Tropidurus torquatus ( Wied-Neuwied, 1820 )

Kunz, Tobias Saraiva & Borges-Martins, Márcio, 2013, A new microendemic species of Tropidurus (Squamata: Tropiduridae) from southern Brazil and revalidation of Tropidurus catalanensis Gudynas & Skuk, 1983, Zootaxa 3681 (4), pp. 413-439 : 432-436

Tropidurus torquatus ( Wied-Neuwied, 1820)

( Figs. 18–20 )

Stellio torquatus Wied-Neuwied, 1820: 106 . Type locality: not given. Restricted to the coast of the State of Rio de Janeiro according to Rodrigues (1987), based on Wied’s itinerary. Neotype, MZUSP 54907, from Lagoa do Paulista, Restinga de Jurubatiba National Park, designated by Rodrigues (1987).

Agama operculata Lichtenstein, 1822: 243 . Type locality: not given (fide Rodrigues 1987).

Agama brasiliensis Raddi, 1822: 59 . Type locality: surroundings of the city of Rio de Janeiro.

Agama tuberculata Spix, 1825: 12 . Type locality: States of Rio de Janeiro and Bahia. Lectotype, ZSMH 531/0, designated by Hoogmoed & Gruber (1983).

Trop [idurus] Tuberculata — Gray, 1831: 41 (partim).

Tropidurus torquatus —Wied-Neuwiedi, 1825: 139

Ecphymotes torquatus — Duméril & Bibron, 1837:344 (partim)

Tropidurus microlepidotus Fitzinger, 1843: 72 . Type locality: not given. Restricted to the State of Rio de Janeiro, according to the itinerary of the frigate Novara (fide Rodrigues, 1987).

Taraguira darwinii Gray, 1845: 220 . Type locality: Abrolhos Archipelago.

Tropidurus torquatus torquatus — Cei, 1993: 316.

Neotype. MZUSP 54907; sex not informed; from Lagoa do Paulista, Parque Nacional da Restinga de Jurubatiba, north coast of the State of Rio de Janeiro (22°14’S e 41°33’W); collected 15 December 1979 by P. E. Vanzolini; designated by Rodrigues (1987).

Diagnosis. Tropidurus torquatus is diagnosable from all congeners, except those of the torquatus group, by lacking a middorsal scale row and by having imbricate and keeled dorsal scales. From those of the torquatus group, except for T. catalanensis and T. imbituba sp. nov., it can be distinguished by having two mite pockets under the oblique neck fold, the upper pocket is well developed and the lower is poorly developed; two or three axillary pockets; and inguinal granular pocket present (type “A” sensu Rodrigues 1987). As presently recognized T. torquatus can be distinguished from T. catalanensis and T. imbituba only on the base of coloration characters which are best developed on adult males. From T. catalanensis it differs in not having yellow coloration on chest and mental region, while from T. imbituba it differs in not having orange-bronze coloration on the venter and sides of the trunk.

Variation. Head length 20–25% of SVL; tail 57–64% of total length. Frontal scales of the head generally flat, eventually concave and, in this case, most frequently in males from Cerrado and the mountain ranges of Rio de Janeiro. Number of scales between the nasals 2–4; rostral contacting nasals in 25.4% of the specimens from Cerrado and in 60.6% and 70.6%, respectively, of specimens from coast and from the mountain ranges of Rio de Janeiro; 9–15 scales between rostral and interparietal; 2–4 scale rows of enlarged supraoculars, generally three; 6– 9 enlarged supraoculars in the first row of specimens from the coast and from the mountain ranges of Rio de Janeiro and 6–11 in specimens from Cerrado (examples of the supraocular scuttelation patterns in some populations are in Rodrigues 1987, p. 194, fig. 28C-H; Cei 1993, p. 319, fig. 49B; and Cei 2003, p. 135, fig. 2G); 4–7 supralabials to below the center of eye, generally 5, followed by 3–7 small elongated scales to the commissure of mouth; supralabials contacting subocular in 51.6% of specimens from the coast; in 82.4% of specimens from the mountain ranges of Rio de Janeiro; and in 41% of specimens from Cerrado; in the three cases, contact is generally with the 5th supralabial, or less frequently with the 4th, 6th or 7th; in some cases subocular is in contact with both 5th and 6th or 6th and 7th supralabials; 7–11 infralabials, generally 8–9; preauricular fringe with 5–8 spines (4–8 according to Rodrigues 1987), generally 6–7; 8–14 scales between the subocular and the preauricular fringe; mental generally triangular or polygonal in the coast and in the mountain ranges of Rio de Janeiro; mental generally presents a bell-like shape (laterally curved) in specimens of the Cerrado group of populations; 3–5 postmentals (3–4 in Rodrigues 1987); first postmentals contacting one another in 6.9% of specimens; 2–4 canthals, generally 3; 42–58 gular scales (51.1 ± 3.3, n=70) in specimens from Cerrado, 49–63 (57.6 ± 4.6, n=16) in those from the mountain ranges of Rio de Janeiro and 41–61 (49.4 ± 3.9, n=32) in the coast.

Two mite pockets under the oblique neck fold, the upper pocket is well developed and the lower is poorly developed; antehumeral fold present; two or three axillary pockets; inguinal granular pocket present; 26–35 subdigital lamellae on the fourth toe; 17–33 scales on the dorsum of tibia (20–33 in the coastal populations and 17– 24 in those from Cerrado and the mountain ranges of Rio de Janeiro). Dorsals imbricate, mucronate and keeled, 95–140 (89–144 according to Rodrigues 1987); 95–125 in coastal populations (102–121 in males, 102–115 in females; we couldn’t determine the sex of the specimen MHNCI 3648, a juvenile from Ilhéus, Bahia, which presents 95 dorsals, the lowest value considering all analyzed groups); 107–140 in the mountain ranges of Rio de Janeiro (107–128 in males, 124–140 in females); 100–136 in Cerrado (100–120 in males, 102–136 in females). 86–130 scale rows around midbody (76–143 according to Rodrigues 1987); 92–117 in coastal populations (92–117 in males, 102–117 in females); 100–130 in the mountain ranges of Rio de Janeiro (100–119 in males, 123–130 in females); 86–114 in specimens from Cerrado (86–105 in males, 91–114 in females). 70–97 smooth ventrals.

Color in life. The dorsal pattern in coastal populations is light gray with a very variable pattern of white and dark spots ( Figure 18 ). In the populations from the mountain ranges of Rio de Janeiro and Cerrado, ground color is browner or greenish. Those from the mountain ranges of Rio de Janeiro are very dimorphic in coloration, adult males showing numerous black and cream spots ( Figure 19 ), while females are almost uniform gray or brown, with few spots. A black antehumeral band occupying 5–15 scale rows in coastal populations, separated or not from the opposite one by up to 25 scales (mean 14,2 ± 5,9); this band occupies 5–13 scale rows in specimens from the mountain ranges of Rio de Janeiro and is separated or not from its opposite one by up to 18 scales (11,5 ± 6,3); and 3–10 scale rows, separated or not from its opposite one by up to 18 scales (10,2 ± 5,4) in specimens from Cerrado. Venter white or cream. Adult males whit femoral and preanal patches of black pigmented glandular scales. Presence of midventral patches variable in specimens of all populations analyzed but usually present in the largest males ( Rodrigues 1987 stated that this condition was variable only in the coastal populations and always present in the “inland” populations). Adults with black gular patches, more developed in males. Chin white or cream, with dark reticulations enclosing white or cream blotches more evident in juveniles.

Distribution. From north and northwestern state of São Paulo, northward to the limits of the Cerrado formations in Tocantins. West to Mato Grosso and east it reaches the coast from the city of Rio de Janeiro to Bahia (about 50 km south of Salvador, according to Rodrigues 1987). Recently the species was reported for the coast of São Paulo but it was considered a recent occupation as a result of the human expansion in areas along the coast ( De Sena et al. 2008).

Remarks. At least three forms are included under Tropidurus torquatus . From east to west, a group of lizards inhabits mainly the coastal restingas from Bahia to Rio de Janeiro. This form corresponds to Tropidurus torquatus strictu sensu, since the species was described based on lizards from these populations. There is another form in the mountainous region in the border of Rio de Janeiro and Minas Gerais. This region was, originally, predominantly covered with Atlantic Rainforest and it is supposed that these populations were originally restricted to open enclaves in the midst of forests, such as large rock outcrops. Westward, in the Brazilian central plateau, there is a third set of populations inhabiting mainly the Cerrado domains.

Coastal populations from eastern Brazil seem to have a continuous distribution along the coastal restingas, where these lizards are very abundant. However, these populations are not morphologically homogenous as evidenced by tibia counts, which shows clinal variation, diminishing from south to north. Rodrigues (1987) provided a plausible explanation for the great variation found in the coastal populations, based on sea level fluctuations.

It is important to note that the coastal distribution of torquatus includes several islands and, there, most of the available habitats are rocks. One of these populations, from Abrolhos Archipelago, was described by Gray (1845) as Taraguira darwinii . A difference between the saxicolous populations, including the insular ones, and the coastal sand-dwelling populations is the shape of the claws, thick and strongly curved in the saxicolous forms and long and thin in the psammophilous forms. However, some specimens collected in the coast in habitats other than the restingas also presented thicker claws, although all other characteristics are still identical to those from the restingas. Therefore, it is a much variable characteristic with little value in understanding differentiation patterns, as already noted by Rodrigues (1987). Rodrigues found lizards with long and thin claws only in the coast. However, we found lizards with this character in our group from Central Brazil (Cerrado group). These populations are probably the most generalist form in relation to habitat use and seem to have a continuous distribution in this broad area, which can explain the relative homogeneity of this group. In many areas of Cerrado, T. torquatus uses tree trunks, termite mounds, and even urban areas as much as, or more frequently than rocks ( Rodrigues 1987; Nogueira 2006; personal observations). According to Nogueira et al. (2005) and Nogueira (2006), the Cerrado populations of T. torquatus are associated with the border of galery forests, where they use mainly tree trunks, rarely occurring in the large and continuous areas of open Cerrados (savannas).

Taxonomic and biogeographic comments. In the first taxonomic revision of the Tropidurus torquatus group, Rodrigues (1987) noted morphological differences between the coastal populations of torquatus and the remaining “inland” populations. However, as he could not find a geographic structure for all the variation found, he preferred to keep them as a single species. This author based his analysis on the sympatric occurrence of lizards with distinct fold and pocket patterns, delimiting very large and heterogeneous areas of simpatry, as he pointed out. Besides lacking a representative group of the southernmost populations attributable to torquatus , the joint analysis of morphologically distinct populations, grouped within large sympatry areas, may have prevented the recognition of geographic patterns.

Geographic distribution of most Tropidurus species, particularly those in the torquatus group, is strongly associated to the great South American diagonal of open formations (from northeast to southwest, respectively, the Caatinga, Cerrado and Chaco) ( Vanzolini 1972, 1974; Rodrigues 1987; Zanella 2010). Some evidences suggest recent speciation processes in the genus. Harvey and Gutberlet (1998) described three new Tropidurus species from the Serranía de Huanchaca, Bolivia, an isolated mountain range overlapping the northeastern border of Bolivia and Brazil, in a transitional area between the Amazon and the Cerrado morphoclimatic domains. These authors suggested that their new species probably shared recent ancestors with Tropidurus currently distributed in open formations at low elevations along the Andes of Argentina and southern Bolivia, forming pairs of possible sister species ( T. etheridgei – T. chromatops , T. melanopleurus Boulenger, 1902 – T. callathelys Harvey & Gutberlet, 1998 and T. spinulosus [cope, 1862]– T. xanthochilus Harvey & Gutberlet, 1998 ), and that a single historical event may account for the evolution of each species pair from common ancestors previously distributed across open formations of south-central South America. Relationship between the last two pairs, involving species of the spinulosus group, was later corroborated on the basis of molecular and morphological evidence ( Harvey & Gutberlet 2000; Frost et al. 2001).

Rodrigues (1987) attributed the great variability of the coastal populations of torquatus to the events of marine transgressions and regressions occurred after the last glaciation. It is possible that these events have also been of great importance to the evolution of T. imbituba , since in the recent past the hill where these lizards are isolated was an island in the periods of marine transgressions, when the ocean was above the current level ( Horn Filho 2003; Leal 2005; Vieira et al. 2009). As Rodrigues pointed out, the physiognomy of the South American landscapes was very different at the end of the last glaciation, when the climate was cooler and arid, and the emerged area of the continent was much larger as a consequence of marine regression. One could hypothesize that a common ancestor of the coastal T. torquatus and T. imbituba had, in this period, a broad distribution along the Brazilian coast. However, the fact that imbituba was not able to colonize the adjacent restingas suggests that his ancestor was already a saxicolous form that possibly inhabited the granitic slopes of the Serra do Mar and Serra Geral, currently covered by the Atlantic Rain Forest. With the gradual shift to a more humid climate and the expansion of forests, these lizards had their habitat drastically reduced. Thus, T. imbituba is probably more related to T. catalanensis or the “inland” populations of T. torquatus than to the psammophilous form of the eastern Brazilian coast ( T. torquatus stricto sensu).

It is interesting to note that the restingas and sand dunes of the coastal plains of southern Santa Catarina share with the plains of Rio Grande do Sul and Uruguay typical faunistic elements of open areas of the Pampa Biome, while the hills and rocky seashores that break the continuity of these coastal plains present a typical fauna of the Atlantic Rainforest ( Ghizoni et al. 2009; Kunz & Ghizoni 2011; Kunz et al. 2011). It seems that T. imbituba represents a relictual element of a period of semi-arid climate, when a typical open area fauna was able to occupy those hills, currently covered by forests. Other reptile species from open areas that inhabit these rocky seashores, like the skink Aspronema dorsivittatum ( Cope, 1862) (Mabuyidae) , are also able to occupy the sandy restingas.

Rodrigues (1987) found two diversity patterns in the torquatus group. In the first group, from eastern South America, there is a large amount of species with restrict geographic distributions. In the second group, there is a low diversity but the species have wide geographic distributions. According to Rodrigues, species of the second group are related to Central Brazil, noting that none can be treated, without restrictions, as an animal of the great South American diagonal of open formations. Still, from northeast to southwest, there is T. hispidus inhabiting mainly the Caatingas, T. itambere , T. oreadicus and the populations of T. torquatus from central Brazil with wide distributions in the Cerrado Domains (noting that the torquatus distribution in Central Brazil seems to be more related to gallery forests than to the extensive savanna-like formations in the Cerrado; see Nogueira et al. 2005 and Nogueira 2006), and T. etheridgei mainly in the Chaco Domains, with some populations in the Cerrado.

Tropidurus catalanensis has a parapatric distribution in relation to T. etheridgei (which occurs exclusively west of the Paraná Basin), with a more continuous distribution and less restricted to the river valleys in the Pampa Domains of southwestern Rio Grande do Sul and northwestern Uruguay, reaching marginally the Chaco Domains in the vicinities of the Paraná and Paraguay rivers (in the border of Argentina and Paraguay) and restricted to the river valleys and canyons of the Paraná-Uruguay Basin in the Paranaense Forest (sensu Morrone, 2001).

The distribution patterns of T. catalanensis and T. imbituba sp. nov. seem to corroborate the two distribution patterns found by Rodrigues, and these can be strengthened with the recognition of specific status for the torquatus populations from the mountain ranges of southeastern Brazil (with probably restricted distribution) and for those from Central Brazil (with a broader distribution). However, further conclusions on the relationship of the species associated with the Tropidurus torquatus complex and their distributional patterns must await ongoing phylogenetic and phylogeographic studies of the Tropidurus torquatus group.

Figures

FIGURE 18. Tropidurus torquatus from Arraial do Cabo, north coast of Rio de Janeiro. Adult male (UFRGS 5756; SVL 70 mm).

FIGURE 19. Tropidurus torquatus from Comendador Levy Gasparian, mountain ranges of Rio de Janeiro, near to the border with Minas Gerais. Adult male (UFRGS 5829; SVL 126.6 mm). Photo by Rafael Balestrin.

FIGURE 20. Tropidurus torquatus from Brasília National Park, Brasília, Distrito Federal, Central Brazil (not collected).