Total Evidence, Sequence Alignment, Evolution of Polychrotid Lizards, and a Reclassification of the Iguania (Squamata: Iguania) Author(s): DARREL R. FROST, RICHARD ETHERIDGE, DANIEL JANIES, and TOM A. TITUS Source: American Museum Novitates, Number 3343:1-39. 2001. Published By: American Museum of Natural History DOI: http://dx.doi.org/10.1206/0003-0082(2001)343<0001:TESAEO>2.0.CO;2 URL: http://www.bioone.org/doi/full/10.1206/0003-0082%282001%29343%3C0001%3ATESAEO %3E2.0.CO%3B2

BioOne (www.bioone.org) is a nonprofit, online aggregation of core research in the biological, ecological, and environmental sciences. BioOne provides a sustainable online platform for over 170 journals and books published by nonprofit societies, associations, museums, institutions, and presses. Your use of this PDF, the BioOne Web site, and all posted and associated content indicates your acceptance of BioOne’s Terms of Use, available at www.bioone.org/page/terms_of_use. Usage of BioOne content is strictly limited to personal, educational, and non-commercial use. Commercial inquiries or rights and permissions requests should be directed to the individual publisher as copyright holder.

BioOne sees sustainable scholarly publishing as an inherently collaborative enterprise connecting authors, nonprofit publishers, academic institutions, research libraries, and research funders in the common goal of maximizing access to critical research. PUBLISHED BY THE AMERICAN MUSEUM OF NATURAL HISTORY CENTRAL PARK WEST AT 79TH STREET, NEW YORK, NY 10024 Number 3343, 38 pp., 6 ®gures, 1 table, 4 appendices June 22, 2001

Total Evidence, Sequence Alignment, Evolution of Polychrotid Lizards, and a Reclassi®cation of the Iguania (Squamata: Iguania)

DARREL R. FROST,1 RICHARD ETHERIDGE,2 DANIEL JANIES,3 AND TOM A. TITUS4

ABSTRACT Using the techniques of direct optimization and sensitivity analysis, the phylogenetics of polychrotid lizards were examined on the basis of both molecular and morphological data (ca. 1040 bp of 12S rDNA, valine tDNA, and 16S rDNA, and 82 characters of morphology). A sensitivity analysis of sequence alignment and morphological change cost functions demon- strated that equal weighting provided the most parsimonious solution for all data. The Poly- chrotidae is found not to be monophyletic, containing instead the Corytophanidae as the sister taxon of Anolis plus Polychrus. Based on these and other results over the last 12 years, the taxonomy of the Iguania is reformulated, with the Iguania composed of two subsidiary taxa, Acrodonta and Pleurodonta, the Acrodonta containing the likely paraphyletic and basally un- resolved ``Agamidae'' as well as the Chamaeleonidae, and the Pleurodonta containing the Corytophanidae, Crotaphytidae, Hoplocercidae, Iguanidae, Leiocephalidae (newly elevated from its former status as a subfamily of the Tropiduridae), Leiosauridae (new taxon including Anisolepis, Aperopristis, Diplolaemus, Enyalius, Leiosaurus, Pristidactylus, and Urostrophus), Liolaemidae (newly elevated from its former status as a subfamily of the Tropiduridae), Oplur- idae, Phrynosomatidae, Polychrotidae (restricted to Anolis and Polychrus), and Tropiduridae (excluding the former subfamilies Leiocephalinae and Liolaeminae).

1 Associate Curator, Division of Vertebrate Zoology (Herpetology), American Museum of Natural History. 2 Research Associate, Division of Vertebrate Zoology (Herpetology), American Museum of Natural History; and Professor Emeritus, Department of Biology, San Diego State University, San Diego, CA 92187±0057. 3 Research Scientist, Division of Invertebrate Zoology, American Museum of Natural History. 4 Research Associate, Postlethwait Laboratory, Institute of Neuroscience, University of Oregon, Eugene, OR 94703.

Copyright ᭧ American Museum of Natural History 2001 ISSN 0003-0082 / Price $4.70 2 AMERICAN MUSEUM NOVITATES NO. 3343

INTRODUCTION ognized that taxon formally as the Polychro- The polychrotids form a large component tidae. of the lizard taxon Iguania, of which the larg- Because of its ubiquity, diversity, and availability in collections, Anolis has enjoyed est amniote genus Anolis5 (Ͼ300 species) is the polychrotids' most conspicuous member considerable attention regarding its taxono- in terms of local abundance, species diver- my and phylogeny (e.g., Cannatella and de sity, and distribution, being found from the Queiroz, 1989; Etheridge, 1959; Guyer and southeastern United States and northwestern Savage, 1986; Hass et al., 1993; Jackman et Mexico south through the Antilles and Cen- al., 1997, 1999; Poe, 1998). The remaining tral and South America to Bolivia, Paraguay, taxa within the polychrotid clade have not and Brazil. Nevertheless, Anolis (sensu lato) received much taxonomic attention in the last is only one genus in this group, with six oth- 30 years (subsequent to the summary of Pe- er smaller genera found primarily in austral ters and Donoso-Barros, 1970). Paull et al. South America. Most similar in general ap- (1976), Etheridge and de Queiroz (1988), pearance to Anolis is Polychrus, which is the Williams (1988), Frost and Etheridge (1989), only putative anole relative not restricted to Macey et al. (1997), and Schulte et al. (1998) southern South America. Polychrus is com- addressed these taxa (at least in part) as part posed of six species distributed from Nica- of larger problems of iguanian relationships. ragua to southern Brazil, Uruguay, and Etheridge and Williams (1991) reported on northern Argentina. Much less conspicuous the para-anoles. Jackson (1978) reviewed the in the literature (and in collections) are the systematics of Enyalius, and Etheridge and leiosaur polychrotids of southern South Williams (1985) provided a preliminary sum- America: Enyalius, Diplolaemus, Leiosau- mary of Pristidactylus. Other than these con- rus, and Pristidactylus. Enyalius contains six tributions, the remaining taxonomic summa- to nine species, depending on author (Eth- ries in the last 30 years have been more re- eridge, 1969; Jackson, 1978), and is found in stricted taxonomic papers: Cei (1973a) on central and southern Brazil. Diplolaemus generic limits among pristidactylines (i.e., contains three poorly distinguished Patagon- Leiosaurus, Diplolaemus, Pristidactylus, and ian species. Leiosaurus has three species [at that time recognized] Cupriguanus), Cei found in arid Argentina, and Pristidactylus, and Castro (1975) on the serology of Cupri- Pristidactylus), and Lamborot and found in western Argentina and central guanus (ϭ Diaz (1987) on speciation of Pristidactylus Chile, contains seven species (Etheridge and in Chile, as well as single species studies de Queiroz, 1988; Etheridge and Williams, (Cei, 1973b, on Pristidactylus fasciatus), 1985). Also found in southern South Amer- species descriptions (Donoso-Barros, 1975, ica are the para-anoles (Urostrophus and An- of Cupriguanus [ Pristidactylus] alvaroi; isolepis). Urostrophus has two species found ϭ Lamborot and Diaz, 1987, of Pristidactylus disjunctly in southern Bolivia to northern Ar- volcanensis), and regional faunal works (Cei, gentina and in southeastern Brazil, and Ani- 1986, 1993; Avila-Pires, 1995). The major solepis has three species in southeastern Bra- thrust of this study is to provide a basic clad- zil, Uruguay, northeastern Argentina, and ogram of the nominal species of non-anole central Paraguay (Etheridge and Williams, polychrotids so that progress in the study of 1991). the subsidiary taxa can be promoted. Nev- The polychrotid iguanians were ®rst pro- ertheless, like any such paper our objectives posed as a monophyletic group (as the ano- are several: loid iguanids) by Etheridge and de Queiroz (1) We test the monophyly of the Poly- (1988), even though the notion of such a chrotidae, which although it has been consid- group had currency somewhat earlier (e.g., ered monophyletic by most recent authors Etheridge in Paull et al., 1976). Subsequent- (e.g., Etheridge and de Queiroz, 1988; Frost ly, of course, Frost and Etheridge (1989) rec- and Etheridge, 1989), others (e.g., Williams, 5 For our purposes we use Anolis (sensu lato) to in- 1988) have disputed this. clude Chamaelinorops, Chamaeleolis, Norops, and (2) We provide an improved approxima- Phenacosaurus (Hass et al., 1993; Poe, 1998). tion of phylogeny and a taxonomy consistent 2001 FROST ET AL.: POLYCHROTID LIZARDS 3 with that phylogeny for the polychrotids not tus, A. meridionalis, A. ortonii, A. roquet, Di- only to elucidate relationships within this plolaemus darwini, Enyalius bilineatus, E. poorly known group, but also to provide a leechii, Leiosaurus catamarcensis, L. paron- more highly corroborated outgroup structure ae, Polychrus acutirostris, P. femoralis, P. for further studies by others on the huge tax- gutturosus, P. marmoratus, Pristidactylus on Anolis. Etheridge and de Queiroz (1988) scapulatus, and Urostrophus gallardoi. and Frost and Etheridge (1989) provided a Mitochondrial genes encoding the 12S ®rst and second estimate on the intergeneric rDNA, valine tDNA, and the 5Ј end of the phylogeny, but these were clearly prelimi- 16S rDNA were ampli®ed using the poly- nary. merase chain reaction. Genomic DNA ex- (3) We provide the morphological evi- traction, primers, and ampli®cation protocols dence in table form, also providing our se- were identical to those in Titus and Frost quences through GenBank, so that others (1996). Ampli®ed DNA was electrophoresed may evaluate our results. on 1% agarose gels and puri®ed for sequenc- (4) We comment brie¯y on various aspects ing using the Geneclean II kit (Bio 101, Inc.). of systematic methods as they present them- Thermal cycle sequencing was done follow- selves. In particular these issues are veri®- ing Titus and Frost (1996). Automated se- cationism in systematics, the assumptions of quencing was performed in the University of sequence alignment, and the justi®cation for Oregon Molecular Biology Sequencing Fa- cost functions of molecular changes as they cility utilizing the Big Dye Terminator Cycle relate to analysis of morphology. The use of Sequencing Kit with AmpliTaq FS (Perkin- sensitivity analyses as an empirical tool is Elmer) and an ABI PRISM 377 DNA Se- also explored. quencer (Perkin-Elmer) following the man- (5) We also discuss the progress in under- ufacturer's speci®cations. standing iguanian lizards and the classi®ca- All sequences have been deposited with tory solutions so far offered. We suggest a GenBank (see appendix 1 for accession num- classi®cation that we think will promote con- bers). tinued research. The data sets used re¯ect comparative dif- MORPHOLOGICAL DATA ferences of mitochondrial DNA and mor- phology, analyzed, we think in a new way, Skeletons, alcoholics, hemipenes, and at least for herpetologists, that allows truly cleared and double-stained specimens of simultaneous analysis of molecular sequence most non-anole ingroup taxa were examined and morphology data. for interspeci®c variation that could be hy- pothesized to be apomorphies relative to the MATERIALS AND METHODS outgroups (see Analytical Methods). Com- parison of specimens yielded 82 characters MOLECULAR DATA (numbered 0±81 in appendix 2 to allow easy Mitochondrial DNA sequences were ob- interpretation of our results) that were ana- tained for 16 ingroup terminals (and four lyzed separately and jointly with the molec- outgroup taxa) representing all nominal gen- ular data. (See appendix 2 for individual dis- era (see appendix 1 for GenBank accession cussion of transformation series, and appen- numbers). Eumeces egregius (as a surrogate dix 3 for the morphological data matrix.) Al- for the Scleroglossa, all noniguanian squa- though every attempt was made to get mates), Basiliscus basiliscus (Corytophani- representatives of all non-anole polychrotid dae), Oplurus cychlurus (Opluridae), and species, some absences are notable. We did Leiocephalus barahonensis (Tropiduridae: not have complete skeletal material for En- Leiocephalinae) were chosen as out-taxa to yalius bibroni (although notes of R. E. Eth- maximize the test of monophyly of the Po- eridge allowed some of the cells to be ®lled lychrotidae (see discussion below under in), Pristidactylus alvaroi, Pristidactylus val- Choice of Out-taxa). Ingroup taxa selected eriae, and Pristidactylus fasciatus (although for molecular analysis were Anisolepis lon- x-rays and a mandible were available). We gicauda, Anolis carolinensis, A. fuscoaura- did not attempt to discriminate the three spe- 4 AMERICAN MUSEUM NOVITATES NO. 3343 cies of Diplolaemus. The monophyly of the 1969; Farris, 1983; Farris and Kluge, 1985, genus is easily supported, but the species are 1986) of molecular and morphological data similar, and the species limits are murky at for most non-anole polychrotids and several best. Anolis (sensu lato) was not sampled out-taxa. The usual procedure for mixed densely. Instead, we used taxa that we hope morphology and molecular data analysis is sit near the base of the taxon and which will to align the DNA strands and employ either not affect either the placement of Anolis in character or taxonomic congruence to rec- our tree(s) or the character distributions that oncile the data sets on a single set of solu- diagnose these taxa. tions, character congruence being preferred for a number of reasons (Kluge, 1989). Nev- CHOICE OF OUT-TAXA ertheless, the procedure of alignment, regard- Out-taxa were selected on the basis of var- less of method used (e.g., CLUSTAL [Hig- ious lines of evidence of evolutionary prox- gins et al., 1992], MALIGN [Wheeler and imity and on their ability as potential falsi- Gladstein, 1994], or ``by eye'', which is least ®ers of polychrotid monophyly (Frost and satisfactory because it is not repeatable), op- Etheridge, 1989; Hallermann, 1994; Macey timizes cost functions (i.e., the cost to the et al., 1997; Titus and Frost, 1996; Schulte parsimony measure of changes by transver- et al., 1998). These taxa included (1) oplurids sions, transitions, and insertions/deletions) (inducted for purposes of morphology from on a single implied tree. When this aligned Chalarodon ϩ Oplurus [Blanc, 1977; Titus molecular data set is combined with another and Frost, 1996], with the molecular evi- data set, such as from morphology, the ®nal dence tied to this hypothetical taxon being tree obtained may actually have more ef®- Oplurus cychlurus); (2) Leiocephalus (coded cient alignments available for that particular morphologically as the ancestor of Leioce- summary tree than allowed by the initial (ഠ phalus as hypothesized by Pregill, 1992), locally optimal) alignment. In other words, with the molecular evidence being derived the ``distortion'' by the morphological data from Leiocephalus barahonensis; and (3) to the implied tree of aligned sequences may corytophanids (inducted as the ancestor of allow more parsimonious sequence solutions Basiliscus ϩ [Corytophanes ϩ Laemanctus] than are allowed by the initial alignment es- on the basis of Lang [1989] and independent timate of site homologies. This additional an- specimen examination) being represented for alytical step of alignment being followed by molecular evidence by Basiliscus basiliscus. inclusion for general analysis with another Finally, a noniguanian out-taxon was includ- data partition such as morphology has gen- ed to help root the entire network, with this erally been considered unavoidable. Never- being coded as the inducted ancestral scler- theless, the standard method of combining oglossan for morphology (based on our ob- aligned sequences with morphology retains a servations and those of Estes et al., 1988) component of taxonomic congruence (sensu and represented by Eumeces egregius (Scin- Kluge, 1989) that is not appropriate if we are cidae) for purposes of molecular evidence. serious about a total-evidence approach to Many of the morphological characters em- analysis. ployed here fail to be de®nable far away To avoid this particular problem, in this from the ingroup, so many cells in the scler- study we employed the method of direct op- oglossan line were left as question marks timization (Wheeler, 1996) as implemented where homology was dubious. A number of by the computer program POY (Gladstein morphological characters that corroborate the and Wheeler, 1997±2001; Janies and Wheel- monophyly of the Iguania were not included er, 2000), which allows morphological (Estes et al., 1988) for no reason other than changes to be simultaneously optimized with they were not needed to support the ingroup the molecular classes of change (transver- relationship with the near outgroups. sions, transitions, indels) on multiple poten- tial trees. This simultaneous analysis allows ANALYTICAL METHODS morphological characters to in¯uence align- The general analytical method employed is ment optimization directly, thereby allowing a parsimony analysis (Kluge and Farris, more ef®cient solutions the transformations 2001 FROST ET AL.: POLYCHROTID LIZARDS 5 implied by separate alignment and analysis. as the optimality criterion for choosing Put another way, POY optimizes cost func- among various topologies that are produced tions for all of the data onto various topol- and for choosing costs that minimize incon- ogies and calculates total parsimony costs for gruence among data sets inasmuch as there all cost functions. POY therefore has the ca- is no other empirical justi®cation for cost pacity to ®nd more parsimonious trees than functions. Janies suggests that there is no does the procedure of alignment followed by theoretical reason for picking any particular total evidence analysis. (For documentation, set of cost functions for sequence alignment a summary of this method, and access to the and that minimizing incongruence between program and command scripts see ftp.amnh. morphology and molecular sequence data org/pub/molecular/poy [Janies and Wheeler, may provide a general means of empirically in press].) investigating optimal alignment costs among POY also allows simultaneous analysis of different analyses. To foreshadow the results, data partitions under various cost regimes. our difference of opinion ultimately made no That is, different weighting functions for difference in this particular study. morphological change, transversions, transi- POY is implemented at the American Mu- tions, and indels can be applied to allow the seum of Natural History on a cluster of 256 general exploration of data partitions. In the UNIX-based work-stations integrated into a case of this analysis, various differential parallel virtual machine (Geist et al., 1993). weights of alignment cost functions were in- A total of 12 parameter sets were explored. vestigated in a sensitivity analysis to discov- The ratio of weights among indels and trans- er at what alignment cost functions incon- version or transition weights ranged from 1 gruence between morphology and molecules to 4. The transversion : transition ratios was minimized (Wheeler, 1995). ranged from 0.5 to 2. Some parameter sets The measure used to optimize congruence were set speci®cally to examine transversion was the Mickevich-Farris extra steps index parsimony (i.e., transitions were set at 0 cost (MFES) (Mickevich and Farris, 1981), which yielding a transversion: transition ratio of ϱ). measures the number of extra steps that oc- Changes in morphological data were pegged cur in an analysis of combined data versus to the cost of indels. Sequence data were also separate analysis of individual partitions. As analyzed separately. The addition of taxa (in- character incongruence among data partitions cluding putative outgroups) was randomized increases, MFES increases. The number of during the build and swapping processes for extra steps is normalized by the length of the molecular data and during swapping for mo- combined analysis so when parameter sen- lecular and morphological data. Tree search- sitivity analyses (Wheeler, 1995) are con- es included TBR and SPR swapping. ducted on the same data partitions (as in this The g1 statistic was not employed as a study), MFES scores are comparable despite measure of information content for reasons different weighting schemes. detailed by KallersjoÈ et al. (1992). Maximum Of course, the issue of differential weight- likelihood approaches have also not been ing and sensitivity analysis has exposed employed in this analysis because these among us some philosophical disagreements. equate to parsimony estimates when no gen- Frost and Janies are of somewhat different eral model of evolution is imposed (Tuf¯ey minds about the cost functions to employ in and Steel, 1997; Steel and Penny, 2000), the POY. Frost favors morphology, transver- assumption of evolutionary process that we sions, transitions, and indels to mutually in- think the most conservative. The measure of teract with a relative cost of 1 inasmuch as tree stability here employed is Bremer sup- he sees these all as marks of history and not port (or decay index) (Bremer, 1994). Inas- ``kinds'' of characters. He regards anything much as all measures of tree stability are fun- else as an attempt to reduce evidentiary am- damentally rules-of-thumb, Bremer support biguity rather than provide a logically and has the advantage of not appearing to be a philosophically sound framework for analy- parametric statistic as do bootstrap values. sis. Janies, on the other hand, argues that Nevertheless, like Wilkinson et al. (2000), character partition congruence must be used we appreciate that problems exist with Bre- 6 AMERICAN MUSEUM NOVITATES NO. 3343 mer values also, so these numbers should be considered carefully within the context of discussion.

RESULTS We will discuss the molecular-only and morphological-only results brie¯y before moving on to the combined analysis and dis- cussion of previous hypotheses.

MOLECULAR-ONLY ANALYSIS The molecular results shown (®g. 1) are based on relative cost functions of 1:1:1 (transversion : transition : indel). (The reason for this approach will be provided in the dis- cussion of the morphology plus molecular data sensitivity analysis below.) Only one tree was obtained for the 21 terminals (length ϭ 2332; CI ϭ 0.23; RI ϭ 0.68). If we arbi- trarily take a Bremer support of 7 and above as ``strong'' and below 7 as ``moderate to Fig. 1. Single tree obtained on molecular-only weak'', we can make the following obser- data (length ϭ 2332; CI ϭ 0.23; RI ϭ 0.68). vations with declining levels of con®dence Numbers on branches are Bremer values. on the molecular evidence alone: The monophyly of the polychrotids is re- MORPHOLOGY-ONLY ANALYSIS jected by the molecular evidence. Leioce- The morphology-only analysis (®g. 2) phalus, Basiliscus, and Oplurus (all nonpo- generally exposes much lower Bremer sup- lychrotids) were suggested by Etheridge (in port numbers than those found in the molec- Paull et al., 1976), Etheridge and de Queiroz ular analysis. Beyond being irrelevant, be- (1988), and Frost and Etheridge (1989) to be cause the numbers are not necessarily com- distantly related to the polychrotids, but in parable between the molecular and morpho- this analysis are imbedded within the Poly- logical analyses, this pattern of low Bremer chrotidae, being most closely related to Po- numbers is not due especially to con¯ict, but lychrus by apparently strong evidence. On to low numbers of characters along many of the face of it, this suggests that, as worried the stems, which is the normal turn of events about by other authors (most notably by Wil- in morphological studies in which individu- liams, 1988), Polychrus has little to do with ation of character states tends to be the cen- the other ``polychrotids''. Furthermore, An- tral problem. The number of equally parsi- olis, never considered a controversial mem- monious trees obtained was 192 (length ϭ ber of the group, is placed outside the re- 276; CI ϭ 0.443; RI ϭ 0.74). Salient features maining polychrotids, oplurids, leiocepha- of the morphology-only analysis are: lines, and corytophanids. The Polychrotidae obtains as a monophy- The leiosaurs plus the para-anoles (Uros- letic group under the most parsimonious ar- trophus, Anisolepis, Enyalius, Leiosaurus, rangement. With a Bremer support of 3 on Pristidactylus and Diplolaemus) are a highly the critical stem according to our arbitrary corroborated group with a decay index of 32, cutoff, this is not highly corroborated. an enormous number. Urostrophus plus An- Anolis is most parsimoniously placed as isolepis is placed as the sister taxon of Eny- the sister taxon of Polychrus, with the evi- alius, and Diplolaemus is placed as the sister dence for this being moderate. The structure taxon of Leiosaurus plus Pristidactylus. discovered within Anolis and Polychrus is 2001 FROST ET AL.: POLYCHROTID LIZARDS 7 substantially different from that obtained in equal-weighting analysis shows the most the molecule-only analysis, although the congruence by far between molecular evi- number of morphological characters critical dence and morphological evidence, this will to this structure is low. be the only analysis discussed here (and pre- The leiosaurs plus para-anoles group is viously in the molecular-only analysis). only weakly united, although the leiosaurs Figure 4 shows the strict consensus of the form a monophyletic group (with an alter- three most parsimonious trees (length of each native rooting of their component of the tree ϭ 2617; CI ϭ 0.25; RI ϭ 0.69) discov- overall network found in the molecule-only ered by the combined analysis, as well as the analysis). The Chilean species of Pristidac- corresponding Bremer support for these tylus (P. alvaroi, P. torquatus, P. valeriae, stems, and ®gure 5 shows this same tree with P. volcanensis) plus P. fasciatus, from Ar- the taxa (internal stems and terminals) iden- gentina, are found to be a monophyletic sub- ti®ed to correspond to the summary of group of the genus. change presented in appendix 4. To abbre- No particular structure within Enyalius viate discussion, the reader is referred to this was discovered, nor were the para-anoles re- table for a summary of the morphological ev- covered consistently as a monophyletic idence as well as the number and kind of group. Much of this lack of intrageneric res- molecular characters in support of each stem. olution was due seemingly to a lack of ®rmly As expected, the complementarity of the placed near neighbors in the analysis. molecular and morphological data is evident in the resolution obtained. With the excep- MORPHOLOGICAL PLUS MOLECULAR tion of lack of resolution at two nodes in En- DATA ANALYSIS yalius (5 and 6), three nodes in Pristidactylus (20, 23, and 24), and the decisive node for As described above a sensitivity analysis Leiosaurus monophyly (16), resolution of the was performed. A text summary of the re- 43 terminal taxa is dichotomous. Again, as- sults of the sensitivity analysis is provided in suming an arbitrary number of 7 for Bremer table 1, but more intelligibly these data are support, virtually all stems meet the criterion provided graphically in ®gure 3. In this for strong support. graphic, as the color goes to red (color cor- POLYCHROTID MONOPHYLY. The most im- responds to the z-axis, which represents data mediate result of the total evidence analysis set congruence), the congruence metric for is that the Polychrotidae as hypothesized by the best trees among the morphological and Etheridge and de Queiroz (1988) and Frost molecular data set increases. In this case and Etheridge (1989) is not monophyletic, maximum congruence between the molecular with the Corytophanidae (or basiliscines) be- and morphological data sets is achieved ing imbedded within it. This result is hardly when the ratios of change costs of transver- bewildering. There are rather few truly ar- sions, indels, transitions, and morphological boreal clades of iguanian lizards, and to ®nd 6 change are all equal to one. Because the the otherwise enigmatic corytophanids in

6 This result has some rather serious implications One possible explanation for why a posteriori differ- about the justi®cation for a priori character weighting ential character weighting could be arrived at by a sen- based on notions of inherent rates of change in particular sitivity analysis of the interaction of morphology and classes of modi®cation (e.g., routinely weighting trans- molecular evidence extends from sampling density of versions over transitions because of their generally av- terminal taxa/patristic distance among terminal. As tax- erage lower rate of appearance), particularly because on sampling density increases, one expects that long there is no logical basis to assume that the average rate branches will be partitioned, and that the partitioned of change in such classes as transversions and transitions components will become increasingly informative (i.e., should translate into differential weights for these class- more and more apomorphies will be correctly identi- es of character change (see Broughton et al., 2000). The ®ed), thereby bringing the ratio of various costs back to critical reader will nevertheless have noted that we also one. The number of studies so far that have performed use classes of characters in our sensitivity analysis (i.e., sensitivity analyses is small (e.g., Edgecomb et al., 1999; transversions, transitions, indels, and morphological Giribet et al., 2000; Janies and Mooi, 1999; O'Leary, character shifts), a central and seemingly inescapable as- 1999; Wheeler, 1995) thus, conjecture aside, we look pect of all scienti®c generalizations and methods (Frost forward to informed generalizations to be made as the and Kluge, 1994; Frost, 2000). set of examples expands. 8 AMERICAN MUSEUM NOVITATES NO. 3343

this position is not surprising. Indeed, in overall external appearance Laemanctus, the most generally plesiomorphic member of the corytophanids, is very similar in general body plan and habitus to some species of Po- lychrus, and young Basiliscus are routinely misidenti®ed in research collections as An- olis spp., a similarity explained in this ar- rangement by homology rather than conver- gence. Furthermore, the propensity for skull casquing in corytophanids and polychrotids is explained by this relationship. The stem (39, ®g. 5) supporting the Corytophanidae (Basiliscus basiliscus as the molecular sur- rogate for this taxon) with Anolis and Poly- chrus is supported by morphological char- acters 14.2 (fragmented suboculars) and, triv- ially, 33.2 (multicarinate subdigitals) and by 8 molecular transitions, 6 insertions, and 7 transversions. The loss of the standard po- lychrotid synapomorphy, calci®ed endolym- phatic sacs extending into the nuchal mus- culature (55.1), in the corytophanids, we judged to have been discovered as well as the loss of the typical polychrotid scale mi- crostructure (which is also lacking in Poly- chrus). Although the evidence in support of a monophyletic Corytophanidae plus the Po- lychrotidae (stem 26, ®g. 5) is not well cor- roborated by Bremer support of 5 we consid- er this a very interesting result further sup- porting the special relationship of coryto- phanids and polychrotids ®rst suggested by Hallermann (1994). LEIOSAURS PLUS PARA-ANOLES. Substantial evidence suggests that leiosaurs and the equally austral para-anoles form a monophy- letic group (stem 14; Bremer support of 11). This taxon is composed of two major groups, the arboreal Enyalius plus para-anoles (stem 9; Bremer support of 13) and the seemingly primitively terrestrial pristidactylines: Leio- saurus, Diplolaemus, and Pristidactylus (stem 17; Bremer support of 14). The phy- logenetic transition from Leiosaurus, Diplo- Fig. 2. Strict consensus of 192 equally parsi- laemus, and Argentinian Pristidactylus (P. monious trees based on morphology alone (length scapulatus, P. fasciatus, P. casuhatiensis)to ϭ 276; CI ϭ 0.443; RI ϭ 0.74). Numbers on the Chilean Pristidactylus (P. torquatus, P. branches are Bremer values. valeriae, P. volcanensis, and P. alvaroi)is unexpected inasmuch as both Etheridge (in Paull et al., 1976: 15) and Etheridge and de Queiroz (1988) suggested on the basis of TABLE 1 Sensitivity Analysis Under 12 Different Weighting Regimes Gap cost ϭ cost of adding a gap into the sequence; tv cost ϭ cost of making a transversion; ts cost

ϭ cost of making a transition; tv/ts ϭ ratio of transversion and transition costs; log2 tv/ts ϭ log (base 2) of tv/ts; gap/change ϭ ratio of gap cost to maximum cost of tv or ts change; log2 gap/change ϭ log (base 2) of gap/change; mm ϭ length of morphological ϩ molecular data on the shortest tree(s) based on analysis of combined morphological and molecular data; mol ϭ length of molecular data on the shortest tree(s) based on analysis of molecular data only; morph ϭ length of morphological data on the shortest tree(s) based on analysis of morphological data only; MFES ϭ Mickevich-Farris extra steps index (0 ϭ no extra steps; lowest MFES score is shown in boldface). See ®gure 3.

Fig. 3. Sensitivity analysis graphic. Y-axis represents the logarithm of the ratio of transversion : transition weights. The x-axis represents the logarithm of the ratio of the indel cost versus the maximal cost of a molecular change. The colors represent the z-axis, which is congruence between the molecular and morphological data partitions. Red is good, blue is bad. 10 AMERICAN MUSEUM NOVITATES NO. 3343

Fig. 4. Strict consensus of the three most par- Fig. 5. The same consensus as shown in ®g- simonious trees (length ϭ 2617; CI ϭ 0.25; RI ϭ ure 4, with stems and taxa identi®ed for ready 0.69) for all data, showing Bremer support. comparison with evolutionary changes presented in appendix 4 (change list). Stems 5, 6, 16, 20, 23, 24 represent alternative relationships not re- very preliminary evidence that the Chilean jected by the data. species were plesiomorphic within Pristidac- tylus because they tend ecologically toward arboreality, arguably the primitive habit of The Enyalius ϩ para-anole clade (stem 9; the group. Nevertheless, the data support ap- Bremer support of 13) is also well supported. pears strong for our conclusion that pristi- With the exception of the placement of Ani- dactylines are primitively terrestrial and sec- solepis grilli as the sister taxon of the re- ondarily arboreal. The inability to obtain a maining Anisolepis (stem 13; Bremer support decisively monophyletic Leiosaurus (sensu of 6), all other stems within the para-anoles lato) is also unexpected. Without molecular are well supported (®g. 4). Nevertheless, this data or a more strenuous look at the anatomy arrangement is hardly unexpected and is sup- of L. bellii, we have no way of resolving this ported by one unambiguous morphological polytomy (see comment in Conclusions). character. Structure within Enyalius (stem 8; 2001 FROST ET AL.: POLYCHROTID LIZARDS 11

Bremer support of 13) is less resolved than to make substantial taxonomic sampling. Our in other taxa, and our suspicion is that re- purpose to was to assure the appropriate gardless of the levels of support for various placement of Anolis within the larger phy- taxa, considerable work remains in the group logenetic framework and we think this has to delimit species. The results of our analysis been accomplished. of Enyalius correspond roughly to the tax- onomy suggested by Jackson (1978), al- CONCLUSIONS though his preferred phylogeny is so much at variance with ours that detailed discussion DIRECT OPTIMIZATION. We think that the is not warranted. (We note, however, that our method of direct optimization has performed phylogeny is roughly comparable to his phe- well, especially with its ability to truly si- netic network based on his meristic-morpho- multaneously cooptimize morphological and metric data set, when rooted between E. bil- molecular sequence data. That the empirical ineatus and E. iheringii, and not on his anal- results of this study indicated that the con- ysis of cranial measures; likely because on- gruence of the molecular and morphological togenetic variation in his osteological data was minimized by setting analytical samples vitiated any results he obtained.) costs at 1 for all changes (morphological Jackson (1978) looked at most of the avail- change, insertions and deletions, transver- able alcoholic material of Enyalius at the sions, and transitions) should make a number time of his writing and suggested that E. bi- of workers reevaluate what the logical justi- broni, E. pictus, and E. catenatus were in- ®cation is for differential weighting schemes tergrading subspecies of one species, as were and assertions of general models of evolu- E. boulengeri and E. brasiliensis. In the ®rst tion. We do not expect to have a large impact case our data do not reject that view. In the on the veri®cationist approach to phyloge- second case, however, our data indicate that netic inference which is increasingly popular E. leechii is imbedded within that second and seems to re¯ect the effects of many pop- group, suggesting that a conservative species ulation biologists entering the ®eld and taxonomy is warranted. bringing with them epistemologically inap- ANOLIS AND POLYCHRUS. Anolis monophyly propriate ways of thinking about historical is corroborated (stem 38; Bremer support of problems. Nevertheless, the ratio of 1:1:1:1 12) as is that of Polychrus (stem 30; Bremer is an empirical result of this study that should support of 29). not be dismissed casually. Furthermore, we Polychrus, despite being the perennial ex- suggest that because direct optimization con- ample of an enigmatic taxon (e.g., Vanzolini, sistently ®nds globally more parsimonious 1983; Williams, 1988) composed in part of resolutions for all data under consideration, poorly diagnosed species has not attracted it has a bright future. the kind of attention that it warrants. As far COMPARISON WITH PREVIOUS HYPOTHESES. as we know, this is the ®rst phylogenetic hy- Only two studies have dealt speci®cally with pothesis of the component species in this the phylogeny of all polychrotid genera: Eth- highly corroborated monophyletic taxon. eridge and de Queiroz (1988) and Frost and Nevertheless, except for the ®rm association Etheridge (1989). The results of these studies of the western species P. peruvianus, P. mar- are presented in ®gure 6, and because they moratus, and P. liogaster, we think that the present a natural evolution of increasing data, last word has hardly been said about the phy- suf®ce it to say that we think we have made logenetics of and species limits within this progress. The Etheridge and de Queiroz taxon. (1988) cladogram was done prior to easy Anolis (sensu lato) is, not surprisingly, computational analysis. The Frost and Eth- monophyletic and highly corroborated (®g. eridge (1989) analysis was done with a sec- 5: stem 38, Bremer support ϭ 12) by this ond-generation computer program (PAUP analysis. Our primary focus in this study was 2.4.1; Swofford, 1985), and the lack of res- on the relationships of the non-anole poly- olution in that study is salutary only in that chrotids, so we did not attempt to exhaust the it pointed the way to further research. More possibilities for morphological characters or recent molecular studies by Macey et al. 12 AMERICAN MUSEUM NOVITATES NO. 3343

agnosed by the features noted in appendix 4, stem 14) to contain ``Leiosaurus'', Diplolae- mus, Pristidactylus, Enyalius, and the para- anoles (Anisolepis and Urostrophus). Within this taxon we recognized two tribes: (1) Leio- saurinae (stem 17) to contain ``Leiosaurus'' (see below), Diplolaemus, and Pristidactylus, and (2) Enyaliinae (stem 9) to contain Eny- alius, Anisolepis, and Urostrophus. The issue of ``Leiosaurus'' is frustrating because we think that subsequent work will show this to be a monophyletic taxon. Nev- ertheless, at this juncture we do not have the evidence to reject the hypothesis of paraphyly. Rather than use the metataxon convention (Estes et al., 1988), which we believe has been abused to the point of now merely being a tool for promoting the recognition of para- phyletic groups (e.g., see Tudge [2000: 82± 87], for the use of the metataxon convention in defense of recognizing unambiguously par- aphyletic groups), our inclination is to return to the taxonomy of this group as it was before Gallardo (1961) and Etheridge and de Quei- roz (1988) synonymized Aperopristis (A. ca- tamarcensis and A. paronae) with Leiosaurus (L. bellii). This minor taxonomic change ren- ders a non-misleading taxonomy, and one we hope will be changed in the near future. The name Polychrotidae7 is attached to Fig. 6. Phylogenetic hypotheses for the Po- stem 32 (®g. 5), subtending Polychrus and lychrotidae of Etheridge and de Queiroz (1988), Anolis (sensu lato). Because this is at vari- and Frost and Etheridge (1989). Circles represent ance with previous taxonomy, at this point hypothesized rooting points. the issue of Iguanian taxonomy in general must be addressed. COMMENTS ON IGUANIAN PHYLOGENETICS AND (1997) and Schulte et al. (1998) dealt with a TAXONOMY. Twelve years ago Frost and Eth- restricted set of terminal taxa, but beyond eridge (1989) noted that no evidence for the providing evidence for the pleurodont igu- monophyly of the Iguanidae (sensu lato; Bou- anians being monophyletic, they did provide lenger, 1885) existed, nor was there decisive molecular evidence in support of Haller- evidence for the monophyly of the Agamidae mann's (1994) earlier suggestion that the cor- with respect to the chameleons. Their (our) so- ytophanids were intimately involved with the lution was to render a scienti®cally conserva- polychrotids. We think that all of these stud- tive taxonomy in the sense of sticking close to ies show increasing understanding of the the evidence rather than bending to social con- phylogenetics of the group. servatism. The strength of the study was that TAXONOMIC CONCLUSIONS. Inasmuch as ev- the ``dirty laundry'' of Iguanian systematics idence for a monophyletic Polychrotidae sen- was displayed for all to see and to invite fur- su Frost and Etheridge (1989) is lacking, we cannot continue its recognition in its original 7 We consider Polychrotidae Fitzinger (1843) to have form. (See comment on Iguanian phylogenet- priority over Corytophanidae Fitzinger (1843), under ics and taxonomy, following this section for provisions of article 24.2 (First Revisor Principle) of the context.) We recognize the Leiosauridae (di- International Code of Zoological Nomenclature (1999). 2001 FROST ET AL.: POLYCHROTID LIZARDS 13 ther scrutiny. This approach met with some mates generallyÐsee Northcutt, 1978, and criticism, most of which was related to discom- Harris et al., 1999), and suggest that the fol- fort with phylogenetic systematics (e.g., BoÈh- lowing taxonomy (for living taxa) of iguanian me, 1990), but for the most part it was met lizards makes the best provision for promoting with considerable enthusiasm and adopted further progress as well as the best statement wherever monophyly was taken seriously, of the state of our understanding of the phy- which we were grati®ed to see was in most of logeny of the group. (Where the group content the Western Hemisphere. The ®rst scienti®c is the same as in Frost and Etheridge, 1989, criticism of this arrangement was that of Ma- no notations are provided.) cey et al. (1997) in which they presented mo- Iguania Cope, 1864 lecular evidence for the monophyly of the Ig- Acrodonta Cope, 1864 (ϭ Chamaeleoni- uanidae (sensu lato). We applaud this effort dae of Frost and Etheridge, 1989). and for those who wish to reclaim the older Chamaeleonidae Ra®nesque, 1815 (ϭ taxonomy on this basis, we have no scienti®c Chamaeleoninae of Frost and Ether- reason to dispute this reclamation. Neverthe- idge, 1989) less, in the intervening years since 1989, the ``Agamidae'' (see note at end of list) family-group names of Frost and Etheridge Spix, 1825 (ϭ Leiolepidinae plus (1989) have widespread acceptance and usage. Agaminae of Frost and Etheridge, Further, we suspect strongly that additional 1989; diagnostically equivalent to work on the fossil record will make for added the Acrodonta as well as the Cha- ambiguity in the phylogenetic record regarding maeleonidae sensu Frost and Ether- the status of the Iguania as well as the evidence idge, 1989) of monophyly of the Iguanidae (sensu lato). Pleurodonta Cope, 1864 (ϭ Iguanidae So, as long as the canon of monophyly is not sensu Boulenger, 1885) violated, the choice of ranks is entirely a sub- Corytophanidae Fitzinger, 1843 jective decision dependent upon what we be- Crotaphytidae Smith and Brodie, 1982 lieve will lead to the least confusion and make Hoplocercidae Frost and Etheridge, for the most progress. We therefore recom- 1989 mend the resurrection of the name Pleurodonta Iguanidae Oppel, 1811 (Cope, 1864) for the monophyletic group for- Leiocephalidae Frost and Etheridge, merly known as the Iguanidae (sensu lato; 1989 (ϭ Leiocephalinae of the Tro- Boulenger, 1885). This provides symmetry piduridae of Frost and Etheridge, with the Acrodonta (Cope, 1864), its putative 1989) sister taxon and, if pleurodont iguanian mono- Leiosauridae New phyly is falsi®ed, or found to be unlikely (un- Leiosaurinae New der the approach promoted by Macey et al., Enyaliinae New 1997), we will not have to make major chang- Liolaemidae Frost and Etheridge, 1989 es in our taxonomy. Further, the molecular ev- (ϭ Liolaeminae of the Tropiduridae idence presented by Titus and Frost (1996), of Frost and Etheridge, 1989). Macey et al. (1997), and Schulte et al. (1998) Opluridae Moody, 1983 suggests that the weak morphological evidence Phrynosomatidae Fitzinger, 1843 for the monophyly of the Tropiduridae is deep- Polychrotidae Fitzinger, 1843 (as de- ly questionable and certainly rejected by the ®ned above) molecular evidence, so we will not persist in Tropiduridae Bell, 1843 (ϭ Tropiduri- recognizing that nominal taxon. Obviously, in nae of the Tropiduridae of Frost and the future, fossils will have to be taken into Etheridge, 1989) account, and the set of possible terminals will have to be sampled densely, both with respect NOTE ON ``AGAMIDAE.'' Macey et al. to molecular and morphological characters an- (1997) argued that within the framework of alyzed simultaneously, before we can arrive at systematic con®dence estimates that the a comprehensive taxonomy. We still consider Agamidae (in the sense of nonchameleon ac- Iguanian phylogenetics to be largely an open rodont iguanians) should be considered a me- ®eld (indeed, we think this extends to squa- tataxon (cf. Estes, 1988). We suggest that the 14 AMERICAN MUSEUM NOVITATES NO. 3343 metataxon convention does not apply, and ACKNOWLEDGMENTS that this issue of taxonomic disarray can be addressed more clearly by applying the quo- We thank J. Faivovich, T. Grant, M. Ma- tation convention for paraphyly developed honey, G. Pregill, W. Wheeler, and especially by Wiley (1981: 213). Frost and Etheridge B. Hollingsworth and M. Donnelly for con- (1989) found no decisive evidence for a structive criticism of the manuscript and to monophyletic Agamidae to the exclusion of G. Giribet and W. Wheeler for advice on the the chameleons. In one of their two most par- use of POY. K. de Queiroz provided advice simonious topologies, chameleons were im- regarding appropriate exemplars of Anolis bedded within the Agamidae (in the tradi- (sensu lato) to represent that taxon in our tional sense of nonchameleon acrodonts). In more general analysis. Access to morpholog- the other arrangement, chameleons were the ical specimens was provided by L. Coloma (Escuela PoliteÂcnica Nacional, Quito, Ecua- sister taxon of other living acrodonts; in oth- dor); D. B. Wake and B. Stein (Museum of er words, only one of two topologies sup- Vertebrate Zoology, University of California ported the Agamidae in the sense of Macey [MVZ]); J. Cadle and J. Rosado (Museum of et al. (1997). If this were the only basis on Comparative Zoology, Harvard University which to derive a taxonomy, we suppose that [MCZ]; R. McDiarmid, R. I. Crombie, and authors could have opted for applying the R. Reynolds (National Museum of Natural metataxon convention, which, at least as History, U.S. Geological Survey and Smith- originally formulated, was to be applied in sonian Institution [USNM]); E. N. Arnold cases of evidentiary con¯ict as in this ex- and B. Clarke (Natural History Museum, ample. However, Macey et al. (1997: ®g. 5a, London [BM]); L. Vitt and J. Caldwell b) presented new molecular evidence for (Oklahoma State Museum of Natural History ``agamid'' paraphyly and none for its mono- [OMNH]); M. T. Rodrigues and H. Zaher phyly. So, even though they reformulated the (Museu de Zoologia, Universidade de SaÄo notion of metataxon in the form of a con®- Paulo [MZUSP]); A. Almendariz (PoliteÂcni- dence measure, the metataxon convention is ca Universidad CatoÂlica, Quito, Ecuador not appropriate (at least in a deductive ap- [PUC]); A. G. Kluge, R. I. Nussbaum, and proach), because the evidence pointed to G. Schneider (Museum of Zoology, Univer- ``agamid'' paraphyly more strongly in 1997 sity of Michigan [UMMZ]); W. E. Duellman than it did in 1989. Formulating a taxonomy and J. Simmons (Museum of Natural Histo- that suggests something different is not con- ry, University of Kansas [KU]); R. Laurent servative in any evidentiary sense and, fur- and F. Lobo (Instituto Miguel Lillo, Tucu- ther, no one attached to the canon of mono- maÂn, Argentina [FML]); R. Inger and H. phyly would formulate the taxonomy of ac- Voris (Field Museum of Natural History rodonts still adhered to by so many. The base [FMNH]); and Juan Carlos Ortiz (Departa- of the cladogram of acrodonts is poorly un- mento de ZoologõÂa, Universidad de Concep- derstood, and we consider it unlikely that fu- cioÂn, Chile [IZUC]). ture work will render the ``agamids'' mono- Tissues were graciously supplied by W. E. phyletic with respect to the chameleons, es- Duellman and C. Raxworthy (KU), C. J. pecially when fossils are ®nally taken into Cole and C. Myers (AMNH), D. Dittman account. For this reason we suggest that as (Museum of Natural Science, Louisiana State an evidentiarily conservative taxonomy that University [LSU]), M. T. Rodriguez the ``Agamidae'' be placed in quotations to (MZUSP), F. B. Cruz and S. E. Torres denote its paraphyletic status (Wiley, 1981: (FML), B. Alvarez de Avanza (Universidad 213), pending resolution of the placement of Nacional del Nordeste, Corrientes, Argentina the problematic taxa, Uromastyx, Leiolepis, [UNNEC]), L. Vitt and J. Caldwell (OMNH), Physignathus (including Hydrosaurus), and and L. Fitzgerald (Texas A&M University the species of ``Hypsilurus'', not to mention [TCWC]). J. Williams (Museo de La Plata, a relatively large number of fossil taxa in- Argentina) and M. R. Cabrera (Universidad cluding Mimeosaurus, Isodontosaurus, Pris- Nacional de CoÂrdoba, Argentina) provided cagama, and Arretosaurus. information on life history and behavior in 2001 FROST ET AL.: POLYCHROTID LIZARDS 15

Urostrophus and Anisolepis. This project Cannatella, D. C., and K. de Queiroz was funded in part by National Science 1989. Phylogenetic systematics of the anoles: Foundation grant DEB-9220870 to D. Frost Is a new taxonomy warranted? Syst. and T. Titus, and by National Aeronautical Zool. 38: 57±69. and Space Administration grants NAG5- Cei, J. M. 8443 to D. Frost and W. Wheeler and NAG2- 1973a. Comentarios sobre algunos geÂneros de iguaÂnidos: Diplolaemus, Leiosaurus, 1339 to D. Janies and W. Wheeler. Aperopristis y Cupriguanus. Physis (Buenos Aires) (C)32(85): 269±276. REFERENCES 1973b. DistribucioÂn geogra®ca y caracteres poblacionales de Cupriguanus fasciatus Anonymous (D'Orbigny) (Sauria, Iguanidae). Phy- 1999. International Code of Zoological No- sis (Buenos Aires) (C)32(85): 255±262. menclature. 4th ed. London: Interna- 1986. Reptiles del centro, centro-oeste y sur tional Trust for Zoological Nomencla- de la Argentina. Herpetofauna de las ture. zonas aridas y semiaridas. Mus. Reg. Arnold, E. N. Sci. Nat. Monogr. (Torino) 4: 527 pp. 1984. Variation in the cloacal and hemipenial 1993. Reptiles del noroeste, nordeste y este muscles of lizards and its bearing on de la Argentina. Herpetofauna de las their relationships. Symp. Zool. Soc. selvas subtropicales, puna y pampas. London 52: 47±85. Mus. Reg. Sci. Nat. Monogr. (Torino) Avila-Pires, T. C. S. de 14: 949 pp. 1995. Lizards of Brazilian Amazonas (Repti- Cei, J. M., and L. P. Castro lia: Squamata). Zool. Verh. (Leiden) 1975. A serological contribution to the taxo- 299: 1±706. nomic status of Cupriguanus, a South Bell, T. American genus of iguanid lizards. Ser- 1843. Zoology of the voyage of the H.M.S. ol. Mus. Bull. 51: 5±6. Beagle, under the command of Captain Cope, E. D. Fitzroy, R.N., during the years 1832 to 1864. On the characters of the higher groups 1836. Edited and superintended by of Reptilia SquamataÐand especially Charles Darwin . . . naturalist to the of the Diploglossa. Proc. Acad. Nat. expedition. Pt. 5. Reptiles. vi ϩ 51 pp., atlas. London: Smith, Elder. Sci. Philadelphia 16: 224±231. Blanc, C. P. Donoso-Barros, R. 1977. Reptiles, Sauriens Iguanidae. Faune 1975. Nuevos reptiles y an®bios de Chile. Madagascar 45: 1±194. Bol. Soc. Biol. ConcepcioÂn 47(1974): BoÈhme, W. 221. 1988. Genitalmorphologie der Sauria: Funk- Edgecomb, G., G. Giribet, and W. Wheeler tionelle und stammesgeschichtliche As- 1999. Ecdyzoa versus Articulata, dos hipoÂtes- pekte. Bonn. Zool. Monogr. 27: 1±276. is alternatives sobre la posicioÂn de los 1990. Review ``Phylogenetic analysis of the artroÂpodos en el reino Animal. Bol. Iguania (Squamata) by D. R. Frost and Soc. Entomol. Aragon. 26: 145±160. R. E. Etheridge''. Z. Zool. Syst. Evo- Estes, R., K. de Queiroz, and J. Gauthier lutionsforsch. 28(4): 315±316. 1988. Phylogenetic relationships within Squa- Boulenger, G. A. mata. In R. Estes and G. Pregill (eds.), 1885. Catalogue of the lizards in the British Phylogenetic relationships of the liz- Museum (Natural History). 2nd ed. ard families: essays commemorating Vol. 2. London: Taylor and Francis. Charles L. Camp: 119±281. Stanford, Bremer, K. CA: Stanford Univ. Press. 1994. Branch support and tree stability. Cla- Etheridge, R. distics 10: 295±304. 1959. The relationships of the anoles (Repti- Broughton, R. E., S. E. Stanley, and R. T. Durrett lia: Sauria: Iguanidae): an interpretation 2000. Quanti®cation of homoplasy for nucle- based on skeletal morphology. Ph.D. otide transitions and transversions and diss., Univ. Michigan, Ann Arbor. a re-examination of assumptions in 1969. A review of the iguanid lizard genus weighted phylogenetic analysis. Syst. Enyalius. Bull. Br. Mus. (Nat. Hist.) Biol. 49(4): 617±627. Zool. 18(8): 231±260. 16 AMERICAN MUSEUM NOVITATES NO. 3343

Etheridge, R., and K. de Queiroz Giribet, G., D. Distel, M. Polz, W. Sterrer, and W. 1988. A phylogeny of Iguanidae. In R. Estes Wheeler and G. Pregill (eds.), Phylogenetic re- 2000. Triploblastic relationships with empha- lationships of lizard families: essays sis on the acoelomates, and the position commemorating Charles L. Camp: of Gnathostomulidae, Cycliophora, 283±368. Stanford, CA: Stanford Univ. Plathyhelminthes, and Chaetognatha; a Press. combined approach of 18S rDNA and Etheridge, R., and E. E. Williams morphology. Syst. Biol. 49: 1±24. 1985. Notes on Pristidactylus (Squamata: Ig- Gladstein, D. S., and W. C. Wheeler uanidae). Breviora 483: 1±18. 1997±2001. POY: The optimization of align- 1991. A review of the South American lizard ment characters. Program and docu- genera Urostrophus and Anisolepis mentation. Am. Mus. Nat. Hist., New (Squamata: Iguania: Polychrotidae). York. [Available at ftp.amnh.org/pub/ Bull. Mus. Comp. Zool. 152(5): 317± molecular.] 361. Guyer, C., and J. M. Savage Farris, J. S. 1986. Cladistic relationships among anoles 1983. The logical basis of phylogenetic anal- (Sauria: Iguanidae). Syst. Zool. 35: ysis. In N. I. Platnick and V. A. Funk 509±531. (eds.), Advances in cladistics 2: 7±36. Hallermann, J. New York: Columbia Univ. Press. 1994. Zur Morphologie der Ethmoidregion Farris, J. S., and A. G. Kluge der Iguania (Squamata)ÐEine verglei- 1985. Parsimony, synapomorphy, and explan- chend-anatomische Untersuchung. Bonn. atory power: a reply to Duncan. Taxon Zool. Monogr. 35: 133 pp. 34: 130±135. Harris, D. J., E. A. Sinclair, N. L. Mercader, J. C. 1986. Synapomorphy, parsimony, and evi- Marshall, and K. A. Crandall dence. Taxon 35: 298±305. 1999. Squamate relationships based on c-mos Fitzinger, L. I. nuclear DNA sequences. Herpetol. J. 1843. Systema Reptilia. Fasciculus primus. 9(4): 147±151. Amblyglossae. Wien: BaumuÈller and Hass, C. A., S. B. Hedges, and L. R. Maxson Seidel. 1993. Molecular insights into the relation- Frost, D. R. ships and biogeography of West Indian 1992. Phylogenetic analysis and taxonomy of anoline lizards. Biochem. Syst. Ecol. the Tropidurus group of lizards (Iguan- 27(1): 97±114. idae: Tropiduridae). Am. Mus. Novita- Higgins, D. G., A. J. Bleasby, and R. Fuchs tes 3033: 68 pp. 1992. CLUSTAL V: Improved software for 2000. Species, descriptive ef®ciency, and pro- multiple sequence alignment. Comput. gress in systematics. In R. D. Bruce, R. Appl. Biosci. 8: 189±191. G. Jaeger, and L. Houck (eds.), Biology Hoogmoed, M. of plethodontid salamanders: 7±30. 1973. Notes on the herpetofauna of Surinam. New York: Plenum Press. IV. The lizards and amphisbaenians of Frost, D. R., and R. Etheridge Surinam. Biogeographica 4: ix ϩ 419 1989. A phylogenetic analysis and taxonomy pp. of iguanian lizards (Reptilia: Squama- Jackman, T. R., A. Larson, K. de Queiroz, and J. ta). Univ. Kansas Mus. Nat. Hist. Misc. B. Losos Publ. 81: 65 pp. 1999. Phylogenetic relationships and tempo Frost, D. R., and A. G. Kluge of early diversi®cation in Anolis liz- 1994. A consideration of epistemology in sys- ards. Syst. Biol. 48(2): 254±285. tematic biology, with special reference Jackman, T. R., J. B. Losos, A. Larson, and K. de to species. Cladistics 10: 259±294. Queiroz Gallardo, J. M. 1997. Phylogenetic studies of convergent 1961. Estudio zoologeogra®co del geÂnero adaptive radiations in Caribbean Anolis Leiosaurus (Reptilia: Sauria). Physis lizards. In T. J. Givnish and K. Systma (Buenos Aires) 22(63): 113±118. (eds.), Molecular evolution and adap- Geist, A., J. J. Beguelin, W. Dongarra, R. Jiang, tive radiation: 535±547. New York: V. Manchek, and V. S. Sunderam Cambridge Univ. Press. 1993. PVM 3 user's guide and reference man- Jackson, J. F. ual. Tech. Rep. ORNL/TM-12187, Oak 1978. Differentiation in the genera Enyalius Ridge National Laboratory, Tennessee. and Strobilurus (Iguanidae): implica- 2001 FROST ET AL.: POLYCHROTID LIZARDS 17

tions for Pleistocene climatic changes 195±212. Cambridge, MA: Museum of in eastern Brazil. Arq. Zool. (SaÄo Pau- Comparative Zoology. lo) 30(1): 79 pp. Northcutt, R. G. Janies, D., and R. Mooi 1978. Forebrain and midbrain organization in 1999. Xyloplax is an asteroid. In M. D. C. lizards and its phylogenetic signi®- Carnevali and F. Bonasoro (eds.), Echi- cance. In N. Greenberg and P. D. noderm research 1998: 311±316. Rot- MacLean (eds.), Behavior and neurol- terdam: Balkema. ogy of lizards, an interdisciplinary col- Janies, D., and W. Wheeler loquium: 11±64. Washington, DC: U.S. 2000. POY. Documentation, sample data sets, Natl. Inst. Mental Health. and shell scripts for sensitivity analyses O'Leary, M. for POY, software for direct optimiza- 1999. Parsimony analysis of total evidence tion of DNA and character data. [http: from extinct and extant taxa and the ce- //research.amnh.org/ϳdjanies/poy.pdf.] tacean-artiodactyl question (Mammalia, In press. Theory and practice of parallel direct Ungulata). Cladistics 15: 315±330. optimization. In R. Desalle, G. Giribet, Oppel, M. and W. Wheeler (eds.), Techniques in 1811. Die Ordnungen, Familien und Gattun- molecular systematics and evolution. gen der Reptilien als Prodrom einer na- Basel: Birkhauser. turgeschichte Derselben. MuÈnchen: J. KallersjoÈ, M., J. S. Farris, A. G. Kluge, and C. Lindauer. Bult Paull, D., E. E. Williams, and W. P. Hall 1992. Skewness and permutation. Cladistics 1976. Lizard karyotypes from the GalaÂpagos 8: 275±287. Islands: chromosomes in phylogeny Kluge, A. G. and evolution. Breviora 441: 1±31. 1989. A concern for evidence and a phylo- Peracca, M. G. genetic hypothesis of relationships 1897. Viaggio del Dott. A. Borelli nel Chaco among Epicrates (Boidae, Serpentes). Boliviana e nella Republica Argentina. Syst. Zool. 38: 7±25. Rettili e An®bi. Boll. Mus. Zool. Anat. Kluge, A. G., and J. S. Farris Comp. Univ. Torino 12(274): 1±19. 1969. Quantitative phylogenetics and the evo- Peters, J. A., and R. Donoso-Barros lution of anurans. Syst. Zool. 18: 1±32. 1970. Catalogue of the neotropical Squamata: Lamborot, M., and N. F. Diaz Part II. Lizards and amphibisbaenians. 1987. A new species of Pristidactylus (Sau- U.S. Natl. Mus. Bull. 297(2): vii ϩ 293 ria: Iguanidae) from central Chile and pp. comments on the speciation in the ge- Peterson, J. A. nus. J. Herpetol. 21(1): 29±37. 1983. The evolution of the subdigital pad in Lang, M. Anolis. I. Comparisons among the an- 1989. Phylogenetic and biogeographic pat- oline genera. In A. G. C. Rhodin and terns of basiliscine iguanians (Reptilia: K. Miyata (eds.), Advances in herpe- Squamata: ``Iguanidae''). Bonn. Zool. tology and evolutionary biology: es- Monogr. 28: 172 pp. says in honor of Ernest E. Williams: Macey, J. R., A. Larson, N. B. Ananjeva, and T. 245±283. Cambridge, MA: Museum of J. Papenfuss Comparative Zoology. 1997. Evolutionary shifts in three major Peterson, J. A., and E. E. Williams structural features of the mitochondrial 1981. A case history of retrograde evolution: genome among iguanian lizards. J. the onca lineage in anoline lizards. II. Mol. Evol. 44: 660±674. Subdigital ®ne structure. Bull. Mus. Mickevich, M. F., and J. S. Farris Comp. Zool. 149: 215±268. 1981. The implications of congruence in Poe, S. Menidia. Syst. Zool. 30: 351±370. 1998. Skull characters and the cladistic rela- Moody, S. M. tionships of the Hispaniolan dwarf twig 1983. The rectus abdominis muscle complex Anolis. Herpetol. Monogr. 12: 192±236. of the Lacertilia: terminology, homol- Pregill, G. K. ogy, and assumed presence in primitive 1984. An extinct species of Leiocephalus iguanian lizards. In A. G. C. Rhodin from Haiti (Sauria: Iguanidae). Proc. and K. Miyata (eds.), Advances in her- Biol. Soc. Washington 97: 827±833. petology and evolutionary biology: es- 1992. Systematics of the West Indian lizard says in honor of Ernest E. Williams: genus Leiocephalus (Squamata: Igu- 18 AMERICAN MUSEUM NOVITATES NO. 3343

ania: Tropiduridae). Univ. Kansas Mus. Tudge, C. Nat. Hist. Misc. Publ. 84: 1±69. 2000. The Variety of life. New York: Oxford Ra®nesque, C. S. Univ. Press. 1815. Analyse de la Nature ou tableau de Tuf¯ey, C., and M. Steel l'universe et des corps organiseÂs. Pa- 1997. Links between maximum likelihood lermo: J. Barravecchia. and maximum parsimony under a sim- Schulte, J. A., III, J. R. Macey, A. Larson, and T. ple model of site substitution. Bull. J. Papenfuss Math. Biol. 59: 581±607. 1998. Molecular tests of phylogenetic taxon- Vanzolini, P. E. omies: a general procedure and exam- 1983. Guiano±Brasilian Polychrus: distribu- tion and speciation (Sauria: Iguanidae). ple using four subfamilies of the lizard In A. G. C. Rhodin and K. Miyata family Iguanidae. Mol. Phylogenet. (eds.), Advances in herpetology and Evol. 10(3): 367±376. evolutionary biology: essays in honor Smith, H. M. of Ernest E. Williams: 118±131. Cam- 1949. Handbook of lizards. Lizards of the bridge, MA: Museum of Comparative United States and Canada. Ithaca: Cor- Zoology. nell Univ. Press. Wheeler, W. C. Smith, H. M., and E. D. Brodie, Jr. 1995. Sequence alignment, parameter sensi- 1992. A guide to ®eld identi®cationÐreptiles tivity, and the phylogenetic analysis of of North America. New York: Golden molecular data. Syst. Biol. 44: 321± Press. 332. Spix, J. B. von 1996. Optimization alignment: The end of 1825. Animalia nova sive species nova lac- multiple sequence alignment in phylo- ertarum quas in itinere per Brasiliam genetics? Cladistics 12: 1±10. annis MDCCCXVII±MDCCCXX jussu Wheeler, W. C., and D. S. Gladstein et auspicius Maximiliani Josephi I Bar- 1994. MALIGN: A multiple sequence align- variae Regis suscepto collegit et des- ment program. J. Hered. 85: 417. criptsit Dr. J. B. de Spix. Lipsiae: T. O. Wiley, E. O. Weigel. 1981. Phylogenetics. The theory and practice Steel, M., and D. Penny of phylogenetic systematics. New 2000. Parsimony, likelihood, and the role of York: Wiley. Wilkinson, M., J. L. Thorley, and P. Upchurch models in molecular phylogenetics. 2000. A chain is no stronger than its weakest Mol. Biol. Evol. 17(6): 839±850. link: double decay analysis of phylo- Swofford, D. L. genetic hypotheses. Syst. Biol. 49(4): 1985. PAUPÐPhylogenetic analysis using 754±776. parsimony. Vers. 2.4.1. User's manual. Williams, E. E. Privately Published. 1988. A new look at the Iguania. In P. E. Van- Titus, T. A., and D. R. Frost zolini and W. R. Heyer (eds.), Proceed- 1996. Molecular homology assessment and ings of the Workshop on Neotropical phylogeny in the lizard family Opluri- Distribution Patterns held 12±16 Janu- dae (Squamata: Iguania). Mol. Phylo- ary 1987: 429±488. Rio de Janeiro: genet. Evol. 6(1): 49±62. Academia Brasileira de CieÃncias.

APPENDIX 1

Specimens Examined, Molecular Data Vouchers, and GenBank Accession Numbers

MORPHOLOGICAL SPECIMENS. Anatomical speci- amined by Etheridge previously (and the source mens of ingroup taxa examined by Frost and Eth- of a considerable volume of notes) are listed in eridge are noted below. In addition to these listed Etheridge (1959, 1969), Etheridge and de Queiroz specimens, other specimens of the genera Diplo- (1988), and Etheridge and Williams (1985, 1991). laemus, Leiosaurus, Pristidactylus, Anisolepis, Outgroup taxa specimens are listed in Frost and Urostrophus, Enyalius, and Anolis (sensu lato) ex- Etheridge (1989) and Titus and Frost (1996) as 2001 FROST ET AL.: POLYCHROTID LIZARDS 19 well as those available in AMNH and REE (R. E. ble-stained with dry skull). E. perditus: AMNH Etheridge osteology; now mostly in the AMNH) 133744±47 (alcoholics); MCZ 163788, USNM collections. 247877, MZUSP 43017±18 (skeletons); AMNH Anisolepis grilli: USNM 73504 (alcoholic); 119749 (skull); AMNH 119749 (postcranial skel- MCZ 13319; BMNH 1946.8.12.38, REE 1952 eton). E. pictus: SDSU 2221, AMNH 131859±60 (skeletons); AMNH 120468 (skull and cleared (alcoholics); MZUSP 8826, 59183±85 (skeletons); and double-stained postcranium). A. longicauda: MCZ 163784 (skull); MZUSP 42686 (skull). BMNH 1946.8.9.2 (alcoholic); BMNH 98.11.3.1 Leiosaurus bellii: SDSU 2228, 2230, AMNH (skeleton). A. undulata: BMNH 1946.8.5.90±91, 17004 (alcoholics); REE 2410 (skeleton). Pristi- USNM 65545 (alcoholics); MCZ 84033, 59273, dactylus achalensis: SDSU 2380±81, 2385±86 59274 (skeletons). Anolis carolinensis: AMNH (alcoholics); MCZ 86628, REE 2488, MVZ (many alcoholic specimens); AMNH 70089, 92966±69, 93006 (skeletons). P. alvaroi: IZUC 70102±05, 75749, 141111 (skeletons). A. eques- 8632, 12104 (alcoholics). P. casuhatiensis: MCZ tris: AMNH 57962, 58352±53, 78074, 78147, 162925 (alcoholic); MCZ 162924 (skeleton). P. 89355 (alcoholics); AMNH 72634, 73848, fasciatus: MVZ 127058, 127061 (alcoholics); 74278±79, 1141137 (skeletons); AMNH 74618 REE 127061 (mandible only). P. scapulatus: (skull); AMNH 126049 (skull and cleared and SDSU 3392±93, 3395±96 (alcoholics); REE double-stained postcranium). A. fuscoauratus: 2381±82, 2509 (skeletons). P. torquatus: SDSU AMNH 125135±52 (alcoholics); AMNH 101383 2249, 2251, AMNH 131856±57 (alcoholics); (skull and cleared and double-stained postcran- MCZ 3586, REE 2766±2767; MZUSP 6982±84, ium). A. meridionalis: AMNH 98413±25 (alco- 65667±78 (skeletons). P. valeriae: USNM holics); AMNH 98421, 98424 (skulls and cleared 165602, IZUC 12037, 12440±41 (alcoholics). P. and double-stained postcrania). A. ortonii: AMNH volcanensis: MCZ 169549 (alcoholic); MCZ 64860±61, 91637, 125354 (alcoholics); AMNH 169550 (skeleton). 56886, 114923 (skulls and cleared and double- Polychrus acutirostris: AMNH 17006, stained postcrania). A. roquet: AMNH 100458±92 104547±49, 62141±42, 75303±04, AMNH 82299, (alcoholics); AMNH 93806, 94598 (skeletons); 75303±04, 101461, 17006, 90274, 104546, AMNH 100472, 100476 (skulls and cleared and 101462, 38806 (alcoholics); AMNH 104549 double-stained postcrania). A. vermiculatus: (skull and cleared and double-stained post- AMNH 76521±26, 144670±82 (alcoholics); cranium); REE 568, 4412, 4488 (skeletons). P. AMNH 70093 (skeleton); AMNH 63062 (skull femoralis: FMNH 34303, 81404±05, (alcoholics); and cleared and double-stained postcranium). FMNH 81405, PUC 7302 (skulls and cleared and Aperopristis catamarcensis: SDSU 1005 (alco- double-stained postcrania). P. gutturosus: AMNH holic); MCZ 96709 (skeleton); FML 0670.2 (skull 13426, 13518, 32674±77, 108990, 13424±25, and cleared and double-stained postcranium). A. 13536, 103744, 104527, 120009, 16338, 16391 paronae: MCZ 162923 (skeleton); FML 00035 (alcoholics); AMNH 32675±76 (skulls and (alcoholic/subsequently skull and cleared and cleared and double-stained postcrania). P. liogas- double-stained postcranium). Diplolaemus bi- ter: AMNH 1679, 6764±65, 22509, 23141, broni: AMNH 80042, 46427±29 (alcoholic). Di- 37812,, 56416, 101452, 101459±60 (alcoholics); plolaemus darwini: AMNH 17005 (alcoholic); AMNH 101460 (skull and cleared and double- MVZ 93047, 93035±36, 200505±06 (skeletons). stained postcranium). P. marmoratus: AMNH Enyalius bibroni: AMNH 131861 (alcoholic). E. 13420±23, 13415±16, 29326, 32279±81, 57211± bilineatus: AMNH 120471, UMMZ 65946 (al- 15, 57217±23, 57225, 57227±30 (alcoholics); coholics); REE 1658, 1958, MZUSP 43019, REE 346, 2283, 2496, MVZ 174843; AMNH 43022±25 (skeletons). E. boulengeri: MCZ 79025 141084, 141130 (skeletons); AMNH 71170±71 (alcoholic); MCZ 163781; MCZ 163781, MZUSP (skulls). P. peruvianus: AMNH 28633±35 (alco- 17452±54, 43020 (skeletons). E. brasiliensis: holics); MVZ 82413 (skeleton). Urostrophus AMNH 62143, MCZ 3317, 4251, UMMZ 108627 gallardoi: BMNH 1902.5.22.4 (alcoholic); MCZ (lot of 3) (alcoholics); MZUSP 3232, 43023, REE 162920 (skeleton). U. vautieri: BMNH 1960 (skeletons); MZUSP 43046 (skull). E. ca- 57.10.28.66 (alcoholic); MCZ 7319, 84036, REE tenatus: MCZ 7320, 3717, 4251, UMMZ 204084 2507, BMNH 94.9.15.3 (skeletons). (alcoholics); REE 1961; MZUSP 43016, 78382 MOLECULAR VOUCHERS AND GENBANK ACCES- (skeletons). E. iheringii: SDSU 2222±23, AMNH SION NUMBERS. Anisolepis longicauda: UNNEC 74964, 120469, UMMZ 108628±29, 204085 (al- 891 (GenBank AF338336). Anolis carolinensis: T. coholics); MCZ 6316, REE 1959, MZUSP 702, Jackman unnumbered from Bimini, 12 April 1992 43024, 43025, 42701, 80630 (skeletons). E. lee- (GenBank AF338324). A. fuscoauratus: KU chii: BMNH 1946.8.9.7, OMNH 36690±92 (al- 214946 (W. E. Duellman 57670) (GenBank coholics) (36690 subsequently cleared and dou- AF338337). A. meridionalis: USNM Field (Lee 20 AMERICAN MUSEUM NOVITATES NO. 3343

Fitzgerald) 166692 (GenBank AF338332). A. or- AF338340). E. leechii: LSU H13958 (formerly L. tonii: Mus. Javier Prado, Lima, unnumbered (W. Vitt) (GenBank AF338342). Leiocephalus bara- E. Duellman 57705) (GenBank AOU39561). A. honensis: T. Titus unnumbered (GenBank roquet: AMNH (C. J. Cole 6387) (GenBank AF338327). Oplurus cychlurus: UMMZ 197023 AF338339). Aperopristis paronae: FML unnum- (GenBank AF338334). Polychrus acutirostris: bered (S. Torres) (GenBank AF338328). A. cata- UNNEC 1368 (GenBank AF338331). P. femor- marcensis: FML unnumbered (S. Torres) (Gen- alis: L. A. Coloma 2568 (GenBank AF338335). Bank AF338341). Basiliscus basiliscus: MVZ P. gutturosus: OMNH unnumbered (J. P. Caldwell 137675 (GenBank AF338330). Diplolaemus 10187) (GenBank AF338338). P. marmoratus: darwini: MVZ (R. D. Sage 13041) (GenBank AMNH (C. J. Cole 6513) (GenBank AF338329). AF33826). Eumeces egregius: GenBank Pristidactylus scapulatus: SDSU 3448 (GenBank AB016606. Enyalius bilineatus: MZUSP unnum- AF338333). Urostrophus gallardoi: FML unnum- bered (M. Rodrigues LG814) (GenBank bered (S. Torres) (GenBank AF338325).

APPENDIX 2

Morphological Transformation Series

Morphological characters were drawn from di- maximum longitudinal diameter of the orbit (as rect observation of alcoholic, dry skeletal and measured from the anterior to the posterior mar- double cleared-and-stained specimens (see appen- gin of the ciliary patch) in Enyalius. The reverse dix 1). General nomenclature of squamation fol- is true in all other taxa. lows Smith (1949). Because POY treats all char- 2. Mental scale (Williams, 1988; Etheridge and acters showing multistate cells (polymorphism) as de Queiroz, 1988; Frost and Etheridge, 1989): (0) nonadditive, where this might have presented a undivided; (1) divided. In Polychrus and the ano- problem in analysis, we transformed additive mul- les the mental scale is partly or completely divid- tistates into sequentially numbered bistate char- ed by a median groove. In Leiosaurus paronae acters. This is re¯ected in the numbering system the mental is so reduced in size, or fragmented, of the characters below. Also, because POY starts that it is not evident whether a mental scale is counting characters with ``0'' rather than with the present so it is coded as unknown. traditional ``1'', we also started our numbering 3. Nasal scale±canthus rostralis: (0) canthus convention with 0 so as to make direct interpre- rostralis relatively straight with nasal scale de®- tation of output (appendix 4) relatively straight- nitely below it; (1) anteriorly the canthus de¯ect- forward. ed medially, with the anteriormost canthals re- duced; no development of a postnasal/subnasal SQUAMATION AND FORM OF HEAD ridge involving anterior loreals; (2) canthal ridge 0. Rostral sutures (Etheridge and de Queiroz, involves anteriormost loreals to form a ridge pos- 1988): (0) rostral scale without sutures, undivid- terior to the nasal scale; canthals either very ed; (1) rostral scale with a pair of posterior su- weakly ridged or no trace of anteriormost canthal tures; (2) rostral scale divided medially. The ros- ridge; (3) canthus obsolete. Although an argu- tral scale is divided medially into a pair of scales ment could be made for ordered 0±1±2, condition subequal to the adjacent anterior supralabials in 3 cannot be placed in this series and could con- Leiosaurus catamarcensis and L. paronae. In Po- ceivably be intercalated between any two of the lychrus acutirostris the rostral scale exhibits pos- other characters. We have therefore treated this terior sutures that do not partition the rostral. Be- set of characteristics as nonadditive. cause we had no ontogenetic reason to assume a 4. Nasal scale±postrostral scale contact: (0) in morphocline, this character is treated as nonad- contact; (1) separated. Contact is intraspeci®cally ditive. variable in Pristidactylus achalensis and P. tor- 1. Snout, orbit relative lengths (Etheridge, quatus. 1969): (0) snout length greater than orbital di- 5. Nasal±labial contact: (0) nasal scale not in ameter; (1) orbital diameter greater than snout broad contact with supralabial(s); (1) nasal scale length. The length of the snout (as measured from in broad contact with second or third supralabial the anterior border of the rostral to the anterior scales. corner of the orbit, as determined by the ciliary± 6. Nasal scale, nostril: (0) nasal scale with preorbital scale contact zone) is less than the rounded margin; except for suture with suprala- 2001 FROST ET AL.: POLYCHROTID LIZARDS 21 bial, nostril almost as large as scale; (1) nasal (smaller than scales of supraorbital semicircles); scale polygonal, nostril much smaller than scale. (1) in narrow contact, or separated by a single In species of Polychrus, except for P. femoralis, row of small scales; (2) separated by a single row the nasal scale is very large with respect to the of large scales that are about as large as those of size of the nostril and is quite unlike the condition the supraorbital semicircles; (3) more than one found in other polychrotids. pair of scales in broad contact. Although descrip- 7. Nasal scale position: (0) entire nasal scale tions make the characteristics sound as if they are closer to anterior tip of rostral than to anterior part of a single ordered series, there are a number border of orbit; (1) center of nasal scale roughly of sources of variation that are confounded by equidistant to anterior border of orbit and anterior them. For instance, in Polychrus the head shield border of rostral. is generally quite large with seeming reduced su- 8. Superciliary scales (Etheridge and de Quei- praoculars, whereas in Diplolaemus the appear- roz, 1988; Frost and Etheridge, 1989): (0) imbri- ance is of widened supraorbital semicircles and cate and elongate; (1) nonimbricate and short. in Leiosaurus is apparently associated with a gen- 9. Supraocular scales (Etheridge, 1969): (0) not eral reduction of scale size. For this reason we carinate or very weakly carinate (may be rugose consider this transformation nonadditive. or swollen); (1) strongly longitudinally carinate 16, 17. Infralabial scale number: (0) 7±7 or or ``pyramidal.'' fewer; (1) 8±8 to 12±12; (2) 14±14 or more. This 10. Eyelids: (0) not fused at anterior and pos- ordered transformation was cast as two columns terior corners of the meatus, iris clearly visible in the data matrix because of intergeneric varia- around the pupil; (1) partly fused, constricting the tion within the Corytophanidae. ocular meatus to roughly the size of the pupil. 18. Mesoptychial scales (Etheridge, 1969): (0) Polychrus is unique in pleurodont iguanians in not conical; (1) conical, with naked skin evident having this characteristic. between. The mesoptychals of Enyalius are dis- 11. Frontal region (Etheridge and de Queiroz, tinctly conical. In all other taxa they are convex 1988): (0) ¯at or slightly convex; (1) concave. or ¯at; in Polychrus they may be elongate. 12. Head scale striae: (0) ®ne, linear rugosities 19. Gular fold (Etheridge and de Queiroz, on head scales absent; (1) ®ne, linear rugosities 1988; Frost and Etheridge, 1989): (0) present; (1) (striae) present on lateral gulars, infra- and su- absent. A gular fold (sensu Frost, 1992) is absent pralabials, and scales of the snout, present or not in Anisolepis longicauda, Polychrus, and anoles on other head scales. Linear striae may be ob- (except Anolis vermiculatus) but is present in all scured in large adults, especially in Polychrus pe- other taxa. Anolis equestris (especially evident in ruvianus, by large, swollen rugosities. juveniles) exhibits a midventrally narrowly in- 13. Interparietal scale and ``eye'': (0) interpa- complete gular fold that is rendered incomplete rietal scale differentiated, with a distinct, opales- by the well-developed dewlap. (The gular fold in cent ``eye''; (1) interparietal scale and ``eye'' dif- A. vermiculatus apparently is ``allowed'' by the ferentiated in juveniles, obscure or absent in rather small dewlap.) Polychrus femoralis has a adults; (2) interparietal scale and eye undifferen- strongly developed antegular fold, which super- tated in juveniles and adults. We have treated this ®cially looks like a gular fold. However, this fold transformation as additive because condition 1 is is not con¯uent with the dorsolateral fold (which developmentally intermediate between 0 and 2. is normally con¯uent with the gular fold when 14. Subocular scales (Etheridge and de Queiroz, present). Instead, it sits anterior to the position 1988): (0) one greatly elongate, more than three where a gular fold would, thus justifying our sup- times longer than any other; (1) one or more mod- position that this is an antegular fold. erately elongate, none three times longer than 20. Gular crest: (0) absent; (1) a short midven- wide; (2) subequal, none elongate. The holotype tral row of compressed, projecting gular scales of Enyalius bibronii has state 2, but according to that form a short anterior crest; (2) a long row of Jackson (1978), one subocular usually is longer compressed, projecting scales extends most of the than the others, so we have treated this as inter- length of the dewlap. This transformation was speci®cally variable. Other species that are inter- treated as additive for the reason that condition 1 speci®cally variable are Enyalius pictus and Po- is developmentally intermediate between 0 and 2. lychrus peruvianus (1±2), Diplolaemus (0±1), and 21. Antegular fold (Frost, 1992): (0) present; Pristidactylus achalensis and P. torquatus (0±1± (1) absent. 2). The transformation is treated as nonadditive for 22. Dewlap (Etheridge and de Queiroz, 1988; reason that intraspeci®c variation suggests multiple Frost and Etheridge, 1989): (0) absent; (1) pres- means of transitioning between states. ent, small, supported by second ceratobranchials 15. Supraorbital semicircles (Etheridge, 1969): that terminate at the antegular fold, below the lev- (0) separated by two to four rows of small scales el of the clavicles; (2) present, large, supported 22 AMERICAN MUSEUM NOVITATES NO. 3343 by second ceratobranchials that extend well back tylus (Etheridge and Williams, 1985; Lamborot below the sternum. and Diaz, 1987), Diplolaemus, Leiosaurus belli, L. catamarcensis, and in Enyalius catenatus, E. bilineatus, E. iheringii, E. perditus and E. brasi- SQUAMATION OF TRUNK AND TAIL liensis (Etheridge 1969, Jackson, 1978). We have 23. Middorsal scale row (Etheridge and de treated this transformation as nonadditive due to Queiroz, 1988; Frost and Etheridge, 1989): (0) our inability to perceive any kind of morphocline present, continuous or nearly so (middorsals may or developmental progression. Leiocephalus and be separated by medial contact of occasional pairs oplurids are coded as unknown because of a cla- of paravertebral scales at the level of the shoul- distically basal dichotomy in this feature in their ders or hips), forming a distinguishable (if occa- respective cladograms (Titus and Frost, 1996; sionally low) ridge of enlarged keeled scales; (1) Pregill, 1992). present, but discontinuous; (2) absent (with oc- 28. Ventrolateral row of enlarged scales (Eth- casional individuals in some taxa showing very eridge and Williams, 1991): (0) absent; (1) pre- weak development of a line of scales over the sent, interrupted or continuous. In Anisolepis (A. sacral region); in some taxa the appearance of a grilli variably) a distinctive ventrolateral row of dorsal scale row is made by the linear enlarge- enlarged scales is evident. ment of adjacent dorsalmost scales; (3) middorsal 29. Proximal caudal scales: (0) keeled; (1) row formed of tubercles rather than of com- smooth. The dorsal scales of the tail are keeled pressed, keeled, or convex scales. We have treat- in Enyalius (weakly in E. bibronii), Anisolepis, ed this set of characters as nonadditive because and Polychrus peruvianus, and faintly keeled in we can easily envision a rather large number of P. marmoratus, P. acutirostris, and P. guttrosus. different processes producing the observed vari- They are smooth on the dorsal surface of the ation. proximal one third (or more) of the tail in Pris- 24. Paravertebral scale shape: (0) polygonal; tidactylus, Diplolaemus, Leiosaurus belli, and (1) rounded (may be convex or tubercular). The Urostrophus. In some Diplolaemus all caudal paravertebral scales of the body are polygonal in scales are smooth. Enyalius and Anisolepis and are rounded in Pris- 30. Caudal annuli: (0) regular, forming seg- tidactylus, Diplolaemus, Leiosaurus, and Uros- ments of four to six scale rows separated by near- trophus. Both states are present in individuals of ly vertical scale sutures; (1) irregular, vertical su- Polychrus liogaster and Enyalius iheringi. tures, if present con®ned to ventral half of scale. 25. Paravertebral scale surface: (0) unicarinate; This character is largely congruent with the pres- (1) tuberculate; (2) smooth; convex or ¯at; (3) ence/absence of autotomy septa, except in Diplo- mixed unicarinate and multicarinate. We have laemus, in which partially fused septa are not ac- treated these characters as nonadditive due to our companied with scale annuli on the tail. inability to discern any particular morphocline 31. Tail: (0) not prehensile; (1) prehensile. beyond setting the root on the unicarinate con- Urostrophus vauteri and Anisolepis grilli were re- dition. ported to have a prehensile tail by Etheridge and 26. Lateral body scales: (0) not in oblique rows Williams (1991). According to Cabrera (in litt.) of rectangular scales; (1) in oblique rows of large, the tail is prehensile in Urostrophus gallardoi and more-or-less rectangular scales with rounded cor- according to J. Williams (in litt.) the same is true ners, separated by skin beset with small, irregular for Anisolepis longicauda. Hoogmoed (1973: thickenings. The dorsal body scales of these taxa 183) rejected earlier reports of tail prehensility in are smooth or keeled, rounded and imbricate or Polychrus. nonoverlapping; however, the lateral body scales of Polychrus are nearly unique in their oblique SQUAMATION AND MORPHOLOGY OF LIMBS arrangement, as also found in Anolis equestris. 27. Ventral body scales: (0) unicarinate; (1) 32. Supradigital scale shape: (0) not all supra- smooth; (2) mixed unicarinate and multicarinate, digitals of third phalanx of third ®nger at least or all multicarinate. Ventral scales are unicarinate twice as broad as postdigitals of third phalanx; in Leiosaurus paronae (Peracca, 1897), Anisole- (1) all supradigitals of third phalanx at least twice pis (Etheridge and Williams, 1991), and Enyalius as broad as postdigitals of third phalanx. All po- pictus and E. bibronii (Etheridge, 1969). They are lychrotids have expanded supradigitals, as do unicarinate in Polychrus acutirostris, tricarinate oplurids, with interspeci®c variation ranging from in P. peruvianus, and variably unicarinate, 1.5 to about 3 times wide than long. Individuating smooth or tricarinate in P. marmoratus and P. classes of variation more than twice the diameter gutturosus. The ventrals are smooth in Urostro- of the supradigital scales proved impossible. phus (Etheridge and Williams, 1991), Pristidac- 33. Supradigital scale keels: (0) smooth; (1) 2001 FROST ET AL.: POLYCHROTID LIZARDS 23 unicarinate; (2) some or all multicarinate. We chrus, Diplolaemus, Pristidactylus fasciatus, and treat this set of characters as nonadditive due to Urostrophus. It extends to a point between the variation in carination not forming a develop- shoulder and the orbit in Anisolepis, Pristidacty- mental series. lus (except P. fasciatus), and Leiosaurus and to 34. Postdigital scales of third ®nger: (0) single a point anterior to the middle of the orbit in En- lateral row penetrating proximally to penultimate yalius. phalanx; (1) double postdigital row penetrating 41. Femoral pores (Etheridge and de Queiroz, proximally to penultimate phalanx; (2) triple 1988; Frost and Etheridge, 1989): (0) present; (1) postdigital row penetrating proximally to penul- absent. Femoral pores are present in Polychrus, timate phalanx. and are absent in all other polychrotids. 35. Subdigital lamellae of toes (Etheridge and de Queiroz, 1988; Frost and Etheridge, 1989): (0) SCALE ORGANS most with three distinct keels; (1) usually a sin- gle, asymmetrical keel; (2) smooth; (3) most with 42. Scale organ of dorsum (Etheridge and de four to six distinct keels. We treat this as non- Queiroz, 1988): (0) without spinules; (1) spinules additive. present. 36. Proximal subdigital lamellae of third toe: 43. Condition of spinulate scale organs of dor- (0) not swollen, or only weakly; (1) swollen and sum (Etheridge and de Queiroz, 1988): (0) with weakly projecting; (2) swollen and projecting to a tuft of mildly elongate spinules; (1) most or all form a pectinate margin. The justi®cation for with a long ®lament formed by many spinules treating this as an additive multistate is that the twisted together; (2) spinulate but without a cen- pectinate margin is necessarily a subset of the set tral tuft of more elongate spinules. The scale or- of taxa showing swollen subdigital lamellae of gans seen in Enyalius leechii are unique; that is, the third toe. they are spinulate, with the spinules short and 37. Distal subdigital lamellae (Etheridge and de without a central tuft of elongate spinules. Taxa Queiroz, 1988): (0) not divided; (1) longitudinal- lacking scale organs were coded as unknown to ly grooved or divided. The distal subdigital scales avoid implicitly weighting in concert with the are not divided in Polychrus, Anisolepis, Uros- previous character transformation. We treat this trophus, and in those species of Enyalius and set of characters as nonadditive because we have Leiosaurus that have multicarinate subdigital la- no compelling theory of transformation. mellae. The distal subdigital lamellae are divided 44. Subdigital surface microstructure (Peterson longitudinally, often with a median notch, in Pris- and Williams, 1981; Peterson, 1983): (0) honey- tidactylus (including P. fasciatus, which has comb; (1) spinulate; (2) subdigital scale surface keeled subdigital lamellae), Leiosaurus belli, and is covered with minute, rounded knobs, all with those Enyalius that have smooth subdigital la- a blanket of short, blunt, minute projections (En- mellae, or subdigital lamellae with a single keel yalius leechii). We treat this set of characters as (Etheridge and Williams, 1985; Etheridge, 1969). nonadditive because we lack a compelling theory of transformation. 38. Digital pads: (0) absent; (1) present. The absence of a digital pad in Anolis onca and its reduction in A. chrysolepis and A. auratus are GENERAL BODY PLAN AND COLORATION considered to be due to ``retrograde'' evolution 45. Sexual size dimorphism: (0) males larger by Peterson and Williams (1981), and in Cha- than females; (1) females larger than males. Max- maelinorops by Peterson (1983). We have ac- imum snout±vent lengths of males are greater in cepted this view at face value for the purposes of Pristidactylus, except for P. volcanensis, where it this analysis. is unknown. Maximum adult size of females is 39. Third toe length: (0) third toes distinctly greater in Anisolepis, Urostrophus, Diplolaemus, shorter than fourth toe; (1); third and fourth toes Leiosaurus, Enyalius (Etheridge, 1969; Jackson of approximately equal length. Polychrus is char- 1978), and Polychrus (except possibly P. peru- acterized by its unusual possession of third and vianus, which is coded as unknown). fourth toes of equal or subequal length. We have 46. Sexual dichromatism: (0) present; (1) ab- not been able to individuate characters beyond sent. Marked sexual dicromatisim occurs in the this, although note that the difference in length Argentinean species of Pristidactylus, P. torqua- between the toes in question appears to be some- tus (Etheridge and Williams, 1985), and P. val- what greater as a trend in Enyalius than in the eriae. It is apparently absent in P. volcanensis remaining taxa. (Lamborot and Diaz, 1987) and is unknown in P. 40. Hindlimb length: (0) short; (1) medium; (2) alvaroi. It is present in Enyalius except E. bili- long. The fourth toe of the adpressed hindlimb neatus; male E. bibronii are unknown (Etheridge, does not extend beyond the shoulder in Poly- 1969; Jackson, 1978). Sexual dichromatism is ab- 24 AMERICAN MUSEUM NOVITATES NO. 3343 sent in Diplolaemus and Leiosaurus (Etheridge roz, 1988: 346; Frost and Etheridge, 1989): (0) and Williams 1985) and in Anisolepis and Uros- super®cial outline of osseous labyrinth distinctly trophus (Etheridge and Williams, 1991). Sexual above the level of the opisthotics, although only dichromatism is present in Polychrus acutirostris of low to moderate elevation; (1) high elevation and absent in P. marmoratus (Vanzolini, 1983). of the osseous labyrinth above the level of the 47. Black antehumeral bar: (0) absent in adult opisthotic. males; (1) present in adult males. A conspicuous, 55. Calci®ed nuchal endolymphatic sacs (Eth- wide, black vertical bar on each side in front of eridge and de Queiroz, 1988; Frost and Etheridge, the shoulder is present in adult males of Pristi- 1989): (0) absent; (1) present. dacytlus scapulatus, P. achalensis, P. torquatus, 56. Supratemporal bones (Etheridge and de and P. fasciatus; the bars are present but narrower Queiroz, 1988; Frost and Etheridge, 1989): (0) and partly hidden within the antehumeral fold in mostly on lateral side of supratemporal process P. casuhatiensis. In P. valeriae the bars may be of parietal; (1) more-or-less equally on both sides faint or absent but are evident in some individu- of the supratemporal process of the parietal. En- als. P. volcanensis in our single alcoholic speci- yalius species have the anterior tip of the supra- men appears to have a faint dark antehumeral bar. temporal moved posteriorly, causing a more me- 48. Dorsal color pattern: (0) not ¯eur-de-lis; (1) dial exposure of the supratemporal on the par- ¯eur-de-lis. The species of Leiosaurus are con- occipital processes. Additionally, they may have spicuous by their possession of a ¯eur-de-lis dor- a ventral groove on the paroccipital process, sal pattern. which makes evaluation of this feature dif®cult. Frost and Etheridge therefore coded Enyalius as HEMIPENIS ``0'' in error. Urostrophus and Anisolepis seem to have reduced supratemporals, but this is dif®cult 49. Hemipenis (Arnold, 1984; Frost and Eth- to evaluate against the variation in Enyalius. eridge, 1989; BoÈhme, 1988): (0) unicapitate; (1) 57. Crista ventrolateralis of basisphenoid and bilobate. basioccipital in adults: (0) well developed and sharp on basioccipital and basisphenoid; (1) CRANIAL SKELETON rounded or absent on basioccipital. 50. Parietal foramen (Etheridge and de Quei- 58. Sphenoccipital process: (0) long, extending roz, 1988; Frost and Etheridge, 1989): (0) pre- to spheno-occipital tubercle; (1) absent or short, sent; (1) absent. terminating well short of spheno-occipital tuber- 51. Lacrimal bone (Etheridge and de Queiroz, cle. The sphenoccipital process extends posteri- 1988; Frost and Etheridge, 1989): (0) on orbital orly with ontogeny, so it must be evaluated in rim; (1) excluded from orbital rim. Considerable older adults. variation exists in the morphology of the lacrimal 59. Coronoid lateral process (Etheridge and de region. In some cases exclusion of the lacrimal Queiroz, 1988; Frost and Etheridge, 1989): (0) from the orbital margin is obtained by contact of large, extending anterolaterally to overlap den- the prefrontal and jugal on the orbital margin, tary; (1) absent or labial process of coronoid not even though the lacrimal bone is exposed laterally extending anterolaterally, but posterolaterally (P. acutirostris: REE 4412 and 568). The lacrimal along margin of dentary. bone is intraspeci®cally variable in its exclusion 60. Splenial, anterior extent (Etheridge and de from the orbital margin in P. marmoratus. The Queiroz, 1988; Frost and Etheridge, 1989): (0) condition is unknown in Pristidactylus alvaroi extends anteriorly more than 25% length of tooth and P. valeriae. row; (1) extremely short or absent, not extending 52. Postfrontal bone (Etheridge and de Quei- anteriorly more than 25% length of tooth row. roz, 1988: 346; Frost and Etheridge, 1989): (0) 61. Splenial, posterior extent (Frost and Eth- present; (1) absent. Frost and Etheridge coded eridge, 1989): (0) terminates posteriorly anterior Polychrus as ``present'' in error. Listed as un- to anterior edge of mandibular fossa; (1) termi- known in E. bibroni, Pristidactylus alvaroi, P. nates posterior to anterior edge of mandibular fos- fasciatus, and P. valeriae due to unavailability of sa. material. 62. Angular (Etheridge and de Queiroz, 1988; 53. Dermal roof bone rugosities (Etheridge and Frost and Etheridge, 1989): (0) moderate to large de Queiroz, 1988; Frost and Etheridge, 1989): (0) with suture line of contact with splenial on the absent or weak, although indistinct rugosities lingual face of the mandible; (1) absent or re- may be present; (1) strong rugosities that corre- duced to splint; if present, suture or angular with spond to scale outlines extend over parietal and splenial on the inferior margin of the mandible. frontal and adjacent dermal skull bones. 63. Dentary, posterior extent (Pregill, 1984; 54. Osseus labyrinth (Etheridge and de Quei- Etheridge and de Queiroz, 1988; Frost and Eth- 2001 FROST ET AL.: POLYCHROTID LIZARDS 25 eridge, 1989): (0) short, more-or-less at level of num; (1) sternum approaches junction of lateral coronoid apex; (1) extends beyond a point 30% and posterior processes of interclavicle closely. of distance from coronoid apex to anterior edge 71. Sternum, median sternal fontanelle (Frost of articular fossa. and Etheridge, 1989): (0) absent; (1) median. The 64. Posterior mylohyoid foramen (Frost and taxa here coded as having a median sternal fon- Etheridge, 1989): (0) on medial face of mandible; tanelle were coded as not having a median sternal (1) on ventral or ventrolateral face of mandible. fontanelle by Frost and Etheridge (1989), because 65. Retroarticular fossa: (0) well developed, of the dif®culty of coding across all iguanian occupies space greater than half the size of artic- taxa. The apomorphy in that case was limited to ular surface; (1) reduced, occupies less than half the phrynosomatid condition of a very large me- the size of articular surface. Urostrophus and Po- dian fontanelle. In this case the fontanelle is con- lychrus are coded as ``0'', but the condition they siderably smaller though de®nitely present. exhibit approaches ``1''. 72. Scapular fenestra (Etheridge and de Quei- 66. Pterygoid teeth (Etheridge and de Queiroz, roz, 1988; Frost and Etheridge, 1989): (0) absent; 1988; Frost and Etheridge, 1989): (0) present; (1) (1) present. absent. In addition to pterygoid teeth, polychro- 73. Posterior coracoid fenestra (Etheridge and tids variably possess palatine teeth. Frost and Eth- de Queiroz, 1988; Frost and Etheridge, 1989; Eth- eridge (1989) coded all polychrotids as possess- eridge and Williams, 1991): (0) absent; (1) pre- ing palatine teeth, excepting Polychrus. This is in sent, marginal and weak. Widely open posterior error inasmuch as Anisolepis and Urostrophus coracoid fenestrae are not found in the polychro- (intraspeci®cally variably) lack palatine teeth, and tids. However, marginal and weak ones along the the only Pristidactylus for which we have ade- edge of the ``window'' where the fenestra quate samples (P. achalensis) is also variable as ``should'' be are found in the coracoid. are all species of Enyalius. Chamaeleolis has pal- 74, 75. Sternal ribs (Etheridge and de Queiroz, atine teeth, but Anolis does not. Therefore, wide- 1988; Frost and Etheridge, 1989): (0) four; (1) spread intraspeci®c variation and small sample three, with posterior extremity of sternum not sizes make us uncomfortable with coding palatine elongated to form parallel rods continuous with teeth or considering them as evidence for rela- xiphisternal rods, and bearing third pair of ribs tionships. aticulating via synovial joints; (2) two, with pos- 67. Marginal teeth (Etheridge and de Queiroz, terior extremity of sternum elongated to form par- 1988; Frost and Etheridge, 1989): (0) tricuspid allel rods continuous with xiphisternal rods, and with sides varying from parallel to moderately last pairs of ribs aticulating via synovial joints. ¯ared; (1) tapered blunt crown with reduced lat- Because of intraspeci®c variation in species of eral cusps; (2) slender, sharp crown with reduced Anisolepis and within genera (Chalarodon and lateral cusps. Characterization of states within ob- Oplurus) in the Opluridae, this transformation served variation is made dif®cult by intraspeci®c had to be cast in two columns. and interindividual variation and many sources of 76. Postxiphisternal inscriptional ribs (Ether- variation. We treat this set of characters as non- idge and de Queiroz, 1988; Frost and Etheridge, additive because we lack a compelling theory of 1989): (0) none midventrally continuous; (1) cha- transformation. meleon-like, with one or more forming midven- trally continuous chevrons, even if laterally not POSTCRANIAL SKELETON con¯uent with ribs. 77. Postxiphisternal inscriptional chevrons that 68. Clavicle (Etheridge and de Queiroz, 1988; are attached to dorsal ribs: (0) 1±5; (1) 8±11. Ter- Frost and Etheridge, 1989): (0) broad, with blade minal taxa that lack postxiphisternal inscriptional forming a sharp angle; (1) slender, with no dis- ribs attached to dorsal ribs are coded as unknown. tinctive lateral angular blade. Diplolaemus has its 78. Transverse processes of caudal vertebrae clavicle a little less ¯ared than the rest of the (Etheridge and de Queiroz, 1988): (0) do not ex- leiosaurs, but not approaching condition 0. tend posteriorly beyond 16; (1) extend beyond 16. 69. Insertion of clavicle (Lang, 1989; Frost and 79. Caudal autotomy fracture planes (Etheridge Etheridge, 1989): (0) on suprascapula, although and de Queiroz, 1988; Frost and Etheridge, may contact scapula; (1) on scapula away from 1989): (0) present, although occasionally showing suprascapular margin. ventral fusion; (1) absent. 70. Sternum, anterior extent (Frost and Ether- 80, 81. Total caudal vertebrae: (0) 33±44; (1) idge, 1989): (0) sternum does not approach junc- 46±64; (2) 66±87. Because Pristidactylus torqua- tion of posterior and lateral processes of interclav- tus varies across the 0±1 boundary it was coded icle closely for more than 50% of length of ante- as polymorphic, requiring that the transformation rior process anterior to the lateral horns of ster- be cast into two columns. 26 AMERICAN MUSEUM NOVITATES NO. 3343

APPENDIX 3 Data Matrix for Morphology 2001 FROST ET AL.: POLYCHROTID LIZARDS 27

APPENDIX 3Ð(Continued )

Recent issues of the Novitates may be purchased from the Museum. Lists of back issues of the Novitates and Bulletin pub- lished during the last ®ve years are avail- able at World Wide Web site http://nimi- di.amnh.org. Or address mail orders to: American Museum of Natural History Li- brary, Central Park West at 79th St., New York, NY 10024. TEL: (212) 769-5545. FAX: (212) 769-5009. E-MAIL: sci- [email protected]

a This paper meets the requirements of ANSI/NISO Z39.48-1992 (Permanence of Paper). 28 AMERICAN MUSEUM NOVITATES NO. 3343

APPENDIX 4 Summary of Evidence on Each Stem (Terminal Taxon or Internal Branch) Terminal taxa and internal branch numbers are shown in ®gure 5. ``Morphological character'' numbers correspond to the character numbers listed in appendix 2. ``Ancestral state'' is the condition of the character in the stem antecedent to the referenced terminal taxon or internal branch; the ``Descendant state'' in that taxon or stem is also noted. The ``Molecular changes'' column shows the number of different kinds of molecular changes along the referenced stem or terminal taxon. Asterisks indicate that characters (molecular or morphological) are ``De®nite'' (i.e., placed on the stem or terminal taxon regardless of optimization). 2001 FROST ET AL.: POLYCHROTID LIZARDS 29

APPENDIX 4Ð(Continued ) 30 AMERICAN MUSEUM NOVITATES NO. 3343

APPENDIX 4Ð(Continued ) 2001 FROST ET AL.: POLYCHROTID LIZARDS 31

APPENDIX 4Ð(Continued ) 32 AMERICAN MUSEUM NOVITATES NO. 3343

APPENDIX 4Ð(Continued ) 2001 FROST ET AL.: POLYCHROTID LIZARDS 33

APPENDIX 4Ð(Continued ) 34 AMERICAN MUSEUM NOVITATES NO. 3343

APPENDIX 4Ð(Continued ) 2001 FROST ET AL.: POLYCHROTID LIZARDS 35

APPENDIX 4Ð(Continued ) 36 AMERICAN MUSEUM NOVITATES NO. 3343

APPENDIX 4Ð(Continued ) 2001 FROST ET AL.: POLYCHROTID LIZARDS 37

APPENDIX 4Ð(Continued ) 38 AMERICAN MUSEUM NOVITATES NO. 3343

APPENDIX 4Ð(Continued )