Phylogenetic Relationships in the Iguanid Lizard Genus Liolaemus

Biological Journal of the Linnean Society (2000), 69: 75–102. With 7 ﬁgures doi: 10.1006/bijl.1999.0346, available online at http://www.idealibrary.com on

Phylogenetic relationships in the iguanid lizard genus Liolaemus: multiple origins of viviparous reproduction and evidence for recurring Andean vicariance and dispersal

JAMES A. SCHULTE II1∗, J. ROBERT MACEY1, ROBERT E. ESPINOZA2 AND ALLAN LARSON1

1Department of Biology, Box 1137, Washington University, St. Louis, MO, USA 63130–4899. 2Ecology, Evolution and Conservation Biology/MS314, University of Nevada, Reno, NV, USA 89557

Received 2 March 1998; accepted for publication 15 January 1999

Phylogenetic relationships within the iguanid lizard genus Liolaemus are investigated using 1710 aligned base positions (785 phylogenetically informative) of mitochondrial DNA sequences, representing coding regions for eight tRNAs, ND2, and portions of ND1 and COI. Sixty new sequences ranging in length from 1736 to 1754 bases are compared with four previously reported sequences. Liolaemus species form two well-supported monophyletic groups of subgeneric status, Liolaemus and Eulaemus. These subgenera appear to have separated at least 12.6 million years ago based on the amount of molecular evolutionary divergence between them. Hypotheses that species occurring in the Andes, west of the Andes, and east of the Andes, each comprise distinct monophyletic groups are independently rejected statistically. The shortest estimate of phylogeny suggests that Liolaemus originated either in the Andes or the eastern lowlands. Numerous evolutionary shifts have occurred between the Andes, and the eastern and western lowlands, suggesting recurring vicariance and dispersal. Species occurring at high elevations or high latitudes usually have viviparous reproduction. Depending on whether parity mode is considered reversible in Liolaemus, the most parsimonious reconstruction supports at least six independent origins of viviparity or at least three gains followed by three losses of viviparity among the 60 Liolaemus lineages examined.

 2000 The Linnean Society of London

ADDITIONAL KEYWORDS:—Reptilia – Squamata – Iguania – Iguanidae – Liolaemus – phylogeny – reproduction – historical biogeography – viviparity – South America – Andes.

CONTENTS

Introduction ...... 76 Methods ...... 79

∗ Corresponding author. E-mail: [email protected] 75 0024–4066/00/010075+28 $35.00/0  2000 The Linnean Society of London 76 J. A. SCHULTE ET AL.

Specimen information ...... 79 Laboratory protocols ...... 80 Phylogenetic analysis ...... 80 Results ...... 81 Authentic mitochondrial DNA ...... 81 Assessment of homology and sequence alignment ...... 82 Genic variation ...... 82 Phylogenetic analyses ...... 83 Evaluating Andean vicariance ...... 87 Evolution of viviparity ...... 88 Discussion ...... 89 Historical biogeography ...... 89 Dating phylogenetic divergence and vicariance ...... 91 Evolution of parity mode ...... 93 Phylogenetic relationships and taxonomy ...... 93 Acknowledgements ...... 95 References ...... 95 Appendix 1 ...... 98 Appendix 2 ...... 99 Appendix 3 ...... 101

INTRODUCTION

The iguanid lizard genus Liolaemus contains over 160 species distributed in the Andes and adjacent lowlands of South America from Peru to Tierra del Fuego (Cei, 1986, 1993; Donoso-Barros, 1966; Etheridge, 1995). This genus has one of the largest latitudinal, elevational, and climatic distributions among lizards world-wide, with species found from sea level to over 5000 m (Cei, 1986, 1993; Donoso-Barros, 1966). Physiographic factors divide the distribution of Liolaemus into three major regions: (1) Andean highlands, (2) lowlands east of the Andes, and (3) lowlands west of the Andes. We test the hypothesis that the uplift of the Andes divided the ancestral Liolaemus into monophyletic groups corresponding to one or more of the three physiographic regions (Fig. 1). Alternatively, continual uplifting of the Andes over the last 25 million years (Norabuena et al., 1998) may have caused recurring episodes of vicariance superimposed on periodic dispersal events between the Andes, and the eastern and western lowlands. The latter hypothesis predicts that none of the three regions should comprise monophyletic groupings of Liolaemus species. The highest elevation of the Andean mountain chain currently extends up to 7000 m and presents a formidable barrier to dispersal for Liolaemus; however, numerous potential dispersal corridors occur under 5000 m. Because approximately half of the Liolaemus species are live-bearing, the phylogenetic pattern of reproductive mode presents an interesting issue for investigation. Viviparity, the retention within the uterus of a developing neonate until development is complete, appears to have evolved frequently among squamate reptiles in cool environments at high latitudes and elevations (Guillette, 1993; Shine, 1985; Shine & Bull, 1979; Tinkle & Gibbons, 1977). Recent molecular phylogenetic work on iguanid lizards of the Sceloporus scalaris complex has demonstrated that viviparity evolved twice within that species group and possibly reversed to oviparous reproduction once (Benabib, Kjer & Sites, 1997; Creer et al., 1997; Mink & Sites, 1996). An evolutionary analysis of reproductive modes within Liolaemus may reveal whether viviparity has evolved multiple times and potentially could reverse (Lee & Shine, 1998). LIOLAEMUS PHYLOGENETICS 77

Eastern and Western Andean

A High-elevation species form a clade

Andean and Western Eastern

B Eastern lowland species form a clade

Andean and Eastern Western

C Western lowland species form a clade

Figure 1. Hypotheses for the historical biogeography of Liolaemus. A possible outcome of the initial uplifting of the Andes is the geographic fragmentation of Liolaemus populations into major subgroups located in (A) the Andes, (B) the eastern lowlands and/or (C) the western lowlands (Table 1). If Liolaemus populations in these three regions formed a genetically homogeneous entity prior to the Andean uplift, and populations in one or more regions remained isolated following the initial uplifting, one to three major clades of Liolaemus corresponding to any or all of the geographic regions may result. If phylogenetic structure was present in Liolaemus prior to the Andean uplift, or if episodes of vicariance and dispersal between these three regions followed the initial uplift, one or more of the three major areas would contain a phylogenetically heterogeneous grouping of species.

A well supported estimate of phylogeny is used to test these hypotheses of historical biogeography and evolution of viviparity. Samples of approximately one-third of the recognized Liolaemus species, including representatives occurring in all physiographic provinces and exhibiting both reproductive modes, are examined. These species represent the major groups from the most recent taxonomy of Liolaemus (Etheridge, 1995; Table 1). Sixty new mitochondrial DNA sequences are reported for the same region sequenced for a broad sampling of iguanid lizard taxa by Macey et al. (1997b). This sequence extends from the end of the gene encoding ND1 (NADH dehydrogenase 1) to the beginning of the gene encoding COI (cytochrome c oxidase I) and includes all of the ND2 gene, eight tRNA genes, and the origin for light- strand replication (OL). This same region of DNA sequenced for Liolaemus pictus (Macey et al., 1997b) is included in this analysis. Outgroups are based on the phylogenetic hypothesis of Schulte et al. (1998). In 78 J. A. SCHULTE ET AL.

T 1. Taxa included in this study, by region of largest distribution, range in elevation (meters above sea level), and parity mode (O=oviparous, and V=viviparous) primarily based on museum specimens, Cei (1986, 1993), and Donoso-Barros (1966). Key references for parity mode are given to the right. Andean taxa occur above 2500 meters, whereas eastern and western taxa occur exclusively from sea level to 2500 meters on either side of the Andes. Museum catalogue numbers and localities are given in Appendix 1. Subgeneric designations are after Laurent (1984) and correspond to the phylogenetic results of this study. Species sections and series for the subgenus Eulaemus are largely after Etheridge (1995), and are amended to correspond with results of this study. Taxa denoted with a cross are members of the L. wiegmannii group of Etheridge (1995) and here are called the ‘sand-lizard clade’ of the L. boulengeri series Taxon Region Elevation Parity Reference range mode

Outgroup taxa 1. Oplurus cuvieri —— O 2. Phrynosoma douglassii —— V 3. Phymaturus palluma Andean 2500–4000 V Donoso-Barros (1966) 4. P. somuncurensis Eastern 1200 V Cei (1986) Subgenus Liolaemus 5. Liolaemus coeruleus Eastern 1500–1700 O Espinoza, unpubl. data 6. L. alticolor Andean 3000–3600 O Espinoza, unpubl. data 7. L. bitaeniatus Andean 700–2800 O Cei (1993) 8. L. robertmertensi Andean 690–2600 O Cei (1993) 9. L. bibronii Andean 0–3000 O Cei (1986) 10. L. gracilis Eastern 0–1380 O Cei (1986) 11. L. bellii Andean 2000–3200 V Ramı´rez Pinilla (1991a) 12. L. chiliensis Western 0–2100 O Donoso-Barros (1966) 13. L. cyanogaster Western 0–225 V Donoso-Barros (1966) 14. L. pictus – Neuque´n Eastern & 10–1600 V Cei (1986) Western 15. L. pictus – Rio Negro Eastern & 10–1600 V Cei (1986) Western 16. L. zapallarensis Western 0–762 O Espinoza, unpubl. data 17. L. tenuis Western 0–1800 O Donoso-Barros (1966) 18. L. lemniscatus Western 0–1800 O Donoso-Barros (1966) 19. L. monticola Western 500–1800 O Donoso-Barros (1966) 20. L. nitidus Andean 0–3153 O Ramı´rez Pinilla (1991a) 21. L. fuscus Western 0–1800 O Donoso-Barros (1966) 22. L. nigroviridis Andean 500–4000 V Donoso-Barros (1966) 23. L. capillitas Andean 2500–3900 V Cei (1993) 24. L. leopardinus Andean 1800–2700 V Donoso-Barros (1966) 25. L. buergeri Andean 1500–3000 V Donoso-Barros (1966) 26. L. ceii Eastern 1000–2300 V H. Nu´n˜ez, pers. comm. 27. L. petrophilus Eastern 900–1400 V Cei (1986) 28. L. austromendocinus Eastern 900–2310 V Cei (1986) 29. L. elongatus Andean 700–3000 V Cei (1986) Subgenus Eulaemus L. lineomaculatus section 30. L. lineomaculatus Eastern 0–1500 V Donoso-Barros (1966) 31. L. somuncurae Eastern 1200–1600 V Scolaro and Cei (1997) 32. L. magellanicus Eastern 0–1050 V Donoso-Barros (1966) L. montanus section L. montanus series 33. L. ruibali Andean 2370–3000 V Cei (1986) 34. L. andinus – La Rioja Andean 3000–4700 V Ramı´rez Pinilla (1991a) 35. L. famatinae Andean 3700–4200 V Espinoza, unpubl. data 36. L. orientalis Andean 4000–4320 V Espinoza, unpubl. data 37. L. dorbignyi Andean 3200–4400 V Donoso-Barros (1966) 38. L. poecilochromus Andean 3500–4130 V Cei (1993) 39. L. multicolor Andean 3360–4400 V Ramı´rez Pinilla (1991a) 40. L. andinus – Jujuy Andean 3390 V Espinoza, unpubl. data

continued LIOLAEMUS PHYLOGENETICS 79

T 1—continued Taxon Region Elevation Parity Reference range mode

L. boulengeri series 41. L. cuyanus Eastern 400–2000 O Cei (1986) 42. L. ﬁtzingerii Eastern 0–1100 O Cei (1986) 43. L. pseudoanomalus Eastern 990–1700 O Espinoza, unpubl. data 44. L. lutzae‡ Eastern 0 O Rocha (1992) 45. L. occipitalis‡ Eastern 0–200 O Espinoza, unpubl. data 46. L. multimaculatus‡ Eastern 0–1000 O Cei (1986) 47. L. scapularis‡ Eastern 1100–2100 O Ramı´rez Pinilla (1994) 48. L. salinicola‡ Eastern 0–2050 O Espinoza, unpubl. data 49. L. wiegmannii‡ Andean 1335–2910 O Cei (1986) 50. L. melanops Eastern 900–2070 O Ramı´rez Pinilla (1991a) 51. L. rothi Eastern 500–1600 O Cei (1986) 52. L. abaucan Eastern 1200–1900 O Espinoza, unpubl. data 53. L. koslowskyi Eastern 800–2450 O Espinoza, unpubl. data 54. L. quilmes Andean 1600–3000 O Espinoza, unpubl. data 55. L. ornatus Andean 2000–4800 V Ramı´rez Pinilla (1991a) 56. L. albiceps Andean 3060–4020 V Espinoza, unpubl. data 57. L. irregularis Andean 3060–5000 V Ramı´rez Pinilla (1991a) 58. L. uspallatensis Eastern 900–2500 O Cei (1986) 59. L. chacoensis Eastern 700 O Cei (1986) 60. L. olongasta Eastern 900–1770 O Espinoza, unpubl. data 61. L. darwinii – Mendoza Eastern 1380 O? Espinoza, unpubl. data 62. L. darwinii – La Rioja Andean 0–3000 O Cei (1986) 63. L. boulengeri Eastern 0–2000 O Cei (1986) 64. L. laurenti Eastern 800–1100 O Cei (1993)

their study of the major clades within the Iguanidae (sensu Macey et al., 1997b), Phymaturus is well supported as the sister taxon and closest outgroup to Liolaemus. Phymaturus palluma is examined to compare with the previously published sequence of P. somuncurensis (Schulte et al., 1998). This same outgroup structure has been supported by morphological data (Etheridge & de Queiroz, 1988). Monophyly of the subfamily Tropidurinae∗ (sensu Macey et al., 1997b), of which Liolaemus and Phymaturus are members, is statistically neither supported nor rejected (Schulte et al., 1998); therefore, representative species from two other major iguanid clades are used. The previously published sequences (Macey et al., 1997b) of the Old World species, Oplurus cuvieri, and the New World species, Phrynosoma douglassii, are used as outgroups to represent a broad phylogenetic sampling of the Iguanidae.

METHODS

Specimen information

See Appendix 1 for museum numbers, localities of voucher specimens from which DNA was extracted, and GenBank accession numbers for DNA sequences. The sequence for Phymaturus somuncurensis (GenBank accession number AF049865, Schulte et al., 1998) originally had position 35 (Human position 4214, Anderson et al., 1981) in the ND1 gene as a T; however, the correct nucleotide is a G. The GenBank ﬁle has been updated. 80 J. A. SCHULTE ET AL. Laboratory protocols

Genomic DNA was extracted from liver or muscle using the Qiagen QIAamp tissue kit. Amplification of genomic DNA was conducted using a denaturation at 94°C for 35 sec, annealing at 50°C for 35 sec, and extension at 70°C for 150 sec with 4 sec added to the extension per cycle, for 30 cycles. Negative controls were run on all amplifications to check for contamination. Amplified products were purified on 2.5% Nusieve GTG agarose gels and reamplified under the conditions described above. Reamplified double-stranded products were purified on 2.5% acrylamide gels (Maniatis, Fritsch & Sambrook, 1982). Template DNA was eluted from acrylamide passively over three days with Maniatis elution buffer (Maniatis et al., 1982). Cycle-sequencing reactions were run using the Promega fmol DNA sequencing system with a denaturation at 95°C for 35 sec, annealing at 45–60°C for 35 sec, and extension at 70°C for 1 min for 30 cycles. Sequencing reactions were run on Long Ranger sequencing gels for 5–12 hours at 38–40°C. Two primer pairs were used to amplify genomic DNA from the ND1 gene to the COI gene: L3878 and H4980, and L4437 and H5934. Both strands were sequenced using L3878, L4221, H4419b, L4437, L4645, L4882a, L4882b, H5617b, L5638b, H5692, and H5934. DNA from L. multimaculatus was amplified also using the primer pairs L4437 and H5692, and L4882b and H5934, to yield smaller fragments of DNA. All primers are from Macey et al. (1997a) except L3878, which is from Macey et al. (1998b), and L4882b, which is from Schulte et al. (1998). Primer numbers refer to the 3′ end on the human mitochondrial genome (Anderson et al., 1981), where L and H denote extension of light and heavy strands, respectively.

Phylogenetic analysis

DNA sequences were aligned manually. Positions encoding part of ND1, all of ND2, and part of COI were translated to amino acids using MacClade (Maddison & Maddison, 1992) for confirmation of alignment. Alignments of sequences encoding tRNAs were constructed manually based on secondary structural models (Kumazawa & Nishida, 1993; Macey & Verma, 1997). Secondary structures of tRNAs were inferred from primary structures of the corresponding tRNA genes using these models. Unalignable regions were excluded from phylogenetic analyses (see Results). Phylogenetic trees were estimated using PAUP∗ beta version 4.0b1 (Swofford, 1998) with 100 heuristic searches featuring random addition of sequences. Bootstrap resampling (Felsenstein, 1985a) was applied to assess support for individual nodes using 500 bootstrap replicates with 10 random additions per replicate. Decay indices (=‘branch support’ of Bremer, 1994) were calculated for all internal branches of the tree using two methods. First, 100 heuristic searches (with random addition of sequences), which retained suboptimal trees one to five steps longer than the overall shortest trees, were conducted to obtain decay indices of one to five. Nodes that remained uncollapsed in the preceding analysis had decay indices above five. For each of these nodes, a phylogenetic topology containing the single node in question was constructed using MacClade (Maddison & Maddison, 1992) and analysed as a constraint in PAUP∗ beta version 4.0b1 (Swofford, 1998) with 100 heuristic searches featuring random addition of sequences. In these searches, trees that did not contain the imposed constraint were retained. We then obtained the LIOLAEMUS PHYLOGENETICS 81 decay index by tabulating the minimum increase in number of steps resulting from removal of the node of interest from the shortest tree. Wilcoxon signed-ranks tests (Felsenstein, 1985b; Templeton, 1983) were used to examine statistical significance of the shortest tree relative to alternative hypotheses. This test determines whether the most parsimonious tree is significantly shorter than an alternative tree or whether their differences in length are statistically indistinguishable (Larson, 1998). Wilcoxon signed-ranks tests were conducted as two-tailed tests (Felsenstein, 1985b). Tests were conducted using PAUP∗ beta version 4.0b1 (Swofford, 1998) which incorporates a correction for tied ranks. Alternative phylogenetic hypotheses were tested using the most parsimonious phylogenetic topologies compatible with them. To find the most parsimonious tree(s) compatible with a particular phylogenetic hypothesis, phylogenetic topologies were constructed using MacClade (Maddison & Maddison, 1992) and analysed as constraints using PAUP∗ beta version 4.0b1 (Swofford, 1998) with 100 heuristic searches with random addition of sequences. The evolution of viviparity among Liolaemus species was reconstructed using MacClade (Maddison & Maddison, 1992) on a strict consensus of the trees that are equally most parsimonious overall. One outgroup species, Phyrynosoma douglassii,is viviparous. However, most species within the Phrynosomatinae (sensu Macey et al., 1997b), excluding some nested species of Phrynosoma and Sceloporus, are oviparous (Meńdez-de la Cruz, Villagrań-Santa Cruz & Andrews, 1998; Montanucci, 1987; Reeder & Wiens, 1996; Schulte et al., 1998; Sites et al., 1992). Therefore, we treated the ancestral condition of the Phrynosoma outgroup as oviparous. Occurrence in cold environments (defined as latitudes exceeding 40 degrees south and/or elevations above 2500 m) was partitioned on the tree in a similar manner. Forty degrees south latitude corresponds to the division between Patagonia, and Pampas and Grand Chaco east of the Andes (Markgraf, 1993). West of the Andes, regions south of 40 degrees south latitude are dominated by steep mountains, and mixed coniferous and broadleaf forest (Markgraf, 1993). Data on latitudinal and elevational distributions for the 60 Liolaemus taxa (Table 1) were assembled from a combination of sources including (1) primary and secondary literature [with about one third of the data from Cei (1986, 1993), Veloso and Navarro (1988), and Nun˜ez, (1992)], (2) records of catalogued museum specimens, and (3) unpublished field notes (R.E.E.). Data for reproductive modes primarily came from Cei (1986, 1993), Donoso-Barros (1966), and Ramı´rez Pinilla (1991a, b), and through examination of catalogued museum specimens (R.E.E.).

RESULTS

Authentic mitochondrial DNA

Several observations support our conclusion that the DNA sequences analyzed here are from the mitochondrial genome and do not represent nuclear integrated copies of mitochondrial genes (see Zhang & Hewitt, 1996). All sequences reported here show strong strand bias against guanine on the light strand (G=11.8–13.0%, T=23.3–27.0%, A=33.2–36.1%, and C=25.1–30.3%), which is characteristic of the mitochondrial genome but not the nuclear genome (Macey et al., 1997a). Similar 82 J. A. SCHULTE ET AL. strand bias has been reported for other squamate lizards from this region of the mitochondrial genome (Macey et al., 1997a, b, 1998a,1999; Schulte et al., 1998). All genes sequenced appear functional; transfer RNA genes encode tRNAs that form stable secondary structures and protein-coding genes have no premature stop codons. Therefore, we interpret these sequences as authentic mitochondrial DNA.

Assessment of homology and sequence alignment

All regions in protein-coding genes are alignable. Gaps are placed in ND2 gene sequences at amino acid position 320 (positions 1259–1261) in L. lutzae. Gaps are placed in ND2 gene sequences at amino acid position 322 (positions 1265–1267) in L. multimaculatus, L. salinicola, L. scapularis, and L. wiegmannii. Among tRNA genes, several loop regions are unalignable as are noncoding regions between genes. The dihydrouridine (D) and part of the TwC (T) loops for the gene encoding tRNAIle (positions 100–107, 142–146), and the T and D-loops for the gene encoding tRNACys (positions 1604–1610, 1643–1645) are excluded from the analyses. The D-loop from the gene encoding tRNATrp (positions 1353–1362) is excluded from analyses. The loop of the origin for light-strand replication (OL, positions 1574–1582) between the tRNAAsn and tRNACys genes is not alignable and therefore not used for phylogenetic analysis. Noncoding sequences between the ND2 and tRNATrp genes (positions 1334–1339, including the ND2 stop codon), between the tRNATrp and tRNAAla genes (positions 1417–1418), and between the tRNACys and tRNATyr genes (positions 1659–1672) are not used. Three taxa have genes that appear to encode tRNAs with shifted T-stems in the tRNATyr gene. Liolaemus dorbignyi has tRNA position 65 (Kumazawa & Nishida, 1993) deleted (gap placed at alignment position 1681), and the L. darwinii population from Mendoza and L. olongasta have tRNA position 49 deleted (gap placed at alignment position 1697). Sequences from these taxa are aligned to the secondary structure observed in all other taxa for the T-stem of the tRNATyr gene. In the phylogenetic analysis, 64 of the 1774 aligned positions are found to have an ambiguous alignment. Excluded regions comprise less than 4% of the aligned sequence positions.

Genic variation

The 60 new mitochondrial DNA sequences range in size from 1736 to 1754 bases. These sequences are aligned with DNA sequences of the outgroups, Oplurus cuvieri and Phrynosoma douglassii, as well as the sequences from Liolaemus pictus (Macey et al., 1997b) and Phymaturus somuncurensis (Schulte et al., 1998) as 1774 positions (Appendix 2). All newly reported Liolaemus and Phymaturus sequences have a mitochondrial gene order of ND1, tRNAIle, tRNAGln, tRNAMet, ND2, tRNATrp, tRNAAla, Asn Cys Tyr tRNA ,OL (origin for light-strand replication), tRNA , tRNA , and COI. This gene order is typical for vertebrates and helps conﬁrm that iguanids do not contain a derived gene order in this region, in contrast to acrodont lizards (Macey et al., 1997a, b, 1998a). These aligned sequences contain 785 phylogenetically informative characters (parsimony criterion). Diﬀerent levels of variation are observed among the three protein-coding genes LIOLAEMUS PHYLOGENETICS 83

T 2. Distribution of phylogenetically informative and variable positions ND1 atRNAIle tRNAGln tRNAMet Codon positions 1st 2nd 3rd Stem Non-stem Stem Non-stem Stem Non-stem

Informative sites 13 5 27 5 5 14 4 2 5 Variable sites 15 8 27 10 5 22 9 2 6

ND2 atRNATrp tRNAAla tRNAAsn Codon positions 1st 2nd 3rd Stem Non-stem Stem Non-stem Stem Non-stem

Informative sites 178 92 322 19 6 10 4 15 8 Variable sites 208 130 340 23 6 12 8 22 10

atRNACys tRNATyr COI bNon- coding Codon positions regions Stem Non- Stem Non- 1st 2nd 3rd stem stem

Informative sites 7 7 15 13 1 1 5 2 Variable sites 12 8 21 16 2 2 6 4

Total Protein-coding tRNA bNon- All aligned coding sequence Codon positions regions 1st 2nd 3rd Stem Non-stem

Informative sites 192 98 354 87 52 2 785 Variable sites 225 140 373 124 68 4 934 a Not including excluded positions. b Noncoding region one is between the ND1 and tRNAIle genes and contains two variable positions. Noncoding region two is between the tRNAAla and tRNAAsn genes and contains one informative position. Noncoding region three is between the tRNATyr and COI genes and contains one informative position. and eight tRNA genes included in the analysis (Table 2). All eleven genes sequenced provide phylogenetically informative characters. Approximately 80% of the variation and phylogenetically informative sites are from protein-coding regions. First and second positions of codons provide 45% of the informative characters from protein- coding genes, and stem regions of tRNA genes provide over 60% of the phylogenetically informative characters from tRNA genes. The eight tRNA genes each provide between seven and 28 (mean 17) phylogenetically informative characters. In each of the three protein-coding genes, phylogenetically informative sites are found in ﬁrst, second, and third positions of codons. In all eight tRNA genes, phylogenetically informative sites are found in both stem and unpaired regions. Therefore, no single region dominates the phylogenetic analyses.

Phylogenetic analyses

Analyses of the DNA sequence data produce 60 equally most parsimonious trees, each with a length of 5858 steps (Figs 2, 3, and 4). Relative to Oplurus and Phrynosoma, monophyletic grouping of the tropidurine∗ (sensu Macey et al., 1997b) genera, Liolaemus and Phymaturus, is well supported (100% bootstrap, decay index 26). 84 J. A. SCHULTE ET AL.

Oplurus cuvieri Phrynosoma douglassii 100 Phymaturus palluma 29 Phymaturus somuncurensis

Clade 1 91 100 See Fig. 3 Subgenus 8 26 Liolaemus

100 42

Clade 2 96 See Fig. 4 Subgenus 11 Eulaemus

Figure 2. Strict consensus of 60 equally most parsimonious trees obtained from analysis of 1710 aligned base positions (785 phylogenetically informative) of mitochondrial DNA (length=5858, consistency index 0.271). Bootstrap values are presented above branches and decay indices are shown in bold below branches. The two clades recovered largely correspond to the subgeneric designations of Laurent (1984) with clade 1 representing the subgenus Liolaemus and clade 2 representing the subgenus Eulaemus. See Figures 3 and 4 for topologies of clades 1 and 2, respectively.

Monophyly of Phymaturus also is well supported (100% bootstrap, decay index 29) as is the monophyly of Liolaemus (100% bootstrap, decay index 42). Within Liolaemus, two major clades are found: one containing taxa from all three regions (Andes, eastern lowlands, and western lowlands; clade 1, subgenus Liolaemus; Fig. 2), and the other composed of Andean and eastern lowland taxa (clade 2, subgenus Eulaemus; Fig. 2). Both clades receive good support in the phylogenetic analysis (91% bootstrap, decay index 8, and 96% bootstrap, decay index 11, respectively). The ﬁrst dichotomy within clade 1 of Figure 2 (subgenus Liolaemus; Fig. 3) separates a well supported monophyletic group (99% bootstrap, decay index 15) from another poorly supported monophyletic group (decay index 2). Within the well supported group, L. coeruleus is the sister taxon to the remaining species, which are grouped with weak support (decay index 1). These remaining species form two well-supported clades, a primarily Andean clade (100% bootstrap, decay index 16) composed of L. alticolor, L. bitaeniatus, L. robertmertensi, L. bibronii, and L. gracilis, and a cosmopolitan clade (100% bootstrap, decay index 24) composed of L. bellii, L. chiliensis, L. cyanogaster, and L. pictus. Within these two groups, only the grouping of L. bibronii with L. gracilis, LIOLAEMUS PHYLOGENETICS 85

L. coeruleus 72 L. alticolor 99 100 1 L. bitaeniatus 15 16 69 L. robertmertensi 2 96 L. bibronii 8 L. gracilis 1 L. bellii

100 L. chiliensis See 24 L. cyanogaster Fig. 2 98 L. pictus-Neuquén 9 L. pictus-Rio Negro L. zapallarensis 71 L. tenuis 3 56 L. lemniscatus 3 60 100 L. monticola 2 97 19 L. nitidus 9 2 95 L. fuscus 11 L. nigroviridis L. capillitas 100 100 L. leopardinus 31 16 100 L. buergeri 55 15 L. ceii 3 L. petrophilus 100 L. austromendocinus 18 L. elongatus

Figure 3. Topology of clade 1 (subgenus Liolaemus) in Figure 2. Bootstrap values exceeding ﬁfty percent are presented above branches and decay indices are shown in bold below branches. See Figures 2 and 4 for topologies of major lineages and clade 2, respectively. and the clade comprising both L. pictus populations receive strong support (96% bootstrap, decay index 8, and 98% bootstrap, decay index 9, respectively). Within the weakly supported group from the primary dichotomy in clade 1 (Fig. 3), a mainly western lowland clade receives weak support (71% bootstrap, decay index 3), and its sister taxon, an Andean/eastern lowland clade, receives strong support (100% bootstrap, decay index 31). The primarily western lowland clade is composed of L. zapallarensis, L. tenuis, L. lemniscatus, L. monticola, L. nitidus, L. fuscus, and L. nigroviridis, with the last four species forming a well-supported clade (97% bootstrap, decay index 9). In addition, strong support is acquired for the sister-taxon relationships of L. monticola and L. nitidus (100% bootstrap, decay index 19), and L. fuscus and L. nigroviridis (95% bootstrap, decay index 11). The Andean/eastern lowland clade is composed of L. capillitas, L. leopardinus, L. buergeri, L. ceii, L. petrophilus, L. austromendocinus, and L. elongatus. Strong support is acquired for a clade containing L. leopardinus, L. buergeri, and L. ceii (100% bootstrap, decay index 16) and the sister- taxon relationship of the latter two species (100% bootstrap, decay index 15). In addition, the sister-taxon relationship of L. austromendocinus and L. elongatus is well supported (100% bootstrap, decay index 18). Clade 2 (subgenus Eulaemus) in Figure 2 is composed of taxa restricted to the Andes and eastern lowlands (Fig. 4). A monophyletic group (100% bootstrap, decay 86 J. A. SCHULTE ET AL.

100 L. lineomaculatus I 19 89 L. somuncurae 9 L. magellanicus L. ruibali 99 A II 11 97 L. andinus-La Rioja See 100 9 L. famatinae 23 Fig. 2 63 L. orientalis 2 95 L. dorbignyi 9 100 L. poecilochromus 26 100 L. multicolor 15 L. andinus-Jujuy 87 99 L. cuyanus 10 14 L. fitzingerii B L. pseudoanomalus 65 100 L. lutzae 1 7 17 L. occipitalis 97 L. multimaculatus 70 9 97 2 L. scapularis 6 13 53 L. salinicola 3 L. wiegmannii 100 L. melanops 16 L. rothi 86 L. abaucan 6 L. koslowskyi 100 1 L. quilmes 18 100 L. ornatus 57 15 100 L. albiceps 3 26 L. irregularis L. uspallatensis 61 L. chacoensis 2 98 L. olongasta 9 96 L. darwinii-Mendoza Roman numerals = Section 9 100 L. darwinii-La Rioja Letters = Series 23 L. boulengeri Number 1 = Clade L. laurenti

Figure 4. Topology of clade 2 (subgenus Eulaemus) in Figure 2. Bootstrap values exceeding fifty percent are presented above branches and decay indices are shown in bold below branches. Informal taxonomic sections (I=L. lineomaculatus section, II=L. montanus section), series (A=L. montanus series, B=L. boulengeri series), and the sand-lizard clade of the L. boulengeri series (1) are labelled to the right. See Figures 2 and 3 for topologies of major lineages and clade 1, respectively. index 19) containing southeastern lowland species (L. lineomaculatus, L. somuncurae, and L. magellanicus; recognized here as the L. lineomaculatus section) forms the sister taxon to a moderately supported group containing all remaining species of Eulaemus (recognized here as the L. montanus section; 87% bootstrap, decay index 10). Among these remaining species, a well-supported group (100% bootstrap, decay index 23) composed of Andean species from the Puna Plateau (recognized here as the L. montanus series of the L. montanus section) forms the sister group to a weakly supported clade (70% bootstrap, decay index 6) composed primarily of eastern lowland taxa (recognized here as the L. boulengeri series of the L. montanus section). Among taxa from the Puna Plateau, a clade comprising L. ruibali, L. andinus – La Rioja, and L. famatinae receives strong support (99% bootstrap, decay index 11) and forms the sister taxon to all other species (L. orientalis, L. dorbignyi, L. poecilochromus, L. multicolor, and L. andinus – Jujuy) from the Puna Plateau (63% bootstrap, decay index 2). All relationships within both groups from the Puna Plateau receive strong support (at least 95% bootstrap, decay index 9). Among species in the primarily eastern clade corresponding to the L. boulengeri series, three clades can be distinguished: (1) the eastern L. cuyanus and L. fitzingerii (99% bootstrap, decay index 14), (2) a primarily eastern sand-dwelling clade (65% LIOLAEMUS PHYLOGENETICS 87 bootstrap, decay index 7), and (3) remaining species occurring in both Andean and eastern lowland regions (86% bootstrap, decay index 6). In the eastern sand-dwelling clade, L. pseudoanomalus is the sister taxon to the other species (here called the ‘sand- lizard clade’ of the L. boulengeri series) with good support (97% bootstrap, decay index 9). Two well-supported clades are distinguished within the sand-lizard clade: (1) L. lutzae and L. occipitalis (100% bootstrap, decay index 17), and (2) monophyly of a group containing L. multimaculatus, L. scapularis, L. salinicola, and L. wiegmannii (97% bootstrap, decay index 13). However, relationships among the latter four species are not well supported. Within the clade composed of both Andean and eastern lowland species, L. melanops and L. rothi form a monophyletic group (100% bootstrap, decay index 16) whose sister taxon (100% bootstrap, decay index 18) contains all other species (L. abaucan, L. koslowskyi, L. quilmes, L. ornatus, L. albiceps, L. irregularis, L. uspallatensis, L. chacoensis, L. olongasta, L. darwinii, L. boulengeri, and L. laurenti). Two well-supported subgroups of the latter clade include (1) an Andean clade (100% bootstrap, decay index 15) containing L. ornatus as the sister taxon to a clade comprising L. albiceps and L. irregularis (100% bootstrap, decay index 26), and (2) a clade composed of L. chacoensis, L. olongasta, L. darwinnii, L. boulengeri, and L. laurenti (98% bootstrap, decay index 9). The last three species form a well-supported clade (96% bootstrap, decay index 9), with the Mendoza population of L. darwinii as the sister taxon to a clade composed of L. darwinii from La Rioja, L. boulengeri, and L. laurenti (100% bootstrap, decay index 23). Two populations are sampled for each of three different species. In clade 1 (subgenus Liolaemus), the two populations of L. pictus form a monophyletic group with 3.2% sequence divergence. In contrast, neither of the two species examined in clade 2 (subgenus Eulaemus) constitute monophyletic groups. In the L. montanus section, the two populations of L. andinus examined have 8.4% sequence divergence; L. andinus from La Rioja is the sister taxon to L. famatinae, and L. andinus from Jujuy is the sister taxon to L. multicolor. Liolaemus poecilochromus, formerly attributed to L. andinus, is recognized here as a full species and is the sister taxon to a clade containing L. andinus from Jujuy and L. multicolor. In the L. boulengeri series, L. darwinii from Mendoza is the sister taxon to a clade containing L. darwinii from La Rioja, L. boulengeri, and L. laurenti. The two L. darwinii populations show 4.9% sequence divergence. Taxonomic revision awaits further sampling of these species.

Evaluating Andean vicariance

Our molecular phylogenetic analysis rejects the hypothesis that Liolaemus comprises monophyletic groupings corresponding geographically to the Andes, the eastern lowlands, and the western lowlands (Figs 1 and 5). When a representative of the overall shortest trees (A in Appendix 3) is compared to a representative of the shortest alternative trees constraining Andean (B in Appendix 3), eastern-lowland (C in Appendix 3), or western-lowland (D in Appendix 3) species to form monophyletic groups, each alternative hypothesis is rejected in favor of the overall most parsimonious tree (Table 3; comparisons 1, 2, and 3, respectively). Therefore, our phylogenetic estimate requires multiple invasions of each of the three geographic regions investigated. 88 J. A. SCHULTE ET AL.

Phymaturus palluma Phymaturus somuncurensis L. coeruleus L. alticolor L. bitaeniatus L. robertmertensi L. bibronii L. gracilis L. bellii L. chiliensis L. cyanogaster L. pictus-Neuquén L. pictus-Rio Negro L. zapallarensis L. tenuis L. lemniscatus L. monticola L. nitidus L. fuscus L. nigroviridis L. capillitas L. leopardinus L. buergeri L. ceii L. petrophilus L. austromendocinus L. elongatus L. lineomaculatus L. somuncurae L. magellanicus L, ruibali L. andinus-La Rioja L. famatinae L. orientalis L. dorbignyi L. poecilochromus L. multicolor L. andinus-Jujuy L. cuyanus L. fitzingerii L. pseudoanomalus L. lutzae L. occipitalis L. multimaculatus L. scapularis L. salinicola L. wiegmannii L. melanops L. rothi L. abaucan L. koslowskyi L. quilmes L. ornatus L. albiceps L. irregularis L. uspallatensis L. chacoensis L. olongasta L. darwinii-Mendoza L. darwinii-La Rioja L. boulengeri L. laurenti

Figure 5. Phylogenetic reconstruction of Liolaemus distributions. The topology is after Figures 2, 3, and 4. Light branches represent Andean taxa; gray branches represent taxa that occur east of the Andes, and dark branches represent taxa that occur west of the Andes. Hatched branches are equivocal for occurrence in the Andes or the eastern or western lowlands. Liolaemus pictus occurs at low elevations on both sides of the Andes and its origins are uncertain.

Evolution of viviparity

We investigate multiple origins of viviparity, the reversibility of parity mode, and the hypothesis of viviparity being associated with cool environments. The strict consensus tree from the phylogenetic analysis depicts multiple origins of viviparity in Liolaemus (Fig. 6A). When the Wilcoxon signed-ranks test is used to compare a representative overall shortest tree (A in Appendix 3) to a representative alternative tree (E in Appendix 3) depicting a monophyletic grouping of all viviparous taxa, the latter hypothesis is signiﬁcantly longer and therefore rejected in favor of the overall shortest tree (Table 3; comparison 4). LIOLAEMUS PHYLOGENETICS 89

T 3. Results of Wilcoxon signed-ranks tests Hypothesis Treesa Nb Zc P-valued

1. Andean species form a A vs. B 420 16.65 0.001∗ monophyletic group 2. Eastern lowland species form A vs. C 375 15.45 0.001∗ a monophyletic group 3. Western lowland species A vs. D 225 8.38 0.001∗ form a monophyletic group 4. Viviparous species form a A vs. E 285 10.06 0.001∗ monophyletic group 5. Viviparous species in clade 1 A vs. F 191 6.53 0.001∗ form a monophyletic group 6. Viviparous species in clade 2 A vs. G 151 3.43 0.001∗ form a monophyletic group a See Appendix 3 for phylogenetic topologies used in tests. b Number of characters differing in minimum numbers of changes on paired topologies. c Normal approximation for Wilcoxon signed-ranks test. d Asterisks indicate a significant difference between the representative overall shortest tree (A) and a representative alternative tree (B-G) using the two-tailed probability for the Wilcoxon signed-ranks test (Felsenstein, 1985b; Templeton, 1983). A significant result denotes rejection of the stated hypothesis.

Each of the two major clades (subgenera Liolaemus and Eulaemus)ofLiolaemus contains oviparous and viviparous forms. The hypothesis that viviparous lineages form a monophyletic group is tested separately within each of these clades. A representative of the shortest alternative trees showing viviparous taxa forming a monophyletic group is compared to a representative of the overall shortest trees for each major clade. The Wilcoxon signed-ranks test rejects each of the alternative hypotheses (F and G in Appendix 3) in favor of the overall shortest tree (Table 3; comparisons 5 and 6, respectively). These results reject monophyly of viviparous lineages within both major clades of Liolaemus. To evaluate the evolutionary association between viviparity and cool environments, the most parsimonious reconstruction of occurrence at high elevation (above 2500 meters in elevation) and high latitude (exceeding 40 degrees south) is mapped onto the strict consensus of the 60 overall most parsimonious trees using MacClade (Maddison & Maddison, 1992) (Fig. 6B). All six origins of viviparity inferred under the assumption that viviparity is irreversible occur in cold environments, consistent with the hypothesis that viviparity is adaptive in these environments. Because a large portion of the phylogeny of Liolaemus occurs in cold environments, this association between viviparity and cold climates is not signiﬁcant using the concentrated changes test (Maddison, 1990). If reproductive mode is considered reversible, both gains and losses of viviparity occur in cold environments, a result that is equivocal with respect to the hypothesis that viviparity is adaptive in cold environments.

DISCUSSION

Historical biogeography

Phylogenetic relationships among Liolaemus species reject the hypothesis that species occurring in the Andes, in the eastern lowlands, and in the western lowlands 90 J. A. SCHULTE ET AL. Figure 6. Caption on facing page. LIOLAEMUS PHYLOGENETICS 91 comprise distinct geographic clades. The shortest phylogenetic estimate suggests multiple invasions of each of these geographic areas (Fig. 5). Liolaemus appears to have originated either in the low-elevation region east of the Andes or in the Andes. Clade 2 (subgenus Eulaemus) in Figures 2 and 4 is composed entirely of species from these two regions, whereas clade 1 (subgenus Liolaemus) in Figures 2 and 3 contains taxa from these regions and the lowlands west of the Andes. Both major clades require multiple origins of Andean species. Furthermore, clade 1 requires at least two independent origins of western lowland species from either the Andes or the low-elevation region east of the Andes, and at least one independent invasion of the eastern lowlands from the Andes. In clade 2, species occurring at latitudes higher than 40 degrees south on the east side of the Andes (L. lineomaculatus section in Table 1) form the sister taxon to all other species (L. montanus section) in this clade. Species that occur in the Andes on the Puna Plateau (L. montanus series) form the sister group to the rest of the species (L. boulengeri series) in this clade with the exception of species occurring at high latitudes (L. lineomaculatus section). Most of the remaining species (L. boulengeri series) in clade 2 occur east of the Andes and at least three separate invasions of the Andes have occurred among these species. Across the phylogenetic tree, at least fourteen evolutionary shifts in geographic distribution have occurred among species now occupying the three regions. The phylogenetic pattern of Liolaemus suggests recurring vicariance with subsequent dispersal between regions, making further vicariant events possible. The Andes have been uplifting continually for the last 25 million years (Norabuena et al., 1998) and Liolaemus populations have had numerous opportunities to become regionally isolated by vicariance. Subsequent dispersal between regions may have occurred either by climatic changes or shifts in life history, such as the development of viviparous reproduction; this dispersal then would permit further vicariance.

Dating phylogenetic divergence and vicariance

Accumulation of DNA substitutions between species can be used to date evolutionary divergences. Recent studies have shown that the segment of mitochondrial DNA used in this study is evolving at a rate of approximately 0.65% change per million years per lineage in other iguanian and gekkonid lizards, bufonid frogs, and

Figure 6. Parsimony reconstructions of viviparous reproduction and occurrence in cold climates (altitudes above 2500 m and/or latitudes exceeding 40 degrees south). The strict consensus tree as shown in Figures 2, 3, and 4 is presented. A, evolution of viviparity in Liolaemus. The outgroup, Phrynosoma douglassii, is coded as oviparous, a condition observed in most species within the Phrynosomatinae and considered ancestral for that taxon. Light branches represent oviparous lineages; dark branches are viviparous lineages, and hatched branches are lineages equivocal for parity mode. Clades numbered 1, 2, and 6 require independent origins of viviparity and, if viviparity is considered irreversible, clades 3, 4, and 5 each require an additional independent origin of viviparity. Arrows denote branches that potentially have losses of viviparity under the hypothesis that viviparity is reversible. B, parsimony reconstruction of species occurring in cold climates. Light branches represent lineages found below 2500 meters elevation and/or north of 40 degrees south latitude; dark branches denote lineages that have part or all of their distribution above 2500 meters elevation and/or exceeding 40 degrees south latitude, and hatched branches are lineages equivocal for occurrence in cold climates. Arrows and clade designations show evolutionary shifts of parity mode as described in part A. 92 J. A. SCHULTE ET AL.

300

250

200

150

100

Number of pairwise comparisons 50

0 1 234567891011121314 Estimated timing of divergence in million years

Figure 7. Plot of the number of pairwise comparisons among Liolaemus species and their inferred ages. For pairwise comparisons, 1.3% sequence divergence is expected to accumulate over one million years (Macey et al., 1998a). Note that less than two percent of pairwise comparisons are suggested to be younger than six million years. The average pairwise comparisons between the two major clades specify 12.6 million years divergence (marked by an arrow on the x axis) assuming a linear relationship between nucleotide substitutions and time. This divergence is most likely older than 12.6 million years because mitochondrial DNA is expected to show some degree of saturation beyond 10 million years of divergence (Moritz et al., 1987).

ﬁshes (Bermingham, McCaﬀerty & Martin, 1997; Macey et al., 1998a, b, 1999). The mean number of substitutions between all pairwise comparisons of taxa across the two major clades (Fig. 2) is 279, which is equivalent to 16.4% sequence divergence for any pair of lineages. This number corresponds to 8.2% change per lineage and predicts 12.6 million years divergence time between the subgenera using the calibration given above. Moritz, Dowling & Brown (1987) have suggested that mitochondrial DNA begins to saturate at ten million years, and that a linear relationship between nucleotide substitution and time is not expected beyond 10 million years. Therefore, our estimation of 12.6 million years is probably too low and perhaps this initial divergence event in Liolaemus corresponds to an earlier phase of the Andean uplift. The timing of additional divergence events may be inferred from pairwise comparisons. A plot of the number of pairwise comparisons between all Liolaemus sampled reveals that less than two percent of the divergence events post-date the end of the Miocene, approximately 6 million years ago (Harland et al., 1989; Fig. 7). This result suggests that the majority of divergence events studied here precede the Pliocene and Pleistocene climatic changes, further supporting our suggestion of vicariant divergence events caused by the Andean uplift with subsequent dispersal. Fossil evidence documents occurrence of Liolaemus in the Miocene (Albino, 1998). LIOLAEMUS PHYLOGENETICS 93 Evolution of parity mode

Phylogenetic analyses reveal multiple origins of viviparous reproduction among Liolaemus species. The transition from oviparity to viviparity occurs unequivocally three times in the parsimony reconstruction of viviparity (Fig. 6): two times in clade 1 (subgenus Liolaemus) and once in clade 2 (subgenus Eulaemus). In clade 1, viviparity evolves independently in L. nigroviridis and the ancestor of an unresolved group containing L. bellii, L. cyanogaster, and L. pictus. The latter group is compatible with although not supported by the molecular phylogenetic analysis, and could be resolved to require additional gains of viviparity. In clade 2, viviparity evolves independently in the ancestor of a high-elevation group composed of L. ornatus, L. albiceps, and L. irregularis. Further interpretation of the evolution of viviparity in this group depends on whether parity mode is considered reversible. If viviparity is irreversible, it has evolved at least six times in Liolaemus and once in the outgroup, Phymaturus.If viviparity is reversible and the ancestral condition for Liolaemus, at least three independent losses of viviparity with three subsequent changes back to viviparity are inferred for Liolaemus, plus an additional loss of viviparity is inferred for Phymaturus. Several recent studies have addressed the issue of the reversibility of parity mode in squamate reptiles (Benabib et al., 1997; Lee & Doughty, 1997; Lee & Shine, 1998), and agree that no undeniably strong evidence exists for irreversibility of viviparity. Although all six clades that are viviparous occupy cold environments at elevations above 2500 meters and/or latitudes exceeding 40 degrees south, evolutionary association between cool environments and origin of viviparity is equivocal. The association of six independent origins of viviparity with cold environments would be expected by chance alone in Liolaemus according to a concentrated changes test (Maddison, 1990), because a majority of the phylogenetic history of Liolaemus species is inferred to have occurred in cold environments. Furthermore, if viviparity is reversible in Liolaemus, any reversals also would be associated with cold environments and would reject the hypothesis that viviparity is an adaptation to cold climates. If further evidence rejects the hypothesis that viviparity is reversible, the results from Liolaemus analyzed in the broader context of squamate phylogeny may lend support to the hypothesis that viviparity arises primarily in cold environments as an adaptation to those conditions.

Phylogenetic relationships and taxonomy

Morphological data (Etheridge & de Queiroz, 1988; Etheridge, 1995; Frost & Etheridge, 1989) suggest that Liolaemus, Phymaturus, and the monotypic genus, Ctenoblepharys, together compose one of three major subgroups of the Tropidurinae∗ (sensu Macey et al., 1997b). Schulte et al. (1998) report that the Tropidurinae∗ appears nonmonophyletic, although monophyly cannot be rejected statistically. The relationship of Phymaturus and Liolaemus as sister taxa receives strong support both in our analysis and the analysis of Schulte et al. (1998). Phylogenetic analyses including DNA sequence for the monotypic genus Ctenoblepharys suggest that it is outside of the Phymaturus and Liolaemus clade (unpublished data). Therefore, we consider the monophyly of Liolaemus well corroborated. 94 J. A. SCHULTE ET AL. Several subgenera are recognized within Liolaemus (reviewed in Etheridge, 1995). Two of these subgenera, Liolaemus and Eulaemus (Laurent, 1992), are hypothesized to correspond largely to the two major clades recovered in this phylogenetic analysis. In addition, a subgeneric name is available for the ‘sand lizard’ clade, Ortholaemus (Girard, 1857; Laurent, 1984). Our phylogenetic analysis finds this group monophyletic and well supported (97% bootstrap, decay index 9). However, recognition of Ortholaemus would render Eulaemus paraphyletic. Another subgenus, Ceiolaemus (Laurent, 1984), represented in our sampling by L. pseudoanomalus, also is nested within Eulaemus. Hence, our phylogenetic analysis is most compatible with the recognition of only two subgenera, Liolaemus and Eulaemus, corresponding to the two major clades recovered. Etheridge (1995) recognizes numerous species groups within Liolaemus. The L. nitidus group (sensu Etheridge, 1995) is not monophyletic because L. chacoensis, L. lineomaculatus, L. magellanicus, and the species in clade 1 of figure 2 do not form a monophyletic group in our analysis. Clade 1 does correspond to the L. chiliensis group of Etheridge (1995) with the exception of our placing L. chacoensis in clade 2 as proposed by Laurent (1984). Our clade 2 corresponds to the L. signifer group of Etheridge (1995) with the exception of our inclusion of L. lineomaculatus, L. somuncurae, L. magellanicus, and L. chacoensis. We recover a monophyletic L. montanus group (Etheridge, 1995) except for our inclusion of L. chacoensis and L. pseudoanomalus.In our taxonomic sampling, species within Etheridge’s (1995) L. montanus group, excluding the L. boulengeri group, form a monophyletic group composed of L. ruibali, L. andinus – La Rioja, L. famatinae, L. orientalis, L. dorbignyi, L. poecilochromus, L. multicolor, and L. andinus – Jujuy. We find support for Etheridge’s (1995) L. boulengeri group except for our inclusion of L. chacoensis and L. pseudoanomalus. Nested within the L. boulengeri group is the L. wiegmannii group. Support is found for monophyly of the L. wiegmannii group, consisting of L. lutzae, L. occipitalis, L. multimaculatus, L. scapularis, L. salinicola, and L. wiegmannii in our sampling. The subgenus Liolaemus is largely synonymous with the L. chiliensis group of Etheridge (1995) and we recommend recognition of this subgenus. Because the L. nitidus group is not monophyletic, we recommend that this group not be recognized. The subgenus Eulaemus, as defined above, corresponds closely with the L. signifer group of Etheridge (1995) and use of the subgeneric name is recommended. The L. lineomaculatus group of Etheridge (1995) is monotypic, and inclusion of L. somuncurae and L. magellanicus in this group is supported by our analysis; we refer to this clade as the L. lineomaculatus section of the subgenus Eulaemus. The L. montanus and L. wiegmannii groups are found to be monophyletic, and the L. boulengeri group as defined above is monophyletic. We recognize Etheridge’s (1995) L. montanus group as a section of the subgenus Eulaemus, and Etheridge’s (1995) L. boulengeri group as a series of the L. montanus section. The clade found on the Puna Plateau is recognized as the L. montanus series of the L. montanus section. Because the L. wiegmannii group (sensu Etheridge, 1995) overlaps in content with the L. boulengeri series, we refer to the L. wiegmannii group as the ‘sand-lizard clade’ of the L. boulengeri series. A recent study uses behavioral characters associated with burial in sand to reconstruct the phylogeny of the L. boulengeri series (Halloy, Etheridge & Burghardt, 1998). This phylogenetic study uses outgroup taxa from the L. montanus series, which forms the monophyletic sister taxon to the L. boulengeri series in our phylogenetic analysis. The phylogenetic hypothesis presented in Halloy et al. (1998) suggests that L. ornatus is the sister taxon to all remaining members of the L. boulengeri series, LIOLAEMUS PHYLOGENETICS 95 followed by L. darwinii as the sister taxon to all other species except L. ornatus. Each of these species is deeply nested in our phylogenetic analysis with strong support. In addition, these taxa are not closely related, suggesting that the conflicts between phylogenetic reconstruction using the behavioral characters and molecular characters are not simply a difference in the position of the root. None of the branching patterns recovered from behavioral characters are compatible with our phylogenetic analysis. If the molecular phylogeny is correct, numerous parallelisms and reversals of behavioral characteristics associated with sand burial must have occurred in Liolaemus.

ACKNOWLEDGEMENTS

This work is supported by grants from the National Science Foundation (pre- doctoral fellowships to J.A.S. and J.R.M.; DEB-9318642 to J.B. Losos, K. de Queiroz, and A.L.; DEB-9726064 to A.L., J.R.M and Theodore J. Papenfuss), a Porter Fellowship from the American Physiological Society, and grants from the Explorers Club, the Society of Integrative and Comparative Biologists, the Chicago and Upstate [New York] Herpetological Societies, the American Society of Ich- thyologists and Herpetologists, and the Graduate School, Biology Department, and Biology Resources Research Center at the University of Nevada, Reno (to R.E.E). The third author thanks the following for permission to examine specimens in their care: Jose´ Navarro (DBCUCH), Raymond Laurent, Gustavo Scrocchi, Sonia Kretzschmar (FML), Alan Resetar and Harold Voris (FMNH), Herman Nun˜ez (MNHNC), and Richard Etheridge (SDSU). Tissue specimens were kindly provided by Carla Cicero, Fe´lix Cruz, Richard Etheridge, Fred da Rocha, Richard D. Sage, Laura E. Vega, David B. Wake, Jorge D. Williams, and Hussan Zaher. We thank Kevin de Queiroz, Richard Etheridge, and Jonathan B. Losos for valuable comments on an earlier draft of the manuscript, and Kraig Adler for consultation on taxonomic issues.

REFERENCES

Albino AM. 1998. Primer registro fo´sil de Liolaemus (Squamata: Iguania: Tropiduridae). Actas del VII Congreso Argentino de Paleontologıá y Biostratigrafıá (Bahıá, Blanca), p. 57. [Abstract]. Anderson S, Bankier AT, Barrell BG, de Bruijn MHL, Coulson AR, Drouin J, Eperon IC, Nierlich DP, Roe BA, Sanger F, Schreier PH, Smith AJH, Staden R, Young IG. 1981. Sequence and organization of the human mitochondrial genome. Nature 290: 457–465. Benabib M, Kjer KM, Sites JW Jr. 1997. Mitochondrial DNA sequence-based phylogeny and the evolution of viviparity in the Sceloporus scalaris group (Reptilia, Squamata). Evolution 51: 1262–1275. Bermingham E, McCafferty SS, Martin AP. 1997. Fish biogeography and molecular clocks: Perspectives from the Panamanian Isthmus. In: Kocher TD, Stepien CA, eds. Molecular Systematics of Fishes San Diego: Academic Press, 113–128. Bremer K. 1994. Branch support and tree stability. Cladistics 10: 295–304. Cei JM. 1986. Reptiles del centro, centro-oeste y sur de la Argentina. Herpetofauna de las zonas a´ridas y semia´ridas. Museo Regionale di Scienze Naturali Torino, Monografie 4: 1–527. Cei JM. 1993. Reptiles del noroeste, nordeste y este de la Argentina. Herpetofauna de las selvas subtropicales, Puna y Pampas. Bollettino Museo Regionale di Scienze Naturali–Torino, Monografie 14: 1–949. Creer DA, Kjer KM, Simmons DL, Sites JW Jr. 1997. Phylogenetic relationships of the Sceloporus scalaris complex (Squamata). Journal of Herpetology 31: 353–364. 96 J. A. SCHULTE ET AL.

Donoso-Barros R. 1966. Reptiles de Chile. Santiago: Universidad de Chile. Etheridge R. 1995. Redescription of Ctenoblepharys adspersa Tschudi, 1845, and the taxonomy of Liolaeminae (Reptilia: Squamata: Tropiduridae). American Museum Novitates 3142: 1–34. Etheridge R, de Queiroz K. 1988. A Phylogeny of Iguanidae. In: Estes R, Pregill G, eds. Phylogenetic Relationships of the Lizard Families, Essays Commemorating Charles L. Camp. Stanford: Stanford University Press, 283–368. Felsenstein J. 1985a. Confidence limits on phylogenies: an approach using the bootstrap. Evolution 39: 783–791. Felsenstein J. 1985b. Confidence limits on phylogenies with a molecular clock. Systematic Zoology 34: 152–161. Frost DR, Etheridge R. 1989. A phylogenetic analysis and taxonomy of iguanian lizards (Reptilia: Squamata). University of Kansas Museum of Natural History, Miscellaneous Publications 81: 1–65. Girard C. 1857. Descriptions of some new reptiles, collected by the United States Exploring Expedition, under the command of Capt. Charles Wilkes, U.S.N. Fourth Part. – Including the species of saurians, exotic to North America. Proceedings of the Academy of Sciences in Philadelphia 9: 195–199. Guillette LJ Jr. 1993. The evolution of viviparity in lizards. Bioscience 43: 742–751. Halloy M, Etheridge R, Burghardt GM. 1998. To bury in sand: Phylogenetic relationships among lizard species of the boulengeri group, Liolaemus (Reptilia: Squamata: Tropiduridae), based on behavioral characters. Herpetological Monographs 12: 1–37. Harland WB, Armstrong RL, Cox AV, Craig LE, Smith AG, Smith DG, 1989. A Geologic Time Scale 1989. Cambridge: Cambridge University Press. Kumazawa Y, Nishida M. 1993. Sequence evolution of mitochondrial tRNA genes and deep- branch animal phylogenetics. Journal of Molecular Evolution 37: 380–398. Larson A. 1998. The comparison of morphological and molecular data in phylogenetic systematics. In: DeSalle R, Schierwater B, eds. Molecular Approaches to Ecology and Evolution. Basel, Switzerland: Birkhaüser Verlag, 275–296. Laurent RF. 1984. On some iguanid genera related to or previously confused with Liolaemus Wiegmann. Journal of Herpetology 18: 357–373. Laurent RF. 1992. On some overlooked species of the genus Liolaemus Wiegmann (Reptilia Tro- piduridae) from Peru. Breviora 494: 1–33. Lee MSY, Doughty P. 1997. The relationship between evolutionary theory and phylogenetic analysis. Biological Reviews 72: 471–495. Lee MSY, Shine R. 1998. Reptilian viviparity and Dollo’s Law. Evolution 52: 1441–1450. Macey JR, Larson A, Ananjeva NB, Fang Z, Papenfuss TJ. 1997a. Two novel gene orders and the role of light-strand replication in rearrangement of the vertebrate mitochondrial genome. Molecular Biology and Evolution 14: 91–104. Macey JR, Larson A, Ananjeva NB, Papenfuss TJ. 1997b. Evolutionary shifts in three major structural features of the mitochondrial genome among iguanian lizards. Journal of Molecular Evolution 44: 660–674. Macey JR, Schulte II JA, Ananjeva NB, Larson A, Rastegar-Pouyani N, Shammakov SM, Papenfuss TJ. 1998a. Phylogenetic relationships among agamid lizards of the Laudakia caucasia species group: Testing hypotheses of biogeographic fragmentation and an area cladogram for the Iranian Plateau. Molecular Phylogenetics and Evolution 10: 118–131. Macey JR, Schulte II JA, Larson A, Fang Z, Wang Y, Tuniyev BS, Papenfuss TJ. 1998b. Phylogenetic relationships of toads in the Bufo bufo species group from the eastern escarpment of the Tibetan Plateau: A case of vicariance and dispersal. Molecular Phylogenetics and Evolution 9: 80–87. Macey JR, Wang Y, Ananjeva NB, Larson A, Papenfuss TJ. 1999. Vicariant patterns of fragmentation among gekkonid lizards of the genus Teratoscincus produced by the Indian collision: A molecular phylogenetic perspective and an area cladogram for Central Asia. Molecular Phylogenetics and Evolution 12: 320–332. Macey JR, Verma A. 1997. Homology in phylogenetic analysis: Alignment of transfer RNA genes and the phylogenetic position of snakes. Molecular Phylogenetics and Evolution 7: 272–279. Maddison WP. 1990. A method for testing the correlated evolution of two binary characters: Are gains or losses concentrated on certain branches of a phylogenetic tree? Evolution 44: 539–557. Maddison WP, Maddison DR. 1992. MacClade, analysis of phylogeny and character evolution, version 3.0. Sunderland, Mass: Sinauer. Maniatis T, Fritsch EF, Sambrook J. 1982. Molecular Cloning: A Laboratory Manual. Cold Spring Harbor, New York: Cold Spring Harbor Laboratory Press. LIOLAEMUS PHYLOGENETICS 97

Markgraf V. 1993. Climatic history of Central and South America since 18,000 yr B.P.: Comparison of pollen records and model simulations. In: Wright Jr. HE, Kutzbach JE, Webb III T, Ruddiman WF, Street-Perrott FA and Bartlein PJ, eds. Global Climates Since the Last Glacial Maximum. Minneapolis, Minnesota: University of Minnesota Press, 357–385. Meńdez-de la Cruz FR, Villagrań-Santa Cruz M, Andrews RM. 1998. Evolution of viviparity in the lizard genus Sceloporus. Herpetologica 54: 521–532. Mink DG, Sites JW Jr. 1996. Species limits, phylogenetic relationships, and origins of viviparity in the scalaris complex of the lizard genus Sceloporus (Phrynosomatidae: Sauria). Herpetologica 52: 551–571. Montanucci RR. 1987. A phylogenetic study of the horned lizards, genus Phrynosoma, based on skeletal and external morphology. Contributions of the Natural History Museum of Los Angeles County 113: 1–26. Moritz C, Dowling TE, Brown WM. 1987. Evolution of animal mitochondrial DNA: Relevance for population biology and systematics. Annual Review of Ecology and Systematics 18: 269–292. Norabuena E, Leffler-Griffin L, Mao A, Dixon T, Stein S, Sacks IS, Ocola L, Ellis M. 1998. Space geodetic observations of Nazca-South America convergence across the central Andes. Science 279: 358–362. Nun˜ ez H. 1992. Geographical data of Chilean lizards and snakes in the Museo Nacional de Historia Natural Santiago, Chile. Smithsonian Herpetological Information Service 91: 1–29. Ramı´rez Pinilla MP. 1991a. Estudio histolo´gico de los tractos reproductivos y actividad cıćlica anual reproductiva de machos y hembras de dos especies del geńero Liolaemus (Reptilia: Sauria: Iguanidae). Unpubl. Disser. Facultad de Ciencias Naturales, Universidad Nacional de Tucumań, Tucumań, Argentina. Ramı´rez Pinilla MP. 1991b. Variaciones histolo´gicas en los tractos reproductivos de machos de algunas especies de Liolaemus (Reptilia: Sauria: Iguanidae) en diferentes estados de actividad reproductiva. Revista Brasileira de Biologia 52: 133–140. Ramı´rez Pinilla MP. 1994. Reproductive and fat body cycles of the oviparous lizard Liolaemus scapularis. Journal of Herpetology 28: 521–524. Reeder TW, Wiens JJ. 1996. Evolution of the lizard family Phrynosomatidae as inferred from diverse types of data. Herpetological Monographs 10: 43–84. Rocha CFD. 1992. Reproductive and fat body cycles of the tropical sand lizard (Liolaemus lutzae)of southeastern Brazil. Journal of Herpetology 26: 17–23. Schulte JA II, Macey JR, Larson A, Papenfuss TJ. 1998. Molecular tests of phylogenetic taxonomies: A general procedure and example using four subfamilies of the lizard family Iguanidae. Molecular Phylogenetics and Evolution 10: 367–376. Scolaro JA, Cei JM. 1997. Systematic status and relationships of Liolaemus species of the archeforus and kingii groups: a morphological and taxonumerical approach (Reptilia: Tropiduridae). Bollettino Museo Regionale di Scienze Naturali Torino 15(2): 369–406. Shine R. 1985. The evolution of viviparity in reptiles: An ecological analysis. In: Gans C and Billet F, eds. Biology of the Reptilia. Volume 15. New York: Wiley, 605–694. Shine R, Bull JJ. 1979. The evoluation of live-bearing in lizards and snakes. American Naturalist 113: 905–923. Sites JW, Archie JW, Cole CJ, Villela OF. 1992. A review of phylogenetic hypotheses for the lizard genus Sceloporus (Phrynosomatidae): Implications for ecological and evolutionary studies. Bulletin of the American Museum of Natural History 213: 1–110. Swofford DL. 1998. PAUP∗. Phylogenetic Analysis Using Parsimony (∗ and Other Methods), Beta Version 4.0b1. Sunderland, Mass: Sinauer. Templeton AR. 1983. Phylogenetic inference from restriction endonuclease cleavage site maps with particular reference to the evolution of humans and the apes. Evolution 37: 221–244. Tinkle DW, Gibbons JW. 1977. The distribution and evolution of viviparity in reptiles. Miscellaneous Publications of the Museum of Zoology, University of Michigan 154: 1–55. Veloso A, Navarro J. 1988. Lista systeme´tica y distribucioń geogra´fica de anfı´bios y reptiles de Chile. Bollettino Museo Regionale Scienze Naturale Torino 6: 481–539. Zhang D-X, Hewitt GM. 1996. Nuclear integrations: Challenges for mitochondrial DNA markers. Trends in Ecology & Evolution 11: 247–251. 98 J. A. SCHULTE ET AL.

APPENDIX 1

Museum numbers and localities for voucher specimens from which DNA was obtained, and GenBank accession numbers are presented below in phylogenetic order (Table 1). Acronyms are FML for Fundacion Miguel Lillo, Tucuma´n, Argentina; MIC for the personal collection of Miguel I. Christie; MLP for Museo La Plata, La Plata, Argentina; MNRJ for Museu Nacional do Rio de Janeiro, Brazil; MVZ for the Museum of Vertebrate Zoology, University of California at Berkeley; MVZ-RDS for the personal collection of Richard D. Sage being deposited at the Museum of Vertebrate Zoology; MVZ-RM for the personal collection of the second author being deposited at the Museum of Vertebrate Zoology; REE for the personal collection of the third author; UFRGS for Universidade Federal do Rio Grande do Sul, Porto Alegre, Brazil; UNMdP for Universidad Nacional Mar del Plata, Mar del Plata, Argentina; SDSU for San Diego State University, San Diego.

Outgroup Oplurus cuvieri, Madagascar (MVZ-RM 10468, U82685; Macey et al., 1997b); Phrynosoma douglassii, 3.0 miles south of Hwy 60 on Water Canyon Campground Rd., Socorro Co., New Mexico (MVZ 180332, U82686; Macey et al., 1997b); Phymaturus palluma, 2768 m, 32°28′52″S69°09′59″W, 27 km NE Uspallata, Dpto. La Heras, Prov. Mendoza, Argentina (SDSU 3387, AF099216); Phymaturus somuncurensis, Meseta Somuncura´, Dpto. Rı´o Negro, Argentina (SDSU 1648, AF049865; Schulte et al., 1998).

Liolaemus, subgenus Liolaemus L. coeruleus, 1800 m, Portal de Atravesada, Dpto. Picunchea, Prov. Neuqueń, Argentina (SDSU 3692, AF099217); L. alticolor, 3360 m, 22°42′24.4″S65°43′12.4″W, 4.2 km W Abra Pampa on Ruta Prov. 71, Dpto. Cochinoca, Prov. Jujuy, Argentina (SDSU 3574, AF099218); L. bitaeniatus, 1860 m, 10 km E Tafı´ del Valle on Ruta Prov. 307, Dique de la Angostura, Dpto. Tafı´ del Valle, Prov. Tucumań, Argentina (SDSU 3568, AF099219); L. robertmertensi, 690 m, 6.6 km W Ruta 46, 33.1 km S Andalgala´, Dpto. Pomań, Prov. Catamarca, Argentina (SDSU 3498, AF099220); L. bibronii, 1440 m, 34 km W Bardas Blancas on Ruta Prov. 224, Dpto. Malarguë, Prov. Mendoza, Argentina (SDSU 3407, AF099221); L. gracilis, 1380 m, 35°02′36.2″S68°40′42.2″W, SE beach of Embalse Nihuı´l, Dpto. San Rafael, Prov. Mendoza, Argentina (SDSU 3409, AF099222); L. bellii, 2530 m, near Farellones, Reg. Santiago, Chile (SDSU 3719, AF099223); L. chiliensis, 1 km N El Salitral on Ruta Nac. 40, Dpto. Collon Cruz, Prov. Neuqueń, Argentina (MIC 1259, AF099224); L. cyanogaster, Puyehue, N¯ ilque Costas, Prov. Neuqueń, Argentina (MIC 1241, AF099225); L. pictus – Neuqueń, 1475 m, 3 km W Ardillas, Dpto. Los Lagos, Prov. Neuqueń, Argentina (MIC 1108, AF099226); L. pictus – Rio Negro, Bariloche, 44 km west at Rio Castanõ Overo, Dpto. Rio Negro, Argentina (MVZ 162076, U82684; Macey et al., 1997b); L. zapallarensis, no locality data (SDSU 3658, AF099227); L. tenuis, near El Ingenio, about 30 km SE Santiago, Reg. Santiago, Chile (SDSU 3723, AF099228); L. lemniscatus, near El Ingenio about 30 km SE Santiago, Reg. Santiago, Chile (SDSU 3721, AF099229); L. monticola, near El Ingenio, about 30 km SE Santiago, Reg. Santiago, Chile (SDSU 3724, AF099230); L. nitidus, near El Ingenio about 30 km SE Santiago, Reg. Santiago, Chile (SDSU 3720, AF099231); L. fuscus, 1 km NE El Romeral (near Ocoa), Prov. Quillota, Chile (MVZ-RDS 12818, AF099232); L. nigroviridis, 2530 m, near Farellones, Reg. Santiago, Chile (SDSU 3715, AF099233); L. capillitas, 2825 m, 27°26′51.6″S 66°24′40.5″W, 22 km S Mina Capillitas on Ruta Prov. 47, Dpto. Andalgala´, Prov. Catamarca, Argentina, (SDSU 3481, AF099234); L. leopardinus, 2530 m, near Farellones, Reg. Santiago, Chile (SDSU 3717, AF099235); L. buergeri, 2070 m, 19 km N intersect. Ruta Prov. 221 & Ruta Nac. 40 on Ruta 221, Dpto. Malarguë, Prov. Mendoza, Argentina (SDSU 3420, AF099236); L. ceii, Copahue, Dpto. N˜ orquin, Prov. Neuqueń, Argentina (MIC 1139, AF099237); L. petrophilus, Meseta Somuncura´, Prov. Rio Negro, Argentina (MLP 1671, AF099238); L. austromendocinus, 2310 m, 48 km S intersect. Ruta Prov. 221 & Ruta Nac. 40 on 221, Dpto. Malarguë, Prov. Mendoza, Argentina (SDSU 3425, AF099239); L. elongatus, 2768 m, 27 km NE Uspallata, Dpto. La Heras, Prov. Mendoza, Argentina (SDSU 3459, AF099240).

Subgenus Eulaemus, L. lineomaculatus section L. lineomaculatus, Laguna LaPlata, Prov. Chubut, Argentina (MLP 1670, AF099241); L. somuncurae, Laguna Raimundo, Meseta Somuncura´, Prov. Rı´o Negro, Argentina (MLP 1661, AF099242); L. magellanicus, 67 m, lighthouse at Cabo Espiritu Santo, 70 km E Cerro Sombrero, Tierra del Fuego, Magallanes Region, Chile (MVZ 180141, AF099243). LIOLAEMUS PHYLOGENETICS 99

L. montanus section, L. montanus series L. ruibali, 2550 m, gravel ﬂats S of Tocota, Dpto. Iglesia, Prov. San Juan, Argentina (SDSU 3455, AF099244); L. andinus-La Rioja, 3177 m, 28°48′11.3″S68°45′49.4″W, Agua Quemada, Reserva Laguna Brava, Prov. La Rioja, Argentina (REE 265, AF099245); L. famatinae, >3600 m, 29°49′50″S 67°43′30″W, Cueva de Perez, Sierra de Famatina, Dpto. Famatina, Prov. La Rioja, Argentina (REE 193, AF099246); L. orientalis, 4300 m, 15.5 km E Orosmayo on Ruta Prov. 70, Dpto. Rinconada, Prov. Jujuy, Argentina (SDSU 3517, AF099247); L. dorbignyi, 4320 m, 5.2 km E Olacapato on Ruta Nac. 51, Dpto. Los Andes, Prov. Salta, Argentina (SDSU 3443, AF099248); L. poecilochromus, 4130 m, 11.1 km E Olacapato on Ruta Nac. 51, Dpto. Los Andes, Prov. Salta, Argentina (SDSU 3593, AF099249); L. multicolor, 3360 m, 4.2 km W Abra Pampa on Ruta Prov. 71, Dpto. Cochinoca, Prov. Jujuy, Argentina (SDSU 3591, AF099250); L. andinus-Jujuy, 3390 m, 23°21′47.2″S65°47′42.3″W, 16.7 km N turnoﬀ to Emp. a los Colorados, 75.7 km S Abra Pampa on Ruta Nac. 40, Dpto Cochinoca, Prov. Jujuy, Argentina (SDSU 3599, AF099251).

L. boulengeri series L. cuyanus, 1530 m, 32°07′07″S67°35′52″W, Los Medanitos, Dpto. Tinogasta, Prov. Catamarca, Argentina (SDSU 3541, AF099252); L. fitzingerii, Prov. Santa Cruz, Argentina (FML not cat., AF099253); L. pseudoanomalus, 990 m, 28°50′27.3″S67°24′49.2″W, 32 km S Pituı´l, 3.7 km E Ruta Nac. 40 on road to Antinaco, Dpto. Famatina, Prov. La Rioja, Argentina (SDSU 3539, AF099254); L. lutzae, Bahıáde Sepetiba, Restinga de Marambala, Estado de Rio de Janeiro, Brazil (MNRJ 4720, AF099255); L. occipitalis, sea level, 29°58′S50°08′W. Tramandaı´, Estado de Rio Grande do Sul, Brazil (UFRGS 2753, AF099256); L. multimaculatus, Costa Bonita, Partido de Necochea, Prov. Buenos Aires, Argentina (UNMdP 407, AF099257); L. scapularis, 1650 m, 26°04′00″S65°54′38″W, edge of dunes at “Los Medanos,” 6.7 km E intersection of Rutas Nac. 68 and 40, Dpto. Cafayate, Prov. Salta, Argentina (SDSU 3560, AF099258); L. salinicola, 1530 m, 32°07′00″S–67°45′02″W, Los Medanitos, Dpto. Tinogasta, Prov. Catamarca, Argentina (SDSU 3531, AF099259); L. wiegmannii, 1565 m, Sierra Medina, 18 km W intersection Rutas Prov. 310 and 306, Dpto. Burrayacu, Prov. Tucumań, Argentina (SDSU 3494, AF099260); L. melanops, 1000 m, 7 km N Zapala on Ruta 231, Dpto. Zapala, Prov. Neuqueń, Argentina (SDSU 3704, AF099261); L. rothi, Laguna Raimundo, Meseta Somuncura´, Prov. Rıó Negro, Argentina (MLP 1662, AF099262); L. abaucan, 1650 m, 27°47′05.6″S67°39′26.8″W, 10.8 km S Fiambala´ on Ruta Prov. 45, Dpto. Tinogasta, Prov. Catamarca, Argentina (SDSU 3532, AF099263); L. koslowskyi, 1290 m, 28°31′55.1″S–67°21′46.5″W, 8.2 km E Pituı´l, Dpto. Famatina, Prov. La Rioja, Argentina (SDSU 3598, AF099264); L. quilmes, 1770 m, 26°28′04″S66°02′16″W, 8 km W Ruta 40, Ruinas de Quilmes, Dpto. Tafı´ del Valle, Prov. Tucumań, Argentina (SDSU 3558, AF099265); L. ornatus, 3360 m, 22°42′24″S65°43′12″W, 4.2 km W Abra Pampa, Dpto. Cochinoca, Prov. Jujuy, Argentina (SDSU 3521, AF099266); L. albiceps, 4020 m, 24°20′20″S66°13′13.2″W, 9.4 km S intersection Ruta Nac. 40 and 51 on Ruta 40, Dpto. Los Andes, Prov. Salta, Argentina (SDSU 3380, AF099267); L. irregularis, 3450 m, 24°02′32.8″S–66°16′21.6″W, 18.3 km N Pueblo Nuevo, N of San Antonio de los Cobres, Dpto. Los Andes, Prov. Salta, Argentina (SDSU 3525, AF099268); L. uspallatensis, 2370 m, 32°28′39.1″S69°13′55.8″W, 20 km NE Uspallata, Dpto. Las Heras, Prov. Mendoza, Argentina (SDSU 3465, AF099269); L. chacoensis, Prov. Santiago del Estero, Argentina (FML 3640, AF099270); L. olongasta, 1700 m, 28°39′25.7″S68°20′11.4″W, sand dunes just west of Bajo Jaguë´ on Ruta Prov. 26, Dpto. Vinchina, Prov. La Rioja, Argentina (SDSU 3546, AF099271); L. darwinii-Mendoza, 1380 m, 35°02′36.2″S68°40′42.2″W, S shore of Embalse Nihuı´l, Dpto. San Rafael, Prov. Mendoza, Argentina (SDSU 3472, AF099272); L. darwinii-La Rioja, 900 m, 30°12′20″S67°20′26.6″W, 51 km W Patquia on Ruta Nac. 160, Dpto. Independencia, Prov. La Rioja, Argentina (SDSU 3477, AF099274); L. boulengeri, 1380 m, 35°02′36″S68°40′42″W, S shore Embalse Nihuı´l, Dpto. San Rafael, Prov. Mendoza, Argentina (SDSU 3469, AF099275); L. laurenti, 1020 m, 28°16′00″S67°25′01.9″W, 20.2 km E Co- pacabana on Ruta Nac. 60, Dpto. Tinogasta, Prov. Catamarca, Argentina (SDSU 3530, AF099273).

APPENDIX 2

Sequence alignment

Gap positions are given so the alignment used in the phylogenetic analysis can be reconstructed. The first number given is the base position before a gap position and the second number in parentheses is 100 J. A. SCHULTE ET AL. the number of gaps placed in a particular position. If only a single gap is placed at a particular position, no second number in parentheses is given. Base positions (first number) correspond to numbering in GenBank accessions which begin with position 4180 and end with position 5933 of the human mitochondrial genome (Anderson et al., 1981). Taxon names appear as in Table 1. Oplurus cuvieri: 83(3), 103, 141, 240, 1331(2), 1350(4), 1386(5), 1400, 1470, 1522, 1589, 1597, 1622, 1637(12). Phrynosoma douglassii: 85, 104(2), 155, 211, 240, 1324(3), 1325(5), 1344(4), 1370, 1375(2), 1396, 1519, 1557(2), 1582(3), 1615(2), 1629(13). Phymaturus palluma: 78(3), 103, 141, 210, 239, 1327(5), 1346(4), 1372, 1378, 1380, 1398, 1468, 1520, 1559, 1586, 1594, 1634(13). Phymaturus somuncurensis: 78(3), 103, 211, 240, 1321(3), 1325(5), 1344(4), 1370, 1376, 1378, 1396, 1466, 1518, 1557, 1584, 1592, 1617, 1630(14). L. coeruleus: 83(3), 103, 141, 210, 239, 1328(4), 1347(4), 1373, 1379, 1381, 1399, 1469, 1521, 1559(2), 1586, 1594, 1618(2), 1643(2). L. alticolor: 83(3), 103, 141, 210, 239, 1328(4), 1349(2), 1375, 1381, 1383, 1401, 1471, 1523, 1561(2), 1597, 1621(2), 1643(5). L. bitaeniatus: 83(3), 103, 141, 210, 239, 1327(5), 1348(2), 1374, 1380, 1382, 1400, 1470, 1522, 1560(2), 1596, 1620(2), 1643(4). L. robertmertensi: 83(3), 103, 141, 210, 239, 1328(4), 1349(2), 1375, 1381, 1383, 1401, 1471, 1523, 1562, 1589, 1597, 1644(6). L. bibronii: 83(3), 103, 141, 210, 239, 1328(4), 1349(2), 1375, 1381, 1383, 1401, 1471, 1523, 1562, 1589, 1597, 1647(3). L. gracilis: 83(3), 103, 141, 210, 239, 1328(4), 1349(2), 1375, 1381, 1383, 1401, 1471, 1523, 1562, 1598. L. bellii: 83(3), 103, 141, 210, 239, 1328(4), 1347(4), 1373, 1379, 1381, 1399, 1469, 1521, 1559(2), 1595, 1619(2), 1644(2). L. chiliensis: 83(3), 103, 141, 210, 239, 1328(4), 1347(4), 1373, 1379, 1381, 1399, 1469, 1521, 1559(2), 1586, 1594, 1618(2), 1643(2). L. cyanogaster: 83(3), 103, 141, 210, 239, 1327(5), 1348(2), 1374, 1380, 1382, 1400, 1470, 1522, 1561, 1597, 1621(2), 1646(2). L. pictus – Neuqeń: 83(3), 103, 141, 210, 239, 1328(4), 1349(2), 1375, 1381, 1383, 1401, 1471, 1523, 1562, 1589, 1597, 1621(2), 1646(2). L. pictus – Rio Negro: 83(3), 103, 141, 210, 239, 1328(4), 1349(2), 1375, 1381, 1383, 1401, 1471, 1523, 1561(2), 1588, 1596, 1620(2), 1645(2). L. zapallarensis: 83(3), 103, 141, 210, 239, 1353(2), 1379, 1385, 1387, 1405, 1475, 1527, 1565(2), 1592, 1600, 1624(2), 1643(8). L. tenuis: 83(3), 103, 141, 210, 239, 1354, 1380, 1386, 1388, 1477, 1529, 1568, 1594(2), 1602, 1627, 1644(10). L. lemniscatus: 83(3), 103, 141, 210, 239, 1354, 1380, 1386, 1388, 1477, 1529, 1567(2), 1594, 1602, 1626(2), 1642(11). L. monticola: 83(3), 103, 141, 210, 239, 1381, 1387, 1389, 1478, 1530, 1568(2), 1594(2), 1602, 1626(2), 1650(3). L. nitidus: 83(3), 103, 211, 240, 1382, 1388, 1390, 1479, 1531, 1569(2), 1596, 1604, 1628(2), 1652(3). L. fuscus: 83(3), 103, 141, 210, 239, 1354, 1380, 1386, 1388, 1477, 1529, 1567(2), 1603, 1627(2), 1649(5). L. nigroviridis: 83(3), 103, 141, 210, 239, 1354, 1380, 1386, 1388, 1477, 1529, 1567(2), 1603, 1627(2), 1650(4). L. capillitas: 83(3), 103, 141, 210, 239, 1327(5), 1349, 1375, 1381, 1383, 1472, 1524, 1562(2), 1589, 1597, 1621(2), 1643(5). L. leopardinus: 83(3), 103, 141, 210, 239, 1327(5), 1349, 1375, 1381, 1383, 1472, 1524, 1563, 1590, 1598, 1622(2), 1643(6). L. buergeri: 83(3), 103, 141, 210, 239, 1327(5), 1376, 1382, 1384, 1473, 1525, 1563(2), 1599, 1623(2), 1645(5). L. ceii: 83(3), 103, 141, 210, 239, 1327(5), 1349, 1375, 1381, 1383, 1472, 1524, 1562(2), 1589, 1597, 1621(2), 1643(5). L. petrophilus: 83(3), 103, 141, 210, 239, 1327(5), 1349, 1375, 1381, 1383, 1472, 1524, 1562(2), 1589, 1597, 1621(2), 1643(5). L. austromendocinus: 83(3), 103, 141, 210, 239, 1327(5), 1349, 1375, 1381, 1383, 1472, 1524, 1562(2), 1589, 1597, 1621(2), 1643(5). L. elongatus: 83(3), 103, 141, 210, 239, 1327(5), 1349, 1375, 1381, 1383, 1472, 1524, 1562(2), 1589, 1597, 1621(2), 1643(5). L. lineomaculatus: 83(3), 103, 141, 210, 239, 1353(2), 1379, 1385, 1387, 1476, 1528, 1567, 1594, 1602, 1626(2), 1642(11). L. somuncurae: 83(3), 103, 141, 210, 239, 1351(4), 1377, 1383, 1385, 1403, 1472(2), 1524, 1562(2), 1589, 1597, 1621(2), 1636(12). L. magellanicus: 83(3), 103, 141, 210, 239, 1351(4), 1377, 1383, 1385, 1403, 1473, 1525, 1563(2), 1599, 1624, 1641(10). L. ruibali: 83(3), 103, 141, 210, 239, 1353(2), 1379, 1385, 1387, 1476, 1528, 1566(2), 1593, 1601, 1625(2), 1642(10). L. andinus – La Rioja: 83(3), 103, 141, 210, 239, 1354, 1380, 1386, 1388, 1477, 1568(2), 1595, 1603, 1627(2), 1645(9). L. famatinae: 83(3), 103, 141, 210, 239, 1354, 1380, 1386, 1388, 1477, 1529, 1568, 1595, 1603, 1627(2), 1645(9). L. orientalis: 83(3), 142, 211, 240, 1332, 1354, 1380, 1386, 1388, 1477, 1529, 1567(2), 1594, 1602, 1626(2), 1643(10). L. dorbignyi: 83(3), 142, 211, 240, 1355, 1381, 1387, 1389, 1478, 1530, 1568(2), 1595, 1603, 1627(2), 1644(10), 1652. L. poecilochromus: 83(3), 103, 141, 210, 239, 1328(4), 1350, 1376, 1382, 1384, 1473, 1525, 1564, 1591, 1599, 1623(2), 1641(9). L. multicolor: 83(3), 103, 141, 210, 239, 1328(4), 1350, 1376, 1382, 1384, 1473, 1525, 1564, 1591, 1599, 1624, 1642(9). L. andinus – Jujuy: 83(3), 103, 141, 210, 239, 1328(4), 1350, 1376, 1382, 1384, 1473, 1525, 1564, 1591, 1599, 1624, 1642(9). L. cuyanus: 83(3), 103, 141, 210, 239, 1330(2), 1352, 1378, 1384, 1386, 1475, 1527, 1565(2), 1592, 1600, 1625, 1649(3). L. fitzingerii: 83(3), 103, 141, 210, 239, 1330(2), 1352, 1378, 1384, 1386, 1475, 1527, 1565(2), 1592, 1600, 1625, 1642(10). L. pseudoanomalus: 83(3), 103, 141, 210, 1328(5), 1347(4), 1373, 1379, 1381, 1419, 1469, 1521, 1559(2), 1586, 1594, 1619, 1641(5). L. lutzae: 83(3), 103, 141, 210, 239, 1251(3), 1327(2), 1349, 1375, 1381, 1383, 1415, 1471, 1523, 1561(2), 1588, 1596, 1621, 1637(11). L. occipitalis: 83(3), 103, 141, 210, 239, 1330(2), 1352, 1378, LIOLAEMUS PHYLOGENETICS 101

1384, 1386, 1418, 1474, 1526, 1564(2), 1591, 1599, 1624, 1640(11). L. multimaculatus: 83(3), 103, 141, 210, 239, 1257(3), 1327(2), 1349, 1375, 1381, 1383, 1472, 1499, 1523, 1561(2), 1588, 1596, 1621, 1638(10). L. scapularis: 83(3), 142, 211, 240, 1258(3), 1328(2), 1350, 1376, 1382, 1384, 1473, 1525, 1563(2), 1599, 1624, 1640(11). L. salinicola: 83(3), 103, 141, 210, 239, 1257(3), 1327(2), 1349, 1375, 1381, 1383, 1472, 1524, 1562(2), 1598, 1623, 1638(12). L. wiegmannii: 83(3), 103, 141, 210, 1258(3), 1325(5), 1347, 1373, 1379, 1381, 1470, 1522, 1560(2), 1587, 1595, 1620, 1637(10). L. melanops: 83(3), 103, 141, 210, 239, 1354, 1380, 1386, 1388, 1426, 1476, 1528, 1566(2), 1602, 1627, 1640(14). L. rothi: 83(3), 103, 141, 210, 239, 1354, 1380, 1386, 1388, 1477, 1529, 1567(2), 1594, 1602, 1627, 1644(10). L. abaucan: 83(3), 103, 141, 210, 239, 1354, 1380, 1386, 1388, 1477, 1529, 1567(2), 1603, 1628, 1645(10). L. koslowskyi: 83(3), 102(2), 140, 209, 238, 1353, 1379, 1385, 1387, 1476, 1528, 1566(2), 1602, 1627, 1644(10). L. quilmes: 83(3), 102(2), 140, 209, 238, 1353, 1379, 1385, 1387, 1476, 1528, 1566(2), 1602, 1627, 1644(10). L. ornatus: 83(3), 102(2), 140, 209, 238, 1353, 1379, 1385, 1387, 1476, 1528, 1566(2), 1602, 1627, 1644(10). L. albiceps: 83(3), 102(2), 140, 209, 238, 1353, 1379, 1385, 1387, 1476, 1528, 1566(2), 1602, 1627, 1644(10). L. irregularis: 83(3), 102(2), 140, 209, 238, 1353, 1379, 1385, 1387, 1476, 1528, 1566(2), 1602, 1627, 1644(10). L. uspallatensis: 83(3), 102(2), 140, 209, 238, 1353, 1379, 1385, 1387, 1476, 1528, 1566(2), 1602, 1627, 1644(10). L. chacoensis: 83(3), 103, 141, 210, 239, 1354, 1380, 1386, 1388, 1477, 1529, 1567(2), 1603, 1628, 1645(10). L. olongasta: 83(3), 103, 141, 210, 239, 1354, 1380, 1386, 1388, 1477, 1529, 1567(2), 1603, 1628, 1645(10), 1669. L. darwinii-Mendoza: 83(3), 103, 141, 210, 239, 1354, 1380, 1386, 1388, 1477, 1529, 1567(2), 1603, 1628, 1644(11), 1668. L. darwinii-LaRioja: 83(3), 103, 141, 210, 239, 1354, 1380, 1386, 1388, 1477, 1529, 1567(2), 1603, 1628, 1644(11). L. boulengeri: 83(3), 103, 141, 210, 239, 1354, 1380, 1386, 1388, 1477, 1529, 1567(2), 1603, 1628, 1644(11). L. laurenti: 83(3), 103, 141, 210, 239, 1354, 1380, 1386, 1388, 1477, 1529, 1567(2), 1603, 1628, 1644(11).

APPENDIX 3

Trees used in Templeton tests

Numbers correspond to species and populations in Table 1. Lengths of trees and consistency indices (Swoﬀord, 1998) are given in parentheses. A representative shortest tree from analysis of the mitochondrial DNA is shown as A, and B-G are the shortest alternative trees recovered from phylogenetic analyses imposing constraints. A. Representative of the 60 equally most parsimonious trees from analysis of the DNA sequence data (5858 steps, CI 0.271): (1, (2, ((3, 4), (((5, (((6, 7), (8, (9, 10))), (((11, 12), (14, 15)), 13))), ((16, (17, (18, ((19, 20), (21, 22))))), (23, (((24, (25, 26)), (28, 29)), 27)))), ((30, (31, 32)), (((33, (34, 35)), (36, (37, (38, (39, 40))))), ((41, 42), ((43, ((44, 45), ((46, 47), (48, 49)))), ((50, 51), (52, (((53, 54), (55, (56, 57))), (58, (59, (60, (61, ((62, 64), 63)))))))))))))))). B. Representative of the 24 equally most parsimonious trees constrained to show a monophyletic grouping of Andean species (6465 steps, CI 0.246): (1, (2, ((3, 4), ((((5, (10, ((12, (14, 15)), 13))), ((26, 27), 28)), (16, (17, (18, (19, 21))))), ((((((((((((((((((((6, 7), (8, 9)), 11), ((20, 22), (23, ((24, 25), 29)))), (((33, (34, 35)), (37, (38, (39, 40)))), 36)), 49), (54, (55, (56, 57)))), 61), 62), 64), 63), 60), 59), 58), 53), 52), (50, 51)), (43, ((44, 45), ((46, 47), 48)))), (41, 42)), (30, (31, 32))))))). C. Representative of the 6 equally most parsimonious trees constrained to show a monophyletic grouping of eastern lowland species (6326 steps, CI0.251): (1, (2, ((3, 4), ((((((((((((((((((5, 10), ((26, 27), 28)), (30, (31, 32))), ((41, 42), (43, ((44, 45), ((46, 47), 48))))), (50, 51)), 52), 53), 58), 59), 60), 63), 64), 62), 61), (54, (55, (56, 57)))), ((33, (34, 35)), (36, (37, (38, (39, 40)))))), 49), (((((6, 7), (8, 9)), ((11, 13), (12, (14, 15)))), (23, ((24, 25), 29))), (16, (17, (18, ((19, 20), (21, 22)))))))))). D. Representative of the 10 equally most parsimonious trees constrained to show a monophyletic grouping of western lowland species (6016 steps, CI 0.264): (1, (2, ((3, 4), (((((5, ((6, 7), (8, (9, 10)))), (11, ((12, (13, (16, (17, (18, (19, 21)))))), (14, 15)))), (((23, 27), (24, (25, 26))), (28, 29))), (20, 22)), ((30, (31, 32)), (((33, (34, 35)), (36, (37, (38, (39, 40))))), (((41, 42), (43, ((44, 45), ((46, 47), (48, 49))))), ((50, 51), (52, (((53, 54), (55, (56, 57))), (58, (59, (60, (61, ((62, 64), 63))))))))))))))). E. Representative of the 14 equally most parsimonious trees with all viviparous taxa constrained to form a monophyletic group (6096 steps, CI 0.260): (1, (2, ((3, 4), ((((5, ((6, 7), (8, (9, 10)))), (((11, (13, (14, 15))), ((22, ((30, (31, 32)), (((33, (34, 35)), (36, (37, (38, (39, 40))))), (55, (56, 57))))), (23, ((24, (25, 102 J. A. SCHULTE ET AL.

26)), (27, (28, 29)))))), 12)), (16, ((17, ((19, 20), 21)), 18))), ((((41, 42), ((50, 51), (((52, 54), 53), (58, (59, (60, (61, ((62, 64), 63)))))))), ((44, 45), ((46, 47), (48, 49)))), 43))))). F. Representative of the 15 equally most parsimonious trees with all viviparous species in clade 1 (subgenus Liolaemus) constrained to form a monophyletic group (5969 steps, CI 0.266): (1, (2, ((3, 4), ((((5, ((6, 7), (8, (9, 10)))), (((11, (13, (14, 15))), (22, (23, ((24, (25, 26)), (27, (28, 29)))))), 12)), (16, ((17, ((19, 20), 21)), 18))), ((30, (31, 32)), (((33, (34, 35)), (36, (37, (38, (39, 40))))), (((41, 42), (43, ((44, 45), ((46, 47), (48, 49))))), ((50, 51), (52, (((53, 54), (55, (56, 57))), (58, (59, (60, (61, ((62, 63), 64))))))))))))))). G. Representative of the 32 equally most parsimonious trees with all viviparous species in clade 2 (subgenus Eulaemus) constrained to form a monophyletic group (5906 steps, CI 0.269): (1, (2, ((3, 4), (((5, (((6, 7), (8, (9, 10))), (((11, 13), (14, 15)), 12))), ((16, (17, (18, ((19, 20), (21, 22))))), (23, (((24, (25, 26)), (28, 29)), 27)))), ((((((((30, (31, 32)), ((33, (34, 35)), (36, (37, (38, (39, 40)))))), (55, (56, 57))), ((52, 54), (53, (58, (59, (60, (61, ((62, 64), 63)))))))), (50, 51)), (41, 42)), ((44, 45), ((46, 47), (48, 49)))), 43))))).