The typical circumorbital series of reptiles consists of five bones: prefrontal, postfrontal, postorbital, jugal, and lacrimal (Romer, 1956, 1970; Evans, 2008). In lepidosaurs, this general orbital pattern is modified by the integration of bones from the tooth bearing (maxilla) and longitudinal series (frontal) and the development of neomorphic elements such as palpebral, supraorbital, and parafrontal bones (Fig. 1).
The prefrontal in lepidosaurs is the only element that invariably forms part of the orbital margin; the contact of this bone with the jugal excludes the maxilla from the orbit in some forms (e.g., Anolis; Fig. 1). In some cases the lacrimal lies between these two bones producing the same result (e.g., Tupinambis and Varanus; Fig. 1). The frontal generally participates in the orbit of lepidosaurs, but is secondarily excluded from the orbit by contact of the prefrontal with the postfrontal (or postorbitofrontal) in representatives of virtually all squamatan clades (Conrad, 2008; Conrad et al., 2008; Evans, 2008), early reptiliomorphs (e.g., Limnoscelis), some testudines and other archosaurians (Romer, 1956; Carroll, 1988; Benton, 2005). While both the maxilla and the frontal are sometimes excluded from the orbital margin, they participate in the wall of the orbit (eye socket).
Element loss and fusion are the two process responsible for changes in presence of bones and, consequently, orbital configuration. The presence of the postfrontal, postorbital, palpebral, lacrimal, and jugal is inconstant among lepidosaurs and this also results in differences in orbital arrangement. The configuration of orbital bones and the resulting contacts are variable, especially along the posterior margin of the orbit, where lepidosaurs may present closed (e.g., rhynchocephalians, iguanians, some lacertoids), semiclosed (e.g., Anguis, Vanzosaura, varanids) or open orbits (e.g., gekkotans, and some amphisbaenians). The posterior closure of the orbit is normally determined by variation in the shape, size, and presence of the jugal (Kearney, 2003), although variation in the postorbital and postfrontal (or postorbitofrontal when fused) also affects this closure. In fossil rhineurid amphisbaenians the orbit is different and there is a strong controversy about the identity of the bones forming the postorbital bar (e.g., Berman, 1972, 1973; Estes, 1983; Kearney, 2003; Montero and Gans, 2008).
In the Gekkota, the orbit is an important area of the skull, in some forms the orbit occupies 41% of the skull length (e.g., Ptenopus; Fig. 1). Orbit size is related to eye size, which in turn is correlated with parameters of behavioral ecology; in gekkotans the eye is larger in nocturnal than in diurnal species and in cursorial than in scansorial species (Werner and Seifan, 2006). Some gekkotans are lidless and the eye is covered by a fixed transparent window or spectacle. When eyelids are present, as in the eublepharid geckos, the nictitating membrane shows different degrees of development among species (Underwood, 1970).
The orbit in gekkotans is always incomplete posteriorly and is combined in a single lateral space together with the supra- and infratemporal fenestrae. This greater development of the orbit produces enormous changes in the bones of the circumorbital series. In this article, we focus on two bones of controversial presence in the orbit of gekkotans, the lacrimal and jugal. The variation in these elements is assessed in the seven major gekkotan clades (i.e., Pygopodidae, Carphodactylidae, Diplodactylidae, Eublepharidae, Gekkonidae, Phyllodactylidae, and Sphaerodactylidae; Fig. 2) by examining museum specimens belonging to 105 extant species. The structural role of the jugal as part of the circumorbital series leads us to also evaluate the configuration of other elements in the series (prefrontal, postorbitofrontal), especially in light of recent morphological treatments of exemplar gekkotans (Conrad and Norell, 2006; Conrad, 2008; Conrad et al., 2008; Evans, 2008).
We acknowledge that higher order relationships among squamates remain controversial and that most data derived from DNA (Townsend et al., 2004; Vidal and Hedges, 2004; Hugall et al., 2007; Vidal and Hedges, 2009) conflict with both more traditional (Estes et al., 1988) and recent (Conrad, 2008; Conrad et al., 2008) morphologically derived hypotheses. In this article, we follow the higher order squamate classification of Conrad (2008) because it is the most current and comprehensive morphological treatment of the group and facilitates the interpretation of morphological data. We follow Gamble et al. (2008) for intra-gekkotan relationships and the allocation of species to gekkotan families.
The presence of the lacrimal is variable among squamates (Conrad, 2008; Fig. 2). In geckos, when present, it is very small and indefinite. Camp (1923) stated that in the Gekkonidae, the lacrimal was crowded within the orbit and lost to view externally, an observation that has been interpreted as a mistaken identification of a reduced jugal (Stephenson and Stephenson, 1956; Evans, 2008). The lacrimal has been reported as rudimentary or very reduced in newly hatched Lygodactylus (Brock, 1932), and the eublepharids Coleonyx variegatus and Eublepharis macularius (Kluge, 1962; Rieppel, 1984a), where this bone shows intraspecific variation and is sometimes asymmetrically present.
The jugal is usually present in lizards; its loss is a rare occurrence that is known exclusively in forms with elongated bodies (e.g., Feyliniidae, Dibamidae, Serpentes and Amphisbaenia, Fig. 2). The presence of this bone has been ambiguously interpreted for the Gekkota. For example, McDowell and Bogert (1954; p. 92) mistakenly describe the jugal as absent for geckos and pygopodids, but they label the bone in their illustrations of Aristelliger lar, Coleonyx variegatus, Pygopus lepidopodus, Lialis burtonis and Aprasia repens. In the gekkotan family Pygopodidae, another group of lizards where body elongation and limb reduction has taken place, it has been proposed that this bone is absent in Delma, Lialis, and Pygopus (Table 1; McDowell and Bogert, 1954; Underwood, 1957; Jollie, 1960; Kluge, 1976; Rieppel, 1984a, b; Conrad, 2008; Evans, 2008).
Table 1. Interpretation of presence (+) or absence (−) of the jugal bone in some papers dealing with osteology of the Pygopodidae
The postfrontal and postorbital fuse to form a single element in representatives of all lizards clades (Conrad, 2008; Fig. 2). Apparently, this fusion also took place in the Gekkota (Daza et al., 2008) where only one bone bounds the orbit posterodorsally (Fig. 1, see below).
In the Squamata, there are three names applied to neomorphic elements present above in the upper edge of the orbit (i.e., palpebral, supraorbital and parafrontal bones). Although the homology of this element has not been corroborated, it is very likely that these structures were acquired independently. By definition, the palpebral is the neomorphic bone located above the upper eyelid in the anterodorsal corner of the orbit (Peters, 1964; Maisano et al., 2002). We restrict the use of palpebral for that element that occurs discontinuously in Autarchoglossa (pal, Fig. 2), being present in some lacertoids, cordyloids, scincoids, and anguimorphs (Estes et al., 1988; Conrad, 2008), supraorbital to the bone present in the orbit of Loxocemus and pythonid snakes (Fig. 1), and parafrontal bones to the multiple ossifications that lie between the prefrontal and postorbitofrontal in some geckos; these unique structures are known only in the gekkotans Aristelliger and Teratoscincus (Bauer and Russell, 1989) and were differentiated from the osteoderms that roof the orbit in a variety of lizards, including the gekkotan genera Geckolepis, Geckonia, and Tarentola (Underwood, 1970, Bauer and Russell, 1989).
Our intention is not to discuss all the soft parts associated with the bones from the circumorbital series in the Gekkota, but we do mention pertinent structures, in particular those that vary across gekkotans and distinguish them from other lizard clades.
MATERIALS AND METHODS
One hundred and five species of gekkotans were examined from nine institutions (Appendix). We reviewed the circumorbital bones in 211 specimens represented by different preparations (skeletonized [Sk], cleared and stained [C&S], high resolution X-ray computed tomography scans [CT], fluid preserved specimens [Et]). The CT scans were available for 27 gekkotans and were performed at the University of Texas High-Resolution X-ray CT Facility. For this preparation the heads of fluid preserved specimens were scanned by means of CT slices taken along the coronal or transverse axis with variable slice thickness and interslice across the different scans. The field of reconstruction is 30 mm with an image resolution of 1024 × 1024 pixels, resulting in an interpixel spacing of 29.3 μm. The files contained QuickTime slice-by-slice animations, QuickTime animations of 3D rotations and QuickTime animations of 3D cutaways along the three orthogonal axes. Additional to the cleared and stained material available in the collections, we prepared additional species following the technique of Hanken and Wassersug (1981). Illustrations were produced with a camera lucida mounted on a dissecting microscope, and all figures were refined and assembled in Adobe® Illustrator® CS3 13.0.2. A list of specimens reviewed is provided in Appendix. Myological gross dissections were performed using the technique of Bock and Shear (1972).
This bone is absent in most of the gekkotans examined, but when present is difficult to distinguish and may be overlooked. This element is easiest to see in CT scans and cleared and stained preparations. We corroborated its presence in both Eublepharis macularius (JFBM 15831 and CM 67524, Fig. 3A–C) and Coleonyx variegatus (YPM 14383, Fig. 3D). The lacrimal is more prominent in E. macularius, where it is broad and flat, lying on top of the maxilla and contacting the jugal laterally. In E. macularius it bounds the posterior edge of the lacrimal foramen (Fig. 3A–C) while in C. variegatus the bone is reduced and occupies only the lateral corner of this foramen, contacting the jugal anterodorsally (Fig. 3D). In Pachydactylus bicolor (CAS 223912, Fig. 3E) we found two small ossifications in the right orbit, adjacent to the lacrimal foramen. One of these ossifications contacts the maxilla and the other the ectopterygoid (ect, Fig. 3E). This bone is broader anteriorly relative to its left counterpart and fails to contact the palatine anterior to the suborbital fenestra. It also presents a sinuous medial edge instead of smooth one. These two small ossifications of P. bicolor seem to be the result of an asymmetrical and irregular development of the right ectopterygoid, as these structures do not resemble the lacrimal seen in the other species.
Among gekkotans, the sphaerodactylids Aristelliger and Teratoscincus possess parafrontal bones or ossa parafrontalia, which are different from integumentary osteoderms present in many lizards (Bauer and Russell, 1989; Fig. 4). These supraorbital elements are a series of small ossifications lying at the same depth as the frontal bone between the prefrontal and the postorbitofrontal. In both position and configuration they are quite different from the osteoderms. In position, these bones differ from osteoderms because they lie beneath the dermis in the plane of the frontal bone. In shape, they are irregular and lack correspondence with overlying epidermal scales.
Quedenfeldtia trachyblephara possesses a continuous mesenchymal sheet, which in position is comparable to the parafrontal bones. This has been interpreted as a putative synapomorphy of a clade of nonminiaturized sphaerodactylids (Daza et al., 2008). This structure differs from the thickened skin of the upper surface of the head that roofs the orbit and fills the gap between the prefrontal and the postfrontal bones in other lizards (Underwood, 1970).
Virtually in all species reviewed, there is a single bone at the posterodorsal corner of the orbit. This bone varies in shape among species, and when present it clasps the mesokinetic frontoparietal suture. Typically, its shape is angulated with a lateral vertex from where the anterior and posterior processes depart, but in Coleodactylus, Tarentola, Ptyodactylus, Thecadactylus, Goggia, Hemidactylus, Stenodactylus, and Tropiocolotes this bone is rounded laterally, these processes being less distinct. In the Gekkota the orbit is incomplete posteriorly, because the postorbital bar typically complete in squamates and formed by the postorbital and/or jugal bar, is missing. In some diplodactylids and carphodactylids the ventral process of the postorbitofrontal and the dorsal process of the coronoid (Fig. 5; Bauer, 1990) are enlarged and come into close approximation when the jaw is closed and in dry skeletal preparations of Rhachodactylus auriculatus these two bones may even contact one another. In other geckos the posterior border of the orbit is completed by a postocular tendon (Stephenson and Stephenson, 1956), and the adductor mandibulae superficialis jugalis (A2-Supj, Fig. 6A,B).
The postorbitofrontal is hypertrophied in Aristelliger (POF, Fig. 4A) and absent in Lygodactylus. In pygopodids the postorbitofrontal contacts or approaches the prefrontal, excluding or limiting participation of the frontal in the orbital margin (Fig. 7A–C; Boulenger, 1885). The only geckos where a contact between prefrontal and postorbitofrontal has been found are Phelsuma lineata (Fig. 7D) and P. madagascarensis (Evans, 2008). The postorbitofrontal of Pygopus, Delma and Lialis is pierced by one or two foramina (Stephenson, 1962; Kluge, 1976) and this has been postulated as indicative of its compound origin (Evans, 2008).
In Chondrodactylus bibronii (Rieppel, 1984a) and Ailuronyx seychellensis, the postorbitofrontal contacts the squamosal, forming a secondarily developed upper temporal arch, with no supra-temporal fenestra due to the apposition of these two bones against the parietal.
In the Gekkota, this bone is much reduced in length and the postorbital process is missing (Fig. 1), nevertheless its complete absence was not corroborated in any of the specimens examined. In the genus Homonota and some miniaturized sphaerodactylids this bone is minuscule. In limbed gekkotans, this bone always contacts the maxilla either ventrally (most geckos) or laterally (eublepharids, Pseudogonatodes, Lepidoblepharis, and Coleodactylus). There have been contradictory reports concerning the presence of the jugal in some pygopodids. From our review of papers dealing with cranial osteology of this group, we found that the confusion lies in the identification of this bone in Delma and Lialis burtonis (Table 1). Two factors that might contribute to this confusion are the higher propensity for the dry skulls of pygopodids to disarticulate, compared with those of geckos, and the topographic position of this bone. Pygopodids differ from fully limbed gekkotans in that the jugal does not approach the anterior margin of the orbit (Evans, 2008), being located more posteriorly and squeezed anteriorly between the maxilla and the ectopterygoid. In Delma borea (USNM 128679, Fig. 8), the jugal is present and firmly contacts the ectopterygoid medially, forming a lateral elevated flange to the flat ectoperygoid. In Delma as in the rest of gekkotans, the ectoperygoid is flat. The ectopterygoid in Lialis burtonis (FMNH 166958) presents an elevated lateral wall, and in this specimen the jugal appears to be absent (Fig. 8B,E). A closer look at the digital sections of the ectopterygoid in both sagittal and lateral planes revealed a lengthwise putative suture, which in some places creates a hollow space (Fig. 8C). On the surface of the ectopterygoid of a skeletonized specimen (JFBM 15829) a longitudinal mark that may indicate a close contact suture between the ectopterygoid and the jugal is clearly visible (Fig. 9). The complex shape of the ectopterygoid in L. burtonis and the presence of a putative suture strongly suggest that in this species, the jugal is not lost, but instead is fused to the ectopterygoid. A detail of the articulation of the ectopterygoid and pterygoid is provided with our interpretation of the portion that corresponds to the jugal (Fig. 8). Additional evidence for the presence of the jugal in Lialis is provided by the muscle adductor mandibulae superficialis jugalis (A2-Supj, Fig. 6). This muscle is attached to the jugal bone in gekkotans (Daza, 2005). In limbed gekkotans, it is immediately behind the eye, bordering the orbit posteriorly (Fig. 6A). In Lialis, this muscle is attached to the lateral flange of the ectopterygoid. A space between the eye and this muscle indicates that the muscle has been displaced posteriorly in association with the relative development of the jugal in pygopodids.
Apparently, the Eublepharidae, uniquely among gekkotans, retain a discrete and unambiguously identifiable lacrimal. Although we did not find it in other species in our sample (the ossifications in P. bicolor are probably an abnormal structural development, see above), it is possible that this element is easily lost during traditional dermestid prepared specimens that comprise the majority of specimens reviewed. The lack of lacrimal in the majority of gekkotans, together with the inconsistent contact of the jugal and the prefrontal generates variation in the configuration of the lacrimal foramen among gekkotans. This does not affect the anterior pathway of the lacrimal duct from the lower posterior side of the lower lid (eublepharids) or spectacle (spectacled gekkotans) toward the palate.
The presence of a palpebral has been proposed by Conrad (2008) as a synapomorphy for Scincogekkonomorpha. This is due to his scoring of palpebral presence in the fossil taxa Ardeosaurus and Bavarisaurus, which he considered basal members of this clade. However, this bone is not present in either Ardeosaurus or Bavarisaurus (Camp, 1923; Cocude-Michel, 1961; Hoffstetter, 1964; Mateer, 1982; Evans, 2003, 2008) being restricted mostly to autarchoglossans. The only scincogekkonomorphs that present comparable structures are members of the sphaerodactylid subclade of crown gekkotans, which present the parafrontal bones. In these taxa, the structure and organization of parafrontal bones differs considerably with the single palpebral bone (Bauer and Russell, 1989). By position and structure, these bones are not homologous, which falsifies this character as a synapomorphy for Scincogekkonomorpha.
The postorbitofrontal is present in almost all gekkotans and can be hypertrophied (e.g., Aristelliger) or reduced (e.g., Lialis), being lost only in Lygodactylus of the taxa examined to date. Evans (2008) concluded that the single ossification in the dorsoposterior corner of the orbit in “geckos” is the postfrontal and that in pygopodids it is a compound bone formed by postfrontal and postorbital. She used two arguments for considering the bone of pygopodids as compound, the observation of two elements in Lialis jicari (Rieppel, 1984a) and the presence of one or more foramina in this bone in some species of Delma, Lialis and Pygopus. Since pygopodids are nested within limbed geckos, we believe that these arguments might apply to the whole clade and we recommend that the term postorbitofrontal also be applied to this bone in limbed gekkotans.
The postorbitofrontal resembles the postfrontal of some squamates in which the two elements are present (e.g., Iguania, Anolis occultus, gobiguanians; autarchoglossans, Cordylus giganteus; teiioids; the fossil gekkonomorph AMNH FR 21444; Conrad and Norell, 2006) and to the element remaining in some other squamates generally regarded as lacking a postorbital (e.g., Lanthanotus borneensis, Acontias meleagris) (J. Conrad, personal communication). However, the identity of this bone is open to discussion and the terms postorbital or postfrontal have been used interchangeably in gekkotan osteology (see discussion in Daza et al., 2008), and we consider that the term postorbitofrontal reflects the most conservative interpretation of the element as it does not imply the loss of either element.
Contact between the postorbitofrontal and the prefrontal, a condition described by Boulenger (1885), is a feature that distinguishes pygopodids from other geckos. McDowell and Bogert (1954) contended that this observation did not apply to Aristelliger (and possibly other geckos). In this particular case, we think that McDowell and Bogert were referring to the bridge between these two elements formed by the parafrontal bones (labeled by them as palpebral bones). In this species, the prefrontal and postorbitofrontal are separated and the parafrontal bones create the illusion of a continuous structure. The only geckos where a contact between prefrontal and postorbitofrontal has been found are Phelsuma lineata and P. madagascarensis and is very likely to be widespread across Phelsuma. The fact that Phelsuma and Lygodactylus (a genus without postorbitofrontal) are sister taxa (Kluge and Nussbaum, 1995; Krüger, 2001; Austin et al., 2004; Han et al., 2004; Feng et al., 2007; Raxworthy et al., 2007) demonstrates the variability of the circumorbital bones, even among closely related taxa.
The approximation of the ventral process of the postorbitofrontal and the coronoid process in Rhacodactylus produces an incomplete bony bar in the posterior border of the orbit. This should be not considered as a secondary formed postorbital bar because it involves an element from the mandible, but it does provide a partial skeletal separation of the eye and the jaw musculature in these geckos.
In related gekkonomorphs where the orbit is complete (e.g., AMNH FR 21444; Conrad and Norell, 2006) or nearly complete (e.g., Ardeosaurus brevipes; Camp, 1923; Cocude-Michel, 1961; Hoffstetter, 1964; Mateer, 1982; Evans, 2003, 2008), the jugal forms most of the postorbital bar. The reduction in the size of the jugal, is the main feature responsible for an incomplete orbit in the Gekkota (Evans, 2008). With the data from dissections and high resolution X-ray computed tomographies we demonstrated that this element is present in all gekkotans examined, and that in Lialis burtonis it is fused to the ectopterygoid. We based this identification in the topology criterion (principe des connexions, Geoffroy Saint-Hilaire, 1818) and the criterion of special quality. The portion of the ectopterygoid that we homologize with the jugal fulfills the properties of this bone in all squamates (Rieppel and Kearney, 2002) in having a relatively slender, three-dimensional structure, and having its anterior end taper off along the dorsomedial surface of the maxilla (unless participating in a well defined contact between the maxilla, lacrimal and prefrontal). The jugal is crucial for gekkotans and possibly varanids, because it serves as an anchor point for the postorbital tendon and the muscle adductor mandibulae superficialis jugalis (A2-Supj, Fig. 4A,B), which extends between the jugal and the postobitofrontal. An alternative interpretation would be the migration of the attachment to an adjacent point onto a contiguous element (e.g., the ectopterygoid) in association with the decrease in size and eventual loss of the jugal. We favor the first explanation given the distinct morphology of the ectoperygoid in Lialis and the presence of putative sutures and hollow space, easily seen with the different preparations. For this reason we suspect that this presents a mechanical restriction that prevents this element from being lost. This hypothesis predicts that the loss of the jugal in other taxa should only occur if the muscle is not present. For the sake of this argument, we can conclude that in all lizards where this muscle has been identified (Agama, Uromastyx, gekkotans, Xantusiidae and Varanidae) the jugal is present, and in the taxa where the jugal is lost, this muscle is absent.
The orbit of the gekkotans is highly modified compared with the rest of Lepidosauria. Our survey of gekkotan circumorbital bones has allowed us to identify some discrepancies with previous morphological analyses attributable to overlooked bones or differences in the interpretation among authors.
The phylogenetic implications of this are important for optimization of morphological characters in future cladistic analysis. The results from this analysis imply that supraorbital ossifications do not support the Scincogekkonomorpha and that special attention to this ossification across the Squamata is required in order to determine its homology. The retention of the jugal is universal among gekkotans, although it is often much reduced and may be fused to the ectopterygoid in some pygopodids. Pygopods also exhibit a derived condition in the contact of the prefrontal and postorbitofrontal. The presence of a lacrimal occurs in only some eublepharid geckos and is absent in all examined representatives of other gekkotan families. Neither the limbed Pacific geckos (Diplodactylidae and Carphodactylidae) nor the highly diverse and species rich “modern” gekkotans (Sphaerodactylidae + Phyllodactylidae + Gekkonidae) are diagnosable by features of the orbital rim, but characters of this portion of the skull do support the monophyly of particular genera (e.g., Phelsuma, Lygodactylus) and may yet prove informative at the intergeneric level.
The authors thank K. de Queiroz from the National Museum of Natural History (Smithsonian Institution) for suggestions to earlier versions of this article and access to specimens. Jack Conrad from the American Museum of Natural History in New York and an anonymous referee provided interesting comments that improved the article. The authors also thank C. McCarthy from Natural History Museum in London, D. Frost and D. Kizirian from the American Museum of Natural History in New York, R. Thomas from the University of Puerto Rico, and T. Gamble from the James Ford Bell Museum, University of Minnesota (Saint Paul, USA) for access to specimens and equipment; L. Benavides and G. Hormiga for access to the photographic facilities at the George Washington University; and finally, A. Herrera from the same institution for discussing her ideas about the homology of the circumorbital bones in some atypical fossil lizards.
AMNH, American Museum of Natural History (New York, USA); BMNH, British Museum (Natural History, London, England); CAS, California Academy of Sciences (San Francisco, USA); CM, Carnegie Museum of Natural History (Pittsburgh, USA); FMNH, The Field Museum (Natural History) (Chicago, USA); JFBM, James Ford Bell Museum, University of Minnesota (Saint Paul, USA); MZSP, Museu de Zoologia da Universidade de São Paulo (São Paulo, Brazil); USNM, United States National Museum of Natural History (Washington, USA); YPM, Yale Peabody Museum of Natural History (New Haven, USA).