PUBLISHED BY THE AMERICAN MUSEUM OF NATURAL HISTORY CENTRAL PARK WEST AT 79TH STREET, NEW YORK, NY 10024 Number 00, 27 pp., 7 figures, 3 tables , 2007

A New Genus of Microteiid Lizard from the Atlantic Forests of State of Bahia, Brazil, with a New Generic Name for Colobosaura mentalis, and a Discussion of Relationships Among the Heterodactylini (Squamata, Gymnophthalmidae)

MIGUEL TREFAUT RODRIGUES1,2, KATIA CRISTINA MACHADO PELLEGRINO3, MARIANNA DIXO4, VANESSA KRUTH VERDADE2, DANTE PAVAN2, ANTOˆ NIO JORGE SUZART ARGOLO5, AND JACK W. SITES, JR.6

ABSTRACT A new genus and species of microteiid lizard is described from a series of specimens obtained in the leaf litter at Una (15u109S, 39u039W) in the Atlantic forest of southern Bahia, Brazil. It is characterized by the presence of prefrontals, frontoparietals, parietals, and interparietal; parietals longer than wide; distinct ear openings and eyelids; two pairs of genials, absence of collar and occipital scales; dorsal scales anteriorly smooth and becoming gradually lanceolate and mucronate posterior to the forelimb; and four regular transverse series of smooth ventrals that are longer than wide, identical in size. A phylogenetic analysis based on external morphology, osteology, and molecular data confirms this new lizard as a member of the Heterodactylini radiation of Gymnophthalminae. The topology recovered by maximum parsimony (MP) analyses reveals that its closest relatives are the sister taxa Colobosaura modesta and Iphisa elegans (BS 5,50%;

1Division of Vertebrate Zoology (Herpetology), American Museum of Natural History. 2Universidade de Sa˜o Paulo, Instituto de Biocieˆncias, Departamento de Zoologia, Caixa Postal 11.461, CEP 05422-970, Sa˜o Paulo, Brazil ([email protected]; [email protected]; [email protected]). 3Universidade Federal de Sa˜o Paulo, Campus Diadema, Sa˜o Paulo, Brazil ([email protected]). 4Universidade de Sa˜o Paulo, Instituto de Biocieˆncias, Departamento de Ecologia ([email protected]). 5Universidade Estadual Santa Cruz, Ilhe´us, Bahia, Brazil ([email protected]). 6Department of Integrative Biology and M. L. Bean Life Science Museum, Brigham Young University, Provo, Utah, US 84602 ([email protected]).

Copyright E American Museum of Natural History 2007 ISSN 0003-0082

American Museum Novitates novi-496-00-01.3d 26/2/07 16:51:49 1 Cust # 496 2 AMERICAN MUSEUM NOVITATES NO. 00

Bremer value 5 2) and the partitioned Bremer indexes indicated that the largest contribution to this relationship comes from morphology; Colobosaura mentalis, for which a new generic name is here proposed, is basal to this radiation. Our analyses confirm a previous hypothesis suggesting Stenolepis as a member of the Heterodactylini radiation and that the clade composed of Colobodactylus and Heterodactylus is the sister group of the clade formed by Colobosaura mentalis- Stenolepis (BS 5 100; Bremer value 5 18), Colobosaura modesta-Iphisa (BS 5,50%; Bremer value 5 1), and the new genus here described. The support for Heterodactylini monophyly, on the basis of combined MP analyses is higher (BS 5 96, Bremer value 5 11) than that previously found in molecular-based studies only. Partitioned Bayesian methodology combining molecular and morphological data sets recovered the new genus as the sister taxon (PP 5 0.94) of the clade (PP 5 0.94) formed by I. elegans-C. modesta (PP 5 0.51) and C. mentalis-S. ridley (PP 5 1.0). An alternative topology demonstrating a paraphyletic Heterodactylini is only weakly supported (PP 5 0.63). Based on the MP topology we discuss tentative scenarios for the evolution of Heterodactylini.

INTRODUCTION have been made to infer relationships among the genera (Myers and Donnelly, 2001; Hoyos, The Gymnophthalmidae represent a large 1998), and one entirely new clade of micro- South and Middle American radiation (41 teiids (characterized by scincoid scales) was genera and about 180 species) of small to recognized (Rodrigues, 1991a, b, c). All of medium-sized lizards occurring in terrestrial, these initiatives were based on morphology. arboreal, fossorial, and semiaquatic habitats Only in this century, the first extensive extending from sea level to the high Andes molecular-based phylogenetic hypothesis was (Pellegrino et al., 2001; Doan, 2003; Doan proposed for Gymnophthalmidae, which were and Castoe, 2005; Rodrigues et al., 2005). recovered as monophyletic (Pellegrino et al., Taxonomic study of these so-called ‘‘micro- 2001). Based on the study of 26 representatives teiid’’ lizards (Ruibal, 1952) has been compli- of the 36 genera recognized at the time, four cated by the rarity of specimens in collections, subfamilies were recognized: Alopoglossinae, which limits studies of geographical and Rhachisaurinae, Cercosaurinae (with two individual variation, and convergence in char- tribes, Cercosaurini and Ecpleopini), and acter complexes (e.g., body elongation, limb Gymnophthalminae (also with two tribes: reduction, earlessness, lack of eyelids, and the Gymnophthalmini and Heterodactylini). fusion/fission of some major head scales), Genera not represented in the molecular which has rendered higher-level taxonomy analysis of Pellegrino et al. (2001) were of the Gymnophthalmidae problematic. Bou- tentatively allocated to the recognized clades lenger (1885) first organized their chaotic on the basis of morphology. Pellegrino et al. taxonomy by recognizing three groups of also demonstrated the paraphyly of several microteiids and another group (macroteiids) microteiid genera, giving support to previous in the Teiidae. After this, several attempts hypotheses suggesting extensive character were made to validate or understand interge- convergence, and making understandable the neric relationships between and within the unsuccessful attempts of previous classifica- boulengerian groups (Ruibal, 1952; MacLean, tions. 1974; Presch, 1980; Harris, 1985; Sullivan and Castoe et al. (2004) using only four of the Estes, 1997; Hoyos, 1998). The relationships five genes explored by Pellegrino et al. (2001) between gymnophthalmids and teiids (macro- and a slightly improved sampling design teiids) were generally recognized, with micro- (twelve additional species and one additional teiids considered either as a distinct family genus), reanalyzed the Pellegrino et al. (2001) (Presch, 1983, 1988; Estes, 1983; Estes et al., data. Their results, based on partitioned 1988) or as a subfamily (MacLean, 1974; Bayesian analyses, were highly consistent Presch, 1978, 1980) of Teiidae. In the last two with those obtained by Pellegrino et al. decades, several new taxa of the Gymno- (2001) with the following higher-level taxo- phthalmidae have been described (Rodrigues, nomic changes: (1) Ptychoglossus was includ- 1991a, 1991b, 1991c, 1997; Myers and ed in Alopoglossinae; (2) Heterodactylini Donnelly, 2001; Kok, 2005); some attempts and Gymnophthalmini were combined in

American Museum Novitates novi-496-00-01.3d 26/2/07 16:51:49 2 Cust # 496 2007 RODRIGUES ET AL.: MICROTEIID LIZARD 3

a Gymnophthalminae without tribal divisions; biodiversity, phylogenetic relationships within (3) Ecpleopini was raised to subfamily status, the family Gymnophthalmidae remain poorly and (4) Bachia was allocated to the new tribe known. We therefore follow Rodrigues et al. Bachiini within the Cercosaurinae. (2005) and use the classification proposed Since then, both schemes have been referred by Pellegrino et al. (2001) for the Gymno- in the literature (Doan and Castoe, 2005; phthalminae until extensive character and Rodrigues et al., 2005). In a paper focused on taxon sampling has been completed. clarification of the phylogenetic taxonomy of Pellegrino et al. (2001) defined the tribe the cercosaurines, Doan and Castoe (2005) Heterodactylini, on the basis of molecular followed the arrangement of Castoe et al. evidence, to include the genera Colobo- (2004). They described two new genera dactylus, Colobosaura, Heterodactylus, Iphisa, (Potamites and Petracola) to accommodate and probably the genus Stenolepis (which was a group of species formerly included in not sampled by Pellegrino et al.; or Castoe et Neusticurus and Proctoporus, respectively, al. 2004) on the basis of morphology. In this and resurrected Riama to allocate another paper we include the latter genus and add group of Proctoporus. Rodrigues et al. (2005) morphological data (completely absent from fully agreed with the reallocation of the studies of both Pellegrino et al., and Ptychoglossus, but continued to follow the Castoe et al.) to extend our studies of relation- proposal by Pellegrino et al. (2001) in other ships within the Heterodactylini. We describe respects because they considered the other an unknown genus and species within this changes proposed by Castoe et al. (2004) clade, and discuss its relationships to the other premature; this decision was based on the heterodactyline genera. The first specimen of evidence provided by morphological charac- the new taxon came to our attention in the ters included in the study of Rodrigues et al. mid-1990s, when the first author received for Although the study by Rodrigues et al. (2005) examination a lizard obtained at Ilhe´us, state was based on a smaller data set, they included of Bahia, Brazil. It was sent by P.E. Vanzolini representatives of the major clades to test the with a note saying that it was possibly a new relationships of the new genus described species of Colobosaura. That specimen was (Dryadosaura) in that paper, and their com- discovered during the process of reorganiza- bined morphological and molecular data re- tion of the herpetological collection of the late covered monophyly for the Heterodactylini Werner Bokermann, now at the Museum of and Gymnophthalmini, as well as a sister- Zoology, University of Sa˜o Paulo. Some years group relationship between them, and all with later, two additional specimens were obtained strong support (their fig. 5B). They also near Itabuna, state of Bahia by one of us (A. recovered the ecpleopines as a monophyletic Argolo). It was not until 2000 that a large group unrelated to Cercosaurini, in agreement series of specimens was collected in pitfall with Castoe et al. (2004). Nevertheless, based traps at Una by M. Dixo during a survey of on the shape of the postorbital, its contact the herpetofauna of the Atlantic forests of with the postfrontal, and the shape of the southern Bahia. Additional specimens were interclavicle all suggesting a close relation- subsequently obtained at Serra do Teimoso in ship between Ecpleopini and Cercosaurini, the same general region. Rodrigues et al. (2005) retained the tribal Vanzolini’s tentative generic allocation was assignment of Pellegrino et al. (2001) while accurate in the context of the taxonomy of awaiting additional evidence. Additionally, that time, and this was also its provisional based on the limited taxonomic and character allocation in the molecular study of Pellegrino sampling of cercosaurines, Rodrigues et al. et al. (2001), where it was referred as Colo- (2005) did not adopt the Bachiini proposal of bosaura sp. and recovered deeply nested within Castoe et al. (2004). the Heterodactylini. Nevertheless, as recom- Although molecular studies have highlight- mended by that study and demonstrated ed and greatly improved our understanding of hereinafter, we allocate this taxon to its own microteiid relationships, geographic distribu- genus to render classification compatible with tions, character evolution, and Neotropical the recovered phylogenetic structure. Our

American Museum Novitates novi-496-00-01.3d 26/2/07 16:51:49 3 Cust # 496 4 AMERICAN MUSEUM NOVITATES NO. 00 study also shows that as currently recognized ing intraspecific polymorphism after the ex- the genus Colobosaura is paraphyletic, and, amination of the MZUSP collections. Hemi- for this reason, a new generic name is penes were examined when partly or totally proposed to allocate the former Colobosaura everted in preserved specimens in order to mentalis. improve descriptions and comparisons, but The discovery of this new taxon in Atlantic due to their limited availability among the forests of southern Bahia, one of the most genera included here, they were not used in the threatened habitats in the world (Myers et al., subsequent analyses. 2000; Rodrigues et al., 2002), reveals further The majority of DNA sequences used here undisclosed biological diversity and refines were those from Pellegrino et al. (2001), but our understanding of heterodactiline relation- new sequences for the same mitochondrial ships. This combined estimate of the evidence (12S, 16S and ND4) and nuclear (c-mos and of phylogenetic relationships of the Hetero- 18S) gene regions were included for the dactilini also raises additional questions about previously unsampled ingroup taxon Ste- historical biogeography and speciation me- nolepis ridleyi (GenBank accession numbers chanisms of this interesting group of lizards. EF 405618-405622). The present molecular partition is composed of 2,333 bp of aligned MATERIAL AND METHODS sequences; some adjustments on the original Pellegrino et al. (2001) alignments for the Snout-vent length was measured to the ribosomal 12S, 16S, and 18S regions were nearest mm with a rule; scale counts and performed manually on the reduced matrix osteological data on cleared specimens (ap- used in this study. This was necessary to pendix 1) were taken with the aid of a stereo- accommodate the sequences of Stenolepis and microscope. Scale counts and scale nomencla- exclude unnecessary gaps. Regions of ambig- ture are according to Rodrigues and Borges uous alignment for the 12S and 16S were (1997), and osteological nomenclature follows excluded from the final analyses on the Presch (1980) and Estes et al. (1988). Sex was reduced matrix. determined by the presence/absence of femoral Phylogenetic inference was first conducted pores and confirmed by dissection in pre- on the separate morphological partition fol- viously opened specimens. All data were taken lowed by a combined analyses with the from preserved specimens housed at MZUSP molecular partitions under equally weighted (Museu de Zoologia, Universidade de Sa˜o parsimony (MP) in PAUP* v4.0b10 Paulo), MNRJ (Museu Nacional, Rio de (Swofford, 2002). For a matrix of 42 morpho- Janeiro), MZUESC (Museu de Zoologia, logical characters (appendix 2), with all char- Universidade Estadual de Santa Cruz), and acter states coded as unordered and all RNHM (Nationaal Natuurhistorisch Mu- transformations uniformly weighted, a search seum, Leiden). To compare abundance of with the branch-and-bound algorithm was males and females and total abundance performed. The analysis of the combined data among habitats of the new taxon sampled we set included an MP heuristic search with 1,000 used one-way analysis of variance (ANOVA; replicates of random stepwise addition and Zar, 1996). When significant differences were TBR branch-swapping. obtained, differences between pairs of habitats Nodal support was assessed by bootstrap were tested with a Tukey’s test (Zar, 1996). analysis (BS; Felsenstein, 1985) with a 1,000 Association among habitat categories and random stepwise additions per bootstrap species or sex was detected by chi-square test. pseudoreplicate, and TBR branch-swapping, All analyses were done using Statistica in both branch-and-bound (morphology) and (StatSoft, 1998). Appendix 2 presents the 19 heuristic (combined data) searches; bootstrap species, assigned to 17 genera of Gymno- values greater than 70% (Hillis and Bull, 1993) phthalmidae (including the new one herein were interpreted as strong support for a node. described), used in the present study. External Total and partitioned Bremer support (PBS) morphological characters included in appen- values (Baker and DeSalle, 1997), the latter dix 2 were selected among those not present- representing the contribution of each specified

American Museum Novitates novi-496-00-01.3d 26/2/07 16:51:50 4 Cust # 496 2007 RODRIGUES ET AL.: MICROTEIID LIZARD 5

data partition to each node, were calculated three supraoculars. Dorsal scales anteriorly for all nodes of the combined data-partition smooth, isodiametric or subrectangular in topology using the program TreeRot v. 2.0 occipital region, becoming progressively nar- (Sorenson, 1999). rower, more elongate and rounded toward Modeltest v.3.06 (Posada and Crandall, arm level and then lanceolate, strongly keeled, 1998) was used to select the appropriate model with lateral sides almost juxtaposed. of evolution for each molecular data parti- Occipitals absent. Ventrals longer than wide, tion: 12S (TrN+C), 16S (TrN+I+C), ND4 smooth, in four regular transverse series (TVM+I+C), c-mos (K80+C) and 18S (JC). identical in size. Males with a continuous Using the combined data set (molecular and series of very conspicuous pores with no gap morphology) a partitioned Bayesian analysis between preanal and femoral; pores absent in was implemented in Mr.Bayes 3.0 (Hue- females. Hemipenis without evident biloba- lsenbeck and Ronquist, 2001), under the best tion, sulcus spermaticus single, marginated by fit models of substitution for individual gene a naked area which is also present on the regions and morphological data set as ‘‘stan- opposite side of sulcus. Between naked areas dard’’, with 3,000,000 generations, four two symmetrical series of continuous trans- chains, and trees sampled at intervals of verse rows of comb-like flounces with minute 1,000 generations. Two independent runs were spines or irregular tooth-like structures. A conducted. Trees prior to stationarity were series of enlarged spines near the base of discarded as ‘‘burn-in’’, and a 50% majority- hemipenis. rule consensus tree was obtained from 2,900 CONTENT: Alexandresaurus camacan new data points. Nodes on consensus trees from species; monotypic. both runs with posterior probability (PP ETYMOLOGY: A homage to the Brazilian values) $ 0.95 were considered as evidence naturalist Alexandre Rodrigues Ferreira, born of significant support for a given clade in Salvador, state of Bahia, as a recognition (Huelsenbeck and Ronquist, 2001). for his effort to document the natural history Two species of Alopoglossus, members of of Brazil at the end of the 18th century. After the subfamily Alopoglossinae formally recog- pursuing natural history studies under nized by Pellegrino et al. (2001) and Castoe et Vandelli at the University of Coimbra, al. (2004) and strongly supported as basal to Alexandre headed the famous viagem filoso´fi- all other clades of Gymnophthalmidae in both ca (‘‘philosophical voyage’’) to Brazil at the studies, were constrained to monophyly and end of 18th century. He remained in Brazil used as the outgroup. Selected taxa from until 1789 where, among other places, he Gymnophthalmini, Cercosaurini and Ecpl- traveled extensively along the Rio Negro, eopini were also included to give Stenolepis Madeira and Guapore´ in the Brazilian a chance to fail to be recovered within the Amazon. The travels of Alexandre were part Heterodactylini (see appendix 2). of a larger project to publish what was meant to be a monumental natural history of the RESULTS Portuguese colonies (Farias, 2001). The plan was abandoned by the end of 1807 when the TAXON DESCRIPTION royal family escaped to Brazil against the Alexandresaurus, new genus backdrop of Lisbon’s imminent conquest by Napoleonic forces. Most of the collections DEFINITION: A large and slightly elongate assembled by Alexandre during the more than gymnophthalmid (maximum SVL 70 mm) 35,000 km traveled in Brazil were published with distinctive ear opening and eyelid, large by other naturalists following the request of tail (1.63%–2.36% SVL), and slender penta- part of the material by E´ tienne Geoffroy Saint dactyl limbs; first toe lacking a claw. Hilaire, who was present in Lisbon during the Frontonasal single; prefrontals, frontoparie- subsequent conquest. tals, parietals, and interparietals present. COMPARISONS: Considering the now recog- Parietals and interparietal longer than wide. nized multiple origins of character complexes Collar fold absent. Two pairs of genials and associated with fossoriality in gymnophthal-

American Museum Novitates novi-496-00-01.3d 26/2/07 16:51:50 5 Cust # 496 6 AMERICAN MUSEUM NOVITATES NO. 00 mid lizards, and the chaotic state of their slightly wider than long in Colobosaura taxonomy based on external attributes (until mentalis, whereas the reverse occurs in recently), we compare Alexandresaurus with Colobosaura modesta. The latter species also all other gymnophthalmids while emphasizing differs from Alexandresaurus by having only the external morphological differences with three pairs of genials. As in Alexandresaurus, the genera Iphisa, Colobosaura, Coloboda- the hemipenis of Colobosaura mentalis also has ctylus, Heterodactylus,andStenolepis. Except a naked area on the side of organ opposite to for the latter, this group of genera was the sulcus, as well as flounces of right and left recovered by Pellegrino et al. (2001) as a well- lobes medially interrupted by a naked area, supported clade that included the closest a character absent in the former species. The relatives of Alexandresaurus. Stenolepis, al- only other species referred to Colobosaura was though not included in that study, was Perodactylus kraepelini. It is a young female tentatively allocated to the same radiation. (SVL 5 40 mm) described by Werner in 1910 Most character states for the taxa com- having Puerto Max: Paraguay as type locality. pared here are summarized in appendix 1. We have not examined the holotype, which is Alexandresaurus differs from Iphisa (data for lost from the Hamburg Zoological Museum the latter in parenthesis) by having: two pairs according to Dr. J. Hallermann (in litt.); based of genials (one); several rows of lanceolate and on the original description, this species is keeled dorsal scales (smooth, in two longitu- identical to Colobosaura modesta, an opinion dinal rows); ventral scales regularly transverse advanced long ago by Amaral (1933). The in four longitudinal rows (not in transverse distributional data agree with this interpreta- rows and in two longitudinal rows); and tion because the type locality is in the area of regularly transverse rows of gulars (not occurrence of C. modesta. The only character regular). Alexandresaurus also differs from that until now has precluded synonymy Iphisa in having a naked area opposite to the between C. modesta and C. kraepelini was sulcus spermaticus and large spines at the base Werner’s reference to the tongue of his of the hemipenis (naked area absent and only specimen as having ‘‘oblique plicae like small spines at base of the organ). From Alopoglossus’’. Based on the agreement of all Colobosaura Alexandresaurus can be distin- other characters mentioned in the description guished from Colobosaura by its smooth neck we think that Werner was in error in de- scales and small, smooth, and posteriorly scribing the tongue, evidently a strange char- rounded scales in the first 7–8 dorsal rows, acter and restricted to Alopoglossinae. rarely two enlarged nuchals are present at the Alexandresaurus differs from Stenolepis, central part of the second dorsal row; occipi- Colobodactylus, and Heterodactylus in having tals are never present. In Colobosaura neck a pair of prefrontals (absent). It further differs scales have keels; there are several enlarged from Stenolepis by having small dorsal neck scales, much wider than long, in the first eight scales (large, in 3–4 longitudinal rows), larger dorsal rows, the first two characteristically scales on the sides of neck (small, some almost smooth and enlarged, forming an occipital granular), ventral and gulars in regularly pair separated or not by a smaller scale; the transverse rows (not so), and smooth scales posterior 5–6 dorsal rows are striate or multi- near venter (keeled). In Colobodactylus and carinate. Ventral scales in Alexandresaurus are Heterodactylus the parietals are wider than longer than wide, laterally juxtaposed, round- long in contrast to the longer than wide ed posteriorly, identical in size and shape, parietal scale observed in Alexandresaurus regularly arranged transversally and in four and other Heterodactilini. This difference is longitudinal series. In Colobosaura there are probably due to a possible scale rearrange- also four longitudinal rows of quadrangular ment in Heterodactylus and Colobodactylus in ventral scales, but they are not regularly which parietals (and interparietal in transversal and are irregular in size and shape, Colobodactylus) were divided and their poste- laterally imbricate, and in the first ventral rior regions incorporated into the highly rows midline scales always distinctively wider conspicuous occipitals that characterize both than long; the external ventral rows are genera. This hypothesis, further discussed

American Museum Novitates novi-496-00-01.3d 26/2/07 16:51:50 6 Cust # 496 2007 RODRIGUES ET AL.: MICROTEIID LIZARD 7

under a phylogenetic framework below, re- Alexandresaurus camacan, new species veals that even head scales traditionally used figures 1–4 in taxonomic descriptions may not be strictly homologous. This is another example of the HOLOTYPE: MZUSP 93201, an adult male taxonomic confusion that may result from from Una (15u109S, 39u039W): state of Bahia: uncritical use of morphological characters in Brazil, collected by Marianna Dixo on 01 microteiid systematics. Alexandresaurus also November 1999, field number md-1106. differs from Colobodactylus and Heterod- PARATYPES: All from the state of Bahia, actylus in having five toes and smooth flank Brazil. MZUSP 93179–93200, Una, collected scales (only four toes and flank scales keeled in by M. Dixo between 24 January 1999 and 27 the latter genera). Heterodactylus further February 2000; MZUSP 93331, Ilhe´us, col- differs from Alexandresaurus by the absence lected by W. Bokerman on December 1971; of external ear, absence of frontoparietals, and MZUSP 93462–93464 Ilhe´us (CEPLAC, sede absence or extreme reduction of interparietal. regional), A. Argolo leg., collected respectively Finally, the hemipenis of Heterodactylus is on July 1990, 24 April 1991, and 12 May 1999; very different from that of Alexandresaurus. MZUSP 93227–93228 , Jussari (Serra do Teimoso) collected by M. T. Rodrigues, D. The organ is distinctively bilobate with a cen- Pavan, M. Dixo, and V. K. Verdade on March tripetal sulcus spermaticus surrounded by 2001; MZUESC 2431, Ilhe´us (Campus a naked area. The asulcate side of the organ UESC), collected by C. N. Souza on 8. has a series of continuous and unornamented March 2002; RMNH 29741, from Una, naked flounces containing only one spine at Ilhe´us. each edge, very different from the interrupted ETYMOLOGY: The specific epithet is a hom- flounces containing spines in a comb-like age to the extinct Camacan Indians, a Boto- arrangement observed in Alexandresaurus. cudo group who lived in the exuberant forests Alexandresaurus differs from members of of Bahia from north of Rio Pardo to Ilhe´us. Alopoglossinae by having imbricate, scalelike DIAGNOSIS: A gymnophthalmid with ear papillae covering the tongue (oblique plicae), openings and eyelids, and slender pentadactyl and from Rhachisauridae by having a distinc- limbs lacking the claw on first toe. tive ear opening and eyelid (absent), five toes Frontonasal single; prefrontals, frontoparie- (four), a normal body (clearly elongate), tals, parietals, and interparietals present. presence of frontoparietals (absent), and Parietals longer than wide. Collar fold absent. a nostril at the nasal border (in the center of Two pairs of genials; three supraoculars. the nasal scale). From all other Gymno- Dorsal scales in 30–38 rows, anteriorly phthalminae Alexandresaurus differs by hav- smooth, isodiametric or subrectangular in ing five toes (fifth toe reduced or absent), occipital region, becoming progressively nar- a distinctive eyelid (absent in all Gymno- rower, more elongate and rounded toward the phthalmini except for Tretioscincus), nostril at arm, and posterior to the arm becoming more the nasal border (in the center of the nasal lanceolate, strongly keeled, with lateral sides scale), a wide and flattened clavicle enclosing almost juxtaposed. Occipitals absent. Ventrals a fenestra (simple, boomerang shaped and longer than wide, smooth, in four regular without fenestrae), nasal bones in contact at longitudinal and 15–19 transverse rows, iden- midline but broadly separated anteriorly tical in size. Scales around body 29–35; 11–15 (totally separated by contact between frontal and 16–20 infradigital lamellae under finger and premaxilary), and, among other charac- IV and toe IV respectively. Males with a series ters, by having straight lateral processes of the of 20–25 very conspicuous pores without gap interclavicle (posteriorly oriented). Alexan- between preanal and femoral; pores absent in dresaurus also differs from all Cercosaurinae females. by having a distinctive sternal fontanelle DESCRIPTION OF HOLOTYPE (fig. 2A–C): process (absent), a glossohyal separated from Rostral broad, wider than high, contacting basiyal (fused), nasal bones broadly separated first supralabial, nasal, and frontonasal. anteriorly (in broad contact), and postorbital Frontonasal pentagonal, twice as wide as covering postfrontal (covered by postfrontal). long, contacting rostral, nasal, loreal, and

American Museum Novitates novi-496-00-01.3d 26/2/07 16:51:50 7 Cust # 496 8 AMERICAN MUSEUM NOVITATES NO. 00

Fig. 1. Alexandresaurus camacan (MZUSP 93228), an adult male (57 mm) from Serra do Teimoso, state of Bahia, Brazil. prefrontals. Prefrontals slightly wider than wider anteriorly and indenting suture between long, in broad contact at midline. Frontal preocular and frenocular. Seven supralabials, hexagonal, with almost parallel lateral mar- third and fourth under the eye, fifth the gins, slightly longer than wide, anteriorly highest and contacting two smaller scales indenting the prefrontal and posteriorly the following posteriorly the subocular, the smal- frontoparietal sutures. Frontoparietals pentag- lest one clearly a postocular; seventh suprala- onal, larger than prefrontals, in slight contact, bial the smallest, in contact with the granules strongly indented by the interparietal. surrounding anterior margin of ear. Three Interparietal longer than wide, longer and superciliaries, first largest, wider anteriorly, narrower than frontal, as long as and longer than first supraocular, contacting preo- narrower than parietals. Parietals heptagonal, cular, loreal, first and second supraoculars, marginated laterally by two temporal scales second superciliary and upper eyelid. Central with approximately the same size, anteriorly part of eyelid with a semitransparent un- by the third supraocular and frontoparietal, divided disc surrounded by small and slightly medially by the interparietal, and posteriorly pigmented granular smooth scales. Lower by the first dorsals. Three supraoculars, first eyelid with eight strongly pigmented palpeb- the smallest, second and third about the same rals. Temporal region with few smooth and size, second slightly wider, its longest suture juxtaposed scales, irregular in size and shape, with frontal, third slightly longer, in broad between parietals and supralabials; the larger contact with frontoparietal. Nasal above first as large as the sixth supralabial. Ear opening supralabial, large, longer than high, with the surrounded by a series of very small and nostril in the center and lower part of scale, juxtaposed rounded granules; external audito- indenting suture with labial. Loreal posterior ry meatus large, tympanum distinct, subovoid, to nasal, narrower and diagonally oriented; recessed. Lateral surface of neck with smooth, contacting posteriorly first superciliary, preo- imbricate and mostly longer than wide scales, cular, and frenocular. Frenocular below preo- larger dorsally. All head scales smooth and cular, followed posteriorly by a long subocular juxtaposed with scattered sensorial organs.

American Museum Novitates novi-496-00-01.3d 26/2/07 16:51:50 8 Cust # 496 2007 RODRIGUES ET AL.: MICROTEIID LIZARD 9

Fig. 2. Ventral (A), lateral (B), and dorsal (C) views of the head of the holotype of Alexandresaurus camacan (MZUSP 93201). Scale 5 1 mm.

Mental broad, wider than high. Postmental the level of labial comissure. An enlarged pair heptagonal, wider than long. Two pairs of of symmetric flat and chevron-like scutes genials, both contacting infralabials; the first follows second pair of genials, preventing smaller, in broad contact at midline; second their contact posteriorly. Six infralabials, pair contacting only anteriorly and reaching second the largest. Gulars smooth, imbricate,

American Museum Novitates novi-496-00-01.3d 26/2/07 16:51:57 9 Cust # 496 10 AMERICAN MUSEUM NOVITATES NO. 00

Fig. 3. Dorsal (A) and ventral (B) views of the skull of Alexandresaurus camacan (MZUSP 94252-53). Scale 5 1 mm.

American Museum Novitates novi-496-00-01.3d 26/2/07 16:52:12 10 Cust # 496 2007 RODRIGUES ET AL.: MICROTEIID LIZARD 11

Fig. 4. Right hand (A), right foot (B), hyoid (C), shoulder girdle (D), and pelvic girdle (E) of Alexandresaurus camacan (MZUSP 94252-53). Scale 5 1 mm.

rounded posteriorly, in seven transversal rows; larger and elongate scales, about twice as long at midline similar to ventrals, much smaller as gulars. Collar fold absent. toward side of neck. Gulars followed by Dorsal scales imbricate and arranged in a distinct interbrachial region with seven much regular transversal rows; smooth, isodiametric

American Museum Novitates novi-496-00-01.3d 26/2/07 16:52:33 11 Cust # 496 12 AMERICAN MUSEUM NOVITATES NO. 00 or subrectangular in occipital region, becom- infralabials, yellow at their sutures. Ventral ing progressively narrower, more elongated parts of body and tail creamy yellow, immac- and rounded at arm level and then posteriorly ulate, except in the external rows of gulars and lanceolate, strongly keeled, with lateral sides ventrals, which present irregular black spots. almost juxtaposed. Thirty transverse rows Ventral part of tail becomes gradually darker between interparietal and the posterior level toward the extremity. Limbs dark brown of hind limbs. Lateral scales about the same dorsally, irregularly mottled with yellow; size as dorsals but smooth, strongly imbricate creamy yellow and immaculate ventrally. laterally, less acuminate and more diagonally Hemipenis partly everted at preservation arranged than dorsals; those closer to ventrals except for the apex; bilobation not evident; larger. A distinctive area with granular scales sulcus spermaticus single, marginated by an surrounds the area of arm insertion. Thirty- enlarged naked area that is also present, but one scales around midbody. Ventral scales narrower, on the opposite side of sulcus. smooth, longitudinally imbricate, laterally Between naked areas of the organ, two juxtaposed, longer than wide, rounded poster- symmetrical series of 29 (or more) continuous iorly; 17 transverse rows from interbrachials transverse rows of comb-like flounces with (excluded) to preanals. Six scales in precloacal minute spines or irregular tooth-like struc- region, central and paramedials the largest. tures. Flounces continuous even near basal Total pores 23, continuous, with no gap part of the organ where their ornamentation is between femoral and preanal ones. medially interrupted by a small area without Scales of tail smaller than midbody dorsals, spines. A basal series of very enlarged spines is but otherwise identical; keeled, lanceolate, present near the base of hemipenis in its strongly imbricate longitudinally; those from asulcate face. base of tail slightly larger ventrally, becoming Measurements of the holotype: snout-vent gradually identical around tail toward the length: 66 mm; tail length: 145 mm. extremity. VARIATION: Males are slightly larger than Forelimbs with large, smooth, and imbri- females but have much longer tails: maximum cate scales; those on ventral part of brachium SVL for males and females was 70 mm and much smaller. Anterior and ventral parts of 66 mm. Tail length varied respectively in hind limbs with irregularly large, smooth, and males and females from 1.76 to 2.36 and 1.8 imbricate scales, identical to the correspon- to 2.03 times SVL (fig. 5: r2 males 0.93; r2 dent parts of fore-limbs. Posterior part of hind females 0.29). No sexual differences were limbs with granular, juxtaposed scales, grad- found in squamation. Variation in meristic ing progressively to larger, imbricate, and characters (N 5 28) was the following (mean keeled scales in dorsal part of tibia. Carpal and standard error, respectively, in the paren- and tarsal scales large, imbricate; supradigital thesis): dorsal rows, 28–30 (29 and 0.72); lamellae, smooth imbricate. Palmar and plan- ventrals 15–19 (16.8 and 1.06); scales around tar surfaces with smooth, small granules; midbody 29–35 (31.5 and 1.42); infradigital infradigital lamellae single, 12 on Finger IV lamellae under finger IV 11–15 (12.9 and and 19 on Toe IV. Toes and fingers, except for 0.19); and infradigital lamellae under Toe IV Finger I, clawed, and respectively in the 16– 20 (18.1 and 1.0). All specimens show following relative sizes: 1 , 2 5 5 , 3 5 4 three supraoculars and three superciliaries. All and, 1 , 2 , 5 , 3 , 4. the specimens have seven supralabials, and all Dorsal surfaces of body and tail and lateral have six infralabials except RM 29791, which part of tail dark brown with an irregularly has seven. One specimen (MZUSP 93187) has distributed darker punctuation smaller than four ventral rows fused midventraly resulting one fourth of the scale size. Flanks pre- in two longitudinal rows of ventrals. We note dominantly yellowish to cream, as the ventral that this arrangement is similar to the pattern parts of body and tail, but strongly mottled by found in Iphisa, but differs from it because the an irregularly black pattern. Lateral parts of rows of scales are regularly arranged transver- head with an identical pattern, black blotches sally in Alexandresaurus and not so in Iphisa. concentrated in central parts of supra- and Only males have pores, they are absent in

American Museum Novitates novi-496-00-01.3d 26/2/07 16:53:04 12 Cust # 496 2007 RODRIGUES ET AL.: MICROTEIID LIZARD 13

without overlapping, and extensively covering prefrontal and jugal; 21 maxillary teeth. Prefrontal large, its posterior process long but not reaching the level of middle of orbit; in broad contact with frontal. Lacrimal large, very conspicuous, contacting prefrontal and maxillary and indenting the orbit. Postfrontal and postorbital single. Postfrontal triangular, contacting jugal, frontal, postorbital and jugal, closing posteriorly the orbita. Postor- bital long and wide, contacting squamosal posteriorly, almost closing the supraorbital fenestra. Squamosal long, posteriorly curved and abutting the top of quadrate. Supra- temporal fenestra almost closed and strongly Fig. 5. Regression lines between body length constrained laterally by parietal and post- and tail length for males and females of orbital. Supratemporal present, in straight Alexandresaurus camacan from Una, State of contact with posterior part of parietal and Bahia, Brazil. squamosal. Fifteen scleral ossicles in the eye. Vomer, palatine, pterygoid, ectopterygoid, females. Pores are highly conspicuous, placed and pterygoid present. Vomer, palatine, pre- in distinctively elevate scales, aligned on each maxillary, and maxillary in contact restricting side without gaps between preanal and femo- the fenestra exochoanalis. Infraorbital fenes- ral. Total number of pores varied from 20–25 tra large, bordered posteriorly by ectoptery- (N 5 16). Sexual color dichromatism is goid. Stapes rodlike, wider at the base. Sutures evident; males are characterized by dark between supraoccipital, exoocipital, basiooci- coloration in the flanks mottled with yellowish pital and the otic area of skull are not as or whitish blotches and a reddish venter; visible as those between basioccipital and females and juveniles are more dull colored basisphenoid. In the lower jaw dentary, having unpatterned light brown flanks and articular, splenial, angular, and supraangular a creamy white venter. are distinct; there are 14–15 dentary teeth, OSTEOLOGY (figs. 3–4): The following de- conic anteriorly, bicuspid or tricuspid poster- scription is based on two alizarin-prepared iorly. skeletons (MZUSP 94252–53). Premaxillary as Glossohyal long, fused to basihyal. First long as wide, contacting but not articulating ceratobranchial curved posteriorly; hypohyal laterally with the maxillary. Its dorsal lamina and ceratohyal present. A second short pair of triangular posteriorly, long, just covering ceratobranchials is present and positioned anteriorly the nasals and deeply indenting parallel to the anterior part of trachea. their suture, preventing their anterior contact. Anterior part of clavicle very enlarged, Twelve premaxillary conical teeth. Nasals flattened, enclosing a fenestra. Interclavicule large, longer than wide, wider anteriorly, long, cruciform, with very long lateral pro- diagonally arranged, widely separated anteri- cesses reaching the sternum but not sternal orly, posterior third in midline contact, fenestra. Scapulocoracoid with coracoid, scap- covering frontal anteriorly. Frontal longer ular, and scapulocoracoid fenestrae; suprasca- than wide, wider posteriorly, covering parietal pula present. Sternum with a large fenestra, and articulating with it by a pair of frontopar- receiving three sternal ribs and a xiphisternum ietal tabs. Parietal longer than wide, wider and with two inscriptional ribs. Ilium, ischium, and concave posteriorly and covering laterally the pubis present, the latter with a conspicuous occipital region. Epipterygoid contacting ex- pectinate apophysis. Hypoischium long, larger ternally a descending epipterygoid process of at the base, almost reaching the preanal border; parietal. Maxillary dorsally contacting nasal, a small elongate preischium and a small qua- and lateral parts of frontal and lacrimal, drangular ossificate prepubis are present.

American Museum Novitates novi-496-00-01.3d 26/2/07 16:53:04 13 Cust # 496 14 AMERICAN MUSEUM NOVITATES NO. 00

Twenty-seven procoelous presacral verte- distance from the coast. Uruc¸uca (14u339S, brae, neural spines low, higher anteriorly 39u189W) is located on fertile soils originating hypopophyses present in the first six verte- from volcanic deposition and is intensely brae; zigantrum-zygosphene present. Last cultivated for cacao production. Ilhe´us and presacral vertebra lacking ribs. Two sacral the Reserva Biolo´gica do Una (15u109S, vertebrae. First caudal vertebrae with long 39u039W) are located near the coast on the transverse processes, lacking autotomic septa. sandy soils of Barreiras Formation and receive Humerus and femur are slightly larger than more rainfall. The RPPN (Reserva Particular radius and ulna and tibia and fibula. do Patrimoˆnio Natural; Serra do Teimoso; Remaining elements of the forelimbs and hind 15u199S, 39u319W) is an isolated peak toward limbs as in figure 4. the interior of the state and is covered by a semideciduous forest, along with cacao DISTRIBUTION AND NATURAL HISTORY groves at the lower areas and tropical rain- forest with high humidity levels near the top. Alexandresaurus camacan is currently Pitfall traps were installed at Una, Serra do known only from four localities in the Teimoso, and Uruc¸uca to sample areas of southern Atlantic rainforest region of the state primary and secondary forest and cacao of Bahia: Ilhe´us, Una, Jussari (Serra do groves. At Serra do Teimoso a series of pitfall Teimoso), and Uruc¸uca. The maximum traps also sampled semideciduous forest. straight-line distance between the most distant Alexandresaurus camacan was always ob- of these localities is less than 60 km. Pristine tained in sheltered areas in primary forest or tropical forest with high levels of humidity and cacao groves, never in secondary forests. At rainfall (about 2,000 mm) regularly distribut- Serra do Teimoso the only individual collected ed throughout the year, with a dense leaf litter, was found in semideciduous forest at high large trees reaching up to 30 m high and more elevation (about 900 m) in the humid forest. than 50 cm diameter at breast height, and an No association between habitat categories and abundance of bromeliads and other epiphytes, sex was detected (p . 0.05, chi-square test). was the typical habitat in the Atlantic rain- Alexandresaurus camacan was always found in forest at this latitude. Most of this area is shady areas in primary forest or cacao groves, dominated by lowlands or low hills dissected and never in secondary forests. Total abun- by small- to medium-sized streams, inter- dance (F(2, 27) 5 1,06, p 5 0.36) and female rupted by isolated higher mountains. The abundance (F(2, 27) 5 0.32, p 5 0.73) did not Serra do Teimoso is one these higher moun- vary significantly among habitats (forest in- tains reaching about 1,000 m altitude and terior, forest edge, and cacao groves), while receiving the highest levels of humidity at the male abundance did (F(2, 27) 5 4.70, p 5 top as a result of the orographic effect and 0.018). Male abundance was significantly condensation of air masses coming from the higher in forest edges than in cacao groves (p sea. Currently most of the forest cover is 5 0.03) and marginally higher than in forest restricted to scattered patches of remnant interior (p 5 0.06). forest, the Una reserve (Reserva Biolo´gica de Una—ca. 11,500 ha) being the largest one. PHYLOGENETIC ANALYSES The largest area of original forest was destroyed by agriculture (for cacao planta- Parsimony analyses to determine the tions), cattle ranching, and timber harvest phylogenetic placement of Alexandresaurus activities. In cacao groves (‘‘cabrucas’’ locally) camacan and its relationships within the where large and emergent trees have been Heterodactylini, were first conducted on the preserved to shade the understory cacao partition composed of 42 morphological plants, leaf litter is abundant and humidity is characters (38 of which are parsimony in- higher than in other disturbed areas. formative; appendix 2) with all states coded as The localities where Alexandresaurus cama- unordered. This analysis recovered 211 most can was collected show slight differences parsimonious trees; the poorly resolved strict related to soil type, levels of humidity, and consensus tree (L 5 106; CI 5 0.55 and RI 5

American Museum Novitates novi-496-00-01.3d 26/2/07 16:53:08 14 Cust # 496 2007 RODRIGUES ET AL.: MICROTEIID LIZARD 15

0.72; fig. 6A) left the position of BS 5 63, Bremer value 5 8.0). The mono- Alexandresaurus unresolved within the phyly of the ecpleopines was also strongly Heterodactylini, but, although weakly sup- supported (node 15: BS 5 92, Bremer value 5 ported, recovered the tribe as monophyletic 11). (BS 5 58, Bremer value 5 4.0). Except for Externally, the (C. ocellata (Rhachisaurus a well-supported sister relationship between brachylepis-B. bresslaui)) clade was weakly Colobodactylus taunayi-Heterodactylus imbri- supported (node 2: BS , 50, Bremer value 5 catus (BS 5 95, Bremer value 5 4.0) this tree 6.0), and the Heterodactylini-Gymno- left all other ingroup taxa as an unresolved phthalmini clade (node 10) and the Ecpl- polytomy. Two other well-supported sister- eopini clade (node 15) were recovered with group relationships recovered by the morpho- high support indexes (BS 5 96, Bremer value logical partition were those of Procellosaur- 5 11; BS 5 92, Bremer value 5 11; fig. 6B, inus tetradactylus-Micrablepharus maximiliani respectively). (BS 5 99, Bremer value 5 4.0) and Colo- Figure 7 shows the majority-rule consensus bosauroides cearensis-Anotosaura vanzolinia tree estimated using partitioned Bayesian (BS 5 88, Bremer value 5 2.0). method, which differs from that recovered Combined analyses of the morphological from parsimony analysis (fig. 6B). The and molecular partitions resulted in a single major difference refers to the paraphyly of most parsimonious tree of 2,740 steps, 660 the Heterodactylini with respect to the parsimony-informative characters, and CI 5 Gymnophthalmini, but this alternative topol- 0.47 and RI 5 0.43 (fig. 6B). The recovered ogy is only weakly supported by posterior relationships are very similar to those sug- probabilities (PP 5 0.63). Bayesian tree is gested by the previous study of Pellegrino et unique in recovering a quite well-supported al. (2001) based on molecular data only. The sister relationship (PP 5 0.94) between Heterodactylini was recovered with strong I. elegans–C. modesta (PP 5 0.51) and support (node 8: BS 5 92, Bremer value 5 C. mentalis–S. ridleyi (PP 5 1.0), with 11; see appendix 4). Within the Hetero- Alexandresaurus basal to this grouping (PP dactylini, A. camacan is weakly resolved as 5 0.94). Also, the sister-group relationship the sister group of I. elegans-C. modesta (node between Bachia and Cercosaura is highly 5: BS , 50; Bremer value 5 2.0; see supported (PP 5 0.99). Similar to the MP appendix 4); (C. mentalis + S. ridleyi)is tree (fig. 6B), the support for Ecpleopini recovered as a strongly supported clade (node monophyly is high (PP 5 1.0), although its 6: BS 5 100, Bremer value 5 18, see position is unresolved (fig. 7). appendix 4); and both the monophyly of the TAXONOMIC IMPLICATIONS: Our results clade (Colobodactylus taunayi + H. imbricatus) show that the position of Colobosaura mentalis and that of its sister group (A. camacan, C. in the tree is incompatible with a monophyletic modesta, I. elegans + C. mentali -S. ridleyi) are Colobosaura. This was also indicated in the strongly and moderately well supported (re- analysis of Pellegrino et al. (2001), which spectively, node 3: BS 5 99, Bremer value 5 although based exclusively on molecular data 13, and node 7: BS 5 75, Bremer value 5 5; incorporated a much larger dataset. The sister- appendix 4). The Gymnophthalmini repre- group relationship between morphologically sented by the Procelosaurinus-Micrablepharus highly divergent species like Iphisa elegans and clade (node 9: BS 5 100, Bremer value 5 28) Colobosaura modesta, suggests the recognition is resolved as the sister clade of the of another new genus to allocate Colobosaura Heterodactylini with high support (node 10: mentalis. We propose the following name to BS 5 96, Bremer value 5 11). Slightly accommodate it: different from the Pellegrino et al. (2001) hypothesis are the weakly supported sister Acratosaura, new genus relationship Rhachisaurus-Bachia (node 1: BS , 50, Bremer value 5 1.0) and the placement DEFINITION: A microteiid characterized by of Arthrosaura reticulata as the sister taxon of a distinctive ear opening and eyelid, large tail, the C. cearensis-A. vanzolinia clade (node 13: and slender pentadactyl limbs. Frontonasal

American Museum Novitates novi-496-00-01.3d 26/2/07 16:53:08 15 Cust # 496 16 AMERICAN MUSEUM NOVITATES NO. 00 50%) and decay indexes are presented above and below branches, respectively. The inset . 0.72). (B) Single most parsimonious tree recovered from analyses of combined morphology and 0.43). The internal nodes are numbered (circles) and the partitioned Bremer support indexes are listed 5 5 0.55; RI 0.47, RI 5 5 and its closest relatives. 106; CI 2.740, CI 5 5 Alexandresaurus camacan Fig. 6. Phylogenetic trees inferred from maximum parsimony analyses. (A) Strict consensus of 211 equally parsimonious trees estimated from in appendix 4 forincludes each node. Bootstrap support values ( molecular partitions (L morphology partition (L

American Museum Novitates novi-496-00-01.3d 26/2/07 16:53:08 16 Cust # 496 2007 RODRIGUES ET AL.: MICROTEIID LIZARD 17

Fig. 7. Phylogenetic tree estimated under Bayesian method of the combined morphology and molecular data set and posterior probabilities (above nodes) for clades recovered at the 50% majority rule consensus topology.

single; prefrontals, frontoparietals, parietals, rows; laterally imbricate, irregular in size and and interparietals present. Parietals longer shape, scales of external rows slightly wider. than wide. Collar fold absent. Three pairs of Males with continuous series of very conspic- genials and three supraoculars. Dorsal scales uous pores without gap between preanal and anteriorly smooth, large in the occipital re- femoral; pores absent in females. Frontal bone gion, becoming gradually elongated, lanceo- anteriorly divided. late, and strongly keeled posteriorly. Occi- CONTENT: Acratosaura mentalis, new ge- pitals present. Ventral scales quadrangular, in neric name for Colobosaura mentalis Amaral, longitudinal but not in regularly transverse 1933.

American Museum Novitates novi-496-00-01.3d 26/2/07 16:53:12 17 Cust # 496 18 AMERICAN MUSEUM NOVITATES NO. 00

ETYMOLOGY: From the Greek akratos, (appendix 4). One is the Heterodactylus- meaning ‘‘unmixed, pure’’; here intended to Colobodactylus assemblage; the other assem- reflect its unmixed condition after being bles all other Heterodactylini (respectively dismembered from the polyphyletic Colo- nodes 3 and 7, fig. 6B). This conclusion is bosaura. supported by the condition of the supratem- poral fenestra, the lateral expansion of the DISCUSSION parietals, the postorbital width, shape of postfrontal, and the shape and size of parietal Our study provides evidence for the mono- scale (characters 23, 31, 33, 35, 37 and 39; phyly of the Heterodactylini and its sister appendix 1). The assemblage Colobodactylus- relationship with Gymnophthalmini (nodes 8 Heterodactylus also differs from the remaining and 10 respectively; fig. 6B). Castoe et al. Heterodactylini by their more elongated body (2004) admitted monophyly of Gymno- and degree of limb reduction (characters 17 phthalminae as proposed by Pellegrino et al. and 21). The genera Alexandresaurus, Colo- (2001), but removed their tribal arrangement. bosaura, Iphisa, Acratosaura, and Stenolepis In contrast to Castoe et al. (2004), our are also characterized by a strong sexual parsimony analyses (MP) recover both the dichromatism absent in Colobodactylus and monophyly of the Heterodactylini and Heterodactylus. Reproductive males of all Gymnophthalmini (nodes 8 and 9; fig. 6) and genera in the first clade exhibit a reddish their sister-group relationship (node 10; venter and have a lateral color pattern fig. 6B). The support for Heterodactylini consisting of cream, white, or yellowish monophyly, on the basis of combined analyses blotches against a dark brown or black of molecular and morphological characters, is ground color, which is absent in females. higher (BS 5 96, Bremer value 5 11) than that Alexandresaurus camacan was recovered as previously obtained using molecular data only the sister group of the (Iphisa elegans- (BS 5 71, Bremer value 5 2 in Pellegrino et Colobosaura modesta) clade with weak sup- al., 2001). The partitioned Bayesian analyses port, and the partitioned Bremer index shows we performed here differently from Castoe et that the largest contribution to this relation- al. (2004), included 42 morphological char- ship comes from morphology (node 5, fig. 6B, acters. Agreeing with them, our analyses also appendix 4). The present study also confirms recovered a paraphyletic Heterodactylini with the previously hypothesized inclusion of the respect to the Gymnophthalmini, but there is genus Stenolepis within the Heterodactylini, in low support for such alternative topology (PP contrast to its affinities with Tretioscincus as 5 0.65). More extensive sampling of morpho- suggested by Presch (1980). Acratosaura logical and molecular characters is essential to (Colobosaura mentalis in fig. 6B) is recovered confirm the result of our Bayesian analyses; as the sister group of Stenolepis, and this these studies are in progress, but until they are relationship received high support (BS 5 100, complete we accept the tree topology re- Bremer value 5 18). covered by MP analyses as the best hypoth- Although our taxonomic sampling is limited esis. with respect to total microteiid diversity and is Monophyly of the Heterodactylini is re- focused primarily on the Heterodactylini, the inforced by the unique condition of the shape monophyly of the Ecpleopini was also con- of their interclavicle and the condition of their firmed. Relationships among the genera are nasals (characters 25 and 34, appendix 1). The similar to those obtained by Pellegrino et al. better support now obtained is due not only to (2001), except for the position of Arthrosaura, the inclusion of morphological characters in which was basal and here is recovered as the this analysis but also to the addition of the sister group to the (Colobosauroides- genus Stenolepis, absent in the Pellegrino et al. Anotosaura) clade (node 13; fig. 6B), with (2001) study. Our results also indicate that the morphology making the second largest con- Heterodactylini consists of two strongly sup- tribution to the partitioned Bremer index ported clades, and that morphological char- (node 13, appendix 4). In contrast to acters are the major contribution to each Pellegrino et al. (2001), Castoe et al. (2004)

American Museum Novitates novi-496-00-01.3d 26/2/07 16:53:13 18 Cust # 496 2007 RODRIGUES ET AL.: MICROTEIID LIZARD 19

recovered the Ecpleopini assemblage distantly the Rio Sa˜o Francisco, but its association with related to Cercosaurini, which led them to mountainous areas is evident because it also raise the status of the former to subfamilial occurs in isolated remnant forests of the level. Our results confirm monophyly of the semiarid northeastern Brazil (for example, Ecpleopini with higher support than previous- Serra de Baturite´ [1115 m altitude], state of ly obtained (BS 5 92, BS 5 11 vs. BS 5 75, BS Ceara´, and Pico do Jabre [1044 m], state of 5 6 in Pellegrino et al., 2001) and not related Paraiba). Acratosaura mentalis has a larger to Cercosaurini, a result consistent with the but seemingly insularized distribution; it is position suggested by Castoe et al. (2004). known from some isolated areas in the In a different part of the tree, we recovered Caatingas of northeastern Brazil, and else- the Rhachisaurinae Rhachisaurus brachylepis, where in the Serra do Espinhac¸o, a large as nested with Bachia and Cercosaura, a result mountain range reaching to 2,400 m in we regard as highly suspect. Pellegrino et al. altitude that extends from the state of Minas (2001) recovered the Rhachisaurinae as the Gerais to Bahia. Colobosaura modesta and sister group of Gymnophthalminae, and ap- Iphisa elegans are the only Heterodactylini not pendix 4 shows that the strongest contribution directly associated with mountains or Atlantic for this relationship (node 1) comes from forests; the first is a typical Cerrado lizard morphology, specifically from characters clas- occurring both in open areas and in gallery sically subject to convergence (see appen- forests, and the second is restricted to dix 2). Further, because our taxonomic sam- Amazonian forests. pling of Cercosaurini is limited (of the 13 We are aware that the ideal approach to genera now accepted [Doan, 2003] only Bachia understand the historic biogeography of the and Cercosaura are represented here), we heterodactyline should be based on a well- prefer to keep the scheme adopted by supported phylogenetic hypothesis derived Pellegrino et al. (2001) until more complete from more extensive molecular and morpho- taxonomic sampling and additional data sets logical character sampling than that now provide better resolved phylogenetic hypothe- available. In our hypothesis, for example, it sis upon which sound taxonomic decisions can is evident that nodes 4, 5, and 7 need be made. additional support. However, some considera- Most of the Heterodactylini are restricted tions based on the present topology might be to eastern Brazil. The two species of appropriate at this time and advanced as Colobodactylus (C. dalcyanus and C. taunayi) a preliminary working hypothesis to be tested are restricted to the Atlantic forest of south- by additional data. eastern Brazil (Vanzolini and Ramos, 1977). Assuming the topology of figure 6B, an The genus Heterodactylus includes one species important point to note is the ancestral (H. imbricatus) occurring in the Atlantic position of Atlantic forest forms compared forests of southern Brazil but extending into to those from central Brazilian Cerrados elevated areas in the mountains of the state of (Colobosaura modesta) or Amazonia (Iphisa Minas Gerais, where the only other species (H. elegans). Another interesting issue suggested lundi) seems to be restricted. Apart from some by the resulting topology is the insularized records near the coast, the distribution of distribution of the older genera as originally Colobodactylus-Heterodactylus appears to be hypothesized by Vanzolini and Ramos (1977). restricted to mountainous terrain once cov- They hypothesized specifically that Colo- ered by or influenced by the Atlantic Forest bodactylus and Heterodactylus were more Domain (see Vanzolini and Ramos, 1977). widely distributed during colder episodes of Except for Iphisa elegans and Colobosaura the Quaternary, and are presently isolated in modesta, the other genera of the Hetero- mountains or especially favorable areas. The dactylini are also associated with Atlantic occurrence of Stenolepis ridleyi in a forest forest or contiguous elevated areas. Alex- refuge in northeastern Brazil also suggests andresaurus camacan is restricted to the a range contraction from times when the Atlantic forests of southern Bahia. Stenolepis Atlantic forest extended more to the west than ridleyi occurs in the Atlantic forest north of the present day distribution indicates. The

American Museum Novitates novi-496-00-01.3d 26/2/07 16:53:13 19 Cust # 496 20 AMERICAN MUSEUM NOVITATES NO. 00 sister relationship recovered between accelerated study is needed in order to un- Stenolepis ridleyi and Acratosaura mentalis derstand better the interesting biological confirms this pattern. Although we did not problems they present. have any previous idea of the relationships between entities grouped at node 7 (fig. 6B), ACKNOWLEDGMENTS in hindsight the recovery of Stenolepis and Acratosaura as sister species is not unexpected. We thank Fundac¸a˜o de Amparo a` Pesquisa This hypothesis is consistent with an emerging do Estado de Sa˜o Paulo (FAPESP), Conselho biogeographic pattern based on sister group Nacional de Desenvolvimento Cientı´fico e relationships between Atlantic forest species Tecnolo´gico (CNPq), Comissa˜o Executiva do and those occurring along the Serra do Plano da Lavoura Cacaueira (CEPLAC), Espinhac¸o in other taxa (see for example Universidade Estadual Santa Cruz (UESC), Parrini et al. 1999). Third, why are Iphisa the staff of Restauna project for support and elegans and Colobosaura modesta so wide- the following persons for their valuable help spread compared to their relatives? Is their in the lab or field, collecting specimens and broad distribution across Amazonia and the sending tissues: G. Skuk, V.X. Silva, M.J. Cerrados due, respectively, to recent origin Jesus Silva, D. Borges, M.N. Stevaux, A. and dispersal under favorable conditions? Langguth, C. Alonso, F. Curcio. Thanks are To answer this last question we should have also due to Luciana Lobo for the drawings more data on divergence times for the species and to Hussam Zaher, Carolina Castro-Mello, involved, particularly for those with currently and Gabriel Skuk for access to specimens. J. isolated populations. Calibration of diver- Sites acknowledges support of NSF award gence times could also help us to under- DEB 9815881 for some of the lab work. stand another intriguing feature of the Heterodactylini radiation. The monophyletic REFERENCES assemblage formed by Alexandresaurus, Iphisa, Colobosaura, Acratosaura, and Steno- Amaral, A. 1933. Estudos sobre lacertilios neotro´- lepis consists of monotypic genera, which is picos. Memo´rias do Instituto Butantan 7(1932): surprising when compared with other genera 1–74, 6 pls. Baker, R.H., and R. DeSalle. 1997. Multiple of Gymnophthalmidae occurring in the same sources of character information and the areas. Leposoma, for example, is a much more phylogeny of Hawaiian drosophilids. Syste- speciose genus, and morphological differences matic Biology 46: 654–673. between its species are slight, sometimes Boulenger, G.A. 1885. Catalogue of the lizards in inconspicuous, contrasting with the strong the British Museum (Natural History). Vols I– morphological gaps we observe among the III. London: Trustees of the British Museum. genera under study. In the latter, anteropos- Castoe, T.A., T.M. Doan, and C.L. Parkinson. terior axes of differentiation affecting size, 2004. Data partitions and complex models in shape, and ornamentation of body scales are Bayesian analysis: The phylogeny of gymn- nophthalmid lizards. Systematic Biology 53: clearly observable. This suggests that devel- 448–469. opmental processes might be differentially Doan, T.M. 2003. A new phylogenetic classification associated with or contributing to speciation for the gymnophthalmid genera Cercosaura, among these lineages. Pantodactylus, and Prionodactylus (Reptilia: Since the submission of this paper two Squamata). Zoological Journal Linnean highly distinctive new Heterodactylini were Society 137: 101–115. discovered from widely isolated mountains of Doan, T.M., and T.A. Castoe. 2005. Phylogenetic eastern Brazil. Our initial observations of their taxonomy of the Cercosaurini (Squamata: tentative phylogenetic positions suggest that Gymnophthalmidae), with new genera for species of Neusticurus and Proctoporus. they support the primary split of this in- Zoological Journal Linnean Society 143: teresting radiation, and the biogeographic 405–416. scenario outlined above. Considering the rate Estes, R. 1983. Fossil lizards of the world. of habitat loss in Brazil and the high Handbuch der Pala¨oherpetologie, vol. 10. sensitivity of these lizards to deforestation, Stuttgart: Gustoff Fisher Verlag.

American Museum Novitates novi-496-00-01.3d 26/2/07 16:53:13 20 Cust # 496 2007 RODRIGUES ET AL.: MICROTEIID LIZARD 21

Estes, R., K. de Queiroz, and J. Gauthier. 1988. phthalmidae), and a new classification for the Phylogenetic relationships within Squamata. In family. Biological Journal of the Linnean R. Estes and G.K. Pregill (editors), Phylo- Society 74: 317–340. genetic relationships of the lizard families, Posada, D., and K.A. Crandall. 1998. Modeltest: 119–281. Stanford: Stanford University Press. testing the model of DNA substitution. Farias, M.F. 2001. A imagem u´til. Edival. Lisboa: Bioinformatics 14: 817–818. Universidade Autoˆnoma Editora. Presch, W. 1978. Descriptions of the hemipenian Felsenstein, J. 1985. Confidence limits on phylog- morphology in eight species of microteiid enies: an approach using the bootstrap. lizards (Teiidae: Gymnophthalminae). Her- Evolution 39: 783–791. petologica 34: 108–112. Gatesy, J., and P. Arctander. 2000. Hidden Presch, W. 1980. Evolutionary history of the South morphological support for the phylogenetic American microteiid lizards (Teiidae: Gymno- placement of Pseudoryx nghetinhensis with phthalminae). Copeia 1980: 36–56. bovine bovids: a combined analysis of gross Presch, W. 1983. The lizard family Teiidae: is it anatomical evidence and DNA sequences from a monophyletic group? Zoological Journal of five genes. Systematic Biology 49: 515–538. the Linnean Society, 77: 189–197. Harris, D.M. 1985. Infralingual plicae: support for Presch, W. 1988. Phylogenetic relationships of the Boulenger’s Teiidae (Sauria). Copeia 3: Scincomorpha. In R. Estes and G. Pregill 560–565. (editors), Phylogenetic relationships of the Hillis, D.M., and J.J. Bull. 1993. An empirical test lizard families: 471–492. Stanford: Stanford of bootstrapping as a method for assessing University Press. confidence in phylogenetic analysis. Systematic Rodrigues, M.T. 1991a. Herpetofauna das dunas Biology 42: 182–192. interiores do Rio Sa˜o Francisco: Bahia: Brasil. Hoyos, J.M. 1998. A reappraisal of the phylogeny I. Introduc¸a˜o a`a`rea e descric¸a˜o de um novo of lizards of the family Gymnophthalmidae geˆnero de microteiı´deos (Calyptommatus) com (Sauria, Scincomorpha). Revista Espan˜ola de notas sobre sua ecologia, distribuic¸a˜o e espe- Herpetologia 12: 27–43. ciac¸a˜o (Sauria, Teiidae). Pape´is Avulsos de Huelsenbeck, J.P., and F. Ronquist. 2001. Zoologia, Sa˜o Paulo 37(19): 285–320. MrBayes: Bayesian inference of phylogeny. Rodrigues, M.T. 1991b. Herpetofauna das dunas Bioinformatics 17: 754–755. interiores do Rio Sa˜o Francisco: Bahia, Brasil. Kok, P.J.R. 2005. A new genus and species of II. Psilophthalmus: um novo geˆnero de Gymnophthalmid lizard (Squamata: Gymno- microteiı´deos sem pa´lpebra (Sauria, Teiidae). phthalmidae) from Kaieteur National Park, Pape´is Avulsos de Zoologia, Sa˜o Paulo 37: Guyana. Bulletin l’Institut Royal des Sciences 321–327. Naturelles de Belgique, Biologie: 35–45. Rodrigues, M.T. 1991c. Herpetofauna das dunas MacLean, W.P. 1974. Feeding and locomotor interiores do Rio Sa˜o Francisco: Bahia: Brasil. mechanisms of teiid lizards: functional mor- III. Procellosaurinus: um novo geˆnero de phology and evolution. Pape´is Avulsos de microteiı´deos sem pa´lpebra, com a redefinic¸a˜o Zoologia, Sa˜o Paulo 27(15): 173–213. do geˆnero Gymnophthalmus (Sauria, Teiidae). Myers, C.M., and M.A. Donnelly. 2001. Pape´is Avulsos de Zoologia, Sa˜o Paulo 27: Herpetofauna of the Yutaje´-Corocoro massif, 329–342. Venezuela: second report from Robert G. Rodrigues, M.T., and D.M. Borges. 1997. A new Goelet American Museum—Terramar expedi- species of Leposoma (Sauria, Gymno- tion to northwestern tepuis. Bulletin of the phthalmidae) from a relictual forest of semiarid American Museum of Natural History 261: northeastern Brazil. Herpetologica 53: 1–6. 1–85. Rodrigues, M.T., M. Dixo, and G.M. Accacio. Myers, N., R.A. Mittermeier, C.G. Mittermeier, 2002a. A large sample of Leposoma (Squamata, G.A.B. da Fonseca, and J. Kent. 2000. Gymnophthalmidae) from the Atlantic Forests Biodiversity hotspots for conservation priori- of Bahia, the status of Leposoma annectans ties. Nature 403: 853–858. Ruibal, 1952, and notes on conservation. Parrini, R., M.A. Raposo, J.F. Pacheco, A.M.P. Pape´is Avulsos de Zoologia, Sa˜o Paulo: 42: Carvalha˜es, T.A. MeloJr., S.M. Fonseca, and 103–117. J.C. Minns. 1999. Birds of the Chapada Rodrigues, M.T., M. Dixo, D. Pavan, and V.K. Diamantina, Bahia, Brazil. Cotinga 11: 86–95. Verdade. 2002b. A new species of Leposoma Pellegrino, K.C.M., M.T. Rodrigues, Y. Yonenaga- (Squamata, Gymnophthalmidae) from the rem- Yassuda, and J.W. SitesJr.. 2001. A molecular nant Atlantic forests of the state of Bahia, perspective on the evolution of South Amer- Brazil. Pape´is Avulsos de Zoologia, Sa˜o Paulo: ican microteiid lizards (Squamata, Gymno- 42: 335–350.

American Museum Novitates novi-496-00-01.3d 26/2/07 16:53:13 21 Cust # 496 22 AMERICAN MUSEUM NOVITATES NO. 00

Rodrigues, M.T., M.E.X. Freire, K.C.M. Pel- Vertebrate paleontology in the neotropics. The legrino, and J.W. Sites, Jr. 2005. Phylogenetic Miocene fauna of La Venta, Colombia, relationships of a new genus and species of 100–112. Washington, DC: Smithsonian microteiid lizard from the Atlantic forest of Institution Press. north-eastern Brazil (Squamata, Gymno- Swofford, D. 2002. PAUP*: Phylogenetic analysis phthalmidae). Zoological Journal Linnean using parsimony (*and other methods), Beta Society 144: 543–557. Version 4.0.b10. Sunderland, MA: Sinauer. Ruibal 1952. Revisionary studies of some South Vanzolini, P.E., and A.M.M. Ramos. 1977. A new American Teiidae. Bulletin of the Museum of species of Colobodactylus, with notes on the Comparative Zoology 106(11): 477–529. distribution of a group of stranded microteiid Sorenson, M.D. 1999. TreeRot, version 2. Compu- lizards (Sauria, Gymnophthalmidae). Pape´is ter software and documentation. Boston: Avulsos de Zoologia, Sa˜o Paulo: 31: 19–47. Boston University. Werner, F. 1910. U¨ ber neue oder seltene reptilien StatSoft 1998. STATISTICA for Windows, Release des naturhistorischen museums in Hamburg II. 5.1 J. Tulsa, Oklahoma: StatSoft, Inc. Eidechsen. Mitteilungen aus dem Natur- Sullivan, R.M., and R. Estes. 1997. A reassessment historischen Museum Hamburg 27: 1–46. of the fossil Tupinambinae. In R.F. Kay, R.H. Zar, J.H. 1996. Biostatistical analysis. Old Tappan, Madden, R.L. Cifelli and J.J. Flynn (editors), NJ: Prentice Hall.

American Museum Novitates novi-496-00-01.3d 26/2/07 16:53:13 22 Cust # 496 2007 RODRIGUES ET AL.: MICROTEIID LIZARD 23

APPENDIX 1

MORPHOLOGICAL CHARACTERS AND CHARACTER STATES USED FOR PHYLOGENETIC ANALYSES External morphology and scalation: extremely long and straight; (3) cruciform, central area reduced, lateral processes poster- 1. External ear opening: (0) present; (1) absent iorly oriented; (4) a longitudinal rod-shaped 2. Ornamentation of head scales: (0) smooth; (1) element, lateral processes absent; (5) trans- rugose versal, clavicular, and sternal processes absent. 3. Posterior margin of head scutes: (0) curved; (1) 26. Sternal process of interclavicle: (0) long, reach- straight ing sternal fontanelle; (1) small, not reaching 4. Prefrontal scales: (0) present; (1) absent fontanelle. 5. Frontoparietal scales: (0) present; (1) absent 27. Sternal fontanelle process: (0) absent; (1) 6. Superciliary scales: (0) four or more; (1) three or present less 28. Number of sternal ribs: (0) three; (1) two; (2) 7. Number of temporal scales: (0) six or more; (1) one four 29. Glossohyal: (0) separated from basihyal; (1) 8. Pairs of enlarged genials: (0) three; (1) two; (2) fused one 9. Size of interparietal scale: (0) reaching anteriorly 30. Second pair of ceratobranchials: (0) present; (1) the level of the parietal; (1) shorter absent 10. Anterior dorsal scales: (0) keeled; (1) smooth 31. Supratemporal fenestra: (0) almost closed; (1) 11. Posterior dorsal scales: (0) keeled; (1) smooth opened; (2) opened only posteriorly 12. Shape of dorsal scales at midbody: (0) lanceo- 32. Postorbital: (0) distinct; (1) fused to postfrontal late; (1) quadrangular or slightly mucronate; (2) 33. Postorbital width: (0) narrow; (1) wide cycloid 34. Nasals: (0) wide, almost parallel, in broad 13. Ventral scales: (0) strongly imbricate; (1) contact under and with premaxillary; (1) wide, juxtaposed divergent and in contact at midline but broadly 14. Flank scales: (0) lanceolate, imbricate; (1) separated anteriorly by the subtriangular lam- rectangular, smooth, imbricate; (2) quadrangu- ina of premaxillary, in slight contact with lar, juxtaposed; (3) cycloid premaxila; (2) separated by contact between 15. Collar: (0) absent; (1) present frontal and premaxillary. 16. Limbs: (0) normal, slender; (1) stout, compact; 35. Supratemporal fenestra: (0) closed by parietal (2) reduced and postorbital; (1) opened 17. Body form: (0) normal, 29 presacral vertebrae 36. Clavicle: (0) wide, flattened anteriorly and (pv.) or less; (1) elongate, more than 29 pv. enclosing a single fenestra; (1) simple, boomer- 18. Eyelid: (0) present; (1) absent ang shaped, fenestra absent; (2) axe shaped 19. Disc of lower eyelid: (0) divided; (1) single anteriorly, fenestra absent 20. Position of nostril: (0) in nasal border; (1) in 37. Postfrontal: (0) irregular, posteriorly wider center of nasal and longer leaving opened the supratemporal 21. Number of toes: (0) five, without reduction; (1) fenestra; (1) triangular; (2) boomerang- less than five shaped 22. Interparietal scale: (0) present, large; (1) vesti- 38. Supratemporal: (0) angulose, extremely curved gial or absent at the end; (1) slightly curved 23. Parietal scale: (0) longer than wide; (1) short, as 39. Lateral expansions of parietal: (0) present; (1) wide as long absent. 24. Enlarged occipitals: (0) absent, (1) present 40. Postorbital: (0) covers postfrontal; (1) covered by postfrontal; (2) contacts postfrontal without Osteology: overlap 41. Premaxillary dorsal lamina: (0) wide, poster- 25. Interclavicle: (0) cross-shaped, with central area iorly triangular; (1) wide, posteriorly straight; extremely enlarged, lateral processes large but (2) sub-triangular with end pointed toward but pointed; (1) same, but lateral processes short, not reaching frontal; (3) large, subrectangular, ending abruptly, not pointed; (2) cruciform, contacting frontal. central area reduced and lateral processes 42. Frontal: (0) normal; (1) anteriorly divided.

American Museum Novitates novi-496-00-01.3d 26/2/07 16:53:14 23 Cust # 496 24 AMERICAN MUSEUM NOVITATES NO. 00 HARACTERS C APPENDIX 2 ORPHOLOGICAL M ATRIX OF M (Characters 1–42 are defined in appendix 1). 456789101112131415161718192021 0 0 0001010100010000100 0 0 0101010100000010101 0 00 0001000100010000100 01 0001010100000000100 0 011101?100000010101 0 00 0001020112030000100 0 0101010100010000100 1 00 0111101111121110100 01 0000000000001000010 00 0111011100100210101 00 0000000001000000010 0 0001000111121110100 0 0 0000000000000000010 0 10 0000000000000000010 0 0101000112030001–11 0 0 0001000112030001–11 1 0 0010000111120010111 0 1 1000000000000000010 0 1 1000000000000000010 taunayi imbricatus cearensis gaudichaudii maximiliani tetradactylus brachylepis atriventris carinicaudatus camacan Species/Characters 1 2 3 Colobodactylus Colobosaura mentalis Colobosaura modesta Heterodactylus Iphisa elegans Stenolepis ridleyi Selected taxa Anotosaura vanzolinia Arthrosaura reticulata Bachia bresslaui Cercosaura ocellata Colobosauroides Ecpleopus Leposoma osvaldoi Micrablepharus Procellosaurinus Rhachisaurus Outgroup Alopoglossus Alopoglossus Ingroup Alexandresaurus

American Museum Novitates novi-496-00-01.3d 26/2/07 16:53:14 24 Cust # 496 2007 RODRIGUES ET AL.: MICROTEIID LIZARD 25 Continued. APPENDIX 2 0 0 0211010001100100000 0 1 1211010100110001000 0 00 1211010001100100001 01 1211010001100100000 1 1211010100110001000 0 00 1211010001100100000 0 1211010001100100000 0 00 0400101211000?00?10 00 0100001001000100100 00 0000200001002200100 00 0000000100010011120 0 0400000211000?00?10 0 0 0100001001000100110 0 00 0100001001000110110 0 0311000100211100230 0 0 0311000100211200230 0 0 0311000100012111020 0 0 0000000011200?10?20 0 0 0000000011200?10?20 taunayi imbricatus cearensis gaudichaudii maximiliani tetradactylus brachylepis atriventris carinicaudatus camacan Colobodactylus Colobosaura mentalis Colobosaura modesta Heterodactylus Iphisa elegans Stenolepis ridleyi Selected taxa Anotosaura vanzolinia Arthrosaura reticulata Bachia bresslaui Cercosaura ocellata Colobosauroides Ecpleopus Leposoma osvaldoi Micrablepharus Procellosaurinus Rhachisaurus Outgroup Alopoglossus Alopoglossus Species/charactersIngroup Alexandresaurus 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42

American Museum Novitates novi-496-00-01.3d 26/2/07 16:54:25 25 Cust # 496 26 AMERICAN MUSEUM NOVITATES NO. 00

APPENDIX 3

ADDITIONAL SPECIMENS EXAMINED Alcohol. Alopoglossus atriventris: MZUSP 13922–24. Cleared and stained. Alopoglossus atriventris: Estiron, Rio Ampiyacu, Peru. Alopoglossus carini- MZUSP 13922, Peru, Loreto, Rio Ampiyacu. caudatus: MZUSP 53683–84. Cachoeira do Lima˜o, Alopoglossus carinicaudatus: MZUSP 67488, Rio Tapajo´s, Para´. Anotosaura vanzolinia: MZUSP 79544, no further data. Anotosaura vanzolinia: 47754, Serra Negra, Inaja´, Pernambuco; MZUSP MZUSP 60772, 93417, 93418, Cabaceiras, 60773, 60775–77, 60780–85, Cabaceiras, Paraiba; Paraı´ba. Arthrosaura kocki: MZUSP 82951 Vila MZUSP 80151, Xingo´, Alagoas. Arthrosaura reticu- Rica, Mato Grosso. Arthrosaura reticulata: lata: MZUSP 82644–46 Aripuana˜, Mato Grosso. MZUSP 52504, no further data. Bachia bresslaui: Bachia bresslaui: MZUSP 10300. Utiaritı´, Mato MZUSP 93419, Brası´lia, Distrito Federal. Grosso. Cercosaura ocellata: MZUSP 82425–27 Cercosaura ocellata: MZUSP 79556, Chapada dos Juruena, Mato Grosso. Colobodactylus taunayi: Guimara˜es, Mato Grosso. Colobodactylus taunayi: MZUSP 91446–48 Pinhala˜o, Parana´. Colobosaura MZUSP 94254, Saleso´polis, Sa˜o Paulo. mentalis: MZUSP 56933, Gra˜o Mogol, Minas Colobosaura mentalis: MZUSP 79557, Maraca´s Gerais. Colobosaura modesta: MZUSP 81346–62, Bahia; MZUSP 66231, Cabaceiras, Paraı´ba; Caldas Novas, Goia´s. Colobosauroides carvalhoi: MNRJ 9968, no further data. Colobosaura modesta: MZUSP 89453, Parque Nacional da Serra da MZUSP 93420, Serra da Mesa, Goia´s. Capivara, Piauı´. Colobosauroides cearensis: 87580– Colobosauroides cearensis: MZUSP 79595, 93421, 83, Mulungu´, Ceara´. Ecpleopus gaudichaudii: Serra de Baturite´, Ceara´. Ecpleopus gaudichaudii: MZUSP 78975, 79559, Caraguatatuba, and MZUSP 79558–59, 93424, 93425, no further data. MZUSP 81401, Ariri, Sa˜o Paulo. Heterodactylus Heterodactylus imbricatus: MZUSP 79562, no imbricatus: MZUSP 89185–86, Parque Estadual da further data; MZUSP 94255, Caucaia, Ibiu´na, Sa˜o Cantareira (Nu´cleo Pedra Grande), Sa˜o Paulo; Paulo. Iphisa elegans: MZUSP 82654, 82658, MZUSP 3309 Santa Ba´rbara, Serra do Carac¸a, Aripuana˜, Mato Grosso. Leposoma osvaldoi: Minas Gerais. Iphisa elegans: MZUSP 82654–76, MZUSP 93426, UHE Samuel, Rondoˆnia. Aripuana˜, Mato Grosso. Leposoma osvaldoi: Leposoma scincoides: MZUSP 79564-65, no further MZUSP 8270–37, Aripuana˜, Mato Grosso. data. Micrablepharus maximiliani: MZUSP 66233, Micrablepharus maximiliani: MZUSP 50172–74, Joa˜o Pessoa, Paraı´ba; MZUSP 79589-91, no further Exu´, Pernambuco. Procellosaurinus tetradactylus: data. Procellosaurinus erythrocercus: MZUSP MZUSP 71598–604, Alagoado, Bahia. Rha- 74935, 74942, Ibiraba, Bahia. Rhachisaurus brachy- chisaurus brachylepis: MZUSP 54897, 54898, Serra lepis: MZUSP 55533, Serra do Cipo´, Minas Gerais. do Cipo´, Minas Gerais. Stenolepis ridleyi: MZUSP Stenolepis ridleyi: MZUSP 66049, Serra dos 78724, Mulungu´ (Sitio Lorena), Ceara´. Cavalos, Pernambuco.

American Museum Novitates novi-496-00-01.3d 26/2/07 16:53:19 26 Cust # 496 2007 RODRIGUES ET AL.: MICROTEIID LIZARD 27

APPENDIX 4

MEASURES OF SUPPORT FOR ALL INTERNAL NODES OF MOST PARSIMONIOUS TREE RECOVERED FROM COMBINED ANALYSIS OF MORPHOLOGY (MORPH) AND MOLECULAR PARTITIONS (fig. 6) Columns present the bootstraps proportions, total combined analysis, respectively. Zero score indicates and partitioned Bremer values; positive and negative the indifference of a given data set at a specific node partitioned values indicate support for a given re- (Baker and DeSalle, 1997; Gatesy and Arctander, lationship in the combined analysis over the alterna- 2000). Node # 5, highlighted in bold, present the tive relationship in separate analyses, and contradic- support for the relationship between Alexandresaurus tory evidence for a particular relationship in the camacan and its closest relatives (see fig. 6).

Partitioned Bremer Bootstrap Node # support Bremer support 12S 16S ND4 18S c–mos morph

1 , 50 1.0 1.5 0.5 25.0 22.0 1.0 5.0 2 , 50 6.0 20.8 4.2 2.2 21.5 1.8 0.0 3 99 13 1.3 4.0 20.3 0.0 3.0 5.0 4 , 50 1.0 3.0 20.5 23.5 21.0 0.5 2.5 5 , 50 2.0 2.0 0.5 22.5 21.0 0.5 2.5 6 100 18 7.0 1.0 6.0 1.0 3.0 0.0 7 75 5.0 23.0 21.0 3.0 0.0 3.0 3.0 8 92 11 1.0 0.5 5.0 21.0 21.0 6.5 9 100 28 2.5 3.0 6.5 22.0 12 6.0 10 96 11 1.0 5.3 25.3 21.7 7.7 4.0 11 50 6.0 0.0 2.5 1.0 20.5 1.5 1.5 12 100 32 7.0 4.0 4.0 22.0 1.0 18 13 62 8.0 28.0 21.0 14.8 21.5 0.8 3.0 14 64 8.0 28.0 21.0 14.8 21.5 0.8 3.0 15 92 11 26.0 0.5 13 21.0 2.0 2.5 16 100 73 0.0 16 20 2.0 28 7.0

American Museum Novitates novi-496-00-01.3d 26/2/07 16:53:19 27 Cust # 496