Correspondence to: Juan D. Daza, Biology Department, Villanova University, 800 Lancaster Avenue, Villanova PA 19085. E-mail: email@example.comPresent address: Department of Biological Sciences, Sam Houston State University, 1900 Avenue I, Huntsville TX 77341-2116, USA
Gekkota (geckos and pygopod lizards) is a diverse clade of lepidosaurs, represented today by nearly 1,500 described extant species (Bauer, 2013; Uetz et al., 2013). Despite this modern diversity, the group is poorly represented by documented pre-Quaternary fossils (Estes, 1983; Müller and Mödden, 2001; Augé, 2005). The last comprehensive revision of its fossil record was published thirty years ago (Estes, 1983), although material from the Paleogene of Europe has been reviewed more recently (Augé, 2005; Rage, 2013). In the last three decades, several advances have been made towards the understanding phylogenetic position of geckos among squamates (Estes et al., 1988; Lee, 1998; Townsend et al., 2004, 2011; Hugall et al., 2007; Conrad, 2008; Vidal and Hedges, 2009; Bolet and Evans, 2012; Gauthier et al., 2012; Wiens et al., 2012), and there have also been major revisions to the classification of group based on phylogenetic advances (Kluge, 1987; Bauer, 1990; Han et al., 2004; Gamble et al., 2008a,b; Pyron et al. 2013), including a recent multigene phylogeny of virtually all genera (Gamble et al., 2012).
From a paleontological perspective, there have been several major discoveries and changes in the interpretation of the earliest geckos and their relatives. Jurassic genera traditionally included in the Gekkota (e.g. Ardeosaurus, Bavarisaurus, Eichstaettisaurus, Paleolacerta, and Yabeinosaurus; Hoffstetter, 1962, 1964, 1967; Estes, 1983) are now considered to be stem scleroglossans (Evans, 2003; Conrad, 2008; Bolet and Evans, 2010, 2012). Isolated vertebrae from the Middle Jurassic have also been tentatively placed in the Gekkota (Evans, 1998). However, the genus Eichstaettisaurus from the Late Jurassic of Germany and the Early Cretaceous of Spain and Italy has been interpreted as member of the Gekkota by some authors (Evans, 1993, 1994; Gauthier et al., 2012). Its position is supported by at least seven characters, but some of these are neither characteristic of nor unique to geckos (e.g., subolfactory processes of frontal arch beneath the brain but do not contact in the midline, parietals paired, 10 or more premaxillary teeth, 31 or more maxillary teeth, and smooth skull bones; Evans, 2003, 2008; Evans et al., 2004; Daza et al., 2012a). Despite the disadvantage of apomorphy-based definitions, this type of definition is very useful for the study of fossil gekkotans because the phylogenetic positions of the majority of these fossils have not been incorporated into any analyses, thus, they cannot be objectively referred to any clade specified by node- or stem-based definitions.
Whereas some previously recognized fossil genera have been excluded from the Gekkota, new fossil taxa from the Mesozoic have been classified as unambiguous gekkotans, and these include material from the Lower to Late Cretaceous of Myanmar (the amber embedded Cretaceogekko burmae; Arnold and Poinar, 2008), Mongolia (Hoburogekko suchanovi and Gobekko cretacicus; Alifanov, 1989, 1990; Borsuk-Białynicka, 1990; Daza et al., 2012b, 2013a), and West Siberia (Skutschas, 2006). In the Cenozoic, gekkotans are intermittently represented and their fossil record is concentrated in Europe (Rage, 1978; Schleich, 1985; Augé, 2005; Fig. 1). The sole Paleocene record is based on undetermined material from Brazil (Estes, 1970, 1983).
Records from the Eocene include the Baltic amber gecko Yantarogekko balticus (Bauer et al., 2005), numerous specimens from France, including the type specimens of the Eocene Laonogekko lefevrei (Augé, 2003), Rhodanogekko viretti, Cadurogekko rugosus and C. piveteaui (Hoffstetter, 1946; Schleich, 1985; Augé, 2005), and undetermined material from Catalonia (Bolet and Evans, 2012), Sainte Néboule (Rage, 1978), and Condé-en Brie (Augé, 1990). Outside Europe, one vertebra and two left dentary fragments from California, USA, have also been interpreted as gekkotan (Schatzinger, 1975; Golz and Lillegraven, 1977).
Miocene material from Europe includes Gerandogekko arambourgi, G. gaillardi (Hoffstetter, 1946; Kluge, 1967a; Schleich, 1985), several taxa of the still-extant leaf-toed gecko genus Euleptes, known from Germany, the Czech Republic, and Slovakia (Estes, 1969; Müller, 2001; Müller and Mödden, 2001; Čerňanský and Bauer, 2010), Palaeogekko risgovienis from Steinberg in southern Germany (Schleich, 1987), and many indeterminate dentaries, maxillae, and vertebrae from the Miocene at Sansan in southwestern France (Augé and Rage, 2000). Miocene material from outside Europe is represented by disarticulated material from Morocco (Rage, 1976), several miniaturized Sphaerodactylus geckos in amber inclusions from Hispaniola (Böhme, 1984; Schlee, 1990; Kluge, 1995; Grimaldi et al., 2000; Daza and Bauer, 2012a, Daza et al., 2013b), the pygopod Pygopus hortulanus from the Early Miocene of Australia (Hutchinson, 1997), several disarticulated bones from New Zealand (Lee et al., 2009a), and an isolated dentary from Florida (Estes, 1963). Quaternary material is generally more abundant and easily assigned to modern species and will not be reviewed here (see Estes, 1983; Worthy, 1987; Pregill, 1993; Albino, 1996, 2005).
Additional paleontological evidence for the occurrence of geckos comes from reports of complete fossil eggs and eggshell fragments similar to those of extant geckos. In most squamates, the eggshell is generally flexible and poorly mineralized (Kohring, 1991; Pianka and Vitt, 2003), but some geckos produce hard-shelled eggs (Mikhailov 1991, 1997; Andrews, 2004; Kratochvíl and Frynta, 2005)—the spheroidal gekkonoid type egg with a shell organization with rigid layers composed of jagged columns, which are classified in the family Gekkoolithidae in fossil egg parataxonomy (Hirsch, 1996). This type of hard-shelled egg is a derived feature only present among the crown group Gekkota in the non-eublepharid gekkonoids (i.e., Sphaerodactylidae, Gekkonidae, and Phyllodactylidae; Bustard, 1965; Kluge, 1967a; Werner, 1972; Kratochvíl and Frynta, 2005). Fossil gecko-like eggs have been recovered from several localities around the world (see Schleich and Kästle, 1988; Kohring, 1991), the oldest of which are from the Lower Cretaceous of Mongolia (Alifanov, 1989, 1990) and Spain (Kohring, 1991).
Differing interpretations of the definitions and content of Gekkota and the related term Gekkonomorpha have clouded the discussion of putative geckos from the Mesozoic (Bauer, in press a,b; Daza et al., 2013a) but this issue is beyond the scope of the present paper. Most Cenozoic material previously assigned to Gekkota can be confidently placed within this group as presently construed by most authors; however, allocation to less inclusive subclades remains problematic. Until relatively recently, all “modern” geckos were placed within Gekkonidae, with Pygopodidae often considered its sister group (e.g., Kluge 1967a; Moffat 1973; Estes et al., 1988; Lee, 1998). Kluge (1987) recognized that pygopods are embedded within limbed geckos and expanded the use of Pygopodidae to the clade including pygopods and their limbed relatives in Australia and the southwest Pacific. He retained all other spectacle-bearing geckos in Gekkonidae, with eyelid geckos in Eublepharidae, interpreted by him as sister to other crown clade gekkotans. Subsequently, a series of papers based chiefly on molecular phylogenetic data (Han et al. 2004; Gamble et al., 2008a,b) have modified the classification of living gecko groups, recognizing seven families divided in two major clades: Pygopodoidea—Pygopodidae, Carphodactylidae, and Diplodactylidae; and Gekkonoidea—Eublepharidae, Sphaerodactylidae, Phyllodactylidae, and Gekkonidae. Interrelationships among these groups are now well-established (Vidal and Hedges, 2009; Gamble et al., 2011a, 2012; Wiens et al., 2012; Pyron et al., 2013) but fossil material previously allocated to “Gekkonidae” has not been critically reassessed in light of the new phylogenetic paradigm. Such specimens might be referable to Gekkonidae (sensu Gamble et al., 2008b, 2012), but the original familial designation for most described material more likely implies membership in the crown clade Gekkota, exclusive of pygopods.
Although all seven gekkotan families receive strong molecular support (e.g., Gamble et al., 2011a, 2012; Heinicke et al., 2011; Nielsen et al., 2011) morphologically based analyses have yet to establish sets of apomorphic characters of paleontological utility for all groups. Improvements in the standardization of osteological terminology (e.g., Häupl, 1980; Kluge, 1995; Abdala, 1996; Daza et al., 2008; Evans, 2008; Gamble et al., 2011b) and new sources of diagnostic characters, such as those identified through high-resolution X-ray computed tomography (e.g., Conrad and Norell, 2006; Daza et al., 2012c; Daza and Bauer, 2012b; Gauthier et al., 2012), have the potential to rectify this problem. Currently there is morphological evidence for the support of Pygopodidae (Kluge, 1976; Greer, 1989), Carphodactylidae and Diplodatylidae (Bauer, 1990), and Eublepharidae (Grismer, 1988), but among the non-eublepharid gekkonoids, morphological convergence as well as high intrafamilial diversity and heterogeneity makes it difficult to diagnose the component subclades (Kluge, 1987, 1995; Abdala, 1996; Abdala and Moro, 1996; Daza, 2008; Daza et al., 2009). Taxon sampling remains a stumbling block to the identification of relevant diagnostic features, but recent morphological revisions of large numbers of characters has revealed new evidence that is congruent with some of the clades previously supported only by molecular data (e.g., an L-shaped squamosal bone in Carphodactylidae + Pygopodidae, and a characteristic fifth metatarsal shape in the Sphaerodactylidae; Daza and Bauer, 2012a,b).
Imprecise taxonomic allocation of Cenozoic gekkotans has restricted the contribution of these fossils to gekkotan temporal and spatial history. Because the familial affinities of all of the Cenozoic fossil taxa except Sphaerodactylus spp. and Euleptes spp. are unknown, paleontological data have thus far revealed little about evolutionary patterns except that a gecko-like body form was geographically widespread by at least the Eocene. Given that this body form was established perhaps as early as the late Jurassic (i.e., in Eichstaettisaurus schroederi, which may be a close gekkotan relative; see Gauthier et al., 2012), the contribution of isolated and disarticulated Cenozoic fossils must be regarded as limited. Even the discovery of adhesive toe pads on an amber gecko from the earliest Eocene of Russia (Bauer et al., 2005) lost some of its significance with the subsequent discovery of a scansorial system on another amber gecko from the Cretaceous of Myanmar (Arnold and Poinar, 2008). This lack of taxonomic resolution even has implications for molecular systematics. Fossils are the preferred means of calibrating timetrees derived from molecular phylogenetic data. Phylogenetic placement of calibrating fossil taxa may greatly affect the implied ages of lineages in the tree. Thus, if Cretaceogekko is interpreted as a gekkonid sensu stricto, it will imply a greater divergence age for Gekkonidae than if it is interpreted only as a crown group gekkotan. Ideally, timetrees should employ calibrations that are distributed across the history of the group (Kumar and Hedges, 1998; Marjanović and Laurin, 2007; Hedges and Kumar, 2009; Vidal and Hedges, 2009; Pyron, 2011). The present lack of knowledge about the phylogenetic position of Mesozoic and pre-Quaternary Cenozoic fossil geckos, however, precludes the application of a robust temporal scale to an ever more finely resolved gekkotan tree of life. Calibrations used to the present are derived chiefly from Miocene taxa near the tips of branches or from Cretaceous gekkotans of uncertain affinity (e.g., Oliver and Sanders, 2009; Heinicke et al., 2001; Nielsen et al., 2011; Gamble et al., 2011a, 2012; Oliver et al., 2012).
Because the Cretaceous, Paleogene, and early Neogene are likely to encompass the period within which most living families of geckos evolved and diversified (Conrad and Norell, 2006; Gamble et al. 2011a), fossils from these periods may be especially enlightening about gecko evolution. We thus consider that a reevaluation of the gekkotan fossil record in the context of the current understanding of the group's phylogeny is needed. In this revision, we seek to provide a catalog of pre-Quaternary fossil gekkotans, including only those groups that have been consistently recovered as associated with the Gekkota (Conrad and Norell, 2006; Conrad, 2008; Bolet and Evans, 2010; Gauthier et al., 2012; Daza et al., 2012b; Fig. 1). In addition to a review of the literature, we provide new anatomical information about these fossils, including new morphological characters and comparisons of these fossils with a wide sampling of living representatives. In some cases, we were able to allocate these fossil taxa at a familial level (in the current classification scheme), a task that has often previously proved difficult or impossible (Kluge, 1967a; Müller, 2001). We hope to facilitate the ultimate inclusion of these materials in future phylogenetic studies and that this will, in turn, contribute to the more effective exploitation of paleontological data for dating inference and other applications to the understanding of gekkotan evolution.
MATERIALS AND METHODS
We assembled all data available on fossil gekkotans, which includes bibliographical information, primary data from fossils, digital photographs (Figs. 2–8), X-rays, and high-resolution X-ray computed tomography (HRXCT) scans. For a comprehensive list of comparative material, see Daza and Bauer (2010, 2012b). Traditionally prepared modern gekkotan specimens (i.e., cleared-and-stained, skeletonized) were also used for comparison (Wassersug, 1976; Hanken and Wassersug, 1981; Bauer, 1986) and HRXCT data from 60 specimens of living taxa were obtained from material scanned at the HRXCT facilities of the University of Texas at Austin, the Museum of Comparative Zoology at Harvard University, and the American Museum of Natural History in New York. These data are available on Digimorph (http://www.digimorph.org) or were commissioned by the second author and will eventually be available at this site. Digital X-rays of the holotypes of Sphaerodactylus ciguapa and S. dommeli were obtained at the Academy of Sciences, Philadelphia, and the Zoologisches Forschungsmuseum Alexander Koenig, Bonn, respectively. Individual bones from comparative HRXCT material were isolated from the skull with the segmentation tools implemented in Avizo® 6.3.1 (VSG, Visualization Sciences Group, Burlington, Massachusetts, USA). Digital pictures were obtained with a Nikon D80 SLR with 60 mm Micro Nikkor lens, using a tripod and typically long exposures (0.3–1.3 sec), or with a single Nikon remote flash above and to the left of the specimen, usually at F/8 and 1/80 sec. Laonogekko lefevrei, Cadurcogekko piveteaui, C. rugosus, and Gerandogekko arambourgi (Figs. 3, 5, 6) were photographed at the Muséum national d'Historie naturelle de Paris. Rhodanogekko viretti and Gerandogekko gaillardi specimens (Figs. 4, 7) were photographed at the Départment de Sciences de la Terre, Centre de Conservation et d'Etude des Collections-Muséum d'histoire naturelle de Lyon (Départment du Rhône).
For each one of the fossil species, rather than providing redescriptions, we present the major diagnostic features discussed in previous morphological descriptions, but complement these with additional details and illustrations. For each fossil taxon, we provide a standardized revision, including diagnostic morphological characters that support their placement at the least inclusive taxonomic rank possible. Similarities with modern taxa are noted within the context of our sampling and this does not necessarily imply particularly close relationship or preclude that other living gekkotans might show a similar condition.
Codes for Institutional Collections
AMNH, American Museum of Natural History, New York, USA.
BSP, Bayerische Staatssammlung fuür Paläontologie und historische Geologie, Munich, Germany.
CAS, California Academy of Sciences, San Francisco, USA.
GAM, Deutsches Bernstein-Museum in Ribnitz-Damgarten, Germany.
GPIM, Institut für Geowissenschaften der Johannes-Gutenberg-Universität Mainz, Germany.
IPS, Institut de Paleontologia de Sabadell (now Institut Català de Paleontologia Miquel Crusafont), Barcelona, Spain.
MCZ, Museum of Comparative Zoology, Harvard University, Cambridge, USA.
MNHN, Muséum National d'Histoire Naturelle, Paris, France.
NHMB, Museum für Naturkunde, Berlin, Germany.
NMNZ, Museum of New Zealand Te Papa Tongarewa, Wellington, New Zealand.
NHMUK, Natural History Museum, London, United Kindom.
PIN, Paleontological Institute, Russian Academy of Sciences, Moscow, Russia.
PIUW, Paläontologisches Institut, Universität Wien, Austria.
QMF, Queensland Museum Palaeontology, Brisbane, Australia.
SGDB, Geological collection of the Bílina opencast mine, Czech Republic.
SMNS, Staatliches Museum für Naturkunde Stuttgart, Germany.
UCBL, Université Claude Bernard-Lyon, Muséum d'histoire naturelle de Lyon, France.
UCMP, University of California Museum of Paleontology, Berkeley, USA.
USNM, National Museum of Natural History, Smithsonian Institution, Washington DC, USA.
ZFMK, Zoologisches Forschungsmuseum Alexander Koenig, Bonn, Germany.
ZPAL, Institute of Paleobiology, Polish Academy of Sciences, Warsaw, Poland.
RESULTS AND DISCUSSION
Commonly Preserved Fossil Gekkotan Elements
The specimens exhibit different degrees of preservation; thus a detailed comparison among all material is not possible. Frontals, maxillae, and dentaries are the most commonly represented bones in the gekkotan fossil record.
The frontal bone displays marked size and morphological diversity among Gekkota and this is reflected in fossil material. The smallest frontal is that of Sphaerodactylus dommeli (2.09 mm), which is comparable in size to extant miniaturized sphaerodactyls (Daza et al., 2008; Gamble et al., 2011b) or to the burrowing pygopod Aprasia (McDowell and Bogert, 1954; Parker, 1956; Stephenson, 1962), forms having tiny skulls around 4–7 mm long (Rieppel, 1984a, Daza et al., 2008; Gamble et al., 2011b). The largest frontals are those of Rhodanogekko viretti (<13.35 mm) and Cadurcogekko piveteaui (<11.75 mm) which, based on this bone, are estimated to have had skulls as large as those of Thecadactylus rapicauda, Phelsuma grandis, or Pseudothecadactylus australis which might reach 120 mm (SVL), with a skull length of about 29 mm. Among gekkotans the subofactory process of the frontal arches beneath the brain, contacting and fusing ventrally to form a tubular structure for the tracts of cranial nerve I (Wellborn, 1933; McDowell and Bogert, 1954; Kluge, 1962; Bauer, 1990; Rieppel, 1984b; Evans, 2008), although there are a few exceptions (e.g., the Cretaceous Gobekko and the modified extant sphaerodactylid genus Pristurus). The shape of frontal is variable among gekkotans; this bone is generally unpaired, but a suture is visible in Gobekko cretacicus, Coleodactylus brachystoma, Teratoscincus, and some Lygodactylus and Phelsuma species. The anterior border is generally very narrow compared with the posterior border, especially in eublepharids, in contrast with pygopods, which have a very modified skull in which these two borders are subequal in width. The interorbital constriction is variable, with one of the relatively narrowest in the fossil Rhodanogekko viretti. In most gekkotans, the frontal participates in the formation of the orbit but is excluded from it in pygopods and Phelsuma by secondary contact of the prefrontal and postorbitofrontal (Boulenger, 1885; Daza and Bauer, 2010), and in Teratoscincus and Aristelliger by the neomorphic parafrontal bones (Bauer and Russell, 1989a).
The maxilla is also highly heterogeneous among gekkotans, with variation in the number of tooth loci, height and width of the facial process, degree of overlap with the prefrontal, and number of alveolar foramina. Tooth loci are, in general, more numerous among pygopodoids, which typically have more than 25 teeth, and some have more than 60 (Saltuarius cornutus, see also Bauer, 1986). The tooth count is also exceptionally high in the eublepharid Aeluroscalabotes and the gekkonid genus Uroplatus; in the latter, the combined number of maxillary and half of the premaxilla tooth loci can be as high as 78 (Uroplatus fimbriatus NHMUK 1922.214.171.124; Bauer and Russell, 1989b). Maxillae from adults of Aprasia are typically toothless (McDowell and Bogert, 1954, Parker, 1956; Stephenson, 1962), although Evans (2008) mentions the presence of a few teeth in juveniles. Variation in the shape of the facial process influences the amount of contact with the nasal, frontal, and prefrontal. Gobekko cretacicus, pygopodoids, and eublepharids mostly have a low facial process, which develops moderate overlap with the prefrontal, whereas Hoburogekko suchanovi and the non-eublepharid gekkonoids exhibit a higher facial process with more extensive overlap of the prefrontal.
The dentary in virtually all gekkotans and several other families is a tubular structure formed by overgrowth of the groove for the Meckel's cartilage (i.e., the Meckelian canal), which reduces the participation of the splenial bone in the Meckelian wall in the dentary (Kluge, 1967a; Estes et al., 1988; Daza et al., 2008). Hoburogekko suchanovi is the only known gekkotan in which the Meckelian canal remains unfused posteriorly; in this species, the dentary shelf for the postdentary bones extends further anteriorly than in other gekkotans (Daza et al. 2012b).
In the next section, we provide a review of the fossil material referable to the Gekkota, and for each we provide a systematic section in which we provide a provisional classification based on diagnostic characters shared with the comparative material. Material of uncertain placement within Gekkota is indicated as incertae sedis.
Systematic Classification of the Pre-Quaternary Gekkotans
Remarks: The fossil record of the Gekkota consists chiefly of isolated bones. This limits meaningful comparison to a few elements preserved in most of the specimens, but there is not a single one that is universally preserved in all the known fossils (Table 1).
Table 1. Skeleletal elements preserved in fossil gekkotans indicated as darker cells
The fossil material reviewed here is assigned to the Gekkota based on the following diagnostic features: (1) subolfactory process of frontal in contact at midline, observable in almost all gekkotans except Gobekko cretacicus, which exhibits the plesiomorphic squamate condition—a midventral gap between the processes, (2) parietal foramen absent, observable in Gobekko cretacicus, and amber-preserved Sphaerodactylus, (3) postorbital absent or fused to the postfrontal, observable in S. dommeli, apparently present in G. cretacicus, (4) Meckelian canal closed and fused, observable in all gekkotans except in H. suchanovi, (5) eyelids fused into a spectacle or brille, observable in Y. balticus and the Sphaerodactylus spp. in amber, and all extant gekkotans except in the Eublepharidae, (6) specialized subdigital structure, observable in Cretaceogekko burmae, Y. balticus, and the Sphaerodactylus geckos in amber, but with padless representatives in every gekkotan family and obviously in the nearly limbless pygopods (Russell, 1972; Gamble et al., 2008b, 2012).
There are four small to medium sized taxa referred to this clade from this age, Cretaceogekko burmae, Hoburogekko suchanovi, Gobekko cretacicus, and an unnamed specimen from West Siberia. Another fossil that has been referred to the Gekkonomorpha (Conrad and Norell, 2006; Conrad, 2008; Daza et al., 2012b; Gauthier et al., 2012) is the unnamed specimen AMNH FR21444, but its phylogenetic position is highly unstable (Daza et al., 2013a) and its morphology requires reconsideration (J. L. Conrad, personal communication).
Cretaceogekko burmae is the oldest known gekkotan; the material is contained in a piece of amber and shows anatomical structures similar to extant gekkotans (Arnold and Poinar, 2008). Hoburogekko suchanovi and G. cretacicus are represented by cranial elements and despite their great age, they are unusually intact and articulated. These two species are accepted unambiguously as gekkotans (Alifanov, 1989; Borsuk-Białynicka, 1990). The completeness of the material also has facilitated their inclusion on recent phylogenetic analyses, revealing that these two fossils are not assignable to any of the extant subclades (Conrad and Norell, 2006; Conrad, 2008; Daza et al., 2012b, 2013a).
Cretaceogekko burmae Arnold and Poinar, 2008
Classification: Incertae sedis.
Age: Late Early Cretaceous, probably Upper Albian or older, 97–100 Ma (Bisulca et al., 2012).
Locality: Obtained from the mine Noije Bum 2001 Summit Site, in the Hujawng Valley, southwest of Maingkhwan, in the state of Kachin (26°20′N, 96°36′E), Myanmar (Fig. 1).
Holotype: George O. Poinar personal collection # B–V–4, a partial specimen preserved in a piece of amber presenting the basal part of tail, left hind foot and lower crus, and isolated partial right foot, including some toes, which apparently suffered some decay prior to embedding, toes II–IV are translucent (described as “ghostly” in original description), distal portion of toe V partially covered by the tail.
Published figures: Arnold and Poinar (2008, figures 1-5).
Remarks: The specimen is estimated to be 15 mm or smaller in snout-vent length, being as small as Sphaeroactylus ariasae from Isla Beata in the Dominican Republic (14.1–17.9 mm), the smallest extant gecko (Hedges and Thomas, 2001). In contrast to the sphaerodactyl geckos which are generally leaf litter dwellers (Parker, 1926; Kluge, 1995; Vitt et al., 2005; Gamble et al., 2011b) the toes of C. burmae have a large number of lamellae under the toes suggesting that this species was capable of effective scansorial locomotion (Arnold and Poinar, 2008). Other groups of squamates have developed similar toepads (e.g., lygosomine skinks and dactyloid anoles). Among skinks, Lipnia leptosoma and three species of Prasinohaema have evolved toe pads with surface folds and ruffles and only Prasinohaema virens has true setae, which are very similar to those of anoles in morphology, being shorter and simpler in construction than those of geckos (Maderson, 1964; Ruibal and Ernst, 1965; Williams and Peterson, 1982; Bauer and Good, 1986; Irschick et al., 1995; Pianka and Sweet, 2005). Further, the toe pads in anoles do not extend to the base of the toe and are more distally located (being broadest ventral to the antepenultimate phalanx) than in many geckos (De Queiroz et al., 1998; Russell, 2002; Arnold and Poinar, 2008). In Cretaceogekko digits I to IV are subequal (versus exhibiting a marked and steady increase in length—Arnold and Poinar, 2008), a characteristic typical of gekkotans but not other scansorial lizards. In addition, it exhibits ungual symmetry, which is typical in pad-bearing and climbing geckos (Russell and Bauer, 1989).
Systematics: This fossil was assigned provisionally to the Gekkonidae (Arnold and Poinar, 2008). This familial allocation was accepted by Heinicke et al. (2011), who considered that the toepad morphology resembled that of gekkonids more closely than that of phyllodactylids. They also cited the current distribution of Sphaerodactylidae and Phyllodactylidae, which today are distributed in west Gondwanan areas and absent from Southeast Asia (Gamble et al., 2008a,b; Gamble et al., 2011a). Although this approximation is reasonable, it requires that ancestral areas of these groups match current distributions, but among squamates there are many exceptions. For example, Iguanoidea is a primarily American group with some representatives in the Old World (Chalarodon, Oplurus, Brachylophus) which together with fossil remains from Asia and Europe indicate that this group was more widespread in the past (Augé, 2007; Smith, 2009). Likewise, Chamaeleonidae in Europe is today limited to coastal areas in the southwest part of the Iberian Peninsula and in the south of Greece (and these may be human introductions; see Themido, 1945, Böhme et al., 1998; Miraldo et al., 2005), but the fossil record indicates that they reached the Czech Republic and Germany during the Miocene (Moody and Roček, 1980; Čerňanský, 2010, 2011). A similar situation with the sub-Saharan endemic family Cordylidae, which is represented in lower Miocene deposits from Europe in the Czech Republic (Roček, 1984; Čerňanský, 2012). Thus, the placement of Cretaceogekko in the Gekkonidae by previous authors should be regarded as provisional, and we rather refer to it to the Gekkota. This fossil has been used previously as a calibration point for both the crown group Gekkota (Nielsen et al., 2011) and the clade Gekkonidae + Phyllodactylidae (Heinicke et al., 2011).
Hoburogekko suchanovi Alifanov, 1989
Classification: Incertae sedis.
Age: Lower Cretaceous, Aptian-Albian.
Locality: Know only from Höövor (Khobur), Guchin-Us District in the province of Övörkhangai, Gobi Desert, Mongolia (Fig. 1).
Holotype: PIN 3334–500 (Fig. 2A), a partial muzzle unit.
Referred material: PIN 3334/501/502–506, dentaries.
Remarks: The holotype includes a portion of the snout, including the maxillae, prefrontal, frontal, vomer, palatine, and an element in the narial area, previously identified as the septomaxilla, but later reinterpreted as an unknown element or as a foreign element not part of the skull (Daza et al., 2012b). Referred material includes dentary bones, which present the posteriorly open Meckelian canal.
Systematics: Originally this fossil was assigned to the Gekkonidae sensu lato (Alifanov, 1989), but in a recent analysis it was retrieved as part of a basal polytomy within the Gekkota (Daza et al., 2013a). The bone arrangement and shape of the snout is similar to extant terrestrial gekkotans (Agamura, Bunopus, Pristurus, Ptenopus, Teratoscincus), but it exhibits a combination of characters not present in extant gekkotans, such as the open Meckelian canal, very enlarged posterolateral process of vomer, and maxillae with broad shelves to accommodate a large jugal. In all these unusual morphological features, this fossil differs from living families and might represent an early radiation of the Gekkota not referable to any extant lineage.
Gobekko cretacicus Borsuk-Białynicka, 1990
Classification: Incertae sedis.
Age: Late Cretaceous, probably late Santonian and/or early Campanian.
Locality: Know only from Bayanzag (Flamming Cliffs), from sediments of the Djadochta Formation, in the province of Ömnögovi, Gobi Desert, Mongolia (Fig. 1).
Holotype: ZPAL MgR-1114 (Fig. 2B), an articulated skull presenting almost all bones of the cranium and jaw.
Referred material: ZPAL MgR-11/43, ZPAL MgR-IV47, two partial skulls (Fig. 2C–D).
Remarks: Some skull elements not clearly distinguished or absent are the septomaxilla, postorbitofrontal, quadrate, stapes, and the postdentary bones. Referred material includes two partial specimens, which include the snout and jaws, with one also having a minor part of the parietals and the braincase represented. HRXCT data revealed that G. cretacicus and Pristurus spp. are the only known gekkotans that have frontal bones with subolfactory processes arching beneath the brain but not contacting on the midline. The lack of a parietal fontanel suggests that the specimens of G. cretacicus are skeletally mature (Bellairs and Kamal, 1981; Maisano, 2001, 2002). Nevertheless, this taxon exhibits many paedomorphic features such as paired nasals, frontals, and parietal as in the sphaerodactylids Teratoscincus and Coleodactylus brachystoma. A persistent basicranial fenestra, suggested by Borsuk-Białynicka (1990), would be an additional paedomorphic feature only known in this taxon (Daza et al., 2013a).
Systematics: Originally this fossil was assigned to the Gekkonidae sensu lato (Borsuk-Białynicka, 1990), but it has recently been found to be part of a basal polytomy within the Gekkota (Conrad and Norell, 2006; Daza et al., 2013a) or the sister taxon of the Gekkota (Conrad, 2008). This taxon displays unusual morphological features and might represent a clade distinct from any of the extant families and from Hoburogekko.
Unnamed material Skutschas, 2006
Classification: Incertae sedis
Age: Early Cretaceous (Aptian-Albian).
Locality: Shestakovo 1, Kiya River, near Shestakovo settlement in Kemerovo Province, West Siberia, Russian Federation, 55° 54'S, 87° 56'E (Fig. 1; Averianov and Voronkevich, 2002).
Remarks: Disarticulated lizard vertebrae (amphicoelous) and small fragments of dentaries that were tentatively attributed to a new gekkotan genus and species (Skutchas, 2006). This material is currently under study and will be described elsewhere (P. P. Skutschas, personal communication).
The fossil remains from this period are mainly from Europe, although some material has been recovered from Brazil and the United States. Most of the material is disarticulated, except Yantarogekko balticus, which is an amber inclusion. These gekkotans are more similar to extant forms than are the fossils from the Cretaceous; some of the material shares features with the extant Diplodactylidae or Sphaerodactylidae, although this is based on a limited number of characters. Sampling of other families confidently excludes membership in Pygopodidae and Eublepharidae.
Unnamed material Estes, 1970
Classification: Incertae sedis
Age: Late Paleocene.
Locality: Itaboraí deposits, Rio de Janeiro State, Brazil (Fig. 1).
Referred material: Repository unknown (see Remarks).
Remarks: This material was mentioned by Estes (1970, 1983) and identified as Gekkonidae sensu lato. The material was described as too fragmentary to be identifiable to the genus level and there is no mention of the repository of this material. The presence of this material in the New World suggests that the geckos may have been present in South America prior to the opening of the Atlantic (Báez and Gasparini, 1979; Duellman, 1979). This is consistent with a molecular timetree which suggested that the presence of New World sphaerodactyls could be explained by the fragmentation of Gondwana. The arrival of the rest of the modern New World gecko fauna, however, has been explained via Cretaceous/Paleogene land bridges (e.g., Coleonyx), trans-Atlantic dispersal (e.g., phyllodactylids, Lygodactylus klugei, Hemidactylus palaichthus and Hemidactylus brasilianus), or human introductions (Hemidactylus mabouia, Hemidactylus haitianus) (Carranza and Arnold, 2006; Gamble et al., 2011a, 2008a). This might imply sphaerodactylid affinities for the Brazilian specimen, but this remains speculative given that the material has not been critically examined.
Yantarogekko balticus Bauer, Böhme and Weitschat, 2005
Holotype: GAM 1400, a piece of amber enclosing the anterior part of the body, including the head, cervical and thoracic regions and the right forelimb.
Published figures: Bauer et al. (2005, figures 1-3); Weitschat and Wichard (2010, figure 3).
Remarks: The eye is covered by a brille, and its hand bears strongly clawed digits, well-developed, undivided subdigital scansors, and highly asymmetrical digits of the manus (digits I and II much smaller than digits III–V). One side of the head suggests that the skull may be partly eroded or destroyed, but phalangeal elements are visible in the hand where several digits have been severed.
Systematics: The lack of movable eyelids and presence of well-developed scansors excludes placement of Yantarogekko within the Eublepharidae or Carphodactylidae. Today four families include padded representatives: Diplodactylidae, Sphaerodactylidae, Phyllodactylidae, and Gekkonidae, the first being today geographically restricted to Australia, New Zealand, and New Caledonia, and the rest widespread, each including taxa in Mediterranean areas of Europe. Although current geographic distribution is no guarantee of ancestral range, it does appear that the pygopodoid families evolved in relative isolation in the Australasian region (Oliver and Sanders, 2009) and on this basis we regard Yantarogekko as a non-eublepharid gekkonoid, as did Bauer et al. (2005). However, external features of the holotype do not suggest particular affinities to any of the extant groups. An evaluation of the internal anatomy might provide additional information about its familial allocation.
Laonogekko lefevrei Augé, 2003
Classification: Incertae sedis
Age: Lower Eocene.
Locality: Prémontré, Aisne department, eastern Paris Basin, France (Fig. 1).
Holotype: MNHN PMT 5 (Fig. 3A–B), a right maxilla.
Referred material: MNHN PMT 6 (Fig. 3C–D), PMT 7 (Fig. 3E–F), PMT 32, PMT 70–78, b include nine dentaries (five right and four left), two frontals, and five vertebrae (four presacrals and one caudal); it has been estimated that this material represents a minimum of six individuals (Augé, 2003).
Remarks: Augé (2003) diagnosed this species from other gekkotans on the basis of two characters of the maxilla: (1) the presence of an inner ledge along the anterior border of the dorsal process (i.e., facial process) and (2) a notably low sloping posterior edge of the facial process (lspe, Fig. 3A–B, G). Augé also noted that this form was smaller than Cadurcogekko and differs in aspects of the facial process of the maxilla. Augé (2005) specified that Laonogekko, Cadurcogekko, and Rhodanogekko were allied to the Gekkoninae (equivalent to the currently recognized clade Sphaerodactylidae + Phyllodactylidae+ Gekkonidae).
In general, gekkotans have a tall facial process of the maxilla with a steep posterior sloping angle (Schleich, 1987; Augé, 2005), but a low and wide facial process of the maxilla is a plesiomorphic character for diplodactylids and carphodactylids (Bauer, 1990). The overall outline of the facial process of the holotype of L. lefevrei resembles that of the diplodactylids Dactylocnemis pacificus (Fig. 3G), Hoplodactylus duvaucelii, Woodworthia maculata, and, to a lesser extent, Eurydactylodes vieillardi. In these diplodactylids, the facial process is low and forms a concave posterior edge for the nares, from which it rises at an oblique angle, reaching its maximum height at one third the length of the maxilla from its anterior extent. From that point, the facial process slopes down gradually to the posterior end of the maxilla. The posterior margin of the facial process in L. lefevrei is uneven, but it is possible that the fossil is broken (Augé, 2003). The holotype of L. lefevrei has 38 maxillary tooth loci, a similar count to Strophurus ciliaris and Bavayia montana (between 35 and 39) and within the range of variation of the Diplodactylidae (between 30 and 44). The shape of the maxilla of L. lefevrei compared with extant gekkotans suggests that this bone had extensive contact with the nasal(s) bone(s), but only slightly overlapped the anterior portion of the prefrontal bone. In the anteromedial part of the maxilla, there is a remainder of the maxillary lappet, but this apparently was not developed to the extent of developing a broad bridge between the vomer and the premaxilla (Häupl, 1980; Kluge, 1995; Kluge and Nussbaum, 1995).
The dentary of L. lefevrei is typically gekkotan in having a closed Meckelian canal. One of the most complete dentaries assigned to L. lefevrei was reported as having 45 tooth loci (Augé, 2003). The dentary illustrated in the description is also figured here (PMT 6, MNHN; Fig. 3C–D); this bone exhibits a complete tooth row and tooth locus count of 33 (assuming minimal interdental space). Our estimation differs from the original description by 12 fewer positions, therefore falling within the range of variation of Woodworthia maculata and Lucasium damaeum (between 30 and 39 tooth loci). Some gekkonoids also have tooth counts that fall within this interval, for example, the eublepharids Coleonyx variegatus and Eublepharis macularius, the sphaerodactylids Lepidoblepharis xanthostigma and Pseudogonatodes guianensis, the phyllodactylids Phyllodactylus whirshingi, Phyllopezus pollicaris, and Ptyodactylus hasselquistii, and the gekkonids Hemidactylus haitianus and Pachydactylus namaquensis among many others.
Two frontal bones are referred to L. lefevrei; one of them is almost complete (PMT 32, MNHN; Augé, 2003), whereas the other one lacks a large portion of the anterior and posterior parts (PMT 7, MNHN; Augé, 2005; Fig. 3E–F). The frontals are wide, as in pygopodoids and many species of Gekkonidae (Kluge, 1967a; Bauer, 1990). These bones were described as being ornamented in a similar way to Caurcogekko piveteaui, having a series of longitudinal grooves in the anterior part that end up in some sort of network with some anastomosis towards the center of the bone (Augé, 2003). A partial frontal (PMT 7) shows that this sculpturing is the result of erosion of the periosteal bone from the element's dorsal surface, which exposes an irregularly furrowed and pitted spongy bone (Fig. 3E). This sculpturing is not exposed in the right anterolateral surface of this specimen, which still preserves a smooth bone surface with only some superficial furrows. On the HRXCT scans of medium- to large-size gekkotans, it is possible to remove the cortical bone digitally and expose the internal structure, which is very similar to the “sculpturing” in L. lefevrei. This internal structure is not developed in small gekkotans, where bone is more compact (e.g., Aprasia, Pseudogonatodes, Ptenopus).
One of the frontals (PTM 32) preserves the facets for the nasal(s) bone(s). The nasofrontal seam resembles a rounded “W” as in L. damaeum, Strophurus ciliaris, Naultinus elegans, and Pseudothecadactylus lindneri, but again, this is not a character exclusive to diplodactylids, and is also present in many gekkonoids, e.g., Paroedura bastardi, Ptenopus carpi, Aeluroscalabotes felinus, Eublepharis macularis, Euleptes europaea, and Homonota darwinii.
Both frontals of L. lefevrei have marked facets for the prefrontals on the lateroventral margin of the crista cranii. These facets extend to the midsection of the bone and indicate that prefrontals were strongly recessed into the frontal (Augé, 2003), in a similar way to some diplodactylids, carphodactylids, phyllodactylids, and gekkonids. The frontal of L. lefevrei does not bear the two anterior process of the crista cranii, which in diplodactylids and carphodactylids are paired and exhibit an expanded terminus (Bauer, 1990); however, this could be an artifact of damage to these delicate structures.
Systematics: The incompleteness of this material makes it difficult to assign confidently to any of the extant clades, but the combination of a low sloping posterior edge of the facial process of the maxilla, a wide frontal bone, tooth counts, and shape of the nasofrontal seam suggests some similarities with diplodactylid geckos, especially Dactylocnemis pacificus among the comparative taxa examined. Laonogekko lefevrei could also represent an extinct lineage that survived at least until the lower Eocene in Laurasia. Its placement in the Pygopodoidea is inconsistent with current distribution of this clade, which is restricted to Australia, New Caledonia, and New Zealand (Kluge, 1967a,b; Schleich, 1987). On the other hand, it is also possible that the Pygopodoidea was more widespread in the past. The oldest date for a Pygopodoidea-Gekkonoidea split extends to the Middle Jurassic, (95% credibility intervals of divergence times 113.8–175.4 Ma; Gamble et al., 2008a,b, 2011a), consistent with the fragmentation of Pangaea (Veevers, 2004); however, the median value is 144.6 Ma, at the very end of the Jurassic. Unamabiguous gekkotan fossils have not been confirmed before the Aptian (Early Cretaceous), although the genus Eichstaettisaurusis is known from the Late Jurassic of Solnhofen (Broili, 1938; Hoffstetter, 1953) to the Early Cretaceous of Catalonia (Bolet and Evans, 2010) and has been interpreted as having gekkotan affinities (Gauthier et al., 2012).
Rhodanogekko vireti Hoffstetter, 1946
Classification: Incertae sedis
Age: Middle Eocene.
Locality: Siderolitihic deposits (residual ironstones) from the Lutetian of Lissieu, Rhône department, France (Fig. 1).
Published figures: Hoffstetter (1946, figure 1A–C); Estes (1983, figure 14D).
Remarks: This species is known from a single frontal bone and was placed in the Gekkonidae sensu lato based on the subolfactory process of frontal in contact at the midline (Hoffstetter, 1946). This frontal indicates that this was a large animal (>100 mm SVL), with very large eyes. It has been pointed out that the frontal of Rhodanogekko vireti differs from that of other geckos in the highly constricted interorbital portion and the presence of bumps on the dorsal surface, which indicate the presence of osteoderms or at least irregularly sculptured dermal bone (Hoffstetter, 1946; Kluge, 1967a; Estes, 1983; Augé, 2005). Although these two characters appear independently in other gekkotan taxa, the combination of a narrow frontal bone and dermal sculpturing is unique to R. vireti among the Gekkota. Rhodanogekko vireti and Pristurus carteri have a similar frontal with a narrow interorbital constriction, a well-defined nasal shelf bounded posteriorly by a rounded and mostly medially located W-shaped nasofrontal seam (nfs, Fig. 4), and a nearly straight frontoparietal suture (fps, Fig. 4A), but the surface of the bone in the extant genus is smooth and ventrally it develops secondarily unfused frontal subolfactory processes (sop, Fig. 4D).
Among gekkonids, Rhoptropus afer and Hemidactylus lemurinus also have highly constricted frontals, but in both cases the frontoparietal suture is wider and exhibits a different nasofrontal seam. The anterodorsal end of the frontal of R. vireti appears spatulated, but part of this bone might be lost, including the anterolateral and anteromedial process of frontal (flp and fmp Fig. 4D). Another feature of this frontal bone is the presence of a well-developed anterior process of the crista cranii (apcc, Fig. 4B). This structure is absent in the Pygopodidae, Eublepharidae, and some gekkonid species (e.g., Ptenopus carpi, Rhoptropus afer, Tropiocolotes tripolitanus) and in Pristurus spp. (due to the modification noted above), but it is present in all members of Carphodactylidae, Diplodactylidae, Phyllodactylidae, and Spharodactylidae. In R. vireti this process resembles most of the majority of gekkonoids where this is narrow, contrasting with pygopodoid groups and some gekkonids (e.g. Cnemaspis), in which the terminus of this process is broadly expanded (Bauer, 1990).
Gekkotans are regarded as typically having smooth cranial bones (Hoffstetter, 1946; Estes, 1969; Gauthier et al., 2012), but there are some exceptions, among them the Madagascan gekkonids Paroedura bastardi and Matoatoa brevipes, and the New Caledonian diplodactylids Rhacodactylus trachyrhynchus and R. trachycephalus. Rhodanogekko vireti is exceptional in exhibiting an ornamented dorsal surface of the frontal (Hoffstetter, 1946), consisting of deep grooves and some irregular isolated cephalic rugosities.
Systematics: The overall shape of the Rhodanogekko frontals is very similar to the sphaerodactylid genus Pristurus (Fig. 4). In the frontals of R. vireti, the tracts for the cranial nerve I show some reduction in diameter and this could indicate a structural modification related to the highly constricted frontal. This morphology might represent an intermediate condition between the tubular frontals of the majority of gekkotans and the secondarily separated ventral downgrowths of the frontal seen in Pristurus (Fig. 4D). In both Rhodanogekko and Pristurus, the frontal bone has apomorphic features relative to gekkotans in general, but the overall similar shape could either be an indicator of close relationship between these two genera or simply a case of convergence.
Cadurcogekko piveteaui Hoffstetter, 1946
Classification: Incertae sedis
Age: Upper Eocene or Lower Oligocene.
Locality: Phosphorites de Quercy, Lot department, France, deposit not specified (Fig. 1).
Syntype: MNHN, No QU 17164, a right maxilla (Fig. 5A–B); MNHN, No QU 17163, a left dentary (Fig. 5E).
Referred material: abundant material including dentaries (Fig. 5F–G), one partial maxilla (Fig. 5C–D), one frontal bone (Fig. 5H–I) and several vertebrae (including thoracolumbars and caudals).
Cadurcogekko rugosus Augé, 2005
Classification: Incertae sedis
Age: Upper Eocene.
Locality: Sindou D, Phosphorites de Quercy, Lot department, France (Fig. 1).
Holotype: MNHN, No SND 622, a left dentary.
Referred material: a right maxilla, Les Pradigues (MP 17), USTL, PRA 9 (Fig. 5J–K), a fragment of the left dentary, a frontal bone, and two dorsal vertebrae.
Published figures: Augé (2005, figures 63a–b, 64a–b).
Remarks: The genus Cadurcogekko was diagnosed by Rage (1978) and Augé (2005) by four characters: (1) the rugose surface of the facial process of the maxilla (Fig. 5A, C, J), (2) an elongated groove associated with the most posterior vascular foramen of the maxilla (mxg, 4A, C, J), (3) a postnarial anterodorsal depression of the maxilla (mxd, Fig. 5A, J), (4) and ventrally grooved surface of the medial maxillary shelf.
The maxilla in C. piveteaui has a tall facial process, which is partially broken, but its projected outline is likely to have been triangular, suggesting extensive contact with the nasal bones(s) and a slight overlapping of the lateral surface of the prefrontal (Fig. 5A–B). The facial process of C. rugosus is more incomplete than in C. piveteaui, but the preserved portion is similar in these two species. In both species its lateral surface is rough and irregularly pitted, but the bone surface is coarser in C. rugosus. The pleurodont maxillary teeth of these two taxa are typical of gekkotan dentition, being straight with rounded crowns. Cadurcogekko piveteaui has 38 tooth loci (Augé, 2005), but considering that the posterior process is missing, it is estimated that the intact maxilla might have borne between 40 and 44, contrasting with only 27 tooth loci in C. rugosus.
The frontal bones in these two species are very wide (as in Laonogekko) but differ considerably in the amount of sculpturing; in C. piveteaui the dorsal surface has longitudinal furrows in the anterior and medial portions, and two shallow, obliquely oriented depressions separated by a posterior and medial low crest (sd, mdc, Fig. 5H), while in C. rugosus this pattern is more marked and the furrows are deeper. Ventrally, the frontals of both species exhibit a similar posterior notch of the subolfactory process. In C. piveteaui the lateral margins of the frontal are sharply creased (fc, Fig. 5I, L), and these define the outer edges of a pair of posteroventral recesses (fr, Fig. 5I, L). These recesses are well developed in many limbed pygopodoids (although not among pygopodids), being present in all diplodactylid genera reviewed and in some carphodactylids such as Nephrurus deleani, Phyllurus platurus, Carphodactylus laevis. Among the gekkonoids the development of these recesses is uncommon, but they occur in forms with massive skulls such as Haemodracon riebeckii. From the frontal of C. piveteaui it can be inferred that the nasofrontal seam was W-shaped and that the medial process was longer than the lateral process of the frontal (based on a small portion of the articulatory surface for the nasal bone). The nearly transverse frontoparietal suture indicates that the anterior margin of the parietal was straight. The dentary of Cadurcogekko piveteaui was originally described with other skull material as the fossil marsupial frog Amphignathodon sp. (Piveteau, 1927), but in a subsequent revision of this material Hoffstetter (1946) reassigned it to Gekkota. More material has been assigned to this taxon, including more dentaries, maxillaries, several vertebrae, and a frontal bone (Rage, 1988; Augé, 2005). Piveteau (1927) in his description of Amphignathodon noted that the dentary was proportionally larger than in extant anurans and stated that the association an ornamented cranium with a smooth jaw was not unusual as it is present in pelobatid frogs. This observation is also true in some gekkotan taxa with ornamented crania, including the two species of Cadurcogekko. The dentaries of C. piveteaui and C. rugosus differ greatly in length, height, number of tooth loci, and vascular foramina, C. piveteaui having a dentary that is long and low, bears about 49 tooth loci (our estimate is ∼40–45 based on the syntype material, Fig. 5E), instead of 26 tooth loci in the shorter dentary of C. rugosus (Augé, 2005). One unusual character in C. rugosus is the presence of a partially unfused Meckelian canal (Augé, 2005), which among extant gekkotans only occurs through late postovopositional embryonic stages.
Systematics: Kluge (1967a) concluded that neither Cadurcogekko nor Gerandogekko were related to the Sphaerodactylinae (i.e., Gonatodes, Lepidoblepharis, Sphaerodactylus, Coleodactylus, Pseudogonatodes, and Chatogekko, now collectively known as the sphaerodactyls) because these two fossil genera had well-defined impressions for the splenial bone, an element thought to be missing in the sphaerodactyls and Pristurus (see also Kluge, 1995). The presence of a splenial was reviewed recently in the Sphaerodactylidae (Daza and Bauer, 2012a), and an alternative explanation for the loss of this bone was proposed. Among the New World sphaerodactyls and the genus Pristurus the coronoid has an anteroventral foot-like expansion that occupies the position of the splenial (Daza et al., 2008). Since the splenial in all other Old World sphaerodactylids and in Aristelliger does not show any significant reduction in size, it is likely that the splenial, instead of being lost, fuses to the coronoid in Pristurus and the sphaerodactyls (Daza and Bauer, 2012a). The argument for the presence of a shelf for the coronoid in the above mentioned Paleogene fossil forms is weak, as whether or not this bone is lost or fused, the dentary would continue to provide some shelving for the postdentary bones.
Like Laonogekko lefevrei, members of the genus Cadurcogekko share several morphological similarities with limbed pygopodoids. Some extant pygopodoid species (e.g., Pseudothecadactylus australis, Carphodactylus laevis, Saltuarius cornutus, and Nephrurus asper) also have the combination of rugose dermal bones (a character not exclusive to this clade; e.g., also present in several gekkonids, including Chondrodactylus bibronii, Paroedura picta, and Matoatoa brevipes), an elongated groove associated with the posteriormost vascular foramen of the maxilla, and wide frontals with lateral crests and recesses (Fig. 5L).
Unnamed material Augé, 1990
Classification: Incertae sedis
Age: Standard level d'Avenay (MP8 + 9), Cuisien, Lower Eocene.
Locality: Condé-en Brie, Aisne Department, France (Fig. 1).
Remarks: One dentary assigned to the Gekkonidae (Augé, 1990).
Unnamed material Bolet and Evans, 2013
Classification: Incertae sedis
Age: Upper Eocene.
Locality: Sossis, La Pobla de Segur, Catalonia, Spain (Fig. 1).
Referred material: Several bone fragments from the dentary, maxilla, and premaxilla. The material was divided in three forms. Form A includes two fragments from the right dentary (IPS 56176, 56178) and a partial right maxilla (IPS 56065). Form B includes fragments from at least 14 dentaries (left: IPS 56132, 56133; right: IPS 56116, 56172, 56182–85, 56187, 56192; indeterminate: IPS 56185, 56188, 59499, 59511). Form C includes fragments from seven dentaries (left: IPS 56180, 56190; right: IPS 56117, 56173–75, 56177). Additional gekkotan material not assigned to these forms includes a partial fused premaxilla (IPS 59478) and an undetermined number of dentary fragments (left: IPS 56114, 56115; right: IPS 56179, 56181; indeterminate: IPS 56120, 56189, 56191, 56520).
Published figures: Bolet and Evans (2013, figure 3).
Remarks: The three forms described for this material are based on the size, interdental space, and shape of preserved teeth. The dentaries and the only maxilla preserved are so fragmentary that they are merely assigned to the Gekkota on the basis of general characters (e.g., fused Meckelian canal and cylindrical, unicuspid teeth).
Unnamed material Rage, 1978
Classification: Incertae sedis
Age: Upper Eocene.
Locality: La Poche A Phosphate, Sainte Néboule, Lot Department, France (Fig. 1).
Remarks: Two incomplete dentaries and one maxilla are known from this locality (Rage, 1978).
Unnamed material Schatzinger, 1975; Golz and Lillegraven, 1977
Classification: Undetermined Gekkota
Age: Upper Eocene, Mission Valley Formation.
Locality: Mission Valley Formation, San Diego County, California, United States (Fig. 1), three localities, Crumbling Slope (V71180), Poway Pipeline 1 (V72157), and Incredible Chaos (V72179).
Remarks: Three specimens were described from this locality, the centrum of an axis (UCMP 104296), and two fragments of the left dentary (UCMP 113225, 113226). This material is currently missing (Estes, 1983; UCMP, University of California Museum of Paleontology, 2012).
Miocene gekkotans are more similar to extant genera and comparisons allow classifying this material with more certainty in the Sphaerodactylidae, Diplodactylidae, and Pygopodidae, although some forms are referable Gekkonidae. These fossils are also predominantly European, but there are undetermined records from Morocco and United States, one record from a pygopodid from Australia, some disarticulated gecko bones from New Zealand, and several specimens embedded in amber from Hispaniola.
Gerandogekko Hoffstetter, 1946
Gerandogekko arambourgi Hoffstetter, 1946
Age: Lower Miocene, Aquitanian.
Locality: Saint-Gérand-le-Puy, Allier department, France (Fig. 1).
Holotype: MNHN [specimen number not cited in original description or subsequent published references], a frontal (Fig. 6A–B).
Referred material: MNHN, topotypic frontal (Fig. 6C–D) and maxilla.
Published figures: Hoffstetter (1946, figure 3A–C); Estes (1983, figure 14E).
Gerandogekko gaillardi Hoffstetter, 1946
Age: Upper Miocene, Vindobonian.
Locality: La Grive-Saint-Alban, Isère department, France (Fig. 1).
Referred material: UBCL, LGr5841–LGr5842, topotypic maxilla (Fig. 7C–D), several dentaries (Figs. 7E–F, 8A–R) and vertebrae (Fig. 7G–I).
Published figures: Hoffstetter (1946, figure 4A–D); Estes (1983, figure 14F).
Remarks: Hoffstetter (1946) described a new genus to include these two species, but he recognized that they were very modern in appearance and that they might belong to an extant genus. Comparison with extant species indicates that Gerandogekko might be related or referable to the sphaerodactylid genus Euleptes. The similarity between these two genera was pointed out previously by Estes (1969), but he decided to maintain a separate genus in part due to the limited material that he had available for comparison. Osteological material of E. europaea is relatively rare in herpetological collections and its mention in the literature is sporadic (Wiedersheim, 1875; Bauer et al., 1997). A HRXCT scan reconstruction of the skull of a specimen (Fig. 9) is thus provided for comparison with Gerandogekko and the fossil species of Euleptes.
The skull of Gerandogekko is larger than that of E. europaea, but the shapes of the maxilla and dentary in these two genera are similar (Figs. 6, 8). In both genera, the frontal is wide and has a marked interorbital constriction. The posterior end is twice the width of the anterior end (although broken in some specimens of Gerandogekko). The posterolateral margins are straight and oriented towards the interorbital constriction, at which point the edge becomes slightly convex and is more anteriorly oriented. In both genera the dorsal surface of the frontals is smooth, but whereas in E. europaea there are two oblique sulci parallel to the posterolateral edge of the bone, in G. arambourgi and G. gaillardi (su, Figs. 6A–C, E, 7A) these sulci are just insinuated. Although the anterodorsal part of the frontal is broken in all specimens of Gerandogekko, the posterior margin is slightly sinuous as in E. europaea. Ventrally the frontals show some slight differences in the orientation of the anterior processes of the crista cranii. In Gerandogekko these two process are oriented almost horizontally, whereas in Euleptes they are more oblique (apcc, Figs. 6, 7). The margin between the anterior process of the crista cranii is convex anteriorly in all specimens, except in the holotype of G. arambourgi which seems broken at this point (Fig. 6B). The posterior edge of the subolfactory process is concave in all specimens and from this edge, two crests extend toward the posterolateral ends of the element.
The maxillary remains in both species of Gerandogekko correspond to the mid part of the bone. In G. arambourgi this bone has a triangular palatal medial shelf (lamina horizontalis) that might have contacted the palatine. Because of incompleteness of the bone, it is not possible to compare the facial process with that of other Euleptes material (Estes, 1969). The portion of facial process of the maxilla in G. gaillardi (Fig. 7C) is not strongly pierced by vascular foramina as in Euleptes spp.
Most of the multiple dentaries attributed to G. gaillardi (Figs. 7E–F, 8A–R) are similar to those of Euleptes (i.e., number of mental foramina, dorsal and ventral edges, Meckelian notch at the same level, and dentition shape, size, and spacing). In the most complete dentaries (Figs. 7E–F, 8A–B), the posterior part is twice the depth of the anterior symphyseal part, as in Euleptes (Fig. 8S–T). The dorsal edge along the tooth row is mainly straight, while the ventral border is slightly curved. One of the specimens has a more slender dentary, with a more curved ventral border and the posteromedial notch of the Meckelian canal almost below the last dentary tooth (Fig. 8Q–R); at least this specimen may represent a different species.
The dentary in G. gaillardi is comparatively shorter than Euleptes, bearing 25–26 teeth, which is similar to E. gallica (28 tooth loci, Müller, 2001) or E. sp. (29 tooth loci, Müller and Mödden, 2001), and differing from E. europaea (31 tooth loci), although it is possible that the number of dentary tooth loci in G. gaillardi might be higher because the specimen is missing a portion near the mandibular symphysis. In the most complete specimen, the dentary has an anteroventral chip that exposes the ventral opening of the Meckelian canal. There are two to three mental foramina in the specimens, although in many of these, bones are incomplete and the exact number for each one cannot be determined. Nonetheless, based on the eight specimens in which some foramina were visible, the modal value was three, still one foramen less than E. europaea (four foramina), and two to three less than the fossil Euleptes material (five to six foramina; Müller, 2001; Müller and Mödden, 2001). The posterior part of the dentary is broken in the majority of specimens of G. gaillardi, but one specimen (Fig. 8A–B) exhibits a posterior border similar to that of Euleptes europaea (e.g., Fig. 8S–T), in which this bone develops three posterior processes: a short coronoid process (cop, Fig. 8A, T), a surangular process (sap Fig. 8A, T), and a longer compound bone process (cobp, Fig. 8A, T). These processes delineate two anterior notches that have an internal angle of nearly 90 degrees. The nearly complete dentary also indicates that the posteroventral angular process did not extend beyond the coronoid eminence, a condition also seen in many living gekkonoids (e.g., Gekko, Pachydactylus, and Thecadactylus). This is in contrast to some diplodactylids and eublepharids where this process extends more posteriorly (Augé, 2005), although not as far as in miniaturized sphaerodactyls (Daza et al., 2008; Gamble et al., 2011b).
The vertebrae of G. gaillardi are short and amphicoelous (Hoffstetter, 1946) and at least one is diagnosed as a thoracolumbar by the presence of synapophyses. The neural arches are low and rounded, and they have subcentral foramina.
Systematics: Kluge (1967a) eliminated any possible relationship between Gerandogekko and the Eublepharinae or the Sphaerodactylinae (as then construed) based on the extremely wide frontal and the presence of amphiceolous vertebrae. While these arguments are still valid for nearly all sphaerodactyls (except Gonatodes), it does not apply to other taxa now assigned to the Sphaerodactylidae, which do have amphicoelous vertebrae (Camp, 1923; Kluge, 1967a; Hoffstetter and Gasc, 1969; Werner, 1971; Moffat, 1973; Kluge, 1995). Euleptes europaea have wide frontals and amphicelous vertebrae similar to those of Gerandogekko, providing support for the placement of the latter genus within the Sphaerodactylidae. Nonetheless its current generic allocation may be justified by differences in size and dentary morphology.
Hoffstetter (1946) tentatively separated G. arambourgi and G. gaillardi, based largely on geographically and geologically separate occurrences, differences in orbital emargination of the frontal (Estes, 1983) and size (G. gaillardi smaller than G. arambourgi). Estes (1983) noticed that according to the scale bars in Hoffstetter's illustrations (1946) the size relationship was the inverse of that originally reported (i.e., G. gaillardi larger than G. arambourgi). We confirm Estes' observation; for instance, the subolfactory process of the frontal in the most complete specimens measures 2.66 mm in G. gaillardi and 1.86 mm in G. arambourgi. With the material available it is not possible to determine if the two taxa are conspecific. Differences in frontal emargination seem to be weak evidence because similar variation is present in the eight specimens reviewed of E. europaea.
Three fossil Euleptes have been described as E. gallica or E. sp. (Müller, 2001; Müller and Mödden, 2001; Čerňanský and Bauer, 2010), but the differentiation of these relies mainly on the shape of the premaxilla, which is not represented in known Gerandogekko material.
Euleptes Fitzinger, 1843
Euleptes gallica Müller, 2001
Age: Lower Miocene.
Locality: Type locality Montaigu-le-Blin, Allier Basin (MN 2), Allier department, France. Also known from Merkur-North locality, Ústí nad Labem Region, Czech Republic (Fig. 1).
Referred material: Paratypes, NHMB 1053 a premaxilla; NHMB 1054 a dentary. Other material assigned to this species includes Ah–875 SGDB a right dentary and Ah–931
SGDB, 932 SGDB, 933 SGDB three left maxillae (Čerňanský and Bauer, 2010).
Published figures: Müller (2001, figures 1A–F; 2A); Čerňanský and Bauer (2010, figures 1, 2A–B).
Remarks: This species has been assigned to the genus Euleptes on the basis of the presence of an acutely angled profile of the posterior edge of the osseus nares, which almost defines a right angle (Estes, 1969; Müller, 2001; Augé, 2005; Čerňanský and Bauer, 2010; Fig. 9). The overall domed outline of the facial process of the maxilla is also characteristic for Euleptes (Müller, 2001). Euleptes gallica also has been differentiated from other congeneric material by the shape of the ascending nasal process; in E. europaea this structure is slender with a triangular terminus (Bauer et al., 1997; Fig. 9), whereas fossils of E. gallica and E. sp. from Oppenheim Germany have a distinctly stouter and broader process (Müller, 2001).
Table 2 lists 14 characters used in osteological descriptions of Euleptes scored for all described fossil specimens and for the only extant species (Wiedersheim, 1875; Estes, 1969; Bauer et al., 1997; Müller, 2001; Müller and Mödden, 2001; Čerňanský and Bauer, 2010). Although incompleteness of the material only allows direct comparisons for four of the characters, it can be appreciated that this genus exhibits variation in shape of the maxilla, number of tooth loci, and foramina. In terms of size, all fossil maxillae are larger than adult specimens of E. europaea (Fig. 10); snout-vent length estimations for E. gallica range from 60 to 70 mm (Čerňanský and Bauer, 2010) to 85 mm (Müller, 2001). Considering the actual values of the maxilla and SVL lengths on the CT scanned specimen of E. europaea (MCZ R–4463, Table 2), we estimated SVL values for the fossil Euleptes between 47.94 and 58.21 mm. The specimen from Devínska Nová Ves (Estes, 1969) has differences in shape and fenestration in comparison with other Euleptes species, and on this basis we consider it likely to represent a separate species.
Table 2. Morphological variation of premaxilla, maxilla and body characters in the genus Euleptes
Vertical keel along facial process (Carina maxillaris)
Anterior tip of the facial process posterior to the nares
Triangular, prominent, angled
Number of foramina posterior to the nares
Number of supralabial alveolar foramina
Posterior supralabial foramen continued in a groove
Number of mental foramina
3 or more
Anterior opening of Meckelian canal with an anteriorly directed groove
Length of maxilla along toothed edge (mm)
Number of dentary teeth above the splenial notch
Number of maxillary teeth
Number of premaxillary teeth
Number of dentary teeth
Estimated SVL (mm)
Euleptessp.[cf.Phyllodactylussp. Estes, 1969]
Age: Middle Miocene (MN6 of the European land mammal biochronology).
Locality: Type Devínska Nová Ves near the borough of Bratislava, Slovakia (Fig. 1). This area has at least three rich Middle Miocene fossil sites – Sandberg (marine sediments), Bonanza, and Zapfe's fissure fillings or cracks (A. Čerňanský, personal communication). This material comes from the last of these.
Holotype: PIUW (Coll. Zapfe) a left maxilla (Fig. 10C; Estes, 1969, 1983).
Paratype: the anterior part of the left dentary up to the 10th tooth loci.
Published figures: Estes (1969, figure 1A–C); Estes (1983, figure 14G).
Remarks: A medium-sized gecko, estimated to measure 49.65 mm SVL, slightly larger than extant E. europaea (Fig. 9) and similar in size to E. gallica. The facial process of the maxilla is domed as in other Euleptes, with the posterior edge of the osseus nares acutely angled. It differs, however, from other congeners in having a rather rounded (versus angular) tip of the anterior margin of the facial process and a less steeply inclined posterior edge of the same structure. The facial process is also smooth and has fewer vascular foramina (maximum seven and arranged in a single row), and a less marked groove extending posteriorly from the last supralabial foramen than do other Euleptes. The vertical keel along the medial surface of the facial process is faint (Estes, 1969) as in other Euleptes specimens (Fig. 10), except E. gallica from North-West Bohemia (Čerňanský and Bauer, 2010). The referred incomplete dentary has the third mental foramina at the tenth tooth loci instead of position 12 or 13 (E. europaea, Fig. 9) or 8 or 9 (E. gallica, but see E. sp., Müller and Mödden, 2001).
Systematics: Estes (1969) differentiated this fossil from Euleptes europaea (as Phyllodactylus europaeus) based on a slightly taller and more posteriorly extended facial process of the maxilla but he did restrict its classification to cf. Phyllodatylus because of the limited comparative material available. He also discussed the similarity of this fossil to the extinct genus Gerandogekko (see above). We consider that this fossil belongs to the genus Euleptes, but a careful examination of this material is required in order to resolve its species allocation.
Referred material: GPIM N 2000–2002, 2005 a-W, 52 disarticulated skull elements, including premaxillae, maxillae, and dentaries.
Published figures: Müller and Mödden (2001, figure 1a–f).
Remarks: Comparison of the dentary of Euleptes sp. with the abundant material of Gerandogekko gaillardi indicates that the former is more slender and proportionally longer, with more mental foramina, more tooth loci, and lacking the discrete processes for the coronoid and the surangular present in Gerandogekko and E. europaea. Müller (2001) pointed out that the ascending process of the premaxilla in E. gallica and the Oppenheim Euleptes is much stouter than in E. europaea. There are also some differences in the margin of the same process among the fossil Euleptes, with the Oppenheim specimen having a more even lateral margin than E. gallica. Additional differences of Euleptes sp. from the other material includes the number of supralabial and mental alveolar foramina and a slight reduction in number of premaxillary and maxillary tooth loci (Table 2).
Locality: Neville's Garden Site on D Site Plateau at Riversleigh, northwest Queensland, Australia (Fig. 1).
Holotype: QMF 16875, a right dentary.
Published figures: Hutchinson (1997, figure 4A–D).
Remarks: This is the only Pre-Quaternary record for the family, otherwise only known from mid-Holocene deposits (Mead et al., 2008). Almost all mandibular characters used by Hutchinson (1997) to characterize the mandible of the genus Pygopus are observable in this fossil. Nonetheless, it differs from other pygopods in having a relatively short and deep dentary and fewer robust teeth, bearing 21 tooth loci (Hutchinson, 1997). Tooth locus counts from museum specimens and published illustrations of P. lepidopodus and Pygopus nigricans are similar to this fossil, ranging between 19–21 and 16–17, respectively (see also McDowell and Bogert, 1954; Greer, 1989; Hutchinson, 1997).
Pygopus hortulanus differs from extant Pygopus in having the three anterior dentary teeth more upright, not procumbent and projecting beyond the anterior end of the bone, a less marked difference in size between the anterior and the middle dentary teeth, five mental foramina (instead of 6 or 7), small symphyseal area (not vertically expanded), and the presence of a long and distinct groove leading forward from the anterior tip of the coronoid on the lateral dentary surface (Hutchinson, 1997; Lee et al., 2009b).
Systematics: Pygopus was previously interpreted as sister to remaining pygopod genera (Kluge, 1976), although recent analyses place it as the sister taxon of Paradelma (Jennings et al., 2003; Lee et al., 2009b). Hutchinson (1997) subjectively considered this fossil to be the sister of all other Pygopus. This was found to be the most probable phylogenetic position for this fossil based on a recent Bayesian analysis of molecular and morphological data, although it was also recovered in the sampled trees as the sister taxon of either Paradelma or Lialis (Lee et al., 2009b). The phylogenetic uncertainty associated with this fossil contributes to broad credibility intervals when it is used as a calibration to infer divergence dates (Lee et al., 2009b).
Unidentified material Lee et al., 2009a
Age: Lower Miocene
Locality: Manuherikia River, Bed HH1a, Bed HH1b, Bed HH4, Saint Bathans, Otago region, New Zealand (44° 54'28.6”S; 169° 51' 29.6”E), Lee et al. (2009a) (Fig. 1).
Referred specimens: Numerous cranial and postcranial elements, including a right and left maxillae (NMNZ S44152, NMNZ S51004), frontal (NMNZ S43124, Fig. 11A), left dentary and distal portion of femur (NMNZ S42296), left dentary (NMNZ S51324), right compound bone (NMNZ S42731, NMNZ S44003, NMNZ S44338, Fig. 11B), left compound bone (NMNZ S42341, NMNZ S42342, NMNZ S44153, NMNZ S42577), fragment of pterygoid (NMNZ S4415, Fig. 11C), dorsal vertebrae (NMNZ S42343, NMNZ S42616, NMNZ S42687, NMNZ S42730, NMNZ S43125), caudal vertebra (NMNZ S51269), partial humerus (NMNZ S44072), partial femur (NMNZ S50709).
Published figures: Lee et al. (2009a, figure 2a–o [grayscale], figure 2a–o (color), data supplement, appendix 3).
Remarks: The material was assigned to two forms of Gekkonidae sensu lato and was stated by the original authors to be intermediate in size between Hoplodactylus maculatus (now Woodworthia maculata) and Hoplodactylus duvaucelii (Lee et al., 2009a; Nielsen et al., 2011). This material is referable to the Diplodactylidae. This placement is supported by the presence of creased lateral margins and paired posteroventral recesses of the frontal. In the description of the material, these fossils were described as being more similar to members of the genus Hoplodactylus than to Naultilus, this based on similarities of the frontal bone (Lee et al., 2009a). The genus Hoplodactylus was recently found to be made paraphyletic by the exclusion of Naultinus and has since been restricted to H. duvaucelii and the extinct giant gecko H. delcourti. Five additional new or resurrected genera are currently recognized: Dactylocnemis, Woodworthia, Tukutuku, Toropuku, and Mokopirirakau (Nielsen et al., 2011). Among the comparative material available, these fossils share similarities with the extant species Woodworthia maculata including characters of the frontal, compound bone, pterygoid, and postcranial elements (Fig. 11), but the fossil elements are somewhat more heavily built. The fossil frontal bones share with W. maculata a smooth dorsal surface, lateral crests, slender posterolateral processes (plp, Fig. 11A, D), and a slight posteromedial projection (pmp, Fig. 11A, D), although the latter three similarities are also shared with the larger H. duvaucelii (Fig. 11G). The compound bone in all these three forms exhibits similar inclination of the articular facets, but the fossil material more closely resembles W. maculata in the similar development and orientation of the adductor fossa (af, Fig. 11B) and almost identical large and rounded medial processes (mp, Fig. 11B), but differs from it in having a shorter and stouter retroarticular process (rap, Fig. 11B), and in this regard is more similar to H. duvaucelii (Fig. 11H). Other similarities with W. maculata are the presence of a pterygoid with a medial concave margin of the interpterygoid vacuity (iv, Fig. 11C, not present in H. duvaucelii, Fig. 11I), a long facet for the basipterygid process, and a slender humeral shaft, which contrasts with the stouter element in H. duvaucelii.
Sphaerodactylus Wagler, 1830
Sphaerodactylus dommeli Böhme, 1984
Age: late Early Miocene to early Middle Miocene.
Locality: La Toca mine, Cordillera Septentrional, Santiago Province, north of Municipio Santiago de los Caballeros, Dominican Republic (Fig. 1).
Holotype: ZFMK 66238 (Fig. 12), an amber-embedded articulated (SLV 32.10 mm) specimen preserving the integument and nearly complete cranial and postcranial elements, tail vertebrae missing, possibly due to regeneration during life. An abdominal radiopaque element might be an egg-shell (Daza et al., 2013b).
Paratype: SMNS Do–3584 (Fig. 12), a much smaller articulated specimen (SVL 16.5 mm) with autotomized tail.
Published figures: Böhme (1984, figures 1-3); Schlee (1990, figure 15); Daza et al. (2013b, figures 1-3, 6).
Remarks: A medium-sized Sphaerodactylus with most of the integument intact, keeled granular scales on dorsum, and a disproportionally elongated cervical region.
Systematics: The allocation of this fossil to the genus Sphaerodactylus has been problematic, Kluge (1995; citing D. Frost, personal communication) in reference to both the holotype and paratype suggested that this species was probably an Anolis lizard. De Queiroz et al. (1998) commented that at least the holotype does not appear to be an anole. Recently Daza et al. (2013b) confirmed that Böhme (1984) correctly identified the holotype and the paratype as a member of Sphaerodactylus and that this fossil exhibits 11 diagnostic characters of Gekkota and eight additional characters for the genus (Daza and Bauer, 2012a).
Sphaerodactylus ciguapaDaza and Bauer, 2012a
Age: late Early Miocene to early Middle Miocene.
Locality: La Toca mine, Cordillera Septentrional, Santiago Province, north of Municipio Santiago de los Caballeros, Dominican Republic (Fig. 1).
Holotype: MCZ R–186380 (Fig. 12), an amber-embedded skeletonized and articulated specimen preserving patches of integument, most of the left basicranium, the posterior part of the left jaw, partial elements from the hyoid apparatus and most of the postcranial elements.
Published figures: Daza and Bauer (2012a, figures 2-7).
Remarks: Compared to all other known Sphaerodactylus in amber, this specimen is unusual for being mostly skeletonized, having dark colored bones with a few integumentary patches preserved (Daza and Bauer, 2012a). The skeleton of S. ciguapa indicates that the specimen was mature. Scales of the dorsum are granular, subimbricate, and weakly keeled to keeless. It differs from S. dommeli in having a shorter neck and relatively smaller granular scales.
Systematics: This species when compared with living taxa from Hispaniola and the Greater Puerto Rico presented similarities with members of the notatus species group of the argus series and with members of shrevei species group in the cinereous series (Daza and Bauer, 2012a).
Age: late Early Miocene to early Middle Miocene.
Locality: Cordillera Septentrional, Santiago Province, north of Municipio Santiago de los Caballeros, Dominican Republic (Fig. 1).
Referred material: Specimens mentioned in the literature are derived from privately held collections (see Remarks below).
Published figures: Kluge (1995, figures 12A–D); Grimaldi (1996, figures on pages 108, 110); Grimaldi et al. (2000, figures 1a–d, 3, 4).
Remarks: Three additional amber-embedded specimens have been briefly described (Kluge, 1995, Grimaldi 1996, 2000), but at least six more are known from two private collections in Turin and Milan, Italy (Fig. 12). One privately owned amber piece belonging to Ms. Susan Hendrickson contains a small specimen (14 mm SVL) that has been argued to have some affinities with the cinereus series (Kluge, 1995). Due to the reduced ossification of the specimen, Kluge (1995) suggested that it was probably a juvenile, the illustration of the head of this fossil has a very short snout, which is characteristic of juveniles in this genus.
The other two specimens were described by Grimaldi (1996) and Grimaldi et al. (2000). The later belongs to the private collection of Dott. Ettore Morone in Turin Italy (M–1274). This piece of amber contains a nearly complete animal with an autotomized tail (Fig. 12). The integument of this specimen is very well preserved and has the potential of providing good amount of detail about its pholidosis. The head of the specimen was CT scanned revealing a partially disarticulated skull represented by several fragments bones (Grimaldi et al., 2000). New data has been obtained from nearly all amber Sphaerodactylus, and these are currently being studied systematically (Daza et al., unpublished).
Some of these amber specimens preserve their skin or an impression of it, while the interior of the animal varies from having an almost intact skeleton in situ to mainly void with only a few bones preserved. In some cases, bones are greatly displaced as a consequence of postmortem decomposition within the amber. Amber preserved bones may sometimes appear transparent (Kluge, 1995).
Adult size in extant Sphaerodactylus geckos ranges from the smallest forms measuring about 16 mm SVL to large forms that can be 40 mm SVL (Schwartz and Henderson, 1991; Henderson and Powell, 2009). One important aspect in the anatomical study of these geckos in amber is the estimate of the skeletal maturity. Based on body proportions and, in some cases, degree of long bone ossification, some of the fossil animals seem to be juveniles (e.g., Fig 12B, H–J), although some extant species of Sphaerodactylus (e.g., S. ariasae and S. partenophion) might be have comparable SVL to these fossils.
Palaeogekko risgoviensis Schleich, 1987
Classification: Gekkonoidea exclusive of Eublepharidae.
Remarks: Schleich (1987) estimated this species to have a total body length of 5.8 cm. The mandible has no angular and the dentary is almost 1 cm long, with between 21 and 37 tooth loci. In bivariate plots, Paleogekko is differentiated from sampled genera in having a higher density index of tooth loci versus overall dentary length (Schleich, 1987). The number of labial (mental) foramina also varies considerable across specimens, between 4 and 7. The maxilla measures 7 mm and potentially bears between 23 and 29 tooth loci. On the toothed part of the maxilla (par dentalis), there is one foramen for the superior alveolar canal for the maxillary branch of cranial nerve V. The premaxilla bears between 9 and 10 tooth loci and has a long ascending nasal process that bears a flattened keel ventrally (septonasal crest in Daza et al., 2008).
Systematics: Schleich (1987) compared this fossil with 14 extant species from the families Eublepharidae, Phyllodactylidae, and Gekkonidae. He concluded that this species resembled the extant genera Tarentola (Phyllodactylidae) and Mediodactylus (Gekkonidae) in terms of the size of bones, number and location of labial foramina, and number of tooth loci. In addition to the characters used by Schleich, we reviewed the presence of nine additional morphological features of the fossil in some more extant genera: (1) shallow dentary, (2) dentary not extending beyond coronoid eminence, (3) coronoid eminence moderately elevated and not compressed, (4) discrete splenial, surangular foramen not bounded by coronoid, (5) posterior border of the osseus nares steeply sloped, (6) width and height of the facial process of maxilla and anterior border of the facial process of the maxilla straight, (7) dorsal terminus of the same process ending in a posteriorly directed point, (8) ascending nasal process of the premaxilla narrow and elongated, and (9) the same process ending in a blunt terminus. From our comparisons we agree with Schleich in that P. risgoviensis is not related to the Eublepharidae or the Pygopodoidea. Unfortunately, all the characters used for intergeneric comparison are very general and the combination of these features appears in several other genera, for example, in addition to Tarentola and Mediodactylus, these are shared with Gymnodactylus, Goggia, Pachydactylus, and Phelsuma. Another factor that cannot be discounted is that the material assigned to P. risgoviensis might belong to several species, one argument in favor of this idea is that the meristic and morphometric data derived from the numerous specimens is highly heterogeneous (see tables and graphs in Schleich, 1987). An additional revision of this material would help to resolve this issue.
Referred specimens: Two specimens are known from this locality. One large one represented by a frontal (BML 915), fragments of the maxilla, a dentary (BML 914), and vertebrae. A dentary and vertebrae represent the smallest form.
Remarks: Based on comparison of the largest specimen with a limited sample of gekkotans, this material was assigned to Tarentola (Rage, 1976). In the published figures of the dentary and the frontal from the large specimen, it can be seen that this fossil has strongly recurved teeth that are spaced similar to those of Tarentola. T. mauritanica has about nine tooth loci above the dentary notch for the splenial, whereas the Moroccan fossil has three to four and a total of about 20 tooth loci. In this regard it more resembles Tarentola gigas (NHMUK 1906.3.30.31), which has 5 and 22 loci, respectively. A revision of the material is necessary to corroborate the placement of this fossil within Tarentola.
Unnamed material Augé and Rage, 2000
Classification: Undetermined Gekkota
Age: Middle Miocene (Lower Astaracium, MN6).
Locality: SanSan, Gers department, France (Fig. 1).
Referred specimens: 17 maxillae and dentaries (MNHN Sa 23577, MNHN Sa 23580), 34 dorsal vertebrae (MNHN Sa 23578, MNHN Sa 23581), and 4 caudal vertebrae (MNHN Sa 23579).
Remarks: This material was originally described as Lacerta ambigua (Lartet, 1851) but this name was subsequently considered a nomen nudum (Augé and Rage, 2000) and the material redetermined as being gekkonid (sensu lato). The material is considered unambiguously gekkotan, but too fragmentary to be assigned to any of the known families (Augé and Rage, 2000).
Unnamed material Estes, 1963
Classification: Incertae sedis.
Locality: Thomas Farm, Gilchrist County, Florida, USA.
Referred specimen: MCZ 3382, a right dentary.
Remarks: A short dentary attributed to the Gekkonidae sensu lato based on the simple pointed teeth, with striated tips, and a labial facet for the coronoid (Estes, 1963). The placement in the Gekkonidae was confirmed by Kluge (1967a) but without referring this material to any particular clade. Estes (1963) considered this specimen to be a gecko with certainty based on the numerous small, slightly irregular conical teeth, and the deeply incised notch for the coronoid. He also discussed the presence of striations in this fossil and modern forms. The toothed area of this dentary is nearly complete, but in fact the number of tooth loci (only 16) is low for a gekkotan. Such a low number of teeth among extant gekkotans is only known among pygopods (e.g., 16–17 in Pletholax gracilis, Pygopus nigriceps) which have differently shaped dentaries. Unfortunately, the Meckelian wall in the dentary is broken on the specimen, but based on the low number of teeth, we think that this specimen might not be referable to any extant gekkotan genus.
CONTEXT AND CONCLUSIONS
Gecko skulls are small, lightly built, and paedomorpohic, and hence rarely fossilized (Evans, 2003). However, preservation frequencies of individual bones have been remarkably consistent, from subfossil remains back to the early Paleogene. Maxillae and dentaries are the most common elements among all known disarticulated fossil bones from the Early Cretaceous through the Tertiary (Table 1). These same elements are those recovered most frequently from the stomachs of some groups of modern gecko-feeding birds (Kupriyanov et al., 2012), suggesting that they are robust to digestive as well as taphonomic processes. Rage (1978) previously described these tooth bearing bones as the most common material among fossil gecko material, but frontals are also frequently found, which can be explained by their fusion dorsally and ventrally into a tubular structure. In contrast, postcranial material of geckos is relatively rare in the fossil record. Although it is possible that such elements do not preserve as well as some of the skull bones, they are also more difficult to identify, especially if fragmentary, and their paucity may be artifactual. The vertebrae of geckos are relatively distinctive and can often be distinguished from those of other lizards but few characters have been identified to distinguish gekkotan appendicular elements, although some are being developed (Daza and Bauer, unpublished data). Another factor that determines the lack of gekkotan fossils is the small size of these lizards (Bauer and Russell, 1991), although the truly small forms are occasionally preserved in amber inclusions, which account for about one fourth of the total pre-Quaternary fossils known.
Currently the pre-Quaternary fossil record of the Gekkota comprises material of 17 described species. The modern genera Pygopus, Sphaerodactylus, and Euleptes can be traced back to the Miocene; fossils assigned to these groups present sufficient diagnostic morphological characters that allow them to be allocated with confidence to the extant clades Sphaerodactylidae and Pygopodidae (Estes, 1969; Böhme, 1984; Hutchinson, 1997; Müller, 2001; Müller and Mödden, 2001; Čerňanský and Bauer, 2010; Daza and Bauer, 2012a).
The large amount of anatomical information accumulated in the last 40 years, including new comparative material and HRXCT data have helped to overcome previous limitations on the classification of fossil gekkotans (Kluge, 1967a). One example of this is how the new cranial information about Euleptes europaea has helped to better understand the anatomy of the Miocene genus Gerandogekko. The similarities between these two genera indicate that these are highly likely to be closely related. In addition, the material attributed to Euleptes indicates that this genus was more morphologically diverse during the Miocene, being represented by two (or possible three) larger forms that reached areas of central Europe.
The species Palaeogekko risgoviensis is probably the only representative of the Gekkonidae sensu stricto known from the Miocene, although inclusion in the Sphaerodactylidae or Phyllodactylidae cannot be rejected. A revision of the material assigned to this species is required; its great morphological heterogeneity indicates that this taxon might include material assignable to other extant genera.
The affinities of the Paleogene gekkotans are less clear, and the small number of fossils from this epoch can be allocated with less confidence to the modern clades. The poorly represented Rhodanogekko vireti might represent a genus closely related to Pristurus, but the possibility that this form belongs to another extinct lineage of large geckos with ornamented cranial bones cannot be discarded. Laonogekko and Cadurcogekko present several similarities with members of the Pygopodoidea, particularly with the Diplodactylidae. This family of gekkotans might be a good representative of the ancestral bauplan for the Gekkota. The tentative allocation of these two genera to the Pygopodoidea is certainly unexpected, since these would be the first records of this clade outside Australasia. This might indicate that the group was more widespread in the past or, alternatively, that these two forms represent an independent clade that, together with diplodactylids, preserved plesiomorphic features, but that became extinct in Europe before the Miocene. The enigmatic Yantarogekko balticus is very likely to be a non-eublepharid gekkonoid, but its osteology needs to be studied in detail. It has the potential to be a “Rosetta Stone” for better understanding gekkotan history in the Paleogene of Europe.
The oldest gecko (Cretaceogekko) needs to be revised and compared with extant gekkotans with similar toe-pads. New material tentatively referable to this taxon may yield more characters of phylogenetic relevance (Bauer and Daza, unpublished). Mesozoic skeletal fossils of Hoburogekko and Gobekko are clearly gekkotan and offer two alternative combinations of characters toward the consolidation of the gekkotan skull: (1) frontal with subolfactory processes arched beneath the brain but do not contact on midline + posteriorly fused Meckelian canal, and (2) frontal with subolfactory processes in contact on midline + posteriorly unfused Meckelian canal. Embryological data favor the first combination of characters as the ancestral condition for modern gekkotans (Fig. 13).
The paleoclimatic and paleoenvironmental context for fossil gekkotans is potentially informative about their evolution, diversification, and ecomorphology. For example, paleoenvironments for the Cretaceous Hoburogekko and Gobekko are unambiguous yet interpretations of lifestyle are open. The Aptian-Albian habitat of Hoburogekko was mesic, with prominent lakes and rivers (Jerzykiewicz and Russell, 1991). Its skull, however, resembles that of extant geckos of more xeric habitats, which burrow or otherwise retreat into the soil. Gobekko is known from eolian deposits of the Campanian Djadokhta Formation, a primarily xeric yet seasonally damp environment (Jerzykiewicz and Russell, 1991; Jerzykiewicz et al., 1993; Montanari et al., 2013). However, there are stratigraphic ambiguities associated with Djadokhta and nearby Nemegt Formation localities (P. Currie, personal communication). The Nemegt represents an analog of the Okovango Delta and Kalahari (Jerzykiewicz, 1998), with a mesic environment of abundant channels giving way to desert regions.
In contrast with these Central Asian Cretaceous taxa, Paleogene geckos of the Paris Basin and other European sites generally inhabited a wet tropical climate. Laonogekko, Yantarogekko, and the unnamed Condé-en Brie dentary date from the early Eocene, when temperatures were warm enough for varanids, agamids, boids, and tropical plants to flourish in the Arctic (Eberle and Greenwood, 2012; Augé et al., 2012). The local environments were not necessarily uniform, however. Laonogekko is known from a strikingly diverse herpetofaunal assemblage in a warm, well-watered environment, including salamadrids, trionychid turtles, and semiaquatic alligatorids and crocodylids (Augé et al., 1997). In contrast, stellate trichomes preserved in amber with Yantarogecko occur in drought-resistant scrub oaks (Nixon, 2002). These trichomes suggest a drier, more open environment for Yantarogecko, reminiscent of some current Euleptes habitats.
Warm temperatures predominated through the mid-Eocene and even lower Oligocene environments occupied by Rhodanogekko and Cadurcogekko. Rhodanogekko experienced mean annual temperatures of 17–27°, with summer temperatures of 30°, based on stable isotopes of near-shore bivalves and gastropods in the Lutetian (Huyghe et al., 2012). Cadurcogekko is known from the Phosphorites de Quercy, which record drastic shifts from rainforest to savannah ecosystems, and the Grande Coupure turnover of mammamlian communities (Legendre, 1988). Cadurcogekko rugosus is clearly upper Eocene (Augé, 2005) and most likely inhabited a rainforest biome, but C. piveteaui may be lower Oligocene (Hoffstetter, 1946; Augé, 2005) and hence from a drier landscape. If the latter is true, the Cadurcogekko lineage may have had broad environmental tolerance. Alternately, its presence in the two localities may indicate proximate stratigraphic horizons and/or environmental continuity.
Our taxonomic assignments become more confident for Neogene geckos, and extant taxa can strongly inform our interpretations of paleoecology. Similarities of Miocene Gerandogekko and fossil Euleptes suggest the initial radiation of nocturnal leaf-toed sphaerodactylids in Central Europe (Müller and Mödden, 2001) in Mediterranean climates like those experienced by extant Euleptes (Schäfer, 1984) or perhaps a more humid, near tropical climate (Haller-Probst, 1997; Böhme, 2003). Similarly, amber-preserved specimens of Miocene Sphaerodactylus from Hispaniola (Böhme, 1984; Daza and Bauer, 2012) are consistent with radiation into Neotropical ecosystems similar to those today.
We thank Didier Berthet, Jean-Claude Rage, Günter Bechly, and Magdalena Borsuk-Białynicka for access to specimens and assistance. We thank Vladimir Alifanov and Andrej Čerňanský for pictures and comparative material sources. Edward Stanley, Jessie Maisano, Matthew Colbert, Jonathan Losos, and Emma Sherratt for HRXCT data from fossil and comparative specimens. Christopher J. Bell provided key literature. The manuscript benefitted from the comments of Jack L. Conrad, Andrej Čerňanský, and Scott Miller.