Novitatesamerican MUSEUM PUBLISHED by the AMERICAN MUSEUM of NATURAL HISTORY CENTRAL PARK WEST at 79TH STREET, NEW YORK, N.Y

NovitatesAMERICAN MUSEUM PUBLISHED BY THE AMERICAN MUSEUM OF NATURAL HISTORY CENTRAL PARK WEST AT 79TH STREET, NEW YORK, N.Y. 10024 Number 3081, 30 pp., 11 figures, 2 tables December 27, 1993

Unisexual and Bisexual Whiptail Lizards of the Cnemidophorus lemniscatus Complex (Squamata: Teiidae) of the Guiana Region, South America, with Descriptions of New Species

CHARLES J. COLE1 AND HERBERT C. DESSAUER2

CONTENTS Abstract ...... 2 Introduction ..... 2 Acknowledgments ...... 3 Materials and Methods ...... 4 Results and Discussion ...... 5 Karyotypes ...... 5 Biochemical Genetics ...... 5 Hybrid Origins ...... 10 Reproduction ...... 11 Evolutionary Scenano ...... 12 Implications for Biodiversity ...... 13 Taxonomy ...... 15 Alternative Treatments ...... 25 Summary and Conclusions ...... 26 References ...... 27 Appendix-Specimens Examined ...... 29

' Curator, Department of Herpetology and Ichthyology, American Museum of Natural History. 2 Research Associate, Department ofHerpetology and Ichthyology, American Museum ofNatural History; Professor Emeritus, Department of Biochemistry and Molecular Biology, Louisiana State University Medical Center, New Orleans, LA.

ABSTRACT

At least four species have been masquerading A revised diagnosis is presented for C. lemnis- under the name Cnemidophorus lemniscatus in catus, a male lectotype is designated, and the type northern South America. These species were re- locality is restricted to an area in Suriname with vealed clearly by karyotypic and electrophoretic a bisexual population oflizards similar to the Lin- analyses of samples from unisexual and from bi- naean syntypes and consistent with their suspected sexual populations. The four taxa include two dip- provenance. The second bisexual species appears loid bisexual species and two unisexual cryptic to be C. gramivagus McCrystal and Dixon, for species (one diploid and one triploid). The uni- which a revised diagnosis is presented also. New sexual species are parthenogenetic clones that ul- diagnoses and names are presented for two uni- timately originated from hybrids among the bisexual species, a diploid based on specimens from sexual species (the triploids stemming from Venezuela (with paratypes from Brazil also), and backcross hybridization). There are several differ- a triploid based on specimens from Surname. Al- ent diagnosable diploid clones, at least two ofwhich ternative taxonomic treatments for the various probably had origins from separate hybridization diploid unisexual clones found in Brazil are dis- events at different localities. Several other diag- cussed, but none of these is named as a separate nosable diploid clones probably stemmed from species here. For now, we treat all of the diploid mutations within one or more established parthe- clones as one species. nogenetic lineages of very similar hybrid origin. INTRODUCTION Cnemidophorus lemniscatus (Linnaeus) is catus from Brazil by comparing products of a widespread and generally abundant and 57 presumed structural gene loci in types B conspicuous whiptail lizard in much of Cen- through E, using protein electrophoresis. The tral America and South America, mostly in hypothesis was strongly supported, as bisex- bisexual populations (gonochoristic, with ual types D and E had relatively low hetero- separate sexes; fertilization is required for re- zygosity indices and fixed or nearly fixed dif- production). Vanzolini (1970) first reported ferences from each other at 11 of the loci that some populations along the Amazon analyzed. In contrast, the unisexual types B River in Brazil are bisexual and others are and C had relatively high heterozygosity and unisexual (all-female). genotypes consistent with their having a hap- Chromosome studies by Peccinini-Seale loid set of genes each from bisexual types D and Frota-Pessoa (1974) revealed five slightly and E, with minor exceptions. Sites et al. different karyotypes among the Brazilian C. (1990) proposed that at least three species are lemniscatus, which they designated alpha- involved in this complex along the Amazon betically, A through E. Types A-C are minor River in Brazil: bisexual type D, bisexual type variants characterizing separate unisexual E, and the diploid unisexuals ofhybrid origin. populations, all diploids, differing from each A companion study ofmitochondrial DNA other primarily in a centromere position (Vyas et al., 1990) showed that the Brazilian among the second largest pair of macrochro- unisexuals have only one type ofmtDNA and mosomes and in number of microchromo- it is most similar to that occurring in certain somes. Types D and E characterize the bi- bisexual type D lizards (i.e., type D was the sexual populations, which differ from each maternal ancestor). There is variation in other only in centromere position on chro- mtDNA in type D, and none of the type D mosome number 1 (largest pair). Type C of samples examined matched perfectly the the unisexuals has a karyotype that could be mtDNAs of the unisexuals. Nevertheless, composed of one haploid complement each none of the unisexuals had mtDNA similar from bisexual types D and E; i.e., it could to that ofbisexual type E (i.e., type E was the have had a hybrid origin. Unisexual types A paternal ancestor). and B could have been derived from unisex- Separate bisexual and unisexual popula- ual type C by chromosomal aberrations. tions of C. lemniscatus also occur in Suri- Sites et al. (1990) tested the hybrid origin name (Hoogmoed, 1973), but these unisex- hypothesis for diploid unisexual C. emnis- uals are triploids rather than diploids (Serena, 1993 COLE AND DESSAUER: UNISEXUAL AND BISEXUAL WHIPTAIL LIZARDS 3

1985; Dessauer and Cole, 1989). Studies of type D bisexual C. lemniscatus from Suri- karyotypes and protein electrophoresis (46 name and Venezuela compare favorably with loci tested) suggested that the bisexual dip- those from Brazil, as indicated by karyotypes loids in Suriname are similar to type D of and protein electrophoresis? (2) Do the uni- Brazil (same taxon), and the triploids had a sexual samples from Brazil, Suriname, and hybrid origin involving two genomes of type Venezuela have basically the same genotypes, D and one of type E (Dessauer and Cole, including the alleles at heterozygous loci that 1989). This probably resulted from back- were inherited from bisexual type E? (3) What crossing ofa diploid unisexual oftype C (orig- is the minimal number ofspecies represented inally of type D x type E origin) with a male by the samples compared here, and what of type D; i.e., a double hybrid origin of [D names should be applied to them? A review x E] x D. of these questions improves our knowledge Electrophoretic analyses of the samples ofbiodiversity and its origins in Neotropica. from Suriname and from Brazil were con- ducted separately in different laboratories ACKNOWLEDGMENTS (Dessauer and Cole, 1989; Sites et al., 1990, Permits and considerable assistance with respectively), so cross-correlation gels com- logistics in the field were provided by S. A. paring all of the samples together did not J. Malone, F. L. J. Baal, and M. H. S. Miskin exist. Evidence that the genomes of type D (all of the Suriname Forest Service), Giu- and E were basically the same in unisexual seppe Colonnello (Museo de Historia Natural lizards from Brazil and Suriname was pro- La Salle, Caracas, Venezuela), and Karen Pil- vided only by the karyotypes (Dessauer and grim (Wildlife Services Division, Depart- Cole, 1989: fig. 7C), which do not resolve ment of Agriculture, Guyana). We were as- genetic data as well as protein electrophoresis sisted in the field in Suriname by Jeffrey Cole, does. To determine whether the same alleles in Venezuela by Carlos Guaveco and Maria at structural gene loci are shared by lizards Jose Praderio (Museo de Historia Natural La in Brazil and Suriname (type D bisexuals and Salle), and in Guyana by Bruce Hoffman. diploid and triploid unisexuals), samples Fieldwork in Guyana in 1992 was con- would have to be subjected to protein elec- ducted through the Biological Diversity of trophoresis on the same gels. the Guianas Program (BDGP), coordinated While the two independent electrophoretic in Guyana by Deputy Vice-Chancellor Mal- reports discussed above on lizards from Su- colm Rodrigues and Dean Indarjit Ramdass riname and Brazil were in press, CJC, Carlos ofthe University ofGuyana (UG), Turkeyen Guaveco, Maria Jose Praderio, and Carol R. Campus, Georgetown. Mike Tamessar ofUG Townsend collected two new samples of the also greatly facilitated this work. Important C. lemniscatus complex in Venezuela. Spec- logistical support was provided by personnel imens from San Ignacio de Yuruani (State of serving the BDGP, including Vicki Funk, Bolivar) included both sexes, whereas those Bruce Hoffman, Carol Kelloff, and Moham- from Icabaru (State ofBolivar) included only med Ameer (the females, although no unisexual populations "Ambassador of Bourda"). were known for Venezuela. Additional logistical support was provided For the present report, we karyotyped the by Malcolm Chan-A-Sue (Torong Guyana), new samples from Venezuela and we ana- Walter Lachman (National Dairy Develop- lyzed together in one laboratory the samples ment Programme), Diane McTurk (Karan- (bisexual and unisexual) from Brazil, Suri- ambo), and the people of Karanambo and name, and Venezuela for direct, side-by-side Yupukari. comparisons of alleles detected by protein Brazilian specimens used here are a subset electrophoresis. We also compared these an- ofthe same ones studied by Sites et al. (1990), imals morphologically with each other, with not a product of our own fieldwork. These new samples from Guyana, and with samples were collected by D. M. Peccinini-Seale and ofCnemidophorus gramivagus McCrystal and R. R. de Souza. Thanks to the generosity of Dixon from Colombia. The following ques- Wesley Brown and Craig Moritz (University tions are addressed: (1) Do the genotypes of of Michigan, Ann Arbor) and Jack Sites 4 AMERICAN MUSEUM NOVITATES NO. 3081

(Brigham Young University, Provo, Utah), MATERIALS AND METHODS homogenates remaining after their study (Sites EXTERNAL MORPHOLOGY were sent to us for direct com- et al., 1990) AND KARYOTYPES parisons with our specimens. The relevant preserved lizards at the University of Mich- Characters of external morphology were igan Museum ofZoology (UMMZ) were lent determined as described elsewhere (Cole et to us by Arnold Kluge. al., 1988), as were karyotypes (Cole, 1979). Important information and photographs of Sex was determined by dissection and ex- the Linnaean syntypes of Cnemidophorus amination of primary sexual characters, ex- lemniscatus were provided by Sven 0. Kul- cept for specimens that had been eviscerated lander, Curator of Fishes, Swedish Museum or photographed by others, in which case we of Natural History, Stockholm (NRM). The used secondary characters (anal spurs, size of photographs (figs. 8, 9) were taken by Uno femoral pores, massiveness ofhead). Method Samuelsson of their Department of Paleo- is specified where particularly pertinent, and zoology. the specimens examined are listed in the Ap- The following people have helped in var- pendix. ious ways and particularly encouraged our progress on this work: Vicki Funk (Depart- BIOCHEMICAL GENETIcs ment of Botany, National Museum of Nat- ural History [USNM], Washington), Marinus The heart, liver, and kidney, plus samples Hoogmoed (National Museum of Natural of skeletal muscle and blood were collected History, Leiden, The Netherlands), and Hen- from Surinamese and Venezuelan lizards pri- ry Reichart (Suriname Forest Service). Carol or to preservation. Plasma was separated from R. Townsend assisted in every way possible blood cells and all tissues were stored in liq- in the field and in the laboratory, putting as uid nitrogen until transfer to ultracold freez- much effort into this project as anybody. ers in the laboratory. For the protein studies, We thank Teresa C. S. de Avila Pires (Mu- tissues were thawed and homogenized in two seu Paraense Emilio Goeldi), Darrel Frost to three volumes of a 0.25 M solution of (American Museum of Natural History sucrose, which contained 20 mg/liter of di- [AMNH]), Marinus Hoogmoed, and Jack thiotreitol, 50 mg/liter of NAD, and 25 Sites for critical and very helpful reviews of mg/liter of NADP. Homogenates were cen- the manuscript. Funding for fieldwork in Su- trifuged at 5000 g to separate soluble proteins riname was provided by the National Geo- from cell debris. graphic Society (grants 2622-83 and 3163- Electrophoretic analyses were performed 85), and some of the laboratory work was using supernatant solutions from these ho- supported by the National Science Founda- mogenates and from Brazilian lizard homog- tion (BSR-8105454, to CJC). Funding for enates that remained from the study by Sites fieldwork in Venezuela was provided in part et al. (1990). Aliquots ofthese solutions were by gifts to the AMNH from Robert G. Goelet added to slots in starch gels and subjected to of New York and the Phipps Foundation, as vertical gel electrophoresis (Smithies, 1959) this fieldwork directly followed participation overnight at a potential gradient of 6 to 8 by CJC and Carol R. Townsend in the Tap- volts/cm in a cabinet maintained at 4°C. To irapeco Expedition (see acknowledgments in develop data comparable to that of Sites et Donnelly et al., 1992: 832). Funding for field- al. (1990), most analyses were carried out with work in Guyana was provided by the Bio- gels made with a Tris-citrate buffer of pH 8, logical Diversity of the Guianas Program identical in composition, or nearly so, to those through the Smithsonian Institution, Wash- used by Sites et al. For analyses involving ington (Department ofBotany, USNM; Vicki plasma, which was not available for the Bra- Funk, North American Program Director). zilian lizards, a veronal buffer of pH 8.6 was Special funds ofthe AMNH also helped meet used. For analyses ofunspecified muscle pro- field expenses, thanks to Thomas Nicholson teins a phosphate buffer of pH 6 was used. and Michael Novacek. Gel buffers were diluted to an ionic strength 1993 COLE AND DESSAUER: UNISEXUAL AND BISEXUAL WHIPTAIL LIZARDS 5 ofabout 0.01; electrode chambers were filled populations on the Amazon River in Brazil with buffers 10 times that strength. The co- (Peccinini-Seale and Frota-Pessoa, 1974; Sites enzyme NADP (10 mg/liter) was added to et al., 1990), for Boa Vista in Roraima, north- gel buffers used in analyses of sMEP, IDHP, ern Brazil (Sites et al., 1990), and for Suri- and PGDH to stabilize their structures during name (Serena 1985; Dessauer and Cole, 1989; electrophoresis. fig. 2A here). Following electrophoresis, enzymes and Both of the lizards karyotyped in the uni- nonenzymic proteins were localized on gel sexual sample from Icabaru (a total of nine slices by means of histochemical stains, flu- lizards in the sample, all females) similarly orescence, or autoradiography. Table 1 lists had a diploid number of 50 chromosomes, the proteins analyzed, enzyme commission with 26 macrochromosomes and 24 micro- numbers, abbreviations (Shaklee et al., 1989; chromosomes. In contrast, however, the larg- Murphy et al., 1990), and tissue(s) from which est pair of macrochromosomes is hetero- data were extracted. Localization techniques morphic, with one submetacentric to closely followed descriptions by Harris and subtelocentric chromosome and one subtel- Hopkinson (1976). Transferrins were iden- ocentric to telocentric (fig. 1 B). This is karyo- tified by iron59 binding and autoradiography type C, reported previously for certain uni- (Giblett et al., 1959). Myoglobin was detected sexual populations in extreme eastern Brazil, in muscle homogenates by the presence of a near the mouth of the Amazon (Peccinini- brownish band migrating anodally on un- Seale and Frota-Pessoa, 1974). Karyotype C stained gels; its identity was confirmed with occurs in clones that originated from hybrids the benzidine test (see Smithies, 1959). between bisexual lizards of karyotype D and Tissues and tissue homogenates that re- karyotype E (Sites et al., 1990; Vyas et al., mained at the completion of this study were 1990). retained in the frozen tissue collection of the Karyotypes A and B of unisexual lizards Museum ofNatural Science, Louisiana State characterize variant clones that were proba- University, Baton Rouge (Dessauer et al., bly derived by mutations from karyotype C, 1990). described by Peccinini-Seale and Frota-Pes- soa (1974) in the same paper in which they RESULTS AND DISCUSSION described karyotypes A through E for the Cnemidophorus lemniscatus complex. KARYOTYPES Here, we designate karyotype F, that ofthe New data are presented here for specimens triploids in Suriname described and illus- from two populations in the state of Bolivar, trated by Dessauer and Cole (1989: fig. 7C; Venezuela, as follows: (1) San Ignacio de Yu- fig 2B here). This triploid karyotype com- ruani (one female, no males; 2 cells exam- prises two haploid complements of type D ined); and (2) Icabaru (two females, no males; and one of type E, probably resulting from 11 cells examined). The lizard karyotyped in two steps ofhybridization as described above the bisexual sample from San Ignacio de Yu- ([D x E] x D; Dessauer and Cole, 1989). ruani (a total of 12 lizards in the sample, five males and seven females) had a diploid num- BIOCHEMICAL GENETICS ber of 50 chromosomes, with 26 macrochromosomes (13 homomorphic pairs) and 24 Proteins encoded by 32 presumptive struc- microchromosomes (apparently 12 homo- tural gene loci were compared by electropho- morphic pairs). The largest pair of chromo- resis for 12 samples of C. lemniscatus from somes is submetacentric to subtelocentric Brazil, Venezuela, and Suriname (figs. 3, 4; with a secondary constriction and dotlike sat- table 2). This allowed cross-correlating the ellite on the end of the long arm distal to the protein data published previously for Suri- centromere (fig. IA). The smaller macro- name (Dessauer and Cole, 1989) and Brazil chromosomes are all telocentric or subtelo- (Sites et al., 1990), while also comparing the centric. This is known as karyotype D of C. new samples from Venezuela. lemniscatus, reported previously for certain Twenty loci showed no variation across all 6 AMERICAN MUSEUM NOVITATES NO. 3081

A. Rsa*a e

o i. S.0 #aab **

.* 0* A - m .. 0

eta,*

4. A.

^ . a0* a.& I & a

as 0O4

Fig. 1. Diploid karyotypes (2n = 50) of two species in the Cnemidophorus lemniscatus complex, each with one pair of large chromosomes, 12 pairs of microchromosomes, and 12 pairs of telocentric chromosomes ofintermediate size. A, the bisexual C. lemniscatus (AMNH 135091, female), from San Ignacio de Yuruani, Bolivar, Venezuela. Note that the largest pair consists of two similar submetacentric to subtelocentric chromosomes with the secondary constriction nearly terminal on the long arm. B, the unisexual named below as C. cryptus (paratype, AMNH 135090, female), from Icabaru, Bolivar, Ven- ezuela. Note that the largest pair consists of two chromosomes with dissimilar centromere positions, one as in C. lemniscatus (submetacentric to subtelocentric) and one subtelocentric to telocentric (arrow). Bar represents 10 ,gm. 1993 BCOLE AND DESSAUER: UNISEXUAL AND BISEXUAL WHIPTAIL LIZARDS 7 B, lIs Ba A *1 0. a, a' IA _aa 46

es0 a"P 0S& aSo

a ab am a

Kr~~~~~~.

Ut} lIt 311

.** * * a

e04 e ee* *asdb .0 seaa4 kd

!: '. -0

Fig. 2. Karyotypes of two species in the Cnemidophorus lemniscatus complex. A, the diploid C. lemniscatus (AMNH 133303, female) from Christiaankondre, Suriname. Note the diploid number of 50 chromosomes, all similar to those in figure IA. B, the triploid unisexual named below as C. pseudolemniscatus (holotype, AMNH 133304, female), from Albina, Suriname. Note that the triplet oflargest chromosomes includes two that are submetacentric to subtelocentric and one subtelocentric to telocentric (arrow). Bar represents 10 ,im. Modified from Dessauer and Cole (1989: 57).

12 samples compared, whereas 12 loci were of each of types D and E (fig. 3). The only more informative in various ways (table 2). exceptions are for one allele each at IDDH For seven loci (IDDH, sMEP, mIDHP, sAAT, and sMEP (table 2). We suggest that future PEPA, PEPB, and ADA) the bisexual dip- research will reveal that the b-allele ofIDDH loids of type D and type E consistently differ and the a-allele of sMEP occur in other pop- from each other, and for five of these loci all ulation samples of type E that we have not of the diploid unisexuals (B + C Brz and C yet examined, although other explanations Ven in table 2) share one allele characteristic for these data are possible also (Sites et al., 8 AMERICAN MUSEUM NOVITATES NO. 3081

TABLE 1 Presumptive Structural Gene Loci Examined Locus EC no. Abbrev.a Tissue" Oxidoreductases Alcohol dehydrogenase 1.1.1.1 ADH L Glycerol-3-phosphate 1.1.1.8 G3PDH M dehydrogenase L-iditol dehydrogenase 1.1.1.14 IDDH K L-lactate dehydrogenase 1.1.1.27 LDH-1 K, L LDH-2 K, L, M Malate dehydrogenase 1.1.1.37 sMDH M mMDH M Malate enzyme 1.1.1.40 sMEP K, L Isocitrate dehydrogenase 1.1.1.42 sIDHP K, L mIDHP H, M Phosphogluconate 1.1.1.44 PGDH K, L dehydrogenase Superoxide dismutase 1.15.1.1 sSOD K, L mSOD K, L Transferases Aspartate 2.6.1.1 sAAT K, L, M aminotransferase mAAT K, L, M Creatine kinase 2.7.3.2 CK-1 I, K CK-2 I, K Adenylate kinase 2.7.4.3 AK M Hydrolases Esterase D 3.1.1.- ESTDC L, M Acid phosphatase 3.1.3.2 ACPd L, M Peptidases 3.4.-.- PEPAe H, K, L PEPBf M Adenosine deaminase 3.5.4.4 ADA L, R Lyases Aconitate hydratase 4.2.1.3 sAH L mAH L Isomerases Mannose-6-phosphate 5.3.1.8 MPI K, L Glucose-6-phosphate 5.3.1.9 GPI K, L, M Nonenzymic proteins Transferrin TF P Myoglobin MB M Muscle proteins (N = 4) MP-1 to 4 M a Based on Shaklee et al. (1989) and Murphy et al. (1990); s = cytosolic enzyme; m = mitochondrial enzyme. b Tissue in which enzyme was scored: H = heart; I = intestine; K = kidney; L = liver; M = skeletal muscle; P = plasma; R = erythrocytes. c Substrate 4-methylumbelliferyl acetate; inactive with alphanaphthyl esters. d Substrate 4-methylumbelliferyl phosphate. e Substrate phenylalanyl *leucine. f Substrate leucyl glycyl*glycine.

1990; and see below). At any rate, we are sidering each ofthe 12 most informative loci satisfied that the hybrid origin of all of the (table 2), and considering the karyotypes in- unisexual populations is confirmed (Des- volved (see above), the triploids from Suri- sauer and Cole, 1989; Sites et al., 1990). Con- name (F Sur) have traits that confirm an or- 1993 COLE AND DESSAUER: UNISEXUAL AND BISEXUAL WHIPTAIL LIZARDS 9

TABLE 2 Allelesa or Genotypesb at 32 Presumptive Structural Gene Loci in Twelve Samplesc of Cnemidophorus lemniscatus Bisexual Unisexual E D D D D B+C C F Locusd Brz Alt Boa Ven Sur Brz Ven Sur IDDH c b b b, ae b bb bb bbb sMEP b a a a a ab aa aaa mIDHP b a a a a ab ab aab sAAT a b b b b ab ab abb ESTD a a, bf a, bf b a aa aa aaa PEPA a c b b b ac ab abb PEPB b a a a a ab ab aab ADA a b b b b ab ab abbg sAH b b b b b, ah bb bb bbb mAH b b b b a, bi bb bb abb MPI b b b b b, ai bb bb bbb GPI b b b a, bk b bb bb bbb a Alleles are designated in alphabetical sequence in order of decreasing anodal migration. Commas separate alternative alleles at a specified locus for bisexual species. See Sites et al. (1990) for more details on allele frequencies in Brazilian populations. b Genotypes are given for unisexual forms, for which we observed no variation at any locus in the samples we examined. c E Brz = type E animals (N = 15 specimens examined by us) from Manacapuru, Brazil (Sites et al., 1990); D Alt = type D animals (N = 10) from Alter do Chao, Brazil (Sites et al., 1990); D Boa = type D animals (N = 4) from Boa Vista, Brazil (Sites et al., 1990); D Ven = type D (N = 9) from San Ignacio de Yuruani, Venezuela; D Sur = type D (N = 10) from Christiaankondre, Suriname; B + C Brz = types B and C diploids (N = 23) from Oriximina, Brazil (type B) and Capanema and Maruda, Brazil (type C; Sites et al., 1990); C Ven = type C diploids (N = 4) from Icabaru, Venezuela; and F Sur = type F triploids (N = 8) from Albina, Brokopondo, and Brownsweg, Suriname. d Each specimen of all populations tested had the same phenotype at the following loci: G3PDH, LDH-1, LDH- 2, sMDH, mMDH, sIDHP, PGDH, sSOD, mSOD, mAAT, CK-1, CK-2, AK, ACP, MB, and MP-l through MP- 4. In addition, all specimens from Suriname and Venezuela had the same phenotypes for TF, whether bisexual or unisexual or diploid or triploid. e Frequency of b = 0.79. fFrequency of a = 0.94 for type D Alt and 0.87 for type D Boa. g The product of the b-allele of ADA was very difficult to resolve in tissues that had been frozen for several years, indicating that this molecule is unstable over time. h Frequency of b = 0.78. Frequency of a = 0.67. i Frequency of b = 0.95. k Frequency of a = 0.64. igin involving hybridization between a diploid some limited clonal diversity in these pop- type C unisexual lizard x male of type D. ulation samples. The four diploid unisexuals The triploid unisexuals from three locali- from Venezuela (C Ven) are identical to each ties in Suriname (F Sur; Albina, Brokopondo, other and consistently differ from the Bra- and Brownsweg) are all identical to each oth- zilian unisexuals by one allele each at two er. The diploid unisexuals of which we re- loci (sMEP and PEPA). We suggested above ceived homogenates representing three lo- that this may reflect geographic variation of calities in Brazil (B + C Brz; Capanema, alleles at sMEP in type E bisexuals, and geo- Maruda, and Oriximina) also are all identical graphic variation is clearly demonstrated for to each other, even though this includes in- PEPA. At PEPA (table 2; fig. 4), the Brazilian dividuals of two karyotypic clones (types B diploid unisexuals (B + C Brz) from along and C); however, Sites et al. (1990) discussed the Amazon have the c-allele characteristic 10 AMERICAN MUSEUM NOVITATES NO. 3081

loids from Suriname (F Sur) fit this pattern of separate hybrid origins, stemming from backcross hybridization of the northern type of diploid unisexual (C Ven) x a northern bisexual type D male (D Boa, D Ven, or D Sur, table 2; fig. 4), consistent with their geo-

C graphic distribution. This conclusion for the triploids is also supported by the alleles at the sMEP and mAH loci (table 2). In fact, the a-allele of mAH is known to date only for the bisexuals from Suriname (D Sur). Al- leles at other loci (e.g., ESTD) also show geographic variation (table 2).

HYBRID ORIGINS Our results strongly support the conclusions of Sites et al. (1990) and Vyas et al. (1990) and the conclusions and predictions t b a of Dessauer and Cole (1989), as follows: (1) Fig. 3. Electrophoretic patterns of the enzyme diploid lizards with karyotype D and those sAAT for 12 lizards representing three of the species in the Cnemidophorus lemniscatus complex. with karyotype E represent two separate bi- Abbreviations below the gel (a, b) refer to the dif- sexual species that are genetically distinct even ferent allele products. Abbreviations beside the gel ifnot so distinct morphologically; (2) the spe- refer to the species, as follows: c, cryptus (unisexual cies with karyotype D has a broad distribu- karyotype C lizards from Venezuela, named be- tion in the Guiana Region, including parts of low; diploid heterozygotes; individual in lane five Brazil, Venezuela, and Suriname, at least (fig. from the top is the holotype); g, gramivagus (bi- 5), but probably also including parts of Guy- sexual karyotype E lizards from Brazil, diploids ana and French Guiana; (3) currently, the homozygous for the a-allele, referred to this spe- bisexual species with karyotype E is known cies below); and 1, lemniscatus (bisexual karyotype from only two localities along the Amazon, D lizards from Brazil, Suriname, and Venezuela in Brazil (Peccinini-Seale and Frota-Pessoa, [all illustrated on this gel being from Venezuela], diploids homozygous for the b-allele). The diploid 1974; Sites et al., 1990; Vyas et al., 1990; but unisexual lizards from Brazil had the same pattern see section on Taxonomy, below); (4) the uni- of sAAT as those from Venezuela (heterozygous), sexual lizards of karyotypes B and C from and the triploids from Suriname were similar also Brazil are clones that ultimately originated but with two doses of the b-allele (genotype abb). by hybridization between lizards of karyo- Anode is to the right; arrow indicates positions of types D and E (types A and B probably arose sample applications. by mutation in one or more ancestral parthenogenetic lineages of unisexual karyotype ofbisexual type D animals from the Amazon C); and (5) the triploid unisexual populations (D Alt), whereas the Venezuelan diploid un- (here referred to as type F) from Suriname isexuals (C Ven) have the b-allele character- represent a clone that originated by hybrid- istic of bisexual type D animals from more ization between one or more unisexual liz- northern localities (D Boa, D Ven, and D ards of karyotype C and one or more males Sur). This indicates that the diploid unisex- of karyotype D (ultimately, a double hybrid uals along the Amazon and those from Ven- origin of [D x E] x D). ezuela are products of two separate hybrid In addition, we suggest that the newly dis- origins, involving different populations ofthe covered diploid unisexual sample of karyo- same bisexual ancestors. For other examples type C from Icabaru, State of Bolivar, Ven- of geographic variation in allele frequencies ezuela, represents a clone, as all four in Cnemidophorus, see Cole et al. (1988). specimens examined had identical allele Allele combinations at PEPA in the trip- combinations at all 28 loci examined in each 1993 COLE AND DESSAUER: UNISEXUAL AND BISEXUAL WHIPTAIL LIZARDS I1I

b b b

, E~b c b a Fig. 4. Electrophoretic patterns of the enzyme PEPA for 12 lizards representing the four species in the Cnemidophorus lemniscatus complex. Abbreviations, anode, and arrow are as in figure 3, with the following abbreviations added beside the gel: b, diploid unisexual lizards of karyotypes B and C from Brazil; and p, pseudolemniscatus (triploid unisexual lizard ofkaryotype F from Suriname, named below). Note that gramivagus is homozygous for the a-allele, lemniscatus from Venezuela (fourth lane from top) is homozygous for the b-allele, lemniscatus from Alter do Chao, Brazil (second lane from top) is homozygous for the c-allele, cryptus from Venezuela (third lane from top; this is the holotype) is heterozygous with the ab genotype, cryptus from Brazil (top lane and bottom six lanes) is heterozygous with the ac genotype, and the triploid pseudolemniscatus from Suriname is heterozygous with the abb genotype, all indicating that the unisexuals from Brazil had a hybrid origin separate from those of Venezuela and Suriname. animal, with five loci bearing alleles in the tical, and their genes detected by electropho- heterozygous state (H = 18%) and no evi- resis are almost identical also, excepting one dence of Mendelian inheritance. This clone allele each at two loci (sMEP and PEPA, dis- also appears to have originated by hybrid- cussed above; table 2; fig. 4). These excep- ization between lizards of karyotypes D and tional alleles suggest that the diploid unisex- E, having a complete haploid set from each ual lizards along the Amazon and those to of the chromosomes (fig. 1B) and genes de- the north (Venezuela) had separate hybrid tected by protein electrophoresis (table 2; figs. origins, and the triploids from Suriname share 3, 4), with the exception ofonly one allele (at a common ancestry with the northern uni- sMEP, discussed above) out of the loci re- sexual diploid clone (see above). solved in these lizards (up to 32). Diploid unisexual lizards of karyotype C from Venezuela are separated by approxi- REPRODUCrION mately 1800 km from those in eastern Brazil The 12 diploid unisexual Brazilian lizards (fig. 5). Lizards from these two widely sepa- ofkaryotypes B and C available to us and for rated areas appear to be karyotypically iden- which we collected the fullest suites of data 12 AMERICAN MUSEUM NOVITATES NO. 3081

Fig. 5. Map of the Guiana Region of South America showing localities for samples of the Cnemi- dophorus lemniscatus complex compared here. The letters refer to the karyotype patterns as follows: B and C, unisexual diploids (named as C. cryptus below) from Brazil and Venezuela; D, bisexual diploids (C. lemniscatus) from Brazil, Suriname, and Venezuela; E, bisexual diploids (referred to C. gramivagus below) from Brazil; and F, unisexual triploids (named as C. pseudolemniscatus below) from Suriname. The open circle in southern Guyana represents lizards recently collected on the Rupununi Savanna (see Specimens Examined), which are morphologically typical of other specimens of C. lemniscatus with karyotype D. For Suriname, the southwesternmost F represents two sites-Brokopondo and Brownsweg; the easternmost F represents the type locality (Albina) of C. pseudolemniscatus. For Brazil, the easternmost C represents two sites-Capanema and Maruda. The C in southeastern Venezuela represents the type locality (Icabaru) of C. cryptus. all had identical allele combinations at all loci savanna, rather than closed-canopy forest, examined, confirming the conclusions that and their evolutionary history may be similar heterozygosity is fixed and they have clonal to that of the microteiid Gymnophthalmus inheritance (Sites et al., 1990). Because these underwoodi (see Cole et al., 1989, 1990; Des- lizards occur at localities lacking congeneric sauer and Cole, 1989). During a past period males, they are probably parthenogenetic ofrelatively low rainfall, an ancestral diploid (Vanzolini, 1970). Similar observations per- bisexual species may have dispersed widely tain to the diploid unisexual Venezuelan liz- in a previously existing broad savanna re- ards ofkaryotype C, suggesting that they also gion, at a time when forests were localized in have fixed heterozygosity, clonal inheritance, smaller, isolated refugia in the Guiana Re- and parthenogenetic reproduction. gion (e.g., Haffer, 1979; Hoogmoed, 1979; Huber, 1987). Later, during a period ofhigh- er rainfall, forests spread from their refugia EVOLUTIONARY SCENARIO and became connected and continuous over Lizards in the Cnemidophorus lemniscatus a broader area, as the savanna became frag- complex prefer rather open habitats, such as mented into refugia. The Cnemidophorus re- 1993 COLE AND DESSAUER: UNISEXUAL AND BISEXUAL WHIPTAIL LIZARDS 13 mained with their preferred habitat, becom- have a hybrid origin. We now realize that the ing genetically differentiated into separate unisexual species did not appear to be of hy- bisexual species during allopatry, although brid origin because the unisexual taxa and stabilizing selection basically maintained their their bisexual ancestral taxa are morpholog- ancestral morphology, which was successful ically similar species, although they have high in the ancestral habitat; i.e., the ancestor degrees of genetic differentiation. evolved into two or more poorly differenti- If it had not been for the unisexual popu- ated species that continued to live in the an- lations, we and others would not have been cestral habitat. conducting this comparative genetic research With the return of a drier climate again, at this time, which clearly revealed that mor- the savannas spread from their refugia and phologically similar or even cryptic species became connected, while the forest became have been involved in hybridization. For the fragmented into refugia again. As formerly Gymnophthalmus underwoodi complex we isolated savannas regained contact at their now know there are at least three cryptic spe- edges while expanding, the cryptic species of cies (Cole et al., 1989, 1990, 1993; Hoog- Cnemidophorus hybridized in contact zones. moed et al., 1992), and for the Cnemidoph- The male hybrids were sterile or essentially orus lemniscatus complex there are at least so, but one or a few female hybrids were vi- four species (Dessauer and Cole, 1989; Sites able and cloned themselves parthenogeneti- et al., 1990; Vyas et al., 1990; Wright, 1993; cally, forming the diploid unisexual lineages and see Taxonomy, below). that survive today. At least two separate hy- What does this mean in terms ofother spe- bridization events among the bisexual an- cies of lizards and other organisms in the cestors resulted in the origin of diploid par- Guiana Region? Ifthe Evolutionary Scenario thenogens in the southern (Amazon) versus (see above) is correct, morphologically sim- northern (Venezuelan) parts of the Guiana ilar or cryptic species of bisexual taxa differ- Region. Subsequently, backcross hybridiza- entiated genetically during isolation in refug- tion between a northern diploid parthenogen ia, while their morphological similarities may and northern male(s) of one of the bisexual have been maintained by stabilizing selec- ancestors (karyotype D) produced the north- tion. Hybridization occurred during subse- ern triploid clone (Suriname; fig. 6). In ad- quent habitat expansion from the refugia. The dition, one or a few point mutations caused same process would apply, in an inverse fash- some genetic divergence among the diploid ion, for either savanna or forest-dwelling or- unisexual lineages along the Amazon. This ganisms. If hybridization had not occurred scenario is the inverse of the forest refugium among the similar or cryptic bisexual species hypothesis that has been applied to explain and produced the unisexual clones that we the ecological and evolutionary history offor- have investigated, or if the hybrids had been est birds and other organisms in Neotropica inviable or sterile and selected against (now (e.g., Haffer, 1979; Hoogmoed, 1979; Cap- extinct), then in some cases there would have parella, 1988; Vanzolini and Heyer, 1988; been few clues to lead biologists to recognize Barrowclough, 1992). Other scenarios could the similar bisexual species with confidence. apply also for part or all ofthis history, such The distances across Amazonia (north- as dispersal along ecologically disturbed and south and east-west) and adjacent regions are changing river systems and habitat modifi- vast, particularly for organisms of low vagil- cations caused by mankind. ity. Erosion of montane plateaus such as the Guiana Shield (which has been above sea level throughout the entire history of the ver- IMPLICATIONS FOR BIODIVERSITY tebrates), formation and erosion of signifi- Our attention was drawn to the Guiana cant mountains such as the Andes, huge rivers Region by reports of the unisexual popula- (which change course), alternating wetter and tions of Cnemidophorus lemniscatus and drier periods which affect distribution ofma- Gymnophthalmus underwoodi (see Vanzoli- jor types ofhabitats, changes in sea level, and ni, 1970, 1976; Hoogmoed, 1973), and par- human land uses (such as logging or slash- ticularly by the suggestion that these did not and-burn agriculture) are all processes that 14 AMERICAN MUSEUM NOVITATES NO. 3081

0 co

._E CD E 40 0 0 CD0F) co C.)(U 0 (A .Q *.' 4.. 0. E 0 co 0a Lo>. 1.. 0. 0 a. 0 CD

Ven Brz D F C A-C E

I I I I I I I I I

Fisacnemi00PhOru- pc.ofteer isn5-0p f z V a t s sp

00 Cnemidophorus Fig. 6. Diagram of phylogenetic relationships of the species of the Cnemidophorus lemniscatus complex as discussed here. Abbreviations are as follows: letters above each solid terminus, karyotype des- ignations; Brz and Ven, specimens from Brazil and Venezuela, respectively; sp., undesignated species. fragment animal and plant distributions, Comparative morphology suggests that while causing extinction of populations with many species of animals and plants are highly localized distributions. Fragmentation broadly distributed within the huge area of is the precursor for genetic differentiation of Amazonia. We suggest, however, that com- populations that survive in isolation. parative genetic research involving widely 1993 COLE AND DESSAUER: UNISEXUAL AND BISEXUAL WHIPTAIL LIZARDS 15 separated population samples of these taxa and the triploids from Suriname studied by may reveal many more cryptic species mas- Dessauer and Cole (1989), and the bisexual querading under one name within broadly and unisexual samples from Venezuela re- distributed species complexes, particularly for ported here; i.e., most ofthe same specimens organisms of low vagility. In addition, im- for which the chromosomes and proteins were proved understanding of geographic varia- analyzed, plus others (see Appendix; Speci- tion in gene frequencies within species and mens Examined). within complexes of cryptic species will im- The following characters of scutellation prove understanding of the biogeography of were either basically invariant (except as not- tropical regions. ed) among all of the lizards examined or showed local variation that was not taxo- nomically useful: usually two frontoparietals; TAXONOMY usually four (often 5) supraoculars on each All clonal species of Cnemidophorus (and side (usually 5 on type B and type C diploid other teiids, sensu lato, including gymno- unisexual lizards from Brazil); eight rows of phthalmids) that have been analyzed in detail ventral plates across abdomen; mesoptychi- are products ofinterspecific hybridization (for als abruptly enlarged; postantebrachials gran- recent reviews, see Darevsky et al., 1985; ular; one enlarged anal spur on each side of Dessauer and Cole, 1989; Vrijenhoek et al., males, but small nearly granular scales here 1989; Cole et al., 1993; Wright, 1993). Data in females (very rarely one somewhat en- presented here confirm this also for the uni- larged scale); usually type I preanal scales (but sexual populations referred to Cnemidopho- rather consistently type III in specimens of rus lemniscatus, as discussed above, and the type D bisexual lizards from Christiaan- unisexual populations should not be consid- kondre, Suriname); mean total number ofcir- ered as conspecific with their bisexual ances- cumorbital scales usually 12-14 (but usually tors (e.g., Cole, 1985; Frost and Wright, 1988; fewer than 11 in unisexual lizards from Bra- Sites et al., 1990; Wright, 1993). At least four zil); usually a total of 5-10 interlabial scales species exist in the Cnemidophorus lemnis- (but usually 10-12 in unisexual lizards from catus complex discussed in this paper: (1) the Venezuela and 8-11 in type E lizards from diploid bisexual lizards with karyotype D; (2) Brazil); usually one interlabial scale on each the diploid bisexual lizards with karyotype E; side considerably larger than the other(s) (but (3) the diploid unisexual lizards of hybrid consistently two or three in unisexual dip- origin (karyotypes D x E); and (4) the triploid loids from Venezuela); usually 20-26 gular unisexual lizards of further (backcross) hy- scales; mean number ofgranules around mid- brid origin (karyotypes [D x E] x D). Some body usually 102-112 (but more in males of other herpetologists would split the unisexual type D from Christiaankondre, Suriname, and lizards further (discussed below), but for now only about 98 in unisexual lizards of type B we address the issues of applying specific from Brazil and type C from Venezuela); names to these four taxa. mean total number of femoral pores usually In addition to gathering the karyotypic and 41-48; usually two or three scales midven- genetic data discussed above, we examined trally between femoral pore series on oppo- and compared external morphological char- site legs; mean total number of toe lamellae acters on specimens of the Cnemidophorus usually 57-65 (but more in males of type D lemniscatus complex from Brazil, Venezuela, from Christiaankondre, Suriname); usually Guyana, and Suriname, as well as a series of three or four particularly large, tubercular la- Cnemidophorus gramivagus from Colombia. mellae on each toe number three; total num- Specimens ofthe C. lemniscatus complex in- ber of finger lamellae usually 28-37. cluded representatives ofthe diploid bisexual The number of parietal plates varied as type D and type E individuals from Brazil follows. For bisexual samples, the number of studied by Sites et al. (1990) and Vyas et al. parietals is usually five; if six, usually either (1990), the diploid unisexual type B and type the leftmost or rightmost is divided into an- C individuals from Brazil studied by the same terior and posterior portions; ifseven, usually authors, the diploid bisexual type D lizards both sidemost ones are so divided. Only the 16 AMERICAN MUSEUM NOVITATES NO. 3081

14.

Fig. 7. Preanal region ofCnemidophorus gramivagus (left; AMNH 97425; male, body length 79 mm) and Cnemidophorus lemniscatus from Christiaankondre, Suriname (right; AMNH 133293; male, body length 81 mm), showing differences in morphology of enlarged preanal spurs on adult males of similar size. described conditions for six and seven pari- Two characters of scutellation merit fur- etals (usually seven) were seen on the nine ther discussion: position of the nostril, and diploid unisexual lizards from Venezuela; shape of the anal spur in males. only the conditions for seven or eight pari- POSITION OF NoSTRIL: In C. gramivagus the etals (one endmost one divided into thirds) nostril lies low in the suture between the nasal were seen on the 17 diploid unisexual lizards plates (McCrystal and Dixon, 1987), which from Brazil; and for the 16 triploids from is typical ofmembers ofthe lemniscatus spe- Suriname, conditions with five, six, and sev- cies group (Burt, 1931; Lowe et al., 1970). en parietals were seen. Considering the uni- This condition occurs in all 25 of the C. gra- sexual samples from Suriname, all five lizards mivagus we examined and also in all 10 of from Albina had seven parietals, all five liz- the type E diploid bisexual lizards from Brazil ards from Brownsweg had five parietals, and plus the two juvenile type D diploid bisexual the six specimens from Brokopondo varied lizards from Brazil, the nostril being essen- in having five to seven. tially centered (anterior to posterior) in the Similarly, there is variation in the presence nasal suture. Alternatively, all 12 of the dip- (and number) or absence of tiny, irregular loid bisexual type D C. lemniscatus from scales anterior to the eye and above the su- Venezuela and 11 out of 12 (92%) diploid pralabials, between the supralabials and the bisexual type D C. lemniscatus from Suri- typical array of facial plates (these might be name had the nostril low in the nasal suture called suprasupralabials). For bisexual sam- but clearly more anteriad (or the suture was ples, usually there are none of these scales more posteriad) than in C. gramivagus, the (rarely 1-3 on one or both sides). The 17 exceptional lizard having the nostril as in C. diploid unisexual lizards from Brazil consis- gramivagus. Both conditions occurred in the tently had none of these scales, while six of 34 diploid bisexual lizards from Guyana (59% the nine diploid unisexual lizards from Ven- with the typical C. lemniscatus state). All of ezuela had from one to three scales per side the unisexual lizards (diploid and triploid) and similar variation occurred on the trip- have the nostril centered (anterior to poste- loids from Suriname (although the majority rior) in the nasal suture as in C. gramivagus, had none). including those from Suriname (N = 16), 1993 COLE AND DESSAUER: UNISEXUAL AND BISEXUAL WHIPTAIL LIZARDS 17

Venezuela (N = 9), and Brazil (N = 13 of two bold white or yellowish dorsolateral type B + 4 oftype C). Thus, the specific state stripes separated from two faint white or yel- of this character is useful in diagnosing C. lowish dorsal stripes by longitudinal band of gramivagus and the unisexual species, but it dark ground color, sometimes with faint lat- varies in C. lemniscatus. The sharing of this eral stripes." The "band of dark ground col- character in C. gramivagus and the type E or" refers to a particularly dark longitudinal diploid bisexual lizards from Brazil is similar stripe that is conspicuously darker than the to the distribution of the following character ground color elsewhere on the body. We con- states also. firmed these characters in the Colombian C. SHAPE OF ANAL SPUR IN MALES: All 17 of gramivagus we examined, and the type E dip- the male C. gramivagus we examined have loid bisexual lizards from Brazil shared these anal spurs with an elongate point extending characters also. upward relatively close to the side ofthe body McCrystal and Dixon (1987) pointed out (fig. 7). This condition is shared by the two thatjuveniles ofC. lemniscatus have 7-9 (ac- type E diploid bisexual males we examined tually up to 10; Hoogmoed, 1973) bold lon- from Manacapuru, Amazonas, Brazil gitudinal light stripes (4-6 in C. gramivagus); (UMMZ 183883 and UMMZ 184302). How- a solid particularly dark ground color be- ever, in males of C. lemniscatus of type D tween the stripes (one thin, longitudinal, par- from Suriname, Guyana, and Venezuela, the ticularly dark stripe medially bordering the anal spurs are shorter and they project more dorsolateral stripe in C. gramivagus); arms from the side of the body (fig. 7), excepting and legs with numerous light spots (no spots one Venezuelan male with a more elongate on arms, few on legs of C. gramivagus); and and more projecting spine. Again, the type E juveniles of both species may have lateral diploid bisexual lizards share this character light spots but these are more conspicuous in with C. gramivagus. A somewhat enlarged C. lemniscatus than in C. gramivagus. anal scale with very short spine very rarely McCrystal and Dixon (1987) also pointed out occurs in females of C. lemniscatus and of that adults of C. lemniscatus have five lon- unisexual samples. gitudinal light stripes (2 in C. gramivagus, SNOUT-VENT LENGTH: Males of C. gra- but 3 ofthe 5 in large C. lemniscatus may be mivagus attain a larger body size than do fe- faint), faint light spots on the arms (usually males, and C. gramivagus attains a greater none on C. gramivagus), and bold light spots length than does C. emniscatus (McCrystal on the legs (occasionally a few light spots on and Dixon, 1987). Our data suggest also that C. gramivagus). Our specimens ofdiploid bi- males of C. lemniscatus grow larger than fe- sexual type D lizards from Suriname, Guy- males. Although our data on snout-vent length ana, and Venezuela fit C. lemniscatus in these are not as extensive as those of McCrystal characters, although some have 10 light dor- and Dixon (1987), one point merits attention. sal stripes (if the vertebral stripe is split, as For our samples of diploid bisexual type D described by Hoogmoed, 1973), and all retain C. lemniscatus from Suriname, Guyana, and 2-3 very dark fields on each side ofthe body Venezuela, the largest female is 69 mm in contrasting with the light stripes. Ontogenetic snout-vent length, whereas the largest female differences in the number of stripes (more in C. gramivagus from Colombia is 81 mm. The juveniles) involve the lateral stripes breaking eight females we examined of the diploid bi- up into spots as the lizards age. Also, the light sexual type E lizards from Brazil all ranged spots on the arms become less conspicuous from 69 to 85 mm in snout-vent length, again with age, but the light spots on the side of sharing this character with C. gramivagus. the head and neck become more conspicuous. COLOR PATTERN: McCrystal and Dixon The specimens of triploid and diploid uni- (1987: 246-247) included the following in the sexual samples from Suriname, Venezuela, diagnosis ofC. gramivagus: "little to no spot- and Brazil (types B and C) are also similar to ting on the limbs; adults usually with only the C. lemniscatus described above, except two longitudinal white, yellow or yellowish- the largest Brazilian females have only one green dorsolateral stripes, which may fade particularly dark longitudinal dark field on posteriorly or be totally absent;juveniles with each side (below the paravertebral light line). 18 AMERICAN MUSEUM NOVITATES NO. 3081

who had two plantations at "the confluence of the Cottica River and the Perica Creek, Surinam" (Hoogmoed, 1973: 42-43), Hoog- moed so restricted the type locality. Because of the antiquity of the name C. lemniscatus, its widespread application to populations in Central America and South America, the existence ofboth unisexual and bisexual populations in Suriname, and the diverse opinions concerning application of specific names to unisexual clones of lizards (e.g., Cole, 1985, 1990; Walker, 1986; Frost and Wright, 1988), it is desirable to fix this name to a relevant, widespread bisexual form. As the syntypes include both sexes (with locality unspecified), this can be accomplished by selecting one of the males to be the lectotype and by restricting the type locality to Fig. 8. Preanal region showing spurs and fem- a site known to be inhabited by a relevant oral pores on the lectotype ofCnemidophorus lem- bisexual population (the restriction by Hoog- niscatus, NRM 126, male. Photo by Uno Sam- moed, 1973, was to a nearly coastal locality uelsson, Department ofPaleozoology, arranged and in northern Suriname that may well be ap- provided by Sven 0. Kullander, Department of Vertebrate Zoology, Swedish Museum of Natural propriate but from which, to our knowledge, History, Stockholm. no specimens of Cnemidophorus are known). The two male syntypes (NRM 126) are in equally good condition (Sven 0. Kullander, Conservatively, there are four species rep- personal commun.), one measuring 55 mm resented by these specimens: two diploid bi- in snout-vent length (with a tail of 95 mm) sexual taxa, one diploid unisexual of hybrid and the other measuring 69 mm in snout- origin, and one triploid unisexual of back- vent length (with a tail of 140 mm) (Anders- cross hybrid origin. Names, diagnoses, and son, 1900). We hereby designate the largest additional items are presented and discussed male as the lectotype (figs. 8, 9). Dr. Kullan- in the following species accounts, including der informed us that the number NRM 126 designation of a lectotype of C. lemniscatus will now continue to apply to the lectotype and rerestriction of its type locality. but only to that specimen, and the smaller specimen previously included has been re- Cnemidophorus lemniscatus numbered as NRM 20126. (Linnaeus, 1758) A locality of reasonable access near Hoog- Figures 1,2,7, 8,9, 11 moed's (1973: 42-43) restricted type locality DESIGNATION OF LECTOTYPE AND RERE- and for which a bisexual population of C. STRICTION OF TYPE LoCALITY: Hoogmoed lemniscatus is documented by series of spec- (1973: 41-44) discussed South American liz- imens in major museums (Hoogmoed, 1973: ards named by Linnaeus. Lacerta lemniscata 266; AMNH 133292-133303) is the follow- probably was based on specimens from an ing, which we hereby designate as the re- unspecified locality in Suriname, type ma- stricted type locality: South America: Suri- terial of which now resides at the Naturhis- name: Marowijne; Christiaankondre and toriska Riksmuseet (NRM) in Stockholm, Langamankondre, on the west bank of the Sweden. There are three syntypes at NRM, Marowijne River. This locality is north of two males (both formerly NRM 126, but see Albina near the mouth ofthe Marowijne Riv- below) and one female (NRM 127) (Anders- er. The specimens we examined from Chris- son, 1900; Hoogmoed, 1973; Sven 0. Kul- tiaankondre, Suriname, are consistent in im- lander, personal commun.). As these speci- portant details with the lectotype (compare mens "probably all originate from Dahlberg," figs. 7 right, 8, 9, 1 lA), and this locality is 1993 COLE AND DESSAUER: UNISEXUAL AND BISEXUAL WHIPTAIL LIZARDS 19

Fig. 9. Dorsal view ofthe lectotype of Cnemidophorus lemniscatus. Specimen and acknowledgments as in figure 8, body length 69 mm, tail length 140 mm. only about 80 km (linear) E of the one sug- DiAGNOSIS: A cryptic species of the Cne- gested by Hoogmoed. Short distances can be midophorus lemniscatus complex distin- significant, though, as Albina is less than 30 guished from all others in the genus by the km (linear) S of Christiaankondre, and only following combination ofcharacters: abrupt- unisexual Cnemidophorus are known from ly enlarged mesoptychials; granular postan- Albina (see below). tebrachials; usually 2 frontoparietals; usually 20 AMERICAN MUSEUM NOVITATES NO. 3081

5 parietals; usually 4 supraoculars each side; soptychials; granular postantebrachials; usu- nostril usually somewhat anteriad within but ally 2 frontoparietals; usually 5 parietals; still in suture between nasal plates; bisexual usually 4 supraoculars each side; nostril cen- (gonochoristic; both sexes exist); males with tered (anterior to posterior) within suture be- 1 anal spur (shorter than in C. gramivagus tween nasal plates; bisexual (gonochoristic; and projecting more from side ofbody); adults both sexes exist); males with 1 anal spur (more with 2-5 longitudinal light stripes (7-10 in elongate and extending upward closer to side juveniles), 4-6 particularly dark stripes (6 or of body than in C. lemniscatus); adults with more in juveniles), numerous conspicuous 2 longitudinal light stripes (4-6 in juveniles), light spots on legs (same in juveniles), faint 2 particularly dark stripes (dorsal to the dor- light spots on arms (conspicuous in juve- solateral light stripes; one each side; similar niles), conspicuous light spots laterally on in juveniles), few light spots on legs (similar body (inconspicuous or absent in juveniles); in juveniles), no light spots on arms (similar maximum snout-vent length about 98 mm; injuveniles), conspicuous light spots laterally diploid number of 50 chromosomes, largest on body (inconspicuous or absent in juve- two being homomorphic pair ofsubmetacen- niles); maximum snout-vent length over 1 15 tric to subtelocentric chromosomes. mm; diploid number of 50 chromosomes, DISTRIBUTION: The range includes north- largest two being homomorphic pair of sub- ern Brazil, French Guiana, Suriname, Guy- telocentric to telocentric chromosomes. ana, Venezuela, and Colombia, extending DISTRIBUTION: The range includes north- northward through Central America to Gua- western Brazil, Venezuela, and Colombia. temala, plus the Republic of Trinidad and These lizards prefer open and sunny habitats Tobago and other islands ofthe southern West and "readily climbed trees and fence posts" Indies. These lizards prefer open and sunny (McCrystal and Dixon, 1987: 245; Dixon and habitats. Staton, 1977). To date, no locality is known COMMENTS: The most recent, thorough de- where C. gramivagus is sympatric with C. scriptions and comparisons have been pre- lemniscatus, but such localities may be found sented by Hoogmoed (1973: 262-273, who in future fieldwork, particularly in north- also presented a synonymy for C. lemnisca- western Brazil and central Venezuela. tus) and McCrystal and Dixon (1987). Spec- COMMENTS: The most recent, thorough de- imens of the C. lemniscatus complex occur- scription has been presented by McCrystal ring in bisexual populations and referred to and Dixon (1987). Our comparative data on as type D (or karyotype or cytotype D; e.g., scutellation, size, and color pattern presented Peccinini-Seale and Frota-Pessoa, 1974; above lead to the conclusion that specimens Dessauer and Cole, 1989; Sites et al., 1990; of the C. lemniscatus complex occurring in Vyas et al., 1990) are Cnemidophorus lem- bisexual populations and referred to as type niscatus as the name is applied here. Speci- E (or karyotype or cytotype E; e.g., Peccinini- mens with the alternative karyotypes (A-C Seale and Frota-Pessoa, 1974; Sites et al., and E and F) are referred to other species 1990; Vyas et al., 1990) are Cnemidophorus below. gramivagus. This conclusion can be tested in the future by collecting Venezuelan samples Cnemidophorus gramivagus of C. gramivagus to compare in karyotype McCrystal and Dixon, 1987 and protein electrophoresis with fresh sam- Figure 7 ples of the cytotype E lizards of Brazil. We predict that the Venezuelan lizards will be The holotype, paratypes, and etymology found to share the karyotypes and proteins were all described and discussed by Mc- of the Brazilian lizards. Crystal and Dixon, but the diagnosis is modified below in view ofthe data presented and discussed here. Cnemidophorus cryptus, new species DIAGNOSIS: A species of the Cnemidoph- Figures 1, 10 orus lemniscatus complex distinguished from HOLOTYPE: AMNH 135089 (field number all others in the genus by the following com- JC 5738; karyotype number 1756; frozen tis- bination ofcharacters: abruptly enlarged me- sue number 1930), adult female, collected on 1993 COLE AND DESSAUER: UNISEXUAL AND BISEXUAL WHIPTAIL LIZARDS 21

12 April 1989 by CJC, Carol R. Townsend, Carlos Guaveco, and Maria Jose Praderio in Icabaru, State ofBolivar, Venezuela. This locality is adjacent to the border with Roraima, Brazil, about 75 km (linear) WSW Santa Elena de Uairen, State of Bolivar, Venezuela. This is one ofthe specimens for which the full suite of morphological, karyotypic, and biochemical data are available. PARATYPES: AMNH 134231-134237, and AMNH 135090, eight females collected with the holotype. UMMZ 183871-183874, 183876-183882, 184292, and 184294, a series of 13 females from Oriximina, Para, Bra- zil, collected by D. M. Peccinini-Seale and R. R. de Souza. UMMZ 184291, a female from Capanema, Para, Brazil, collected by D. M. Peccinini-Seale and R. R. de Souza. UMMZ 184298-184300, three females from Maruda, Para, Brazil, collected by D. M. Pec- cinini-Seale and R. R. de Souza. ETYMOLOGY: The Latin adjective cryptus is derived from the Greek kryptos, meaning se- cret or hidden, as these lizards, particularly Fig. 10. Dorsal view of a paratype of Cnemi- in the Venezuelan series, are very similar dophorus cryptus, AMNH 134235, female, body morphologically to C. lemniscatus. The Bra- length 49 mm. zilian populations have been masquerading under the name C. lemniscatus until now, morphic pair with one submetacentric to and some systematists would argue that two subtelocentric chromosome and one subtel- or more species are included here under the ocentric to telocentric chromosome. name C. cryptus (see Alternative Treatments, DEscRIPrION OF THE HOLOTYPE: This large- below). ly follows the format and wording of Hoog- DIAGNOSIS: A cryptic species of the Cne- moed's (1973: 267-269) description of C. midophorus lemniscatus complex distin- lemniscatus, but here is based on C. cryptus, guished from all others in the genus by the AMNH 135089. Rostral pentagonal, visible following combination ofcharacters: abrupt- from above, wider than deep. Nostril cen- ly enlarged mesoptychials; granular postan- tered (anterior to posterior) but low in an tebrachials; usually 2 frontoparietals; usually obliquely divided nasal, the anterior parts 6 or 7 parietals (the rightmost and/or left- forming a short median suture behind the most usually being divided into anterior and rostral. A transversely oval frontonasal. A posterior portions); usually 4 or 5 supraocu- pair of irregularly pentagonal prefrontals, lars each side; nostril centered (anterior to forming a short median suture. Frontal hex- posterior) within suture between nasal plates; agonal, longer than wide, wider anteriorly unisexual (only females exist); adults usually than posteriorly. A pair of irregularly pen- with 5-7 longitudinal light stripes (8-9 in ju- tagonal frontoparietals, forming a long me- veniles), 2-6 particularly dark stripes (4-6 or dian suture. Seven irregularly shaped pari- more in juveniles), numerous conspicuous etals in a transverse series, three large, medial, light spots on legs (same in juveniles), faint two on each side. At least five irregular, small light spots on arms (conspicuous in juve- occipitals. Back of head covered with small niles), conspicuous light spots laterally on granules. Supraoculars 4; fourth smallest; body (inconspicuous or absent in juveniles); second and third largest; separated from the maximum snout-vent length about 70 mm; supraciliaries by generally one row of gran- diploid number of 48-50 chromosomes (de- ules, except first supraocular broadly contacts pending on clone), largest two being hetero- first supraciliary. Last supraocular separated 22 AMERICAN MUSEUM NOVITATES NO. 3081 from frontoparietals by a row of small scales Femoral pores 23 (left)/24 (right); at least (circumorbital semicircles, 8 scales on left, 7 two in preanal position; each pore usually on right). Loreal large, more or less trapezoid. surrounded by three small scales (medial one Preocular absent. A row of four suboculars, largest in ring). Midventrally, three scales forming a suborbital ridge; the anterior sub- separate the femoral pore series ofeach side. ocular more or less in a preocular position, Scales on tail large, rectangular, obliquely deeper than long; the third longest. An irreg- keeled, slightly mucronate; imbricate, in ular row offour enlarged postoculars. Supra- transverse rows; keels forming longitudinal ciliaries 6; the second one largest, the anterior ridges. Scales under tail similar but narrower. two elongate, the posterior ones quadrangu- Scales on regenerated part of tail small, ir- lar. A row of enlarged supratemporals. En- regular, keeled, in transverse rows. Tail round larged scales in front of ear opening; central in cross section. region of temple with small, round to oval Scales on upper and anterior surfaces of granules. Ear opening large, surrounded by upper arm, on anterior and lower surfaces of small scales forming smooth margin; external forearm, on anterior and lower surfaces of auditory meatus short, tympanum clearly thighs, and on lower and posterior surfaces visible. Seven large supralabials followed by oflower legs large, smooth, more or less hex- small ones; five supralabials to below the cen- agonal, imbricate; on lower and posterior sur- ter of the eye; the third largest, the sixth faces of upper arm, on posterior and upper smallest. Two tiny irregular scales anterior to surfaces of forearm, on posterior and upper eye, above supralabials, between supralabials surfaces of thighs and on upper and anterior and usual array of facial plates, each side. surfaces of lower legs small, granular, juxta- Lower eyelid with a semitransparent disc of posed (postantebrachials granular). Fourth five enlarged palpebrals. Pupil shape unclear. finger lamellae number 17/17. Fourth toe la- Mental trapezoid with convex anterior mellae number 30/31, usually single but di- margin. A pentagonal postmental. Six pairs vided under articulations. Approximately ofchinshields curving posteriad and dorsally four triangular, strongly enlarged and tuber- to the infralabials; only the members of the cular scales under third toe, proximally. Fin- anterior pair (largest) in contact at midline; gers and toes compressed. Palms and soles from the second pair on separated from the with small, irregular, juxtaposed, flat scales. labials by the interlabial scales (7 on each Upper surfaces of hands and feet with large, side). Seven enlarged anterior infralabials, imbricate flat scales. followed by small scales; third infralabial In preservative, back and flanks with nine largest. light longitudinal stripes (including a single Gulars small, flat, rounded, juxtaposed, vertebral stripe) separated by dark stripes, of slightly larger in the anterior part ofthe throat, which six are black (lowest gray). The dor- smaller in the posterior part, which starts at solateral light stripe starts at posterior corner level of ears. Scales (mesoptychials) in front of eye and continues to tail. The vertebral ofprepectoral fold abruptly enlarged, slightly stripe arises at the back of the head and con- imbricate, smooth. Scales on nape and side tinues to the base of the tail. The stripe be- of neck similar to dorsal and lateral body tween the vertebral and dorsolateral stripes scales. arises at the posterior edge ofthe most lateral Dorsal and lateral scales granular, in in- parietals and continues to the proximal part distinct transverse and oblique rows. Ven- of the tail, merging with the other from the trals large, rectangular, wider than long, im- opposite side. The upper lateral stripe starts bricate, smooth, in eight longitudinal and in on the lower eyelid, passes over the ear open- 30 transverse rows (shoulder to hip); the an- ing and continues to the groin; the lower lat- terior rows in midventral area interrupted by eral stripe passes from the posterior edge of a small triangular area ofsmaller scales. Total the ear opening over the arm to the groin. number of dorsal granules around midbody Side ofhead, neck, and body with white spots. 101. Preanal area with four clearly enlarged Upper surfaces ofhind limbs and arms brown smooth, irregularly shaped, juxtaposed scales or black, with scattered white spots. Back of plus smaller ones (preanal scutes are type I). thighs with a horizontal series of longitudi- 1993 COLE AND DESSAUER: UNISEXUAL AND BISEXUAL WHIPTAIL LIZARDS 23 nally elongate white spots, continuing onto send in Albina, Marowijne District, Surina- base of tail as a white stripe. Ventral parts me. This is one of the specimens for which white. Tail above brown with tan to white the full suite of morphological, karyotypic, dorsolateral stripes on base; lower surface of and biochemical data are available. tail white. Snout-vent length, 60 mm; tail PARATYPES: AMNH 133305-133307 and regenerated. Hind limb approximately twice AMNH 135088, four females collected with as long as arm. When laid along the body the holotype. AMNH 119407-119411 and there is considerable overlap ofarm and hind AMNH 133313, six females from Brokopon- limb. do, Brokopondo District, Suriname. AMNH DISTRIBUTION: The range is disjunct, in- 133308-133312, five females from Brown- cluding Amazonian Brazil and the type lo- sweg, Brokopondo District, Suriname. cality in southeastern Venezuela. These liz- ETYMOLOGY: The name is in reference to ards prefer open and sunny habitats, and to the morphological similarity to C. lemnis- date are not known to be sympatric at any catus, under which name these lizards have locality with other members of the C. lem- been masquerading until now. niscatus complex. DIAGNOSIS: A cryptic species of the Cne- COMMENTS: We apply the name C. cryptus midophorus lemniscatus complex distin- to the diploid clones of hybrid origin involv- guished from all others in the genus by the ing karyotype D x karyotype E lizards ofthe following combination ofcharacters: abrupt- C. lemniscatus complex (Dessauer and Cole, ly enlarged mesoptychials; granular postan- 1989; Sites et al., 1990; Vyas et al., 1990). tebrachials; usually 2 frontoparietals; 5-7 pa- Data presented above suggest that this origin rietals (usually depending on endmost being was of hybridization between C. lemniscatus divided or not into anterior and posterior x C. gramivagus (fig. 6), and Vyas et al. (1990) portions); usually 4 supraoculars each side; presented evidence from mitochondrial DNA nostril centered (anterior to posterior) within suggesting that C. lemniscatus was the ma- suture between nasal plates; unisexual (only ternal parent for at least one such cross. For females exist); adults usually with 5-7 lon- now, we apply the name C. cryptus to lizards gitudinal light stripes (9 in juveniles), 4 par- ofkaryotype C from Venezuela, reported here ticularly dark stripes (6 in juveniles), numer- for the first time, and to lizards ofkaryotypes ous conspicuous light spots on legs (same in A, B, and C from Brazil (Peccinini-Seale and juveniles), faint light spots if any on arms Frota-Pessoa, 1974; although we have not yet (conspicuous in juveniles), conspicuous small personally examined a lizard oftype A). These light spots laterally on body (inconspicuous karyotypic clones can be informally referred or absent injuveniles); maximum snout-vent to as C. cryptus type A, C. cryptus type B, length about 70 mm; triploid number of 75 and C. cryptus type C. Other clones are rec- chromosomes, largest three being hetero- ognizable also, as in variants having different morphic with two submetacentric to subtel- combinations of alleles detected by protein ocentric chromosomes and one subtelocen- electrophoresis (Sites et al., 1990). Our data tric to telocentric chromosome. suggest that the Amazonian lizards and the DESCRIPTIoN OF THE HOLOTYPE: Rather Venezuelan lizards here referred to C. cryptus than repeat the long list oftraits and character had at least two separate hybrid origins in- states that are identical to those described volving the same two bisexual ancestors but above for the holotype ofC. cryptus, we men- at two different localities. tion here only those conditions that differ from that specimen. A large, irregularly hex- Cnemidophorus pseudolemniscatus, agonal frontonasal; supraoculars 4/5, sepa- new species rated from supraciliaries by generally two Figures 2, 11 rows of granules; circumorbital semicircles 5/7; postoculars 5/4; two tiny irregular scales HOLOTYPE: AMNH 133304 (field number anterior to eye, above supralabials, between JC 5227; karyotype number 1731; frozen tis- supralabials and usual array of facial plates, sue number 1355), adult female, collected on on right side only; pupil kidney shaped, in- 10 March 1986 by CJC and Carol R. Town- dentation on ventral side; four interlabial 24 AMERICAN MUSEUM NOVITATES NO. 3081

...--,----.... , .. -._ .7

C D

E F

Fig. 11. Two species of the Cnemidophorus lemniscatus complex, C. lemniscatus (bisexual diploid; A-D) and C. pseudolemniscatus (unisexual triploid; E-F). A, AMNH 133295, male from Christiaan- kondre, Suriname, body length 86 mm. B, AMNH 138079, male from Yupukari, Guyana, body length 71 mm. C, AMNH 133302, female from Christiaankondre, Suriname, body length 63 mm. D, AMNH 133301, female from Christiaankondre, Suriname, body length 69 mm. E, AMNH 133307, female from Albina, Suriname, body length 65 mm. F, AMNH 133309, female from Brownsweg, Suriname, body length 49 mm. scales on each side; granules around midbody this is the largest triploid female we have seen number 105. Preanal area with 5 enlarged from Suriname, such spots are present on all scales; femoral pores 22/23, two midventral others, and tan spots were present in life [see scales separating the series ofeach side; fourth color notes below]). Lower surface oftail white finger lamellae number 18/19; fourth toe la- at base, bluish distally. Snout-vent length, 70 mellae number 32/31. In preservative, back mm; tail regenerated. with five light stripes and four black ones; COLORATION OF THE HOLOTYPE IN LIFE: The lateral light stripes broken up into white spots; following notes are paraphrased from the white spots not presently visible on arms (but fieldnotes of CJC dated 8 April 1986, per- 1993 COLE AND DESSAUER: UNISEXUAL AND BISEXUAL WHIPTAIL LIZARDS 25 taining to specimen number JC 5227 from the morphology of their karyotypes, even Albina, Suriname. Upper body dark brown though all three of these might ultimately to black with tan stripes; trace ofbrown wash share a common ancestor of one F1 hybrid on neck; dorsolateral light stripe yellowish between C. emniscatus x C. gramivagus (type tan; sides brown with yellowish green spots D x type E; Vyas et al., 1990). Each clone and wash of greenish yellow; head tan with could be named as a separate species, choos- light green wash and light spots on sides; arms ing as a significant historic event an aspect brown with tan spots and yellowish green ofthe occurrence and establishment of those wash; hands pale green; top of legs brown chromosome mutations (derived character with numerous large beige spots; anteriorly, states) that resulted in each diagnosable clone. legs bright green; hands and feet brown, feet However, mutations can occur anywhere in with green on anterior edge; tail tan above, the genome, in mitochondrial DNA or in nu- bright chartreuse green on sides, bright yel- clear DNA, affecting the shape of chromo- lowish green below (but light brown posterior somes, allelic variants detected by protein to distally located point ofregeneration); chin, electrophoresis, and other characters such as throat, and underside of arms pale yellowish scutellation and color pattern. Such muta- tan; chest greenish tan; abdomen tan but pos- tions usually produce variants at the level of teriorly becoming bright yellowish green, as organization similar to normal individual on undersides of legs and tail. variation, particularly as seen also in bisexual DISTRIBUTION: These triploid unisexual species with Mendelian patterns of inheri- lizards are known only from Suriname and tance, and we do not consider these as the French Guiana (Serena, 1985; Dessauer and kind ofsignificant historic events upon which Cole, 1989), and usually they are allopatric separate species should be conceived and to populations ofthe diploid bisexual C. lem- named if the practice of taxonomy is to be niscatus (Hoogmoed, 1973; Serena, 1985). useful. We emphasize instead the shared an- COMMENTS: We apply the name C. pseu- cestry of these clones by including them as dolemniscatus to the triploid clone of hybrid variants in a species of hybrid origin (in- origin involving a karyotype C diploid uni- volving the same bisexual ancestral species sexual lizard x karyotype D male of the C. but not necessarily the same ovum). lemniscatus complex (Dessauer and Cole, In the case ofthe diploid unisexual C. cryp- 1989; fig. 2B here, designated as karyotype tus type C, we have presented evidence F). Present data and nomenclature suggest (above) for separate hybrid origins within the that this origin was ofhybridization between same ancestry (C. lemniscatus x C. grami- C. cryptus x C. lemniscatus (fig. 6), as names vagus) for the populations in Brazil versus are applied here. that in Venezuela. We emphasize their similar hybrid ancestry by including them in one species; we do not treat the fact that they ALTERNATIVE TREATMENTS probably stemmed from different F1 hybrid Some herpetologists would prefer taxo- individuals as a rationale for providing dif- nomic treatments different from ours, par- ferent names. ticularly for parthenogenetic lineages. Alter- The evidence that C. cryptus type C ofVen- natives have been discussed or mentioned by ezuela stemmed from a different F1 hybrid Maslin (1968), Morafka (1977), Cole (1985, egg than did the C. cryptus type C of Brazil 1990), Walker (1986), Frost and Wright is in the electrophoretic data (based on geo- (1988), Wright (1993), and others. A review graphic variation in both the unisexual and of all ofthis literature is beyond the scope of bisexual species). In particular, consider the the present paper, but a few points should be locus PEPA (table 2; fig. 4). Cnemidophorus noted with respect to the Cnemidophorus cryptus type C from Venezuela has the b-al- lemniscatus complex. lele ofC. lemniscatus, which characterizes its How many cryptic species are included populations from Venezuela, northern Brazil among the diploid unisexual clones here rel- (Boa Vista, Roraima), and Suriname. How- egated to Cnemidophorus cryptus? Types A, ever, C. cryptus type C from Brazil has the B, and C can be recognized and diagnosed by c-allele of C. lemniscatus, which character- 26 AMERICAN MUSEUM NOVITATES NO. 3081 izes its populations from the Amazon (Alter could also be involved. For another example, do Chao). Allelic variants at sMEP also sup- if the parthenogenetic ancestor of the type C port this conclusion, although alternative ex- unisexual lizards from Venezuela originally planations are possible (e.g., what appear had both the a- and b-alleles at sMEP, gene electrophoretically to be the product of one conversion from the b-allele of sMEP to the allele may actually be two different proteins a-allele (see Hillis et al., 1991) or a mutation with the same mobility, one derived by point in the ancestral b-allele to produce either a mutation within a parthenogenetic lineage). null allele or an allele the product of which The diploid unisexual lizards oftype C from has the same mobility as the a-allele, could Venezuela usually have 4 supraoculars on explain our results (table 2). Either of these each side, those from Brazil usually 5. The events could have occurred in a partheno- number of gulars in Venezuela is 19-21, in genetic common ancestor ofall ofthe type C Brazil 23-25. The total number of circum- lizards, although the hypothesis that the dif- orbital scales in Venezuela is 12-16, in Brazil ferent type C clones in Brazil and Venezuela 8-12. The number of equally large scales had separate hybrid origins is the most par- among the largest interlabials in Venezuela simonious for existing data, particularly con- is 2-3, in Brazil 1. Adults of the Venezuelan sidering PEPA. lizards have (on each side) two or three par- The data suggesting that the diploid uni- ticularly dark stripes among the dark fields, sexual lizards from Venezuela and Brazil share whereas adult Brazilian lizards appear to have the same basic hybrid origin (i.e., the same only one (below the paravertebral light line; ancestral species, C. lemniscatus x C. gra- most Brazilian specimens we examined, mivagus) are particularly robust. We empha- however, were darkened in preservative). size that similar heritage by considering these Although our sample sizes are small (N = parthenogenetic lizards as one species, C. 9 type C from Venezuela, N = 0 type A from cryptus. We are interested in supporting and Brazil, N = 13 type B from Brazil, and N = continuing research on additional evolution- 4 type C from Brazil), and most of the Bra- ary questions and matters concerning popu- zilian lizards available to us are preservative lation genetics, but we do not consider the darkened, it does appear as if the two geo- question of how many separate F1 zygotes graphically separated hybrid events captured produced separate clones as helpful in pro- and cloned different parts of the spectrum of ducing a useful and stable taxonomy, even morphological variation that occurs in the though this is an extremely interesting area bisexual ancestors, and the resulting unisex- of investigation for geneticists, evolutionary ual forms probably could be diagnosed mor- biologists, and biogeographers. phologically. In other instances of separate hybrid origins, particularly those involving separate zygotes in the same clutch of eggs SUMMARY AND CONCLUSIONS produced by the same individual parents, 1. Until recently, at least four species have electrophoretic characters might be the only been masquerading under the name Cnemi- ones to use to diagnose clones ifany evidence dophorus lemniscatus in the Guiana Region is available at all. of northern South America. It also seems appropriate, particularly in 2. Two of these are the following diploid the context of utility and stability of no- bisexual species: C. lemniscatus (Linnaeus, menclature, to question how robust the con- 1758) and C. gramivagus McCrystal and Dix- clusions of separate hybrid origins are for on, 1987. parthenogenetic lineages. For example, mor- 3. The other two are unisexual species that phological characters such as those just dis- are diagnosed and named here as follows: C. cussed could diverge by mutation from a par- cryptus, new species, a diploid species con- thenogenetic common ancestor, as may best sisting of several clones in Brazil and Vene- explain karyotypes A, B, and C in Brazil (Vyas zuela; and C. pseudolemniscatus, new spe- et al., 1990; and see above). A mutation at cies, a triploid species occurring in Suriname PEPA (table 2; fig. 4) and limitations in in- and French Guiana. terpreting electrophoretic results (see above) 4. Karyotypic data and electrophoretic 1993 COLE AND DESSAUER: UNISEXUAL AND BISEXUAL WHIPTAIL LIZARDS 27 data representing 32 gene loci ofthese lizards two genomes from C. lemniscatus and one have been compared in samples from Brazil, from C. gramivagus. Suriname, and Venezuela, and morphologi- 9. The clone of C. cryptus involved in the cal comparisons include these specimens plus hybrid origin of C. pseudolemniscatus was new samples from Guyana. more similar to that ofthe type locality (Ven- 5. Cnemidophorus cryptus is a diploid par- ezuela) than to those occurring to the south, thenogenetic species that originated from hy- along the Amazon. bridization between C. lemniscatus x C. gra- 10. The male of C. lemniscatus involved mivagus. in the origin ofC. pseudolemniscatus was also 6. Several clones of C. cryptus exist, each from a more northern population in South demonstrating fixed heterozygosity in a series America, rather than one along the Amazon. ofkaryotypic and electrophoretic characters. 1 1. The overall morphological similarity 7. Alleles at two loci (sMEP and PEPA) of the lizards in the Cnemidophorus lemnis- suggest that the clone ofC. cryptus from Ven- catus complex has obscured their taxonomy ezuela had a hybrid origin separate from the and relationships for a long time, and many clones in Brazil, although the same species problems remain to be resolved. For these of bisexual ancestors (different populations) and other organisms, karyotypic and partic- were involved. ularly biochemical investigations show con- 8. Cnemidophorus pseudolemniscatus is a siderable potential for revealing relationships triploid parthenogenetic species that origi- in further detail and improving our under- nated from hybridization between C. cryptus standing of the biodiversity and biogeogra- x C. lemniscatus. In terms of the diploid phy of Neotropica. bisexual ancestors, C. pseudolemniscatus has

REFERENCES Andersson, L. G. polyploidy in reptiles. J. Hered. 70: 95- 1900. Catalogue ofLinnean type-specimens of 102. Linnaeus's Reptilia in the Royal Mu- 1985. Taxonomy of parthenogenetic species seum in Stockholm. Bihang Till K. of hybrid origin. Syst. Zool. 34: 359- Svenska Vet.-Akad. Handl. 26: 1-29. 363. Barrowclough, G. F. 1990. When is an individual not a species? 1992. Systematics, biodiversity, and conser- Herpetologica 46: 104-108. vation biology. In N. Eldredge (ed.), Cole, C. J., H. C. Dessauer, and G. F. Barrow- Systematics, ecology, and the biodiver- clough sity crisis, pp. 121-143. New York: Co- 1988. Hybrid origin of a unisexual species of lumbia Univ. Press. whiptail lizard, Cnemidophorus neo- Burt, C. E. mexicanus, in western North America: 1931. A study of the teiid lizards ofthe genus new evidence and a review. Am. Mus. Cnemidophorus with special reference Novitates 2905: 38 pp. to their phylogenetic relationships. U.S. Cole, C. J., C. R. Townsend, H. C. Dessauer, and Natl. Mus. Bull. 154:1-286. L. M. Hardy Capparella, A. P. 1989. A lizard foretold. Nat. Hist., May,1989: 1988. Genetic variation in neotropical birds: 12-17. Implications for the speciation process. Cole, C. J., H. C. Dessauer, C. R. Townsend, and In H. Ouellet (ed.), Acta XIX Congres- M. G. Arnold sus Internationalis Ornithologici 2: 1990. Unisexual lizards of the genus Gym- 1658-1673. Natl. Mus. Nat. Sci., Ot- nophthalmus (Reptilia: Teiidae) in the tawa: Univ. Ottawa Press. Neotropics: Genetics, origin, and sys- Cole, C. J. tematics. Am. Mus. Novitates 2994: 29 1979. Chromosome inheritance in partheno- pp- genetic lizards and evolution of allo- Cole, C. J., H. C. Dessauer, and A. L. Markezich 28 AMERICAN MUSEUM NOVITATES NO. 3081

1993. Missing link found: The second ances- IV. The lizards and amphisbaenians of tor of Gymnophthalmus underwoodi Surinam. The Hague: W. Junk, ix + 419 (Squamata: Teiidae), a South American pp- unisexual lizard of hybrid origin. Am. 1979. The herpetofauna of the Guianan Re- Mus. Novitates 3055: 13 pp. gion. In W. E. Duellman (ed.), The South Darevsky, I. S., L. A. Kupriyanova, and T. Uzzell American herpetofauna: its origin, evo- 1985. Parthenogenesis in reptiles. In C. Gans lution, and dispersal, pp. 241-279. and F. Billett (eds.), Biology ofthe Rep- Monogr. No. 7, Mus. Nat. Hist., Univ. tilia 15: 411-526. New York: Wiley. Kansas, 485 pp. Dessauer, H. C., and C. J. Cole Hoogmoed, M. S., C. J. Cole, and J. Ayarzaguena 1989. Diversity between and within nominal 1992. A new cryptic species of lizard (Sauria: forms of unisexual teiid lizards. In R. Teiidae: Gymnophthalmus) from Ven- M. Dawley and J. P. Bogart (eds.), Evo- ezuela. Zool. Meded. 66(1): 1-18. lution and ecology of unisexual verte- Huber, 0. brates, pp. 49-71. Bull. 466, New York 1987. Neotropical savannas: their flora and State Mus., Albany, v + 302 pp. vegetation. Trends Ecol. Evol. 2(3): 67- Dessauer, H. C., C. J. Cole, and M. S. Hafner 71. 1990. Collection and storage of tissues. In D. Linnaeus, C. M. Hillis and C. Moritz (eds.), Molec- 1758. Systema naturae per regna tria naturae, ular systematics, pp. 25-41. Sunder- secundum classes, ordines, genera, spe- land, MA: Sinauer. cies, cum characteribus, differentiis, Dixon, J. R., and M. A. Staton synonymus, locis. Ed. 10, vol. 1: 1-823. 1977. Arboreality in the teiid lizard Cnemi- Laurentius Salvius, Stockholm. dophorus lemniscatus (Reptilia, Lacer- Lowe, C. H., J. W. Wright, C. J. Cole, and R. L. tilia, Teiidae) in the Venezuelan Llanos. Bezy J. Herpetol. 11: 108-111. 1970. Chromosomes and evolution ofthe spe- Donnelly, M. A., R. W. McDiarmid, and C. W. cies groups ofCnemidophorus (Reptilia: Myers Teiidae). Syst. Zool. 19: 128-141. 1992. A new lizard of the genus Arthrosaura Maslin, T. P. (Teiidae) from southern Venezuela. 1968. Taxonomic problems in parthenoge- Proc. Biol. Soc. Washington. 105: 821- netic vertebrates. Syst. Zool. 17: 219- 833. 231. Frost, D. R., and J. W. Wright McCrystal, H. K., and J. R. Dixon 1988. The taxonomy of uniparental species, 1987. A new species of Cnemidophorus (Sau- with special reference to parthenoge- ria: Teiidae) from the Llanos ofColom- netic Cnemidophorus (Squamata: bia and Venezuela. J. Herpetol. 21: 245- Teiidae). Syst. Zool. 37: 200-209. 254. Giblett, E. R., C. G. Hickman, and 0. Smithies Morafka, D. J. 1959. Serum transferrins. Nature 183(4659): 1977. A biogeographical analysis of the Chi- 1589-1590. huahuan Desert through its herpeto- Haffer, J. fauna. Biogeographica 9: viii + 313 pp. 1979. Quaternary biogeography of tropical The Hague: Dr. W. Junk. lowland South America. In W. E. Duell- Murphy, R. W., J. W. Sites, Jr., D. G. Buth, and man (ed.), The South American her- C. H. Haufler petofauna: its origin, evolution, and dis- 1990. Proteins I: Isozyme electrophoresis. In persal, pp. 107-140. Monogr. No. 7, D. M. Hillis and C. Moritz (eds.), Mo- Mus. Nat. Hist., Univ. Kansas, 485 pp. lecular systematics, pp. 45-126. Sun- Harris, H., and D. A. Hopkinson derland, MA: Sinauer. 1976. Handbook ofenzyme electrophoresis in Peccinini-Seale, D., and 0. Frota-Pessoa human genetics. Amsterdam: North- 1974. Structural heterozygosity in partheno- Holland. genetic populations of Cnemidophorus Hillis, D. M., C. Moritz, C. A. Porter, and R. J. lemniscatus (Sauria, Teiidae) from the Baker Amazonas Valley. Chromosoma 47: 1991. Evidence for biased gene conversion in 439-451. concerted evolution ofribosomal DNA. Serena, M. Science 251: 308-310. 1985. Zoogeography ofparthenogenetic whip- Hoogmoed, M. S. tail lizards (Cnemidophorus lemnisca- 1973. Notes on the herpetofauna of Surinam. tus) in the Guianas: evidence from skin 1993 COLE AND DESSAUER: UNISEXUAL AND BISEXUAL WHIPTAIL LIZARDS 29

grafts, karyotypes, and erythrocyte ar- Vrijenhoek, R. C., R. M. Dawley, C. J. Cole, and eas. J. Biogeogr. 12: 49-56. J. P. Bogart Shaklee, J. B., F. W. Allendorf, D. C. Morizot, and 1989. A list of the known unisexual verte- G. S. Whitt brates. In R. M. Dawley and J. P. Bogart 1989. Genetic nomenclature of protein-cod- (eds.), Evolution and ecology ofunisex- ing loci in fish. Trans. Am. Fish. Soc. ual vertebrates, pp. 19-23. Bull. 466, 118: 218-227. New York State Mus., Albany, v + 302 Sites, J. W. Jr., D. M. Peccinini-Seale, C. Moritz, PP. J. W. Wright, and W. M. Brown Vyas, D. K., C. Moritz, D. Peccinini-Seale, J. W. 1990. The evolutionary history of partheno- Wright, and W. M. Brown genetic Cnemidophorus lemniscatus 1990. The evolutionary history of partheno- (Sauria, Teiidae). I. Evidence for a hy- genetic Cnemidophorus lemniscatus brid origin. Evolution 44: 906-921. (Sauria, Teiidae). II. Maternal origin and Smithies, 0. age inferred from mitochondrial DNA 1959. Zone electrophoresis in starch gels and analyses. Evolution 44: 922-932. its application to studies of serum pro- Walker, J. M. teins. Adv. Protein Chem. 14: 65-113. 1986. The taxonomy of parthenogenetic spe- Vanzolini, P. E. cies of hybrid origin: Cloned hybrid 1970. Unisexual Cnemidophorus lemniscatus populations of Cnemidophorus (Sauria: in the Amazonas Valley: a preliminary Teiidae). Syst. Zool. 35: 427-440. note (Sauria, Teiidae). Pap. Avulsos Wright, J. W. Zool. (Sao Paulo) 23: 63-68. 1993. Evolution of the lizards of the genus 1976. On the presence of males in Gymno- Cnemidophorus. In J. W. Wright and L. phthalmus underwoodi, a presumed all- J. Vitt (eds.), Biology ofwhiptail lizards female lizard species (Sauria, Teiidae). (genus Cnemidophorus), pp. 27-81. Ibid., 29: 177-179. Oklahoma Mus. Nat. Hist., Norman, xiv Vanzolini, P. E., and W. R. Heyer (eds.) + 417 pp. 1988. Proceedings ofa workshop on neotropical distribution patterns. Acad. Brasi- leira Cienc., Rio de Janeiro, 488 pp. APPENDIX Specimens Examined The 170 specimens are referred to by their in- Capanema (UMMZ 184291, e, p); Maruda dividual catalog numbers and initials for their re- (UMMZ 184298-184300, e, p); and homogenates spective collections, as follows: AMNH (Ameri- of karyotype C lizards from either Capanema or can Museum of Natural History); CM (Craig Maruda but unspecified when provided to us (CM Moritz-for some of the Brazilian homogenates 238-241 and CM 243-246, p). used and reported also by Sites et al. [1990]); and UMMZ (University of Michigan Museum of Zo- Cnemidophorus gramivagus ology, with thanks to Arnold Kluge; these specimens also were used andI reported by Sites et al. BRAZIL: Amazonas; Manacapuru (UMMZ [1990]). The lowercase letters following the catalog 183883-183887, e, p; UMMZ 184293, e, p; numbers indicate the kind ofdata taken from each UMMZ 184295 e, p; UMMZ 184301-184303, e, specimen, as follows: e, external morphology; k, p; and CM 221-225, p). COLOMBIA: Depart- karyotype; p, protein electrophoresis. ment of Arauca; Cravo Norte (AMNH 97410- 97434, e). Cnemidophorus cryptus Cnemidophorus lemniscatus Type locality, VENEZUELA: State of Bolivar; Icabaru (holotype, AMNH 135089, e, k, p; AMNH BRAZIL: State ofPara; Alter do Chao (UMMZ 134231-134232, e, p; AMNH 134233, e; AMNH 184296, e, p; UMMZ 184297, e; CM 234-237, p; 134234, e, p; AMNH 134235-134237, e; AMNH CM 242, p; CM 248-249, p; CM 269, p; and CM 135090, e, k). BRAZIL: State of Para; Oriximina 277, p). State of Roraima; Boa Vista (CM 226- (CM 229-233, p; UMMZ 183871-183874, e, p; 228, p; CM 247, p). GUYANA: northern Rupu- UMMZ 183876-18388 1, e, p; UMMZ 183882, e; nuni Savanna; Jouri, 33 km (linear) NW Karan- UMMZ 184292, e, p; and UMMZ 184294, e, p); ambo (AMNH 138109-138110, e); Karasabai Vil- 30 AMERICAN MUSEUM NOVITATES NO. 3081 lage, 42 km (linear) NW Karanambo (AMNH Cnemidophorus pseudolemniscatus 138099-138108, e); Yupukari (on the Rupununi River), 12 km (linear) SSW Karanambo (AMNH Type locality, SURINAME: Marowijne Dis- 138058-138062, e; AMNH 138065-138079, e; trict; Albina (holotype, AMNH 133304, e, k, p; and AMNH 138081-138082, e). SURINAME: AMNH 133305, e, k, p; AMNH 133306-133307, Marowijne District; Christiaankondre, near the e, p; and AMNH 135088, e). Brokopondo District; mouth of the Marowijne River (AMNH 133292- Brokopondo (AMNH 119407-119411, e; AMNH 133293, e, k, p; AMNH 133294-133297, e, p; 133313, e, k, p); Brownsweg (AMNH 133308- AMNH 133298-133299, e; and AMNH 133300- 133309, e, k, p; AMNH 133310 e, p; and AMNH 133303, e, k, p). VENEZUELA: State of Bolivar; 133311-133312, e). San Ignacio de Yuruani (AMNH 134221, e; AMNH 134222-134228, e, p; AMNH 134229- 134230, e; AMNH 135091, e, k, p; and AMNH 135092, e, p).

Recent issues of the Novitates may be purchased from the Museum. Lists of back issues of the Novitates, Bulletin, and Anthropological Papers published during the last five years are available free of charge. Address orders to: American Museum of Natural History Library, Department D, Central Park West at 79th St., New York, New York 10024.

This paper meets the requirements of ANSI/NISO Z39.48-1992 (Permanence of Paper).