the posterior opening of the recessus (Di Dario, 2004; Fig. 4B, por) and the canal enclosed in the anterior region of the extrascapular bone, which results from the merging of the proximal tips of the extrascapular and post-temporal canals. The posterior tubule of the recessus lateralis is dorsal and parallel to the tubule of the posteroventral opening of the recessus, runs along the proximal region of the opercle, and ends in a pore at the posterodorsal border of the bone.

Portions of the infraorbital bones, preopercle, and opercle overlain by superficial tubules of the recessus lateralis and of the infraorbital and preopercular canals lack odontodes (^{Fig. 6}). That may have led Clausen (1959:144) to note that the tubules over the opercle of Denticeps are fringed by odontodes, when in fact odontodes cover the whole region except the paths of tubules. Greenwood (1960) noticed two large and well-defined odontode-free areas on the opercle of Paleodenticeps tanganikae. The position of those areas suggests that in life they were overlain by superficial tubules in a similar way as in Denticeps clupeoides (Greenwood, 1960).

Short tubules were also observed in the extrascapular, post-temporal, and supraorbital canals of Denticeps clupeoides. Two tubules are present in the post-temporal canal. The anterior one projects from a somewhat oblique dorsal opening at the middle of the post-temporal bone (Fig. 4B, optt). That tubule is oriented in a posterodorsal direction, and ends in a pore (Fig. 4A, atpc). The post-temporal canal bears another posterior tubule in the short space located between the supracleithrum and the first lateral-line scale (Fig. 4A, ptpc). As in the case of the anterior tubule, the posterior tubule of the post-temporal canal extends posterodorsally and ends in a pore. The extrascapular canal sends out two tubules during its course. The longer of those tubules (Fig. 4A, ptec) is present in the region of articulation between the extrascapular and the parietal, in the short space where the extrascapular canal is not enclosed by either of these two bones (Figs. 4B and ⁶). It is closely associated to and rests along the dorsal margin of the posteriormost tubule of the supratemporal system. The openings of both tubules are located in roughly the same region, which is slightly posterior to the posterior border of the extrascapular bone. Another short tubule originates at the space between the ossified segments of the extrascapular canal in the parietal and the supraoccipital (Fig. 4A, stec). That tubule projects posteriorly roughly perpendicular to the main axis of the body, and ends in a pore at a region close to the superficial border of the supraoccipital.

The last tubule of a cephalic sensory canal of Denticeps is a delicate structure located at the posterolateral region of the supraorbital cavern (Figs. 4A and ⁶, cv), which is a specialization of the supraorbital canal (Greenwood, 1968a). It extends over the top of the cranium, and opens in a small pore dorsal to the region between the anterior and posterior borders of the chamber of the temporal foramen (Fig. 4A, tcv).

DISCUSSION

Patterson (1970) suggested the presence of the supratemporal system as a synapomorphy of the Clupeoidei, and one which is absent in other teleosts including Denticeps clupeoides (Greenwood, 1968a). Our observations show that a supratemporal system is actually present in Denticeps, and that the structure is therefore a synapomorphy of the Clupeiformes. Differences in the anatomy of the supratemporal system, particularly of the sinus temporalis, between Denticeps and members of the Clupeoidei, explain the difficulties in identifying the structure in that genus previously. Such differences are probably related to the peculiar architecture of the temporal region in Denticeps, especially of its temporal foramen. The supratemporal system is a delicate structure, not preserved in fossils. Nevertheless, as the sinus temporalis is homologous to the parietal branch of the supraorbital canal (Wohlfahrt, 1937; Bamford, 1941; Di Dario, 2004), its absence can be inferred in fossils where that portion of the canal is unmodified (Patterson, 1970). Such is the case for all non-clupeiform clupeomorphs (Grande, 1985). The supratemporal system is also absent in all other known teleosts.

The identification of the supratemporal system in Denticeps also refutes the hypothesis of Di Dario (2004) that the supratemporal system of the Clupeoidei and the supraorbital cavern of Denticeps are homologous. The supraorbital cavern is exclusive to Denticeps among recent teleosts, as originally proposed by Greenwood (1968a). A particular feature of the supraorbital cavern, previously unnoticed, is the presence of an associated tubule (Fig. 4A, tcv). The pore at the distal end of that tubule is the only direct connection between the supraorbital cavern and the environment.

Another specialization of the cephalic sensory canals exclusive to Denticeps clupeoides is the postorbital bulla. The degree of inflation of that portion of the infraorbital canal and related bones, particularly of infraorbital 5, is unique among examined members of the Clupeoidei, other clupeomorphs and other lower teleosts (Grande, 1982a, 1982b, 1985; Chang & Grande, 1997; Chang & Maisey, 2003). This character cannot currently be observed in Paleodenticeps tanganikae. Tentatively, the postorbital bulla is hypothesized as a synapomorphy of the Denticipitoidei, based on the quite similar overall morphology of the circumorbital bones in Denticeps and Palaeodenticeps (Greenwood, 1960:8). Although the function of the postorbital bulla cannot be precisely determined at the moment, it is known that widening of portions of latero-sensory canals is associated with enhanced sensitivity to mechanical waves (Blaxter, 1987; Coombs et al., 1988; Webb, 1989).

Four major patterns of cephalic sensory canals have been identified in teleosts, according to their general anatomy and degree of branching (Coombs et al., 1988; Webb, 1989). Denticeps clupeoides has the simple canal pattern, where canals are typically cylindrical with bony walls, and only primary ramifications (tubules) are present. Such pattern is also present in representative basal teleosts, as in the osteoglossomorph Hiodon (Allis, 1904; Nelson, 1972; Hilton, 2002), and in ostariophysans such as Chanos chanos, cypriniforms (Lekander, 1949; Nelson, 1972; Siebert, 1987), primitive siluriforms (Arratia, 1987; Arratia & Huaquin, 1995), and characiforms (pers. obs.). The Clupeoidei is apomorphically characterized by the presence of a highly-branched canal pattern, forming an elaborate system composed of numerous pores at the end of secondary, tertiary, or even higher-order branching of sensory canals in the soft tissue of the head (Fig. 2). Such a pattern is most probably primitive for the suborder in view of its wide distribution among representative members of the clade, which spam most of the clades recognized for Pristigasteridae, Engraulidae and Clupeidae (Wohlfahrt, 1937; Bamford, 1941; Gunther & Demoran, 1961; Nelson, 1972; Hoss & Blaxter, 1982; Nelson, 1983; Stephens, 1985, 1996; Blaxter, 1987; Webb, 1989). Basal euteleosts typically have reduced, unbranched (i.e., simple) canal patterns (Nelson, 1972; Coombs et al., 1988; Webb, 1989). Complex ramifications similar to the condition found in the Clupeoidei are rare among lower teleosts. Some such cases occur within Siluriformes, where the possession of a dendritic arrangement of latero-sensory canals in the skin of snout, cheek and nape is hypothesized as a synapomorphy of the Pimelodidae, sensu stricto (character 1 of Lundberg et al., 1991), also occurring homoplastically in some other siluriform subclades (de Pinna, 1993). The elopiforms Elops and Megalops also have been reported to have a branched pattern of cephalic sensory canals (Webb, 1989). Nevertheless, the degree of ramification of those canals and the extent to which the network covers the head is much reduced when compared to members of the Clupeoidei (Nelson, 1972; Webb, 1989). Cephalic sensory canals in other elopomorphs are unbranched, except for a few shallow-water anguilliforms (Coombs et al., 1988; Webb, 1989), indicating that the states in the Elopiformes and the Clupeoidei are convergent. Cephalic sensory canals with secondary and tertiary branches are found in basal actinopterygians, as in Lepisosteiformes (Allis, 1904) and Amiiformes (Allis, 1889), and in distal euteleosts, as sparids, exocoetids, labrids, and serranids (Coombs et al., 1988; Webb, 1989). Based on current understanding of actinopterygian relationships, such occurrences are independent of that in Clupeoidei.

The repeated occurrence of superficial ramification of cephalic sensory canals in broadly divergent levels in actinopterygian phylogeny clearly shows that the character is prone to rampant homoplasy at higher levels. Within Clupeoidei, however, canal branching is constant in all species except highly paedomorphic ones, such as Amazonsprattus (pers. obs.) and Sundasalanx (Siebert, 1997). Cases of reversal in miniaturized species is not surprising, given that the latero-sensory canal system expands and branches off ontogenetically (Stephens, 1985; Weitzman & Vari, 1988; Webb, 1989; Arratia & Huaquin, 1995). We note, however, that the typical highly-branched pattern occurs in some other putatively paedomorphic small clupeoids, such as Thrattidion noctivagus (Roberts, 1972).

The condition of superficial sensory tubules on the head of Denticeps clupeoides is also elucidative at other levels of clupeiform phylogeny. The posteroventral opening of the recessus lateralis and its associated tubule are features exclusive of Denticeps among Recent teleosts (Di Dario, 2004). Whether this condition is a synapomorphy for Denticeps or Denticipitoidei cannot be resolved at present, because the relevant anatomical portions are not well-preserved in the recessus lateralis of available specimens of Paleodenticeps tanganikae (Greenwood, 1960).

Homologies between tubules of the latero-sensory system of the head of Denticeps and main trunks of the highly branched canals in members of the Clupeoidei can also be hypothesized, within the framework of a sister group relationship between the two taxa (Greenwood, 1968a; Greenwood et al., 1966; Grande, 1985; Di Dario, 2002, 2004). The origin, angle of orientation, and extent of projection over the opercle, are all strongly suggestive of a homology between the posterior tubule of the recessus lateralis of Denticeps (Fig. 4A, ptrc) and branch 53 identified by Stephens (1985) in Dorosoma (Fig. 2). That branch, which originates between the extrascapular and the posterior opening of the recessus lateralis, also occurs in all members of the Clupeoidei so far studied (c.f. Stephens, 1985). No such tubule or branch is present in non-clupeiform teleosts. Thus, the presence of an unbranched tubule or a branched canal that originates at the junction between the extrascapular bone and the recessus lateralis and extends over the posterodorsal region of the opercle is hypothesized as a synapomorphy of Clupeiformes.

The posterior tubule of the recessus lateralis of Denticeps and branch 53 of the Clupeoidei are the only ramifications of the cephalic canals with identical origins in the bones in the cranium of members of both taxa. The lack of equivalent landmarks for other canal branches complicates futher hypotheses of homology. Two branches are typically present between the posterior opening of the recessus lateralis and the parietal of the Clupeoidei (Stephens, 1985), whereas a single tubule exists at that region in Denticeps clupeoides. The dorsal and more developed of the ramifications of the extrascapular canal in Clupeoidei is branch 62 of Stephens (1985), which further ramifies into branches 61 and 63 (Fig. 2). The dorsal position on the extrascapular bone and the similar degree of development shared between branch 62 in Clupeoidei and the parietal tubule of the extrascapular canal in Denticeps suggest that those ramifications are homologous. However, the alternative hypothesis of homology between the parietal tubule of the extrascapular canal of Denticeps and the ventral, reduced, branch of the extrascapular canal (branch 60) of Clupeoidei cannot be currently refuted. The presence of a different and, in case of members of Clupeoidei, variable number of ramifications of the post-temporal canal (Stephens, 1985) also prevents further proposal of homologies between ramifications of that portion of the latero-sensory system in Denticeps and the Clupeoidei.

Further homologizing of canal branches, however, are possible on the basis of one particular ramification in Denticeps and Clupeoidei which may serve as a useful landmark. In both taxa, only the dorsalmost ramification of the infraorbital canal overlies the opercle. In Denticeps, that ramification is the dorsal tubule of the postorbital bulla, which originates from the relatively small exposed canal area between infraorbital 5 and the recessus lateralis (Figs. 4A and ⁶, dtpb). In the Clupeoidei, branch 24 of Stephens (1985) originates at the junction between infraorbital 5 and the dermosphenotic (infraorbital 6). Branch 24 is the dorsalmost ramification of the infraorbital canal in clupeoids, and the only one which extends onto the opercle. The dorsal tubule of the postorbital bulla in Denticeps and branch 24 share various topological similarities. Both originate at the proximal region of infraorbital 5. Also, they are oriented in a posteroventral direction, and end in a region close to the middle of the opercle. Both ramifications overlie the preopercle and, consequently, the preopercular canal along their course. That set of attributes strongly indicates that the dorsal tubule of the postorbital bulla in Denticeps is homologous to branch 24 in Clupeoidei. Because branch 24 is constant in several species of Clupeoidei (Stephens, 1985) and probably primitive for the clade, it is a potentially useful anatomical landmark. The main difference between the two ramifications is the position of their origin. The dorsal tubule of the postorbital bulla in Denticeps originates relatively closer to the recessus lateralis than branch 24 does in Clupeoidei. That difference, nevertheless, is a result of the extremely reduced and displaced dermosphenotic of Denticeps, which is dorsal to the anterior opening of the recessus lateralis and covers the groove for the proximal portion of the supraorbital canal in the frontal (Greenwood, 1968a; Grande, 1985; Di Dario, 2004). Dermosphenotics in Clupeoidei, contrastingly, are relatively well developed, resting between infraorbital 5 and the recessus lateralis (Grande, 1985; Di Dario, 2004).

If homology is accepted between the dorsal tubule in Denticeps of the postorbital bulla and branch 24 in clupeoids, then other ramifications of the infraorbital canals in the two taxa are easily homologized. Branches 23 to 21, which originate sequentially in the region between the distal end of infraorbital 5 and the proximal end of infraorbital 3 (Fig. 2), are found in all members of the Clupeoidei so far studied (Stephens, 1985). The posteroventral orientation, degree of development, and relations to infraorbital bones and the preopercle of branches 23 to 21 and the dorsal middle, ventral middle, and ventral tubules of the postorbital bulla of Denticeps (Fig. 4A) are strikingly similar, indicating that those ramifications are sequentially homologous. Branches originating between infraorbitals 1 to 3 are variable in members of the Clupeoidei (Stephens, 1985), and therefore difficult to homologize at higher taxonomic levels.

The identification of homologies between branches 21 to 24 and tubules of the postorbital bulla helps to resolve the homology between dermosphenotic of Denticeps and that of other teleosts. The so-called dermosphenotic of Denticeps is so unusual that its homologies have long been controversial (Greenwood, 1968a; Grande, 1985; Di Dario, 2004). Greenwood (1968a) was the first to propose the homology of the bone in Denticeps with the dermosphenotic in other teleosts. Grande (1985) and Di Dario (2004) presented and discussed a series of evidence (e.g., shape of the bone, relations to cephalic canals, absence of neuromasts) contrary to that hypothesis. Di Dario (2004) suggested instead that the so-called dermosphenotic of Denticeps is a neomorph, and that infraorbital 5 of Denticeps is actually homologous to the dermosphenotic of other teleosts, including members of the Clupeoidei. The reduction in the number of infraorbital bones from six, which is the widespread condition among basal teleosts, to five in Denticeps would have resulted from the fusion between infraorbitals 3 and 4 in Denticeps. However, the pattern of ramification of the tubules of the postorbital bulla indicates that infraorbitals 3 to 5 in Denticeps are indeed homologous to similarly-numbered bones in members of the Clupeoidei and other basal teleosts with the primitive condition of six infraorbital bones. These considerations support Greenwood's (1968a) idea that the dermosphenotic in Denticeps is actually homologous to the bone bearing the same name in other lower teleosts.

A reduction in the proximal region of the infraorbital canal with the associated displacement of the dermosphenotic from its primitive position to a location dorsal to the recessus might explain the articulation between infraorbital 5 and the recessus lateralis in Denticeps. The distance between the proximal portion of the infraorbital and preopercular canals in all clupeiforms is markedly reduced, a trait hypothesized as a synapomorphy of the order (Di Dario, 2004). If the proximal portion of the infraorbital canal was shortened in Denticeps, such morphological change would be associated with a shortening of the proximal portion of the preopercular canal. The morphology of the preopercle of Denticeps favors that idea. The vertical distance between the preopercle and the temporal region which houses the recessus lateralis or the temporal canal is markedly smaller in Denticeps than in other clupeomorphs (Grande, 1982a, 1982b, 1985; Chang & Grande, 1997; Chang & Maisey, 2003; pers. obs.). In members of the Clupeoidei, the horizontal distance between the preopercular and infraorbital canals is markedly shorter than the vertical distance between the preopercle and the recessus lateralis (Fig. 2). In Denticeps, on the other hand, the horizontal distance between the proximal portions of the infraorbital and preopercular canals is only slightly shorter than the vertical distance between the preopercle and the recessus lateralis (Fig. 4B).

The proximity between the preopercle and the recessus lateralis in Denticeps may impede the presence of more than one tubule in that region, whereas three developed branches typically exist in the large space between the preopercle and the recessus lateralis in species of Clupeoidei (Stephens, 1985; Fig. 2). According to Stephens (1985), the most well-developed of the aborbital branches of the proximal portion of the preopercular canal is branch 50 (Fig. 2). That branch leaves the preopercular canal shortly after leaving the dorsal opening of the preopercle, and ends in the mid to posterior region of the opercle. The dorsal tubule of the preopercle of Denticeps (Fig. 4A, dtpp) also originates at the dorsal opening of the preopercle, runs in a similar orientation over the opercle, and ends approximately in the middle of the bone, as does branch 50. These features are indirect evidence that the dorsal tubule of the preopercular canal of Denticeps and branch 50 of members of the Clupeoidei are homologous.

ACKNOWLEDGMENTS

We first thank Dr. Philippe Laleye (Université D'Abomey-Calavi, Faculté des Sciences Agronomiques, Cotonou), for logistic help and support during our visit to Benin. We are grateful to Roger Bills (South African Institute for Aquatic Biodiversity, Grahamstown), Timo Moritz (University of Würzburg, Würzburg), Charles Niyonkuru and Céline Dan (both from Université D'Abomey-Calavi) for help and companionship during fieldwork. Additional words of appreciation go to Roger Bills for allowing us to use photographs of Denticeps clupeoides taken during our joint expedition, one of which resulted in Figure 1. The manuscript benefited from comments and suggestions by C. Moreira, M. Toledo-Piza, and N. Menezes. For loans and exchanges of specimens we thank M. Rogers (FMNH), G. Burgess (UF), C. Lucena (MCP), M. Sabaj and J. Lundberg (ANSP), L. Py-Daniel (INPA), D. Nelson and W. Fink (UMMZ), S. Jewett and R. Vari (USNM). Research funding was provided by CNPq (Conselho Nacional de Desenvolvimento Científico e Tecnológico), FAPESP (Fundação de Amparo à Pesquisa do Estado de São Paulo, proc. 99/09781-6 and 05/54162-5), Smithsonian Institution, and NSF (National Science Foundation).

Recebido em: 23.08.2006

Aceito em: 10.10.2006

• Allis, E.P., Jr. 1889. The anatomy and development of the lateral line system in Amia calva. Journal of Morphology, 2:463-566.
• Allis, E.P., Jr. 1904. The latero-sensory canals and related bones in fishes. Journal of Anatomy and Physiology, 21:401-502.
• Arratia, G. 1987. Description of the primitive family Diplomystidae (Siluriformes, Teleostei, Pisces): morphology, taxonomy and phylogenetic implications. Bonner Zoologische Monographien, 24:1-120.
• Arratia, G. 1997. Basal teleosts and teleostean phylogeny. Palaeo Ichthyologica, 7:1-168.
• Arratia, G. 1999. The monophyly of Teleostei and stem-group teleosts. Consensus and disagreements. In: Arratia, G. & Schultze, HP (Eds.), Mesozoic Fishes 2 Systematics and fossil record. Verlag Dr. Friedrich Pfeil, München, p.265-334.
• Arratia, G. & Huaquin, L. 1995. Morphology of the lateral-line system and of the skin of diplomystid and certain primitive loricarioid catfishes, and systematic and ecological considerations. Bonner Zoologische Monographien, 36:1-110.
• Bamford, T.W. 1941. The lateral line and related bones of the herring (Clupea harengus L.). Annals and Magazine of natural History, 8:414-438.
• Blaxter, J.H.S. 1987. Structure and development of the lateral line. Biological Reviews, 6:471-514.
• Chang, M.M. & Grande, L. 1997. Redescription of Paraclupea chetungensis, an early clupeomorph from the Lower Cretaceous of southeastern China. Fieldiana Geology (New Series), 1489:1-19.
• Chang, M.M. & Maisey, J.G. 2003. Redescription of Ellimma branneri and Diplomystus shengliensis, and relationships of some basal clupeomorphs. American Museum Novitates, 3404:1-35.
• Clausen, H.S. 1959. Denticipitidae, a new family of primitive isospondylous teleosts from West African freshwaters. Videnskabelige Meddelelser Dansk Naturhistorisk Forening, 121:141-151.
• Coombs, S.; Janssen, J. & Webb, J.F. 1988. Diversity of lateral line systems: evolutionary and functional considerations. In: J. Atema; R.R. Fay; A.N. Popper & W. Tavolga, Sensory Biology of Aquatic Animals. Springer, New York. p.553-593.
• Di Dario, F. 2002. Evidence supporting a sister group relationship between Clupeoidea and Engrauloidea. Copeia, 2002:496-503.
• Di Dario, F. 2004. Homology between the recessus lateralis and cephalic sensory canals, with the proposition of additional synapomorphies for the Clupeiformes and the Clupeoidei. Zoological Journal of the Linnean Society, 141:257-270.
• Grande, L. 1982a. A revision of the fossil genus Diplomystus, with comments on the interrelationships of clupeomorphs fishes. American Museum Novitates, 2728:1-34.
• Grande, L. 1982b. A revision of the fossil genus Knightia, with a description of a new genus from the Green River Formation (Teleostei, Clupeidae). American Museum Novitates, 2731:1-22.
• Grande, L. 1985. Recent and fossil clupeomorph fishes with materials for revision of the subgroups of clupeoids. Bulletin of the American Museum of Natural History, 181:231-372.
• Greenwood, P.H. 1960. Fossil denticipitid fishes from East Africa. Bulletin of the British Museum (Natural History), 5:1-11, 3 plates.
• Greenwood, P.H. 1968a. The osteology and relationships of the Denticipitidae, a family of clupeomorph fishes. Bulletin of the British Museum (Natural History), Zoology, 16:215-272.
• Greenwood, P.H. 1968b. Notes on the visceral anatomy of Denticeps clupeoides Clausen, 1959, a West African clupeomorph fish. Revue de Zoologie et de Botanique Africaines, 77:1-10.
• Greenwood, P.H.; Rosen, D.E.; Weitzman, S.H. & Myers, G.S. 1966. Phyletic studies of the teleostean fishes, with a provisional classification of living forms. Bulletin of the American Museum of Natural History, 131:339-456.
• Gunther, G. & Demoran, W.J. 1961. Lateral lines and an undescribed sensory area on the head of the Gulf Menhaden, Brevoortia patronus. Copeia, 1961:39-42.
• Hilton, E.J. 2002. Osteology of the extant North American fishes of the genus Hiodon Lesueur, 1818 (Teleostei: Osteoglossomorpha: Hiodontiformes). Fieldiana (Zoology), New Series, 100:1-142.
• Hoss, D.E. & Blaxter, J.H.S. 1982. Development and function of the swimbladder-inner ear-lateral line system in the Atlantic menhaden, Brevoortia tyrannus (Latrobe). Journal of Fish Biology, 20:131-142.
• Ishiguro, N.B.; Miya, M. & Nishida, M. 2003. Basal euteleostean relationships: a mitogenetic perspective on the phylogenetic reality of the "Protacanthopterygii". Molecular Phylogenetics and Evolution, 27:476-488.
• Johnson, G.D. & Patterson, C. 1996. Relationships of lower euteleostean fishes. In: Stiassny, M.L.J., Parenti, L.R. & Johnson, G.D. (Eds.), Interrelationships of Fishes. Academic Press, San Diego, p.251-330.
• Lauder, G.V. & Liem, K.F. 1983. The evolution and interrelationships of the actinopterygian fishes. Bulletin of the Museum of Comparative Zoology, Harvard, 150:95-197.
• Lavoué, S.; Miya, M.; Inoue, J.G.; Saitoh, K.; Ishiguro, N.B. & Nishida, M. 2005. Molecular systematics of the gonorynchiform fishes (Teleostei) based on whole mitogenome sequences: implications for higher-level relationships within the Otocephala. Molecular Phylogenetics and Evolution, 37:165-177.
• Lekander, B. 1949. The sensory line system and canal bones in the head of some Ostariophysi. Acta Zoologica (Stockholm), 30:1-131.
• Lundberg, J.G.; Mago-Leccia, F. & Nass, P. 1991. Exallodontus aguanai, a new genus and species of Pimelodidae (Pisces: Siluriformes) from deep river channels of South America, and delimitation of the subfamily Pimelodinae. Proceedings of the Biological Society of Washington, 104:840-869.
• Nelson, G.J. 1972. Cephalic sensory canals, pitlines, and the classification of esocoid fishes, with notes on galaxiids and other teleosts. American Museum Novitates, 2492:1-49.
• Nelson, G.J. 1983. Anchoa argentivittata, with notes on other Eastern Pacific anchovies and the Indo-Pacific genus Encrasicholina. Copeia, 1983:48-54.
• Nelson, J.S. 2006. Fishes of the World, 4Ş. ed., John Wiley & Sons, Inc. Hoboken, New Jersey.
• O'Connell, C.P. 1955. The gas bladder and its relation to the inner ear in Sardinops caerula and Engraulis mordax. Fishery Bulletin of the Fish and Wildlife Service, 56:503-533.
• Patterson, C. 1970. A clupeomorph fish from the Gault (Lower Cretaceous). Zoological Journal of the Linnean Society, 49:161-182.
• de Pinna, M.C.C. 1993. Higher-level phylogeny of Siluriformes (Teleostei, Ostariophysi), with a new classification of the order. Ph.D. Dissertation, City University of New York, New York.
• Roberts, T. 1972. Osteology and description of Thrattidion noctivagus, a minute, new freshwater clupeid fish from Cameroon, with a discussion of pellonulin relationships. Breviora, 382:1-25.
• Siebert, D.J. 1987. Interrelationships among families of the order Cypriniformes (Teleostei). Ph.D. Dissertation, City University of New York.
• Siebert, D.J. 1997. Notes on the anatomy and relationships of Sundasalanx Roberts (Teleostei, Clupeidae), with descriptions of four new species from Borneo. Bulletin of the Natural History Museum, London (Zoology), 63:13-26.
• Sire, JY.; Marin, S. & Allizard, F. 1998. Comparison of teeth and dermal denticles (odontodes) in the teleost Denticeps clupeoides (Clupeomorpha). Journal of Morphology, 237:237-255.
• Song, J. & Parenti, L.R. 1995. Clearing and staining whole fish specimens for simultaneous demonstration of bone, cartilage and nerves. Copeia, 1995:114-118.
• Stephens, R.R. 1985. The lateral line system of the gizzard shad, Dorosoma cepedianum Lesueur (Pisces: Clupeidae). Copeia, 1985:540-556.
• Stephens, R.R. 1996. Reconstruction of the relationships of Harengula, Herklotsichthys, Opisthonema, Amblygaster, and Sardinella. Copeia, 1996:475-478.
• Taylor, W.R. & van Dyke, G. 1985. Revised procedures for staining and clearing small fishes and other vertebrates for bone and cartilage. Cybium, 9:107-119.
• Teugels, G. 2003. Denticipitidae. In: D. Paugy; C. Lévêque & Teugels, G.G. (Eds.), Poissons d'eaux douces et saumâtres de l'Afrique de l'Ouest. Tome I. Collection Faune et Flore tropicales 40. Institut de Recherche pour le Développement, and Muséum National d'Histoire Naturelle, Paris, France; Musée Royal de l'Afrique Centrale, Tervuren, Belgique. Paris.
• Tracy, H.C. 1920. The clupeoid cranium in its relation to the swimbladder diverticulum and the membranous labyrinth. Journal of Morphology, 33:439-483.
• Webb, J.F. 1989. Gross morphology and evolution of the mechanoreceptive lateral-line system in teleost fishes. Brain, Behavior and Evolution, 33:34-53.
• Weitzman, S.H. 1962. The osteology of Brycon meeki, a generalized characid fish, with an osteological definition of the family. Stanford Ichthyological Bulletin, 8:1-77.
• Weitzman, S.H. 1974. Osteology and evolutionary relationships of the Sternoptychidae, with a new classification of stomiatoid families. Bulletin of the American Museum of Natural History, 153:327-478.
• Weitzman, S.H. & Vari, R.P. 1988. Miniaturization in South American freshwater fishes: an overview and discussion. Proceedings of the Biological Society of Washington, 101:444-465.
• Wohlfahrt, T.A. 1936. Das Ohrlabyrinth der Sardine (Clupea pilchardus Walb.) und seine Beziehungen zur Schwimmblase und Seitenlinie. Zeitschrift für Morphologie und Ökologie der Tiere, 31:371-410.
• Wohlfahrt, T.A. 1937. Anatomische Untersuchungen über die Seitenkanäle der Sardine (Clupea pilchardus Walb.). Zeitschrift für Morphologie und Ökologie der Tiere, 33:381-411.

Publication Dates

• Publication in this collection
  11 Jan 2007
• Date of issue
  2006

History

• Received
  23 Aug 2006
• Accepted
  10 Oct 2006