Tikiguania estesi is widely accepted to be the earliest member of Squamata, the reptile group that includes lizards and snakes. It is based on a lower jaw from the Late Triassic of India, described as a primitive lizard related to agamids and chamaeleons. However, Tikiguania is almost indistinguishable from living agamids; a combined phylogenetic analysis of morphological and molecular data places it with draconines, a prominent component of the modern Asian herpetofauna. It is unlikely that living agamids have retained the Tikiguania morphotype unchanged for over 216 Myr; it is much more conceivable that Tikiguania is a Quaternary or Late Tertiary agamid that was preserved in sediments derived from the Triassic beds that have a broad superficial exposure. This removes the only fossil evidence for lizards in the Triassic. Studies that have employed Tikiguana for evolutionary, biogeographical and molecular dating inferences need to be reassessed.
Lizards and snakes (Squamata) become increasingly common in the fossil record from the mid-Jurassic (170 Ma) , but early fossils of their sister lineage, Rhynchocephalia (the group that includes the living tuatara), are known from the Upper Triassic [2,3]. The squamate–rhynchocephalian divergence therefore probably occurred during the Triassic [4,5], but the origin of crown squamates remains poorly constrained. Confirmation of a Triassic crown squamate would demonstrate that living lineages of lizards and snakes diversified from each other almost immediately after squamates split from rhynchocephalians, elucidate the early evolution of key squamate characters [6,7] and provide a critical calibration for molecular divergence dating studies of reptiles in general.
Tikiguania estesi  is based on an isolated left dentary bone from the lower jaw (described as a ‘nearly complete left mandiblular ramus’ [8, p. 795]), recovered from screen-washing of bulk material excavated from the Late Triassic (Carnian, 216–229 Ma) Tiki Formation of north-central India (Madhya Pradesh). It was identified as an acrodontan lizard, and is currently the only recognized Triassic squamate. (To reduce ambiguity, in this paper, ‘acrodont’ refers to a type of tooth implantation, whereas ‘acrodontan’ refers to the clade consisting of Agamidae and Chamaeleonidae.) Tikiguania has consequently been cited in a wide variety of reptile studies spanning molecular divergence dating , evolutionary morphology  and palaeobiogeography .
The similarity of Tikiguania to living acrodontans (see figure 1) was frequently mentioned , but strangely no explicit comparisons were made with any living squamates. Instead, comparisons emphasized acrodont-toothed rhynchocephalians and Mesozoic squamates: Bharatagama , a possible acrodontan lizard from the Early Jurassic of India, and priscagamids from the early late Cretaceous of Mongolia, currently regarded as a fossil sister group of living acrodontans . These comparisons demonstrated that Tikiguania was different from these ancient taxa, but what did not emerge was that the specimen is almost indistinguishable from certain living agamids, in particular the draconines. We here detail these similarities, and suggest a radical reinterpretation of Tikiguania which significantly impacts on all studies of reptile evolution involving this taxon.
2. Material and methods
Analysis 1: Testing possible affinities with rhynchocephalians (acrodont-toothed reptiles also abundant in the Triassic). Tikiguania was included in a broad-scale morphological and molecular matrix for the major lineages of lepidosauromorphs, with archosauromorphs as the outgroup. The analysis included three nuclear genes, rag-1, c-mos and BDNF (brain-derived neurotrophic factor) (total 4224 aligned sites), widely sequenced across squamates , and 233 morphological traits drawn from previous studies [7,10].
Analysis 2: The affinities of Tikiguania within acrodontans. Tikiguania was inserted into a matrix of 22 mandibular characters scored across a range of acrodontans, along with approximately 1700 bp of mitochondrial data from ND2 and adjacent loci, the longest region currently broadly sampled across acrodontans .
Whereas combining the ‘lepidosauromorph’ and ‘acrodont’ matrices into a single comprehensive dataset is a laudable ideal, splitting the higher level terminal taxa in Analysis 1 into species-level terminals proved impractical, and many subtle characters that are informative and rigorously definable in restricted groups (e.g. agamids) are difficult to score across phylogenetically remote and morphologically divergent taxa (e.g. snakes, dibamids, amphisbaenians, archosauromorphs).
Analysis 1 groups Tikiguania with Acrodonta and thus within crown Squamata (figure 2a). These results should be robust to additional data: the tree is consistent with trees obtained from molecular analyses of lepidosauromorphs with more nuclear genes and/or denser taxon sampling [15,16]. Tikiguana is not related to rhynchocephalians; in particular, the narrow tapering dentary without a posterodorsal ‘coronoid’ extension, reduced and dorsal symphysial contact point, as well as tooth implantation (successively pleurodont/acrodont/pleuroacrodont anteroposteriorly) are typical of acrodontan lizards and not seen in early rhynchocephalians .
Analysis 2 nests Tikiguania deeply within living acrodontans (figure 2), grouping with advanced agamids (i.e. above Leiolepis and Uromastyx) in all maximum parsimony trees (MPTs) and with draconines in most MPTs. This conclusion is consistent with the striking (albeit phenetic) similarity between Tikiguana and living draconine agamids. In particular, Tikiguania has several traits (table 1) placing it deep within acrodontans, some uniquely acquired and unreversed (i.e. consistency index = 1) within the present phylogeny. Tikiguania was described as distinct from the other acrodontans in lacking a coronoid process of the dentary . However, the only agamid explicitly mentioned was Uromastyx, which has a distinct crest rising vertically from the dentary that articulates with the labial surface of the coronoid bone. Other agamids, however, lack such a crest (see electronic supplementary material 2.3): they have at most a somewhat angular posterodorsal dentary margin that articulates with the base of the coronoid bone, not its ascending process.
Phylogenetic analysis of morphological and molecular data confirms that T. estesi is a crown acrodontan lizard with affinities to advanced agamids (agamids excluding Leiolepis and Uromastyx). The presence of an acrodontan lizard in the Tiki formation was not previously seen as problematic, as iguanians (including acrodontans) were widely considered basal squamates [6,7], and mitochondrial DNA studies had indicated a very ancient origin of acrodontans . However, although morphological evidence continues to robustly support a basal split between iguanians and other squamates [6,7,10], an increasing body of molecular and combined analyses [15,16,18,19] nest Iguania (including acrodontans) deeply within squamates. The acrodontan phylogeny of Macey et al.  has largely been supported by subsequent studies using nuclear genes, but this newer work suggests shallower (Cretaceous) branching times [19–22], consistent with a Triassic age for Squamata as a whole [18–22]. Any acrodontan—let alone an advanced agamid—in the Triassic is thus highly unexpected in the light of recent studies.
It is extremely unlikely that Tikiguania is an advanced agamid from the Triassic, and that the draconine jaw ‘morphotype’ has persisted largely unchanged for 216 Myr. Tikiguania came from a depth of 1.5 m within the Tiki Formation mudstone layers. As the specimen was screen washed from a load of five tonnes of excavated material, more precise depositional relationships are unknown. It shares with all of the fossil bones from this deposit a thin coating of haematite and calcite cementation, consistent with the specimen being interred in these sediments for some time, rather than a modern specimen or a reworked fossil . However, a Triassic age for Tikiguania does not necessarily follow. Erosion or fissuring into the Tiki Formation at any time during the Neogene or Quaternary would have allowed more recent faunal remains to have been incorporated into the Triassic mudstones, long enough to develop the characteristic chemical patina. The Tiki Formation is widely exposed at or near the surface across more than 70 000 km2 near Tiki and Beohari (fig. 1 in ), suggesting a long period dating back to the Late Tertiary at least where its sediments would have been sufficiently superficial to capture more recent animal remains. Consistent with this, Tikiguania shows very little damage to fragile bone margins or tooth crowns, and is extremely similar to living draconine agamids that have occupied this region during the Tertiary and still do today.
Molecular studies [9,25,26] that have employed Tikiguania as a calibration point all interpreted Tikiguania as if it were the sister of all other squamates rather than an agamid, but this misreading of its phylogenetic position, ironically, has so far resulted in a calibration point for the origin of squamates in the Late Triassic, as was generally supposed prior to the discovery of Tikiguania . We suggest avoiding this problematic fossil altogether. Morphological and palaeontological studies that have included Tikiguania also need to be revised. Tikiguania would have been evidence for an anomalously early (i.e. Triassic) age for what molecular studies suggest is a highly derived squamate clade (Acrodonta), implying that all major clades of squamates such as iguanians, anguimorphs, snakes, scincomorphs and gekkotans had diverged in the Triassic. However, none of these groups appear unequivocally in the fossil record until substantially later . Indeed, some recent palaeontological and molecular studies of squamate divergence dates have not mentioned Tikiguania, presumably because of its problematic nature [19,20]. Recognizing Tikiguania as essentially modern removes any potential need to assume early diversification and long ghost lineages for all major squamate clades.
We thank the Australian Research Council for funding, Dr. J. I. Mead (East Tennessee State University) for access to specimens and Sam Morrison for access to the Australian Research Collaboration Service compute cloud.
- Received December 13, 2011.
- Accepted January 6, 2012.
- This journal is © 2012 The Royal Society