Variation in Amapasaurus tetradactylus ( Reptilia : Squamata : Gymnophthalmidae ) , a rare Guiana lizard Variação em Amapasaurus tetradactylus ( Reptilia : Squamata : Gymnophthalmidae ) , um lagarto raro das Guianas

The Guiana region is one of the better-preserved parts of Amazonia, but its fauna is still incompletely known. Amapasaurus tetradactylus was described in 1970, on the basis of two specimens from the upper Maracá river basin, Amapá, Brazil. Only between 2004-2006, during a series of expeditions in Amapá, new material of this species was obtained, and in 2008 it was collected also in northern Pará (at present it is also recorded from French Guiana). A total of 37 specimens, from six localities in Brazilian Guiana, is here examined, giving an idea of variation in scale characters and of its habitat. Present data makes it likely that the species is an endemic to the eastern part of the Guiana Shield, possibly in its central-southern portion.


INtRoduCtIoN
The Guiana part of Amazonia harbors large, well-preserved areas of rainforest.Even though studies of this complex environment began some centuries ago, its fauna is still relatively poorly known.Discovery of new species is not uncommon (e.g., Kok, 2011;Sturaro & Avila-Pires, 2011), and understanding of the hidden molecular diversity within species is only beginning (e.g., Fouquet et al., 2012).Knowledge on the diversity of a region, at all levels, is fundamental for the development of conservation policies.We here provide data on an apparently rare (or at least elusive), small lizard of the Guiana region.Cunha (1970) described Amapasaurus tetradactylus (Figure 1) on the basis of two specimens from the upper Maracá river basin, Amapá, Brazil.The species was described in a new genus, mainly due to the presence of four fingers and a frontonasal divided into several scales.Apart from these characters, lepidosis was similar to that of Leposoma, suggested by Cunha (1970) to be its closest relative.For more than 30 years these were the only two specimens known, and they were reported by Avila-Pires (1995) to be in poor condition, especially the paratype.Between 2004Between -2006, however, during a series of expeditions more specimens were collected in three distinct localities in Amapá (Lima, 2008), to which a fourth locality was recently (September 2012) added.In 2008 specimens were collected for the first time in the state of Pará, in two of seven localities studied north of the Amazon (Avila-Pires et al., 2010).Dewynter & Surugue (2012) reported the species from southern French Guiana, extending its distribution further north (Figure 2).
A total of 37 specimens, from six localities in Brazilian Guiana, is now available, apart from the types, and we here analyze variation in scale characters, as well as what can be added in terms of ecological data about this species.

MAtERIAlS ANd MEthodS
Specimens examined are deposited in Museu Paraense Emílio Goeldi (MPEG) or in the Instituto de Pesquisas Científicas e Tecnológicas do Amapá (IEPA).They represent 31 out of the 37 newly collected specimens, all from Brazil.MPEG 25099 (from Óbidos) was not included in the study of variation because it was not available because of hemipenial studies elsewhere.Five specimens (four males, one female; IEPA 1870-1874) were obtained after the manuscript had been closed and were not used for morphometrics, but the collecting locality was inserted in the map.
Among the material examined 28 are from Amapá Field data of specimens from Amapá were obtained by JDL, JRFL and Rafael Cabral dos Santos; of specimens from Pará by TCSAP, MSH and WAR.Snout-vent length (SVL), tail length (specimens from Pará) and weight were measured in the field, prior to fixation.Scale counts were made under a stereomicroscope and follow Avila-Pires (1995).Mean and standard deviation of paired characters were calculated on the basis of counts on each side separately, but frequency distributions shown in graphs consider counts on both sides.In order to estimate geographic variation we make comparisons between specimens from Amapá, which encompasses the easternmost localities, and Pará (western localities).

GENERAl dESCRIPtIoN ANd oBSERVEd VARIAtIoN
Specimens (14 males, 12 females, five of unidentified gender) in the sample ranged from 23 mm SVL, weighing 0.3 g, to 36 mm SVL, 0.6 g (Figure 3).Three specimens from Pará had intact tails that were 1.6-1.7 times SVL.All specimens have the following dorsal head scales, most of which exhibit small pits close to their margins (Figure 4A): rostral, frontonasal divided into 3-6 scales (Figures 5, 6A), paired prefrontals (left prefrontal divided in IEPA 254; Figure 5G), frontal, paired frontoparietals, and a group of one interparietal bordered by one parietal on each side forming a rounded posterior margin.Anterior supraocular single or consisting of up to four additional smaller scales (Figure 6B), followed by three other supraoculars (Figure 7).Four to six supraciliaries (Figure 6C), partially separated from supraoculars by 1-5 scales (Figure 6D); with one exception (IEPA 319), there is always contact between anterior supraciliaries and supraoculars, while posteriorly they may or not be in contact.
On ventral surface of head (Figure 4B), mental is followed by a single, large postmental and usually three pairs (Figures 8A-8D  8C, 8E), and both in contact with infralabials.Some specimens have both pairs in medial contact, with a small scale in the corner between the two pairs (Figure 8A).Third pair varies from relatively small to large and it is always separated medially and from the infralabials.Posteriorly scales are irregularly polygonal to hexagonal, broadly keeled; they may be completely or partially separated by small scales from the gular region.Usually three infralabials (four on one side in IEPA 253), the contact between second and third below centre of eye.Gulars in 8-11 transverse rows (Figures 4B, 6H), of which the collar is composed by 7-10 scales (Figure 6I).Gular fold distinct.
Amapasaurus tetradactylus is mostly brown dorsally, cream or yellow ventrally.None of the males had an orange venter as found in males of Leposoma.A series of cream blotches, outlined by dark brown, are present dorsolaterally, with smaller spots on neck and the lower part of flanks.Up to blotches may be present, but some may be fused, forming a continuous line; blotches may be rounded or more irregular, and the dark brown outline more or less complete.Spots on sides are distinctly smaller, up to 13 on each side but usually less, with much variation in number, form, and how far back they reach.Descriptions of colour in life, all from specimens from Pará (by MSH) -MPEG 25099 (male, SVL 29 mm): Back brown, series of orange-brown dorsolateral spots, lateral line of white ocelli.Belly and chin yellow.Iris orange-brown.MPEG 25100 (male, SVL 29 mm): underside yellow, centre of throat white.MPEG 25101 (female, SVL 33 mm): back dark brown.Dorsolateral band indicated by light brown spots.Series of white ocelli at    dIStRIButIoN ANd hABItAt Known from Amapá and Pará, Brazil, north of the Amazon River, and from southern French Guiana.Specimens from Amapá were collected in four localities -in the municipalities of Laranjal do Jari (two specimens) and Pedra Branca do Amapari (21 specimens), both within the limits of the 'Reserva de Desenvolvimento Sustentável (RDS) do rio Iratapuru'; in the municipality of Serra do Navio (one specimen), in the 'Parque Nacional Montanhas do Tumucumaque' (Lima, 2008); and in the municipality of Mazagão (five specimens), in the 'Floresta Estadual (FLOTA) do Amapá'.In Pará, Amapasaurus tetradactylus was found in two localities -municipality of Óbidos, within the 'Floresta Estadual (FLOTA) Trombetas' (eight specimens) and a single specimen in the municipality of Oriximiná, in the Acarai (or Acari) Mountains, northern part of 'Estação Ecológica (ESEC) Grão-Pará' (Avila-Pires et al., 2010; Figure 2).All were in the leaf litter of well-drained ('terra firme') rainforest, although the forest where the only specimen (IEPA 507) from Serra do Navio was found showed signs of influence of the Anacuí river (with tallest trees about 20-25 m high, vines, and dense undergrowth), which ran parallel to the trail used, but at least 500 m away from it.Leaf litter depth in most places in Amapá varied between 15-30 cm (but see below for an exception).Most specimens (32 out of 37) fell in pitfall traps.In RDS do Rio Iratapuru, 12 specimens were collected in a single pitfall, in one day (eight in the central bucket of a Y-shaped pitfall), during the dry season (July 31, 2005), in a place with a humid, thick layer of leaf litter (c.50-60 cm) at the base of a hill.This seems to indicate high activity and local abundance.
In the field the species is easily confused with sympatric species of Leposoma.At a first glance the most distinctive character is the dorsolateral line of cream blotches in A. tetradactylus.Adult males differ in the color of the belly-yellow in A. tetradactyus, orange in L. guianense.

dISCuSSIoN
Our data show that the divided frontonasal, which together with the presence of only four fingers on the hand make Amapasaurus tetradactylus so distinct, occurs in all specimens and follows a rather regular arrangement, even though there is some variation in the number of scales that substitute the single or double frontonasal of most gymnophthalmids.These characters make the species easily recognizable, while other features of lepidosis, size and habitat are quite similar to those in species of Leposoma.Cunha (1970) already supposed that Amapasaurus and Leposoma were most likely closely related.The similarity between these two genera is strengthened now that Rodrigues et al. (2013) described a species of Leposoma, L. sinepollex Rodrigues, Teixeira, Recoder, Dal Vechio, Damasceno & Pellegrino, 2013, with only four fingers.However, even considering the possibility of Amapasaurus being derived from Leposoma, morphology would include it in the L. parietale group, of Amazonian distribution, while L. sinepollex is part of the L. scincoides group, present only in the Atlantic Forest.Thus, even in this case two independent origins for the reduction of the pollex would have to be considered, as pointed out by Rodrigues et al. (2013).
Most characters did not show variation when specimens from Amapá and Pará were compared.A few characters, however, showed some tendency to differ between the two groups, e.g.scales around midbody, transverse rows of gulars and lamellae under fourth finger.Number of specimens available and geographic representation are still insufficient for a good understanding of geographic variation in the species, and these results should be seen as no more than an indication.Differences between populations due to geographic distance are expected to occur and it is important to take them into consideration in conservation policies.
Amapasaurus tetradactylus seems to be endemic to the eastern portion of the Guiana Shield, south of the divide (Brazil-Guyana+Suriname border), except in the easternmost part, where it reaches northward at least to the Massif Central Guyanais, in French Guiana.Another possibility is that its northern limit coincides with a faunal break observed by Fouquet et al. (2012) in 12 amphibian lineages (within species), that crosses the region in a SE-NW direction, from Amapá to Suriname, passing through French Guiana.At least three of the species examined by these authors (Adenomera heyeri, Allobates femoralis and Rhinella 'margaritifera') showed independent lineages in the central portion of the Guiana Shield south of this break.The northeastern distribution limit of Atelopus hoogmoedi seems also to follow this break zone (Noonan & Gaucher, 2005).
In all cases mentioned above, the area of occurrence includes parts of Suriname and Guyana, where A. tetradactylus has not yet been found.Reasonably good inventories of lizards were made in many parts of Suriname between 1968 and 1990 (Hoogmoed, 1973(Hoogmoed, , 1985;;Hoogmoed, personal observation), although sampling has been minimal in southwestern and southeastern Suriname.Recently some RAPs have been made in areas in southern Suriname (Ouboter et al., 2011;Nielsen et al., 2012Nielsen et al., , 2013)).In Guyana inventories of lizards have been few and far between.The largest inventory effort seems to have been that in Iwokrama forest (Donnelly et al., 2004), but other inventories were made in southern (Señaris et al., 2008), northwestern (Reynolds & MacCulloch, 2012), and central western Guyana (MacCulloch & Reynolds, 2012).Considering the apparent difficulty in collecting this species, however, and considering that many of these inventories did not use pitfall traps (none in Suriname, nor in southern Guyana), presence of A. tetradactylus, especially in the southern part of these countries, cannot be ruled out.For instance, the species was collected in Brazil in the Acarai Mountains, a few hundred meters from the Brazil-Guyana border, making the presence of the species on the Guyana side of the border very likely.All evidence therefore indicates that A. tetradactylus is endemic to the eastern part of the Guiana Shield, possibly in its southern portion.
) of chinshields (two pairs in IEPA 317 [Figure 8E], second and third partially fused on left side in IEPA 507 [Figure 8F]), of which the anterior pair in medial contact (separated by an azygous medial scale in MPEG 27751 [Figure 8D]), the second pair in medial contact (Figures 8A, 8D, 8F) or separated by small scales (Figures 8B-

Figure 9 .
Figure 9. Variation in scale counts in A. tetradactylus, separately for specimens from Amapá (AP) and Pará (PA).A) Dorsals along a middorsal line between interparietal and posterior margin of hind limbs; B) longitudinal rows of ventrals (counted at midbody); C) transverse rows of ventrals between collar and precloacal plate; D) scales around midbody; E) lamellae under fourth finger (both sides); F) lamellae under fourth toe (both sides).