Open Access
How to translate text using browser tools
3 June 2008 Phylogeny And Systematics Of Squamata (Reptilia) Based On Morphology
Jack L. Conrad
Author Affiliations +
Abstract

Squamata (amphisbaenians, “lizards”, mosasaurs, and snakes) is an extremely diverse clade with a rich fossil record. There is little consensus about the interrelationships of the major squamate clades (i.e., Iguania, Gekkota, Scincomorpha, Anguimorpha, Amphisbaenia, and Serpentes), or even the membership of some of these clades. Morphology-based cladistic analyses typically agree only that the major dichotomy in extant squamates is between Iguania and all other taxa. The phylogenetic placement of Amphisbaenia and Serpentes is particularly problematic. Incomplete taxon sampling is likely a major contributing factor to the absence of a consensus about squamate interrelationships. This study examines squamate relationships using 222 ingroup taxa scored for 363 morphological characters. Analysis of these data recovered 2,213 equally short trees with a length of 3,273 steps and a retention index of 0.7164. The results confirm the monophyly of the clades Scleroglossa (extant squamates exclusive of Iguania), Gekkota, Scincomorpha, Lacertoidea, Scincoidea, Anguimorpha, Carusioidea, Platynota, and Varanoidea. Novel results include the identification of a clade containing Scincidae sensu lato, Dibamidae, Amphisbaenia, and Serpentes; identification of a Mesozoic clade containing Bainguis, Eoxanta lacertifrons, Globaura venusta, and Myrmecodaptria; and identification of Dalinghosaurus as a basal shinisaur. A new taxonomic scheme is outlined. The names Iguanomorpha, Scincogekkonomorpha, Evansauria, and Mosasauriformes are applied to the stem-based groups including Iguania, Scleroglossa, Autarchoglossa, and Mosasauria, respectively. The importance of strict rigidity within taxonomy is questioned; taxonomy is most useful as a tool for communication about organisms or groups of organisms.

Introduction

Subject Matter and Goals

Squamata (amphisbaenians, “lizards”, mosasaurs, and snakes) represents a morphologically and ecologically diverse clade with a rich fossil record. The smallest known squamates are no more than 18 mm in snout-to-vent length (Hedges and Thomas, 2001) and the largest known fossil form probably exceeded 17 m in total length (Lingham-Soliar, 1995). Squamates are very speciose in extant faunas with approximately 8,000 species spreading to every country except Iceland (Bauer, 2003; Gans, 2003; Shine, 2003; Uetz, 2007). Moreover, the diversity of form throughout the last 160 million years of known squamate history rivals that of mammals. Different squamates have become adapted for fossoriality, terrestriality, arboreality, and for the near-shore, open-water, and reef marine environments (see Lingham-Soliar, 1992b, 1995, 1999b; Caldwell, 1996, 2000; Caldwell and Lee, 1997, 2004; Caldwell and Cooper, 1999; Lee et al., 1999b; Kley, 2000, 2001, 2006; Caldwell and Albino, 2001; Lee and Scanlon, 2002b; Voris and Murphy, 2002; Bauer, 2003; Gans, 2003; Kearney, 2003a; Shine, 2003; Alfaro et al., 2004; Kearney et al., 2005; Uetz, 2007). Such diversity in morphology, biogeography, and ecology paired with a relatively deep fossil record and numerous extant taxa makes Squamata an extremely attractive group for evolutionary studies. Unfortunately, there is currently no clear picture of squamate phylogenetic relationships. This problem is due, in part, to incomplete sampling of squamates in recent phylogenetic analyses. Incomplete morphological documentation of some problematic taxa is also a major problem. A few landmark works have looked at squamates in fantastic detail (e.g., Parker, 1878, 1879; Bellairs, 1949; Oelrich, 1956; Jollie, 1960) and recent application of high-resolution x-ray tomography has helped to demystify problematic taxa and otherwise hidden morphology (see the recent work on the braincase of Shinisaurus crocodilurus by Bever et al., 2005a and of the cranial anatomy of Rhineura hatcheri by Kearney et al., 2005, for example).

Many extant and fossil taxa are problematic for understanding squamate evolution because they are poorly known morphologically, because they possess no fossil record, and/or because they possess a combination of character states making their referral difficult. Dibamidae, a clade of bizarre fossorial squamates, is one example of such a taxon. That is, Dibamidae is a well-described extant clade whose affinities are in nearly constant flux.

This paper presents a phylogenetic analysis of the squamate groups and offers a new taxonomic scheme. Although future discoveries and incorporation of new evidence will offer variations in parts of the tree, the current study is based on the most extensive morphological data matrix currently available. Broad taxonomic sampling in this analysis is one way to analyze the phylogenetic positions of some particularly problematic taxa (e.g., necrosaurs, snakes, amphisbaenians); that is, to avoid constraining those problematic taxa within or outside of their historical placements.

Taxonomy is a tool for communication about groups of things (organisms, in this case). To this end, I find that taxonomy is most effective when names remain meaningful between phylogenetic hypotheses, even if there are some differences in group membership. Empirical studies always offer the possibility of changes (radical or trivial) in broader scale relationships, and a given taxonomic scheme remains useful only if it is flexible enough to accommodate these changes while maintaining continuity of meaning (see Rieppel, 2005, 2006). Utility is a worthy prize for which rigid structure may be sacrificed. This is especially true in something so subjective (and semantic) as taxonomy. Having said that, I do employ taxon name definitions below with the hope that they will be employed—so long as the terminology remains useful. Thus, the taxonomic scheme accompanying my squamate phylogenetic hypothesis is designed to be useful even in the face of some topological changes to the tree. It is constructed to be useful for neontologists (e.g., herpetologists) and paleontologists alike, and to remain relatively consistent with the current taxonomic usage of both those groups of scientists.

My three overarching aims for this study are:

  1. Produce a morphological phylogenetic data matrix for the major squamate groups, more exclusive groups with problematic history, and numerous fossils of debated history.

  2. Produce a phylogenetic hypothesis of squamate relationships based on this morphological data matrix.

  3. Offer a revised taxonomy taking into consideration previous studies and problematic areas in the current topology, and stabilize existing names that lack clear or precise definitions.

Historical Analyses

History provides a useful frame on which to rest new ideas. Because Squamata is a conspicuous extant clade, it has been the focus of numerous phylogenetic studies and taxonomic treatments over the past 13 decades or so. Some of the more inclusive and influential ones are reviewed here.

Precladistic Studies

Wallace (1876a, 1876b) provides an early comprehensive view of “family”-level squamate systematics, identifying 27 lizard “families” (his Lacertilia) and 24 snake “families”. Importantly, many of these groups are retained in modern systematic treatments, although often with some modifications of membership.

Squamate systematics were extensively analyzed by two papers appearing just at the turn of the last century (Cope, 1900; Fürbringer, 1900a), but the real landmark paper is that of Camp (1923). Using extant taxa and numerous fossil groups, Camp constructed a branching diagram (a “skiogram”) (Camp, 1923: 333) that has been compared to a cladogram (Moody, 1985), identifying several groups that are still supported in modern analyses (fig. 1). That diagram illustrates iguanians branching off from the rest of squamates first, followed by a dichotomy between Gekkota and Autarchoglossa (containing Scincomorpha and Anguimorpha) (Camp, 1923). Romer (1956, 1966), Estes (1983), and Carroll (1988a, 1988b suggested three similar taxonomic schemes for squamates containing several “infraorders”. Carroll's (1988a, 1988ba) incarnation of this scheme is seen in table 1. This system largely corresponds to that in the popular literature (e.g., Whitfield, 1982), and web resources such as the TIGR Reptile Database (Uetz, 2007) and Animal Diversity Web (Myers, 2001). Herpetologists usually employ similar schemes (see, for example, Behler and King, 1979, Kent and Miller, 1997, and Pough et al., 2005).

Figure 1

Camp's “skiogram”; the evolutionary history of squamates as envisioned by Camp (1923), often heralded as the first approximation of a cladogram. Some higher level taxon names (e.g., Rhiptoglossa, Xantusioidea) have been omitted. Note that much of the taxonomy employed here is still used in modern phylogenetic taxonomy. Fossil taxa denoted by daggers (†).

i0003-0090-310-1-1-f01.gif

Table 1

The Taxonomic Scheme of Carroll (1988a)

i0003-0090-310-1-1-t01.gif

Estes et al., 1988

Camp's (1923) systematic scheme as modified by later authors (Romer, 1949, 1956; Carroll, 1988a, 1988ba) remained the mainstay of lizard interrelationships until the widespread use of cladistics over the last two decades. The new standard for squamate relationships is that of Estes et al. (1988). Like Camp (1923), this analysis identified Gekkota as the sister-taxon to the Autarchoglossa, with Iguania representing the most basal extant squamate lineage. The most often cited diagram from Estes et al. (1988: fig. 6) (fig. 2A) was not a consensus cladogram of relationships, but rather, as Estes et al. state, a preferred “conservative” hypothesis of interrelationships (Estes et al., 1988:140). The Estes et al. (1988) data set actually supports a somewhat different phylogenetic topology in which an amphisbaenian-dibamid-snake clade is the sister group to Scleroglossa (fig. 2B).

Figure 2

Hypotheses of squamate interrelationships, based on morphology, as presented by Estes et al. (1988) and separately derived from their data. The cladogram on the left represents the “conservative cladogram of squamate relationships” as reported by Estes et al. (1988: 140, fig. 6). The right side shows the relationships recovered when the Estes et al. (1988) data matrix is run in PAUP* (Swofford, 2001) using a heuristic search (parsimony) and TNT (Goloboff et al., 2003) using the traditional search.

i0003-0090-310-1-1-f02.gif

Estes et al. (1988) did not use fossil taxa as part of the ingroup in their analysis. Despite this, Estes et al. (1988) remains an extremely important study used as the basis of recent morphology-based cladistic analyses, including this one.

Eight additional morphology-based analyses (Wu et al., 1996; Evans and Barbadillo, 1998, 1999; Lee, 1998; Caldwell, 1999a; Lee and Caldwell, 2000; Evans et al., 2005; Evans and Wang, 2005) that have addressed a broad range of squamate taxa and included significant fossil data are reviewed here and used for comparisons below. All of these analyses have drawn from the character list presented in Estes et al. (1988). Lee (1998) and Lee and Caldwell (2000) are essentially one analysis with the latter study including a few additional fossil taxa. Evans and Barbadillo (1998, 1999) are literally the same analysis with similar findings, except that the Evans and Barbadillo (1999) analysis included the fossil taxon Hoyalacerta sanzi. Evans et al. (2005) and Evans and Wang (2005) updated this analysis with broader taxon sampling and recovered a different hypothesis. Finally, I will also review two recent molecular analyses (Townsend et al., 2004; Vidal and Hedges, 2005) that turn the morphological trees on their collective (figurative) heads.

Wu et al. (1996) citing problems with unscoreable characters (“missing data” of their usage, see Kearney and Clark, 2003), modified the analysis of Estes et al. (1988) by excluding non-osteological characters, adding 22 new characters, and including the fossil taxa Adamisaurus magnidentatus, Eoxanta lacertifrons, Globaura venusta, Macrocephalosauridae ( = Gilmoreteiidae; Langer, 1998), Polyglyphanodontidae, Sineoamphisbaena hexatabularis, and Slavoia darevskii. Their most inclusive analysis (for taxa and characters) yielded 28 shortest trees. They found Sineoamphisbaena hexatabularis to be the sister-group to their terminal group “other amphisbaenians” and suggested placement of the Xantusiidae with Scincoidea (sensu Estes et al., 1988) (fig. 3).

Figure 3

Hypothesis of squamate interrelationships, based on morphology, as presented by Wu et al. (1996). Fossil taxa denoted by daggers (†).

i0003-0090-310-1-1-f03.gif

Evans and Barbadillo (1998) included the major groups used by Estes et al. (1988), but added the fossil taxa Ardeosaurus (unspecified species inclusion), Bavarisaurus macrodactylus ( = Homoesaurus macrodactylus of Wagner, 1852), Eichstaettisaurus schroederi, Meyasaurus diazromerali, Paramacellodus (presumably including taxa from several localities, including those described in Prothero and Estes, 1980; Broschinski and Sigogneau-Russell, 1996; Evans and Barbadillo, 1998; Evans and Chure, 1998a; Averianov and Skutchas, 1999), and Scandensia ciervensis to the analysis. Additionally, they not only included Rhynchocephalia, Kuehneosauridae, and a “paliguanid” (Saurosternon) as outgroups, but also the relatively recently described Marmoretta (these outgroups omitted from fig. 4). Characters from this analysis were taken from both Estes et al. (1988) and Gauthier et al. (1988a) with one novel character included, though many of the character states were reported as “parsimony uninformative” (Evans and Barbadillo, 1998). The analysis recovered six most parsimonious trees. A strict consensus leaves only the relative positions of Eichstaettisaurus and Scandensia ciervensis unresolved with respect to more nested squamates (fig. 4). The results showed that Ardeosaurus, Bavarisaurus, and Eichstaettisaurus constitute basal members of Squamata. Additionally, Evans and Barbadillo (1998) retrieved the novel position. A contemporary study (Lee, 1998; see below) found dibamids and amphisbaenians to be close to Gekkota, and snakes as falling within Anguimorpha. A subsequent analysis (Evans and Barbadillo, 1999) included the taxon Hoyalacerta sanzi and resulted in three shortest trees. The variation between these trees occurred only between the relative placements of Hoyalacerta sanzi and Eichstaettisaurus with respect to each other, Ardeosaurus, and Iguania (fig. 5). The rest of the tree was consistent with the earlier study (Evans and Barbadillo, 1998).

Figure 4

Hypothesis of squamate interrelationships, based on morphology, as presented by Evans and Barbadillo (1998). Fossil taxa denoted by daggers (†).

i0003-0090-310-1-1-f04.gif

Figure 5

Hypothesis of squamate interrelationships, based on morphology, as presented by Evans and Barbadillo (1999). Note that this is essentially the same tree as in Figure 4, but with the addition of Hoyalacerta sanzi and the collapse of several nodes at the base of squamates. Fossil taxa denoted by daggers (†).

i0003-0090-310-1-1-f05.gif

Lee (1998) and Lee and Caldwell (2000) analyzed very similar data sets with the latter subdividing some of the taxa included in the former (Xenosauridae, Mosasauroidea, Gekkonidae, and Agamidae) and adding four additional fossil taxa (Adriosaurus, Aphanizocnemus, Dolichosaurus longicollis, and Pachyophis woodwardi). Both character lists draw heavily from Estes et al. (1988), but with additions and modifications. Unlike Evans and Barbadillo (1998), these analyses include numerous limbless terminal taxa and aquatic fossil forms. The Lee (1998) analysis found two most parsimonious trees with the only unresolved node occurring between Scincidae, Cordyliformes (Cordylidae of his usage), and Anguimorpha (fig. 6). Lee and Caldwell (2000) produced 12 most parsimonious trees with the same polytomy as in Lee (1998), and another between Dolichosauridae, Aphanizocnemus, and a clade including Adriosaurus and snakes (fig. 7). Both analyses fail to retrieve a monophyletic Scincomorpha, place amphisbaenians, dibamids, and xantusiids near Gekkota, and hypothesize that snakes are derived from the mosasaurid-varanid clade.

Figure 6

Hypothesis of squamate interrelationships, based on morphology, as presented by Lee (1998) with some of the accompanying taxonomy, especially as it differs from that of Estes et al. (1988). Note that Serpentes exclusive of Pachyrhachis problematicus is constrained to be monophyletic. Fossil taxa denoted by daggers (†).

i0003-0090-310-1-1-f06.gif

Figure 7

Hypothesis of squamate interrelationships, based on morphology, as presented by Lee and Caldwell (2000). This analysis was similar to that of Lee (1998), but several fossil taxa were added near the base of Mosasauroidea, and Agamidae and Gekkota were further divided into their presumed constituent clades. Note that Serpentes exclusive of Pachyrhachis problematicus and Pachyophis woodwardi is constrained to be monophyletic. Fossil taxa denoted by daggers (†).

i0003-0090-310-1-1-f07.gif

Caldwell (1999a) analyzed a set of taxa similar to that of Estes et al. (1988), but added the fossil taxa Coniasaurus, Dinilysia patagonica, Estesia, and Mosasauroidea. Additionally, he divided Gekkonidae sensu Estes et al. (1988) into Eublepharinae and Gekkonoidea sensu Kluge (1987), and extant snakes into Scolecophidia and Alethinophidia to help analyze the position of Dinilysia patagonica. This analysis recovered 18 equally short trees whose strict consensus shows limited resolution (fig. 8). However, this analysis did offer support for some clades questioned by Lee (1998) and later by Lee and Caldwell (2000). Scincomorpha, Scincoidea, and Lacertoidea (all sensu Estes et al., 1988) were supported in the strict consensus. Additionally, this analysis supported a sister-taxon relationship between Dinilysia patagonica and alethinophidians, suggested the paraphyly of Xenosauridae (sensu Estes et al., 1988), and suggested that Lanthanotus borneensis, Estesia, and Varanus were successively more remote outgroups to Heloderma. The latter is significant because Lanthanotus borneensis typically has been considered to be more closely related to Varanus than to Heloderma, and because Estesia was first considered a close relative of the Varanus-Lanthanotus borneensis clade (Norell et al., 1992) and then a monstersaur (Norell and Gao, 1997; Gao and Norell, 1998, 2000; Nydam, 2000).

Figure 8

Hypothesis of squamate interrelationships, based on morphology, as presented by Caldwell (1999a). Fossil taxa denoted by daggers (†).

i0003-0090-310-1-1-f08.gif

Evans et al. (2005) have offered one of the most recent broad-scale analyses of squamate phylogeny. Similar in composition to the analyses of Evans and Barbadillo (1998, 1999), this analysis has added three taxa (Yabeinosaurus tenuis, Parviraptor, and Aigialosaurus) and some new characters. The resulting study is one of the most fossil-inclusive studies so far published and the phylogenetic hypothesis (fig. 9A) shows significant differences from the Evans and Barbadillo (1998, 1999) studies. Importantly, Gekkota is found to be a basal clade within Scleroglossa, the amphisbaenian-dibamid-snake clade is the sister-taxon to Aigialosaurus (and, presumably, other mosasauroids) within Anguimorpha (above the level of Shinisaurus and Xenosaurus), and there is an extinct clade composed of Eichstaettisaurus, Hoyalacerta sanzi, Parviraptor, and Scandensia ciervensis at the base of Anguimorpha.

Figure 9

Hypothesis of squamate interrelationships, based on morphology, as presented by Evans et al. (2005). Fossil taxa denoted by daggers (†).

i0003-0090-310-1-1-f09.gif

Evans and Wang (2005) offer another derivation from the Evans and Barbadillo (1998, 1999) matrices. This analysis includes the fossil taxa Carusia intermedia, Mosasauroidea, and Dalinghosaurus longidigitus, but not Yabeinosaurus tenuis nor Scandensia ciervensis (fig. 9B). Similar to Caldwell (1999a), this study finds mosasauroids to be basal to Scleroglossa. It also recovers a Carusioidea (sensu Gao and Norell, 1998) and suggests that D. longidigitus and a clade consisting of Eichstaettisaurus, Hoyalacerta sanzi, and Parviraptor are successively more remote outgroups to Carusioidea. A monophyletic Scincomorpha is the sister taxon to Gekkota in this analysis and is nested within Anguimorpha as the sister-group to a clade containing snakes, dibamids, amphisbaenians, and non-carusioid anguimorphs.

Molecular Analyses

Townsend et al. (2004) and Vidal and Hedges (2005) have recently offered a higher-level analysis of extant squamate relationships based on molecular evidence (fig. 10). Both of these analyses find Gekkota and Dibamus to be basal radiations of squamates. Townsend et al. (2004) suggest that Gekkota is the basalmost lineage; Vidal and Hedges (2005) suggest that Dibamus (Dibamidae of their usage; because Anelytropsis papillosus was not included, Dibamus is a more accurate taxon indicator) is more basal. Both analyses suggest that cordylids, xantusiids, and scincids form a clade that was the next to diverge. Amphisbaenians are hypothesized to be nested within Lacertiformes (sensu Estes et al., 1988) as the sister-group to a clade composed of snakes, iguanians, and anguimorphs. Although there is some question about the exact placement of snakes, each study suggests that they are close to an Iguania-Anguimorpha clade.

Figure 10

Hypotheses of squamate interrelationships, based on molecular data, as presented by (A) Townsend et al. (2004) and (B) Vidal and Hedges (2005). Note that the “leaf” taxa of each tree have been modified somewhat for this analysis from those presented in the original study. In (A), some clades have been collapsed and the “leaves” are labeled as the most exclusive taxon name for which taxa are sampled (e.g., the clade containing gekkonids, pygopodids, and eublepharids are collapsed into a clade termed “Gekkota” here). The tree in (B) has been modified in a slightly different way. The “leaves” of the tree as presented in the original study (Vidal and Hedges, 2005) usually suggested a broader taxon than was represented by their data. Consequently, the appropriate taxon names are put on the tree for those taxa represented by the study (e.g., “Scincidae” is replaced here by Plestiodon sensu Smith, 2005).

i0003-0090-310-1-1-f10.gif

The basal position of gekkotans and dibamids in these analyses is intriguing in part because of parallels with some historical discussions of the plesiomorphic squamate form. Earlier, noncladistic, discussions of dibamid and gekkotan morphology often characterized them as a puzzling combination of plesiomorphic and apomorphic character states that might be close to the plesiomorphic squamate form (see Estes, 1983; Kluge, 1983, 1987; Rieppel, 1984b; Greer, 1985). This has been influenced, to some degree, by the late fusion (or absence of fusion) of some braincase elements in dibamids and the persistence of notochordal vertebrae in some gekkotans.

Comparisons

Although all of the described morphology-based analyses agree on points such as the monophyly of Gekkota, Varanoidea, and Scleroglossa, there remains virtually no consensus about higher-level relationships. Xantusiidae, Cordyliformes, Scincidae, Dibamidae, Amphisbaenia, Serpentes, Mosasauroidea, and Xenosauridae are extremely problematic. This, despite a common dependence on the Estes et al. (1988) character list. Differing taxonomic inclusions are likely a major contributing factor to the absence of consensus among these phylogenetic hypotheses. Taxonomic sampling in these studies is seemingly dependent upon the specific problem the authors are addressing, probably because no more inclusive data matrices of squamates (including fossils) exist. Wu et al. (1996) were interested in the cladistic position of Sineoamphisbaena hexatabularis as it relates to Amphisbaenia and the Gilmoreteius-type scincomorphs it resembles, so they included teiioid fossils. Evans and Barbadillo (1998, 1999) were examining basal squamates and taxa previously believed to be related with scincomorphs and/or “stem” Gekkota. Accordingly, they included “bavarisaurid” and “ardeosaurid” taxa and Scandensia ciervensis, Hoyalacerta sanzi, and Paramacellodus. Lee (1998), Caldwell (1999a), and Lee and Caldwell (2000) were primarily concerned with the relative position of specific limbless taxa. Thus, Lee (1998) included Mosasauroidea, Pachyrhachis problematicus, and Sineoamphisbaena hexatabularis. Lee and Caldwell (2000) included these taxa and added Adriosaurus, Aphanizocnemus, and Dolichosauridae such that their analyses were sensitive to testing the position of those particular taxa within Anguimorpha. Caldwell (1999a) focused somewhat on snakes and mosasauroids, but helped to balance these taxonomic selections by including Estesia. Evans et al. (2005) and Evans and Wang (2005) both included numerous fossil taxa, but were still closely examining the positions of specific taxa. Comparable analyses not observed in detail here include those also testing the specific placement of new fossil taxa (e.g. Nydam, 2000; Reynoso and Callison, 2000).

There is very little overlap of fossil taxa included in the analyses described above. Sineoamphisbaena hexatabularis was included in three analyses (Wu et al., 1996; Lee, 1998; Lee and Caldwell, 2000) and Mosasauroidea was included in three (Lee, 1998; Caldwell, 1999a; Lee and Caldwell, 2000), but no other taxon appeared in more than two of these six studies. Adamisaurus magnidentatus, Adriosaurus, Aphanizocnemus, Coniasaurus, Dinilysia patagonica, Dolichosauridae, Eoxanta lacertifrons, Estesia, Hoyalacerta sanzi, Paramacellodus, Parviraptor, Meyasaurus diazromerali, Scandensia ciervensis, Slavoia darevskii, and Yabeinosaurus each appeared in only one analysis. Of course, the molecular studies included no fossil taxa at all; this is significant given the findings of Gauthier et al. (1988b) (see below). Different taxonomic sampling in these analyses is equivalent to asking different phylogenetic questions, so different answers (in the form of phylogenetic hypotheses) should be expected. The question then becomes: Are we asking the appropriate questions to find the information we desire? The answer is “yes” when the goal is to place specific fossil taxa within a subset of Squamata. However, to more fully test the relative positions of any group of taxa, more inclusive sampling is necessary.

Broadly Sampling Fossils

Gauthier et al. (1988b) clearly demonstrate that fossils are important for inclusion in a phylogenetic analysis because they affect character polarities throughout the tree. They posit that intermediate forms offer new information regarding character state changes throughout a tree (Gauthier et al., 1988b). They further outline the difficulties with using only extant taxa as outgroups in an analysis; the possibility of polymorphism and the derived condition of many taxa (Gauthier et al., 1988b). Both pose problems, but the latter is especially important because it is easy to think of some extant animals as primitive, even though they are not. The duck-billed platypus (Ornithorhynchus anatinus) represents an ancient and basal lineage of Mammalia. However, nobody actually considers the platypus “primitive”; this extremely derived mammal lacks teeth as an adult, possesses a bizarrely specialized snout and venomous spurs (Manger and Pettigrew, 1995; Attenborough, 2002; Dawkins, 2004) among other features that certainly were not present in truly primitive mammals (Desui, 1991; Luo et al., 2002; Rich et al., 2005).

Further analysis led Gauthier et al. (1988b) to suggest that some taxa were important and others unimportant for reconstructing character polarities within their data matrix. That is, inclusion of some taxa has more effect on tree topology than others and some taxa may be excluded with no extended effects. This seems to imply that only some fossil taxa need to be included in a given analysis, but that is not the intention of Gauthier et al. (1988b). Indeed, the point is that such analyses are context sensitive, meaning that different taxonomic samplings have the potential to produce different results. Additional data (as from fossils) can render the “unimportant” taxa “important.” Gauthier et al. (1988b) show that their so-called “unimportant” taxa could only be identified a posteri, underscoring the necessity of increased taxon sampling.

That increasing taxon sampling is beneficial was recently challenged by an evolutionary model using DNA evidence (Rosenberg and Kumar, 2001). However, subsequent analyses reveal errors in the original interpretations of this model and provide further confirmation for the importance of broad taxonomic sampling (Pollock et al., 2002; Zwickl and Hillis, 2002).

Two major conclusions may be drawn from these studies:

  1. Fossils are important for inclusion in phylogenetic analyses

  2. As many taxa as possible should be included.

When viewed in conjunction, these first two conclusions support a third:

  1. As many fossil taxa as possible should be included in any given phylogenetic analysis.

Even if fossil taxa include a large number of unknown or unscoreable character states, their inclusion may be useful. Incompletely known taxa with unexpected character state combinations are sometimes omitted from analyses because they may reduce resolution in a strict consensus (Nixon and Wheeler, 1992). However, this type of taxonomic deletion is problematic because taxa with few scoreable characters still offer important data for reconstructing character polarities (Kearney and Clark, 2003; Wiens, 2003), although incompletely known taxa whose character scoring completely overlap with another taxon in the matrix, may be safely removed from an analysis (Wilkinson, 1995; Kearney and Clark, 2003). Similarly, soft tissue (or molecular) characters remain important in analyses including large numbers of fossils if taxa and characters are well sampled overall (Kearney and Clark, 2003; Wiens, 2003).

Given the context-sensitive nature of phylogenetic analyses and that the cladistic positions of included taxa are somewhat interdependent, the published phylogenetic analyses of squamates described above have obvious conflicts with one another. Each of those studies includes only a few fossil taxa and the inclusion of many more would be expected to retrieve a different conceivably more accurate picture of the interrelationships (see Pollock et al., 2002; Zwickl and Hillis, 2002).

That being said, the current study does not examine all of the described fossil squamates that are diagnosable to species. Taxa of unquestioned affinity or that are nested members of well supported clades have been omitted in some cases, but may be included in future versions of this data matrix by myself or others.

Materials and Methods

Breadth of the Analysis

This study was undertaken with the goal of analyzing the phylogenetic relationships and interrelationships of major squamate groups with more focus given to problematic groups. The historical analyses described above have helped shape this analysis. Iguanians are a problematic group whose ingroup relationships remain uncertain and without consensus (Etheridge and de Queiroz, 1988; Frost and Etheridge, 1989; Macey et al., 1997; Frost et al., 2001; Schulte et al., 2003; Conrad and Norell, 2007a). As described above, the structure (or monophyly) of various scleroglossan groups have not reached consensus. Snake origins are an important problem for understanding almost any area of squamate phylogeny. Indeed, even if a researcher were interested only in anguimorph phylogenetics, she or he would be forced to examine all of Squamata based on the problem of snake origins in the context of recent phylogenetic analyses.

Since McDowell and Bogert's (1954) landmark study, the conventional wisdom has been that snakes are derived anguimorphs. Importantly, the only broad cladistic analysis of terrestrial anguimorphs (including fossils) have not included mosasaurs or snakes (e.g., Gao and Norell, 1998; fig. 11) and those analyses including mosasaurs and snakes include few, if any, terrestrial fossil forms (Lee, 1997, 1998, 2000; Caldwell, 1999a, 2000; Lee and Caldwell, 2000; Rieppel and Zaher, 2000a) (e.g., see figs. 6Figure 78). Numerous recent papers provide data suggesting that snakes are derived varanoids (Forstner et al., 1995; Lee, 1997, 1998; Lee et al., 1999a; Caldwell, 2000; Lee, 2000; Lee and Caldwell, 2000; Lee and Scanlon, 2001; Scanlon and Lee, 2002; Caldwell and Dal Sasso, 2004), but others have re-examined the issue and found that the characters or character codings supporting this hypothesis are problematic (Zaher and Rieppel, 1999a, 1999b, 2002; Rieppel and Zaher, 2000a, 2000b, 2001; Rieppel et al., 2003). Caldwell (1999a) finds snakes and mosasauroids to fall outside of Anguimorpha (fig. 8), and some studies place snakes with Amphisbaenia, Dibamidae, and Gekkota within Anguimorpha (e.g., Evans and Barbadillo, 1998, Rieppel and Zaher, 2000a).

Figure 11

Hypotheses of anguimorph interrelationships, based on morphology, as presented by Gao and Norell (1998). The cladogram on the left (A) is a more inclusive analysis; (B) is the analysis after the removal of Bainguis parva, Eosaniwa koehni, Palaeosaniwa canadensis, and Restes rugosus. Fossil taxa denoted by daggers (†).

i0003-0090-310-1-1-f11.gif

Suggestions that Gekkota or Gekkota and Scincomorpha are nested within Anguimorpha (Evans and Barbadillo, 1998, 1999) (figs. 4, 5) carry important implications and require broad sampling of anguimorphs. The possibility of geckos being derived anguimorphs (Evans and Barbadillo, 1998), or closely related to scincomorphs (Caldwell, 1999a), or forming the sister group to Autarchoglossa (Estes et al., 1988) is important for understanding anguimorph outgroups. Evans and Wang (2005) suggested not only that geckos were nested anguimorphs, but also that scincomorphs may be.

Given all of this, the current study samples most densely within Iguania and Anguimorpha. Snakes, a very diverse group representing over one-third of extant squamates, are not densely sampled, but snake monophyly is virtually unquestioned. McDowell and Bogert (1954) were the last to suggest non-monophyly of snakes, suggesting that some scolecophidians might be more closely related to anguid lizards than to other snakes (see below). The analysis presented below is sensitive to that possibility.

Taxon Sampling

Many taxa included in this analysis are coded at the species level, but some are scored at supraspecific levels, including relatively inclusive levels such as “family” or, less commonly, at groups more inclusive than previously outlined “families”. Except in the cases of very small, morphologically homogenous clades, and coursely sampled morphologic character lists, at least some of the morphological diversity for a given group will not be coded when taxa are collapsed and coded supraspecifically. This type of coding requires assumptions regarding monophyly of the group and, usually, some of the phylogenetic relationships within the group. In the current analysis, I have focused on groups that are more problematic. For instance, macrostomatan snakes are universally identified as a monophyletic group. However, McDowell and Bogert (1954) suggested that scolecophidians were polyphyletic with Typhlopidae (including Anomalepidae given their usage) being nested within Anguidae (Anguinidae of their usage). Thus, in the current analysis, representatives of the three scolecophidian groups and ten other snake taxa are included and anguids are also sampled heavily. By contrast, most extant macrostomatans were included as a single terminal group here.

Published analyses have guided my selection of species exemplars for scoring supraspecific groups in this analysis. Given a clade with some resolution of basal taxa, collapsing the taxa into one unit is beneficial logistically because it reduces the total number of included taxa. Even so, future versions of this analysis will include more species-level codings, therefore relying less on the hypotheses of prior analyses. Below, explanations are offered both for taxon selection and character scoring. The explanations for taxon sampling specifically address those taxa that appear only as generic names.

The Outgroup

Carroll (1988b) describes a “Suborder” Eolacertilia as the sister taxon to other squamates, but Eolacertilia has been rejected as paraphyletic or polyphyletic (Estes et al., 1988; Gauthier et al., 1988a). As described by Estes (1983), this “Eolacertilia” includes Paliguanidae, Kuehneosauridae, Fulengia, Lacertulus, and Litakis. Carroll (1988b) includes these taxa and Colubrifer.

I have omitted the so-called eolacertilians for various, case-specific reasons. Carroll (1975, 1977, 1988a, 1988b) regards “paliguanids” as very important for understanding the origin and early evolution of lepidosaurs. Other authors have followed this opinion and have described new “paliguanids”, including Blomosaurus (Tatarinov, 1978) and Kudnu (Bartholomai, 1979). Even so, “Paliguanidae” is widely regarded as a paraphyletic taxon and, unfortunately, the preservation of specimens constituting the known “paliguanid” genera (including Paliguana, Palaeagama, and Saurosternon) makes it impossible to characterize them except through plesiomorphy (Benton, 1985; Gauthier et al., 1988a; Rieppel, 1994). Thus, their position within Lepidosauromorpha is currently impossible to ascertain with any kind of precision. Fulengia is not a lepidosaur, but instead is a juvenile sauropodomorph dinosaur (Evans, 1989). Litakis is based only on a dentary fragment lacking both the surangular margin and the symphysial portion (Estes, 1964). Thus, although probably representing a taxon of interest, Litakus is far too fragmentary to be informative for this analysis. Lacertulus has not been convincingly shown to be a squamate or lepidosaur (Carroll and Thompson, 1982; Estes, 1983; Carroll, 1988a; Gauthier et al., 1988a). Colubrifer has been shown to be a procolophonian, probably Owenetta (Evans, 2001).

Kuehneosauridae is a monophyletic assemblage (Robinson, 1962, 1967) and usually considered to be close to squamate ancestry (Robinson, 1962, 1967; Gauthier et al., 1988a; Rieppel, 1994; Evans, 2003). However, recent analyses have brought even this relationship into question, suggesting kuehneosaurids are the outgroup to Sauria (the archosauromorph-lepidosauromorph clade) (Müller, 2003, 2004b). Thus, kuehneosaurids have been omitted from this analysis.

Rhynchocephalia is the immediate sister-group to Squamata and is used as the outgroup in this analysis. Sphenodon punctatus and S. guntheri represent the only two living species of this ancient and historically diverse clade. Although Sphenodon differs from squamates in numerous aspects of its morphology that have been interpreted as plesiomorphic for Lepidosauria (Romer, 1966; Carroll, 1988b), comparisons with Triassic and Jurassic rhynchocephalians demonstrate that some of its supposedly plesiomorphic characteristics are actually derived (Evans, 1980, 1981; Fraser, 1982; Fraser and Walkden, 1984; Whiteside, 1986; Fraser, 1988; Carroll and Wild, 1994; Sues et al., 1994; Reynoso, 1996; Wilkinson and Benton, 1996; Evans and Sigogneau-Russell, 1997; Reynoso, 2000, 2003; Evans et al., 2001; Fraser, 2002). There is general agreement that Gephyrosaurus, Diphydontosaurus, Planocephalosaurus, and Rebbanasaurus are the most basal Rhynchocephalia (Sues et al., 1994; Reynoso, 1996; Wilkinson and Benton, 1996; Evans and Sigogneau-Russell, 1997; Reynoso, 2000; Evans et al., 2001). These taxa have been used to score the group wherever possible, with Sphenodon being used to supplement these codings.

Marmoretta is a recently described lepidosauromorph from the Middle Jurassic of Skye, Scotland (Evans, 1991; Waldman and Evans, 1994). It resembles kuehnoesuarids, but is somewhat less specialized based on the known remains (Evans, 1991; Waldman and Evans, 1994). However, given the uncertainty regarding the placement of kuehneosaurids, this taxon was not included as an outgroup. Rhynchocephalia stands as the sole outgroup for the phylogenetic analysis performed here, but character codings are given for Marmoretta and Kuehneosauridae (see appendix 2).

Stem Squamates

Most known squamates fit within one of the seven major radiations (Iguania, Gekkota, Lacertoidea, Scincoidea, Anguimorpha, Amphisbaenia, and Serpentes), but some fossil taxa defy placement within any of these groups. Recent descriptive and phylogenetic work suggests that some fossil taxa fall outside of the crown-group represented by this framework. Among these are Huehuecuetzpalli mixtecus (Reynoso, 1998), Hoyalacerta sanzi (Evans and Barbadillo, 1999; Evans et al., 2004), Scandensia ciervensis, “bavarisaurids”, and “ardeosaurids” (Evans and Barbadillo, 1998, 1999; however, see Conrad, 2004c; Conrad and Norell, 2006a, 2007b). Accordingly, these taxa have been included in this analysis with the latter two taxa being represented by most of their constituent taxa: Ardeosaurus brevipes, Bavarisaurus macrodactylus, Eichstaettisaurus, and Yabeinosaurus tenuis are not closely related to Gekkota (Evans and Barbadillo, 1997, 1998, 1999; Evans et al., 2005; Conrad and Norell, 2006a) (see figs. 4, 5, 9), contra earlier some earlier taxonomic treatments (e.g., Kluge, 1967, 1983, Estes, 1983). These taxa are included individually in the current analysis. Published data were used for scoring A. brevipes (Hoffstetter, 1966; Mateer, 1982; Estes, 1983; Evans and Barbadillo, 1998), B. macrodactylus (Estes, 1983; Evans, 1994b; Evans and Barbadillo, 1998), and Y. tenuis (Endo and Shikama, 1942; Young, 1958; Evans et al., 2005). Note that much of the data for Y. tenuis comes from an excellent recent study by Evans et al. (2005). Eichstaettisaurus is scored as a single taxon based on published descriptions (Hoffstetter, 1966; Estes, 1983; Evans and Barbadillo, 1998; Evans et al., 2004) and photos kindly supplied by Sterling J. Nesbitt. The “bavarisaurid” Palaeolacerta bavarica is not included in the current analysis because it lacks complete, available, descriptions. Tijubina pontei is an incompletely known taxon that is apparently similar to Huehuecuetzpalli mixtecus (Bonfirm-Júnior and Marques, 1997; Bonfirm-Júnior and Avilla, 2002; Bonfirm-Júnior and Rocha-Barbosa, 2006). It may be included in future versions of this analysis, pending further studies.

Iguania

Monophyly of Iguania as defined by Estes et al. (1988) is universally accepted. However, the precise relationships of the iguanian clades remain problematic. Importantly, most analyses not treating Iguania as a single taxon assume a basal dichotomy of iguanians with an “iguanid” group forming the sister group to Acrodonta (agamas and chameleons) and usually break Iguania into the three taxa Iguanidae, Agamidae, and Chamaeleontidae (Estes et al., 1988; Wu et al., 1996; Lee, 1998, 2000; Caldwell, 1999a; Caprette et al., 2004). The former two taxa are usually highlighted with an asterisk (*) to indicate their possible paraphyly. Lee and Caldwell (2000) divided Agamidae into two taxa in an attempt to eliminate paraphyletic taxa, but (like many other analyses) coded all nonacrodont iguanians as single unit.

Current understanding of iguanian phylogeny is rudimentary at best. Only three morphological analyses have been performed with the hopes of sorting out the broad-scale iguanian interrelationships. The first of these (Etheridge and de Queiroz, 1988) excluded acrodontans. A superior study (from a taxonomic sampling perspective) considered 35 extant iguanians and one fossil (Frost and Etheridge, 1989). Although Frost and Etheridge (1989) did not fully resolve the interrelationships of the iguanian groups (fig. 12), the strict consensus of their trees did show support for a number of clades that are used in this analysis, including Corytophanidae, Phrynosomatidae, Crotaphytidae, Iguanidae (sensu stricto) and Tropidurinae (sensu stricto). A recent molecular study (Schulte et al., 2003) (fig. 13) provides independent support for many (but not all) of the groups identified by Frost and Etheridge (1989).

Figure 12

Hypothesis of iguanian interrelationships, based on morphology, as presented by Frost and Etheridge (1989). As recently described by Schulte et al. (2003), iguanian relationships are a problematic area, seeming to defy any attempt to recover strong node supports. The letters in parentheses are implemented for easy comparison with Figure 11 (below). Fossil taxon denoted by a dagger (†).

i0003-0090-310-1-1-f12.gif

Figure 13

Hypothesis of iguanian interrelationships, based on molecular data, as presented by Schulte et al. (2003). The gray letters correspond with those in Figure 10. They refer to clades recovered by Frost and Etheridge (1989): (E) Leiocephalinae; (I) Liolaeminae; (P) Polychrotidae, polyphyletic here; (T) Tropidurinae. Note that Tropiduridae sensu Frost and Etheridge (1989) included leiocephalines, liolaemines, and tropidurines, but these taxa do not form a clade in the Schulte et al. (2003) analysis.

i0003-0090-310-1-1-f13.gif

Despite the consensus that Acrodonta is monophyletic, questions remain regarding the relationships of Agaminae (sensu Frost and Etheridge, 1989: 32–33), Leiolepis, Physignathus, Uromastyx, and Chamaeleonidae (Frost and Etheridge, 1989; Macey et al., 1997, 2000; Honda et al., 2000). The monophyly of Chamaeleonidae, on the other hand, has never been questioned. Indeed, the clade seems to be universally viewed as a very distinctive radiation of peculiar squamates (Hillenius, 1978; Moody and Rocek, 1980; Rieppel, 1981b, 1987; Estes et al., 1988; Frost and Etheridge, 1989; Macey et al., 2000; Townsend and Larson, 2002; Bauer, 2003; Uetz, 2007). Brookesia superciliaris and Rhampholeon spectrum were used in this analysis because of their apparently basal position within Chamaeleonidae in morphological- and molecular-based studies (Rieppel, 1981b, 1987; Townsend and Larson, 2002). Three representative “agamids” were coded in this analysis: Agama agama, Physignathus cocincinus, and Uromastyx (coded based primarily on U. aegyptius).

Priscagamids are a group of Late Cretaceous iguanians showing similarities with extant agamas. Three priscagamines were included in the present analysis: Priscagama gobiensis (probably including Chamaeleognathus iordanskyi and Cretagama Białynickae of Alifanov, 1996, and maybe Pleurodontagama aenigmatodes of Borsuk-Białynicka, 1996 based on data in Gao and Norell, 2000); Mimeosaurus crassus (probably including Gladidenagama semiplana); and Phrynosomimus asper. Data used for coding the individual priscagamids comes from previous descriptive studies (Borsuk-Białynicka and Moody, 1984; Alifanov, 1989b, 1996; Borsuk-Białynicka, 1996; Gao and Norell, 2000) and observation of specimens (see appendix 1).

Tikiguania estesi, known only from a dentary, is highly important in that it may be a Triassic iguanian (Datta and Ray, 2006), but is too incomplete to be coded meaningfully here. Its relationships must be tested by an analysis sampling basal rhynchocephalians and acrodontans more intensely.

Three representative hoplocercids were scored. These were Enyalioides (based on E. palpebralis and E. laticeps), Hoplocercus spinosus, and Morunasaurus annularis.

Camp (1923) considered Euposaurus to be a relative of anguimorphs (see fig. 1), Carroll (1988b) hypothesized that it was an iguanian, and Gauthier et al. (1988a: 97) suggested it to be a rhynchocephalian possibly close to “clevosaurs”. Evans (1993) has shown that the three species of Euposaurus represent a non-diagnostic lepidosaur and two relatively derived rhynchocephalians.

Gekkota

Gekkota is similar to Iguania in that there is complete consensus regarding the monophyly of a clade including Pygopodinae, Diplodactylinae, Gekkoninae, and Eublepharinae, but gekkotan interrelationships remain problematic. Few fossil forms represent this group and previously attributed forms (“ardeosaurids” and “bavarisaurids”) have been removed (Evans and Barbadillo, 1997, 1998), leaving the scoring of Gekkota to rest mainly upon extant forms. The extant taxa are scored based on a number of studies dating from the last 40 years (Kluge, 1967, 1969, 1974, 1983, 1987; Estes, 1983; Estes et al., 1988; Schwenk, 1988; Bauer, 1989; Rieppel, 1992; Hutchinson, 1997; Uetz, 2007) as well as upon preserved specimens. Three representative species are coded for Eublepharidae. These were chosen based on the phylogenetic hypotheses of morphology- (Grismer, 1988) and molecular-based studies (Ota et al., 1999). Included are two basal species (Aeluroscalabotes felinus and Coleonyx mitratus) and a more derived eublepharine (Hemitheconyx caudicinctus). The representative gekkonines were chosen similarly, based on the phylogenetic hypotheses presented by Kluge (1967, 1983, 1987) and Han et al. (2004).

Numerous fossil forms, besides those mentioned above, have been attributed to the Gekkota, but most are too incomplete for species level diagnosis. Noteworthy, though, are Gobekko cretacicus, Hoburogekko suchanovi, and Pygopus hortulanus. Of these three taxa, only Gobekko cretacicus is included in the current analysis, being represented by a relatively complete skull (Borsuk-Białynicka, 1990). Hoburogekko is represented by a partial skull and mandible showing enough features to demonstrate that it is a gekkotan (Alifanov, 1989a, 2000), but there are no characteristics that may diagnose it specifically or distinguish it from gekkonines, pygopods, or diplodactylines (Conrad and Norell, 2006a). Pygopus hortulanus is represented only by a dentary resembling (but distinct from) extant Pygopus species (Hutchinson, 1997). Although not included as a separate taxon in this analysis, this fossil offers important biostratigraphic information regarding the age of the pygopod lineage.

Scincomorpha

Scincomorpha is a diverse and speciose assemblage of lizards that may or may not represent a monophyletic group exclusive of Gekkota and/or Anguimorpha. Recent studies testing scincomorph monophyly in the broader context of squamates have produced both support for (Estes et al., 1988; Presch, 1988; Wu et al., 1996; Evans and Barbadillo, 1998; Caldwell, 1999a; Vicario et al., 2003; Evans and Wang, 2005; Evans et al., 2005) (see figs. 2Figure 3Figure 45, 8, 9) and evidence against it (Lee, 1998, 2000; Lee and Caldwell, 2000; Townsend, 2002; Townsend et al., 2004; Vidal and Hedges, 2005) (see figs. 6, 7, 10). Most of these studies agree that Lacertidae, Gymnophthalmidae, and Teiidae form a clade (Lacertiformes of Estes et al., 1988). Presch (1988) places Lacertidae with Scincidae, Cordyliformes [as used by, for example, Lang, 1991; Mouton and Wyik van, 1997; Cooper and Steele, 1999; Odierna et al., 2002; Lamb et al., 2003; Cordylidae of Presch's (1988) usage and that of some other authors; Cordylidae and Gerrhosauridae], and Xantusiidae. Xantusiidae is particularly problematic and has been suggested as having affinities with the lacertiforms (Estes et al., 1988; Caldwell, 1999a) (figs. 2 and 8, respectively), with Cordyliformes and Scincidae (Presch, 1988; Evans and Barbadillo, 1998; Vicario et al., 2003; Townsend et al., 2004; Vidal and Hedges, 2005) (see figs. 4, 10), or with Gekkota (Lee, 1998, 2000; Lee and Caldwell, 2000) (figs. 6, 7). Scincidae and Cordyliformes are typically suggested as being closely related, but some analyses have found no support for this clade, instead recovering a polytomy between these taxa and Anguimorpha (Lee, 1998, 2000; Lee and Caldwell, 2000).

Polyglyphanodontidae is a Cretaceous radiation of teiid-like lizards that has been suggested as forming a subfamily of Teiidae, possibly close to the Teiinae (Estes, 1983; Gao and Norell, 2000; though see Sulimski, 1975; Alifanov, 1993a). More recently, it has been shown that Sineoamphisbaena hexatabularis, previously identified as a basal amphisbaenian (Wu et al., 1996; Lee, 1998), probably represents a derived member of this radiation (Kearney, 2003a, 2003b).

Because of the disagreement about placement of polyglyphanodontines and Sineoamphisbaena hexatabularis with regard to Teiidae and other squamates (respectively), Teiidae is broken into Teiinae and Tupinambinae. These two taxa are well recognized as sister taxa and they have been scored based on numerous specimens (appendix 1) and literature (Estes, 1964, 1983; Estes et al., 1988; Uetz, 2007). Adamisaurus magnidentatus, Gobinatus, Cherminsaurus, Gilmoreteius ( = Macrocephalosaurus), Polyglyphanodon, and Sineoamphisbaena hexatabularis are also included, as well as the possible teiid Chamops (Estes, 1964, 1983; Gao and Fox, 1996) are included here. Bicuspidon numerosus (Nydam and Cifelli, 2002) and Peneteius aquilonius (Nydam et al., 2000) are also probable teiioids known from remains of similar incompleteness as Chamops. These taxa may be included in future iterations of this matrix. Chamops is included here and B. numerosus and P. aquilonius are not, in part, because the latter taxa are unquestioned as polyglyphanodontines, whereas there is some question as to the placement of Chamops within squamates (summarized in Estes, 1983; Gao and Fox, 1996).

Gymnophthalmidae is retained as a separate taxon following most recent studies and scored based mainly on published accounts (Presch, 1976, 1983, 1988; Estes et al., 1988; Kizirian, 1996; Kizirian and McDiarmid, 1998; Kizirian and Cole, 1999; Montero et al., 2002; Bell et al., 2003). The morphology of this important and intriguing group is understudied and deserves more attention.

Lacertidae is included as a single taxon. Data for Lacertidae comes not only from observations of preserved specimens, but also from literature regarding extant forms (Estes, 1983; Estes et al., 1988; Borsuk-Białynicka et al., 1999; Müller, 2001; Barbadillo and Martínez-Solano, 2002) and regarding the fossils Succinilacerta (Borsuk-Białynicka et al., 1999) and Dracaenosaurus croizeti (Müller, 2004a).

Representatives from each of the three extant xantusiid genera (Cricosaura, Lepidophyma, and Xantusia) are included in this analysis. Coding is based in part from observation of specimens and also on the literature (Rieppel, 1984a; Peterson and Bezy, 1985; Estes et al., 1988; Maisano, 2003a, 2003b, 2003c, 2003d). Palaeoxantusia is included and scored with the assumption that all three named species (P. fera, P. allisoni, and P. kyrentos) represent an exclusive, monophyletic, clade (Hecht, 1956; Schatzinger, 1980; Estes, 1983).

Cordyliformes includes only a few extant taxa and very few fossil forms. This is only one of six taxa scored above the level of “family” in this analysis, this because of the relatively low diversity known for the clade and its universal acceptance as monophyletic (McDowell and Bogert, 1954; Romer, 1956; Presch, 1988; Lang, 1991; Mouton and Wyik van, 1997; Odierna et al., 2002; Maisano, 2003e; Uetz, 2007).

Scincidae represents one of the most speciose and morphologically diverse “families” of squamates, yet in most cladistic analyses they are coded as a single taxon. There remains some debate about the topology of scincid interrelationships, but several monophyletic groups may be recognized. Greer (1970) recognized the four “subfamilies” Feyliniinae, Acontinae, Scincinae, and Lygosominae, but suggested that his Scincinae might be paraphyletic. Rieppel (1981a, 1982,; 1984c) described the skull and jaw adductor musculature in Acontinae and Feyliniinae, concluding that the former were derived scincids, but that the latter represented the sister lineage to the Scincidae; the Feyliniidae. Based on similarities in brain morphology, Northcutt (1978) suggested a close relationship between dibamids, scincids, and snakes (fig. 14). Hallermann (1998) alone has used morphology to test the monophyly of a clade containing the main lineage of skinks, acontines, and feyliniines, but was unable to resolve the interrelationships of these groups (fig. 15). Recent molecular work (Whiting et al., 2003) suggests that Acontinae is the sister taxon to other scincids, with Feyliniinae forming a sister group to a clade including Scelotes and Proscelotes (fig. 16), genera regarded as scincines by Greer (1970). Although this paints a confusing picture of scincid interrelationships, in reality these three studies are complementary and seem to identify four major clades of “scincids,” including a Feylinia clade, an Acontias clade, a Scelotes-Proscelotes clade, and a larger radiation including genera such as Scincus, Eumeces, Lygosoma, Mabuya, Plestiodon, and Tiliqua. These four clades are referred to here as Feyliniidae, Acontidae, Scelotidae, and Scincidae, respectively, and taxa were selected for scoring based on the topology presented in Greer (1970) and Whiting et al. (2003). These were scored based on observations of specimens (appendix 1) and published data (Greer, 1970; Haas, 1973; Rieppel, 1981a, 1982, 1984c).

Figure 14

Northcutt's (1978) vision of squamate interrelationships based on his studies of brain morphology.

i0003-0090-310-1-1-f14.gif

Figure 15

Hypothesis of squamate interrelationships, based on morphology, as presented by Hallermann (1998).

i0003-0090-310-1-1-f15.gif

Figure 16

Hypothesis of scincomorph interrelationships, based on molecular data, as presented by Whiting et al. (2003). Suprageneric taxon names in quotation marks are the “family” names applied to the scincoid groups in the text of the current study. This tree was used to guide selection of exemplars for the identified “family” groups. Note that Plestiodon is used following Smith (2005).

i0003-0090-310-1-1-f16.gif

In addition to these extant clades of scincomorphs, many additional fossil taxa have been referred to the group, but whose familial affinities are less clear than those described above. Many of the appropriately complete representatives are included in this analysis to offer some additional context for the more clearly resolved fossils, to aid in reconstructing scincomorph nodes, and thus help determine if Scincomorpha is holophyletic. Included in this analysis are Becklesius hoffstetteri, Eoxanta lacertifrons, Eolacerta robusta, Globaura venusta, Meyasaurus diazromerali, Paramacellodus oweni, Parmeosaurus scutatus, Pseudosaurillus (P. becklesi and P. sp. of Estes, 1983), Sakurasaurus shokawensis, Slavoia darevskii, Tchingisaurus multivagus, and Tepexisaurus tepexii. Several of these taxa have been redescribed based on new material recovered from Cretaceous rocks in the Gobi desert over the last several years (Gao and Norell, 2000). Gao and Norell (2000) described several new scincomorph taxa from the Gobi, including Hymenosaurus, Parmeosaurus scutatus, and Tchingisaurus multivagus. Eolacerta was recently redescribed (Müller, 2001) and primary coding for this taxon is based on that excellent paper. Evans and Barbadillo (1996) have shown that specimens referred to Meyasaurus diazromerali and Ilerdaesaurus represent a the single taxon (Meyasaurus diazromerali) included here. Becklesius hoffstetteri includes most of the material referred to Macellodus brodiei (Estes, 1983). Monophyly of Pseudosaurillus as described by Estes (1983) is ambiguous, and thus P. becklesi and P. sp. are included as separate taxa.

Numerous fossil scincomorphs are represented by partial mandibles or skull bits; these are omitted from the present analysis, but may be included in future versions of the data set. Notable examples of this include Estescincosaurus ( = Sauriscus) (Sullivan and Lucas, 1996), Peneteius, Leptochamops, Contogenys, and Palaeoscincosaurus.

Anguimorpha

Anguimorpha is represented by only about 181 extant species in the five clades Xenosaurus, Shinisaurus, Anguidae, Heloderma, and Varanidae. Even so, extant anguimorphs rival scincomorphs in morphological diversity. The genus Varanus alone rivals terrestrial mammals in its size range (Pianka, 1995). The interrelationships and inclusion or exclusion of other taxa, both extant and fossil, are more contentious. Most studies suggest a monophyletic Varanoidea including Heloderma, Lanthanotus borneensis, and Varanus to the exclusion of Anguidae and Xenosaurus (e.g. Lee, 1998; fig. 6), but recent molecular studies call even this into question (Macey et al., 1999; Townsend, 2002; Townsend et al., 2004; Vidal and Hedges, 2005) (see fig. 10).

In addition to these extant anguimorph clades, two major fossil groups have been identified. These are the Mosasauroidea and the Necrosauridae ( = Parasaniwidae of Estes, 1964). Besides these, a number of miscellaneous taxa that seem to defy placement in any previously defined group have been described.

Many recent phylogenetic analyses have coded Xenosauridae as a single taxon including Shinisaurus (Estes et al., 1988; Wu et al., 1996; Evans and Barbadillo, 1998; Lee, 1998, 2000). New data show this to be potentially misleading. It relies upon and incorporates erroneous morphological characterizations for Shinisaurus, Xenosaurus, or both (Conrad, 2004a, 2006a, 2006b). Consequently, Shinisaurus and Xenosaurus are here scored as separate taxa. Xenosaurus is scored from observations of both X. platyceps and X. grandis as well as from the literature (Barrows and Smith, 1947; McDowell and Bogert, 1954; King and Thompson, 1968; Rieppel, 1980a; Gao and Norell, 1998; Ramos et al., 2000; de Oca et al., 2001). Additionally, four fossil taxa have been referred to the Xenosauridae. Restes rugosus (Gauthier, 1982;  = Exostinus rugosus of Gilmore, 1942a) is represented by most of the dermal bones of the skull roof and included here, coded from the literature (Estes, 1975, 1983; Gauthier, 1982). Gauthier (1982) questioned the monophyly of Exostinus serratus and E. lancensis. Both are included here, with codings for E. serratus based mainly on Gilmore (1928) and Estes (1964, 1983), and E. lancensis based on Gilmore (1928), Gauthier (1982), Estes (1964, 1976, 1983), and Gao and Fox (1996). A fossil genus based on incomplete dentaries from the Cretaceous, Oxia, has additionally been referred to Xenosauridae (Gao and Nessov, 1998), but this material is non-diagnostic and is omitted here.

Recent analyses suggest that Carusia (the senior synonym of Shinisauroides as shown by Gao and Norell, 1998), sometimes considered an unusual scincomorph (Borsuk-Białynicka, 1985; Alifanov, 2000), is actually an anguimorph close to Xenosauridae (Gao and Hou, 1996; Gao and Norell, 1998, 2000; but see Conrad, 2006b) (fig. 11). This taxon is included in the current analysis. The possible carusioid relative Dalinghosaurus longidigitus is included based on published data (Ji, 1998; Ji and Ji, 2004; Evans and Wang, 2005).

Shinisaurus crocodilurus is included and its osteology and external morphology is scored almost exclusively from observations of skeletonized and preserved specimens, supplemented by data from the excellent study of Bever et al. (2005a). Muscle characters and other morphological data are derived from the published literature (McDowell and Bogert, 1954; Haas, 1960; Rieppel, 1980a; Zhang, 1991; Zhao et al., 1999). The recently described Bahndwivici ammoskius is included based on direct observations summarized by Conrad (2006b).

Monophyly of Anguidae is unquestioned, but no cladistic analysis has ever tested the relationships and interrelationships of the entire anguid clade. Based on molecular data (Macey et al., 1999; Wiens and Slingluff, 2001) and overall similarity (Uetz, 2007), the extant genera may be divided into the Gerrhonotinae (Abronia, Barisia, Coloptychon, Elgaria, Gerrhonotus, and Mesaspis), the Diploglossinae (Celestus, Diploglossus, and Ophiodes), and the Anguinae (Anguis, Ophisaurus, and Pseudopus), with Anniella representing its own lineage or nested in one of the others. Indeed, Macey et al. (1999) and Wiens and Slingluff (2001) demonstrate paraphyly of Ophisaurus with respect to both Pseudopus and Anguis, and paraphyly of Diploglossus with respect to the other diploglossines (fig. 17). More broadly, a fossil radiation close to extant anguids, the glyptosaurs, is usually overlooked entirely. For example, Gao and Norell (1998) include representative members of the four major clades of extant Anguidae in their analysis, but did not address glyptosaurs.

Figure 17

Hypotheses of anguid interrelationships, based on molecular data, as presented by (A) Macey et al. (1999) and (B) Wiens and Slingluff (2001).

i0003-0090-310-1-1-f17.gif

Because of the poorly understood relationships of this clade as a whole, numerous taxa are included here. Included here are the extant taxa Anguis fragilis, Anniella pulchra, Ophisaurus ventralis, O. attenuatus, Pseudopus apodus, Dopasia harti, Diploglossus millepunctatus, Celestus costatus, Ophiodes sp., Gerrhonotus liocephalus, Abronia deppii, and Barisia imbricata from specimens (see appendix 1). Additional data was taken from the literature for A. pulchra (Coe and Kunkel, 1906; McDowell and Bogert, 1954; Bellairs, 1970; Rieppel, 1978, 1980b; Gao and Norell, 1998), and A. fragilis (Bellairs, 1970; Rieppel, 1980b; Iordansky, 1997; Gao and Norell, 1998). Published codings and descriptions were used for G. multicarinata, Ab. mixteca, P. apodus, Di. lessonae, and Ophiodes striatus (Meszoely, 1970; Rieppel, 1980a; Gao and Norell, 1998).

Taxonomy and scoring of fossil taxa comes from their associated reference literature and observation of specimens. Taxa included paired with the literature that was used to code or supplement their codings here are Apodosauriscus minutus (Gauthier, 1982), Arpadosaurus gazinorum (Meszoely, 1970), Bainguis parvus (Borsuk-Białynicka, 1984; Gao and Hou, 1996), Glyptosaurus sylvestris sensu Sullivan (1979, 1986, 1989) (Gilmore, 1928; Sullivan, 1979,1986, 1989; Estes, 1983), Helodermoides tuberculatus (Gilmore, 1928; Sullivan, 1979, 1986, 1989; Estes, 1983), Melanosaurus maximus (Estes, 1983), Odaxosaurus sensu Gauthier (1982) (Meszoely, 1970; Sullivan, 1979; Gauthier, 1982; Estes, 1983), Ophisauriscus quadrupes (Sullivan et al., 1999), Paragerrhonotus ricardensis (Estes, 1963, 1983), Paraglyptosaurus princeps (Sullivan, 1979), Parophisaurus pawneensis (Sullivan, 1987), Peltosuarus granulosus sensu Estes (1983) (Gilmore, 1928; Estes, 1964, 1983), Proglyptosaurus huerfanensis sensu Sullivan (1989) (Sullivan, 1979, 1989), Proxestops jepseni sensu Gauthier (1982) (Estes, 1964, 1983; Gauthier, 1982), and Xestops vagans sensu Sullivan (1979) (Meszoely, 1970; Meszoely et al., 1978; Sullivan, 1979; Estes, 1983) (see appendix 1).

Noteworthy exclusions from this analysis include the gerrhonotine genera Coloptychon and Mesaspis among extant forms. Placosaurus and Paraplacosauriops were recently reviewed (Augé and Sullivan, 2006; Sullivan and Augé, 2006) and will be included in future versions of this analysis. Eodiploglossus (Gauthier, 1982) awaits redescription.

Heloderma includes the extant beaded lizard (H. horridum) and the Gila monster (H. suspectum). These species and the fossil H. texana are included in the present analysis as individual taxa in order to determine their positions relative to one another. Extant Heloderma are scored primarily based on observations of specimens (appendix 1). Heloderma texana was scored based on published descriptions (Estes, 1983; Pregill et al., 1986) and digital scans (Maisano, 2001a).

Although Heloderma is the sole extant genus of Helodermatidae, a number of fossil taxa have been associated with this clade. The following taxa are included based on observations of specimens and published data: Estesia mongoliensis (Norell et al., 1992; Norell and Gao, 1997; Gao and Norell, 1998, 2000; Nydam, 2000), Eurheloderma gallicum (Hoffstetter, 1957; Estes, 1983; Norell and Gao, 1997; Gao and Norell, 1998; Nydam, 2000), Gobiderma pulchrum (Borsuk-Białynicka, 1984; Gao and Norell, 1998, 2000), Lowesaurus matthewi (Gilmore, 1928; Estes, 1983; Pregill et al., 1986; Gao and Norell, 1998), Paraderma bogerti (Estes, 1964; Gao and Fox, 1996; Nydam, 2000), and Primaderma nessovi (Cifelli and Nydam, 1995; Nydam, 2000). Both extant species of Heloderma are included and are scored based on observations of specimens and also on published data (McDowell and Bogert, 1954; Bogert and Del Campo, 1956; Rieppel, 1980a; Pregill et al., 1986; Bernstein, 1999).

Varanidae includes at least 45 extant species (Pianka, 1995; Fuller et al., 1998; Ast, 2001) and possibly more than 60 (Uetz, 2007) in the two genera Lanthanotus borneensis and Varanus. Although the affinities of the monospecific Lanthanotus borneensis have been problmatic in the past (McDowell and Bogert, 1954), a consensus opinion has arisen that it is the extant sister taxon to Varanus (Rieppel, 1980a; Pregill et al., 1986; Lee, 1997, 1998; Evans and Barbadillo, 1998; Gao and Norell, 1998; Lee and Caldwell, 2000; though see Caldwell, 1999a) (figs. 6Figure 7Figure 89, 11). Lanthanotus borneensis was scored based on observations of specimens and on published descriptions (McDowell and Bogert, 1954; Haas, 1973; Rieppel, 1980a, 1980b, 1983; Maisano, 2001b; Maisano et al., 2002).

Recently, molecular data have been employed to identify clades within Varanus, with some consistency of results (Fuller et al., 1998; Ast, 2001, 2002; Pepin, 2001). Ast (2002) is the most recent of these analyses and also the most species-inclusive. Multiple species of Varanus are included in the present analysis to help identify some of the broader relationships among Varanus, because of their relatively certain monophyly based on comparisons of various analyses, and to test their monophyly exclusive of similar taxa such as Megalania prisca, Saniwa ensidens, and Saniwides mongoliensis; something not previously tested. Species used for coding these taxa were selected based on the cladistic relationships suggested by molecular data (Ast, 2001, 2002; Pepin, 2001) and were scored based on observations of specimens and published data (Mertens, 1942a, 1942b, 1942c; Bellairs, 1949; McDowell and Bogert, 1954; Bellairs, 1970; Haas, 1973; Rieppel, 1980a; Jenkins and Goslow, 1983; Zaher and Rieppel, 1999a). In addition to these other taxa, one fossil species described as Varanus rusingensis (Clos, 1995) is included because it is the earliest known specimen that may be reliably referred to Varanus.

Numerous fossils have been referred to the Varanidae and many are included here. Megalania prisca is the largest-known terrestrial lizard and may belong within crown group Varanus (Hecht, 1975; Molnar, 1990, 2004; Lee, 1995). This taxon is incomplete, but the characters for which it may be coded show non-synonymous coding with other observed taxa.

Several species have been referred to Saniwa, but only the type species (S. ensidens) and a possible “necrosaurid” (below) are included here based on published descriptions (Gilmore, 1928; Estes, 1983; see also the necrosaurid literature used below). Estes (1983) regarded S. agilis as a probable synonym of S. ensidens. The type specimens for S. brooksi, S. crassa, S. grandis, S. orsmaelensis, and S. paucidens are isolated vertebrae or a series of several vertebrae that are probably too incomplete for generic or specific diagnosis (Estes, 1983; Augé, 2005); these taxa are not included here. Vertebrae from the Eocene of Kirghizia have been referred to ?Saniwa sp. (Averianov and Danilov, 1997), but these lack varanid characteristics and are otherwise non-diagnostic. Saniwa australis is a nomen dubium (Báez and de Gasparini, 1977; Estes, 1983) and is not included.

A European fossil broadly resembling varanids has been described as Saniwa feisti (Stritzke, 1983). Although probably not representing a Saniwa, this taxon is relatively completely known (Stritzke, 1983; Keller and Schaal, 1992) and is included in the analysis.

Saniwides mongoliensis, Telmasaurus grangeri, and Cherminotus longifrons, all from the Gobi, show varanid affinities and are included here based observations of specimens and published descriptions (Gilmore, 1943; Estes, 1983; Borsuk-Białynicka, 1984; Gao and Norell, 2000). Gao and Norell (2000) have added two new Gobi lizards to the Varanidae. One of these, Aiolosaurus, is included in this analysis; the other was deemed too incompletely known by Gao and Norell (2000) to be named and is not included here.

Palaeosaniwa canadensis represents a problematic taxon identified variably as a varanid (Gilmore, 1928; Estes, 1964, 1983) or a helodermatid (Balsai, 2001). Regardless, it is included in this analysis based on the original descriptions (Gilmore, 1928; Estes, 1964, 1983) and on a recently described new specimen (Balsai, 2001).

Various fossils too incomplete to be included here have been referred to the Varanidae. Noteworthy is the Middle Miocene Iberovaranus catalaunicus (Hoffstetter, 1969; Estes, 1983).

Dolichosauridae is a group of Cretaceous lizards with reduced limbs and an overall morphology that looks to be intermediate between terrestrial varanids and the fully aquatic mosasaurs (Romer, 1966; Carroll, 1988b; Caldwell, 2000; Lee and Caldwell, 2000). Most of the taxa referred to as dolichosaurids have been recently redescribed and all are included here at the generic level or below. Included are Adriosaurus suessi ( = Acteosaurus crassicostatus) (Lee and Caldwell, 2000; Caldwell and Lee, 2004), Coniasaurus (based on both C. crassidens and C. gracilodens) (Bell et al., 1982; Caldwell, 1999b; Caldwell and Cooper, 1999), Dolichosaurus longicollis (Caldwell, 2000), Eidolosaurus trauthi (Nopcsa, 1923a), and Pontosaurus ( = Hydrosaurus) (Kornhuber, 1873; Dal Sasso and Pinna, 1997). Dal Sasso and Pinna (1997) described Aphanizocnemus, demonstrating some affinities with Adriosaurus and Dolichosaurus longicollis (Lee and Caldwell, 2000).

Aigialosauridae is usually considered a paraphyletic assemblage of basal mosasauroids forming successively more proximal outgroups to Mosasauridae (Bell, 1997; Lee, 1997; Caldwell, 1999a; Dutchak, 2005, 2006; Dutchak and Caldwell, 2006), but some analyses (Caldwell, 1996, 2000; Bardet et al., 2003) recover an aigialosaurid clade. Because of this ambiguity and to avoid misinterpretation of the basal character polarities for a single-taxon aigialosaur group, the aigialosaur species are included individually here. Aigialosaurus dalmaticus is the only species of Aigialosaurus included and it is scored from Carroll and DeBraga (1992). There is some question as to the generic and/or specific distinctiveness of Opetiosaurus buchichi from A. dalmaticus (Carroll and DeBraga, 1992; Caldwell, 1996, 2000), although recent work suggests more convincingly that they are distinct (Dutchak, 2005, 2006; Dutchak and Caldwell, 2006). Both are included based primarily on the descriptions of Carroll and DeBraga (1992), Dutchak (2005), and Dutchak and Caldwell (2006). Carsosaurus marchesetti seems to represent a taxon distinct from Aigialosaurus and Opetiosaurus and is here included based on recent descriptive works (Carroll and DeBraga, 1992; Caldwell, 1996, 2000). Tethysaurus nopcsai is a recently described lizard of a similar “grade” as aigialosaurids (Bardet et al., 2003) also included in the analysis.

Proaigialosaurus has been lost (Carroll and DeBraga, 1992) and is not included in the present analysis. The recently described Dallasaurus turneri (Bell and Polcyn, 2005) and Russellosaurus coheni (Polcyn and Bell, 2005) will be included in future analyses.

Mosasauridae is a Late Cretaceous radiation of large aquatic lizards. Existing phenetic groupings and cladistic systematic analyses (Russell, 1967; Carroll, 1988b; DeBraga and Carroll, 1993; Bell, 1997) have been limited in scope to mostly North American and European taxa and have had conflicting results (DeBraga and Carroll, 1993; Bell, 1997) (figs. 18, 19). Given this, scoring a single-taxon Mosasauridae would be difficult; so multiple mosasaurid taxa are scored for this analysis based on previous descriptive and phylogenetic studies. Halisaurus arambourgi, H. platyspondylus and Eonatator sternbergii ( = Halisaurus sternbergii) are included as separate taxa based on recent descriptive works (Russell, 1967; DeBraga and Carroll, 1993; Holmes and Sues, 2000; Bardet and Suberbiola, 2001; Bardet et al., 2005). Debate continues over the distinctiveness of these taxa at both the generic and specific level (Caldwell, 1996; Lingham-Soliar, 1996; Bell, 1997; Holmes and Sues, 2000; Bardet and Suberbiola, 2001), but they are not identical in coding and so each is included.

Figure 18

Hypothesis of mosasaur interrelationships as presented by DeBraga and Carroll (1993). Note that Helodermatidae and Varanidae include extant taxa, but all other taxa are fossils.

i0003-0090-310-1-1-f18.gif

Figure 19

Hypothesis of mosasaur interrelationships as presented by (A–B) Bell (1997) and (C) Bell and Polcyn (2005) and Polcyn and Bell (2005). Note that the taxon labeled “mosasaurines (Bell 1997)” refers to Clidastes, Globidensini, and Plotosaurini from (A). All displayed taxa are fossils.

i0003-0090-310-1-1-f19.gif

Clidastes liodontus is included based mainly on data from Russell (1967). Multiple species of Clidastes have been described (Russell, 1967; Bell, 1997) and may form a paraphyletic grade (Bell, 1997; Christiansen and Bonde, 2002). Only C. liodontus is used here based mainly on accessibility of descriptions and specimens.

Ectenosaurus clidastoides was initially considered a species of Platecarpus, but Russell (1967) identified it as representing an independent lineage. This species is included and scored based on photos of Sternberg Museum VP 40 generously provided by M. J. Everhart (personal commun.) and data in Russell (1967) and DeBraga and Carroll (1993).

Data for both species of Globidens (G. alabamaensis and G. dakotensis) were used to score the single-taxon Globidens (Gilmore, 1911; Russell, 1967, 1975; Bell, 1997; Lingham-Soliar, 1999a). Carinodens is not included because it is very incompletely known and strongly resembles Globidens in scoreable features (Lingham-Soliar, 1999a). Carinodens and other mosasaurids not included here may be included in future incarnations of this analysis.

Goronyosaurus nigeriensis is an overlooked taxon that has appeared in no previous analyses of squamate or mosasaur relationships. Originally described as Mosasaurus nigeriensis (Azzaroli et al., 1975), Goronyosaurus is currently a monospecific taxon whose anatomy was reviewed and clarified by Soliar (1988). This mosasaurid was coded based on descriptive works (Swinton, 1930; Azzaroli et al., 1975; Soliar, 1988; Lingham-Soliar, 1991, 1999b, 2002).

Hainosaurus is included here based on a recent redescription of H. bernardi from a nearly complete skeleton (Lingham-Soliar, 1992a) and on the account given by Russell (1967). Hainosaurus gaudryi (Lingham-Soliar, 1992a) and H. pembinensis (Nicholls, 1988; Lingham-Soliar, 1992a) may or may not be distinct from H. bernardi; only H. bernardi is included in this analysis.

Lakumasaurus antarcticus represents the only relatively complete squamate currently known from Antarctica. It has been briefly described as a basal tylosaurine and this description was used for coding character states in this analysis (Novas et al., 2002).

Moanasaurus mangahouangae is a mosasaurid from New Zealand that encompasses specimens previously attributed to Mosasaurus flemingi and Rikisaurus tehoensis. This taxon is scored based on published descriptions of these specimens (Wiffen, 1980, 1990).

Under its recent usage, the name Mosasaurus represents a speciose assemblage of mosasaurids probably not representing a monophyletic group to the exclusion of Plotosaurus (Bell, 1997; Christiansen and Bonde, 2002; Bell and Polcyn, 2005; Polcyn and Bell, 2005). Russell (1967) reviewed all of the then-described species of Mosasaurus. Based on this review, M. conodon, M. ivoensis, and M. missouriensis are relatively incompletely known taxa that are difficult to differentially diagnose. The position of these taxa must be further analyzed elsewhere. The two species of Mosasaurus included in this analysis, M. hoffmanni and M. lemonnieri, are based on reasonably complete specimens. Mosasaurus maximus has been convincingly shown to be synonymous with M. hoffmanni (Mulder, 1999) and so specimens and descriptions of both species are used for coding M. hoffmanni (Russell, 1967; Lingham-Soliar, 1995; Bell, 1997; Mulder, 1999). Lingham-Soliar (2000) rejects the synonymy of M. lemonnieri with M. conodon and his descriptive work is used for coding M. lemonnieri here.

Platecarpus is another relatively speciose group of mosasaurids probably not representing a holophyletic clade with respect to other taxa. Plioplatecarpus has been demonstrated to fall within the Platecarpus radiation (Bell, 1997; Christiansen and Bonde, 2002; Bell and Polcyn, 2005; Polcyn and Bell, 2005). As described by Russell (1967), Platecarpus tympaniticus probably represents the senior subjective synonym for P. ictericus and P. coryphaeus ( = P. abruptus) (but see Christiansen and Bonde, 2002). Platecarpus planifrons is not included in this analysis, pending full description of a new specimen. Here, Platecarpus has been scored based on P. tympaniticus and its subjective synonyms based on published data (Williston, 1910; Russell, 1967) and examination of specimens.

Plioplatecarpus is presumed to be a monophyletic radiation of mosasaurids, but is known from relatively few good specimens. Here, Plioplatecarpus was scored using data presented in previous descriptive and phylogenetic works of various specimens and species, especially P. primaevus sensu Holmes (1996) (Russell, 1967; Burnham, 1991; Lingham-Soliar, 1992b; Holmes, 1996; Bell, 1997; Christiansen and Bonde, 2002).

Plotosaurus ( = Kolposaurus of Camp, 1942) is known from two species, Plotosaurus bennisoni and P. tuckeri, that may form a single species. Apparently, they are a monophyletic group scored here as a single taxon based on the original description (Camp, 1942).

Prognathodon is a speciose genus of large-bodied mosasaurids (Russell, 1967; Lingham-Soliar and Nolf, 1989; Bell, 1997; Christiansen and Bonde, 2002). The monophyly of this genus has been tested and corroborated (Christiansen and Bonde, 2002). Recent studies (Bell, 1997; Bell and Polcyn, 2005; Polcyn and Bell, 2005) have suggested that the poorly known Plesiotylosaurus crassidens (not included here) may belong within this radiation. Two representative species of Prognathodon, P. overtoni and P. solvayi, have been scored individually here based on previous phylogenetic codings and descriptions (Bell, 1997; Christiansen and Bonde, 2002) and published descriptions (Russell, 1967; Lingham-Soliar and Nolf, 1989).

Based on a recent analysis (Novas et al., 2002), Tylosaurus proriger and T. nepaeolicus may not represent a monophyletic assemblage with respect to Hainosaurus. Because of this uncertainty, these two species must be treated separately. However, T. nepaeolicus is in need of redescription, so only T. proriger is included here and coded based on observations of specimens (appendix 1) and published literature (Osborn, 1899; Russell, 1967; Bell, 1997).

Various other mosasaurs have been described and named based on inferior remains and are omitted here. Notable among these are Amphekepubis and Pluridens. Amphekepubis is a pelvis with a few associated vertebrae that may or may not be diagnostic (Russell, 1967). Pluridens is represented by a single dentary from Niger that is probably diagnostic of a new taxon (Lingham-Soliar, 1998), but it is too incomplete for meaningful inclusion here.

Necrosauridae is generally acknowledged as a problematic group that may not represent a monophyletic assemblage (Pregill et al., 1986; Norell et al., 1992; Evans, 1994a; Lee, 1997; Gao and Norell, 1998; Conrad, 2005b). Although incompletely known, most “necrosaurids” preserve enough informative morphology for specific diagnosis and are included in this analysis. Included taxa have been scored primarily on published descriptions (but see appendix 1). Besides Necrosaurus, the taxa included here and their associated references are Colpodontosaurus (Estes, 1964, 1983), Eosaniwa (Estes, 1983; Gao and Norell, 1998; Rieppel et al., 2007), Parasaniwa (Gilmore, 1928; Estes, 1964, 1975, 1976; Gao and Fox, 1996; Gao and Norell, 1998), Parviderma (Borsuk-Białynicka, 1984; Gao and Norell, 1998), and Proplatynotia (Borsuk-Białynicka, 1984; Gao and Norell, 1998). Necrosaurus itself may or may not represent a monophyletic group. Necrosaurus cayluxi, N. eucarinatus, and Saniwa feisti (possibly a Necrosaurus, see above) have been coded as separate taxa based on published data (Hoffstetter, 1943; Haubold, 1977; Rage, 1978; Estes, 1983; Stritzke, 1983; Keller and Schaal, 1992; Augé, 2005).

Several taxa of a similar “grade” to necrosaurs have been described and are included here. Paravaranus angustifrons was suggested as having possible affinities with Mosasauroidea by Alifanov (2000); this was independently confirmed by a recent analysis (Rieppel et al. 2007). It is included here and was coded based on its original description (Borsuk-Białynicka, 1984). Dorsetisaurus purbekensis was coded based on Hoffstetter (1967a). Parviraptor is a problematic taxon of this “grade,” that may or may not represent a single taxon (Evans, 1994a). A possible anguimorph from Portugal, Lisboasaurus estesi (Seiffert, 1973) needs re-evaluation and redescription and is not included here.

Dibamidae

Dibamidae exhibits low diversity, including only about 18 species in Dibamus and Anelytropsis papillosus (Das and Lim, 2005; Uetz, 2007). Dibamids are important for inclusion in this analysis because of their uncertain affinities. Importantly, no broad-scale cladistic analysis has addressed the monophyly of Dibamidae, despite some relatively plesiomorphic character states in Anelytropsis papillosus and their disparate ranges (Rieppel, 1984b; Greer, 1985; Uetz, 2007). Consequently, published descriptions were used to score Dibamus novaeguineae (Gasc and Renous, 1979; Rieppel, 1984b; Greer, 1985) and Anelytropsis papillosus (Greer, 1985) separately in this analysis. There are no known fossil dibamids.

Amphisbaenia

Amphisbaenia is a bizarre clade of mostly limbless squamates whose interrelationships have been recently re-analyzed very thoroughly (Kearney, 2001, 2003a; Kearney and Stuart, 2004). These analyses are the basis for dividing amphisbaenian diversity into the five taxonomic units (Rhineuroidea, Trogonophidae, Amphisbaenidae, Bipes, and Blanus) scored here. Also based on the topology of Kearney (2001, 2003a), taxa near the bases of these major lineages were used to score those lineages from the literature (Zangerl, 1944; Gans, 1960; Montero and Gans, 1999; Kearney, 2001, 2002, 2003a; Maisano, 2003f, 2003g, 2003h, 2003i; Kearney and Maisano, 2004) and from direct observations of specimens. Taxa excluded from consideration also follows that of Kearney (2003a). Future versions of this analysis will include multiple species from each of the major amphisbaenian lineages.

Serpentes

Although the monophyly of snakes ( = Ophidia sensu Caldwell and Lee, 1997) is unquestioned, the interrelationships of the major snake taxa are not. Because Cretaceous limbed marine snakes may represent primitive snakes (Caldwell and Lee, 1997; Lee, 1997, 1998; Lee et al., 1999a, 1999b; Caldwell, 2000; Lee and Caldwell, 2000; Rage and Escuillé, 2000; Scanlon and Lee, 2000) or derived alethinophidians (Zaher and Rieppel, 1999b, 2002; Rieppel and Zaher, 2000a, 2000b, 2001; Tchernov et al., 2000; Rieppel et al., 2003), reliably reconstructing the ancestral states for snakes and scoring them as a single taxon is not a viable option if the idea is to test the phylogenetic placement of snakes within squamates. For the purposes of this study, snakes are broken up into 12 separate taxa: Anilioidea, Dinilysia patagonica, Eupodophis descouensi, Haasiophis terrasanctus, Leptotyphlops goudottii, Liotyphlops albirostris, “other macrostomatans”, Pachyophis woodwardi, Pachyrhachis problematicus, Typhlops lineolatus, Wonambi naracoortensis, and Xenopeltis unicolor. All taxa based mainly upon extant species (Anilioidea, macrostomatans, and scolecophidians) were scored via a combination of observations of specimens (appendix 1) and published literature.

Three representative scolecophidians are included based on specimens and published data (Evans, 1955; List, 1966; Parker and Grandison, 1977; Scanlon and Lee, 2000; Tchernov et al., 2000; Kley, 2006). Anilioidea is of questionable monophyly, supported by recent analyses (Scanlon and Lee, 2000; Tchernov et al., 2000; Lee and Scanlon, 2002a) and is tentatively accepted here. Most of the coding for Anilioidea comes from published literature (Rieppel, 1977, 1979; Rieppel and Zaher, 2000a, 2002; Scanlon and Lee, 2000; Tchernov et al., 2000; Lee and Scanlon, 2002a). Tchernov et al. (2000) found Pachyrhachis problematicus and Haasiophis terrasanctus to be nested within Macrostomata. To make this analysis sensitive to that possibility, extant Macrostomata was broken into Xenopeltis [with supplementary codings from the literature (Scanlon and Lee, 2000; Tchernov et al., 2000; Lee and Scanlon, 2002a)] and a group informally termed “other macrostomatans” to receive all other extant taxa traditionally considered to be macrostomatans (see Rieppel and Zaher, 2000a; Scanlon and Lee, 2000; Tchernov et al., 2000; Lee and Scanlon, 2002a; codings also based on data taken from these studies).

Fossil snakes were coded based on the following literature: Dinilysia patagonica (Estes et al., 1970; Rage, 1984; Caldwell and Albino, 2001, 2003), Eupodophis descouensi (Rage and Escuillé, 2000; Rieppel et al., 2003; Rieppel and Head, 2004), Haasiophis terrasanctus (Tchernov et al., 2000; Rieppel et al., 2003), Pachyophis woodwardi (Nopcsa, 1923a; Lee et al., 1999b; Caldwell and Albino, 2001; Rage and Escuillie, 2002; Rieppel et al., 2003; Rieppel and Head, 2004), Pachyrhachis problematicus (Zaher and Rieppel, 1999b, 2002; Lee and Caldwell, 2000; Rieppel and Zaher, 2000a; Caldwell and Albino, 2001; Polcyn et al., 2005a, 2005b), Wonambi naracoortensis (Scanlon and Lee, 2000; Rieppel et al., 2002).

Data Analysis

Principle Tree Searches

The data matrix used in this analysis is very large and a PAUP* (Swofford, 2001) analysis of the data set would take months. Goloboff et al. (2003) offer an alternative, a program called T.N.T. (tree analysis using new technology), which is very efficient at analyzing this type of data set (Hovenkamp, 2004). T.N.T. was used here with the specifications given below.

Consensus Trees and Apomorphy Lists

T.N.T. does not offer the option of reconstructing Adams consensus trees. The principle trees were exported from T.N.T. to PAUP* (Swofford, 2001) for reconstruction of Adams consensus trees and for the reconstruction of an apomorphy list.

Viewing Trees

Principle trees were primarily viewed in TreeViewX, version 0.4 (Page, 2004). This program was used to quickly view the principle trees and to discover the alternative placements of some volatile taxa (e.g., basal varanoids; see below).

Deformation Comparisons and Decay Indices

Deformation comparisons (see below) were performed using the program Mesquite (Maddison and Maddison, 2006), a program which allows easy movement of branches within a tree with simultaneous reports of length change. Decay indices (Bremer support) were calculated using the “Suboptimal Search” criteria in T.N.T. (Goloboff et al., 2003).

Institutional Abbreviations

AMNH, American Museum of Natural History; FMNH, Field Museum of Natural History; GM, Geiseltal Museum of the Martin-Luther-Universität in Halle/Saale (Germany); IGM, Institute of Geology, Mongolian Academy of Sciences, Ulaanbaatar, Mongolia; REE, Richard E. Etheridge Collection; UF, University of Florida, Florida State Museum.

Character List for Squamata

Below is a list of morphological characters used in this study. Each character and some of the character states are followed by an abbreviation identifying the publication from which the character or character state was taken. This reference does not always correspond with the original use of a particular character or character state, but rather the specific study used to derive the character as used in this study. The abbreviations are listed below with their corresponding study listed afterward. A dash (-) and a number representing the character number from the original study follow these abbreviations in most cases. Some characters come from studies that did not include character lists; a dash and number do not follow abbreviations associated with these characters. Abbreviations are present after all of the character descriptions, but some character states are followed by abbreviations indicating that these character states were not originally identified for that character or were not identified for that character in the cited study.

Citation Abbreviations

AM, Abdala and Moro, 2003; B, Bell, 1997; B82, Branch, 1982; B86, Beuchat, 1986; BB, Borsuk-Białynicka, 1983; C99, Caldwell, 1999a; CDB, Carroll and DeBraga, 1992; CN, Conrad and Norell, 2006a; CRG, Conrad et al., 2007; C06, Conrad, 2006b; DBC, DeBraga and Carroll, 1993; E, Estes et al., 1988; E83, Estes, 1983; Eagam, characters from Estes et al., 1988 that do not appear in the larger list of characters, but that are found only in the section describing Agamidae*; Eanguim, characters from Estes et al., 1988 that do not appear in the larger list of characters, but that are found only in the section describing Anguimorpha; EB98, Evans and Barbadillo, 1998; Echam, characters from Estes et al., 1988 that do not appear in the larger list of characters, but that are found only in the section describing Chamaeleonidae; EdQ88, Etheridge and de Queiroz, 1988; Egek, characters from Estes et al., 1988 that do not appear in the larger list of characters, but that are found only in the section describing gekkotans; FE, Frost and Etheridge, 1989; Ga82, Gauthier, 1982; Ga84, Gauthier, 1984; Ga88, Gauthier et al., 1988a; GN98, Gao and Norell, 1998; GN00, Gao and Norell, 2000; Gr85, Greer, 1985; Gr88, Grismer, 1988; H93, Harvey, 1993; Ke, Kearney, 2003a; Kl87, Kluge, 1987; L98, Lee, 1998; LC00, Lee and Caldwell, 2000; M70, Meszoely, 1970; MB54, McDowell and Bogert, 1954; McG, McGuire, 1996; NG, Norell and Gao, 1997; PGG86, Pregill et al., 1986; R80, Rieppel, 1980a; R84, Rieppel, 1984a; R80L, Rieppel, 1980b; Rs80, Rieppel, 1980c; RZ, Rieppel and Zaher, 2000a; S, Schwenk, 1988; TC00, Tchernov et al., 2000; Y76, Yatkola, 1976.

Character Descriptions

1. Skull, percentage of total length made up by antorbital snout (DBC-2): (0) <30%; (1) >30%; (2) >45%; (3) >50%. The structure of this character allows that it be ordered. Logically, if a snout is 50% of the total skull length, then it is also more than 30% or 45%. The character states used for this character are somewhat arbitrarily delimited, but are descriptive. They largely follow the character states put forward in DeBraga and Carroll (1993).

2. Skull, rostrum anterior to the bony external nares (new/extensively modified): (0) short, absent; (1) four tooth positions long or more. This character refers to the amount of the premaxilla extending anterior to the anterior margin of the septomaxilla. Most lepidosaurs possess a small portion of premaxilla anterior to the septomaxilla, but certain forms [e.g., Huehuecuetzpalli mixtecus (fig. 20), Varanus] possess a more significant rostrum.

Figure 20

Reconstruction of the skull of Huehuecuetzpalli mixtecus mixtecus in left lateral view, modified after Reynoso (1998). Missing portions are reconstructed as semi-opaque shadows.

i0003-0090-310-1-1-f20.gif

3. Skull, muzzle shape (NG-33): (0) tapering; (1) blunt and rounded. This character describes the presence or absence of linear, subparallel, lateral surfaces of the snout in dorsal view. The absence of flat, subparallel surfaces is considered the derived state for this analysis.

4. Skull, supratemporal crest (Ga82-25): (0) absent; (1) present, forming a distinct angle between the dorsal and lateral faces of the skull (a postorbital canthal crest).

5. Skull, interorbital septum (Ke-27): (0) present; (1) absent. I treat this character exactly as in Kearney (2003a).

6. Nares, posterior elongation invading contact between prefrontal and nasals or such that they open extensively dorsally (E-2): (0) absent (fig. 21); (1) present (fig. 22B). This differs somewhat from the description given by Estes et al. (1988: character 2). They discuss only the condition of the posterior nareal border approaching the frontal. Theoretically, the frontal could approach the naris without the latter being greatly posteriorly expanded. As used here, this character also helps to identify the anterior elongation of the premaxillary process of the maxilla without overlapping with the character describing the anteroposterior placement of the nasal process of the maxilla. There are some difficulties in scoring the condition of the external naris in Heloderma (Pregill et al., 1986). Based on the work of Pregill et al. (1986) and similarities the nasal-prefrontal-maxillary morphology in Estesia mongoliensis (see Norell et al., 1992; fig. 4), each of these taxa has been scored with the plesiomorphic state.

Figure 21

Skull of Shinisaurus crocodilurus in (A) right lateral (reversed to be left lateral) view, (B) dorsal view, and (C) ventral view. (A) and (B) UF 62316, and (C) UF 62497. Modified after Conrad (2004a).

i0003-0090-310-1-1-f21.gif

Figure 22

Skulls of (A) Paravaranus angustifrons in dorsal view and Proplatynotia longirostris in (B) dorsal and (C) left lateral views. Modified after Borsuk-Białynicka (1984). Missing portions are reconstructed as semi-opaque shadows.

i0003-0090-310-1-1-f22.gif

Caldwell et al. (1995) describe apparent narial retraction as the result of snout elongation and/or topological changes in some skull roofing bones. The current character takes this into account in describing the reduction of contact between the nasals and prefrontals. Other characters described by Caldwell et al. (1995) are also accounted for in this character list (see characters relating to the nasals, prefrontals, and maxillae). Note, however, that Caldwell's (1995) characterization of the processes leading to apparent narial retraction in Varanus (in contrast to mosasauroids or other taxa with apparent narial retraction) includes morphological characteristics that are present in only some Varanus. For example, not all Varanus possess anteriorly elongate nasals (see figures in Mertens, 1942b).

7. Dermal sculpturing (E-129): (0) irregular (vermiculate); (1) pitted; (2) bumps/hornlets. Treatment of this character, again, differs somewhat from that of Estes et al. (1988). They coded for the presence or absence of vermiculate sculpturing, but here sculpturing is described with three character states and applies not only to the osteoderms, but also the dermal skull roofing bones. If there is no sculpturing whatever, then the taxon is scored “-” or “inapplicable” for this character because the three following characters (8–10) code for the presence or absence of sculpturing on various skull bones.

8. Dermal sculpturing, maxilla (CN-5): (0) absent; (1) present.

9. Dermal sculpturing, prefrontal (CN-6): (0) absent; (1) present.

10. Dermal sculpturing, parietal/frontal (E-129): (0) absent; (1) present on frontal and parietal.

11. Premaxilla, fusion into single element (LC00-1): (0) absent, paired premaxillae; (1) present. Estes et al. (1988) define this character based on ontogeny. Very little ontogenetic data are known for most fossil taxa, making this character virtually impossible to score for any taxa without fused premaxillae. Given a taxon with paired premaxillae, one might appeal to the immaturity of the specimen. I have chosen to leave out the ontogenetic component of the character description and implemented the following conventions for scoring it. When taxa appear, from other indicators, to represent adults, I have scored them for this character. When they appear to be somatically immature, but with fused premaxillae, I have scored them as possessing character state “1”. When they appear immature and possess paired premaxillae, I have refrained from scoring them and thus leave the “?” in place. Thus, the only conditions under which fossil taxa are coded as “0” for this character are when they appear to be adults and have paired premaxillae.

12. Premaxilla, mediolateral breadth of nasal process (B-4): (0) absent; (1) broad, widest plane; (2) narrow, narrowest plane; (3) narrow at its base, but spatulate posteriorly. This character is an attempt at quantifying the relative breadth of the nasal process of the premaxilla. Bell (1997) presented this character making comparisons between the internarial bar ( = nasal process) of the premaxilla and the premaxillary rostrum of mosasauroids. Here, it is somewhat reformulated to be comparable to other groups of squamates.

13. Premaxilla, external contact with the frontal(s) (LC00-2): (0) absent, (1) present, (2) contact overlain by nasals. This character is modified from that of Lee and Caldwell (2000) in that the contact is specified, here, to be external. It retains the same distribution as in the previous study, but allows for the possibility of a contact with the frontal ventrally, invisible in dorsal view.

14. Premaxilla, incisive process (GN98-46): (0) single; (1) bilobed or bipartite; (2) absent.

15. Premaxilla, rostrum anterior to the premaxillary teeth (DBC-4): (0) absent (fig. 23A, D); (1) present, conical and short (extending for about 1 tooth position) (fig. 23B, E); (2) cylindrical and elongate (extending for about the length of two tooth positions) (fig. 23C, F). This character was modified by Bell (1997) who divided it into two characters, effectively ordering the three character states above. I do not use the latter approach. The rostrum anterior to the premaxillary teeth is not necessarily homologous in states (1) and (2).

Figure 23

Line drawings and photos of selected mosasaurid premaxillae in lateral view, showing the relative lengths of the rostrum (see character 15) as illustrated by the gray arrows. (A), (D) Platecarpus tympaniticus, character state 15(0). (B), (E) Clidastes sp., character state 15(1). (C), (F) Tylosaurus proriger, character state 15(2). (A–C) redrawn after Russell (1967). Photos of (D) AMNH FR1532; (E) AMNH FR14791; and (F) AMNH FR3451 (reversal of right lateral view).

i0003-0090-310-1-1-f23.gif

16. Premaxilla, contact with maxilla (L98-3; RZ-3): (0) sutural; (1) nonsutural (fig. 24).

Figure 24

Skull and mandibles of Boa constrictor (AMNH R73614) in (A) left lateral view, (B) dorsal view, and (C) ventral view.

i0003-0090-310-1-1-f24.gif

17. Premaxilla-maxilla aperture (M70): (0) absent; (1) present (fig. 22B). This structure, a hole between the premaxilla and maxilla, has been referred to by a variety of names, including premaxillary foramen (Meszoely, 1970). Gao and Norell (1998: 44–45) favor the term used here, which helps to avoid confusion with the premaxillary foramen of lizards lacking the hole between the premaxilla and maxilla.

18. Premaxilla, contact with the nasal (FE-1): (0) premaxilla(e) overlaps the nasal(s); (1) nasal(s) overlap the premaxilla(e); (2) premaxilla(e) does not reach the nasal(s).

19. Nasals, presence as discrete elements (DBC-12): (0) present; (1) absent. This is a modification of DeBraga and Carroll (1993: character 12), Bell (1997: character 8) and Lee and Caldwell (2000: character 21). Although there is some ambiguity left in the character by the current wording, it allows for either of the two possibilities (nasals fused to the premaxilla or nasals absent) to be scored. Because it is not believed that these two possibilities co-exist as independent character states, but instead that either one or the other is the case for all taxa for which the nasals are indistinct (known only within mosasauroids), the correct identification of one or the other of these character states is not imperative.

20. Nasals (E-3): (0) paired (e.g., fig. 22B); (1) fused to one another (fig. 22A). Estes et al. (1988) define this character (like character 11) based on ontogeny. This character is treated in much the same manner as character 11 (above).

21. Nasals, internasal contact (GN98-2): (0) extensive; (1) less than one-half of their length.

22. Nasal, anterior border (Gr88-2): (0) concave, forming the posterior border of the external naris; (1) lacking anterolateral narial process. Some squamates possess an anterolateral prong of the nasal bone such that the nasal forms the entire posterior border of the external naris (state 0). Other taxa lack this process and the posterior border of the external naris has contributions from the prefrontal, frontal, or maxilla (state 1).

23. Nasofrontal suture, articulated shape in dorsal view (C06-7): (0) M-shaped (nasals forming a posterior wedge); (1) frontal forms an anterior wedge (fig. 21B); (2) transverse (fig. 24B); (3) W-shaped (fig. 22A, B). Some authors discuss the presence of a nasal shelf of the frontal as a synapomorphy of iguanians (Estes et al., 1988). However, such a shelf is present in many squamates, including all of the observed taxa having states 1 and 3 in this analysis.

24. Nasofrontal fontanelle (new/extensively modified): (0) absent (e.g., figs. 21B, 24B), (1) present (fig. 25).

Figure 25

Skull of Corytophanes cristatus (FMNH 22093) in dorsal view. Note the presence of a nasofrontal fontanelle (see character 23).

i0003-0090-310-1-1-f25.gif

25. Maxilla, anteromedial process lying between vomers and premaxillae (Egek-12): (0) absent; (1) present. Estes et al. (1988) include this character in their diagnosis of Gekkota, although it did not make it into their larger list of characters for their overall analysis.

26. Maxilla, strong medial processes posterior and posteroventral to the nasal process of the premaxilla (FE-2): (0) absent; (1) present. Although this character was not explicitly defined by Frost and Etheridge (1989), it seems a natural outgrowth of character 2 in their analysis.

27. Maxillae, contact at midline behind nasal process of premaxilla (FE-2): (0) absent; (1) present.

28. Maxilla, nasal process (RZ-5): (0) at or anterior to midpoint of maxilla (fig. 26A); (1) posterior to midpoint of maxilla (fig. 26B); (2) dorsal and ventral margins subparallel (TC00-29) (fig. 24A); (3) maxilla very short, presence or absence of a nasal process implicit. Rieppel and Zaher (2000a), believed this character to be the same as describing a presence or absence of a retracted naris. However, Dinilysia patagonica and anilioids possess a retracted external naris (see character 6), but not a posteriorly positioned maxillary nasal process. Conversely, Hemitheconyx caudicinctus, Proplatynotia longirostrata, Estesia mongoliensis (see description under character 6, above), and rhineurids possess a posterior nasal process, but not a retracted naris. Thus, this character varies independently from other characters in the analysis. State 2 was added based on descriptions in the supplementary data of Tchernov et al. (2000).

Figure 26

Some paired dermal skull bones of (A) Shinisaurus crocodilurus (UF 57112) and (B) Heloderma suspectum (AMNH R142627) in left lateral view for comparison. Note that Heloderma lacks a postorbital and has a very reduced squamosal that is not figured here.

i0003-0090-310-1-1-f26.gif

29. Maxilla, nasal process inclination (CN-17): (0) steeply inclined, posterior border of the naris distinct from ventral border (e.g., Figs. 26, 27); (1) weakly inclined, posterior border of the naris not distinct from the ventral border (no strong angle between the two faces) (figs. 28B, 29C, E). This character may, at first, seem to be correlated with a retracted naris, but this is not the case. A variety of taxa from each of the major squamate clades possess a strongly angled anterior margin of the nasal process in lateral view (e.g. Leiolepis belliana, Pygopus lepidopus, Tupinambis nigropunctatus, Cordylus polyzonus, and Pseudopus apodus). However, many taxa show much gentler slope to the anterior margin of the nasal process, including Hemitheconyx caudicinctus, Xantusia henshawi, and Gobiderma pulchrum without concomitant posterior placement of the maxillary nasal process.

Figure 27

Skull of Corytophanes cristatus (FMNH 22093) in left lateral view.

i0003-0090-310-1-1-f27.gif

Figure 28

Reconstructed skulls in left lateral view and frontals and parietals in dorsal view of two “necrosaurs”. (A) Parasaniwa wyomingensis; and (B)Necrosauruseucarinatus. Drawn after (A) Estes (1964), and (B) Kuhn (1940) and Estes (1983). Missing portions are reconstructed as semi-opaque shadows.

i0003-0090-310-1-1-f28.gif

Figure 29

Skulls of a modern gecko and two taxa traditionally considered closely related to gekkotans. (A–C) Skull of Hemitheconyx caudicinctus in ventral, dorsal, and left lateral view (with lower jaw), respectively. (D) Skull of Eichstaettisaurus schroederi in dorsal view. (E) Skull and lower jaw of Xantusia henshawi in left lateral view. Note that the ectopterygoid is visible posterior to the maxilla and ventral to the jugal in (E). (A–C) modified after Rieppel (1984a) and Maisano (2003j), (D) redrawn from Evans et al. (2000), Evans et al. (2004), and photos kindly supplied by Sterling Nesbitt, and (E) modified after Maisano (2003a).

i0003-0090-310-1-1-f29.gif

30. Maxilla, overlap of prefrontal (new/extensively modified): (0) only anteriorly; does not include a supraorbital component (fig. 30A, C, D); (1) extensive, extends beyond the lacrimal and/or lacrimal foramen (fig. 30B). In a few taxa, the maxillary overlap of the prefrontal has increased such that the former overlaps the latter in a way that it partly encircles the orbit.

Figure 30

Snouts of selected squamates in left lateral view to show relative snout lengths and details of the arrangements of snout bones. (A) Huehuecuetzpalli mixtecus mixtecus, (B) Dibamus novaeguinea, (C) Xantusia henshawi, and (D) Heloderma suspectum. Not to scale but drawn to the same approximate depth. Redrawn or modified from (A) Reynoso (1998), (B) Rieppel (1984b), (C) Maisano (2003a), and (D) Rieppel (1980a).

i0003-0090-310-1-1-f30.gif

31. Maxilla, contact with vomer posterior to the fenestra vomoeronasalis externa (E-42): (0) absent (paleochoanate condition); (1) present (neochoanate condition).

32. Maxilla, palatine flange (CN-19): (0) medial flaring absent (fig. 21C); (1) medially flared from the lateral border of the internal nares (fig. 29A); (2) present, expanded posteromedially beyond the posterolateral process of the maxilla (fig. 31C). This character addresses the palatal portion of the maxilla. Various squamates possess a medial, usually obtuse and pointed projection of the maxilla that contacts the palatine and carries the posterior part of the infraorbital canal (see Oelrich, 1956; Conrad, 2004a; Bever et al., 2005a, 2005b). Feyliniids and dibamids alone are known to have state 2 in which this palatal ramus of the maxilla extends posteriorly beyond the lateral maxillary exposure.

Figure 31

The skull of Dibamus novaeguineae in (A) left lateral view (with mandible), (B) dorsal view, and (C) ventral view. Modified primarily after Rieppel (1984b) with consideration of data from Greer (1985).

i0003-0090-310-1-1-f31.gif

33. Maxilla, posterior extent of tooth row (E-27): (0) beyond anterior one-fourth of the orbit (e.g., fig. 27); (1) terminates at anterior border of the orbit (fig. 31).

34. Prefrontal, dorsolateral tuberosity (Ga84): (0) absent; (1) present.

35. Prefrontal, supraorbital ridge (DBC-13): (0) absent; (1) present. The supraorbital ridge is a laterally projecting shelf on the dorsolateral margin of the prefrontal, extending antriorly from the orbit.

36. Prefrontal, pares frontales contact at midline (Gr88-011): (0) absent; (1) present.

37. Prefrontal, contact with postorbitofrontal (E-5): (0) absent; (1) present.

38. Prefrontal, blocks contact between maxilla and nasal (C06-10): (0) absent (e.g., fig. 29B); (1) present, extends anteriorly to reach the naris (fig. 21B); (2) present, contacts the premaxilla.

39. Prefrontal, contact with jugal (R80-24): (0) absent; (1) present.

40. Prefrontal, subpalpebral fossa (C06-11): (0) absent (fig. 26B); (1) present (fig. 26A).

41. Lacrimal (E-28): (0) present, large and extending for more than one-half the distance to the external naris; (1) present, discrete, and limited to orbital margin (fig. 30D); (2) present on orbital margin, but fused to the prefrontal (fig. 30A, C); (3) absent (figs. 24A, 30B); (4) present, but reduced to a nubbin that is supported by soft tissue and fails to contact the prefrontal. Within Rhynchocephalia, Gephyrosaurus possesses a distinct lacrimal (Evans, 1980) (fig. 32A), but most sphenodontidans including Diphydontosaurus (Whiteside, 1986), Planocephalosaurus (Fraser, 1982), Paleopleurosaurus (Carroll, 1985), and Sphenodon, possess a fused lacrimal-prefrontal. Character state (0) is included here based on previous ideas about the close relationship of Kuehneosauridae with squamates (see Romer, 1956, 1966; Robinson, 1967; Estes, 1983; Gauthier, 1984; Estes et al., 1988; Gauthier et al., 1988a). This relationship has been recently challenged (Müller, 2003, 2004b). The character state is retained for easy inclusion of kuehneosaurids or the “paliguanid” Paliguana whitei* (see Carroll, 1975, 1977, 1988a; Gauthier, 1984) in later analyses. State 4 is difficult to assess in some cases. I have observed it only in Hoplocercus spinosus and in Enyalioides and it is clear that the very reduced lacrimal might easily be overlooked or lost because it was held in place almost exclusively by soft tissue. High-resolution x-ray tomography might serve as a tool to search for a vestigial/rudimentary lacrimal in various taxa apparently possessing state 3 of this character just as it has in the identification of the palpebral in Lanthanotus borneensis (Maisano, 2001b; Maisano et al., 2002) and the confirmation of its absence in Heloderma (see Bonine, 2005).

Figure 32

Skull of a basal rhynchocephalian, Gephyrosaurus bridensis, in (A) left lateral, (B) dorsal, and (C) ventral views. All redrawn and modified after Evans (1980). Palate modified after the individual elements drawn in Evans (1980), not the composite reconstruction. Missing portions are reconstructed as semi-opaque shadows.

i0003-0090-310-1-1-f32.gif

42. Lacrimal, posterolateral flange (new/extensively modified): (0) absent; (1) present.

43. Lacrimal, foramen (RZ-10): (0) single; (1) double.

44. Lacrimal foramen, size (FE-6): (0) small; similar in size to palatine foramen; (1) large; distinctly larger than the palatine foramen.

45. Jugal (RZ-12): (0) present; (1) absent (figs. 30B, 31).

46. Jugal, anterior extension (LC00-18): (0) no further than if forming the anterior border of the orbit; (1) extends anteriorly beyond the margin of the orbit and not contiguous with the prefrontal and/or lacrimal suture.

47. Jugal, shape (GN98-13): (0) angulated (e.g., fig. 21B); (1) little angulation; curved (figs. 20, 29E); (2) reduced to a small splint barely extending beyond the posterior margin of the maxilla (fig. 29C).

48. Jugal, posteroventral process (GN98-14): (0) present (fig. 26); (1) absent (fig. 29E).

49. Jugal, postorbital branch (Ga82-27): (0) without anterior or posterior flanges (fig. 26); (1) dilated (fig. 27).

50. Jugal, postorbital process rugosities (GN98-17): (0) absent; (1) present.

51. Jugal, contact with the postorbitofrontal (E-32): (0) present; (1) absent.

52. Jugal, relationship to maxilla (C06-14): (0) mostly medially; (1) mostly dorsally; (2) jugal reduced and lying mostly posterior to the maxilla. In some taxa, the posteroventral part of the maxilla laterally overlies the jugal (state 0). In others, the jugal lies mostly dorsal to the posteromedial part of the maxilla, the latter being mediolaterally broadened (state 1; fig. 26). The anterior (suborbital) ramus of the jugal is reduced in some taxa such that the bone lies mostly posterior to the maxilla and has limited dorsal or medial overlap with it.

53. Jugal-squamosal contact (E-18): (0) absent (fig. 29E); (1) present (fig. 27). Rhynchocephalians are coded as apomorphic for this condition even though Sphenodon does not possess this contact. The relatively basal rhynchocephalians Gephyrosaurus, Planocephalosaurus, Clevosaurus, Paleopleurosaurus, and Pleurosaurus show state 1.

54. Quadratojugal (new/extensively modified): (0) present; (1) absent.

55. Frontals (E-6): (0) separate in adults (figs. 22B, 31B); (1) fused in adults (figs. 22A, 32B, 33, 34). Basal rhynchocephalians (Gephyrosaurus, Diphydontosaurus, Planocephalosaurus) possess fused frontals (Evans, 1980; Fraser, 1982; Whiteside, 1986) (fig. 32B). Paleopleurosaurus appears to be the most basal rhynchocephalian possessing paired frontals (Carroll and Wild, 1994) like those of Sphenodon.

Figure 33

Ventral views of the frontal (top) and parietal (bottom) of Shinisaurus crocodilurus (UF 57112).

i0003-0090-310-1-1-f33.gif

Figure 34

Dorsal view of part of an articulated skull of Mosasaurus hoffmanni (drawn after Lingham-Soliar, 1995).

i0003-0090-310-1-1-f34.gif

56. Frontal, anterior constricted neck (DBC-19): (0) absent; (1) present.

57. Frontals, shape as a unit (CN-33): (0) anterior and posterior borders subequal in width (figs. 31B, 32B); (1) rhomboid (fig. 34); (2) concave lateral margins, minimum width less than three-fifths of the posterior border width (fig. 33); (3) tapering posteriorly. The condition “triangular” in Lowesaurus matthewi (Pregill et al., 1986) is questionable because of breakage; scored trapezoidal here. State 2 is not identical to the derived state of character 58. This character refers to the shape of the frontal unit as a whole; specifically comparing the anterior and posterior widths of the unit. Shinisaurus crocodilurus possesses concave lateral margins on the frontal (character 57, state 2), but because the frontal does not expand mediolaterally anterior to this lateral margin, it is also scored as state (0) for character 58. Other taxa share this combination of character states.

58. Frontals, constriction between orbits (E-7): (0) absent, interorbital margin linear; (1) present, anterior portion of the frontal is hourglass shaped.

59. Frontal, dorsal keel (B-12): (0) absent (fig. 29); (1) low, weakly developed (fig. 34); (2) tall, well developed.

60. Frontal, dorsoventral inflation (CRG-36): (0) absent; (1) present. Some taxa (e.g., Anolis carolinensis; see The Deep Scaly Project, 2006) possess dorsoventrally inflated frontals with large internal cavities. These cavities are entirely within the frontal and do not form canals between the frontals and other bones, nor do they house the olfactory tract.

61. Frontals, subolfactory processes (E-10): (0) ventral downgrowths; (1) partly surrounding the olfactory tracts; (2) contact the parasphenoid (RZ-54). State (1) describes the ventromedial growth of the subolfactory processes wherein they grow toward, and may approach, one another. State (2) addresses a condition wherein the frontal partly underlies the olfactory tract, but the processes contact part of the braincase (the parasphenoid portion of the sphenoid). Taxa with state (2) cannot be scored for character 62 (below) and are coded as (-) for that character.

62. Frontals, subolfactory processes contact at midline (E-10): (0) absent; (1) present (fig. 29A).

63. Frontal, medial pillar separating the olfactory tracts (Rs80): (0) absent, (1) present.

64. Frontals, contact between the medial pillar and lateral subolfactory flanges (Rs80): (0) absent, (1) present.

65. Frontals, contact the maxilla anteriorly (E-4): (0) absent (figs. 21B, 24B); (1) present (figs. 22B, 29B, 31B).

66. Frontals, participation in the orbitonasal foramen (C99-7): (0) absent, prefrontals with large contributions; (1) present, prefrontals largely blocked from the orbitonasal fenestra. Note that character state (1) is not redundant with the presence of a frontal-palatine contact (see below). State (1) suggests the presence of strong descending processes of the frontals along the medial surfaces of the prefrontals that may or may not co-occur with the frontal-palatine contact.

67. Frontal, contact with palatines (Ga82-82): (0) absent; (1) present.

68. Frontal, invaded by external nares (B-5): (0) absent; (1) present. This is a further transformation of character 6.

69. Frontals, parietal tabs (C99-9): (0) absent; (1) present (fig. 33); (2) present, elaborated into dorsomedial extensions on top of the parietals (fig. 34).

70. Frontoparietal suture, dorsal view (EB98-131): (0) U-shaped, anteriorly arched; (1) transverse; (2) W-shaped; (3) U-shaped, posteriorly arched. Most other rhynchocephalians possess a U- or even W-shaped frontoparietal suture.

71. Frontoparietal fontanelle (new/extensively modified): (0) absent; (1) present.

72. Parietal, lateral flange at the frontoparietal suture (new/extensively modified): (0) gently curved, laterally tapering; (1) with broad, squared, lateral tabs such that the postfrontal margin of the frontal is parallel with the postfrontal margin of the parietal.

73. Parietals (E-21): (0) paired (fig. 29D); (1) fused (fig. 29B). Rhynchocephalians appear to have fused parietals primitively. This is the case for Gephyrosaurus (Evans, 1980) (fig. 32B), Planocephalosaurus (Fraser, 1982), and Paleopleurosaurus (Carroll and Wild, 1994). Diphydontosaurus has been reported to possess fused parietals (Whiteside, 1986), although the same author illustrated a midline suture. Regardless, rhynchocephalians are scored with state (1).

74. Parietal, frontal tabs (C99-17): (0) absent; (1) present within the contact and visible dorsally; (2) present on the ventral surface.

75. Parietal, median adductor crest expressed as a keel (FE-10): (0) absent; flat parietal table extends to the posterior margin; (1) present. Taxa in which the jaw adductors originate from the ventral surface of the parietals cannot be scored for this character because of redundancy with character 86 (see below). Note that this median crest is not homologous with the sagittal crest seen in some taxa wherein the jaw adductors do not contribute to the crest.

76. Parietal, decensus parietalis (E-23): (0) weakly developed/absent (fig. 29C); (1) present as anteroposteriorly elongate crest (fig. 33); (2) present, anteroposteriorly narrow ventral projection (figs. 29E, 35, 36A).

Figure 35

Left lateral views of the skulls of (A) Myrmecodaptria microphagosa; and (B) Globaura venusta. Modified after (A) Gao and Norell (2000) and (B) Borsuk-Białynicka (1988).

i0003-0090-310-1-1-f35.gif

Figure 36

Skull of Scincus scincus (AMNH R2245) in (A) left lateral view (with mandible), (B) dorsal view, (C) ventral view, and (D) mandible in medial view. Note the presence of a partly developed secondary palate [character state 114(1)].

i0003-0090-310-1-1-f36.gif

77. Parietal, pineal foramen (E-26): (0) within parietal; (1) within frontal; (2) at frontoparietal suture; (3) absent.

78. Parietal fossa, posterior margin (CN-46): (0) open, crests extend posterolaterally; (1) closed, crests meet at midline; (2) absent. The parietal fossa is a dorsal concavity on the ventral surface of the parietal occurring posterior to the pineal foramen (when the latter is present) and is flanked by the cristae postfovealis. In some taxa (e.g., Xenosaurus, Glyptosaurus) these crests extend posteromedially to and contact one another at midline posterior to the parietal fossa.

79. Parietal, posterior flange (not associated with a sagittal, jaw adductor, crest) (CN-47): (0) absent; (1) present.

80. Parietal, supratemporal processes length from the level of the parietal notch compared to the parietal anterior to that point (Ga82-30): (0) greater than one-half (fig. 28B); (1) less than one-half (fig. 29B); (2) absent (fig. 24).

81. Parietal, dorsal margin of the supratemporal process (GN98-33): (0) narrow and bladelike; (1) broad and flat.

82. Parietal, transverse posterior margin between the supratemporal processes (CN-50): (0) present (fig. 36B); (1) absent (fig. 33).

83. Parietal, nuchal fossa (GN00): (0) absent (fig. 22B); (1) present, visible in dorsal view (fig. 22A); (2) present and extending substantially onto the skull table (fig. 36B).

84. Parietal, contact with supratemporal arch (C99-15): (0) only at the anterior and posterior extremes (figs. 21B, 32B); (1) increased contact anteriorly and posteriorly (fig. 36B).

85. Parietal, contact with the supraoccipital (TC00-37): (0) no bony contact (contact only via the processus ascendens tecti synotici) (fig. 36B); (1) bony contact present and extensive; the supraoccipital becomes incorporated into the skull roof (figs. 24B, 31).

86. Parietal, origin of the jaw adductor musculature (E-54): (0) dorsally (figs. 21, 22, 24); (1) ventrally (fig. 34B, D, 36B).

87. Supratemporal (Egek-10): (0) present (figs. 21B, 24B, 36); (1) absent (figs. 29B, 31).

88. Supratemporal, length relative to depth (extensively modified after Ga82-86; CN-55): (0) less than 2.5 times as long as deep; (1) more than 3 times as long as deep. The arbitrary numbers implemented here are used as an alternative to the even more arbitrary descriptions sometimes used, including “elongate” and “short”.

89. Postfrontal/postorbital, forking of medial surface (RZ-25): (0) absent; (1) present.

90. Postorbital/postfrontal tubercle (McG-7): (0) absent (figs. 25, 29D); (1) present. In some taxa, an anterodorsal tuberosity is present near the postorbital-postfrontal contact at the posterolateral margin of the orbit. This tubercle marks a point of strong connection between the integument and skull bones during dissection.

91. Postfrontal (E-12): (0) present; (1) absent.

92. Postfrontal shape (new/extensively modified): (0) anteroposteriorly elongate (fig. 29B); (1) irregularly shaped, not elongate in mediolateral or anteroposterior planes (fig. 25); (2) mediolaterally developed bar bordering the orbit and supratemporal fenestra (fig. 22A). Note that state (1) appears primarily in iguanians and has not yet been observed in taxa lacking a postorbital.

93. Postfrontal, contact with the parietal (E-15): (0) absent; (1) present; (2) present for more than one-half the parietal table length. This ordered character addresses the presence or absence of a postfrontal-parietal contact and amount of contact between these bones.

94. Postorbitofrontal, fusion (E-14): (0) absent; postorbital and postfrontal exist as distinct elements (figs. 22, 25); (1) present (figs. 21B, 34).

95. Postorbital (may be fused with the postfrontal) (E-16): (0) present (figs. 21, 35B); (1) absent (figs. 35A, 36).

96. Postorbital, posterior extent (E83): (0) less than one-half the length of the supratemporal fenestra; (1) more than one-half the length of the supratemporal fenestra; (2) more than ¾ the length of the supratemporal fenestra; (3) contacts the supratemporal, partly or completely blocking the squamosal from contacting the supratemporal fenestra.

97. Postorbital, contribution to the postorbital bar (E-17): (0) one-half or more; (1) less than one-half.

98. Squamosal (RZ-40): (0) present; (1) absent.

99. Squamosal, contact with postorbitofrontal (or postorbital or postfrontal)—completion of the supratemporal arch (RZ-38): (0) present (e.g., fig. 29D); (1) absent (fig. 29B, C).

100. Squamosal, dorsal process (C99-25): (0) present; (1) absent.

101. Palpebral bone (E-36): (0) absent; (1) present, a single ossification articulating with or located near the prefrontal.

102. Septomaxilla (Echam-12): (0) present; (1) absent.

103. Septomaxilla, medial flange (RZ-87): (0) short/absent; (1) long.

104. Septomaxilla, contact with the osseous nasal cavity roof (FE-54): (0) absent; (1) present.

105. Palate, orientation of the ectopterygoid (CN-68): (0) mostly mediolaterally; (1) oriented anterolaterally (at more than 30 degrees from perpendicular to sagittal for the skull). Future work should address the utility of this character and the way it is scored in taxa for which the maxillary tooth row terminates anterior to the orbit (e.g., Varanus).

106. Vomers, fusion: (0) absent (fig. 37B); (1) present, intervomerine suture lost (fig. 37A).

Figure 37

(A) Skull of Uromastyx sp. (AMNH R73350) in ventral view; note the fusion of the vomers and broad interpertygoid vacuity [character states 106(1) and 123(0), respectively]. (B) Posterolateral view through the right orbit of Uromastyx sp. (AMNH 73350) showing the midline contact of the palatal bones and the absence of a secondary palate [character state 114(0)]. (C) Skull of Lacerta viridis (AMNH R1148) in ventral view. Note the absence of a distinct palatine flange of the maxilla and narrow pyriform recess [character states 32(0) and 123(1), respectively]. (D) Posterolateral view through lateral temporal vacuity of Tiliqua nigrolutea (AMNH R99684) with the presence of a secondary palate indicated by an arrow [character state 114(2)]. Note that (B) and (D) are not to scale.

i0003-0090-310-1-1-f37.gif

107. Vomer, shape (RZ-93): (0) platelike (broad, flat); (1) rodlike (narrow and sub-cylindrical).

108. Vomer, articulation with the palatine (R84): (0) relatively broad (subequal in breadth to the contact with the maxilla) (figs. 32C, 37C); (1) relatively narrow (about one-half the breadth of the contact with the maxilla) and movable (e.g., fig. 29A); (2) absent (TR-47) (fig. 24C).

109. Vomerine teeth (Ga88-22): (0) present; (1) absent.

110. Palatines, medial expansion anteriorly (Ga82-84): (0) absent; (1) present. As used here, this character is meant to describe the anteromedially oriented margin of the palatine in some taxa. This results in an anterior constriction (or even closure) of the pyriform recess in some taxa. Estes et al. (1988) used a similarly worded character (E-43) to describe a secondary palate. The secondary palate is treated separately in the present analysis (see character 114 below).

111. Palatine, length (GN98-51): (0) longer than wide; (1) subequal in length and width; (2) deeper than long.

112. Palatine, length relative to the vomer (L98-98): (0) subequal; (1) only two-thirds the length. Character state (1) has been modified from Lee (1998).

113. Palatines, choanal groove (RZ-101): (0) very short/absent; (1) distinct, elongate.

114. Palatine, secondary palate formed around choanal groove (E-43): (0) absent (fig. 21C); (1) present, ventromedial fold partly hides the choanal groove (fig. 36C); (2) present, ventromedial processes hide most or all of the dorsomedial processes (fig. 31C). The character states here have been modified to help identify the degree of the secondary palate formation. Of course, the cutoff points for the individual character states are somewhat arbitrary, but they are descriptive.

115. Palatine, teeth (E-82): (0) present, patches; (1) absent; (2) present, single line. Although some authors may treat this as two characters or as an ordered character to help emphasize the difference between the presence and absence of teeth, it is not clear that it is a larger evolutionary change to go from patches of teeth to teeth being absent than from patches of teeth to a single line. Consequently, this character is treated as unordered, as is character 118 addressing pterygoid teeth. Lepidosaurs apparently possess patches of palatine teeth plesiomorphically (Evans, 1980; Estes, 1983; Whiteside, 1986; Estes et al., 1988; Fraser, 1988; Gauthier et al., 1988a; Wilkinson and Benton, 1996). Note that studies by Mahler and Kearney (2006a, 2006b) found evidence for non-independence of palatal teeth on different palatal bones. However, this character (115) and character 118 do not co-vary in every circumstance, so they are treated separately.

116. Pterygoid, contact with jugal (GN98-54): (0) absent; (1) present.

117. Pterygoid, ventromedial process (GN98-32): (0) absent; (1) present.

118. Pterygoid, teeth (E-83): (0) arranged in multiple rows or patches (fig. 36C); (1) in a single line; (2) absent (fig. 21C).

119. Pterygoid, contact with vomer: (0) present (fig. 32C); (1) absent (fig. 29A).

120. Pyriform recess, midline contact of vomers (C06-22): (0) present, invaded by pyriform recess (fig. 21C); (1) present, contact for their length (figs. 24C, 34C); (2) present, contact anteriorly and posteriorly) (fig. 37A); (3) absent. This and the following two characters are used to define the forward extent of the pyriform recess.

121. Pyriform recess, midline contact of palatines (CN-81): (0) absent; (1) present.

122. Pyriform recess, midline contact of pterygoids (new/extensively modified): (0) absent; (1) present. Scoring of Saniwa ensidens is based on Gilmore's (1928: 61) discussion of skull distortion and his plate 4.

123. Pyriform recess, broadest point compared to the distance from basicranial joint to quadrate (CN-82): (0) greater than one-half; (1) less than one-half.

124. Ectopterygoid, contact with the palatine anterior to the suborbital fenestra (E-45): (0) absent (fig. 38C); (1) present (fig. 38A); (2) present, contact broader than suborbital fenestra (fig. 38B); (3) present, closes suborbital fenestra (Ke-99). This character is modified to accommodate the special morphology of dibamids and some skinks or skink-like taxa.

Figure 38

Ventral view of the anterior part of the skull in (A) Gekko gecko (AMNH R141109), (B) Heloderma suspectum (AMNH R74778), and (C) Shinisaurus crocodilurus (UF 62497). (D) Posteroventral of the skull of Heloderma horridum (FMNH 98468) highlighting the pronounced increase in tooth length between the premaxilla and maxilla. (E) Medial view of the right mandible of Heloderma suspectum (AMNH R74778).

i0003-0090-310-1-1-f38.gif

125. Ectopterygoid, contact with the palatine posterior to the suborbital fenestra (new/extensively modified): (0) absent; (1) present (fig. 31C).

126. Ectopterygoid, lateral exposure posterior to the maxilla (MB54): (0) absent (e.g., fig. 29C); (1) present (e.g., figs. 29E, 36A).

127. Braincase, ventral sagittal ridge or crest on the sphenoid and basioccipital (TC00-77): (0) absent; (1) present.

128. Braincase, spheno-occipital epiphyses (NG-27): (0) absent; (1) present. Kearney (2003a: 32) gives a summary of the treatment of these elements in the literature. This character should be further examined (perhaps in a developmental framework) to determine its exact utility.

129. Braincase, fenestra ovalus location (NG-25): (0) above/slightly posterior to the spheno-occipital tubercle; (1) anterior to the spheno-occipital tubercle.

130. Braincase, azygous orbitosphenoid (Ke-105): (0) absent; (1) present. Originally, state (1) of this character was considered an amphisbaenian synapomorphy (Kearney, 2001, 2003a). However, new morphological studies suggest a narrower distribution for it (Kearney et al., 2005).

131. Braincase, ossified part of the occipital condyle (Egek-9): (0) single unit made of basioccipital and otooccipitals; ovoid or subovoid (e.g., figs. 36, 39); (1) bipartite, constructed primarily by otooccipitals with little contribution from the basioccipital (fig. 29A); (2) formed only by the basioccipital (GN00). Although Kluge (1987) takes issue with this character, the phenotypes for states (0) and (1) are readily observed in exemplars of skeletonized and CT-scanned squamate samples (see Maisano, 2003k; Kley, 2004; and Conrad and Norell, 2006a).

Figure 39

The braincases in two autarchoglossans. (A) Posterolateral view of the posterior part of the skull in Lacerta viridis (AMNH R1148) in left posterolateral view. (B) Posteroventral view of the braincase in Varanus sp. (AMNH, uncatalogued specimen). Note the absence of an enlarged crista tuberalis in (A) [character state 149(0)] and its presence in (B) [character state 149(1)].

i0003-0090-310-1-1-f39.gif

132. Braincase, anterior extension of crista prootica (E-52): (0) terminates on or just ventral to the inferior process; (1) extends onto the basipterygoid process; (2) crista prootica absent. This character is treated as unordered because there is no clear nested set of primary homologies inherent to the character and because it is possible to hypothesize a direct transformation from state (0) to state (2).

133. Braincase closure (Ke-29): (0) open (e.g., fig. 36); (1) parietal downgrowths and anterior extensions of prootics (fig. 40); (2) parietal downgrowths (figs. 24A, 31A).

Figure 40

Skull of Anguis fragilis in left lateral view. Compare the closure of the braincase with that of Dibamus novaeguineae (fig. 31), which lacks the anterior projections of the prootics. Modified after Rieppel (1980a).

i0003-0090-310-1-1-f40.gif

134. Epipterygoid (E-47): (0) present; (1) absent.

135. Supraoccipital, processus ascendens tecti synotici (Egek-11): (0) present; (1) absent.

136. Prootic, supratrigeminal process (E-50): (0) absent or faint ridge; (1) distinct, anterior process visible in lateral view; (2) extensive, with a downgrowth that closes the trigeminal foramen.

137. Prootic, crista alaris (Ke-79): (0) absent; (1) present, short (dorsoventral depth greater than anteroposterior length); (2) present, elongate (dorsoventral depth less than anteroposterior length).

138. Prootic, crista prootica (RZ-66): (0) well-developed; laterally and ventrally projecting parts; (1) reduced; extending mostly laterally; (2) absent.

139. Prootic, perforation of the crista prootica (CN-91): (0) absent; (1) present. Although this character state had been illustrated previously (e.g., Rieppel, 1984a; Grismer, 1988), it was not used in a phylogenetic analysis until very recently (Conrad and Norell, 2006a). This perforation carries a branch of the trigeminal nerve.

140. Prootic, entocarotid fossa (GN98-30): (0) present as distinct a fossa within the recessus vena jugularis; (1) reduced/absent. As defined here, this refers to a depression on the lateral surface of the braincase posterior to the Vidian canal as opposed to the carotid fossa (below), which is located on the anterior surface of the sphenoid.

141. Prootic, external facial foramen (NG-20): (0) single; (1) double (presence of an external facialis canal). Conrad (2004a) and Bever et al. (2005a) term this character differently for some taxa, including Shinisaurus crocodilurus. Bever et al. (2005a) point out that the double external opening of the facial foramen in some specimens of Shinisaurus and some Varanus exanthematicus is the result of the fusion between a ventral flange of the crista prootica with a lateral extension of the prootic; the medial opening of the facial foramen is single. Thus, they argue that there is not a true bifurcation of the facial foramen, but rather that a superficial canal is formed near the external surface of the braincase (Bever et al., 2005a: 15–16). Certainly, the account of the morphology offered by Bever et al. (2005a) is accurate and I accept this interpretation of the morphology. Regardless of whether this morphology is termed a “double facialis foramen” (Rieppel and Zaher, 2000b: 504), a bifurcated facial foramen (Conrad, 2004a), or a facialis canal (Bever et al., 2005a), this character is coded here based on the apparent phenotype of the external opening of the facial foramen.

142. Sphenoid, carotid fossa (NG-22): (0) present; (1) absent. Because the basisphenoid and parasphenoid usually fuse in squamates, that compound structure is here referred to as the sphenoid, following some recent studies (Bever et al., 2005a; Conrad and Norell, 2006a).

143. Sphenoid, posterolateral ventral flanges laterally overlying basioccipital (BB): (0) absent (fig. 31C); (1) present (fig. 36C); (2) fusion of basioccipital to the sphenoid. Many squamates show a condition in which tapering processes of the sphenoid extends posterolaterally along the ventral surface of the braincase and ventrally overlies the basioccipital. These processes of the sphenoid often extend onto the spheno-occipital tubercles. The presence or absence of these processes cannot be determined in taxa wherein the sphenoid and basioccipital fuse. This character possesses no clear set of nested homologies and is left unordered.

144. Sphenoid, enclosure of the lateral head vein (E-52): (0) absent/incomplete; (1) present. Here, character 52 (E-52) of Estes et al. (1988) is broken up into two characters. This character (the current study's character 144) addresses only the actual encircling of the lateral head vein by the sphenoid as discussed by Conrad and Norell (2006a).

145. Sphenoid, anterior opening of Vidian canal (NG-18): (0) ventral to dorsum sella; (1) in the floor of the braincase, dorsal to the dorsum sella. The derived state (1) has been described and illustrated by Rieppel (1978).

146. Sphenoid, relationship with the posterior opening of the Vidian canal (E-53): (0) houses it; (1) shares it with the prootic; (2) posterior opening of the Vidian canal occurs within the prootic; (3) sphenoid and parietal share the posterior opening of the Vidian canal.

147. Parasphenoid, teeth (GA88-31): (0) present; (1) absent. This character is informative only as characterizing lepidosaurs; that is, with the inclusion of kuehneosaurids (see appendix 1)

148. Basipterygoid processes, length (RZ-74): (0) long, extending beyond main body of the sphenoid; (1) short, expressed as short nubbins that are more than 2 times as wide as long and not extending anterior to the main body of the sphenoid; (2) absent.

149. Basioccipital, crista tuberalis development (RZ-79): (0) medially concave (fig. 39A); (1) medially flat, inclusion in the paroccipital process more lateral than on the spheno-occipital tubercle (fig. 39B). This character describes the “webbing” of bone that extends between the ventrolateral margins of the braincase posteriorly and the paroccipital processes of the otooccipital. In some cases, this bony lamina may be extensive such that its lateral margin is more or less linear and extends diagonally from the basioccipital to the paroccipital processes.

150. Basioccipital, spheno-occipital tubercle (NG-23): (0) short and ventrally directed; (1) elongate and posterolaterally directed.

151. Basioccipital, location of the spheno-occipital tubercle (RZ-76): (0) posteriorly, crista tuberalis nearly vertical; (1) anteriorly, crista tuberalis posterodorsally inclined.

152. Basioccipital, canal or groove for basalar artery (DBC-34): (0) absent; (1) present.

153. Otooccipital, hypoglossal foramen (L98-69): (0) separated from vagus foramen; (1) both the hypoglossal and vagus nerves passing through internally subdivided canal or completely confluent. Lee (1997, 1998) and Rieppel and Zaher (2000a) left a small chance of confusion about the scoring of this character in their version of state (1) for this character. Lee described it as the condition wherein the hypoglossal foramen is “very close to or confluent with [the] jugular foramen on [the] external surface of the braincase,” (Lee, 1998: 392). “Very close” is open to interpretation. The presence or absence of confluence may even be problematic in the case of this character because the hypoglossal and vagus nerves often share a canal that is further subdivided for these nerves. Thus, the wording of state (1) as presented here.

154. Otooccipital, closure of the occipital recess (RZ-70): (0) open; (1) closed. Rieppel and Zaher (2000a; 2000b) offer discussions of this feature and related surrounding structures (e.g., the crista circumfenestralis of snakes), clarifying problematic areas of primary homology through detailed anatomical descriptions.

155. Otooccipital, ventral view of the occipital recess (CN-98): (0) hidden by spheno-occipital tubercle in ventral view; (1) visible in ventral view. This character is designed to further describe the relative development and orientation of the spheno-occipital tubercle.

156. Stapes, internal (quadrate) process lost (E-141): (0) present; (1) absent.

157. Stapes, shape of shaft (TC00-60): (0) straight; (1) angulated.

158. Extracolumella, anterior elongation (Ke-82): (0) absent; (1) present.

159. Quadrate, suspension (RZ-49): (0) monimostylic (fig. 32A); (1) streptostylic and supported by the squamosal, supratemporal, and paroccipital process of the otooccipital (e.g., fig. 36A); (2) suspended mainly from supratemporal (figs. 24, 40); (3) suspended mainly from otooccipital (fig. 31); (4) suspended mainly from squamosal. This character was discussed at some length by Rieppel and Zaher (2000a) who further listed historical references regarding this morphological area in various squamates.

160. Quadrate, pterygoid lappet (E-37): (0) present; (1) absent.

161. Stapes, position of stapedial artery (E-145): (0) anteriorly; (1) pierces stapes; (2) posteriorly. This character is scored based primarily on data in Estes et al. (1988). The perforate stapes in Kuehneosauridae (Evans, 1980; Gauthier et al., 1988a) indicates character state 1 for that taxon.

162. Quadrate, tympanic crest (RZ-51): (0) greater than or equal to the length of the posterior crest of the quadrate; (1) shorter than posterior crest of the quadrate at the dorsal head; (2) tympanic crest absent. The wording of this character as presented here is meant to reduce ambiguity. Essentially, the quadrate crest is broad/extensive (0), reduced (1), or absent (2).

163. Quadrate, suprastapedial process (DBC-39): (0) absent (e.g., figs. 36, 40); (1) present (fig. 41).

Figure 41

Skull of Mosasaurus hoffmanni, the largest known squamate, in left lateral view. Modified after Lingham-Soliar (1995).

i0003-0090-310-1-1-f41.gif

164. Quadrate, infrastapedial arch (DBC-41, 43): (0) absent (e.g., figs. 36, 40); (1) present (fig. 41); (2) present, contacts the suprastapedial process.

165. Extracolumellar tissue, calcification (LC00-65): (0) absent; (1) present. Coding of this character is somewhat tentative, although it appears that no extant squamates possess a calcified portion of the extracolumellar tissue and only Aigialosaurus dalmaticus and Platecarpus tympaniticus have been reported as possessing such among fossil taxa.

166. Mandible; fusion of articular, prearticular, and surangular (RZ-129): (0) absent (e.g., fig. 35); (1) present (articular-prearticular-surangular as a single unit) (fig. 36A, D). Although it may be at least partly related to ontogeny, this character is retained and it was scored as follows: If adults and/or late juveniles (staged based on other morphological indicators) lack fusion of the articular, prearticular, and surangular as observed and/or reported in the literature, they are coded with state (0); state (1) is coded for any taxon showing fusion of these elements. Gephyrosaurus (Evans, 1980) shows the unfused condition, but Diphydontosaurus (Whiteside, 1982) and other basal sphenodontidans possess the derived condition.

167. Mandible, symphysis (RZ-110): (0) present; (1) absent (fig. 24C). In extant squamates, the absence of a bony symphysis between the dentaries is accompanied by an elastic, soft-tissue connection between them.

168. Mandible, adductor fossa orientation (RZ-137): (0) medial margin low and rounded; (1) distinct vertical flange.

169. Mandible, adductor fossa expansion (E-81): (0) absent; (1) present.

170. Mandible, intramandibular septum (L98-116): (0) absent; (1) present. This character is modified from the form in Lee (1998) and Rieppel and Zaher (2000a), each of which described the posterior extent of the intramandibular septum.

171. Mandible, intramandibular septum ventral margin (M70): (0) absent; (1) posteroventral margin sutured; (2) posteroventral margin free. Lowesaurus matthewi is scored based on the description and figures in Yatkola (1976).

172. Mandible, anterior surangular foramen (CN-109): (0) present; (1) absent.

173. Mandible, external border of the anterior surangular foramen (CN-110): (0) formed only by the surangular; (1) margin with coronoid contribution; (2) with dentary contribution; (3) with coronoid and dentary contribution; (4) anterior surangular foramen absent. Because there is no clear set of nested homology statements for this character, it is treated as unordered.

174. Mandible, groove associated with anterior surangular foramen (CN-111): (0) absent; (1) present.

175. Mandible, posterior mylohyoid foramen (CN-112): (0) present; (1) absent.

176. Mandible, position of posterior mylohyoid foramen (FE-24): (0) anterior to the coronoid apex (fig. 38D); (1) posterior to the coronoid apex.

177. Mandible, glenoid (DBC-54): (0) formed at least primarily by articular; (1) formed equally by articular and surangular. This character cannot be scored for taxa in which the articular-prearticular and surangular are fused.

178. Dentary, shape of long axis (DBC-67): (0) ventrally convex (e.g., fig. 21A); (1) straight (e.g., fig. 36A).

179. Dentary, anteroventral surface (Ga82-59): (0) narrow, depth greater than width; (1) broader than tall with splenial and Meckel's canal slightly visible laterally.

180. Dentary, posterior extent (EB98-213): (0) to the level of the posterior margin of the coronoid process (eminence) (e.g., fig. 36A); (1) extends to or beyond the midpoint of mandible between the coronoid eminence and the articular condyle (fig. 31A).

181. Dentary, Meckel's canal (E-55): (0) open; (1) partly closed; (2) closed and fused.

182. Dentary, subdental shelf (E-58-59): (0) present; (1) absent; (2) present, enlarged (E-58). The subdental shelf is a lingual extension of the dentary originating from the area on which the dentary teeth attach. Estes et al. (1988) note that the subdental shelf is present in Gephyrosaurus, but suggest that other rhynchocephalians lack it. However, descriptions and figures in Whiteside (1986) suggest that a shelf is also present in Diphydontosaurus.

183. Dentary, contribution to the anterior inferior alveolar foramen (E1964): (0) dentary does not contribute; (1) dentary contributes to dorsal border; (2) dentary forms anterior and dorsal border; (3) discrete foramen absent. Eurheloderma gallicum is scored from figure 3B in Hoffstetter (1957).

184. Dentary, posterodorsal coronoid process(es) (RZ-113): (0) large and extensively overlying the coronoid eminence of the coronoid (fig. 31A); (1) small (not approaching the dorsal terminus of the coronoid eminence nor significantly overlapping it) (fig. 36A); (2) absent (fig. 35A).

185. Dentary, angular and surangular processes (E-63): (0) absent (fig. 29E); (1) present and distinct (fig. 29C). This is modified from Estes et al. (1988) in which a third character state (notches present, reduced) was included. Here this has been treated strictly as a binary character. A wavy suture is coded as (0); only the presence of distinct notches is considered to represent state (1). The structure of the intramandibular hinge and anterior end of the surangular in mosasauroids and their closest relatives makes this character inapplicable to that group (see character 191). The compound bone in snakes tapers anteriorly and helps to complete the hinge mechanism laterally.

186. Dentary, angular process compared to surangular process (Ga82-41): (0) angular and surangular processes terminate at about the same posterior level; (1) angular process terminates more anteriorly; (2) angular process extends more posteriorly (figs. 29C, 36A).

187. Dentary, principle support (RZ-119): (0) coronoid, surangular, and prearticular; (1) prearticular; (2) surangular.

188. Splenial (Ke-124): (0) present, discrete; (1) absent; (2) present, fused to the postdentary bones (Gr85).

189. Splenial, extent of anteromedial walling of Meckel's canal (E-67): (0) extends for more than two-thirds of the dentary; (1) extends for less than one-half of the dentary.

190. Splenial, posterior extent (E-66): (0) extends posterior to the apex of the coronoid (fig. 36D); (1) terminates at, or anterior to, the coronoid apex (fig. 38E).

191. Splenial, overlap with postdentary bones (RZ-121): (0) overlap, no hinge with angular (fig. 36D); (1) abutting, splenial receives angular (figs. 41, 42A); (2) abutting, angular receives splenial (fig. 42B). This character follows the descriptions given in Rieppel and Zaher (2000a) and is unordered. Thus, no assumption is made about whether or not the condition seen in snakes (state 2) is derived from the plesiomorphic condition or from the mosasaur-style intramandibular hinge (state 1). Although Rieppel and Zaher (2000a) score Scolecophidia as plesiomorphic for this character, data in List (1966) and Kley (2004) suggest state 2 for at least some members of this taxon.

Figure 42

Diagrammatic illustration of the construction of the intramandibular joint in mosasaurs and snakes. (A) The mosasaur condition in which the splenial receives a projection of the angular. (B) The snake condition in which the angular receives a projection of the splenial. Modified after Rieppel and Zaher (2000a: figs. 7, 8).

i0003-0090-310-1-1-f42.gif

192. Coronoid, height of coronoid at process relative to the length of the mandible (CN-125): (0) short, broad; (1) tall, narrow. This is an imperfect attempt to quantify the character states “coronoid process tall” and “coronoid process short”. Every attempt has been made, though, for internal consistency within the present analysis.

193. Coronoid, posterior extent of the labial flange (CN-126): (0) absent (fig. 31A); (1) extends mostly labially (fig. 29C, E), does not overlap the posterior margin of the coronoid process in lateral view; (2) extends beyond the posterior margin of the coronoid process in lateral view

194. Coronoid, anterior end (E-70): (0) clasps the dentary; (1) butts against dentary.

195. Coronoid, ventral margin (RZ-128): (0) flat or concave; (1) dorsally convex.

196. Coronoid, long and low anterior process (E-69): (0) absent (e.g., fig. 35); (1) present (figs. 20, 41).

197. Coronoid, medially exposed contact with the anterior inferior alveolar foramen (new): (0) absent (fig. 36D); (1) present (fig. 38E). In some taxa, the anterior ramus of the coronoid is exposed anteriorly between the splenial and the dentary to the level of the anterior inferior alveolar foramen and contributes to its margin.

198. Coronoid, posterior overlap by surangular (E-71): (0) absent (figs. 29C, 30); (1) present (figs. 20, 29E, 35, 36).

199. Surangular, anterodorsal buttress of coronoid (DBC-57): (0) absent (e.g., fig. 36A); (1) present (fig. 41).

200. Surangular, anterior border when disarticulated (E-61): (0) tapering; (1) expanded anterodorsally with vertical anterior margin.

201. Surangular, anterior extension into mental canal (Ga82-75): (0) absent; (1) present.

202. Angular (RZ-132): (0) present; (1) absent.

203. Prearticular crest (E-73): (0) absent; (1) present.

204. Prearticular crest with imbedded angular process (E-73): (0) absent; (1) present.

205. Articular, orientation of the retroarticular process along its long axis (E-75): (0) posteriorly directed; (1) medially deflected.

206. Articular, medial offset of retroarticular process with lateral notch (E-77): (0) absent; (1) present.

207. Articular, retroarticular process with posterior broadening (E-78): (0) absent; (1) present.

208. Articular, presence of a deep fossa on the dorsal or dorsomedial surface of the retroarticular process (the retroarticular process pit) (E-74): (0) present; (1) absent.

209. Articular, tubercle on the medial margin of the retroarticular process (E-76; GN00): (0) absent; (1) present as a tubercle; (2) present, elaborated into a fingerlike process.

210. Articular, torsion of the retroarticular process (GN98-70): (0) absent; (1) present; (2) present and strongly twisted.

211. Dentition, spacing (C06-33): (0) closely spaced (fig. 26A); (1) widely separated; spaces between tooth bases greater than one-half the width of a tooth shaft (figs. 22C, 26B); (2) tightly packed (ctenodont) (figs. 20, 29A, C, 38A).

212. Dentition, form of middle and posterior marginal teeth (Ga82-34): (0) straight, pointed; (1) triangular; (2) trenchant, curved; (3) incipient cusps on posterior teeth; (4) teeth with multiple crowns; (5) globidont; (6) squared dorsal margin. Premaxillary teeth and anterior maxillary and dentary teeth tend to be the most variable in their form, whereas more posterior marginal teeth are usually more uniform with one another (although often different from the more anterior teeth).

213. Dentition, waist on marginal teeth (new/extensively modified): (0) absent (e.g., fig. 38E); (1) present (e.g., fig. 43).

Figure 43

The skull of Dorsetisaurus purbeckensis as reconstructed based on the individual elements illustrated and described in Hoffstetter (1967a). Note that the dentary teeth are much larger than the corresponding maxillary teeth. Reconstructed portions are shown as semi-opaque shadows.

i0003-0090-310-1-1-f43.gif

214. Dentition, marginal tooth implantation (RZ-146): (0) labially pleurodont (e.g., fig. 37C); (1) acrodont (fig. 37A); (2) modified pleurodont; (3) enclosed by expanded interdental ridge; (4) subacrodont. The terminology used in this character and its scoring are based on several recent studies and may be slightly different from traditional usage (Zaher and Rieppel, 1999a; Rieppel and Zaher, 2000a).

Borsuk-Białynicka (1996) described a condition of tooth permanency in some squamates that is, apparently, related to acrodonty and posterior extension of the dentary as exemplified by Pleurodontagama aenigmatodes (possibly a young Priscagama gobiensis as discussed by Gao and Norell, 2000) (Borsuk-Białynicka, 1996). Posterior extent of the dentary and presence of acrodont dentition do not invariably co-vary, so both of these characters are included in the current analysis.

215. Dentition, caniniform teeth (Eagam-1): (0) absent; (1) present.

216. Dentition, anterior marginal teeth (DBC-53): (0) generally perpendicular to the long axis of the jaw; (1) procumbent.

217. Dentition, expanded bases on marginal teeth (RZ-149): (0) absent; (1) present, main shafts of teeth somewhat separated.

218. Dentition, plicidentine (E-86): (0) absent; (1) present. Proplatynotia and Paravaranus coded as per Gao and Norell (1998). Note that there is some question about the homology of basally ridged teeth in some taxa and the relationship of this to true plicidentine (Kearney and Rieppel, 2006). Presence of plicidentine can only be assessed in the context of a broken/sectioned tooth clearly showing the presence or absence of dentine folds or, in some cases, in high-resolution x-ray computed tomography scans (Kearney and Rieppel, 2006).

219. Dentition, crown striations (M70): (0) absent; (1) present. Striations are present most commonly in anguids, but are also present in the pleurodont teeth of Diphydontosaurus (Whiteside, 1986) and many mosasaurs (e.g., Mosasaurus hoffmanni; fig. 41). Because of possible asymmetric wear on teeth, crown striations are coded as a single character in the present analysis.

220. Dentition, venom groove (PGG86-30): (0) absent (fig. 36D); (1) present (fig. 38E). Nydam (2000) carefully reviews the morphology of the various relevant taxa for this character and the codings given in this analysis are based mainly on that study.

221. Dentition, replacement (E-85): (0) develop lingually, large resorption pit; (1) posterolingually, resorption pit; (2) posterolingually, no resorption pit.

222. Dentition, enlarged median premaxillary tooth more than half again the diameter of the other premaxillary teeth (Ke-114): (0) absent; (1) present. Kearney (2001, 2003a, 2003b) takes issue with the previous coding of this character, her discussions are used as a basis for the codings in the present analysis.

223. Dentition, premaxillary teeth compared to maxillary teeth (RZ-156): (0) similar (e.g., fig. 38A); (1) markedly smaller (e.g., fig. 38B–D); (2) absent. Note that snakes sometimes lack premaxillary teeth (state 2).

224. Dentition, maxillary teeth (new): (0) present; (1) absent.

225. Dentition, dentary teeth (new): (0) present; (1) absent. Coding of the presence or absence of maxillary and dentary teeth is not redundant with other dental characters. Taxa lacking maxillary and dentary teeth are coded as “-” or “unknown due to change” in this analysis for characters associated specifically with maxillary or dentary teeth. No taxon included in this study lacks both maxillary and dentary teeth.

226. Dentition, chisel shaped posterior teeth (M70): (0) absent; (1) present.

227. Hyoid, second ceratobranchial (E-91): (0) present; (1) absent.

228. Hyoid, second epibranchials (E-90): (0) present; (1) absent.

229. Epihyal, shape (Egek-20): (0) small, triangular; (1) large, winglike.

230. Notochord, in adults (Kl87-2): (0) persistent; (1) obliterated.

231. Vertebrae, centrum morphology (Kl87-1): (0) amphiplatyan; (1) amphicoelous; (2) procoelous. Examination of published descriptions and specimen observations confirm that a persistent notochord is not always concomitant with amphicoelous vertebrae (see, for example, Evans, 1994a).

232. Vertebrae, neural spines (GN98-78; Ke03-134): (0) short and broad; (1) tall and narrow; (2) absent. State (2) is used for those taxa in which there is no projection above the level of the neural canal roof.

233. Vertebrae, precondylar constriction (E-94): (0) absent; (1) present, weakly constricted; (2) strongly constricted to less than 80% of the maximum condylar diameter.

234. Vertebrae, obliqueness of condyles (E-92): (0) absent/weak, posterior apex of condyle visible; (1) moderate, articulating condylar surface slightly visible in ventral view; (2) strong, articulating surface not visible in ventral view.

235. Vertebrae, zygosphenes-zygantra (RZ-169): (0) absent; (1) present, zygosphene articular surface faces dorsolaterally (fig. 44A); (2) present, zygosphene articular surfaces face ventrolaterally (fig. 44B). Rieppel and Zaher (2000a) pointed out the difference in structure between some taxa possessing these accessory articulations. There is no reason to assume primary homology between the two types of accessory articulation described here. Thus, there is no clear nested set of homologies and this character is considered unordered. Retaining this character as unordered further allows assessment of whether the presence/absence of zygosphenes-zygantra is a viable character; that is, whether or not one morphology of zygosphenes evolved from the other.

Figure 44

Anterior view of two vertebrae bearing zygosphenes. (A) Necrosaurus cayluxi possesses dorsolaterally oriented zygosphenes [character state 235(1)]. (B) Natrix natrix possesses ventrolaterally oriented zygosphenes [character state 235(2)]. Modified after (A) Rage (1978) and (B) Parker and Grandison (1977).

i0003-0090-310-1-1-f44.gif

236. Vertebrae, presacral number (E-105, 106): (0) 25 or fewer; (1) 26; (2) 27 or more. More work is needed to help identify the informative character states for this character. Although the character is informative as used here and similarly used in recent publications (Evans and Barbadillo, 1998; Gao and Norell, 1998; Lee, 1998; Lee and Caldwell, 2000), its conception is imperfect.

237. Presacral vertebrae, notching of synapophyses: (0) absent; (1) present.

238. Presacral vertebrae, length of transverse processes: (0) short, subequal or shorter than centrum; (1) more than the length of the centrum. This character is informative only with the inclusion of Marmoretta and Kuehneosauridae (see appendix 1).

239. Atlas, dorsal margin (CN-158): (0) horizontal; (1) posteroventrally inclined. This character is used to describe the dorsal margin of the atlas neural spine when in articulation and with the skull held horizontally.

240. Atlas, lateral process (CN-159): (0) well defined with some posterior overlap of the axis; (1) small, a “hill-like” projection; (2) absent.

241. Cervical vertebrae, length relative to the dorsal vertebrae (L97-103): (0) cervical vertebrae subequal to or shorter than the dorsal vertebrae; (1) more elongate than the dorsal vertebrae.

242. Cervical vertebrae (E-107, 108): (0) 8; (1) 7 or fewer; (2) 9; (3) 10 or more.

243. Cervical, intercentra (E-97): (0) intervertebral; (1) sutured to the posterior part of the preceding centrum; (2) fused to posterior part of preceding centrum; (3) fused to the succeeding centrum (CN-161); (4) absent.

244. Cervicals, hypapophyseal keel (EB98-149): (0) absent; (1) present.

245. Dorsal vertebrae, pachyostosis (LC00-196): (0) absent; (1) present.

246. Dorsal vertebrae, intercentra (EB98-86): (0) present; (1) absent.

247. Sacral vertebrae, functional (DBC-72): (0) present; (1) absent. “Functional” sacral vertebrae are defined here as those whose sacral ribs contact the ilium, anchoring the pelvis to the vertebral column.

248. Cloacal vertebrae, lymphapophyses (LC00-209): (0) absent; (1) present.

249. Caudal vertebrae, dorsoventral height (including the neural spines and chevrons) (CDB—mosasauroid character 9): (0) unexpanded; (1) expanded, creating a sculling organ (depth of proximal and mid caudal vertebrae, including chevrons, greater than 3 times length of centrum).

250. Caudal vertebrae, zygapophyses and transverse processes (CDB5): (0) well developed, zygapophyses extending more one-fourth the length of centrum; (1) reduced, creating greater flexibility of the trunk and tail. This character complex describes the flexibility of the tail. Greater flexibility is typically associated with the further development of the tail as a sculling organ (see character 249).

251. Caudal vertebrae, transverse processes (E-100-102): (0) single (fig. 45A, B, D); (1) double, diverging (fig. 45C); (2) double converging; (3) absent.

Figure 45

Dorsal view of the various caudal vertebrae possessing autotomy planes. (A) Shinisaurus crocodilurus (UF 57712) possesses character states 251(0) and 252(0); (B) Gekko sp. possesses character states 251(0) and 252(1); (C) Dipsosaurus dorsalis possesses character states 251(1) and 252(0); and (D) Anolis sagrei possesses character states 251(0) and 252(3). The vertebra in (A) is modified from Conrad (2006a), B–D are redrawn from Estes et al. (1988). Not to scale.

i0003-0090-310-1-1-f45.gif

252. Caudal vertebrae, autotomy planes (E-103): (0) present on (or between) the transverse process(es) (fig. 45A, C); (1) present posterior to the transverse process(es) (fig. 45B); (2) absent; (3) present anterior to transverse processes (fig. 45D). Species and even specimens are sometimes polymorphic for this character.

253. Caudal vertebrae, pedestals for chevrons (RZ-183): (0) bulges; (1) well-developed, expressed as relatively deep and discrete pedestals.

254. Chevrons, position (RZ-185): (0) at the posteroventral margin of the centrum; (1) anterior to the posteroventral margin of the centrum; (2) fused to vertebrae (GN98-83).

255. Ribs, anteroventral pseudotuberculum (LC00-207): (0) absent; (1) present.

256. Ribs, posterodorsal pseudotuberculum (LC00-208): (0) absent; (1) present.

257. Ribs, expansion and flattening of the anterior presacral ribs (CN-166): (0) absent; (1) present.

258. Ribs, postxiphisternal inscriptional ribs (FE-40): (0) contacting the dorsal ribs, not contacting at midline; (1) contacting dorsal ribs, one or more pairs confluent at midline; (2) free dorsally, confluent ventrally. Some taxa scored for this character based on Torres-Carvajal (2004).

259. Postcloacal bones (E-125): (0) absent; (1) present.

260. Clavicles, shape (RZ-196): (0) rodlike; (1) expanded proximally with notch or fenestra; (2) absent.

261. Clavicle, angulation (E-116): (0) straight, without angulation; (1) strongly curved/angled.

262. Coracoid, anterior (primary) coracoid emargination (E-112): (0) absent; (1) present.

263. Coracoid, posterior (secondary) emargination (E-113): (0) absent; (1) present.

264. Epicoracoid cartilage, contact with suprascapula (E-114): (0) present; (1) absent.

265. Scapula, size relative to the coracoid (DBC-95): (0) scapula subequal to, or larger than the corocaoid; (1) scapula smaller than the coracoid; (2) scapula and coracoid absent.

266. Scapula, secondary scapular fenestra formed by a scapular epicoracoid bar (E-111): (0) absent; (1) present.

267. Sternum, rib attachments (E-109): (0) five; (1) four; (2) three; (3) two or fewer. Taxa lacking sterna cannot be scored for this character.

268. Interclavicle (E-118): (0) present; (1) absent.

269. Interclavicle, anterior process (E-120): (0) absent; (1) present, single; (2) present, double.

270. Interclavicle, lateral arms (E-119): (0) present; (1) absent.

271. Sternum (Ke-146): (0) present, articulates with pectoral girdle; (1) present, reduced and does not articulate with the pectoral girdle; (2) absent.

272. Sternum, proximity to the lateral arms of the interclavicle (FE-33): (0) separated by more than one-third the posterior process of the interclavicle; (1) separated by one-third or less the length of the posterior process. Xenosaurus is coded as polymorphic; an illustration in Renous-Lécuru (1968) indicates derived condition for X. grandis, but X. platyceps has the plesiomorphic state.

273. Sternum, fontanelle (E-121): (0) absent; (1) present. Some taxa were coded based on Renous-Lécuru (1968).

274. Xiphisternum, branching (this character is derived from figures and descriptions in Renous-Lécuru, 1968; Estes et al., 1988; and Etheridge and de Queiroz, 1988; CN-179): (0) more than one branching; (1) one branching; (2) unbranched.

275. Humerus, shape (DBC-104): (0) elongate, sub-cylindrical, and twisted such that distal ends at right angles to one another; (1) flattened; hourglass-shaped; (2) flattened—with square, ends expanded but equal; (3) flattened; rhomboid, distal end more expanded; (4) humerus absent.

276. Humerus, deltopectoral crest (DBC-107): (0) single continuous projection; (1) separate, but joined by a lamina; (2) separate, not joined by a lamina. Ordered by DeBraga and Carroll (1993), but not ordered here.

277. Ectepicondylar foramen (EB98-103): (0) foramen (fig. 46A); (1) groove; (2) absent altogether (fig. 46B).

Figure 46

Left humeri of (A) Telmasaurus grangeri (AMNH FR6643) in anterodorsal view; and (B) Lacerta viridis (AMNH R1148) in dorsal view. Note the absence of a ectepicondylar foramen and groove [character state 277(2)] in (B).

i0003-0090-310-1-1-f46.gif

278. Forelimb, zeugopodium (CN-181): (0) present; (1) absent.

279. Radius, preaxial ridge (DBC-113): (0) absent; (1) thin, rounded, lamina extending for more than one-half the anterior margin; (2) present, rounded and extending for less than one-half the anterior edge; (3) present, greatly expanded with an anteroproximal apex.

280. Ulna, articulation with intermedium (DBC-111): (0) absent, intermedium does not contact the ulna; (1) present, no facet; (2) present, with a distinct intermedium facet on the ulna.

281. Carpus, intermedium (Egek-24): (0) present; (1) absent.

282. Manus, first metacarpal (DBC-116): (0) similar in robustness to other metacarpals; (1) robust, more than 1.5 times wide as the other metacarpals.

283. Manus, orientation of the fifth digit relative to the others (DBC-119): (0) fifth digit not greatly divergent; (1) at greater than 70 degrees from fourth digit.

284. Pelvis (RZ-207): (0) fused into a single ossification such that the sutre lines have become indistinct; (1) strongly sutured, but with distinct suture lines visible; (2) nonsutural contacts. Many taxa have fused the pelvis so tightly that sutures are not visible in the acetabulum or on the medial surface of the pelvis (state 0). Many others have sutures visible (state 1). Other taxa, especially those adapted to aquatic lifestyle and/or limb-reduced taxa, have lost sutural contacts altogether (2). In the latter, the bones of the pelvis may contact one another, but are not sutured.

285. Pubis, relative length of the symphysial portion compared to the tubercular portion (E-124): (0) shorter than; (1) subequal to, slightly longer than; (2) more than one-half again as long.

286. Pubis, distal shape (DBC-125): (0) expanded and fanlike; (1) slender. This character is problematic because the derived state as described by DeBraga and Carroll (1993) seems to be plesiomorphic. This character is included here with some reservations, but is coded consistently within the analysis. Taxa that lack an ossified pubis are not coded, even if a cartilaginous element is present.

287. Ilium, anterior process (RZ-208): (0) present; (1) absent.

288. Femur, distal condyles (DBC-132): (0) separate and distinct such that the distal part of the femur is a single convex entity; (1) confluent.

289. Femur, shape (DBC-130): (0) cylindrical with moderately expanded proximal and distal ends; (1) flattened, breadth of distal end more than one-quarter the bone's length; (2) femur absent.

290. Pes (CN-186): (0) present with digits and/or metatarsals (fig. 47A); (1) absent (fig. 47B); (2) tarsal element(s) only. Gasc and Renou (1979) illustrate a single distal element in Dibamus.

Figure 47

Hind limbs (A) Haasiophis terrasanctus and (B) Blanus cinereus. The hind limb of Haasiophis terrasanctus, a snake is well developed with tarsals, metatarsals, and phalanges [character state 290(0)]. The hind limb of Blanus is composed only of a femur without distal elements such as a pes [an exemplar of character state 290(1)]. Note that the dark gray represents unknown or heavily reconstructed portions and the light gray (in B) represents cartilaginous elements. The gray outlines represent portions that are hypothesized to have originally been present, but for which there is not direct evidence. Modified after (A) Tchernov et al. (2000) and Rieppel et al. (2003), and (B) Kearney (2002).

i0003-0090-310-1-1-f47.gif

291. Pes, relative positions of the medial and lateral plantar tubercles on metatarsal V (R80L; C06-44): (0) even with one another or overlapping levels; (1) lateral tubercle distally placed; (2) lateral tubercle distally placed, approaching condyle; (3) greatly shortened metatarsal V precludes identification. Rieppel (1980b) originally identified the distal placement of the lateral plantar tubercle in the varanid Lanthanotus borneensis. Later, the same character state was observed in shinisaurids (Conrad, 2005a, 2006a, 2006b). This character has now been scored across squamates and the derived states have been observed in various taxa.

292. Astragalus and calcaneum, fusion (L98-215): (0) separate; (1) fused; (2) absence of a bony calcaneum. Lee and Caldwell (2000) scored Aphanizocnemus as possessing a separate astragalus and calcaneum, but this is ambiguous based on the description by Dal Sasso and Pinna (1997). Many aquatic taxa apparently lack a calcaneum (Russell, 1967; Bell, 1997; however see Caldwell, 1996). However, because ossification of various distal appendicular elements is reduced in these forms, it seems possible that a cartilaginous calcaneum might have originally been present.

293. Egg teeth (KL87-15): (0) single; (1) double; (2) absent.

294. Femoral/precloacal pores (E-144): (0) absent; (1) present.

295. Integument, gular fold with distinctive midventral squamation (FE-47): (0) absent (figs. 48, 49B); (1) present (fig. 50).

Figure 48

Uromastyx dispar maliensis ( = Uromastyx maliensis of the usage of Joger and Lambert, 1996). (A) Broad view of the body except the tail tip. (B) Detail of the head. Note the character states 198(1) (compare with fig. 51) and 300(1) (compare with fig. 50), among others visible in this figure. Specimen housed at Dickerson Park Zoo in Springfield, Missouri. Photo by R. M. Shearman.

i0003-0090-310-1-1-f48.gif

Figure 49

The head of two species of Varanus. (A) Dorsolateral view of the head and neck in Varanus komodoensis exhibiting character states 3(1), 298(1), 300(1), and 310(3). (B) Ventrolateral view of the head of Varanus salvator exhibiting character states 3(0) and 295(0) (compare with fig. 50). Specimens housed at (A) the Audubon Zoo in New Orleans, Louisiana, and (B) photographed at AMNH, part of the collection from Clyde Peeling's Reptiland in Allenwood Pennsylvania. Photos by R. M. Shearman.

i0003-0090-310-1-1-f49.gif

Figure 50

Chamaeleo calyptratus, exhibiting character state 295(0), 298(1), and 300(0). Specimen photographed at AMNH, part of the collection from Clyde Peeling's Reptiland in Allenwood Pennsylvania. Photo by R. M. Shearman.

i0003-0090-310-1-1-f50.gif

Figure 51

Tiliqua rugosa asper (sensu Shea 1988;  = Trachydosaurus rugosus). (A) Broad view of the body. (B) Detail of the head. Note the presence of character states 298(2), 305(1), 306(1), 307(1), 308(1), and 314(0). Specimen housed at the San Diego Zoo in Balboa Park, California. Photo by R. M. Shearman.

i0003-0090-310-1-1-f51.gif

296. Integument, annular rings (dermal/epidermal) in the body squamation (Ke-3): (0) absent; (1) present.

297. Integument, scale organ ornamentation (FE-52; H93): (0) absent; (1) spinules; (2) spikes. Original character from Frost and Etheridge (1989); character state “spikes” added from Harvey (1993).

298. Squamation, cephalic scales (E-147; M70): (0) absent; (1) small and irregularly shaped (figs. 48, 49); (2) enlarged plates (fig. 51). The character state (0) absent does not occur in any taxon used here, although some gekkotans possess only lightly keratinized dermal tubercles that do not follow the traditional definition of a scale. This character is considered unordered.

299. Squamation, contact between frontal and parietal scale (M70): (0) absent; (1) present. Meszoely et al. (1978) report that the plesiomorphic condition occurs in Peltosaurus. Lacertids are coded here with consideration of data from Succinilacerta succinea, an Eocene lacertid preserved in Baltic amber (Borsuk-Białynicka et al., 1999).

300. Squamation, middorsal scale row (E-146): (0) differing from surrounding scales, elongate with apices (fig. 50); (1) similar to surrounding scales (figs. 48, 49, 51, 52).

Figure 52

Heloderma suspectum (H. s. cinctum sensu Bogert and Del Campo, 1956); anterior part of the body in dorsolateral view highlighting character states 298(1), 300(1), 305(1), 306(0), and 310(2). Specimen photographed at AMNH, part of the collection from Clyde Peeling's Reptiland in Allenwood Pennsylvania. Photo by R. M. Shearman.

i0003-0090-310-1-1-f52.gif

301. Squamation, cycloid scales (E-148): (0) absent; (1) present.

302. Squamation, cephalic scale fusion (Ke-5): (0) absent; (1) present. This character addresses the presence or absence of expansive head shields that incorporate (and obscure the boundaries of) individual head scales such as labials or superciliaries (or others).

303. Squamation, imbrication (M70): (0) absent (fig. 49); (1) present (fig. 51).

304. Squamation, lateral fold in body (MB54): (0) absent; (1) present.

305. Squamation, dorsal body osteoderms (E-127): (0) absent (figs. 48, 50); (1) present (figs. 49, 51, 52).

306. Squamation, dorsal compound osteoderms (E83): (0) absent; (1) present. Taxa lacking dorsal body osteoderms cannot be scored for this character.

307. Squamation, ventral body osteoderms (E-126): (0) absent; (1) present.

308. Squamation, ventral compound osteoderms (E83): (0) absent; (1) present. Taxa lacking ventral body osteoderms cannot be scored for this character.

309. Osteoderms, grooves separating osteoderms on maxilla (Y76-5): (0) absent (fig. 26A); (1) present (fig. 26B).

310. Squamation, osteoderm thickening (C06-48): (0) absent, osteoderms thin plates or noncalcified (fig. 51); (1) present, irregularly shaped (fig. 22B); (2) present, polygonal mounds (fig. 52); (3) absent, osteoderms wormlike (vermiform). State 3 has been carefully described in the literature (Smith, 1935; McDowell and Bogert, 1954; Auffenberg, 1981; Erickson et al., 2003). Fig. 49 has photos of monitor lizards that would, presumably, possess vermiform osteoderms, but their presence is not obvious from external view (unlike the large thickened osteoderms in taxa such as Heloderma; fig. 52)

311. Squamation, bony tubercles (new/extensively modified): (0) absent; (1) present as individual bony tubercles; (2) large osteoderms covered with individual bony tubercles. Tubercular osteoscutes and bony tubercles have been variably described for numerous taxa and are present in xenosaurids (state 1) and glyptosaurs (state 2), for example.

312. Squamation, keeled osteoderms on body (C06-47): (0) absent; (1) present.

313. Eyeball (Ke-10): (0) complete and exposed; (1) reduced, covered by a head scale; (2) reduced, not externally visible. Although most of the fossil taxa included in this analysis probably possessed state (0), it is impossible to be sure for all except some of those preserved in amber. To minimize assumptions about fossil taxa, this character has been left as “?”s for nonamber specimens.

314. Eye, movable eyelid (Ke-11): (0) present; (1) absent, eyelids fused into a spectacle or brill.

315. Eye, scleral ossicles (Ke-12): (0) present; (1) absent. The number, orientation, morphology, and interrelationships of squamate scleral ossicles have not been included in this matrix as yet. Future iterations of this analysis will address those morphological features, but they require further consideration.

316. Glossus, filamentous tongue papillae (S-6, 7, 8, 9): (0) absent; (1) peglike; (2) individual papillae dorsally asymmetrical, forming points.

317. Glossus, division of the foretongue (E-137): (0) absent; (1) notched more than 10% of length; (2) notched more than 20%; (3) notched more than 40%; (4) notched more than 50% of length. The character states delimited here are slightly modified from Estes et al. (1988). The states for this character both here and in Estes et al. (1988) are arbitrary in their delimitation.

318. Glossus, foretongue retracts within hindtongue at zone of invagination (E-136): (0) absent; (1) present. Many of the tongue characters listed here (including this one) are taken from or modified after both Estes et al (1988) and Schwenk (1988).

319. Glossus, cross-section of tongue (E-138): (0) rounded and glandular; (1) flattened foretongue; (2) keratinized and mushroom shaped foretongue cross-section.

320. Gland of Gabe (GN98-102): (0) absent; (1) present.

321. Ear, external ear opening (auricular depression or canal) (Ke-13): (0) present; (1) absent. For a further discussion of this character, see Greer (2002).

322. Inner ear, thickening of the neural limbus of cochlear duct (Eanguim-14): (0) absent; (1) present. This character is scored based on data presented in Miller (1966), Wever, (1978), and Estes et al. (1988).

323. Inner ear, ciliary restraint for hair cells (E-140): (0) tectorial, lacking sallet systems; (1) tectorial and sallet; (2) more than one-half hair cells inertial.

324. Seromucus glands in inferior labial glands (Eanguim-16): (0) absent; (1) present.

325. Endolymphatic sacs, extension into the nuchal musculature (Kl67-C; EdQ88-35): (0) absent; (1) present, the endolymphatic sacs exit through an aperture between the supraoccipital and parietal; (2) present, the endolymphatic sacs exit through the epiotic foramen; (3) present, the endolymphatic sacs exit through the vagus foramen.

326. Hemipenis, symmetry (B82): (0) present; (1) absent.

327. Hemipenis, sulcus (B82): (0) simple; (1) divided.

328. Hemipenis, dorsal asulcal ornamentation (B82): (0) absence; (1) simple flounces; (2) bifurcated flounces.

329. Hemipenis, m. retractor lateralis posterior substantial situation within hemipeneal sheath (FE-63): (0) absent; (1) present.

330. Hemipenis, horns (B82): (0) absent; (1) present, simple; (2) present, multicusped.

331. Hemibacula (GN98-99): (0) absent; (1) present.

332. Neurology, ulnar nerve position (J72; E-142): (0) “lacertid” style; (1) “varanid” style.

333. Neurology, dorsal leg muscles (J72; E-143): (0) peroneal nerve present; (1) peroneal nerve absent, interosseus innervation.

334. M. anterior mandibulae externus (MAME) profundus origin (GN98-93): (0) supratemporal and parietal; (1) supratemporal only. Data used here were derived from various sources in the literature for different squamate groups (Haas, 1960, 1973; Rieppel, 1980a, 1980d, 1980e, 1981a, 1981b, 1982, 1984a; Estes et al., 1988; Gao and Norell, 1998; Abdala and Moro, 2003).

335. Meatal closure muscle (Kl87-12): (0) absent; (1) present, L-shaped; (2) present, O-shaped.

336. M. extracolumellaris (E-135): (0) absent; (1) present.

337. Myology, anterior extension of m. adductor mandibulae posterior (E-131): (0) no further than the posterior margin of Meckel's canal; (1) anterior to the posterior one-fourth of the dentary.

338. Myology, m. pseudotemporalis superficialis origin (E-132): (0) lateral and anterior margins of the supratemporal fossa; (1) also along the medial margin of the temporal fenestra.

339. M. levator pterygoidii, insertion (GN98-96): (0) extends posteriorly beyond the columellar fossa; (1) restricted anterior to the columellar fossa.

340. M. pseudotemporalis profundus, anterior head (E-133): (0) absent; (1) present, not expanded; (2) present, expanded.

341. M. pseudotemporalis superficialis, origin (E-132): (0) limited to the anterior one-half of the supratemporal fenestra; (1) extends far posteriorly, onto the posterior one-third of the supratemporal fenestra.

342. Bodenaponeurosis, base contact with mandibular fossa (GN98-98): (0) present; (1) absent, attached only to the caudomesial edge of the coronoid Coding from Gao and Norell, 1998 after Lakjer (1926), Haas (1973), and Rieppel (1980a).

343. M. constrictor colli coverage of first ceratobranchials (GN98-92): (0) absent; (1) present.

344. M. genioglossus lateralis, morphology (GN98-95): (0) not separate bundles, not inserting on the hyobranchials; (1) separate bundles, some inserting on the hyobranchials.

345. M. rectus abdominis lateralis (E-134): (0) absent; (1) present.

346. M. episterno-cleido-mastoideus insertion (GN98-91): (0) mainly on the paroccipital process; (1) extensively on parietal.

347. Urinary bladder (B86): (0) present, complete; (1) present, vestigial; (2) absent. Coded largely after data summarized by Beuchat (1986).

348. M. levator anguli oris (AM03-3): (0) present; (1) absent.

349. M. levator anguli oris, aponeurosis (AM03-10): (0) present; (1) absent.

350. Adductor mandibulae externus, tendinous system (AM03-12): (0) absent; (1) present.

351. M. adductor mandibulae posterior (AM03-32): (0) present; (1) absent.

352. M. pseudotemporalis superficialis (AM03-36): (0) present; (1) absent.

353. M. protractor pterygoidei (AM03-51): (0) present; (1) absent.

354. M. retractor pterygoidei (AM03-53): (0) present; (1) absent.

355. M. intermandibularis anterior superficialis (AM03-54): (0) absent; (1) present.

356. M. intermandibularis anterior profundus aponeurosis (AM03-55): (0) absent; (1) present.

357. M. depressor mandibulae profundus (AM03-65): (0) present; (1) absent.

358. M. mandibulohyoideus II (AM03-76): (0) absent; (1) present.

359. M. mandibulohyoideus III (AM03-81): (0) absent; (1) present.

360. M. branchiohyoideus aponeurosis (AM03-83): (0) absent; (1) present.

361. M. ceratohyoideus (AM03-84): (0) absent; (1) present.

362. Muscle “X” (AM03-86): (0) absent; (1) present.

363. M. sternohyoideus (AM03-92): (0) absent; (1) present.

364. Biogeography: (0) global; (1) Madagascar; (2) South America; (3) North America/Central America; (4) Europe/western Asia; (5) sub-Saharan Africa; (6) northern Africa/Arabia; (7) India; (8) East Asia; (9) Australia.

The Characters of Estes et al. (1988)

Many of the characters used in this analysis are directly or indirectly derived from the character list provided by Estes et al. (1988). Estes et al. (1988) used 148 morphological characters in their analysis, 139 of which are covered by the characters used in this analysis. Each of the nine excluded characters will be discussed here.

Estes et al. (1988) Character 40

“Median contact of septomaxillae: (0) separated by a gap filled by the cartilaginous internarial septum; (1) septomaxillae meet or nearly meet on midline in a raised crest” (Estes et al., 1988: 129). The character states included in this description are vague and, as pointed out by Rieppel and Zaher (2000a), an internarial septum always separates the septomaxillae.

Estes et al. (1988) Character 51

“Opisthotic-exoccipital fusion: (0) bones remain separate or fuse to exoccipitals relatively late in postembryonic ontogeny; (1) fuse to exoccipital in embryo or in early postembryonic ontogeny, or the two bones develop from a single ossification center” (Estes et al., 1988: 130). This character was omitted because of the extremely limited data available to determine character states. Additionally, the derived state apparently refers to two non-homologous conditions.

Estes et al. (1988) Character 57

“Meckel's canal exposure ventrally: (0) opens medially for entire length; (1) opens ventrally anterior to anterior inferior alveolar foramen” (Estes et al., 1988: 130). Scoring of this character is difficult and may be variable within a specimen. Moreover, the ventral surface of the mandible is dependent upon the orientation of the naturally articulated mandible, something not always immediately apparent.

Estes et al. (1988) Character 62

“Medial view of prearticular with dentary and splenial removed: (0) prearticular extends nearly to anterior end of surangular, well anterior to coronoid bone; (1) reduced not extending well anterior to the coronoid bone” (Estes et al., 1988:131). The character states of this character are dependent upon the anterior extensions of the coronoid and the surangular, each of which being the subject of other independently varying characters.

Estes et al. (1988) Character 88–89

“Scleral ossicle number I: (0) more than 14 ossicles; (1) 14 ossicles or fewer […] Scleral ossicle number II: (0) 14 ossicles ore more; (1) fewer than 14 ossicles” (Estes et al., 1988: 132). These two characters possess overlapping character states. Inclusion of this character could be accomplished in the future with further subdivision of the number of scleral ossicles.

Estes et al. (1988) Character 117

“Dorsal articulation of clavicle: (0) articulates with scapula; (1) articulates with suprascapula” (Estes et al., 1988: 133). Scoring this character relies heavily on the use of skeletonized specimens. Such specimens are usually desiccated, meaning that the clavicles may be pulled out of position through the shrinkage of soft tissue, introducing the potential for erroneous observations of character states.

Estes et al. (1988) Character 123

“Notching of distal tibial epiphyses: (0) gently convex for astragalocalcaneal articulation; (1) tibial epiphysis more or less distinctly notched, fitting onto a ridge on the astragalocalcaneum” (Estes et al., 1988: 133). The character states described for this character are vague and allow for much interpretation. However, a modified version of this character may be implemented in future analyses.

Estes et al. (1988) Character 130

“Epiphysis fusion: (0) fuse to diaphyses at same time or after fusion of braincase elements; (1) fuse to diaphyses prior to fusion of braincase elements” (Estes et al., 1988: 133). As with character 51 in Estes et al. (1988) (above), there are little data available to aid in meaningfully and accurately scoring this character and so it is omitted.

Analyzing the Data

The morphological data set includes 363 morphological characters scored in 222 ingroup taxa (appendix 2) with the outgroup Rhynchocephalia. Seventeen characters were identified as parsimony uninformative. Multistate characters were treated as ordered only if they formed a clear set of nested homology statements (characters 1, 32, 59, 83, 93, 96, 111, 114, 124, 136, 137, 162, 164, 181, 280, 285, 313, 317, 323, and 330 were considered ordered in this analysis). The taxon-character matrix was analyzed via a new technology search using the ratchet option in the computer program T.N.T. (Goloboff et al., 2003) with the option set to find the shortest tree 1,000 times. The resulting trees were saved to RAM and two additional ratchet runs, each of 1,000 iterations were performed on these trees. The resulting 2,213 trees were used for the strict consensus cladogram (figs. 53, 54). The principle trees from the T.N.T. analysis were exported to the computer program PAUP* (Swofford, 2001) and Adams consensus trees were computed. Adams trees and strict consensus trees are reported here (figs. 55, 56) because they identify the groupings consistent within all trees. Because the current analysis does not include an exhaustive search, no majority rule trees are reported. Synapomorphies for each node in the Adams consensus tree were identified by PAUP* (Swofford, 2001) and unambiguous synapomorphies are reported below.

Figure 53

Hypothesis of squamate interrelationships based on the current study. This figure shows entire tree at once; portions of the tree are shown individually in Figures 54 and 55. Note the presence of a clade containing cordyloids, lacertoids, and scincoids, to the exclusion of all other squamates. Serpentes (snakes) are shown here within Scincoidea (see also figs. 54C, 55C, and 56C). This figure also illustrates that anguimorphs are sampled most heavily within this analysis.

i0003-0090-310-1-1-f53.gif

Figure 54

(A–F) Hypothesis of squamate interrelationships based on the current study; the strict consensus of 2,213 shortest recovered trees from the analysis described in the text. Each tree had a length of 3,273 steps, CI of 0.1499, and RC of 0.7164. The numbers represent Bremer support (decay indices) for individual nodes. Pluses (+) indicate indices of 8 or more.

i0003-0090-310-1-1-f5401.gif

Figure 54

Continued.

i0003-0090-310-1-1-f5402.gif

Figure 54

Continued.

i0003-0090-310-1-1-f5403.gif

Figure 54

Continued.

i0003-0090-310-1-1-f5404.gif

Figure 54

Continued.

i0003-0090-310-1-1-f5405.gif

Figure 54

Continued.

i0003-0090-310-1-1-f5406.gif

Figure 55

(A–F) Hypothesis of squamate interrelationships based on the current study; the Adams consensus with accompanying taxonomic scheme. Following (Brochu, 1999), parentheses indicate stem-based clade names, dots indicate node-based names.

i0003-0090-310-1-1-f5501.gif

Figure 55

Continued.

i0003-0090-310-1-1-f5502.gif

Figure 55

Continued.

i0003-0090-310-1-1-f5503.gif

Figure 55

Continued.

i0003-0090-310-1-1-f5504.gif

Figure 55

Continued.

i0003-0090-310-1-1-f5505.gif

Figure 55

Continued.

i0003-0090-310-1-1-f5506.gif

Figure 56

(A–E) Temporally calibrated hypothesis of squamate interrelationships based on the Adams consensus with some taxa collapsed in the interest of space. Temporal ranges are approximate.

i0003-0090-310-1-1-f5601.gif

Figure 56

Continued.

i0003-0090-310-1-1-f5602.gif

Figure 56

Continued.

i0003-0090-310-1-1-f5603.gif

Figure 56

Continued.

i0003-0090-310-1-1-f5604.gif

Each of the 2,213 most parsimonious recovered trees from the T.N.T. (Goloboff et al., 2003) analysis have a length of 3,273 steps, a consistency index (excluding uninformative characters) of 0.1499 and a retention index of 0.7164 as reported by PAUP* (Swofford, 2001). Note that consistency index (CI) is inversely correlated with the number of included taxa in a given analysis (Klassen et al., 1991:446).

Phylogeny and Taxonomy

The revised phylogeny and taxonomy of Squamata presented here is intended to improve the current state of systematic understanding. Taxon name definitions used here are intended to follow the most common usage and allow for easy incorporation of new discoveries while also allowing for revision of phylogenetic hypotheses. This approach is in contrast to the approach taken by Lee (1998) who suggests that the definitions of taxon names should be specific to each phylogenetic hypothesis (see the Introduction, above).

Importantly, the name Squamata itself is not defined here. Squamates are constrained to be monophyletic by outgroup and ingroup choice and so no diagnosis may be offered. Estes et al. (1988) defined the taxon as a crown group, but common usage of the name incorporates all non-rhynchocephalian members of the crown-node Lepidosauria. That is, my perception is that lepidosaurs are regarded as always being either rhynchocephalians or squamates, regardless of whether they fall within the squamate crown. The taxa included under Squamata would remain the same in the current analysis no matter if a node-based or stem-based definition were used.

Iguanomorpha Sukhanov, 1961

(figs. 54A, 55A, 56A)

Definition

All taxa sharing a more recent common ancestor with Iguana iguana than with Gekko gecko, Scincus scincus, or Varanus varius.

Diagnosis

Iguanomorphs are united by two unambiguous synapomorphies in the current analysis, 230(1) notochordal canal obliterated by centrum ossification, and 231(2) procoelous vertebrae.

Comments

Sukhanov (1961), following earlier authors, described a basal dichotomy between iguana-like lizards and all other squamates. He referred to all iguana-like squamates as Iguanomorpha, a name that works well as a stem-defined group to include the crown iguanians and their fossil relatives. Hoyalacerta sanzi falls on the iguanomorph stem in this analysis, rather than in a position basal to other squamates as previously suggested (Reynoso, 1998; Evans and Barbadillo, 1999). Thus, it gives insight into character polarization at the base of the Iguanomorpha and is important for determining interrelationships within Iguania.

Iguania Cuvier, 1817

(figs. 54A, 55A, 56A)

Definition

The most recent ancestor of extant taxa more closely related to Iguana iguana than to Gekko gecko or Varanus varius and all descendants of that ancestor.

Diagnosis

Iguania, as defined here, may be diagnosed by 11(1) premaxilla fusion (unpaired premaxilla), 34(1) presence of a prefrontal tuberosity, 117(1) ventromedial processes of the pterygoids, 244(1) hypapophyseal keels present on the cervical vertebrae, and 252(1) caudal autotomy planes present posterior to the transverse processes.

Comments

Estes et al. (1988) defined Iguania as a node-based name describing “[t]he last common ancestor of Iguanidae*, Agamidae* and Chamaeleontidae and all of its descendants,” with the caveat that Iguanidae and Agamidae might represent paraphyletic taxa. Even so, their definition was apparently intended to include all extant taxa closer to Iguana than to geckos, skinks, etc. The revised definition does that more precisely and does not rely upon metataxa.

Iguania exclusive of Phrynosomatidae

(figs. 55A, 56A)

Diagnosis

The clade containing crown iguanians exclusive of phrynosomatids is diagnosed by 26(1) strong medial processes of the maxillae behind the nasal process of the premaxilla, 29(1) weakly inclined anterior margin of the maxillary nasal process, 83(1) presence of a nuchal fossa on the parietal table, 104(1) presence of a contact between the septomaxilla and the osseous roof of the nasal capsule, 181(2) closed and fused Meckel's canal, 261(1) angulated clavicle, and 263(1) a posterior coracoid emargination.

Comments

The current analysis suggests that the basal dichotomy in Iguania is between Phrynosomatidae and other iguanians, the latter including acrodontans. Other recent analyses have suggested a basal dichotomy between Chamaeleontiformes (see below) and Pleurodonta (sensu Conrad et al., 2007; Conrad and Norell, 2007a; and see preliminary data in Conrad, 2005a) ( = Iguanidae sensu lato). The different results appear to be caused, at least in part, by the inclusion of Isodontosaurus (also present in Conrad and Norell, 2007a) and by the inclusion of additional non-osteological character states (see below).

Opluridae + Tropidurinae + Liolaemus + Leiocephalus

(figs. 55A, 56A)

Diagnosis

This clade is supported by six unambiguous synapomorphies in the current analysis. These are presence of 9(1) dermal sculpturing on the prefrontal, 29(0) steeply inclined anterior margin of the maxillary nasal process, 121(1) midline contact of the palatines, 143(1) posterolateral sphenoid flanges ventrolaterally overlying the basioccipital, 180(1) elongate lateral portion of the dentary extending along the mandible to a point at least half way between the coronoid eminence and the mandibular glenoid, 185(0) dentary without a notch distinguishing coronoid and surangular processes, and 190(0) splenial extending far anteriorly.

Comments

The current analysis does not recover unambiguous support for a monophyletic Tropiduridae sensu Frost and Etheridge (1989). Recent phylogenetic analyses have suggested the polyphyly of tropidurids and polychrotids (Frost et al., 2001; Schulte et al., 2003). However, two recent morphological analyses suggest tropidurid monophyly (Conrad et al., 2007; Conrad and Norell, 2007a). The first run of the current analysis also supports tropidurid monophyly, but the additional 1000 ratchet replicates lost support for this clade in each, the strict and the Adams consensuses. The presence of a clade containing all tropidurids sensu lato and oplurids is consistent with some networks recovered Frost and Etheridge (1989).

Opluridae Moody, 1983

(figs. 55A, 56A)

Diagnosis

Oplurids (here represented by two specimens of Oplurus quadrumcinctus, O. cyclurus, and Chalarodon madagascariensis) are united in this analysis by five unambiguous character states. These are 47(0) presence of an angulated jugal, 60(1) a dorsoventrally inflated frontal, 263(0) absence of a posterior coracoid emargination, 332(1) “varanid” style ulnar nerve position, and 333(1) absence of a peroneal nerve.

Comments

I refrain from defining Opluridae here, but it should be used to include Oplurus madagascariensis minimally in the future. The diagnosis above describes the taxa included in this analysis. A recent phylogenetic analysis (Conrad et al., 2007) suggest that Oplurus is the sister taxon to tropidurids (sensu Frost and Etheridge, 1989) and that Chalarodon is the sister-taxon to a clade containing the fossil taxa Igua and Polrussia. Titus and Frost (1996) recovered oplurid monophyly as does the current study, but this problem warrants further investigation.

Crotaphytidae + Iguanidae + Polychrotiformes + Hoplocercidae + Chamaeleontiformes

(figs. 55A, 56A)

Diagnosis

This clade may be diagnosed by three unambiguous synapomorphies. These are 9(1) dermal sculpturing on the prefrontal, 29(0) steeply inclined anterior margin of the maxillary nasal process, and 90(1) presence of a postorbital tuberosity.

Chamaeleontiformes + Hoplocercidae + Polychrotiformes

(figs. 55A, 56A)

Diagnosis

This clade is supported by five unambiguous synapomorphies: 21(1) nasals that are in contact for less than half their length in dorsal view, 52(0) jugal lying mostly medial (rather than dorsal) to the maxilla, 104(0) no contact between the septomaxilla and the osseous nasal cavity roof, 123(1) pyriform recess narrow (see the character description above), and 263(0) absence of a posterior coracoid emargination.

Polychrotiformes comb. nov.

(figs. 54A, 55A, 56A)

Definition

All taxa sharing a more recent common ancestor with Polychrus marmoratus than with Iguana iguana, Phrynosoma orbiculare, or Chamaeleo chamaeleon.

Diagnosis

The clade formed by Polychrotidae and Corytophanidae is joined by 47(0) presence of an angulated jugal, 174(1) presence of a groove extending anterior to the anterior surangular foramen, 260(1) clavicles proximally expanded and with a notch or fenestra, and 273(0) absence of a sternal fontanelle.

Comments

Polychrotiformes as defined here includes a monophyletic Polychrotidae (sensu Frost and Etheridge, 1989 and Conrad et al., 2007;  = “anoloids” of Etheridge and de Queiroz, 1988) and the Corytophanidae ( = “basiliscines” of Etheridge and de Queiroz, 1988). Frost et al. (2001) previously alluded to a close relationship between polychrotids and corytophanids. Macey et al. (1997) and Frost et al. (2001) suggested the paraphyly of polychrotids with respect to corytophanids, but the taxonomic sampling of these studies was extremely limited, raising questions about the “reclassification” of iguanians proposed therein.

Note that the possible corytophanid Geiseltaliellus is currently being redescribed and that this taxon may be important for more completely understanding the plesiomorphic morphology of corytophanids (Smith, 2004).

Polychrotidae (Fitzinger, 1843)

(figs. 54A, 55A, 56A)

Diagnosis

Seven unambiguous synapomorphies support polychrotid monophyly in this analysis. These are: 26(0) absence of strong medial processes of the maxilla posterior to the premaxillary nasal process, 252(3) caudal autotomy planes located anterior to the transverse processes (fig. 45D), 258(1) one or more pair(s) of postxiphisternal inscriptional ribs confluent at midline, 274(0) multiple xiphisternal branchings, 316(1) presence of peglike filamentous tongue papillae, 325(1) endolymphatic sacs extending into the nuchal musculature via the space between the supraoccipital and parietal, and 327(1) divided hemipeneal sulcus.

Comments

Recent molecular studies have suggested the non-monophyly/non-holophyly of Polychrotidae (Frost et al., 2001; Schulte et al., 2003), but morphological data continue to show support for polychrotid monophyly (Frost and Etheridge, 1989; Schulte et al., 2003; Conrad, 2005a; Conrad and Norell, 2007a; Conrad et al., 2007). These issues are discussed more fully by Schulte et al. (2003) and Conrad et al. (2007). Despite the presence of numerous unambiguous morphological synapomorphies supporting this clade here, I follow the relatively prudent approach exemplified by Schulte et al. (2003) and refrain from defining Polychrotidae. The name is used here in the same sense as Conrad et al. (2007); all descendants of the last common ancestor of the polychrotids identified by Frost and Etheridge (1989).

Hoplocercidae Frost and Etheridge, 1989

(figs. 54A, 55A, 56A)

Diagnosis

The current analysis recovers five unambiguous synapomorphies for this clade. These are 44(1) enlarged lacrimal foramen (see character description above), 60(1) a dorsoventrally inflated frontal, 190(0) splenial extending anteriorly for more than two-thirds the dentary tooth row, 258(1) one or more pair(s) of postxiphisternal inscriptional ribs confluent at midline, and 285(2) symphysial portion of the pubis more than half again as long as the tubercular portion.

Comments

This analysis includes Morunasaurus annularis, Hoplocercus spinosus, and a composite Enyalioides scored from E. palpebralis and E. laticeps. A recent study by Wiens and Etheridge (2003) with extensive sampling of hoplocercid taxa suggests the monophyly of Enyalioides. Even so, future versions of this analysis will eliminate the composite coding of Enyalioides and likely include more species of Morunasaurus. Importantly, the Adams consensus (figs. 55, 56) supports the topology of Wiens and Etheridge based on their analysis using mixed scaling of meristic characters (Wiens and Etheridge, 2003: fig. 4).

Given the absence of a fossil record for this clade and the somewhat limited sampling for it, I refrain from attaching a definition to this taxon name. The term “Hoplocercidae” is usually used to refer to all species within Enyalioides, Hoplocercus, and Morunasaurus; it is used here in that context.

Chamaeleontiformes comb. nov.

(figs. 54A, 55A, 56A)

Definition

All taxa sharing a more recent common ancestor with Chamaeleo chamaeleon than with Hoplocercus spinosus, Polychrus marmoratus, or Iguana iguana.

Diagnosis

This stem-based taxon diagnosed by the following unambiguous synapomorphies: 27(1) absence of midline contact of the maxillae behind the premaxillary nasal process, 90(0) absence of a postorbital tuberosity, 117(0) absence of ventromedial processes (basipterygoid buttresses) on the pterygoid, 118(2) absence of pterygoid teeth, and 189(0) a shortened splenial (see character description above).

Comments

Chamaeleontiformes, as defined here, is essentially equivalent to Chamaeleonidae as described by Frost and Etheridge (1989). Frost and Etheridge considered their Chamaeleonidae to be “equivalent to Acrodonta of Estes et al. (1988)” (Frost and Etheridge, 1989: 32), but explicitly included the Priscagama gobiensis and Priscagaminae as incertae sedis. Estes et al. (1988) defined Acrodonta as a crown group and in both Frost and Etheridge (1989) and, in the current analysis (figs. 16A, 17A, 18A), the Priscagama gobiensis-like taxa fall outside the radiation of acrodonts. Thus, Chamaeleontiformes is used to name the clade Frost and Etheridge (1989) recognized as Chamaeleonidae.

Isodontosaurus gracilis is a chamaeleontiform according to the current analysis. Isodontosaurus bears an unusual combination apomorphic features making it somewhat problematic for phylogenetic placement (Gao and Norell, 2000) and also make it a reasonable intermediate between “iguanids*” and higher chamaeleontiforms. A more complete morphological treatment may add further evidence to support this phylogenetic hypothesis.

A recent analysis by Conrad and Norell (2007a) identifies a dichotomy between chamaeleontiforms and pleurodontans. In that analysis Isodontosaurus gracilis is found to be a basal iguanomorph. Further investigation may help to sort out the differences between this analysis and that one.

Although Chamaeleo chamaeleon is used in the definition of this taxon, it does not appear in the phylogenetic analysis. However, the monophyly chamaeleonids has never been questioned, and Chamaeleonidae is consistently cited as an unmistakable natural group (see the section on taxon sampling above; also see Hillenius, 1978; Moody and Rocek, 1980; Rieppel, 1981b, 1987; Estes et al., 1988; Frost and Etheridge, 1989; Macey et al., 2000).

All of the chamaeleontiforms currently included in the analysis are from Africa, Asia, Australia, or Europe. The fossil priscagamids and the potential chamaeleontiform Arretosaurus ornatus (see below) are from Mongolia. Tinosaurus and Pseudotinosaurus were not included in the present analysis, but show some chamaeleontiform and/or acrodontan characteristics, including acrodont dentition with heterodonty (Marsh, 1872; Estes, 1983; Rage, 1987; Alifanov, 1993b; Augé and Smith, 1997; Li and Xue, 2002; Augé, 2003). The various species of Tinosaurus and Pseudotinosaurus are all poorly known; they are represented by fragmentary maxillae and dentaries that may or may not be diagnostic at the generic or specific levels and probably do not form monophyletic groups. Even so, Tinosaurus stenodon (Marsh, 1872) is significant in that it probably represents the only known American chamaeleontiform.

Priscagamidae + Acrodonta

(figs. 54A, 55A, 56A)

Diagnosis

An unnamed clade composed of chamaeleontiforms exclusive of Isodontosaurus is united by three unambiguous synapomorphies. These are 91(1) absence of a postfrontal, 214(1) presence of acrodont dentition, and 215(1) presence of caniniform teeth.

Priscagamidae Borsuk-Białynicka and Moody, 1984

(figs. 54A, 55A, 56A)

Definition

All taxa sharing a more recent common ancestor with Priscagama gobiensis than with Agama agama.

Diagnosis

Phrynosomimus asper, Mimeosaurus crassus, and Priscagama gobiensis are united in this analysis to the exclusion of all other chamaeleontiforms based on three unambiguous synapomorphies: 8(1) presence of dermal sculpturing on the maxilla, 50(1) dermal rugosities on the postorbital process of the jugal, and 97(1) postorbital extending ventrally for less than one-half the orbital margin.

Comments

Preliminary analyses (Conrad, 2005a) suggests that Arretosaurus ornatus shows some affinities with priscagamids. Inclusion of that taxon in a comprehensive analysis requires a re-visitation of its morphology.

Priscagama gobiensis + Mimeosaurus crassus

(figs. 54A, 55A, 56A)

Diagnosis

Priscagama gobiensis and Mimeosaurus crassus share two unambiguous synapomorphies to the exclusion of Phrynosomimus. These are 29(1) weakly inclined anterior margin of the maxillary nasal process and 117(1) ventromedial processes of the pterygoids.

Comments

Estes et al. (1988) identified three potential character states uniting Agamidae sensu lato. Of these, only the presence or absence of caniniform anterior maxillary and dentary teeth may be scored for priscagamids and only Mimeosaurus crassus shows the apomorphic state. Although this may indicate a relationship between Mimeosaurus crassus and agamids, the current analysis suggests priscagamid affinities.

Acrodonta Estes et al., 1988

(figs. 54A, 55A, 56A)

Definition

This taxon is defined here as all descendants of the last common ancestor of all extant taxa more closely related to Agama agama and Chamaeleo chamaeleon than to Gekko gecko, Varanus varius, Hoplocercus spinosus, Corytophanes cristatus, Polychrus marmoratus, or Iguana iguana. This is in keeping with the definition originally proposed by Estes et al. (1988), who defined Acrodonta as all a node attached to their metataxon Agamidae* and the clade Chamaeleonidae.

Diagnosis

The acrodontans included in this analysis are united by five unambiguous synapomorphies. These are 14(2) a bilobed premaxillary incisive process, 182(1) absence of a dentary shelf, 188(1) absence of a splenial, 261(0) straight (non-angulated) clavicle, and 272(1) a sternum that extends anteriorly, approaching the lateral arms of the interclavicle.

Comments

Acrodonta here is scored from only a few species (two chamaeleonids and three “agamids”). Future versions of this data matrix will sample more broadly from within acrodontans and analyze the apparent paraphyly of the “agamids” recovered here.

Scincogekkonomorpha Sukhanov, 1961

(figs. 54B, 55B, 56A)

Definition

All taxa sharing a more recent common ancestor with Gekko gecko and Scincus scincus than with Iguana iguana.

Diagnosis

Scincogekkonomorphs are united by the 41(3) absence of the lacrimal (with notable reversals), 86(1) ventral origin of the jaw adductor musculature on the parietal, 92(0) an anteroposteriorly elongate postfrontal component of the postorbitofrontal, 100(1) absence of a dorsal process on the squamosal, and 101(1) presence of a palpebral.

Comments

Scincogekkonomorpha is the sister taxon to Iguanomorpha in Sukhanov's (1961) taxonomic scheme. Scincogekkonomorpha has been used in some recent studies as a clade similar to or equivalent with Scleroglossa (Russell, 1988; Gao and Norell, 1998, 2000; Reynoso and Callison, 2000). Although Scleroglossa has been considered the sister taxon to iguana-like lizards, Estes et al. (1988) defined Scleroglossa as node-based taxon, anchoring it to extant taxa. Scincogekkonomorpha, as implemented here, includes scleroglossans and their extinct sister-taxa. Use of this name is especially appropriate given the current topology of relationships (figs. 54Figure 5556), showing that the stem of this clade is mostly made up of taxa previously considered stem-geckos (e.g. Bavarisaurus macrodactylus and Eichstaettisaurus schroederi) or variably considered scincomorphs or geckos (Ardeosaurus brevipes) (see below).

Eichstaettisaurus schroederi + Scandensia ciervensis + Scleroglossa

(figs. 55B, 56A)

Diagnosis

This clade (Scincogekkonomorpha exclusive of Ardeosaurus and Bavarisaurus) is diagnosed by 61(1) subolfactory processes of the frontal(s) partly surrounding the olfactory tract, 88(0) plesiomorphically possessing an anteroposteriorly short supratemporal, 97(1) postorbital extending ventrally for less than one-half the orbital margin, 260(1) clavicles proximally expanded and with a notch or fenestra, and 261(1) an angulated clavicle.

Comments

Ardeosaurus, Bavarisaurus, and Eichstaettisaurus (fig. 29D) all have been considered basal members of the gecko-lineage (see Hoffstetter, 1964, 1967b, Kluge, 1967, 1983, 1987, Estes, 1983). Recent analyses have suggested that these taxa have nothing to do with gekkotans, but instead form a paraphyletic assemblage near the base of Squamata and/or Iguania (Evans and Barbadillo, 1997, 1998, 1999, Evans and Chure, 1998b, Evans et al., 2005), or that they are close to scleroglossans (Conrad, 2004c; Conrad and Norell, 2006a). Evans and Wang (2005) re-analyzed the data matrices of Evans and Chure (1998b) and Evans and Barbadillo (1997, 1998, 1999) after adding some taxa. They recovered the basal positions for Bavarisaurus and Ardeosaurus, but found Eichstaettisaurus to be a carusioid anguimorph (Evans and Wang, 2005). The current analysis includes all of the taxa used in the most recent of these studies (Evans and Wang, 2005; Evans et al., 2005), but supports a phylogenetic hypothesis more similar to that of Evans et al. (2005) and Conrad and Norell (2006a) who suggest that Eichstaettisaurus is a basal scincogekkonomorph.

Scleroglossa Estes et al., 1988

(figs. 54B, 55B, 56A)

Definition

Estes et al. (1988) defined Scleroglossa as a node-based taxon including Gekkota and Autarchoglossa. I modify this definition here as follows: All descendants of the last common ancestor of Gekko gecko, Scincus scincus, and Anguis fragilis.

Diagnosis

The unnamed, unresolved clade containing Scandensia ciervensis, Gekkonomorpha (sensu Conrad and Norell, 2006a), and Evansauria (see below) is diagnosed by 65(1) frontals with an anterior maxillary contact and 269(1) presence of a single anterior process of the interclavicle.

Comments

The diagnosis of this clade is certainly influenced by the lack of resolution between its constituent groups and the relatively limited morphological understanding of Scandensia ciervensis (Evans and Barbadillo, 1998).

Gekkonomorpha Fürbringer, 1900b

(figs. 54B, 55B, 56A)

Definition

Following Conrad and Norell (2006a), Gekkonomorpha is defined as all taxa sharing a more recent common ancestor with Gekko gecko than Iguana iguana, Lacerta viridis, Scincus scincus, Anguis fragilis, or Varanus varius.

Diagnosis

Gekkonomorphs plesiomorphically share the following unambiguous synapomorphies: 79(1) posteromedial parietal flange, 131(1) occipital condyle bipartite and constructed primarily by the exoccipital portions of the otooccipitals, 136(2) presence of a distinct supratrigeminal process that anteriorly closes the trigeminal foramen, 144(1) sphenoid enclosing the lateral head vein, 151(1) anterior location for the spheno-occipital tubercle, and 207(1) retroarticular process that is posteriorly expanded (broadened).

Comments

Cope (1900) and Fürbringer (1900a) used Nyctisaura to receive Gekkoninae and Eublepharinae. Lee (1998) recently redefined Nyctisaura such that it applies to a clade containing Xantusiidae and Gekkota to the exclusion of Rhynchocephalia, Iguanidae, Agamidae, Leiolepis, Uromastyx, Chamaeleonidae, Lacertidae, Teiidae, Gymnophthalmidae, Cordylidae, Scincidae, Anguidae, Xenosauridae, Heloderma, Lanthanotus borneensis, and Varanus. If not for this latter definition, Nyctisaura might be invoked for the clade referred to here as Gekkonomorpha. Gekkonomorpha is a name that is sometimes used informally to refer to extant geckos (Withers et al., 2000; Seligmann, 2002; Werner et al., 2002, 2004, 2005; Persaud et al., 2003; Gehr and Werner, 2005) and has been used more formally as well (Fürbringer, 1900b; Conrad and Norell, 2006a). Thus, this name is available and useful as a name for the stem-based group including geckos and their fossil relatives.

Parviraptor + Gobekko cretacicus + Gekkota

(figs. 54B, 55B, 56A)

Diagnosis

The current analysis suggests that a clade formed by Parviraptor, Gobekko cretacicus, and Gekkota may be diagnosed by a single unambiguous synapomorphy: 108(1) vomer-palatine contact very narrow mediolaterally.

Comments

The placement of Parviraptor (including P. estesi and P. cf. estesi of Evans, 1994a) within Gekkonomorpha is somewhat unexpected. Parviraptor was originally considered to be an anguimorph with platynotan affinities (Evans, 1994a; Evans and Wang, 2005). Recent studies have suggested that this is, instead, a basal scincogekkonomorph (Evans et al., 2005) or a basal member of the autarchoglossan lineage (Conrad and Norell, 2006a). Discovery of more complete remains of this intriguing animal would be a windfall for squamate anatomy and systematics.

Parviraptor Evans, 1994a

(figs. 54B, 55B, 56A)

Diagnosis

Parviraptor estesi and P. cf. estesi are joined in this analysis by two unambiguous synapomorphies. These are 83(2) presence of an expansive nuchal fossa that extends well onto the parietal table and 214(2) “modified pleurodont” dentition (sensu Zaher and Rieppel, 1999a).

Comments

I refrain from defining the clade Parviraptor as constituted here, preferring to leave that to the original describer of the specimens included in it. Parviraptor estesi and P. cf. estesi, as described by Evans (1994a), are each represented by incomplete remains that may be diagnosed from one another and probably separated by several million years (Evans, 1994a). However, the present analysis supports Evans' (1994a) suggestion that these two taxa form a clade.

Gobekko cretacicus + Gekkota

(figs. 54B, 55B, 56A)

Diagnosis

Gekkota and Gobekko cretacicus are united by 77(3) absence of a pineal foramen, 159(3) quadrate suspended mostly from the opisthotic, and 181(2) closed and fused Meckel's canal.

Comments

The phylogenetic placement of Gobekko cretacicus has been cladistically analyzed only for the first time recently. A preliminary analysis of its relationships suggested the placement of this taxon as the sister-group to Gekkota (Conrad, 2005a), a view supported here. However, another recent study has suggested that Gobekko cretacicus falls within the gekkotan crown group (Conrad and Norell, 2006a), with the implication that crown-group gekkotans have been distinct since the Cretaceous. Gobekko cretacicus is known from an incomplete skull. A new study of the specimen and/or the discovery of more complete remains might help to more confidently place this animal in the context of gekkonomorph evolution.

Gekkota Cuvier, 1817

(figs. 54B, 55B, 56A)

Definition

The most recent ancestor of extant taxa more closely related to Gekko gecko than to Xantusia vigilis, Scincus scincus, or Varanus varius and all descendants of that ancestor.

Diagnosis

The present analysis yields two unambiguous synapomorphies for Gekkota, including 55(1) fused frontals and 80(1) short parietal supratemporal processes (see character description above).

Comments

Estes et al. (1988) defined Gekkota as all descendants of the common ancestor of Gekkonidae and Pygopodidae. However, they considered Gekkonidae to include all limbed gekkotans. The revised definition used here maintains the apparent intention of Estes et al. (1988) to include all extant geckos and pygopods, but does not make the same assumption about the primary gekkotan dichotomy.

Bauer et al. (2005) recently described an Eocene gecko preserved in Baltic Amber that represents the earliest known member of the crown gekkotan radiation. However, this taxon, Yantarogekko balticus, is not included in the current analysis because the current morphological data set is not sensitive enough to discriminate between Y. balticus and other gekkonids based on external morphology. Even so, the presence of Y. balticus is acknowledged in fig. 56A.

Gekkonidae + Eublepharidae

(figs. 54B, 55B)

Diagnosis

Gekkonids and eublepharids are joined in this analysis by only two unambiguous synapomorphies. These are 73(1) fused parietals and 284(0) pelvic elements co-ossified (sutures lost).

Comments

Numerous hypotheses for the interrelationships of the gekkotan clades have been put forward. Traditionally, a dichotomy has been described between limbed and limbless gekkotans (Kluge, 1967, 1983; Estes et al., 1988; Wu et al., 1996; Lee, 1998). However, Kluge (1987) suggested that the basal gekkotan dichotomy lay between eublepharids and all other gekkotans, with diplodactylines being the sister group to pygopodines. The latter hypothesis has been supported by one recent morphological study of gekkonomorph relationships (Conrad and Norell, 2006a). However, a recent molecular analysis (Han et al., 2004) with extensive taxonomic sampling offers some support for a new hypothesis in which eublepharids and gekkonids form the sister-clade to a Pygopodidae sensu Kluge (1987; see below). A recent analysis of diplodactylines raises questions about the monophyly of that group as it has been traditionally conceived (Donnellan et al., 1999). Importantly, the present phylogenetic analysis offers the same topological hypothesis for extant “families” as Han et al. (2004). Future versions of this data matrix will include more extensive sampling of gekkonids.

Gekkonidae Gray, 1825

(figs. 54B, 55B)

Definition

All taxa sharing a more recent common ancestor with Gekko gecko than with Pygopus lepidopus or Eublepharis hardwickii.

Diagnosis

The current analysis reveals six unambiguous synapomorphies for a clade minimally containing Teratoscincus microlepis, Gekko gecko, Gonatodes albogularis, and Pachydactylus bibronii. These are 26(1) the presence of strong medial processes of the maxillae posterior to the nasal process of the premaxilla, 104(1) presence of a contact between the septomaxilla and the osseous roof of the nasal capsule, 106(1) fused vomers, 113(0) absence of a distinct choanal groove on the palatine, 242(1) presence of seven or fewer cervical vertebrae, and 325(3) endolymphatic sacs extending into the nuchal musculature through the vagus foramen.

Comments

Note that Gekkonoidea as used by Kluge (1987) and defined by Conrad and Norell (2006a) is taxonomically equivalent to Gekkonidae as defined here and with the present phylogenetic hypothesis. However, as discussed earlier, this result is acceptable to keep the meaning of the name consistent.

Conrad and Norell (2006a) suggested that Teratoscincus was closely related to pygopodids, but cautioned that a more inclusive analysis might reveal a different hypothesis. Indeed, the current analysis places Teratoscincus microlepis in a more traditional position as a member of the Gekkonidae. However, I will echo the cautionary statement of Conrad and Norell (2006a) in warning that the current sampling of gekkonids is relatively sparse and that more complete sampling is in order for future versions of this analysis. With that in mind, I will not further discuss gekkonid interrelationships here.

Eublepharidae Boulenger, 1883

(figs. 54B, 55B)

Definition

All taxa sharing a more recent common ancestor with Eublepharis hardwickii than with Gekko gecko or Pygopus lepidopus.

Diagnosis

The eublepharid taxa included in this analysis are joined by four unambiguous synapomorphies. These are 32(1) presence of a palatine flange of the maxilla, 36(1) pares frontales contact at midline, 74(2) frontal tabs of the parietal extending anteriorly onto the ventral surface of the frontal, and 211(2) teeth closely packed together (ctenodont).

Comments

Eublepharis hardwickii, the type species for the type genus of Eublepharidae, is used in the definition of this group, even though Eublepharis does not appear in the present analysis. Morphological (Grismer, 1988) and molecular (Ota et al., 1999; Han et al., 2004) studies recover the monophyly of a clade including Eublepharis, Aeluroscalabotes felinus, Coleonyx mitratus, and Hemitheconyx caudicinctus. Thus, E. hardwickii is presumed to be a member of the clade represented by A. felinus, C. mitratus, and H. caudicinctus in the present analysis. This relationship will be further analyzed elsewhere.

Pygopodidae Gray, 1845

(figs. 54B, 55B)

Definition

The most recent ancestor of extant taxa more closely related to Pygopus lepidopodus than to Gekko gecko or Eublepharis hardwickii and all descendants of that ancestor.

Diagnosis

This analysis recovers six unambiguous synapomorphies for Pygopodidae, including 14(2) bilobed premaxillary incisive process, 22(1) absence of anterolateral processes on the nasals, 39(0) absence of a prefrontal-jugal contact, 114(1) presence of ventromedial fold on the palatine partly hiding the choanal groove in ventral view (rudimentary secondary palate), 185(0) dentary without a notch distinguishing coronoid and surangular processes, and 234(1) moderately oblique vertebral condyles and cotyles (see character description above).

Comments

The current definition is a slight modification of that proposed by Kluge (1987) such that the anchor taxa are more specific. This is a more taxonomically inclusive definition than that of Estes et al. (1988) and includes diplodactyline geckos.

Pygopodinae Gray, 1845

(figs. 54B, 55B)

Definition

All taxa sharing a more recent common ancestor with Pygopus lepidopodus than with Carphodactylus laevis or Diplodactylus vittatus.

Diagnosis

In the present analysis, pygopodines are diagnosed by the following unambiguous synapomorphies: 21(1) nasals that are in contact for less than half their length in dorsal view, 33(1) a maxillary tooth row terminating at the level of the anterior border of the orbit, 57(0) anterior and posterior borders of the frontal of subequal width, 136(0) absence of a supratrigeminal process on the prootic, 143(2) fusion of the sphenoid and basioccipital, 151(0) posterior positioning of the spheno-occipital tubercle, 180(1) elongate lateral portion of the dentary extending along the mandible to a point at least half way between the coronoid eminence and the mandibular glenoid, 207(0) no posterior broadening of the retroarticular process, 230(1) notochordal canal obliterated by centrum ossification, 303(1) deeply imbricating scales, and 340(1) presence of an unexpanded head of the M. pseudotemporalis profundus.

Evansauria tax. nov.

(figs. 54B, 55B, 56B)

Definition

All taxa sharing a more recent common ancestor with Lacerta viridis and Varanus varius than with Gekko gecko or Iguana iguana.

Etymology

Named in honor of Susan E. Evans and her extensive work on the evolutionary history and systematics of lepidosaurs.

Diagnosis

This taxon is diagnosed by 76(2) presence of an anteroposteriorly narrow decensus parietalis, 146(1) posterior opening of the Vidian canal located on the sphenoid-prootic suture, 185(0) dentary without a notch distinguishing coronoid and surangular processes, 209(1) presence of a tubercle on the medial margin of the retroarticular process, 230(1) notochordal canal obliterated by centrum ossification, 234(1) moderately oblique vertebral condyles and cotyles (see character description above), and 257(0) no expansion of the anterior presacral ribs.

Comments

This taxon has been unidentified until now, mostly because most of the basal members of the clade have not previously appeared in cladistic analyses together. Importantly, each of the non-autarchoglossan evansaurs included here have been allocated to the clades Gekkota (Myrmecodaptria microphagosa), Lacertoidea (Eoxanta lacertifrons lacertifrons, Globaura venusta), Scincoidea (Parmeosaurus scutatus multivagus, Tepexisaurus tepexii), or Anguidae (Bainguis parvus) (see Borsuk-Białynicka, 1988; Gao and Hou, 1996; Gao and Norell, 1998, 2000; Reynoso and Callison, 2000).

Conrad and Norell (2006a) suggested that Myrmecodaptria microphagosa is a basal member of the autarchoglossan lineage; a hypothesis supported here, but in a slightly different phylogenetic context. Wu et al. (1996) suggested that Globaura venusta and Eoxanta lacertifrons lacertifrons are basal scincoids (contra Borsuk-Białynicka, 1988).

Bainguidae Borsuk-Białynicka, 1984

(figs. 54B, 55B, 56B)

Definition

All taxa sharing a more recent common ancestor with Bainguis parvus than with Lacerta viridis, Scincus scincus, or Varanus varius.

Diagnosis

Bainguids, as defined here, are united by five unambiguous synapomorphies, including 53(1) presence of a jugal-squamosal contact, 96(2) postorbital extending posteriorly for more than 3/4 the length of the supratemporal fenestra, 121(1) midline contact of the palatines, 193(1) coronoid labial flange present, but not greatly developed posteriorly, and 203(1) presence of a prearticular crest.

Comments

Bainguidae was originally erected as a monospecific “family” described as possessing both scincomorph and anguimorph affinities, but conservatively placed as a member of a “preanguimorphan grade” (Borsuk-Białynicka, 1984). Later, Bainguis parvus was shown to have affinities with Anguidae (Borsuk-Białynicka, 1991; Gao and Hou, 1996; Gao and Norell, 1998) and Bainguidae was synonymized with Anguidae (Alifanov, 2000). Gao and Norell (1998) alone have cladistically analyzed the position of Bainguis parvus. They provided evidence of anguine affinities for Bainguis, but cautioned that this poorly known taxon could not be scored for many of the anguid characters (Gao and Norell, 1998).

The current analysis is the first to include Bainguis parvus in a cladistic analysis with non-anguimorph fossils such as Myrmecodaptria and Eoxanta lacertifrons. Thus, this is the first analysis in which Bainguidae, as it is currently used, had the potential to be discovered. Given the relative completeness of our knowledge of B. parvus, prudence is exercised in the current definition of Bainguidae.

Note that the phylogenetic position of Bainguidae within Autarchoglossa is not strongly supported. A secondary analysis of only osteological characters (see below) suggested that bainguids are lacertoids. Further analysis of the position of Bainguidae is appropriate given this result and the convergences between this clade and the plesiomorphic condition for Scincomorpha and some of its constituent clades.

Bainguis parvus + Eoxanta lacertifrons + Myrmecodaptria

(figs. 54B, 55B)

Diagnosis

Bainguids exclusive of Globaura venusta are united by 143(1) the presence of posterolateral sphenoid flanges ventrolaterally overlying the basioccpipital and 178(0) a dentary that is ventrally convex along its long axis.

Eoxanta lacertifrons + Myrmecodaptria

(figs. 54B, 55B)

Diagnosis

This unnamed clade is united to the exclusion of other bainguids within Bainguidae by 209(0) the absence of a tubercle on the medial surface of the retroarticular process.

Parmeosaurus scutatus + Autarchoglossa

(figs. 54B, 55B, 56B)

Diagnosis

Parmeosaurus scutatus and autarchoglossans share two unambiguous synapomorphies: 108(1) vomer-palatine contact very narrow mediolaterally and 143(2) fusion of the sphenoid and basioccipital.

Autarchoglossa Wagler, 1830

(figs. 54C, 55C, 56B)

Definition

Lacerta viridis, Scincus scincus, Anguis fragilis, and all descendants of their last common ancestor.

Diagnosis

The current analysis recovers five unambiguous autarchoglossan synapomorphies: 10(1) presence of dermal sculpturing on the frontal and parietal, 11(1) premaxilla fused (unpaired), 47(0) presence of an angulated jugal, 48(0) presence of a posteroventral process on the jugal, and 166(1) articular-prearticular fused to the surangular.

Comments

Estes et al. (1988) defined this clade phylogenetically following the general traditional usage of the name (e.g., Wagler, 1830; Camp, 1923); that definition is only made more precise here by implementing more specific anchor taxa.

Scincomorpha Camp, 1923

(figs. 54C, 55C, 56B)

Definition

All taxa sharing a more recent common ancestor with Scincus scincus than with Gekko gecko, Anguis fragilis, or Varanus varius.

Diagnosis

This clade is diagnosed by 114(1) presence of ventromedial fold on the palatine partly hiding the choanal groove in ventral view (rudimentary secondary palate), 285(2) symphysial portion of the pubis more than half again as long as the tubercular portion, 306(1) presence of compound osteoderms dorsally 308(1) and ventrally, 319(2) keratinized tongue with mushroom-shaped foretongue in cross section, 323(2) inner ear with more than one-half of the hair cells inert, 327(1) divided hemipeneal sulcus, and 360(0) absence of an aponeurosis for the M. branchiohyoideus.

Comments

Although Estes et al. defined Scincomorpha as “[t]he last common ancestor of Scincidae, Cordylidae, Xantusiidae, Lacertidae, Teiidae, and Gymnophthalmidae, and all of its descendants” (Estes et al., 1988: 207), the definition is expanded here to include stem taxa. Such is more similar to the original usage of Camp (1923) who hypothesized that Amphisbaenia is the sister taxon to scincomorphs sensu Estes et al. (1988), but included them in his Scincomorpha (see fig. 1). Scincomorpha is the sister group of Anguimorpha within the node-based Autarchoglossa. The name Leptoglossa (Cope, 1900) is available for the crown group scincomorph node, is suited for that purpose in identifying a soft-tissue characteristic, and has been implied as such in the past (Romer, 1956; Hu et al., 1984; Carroll, 1988b). Although the constituent taxa of that clade is the same as Scincomorpha in the present study's topology and Leptoglossa is not used or defined here, it is suggested for potential future use at that node.

Estes et al. (1988) considered Scincomorpha to be a dichotomy between Scincoidea and Lacertoidea, but scincomorph monophyly has been challenged more recently (Lee, 1998, 2000, 2005a, 2005b; Lee and Caldwell, 2000; Townsend et al., 2004; Vidal and Hedges, 2005). Lee (1998) and Lee and Caldwell (2000) suggested that scincomorphs are paraphyletic with respect to both a gekkotan-amphisbaenian-dibamid clade and to anguimorphs, representing three or four different lineages (Xantusiidae with “Annulata”, Lacertiformes, Scincidae, and Cordylidae) (figs. 6, 7), although Lee (2005a) later found morphological support for a Scincidae-Cordyliformes clade. Molecular studies (Townsend et al., 2004; Vidal and Hedges, 2004, 2005) (fig. 10) and combined morphological-molecular analyses (Lee, 2005a, b) also suggest the non-monophyly of Scincomorpha. Importantly, Xantusiidae has been placed near Gekkota (Lee, 1998, 2000, 2005a, b; Lee and Caldwell, 2000) (figs. 6, 7), with Lacertiformes (Estes et al., 1988; Caldwell, 1999a) (figs. 2, 8), or associated with Scincidae and/or Cordyliformes (Wu et al., 1996; Evans and Barbadillo, 1997, 1998; Townsend, 2002; Vicario et al., 2003; Townsend et al., 2004; Vidal and Hedges, 2004, 2005) (figs. 3, 4, 10).

The taxonomy laid out for scincomorphs here is sensitive to the varying hypotheses of scincomorph interrelationships. The basic dichotomy recovered in this analysis is between a clade containing Slavoia darevskii and Scincidae sensu lato and a clade containing Cordyliformes, Xantusiidae, Lacertiformes, and several fossil relatives of these taxa. I apply names to three major stem-based clades below: Scincoidea, Cordyloidea, and Lacertoidea. Cordyloidea may become a subjective junior synonym of Scincoidea if the two are shown to form a clade exclusive of lacertoids, but Cordyloidea will never fall within Lacertoidea (see the definitions of these taxa below). This would more-or-less follow the usage of Estes et al. (1988). Xantusiids may move around the tree without disrupting the general meaning of any of the scincomorph group names, or of Gekkonomorpha (see above). Although I have some confidence in clades described herein, I do not delude myself by thinking that alternative phylogenetic hypotheses are impossible or that they will not be proposed. The taxonomic framework offered here will allow continued communication about the groups to which they refer even in different phylogenetic contexts.

Lacertoidea + Cordyloidea

(figs. 54C, 55C, 56B)

Diagnosis

Lacertoids and cordyloids are united in this analysis by 32(0) absence of a maxillary palatine flange, 80(1) short parietal supratemporal processes (see character description above), 96(2) postorbital extending posteriorly for more than 3/4 the length of the supratemporal fenestra, 251(1) double, divergent caudal transverse processes, 269(0) absence of an anterior process of the interclavicle, and 294(1) presence of large pores on the scales anterior to the cloaca and/or on the thigh.

Comments

Estes et al. (1988) suggested a close relationship between scincids sensu lato and Cordyliformes ( = Cordylidae of their usage and sensu lato) and found Xantusiidae to be more closely related to Lacertiformes. Recent molecular work has also suggested a close relationship between cordyliforms and scincids sensu lato, but further suggests that xantusiids are the sister-taxon to cordyliforms (Vicario et al., 2003; Townsend et al., 2004; Vidal and Hedges, 2005).

The current result of a lacertoid-cordyloid clade is somewhat unexpected. This may partly explain the molecular similarities between cordyliforms and xantusiids. Inclusion of fossil cordyloids and xantusiids also help to reveal this relationship by modifying character polarities and contributing to reconstruction of hypothetical ancestors. However, because of the consistent lack of consensus regarding the relationshiops of cordyloids, xantusiids, and lacertiforms to one another and to other squamate groups, it would not be surprising if further investigations with additional character data provided an hypothesis different from this one. With this in mind, no name is applied to this grouping of lacertoids and cordyloids.

Cordyloidea Fitzinger, 1826

(figs. 54C, 55C, 56B)

Definition

All taxa sharing a more recent common ancestor with Cordylus cordylus than Lacerta viridis or Anguis fragilis.

Diagnosis

Cordyloids share three unambiguous synapomorphies: 8(1) presence of dermal sculpturing on the maxilla, 166(0) absence of fusion between the articular-prearticular and the surangular, and 219(1) striated tooth crowns.

Comments

Cordylidae sensu lato is now typically split into two “families” (Cordylidae sensu stricto and Gerrhosauridae), collectively referred to as Cordyliformes (Lang, 1991; Harvey and Gutberlet, 1995; Mouton and Wyik van, 1997; Cooper and Steele, 1999; Odierna et al., 2002; Lamb et al., 2003). The monophyly of a cordylid-gerrhosaurid clade exclusive of other extant taxa is unquestioned, rendering the usage of Cordyliformes or Cordylidae sensu lato somewhat semantic. Regardless, I use Cordyliformes here for the crown group, following recent usage, and employ Fitzinger's (1826) term Cordyloidea for the stem-based taxon including cordyliforms and their proximal fossil outgroups.

The current analysis is unable to resolve the relationships between cordyliforms, Ornatocephalus metzleri, and Sakurasaurus shokawensis. An analysis by Weber (2004) placed O. metzleri in an unresolved trichotomy with cordyliforms and scincids, with paramacellodids as the outgroup. The current analysis does not recover a monophyletic scincoid clade as conceptualized by Weber (2004), but does support the relationship between O. metzleri and cordyliforms. Evans and Manabe (1999) were prudent in their placement of S. shokawensis, suggesting only that it belonged within Scincomorpha; a hypothesis consistent with the current topology. The possibility that one or both of these fossil taxa belong within Cordyliformes will be analyzed elsewhere.

Lacertoidea Camp, 1923

(figs. 54C, 55C, 56B)

Definition

All taxa sharing a more recent common ancestor with Lacerta viridis than with Scincus scincus, Cordylus cordylus ( = Zonurus cordylus), or Anguis fragilis.

Diagnosis

Lacertoids are diagnosed by 77(3) absence of a pineal foramen (with notable apparent reversals), 110(0) absence of anteromedial “step” to the palatine, 203(1) presence of a prearticular crest, 205(0) retroarticular process lacking medial deflection, 209(0) absence of a tubercle on the medial surface of the retroarticular process, 210(0) retroarticular process without torsion, 212(4) teeth with divided crowns/cusps, 303(0) body scales not deeply imbricating, 305(0) dorsal body osteoderms absent, 307(0) ventral body osteoderms absent, 337(0) M. adductor mandibulae posterior extending only to the posterior margin of Meckel's canal, and 341(1) origin of the M. pseudotemporalis superficialis extending onto the posterior one-third of the margin of the supratemporal fenestra.

Comments

Estes et al. (1988) gave this taxon name a node-based definition including Xantusiidae and lacertiform taxa. Although the current topology would work with that definition and maintain the integrity of the taxon name as intended by Estes et al. (1988), Xantusiidae remains a problematic clade with a limited fossil record (Hecht, 1956; Estes, 1983). The relationships of xantusiids have been questioned and recent studies have suggested that their affinities lie with gekkotans, amphisbaenians, and dibamids (Lee, 1998, 2000, 2005a, 2005b; Lee and Caldwell, 2000) or with scincoids (Wu et al., 1996; Evans and Barbadillo, 1997, 1998; Vicario et al., 2003; Vidal and Hedges, 2004, 2005). If the former hypothesis is correct, then Lacertoidea sensu Estes et al. (1988) would become synonymous with Scleroglossa; if the latter is correct, then it would become synonymous with Scincomorpha sensu Estes et al. (1988). Even though the current topology supports the placement of Xantusiidae with many unambiguous synapomorphies, I find it prudent to modify the definition of Lacertoidea to allow for differing topologies with minimal disruption of taxonomy (see Scincomorpha above).

Xantusiidae Baird, 1859

(figs. 54C, 55C, 56B)

Definition

Xantusia vigilis, Cricosaura typica, Lepidophyma flavimaculatum, Palaeoxantusia kyrentos, and all descendants of their last common ancestor.

Diagnosis

Xantusiids, as defined here, are united by 89(0) contact between supratemporal arch bones and frontal and parietal unforked, 93(2) postfrontal contacting the parietal for more than one-half the length of the parietal table, 94(1) fused postorbitofrontal, 114(0) no development of the secondary palate, 126(1) ectopterygoid exposed on the lateral skull surface posterior to the maxilla, 180(1) elongate lateral portion of the dentary extending along the mandible to a point at least half way between the coronoid eminence and the mandibular glenoid, 181(2) closed and fused Meckel's canal, and 189(1) splenial not extending anterior to the midpoint of the dentary tooth row.

Xantusia + Cricosaura + Lepidophyma

(figs. 54C, 55C)

Diagnosis

The extant xantusiids included in this analysis are united by seven unambiguous synapomorphies. These are 41(2) fusion of the lacrimal and prefrontal, 47(1) jugal curved (rather than angulated), 88(1) an elongate supratemporal, 97(0) postorbital with a robust descending process contributing at least one-half the posterior orbital border, 132(1) crista prootica extending anteriorly onto the basipterygoid process, 183(1) dentary forming the dorsal border of the anterior inferior alveolar foramen, and 202(1) absence of an angular.

Xantusia + Cricosaura

(figs. 54C, 55C)

Diagnosis

According to the current analysis, Xantusia and Cricosaura share one unambiguous synapomorphy uniting them to the exclusion of Lepidophyma: 48(1) posteroventral process of jugal absent.

Comments

Morphological data have previously suggested a sister taxon relationship between Cricosaura and Lepidophyma. Molecular data suggest a sister-group relationship between Xantusia and Lepidophyma (Hedges et al., 1991; Hedges and Bezy, 1993; Vicario et al., 2003). Recovery of a novel phylogenetic topology here may be related to the broader sampling of outgroup taxa compared to previous analyses. Even so, the composite coding of Xantusia in this matrix will be eliminated in future versions of this analysis to further analyze the current phylogenetic hypothesis.

Lacertiformes Estes et al., 1988

(figs. 54C, 55C, 56B)

Definition

Lacerta viridis, Teius teyou, and all descendants of their last common ancestor.

Diagnosis

Lacertiforms are united by 52(0) jugal lying mostly dorsal (rather than medial) to the maxilla, 160(0) presence of a pterygoid lappet on the quadrate, 169(1) expanded adductor fossa, 184(2) absence of a posterodorsal coronoid process on the dentary, 198(0) absence of a posterior overlap of the coronoid by the dentary, 287(0) presence of an anterior iliac process, 317(3) foretongue notched for more than 40 percent of its length, and 323(1) ciliary restraint system for the inner ear composed of tectorial and sallet systems.

Comments

The definition applied to this taxon name follows that of Estes et al. (1988), but substitutes more specific anchor taxa. Although most recent phylogenetic analyses recover a clade containing Lacertidae and Teiioidea to the exclusion of other extant taxa, at least two do not (Evans and Barbadillo, 1998, 1999). Both of these studies suggest that lacertids are more closely related to a clade containing xantusiids, Paramacellodus, scincoids, and cordylids than to teiioids. In this case, Lacertiformes includes scincoids and cordyloids and possibly would be considered a synonym of Leptoglossa (see Scincomorpha above).

Note that Chamops segnis is problematic. Observation of the principle trees saved from the analysis reveals that C. segnis may be the immediate outgroup of Lacertiformes, the sister-taxon to Lacertidae, a basal teiioid, the sister-taxon to Gymnophthalmidae, a basal macroteiid, the sister-taxon to Polyglyphanodontidae, or the sister-taxon to Teiidae. Chamops segnis may only be scored for characters 52 and 287 among those contributing to the lacertiform diagnosis and shares the lacertiform condition in both. A more complete understanding of this taxon may help to resolve this issue. Prototeius stageri is currently known from various skull bits associated based on size and proximity (Denton and O'Neill, 1995). Further study or more complete (articulated) remains are necessary to demonstrate that the known specimens belong to a single species.

Teiioidea Estes et al., 1988

(figs. 54C, 55C, 56B)

Definition

All taxa sharing a more recent common ancestor with Teius teyou than with Lacerta viridis.

Diagnosis

Teiioids are diagnosed by 80(0) elongate supratemporal processes (see character description above), 84(0) no increased contact between the supratemporal arches and the parietal, 121(1) midline contact of the palatines, 204(1) prearticular crest with imbedded angular crest, 263(1) a posterior coracoid emargination, 277(2) absence of an ectepicondylar groove and foramen, 317(4) foretongue notched for more than 50 percent of its length, 333(1) peroneal nerve absent, and 347(2) absence of the urinary bladder.

Comments

Estes et al. (1988) defined this clade as a crown, node-based taxon anchored to teiids and gymnophthalmids. That definition is equivalent to Boulenger's (1885–1887) conception of Teiidae. Upon considering fossil taxa and implementing cladistic methodology, Presch (1983) questioned the holophyly of Teiidae sensu Boulenger and raised Gymnophthalminae to “family” rank, a convention followed by many subsequent studies (e.g., Estes, 1983; Estes et al., 1988; Presch, 1988; Schwenk, 1988; Kizirian, 1996; Evans and Barbadillo, 1998; Kizirian and McDiarmid, 1998; Lee, 1998; Kizirian and Cole, 1999; Reynoso and Callison, 2000; Montero et al., 2002; Bell et al., 2003). The definition of Teiioidea is slightly expanded here to include fossil stem taxa that may be more closely related to teiids and gymnophthalmids than to lacertids.

Macroteiida tax. nov.

(figs. 55C, 56B)

Definition

Teius teyou, Polyglyphanodon sternbergii, and all descendants of their last common ancestor.

Etymology

Macro, (Latin) large; “-teiida” referring to the Teiidae, the extant radiation of macroteiids. Referring to “macroteiids”, the common name given to extant, non-gymnophthalmid, teiioids (e.g., Presch, 1974, 1976, 1983, 1988; Vitt, 1982; Estes, 1983; Krause, 1985; Estes et al., 1988; Schwenk, 1988; White, 1990).

Diagnosis

Macroteiids form a clade diagnosed by 66(0) prefrontals with large contributions to the orbitonasal fenestra, 92(2) postfrontal developed as a mediolaterally elongate bar forming the anterior margin of the supratemporal fenestra, 114(0) no development of the secondary palate, 124(1) ectopterygoid contacting the palatine in the suborbital fenestra, 166(0) absence of fusion between the articular-prearticular and the surangular, and 269(1) presence of a single anterior process of the interclavicle.

Comments

The informal term “macroteiid” is formalized here as Macroteiida to encompass not only crown group teiids, but to include all of the known larger-bodied teiioids (fig. 55).

Teiidae Gray, 1827

(figs. 54C, 55C, 56B)

Definition

Teius teyou, Tupinambis teguixin, and all descendants of their last common ancestor.

Diagnosis

Teiidae share five unambiguous synapomorphies to the exclusion of polyglyphanodontids: 8(1) presence of dermal sculpturing on the maxilla, 100(0) presence of a dorsal process on the squamosal, 240(0) well-developed atlantal lateral processes, 243(3) cervical intercentra fused to the succeeding vertebrae, and 285(1) symphysial and tubercular portions of the pubis of subequal length.

Comments

Estes et al. defined Teiidae as “the last common ancestor of the Teiinae and Tupinambinae … and all organisms sharing a more recent common ancestor with these taxa than with any other extant organisms” (Estes et al., 1988: 215). Polyglyphanodontids would be considered teiids under that definition. Indeed, polyglyphanodontids have been considered part of the Teiidae in the past (Estes, 1983; Presch, 1983; Gao and Norell, 2000; Nydam and Cifelli, 2005). However, Polyglyphanodon-like teiioids have also often been considered to constitute up to four separate “families” from teiids; Adamisauridae, Gilmoreteiidae ( = Macrocephalosauridae; Langer, 1998), Mongolochamopidae (see Alifanov, 2000), and Polyglyphanodontidae (see Gilmore, 1942b; Sulimski, 1972, 1975, 1978; Alifanov, 1993a, 2000; Langer, 1998). Because of this precedent and because polyglyphanodontids sensu lato (see usage below) do not fall within the crown group Teiidae, I use the above definition for Teiidae.

Polyglyphanodontidae Gilmore, 1942b

(figs. 54C, 55C, 56B)

Definition

Adamisaurus magnidentatus, Gilmoreteius ferrugenous, Gobinatus arenosus, Polyglyphanodon sternbergii, and all descendants of their last common ancestor.

Diagnosis

Polyglyphanodontids are diagnosed by 9(0) absence of dermal sculpturing on the prefrontal, 77(0) pineal foramen lying within the parietal, 119(0) presence of a pterygoid-vomer contact, 122(1) pterygoids contact at midline, 213(1) marginal teeth waisted, 261(0) straight (non-angulate) clavicle, and 266(1) secondary scapular fenestra present.

Comments

Estes (1983) and Presch (1983) considered the taxa contained here in Polyglyphanodontidae to be a subclade of Teiidae (see comments above for Teiidae). New data regarding the dentition of Polyglyphanodon sternbergii from an excellent study by Nydam and Cifelli (2005) will be incorporated into future versions of this analysis.

Gobinatus arenosus + Tchingisaurus multivagus

(figs. 54C, 55C)

Diagnosis

Gobinatus arenosus and Tchingisaurus multivagus form a clade exclusive of other polyglyphanodontids based on the presence of three unambiguous synapomorphies: 83(1) presence of a nuchal fossa on the parietal, 100(0) presence of a dorsal process on the squamosal, and 182(0) presence of a subdental shelf.

Comments

The principle trees recovered in this analysis revealed two competing hypotheses for the placement of this clade; as the sister group to Adamisaurus magnidentatus magnidentatus, or as the sister taxon to a clade containing Erdenetesaurus robinsonae and Cherminsaurus kozlowskii. A forthcoming revision of the morphology of Adamisaurus magnidentatus may help resolve these interrelationships.

Polyglyphanodontinae Gilmore, 1942b

(figs. 54C, 55C)

Definition

All taxa sharing a more recent common ancestor with Polyglyphanodon sternbergii than with Adamisaurus magnidentatus or Teius teyou.

Diagnosis

The current analysis reveals five unambiguous synapomorphies uniting polyglyphanodontines as defined here. These are 39(1) presence of a jugal-prefrontal contact, 47(1) jugal curved (rather than angulated), 58(0) linear interorbital margins of the frontal, 96(1) postorbital extends posteriorly for more than one-half the length of the supratemporal fenestra, and 208(1) absence of a retroarticular process pit.

Comments

Among the taxa included in this analysis, the strict consensus tree presented here (fig. 54C) supports only the inclusion of Polyglyphanodon sternbergii, Gilmoreteius chulsanensis, Sineoamphisbaena hexatabularis, and Darchansaurus estesi within Polyglyphanodontinae as described here. The Adams Rule tree (fig. 55C) shows that all of principle trees are consistent with the inclusion of Erdenetesaurus robinsonae and Cherminsaurus kozlowskii. However, the inclusion of Gobinatus arenosus and Tchingisaurus multivagus clade can be neither confirmed nor denied by the current data. Gobinatus arenosus possesses only two of the synapomorphies for Polyglyphanodontinae listed in the diagnosis above (prefrontal-jugal contact and curved jugal); T. multivagus shares none of them.

Erdenetesaurus + Cherminsaurus

(figs. 54C, 55C)

Diagnosis

Erdenetesaurus robinsonae and Cherminsaurus kozlowskii share two unambiguous synapomorphies according to this analysis: 166(1) articular-prearticular fused to the surangular and 185(1) presence of distinct subcoronoid and surangular processes of the dentary.

Comments

As described and diagnosed by Sulimski (1975), E. robinsonae and C. kozlowskii are very similar. By the same token, cranial osteology may be very conserved in otherwise morphologically different species among extant squamates (e.g., closely related species of Anolis, Tiliqua, and Varanus).

Polyglyphanodon + Sineoamphisbaena hexatabularis + Darchansaurus + Gilmoreteius

(figs. 54C, 55C)

Diagnosis

This unnamed clade is united by 9(1) dermal sculpturing on the prefrontal, 26(0) absence of strong medial processes of the maxilla posterior to the premaxillary nasal process, and 57(0) anterior and posterior borders of the frontal of subequal width.

Sineoamphisbaena hexatabularis + Darchansaurus + Gilmoreteius

(figs. 54C, 55C)

Diagnosis

The unnamed clade containing Sineoamphisbaena hexatabularis, Darchansaurus, and Gilmoreteius chulsanensis are united in this analysis by two unambiguous synapomorphies: 131(1) occipital condyle bipartite and constructed primarily by the exoccipital portions of the otooccipitals and 213(0) unwaisted marginal teeth.

Darchansaurus + Gilmoreteius

(figs. 54C, 55C)

Diagnosis

This unnamed clade is united by 96(3) a postorbital that extends posteriorly to contact the supratemporal and 215(1) the presence of caniniform teeth.

Scincoidea Oppel, 1811

(figs. 54C, 55C, 56B)

Definition

All taxa sharing a more recent common ancestor with Scincus scincus than with Lacerta viridis or Anguis fragilis.

Diagnosis

This stem-based taxon is diagnosed by 14(2) bilobed premaxillary incisive process, 31(1) neochoanate condition, 83(1) presence of a nuchal fossa on the parietal, 96(0) postorbital extends posteriorly for less than one-half the length of the supratemporal fenestra, 114(2) ventromedial palatal folds of the palatine hiding most or all of the dorsomedial processes of the palatine, 121(1) midline contact of the palatines, and 195(0) straight ventral margin of the coronoid in medial view.

Comments

Estes et al. (1988) assigned the name Scincoidea as a node using Cordylidae and Scincidae as anchor taxa. However, differing phylogenetic hypotheses have yielded a variety of placements for Cordyliformes, Xantusiidae, and Paramacellodidae (see above) (figs. 1Figure 2Figure 3Figure 4Figure 5Figure 6Figure 7Figure 8Figure 910). Moreover, given the present topology (figs. 53Figure 54Figure 5556) and the Estes et al. (1988) definition, Scincoidea would include Lacertoidea. Implementation of the Estes et al. (1988) definition would require Scincoidea to become synonymous with Scincomorpha given the current analysis, or possibly with Lee's (1998) Diploglossa given some other recent analyses (Lee, 1998, 2000; Lee and Caldwell, 2000). The current definition ensures the presence of a Scincoidea including skinks and all their closest relatives.

Scinciformes comb. nov.

(figs. 54C, 55C)

Definition

Scincus scincus, Scelotes bipes, Acontias meleagris, Feylinia currori, and all descendants of their last common ancestor.

Diagnosis

Scinciformes, as defined here and in the context of the current topology, are united by six unambiguous synapomorphies. These are 83(2) presence of an expansive nuchal fossa that extends well onto the parietal table, 181(1) Meckel's canal partly closed, but unfused, 192(0) coronoid process short and broad, 207(1) retroarticular process that is posteriorly expanded (broadened), 221(1) replacement teeth occur posterolingually with a small resorption pit, and 277(2) absence of an ectepicondylar groove and foramen.

Comments

Scinciformes, as defined here, includes all the taxa usually considered Scincidae, plus Dibamidae, Amphisbaenia, and Serpentes (figs. 53Figure 54Figure 5556). Scincidae sensu lato, the Scincidae of most recent authors (Estes et al., 1988; Wu et al., 1996; Evans and Barbadillo, 1998; Lee, 1998, 2000; Caldwell, 1999a, 2000; Lee and Caldwell, 2000; Reynoso and Callison, 2000), is paraphyletic in the current topology with respect to the Dibamidae, Amphisbaenia, and Serpentes. Rather than simply including dibamids, amphisbaenians, and snakes within Scincidae to maintain the “subfamilies” Feyliniinae, Acontinae, and Scelotinae, the traditional “subfamilies” are treated as “families” of scinciforms in the current taxonomic scheme.

The recovery of a clade containing skinks, dibamids, amphisbaenians, and snakes is not totally surprising. Morphologists have noted, for decades, the similarities between various limbless and limb-reduced skinks and all, or subsets, of these groups (Camp, 1923; Senn and Northcutt, 1973; Northcutt, 1978; Rieppel, 1980d, 1981a, 1984b; Greer, 1985) (see figs. 1, 14). Hallermann (1998) offered the first study to analyze the monophyly of scincids sensu lato (fig. 15). However, this is the first morphological analysis to include the appropriate taxonomic sampling (including both extant and fossil groups) to analyze this possible relationship. Even so, further analysis is necessary, especially in light of recent molecular studies suggesting a possible relationship between Serpentes and Iguania and between Amphisbaenia and Lacertidae (Harris et al., 2001; Townsend et al., 2004; Vidal and Hedges, 2004, 2005).

The current phylogenetic hypothesis suggests that the individual clades of scinciforms (including the clades usually considered to constitute Scincidae sensu lato) must have been distinct by the end of the Early Cretaceous (fig. 56B). This hypothesized antiquity is somewhat surprising based on the nature of the fossil record of scincids, scelotids, feyliniids, acontids, and dibamids. However, recent molecular and molecular/biogeographic studies suggest that some nested scincid (sensu stricto) lineages may have been distinct by the beginning of the Oligocene (Hickson et al., 2000). Also, the morphology of the Cretaceous Contogenys is consistent with extant scincids.

Scelotidae + Scincophidia

(figs. 54C, 55C, 56B)

Diagnosis

Scinciforms exclusive of Scincidae (in the current sense) are united by 10(0) absence of dermal sculpturing from the frontal and parietal, 26(0) absence of strong medial processes of the maxilla posterior to the premaxillary nasal process, 52(1) jugal lying mostly medial (rather than dorsal) to the maxilla, 69(0) absence of parietal tabs on the frontal, 74(0) absence of frontal tabs on the parietal, 175(1) absence of a clear posterior mylohyoid fenestra, 267(3) two or fewer sternal ribs, and 302(1) fusion of cephalic scales.

Scincophidia tax. nov.

(figs. 54C, 55C, 56B)

Definition

Feylinia currori, Acontias meleagris, Dibamus novaeguineae, and all descendants of their last common ancestor.

Etymology

Skinkos (Greek; a kind of lizard), in reference Scincus and to the larger, including clade (Scincoidea); fidi (Greek; “snake”) in reference to the name (Ophidia) sometimes used for snakes, as a suffix for snake group names (e.g., Scolecophidia), or as a descriptor in naming snakelike taxa (e.g., Ophisaurus, Ophiodes). The name refers to the body form of these squamates, elongate and often limbless.

Diagnosis

This clade is diagnosed by 28(1) midpoint/apex of the maxillary nasal process posterior to the midpoint of the maxilla, 57(0) anterior and posterior borders of the frontal of subequal width, 81(1) broad, flat dorsal margins to the supratemporal processes of the parietal, 95(1) absence of the postorbital, 117(1) ventromedial processes of the pterygoids, 181(2) closed and fused Meckel's canal, 261(0) straight (nonangulate) clavicle, and 268(1) absence of the interclavicle.

Comments

A possible relationship between scincomorphs and dibamids has been recognized for more than a century (Cope, 1900; Fürbringer, 1900a; Camp, 1923; Rieppel, 1984b), but recent cladistic studies (Wu et al., 1996; Evans and Barbadillo, 1998; Lee, 1998, 2000; Caldwell, 1999a; Lee and Caldwell, 2000) have suggested that dibamids probably belong to the gekkonomorph lineage (figs. 3, 4, 6Figure 78). Importantly, the older studies considered independent groups of Scincidae sensu lato, but the more recent studies (those which suggest dibamids are related to gekkotans) have treated Scincidae sensu lato as a single taxon. Acontidae, Amphisbaenia, Dibamidae, Feyliniidae, and Serpentes form a clade under the current phylogenetic hypothesis and are included here in the Scincophidia (figs. 57, 58). However, only acontids, feyliniids, and dibamids are used as anchor taxa because their relationship has been the longest recognized and because of the relatively volatile nature of amphisbaenians and snakes in phylogenetic analyses. If dibamids should be shown to be gekkonomorphs, then Scincophidia becomes a junior subjective synonym for Scleroglossa. In this way, the taxon name may be effectively eliminated in the event of the gekkonomorph-dibamids hypothesis is shown to be a strong one.

Figure 57

The skulls of some representative, nonsnake scincophidians in left lateral view. (A) Feylinia elegans, (B) Acontias plumbeus, (C) Dibamus novaeguineae, and (D) Diplometopon zarudnyi. Note the progressive development of the descending processes of the frontals, descending processes of the parietals, and coronoid process of the dentary. Modified after (A–B) Rieppel, 1981a, (C) Rieppel (1984b) and Greer (1985), and (D) Maisano et al. (2005; 2006).

i0003-0090-310-1-1-f57.gif

Figure 58

The skulls of three snakes that have, at some point, been considered close to the ancestral morphology for Serpentes. (A) Leptotyphlops goudottii dulcis, (B) Haasiophis terrasanctus, and (C) Cylindrophis ruffus. Leptotyphlyops dulcis is an extant burrowing form with greatly reduced eyes, H. terrasanctus is an aquatic Cretaceous fossil snake with legs, and C. ruffus is an extant burrower that is less specialized for fossoriality than L. dulcis. Modified after (A) Kley (2004; 2006), (B) Tchernov et al. (2000) and Rieppel et al. (2003), and (C) Rieppel (1983).

i0003-0090-310-1-1-f58.gif

Acontidae + Dibamidae + Amphisbaenia + Serpentes

(figs. 54C, 55C, 56B)

Diagnosis

These taxa form a clade within scincophidians based on 67(1) presence of a frontal-palatine contact, 83(0) absence of a parietal nuchal fossa (interpreted here as a reversal), 86(0) dorsal origin of the jaw adductor musculature on the parietal, 153(1) hypoglossal and vagus foramina confluent or subdivided within a single canal, 162(1) short tympanic crest on the quadrate, 184(0) presence of a large, posterodorsal, coronoid process of the dentary, 192(1) presence of a tall, narrow, coronoid process, 198(0) absence of a posterior overlap of the coronoid by the dentary, 209(0) absence of a tubercle on the medial surface of the retroarticular process, and 321(1) absence of an external ear.

Dibamidae + Amphisbaenia + Serpentes

(figs. 54C, 55C, 56B)

Diagnosis

This mostly limbless clade of scincophidians is diagnosed by 5(1) presence of an interorbital septum, 41(3) absence of the lacrimal, 98(1) absence of a squamosal, 101(0) absence of a palpebral, 132(2) absence of a crista prootica, 133(2) braincase closure primarily through downgrowth of the parietals, 137(0) absence of the crista alaris prootica, 140(1) entocarotid fossa indistinct/absent, 145(1) anterior opening of the Vidian canal opens dorsally on the dorsum sella, 151(1) anterior location for the spheno-occipital tubercle, 155(1) occipital recess visible in ventral view (not hidden by spheno-occipital tubercles), 159(3) quadrate suspended mostly from the opisthotic, 162(2) absence of a tympanic crest, 232(2) absence of distinct neural spines, 256(1) presence of a posterodorsal pseudotuberculum on the ribs, 305(0) dorsal body osteoderms absent, 307(0) ventral body osteoderms absent, 313(1) eyeball reduced and covered externally by a head scale, and 319(1) a flattened foretongue (see character description above, Estes et al., 1988, and Schwenk, 1988).

Dibamidae Boulenger, 1884

(figs. 54C, 55C, 56B)

Definition

Anelytropsis papillosus, Dibamus novaeguineae, and all descendants of their last common ancestor.

Diagnosis

Dibamus and Anelytropsis papillosus form a clade diagnosed by 12(2) mediolateral breadth of the premaxillary nasal process less than the dorsoventral depth, 32(2) palatine flange expanded posteromedially beyond the posterolateral process of the maxilla, 33(1) maxillary tooth row terminates at the level of the anterior border of the orbit, 146(2) prootic alone houses the external posterior opening of the Vidian canal, 188(2) splenial present, but fused to the postdentary bones, and 234(0) absence of oblique vertebral condyles.

Comments

Only one of about 20 named species of Dibamus (Uetz, 2007) was included in this analysis (D. novaeguineae). More species, possibly one including more plesiomorphic features (e.g., D. bourreti; see Iordansky, 1985), will be included in the future. However, available material and descriptions suggests that Anelytropsis papillosus retains more plesiomorphic features than any observed Dibamus.

Amphisbaenia + Serpentes

(figs. 54C, 55C, 56B)

Diagnosis

Amphisbaenians and snakes are hypothesized to form a clade exclusive of other squamates in this analysis; a clade diagnosed by 28(0) midpoint/apex of maxillary nasal process at or anterior to the midpoint of the maxilla, 79(0) absence of a midline parietal flange, 110(0) absence of medial “step” to the palatine, 114(0) no development of the secondary palate, 117(0) absence of ventromedial processes (basipterygoid buttresses) on the pterygoid, 131(1) occipital condyle bipartite and constructed primarily by the exoccipital portions of the otooccipitals, 134(1) absence of an epipterygoid, 182(1) absence of a dentary shelf, 254(2) chevrons fused to the vertebrae, 284(2) pelvic bones with nonsutural contacts, 289(2) absence of a femur (note that this is reversed both within Amphisbaenia and Serpentes), and 317(4) foretongue divided for more than 50 percent of its length.

Comments

The sister-group relationship between amphisbaenians and snakes was recently recovered by Evans and Wang (2005), a result duplicated here. Other recent studies have suggested that amphisbaenians and dibamids form the sister-taxon to snakes (Evans and Barbadillo, 1997, 1998, 1999; Evans et al., 2005). This hypothesis is consistent with the terrestrial origin of snakes, a hypothesis further supported by the recent discovery of Najash rionegrina (Apestiguía and Zaher, 2006); a taxon that will be included in future versions of this analysis.

Amphisbaenia Gray, 1844

(figs. 54C, 55C, 56B)

Diagnosis

The current analysis recovers the following unambiguous synapomorphies for Amphisbaenia: 32(0) absence of a maxillary palatine flange, 37(1) prefrontal-postfrontal/postorbitofrontal contact present, 67(0) frontal not contacting the palatines, 74(1) frontal tabs of the parietal present dorsally, 108(0) broad vomer-palatine contact, 123(1) pyriform recess narrow (see the character description above), 124(3) ectopterygoid-palatine contact closes the suborbital fenestra, 128(1) spheno-occipital epiphyses present, 158(1) streptostylic quadrate suspension, 222(1) presence of an enlarged medial premaxillary tooth, 296(1) presence of annular rings along the length of the body, 301(0) presence of cycloid scales, and 303(0) body scales not deeply imbricated.

Comments

Estes et al. (1988) defined Amphisbaenia as a node. This would mean that fossil taxa related to amphisbaenians, but falling outside of the crown-group, would not be considered part of Amphisbaenia. The matter is largely semantic, but neither this analysis, nor that of Estes et al. (1988) include taxonomic sampling complete enough to warrant a definition of the name Amphisbaenia. Kearney (2003a) performed the most complete recent analyses, by far, of amphisbaenians, but did not formally define Amphisbaenia. I refrain from attaching a definition to the name.

Amphisbaenidae + Bipes + Blanus

(figs. 54C, 55C)

Diagnosis

The current analysis recovers two unambiguous synapomorphies joining amphisbaenids, Bipes biporus, and the genus Blanus. These are 91(1) absence of a postfrontal and 252(0) caudal autotomy planes present on the transverse processes.

Bipes + Blanus

(figs. 54C, 55C)

Diagnosis

This unnamed clade is diagnosed, in this analysis, by 119(0) presence of a pterygoid-vomer contact, 158(0) absence of an anteriorly elongate extracolumella, 260(0) rodlike clavicles, 265(1) scapula shorter than the coracoid, and 289(0) cylindrical femur with moderately expanded proximal and distal ends.

Serpentes Linnaeus, 1766

(figs. 54C, 55C, 56B)

Definition

Typhlops lineolatus ater, Anomalepis mexicanus, Vipera aspis, Python molurus, and all descendants of their last common ancestor.

Diagnosis

This clade is diagnosed by 16(1) nonsutural contacts between the premaxilla and maxilla, 29(1) weakly inclined anterior margin of the maxillary nasal process, 31(0) paleochoanate condition, 61(3) subolfactory processes of the frontal contact the parasphenoid, 65(0) frontal fails to contact the maxilla, 91(1) absence of a postfrontal, 135(1) absence of a processus ascendens tecti synotici, 150(2) elongate, posterolaterally directed, spheno-occipital tubercles, 167(1) absence of a mandibular symphysis, 180(0) dentary extends posteriorly no further than the coronoid process, 181(0) Meckel's canal open, 184(2) absence of a posterodorsal coronoid process on the dentary, 187(2) dentary primarily supported by the surangular, 190(1) posterior terminus of the splenial at or anterior to the coronoid apex, 191(2) presence of an intramandibular joint in which the angular receives the splenial, 194(1) anterior end of coronoid abutting the dentary (rather than overlapping it), 210(0) retroarticular process without torsion, 214(2) “modified pleurodont” dental attachment (sensu Zaher and Rieppel, 1999a), and 235(2) presence of zygosphenes which face ventrolaterally (fig. 44B).

Comments

The definition of Serpentes as a node containing extant snakes here is in keeping with those of previous such definitions and applications of the group name (Rage, 1984; Estes et al., 1988; Lee, 1997, 1998, 2000, 2001; Lee and Caldwell, 2000; Caldwell, 1999a; Apestiguía and Zaher, 2006). Under this definition, Ophidia might be invoked as a stem-based name to include taxa that would be identified as snakes, but which fall outside of the crown group (e.g., Najash rionegrina; Apestiguía and Zaher, 2006).

Liotyphlops albirostris + Typhlops lineolatus

(figs. 54C, 55C, 56B)

Diagnosis

Liotyphlops albirostris and Typhlops lineolatus (representatives of Anomalepidae and Typhlopidae, respectively, in this analysis) are joined by three unambiguous synapomorphies. These are 28(3) extremely foreshortened maxilla, 111(2) palatine deeper than long, and 112(1) vomer much longer than the palatine.

Comments

The current analysis finds support for Scolecophidia in a majority of the principle trees, but it is not supported in the strict nor the Adams consensus (figs. 54 and 55, respectively). In this analysis, Leptotyphlops goudottii, the only other putative scolecophidian included in this analysis, is recovered as the sister-taxon to a clade composed of Liotyphlops albirostris and Typhlops lineolatus, as the sister taxon to all other snakes, or as the sister taxon to a clade containing Alethinophidia, Dinilysia patagonica, Pachyophis woodwardi, and Wonambi naracoortensis. Future analyses (with more inclusive taxon and character sampling) will more completely address this issue.

Note that McDowell and Bogert (1954) challenged the idea that typhlopids and anomalepids are snakes and suggested that they are anguids. This hypothesis was never widely accepted and no subsequent study has recovered non-holophyly of snakes. Snake holophyly is analyzed and supported by this analysis.

Oculatophidia tax. nov.

(figs. 54C, 55C, 56B)

Definition

All taxa sharing a more recent common ancestor with Coluber constrictor and Anilius scytale than with Leptotyphlops goudottii bilineatus, Anomalepis mexicanus, or Typhlops lineolatus jaimacensis.

Etymology

Oculatus (Latin; “having eyes” or “conspicuous”); fidi (Greek; “snake”) in reference to the name (Ophidia) sometimes used for snakes and as a suffix for snake group names. The name used here is in reference to the basal dichotomy between blindsnakes (scolecophidians) and all other snakes (the group here named; nonblind snakes), and in reference to the more conspicuous nature of these snakes (as compared to blindsnakes) today and in the fossil record.

Diagnosis

This clade is diagnosed by 6(1) posteriorly elongated nares, 74(1) frontal tabs of the parietal present dorsally, 87(0) presence of a supratemporal, 115(2) palatine teeth arranged in a single line, 118(1) pterygoid teeth arranged in a single line, 124(0) absence of a contact between the ectopterygoid and palatine anterior to the suborbital fenestra, 131(0) occipital condyle a single unit made of the otooccipitals and basioccipital, 159(2) quadrate suspended mainly from the supratemporal, 163(1) presence of a quadrate suprastapedial process, 168(1) adductor fossa with a distinct vertical flange, 192(0) coronoid process short and broad, 214(3) tooth bases enclosed by an expanded interdental ridge, and 232(0) presence of short and broad neural spines.

Comments

The interrelationships between Dinilysia patagonica, Alethinophidia, and a clade containing Pachyophis woodwardi and Wonambi naracoortensis are unresolved in this analysis. Principle trees variably suggest that Dinilysia patagonica is the basal most member of this clade or that it is the sister-taxon to the Pachyophis woodwardiWonambi naracoortensis clade. Future inclusion of more putative “madtsoiids” and Najash may help to more completely resolve these relationships.

Pachyophis woodwardi + Wonambi naracoortensis

(figs. 54C, 55C, 56B)

Diagnosis

Pachyophis woodwardi and Wonambi naracoortensis share one unambiguous synapomorphy: 186(1) angular process of dentary terminates anterior to the level of the coronoid process.

Comments

The relationship between Pachyophis woodwardi and Wonambi naracoortensis is only weakly supported, probably relating to the relative completeness of both taxa and the small amount of known morphological overlap. This possible relationship requires further study.

Alethinophidia Nopcsa, 1923b

(figs. 54C, 55C, 56B)

Definition

The most recent ancestor of extant taxa more closely related to Coluber constrictor and Anilius scytale than to Leptotyphlops goudottii bilineatus, Anomalepis mexicanus, or Typhlops lineolatus jaimacensis.

Diagnosis

Alethinophidians are diagnosed in this study by 110(1) palatines with medial expansion anteriorly, 126(0) absence of ectopterygoid exposure on the lateral surface of the skull behind the maxilla, and 143(1) posterolateral sphenoid flanges ventrolaterally overlying the basioccipital.

Comments

Rage (1984) described a basal dichotomy in Serpentes between Scolecophidia and Alethinophidia, an approach to modern snakes informally applied by various authors (see Lee, 2001; Scanlon and Lee, 2002; Kelly et al., 2003). However, Dinilysia patagonica has been excluded from Alethinophidia whenever it is mentioned explicitly (e.g., Caldwell, 1999a; Rieppel et al., 2002, 2003; Apestiguía and Zaher, 2006). The current definition of Alethinophidia reflects the latter usage in which Alethinophidia is a node-based name including extant non-scolecophidian snakes.

Macrostomata Müller, 1831

(figs. 54C, 55C, 56B)

Definition

All taxa sharing a more recent common ancestor with Coluber constrictor than with Anilius scytale, Cylindrophis ruffus, or Uropeltis ceylanicus.

Diagnosis

The six unambiguous macrostomatan synapomorphies recovered in this analysis are 1(1) snout makes up greater than 30 percent of the total skull length, 28(2) dorsal and ventral margins of the maxilla subparallel, 64(1) presence of a contact between the medial frontal pillar and the lateral subolfactory flanges, 107(1) a rod-shaped vomer, 137(1) presence of a short crista prootica, and 178(0) dentary ventrally convex along its long axis.

Comments

The definition of Macrostomata offered here is in line with the common usage of the name. Macrostomata is typically regarded as including all non-anilioid alethinophidians. Anilioid monophyly is not analyzed here, thus the type species from each of the three anilioid clades (Anilius, Cylindrophis, and Uropeltidae) are used in the definition. Anilioid monophyly will be further analyzed elsewhere.

Pachyrhachidae + “neo-Macrostomata”

(figs. 54C, 55C)

Diagnosis

Pachyrhachids and more derived macrostomatans form a clade diagnosed by 108(2) absence of an articulation between the vomer and palatine, 127(1) presence of a ventral sagittal ridge on the sphenoid and basioccipital, and 163(0) absence of a suprastapedial process (interpreted as a reversal here).

Pachyrhachidae comb. nov.

(figs. 54C, 55C)

(Pachyrhachis problematicus from Haas, 1979)

Definition

All taxa sharing a more recent common ancestor with Pachyrhachis problematicus than with Coluber constrictor.

Diagnosis

Pachyrhachis problematicus, Haasiophis terrasanctus, and Eupodophis descouensi form a clade exclusive of other macrostomatans, diagnosed by 95(0) presence of a postorbital, 192(1) presence of a tall, narrow, coronoid process, and 290(0) presence of a pes with metatarsals and digits (see fig. 47).

Anguimorpha Fürbringer, 1900a

(figs. 54C, 55C, 56B)

Definition

All taxa sharing a more recent common ancestor with Anguis fragilis and Varanus varius than with Scincus scincus, Cordylus cordylus, or Iguana iguana.

Diagnosis

Anguimorphs are diagnosed by two unambiguous synapomorphies; 8(1) presence of dermal sculpturing on the maxilla and 209(0) absence of a tubercle on the medial surface of the retroarticular process.

Comments

Estes et al. (1988) defined Anguimorpha as a crown-node, but the taxon is typically used to receive all nonscincomorph autarchoglossans and it is so defined here. Importantly, no previous analysis has addressed noncrown anguimorph relatives. Many of the basal most members of this clade were originally described as scincomorphs (e.g., “Pseudosaurillus” sp., Paramacellodus oweni, Becklesius hoffstetteri, Pseudosaurillus becklesi, Meyasaurus diazromerali, and Eolacerta robusta) or gekkotans (Yabeinosaurus tenuis) (see Endo and Shikama, 1942; Hoffstetter, 1967a; Prothero and Estes, 1980; Estes, 1983; Borsuk-Białynicka, 1985; Broschinski and Sigogneau-Russell, 1996; Evans and Barbadillo, 1998; Evans and Chure, 1998a; Averianov and Skutchas, 1999; Reynoso and Callison, 2000; Müller, 2001). The incomplete nature of taxa such as “‘Pseudosaurillus’ sp.” (sensu Estes, 1983), Paramacellodus oweni, and Becklesius hoffstetteri, is probably to blame for the relative paucity of unambiguous anguimorph synapomorphies.

Anguimorpha exclusive of “Pseudosaurillus” sp.

(figs. 54D, 55D, 56C)

Diagnosis

The clade containing all anguimorphs except “Pseudosaurillus” sp. may be diagnosed by 182(0) presence of a subdental shelf.

Anguimorpha exclusive of “Pseudosaurillus” sp. and Paramacellodus oweni

Diagnosis

This clade is supported by only one unambiguous synapomorphy: 166(0) absence of fusion between the articular-prearticular and the surangular.

Eolacerta robusta + Meyasaurus diazromerali + Pseudosaurillus beckelsi + Yabeinosaurus tenuis + Anguiformes

(figs. 54D, 55D, 56C)

Diagnosis

The clade formed by this unresolved trichotomy is here diagnosed by three unambiguous synapomorphies, including 26(0) absence of strong medial processes of the maxilla posterior to the premaxillary nasal process, 176(0) posterior mylohyoid foramen located anterior to the coronoid apex, and 185(1) presence of distinct subcoronoid and surangular processes of the dentary.

Eolacerta robusta + Meyasaurus diazromerali + Pseudosaurillus beckelsi

(figs. 55D, 56C)

Diagnosis

This unnamed clade is diagnosed, in this analysis, by 21(1) nasals that are in contact for less than half their length in dorsal view and 203(1) presence of a prearticular crest.

Meyasaurus diazromerali + Pseudosaurillus becklesi

(figs. 55D, 56C)

Diagnosis

These taxa share 182(2) presence of an enlarged subdental shelf and 277(2) absence of an ectepicondylar groove and foramen.

Yabeinosaurus tenuis + Anguiformes

(figs. 55D, 56C)

Diagnosis

Yabeinosaurus tenuis and anguiforms form a clade exclusive of all other squamates diagnosed by 26(0) absence of strong medial processes of the maxilla posterior to the premaxillary nasal process, 176(1) position of the posterior mylohyoid foramen posterior to the coronoid apex, and 185(1) presence of distinct subcoronoid and surangular processes of the dentary.

Comments

Evans et al. (2005) recently redescribed Y. tenuis and inserted it into some recent phylogenetic data matrices. They found Y. tenuis to be a basal scleroglossan or as the immediate outgroup to crown-squamates (Evans et al., 2005). The current analysis produces a markedly different result, suggesting that Y. tenuis is a basal anguimorph, close to the crown group.

Anguiformes Conrad, 2006b

(figs. 54D, 55D, 56C)

Definition

Anguis fragilis, Varanus varius, and all descendants of their last common ancestor.

Diagnosis

This analysis recovers five unambiguous synapomorphies for anguiforms. These are 74(0) absence of frontal tabs on the parietal, 76(1) presence of an anteroposteriorly elongate decensus parietalis, 198(0) absence of a posterior overlap of the coronoid by the dentary, 221(1) replacement teeth occur posterolingually with a small resorption pit, and 260(0) clavicles lacking proximal expansion; thus, rod-like.

Comments

The phylogenetic topology presented in this analysis differs from that of Conrad (2006b) in the placement of Carusia and Xenosauridae. The earlier analysis found Carusia and Xenosauridae to be successively more proximal outgroups to a clade containing anguids, shinisaurids, and varanoids (Conrad, 2006b), but the present study recovers a monophyletic Carusioidea (sensu Gao and Norell, 1998, 2000; see below) as the sister-taxon to a clade containing modern anguids (figs. 54D, 55D). Regardless, the name Anguiformes is retained and maintains its meaning as a clade containing anguids and varanoids (among others) to the exclusion of more basal anguimorphs (e.g., Yabeinosaurus).

Note that Diploglossa might have been an alternative name for this clade or even for the clade described above as Anguimorpha if not for re-definition of that name (Lee, 1998). That definition, “The least inclusive clade containing [Cordylidae, Scincidae, Anguidae, Xenosauridae, Helodermatidae, Lanthanotus borneensis, and Varanus] to the exclusion of [Rhynchocephalia, Iguanidae, Agaminae, Leiolepis, Uromastyx, Chamaeleonidae, Xantusiidae, Eublepharinae, Diplodactylinae, Gekkoninae, Pygopodidae, Lacertidae, Teiidae, and Gymnophthalmidae]” (Lee, 1998: 436) describes a paraphyletic group in this phylogenetic hypothesis. If that group name is made monophyletic by amending the definition to fewer excluded taxa, then the name becomes a synonym for Autarchoglossa. However, traditional usage of Diploglossa (see, for example Romer, 1956; Estes, 1983; Estes and Pregill, 1988; Uetz, 2007) does not intend the inclusion of scincomorphs such as Cordyliformes or Scincidae.

Carusioidea + Anguidae

(figs. 54D, 55D, 56C)

Diagnosis

The current phylogenetic hypothesis suggests a sister-group relationship between carusioids and anguids based on 69(0) absence of parietal tabs on the frontal, 285(2) symphysial portion of the pubis more than half again as long as the tubercular portion, and 351(1) absence of the M. adductor mandibulae.

Comments

Caldwell (1999a) recently recovered support for a monophyletic clade including xenosaurids and anguids in a clade roughly corresponding with the traditional usage of Anguioidea (McDowell and Bogert, 1954; Romer, 1956; Carroll, 1988b; Uetz, 2007). Importantly, analysis of a preliminary version of this data set did not recover a carusioid-anguid clade (Conrad, 2005a), nor did other recent osteological studies (Wu et al., 1996; Evans and Barbadillo, 1997, 1998; Gao and Norell, 1998; Lee, 1998, 2000, 2005a; Evans and Barbadillo, 1999; Conrad, 2004b, 2006b; Evans and Wang, 2005; Evans et al., 2005; ). Importantly, Townsend (2002) offered preliminary results of an analysis based on molecular data supporting a similar hypothesis, published in full by Townsend et al. (2004). Although these analyses find Heloderma as the sister-taxon to Anguidae, a hypothesis not supported by the present morphological data set, they did consistently recover a clade containing Xenosaurus and Anguidae to the exclusion of Shinisaurus and Varanidae in their maximum-likelihood trees.

The greater taxonomic and character sampling of the present study as compared to previous morphological analyses suggests greater credibility of the current phylogenetic hypothesis. This is bolstered by a general similarity between the phylogenetic topology supported by molecular data. Even so, the monophyly of the carusioid-anguid taxon and of Varanoidea should be further analyzed.

Carusioidea Gao and Norell, 1998

(figs. 54D, 55D, 56C)

Definition

Carusia intermedia, Xenosaurus grandis, and all descendants of their last common ancestor.

Diagnosis

Carusia intermedia and xenosaurids share the following unambiguous synapomorphies: 50(1) dermal rugosities on the postorbital process of the jugal, 53(1) presence of a jugal-squamosal contact, 55(1) fused frontals, 61(0) subolfactory processes present as simple ventral downgrowths, 94(1) fused postorbitofrontal, 96(0) postorbital extends posteriorly for less than one-half the length of the supratemporal fenestra, 100(0) presence of a dorsal process on the squamosal, 115(1) absence of palatine teeth, 166(1) articular-prearticular fused to the surangular, and 193(2) presence of a well-developed coronoid labial flange that extends posterior to the main body of the coronoid process in lateral view.

Comments

Gao and Norell (1998) defined Carusioidea as a node anchored to Carusia intermedia and their Xenosauridae, the latter including Shinisaurus crocodilurus according to their analysis. Here, the definition is modified such that it is anchored to the type species of the Xenosauridae and Carusia.

The carusioid clade has remained largely untested since Gao and Norell (1998) proposed it. Conrad (2006b) found no support for the clade. The current analysis incorporates much broader taxonomic and character sampling, though, and recovers a monophyletic Carusioidea, albeit without the inclusion of Shinisaurus crocodilurus.

Xenosauridae Cope, 1886

(figs. 54D, 55D, 56C)

Definition

All taxa sharing a more recent common ancestor with Xenosaurus grandis than with Anguis fragilis, Carusia intermedia, or Varanus varius.

Diagnosis

Xenosaurids share two unambiguous synapomorphies: 126(1) ectopterygoid exposed on the lateral skull surface posterior to the maxilla and 212(3) presence of “shoulders” or incipient cusps on the posterior marginal teeth.

Comments

Common usage and phylogenetic definitions of Xenosauridae have been anchored to Xenosaurus and Shinisaurus (McDowell and Bogert, 1954; Gauthier, 1982; Estes, 1983; Estes et al., 1988; Presch, 1988; Wu et al., 1996; Lee, 1998; Lee and Caldwell, 2000). However, Shinisaurus shows morphological affinities with platynotans (see below) and no special affinity for xenosaurids (Conrad, 2004a, 2004b, 2005a, 2006a, 2006b) (figs. 54Figure 5556). Recent molecular work supports a similar hypothesis, suggesting that Shinisaurus is related to varanids (Townsend et al., 2004) (fig. 10A).

Restes rugosus is usually considered to have affinities with Xenosaurus and was initially considered a species of Exostinus (Gilmore, 1942a; Estes, 1965, 1975). Although Conrad (2006b) was unable to resolve the placement of R. rugosus, perhaps due to the incompleteness of that and other taxa, the current analysis supports the hypothesis that it is a xenosaurid.

The current definition of Xenosauridae takes into account the problematic relationships of both Shinisaurus and Restes. Although the data analyzed here strongly suggest that Shinisaurus is not a member of the xenosaurid radiation, the current definition of Xenosauridae would allow inclusion of Shinisaurus should the topology of, for example, Lee and Caldwell (2000) (fig. 7) prove accurate.

Note that Xenosaurus is included here as a composite taxon based on observations of X. grandis and X. platyceps. Future versions of this data matrix may include codings for some individual Xenosaurus species.

Exostinus Cope, 1873

(figs. 54D, 55D)

Definition

Exostinus serratus, E. lancensis, and all descendants of their last common ancestor.

Diagnosis

The two included species of Exostinus are united by 310(1) presence of thickened, irregularly shaped, osteoderms.

Comments

Exostinus serratus and E. lancensis are nearly identical in the 42 characters for which they may both be scored. Exostinus lancensis possesses more well-developed subolfactory processes than E. serratus, but this is the only observed coding difference between the two for this analysis. Given the incomplete nature of the remains for both species, however, more difference might appear with the discovery of more complete remains for one or both species. Indeed, a preliminary review of available material suggests that Exostinus may be a paraphyletic assemblage of three or more taxa (Bhullar, 2007).

Anguidae Gray, 1825

(figs. 54D, 55D, 56C)

Definition

Anguis fragilis, Gerrhonotus liocephalus, Diploglossus fasciatus, and all descendants of their last common ancestor.

Diagnosis

Anguids form a clade diagnosed by 9(0) absence of dermal sculpturing on the prefrontal, 10(0) absence of dermal sculpturing from the frontal and parietal, 31(1) neochoanate condition, 58(0) linear interorbital margins of the frontal, 67(1) presence of a frontal-palatine contact, 171(2) free posteroventral margin of the intramandibular septum, 183(2) dentary contributing to dorsal and anterior margin of the anterior inferior alveolar foramen, 207(1) retroarticular process that is posteriorly expanded (broadened), 210(2) strongly twisted retroarticular process, 254(2) chevrons fused to the vertebrae, and 304(1) presence of a lateral body fold.

Comments

The definition of Anguidae used here follows that of Estes et al. (1988), but uses more specific anchor taxa. Note that if glyptosaurines are found to be outside of crown-anguids, then they must be considered distinct from Anguidae following this definition. However, the current analysis supports the general hypothesis of Gauthier (1982) that glyptosaurines are nested within extant anguids (figs. 54Figure 5556).

Gerrhonotinae McDowell and Bogert, 1954

(figs. 54D, 55D, 56C)

Definition

All taxa sharing a more recent common ancestor with Gerrhonotus liocephalus than with Anguis fragilis, Anniella pulchra, or Diploglossus fasciatus.

Diagnosis

The current analysis recovers five unambiguous gerrhonotine synapomorphies based on the included taxa, these are 88(1) supratemporal elongate, 173(1) coronoid contribution to the external border of the anterior surangular foramen, 186(1) angular process of dentary terminates anterior to the level of the coronoid process, 251(2) double, converging, caudal transverse processes, and 312(1) presence of keeled body osteoderms.

Comments

The position of Parophisaurus pawneensis ( = ?Xestops pawneensis Gilmore, 1928;  = Pancelosaurus pawneensis Meszoely, 1970;  = Odaxosaurus pawneensis Meszoely et al., 1978) as a basal gerrhonotine in this analysis is unexpected. Sullivan (1987) regarded P. pawneensis as a proximal outgroup to the North American anguine Ophisaurus, but did not offer a cladistic analysis. Sullivan (1987) further cautioned, however, that P. pawneensis is a difficult taxon to interpret because it is relatively plesiomorphic in much of its known morphology. A preliminary phylogenetic study of squamates suggested that P. pawneensis is the sister-taxon to a clade containing Anguinae and Diploglossinae (Conrad, 2005a). Comparisons of Parophisaurus pawneensis (AMNH FR8711) with extant taxa reveal that Parophisaurus pawneensis more closely resembles Ophisaurus ventralis than Gerrhonotus liocephalus in muzzle shape, dermal sculpturing, possessing extensive internasal contact, and in the relative contribution of the dentary to the anterior inferior alveolar foramen margin, but this parsimony analysis interprets those similarities as plesiomorphies or convergence.

Gerrhonotines are interpreted by this analysis to be the basalmost lineage of Anguidae. Although P. pawneensis is from the Middle Oligocene, the gerrhonotine lineage must have been distinct by the Late Cretaceous.

Paragerrhonotus ricardensis + Extant Gerrhonotines

(figs. 54D, 55D)

Diagnosis

This clade is supported by 7(1) presence of pitted dermal sculpturing, 21(1) nasals that are in contact for less than half their length in dorsal view, 55(1) fused frontals, and 183(1) dentary forming the dorsal border of the anterior inferior alveolar foramen.

Extant Gerrhonotine radiation

(figs. 54D, 55D)

Diagnosis

The extant gerrhonotines used in this analysis are considered to form a clade exclusive of Paragerrhonotus ricardensis based on the presence of three unambiguous synapomorphies. These are 61(0) subolfactory processes present as simple ventral downgrowths, 219(1) striated tooth crowns, and 226(1) chisel shaped posterior teeth.

Comments

The present analysis does not sample densely enough to meaningfully analyze the interrelationships of extant gerrhonotines. Therefore, this clade is not named and the further relationships of the gerrhonotines are not reported here.

Anguinae + Diploglossinae + Glyptosaurinae

(figs. 54D, 55D, 56C)

Diagnosis

This unnamed clade is diagnosed by 8(0) absence of dermal sculpturing on the maxilla and 120(0) vomer contact posteriorly invaded by the pyriform recess.

Anguinae Gray, 1825

(figs. 54D, 55D, 56C)

Definition

All taxa sharing a more recent common ancestor with Anguis fragilis than with Diploglossus fasciatus, Gerrhonotus liocephalus, or Glyptosaurus sylvestris.

Diagnosis

Anguines are diagnosed by 74(1) frontal tabs of the parietal present dorsally, 84(1) increased contact between the parietal and supratemporal arch, resulting in reduction of the supratemporal fenestra, 117(1) presence of ventromedial processes of the pterygoids, 140(1) entocarotid fossa indistinct/absent, 233(1) presence of precondylar constriction, 275(4) absence of a humerus, and 290(1) pes absent.

Comments

Sullivan (1987) suggested that anguines were non-monophyletic, but with respect to which groups, he gave no indication. Indeed, Sullivan (1987: fig. 9) produced a phylogenetic diagram showing a monophyletic Anguinae (fig. 59). Ironically, the current topology (fig. 54Figure 5556) suggests that Sullivan's (1987) Anguinae is polyphyletic because Parophisaurus, nested taxa within Anguinae according to Sullivan, is a basal gerrhonotine according to the current analysis.

Figure 59

Relationships of the anguine anguids according to Sullivan (1987). Fossil taxa denoted by daggers (†).

i0003-0090-310-1-1-f59.gif

Anguinae exclusive of Ophisaurus ventralis

(figs. 54D, 55D)

Diagnosis

Anguines exclusive of Ophisaurus ventralis form a clade diagnosed by 162(1) short tympanic crest on the quadrate, 183(1) dentary forming the dorsal border of the anterior inferior alveolar foramen, 239(0) horizontal dorsal margin of the atlas in lateral view, 264(1) absence of a contact between the epicoracoid and suprascapula, 265(1) scapula shorter than the coracoid, and 284(2) pelvic bones with nonsutural contacts.

Anguinae exclusive of Ophisaurus ventralis and “O.attenuatus

(figs. 54D, 55D)

Diagnosis

This unnamed clade is diagnosed by 126(1) ectopterygoid exposed on the lateral skull surface posterior to the maxilla, 186(1) angular process of dentary terminates anterior to the level of the coronoid process, and 251(2) double, converging, caudal transverse processes.

Anguini Augé, 2005

(figs. 54D, 55D)

Definition

Anguis fragilis, Pseudopus apodus, and all descendants of their last common ancestor.

Diagnosis

This clade is diagnosed by 47(1) jugal curved (rather than angulated), 67(0) frontal not contacting the palatines, 240(1) lateral processes of the atlas reduced, and 269(0) absence of an anterior process of the interclavicle.

Comments

Anguins include Anniella and Pseudopus apodus; one of the more often sampled and illustrated species of “Ophisaurus”. Anniella is sometimes separated into its own “family” Anniellidae (for a history of the phylogenetic and taxonomic placements of Anniella, see Gao and Norell, 1998). Importantly, the current analysis does not find a close relationship between Anniella and Apodosauriscus minutus, in contrast to the findings of Gauthier (1982). However, additional, undescribed, remains of Apodosauriscus (mentioned in Gauthier, 1982) will add critical data to the analysis and may bolster Gauthier's (1982) hypothesis upon their description. Even so, the current phylogenetic hypothesis reflects available data, suggesting that Anniella is an anguin close to Anguis.

Pseudopus apodus + Ophisauriscus quadrupes

(figs. 54D, 55D)

Diagnosis

This unnamed clade is supported by 61(0) subolfactory processes present as simple ventral downgrowths, and 312(1) presence of keeled body osteoderms.

Comments

Sullivan et al. (1999) recently redescribed Ophisauriscus quadrupes and offered important new details of its morphology from x-rays. These data help identify the presence of a limbed anguid (Ophisauriscus) nested within Anguini.

Anguis fragilis + Anniella pulchra

(figs. 54D, 55D)

Diagnosis

This unnamed clade is supported by 14(0) single-lobed incisive process, 23(1) frontal forms a single anterior wedge, 39(0) absence of a prefrontal-jugal contact, 82(0) presence of a transverse posterior margin of the parietal between the supratemporal processes, 88(1) elongate supratemporal, 115(1) absence of palatine teeth, 175(1) absence of a clear posterior mylohyoid foramen, 183(2) dentary contributing to dorsal and anterior margin of the anterior inferior alveolar foramen, 189(1) splenial not extending anterior to the midpoint of the dentary tooth row, 211(1) marginal teeth widely separated, 233(0) absence of precondylar constriction in the vertebrae, and 304(0) absence of a lateral fold in the body squamation.

Comments

As described by Gao and Norell (1998), the placement of Anniella (A. pulchra included here) has been problematic. However, the present analysis duplicates the findings of Gao and Norell (1998) adding support for the hypothesis that Anguis and Anniella are sister taxa.

Diploglossinae + Glyptosaurinae

(figs. 54D, 55D, 56C)

Diagnosis

The unnamed clade including diploglossines and glyptosaurines is diagnosed in this analysis by 166(1) articular-prearticular fused to the surangular, 201(1) angular extends anteriorly into the mental canal, and 254(1) anterior positioning of the chevrons.

Diploglossinae Cope, 1864

(figs. 54D, 55D, 56C)

Definition

All taxa sharing a more recent common ancestor with Diploglossus fasciatus than with Anguis fragilis, Glyptosaurus sylvestris, or Gerrhonotus liocephalus.

Diagnosis

The current analysis recovers two unambiguous synapomorphies for Diploglossinae as defined here: 178(0) dentary ventrally convex along its long axis and 186(1) angular process of dentary terminates anterior to the level of the coronoid process.

Comments

The current analysis suggests a sister-group relationship between extant diploglossines and Apodosauriscus minutus, the latter having been originally described as close to Anniella. The current phylogenetic hypothesis should be further analyzed. Importantly, Wiens and Slingluff (2001) recovered a sister-taxon relationship between Anniella and diploglossines and Gauthier (1982) suggested that Anniella and Apodosauriscus are each other's closest known relatives.

Celestus costatus + Diploglossus millepunctatus + Ophiodes sp.

(figs. 54D, 55D)

Diagnosis

The extant diploglossines included in this analysis form a clade diagnosed by 7(1) presence of pitted dermal sculpturing, 57(0) anterior and posterior borders of the frontal of subequal width, 61(0) subolfactory processes present as simple ventral downgrowths, 183(0) dentary does not contribute to the anterior inferior alveolar foramen margin, and 301(1) presence of cycloid scales.

Comments

The current analysis does not sample densely enough from diploglossines for further discussion of their relationships.

Glyptosaurinae Marsh, 1872

(figs. 54D, 55D, 56C)

Definition

All taxa sharing a more recent common ancestor with Glyptosaurus sylvestris than with Anguis fragilis, Diploglossus millepunctatus, or Gerrhonotus liocephalus.

Diagnosis

Glyptosaurinae is diagnosed by 10(1) presence of dermal sculpturing on the frontal and parietal, 82(0) presence of a transverse posterior margin of the parietal between the supratemporal processes, and 212(6) squared dorsal tooth margins.

Comments

Gauthier (1982) and Conrad (2006b) provided the only prior cladistic analyses of glyptosaurine relationships with other anguid groups, but both of those analyses used composite codings for said groups. Based on preliminary phylogenetic analyses (including those reported in Conrad, 2004b, 2005a), Peltosaurus granulosus and Glyptosaurinae were treated separately and as individual taxa. However, the current analysis reveals that P. granulosus is nested within Glyptosarinae, as suggested by earlier studies (e.g., Gilmore, 1928; Meszoely, 1970; Meszoely et al., 1978; Sullivan, 1979, 1986; Estes, 1983).

Glyptosaurines exclusive of Odaxosaurus piger

(figs. 54D, 55D)

Diagnosis

Glyptosaurines exclusive Odaxosaurus piger form a clade diagnosed by 7(1) presence of pitted dermal sculpturing, 8(1) presence of dermal sculpturing on the maxilla, 171(1) sutured posteroventral margin of the intramandibular septum, 311(2) large osteoderms covered with bony tubercles, and 312(1) presence of keeled body osteoderms.

Comments

Glyptosaurines are often subdivided into a two “tribes”; the plesiomorphic Melanosaurini and the apomorphic Glyptosaurini (Meszoely, 1970; Meszoely et al., 1978; Sullivan, 1979, 1986; Augé and Sullivan, 2006; Sullivan and Augé, 2006). Melanosaurins sensu Sullivan (1979) were initially considered to form a monophyletic group. More recently, melanosaurins sensu lato have been suggested as a paraphyletic assemblage (Estes, 1983; Sullivan, 1986; Augé and Sullivan, 2006; Sullivan and Augé, 2006), a view supported by the current analysis (figs. 54D, 55D).

Proxestops jepseni + Xestops vagans

(figs. 54D, 55D)

Diagnosis

This unnamed clade is supported by one unambiguous synapomorphy: 173(2) dentary contributing to the margin of the anterior surangular foramen.

Comments

Neither of these species is well preserved. Gauthier (1982) suggested that Proxestops jepseni represents a good potential structural intermediate between the Cretaceous anguid Odaxosaurus piger and the Eocene Xestops vagans, a view supported by Estes (1983). The present analysis offers a phylogenetic hypothesis consistent with that scenario. Note that this analysis recovers autapomorphies for X. vagans, but none for P. jepseni. Although the fragmentary nature of known P. jepseni material is probably at least partly responsible for this result, it supports Gauthier's (1982) hypothesis.

Peltosaurus granulosus + Melanosaurini + Glyptosaurini

(figs. 54D, 55D)

Diagnosis

This unnamed clade is united by 55(1) fused frontals, 84(1) increased contact between the parietal and supratemporal arch, resulting in reduction of the supratemporal fenestra, and 170(0) absence of a distinct intramandibular septum.

Melanosaurini + Glyptosaurini

(figs. 54D, 55D)

Diagnosis

Melanosaurins and glyptosaurins form a clade to the exclusion of Peltosaurus granulosus diagnosed by 57(1) frontals rhomboid in dorsal view and 78(1) parietal fossa closed posteriorly.

Melanosaurini Sullivan, 1979

(figs. 54D, 55D)

Definition

All taxa sharing a more recent common ancestor with Melanosaurus maximus than with Glyptosaurus sylvestris.

Diagnosis

Melanosaurus maximus and Arpadosaurus gazinorum are united in this analysis based on one unambiguous synapomorphy: 88(1) supratemporal 3 times as long as deep.

Comments

Estes (1983) suggested that any re-evaluation of the interrelationships of A. gazinorum and M. maximus must consider the possibility of the synonymy of the two. Indeed, the present analysis finds no diagnostic characters to separate these two taxa. However, the absence of autapomorphies for these taxa may be related to the incompleteness of A. gazinorum remains. If they are considered conspecific specimens, then Melanosaurini is taxonomically synonymous with Melanosaurus maximus in the present topology. However, the current definition of Melanosaurini allows for the addition of taxa if future discoveries produce new taxa more closely related to M. maximus than to Glyptosaurus sylvestris. The current definition also maintains the perceived spirit of usage set forth in Sullivan (1979).

Glyptosaurini Sullivan, 1979

(figs. 54D, 55D)

Definition

All taxa sharing a more recent common ancestor with Glyptosaurus sylvestris than with Melanosaurus maximus.

Diagnosis

This clade is diagnosed by 50(1) dermal rugosities on the postorbital process of the jugal, 80(1) short parietal supratemporal processes (see character description above), 298(1) small and irregularly shaped cephalic scales and 310(2) osteoderms thickened and expressed as polygonal mounds.

Glyptosaurins exclusive of Glyptosaurus sylvestris

(figs. 54D, 55D)

Diagnosis

This unnamed clade is diagnosed by 47(1) jugal curved (rather than angulated) and 48(1) posteroventral process of jugal absent.

Helodermoides tuberculatus + Paraglyptosaurus princeps

(figs. 54D, 55D)

Diagnosis

Helodermoides tuberculatus and Paraglyptosaurus princeps form a clade diagnosed by the unambiguous synapomorphies 84(0) no increased contact between the supratemporal arches and the parietal and 226(0) posterior marginal teeth not chisel-shaped (reversal of an anguid synapomorphy).

Platynota Duméril and Bibron, 1839

(figs. 54E, 55E, 56C)

Definition

All taxa sharing a more recent common ancestor with Varanus varius and Heloderma horridum than with Anguis fragilis or Xenosaurus grandis.

Diagnosis

In the context of the current topology, this clade is diagnosed by 32(0) absence of a maxillary palatine flange, 86(0) dorsal origin of the jaw adductor musculature on the parietal, 178(0) dentary ventrally convex along its long axis, and 190(1) posterior terminus of the splenial at or anterior to the coronoid apex.

Comments

Platynota has been considered by some to be an equivalent for Varanoidea (McDowell and Bogert, 1954; Romer, 1956; Estes, 1983; Carroll, 1988b). However, this term has also been used to refer to a more inclusive group than Varanoidea (Rieppel, 1980a; Pregill et al., 1986; Lee, 1997; Bernstein, 1999; Conrad, 2004a, 2006a, 2006b) as defined here. Gao and Norell (1998, 2000) defined this taxon as a node anchored to Monstersauria and Varanidae, synonymous with the traditional definition of Varanoidea. Gao and Norell (1998, 2000), in turn, applied the taxon name Varanoidea to a node anchored by Telmasaurus grangeri and Varanidae of their definition.

Molecular data offer phylogenetic hypotheses in which Xenosaurus grandis and anguids are closely related to Heloderma suspectum, and Shinisaurus is closely related to varanids (Townsend et al., 2004; Vidal and Hedges, 2004). The study of Vidal and Hedges (2005) possessed weaker taxonomic sampling and did not analyze the phylogenetic position of Shinisaurus or Xenosaurus, but did support a sister-group relationship between Heloderma suspectum and Anniella pulchra. Given such topologies, Platynota and Varanoidea (below) are invalidated given their traditional understandings and their definitions here.

Platynota exclusive Dorsetisaurus pubeckensis

(figs. 54E, 55E, 56C)

Diagnosis

Platynotans exclusive of Dorsetisaurus purbekensis form a clade diagnosed by 55(1) fused frontals, 88(1) elongate supratemporal, 182(1) absence of a dentary shelf, 212(2) trenchant marginal teeth, and 217(1) presence of expanded bases on the marginal teeth.

Shinisauria comb. nov.

(figs. 54E, 55E, 56C)

Definition

All taxa sharing a more recent common ancestor with Shinisaurus crocodilurus than with Anguis fragilis, Heloderma suspectum, and Varanus varius.

Diagnosis

Shinisaurids and Dalinghosaurus longidigitus form a clade diagnosed by 3(1) presence of a blunt muzzle, 23(1) frontal forms a single anterior wedge, 34(1) presence of a prefrontal tuberosity, 38(1) anterior flange of the prefrontal extending anteriorly to the margin of the external naris, 50(1) dermal rugosities on the postorbital process of the jugal, 61(0) subolfactory processes present as simple ventral downgrowths, 138(1) reduced crista prootica without a lateral descending portion, and 307(0) ventral body osteoderms absent.

Comments

Evans and Wang presented analyses suggesting that D. longidigitus is the sister taxon to carusioids, the sister-taxon to Eichstaettisaurus in a clade that is the outgroup to carusioids, or that D. longidigitus is a basal scleroglossan (Evans and Wang, 2005: figs. 10, 12b, and 11b, respectively). The phylogenetic data matrices upon which these analyses were based (those presented in Gao and Norell, 1998; Lee, 1998; and a combination of Evans and Barbadillo, 1997, 1998, 1999; Evans and Chure, 1998b respectively) incorporated incomplete data from early descriptions of Shinisaurus crocodilurus specimens as discussed by Conrad (2003, 2004a, 2006a, 2006b). Incorporation of corrected morphological data for S. crocodilurus as well as the addition of Bahndwivici ammoskius (Conrad, 2006b) are partly responsible for the novel hypothesis presented here (figs. 54E, 55E, 56D).

Shinisauridae Ahl, 1930

(figs. 54E, 55E, 56C)

Definition

Shinisaurus crocodilurus, Bahndwivici ammoskius, and all descendants of their last common ancestor.

Diagnosis

Shinisaurids are diagnosed by 58(0) linear interorbital margins of the frontal and 205(0) retroarticular process lacking medial deflection (a reversal).

Comments

Shinisaurus has traditionally been considered a member of Xenosauridae (see above). Fan (1931) initially proposed the family Shinisauridae for Shinisaurus crocodilurus alone, but this group was demoted to subfamily by McDowell and Bogert (1954), a convention followed by many subsequent authors (Rieppel, 1980a; Gauthier, 1982; though see Hu et al., 1984). Given the extremely long missing history of the shinisaurid lineage and its potential to receive morphologically divergent taxa, the name Shinisauridae is applied here at the node containing S. crocodilurus and Bahndwivici ammoskius following Conrad (2006b).

Parasaniwa wyomingensis + Parviderma inexacta + Varanoidea

(figs. 54E, 55E, 56C)

Diagnosis

Platynotans exclusive of shinisaurs are diagnosed by 69(0) absence of parietal tabs on the frontal, 184(2) absence of a posterodorsal coronoid process on the dentary, 185(0) dentary without a notch distinguishing coronoid and surangular processes, 192(0) coronoid process short and broad, 194(1) anterior end of coronoid abutting the dentary (rather than overlapping it), 200(1) anterodorsally expanded surangular (vertical anterior margin), and 218(1) presence of plicidentine (as evidenced by internal view of teeth).

Comments

Although the Adams consensus for this analysis shows an unresolved trichotomy between Parasaniwa wyomingensis, Parviderma inexacta, and Varanoidea (fig. 55E), the volatile nature of some of the more incompletely known taxa that optimize in that tree as basal monstersaurs, mosasauriforms, or varaniforms, cause the strict consensus tree to show much less resolution (fig. 54E). The lack of resolution in the strict consensus tree derives mainly from the topological volatility of Colpodontosaurus cracens, Eosaniwa koehni, Necrosaurus cayluxi, “Necrosauruseucarinatus, Palaeosaniwa canadensis, Paravaranus angustifrons, Parviderma inexacta, and “Saniwafeisti. All of these taxa, except N. cayluxi, P. canadensis, and “S.feisti variably are recovered as basal platynotans (outside of Varanoidea), as basal goannasaurs and mosasauriforms (see below), or as basal mosasauriforms. Palaeosaniwa canadensis and P. inexacta are recovered in some trees as a clade near the base of Goannasauria (see below). “Necrosauruseucarinatus is the basalmost monstersaur in some trees. Proplatynotia longirostrata is variably an outgroup to varaniforms or a basal goannasaur. “Saniwafeisti and N. cayluxi are always contiguous on the tree, either as a clade or as a paraphyletic group. They may be a proximal outgroup(s) to goannasaurs, basal varaniforms, or basal varanines.

The Adams consensus tree shows that all of the taxa discussed above are always members of Platynota above the level of shinisaurs. All of the shortest recovered trees are consistent with the possibility that P. canadensis and “N.eucarinatus are monstersaurs, that “S.feisti, N. cayluxi, P. longirostrata, and Saniwides mongoliensis are goannasaurs, and that E. koehni and P. angustifrons are mosasauriforms; this is, therefore, reflected in the Adams consensus tree (figs. 55E, 56E).

Varanoidea Camp, 1923

(figs. 54E, 55E, 56D)

Definition

Heloderma horridum, Lanthanotus borneensis, Varanus varius, and all descendants of their last common ancestor.

Diagnosis

Crown group platynotans form a clade (figs. 53E, 54E, 55E, 56E) diagnosed by 26(1) strong medial processes of the maxillae behind the nasal process of the premaxilla, 29(1) weakly inclined anterior margin of the maxillary nasal process, 55(0) paired (unfused) frontals, 111(1) palatine subequal in length and width, and 189(1) splenial not extending anterior to the midpoint of the dentary tooth row.

Comments

The current definition of Varanoidea follows that of Estes et al. (1988), Norell et al. (1992), Lee (1997, 1998), Norell and Gao (1997), and Bernstein (1999), but uses the more specific anchor taxa as in Conrad (2006b). Gao and Norell (1998, 2000) used this taxon name in a more restricted sense, but theirs was not the common usage. See further comments at Platynota regarding usage of this definition.

Monstersauria Norell and Gao, 1997

(figs. 54E, 55E, 56D)

Definition

All taxa sharing a more recent common ancestor with Heloderma horridum than with Varanus varius.

Diagnosis

As defined in this analysis with the taxa included in Adams rule tree, this stem-based clade is diagnosed by 214(2) “modified pleurodont” dentition (sensu Zaher and Rieppel, 1999a) and 232(1) tall, narrow vertebral neural spines.

Comments

Gao and Norell (1998) defined Monstersauria as a node-based taxon anchored to Gobiderma pulchrum and Heloderma suspectum. They re-iterated this definition later (Gao and Norell, 2000), but described their intention to include in this group “…Helodermatidae and its closely related fossil taxa…” (Gao and Norell, 2000: 92–93). Thus, their definition was used to encompass all of the then-recognized fossil relatives of Heloderma. Although the principle trees recovered in this analysis always recover monsteraurian status for Gobiderma pulchrum, Paraderma bogerti, and Primaderma nessovi, the principle trees support various positions for these taxa. Some trees support the hypothesis that G. pulchrum shares a more recent common ancestor with Heloderma than Paraderma bogerti or Primaderma nessovi, which, under the node-based definition, would place them outside of Monstersauria. The current definition is intended to maintain and stabilize the originally intended use of the name Monstersauria.

Gobiderma pulchrum + Paraderma bogerti + Primaderma nessovi + Helodermatidae

(figs. 55E, 56D)

Diagnosis

This clade is diagnosed by 32(1) presence of a palatine flange of the maxilla, 160(0) presence of a pterygoid lappet on the quadrate, and 171(2) free posteroventral margin of the intramandibular septum.

Paraderma bogerti + Primaderma nessovi

(figs. 55E, 56D)

Diagnosis

These taxa are united in the current analyses by 309(1) presence of grooves separating the osteoderms on the maxilla.

Comments

Monophyly of Paraderma bogerti and Primaderma nessovi to the exclusion of all other squamates is not recovered in all of the principle trees in this analysis (see comments above and below), but the close relationship of these two taxa is recovered by the Adams consensus. Note, though, that the only unambiguous synapomorphy for this clade is homoplastic in that it also forms part of the diagnosis for the clade of helodermatids including Lowesaurus matthewi and Heloderma (below).

Helodermatidae Gray, 1837

(figs. 54E, 55E, 56D)

Definition

Heloderma horridum, Lowesaurus matthewi, Eurheloderma gallicum, and all descendants of their last common ancestor.

Diagnosis

The node here referred to as Helodermatidae is diagnosed by seven unambiguous synapomorphies in the current analysis. These are 37(1) prefrontal-postfrontal/postorbitofrontal contact present, 57(1) frontals rhomboid in dorsal view, 77(3) absence of a pineal foramen, 115(1) absence of palatine teeth, 118(2) absence of pterygoid teeth (reversed in some Heloderma suspectum), 182(1) absence of a dentary shelf (fig. 38E), and 220(1) presence of a distinct venom groove in the dentary teeth (fig. 38E).

Comments

Helodermatidae as defined here generally follows common usage (McDowell and Bogert, 1954; Bogert and Del Campo, 1956; Yatkola, 1976; Estes, 1983; Norell et al., 1992; Gao and Fox, 1996) and is similar to the usage of Pregill et al. (1986) who first applied this term to a node on a cladogram. This differs from the usage of some subsequent authors who prefer a crown based application of Helodermatidae (Norell and Gao, 1997; Gao and Norell, 1998, 2000).

Estes (1983) tentatively included Paraderma bogerti in his treatment of Helodermatidae, citing plesiomorphic characteristics and the then-forthcoming study by Pregill et al. (1986). Estes (1983) further cautioned that a parietal he had referred to P. bogerti (Estes, 1964) might be that of Palaeosaniwa canadensis. However, the parietal of P. canadensis is inconsistent with that morphology and the parietal is here tentatively considered to belong to P. bogerti. Pregill et al. (1986) included Heloderma, Lowesaurus ( = Heloderma) matthewi, Eurheloderma gallicum, and Paraderma bogerti in their Helodermatidae. The latter usage is consistent with that of Estes et al., who defined it as a stem including “Heloderma, and all organisms sharing a more recent common ancestor with this taxon than with any other extant organisms,” (Estes et al., 1988: 228). Given the phylogenetic topology presented here (figs. 54E, 55E) and a crown-based definition for Helodermatidae, helodermatids would be a subclade of Heloderma. Consequently, maintaining Helodermatidae in a traditional sense is advocated here.

The current application of Helodermatidae to the node specified above is beneficial given that all of the anchor taxa for the name were originally described as helodermatids (as opposed to “necrosaurids”/“parasaniwids”; e.g., Paraderma bogerti and Gobiderma pulchrum) and are retained as a monophyletic group in all of the principle trees recovered by this study (see also Augé, 2003). Estesia mongoliensis is variable in its placement in the current analysis. Different principle trees place it as the sister-taxon to L. matthewi, E. gallicum, the L. matthewi + Heloderma clade, or as the sister-taxon to Helodermatidae as defined here. This phylogenetic hypothesis owes, in part, to the relative quality of known E. gallicum and L. matthewi. Thus, E. mongoliensis may or may not be a helodermatid as the name is defined here, but it is certainly a monstersaur.

Lowesaurus matthewi + Heloderma

(figs. 55E, 56D)

Diagnosis

This unnamed clade is diagnosed by 62(1) midline contact of the frontal subolfactory processes and 309(1) presence of grooves separating the osteoderms on the maxilla.

Comments

Lowesaurus matthewi was initially described as Heloderma matthewi (Gilmore, 1928) and was, for many years, an unquestioned Heloderma species (Bogert and Del Campo, 1956; Pianka, 1967; Yatkola, 1976; Estes, 1983). However, Pregill et al. (1986) showed that they could recover no support for a sister-group relationship between the matthewi species and extant Heloderma exclusive of Eurheloderma gallicum, and erected a new genus for the Oligocene species.

Heloderma Wiegmann, 1829

(figs. 54E, 55E, 56D)

Definition

Heloderma suspectum (fig. 52), H. horridum, H. texana, and all descendants of their last common ancestor.

Diagnosis

Two unambiguous synapomorphies diagnose Heloderma as defined here. These are 57(0) anterior and posterior borders of the frontal of subequal width and 124(2) presence of a broad ectopterygoid-palatine contact anterior to the suborbital fenestra.

Heloderma horridum + H. suspectum

(figs. 54E, 55E, 56D)

Diagnosis

Extant members of Heloderma form a clade to the exclusion of Heloderma texana. This clade is diagnosed by 32(0) absence of a maxillary palatine flange.

Goannasauria tax. nov.

(figs. 55E, 56D)

Definition

All taxa sharing a more recent common ancestor with Varanus varius than with Heloderma suspectum.

Etymology

Goanna, a modification of “iguana” (derived from iwana, Arawak). Goanna is a common name for some members of the genus Varanus, especially in Australia. Sauros, (Greek) reptile. This taxon name is applied such that it is complementary to the Monstersauria within the Varanoidea. Thus, monstersaurs are Gila monsterlike varanoids and goannasaurs are goannalike varanoids.

Diagnosis

Goannasaurs are diagnosed by 2(1) presence of pre-septomaxillary rostrum, 8(0) absence of dermal sculpturing on the maxilla, 31(1) neochoanate condition, 47(1) jugal curved (rather than angulated), 65(1) frontals with an anterior maxillary contact, 113(0) absence of a distinct choanal groove, 178(1) straight main axis of dentary, 214(2) “modified pleurodont” dentition (sensu Zaher and Rieppel, 1999a), and 233(1) presence of precondylar constriction of the vertebrae.

Comments

Traditionally, no name has been given to the clade containing all the taxa more closely related to monitor lizards than to Heloderma. Recently, Gao and Norell (1998, 2000) redefined Varanoidea for this purpose, but this usage has not gained wide acceptance (see Platynota and Varanoidea above), nor did those papers specifically address mosasaurs. However, given the current topology (figs. 54E, 55E, 56E) (and some of previous analyses), naming this well supported clade seems helpful. The name Goannasauria is used here to balance the other varanoid stem, Monstersauria. Note that mosasauriforms may be goannasaurs or they may fall outside of Varanoidea. Goannasauria retains its general meaning (taxa more closely related to monitor lizards than Heloderma) regardless of this topological variability. Note that the strict consensus tree (fig. 54E) collapses the node containing Lanthanotus borneensis and Varanus exclusive of helodermatids. This is because of uncertainty regarding the phylogenetic placement of Palaeosaniwa canadensis, which is a monstersaur in many trees, but sometimes falls on the goannasaur lineage (for further discussion, see the comments for the clade Parasaniwa wyomingensis + Parviderma inexacta + Varanoidea, above).

Varaniformes comb. nov.

(figs. 54E, 55E, 56D)

Definition

All taxa sharing a more recent common ancestor with Varanus varius than with Heloderma suspectum or Mosasaurus hoffmanni.

Comments

Because the interrelationships of Varaniformes, Mosasauriformes, and Monstersauria are unresolved in this analysis, Varaniformes and Goannasauria are identical in their taxonomic content in the consensus trees presented here (see figs. 54Figure 5556). Many of the principle trees support a sister-group relationship between dolichosaurs and a clade containing varanids (usage below) and their fossil outgroups. In order to communicate easily about the nature of the latter group, the name Varaniformes is being applied as a stem-defined taxon name. Note that the principle trees recovered in this analysis always support the placement of Saniwides mongoliensis and Telmasaurus grangeri as proximal outgroups to Varanidae. Proplatynotia longirostrata, Necrosaurus cayluxi, and “Saniwafeisti are always goannasaurs, but, in some trees, are the outgroup to a clade containing mosasauriforms and varaniforms, and in others are basal varaniforms (fig. 54E; indicated also by the Adams consensus tree fig. 55E). Thus, although varaniforms cannot be diagnosed from other goannasaurs based on the current Adams consensus, it is a useful group based on goannasaur interrelationships.

Telmasaurus grangeri + Varanidae

(figs. 55E, 56D)

Diagnosis

Varanidae sensu stricto (below) forms a clade with Telmasaurus grangeri based on 9(0) absence of dermal sculpturing on the prefrontal, 43(1) presence of paired lacrimal foramina, 124(1) ectopterygoid contacting the palatine in the suborbital fenestra and 140(1) entocarotid fossa indistinct/absent.

Comments

This clade fits exactly with the definition Gao and Norell (1998) gave to the taxon name Varanoidea. Varanoidea as defined above and used by prior authors (e.g., Rieppel, 1980a; Pregill et al., 1986; Lee, 1997; Bernstein, 1999; Conrad, 2004a, 2006a, 2006b) is much more inclusive than the definition implemented by Gao and Norell (1998).

Varanidae Gray, 1827

(figs. 55E, 56D)

Definition

Varanus varius, Lanthanotus borneensis, and all descendants of their last common ancestor.

Diagnosis

Varanids are united by 115(1) absence of palatine teeth, 118(1) pterygoid teeth arranged in a single line, 149(1) presence of an expansive crista tuberalis (see character description above), and 151(1) anterior location for the spheno-occipital tubercle.

Comments

The present definition of Varanidae follows that of common usage and previous phylogenetic definitions (Estes et al., 1988; Pianka, 1995; Lee, 1998; Bernstein, 1999; Conrad, 2004a, 2006a, 2006b). The clade as defined here was left unnamed in Lee (1997), a study in which the name Varanidae was defined as all taxa sharing a more recent common ancestor with Varanus than with Lanthanotus borneensis. Caldwell (1999a) produced a hypothesis of relationships differing from that of the current analysis (fig. 8), but that is one of the only recent studies to do so.

Lanthanotinae Steindachner, 1878

(figs. 54E, 55E, 56D)

Definition

All taxa sharing a more recent common ancestor with Lanthanotus borneensis than with Varanus varius.

Diagnosis

Lanthanotines are diagnosed by 3(1) presence of a blunt muzzle, 37(1) prefrontal-postfrontal/postorbitofrontal contact present, 83(1) presence of a nuchal fossa on the parietal, and 193(0) absence of a labial flange of the coronoid.

Aiolosaurus oriens + Cherminotus longifrons

(figs. 54E, 55E, 56D)

Diagnosis

This unnamed lanthanotine clade is diagnosed by 12(2) mediolateral breadth of the premaxillary nasal process less than the dorsoventral depth and 43(0) presence of a single lacrimal foramen (a reversal).

Comments

The current analysis is the first to cladistically test the position of Lanthonotus borneensis with respect to fossil taxa such as Aiolosaurus oriens and Cherminotus longifrons. Cherminotus longifrons was originally considered to be a close relative of L. borneensis (Borsuk-Białynicka, 1984; Gao and Norell, 1998), but the specific position for A. oriens within Varanidae was not hypothesized in its original description (Gao and Norell, 2000). Gao and Norell (2000) showed that C. longifrons does not possess some of the character states that were used to join it with L. borneensis in the studies of Borsuk-Białynicka (1984) and Gao and Norell (1998). Despite this, the current analysis suggests that both C. longifrons and A. oriens are lanthanotines.

Varaninae Camp, 1923

(figs. 54E, 55E, 56D)

Definition

All taxa sharing a more recent common ancestor with Varanus varius than with Lanthanotus borneensis.

Diagnosis

Varanines are diagnosed by 42(1) presence of a posterolateral flange on the lacrimal, 44(1) enlarged lacrimal foramen (see character description above), 107(1) rod-shaped vomer, 194(0) anterior end of the coronoid clasping (rather than abutting) the dentary, 235(0) absence of true zygosphenes-zygantra, 257(1) presence of expansion/flattening of the anterior presacral ribs, and 281(1) absence of an intermedium in the manus.

Varanus White, 1790

(figs. 54E, 55E, 56D)

Diagnosis

The species of Varanus included in this study form a clade diagnosed as primitively possessing the following synapomorphies: 1(2) elongate antorbital snout, 12(2) nasal process of premaxilla narrowest mediolaterally, 17(1) presence of a premaxilla-maxilla aperture, 69(1) presence of a frontal-palatine contact, 74(1) presence of frontal tabs on the parietal dorsally, 118(2) absence of pterygoid teeth, and 243(2) fusion of the cervical intercentra to the posterior part of the preceding centrum.

Comments

Extant Varanus (e.g., fig. 49) are extremely diverse morphologically and ecologically (Mertens, 1942a; Irwin, 1994; Pianka, 1995; Ast, 2001, 2002; Pepin, 2001) and numerous monophyletic groups have been identified in separate morphological and molecular studies with a fair degree of consistency (Fuller et al., 1998; Ast, 2001, 2002; Pepin, 2001). Although these clades are often considered “subgenera,” they are typically geographically and morphologically distinct and could be considered relatively speciose genera of their own. Although this issue is not further addressed here, it is a topic deserving more attention.

The current analysis is the first to analyze the relative phylogenetic positions of Varanus rusingensis and Megalania prisca with respect to numerous extant species. Both are within the extant Varanus radiation. Varanus rusingensis is nested within a basal clade of Varanus. “Megalaniaprisca (hereafter referred to as Varanus priscum) is actually a species of Varanus (as the name is applied here) and deeply nested within that clade. In this analysis, Varanus priscum is suggested to form a clade with V. salvadorii, with V. komodoensis (fig. 49A) as the sister-taxon to that clade.

The phylogenetic positions of the major Varanus clades will be further analyzed using morphology and incorporating more complete taxon sampling elsewhere.

Mosasauriformes comb. nov.

(figs. 54E, 55E, 56D)

Definition

All taxa sharing a more recent ancestor with Mosasaurus hoffmanni (fig. 41) than with Varanus varius or Heloderma suspectum.

Diagnosis

Mosasauriforms are united by 61(0) subolfactory processes present as simple ventral downgrowths, 107(1) rod-shaped vomers, 172(1) absence of a distinct anterior surangular foramen, 178(1) straight main axis of dentary, 218(0) absence of plicidentine, and 303(1) deeply imbricating scales.

Comments

Cope (1869, 1870, 1872, 1878) erected Pythonomorpha to include mosasauroids to the exclusion of his Ophidia (snakes) which he thought to be their nearest relatives. Recently, the taxon name Pythonomorpha has been revived and defined to receive mosasauroids and snakes (Lee, 1997, 1998; Caldwell, 1999a; Lee and Caldwell, 2000; Rage and Neraudeau, 2004). That usage of Pythonomorpha is synonymous with Autarchoglossa in the present phylogenetic hypothesis. The definition given by Lee (1998) is not tenable given the present topology. Implementation of Cope's term Pythonomorpha would be preferable to creating a new taxon name or new combination for this clade, but the ubiquity of the recent usage of Pythonomorpha sensu Lee (1997) would be difficult to overcome. Therefore, rather than trying to reformulate or redefine Pythonomorpha to restore it to monophyly, the name Mosasauriformes is used here.

Paravaranus angustifrons + Mosasauria

(figs. 55E, 56D)

Diagnosis

These taxa are united, to the exclusion of Eosaniwa koehni, by 6(1) posteriorly elongated nares, 8(0) absence of dermal sculpturing on the maxilla, 9(0) absence of dermal sculpturing on the prefrontal, 10(0) absence of dermal sculpturing from the frontal and parietal, 47(1) jugal curved (rather than angulated), 55(1) fused frontals, 115(1) absence of palatine teeth and 189(0) a shortened splenial (see character description above).

Mosasauria Marsh, 1880

(figs. 54F, 55F, 56D)

Definition

Dolichosaurus longicollis, Coniasaurus crassidens, Coniasaurus gracilodens, Adriosaurus suessi, Mosasaurus hoffmanni, and all descendants of their last common ancestor.

Diagnosis

Mosasaurs are united by 66(0) prefrontals with large contributions to the orbitonasal fenestra, 69(1) presence of parietal tabs on the frontal, 168(1) adductor fossa with a distinct vertical flange, 191(1) presence of an intramandibular joint in which the splenial receives the angular, and 214(2) “modified pleurodont” dentition (sensu Zaher and Rieppel, 1999a).

Comments

There is no existing phylogenetic definition for Mosasauria. Traditionally, this taxon has been treated as an equivalent of Pythonomorpha (Marsh, 1880; Cope, 1900; Fürbringer, 1900a; Osborn, 1903a, 1903b, 1904; Hay, 1905; Camp, 1923; Russell, 1967), but given the current definition of that taxon (see Mosasauriformes above) a distinction between the two is worthwhile. All of the taxa included in this clade according to the current analysis are generally believed to be semi- to fully aquatic and are universally considered close relatives of mosasaurids; thus the name Mosasauria is defined here such that these taxa are included.

Dolichosauridae Gervais, 1852

(figs. 54F, 55F, 56D)

Definition

All taxa sharing a more recent common ancestor with Dolichosaurus longicollis than with Mosasaurus hoffmanni.

Diagnosis

Dolichosauridae as defined here in the context of the present analysis includes only Dolichosaurus longicollis and Aphanizocnemus, and is united by 243(2) cervical intercentra fused to the posterior part of the preceding intercentrum.

Comments

Lee and Caldwell (2000) coded Dolichosauridae based on Coniasaurus and Dolichosaurus longicollis, but their analysis does not resolve the position of Aphanizocnemus with respect to their Dolichosauridae, Adriosaurus, and/or snakes. Carroll (1988b: 618) included Adriosaurus, Dolichosaurus longicollis, Eidolosaurus trauthi, and Pontosaurus in his Dolichosauridae, but placed Coniasaurus tentatively with the Aigialosauridae.

The current analysis offers no unambiguous character support for uniting Coniasaurus and Dolichosaurus longicollis. The principle trees recovered in this analysis always recover a basal position for a clade containing Dolichosaurus longicollis and Aphanizocnemus, but the relative phylogenetic placement of Coniasaurus is more problematic. Coniasaurus is recovered as the sister taxon to Dolichosaurus longicollis, as the sister taxon to all other mosasaurs, as a nested mosasauroid (above the level of adriosaurids), or as the sister-taxon to Opetiosaurus within mosasauroids. Note that the existing cranial remains are very limited in most of the basal mosasaur taxa and that Coniasaurus is known primarily from partial skulls. Future inclusion of a Judeasaurus tchernovi, recently described by Haber and Polcyn (2005), may help to resolve this problem.

Mosasauroidea Camp, 1923

(figs. 54F, 55F, 56D)

Definition

All taxa sharing a more recent common ancestor with Mosasaurus hoffmanni than with Dolichosaurus longicollis.

Diagnosis

Mosasauroids are united by 163(1) presence of a quadrate suprastapedial process, 187(1) dentary suspended primarily from the prearticular, 249(1) lateral compression and dorsoventral deepening of the tail into a sculling organ, and 265(1) scapula shorter than the coracoid.

Comments

Marsh (1880) discusses “mosasauroids” in an informal fashion referring to Tylosaurus, Lestosaurus, and Holosaurus (the latter two taxa probably synonymous with Clidastes). Camp (1923) formalized Mosasauroidea and originally included only Mosasauridae. Recently, the name Mosasauroidea has been used to refer to clades of varying inclusiveness. Most treatments of the group have included “aigialosaurs” and Mosasauridae (Caldwell et al., 1995; Bell, 1997, 2002; Dal Sasso and Pinna, 1997; Lee, 1997, 1998; Caldwell, 1999a; Lee and Caldwell, 2000; Bardet et al., 2003). This is problematic because the monophyly of “aigialosaurs” has been questioned with respect to Mosasauridae, Adriosaurus, and Dolichosauridae (Bell, 1997; Dal Sasso and Pinna, 1997; Lee, 1997, 1998; Caldwell, 1999a; Lee and Caldwell, 2000; Bardet et al., 2003) and, consequently, because the composition of the “aigialosaur” group is uncertain. As originally conceived, Aigialosauridae included, among other taxa, Pontosaurus (Kramberger, 1892). Carroll (1988b) includes Coniasaurus in Aigialosauridae and places aigialosaurids in Mosasauroidea. Some of the studies mentioned above questioning the monophyly of the “aigialosaur” group have specifically omitted Dolichosaurus longicollis and/or Coniasaurus from their Mosasauroidea (Caldwell, 1999a; Lee and Caldwell, 2000) and others have specifically included them in discussing mosasauroids (Lee et al., 1999a).

Obviously, anchoring Mosasauroidea to “aigialosaurs” leaves a great deal of ambiguity; few studies have analyzed the monophyly of the clade and there is little support for it in general. Here, Mosasauroidea is defined as a stem-based taxon to help alleviate some of this uncertainty and to attempt to retain all of the “aigialosaurs” as mosasauroids (figs. 55F).

Adriosaurus + Pontosaurus

(figs. 54F, 55F)

Diagnosis

These taxa form a clade diagnosed by 33(0) maxillary tooth row extending posteriorly beyond the anterior one-fourth of the orbit, 207(1) retroarticular process that is posteriorly expanded (broadened), and 245(1) presence of pachyostotic dorsal vertebrae and ribs.

Aigialosaurus + Carsosaurus marchesetti + Eidolosaurus trauthi + Mosasauridae

(fig. 55F)

Diagnosis

The clade containing mosasauroids exclusive of Adriosaurus, Opetiosaurus, and Pontosaurus is diagnosed, in this analysis, by 13(1) premaxilla possessing an external contact with the frontal and 46(1) jugal extending well anterior to the level of the orbit.

Comments

The current analysis offers only very weak support for an aigialosaurid clade. Instead, “aigialosaurs” are a paraphyletic assemblage that is intermediate between dolichosaurids and mosasaurids. Some of the principle trees in this analysis support close relationships between Carsosaurus marchesetti and Opetiosaurus, between Carsosaurus marchesetti and Aigialosaurus, or between Aigialosaurus, Carsosaurus marchesetti, and Opetiosaurus. However, these relationships are not supported in the strict or Adams consensus trees (see comments below).

Eidolosaurus trauthi + Mosasauridae

(figs. 54F, 55F)

Diagnosis

Eidolosaurus trauthi, Tethysaurus nopcsai, and Mosasauridae form a clade diagnosed by 288(1) lack of distinction between the distal femoral condyles and 289(1) femur flattened and shortened such that the ends are more than one-quarter the length of the ends.

Mosasauridae Gervais, 1853

(figs. 54F, 55F)

Definition

Halisaurus platyspondylus, Tylosaurus proriger, Mosasaurus hoffmanni, and all descendants of their last common ancestor.

Diagnosis

Mosasaurids are diagnosed by two unambiguous synapomorphies: 247(1) absence of a functional sacrum and 275(1) presence of a flattened “hourglass-shaped” humerus.

Comments

The definition of Mosasauridae implemented here follows that of Bell (1997) who divided mosasaurids into Halisaurus and Natantia. The sister-group relationship of Halisaurus and all other mosasaurids has been recovered in several analyses (DeBraga and Carroll, 1993; Caldwell, 1996; Bardet and Suberbiola, 2001; Bardet et al., 2003, 2005), but has been questioned recently (Bell and Polcyn, 2005; Polcyn and Bell, 2005). This problem will require further examination given the recent description of new taxa such as Russellosaurus coheni (Polcyn and Bell, 2005) and Dallasaurus turneri (Bell and Polcyn, 2005). Note, however, that although some existing phylogenetic studies of mosasauroids have sampled densely for species than the current analysis (Bell, 1997; Bell and Polcyn, 2005; Polcyn and Bell, 2005), they have limited their taxonomic sampling to mainly North American and European taxa. The current analysis benefits from more complete taxonomic sampling from all over the world (e.g., Lakumasaurus antarcticus from Antarctica and Goronyosaurus nigeriensis from Africa), but does not include many of the American species represented in earlier analyses. A more definitive phylogenetic and taxonomic analysis should include all of these data.

Halisaurinae Bardet et al., 2005

(figs. 54F, 55F)

Definition

All taxa sharing a more recent common ancestor with Halisaurus platyspondylus than Mosasaurus hoffmanni.

Diagnosis

Halisaurus platyspondylus, H. arambourgi, and Eonatator sternbergii are united in this analysis by five unambiguous synapomorphies. These are 151(1) anterior location for the spheno-occipital tubercle, 164(2) presence of an infrastapedial process (also within plioplatecarpines), 172(0) presence of a distinct anterior surangular foramen, 235(0) absence of true zygosphenes-zygantra, and 288(0) presence of two distinct distal femoral condyles (a reversal of a synapomorphy uniting Mosasauridae and Eidolosaurus trauthi).

Comments

Halisaurinae is problematic. All of the halisaurines in this analysis have been considered Halisaurus at some point. Recently, Bardet et al. (2005) proposed a new genus, Eonatator, to receive the species originally described as Clidastes sternbergii and later placed in Halisaurus. The apparent reason for this change is strictly taxonomic. Bardet et al. (2005) found Eonatator sternbergii to be the sister-taxon to the clade Halisaurus; thus, the species sternbergii might as easily be retained in Halisaurus. However, it is not problematic to separate sternbergii at the generic level in their analysis or in the context of other recent analyses (Bell and Polcyn, 2005; Polcyn and Bell, 2005), because they also support sternbergii as the sister-taxon to other “Halisaurus.” The current analysis does not resolve the interrelationships between the three included halisaurine species (figs. 54F, 55F), but the future inclusion of the Halisaurus ortliebi might clarify these relationships.

The name Halisaurinae was recently defined as a stem-based name including taxa closer to Halisaurus than to Mosasaurus (Bardet et al., 2005), a definition maintained here but with more specific anchor taxa. Bell and Polcyn (2005) applied the name Halisauromorpha to a clade containing all of the taxa they considered to belong to Halisaurus (including Eonatator sternbergii) and cite the usage in Bell's unpublished dissertation (Bell, 1993), but do not formally define the group.

The current analysis leaves the placement of Tethysaurus nopcsai unresolved with respect to halisaurines and natantians (figs. 54F, 55F). The principle trees recovered in this analysis recover the placement of Tethysaurus nopcsai as the sister-taxon to Mosasauridae, as a basal halisaurine, or as a basal natantian. Bardet et al. (2003) originally recovered a sister-taxon relationship between Tethysaurus nopcsai and Mosasauridae. Two more recent analyses have suggested that Tethysaurus nopcsai forms a clade with Russellosaurus coheni and Yaguarasaurus columbianus near the base of Russellosaurina (see comments for Russellosaurina below).

Natantia Owen, 1849–1884

(figs. 54F, 55F)

Definition

All taxa sharing a more recent common ancestor with Mosasaurus hoffmanni, Tylosaurus proriger, and Plioplatecarpus marshi than with Halisaurus platyspondylus.

Diagnosis

Mosasaurids exclusive of halisaurines form a clade diagnosed in this analysis by 88(0) anteroposteriorly short supratemporal, 183(3) absence of a separate anterior inferior alveolar foramen, 212(0) teeth straight and pointed (conical), and 234(0) absence of oblique vertebral condyles.

Comments

Bell (1997) revived this name first proposed by Owen (1849–1884) to receive all of the non-Halisaurus taxa he (Bell) included in his analysis, but did not formally define the taxon name. It is unclear whether Bell (1997) meant for Natantia to include only his “Russellosaurinae” ( = Russellosaurina of Polcyn and Bell, 2005 and Bell and Polcyn, 2005) and Mosasaurinae or if he meant for stem-taxa of this radiation to be included as well. Although this question does not affect the taxonomic inclusiveness of Natantia under the current topology, it would become important were members of the stem-group (currently unknown) to be discovered. Because the apparent intention of Owen (1849–1884) was for the name to apply to all of the then-known mosasaurs, the more inclusive, stem-based, definition is implemented here. Natantia is used with some reservation given the absence of a clear Natantia in the strict consensus tree. However, the current definition allows Natantia to become a junior subjective synonym to Mosasauridae if Halisaurus is nested within Mosasauridae as suggested by some recent analyses (Bell and Polcyn, 2005; Polcyn and Bell, 2005).

Note that the name Natantia has also been applied to a group of decapods (Boas, 1880), but Owen's (1849–1884) usage has priority.

Mosasaurinae Gervais, 1853

(figs. 54F, 55F)

Definition

All taxa sharing a more recent common ancestor with Mosasaurus hoffmanni than with Tylosaurus proriger or Plioplatecarpus marshi.

Diagnosis

Mosasaurines are united by 53(1) frontal possessing an anterior constricted neck, 177(1) anterodorsal buttress for the coronoid, 250(2) humerus flattened with expanded, but equal, ends, 251(1) divided deltoid and pectoral crests joined by a lamina, 255(2) ulna possessing a facet articulating with the intermedium, and 257(1) notably robust first metacarpal of the manus.

Moanasaurus mangahouangae + Mosasaurus hoffmanni + Globidens dakotensis + Plotosaurini

(figs. 54F, 55F)

Diagnosis

In the present analysis, mosasaurines exclusive of Clidastes are united by 28(2) dorsal and ventral margins of the maxilla subparallel and 69(2) presence of large parietal tabs of the frontal that extend well onto the dorsal surface of the parietal (see fig. 34).

Mosasaurus hoffmanni + Globidens dakotensis + Plotosaurini

(figs. 54F, 55F)

Diagnosis

The present analysis recovers only one synapomorphy to support this clade: 219(1) striated tooth crowns.

Globidens dakotensis + Plotosaurini

(figs. 54F, 55F)

Diagnosis

This derived mosasaurine clade is united by 46(0) jugal extends no further anteriorly than if forming the anterior margin of the orbit.

Plotosaurini Russell, 1967

(figs. 54F, 55F)

Definition

All taxa sharing a more recent common ancestor with Plotosaurus bennisoni than with Mosasaurus hoffmanni or Globidens dakotensis.

Diagnosis

Plotosaurins are united by 61(1) subolfactory processes of the frontal(s) partly surrounding the olfactory bulbs.

Comments

Bell (1997) synonymized Mosasaurini with Plotosaurini based on the paraphyly of the nominal taxon Mosasaurus as Russell (1967) and several others had conceived of that clade. However, no taxon name at the level of “superfamily” or below may contain the type species of Mosasaurus (M. hoffmanni) and have their basis in a generic name other than Mosasaurus (International Commission for Zoological Nomenclature, 2000).

The current topology would yield Mosasaurini Gervais, 1853 exclusive of a Globidensini Dollo, 1924 and Plotosaurini Russell, 1967 monogeneric. Similarly, the current topology and taxonomic sampling would yield a Globidensini exclusive of Mosasaurus and Plotosaurini monogeneric (although Carinodens almost certainly would belong to Globidensini). Thus, neither Mosasaurini nor Globidensini are used here, and Plotosaurini is redefined. Russell (1967) originally included only Plotosaurus in his Plotosaurini. The name is applied here to all taxa that would not be subsumed by Globidensini or Mosasaurini if those clades are someday deemed necessary.

Goronyosaurus nigeriensis + Plotosaurus bennisoni

(figs. 54F, 55F)

Diagnosis

Goronyosaurus and Plotosaurus form a clade to the exclusion of “Mosasauruslemonnieri diagnosed by 15(0) absence of a rostrum anterior to the premaxillary teeth, 33(0) maxillary tooth row extending posteriorly beyond the anterior one-fourth of the orbit, 65(1) frontals with an anterior maxillary contact, 89(0) contact between the supratemporal arch bones and frontal and parietal unforked, 146(1) posterior opening of the Vidian canal located on the sphenoid-prootic suture, 235(0) absence of true zygosphenes-zygantra, and 276(2) distinct and completely separate deltoid and pectoral crests on the humerus.

Russellosaurina Polcyn and Bell, 2005

(figs. 54F, 55F)

Tentative Definition

All taxa sharing a more recent common ancestor with Russellosaurus coheni, Plioplatecarpus marshi, and Tylosaurus proriger than with Mosasaurus hoffmanni.

Diagnosis

Russellosaurinans, defined here in the context of the present analysis, are united by 83(0) absence of a parietal nuchal fossa and 249(0) dorsoventral height of tail vertebrae and chevrons less than three times the length of the associated centrum.

Comments

Bell (1997) applied the name Russellosaurinae to the radiation including Tylosaurus-like and Plioplatecarpus-like natantians, explaining that the name was based on a forthcoming description of a new taxon, “Russellosaurus.” Polcyn and Bell (2005) described Russellosaurus coheni and, recognizing the illegitimacy of Russelosaurinae as conceived of by Bell (1997), proposed the taxon name Russellosaurina for a clade minimally containing Plioplatecarpus and Tylosaurus. Unfortunately, the definition Polcyn and Bell (2005) intended for Russellosaurina is frustratingly ambiguous. Polcyn and Bell explicitly define the taxon name as, “Plioplatecarpinae, Tylosaurinae, their common ancestor and all descendants,” in their abstract (Polcyn and Bell, 2005: 321); a clear and concise, node-based definition for the clade name. However, under the “definition” for the taxon name, they define it as, “[a]ll mosasaurs more closely related to Tylosaurinae and Plioplatecarpinae, the genus Tethysaurus nopcsai, their common ancestor and all descendants than to Mosasaurinae,” (Polcyn and Bell, 2005: 322); an explicitly stem-based definition. Subsequently, they include plioplatecarpines, tylosaurines, “and closely related forms” (Polcyn and Bell, 2005: 322). Then, they define it to include some closely related forms (Polcyn and Bell, 2005), but in a way that is incongruous with the stem-based definition they offered on page 322. Further complicating the issue is the usage of Russellosaurina in Bell and Polcyn who first treated mosasaurines and russellosaurines as apparent sister-taxa (Bell and Polcyn, 2005: 188), then considered Russellosaurina to be limited to “Plioplatecarpinae plus Tylosaurinae” (Bell and Polcyn, 2005: 189). In their preferred cladogram, Bell and Polcyn (2005: fig. 7) included plioplatecarpines, tylosaurines, R. coheni, Tethysaurus nopcsai, and Yaguarasaurus columbianus in Russellosaurina.

The definition offered above is in keeping with that given in the definition of Polcyn and Bell (2005), but anchored to different taxa so that the taxon name retains its meaning regardless of the position of Tethysaurus nopcsai. The sister-group relationship between a Plioplatecarpus group and a Tylosaurus group exclusive of Mosasaurus is stable and, given the current definition, Russellosaurina will always be referable to that clade. Should new data suggest that Mosasaurus is nested within the Tylosaurus-Plioplatecarpus dichotomy, then Russellosaurina would be invalidated. This is acceptable given that the clade intended and understood when the Russellosaurina is invoked would be changed beyond recognition.

Tylosaurinae Williston, 1897

(figs. 54F, 55F)

Definition

All taxa sharing a more recent common ancestor with Tylosaurus proriger than with Mosasaurus hoffmanni or Plioplatecarpus marshi.

Diagnosis

Tylosaurines are united by 15(2) presence of an elongate, cylindrical, premaxillary rostrum, 162(1) short tympanic crest on the quadrate, 262(0) absence of anterior coracoid fenestra, and 279(2) presence of a short preaxial ridge on the radius.

Hainosaurus bernardi + Lakumasaurus antarcticus

(figs. 54F, 55F)

Diagnosis

Hainosaurus bernardi and Lakumasaurus antarcticus are united in this analysis (to the exclusion of Tylosaurus) by 12(1) nasal process of the premaxilla narrowest mediolaterally and 192(1) presence of a tall, narrow, coronoid process.

Comments

Novas et al. (2002) found Lakumasaurus antarcticus to be a basal tylosaurine, an issue that deserves further attention given the differing paleobiogeographical implications of the two hypotheses. Inclusion of Tylosaurus nepaeolicus may be helpful in resolving this issue.

Plioplatecarpinae Dollo, 1884

(figs. 54F, 55F)

Definition

All taxa sharing a more recent common ancestor with Plioplatecarpus marshi than with Tylosaurus proriger or Mosasaurus hoffmanni.

Diagnosis

Plioplatecarpines are united by 41(2) fusion of the lacrimal and prefrontal, 152(0) presence of a canal or groove for the basalar artery, 216(1) procumbent anterior marginal teeth, 275(3) presence of a flattened, rhomboid, humerus, and 283(1) manual digit V set off from other digits by 70 degrees or more.

Plioplatecarpus primaevus + Ectenosaurus clidastoides + Prognathodon

(figs. 54F, 55F)

Diagnosis

Five plioplatecarpines species were included in this analysis and Platecarpus tympaniticus is the basalmost taxon. The other four species are united by 72(1) anteroposteriorly broad lateral tabs of the parietal, 146(1) posterior opening of the Vidian canal located on the sphenoid-prootic suture, 164(2) presence of an infrastapedial process and 240(1) lateral processes of the atlas reduced.

Ectenosaurus clidastoides + Prognathodon

(figs. 54F, 55F)

Diagnosis

Ectenosaurus and Prognathodon are united to the exclusion of Plioplatecarpus in a clade diagnosed by 37(0) absence of a prefrontal-postorbitofrontal contact, 89(0) contact between supratemporal arch bones and frontal and parietal unforked, and 192(1) presence of a tall, narrow, coronoid process.

Prognathodon solvayi + Prognathodon overtoni

(figs. 54F, 55F)

Diagnosis

These two species of Prognathodon included in this analysis form a clade diagnosed by 178(0) dentary ventrally convex along its long axis and 219(1) striated tooth crowns.

Secondary Analyses

Osteology-Only Analysis

Some recent analyses based on subsets of this data set have recovered somewhat different phylogenetic hypotheses for specific parts of the squamate tree. Among these are analyses of Iguania (Conrad and Norell, 2007a; Conrad et al., 2007), Gekkonomorpha (Conrad and Norell, 2006a), and Anguimorpha (Conrad, 2006b). Because these analyses each focused on the placement of specific fossil taxa, they relied heavily upon osteological characters. These differing phylogenetic hypotheses (when compared with the current analysis) are also important because the current phylogenetic hypothesis is based on data obtained, in part, while researching those studies. Additionally, the current analysis relies heavily upon fossil taxa and non-osteological characters cannot be scored for those fossils. Because of all these things, both individually and in concert, I performed an osteology-only analysis. The results of this analysis are presented as fig. 60 (note that, where the osteology analysis is identical with the full analysis, some taxa have been collapsed into larger clades in the figure). Areas of divergence between the osteological analysis and the full analysis will be highlighted below.

Figure 60

(A–D) Hypothesis of squamate relationships based on the current data matrix with non-osteological characters omitted; Adams consensus. Some taxa have been collapsed where the topology of the clade in this analysis is identical to that presented for the full data matrix (see fig. 55); those taxa have been marked with pound signs (#). Fossil taxa denoted by daggers (†).

i0003-0090-310-1-1-f6001.gif

Figure 60

Continued.

i0003-0090-310-1-1-f6002.gif

Figure 60

Continued.

i0003-0090-310-1-1-f6003.gif

Figure 60

Continued.

i0003-0090-310-1-1-f6004.gif

The osteology-only analysis was run exactly as was the full analysis of all the characters and taxa. A total of 3,973 equally short trees were recovered, each with a length of 3,034 steps. Each of these trees had a consistency index (excluding uninformative characters) of 0.1371 and a retention index of 0.7100. Note that because the character/taxon ratio has decreased, a decrease in consistency index is also expected.

Iguania (fig. 60a)

Most cladistic analyses (morphology-based and molecular) have suggested a basal dichotomy between non-acrodontan iguanians ( =  Pleurodonta;  =  Iguanidae sensu lato) and Acrodonta (see figs. 2, 3, 6, 7, 13; see also Conrad et al., 2007, Conrad and Norell, 2007a), but analysis of this hypothesis has generally been relatively weak. However, the full analysis presented above suggests that Acrodonta is nested within non-acrodontan iguanians. The osteology-only analysis reproduces the hypothesis of a basal dichotomy between Acrodonta and Pleurodonta (sensu Schulte et al., 2003; Conrad and Norell, 2007a; Conrad et al., 2007), but still suggests that hoplocercids are close to acrodontans. Additionally, this analysis supports the presence of a Cretaceous radiation of iguanians from the Gobi (Conrad and Norell, 2007a), a hypothesis that was neither supported nor denied by the full analysis (figs. 54Figure 5556).

Opluridae is problematic. Opluridae is the monophyletic sister-taxon to a monophyletic Tropiduridae sensu Frost and Etheridge (1989) in the full analysis. Recent analyses have questioned this hypothesis (Conrad and Norell, 2007a; Conrad et al., 2007), as does the osteological data presented here (fig. 60A). Indeed, the osteology-only analysis suggests that Chalarodon madagascariensis is nested within Polychrotidae sensu Frost and Etheridge (1989) and that Oplurus is nested within Tropiduridae. Given the distribution of extant iguanians, this topology implies separate invasions of Madagascar by American clades of iguanians. However, many nested fossil “pleurodontans” are from Asia (e.g., Ctenomastax, Igua, Polrussia), suggesting that the biogeography of the group is more complex than it might appear based on extant taxa alone. Although this cannot be considered an argument in favor of oplurid polyphyly, it does offer some plausibility to the hypothesis.

The osteology-only analysis shows more complete resolution of the iguanian clades than the full analysis. For instance, it demonstrates the Cretaceous Gobi clade and resolves the tree supporting a hypothesis that Igua and Polrussia form a clade close to the tropidurid-Oplurus clade.

Gekkonomorpha (fig. 60b)

The full analysis suggested that Parviraptor (P. estesi and P. cf. estesi as described by Evans, 1994a) is a basal genus of gekkonomorph, falling between AMNH FR21444 and Gobekko cretacicus on the tree (figs. 54Figure 5556). However, Conrad and Norell (2006a) suggest that Parviraptor is a basal member of the lineage including Autarchoglossa and its stem taxa (Evansauria; see above). The osteology-only analysis, instead, suggests that Parviraptor is a scincogekkonomorph basal to the gekkonomorph-evansaur split at the level of (and in a polytomy with) Scandensia ciervensis. Gobekko cretacicus and AMNH FR21444 are recovered as basal gekkonomorphs and the topology within Gekkota is identical to that of the full analysis.

Conrad and Norell (2006a) also suggest that Gobekko cretacicus is nested within Gekkota and that pulls the minimum divergence time of the primary gekkotan lineages fall in the Cretaceous. The placement of Gobekko cretacicus as a proximal outgroup to Gekkota does not refute that hypothesis, but it does remove all of the evidence supporting it.

Scincomorpha (fig. 60b)

The osteology-only analysis does not recover a monophyletic Scincomorpha. Instead, Tepexisaurus tepexii, Scincoidea (including snakes, amphisbaenians, dibamids, acontids, and feyliniids), and a clade composed of Lacertoidea, Cordyloidea, Pseudosaurillus becklesi, “Pseudosaurillus” sp. sensu Estes, 1983, and Anguimorpha form a polytomy. In this topology, Parmeosaurus scutatus is a scincoid, Slavoia darevskii is the outgroup to Scincophidia, and Bainguidae is a basal radiation of Lacertoidea. Inclusion of bainguids in Lacertoidea and of Parmeosaurus scutatus in Scincoidea based on osteology is more in line with the traditional views of these taxa.

Carusioidea and Anguidae (fig. 60c)

Conrad (2006b) did not recover a monophyletic Carusioidea, but suggested that Carusia intermedia was a basal member of Anguimorpha (outside of the crown group). The current full analysis and the osteology-only analysis each recovers a Carusioidea, but the placement of that group varies between the two phylogenetic hypotheses. The full analysis places carusioids and anguids as a clade exclusive of Platynota, but the osteology-only analysis suggests that Carusioidea and the anguid clade (including glyptosaurs, see below) are successively more proximal outgroups to Platynota. The latter hypothesis is more similar to that of Gao and Norell (1998) (fig. 11) and Conrad (2005a, 2006b). Note that Shinisaurus is never a carusioid in the current analyses (figs. 54Figure 5556, 60) (contra Gao and Norell, 1998). The differences in topology between the present and previous analyses are probably related to the inclusion of numerous basal anguimorphs and/or scincomorphs (e.g., Becklesius, Paramacellodus, Parmeosaurus scutatus, Pseudosaurillus) and their effects on character polarities near the base of the anguimorph tree.

The full and osteology-only analyses show numerous minor differences in the placements of fossil anguids (including glyptosaurs). The phylogenetic placements of Apodosauriscus and Parophisaurus are unresolved in the osteology-only analysis. Parophisaurus is recovered in a trichotomy with Anguidae sensu stricto and glyptosaurs. Numerous differences exist between the two trees in the placements of specific glyptosaurs. Notably, the osteology-only analysis suggests that the “melanosaurs” are more closely related to Glyptosaurus and Proglyptosaurus than the latter two taxa are to the Helodermoides-clade. The addition of data for Placosaurus and related, non–North American glyptosaurines, may help resolve this issue.

The full analysis suggests that glyptosaurs are deeply nested within Anguidae, the sister group to diploglossines (fig. 54D, 55D, 56D), but the osteology analysis suggests that glyptosaurs fall outside of the anguid crown group. If the latter is true, then taxonomy becomes an issue for this group. As described above, anguids traditionally are defined as a crown clade, but glyptosaurs are usually considered anguids. Camp (1923) considered Anguidae and Glyptosauridae to represent distinct “families” (fig. 1), but that distinction has not been widely followed since McDowell and Bogert (1954). Regardless, this semantic issue will be investigated further elsewhere.

Platynota (fig. 60d)

The osteology-based analysis is generally quite similar to the full analysis for Platynota. Dorsetisaurus and shinisaurs remain platynotans in both hypotheses. Minor differences include the placement of some taxa and polytomies, such as the specific placement of Primaderma, within monstersaurs and of “Saniwafeisti and Necrosaurus cayluxi within goannasaurs. Importantly, the osteology-only analysis does not recover a mosasauriform position for Eosaniwa and leaves the placement Paravaranus unresolved with respect to varaniforms and mosasauriforms. However, the placement of mosasaurs within Goannasauria is confirmed.

Deformation Comparisons

Several previous analyses of squamate relationships are described above and major differences between the phylogenetic hypotheses are highlighted. The current phylogenetic hypothesis was further compared with two morphological analyses (Evans and Barbadillo, 1999; Lee and Caldwell, 2000) and two molecular analyses (Townsend et al., 2004; Vidal and Hedges, 2005) through deformation analyses of tree topology and length. Branches of the tree were manipulated and changes in tree length given the current data matrix were reported in the computer program Mesquite (Maddison and Maddison, 2006). Only the branches that were represented in each cited analysis (e.g., Evans and Barbadillo, 1999) were moved; the rest of the topology stayed as is within the current Adams consensus hypothesis (fig. 55).

Morphology

The current phylogenetic hypothesis, reformed such that it is consistent with that of Evans and Barbadillo (1999), lengthens the tree by 72 steps. To retrieve the topology presented by Lee and Caldwell (2000) requires an additional 106 steps. Making the limbed, Cretaceous, snakes form a basal clade outside of crown-group snakes (and changing nothing else) requires a tree 22 steps longer. Moving this clade into a position as the sister group to a polytomy with dolichosaurids, the Adriosaurus-Pontosaurus clade, and higher mosasauroids requires another 37 steps, for a total of 59 steps longer than the Adams tree presented here. Forcing shinisaurs into the more traditional position of being xenosaurids requires the addition of 10 steps.

Molecular Data

Only subtle differences are present between the studies of Townsend et al. (2004) and Vidal and Hedges (2005), but the taxonomic inclusiveness does vary. Forcing the current topology to reflect that of Townsend et al. (2004) requires an additional 175 steps. Deforming the Adams consensus of the current analysis to resemble that of Vidal and Hedges (2005) adds 171 steps to the tree.

Bremer Support

Numerous fossil taxa (e.g., Chamops, Sakurasaurus shokawensis, Restes) included in this analysis are so poorly known that only a few characters may be scored for them. These taxa were included if they were diagnostic from all other taxa in some way for reasons described above in the materials and methods section. However, the incomplete nature of these fossils means that they may be somewhat volatile within the phylogenetic tree (as evidenced in comparing the strict and Adams consensuses; figs. 54, 55). In many cases only one additional step is necessary to change their position in the phylogenetic topology and, thus, collapse a number of nodes. One way to deal with this problem would be to delete these problematic taxa. However, again as described above, their deletion might be deleterious to the analysis as a whole and is not desirable. Bremer supports are listed for the strict consensus tree (fig. 55).

Discussion and Conclusions

Phylogenetic Hypothesis

Squamate relationships as identified from the present analysis differ somewhat from all previous analyses. Some of the major differences will be highlighted below.

Basic Tree Structure

Importantly, if all extinct taxa are ignored, the basic structure of the tree (fig. 61) is similar to that of Estes et al. (1988) (fig. 2) with similarly applied names to the major clades. Iguania, Scleroglossa, Gekkota, Autarchoglossa, Scincomorpha, Lacertoidea, Scincoidea, Amphisbaenia, Dibamidae, Serpentes, and Anguimorpha are recovered as monophyletic. However, the current analysis resolves the position of amphisbaenians, dibamids, and snakes whereas Estes et al. (1988) does not (fig. 2A). Moreover, several new clades are recognized, including the Bainguidae and the Scincophidia.

Figure 61

The current phylogenetic hypothesis (Adams consensus) reduced to display only the major extant squamate clades. The accompanying higher taxonomy (in shades of gray along the right side of the cladogram) demonstrates that the taxonomy proposed here would create minimal disturbance for herpetologists and other neontologists.

i0003-0090-310-1-1-f61.gif

Gekkonomorphs, Scincomorphs, and Snake Origins

In contrast to the hypotheses put forward by Evans and Barbadillo (1998, 1999), the amphisbaenian-dibamid-snake clade is not closely related to geckos in this analysis. Xantusiidae are basal members of the Lacertoidea, in contrast the findings of some recent analyses (Presch, 1988; Evans and Barbadillo, 1998; Lee, 1998, 2000; Lee and Caldwell, 2000; Vicario et al., 2003).

Scincomorpha is found to be monophyletic, in contrast to some more recent analyses (Lee, 1998, 2000; Lee and Caldwell, 2000; Townsend et al., 2004; Vidal and Hedges, 2005). According to the present study, Scincomorpha includes Dibamidae, Amphisbaenia, and Serpentes. Thus, the present analysis does not support a close relationship between snakes and mosasaurs as has been suggested by some morphological analyses (Lee, 1997, 1998, 2000; Caldwell, 1999a; Lee and Caldwell, 2000; Lee and Scanlon, 2001; Scanlon and Lee, 2002; Caldwell and Dal Sasso, 2004) or with an anguimorph-iguanian group as suggested by molecular data (Townsend et al., 2004; Vidal and Hedges, 2005). Instead, snakes are nested in a group of limbless and limb-reduced scincoids, including feyliniids, acontids, dibamids, and amphisbaenians. Numerous synapomorphies support this hypothesis, but the lack of a fossil record for most clades is somewhat worrisome. It is possible that future discovery of basal members of any of those clades (Feyliniidae, Acontidae, or Dibamidae) may show that the known extant taxa are convergent in their morphology. Additionally, more inclusive taxon sampling will be necessary to analyze the position of limb-reduced gymnophthalmids and lacertids with regard to amphisbaenians, dibamids, and snakes.

Anguimorpha

Among anguimorphs, Xenosauridae is decidedly distinct from shinisaurs. Shinisaurs are found to be basal platynotans.

Although Caldwell (1999a) suggested that mosasaurs might fall outside of Scleroglossa, the current analysis supports the more common placement of mosasaurs as derived varanoids (contra Caldwell, 1999a). Importantly, some “necrosaurid” taxa are more closely related to the mosasaur clade than to any extant radiation (e.g., Varanidae or Shinisauridae).

Why the Differences?

Differences in topology between this and other recent analyses of squamate phylogeny (e.g., Estes et al., 1988; Wu et al., 1996; Evans and Barbadillo, 1997, 1998; Lee, 1998, 2000; Caldwell, 1999a; Lee and Caldwell, 2000) probably result from more taxon sampling in this analysis and, perhaps to a lesser degree, from character selection. Character selection probably bears less of the impact on differences in the topological tree than does taxonomic selection for many reasons. First, the present analysis and all those listed above draw heavily from the data set of Estes et al. (1988). Second, in addition to the inclusion of several new characters, the present analysis has been designed with the intention of including all of the nonredundant, informative, characters used in the described earlier studies. Thus, there is extensive overlap between previous analyses and this analysis. Third, taxonomic selection has varied widely in the previous analyses described and the current analysis has been designed with the intention of including all of the previously analyzed taxa. Fourth, new taxa have been incorporated in this analysis (e.g., Lakumasaurus, Parmeosaurus scutatus, Temujinia) that were not available to those researchers creating the earlier data matrices.

Scincophidia, tax. nov.

According to this analysis, Feyliniidae and Acontidae, and the more problematic Dibamidae, Amphisbaenia, and Serpentes form a clade termed Scincophidia within Scincidae sensu lato. This topology seemingly represents a marriage of thought between the traditional anatomical studies suggesting that dibamids and amphisbaenians are scincomorphs and the recent cladistic studies identifying a potential relationship between dibamids, amphisbaenians, and snakes. It is unsurprising that the scincophidian clade has been previously unrecognized given the taxonomic sampling of earlier analyses.

Lee (1998, 2000) argues that the similarities between snakes, dibamids, and amphisbaenians is an example of convergence influenced by a fossorial lifestyle. He states that “nearly all of the characters supporting this arrangement are correlated with head-first burrowing … and invariably co-occur in other tetrapods with similar habits” (Lee, 1998: 369). It follows, then, that the unusual suite of characteristics associated with burrowing will cause unrelated fossorial forms to cluster together on a cladogram and that all fossorial squamates would share most or all of these character states, possibly recovering an erroneous topology. This is certainly a legitimate concern requiring analysis.

Importantly, the only specialized headfirst burrowers Lee (1998, 2000) includes in his analyses are Pygopodidae, Amphisbaenia, and Dibamidae. He codes scincids sensu lato, Anguidae, and Serpentes (exclusive of Pachyrhachis problematicus) without breaking them into constituent clades. In doing so, he eliminates four major fossorial squamate radiations included here (Feyliniidae, Acontidae, anguines, and Scolecophidia) that could be used to further analyze his hypothesis of a convergent ecomorph. He also constrains Pachyrhachis problematicus to fall outside of crown-group Serpentes and, through his character codings, indicates his a priori assessment that fossorial snakes are not basal. Finally, Lee's (1998, 2000) answer to the perceived problem of the fossorial ecomorph is to downweight all of the characters he considers to represent fossorial adaptations. This requires an a priori judgment about which characters are associated with headfirst burrowing.

A New Test for the Fossorial Ecomorph

The phylogenetic analysis above presents a hypothesis in which headfirst, limb-reduced burrowing appears no fewer than four times (within Gekkota, twice within Anguidae, and within Scincoidea), demonstrating that Lee's (1998, 2000) ecomorph problem is not a major concern for the current data matrix; and even this hypothesis neglects to assess the placement of the limbless lacertoids and cordyliforms. A further analysis of the current data matrix was used here to determine the role of the fossorial ecomorph in the current analysis.

If Lee's (1998, 2000) strategy of downweighting characters contributing to the fossorial ecomorph is accepted, then making assumptions about which character states to include and therefore which characters to downweight or delete is problematic. In the context of the present analysis, it is unnecessary to determine exactly which characters might contribute to a fossorial lifestyle because extant forms are readily recognized as fossorial or not. Nullifying the impact of the fossorial ecomorph may be accomplished by deleting all taxa in the current analysis except for the limb-reduced, fossorial forms (extinct or extant and fossorial; see below). This was accomplished in two analyses by deleting taxa of variable limb robustness and two analyses including only those taxa that are suspected of being closely related to snakes. The limbed outgroup Rhynchocephalia was always retained. All limbless snakes were always retained because even snake taxa that are not exclusively fossorial practice some headfirst burrowing. (The strictly marine hydrophiines do not burrow, but are universally considered derived colubroids and were not specifically considered here.) Four analyses were then run with taxonomic inclusions as listed below.

  1. The first analysis included the ingroup taxa Acontidae, Amphisbaenidae, Anilioidea, Blanus, Dibamidae (Anelytropsis papillosus and Dibamus), Dinilysia patagonica, Feyliniidae, limbless anguids (Anniella, Anguis, Dopasia, Ophiodes, “Ophisaurusattenuatus, Ophisaurus ventralis, and Pseudopus), limbless macrostomatans, Pygopodinae (Aprasia, Delma, Pletholax, and Pygopus), Rhineuroidea, Scelotidae, Trogonophidae, and Wonambi naracoortensis.

  2. The second analysis included all the taxa from the first analysis and the bipodal taxa Bipes and pachyrhachids.

  3. The third analysis was used to analyze the position of mosasaurs when most limbed forms are deleted. All taxa were deleted except for those listed in the first two analyses and all of the Mosasauria.

  4. The final analyze excluded the limbless anguids, but included all of the other taxa included above and added the limbed gekkotans (Diplodactylinae, Eublepharidae, and Gekkonidae), Scelotidae, and Scincidae.

Snakes, amphisbaenians, and dibamids formed a clade in all of these analyses. Successively more distant outgroups in the first three analyses were pygopodines and a clade including Acontidae, Feyliniidae, and Scelotidae. The fourth analysis recovered a clade containing Scincidae sensu lato as the sister group to the snake-amphisbaenian-dibamid clade with a monophyletic Gekkota as the next outgroup. Limbless anguids were invariably monophyletic, but were found to be closer to the other limbless taxa than to the mosasaurs in the third analysis. Mosasaurs were monophyletic in both analyses in which they were included and always represented the basalmost ingroup lineage.

These results demonstrate the cohesiveness of the amphisbaenian-dibamid-snake clade even after nullification of fossorial/limbless characters. Snakes are not mosasaurs or even anguimorphs in any iteration of these analyses. Although the topology of the tree is somewhat different from that of the full analysis, this is not unexpected given the number of deleted taxa.

Certainly, being fossorial is a contributing factor to the morphology of scincophidians, but ecology is expected to be represented in morphology and phylogeny. Varanoids are usually predators of relatively large prey, chameleons are specialized for an arboreal existence, and gekkonids are typically crepuscular or nocturnal predators. These animals show heritable morphological adaptations for these behavioral and ecological aspects of their biology. The same should not be surprising in the fossorial clade Scincophidia.

Basal Scincogekkonomorphs and Evansaurs

The current phylogenetic hypothesis invites re-interpretation of some known squamate radiations pursuant to more precise understandings of geckos, scincomorphs, and necrosaurs. Based on current evidence, the traditional understanding of the diagnostic characters of these groups is insufficient for referral of many fossil taxa. That is to say, for example, that many taxa that have been described as scincomorphs do not represent members of a monophyletic Scincomorpha. Instead, these misidentified taxa are important transitional forms representing intermediate morphologies between the major squamate clades. These misidentified taxa demonstrate the incremental acquisition of character states along a much broader span of squamate phylogeny than to which they are usually attributed. For example, Ardeosaurus and Eichstaettisaurus represent not true geckos, but basal scincogekkonomorphs possessing some characteristics usually attributed to the Gekkota.

Ardeosaurus, and Eichstaettisaurus are representative examples of many taxa close to the main trunk of the squamate family tree whose morphologies show a mosaic pattern of primitive and derived character states when placed in the context of Scincomorpha proper and Gekkota proper.

Basal Gekkonomorpha?

Identification of “ardeosaurs” and “bavarisaurs” as basal scincogekkonomorphs rather than as stem-geckos demonstrates the poor quality of the gecko fossil record. Gobekko cretacicus remains the only well-preserved basal gekkonomorph described to date, but even this taxon seems very like modern geckos (see Borsuk-Białynicka, 1990; Conrad and Norell, 2006a) and offers little in the way of a transitional form between the basal scleroglossan condition and Gekkota. However, a new taxon from the Aptian-Albian of Mongolia (AMNH FR21444) was included in the current analysis and helps to polarize character states for Gekkonomorpha. This currently unnamed taxon possesses primitive characteristics such as a complete supratemporal arch and a toothed pterygoid, but also possesses characteristics shared with geckos and Gobekko cretacicus. Intermediate taxa such as this reduce the number of character states that may be used to diagnose a given clade by expanding the distribution of some characters and helping to bridge morphological gaps between previously known taxa. They also offer important insights into the relative timing of synapomorphy acquisition for clades and character evolution.

It is worth noting that Sereno (2006) does not consider AMNH FR 21444 (fig. 62) to represent a basal gekkonomorph, but suggests it as a possible basal squamate. He bases his assertion on “the narrow width of the nasals, the simple transverse frontoparietal suture, broad pyriform recess, and absence of [a] pterygoid-vomer contact” (Sereno, 2006:124A). Because Sereno (2006) is the only other study currently addressing this specimen, I will discuss this hypothesis and the characters used to support it here.

Figure 62

AMNH FR21444, a basal gekkonomorph whose morphology was recently described by Conrad and Norell (2006b). These are (A) dorsal and (B) ventral views of high-resolution x-ray computed tomography scans of the skull (anterior toward the top). The scans from which this figure was constructed are available online (Conrad and Norell, 2007b).

i0003-0090-310-1-1-f62.gif

Nasal width is difficult to assess outside the context of some comparison (that is, narrow relative to what?). Evans and Barbadillo (1998) compared nasal width to the width of the external nares, but it is unclear if Sereno (2006) is also making that comparison. Regardless, no other group or species otherwise hypothesized to be near the basal squamate condition (by this or other studies) has particularly narrow nasals (e.g., iguanians, Bavarisaurus, Eichstaettisaurus (fig. 29), dibamids (fig. 31), or basal rhynchocephalians (fig. 32)); indeed, narrow nasals seem to be a varanoid characteristic. Moreover, the nasals are not preserved in AMNH FR 21444.

The suggestion that AMNH FR 21444 has a transverse frontoparietal suture (Sereno, 2006) is erroneous (see illustrations and CT data in Conrad and Norell, 2006a, 2007b). Instead, this animal possesses a gently anteriorly arched frontoparietal suture.

A broad pyriform recess (as defined above, character 123) is plesiomorphic for iguanomorphs, gekkonomorphs, scincomorphs, and anguimorphs, with reversals in most of these groups. Presence of this character state in AMNH FR 21444 does not suggest that it is close to the basal squamate.

The specimen AMNH FR 21444 lacks as vomer-pterygoid contact as described by Sereno (2006). However, this character state is present in the majority of squamates (reversals within chamaeleontiforms and amphisbaenians, and in polyglyphanodontids and Shinisaurus crocodilurus).

Thus, none of the character states suggested by Sereno to place AMNH FR 21444 “just outside Squamata…” or “…at a basal position within Squamata,” (Sereno, 2006: 124A) actually support that hypothesis. Indeed, these character states would not be useful for placing any taxon at the base of Squamata. Given the six unambiguous synapomorphies uniting AMNH FR 21444 with other gekkonomorphs listed above, it currently is most prudent to consider AMNH FR 21444 a basal gekkonomorph.

Taxonomic Considerations

Stringency

Taxonomy is a tool for communicating about groups of things. Phylogenetic taxonomy has been, and will continue to be, an important tool with which to discuss organisms in a phylogenetic framework. Ideally, a single taxonomic scheme would be used for every given taxonomic group and that taxonomy would be based on the one true phylogeny of the group. Unfortunately, we are unlikely ever to know the one true phylogeny for any group with more than a few species. Therefore, taxonomists must be careful to make their nomenclatural schemes strict enough to be meaningful, but not so rigid that they are useless if some taxonomic content changes based on new discoveries and/or analyses.

Lee (1998) presents a taxonomic scheme in which any shifting of taxa between groups invalidates the meaning of two or more taxon names. This, or any similar stringency in taxonomy, makes taxonomy less useful as a tool for discussion of ideas or phylogeny.

Continuity and Superfluous Taxonomy

New phylogenetic hypotheses sometimes require revisions in taxonomy, but the taxonomy of the squamate “families” has been relatively stable for well over 100 years.

Wallace (1876a, 1876b) supplied a then-comprehensive list of squamate taxa including 2,256 species in 52 families. There is a remarkable correlation between that family list and the current understanding of squamate families, despite the fact that around 8,000 species of squamates are currently named (Uetz, 2007) (fig. 63). Of course, there are some differences, but most of these include further subdivision of currently recognized families by Wallace (1876a, 1876b), or vice versa rather than substantive differences in the included taxa. Other major differences include the recognition of the families Anguidae and Dibamidae, but Wallace's (1876a, 1876b) system remains useful even now. Similarly, Camp's (1923) suprafamilial taxonomy remains useful (see part of it in fig. 1).

Figure 63

Comparison of “family”-level squamate taxonomies from Wallace (1876a; 1876b) and current usage (modified after Behler and King, 1979, Whitfield, 1982, Bauer, 2003, Gans, 2003, Pianka and Vitt, 2003, Shine, 2003, and Uetz, 2007). There is remarkable consistency between the two lists, especially given the separation of approximately 130 years and since more than three times as many squamate species are now recognized. The major differences are mostly the result of identifying new squamate clades (in many cases, through the discovery of species) and subdivision of “families” or lumping them together. Clearly, taxonomy may be relatively constant and remain informative.

i0003-0090-310-1-1-f63.gif

Vidal and Hedges (2005) recently proposed a radically different phylogenetic hypothesis for squamate interrelationships (fig. 12B), and applied new taxonomy to some groups. The phylogenetic hypothesis of Vidal and Hedges (2005), similar to that of Townsend et al. (2004), is important and intriguing given the dissimilarities between those hypotheses and the usual ideas of squamate interrelationships (for example, compare with figs. 2Figure 3Figure 4Figure 5Figure 6Figure 7Figure 89, 53Figure 54Figure 5556). However, much of the new taxonomy presented by Vidal and Hedges (fig. 64) is unhelpful and gratuitous. Moreover, Vidal and Hedges (2005) offer no explanation for most of this taxonomy, leading to several problematic situations. Vidal and Hedges (2005) sampled within Dibamus, but applied to that branch the name Dibamidae (a name used to describe both Dibamus and Anelytropsis papillosus) and Dibamia. Thus, the name Dibamia becomes an apparent synonym of Dibamidae or, conceivably, Dibamus. Lacertidae is also labeled Lacertiformata. The clade containing teiids and Gymnophthalmus underwoodi is labeled both Teiioidea and Teiformata. The clade formed by Teiioidea, Lacertidae, and Amphisbaenia in their tree is labeled Laterata (Vidal and Hedges, 2005), even though this clade is essentially the same as the traditional idea of Lacertoidea (minus xantusiids; but see the usage of Lee, 1998 and Vicario et al., 2003). The Scinciformata of Vidal and Hedges (2005) is essentially the Scincoidea of Vicario et al. (2003) and is almost exactly the Scincoidea of Townsend et al. (2004).

Figure 64

The taxonomy of Vidal and Hedges (2005) as they apply it to their phylogenetic hypothesis. Taxonomy in gray to the right of the cladogram highlights redundant taxa and taxa applied inappropriately with respect to the content included in the analysis.

i0003-0090-310-1-1-f64.gif

Conclusions

The current study was undertaken with the intentions of supplying an extensive morphological phylogenetic data matrix for squamates, offering a phylogenetic hypothesis based on that matrix, and providing an updated and useful taxonomy. The data matrix provided here is the most taxonomically inclusive so far offered and it may be useful for morphologists as well as for systematists. The provided phylogenetic hypothesis will not be the last word on the subject of squamate phylogeny. Indeed, the matrix is already being expanded both in taxa and characters. Others may well analyze the provided data matrix differently, obtaining a different result. The phylogenetic hypothesis provided herein is no more than an accurate description of the data as it was analyzed. The taxonomy I provide reflects the usage of taxon names as I perceive them to be most often used. The definitions of existing taxon names in this study are offered only as a tool; a reference for the taxonomy as a whole. Taxonomy is most useful as a tool for discussing groups of animals, phylogenetic hypotheses, and ideas about evolutionary history.

Squamata is a clade of extraordinary diversity now and throughout its 210 million year history. The wide geographic distribution of squamates, their prominence in modern and fossil ecosystems, and their remarkable morphological diversity must rank them as one of the most important vertebrate clades for continued scientific study.

Skulls of three squamates suggested as close to the origins of snakes: Scincus scincus (Scincomorpha) (top); Cylindrophis ruffus (Serpentes) (middle); Mosasaurus hoffmanni (Anguimorpha) (bottom).

i0003-0090-310-1-1-f81.gif

Acknowledgments

This work would not have been possible without the support, guidence, and kindness of Olivier Rieppel, Mark Norell, Michael LaBarbera, and Neil Shubin. Rebecca Shearman has helped with various aspects of developing ideas, figure construction, and reading earlier drafts. This study also benefited from conversations with D. Allen, A. Balanoff, B.-A. Bhullar, S. Conrad, T. Conrad, J. Gauthier, P.S. Hood, J. Hopson, C. Ingram, C. Kammerer, M. Kearney, D. Kizirian, R. Lewis, T. Macrini, J. Maisano, S. Nesbitt, A. Shearman, T. Shearman, K. Smith, and A. Turner. I thank the Florida State Museum (University of Florida) for their loan of Shinisaurus and Xenosaurus material and R. Etheridge for his loan of polychrotid material. H. Voris, A. Resetar, and J. Ladonski (FMNH herpetological collections), W. Simpson (FMNH vertebrate paleontological collections), D. Frost and D. Kizirian (AMNH herpetological collections), and C. Mehling (AMNH fossil reptile collections) were all hospitable and helpful while I visited specimens in their care. R.L. Nydam and a second reviewer were kind enough to lend their time and considerable expertise reviewing and improving this manuscript. Any mistakes herein are, of course, my own. This research was supported, in part, by a Doctoral Dissertation Improvement Grant (DDIG 0408064) and a Kalbfleisch Postdoctoral Fellowship (AMNH).

References

1.

V. Abdala and S. A. Moro . 2003. A cladistic analysis of ten lizard families (Reptilia: Squamata) based on cranial musculature. Russian Journal of Herpetology 10:53–78. Google Scholar

2.

E. Ahl 1930. Beitrage zur Lurch und Kriechtierfauna Kwangsi: Section 5, Eidechsen. Sitzungsberichte der Gesellschaft der Naturforschenden Freunde zu Berlin 1930:326–331. Google Scholar

3.

M. E. Alfaro, D. R. Karns, H. K. Voris, E. Abernathy, and S. L. Sellins . 2004. Phylogeny of Cerberus (Serpentes: Homalopsinae), and phylogeography of Cerberus rynchops: diversification of a coastal marine snake in Southeast Asia. Journal of Biogeography 31:1277–1292. Google Scholar

4.

V. R. Alifanov 1989a. More ancient gekkos (Lacertilia, Gekkonidae) from the Lower Cretaceous of Mongolia. Paleontologicheskii Zhurnal 1:124–126. Google Scholar

5.

V. R. Alifanov 1989b. New priscagamids (Lacertilia) from the Upper Cretaceous of Mongolia, and their systematic postion among Iguania. Paleontological Journal 1989:68–80. Google Scholar

6.

V. R. Alifanov 1993a. New lizards of the family Macrocephalosauridae (Sauria) from the Upper Cretaceous of Mongolia, critical remarks on the systematics of the Teiidae (sensu Estes, 1983). Paleontological Journal 27:70–90. Google Scholar

7.

V. R. Alifanov 1993b. Revision of Tinosaurus asiaticus Gilmore (Agamidae). Paleontological Journal 27:148–154. Google Scholar

8.

V. R. Alifanov 1996. Lizard families Priscagamidae, and Hoplocercidae (Sauria, Iguania): phylogenetic position, and new representatives from the Late Cretaceous of Mongolia. Paleontological Journal 1996:100–118. Google Scholar

9.

V. R. Alifanov 2000. The fossil record of Cretaceous lizards from Mongolia. In M. J. Benton, M. A. Shishkin, D. M. Unwin, and E. N. Kurochkin , editors. The age of dinosaurs in Russia and Mongolia. 368–389.Cambridge Cambridge University Press. Google Scholar

10.

S. Apestiguía and H. Zaher . 2006. A Cretaceous terrestrial snake with robust hindlimbs, and a sacrum. Nature 440:1037–1040. Google Scholar

11.

J. C. Ast 2001. Mitochondrial DNA evidence, and evolution in Varanoidea (Squamata). Cladistics 17:211–226. Google Scholar

12.

J. C. Ast 2002. Evolution in Squamata (Reptilia). Ann Arbor Ph.D. dissertation, University of Michigan. 276. Google Scholar

13.

D. Attenborough 2002. The life of mammals. Princeton, NJ Princeton University Press. 320. Google Scholar

14.

W. Auffenberg 1981. The behavioral ecology of the Komodo monitor. Gainesville University Presses of Florida. 406. Google Scholar

15.

M. Augé 2003. La faune de Lacertilia (Reptilia, Squamata) de l'Éocène inférieur de Prémontré (Bassin de Paris, France). Geodiversitas 25:539–574. Google Scholar

16.

M. Augé 2005. Évolution des lézards. Mémoires du Muséum National d'Histoire Naturelle 192:1–369. Google Scholar

17.

M. Augé and R. Smith . 1997. Les Agamidae (Reptilia, Squamata) du Paleogene d'Europe occidentale. Belgian Journal of Zoology 127:123–138. Google Scholar

18.

M. Augé and R. M. Sullivan . 2006. A new genus, Paraplacosauriops (Squamata, Anguidae, Glyptosaurinae), from the Eocene of France. Journal of Vertebrate Paleontology 26:133–137. Google Scholar

19.

A. O. Averianov and I. G. Danilov . 1997. A varanid lizard (Squamata: Varanidae) from the Early Eocene of Kirghizia. Russian Journal of Herpetology 4:143–147. Google Scholar

20.

A. O. Averianov and P. P. Skutchas . 1999. Paramacellodid lizard (Squamata, Scincomorpha) from the Early Cretaceous of Transbaikalia. Russian Journal of Herpetology 6:115–117. Google Scholar

21.

A. Azzaroli, C. De Giuli, G. Ficcarelli, and D. Torre . 1975. Late Cretaceous mosasaurs from the Sokoto district, Nigeria. Accademia Nazionale dei Lincei, Rendiconti (8): Classe di Scienze Fisiche Matematiche e Naturale 13:21–34. Google Scholar

22.

A. Báez and Z. de Gasparini . 1977. Origines y evolución de los anfibios y reptiles del Cenozoico de America del Sur. Acta Geologica Lilloana 14:149–232. Google Scholar

23.

S. F. Baird 1859. Description of new genera and species of North American lizards in the museum of the Smithsonian Institution. Proceedings of the Academy of Natural Sciences of Philadelphia 1858:253–256. Google Scholar

24.

M. J. Balsai 2001. The phylogenetic position of Palaeosaniwa, and the early evolution of platynotan (varanoid) anguimorphs. Philadelphia Ph.D. dissertation, University of Pennsylvania. 253. Google Scholar

25.

L. J. Barbadillo and I. Martínez–Solano . 2002. Vertebral intercentra in Lacertidae: variation and phylogenetic implications. Copeia 2002:208–212. Google Scholar

26.

N. Bardet and X. P. Suberbiola . 2001. The basal mosasaurid Halisaurus sternbergii from the Late Cretaceous of Kansas (North America): a review of the Uppsala type specimen. Comptes Rendus de l'Académie des Sciences Série IIa Earth and Planetary Sciences 332:395–402. Google Scholar

27.

N. Bardet, X. P. Suberbiola, and N-E. Jalil . 2003. A new mosasauroid (Squamata) from the Late Cretaceous (Turonian) of Morocco. Comptes Rendus Palevol 2:607–616. Google Scholar

28.

N. Bardet, X. P. Suberbiola, M. Iarochene, B. Bouya, and M. Amaghzaz . 2005. A new species of Halisaurus from the Late Cretaceous phosphates of Morocco, and the phylogenetical relationships of the Halisaurinae (Squamata: Mosasauridae). Zoological Journal of the Linnean Society 143:447–472. Google Scholar

29.

S. Barrows and H. M. Smith . 1947. The skeleton of the lizard Xenosaurus grandis (Gray). University of Kansas Science Bulletin 31:227–281. Google Scholar

30.

A. Bartholomai 1979. New lizard-like reptiles from the Early Triassic of Queensland. Alcheringa 3:225–234. Google Scholar

31.

A. M. Bauer 1989. Extracranial endolymphatic sacs in Eurydactylodes (Reptilia: Gekkonidae), with comments on endolymphatic function in lizards. Journal of Herpetology 23:172–175. Google Scholar

32.

A. M. Bauer 2003. Lizards. In H. G. Cogger and R. G. Zweifel , editors. Encyclopedia of reptiles and amphibians. 126–173.San Francisco Fog City Press. Google Scholar

33.

A. M. Bauer, W. Böhme, and W. Weitschat . 2005. An Early Eocene gecko from Baltic amber, and its implications for the evolution of gecko adhesion. Journal of Zoology (London) 265:327–332. Google Scholar

34.

J. L. Behler and F. W. King . National Audubon Society field guide to North American reptiles and amphibians. 1979. New York Alfred A. Knopf. 743. Google Scholar

35.

B. A. Bell, P. A. Murry, and L. W. Osten . 1982. Coniasaurus Owen, 1850 from North America. Journal of Paleontology 56:520–524. Google Scholar

36.

C. J. Bell, S. E. Evans, and J. A. Maisano . 2003. The skull of the gymnophthalmid lizard Neusticurus ecpleopus (Reptilia: Squamata). Zoological Journal of the Linnean Society 139:283–304. Google Scholar

37.

G. L. Bell Jr 1993. A phylogenetic analysis of Mosasauroidea (Squamata). Austin University of Texas at Austin. 293. Google Scholar

38.

G. L. Bell Jr 1997. A phylogenetic revision of North American and Adriatic Mosasauroidea. In J. M. Callaway and E. L. Nicholls , editors. Ancient marine reptiles. 293–332.San Diego, CA Academic Press. Google Scholar

39.

G. L. Bell Jr 2002. The questionable monophyly of Mosasauridae. Journal of Vertebrate Paleontology 22:Suppl. 335A. Google Scholar

40.

G. L. Bell Jr and M. J. Polcyn . 2005. Dallasaurus turneri, a new primitive mosasauroid from the Middle Turonian of Texas, and comments on the phylogeny of Mosasauridae (Squamata). Netherlands Journal of Geosciences 84:177–194. Google Scholar

41.

A. Bellairs 1949. Observations on the snout of Varanus, and a comparison with that of other lizards and snakes. Journal of Anatomy 83:116–146. Google Scholar

42.

A. Bellairs 1970. The life of reptiles. New York Universe Books. 590. Google Scholar

43.

M. J. Benton 1985. Classification and phylogeny of the diapsid reptiles. Zoologica Scripta 84:97–164. Google Scholar

44.

P. Bernstein 1999. Morphology of the nasal capsule of Heloderma suspectum with comments on the systematic position of helodermatids (Squmata: Helodermatidae). Acta Zoologica (Stockholm) 80:219–230. Google Scholar

45.

C. A. Beuchat 1986. Phylogenetic distribution of the urinary bladder in lizards. Copeia 1986:512–517. Google Scholar

46.

G. S. Bever, C. J. Bell, and J. A. Maisano . 2005a. The ossified braincase and cephalic osteoderms of Shinisaurus crocodilurus (Squamata, Shinisauridae). Palaeontologia Electronica 8:1–36. Google Scholar

47.

G. S. Bever, C. J. Bell, and J. A. Maisano . 2005b. Shinisaurus crocodilurus. Austin, TX Digital Morphology.  http://digimorph.org/specimens/Shinisaurus_crocodilurus/adult. Accessed 2006.  Google Scholar

48.

B-A. Bhullar 2007. The enigmatic fossils Exostinus, and Restes: resolving the stem, and the crown of Xenosaurus, the knob-scaled lizards. Journal of Vertebrate Paleontology 27:Suppl. 348A. Google Scholar

49.

J. E. V. Boas 1880. Studier over deapodernes Slægtskabsforhold. Kongelige Danske Videnskabernes Selskabs Skrifter, 6 Række, Naturvidenskabelig og Mathematisk Afdeling 1:23–210. Google Scholar

50.

C. M. Bogert and R. M. Del Campo . 1956. The Gila monster and its allies: the relationships, habits, and behavior of the lizards of the family Helodermatidae. Bulletin of the American Museum of Natural History 109:11–238. Google Scholar

51.

F. C. Bonfirm-Júnior and L. S. Avilla . 2002. Phylogenetic position of Tijubina pontei Bonfim [sic] & Marques, 1997 (Lepidosauria, Squamata), a basal lizard from the Santana Formation, Lower Cretaceous of Brazil. Journal of Vertebrate Paleontology 22:Suppl. 337A–38A. Google Scholar

52.

F. C. Bonfirm-Júnior and R. B. Marques . 1997. Um novo Lagarto do Cretáceo do Brasil. Anuário do Instituto de Geociêcias 20:233–240. Google Scholar

53.

F. C. Bonfirm-Júnior and O. Rocha-Barbosa . 2006. A paleoautoecologia de Tijubina pontei Bonfirm-Júnior & Marques, 1997 (Lepidosauria, Squamata) basal da formação, Santana, Aptiano da Bacia do Araripe, Cretáceo Inferior do nordeste do Brasil. Anuário do Instituto de Geociências 29:52–63. Google Scholar

54.

K. Bonine 2005. Heloderma suspectum. Austin, TX Digital Morphology.  http://digimorph.org/specimens/Heloderma_suspectum/adult/. Accessed 2006.  Google Scholar

55.

M. Borsuk-Białynicka 1983. The early phylogeny of Anguimorpha as implicated by craniological data. Acta Palaeontologia Polonica 28:5–105. Google Scholar

56.

M. Borsuk-Białynicka 1984. Anguimorphans and related lizards from the Late Cretaceous of the Gobi Desert, Mongolia. Palaeontologia Polonica 46:5–105. Google Scholar

57.

M. Borsuk-Białynicka 1985. Carolinidae: a new family of xenosaurid-like lizards from the Upper Cretaceous of Mongolia. Acta Palaeontologica Polonica 30:151–176. Google Scholar

58.

M. Borsuk-Białynicka 1988. Globaura venusta gen. et sp. n., and Eoxanta lacertifrons lacertifrons gen. et sp. n. – non-teiid lacertoids from the Late Cretaceous of Mongolia. Acta Palaeontologica Polonica 33:211–248. Google Scholar

59.

M. Borsuk-Białynicka 1990. Gobekko cretacicus gen. et sp. n.: a new gekkonid lizard from the Cretaceous of the Gobi Desert. Acta Palaeontologica Polonica 35:67–76. Google Scholar

60.

M. Borsuk-Białynicka 1991. Questions and controversies about saurian phylogeny, a Mongolian perspective. In Z. Kielan-Jawarowska, N. Heintz, and H. A. Nacrem , editors. 5th Symposium on Mesozoic Terrestrial Ecosystems and Biota (extended abstracts). 9–10.Oslo University of Oslo. Google Scholar

61.

M. Borsuk-Białynicka 1996. The Late Cretaceous lizard Pleurodontagama and the origin of tooth permanency in Lepidosauria. Acta Herpetologica Polonica 41:231–252. Google Scholar

62.

M. Borsuk-Białynicka, M. Lubka, and W. Böhme . 1999. A lizard from Baltic amber (Eocene) and the ancestry of the crown group lacertids. Acta Palaeontologica Polonica 44:349–382. Google Scholar

63.

M. Borsuk-Białynicka and S. M. Moody . 1984. Priscagaminae: a new subfamily of the Agamidae (Sauria) from the Late Cretaceous of the Gobi Desert. Acta Palaeontologica Polonica 29:51–81. Google Scholar

64.

G. A. Boulenger 1883. Remarks on Nyctisauria. Annals and Magazine of Natural History (series 5) 12:308. Google Scholar

65.

G. A. Boulenger 1884. Synopsis of the families of existing Lacertilia. Annals and Magazine of Natural History (series 5) 14:117–122. Google Scholar

66.

G. A. Boulenger 1885–1887. Catalogue of the lizards in the British Museum (Natural History). 2nd ed. London British Museum (Natural History), 3 v. Google Scholar

67.

W. R. Branch 1982. Hemipeneal morphology of platynotan lizards. Journal of Herpetology 16:16–38. Google Scholar

68.

C. A. Brochu 1999. Phylogenetics, taxonomy, and historical biogeography of Alligatoroidea. Memoir: Society of Vertebrate Paleontology 6:9–100. Google Scholar

69.

A. Broschinski and D. Sigogneau-Russell . 1996. Remarkable lizard remains from the Lower Cretaceous of Anoual (Morocco). Annales de Paléontologie 82:147–175. Google Scholar

70.

D. A. Burnham 1991. A new mosasaur from the Upper Demopolis Formation of Sumter County, Alabama. M.Sci. thesis, University of New Orleans. 63. Google Scholar

71.

M. W. Caldwell 1996. Ontogeny and phylogeny of the mesopodial skeleton in mosasauroid reptiles. Zoological Journal of the Linnean Society 116:407–436. Google Scholar

72.

M. W. Caldwell 1999a. Squamate phylogeny and the relationships of snakes and mosasauroids. Zoological Journal of the Linnean Society 125:115–147. Google Scholar

73.

M. W. Caldwell 1999b. Description and phylogenetic relationships of a new species of Coniasaurus Owen, 1850 (Squamata). Journal of Vertebrate Paleontology 19:438–455. Google Scholar

74.

M. W. Caldwell 2000. On the aquatic squamate Dolichosaurus longicollis Owen, 1850 (Cenomanian, Upper Cretaceous) and the evolution of elongate necks in squamates. Journal of Vertebrate Paleontology 20:720–735. Google Scholar

75.

M. W. Caldwell and A. Albino . 2001. Paleoenvironment and paleoecology of three Cretaceous snakes: Pachyophis woodwardi, Pachyrhachis problematicus, and Dinilysia patagonica. Acta Palaeontologica Polonica 46:203–218. Google Scholar

76.

M. W. Caldwell and A. Albino . 2003. Exceptionally preserved skeletons of the Cretaceous snake Dinilysia patagonica Woodward, 1901. Journal of Vertebrate Paleontology 22:861–866. Google Scholar

77.

M. W. Caldwell, R. L. Carroll, and H. Kaiser . 1995. The pectoral girdle and forelimb of Carsosaurus marchesetti (Aigialosauridae), with a preliminary phylogenetic analysis of mosasauroids, and varanoids. Journal of Vertebrate Paleontology 15:516–531. Google Scholar

78.

M. W. Caldwell and J. A. Cooper . 1999. Redescription, palaeobiogeography, and palaeoecology of Coniasaurus crassidens Owen, 1850 (Squamata) from the Lower Chalk (Cretaceous; Cenomanian) of SE England. Zoological Journal of the Linnean Society 127:423–452. Google Scholar

79.

M. W. Caldwell and C. Dal Sasso . 2004. Soft-tissue preservation in a 95 million year old marine lizard: form, function, and aquatic adaptation. Journal of Vertebrate Paleontology 24:980–985. Google Scholar

80.

M. W. Caldwell and M. S. Y. Lee . 1997. A snake with legs from the marine Cretaceous of the Middle East. Nature 386:705–709. Google Scholar

81.

M. W. Caldwell and M. S. Y. Lee . 2004. Reevaluation of the Cretaceous marine lizard Acteosaurus crassicostatus Calligaris, 1993. Journal of Paleontology 78:617–619. Google Scholar

82.

C. L. Camp 1923. Classification of the lizards. Bulletin of the American Museum of Natural History 48:11289–480. Google Scholar

83.

C. L. Camp 1942. California mosasaurs. Memoirs of the University of California 13:1–68. Google Scholar

84.

C. L. Caprette, M. S. Y. Lee, R. Shine, A. Mokany, and J. F. Downhower . 2004. The origin of snakes (Serpentes) as seen through eye anatomy. Biological Journal of the Linnean Society 81:469–482. Google Scholar

85.

R. L. Carroll 1975. Permo-Triassic “lizards” from the Karroo. Palaeontologia Africana 18:71–87. Google Scholar

86.

R. L. Carroll 1977. The origin of lizards. In S. M. Andrews, R. Miles, and A. D. Walker , editors. Problems in vertebrate evolution. 359–396.London Academic Press. Google Scholar

87.

R. L. Carroll 1988a. Late Paleozoic and Early Mesozoic lepidosauromorphs, and their relation to lizard ancestry. In R. Estes and G. Pregill , editors. Phylogenetic relationships of the lizard families. 99–118.Palo Alto, CA Stanford University Press. Google Scholar

88.

R. L. Carroll 1988b. Vertebrate paleontology and evolution. New York W. H. Freeman. 698. Google Scholar

89.

R. L. Carroll and M. DeBraga . 1992. Aigialosaurs: mid-Cretaceous varanoid lizards. Journal of Vertebrate Paleontology 12:66–86. Google Scholar

90.

R. L. Carroll and P. Thompson . 1982. A bipedal lizardlike reptile from the Karoo. Journal of Paleontology 56:1–10. Google Scholar

91.

R. L. Carroll and R. Wild . 1994. Marine members of the Sphenodontia. In N. C. Fraser and H-D. Sues , editors. In the shadow of the dinosaurs: Mesozoic tetrapods. 70–83.New York Cambridge University Press. Google Scholar

92.

P. Christiansen and N. Bonde . 2002. A new species of gigantic mosasaur from the Late Cretaceous of Israel. Journal of Vertebrate Paleontology 22:629–644. Google Scholar

93.

R. L. Cifelli and R. L. Nydam . 1995. Primitive helodermatid-like platynotan from the Early Cretaceous of Utah. Herpetologica 51:286–291. Google Scholar

94.

L. M. Clos 1995. A new species of Varanus (Reptilia, Sauria) from the Miocene of Kenya. Journal of Vertebrate Paleontology 15:254–267. Google Scholar

95.

W. R. Coe and B. W. Kunkel . 1906. Studies on the California limbless lizard, Anniella. Transactions of the Connecticut Academy 12:23–53. Google Scholar

96.

J. L. Conrad 2003. Morphology of Shinisauridae compared with other Anguimorpha (Squamata, Reptilia). Integrative and Comparative Biology 42:140. Google Scholar

97.

J. L. Conrad 2004a. Skull, mandible, and hyoid of Shinisaurus crocodilurus Ahl (Squamata, Anguimorpha). Zoological Journal of the Linnean Society 141:399–434. Google Scholar

98.

J. L. Conrad 2004b. Re-analysis of anguimorph (Squamata: Reptilia) phylogeny with comments on some problematic taxa. Journal of Vertebrate Paleontology 24:Suppl. 347A. Google Scholar

99.

J. L. Conrad 2004c. Is the ‘stem-gecko’ body plan really plesiomorphic for Squamata. Journal of Morphology 260:284. Google Scholar

100.

J. L. Conrad 2005a. Shinisaur osteology and the evolution of Squamata. Ph.D. dissertation, University of Chicago. 682. Google Scholar

101.

J. L. Conrad 2005b. Evolution of ‘necrosaurids’ (Reptilia, Squamata), and their utility in reconstructing ancestral character states. Integrative and Comparative Biology 125:meeting abstracts143. Google Scholar

102.

J. L. Conrad 2006a. Postcranial skeleton of Shinisaurus crocodilurus (Squamata: Anguimorpha). Journal of Morphology 267:759–775. Google Scholar

103.

J. L. Conrad 2006b. An Eocene shinisaurid (Reptilia, Squamata) from Wyoming, U.S.A. Journal of Vertebrate Paleontology 26:113–126. Google Scholar

104.

J. L. Conrad and M. A. Norell . 2006a. High-resolution x-ray computed tomography of an Early Cretaceous gekkonomorph (Squamata) from Öösh (Övörkhangai; Mongolia). Historical Biology 18:405–431. Google Scholar

105.

J. L. Conrad and M. Norell . 2006b. A complete Cretaceous iguanian (Squamata) from the Gobi. Journal of Vertebrate Paleontology 26:Suppl. 351A–52A. Google Scholar

106.

J. L. Conrad and M. Norell . 2007a. A complete Late Cretaceous iguanian (Squamata: Reptilia) from the Gobi and identification of a new iguanian clade. American Museum Novitates 3584:1–47. Google Scholar

107.

J. L. Conrad and M. A. Norell . 2007b. Cretaceous gekkonomorph, AMNH 21444. Austin, TX Digital Morphology.  http://www.digimorph.org/specimens/AMNH_21444/. Accessed 2007.  Google Scholar

108.

J. L. Conrad, O. Rieppel, and L. Grande . 2007. An Eocene iguanian (Squamata: Reptilia) from Wyoming, U.S.A. Journal of Paleontology 81:1375–1383. Google Scholar

109.

W. E. Cooper Jr and L. J. Steele . 1999. Lingually mediated discriminations among prey chemicals and control stimuli in cordyliform lizards: presence in a gerrhosaurid, and absence in two cordylids. Herpetologica 55:361–368. Google Scholar

110.