In gnathostomes, the lower jaw is raised and the mouth is closed by contraction of the adductor mandibulae muscle complex (Olson, 1961). This complex has been intensively studied, and different criteria to name individual muscles and their bundles have been used (e.g., Lubosch, 1914; Lakjer, 1926; Edgeworth, 1935; Kesteven, 1944). It is clear that the nomenclature applied in anatomical studies of tetrapod jaw muscles over the past 100 years is inconsistent. One of the most commonly followed systems is that of Lakjer (1926) who proposed three different sections of the adductor muscles according to the position of their divisions in relation to the Bodenaponeurosis (the strong tendinous sheet that attaches the main jaw muscles to the coronoid). Before Lakjer, Luther (1914) instead used nerves as the main homology criterion for these muscles. Edgeworth (1935) proposed a nomenclature that was mainly based on ontogenetic data and considered Sphenodon to have the plesiomorphic condition of cranial muscles within extant sauropsid groups, this view was clearly influenced by the then prevalent view of Sphenodon as a primitive or relict reptile (but see Watson, 1914), a perception that has gradually been overturned (Whiteside, 1986; Evans, 2008; Jones, 2009; Jones et al., 2009b; Evans and Jones, 2010). The heterogeneity of names applied to these muscles, inconsistent nomenclature, and incorrect homology inferences has caused confusion when comparing the primary data of different authors, and therefore, it has been difficult to establish the correspondence across taxa.
The Lepidosauria, encompassing Rhynchocephalia (represented by two extant species of tuatara, or one according to Hay et al., 2009) and Squamata (nearly 7200 species of lizards, amphisbaenians and snakes), includes a diverse and speciose living clade of reptiles (Vitt and Caldwell, 2008; Conrad, 2008; Evans, 2008; Uetz, 2010, 2011). For such a diverse group, the problems of maintaining a consistent- and homology-based muscle nomenclature are exacerbated. However, a unifying nomenclature system for all tetrapods is in the process of being established. Diogo and Chardon (2000) and Diogo (2004, 2007, 2008) proposed homologies among different adductor mandibulae sections of teleosteans and other osteichthyan fishes mainly based on topology and the course of the mandibular branch of the trigeminal nerve, following the general criterion of Winterbottom (1974). The nomenclature system is based on the homology established between the undivided adductor mandibulae such as that found in the coelacanth Latimeria and the “levator anguli oris,” “adductor externus,” and “adductor posterior” of lizards (e.g., the genus Timon, Lacertidae, Diogo et al., 2008). The proposed names for these muscles were: levator anguli oris mandibularis, A2, or adductor externus complex including the adductor posterior. Diogo et al. (2008) proposed explicitly the name levator anguli oris mandibularis to avoid confusion with the facial muscle levator anguli oris facialis, which is innervated by facial nerve (cranial nerve VII) instead of trigeminal nerve. These authors also postulated that the muscles “pseudotemporalis” (PS) and “pterygomandibularis” (PTM) of lizards derived mainly from the adductor mandibulae A3′/A3″ of bony fishes. Diogo et al. (2008) also proposed that the small muscle that occasionally occupies the adductor fossa of the mandible of the lizards (see e.g., Rieppel, 1980a, b) corresponds to an additional jaw adductor, the A-ω, which is also present in bony fishes.
The goal of this article is thus to review the major subdivisions of the adductor mandibulae complex of lepidosaurians and to apply a revised nomenclature derived from observations of homologous muscles of the sarcopterygian fish Latimeria (Diogo and Chardon, 2000; Diogo, 2004, 2007, 2008). To this end, we provide a general overview of lepidosaur jaw adductor muscles based on dissections of 66 representative species and the revision of all relevant literature. This expansion of this nomenclatural system to lepidosaurians is a further attempt to provide a nomenclature applicable to all vertebrates and should facilitate broad comparisons across vertebrate groups and provide a better understanding of jaw muscle evolution. The general muscle complex is described according to the proposed nomenclature, from superficial to deep in this order: LAO and RAO (levator and retractor anguli oris muscles), A2-SUP and A2-PRO (adductor mandibulae externus superficialis and profundus), adductor mandibulae externus medialis (A2-M), A2-PVM (adductor posterior), PS, PTM, and A-ω (intramandibularis).
Abbreviations used: A2-M = m. adductor mandibulae externus medialis A2-PRO = m. adductor mandibulae externus profundus A2-PVM = m. adductor mandibulae externus posterior A2-SUP = m. adductor mandibulae externus superficialis A2-SUP-I = m. adductor mandibulae externus superficialis insertion A2-SUPj = m. adductor mandibulae externus superficialis jugalis A2-SUPm = m. adductor mandibulae externus superficialis mandibularis A2-SUP-o = m. adductor mandibulae externus superficialis origin A3′ = subdivision of m. adductor mandibulae A3′ LAO = m. levator anguli oris LAO-o = m. levator anguli oris origin POT = postorbital tendon PTMA = m. pterygomandibularis atypicus PTM = m. pterygomandibularis PSs = m. pseudotemporalis superficialis PSp = m. pseudotemporalis profundus RAO = m. retractor anguli oris RAO-o = m. retractor anguli oris, origin RP = rictal plate
MATERIALS AND METHODS
We reviewed the jaw adductor muscles in various extant lepidosaurian families. Representative species of 23 families were dissected, and the anatomy of 10 families was gleaned from published descriptions (Appendix 1). Specimens were dissected from eight collections (OUA, FML, MACN, AMNH, UPRRP, MZUSP, AH, RT, see appendix for more information and institution acronyms). Specimens dissected were drawn, small specimens with the aid of a dissecting microscope equipped with camera lucida. Each muscle layer was illustrated and removed to expose subjacent layers. Later these drawings were combined to produce scene files as renderings superimposed on reconstructions generated from high-resolution computed tomography scans (available at the Digital Morphology library, University of Texas, http://digimorph.org/).
This contribution intends to facilitate broad comparisons across vertebrate groups and provide a better understanding of jaw muscle evolution. The general muscle complex are described according to the proposed nomenclature, from superficial to deep in this order (historical names in parenthesis); LAO and RAO, A2-SUP and A2-PRO (adductor mandibulae externus superficialis and profundus), A2-M, A2-PVM (adductor posterior), PS, PTM, and A-ω (intramandibularis).
In the present work, the phylogenetic definition of homology, as proposed by Patterson (1982) is followed: homology is equivalent to synapomorphy. Therefore, the inference of homology would be at least a two-step procedure (Rieppel 1988a; de Pinna, 1991): Primary homology hypotheses and secondary homology hypotheses. Primary homology hypotheses are conjectures or hypotheses about common origin of characters that are established after analyses of criteria such as function, topology, and ontogeny (i.e., after the so-called test of similarity). In this study, the same method used in previous works (e.g., Diogo 2007, 2008) is used, and thus, all the informations gathered from dissections or from the literature are used to formulate such primary homology hypotheses (e.g., the innervations of the muscles, the relationships with other muscular structures, the relationships with hard tissues, orientation of their fibers, and their function). The primary homology hypotheses have to pass the second or hard test of homology (i.e., the test of phylogenetic conjunction and congruence) before they can actually be considered solid hypotheses of homology. However, under this philosophy, inference of homology is tree dependent (Wheeler, 2005; but see Rieppel and Kearney, 2002). Thus, under the phylogenetic definition of homology, it is the test of phylogenetic conjunction and congruence that ultimately determines whether a hypothesis can be considered a solid hypothesis of homology. The general homology hypotheses proposed previously (Diogo et al., 2008) using the more inclusive groups Actinistia, Dipnoi, Amphibia, Reptilia, and Mammalia have been applied in this work. The characters that were found to be the most variable were optimized in recent phylogenies for lepidosauria based on morphology using parsimony (Conrad, 2008) and molecular data with Bayesian inference (Wiens et al., 2010). Optimization was done in WinClada (Nixon, 1999) using slow and fast optimization functions.
Since jaw muscle nomenclature varies among authors, we reviewed the nomenclature used in seven publications dealing with lepidosaurian jaw muscles (Lakjer, 1926; Haas, 1973; Gomes, 1974; Rieppel, 1980a; Zaher, 1994; Herrel et al., 1999a; Jones et al., 2009a), and synonymized these terms with the proposed nomenclature.
Considering the variation from a general pattern, the jaw muscles of lepidosaurs show at least five different patterns of organization (Fig. 1). According to the homology hypothesis used in this work, the jaw musculature of Lepidosauria was compared with that of the other sarcopterygian vertebrates (Figs. 2A and 3); six characters associated with jaw muscles were found to be the most variable among lepidosaurs (Fig. 2B). These characters showed slightly less homoplasy in the morphological (Fig. 2C) than the molecular tree (Fig. 2D). Fast and slow optimizations were identical in the molecular topology, whereas in the morphological topology, the optimization of RAO fluctuated. Although there is not much difference between these two hypotheses in terms of number of steps, we favored a discussion of the results onto a morphological framework (Conrad, 2008), because it is the most updated and comprehensive morphological treatment of the group and facilitates the interpretation of morphological data (Daza and Bauer, 2010). The proposed nomenclature and the equivalent terms used in seven publications dealing with lepidosaurian jaw muscles (Lakjer, 1926; Haas, 1973; Gomes, 1974; Rieppel, 1980a; Zaher, 1994; Herrel et al., 1999a; Jones et al., 2009a) is presented in Table 1.
Table 1. Comparison between different terminologies employed to name the jaw adductor muscles and the proposed in this study
mm. adductor mandibulae externus profundus (MAMEP)
m. Adductor Mandibulae Externus Profundus (mAMEP)
M. adductor mandibulae posterior (Add. post.)
M. adductor mandibulae posterior (MAMP)
M. adductor mandibulae posterior (amp)
m. adductor mandibulae posterior (map)
M. adductor mandibulae posterior (ap)
mm. adductor mandibulae posterior (MAMP)
m. Adductor Mandibulae Posterior (mAMP)
M. adductor mandibulae externus profundus hinterer Kopf? (IIIb.)
M. pseudotem poralis
M. pseudotemporalis (pt)
m. pseudotemporalis (MPsT)
m. Pseudotemporalis (mPst)
M. pseudotemporalis superficialis (pseud. sup.)
m. pseudotemporalis superficialis (MPSTS)
M. pseudotemporalis superficialis (pss)
m. pseudotemporalis superficialis (ps.s)
m. pseudotemporalis superficialis (MPsTS)
m. Pseudotemporalis Superficialis (mPstS)
M. pseudotemporalis profundus (pseud. prof.)
m. pseudotemporalis profundus (MPSTP)
M. pseudotemporalis profundus (psp)
m. pseudotemporalis profundus (ps.p)
Deep division of m. pseudotem poralis
m. Pseudotemporalis profundus (mPstP)
M. pterygoideus (Pter.)
M pterygoideous (MPT)
M. pterygomandibularis (ptm)
m. pterygoideous (m. pt)
m. pterygoideous (MPt)
m. Pterygoideous (mPt)
M pterygoideous typicus (MPTT)
m. Pterygoideous Typicus (mPtTy)
M. pterygoideus superficialis (Pter. sup.)
m. pterygoideus superficial portion
m. pterygoideous lateralis (MPtlat)
m. pterygoideous anterior (MPtant)
M. pterygoideus extra (Pter. ext.)
m. pterygoideous externus (sic) (MPtext)
M. pterygoideus prufundus (Pter. prof.)
m. pterygoideus profundus
m. pterygoideous medialis (MPtmed)
M pterygoideous atypicus (MPTA)
m. Pterygoideous Atypicus (mPtAt)
M. intramandibularis (Intra.)
All jaw muscles are covered by an aponeurosis, which is variable in size, extension, and pigmentation. In Sphenodon, this aponeurosis forms a relatively rigid sheet or fascia covering entirely the lower temporal fenestra (Fig. 4A; see also Fig. 20 in Jones et al., 2009a). Among squamates, this aponeurosis is well developed in iguanoids (e.g., tropidurids, leiosaurids, Iguana, Anolis; Fig. 5) and shows notable reduction in many scleroglossans, being persistent in some of them (e.g., Xantusiidae, Teiidae, and Varanidae; Figs. 6, 7, 8, 6–8). Some teiids such as Tupinambis or the liolaemid Phymaturus have some fibers of the LAO originating from this aponeurosis. In most lizards, the lower border of this aponeurosis is attached to the quadrato-maxillary ligament (Rieppel, 1980a), although the latter can be present without aponeurosis (see also Rieppel, 1981; Iordanski, 1996).
Superficial Lateral Muscles: LAO and RAO
LAO and RAO are muscles that mainly elevate and retract the corner of the mouth, respectively. Although the LAO participates in jaw adduction because it is connected to the lower jaw indirectly via the rictal plate (RP), there is no a direct homology between this muscle and the A2 of sarcopterygian taxa. For this reason, we agree with Diogo et al. (2008) in considering LAO and RAO as separate muscles from the adductor complex or A2.
LAO and RAO (Fig. 1) have been considered part of the adductor externus mandibulae complex in previous works (Lakjer, 1926; Oelrich, 1956; Haas, 1973; Gomes, 1974; Rieppel, 1980a; Jones et al., 2009a).
Sphenodon possesses both RAO and LAO, arising from the anterior (LAO) and posterior (RAO) margins of lower temporal fenestra (Fig. 4B). Both LAO and RAO are discrete from the underlying A2-SUP muscle (see below) and are inserted in the RP (Fig. 4A). In most families dissected, the LAO normally originates from the jugal, postorbital, squamosal, and quadrate and inserts into the RP (Figs. 4, 7, 8). LAO is variable in size and shape, being very small and triangular in Podarcis, Chalcides and Mabuya, but wide and triangular in most teiids. No LAO was observed in Xantusia vigilis (Fig. 6) or any of the examined gekkotans (e.g., Hemidactylus, Sphaerodactylus, or Lialis; Figs. 9, 10). However, we observed in Coleonyx variegatus some muscular fibers overlying the POT that connects the postorbitofrontal and jugal, which in this species might represent a relict of this muscle.
In Amphisbaena alba, there are two muscles inserted onto the RP (Fig. 11A), the most anterior muscle, which is located posterior to the orbit is the LAO. The other muscle that inserts onto the RP with the participation of the quadrate-maxillary ligament has been identified as RAO (Fig. 11; Rieppel, 1979).
LAO has frequently been reported in snakes (Cundall, 1986; Zaher, 1994; Cundall, 1995; Moro, 1997; Borczyk, 2006) originating from the parietal and postorbital bones and inserting onto the RP. In Waglerophis merremi, we confirmed the presence of LAO but RAO was not found.
Adductor Mandibulae Externus Complex: A2
Our observations of the jaw muscles of lepidosaurs are congruent with the idea that the adductor A2 of Latimeria is homologous with the adductor externus complex of lepidosaurs (Diogo et al., 2008). The A2 in lizards is divided into three layers: A2-SUP (adductor mandibulae externus superficialis), A2-M, and adductor mandibulae externus profundus (A2-PRO).
The A2-SUP is generally undivided (Figs. 1, 4, 7, 12). In Sphenodon, the A2-SUP is a broad sheet, which mainly inserts directly on the mandible, with some fibers inserting on the lateral surface of the Bodenaponeurosis. In the Squamata the A2-SUP originates from the jugal, postorbital squamosal, and quadrate, but in many taxa it also could originate from the parietal bone (e.g., teiids and tropidurids; Abdala and Moro, 2003). The A2-SUP inserts on the several structures: the coronoid process, the Bodenaponeurosis, the articular, and the angular bones of the jaw. In the majority of squamates (e.g., Anolis cuvieri, Ameiva exsul, leiosaurids, Amphisbaena, Heloderma Figs. 5, 7, 11, 12) the A2-SUP is undivided. We found an A2-SUP with a subdivision in xantusiids, varanids, and some gekkotans (Figs. 6, 8, 9, 10, 8–10); one of the divisions inserts onto the jugal and the other onto the jaw. Following the nomenclature used in the present work, we propose that these bundles should accordingly be named as A2-SUPj (the j referring to jugalis) and A2-SUPm (the m referring to mandibularis; see also Daza and Bauer, 2010). The orientation, origin, and insertion of the A2-SUPm are essentially similar to those of an undivided A2-SUP. This A2-SUPm occupies most of the space of the lower temporal fenestra. The A2-SUPj mainly occupies the anterior portion of the lower temporal fenestra. In gekkotans where the orbit and lower temporal fenestra are confluent, the A2-SUPj forms the posterior boundary of the orbit. In limbed gekkotans, the A2-SUPj is closely packed to the eye sclera but in pygopodids such as Lialis burtonis, there is a gap between the A2-SUPj and the eye (Fig. 10A) and a ring of connective tissue surrounds the orbit. The A2-SUPj arises from the postorbitofrontal and parietal and is inserted through a tendinous connection onto the jugal in Sphaerodactylus and Hemidactylus (Fig. 9), but in Lialis, it is also inserted onto the labial surface of the coronoid. In xantusiids, the A2-SUPj is also present, but its fibers run parallel to the jugal, and have a fleshy insertion onto the mandible (Fig. 6). In Varanus griseus, we found a thin muscle behind the orbit that is tentatively identified as an A2-SUPj (Fig. 8). In some gekkotans and varanids, there is also a well-developed POT extending between the jugal and postobitofrontal (Fig. 13), which plays a role in the closing of the orbit posteriorly (Mertens, 1942; Daza and Bauer, 2010). In snakes, the A2-SUP corresponds to a composite muscle of two bundles, named as adductor externus superficialis and adductor externus temporalis (Fig. 1, Table 1 in Zaher, 1994).
The inner and adjacent layer to the A2-SUP is the A2-M (Figs. 4B,C, 7, 8A–C, 10B, 11A–C, 12C). The A2-M most commonly arises from the posterior process of the parietal, prootic, squamosal, and quadrate and inserts on the Bodenaponeurosis (Figs. 2C,D, 8C, 10B, 11D, 12C), and the coronoid. It occupies part of the supratemporal fenestra.
In Sphenodon, the A2-M presents 4 or 5 distinct muscle heads, which can be recognized by the direction of their fibers or their relations with the surrounding vessels and nerves (A2Ma-d, Fig. 4C). These muscle heads converge on the lateral aspect of the Bodenaponeurosis (Haas, 1973). One of the parts of the A2-M is oriented rostrocaudally and the others lateromedially (Fig. 4B,C). Our observations from the dissections are consistent with the detailed description provided by Haas (1973). A separate, deeper portion, the “MAMEP” sensu Haas (1973) inserts on the medial faces of the Bodenaponeurosis.
The A2-M is present in all dissected squamates; although it should be noted that in limbed gekkotans there are no visible different layers of the A2, except by the A2-SUPj and A2-SUPm (see also Herrel et al., 1999b). In the pygopodid Lialis burtonis, the A2-M is separated from the A2-SUP layers and from the A2-PRO (See below). In Lialis, the A2-M is well developed, and it is attached to a narrow Bodenaponeurosis and the lingual side of the jaw (Fig. 10B), consequently a discrete A2-M is polymorphic for the Gekkota. The A2-M is also well developed in cordyliformes, which have a broad origin on the parietal, prootic and quadrate and a huge Bodenaponeurosis separating it and the A2-PRO. In amphisbaenids, the A2-M is also well developed and its fibers inserts on the lateral surface of the hypertrophied Bodenaponeurosis (Fig. 11B,D).
The deeper portion of the A2 is the A2-PRO and corresponds to the “adductor mandibulae externus profundus” (Haas, 1973; Rieppel, 1980a, 1980b). Sphenodon has a well-developed A2-PRO (Fig. 4D). Although this structure is usually composed by a single bundle, in taxa such as Lanthanotus borneensis it may be a complex section that is subdivided into various heads, all of which showing the same aponeurotic insertion (Haas, 1973). The A2-PRO is present in almost all squamates (Haas, 1973; this work) but likewise the A2-M of limbed gekkotans, the A2-PRO is also not differentiated in these taxa. In Lialis the A2-PRO is discrete (Fig. 10C). In snakes, the A2-PRO is always present, being usually divided into three layers (Fig. 1; Zaher, 1994; Moro, 1997).
The A2-PVM is usually designated in the literature as “adductor mandibulae posterior” (Gomes, 1974; Rieppel, 1980a; Herrel et al., 1996, 1999a, b); in Sphenodon this muscle runs from the pterygoid lamella of the quadrate to the medial edge of the mandibular fossa (Jones et al., 2009a). It lies in the same plane and have fibers oriented similarly to the PTM muscle (Fig. 4E). A similar arrangement is found in some lizard groups such as Chamaeleonidae (Poglayen-Neuwall, 1953a). The A2-PVM usually runs from the anterolateral surface of the quadrate to the medial side of the mandible, commonly around the mandibular fossa and medial to the mandibular nerve. The A2-PVM is present in all sauropsid groups (i.e., the sauropsid pattern) and typically arises caudal to the cranial nerve V3 as this nerve leaves the braincase and the pterygoquadrate bar and then is crossed superficially by a branch of this nerve (the caudal ramus sensu Holliday and Witmer, 2007). In Podarcis a thin layer lateral to the cranial nerve V3 could be representing the A2-PVM, which in the specimens dissected was also closely associated with the Meckelian cartilage. In snakes the A2-PVM is usually present and divided in three layers (e.g., Zaher, 1994; Moro, 1997, this work).
Adductor Internus Complex: PS and PTM
According to the homology hypothesis with Latimeria, these two muscles are derived from the A3′/A3″ (Diogo et al., 2008). The adductors A3′ and A3″ of Latimeria correspond to the A3′ of dipnoans such as Lepidosiren and Neoceratodus (Fig. 3) which in turn corresponds to the PS + PTM of amphibians and reptiles, including lepidosaurs. However, there is no evidence for the homology between the A3′ and A3″ muscles of Latimeria and the PS and PTM muscles of lepidosaurs. For this reason, the names PS and PTM (pterygotemporalis) are used for these muscles in amphibians and reptiles (see also Diogo et al., 2008).
In Sphenodon, the PS is divided into the superficial (PSs) and deep bundles (PSp). Most lizards analyzed also exhibit both divisions; however, in gekkotans, this muscle is mainly an undivided structure. In Gekko vittatus, we found a PS with two divisions, similar to that described by Poglayen-Neuwall (1953b), while in the rest of the geckos dissected we found a mainly undivided PS. The PS is also undivided in Lialis burtonis, so it seems to be a widespread character for Gekkota. Based on our observations, it can be stated that the PS is generally divided in lizards.
In Sphenodon and most lepidosaurs, the A2-M and the PSs are visible in dorsal view, sharing the supratemporal fenestra (Poglayen-Neuwall, 1953a; Rieppel, 1981; Abdala and Moro, 2003; Jones et al., 2009a). In lizards with confluent temporal fenestrae (e.g., geckos), the area of the supratemporal fenestrae is usually covered by the anterior portion of the spinalis capitis (SC) muscles (see also Al Hassawi, 2007; Fig. 14). The SC closes the post-temporal fossae or inserts into the nuchal fossa (when this fossa is present), but in some forms it can extend forward and thus covers a significant portion of the skull table (e.g., Sphaerodactylus roosevelti; Fig 14C). In cases where the PSp is present, it is usually thin and very wide, wrapping the epipterygoid bone, anteriorly, laterally, and posteriorly (Figs. 4E, 8D, 12D). In amphisbaenians, both layers of the PS are usually present, but its association with the epipterygoid is unclear because this bone is only know in Blanus and Rhineura and its homology with the epipterygoid on other lizards is unknown (Montero and Gans, 2008). In snakes, the PS is usually undivided (Zaher 1994).
The PTM (Figs. 1, 4, 7, 9, 10, 14D) in Sphenodon possesses two components, one dorsal referred to as “pterygoideus atypicus” and one ventral, the “pterygoideus typicus” (PTMA and PTMT in Fig. 4D–F). The PTMT of Sphenodon is a very complex muscle, which consists of at least three divisions (Gorniak et al., 1982; Jones et al., 2009a). The PTMA in Sphenodon arises from the dorsal surface of the palatine and the adjacent interorbital septum medially and inserts just caudal to the coronoid process, separated from the more caudal insertion of PTMT. In the Squamata, there is typically only one bundle that corresponds to the PTMT, which arises from the pterygoid and inserts into the entire ventral surface of the articular bone, including the retroarticular and the angular processes. This belly clearly corresponds to the “pterygoideus ventralis” (sensu Holliday and Witmer, 2007) of crocodilians. The components of PTM could be labeled PTM-d and PTM-v. A typical PTM is present in snakes, and also a “PTM accesorius” (see also Zaher, 1994), which is absent in lizards.
The Intramandibularis Muscle: A-ω
The A-ω (intramandibularis) of Latimeria and lepidosaurs is homologous. The close association of the ventral insertions of the PS muscle, the A2-PVM, and the Meckelian cartilage are the defining criteria for recognizing the homologies of the small A-ω (i.e., intramandibularis, Diogo et al., 2008). This muscle is absent in Sphenodon (Jones et al., 2009a, this work). In some lepidosaur groups analyzed, such as teiids, there is an extension of the A2-PVM into the posterior portion of the Meckelian groove (Oelrich, 1956; Haas, 1973; Rieppel, 1980b; Abdala and Moro, 2003); we agree with Diogo et al. (2008) and propose that this muscle corresponds to the A-ω. This structure is absent in snakes and amphisbaenians.
Character Variation and Distribution
From the dissection of specimens and revision of literature, we were able to identify six morphological characters that showed variations within the Lepidosauria (Fig. 2B):
1LAO: (0) small or absent; (1) present, well developed, previously discussed (Gomes, 1974; Rieppel, 1980a, 1984; Herrel et al., 1999a; Abdala and Moro, 2003; Jones et al., 2009a).
2RAO: (0) absent; (1) present, previously discussed (Gomes, 1974; Rieppel, 1980a, 1984; Herrel et al., 1999a; Abdala and Moro, 2003; Jones et al., 2009a).
3A2-SUP divided: (0) absent; (1) present, with a distinct anterior section inserting on the jugal and a posterior section inserted in the mandible, previously discussed (Poglayen-Neuwall, 1953a; Haas, 1973; Gomes, 1974; Rieppel, 1980a, 1984; Herrel et al., 1999a; Abdala and Moro, 2003; Jones et al., 2009a).
4Number of divisions of A2-M: 4 (0); less than 4 (1); undivided (2), previously discussed (Haas, 1973; McDowell, 1986; Rieppel, 1988b; Zaher, 1994; Abdala and Moro, 2003; Jones et al., 2009a).
5PTMA: absent (0); present (1), previously discussed (Haas, 1973; Abdala and Moro, 2003; Jones et al., 2009a).
6Postorbital tendon between jugal and postorbital (postorbitofrontal): absent, this contact being bony by abutting of the jugal onto the postorbital (postorbitofrontal) or being totally absent (0); present, a postorbital ligament connects these two elements (1), previously discussed (Mertens, 1942; Daza and Bauer, 2010).
The scores of these six characters for 19 main clades of lepidosaurs are:
Lepidosaurs presented great variation in the jaw muscle structures, Sphenodon presenting one of the most complex jaw muscles arrangements. Sphenodon differs from Squamata in the possession of multiple divisions of the A2-M and the PTMA. Possession of RAO is homoplasic for Sphenodon and the clade formed by amphisbaenians and snakes, it is more likely that this muscle is not homologous for these taxa. Squamata presents a reduction in the number of divisions of A2-M, and limbed gekkotans present an undivided A2-M. Among the Squamata, Gekkotans developed several changes in the jaw muscles, one of them is reduction or lost of LAO, a character that is also common for Scincoidea. The A2-SUP is undivided in the majority of squamates but appears entirely or partially divided four lizard clades: Agamidae, Xantusiidae, Gekkota, and Varanidae, the later two having an incomplete eye socket. The development of a partially or complete divided A2-SUP and a POT in Gekkota and the Varanidae could be influenced with the lost of contact between the jugal and the postorbital (Daza and Bauer, 2010). These two clades present a reduced jugal bone and only one bone in the posterodorsal edge of the eye socket, termed the postorbitofrontal. For an alternative interpretation of this bone in the Gekkota see Wise and Russell (2010).
Superficial Lateral Muscles: LAO and RAO
The A2 of Latimeria corresponds to the LAO, RAO, and A2-PVM of amphibians and reptiles. As none of these three latter muscles corresponds directly to the A2, they are designated previously as “levator anguli oris mandibularis” (LAO), “retractor anguli oris” (RAO) and adductor mandibulae “A2-posteroventral muscle” (A2-PVM), respectively (Diogo et al., 2008; Diogo and Abdala, 2010).
It has been considered that the m. adductor mandibulae externus superficialis (A2-SUP) is divided into three portions: the m. levator anguli oris, the m. RAO, and the A2-SUP sensu stricto (Lakjer 1926, Holliday and Witmer 2007, Jones et al., 2009a). The presence of LAO and RAO is in general highly variable; RAO is present in sarcopterygian taxa such as lungfishes (e.g., Diogo et al., 2008), Sphenodon, and Uromastyx acanthinurus (Herrel et al., 1999a), Lanthanotus borneensis (Haas, 1973; labeled as “levator anguli oris 1c”), and amphisbaenians (Rieppel, 1979). RAO is absent in amphibians, crocodiles, and turtles (Ecker and Haslam, 1889; Schumacher, 1973; Duellman and Trueb, 1994; Diogo and Abdala, 2010). The absence of RAO has been described in snakes (Cundall, 1986, 1995; Zaher, 1994; Moro, 1997; Borczyk, 2006), although Rieppel (1980c) reports the presence of RAO in Natrix and Elaphe. Considering the recent morphological and molecular phylogenies, the presence of this muscle is convergent between Sphenodon and the squamatan groups (Fig. 3), and as it is not widespread in the later group, it could be inferred that is not an essential muscle to perform an efficient feeding mechanism.
LAO is present in almost all lepidosaurs and may vary in size and being dimorphic in different species. In Laudakia stellio, LAO has been reported as being sexually dimorphic because tends to be absent in females (Herrel et al., 1995). It is lost or small in size and confined toward its insertion area in some Gekkota and Scincoidea (Figs. 1, 9, 10; Rieppel, 1981; Rieppel, 1984; Abdala and Moro, 1996; Abdala and Moro, 2003). The identification of LAO in gekkotans is controversial. It has been described as poorly defined but generally present (Rieppel, 1984). Gomes (1974) described this muscle in Gekko gecko and Uroplatus fimbriatus; (Haas, 1973) considered that gekkonids possesses a well-developed LAO. Abdala and Moro (1996) did not describe this muscle in any of the South American gekkotans from their study.
In Amphisbaena fuliginosa and Trogonophis wiegmanni, it has been proposed that the LAO is incorporated with the A2-SUP (Rieppel, 1979). In Amphisbaena alba, we were unable to identify an A2-SUP layer (whose insertion usually includes the coronoid), we only found one muscle occupying the anterior and superficial layer of the jaw muscles, which we identify as LAO.
The LAO is present in snakes (Cundall, 1986; Zaher, 1994; Cundall, 1995; Moro, 1997; Borczyk, 2006), but Zaher (1994) questioned its homology with the LAO of lizards. Optimization of the character related to the presence of LAO in the morphological tree, shows that LAO is small or lost in the Gekkota and the Scincoidea. In the molecular tree, LAO is small or lost independently in four clades Gekkota + Dibamidae, Serpentes, Amphisbaenia and Scincidae.
Adductor Mandibulae Externus Complex: A2
Among lepidosaurs, this complex reaches its maximum complexity in Sphenodon (Figs. 1, 4). In Squamata, this complex is more simplified, presenting some reductions, but generally preserving the basic division: A2-SUP, A2-M, and A2-PRO.
In agamids, such as Laudakia stellio and Uromastyx acanthinurus, the A2 has been described as composed by two bundles (“adductor mandibulae externus superficialis anterior and posterior”; Herrel et al., 1999a). Herrel et al. (1999a) illustrated a subdivided muscle labeled as MAME 3 and MAME 4; these divisions might correspond to similar divisions of the A2-SUP in the Acrodonta. Poglayen-Neuwall (1953a), Gomes (1974), and Rieppel (1984) suggest that the most anterior of this division is a poorly defined LAO. However, as pointed out above, our observations indicate that these bundles are subdivisions of the A2-SUP, because the LAO is generally superficial to the A2-SUP.
The POT described above is present in many geckos (e.g., Gekko gecko, Rhacodactylus auriculatus, Uroplatus), and in some cases, it is a very prominent structure that can be preserved in some skeletonized preparations (Daza and Bauer, 2010). This structure was described in geckos as a tendon that inserts posterolaterally in the jugal and goes to the postorbitofrontal (Häupl, 1980). Nonetheless the presence of a POT and a divided A2-SUP muscle appears to be variable in gekkotans and need to be corroborated in more species and specimens from each family. Loss of bony contact between the jugal and the postorbitofrontal in both Gekkota and Varanidae may be symptomatic of reduced compressive strain at the posterodorsal corner of the orbit, in similar fashion to the results obtained by modeling a fused versus unfused postorbital–parietal sutures (Moazen et al., 2009). The development of a POT could be a response to compensate this loss in this area of the skull, which could be of significance for cranial kinesis in lizards (Iordanski, 1996). Optimization of this character in both the morphological and molecular phylogenies suggests that it is an autapomorphy of the clade Gekkota, and is polymorphic for Varanidae because of its absence in Lanthanotus borneensis (McDowell and Bogert, 1954).
The Bodenaponeurosis, which connects the coronoid process of the lower jaw with the A2-M complex (Bodenaponeurose in Lakjer, 1926; basal aponeurosis in Haas, 1973), is highly variable in shape and size in all taxa considered. It is a structure that has been used to categorize the A2-M complex organization Lakjer (1926). His proposal of A2-M organization has been widely followed by most subsequent authors, with some exceptions (Iordanski, 1970), and was also considered in this work to settling the divisions of the A2 complex.
Different criteria have been used to recognize the subdivisions of the A2-M in snakes (Haas, 1973; McDowell, 1986; Rieppel, 1988b; Zaher, 1994), but most authors agree that this structure is always divided into at least two bundles. Results of this work are in accordance with those authors. The number of divisions of the A2-M is another of the characters optimized in the available phylogenies. According to the optimization in the morphological tree, an A2-M with less than four divisions is a synapomorphy of Squamata. Presence of an undivided A2-M could represent a distinct character for gekkonids sensu lato and eublepharids. A revision of more carphodactylids and diplodactylids, which are closely related to pygopodids, is necessary to understand the variation within the Gekkota. Optimization on the molecular tree indicates that the A2-M divided in more than four divisions is an autapomorphy of Rhynchocephalia, and the undivided A2-M an autapomorphy of Gekkota.
The distribution of the A2-PRO in the Gekkota is not clear. Rieppel (1984) considered that some muscular fibers present in gekkonids sensu lato may correspond to the A2-PRO (following the terminology of this work). However, Abdala and Moro (1996) and Herrel et al. (1999b) did not find the A2-PRO in any of the gekkotans reviewed by them. In general, results of this work indicate that there is a trend toward the lack of divisions of the A2 complex in limbed gekkotans, however, since the A2 is markedly divided in the pygopodid Lialis burtonis, there is the possibility that other gekkotans present similar divisions.
The tetrapod adductor posterior has been considered for most authors an independent muscle with a morphological identity separate from the various configurations of the A2 (Oelrich, 1956; Gomes, 1974; Rieppel, 1980a; Holliday and Witmer, 2007; Jones et al., 2009a). This concept is introduced for birds by Barnikol (1954) and is considered for amphibian groups (e.g., Johnston, in press). Iordanski (2010) considered that in Lacertilia the posterior adductor is inseparable from the external adductors. Here, we interpret the tetrapod adductor posterior as a component of the A2 as A2-PVM.
The presence of an A2-PVM in Sphenodon has not been always recognized (Gorniak et al. 1982). The presence of this muscle was reviewed and corroborated by Jones et al. (2009a) and by the results from this work. In geckos, Brock (1932) observed a distinct A2-PVM in the stage near hatching of Cyrtopodion kotschyi specimens. Herrel et al. (1999b) found a reduced A2-PVM in geckos. Haas (1973) suggested that the frequent loss of this muscle in squamates is probably related to the angle and topological relation between the quadrate and the mandible in these reptiles, and he stated that in terms of its innervations the A2-PVM seems to be more similar to the A3 of fishes than to the A2. Interestingly, in the turtle Chelydra the A2-PVM seemingly develops ontogenetically from the A3, whereas in the lizard Podarcis the A2-PVM seems to be derived from the A2 (Rieppel, 1990). Our observations and comparisons support the hypotheses that the A2-PVM is homologous within all sauropsid groups (Holliday and Witmer, 2007; Diogo et al., 2008). In all sauropsids there is a muscle located caudal to the mandibular branch of the trigeminal nerve (V3), mesial to the rest of the A2 complex, and lateral to the PTM, that corresponds to the A2-PVM.
Adductor Internus Complex: PS and PTM
The PS is present in Sphenodon and all squamate taxa. Gekkotans exhibit a PS undivided, reproducing the general trend in this group to the simplification of the muscular complexes. Abdala and Moro (1996) also reported an undivided PS in gekkonoideans, but the muscle described by them actually seems to correspond to the PSs. In a description of the jaw muscles of Tarentola, Poglayen-Neuwall (Poglayen-Neuwall, 1953a) stated that it has a PSs composed of only a few fibers.
Oelrich (1956) and Gomes (1974) considered the PSp almost indistinguishable in iguanids, but Moro and Abdala (1998) found this structure in liolaemids. Abdala and Moro (2003) also found that in teiids the PSp is generally very developed (see also Rieppel, 1980b), but that it could be absent in some taxa such as Echynosaura, Calyptommatus, and Lygodactylus (Abdala and Moro, 2003). In consequence, according to our dissections, the presence of the PSp seems to be generalized in most squamates.
Various authors (Poglayen-Neuwall, 1953c; Schumacher, 1973; Holliday and Witmer, 2007) consider that reptiles such as turtles also have more than one belly of the PTM, which has seemingly been lost in squamates. In the agamids Uromastyx and Laudakia, the PTM has been described as composed of four and three sections, respectively, (Herrel et al., 1999a) although Throckmorton (1976) and Moazen et al. (2008) reported only three divisions in Uromastyx. This in addition to the presence of a divided A2 is additional evidence of the highly derived muscular design of these taxa (see also Moazen et al., 2008, 2009; Iordanski, 2010). Abdala and Moro (1996) described Gymnodactylus geckoides with a PTM and an independent dorsal muscle. However, none of the squamates dissected possesses a well-developed and distally located PTM dorsalis such as that of crocodiles (“pterygoideus dorsalis,” Holliday and Witmer, 2007) or the PTMA of Sphenodon. Wu (2003) provided a mechanical explanation for the loss of the PTM dorsalis (PTMA) in lizards based on the need for anterior movement of the opening jaw in these reptiles, but this awaits further analysis. In amphisbaenians the PTM is small (Rieppel, 1979). According to Haas (1973) and Rieppel (1979), Trogonophis has only one muscle, which corresponds to the pars ventralis (Holliday and Witmer, 2007).
The PTM is widespread in the Lepidosauria, being very complex in agamids, on the other hand, the PTMA is known only in the Rhynchocephalia. However, Sphenodon is the only member of this clade in which this can be observed, and this species is currently viewed as derived rather than plesiomorphic within this group (Jones, 2008). Nevertheless, the presence of a dorsal pterygoideus component in this taxon coupled with its presence in archosaurs and possibly chelonians, suggests the dorsal pterygoideus may be a plesiomorphic feature of amniotes that has been lost in squamates.
The Intramandibularis Muscle: A-ω
According to Iordanski (2008), there are two main types of the intramandibular muscles: “lacertiloidan,” where a portion of the external or posterior jaw adductor (A2-PVM in our system) enters the primordial canal (Meckelian groove) of mandible, and “crocodiloidan,” where the ventral belly of the anterior portions of the internal jaw adductor (A3) separated from the corresponding dorsal belly by the intramuscular tendon, which often has its own attachment to mandible. This structure is absent in snakes and amphisbaenians (Rieppel, 1979; Zaher, 1994; Moro, 1997; Borczyk, 2006; this work). Holliday and Witmer (2007) considered that this muscle is in fact part of the A2-PVM, which has intimate developmental ties to the Meckelian cartilage during the development of the mandible, or of the PSs. In the scheme, we use here, we prefer to consider the fibers occupying the majority of the medial mandibular fossa as a lacertiloidean type of A-ω present in lepidosaurs, homologous with the muscle of the same name in Latimeria.
Lepidosaurs demonstrate considerable variation in their jaw muscles. This variability provides a source of phylogenetic and functional information whose homologies need to be correctly evaluated with proper consideration of outgroup taxa. In general terms, the jaw adductor muscles of lizards are divided in one externus and one internus complex. LAO and RAO are two superficial muscles independent of these complexes, presenting high variability among reptiles and within Lepidosauria. The only muscle occupying the anterior and superficial layer of the jaw muscles in amphisbaenians is identified as LAO. According to the distribution of RAO among lepidosaurs, it is likely that this muscle is convergent between Sphenodon and the clade formed by amphisbaenians and snakes. The A2 or externus jaw complex is formed by four layers: A2-SUP (adductor mandibulae externus superficialis), A2-M, A2-PRO, and A2-PVM (adductor mandibulae posterior). The V3 and V2 branches of the trigeminal nerve embrace the first three of these muscles. The A2-PVM is located caudal to the V3 branch of the mandibulary nerve; although this muscle is included within the A2 complex, an alternative interpretation for this muscle was presented. The A2-SUP is entirely or partially divided in at least four lizard clades: Agamidae, Xantusiidae, Gekkota, and Varanidae. It is possible that the A2-SUPj and the development of a POT play a role in completing the orbit in those clades with incomplete eye socket (Gekkota and Varanidae). Apparently Amphisbaenia is the only lepidosaur group where the A2-Sup is lost, but this interpretation needs support from with embryological studies; it is still unclear if during development this muscle is integrated to the A2-SUP or it is lost indeed. There is some discrepancy between the A2-M muscles arrangement in the Gekkonoidea and the Pygopodoidea. A revision of more representative species from the Pygopodoidea is required to understand better the variation within the Gekkota.
A complete revision of structures associated to the A2 muscles such as the Bodenaponeurosis has the potential for the discovery of additional characters for lepidosaur systematics. The internal jaw muscle complex is formed by PS and PTM; these muscles are between the V2 and V1 branches of the trigeminal nerve. The PS seems to be universally present in the Lepidosauria, but it needs to be evaluated in the Gekkota. Within Lepidosauria, PTMA is only known in Sphenodon, a PTM dorsalis such as that of crocodiles is completely lost in the Squamata.
The nomenclature proposed for the jaw muscles in lepidosaurs is equivalent to that used for the homologous muscles of the sarcopterygian fish Latimeria. Hopefully this contribution will promote and facilitates future discussions in the subject of amniote feeding mechanics and squamate phylogeny.
The authors thank Richard Thomas (Universidad de Puerto Rico), Anthony Herrel (Musée National d'Histoire Naturelle, Paris), Darrel Frost and David Kizirian (American Museum of Natural History of New York), and Colin McCarthy (Natural History Museum of London) for providing the specimens dissected in this work. They also thank Jessie A. Maisano and Chris Bell (The University of Texas at Austin) and Aaron M. Bauer (Villanova University) for access to their CT scan data. Esteban Lavilla, Sonia Kretschmar, and Marta Canepa provided specimens from the Herpetological collection at the Fundación Miguel Lillo. The authors also thank Marc Jones and an anonymous reviewer for their suggestions to improve this manuscript, which are very much appreciated.
List of Species Dissected and Observed from Literature
Numbers between parentheses indicate the amount of specimens dissected. Institutional abbreviations: Anatomy Department, University of Otago, Dunedin, New Zealand (OUA); Field Museum of Natural History, Chicago, USA (FMNH); Fundación Miguel Lillo, Tucumán, Argentina (FML); Museo Argentino de Ciencias Naturales, Buenos Aires, Argentina (MACN); Natural History Museum, London, UK (BMNH); Herpetology Department at the American Museum of Natural History, New York, USA (AMNH); Los Angeles County Museum of Natural History, Los Angeles, USA (LACM); Museo de Zoologia, Universidad de Puerto Rico, Puerto Rico (UPRRP); Museo de Zoologia, Universidad de São Paulo, São Paulo, Brazil (MZUSP); Proyecto Tupinambis en FML (PT); San Diego State University, San Diego, USA (SDSU); private collection of Anthony Herrel (AH); private collection of Richard Thomas (RT); Texas Memorial Museum, Texas, USA (TMM); Yale Peabody Museum, New Haven, USA (YPM).