The Second Ural Centrum in Siluriformes and Its Implication for the Monophyly of Superfamily Sisoroidea (Teleostei, Ostariophysi)

2004·0.71 MB·English

by  Mário C. C. De Pinna

PUBLISHED BY THE AMERICAN MUSEUM OF NATURAL HISTORY
CENTRAL PARK WEST AT 79TH STREET, NEW YORK, NY 10024
Number 3437, 23 pp., 14 figures, 1 table May 14, 2004
The Second Ural Centrum in Siluriformes and Its
Implication for the Monophyly of Superfamily
Sisoroidea (Teleostei, Ostariophysi)
MA´RIO C. C. DE PINNA1 AND HEOK H. NG2
ABSTRACT
Awell-developedseconduralhalf-centrumisapeculiarcharacteristicofthecaudalskeleton
of certain siluriforms, not seen in other Recent ostariophysans. Although the character has
beenpreviouslyrecordedintheliterature,itsexacttaxonomicdistribution,structure,ontogeny,
and phylogenetic meaning within siluriforms are not yet determined. In this paper, the degree
of development of the second ural half-centrum is surveyed across the order. The ontogeny
of the relevant axial structures in the posterior region of the vertebral column is reported in
representative siluriform taxa. The condition where the second ural centrum is well formed
and forms a complete intervertebral joint anteriorly with the compound caudal centrum is
considered derived withinsiluriforms,acharacterstatehomoplasticwiththeprimitivestatein
more distant teleostean outgroups. Various catfishes display that derived condition, which is
informative about relationships at different levels within the group. The most inclusive of
those clades is the superfamily Sisoroidea, including the families Amblycipitidae, Akysidae,
Sisoridae, Erethistidae, and Aspredinidae. The placement of theneotropical Aspredinidaeinto
anotherwiseexclusivelyAsiancladehasimportantbiogeographicalimplicationsandthestruc-
ture of the second ural centrum provides additional support to that hypothesis. Based on the
currently available knowledge on the relationships among catfishes, a well-developed second
ural centrum is hypothesized to be a result of six different events in siluriforms.
INTRODUCTION the Teleostei and its subgroups. Numerous
papers have documented and interpreted the
Thecaudalskeletonhasbeenarichsource structure of the internal supports of the cau-
of taxonomically informative variation for dal fin in bony fishes, both before and after
1Museu de Zoologia,Universidade deSa˜o Paulo,CaixaPostal 42594, Sa˜o Paulo-SP 04299-970Brazil.Research
Associate,DivisionofVertebrateZoology(Ichthyology),AmericanMuseamof NaturalHistory.
2School of Biological Sciences, National University of Singapore, 10 Kent Ridge Crescent, 119260 Singapore.
Presentaddress:MuseumofZoology,UniversityofMichigan,AnnArbor,MI48109,USA.
Copyright(cid:113)AmericanMuseumofNaturalHistory2004 ISSN0003-0082
2 AMERICAN MUSEUM NOVITATES NO. 3437
theadventofphylogeneticthinking(seeMo- lationshipinSouthAmericanfreshwaterfish-
nod, 1968; Schultze and Arratia, 1989 and es. Although a number of trans-Pacific rela-
references therein). This long record of in- tionships are known for the North American
vestigation has made the caudalskeletonone freshwater fish fauna, including both Recent
of the most important sources of characters and Fossil taxa (Grande, 1994: 68–74), no
for elucidating phylogenetic relationships such case was known for South America. In
among teleostean fishes. Lundberg and Bas- the process of the study, taxonomiccoverage
kin (1969) were the first to study in detail was expanded to encompass the whole Sil-
the large degree of variation in the caudal uriformes,andrevealeddatarelevantalsofor
skeleton of the orderSiluriformes,andfound understanding relationships of other sub-
many morphological patterns diagnostic for groups within the order.
monophyletic groups therein. Since then,
caudal skeleton characters have been regu- MATERIALS AND METHODS
larly used in studies of relationships among
Comparative material representing all cur-
catfishes. Despite relatively intense research
rently recognized families of Siluriformes
focusing on the caudal skeleton and its var-
was examined as cleared and stained prepa-
iations within Siluriformes, it appears to us
rations, listed in appendix 1. Representatives
that a number of additional potentially infor-
chosen included the genera considered prim-
mativecharactersstillawaitdiscoveryinthat
itive or basal in their respective families,
complex.
whenever that information is known and
The aim of this paper is to report on a
specimens were available. Most specimens
survey of a specific part of the caudal skel-
were prepared according to the method of
eton of siluriforms, namely the degree of de-
Taylor and Van Dyke (1985), which stains
velopment of the second ural centrum (U ).
2 cartilage and bone, but some were prepared
The full development of the anterior half of
previously and were stained for bone only.
that centrum, where it forms a complete in-
The embryonic specimen of Pterobunoce-
tervertebral joint anteriorly and has a well-
phaluswaspreparedinsidetheegg,bypunc-
defined ossification center, occursconsistent-
turing the egg case before the clearing and
ly in the adults of a few groups of catfishes,
staining procedure. Illustrations were pre-
as first reported by Lundberg and Baskin
pared with the aid of a microvideo system
(1969). This feature is unusual in siluriforms
attached to astereomicroscope.Imagestaken
and unique among other Recent ostariophy-
by the video system were printed with a col-
sans, yet it remains poorly understood as to
or video printer and then traced by hand,
its exact distribution and phylogeneticimpli-
againstcomparisonwiththeactualspecimen.
cations. Accordingly, we here conduct a
Anatomical terminology follows Lundberg
comparative survey of the U in various rep-
2 and Baskin (1969) and, for terms not treated
resentatives of all siluriformfamilies,aiming
therein, Monod (1968).
at documenting the distribution ofthetraitin
as much detail as possible. On the basis of
ANATOMICAL ABBREVIATIONS
that, we propose that the degree of devel-
EP epural
opment of the second ural centrum is a well-
HS hemal spine
defined character informative at various lev-
HY hypural n
els in the phylogeny of siluriforms. n
LHP lower hypural plate
This study started as an evaluation of the
NS neural spine
significance of the second ural centrum for PH parhypural
the monophyly of the superfamily Sisoro- PU preural centrum n
n
idea, defined by de Pinna (1996) to include U ural centrum n
n
the Asiatic catfish families Amblycipitidae, UN uroneural
Akysidae, Erethistidae, and Sisoridae, plus UHP upper hypural plate
the neotropical Aspredinidae. The possible
affinities of Aspredinidae with Asiatic taxa INSTITUTIONAL ABBREVIATIONS
has important biogeographic implications, AMNH American Museum of Natural History,
since it is the first case of a trans-Pacific re- New York
2004 DE PINNA AND NG: URAL CENTRUM IN SILURIFORMES 3
ANSP Academy of Natural Sciences, Phila- morphs, U is most markedly developed in
2
delphia Denticeps clupeoides, sister group to all oth-
BMNH The Natural History Museum, London er Clupeiformes and sole Recent representa-
CAS California Academy of Sciences, San
tiveofsuborderDenticipitoidei.Thecentrum
Francisco
is also well differentiated in the majority of
CMK Collection Maurice Kottelat, Cornol
other clupeomorphs, though usually smaller
DUVC VertebrateCollection,DukeUniversity,
in relative size than that observed in Denti-
Durham
FMNH Field Museum, Chicago ceps. The degree of development of U2 is
MCP Museu de Cieˆncias da Pontif´ıcia Univ- variable in most groups of lower Neoteleos-
ersidade Cato´lica do Rio Grande do tei,includinggroupssuchasaulopiformsand
Sul, Porto Alegre protacanthopterygians (sensu Johnson and
MCZ MuseumofComparativeZoology,Har- Patterson, 1996). A second ural centrum is
vard University, Cambridge not differentiated (or not independent) in
MNHN Museum National d’Histoire Naturelle,
adult acanthopterygians. Among ostariophy-
Paris
sans, a well-developed autogenous second
MRAC Musee Royal de l’Afrique Centrale,
ural centrum is present in some fossil gono-
Tervuren
rynchiforms, such as Dastilbe, Tharrhias,
MZUSP MuseudeZoologiadaUniversidadede
Sa˜o Paulo, Sa˜o Paulo and Parachanos (Poyato-Ariza, 1996), but
NRM Swedish Museum of Natural History, not in Recent adult gonorynchiforms. Chan-
Stockholm os has bilateral flanges extending dorsopos-
RUSI South African Institute for Aquatic teriorlyfromitscomplexcentrum(illustrated
Biodiversity, formerly the JLB Smith in Poyato-Ariza, 1996: fig. 19 and labeled
Institute of Ichthyology, Grahamstown ‘‘postero-lateral process of caudal terminal
UMMZ Museum of Zoology, University of
complex’’). Topological similarities suggest
Michigan, Ann Arbor
that these structures may be remnants of the
USNM National Museum of Natural History,
urostyle (sensu Monod, 1968) of a primitive
Smithsonian Institution, Washington
U , itself fused or greatly reduced and ad-
D.C. 2
ZSM Zoologische Staatssammlung, Munich pressed to the posterior margin of PU1 (cid:49) U1.
A well-developed U is also present in some
2
siluriforms, as detailed in this paper, but not
THE SECOND URAL CENTRUM
in other otophysans.
IN SILURIFORMES AND
Lundberg and Baskin (1969) recognized
OTHER TELEOSTEI
forthefirsttimethatafewSiluriformeshave
The ural centra in the caudal skeleton of awell-developedU asadults.Inthosecases,
2
adult teleosts are often compound structures the centrum forms a normal vertebral joint
formedontogeneticallybythefusionofauto- anteriorlywiththecompoundcaudalcentrum
and chordacentra precursors corresponding (PU (cid:49) U ), with a clearly recognizablecon-
1 1
to multiple individual centra. It has been ar- ical articular surface intermediated with car-
gued that the so-called second ural centra in tilage. Since its discovery, the presence of a
adult teleosts can be formed ontogenetically well-developed U has been consideredtobe
2
in markedly different ways and are not nec- of phylogenetic significance. Lundberg and
essarily homologous in various lineages Baskin observed that all other adult ostario-
(Schultze and Arratia, 1988). A structure physans lack a U autocentrum, and that its
2
identifiable as a second ural centrum (U ) is presenceinsomecatfishesindicatesthat‘‘the
2
well developed in adults of the majority of potential for its independent redevelopment
lower teleosts, and this is probably the prim- has not been lost, at least in Siluriformes’’
itiveconditionforthegroup.Suchisthecase (Lundberg and Baskin, 1969: 17). In light of
in most, or at least basal members of, osteo- the phylogeny currently accepted for ostar-
glossomorphs, elopomorphs, clupeomorphs, iophysans (Fink and Fink, 1996) and lower
and esocoids (Monod, 1968; Schultze and teleosts (Johnson and Patterson, 1996; Arra-
Arratia, 1988; Arratia and Schultze, 1992), tia, 1997a, 1997b), it seems certain that the
as well as most basal fossil teleosts (Patter- presence (or the redevelopment) of a well-
son and Rosen, 1977). In Recent clupeo- developed U in some catfishes shouldbein-
2
4 AMERICAN MUSEUM NOVITATES NO. 3437
terpreted asapomorphic.TheDiplomystidae, nous U (Hilton, 2002) similar to thatin bas-
2
considered as sister group to all other fossil al Elopomorphs (such as Elops) and quite
and Recent siluriforms (Lundberg and Bas- differentfromthesiluriformcondition.Some
kin, 1969; Grande, 1987; Arratia, 1987; Mo, paracanthopterygians, such as percopsiforms
1991; de Pinna, 1998), do not have a well- (Rosen and Patterson, 1969: fig. 16) also
developed U as adults. Likewise, species of have a condition similar to that in siluri-
2
the fossil family Hypsidoridae,consideredas forms, though there may be some question
basal siluriforms, also lack an identifiableU as to the actual homology of the second ural
2
(Grande, 1987; Grande and de Pinna, 1998). centrum in paracanthopterygians.
Finally, no instances of a well-developed U The well-developed condition of the U in
2 2
areknowningymnotiforms,characiforms,or catfishes can be associated with some inter-
cypriniforms, closest relatives of siluriforms. esting structural modifications. In sometaxa,
Within Siluriformes, Lundberg and Baskin such as aspredinids and akysids, the U and
2
reported the presenceof awell-developedU associated upper hypural plate are extremely
2
in Helogeninae (Cetopsidae), Aspredinidae, well developed, and are oriented nearly on
Amblyceps (Amblycipitidae),andsomeChil- the horizontal axis. In such cases, it seems
oglanis (Mochokidae). The authors consider like the upper hypural plateistheactualcau-
that each of thosehaveclosestrelativeslack- dal-finsupport,whilethelowerhypuralplate
ing a full U , indicating that thestructurehas appears superficially to be a large hemal
2
been redeveloped four independent times in spine. This arrangement must have impor-
siluriforms. tant, yet unexplored, biomechanical impli-
The U in siluriforms, when well devel- cations in the movement ofthe upperlobeof
2
oped and in adults, is always entirely fused the caudal fin in those fishes.
and continuous with the upper hypural plate. The degree of U development in adults
2
No siluriforms have a U independent of hy- varies markedly among various catfish taxa.
2
pural ossifications like in the primitive con- Three basic types can be recognized:Type1:
dition of other lower teleosts (which also in- there is little or no trace of U ossification.
2
cludemanycasesoffusion).Thesecondural Hypural 3 (or the upper hypural plate) has
centrum in siluriforms, when present, tapers no obvious proximal thickening and fits into
posteriorly, is fused with one or more of the a cavity on the posterodorsal end of PU (cid:49)
1
upper hypurals, and does not have an artic- U , immediately ventral to the base of the
1
ular surface on its posterior end. It should, uroneural (fig. 1). This is the condition pre-
thus, be more rigorously called a half-cen- sent in most catfishes and other ostariophy-
trum, though for brevity in this paper it is sans. Type 2: the proximal end of hypural 3
just called a centrum. Also, it always lacks or upper hypural plate has a small bony
the peculiar dorsoposterior process (oftentu- thickening, often having a tight contact, or
bularorsemitubular)calledaurostyle(inthe even forming an incipient articulation, but
terminology of Monod, 1968). A vestige of without a well-defined cartilage disk, with
the urostyle may be represented by a ridge the posterior portion of PU (cid:49) U (fig. 2).
1 1
along the middorsal line of the centrum and Type 3: a well-formed, obvious vertebral
of the associated hypural. In some aspredi- centrum is attached to the base of hypural 3
nids (such as Pterobunocephalus), the ridge or a compound upper hypural plate. This
may be quite deep anteriorly, indeed resem- centrum has a well-defined articulation an-
bling an attenuated urostyle. This ridge ac- teriorly with the posterior portion of PU (cid:49)
1
commodates the ventral margin of the uro- U , intermediated by cartilage, and a conical
1
neurals and seems to form a floor for the op- ossification center visible by transparency
isthural cavity. Overall, the condition of the (fig. 3). In anterior view, the similarity in
caudal skeleton in catfishes with a well de- structure between the U and a normal ver-
2
veloped U is markedly similar to that in tebral centrum is evident (fig. 4).
2
some osteglossomorphs, such as Pantodon, Type 1 U is present in outgroup ostario-
2
Scleropages, and Heterotis (cf. Monod, physans, and is considered as the primitive
1968).Basalmembersofosteoglossomorphs, stateforcatfishes.Type2seemstobesubject
such as Hiodon, however, have an autoge- to considerable degree of intraspecific vari-
2004 DE PINNA AND NG: URAL CENTRUM IN SILURIFORMES 5
Fig. 1. Caudal skeleton ofDiplomystesmesembrinus(Diplomystidae;MZUSP62595,149mmSL).
Scale bar (cid:53) 1 mm.
ation; its delimitation as a discrete character
state is difficult and we consider its phylo-
genetic significance to be uncertain at this
point. In fact, the homology of the bony
thickening at the base of the upper hypural
with a ural centrum remains to be directly
demonstrated by ontogenetic data. Type 3 is
a condition derived within siluriforms; it
seems to form a well-defined characterstate,
with few cases of intermediate conditions
and little or no relevant intraspecific varia-
tioninthesamplesexamined.Althoughmore
in-depth studies may find a basis for consid-
ering U types 1, 2, and 3 as a multistate
2
characterorderedinthatsequence,wedonot
considerthattobewarrantedatpresent.Only
type 3 can currently be considered as objec-
tive evidence of relationship.
ONTOGENETIC DATA
Fig. 2. Caudal skeleton of Helicophagus
waandersii(Pangasiidae;UMMZ186797,77mm The ontogenetic fate of the second ural
SL). Scale bar (cid:53) 1 mm. centrum seems to differ in various catfish
6 AMERICAN MUSEUM NOVITATES NO. 3437
Fig. 3. Caudal skeletons in Akysidae: A. Akysis recavus (MZUSP 75128, 24 mm SL), B. Breiten-
steinia insignis (AMNH 58378, disarticulated adult specimen), C. Parakysis grandis (CMK 7915, 44
mm SL). Scale bar (cid:53) 1 mm.
taxa. Lundberg and Baskin (1969) showed Fujita (1992) and Ichiyanagi and Fujita
that the most common condition in siluri- (1995), on the basis of observations on spe-
formsistohaveaU ,orsomeremnantthere- cies of Clariidae (Siluriformes) and Cobiti-
2
of, fused to the bases of hypurals 3 and 4, dae (Cypriniformes), respectively. In their
but noticed that in trichomycterids it isfused view,thestructurecalledPU (cid:49)U byLund-
1 1
to the base of hypural 3 only. According to berg and Baskin (1969) was considered to
those authors, any remnants of the second include the second ural centrum as well (as
ural centrum are associated with the base(s) PU (cid:49) U ). Thus, the vertebral centrum
1 1(cid:49)2
of upper hypurals, and not with the com- material associated with hypural 3, consid-
pound centrum. Arratia (1983) confirmed ered by Lundberg and Baskin to be the U ,
2
that U is fused to the base of hypural 3 in would actually be homologous to a more
2
some Trichomycterinae, but argued that in posterior ural centrum (U ). This idea was
3
Nematogenyidae it is instead fused to the notbasedondirectobservationoffusion,but
posterior portion of the compound centrum. apparently inferred on the basis of the joint
An alternative interpretation, similar to Ar- association of parhypural, hypural 1, and hy-
ratia’s on Nematogenyidae,wasadvancedby pural 2 to a single centrum element during
2004 DE PINNA AND NG: URAL CENTRUM IN SILURIFORMES 7
tariophysans and other teleosts is available.
We would further note that hypural 1 origi-
nates,inatleastonesiluriform(Fujita,1992)
and one cypriniform (Ichiyanagi and Fujita,
1995), as a basal anteroventral offshoot of
the cartilaginous hypural 2, rather than as an
independent element (this information dis-
agrees with Arratia, 1997b: 309, who stated
that hypurals 1 and 2 are not connected by
cartilage in juvenile ostariophysans). This is
perhaps also the case in juveniles of many
other teleosts,wherecartilaginoushypurals1
and 2 are fused attheirbasesatastagewhen
all other hypurals are independent (cf. Mo-
nod,1968:figs.30,32,61;PattersonandRo-
sen, 1977: figs. 26, 28). Although separate
chondrification and subsequent cartilage fu-
sion in early development has been docu-
mentedforsometaxa(cf.Fujita,1994,inthe
cichlid Tilapia and Arratia and Schultze,
1992, in the Salmonid Onchorhynchus), it
seems possible that hypural 1 in many, per-
haps most, teleosts isjustanontogeneticoff-
shoot from hypural 2. If so, it should not be
expected to be associated with its own cen-
Fig. 4. Caudal skeleton in anterior view: A.
trum in the vertebral series, and the centrum
Parakysis grandis (Akysidae; MZUSP 63109,
thatsupportshypurals1and2inthemajority
37.8 mm SL), B. Agmus sp. (Aspredinidae;
AMNH 97160, disarticulated adult specimen). of teleosts would indeed be a singleelement.
Scale bars (cid:53) 1 mm. In that case, the compound centrum for the
parhypural plus hypurals 1 and 2 in siluri-
formsandotherostariophysanswouldnotin-
ontogeny, thereby implying that this element volve three or more centra, but only two
iscompoundfromthreeprimitivecentra.The (PU (cid:49) U ), as normally accepted. However,
1 1
compound nature of the teleostean U has general conclusions along these lines need
1
been suggested in the past (Nybelin, 1971) further information on early chondral differ-
on the basis of its association with hypurals entiation of the elements of the caudal skel-
1 and 2 in the vast majorityofteleosts,fossil eton,notyetavailableformostrelevanttaxa.
and Recent (Patterson and Rosen, 1977; de The development of the second ural cen-
Pinna, 1996; Arratia, 1997b). Schultze and trum in catfishes in general is still poorly
Arratia(1988)wentfurther,andproposednot documented, and current ideas are based on
only that U was formed by fusion of two observations on few taxa. Ontogeny of U in
1 2
centra (as U ), but also that U was formed groups where it displays a type 3 degree of
1(cid:49)2 2
by fusion of three other centra (as U ). development is virtually unknown, in great
3(cid:49)4(cid:49)5
Such complex fusions have not been corrob- part due to the difficulty in obtaining growth
orated by other authors (Hilton, 2002), and series of the relevant taxa. Juveniles of the
have also not been confirmed in the material superfamily Sisoroidea are rare in collec-
examined for this paper. While some of the tions. For this study, we obtained juvenile
individualuralandpreuralcentramayindeed sisoroid material of Gagata (Sisoridae) and
be compound in origin, evidence for that is Aspredo (Aspredinindae), plus late embry-
controversial. Therefore, we prefer to main- onic material of Pterobunocephalus (Aspre-
tain theterminologyofLundbergandBaskin dinidae). Although no growth series were
until more direct evidence of the exact com- available, observation of juvenile conditions
positionofthecompounduralcentruminos- is nonetheless informative about the forma-
8 AMERICAN MUSEUM NOVITATES NO. 3437
Fig. 5. Ontogenetic changes in the second ural centrum of Aspredinidae. A. Pterobunocephalussp.
(MZUSP 57208, approx. 5.5 mm SL), late embryo inside eggshell, B. Pterobunocephalus sp. (MZUSP
uncat., 63.0 mm SL), adult. Heavy stippling represents cartilage, light stipling represents a thin miner-
alized layer, and open areas represent notochord (in A) or bone (in B). Scale bars (cid:53) 0.5 mm (A) and
1 mm (B).
tion of theseconduralcentruminthosetaxa. the possibility that the structure identified as
In the smallest specimen examined, a Pter- U might be a centrum-like specialization of
2
obunocephalus embryo still inside an egg- the proximal part of the upper hypural com-
shell, hypurals 3 and 4 are already fused to plex, that is, not really a centrum,butsimply
U (fig. 5A). Still, there is little doubt that a hypural modification gross-morphological-
2
the structure is actually a ural centrum, be- ly similar to a centrum in adults.
cause it forms a continuous series with other In catfishes without a well-developed U ,
2
centra anterior to it. Concomitantly, its ho- very little centrum material, if any, is asso-
mology with therespectivecentruminadults ciated with the bases of hypurals 3 and 4. In
is also obvious (fig. 5B). Similar situations very small juvenile Trichogenes longipinnis
are seen in later-stage juveniles of Gagata
(Trichomycteridae) examined, no ossifica-
(fig. 6) and Aspredo (MZUSP 77694). The
tions can be clearly identified as U material
configurations in these specimens rule out 2
duringtheossificationofthecaudalskeleton.
At the earliest stage available (fig. 7A), the
limits between centra can be seen as mem-
branous separations, and the region corre-
sponding to U is independent of any hypur-
2
als. It remains so as ossification progresses
(fig.7B).WhateverlittleU materialthatgets
2
ossified seems to be incorporated into the
posterior region of the compound centrum
and/or the mesial surface of the uroneural.
The bone thickening at the base of hypural
3 in adults is entirely composed of hypural
material, since an evident thickening is al-
ready evident in the cartilaginous phase of
the hypural, before the onset of ossification.
ThesamesituationholdsforjuvenileClarias
gariepinnus (NRM 15319). Thus, the gen-
Fig. 6. Caudal skeleton of juvenile specimen
erally held assumption that U is fusedtothe
of Gagata melanopterus (Sisoridae; MZUSP 2
bases of hypurals 3 (trichomycterids) or 3
52865, 16.0 mm SL). Stippling represents carti-
lage, open areas represent bone or notochord (in and 4 (all other catfishes) does not seem on-
dotted lines). Scale bar (cid:53) 1 mm. togenetically corroborated in the taxa exam-
2004 DE PINNA AND NG: URAL CENTRUM IN SILURIFORMES 9
Fig. 7. Development of the caudal skeleton of Trichogenes longipinnis (Trichomycteridae;MZUSP
80933). A. at 9.1 mm SL, B. at 18.2 mm SL. Stippling represents cartilage, open areas represent
membrane (A) or bone (B). Scale bars (cid:53) 0.5 mm (A) and 1 mm (B).
ined. The fateofU materialmaynotbeuni- AKYSIDAE
2
form among various catfish taxa.
A previously unreported fully developed
Also, we found only little evidence of fu-
U forming an intervertebral joint is present
sion between PU and U . The region cor- 2
1 1 and very conspicuous in all akysids exam-
responding to the twocentrabeforetheonset
ined, and is the primitive condition for the
of ossification is always continuous in all ju-
family (fig. 3). In all cases, the second ural
venilesexamined.Ontheotherhand,alarger
centrum is very well developed, and almost
juvenile of Trichogenes (fig. 7B) shows a
the same size as the anterior half of the sec-
tenous line of suture at the expected limit
ond preural vertebra. Also, the centrum and
between PU and U which was not evident
1 1 associated upper hypural platearepositioned
before ossification (fig. 7A). Perhaps fusion
nearly along the longitudinal axis of the ver-
of the precursors of the two centra occurs
tebral column.
very early in development, yet beginning of
ossificationstillrevealsavestigeoftheprim-
itive separation between them. In adults, fu- AMBLYCIPITIDAE
sion is again complete.
Lundberg and Baskin (1969) reported
some variation in U development in this
THE SECOND URAL CENTRUM IN 2
family, with Liobagrus showing a well-de-
SILURIFORM FAMILIES
veloped condition and Amblyceps lacking it.
We were able to examine several taxa not Our observations confirm theirs. In Lioba-
available to other authors and to verify that grus reini, three of the four specimens ex-
U forms a full vertebral joint anteriorly in amined show a very conspicuous full devel-
2
several additional catfish taxa not previously opment of U (fig. 8), while a fourth speci-
2
recorded. Appendix 1 summarizes our re- men shows a rather reduced condition. In
sults,withalistofthematerialexaminedand Amblyceps, the centrum is poorly developed.
an indication of the degree of development WedidnotconfirmtheobservationsbyChen
of the second ural centrum for Siluriforms (1994: 111) that the U is fused to the com-
2
(types 1, 2, or 3; see above). Below we pro- pound centrum in Amblyceps. In the speci-
vide a discussion on each of the catfish fam- mens we examined, the two are clearly sep-
ilies known to have at least one instance of arate, and the U is reduced. However, there
2
a well-developed U . seems to be some variation in this fusion
2
10 AMERICAN MUSEUM NOVITATES NO. 3437
Fig. 8. Caudal skeleton of Liobagrus anguil-
licaudatus (Amblycipitidae; AMNH 11069, dis- Fig. 9. Caudal skeleton of Leptoglanis brevis
articulated adult specimen). Scale bar (cid:53) 1 mm. (Amphiliidae,UMMZ199990,29mmSL).Scale
bar (cid:53) 1 mm.
among amblycipitids, since Lundberg and
Baskin (1969: 17) reported that 3 out of 29 mus (fig. 10A), the limit between U and the
2
specimens examined of Liobagrus anguilli- compound centrum is partly covered by ad-
caudatusshowed fusionbetweenU andPU ditional ossification, but still clearly visible.
2 1
(cid:49) U . In Xiurenbagrus, sister group to all In representatives of the tribe Hoplomyzon-
1
other amblycipitids (Chen, 1994; Chen and tini examined here (Ernstichthys and Hoplo-
Lundberg, 1995), U is reported as well de- myzon), the whole portion posterior to the
2
veloped and forming a full intervertebral compound centrum is heavily covered by
joint with the compound centrum (Chen, thick superficial ossification, and a separate
1994: 111; Chen and Lundberg, 1995: 790). U is not evident. Considering that all other
2
The presence of a well-developed U in Xiu- aspredinid representatives have a well-devel-
2
renbagrus and Liobagrus indicates that this oped U , and furtherthathoplomyzontinsare
2
is the plesiomorphic condition for amblycip- not basalin thefamily(Friel,1994),itseems
itids, and that the reducedstateinAmblyceps thattheabsenceofanevidentU inthattribe
2
is a reversal. issecondary,andthattheprimitivecondition
forthefamilyAspredinidaeistohaveawell-
AMPHILIIDAE developed second ural centrum. It is likely
that the condition in hoplomyzontins is a re-
A well-developed U was previously un-
2 sult simply of additional ossifications in the
reported in amphiliids. We found the struc-
caudal skeleton, and that the normal aspre-
ture forming an anterior intervertebral joint
dinid condition is present but obscured in
in Leptoglanis brevis (fig. 9) and Zairei-
members of that tribe. Intermediate condi-
chthys zonatus.Similaryetslightlylesswell-
tions such as that in Agmus demonstrate that
developed conditions are seen in the very
such a configuration is likely. The condition
similar L. rotundiceps and in L. camerunen-
in the basal undescribed genus which is the
sis. All other amphiliids examined,including
sister group to all other aspredinids (referred
L. xenognathus, lack a well-developed sec-
to as ‘‘Pseudobunocephalus’’by Friel,1994)
ond ural centrum.
seemstobesimilartothatinhoplomyzontins
(cf. Friel, 1994: fig. 27). The condition of U
ASPREDINIDAE 2
in such cases should be examinedinjuvenile
Most aspredinids examined have a well- specimens, beforethe onsetofsuperficialos-
developed U (fig. 10). In some, such as Ag- sification on the caudal skeleton. Lundberg
2