Foliesaurus boutersemensis, Augé & Smith, 2009

Augé, Marc & Smith, Richard, 2009, An assemblage of early Oligocene lizards (Squamata) from the locality of Boutersem (Belgium), with comments on the Eocene-Oligocene transition, Zoological Journal of the Linnean Society 155 (1), pp. 148-170 : 151-160

FOLIESAURUS BOUTERSEMENSIS SP. NOV.

Holotype: An incomplete right dentary lacking its anterior part, IRSNB R 245 , previously BOU-AR-40- RS, Figure 4 (only known specimen).

Type locality and horizon: Boutersem-TGV locality, Belgium, early Oligocene ( MP 21).

Etymology: From the locality of Boutersem, Belgium.

Diagnosis: Foliesaurus boutersemensis differs from all other scincoid species by the following combination of character states: extremely narrow Meckelian canal and straight ventral margin of the dentary; coronoid process of the dentary well developed and extended above the level of the tooth row; teeth robust, conical, unicuspid, tapering, with high crowns, their size increasing steadily towards the rear of the tooth row.

Description and discussion

The dentary ( Fig. 4) is slender and rather elongated, its anterior part is missing and it measures 3.4 mm (maximum length). The ventral edge is straight. No articular facet for the coronoid is present along the dorso-lateral side of the specimen. The dentary has a prominent coronoid process projecting posterodorsally at high angle above the level of the tooth. Below this process, a small part of the posteroventral process is preserved. The subdental shelf is prominent medially and anteriorly it occupies a position close to the ventral margin of the jaw. This position results in a great restriction of the Meckelian canal, which is exposed only ventrally in its anterior part. The subdental shelf has a weakly developed dorsal ridge that borders the sulcus dentalis. The splenial facet on the ventral margin of the subdental shelf is very limited. Two alveolar foramina are visible on the labial side of the dentary.

Eight tooth positions are preserved, including five teeth. More than one-third of the tooth extends above the level of the dental parapet. The teeth are robust, high-crowned and unicuspid. The tooth crown is pointed and not recurved. The tooth shaft is columnar and the last preserved tooth shows a replacement pit on its lingual side. The size of the tooth increases steadily towards the rear of the tooth row. This dental form is somewhat reminiscent of the teeth seen in amphisbaenid lizards, namely Blanus or Amphisbaena , but the general morphology of the dentary seems to rule out such a relation.

Discussion

The above-described new genus and species is only known from the holotype. The following combination of primitive and derived features shows strong evidence of scincoid affinities: presence of a prominent subdental shelf and of a sulcus dentalis; extremely narrow Meckelian canal and straight ventral margin of the dentary; absence of an articular surface for the coronoid lateral process on the dorso-lateral side of the dentary. Foliesaurus is very distinct from the scincoid taxa known in the European Paleogene. Perhaps its most unusual characteristic is its robust, conical and high-crowned posterior teeth. The tendency toward closure of the Meckelian canal could be a characteristic of Scincidae as many scincid lizards show a closed Meckelian canal, in contrast to cordylids, which all have an open Meckelian canal.

However, an assignment to scincid lizards is premature and more fossils are needed to reach a final decision.

If correctly attributed, this new species marks the first appearance of scincoid lizards in the European Oligocene. Folie, Sigé & Smith (2005) have described a basal scincoid lizard, Scincoideus haininensis , which is close to the origin of cordylids and scincids, from the middle Palaeocene of Hainin, Belgium. A potential scincid lizard has been named from the Palaeocene of Cernay ( MP 6, France, Augé, 2005), but it lacks true scincid apomorphies and it may be considered as basal in Scincoidea . Another basal scincoid lizard, Ornatocephalus metzleri , has been recently described from the Middle Eocene of Messel ( Weber, 2004). Böhme & Lang (1991) demonstrated that ‘ Lacerta ’ rottensis (latest Oligocene, MP 30, Germany) should be referred to Cordyliformes, i.e. to the Cordylidae ( Estes et al., 1988) . The presence of one scincoid species in the European early Oligocene accords well with the hypothesis of a progressive replacement of basal scincoid lizards by more derived scincoids (scincids?, cordylids?) during the Palaeogene in Europe ( Folie et al., 2005).

ANGUIMORPHA FÜRBRINGER, 1900

ANGUIDAE GRAY, 1825

ANGUINAE GRAY, 1825

Dopasia Gray, 1853

Type species: Dopasia gracilis Gray, 1853

Dopasia roqueprunensis ( Augé, 1992)

Ophisaurus roqueprunensis Augé, 1992: 159–175 .

Ophisaurus roqueprunensis Rage & Augé, 1993: 202 . Dopasia roqueprunensis Sullivan & Holman, 1996: 367 .

Ophisauriscus roqueprunensis Sullivan, Keller & Habersetzer, 1999: 102 .

Dopasia roqueprunensis Augé, 2005: 234 .

Referred specimens (Boutersem-TGV locality): Nearly complete left dentary, IRSNB R 246, previously BOU- AR-26-RS ( Fig. 5 ); incomplete left dentary; incomplete left maxilla; seven left dentary fragments; three right dentary fragments; two right maxillary fragments; left maxillary fragment; incomplete parietal, IRSNB R 247, previously BOU-AR-38-RS ( Fig. 6 ); numerous osteoderms; dorsal and caudal vertebrae.

Known distribution: Early Oligocene, Boutersem, Hoogbutsel ( MP 21), Belgium; early Oligocene–late Oligocene, Phosphorites du Quercy ( MP 23– MP 28), Roqueprune ( MP 23), Belgarric ( MP 25), Rigal-Jouet ( MP 25), Pech du Fraysse ( MP 28), France.

Diagnosis (emended): Dopasia roqueprunensis differs from Ophisauriscus quadrupes in having a parietal notch; low tooth count (11–12 tooth positions on the dentary); tooth bases less expanded. D. roqueprunensis differs from all other anguines by the following combination of characters: low tooth count; teeth recurved, sharply pointed and broadly based; shallow coronoid notch on the posterior border of the dentary; prominent intramandibular septum that does not reach the ventral margin of the dentary; osteoderms rectangular to subrectangular, some with a keel.

Description

Dentaries: These early Oligocene specimens ( Fig. 5 ) differ only in size from the holotype of Dopasia roqueprunensis . They are long and slender. The labial surface of the dentary is smooth with five alveolar (mental) foramina. Posteriorly, the coronoid process and the supra-angular process reach nearly the same level. A faint angular process is present. The Meckelian canal opens exclusively ventrally for most of its length. Posteriorly an intramandibular septum divides the Meckelian canal, and its ventral border fuses with the dentary wall. The subdental shelf is reduced and it is notched by the anterior inferior alveolar foramen.

Eleven or 12 tooth positions are preserved on the complete specimen, and four teeth are complete. They are unstriated, keeled, recurved and sharply pointed. The tooth replacement is alternate and interdental. The tooth base is wide and without infoldings. These teeth are similar to those of Recent Anguis fragilis and Dopasia harti and to those described for Ophisauriscus quadrupes ( Sullivan et al., 1999) .

Parietal ( IRSNB R 247, Fig. 6 ): The parietal is fused into a single element. Its anterior part is missing and the supratemporal processes are badly damaged. The preserved lateral and posterior margins are concave. Faint dermal sculpturing covers the anterior part of the parietal table. This osteodermal crust seems to be not indented for epidermal scutes. A small parietal notch cuts the posterior margin of the parietal. The supratemporal processes are located at the posterior portion of the parietal table and they seem to be widely divergent.

Osteoderms: The body osteoderms of Dopasia roqueprunensis are rectangular to subrectangular elements with straight margins. Some have a prominent medial keel. Anteriorly, a gliding surface is present. A marked ornamentation covers the rest of the dorsal surface of the osteoderm. This ornamentation consists of relatively deep pits and grooves.

Sullivan et al. (1999) transferred Dopasia roqueprunensis to the synonymy of the middle Eocene anguid genus Ophisauriscus (type and only species Ophisauriscus quadrupes Kuhn, 1940 ), known from the localities of Messel and Geiseltal ( Germany). However, the new specimens from Boutersem allow critical comparison between Dopasia roqueprunensis and Ophisauriscus quadrupes . According to Meszoely & Haubold (1975) and Sullivan et al. (1999), the posterior border of the parietal table of Ophisauriscus lacks the medial parietal notch that is present in a number of Recent Anguinae and other anguids (see also Klembara, 1981). Although D. roqueprunensis and O. quadrupes share the same general tooth morphology, obvious differences occur. Teeth of Ophisauriscus are also recurved and sharply pointed but their bases seem less expanded than those of D. roqueprunensis . According to Meszoely & Haubold (1975) tooth count in Ophisauriscus is low (11 dentary teeth) but, apparently, they did not include the unoccupied tooth positions in their count. Furthermore, Sullivan et al. (1999: 104) report a tooth count of 14–15 fang-like teeth on the maxilla of the specimen GM CeIII-4038. Clearly, O. quadrupes has more teeth (at least 15 tooth positions on both the dentary and the maxilla) than D. roqueprunensis (11 or 12 tooth positions on the dentary).

According to Sullivan et al. (1999), the genus Ophisaurus is restricted to North America. Old World anguin lizards previously referred to this genus belong to either the Pseudopus ( Klembara, 1981) or the Ophisauriscus – Anguis – Dopasia lineages. The genus Pseudopus possesses distinctly inflated, posterior teeth. The genus Dopasia is distinguished from Ophisauriscus by the characters already cited in the diagnosis. The genus Anguis differs from Dopasia by its low tooth count (8–10 in Anguis ; 11–14 in Dopasia ).

Hence, D. roqueprunensis is certainly a valid species, sufficiently distinct from O. quadrupes to be placed within a distinct genus. According to Sullivan & Holman (1996), the Oligocene specimens are very similar to the Recent Dopasia harti ; accordingly, ‘ Ophisaurus ’ roqueprunensis has been transfered to Dopasia .

PLATYNOTA BAUR, 1890

VARANOIDEA BOULENGER, 1891

Necrosaurus Filhol, 1876

Type species: Palaeovaranus cayluxi Filhol, 1873 Diagnosis: Necrosaurus differs from all other varanoid species by the following combination of character states: presence of oval, keeled osteoderms; no ridges on the ventral surface of the parietal; premaxilla with a long, arched nasal process (contra Estes, 1983); frontal azygous, not narrowed between the orbits; parietal azygous, adductor musculature extending onto dorsal surface of parietal, sometimes with a sagittal crest; teeth trenchant, blade-like, recurved; tooth bases dilated, striated, with plicidentine; tooth replacement alternate and interdental.

Necrosaurus sp.

Referred specimens: Nearly complete dentary; incomplete left dentary; incomplete left maxilla (BOU-AR- 36-RS); fragment of maxilla ( IRSNB R 248, previously BOU-AR-33-RS, Fig. 7 ); fragment of right dentary; complete frontal ( IRSNB R 249, previously BOU-AR- 42-RS, Fig. 8 ); numerous osteoderms ( IRSNB R 250, previously BOU-AR-28-RS, Fig. 9 ); dorsal and caudal vertebrae.

As used here, Platynota is a group containing the crown-group Varanoidea. We follow Rieppel (1980), Pregill, Gauthier & Greene (1986) and Rieppel, Conrad & Maisano (2007) in using Varanoidea as an inclusive group, including Varanidae , Helodermatidae as well as Monstersauria and perhaps marine taxa ( Mosasauridae , Dolichosauridae , Aigialosauridae ).

Description

Maxilla (BOU-AR-36-RS): The specimen is an incomplete left maxilla bearing three teeth. The lateral Frontal ( IRSNB R 249, Fig. 8 ): The nearly complete frontal is azygous, constriction between the orbits is absent and the lateral margins of the frontal are parallel, although the posterior part of the frontal expands markedly for the parietal contact. The entire dorsal surface is covered with fused, oval osteoderms. The frontal has a triradiate anterior margin. Small facets are present at each posterior corner of the frontal. These facets certainly receive matching tabs of the parietal. This character suggests a differenciation of the mesokinetic line of flexion in Necrosaurus , as in Varanus . Medial to these facets, there is very slight digitation of the posterior part of the frontal. On the ventral surface of the frontal, the descending processes (cristae cranii) are well developed. They approach each other, but do not meet below the olfactory tracts.

surface of the bone is rather rugose but devoid of osteoderms, and a row of foramina parallels the tooth row near the inferior margin of the bone. The anterior margin of the ascending (nasal) process of the maxilla is relatively steep and deflected medially but its dorsal part is broken.

The teeth are broad-based and widely spaced; the first maxillary tooth is smaller than the succeeding ones. Scars on the dental shelf indicate that replacement of marginal teeth was alternate. Enamel of tooth bases is infolded, and striations are widely spaced. The longitudinal striation of the tooth surface near its expanding base is interpreted by Caldwell (2003) as an indication of the presence of plicidentine. The broken tooth on specimen IRSNB R 248 ( Fig. 7 ) shows basal infolding of the dentine and confirms the presence of plicidentine. Tooth crowns are long, gracile, recurved and trenchant. The anterior and posterior tooth margins are offset from the main shafts (they form keels).

Osteoderms ( IRSNB R 250, Fig. 9 ): The osteoderms are oval elements with more or less irregular margins. They bear a prominent medial keel that extands the full length of the osteoderm. From the keel radiates a pattern of deep pits or grooves and marked ridges. Gliding surfaces are absent.

Several features of the marginal dentition indicate relationship to Varanoidea: Teeth broad-based, trenchant, with plicidentine and tooth replacement without resorption pits.

There is a close morphological similarity between the teeth of Necrosaurus and Varanus . The presence of small facets at each posterior corner of the frontal is also a derived character shared by Varanus and Necrosaurus .

These features are common among necrosaurs, a group of Platynota that lack several characters shared by recent varanoids (external nares not retracted, descending processes of frontal not meeting below the olfactory tract). Owing to the lack of synapomorphies, the Necrosauridae are generally regarded as a paraphyletic group known from the Late Cretaceous through Early Oligocene ( Hoffstetter, 1943, 1954; Estes, 1983; Rage, 1988; Augé, 2005). Many poorly known varanoid lizards have been included in the family, but Necrosauridae is a problematic, certainly not monophyletic assembly of lizards and references to Necrosauridae are presently inappropriate.

However, several characters serve to diagnose the genus Necrosaurus and to separate it from other platynotan lizards. Among them, the principal derived character states of Necrosaurus include: presence of oval, keeled osteoderms both on the body and on the skull roof; tooth bases with infoldings, indicating the presence of plicidentine, but striations widely spaced; lack of marked processes (reliefs) on the ventral side of the parietal.

The oldest putative record of Necrosaurus is from the European Palaeocene (locality of Cernay, France, MP 6), and the genus is known throughout the Eocene and the early Oligocene of Europe ( Rage & Augé, 1993; Augé, 2005). Estes (1983) cited two species of Necrosaurus in the European Eocene ( Necrosaurus eucarinatus , N. cayluxi ), along with other specimens from various strata of the European Palaeogene. The early Oligocene specimens from Boutersem appear to differ from these species, and it seems that the Boutersem fossils belong to a new species. Hecht & Hoffstetter (1962) already noted some features that suggest that the early Oligocene specimens may represent a new species of Necrosaurus . However, a renewed evaluation of the material assigned to Necrosaurus is necessary before diagnosing a new species.

LACERTILIAN FAUNAS ACROSS THE EOCENE/ OLIGOCENE BOUNDARY IN WESTERN EUROPE

It has been known for almost a century that a major turnover occurred in the European mammalian faunas near the Eocene/Oligocene boundary (known as the Grande Coupure, see Stehlin, 1909). The Quercy localities represent a long biochronological succession from this period, and contain, next to mammals, an amazing assemblage of squamate reptiles. Their study has revealed an important change among Lacertilia across the Eocene/Oligocene boundary ( Rage, 1984, 1986; Augé, 1993; Milner, Milner & Evans, 2000). However, the early levels of the Oligocene are poorly represented in the localities from the Phosphorites du Quercy. Lizards from Boutersem confirm that the Eocene–Oligocene (E-O) transition represents, together with the Palaeocene–Eocene transition, the most critical turnovers in the Cenozoic history of lizards.

Diversity

During the late Eocene, the lizard faunas were abundant and diverse. Nine families and 17 species are present in the standard level MP 19 (mammalian standard level of Escamps), before the E-O transition ( Augé, 2005). Unfortunately, no lizard remains are known from the last Eocene standard level (St Capraise, MP 20), which is very poorly represented among all taxa.

At the family and the species levels, lizards were severely affected by the Grande Coupure. A drop in diversity is protracted between the MP 20 and MP 21 levels (early Oligocene) and a low diversity appears in the MP 21 level (five families, eight species). Four families encountered in the European upper Eocene became extinct or temporarily absent between MP 19 and MP 21 ( Iguanidae *, Gekkonidae , Glyptosaurinae and Helodermatidae ). However, gekkonid and helodermatid lizards are found in the following MP 22 standard level. At lower taxonomic levels, estimates of species-level extinctions range as high as 80%. According to Rage (2006), they include members of virtually all the families present in the upper Eocene ( Table 2a; the number of specimens known from each level is plotted on Table 2b). Rage & Augé (1993) have estimated that between 66 and 80% of the lizard and snake species may have been lost in Europe over this period. During that time, species-level originations are not especially high (29%). Thus, the E-O event provides evidence of a high rate of extinction that is not matched by originations.

By the Oligocene, the lizard world had changed radically and a striking decline in families and species numbers occurred. Lizard diversity remained relatively constant and low by Eocene standards.

Of the 12 species that occur after the Grande Coupure (standard levels MP 21 + MP 22), seven have no close relatives in the preceding Eocene fauna and five survived the event unchanged or slightly modified. Hence, 58% of the early Oligocene fauna could have been the result of dispersal from elsewhere in Europe or another continent. Survivors once diminished remained so until extinction a few million years later (e.g. Necrosaurus , Helodermatidae , Plesiolacerta and Helvetisaurus ).

Interpreting faunal change

What caused this change in the lizard fauna during the E-O event? The two main physical parameters that appear to have controlled this phenomenon are climate change and the presence or absence of land bridges between land masses. The absence of land connections may prevent or inhibit geographical range extensions and cause widespread endemism.

The E-O period corresponds to a well-established global event, the so-called Terminal Eocene Event, which certainly post-dates the E-O boundary by 0.2 Myr ( Pomerol & Premoli-Silva, 1986; Prothero & Berggren, 1992; Prothero, 1994; Culver & Rawson, 2000; Prothero, Ivany & Nesbitt, 2003; Zanazzi et al., 2007). During this interval, marine deep waters experienced a progressive cooling of as much as 10 °C ( Miller, 1992). This resulted in severe climatic and atmospheric disruptions, including a steepened latitudinal temperature gradient and increasing seasonality ( Ivany, Lohmann & Patterson, 2003; Mosbrugger, Utescher & Dilcher, 2005). However, the first major step in the climate change of the Cretaceous through middle Eocene ‘Greenhouse world’ was the profound cooling event at the end of the middle Eocene. By any standard, the end of the middle Eocene Oligocene

Iguanidae *

Cadurciguana hoffstetteri Geiseltaliellus lamandini Geiseltaliellus grisolli Geiseltaliellus sp. Pseudolacerta mucronata Pseudolacerta quercyini Agamidae * Uromastyx europaeus Quercygama galliae Gekkonidae

Cadurcogekko piveteaui Cadurcogekko rugosus Gekkonidae sp. Lacertidae

Plesiolacerta lydekkeri Dormaalisaurus girardoti Quercycerta maxima Gracilicerta sindexi Escampcerta amblyodonta Lacerta s.l. filholi Mediolacerta roceki Pseudeumeces cadurcensis Teiidae ? Brevisaurus smithi Cordylidae ? Eocordyla mathisi Cordylidae sp.

Scincoidea Ayalasaurus tenuis Scincidae ? Orthoscincus malperiensis Scincoidea undetermined Foliesaurus

Anguidae

Anguinae

Helvetisaurus

Dopasia roqueprunensis Dopasia coderetensis Anguis sp. Glyptosaurinae

Placosaurus Paraplacosauriops Paraxestops Helodermatidae

Eurheloderma gallicum Helodermatidae sp. Necrosauridae

Necrosaurus eucarinatus Necrosaurus cayluxi Necrosaurus sp. Amphisbaenidae

Amphisbaenidae sp.

?

??

?

Date marks the Grande Coupure (-33.5 Ma). Eocene was a dramatic cooling event (Berggren & Prothero, 1992; Berggren et al., 1995). The first icesheets in Antarctica are now believed to have formed in the mid-Eocene ( Hurley & Fluegeman, 2003), and a large glaciation occurred 42 Mya ( Tripati et al., 2005). Furthermore, the glaciation is not limited to the Antarctic; Moran et al. (2006) show that arctic ice developed much earlier than previously believed (45 Mya). The beginning of the Oligocene records the first major expansion of glacial ice on Antarctica ( Zachos et al., 1992, 2001). Terrestrial palaeoclimate proxies also recorded this climate step. Palaeoprecipi- tation and palaeotemperature estimates based on palaeosols in North America suggest that changes associated with the E-O transition are a small part of a long decline rather than a sudden climate shift ( Sheldon et al., 2006). New studies of palaeosols on the Isle of Wight ( UK) yield a similar result and suggest relatively steady annual climatic conditions across the E-O transition.

The E-O event also involved a major eustatic sealevel fall ( Gély & Lorenz, 1991; Steurbaut, 1992; Hooker et al., 2004). Faunas had evolved independently in island Europe during the Eocene after the split from North America. The sea-level fall allowed the dispersal of new mammalian faunas, certainly from Asia. The flood of Asian taxa was thought to have crossed the Turgai Straits between Europe and Asia ( Vianey-Liaud, 1976). Unfortunately, our knowledge of the lizard fauna during the E-O event in Eastern Europe and Asia is very limited.

Climate change is a key forcing mechanism for global biotic change. Due to their physiological requirements, vertebrate ectotherms are often considered good climatic indicators ( Markwick, 1998; Böhme, 2003). A temperature drop would increase the environmental pressures on terrestrial vertebrates, presumably having more adverse effects on terrestrial ectotherms than endotherms.

However, this simple picture is affected by several shortcomings ( Augé, 2003):

1. The timing of the turnover in Europe suggests that it was less directly connected with climate than with land bridge connections. Three major episodes of climatic deteriorations occurred during this period, one at the middle–late Eocene transition ( Châteauneuf, 1980, 1986; Collinson & Hooker, 1987; Ollivier-Pierre et al., 1987; Schuler, 1990; Collinson, 1992), one in the earliest Oligocene ( Wolfe, 1992) and the last at the end of the early Oligocene ( Miller et al., 1991). Two of these climatic shifts do not seem to have resulted in fundamental changes for lizard faunas. For example, species diversity reaches a peak in MP17 (10–11? families; 22 species), just after the first climatic deterioration. Furthermore, according to Sheldon et al. (2006), changes associated with the E-O transition are not a sudden climate shift but rather a steady decline.

2. A growing body of data conflicts with the notion that all organisms are highly sensitive to climatic changes ( Prothero, 1999). Some past climatic deteriorations did not affect lizard faunas dramatically. During the North American Pleistocene, the herpetofauna ( Squamata + Amphibia) was quite stable, compared with the avian and mammalian faunas ( Kurten & Anderson, 1980). No extinct families or genera of Pleistocene reptiles are currently recognized in the North American continental Pleistocene. Lizards had only one unquestioned extinct species among 46 identified taxa. However, beyond general taxonomic stability, wide geographical changes and range adjustments occur in lacertilian populations ( Holman, 1995).

3. In the European Pleistocene, only three species of oceanic island lizards become unquestionably extinct, and no mainland form is considered to be an extinct species ( Holman, 1998). There is still no doubt that the herpetofaunas of Europe and North America have been strikingly more stable since the beginning of the Pleistocene than the endothermic faunas.

4. Changes in the Pleistocene mammal fauna of Mediterranean Europe were severe. The lower Pleistocene mammal fauna was still largely tropical. Only a few subtropical taxa survived after the large extinctions of the first glaciations (Mindel glacial, about 1 Mya) ( Kurten, 1968; Cheylan, 1991; Blondel & Vigne, 1993). In contrast, Blondel (1995) argues that these events had less adverse effects on the herpetofauna than on mammals. In the same vein, Bailon (1991) reports several local extinctions among the herpetofauna during the Pliocene–Pleistocene of France and Spain. However, only one species is unquestionably extinct.

Although climate change is certainly involved in squamate turnover, insights into the causes of extinction during the E-O event are also gained from an examination of the European palaeogeography and of the biology of the lizards. For the most part of the Eocene, Europe was an archipelago of large and small islands.

The basic pattern of biodiversity on islands relates the number of species on an island to both the area of the island and its distance from a source area ( McArthur & Wilson, 1967; Diamond, 1973). McArthur and Wilson suggested that the extinction rate on an island increases as the island becomes smaller and decreases with its distance from the source area. They reasoned that extinction rates on islands would be higher because of smaller average population sizes (small populations are more extinction-prone). It is clear that extinction affected some groups more than others ( James, 1995; Roughgarden, 1995). The comparative energetics of the various classes of vertebrates on an island have an influence on their extinction rate. Lizards are much less extinction-prone than birds and mammals on the same island. The reason is that lizards have much lower metabolic rates and tend to have higher population densities ( Wilcox, 1980). Lizards can be as much as 100 times more abundant than birds on the same food supply, and lizards have smaller spatial requirements than homeotherms of similar mass ( Schoener & Schoener, 1978). Obviously, small populations lead to extinction, and conservation biologists adopted the idea and wrote about ‘minimum population size’ ( Preston, 1962; Soulé, 1987; Remmert, 1994; Caughley & Gunn, 1996; Hubbell, 2001). Good evidence for the importance to extinction of population size comes from both palaeontology and biogeography ( Diamond, 1984; Jablonski, 1986; Pimm, Jones & Diamond, 1988).

References: Nowak, 1991; Burness et al., 2001; Molnar, 2004.

Moreover, an insular vertebrate fauna typically has fewer large mammal predators than its continental counterpart ( Marquet & Taper, 1998). Lizards become top predators in the absence of the mammals that usually prey on them. Island populations are strongly driven by release from ecological constraints (e.g. competition, predation). Birds are also key factors that affect ecological release. The importance of birds can be broadly explained by greater insular extinction rates for birds than for lizards ( Case, 1975; Wright, 1981). Differential extinction rates between taxa can result in lizard communities receiving more energy on islands than on mainlands. Buckley & Jetz (2007) demonstrate that, at least for lizards, higher densities on islands are a ubiquitous and global phenomenon. For example, populations of lizards in the genus Anolis are nearly an order of magnitude more dense on Caribbean islands than in the adjacent Central American mainland despite reasonably similar environmental conditions ( Buckley & Jetz, 2007). This shift to ectothermy ( McNab, 1994a, 2002a) accounts for the replacement of mammalian carnivores on islands by large reptiles, including, for example, the large varanid on Komodo. Indeed, varanids are the principal predators on islands from Wallace’s line to New Guinea and the Solomons. New Caledonia has yielded a large fossil Varanus ( Flannery, 1991) and Flannery (1991, 1993) has suggested that the dominance of giant varanid, snakes and land crocodiles as predators on Australia reflected the isolation and the low productivity of this continent. The presence of terrestrial crocodiles has been reported on Fiji ( Worthy, 2001) and in Australia ( Quinkana, Flannery, 1993 ). Australia, uniquely among the continents, evolved a varanid lizard, the now-extinct 200–400-kg Megalania prisca , as one of its top carnivores. Australia is the smallest continent and it would have had difficulty supporting a large endothermic carnivore ( Burness, Diamond & Flannery, 2001; Maurer, 2002; Molnar, 2004, see Table 3). Furthermore, although one of the native Australian top carnivores (the marsupial lion) was a mammal, it is not a placental but a marsupial. It is noteworthy that the metabolic rates of marupials are much lower than those of placentals ( Dawson & Hulbert, 1970; McNab, 1978). Wroe (2002) brought into question the contention that reptiles dominated Australia’s large terrestrial carnivore niches. However, in a subsequent paper, Wroe, Argot & Dickman (2004) acknowledge that, in Australia, terrestrial (flightless) birds and reptiles may have occupied a number of niches taken by mammals elsewhere. Wroe et al. (2004) attribute the paucity of mammalian carnivores in Australia to isolation and limited landmass area. In the same vein, Hickson, Slack & Lockhart (2000) argue that the large number of lizards species in New Zealand can be explained by the absence of mammals, which would have been both competitors and predators.

The evolution of a flightless condition in birds is linked with past and present isolated areas. New Zealand is a classic example: some 30–35 species and subspecies (or 25–35% of terrestrial and freshwater birds) were flightless before the arrival of Polynesians ( Atkinson & Millener, 1991; Bell, 1991). Energy

Figures

Figure 5. Dopasia roqueprunensis, nearly complete left dentary, IRSNB R 246, previously BOU-AR-26-RS. 5a, lateral view; 5b, medial view.

Figure 6. Dopasia roqueprunensis, incomplete parietal, IRSNB R 247, previously BOU-AR-38-RS. 6a, dorsal view; 6b, ventral view.

Figure 7. Necrosaurus sp., fragment of maxilla, IRSNB R 248, previously BOU-AR-33-RS, medial view.

Figure 8. Necrosaurus sp., complete frontal, IRSNB R 249, previously BOU-AR-42-RS. 8a, dorsal view; 8b, ventral view.

Figure 9. Necrosaurus sp., osteoderm, IRSNB R 250, previously BOU-AR-28-RS, dorsal view.