Evolution of Squamata Reptiles in Patagonia based on the fossil record

Abstract

The Squamata, a vertebrate group that includes ‘lizards’, amphisbaenians and snakes, have inhabited Patagonia since the Late Cretaceous. The Mesozoic record comprises an iguanian, a scleroglossan and a great diversity of snakes that include, among others, the most primitive terrestrial forms. Snakes were also well diversified in the Palaeogene, with fossils of madtsoiids and boids (including the oldest record of the extant Boa) found as far south as 46°S, following the then dominant subtropical conditions in the region. The Palaeogene record comprises small, medium-sized and large forms. The only boids that persisted at the beginning of the Neogene are all small forms, suggesting colder temperatures than during the Palaeogene. The oldest South American ‘colubrids’ made their first appearance at this time, reaching latitudes south of the current Patagonian distribution (50°S). The Neogene record also includes the first occurrence of extant iguanians (Pristidactylus and Liolaemus) and teiids (Tupinambis). The latter reached the southernmost distribution of its evolutionary history during the Miocene. The distribution of boids, ‘colubrids’ and tupinambine teiids at higher latitudes than at present indicates warmer conditions during the Miocene. The uplift of the Patagonian Andes, followed by a decrease in temperature and an increase in desertification, induced a strong contraction in the distribution of tupinambine teiids to northern regions of Patagonia and even forced the complete disappearance of boids from this region. In contrast, these changes appear to have favoured the diversification of iguanian genera on both sides of the Patagonian Andes. © 2011 The Linnean Society of London, Biological Journal of the Linnean Society, 2011, 103, 441–457.

Los Squamata (‘lagartos’, anfisbenas y serpientes) habitan la Patagonia desde el Cretácico tardío. El registro Mesozoico comprende un iguanio, un esclerogloso, y una gran diversidad de serpientes que incluye las formas terrestres más primitivas. Las serpientes estaban bien diversificadas durante el Paleógeno, con madtsoideos y boidos (incluyendo el registro más antiguo de la actual Boa) que alcanzaron altas latitudes (46 °S) siguiendo las condiciones subtropicales dominantes en la región. En contraste con el Paleógeno, que incluye formas pequeñas, medianas y grandes, los únicos boidos que persisten a comienzos del Neógeno son todas formas pequeñas que sugieren temperaturas más bajas que durante el Paleógeno. Los más antiguos ‘colúbridos’ sudamericanos son del Neógeno, cuando alcanzan latitudes (50 °S) al sur de su actual distribución. El registro del Neógeno también incluye la aparición de géneros actuales de lagartos iguanios (Pristidactylus y Liolaemus) y teidos (Tupinambis). Tupinambis alcanza la distribución más austral durante el Mioceno. La presencia de boidos, ‘colúbridos’ y teidos tupinambinos a mayores latitudes que en el presente indica condiciones más cálidas durante el Neógeno. El levantamiento de los Andes Patagónicos, seguido por la disminución de la temperatura y el incremento de la aridez, produjo una fuerte contracción en la distribución de teidos tupinambinos hacia el norte y la desaparición total de boidos de Patagonia. En contraste, estos cambios parecen haber favorecido la diversificación de los géneros de iguanios a ambos lados de los Andes.

INTRODUCTION

The Squamata, by far the most diverse group of modern reptiles, includes more than 7000 extant species distributed among Ophidia (snakes) and ‘Lacertilia’ (‘lizards’ and amphisbaenians) (Zug, Vitt & Caldwell, 2001). The Patagonian fossil record within Squamata includes fossils of extant families and genera, as well as of extinct forms. Both types of fossils provide valuable information; those related to extant groups provide useful data to estimate the antiquity of some genera and their past distribution, and those of extinct forms illustrate the diversity of Patagonian squamatans in the past, and often contribute to the elucidation of phylogenetic relationships.

Despite its geologically and geographically biased nature, the fossil record of squamatans has been used to infer the history and radiation of the clade around the world. The record relies mainly on microvertebrate assemblages, supplemented by rare skeletons from lacustrine and other fine-grained deposits (Evans & Jones, 2010). With some exceptions, squamatans are relatively small animals with gracile skeletons that require a suitable depositional environment for preservation (low-energy, fine-grained sediments) and a search strategy appropriate for small fossils (Evans, 2003). Only the specimens found complete and fully articulated provide valuable data for the understanding of morphological evolution and phylogeny, but the fragmentary remains (mostly mandibles and vertebrae) support a comprehensive understanding of temporal and geographical radiation, as well as inferences of palaeoenvironments. For instance, the body size of fossil reptiles has been used to deduce palaeotemperatures (Head et al., 2009), inferences that follow from the fact that ambient temperature regulates maximum body size in poikilotherm vertebrates (Makarieva, Gorshkov & Li, 2005).

A number of previous studies have attempted to summarize and interpret the fossil record of South American herpetofauna. For instance, Báez & Gasparini (1977, 1979) provided a review of the fossil amphibians and reptiles for the whole of South America, and Gasparini, de la Fuente & Donadío (1986) concentrated on the Cenozoic reptiles of Argentina. These early reviews were, however, based on scarce, fragmentary and chronologically discontinuous records of squamatans supported by fossils mostly studied in the 19th century and lacking strong taxonomic identifications. As such, they did not provide the necessary data for a robust evaluation of the origin and evolution of squamatans in South America. A subsequent review (Albino, 1996a) listed several new findings, and provided better identifications at lower systematic levels of snakes in particular, but only contributed with general assumptions about the impact of palaeoenvironmental modifications in the evolution of squamatans in Patagonia. More recently, Albino (2007) discussed some palaeobiogeographical implications of the fossil record of Patagonian Mesozoic Squamata, and included a discussion on the evolution of basal snakes based on the possibility to ingest large prey. Although these last two studies (Albino, 1996a, 2007) give evidence for a higher diversity among fossil squamatans in South America (and, in particular, in Argentina) than previously known or suspected, they do not provide information on the origin and evolution of extant genera of the present Patagonian herpetofauna, which is one of the objectives attempted in this article.

Our current understanding of the squamatan fossil record in Patagonia is described herein, focusing on the biogeographical changes that have taken place among ‘lizards’ and snakes, first during the Mesozoic and then during the Palaeogene and Neogene. An attempt is also made to interpret the significance of the palaeoclimatic changes that took place in Patagonia for the diversity and distribution of some elements of the extant herpetofauna. The adopted phylogeny of Squamata is shown in Figure 1A, and the details of the phylogenetic relationships within the snakes are shown in Figure 1B.

Figure 1. A, Squamatan phylogeny (from Estes, De Queiroz & Gauthier, 1988). B, Snake phylogeny (modified from Apesteguía & Zaher, 2006).

REVIEW OF THE MESOZOIC RECORD

The Mesozoic record of squamatans is diverse and abundant, and includes both marine and continental taxa. Among the former, Mosasauridae were large predatory ‘lizards’ that inhabited warm, epicontinental seas during the Cretaceous. Nearly 20 genera are recognized, the largest specimens of which exceed 10 m long. Antarctic and Patagonian mosasaur remains are mostly limited to isolated teeth and fragmentary bones assigned to mosasaurines, plioplatecarpines and tylosaurines (Martin et al., 2002). Although the record is poor and fragmentary, the mere existence of Patagonian mosasaurine fossils demonstrates that the group inhabited the seas of the first Atlantic transgression over the northern part of Patagonia. Indeed, this record provides evidence for the latest presence of marine Mesozoic reptiles in the world (Gasparini et al., 2001, 2007).

With respect to continental squamatans, there are some remains from the Jurassic of Patagonia that were originally considered to be of ‘lizards’ (Evans, 2003), but have more recently been reinterpreted as Rhynchocephalia, the sister group of Squamata (Albino, 2007). The tuataras (genus Sphenodon) of New Zealand are the only surviving rhynchocephalians but, during Triassic and early Jurassic times, this group was relatively well represented, showing considerable diversity and a worldwide distribution (Evans, 2003). The record of rhyncocephalians in the Jurassic and Cretaceous of Patagonia (Apesteguía & Novas, 2003; Simón & Kellner, 2003; Martinelli & Forasiepi, 2004; Apesteguía & Rougier, 2007) suggests that they were an important component of the Mesozoic terrestrial ecosystems of South America, in contrast with the scarce presence of ‘lizards’.

The earliest unquestionable continental ‘lizard’ from the Patagonian Mesozoic is represented by an incomplete frontal of an iguanian found in early Late Cretaceous deposits (Fig. 2, 1) (Apesteguía, Agnolin & Lio, 2005). In addition, one incomplete dentary corresponding to a scleroglossan was found in Late Cretaceous rocks (Fig. 2, 2) (Brizuela, 2010). Indeterminate Late Cretaceous dentaries were also mentioned (Leanza et al., 2004; Albino, 2007), but they have been later recognized as Recent remains (S. Apesteguía, pers. comm.).

Figure 2. Distribution of Patagonian Mesozoic squamatans (modified from Albino, 2007): *, continental ‘lizards’; •, snakes; 1, La Buitrera (Apesteguía et al., 2005; Apesteguía & Zaher, 2006); 2, Cinco Saltos (Heredia & Salgado, 1999; Brizuela, 2010); 3, Neuquén (Smith-Woodward, 1901; Rage & Albino, 1989; Caldwell & Albino, 2002); 4, Paso Córdoba (Caldwell & Albino, 2002); 5, Barreales Norte (Albino, 2007); 6, Aguada Toledo (Scanferla & Canale, 2007); 7, Puesto La Rinconada (Filippi & Garrido, 2010); 8, La Colonia (Albino, 2000); 9, Los Alamitos (Albino, 1986); 10, Salina de Trapalcó and Santa Rosa areas (Albino, 2000; Martinelli & Forasiepi, 2004; Gómez et al., 2008); 11, El Palomar (Albino, 2000).

Although scarce, the Patagonian Cretaceous records of ‘lizards’, together with additional findings in Brazil (Estes & Price, 1973; Evans & Yabumoto, 1998; Bonfim-Júnior & Avilla, 2002; Candeiro et al., 2009), prove that different lineages of basal squamatans lived in South America during the Mesozoic. The presence of an acrodont iguanian in deposits of early–middle Jurassic age in India provides support for an early radiation of Iguania and thus for the separation of Iguania and Scleroglossa (see Fig. 1A) at or before the end of the Triassic, supporting the hypothesis that the first diversification of squamatans had already taken place before the fragmentation of the Pangaean supercontinent (Evans, 2003). The iguanian and scleroglossan remains found in the Patagonian Mesozoic are consistent with this hypothesis.

Iguanians have an important and poorly known Gondwanan history, followed by an expansion into South America during its isolation period. Living pleurodont iguanians are mainly distributed through the Americas, with outliers in Fiji, Tonga and Madagascar (Evans, 2003). The material found in the Cretaceous of Patagonia was identified as a pleurodont iguanian, supporting an early presence of this kind of ‘lizard’ in South America. Nevertheless, pleurodont iguanians are also present in the Mesozoic of Laurasian territories, suggesting that it is premature to attempt to reconstruct centres of origin or dispersal of this lineage (Evans, 2003).

In contrast with the record for ‘lizards’, remains of snakes, including those of the most primitive forms of terrestrial snakes, are common and diverse in Mesozoic deposits from Patagonia. This great diversity indicates a variety of palaeoecological roles and palaeoenvironments within the Patagonian Cretaceous ecosystems. The oldest of these snakes is Najash rionegrinaApesteguía & Zaher 2006, represented by cranial elements and a nearly complete and articulated postcranial skeleton that includes a sacrum associated with robust and well-developed hind limbs (Apesteguía & Zaher, 2006; Zaher, Apesteguía & Scanferla, 2009). A phylogenetic analysis of this fossil snake, found in sedimentary rocks of early Late Cretaceous age (Fig. 2, 1), shows Najash to be the most basal snake, lying outside the clade consisting of all living snakes (Fig. 1B) (Apesteguía & Zaher, 2006). According to Apesteguía & Zaher (2006), both cranial and vertebral morphological traits of Najash show adaptations to a subterranean life, perhaps as a surface-dwelling species that would occasionally use tunnels produced by burrowers. However, the use of tunnels constructed by burrowers does not require special morphological adaptation; thus, this habit cannot be established with certainty in the case of Najash. The vertebral size of this snake and the estimated body length of around 2 m (Apesteguía, 2007) are uncommon in subterranean snakes, which have smaller vertebrae (and thus smaller diameter) and shorter bodies (Albino & Caldwell, 2003). Therefore, a surface way of life cannot be rejected for Najash (terrestrial, semi-aquatic or semi-fossorial), and only extreme habits (aquatic and fossorial) can be obviated. Although the cranial morphology suggests that Najash had only limited unilateral movements of the mandibles, and lacked the attributes of macrostomatan snakes to ingest large prey, Apesteguía (2007) suggests, presumably based on the associated fauna, that this snake preyed on a variety of small vertebrates, such as ‘lizards’, micro-mammals and dinosaur hatchlings. This assumption is well supported by the snake body size.

Another significant snake from the Late Cretaceous is Dinilysia patagonica Smith-Woodward 2001. This species is known from fragmentary skulls and portions of articulated vertebral columns recovered from several sites in Patagonia (Fig. 2, 2–7) (Heredia & Salgado, 1999; Albino, 2007; Scanferla & Canale, 2007; Filippi & Garrido, 2010). Dinilysia was a medium-sized snake (1.5–2 m in length), with a relatively large head (approximately 10 cm in length), and large, dorsally exposed orbits (Albino & Caldwell, 2003; Albino, 2007). The species was interpreted as a partially terrestrial snake whose morphology may have been adaptable to semi-aquatic (seasonal lagoons and streams) or semi-fossorial (dune fields and interdune basin deposits) habits, although the body size of Dinilysia is more compatible with a semi-aquatic life style (Albino & Caldwell, 2003; Albino, 2007). Prior to the discovery of Najash, Dinilysia was postulated to be the sister group of all modern snakes (Lee & Scanlon, 2002), but an alternative analysis considered it in a more advanced position, as the sister group of alethinophidians (Rieppel, Kluge & Zaher, 2002). Both hypotheses place Dinilysia as a relatively basal snake, prior to the macrostomatan origin (Fig. 1B). Although Dinilysia had a relatively consolidated skull, its body size and head size were larger than those of living primitive snakes, and the elongation of the supratemporal that carries the quadrate posteriorly, enlarging the gape size, distinguished this snake from all living basal forms (Fig. 3). These characters would have improved the ability to prey on items larger than those possible for primitive extant snakes, such as scolecophidians and basal alethinophidians (Albino, 2007). Nonfossorial snakes, with bodies > 1 m long and consuming prey of diverse shapes and sizes, thus appear to have developed early in snake phylogeny (pre-macrostomatan), regardless of which hypothesis about the phylogenetic position of Dinilysia is embraced (Albino & Caldwell, 2003; Albino, 2007). The proposed eating role of the most basal snake Najash supports this idea, and suggests that the earliest evolution of terrestrial snakes occurred in Gondwanan territories. Recent discoveries in Late Cretaceous deposits from western India also corroborate this assumption. A partial skeleton of a 3.5-m-long surface living snake, named Sanajeh indicusWilson et al. 2010, was found in association with a sauropod dinosaur egg clutch, signifying that this snake frequented nesting grounds and preyed on hatchling sauropods (Wilson et al., 2010). Sanajeh indicus lacks the specializations of modern egg-eaters and of macrostomatans, and skull and vertebral synapomorphies place it in a pre-macrostomatan position in snake phylogeny (Wilson et al., 2010). These results suggest that large body size and jaw mobility allowed some non-macrostomatan snakes a greater diversity of prey items than previously suspected on the basis of extant basal snakes, corroborating the hypothesis of Albino (2007).

Figure 3. Skull of the Cretaceous snake Dinilysia patagonica; st, supratemporal bone.

The record of Patagonian Cretaceous snakes is enriched with the presence of some extinct genera belonging to the Madtsoiidae, which were found in several sites (Fig. 2, 8–11). The madtsoiids are interpreted as a probably monophyletic group of basal alethinophidians (Scanlon, 1993, 2005; Albino, 1996a), including small, medium-sized and large forms. Currently madtsoiids comprise at least 10 genera (see LaDuke et al., 2010), five of which have representatives in Patagonian deposits from the Late Cretaceous to Eocene (Simpson, 1933; Hoffstetter, 1959; Albino, 1986, 1993, 1994, 1996a, 2000, 2007; Martinelli & Forasiepi, 2004). Although madtsoiids are frequent in continental Patagonian deposits of Cretaceous age, the phylogenetic information they provide is limited because the remains mostly consist of isolated and fragmentary vertebrae. The genera of madtsoiids currently recognized in the Cretaceous of Patagonia include Alamitophis (with two species, A. argentinusAlbino, 1986 and A. elongatusAlbino, 1994), Patagoniophis (P. parvusAlbino, 1986) and Rionegrophis (R. madtsoioidesAlbino, 1986). Among them, Alamitophis and Patagoniophis are also recorded from the early Eocene of Australia (Scanlon, 1993, 2005). Also, a rib fragment referred as cf. Madtsoia sp. was found from the early Eocene of Australia (Scanlon, 2005). The genus Madtsoia is a madtsoiid well documented in Palaeogene deposits of Patagonia (see the following section). Thus, the distribution of these madtsoiid genera suggests a biogeographical continuity between Australia and southern South America across Antarctica between the latest Cretaceous and the early Eocene (Scanlon, 1993, 2005; Albino, 2000).

Another Patagonian Mesozoic snake is the anilioid Australophis (A. anilioidesGómez, Báez & Rougier, 2008) (Fig. 2, 10). Anilioidea is a probable paraphyletic group of basal alethinophidians, prior to the macrostomatan divergence (Fig. 1B) (Lee & Scanlon, 2002). The fossil history of the group is large with at least six extinct genera. In spite of its numerous records in most continents, the fossils assigned to the anilioids usually consist of isolated vertebrae, and the characters that support the identifications are mostly primitive. Australophis is morphologically closer to Palaeocene Hoffstetterella from Brazil and extant South American Anilius than to any other snake, probably indicating an early divergent lineage (Gómez et al., 2008). It is a small snake (probably less than 1 m long), with fossorial habits as the remaining anilioids, and with strong restrictions to the ingestion of prey that outsize its head diameter.

Only one vertebra has been recorded in Late Cretaceous Patagonian beds that is distinct from Najash, Dinilysia and madtsoiids, and that lacks morphological affinities with any extant group (Fig. 2, 8) (Albino, 2000). This vertebra has been referred to as belonging to Serpentes incertae sedis, although some plesiomorphic features were recognized (Albino, 2000, 2007). According to Scanlon & Hocknull (2008), this vertebra rather resembles a specimen of cf. Coniasaurus sp. of Australia, an aquatic ‘dolichosaur’ related to mosasaurs and snakes. If this is true, it would represent the only post-Cenomanian record of ‘dolichosaurs’, suggesting a persistence of these squamatans in Gondwana. The Patagonian deposit in which this vertebra was found (second facies association of La Colonia Formation) was interpreted to have been formed in an estuary, tidal flat or coastal plain environment; it was influenced both by occasional high freshwater streamflow from the continent and tidal currents from the sea (Albino, 2000). This palaeoenvironment is compatible with the aquatic habits of Coniasaurus. Although the area of origin of ‘dolichosaurs’ is probably the western Tethys (southern Europe and the Levant), Scanlon & Hocknull (2008) suggest that marine Coniasaurus-like forms expanded their range and entered rivers and estuaries along the northern and southern coasts of Tethys, ultimately establishing multiple distinct populations in freshwater environments (rivers and lakes) as far apart as Japan and Australia. Unfortunately, the only known vertebra of a possible Coniasaurus from Late Cretaceous Patagonia provides scarce morphological information to corroborate the presence of ‘dolichosaurs’, but it alerts to the possibility of colonization of South America by this group.

PALAEOGENE DIVERSITY

Despite the frequent presence of squamatans in Patagonian Cretaceous localities and the near-perfect preservation of some fossils providing relevant phylogenetic information, this group is represented in the Palaeogene with badly preserved remains. Nevertheless, the importance of this record is herein revised under a modern biochronological context based on the present calibration of the Palaeogene South American Land Mammal Ages (SALMAs) (Gelfo et al., 2009).

Although ‘lizards’ are absent from the Patagonian Palaeogene, snakes comprise representatives of madtsoiids and boids (Simpson, 1933, 1935; Hoffstetter, 1959; Albino, 1993). Palaeocene faunas of Patagonia, such as the well-known Peligran SALMA, have not provided remains of squamatans thus far, although the group, including ‘lizards’ and snakes, is present in other Palaeocene South American sites (De Muizon et al., 1983; Rage, 1992; Head et al., 2009). The records of Palaeogene snakes, the size diversity and their implications are summarized below.

The earliest record of snakes for the Patagonian Cenozoic (indeterminate madtsoiids and boids) comes from the Las Flores Formation, corresponding to the Early Eocene Itaboraian SALMA (Fig. 4, 1) (Albino, 1993). An extinct snake belonging to the genus Madtsoia (Madtsoiidae) is present in various deposits (Fig. 4, 2–5) corresponding to the early Eocene (Riochican SALMA) and middle–late Eocene (Casamayoran SALMA) (Simpson, 1933, 1935; Hoffstetter, 1959; Albino, 1993). A second Palaeogene genus, the boid snake Chubutophis, is restricted to the middle–late Eocene (Fig. 4, 6) (Albino, 1993). Both extinct snakes were very large snakes, Madtsoia reaching nearly 10 m in length (Hoffstetter, 1959) and Chubutophis probably even a little longer (Albino, 1993). A third example of a very large snake comes from an indeterminate Palaeogene deposit in the southern limit of the province of Chubut (Fig. 5) (Albino, 1991). Although the fragmentary preservation of this vertebra prevents a correct estimation of body size and taxonomic affinity, the vertebral centrum is larger than those of known vertebrae of Chubutophis and Madtsoia (Albino, 1991), making it the largest snake, both extinct and extant, recognized in Argentina. The large size of all of these snakes implies easy access to large prey (Albino, 1993). Apart from these snakes, Patagonian Palaeogene predators were limited to crocodiles, phorusrhacoid birds and borhyaenid mammals, all of large size; thus, the absence of other predatory forms and the notable diversity of large and medium-sized herbivore mammals could have played a fundamental role favouring the increase in predator size (Albino, 1993).

Figure 4. Distribution of Patagonian Palaeogene snakes (modified from Albino, 1993): 1, Las Flores (Albino, 1993); 2, Gaiman (Simpson, 1935; Hoffstetter, 1959); 3, Cañadón Hondo (Albino, 1993); 4, Cañadón Vaca (Simpson, 1933; Albino, 1993); 5, Colhué Huapi (Albino, 1993); 6, Valle Hermoso (Albino, 1993); 7, Aguada de Batistín (Albino, 1993).

Figure 5. A, Mid-trunk vertebra of the extant snake Boa constrictor. B, Mid-trunk vertebra of Chubutophis grandis from the Palaeogene of Patagonia. C, Vertebral centrum of an undetermined Patagonian Palaeogene snake.

Aside from these large snakes, medium-sized boids have been recorded from the middle–late Eocene of Patagonia (Fig. 4, 5, 7) (Albino, 1993). At least one specimen of these medium-sized snakes was referred to as Boa, constituting the oldest record of this extant genus (Albino, 1993; Albino & Carlini, 2008). Phylogenetic information suggests that Boa is the most basal extant lineage among Neotropical boids (Burbrink, 2005). Thus, the Eocene record implies that the initial cladogenesis of extant Neotropical boids would have occurred at least at the beginning of the Cenozoic (Albino & Carlini, 2008).

At present, the largest South American boid snakes do not extend south of 34°S, the latitude reached by the medium-sized Boa constrictor occidentalis Linnaeus 1758. The palaeontological record of Patagonian snakes demonstrates, however, that large and medium-sized snakes were not only much more widely distributed during the Palaeogene than at present, with some even reaching 46°S and probably beyond, but also included some of the largest snakes ever known (Albino, 1993).

The existence of large and medium-sized snakes and their wide latitudinal distribution during the Palaeogene in Patagonia suggest that regional palaeotemperatures were, at that time, significantly higher than today (Albino, 1993). Extant Neotropical boids live in areas in which mean annual temperatures lie above 16.5–17 °C. The largest Patagonian Palaeogene snakes mentioned above were probably larger than the largest snake in the modern neotropics, the extant green anaconda Eunectes murinus (Linnaeus 1758), with a maximum verifiable body length of 7.3 m (Head et al., 2009). This snake lives in temperatures for modern Neotropical lowland rainforests of 26–27 °C (Head et al., 2009). The Patagonian Palaeogene snakes were nearly as large as those from the northwestern South American Palaeocene, where a giant snake was recovered in deposits from Colombia. This snake, Titanoboa, was probably an aquatic boid with a body size estimated to be 13 m in length (Head et al., 2009), indicating a probable minimum mean annual temperature of 30–34 °C. Palaeocene temperatures in the southern ocean were about 20 °C, a typical value for subtropical oceans (Shackleton & Kennett, 1975), and fossil Patagonian Palaeogene mammals suggest the existence of tropical and subtropical forests (Ortiz Jaureguizar & Cladera, 2006), supporting the inferences based on the presence of these snakes in Patagonia.

A final example of a Patagonian Palaeogene snake is the relatively small boid Waincophis, known since the early Eocene Itaboraian SALMA of Brazil (Rage, 2001), and also found in the middle–late Eocene of Patagonia (Fig. 4, 3) (W. australisAlbino, 1987). It is a snake with probable semi-fossorial or fossorial habits, probably inhabiting loose soils.

The fossil record of squamatans is not homogeneous throughout the Palaeogene. Currently, a large temporal gap exists between the Casamayoran (middle–late Eocene) and the Colhuehuapian (early Miocene) of Patagonia, where no squamatans have been documented, except for some undescribed snake remains (Simpson, 1933; Simpson, Minoprio & Patterson, 1962; Albino, 1996a). The notable diversity of Palaeogene snakes from Patagonia, with at least one genus of madtsoiid and three genera of boids, contrasts with the absence of ‘lizards’ during the same period. A similar pattern is observed in Patagonian Mesozoic deposits, where snakes are represented by more than six genera and continental ‘lizards’ by only two taxa. However, the rich and diverse Itaboraian ‘lizard’ fauna of Brazil (Carvalho, 2001) demonstrates that the principal clades were diversified in South America since the early Eocene. These deposits are also rich in snake diversity, including both madtsoiids and boids (Albino, 1990, 1993; Rage, 1998, 2001). It is not possible to know at present whether the difference between Patagonian and Brazilian squamatan fauna reflects the lack of ‘lizards’ in southernmost Palaeogene environments or poorer samplings of small remains in Patagonian Palaeogene deposits than in those of Brazil. Early Eocene deposits of northwestern Argentina provided an association of a ‘lizard’ and a snake, strongly suggesting that ‘lizards’ were present in the Palaeogene of southern South America (Albino, 1989; Brizuela, 2010) and probably also in Palaeogene Patagonia.

THE NEOGENE AND EXTANT TAXA

Both ‘lizards’ and snakes are present in Neogene deposits of Patagonia. The ‘lizard’ record comprises the earliest representatives of Iguania and Teiidae Scleroglossa in the Cenozoic of this territory, including the first occurrence of extant genera. Relative to the Palaeogene, snakes have a moderately diverse Neogene record with the presence of boids and the more advanced ‘colubrids’.

The fossil-bearing beds of the Sarmiento Formation at Gaiman in Chubut Province (Fig. 6, 1) have provided one of the most relevant associations of Miocene squamatans, including the extant pleurodont iguanians Pristidactylus and Liolaemus, the modern teiid Tupinambis, the extinct boids Waincophis and Gaimanophis, and indeterminate ‘colubrids’ (Albino, 1996b, c, 2008; Brizuela & Albino, 2004). The remains were found associated with mammals characteristic of the Colhuehuapian SALMA, the calibration of which was proposed by Madden (2004) to be between 20 and 18.5 Ma at the Gran Barranca site (early Miocene). Other important localities of Miocene age in Patagonia are located in Santa Cruz Province (early–middle Miocene), where indeterminate iguanians, the teiid Tupinambis (and a probable Tupinambis) and ‘colubrids’ were identified (Fig. 6) (Ameghino, 1889, 1893; Albino, 1996c; Brizuela & Albino, 2008a). Among iguanians, Ameghino (1889) described the extinct genus Erichosaurus (under revision at present), whereas other remains have not yet been described.

Figure 6. Distribution of Patagonian Neogene squamatans: ★, continental ‘lizards’; •, snakes; 1, Gaiman (Albino, 1996b, c, 2008; Brizuela & Albino, 2004); 2, La Cueva (Ameghino, 1889); 3, Monte León (Ameghino, 1893; Brizuela & Albino, 2008a); 4, Monte Observatorio (Albino, 1996c); 5, Cañadón del Tordillo (Albino, 1996b; Brizuela & Albino, 2008b); 6, Río Pinturas (Albino, 1996c); 7, Pilcaniyeu Viejo (Albino, 1996b); - - -, southern range limit of extant tupinambine teiids; , southern range limit of extant ‘colubrids’.

‘Lizards’ of the genus Pristidactylus are endemic to Argentina and Chile, with 10 species now identified. Among them, four species occur west of the Andes, in Chile, whereas six are from the eastern side, in Argentina (Cei, Scolaro & Videla, 2001). Based on the distribution on both sides of the Andes, Cei (1986) proposed the origin of the genus in the Cretaceous or in the early Cenozoic, before the uplift of the mountain ranges. There are, however, no current records attributable to Pristidactylus in the Cretaceous or Palaeogene to support this hypothesis. Nevertheless, its presence in the early Miocene of Gaiman indicates that, at the beginning of the Neogene, this genus was well differentiated. The uplifting of the Andes appears to have produced vicariance on Pristidactylus populations. The persistence of the Nothofagus and Chusquea forests in southern Chile favoured the differentiation of the Chilean species of Pristidactylus at the west of the Patagonian Andes, whereas environmental modifications that took place on the eastern side, such as the reduction and disappearance of forests, contracted the Pristidactylus distribution in Patagonia to discontinuous rocky habitats of volcanic origin (Albino, 2008).

The other iguanian genus represented in Miocene deposits of Patagonia is Liolaemus, which includes around 224 extant species, collectively ranging in distribution from the high Andean mountains of Peru and Bolivia in the north to northern Tierra del Fuego Island in the south, and from the Pacific beaches in the west to the sandy Atlantic Ocean beaches of Argentina, Uruguay and Brazil in the east (Cei, 1986, 1993). Species within this genus comprise one of the most phenotypically diverse groups among ‘lizards’, exhibiting the highest latitudinal, altitudinal and climatic distribution known for ‘lizards’, with species inhabiting areas even over 5000 m a.s.l. (Cei, 1986, 1993). Using molecular analysis, Schulte et al. (2000) placed the initial divergence event within the Liolaemus lineage before 12.6 Ma. The earliest record for Liolaemus corresponds to the early Miocene of Patagonia (Gaiman), suggesting a minimum age for the genus of approximately 20–18.5 Ma (Albino, 2008), supporting the hypothesis of Schulte et al. (2000). The shortest estimate of phylogeny obtained by Schulte et al. (2000) suggests that Liolamus originated either in the Andes or the eastern lowlands. Numerous evolutionary shifts have occurred between the Andes, and the eastern and western lowlands, suggesting recurring vicariance with subsequent dispersal between regions, making further vicariance events possible. Thus, the earliest phase of the Andean uplift would have produced the initial divergence within this genus, followed by further divergence that mostly occurred before the Pliocene (Schulte et al., 2000). This suggests that changes in sea level, geodynamic events and climatic fluctuations occurring throughout the Miocene (Ortiz Jaureguizar & Cladera, 2006) could have favoured the diversification of the genus in the principal supraspecific clades (Albino, 2008).

Both Pristidactylus and Liolaemus are successful in the present ecosystems of Patagonia, contrary to the unique teiid genus recorded in the Miocene locality of Gaiman (Tupinambis).

Tupinambis is the largest genus of tupinambine teiid, having the broadest range of distribution through South America and exploiting tropical and subtropical environments. The earliest record of Tupinambis in the early Miocene of Gaiman (Brizuela & Albino, 2004) implies that the minimum age of the genus is approximately 20–18.5 Ma. Thus, these results suggest that the genus probably originated in a tropical or subtropical environment and spread all along South America in the middle Cenozoic. The southernmost distributional limit of the genus, and therefore of tupinambine teiids, is the Río Negro (Cei, 1986, 1993), with the exception of isolated populations of Tupinambis rufescens (Günther 1871) that live further south, in the Bajo del Gualicho area (Cei & Scolaro, 1982). Fossils referable to Tupinambis (Fig. 7) (and probable Tupinambis) were found in several Miocene localities of Patagonia, including Gaiman in the province of Chubut (early Miocene), Monte León in the province of Santa Cruz (early–middle Miocene) and Cañadón del Tordillo in the province of Neuquén (middle Miocene) (Fig. 6, 1, 3, 5) (Brizuela & Albino, 2004, 2008a,b). Monte León (50°19′S) and Cañadón del Tordillo (70°05′W) are the southernmost and westernmost localities, respectively, at which specimens of tupinambine teiids have been found (Fig. 6). All of these localities are located south of the present limit of distribution of the group and would have been characterized by climatic conditions warmer than today (Brizuela & Albino, 2008a, b). According to the presence of extant tupinambines, these conditions would have included temperatures above 14 °C (Brizuela, 2010). Palaeoclimatic studies show a climatic optimum during the middle Miocene, between 17 and 15 Ma (Böhme, 2003), with a slight increase in oceanic temperatures (Zachos et al., 2001). The southernmost extension in the distribution of tupinambine teiids coincides with the beginning of this optimum.

Figure 7. A–C, Teeth of Tupinambis sp. from the Miocene of Patagonia. D–F, Teeth of extant Tupinambis merianae. Scale bars equal 0.5 mm.

Apart from these ‘lizards’, South American ‘colubrids’ made their first appearance during Miocene times (Albino, 1996c). Remains of ‘colubrids’ were recovered from the locality of Gaiman (Fig. 6, 1), suggesting that the group entered South America, starting at least in the early Miocene, and presumably entering from North or Central America by dispersal across the islands that preceded the formation of the Panama isthmus (Albino, 1996c). Early–middle Miocene ‘colubrids’ were also found at latitudes as far south as 50°S, where they are not found today (Fig. 6, 4, 6). Such records constitute the southernmost location data for the group (Albino, 1996c), perhaps not surprisingly coinciding with the beginning of the Miocene climatic optimum.

Some of the palaeontological sites already mentioned also provide Neogene boid remains (Albino, 1996a, b). The small boid Waincophis, known since the Eocene, is reported in the Miocene localities of Gaiman (Fig. 6, 1), Cañadón del Tordillo (Fig. 6, 5) and Pilcaniyeu Viejo (Fig. 6, 7) (Albino, 1996b). The recovered vertebrae are smaller than those reported from the Eocene locality (Albino, 1987). Also, another boid found in the early Miocene of Gaiman is Gaimanophis, which would have had a slightly smaller size than Waincophis (Albino, 1996b). These snake fossils recovered from Miocene sediments of Patagonia are found outside the distributional range of modern boids, suggesting warmer climatic conditions at that time than at present. Nevertheless, the small size of these snakes contrast with the larger size of Palaeogene Patagonian snakes. Taking into account the fact that the snakes of the colder regions are generally of small size (Parker & Grandison, 1977), the presence of boids in the Patagonian Neogene that are smaller than those present in earlier periods would suggest a climatic deterioration, an interpretation that is in line with known changes in mean global annual temperatures reported for this period (Zachos et al., 2001).

Pascual et al. (1996) considered that the mammal history in South America would have been drastically affected during the Miocene by geodynamic events (Andean diastrophism), sea level changes and related climatic fluctuations, which would have caused the extinction of some higher latitude groups (e.g. primates) and an increase in the diversity of large cursorial forms following the expansion of the plains. These changes would have influenced significantly the evolutionary history and biogeography of South American squamatan taxa, as they did for mammals. The progressive decrease in temperatures through the middle–late Miocene and later (Ortiz Jaureguizar & Cladera, 2006) would have restricted the distribution of tupinambine teiids and boids to environments of more temperate climate. It is possible that the middle and late Miocene Atlantic marine transgressions (informally known as the ‘Paranaean Sea’; Pascual et al., 1996), that extended from northern Patagonia to southern Paraguay and from the eastern slopes of the Andes to western Uruguay, have favoured the diversification of Tupinambis (Albino, Brizuela & Montalvo, 2006). It is probable that T. rufescens differentiated in the most arid environments of the south, and then, following the development of widespread and varied plains, dispersed to the north, whereas T. merianae differentiated in the warm and humid environments in eastern Brazil, south of Amazonia, and colonized the Chaco biome during more recent periods (Albino et al., 2006). The southernmost present population of T. rufescens in northern Patagonia (Cei & Scolaro, 1982) is probably a relic of the more widespread Miocene distribution.

During late Miocene–early Pliocene times, the widespread surface flooded by the ‘Paranaean Sea’ was replaced by likewise ample, extended plains (Pascual et al., 1996). This period is known as ‘the Age of the Southern Plains’ (Pascual et al., 1996). The beginning of these new habitats is correlated with the ‘Quechua Phase’ of the Andean diastrophism and the uplifting of the Patagonian Andes and the Principal Cordillera (Ortiz Jaureguizar & Cladera, 2006). During this period, the climate was cooler than previously, and seasonality was more marked, with a more varied environmental subdivision (Pascual et al., 1996). The new available environments would have favoured the iguanian diversification on both sides of the Andes, especially for Liolaemus, in which the most divergent events would have arisen during the entire Miocene (Schulte et al., 2000).

QUATERNARY

The present Patagonian herpetofauna includes around 85 species of ‘lizard’ (Scolaro, 2005, 2006), 65% of which correspond to the Liolaemus genus. Other ‘lizards’ inhabiting Patagonia are the pleurodont iguanians Diplolaemus, Leiosaurus, Pristidactylus and Phymaturus, the geckonid Homonota and the teiid Cnemidophorus (Scolaro, 2005, 2006). Tupinambis, Teius, Stenocercus and Ophiodes are marginal in this region (Scolaro, 2005, 2006). With the exception of Liolaemus, Pristidactylus and Tupinambis, none of these ‘lizards’ are represented in the Patagonian fossil record.

Some Patagonian archaeological sites provide remains of Quaternary pleurodont iguanians, probably mostly corresponding to the Liolaemus genus, but identifications of the remains at specific levels are impossible at the moment (A. M. Albino, pers. observ.) preventing the reconstruction of the latest biogeographical history of the genus based on the fossil record. It is not possible to deduce whether Pleistocene climatic changes (e.g. glaciations) could influence the mechanisms and processes of divergence in the evolutionary history of Patagonian Liolaemus, but, interestingly, a large proportion of species occur in high-latitude and high-elevation environments, where many taxa appear to be still undescribed (Pincheira-Donoso et al., 2008). In these cold habitats, the establishment of Pleistocene icefields could have been responsible for active speciation events, and would therefore explain the high species' richness observed (Cei, 1986).

According to phylogenetic analysis, viviparity appeared on multiple occasions in the evolutionary history of the genus associated with cool environments at high elevations or high latitudes (Schulte et al., 2000). Thus, this mode of reproduction would have been an effective strategy for surviving the strong Quaternary climatic changes. Nevertheless, the evolution of the genus is characterized by ancient and recent allopatric divergence across ecologically and geologically complex landscapes, incipient speciation, secondary contact and discordance between molecular and morphological patterns of variation (Ávila, Morando & Sites, 2006), making it difficult to reconstruct the evolutionary history.

CONCLUSIONS

The present study suggests that ‘lizards’ and snakes differ in their abundance and diversity in the Patagonian region through time. This difference could be caused, in part, by a biased fossil record. ‘Lizards’ are scarce in the Patagonian Mesozoic and completely absent from Palaeogene deposits, although diverse in other South American sites of the same age. They are relatively frequent in the Patagonian Neogene, where the emergence of some extant genera of iguanians and teiids is now well supported. Snakes are, instead, diverse and abundant in the Cretaceous and Palaeogene of Patagonia, and sparser in the Neogene.

The fossil record of Patagonian squamatans includes extant as well as extinct forms. Among extinct forms, the Cretaceous diversity includes the most basal terrestrial snakes, at least three genera of madtsoiids and a probable ‘dolichosaur’. An extinct genus of the present anilioid snakes is also recorded. None of these Cretaceous forms are phylogenetically related to the extant Patagonian herpetofauna. Contrary to Mesozoic snakes, pleurodont iguanians and scleroglossans are now present in Patagonian territories, with indeterminate forms.

The Patagonian Palaeogene palaeoclimate was dominated by the predominance of a subtropical rain forest, developed under moist and temperate climates (Ortiz Jaureguizar & Cladera, 2006), favouring a relatively great diversity of small, medium-sized and large genera of boid and madtsoiid snakes at high latitudes. Neotropical extant boids would have emerged during the early Palaeogene, with an extended distribution to southern latitudes than at present.

Although known since the Late Cretaceous, the Patagonian fossil record for pleurodont iguanians is scarce and fragmentary, a situation that contrasts with the significance of the group in the present Neotropical herpetofauna. Extant pleurodont iguanians appear for the first time in early Miocene deposits, with genera that are successful in the present ecosystems on both sides of the Patagonian Andes. The scleroglossans are also known in Patagonia since the Late Cretaceous, but tupinambine teiids appear for the first time in this territory at the beginning of the Miocene. They reached the southernmost distribution in its evolutionary history during the Miocene Climatic Optimum, but, at present, they persist only with isolated populations in northern Patagonia. Most advanced snakes (‘colubrids’) appear in South America for the first time at the beginning of the Miocene. The early Neogene climatic conditions, with temperatures warmer than today, permitted the distribution of tupinambine teiids, boids and ‘colubrids’ south of the current range of distribution (Fig. 6). Nevertheless, Neogene boids persisted in Patagonia with extinct genera of smaller size than those from the Palaeogene, suggesting colder palaeotemperatures during the Miocene than in earlier periods. Warm and forested habitats were well developed in Patagonia at this time, although some evidence indicates the existence of drier climatic events, or complex environments represented by dune deposits mixed with areas of highland wet forests (Ortiz Jaureguizar & Cladera, 2006). These environments did not restrict the presence of tupinambine teiids, boids and ‘colubrids’, which appear to be more dependent on temperature than on humidity.

In summary, the geographical, and subsequent climatic and environmental, changes that occurred during the Neogene in Patagonia appear to have had a differential influence on the squamatan fauna, producing a strong retraction in the distribution of tupinambine teiids and boid snakes, and, less noticeably, of ‘colubrids’. In contrast, the geographical fragmentation and emergence of drier environments in Patagonia, generated as a consequence of the Miocene changes, would have favoured the differentiation of new adaptive types of pleurodont iguanians.

ACKNOWLEDGEMENTS

This paper is based on my presentation at the Symposium, ‘Palaeogeography and Palaeoclimatology of Patagonia: Effects on Biodiversity’, held at the La Plata Museum in May 2009. I thank Jorge Rabassa, Eduardo Tonni, Alfredo Carlini and Daniel Ruzzante, the symposium organizers, for their invitation. I also thank Daniel Ruzzante and Jorge Rabassa for their comments on previous drafts. This work was partially supported by PIP-CONICET N°112–200901-00176.