the hyal and ventral branchial muscles in caecilian and salamander larvae: homologies and evolution

The Hyal and Ventral Branchial Muscles in Caecilianand Salamander Larvae: Homologies and Evolution

Thomas Kleinteich1,2* and Alexander Haas1

1Biozentrum Grindel und Zoologisches Museum, Universitat Hamburg, Hamburg 20146, Germany2Friday Harbor Laboratories, University of Washington, Friday Harbor, Washington 98250

ABSTRACT Amphibians (Lissamphibia) are character-ized by a bi-phasic life-cycle that comprises an aquaticlarval stage and metamorphosis to the adult. The ancestralaquatic feeding behavior of amphibian larvae is suctionfeeding. The negative pressure that is needed for ingestionof prey is created by depression of the hyobranchial appa-ratus as a result of hyobranchial muscle action. Under-standing the homologies of hyobranchial muscles in am-phibian larvae is a crucial step in understanding the evolu-tion of this important character complex. However, theliterature mostly focuses on the adult musculature andterms used for hyal and ventral branchial muscles in dif-ferent amphibians often do not reflect homologies acrosslissamphibian orders. Here we describe the hyal and ven-tral branchial musculature in larvae of caecilians (Gymno-phiona) and salamanders (Caudata), including juveniles oftwo permanently aquatic salamander species. Based onprevious alternative terminology schemes, we propose aterminology for the hyal and ventral branchial musclesthat reflects the homologies of muscles and that is suitedfor studies on hyobranchial muscle evolution in amphib-ians. We present a discussion of the hyal and ventral bran-chial muscles in larvae of the most recent common ancestorof amphibians (i.e. the ground plan of Lissamphibia).Based on our terminology, the hyal and ventral branchialmusculature of caecilians and salamanders comprises thefollowing muscles: m. depressor mandibulae, m. depressormandibulae posterior, m. hyomandibularis, m. branchio-hyoideus externus, m. interhyoideus, m. interhyoideus pos-terior, m. subarcualis rectus I, m. subarcualis obliquusII, m. subarcualis obliquus III, m. subarcualis rectus II-IV,and m. transversus ventralis IV. Except for the m. bran-chiohyoideus externus, all muscles considered herein canbe assigned to the ground plan of the Lissamphibia withcertainty. The m. branchiohyoideus externus is either apo-morphic for the Batrachia (frogs 1 salamanders) or sala-mander larvae depending on whether or not a homologousmuscle is present in frog tadpoles. J. Morphol. 272:598–613, 2011. � 2011 Wiley-Liss, Inc.

KEY WORDS: Lissamphibia; cranial muscle homology;Batrachia hypothesis; amphibian larvae; caecilians

INTRODUCTION

Recent hypotheses on the relationships withinthe extant amphibians (Lissamphibia) suggest thatcaecilians (Gymnophiona) are the sister taxon tosalamanders (Caudata) plus frogs (Anura) (Batra-chia hypothesis; Trueb and Clothier, 1991; Zardoya

and Meyer, 2000, 2001; Frost et al., 2006; Roelantset al., 2007). The three groups of extant amphib-ians are characterized by a complex life cycle withlarvae and metamorphosis and it is most parsimo-nious to assume that the most recent commonancestor of amphibians had an aquatic larval stage(Wake, 1993; Duellman and Trueb, 1994; Wilkinsonet al., 2002; Schoch, 2009). All salamander larvaestudied so far use suction for prey capture (Debanand Wake, 2000; Deban et al., 2001; O’Reilly et al.,2002). In caecilians, suction feeding has been docu-mented for larvae of a species within the genusEpicrionops (Rhinatrematidae) and is hypothesizedto be the feeding mode in other caecilian taxa withaquatic larvae (O’Reilly, 2000; O’Reilly et al., 2002).Frog tadpoles use mucous entrapment suspensionfeeding, which, however, is a derived feeding modewithin amphibian larvae (O’Reilly et al., 2002).

The negative pressure needed for suction feedingis generated by depression of the hyobranchial ap-paratus and by virtue of hyal and branchial muscu-lature contraction (Deban and Wake, 2000). Besidessuction, the hyobranchial musculature is importantfor tongue movements and ventilation and thus hasa double function in feeding and breathing (Wake,1982; Roth and Wake, 1985). A previous study onthe hyobranchial apparatus in anuran tadpoles(Haas, 1997) showed the transformations of larvalhyobranchial characters in anuran evolution; par-ticularly, the ventral branchial muscles of the sub-arcualis muscle system showed a substantial degreeof character state evolution.

Homologies among muscles of the hyobranchial ap-paratus in amphibian larvae are obscured by alterna-

Contract grant sponsor: German Research Foundation (DFG);Contract grant number: HA2323/10-1.

*Correspondence to: Thomas Kleinteich, University of Washing-ton, Friday Harbor Laboratories, 620 University Road, Friday Har-bor, WA 98250. E-mail: [email protected]

Received 4 August 2010; Revised 31 October 2010;Accepted 29 November 2010

Published online 3 March 2011 inWiley Online Library (wileyonlinelibrary.com)DOI: 10.1002/jmor.10940

JOURNAL OF MORPHOLOGY 272:598–613 (2011)

� 2011 WILEY-LISS, INC.

tive, synonymous schemes of muscle names. The di-versity in muscle terminology is based historically onthe fact that comparative anatomists at the begin-ning of the last century mainly used the terms asdescriptions of either topology or function (see Hoyosand Dubois, 2004). Recent studies on the cranialmusculature in amphibians (Bauer, 1997; Haas,1997, 2001, 2003; Kleinteich and Haas, 2007; Diogoet al., 2008a,b) apply terms that were mostly derivedfrom the studies by Druner (1901, 1904) and Edge-worth (1920, 1935). In Druner’s (1901, 1904) andEdgeworth’s (1935) terminologies, homologousmuscles that differed in function or topology betweenspecies could be assigned different names. Today,morphological character states feed into analyses onmorphological character evolution (Rieppel and Kear-ney, 2002; Haas, 2003); names of structures can serveas semantic proxies for homology statements in clad-istics, i.e. primary homologies (De Pinna, 1991;Brower and Schawaroch, 1996; Hoyos and Dubois,2004; but see Vogt et al., 2010 for a different view).The lack of well-founded hypotheses on homologymakes it difficult to apply the terminology of Druner(1901, 1904) or Edgeworth (1935) to questions onmorphological evolution in a phylogenetic context.

In amphibians, assignment of primary homolo-gies (sensu De Pinna, 1991) to the hyobranchialmusculature is further complicated by the factthat the divergence of the three amphibian groupsis supposed to be very old (351 – 266 mya; Marja-novic and Laurin, 2007) and each lineage hasevolved highly specialized morphologies (Carroll,2007). Further, previous studies on the morphologyand evolution of hyal and branchial muscles inamphibian larvae (Fischer, 1864; Druner, 1901,1904; Litzelmann, 1923; Edgeworth, 1920, 1935;Piatt, 1938, 1939, 1940; Fox, 1959; Bauer, 1997;Haas, 1997, 2003) have neglected caecilian larvae,although their phylogenetic position as sister-group to the Batrachia (Frost et al., 2006; Roelantset al., 2007) is crucial to explore questions on theevolution of the hyobranchium in amphibians.Larval caecilian musculature is only partially cov-ered in Edgewoth (1920, 1935) and Muller (2007);a list of hyobranchial muscles in larvae of Epicrio-nops bicolor was provided by Wake (1989) in astudy on the development of the skeletal elementsof the hyobranchium. Only recently, the cranialmusculature was described for Ichthyophis koh-taoensis and included information on all cranialmuscles, except for the eye muscles, in a larvalcaecilian (Kleinteich and Haas, 2007).

The present study describes the hyal and ven-tral branchial muscles in larvae of the caecilianEpicrionops bicolor, a species within the Rhinatre-matidae, the sister-group to the remainder caecil-ians (San Mauro et al., 2004, 2005; Frost et al.,2006; Roelants et al.; 2007, Zhang and Wake,2009a). Further, we examined the hyal and ventralbranchial muscles in salamander larvae of the spe-

cies Salamandrella keyserlingii (Hynobiidae) andDesmognathus quadramaculatus (Plethodontidae),and in juvenile specimens of the paedomorphicspecies Siren intermedia (Sirenidae) andAmphiuma means (Amphiumidae). The aims ofthis study are to 1) to reevaluate the terms forhyal and ventral branchial muscles in amphibianlarvae that were mainly based on the studies byDruner (1901, 1904) and Edgeworth (1935) and toestablish a terminology that can be universallyapplied to caecilians and salamanders and 2) toinfer the hyal and ventral branchial muscles inthe most recent common ancestor of amphibians.

MATERIALS AND METHODS

We examined larval specimens of the caecilian Epicrionopsbicolor Boulenger, 1883 (Gymnophiona: Rhinatrematidae), lar-vae of the salamander species Salamandrella keyserlingiiDybowski, 1870 (Caudata: Hynobiidae) and Desmognathusquadramaculatus (Holbrook, 1840) (Caudata: Plethodontidae),and juveniles of the neotene salamander species Siren interme-dia Barnes, 1826 (Caudata: Sirenidae) and Amphiuma meansGarden in Smith, 1821 (Caudata: Amphiumidae). Table 1 con-tains a list of specimens that were examined herein.

Specimens have been available as serial sections, enzymecleared and stained animals, or were dissected manually. Table 1shows the preparation techniques that were applied for eachspecimen. Specimens MHW341 and MHW367 have been verti-cally bisected along the body axis before sectioning; the twohalves of the body were serially sectioned in different planes ofsection. Serially sectioned animals were stained in Azan standardstain (Bock, 1989). The serial sections of the Epicrionops bicolorspecimens were stained alternating in Picro Ponceau, Haema-laun Eosin, and Azan stain. Enzyme clearing and staining fol-lowed the procedure in Dingerkus and Uhler (1977); bones werestained red with Alizarin red, cartilages were stained blue byAlcian blue. Before dissection by hand, the cartilages of speci-mens ZMH A09702 and ZMH A10831 were stained blue withAlcian blue by following the first steps of the Dingerkus andUhler (1977) protocol but not taken to the enzyme macerationstep. Pencil drawings of the dissected specimens were created ata dissecting microscope and redrawn on a computer with the vec-tor graphics software Inkscape 0.45 (open source).

A 3D computer reconstruction was built from serial histologi-cal sections for Salamandrella keyserlingii ZMH A09801. Digi-tal photographs were taken for every third section of the speci-men with a Canon PowerShot S50 digital camera that wasmounted on a Leica MZ 9.5 microscope. The general procedureof digitizing the structures in histological sections, alignment ofthe contours, and modeling of objects followed Haas and Fischer(1997). We used Alias1 MayaTM 6.0 for drawing contour lines,alignment, modeling and rendering. The enzyme cleared andstained specimens of S. keyserlingii were used as reference forthe alignment of contours and for comparison during the model-ing process. Bones, cartilages, and muscles have been recon-structed for the S. keyserlingii specimen; teeth, although pres-ent, have been omitted for 3D modeling.

Innervation often is crucial to identify muscles and to estab-lish homologies. Cranial nerve innervations of cranial musclesare well established in amphibians (Fischer, 1864; Strong, 1895;Edgeworth, 1935; Bauer, 1997; Haas, 1997; Schlosser and Roth,1995, 1997) and our study takes advantage of this available in-formation; however, we confirmed the innervation patterns byidentification of cranial nerves in the serial sections examined.

There has been a long and still ongoing debate on how todefine the term ‘‘homology’’ and what the criteria are to identifyhomologies (reviewed by Hoßfeld and Olsson, 2005). Here wefollow the approaches by de Pinna (1991) and Brower and Scha-

AMPHIBIAN HYOBRANCHIAL MUSCLE HOMOLOGIES 599

Journal of Morphology

waroch (1996) (see also Richter, 2005). De Pinna (1991) sug-gested differentiating between primary homologies, which areuntested homology hypotheses and secondary homologies, whichresult from phylogenetic reconstructions and can either be syn-apomorphies or synplesiomorphies of different taxa. Brower andSchawaroch (1996) pointed out that identification of primaryhomologies is a two-step process that involves first the identifi-cation of topographical identities and then second characterstate identities. We identify topographical identities of cranialmuscles herein by considering 1) the innervation patterns, 2)the relationships to other cranial elements and muscles, 3) ori-gins and insertions, 4) the ontogeny, and 5) conjunction (i.e.,multiple homologues may not exist in the same organism;Schuh, 2000). Although individual criteria might be differentbetween corresponding structures in different taxa, the mosaic-like integration of our criteria and the most parsimonious inter-pretation of the results allows to identify primary homologies;e.g., if muscles differ in their origin between different taxa butare otherwise similar (innervation, position relative to othermuscles and skeletal structures, ontogeny) and if they pass theconjuncture test, we still consider them topographical identicaland presence of these muscle is hypothesized to be homologous(i.e., a primary homology). In this study, we do not go beyondthe state of primary homologies because we did not include aphylogenetic analysis.Currently, there is no terminology for hyal and ventral bran-

chial muscles available that can be applied for caecilians andsalamanders (or Batrachia) simultaneously. Recent works toinfer tetrapod muscle homology (Diogo et al., 2008a,b) havetheir merits, but leave out the specific terminology and homol-ogy problems within the Lissamphibia. Table 2 shows a list ofsynonyms for the hyal and ventral branchial musculature; syn-onymous terms are discussed in the discussion section of thischapter.There has been a long debate, whether the distalmost ele-

ments of the branchial arches in salamanders are homologousto the epi- or the ceratobranchials of other vertebrates. Hereinwe follow Reilly and Lauder (1988) who concluded that themost medial paired elements in the hyobranchial apparatus ofsalamanders are homologous to hypobranchials and that thedistal elements are homologous to ceratobranchials. For otherskeletal terms we follow Duellman and Trueb (1994).

RESULTSM. Depressor Mandibulae Group

The m. depressor mandibulae group comprisesfour muscles; i.e., the m. depressor mandibulae,

the m. depressor mandibulae posterior, the m.hyomandibularis and the m. branchiohyoideusexternus. Muscles of this group are innervated bycranial nerve VII (n. facialis).

Caecilians. In Epicrionops bicolor, the m. de-pressor mandibulae originates from the lateralface of the squamosal, the dorsal surface of the pa-rietal, and parts of this muscle originate from thetrunk fascia. The m. depressor mandibulae in E.bicolor larvae inserts along the dorsal edge of theretroarticular process of the lower jaw (Fig. 1A).

The m. depressor mandibulae posterior in Epi-crionops bicolor has its origin at the lateral side ofthe otic capsule and at the dorsal trunk fascia. Itinserts distally on the ceratohyal and wrapsaround the dorsal tip of the ceratohyal like a hood(Fig. 1B). Some fibers however attach to the dorsaledge of the retroarticular process of the lower jaw,medial to the m. depressor mandibulae.

The m. hyomandibularis is a voluminous musclethat originates from the lateral face of the cera-tohyal in Epicrionops bicolor (Fig. 1B). The m.hyomandibularis inserts ventrolaterally on thepseudoangular along an area that reaches fromthe mandibular joint to the caudal tip of the retro-articular process (Fig. 1A). The m. ceratoyhoideusexternus is absent in larval E. bicolor.

Salamanders. In all salamanders examinedherein, the m. depressor mandibulae and the m.depressor mandibulae posterior are only incom-pletely separate from each other. The m. depressormandibulae posterior can be identified as a layer ofmuscle fibers that lies close to the mediocaudalside of the m. depressor mandibulae (Fig. 2). Bothmuscles originate from the lateral face of the squa-mosal and the otic capsule. In Amphiuma meansand Siren intermedia, the m. depressor mandibulaeinserts on the dorsal edge of the retroarticularprocess immediately caudal to the mandibularjoint (Fig. 3); in Salamandrella keyserlingii andDesmognathus quadramaculatus the m. depressor

TABLE 1. Specimens used in this study

ID Species TL (mm) Preparation StainThickness

(lm)

MHW341 Epicrionops bicolor 89 Serial section frontal Picro Ponceau/H1E/Azan 10MHW341 Epicrionops bicolor 89 Serial section sagittal Picro Ponceau/H1E/Azan 10MHW367 Epicrionops bicolor 161 Serial section transversal Picro Ponceau/H1E/Azan 10MHW367 Epicrionops bicolor 161 Serial section frontal Picro Ponceau/H1E/Azan 10ZMH A09801 Salamandrella keyserlingii 26 Serial section transversal Azan 8ZMH A10832 Salamandrella keyserlingii 29 Cleared and stained Alcian blue/Alizarin red —ZMH A10832 Salamandrella keyserlingii 31 Cleared and stained Alcian blue/Alizarin red —ZMH A10831 Salamandrella keyserlingii 18a Dissection by hand Alcian blue —MVZ226908 Desmognathus quadramaculatus 59 Serial section transversal Azan 10MVZ213004 Desmognathus quadramaculatus 57 Serial section transversal Azan 10ZMH A08377 Amphiuma means 105 Serial section transversal Azan 10ZMH A09702 Siren intermedia 216 Dissection by hand Alcian blue —ZMH A09701 Siren intermedia 142 Serial section transversal Azan 10

aOnly snout vent length available.MHW 5 University of California, collection of Marvalee Wake; MVZ 5 Museum of Vertebrate Zoology, Berkeley, ZMH 5 ZoologicalMuseum Hamburg.

600 T. KLEINTEICH AND A. HAAS


mandibulae inserts on the ventral edge of the artic-ular bone via a tendon that reaches rostrad beyondthe mandibular joint (Fig. 4A). The m. depressormandibulae posterior shares its insertion with them. depressor mandibulae in A. means, S. keyserlin-gii, and D. quadramacualtus. In D. quadramacula-tus, however, some fibers of the m. depressor man-dibulae posterior are attached to the distal tip ofthe ceratohyal (Fig. 2); yet in the S. intermediaexamined, the m. depressor mandibulae posteriorinserts exclusively on the distal part of the cera-tohyal and has no insertion on the lower jaw.

The m. hyomandibularis is present in Sirenintermedia, Amphiuma means, and Desmognathusquadramaculatus; it is absent in Salamandrellakeyserlingii. In S. intermedia, the m. hyomandibu-laris originates from the ceratohyal by coveringthe dorsal edge and lateral face of it entirely (Figs.2, 3, and 5). In A. means and D. quadramaculatus,the m. hyomandibularis originates from the lateralface of the distal part of ceratobranchial I; in A.means an additional lateral layer of muscle fibersreaches beyond the ceratobranchial I and origi-nates from the fascia of the dorsal trunk muscula-ture (Fig. 5). The fibers of this muscle run rostradand ventrad. The m. hyomandibularis inserts ven-trally on the distalmost tip of the retroarticular

process in S. intermedia and A. means (Figs. 3 and5); in D. quadramaculatus it inserts via a tendonon the ventral edge of the articular, ventral androstral to the mandibular joint.

The m. branchiohyoideus externus is present inSalamandrella keyserlingii, Siren intermedia, andDesmognathus quadramaculatus, but absent inAmphiuma means. This muscle originates fromthe lateral face of the distal part of ceratobranchialI, where it covers most of it laterally (Figs. 3, 4,and 5). The m. branchiohyoideus externus insertsalong the ventral side of the ceratohyal from proxi-mal to distal, following the bended shape of theceratohyal (Figs. 3 and 4).

The Ventral Hyal Muscles

This group contains the m. interhyoideus andthe m. interhyoideus posterior. Both are inner-vated by cranial nerve VII (n. facialis).

Caecilians. In Epicrionops bicolor, the m. inter-hyoideus has two areas of origins. A rostral bundleof muscle fibers is attached by a fascia to the ven-tral edge of the ceratohyal; a caudal fiber bundleoriginates from the dorsal fascia (Fig. 1B). Thefiber bundles of both parts of the muscle run ven-trad where they merge into one sheath. The m.

TABLE 2. Synonymous terms for hyal and ventral branchial muscles in caecilians and salamanders

This study Druner (1901, 1904) Edgeworth (1935) Other examples

M. depressormandibulae

M. cephalodorsomandibularis(superficial layer)

M. cephalohyomandibularis

M. depressor mandibulae Anterior portion of digastricus(Fischer, 1864; Wilder, 1891);M. digastricus (Fox, 1954; 1959);M. depressor mandibulae anterior(Bauer, 1997; Diogo et al., 2008a,b)

M. depressor mandi-bulae posterior

M. cephalodorsomandibularis(deep layer)

M. depressor mandibulae

M. levator hyoidei M. levator hyoideiM. hyomandibularis M. ceratomandibularis

Part of M. cephalohyo-mandibularis

M. hyomandibularisM. branchiomandibularis

Posterior portion of digastricus(Wilder, 1891); M. ceratohyoideusexternus (Norris and Hughes, 1918);posterior depressor mandibulae(Erdman and Cundall, 1984);

M. branchiohyoideusexternus

M. ceratohyoideus externus M. branchiohyoideusexternus

M. branchiohyoideus(Diogo et al., 2008 a,b)

M. interhyoideus M. interhyoideus M. interhyoideus M. intermaxillaris posterior(Wilder, 1891); M. interhyoideusanterior (Erdman andCundall, 1984)

M. interhyoideusposterior

M. interbranchialis I M. interhyoideus posterior M. gularis (Eaton, 1936)

M. subarcualisrectus I

M. ceratohyoideus internus M. subarcualis rectus I

Mm. subarcualeobliqui II and III

Mm. subarcuale obliquiII and III

Mm. subarcuale obliquiII and III

Adductor arcuate musclesII and III (Fox, 1959)

Mm. ceratohypobranchialesII and III

Mm. subarcuale rectiII and III

M. subarcualisrectus II–IV

M. subarcualis rectusI, II and III

Mm. subceratobranchiales

M. subarcualis rectus IV Constrictor arcuum branchiarum(Wilder, 1891); constrictor arcuatemusculature I–IV (Fox, 1959)

M. transversusventralis IV

M. interbranchialis IV M. transversus ventralis IV Hyotrachealis (Fischer, 1864;Wilder, 1891)



interhyoideus meets its contralateral counterpartin the ventromedian plane.

The m. interhyoideus posterior in Epicrionopsbicolor is a superficial and wide muscle. It origi-nates along a fascia that covers the posterior part ofthe m. depressor mandibulae, the m. depressormandibulae posterior and the dorsal trunk muscu-lature laterally. The m. interhyoideus posteriorfibers run in ventral and caudal direction. The mus-cle inserts in a ventromedian raphe with the respec-tive muscle of the contralateral side (Fig. 1B).

Salamanders. The m. interhyoideus and m.interhyoideus posterior are present in all salaman-ders examined. Both muscles are confluent and onlyincompletely separate. The m. interhyoideus origi-nates from the ventrolateral edge of the ceratohyal;in Amphiuma means, some muscle fibers go beyondthe ceratohyal and take their origin from the ven-tral edge of the squamosal. The m. interhyoideusposterior originates caudal and medial to the m.interhyoideus (Figs. 2 and 5). The fibers of the m.interhyoideus posterior are attached laterally to thefascia of the m. branchiohyoideus externus (Fig. 5)or, in A. means, where a m. branchiohyoideus exter-nus is absent, to the m. hyomandibularis. The m.interhyoideus and the m. interhyoideus posterior,respectively, attach to their counterparts from theopposite side in a median raphe.

The Ventral Branchial Muscles

The ventral branchial muscles comprise the m.subarcualis rectus I, the mm. subarcuales obliquiII and III, the m. subarcualis rectus II-IV, and the

m. transversus ventralis IV. The m. subarcualisrectus I is innervated by cranial nerve IX (n. glos-sopharyngeus), all other muscles of this group areinnervated by the cranial nerve X (n. vagus).

Caecilians. The m. subarcualis rectus I origi-nates along the entire lateral face of ceratobran-chial I (Fig. 1A). The fibers of the m. subarcualisrectus I are obliquely oriented and run in rostraland ventral direction. The m. subarcualis rectus Iinserts on the ceratohyal over a ventral area thatextends from the distal region of the ceratohyal tothe entire ventral surface in the proximal region ofit (Fig. 1A).

In Epicrionops bicolor, the mm. subarcuales obli-qui II and III are absent.

The m. subarcualis rectus II-VI is a thin musclethat originates from the lateroventral side of theceratobranchial IV (Fig. 6). Its fibers run rostradand cover two segments of the hyobranchium.Some fibers insert on the lateral face of cerato-branchial III, the remainder fibers are attached toceratobranchial II.

The m. transversus ventralis IV originates fromthe distal tip of ceratobranchial IV (Fig. 6). Itsfibers run caudad and ventrad and insert on thelateral wall of the trachea, caudal to the trachealcartilages.

Salamanders. The m. subarcualis rectus I isvariable in the salamanders examined. In Sala-mandrella keyserlingii and Desmognathus quadra-maculatus, the m. subarcualis rectus I is a thinbundle of muscle fibers on the ventral side of thebranchial apparatus (Figs. 2 and 4B); in Sirenintermedia and Amphiuma means, the same mus-cle is more voluminous and extends dorsolaterally

Fig. 1. Epicrionops bicolor (MHW367). (A) transverse section in plane with the otic capsule. B: transverse section through thefirst vertebra. Bones, cartilages, and muscles have been outlined and highlighted. (A) The m. depressor mandibulae inserts on theretroarticular process of the lower jaw. The m. hyomandibularis is a voluminous muscle that connects the lower jaw and the cera-tohyal. The m. interhyoideus posterior has no insertion on the lower jaw. (B) The m. depressor mandibulae posterior attaches tothe ceratohyal. The m. interhyoideus is clearly separated into a rostral and a caudal layer.



(Figs. 2 and 5). In S. keyserlingii the m. subarcua-lis rectus I takes origin from the ventromedialedge of ceratobranchial I (Fig. 4); in D. quadrama-culatus this muscle originates ventrolaterally fromceratobranchial I. In S. intermedia, the m. subar-cualis rectus I originates along the lateral face ofceratobranchial I and hypobranchial I (Fig. 7);whereas in A. means, the origin is along the dorso-lateral edge of ceratobranchial I and II.

In Salamandrella keyserlingii and Desmogna-thus quadramaculatus the m. subarcualis rectus Iinserts with a tendon on the medial side of theproximal tip of the ceratohyal (Fig. 4). In S. inter-media it inserts on the ventrolateral edge of basi-branchial I and the ventromedial face of the proxi-

mal part of the ceratohyal (Fig. 7). The m. subar-cualis rectus I in A. means inserts moreextensively on the ventral face of the ceratohyal(Fig. 2) and covers the proximal parts of the carti-lage ventrally and laterally.

The m. subarcualis obliquus II is a thin muscle.It originates from the ventral side of ceratobran-chial II in its proximal region, immediately caudalto the articulation between ceratobranchial II andceratobranchial III (Fig. 4) in all salamander speci-mens examined. In Salamandrella keyserlingii andDesmognathus quadramaculatus, this muscleinserts via a tendon on the lateral side of the cau-dal tip of the basibranchial I (Fig. 4). In Sirenintermedia, the m. subarcualis obliquus II inserts

Fig. 2. Salamandrella keyserlingii (ZMH A09801), Siren intermedia (ZMH A09701), Amphiuma means (ZMH A08377), and Des-mognathus quadramaculatus (MVZ226908). Transverse sections in plane with the otic capsule. Bones, cartilages, and muscleshave been outlined and highlighted. The m. depressor mandibulae and the m. depressor mandibulae posterior can be barely sepa-rated in transverse histological sections, except in S. intermedia. However, the dashed line shows the boarder between the twomuscles that can be traced through consecutive sections. In S. intermedia and D. quadramaculatus, the fibers of the m. depressormandibulae posterior are attached to the ceratohyal.



ventrolaterally on the fascia that covers the m.rectus cervicis (Fig. 7). This insertion site is imme-diately medial to hypobranchial I. In Amphiumameans this muscle inserts on the ventral side ofthe proximal tip of the ceratobranchial I.

In all salamander specimens examined, the m.subarcualis obliquus III originates from the ven-trolateral side of the proximal part of ceratobran-chial III (Figs. 4 and 5). Its fibers run rostrad andmediad, and merge with the m. subarcualis obli-quus II.

The m. subarcualis rectus II-IV originates fromthe ventromedial face of the proximal ceratobran-chial IV in Salamandrella keyserlingii and Des-mognathus quadramaculatus (Figs. 4 and 5). InSiren intermedia and Amphiuma means, the originof this muscle is located on the ventrolateral sideof distal ceratobranchial IV (Fig. 5). In all sala-mander specimens examined, the fibers of the m.subarcualis rectus II-IV are oriented rostrad. Bun-dles of muscle fibers attach to the ventrolateralfaces of the ceratobranchials I, II, and III, respec-tively (Fig. 4); except for S. intermedia, where them. subarcualis rectus II-IV inserts exclusively onceratobranchial I (Fig. 7).

The m. transversus ventralis IV is similar in allspecimens examined, except for Amphiuma means.In most taxa, the muscle originates from the ven-tromedial face of the proximal parts of ceratobran-chial IV, runs ventromediad and rostrad. It insertstogether with its contralateral counterpart fromthe fascia of the m. rectus cervicis in the dorsome-dial region of the muscle. The insertion site is im-mediately rostral to the arytaenoid cartilages andthe laryngeal muscles (Figs. 4 and 5). In A. means,this muscle originates from the distal tip of cerato-branchial IV. Its fibers run in ventromedial andcaudal direction and attach to the linea alba, cau-dal to the arytaenoid cartilages.

DISCUSSIONTerminology of Cranial Muscles in Caeciliansand Salamanders

We consider a consistent use of anatomicalterms necessary for studies on the evolution ofcharacter complexes in animals (sensu Hoyos andDubois, 2004). However, we do not follow the sug-gestion of Hoyos and Dubois (2004) to apply taxo-nomical principles to the terminology of muscles.Our approach is to reevaluate the historical termsto place them in context with modern concepts ofprimary homologies that can then be tested inphylogenetic analyses. Most of the terms appliedherein are derived from the studies of Druner(1901, 1904) and Edgeworth (1935) since thesestudies have been frequently cited and referred toin previous articles on amphibian cranial muscles.

Fig. 3. Siren intermedia (ZMH A09702), drawing of skinnedspecimen in lateral view. Parts of the jaw closing musculature(mm. levatores mandibulae) have been removed. The m. hyo-mandibularis and the m. depressor mandibulae are in closeproximity to each other. Both muscles insert on the dorsal edgeof a small retroarticular process of the lower jaw.

Fig. 4. Salamandrella keyserlingii (ZMH A09801), 3D recon-struction, teeth omitted, jaw closing and dorsal branchialmuscles removed; the following colors have been assigned to thematerials: ocher to bone, blue to cartilage, dark red to muscles.(A) lateral view. The m. depressor mandibulae inserts with atendon on the ventral edge of the lower jaw. (B) ventral view,m. interhyoideus and m. interhyoideus posterior removed. Them. branchiohyoideus externus and the m. subarcualis rectus Ihave a very similar orientation. The m. subarcualis rectus II-IVinserts on the ceratobranchials III, II, and I. The mm. subar-cuale obliqui II and III share their insertion.



M. depressor mandibulae and m. depressormandibulae posterior. These muscles have beenextensively discussed for salamander specimens inBauer (1997) and a comprehensive list of syno-nyms was provided there. Here, we follow Bauer’s(1997) terminological suggestions for the m. de-pressor mandibulae posterior but prefer m. depres-sor mandibulae over m. depressor mandibulae an-terior. Druner (1901) applied the term m. cephalo-dorsomandibularis for the jaw opening muscle anddistinguished deep and superficial layers of thismuscle. The deep layer of Druner’s m. cephalodor-somandibularis is equivalent to the m. depressormandibulae posterior and the superficial layer tothe m. depressor mandibulae. Confusingly, in thesecond part of his extensive study, Druner (1904)examined Siren lacertina and defined a m. cepha-lohyomandibularis and a m. levator hyoidei exclu-

sively for this taxon. The m. cephalohyomandibula-ris originates from the dorsocaudal region of theskull and the ceratohyal and inserts on the lowerjaw. Druner (1904) mentioned that the m. cephalo-hyomandibularis in S. lacertina represents the m.cephalodorsomandibularis (i.e. the m. depressormandibulae; Table 2) in other salamanders. Obvi-ously, Druner’s (1901, 1904) concern was mainly tocreate a terminology that reflects topology in a de-scriptive fashion, rather than one that denotes cor-respondence and sameness. The same applies tothe m. levator hyoidei in S. lacertina, that corre-sponds to the deep layer of the m. cephalodorso-mandibularis in Druner (1901, 1904) and, thus, isthe homologue to the m. depressor mandibulaeposterior herein (Table 2). We suggest to apply theterm m. depressor mandibulae posterior becausethis muscle is separate from the m. depressor

Fig. 5. Salamandrella keyserlingii (ZMH A09801), Siren intermedia (ZMH A09701), Amphiuma means (ZMH A08377), and Des-mognathus quadramaculatus (MVZ226908). Transverse sections at the level of the branchial basket. Bones, cartilages, and muscleshave been outlined and highlighted. The two larval specimens of S. keyserlingii and D. quadramaculatus are very similar. Themost notable difference between the two species is the absence of the m. hyomandibularis in S. keyserlingii.



mandibulae only incompletely and merges withthe m. depressor mandibulae during metamorpho-sis (Norris and Hughes, 1918; Edgeworth, 1935;Eaton, 1936; Piatt, 1938; Fox, 1959; Bauer, 1997).

The studies on salamander cranial musculatureby Fischer (1864), Wilder (1891), and Fox (1954,1959) used the term m. digastricus as synonym forthe m. depressor mandibulae. This name refers tothe jaw opening muscle in human anatomy, which

however is in a much different position and inner-vated in part by the fifth cranial nerve (mylohyoidnerve of the nervus trigeminus ramus mandibula-ris) (Putz and Pabst, 2000).

M. hyomandibularis and m. branchiohyoi-deus externus. Edgeworth (1935) used the termm. hyomandibularis to describe a muscle that con-nects the ceratohyal and the lower jaw in caecil-ians and Siren. In salamanders other than Siren,Edgeworth (1935) used the name m. branchioman-dibularis for a muscle that he considered homolo-gous to the m. hyomandibularis but that originatesfrom the ceratobranchial I instead of the cera-tohyal. We consider those names in Edgeworth’s(1935) study as synonyms and consider it mostparsimonious to hypothesize that this muscleshifted its origin in the evolution of salamanders(Fig. 8). Druner (1901, 1904) applied the term m.ceratomandibularis for the muscle that originatesfrom ceratobranchial I in salamander larvae andinserts on the lower jaw. We conclude that Dru-ner’s (1901, 1904) m. ceratomandibularis is a syno-nym to Edgeworth’s (1935) m. hyomandibularis. InSiren lacertina where the m. hyomandibularisoriginates from the ceratohyal instead of the cera-tobranchial I (Edgeworth, 1935; Bauer, 1997), Dru-ner (1904) described the m. cephalohyomandibula-ris to originate from the dorsal part of the cera-tohyal. We suggest to interpret Druner’s (1904) m.cephalohyomandibularis as compound muscle con-taining homologous parts to the m. depressor man-dibulae as well as the m. hyomandibularis. In S.intermedia we found both muscles in close proxim-

Fig. 6. Epicrionops bicolor (MHW367), transverse sectionimmediately rostral to the tracheal cartilages. Bones, cartilages,and muscles have been outlined and highlighted. The m. trans-versus ventralis IV originates from the lateroventral face ofceratobranchial IV, immediately ventral to the origin of m. sub-arcualis rectus II-IV.

Fig. 7. Siren intermedia (ZMH A09702), drawing of skinned specimen in ventral view, m.intermandibularis, m. interhyoideus, m. interhyoideus posterior, and m. geniohyoideus cut onthe right side of the animal. The mm. subarcuales obliqui II and III insert on the fascia of them. rectus cervicis in S. intermedia.



ity (Fig. 3) and difficult to delimitate from eachother.

Erdman and Cundall (1984), in their study onAmphiuma cranial morphology, used the nameposterior depressor mandibulae for a muscle thatoriginates from ceratobranchial I and inserts ontothe lower jaw. Based on our examinations of thistaxon, we consider their posterior depressor man-dibulae to be the m. hyomandibularis of other sala-manders and caecilians.

The presence of a second n. facialis innervatedmuscle in a position similar to the m. hyomandibu-laris (Fig. 8) caused some confusion, especially instudies on caecilian cranial muscles. This musclewas called m. branchiohyoideus externus by Edge-worth (1935) and we adopted this term herein.Druner (1901) used the synonymous term m. cera-tohyoideus externus for this muscle. The m. bran-chiohyoideus externus spans from ceratobranchialI, close to the origin of the m. hyomandibularis, tothe ceratohyal in salamanders. In caecilians, onlyone muscle is present and homologies have beenambiguous: either the m. hyomandibularis or them. branchiohyoideus externus must be absent com-

pared to salamanders. For larvae of the caecilianIchthyophis beddomii, Norris and Hughes (1918)mentioned a m. ceratohyoideus externus; giventhat this name is synonymous to m. branchiohyoi-deus externus, one would conclude that the m.hyomandibularis is absent in caecilian larvae.However, the same muscle in caecilian larvae wasdescribed as m. hyomandibularis in larvae of I.glutinosus (Edgeworth, 1935) and I. kohtaoensis(Kleinteich and Haas, 2007). By direct comparisonof caecilian and salamander patterns (Fig. 8), wefound that there is no muscle in caecilians thatboth is innervated by cranial nerve VII and thatconnects ceratobranchial I and ceratohyal (see alsoEdgeworth, 1935; Kleinteich and Haas, 2007).Thus, we do not follow Norris and Hughes’ (1918)interpretation and conclude that the m. branchio-hyoideus externus is absent in caecilians.

M. interhyoideus and m. interhyoideus pos-terior. The term m. interhyoideus is used almostconsistently throughout the literature (Druner,1901, 1904; Edgeworth, 1935; Piatt, 1938, 1940;Fox, 1959; Nussbaum, 1977; Kleinteich and Haas,2007). For the m. interhyoideus posterior, however,

Fig. 8. Schematic representation of the mandibular, hyal, and branchial arches and the associated ventral musculature in differ-ent caecilian and salamander taxa. Identical names indicate primary homologies. The scheme for Ichthyophis kohtaoensis isderived from the descriptions in Kleinteich and Haas (2007). The phylogeny has been adopted form Roelants et al. (2007); taxonnames for the Gymnophiona follow Wilkinson and Nussbaum (2006). Taxa that were examined in this study are highlighted inbold face. Abbreviations: HM, m. hyomandibularis; BHE, m. branchiohyoideus externus; SRI, m. subarcualis rectus I; SOII, m. sub-arcualis obliquus II; SOIII, m. subarcualis obliquus III; SRII-IV, m. subarcualis rectus II-IV; Caecilians do not have a m. branchio-hyoideus externus; the muscle that connects the hyal and the first branchial arch is innervated by cranial nerve IX and thus ho-mologous to the m. subarcualis rectus I. The mm. subarcuales obliqui II and III are absent in Epicrionops bicolor. In Salaman-drella keyserlingii there is no ventral muscle that inserts on the mandibular arch; a m. hyomandibularis is absent. In Amphiumameans, only one muscle is present that connects the hyal and first branchial arches. Based on the innervation pattern, the m. bran-chiohyoideus externus is absent.



Druner (1901) used the synonym m. interbranchia-lis I; again for descriptive, topographic reasons.This seems confusing because this posterior mus-cle sheath is weakly delimited from the m. inter-hyoideus and also innervated by cranial nerve VII(Druner, 1901). The m. interhyoideus posteriordevelops in proximity to the m. interhyoideus butfrom a separate anlage in embryonic developmentof the axolotl (Ericsson and Olsson, 2004). Anothersynonym that has been coined for the m. interhyoi-deus posterior is ‘‘m. gularis’’ (Eaton, 1936, 1937).Piatt (1940) discussed the synonymy of the m.interhyoideus posterior and the m. gularis and hesuggested to apply the later term to a muscle thatemerges during metamorphosis in some salaman-ders and that is not present in larvae. Thus, wesuggest using the term m. interhyoideus posteriorin larval stages for stability in terminology and toaccount for its proximity and spatial relations tothe m. interhyoideus.

M. subarcualis rectus I, m. subarculis obli-quus II, m. subarcualis obliquus III, and m.subarcualis rectus II-IV. The term m. subarcua-lis rectus I as applied herein, describes a musclethat originates from ceratobranchial I and insertson the ceratohyal and that is innervated by cranialnerve IX. Druner (1901, 1904) used the synony-mous term m. ceratohyoideus internus to accountfor the position of the m. subarcualis rectus Imedial to the m. ceratohyoideus externus (i.e. them. branchiohyoideus externus herein) in salaman-der larvae. Although both muscles, the m. bran-chiohyoideus externus and the subarcualis rectusI, connect ceratobranchial I to the ceratohyal, theyreceive neural inputs via different routes; the m.branchiohyoideus externus is innervated by cra-nial nerve VII instead of cranial nerve IX (Fig. 8).

Druner (1901) further applied the term mm.ceratohypobranchiales II and III to the ventrome-dial muscles in salamander larvae that originatefrom the ventral edges of the ceratobranchials IIand III and run rostrad to insert together on thebasibranchial I. Another group of muscles, thatoriginate from ceratobranchials I, II, and III andinsert on ceratobranchial IV was called mm. sub-ceratobranchiales in Druner (1901). Confusingly,later Druner (1904) used the terms mm. subar-cuales obliqui synonymously to mm. ceratohypo-branchiales II and III, and mm. subarcuales rectiI, II, and III as synonyms to mm. subceratobran-chiales, respectively. Herein we use the terms m.subarcualis obliquus II and m. subarcualis obli-quus III (Druner, 1904; Edgeworth, 1935) for thesehave become the most commonly used terms in theliterature. The mm. subarcuales recti in Druner(1904) are considered homologous to and termedthe m. subarcualis rectus II-IV in this study.

Edgeworth (1935) adopted the terms from Dru-ner’s second study (1904). Unfortunately, Edge-worth (1935) used the names m. subarcualis obli-

quus II and m. subarcualis obliquus III as syno-nyms for the m. subarcualis rectus II and III.Further, it is important to note that Edgeworth(1935) defined the caudal attachment of themuscles as origin rather than their insertion. Con-sequently, the consecutive numbering that reflectsthe origins of those muscles differs to Druner(1904). In Edgeworth (1935), the m. subarcualisrectus I originates from ceratbranchial I andinserts onto the ceratohyal; in Druner (1904) the‘‘m. subarcualis rectus I’’ originates from cerato-branchial I and inserts on ceratobranchial IV. Dru-ner’s (1904) m. subarcualis rectus I is actuallythe m. subarcualis IV in Edgeworth’s (1935) sys-tem. Although Edgeworth (1935) adopted the termm. subarcualis rectus I from Druner (1904), heapplied it to a muscle that is actually Druner’s(1904) m. ceratohyoideus internus. This was previ-ously pointed out already by Hilton (1957) andLawson (1965).

To avoid potential conflicts between Druner’s(1904) and Edgeworth’s (1935) terminology whenapplied to in Ichthyophis kohtaoensis, Kleinteichand Haas (2007) suggested using the term m. sub-arcualis rectus for muscles that are located on thelateral side of the branchial apparatus and m. sub-arcualis obliquus for muscles that connect theceratobranchials ventrally. Based on Kleinteichand Haas (2007) the ventral branchial muscula-ture in amphibian larvae contains a m. subarcualisrectus I that connects ceratobranchial I to theceratohyal, a m. subarcualis obliquus II, that con-nects ceratobranchial II to I; a m. subarcualis obli-quus III, in between ceratobranchial III and II;and a m. subarcualis rectus II-IV, that runs fromceratobranchial IV to ceratobranchial I (cerato-branchial II in I. kohtaoensis), with additionalinsertion sites on ceratobranchials II and III,respectively. Herein we adopt the terminology thatwas suggested by Kleinteich and Haas (2007) andapplied it to salamander larvae.

M. transversus ventralis IV. This term wasadopted from Edgeworth (1935) and is synonym tom. interbranchialis IV in Druner (1901, 1904) andthe m. hyotrachealis in Fischer (1864) and Wilder(1891).

Evolution of Hyal and Ventral BranchialMuscles in Amphibian Larvae

The confusing terminologies used by variousauthors impede the understanding of evolutionarytransformations of cranial musculature. Ourapproach seeks to clarify the homologies within theLissamphibia, to use a terminology that reflectshomology hypotheses, and to apply the same nameto equivalent structural units in larval and postme-tamorphic stages. This implies the inference of themuscle set of the most recent common ancestor ofthe Lissamphibia. Based on a phylogenetic frame-



work, we use the term ground plan (sensu Hennig,1966) of the Lissamphibia to describe the characterstates that are present at the last node that isshared by the Gymnophiona, Anura, and Caudata.This is an internal node and thus refers to a hypo-thetical taxonomic unit, which cannot be directlyobserved. First preliminary discussions on larvalmuscles in the most recent common ancestor, i.e., forthe ground plan of the Lissamphibia, were presentedin Haas (2001) for the jaw closing muscles and inKleinteich and Haas (2007) for the entire cranialmusculature, except for the eye muscles. Here, wewill complement our previous contributions withresults on the hyobranchial musculature. The phy-logeny that will be used for discussion (Fig. 8) isbased on Roelants et al. (2007). Figure 8 further pro-vides a schematic representation of the homologyhypotheses for the ventral branchial muscles of thespecies studied herein plus Ichthyophis kohtaoensis(data from Kleinteich and Haas, 2007).

Most of the hyal and branchial muscles were pres-ent in all caecilian and salamander species exam-ined (see Haas, 1997, 2003 and discussion below foranurans). We conclude that these muscles werepresent in larvae of the most recent common ances-tor of the Lissamphibia, comprising: m. depressormandibulae, m. depressor mandibulae posterior, m.interhyoideus, m. interhyoideus posterior, m. subar-cualis rectus I, m. subarcualis rectus II-IV, m. trans-versus ventralis IV. However, the presence of thefollowing muscles in the larval ground plan of theLissamphibia needs to be discussed more thor-oughly: the mm. subarcuales obliqui II and III(absent in Epicrionops bicolor), the m. hyomandibu-laris (absent in Salamandrella keyserlingii), andthe m. branchiohyoideus externus (absent in E.bicolor and Amphiuma means) (Fig. 8).

Mm. subarcuales obliqui II and III. In caecil-ians, these muscles are present in larvae of Ichthyo-phis kohtaoensis (Kleinteich and Haas, 2007) andhave been described for adult individuals of the Ich-thyophidae, Uraeotyphlidae, and Caeciliidae (Nuss-baum, 1977, 1979; Wilkinson and Nussbaum, 1997).Further, the mm. subarcuales obliqui II and IIIhave been described in frog tadpoles (Sokol, 1977;Haas, 1997, 2003). By assuming homology (innerva-tion, topographic correspondence) between the mm.subarcuales obliqui II and III in caecilians otherthan Epicrionops bicolor, in frogs, and in salaman-ders, it is most parsimonious to assign thesemuscles to the ground plan of the Lissamphibia.The loss of the mm. subarcuales obliqui II and III inE. bicolor, thus, is an autapomorphic characterpotentially at the level of the Rhinatrematidae, aswas previously suggested (Nussbaum, 1977; Wilkin-son and Nussbaum, 2006).

M. hyomandibularis. Kleinteich and Haas(2007) concluded that this muscle is part of theground plan of the Lissamphibia. Furthermore,the presence of the m. hyomandibularis was

assumed to be ancestral for salamanders. How-ever, based on the phylogeny, that proposed a sis-ter-group relationship of the Hynobiidae (m. hyo-mandibularis is absent) plus Cryptobranchidae tothe remainder salamanders (Roelants et al., 2007),absence could be ancestral, depending on the char-acter state in the Cryptobranchidae, the Anura,and the correctness of the phylogenetic hypothesis(see Wiens et al., 2005 for variants depending ondatasets and types of analyses). Crucial taxa tosolve this question are species within the Crypto-branchidae, as well as basal anurans.

Druner (1904) examined Andrias japonicus(Cryptobranchidae; old synonym Cryptobranchusjaponicus) and did not mention a m. hyomandibu-laris. Edgeworth (1923) in his study on hyobran-chial architecture in Cryptobranchus alleghanien-sis, Andrias japonicus (both Cryptobranchidae)and Hynobius naevius (Hynobiidae) (Menopomaalleghaniense, Cryptobranchus japonicus, andEllipsoglossa naevius in Edgeworth) did not men-tion a m. hyomandibularis, either. Further, Eaton(1936) states that the m. hyomandibualris (Eatonused the term m. ceratomandibualris) is absent inthe Cryptobranchidae. This suggests that absenceof the m. hyomandibularis is ancestral for theHynobiidae plus Cryptobranchidae. However,Eaton did not study early larval forms and grown-up specimens of paedomorphic Cryptobranchidaeare not ‘‘true’’ larvae. It is well documented thatthe m. hyomandibularis becomes part of the m. de-pressor mandibulae in adult salamanders (Edge-worth, 1935; Eaton, 1936; Bauer, 1997) and thismight be the case in later stages of A. japonicusand C. alleghaniensis.

In anuran tadpoles, there are five muscles thatconstitute to the m. depressor mandibulae complex(Takisawa and Sunaga, 1951; Edgeworth, 1935; deJongh 1968; Sokol, 1975; Haas 2003) and althoughhomologies for those muscles have not been estab-lished yet, the m. hyoangularis of anuran tadpolesoriginates from the ceratohyal and inserts onto theretroarticular process of Meckel’s cartilage. The m.hyoangularis is, therefore, in the correct topo-graphic position to be a tentative m. hyomandibu-laris homologue. If this was true, then it would bemost parsimonious (assuming the least number ofevolutionary transformations) that the m. hyoman-dibularis is ancestral to the Lissamphibia andeach of the three lissamphibian subgroups. Accord-ingly, its loss in the Hynobiidae plus Cryptobran-chidae would be an apormorphy for the Crypto-branchoidea clade. Furthermore, the origin of thismuscle from the ceratohyal (as found in caecilians,frogs, and Siren intermedia) would be ancestral forthe Lissamphibia; the shift of the origin of the m.ceratomandibualris to the ceratobranchial I wouldthen be an autoapomorphic character for the Sala-mandroidea sensu Zhang and Wake (2009b), i.e.the monophyletic group that comprises the Ambys-



tomatidae, Dicamptodontidae, Salamandridae, Pro-teidae, Rhyacotritonidae, Amphiumidae, and Ple-thodontidae (Fig. 8; in Roelants et al., 2007, theDicamptodontidae are considered to be nestedwithin the Ambystomatidae).

The M. branchiohyoideus externus was originallydescribed as being unique for salamander larvae(Druner, 1901; Edgeworth, 1935). This is also sup-ported herein; the m. branchiohyoideus externus ispresent in the salamanders and absent in the cae-cilian examined. However, it is unclear if anurantadpoles possess a homologue. Thus, the questionremains ambiguous as to whether the m. branchio-hyoideus externus is apomorphic to the Batrachia(Caudata 1 Anura) or to the Caudata alone. Inanuran tadpoles, the m. constrictor branchialis I(sensu Schlosser and Roth 1995; Haas, 1997) origi-nates from the distal parts of ceratobranchial I,adjacent to the origin of the first branchial arch le-vator. It inserts on the posterolateral tip of theceratohyal. It is present in Ascaphus truei, all spe-cies of Alytidae and Bombinatoridae, Spea bombi-frons, all three species of the genus Scaphiopus,Heleophryne natalensis, and Pelodytes caucasicus(Pusey, 1943; Gradwell, 1971; Sokol, 1975; Haas,1997; Haas, 2003). It is innervated by cranialnerve IX, and not c. n. VII as the branchiohyoi-deus externus of salamanders. The interpretationof the muscle as the most anterior of the m. con-strictor branchialis series in anuran tadpoles isambiguous as was discussed in detail elsewhere(Cannatella, 1999). However, the alternative hy-pothesis, that this anuran larval muscle is homolo-gous to the caudate m. branchiohyoideus externuswith the same topographic relations, would requireexplaining the difference in innervation.

Within salamanders, the absence of the m. bran-chiohyoideus externus in Amphiuma means clearlyis the derived condition. Druner (1904) and Erd-man and Cundall (1984) reported this muscleabsent in A. tridactylum as well. Edgeworth (1935)claimed that the absence of this larval muscle isrelated to direct development and, thus, the ab-sence of a distinct larval stage in Amphiuma.

Caecilians and salamanders are mostly consist-ent in their sets of homologous muscles, but theydiffer notably in the way the hyal and branchialmuscles are organized.

Epicrionops bicolor larvae are very similar to lar-vae of Ichthyophis kohtaoensis (Kleinteich andHaas, 2007) in their hyal and branchial muscula-ture. Wake (1989) did not describe a m. transver-sus ventralis IV in E. bicolor larvae, contrary toour results (see Fig. 6). The only differences wefound between larvae of the two species is the ab-sence of the mm. subarcuales obliqui and the ab-sence of an attachment of the m. interhyoideus pos-terior to the lower jaw in E. bicolor. The insertionof the m. interhyoideus posterior on the ventralside of the retroarticular process is autapomorphic

for adult caecilians and contributes to their uniquedual jaw closing mechanism (Bemis et al., 1983;Nussbaum, 1983; Summers and Wake, 2005; Klein-teich et al., 2008). Larvae of E. bicolor show theancestral character state, i.e. the m. interhyoideusposterior is an exclusively branchial muscle andwill shift its insertion at metamorphosis. In I. koh-taoensis the unique caecilian jaw closing mecha-nism is present in larvae; i.e. the ancestral adultcharacter state is expressed earlier in development.

Salamandrella keyserlingii and Desmognathusquadramaculatus both have larvae that are sur-prisingly congruent in their cranial musculaturedespite being distantly related in salamander phy-logeny (Frost et al., 2006; Roelants et al., 2007)and despite evidence that D. quadramaculatus re-evolved a free-living aquatic larval stage fromdirect developing ancestors (Chippindale et al.,2004; Mueller et al., 2004). Our descriptions of themusculature in salamander larvae are consistentwith previous studies (Druner, 1901, 1904; Litzel-mann, 1923; Edgeworth, 1935; Piatt, 1938; Fox,1959; Bauer, 1997). The juveniles of paedomorphicSiren intermedia and Amphiuma means examineddiffer from other salamander larvae in that theinsertion of the m. depressor mandibuale is at thedorsal edge of the retroarticular process, whereasin S. keyserlingi and D. quadramaculatus the m.depressor mandibulae inserts via a tendon on theventral surface of the lower jaw. The situation inS. intermedia and A. means is comparable tolarval and adult caecilians (Lawson, 1965; Wilkin-son and Nussbaum, 1997; Kleinteich and Haas,2007; Muller et al., 2009). Further, in the two pae-domorphic species the m. subarcualis rectus I hasa fleshy insertion that was proposed to be typicalfor adult salamanders (Druner, 1901) and is simi-lar to the condition in caecilians.

The m. subarcualis rectus II-IV in Siren interme-dia differs from the homologous muscle in othersalamanders by not having fibers that insert onceratobranchials II and III. This was previouslymentioned by Eaton (1936) and considered to bethe derived condition within salamanders; fibersthat attach to ceratobranchials II and III were con-sidered the ancestral character state. However,Haas (1997) did not find distinct heads of the m.subarcualis rectus II-IV that insert on ceratobran-chials II and III in frog tadpoles (comparable toS. intermedia), suggesting that the absence ofthese heads actually might be ancestral for theBatrachia. Unlike frogs and salamaders, caecilianslack an insertion of the m. subarcualis rectus II-IVat ceratobranchial I and larvae of both speciesexamined so far (Ichthyophis kohtaoensis Klein-teich and Haas, 1997 and Epicrionops bicolor) dif-fer in the presence of an insertion of the m. subar-calis rectus II-IV at ceratobranchial III (Fig. 8).The character state for the m. subarcualis rectusII-IV in the ground plan of the Gymnophiona and



Lissamphibia cannot be reconstructed with cer-tainty at the moment.

The m. depressor mandibulae posterior shows anontogenetic relocation of muscle fibers from an inser-tion at the ceratohyal to the lower jaw in salamanders(Litzelmann, 1923; Edgeworth, 1935; Piatt, 1938;Fox, 1959; Bauer, 1997). However, the relocation ofm. depressor mandibulae posterior muscle fibers insalamander larvae apparently happens at develop-mental stages that were younger than the ones exam-ined herein. In Desmognathus quadramaculatus lar-vae examined, only few fibers of the m. depressormandibulae posterior attached to the ceratohyal, andin Salamandrella keyserlingii, there was no connec-tion to the ceratohyal. Caecilians also show an ontoge-netic transition of the m. depressor mandibulae poste-rior muscle fibers from the ceratohyal to the lowerjaw (Edgeworth, 1935). However, in caecilian larvaethe entire m. depressor mandibuale posterior firstinserts on the ceratohyal and the shift to the lowerjaw does not occur before metamorphosis. Interest-ingly, the paedomorphic salamander Siren intermediashows the same condition as caecilian larvae, whichled to the assumption that the insertion of the m. de-pressor mandibuale posterior entirely to the cera-tohyal is ancestral for amphibian larvae (Edgeworth,1935; Bauer, 1997). However, recent studies differ inthe position of the Sirenidae in salamander phylog-eny: Frost et al. (2006) assume a sister group relation-ship of the Sirenidae and Proteidae that is nestedwithin a clade that further comprises the Ambysto-matidae and Salamandridae. Roelants et al. (2007;Fig. 8), considered the Hynobiidae plus Cryptobran-chidae to be in a sister-group relationship to the Sire-nidae plus all other salamanders. Zhang and Wake(2009b) found the Sirenidae to be the sister taxon toall other salamanders as had been proposed(Duellman and Trueb, 1994). Finally, Wiens et al.(2005) retrieved the Sirenidae at these aforemen-tioned positions depending on the dataset includedand the algorithm of reconstruction applied. The cur-rent lack of a robust framework for salamanderhigher-level phylogeny makes it impossible to recon-struct the ontogenetic transformation of the m. de-pressor mandibulae posterior in larvae of the mostrecent common ancestor of amphibians. However,although the developmental timing of the shift in theinsertion of the m. depressor mandibulae posterior inthe ground plan of the Lissamphibia remains obscure,it can be stated with some certainty, that the processof shifting itself is ancestral for metamorphosing am-phibian larvae adding to the characteristics of am-phibian metamorphosis (Reiss, 2002; Schoch 2009).

ACKNOWLEDGMENTS

Specimens studied here have kindly been madeavailable by the Zoological Museum Hamburg(ZMH), the Museum of Vertebrate Zoology Berkeley(MVZ), and by Marvalee H. Wake (MHW, Univer-

sity of California, Berkeley). In particular, theauthors are grateful to Marvalee H. Wake (Berke-ley) for granting to access the Epicrionops bicolorspecimens examined therein. The preparation ofhistological serial sections of salamander larvae byKatja Felbel and Rommy Petersohn (both Jena) ishighly appreciated. Further, they thank ThomasDejaco (Innsbruck) and Tamer Fawzy (Hamburg)for their help in digitizing the histological serial sec-tions of the salamander specimens. Two anonymousreviewers helped to improve an earlier version ofthe manuscript. Thomas Kleinteich is grateful tothe Studienstiftung des deutschen Volkes and to theVolkswagen Foundation for funding his research onamphibian feeding systems. Further support of thiswork was funded by the German Research Founda-tion (DFG) grant HA2323/10-1.

LITERATURE CITED

Bauer WJ. 1997. A contribution to the morphology of visceraljaw-opening muscles of urodeles (Amphibia: Caudata). J Mor-phol 233:77–97.

Bemis WE, Schwenk K, Wake MH. 1983. Morphology and func-tion of the feeding apparatus in Dermophis mexicanus(Amphibia: Gymnophiona). Zoo J Linn Soc 77:75–96.

Bock P. 1989. Romeis mikroskopische Technik. Baltimore:Urban und Schwarzenberg. 697 p.

Brower AVZ, Schawaroch V. 1996. Three steps of homologyassessment. Cladistics 12:265–272.

Cannatella D. 1999. Architecture: Cranial and axial musculos-kelton. In: McDiarmid RW, Altig R, editors. Tadpoles: TheBiology of Anuran Larvae. Chicago: University of ChicagoPress. pp 52–91.

Carroll RL. 2007. The palaeozoic ancestry of salamanders, frogsand caecilians. Zoo J Linn Soc 150:1–140.

Chippindale PT, Bonett RM, Baldwin AS, Wiens JJ. 2004. Phy-logenetic evidence for a major reversal of life-history evolu-tion in plethodontid salamanders. Evolution 58:2809–2822.

De Jongh HJ. 1968. Functional morphology of the jaw appara-tus of larval and metamorphosing Rana temporaria L. Neth-erlands J Zool 18:1–103.

De Pinna MGG. 1991. Concepts and tests of homology in thecladistic paradigm. Cladistics 7:367–394.

Deban SM, Wake DB. 2000. Aquatic feeding in salamanders. In:Schwenk K, editor. Feeding: Form, Function and Evolution inTetrapod Vertebrates. San Diego: Academic Press. pp 65–94.

Deban SM, O’Reilly JC, Nishikawa KC. 2001. The evolution ofthe motor control of feeding in amphibians. Am Zool 41:1280–1298.

Dingerkus G, Uhler DL. 1977. Enzyme clearing of alcian bluestained whole small vertebrates for demonstration of carti-lage. Stain Tech 52:229–232.

Diogo R, Abdala V, Lonergan N, Wood BA. 2008a. From fish tomodern humans-comparative anatomy, homologies and evolu-tion of the head and neck musculature. J Anat 213:391–424.

Diogo R, Hinits Y, Hughes SM. 2008b. Development of mandib-ular, hyoid and hypobranchial muscles in the zebrafish:Homologies and evolution of these muscles within bony fishesand tetrapods. BMC Dev Biol 8:24.

Druner L. 1901. Studien zur Anatomie der Zungenbein-, Kie-menbogen- und Kehlkopfmuskulatur der Urodelen. I. Theil.Zool Jahrb Anatomie Ontogenie 15:435–622.

Druner L. 1904. Studien zur Anatomie der Zungenbein-, Kie-menbogen-, und Kehlkopmuskeln der Urodelen. II. Theil.Zool Jahrb Anatomie Ontogenie 19:361–690.

Duellman WE, Trueb L. 1994. Biology of amphibians. Balti-more, London: John Hopkins University Press. 696 p.



Eaton TH. 1936. The myology of salamanders with particularreference to Dicamptodon ensatus (Eschscholtz). I. Muscles ofthe head. J Morphol 60:31–75.

Eaton TH. 1937. The gularis muscle in Urodela. J Morphol60:317–324.

Edgeworth FH. 1920. On the development of the hyobranchialand laryngeal muscles in Amphibia. J Anat 54:125–162.

Edgeworth FH. 1923. On the larval hypobranchial skeleton andmusculature of Cryptobranchus, Menopoma, and Ellipso-glossa. J Anat 57:97–105.

Edgeworth FH. 1935. The cranial muscles of vertebrates. Lon-don: Cambridge University Press. 493 p.

Erdman S, Cundall D. 1984. The feeding apparatus of the sala-mander Amphiuma tridactylum: Morphology and behavior. JMorphol 181:175–204.

Ericsson R, Olsson L. 2004. Patterns of spatial and temporalvisceral arch muscle development in the Mexican axolotl(Ambystoma mexicanum). J Morphol 261:131–140.

Fischer JG. 1864. Anatomische Abhandlungen uber die Peren-nibranchiaten und Derotremen. Hamburg: O. Meissner.170 p.

Fox H. 1954. Development of the skull and associated struc-tures in the Amphibia with special references to the urodeles.Trans Zool Soc Lond 28:241–304.

Fox H. 1959. A study of the development of the head and phar-ynx of the larval urodele Hynobius and its bearing on the evo-lution of the vertebrate head. Philos Trans R Soc Lond B BiolSci 242:151–204.

Frost DR, Grant T, Faivovich J, Bain RH, Haas A, HaddadCFB, De Sa RO, Channing A, Wilkinson,M, Donnellan SC,Raxworthy CJ, Campbell JA, Blotto BL, Moler P, Drewes RC,Nussbaum RA, Lynch JD, Green DM, Wheeler WC. 2006. Theamphibian tree of life. Bull Am Mus Nat Hist 297:1–370.

Gradwell N. 1971. Ascaphus tadpole: Experiments on the suc-tion and gill irrigation mechanism. Can J Zool 49:307–332.

Haas A. 1997. The larval hyobranchial apparatus of discoglossidfrogs: Its structure and bearing on the systematics of theAnura (Amphibia: Anura). J Zoolog Syst Evol Res 35:179–197.

Haas A. 2001. Mandibular arch musculature of anuran tad-poles, with comments on homologies of amphibian jawmuscles. J Morphol 247:1–33.

Haas A. 2003. Phylogeny of frogs as inferred from primarilylarval characters (Amphibia: Anura). Cladistics 19:23–89.

Haas A, Fischer MS. 1997. Three-dimensional reconstruction ofhistological sections using modern product-design software.Anat Rec 249:510–516.

Hennig W. 1966. Phylogenetic systematics. Urbana, IL: Univer-sity of Illinois Press. 263 p.

Hilton WA. 1957. Head muscles of salamanders. Bull SouthernCalif Acad Sci 56:1–13.

Hoßfeld U, Olsson L. 2005. The history of the homology conceptand the ‘‘Phylogenetisches Symposium.’’ Theory Biosci124:243–253.

Hoyos JM, Dubois A. 2004. An approach to the nomenclature ofanuran musculature. Cuad Herpetol 18:65–72.

Kleinteich T, Haas A. 2007. Cranial musculature in the larva ofthe caecilian, Ichthyophis kohtaoensis (Lissamphibia: Gymno-phiona). J Morphol 268:74–88.

Kleinteich T, Haas A, Summers AP. 2008. Caecilian jaw-closingmechanics: Integrating two muscle systems. J R Soc Interface5:1491–1504.

Lawson R. 1965. The anatomy of Hypogeophis rostratus Cuvier(Amphibia: Apoda or Gymnophiona). Part II. The muscula-ture. Proc Univ Newcastle Phil Soc 1:52–63.

Litzelmann E. 1923. Entwicklungsgeschichtliche und verglei-chend-anatomische Untersuchungen uber den Visceralap-parat der Amphibien. Z Anat Entwicklungsgesch 67:457–493.

Marjanovic D, Laurin M. 2007. Fossils, molecules, divergencetimes, and the origin of lissamphibians. Syst Biol 56:369–388.

Mueller RL, Macey JR, Jaekel M, Wake DB, Boore JL. 2004.Morphological homoplasy, life history evolution, and historicalbiogeography of plethodontid salamanders inferred from com-

plete mitochondrial genomes. Proc Natl Acad Sci USA101:13820–13825.

Muller H. 2007. Developmental Morphological Diversity in Cae-cilian Amphibians. Systematic and Evolutionary Implications.Leiden: Leiden University Press. 266 p.

Muller H, Wilkinson M, Loader SP, Wirkner CS, Gower DJ.2009. Morphology and function of the head in fetal and juve-nile Scolecomorphus kirkii (Amphibia: Gymnophiona: Scoleco-morphidae). Biol J Linn Soc Lond 96:491–504.

Norris HW, Hughes SP. 1918. The cranial and anterior spinalnerves of the caecilian amphibians. J Morphol 31:490–557.

Nussbaum RA. 1977. Rhinatrematidae: A new family of caecil-ians (Amphibia: Gymnophiona). Occas Pap Mus Zool UnivMich 682:1–30.

Nussbaum RA. 1979. The taxonomic status of the caecilian ge-nus Uraeotyphlus Peters. Occas Pap Mus Zool Univ Mich687:1–20.

Nussbaum RA. 1983. The evolution of an unique dual jaw-clos-ing mechanism in caecilians (Amphibia: Gymnophiona) andits bearing on caecilian ancestry. J Zool London 199:545–554.

O’Reilly JC. 2000. Feeding in caecilians. In: Schwenk K, editor.Feeding: Form, Function and Evolution in Tetrapod Verte-brates. San Diego: Academic Press. pp 149–166.

O’Reilly JC, Deban SM, Nishikawa KC. 2002. Derived live his-tory characteristics constrain the evolution of aquatic feedingbehavior in adult amphibians. In: Aerts P, D’Aout K, HerrelA, Van Damme R, editors. Topics in Functional and EcologicalVertebrate Morphology. Maastricht: Shaker Publishing.pp 153–190.

Piatt J. 1938. Morphogenesis of the cranial muscles of Amblys-toma punctatum. J Morphol 63:531–587.

Piatt J. 1939. Correct terminology in salamander myology I.Intrinsic gill musculature. Copeia 1939:220–224.

Piatt J. 1940. Correct terminology in salamander myology II.Transverse ventral throat musculature. Copeia 1940:9–14.

Pusey HK. 1943. On the head of the liopelmid frog Ascaphustruei. I. The chondrocranium, jaws, arches and muscles of apartly-grown larva. Q J Microsc Sci 84:105–185.

Putz R, Pabst R. 2000. Sobotta, Atlas der Anatomie des Men-schen. Band 1. Jena, Munchen: Urban & Fischer. 440 p.

Reilly SM, Lauder GV. 1988. Atavisms and the homology of hyo-branchial elements in lower vertebrates. J Morphol 195:237–245.

Reiss JO. 2002. The phylogeny of amphibian metamorphosis.Zoology 105:85–96.

Richter S. 2005. Homologies in phylogenetic analyses-conceptand tests. Theory Biosci 124:105–120.

Rieppel O, Kearney M. 2002. Similarity. Biol J Linn Soc 75:59–82.

Roelants K, Gower DJ, Wilkinson M, Loader SP, Biju SD, Guil-laume K, Moriau L, Bossuyt F. 2007. Global patterns of diver-sification in the history of modern amphibians. Proc NatlAcad Sci USA 104:887–892.

Roth G, Wake DB. 1985. Trends in the functional morphologyand sensorimotor control of feeding behavior in salamanders:An example of the role of internal dynamics in evolution.Acta Biotheor 34:175–191.

San Mauro D, Gower DJ, Oommen OV, Wilkinson M, ZardoyaR. 2004. Phylogeny of caecilian amphibians (Gymnophiona)based on complete mitochondrial genomes and nuclear RAG1.Mol Phylogenet Evol 33:413–427.

San Mauro D, Vences M, Alcobendas M, Zardoya R, Meyer A.2005. Initial diversification of living amphibians predated thebreakup of Pangaea. Am Nat 165:590–599.

Schlosser G, Roth G. 1995. Distribution of cranial and rostralnerves in tadpoles of the frog Discoglossus pictus (Discoglos-sidae). J Morphol 226:189–212.

Schlosser G, Roth G. 1997. Evolution of nerve developmentin frogs. I. The development of the peripheral nervous systemin Discoglossus pictus (Discoglossidae). Brain Behav Evol 50:61–93.

Schoch R. 2009. Evolution of life cycles in early amphibians.Ann Rev Earth Planet Sci 37:135–162.



Schuh RT. 2000. Biological Systematics: Principles and Applica-tions. Ithaca: Cornell University Press. 236 p.

Sokol OM. 1977. The free swimming Pipa larvae, with a reviewof pipid larvae and pipid phylogeny (Anura: Pipidae). J Mor-phol 154:357–426.

Sokol OM. 1975. The phylogeny of anuran larvae: A new look.Copeia 1975:1–23.

Strong OS. 1895. The cranial nerves of Amphibia. J Morphol10:101–230.

Summers AP, Wake MH. 2005. The retroarticular process,streptostyly and the caecilian jaw closing system. Zoology108:307–315.

Takisawa A, Sunaga Y. 1951. Ueber die Entwicklung des m. de-pressor mandibulae bei Anuren im Laufe der Metamorphose.Okajimas Folia Anat Jpn 23:273–293.

Trueb L, Cloutier R. 1991. A phylogenetic investigation of theinter- and intrarelationships of the Lissamphibia (Amphibia:Temnospondyli). In: Schultze HP, Trueb L, editors. Origins ofthe Higher Groups of Tetrapods. Ithaca: Cornell UniversityPress. pp 223–313.

Vogt L, Bartolomaeus T, Giribet G. 2010. The linguistic problemof morphology: Structure versus homology and the standardi-zation of morphological data. Cladistics 26:301–325.

Wake DB. 1982. Functional and developmental constraints andopportunities in the evolution of feeding systems in urodeles.In: Mossakowski D, Roth G, editors. Environmental Adapta-tion and Evolution. Stuttgart, New York: Gustav Fischer.pp 51–66.

Wake MH. 1989. Metamorphosis of the hyobranchial apparatusin Epicrionops (Amphibia: Gymnophiona: Rhinatrematidae):Replacement of bone by cartilage. Annales des SciencesNaturelles Zoologie 10:171–182.

Wake MH. 1993. Evolution of oviductal gestation in amphib-ians. J Exp Zool 266:394–413.

Wiens JJ, Bonett RM, Chippindale PT. 2005. Ontogeny discom-bulates phylogeny: Paedomorphosis and higher-level salaman-der relationships. Syst Biol 54:91–110.

Wilder HH. 1891. A contribution to the anatomy of Siren lacer-tina. Zool Jahrb Anatomie Ontogenie 4:653–696.

Wilkinson M, Nussbaum RA. 1997. Comparative morphologyand evolution of the lungless caecilian Atretochoana eiselti(Taylor)(Amphibia: Gymnophiona: Typhlonectidae). Biol JLinn Soc Lond 62:39–109.

Wilkinson M, Nussbaum RA. 2006. Caecilian phylogeny andclassification. In: Exbrayat JM, editor. Reproductive Biologyand Phylogeny of Gymnophiona (Caecilians). Enfield: SciencePublishers. pp 39–78.

Wilkinson M, Richardson MK, Gower DJ, Oommen OV. 2002.Extended embryo retention, caecilian oviparity and amnioteorigins. J Nat Hist 36:2185–2198.

Zardoya R, Meyer A. 2000. Mitochondrial evidence on the phy-logenetic position of caecilians (Amphibia: Gymnophiona).Genetics 155:765–75.

Zardoya R, Meyer A. 2001. On the origin of and phylogeneticrelationships among living amphibians. Proc Natl Acad SciUSA 98:7380–7383.

Zhang P, Wake MH. 2009a. A mitogenomic perspective on thephylogeny and biogeography of living caecilians (Amphibia:Gymnophiona). Mol Phyl Evol 53:479–491.

Zhang P, Wake DB. 2009b. Higher-level salamander relation-ships and divergence dates inferred from complete mitochon-drial genomes. Mol Phyl Evol 53:492–508.



the hyal and ventral branchial muscles in caecilian and salamander larvae: homologies and evolution

Documents