Molecular Phylogenetics and Conservation of Tupinambis (Sauna: Teiidae)

Gopeia,1999(4), pp. 894-905

LEE A. FITZGERALD,JOSEPH A. COOK, AND A. Luz AQuINO " At least five species of Tupinambislizards (Sauna: Teiidae) occur in South America east of the Andes. In Argentina and Paraguay, two species of Tupinambishave been of commercially exploited for the leather trade since the 1970s. Current quotas total 1.35 million skins/yr. Despite tbjs large and sustained harvest, few studies have examined genetic or morphological variation in these lizards. We sequenced parts of the mitochondrial cytochrome b (300bp) and ND4 (375bp) genes and examined tbjs variation in light of morphological characters traditionally used to identify these species. DNA sequences provided a preliminary view of intraspecific and interspecific variation and were used to explore evolutionary relatiouships among 17 individuals representing7: menanae,7: rufescens,and 7: duseni from Paraguay,and 7: longilineusand 7: teguixin from Roraima, Brazil, and Cuyabeno, Ecuador. Kentropyx viridistriga, Ameiva ameiva, and Cnemidophorusocellifer were included as outgroups. Maxjmum-parsimony and neighbor-joining analyses revealed two distinctive groups within Tupinambil; a northern South American and Amazonian clade (7: teguixin, 7: longilineus) and a clade (7: duseni, 7: rufescens,and 7: menanae) that is distributed primarily south of Amazonia. Although genetically similar and previously cousidered synonymous with 7: rufescens,7: duseni is morphologically distinct based on squa- mation, coloration, and morphometrics. This incongruence between the molecular data and morphologysuggests that 7: duseni and 7: rufescensmay have undergone extensive and recent morphological evolution or there has been introgression of mitochondrial DNA between these species. Sequence divergence between 7: teguixin from Brazil and Ecuador was similar to that found between 7: rufescensand 7: merianae and may indicate these 7: teguixin populations are not conspecific. The 7: teguixin clade was sister to 7: longilineus. These findings, combined with the large- scale commercial exploitation of the genus, suggest an urgent need to address geographic variation and the systematics of species of Tupinambis.

A T least five species of tegu lizards of the ge- quotas total 1,350,000 skins/year (Fitzgerald et nus Tupinambis (Sauria: Teiidae) occur in al., 1991, 1994). Thousands of campesinos and South America east of the Andes (Cei, 1993; indigenous people hunt tegus for income and Avila-Pires, 1995; Colli et al., 1998). Tegus are food, and the export value of the processed the largest members of the Teiidae and occur skins is important to the national economies over a broad geographic range extending from where the trade is legal (Fitzgerald, 1994). An Colombia and Venezuela throughout Amazonia unquantified number of Tupinambisskins enter and reaching as far south as northeastern Pa- the international trade from countries where tagonia (Cei and Scolaro, 1982; Peters and Don- harvesting and exportation are currently illegal. oso-Barros, 1986). Considerable morphological A variable number of tegus enter the live pet variation exists within the genus, and although trade from throughout the distribution of the two or three species may be sympatric, these genus. generalist omnivores appear to segregate eco- Despite the economic and social importance logically by habitat type (Presch, 1973; Norman, of tegus as a resource, and the high rate of har- 1986; Fitzgerald et al., 1991). vest of these lizards since 1980, few studies have . ' Tupinambis teguixin and 7: rufescensare listed examined genetic and morphological variation i in Appendix II of CITES and species of this ge- among the species (Avila-Pires, 1995; Colli et al., .1 nus are among the most heavily exploited rep- 1998), and none has included the exploited 1 tiles in the world (Luxmoore et al., 1988; Fitz- populations in the Southem Cone of South gerald, 1994). An average of 1.9 million skins America. The taxonomy of the genus remains annually were traded in the 1980s, mostly to unresolved, causing confusion for management cowboy boot manufacturers in the United agencies charged with monitoring the interna- States. In Argentina and Paraguay, where the tional trade in tegu skins. It is likely that species trade is legal and monitored, current export of TuPinambisremain undescribed (Avila-Pires,

C 1999by the AmericanSociety of Ichthyologistsand Herpetologists

, .

j (1.; . r ,\ "] FITZGERALD ET AL.-MOLECUlAR PHYLOGENYOF TUPINAMBIS 895

1995), for example, and the lack of consistent so 40 taxonomy and genetic and morphological markers useful for forensics pose impediments to management and long-term conservation. , Characterization of geographic variation and species limits is necessary to quantify the trade effectively, enforce harvest quotas, and monitor " trends in the take by geographic area. The taxonomic history of the genus Tupinam- his Daudin 1802, and particularly that of 7: teguixin (Linnaeus 1758) and 7: duseniLonnberg in Lonnberg and Andersson 1910, is tortuous. Four species were recognized by Peters and Donoso-Barros (1970): 7: teguixin, 7: nigropunc- talus Spix 1825, 7: rufescensGunther 1871, and 7: duseni.Presch (1973) reviewed the genus and recognized only two species: 7: rufescens,the red tegu, which included the single known specimen of 7: duseni (following Burt and Burt, 1931), and 7: teguixin, comprised of all other tegus, including 7: nigropunctatusfrom northern populations and 7: teguixin (sensu Boulenger, 1885) from the Southern Cone region of South America. Presch's (1973) arrangement for 7: teguixin, however, was not followed by su~sequent Fig. 1. Collecting localities of Tupinambisused in authors (e.g., Hoogmoed, 1973; CeI, 1986, this study.The map of Paraguayshows the distribu- 1993). In her recent review of available names tion of speciesof TuPinambisbased on holdingsin the and type material for "teguixin" forms, Avila-Pi- Museo Nacional de Historia Nacional de Paraguay res (1995) clarified the situation by renaming 7: (MNHNP). Symbolsare as follows: Closedcircles = nigropundatus as 7: teguixin and renaming 7: te- 7: merianae;.opensquares = 7: rujescens;,o.pen triangle.s guixin (sensu Boulenger, 1885) as 7: menanae. = 7: dusenz;cl~s~d squares = 7: teguzXln;closed trI- Hence, northern South American and Amazo- angle = 7: longlhneus. nian populations are 7: teguixin, whereas the larger southern form that is exploited for skins vine mitochondrial genome; Anderson et al., is 7: merianae. 1982) and ND4 (375 bp, positions 11204-11578 Two additional species of TuPinamhishave re- of the bovine mitochondrial genome; Anderson cently been described. Avila-Pires (1995) de- et al., 1982) sequences were obtained from the scribed 7: longilineusfrom Rondonia, Brazil, and mitochondria of 17 individuals representing 7: Manzani and Abe (1997). descri?ed 7: quadrili- merianae,7: rufescens,7: duseni, 7: longilineus,and neat~s from central Brazil. C.<>lliet al: \1998) 7: teguixin and from one individual each of Ken- proVIded a more ~ompre?~nSIVedescnptIon of tropyx viridistriga, Ameiva ameiva, and Cnemido- the ~ame taxon, WIth additIonal data..Although phorus ocellifer.The 7: longilineusand 7: teguixin Colli et al. (1998) compared five specIesof Tup- were from Brazil and Ecuador and the remain- inamhis usi!lg morphological and allozyrne elec- der of the material was from Paraguay (Fig. 1). trophoretic data, reconstruction of evolutionary DNA was extracted from dried skin samples rela~o.nships amoni? TuPinambis species using or from tissues frozen or preserved in alcohol. explIcIt phyloge?etIc meth.ods h~ not been The sequence data for these individuals have completed. Herem, we proVIde an mdependent been submitted to the GenBank Data Libraries view of phylogenetic relations and species limits (accessionsAF151174-AF151213). . among five nominal species of TuPinamhisbased on sequences of the mitochondrial cytochro~e DNA extraction, amPlification, and sequencing.- b (cyt b) and ND4 i?enes.We relate ~e genetIc Tissues were subjected to proteinase K diges- da~ to ~orphologIcal features preVIously used tion, NaCI precipitation of proteins, and DNA to Identify these taxa. precipitation with ethanol, following a protocol modified from Miller et al. (1988). Partial seg- MATER1AL5AND METHODS ments of the cyt band ND4 genes were ampli- Specimensexamined.-Partial cyt b (300 bp, cor- fied by the polymerase .chain reactio.n (PCR; responding to positions 14618-14917 of the bo- usually 30 cycles alternatIng denaturatIon at 93

. , .

jr .~ , .

896 COPEIA, 1999, NO.4

C for 1 min, annealing at 45 C for 1 min, ex- outgroup criterion, with Ameiva,Cnemidophorus, tension at 72 C for 1.5 mill) using the following and Kentropyxdefined as outgroups, collectively. primer combinations: for cyt b, UZI (5'-3', To test for the monophyly of the genus Tupi- AGC CCC ATC CAA CAT CTC TGC 1TG ATG nambiswith respect to these three genera, sep- AAA)-UZ2 (5'-3', TGA CTG TGG CAC CTC arate MP analyses (equal weights) were con- AGA ATG ATA 1TT GGC CTC A» and for ducted with each genus defined as the sole out- ND4, ND4-LEU (Arevalo et al., 1994). UZl- group. LIZ2 were modified from the universal cyt b Distance-based NJ trees using Kimura 2-pa- primers of Kocher et al. (1989) based on the rameter distances were computed and assessed chicken sequence (Desjardins and Morais using PAUP*. MP and NJ trees were boot- 1990). All PCR experiments included negative strapped (500 replicates) to assessthe strength controls. of associations and all trees were rooted for pre- Aliquots (5 ILl) of the PCR products and neg- sentation. Decay indices (Bremer, 1988) were ative controls were visualizedin agarosemini- computed for the MP tree using TreeRot (So- gels. Each remaining product was precipitated renson, 1996). with polyethylene glycol, recovered by vacuum cenbifugation, and resuspended in IX TE buff- Morphologicalanalyses.-We measured specimens er. The clean products were used as templates of 7: merianae, 7: rufescens,and 7: duseni from in cycle sequencing utilizing a Perkin-Elmer kit Paraguay. We determined the sex of each spec- (Fst-RR, 402119). Cycle sequencing products imen and measured snout-vent length (SVL), were purified with Sephadex G-50 (Sigma) in head length, head width, tip of snout to eye, reusable columns (Princeton Separations) and length of the pes, and length of the fourth toe. run on 2% polyacrylamidegels using an auto- We counted the number of ventral scalerows, mated sequencer (Applied Biosystems, Inc., ventrals across midbody, subdigitallamellae on 373). Both heavy and light strands were se- the fourth toe, femoral and preanal pores, and quenced in all cases. labials. Descriptive statistics of scale counts were compared to values reported in the literature Sequenceand phylogenetic analysis.-Partial se- for species of Tupinambis (Presch, 1973; Avila- quences were examined, assembled based on Pires, 1995). Ten 7: duseniwere hatchling clutch overlapping regions and aligned by eye using mates. We tabulated the scale counts for these the Sequence Navigator program (Applied Bio- individuals but excluded them from statistical systems, Inc., verso 1.01). Kimura 2-parameter analyses. distance values, nucleotide composition and The individuals used in the study varied sub- numbers of variable and phylogenetically infor- stantially in body size (83-423 mm SVL). Prior mative sites for nucleotides were obtained using to analysis, individuals < 200 mm SVL were de- PAUP* (vers. 4.0bl, D. Swofford, 1998, un- leted from the dataset to reduce bias in statis- publ.). tical analyses.Analyses of covariance, using SVL Phylogenetic analysesused two primary meth- as the covariate, were performed on five mea- ods: maximum parsimony (MP) and neighbor surement variables ( loglo-transformed) to test joining (NJ). Weighted and unweighted MP for differences in the relative size of traits trees were generated using PAUP*. Weighting among the samples from Paraguay. In casesof schemes comprised equal weights or differential overall significance, contrasts were computed to weights for first, second, and third codon posi- test for differences in least-square means be- tions (weighted 2, 5, and 1, respectively) in com- tween species. bination with a stepmatrix awarding greater To remove the confounding effects of body weight to transversions than transitions (10:1). size for multivariate analyses,loglo-transformed The latter scheme attempts to accommodate observations were pooled, and each morpho- known biases in the evolution of cyt b and ND4 metric character was regressed against SVL. Re- genes (Irwin et al., 1991; Arevalo et al., 1994). siduals resulting from the regressions were re- We used the branch-and-bound search algo- tained as size-adjustedvariables and allowed sta- rithm for the weighting scheme (and equal tistical comparisons of morphological shape var- weights) in PAUP*. Because the topologies of iation among individuals of different size these trees were identical, we used only the (Miles, 1994; Losos et al., 1998). A canonical equal-weightdataset in subsequentanalyses. Val- discriminant analysis(PROC Candisc, SAS Insti- ues of skewness were computed from 10,000 tute, Cary, NC, 1989, unpubl.) was performed random, trees and used to assessoverall phylo- on the five measurement variables to evaluate genetic signal in the dataset (Hillis and Huel- morphological differences among species. Cor- senbeck, 1992). Trees were rooted using the relations between canonical variates axis scores

. ~ .

:~ ~~",/J:c ~~!ifC!"c:-:!(".,.;c FITZGERALD ET AL.-MOLECULAR PHYLOGENY OF TUPlNAMBIS 897

and original variables were used to assessthe dicated by the highly significant G1 value relative importance of each trait. Mahalanobis' (-0.4~50) obtained from 10,000 random trees distances (D2) were computed, and resulting P- (P < 0.01, table 2 of Hillis and Huelsenbeck, values allowed us to accept or reject the null 1992). Of the 675 characters, 205 were parsi- hypothesis of no significant morphometric var- moniously informative. Maximum-parsimony iation among the three Paraguayan taxa (Miles, analyses recovered a single most-parsimonious 1994). tree (Fig. 2A) when characters were unweighted c'". andThe the fivefive scalesize-adjusted characters body were measurementsused in a clus- permutations,(541 steps in length).the same Under most-parsimonious several weighting tree ter analysis to assessmorphological similarity was consistently recovered. One of the weight- among the species. Cowers' median method for ing schemes (transversion/transition weighted computing distances (a weighted pair-group 10/1 and first, second, and third position codon method using centroids) was used to minimize sites weighted 2/5/1, respectively) produced an effects of combining measurement and meristic additional shortest tree, but it differed only with data (SAS Institute, Cary, NC, 1989, unpubl.; D. respect to the relationships among the out- B. Miles, pers comm). group taxa. The neighbor-joining analysis produced a tree with identical topology (Fig. 2B). RESULTS These analyses suggest the following relationships: (1) species of Tupinambisform a mono- Compositionalbiases, and patterns and levelsof var- phyletic group with respect to the representa- iation.-AlI sequences obtained follow the pat- tives of Ameiva, Cnemidophorus,and Kentropyx terns of compositional bias common among ver- used in this study; (2) there are two primary tebrate mitochondrial cyt b (e.g., Irwin et al., and deeply divergent clades in Tupinambis (a 1991; Petren and Case, 1997) and ND4 (e.g., southern group consisting of 7: merianaeand 7: Benabib et al., 1997) genes. Average base fre- rufescens,and 7: duseni and a northern/ Amazo- quencies for the coding strand were: A (28.8% nian clade comprised of 7: teguixin and 7: lon- overall, 27.5% cyt b, 29.9% ND4); C (28.1%, gilineus); (3) within the southern clade, 7: meri- 27.8%,28.3%); G (15.3%, 15.0%, 15.5%); and anaeis sister to a clade composed of 7: rufescens T (27.8%, 29.7%, 26.3%). There is a deficiency and 7: duseni (which show very low levels of di- in guanines in the light strand of all individuals vergence); (4) within the northern clade, the (14.5-16.3% overall, 13.7-16.7% crt b, 14.4- level of divergence between 7: teguixinsamples 16.5% ND4). Observed substitutions were most from Roraima, Brazil, and Cuyabeno, Ecuador, abundant in third positions. and least common is comparable to levels of divergence found in second positions. Observed, uncorrected among other interspecific comparisons. transitions outnumber transversions in pairwise comparisons of haplotypes that exhibit relatively Mo1fIhology.-Morphological analysesquantified little divergence (e.g., 9:1 in crt band 13:1 in several consistent differences among the three ND4 for the comparison of two samples of 7: speciesof Tupinambisin Paraguay.There was sig- dusen,). These substitutions and biases are ex- nificant variation in snout-eye length (n = 48, pected for functional mitochondrial cyt b and F2.43= 5.224,P < 0.008), and contrastsshowed ND4 genes. We subsequently combined these 7: duseni had a relatively longer snout-eye data for all phylogenetic analyses. length than the other two species (P < 0.006). Distance values for pairwise comparisons As noted in Unnberg and Andersson (1910), ranged from 0.0 for two specimens of 7: duseni length of the fourth toe also varied statistically (tissue 766 and tissue 777) and two 7: teguixin among the species (n = 48, F2.43= 5.224,P < (LSUMNS 12431 and LSUMNS 12450) to 0.424 0.008). TuPinambisduseni had a relatively short- for comparisons between Kentropyxand 7: meri- er fourth toe than the other species (P < anae (tissue 779). Samples 777 and LSUMNS 0.005), whereas the fourth toe in 7: rufescenswas 12450 subsequently were excluded from further shorter than in 7: merianae(P < 0.02). Adjusted , phylogenetic analyses of the sequence data. mean head length, head width, and hind foot Largest intraspecific distances (0.099) were length did not differ among the species (n = found within 7: teguixin (LSUMNS 12405 and 48; df = 2,43; P> 0.13, P> 0.37, P> 0.95, LSUMNS 12715) and were nearly as large as the respectively). distance (0.109) between comparisons of 7: meT- The canonical discriminant analysis corrobo- ianae (tissue 649) and 7: rufescens(tissue 740). rated the univari~te comparisons. Canonical variate axis 1 was statistically significant (F1O.82= Relationshipsamong species ofTupinambis.-Strong 5.057; P < 0.001) and explained 92.1% of the phylogenetic signal in the sequence data was in- variance in the dataset. The relative length of

~ .

. jr '] - ,\ "') 898 COPEIA, 1999, NO.4

55/ 1 Ameiva 640 65

Cnemidophorus 784 88 / T. merianae 649 100 / 16 T. merianae 756 100 T. merianae 779 81 / T. merianae 780 4 100 100/22 T. rufescens 740 T. duseni 766 T. rufescens 743 100

T. duseni 778 100 99/15 T. teguixin 12405 100 T. teguixin 12431 76 T. teguixin 12678 100 100/16 100/12 T. teguixin 12703 100 86/2 T. teguixin 12715 T. fongifineus 14135 100 100/19 T. fongifineus 14136 Kentropyx 765 A B Fig. 2. (A) Results of unweighted maximum-parsimony analysis of tegu relationships with Kentropyx,Ameiva, and Cnemidophorusdesignated as outgroup taxa. This branch-and-bound analysis produced a single tree (length 541 steps, CI = 0.691, RI = 0.824). Percentage of 500 bootstrap pseudoreplicates are reported near each node followed by decay indices. (B) Neighborjoining tree of relationships between species of Tupinambisand three other genera of macroteiids using pairwise Kimura two-parameter distances and associated gamma distribu- tions. Numbers above branches represent percentage values from 500 bootstrap replicates. Numbers following the names identify tissue specimen vouchers and are listed in Materials Examined.

the fourth toe and snout-eye length were < 0.317). Canonical variate axis 2 represented strongly correlated with CV axis 1 (Table 1), a morphological gradient based on relative and 7: duseni formed a distinct group on the low head length and snout-eye length, but it did end of this axis (Fig. 3). Although CV axis 2 not separate the species well (Table 1). Mahal- explained an additional 7.9% of the variance, it anobis' distances with corresponding P-values was not statistically significant (F4.42= 1.217; P showed 7: duseni occupied distinct multivariate space from 7: rufescens and 7: merianae (P < 0.002 for both comparisons; Fig. 3). TuPinambis TABLE 1. SUMMARYOF A CANONICALDISCRIMINANT rufescens and 7: merianae did not occupy statisti- ANAL)SISBASED ON FivE SIZE-ADJUSTEDMORPHOLOGI- cally distinct morphological space in this analy- CAL VARIABLEs.Values shown are total-sample corre- sis. lations between the canonical variables and the orig- Scale characters were also informative for dis- inal variables. tinguishing the species. TuPinambis merianae had

Vanable' cvaxIS. 1 cv axIS. 2 a high number of femoral'. + precloacal . pores (20-25), a charactenstic not shared WIth Its con- Head length -0.077 0.840 geners (Table 2). The number of ventral scales Snout~ye length -0.519 0.614 in a single row across the body correlated well Head WIdth -0.295 0.116 with the northern and southern clades identi- Hind foot -0.033 0.194 fied in the phylogenetic analyses. Tupinambis Fourth toe 0.939 0.234. d . . 11 h d Elgenva I ue 1 ..342 0 116 menanae, 7: rufescens, an 7: dusenz a s are a h . h b f 1 .db d ( . P . I . d 921 ot 7 9ot Ig num er 0 ventra s across ml 0 y mm- ercent variance exp ame . 70 . 70 . imum ventrals in a row for these species = 34),

, .

:i! 9

, FITZGERALD ~T AL.-MOLECULAR PHYLOGENY OF TUPlNAMBIS 899

3 clade (1: merianaeand 1: rufescens)and the Am- . azonian forms is indicative of an old split within N 2 . . . . Tupinambis.The generality of this deep north- .~ south biogeographic split in other South Amer- ,- ~ 1 . .. ican taxa should be investigated further. Given - . .. &. '. .~ ..,-. . a the early cladogenesIs of these pnmary cIad es,

0 .

M,h,',nob;.. O;o"no» maW al dd . T_os TOMri,n..

T""'n! ('-00021828 "-00011 atIves 0 r-.~~gc-1..".'.... f ...... ciallythena, IS .inclusion Callopistes,CrUCI "" .upznam. oftootheraTeius,bis(P ressmg rescmacroteiidCrocodilurus, h1974SII ,th eIssue.; genera, andu 0Ivan.fDracae-mono-espe- and ~.1. .phyly.eses basedThese ongenera morphology.arethought Tupinambistobeclose longili-rel- 1343...neus. wasf the sIster taxon to 1: teguzxznfrom Ro- Fig.-2 Tro"""',3. A plot of the187. positions of individual Tupi- Estes,raima, 1997).Brazil, a finding consistent with the con- nambison CV axis 2 versus CV axis 1. The analysiswas clusions of Avila-Pires (1995) and Colli et al. ~onducte~-6 us~g-4 five ~e-adjusted...-2 morpholoi?ical02 v~- (1998).The geneticInterestingly, data agreethe splitwith amongprevious 1: hypoth-teguixin Iables .on mdIVIdualsCanonical 200 variatemm ~.axIs Can°n.Ical ~n- from Cuyabeno,. Ecuador and Roraima, .. Brazil, ate axIS 1 represented a morphologIcal gradIent With mdlVIdualshaVIng long fourth toe and short snout- . . para e to. . J.erences e een 'r .. ..,-'n eyelength on the high end, and short fourth toe and guzxzn and 1: longzlzneusan~ even to that be-

iate axis 2 representeda gradient of head length but analYSIs,Including samples of Tupznambisfrom did. ..not separate.the specieswell. Closed circles = 1: wasthroughoutcom blAmazoniadifr and northernbtw South merianae;closed triangles = 1: rnfescens;closed squares America, might reveal significant variation with- = 1: duseni. Mahalanobis'distances with correspond- in the currently defined taxon, 1: teguixin. longing P-valuessnout-eyeare length shownonin the lowinset.end. Canonicalvar-tween.In spite1:.rufescer:s of morphologicaland1:menanae. similarities~ex~anded between 1: teguixin and 1: merianae,the latter taxon is apparently the sister group to 1: rufescensand 1: whereas 1: teguixin, 1: longilineus,and 1: quadri- duseni,and part of the distinctive clade in south- lineatus have a much smaller number (maxi- ern South America. We hypothesize that, during mum ventrals in a row for these species = 28). or after the north-south split of the Tupinambis The cluster analysis using five size-adjusted clade, 1: rufescensderived from a 1: merianae-like morphological variables and the five scale ancestor. TuPinambisrufescens is restricted to the counts from Paraguayan material produced a dry chaco of Bolivia, Paraguay, and Argentina. dendrogram that described well the southern Hence, the morphological and ecological spe- clade of TuPinambis(Fig. 4). Morphological dis- cialization to the xeric chaco environment ap- tances among the 1: merianaewere generally pears to be a derived condition in this genus. mostly small. One individual each of 1: rufescens Further work is needed to resolve the rela- and 1: duseniclustered with five 1: merianaethat tionship between 1: rufescensand 1: duseni.Burt were relatively distant from their conspecifics. and Burt (1931) considered the single speci- These individuals clustered together because men of 1: duseni a geographical variant of 1: the five 1: merianaepossessed a relatively high rufescens,distinguishable only by relative sizesof number of ventral scales across midbody, and limb elements. They apparently did not place the 1: rufescensand 1: duseni each had a rela- importance on the striking differences in col- tively high number of femoral pores. The re- oration and scalation of the type specimen of 1: maining three 1: duseni clustered together, and duseni. Peters and Donoso-Barros (1970) did this group was paired with the remaining seven recognize 1: duseni as distinct, but Presch 1: rufescens. (1973) followed Burt and Burt (1931), retaining 1: duseniwithin 1: rufescens.He did not examine DISCUSSION the type specimen of 1: duseni, however. Al- though Avila-Pires (1995) also recognized 1: du- These mitochondrial data support some pre- seni, Colli et al. (1998) preferred to retain 1: vious hypotheses of evolutionary relationships, duseni as a junior synonym of 1: rufescensuntil provide evidence for unsuspected relationships further information is available. and suggest substantial and previously unde- TuPinambisduseni is yellowish in color and, in tected geographic variation in these large liz- addition to quantified differences in head shape ards. Deep divergence between the southern and fourth toe, it is also distinct from 1: rufescens

;lr 900 COPEIA, 1999, NO.4

? os ~ ~ ~ 00 0 ~1"- 8 '" - - G'I - ~ .0 I I I - r ~ ~ ~ ~I~ I ~=~ C() 1"- 0 1"-~C() ~ ...'- ~ . ~ .. ~ os ~ "'" "" . os .~ ;0+ '-- '-- 8 ~'" -~ -~ -~ ~ -.. -~~~ - Z.. .g..j...j.8.- - -J, - --J, J, "' " ., .. .. ., ~ ~ - 0 00 i-' 1"- os ~ "" - - -- :au .S ~ ... ~ 1"- ~ ~ ~ u G'I - - 1"- ~~ - ~ I I I - - I 8 "U;r;- c:i ~ I )...~oo Z~ 5.~"" Oti .t) ~ ~ 0; ~ 'i '" G'I . v~ ~at) . + 8. ~ -~ -~ C() .. - ~ I I I ~_ .6 ~ ~ 0 G'I "" -~ - I.' - :;. 0 G'I -.,- 6 -t=.:"" E-' s , - .'- -.: 0 ,~ 1"- ~ ...,,- "" ~ - - - :s

~ G'I ~ 00 C() toO~ S.$~ ,,' C()I G'II G'II ~ C()I - ~~~C() ~ ~ I" - - 1"- I ~ 1"- - z ,,?j~r-: r-: Oti 0" "i' 00 "'i' -.: :E~ - G'I G'I G'I "'" 00 G'I 1"- >- .~ S ~~ 'I 'I 'I C() G'I 'I G'I ~ ..~.."", ~ ~ .,~., i:: ~ ':8 G'I G'I -G'I c:I, - G'I "" .A =" , G'I ., ~. .5 '" - 0 ~ 1"- '" "" - - - ~ z' ~ ~ .~ .?j ~ ~ ~ 00 ~ ,g" '? ,. 00 "i' ~ .- ~ ~ ~ G'li ~ ~ . S 1"- - - ""~ .~ ~ ~ ~ IC 0 00 ~ ~ "i' ri; _ 8 0 "" "" "" . v~ G'I"" ~ E;:;~ I ~'"~~~"" I I I G'I ;.: ~_.I v, -~ 1"- G'I "" ~ 0 ., , ~ ] ~ ~. ~ ~ 0:"."" ~ >i =- 0 00 1"- -.: o!! "" :9 :;"" - - - 0 " ~ ~ ] ~ ~~"" 1"- ~ 0 ~"'i' "'i' "'i' 00 "'i' .:s ~ 0~~~C() ~ toO '"G'I""~ I ""~ ~ ~ ~Oti"",~~~"","'i' ~ ~ 01 C() C() C() .t) u', C() at) ~ :;. ]rXJ~;.:x~] ~ = C() C() C() a!s C() "" = .." ~ .. ., .. C() .. 0 "" ""OO~ - - 1"-- "0" ~ .?i-§" 0 u _u ~ ~] . ~.2;-;:: "O£~ U .;;-~~~ ... ~ 8.Coj, 0 ~,,~~ ~'""O" 'd ~~£] ~ .: " 8.'" '" f/) ~ ~ "- - ~ ~ " ; E-' "" ~~'"N ~ f::' -.: 8.~ ; ~ = "~ (/) ~ ~ ~ ~ ~ -~ -~ ,,'"..."~=u~c ~ ~ ~ " oS"~ "0 c: . ~ "60 u i " t ~ ~ ;:~ ~~oo~ -~...t: ~ ;: ~ .N .N ~ ~ ~ .~ ~.J:) g Co,:: 8. uii\ .~ ~ ~ .!1~.~~ 1~< "O.S~ ~ "" ~ ~-.§ ~;"O- ~ ~ ~ M~ ~- ~~"-"0 0 ~ "' ~ ~ '" '" . ~ ~~;= >- >- >- ~ ~ ~'"~" :S "0 :S "0 :S "0 :S .:= :S .:= :S :.~ "~8.] b ~ ~ 3 ~3 ~ 3 ~~ ~~ ~ ~; ~-=i~ ~ e ;:~"'~"'~"'~~~~~.~N "';: "';: "';::-;::-;:= ~ "E.§~a oJ §"., .~.~.~.~ ~.~ ".§. > .~ > .~ 0 ~ ~~"~'" =-" ~ ,,~ ~ -5 ~ -5 ~ -5 ~ < ~ < ~ u ~ ... '" ~ ~ ~ E;:; E;:; E;:; E=: E=: E=: .~u~

, .

:it FITZGERALD ET AL.-MOLECUlAR PHYLOGENYOF TUPlNAMBIS 901

RescaledDistance whereas all Paraguayan records of 7: rufescens

Species ~ ~ :,~ 't 2,~.:; are far west of the Paraguay River (Fig. 1). Sum- T _rian.. marizing, 7: duseni differs from 7: rufescensin ~: =~::: many morphological and ecological characters, '.. ~: =~::: and there is presumably some heritable genetic T. _rianae component to this variation. Tupinambis duseni T. ~"ana. T. ~iana. may have split recently from a 7: rufescensan- ~ =~::: cestorand is experiencingrapid morphological ~: ::~::: divergence in the isolated cerrado habitats. The ~ =~::: paraphyletic relationship indicated by our lim- ~ ::~::: ited sampling scheme suggests the possibility of ;: ::~::: incomplete lineage sorting (Avise,1994; Patton ~:=~::: and Smith, 1994). Thus, these taxa do not ap- ;: ::~::: pear to be reciprocallymonophyletic because of ;: ::~::: the retention of ancestralalleles. Alternatively, T _riana. the similarity among the mitochondrial gene se- T ~"ana. T. ~ianae quences examined could have resulted from in- ~: ::~::: trogression of mitochondrial genes between the ~ ::~::: two taxa. This explanation appears less likely as ~: =~::::. these species are not sympatric in Paraguay. Fi- ~ ::~ nally, we know very little about the occurrence ;: ::~ of 7: rufescensin Brazil, and how Brazilian 7: ru- ;: =::~= fescenscompare to 7: rufescensfrom the Chaco. ~ =::~:.,u: The recordsin Presch(1973) are unreliable be- ;: =::~:: causeof the general confusionof northern and T. ruL..a.". southern forms of TuPinambis, the previous Fig. 4. A dendrogramof morphologicaldistances identi~ca~~n o~ the. specimens as 7: teguixin, among 44 individual Tupinambisconstructed from a and hIS mISIdentification of sepsmarmoratus Lau- cluster analysisusing five size-adjustedmeasurement renti 1768 as 7: rufescens(a juvenile 7: teguixin variablesand five scale characters.Gowers' median according to Avila-Pires, 1995). method wasused for computingdistances (a weighted Although 7: merianaeis sympatric with 7: ru- p,air-groupmethod using cen~oids). Mo~holo~cal fescensin the transition zone of the Paraguayan distances were co~puted usIng Gowers median Chaco, 7: duseni is allopatric to populations of method to reduce biasesthat could result from com- rr.J; b. . d .. . bl 1. ~escens. Ab e e. t al (1992) repor t e d sympa tr y mmg measurement an menstlc vana es. rr d . rr. among 1. usenz, 1. menanae,an d rr1. teguzxzn. . at Baliza, Goias, Brazil. We suggest that 7: duseniis and 7: merianaein having large, convex scales a valid taxon exhibiting rapid morphological di- on the anterior part of the dorsum. The de- vergence from a 7: rufescens-likeancestor. Larger scription of 7: duseni (UJnnberg and Andersson, samples of 7: rufescensfrom throughout its 1910) points out these features and is entirely range, especially from Brazil, should clarify consistent with the specimens collected in Par- whether 7: duseni is distinctive, and define the aguay. One of us (ALA) examined the holotype extent of morphological and genetic variation (NRM 2886) and found no differences between within 7: rufescens. it and Paraguayan material. TuPinambisduseni was previously known only Implications for conservation, management,and from the type locality, Parana, Brazil. Our dis- trade.- This phylogeny, along with the taxonom- covery of 7: duseniat three localities in Paraguay ic clarifications by Avila-Pires (1995), carries se- since 1993 provided additional material that rious implications for conservation of these ex- gave insight into, but did not resolve, the ques- ploited lizards. First, all species of Tupinambis tion of the specific status of 7: duseni.The minor are managed as a single entity at this time, with genetic divergence between 7: rufescensand 7: quotas allocated for the entire genus within a duseni suggests the two forms are very closely country. Though harvest monitoring programs related. Despite this close genetic association, identify skins according to species and sex (Fitz- TuPinambisduseni exhibits consistent and strik- gerald et al., 1991, 1994), skins are mixed during differences from 7: rufescensin coloration, ing the tanning process making it difficult to shape of the head, and limb morphology. Ad- verify quantities of each species that are export- ditionally, 7: duseni has been found only in iso- ed. Although species of Tupinambisare morpho- lated cerrado habitats in eastern Paraguay, logically similar, our results indicate substantial

. '0

'. ..-j!, 902 COPEIA, 1999, NO.4

genetic variation exists within them, and this and comparisons between 7: merianaeand 7: te- variation is likely to be associated with geo- guixin should be interpreted with caution (Har- graphically structured populations. The avail- vey and Pagel, 1991). The northern/Amazonian ability. of phylogenetic information should clade, though less studied, contains smaller spe- make It easier to move toward management of cies with a correspondingly smaller clutch size these lizards at the level of individual species (Reese, 1922). In Venezuela and Ecuador, 7: te- and identify species and populations that may guixin nests in termite mounds (Reese, 1922; need distinct management practices. Beebe, 1945; Dixon and Soini, 1975). Hence, .A striking example of the relevance of system- management should account for phylogenetic a.tlcs ~o the conserv~tion. of Tupinambis is the similarity, particularly when attempting to pre- situation of 7: dusenJ.This taxon was recently dict species specific responses to exploitation discovered in Paraguay where it is apparently and habitat change when faced with limited confined to patchy cerrado habitats. A substan- data on population ecology. tial portion of cerrado vegetation has been con- verted to pasture for cattle grazing, yet the im- pact of this habitat modification on populations MATERIALS,EXAMINED of 7: duseni is unknown. Meanwhile, the species has been harvested for the skin trade since be- Corresponding voucher specimens are listed fore 1980 along with 7: merianaeand 7: rufescens. by .species, ca~lo? number, -:md l?cality: Insti- Skins of 7: duseni were identified in samples of tutlonal abbreVIations are as listed In LeVIton et the yearly harvest used for monitoring the liz- al. (19~5).. . . ards in Paraguay. An estimated 5000 7: duseni TupJnambJs teguJxJn (LSUMNS H12405, entered the trade in 1996 (M. A. Mieres and A. H12450, H12431): BRAZIL; Roraima; Fazenda L. Aquino, Report on the Tupinambisharvest to Nova Es~~ran~a; 44 km W BR-174 on BR-210; CITES-py; Ministerio de Agricultura y Ganad- 7: teguJxJn (LSUMNS H12678, LSUMNS eria, Paraguay, 1997, unpubl.). H12703, LSUMNS H12715): ECUADOR; Suc- Second, a taxonomy based on the partition- umbios Province; Reserva Faunistica Cuyabeno; ing of geographic and genetic variation and the Estacion Biologica da Universidad Catolica. delineation of species limits will facilitate inter- TuPinambis longilineus (LSUMNS H14135, national monitoring (e.g., CITES) of the skin LSUMNS H14136): BRAZIL; Amazonas; Rio trade. TuPinambisteguixin does not occur in Ar- Ituxi at the Madeireira Scheffer; S 8°20'47.0" W gentina and Paraguay, for example, so any 7: 65°42'57.9". teguixin in the lizard trade must originate from Tupinambisduseni (USNMField 166766 (tissue elsewhere. Our preliminary data provide a 766), MNHNP 5074-5): PARAGUAY; Concep- framework for forensic identification of species cion; 17 km S 20 km E Yby Yau; Estancia Siete and the geographic origins of skins. The mor- Laguna; S 23°06.586' W 56°20.001'; 7: duseni phological analysesindicated, for example, that (MNHNP 6181-94): Concepcion; Paso Barreto; any skins with fewer than 28 ventral scales in a 7: duseni (USNMField 166778 (tissue 778), row belong to a species from the northern/ Am- MNHNP 4342-3, 5076): Amambay; Colonia Yby- azonian clade, and could not have originated cui; S 23°38' W 55°32'; 7: duseni (NRM 2886 ho- from Paraguay or Argentina. Genetic markers lotype): BRAZIL; Parana. identified in this study may identify taxa with a Tupinambis rufescens[MNHNP 5068, USNM- high degree of confidence. The substantial in- Field 166740 (tissue 740)]: PARAGUAY; Presi- traspecific genetic variation found in this study dente Hayes; Estancia Santa Elisa; S 22°24'33" indicates, however, that more complete geo- W 58°56'33"; 7: rufescens[MNHNP 4141, 5073, graphic sampling of genetic and morphological USNMField 166743 (tissue 743)]: Presidente variability is necessary to manage and monitor Hayes;Juan de Zalazar; S 23°04'44"W 59°14'12"; effectively such a heavily harvested group of 7: rufesGens(MNHNP 4145): Reserva Indigena species. Casanillo; S 22°11'02" W 59°19'07"; 7: rufescens Finally, a .phylogeny of TuPinambisallows pre- (MNHNP 4340-1): Boqueron; 7 ma. Division de dictions to be made about the potential for sus- Infanteria Tte 1° Alfredo Stroessner; S 22°39.5' I tainability of these little-studied lizards. The S 61°30.5'; 7: rufescens(MNHNP 4339): Boquer- clade containing 7: rufescensand 7: merianaeis on; Neuland; 7: rufescens(MNHNP 4338): Bo- charact~rized b~ lar?e body size, relatively large queron; 5 km W Filadelfia; 7: rufescens(MNHNP clutch SIZe,nesting In burrows, nest attendance, 4146-7): Boqueron; 60 km ENE Filadelfia; 7: and a seasonal activity pattern (Fitzgerald et al., rufescens(MNHNP 4337): Boqueron; Colonia 1991, 1993). Based on our results, this group is Campo Alegre; 25 km W de Neuland; 7: rufes- distant from the northern/Amazonian clade, GenS(MNHNP 3044): Boqueron; Filadelfia; 7:

I ,: " . '~

:~ FITZGERALD ET AL.-MOLECUlAR PHYLOGENYOF TUPlNAMBIS 903

rufescens(MNHNP 3036, 3040): Chaco; 70 kIn idente Hayes; Juan de Zalazar; 7: merianae NW Filadelfia on road to Madrejon. (MNHNP 4142-4): Presidente Hayes; Juan de Tupinambis merianae (MNHNP 3984-4000, Zalazar; S 23°04'36" W 59°14'07"; 7: merianae 4345): PARAGUAY; Alto Paraguay; 30 kIn SW (MNHNP 5070): Presidente Hayes;Juan de Zal- Bahia Negra; Colonia Potrerito; 7: merianae azar; 2 kIn E administration; S 22°26'40" W (MNHNP 3043): Alto Paraguay; Bahia Negra; 7: 58°42'40"; 7: merianae (MNHNP 4328): Presi- merianae(MNHNP 4326): Puerto Casado; 7: mer- dente Hayes; kIn 75.5 Ruta Trans-Chaco; 7: mer- 0 ianae (MNHNP 4426): Alto Parana; 15 kIn SW ianae (MNHNP 6218, 6228): Presidente Hayes; Santa Fe; 7: merianae (MNHNP 3039): Amam- Reserva Indigena Casanillo; S 22°11'02" W bay; Parque Nacional Cerro Cora; S 22°38'03" 59°19'07"; 7: merianae[USNMField 166649 (tis- W 56°8'03"; 7: merianae(6205): Canindeyu; Re- sue 649 )] : PARAGUAY; Presidente Hay es' Ruta ° , 55° 64" " ser~ Mbaracayu; S 24 08.004 ~ 31.6 ; 7: Pozo Colorado to Concepcion; S 23°31'07.0" W menanae ~MNHNP 3136): Canm.rleyu; Res,erva 57°50'56.7"; 7: merianae (MNHNP 4329-30, Mbaracayu; 1 kIn S ~esto Lagumta; 7: menanae 4333-5, 6180, 6196-7, 6206-7, 6218, 6228): (MNHNP 3042): Canmdeyu; Reserva Mbaraca- P - " reSI.d en te H ayes; kIn 250-253 Ru ta Trans-ch a- yu; 3.1 km NE N,andurocal; 7: menanae co; S 23°38.652' W 58°41.958'; 7: merianae (~NHNP 4011): Canmd~yu; Reserva Mbaraca- (MNHNP 4336): Presidente Hayes; 10 kIn N of yu; Puesto Morena; 7: menanae (MNHNP 4008): . - Ruta Mlltar' l ' at kIn 50; C0 1oma. Makthl awmya;' S Canmdeyu; Reserva Mbaracayu; Puesto Carapa; ° 7'38" 58°2 '10". ' 7: merianae (USNMField 167290): Canindeyu; 232 W. 0 , 7: menanae.(MNHNP Reserva Mbaracay(1'S 24°08' W 55°31'" in forest 5069): Presldente Hayes; Estancla Campo SW of reserve headquarters; 7:' merianae Verde; ,S 22°26:40" W 58°42:40". . (USNMField 167291): Central; Asuncion; 7: mer- Amezva amezva [USNMFleld 166640 (tIssue ianae (MNHNP 4346): Central; San Lorenzo; S 640)]: PARAGUAY; Presidente Hayes; Ruta 25°20'6" W 57°31'05"; 7: merianae[USNMField Trans-Chaco km 75; Cabana; S 24°50'17.9" W 166779 (tissue 779), USNMField 166780 (tissue 57°46'40.8". 780), MNHNP 6234-5]: Caazapa; Parque Na- Kentropyxviridistriga [USNMField 166765 (tis- cional Caaguazu; Admnistracion S 26°05'50.9" sue 765)]: PARAGUAY; Concepcion; 17 kIn S, W 55°26'31.3"; 7: merianae (MNHNP 6236): 20 kIn E Yby Yau; Estancia Siete Lagunas; S Concepcion; 17 kIn S 20 kIn E Yby Yau; near 23°06.173' W 56°21.568'. Estancia 7 Lagunas; S 23°10.131' W 56°19.156"; CnemidoPhorusocellifer [USNMField 166784 7: merianae [USNMField 166756 (tissue 756)]: (tissue 784)]: PARAGUAY; Amambay; Parque Concepcion; 30 kIn N Estancia San Fernando; Nacional Cerro Cora. Parque Nacional Serrania San Luis; Retiro Lau- rel; S 22°40'30.2" W 5~21'02.2"; 7: merianae (MNHNP 5071-2, 6177-9, 6198-204, 6208-17, ACKNOWLEDGMENTS 6219-27,6229-33): Concepcion; near Paso Bar- reto; S 23°03.494' ~ 5~07.315";. 7: merianae We are especially indebted to G. Terol for his (MNHNP 6176): CordIllera; Estancla Sombrero; leadership in the field and museum. We also 25°04' S 56°36' W; 7: merianae(MNHNP 4364): thank R. Palacios M. Mieres C. Morales and S. C~rdi!l~ra; Es:ancia Ypoa; S 25°55'5" W Frutos for invalu~ble help i~ the field. B. Tan- 59 25 9 ; 7: menanae(MNHNP 4344, 6176): Cor- ner and]. Demboski gave excellent laboratory

dillera; Estancia Sombrero; 8 kIn NW Cleto " Romero; S 25°04'26" W 56°36'08"; 7: merianae asslstance.m Alaska. V. Mor ales an.d C. Conroy (MNHNP 3041): Itapua; 10 kIn W Centro de help~d WIth the ~gures. Key specImens ~ere D 11 F tal. ..,. , (MNHNP 3977) ' proVIdedby L. J. Vltt andJ. P. Caldwell(NatIon- esarro 0 ores , ~. menanae . al S . F d . M . . . 15 kIn SSW S tI . . S2.-.o15'8" W Clence un atIon grant DEB-9200779) . F . H . ISlones, an ago, I Sh 1 d D D. ~ .1' d 1 f . 56°49' 8"; 7: merianae (MNHNP 3815): Misiones; e don an . lttmann J.aClltate oans 0 tIs- Yabebyry; 7: merianae(MNHNP 3052): Misiones; sue from LSUMNS. M. Motte provided addition- , San Ignacio; S 26°53'01" W 57°01'06"; 7: meri- al materi~ .from MNH~P ~n Paraguay. S. Kul- anae (MNHNP 6175): Paraguari; Lago Ypoa; S lander facIlItated examInatIon of the holotype 25°56'71" W 57°26'80"; 7: merianae (MNHNP of 7: duseni. T. Titus provided the LIZ1-LIZ2 4365): Paraguari; Lago Ypoa; Estancia Ypoa; 32 primers. N. Scott, G. Colli, F. Lobo, and two kIn W Valle Apua; 7: merianae(MNHNP 2838- anonymous reviewers gave useful comments on 9, 3037, 4327): Paraguari; Parque Nacional Yby- the manuscript. Funding was provided by a cui; S 26°5'6" W 56°50'2"; 7: merianae(MNHNP grant from World Wildlife Fund. F. W. King 6195): Presidente Hayes; E;stanciaVilla Rey, Cas- helped with administration of funds donated in co Principal; 7: merianae(MNHNP 6237): Pres- Paraguay.

:~ I 904 COPEIA, 1999, NO.4

LrrERATURE CUED pie: a sustainable use approach to conservation and development. Conserv. Bioi. 8:12-15. ME, A. S., O. S. PESANTES,ANDP. R. MANZANI. 1992. -,J. M. CHANI,AND O. E. DONADIO. 1991. Tup- Considera~oes sobre 0 genero Tupinambis (Lacer- inambis lizards in Argentina: implementing man- tilia, Teiidae). Resumos XII Gong. Latino-Am. agement of a traditionally exploited resource, p. Zool., XIX Congreso Brasil. Zool., Belem, Para:128. 303-346. In: Neotropical wildlife use and conser- ANDERSON,S., M. H. L. DE BRUlJN,A. R. COUL50N,I. vation.J. H. Robinson and K. Redford (eds.). Univ. C. EPERON,F. SANGER,AND I. G. YOUNG.1982. Com- of Chicago Press, Chicago. plete sequence of bovine mitochondrial DNA: con- -, F. B. CRUZ,AND G. PEROTTI. 1993. The re- served features of the mammalian mitochondrial productive cycle and the size at maturity of Tupi- genome. J. Mol. Bioi. 156:683-717. nambis rufescens(Sauria: Teiidae) in the dry chaco AREvALo, E., S. K. DAVIS,AND J. W. SITESJR. 1994. of Argentina. J. Herpetol. 27:70-78. Mitochondrial DNA sequence divergence and phy- -, G. PORINI,AND V. LICHTSCHEIN.1994. Aspec- logenetic relationships among eight chromosome tos hist6ricos, estado actual, y perspectivas futuras races of the Sceioporusgrammicus complex (Phry- del manejo de Tupinambisen Argentina. Intercien- nosomatidae) in central Mexico. Syst. Bioi. 43:387- cia 19:166-170. 418. HARVEY,P. H., ANDM. PAGEL.1991. The comparative AVILA-PIRES,T. C. S. 1995. Lizards of Brazilian Ama- method in evolutionary biology. Oxford Univ. zonia (Reptilia: Squamata). Zool. Verh. 299:1-706. Press, Oxford. AVISE,J. C. 1994. Molecular markers, natural history, HILLIS, D. M., ANDJ. P. HUELSENBECK.1992. Signal, and evolution. Chapman and Hall, New York. noise, and reliability in molecular phylogenetic BEEBE,W. 1945. Field notes on the lizards of Kartabo, analysis.J. Hered. 83:189-195. British Guiana, and Caripito, Venezuela. Part 3. Tei- HOOGMOED,M. S. 1973. Notes on the herpetofauna idae, Amphisbaenidae and Scincidae. Zoologica 30: of Surinam IV. The lizards and amphisbaenians of 7-32. Surinam. Biogeographica 4:v-419. BENABIB,M., K. M. K;JER,AND J. W. SITESJR. 1997. IRWIN,D. M., T. D. KOCHER,AND A. C. WIL50N. 1991. Mitochondrial DNA sequence-basedphylogeny and Evolution of the cytochrome b gene of mammals. J. the evolution of viviparity in the Sceioporusscalaris Mol. Evol. 32:128-144. group (Reptilia, Squamata). Evolution 51:1262- KOCHER,T. D., W. K. THOMAS,S. V. EDWARDS,S. 1275. PDDBO,F.X. VILLABLANCA,AND A. C. WILSON. 1989. BOULENGER,G. A. 1885. Catalogue of the Lizards in Dynamics of mitochondrial DNA evolution in ani- the British Museum (Natural History). 2d ed. Vol. mals: amplification and sequencing with conserved 1. Trustees of the British Museum, London. primers. Proc. Nad. Acad. Sci. USA 86:6196-6200. BREMER,K. 1988. The limits of amino acid sequence LEVITON,A. E., R. H. GIBBSJR., E. HEAL, AND C. E. data in angiosperm phylogenetic reconstruction. DAWSON.1985. Standards in herpetology and ich- Evolution 42:795-803. thyology. Part I. Standard symbolic codes for insti- BURT,C. E., AND M. D. BURT. 1931. South American tutional resource collections in herpetology and lizards in the collection of the American Museum ichthyology. Copeia 1985:802-821. of Natural History. Bull. Am. Mus. Nat. Hist. 61: WNNBERG,E., AND L. G. ANDERSSON.1910. A new 227-395. lizard and a new frog from Parana. Arkiv. Zoologi GEl, J. M. 1986. Reptiles del centro, centro-oeste y 6:1-11. sur de la Argentina: herpetofauna de las zonas ar- Losos,J. B., T. R.JACKMAN,A. LARSoN,K. DEQUEIROZ, idas y semiaridas. Monografie IV, Mus. Reg. Sci. AND L. RODRIGUEZ-ScHETTINO.1998. Contingency Nat., Torino, Italia. and determinism in replicated adaptive radiations -. 1993. Reptiles del noroeste, nordeste, yeste of island lizards. Science 279:2115-2118. de la Argentina: herpetofauna de las selvassubtrop- LUXMOORE,R., B. GROOMBRIDGE,AND S. BROAD.1988. icales, Puna y Pampas. Monografie XIV, Mus. Reg. Significant trade in wildlife: A review of selected Sci. Nat., Torino, Italia. species in CITES Appendix II. Vol. 2. reptiles and -, ANDJ. A. ScOLARO. 1982. A population of invertebrates. International Union for the Conser- Tupinambisfrom northern Patagonia, south of the vation of Nature and Natural Resources (IUCN) , Rio Negro, Argentina. Herpetol. Rev. 13:26. Gland, Switzerland. COLLI, G. R., A. K. PERESJR., AND H. J. DA CUNHA. MANZANI,P. R., AND A. S. ME. 1997. A new species 1998. A new species of Tupinambis (Squamata: Tei- of TuPinambis Daudin, 1802 (Squamata, Teiidae) idae) from central Brazil, with an analysis of mor- from central Brazil. Bol. Mus. Nac., n.s. 382:1-10. phological and genetic variation in the genus. Her- MILES,D. B. 1994. Covariation between morphology petologica 54:477-492. and locomotor performance in Sceloporine lizards, DESJARDINS,P., AND R. MORAis. 1990. Sequence and p. 207-235. In: Lizard ecology: historical and ex- gene organization of the chicken mitochondrial ge- perimental perspectives. L. J. Vitt and E. R. Pianka nome. J. Mol. Bioi. 212:599-634. (eds.). Princeton Univ. Press, Princeton, NJ. DIXON,J. R, AND P. SOINI. 1975. The reptiles of the MILLER,S. A., D. D. DYRES,AND H. F. POLESKY.1988. upper Amazon Basin, Iquitos region, Peru 1. Liz- A simple salting procedure for extracting DNA ards and amphisbaenians. Milwaukee Pub. Mus. from human nucleated cells. Nucleic Acids Res. 16: Contrib. Bioi. Geol. 4:1-58. 215. FITZGERALD,L. A. 1994. Tupinambislizards and peo- NORMAN,D. R. 1986. Man, anacondas, and tegus in

~ .

j ! FITZGERALD ET AL.-MOLECULAR PHYLOGENY OF TUPINAMBIS 905

eastern Paraguay.Unpubl. master'sthesis, Tulane REESE,A. M. 1922. A note on the breedinghabits of Univ., New Orleans,LA. tegu. Copeia 110:69-72. PATrON,J. L., AND M. F. SMITH. 1994. Paraphyly, po- SoRENSON,M. D. 1996. TreeRot. Univ. of Michigan, lyphyly, and the nature of species boundaries in Ann Arbor. {. pocket gophers (genus Thomomys).Syst. Bioi. 43: SUUJVAN,R. M., ANDR. ESTES.1997. A reassessment 11-26. of the Tupinambinae,p. 100-112.In: Vertebratepa- PETERS,J. A., AND R. DONoso-BARROS.1970. Cata- leontology in the Neotropics: the Miocene fauna of logue of the Neotropical Squamata. Part II. Lizards La Venta, Colombia. R. F. Kay, R. H. Madden, R. L. ('"' and Amphisbaenians. Bull. U.S. Nat!. Mus. 297:1- Cifelli, and J. J. Flynn (eds.). Smithsonian Institu- 293. tion Press,Washington, DC. -, AND-. 1986. Catalogue of the Neotrop- ical Squamata. Part II. Lizards an Amphisbaenians. Rev. ed. Smithsonian Institution Press,Washington, (lAF) DEPARTMENT OF WILDUFE AND FISHERIES DC. ScIENCES,TEXAS COOPERAnVE WILDUFE COL- PETREN,K., ANDT.J. CASE.1997. A phylogenetic anal- LECrlON, TEXAS A&M UNIVERSnY, COLLEGE ysis of body size evolution and biogeography in STA'I10N, TEXAS 7784~2258; (jAC) UNlVERSI- chuckwallas (Sauromalus)and other iguanines. Evo- TV OF ALAsKA MUSEUM, 907 YuKON DRIVE, lution 51:206-219. FAIRBANKS, ALAsKA 99775-6960; AND (AlA) PRESCH,W. 1973. A review of the tegus, lizard genus CITES-PARAGUAY MINISTERIO DE AGRICULTU-

Tupi~mbis (Sauria: Teiidae) from South America. RA Y GANADERlA, GRAL. AQUINO 1888 Y AuRA, Copela 1973:740-746. As . P . -. 1974. Evolutionaryrelationships and bioge- UNCION, ~GUAY. E-~a1:1 ( lAF) lfi tzger- ography of the macroteiid lizards (family Teiidae, [email protected]. Send repnnt requests to lAF. subfamily Teiinae). Bull. So. Calif. Acad. Sci. 73:23- Submitted: 12 Aug. 1998. Accepted: 9 April 32. 1999.

~. ,.

:~ "J

-.,