Bulletin of the Museum of Comparative Zoology Volume 160, Number 1 7 October 2010 The genus Siro Latreille, 1796 (Opiliones, Cyphophthalmi, sironidae) in North America with a phylogenetic analysis based on molecular data and the description of four new species GONZALO GIRIBET AND WILLIAM A. SHEAR HARVARD UNIVERSITY | CAMBRIDGE, MASSACHUSETTS, U.S.A. BULLETIN OF THE Museum of Comparative Zoology BOARD OF EDITORS Editor: Jonathan Losos Managing Editor: Adam Baldinger Editorial Assistant: Samantha Edelheit Associate Editors: Andrew Biewener, Scott Edwards, Brian Farrell, Farish Jenkins, George Lauder, Gonzalo Giribet, Hopi Hoekstra, Jim Hanken, Jim McCarthy, Naomi Pierce, and Robert Woollacott Publications Issued or Distributed by the Museum of Comparative Zoology Harvard University Bulletin 1863- Breviora 1952- Memoirs 1865-1938 Johnsonia, Department of Mollusks, 1941-1974 Occasional Papers on Mollusks, 1945- General queries, questions about author guidelines, or permissions for MCZ Publications should be directed to the editorial assistant: MCZ Publications Museum of Comparative Zoology Harvard University 26 Oxford Street Cambridge, MA 02138 meczpublications@mcz.harvard.edu EXCHANGES AND REPRINTS All of our publications are offered for free on our website: http://www.mcz.harvard.edu/Publications/index.html To purchase individual reprints or to join our exchange program, please contact Susan DeSanctis at the Ernst Mayr Library: mayrlib@oeb.harvard.edu. This publication has been printed on acid-free permanent paper stock. © The President and Fellows of Harvard College 2010. THE GENUS S/RO LATREILLE, 1796 (OPILIONES, CYPHOPHTHALMI, SIRONIDAE), INNORTH AMERICA WITH A PHYLOGENETIC ANALYSIS BASED ON MOLECULAR DATA AND THE DESCRIPTION OF FOUR NEW SPECIES GONZALO GIRIBET' AND WILLIAM A. SHEAR? Asstract. The North American fauna of the Laur- asian family Sironidae is examined phylogenetically and compared with species from Europe and Japan. The North American clade is not resolved as monophyletic. The phylogenetic analyses and detailed morphological study identified four cryptic species of sironids in the western United States, formerly considered within the geographical and morphological range of Siro acaroides (Ewing, 1923). These four species are described as Siro boyerae sp. nov., Siro calaveras sp. nov., Siro clousi sp. nov., and Siro shasta sp. nov. We also provide new localities for the previously known species in the western United States. Siro boyerae sp. nov. forms a clade with Siro kamiakensis (Newell, 1943) and with the East Coast species Siro exilis Hoffman, 1963, characterized by the presence of narrow coxae IIT that do not meet along the midline. The affinities of S. calaveras sp. nov., S. clousi sp. nov., and S. shasta sp. nov. remain largely unresolved, but S. clousi sp. nov., is not related to S. acaroides despite being found sympatrically. INTRODUCTION The cyphophthalmid genus Siro currently includes a series of species found in North America and continental Western Europe (Giribet, 2000; Juberthie, 1970; Novak and Giribet, 2006; Shear, 1980). The status of the European members of the genus Siro has been recently revised, and the radiation of species related to Cyphophthalmus dur- icorius Joseph, 1868, in the Balkans and adjacent geographic areas seems to be ‘Museum of Comparative Zoology, Department of Organismic and Evolutionary Biology, Harvard Uni- versity, 26 Oxford Street, Cambridge, Massachusetts 02138. * Department of Biology, Hampden-Sydney College, Hampden-Sydney, Virginia 23943. unrelated to Siro rubens Latreille, 1804, and therefore considered a different genus (Boyer et al., 2005; Karaman, 2008; Muri- enne et al., 2010). In this article, we restrict the concept of Siro to a clade of recent Western European species composed of S. rubens; Siro carpaticus Rafalski, 1956; Siro crassus Novak & Giribet, 2006; and Siro valleorum Chemini, 1990, and to a clade of several North American species: Siro acar- oides (Ewing, 1923); Siro exilis Hofman, 1963; Siro kamiakensis (Newell, 1943); and Siro sonoma Shear, 1980. The four previ- ously known North American species were revised by Shear (1980) and _ profusely illustrated by de Bivort and Giribet (2004: figs. 10-39). Siro acaroides was described in 1923 as the type of the new genus Holosiro Ewing, 1923, this species being the first cyphophthalmid discovered in the New World (Ewing, 1923). Later, it was recognized that the species could not be easily distinguished from the European Siro at the generic level, and Holosiro was considered a junior synonym of Siro (Newell, 1943). In the same article, Newell described a new species of American sironid in the new genus Neosiro Newell, 1943, for the species Neosiro kamiakensis. The new genus was based on the divided fourth tarsus of the male. Both species inhabit western North America, each originally described from single locali- ties: S. acaroides from Benton County, southwestern Oregon, and N. kamiakensis from Whitman County in western Washing- ton. An eastern North American species, Siro Bull. Mus. Comp. Zool., 160(1): 1-33, October, 2010 ih 2 Bulletin of the Museum of Comparative Zoology, Vol. 160, No. 1 exilis, found in the Appalachian Mountains along the boundary ieee Virginia and West Virginia, was subsequently added to the list (Hoffman, 1963). Meanwhile, Davis (1933) had described Siro americanus from northwestern Florida; after an unwarranted sojourn in the genus Parasiro Hansen and Sgrensen, 1904 (Hin- ton, 1938), this species was made the type of the new genus Metasiro Juberthie, 1960 eee 1960), within the family Sironi- ae (or Sironinae of Juberthie, 1970) (Giribet, 2000; Juberthie, 1970; Shear, 1980). Later on, Hoffman (1963) proposed the synonym genus Floridogovea Hoffman, 1963, for Metasiro. Based on ample mor- phological and molecular evidence, Meta- siro is now considered a member of Neogoveidae (Boyer et al., 2007b; Giribet, 2007). In 1980, Shear had access to a wide range of collections that had been assembled since 1947 (Shear, 1980). Contrary to the asser- tions of Ewing (1923) and Newell (1947), Shear (1980) proposed that S. acaroides was widely distributed in the Coast Ranges from northern California to Puget Sound and that N. kamiakensis seine in at least one more locality in western Washington (Mt. Spokane) and at three places in Kootenai and Idaho Counties, in northern Idaho. Furthermore, Shear argued that the pre- ponderance of characters of N. kamiakensis were consistent with a placement in the genus Siro and so synonymized Newell’s genus Neosiro. Shear also added a distinc- tive fourth species of Siro, S. sonoma Shear, 1980, from Sonoma County in Northern California. Some loose ends were mentioned in Shear’s 1980 paper. In particular, a single female specimen from Calaveras County, California, in the Sierra Nevada Mountains, seemed clearly to be a new species, but Shear was reluctant to describe it from a single female example. Now additional material from that same collection has become available, and it is clear that this population represents a new, fifth species of American Siro. Additional material has also been recently collected by G. Giribet, S. Boyer, and R. M. Clouse in Calaveras Big Trees State Park, which was suitable for molecular work. The map Shear published in 1980 did not correlate well with the list of localities given; a location for S. acaroides is shown signif- icantly south of the California/Oregon border, but only Del Norte County records are listed in the text. This map symbol was added late in the preparation of the paper and referred to Shasta County specimens that were then considered S. acaroides. They are of a yet another new species. A field trip through Idaho, Washington, Oregon, ae Northern California by G. Giribet, S. Boyer, and R. M. Clouse in June 2005 yielded numerous collections of Cy- aa a including all known species or the western United States, with the exception of the elusive S. sonoma. The aim of this trip was to obtain more specimens of the new species from Calaveras County and Shasta County, as well as to revisit other cyphophthalmid localities to obtain speci- mens suitable for molecular work for all the NW USS. species. Two specimens of S: sonoma were collected by G. Giribet, T. Briggs, and D. Ubick in Monte Rio, December 2001. Phylogenetic analysis of the new specimens further revealed the presence of multiple cryptic lineages in the previously considered widespread eee S. acaroides. Two of these species that could be characterized morpho- logically are described here. The new species double the number of known American sironids but also indicate that our knowledge of the American sironid fauna is still in its infancy. California has not been intensively ex- plored for cyphophthalmids. They are most easily collected from Berlese samples; the success of this method was demonstrated by the many specimens and new records of S. acaroides obtained by Ellen Benedict (Shear, 1980). We have also been successful collecting many live specimens by sifting with a 4-mm mesh size or via extraction with Winkler apparatus. But other than these PHYLOGENETIC ANALYSIS OF SIRO IN NorRTH AMERICA © Giribet and Shear ° examples, most specimens have been obtained after occasional direct collecting. We predict that a thorough search of proper habitats in the Sierra Nevada, and both northern and southern Coast Ranges in California, will yield more new species of sironids. The distribution pattern of soil- dwelling organisms with species in the Appalielne in the east and the Coast Ranges and northern Idaho in the west often includes the central Rocky Mountains as well; Siro might be expected to turn up in Utah, Colorado, or New Mexico. With 43 extant species of sironids in Europe, it seems reasonable to expect that North America eventually could pe shown to have more species than the 10 we know now. MATERIALS AND METHODS Abbreviations for Repository Institutions AMNH American Museum of Natural History, New York, New York, USA The Natural History Museum, London, United Kingdom California Academy of Sciences, San Francisco, California, USA Field Museum of Natural History, Chicago, Illinois, USA (usually, FMNH for Field Museum of Natural History) Essig Museum of Entomology, U.C. Berkeley, Berkeley, California, USA Field Museum of Natural History, Chicago, Illinois, USA (in some labels, CNHM for Chicago Natural History Museum) Museum of Comparative Zoology, Harvard University, Cambridge, Massachusetts, USA Muséum dhistoire naturelle, Geneva, Switzerland Senckenberg Museum, Frankfurt am Main, Frankfurt, Germany BMNH CAS CNHM EME FMNH MCZ MHNG SMF Morphological Methods For each species, the male holotype and a female paratype were photographed using a JVC KY-F70B digital camera mounted on a Leica MZ 12.5 stereomicroscope. A series of images (ca. 10) were taken at different focal planes and assembled with the dedi- cated software package Auto-Montage Pro Version 5.00.0271 (Syncroscopy, Freder- ick, Maryland, USA). Each specimen was photographed in dorsal, ventral, and lateral views, and when available, the holotype was always photographed. Full body measurements of the holotype and a female paratype were then taken from these photographs in Adobe Photoshop CS3 with the “Analysis” menu and were recorded in a spreadsheet. Total body length refers to the distance between midpoint of anterior and midpoint of posterior margin of the dorsal scutum. Body width refers to the maximum width, whether recorded in the prosomal or in the opisthosomal region. One male and one female specimen of each species were examined with a FEI Quanta 200 SEM (Peabody, Massachusetts, USA). Appendage and body part measure- ments were taken from the digital micro- graphs in Adobe Photoshop CS3 with the “Analysis” menu and were recorded in a spreadsheet. Measurements of the chelic- era, palp, and leg articles were mostly taken on their dorsal side, from the midpoint of the anterior margin to the midpoint of the posterior margin. Depths were measured on the lateral side at the widest portion, except for tarsus IV of the male, which was measured behind the adenostyle. Tarsal length does not include the claw. The position of the adenostyle on tarsus IV is given at the more clearly marked distal point, where it abruptly rises from the dorsal surface of the tarsus. Finally, some body measurements were taken with an ocular micrometer on an Olympus SZH dissecting microscope, at 50x. Measurements of appendages tempo- rarily mounted on microscope slides were 4 Bulletin of the Museum of Comparative Zoology, Vol. 160, No. 1 Figure 1. Map of the NW United States with the sampled localities for Siro acaroides (red), S. boyerae sp. nov. (navy blue), S. calaveras sp. nov. (yellow), S. clousi sp. nov. (white), S. kamiakensis (black), S. shasta sp. nov. (orange), and S. sonoma (green). For details on the collecting localities, see Table 1 and Supplemental Appendix 1. taken with an ocular micrometer on an Olympus BX50 compound microscope, at 100 with Nomarski differential interfer- ence contrast. Drawings were made using the latter microscope, equipped with a drawing tube; Nomarski contrast was used to clarify details of the spermatopositors. Molecular Sampling To evaluate the phylogenetic position of the new species and for testing the validity of the “widespread” species S. acaroides, we undertook phylogenetic analyses of molecu- lar data from specimens of all American sironids (see distribution map in Fig. 1) and multiple representatives of other sironid genera, maeladi Suzukielus from Japan, Paramiopsalis and Parasiro from the Iberian penineike Siro from France and Italy, and Cyphophthalmus from multiple localities in the Balkans (Table 1). Molecular data were obtained from freshly collected specimens preserved in 96% EtOH at —80° C. DNA from pre- served tissues was extracted with the use of the Qiagen DNeasy® Tissue Kit following eae protocols described, for example, by Boyer et al. (2005). Three different loci were chosen for this study. Ribosomal sequence data of complete 18S rRNA and a ca. 2.1-kb fragment of 28S rRNA were 5 PHYLOGENETIC ANALYSIS OF SIRO IN NorTH AMERICA © Giribet and Shear CLIEISOd 6LOBEIAV ZIIEISOd FIIETSOd CTISTsOd SLOBS9OAV OSS6E9AV ITTETSOd OLTETSOd 60ISTSOd 9CC6E9AV ZLSBEIAV ILS6SOAV SLSST6AV solstcOod 1OO -LEISLSOG & 9 5 Wa ae Pn. SO 4 BAUR g4 7 - eos - a 4 rT Wo aad eS He id } / \ gare Ey 4 b . Figure 9. Siro boyerae sp. nov. (A) Spermatopositor, ventral view. (B) Spermatopositor, dorsal view. (C) Ovipositor. (D) Ovipositor sensory organ. (C, Scale bar 200 um; D, scale bar 50 um.) kensis by its undivided male 4th tarsus. Likewise, it can easily be separated from S. shasta because the latter lacks ornamenta- tion in the legs. Description of Male. Small sironid of uniform chestnut brown color; total length of holotype 1.69 mm, maximum width at 3rd opisthosomal segment at 0.91 mm, body L/ W = 1.84 (Fig. 6A). Anterior margin of dorsal scutum slightly convex; prosomal region almost semicircular. Ozophores con- ical, of type II (sensu Juberthie, 1970), with subterminal ozopore (sensu Novak and Giribet, 2006); maximum width across ozophores 0.75 mm. Eyes absent. Trans- verse prosomal sulcus little conspicuous; transverse opisthosomal sulci inconspicu- ous. Dorsal scutum with maximum height at around segments 4-5 (Fig. 6C). Ventral prosomal complex (Figs. 6B, 7A, C) with coxae I-II free, coxae III-IV fused; coxae II and IV meeting along the midline, but not coxae IIT; coxae IV meeting along the midline for a distance greater than gonostome length; sternum absent; coxal pores clearly visible between coxae III and IV. Projections of coxae IV endites present in the anterior portion of gonostome wall (Fig. 7C). Male gonostome sub-semicircu- lar, with slightly concave posterior margin, wider than long (0.12 X 0.07 mm), and delimited laterally and anterolaterally by the elevated endites of coxae IV. Spiracles (Fig. 7H) of circular pug (sensu. Giribet an Boyer, 2002), circular to oval in shape in male, with a maximum diameter of 0.07 mm. Ventral opisthosomal region (Fig. 7A) without conspicuous modifications other than in the anal Se Opisthosomal tergite IX and sternites 8 and 9 fused into a broad corona analis (Fig. 7E); tergite VIII without modifi- cations. Anal plate oval, 0.21 x 0.15 mm, only ornamented in the sides, with a rather inconspicuous longitudinal central ridge that leaves two depressions laterally. Three anal gland pores on tergite VIII of males (Fig. 7G). Cuticle with tuberculate-microgra- nular surface (sensu Murphree, 1988; this is referred to as “ornamented” hereafter), nearly uniform in dorsal areas and in ventral areas, including coxae. Chelicerae (Fig. 8A) relatively short and robust; basal article in males 0.40 mm long, 0.14 mm wide, without a ventral process or a dorsal crest; 2nd article 0.59 mm long, 0.15 mm wide; movable finger 0.20 mm long; all articles with few setae, the proximal one almost entirely granulated but with sparse granulation; 8 uniform denticles on PHYLOGENETIC ANALYSIS OF SiRO IN NortH AMERICA ¢ Giribet and Shear 17 TABLE 3. LEG MEASUREMENTS (LENGTH/WIDTH, MM) IN S/RO BOYERAE SP. NOV. MEASUREMENTS REFER TO MALE PARATYPE MOUNTED FOR SEM. Leg Trochanter Femur Patella Tibia Metatarsus Tarsus Total I ().14/0.11 0.43/0.12 0).22/0.12 0.28/0.12 0.18/0.10 0).38/0.13 1.65 II 0.11/0.10 0.33/0.11 0.19/0.11 O:22/0512 ().15/0.09 0).33/0.11 35 III ().14/0.09 0223/0311 0.16/0.10 0.19/0.11 ().09/0.07 ().29/0.08 1.14 IV 0.22/0.10 0.33/0.11 0.21/0.11 0.23/0.12 ().16/0.09 0.34/0.13 Lal the cutting edge of each cheliceral finger (Fig. 8B). Second cheliceral segment not ornamented. Palp (Fig. 8C) 1.11 mm long, smooth, slightly ornamented on trochanter. Mea- surements of palpal article length in SEM male paratype (mm): trochanter 0.165, femur 0.311, patella 0.205, tibia 0.212, tarsus 0.220; claw 0.037 mm long. Legs relatively robust; leg formula I-IV- II-III (measurements in Table 3; Figs. 8D- G). Tarsus I with a concentration of setae, but not forming a distinct solea. Except for the tarsi I-IV and metatarsi I-II, all articles ornamented (Figs. 8D-G). Tarsus IV of male entire, with a narrow lamelliform adenostyle (Fig. 8G), subcylindrical at the base, with lateral pore; proximal margin at 40% of tarsal length. Claws hooked, smooth, without dentition or lateral pegs. Spermatopositor (Figs. 9A, B) short, typ- ical of sironids, smooth; with movable fingers, slightly curved outward, ending as hooks, longer than the membranous ata lobe; microtrichial formula: 4, 6, 5+5 (n = 1). Gonopore complex not observed. Description of Female. Total length 1.94 mm, maximum width 0.95 mm (L/W = 2.05; Fig. 6D). Ventral prosomal complex (Figs. 7B, D) only with coxae I-II meeting along the midline, coxae III delimiting the anterior part of gonostome. Female gonos- tome semicircular anteriorly, wider than long (Fig. 7D). Gonostome of female form- ing a tube. Corona analis not Sara or forming a tube (Fig. 7F). Female anal plate unmodified. Tarsus of leg IV (Fig. 8H) without modifications, narrower than that of males. Ovipositor (Figs. 9C, D) 0.84 mm long, typical of Siro (see Juberthie, 1967), com- posed of two apical lobes and 20 circular articles (n = 1), each with 8 short setae equal in length; these setae slightly longer toward the terminus; most een article without setae. Apical lobes (Fig. 9D) each with a long terminal seta and ca. 12 setae slightly increasing in length toward the tip; sensitive processes with multibranching setae with 6 endings. Because of SEM examination, we have not studied the receptaculum seminis. Siro calaveras Giribet & Shear sp. nov. (Figs. 10-13) Type Specimens Holotype. Male (MCZ 92898, ex MCZ DNA101623) from North Grove (41°03’49"N, 122°21'37"W), Calaveras Big Trees State Park, Calaveras Co., CALIFOR- NIA, collected 23.vi.2005 by S. L. Boyer, R. M. Clouse, & G. Giribet; litter sifting (Figs. 10A-C). Paratypes. Three males and 3 females (MCZ 92899, ex MCZ DNAI101623):; same collecting data as holotype; 1 male, 3 females, 1 juvenile (MCZ DNA101623), same collecting data as holotype (1 male and 1 female used for DNA extraction); 2 females, 2 juveniles (AMNH), from North Grove, Sines Big Trees State Park, Calaveras Co., CALIFORNIA, collected 5.ii.1958 by L. M. Smith & R. O. Schuster; 8 males, 11 females, 3 juveniles (AMNH), from North Grove, Calaveras Big Trees State Park, Calaveras Co., CALIFORNIA, collected 10.iii.1958 by L. M. Smith & R. O. Schuster, rotten log; 4 males (3 dissected for genitalia), 2 females (AMNH), from North Grove, Calaveras Big Trees State Park, Calaveras Co., CALIFORNIA, collected 5.ii.1958 by L. M. Smith & R. O. Schuster; 18 Bulletin of the Museum of Comparative Zoology, Vol. 160, No. 1 Figure 10. Siro calaveras sp. nov. (A-C) Holotype male (MCZ 92898) in dorsal (A), vent ee ral (B), and lateral (C) view. (D-F) Paratype female (MCZ 92899) in dorsal (D), ventral (E), and lateral (F) view. Scale bars 500 um. 1 male, 1 female in SEM stubs (MCZ, ex AMNH), from North Grove, Calaveras Big Trees State Park, Calaveras Co., CALIFOR- NIA, collected 5.iii.1958 by L. M. Smith & R. O. Schuster; 1 female (CAS), from South Grove, Calaveras Big Trees State Park, Calaveras Co., CALIFORNIA, collected 13.vi.1956 by B. J. Adelson. Etymology. The species epithet is a noun in apposition, after Calaveras Co., Califor- nia. Diagnosis. Siro calaveras (Fig. 10) is a more slender species (L/W = 1.99) than S. acaroides (L/W = 1.50), to which it is similar in body length, and the legs are also roportionally shorter, as is male tarsus IV L/W = 2.0 as opposed to 2.9 in S. acaroides); S. peenas ta has a rather smooth palpal trochanter, whereas that of S. cala- veras is ornamented; the number of sper- matopositor microtrichiae of both species is the same. Siro calaveras is distinctly smaller than S. shasta (1.5 vs. 2.6 mm long), which lacks leg ornamentation and differs in spermatopositor microtrichiae. The species can easily be separated from S. sonoma by the unmodified ventral surface of its male tarsus IV and from S. kamiakensis by its undivided male tarsus IV. Unlike S. exilis, S. calaveras males have a concave 8th tergite. Finally, S. calaveras has a unique body profile (Fig. 1OA) among North American species, with the body widest across the posterior opisthosomal part, rather than opisthosomal tergites 2 or 3. Description of Male. Slender small sironid of uniform chestnut brown color; total length of holotype 1.53 mm, maximum width at prosoma at 0.77 mm, body L/W = 1.99 (Fig. 10A). Anterior margin of dorsal scutum slightly convex; prosomal region sub-semicircular. Ozophores conical, of PHYLOGENETIC ANALYSIS OF SzRO IN NortH AMERICA © Giribet and Shear 19 Figure 11. type II (sensu Juberthie, 1970), with sub- terminal ozopore (sensu Novak and Giribet, 2006), and entirely omamented; maximum width across ozophores 0.72 mm. Eyes absent. Transverse prosomal sulcus incon- spicuous; transverse opisthosomal sulci in- conspicuous. Dorsal scutum with maximum height posterior end, but very similar along the length of the animal (Fig. 10C). Siro calaveras sp. nov. paratype male (MCZ 92899) and female (MCZ 92899). (A) Paratype male in ventral position. (B) Paratype female in ventral position. (C) Male ventral thoracic complex. (D) Female ventral thoracic complex. (E) Male anal region. (F) Female anal region. (A, B, Scale bars 500 um; C, D, scale bars 300 um; E, F, scale bars 100 um.) Ventral prosomal complex (Figs. 10B, 11A, E) with coxae I-II free, coxae IIJ-IV fused; coxae II, III, and IV meeting along the midline; coxae IV meeting along the midline for a distance greater than gonos- tome length; sternum absent; coxal pores clearly visible between coxae III and IV. Projections of coxae IV endites present in the anterior portion of gonostome wall 20 Bulletin of the Museum of Comparative Zoology, Vol. 160, No. 1 Figure 12. Siro calaveras sp. nov. paratype male (MCZ 92899) and female (MCZ 92899). (A) Left palp of male. (B) Metatarsus and tarsus | of male. (C) Metatarsus and tarsus II of male. (D) Metatarsus and tarsus III of male. (E) Metatarsus and tarsus IV of male. (F) Detail of adenostyle. (G) Metatarsus and tarsus IV of female. (H) Spiracle. (A, Scale bar 200 um; B-E, G, scale bars 100 um; F, scale bar 50 um; H, scale bar 20 um.) (Fig. 11C). Male gonostome semicircular, 2002), circular to oval in shape in male, with straight posterior margin, wider than with a maximum diameter of 0.06 mm. long (0.10 X 0.06 mm), and delimited Ventral opisthosomal region (Fig. 11A) laterally and anterolaterally by the elevated without conspicuous modifications other endites of coxae IV. Spiracles (Fig. 12H) of _ than in the anal plate. Opisthosomal tergite circular type (sensu Giribet and Boyer, IX and sternites 8 and 9 fused into a broad PHYLOGENETIC ANALYSIS OF S7RO IN NortH AMERICA ¢ Giribet and Shear 2] Figure 13. Siro calaveras sp. nov. (A) Spermatopositor, ventral view. (B) Spermatopositor, dorsal view. corona analis (Fig. 11E); tergite VIII slight- ly bilobed. Anal plate oval, 0.22 < 0.17 mm, concave, mostly smooth, without ornamen- tation or a longitudinal carina. Three anal gland pores on tergite VIII of males (Fig. L1E). Cuticle with tuberculate—micro- granular surface (sensu Murphree, 1988), nearly uniform in dorsal areas and in ventral areas including coxae. Chelicerae robust; basal article in males 0.73 mm long, 0.17 mm wide, with a ventral process and a dorsal crest; 2nd article 0.60 mm long, 0.17 mm wide; movable finger 0.23 mm long; all articles with few setae, the proximal one almost entirely granulated with dense granulation; denticles on the cutting edge of each cheliceral finger uniform. Second cheliceral segment not ornamented. Palp (Fig. 12A) 1.13 mm long, smooth, slightly ornamented on trochanter. Mea- surements of palpal article length in SEM male paratype (mm): trochanter 0.16, femur 0.31, patella 0.20, tibia 0.22, tarsus 0.24; claw 0.04 mm long. Legs relatively robust; leg formula IV-I-II-III (measurements in Table 4; Figs. 12B—E). Tarsus I with a concentration of setae, but not forming a distinct solea. Except for tarsi I-IV and metatarsi I-II, all articles ornamented (Figs. 12B—E). Tarsus IV entire, globose, with a lamelliform adenostyle (Figs. 12E, F), subcylindrical at the base, with lateral pore (Fig. 12F); proximal margin at 31% of tarsal length. Claws hooked, smooth, without dentition or lateral pegs. Spermatopositor (Figs. 13A, B) short, me of sironids, smooth; with movable ingers, slightly curved outward, ee as hooks, longer than the membranous median lobe; microtrichial formula: 3, 4, 5+5 (n = 1). Gonopore complex not observed. Description of Female. Total length 1.74 mm, maximum width 0.80 mm (L/W = 2.18; Fig. 10D). Ventral prosomal com- plex (Figs. 11B, F) only with coxae [II meeting along the midline, coxae III delim- iting the anterior part of gonostome. Female gonostome subtrapezoidal, wider than long. Gonostome of female forming a tube. Corona analis not protruding or forming a tube (Figs. l1OD-F, 11B). Female anal plate (Fig. 11F) with modifications, slightly raised in the mid-posterior section, and forming two concave lateral areas; ornamentation is sparse. Tarsus of leg IV without modifica- tions, narrower than that of males. Ovipositor not studied. Siro clousi Giribet & Shear sp. nov. (Figs. 14-16) Type Specimens Holotype. Male (MCZ DNA101871 ex DNA101616large; used for DNA study) TABLE 4. LEG MEASUREMENTS (LENGTH/WIDTH, MM) IN S/RO CALAVERAS SP. NOV. MEASUREMENTS REFER TO MALE PARATYPE MOUNTED FOR SEM. Leg Trochanter Femur Patella I 0.20/0.10 ().44/0.10 0.22/0.11 Il 0.13/0.10 0.38/0.11 0.19/0.11 Ill 0.16/0.09 0).26/0.10 0.17/0.11 IV 0.25/0.10 ().40/0.11 0.24/0.12 Tibia Metatarsus Tarsus Total 0.30/0.11 ().20/0.09 0.39/0.12 iio 0.23/0.11 0.16/0.09 0.36/0.11 1.45 0.21/0.11 ().14/0.07 0).29/0.09 23 0.28/0.11 0).28/0.10 0.33/0.16 1.78 22 Bulletin of the Museum of Comparative Zoology, Vol. 160, No. 1 a ME i Ape f a. Figure 14. Siro clousi sp. nov. (A-C) Holotype male (MCZ DNA101871) in dorsal (A), ventral (B), and lateral (C) view. (D—F) Paratype female (MCZ DNA101871) in dorsal (D), ventral (E), and lateral (F) view. Scale bars 500 um. from Olalla Road, Lincoln Co., OREGON, collected 20.vi.2005 by S. L. Boyer, R. M. Clouse, & G. Giribet. Paratypes. Two females (1 female for DNA work; MCZ DNAI01871 ex DNA 101616large). Etymology. The species is named after cyphophthalmid biologist Ronald M. Clouse, who assisted collecting the type material of the species. Diagnosis. Siro clousi sp. nov. is easily distinguished from S. kamiakensis in that the latter has a divided male tarsus IV, from S. sonoma in that the latter has a mesal modification in the male tarsus IV, and from S. shasta in that the latter lacks ornamen- tation on the legs. The species is larger than S. acaroides, which live sympatrically and can also be distinguished from it by the spiracles, which are open in S. acaroides. The presence of the anal carina also distinguishes it from S. acaroides, S. boy- erae, S. calaveras, and S. shasta. The presence of coxae III meeting along the midline also distinguishes it from S. boy- erae, S. exilis, S. kamiakensis, and S. sonoma. Description of Male. Medium-sized sir- onid of uniform chestnut brown color; total length of holotype 1.89 mm, maximum width at 3rd opisthosomal segment at 1.05 mm, body L/W = 1.80 (Fig. 6A). Anterior margin of dorsal scutum straight or slightly concave; prosomal region trape- zoidal. Ozophores conical, of type II (sensu Juberthie, 1970), with subterminal ozopore (sensu. Novak and Géiribet, 2006), and entirely ornamented, with spiral ormamen- PHYLOGENETIC ANALYSIS OF S7RO IN NortH AMERICA ¢ Giribet and Shear 23 Figure 15. Siro clousisp. nov. paratype male (MCZ DNA101871). (A) Paratype male in ventral position. (B) Male ventral thoracic complex. (C) Male anal region. (D) Spiracle. (E) Detail of the ozophore ornamentation. (F) Detail of the anal gland openings. (A, scale bar 400 um; B, scale bar 200 um, C, scale bar 100 um; D, F, G, scale bars 20 um; E, scale bar 40 um.) 24 Bulletin of the Museum of Comparative Zoology, Vol. 160, No. 1 . * Yoo ce erent a ve o ©. wi og . ’ 8 Pyare) Figure 16. Siro clousi sp. nov. paratype male (MCZ DNA101871). (A) Left chelicera. (B) Left palp. (C) Metatarsus and tarsus I. (D) Metatarsus and tarsus II. (E) Metatarsus and tarsus III. (F) Metatarsus and tarsus IV. (G) Detail of the adenostyle. (H) Detail of leg | cuticle ornamentation. (A, B, scale bars 200 m; C-F, scale bars 100 um; G, scale bar 50 um; H, scale bar 5 um.) tation (sensu de Bivort and Giribet, 2004): maximum width across ozophores 0.95 mm. Eyes absent. Transverse prosomal sulcus inconspicuous; transverse opisthosomal sul- ci inconspicuous. Dorsal scutum with max- imum height at around segments 4-5 (Fig. 14C). Ventral prosomal complex (Figs. 14B, 15A, B) with coxae I-II free, coxae IIJ-IV fused; coxae II, II, and IV meeting along the midline; coxae IV meeting along the midline for a distance greater than gonos- tome length; sternum absent; coxal pores clearly visible between coxae III and IV. PHYLOGENETIC ANALYSIS OF SiRO IN NortuH AMERICA ¢ Giribet and Shear 25 TABLE 5). LEG MEASUREMENTS (LENGTH/WIDTH, MM) IN S7RO CLOUSI SP. NOV. MEASUREMI NTS REFER TO MALE PARATYPE MOUNTED FOR SEM. Leg Trochanter Femur Patella Tibia Metatarsus Tarsus Total I 0.17/0.14 0.62/0.13 0.31/0.12 0.41/0.12 ().26/0.10 0.51/0.12 2.28 II 0.15/0.13 0.57/0.13 0.36/0.13 0.35/0.13 0.23/0.09 0.46/0.11 ae Ill 0.16/0.13 0.42/0.12 0.24/0.13 0.28/0.13 0.23/0.09 ().42/0.10 LS IV 0.25/0.13 0.54/0.13 0).29/0.14 0.32/0.14 ().25/0.11 0.46/0.17 PFI Projections of coxae IV endites present in the anterior portion of gonostome wall (Fig. 15B). Male gonostome semicircular, with straight posterior margin, wider than long (0.13 X 0.07 mm), and delimited laterally and anterolaterally by the elevated endites of coxae IV. Spiracles (Fig. 15D) of circular type (sensu Giribet and Boyer, 2002), circular to oval in shape in male, with a maximum diameter of 0.05 mm. Ventral opisthosomal region (Fig. 15A) without conspicuous modifications other than in the anal plate. Opisthosomal tergite IX and sternites 8 and 9 fused into a broad corona analis (Fig. 15C). Anal plate oval, 0.24 X 0.18 mm, mostly ornamented, with a conspicuous longitudinal central ridge that leaves two lateral depressions deprived of ornamentation. Three anal gland pores on tergite VIII of males (Fig. 15F). Cuticle with tuberculate-microgranular surface (sensu Murphree, 1988), nearly uniform in dorsal areas and in ventral areas including coxae. Chelicerae (Fig. 16A) robust; basal article in males 0.69 mm long, 0.21 mm wide, with a ventral process A a dorsal crest; 2nd article 0.84 mm long, 0.18 mm wide; movable finger 0.30 mm long; all articles with few setae, the proximal one almost entirely granulated with dense granulation; aentieles on the cutting edge of each cheliceral finger uniform. Second cheliceral segment not ornamented. Palp (Fig. 16B) 1.65 mm long, smooth, slightly ornamented on trochanter. Mea- surements of palpal article length in SEM male Seeati (mm): trochanter 0.23, femur 0.46, patella 0.30, tibia 0.35, tarsus 0.30; claw 0.05 mm long. Legs relatively robust, leg formula I-II-IV-III (measurements in Table 5; Figs. 16C-—F). Tarsus I with a concentration of setae, but not forming a distinct solea. Except for the tarsi I-IV and metatarsi [-IJ, all articles ornamented (Figs. 16C—F). Tar- sus IV of male entire, with a narrow lamelliform adenostyle (Fig. 16F), subcy- lindrical at the base, with lateral pore (Fig. 16G); proximal margin at 39% of tarsal length. Claws hooked, smooth, without dentition or lateral pegs. Spermatopositor not studied. Description of Female. Total length 2.04 mm, maximum width 1.02 mm (L/W = 1.99; Fig. 14D). Ventral prosomal com- plex (Fig. 14E) only with coxae I-II meeting along the midline, coxae III delimiting the anterior part of gonostome. Female gonos- tome semicircular anteriorly, wider than long. Gonostome of female forming a tube. Corona analis not protruding or forming a tube (Figs. 14D-F). Female anal plate unmodified. Tarsus of leg IV without modifications, narrower than that of males. Ovipositor not studied. Notes. Siro clousi sp. nov. is sympatric with the more widespread species S. acaroides, a considerably smaller species. Originally the specimens collected in the type locality, Olalla Road, were labeled as “large” and “small,” but assigned the same MCZ DNA collection number, the vial containing the type material of the new species along with 4 males and 5 females of S. acaroides. Siro shasta sp. nov. (Figs. 17-20) Type Specimens Holotype. Male (AMNH) from 8 mi south of Dunsmuir, Shasta Co., CALIFORNIA, 26 Bulletin of the Museum of Comparative Zoology, Vol. 160, No. 1 if AN 0.50 mm ny ae Figure 17. Siro shasta sp. nov. (A-C) Holotype male (AMNH) in dorsal (A), ventral (B), and lateral (C) view. (D—F) Paratype female (AMNH) in dorsal (D), ventral (E), and lateral (F) view. Scale bars 500 um. collected 11.vii.1954 by R. O. Schuster & E. E. Gilbert (Figs. 6A—C). Paratypes. Eight males, 6 females (AMNH), same collecting data as holotype; 1 male (MCZ 92985, ex AMNH, on SEM stub), 1 female (MCZ 92986, ex AMNH, on SEM stub), same collecting data as holo- type; 3 males, 5 females (AMNH) from North of Hazel Creek, Shasta Co., CALI- FORNIA, collected 26.vi.1954 by R. O. Schuster & B. Adelson; 1 female (MCZ DNA101622) from Sims bridge, Shasta National Forest, Shasta Co., CALIFOR- NIA, collected 22.vi.2005 by S. L. Boyer, R. M. Clouse, t+ G. Giribet. Etymology. The species epithet is a noun in apposition, after Shasta Co., California. Diagnosis. Siro shasta (Fig. 17) is notably larger, with longer, thinner legs, than any of the other western North American Siro species, being about one-third longer than S. acaroides (2.6 vs. 1.5 mm); the body is more robust (L/W = 1.7) than that of S. calaveras, n. sp. (L/W = 1.9) and the 8th tergite of the male is noticeably more concave and bilobed than in any other North American species; in S. exilis the 8th tergite is convex. The entire 4th tarsus of the male differentiates the species from S. kamiakensis. The leg ornamentation differs considerably from all other North American species, being so sparse that it is barely noticeable elsewhere than on the trochanter and the dorsal part of the femur, whereas in all other species, legs I and II have a smooth tarsus and metatarsus only and legs III and IV have a smooth tarsus only. The ventral prosomal complex resem- bles mostly that of S. acaroides, S. calaveras, and S. clousi, in that the endites of coxae III meet along the midline, but not so in S. boyerae, S. exilis, or S. kamiakensis. The spiracles are similar to those of S. acaroides, in the form of an open circle, but differ from all other North American sironids, which have circular spiracles. The microtrichia of PHYLOGENETIC ANALYSIS OF SiRO IN NortH AMERICA ¢ Giribet and Shear 2 3 ~ — i a a ~< Sher _L> 4 ne re Figure 18. Siro shasta sp. nov. paratype male (MCZ 92985) and female (MCZ 92986). (A) Paratype male in ventral position. (B) Paratype female in ventral position. (C) Male ventral thoracic complex. (D) Female ventral thoracic complex. (E) Male anal region. (F) Female anal region. (A, B, scale bars 1 mm; C, D, scale bars 300 um; E, scale bar 100 um; F, scale bar 200 um.) the penis is unique among North American Siro, with 4 apical, 6 ventral, and 10 dorsal microtrichiae, whereas the unusually large movable fingers of the penis are like those of the western species group. Description of Male. Large sironid of uniform chestnut brown color; total length of holotype 2.33 mm, maximum width at 3rd opisthosomal segment at 1.38 mm, body L/W = 1.69 (Fig. 17A). Anterior margin of dorsal scutum bilobed; prosomal region sub-semicircular. Ozophores coni- cal, of type II (sensu Juberthie, 1970), with subterminal ozopore (sensu Novak and Giribet, 2006), and entirely ornamented; maximum width across ozophores 1.02 mm. 28 Bulletin of the Museum of Comparative Zoology, Vol. 160, No. 1 \ 3 ~ AN wh EI eo 3 Figure 19. Siro shasta sp. nov. paratype male (M male. (C) Metatarsus and tarsus | of male. (D) Metatarsus and tarsus Il of male. (E) Metatarsus and tarsus Ill of male. (F) Metatarsus and tarsus IV of male. (G) Detail of adenostyle. (H) Metatarsus and tarsus IV of female. (A, scale bar 500 um; B-F, H, scale bars 300 um; G, scale bar 50 um.) Eyes absent. Transverse prosomal sulcus Ventral prosomal complex (Figs. 18A, E) inconspicuous; transverse opisthosomal sul- with coxae I-II free, coxae III-IV fused; ci inconspicuous. Dorsal scutum with coxae II, III, and IV meeting along the maximum height at around segments 4-5 midline; coxae IV meeting along the midline (Fig. 17C). for a distance slightly greater than gonos- PHYLOGENETIC ANALYSIS OF SiRO IN NortH AMERICA ¢ Giribet and Shear 29 Figure 20. Siro shasta sp. nov. (A) Spiracle (scale bar 20 um). (B) Spermatopositor, ventral view. (C) Spermatopositor, dorsal view. tome length; sternum absent; coxal pores clearly visible between coxae III and IV. Projections of coxae IV endites present in the anterior portion of gonostome wall (Fig. 18C). Male gonostome almost oval in shape, with the posterior margin forming a concave lip, wider than long (0.14 xX 0.10 mm), and delimited laterally and anterolaterally by the elevated endites of coxae IV. Spiracles (Fig. 20A) in the shape of an open circle (sensu Giribet and Boyer, 2002), with a maximum diameter of 0.10 mm. Ventral opisthosomal region (Fig. 18A) without conspicuous modifications other than in the anal ee Opisthosomal tergite IX and sternites 8 and 9 fused into a broad corona analis (Fig. 18E). Anal plate oval, 0.32 xX 0.21 mm, completely depressed and smooth, except for the anterior and lateral rims. Three anal gland pores on tergite VIII of males. Tergite VIII depressed posteriorly, forming a Wiel posterior end. Cuticle with tubercu- late-microgranular surface (sensu Murphree, 1988), nearly uniform in dorsal areas and in ventral areas including coxae. Chelicerae (Fig. 19A) robust; basal article in males 0.85 mm long, 0.23 mm wide, with a ventral process but without a dorsal crest; 2nd article 1.07 mm long, 0.19 mm wide; movable finger 0.35 mm long; all articles with few setae, the SE one almost entirely granulated he with sparse granulation; denticles on the cutting edge of each cheliceral finger uniform, triangular, with 8 denticles in the moveable finger. Second cheliceral segment smooth, not ornamented. Palp (Fig. 19B) 1.91 mm long, smooth, slightly eee on trochanter. Mea- surements of palpal article length in SEM male paratype (mm): trochanter 0.27, femur 0.55, patella 0.29, tibia 0.40, tarsus 0.40; claw 0.06 mm long. Legs slender, leg formula I-IV-II-III (measurements in Table 6; Figs. 19C-—F). Tarsus I with a concentration of setae, but not forming a distinct solea. Tarsi I-IV and metatarsi I-II smooth, all other articles presenting sparse ornamentation, to the point that metatarsi II-IV are almost smooth, but present a few tuberculate structures (Figs. 19E, F). Tarsus IV of male entire, swollen, math a small lamelliform adenostyle (Fig. 19F), subcylindrical at the base, with lateral pore (Fig. 19G); proximal margin at 35% of tarsal length. Claws hooked, smooth, without dentition or lateral pegs. Spermatopositor (Figs. 20B, C) short, typical of sironids, smooth; with movable 30 Bulletin of the Museum of Comparative Zoology, Vol. 160, No. 1 fingers, slightly curved outward, ending as hooks, not much longer than the membra- nous median lobe; microtrichial formula: 4, 6, 5+5 (n = 1). Gonopore complex not observed. Description of Female. Total length 2.40 mm, maximum width 1.29 mm (L/W = 1.85; Fig. 17D). Ventral prosomal com- plex (Figs. 18B, F) only with coxae [-II meeting along the midline, coxae III delim- iting the anterior part of gonostome. Female gonostome near circular, wider than long. Gonostome of female forming a tube. Corona analis not protruding or forming a tube, with most a its surface deprived of macrotuberculate ornamentation, only pre- senting microtuberculate one (Fig. 18F). Tarsus of leg IV without modifications (Fig. 19H), narrower than that of males. Ovipositor not studied. Siro acaroides (Ewing, 1923) Holosiro acaroides Ewing 1923: 338. Siro acaroides: Newell 1947: 354; Shear 1980: 10. The following new records establish a new southern limit for the distribution of S. acaroides, by about 50 mi. The records for the locality, “18 miles south of Klamath,” were labeled as being from Del Norte Co., but that distance south of Klamath would be well into Humboldt Co., and the records are so given here. It is possible that S. acaroides occurs much farther south; there is a single juvenile (CAS) known from Mendocino Co., 2 mi south of Rockport, collected by C. W. O’Brien, 2.ii.1962. However, the possibility that this is another species cannot be dismissed. We have also found that S. acaroides exhibits at least one unique character when compared with other North American Siro: the palpal trochanter without tubercles (de Bivort and Giribet, 2002: fig. 16a). CALIFORNIA: Del Norte Co.: One male (CAS) near Crescent City, Smith River, collected 9.xi.1956 by J. Schuh; 2 males, 1 female (CAS) 5 mi south of Crescent Si collected 9.ix.1958; 7 males (1 for DNA work) and 6 females (MCZ DNA101620) from Kings Valley, near Crescent City, collected 21.vi.2005 by S. L. Boyer, R. M. Clouse, & G. Giribet. Humboldt Co.: Thirteen males, 8 females (CAS) 18 mi south of Klamath, collected 13.viii.1953; 28 males, 28 females (CAS) collected 19.ix.1953; 5 males, 4 females (CAS) Big Lagoon, collected 13.viii.1953 by G. A. Marsh & L. O. Schuster; 3 males, 3 females (CAS) _ freshwater, collected 18.viii.1952 by G. A. Marsh & L. O. Schuster; 5 males (CAS) Prairie Creek Redwoods State Park, collected 8.ix.1958 by L. M. Smith; 7 males (1 for DNA work) and 9 females (1 for DNA work; MCZ DNAI101621) from Lady Bird Johnson Grove, Redwood National and State Parks, collected 21.vi.2005 by S. L. Boyer, R. M. Clouse, & G. Giribet. OREGON: Benton Co.: Thirteen males (1 for DNA work) and 19 females (MCZ DNAI101618) from MacDonald State For- est, collected 20.vi.2005 by S. L. Boyer, R. M. Clouse, & G. Giribet. Douglas Co.: Eleven males (2 for DNA work) and 4 females (MCZ DNAI101619) from Elliot State Forest, Umpqua State Scenic Corri- dor, collected 20.vi.2005 by S. L. Boyer, R. M. Clouse, ¢+ G. Giribet. KEY TO MALES OF NORTH AMERICAN SIRONIDAE lia) Male 4th tarsus*divided' ae S. kamiakensis (Newell, 1943) TABLE 6. LEG MEASUREMENTS (LENGTH/WIDTH, MM) IN S/RO SHASTA SP. NOV. MEASUREMENTS REFER TO MALE PARATYPE MOUNTED FOR SEM. Leg Trochanter Femur Patella I 0.27/0.18 0.83/0.14 0.38/0.16 I] 0.18/0.15 0.68/0.16 0.28/0.16 Ill 0.25/0.16 OVO 0).26/0.17 IV 0.43/0.15 ().72/0.16 0.38/2.0 Tibia Metatarsus Tarsus Total 0.58/0.16 0.31/0.12 0).70/0.16 R07 0).45/0.16 ().26/0.11 0.62/0.13 2.47 0.41/0.16 ().25/0.11 0.57/0.13 OS, 0.50/0.18 0.31/0.13 0.64/0.21 2.98 a8 oc Figure PHYLOGENETIC ANALYSIS OF Siro IN NortTH AMERICA ¢ Giribet and Shear ea 3] 7 as ad enw 21. Male ventral thoracic complex of described North American sironids. (A) Siro acaroides (arrowhead indicates the area of contact of the endites of cozae Ill). (B) S. exilis. (C) S. kamiakensis. (D) S. sonoma. (A—D, scale bars 100 um.) tb: 2a. 2b. oa: 3b. Aa. Male 4th tarsus not divided Male 4th tarsus with a ventral lobe mesally excavated .................+.. Ree eels S. sonoma Shear, 1980 Male 4th tarsus without a ventral lobe Male tergite VIII convex, with SATO ace erect acoder otoatetct ie vesereeeee-d. EXtlis Hoffman, 1963. Male tergite VIII concave, without a knob Legs with very sparse ornamenta- tion S. shasta sp. nov. Ab. 5a. 5b. 6a. 6b. Ta. Legs with basal articles ornament- ed until tibia I and II and until metatarsi III and IV Spiracles in the form of an open circle S. acaroides (Ewing, 1923). Spiracles circular 6 Endites of coxae III not meeting along the midliney tc. qco1ty-o5'- isc che ee S. boyerae sp. nov. Endites of coxae III meeting along the midline eee rere eee eee ee eeeeeeeoes eo eee reese e eee eee eee eee 32 Bulletin of the Museum of Comparative Zoology, Vol. 160, No. 1 eee eee eee eee eee eee eeeee ACKNOWLEDGMENTS We are indebted to many students in the Giribet laboratory that assisted with this research. Sarah Boyer and Tone Novak provided comments that helped to improve this article. Sarah Boyer participated in collecting trips to Japan and the NW United States and generated sequence data for some outgroups; Ron Clouse participated in the collecting trip to the NW United States; Prashant Sharma, Ligia Benavides, and Ben de Bivort assisted with SEM; Ligia Benavides generated the automontage im- ages for the types; and Joey Pakes assisted with the molecular work. We are also indebted to Tom Briggs, Salvador Carranza, Michele Nishiguchi, Nobuo Tsurusaki, and Darrell Ubick for assisting with fieldwork in Sonoma Co. (California), Japan, and Spain. Marco Valle, Ivo Karaman, and Plamen Mitov provided specimens of Cyphophthal- mus and S. valleorum. Finally, the arachnid curators and curatorial staff of the AMNH (Lorenzo Prendini and Norman Platnick), CAS (Charles Griswold and Darrell Ubick), FMNH (Petra Sierwald), and EME (Jerry Powell) are acknowledged for their support and long-term loans, which we promise will be returned some day. This material is based on work supported by the National Science Foundation under Grant 0236871. LITERATURE CITED Boyer, S. L., J. M. BAKER, and G. Giripet. 2007a. Deep genetic divergences in Aoraki denticulata (Arach- nida, Opiliones, Cyphophthalmi): a widespread ‘mite harvestman’ defies DNA taxonomy. Molecu- lar Ecology 16: 4999-5016. Boyer, S. L., R. M. Crouse, L. R. BENavipes, P. SHARMA, P. J. SCHWENDINGER, I. KARUNARATHNA, and G. GiriBeT. 2007b. Biogeography of the world: a case study from cyphophthalmid Opiliones, a globally distributed group of arachnids. Journal of Biogeography 34: 2070-2085. Boyer, S. L., I. KARAMAN, and G. GiriBeT. 2005. The genus Cyphophthalmus (Arachnida, Opiliones, Cyphophthalmi) in Europe: a phylogenetic ap- proach to Balkan Peninsula biogeography. Molec- ular Phylogenetics and Evolution 36: 554-567. Cattey, K. M. 1994. Descriptions of new Hypochilus species from New Mexico and California with a cladistic analysis of the Hypochilidae (Araneae). American Museum Novitates 3088: 1-27. Davis, N. W. 1933. A new opilionid from Florida (Arachnida, Cyphophthalmi). Journal of the New York Entomological Society 41: 49-53. DE Bivort, B. L., and G. GiriBer. 2004. A new genus of cyphophthalmid from the Iberian Peninsula with a phylogenetic analysis of the Sironidae (Arachnida : Opiliones : Cyphophthalmi) and a SEM database of external morphology. Invertebrate Systematics 18: 7-52. De Laet, J. E. 2005. Parsimony and the problem of inapplicables in sequence data, pp. 81-116. In V. A. Albert (ed.), Parsimony, Phylogeny, and Geno- mics. Oxford: Oxford Univ. Press. Ewinc, H. E. 1923. Holosiro acaroides, new genus and species—the only New World representative of the mite-like phalangids of the suborder Cy- phophthalmi. Annals of the Entomological Society of America 16: 387-390. Farris, J. S. 1997. The future of phylogeny reconstruc- tion. Zoologica Scripta 26: 303-311. Farris, J. S., V. A. ALBERT, M. KALLERSJO, D. Lipscoms, and A. G. Kiuce. 1996. Parsimony jackknifing outperforms neighbor-joining. Cladistics 12: 99- 124. Farris, J. S., M. KALLERSJO, A. G. KLUGE, and C. BULT. 1995. Constructing a significance test for incon- gruence. Systematic Biology 44: 570-572. Foutmer, O., M. Buack, W. Hoen, R. Lutz, and R. C. VRIJENHOEK. 1994. DNA primers for amplification of mitochondrial cytochrome c oxidase subunit I from diverse metazoan invertebrates. Molecular Marine Biology and Biotechnology 3: 294-299. GiriBet, G. 2000. Catalogue of the Cyphophthalmi of the world (Arachnida, Opiliones). Revista Ibérica de Aracnologia 2: 49-76. . 2003. Stability in phylogenetic formulations and its relationship to nodal support. Systematic Biology 52: 554-564. . 2005. Generating implied alignments under direct optimization using POY. Cladistics 21: 396— 402. . 2007. Neogoveidae Shear, 1980, pp. 95-97. In R. Pinto-da-Rocha, G. Machado, and G. Giribet (eds.), Harvestmen: The Biology of Opiliones. Cambridge, Massachusetts: Harvard Univ. Press. GiriBet, G., and S. L. Boyer. 2002. A cladistic analysis of the cyphophthalmid genera (Opiliones, Cy- phophthalmi). The Journal of Arachnology 30: HO=I2s: GiriBeEt, G., G. D. EpcECoMBE, W. C. WHEELER, and C. Baspitr. 2002. Phylogeny and systematic position of Opiliones: a combined analysis of chelicerate relationships using morphological and molecular data. Cladistics 18: 5—70. GirIBET, G., L. Voct, A. PEREZ GONZALEZ, P. SHARMA, and A. B. Kury. 2010. A multilocus approach to harvestman (Arachnida: Opiliones) phylogeny with PHYLOGENETIC ANALYSIS OF SrRO IN NortH AMERICA © Giribet and Shear He emphasis on biogeography and the systematics of Laniatores. Cladistics 26. Available online only doi:10.111 l/j 1096-0031.2009.00296.x GorEBEL, A. M., T. A. RANKER, P. S. Corn, and R. G. OLMSTEAD. 2009. Mitochondrial DNA evolution in the Anaxyrus boreas species group. Molecular Phylogenetics and Evolution 50: 209-225. Gotosorr, P. A. 1999. Analyzing large data sets in reasonable times: solutions for composite optima. Cladistics 15: 415-428. . 2002. Techniques for analyzing large data sets, pp. 70-79. In R. DeSalle, G. Giribet, and W. Wheeler (eds.), Techniques in Molecular System- atics and Evolution. Basel: Brikhiuser Verlag. Hinton, B. E. 1938. A key to the genera of the Suborder Cyphophthalmi with a description and figures of Neogovea immsi, gen. et sp.n. (Arachni- da, Opiliones). Annals and Magazine of Natural History Ser.1] 2: 331-338. HorrMan, R. L. 1963. A new phalangid of the genus Siro from eastern United States, and taxonomic notes on other American sironids (Arach., Opi- liones). Senckenbergiana Biologica 44: 129-139. JuBertuie, C. 1960. Contribution a l'étude des opilions cyphophthalmes: description de Metasiro gen. n. Bulletin du Muséum National d’Histoire Naturelle 32: 235-941. . 1967. Siro rubens (Opilion, Cyphophthalme). Revue d’Ecologie et de Biologie du Sol 4: 155-171. . 1970. Les genres dopilions Sironinae (Cy- phophthalmes). Bulletin du Muséum National dW Histoire Naturelle 41: 1371-1390. KaraMan, I. M. 2008. Cyphophthalmi of Serbia (Arachnida, Opiliones), pp. 97-118. In D. Pavice- vic and M. Perreau (eds.), Advances in the studies of the fauna of the Balkan Peninsula. Papers dedicated to the memory of Guido Nonveiller. Belgrade: Institute for Nature Conservation of Serbia. Linton, E. W [Internet]. MacGDE: Genetic Data Environment for MacOS X. Mt. Pleasant (MI): Central Michigan University; c2005 [updated 2009 Jun 3; cited 2010 June 09]. Available from: http:// macgde.bio.cmich.edw/. MALLaTr, J., and C. J. WINCHELL. 2002. Testing the new animal phylogeny: first use of combined large- subunit and small-subunit rRNA gene sequences to classify the protostomes. Molecular Biology and Evolution 19: 289-301. Mickevicn, M. F., and J. S. Farris. 1981. The implications of congruence in Menidia. Systematic Zoology 30: 351-370. MUuRIENNE, J., I. KARAMAN, and G. GrriBetT. (2010). Explosive evolution of an ancient group of Cyphophthalmi (Arachnida: Opiliones) in the Balkan Peninsula. Journal of Biogeography 37: 90-102. MurPuReE, C. S. 1988. Morphology of the dorsal integument of ten opilionid species (Arachnida, Opiliones). The Journal of Arachnology 16: 237- 252. NEWELL, I. M. 1943. A new sironid from North America (Opiliones, Cyphophthalmi, Sironidae). Transac- tions of the American Microscopical Society 62: 416-499. . 1947. The rediscovery and clarification of Siro acaroides (Ewing) (Opiliones, Cyphophthalmi, Sironidae). Transactions of the American Micro- scopical Society 66: 354-365. Novak, T., and G. Giripet. 2006. A new species of Cyphophthalmi (Arachnida, Opiliones, Sironidae) from Eastern Slovenia. Zootaxa 1330: 27-42. SCHWENDINGER, P. J., and G. GiriBet. 2005. The systematics of the south-east Asian genus Fangensis Rambla (Opiliones : Cyphophthalmi : Stylocelli- dae). Invertebrate Systematics 19: 297-323. SuearR, W. A. 1977. The opilionid genus Neogovea Hinton, with a description of the first troglobitic cyphophthalmid from the western hemisphere (Opiliones, Cyphophthalmi). The Journal of Ara- chnology 3: 165-175. . 1980. A review of the Cyphophthalmi of the United States and Mexico, with a proposed reclassification of the suborder (Arachnida, Opi- liones). American Museum Novitates 2705: 1-34. WHEELER, W. C. 1995. Sequence alignment, parameter sensitivity, and the phylogenetic analysis of molec- ular data. Systematic Biology 44: 321-331. 1996. Optimization alignment: the end of multiple sequence alignment in phylogenetics? Cladistics 12: 1-9. . 2003. Implied alignment: a synapomorphy- based multiple-sequence alignment method and its use in cladogram search. Cladistics 19: 261-268. WHEELER, W. C., D. GLADSTEIN, and J. De Laer. 2004. POY Version 3.0. New York: American Museum of Natural History. ~——s eel jyge f aw “ i 5 ious -é US ISSN 0027-4100 MCZ Publications Museum of Comparative Zoology Harvard University 26 Oxford Street Cambridge, MA 02138 mczpublications@mcz.harvard.edu © The President and Fellows of Harvard College 2010. Bulletin of the Museum of Comparative Zoology Volume 160, Number 2 15 December 2011 Anolis chrysolepis Dumeril & Bibron, 1837 (Squamata: Polychrotidae) revisited: Molecular phylogeny and taxonomy of the Anolis chrysolepis species group ANNELISE B. DPANGIOLELLA, TONY GAMBLE, TERESA C.S. AVILA-PIRES, GUARINO R. COLLI, BRICE P. NOONAN, AND LAURIE J. VITT HARVARD UNIVERSITY | CAMBRIDGE, MASSACHUSETTS, U.S.A. BULLETIN OF THE Museum of Comparative Zoology BOARD OF EDITORS Editor: Jonathan Losos Managing Editor: Adam Baldinger Editorial Assistant: Samantha Edelheit Associate Editors: Andrew Biewener, Scott Edwards, Brian Farrell, Farish Jenkins, George Lauder, Gonzalo Giribet, Hopi Hoekstra, Jim Hanken, Jim McCarthy, Naomi Pierce, and Robert Woollacott Publications Issued or Distributed by the Museum of Comparative Zoology Harvard University Bulletin 1863- Breviora 1952- Memoirs 1865-1938 Johnsonia, Department of Mollusks, 1941-1974 Occasional Papers on Mollusks, 1945- General queries, questions about author guidelines, or permissions for MCZ Publications should be directed to the editorial assistant: MCZ Publications Museum of Comparative Zoology Harvard University 26 Oxford Street Cambridge, MA 02138 mczpublications@mcz.harvard.edu EXCHANGES AND REPRINTS All of our publications are offered for free on our website: http://www.mcz.harvard.edu/Publications/index.html To purchase individual reprints or to join our exchange program, please contact Susan DeSanctis at the Ernst Mayr Library: mayrlib@oeb.harvard.edu. This publication has been printed on acid-free permanent paper stock. © The President and Fellows of Harvard College 2011. ANOLIS CHRYSOLEPIS DUMERIL AND BIBRON, 1837 (SQUAMATA: IGUANIDAE), REVISITED: MOLECULAR PHYLOGENY AND TAXONOMY OF THE ANOLIS CHRYSOLEPIS SPECIES GROUP ANNELISE B. D’ANGIOLELLA,' TONY GAMBLE,? TERESA C. S. AVILA-PIRES,? GUARINO R. COLLI,* BRICE P. NOONAN,? AND LAURIE J. VITT® Asstract. The Anolis chrysolepis species group is distributed across the entire Amazon basin and currently consists of A. bombiceps and five subspecies of A. chrysolepis. These lizards are characterized by moderate size, relatively narrow digital pads, and a small dewlap that does not reach the axilla. We used the mitochondrial gene ND2 to estimate phylogenetic relationships among putative subspecies of A. chryso- lepis and taxa previously hypothesized to be their close relatives. We also assessed the congruence between molecular and morphological daiasen to evaluate the taxonomic status of group members. On the basis of the two datasets, we present a new taxonomy, elevating each putative subspecies of A. chrysolepis to species status. We provide new morphological diagnoses and new distributional data for each species. Key words: Anolis, Amazon, mole- cular phylogeny, taxonomy Touanidae, le Resumo. O grupo de espécies Anolis chrysolepis atualmente consiste em A. bombiceps e cinco sub- espécies de A. chrysolepis, ocupando toda a Bacia Amazonica. Esses lagartos sao caracterizados por tamanho moderado, Acrvelse digitais relativamente estreitas e um papo extensivel que nao chega as axilas. Nos utilizamos 0 gene mitocondrial ND2 para estimar as relacdes filogenéticas entre as subespécies de A. chrysolepis e taxons previamente considerados_ par- entes proximos. Nos também determinamos a con- gruéncia entre conjuntos de dados morfoldgicos e moleculares, para avaliar o status taxondmico dos membros desse grupo. Com base nos dois conjuntos de dados, apresentamos uma nova taxonomia, elevando ‘Programa de Pos-Graduacaéo em Zoologia UFPA- MPEG, Belém, PA, Brazil. Author for correspondence (annelise. dangiolella@gmail.com). * University of Minnesota, Minneapolis, Minnesota. * Museu Paraense Emilio Goeldi, Belém, PA, Brazil. * Universidade de Brasilia, Brasilia, DF, Brazil. ° The University of Mississippi, University, Mississippi. ° University of Oklahoma, Norman, Oklahoma. Bull. Mus. Comp. Zool., 160(2): 35-63, December, cada subespécie de A. chrysolepis ao status de espécie. Fornecemos novas diagnoses morfologicas e novos dados de distribuic&o para cada espécie. Palavras-chave: Anolis, AmazOnia, Iguanidae, filo- genia molecular, taxonomia The Pleistocene Refuge Hypothesis pro- posed almost simultaneously by Haffer (1969) and Vanzolini and Williams (1970) posits that patches of lowland tropical forest that existed during dry periods in the Pleistocene served as core areas for speci- ation in birds and in the lizard complex Anolis chrysolepis, respectively. Although the Pleistocene Refuge Hypothesis has been falsified for mente of the A. chrysolepis species group because diversification oc- curred much earlier (15 mya) than the Pleistocene (Glor et al., 2001), relationships among all members of the group have not been worked out and related taxa (e. CoA meridionalis and A. bombiceps) have not been properly placed with reference to the A. chrysolepis complex, and current names do not accurately reflect the evolutionary history of the group (Glor et al., 2001; Nicholson et al., 2005). Because the A. chrysolepis species group has been and continues to be a model for evolutionary (Nicholson et al., 2006, 2007: Schaad and Poe, 2010) and ecological (Vitt and Zani, 1996: Vitt et al., 2001, 2008) studies, it is critical that their relationships be properly understood. Here we present a phylogenetic hypothesis for the chrysolepis species 2011 39 36 Bulletin of the Museum of Comparative Zoology, Vol. 160, No. 2 group using a much larger set of samples than was available previously and provide species names for taxa that can be identified as independent evolutionary lineages. Fol- lowing de Queiroz (2007), we consider independent evolutionary lineages, here recognized on the basis of Sene milices, analogous to species. Results of this study shoul be directly applicable to phylogeo- graphic and phyloecological studies of the A. chrysolepis species group. The A. chrysolepis group comprises two species: A. chrysolepis Duméril and Bibron, 1837, and A. bombiceps Cope, 1876. Anolis chrysolepis is currently composed of five subspecies: A. chrysolepis chrysolepis, in eastern Guiana (Brazil, French Guiana, Suriname, and southern Guyana); A. chry- solepis planiceps Troschel, 1848, in western Guiana (Brazil, Suriname, northwestern Guyana, Venezuela, and Trinidad); A. chry- solepis scypheus Cope, 1864, in western Amazonia (Colombia, Ecuador, Peru, and northwestern Brazil); A. chrysolepis tandai Avila-Pires, 1995, in southwestern Amazo- nia (Brazil and Peru); and A. chrysolepis brasiliensis Vanzolini and Williams, 1970, in Brazil, from Maranhao and enclaves of open vegetation in southern Para south to Sao Paulo (Vanzolini and Williams, 1970; Avila- Pires, 1995: Icochea et al., 2001: Santos- Jr et al., 2007). Anolis bombiceps occurs in western Amazonia, in Peru, Colombia, and Brazil, at least in partial sympatry with A. c. scypheus and perhaps also with A. c. tandai (Avila-Pires, 1995). Members of the A. chrysolepis group are characterized by their moderate size (up to 83 mm snout—vent length); short heads; supraorbital semicir- cles usually forming a pronounced ridge; relatively narrow digital pads, with distal lamellae under phalanx ii tor ‘ming a slightly prominent border; a dewlap that does not reach the axilla and is present in both sexes (but smaller in females); and keeled, imbri- cate ventral scales that are distinctly larger than dorsals. The A. chrysolepis species group was examined morphologically by Vanzolini and Williams (1970), who recognized four sub- ‘the name A. nitens. species of A. chrysolepis and a_ distinct species, A. bombiceps. Vanzolini and Wil- liams (1970: 13) believed the level of differentiation between the subspecies were “closest to species difference, and indica- tive, perhaps, of past and future potential species formation.” Anolis chiysdleg was later examined by Avila-Pires (1995) under She described another subspecies, A. n. tandai, and observed that most specimens occurring in areas of intergradation according to Vanzolini and Williams (1970) could be assigned to one of the recognized subspecies. Very little subsequent taxonomic research has been conducted on the species of the A. chrysolepis group. One molecular phyloge- netic study included three of the described A. chrysolepis subspecies and found they formed a weakly supported clade (Glor et al., 2001). Glor et al. (2001: 2664) concluded that, “further study of geographical genetic interactions among these subspecies proba- bly will reveal that they are distinct species. Additional molecular phylogenetic research, with broad outgroup sampling, recovered a well-supported an consisting of A. onca, A. annectans, A. lineatus, A. auratus, A. meridionalis, and A. chrysolepis, al- though A. chrysolepis was represented by just a single individual from Roraima, Brazil (Nicholson et al., 2005). Members of this clade were included in another phylogenetic analysis (Nicholson et al., 2006), using the same three A. chrysolepis subspecies of Glor et al. (2001), which recovered a paraphyletic A. chrysolepis. Nicholson et al. (2006) found that A. c. tandai was more closely related to A. meridionalis and the A. onca + A. annectans clade, whereas A. c. scypheus and A. c. planiceps formed a clade that was the sister group to the remaining species + A. auratus. Like Glor et al. (2001), Nicholson et al. (2006) stressed the need for additional research into the systematics of A, chrysolepis and the possible existence of cryptic species. Anolis bombiceps has not been included in any molecular studies so far. The name A. chi ysolepis has a long and confusing history, with both A. nitens ae A. ANOLIS CHRYSOLEPIS SPECIES GROUP ¢ D’Angiolella et al. =) chrysolepis considered valid names for the species (Hoogmoed, 1973; Avila-Pires, 1995; Myers and Donnelly, 2008). Myers and Donnelly (2008) presente ‘da detailed history of the use of these names, and Myers (2008) requested the International Com- mission of Zoological Nomenclature (ICZN ) to give precedence of A. chrysolepis Du- miceil and Bibron, 1837, over Draconura nitens Wagler, 1830, which was accepted (ICZN, 2010). In the present work, we analyzed mito- chondrial DNA from the protein coding gene ND2 and associated tRNA and mor- phological data from all five described subspecies of A. chrysolepis and related taxa to 1) recover the phylogenetic relation- ships among subspecies of A. chrysolepis ye test previous phylogenetic hypotheses, 2) evaluate the taxonomic status of de- scribed subspecies of A. chrysolepis, and 3) present a revised taxonomy that Incorpo- rates this phylogenetic information. MATERIALS AND METHODS Taxon Sampling and DNA Sequencing We sampled representatives of each of the five abegecies of A. chrysolepis (Ta- ble 1, Figure 1). Species previously shown to be closely related to A. chrysolepis were also included either from newly sequenced samples (e.g., A. bombiceps) or from previ- ously published GenBank material (Glor et al., 2001: Nicholson, 2002; Nicholson et al., 2005, 2006). Genomic DNA was extracted from muscle, liver, or tail a using DNeasy Blood and Tissue Kit (Qiagen, Valencia, California). Polymerase shai re- action was used to amplify portions of the mitochondrial protein-coding gene ND2 (NADH Mids eabumet 2) and adjacent tRNAs with primers LVT_Met- f.6_AnCr (AAGCTATTGGGCCCATACC) ancdeby las, AnCr (AAAGTGYTPGAC- TTGCATTCA) (Rodriguez Robles et al., 2007). Polymerase chain reaction cleanup and DNA sequencing was performed by Agencourt Bioscience (Beverly, M Massachu- setts). Sequences were edited and aligned using SHOUENCHIER. ver:. 4.2 (Gene Codes, Ann Arbor, Michigan). ND2_ se- quences were translated into amino acids using MacClade ver. 4.08 (Maddison and Maddison, 1992) to confirm alignment and gap placement and ensure there were no premature stop codons. Phylogenetic Analyses We analyzed the ND2 data using parsi- mony in PAUP ver. 4.0b10 (Swofford, 2001). Parsimony analysis was conducted using a heuristic search with 1,000 random taxon additions and tree bisection and reconnection (TBR) branch swapping and all characters equally weighted. We con- ducted 1,000 bootstrap replicates with 25 random additions per replicate to assess nodal support (Felsenstein, 1985). Mitochondrial DNA (mtDNA) has been widely used to recover phylogenetic rela- tionships among species and to delimit species (Avise et al., 1998; Grau et al., 2005: Gamble et al., 2008: Fenwick et al., 2009), and because of its shorter coalescent times, it is considered a good indicator of © population history and species limits (Avise et al., 2000; Wiens and Hollingsworth, 2000; Wiens and Penkrot, 2002: Tink and Barrow- clough, 2008; Barrowclough and Zink, 2009). However, the high substitution rate of mitochondrial DNA. makes saturation, especially at third codon positions, a possi- ble problem for accurate phylogenetic reconstruction (Jukes, 1987; Yoder et al., 1996, Glor et al., 2001, Hinder and: Tur ell. 2003). One way to minimize the effects of saturation is to use model-based phyloge- netic methods like maximum likelihood (ML) and Bayesian analyses (Felsenstein, 1978; Jukes, 1987; Huelsenbeck et al., 2001, Lartillot et al., 2007). Additionally, the use of partitioned model-based analyses, with separate models of molecular evolution for each gene or codon, can minimize phyloge- netic error (Bull et al., 1993: Lemmon and Moriarty, 2004; Nylander et al., 2004; Brandley et al., 2005). We conducted Bayesian analyses using MrBayes 3.1.2 38 Bulletin of the Museum of Comparative Zoology, Vol. 160, No. 2 TABLE 1. MATERIAL EXAMINED FOR THE MOLECULAR PHYLOGENETIC ANALYSIS, INCLUDING TAXON NAME, MUSEUM NUMBERS, SPECIMEN LOCALITY, Anolis Taxon A. tandai A. chrysolepis A. scypheus A. planiceps A. brasiliensis A. auratus A. bombiceps A. fuscoauratus A. meridionalis . lineatus _ onca . annectens sericeus _ isthmicus . laeviventris sagrei utilensis . grahami loveridgei . uniformis . crassulus . carolinensis poybs ete eS Se ID No. MPEG 22285 MPEG 25029 LSUMZ H-14098 MPEG 25060 LSUMZ H-16398 LSUMZ H-16474 LSUMZ H-13599 MPEG 26590 MPEG 26568 MPEG 26563 BPN780 MPEG 26584 BPN 1587 BRINEVSTS BPN 1874 LSUMZ H-12592 LSUMZ H-12543 LSUMZ H-12989 LSUMZ H-12300 BPN 1080 BPN 1082 BPN 228 BPN 96 CHUNB 45077 CHUNB 45075 CHUNB 08842 CHUNB 43282 CHUNB 34542 CHUNB 27158 CHUNB 11521 CHUNB 37528 CHUNB 37527 CHUNB 52471 GRC 16378 LSMUZ H-13928 KU 222145 LSUMZ H-13566 LF 166692 LSUMZ H-—5450 CIEZAH1156 CIEZAH1160 LACM7069 MFO191 MVCFC12252 KdQ1797 LDW 12480 JBL 250 USNM10683 n/a MZFC6458 CCA 8051 AND GENBANK NUMBERS. Locality Itaituba, Para, Brazil Juruti, Para, Brazil Rio Ituxi, Amazonas, Brazil Coari, Amazonas, Brazil Rio Solim6es, Amazonas, Brazil Rio Solim6es, Amazonas, Brazil Rio Jurua, Acre, Brazil Trombetas, Parad, Brazil Faro, Para, Brazil Acari, Para, Brazil Ralleighvallen, Suriname Maicuru, Para, Brazil Saul, French Guiana Saul, French Guiana Nouragues, French Guiana Reserva Faunistica Cuyabeno, Sucumbios Province, Ecuador Reserva Faunistica Cuyabeno, Sucumbios Province, Ecuador Reserva Faunistica Cuyabeno, Sucumbios Province, Ecuador Rio Ajarani, Roraima, Brazil Kartabo, Guyana Kartabo, Guyana Imbaimadai, Guyana Kartabo, Guyana Caseara, Tocantins, Brazil Minacu, Goias, Brazil Parauapebas, Para, Brazil Brasilia, Distrito Federal, Brazil Novo Progresso, Para, Brazil Mateiros, Tocantins, Brazil Palmas, Tocantins, Brazil Sao Domingos, Goias, Brazil Parana, Tocantins, Brazil Peixe, Tocantins, Brazil Alto Paraiso, Goias, Brazil Alter do Chao, Para, Brasil. 1.5 km N of Teniente Lopez, Loreto, Peru Rio Jurua, Acre, Brazil Reserva Mbaracayu, Canindeyu, Paraguay Netherlands Antilles Estado Falc6n, Venezuela Estado Faleén, Venezuela Costa Rica Mexico Guatemala La Habana, Cuba Honduras Discovery Bay, Jamaica Honduras Belize Mexico Unknown GenBank N191547 N191546 J J ip JN191545 JN191543 JN191544 JN191548 JN191532 JN191530 JN191531 ]N191534 ]N191533 JN191536 JNI91541 JN191540 AF337804 AF337802 N191568 — ]N191559 JN191557 JN191555 ]N191556 ]N227868 JN191565 JN191563 JN191562 ]N191560 JN191561 JN191571 ]N191570 AF337786 AY909760 AF 294287 DQ377357 DQ377345 AY909778 AY909762 AY909756 AF337778 AY909785 AF055938 AY909759 AY909784 AY909748 NC010972 ANOLIS CHRYSOLEPIS SPECIES GROUP * D’Angiolella et al. 39 (Huelsenbeck and Ronquist, 2001) on both the partitioned and unpartitioned datasets. Data were Rettioned by codon with a fourth partition for tRNAs and the optimal partitioning strategy selected using Bayes Factors calculated oar the harmonic mean likelihood values (Nylander et al., 2004; Brandley et al., 2005). We estimated the best fit model of sequence evolution for the data as a whole and for each partition separately using AIC scores in Modeltest (Posada, 2008), Bayesian analy ses were initialized with a neighbor-joining tree and two separate analyses conducted for each partitioning strategy. Each analysis consist- ed of seven heated chains and one cold chain run for 2 million generations, with sampling every 1,000 generations. Post- burnin convergence was checked by visual inspection of likelihood values by generation using Tracer 1.5 (Rambaut and Drummond, 2009) and visual inspection of split frequen- cies using AWTY (Nylander et al., 2008). We also conducted partitioned Maximum Likelihood analysis, with data partitioned as above using RAXML ver. 7.0.4 (Stamatakis, 2006) using the GTR+GAMMA model for all partitions. We conducted 1,000 “fast bootstrap” replicates and 10 separate max- imum likelihood searches. Bootstrap values =70 were considered as indicating strong support for both parsimony and ML anal- yses. We calculated net among group distances (Nei and Li, 1979) between major lineages of the “A. chrysolepis species group” using MEGA 4 (Kumar et al., 2008). We calcu- lated both uncorrected p-distances and corrected distances using the GTR model. On the basis of our best ML tree, we compared alternative phylogenetic hypoth- eses using the Shimodaira-Hasegawa (SH) test (Shimodaira and Hasegawa, 1999) and the Approximately Unbiased (AU) test (Shimodaira, 2002). Three alternative hy- potheses were considered: 1) monophyly of A. chrysolepis subspecies, excluding A. bombiceps and A. meridionalis; 2) mono- phyly of the A. chrysolepis ae + A. bombiceps, excluding only A. meridionalis; and 3) monophyly of all A. c. tandai specimens, as identified by morphological data. We used RAxML7.0.4 (Stamatakis, 2006) to compute per-site log likelihoods that were input into CONSEL (Shimodaira and Hasegawa, 2001) to calculate P values. We also tested alternative phylogenetic hypotheses in a Bayesian framework and calculated the Posterior Probabilities of alternative hypotheses using the tree filter option in PAUP*. Morphological Analyses We collected morphological and morpho- metric data from 403 specimens (Appendix 1) from the following zoological collections: MZUSP, Museu de Zoologia da Universi- dade de Sido Paulo: CHUNB. Colecao Herpetologica da Universidade de Brasilia: MPEG, Museu Paraense Emilio Goeldi: MCZ, Harvard Museum of Comparative Zoology; and KU, University of Kansas. Measurements were recorded with digital calipers to the nearest 0.1 mm on the right side of the body, except when specimens were damaged (in this case, the left side was used). Sele and measurement terminology follows Avila-Pires (1995). We recorded the following morphometric data: snout-vent length (SVL), tail length (from posterior edge of precloacal plate), head width, head height, mouth length (from tip of snout to posterior margin of mouth), distance between orbits (mini- mum), ear-opening diameter, distance be- tween nostrils (minimum), distance from mouth to ear (from anterior margin of ear- opening to posterior margin of mouth), snout length (from tip of snout to anterior margin oe orbit), interparietal length, tibia length, foot length (from toe IV ee to the heel. fourth toe length (from toe IV nail to toe base), and fount toe maximum width. Additionally, we recorded the following meristic characters: scales around midbody, postrostrals, supralabials, infralabials, loreals (under second canthal), canthals, scales between second canthals, scales between supraorbital semicircles (minimum), scales A() Bulletin of the Museum of Comparative Zoology, Vol. 160, No. 2 between interparietal and supraorbital semicircles (minimum), postmentals, fourth finger lamellae, and fourth toe lamellae. A few measurements and scale counts could not be assessed for all specimens analyzed. In multivariate analysis, cases with missing observations will be dr opped, weak- ening She analysis because of loss of TOA and. degrees of freedom. To avoid simply deleting entire rows of data, missing observations can be estimated using a variety of methods, including mean substitution, regression, expectation maxi- mization, maximum likelihood and multiple imputation (Tabachnick and Fidell, 2001, Quinn and Keough, 2002). Among these approaches for imputing values to missing observations, multiple imputation is the most robust and also makes fewer assump- tions about the pattern of missing observa- tions (Rubin, 1996; Van Buuren et al., 2006). Therefore, we imputed missing data using multivariate imputations by chained equations (Van Buuren et al., 2006), as implemented by package mice in R v. 2.12.0 (R Development Core Team, 2009). To partition the total morphometric variation between size and shape variation, we defined body size as an isometric size variable (Rohlf and Bookstein, 1987) fol- lowing Somers (1986): we calculated an isometric eigenvector, defined a priori with values equal to p °°, where p is the number of variables (Jolicoeur, 1963), and obtained scores from this eigenvector, hereafter called body size, by postmultiplying the n p matrix of log-tr: ansformed data, where n is the number ot observations, by the p x 1 isometric eigenvector. To remove the ef- fects of bode size from the log-transformed data, we used Burnaby’s method (Burnaby, 1966): we postmultiplied the n X p matrix of the log-transformed data by a p X p sy mmetric matrix. L, defined as: L=1, Viviana where sap Kp identity matrix, V is the isometric size eigenvector defined above, and V' is the transpose of matrix V (Rohlf and Bookstein, 1987). Hereafter, we refer to the resulting size-adjusted variables as shape variables. To identify morphometric and meristic variables that best discriminate among species, we used a stepwise discriminant analysis coupled with 100-fold cross-valida- tion to measure classification performance (Quinn and Keough, 2002) using the package klaR in R v. 2.12.0 (R Development Core Team, 2009). RESULTS Phylogenetic Analyses We sequenced 1,088 base pairs of the mitochondrial ND2 gene and adjacent tRNAs, which contained 82 variable sites and 633 parsimony-informative characters. Thirty-nine new mtDNA sequences from 34 localities (Fig. 1) are reported and aligned with 14 previously published sequences. A comparison of the partitioned Bayesian analyses to the unpartitioned analyses strongly favored the partitioned strategy (Bayes. Factors > 860). We observed con- vergence among multiple Bayesian runs and ailied post-burnin samples (burnin = 1,000) to estimate model parameters and tree topology (Fig. 2). The partitioned ML analysis produced a single tree (Fig. 3, In L = -—16,649.1489) that had a similar topology to the partitioned Bayesian con- sensus tree at well-supported nodes. The Parsimony analysis produced 54 equally most parsimonious trees (TL = 3,832, CI =. 0.337683, Rl = -0:6382552) 5 hGa— 0.230486, HI = 0.662317; Fig. 4), Subspe- cies formed strongly supported monophy- letic groups in all analyses, with the exception of specimens of A. c. tandai from Acre. All analyses also recovered a para- phyletic A. chi ysolepis with regard to A. bombiceps and A. meridionalis (Figs. a Sampled individuals of A. chi ysolepis, A bombiceps, and A. meridionalis were mem- bers of one of two clades: one (Clade A) composed of A. c. chrysolepis, A. c. tandai, ANOLIS CHRYSOLEPIS SPECIES GROUP * D’Angiolella et al. 4] Bo . ‘a Country Rivers Tissues O A.c.chrysolepis A. c. planiceps OA. c. scypheus sy A. ¢. tandai 3 A. meridionalis Specimens 4 A. bombiceps w A. c brasiliensis @ A.c.chrysolepis A A.c.planiceps @ A c.scypheus % Ac. tandai 3 A. meridionalis Both tissues ana specimens Ww A. ¢. brasiliensis @ A. c.chrysolepis a A. c.tandai — A. bombiceps 80° 70° Figure 1. 50° 40° Distribution of material examined of A. c. chrysolepis, A. c. scypheus, A. c. tandai, A. c. brasiliensis, A. c. planiceps, A. bombiceps and A. meridionalis. Symbols may represent more than one locality. and Anolis meridionalis and another (Clade B) composed of A. c. brasiliensis, A. c. planiceps, A. c. scypheus, and A. bombiceps. Relationships among taxa in clade B were similar across all trees, with an A. bombiceps + A. c. scypheus clade and an A. c. planiceps + A. c. brasiliensis clade that are sister taxa. Relationships within Clade A varied de- pending on the analysis. Parsimony analysis recovered A. c. tandai from Acre (LSUMZ H13599) as the sister taxon of the A. c. tandai + A. c. chrysolepis clade. The ML and Bayesian trees, on the other hand, recovered the Acre A. c. tandai as the sister taxon of A. c. chrysolepis, but with low bootstrap support. The A. c. chrysolepis + A. c. tandai Ae was well supported in all analyses, whereas the A. c. pain +A. c. tandai + A. meridionalis clade received poor nodal Lees Uncorrecte airwise distances among lineages in the x chrysolepis species group ranged from 5.0% Bees A. c. tandai and A. c. chrysolepis to 22.1% between A. scypheus and A. mei ec (Table 2). Both the SH and AU tests (Table 3) found that the alternative hypothesis of a monophy- letic A. chrysolepis, excluding both A. bombi- ceps and A meridionalis, resulted in a eeeeenee worse tree than the ML tree. The ML tree constrained to exclude just A. meridionalis was not significantly worse than our best ML tree. Sama both tests found no. significant difference between a tree constr aining a monophyletic A. c. tandai and our best ML tree. Bayesian Posterior Prob- abilities of alternative hypotheses showed little to no support (e.g., Posterior Probabil- ities < 0.05) fora monophy letic A. chrysolepis excluding A. bombiceps and A. meridionalis, as well as a monophyletic A. c. tandai. The Bayesian Posterior Probability of a monophy- letic A. chrysolepis + A. bombice pS, excluding A. meridionalis, received moderate support. 42 Bulletin of the Museum of Comparative Zoology, Vol. 160, No. 2 A. crassulus A. carolinensis a a A. uniformis A. fuscoauratus A.ortonii A. laeviventris o A. sericeus A. isthmicus A. utilensis A. sagrei A. grahami A. loveridgei A. onca A. annectens A. auratus oe = — — generation C. PE O Posterior Probability = s 0.95 - 0.99 ; ® Posterior Probability = 1.00 2 * = = = 0.08 substitutions/site run1 A. lineatus Figure 2. Results of the partitioned Bayesian analysis. a. Phylogeny of the Anolis chrysolepis species group and outgroups. Bayesian posterior probabilities >0.95 are indicated by circles at nodes. b. Trace plot of post-burnin log likelihood values for the two Bayesian runs. c. Bivariate plot of the split frequencies for the two Bayesian runs. Morphological Analyses The stepwise discriminant analysis ap- plied on body size and all shape variables selected tibia length, interparietal length, and snout-vent length (all size-adjusted) as the most powerful discriminators of A. chrysolepis spp., A. bombiceps, and A. meridionalis, with a classification accuracy of 0.67 based on cross-validation. The first two linear discriminant functions based on these three variables explained about 99% of the total variation, the first function mainly ea eo a contrast between relative tibia aan (—) versus relative SVL (+), and the second function repre- senting primarily the variation in interpari- etal length (Table 4, Fig. 5). Results indi- cate that A. meridionalis and A. c. brasiliensis have short tibias and elongate bodies relative to total body size, whereas A. c. tandai and A. c. chrysolepis have long tibias and short bodies relative to total body size, and A. bombiceps, A. c. planiceps, and A. c. scypheus have intermediate values of these variables. Additionally, A. c. planiceps has the longest, and A. c. chrysolepis the shortest, interparietal relative to its body size. Morphologically, A. c. chrysolepis and A. c. tandai are very similar, whereas A. meridionalis is the most divergent species, ANOLIS CHRYSOLEPIS SPECIES GROUP * D’Angiolella et al. 43 A. carolinensis A. uniformis A. crassulus A. fuscoauratus A. ortonil A. laeviventris A. sericeus A. isthmicus A. sagrel A. utilensis A. loveridgei © Bootstrap = 70% - 89% @ Bootstrap = 90% - 99% ® Bootstrap = 100% 0.09 substitutions/site A. grahami A. auratus A. onca A. annectens A. lineatus Figure 3. Partitioned Maximum Likelihood phylogeny of the Anolis chrysolepis species group and outgroups. Bootstrap values >70% are indicated by circles at nodes. Photo: Anolis brasiliensis from Sao Domingos, Goias, Brazil. Tony Gamble. followed by A. c. brasiliensis. Nevertheless, classification accuracy based on morphology was moderate. The stepwise discriminant analysis applied on meristic counts selected canthals, fourth toe lamellae, and scales between second canthals as the most powerful discriminators of the species, with a classification accuracy of 0.83 based on cross-validation (Fig. 6). The first two linear discriminant functions based on these three variables explained about 93% of the total variation. The first function mainly represented a contrast be- tween canthals and scales between second canthals (—) versus fourth toe lamellae (+), whereas the second function primarily rep- resented the variation in fourth toe lamellae and canthals (Table 5, Fig. 6). Results indicate discrimination 1) in the number of canthals among A. c. planiceps and A. meridionalis (small); A. bombiceps, A. c. tandai, and A. c. chrysolepis (large); and A. c. brasiliensis and A. c. scypheus (intermediate); 2) in the number of fourth toe lamellae among A. c. chrysolepis and A. meridionalis (small); A. c. brasiliensis, A. c. planiceps, and A. c. scypheus (large); and A. bombiceps and A. c. tandai (intermediate); and 3) in the number of scales between second canthals among A. meridionalis (few), A. c. scypheus and A. tandai (large), and the remaining species (intermediate). Overall, A. c. chrysolepis, A. c. tandai, and A. bombiceps are more similar, the same happening with A. c. planiceps, A. brasiliensis, and A. c. scy- pheus. Anolis meridionalis is the most divergent species. Classification accuracy based on meristic counts was relatively good. 44 Bulletin of the Museum of Comparative Zoology, Vol. 160, No. 2 5 MPEG_26568 Faro, PA MPEG_26563 Serra do Acari, PA MPEG_26590 Trombetas, AM ; MPEG26584 Maicuru, PA Anolis c. BPN780 Suriname chrysolepis BPN1587 French Guiana BPN1874 French Guiana C) BPN1979 French Guiana LSUMZ14098 Rio Ituxi, AM A O LSUMZ16398 Rio Solimées, AM : LSUMZ16474 Rio Solimées, AM | Anolis c. VY MPEG25060 Coari, AM é MPEG25029 Juruti, PA tandai MPEG22285 Itaituba, PA LSUMZ13599 Rio Jurua, AC A. meriodionalis 5 BPN1082 Guiana BPN1080 Guiana Anolis c. BPN228 Guiana 2 BPN9Q6 Guiana planiceps LSUMZ12300 Rio Ajarani, RR CHUNB43282 Brasilia, DF CHUNB34542 Novo Progresso, PA oe CHUNB08842 Parauapebas, PA CHUNB45077 Caseara, TO CHUNB45075 Minacu, GO . CHUNB52471 Peixe, TO Anolis c. B GRC16378 Alto Paraiso, GO li i CHUNB37527 Parana, TO brasiliensis CHUNB_37528 Sao Domingos, GO CHUNB.27158 Mateiros, TO CHUNB11521 Palmas, TO LSUMZ12543 Ecuador LSUMZ12989 Ecuador - © KU222147 Peru Anolis c. scypheus 0 LSUMZ12592 Ecuador . bombiceps . auratus . lineatus onca _ annectens . fuscoauratus ortonii . sericeus _isthmicus . laeviventris © Bootstrap = 70% - 89% sagrei . utilensis . grahami — @ Bootstrap = 90% - 99% . loveridge! _ uniformis _ crassulus ® Bootstrap = 100% . carolinensis bb bhDEEPPEPSPEEESESESESESpySE Figure 4. Maximum Parsimony consensus phylogeny of the Anolis chrysolepis species group and outgroups. TABLE 2. NET BETWEEN GROUP DISTANCES FOR ND2 AMONG THE ANOLIS CHRYSOLEPIS GROUP. DISTANCES ABOVE THE DIAGONAL ARE UNCORRECTED P-DISTANCES. DISTANCES BELOW THE DIAGONAL WERE MAXIMUM LIKELIHOOD—CORRECTED USING THE GTR MODEL. AEC: AG: TENT: Teo /a\, (C lean A. chrysolepis tandai planiceps brasiliensis scypheus bombiceps — meridionalis A. c. chrysolepis — 0.050 0.195 On73 0,192 OAS9 0.201 A. c. tandai 0.059 — 0.169 0.149 0.167 0.170 Oa A. c. planiceps 0.274 0.233 _ 0.076 0.128 0.136 0.193 A. c. brasiliensis 0.249 0.209 0.093 _— 0.114 0.121 0.193 A. c. scypheus 0.270 0.229 0.160 0.155 — 0.104 0.221 A. bombiceps 0.255 0.228 0.156 0.154 0.123 — 0.209 A. mridionalis 0.267 0.232 0.279 0.270 0.298 0.267 — ANOLIS CHRYSOLEPIS SPECIES GROUP ¢ D’Angiolella et al. 45 TABLE 3. COMPARISONS OF MAXIMUM LIKELIHOOD (ML) TREE SCORES (—LNL ) AND P VALUES OF THE SH AND AU TESTS BETWEEN OUR BEST ML TREE AND THE CONSTRAINED TREES. BAY ESIAN POSTERIOR PROBABILITIES OF ALTERNATIVE HYPOTHESES ARE ALSO SHOWN. Difference SH AU Bayesian Posterior Hypothesis —In L —In L Test (P) Test (P) Probability Optimal tree — 16,900.3198 n/a n/a n/a n/a Monophyletic A. chrysolepis group — 16,963.9620 — 63.6423 <().001 <(0.000] 0.000 Monophyletic A. chrysolepis group + A. bombiceps — 16,900.9392 —0.6194 0.867 0.57] 0.219 Monophyletic A. c. tandai — 16,903.4524 —3.1326 0.726 0.420 (0.020 DISCUSSION The molecular Be enetic analyses re- covered six species- leve cca as part of the A. chrysolepis species group. These taxa can also be morphologically distinguished on the basis of morphometric and meristic charac- ters. Even though we cannot infer relation- ships among these taxa on the basis of the meristic discriminant analysis, the results of this analysis are consistent with the exis- tence of two clades: one containing A. c. tandai, A. c. chrysolepis, and A. bombiceps and another clade containing A. c. brasi- liensis and A. c. planiceps. Meristic charac- ters in A. c. scypheus appear to be intermediate between these two groups, which is also consistent with it being (together with A. bombiceps) the sister clade to A. c. brasiliensis + A. c. planiceps. Anolis meridionalis was quite distinct from other members of the A. chrysolepis species group on the basis of meristic characters. We define the A. chrysolepis species group as the clade originating with the most recent common ancestor of A. c. chrysolepis and A. brasiliensis. Anolis meridionalis has not historically been allied with the A. chrysolepis species group because of its unique morphology. In particular, A. mer- idionalis differs from other members of the A. chrysolepis species group by having digital dilatations on phalanx ii and iii continuous with scales under phalanx i, instead of forming the prominent border observed in the A. chrysolepis subspecies and A. bombiceps. Although the node leading to the A. chrysolepis species group, including A. meridionalis, was well support- ed in the ML and Bayesian analyses, the presence of A. meridionalis in clade A received poor support in all phylogenetic analyses. For this reason, we could not reject the alternative hypothesis of a mono- phyletic A. chrysolepis group exclusive of A. meridionalis. This means that inclusion of A. meridionalis in the A. chrysolepis species group is still uncertain. Future phylogenetic analyses that include additional A. meridio- TABLE 4. LINEAR DISCRIMINANT ANALYSIS OF THREE MORPHOMETRIC VARIABLES THAT BEST DISTINGUISH THE SPECIES AND SUBSPECIES OF ANOLIS STUDIED. VALUES REPRESENT MEANS OF SCALED, SIZE-ADJUSTED VARIABLES FOR EACH SPECIES AND COEFFICIENTS OF VARIABLES ON FIRST AND SECOND LINEAR DISCRIMINANT FUNCTIONS (LDF 1, LDF Anolis Species Tibia Length A. bombiceps 0.57. A. c. brasiliensis —0.91 A. c. chrysolepis 0.67 A. meridionalis = 167 A. c. planiceps —0.02 A. c. scypheus =(0105 A. c. tandai 0.87 LDF 1 (0.86) sa ry) EDE2Z-(O13) —0.60 2). PROPORTION OF TOTAL VARIATION EXPLAINED BY EACH LDF IN PARENTHESES. Interparietal Length Snout—Vent Length —0.35 0.78 0.46 0.15 = 1.02 0.14 0.58 1.46 O79 =0!59 —0.32 =O —0.60 0.11 0.53 1.05 sd Le | 0.49 46 2- A. c. chrysolepis 4 a —_————————_4 a i O-A c tandai LDF 2 | A © Sn | -2 0 A. bombiceps Bulletin of the Museum of Comparative Zoology, Vol. 160, No. 2 A. meridionalis I, c. brasiliensis ————— A. c. scypheus | A. c. planiceps 2 4 6 LDF 1 Figure 5. Means (open circles) and standard deviations (error bars) of scores on first (LDF 1) and second (LDF 2) linear discriminant functions of tibia length, interparietal length, and snout—vent length (all size-adjusted; see text for details) for seven subspecies and two species of Anolis. nalis samples and data from nuclear loci may help resolve this issue. All described taxa in the molecular analyses formed well-supported, monophy- letic groups, with the exception of A. c. tandai. The A. c. tandai individual from Acre fit the morphological diagnosis we present in this study but was either the sister taxon to A. c. chrysolepis (ML and Bayesian analyses) or the sister taxon to the A. c. chrysolepis A. c. tandai A. bombiceps 4 a 2 - A. c. chrysolepis + Ac. tandammelade (parsimony analysis). The apparent para- phyly of A. c. tandai may be due to several phenomena, none of which are mutually exclusive. One possibility is phylogenetic error due to incomplete taxonomic sampling or lack of data (Graybeal, 1998; Mitchell et al., 2000). It is also possible that individuals from the Acre population represent an as yet undescribed, morphologically cryptic A. meridionalis A. c. planiceps A. c. brasiliensis A. c. scypheus 0 1 2 3 4 EBFal Figure 6. Means (open circles) and standard deviations (error bars) of scores on first (LDF 1) and second (LDF 2) linear discriminant functions of canthals, fourth toe lamellae, and scales between second canthals for seven species of Anolis. ANOLIS CHRYSOLEPIS SPECIES GROUP * D’Angiolella et al. AT TABLE 5. LINEAR DISCRIMINANT ANALYSIS OF THREE MERISTIC COUNTS THAT BEST DISTINGUISH THE SPECIES AND SUBSPECIES OF ANOLIS STUDIED. VALUES REPRESENT MEANS OF SCALED VARIABLES FOR EACH SPECIES AND COEFFICIENTS OF VARIABLES ON FIRST AND SECOND LINEAR DISCRIMINANT FUNCTIONS (LDF 1, LDF 2). Proportion OF TOTAL VARIATION EXPLAINED BY EACH LDF IN PARENTHESES. Scales Between Second Anolis Species Canthals A. bombiceps 1.13 c. brasiliensis —(0.09 A. c. chrysolepis 0.41 A. meridionalis —0.77 A. c. planiceps —1.48 A. c. scypheus 0.09 A. c. tandai 0.96 DE al O274) =—135 DE 2 (Ows)) —0.92 species. Incomplete lineage sorting can also result in discordance between individual gene trees and the species tree because of the retention and/or sorting of ancestral polymorphisms, particularly Salen popula- tions have diverged recently, have a large effective population size, or both (M Madde son, 1997: Ballard and Whitlock, 2004; Maddison and Knowles, 2006). Additional phylogenetic analyses incorporating nuclear genes and additional taxa, as well as using methods that incorporate coalescent pro- cesses and incomplete lineage sorting, would be useful in clarifying relationships among A. c. tandai populations. Our results show broad congruence among molecular and morphological data sets that are consistent with independent evolutionary lineages. Most importantly, each of these time: ages is morphol seically diagnosable. anche distances among. sis- ter taxa in the A. chrysolepis group were also comparable to ND2 distances among SISter Species: In other squamate taxa (Macey et al., 1998, 1999; Glor et al., 2001; Oliver et al., 2009). Therefore, we elevate each subspecies to species status under the general lineage species concept (de Queiroz, 1998, 1999, 2005, 2005a, 2005b, 2007). To facilitate future studies, each species, including A. bombiceps and A. meridonalis, is di: agnosed below and an identification key is provided, considering morphological data collected for this study oO as well as data from the literature. Table 6 Fourth Toe Lamellae Canthals —0.63 — 022 0.72 —().67 =150 0.2] = 1,50 =0D 0.42 —0.56 0.56 0.99 —(.66 0.75 ().92 —0(0.69 = 36 —0.20 compares the main meristic and morpho- metric characters. Taxonomy/Species Accounts All descriptions of color pattern are based on literature. photographs of live animals. and preserved specimens. Anolis chrysolepis Dumeril and Bibron, 1837. Anolis chrysolepis Dumeril and Bibron, 1837:94 (lectotyoe MHNP 2456, type locality: La Mana, French Guiana); Cunha, 1961:60; Peters and Donoso- Barros, 1970:61; Avila Pires et al., 2010:94. Anolis chrysolepis chrysolepis; Vanzolini and Williams, 1970:85; Hoogmoed, 1973:112; Hoogmoed and Avila-Pires, 1989:168. Norops nitens chrysolepis; Savage and Guyer, 1991:366. Anolis nitens chrysolepis; Avila-Pires, 1995:75. Abbreviated Description. Maximum SVL 74 mm. Vertebral region with distinctly enlarged scales, middorsal row largest; number of rows of enlarged scales increases posteriorly. Scales on upper arms smaller than, to subequal to, vertebral scales. Supraorbital semicircles with scarcely en- larged scales. Supraocular scales keeled, slightly larger than or subequal to scales 48 Bulletin of the Museum of Comparative Zoology, Vol. 160, No. 2 TABLE 6. COMPARISONS OF MERISTIC CHARACTERS, BODY PROPORTIONS AND MEASUREMENTS (IN MILLIMETERS) AMONG THE SPECIES OF MEMBERS OF THE ANOLIS CHRYSOLEPIS GROUP. A. A. A. A. A. A. A. Character* chrysolepis tandai scypheus planiceps __ brasiliensis bombiceps _meridionalis No. of specimens 50 96 49 12 116 11 9 max. svl (mm) 74 70 80 76 69 74 56 midbody 101-156 91-150 PO = 174! 110-149 JUPAS olen 79-106 midbody median and = 123.18 + 134.06 + 145.94 + 125 845 129M Ops LES es O29) + standard deviation 1258 7.43 14.59 9.46 10.78 3.08 9.39 values slabials MENS 10-14 10=14 10-14 NO=Ks 10=12 9-11 ilabials 11-15 10-14 10-13 9-11 S12 10=13 811 prostrals A=H| 5-9 57] ai 5=1/ 5=6 5 loreals 5—LO py) (0) 6-9 6-10 6 6=7) canthals 6-10 6-11 6-11 6-8 b=9 9-11 5=6 scales bet 2* canthals 9-14 10=15 9-15 8-14 8-14 9-12 8-10 scales bet 1-4 4 15 i=3 0-3 13 O= i semicirsorbits interp—semicirsorbit 15) oa 225 13 14 9-4 0=2 pmentals 4-8 4-8 4-8 4-8 4-8 6-8 os lam-4tg 7 14-19 LoS 14-20 la=20) i=19 10=12 lam-4toe 18-26 20=8 26-50 19-34 WF 89) 2i1=29 18-24 tail/svl =i AS 1:66=2208 S847) 22 = 2704: 1.69-2.18 1.97-2.38 mouth/svl OF3=025" > ONIS=0724' 0.18—0.24 0.19-0.26 0.18—0.26 0.19-0.21 0.19-0.21 interp/head-w OOT=Os19 008-004 OQ. 11=0°25 0.16—0.32 0: 11=0:30 0.10—0.28 0.16—0.30 tibia/svl 0.20-0.38 0.30—0.43 0.26—0.38 0.27—0.36 0.26—0.34 O:29=0'36 022-027 Interpatietal 0646p" 0'842022 OS rai ().98—2.79 0.92-2.65 i298 126 =2207 head-w 593-105 5.371141 567=-13.08 4.95=10014 “7225-10 7>" “S96-l0io Forsa—or0m head-alt 5.00-8.39 4.33—9.65 AOSV OOS 26-9939 DELO i.O2=01303 4.88-6.36 orbdist 4.01-6.39 4 = ls) 10 4—oroll Solo 4.39-7.26 6.09-6.65 4,02—4.86 eardiam 0402-4 Os7>—1s9 0.40-2.23 OAT—le71 0).87—2.00 0.88—1.68 0).84—1.09 nostrilsdis pS AL NSO) 7 1 1LO=2'95 LO=2276 [A0=2732 2,022.30 Ne) 78 mouth to ear IETS =3103 1.6-2.92 1.09-—3.09 1:06=3.52 0.96—2.72 WAT lis) I ero 1e'5)8) snout 297-6169 3:35—7-00 Deo Nala 2 69=727 3 92—os14 4.67-6.46 4.8—5.82 max. toe IV 20.94 24.42 26.91 YSeel 21.63 2a 14.12 max. foot 25.63 30.24 31.46 31.89 26.48 22.66 17.39 toe IV width Oba 30) Ory 9 0.61-1.48 0.51—-1.46 0.92-—1.65 0.96-1.26 0.61—0.91 * Abbreviations: max. sv! = maximun snout-vent length; midbody = number of scales around midbody; slabials = total number of supralabials; ilabials = total number of infralabials; prostrals = total number of postrostrals; scales bet 2* canthals = number of scales on the snout between the second canthals; scales bet semicirsorbits = minimum number of scales between supraorbital semicircles; interp—semicirsorbit = minimum number of scales between any of the supraorbital semicircles and interparietal; pmentals = number of postmentals; lam-4fg = number of expanded lamellae under the fourth finger; lam-4toe = number of expanded lamellae under fourth toe; tail/svl, mouth/svl, tibia/svl = respectively, the rates of the tail, mouth, and tibia length with the snout—vent length; interp/ head-w = the rate of interparietal width with head width; interparietal = interparietal width; head-w = head width; head-alt = head height; orbdist = minimum distance between orbits; eardiam = ear diameter; nostrilsdis = minimum distance between nostrils; mouth to ear = minimum distance between mouth and ear; snout = from the tip of snout to anterior margin of orbit; max. toe lV = from toe IV base to the heel; max. foot = fourth toe length from toe nail to toe base; toe IV width = fourth toe width. on snout, grading into granules laterally and oaateee Interparietal subequal to or slightly arger than adjacent scales (Fig. 7A, B). Color in Preservative. Color pattern sexually dimorphic. Male dorsal ei pale or grayish-brown with or without a wide light vertebral band bordered by a grayish- brown irregular band laterally. Paired trian- gular spots may be present tee: back; most specimens with paired triangular spots on sacral region. Female dorsal pattern less variable. A thin dark brown line begins at posterior corner of each eye at each side, converging toward neck and continuing along the body, where they delimit a lighter or ee Gore vertebral band that darkens cat expands laterally on tail. Dewlap Color in Life and Preservative. In preservative, male dewlap skin royal blue, dark blue, or blackish, with light scales or blue scales toward rim. Female dewlap cream, similar to surrounding area; scales may be darker at edge. In life, male dewlap skin usually royal blue or blackish blue with ANOLIS CHRYSOLEPIS SPECIES GROUP * D’Angiolella et al. 49 rote Figure 7. Anolis chrysolepis species group. A) Anolis chrysolepis female from Nassau Plateau, Suriname (Photo: Robert Langstroth), B) Anolis chrysolepis male from Faro, Para, Brazil (Photo: Waldima Rocha), C) Anolis tandai female from Rio Jurua, Acre (Photo: Laurie J. Vitt), D) Anolis tandai male from Rio Jurua, Acre, Brazil (Photo: Laurie J. Vitt), E) Anolis tandai female from Rio Ituxi, Amazonas, Brazil (Photo: Laurie J. Vitt), F) Anolis tandai male from Amazonas, Brazil (Photo: Laurie J. Vitt). light scales or blue scales along rim. Avila- Pires (1995) mentioned a cob alt-blue juve- nile male dewlap (RMNH 24673) with white to orange scales, surrounded by a spectrum-orange area that extended through most of ventral surface of head. Female dewlap skin usually yellowish to orange with gray or cream scales; an orange lateral area extended through most of ventral surface of head may aie present. Hoogmoed and Avila-Pires (1991) men- Honed a female from French Guiana with yellow dewlap with orange scales, present- ing a bluish area toward ae rim. ‘Comparison with Other Species from the A. chrysolepis Species Group. This species 50 Bulletin of the Museum of Comparative Zoology, Vol. 160, No. 2 has proportionally the smallest interparietal length among the species of the group Gable 6). It differs from its sister taxon A. ae mainly by a lower number of postrostral scales (4-7 in A. chrysolepis and 5-9 in A. tandai) and by female dewlap color, generally cream in A. chrysolepis; cream with a large central blue spot in A. tandai. Distribution. Southern Guyana, Suri- name, French Guiana and northern Brazil, in the states of Amapa and Para. Anolis tandai Avila-Pires, 1995. Anolis chrysolepis; Vanzolini, Gascon and Pereira, 1993:181. Anolis nitens tandai Avila-Pires, 1995:80 (holotype MPEG 15850, type locality: Rio Urucu, Amazonas state, Brazil); lcochea et al., 2001:140; Vitt et al., 2001:401; Santos-Jr et al., 2007:9; Avila-Pires et al., 2009:116. Abbreviated Description. Maximum SVL 70 mm. Vertebral region with slightly en- larged scales; number of rows of enlarged scales increases posteriad. Scales on upper arms smaller than, or subequal to, vertebral scales. Supraorbital semicircles with scarcely enlarged scales. Supraocular scales weakly to distinctly keeled, approximately same size as scales on snout, laterally and posteriorly grading into granules, anteriorly surrounded by smaller scales. Interparietal moderately small, larger than adjacent scales (Fig. 7C— F). Color in Preservative. Color pattern sexually dimorphic. In males, vertebral region “usually distinct from flanks, with unclear limits between these areas. A pair of subtriangular dark spots present on sacral region. Some specimens may present sinu- ous lines, assuming subtriangular shapes along dorsum. Females usually with a well- delimited vertebral band, similar to Anolis chrysolepis females; occasionally dorsal pat- tern similar to males. Dewlap Color in Life and Preservative. In preservative, male dewlap royal blue or blackish-blue with light scales. Female dewlap with central blue spot surrounded ISSG21c; by a cream area; scales usually light colored. In life, male dewlap skin frequently blue or blackish, with light scales. Avila-Pires (1995) mentioned the dewlap in MPEG 15986 as “ultramarine with cream-color scales on rim.” Dewlap in females, when extended, presents a large and central blue spot, surrounded by a cream area. Scales are frequently cream to orange. When not extended, dewlap presents a light rim and is blue laterally. Avila-Pires (1995) de- scribed the holotype MPEG 15850 female dewlap color as “sulphur-yellow with a large indigo-blue spot.” Comparison with Other Species from the A. chrysolepis Species Group. As already mentioned by Avila-Pires (1995), this spe- cies has the longest tibia in relation to SVL (0.30-0.43). For differences with A. chryso- lenis, see above. Avila-Pires (1995) also mentioned the possible sympatry with A. bombiceps, which also has a blue or blackish blue dewlap (with no sexual dimorphism), but they can be distinguished by female dewlap color (a central blue spot, surround- ed by a pale area in A. tandai), by the minimum number of scales between supra- orbital semicircles (1-4 in A. tandai and 1-2 in A. bombiceps) and by the number of postmentals (4-8 in A. tandai and 6-S in A. bombiceps). Distribution. South of the Amazon River and west of the Tapajos River, in Brazil (states of Para, Amazonas, Rondonia, Acre, and north of Mato Grosso), and in Peru. Anolis planiceps Troschel, 1848. Anolis planiceps Troschel, 1848:649 (holo- type ZMB 529, type locality: Caracas, Venezuela). Anolis chrysolepis planiceps; Vanzolini and Williams, 1970:85; Hoogmoed, 1973: 125; Myers and Donnelly, 2008:100. Anolis chrysolepis; Beebe, 1944:97; O’Shea, 1989:69; Zimmerman and Rodrigues, 1990:449; Martins, 1991:182. Anolis eewi Roze, 1958:311 (holotype FMNH 74040, type locality: Chimanta- tepui, Bolivar, Venezuela). ANOLIS CHRYSOLEPIS SPECIES GROUP * D’Angiolella et al. Figure 8. Anolis chrysolepis species group. A) Anolis planiceps from Roraima, Brazil (Photo: Laurie J. Vitt), B) Anolis planiceps from Guatopo, Venezuela (Photo: Laurie J. Vitt), C) Anolis planiceps from Cuyuni-Mazaruni, Guyana (Photo: Robert Langstroth), D) Anolis scypheus from Ecuador (Photo: Laurie J. Vitt), E) Anolis scypheus from Ecuador (Photo: Laurie J. Vitt), F) Anolis scypheus from Ecuador (Photo: Laurie J. Vitt). Anolis nitens: Boulenger, 1885:91; Beebe, 1944:200. Norops nitens nitens; Savage and Guyer, 1991:366. Anolis nitens nitens; Avila-Pires, 1995:70; Vitt et al., 2008:84. Abbreviated Description. Maximum SVL 76 mm. Double row of enlarged vertebral scales extending from nape to base of tail; few to several rows of weakly keeled scales, increasing in number caudally, forming a gradual transi- tion between double row of enlarged scales and Ol bo granules on flanks. Scales of upper arms markedly larger than vertebral scales. Supraor- bital Pie sisc with enlarged scales, forming pronounced ridge in some specimens. Suprao- cular region with distinct group of enlarged, weakly keeled, scales surrounded by smaller scales. Interparietal distinctly larger than adja- cent scales (Fig. SA-C). Color in Preservative. No sexual dimor- phism in color pattern. Specimens usually have many chevrons along back, with tips directed posteriorly, sometimes forming the posterior border of rhomboid figures. A pair of triangular spots commonly present on sacral region. Myers and Donnelly (2008) Cesena? the color pattern of two adult males and one adult female as “orange with white or grayish white scales in peel rows. scales Se ee gray or blackish gray in distal rows.’ Dewlap Color in Life and Preservative. Dewlap red, fading rapidly in preserved specimens, appearing cream-white, with light scales. A lateral lavender area may be present as mentioned by Avila-Pires ( 1995). In life, dewlap skin orange to sek with grayish to cream scales. Mi ers and Donnelly (2008) found variation in the dew lap of four juveniles, including a female that had “a large bluish black basal spot on the dewlap, which had a bright orange periphery and mostly white AS (only a few dark scales). Comparison with the Other Species from the A. chrysolepis Species Group. This species has the proportionately largest interparietal scale. It differs from its sister taxon A. brasiliensis mainly by dewlap color (red in A. planiceps and blue or grayish/ baakseh blue in A. brasiliensis) and body size (A. planiceps reaches 76 mm, w hereas A. brasiliensis reaches 69 mm). Distribution. Venezuela, Trinidad, Guy- ana, and the states of Roraima and Amazo- nas on the northern part of Brazil. Anolis brasiliensis Vanzolini and Williams, 1970. Anolis chrysolepis; Amaral, 1937:1722. Anolis chrysolepis brasiliensis; Vanzolini and Williams, 1970:85 (holotype MZUSP Bulletin of the Museum of Comparative Zoology, Vol. 160, No. 2 10319, type locality Barra do Tapirapés, Mato Grosso, Brazil); Williams and Van- zolini, 1980:99; Vanzolini, 1981:253, 1986:3; Cunha et al., 1985:23. Norops nitens brasiliensis; Savage and Guyer, 1991:366. Anolis nitens brasiliensis; Avila-Pires, 1995:70; Werneck and Colli, 2006:1987. Abbreviated Description. Maximum SVL 69 mm. Double row of enlarged vertebral scales from nape to base of tail; few to several rows of dorsal scales with weak keels, increasing in number caudally, grad- ually transitioning between double row of enlarged scales end granules on flanks. Beles of upper arms markedly larger than vertebral scales. Scales on snout from moderately keeled to smooth, heteroge- neous in size, with no distinction between anterior and posterior scales. Supraorbital semicircles with enlarged, generally smooth scales. Supraocular region with most scales large and weakly keeled, surrounded by ell scales. Interparietal distinctly larger than adjacent scales (Fig. 9A—D). Color in Preservative. No sexual dimor- phism in color pattern. Dorsal color grayish- brown or pale white, either uniform or not. A light vertebral band may be present. aie narrow with undefined margins or wide; in both cases surrounded by darker area. A pair of triangular spots on sacral region commonly present. may be accom- panied by second pair at the ‘base of tail. Ventral region usually pale-white, may be marbled ah brown spots. Dewlap Color in Life and Preservative. Dewlap blue or grayish-blue, with light or grayish scales. In fic dewlap usually gr Satria s ee or blackish blue, with dark oe varying from light-cream to dark gray. Some specimens Foi Tocantins state show the dewlap skin grayish-green tending to yel- lowish-beige along rim, with scales grayish- brown or pale-cream tending to brownish along rim. Some irregular light-blue lines may be present ( (Fig. SD). Vanzolini and Williams (1970) do not describe the dew lap color re refer to the frontispiece plate ANOLIS CHRYSOLEPIS SPECIES GROUP * D’Angiolella et al. Ul (Te) Figure 9. Anolis chrysolepis species group. A) Anolis brasiliensis from Cantao, Tocantins, Brazil (Photo: Laurie J. Vitt), B) Anolis brasiliensis from Cantao, Tocantins, Brazil (Photo: Laurie J. Vitt), C) Anolis brasiliensis from Barra do Ouro, Tocantins, Brazil (Photo: Itamar Tonial), D) Anolis brasiliensis from Jalapao, Tocantins, Brazil (Photo: Laurie J. Vitt), E) Anolis bombiceps from Peru (Photo: Young Cage), F) Anolis meridionalis from Tocantins, Brazil (Photo: Itamar Tonial). representing the dewlap color in life of a male as green with a brown edge along rim. Avila- Bice (1995) observed in specimens from Carajas, Southern Para, “a blue dewlap, lighter in females, with scales varying from light to dark gray or cream “and the surrounding area may be chrome-orange. Comparison with the Other Species “from the A. chrysolepis Species Group. Anolis brasiliensis, along with A. bombiceps, has the largest toe IV among the other species of the A. chrysolepis group. Anolis brasi- liensis differs from A. planiceps, mainly by dewlap color (red in A. planiceps and blue or grayish/blackish blue in A. brasiliensis) aad body size (A. planiceps reaches 76 mm, whereas A. brasiliensis reaches 69 mm). ei brasiliensis is broadly sympatric with 54 Bulletin of the Museum of Comparative Zoology, Vol. 160, No. 2 A. meridionalis, although they occur in different habitats (see ne meridionalis de- scription below) and they differ mainly by digital dilatations under phalanx ii and iii that form a prominent border in A. brasi- liensis. Distribution. Brazil, in southern Para, Tocantins, Piaui, Maranhao, Ceara, Goias, Mato Grosso, Minas Gerais, S40 Paulo, and Distrito Federal. Two individuals housed in MCZ under the numbers R-60580 and R-60581 are labeled as paratypes of A. c. brasiliensis but have as localities Rio Jurua, Brazil, and Loreto, Peru, respectiv ely. Vanzolini and Williams (1970) did not mention those individuals, which mean that they are not paratypes. Moreover, even though they have typical A. brasiliensis characteristics, it is extremely unlikely that the species occurs in these localities. Given their questionable data, we did not consider these individuals in the morphological analyses or elsewhere in this study. Anolis scypheus Cope, 1864. Anolis chrysolepis; Guichenot, 1855:15. Anolis scypheus Cope, 1864:172 (holotype BM 1946.8.855, type locality: “Caracas” according to Boulenger, 1885, but con- sidered an error by Vanzolini and Wil- liams, 1970:85); Boulenger, 1885:90; Goeldi, 1902:16, 32; Cunha, 1961:67; Peters and Donoso-Barros, 1970:66. Anolis incompertus incompertus Barbour, 1932:99 (holotype MCZ 32309, type locality: Villavicencio, Meta, Colombia). Anolis chrysolepis scypheus: Vanzolini and Williams, 1970:85; Vanzolini, 1986:3. Norops nitens scypheus; Savage and Guyer, 1991:366. Anolis nitens scypheus; Avila-Pires, 1995: 78). Abbreviated Description. Maximum SVL 80 mm. Vertebral scales forming double row of enlarged dorsals along back. Scales of upper arms small relative to other species, larger than vertebral scales. Scales on snout relatively small, with raised surface. Supra- orbital semicircles with enlarged scales, generally forming pronounced ridge. A Saal group of enlarged supraocular scales, grading into granules posteriorly and een ally, aniternenle surrounded by smaller scales. Interparietal distinctly larger than surround- ed scales (Fig. 8D- F). Color in Presemuctine. No sexual dimor- phism in color pattern. Dorsum usually with caudally directed chevrons that may form the posterior border of rhomboid figures, similar to the pattern described for A planiceps; or, it may show a broad band with lateral expansions (narrower at nape, extending caudally). A pair of subtriangular dark spots may be present on sacral region. Dewlap Color in Life and Preservative. Dewlap color in preservative usually pale along rim with pale-cream or blackish scales, and blue with light scales in the center (Fig. 7). In life, dewlap skin red along rim Gic red color vanishes very easily in preserved specimens) with red or black- ish scales, and blue with pale-cream or light- brown scales laterally. Avila-Pires (1995) described the dewlap color of RMNH 24653 as “cobalt-blue with red rim, scales white with orange center.’ i Comparison aa the Other Species from the A. chrysolepis Species Group. Anolis c. scypheus presents the proportionally maxi- mum values of head width, head height, ear diameter, minimum distance hanucen nos- trils and SVL among all species of the A. chrysolepis species group. For differences with A. bombiceps, see A. bombiceps diag- nosis below. Distribution. Amazonian Colombia, Ecu- ador, Peru, and the northwestern part of Amazonas state in Brazil. Anolis bombiceps Cope, 1876. Anolis bombiceps Cope, 1876:168 (type apparently lost, type locality: Nauta, Peru); Goeldi, 1902:16, 32; Peters and Donoso-Barros, 1970:49; Vanzolini and Williams, 1970:86; 1986:28; Avila-Pires, 1995:54. ANOLIS CHRYSOLEPIS SPECIES GROUP ® D’Angiolella et al. Norops bombiceps; Savage and Guyer, 1989:110. Abbreviated De scription. Maximum SVL 74 mm. Vertebral scales not or only slightly enlarged. Scales on upper arm subequal to, or slightly larger than, vertebral scales. eenice on snout anteriorly small, weakly to distinctly keeled, posteriorly larger, flat, usually uni- or multicarinated. Supraorbital semicircles with enlarged, keeled scales, forming pronounced ridge. Supraocular region Syith a group of distncdy enlarged senles Sarrounded posteriorly and inter alls by granules. Interparietal scale distinctly larger than adjacent scales (Fig. 9E). Color in Preservative. No sexual dimor- phism in color pattern. Dorsal color usually brown, with irregular dark spots between hind limbs and irregular figures across limbs. V-shaped lines along back, with apex directed posteriorly, may be present. Dewlap Color in Life and Preservative. Dewlap deep blue or blackish in pre- servative, with light or dark scales. In life, dewlap skin deep blue with light or dark scales. Comparisons with the Other Species from the A. chrysolepis Species Group. Anolis bombiceps is sympatric with A. scypheus and may be sympatric with A. tandai, of which it can be distinguished by dewlap color (deep blue or blackish in A. bombi- ceps, blue on central region and red along rim in A. scypheus, and blue on central region and cream along rim on females of A. tandai; males of A. Rae and A. bombiceps have similar dewlap colors), number of loreals (6-7 in A. bombiceps, 5-9 in A. tandai, and 7-10 in A. scypheus), number of scales between the second canthals (9-12 in A. bombiceps, 10-15 in A. tandai, and 9-15 in A. scypheus), number of scales between supraorbital semicircles (1-2 in A. bombi- ceps, 1-4 in A. tandai, and 1-5 in A. scypheus) and postmentals (6-8 in A. bombiceps and 4-8 in A. tandai and A. scypheus). Distribution. Amazonian Colombia. Ecuador, and Peru and in the State of Ul Ul Amazonas, Brazil. In this study, we reported A. bombiceps from two Brazilian localities, both in Amazonas state: Apui, very close fo the Colombian border a Sao Gabriel da Cachoeira. Anolis meridionalis Boettger, 1885. Anolis meridionalis Boettger, 1885a:437 (holotype lost, original type locality: Para- guay; neotype MNHN Paraguay 6608, type locality: Colonia Ybycui, Estancia Ybycui, Departamento Canindeyu, Para- guay, according to designation by Motte and Cacciali, 2009); Vanzolini and Wil- liams, 1970:8; Peters and Donoso-Bar- ros, 1970:60; Motte and Cacciali, 2009:19; Langsthroth, 2006:154; Nichol- son et al. 2006:2. Anolis holotropis Boulenger, 1895:522 (ho- lotype unknown, type locality: “Province Matto Grosso, Brazil’). Norops sladeniae Boulenger, 1903:69 (ho- lotype unknown, type locality: “Chapada, Matto Grosso” = Chapada dos Guimar- aes, Mato Grosso, Brasil, according to Vanzolini and Williams, 1970). Anolis steinbachi Griffin, 1917:308 (holo- type CM 988, type locality “Provincia del Sara, Bolivia’). Norops marmorata Amaral, 1932:63 (holo- type MZUSP 737, type locality “Jaguara, Rio Grande, Minas Gerais, Brasil’). Anolis chrysolepis meridionalis; Hellmich, 1960:22 (partim, according to Vanzolini and Williams, 1970). Norops meridionalis; Nicholson 2002:97. Abbreviated Description. Maximum SVL 56 mm. Specimens analyzed presented vertebral scales anteriorly small, keeled, posteriorly increasing in size and forming a gradual transition iowned the granules on flanks: they do not form rows OF enlarged scales along dorsum. The neotype description provided ie Motte and Cacciali (2009) states that vertebral scales form eight rows of enlarged scales along back that are distinctly larger than scales on the flanks. Scales on upper arm keeled, markedly larger than Io) Bulletin of the Museum of Comparative Zoology, Vol. 160, No. 2 vertebral scales. Scales on snout anteriorly weakly to distinctly keeled, uni- or multi- carinate and slightly larger than the flat posterior scales. Supraorbital semicircles with enlarged, weakly keeled scales. Suprao- cular region with a group of distinctly enlarged ale surrounded posteriorly and later ally by granules and anteriorly by smaller scales. Interparietal scale distinctly larger than adjacent scales. Loreals 6-7; canthals 5-6; S—10 scales between second canehale. v= 1 scales between supraorbital eonencles O= 2 scales between supraorbital semicircles and interparietal, postmentals 3-8 (Fig. 9F). Color in Preservative. No sexual dimor- phism in color pattern observed. Dorsal color usually grayish-brown, with a light cream, sometimes tending to reddish or light-orange, vertebral band. Dark brown V-shaped lines (apex directed posteriorly) may be present, as well as dark-brown irregular figures and spots across the limbs and in paravertebral region. A pair of subtriangular dark spots on sacral region frequently present. Dewlap Color in Life and Preservative. When extended, dewlap skin in preservative blue or grayish-blue in center and pale or cream along rim, with light or dark scales. When not acadcdl grayish-blue laterally, with the center cream or beige. In life, dewlap skin deep blue in center and orange to sale yellow along rim. Scales may be darker on the center, tending to cream or beige on the border, changing to orange or pale yellow on anterior base of dewlap, along rim (Fig. 8F). Langstroth (2006) presented a dewlap photo of a male A. meridionalis from near the Zapocos Reser- voir in Bolivia showing a deep blue skin with irregular grayish lines and light or dark scales, tending to grayish-green along rim. Comparisons with the Other Species from the A A. chrysolepis Species Group. Anolis meridionalis is sympatric with A. brasilien- sis, of which it can be distinguished mainly by the digital dilatations on ‘phalanx ii and iii, which are continuous with scales under phalanx i and do not form the prominent border observed in A. brasiliensis. Anolis meridionalis can also be distinguished from A. brasiliensis by smaller body size (A. meridionalis reaches 56 mm, whereas A. brasiliensis reaches 69 mm), by the smaller number of scales between supraorbital semicircles (O—1 in A. meridionalis and 0-3 in A. brasiliensis), by the smaller number of scales between supraorbital semicircles and interparietal (1-4 in A. brasiliensis and 0-2 in A. meridionalis), and by the smaller number of fourth finger and toe lamellae (15-20 and 25-32, respectively, in A. brasiliensis, 10-12 and 18-24 in A. mer- idionalis). Besides these morphological dif- ferences, these species do not occur in the same habitat: A. meridonalis is commonly found in open areas densely covered by grass in Brazilian Cerrado (Vanzolini and Williams, 1970), whereas A. brasiliensis is a typical inhabitant of gallery forests in the same biome (Vitt et al., 2008). Vitt et al. (2008: 146) found “only two specimens of A. brasiliensis outside of forested habitat in typical Cerrado, and both were inside termite nests and inactive.” Distribution. Central Brazil, and Bolivia. Paraguay, IDENTIFICATION KEY TO THE SPECIES OF THE ANOLIS CHRYSOLEPIS (GROUP la. Digital dilatations under phalanx ii and iii are continuous with scales under phalanx i, not forming a prominent border; dewlap skin blue or grayish-blue in center and orange to yellowish- -orange (it can be pale or cream in preserved specimens) along rim, with light or dark sca- less FEL bee A. meridionalis 1b. Digital dilatations under phalanx ii form a prominent border over sealesiunderpbalanx | Ye sraee 2 2a. A weakly to distinctly double row of enlarged vertebral scales from nape to base of tail; scales on upper arm larger than vertebral Scales |..2in. i eee ee eee 3 ANOLIS CHRYSOLEPIS SPECIES GROUP ¢ D’Angiolella et al. 2b. Vertebral area with slightly or distinctly enlarged Scales of which the two tenia rOWS may be larger than adjacent ones; scales on upper arm small- er than, to sub equal to, verte- forall S@AleS: cs. 4 Bae dedesscrestescauahen. double row of weakly enlarged vertebral scales: scales on upper arm slightly larger than vertebral scales; dewlap red along rim (it can be pale in some specimens), with red or blackish scales, and blue with light scales in the COIMUCT i seeds ncdoncensce A. scypheus 3b. A double row of distinctly enlar ged vertebral scales from nape to base of tail; scales on upper arm distinctly larger than vertebral BOSS Teens 6h We te. ta atwnsioereAeacws Aa. Supraorbital semicircles with en- Ab. larged, keeled scales, forming a pronounced ridge in some spec- imens; supraocular scales with a group of enlarged scales, sur- rounded in all their extension by distinctly smaller scales; dewlap skin red, but in preserved spec- imens it usually appear as pale, with light scales ........ A. planiceps Supraorbital semicircles with en- larged and generally smooth scales; most supraocular scales large, grading laterally into smallest Seiek: dewlap skin in preserved specimens blue or grayish-blue with light or grayish SCMesici sd. neky..U-b..s. A. brasiliensis 5a. Vertebral area with scales not or only slightly enlarged; interpari- etal scale dlicuaetl larger than adjacent scales; dewlap skin deep blue or blackish-blue with light and grayish scales ...... A. bombiceps 5b. Vertebral area with scales slightly to distinctly enlarged: interpari- etal subequal to or “slightly larger thant adjacentiscales) (2)... /44.14.:..-- Ol 6a. Vertebral area with scales distinctly enlarged along back, numbers of enlarged rows increasing poste- riorly: inte rparietal subequal to or slightly larger than adjacent seals: male ae wlap royal blue, dark blue, or blackish, with light scales or blue scales toward fhe rim; female dewlap skin pale vellow to orange (it can be pale or cream in preserved speci- mens), and toward the rim scales may be darker ........ A. chrysolepis Gb. Vertebral area with scales slightly enlarged along back, ee ee of enlarged rows increasing poste- riorly; interparietal larger than adjacent scales: male dew lap royal blue or blackish-blue, with light scales; female dewlap with a central blue spot, surrounded by a cream area; scales usually FUGUE cern eee eer A. tandai ACKNOWLEDGMENTS We thank M. S. Hoogmoed, T. Mott, D. Mulcahy, I. Sampaio, A. Garda, J. Losos, and A. Aleixo for rey ‘iewing the manuscript: . Losos (Harvard Museum of Comparative Zoology —MCZ), R. Brown (University of Kansas—KU), and R. Brumfield and D. Dittman (Louisiana State University— LSUMZ) for the loan of specimens and tissues; K. Kozak for the free pass to the molecular lab; G. D’Angiolella, who helped with map construction; and M. J. Sturaro for help with figure plates. ABD was financially supported by Conselho Nacional de Desen- volvimento _Cientifico e —Tecnoldgico (CNPq) and received an seh Mayr Travel Grant to visit the MCZ, where J. Rosado was of great assistance. W ork in the herpetological lab of MPEG was supported by PPBIO Amazonia oriental-MCT/M PEG and by CNPq project 473177/2006-4. TG was supported by NIH grant T32DE007288 from the National lnstitate of Dental & Craniofacial Research, TCAP and GRC by research fellowships (respectively, 304199/ 58 Bulletin of the Museum of Comparative Zoology, Vol. 160, No. 2 2010-9 and 302343/88-1) from CNPq. Fieldwork conducted by LJV and J. P. Caldwell was supported by NSF grants DEB 9200779, DEB 9505518, and DEB 0415430; those in Brazil were covered by Portaria MCT 170, de 28/09/94, and by IBAMA permit 073/94-DIFAS. Fieldwork conducted Dy GRC was covered by IBAMA permit 027/2003-CGFAU/LIC and sup- yorted by Programa Nacional de Diversi- dade Biol6gica. PROBIO/MMA, project Paisagens e biodiversidade: uma perspec- tiva integré ada para o inventario e conserva- cao da Serra do Cachimbo.” APPENDIX 1 Specimens examined (number of specimens of each species in parentheses). Anolis chrysolepis (50). SURINAM: Sipaliwini: Tafelberg Nature Reserve (MCZ R-154861), Nassau: Nassau Blatean (MZUSP 29658), Saramacea: 1 km from Coppename River (MCZ R-155217). FRENCH GUIANA: Sophie trail S from St. Elie gold pits (MCZ R-77548), lower Matarony River, tributary of Approa- gue River (MCZ R- 118644, R-118641, R-110388), Saul (MCZ R-146767), Sinnamary, Petit Sant River (MPEG 15829, 15845). GUIANA: Onora (MCZ R- 63350), Shudikar-won (MCZ R-65349). BRAZIL: Amapa: Porto Platon (MCZ 85010), Serra do Navio (MCZ R-79146, R-85009: MPEG 19676, 19678, 19679, 19680, 19681, 19682, 19683, 19685, 19686, 19687, 19688, 19690), Mazagao: Rio Maraca (MZUSP 8845, 8846; CHUNB 56761), Laranjal do Jari (CHUNB 56767; MZUSP 83222, 83224), Para: Aldeia dos indios Tirid (MZUSP 13141), Oriximina: FLONA Saraca- Taquera: Plat6 Aviso (MPEG 27114, 27217), Platé Almeidas (MPEG 26864), Estacaéo Ecoldgica Grao- Para, Serra do Acari (MPE@ 26573) 273674. 27375. 27376, 27377, 26587, 26578, 26593), Porto Trombe- tas: Plato Bela Cruz (MPEG 24216), Plato Greig (MPEG 24218), Faro: Floresta Estadual de Faro (MPEG 26597). Anolis planiceps (72). VENEZUELA: Portu- guesa (MCZ R-176492, R-176491, R-176493, R- 176494, R-176495, R-176496): Acarigua: Environs of Agua Blanca (MCZ_ R-123708), Distrito Federal: Puerto la Cruz (MCZ R-48782), Amazonas: Puerto Ayacucho (MCZ R-58328, R-84073), Sucre: Cunama- coa, San Rafael (MCZ R-48781), Peninsula of Paria, Yacua (MCZ R-43856, R-43860, R-43857, R-43858), Aragua: Palsamencho Parque, HPittier between Por- tochuelo and Ocumare (MCZ R-101817), Carabobo: Hacienda San Esteban, North of Puerto Cabello (MCZ R-107688), Goaiguaza: Miquija (MCZ R-121759, R- IPAs ioe 12 121762. R-121760" Ral20 757). Salone Was Quiguas (MCZ R-121753, R-81302), Faleén: Acosta, Pauji, (MCZ R-49035, R-48723, R-49037, R-48725, R-48724, R-49036), Riecito (MCZ R-49050), Miranda: Guatapo (MCZ R-100438), Bolivar: Arabupu, Mt. Roraima (MCZ R-34868). TRINIDAD: Nariva Swamp, Cascadoux Trace (MCZ R-60800), Huevas Id. (MCZ R-100119), Sangre Grande: Toco (MCZ R- 10746, R-10747), La Seiva (MCZ R-8998, R-8999, R- 9000), Trinidad: Chacachacare (MCZ R-100118), Maruga (MCZ R-31495), Palmiste: 3 km S San Fernando (MCZ R-85011, R-85012, R-85013, R- 85014, R=85015, R=85016, R=85017, R=e85s0i6. Re 100120;,. R=100121., R-100122. R-81303> R=813045 Re 81501, R-85474, R-85475, R-85473). BRAZIL: Ama- zonas: Serra da Neblina (MCZ R-86763), Manaus: Reserva Ducke (MCZ R-92682), Camp Gaviao, approx. 100 km N of Manaus (MCZ R-168987). GUYANA: Pomeroon-Supenaam: Dawa (MCZ_ R-123745), Cuyuni-Mazaruni: Kaburi, 30 miles from Bartica (MCZ R-81306), Mazaruni River (MCZ R-39690), Kamakusa (MCZ R-65352). Anolys scypheus (49). PERU: Loreto: (KU 222147), Rio Pacaya, Cahuana (MCZ R-160775), Galicia: west bank of Rio Tapiche (MCZ R-157245), Pampa Hermosa: near mouth Cushcbatey River (MCZ R-57369, R-57373, R-57372), Amazonas: Mouth of R.Santiago Maranon (MCZ R-57371), Ucayali: River Tamaya (MCZ R-57374), Madre de Dios: Cocha Cashu, NW of mouth of Rio Manu (MCZ B-178177). ECUADOR: Napo: Trail from Laguna Taracoa, S bank Rio Napo, 30 km downriver from Coca (MCZ R- 154571), Hacienda ‘Primavera’ N bank Rio Napo 30 km from Coca (MCZ 92612, 154567, 154568), Prov. S side of Rio Napo, 6.5 km ESE of Puerto Misahualli, at La Cruz Blanca on ASuarez’s land (MCZ R-171889), Sucumbios: Santa Cecilia (MCZ R-92601, R-92602, R-92603, R- 92604, R-92605, R-92608), Santa Cecilia, Rio Agvaratis (MCZ R-100010), Limon Cocha (MCZ R-156822, R-85098, R-85099, R-85100, R-85101, R- 92610, R-92611, R-92612), Chimborazo: Riobamba (MCZ R-29290). COLOMBIA: Meta: Villavicencio: San Martim (MCZ R=32309; R=32310: “R=32312. Re 32313, R-32314, R-32315, R-32317), Bairro Povenir (MZUSP 36099, 36100). BRAZIL: Amazonas: Aca- naui (MZUSP 47213, 47214), Rio Japura: Serrinha (MZUSP 46793, 46794), Costa da Altamira (MZUSP 47301, 47302, 47303); Lago Amana (MZUSP 60462). Anolis tandai (96). BRAZIL: Rondonia: Espigao do Oeste (MPEG 21479, 21483, 21488, 21899), Amazonas: Estirao do Equador (MPEG 901), Rio Urucu (holotype: MPEG 15850), Benjamin Constant (paratypes: MPEG 15933, 15938, 15949, 15986, 15987, 15995: MZUSP 15898), Madeireira Sheffer, Rio Ituxi (MPEG 20372, 20473, 20474), Careiro da Varzea, estrada para Altazes (MPEG 18918, 18920, 18928, 18931, 18934, 18935), Tapaua (MZUSP 37763, 37764, 37765), Beruré (MZUSP 38101, 38102), Novo Ar- ipuana (MZUSP 42400), Borba (MZUSP 41077, 9145), Rio Jurua (MZUSP 700), Barreira do Matupiré, Rio Madeira, (MZUSP 42150), Maués, Braganca/Sao Tomé, Rio Paraconi (MPEG 27673, 27674, 27675, 27676), Acre: Rio Jurua, approx. 30 km North of Porto Walter of Walter (MPEG 20655, 20656, 20658, 20659, ANOLIS CHRYSOLEPIS SPECIES GROUP * D’Angiolella et al. 59 20662: 20663. 20665, MZUSP 53270.. 538271, 53272). Alto Purus (MZUSP 2413, 2514), “Seringal Santo Antonio, proximo a Manoel Urbano (MZUSP 32097), Parque Nacional Serra do Divisor, Estirao do Panela (MZUSP 88656), Estirao do Equador (MZUSP 899), Para: Parque Nacional da Amazonia, Itaituba (MPEG 21986, 21987, 21988, 21989, 21990, 21991), Parque Nacional da Amaz6Onia, Uruad (MZUSP 52533, 52536, 52538, 52539), Cachoeira da Montanha (MZUSP 53657, 53658), Buburé (MZUSP 53692, 53693, 533695), Barreirinha, proximo a Sao tae Rio Tags: (MZUSP 20679), Mato Grosso: Aripuana (MZUSP 82590, 82585, 82589, 82587, 82591, 82593, 82586, 82592, 82588, 82594, 81502, 81512, 81501, 81520, 81505, 81497, 81508, 81514, 81494, 81491), ), Juruena (MZUSP 82402, 82404, 82397, 82403, 82400). PERU: Loreto: Rio Orosa (MZUSP 56658), Alto Curanja, Igarapé Champuia (MZUSP 3324, 3324). Anolis brasiliensis (116). BRAZIL: Mato Grosso: Barra do Tapirapés (holotype: MZUSP 10319, para- types: MCZ R-98284, R-98285, R-98286, R-98287, R- 98288, R-98289), Porto Alegre do Norte (CHUNB 47842) Vila Rica (MZUSP 82874, 82884, 82882), Tocantins: Caseara (CHUNB 44989, 44992, 44499, 45000, 45009, 45036, 45028, 45026, 45022, 45030), Pium (CHUNB 24754, 24756), Palmas (CHUNB 95082: 24216. 11521, 19522 24915. 16947, 25092. 11520, 24262, 15230, 16142, 16948), Peixe (CHUNB D247 1. 52472. b2474). Porto Alegre do Tocantins (CHUNB 38918), Porto Nacional (CHUNB 47741, AT (42, 47740), Mateiros (CHUNB 27161, 27158, 27162), Usina Hidroelétrica Luis Eduardo Magalhaes (MZUSP 95466, 95467), Goids: S4o Domingos (CHUNB 35282, 35257, 35288, 35319, 35285, 35314, 43823, 35315, 35311, 35258, 43824, 43832, 33020), Minacu (CHUNB 11019, 11021, 10967, 10983, 49706, 48617, 10987, 49703, 10986, 10970, 10999, 08655, 29309), Colinas do Sul (CHUNB 50396, 50393, 50395, 44694), Valparaiso (CHUNB 08556, 08555), Luziania (CHUNB 47427, 43408, 40865, 43407, 47426), Trés Ranchos (CHUNB 44740), Goiania (CHUNB 57318), Minas Gerais: Unai (CHUNB 30888, 32873, 32867, 36287), Distrito Federal: Brasilia (CHUNB 49614), Piaui: Ribeiro Golcalves (CHUNB 57028), Maranhao: Carolina (CHUNB 51972. 51973, 51974, 51975), Gancho do Arari (MZUSP 60682), Para: Novo Pro- gresso (CHUNB 34542), Parauapebas (CHUNB 08855, 11132), Canaa dos Carajés (MPEG 25166), Itaituba, Jacareacanga (CHUNB 56398), Ceara: Arajara (MZUSP. 51698, 51699, 51703, 51706), Sao Paulo: Bueno de Andrade (MZUSP 4384), Ibarra (MZUSP 4487), Itapura (MZUSP 551). Anolis meridionalis (9). BRAZIL: Minas Gerais: Paracatus (CHUNB” 26140, 26146), Buritizeiro (CHUNB 44522), Rondoénia: Vilhena (CHUNB 11715), Tocantins: Palmas (CHUNB 12022, 12026), Goias: Luziania (CHUNB 43404, 43405), Distrito Federal: Brasilia (CHUNB 43282). Anolis bombiceps (11). PERU: Loreto (MZUSP Islopy Lali3is 13138), NAS L5ikin Ns of Teniente Lopez, 310 (KU 222145). COLOMBIA: Amazonas: 6 km NW of Leticia (MCZ R-112267). ECUADOR: Pastaza: Conambo (MZUSP 13140). BRAZIL: Ama- zonas: Apui (CHUNB 38340); Sao Gabriel da Cachoeira: Missao Taraqua, Rio Uapés (MPEG 17819), Uatuma (MZUSP 17439), Igarapé Belém, Rio Solimées (MZUSP. R-13026, 13027). Anolis onca (1). VENEZUELA: on Caribean Sea, between Urumaco and Coro R=132780), Anolis lineatus (1). NETHERLANDS ANTIL- LES: Curacao: Cos Cora (MCZ R-83027). Anolis sagrei sagrei (1). JAMAICA: Westmore- land: Savanna La Mar (MCZ R-161754). Anolis auratus (1). COLOMBIA: Pozo Colorado, 11 km W of Santa Maria 104546). Faleén: Rio Seco (MCZ Magdalena: (MCZ R- LITERATURE CITED AmaraL, A. 1932. Estudos sobre lacertilios neotropicais. I. Novos géneros e espécies de lagartos do Brasil. Memorias do Instituto Butantan 7: 51-74. AMARAL, A. 1937. Herpetological collection from Central Brazil. Compte Rendus XIle. Congres International de Zoologique. Lisbon 1935: 1720- RZ. AviLa-PirEs, T. C. S. 1995. Lizards of Brazilian Amazonia (Reptilia: Squamata). Zoologische Ver- handelingen Leiden, 299: 1-706. AviLA Pires, T. C. S., M. S. HooGMoeb, and W. A. Rocua. 2010. Notes on the herpetofauna of northern Para, Brazil: a forgotten part of the Guianan Region, I. Herpetofauna. Boletim Museu Paraense Emilio Goeldi. Ciéncias Naturais 5: 13- 112, Avita Pires, T. C. S., L. J. Vit, S. S. Sartorius, and P. A. Zani. 2009. Squamata (Reptilia) from four sites in southern Amazonia, with a biogeographic analysis of Amazonian lizards. Boletma Museu Paraense Emilio Goeldi. Ciéncias Naturais 4: 99— 118. AvisE, J. C. 2000. Phylogeography: The History and Formation of Species. Cambridge: Harvard Uni- versity Press. AvisE, J. C., D. Waker, and G. C. Jouns. 1998. Speciation durations and Pleistocene: effects on vertebrate phylogeography. Proceedings of the Royal Society of London. B 265: 1707- 1712. BaLLarD, J. W., and M. C. Wuittockx. 2004. The incomplete mati history of mitochondria. Mo- lecular Ecology 13: 729-744. Barsour, T. 1932. New Anoles. Proceedings of the New England Zoological Club 12: 97-102. BarRowCLouGH, G. F., and R. M. Zink. 2009. Funds enough and time: mtDNA, nuDNA and the discovery of divergence. Molecular Ecology 18: 2934-2936. BreeBr, W. 1944 Field notes on the lizards of Kartabo, British Guiana, and Caripito, Venezuela. Part 2. Iguanidae. Zoologica 29: 195-216. 60 Bulletin of the Museum of Comparative Zoology, Boetrcer, O. 1885a. Liste von Reptilien und Batra- chiern aus Paraguay. Zeitschrift ftir Naturwis- senschaften, Halle 58: 213-248. Bortrcer, O. 1885b. Berichtigung der Liste von Reptilien und Amphibien am Paraguay von Dr. O. Boettcher. Zeitschrift fiir Naturwissenschaften, Halle 58: 436-437. Boetrcer, O. 1895. Description of a new Anolis from Brazil. Annals and Magazine of Natural History 15: Dy yey Boetrcer, O. 1903. List of the batrachians and reptiles collected by M.A. Robert at Chapada, Matto Grosso, and presented by Mrs. Percy Sladen to the British Museum. (Percy Sladen Expedition to Central Brazil). Proceedings of the Zoological Society of London 2: 69-70. BouLENGER, G. A. 1885. Catalogue of the Lizards in the British Museum (Natural History). ). Vol. 2. London: Taylor & Francis. BRANDLEY, M. C., A. Scumirz, and T. W. REEDER, 2005. Partitioned Bayesian analysis, partition choice, and the phylogenetic relationships of scincid lizards. Systematic Biology 54: 373-390. Burt | oj fe C. W. CuNNINGHAM, and D. L. Sworrorp. 1993. Partitioning and combining data in phylogenetic analysis. Systematic Zoology 42: 384-397. Burnasy, T. P. 1966. Growth-invariant discriminant functions and generalized distances. Biometrics 22: 96-110. . Core, E. D. 1864. Contributions to the herpetology of tropical America. Proceedings of the Academy of Natural Sciences of Philadelphia 16: 166-181. Corr, E. D. 1876. Report on the reptiles brought by Professor James Orton from the middle and. upper Amazon, and western Peru. Journal of the Acad- emy of Natural Sciences of Philadelphia 8: 159-183. Cunua, O. R. 1961. Lacertilios da Amazonia II. Os lagartos da Amazonia brasileira, com especial referéncia aos representados na colecao do Museu Goeldi. Boletim do Museu Paraense Emilio Goeldi 39>1= 189: Cunna, O. R, F. P. Nascimento, and T. C. S. Avita- Pires. 1985. Os Répteis da area de Carajas, Para, Brasil (Testudines e Squamata). I. Publicacées Avulsas do Museu Paraense Emilio Goeldi 40: 9-92. DE Queiroz, K. 1998. The general lineage concept of species, species criteria, and the process of speciation: a conceptual unification and termino- logical recommendation. In D. J. Howard and S. H. Berlocher (eds.), Endless Forms: Species and Speci- ation. New Yous Oxford University Press. bE Queiroz, K. 1999. The general lineage concept of species and the defining properties Bf the species category, pp. 49-89. In R. A. Wilson (ed.), ), Species: New Interdisciplinary Essays. Cambridge, Massa- chusetts: MIT Press. DE Quetroz, K. 2005a. Ernst Mayr and the modern concept of species. Proceedings of the National Academy of Science 102: 6600- “6607. Vol. 160, No. 2 DE Queiroz, K. 2005b. A unified concept of species and its consequences for the future of taxonomy. Proceedings of the California Academy of Science 18: 196-215. DE QuetRoz, K. 2007. Species concept and species delimitation. Systematic Biology 56: 879-886. Dixon, J. R., and P. Sort. 1986. The reptiles of the upper Amazon Basin, Iquitos Region, Peru. Milwaukee: Milwaukee Public Museum Contribu- tions in Se and Geology 4: 1-58. DumeriL, A. M. C., and G. Bisron. 1837. Erpétologie Général ou Histoire Naturelle compléte des Reptiles. Vol. 4. Paris: Librairie Encyclopédique de Roret. FELSENSTEIN J. 1978. Cases in which parsimony or compatibility methods will be positively mislead- ing. Systematic Zoology 27: 401-410. FELSENSTEIN, J. 1985. Confidence limits on phylogenies, an approach using the bootstrap. Evolution 39: ooo Fenwick, A. M., R. L. GuTBeR_et, TB Jr, J. A. Evans, and C. L. Parkinson. 2009. Morphological and molecular evidence for phylogeny and classification of South American pitvipers, genera Bothrops, Bothriopsis, and Bothrocophias (Serpentes: Viper- idae). Zoological Journal of the Linnean Society 156: 617-640. Forses, E. 1846. On the connection between the distribution of the existing fauna and flora of the British Isles, and the geological changes which have affected their area, especially during the epoch of the northern drift. Memoirs of Geological Survey, Great Britain 1: 336-432. GamMBLE, T., A. M. Simons, G. R. Cou, and L. [VERE 2008. Tertiary climate change and the diversifica- tion of the Amazonian gecko genus Gonatodes (Sphaerodactylidae, Squamata). Molecular Phylo- genetics and Evolution 46: 269-277 Gascon, C., and O. S. PEREeira. 1993. Preliminary checklist of the herpetofauna of the upper Rio Urucu, Amazonas, Brazil. Revista Brasileira de Zoologia 10: 179-183. Gror.. R. E. L. J., Vire and A: Lassons 00s molecular phylogenetic analysis of diversification in Amazonian Anolis lizards. Molecular Ecology 10: 2661—266S. Goetpr, E. A. 1902. Lagartos do Brasil. Boletim do Museu Paraense 3: 499-560. Grau, E. T., S. L. Perea, L. F. Sicverra, .. HOrmne: and A. WayntaL. 2005. Molecular phylogenetics and biogeography of Neotropical piping guans (Aves: Galliformes): Pipile Bonaparte, 1856 is synonym of Aburria Reichenbach, 1853. Molecular Phylogenetics and Evolution 35: 637-645. GrayBEAL, A. 1998. Is it better to add taxa or characters to a difficult phylogenetic problem? Systematic Biology 47. 9-17. GriFFIN, L. E. 1917. A list of the South American lizards in the Carnegie Museum, with descriptions of four new species. Annals of the Carnegie Museum 2: 304-320. ANOLIS CHRYSOLEPIS SPECIES GROUP * D’Angiolella et al. 6] GuicHENot, A. 1855. Reptiles, pp. 1-95. In F. de Castelnau (ed.), Animaux nouveaux ou rares recueil- lis pendant Vexpédition dans les parties centrales de l'Amérique du Sud, de Rio de Janeiro a Lima, et de Lima au Para; exécutée par ordre du gouvernement francais pendant les années 1843 a 1847, sous la direction du Comte Francis de Castelnau. Paris: Chez P. Bertrand, Libraire Editeur. HarFrFer, J. 1969. ae in Amazonian forest birds. Science. 165: 131-13 Hetumicu, W. 1960. Die aa der Grand Chaco und seiner Randgebiete. Abhandlungen Bayerische Akademie der Wisse nschaften (1 LF) , 101: 1-131. Hoocmoep, M. S. 1973. Notes on the Herpe tofauna of Surinam IV. The Lizards and Amphisbaenians of Surinam. Biogeographica, 4. The Hague: Dr. W. Junk Publishers. HoocMoep, M. S., and T. C. S. Avita-Pires. 1989. Observations on the nocturnal activity of lizards on a marshy area in Serra do Navio, Brazil. Tropical Zoology 2: ee HoocMmoen, M. S., oad T. C. S. Avita-Prres. 1991. Annotated eae of the herpetofauna of Petit Saut, Sinnamary River, French Guiana. Zoolo- gische Verhandelingen Leiden 65: 53-58. Hupson, R. R., and M. Ture.ii. 2003. Stochasticity overrules the “three-times rule”: genetic drift, genetic draft, and coalescence times for nuclear loci versus mitochondrial. Evolution 57: 182-190. HUELSENBECK, J. P., and F. Ronguist. 2001. MRBAYES: Bayesian inference of phylogeny. Bioinformatics 17: 754-755. HUELSENBECK, J. P., F. Ronquist, R. NiExson, and J. P. Bo.LBack. 2001. Bayesian inference of phylogeny and its impact on evolutionary biology. Science 294: 2310-2314. IcocHEA, J., E. Quispitupac, A. PorTILLA, and E. Ponce. 2001. The amphibian and reptile fauna of the lower Urubamba region, Peru. In F. Dallmeier, A. Alonso, and P. Campbell (eds.), Biodiversity of the Lower Ghana Region, Peru. SIMAB series 7. Washington, DC: Smithsonian Institution/MAB Biodivetsity Program. ICZN. 2010. OPINION chrysolepis Duméril & Bibron, Squamata): precedence given over Draconura nitens Wagler, 1830. Bulletin of Zoological Nomen- clature 67: 266-268. Joticozur, P. 1963. Multivariate generalization of allometry equation. Biometrics 19: 497-499. Jukes, T. H. 1987. Transitions, transversions, and the molecular evolutionary clock. Journal of Molecular Evolution 26: S7-98. Kumar, S., M. Net, J. Duptey, and K. Tamura. 2008. MEGA: a biologist- centric software for evolution- ary analysis of DNA and protein sequences. Briefings in Bioinformatics 9: 299-306. Lancstrotu, R. 2006. Notas sobre Anolis meridionalis Boettger, 1885 (Squamata: Iguania: Polychrotidae) en Bolivia y comentarios sobre Anolis steinbachi. Kempffiana 2: 154-172. 2256 (Case 3446) Anolis 1837 (Reptilia, LARTILLOT, N., H. BRINKMANN, and H. Puiuipre. 2007. Suppression of long-branch attraction artifacts in the animal phylogeny using a site-heterogeneous model. BMC Evolutionary Biology 7: 1-14. Lemmon, A. R., and E. C. Moriarty. 2004. The importance of proper model assumption in Bayes- ian phylogenetics. Systematic Biology 53: 265-277. Macey, J. R., J. A. SCHULTE, N. B. ANANJEvA, A. LARSON, N. Rastecar-Pouyant, S. M. SHamMakoy, and T. J. Parenruss. 1998. Phylogenetic relationships among agamid lizards of the Laudakia caucasia species group: testing hypotheses of biogeographic frag- mentation and an area cladogram for the Iranian Plateau. Molecular Phylogenetics and Evolution LO: t16—f31. Macey, J. R., Y. Wanc, N. B. Ananyeva, A. Larson, and T. J. Papenruss. 1999. Vicariant patterns of fragmentation among gekkonid lizards of the genus Teratoscincus produced by the Indian pollesion: a molecular phylogenetic perspective and an area cladogram for Central Asia. Molecular Phyloge- netics nd Evolution 12: 320-332. Mappison, W. P. 1997. Gene trees in species trees. Systematic es 46: 523-536. Mappison, W. P., and L. L. KNowtes. 2006. Inferring phylogeny despite incomplete lineage sorting. Systematic Biology 55: 21-30. Mappison, W. P., and D. R. Mappison. 1992. MacClade Version 3: Analysis of Phylogeny and Character Evolution. Sunderland Massachusetts: Sinauer Associates. Martins, M. 1991. The lizards of Balbina, central Amazonia, Brazil: a qualitative analysis of resource utilization. Studies on Neotropical Fauna and Environment 26: 179-190. MITCHELL, A., C. Mirrer, and J. C. RecrEer. 2000. More taxa or more characters revisited: combining data from nuclear protein-encoding genes for phyloge- netic analyses of Noctuonler. (Insecta: Lepidop- tera). Systematic Biology 49: 202-224. Motrte, M. and P. Cacctaur. 2009. Descripcion de un neotipo para Anolis meridionalis Boettger, 1885 (Sauria: Polychrotidae). Cuadernos de Herpetolo- gta 23: 19-24. Myers, C. W. 2008. Anolis chrysolepis Duméril & Bibron, 1837 (Reptilia, Squamata): proposed precedence over Draconura nitens Wagler, 1830. Bulletin of Zoological Nomenclature 65; 205-213. Myers, C. W., and M. A. DonnELLY. 2008. The summit herpetofauna of Auyantepui, Venezuela: report from the Robert G. Goelet American Museum— Terramar Expedition. Bulletin of the American Museum of Natural History 308. Ne, M., and W. H. Li. 1979. Mathematical model for studying genetic variation in terms of restriction endonucleases. Proceedings of National Academy of Sciences 76: 5269-5273. NIcHOLson, K. E. 2002. Phylogenetic analyzes and a test of the current infrageneric classification of Norops (Beta Anolis). Herpetological Monographs 16: 93- 120. 62 Bulletin of the Museum of Comparative Zoology, Vol. 160, No. NiIcHOLsoN, K. E., R. E. Gior, J. J. KoLBe, A. Larson, S. B. Hepces, and J. B. Losos. 2005. Mainland colonization by island lizards. Journal of Biogeog- raphy 32: 929-938. Nicuoison, K. E., L. J] Koxse, and J. B. Losos. 2007. Evolution of Anolis lizard dewlap diversity. PLoS ONE 2: e274. NicHouson, K. E., A. Myares-Urrutia, and A. Larson. 2006. Molecular phylogenetics of the Anolis onca series: a case history in retrograde evolution revisited. Journal of Experimental Zoology (Molec- ular and Developmental Evolution) 306: 450- 459. NYLANDER, J. A. A., F. Ronguist, J. P. HUELSENBECK, and J. L. Nieves-ALprey. 2004. Bayesian phyloge- netic analysis of combined data. Systematic Biology 53: 47-67. NYLANDER, J. A. A., J. C. WiLGeNBuscH, D. L. WaRREN, and D. L. Sworrorp. 2008. AWTY (are we there yet?): a system for graphical exploration of MCMC convergence in Bayesian phylogenetics. Bioinfor- matics 24: 581-583. O.tiveER, P. M., M. Apams, M. S. LEE, M. N. HutcuHinson, and P. Doucuty. 2009. Cryptic diversity in vertebrates: molecular data double estimates of species diversity in a radiation of Australian lizards (Diplodactylus, Gekkota). Pro- ceedings of Biological Sciences 276: 2001-2007. O'SHEA, M. 1989. The herpetofauna of Ilha de Maraca State of Roraima, northern Brazil. Proceedings of the 1988 U.K. Herpetological Societies, Symposium on Captive Breeding Reptiles: 51-72. Peters, J. A., and R. Donoso-Barros. 1970. Catalogue of the Neotropical squamata. Part II. Lizards and amphisbaenians. United States National Museum Bulletin 297: 1-293. Posapa, D. 2008. jModelTest: phylogenetic model averaging. Molecular Biology and Evolution 25: 1253-1256: Quinn, G. P., and M. J. Keoucu, 2002. Experimental Design and Data Analysis for Biologists. Cam- bridge. UK: Cambridge University Press. R DEVELOPMENT Core TEAM. 2009. R: A Language and Environment for Statistical Computing. Vienna, Austria: R Foundation for Statistical Computing. RampavT, A., and A. J. DrumMonp. 2009. Tracer v1.5. Available from http://beast.bio.ed.ac.uk. RODRIGUEZ-ROBLES, J. A., T. JEskova, and M. A. Garcia. 2007. Evolutionary relationships and_ historical biogeography of Anolis desechensis and Anolis monensis, two lizards endemic to small islands in the Caribbean Sea. Journal of Biogeography 34: 1546-1558. Rou_r, F. J., and F. L. Booxstern. 1987. A comment on shearing as a method for “size correction.” Systematic Zoology 36: 356-367. Roze, J. A. 1958. Los reptiles del Chimanta Tepui (Estado Bolivar, Venezuela) colectados por la expedicion botanica del Chicago Natural History Museum. Acta Biologica Venezuelica 2: 299- 314. Rusn, D. B. 1996. Multiple imputation after 18+ years. Journal of the American Statistical Association 91: A73-AS9. SanTos-JR, A. P., J. G. Frora, and F. R. V. Riperro. 2007. Reptilia, Squamata, Polychrotidae, Anolis nitens tandai: distribution extension, new State record, and geographic distribution map. Check List 3: 9-10. SavacE, J. M., and C. Guyer. 1989. Infrageneric classification and species composition of the anole genera, Anolis, Ctenotus, Dactyloa, Norops and Semiurus (Sauria: Iguanidae). Amphibia-Reptilia 10: 105-116. SavaGE, J. M., and C. Guyer. 1991. Nomenclatural notes on Peale Sauria: Polychridae): stability over priority. Journal of Herpetology 25: 365-366. ScHaaD, E. W., and S. Por. 2010. Patterns of ecomorphological convergence among mainland and island Anolis lizards. Biological Journal of the Linnean Society 101: 852-859. SHimoparra, H. 2001. CONSEL: for assessing the confidence of phylogenetic tree selection. Bioinfor- matics 17: 1246-1247. Suimopaira, H.. 2002. An approximately unbiased test of phylogenetic tree selection. Systematic Biology 51: 492-508. SHimopaira, H., and M. Hasecawa. 1999. Multiple comparisons of log-likelihoods with applications to phylogenetic inference. Molecular Biology and Evolution 16: 1114-1116. Somers, K. M. 1986. Multivariate allometry and removal of size with principal components analysis. Systematic Zoology 35: 359-368. STAMATAKIS, A. 2006. RAxML-VI-HPC: maximum likelihood-based phylogenetic analyses with thou- sands of taxa and mixed models. Bioinformatics 22: 268S—2690. SworrorD, D. L. 2001. PAUP. Phylogenetic Analysis Using Parsimony (and Other Methods). Version 4.0. Sunderland, Massachusetts: Sinauer Associates. TABACHNICK, B. G., and L. S. Fimexy. 2001. Using Multivariate Statistics. Needham Heights, Massa- chusetts: Allyn and Bacon. TroscHEL, F. H. 1848. Amphibien. In R. Schomburgk (ed.), Reisen in British Guiana. In Den Jahren. Im Auftrag Sr. Majestat des Konigs von Preussen ausgefuhrt von Richard Schomburgk. 3: 1840- 1844. Versucheiner fauna und flora von British Guiana. VAN Buuren, S., J. P. L. Branp, C. G. M. Grooruuts- OupsHoorn. and D. B. Rusin. 2006. Fully conditional specification in multivariate imputa- tion. Journal of Statistical Computation and Simulation 76: 1049-1064. VANZOLINI, P. E., and E. E. Wituiams. 1970. South American anoles: the geographic differentiation and evolution of the Anolis chrysolepis species group (Sauria: Iguanidae). Part 1 and part 2. Arquivos de Zoologia (Sao Paulo) 19: 1-124. VANZOLINI, P. E. 1981. The vanishing refuge: a mechanism for ecogeographic speciation. Papéis Avulsos de Zoologia (Séo Paulo) 34: 251-255. ANOLIS CHRYSOLEPIS SPECIES GROUP * D’Angiolella et al. 63 VANZOLINI, P. E. 1986. Levantamento herpetologico da area do estado de Rondénia sob a influéncia da rodovia BR 364. Programa Polonoroeste, subpro- grama Ecologia Animal, Relatorio de Pesquisa no. Ll. Brasilia: Ministério de Ciéncia e Técnologia— CNPq, 1. J., W. E. Macnussum, T. C. S. Avita-Pires, and A P. Lima, 2008. Guide to the Lizards of Reserva Adolpho Ducke, Central Amazonia. Manaus: At- tena Design Editorial. Witt, a j.s S. SaRTORIUS, T. C. S. AvILA-Pires, and M. C. Esposito. 2001. Tite on the leaf litter: the ecology of Anolis nitens tandai in the brazilian Amazon. Copeia 2001: 401-412. Witt Ek; js Ds B.. Surparp, G. H.C. Vinins, J. P. CALDWELL, G. R. Cott, and D. O. Mesouita. 2008. Ecology of Anolis nitens brasiliensis in cerrado oodlands of Cantao. Copeia 2008: 144-153. Virt, L. J., and P. A. Zani. 1996.-Ecology of the South American Lizard Norops chrysolepis (Polychroti- dae). Copeia 1996: 56-68. WAGLER, J. G. 1830. Nattirliches System der Amphibien, mit vorangehender Classification der Sdugthiere und Vogel. Munich: ].G. Cotta. WERNECK, F. P., and. G. R. Cour. 2006. The lizard assemblage from seasonally dry tropical forest enclaves in the Cerrado biome and its association VITT, with the Pleistocenic Arc. Journal of Biogeography 33: 1983-1992. WIENS, J. J., and B. HoOLLiIncswortH. 2000. War of the iguanas: ee molecular and morphology phylogenies and long- ‘branch attraction in iguanid lizards. Systematic Biology 49: 143-149. Wiens, J. J., and T. A. Penkror. 2002. Delimiting species using DNA and morphological variation and accordant species limits in spiny lizards (Sceloporus). Systematic Biology 51: 69-91. WILuAMs, E. E., and P. E. VANzouini. 1980. Notes and biogeographic comments on anoles from Brasil. Papéis Avulsos de Zoologia, Sdo Paulo 34: 99- 108. Yoper, A. D., R. Vitcatys, and M. Ruvoto. 1996. Molecular evolutionary dynamics of cytochrome b in strepsirrhine primates: the phylogenetic signif- icance of third-position transversions. Molecular Biology and Evolution 13: 1339-1350. ZIMMERMAN, B. L., and M. T. Ropricues. 1990. Frogs, snakes, and lizards of the INPA-WWFE reserves near Manaus, Brazil, pp. 426-454. In A. H. Gentry (ed.), Four Neotropical Rainforests. London: Yale University Press. ZINK, R. M., and G. F. BarrowcLouGcH. 2008. Mito- cligndial DNA under siege in avian phylogeogra- phy. Molecular Ecology 17. 2107-2701. ee US ISSN 0027-4100 MCZ Publications Museum of Comparative Zoology Harvard University 26 Oxford Street Cambridge, MA 02138 mczpublications@mcz.harvard.edu © The President and Fellows of Harvard College 2011. Volume 160, Number 3 15 December 2011 The power and utility of morphological characters in systematics: A fully resolved phylogeny of Xenosaurus and its fossil relatives (Squamata: Anguimorpha) B.-A.S. BHULLAR HARVARD UNIVERSITY | CAMBRIDGE, MASSACHUSETTS, U.S.A. BULLETIN OF THE Museum of Comparative Zoology BOARD OF EDITORS Editor: Jonathan Losos Managing Editor: Adam Baldinger Editorial Assistant: Samantha Edelheit Associate Editors: Andrew Biewener, Scott Edwards, Brian Farrell, Farish Jenkins, George Lauder, Gonzalo Giribet, Hopi Hoekstra, Jim Hanken, Jim McCarthy, Naomi Pierce, and Robert Woollacott Publications Issued or Distributed by the Museum of Comparative Zoology Harvard University Bulletin 1863- Breviora 1952- Memoirs 1865-1938 Johnsonia, Department of Mollusks, 1941-1974 Occasional Papers on Mollusks, 1945- General queries, questions about author guidelines, or permissions for MCZ Publications should be directed to the editorial assistant: MCZ Publications Museum of Comparative Zoology Harvard University 26 Oxford Street Cambridge, MA 02138 meczpublications@mcz.harvard.edu EXCHANGES AND REPRINTS All of our publications are offered for free on our website: | http://www.mcz.harvard.edu/Publications/index.html | To purchase individual reprints or to join our exchange program, | please contact Susan DeSanctis at the Ernst Mayr Library: mayrlib@oeb.harvard.edu. This publication has been printed on acid-free permanent paper stock. © The President and Fellows of Harvard College 2011. THE POWER AND UTILITY OF MORPHOLOGICAL CHARACTERS IN SYSTEMATICS: A FULLY RESOLVED PHYLOGENY OF XENOSAURUS AND ITS FOSSIL RELATIVES (SQUAMATA: ANGUIMORPHA) B.-A. S. BHULLAR' ABSTRACT. Xenosaurus is an enigmatic Mexican and Central American lizards distinguished by knob-like scalation and flattening of the head and body associated with living in epicks. within cliff faces. The position of Xenosaurus within the larger clade Anguimorpha is difficult to determine owing to a combination of primitive features and a unique, highly modified anatomy that obscures useful characters. Evidently, the phylogenetic stem of Xenosaurus represents a long independent history of evolution. Fortunately, seve1 mal fossil taxa exist that can elucidate this history. These taxa include the extinct Exostinis lancensis (Cretaceous), Exostinus serratus (Oligocene), and Restes rugosus (Paleocene), the latter two known from substantial, cranial material (Bhullar, 2007; 2010). Using osteological and alcohol-preserved specimens, fascile. and high- resolution x-ray CT scans thereof, I attempted to reconstruct the relationships of the three fossil taxa and the six extant species of Xenosaurus that - are available in U.S. collections. Despite the considerable phylogenetic importance of Xenosaurus and its stem, this is the first phylogenetic analysis of the group. An exhaustive search of the skeleton, including osteoderms embedded in the skin visualized using CT scanning, yielded 274 new characters, substantially more than have been used previously in gross anatomy—based analyses of such a restricted group of reptiles. The great number of characters is largely the result of the availability of disarticulated skeletal material, CT scans showing internal bone structure and bones embedded in the skin, and attention to subtle anatomical differences whose validity could be assessed in terms of intraspe- cies variation because of the av ailability of large sample sizes for certain taxa. My results suggested that R. rugosus is sister to the other xenosaurs, resolving a polytomy with other Anguimorpha recovered by previous work. Exostinus ' Department of Geological Sciences, Jackson School of Geosciences, The University of Texas at Austin, Austin, Texas; and Department of Organisinic and Evolutionary Biology, Harvard University, 16 Divinity Avenue, Bio Labs Room 4110, Cambridge, Massachu- setts 02138 (bhartanjan.bhullar@gmail. com). Bull. Mus. Comp. Zool., clade of lancensis is problematic in that it may represent several distinct taxa, but it was recovered as sister to E. serratus + Xenosaurus, making Exostinus paraphyletic. Exostinus serratus emerged as sister to Xenosaurus. Xenosaurus comprises a northern clade consisting of Xenosaurus newmanorum and Xenosaurus platyceps: the remaining taxa are united as a southern clade. Within the southern clade, Xenosaurus agrenon and Xenosaurus rectocollaris are sister to Xenosaurus grandis and Xenosaurus rackhami. North—south splits within Xenosauridae mirror those of several other lizard clades and may be the legacy of the equatorial contraction of early Tertiary tropical forests. The fully resolved nature of the phylogeny and the congruence of the extant portion with molecular results mmdiaates the continued relevance and efficacy of morphological systematics when an exhaustive anatomical analysis is perfor med to search for new characters. Key words: Xenosaurs, Xenosauridae, squamates, lizards, extinct, paleontology, Anguimorpha, systemat- ics, Shinisaurus INTRODUCTION Phylogenetic analysis using gross mor- phologic: al characters, while once the stan- dard approach to systematics, has in the last two decades been complemented by the huge number of characters (albeit with a very limited number of character states) available from the morphology of nucleic acids, or “molecular” data. Ancient fossil taxa, of course, can generally be included in phylogenetic analyses only when morpho- logical data are in play, but for extant organisms the huge number of individual ohiaroiee avaiable from DNA sequences compared with the characters available from gross morphology has led to suggestions that morphological analyses are woefully i inaccu- o rate and obsolete (Scotland et al. 2003), 160(3): 65-181, December, 2011 65 66 Bulletin of the Museum of Comparative Zoology, Vol. 160, No. 3 Indeed, early morphology-based analyses in particular often ee only a few dozen characters—but still their resolving power could be considerable (Gauthier et. al., 1988). However, far longer character lists and far better resolution can be achieved with the detailed consideration of the near- infinite aspects of organismal morphology at all scales. Here, I employ an exhaustive search for the characters most relevant to fossil taxa—those derived from skeletal anatomy—in an attempt to resolve the phylogeny of the enigmatic clade Xeno- saurus and its fossil relatives. Xenosaurus (“knob-scaled lizards”) con- sists of unusual flat-bodied crevice-dwelling lizards distributed throughout Mexico and northern Central America. The relation- ships of the species within Xenosaurus and extinct taxa with affinities to the clade have not been treated in a published phylogenet- ic analysis. Two studies, by Gauthier (1982) and Conrad (2008: see also Conrad et al., 2011), presented hypotheses of relation- ships along the phylogenetic stem of Xeno- saurus, although both dealt primarily with relationships within the larger clade Angu- imorpha as a whole. In the first of these studies, Gauthier (1982) suggested that the Oligocene fossil taxon Exostinus serratus is sister to Xenosaurus to the exclusion of the Cretaceous Exostinus lancensis and the Paleocene Restes rugosus (Exostinus rugo- sus prior to reassignment in that work) but left the relationships of the latter two taxa unresolved. These taxa are described fully under Materials and Methods. A preferred hypothesis in which R. rugosus is sister to the remaining taxa and E. lancensis is sister to E. serratus was provided by Conrad (2008). Finally, an unpublished master’s thesis (Canseco Marquez, 2005) included a hypothesis of relationships among the extant species of Xenosaurus based on squamation. That study included more species of Xenosaurus than were available to me, but for those which were included by Canseco Marquez (2005) and by me in this study, the phylogenetic hypotheses recov- ered for relationships within Xenosaurus is identical. I did not include the characters from that study here in deference to the author of the work, who is preparing it for publication. Hypotheses of the phylogeny of Angui- morpha as a whole were provided by McDowell and Bogert (1954) and Gauthier (1982). Those hypotheses were not explicitly tested by the authors, although Gauthier (1982) presented his hypothesis with an explicitly cladistic frame of reference. An- guimorpha was subjected to explicit phylo- genetic analysis by several authors, some of whom used gross anatomical characters (Rieppel, 1980; Estes et al., 1988; Mee: 1998: Evans and Barbadillo, 1998: Gao and Norell, 1998: Conrad, 2005, 2008: Conrad et al., 2011) and others of whom used nucleic acid structure (Macey et al., 1999; Wiens and Slingluff, 2001; Townsend et al., 2004; Conrad et al., 2011). In a cladistic framework, the phylogeny posited by McDowell and Bogert (1954) has an initial split between Varanoidea (here Platynota) and Diploglossa, the latter of which includes the remainder of Anguimor- pha. The arrangement suggested that Var- anoidea is split between Heloderma and Lanthanotus ieee + Varanus, whereas Diploglossa is split between Diploglossinae and a trichotomy of Gerrhonotinae + Anguinae + Xenosauridae. Xenosauridae was used in the sense of Shinisaurus crocodilurus + Xenosaurus. A similar phy- logeny was presented by Rieppel (1980), with Varanoidea sister to a clade whose first split is between Gerrhonotinae and all other taxa, with the latter then split into Xeno- sauridae and Diploglossinae + Anguinae. Xenosauridae was not nested within Angu- idae in the topology presented by Gauthier (1982). Instead, he posited a trichotomy of Varanoidea, Xenosauridae, and Anguidae, the last consisting of Anguinae and Ger- rhonotinae + Di Aocloaane The trichoto- my was ee, by Estes et al. (1988) to Anguidae on one branch and Xenosauridae + Varanoidea on the other. The same topology was recovered by Lee (1998). According to Evans and Barbadillo (1998), the initial split is between Xenosauridae and the remaining anguimorphs (in addition, Gekkota is nested within Anguimorpha, an unusual result that was not recovered other studies). This topology, without the nested Gekkota, was also recovered by Gao and Norell (1998), as was the placement of the Mongolian taxon Carusia intermedia (Borsuk- Bialynicka, 1984) as the sister to Xenosauridae. F inally, a similar topology was recovered by Conrad (2005, 2008), save that S. crocodilurus and its extinct relative Bahndwivici ammoskius are sister to Var- anoidea, leaving the initial split in Angui- morpha Bentecn Xenosaurus plus its al hed extinct taxa and the remainder of Angui- morpha. Carusia intermedia emerged in the 2005 study as the sister taxon to Anguimor- pha, and in the 2008 and 2011 studies as the sister to Xenosaurus and its extinct relatives. Within Anguidae, Anguinae and Diploglos- sinae are sister taxa to the exclusion of Gerrhonotinae. Among the studies based on nucleic acid structure, the same data set was used by Macey et al. (1999) and Wiens and Slingluff (2001). Both groups recovered the same overall topology for Anguimorpha. The initial split according to hose studies is a trichotomy among Varanus, Heloderma, and the remainder oF Anguimorpha. The latter clade is split between S. crocodilurus and Xenosaurus + Anguidae. Within Anguidae, Anguinae and Gerrhonotinae are sister taxa to the exclusion of Diploglossinae. Several years after the publications of those focused studies on Anguimorpha, the first molecular structure-based phylogenies of all of Squa- mata appeared (Townsend et al., 2004; Vidal and Hedges, 2005). In general, most of the trees recovered in those analyses have the initial split in Anguimorpha between S. crocodilurus + Varanidae and Anguidae + Heloderma + Xenosaurus in some combina- tion. An Anguidae + Heloderma clade appears more often than a Heloderma + Xenosaurus clade. The molecular structure— based phylogenies are thus broadly congru- ent with the newest gross anatomy—based phylogenies, save for the unprecedented XENOSAUR PHYLOGENY ¢ Bhullar 67 nonmonophyly of Varanoidea, a clade sup- ported by numerous gross anatomical apo- morphies (summarize d by Gao and Norell, 1998: Conrad, 2005, 2008: Conrad et E-al. 2011), Regarding outgroups to Anguimorpha, the gross anatomy—based and molecular structure-based phyloge nies differ striking- ly. All of the gross anatomy—based phylog- enies have Anguimorpha as part of < monophyletic Scleroglossa (sensu Estes x al., 1988), with Iguania sister to that large ode: The appropriate outgroups to Angui- morpha would thus ee found within “Scincomorpha” (whose monophyly is not universally supported), at least a part of which is, in most gross anatomy—based phylogenies, more closely related to Angui- morpha than is Gekkota. The walccnlar structure-based phylogenies, on the other hand, generally recover Iguania as more closely allied to Anguimorpha than any “scincomorph” clade, sometimes with Ser- pentes intervening. Given the potential importance of Xeno- saurus and its extinct relatives to resolving the phylogeny of Anguimorpha and ther eby of Squamata as a whole, this study is directed at resolving the problematic ane tionships within the highly autapomorphic crown clade and its stem. The extinct taxa are particularly important to my analyses because they can break up the phylogenetic “long branch” leading to Xenosaurus (Gau- thier et al., 198S). Two other factors were important to the viability of the study. First, museum collections in the United States have sufficient skeletal and wet-preserved specimens of extant Xenosaurus to allow a reasonable sampling of taxa and some assessment of intraspecific variation. Fur- thermore, some of the skeletal specimens I used were disarticulated. Many of the characters identified here would have been impossible to see in articulated skeletons. Finally, high-resolution CT scanning tech- nology allowed the digital disarticulation of fossils in matrix and the visualization of osteoderms within the skin of extant spec- imens. 68 Bulletin of the Museum of Comparative Zoology, Vol. 160, No. 3 MATERIALS AND METHODS Ingroup The focus of the present study is Xeno- saurus and its extinct relatives, also referred to herein as “xenosaurs” (Shinisaurus and its extinct relatives are likewise called “shinisaurs’). The term Xenosauridae_ is used exclusively to refer to the “traditional” clade, including xenosaurs and shinisaurs as sister taxa, a topology only supported by one of the two analyses performed here. The characters included are limited in large part to those pertinent to relationships athe the xenosaur ingroup. Most of the 274 charac- ters I describe and score (Table 1) are new or newly defined; derivation from previous literature is noted in the character descrip- tions. The extinct taxa are particularly important in light of the highly derived nature of the crown clade: they provide the opportunity to break up “long branches” and resolve problematic cludes (Rowe, 1986; Gauthier et al., 1988). Xenosaurus. Twelve extant species of Xenosaurus were included in the unpub- lished external anatomy—based phylogenetic analysis by Canseco Marquez (2005): two of those are undescribed, and I was unable to obtai Xenosaurus arboreus, Xeno- saurus sanmartinensis, Xenosaurus phalar- oanthereon, and Xenosaurus penai—from (Was. aolleetions, Most species of Xenosaurus were once Classified as subspecies of Xeno- saurus grandis Gray 1856. Summaries of the ta ononne history of the group were pro- vided by King and Thompson (1968), Ballinger Chea: ” (2000). and Canseco Mar- quez (2005). The taxa used in my study (see Table 1) are Xenosaurus newmanorum Tay- lor 1949 (Fig. 1), Xenosaurus platyceps King and Thompson 1968 (Fig. 2), Xeno- saurus rackhami King and Thompson 1968 (as Xenosaurus srandis rackhami) (Fig. 3), X. grandis (Fig. 4), Xenosaurus agrenon King and Thompson 1968 (as Monescucns srandis agrenon), and Xenosaurus rectocol- ines Sint and Iverson 11993 (Fig 5): Alcoholic specimens were available for all taxa, and skeletal material was available for all bate X. rectocollaris. To supplement the skeletal material, I used high-resolution x- ray CT scans of the heads and bodies of all species save X. agrenon, for which the wet- preserved specimens were filled with metal shot that interfered with the scanning process. The scans were performed at the High-Resolution CT Scanning Facility at Tike University of Texas at Austin (UTCT) and allowed visualization of the articulated skeletons with the osteodermal armor in place. Exostinus serratus. The included speci- mens, from the Oligocene of Colorado and Wyoming, were described by Bhullar (2008; 2010), as was the history of this taxon. Exostinus lancensis. This is the most problematic taxon included in the study. It was described by Gilmore (1928) and Estes (1964) on the basis of fragmentary material from the Late Cretaceous of Wyoming and Montana. A summary of the history of the taxon until 1983 was provided by Estes (1983). Following that work, the only additional information on the anatomy of the taxon was provided by Gao and Fox (1996) on the basis of specimens from the western interior of Canada. Scoring here is based on those descriptions and on a number of undescribed specimens from the western United States in the collections of the American Museum of Natural History. For the purposes of this study, all of these disparate specimens are assumed to represent the same taxon. However, E. lancensis requires additional study. In particular, most of the American Museum specimens from the Lance Formation of Wyoming conform well to the descriptions of E. lancensis in the literature. However, frontals that are clearly associated with the characteristic parietals based on size, osteo- dermal sculpturing, and fit of the articular surfaces are large and unfused, whereas Gao and Fox (1996) described the anterior portion of a fused frontal that is nearly the same size. For the purposes of this study, frontal fusion is assumed to be ontogenet- ically variable in E. lancensis, but work on this taxon continues. Restes rugosus. This taxon was described as Exostinus rugosus by Gilmore (1942) and further by Estes (1965). It was renamed R. rugosus by Gauthier (1982). It is known primarily from a single well-preserved but disarticulated specimen, YPM PU 14640 from the late Paleocene of Wyoming. Isolated frontals that Gauthier (1982) re- ferred to the taxon were suggested to belong instead to a shinisaur or a platynotan by Smith (2006b). Restes rugosus was never fully described, nor has it been examined since the work of Gauthier (1982). The analysis here utilizes a low-resolution CT scan of the YPM PU 14640 block, which I found to contain a surprising amount of unexposed material. I had access to one additional specunen, an anterior section of a maxilla (UMMP 73565) from the Paleocene of Wyoming. Outgroup Anguimorpha. Because of the unresolved state of the larger clade Anguimorpha, it was necessary to use a variety of angui- morph taxa in the analysis. In general, the focus was upon extant taxa to minimize missing data, and because I was not attempting to resolve anguimorph relation- ships definitively. However, because S. crocodilurus was traditionally allied to Xenosaurus in Xenosauridae, the analysis includes two described fossil shinisaurs—B. ammoskius from the Eocene of Wyoming (Conrad, 2005, 2008) and Merkurosaurus ornatus from the Miocene of the Czech Republic (Klembara, 2008). These are the only included taxa that I scored entirely from the literature. I also included three anguids—the diploglossine Celestus ennea- grammus (part of the clade sister to all other diploglossines; Macey et.al... 1999), the anguine Ophisaurus pe ntronS. and the gerrhonotine Elgaria multicarinata. 1 ex- cluded the problematic glyptosaurs (see Conrad, 2005, 2008) and Anniella (Wiens and Slingluff, 2001; Conrad, 2005, 2008). Within Varanidae, L. borneensis and the monitor lizard species Varanus niloticus and XENOSAUR PHYLOGENY ¢ Bhullar 69 Varanus exanthematicus were used (Fuller et al., 1998; Ast, 2001; Pianka and King, 2004). Helodermatidae is scored as a composite taxon. Heloderma suspectum suffices for scoring most characters of this clade. However, for two characters (96 and 97), the derived nature of H. suspectum affected the resolution of xenosaur phylog- eny. The anterior (frontal and pre-orbital) skull roof osteoderms of extant Heloderma are highly domed, but those of Primaderma nessovi Nydam 2000, Gobiderma pulchra Borsuk- Bialy ynicka 1984, and Eurheloderma gallicum, primitive taxa on the stem of Heloderma, are flat and plate-like (Hoff- stetter, 1957; Gao and Norell, 2000; Nydam, 2000). The domed state of the extant taxa rendered ambiguous the nature of the flattened skull roof osteoderms of R. rugo- sus, whereas scoring the primitive state fox the helodermatid lineage results in a most parsimonious hypothesis that R. rugosus is primitive relative to Xenosaurus +. E. lancensis and pulls it out of a trichotomy with those taxa. In these cases,:as noted, scoring was performed as though the ancestr al states for the entire helodesniacd lineage were being considered. When great- er access to the skeletal material of these stem taxa is possible, they should be broken out and scored separ: ately, Outside of Anguimorpha. One non-angui- morph was used as an outgroup. A possible choice according to Conrad (2005) would be C. intermedia (Gao and Norell, 1998, 2000). However, C. intermedia is not universally agreed to lie outside of Angui- morpha and ial cannot be as thoroughly scored as an extant taxon. Because of the overwhelming molecular scale support for an Anguimorpha + Iguania clade, I chose the iguanian Pristidactylus torquatus as my additional outgroup. No previous gross ana- tomical phylogenetic analysis has used this topology. Variation The assessment of intraspecies variation, including but not limited to ontogenetic (OOOO Oe Or 0-10 © O- 0 << © © 1 © © 0 © 64. © & © & OF 0 0 0 snyonb.1o4 (ie Oe Omer cor UcoeO 420i 0 - bf SOO sO Of 1 ST 0) 900 o> oO <2 Hoy B05 OF © Bor sompueyunss (Onn MO Ol ie. e220 0) O- Tf 200 Od 20. © TL © 040 we Oo. 2 Sos 10° Oo W a0 snoyoj (ee MO Oh On ae OP) 0" O° sm 0. OO «6 OF © Or 0? 8 0 OO CO MO oT sisuaaUsog a Ae OG Ores st? Ge Oe 0 8 20 10 0° 2 0 Of Th Eh GO 0 O' F € oO £20 O79 T unjoadsns (ACO OnE(r MOla Ge ae) Ol OO) Tf “ON Oh @ 080) 0 90) © 00 od 0 € 9 0 4 80 OF @ TP. anumoisnene (OO Om OL 0m Oo Mann) “OS OL BORN On «+0 210) 0 0 6" O Om “2 0 2 Oe ke <0" OF rT sypajuaa tc 0 0 © O07 O97 42 & © 0 O50 OF 0 OF 0 Soe Oe 0 ian am (OO it Or sa ce eo Om Oo ace OF 0a 0 Lo) Th SO. £0 SO. o£ Oo € =O. & Fem O26 SNM YPOIOLD COO Oe sO co el SOG 20ers. MOR di OE. dyed? “nd ~aderd A dr id do aid. Sudee oc Sch es. es aid SNIYSOULUD (Mane a = (ie te et HOF We G ee Ol) Or ZOMmIGN ONGC Ce 1G 9G \ce pe co sce 1c 0G or sr fy OF Gra tr Gh Gr Ip Op Ge Se JG OG ces Po ce ce S (an eek .0 <0, OF sc Tf O° 0 10> 0 se es co Oe 606 OO OP Oy Oe oe snypnb1o} N 0 0-0 28 Oe SO Se Oe es Te a a OO ea S (Onan Ogu RENT ce Tiare Oe ROM Ok COL 2 “OM SOM Ore le eo. Seo. ck “Oe SOME ak, 20SP On etOmROLO snoyopiu s (een Mes EN rce a( eC: xOised anu abe Je One ce On Tt) Olle ody ce VOr 1c OR 2) mule ecile | Oar aoc) sqsudaaUlog = i OMe aero. Ble STOLE KO. Oe. 6 Oe Ca 0 8 Len O) u0fee nce, © OP ae SHOR 0" <0) 0h SOn SOO, Ole nO wnjoadsns a 0 1 6 @ OO 6) ew et) Fe Oe ee On iy Oe ee Ta Oy al OO OR Se S OM Cr nOhOo NO. On Oe. =O! Ore On 0) 50 sce od, FO Bee 0m SO. V0 a0E 0. (02 Oe Oued SHE S OY O Tt 8 0 6 0 Oe Oe Oe ee ei ee a OF OOO OO ke einen = (en sO On Mm SO we soho One & O00 Oo SO << rl: oi Oo fos Me oo Alte 1 role O SNIN]YPOIOLI = ee ee Ole crs ceed) Cath o Beate Ch Ore Gh Gee, ied Ge ade cee tocar ca a meee (dc On smpysouluD = er ee eer eo Oe hee) en. er Ye ce 0} FOr OH VON ey 20m 40 Oru W0e Se ar 400.0 snyvuso E eee ac Nea Ih ee el O- Teo Tf TO i i eae Mow Js, Se) ae sypupss = PO eect iT We Ie ak te To Oe TL JO FT ab Of ee eo Be eS ee Sez UDY YODA 8 ee Oe ice en ibe eo Mee O° at uO. el. @ Jhe eat, gGa MO. ceo 20 oe eG es SMUD]JOIOJIAA S nn iia ce ie nie ole Tl) ili). 2e0° 0 TO E«Za eGo ee Po o0" mh ot 41] z UOUIIDD > eer Me Cuil: Ite Cees 0 Ste, ao 0 HAR On theo eee oe ese ue a, Oe sdaohyjd < cn cen 0h iemceusl he Gm) Ble i Lh. 1. 0 ib sk Ol Prol |G. ore. ce a *O) Psp 1 2g. emounumel S Cr CoC mC) MC ce Me mice coeds Gilad ¢ Wd «@g “di ied do 8g ¢@ ode ¢ 6 8G ead SSE) = Ree CeCe PCa a cee, PCG eee ee He dy od gE Gh ad de od ed oe Pa SE GS 1S snsoond oo re Ta (na ace a me ae oe Se LI <- O RO: XO 0 0k LO Roce 8S 1. AO Ee E80 @ SNIDIAIS mCCmCcmneG ic mUGmeCE vomccn Comic 0G Glacl Z19l Oe 1 Oyo ie Oro ie Sea 0, G&G f OF fF (5 -O fF “Or ce © SY ROO Le! eS ee Oe LO Oh 0 Or On 1h Oe kU 1h Ss bt & bh & Lh OO Rb. Oe oO GO UOUAISD S een cee sie KORO! aie Oh Oe 9 SG ip On OL’ 0, *“S._G -G O. & LE OW “°F 0s 4h. lm “kh sdaohyojd = Oma Or sOh al Ake Oy. Ope Sim 90) “I 0k OL 0 8. OI) 8S Os. Oe OO OO Oe OL kN O WNLOUDUNIU = eRe ceric = cee ce Sd adc ice eG cea dy OG red OG OG OG Oe SO oe Oe SISUDIUD) S ee ercueptcee ice TniGw mice Bicime ce wach. cp’ ade ch da fedusdy din ch Gd Cease ee vie ie te VOF “ie Ge Or Ok Oh 0, Oy Li OL be OF Fe Oy AG I ee UOUIIDD < . Cree ah ORs Te ioe OF i toe 10) Ik aly OF Te sypaquaa Owe. 0 0: 05 1 0 G0) 0 “2 0 0 OO U0 090) 0 6 TT 0 © OO 0 SIP OO Vera Came RO Mods 0) be OPO 10) 20a ee PAK PFO OLE Ome Olen’ <0> OF al ee RO.- Tt 40'- 0 . Os raat SNANIPOIOLI ene Ueto CSc 3 10, 7e- Ol ded OO Or, 027, ees 0 Pe Se 60m ed > eS eee smysouUnUD Geom ee red eG de eS cd Lice te ee Na ere. Se. Oe CT 20> TO, Re A ee ieee Sao (eee Ae eo ol Lol Wis Ae oe SeKO Oe ta momar ale We rG.> OY NOS Te 10%" 0) ee ao spupis CR ONer in ame eG. TS eo el a On Oo ibe Mid? Vea yO." 20 = i 200) Wee aero MUDYYID CeO et ob Oe Tal. ot Oe ea a SOne ential “Oa a: 250: Ob Oh 9% SiO) Ae tee nO SLID ]JOIOJIOL Pek eek arc, ie Sb Re dO ce = Oem Ele gO: AO. 40: 4:0 6. S -skO SORE I> a > wah ea WOULD mec eC ol oo. Ly Lh ee re CO a ee eee a Gs Or OO" 0 O29 205 Oy a0. > ie oe el sdaohjojd DeOemir dy i “Co ho TY fo = (20. Ee. 0830) ier Sie 0 OOe O40) 3S S05 OOS 1k Or eon Crcmmiom tear we cd ae i A cae, gua mecha ice O; cil UE Or 2C. di) Sc) oe ee mire SISUDIUD] CeprO males vac. Vee” a5 Oy “OS GO ce 50; Shera ee sree a0 Or T= Sie BOL I £0e Ted: me leaey nO snsop nt Gua Cmca cs od) ca. aay ad ee 07 kOe nO OR baie: Oy: .OS~ al Se le oa diay (On. Sol a OD Rear De ee 0 SNYDALAs ! ( 9IZ SIZ FIT FIZ TIT IZ OLS 60S 80ST LOT 9NST SOT FOS OT GOST LOT 00ST GEI SBI LET Y6T SGI FET SGI Z6T IGT OGI GSI SST LST “dds snunvsouay “CHNNILNO?) Palaival: CONTINUED. TABLE 1. Yenosaurus spp. 949 250 Bulletin of the Museum of Comparative Zoology, Vol. 160, No. 3 naa oeeooocona onde ooe| VallaOn is an Imperane neta looked aspect of character discovery for Tee OOASCAMAAAOCOSSSS OS) phylogenetic analysis (Barahona ame dillo, 1998: Wible, 2003: Bever, 2005, 2006, sae eS Ce ty oa 2008). Large sample sizes representing a naoaooonceoeoceoooeooocoe| good sample ot the populanontaee species were not possible to obtain for all LAA aAAANIAILaAacoooesSos| taxa here examined, but such samples were available for enough taxa to phylogenetically PDE EE TE ODOT | _ bracket several clades’ Spectieatyaeaaneen aeounnanad neesecaka| 08 10 of mere specimens -paameeeeeeee = postnatal ontogeny, were available for V. SO HOANANAa ACA oCOoOSoCoa ag exanthematicus, E. multicarinata, H. suspec- tum, X. platyceps, and X. grandis. Lalso had a LOAN AAA ore ramiece.| Jarge sample Of X Chait | s-ou == mentary Information), although very young individuals were lacking. When statements are made in the character descriptions regarding ontogenetic variation, they refer wAANANA WWMM A-02 FH OOOOFa4.4. x naa Soa. naan aoooocaaa.| to the taxa listed above as being present as large samples or clades bracketed by them. SO Observed variation is noted in the charac- ter descriptions. For most taxa scored, PIE TEE IEEE IEEE EE EEE SES | multiple specimens were available for exam- eaaonananoaanceee San. ~| |ialiom Gee specimen Tsim sien eee Information). As might be expected, in Aaa ceorwnanaaaaanaaaaa.| general, sample sizes were low for the extinct taxa—one in the cases of R. rugosus (save for AO DFTA ANAAA WOON OANAA So a small fragment of maxilla, UMMP 73565) and B. ammoskius. The weakest part of the Br 2 9 ee i 4 i i he ae em i mem by fee 9 5 5 Fo A : ° ° extant ingroup in terms of variation assess- oawoocaauanaanooceSnaa.w| Ment was the X Gerenen 7 7. eee clade. For each of these taxa, I had only a Aaannecconaacocecucaaa.| single skeletal specimen (but multiple wet specimens). However, I discovered multiple As OF ODA OO O00 SOCIO A ee | SunGue SyaapOmorphics +On tic hss their position in my results is consistent, as mentioned above, with that recovered by other researchers using different sources of a-02.0 Ooao.namna.arddcdceqoec°c”c”oodrz a = ) data from gross anatomy. In general, I was Aaa nH ooooaaooSo4Saaa.| conservative in defining characters; most of the characters described in this work vary TeKOOeAAAE A AOOS Ooo ose-| httle qwithat ‘Species. 1or | whieh mana waacoawonaaaduaonoaaa.| imdividuals were available. Any variation that ee was present is noted. : a < Another aspect of variation is the assess- ee ae eee _| ment of the relative ontogenetic age of the 228§ Ses § o g¥S8 $858 2288] specimens examined, which affects the Eggs SPHSSSEEES8S5 882828] scoring of some characters. Within Squa- SESSA PLE HS §5 E85 35 8S! mata and more specifically Anguimorpha, XENOSAUR PHYLOGENY ¢ Bhullar io Figure 1. Cranium of Xenosaurus newmanorum, CT scan of UMMZ 126056. A, Dorsal, anterior to the left. B, Ventral, anterior to the left. C, Left lateral, anterior to the left. Illustrates characters 114(1), 151(0), 184(0), 247(1), 248(1), 249(0), 251(0), 252(0), 255(1), 256(0), 257(0), 258(3), 259(0), 260(0), 261(1), 262(0), 263(0), 264(2). several general ontogenetic trends can be Bracketed as ancestr ‘i for the clade (Bever et al., 2005: Bhullar, 2006; Bhullar and Smith, 2008, for Anguimorpha). Ontogenet- ic age can thus be inferred for fossils and extant specimens without age data. Merkur- osaurus ornatus specimens range from “subadults” with unfused frontals and pro- portionally shorter supratemporal processes of the parietal to large, mature adults with fused frontals, heavy resteodeemial sculptur- ing, and long supratemporal processes (Klembara, 2008). The single known spec- imen of B. ammoskius appears to be a relatively mature adult based on the prom- inent osteodermal sculpturing on the skull roof and the proportionally great length of the supratemporal processes. The: YaleoR: rugosus likewise has fully fused frontals and strong osteodermal sculpturing upon both the frontal and the maxillae, suggesting that it is a large, mature individual. Most specimens of E. lancensis have parietals that are heavily sculptured and rectangular in overall outline and are lar ge compared with those of other xenosaurs, suggesting fairly advanced ontogenetic stages, Tat the frontals are unfused save in a fawn individ- uals, indicating immaturity (taking into account the cautionary notes regarding this taxon made earlier). Finally, the known specimens of E. serratus are “approximately the same size as adults of other xenosaurs and bear heavy osteodermal sculpturing and generally * ‘mature adult” proportions of the Genel elements. 76 Bulletin of the Museum of Comparative Zoology, Vol. 160, No. 3 Figure 2. data processing). A, Dorsal, anterior to the left. B, Ventral, anterior to the left. C, Left lateral, anterior to the left. Illustrates characters 26(1), 248(0), 250(0), 253(0), 254(1), 258(4), 260(1), 262(1), 263(1). Analyses Instead of a single preferred starting hypothesis, I use two, generating results referred to as Analysis 1 and Analysis 2 in the descriptions and discussions that follow. Starting conditions for Analysis 1 were inspired by phylogenies of Anguimorpha on the basis of nucleic acid structure, notably that of Townsend et al. (2004), which remains the most thorough and thoughtful analysis of these data. To gener- ate Analysis 1, I used a constraint tree fixing extant taxa in the topology discussed previ- ously for molecular studies (with an eee dae + Helodermatidae clade instead a Helodermatidae + Xenosaurus clade or a trichotomy) and leaving relationships among Cranium of Xenosaurus platyceps, CT scan of UTA 23594, courtesy of Deep Scaly Project (2007; images from original xenosaurs, as well as among Shinisaurus and its extinct relatives, free to vary. To generate Analysis 2, I initially intended to constrain relationships based on a preferred gross anatomy—based phylogenetic hypothesis for Anguimorpha, but I found that my charac- ters alone, with P. torquatus spe cified as the outgroup, generated a hypothesis identical to that of Gone 2005, 2008), with one major difference: Senne us and its extinct relatives are not sister to Varanoidea but are united with xenosaurs in the traditional Xenosauridae. This result obtained with overwhelming support even after the exper- imental addition of the characters uniting shinisaurs with varanoids in Conrad’s anal- yses (2005, 2008; results not shown). ie 4 es XENOSAUR PHYLOGENY © Bhullar wage Figure 3. Cranium of Xenosaurus rackhami, CT scan of UTEP 4555. A, Dorsal, anterior to the left. B, Ventral, anterior to the left. C, Left lateral, anterior to the left. Illustrates characters 28(0), 123(0), 133(1), 135(2), 249(6), 252(3), 256(3), 264(3). Both analyses proceeded using the parsi- mony heuristic search epuon im PAUP* vy. 4.0b10, with default options save that tree bisection and reconnection (TBR) branch swapping was enabled with 1,000 random addition sequences (Swofford, 2001). Boot- strap analyses were run with default settings save the parsimony settings noted above anid the specification of 200 replicates. E Each set of conditions yielded a single most parsimo- nious tree under all three character state optimization options, mercifully obviating calculations of consensus. Ingroup relationships were identical in the results of Analysis 1 and Analysis 2. Because Analysis 2 was essentially uncon- strained save for the specification of an uncontroversial outgroup, I was able to use a nonparametric Wilcoxon signed ranks test (Templeton test), automated through PAUP®*, to assess whether the overall tree topologies (including the outgroups) were significantly different in terms of character support and evolution (Templeton, 1983; Larson, 1994). Inapplicable and unknown data were both scored as ? under the default settings in PAUP* v. 4.0b10. Failing to distinguish between the two can be problematic be- cause character states from applicable taxa can “bleed” to inapplicable taxa during optimization, but the alternative strategy of ‘absence coding” is also sroblematic (Strong and Lipscomb, 1999). Specifically, in such a case, inapplicability can be es as a synapomorphy when multi- ple taxa are scored as inapplicable for a single character. In several cases within this 78 Bulletin of the Museum of Comparative Zoology, Vol. 160, No. 3 Figure 4. Cranium of Xenosaurus grandis, CT scan of FMNH 123704, courtesy of Deep Scaly Project (2007; images from original data processing). A, Dorsal, anterior to the left. B, Ventral, anterior to the left. C, Left lateral, anterior to the left. Illustrates characters 135(1), 248(3), 250(1), 251(1), 252(2), 253(2), 256(2), 260(3). analysis, this presence/absence scoring would be redundant with another character. Multistate characters are discretizations of quantitative continua (counts or mea- surements) and were run as_ ordered, although the unordered setting was used to test the robusticity of the results. Mesquite v. 2.01 (Maddison and Maddison, 2008) was used to construct the character matrix and to trace character evolution. The process of discretizing continua necessarily leads to artificial bine. and the cutoffs for each character state are different to capture different levels of quantitative distinction among taxa. However, when more than two character states exist, the intervals are even. Even intervals prevent “cooking” of the data by selectively expanding some bins to encompass extra taxa. I first described most of the characters expressed as numerical values in less precise or more “qualitative” terms (all valid “qualitative” characters can be expressed in quantitative terms; see Wiens, 2001). The numerical values serve to give additional precision to these differ- ences: essentially, after describing the dif- ferences in the states of a character in com- parative or “qualitative” terms, I searched for a numerical expression of these differ- ences. This process explains why the cutoffs for different values are not always “stan- dard” intervals of 5 or 10, for instance (though, as noted, they are always regular intervals). The logical next step would be to attempt quantification of all characters and to score them using more advanced numer- XENOSAUR PHYLOGENY ¢ Bhullar 79 Figure 5. Cranium of Xenosaurus rectocollaris, CT scan of UF 51443. A, Dorsal, anterior to the left. B, Ventral, anterior to the left. C, Left lateral, anterior to the left. Illustrates characters 87(1), 123(1), 135(0), 165(2), 248(2), 249(3), 252(1), 253(1), 254(0), 255(0), 256(1), 257(1), 258(2), 259(1), 260(2), 264(4). ical methods such as gap-weighting (Wiens, 2001). For the purposes of thie present study, however, this analysis takes a middle ground, Institutional Abbreviations AMNH, American Museum of Natural History, New York; CAS, California Acade- my of Sciences, San Francisco; NAUQSP- JIM, Northern Arizona University Quater- nary Sciences Program, Jim I. Mead collec- tion (now houcede at East Tennessee State University, Johnson City); FMNH, Field Museum of Natural History, Chicago; MCZ, Harvard Museum of Compar ative “Zoology, Cambridge; MVZ, Museum of Nestebvate Zoology, “The University of California at Berkel TMM, Vertebrate Paleontology Laboratory, Texas Natural Science Gentcn The University of Texas at Austin; UF, Florida Museum of Natural History, Gains- ville; TNHC, Texas Natural History Collec- tions, The University of Texas at Austin: UMMZ, University of Michigan Museum of Zoology, Ann Arbor: USNM, National Mansi of Natural History, Washington, DC: UTA, The University of Texas at Arlington Herpetological Collections, Ar- lington: UTEP, The University of Texas at El Paso Natural History Collections, E] Paso; YPM, Yale Peabody Museum of Natural History, New Haven, Connecticut. CHARACTER DESCRIPTIONS Extended descriptions of characters and their distribution are contained in the electronic supplementary material. All char- SO Bulletin of the Museum of Comparative Zoology, Vol. 160, No. 3 58 60 100 100 100 100 Xenosaurus newmanorum Xenosaurus platyceps 100 99 Xenosaurus agrenon 100 Xenosaurus rectocollaris 100 Xenosaurus rackhami Xenosaurus grandis Exostinus serratus Exostinus lancensis Restes rugosus 100 Ophisaurus ventralis 100 Elgaria multicarinata Celestus enneagrammus Helodermatidae 60 Bahndwivici ammoskius 100 Shinisaurus crocodilurus 100 Merkurosaurus ornatus Lanthanotus borneensis 100 100 Varanus niloticus Varanus exanthematicus Pristidactylus torquatus Figure 6. Analysis 1 tree with bootstrap values greater than 50%. acters refer \ large “adult” individuals, except where noted. Measurements are given to define some character states; most states, however, are illustrated, and the figures should be used as primary guides in scoring speci- mens. Intraspecific variation was accounted for by measurements of additional individ- uals and by simple visual comparison when possible. Where angles to a plane or an edge are given, the acute component is provided. Comments on variation follow the charac- ter descriptions when variation that would affect scoring was evident. Given the low sample sizes of some of the taxa included in this study, the comments are certainly incom- plete. An intensive study of variation in the characters used in this study is warranted but beyond the scope of the current work. to the condition in relatively Description of the evolution of characters references both terminal taxa and clades that are supported by both resultant phylogenetic hypotheses (Figs. 6, 7), i.e., Anguimorpha, Anguidae, Varanidae, S$. crocodilurus + B. ammoskius + M. ornatus, and all subclades within the clade Xenosaurus + E. serratus. Varanus exanthematicus + Varanus niloticus is referred to as Varanus where scorings do not differ in Varanus salvator, a taxon on the other branch from the basal split of Varanus (Fuller et al., 1998: Ast, 2001). Bahndwivici ammoskius was scored based on the descrip- tion by Conrad (2006), and M. ornatus from that by Klembara (2008). When character states are identified synapomorphies of various clades, these state- ments are to be understood given the limitations of the taxon sampling for this study; 70 D3 87 95 97 XENOSAUR PHYLOGENY ¢ Bhullar S] 100 Xenosaurus newmanorum Xenosaurus platyceps 99 Xenosaurus agrenon 100 Xenosaurus rectocollaris 100 Xenosaurus rackhami Xenosaurus grandis Exostinus serratus Exostinus lancensis Restes rugosus Bahndwivici ammoskius Shinisaurus crocodilurus Merkurosaurus ornatus Elgaria multicarinata Ophisaurus ventralis Celestus enneagrammus Helodermatidae Lanthanotus borneensis 90 100 Varanus niloticus Varanus exanthematicus Pristidactylus torquatus Figure 7. Analysis 2 tree with bootstrap values greater than 50%. thus, most stated synapomorphies of clades outside of Xenosaurus and its stem are likely to represent synapomorphies of more inclusive clades or may not be valid with the inclusion of the vast array of additional extant and extinct taxa within Anguimorpha (Conrad et al., 2011). Most characters are new; if a version of the character appeared in a previous analysis, the original analysis is cited parenthetically atter the short description. The only characters that appeared in some form in previous analyses are 62, 91, 117, 124, 134, 145, 149, 189, 218, 294. 93]. JAG. and. 272. of ase: only 91, 149, and 189 are essentially unmodified. The full taxon—character matrix is depicted in Table 1. Dentition, General 1. Dentition: Tooth form (0) entirely uni- cuspid with pointed apices (Fig. 8A); (1) chisel-shaped (see Gao and Fox, 1996); (2) chisel-shaped, some bicuspidity (Fi ig. SB). VW aionon: Considerable ontogenetic variation occurs in tooth form. Bicus- pidity develops during postnatal ontog- eny in Xenosaurus aad in Anguidae. However, whereas early (postnatal)- stage anguid teeth are nearly conical, in “early stage Xenosaurus, the chisel shape already obtains. The teeth of iguanians in general do not undergo a unicuspid-to-multicuspid transition dur- ing postnatal ontogeny, although the extremity of development of fhe cusps may increase with ontogenetic age. Moreover, in no observed taxon do the teeth progress during ontogeny from multicuspid to unicuspid. Thus, the fossils with bicuspid teeth rugosus 82 Bulletin of the Museum of Comparative Zoology, Vol. 160, No. 3 Figure 8. A, Right dentary of Shinisaurus crocodilurus MVZ 204291, medial, anterior to the left. Illustrates character 1(0). B, Left dentary of Xenosaurus newmanorum uncatalogued NAUQSP-JIM specimen, medial, anterior to the right. Illustrates character 1(2). C, Premaxilla of Elgaria multicarinata TMM-M 8993, ventral, anterior to the top. Illustrates characters 2(0), 14(0). D, Premaxilla of Xenosaurus newmanorum uncatalogued NAUQSP-JIM specimen, ventral, anterior to the top. Illustrates characters 2(1), 14(4). and E. serratus are scored as such and the inferred late ontogenetic stages of the fossil shinisaurs are beyond theese at which the transition from unicuspid to multicuspid occurs in any observed extant specimen. Little variation occurs across broad ontogenetic stages in the general form of the teeth. Evolution. Under both phylogenetic analyses, state (0), unicuspid teeth, is plesiomorphic within Anguimorpha as suggested by Gauthier (1982). Further- more, under most hypotheses for the placeme nt of Anguimorpha that have been suggested, having unicuspid teeth iS a synapomorphy along the angui- morph stem. If C. intermedia is on the anguimorph stem as suggested by Con- rad (2005) and if the blunt, multicuspid state of the teeth of that taxon (Gao and Norell, 1998) is homologous to that of most other Scleroglossa,. chien this syn- apomorphy can be constrained to the internode between C. intermedia and Anguimorpha. However, various Meso- zoic squamates with unicuspid teeth are placed by some phylogenetic analyses as a sister clade to all other Scleroglossa (Evans et al., 2005; but see Conrad, 2008), potentially complicating the opti- mization of unicuspidity. The chisel shape of the tooth crowns is a synapomorphy of Xenosaurus and its fossil relatives. The presence of chisel-shaped teeth without bicuspidity in lancensis is either a primitive character of that taxon with biscupidity a synapomorphy of Xeno- saurus and an autapomorphy of R. rugosus, or the chisel-shaped state lacking bicuspidity is an autapomorphy of ou lancensis with the bicuspid chisel-shaped morphology as a synapomorphy of Xeno- saurus + Retes rugosus. Premaxilla The premaxilla is unknown for R. rugosus and E. lancensis. 9) aol Premaxilla: Curvature of rostral arc in horizontal plane (0) relatively broad (Fig. 8C); (1) relatively acute (Fig. 8D). Variation. The premaxillae of most taxa become slightly less acute during ontogeny, but even early neonates ot non-xenosaurs do not show the acute morphologies of adult Xenosaurus. Evolution. Acute curvature is a synapo- morphy of Xenosaurus under both analyses. Premaxilla: Height of dentigerous arc at suture of dermal marae’ to lateral (dermal) surface of maxilla: (0) dorso- ventrally short, about one-fifth or less of mediolateral width between contralateral sulumess (Pig. OA): (1) tall about one quarter or more of mediolateral width between contralateral sutures (Fig <3), Evolution. Under Analysis 1, two alternative hypotheses of character evo- lution each require three steps. In the first hypothesis of character evolution, the proportional tallness of the dentig- erous arc is a synapomorphy of Angui- morpha. The shortness of the dentiger- ous arc is then a synapomorphy of Anguidae + Helodermatidae and of Varanus. In the second hypothesis of character evolution, relative tallness is independently a synapomorphy Gin Xe= XENOSAUR PHYLOGENY ¢ Bhullar §3 nosaurus + E. serratus, S. crocodilurus + M. ornatus, and L. borneensis. Under Analysis 2, relative tallness is a synapo- morphy of Xe ‘nosauridae, independently of L. borneensis. Premaxilla: Angle to the horizontal in transverse plane of lateral edges of rostral body connecting portion (0) low, 40° or lower (Fig. 9A); (1) high, greater than 40° (Fig. 9C). Evolution. Taking into account the relatively vertical condition in P. torqua- tus and the wide distribution of that condition in other non-anguimorph Squamata and sphenodontians (Evans, 1980; Fraser, 1982: Whiteside, 1986), a low angle to the horizontal of the lateral edges is a synapomorphy of Anguimor- pha under both analyses. According to Analysis 1, a high angle to the Horcontal within Aguimorpha is independently a synapomorphy of E. serratus + Xeno- saurus and of S. crocodilurus + M. ornatus. According to Analysis 2 2, a high angle to the horizontal is a synapomor- phy of Xenosauridae. Premaxilla: Dersolateral angle of den- tigerous arc (0) formed by meeting of relatively straight edges, not produced into wedge (Fig. 9A); wal ) produced into small wedge (Fig, QB). Variation. The degree of development of the produced wedges varies within broad ontogenetic stages of X. rackhami and X. grandis but is always greater than that of The taxa scored as (0 i" Evolution. Under both analyses, pro- duction into wedges is a synapomorphy of X. grandis + X. rackhami and of oredr + M. ornatus. Premaxilla: Major premaxillary ethmoid canals (0) partially or completely bound- ed by connective tissue only, not fully surrounded by ossified premaxilla (Fig. 9D, left side); (1) bony, surround- ed by ossified premaxilla (Fig. 9D, right side). 84 Bulletin of the Museum of Comparative Zoology, Vol. 160, No. 3 Figure 9. Variation. The enclosure does not transform during postnatal ontogeny, but its does vary to some extent w athin taxa. Specifically, in one S. crocodilurus specimen examined, UF 71623, the enclosure is incomplete on one side. Enclosure is also present on one side in a single very large individual of E. multi- carinata, "TMM- M 8993 (Fig. 9D). In general, extremely large idivihials oat eect within anguimorphs, tend to devel- op extra flanges of bone that sometimes close pre Sealy open grooves for nerves and vasculature ( (personal observation). Evolution. Under Analysis 1, enclo- sure of the medial ethmoidal canals is a synapomorphy of Xenosaurus + E. serra- A, Premaxilla of Heloderma suspectum TMM-M 9001, anterior. Illustrates characters 3(0), 4(0), 21(0). B, Premaxilla of Xenosaurus grandis NAUQSP-JIM 1460, anterior. Illustrates characters 3(1), 5(1), 7 12(1). C, Premaxilla of Exostinus serratus, CT scan of USNM v16565, anterior. Illustrates characters 4(1), 8 13(0). D, Premaxilla of Elgaria multicarinata TMM-M 8993, posterior. Illustrates characters 6(0), 6(1), 19(0). 5(0), 7(0), 8(0), 9(0), (2), 9(2), 10(1), 11(2), (3), 9(1), 11(1), 12(0), tus and of S. crocodilurus + Varanidae. Under Analysis 2, enclosure is a synap- omorphy of Xenosauridae and an auta- pomorphy of L. borneensis. The hypoth- esis that Varanidae and its stem ancestrally showed enclosure of the canals is complicated by the existence of taxa along the stems of the two varanoid lineages fend perhaps Varanoidea as a whole) from the Cretaceous of Mongolia and the early Paleogene of Europe, including Estesia mongoliensis, Aiolo- saurus oriens, G. pulchra. and Necro- SAUTUS SP. that do not appear to show the morphology (Estes, 1983; Borsuk-Bialy- nicka, 1984: Norell et al., 1992: Norell and Gao, 1997: Gao and Norell, 1998, XENOSAUR PHYLOGENY ¢ Bhullar 85 Figure 10. A, Premaxilla of Shinisaurus crocodilurus UF 72805, anterior. Illustrates characters 7(1), 9(0), 11(0). B, Premaxilla of Xenosaurus newmanorumM uncatalogued NAUQSP-JIM specimen, anterior. Illustrates characters 8(4), 10(0), 13(1). C, Premaxilla of Xenosaurus newmanorum uncatalogued NAUQSP-JIM specimen, posterior. Illustrates characters 15(0), 16(1), 17(1), 18(1), 22(2), 23(1). D, Premaxilla of Xenosaurus grandis NAUQSP-JIM 1460, posterior. Illustrates characters 15(1), 16(0), 18(0), 19(1). ~ pulchra (Norell and G 2000). However, the issue is resolved if Varanoidea is not monophyletic, as in Analysis 1, and most or all of these taxa lie along the helodermatid stem, as was suggested for E. mongoliensis and G. Gao, 1997: Gao and Norell, 1998: Conrad, 2004, 2005, 2008). Premaxilla: Number of anterior forami- na collinear with main row of maxillary labial foramina (0) none (Fig. 9A); (1) two (Fig. 10A); (2) more than two (Fig, OB). Variation. Other than the variation in the degree of enclosure noted for Char- acter 6, the number of lower foramina does not vary in observed specimens of those taxa scored as possessing two (but see variation in upper foramina noted under Character S). Slight variation oc- curs in the number of foramina within the species of Xenosaurus—trom three to five, for example, in X. rackhami. No variation that would affect scoring was evident. Evolution. Under both analyses, the possession of several lower anterior fo- ramina is a synapomorphy of E. serratus + X. grandis. Under Analysis 1, the ances- tral state for the entire group, for Xenosaurus + Anguidae, and for S. crocodilurus + Varanidae is ambiguous between no foramina and two foramina. The ancestral state for Anguidae + Helodermatidae is no foramina. Under S6 . Premaxilla: Bulletin of the Museum of Comparative Zoology, Vol. 160, No. 3 Analysis 2. a lack of foramina is ancestral for the entire group and for Anguimorpha and two foramina is a synapomorphy of Varanidae: the ancestral state for Xeno- sauridae is ambiguous between no foram- ina and two foramina. Premaxilla: Number of anterior forami- na dorsal to main row of maxillary labial foramina (0) none (Fig. 9A); (1) one; (2) two; (3) three (Fig. 9C); (4) four or more (Fig. 10B). Evolution. Under both analyses, the presence of one to three dorsal foramina is a synapomorphy of E. serratus + Xenosaurus and an increased number of such foramina is a further synapo- morphy of Xenosaurus. Absence of foramina is an autapomorphy of X. rectocollaris. Premaxilla: Anterior surface of premax- illa just dorsal to teeth (0) flush with remainder of anterior surface (Fig. QA): (1) raised into supradental thickening (Fig. 9C); (2) supradental thickening pronounced (Fig. 9B). Variation. Little noticeable variation was evident, save that very young individuals of Xenosaurus have a slightly less pronounced thickening. Evolution. Under both analyses, the presence of the premaxillary supra- dental thickening is a synapomorphy of Xenosaurus + E. serratus. The pres- ence of a more pronounced thickening is a synapomorphy of Xenosaurus. Surface ventral to rostral osteoderm (0) rugose (Fig. LOB); (1) bearing discrete fused osteoderms (Fig. 9B). Variation. Osteoderms fuse to the premaxilla during postnatal ontogeny, but are present in mid-sized and large specimens within the species that pos- sess them. Evolution. Under both analyses, ru- gosity of the surface ventral to the | 13. 14. . Premaxilla: rostral osteoderm is a possible synapo- morphy of E. serratus + Xenosaurus; fusion of osteoderms in this region is a synapomorphy of the southern clade of Xenosaurus. Premaxilla: Rostral osteoderm (0) absent (Fig. OA); (1) oval and mediolaterally narrow (Fig. 9C); (2) rounded and mediolaterally wide (Fig. 9B). Variation. Other than the ontogenetic fusion of osteoderms to the premaxilla, I observed no intraspecific variation in this character. Evolution. Under both analyses, presence of a narrow rostral osteoderm is a synapomorphy of E. serratus + Xenosaurus. A rounded shape to the osteoderm is a synapomorphy of the southern clade of Xenosaurus. Distinct flanking osteo- derms dorsolateral to rostral osteoderm (0) present (Fig. 9C); (1) absent (Fig. 9B). Evolution. Under both analyses, the presence of flanking osteoderms is a synapomorphy of E. serratus + Xeno- saurus. Their absence is a synapomor- phy of the southern clade of Xeno- SAUTUS. Premaxilla: Rostral osteoderm and flank- ing osteoderms (when present) (0) smoothly domed or weakly keeled (Fig. 9C); (1) strongly keeled (Fig. LOB). Variation. The keel appears to form first during osteoderm development (un- published been ation), and as evidenced by juvenile X. platyceps, this character does not transform postnatally. Evolution. Under both analyses, a relatively smooth and rounded morphol- ogy is the primitive state for the rostral percadee m, with strong keeling a synap- omorphy of the northern clade of Xenosaurus. Premaxilla: Medial edges of vomerine processes oriented in a horizontal plane lis) 16. (0) more mediolaterally (Fig. 8C); (1) more anteroposte riorly (Fig. SD). Evolution. Under Analysis 1, medio- lateral orientation is a synapomorphy of Anguidae and of S. crocodilurus (or a more inclusive shinisaur-related clade). Under Analysis 2, more ambiguity exists. With one possible hypothesis of charac- ter evolution, an anteroposterior orien- tation is ancestral, with a mediolateral orientation an autapomorphy OL oS. crocodilurus. A mediolateral orientation is then either a synapomorphy of Anguidae + Varanoidea with an antero- posterior orientation a synapomorphy of Varanidae, or a mediolateral orientation is a synapomorphy of Anguidae and an autapomorphy of Helodeematidae. The second hypothesis of character evolution is that a mediolateral orientation is ancestral. An anteroposterior orientation is thus an autapomorphy of P. torquatus and a synapomorphy of E. serratus + Xenosaurus on the one hand and Varanidae on the other. Premaxilla: Medial edges of vomerine processes near midline convergence (0) straight, without dorsoventral bow (Fig. 10C); (1) with dorsoventral bow (Fig. LOD). Evolution. Under both analyses, the bowing or inflection is a synapomorphy of X. rackhami + X. grandis. Premaxilla: Stalk of incisive process (0) relatively long, length similar to or greater than diameter (Fig. 10D); (1) relatively short, length shorter than diameter (Fig. 10C). Variation. Some variation does occur across ontogenetic stages within taxa: in particular, a few X. rackhami and X. grandis show shorter processes. However, the vast majority of specimens where adequate samples are available have incisive processes of approximately uniform — relative length. individuals of We 1D: 20, XENOSAUR PHYLOGENY ¢ Bhullar 87 Evolution. Under both analyses, shortness of the stalk of the incisive process is a synapomorphy of the northern clade of Xenosaurus and an autapomorphy of X. rectocollaris. Under Analysis 1, ae SS iS a synapomorphy of Helodermatidae + Anguidae ee an autapomorphy of L. barons: Under Analysis 2, ee Ss 1S a synapomorphy of Anguidae + Varanoi- dea. Within this clade, a longer morphol- ogy is a synapomorphy of Varanus. Premaxilla: Accessory processes dorsal to vomerine processes (0) absent (Fig. 11A); (1) present (Fig. 10C). Evolution. Under both analyses, pres- ence of the accessory dorsal processes is a synapomorphy of Xenosaurus. . Premaxilla: Accessory processes dorsal to vomerine processes ((Q)) relatively short, nub-like (Fig. 10C); (1) relatively long and slender (Fig. LOD). Evolution. Under both analyses, rela- tive length of the processes is a synap- omorphy of the northern clade of Xenosaurus. Premaxilla: Fossa for rostral process of nasal septum (0) relatively mediolateral- ly narrow (Fig. 9D); (1) relatively me- diolaterally wide (Fig. LOD). Variation. Although the shape of the fossa varies somewhat among indi- viduals, no variation so great as to affect the scoring of the character was evident. Evolution. Under both analyses, rela- tively great width of the fossa is a synapo- morphy of E. serratus + Xenosaurus. Premaxilla: Angle of rise of nasal process in sagittal plane (0) relatively low (Fig. TLB): (1) intermediate (Fig. 11C); (2) relatively great (Fig. 11D). Evolution. Under both analyses, a moderate angle is ancestral and a high angle is a synapomorphy of S. ween hans + M. ornatus and an autapomorphy of Helodermatidae. Also under both analyses, a low angle is a synapomorphy 88 Bulletin of the Museum of Comparative Zoology, Vol. 160, No. 3 Figure 11. A, Premaxilla of Exostinus serratus, CT scan of USNM v16565, posterior. Illustrates characters 17(0), 22(1), 23(0), 24(1). B, Premaxilla of Elgaria multicarinata TMM-M 8993, left lateral, anterior to the left. Illustrates character 20(0). C, Premaxilla of Exostinus serratus, CT scan of USNM v16565, left lateral, anterior to the left. Illustrates character 20(1). D, Rostrum of Shinisaurus crocodilurus MVZ 204291, left lateral, anterior to the left. Illustrates character 20(2). of the northern clade of Xenosaurus and of Varanus. Under Analysis 1, a low angle is a synapomorphy of E. multicarinata + O. ventralis. Under Analysis 2, a low angle is either a synapomorphy of Anguidae with a moderate angle an autapomorphy of C. enneagrammus or an autapomorphy of E. multicarinata and of O. ventralis. . Premaxilla: Mediolateral width of nasal process at base (0) two to three tooth positions (Fig. 9A); (1) between four and five tooth positions (Fig. 12A); -(2) five tooth positions (Fig. 11A); (3) be- tween five and six tooth positions (Fig. LOD); (4) six or more tooth posi- tions (Fig. 10C). Variation. The values scored for this character are approximate; it can be difficult to determine precisely whether all or most of a tooth lies within the span of the nasal process. The character may need refining in the future. Neverthe- less, at least comparatively, the scorings are sound. Evolution. Under both analyses, a width of about five tooth positions is ancestral. This is largely a consequence of the state in P. torquatus and may seem unusual given the classical gross anatomy—based hypothesis of squamate phylogeny, in that many non-angui- morph scleroglossans have narrow nasal XENOSAUR PHYLOGENY © Bhullar 89 Figure 12. A, Premaxilla of E/garia multicarinata TMM-M 8958, posterior and slightly left lateral. Illustrates characters 21(1), 22(0), 24(0). B, Rostrum of Shinisaurus crocodilurus MVZ 204291, dorsal, anterior to the left. Illustrates characters 26(0), 67(0). C, Left nasal of Xenosaurus grandis NAUQSP-JIM 1460, ventral, anterior to the left. Illustrates character 27(0). D, Left nasal of Xenosaurus platyceps UF 45622, dorsal, anterior to the left. Illustrates characters 27(1), 28(1), 29(2). processes. However, in support of the inference here presented, a relatively wide-based nasal process is shared by a host of primitive iguanians on the one hand (Polychrotinae, Corytophaninae, Hoplocercinae, Leiolepis, ee Uromas- tyx, among others) and by C. interme- dia, sometimes aes to be the extinct sister taxon to Anguimorpha (Gao and Norell, 1998, 2000: Conrad, 2005, 2008). Under both analyses, a width of between five and _ six tooth positions is a synapomorphy of Xeno- sadurus and six or more is a synapomor- phy of the northern clade of Xenosaurus, as well as an autapomorphy of L. borneensis. Under Analysis 1, a width of two to three tooth positions is a synapomorphy of Anguidae, and a width of between four and ae tooth positions is an autapomorphy of E. multicarinata. A width of between four and five tooth positions is also a synapomorphy of Varanus, and a width of two to three an autapomorphy of V. niloticus. However, this result might be skewed by the unusual width of the snout of V. ex- anthematicus. Under Analysis 2, a width of between four and five tooth positions is a synapomorphy of Anguidae + Var- anoidea; a width of two to three is a synapomorphy of O. ventralis + C. enneagrammus and an autapomorphy of both Helodermatidae and V. niloticus. 90 irae: Bulletin of the Museum of Comparative Zoology, Vol. 160, No. 3 Premaxilla: Dorsolateral flare at con- tact with dermal surface of nasal (0) absent or minimal (Fig. 12A); (1) present with distinct loreal wedge separated (Fig. 11A); (2) pronounced (Fig. 10C). Evolution. Under both analyses, a pronounced premaxillary flare is a syn- apomorphy of the northern clade of Xenosaurus. Under Analysis 1, the flaring morphology is a synapomorphy of E. serratus + Xenosaurus. Additional- ly, either the flare is a synapomorphy of M. ornatus + S. crocodilurus and no flare is an autapomorphy of B. ammos- kius, or the flare is an autapomorphy of both ornatus and S. crocodilurus. Under Analysis 2, the flaring morphol- ogy is a sy napomorphy of Xenosauridae, aril the lack of a flare is an autapomor- phy of B. ammoskius. . Premaxilla: Anterior portions of nasal facets (0) shallowly impressed (Fig. 11A); (1) deeply impressed (Fig. 10C). Evolution. Under both analyses, deep impression is a synapomorphy of Xenosaurus and an autapomorphy of M. ornatus. Keel between nasal facets (0) dorsoventrally extensive (Fig. 12A); i restricted to dorsal portion of nasal process) (ie: 1A). Evolution. Under Analysis 1, relatively great extent of the keel is unambiguous- by a synapomorphy of E. Pr eran O. ventralis. At the initial split of the tree, the ancestral state is ambiguous. If the ancestral state is great extent, then restriction is either a ‘synapomorphy of Anguimorpha with great extent an auta- pomorphy of Var ee! or restriction is a synapomorphy of both Anguidae Xenosaurus and S. crocodilurus + M. ornatus. If the ancestral state is restric- tion, then great extent is an autapomor- phy of P. torquatus and a synapomorphy of Varanidae. Under Analysis 2, great extent is ancestral and restriction is a synapomorphy of Xenosauridae and an autapomorphy of both C. enneagram- mus and Helodermatidae. _ Premaxilla: Tooth position count, aver- age rounded to nearest integer, (Oe izctals) 8: (2) 9: (3) more than nine. Variation. Premaxillary tooth position count does vary, and the character scored is the average value for the individuals examined. However, even in those taxa whose count did vary, the variants always possessed one fewer tooth position than the average (and modal) number. Furthermore, these nonmodal variants were rare. Evolution. Under both analyses, hav- ing more than nine tooth positions is an autapomorphy of L. borneensis. Under Analysis 1, the ancestral state is ambig- uous. If it is eight or nine mtooern positions, then having seven tooth posi- tions is an autapomorphy of P. torqua- tus, and the other transformations are as described below for the case in which eight or nine is ancestral for Anguimor- pha. If seven tooth positions is the ancestral state, an increased number is a synapomorphy either of Anguimorpha OL OL Anguidae + Xenosaurus. Two possibilities obtain if an increased num- ber of either eight or nine is a synapo- morphy of Anguimorpha. If eight, then having nine tooth positions isa synapo- morphy of Anguidae + Xenosaurus and of Varanidae, and having seven tooth positions is a synapomorphy of S. crocodilurus + M. ornatus. If nine, then only the two-step transformation along the Shinisaurus stem need occur. Nasal The nasal is unknown for E. lancensis, R. rugosus, and M. ornatus. 26. Nasal: Contacts contralateral nasal dermal surface (0) extensively, along. at least a quarter of its medial margin (Fig. 12B); (1) barely or not at all (Fig. 2A). XS) | Variation. In younger individuals of V. exanthematicus, the peeale are barely in contact, but in medium to large individ- uals, they show extensive contact. The nasals of Varanus in general are unusual in their morphology and their contacts, concomitantly with the retracted nares characteristic of the clade (McDowell and Bogert, 1954; Estes et al., 1988). Evolution. Under both analyses, a small amount of contact is a synapomor- phy of Xenosaurus and an autapomor- phy of V. niloticus: Nasal: Posterior tapered portion (in horizontal plane) (0) mediolaterally wid- er at base than long along long axis of o nasal (Fig. 12C); (1) longer than wide (Fig. 12D). Evolution. Under Analysis 1, the longer-than-wide morphology is ances- tral, with wider-than-long a synapomor- phy of Xenosaurus + Anguidae and, within that clade, longer than-wide a synapomorphy of the southern clade of Xenosaurus. Under Analysis 2, the ancestral state for Anguimorpha_ is ambiguous. If the ancestral state is wider-than-long, then the longer-than- wide morphology is a synapomorphy of B. ammoskius + S. crocodilurus and the southern clade of Xenosaurus. If the ancestral state is longer- than-wide, then wider-than-long is a synapomorphy of Anguidae. Baatheraiore. wider-than- long is either a synapomorphy of E. serratus + Xenosaurus and longer-than- wide a synapomorphy of the southern clade of Xenosaurus, or wider-than-long is an autapormophy of E. serratus and a synapomorphy of the northern clade of Xenosaurus. . Nasal: Osteoderms of lateral row poste- rior to enlarged anterolateral osteoderm (0) comparable in size to or smaller than osteoderms of medial row (Fig. 3A); (1) larger and more prominent than osteo- derms of medial row: anterolateral osteoderm especially large (Fig. 12D). XENOSAUR PHYLOGENY ¢ Bhullar 9] Evolution. Under both analyses, greater size and prominence of the lateral row is a synapomorphy of the northern clade of Xenosaurus. bo CO Nasal: Anterior portion of lateral edge corresponding to underlying lateral bend of cartilaginous nasal capsule (0) with no change in angle or with deflection in horizontal plane away from the midline of a few degrees relative to the posterior portion of ‘the lateral edge (Fig. 13A); (1) with slight deflection from posterior portion of about 10°; (2) with notable bend, deflecting from the posterior portion by about 40 (Fig. 12D). Evolution. Under both analyses, an- cestral character states are ambiguous between a change in angle of 10° and 40° across most of the tree. Under Analysis 1, the only unambiguous syn- apomorphy is the lack of a change in angle uniting Varanus. Under Analysis 2 the ambiguity along the stem of Varanoidea shifts to no change in angle/slight change in angle. Septomaxilla Unfortunately, the septomaxilla is un- known in all of the extinct taxa. In general, this element is neglected in studies of squamate osteology. It received the most attention in work dealing with the complex olfactory organs of lizar ds (Stebbins, 1948: Bellairs, 1949: Hallerman, 1994; Bernstein, 1999). A few characters are used in current phylogenetic works on Serpentes (Lee and Scanlon, 2002). In the course of this investigation, I was able to identify a mame: of informative characters in the septomaxilla. In part, this stems from the unusually complex septomaxilla of Xeno- saurus, in and upon which several canals and ridges are formed that do not appear on the septomaxillae of most other squa- mates. 30. Septomaxilla: Posteromedial corner of cupular portion (0) with a long para- Bulletin of the Museum of Comparative Zoology, Vol. 160, No. 3 Figure 13. A, Nasals of Heloderma suspectum TMM-M 9001, dorsal, anterior to the left. Illustrates character 29(0). B, Right septomaxilla of Elgaria multicarinata TMM-M 8958, dorsal, anterior to the left. Illustrates characters 30(0), 31(0), 34(0). C, Left septomaxilla of Xenosaurus newmanorum NAUQSP-JIM uncatalogued specimen, dorsal, anterior to the left. Illustrates characters 30(2), 31(1), 34(1), 36(0), 38(0), 41(1). D, Right septomaxilla of Elgaria multicarinata TMM-M 8958, medial, anterior to the left. Illustrates character 32(0). septal process, several times as long as mediolaterally wide at base (Fig. 13B): (1) sending back moderate- length or short process, about twice as long as mediolaterally wide at base; (2) process absent or reduced to minor projection of corner, septomaxilla roughly an equilat- eral triangle in overall shape (Fig. 13C). Variation. The process becomes lon- ger during postnatal ontogeny, but its proportions are consistent with the scoring for most of ontogeny. Evolution. Under both analyses, moderate length of the process is ancestral and absence is a synapomor- phy of Xenosaurus. Under Analysis 1, ou, great length is a synapomorphy of Varanidae. Great length is also either a synapomorphy of Anguidae with moderate length an auto of C. enneagrammus, or great length is an autapomorphy of Helodermatidae and O. senna + FE. multicarinata. Under Analysis 2, great length .is a synapo- morphy at Anguidae + Varanoidea, and moderate length is an autapomorphy of C. enneagrammus. Septomaxilla: Deflection of bone away from rostral process of nasal septum to form distinct anteromedial surface (0) absent (Fig. 13B); (1) pres- ent (Fig. 13C). dy XENOSAUR PHYLOGENY © Bhullar 93 Figure 14. A, Left septomaxilla of Xenosaurus rackhami UTEP-OC “MALB” 388, lateral, anterior to the left. B, Left septomaxilla of Xenosaurus newmanorum NAUQSP-JIM uncatalogued specimen, medial, anterior to the right. Illustrates characters 32(1), 40(1). C, Right septomaxilla of E/garia multicarinata TMM-M 8958, lateral, anterior to the right. Illustrates characters 33(0), 37(0), 39(0). D, Left septomaxilla of Xenosaurus newmanorum NAUQSP-JIM uncatalogued specimen, lateral, anterior to the left. Illustrates characters 33(1), 37(1), 39(1), 40(0). Evolution. Under both analyses, the sure is not ancestral, it is a synapomor- presence of the anteromedial surface is phy of Anguimorph aora synapomorphy a svnapomorphy of Xenosaurus. of Xenosaurus and of Anguidae + Varanoidea. Under any of ies hypoth- eses of character evolution, lack of enclosure is a synapomorphy ot Vara- Bu: Septomaxilla: Medial ethmoidal nerves and vasculature (0) running in connec- tive tissue near surface of septomazxilla, nus. not enclosed by bone (Fig. 13D); (1) enclosed by septomaxilla (Fig. 3.32B). 33. Septomaxilla: Lateral ethmoidal nerves y Sey g I Eooinhon Under Analvee 1. eaclo- and vasculature (0) running in connec- sure is a synapomorphy of Xenosaurus + tive tissue near surface of septomaxilla, Anguidae and an autapomorphy of L. not enclosed by bone (Fig. 14C); (1) borneensis. Under Analysis 2, enclosure enclosed by septomaxilla (Fig. 14D). may be ancestral, in which case lack of Evolution. Under Analysis 1, lateral enclosure is an autapomorphy ot -P. enclosure is either a synapomorphy of torquatus and S. crocodilurus. If enclo- Anguidae + Xenosaurus, with lack of 94 Bulletin of the Museum of Comparative Zoology, Vol. 160, No. 3 Figure 15. A, Right septomaxilla of E/garia multicarinata TMM-M 8958, ventral, anterior to the left. Illustrates characters 35(0), 42(0). B, Right septomaxilla of Elgaria multicarinata TMM-M 8958, lateral, anterior to the right. Illustrates characters 35(1), 42(1). C, Left septomaxilla of Xenosaurus grandis NAUQSP-JIM 1460, dorsal, anterior to the left. Illustrates characters 36(1), 38(1), 41(0). D, Left septomaxilla of Xenosaurus grandis NAUQSP-JIM 1460, lateral, anterior to the left. Illustrates character 37(0). . Septomaxilla: enclosure a synapomorphy of Angui- dae, or lateral enclosure is a synapo- morphy of Xenosaurus and an autapo- morphy of Helodermatidae. Under Analysis 2, only the latter hypothesis of character evolution is most parsimo- nious. Dorsal anterolateral and (0) absent or low (Fig. 13B); and sharp (Fig. 13C). Evolution. Under both analyses, a high, sharp morphology is a synapomor- phy o Xenosaurus. ridges following anteromedial edges (1) high . Septomaxilla: Ventral surface posterior to vomeronasal cupula (0) smooth 6. (Fig. 15A); (1) impressed into longitudi- nal grooves by vomeronasal nerve (Fig. 15B). Variation. The sulci become more deeply incised during ontogeny but are present in even the youngest individuals examined. Evolution. Under both analyses, the presence of vomeronasal nerve im- pressions is a synapomorphy of Xeno- SaAUTUS. Septomaxilla: Anteroposterior length of flattened dorsal region near anterior apex, in front of capsular depression. Measured as proportion of total antero- posterior length of septomaxilla from 38. 40. . Septomaxilla: anterior apex back along anteroposterior axis (0) "0 So,on ereater (Bie, 13C); (1) less than 0.35 (Fig. 15C). Variation. The lamina of bone poste- rior to the cupular region of the septo- maxilla lengthens slightly during ontoge- ny, so the ratios above should be t: hea to apply to fairly mature individuals. Evolution. Because no other squa- mate possesses the flattened region, the ancestral state within Xenosaurus is ambiguous. Ratio of dorsoventral height of anterolateral face at shortest level to anteroposterior length of face (O) less than 0.10 (Fig. 14C): (1) 0.10 to 0.50 (Fig. 14D); (2) greater than 0.50 (Fig. 15D). Evolution. A dorsoventrally short an- terolateral face with a height-to-length ratio of less than 0.10 is ancestral under both analyses. Also under both analy- Secmeneeratio of O.10: to 0.50 is a synapomorphy of Xenosaurus and a ratio greater than 0.50 is a synapomor- phy of the southern clade of Xeno- Saurus. Septomaxilla: Anterolateral dorsal ridge continuing from lateral wing (0) broa and rounded (Fig. 13C); (1) narrow and sharp (Fig. 15C). Evolution. The ancestral state within Xenosaurus is ambiguous. . Septomazxilla: Ratio of anteroposterior length of lateral wing to anteropos- terior length of septomaxilla along anteroposterior axis from level of ante- rior apex to level of back of lateral wing (0) greater than 0.55 (Fig. 14C); (1) 0.55 or less (Fig. 14D). Evolution. Under both analyses, a ratio of 0.55 or less is a synapomorphy of the northern clade of Xenosaurus and an autapomorphy of Helodermatidae. Septomaxilla: Anterolateral ethmoid ca- nal (0) open only at round terminal 4]. Septomaxilla: XENOSAUR PHYLOGENY © Bhullar 95 foramina (Fig. 14D); (1) open along much of anterolateral face as long fissure (Fig. 14A). Evolution. Under both analyses, closure is the primitive state and the existence of the fissure is a synap- omorphy of the southern clade. of Xenosaurus. Septomaxilla: Canal paralleling and running lateral to the medial edge of the septomaxilla or to the antero- medial dorsal ridge (when visible) (0) absent (Fig. 15C); (1) present (Fig. 13C). Evolution. Under both analyses, presence of the canal is a synapo- morphy of the northern clade of Xenosaurus. Ratio of mediolateral width of vomeronasal cupula at its widest level to width of septomaxilla at that level (0) 0.45 or greater (Fig. 15A); (1) less than 0.45 (Fig. 15B). Evolution. Under both analyses, a ratio of less than 0.45 is a synapomorphy of the northern clade of Xenosaurus. Maxilla The maxilla is only partially preserved in E. lancensis and M. ornatus. 43. Maxilla: Medial premaxillary process, portion projecting beyond dental gutter, (QO) anteroposteriorly very short, “about twice as tall at posterior end as long (Fig. 16A); (1) short, about as tall as long: (2) intermediate length, about two-thirds as tall as long (Fig. 16B);: (3) long, about half as fall as long (Fig. 16C). Variation. The medial premaxillary process becomes slightly dorsoventrally taller during ontogeny, but despite this, no variation that would affect scoring was evident. Evolution. Under both analyses, a very short process is ancestral for the Bulletin of the Museum of Comparative Zoology, Vol. 160, No. we) Figure 16. A, Left maxilla of Shinisaurus crocodilurus UF 72805, medial, anterior to the right. Illustrates characters 43(0), 44(1), 46(2), 63(1). B, Left maxilla of Exostinus serratus, CT scan of USNM v16565, medial, anterior to the right. Illustrates characters 43(2), 45(0), 46(1), 60(1), 63(0). C, Left maxilla of Xenosaurus platyceps UF 45622, medial, anterior to the right. Illustrates characters 43(3), 45(1), 55 (n for numerator, d for denominator), 56 (n for numerator, d for denominator), 60(0). D, Left maxilla of Elgaria multicarinata TMM-M 8958, anterior. Illustrates character 44(0). entire group and for Anguimorpha anda long process is a synapomorphy of the are ambiguous among very short, short, and intermediate lengths. The long mor- Aor Hen clade of Xenosaurus. The phology is a synapomorphy of O. ventralis ancestral states for Xenosaurus + E. + C. enneagrammus and an autapomor- serratus and Xenosaurus are ambiguous between short and intermediate. Under Analysis 1, the ancestral state for Angu- idae + Xenosauridae is ambiguous among very short, short, and intermedi ate, as is that for Xenosaurus + R. rugosus. The long morphology is a synapomorphy of Anguidae + Heloder- matidae, and itermediate length is an autapomorphy of E. multicarinata. Un- der Analysis 2, the ancestral states for Anguidae + Varanoidea and Varanoidea 44. phy of . Along the other branch of Anguimorpha, the ancestral state for Xenosauridae and for Xenosaurus + R. rugosus is the very short morphology. Maxilla: Major anterior foramen for contents of infraorbital canal (ethmoidal nerve and accompanying structures) (0) exiting onto lamina intercristalis, lateral to crista transversalis (Fig. 16D); (1) exiting posteromedial to crista transver- salis (Fig. 16A). 45. M A6. Evolution. Under both analyses, the ancestral state of the character is ambiguous. However, the medial fora- men is not widespread outside of Anguimorpha, and I tentatively suggest that the ancestral state is for the exit to be in the upper region of the lamina intercristalis. Under Analysis 1, a medial exit is then a synapomorphy of Angui- morpha and a lateral exit is a synapo- morphy of Anguidae. Under Analysis 2, the ancestral state for Anguimorpha is ambiguous. Maxilla: Anterior end of lacrimal recess (O) relatively posterior, greater than one quarter of the way to the posterior end of the facial process (Fig. 16B); (1) relatively anterior, one quarter of the way back or less (Fig. 16C). Evolution. Under both analyses, the relatively anterior morphology is ances- tral for the total group and the relatively posterior morphology is ancestral for Xenosaurus + R. rugosus, with the relatively anterior morphology a synap- omorphy of Xenosaurus. Under Analysis 1, the ancestral state for Anguidae + Xenosaurus and for the éladis within Ame nicae is ambiguous. Under Analysis 2, the relatively posterior morphology is a synapomorphy of O. ventralis + C. enneagrammus. Maxilla: Anterior portion of dorsal margin of lacrimal recess—posterodor- sal rise in plane of facial process (0) shallow, 30° or less (Fig. LWA)» (1) moderate, between 30° and 35° (Fig. 16B); (2) steep, 35° and greater (Fig. 16A). Evolution. Under both analyses, a shallow angle of 30° or less is ancestral. In the absence of data for B. ammoskius, a steep angle of 35° or greater is an autapomorphy of S. ey ODOCINRS An angle between 30° and 35° is a synap- omorphy of Xenosaurus, and an angle of 35° or greater is a synapomorphy of X. agrenon + X. rectocollaris. 47. M 48. M 49. XENOSAUR PHYLOGENY ¢ Bhullar 97 Maxilla: Articular facet posterior to facial process (O) mediolaterally narrow, hardly differentiated from sharp dorsal edge of maxilla (Fig. 17B); (1) medio- laterally wide, with a distinct dorsally facing table (Fig. 17C). Evolution. Under Analysis 1, the wide morphology is a synapomorphy of E. serratus + Xenosaurus and of S. crocodi- lurus + B. ammoskius. Under Analysis 2, the ancestral state for Xenosauridae is ambiguous. Maxilla: Mediolaterally expanded facet posterior to facial process: (0) for lacrimal; (1) for jugal. Evolution. Because of widespread missing data, the ancestral state for the character is ambiguous at all levels. Maxilla: Width of palatal shelf at widest level (measured in horizontal plane perpendicular to axis of toothrow) (0) less than one-fifth length of maxillary toothrow (Fig. 17D); (1) one-fifth length of maxillary toothrow or greater OD ) (Fig. 18A). Evolution. Under both analyses, a wide palatal shelf is an autapomorphy of E. serratus. Maxilla: Posterior end of tooth row (0) relatively straight, collinear with more anterior portion (Fig. 17D); (1) medially deflected (Fig. 18A). Evolution. Under both analyses, me- dial deflection is an autapomorphy of E. serratus. . Maxilla: Infraorbital canal (0) round for entire length, confined to medial, steep portion oF palatal shelf (see E. multi- carinata, Vhe Deep Scaly Project [2007], coronal slices 35-124): (1) me- diolaterally expanded and oval for por- tion of length, extending above flat- tened, howeoutal lateral portion of palatal shelf (Fig. 18B). Evolution. Under both analyses, an expanded canal is a synapomorphy of 98 Bulletin of the Museum of Comparative Zoology, Vol. 160, No. 3 Figure 17. A, Left maxilla of Restes rugosus, CT scan of YPM PU 14640, medial, anterior to the right. Illustrates character 46(0). B, Left maxilla of Elgaria multicarinata TMM-M 8958, dorsal, anterior to the left. Illustrates character 47(0). C, Left maxilla of Xenosaurus grandis NAUQSP-JIM 1460, dorsal, anterior to the left. Illustrates character 47(1). D, Skull of Xenosaurus newmanorum, CT scan of UMMZ 126056, ventral, anterior to the left. Illustrates characters 49(0), 50(0), 93(1). Xenosaurus + R. rugosus and of Var- anus, and an autapomorphy of C. enneagrammus. 2. Maxilla: ae aa thickening (0) ab- sent (Fig. 18C); (1) weak, lose ridge above Bah (Hig. WSID) (2) strong, rounded ridge above teeth, dice ed dorsally by groove (Fig. 19A). Variation. The ridge becomes stron- ger with increasing age. Evolution. Under both analyses, weak development is an autapomorphy of E. serratus. Under Analysis 1, the ancestral state for the entire group is ambiguous between absence of the thickening and the weakly developed morphology, as are 03. M the ancestral states for Anguimorpha and Anguidae + Helodermatidae. Strong de- velopment is a synapomorphy of Xeno- saurus + R. rugosus and of S. crocodilurus + M. ornatus. Under Analysis 2, absence is the ancestral state for the entire sea and for Anguimorpha and Anguidae + Varanoidea. Within Varanoidea, weak development is an autapomorphy of Helodermatidae. Strong development is a synapomorphy of Xenosauridae. axilla: Major labial foramina: (0) pos- teriormost one or more foramina abruptly larger than others (Fig. 19B); (1) nearly the same size, aati only a subtle and gradual trend of posterior enlargement (Fig. 19A). 99 XENOSAUR PHYLOGENY ¢ Bhullar Figure 18. A, Left maxilla of Exostinus serratus, CT scan of USNM v16565, ventral, anterior to the left. Illustrates characters 49(1), 50(1). B, Cutaway near posterior end of infraorbital canal of rostrum of Exostinus serratus, CT scan of USNM v16565, anterior. Illustrates character 51(1). C, Left maxilla of Elgaria multicarinata TMM-M 8993, lateral, anterior to the left. Illustrates characters 52(0), 54(0), 59(0), 61(0), 62(1). D, Left maxilla of Exostinus serratus, CT scan of USNM v16565 lateral, anterior to the left. Illustrates characters 52(1), 54(1), 57(0), 58(0), 59(1), 61(3), 62(4). O4. saurus + R. Evolution. Under both analyses, rough equivalence in size along the entire row is a synapomorphy of Xeno- rugosus. Maxilla: Foramina (0) concentrated in single row of labial foramina (Fig. 1$C); (1) present in two rows—ventral labial foramina and dorsal row upon lower portion of facial process (Fig. 18D). Evolution. ancestral state for Under both analyses, the Xenosaurus +. E. serratus and for the southern clade of Xenosaurus is ambiguous, such that presence of the dorsal row could either be an autapomorphy of E. serratus and a synapomorphy of the southern clade of Xenosaurus or a synapomorphy of OO. Xenosaurus + E. serratus, with lack of a dorsal row a synapomorphy of the northern clade of Xenosaurus. Maxilla: Ratio of length of facial process along base (at dorsoventral level of inflection toward horizontal between rami of lacrimal) to total length of maxilla from anterior edge of cial process at that dJorcoventcal level to Bestcuel end of maxilla (Fig. 16C) (0) 0.20 to less than 0.30; (1) 0,30 to less: than 040: (2) 0:40 to less than 0.50: (3) 0:50 to less than 0.60; (4) 0.60 to less than 0.70; (5) 0.70 to less than 0.80. Variation. As the proportionally smaller orbit becomes during growth, Bulletin of the Museum of Comparative Zoology, Vol. 160, No. 3 Figure 19. A, Left maxilla of Xenosaurus platyceps UF 45622, lateral, anterior to the left. Illustrates characters 52(2), 53(1), 57(1), 61(2). B, Left maxilla of Elgaria multicarinata TMM-M 8958, lateral, anterior to the left. Illustrates character 53(0). C, Left maxilla of Xenosaurus rackhami UTEP-OC “MALB” 388, lateral, anterior to the left. Illustrates character 58(2). D, Left maxilla of Xenosaurus grandis NAUQSP-JIM 1460, lateral, anterior to the left. Illustrates character 58(3). the posterior suborbital portion of the maxilla shortens relative to the remain- der of the bone. However, this transfor- mation occurs largely during early onto- genetic stages, and I scored this character on relatively large individuals. Evolution. Under both analyses, state 1 is autapomorphic for X. agrenon, X. grandis, and B. ammoskius, state 4 is autapomorphic for O. ventralis, and state 5 is autapomorphic for L. bor- neensis. Under Analysis 1, the ancestral state for the entire group is ambiguous among state 0, state 1, state 2, and state 3. The ancestral state for Anguimorpha is ambiguous between state 2 and state 3, as is that for S. crocodilurus + Vara- nidae, Varanidae, Varanus, Xenosaurus + Anguidae, and Xenosaurus + R. rugosus. The ancestral state for S. crocodilurus + B. ammoskius is state 2, as is that for E. serratus + Xenosaurus and all nodes therein. Finally, the ancestral state for Anguidae is state 3, making state 2 an autapomorphy of E. multicarinata. Under Analysis 2, the ancestral state for the entire group is ambiguous among state 0, state 1, and state 2. The ancestral state for Xeno- sauridae and all nodes therein is state 2, and state 3 is autapomorphic for R. rugosus. The ancestral state for Angui- dae + Varanoidea and for Anguidae is ambiguous between state 2 and state 3. The ancestral state for O. ventralis + C. enneagrammus is state 3, as is that for 06. Varanoidea, making state 2 morphic for V. exanthematicus. autap )- Maxilla: Ratio of length of facial process along base (at Horo entra level of posterior inflection toward horizontal) to greatest height posterior to nasal facet (Fig. 16C) (0) 0.50 to less than 0.75; (1) 0.75 to less than 1.00: (2) 1.00 to less thramele257 (3) 1-25 tovless than 1.50: (4) es>0Gtomless than 175245) 1.75 to less than 2.00: (6) 2.00 to less than 2.25. Evolution. Under both analyses, state Q is a synapomorphy of Varanus and Statvenvo. 1S. am autapomorphy (oie Op borneensis and of O. ventralis. State 2 is an autapomorphy of C. enneagram- mus and a synapomorphy of S. crocodi- lurus + B. ammoskius. Furthermore, the ancestral state for Xenosaurus is ambig- uous between state 3 and state 4. wath Xenosaurus, state 5 is autapomorphic for X. newmanorum, and state 2 is synapo- morphic for X. agrenon + X. rectocol- laris, with state 1 an autapomorphy of X. agrenon. Under Analysis 1, the ancestral state for the entire BEOUD is ambiguous among state 1, state 2, and state 3. The ancestral state for Anguimorpha is state 2. State 4 is a synapomorphy of Xeno- saurus + Anguidae. Under Analysis 2, the ancestral state for the entire group is ambiguous among state 0, state 1, state 2, ae 3. and efite A. The ane eatral states for Anguimorpha and Xenosaur- idae are ambiguous among the last three of those states. That for Anguidae + Varanoidea is state 4. The ancestral state for Xenosaurus + R. rugosus is ambigu- ous between state 3 and state 4. . Maxilla: Narial margin of facial process (QO) dorsoventrally tall, horizontal portion about as extensive as or less extensive than vertical portion (Fig. 18D); (1) dorsoventrally short, homzontal margin more extensive than vertical portion (Fig. 19A). Evolution. Under both analyses, a proportionally short vertical margin is a o9. M 60. XENOSAUR PHYLOGENY ¢ Bhullar 10] synapomorphy of Xenosaurus. Under Analysis 1, a short vertical margin is a synapomorphy of Varanidae and an autapomorphy of Helodermatidae. Un- der Analysis 2, a short vertical margin is a synapomorphy of Varanoidea. Maxilla: Vertical portion of narial margin of facial process (0) failing to tilt anterodorsally beyond the edical by more than a few degrees (Fig. 18D); (1) tilting slightly be ond the vertical to form a dorsal ove srhang of the naris by the facial process mit a slight anterior eminence (Fig. 19A); (2) tilting beyond the vertical to form a dorsal overhang with a pronounced but rounded anterior eminence (Fig. 19C); (3) tilting beyond the vertical to form a marked dorsal overhang with a pronounced and sharply pointed anterior eminence (Fig. 19D). Variation. Smaller individuals of X. rackhami do not show as much anterior projection of the overhang as large adults, which were scored. Evolution. Under both analyses, a slight overhang is a synapomorphy of Xenosaurus, a moderately pronounced overhang is a synapomorphy Or rpehhane + X. grandis, and a very pronounced overhang is an autapomor- phy of X. grandis. Maxilla: Posterior portion of nasal facet upon facial process (0) curving smoothly in transverse plane, following general curvature of facial process (Fig. 18C): (1) distinctly folded toward the vertical relative to remainder of facial process, forming an upturned posterior tab of the nasal facet (Fig. 18D). Evolution. Under Analysis 1, the upturned tab is a synapomorphy of Xenosaurus + R. rugosus and of S. crocodilurus + B. ammoskius. Under Analysis 2, the upturned tab is a synapomorphy of Xenosauridae. Manilla: Highest point of facial process: (O) prominent, posterior corner between Bulletin of the Museum of Comparative Zoology, Vol. 160, No. 3 Figure 20. A, Left maxilla of Heloderma suspectum TMM-M 9001, lateral, anterior to the left. Illustrates character 61(1). B, Left maxilla of Shinisaurus crocodilurus MVZ 204291, lateral, anterior to the left. Illustrates characters 62(0), 118(1). C, Left maxilla of Restes rugosus YPM PU 14640, lateral, anterior to the left. Illustrates character 62(2). D, Fragmentary left maxilla of Exostinus lancensis AMNH 8498, lateral, anterior to the left. Illustrates character 62(3). prefrontal facet and orbital (posterior) edge (Fig. 16C); (1) dorsal edge of facial process Foal horizontal (Fig, 17A); (2) upturned tab of nasal facet ( Fig. 16B). Evolution. Under both analyses, the ancestral state for the entire group is for the highest point to be the corner between the prefrontal facet and orbital edge. The ancestral state for Xenosaurus + R. rugosus is a horizontal dorsal edge of the facial process, with the upturned tab of the nasal facet the highest point autapomorphically in E. ser ratus and X. rectocollaris. The corner between the prefrontal facet and orbital edge the highest point sy napomorphically in the northern clade of Xenosaurus. Under 61. Analysis 1, the ancestral state for Angu- idae + Xenosaurus is ambiguous oe tween the prefrontal facevonnte| edge corner and a horizontal dorsal edge, as is the ancestral state for Anguidae + Helodermatidae. A horvonts dorsal edge is an autapomorphy of L. borneen- sis. Under Analvsis 2. a horizontal dorsal edge is a synapomorphy of Xenosaurus + R. rugosus, and the ancestral state for Varcmendes is ambiguous between the prefrontal facet/orbital edge corner and a horizontal dorsal edge. Maxilla: Canthal crest (0) absent (Fig. 18C); (1) minimally developed: abrupt medial fold of facial proees but no projecting ridge (Fig. 20A); (2) projecting as a low ridge (Fig. 19A); (3) proje cting as an extensive, strong ridge (Fig. 18D). Variation. The crest, when present, becomes somewhat stronger with age, and the scoring here represents the condition in relatively large individuals. Evolution. Under both analyses, the ancestral state for the entire group is ambiguous between slight and moderate development of the crest. The ancestral state for Anguimorpha is moderate development. Also under both analyses, lack of a crest is a synapomorphy of Anguidae, a moderately developed crest is an autapomorphy of X. platyceps, and an extensive crest is an autapomorphy of E. serratus. Under Analysis 1, the ancestral state for S. crocodilurus + Varanidae is ambiguous between lack of a crest and slight development. Under Analysis 2, lack of a crest is a synapo- morphy of S. crocodilurus + M. ornatus and of Varanus. . Maxilla: Osteoderms upon facial process (0) absent (Fig. 20B); (1) present as nearly continuous sculptured plate or a small number of rectangular plates (Fig. 18C); (2) present as sculptured plate with sculpture concentrated into low mounds (Fig. 20C); (3) present as several low, polygonal (generally more edges than rectangles) tesserae (Fig. 20D); (4) present as pronounced mounds (Fig. 18D). This character was described and scored as “rugosity absent or present” by Conrad (2005, 2008). Variation. Fusion of osteoderms to dermal elements proceeds during post- natal ontogeny, and mound-shaped os- teoderms develop from flat lattices of bone (personal observation). Thus, rel- atively large individuals were scored. Evolution. Under both analyses, the ancestral state for the entire group and for Anguimorpha is no osteodermal fusion. The ancestral state for Xeno- saurus + R. rugosus is a sculptured 63. 64. XENOSAUR PHYLOGENY ¢ Bhullar 103 plate concentrated into scattered low mounds. The low, polygonal, tesselated morphology is a synapomorphy of E. lancensis + Xenosaurus, and the pro- nounced mound-shaped morphology is a synapomorphy of E. serratus + Xeno- saurus and an autapomorphy of Helo- dermatidae. The latter result is consis- tent with the presence of low, polygonal tesserae instead of pronounced mounds in primitive helodermatids (Hoffstetter, 1957). Under Analysis 1, the ancestral state for Anguidae + Xenosaurus is ambiguous among all four states. The ancestral state for Anguidae is ambigu- ous between state 0 and state 1. Under Analysis 2, the ancestral state for Xeno- sauridae is the absence of fused osteo- derms; presence of a few large sculp- tured plates is an autapomorphy of E. multicarinata. Note, however, that the apparently primitive (Conrad, 2005, 2008) glyptosaurid anguids generally have large sculptured plate- like osteo- derms fused to the facial process (Mes- zoely, 1970; personal observation). Maxilla: Tooth height (0) short, less than half of tooth extending beyond margin of bone (Fig. 16B); (1) tall, half or more of tooth extending beyond margin of bone (Fig. 16A). Evolution. Under both analyses, the ancestral state is ambiguous at all nodes whose branches have mixed states. However, this is largely a result of incomplete taxon sampling, as few non- anguimorph squamates have tall teeth. Maxilla: Tooth count (average, rounded toi mearest integer) (0) 7; (1)8; (2) 9; (3) 102 (4) ly (3) 12: (6) 13: 7) 142 (8) 15; (9) 6: UA) V7: (By LS: Variation. As noted, average values were scored. Ontogenetic increase in tooth number was previously document- ed for some squamates (Edmund, 1969; Ananjeva et al., 2003), but it was not in abundant evidence for the taxa I examined. 104 Bulletin of the Museum of Comparative Zoology, Vol. 160, No. 3 Figure 21. 68(2), 69(2), 70(1), 71 65(1), 68 71(0), 72(0). D, Left prefrontal of Restes rugosus YPM PU 14640, lateral, anterior to the left. Illustrates characters 66(1), 69(1). Evolution. Under both analyses. a count of seven is an autapomorphy of Helodermatidae, nine is an autapomor- phy of V. exanthematicus, 12 is an autapomorphy of E. serratus and S. crocodilurus, and 17 is an autapomorphy of E. multicarinata. Additionally, a count of 16 is a synapomorphy of Xenosaurus and a count of 18 is a synapomorphy of the northern clade of Xenosaurus, as well as an autapomorphy of X. rectocol- laris. Under Analysis 1, the ancestral state for the entire group is ambiguous from 13 to 16, and that for Anguimorpha is ambiguous between 13 and 14. The ancestral state for S. crocodilurus + Varanidae is ambiguous from 12 to 14, and a count of 15 is an autapomorphy of B. ammoskius. A count of 11 is a A, Left prefrontal of Exostinus serratus, CT scan of USNM v16565 lateral, anterior to the left. Illustrates characters 65(0), (1). B, Left prefrontal of Xenosaurus grandis NAUQSP-JIM 1460, lateral, anterior to the left. Illustrates characters (0). C, Left prefrontal of E/garia multicarinata TMM-M 8958, lateral, anterior to the left. Illustrates characters 66(0), 69(0), 70(0), Sy napomorphy of Varanidae. Along the other major anguimorph lineage, a count of 16 is a synapomorphy AG E. multicarinata + O. ventralis. Under Analysis 2, the primitive state for the entire group is ambiguous from, Zo 16. A count of 11 is a synapomorphy of Varanoidea. Prefrontal The prefrontal is unknown for E. lancen- sis and M. ornatus. 65. Prefrontal: Frontal process (0) relatively short and broad-based at divergence from main body of prefrontal, ahant twice as long along long axis as wide at base perpendicular to long axis (Fig. 21A); (1) relatively long and nar- LO5 XENOSAUR PHYLOGENY ¢ Bhullar Figure 22. A, Left prefrontal of Restes rugosus YPM PU 14640, dorsal, anterior to the left. Illustrates characters 67(1), 68(1). B, Left prefrontal of Xenosaurus newmanorum NAUQSP-JIM uncatalogued specimen, dorsolateral, anterior to the left. Illustrates character 68(3). C, Left prefrontal of Xenosaurus newmanorum NAUQSP-JIM uncatalogued specimen, lateral, anterior to the left. Illustrates character 72(1). D, Left vomer of Xenosaurus grandis NAUQSP-JIM 1460, dorsolateral, anterior to the left. E, Left vomer of Xenosaurus newmanorum NAUQSP-JIM uncatalogued specimen, same view. D and E illustrate characters 73(0), 73(1), 74(1), 74(2), 75(1), 76(2). 66. row-based, about two-and-a-half times as long as wide (Fig. 21B). Variation. The frontal process be- comes relatively shorter and more stout with age, and this character was scored on relatively large individuals. Large adults were available for all taxa scored as having long frontal processes. Evolution. Under both analyses, a long frontal process is an autapomorphy of O. ventralis and of X. newmanorum and a synapomorphy of X. grandis + X. rackhami. Prefrontal: Row of osteoderms along lateral edge of dermal surface (0) absent (Hig. 21); (1) present (Fig.:211)). i ood ies Variation. Osteoderms fuse to the prefrontal postnatally. Evolution. Under Analysis 1, presence of a lateral row of osteoderms is a synapomorphy of S. crocodilurus + B. ammoskius and of Xenosaurus +. R. rugosus. Under Analysis 2, presence of the lateral row is a synapomorphy of Xenosauridae. Prefrontal: Distinct osteodermal pattern of two small osteoderms adjacent to nasal facet and longer, lateral row of four with anteriormost laterally displaced, (0) ab- sent (Fig. 12B); (1) present (Fig. 22A). Variation. Osteoderms fuse to the prefrontal postnatally. 106 68. 69. 70. Bulletin of the Museum of Comparative Zoology, Vol. 160, No. 3 Evolution. Under both analyses, the “two and four” pattern is a synapomor- phy of Xenosaurus + R. rugosus. Prefrontal: Osteoderms (0) low, barely domed, with flat appearance (Fig. 21B); (1) moderately pronounced, with some doming (Fig. 22A); (2) pronounced with domes and low keels (Fig. 21A); (3) pronounced with high domes and sharp keels (Fig. 22B). Variation. Osteoderms become more domed during postnatal ontogeny, but as mentioned in the description for character 13, the keel forms early. Evolution. Under both analyses, the moderately pronounced morphology is ancestral. The low, barely domed morphology is a synapomorphy of S. crocodilurus + B. ammoskius. The ancestral states for Xenosaurus +. E. serratus and for Xenosaurus are am- biguous between the moderately pro- nounced morphology and the domed and moderately keeled morphology. The low morphology is a synapomor- phy of the southern clade of Xeno- saurus and the highly domed and highly keeled morphology a synapo- morphy of the northern clade. Prefrontal: Canthal crest (0) absent (Fig. 21C); (1) present as distinct ridge (Fig. 21D); (2) present as sharp, pro- nounced ridge (Fig. 21A). Evolution. Under both analyses, a sharp, pronounced morphology is an autapomorphy of E. serratus. Under Analysis 1, a distinct crest is a synapo- morphy of S. crocodilurus + B. ammos- Kius and of Xenosaurus + R. rugosus. Under Analysis 2, it is a synapomorphy of Xenosauridae. Prefrontal: Emargination or straight edge in maxillary flange (0) dorsoven- trally extensive, representing over half of entire edge length of flange (Fig. 21C); (1) dorsoventrally restricted, less than half of entire edge length (Fig. 21A). Variation. The maxillary flange is extremely thin early in postnatal on- togeny, and in young individuals, it may be damaged or only partially ossified. This character is best scored on relatively large individuals. Evolution. Under both analyses, the restricted morphology is an autapomor- phy of E. serratus and of Helodermatidae. 71. Prefrontal: Lacrimal foramen (0) rela- tively unconstricted, margins remain divergent (Fig. 21C); (1) relatively constricted, margins become nearly parallel (Fig. 21A). Evolution. Under both analyses, the parallel morphology is an autapomorphy of E. serratus. ~l ho . Prefrontal: Lacrimal facets (0) relatively smooth (Fig. 21C); (1) adorned with complex ridges and bumps (Fig. 22C). Variation. When the complex adorned morphology is present, it tends to be more pronounced in older indi- viduals. Evolution. Under both analyses, the adorned morphology is a synapomorphy of Xenosaurus + R. rugosus. Vomer The vomer is unknown in F. serratus, E. lancensis, R. rugosus, and M. ornatus. It is not visible in the single known specimen of B. ammoskius (Conrad, 2006). 73. Vomer: Anterior facet for medial pre- maxillary process of maxilla (0) less than three times as long along its long axis as tall perpendicular to” lone Mads (Fig. 22D); (1) three or more times as long as tall (Fig. 22E). Evolution. Under both analyses, the elongated morphology is a synapomor- phy of the northern clade of Xenosaurus. 74. Vomer: Anterior facet for medial pre- maxillary process of maxilla (0) abruptly diverging ascending portion of ventral edge absent (Fig. 23A); (1) abruptly XENOSAUR PHYLOGENY ¢ Bhullar 107 79(1), 80(1) Figure 23. A, Left vomer of Varanus exanthematicus TMM-M 8956, dorsolateral, anterior to the left. Illustrates characters 74(0), 75(0). B, Rostrum of Pristidactylus torquatus CAS 85234, ventral, anterior to the left. Illustrates characters 76(0), 78(0), 87(0). C, Left vomer of Xenosaurus rackhami UTEP-OC “MALB” 388, ventromedial, anterior to the left. D, Left vomer of Xenosaurus platyceps UF 45622, same view. C and D illustrate characters 76(1), 77(0), 78(1), 78(2), 79(0), 81(0), 81(1), 82(0), 82(1), 83(0), 83(1). E, Left vomer of Xenosaurus newmanorum NAUQSP-JIM uncatalogued specimen, same view. F, Left vomer of Xenosaurus platyceps UF 45622, same view. E and F illustrate characters 76(2), 79(1), 80(0), 80(1). diverging ascending portion of ventral Varanidae is ambiguous between ab- edge relatively short, beginning over sence and the short morphology. The halfway to medial limit of facet ancestral state for Varanidae is ab- (Fig. 22E); (2) ascending portion rela- sence. Along the other major branch of tively long, beginning less than halfway Anguimorpha, the ancestral state for Peomedaibimite: facet (Fig. 22D). Xenosaurus + Anguidae is ambiguous between absence and the short mor- phology, but the short morphology is the ancestral state for Xenosaurus and for Anguidae. Evolution. Under both analyses, the ancestral states for the entire group and for Anguimorpha are ambiguous between absence and the short mor- phology. The ancestral state for the 75. Vomer: Anterior process from antero- southern clade of Xenosaurus is am- medial corner (0) absent (Fig. 23A); (1) biguous between the short and long present, short—about as long along long morphologies, and the long morpholo- axis as wide perpendicular to it gy is an autapomorphy of E. multi- (Fig. 22D); (2) present, long—longer carinata. Under Analysis 1, the ances- along long axis than wide perpendicular tral state for S. crocodilurus + to long axis (Fig. 22E). 108 76. 78. . Vomer: Bulletin of the Museum of Comparative Zoology, Vol. 160, No. 3 Evolution. Under both analyses, pres- ence of a short process is a synapomor- phy of Xenosaurus, and a long process is a synapomorphy of the nOriNenn clade of Xenosaurus. Vomer: Lateral parasagittal crest on ventral surface (0) short, extends about a third of the way to the back of the vomer (Fig. 23B); (1) intermediate length, extends about half of the way to the back of the vomer (Fig. 23C); (2) long, extends about two-thirds of the way to the back of the vomer (Fig. 23F). Evolution. Under both analyses, the ancestral state for the entire group is ambiguous between short and interme- diate length. The ancestral state for Xenosaurus is the intermediate mor- phology, with the short morphology an autapomorphy of X. agrenon and the long morphology a synapomorphy of the atom clade. Under Analysis 1, the ancestral state for Anguimorpha is ambiguous between short and interme- dates length. The ancestral state for NenOsnns + Anguidae is intermediate length, that for Anguidae is ambiguous hereon niernediate length and the long morphology, and that for Varanidae is ake short morphology. Under Analysis 2, intermediate length is ancestral for Anguimorpha. The long morphology is a synapomorphy of O. ventralis + C. enneagrammus, and the short morphol- ogy is a synapomorphy of Varanidae. Lateral parasagittal crest (0) extends largely dorsoventrally (Fig. 23C); (1) folds medially as a tall flange, underhanging much of the vomer (Fi ig. 24A). Evolution. Under both analyses, the medial folding is an autapomorphy of X. agrenon. Vomer: Mediolateral width of lateral flange bordered medially by lateral parasagittal crest at widest anteroposte- rior level (0) narrow, less than half that of SO. Sk . Vomer: remainder of vomer (Fig. 23B); (1) inter- mediate, one half of that of remainder of vomer to just under equal to that of remainder of vomer (Fig. 23D); (2) wide, equal to or wider than that of remainder of vomer (Fig. 23C). Evolution. Under both analyses, the narrow morphology is ancestral for the entire group and for Anguimorpha. The intermediate morphology is ancestral for Xenosaurus and the wide morphology is a synapomorphy of the southern clade of Xenosaurus. Under Analysis 1, the ancestral state for Xenosaurus + Angu- idae is ambiguous between narrow and intermediate, and the ancestral state for Anguidae is ambiguous between inter- mediate and wide. “Under Analysis 2, the ancestral state for Anguidae + V‘ apneidee and for Xenosauridae is narrow, making the intermediate morphology a synapo- morphy of Anguidae and the wide a synapomorphy of O. ventralis + C. enneagrammus. Medial parasagittal crest (0) ends posterior to vomerine foramen (Fig. 23C); (1) ends at level of, or ante- rior to, vomerine foramen (Fig. 23F). Evolution. Under both analyses, ter- mination at or anterior to the vomerine foramen is a synapomorphy of the northern clade of Xenosaurus. Vomer: Medial parasagittal crest extend- ing to or past level of vomerine foramen (0) passes lateral to foramen (Fig. 23E); (1) passes medial to foramen (Fig. 23F). Evolution. Given the restricted distri bution of the character states, it is impossible to infer the ancestral state of this character. Vomer: Vomerine foramen (0) small, mediolateral width less than one-eighth mediolateral width of vomer at same level (Fig. 23D); (1) large, mediolateral width one-eighth mediolateral width of vomer at same level or greater (Fig. 23C). XENOSAUR PHYLOGENY ¢ Bhullar 109 Figure 24. A, Rostrum of Xenosaurus agrenon UTACV 145008, ventral, anterior to the left. Illustrates character 77(1). B, Left palatine of Xenosaurus newmanorum NAUQSP-JIM uncatalogued specimen, dorsal, anterior to the top. C, Left palatine of Xenosaurus grandis NAUQSP-JIM 1460, same view. D, Left palatine of E/garia multicarinata TMM-M 8958, same view. B through D illustrate characters 84(0), 84(1), 85(0), 85(1), 85(2), 86(0), 86(1), 88 (n for numerator, d for denominator), 89(0). 3. Vomer: Evolution. Under both analyses. a large foramen is an autapomorphy of X. rackhami. 2. Vomer: Posterior end of palatine process (0) nearly straight in horizontal plane (Fig. 23D); (1 ) bowed laterally (Fig. 23C). Evolution. Under both analyses, lat- eral bowing is a synapomorphy Ol X. grandis + xX rackhami. Palatine process, width mea- sured by ratio of mediolateral width to widest mediolateral width of vomer, (0) mediolaterally wide, ratio 0.44 or greater (Fig. 23D); (1) mediolaterally narrow, ratio less than 0.44 (Fig. 236). Evolution. Under both analyses. the ancestral state for the entire group is the wide morphology, as is that for Anguimorpha. The ancestral state for the southern clade of Xenosaurus is ambiguous, with the narrow morphol- ogy either a synapomorphy of the entire clade and the wide morphology an autapomorphy of X. rectocollaris or the narrow morphology an autapomor- phy of X. agrenon and a synapomorphy Oi Xe srandis + X. rackhami. Under AAlwsis l, on narrow morphology is an autapomorphy of S. crocodilurus. Under Analysis 2, the ancestral state for Xenosauridae is ambiguous. Palatine The palatine is unknown in E. serratus, E. lancensis, R. rugosus, and M. ornatus. It is 110 Bulletin of the Museum of Comparative Zoology, Vol. 160, No. 3 not visible in the single known specimen of B. ammoskius (Conrad, 2006). 84. 85. S6. 87. Palatine: Choanal margin (0) straight or convex anteriorly in horizontal plane (Fig. 24D); (1) ‘concave anteriorly in horizontal plane (Fig. 24C). Evolution. Under both analyses, ante- rior concavity is a synapomorphy of the northern clade of Xenosaurus and an autapomorphy Ore grandis. Under Analysis 1, anterior concavity is an autapomorphy of Helodermatidae and of L. borneensis. Under Analysis 2, the ancestral states for Varanoidea and Var- anidae are ambiguous, with anterior concavity either a synapomorphy of Varanoidea and convexity or linearity a synapomorphy of Varanus or anterior concavity an autapomorphy of Heloder- matidae and L. borneensis. Palatine: Eminence in choanal margin (0) absent (Fig. 24D); (1) small and located medial to the mediolateral midpoint of the margin (Fig. 24C); (2) large, broadly curved, and located around the mediolateral midpoint of the margin (Fig. 24B). Evolution. Under both analyses, ab- sence of the eminence is ancestral for the entire group and the small, medial morphology is a synapomorphy of Xe- nosaurus. The large, broad morphology is a synapomorphy of the northern clade of Xenosaurus. Palatine: Maxillary process (0) about as anteroposteriorly long as mediolaterally wide (Fig. 24D); (1) distinctly anteropos- teriorly longer than mediolaterally wide (Fig. 24C). Evolution. Under both analyses, the longer-than-wide morphology is a syn- apomorphy of Xenosaurus. Palatine: Pterygoid process ventral surface (0) bearing teeth (Fig. 23B); (1) bearing midline ridge or groove (Fig. 5B); (2) smooth (Fig. 25A). Presence or absence 88. of palatine teeth was character 82 of Estes et al. (1988). Variation. I did not have ontogenetic series for taxa that possess palatine teeth, but the palatal teeth upon the pterygoid of E. multicarinata increase in number during postnatal ontogeny. Evolution. Under both analyses, the ancestral state for the entire group is ambiguous among the three states, but the ancestral state for Anguimorpha is absence of teeth and ridges. The presence of ridges in X. rectocollaris, and of teeth in O. ventralis and L. borneensis, is autapomorphic for each of these taxa. Palatine: Pterygoid process, ratio of length along long axis beginning at posterior end of base of vomerine process to width perpendicular to long axis at constriction behind base of vomerine process (Fig. 24C) (0) 1.1 or less; (1) greater than J.1 to wee) greater than 1.2 to 1.3; (3) greater than 1.3 to 1.4; (4) greater than 1.4 to 1.5; (5) greater than 1.5. Variation. The pterygoid process in some but not all taxa becomes rela- tively wider with age, and all scorings were performed on relatively large individuals. Evolution. Under both analyses, the ancestral state for Xenosaurus is ambig- uous between state | and state 2. State 0 is a synapomorphy of X. agrenon + X. rectocollaris, and state 4 is an autapo- morphy of X. rackhami. Under Analysis 1, the ancestral state for the entire group is ambiguous among all states, as is that for Anguimorpha and Xenosaurus + Anguidae, as well as Shinisaurus + Varanidae. State 0 is a synapomorphy of Varanidae and an autapomorphy of Helodermatidae. Under Analysis 2, state 5 is ancestral for the entire (eroup: Anguidae + Varanoidea, and Xenosaur- idae. State 0 is a synapomorphy of Varanoidea. XENOSAUR PHYLOGENY ¢ Bhullar 11] Figure 25. A, Left palatine of Xenosaurus rackhami UTEP-OC “MALB” 388, ventral, anterior to the left. B, Left palatine of Xenosaurus newmanorum NAUQSP-JIM uncatalogued specimen, same view as A. A and B illustrate characters 87(2), 90(0), 90(1). C, Left palatine of Xenosaurus platyceps UF 45622, dorsal, anterior to the left. Illustrates character 89(1). D, Frontal of Elgaria multicarinata TMM-M 8958, dorsal, anterior to the left. Illustrates characters 91(0), 94(0), 97(0), 104(1), 108(0). E, Anterior tip of frontal of Exostinus serratus, CT scan of USNM v16565, ventral, anterior to the left. Illustrates character 93(0). 89. 20) Palatine: Dorsomedial tongue for ptery- goid articulation (0) anteroposteriorly longer than mediolaterally wide at base (Fig, 24B): (1) mediolater ‘ally wider at base than long (Fig. 25C). Evolution. Under both analyses, the wider-than-long morphology is anauta- pomorphy of xX platyceps. Under Anal- ysis 1, the wider-than-long morphology is a synapomorphy of Var ac and an anteponioyphiy of Helodermatidae. Un- der Analysis 2, it is a synapomorphy of Varanoidea. Palatine: Anterior to mediolateral di- vergence of dorsomedial and ventrolat- sal tongues for pterygoid articulation, medial edge of dorsoventrally diverging ventrolateral tongue (0) raised into sharp, underhanging ridge (Fig. 25B); (1) raised into ow “ridge with little if any underhang (Fig. 25A). Evolution. Under both analyses, the low, rounded morphology is a synapo- morphy of the southern clade of Xeno- SAUTUS. Frontal 91. Frontal: Frontals (O) remain unfused throughout ontogeny (Estes et al., 1988, Fig, 13B): (1) hee at some point during ontogeny (Fig. 25D). Variation. Frontal fusion occurs dur- ing ontogeny; in the taxa (specified in Weterale and Methods) for which very 112 93. _ Frontal: Bulletin of the Museum of Comparative Zoology, Vol. 160, No. 3 young postnatal specimens were avail- able, a seam and sometimes a split at the anterior end still remained. Evolution. Under both analyses, the fused morphology is ancestral for the entire group. Under Analysis 1, pairing is a synapomorphy of Var ‘anidae and of Anguidae + Helodermatidae, with the fused morphology an autapoInamnay of E. multicarinata. Under Analysis 2, the ancestral state for Anguidae + Wanner dea is ambiguous; either the paired morphology is a synapomorphy of that clade, and fused morphology is an autapomorphy of E. multicarinata, or the paired morphology is a synapomor- phy of O. ventralis + C. enneagrammus and of Varanoidea. Frontals in taxa where fusion occurs (0) remain separate for some period of time postnatally; (1) fuse prenatally. Variation. This character depends upon the recognition of ontogenetic variation. It is possible that unrecog- nized individual variation occurs regard- ing when and whether fusion occurs, in particular among the extinct taxa. Evolution. Under both analyses, the ancestral state for the entire group is prenatal fusion. Postnatal fusion is an autapomorphy of E. lancensis and of M. ornatus. Frontal: In taxa with fused frontals, frontals fuse and raised seam at line of fusion vanishes (0) only at most ad- vanced ontogenetic stages (Fig. 25E); (1) by attainment of approximately two- thirds of “adult Size of frontals ( Fig. 7D): Variation. As noted, this character is scored based on the nature of inferred ontogenetic variation. Evolution. Under both analyses, the ancestral state for the entire group and for Anguimorpha is early disappear- ance of the seam. Under Analysis 1, 94. Jo 96. the ancestral state for E. lancensis + Xenosaurus and for E. serratus + Xeno- saurus is ambiguous, and late disappear- ance is an uepOn eee of M. ornatus. Under Analysis 2 2. the ancestral state for Xenosauridae is ambiguous. Frontal: Constriction (0) weak, ratio of widest mediolateral width of frontal table anterior to level of greatest con- striction to width at greatest constriction less than 1.15 (Fig. 25D); (1) moderate, ratio 1.15 to 1.7; (2) strong, ratio greater than 1.7 (Fig. 26A). Variation. In many taxa, including S. crocodilurus, the frontals in younger individuals are considerably more con- stricted than in older individuals, con- comitant with the relatively larger size of the eyes. Individuals scored here were relatively large adults. Evolution. Under both analyses, the ancestral state for the entire group and for Anguimorpha is moderate constric- tion, and weak constriction is an auta- pomorphy of S. crocodilurus. Under Analysis 1, weak constriction is a synap- omorphy of Varanidae and of Anguidae + Helodermatidae, and strong constric- tion is a synapomorphy of E. serratus + Xenosaurus and an autapomorphy of M. ornatus. Under Analysis 2, weak con- striction is a synapomorphy of Anguidae + Varanoidea, and the ancestral state for Xenosauridae is ambiguous between weak and strong constriction. Frontal: Osteoderms (0) lacking keels or with weak keels (Fig. 26A); (1) strongly keeled, at least in part (Fig. 26B). Variation. Osteoderms fuse to the frontal postnatally in most taxa, but the keel may be the first pari serjeaeh osteoderm to form. Evolution. Under both analyses, the keeled morphology is a synapomorphy of the northern clade of Xenosaurus. Frontal: Orbital rows of osteoderms along lateral edges (0) flat and rectan- XENOSAUR PHYLOGENY © Bhullar 113 Figure 26. A, Frontal of Xenosaurus grandis NAUQSP-JIM 1460, dorsal, anterior to the left. Illustrates characters 94(2), 95(0), 102(1), 103(1). B, Frontal of Xenosaurus newmanorum NAUQSP-JIM uncatalogued specimen, dorsal, anterior to the left. Illustrates characters 95(1), 96(1), 97(3), 98(1), 99(1), 100(0), 102(0), 103(0), 108(1). C, Frontal of Restes rugosus YPM PU 14640, dorsal, anterior to the left. Illustrates characters 96(0), 97(2), 98(0), 99(0). D, Frontal of Exostinus serratus, CT scan of AMNH 1608, dorsal, anterior to the left. Illustrates characters 98(2), 100(1), 108(2). OG gular (Fig. 26C); (1) small and domed (Fig. : 26B). Variation. Osteoderms fuse to the frontal postnatally in most taxa. Evolution. Under both analyses, the ancestral state for the entire group is ambiguous and the ancestral state for Anguimorph ais flat and rectangular. The domed morphology i isa synapomorphy of E. lancensis + Xenosaurus, and the flat, rectangular morphology is an autapomor- phy of y rectocollaris. Frontal: Osteoderms in center of expand- ed posterior portion of frontal (0) all flat ee like (Fig. 25D); (1) some flat and plate-like, others broad and domed (Klem- bara, 2008); (2) some flat and plate-like, C): (3 ) all others small and domed (F . o, 26 ig. small and domed (Fig. 26B). Variation. ae es become fused to the frontal and become more domed during ontogeny. Evolution. Under both analyses, the ancestral state for the entire group is ambiguous among all three states, and that fe Anguimorpha is ambiguous be- tween state 1 and state 2. State 2 is ancestral for R. rugosus + Xenosaurus, and state 3 is a synapomorphy of E. lancensis + Xenosaurus and an autapomorphy of B. ammoskius. Under Analysis 1, the ances- tral states for S. crocodilurus + Varanidae and for Xenosaurus + Anguidae are ambiguous like Anguimorpha. State 0 is 114 DS). 100. Bulletin of the Museum of Comparative Zoology, Vol. 160, No. 3 ancestral for Anguidae + the heloderma- tid lineage. Under Analysis 2, the ances- tral state for Xenosauridae is ambiguous like Anguimorpha, and the ancestr al state for Anguidae + Varanoidea is state 0. Frontal: Region of strong development of orbital rows of osteoderms along lateral edges (0) restricted, extending anteriorly von up to greatest Sone tion of frontal table or “ending posterior tom (Fie, 26C) Ch) intermediate ex- tending anteriorly past greatest con- striction of frontal table anil for less than a third of the length of the prefrontal facet (Fig. 26B); (2) exten- sive, extending anteriorly past greatest constriction of frontal table and for about a third of the length of the prefrontal facet (Fig. 26D). Variation. Osteoderms fuse to the frontal during postnatal ontogeny. Evolution. Under both analyses, the ancestral state for the entire group is a restricted extent. The intermediate morphology is a synapomorphy of S. crocodilurus + B. ammoskius and of E. serratus + Xenosaurus, and the exten- sive morphology is an autapomorphy of E.serratus. Frontal: Transverse rows of osteoderms in posterior expanded portion (0) two to three (Fig. 26C); (1) four or more (Fig. 26B). Evolution. Under both analyses, the ancestral state for the entire group is ambiguous, and four or more rows is a synapomorphy of E. serratus + Xenosaurus. Frontal: Posteriormost transverse row of osteoderms in posterior expanded portion (0) similar in prominence to more anterior rows (Fig. 26B); (1) less prominent than more anterior rows (Fig. 26D). Variation. No variation affecting scor- ing was evident, save, most likely, the ontogenetic variation already noted in pateotlen m characters. Gis 103. 104. Evolution. Under both analyses, the reduced prominence of the posterior- most row is either a synapomorphy of E. lancensis + Xenosaurus, with similar prominence a synapomorphy of Xeno- SQUTUS, OF It 1S aM autapomorphy Gt Es lancensis and of E. serratus. Frontal: Distance from posterior end of prefrontal facet to posterior end of frontal (0) relatively great, 2.25 or more times length of posterior sharply taper- ing portion of prefrontal facet (Fig. 27A); (1) relatively small, fewer than 2.25 times length of posterior portion of facet (Fig. 27B). Evolution. Under both analyses, a relatively small distance is an autapo- morphy of E. serratus. Under Analysis 1, that character state is also an autapomorphy of Helodermatidae and a synapomorphy of Varanidae. Under Analysis 2, it is a synapomorphy of Varanoidea. .Frontal: Anterior mediolaterally ta- pered tip of frontal (0) mediolaterally wider at base than anteroposteriorly long (Fig. 26B); (1) longer than wide (Fig. 26A). Evolution. Under both analyses, lon- ger than wide is a synapomorphy of the southern clade of Xenosaurus. Frontal: Internasal spine (0) closely approaching anterior tip of frontal (Fig. 26B); (1) halting posterior to anterior tip with frontal extending beyond it for half or more of the length of the spine, forming an extensive anterior lamina (Fig. 26A). Evolution. Under both analyses, a posterior termination of the spine is a syn- apomorphy of X. rackhami + X. grandis. Frontal: Posteriorly, medial and lateral edges of nasal facets converge (0) at a relatively wide angle of 70° or greater (Fig. 27C); (1) at a relatively narrow angle of less than 70° (Fig. 25D). XENOSAUR PHYLOGENY © Bhullar EAs Figure 27. A, Frontal of Xenosaurus newmanorum NAUQSP-JIM uncatalogued specimen, left lateral, anterior to the left. Illustrates characters 101(0), 106(0), 107(0). B, Frontal of Exostinus serratus, CT scan of AMNH 1608, left lateral, anterior to the left. Illustrates characters 101(1), 106(1), 107(1). C, Anterior tip of frontal of Exostinus serratus, CT scan of USNM v16565, dorsal, anterior to the left. Illustrates character 104(0). D, Frontal of Restes rugosus, CT scan of YPM PU 14640, ventral, anterior to the left. Illustrates character 105(0). 105. Evolution. Under Analysis 1, the ancestral states for all clades with mixed character distribution are ambiguous. Under Analysis 2, the ancestral state for the entire group is a relatively narrow angle, as is that for Anguimorpha. The aniceetral state for Xenosauridae is ambiguous. On the other Anguimorpha, the ancestral state for Anguidae + Varanoidea is a relatively narrow angle, with the wider morphol- ogy a synapomorphy of C. enneagram- mus + O. ventralis and an autapomorphy of Lanthonotus borneensis. Frontal: Ventral edges of cristae cranii, at closest approach in a_ horizontal plane, separated by (0) greater than branch of 106. SAUTUS. one-third of the mediolateral width of the frontal at that anteroposterior level (Fig. 27D); (1) one-third or less of the mediolateral width of the frontal (Fig. 28A). Evolution. Under both analyses, wide separation is the primitive state for the entire group and_ for Anguimorpha. Under Analysis 1, close approach is a synapomorphy ar S. crocodilurus + Varanidae and of FE. lancensis + Xeno- Under Analysis 2, close ap- proach is a synapomorphy of Varani- dae, and the ancestral state for Xenosauridae is ambiguous. Frontal: Cristae cranii deepen anterior to expanded portion of frontal to (0) 116 Bulletin of the Museum of Comparative Zoology, Vol. 160, No. 3 Figure 28. A, Frontal of Exostinus serratus, CT scan of AMNH 1608, ventral, anterior to the left. Illustrates character 105(1). Left palpebrals, dorsal, anterior to the top: B, Elgaria multicarinata TMM-M 8958; C, Xenosaurus newmanorum NAUQSP-JIM uncatalogued specimen; D, Xenosaurus agrenon UTACV 145008; E, Xenosaurus grandis NAUQSP-JIM 1460. B through D illustrate characters 109(0), 109(1), 110(0), 110(1), 111(0), 111(1), 112(0), 112(1), 113(0), 113(1). F, Rostrum of Elgaria multicarinata CAS 85234, left lateral, anterior to the left. Illustrates character 114(0). less than twice their dorsoventral portion of ventral edges (0) less than 40° height along the expanded portion (Fig. 27A); (1) 40° or greater (Fig. 27B). (Fig. QTA): (T) equal to or greater than twice their dorsoventral height along the expanded portion (Fig. 27B). Evolution. Under both analyses, a steep slope is an autapomorphy of E. serratus. Under Analysis la ‘Sheep Evolution. Under both analyses, shal slope is a synapomorphy of Varanidae low cristae cranii are ancestral for the and an autapomorphy of Helodermati- entire group and for Anguimorpha, and dae. Under Analysis 2, it is a synapo- the deep morphology is an autapomor- morphy of Varanoidea. phy of E. lancensis. Under Analysis 1, the deep morphology is a synapomor- phy of Varanidae and of Anguidae + Helodermatidae. Under Analysis 2 2. the deep morphology is a synapomorphy of Anguidae + Varanoidea. 108. Frontal: Cristae cranii project laterally beyond dermal table of frontal (0) not at all (Fig. 25D); (1) only in region of greatest constriction (Fig. 26A); (2) extensively beginning in region of greatest constriction and extending 107. Frontal: Angle of posterior descent of nearly to anterior tip of frontal cristae cranii just behind horizontal (Fig. 96D). Evolution. Under both analyses, no projection is the ancestral state for the entire group and for Anguimorpha. Moderate projection is a synapomorphy of E. serratus + Xenosaurus, and exten- sive projection is an autapomorphy of E. serratus and a synapomorphy of Varanus. Under Analysis 1, moderate projection is a synapomorphy of Varanidae and an a epomorphy of Helodermatidae. Un- der Analysis 2, moderate projection is a synapomorphy of Varanoidea. Palpebral The palpebral is unknown for M. ornatus, B. ammoskius, E. lancensis, and E. serratus. It is absent in P. torquatus and Helodermatidae and is so reduced in L. bormeensis that its mor- phology in relation to the characters described here is unscorable (Maisano et al., 2002). 109. 110. Ie Palpebral: Overall shape: (0) mediolat- erally elongate triangle with lateral constriction caused by step in posterior margin (Fig. 28B); (1) mediolaterally shorter triangle without lateral constric- tion caused by step in posterior margin (Fig. 28C). Evolution. Under both analyses, the mediolaterally elongate morphology is a synapomorphy of Anguimorpha, and the more equilateral morphology is a syn- apomorphy of R. rugosus + Xenosaurus and an autapomorphy of O. ventralis. Palpebral: Posterior edge (0) with relatively straight or smoothly curving margin (Fig. 28E); (1) with wavy mar- gin (Fig. 28D); (2) with ragged margin (Fig. 28C). Evolution. Under both analyses, the wavy morphology is an autapomorphy of X. agrenon, and the ragged morphol- ogy is a synapomorphy of the northern clade of Xenosaurus. Palpebral: Fused osteoderms (0) absent (Fig. 28B); (1) present as slight dorsal rugosities (Fig. 28E); (2) present across most of dorsal surface; less coverage LS: XENOSAUR PHYLOGENY © Bhullar 117 upon anterior two-thirds and along anterior edge (Fig. 28D); (3) strong across dorsal surface, including anterior portion, with distinct, tall row along anterior edge (Fig. 28C). Variation. Osteodermal fusion to dermal elements, including the palpe- bral, occurs postnatally, although rela- tively small specimens of X. platyceps already show some dorsal rugosity. Evolution. Under both analyses, a slight dorsal rugosity is a synapomorphy of R. rugosus + Xenosaurus, a moder- ately strong COV ering of osteoderms is a synapomorphy of Xenosaurus, and a strong covering is a synapomorphy of the pOnLiCa olde of Xenosaurus. Palpebral: Foramen near anterior edge, just lateral to mediolateral level of apex of slight concavity in posterior margin (0) absent (Fig. 98C): (1 ) present (Fig, { I8E). Evolution. Under both analyses, presence of the foramen is a synapo- morphy of X. rackhami + X. grandis. Palpebral: Strong s-curve to medial edge in horizontal plane, with anterior emargi- nation and posterior bulge accompanied by dorsoventral deepening (0) absent (Fig. 28B); (1) present (Fig. 28D). Variation. In some large adult E. multicarinata, the medial edge of the palpebral displays a slight s-curve, but not to the extent of the faa scored as (1). Evolution. Under both analyses, the ancestral state for the entire group is ambiguous, as is that for Anguimorpha. Under Analysis 1, the ancestral state for R. rugosus + Xenosaurus is presence of the s-curve, with all mixed nodes ambiguous. Under Analysis 2, the ancestral state for Xenosauridae is presence of an s-curve, and that for Anguidae + Varanoidea is absence. Lacrimal The lacrimal is unknown for all of the extinct taxa save B. ammoskius, but some of 118 Bulletin of the Museum of Comparative Zoology, Vol. 160, No. 3 Figure 29. NAUQSP-JIM uncatalogued specimen; C, Xenosaurus agrenon UTACV 145008; D, Xenosaurus grandis NAUQSP-JIM 1460. A through D illustrate characters 115(0), 115(1), 116(0), 116(1), 116(2), 116(3), 117(0), 117(1), 120(0), 120(1), 121(0), 121(1), 122(0), 122(1). Left lacrimals, medial, anterior to the right: E, Xenosaurus platyceps UF 45622; F, Xenosaurus grandis NAQSP- JIM 1460; E and F illustrate characters 119(0), 119(1), 122(0), 122(1). its basic dimensions can be inferred from the morphology of the maxilla and prefron- tal. This is the method by which character 114 is scored for all taxa. 114. ES. Lacrimal: Size relative to maxilla (0) small, around one-quarter length of maxilla or smaller (Fig. 28F); (1) large, around one- third length of maxilla (Fig. 1C). Evolution. Under both analyses, a relatively large lacrimal is a synapomor- phy of R. rugosus + Xenosaurus. Lacrimal: Angle between antorbital and suborbital rami measured along lower margin (0) high, greater than 135 (Fig. 29A); (1) low, less than 135 (Fig. 29C). IG: Left lacrimals, lateral, anterior to the left: A, Elgaria multicarinata TMM-M 8958; B, Xenosaurus newmanorum Evolution. Under both analyses, an angle of less than 135° 1s a synapomor- phy of X. rackhami + X. grandis. Lacrimal: Fused osteoderms (0) absent (Fig. 29A); (1) present as slight rugos- ities (Fig. 29C); (2) present as low mounds (Fig. 29B); (3) present as tall mounds, some of which bear keels (Fig, 29D). Variation. Osteoderms become fused to dermal elements postnatally. Evolution. Under both analyses, the presence of fused osteoderms as slight rugosities is a sy) napomorphy of ee. saurus, low mounds is an autapomorphy of X. grandis, and tall, sometimes keeled IN IS: I). 1 ] ) lle mounds is a sy napomorphy of the north- ern clade of Xenosaurus. Lacrimal: Subpalpebral fossa (0) absent (Fig. 29A); (1) present (Fig. 29B). Evolution. Under Analysis 1, the fossa is a synapomorphy of Xenosaurus and of S. crocodilurus + B. ammoskius. Under Analysis 2, it is a synapomorphy of Xenosauridae. Lacrimal: Subpalpebral fossa (0) con- tained entirely in lacrimal (Fig. 1C); (1) extending onto adjacent elements (Fig. 20B). Evolution. Under both analyses, most nodes optimize ambiguously. The ancestral state for the southern clade of Xenosaurus is containment within the lacrimal. Lacrimal: Lacrimal foramen (0) large, expanse at greatest extent is half or more of length of antorbital ramus of lacrimal (Fig. S9F ): (1) small, less than half as extensive as antorbital ramus (Fig. 29E). Evolution. Under both analyses, the small, constricted morphology is an autapomorphy of X. platyceps. .Lacrimal: Dorsal and ventral prefrontal processes (0) project as prominent wedges beyond main body of lacrimal in sagittal plane (Fig. 29A); (1) barely project if at all bevonttvao of lacrimal in sagittal plane (Fig. 29B). Evolution. Under both analyses, lack of projection or minor projection is ancestral for the entire group, and projection is an autapomorphy of X. newmanorum, E. multicarinata, and X. grandis. Under Analysis 1, projection is an autapomorphy of ieledesmatidac and a synapomorphy of Varanidae. Under Analysis TENORS) syhapomorphy of Varanoidea. Lacrimal: Lacrimal canal (0) deeply incised into medial surface of lacri- mal, with strong overhanging and/or underhanging ridges, the inter formed XENOSAUR PHYLOGENY © Bhullar 119 by flange of ventral prefrontal process (Fig. 29A); (1) shallowly incised, with marginal ridges only in anteriormost portion (Fig. 29B). Evolution. Under both analyses, the ancestral state for the entire group is ambiguous. Under Analysis 1, shallow incision is ancestral for “Anguimorpha, and deep incision is a synapomorphy of Anguidae. Under Analysis 2, the ances- eral state for Anguimorpha is ambigu- ous, as are those for Anguidae. + Varanoidea and Varanoidea. -Lacrimal: Posteriorly, ventral prefrontal I process (0) diminishes mediolatet ‘ally (Fig. 29D); (1) expands mediolaterally (Fig. 298). Evolution. Under both analyses, me diolateral expansion along a posterior cline is a synapomorphy of hie northern clade of Xenosaurus. Jugal The jugal is unknown for M. ornatus. 123. Jugal: Postorbital ramus (0) long axis iis 4. relatively straight (Fig. 3C); (1 ) long axis broadly curved in sagittal plane (Fig. 5C). Evolution. Under both analyses, sag- ittal curvature is an autapomorphy of X rectocollaris. Jugal: Fused osteoderms or osteoder- il sculpturing (0) absent (Fig. 30A); (1) present on ventral two- thirds of postorbital ramus (Fig. 30B); (2) pres- ent on all of postorbital ramus (Fig. 30C). This character was scored as presence or absence of sculpturing by Estes et al. (1988). Variation. Osteoderms fuse to the jugal during early postnatal ontogeny. Evolution. Under both analyses, the ancestral state for the entire group and for Anguimorpha is a lack of sculptur- ing or fused osteoderms and that for R. rugosus + Xenosaurus is partial cover- 120 Bulletin of the Museum of Comparative Zoology, Vol. 160, No. 3 Figure 30. A, Left jugal of Elgaria multicarinata TMM-M 8958, lateral, anterior to the left. Illustrates character 124(0). B, Right jugal of Restes rugosus YPM PU 14640, lateral, anterior to the right. Illustrates characters 124(1), 125(0), 132(2), 133(0). C, Left jugal of Exostinus serratus, CT scan of AMNH 1608, lateral, anterior to the left. Illustrates characters 124(2), 125(1), 132(1). D, Left jugal of Xenosaurus newmanorum NAUQSP-JIM uncatalogued specimen, medial, anterior to the right. Illustrates characters 126(0), 128(1). 125. age. Having osteoderms upon the entire dermal surface is a synapomor- phy of E. serratus + Xenosaurus. Under Analysis 1, partial coverage is an autapomorphy of S. crocodilurus. The ancestral state for Xenosaurus + Angu- idae and for Anguidae + Helodermati- dae is ambiguous between absence and partial coverage. Under Analysis 2, partial coverage is an autapomorphy of Helodermatidae; the ancestral state for Xenosauridae and for S. crocodilurus + M. ornatus is ambiguous between ab- sence and partial coverage. Jugal: Fused osteoderms (0) relatively unconsolidated, vermiculate plate with a few mounds (Fig. 30B); (1) mostly 1 9) nd consolidated into discrete osteoderms (Fig. 30C). Variation. Osteoderms fuse to the jugal during early postnatal development. Evolution. Under both analyses, all mixed internal nodes are ambiguous regarding ancestral states. 6. Jugal: Ridge between orbital and ad- ductor surfaces (0) at or posterior to midline of postorbital ramus of jugal (perpendicular to its long axis), orbital surface relatively sagittally oriented (Fig. 30D); (1) anterior to midline, orbital surface relatively transversely oriented (Fig. 31A). XENOSAUR PHYLOGENY © Bhullar 12] Figure 31. 129(1), 131(0). 129(0), 130(0). Evolution. Under both analyses, the ancestral state for the entire group is a location at or posterior to the midline. The ancestral states for the southern clade of Xenosaurus and for S. croco- dilurus + M. ornatus are ambiguous and a location anterior to the midline is an autapomorphy of Helodermatidae. . Jugal: Base of postorbital ramus (0) less extensive perpendicular to its long axis than base of suborbital ramus or approximately as extensive (Fig. 31A); (1) more extensive than base of subor- bital ramus (Fig. 31B). Evolution. Under both analyses, the more extensive morphology is an autapo- morphy of X. platyceps and of X. grandis. 128. 129. A, Left jugal of Xenosaurus agrenon UTACV 145008, medial, anterior to the right. Illustrates characters 126(1), 127(0). B, Left jugal of Xenosaurus platyceps UF 45622, medial, anterior to the right. Illustrates characters 127(1), 130(1), 131(1). C, Left jugal of Exostinus serratus, CT scan of AMNH 1608, medial, anterior to the right. Illustrates characters 126(0), 128(0), D, Left jugal of Shinisaurus crocodilurus UF 72805, medial, anterior to the right. Illustrates characters Jugal: Adductor surface (0) expands ventrally (Fig. 31C); (1) diminishes ventrally, resulting in a longer quadra- tojugal process (Fig. 30D). Evolution. Under both analyses, ven- tral diminishment is a synapomorphy of Xenosaurus. Jugal: Foramen piercing table at junc- tion of maxillary, orbital, and adductor surfaces (0) absent (Fig. 31D); (1) present (Fig. SUG): Evolution. Under both analyses, the ancestral states for the entire group and for Anguimorpha are absence of the foramen. Under Analysis 1, the ancestral state for S. crocodilurus + Varanidae is 130. eile Bulletin of the Museum of Comparative Zoology, Vol. 160, No. 3 absence of the foramen; that for Anguidae + Xenosaurus is ambiguous, as is that for Helodermatidae + ‘Angu- idae and all nodes therein. Wades Analysis 2, presence of f the foramen is a synapomorphy of E. serratus + Xenosaurus and of C. enneagrammus + O. ventralis. Jugal: Foramen piercing adductor sur- face just posterior to table at junction of maxillary, orbital, and adductor surfac- es (0) \ absent (Fig. 31D); (1) present (Fig. 31B). 7 Evolution. Under both analyses, the ancestral states for the entire group and for Anguimorpha are ambiguous. Un- der Analysis 1, the aniCostale state for S. crocodilurus + Varanidae and for An- guidae + Helodermatidae is absence of the foramen. Under Analysis 2, the ancestral state for Anguidae + Varanoi- dea is absence, and that for Xenosaur- idae is ambiguous. Jugal: Anterior expansion at meeting of anterior and dorsal edges of postorbital ramus (0) absent (Fig. 31C); (1) pres- ent (Fig. 31B). Evolution. Under both analyses, the ancestral states for R. rugosus + Xeno- saurus and E. lancensis + Xenosaurus were ambiguous. The ancestral state for Xenosaurus is expansion. Lack of ex- pansion is an autapomorphy Oi OG rectocollaris. 2.Jugal: Tip of postorbital ramus (0) without strong change in sagittal angle at level of postorbital facet (Estes et al., 1988, Fig. 13B); (1) moderately project- ed posteriorly at beginning of level of postor bital facet by less than 20° to long axis of majority of postorbital ramus (Fig. 30€): (2 ) strongly projected poste- riorly beginning at level of postorbital facet by 20° or more to long axis of majority of postorbital ramus (Fig. 30B). Evolution. Under both analyses, the ancestral state for the entire group and for Anguimorpha is moderate posterior projection, as is that for shinisaurs. The ancestral state for Varanidae is lack of projection, and that for R. rugosus + Xenosaurus is ambiguous between moderate and strong projection. Under Analysis 1, the ancestral state for S. euaread inns + Varanidae is moderate projection, as is that for Xenosaurus + Anguidae. The ancestral state for An- guidae + Helodermatidae and for Anguidae is ambiguous between no al moderate projection. Under Anal- ysis 2, the ancestral state for Angui- morpha, Anguidae + Varanoidea, and Xenosauridae is moderate projection. A lack of projection is an autapomorphy of C. enneagrammus and a synapomor- phy of Var: nnodes 133. Jugal: Posterior edge of postorbital ramus (Q) straight or smoothly curved (Fig. 30B); (1) abruptly angled toward the vertical at level of postorbital facet (Hig, 3€): Evolution. Under both analyses, abrupt vertical an Sune is a synapo- morphy of X. rackhami + X. grandis. Postorbital/Postfrontal The postorbital and postfrontal are un- known in M. ornatus, E. serratus, and E. lancensis. The postorbital is nko in B. ammoskius. Only the postfrontal ossifies in Helodermatidae and L. borneensis. 134. Postorbital/Postfrontal: Postorbital and postfrontal (0) remain separate well after hatching/birth or throughout on- togeny (Fig. 32A); (1) fuse in late prenatal or early postnatal ontogeny (Fig. 32B) (see character 14 of Estes et al., 1988). Variation. As noted, this character involves the assessment of the ontoge- netic stage of a specimen. Justifications for age assessments of the fossils were given in Materials and Methods. Evolution. Under both analyses, the ancestral state for the entire group XENOSAUR PHYLOGENY ¢ Bhullar 123 136(1) Figure 32. Skulls, dorsal, anterior to the left: A, Xenosaurus newmanorum, CT scan of UMMZ 126056; B, Xenosaurus rackhami, CT scan of UTEP 4555. A and B illustrate characters 134(0), 134(1), 136(1), 138(0), 139(1), 139(2), 140(0), 140(1), 141(0), 141(1), 142(1), 165(1), 165(3), 166(0), 166(1). and for Anguimorpha is lack of fusion, State O is a synapomorphy ot xX. and early fusion is a synapomorphy of agrenon + X. rectocollaris and an the southern clade of Xenosaurus. autapomorphy of X. platyceps. Under Analysis 1, the ancestral state : Y x F . “ yA ED) , * ate aa = P i! egitg ie FIG reais! 4 WaranGides js 136. Postorbital/Postfrontal: Dorsal ridge ambiguous. Under Analysis 2, early adjacent to lateral edge of postorbital 2 ° ‘ . x 7 7 z ( ‘ NS ‘4 29 pA . rAca fusion is a synapomorphy of Varanus (0) ug (Fig. 335A); (1) present and an autapomorphy of S. crocodilurus. (Hig. 325). Evolution. Under both analyses, ab- sence of a ridge is the ancestral state for the entire group and for Anguimorpha. Under Analysis 1, presence of a ridge is 135. Postorbital/Postfrontal: Angle of post- frontal “clasp” of frontoparietal suture (taken between long axes of anterior and posterior processes of postfrontal) et AER Ten ae (0) less than 80° (Fig. 5A); (1) 80° to a meals of sat anc = Be sneer co ae) toreater than autapomorp Mi of S. crocodilurus. Un- 85° (Fig, 3A). : der Analysis 2, the ancestral state for Xenosauridae is ambiguous. Evolution. Under both analyses, state 2 is ancestral for the entire group. The . Postorbital/Postfrontal: Postorbital (0) ancestral state for Xenosaurus is am- relatively flat, lying largely in horizontal biguous between state 2 and state 1. plane (Fig. 33A); (1) bearing deep — ow) ~l 124 Bulletin of the Museum of Comparative Zoology, Vol. 160, No. 3 Figure 33. A, Temporal region of E/garia multicarinata CAS 85234, dorsolateral, anterior to the left. Illustrates characters 136(0), 137(0), 139(0), 142(0). B, Left postorbital of Xenosaurus newmanorum NAUQSP-JIM uncatalogued specimen, dorsolateral, anterior to the left. Illustrates character 137(1). C, Left pterygoid of Shinisaurus crocodilurus UF 72805, ventral, anterior to the left. Illustrates characters 143(0), 145(0). D, Left pterygoid of Xenosaurus newmanorum NAUQSP-JIM uncatalogued specimen, ventral, anterior to the left. Illustrates characters 143(1), 145(2). flange descending from lateral edge, inflection of medial edge in horizontal creating a epraees in the sagittal plane, plane (Fig. 32B). perpendicular to the main body of the Evolution. Under both analyses, the bone (Fig. 33B). ancestral state for the entire group and Evolution. Under both analyses, the for Anguimorpha is pointed, and the flat morphology is ancestral for the squared-off morphology is a synapo- entire group and for Anguimorpha. morphy of the southern clade of Under Analysis 1, presence of a flange Xenosaurus and an autapomorphy of sma synapomorphy of Xenosaurus and S. crocodilurus. Under Analysis 1, the an autapemenpny of S. crocodilurus. squared-off morphology is a synapo- Under Analysis 2, the ancestral state for morphy of Varanus and an autapomor- Xenosauridae is ambiguous. phy of Helodermatidae. Under Analysis 138. Postorbital/Postfrontal: Tip of postor- Pike ancestral stale tor Varies a ¢ guoUs. bital process of postfrontal (0) pointed, formed by meeting of smoothly curv- 139. Postorbital/Postfrontal: Narrowness_ of ing medial and lateral edges (Fig. 32A): postorbital measured by ratio of antero- (1) squared-off because of posterior posterior length to mediolateral width 140. just posterior to divergence of posterior process of ae (0) 4.0 or greater (Fig. 33A); (1) between 2.5 aa 4.0 (Fig, 32A); (2) 2.5 or less (Fig. 32B). Evolution. Under both analyses, the ancestral state of the entire group and of Anguimorpha is the narrow mor- phology, and the wide morphology is a synapomorphy of X. rackhami a xX grandis. Under Analysis 1, intermediate width is a synapomorphy of Xenosaurus and an autapomorphy of S. crocodi- lurus. Under Analysis 2, the ancestral state for Xenosauridae is ambiguous between the narrow and intermediate morphologies. Postorbital/Postfrontal: Posterior end of squamosal process of postorbital (0) pointed (Fig. 32A); (1) rounded (Fig. 32B). Evolution. Under both analyses, the rounded morphology is an autapomor- phy of X. rackhami. 143. 144. 145. XENOSAUR PHYLOGENY ¢ Bhullar 125 Pterygoid: Margin of pterygoid border- ing infraorbital fenestra (0) smoothly eied (Fig. 33C); (1) bearing small eminence just medial of posterior apex (Fig. 33D). Evolution. Under both analyses, presence of an eminence is a synapo- morphy of the northern clade of Xenosaurus and an autapomorphy of Helodermatidae. Pterygoid: Medial and lateral edges of vomerine process posterior to oblique anterior idee (0) weakly divergent (by 20° or less) in horizontal plane (Fig. 34A); (1) strongly divergent (by greater than 20°) in horizontal plane (Fig, 34B). Evolution. Under both analyses, strong divergence is an autapomorphy of R. rugosus and of C. enneagrammus. Pterygoid: Bears (0) large row or patch of teeth (Fig. 33C); (1) one or two small teeth, sometimes bilaterally asymmetri- 141. Postorbital/Postfrontal: Posterior end of cal (Fig. 34B); (2) no teeth ( Fig. 83D), squamosal process of postorbital (0) Presence or absence of pterygoid teeth gently curved medially in horizontal was character 83 of Estes et al. (1988). ae or not curved medially Variation. Pterygoid teeth often (Fig. 32A); (1) strongly curved medially increase in number with age (personal mM horizontal plane, resulting in sharp observation), and observations on this change in angle of medial edge character were made using relatively (Fig. 32B). large individuals. Additionally, when Evolution. Under both analyses, tecth are highly reduced, their presence strong curvature is a synapomorphy of can vary Fon side to side in an X. rackhami + X. grandis. individual or from individual to individ- 142. Postorbital/Postfrontal: Lateral edge of pee pee ee ae postorbital (0) straight or nearly straight Evolution. Under both analyses, re- in horizontal plane (Fig. 33A); (1) duction of eee dentition is a broadly acd in horizontal plane synapomorphy of Xenosaurus + R. (Fig. 32B). rugosus and an autapomorphy of C. Teepe Miadee bee analyses, enneagrammus. Absence of pterygoid arene luceste Ulan onncaiais 4s denteot is’a SEDOmQp ay of Xeno- ambiguous. saurus and of Varanus. Pterygoid Ectopterygoid The pterygoid is unknown in E. lancensis The ectopterygoid is unknown for all extinct taxa in the study (not sufficiently exposed in B. ammoskius). and E. serratus and is not substantially visible in B. ammoskius. 126 Bulletin of the Museum of Comparative Zoology, Vol. 160, No. 3 Figure 34. Illustrates character 144(0). Right pterygoids, ventral, anterior to the left: A, Xenosaurus newmanorum NAUQSP-JIM uncatalogued specimen. B, Restes rugosus, CT scan of YPM PU 14640. Illustrates characters 144(1), 145(1). Left ectopterygoids, maxillary articulation surfaces, anterior to the left: C, Elgaria multicarinata TMM-M 8958; D, Xenosaurus newmanorum NAUQSP-JIM uncatalogued specimen. C and D illustrate characters 146(0), 146(1). E, Parietal of Xenosaurus rackhami UTEP-OC “MALB” 388, dorsal, anterior to the left. Illustrates characters 147 (n for numerator, d for denominator), 150(0), 151(1), 153(1), 154(1), 156(1). 146. Ectopteryg goid: Prominent descending projection of ae process (0) absent (Fig. 34C); (1) present (Fig. 34D). Evolution. Gane both analyses, presence of a descending projection is a synapomorphy of Xenosaurus. Parietal The parietal is unknown in R. rugosus and Ee serratus. 147. Parietal: Ratio of anteroposterior length along midline (to apex of meeting of supratemporal processes) to mediolat- eral width at frontoparietal suture (Fig. 34E) (0) less than 0.70; (1) 0.70 to less than 0.75; (2) 0.75 to less than (0.80: (3) 0.80 to less than 0.85; (4) 0.85 i less than 0.90; (5) 0.90 to less than 0.95: (6) 0.95 to less than 1.00: (7) 1.00 to less than 1.05; (8) 1.05 to less than 1.10: (9), 1.10 to less than 1:152A) as to less than 1.20; (B) 1.20 to less than ].25; (C) 1.25 or greater. Variation. The length-to-width ratio of the parietal body increases with growth, especially in the earlier parts of ontogeny (personal observation). Thus, care must be taken “to: use relatively large adult individuals in scoring this character. Evolution. Under both analyses, the ancestral state for the entire group is ambiguous among state 1, state 2, state 148. 3, and state 4. The ancestral state for Anguimorpha is ambiguous between state 3 and state 4. Additionally, the ancestral state for S. crocodilurus + M. ornatus is state 5. State 6 is a synapo- morphy of S. crocodilurus + B. ammos- kius, and state 9 is an autapomorphy of S. crocodilurus. The ancestral state for Anguidae is ambiguous between state 3 and state 4. State 7 is an autapomorphy of O. ventralis. State C is an autapo- morphy of E. lancensis. The ancestral state for Xenosaurus is ambiguous between state 3 and state 4. A reduced ratio is a synapomorphy of the southern clade of Xenosaurus, whose ancestral state is ambiguous between state 1 and state 2. Giniel 0 is a synapomorphy of X. agrenon + X. rectocollaris. State 8 is an autapomorphy of L. borneensis, and state 0 is an autapomorphy of V. exanthematicus. Under Analysis 1, the ancestral state for crocodilurus + Varanidae is ambiguous between state 4 and state 5, and that for Varanus is state 4. The ancestral state for Xeno- saurus + Anguidae and for Anguidae + Helodermatidae is ambiguous between state 3 and state 4. Under Analysis 2, the ancestral state for Anguidae + Varanoidea is ambiguous between state 3 and state 4, as is that for Varanoidea and all nodes therein. The ancestral state for Xenosauridae is ambiguous among state 3, state 4, and state 5, as is that for Xenosaurus + E. lancensis. Parietal: Ratio of anteroposterior length of supratemporal processes beginning at apex of their meeting to medielateral width at widest separation of processes (Fig. 35A) (0) less than 0.20; (1) 0.20 to ae than 0.25; (2) 0.25 to less than 0.30: (3) 0.30 to less than 0.35; (4) 0.35 to less than 0.40; (5) 0.40 to less than 0.45: (6) O45 to less, than 0.50.47). 0:50 or greater. Variation. The posterior portion of the parietal becomes relatively more elongate with growth, especially in 149. XENOSAUR PHYLOGENY ¢ Bhullar 1:27 early ontogeny (personal observation), so it is important to score this character on relatively large adults. Evolution. Under both analyses, the ancestral state for Varanidae is state 5, and state 7 is an autapomorphy Ores borneensis. State 7 is a synapomorphy of Anguidae. The ancestral state for S crocodilurus + M. ornatus is state 4, and state 3 is an autapomorphy of S. crocodilurus. Finally, the ancestral state for Xenosaurus is ambiguous between state 2 and state 3. The ancestral state for the northern clade of Xenosaurus is state 2, and state 0 is an autapomorphy of X. newmanorum. In the southern clade, state 1 is an autapomorphy of X. grandis. Under Analysis 1, the ancestral state for the entire group, Anguimorpha, S. croco- dilurus + Varanidae, and Xenosaurus + Anguidae is ambiguous between state 4 arid state 5. Cinder Analysis 2, the ancestral state for the entire group, Anguimorpha, Anguidae + Varanoidea, and Varanoidea is state 5. State 4 is an autapomorphy of Helodermatidae. On the other branch of Anguimorpha, state 4 is a synapomorphy of Xenosauridae. Parietal: Attachment areas for adductor musculature (0) dorsolateral or lateral, without extensive overhanging flange: (1) ventral, roofed over by flange of parietal table (Fig. 35B) (see character 54 of Estes et al., 1988). Variation. Although the overhanging flange in those taxa with ventral origin often becomes relatively more exten- sive with age, I did not observe intraspecies variation that would alter the scoring of the character. Evolution. Under both analyses, the ancestral state for Xenosaurus +. E. lancensis is ventral origin. Under Anal- ysis 1, the ancestral state for the entire group and for Anguimorpha is dorsal origin, and ventral origin is a synapo- morphy of Monosgurins + Anguidae. 128 Bulletin of the Museum of Comparative Zoology, Vol. 160, No. 3 Figure 35. A, Parietal of Xenosaurus agrenon UTACV r45008, dorsal, anterior to the left. Illustrates characters 148 (n for numerator, d for denominator), 150(1), 152(1), 153(0), 155(2), 156(0). B, Parietal of Elgaria multicarinata TMM-M 8958, ventral, anterior to the left. Illustrates characters 149(1), 152(0), 155(0), 157(0). C, Parietal of Xenosaurus platyceps UF 45622, ventral, anterior to the left. Illustrates characters 154(2), 155(1), 157(1). D, Parietal of Xenosaurus newmanorum NAUQSP-JIM uncatalogued specimen, ventral, anterior to the left. Illustrates characters 154(3), 158(1). 150. Note that this is dependent upon the use of an iguanian as the immediate anguimorph outgroup. Most sclero- glossans show ventral origin (Estes et al., 1988). Under Analysis 2. the ancestral state for all mixed nodes is ambiguous. Parietal: Lateral edges forming margins of supratemporal feneeeh ae (0) strongly curved in a horizontal plane (Fig. 34F): (1) weakly curved, especially anterior to apex of temporal emargination (Fig, .35A). Variation. In the taxa showing lateral or dorsolateral origin of the adductor musculature (character 149-0), the 151. degree of emargination increases with age as the Beane becomes relatively Salle and the adductor chamber relatively larger. Evolution. Under both analyses, rel- atively weak curvature is an autapo- morphy of X. agrenon. Parietal: Number of osteoderms on each side lateral to midline osteoderm just posterior to parietal foramen (0) three (Fig. 1A); (1) two (Fig. 34E); (2) one (Gone 2006, fig. 2A). Evolution. Under both analyses, the ancestral state for the entire group is ambiguous between three osteoderms and two osteoderms. The ancestral 153. 154. 2. Parietal: state for shinisaurs is two osteoderms and one osteoderm is an autapomorphy ot M. ornatus. Parietal foramen (0) set far back from frontoparietal suture, con- siderably posterior to anterolateral ex- tensions of parietal bearing postfrontal facets (Fig, 35B); (1) set close to frontoparietal suture, at anteroposterior level of anterolateral extensions bearing postfrontal facets (Fig. 35A). Variation. State 0 reportedly occurs in some S. crocodilurus (J. L. Conrad, personal communication), although not in the specimens examined. Evolution. Under both analyses, the ancestral state for the entire group and for Anguimorpha is ambiguous. Under Analysis 1, the aeceteal state for Xenosaurus + Anguidae and for Helo- dermatidae + Anguidae is ambiguous. The ancestral state for E. Mideiotr inata + O. ventralis is relatively posterior placement. Under Analysis 2, the an- cestral state for S. crocodilurus + Varanoidea is the relatively posterior jlacement, with placement close to the eel suture an autapomorphy of S. crocodilurus. The ancestral state for Anguidae + Helodermatidae is ambiguous. Under Analysis 2, the ancestral state for Anguidae + Varanoi- dea is relatively posterior placement. Parietal: Parietal foramen (0) relatively large (Fig. 35A); (1) small, barely larger than tiny nutrient foramina (Fig. 34E). Evolution. Under both analyses, the small morphology is an autapomorphy of X. rackhami. Parietal: Anterior edge in horizontal plane (0) convex (Conrad, 2006, fig. 2A); (1) relatively straight (Fig. 34). (2) slightly concave (Fig. 35C): (3) strongly concave (Fig. 35D). Evolution. Under both analyses, the ancestral state for the entire group is ambiguous among straight, slightly 155. XENOSAUR PHYLOGENY ¢ Bhullar 129 concave, and strongly concave. The an- cestral state for auinieanie is_ relatively straight and a convex morphology is an aut ipomorphy of B. ammoskius. The ancestral state for Anguidae and xenosaurs is also straight, wake: convex an autapo- morphy of E. lancensis, slightly concave a ea geal of thé noreeea clade of Xenosaurus, and strongly concave an autapomorphy of X. newmanorum. Strongly concave is an autapomorphy of ie omnconss, Under Analysis 1, the relatively straight morphology is ancestral for Anguimorpha, and slightly concave is an autapomorphy of Holodemnandae: Under Analysis 2, the ancestral state for Anguimorpha is ambiguous between rel- atively straight and slightly concave. Parietal: Notch in posterior edge of parietal at meeting of medial edges of supratemporal processes (0) absent (Fig. 35B); (1) weak, more than four times as mediolaterally wide as antero- posteriorly long (Fig. 35C): (2) strong, four times or (eee as macdiolaicrally wade as anteroposteriorly long (Fig. 35A). Evolution. Under both analyses, the ancestral state for the entire group and for Anguimorpha is ambig- uous between weakly notched and strongly notched, as are the ancestral states for Varanidae and Anguidae. The absence of a notch is a synapo- morphy of shinisaurs and an autapo- morphy of E. multicarinata, Heloder- matidae, and V. exanthematicus. The ancestral state for E. lancensis + Xenosaurus is weakly notched. The strongly notched morphology is a synapomorphy of X. agrenon + X. rectocollaris, and aeence of a notch is a synapomorphy of X. rackhami + X. grandis. Under Analysis 1, the ances- tral state for shinisaurs + Varanidae is ambiguous between weakly and strongly notched, as is that for Xeno- saurus + Anguidae. Under Analysis 2, the ancestral state for Xenosauridae is weakly notched. 130 Bulletin of the Museum of Comparative Zoology, Vol. 160, No. 3 158(0) Figure 36. Parietals, ventral, anterior to the left: A, Heloderma suspectum TMM-M 9001. Illustrates character 157(2). B, Xenosaurus grandis NAUQSP-JIM 1460. Illustrates character 158(0). Left supratemporals, lateral, anterior to the left: C, Xenosaurus newmanorum NAUQSP-JIM uncatalogued specimen; D, Xenosaurus grandis NAUQSP-JIM 1460. E, Xenosaurus agrenon UTACV r45008. C through E illustrate characters 159(0), 159(1), 160(0), 160(1), 161(1), 162(1), 163(0), 163(1), 163(2). 156. Parietal: Dorsal fossae a m. articulo- Xenosaurus + Anguidae is the shallow parietalis attachment (0) shallow, only morphology. Winder Analysis 2, the slightly stepped dow n ee parietal ancestral state for Anguimorph a is the tab ale (Fig. 395A); (1) deep, divided from shallow morphology, and the deep mor- parietal fable by sharp ridge along at phology is a synapomorphy of Varanidae. least part of fossa (Fig. 34). —_ ; 157. Parietal: Ventral excavations on supra- _ Variation. The relative depth of the temporal processes for transversospina- fossae increases somewhat with age, lis group muscles (0) widely senile most prominently in the early stages of by sulcus processi ascendenne (Fig. postnatal ontogeny. 35B): (1) moderately separated; Evolution. Under both analyses, the bridged by flange of parietal extending ancestral condition for the entire group is posteriorly past sulcus processi ascen- ambiguous, and the deep morphology is dentis but do not approach each other a synapomorphy of X. rackhami + X. closely, leaving between them a broad grandis and an autapomorphy of C. triangular wedge of cerebral table enneagrammus. Under Analysis 1, the (Fig. 35C): (2) approach each other eeeaal state for Anguimorph a is am- closes leaving between them only a biguous, and the ancestral state for thin ridge of Srebial table (Fig. 36A). Evolution. Under both analyses, widely separated is the ancestral state for the entire group and for Anguimor- pha. The ancestral state for E. lancensis + Xenosaurus is moderately separated, ee that for the southern clade of Xenosaurus is ambiguous between moderately separated and closely ap- proaching. Closely approaching is an ee ey of Helodermatidae. Un- der Analysis 1, the ancestral state for Xenosaurus + Anguidae is ambiguous between closely approaching and ‘mod- erately separated. Under Analysis 2, the ancestral state for eenecaucidae is widely separated, and moderately sep- arated is a synapomorphy of Xeno- saurus + E. lancensis. 158. Parietal: Ventral ridges contacting tae- niae marginales (0) ae bound adductor attachment surface medially for entire length (Fig. 35D); (1 ) diverge just pos- terior to level of prefrontal facet from ridge delineating medial margin of ane attachment surface (Fig. 36B). Evolution. Under both analyses, di vergence is a synapomorphy of the northern clade of Xenosaurus. Supratemporal The supratemporal is unknown for all extinct taxa save B. ammoskius, in which a negligible amount of its morphology is visible. 159. Supratemporal: Anterior edge (0) dor- soventrally short, less than half greatest height of supratemporal (Fig. 36D); (1) dorsoventrally tall, more than half greatest dorsoventral height of supra- temporal (Fig. 36C). Evolution. Under both analyses, the ancestral state for the entire group and for Anguimorpha is the short morphol- ogy. Under Analysis 1, the ancestral state for Xenosaurus is ambiguous, and the tall morphology is an autapomorphy of S. crocodilurus. Under Analysis 2, the tall morphology is a synapomorphy of Xenosauridae and the short mor- 160. 161. 162. 163: XENOSAUR PHYLOGENY ¢ Bhullar 13] phology is a synapomorphy Ob rackhami + X. grandis. Supratemporal: Foramen piercing lat- eral surface near ridge dividing squa- mosal facet and adductor attachment surface about two-thirds of the i to posterior end of supratemporal (0) pre- sent (Fig. 36E); (1) absent (Fig. 36D). Evolution. Under both analyses, the ancestral state for the entire group is ambiguous, and absence of the foramen isa synapomorphy of X. rackhami + X. grandis. Under Analysis 1, the ancestral state for Anguimorpha is ambiguous, and the ancestral state for Xenosaurus + Anguidae is pigsaice of the foramen. Under Analysis 2, the ancestral state for Anguimorpha is presence of the fora- men, and absence is a synapomorphy of Varanidae and of X. Pain + X. grandis. Supratemporal: Foramen piercing lat- eral surtace (0) above (Fig. 37A); (1) below ridge dividing squamosal facet and adductor attachment surface (Fig, 36E). Evolution. Under both analyses, a ventral position is a synapomorphy of Xenosaurus. Supratemporal: Ventral flange wrap- ping under supratemporal process of parietal (0) smoothly curved along ventromedial edge (Fig. 37A); (1) pro- duced into wedge (Fig. 36C). Evolution. Under both analyses, the ancestral state for the entire group and Anguimorpha is smoothly curved. Under Analysis 1, production into a wedge is an autapomorphy of S:. crocodilurus and a synapomorphy of Xenosaurus. Under Analysis 2, produc- tion into a wedge is a synapomorphy of Xenosauridae. Supratemporal: Wedge-shaped ventral flange wrapping under supratemporal process of parietal (0) small and nub- like (Fig. 36E); (1) squat, obtuse wedge 132 Bulletin of the Museum of Comparative Zoology, Vol. 160, No. 3 of 169(3) Figure 37. A, Left supratemporal of Elgaria multicarinata TMM-M 8958, lateral, anterior to the left. Illustrates characters 161(0), 162(0). Left squamosals, ventral, anterior to the left: B, Elgaria multicarinata TMM-M 8958. Illustrates character 164(0). C, Xenosaurus newmanorum NAUQSP-JIM uncatalogued specimen. Illustrates characters 164(1), 167, 169(3). Left squamosals, lateral, anterior to the left: D, TMM-M 8956. Illustrates characters 165(0), 168(0), 169(0). E, Xenosaurus newmanorum NAUQSP- JIM uncatalogued specimen; F, Xenosaurus grandis NAUQSP-JIM 1460. E illustrates 168(1) and F illustrates 168(2). (Fig: 36€): (2) sharp, more acute W edge ( (rc 51D): Evolution. Under both analyses, the ancestral state is squat and obtuse. The small, nub-like morphology is a synapo- morphy of X. agrenon + X. rectocollaris, and the sharp, acute morphology is a syn- apomorphy of X. rackhami + X. grandis. Squamosal The squamosal is unknown for all of the extinct taxa. 164. Squamosal: Posterior portion (0) unex- panded mediolaterally (Fig. 37B); (1) mediolaterally expanded (Fig. 37C) The presence or absence of a canthal 165. crest in the temporal 1 region was men- tioned as a possible sy napomorphy of Xenosauridae by Estes et al. (1988). Evolution. Under both analyses, ex pansion is the ancestral state for the entire group and for Anguimorpha. Under Analysis 1, narrowness is a synapomorphy of Varanidae and of Anguidae + Helodermatidae. Under Apale sis 2, narrowness is a synapomor- phy of Anguidae + Var anoidea. Squamosal: Dorsal ridge adjacent to lateral edge (0) absent (Fig. 37D); (1) relatively low (Fig. 32A); (2) interme- diate height ( Fig. 5A); (3) pronounced (Fig. 32B). 166. Low: 168. Evolution. Under both analyses, presence of a low ridge is a syni apomor- phy of Xenosaurus, presence of a ridge of intermediate height is a synapomor- phy of the southern clade of Xenosaurus. and presence of a sharp ridge is an autapomorphy of X. rac ae Squamosal: Dorsal ridge adjacent to lateral edge (0) most prominent poste- riorly (Fig. 32A); (1) most prominent anteriorly. (Fig. 32B). Evolution. Under both analyses, an- terior prominence is an auta yomorphy of X. agrenon and of X. ln Squamosal: Width, assessed by ratio of mediolateral width at anteroposterior level of closure of supratemporal fe- nestra to anteroposterior length of squamosal (Fig. 37C) (0) 0.15 or less; Cheats to: Or30: (2) 0.30 to: 0.40: (3) 0.40 or above. Evolution. Under both analyses, the ancestral state for the entire group and for Anguimorpha is ambiguous be- tween state 1 and state 2. The ancestral state for Xenosaurus is state 2, and state 3 is a synapomorphy of X. rackhami + X. grandis. Under Analysis 1, the eel state for Shinisaurus + Var- anidae is state 1, and state 0 is a synapomorphy of Varanidae. Under Analysis 2, state 0 is a synapomorphy of Anguidae + Varanoidea. Squamosal: Suspensorial end curvature in a sagittal plane (0) without abrupt terminal hook (Fig. 37D); (1) terminal hook present and reaches the vertical or only slightly beyond the vertical (Fig. 37E); (2) terminal hook progress- es well beyond the vertical, folding under the remainder of the squamosal and extending anteriorly for a short distance (Fig. 37F). Evolution. Under both analyses, cur- vature of the hook to near the vertical is a synapomorphy of Xenosaurus, and curva- ture well beyond the vertical is a synap- omorphy of X. rackhami + X. grandis. XENOSAUR PHYLOGENY ¢ Bhullar 133 169. Squamosal: Ventral ridge dividing at- tachment surface for m. anguli oris la and m. adductor mandibularis externus superticialis 1b from that for m. adduc- tor mandibularis externus medialis and m. adductor mandibularis externus profundus (0) absent (Fig. ee a) uniformly weak (Fig. 38A); (2) sharp only in poste rior portion of squamosal (Fig. 38B); (3) sharp and well-defined for most of length of squamosal (Fig. 37C). Evolution. Under both analyses, ab- sence of a ridge is ancestral for the entire group id for Anguimorpha. A sharp ridge is an autapomorphy of E. eaCarinare: and a ridge - mixed prominence is ancestral tox X, rack hami + X. grandis, ipa a weak ridge is an autapomorphy of X. rack- hawt. Under Analysis 1, a sharp ridge is an autapomorphy of S. croc Bciiarus: The ancestral state for Xenosaurus is ambiguous between a ridge of mixed prominence and a sharp Tidge, as is that for the southern clade Of Xeno- saurus. Under Analysis 2, a sharp ridge is a synapomorphy of Nenosaundac. and a ridge of mixed prominence is a synapomorphy of X. rackhami + X. srandis. Quadrate The quadrate is unknown in the fossil xenosaurs and in M. ornatus. 170. Quadrate: Two-thirds or more of the way down its dorsoventral height, lateral edge of tympanic crest ~(0) abruptly angles medially in plane of the crest, possibly associated with attachment of lateral collateral ligament and other connective tissue (Fig. 38C); (1) curves smoothly without abrupt medial angulation (Fig, 38E). Evolution. Under both analyses, lack of a medial deflection is a synapomor- phy of Varanidae and of X. agrenon + X. Go 134 Bulletin of the Museum of Comparative Zoology, Vol. 160, No. 169(1) 170(0) — Figure 38. Left squamosals, ventral, anterior to the left: A, Xenosaurus rackhami UTEP-OC “MALB” 388. Illustrates character 169(1). B, Xenosaurus grandis NAUQSP-JIM 1460. Illustrates character 169(2). Left quadrates, posterior: C, Elgaria multicarinata TMM-M 8958; D, Xenosaurus newmanorum NAUQSP-JIM uncatalogued specimen. C and D illustrate characters 170(0), 173(0), 173(1). E, Xenosaurus agrenon UTACV r45008. Illustrates character 170(1). rectocollaris. Under Analysis 1, the bump (Fig. 39A); (1) pronounced ancestral state for Anguidae is ue, wedge (Fig. 39B). uous. Under Analysis 2, lack of < deflection is a synapomorphy of C. enneagrammus + O. ventralis. Evolution. Under both analyses, the ancestral state for the entire group is ambiguous; that for Anguimorpha is 171. Quadrate: Anteromedial emargination the small, bump-like morphology; and of dorsal surface of cephalic condyle (O) the wedge morphology is a synap- relatively deep, enclosing angle oos omorphy of X. rackhami + X. grandis. or more (Fig. 39A); (1) relatively shal- |. ae ieee 2 gee ea andusne acetone nes: 173. Quadrate: Depression for tympanic cavity (0) strong, deeply concave pos- (Fig. 39C). teriorly in horizontal plane (Fig. 38C); Evolution. Under both analyses, the (1) weak, weakly concave posteriorly i in shallow morphology is a synapomorphy horizontal plane (Fig. 38D). of X. agrenon + X. rectocollaris. Variation. The depression for the 172. Quadrate: Posterior eminence from tympanic cavity deepens with age, but lateralmost portion of dorsal surface of relative differences can. still We ob- cephalic condyle (0) small, rounded served in early postnatal individuals. 174(0) XENOSAUR PHYLOGENY ¢ Bhullar 135 Figure 39. Left quadrates, dorsal, anterior to the top: A, Xenosaurus newmanorum NAUQSP-JIM uncatalogued specimen; B, Xenosaurus grandis NAUQSP-JIM 1460. A and B illustrate characters 171(0), 172(0), 172(1). C, Xenosaurus agrenon UTACV r45008. Illustrates character 171(1). D, Braincase of Xenosaurus newmanorum NAUQSP-JIM uncatalogued specimen, dorsal, anterior to the top. Illustrates characters 174, 181(0). Evolution. Under both analyses, the shallow morphology isa synapomorphy of Xenosaurus and of Varanidae. Braincase The braincase is unknown for all extinct taxa save B. ammoskius, in which it is largely obscured. 174. Braincase: Ratio of greatest mediolat- eral width to anteroposterior length from posterior end of paroccipital process anteriorly to level of anterior end of alar process (Fig. 39C) (0) 1.20 or less; (1) between 1.20 and 1.50; (2) 1.50 or greater. Variation. The proportions of the braincase vary significantly during ontogeny (Barahona and Barbadillo, 1998: Bever et al., 2005). In particular, the paroccipital processes become rel- atively longer with the relative expan- sion of he adductor chamber relative to the brain, and the alar processes of the prootic lengthen. This character must be ev Alwated on relatively large, adult individuals with fused ivaimeases. Evolution. Under both analyses, a ratio of 1.20 or less is an autapomorphy of X. newmanorum. Under Analysis 1, the ancestral state for the entire group and for Anguimorpha is 1.50 or greater. A ratio of 1.20 to 1.50 is an autapomor- phy of S. crocodilurus, a synapomorphy of Anguidae, and a synapomorphy of X. rackhami + X. grandis. A ratio of 1.20 or 136 Bulletin of the Museum of Comparative Zoology, Vol. 160, No. 3 Figure 40. A, Braincase of Xenosaurus agrenon UTACV 145008, anterior. Xenosaurus platyceps UF 45622, anterior. Illustrates character 175(1). C, Braincase of Xenosaurus grandis NAUQSP-JIM 1460, ventral, anterior to the left. Illustrates character 176(0). D, Braincase of Xenosaurus agrenon UTACV r45008, ventral, anterior to the left. Illustrates character 176(1). less is a synapomorphy of E. multi- carinata + O. ventralis. Under Analysis 2, the ancestral state for the entire group and Anguimorpha is ambiguous be- tween 1.20 to 1.50 and 1.50 or greater. The ancestral state for Anguidae is ambiguous between 1.20 or less and L20:towe a0: 5S, Pee Carotid fossa and retractor pits ) shallow, barely excavated (Fig. 40A); 1) deeply excav ated (Fig. 40B). Variation. The carotid fossa and retractor pits deepen somewhat with age, and only relatively large adults were scored. Evolution. Under both analyses, the shallow morphology is the ancestral state for the entire group and _ for 6: Illustrates character 175(0). B, Braincase of Anguimorpha and the ancestral state for Xenosaurus is ambiguous. Under Analysis 1, the ancestral state for Xeno- saurus + Anguidae and all mixed nodes therein is ambiguous. Under Analysis 2, the ancestral state for Xenosauridae and for Anguidae + Varanoidea is the shallow morphology. The deep mor- phology is an autapomorphy of O. ventralis and of Helodermatidae. Braincase: Domed portion of ventral surface spanning sphenoid/basioccipital suture (Q) divided into bilateral swell- ings with median groove (Fig. 40C); (1) projected into a single swelling without midline groove, bear ing small tubercle slightly posterior to apex of dome in basioccipital region (Fig. 40D). XENOSAUR PHYLOGENY © Bhullar 137 Figure 41. A, Braincase of Xenosaurus newmanorum NAUQSP-JIM uncatalogued specimen, left lateral, anterior to the left. Illustrates character 170(0). B, Braincase of Xenosaurus rackhami UTEP-OC “MALB” 388, left lateral, anterior to the left. Illustrates character 177(1). C, Braincase of Xenosaurus grandis NAUQSP-JIM 1460, left ventrolateral, anterior to the left. Illustrates character 178(0). D, Braincase of Xenosaurus newmanorum NAUQSP-JIM uncatalogued specimen, left ventrolateral, anterior to the left. Illustrates character 178(1). Evolution. Under both analyses, Variation. The sphenooccipital tu the undivided morphology is an bercles become more prominent with autapomorphy of C. enneagrammus age, and their terminal shape forms in the and a synapomorphy of X. agrenon + later stages of postnatal ontogeny; this ie ee ad Under Analy sis: tl. aharacter should be scored on “relatively the ancestral state for the entire large adults with fused braincases. group and for Anguimorpha is am- biguous. The ancestral state for Xe- nosaurus + Anguidae is the divided morphology. ade Analysis 2, the ancestral state for the entire group, 178. Braincase: Recessus vena jugularis (0) Evolution. Under both analyses, the pointe d morphology is an autapomor- phy of O. ventralis and of X. rackhami. Anguimorpha, Xenosauridae, and An- ends anteriorly behind anterior tip of guidae + Varanoidea is the divided alar process of sphe noid (Fig. ALG s(1) morphology. extends along entirety of lateral surface 177. Braincase: Sphenooccipital tubercles of alar process (Fig. 41D). (0) rounded and blunt (Fig. 41A); (1) Variation. The recessus vena jugu- acutely pointed (Fig. 41B). laris and the alar process of the sphenoid 138 Bulletin of the Museum of Comparative Zoology, Vol. 160, No. 3 Figure 42. A, Braincase of Xenosaurus newmanorum NAUQSP-JIM uncatalogued specimen, posterior. Illustrates characters 179(0), 180(0). B, Braincase of Xenosaurus grandis NAUQSP-JIM 1460, posterior. Illustrates characters 179(1), 180(1). C, Braincase of Xenosaurus rackhami UTEP-OC “MALB” 388, dorsal, anterior to the top. Illustrates character 181(1). D, Braincase of Xenosaurus newmanorum NAUQSP-JIM uncatalogued specimen, posterior. Illustrates characters 182(0), 183(1). LAS). both develop largely after hatching. This character should be scored on relatively large adults. Evolution. Under Analysis 1, exten- sion to the anterior end of the alar process is an autapomorphy of S. crocodilurus and a synapomorphy of the northern clade of Xenosaurus. Under Analysis 2, the ancestral state for Xenosauridae is ambiguous and the ancestral state for the southern clade of Xenosaurus is ending posterior to the tip of the alar process. Braincase: Constriction of odontoid recess (notochordal groove) in dorsal margin of occipital condyle, measured by anole through highest points of notch ad deepest point (0) greater 180. than 120° (Fig. 42B). Variation. The recess becomes somewhat more constricted with age and should be scored on relatively large adults. (Fig. 422A); (1) 120° or less Evolution. Under both analyses, an angle of 120° or less is an autapomor- phy of O. ventralis and a synapomorphy of X. rackhami + X. grandis. Braincase: Cava capsulares within braincase (0) leave relatively wide space between them—mediolateral separa- tion at closest approach greater than one-quarter mediolateral width of fo- ramen magnum at that dorsoventral level (Fig. 42A); (1) approach closely— mediolateral separation at closest ap- 183(0) \ XENOSAUR PHYLOGENY ¢ Bhullar 139 Figure 43. A, Braincase of Xenosaurus agrenon UTACV r45008, posterior. Illustrates characters 182(1), 183(0). B, Braincase of Xenosaurus grandis NAUQSP-JIM 1460, posterior. Illustrates character 182(2). C, Left dentary of Xenosaurus rackhami UTEP- OC “MALB” 388, medial, anterior to the right. Illustrates characters 185(0), 197(1). D, Right dentary of Exostinus serratus, CT scan of AMNH 1608, lateral, anterior to the right. Illustrates characters 185(1), 186(1), 195(1), 196(1). SW. proach one-quarter or less mediolateral width of foramen magnum at that dorsoventral level (Fig. 428). Variation. The approach of the cava capsulares increases somewhat with age and should be scored in relatively large adults. Evolution. Under both analvses. close approach is an autapomorphy of O. ventralis and of X. grandis. Braincase: Paroccipital processes (QO) distinctly Beyer ated aly directed (Fig. 39D): (1) only slightly posterolat- erally directed, nearly Greate medio- laterally (Fig. 42C). Variation. The posterior direction of the paroccipital processes tends to increase with age and should be scored in relatively large adults. Evolution. Under Analysis 1, the ancestral state for the entire group and Anguimorpha, as well as all internal mixed nodes, is ambiguous. Under Analysis 2, the aneecal state for the entire group and for Anguimorpha is lateral extension. Posterolateral exten- sion is a synapomorphy of C. ennea- srammus + O. ve ere and an autapo- morphy of L. borneensis. The ancestral state for Xenosauridae is ambiguous. . Braincase: Expanded tips of paroccipi- tal Be Smear ae projections (0) ee (Fig. 42D); (1) present, squared off or blunt (Fig. 43A); (2) present, long and pointed (Fig. 43B). 140 183: Bulletin of the Museum of Comparative Zoology, Variation. The projections become more prominent during early postnatal ontogeny and Btatial be scored in relatively large adults. Evolution. Under both analyses, ab- sence of projections is an autapomorphy of S. crocodilurus and of X. newma- norum. Under Analysis 1, the ancestral state for the entire group and _ for Anguimorpha is ambiguous between squared off and long and pointed, as is that for S. crocodilurus + Varanidae, Xenosaurus + De radae. and Xeno- saurus. Under Analysis 2, the ancestral state for the entire group and for Anguimorpha is long and pointed. Blunt or squared-off is an autapomorphy of Helodermatidae and a synapomorphy of Xenosauridae. Within Xenosauridae, the long and pointed state is a synapomor- phy of X. rackhami + X. grandis. Braincase: Expanded tips of paroccipi- tal processes—dorsal projections (0) present as blunt or squared off tab (Fig. 43A); (1) low, barely divergent (Fig. 42D). Variation. Dorsal projections tend to become more prominent during post- natal ontogeny and should be scored in relatively lar ge adults. Evolution. Under Analysis 1, the ancestral state for the entire group and all mixed nodes therein is ambig- uous. Under Analysis 2, the ancestral state for the entire croup and for Anguimorpha is presence of relatively extensive projections. The low and rounded morphology is a synapomor- phy of Varanoidea and the northern clade of Xenosaurus and an autapo- morphy of C. enneagrammus. Dentary 184. Dentary: Anterior tip of dentary (0) relatively pointed in sagittal plane (Fig. 1A); (1) blunt, truncated, with steeply rising “chin” (genioglossus at- tachment area). 185. 186. Vol. 160, No. 3 Evolution. Under both analyses, the ancestral state for the entire group and for Anguimorpha is the relatively point- ed morphology. The ancestral state for E. lancensis + Xenosaurus is ambiguous, and the ancestral state for Xenosaurus is pointed. Dentary: Posterior, rising section of dorsal edge of dentary extends for (0) more than six tooth positions (Fig. 43C); (1) six or fewer tooth positions (Fig. 43D). Evolution. Under Analysis 1, the ancestral state for the entire group is ambiguous, and the ancestral state for Anguimorpha is six or fewer tooth positions. More than six tooth posi- tions is a synapomorphy of O. ventralis + FE. multicarinata and an autapomor- shy of M. ornatus. The ancestral state Fox Xenosaurus + E. lancensis is ambiguous, as is that of Xenosaurus + E. serratus. Under Analysis) 27ape ancestral state for the entire group and for Anguimorpha is more than six tooth positions. Six or fewer tooth positions is a synapomorphy of Var- anoidea and S. crocodilurus + B. ammoskius and an autapomorphy of C. enneagrammus, R. rugosus, and E. serratus. Dentary: Groove anterior to coronoid facet on lateral surface of dentary (0) shallow and short, not approaching posteriormost mental foramen (Fig. 44A); (1) deep and long, approaching or running to posteriormost mental foramen (Fig. 43D). Evolution. Under both analyses, the ancestral state for the entire group and for Anguimorpha is the shallow, short morphology. Under Analysis > the deep, long morphology is a synapo- morphy of Xenosaurus + R. rugosus and an autapomorphy of S. Croounnie: Under Analysis 2, the deep lone morphology is a synapomorphy of Xenosauridae. 186(0) 187(1) XENOSAUR PHYLOGENY ¢ Bhullar 14] 187(0) Figure 44. A, Left dentary of E/garia multicarinata TMM-M 8958, lateral, anterior to the left. Illustrates characters 186(0), 188(0). B, Left dentary of Xenosaurus grandis NAUQSP-JIM 1460, lateral, anterior to the left. Illustrates characters 187(0), 190(0). C, Left dentary of Xenosaurus newmanorum NAUQSP-JIM uncatalogued specimen, lateral, anterior to the left. Illustrates character 187(1). D, Left dentary of Xenosaurus agrenon UTACV r45008, lateral, anterior to the left. Illustrates characters 188(1), 189(1), 190(1). liege 188. Dentary: Lateral coronoid facet (0) shallowly impressed, surrounded dor- sally and anteriorly by low ridge (Fig. 44B); (1) deeply impresse d. Sur: rounded by sharp ridge (Fig. 44C). Evolution. Under both analyses, deep impression with a sharp ridge is a synapomorphy of the northern cade of Xenosaurus and an autapomorphy ot C. enneagrammus. Dentary: Coronoid process (0) extend- ing posteriorh ly beyond surangular pro- eess (Hie: A4A): (1) ending anterior to tip of surangular process (Fig. 44D). Evolution. Under both analyses, the ancestral state for the entire group is ambiguous, and that for Anguimorpha is extension beyond the surangular 189. ISe, process. Ending anterior to the suran- gular process is a synapomorphy of X. agrenon + X. rectocollaris and an auitapomorphy of S. crocodilurus and of Helodermatidae. Dentary: Surangular notch between surangular and angular process (0) absent (Estes et al., 1988, fig. 11B); (1) present (Fig. 44D). Presence or absence of surangular notch was character 63 of Estes et al. (1988). Evolution. Under both analyses, the ancestral state of the entire group is ambiguous. Dentary: Tip of angular process (Q) pointed (Fig. is (1) blunt and slightly bifurcated ( Fig. 44D). 142 Bulletin of the Museum of Comparative Zoology, Vol. 160, No. 3 —192(1) Figure 45. A, Left dentary of Shinisaurus crocodilurus UF 72805, medial, anterior to the right. Illustrates characters 191(0), 193, 195(2). B, Left and right dentaries of Restes rugosus YPM PU 14640, lateral and medial, respectively, anterior to the left. Illustrates character 191(1). C, Left dentary of Xenosaurus rectocollaris, CT scan of UF 51443, dorsal, anterior to the left. Illustrates characters 192(0), 194(0). D, Right dentary of Exostinus serratus, CT scan of AMNH 1608, dorsal, anterior to the left. Illustrates characters 192(1), 194(1). oe Io2: Evolution. Under both analyses, a blunt, slightly bifurcated angular pro- cess is a synapomorphy of X. agrenon + X. rectocollaris. Dentary: Suprameckelian lip (0) be- comes ‘dorsoventrally taller anteriorly, but Meckel’s groove remains fairly widely open (Fig. 45A); (1) becomes dor soventrally tall anteriorly, restricting Meckel’s groove to thin slit (Fig. 45B). Evolution. Under both analyses, the ancestral state for Xenosaurus + R. rugosus and Xenosaurus + E. lancensis was ambiguous. Dentary: Wedge- shaped process extend- ing posteriorly from suprameckelian lip 193. near posterior end of dental gutter (0) absent (Fig. 45C); (1) present (Fig. 45D). Evolution. Under both analyses, the presence of a wedge-shaped process is an autapomorphy of E. serratus. Dentary: Apex of posterior u-shaped emargination in intramandibular septum (Fig. 45A) at level of (0) third or fewer tooth position from back of dentary; (1) fourth tooth position from back; (2 ) fifth tooth position from back; (3) sixth or more tooth position from back. Evolution. Under both analyses, the ancestral state for Xenosaurus + R. rugosus is at the level of the fourth tooth position from the back, and the 194. IN Os state for Xenosaurus + E. lancensis and Xenosaurus + E. serratus is ambiguous between the fourth and_ fifth tooth position from the back. The ancestral state for Xenosaurus is the fifth tooth position from the back. A position at the level of the fourth tooth position from the back is an autapomorphy of X. grandis, and one at the level of the sixth from the back is an autapomorphy of X. new- manorum. Finally, a position at the level of the sixth from the back is an autapomorphy of E. multicarinata. Un- der Analysis 1, the ancestral state for the entire group, for Anguimorpha, S. croco- dilurus + Varanidae, Xenosaurus + An- guidae, and Anguidae + Helodermatidae, is the third or fewer tooth position from the back. A position at the level of the fourth tooth position from the back is a synapomorphy of Xenosaurus + R. rugo- sus, of E. multicarinata + O. ventralis ead of S. crocodilurus + M. ornatus. Under Analysis 2, the ancestral state for the entire group and for Anguimorpha is ambiguous between the third and fourth ai positions from the back. That for Xenosauridae is the fourth tooth position from the back, and that for Varanoidea is the third from the back. Dentary: Posterior end of tooth row in horizontal plane (0) relatively straight or with slight medial inflection (Fi ig. 45C); (1) with marked medial inflection ( Fig. 45D). Evolution. Under both analyses, me- dial inflection is an autapomorphy of E. serratus. Dentary: Tooth height (0) short, with one-third or less of most teeth extend- ing above dorsal margin of dentary (Fig. 45A); (1) intermediate height, between one-third and half of most teeth extending above dorsal margin of dentary (Fig. 46A); (2) tall, with half or more of most teeth extending above dorsal margin of dentary (Fig. 4B). Evolution. Under both analyses, the ancestral state for the entire group is 196. 1 9 eS . Dentary: XENOSAUR PHYLOGENY © Bhullar 143 ambiguous between tall and interme- diate height, and the ancestral state for Anguimorpha is intermediate height. The ancestral state for Xenosaurus + R. rugosus is intermediate height, and the alert morphology is a syna somorphy of Xenosaurus. Under An: sl sis 1, the ancestral state for Xenosaurus + Angu- idae is intermediate height. The tall morphology is a synapomorphy of S. crocodilurus + Var ae and an auta- pomorphy of Helodermatidae. Under Analysis 2, the ancestral state for Anguidae + Varanoidea and for Xeno- sauridae is intermediate height. The tall morphology is a synapomorphy of shinisaurs and of Varanoidea. Dentary: Tooth height (0) declines dras- tically posteriorh ly, atk two or three teeth anterior to most posterior tooth about half to two-thirds the height of tallest dentary teeth (Fig. 46A); oa ) declines less, with teeth just anterior to most posterior tooth nearly the same height as tallest dentary teeth (F (F ig. 43D). Evolution. Under both analyses, a precipitous decline is the ancestral state for the entire group and for Anguimor- pha. Lack of a decline is ancestral for Xenosaurus + R. rugosus and_ for shinisaurs. A precipitous decline is a synapomorphy of Xenosaurus. Under Analysis 1, lack of a decline is a synapomorphy of Xenosaurus + R. rugosus and of shinisaurs. Under Anal- eee , lack of a decline is a synapomor- phy as Xenosauridae. Shafts of teeth (0) do not change drastically in diameter posteri- orly, save at anterior tip in premaxil- lary—maxillary occlusal transition (Fig. 46A); (1) increase markedly in diameter posteriorly (Fig. 43C). Variation. The differentiation of shaft widths becomes more pro- nounced with age. Evolution. Under both analyses, a marked posterior increase in diameter 144 Bulletin of the Museum of Comparative Zoology, Vol. 160, No. 3 Figure 46. A, Right mandible of Xenosaurus platyceps UF 45622, medial, anterior to the left. Illustrates characters 195(0), 196(0), 197(0), 200(1), 201(1), 202(1), 203(0), 204(0). B, Left mandible of E/garia multicarinata CAS 54241, medial, anterior to the right. Illustrates characters 195(1), 200(0), 201(0). C, Right mandible of Xenosaurus platyceps UF 45622, lateral, anterior to the right. Illustrates characters 198(0), 199(1), 205(2), 206(1), 208(1). D, Right mandible of Xenosaurus newmanorum NAUQSP-JIM uncatalogued specimen, lateral, anterior to the right. Illustrates characters 198(1), 206(0), 207(1). isa synapomorphy of X. rackhami + X. erandis. Coronoid The coronoid is unknown for E. lancensis and M. ornatus. 198. IS): Coronoid: Anterolateral process (0) anteroposteriorly longer than dorsoven- trally tall at base (Fig. A6C): (1) taller than long (Fig. 46D). Evolution. Under both analyses, taller than long is an autapomorphy of X. new- manorum, of X. agrenon, and of X. grandis. Coronoid: Posterolateral process: (0) ventral margin directed posterodor- sally, resulting in taper for entire length 200. (Fig. 47A); (1) ventral margin directed Siar aight posterior ly for most of length, resulting in a dorsoventr ally extensive process (Fig. A6C). Evolution. Under both analyses, a straight ventral margin and _ tall pos- terolateral process is a synapomorphy of Xenosaurus. Coronoid: Posteromedial process (0) trending posteroventrally without strong ‘bend toward the vertical (Fig. 46B); (1) becoming nearly vertical at tip, resulting in highly bowed ap- pearance of Sorancid arch (Fig. 47B). Evolution. Under Analysis Le. the ancestral state for the entire group and 145 XENOSAUR PHYLOGENY © Bhullar Figure 47. A, Left mandible of Elgaria multicarinata CAS 54241, dorsolateral, anterior to the left. Illustrates characters 199(0), 205(0), 207(0), 208(0). B, Left coronoid of Xenosaurus agrenon UTACV 145008, medial, anterior to the right. Illustrates characters 200(1), 201(1), 202(0), 203(1). C, Right mandible of Restes rugosus, CT scan of YPM PU 14640, medial, anterior to the left. Illustrates character 204(1). D, Right mandible of Xenosaurus rackhami UTEP-OC “MALB” 388, lateral, anterior to the right. Illustrates character 205(1). 201. for Anguimorpha is ambiguous. The ancestral state for Anguidae + Heloder- matidae is the helntin . straight mor- phology. Under Analysis 2 _the ancestral state for the entire group and for Anguimorpha is the bowed morphol gy. The straight morphology is a synapo- morphy of Anguidae + V aranoidea and an autapomorphy of R. rugosus. Coronoid: Anteromedial process (0) with abrupt constriction of dorsal and ventral margins about a third of the way toward anterior tip followed by contin- ued taper (Fig. 46B); tapering without abrupt constriction or with slight step in dorsal, but not ventral, margin (Fig. 47B). 202. Coronoid: (1) gradually Evolution. Under both analyses, a gradual t taper is a synapomorphy of Soins auiietis and of Varanus. Anteromedial process (Q) more than twice as long along long axis as tall perpendicular to long axis at widest level (Fig. 47B); (1) twice as long as. tall less, with thickened process restricting exposed portion anteriorly (Fig. AGA), Evolution. Under both analyses, twice as long as tall or shorter is a synapomorphy of the northern clade of Xenosaurus. Under Analysis 1, twice as long as tall or shorter is a synapomor- phy of Varanus and an autapomorphy of Helodermatidae. Under Analysis 2, 146 204. 205. 206. Bulletin of the Museum of Comparative Zoology, Vol. 160, No. 3 twice as long as tall or shorter is a synapomorphy of Varanoidea. .Coronoid: Medially facing flange ex- tending posteriorly from posteromedi- al process for attachment of bodena- poneurosis; ratio of greatest width perpendicular to long axis to length along long axis, (0) 0 25m ommless (Fig. 46A); (1) greater than 0.25 (Fig. 47B). Evolution. Under both analyses, an extensive flange ise synapomorphy of X. agrenon + X. rectocollaris. Coronoid: Posterior end of posterome- dial process (0) extends for short distance along prearticular (ventrome- dial) rim of ei: fossa (Fig. 46A); (1) ends anterior to adductor fossa (Fig. 47C). Evolution. Under both analyses, ter- mination anterior to the adductor fossa is a synapomorphy of the southern clade of Xenosaurus and an autapomorphy of R. rugosus. Coronoid: Anterior surangular foramen (0) not overlapped by coronoid (Fig. 47A); (1) partially overlapped by coronoid, forming emargination in ven- tral margin of bone (Fig. 47D); (2) fully overlapped by coronoid, piercing through the bone (Fig. 46C). Evolution. Under both analyses, par- tial overlap is an autapomorphy of L. borneensis. Complete overlap is a synapomorphy of Xenosaurus, and par- tial overlap is an autapomorphy of X. rackhami. Coronoid: Anterior surangular foramen in coronoid (0) single (Fig. 46D); (1) double (Fig. 46C). Evolution. Under both analyses, the paired morphology is an autapomorphy OFX platyceps. Surangular/Prearticular/Articular The surangular/prearticular/articular com- plex is unknown in E. lancensis and M. ornatus. 207. Surangular/prearticular/articular: Strong dorsal crest on surangular lateral to m. adductor externus medialis attachment table (0) absent (Fig. 47A); (1) present (Fig. 46D). Variation. The crest becomes more prominent with age, and only relatively large adults should be scored. Evolution. Under both analyses, presence of a crest is a synapomorphy of Xenosaurus and an autapomorphy of Helodermatidae. 208. Surangular/prearticular/articular: Pos- terior surangular foramen (0) anterior to articular (Fig. 47A); (1) at antero- posterior level of anterior edge of articular (Fig. 46C). Evolution. Under both analyses, a position at the anterior edge of the articular is a synapomorphy of Xenosaurus. 209. Surangular/prearticular/articular: Fossa ventral to arc of medial coronoid facet (subcoronoid fossa) (0) shallow, only slightly impressed (Fig. 48A); (1) deep, forming a strongly impressed bowl- shaped cavity (Fig. 4SB). Evolution. Under both analyses, a deep subcoronoid fossa is a synapo- morphy of Xenosaurus + E. serratus. 210. Surangular/prearticular/articular: Sub- coronoid fossa (0) with no medial overhang at posterior end or only slight ridge developed there (Fig. 48A); (1) posteriorly continuing into bone as a blind pocket, sometimes pierced by a foramen, such that a portion of the medial wall of the surangular bearing the facet for the posteromedial process of the coronoid overhangs it (Fig. 48B). Evolution. Under both analyses, lack of overhang is the ancestral state for the entire group and for Anguimorpha, and presence of an overhang is the ancestral state for Xenosaurus + R. rugosus. Under Analysis 1, the ancestral state for Xeno- saurus + Anguidae and for Anguidae + Helodermatidae is ambiguous, and over- XENOSAUR PHYLOGENY ¢ Bhullar 147 Figure 48. A, Left postdentary elements of Elgaria multicarinata TMM-M 8958, medial, anterior to the right. Illustrates characters 209(0), 210(0). B, Left postdentary elements of Xenosaurus grandis NAUQSP-JIM 1460, medial, anterior to the right. Illustrates characters 209(1), 210(1). C, Cervical vertebrae of Shinisaurus crocodilurus MVZ 204291, left lateral, anterior to the left. Illustrates characters 212(0), 213(1). D, Cervical vertebrae of Heloderma suspectum CAS 513, right ventrolateral, anterior to the right. Illustrates characters 212(1), 213(0). hang i is an autapomorphy of L. borneen- sis. aWiader Analysis 2, overhang is a synapomorphy of Xenosaurus + R. Pugo= sus, and the ancestral state for Varanoi- dea and Varanidae is ambiguous. Sur angular, /preartic wlar/articular: Subcoro- noid {0540 ) without foramina; (1) pierced by one or more small foramina; (2) pierced by one large posterior foramen serving as exit for batial from adductor fossa. Evolution. Under both analyses, the ancestral state for the entire group is ambiguous, and that for Anguimorpha is one or more small foramina. A single large posterior foramen is a synapo- morphy of Xenosaurus. 212. Cervical skeleton: Cervical Skeleton Most or all of the cervical skeleton is lacking in all fossils save B. ammoskius. Intercentra three and four (0) dorsoventrally taller than anteroposteriorly long (Fig. A8C): (1) anteroposteriorly longer ‘han dorsoven- trally tall (Fig. 48D). Evolution. Under both analyses, the ancestral state for the entire group and for Anguimorpha is taller than long. Under Analysis 1, the ances- tral state for Xenosaurus + Anguidae and for Anguidae + Helodermatidae is ambiguous. Under Analysis 2, long- er ehant tall is c FIRS OD v : a Sig ie, Go Figure 50. A, Caudal vertebrae of Shinisaurus crocodilurus MVZ 204291, left lateral, anterior to the left. Illustrates character 218(0). Caudal vertebrae, dorsal, anterior to the left: B, Xenosaurus newmanorum, CT scan of UMMZ 126057; C, Xenosaurus platyceps, CT scan of UF 25005. B and C illustrate characters 218(1), 268(0), 268(1). Evolution. Under Analysis 1, the ancestral state for the entire group and for Anguimorpha is ambiguous. Under Analysis 2, the ancestral state for the entire group and for Anguimorpha is presence of autotomy planes, and their absence is a synapomorphy of Xenosaurus and of Varanoidea. Pectoral Girdle The pectoral girdle is unknown for all extinct taxa save B. ammoskius. 219.Pectoral girdle: Scapulocoracoid emargination: (0) dorsal and ventral margins strongly div erging for most of length (Fig. B5lAY (1) dora and ventral margins weakly diverging or nearly parallel for most of length (Fig. 51B). Evolution. Under both analyses, weak or no divergence is a sy napomor- phy of the fiortGen clade of Xeno- saurus and of Varanus, as well as an autapomorphy of C. enneagrammus. . Pectoral girdle: Scapulocoracoid emar- gination; ratio of greatest height per- pendicular to long axis to length along long axis (0) 0.70 or greater (Fig, 5): (1) ess than 0.70 (Fig. SLD: Evolution. Under both analyses, the ancestral state for Anguidae is ambigu- ous, and a ratio of less than 0.70 is a synapomorphy of the northern clade of Xenosaurus. Figure 51. 221(1)~ XENOSAUR PHYLOGENY © Bhullar 15] A, Left scapulocoracoid of E/garia multicarinata TMM-M 8993, lateral, anterior to the left. Illustrates characters 219(0), 222(0), 223. B, Left scapulocoracoid of Xenosaurus newmanorum NAUQSP-JIM uncatalogued specimen, lateral, anterior to the left. Illustrates characters 219(1), 221(0), 224(1). C, Left scapulocoracoid of Xenosaurus rackhami UTEP-OC “MALB’” 388, lateral, anterior to the left. Illustrates characters 220(0), 221(1), 222(1), 225(0). D, Left scapulocoracoid of Xenosaurus platyceps UF 45622, lateral, anterior to the left. Illustrates characters 220(1), 224(0), 225(1). 221. . Pectoral girdle: Pectoral girdle: Sharp hook-like over- hang of scapulocoracoid emargination by scapula (0) absent (Fig. 5B): (1) present (Fig. 51C). Evolution. Under both presence of an overhang is morphy of X. rackhami +X. analyses, a synapo- g -andis. Length of coracoid emargination along long Aas from apex of emar gination to anterior end of dorsal margin (0) equal to or than length of remainder of OF eater scapulocor acoid along line of axis con- tinued posteriorly Foti apex (Fig. 51A); (1) less than length of remainder of scapulocoracoid (Fig. 51C). anterior . Pectoral Evolution. Under Analysis 1, the ancestral state for the entire group and for Anguimorpha is ambiguous, as is that for all mixed ater nodes. Under Analysis 2, the ancestral state for the entire group and Anguimorpha is as long or longer than the remainder of the scapulocoracoid. Shorter than the remainder is a synapomorphy of Xeno- SAUTUS and of Var anoidea. girdle: Anterior coracoid emargination: ratio of greatest height perpendicular to long axis to length along long axis (Fig. 51B) (O) 0.50 or less: (1) between 0.50 and 0.85; (2) 0.85 or greater. 1 bo 229. o2 .Pectoral girdle: Bulletin of the Museum of Comparative Zoology, Vol. 160, No. 3 Evolution. Under both analyses, the ancestral state for the entire group and for Anguimorpha is between 0.50 and 0.85. The state2050-or lessvis an autapomorphy of X. newmanorum, whereas 0.85 or greater is an autapo- morphy of X. platyceps and of L. borneensis. Under Analysis 1, the an- cestral state for Anguidae + Heloderma- tidae and for Anguidae is ambiguous between 0.50 or less and between 0.50 and 0.85. The state 0.50 or less is a synapomorphy of Varanus. Under Anal- ysis 2, the ancestral state for Anguidae + Var: angen is ambiguous between states 0.50 or less and bemueen 0.50 and 0.85. Posterior coracoid emar gination (0) absent, ventral margin of coracoid curves smoothly (Fig. 51D); (1) present at least as abrupt anterior change in angle of ventral margin of coracoid toward the horizontal and straightening of curve (Fig. 51B) (see character 112 of Estes et al., 1988). Evolution. Under both analyses, the ancestral state for the entire group is ambiguous, and that for Anguimorpha is absence of the emargination. Pres- ence of the emargination is an autapo- morphy of X. newmanorum, of C. enneagrammus, and of L. borneensis. Pectoral girdle: Posterior end of cora- coid (0) terminating in a point (Fig. 51C); (1) squared-off or blunt, attenuate (Fig. 51D). Evolution. Under both analyses, the squared-off or blunt morphology is a synapomorphy of the northern clade of Xenosaurus. 226. Pectoral girdle: Clavicle, flattened tip of sternal ramus (0) in plane parallel or oblique to that of flattened portion of scapular ramus; (1) in plane nearly perpendicular to that of flattened portion of scapular ramus. Evolution. Under both analyses, the nearly perpendicular morphology is a bo ~l synapomorphy of the northern clade of Xenosaurus. . Pectoral girdle: Clavicle, corner be- tween rami (0) projected into long, narrow process (Fig. 52A); (1) pro- duced into distinct, moderately lon eminence (Fig. 52B); (2) oda only slightly into small thickest or bump (Fig. 52C). Evolution. Under both analyses, the ancestral state for the entire group and for Anguimorpha is ambiguous be- tween production into a moderately long eminence and slight production. The long morphology is an autapomor- phy of S. crocodilurus and of E. multi- carinata. Under Analysis 1, the ances- tral state. for (lh) spresemta ian WG,)s Evolution. Under both analyses, the presence of rims is a synapomorphy of Xenosaurus. It is also an autapomorphy of X. grandis and of Helodermatidae. Cranial osteoderms: Lateral temporal osteoderms, spacing—(0) closely packed (Fig. 1C); (1) separated by small gaps (Fig. 5C); (2) separated by gaps of one scale or more (Fig. 4C); (3) present only as small slivers of ossifica- tion (Fig. 3C): (4) absent. Evolution. Under both analyses, the ancestral state for the entire group is ambiguous among the last four states. That for Anguimorpha is ambiguous bo between separation by small and large gaps. Presence as only small slivers of ossification is autapomorphic for X. rackhami. Under Analysis 1, the ances- tral state for S. crocodilurus + Varanidae is ambiguous between separation by small and large gaps. The ancestral state for Xenosaurus + Anguidae is ambigu- ous between close packing and separa- tion by small gaps. Close packing is the ancestral state for Anguidae + Heloder- matidae. Separation by small gaps is ancestral for the southern clade of Xenosaurus, and separation by large gaps is a synapomorphy of X. rackhami + X. grandis. Under Analysis 2, the ancestral state for Anguidae + Varanoi- dea is ambiguous between closely packed and separated by small gaps. Cranial osteoderms: Lateral temporal osteoderms, coverage of quadratojugal region posterior to “quadratojugal pro- cess of jugal (0) present (Fig. 1C); (1) absent (Fig. 5C). Evolution. Under both analyses, noncoverage of the quadratojugal re- gion is a synapomorphy of the southern clade of Xenosaurus. 258. Cranial osteoderms: Mandibular osteo- derms (0) absent; (1) most of dentary exposed; only portions of postdentary bones extensively covered; (2) covering part of lateral face of mandible, includ- ing dentary and postdentary bones, but leaving large areas of dentary and postdentary bones exposed (Fig. 5B, C); (3) covering most of lateral face oF mandible save for retroarticular process and pterygoideus insertion on surangu- lar, present in three major longitudinal rows (Figs. 1B, C); (4) as with state 3 but with four major longitudinal rows (Higse 2 BiG) Evolution. Under both analyses, the ancestral state for Anguimorpha is ambiguous between state 1 and state 2. The ancestral state for Xenosaurus is 2, and state 3 is a synapomorphy of 209. 260. the northern clade of Xenosaurus, with state 4 an autapomorphy of X. platy- ceps. State 3 is also a synapomorphy of Anguidae. The ancestral state of most other nodes under both analyses is ambiguous between. state 1 and state 2. Cranial osteoderms: Mandibular osteo- derms (0) with strong, sharp. keels (Fig. 1B, C); (1) unkeeled or subtly keeled (Fig, ob, ©). Evolution. Under both analyses, strong keeling of the mandibular osteo- derms is a synapomorphy of the north- ern clade of Xenosaurus. Cranial osteoderms: Intermandibular osteoderms (0) present, well-developed (Fig. 1B); (1) present as multiple small ossifications (Fig. 2B); (2) present as one or two small ossifications (Fig. 5B); (3) absent (Fig. 4B). Evolution. Under both analyses, the well-developed morphology is an auta- pomorphy of X. newmanorum, and absence is a synapomorphy of X. sal + X. grandis. Under Analysis 1, the ancestral state for the entire group is ambiguous among multiple small paciecauonus. one or two small ossifications, and absence, and that for Anguimorpha is ambiguous between multiple and one or two small ossifica- tions. The well-developed morphology is a synapomorphy of Anguidae + Helodermatidae and an autapomorphy - of L. borneensis. Under Analysis 2. the 261. ancestral state for the entire group is ambiguous between one or two small ossifications and absence. That for Anguimorpha is one or two small ossifications. The well-developed mor- phology is a synapomorphy of Anguidae + Varanoidea, and presence as multiple small ossifications is a synapomorphy of the northern clade of Xenosaurus. Cranial osteoderms: Intermandibular osteoderms. well-developed portion 26 263. 2. Cranial XENOSAUR PHYLOGENY © Bhullar 16] (0) in posterior half of mandible; (1) in anterior half of mandible (Fig. 1A). Evolution. Under An: aly sis 1, poste- rior deve saa nt is ance steal and development in the anterior half of the mandible is a synapomorphy of the northern clade of Xenosaurus or of all xenosaurs. Under Analysis 2, the an- cestral state is ambiguous across most of the tree, but the ancestral state of Varanoidea is posterior development. osteoderms: Intermandibular osteoderms, lateral rows (0) present along entirety of tooth-bearing region of dedtarics (Fig. 1B); (1) present only along anterior portion (Fig. 2B), Evolution. The ancestral state of the character is entirely ambiguous. Cranial osteoderms: Intermandibular osteoderms (0) spanning mandibular rami for large portion of anteroposteri- or extent (Fig, 1B); (1) developed only laterally for most of extent (Fig. 2B). Evolution. Under both analyses, the ancestral state for Anguimorpha is ambiguous. Under Analysis 1, most atornal nodes are ambiguous, save that the ancestral state fog Anguidae + Helodermatidae is extensive “develop- ment. Under Analysis 2, the ancestral state for Anguidae + Varanoidea is extensive development, and that for Xenosauridae is lateral development, making the extensive development in X. newmanorum an autapomorphy of that taxon. Cervical Osteoderms 264. Cervical osteoderms: Cervical osteo- derms (0) present dorsally in each scale (The Deep Scaly Project, 2007); (1) present dorsally in regularly spaced pattern, but not in each scale: (2) present dorsally as small ossifications concentrated anteriorly with occasional larger osteoderms interspersed (Fig. 1A); (3) present dorsally as small 162 Bulletin of the Museum of Comparative Zoology, Vol. 160, No. 3 hr Sg om, et tA Bees SAAS WOOY ol 2 Figure 57. Left pectoral limbs, extensor view, anterior to the left: A, Xenosaurus newmanorum, CT scan of UMMZ 126057; B, Xenosaurus platyceps, CT scan of UF 25005; C, Xenosaurus rectocollaris, CT scan of UF 51443. A illustrates 265(1) and 266(1). B illustrates 265(0) and 266(1). C illustrates 265(0) and 266(2). ossifications (Fig. 3A); (4) absent Pectoral Osteoderms (Fig. 5A). Evolution. Under both analyses, the ancestral state for the entire group is ambiguous among the last four states, and that for Anguimorpha is ambiguous on flexor surface (Fig. 57B); (1) present between | and 2. The ancestral state for on central two-thirds or more of Xenosaurus is 2. and state 3 is a extensor surfaces, but only marginally synapomorphy of the southern clade on flexor surfaces (Fig. 57A); (2) pres- of Xenosaurus. State 4 is an autapo- ent only as narrow central band (about morphy of X. rectocollaris. Under central one-third) of extensor surfaces, Analysis 1, the ancestral state for S. save for a few scattered ossifications crocodilurus + Varanidae is 2, and that over elbow (Fig. 57C); (3) absent. 265. Pectoral osteoderms: Osteoderms on pectoral limb (0) present on most of extensor surfaces of stylopod and zeugopod and more lightly distributed for Anguidae + Helodermatidae is 0. Under Analysis 2, the ancestral state for Anguidae + Varanoidea is ambiguous between O and 1, and that for Xeno- sauridae is ambiguous between 1 and 2. Evolution. Under both analyses, the ancestral state for the northern clade of Xenosaurus is 1, and state O is an autapomorphy of X. platyceps. The ancestral state for the southern clade 66. of Xenosaurus is Z and state 3 is a synapomorphy of rackhami + X. grandis. Under oy sis 1, the ancestral state for the entire group is ambiguous among 1, 2, and 3. That for Anguimor- pha is ambiguous between | and 2. The ancestral state for Anguidae + Heloder- matidae is 0. Under Analysis 2, the ancestral state for the entire group is ambiguous between 2 and 3, and that for Anguimorpha is 2. State 0 is a synapo- morphy of Anguidae + Varanoidea, and state 1 is a synapomorphy of the northern clade of Xenosaurus. Pectoral osteoderms: Osteoderms on pectoral limb (0) present as small nubs of ossification; (1) present as large rounded ossifications (Fig. 57A); (2) present as large rounded ossifications and in places as cone-shaped spicules (F1ig.-57 3B). Evolution. Under both analyses, the ancestral state for the entire group is presence as large rounded ossifications. Presence as small nubs of ossification is an autapomorphy of S. crocodilurus, and presence as large rounded ossifi- cations plus cone-shaped ossifications is an autapomorphy of X. platyceps. Pelvic Osteoderms 267. Pelvic osteoderms: Osteoderms on pel- vic limb (0) present on extensor surface of stylopod and scattered on flexor surface (McDowell and Bogert, 1954, plate 4); (1) present only on extensor surface of stylopod (Fig. 54A); (2) absent. Evolution. Under both analyses, the ancestral state for the northern clade of Xenosaurus is extensor presence only, and presence on the extensor and flexor surfaces is an autapomorphy of X. platyceps. Under Analysis 1, the ancestral state for the entire group and for Anguimorpha is ambiguous be- tween presence on the extensor surface and absence. Presence on the flexor XENOSAUR PHYLOGENY ¢ Bhullar 163 and extensor surfaces is a synapomor- phy of Anguidae + Aelodenn sides and an autapomorphy of L. borneensis. Under Analysis 2, the ancestral state for the entire group and for Anguimor- pha is absence, and presence on the extensor and flexor surfaces is a synapomorphy of Anguidae + Varanoi- ie Presence on the extensor surface is a synapomorphy of the northern clade of Xenosaurus. Caudal Osteoderms 268. Caudal osteoderms: Caudal osteoderms (0) in complete or nearly complete rings on anterior portion of ae scattered over more posterior portion (Fig. 50B); (1) scattered in anterior portion of tail, sparse in more posterior portion (Fig. 50C); (2) absent. Evolution. Under both analyses, the ancestral state for the entire group is ambiguous between scattered and ab- sent. That for Anguimorpha is scat- tered, as is that for Xenosaurus. Ab- sence of cadual osteoderms is a synapomorphy of the southern clade of Xenosaurus, and complete rings thereof is an autapomorphy of X. new- manorum. Under Analysis 1, complete rings is a synapomorphy of Anguidae + He lodermatidac and an autapomorphy of L. borneensis. Under Analysis 2, complete rings is a synapomorphy of Anguidae + Varanoidea. Scalation Scalation is not informatively preserved on any of the extinct taxa. 269. Scalation: Canthus temporalis (0) scales no larger or more prominent than lateral and dorsal temporal scales, or barely so (Fig. 58A); (1) scales mark- edly larger and more prominent, can- thus well-developed (Fig. 58B). Evolution. Under both analyses, the ancestral state for the entire group and for Anguimorpha is a lack of we) 164 Bulletin of the Museum of Comparative Zoology, Vol. 160, No. 269(0) Figure 58. Temporal and tympanic regions, left lateral, anterior to the left: A, Xenosaurus newmanorum UF 25006. Illustrates characters 269(0), 270(1). B, Xenosaurus rackhami UTEP 4555. Illustrates characters 269(1), 270(2). Necks, dorsal, anterior to the left: C, Xenosaurus newmanorum UF 25006. Illustrates character 271(0). D, Xenosaurus rectocollaris UF 51443. Illustrates character 271(1). prominence. Under Rees I “thre Variation. In S. crocodilurus, small ancestral state for S. crocodilurus + juveniles have a tympanum that ap- Varanidae is cee and promi- pears naked to the eye (Sprackland, nence of the eanthas iss al synapomor- 1993; Miagdefrau, 1997: Bever et al., phy of the southern clade of Xeno- 2005). It is possible that a thin layer of saumus. “Under — Analysis) (25) the scales covers “it, as in juvenilemec ancestral state for Xenosauridae is grandis. Relatively large adults should ambiguous, and prominence is an ie scored for this Akane autapomorphy of L. borneensis. Evolution. Under both analyses, the 270. Scalation: Tympanum (0) unscaled: (1) nondifferentiated covering of thick covered with thin scales. becoming scales nakae by borneensis is an autapo- thinner toward middle of tympanum morphy of that taxon. Under Analysis 1, (Fig. 58A); (2) covered with thick scales the ancestral state for the entire group bie still differentiated from surround- and for Anguimorpha is ambiguous ing skin (Fig. 58B); (3) covered with between unscaled and covered with thick scales and undifferentiated from thin scales. A thick scaly covering is an surrounding skin (McDowell and Bo- autapomorphy of S. crocodilurus. The gert, 1954, plate 1). ancestral state for Anguidae + Helo- _Scalation: _Scalation: dermatidae is unscaled, and the ances- tral state for Xenosaurus is thinly scaled, with thickly scaled but differ- entiated a synapomorphy of the south- em place of Xenosaurus. Under Anal- ysis 2, the ancestral state for the entire group and for Anguimorpha is un- scaled, and increased scalation is a synapomorphy of Xenosauridae, whose ancestral state is ambiguous between lightly scaled and heavily scaled but o differentiated. neck (O) con- to underlying 58C); (1) pufts that neck ap- differentiated Skin around forms relatively tightly muscular structure (Fig. out considerably, such pears wide and little from back of head (Fig. 58D). Evolution. Under both analyses, the loose collar-like morphology is an autapomorphy of X. rectocollaris. Lateral fold (0) absent; (1) present, relatively weakly developed, discontinuous; (2) present, well-devel- oped, continuous along body. (Scored as lateral fold absent or present in Estes et al., 1988). Evolution. Under both analyses, lack of a fold is the ancestral state for the entire group and _ for Anguimorpha. Presence of a weak fold is a synapo- morphy of Xenosaurus, and a strong fold is a synapomorphy of the southern clade of Xenosaurus. Under Analysis 1, a well-developed fold is a synapomor- phy of E. multicarinata + O. ventralis. Under Analysis 2, the ancestral state for Anguidae is ambiguous among absent, weakly developed, and strongly developed. .Scalation: Dark markings on venter (0) 9 absent; (1) present peripherally: (2) present across most of venter. Evolution. Under both analyses, lack of markings is the ancestral state for the entire group and for Anguimorpha. Under Analysis 1, peripheral markings XENOSAUR PHYLOGENY ¢ Bhullar 165 are an autapomorphy of S. crocodi- lurus and the ancestral state for the southern clade of Xenosaurus is am- biguous among absence, peripheral presence, and extensive presence, as is the state for X. agrenon + X. rectocollaris. .Scalation: Epidermal ridge microstruc- ture (O) polygonal; (1) not arranged in regular polygons. Evolution. This character was scored from the work of Harvey (1991, 1993). Under both analyses, the ancestral state for the entire group and for Anguimor- pha is polygonal microstructure. Under Analysis 1, a lack of polygonal arrange- ment is a synapomorphy of Anguidae + Helodermatidae. Under Analysis 2, the ancestral state for Anguidae + Varanoi- dea is ambiguous. RESULTS Ingroup Topology and Effects of Ordering Analysis 1 and Analysis 2 yielded the same fully resolved topology for Xenosaurus and its extinct relatives (Figs. 6, 7). In the single recovered topology, R. rugosus is sister to all other xenosaurs, and E. eae ensis is sister to E. serratus + Xenosaurus. Within Xenosaurus, the northern clade of X. new- manorum + X. platyceps is sister to the southern clade, which is divided into two additional clades: X. agrenon + X. rectocol- laris and X. rackhami + X. grandis. Boot- strap values greater than 50% are given at the internal nodes. For both trees, all internal nodes for the ingroup had bootstrap values greater than 50%. Analysis 1 yielded a single most parsimo- nious tree with a length of 924 steps (Fig. 6). Of the 274 characters used in the analysis, 253 were parsimony-informative and 21 were parsimony-uninformative. The tree consistency index was 0.4708 (0.3147 rescaled and 0.4573 excluding uninforma- tive characters), the homoplasy index was 0.5292 (0.5427 excluding uninformative characters), and the retention index was 166 0.6685. The parametric bootstrap test re- sulted in values of greater than 50% for all ingroup nodes and for the nodes within the shinisaur clade. The other nodes were constrained. Analysis 2 yielded a single most parsimo- nious tree with a length of 875 steps (Fig. 7). Of the 274 characters used in the analysis, 253 were parsimony-informative and 21 were parsimony-uninformative. The tree consistency index was 0.4971 (0.3488 rescaled and 0.4836 excluding uninforma- tive characters), the homoplasy index was 0.5029 (0.5164 excluding uninformative characters), and the retention index was 0.7017. The parametric bootstrap test re- sulted in values of greater than 50% for all ingroup nodes, for Xenosauridae, for Angui- dae + Varanoidea, and for Varanoidea and all nodes therein. A comparison of the Analysis 1 and Analysis 2 trees using a Templeton test with the Analysis 2 tree as the Bet or uncon- strained instance (see Materials and Meth- ods) demonstrated that the toplogies were significantly different (P < 0.0001, Wil- coxon signed-rank statistic 556.0, N = 70, Z = —4,3500). As expected, the tree length of the unconstrained tree (Analysis 2) was less than that of the constrained tree (Analysis 1), and the consistency and retention indices were higher in the unconstrained tree, which also had a lower homoplasy index. The number of parsimony-informa- tive and parsimony-uninformative charac- ters was consistent between the two analy- ses, despite the differing outgroup topologies. Finally, in both analyses, boot- strap values for the eran nodes were all reater than 50%, with the lowest support the only values less than 80%) being for the Xenosaurus + E. serratus node and the Xenosaurus + E. lancensis node. The internode bounded by these nodes was the only branch to collapse when characters were run unordered, as described below. Ingroup topologies were identical between the two trees. When the matrix was run with all characters unordered, the internode be- Bulletin of the Museum of Comparative Zoology, Vol. 160, No. 3 tween E. serratus + Xenosaurus and E. lancensis + Xenosaurus collapsed. This somewhat unexpected result emphasizes that the position of R. rugosus as sister to all other xenosaurs here included is fairly robust, and that E. lancensis is known from specimens representing a relative paucity of phylogenetic information. At a glance, the loss of resolution of E. serratus + Xeno- saurus seems absurd; E. serratus is nearly identical to Xenosaurus in several unique features. Notable among these is the highly domed form of the maxillary osteoderms, as opposed to the primitively flat osteoderms of E. lancensis. That distinction is repre- sented in the scorings of character 62, a multistate character representing a morpho- cline from a flat plate-like morphology to a broken-up domed morphology. The nature of the problem lies in the status of E. lancensis as a “transitional form,” uniquely displaying a broken-up but undomed mor- phology. In an unordered scheme, the transitional nature of this morphology is not recognized, and it instead becomes simply an autapomorphy. Indeed, a simpli- fication of the character scoring to two states (not domed and Xenosaurus-like, or domed and Xenosaurus-like) results in a majority-rule consensus identical to the ingroup tree produced by the ordered data. Two other multistate characters, 46 (the steepness of the slanted dorsal margin of the lacrimal recess) and 108 (the lateral projec- tion of the cristae cranii), likewise have states that are unambiguous synapomor- phies of E. serratus + Xenosaurus, whereas E. lancensis has an intermediate state. Outgroup Topology The alliance of S. crocodilurus and the fossil B. ammoskius was established by Conrad (2005, 2006) and Conrad et al. (2011). My study also agrees with Conrad et al. (2011) in placing M. ornatus Klembara 2008 as sister to a Shinisaurus + Bahndwi- vici clade in a phylogenetic analysis. All other relationships were set using a con- straint tree to generate Analysis 1 (see Materials and Methods). However, Analysis 2 was generated by specifying only P. torquatus as an outgroup for the analysis and therefore generated a hypothe sis of anguimorph relationships (Fig. 7). Under Analysis 2, the initial split of Anguimorpha i is between the traditional Xenosauridae, in- cluding the Xenosaurus clade and_ the shinisaurs, and an Anguidae + Varanoidea clade. Anguidae consists of an Anguinae + Diploglossinae clade to the exclusion of Gerrhonotinae, and Varanoidea_ has_ its traditional topology of Helodermatidae + Varanidae, with Varanidae consisting of L. borneensis and Varanus. Character States Supporting Clades and Terminal Taxa Following are lists of character states supporting the focal clades in this study (synapomorphies) and terminal taxa (auta- pomorphies) under each hypothesis. Starred states (*) are unambiguous. Clade number- ing is arbitrary, but clades common to both hypotheses (clades 1 to 22) have the same number under both listings. Character states unique to one hypothesis are denoted with a capital “U” (not marked as such when state being transitioned from is different), and states that differ between analyses in ambiguity are denoted with a lowercase “u.” When dealing with scores of ?, accelerated transformation (ACCTRAN) optimization is assumed. My character descriptions are worded with neither accelerated transition nor delayed transformation (DELTRAN) optimization in mind, and they are thus more complete guides to character distribu- tion than the following lists. One exception obtains in the case of the character descrip- tions and in the lists below—when data are truly missing, instead of a character being inapplicable to the taxon (see Strong and Lipscomb, 1999), I use a DELTRAN-type assumption, placing ambiguous synapomor- phies at more exclusive nodes instead of assuming early transformation. Those trans- formations are listed in brackets following the other transformations optimized at the XENOSAUR PHYLOGENY ¢ Bhullar 167 nodes in question. Specifically, those nodes are all of those for which one branch is completely extinct and therefore represented by incomple te fossils: Xenosaurus + R. rugosus, Xenosaurus + E. lancensis, Xeno- saurus + E. serratus, S. crocodilurus + M. ornatus, and S. crocodilurus + B. ammoskius. Additionally, I use the same approach for the osteodermal characters of X. agrenon at the X. agrenon + X. rectocollaris node because data were unavailable regarding the osteo- derms of X. agrenon. Analysis 1 1. Xenosaurus + Restes rugosus: | 0-2*; 51 Oot. 52, 1-9 a3 OL. 59 02) UF. 64 6-7 U: 66 0-1 U*: 67 0-1*: 69 0-1 U*: 72 O2t*97 1-2-109 0-1*. Li Oal*: 1130-1 U: 114 0-1*: 131 0-1; 132 1-2: 145 0-1*: 184 0-1: 186 0-1 U*:; 191 0-1; 193 0-1 U*; 196 0-1 U*:; 214 0-1*: Restes rugosus: 43 2-0 U; 125 1-0 U; 144 0-1*: 204 0-1* 3. Xenosaurus + Exostinus lancensis: 62 2- ot: 93 1-0: 96:0_1*: 97 2-3*: 100021. 105 O-1) Us. 185 1-0: U; 193 122 (152°021); 4. Exostinus lancensis: 1 2-1; 92 1-0*; 147 4-G*: 154 1-0: 246 1-2* 5. Xenosaurus + Exostinus serratus: 46 O- 1*. 54 0-1; 62 3-4*; 93 1-0: 98 O-1*; 99 0- 1*. 108 0-1*; 191 1-0 [7 1-2; 8 0-3; 9 0-1: TRO-1e19 0-1. 22 021 VU 47 O21 Ue 55 3 2 U;: 63 1-0; 130 0-1; 68 1-2; 94 1-2 U; Lil 1-2; 124 1-2, 200 0-1 U; 209 0-1) 6. Exostinus serratus: 49 0-1*; 50 O-1*; 52 Zale 60 1-27. G61 123": 64 7-5*:-69:. 1-2*: 70 0-1*: 71 0-1: 98 1-2*- 101 O-1*: 106 O-1*: 107 (O=1%= 106: 12242 Tok 1262 132 2-1: 185 0-1 U; 192 0-1*; 193 2-0*:; 194 0-1* Xenosaurus: 2 0-1*; 8 3-4*: 9 1-2*. 10 0- 1; 17 0-17: 21 2-3*: 23 0-1*; 26 0-1*; 45 0-1*: 56 4-3; 57 0-1*; 58 O0-1*; 64 7-9*: 93 0-1*: 100 1-0; 104 0-1; 128 0-1*: 184 1-02.195- 120*.196 1-0*: 201)0-17 130 I- 2 3 0212 34-021: 3b Oke 37 021.-75.021: 65 051;86 0-1: 116 0-1: 117 0-1°U2 135 =e 136:0-12 137 0-1 U: 1389:0-1 Us 142 1; 145 1-2; 146 0-1; 147 4-3; 148 4-2; bo —~ 2 ()- 168 nO: br. Bulletin of the Museum of Comparative Zoology, Vol. 160, No. 3 159 0-1 U; 161 0-1; 162 0-1 U; 165 0-1; 167 1-2; 168 0-1; 169 0-3 U; 173 0-1; 199 0-1; 205 0-2; 207 0-1; 208 0-1; 211 1-2; 213 0-1 U; 228 0-1; 238 0-1; 243 0- 1; 248 3-2; 260 0-1 U; 261 0-1 U; 263 0- 1 U; 264 1-2; 265 0-1 U; 268 0-1 U; 272 0-1] _ Northern clade of Xenosaurus: 13 0-1*; 16. 0214s 18-02% 20 se0 tls =45 al 9% 98 021% e209 gle? 39 0a eA Oa AD, Qa1* 543,925" 24. 1-0 = 60 mlE0s noe O=B*: 68 22375 (5 On > Ie se Wome 22 79 ORF 284 021s Sor 25595) 0S LO: 022%. hh 2-33. 16. te 3s Os (= Le 143 0215. 154 12s Las 02l wire 02h Us ST 021s. 202, 02.19 e 0a 220 0-14, 22500 I 22680 aI 22 0- 139 OR AS eal eS Oe ol Oeil. 2os JOM ase 23 oo) wer: 200 Qa) Xenosaurus newmanorum: 56 3-5*; 65 Onl S812) 135 1225147 3-4 AS 2-08: 545223 1 742-055 AS Ie OS 2a Oe: 19S O21 223 10s O04 Ol s4 eos: 230 O-1e: 2350 0-260) I-Oes2 6a. l-Onue 268 120) uw Xenosaurus platyceps: Ol 1-22) SOn0rI: SOO SOA ETS OA OO ae 27 po pleOt 193° 2-3°2206 O12 14 es NS) Se DIS NA BY) (Us ilie O8yl 12 A000 24581202 25402258 = 202702 265) 1-0 2660-2 Or Os Southern clade of Xenosaurus: 11 1-2*: PCE lis OSI wes Beslle oy Iles Bis} Oe IA0 021 A309 Fl 65-2205) (4a aio We corey WEIL Ue S10) (lie IOPAC Ses IAS (alle 1347021 s 1367025. 147 sal a0 le 157 V2: Nos 1222 Se 12000 204 0a: 2-0 5242 OAS A> Oo = 250102 ues 252,051 2a s0=l Us 256 OSU 22 5n0- I= 260 We2 Us] 26402-3" 2265 Ne2 we: 268) 1=2*) 269) 021 Us 270 Ie2 Us 272 (ee 2738022 . Xenosaurus adgrenon + Xenosaurus rec- tocollaris: 46 1-2*; 56 3-2*; 88 1-0*; 135 1207; 1497120 WAG Te0e isa l-2e 6s 1-0* 170 (0212 OSI Sif. 1202876 Q-1*. 188 0-1*; 190 0-1*; 203 O-1*: 216 (-1*: 13. 14. 15. 16. igi iS: nD: Xenosaurus agrenon: 55 2-1*; 56 2-1*; TA .2-\= 76 120*: 77 0-17: 110021 ase 3: 1500-1*- 166 0-1*2 181 120: 19Sale: 2G ON 239) O=1 Xenosaurus rectocollaris: 8 4-0*; 16 O- 1*: 60 1-2*: 64 9-B*: 83 11-0287 2aln96 1-O0*: 123 O=1*: 131 120%: 145 221 eee [=2*; Zid. 1-0*: 229 O-1*2 234.1-052730 O17 27 10-1= 2732-0 [264.3244 Xenosaurus grandis + Xenosaurus rack- hami: 5 0-1*; 15 0-1*; 58 1-2*; 65 0-1*: §2.0=1*: 103: 0-1 112 Oa. soa 133 0-14: 139 1-2*- 141) O21: sea 155 1-0*:. 156 021*: 159 1-0-0. 16002 163 1-2: 167 2-3*. 16S 122%. 169re=2au: 172 O21 174 2A a7 9 Ol eo u: 197 O-1*. 22) O21. 237 22lsnee24s 2-3; 252 122%. 253) 122%. 26 eZ a wow 2-357) 2600-55 Xenosaurus rackhami: 56 3-4: 81 O-1*: 88) 1-4*: 126 1202 135 122) 40°02 ales 2-3: 153 O21" 165 2-37; 1667021. ee J-1*. 177 O21: 205 221. 240 G2 45 Be6t1 252 Q=3") 256.225 Xenosaurus grandis: 5D 2-1* 56 Zeon oe O21"; 116 122%: 120 O21 127 0A aie 1-22 157° 221. 180) 0-1 1S aeons 9-1 198 O2N* 255: Oz Shinisaurus crocodilurus + Merkuro- saurus ornatus: 20 I22* 2250- wes 9-0: 52 1-2 U* 155° 120%. 19302 196 0-1 U* Shinisaurus crocodilurus + Bahndwivici ammoskius: 47 0-1 U*; 98 0-1*; 147 5-6* [29 1-2; 55 3-2 U; 59 0-1 U; 66 0-1 U; 69 0-1 U2 117 0-1 U: 200 02 U2 2133 0nwe 233102 . Merkurosaurus ornatus: 7 1-0; 23 0-1*: * 92120* "94 te? U= Saale Oru . Bahndwivici ammoskius: 22. 1-0 U; 55 2-1*; 64 5-8* U; 97 1-3*: 134 L-OWU2 vol 2-3*; 154 1-0*; 240 0-1* . Shinisaurus crocodilurus: 5 O-1*;, 94 1- OF 124 021 U*) 138 0-1*; 147629743 A-3*; 152 0-1 u* [14 1-0 U; 46 0-2; 83 0- 1 U; 88 1-5 U; 93 0-1; 113 0-1 U; 126 0- 12136 0-1 Us 137 02) U; 139)015U Esse 0-1 U; 169 0-3 U; 174 2-1 U; 178 0-1 U; bo WwW 3. Xenosaurus + Anguidae: Zao: We2 Zoo. - 1s: 260 0-2. .U: 2OD0=21U. 266-120: 268 0-1 U: Us| 2 V=OFs 32; Oe 1*. 33 0-1; 43 0-2; 45 1-0: 56 2-4*: 60 O- 627082 (5: Ole 1240-1125 O21 129 OSA OH tt57 O=1< 1750-1 210 0-1: 202021 231 1-0; 249 3-2 U; 250 0-1 U; 25: (- ( 263 D3 Analysis 2 Ne 6. Xenosaurus + Restes rugosus: | 0-2*; 45 IOs 910-1 53 021". 60 0-1 UF 62 Of? UG O-1*: 72 0-1": 97 1-22 109 O- eal =A Oat) 3) 0-1: 132, 1-2: LA DIOET ss 19) O12 10 O21 UF: 214.051" Restes rugosus: 55 2-3 U*; 105 1-0 U; 136 1-0 U; 137 1-0 U; 189 1-0 U; 144 0- I* 165 0-1 UF, 200: 1-0 U*-; 204-0-1* Xenosaurus + Exostinus lancensis: 62 2- o*: 95 120: 96.021%: 97 2-3*- 100 0-1- 111 IO Oa AOS 122; (152 0212 157 Oe baie T84-0=1 | Exostinus lancensis: 1 2-1: O-GP 2154 10%. 246. 1-2" Xenosaurus + Exostinus serratus: 46 0-1*: 54 0-1; 62 3-4*; 93 1-0; 98 0-1*; 99 0-1*: 10802153191) 120: [7 122-8 0-3; 9: 0A® 11 OS 14 OSI 219 :021%43:0-2 U263 1-0: 68 $29 12412129 051 U2 13001: 209:0=1| Exostinus serratus: 49 0-1*:; 50 O-1*; 52 2-602 ON 1-364. 7-58: 69 il-2*: WO OF = 7102129530" 101 O=1= 106 O21 107 031) 1081-2" 13) 1-0: 132 2-1: 185 0-1 u*; 192 O-1*, 193 2-0*: 194 O=12 Xenosaurus: 2. 0-1*: § 3-4*: 9 1-2*: 10 0- ra Oa 223.23 O21 26 O-14245 O21*56 423.57 0-1% 58. 021%: 64 7-9*: 93 0-1*; 100 1-0; 104 0-1; 128 0-1*:; 184 1-0; 195 1-0*; 196 1-0*; 201 0-1* [30 1- 2 SO Oa 33'0- We 34-0- 1 35 0-1; 37 0- i775 O21: 78:0-h-U: 85021. 86 0-1; 88 5- UE G 02135 )- I 142 0-145 122: IAG 0s Ayes -32 1484-2 lol O=12 165 OAs LOE? 16Ss0- 12 1730-12 175. 0-1 U; 199 0-1; 205 0-2; 207 0-1; 208 0-1; Oe OO UE 2 2 2810-1 We72 2) 021228 0212 238 0=1--243 0-1. JAS O-2 204 1-2: 272 Oa 92 1-0*; 147 _—— 8. 2 10. Lt. 13. 14. XENOSAUR PHYLOGENY © Bhullar 169 Northern clade of Xenosaurus: 13 0-1*: 16:02). 18 0-1 20: 1-0* 21) 3-4%2 227. Ie F938 O21) 29 1-2--39 0-1": rf (-1*:; 42 O21: 43: 9-3*- 54 1-0- 60) 1-0#: 64 9-B*: 66:225°: vo OF: 7> 1-22.76 1-7 79: 0- PO: 1-0 Us 84 0-1 65: 1-2%. 9s 0-1*- 110 O-2F: HI OeOF V6 1-3F 192.0-1* 143 0-1*. 154 1-2*, 158 0-1*. 183 0-] Ue 138 7-0-1*- 202 0-219 02 1*-220 0- [se 2050-1 226 021". 229 0-1 239.0- It 9AS 2-17. 2A0 2-0; 250 1-0) U. 25] Q-1*. 2531-0 U; 255. 0-1*: 256 1-0 U: 25S 223" 259 1-0": 260 2-1 U: 265 2-1 Ue Zor 271*. 269 120 Ue 2702-1 Ue 273 1-0 U Xenosaurus newmanorum: 56 3-5*; 65 0-1 8S 1-2: 135 1-2: 147 3-4. 148 2-0#: 154 2-3": 174 1-0*: 18] 1-0; 182 1-0*: 19S 021" 225 J-047224 021%. 234 1-0*: 230 O-1*: 237 OF1*2 260 1-0*. 263 1-0 ut: 268 1-0-0" Xenosaurus platyceps: 61 1-2*; 80 0-1; 89 0-1*:; 118 0-1; 119 0-1*; 120 (- 1* 127 Oeil 135 1-0" 174 1-2. 193. 2-3*-206 Qales Di 1-2 lea 2 ODS ee 282 O-1*: 254 1-2%: 240 0-1; 248 120%: 254 Q-1*: 258 3-4*- 262 0-1: 265 1-0 u*: 266 £22F)267 I0z Southern clade of Xenosaurus: 11 1-2*: 120-17. 27 O21 wu: 36:0-1<. 37 1-2%288 02 IAG 02142435. 2-I. 68 2-0" 7 1222 75 1 2*. 90 O0-1*; 102 O-1*> 126 '0-I: 134'0-J*: 138 O-1% 147 3-1*: 1510-1: 157 122: 165 L2*) 178 10 U. DOA Olly 14 0: 2A? Onl 245 0212249 2-3 U: 2520-1: 257 O-1*; 264 2-3*; 268 1-2*; 272 1-2*> 273 1-2 . Xenosaurus agrenon + Xenosaurus rec- ram) tocollaris: 46 1-2*; 56 3-2*; 88 1-0*; 135 1-0*; 142 1-0; 147 1-0*; 155 1- 2°; 63. [- O*; 170 0-1*; 171 0-1*; 174 1-2 U: 175 1- 0: 176 0-1*; 158 0-1*; 190 0-1*; 203 0- 216 Q21* Xenosaurus agrenon: 55 2-1*; 56 2-1*; 74 2-1; 76 1-0*; 77 0-1*; 110 0-1*; 148 2- 350021: 166 0-1*: 1S 1-0; 198 0-1; 217 0-1; 239 0-1* Xenosaurus rectocollaris: 2-0*: 16 O-1*: 60) 1-2*. 64 9-B*; 83: 1-0; 87 2-I*; 96. 1- OF» 123. 0-1*; 131 1-0*; 148 2-1* 214 170 15. 16. le nS: iD: Bulletin of the Museum of Comparative Zoology, Vol. 160, No. 3 e225: 1=0*:-- 229) 0-1: 234 1-022 36 OF. 271 0-1: 273 220) P6453-41 Xenosaurus grandis + Xenosaurus rack- ham 5. O-l*: 15 O21 582 “651021 620214. 103 OS ID OAs taeda: 133) OL1*: 1391-22 Tals Os soiled. 155 1-0*: 156 0-1*; 159 120 ut. 160 0-1: 163 122% 167 223%: 6S 1E2* 69) 3-2: W772 O=1*) 179 O21 1S ME? is: 197 OA: DONO 22s tie 468 ale: 253 1-2 F 25612 * 260) 225432601245" Xenosaurus rackhami: 56 3-4; 81 0-1*: 88 1-4*: 126 1-0: 135 1-2: 140 0=1*: 148 2-3: 15302142 165 223%: 166.0212 169 2- Tt. 177 01% 205 221: 240022272497 3= Of: 252 2-33 256.2-3% Xenosaurus grandis: 55 2-1*; 58 2-3*; 84 O=1*: 116) 122%: 12010214 127 051% 147 1-2-9157 221-180. 021*5 1Si 12051938 2- 1* 19S 02145255. 021+ Shinisaurus crocodilurus + Merkuro- saurus ornatus: 20 1-2%* 252-0: 61 120 U*: 149 1-0 U; 151 0-2 U: 155 1-0*: 195 Jeo Shinisaurus crocodilurus + Bahndwivici ammoskius: 94 2-1 U: 98 0-1*; 147 5-6*:; 185 0-1 U*: [27 0-11 U: 29 1-2: 48 0-1 U: 56 422,068: 120) US 1s 02a 1267021: IES VEIL . Merkurosaurus ornatus: 7 1-0; 23 0-1*:; 47 1-0 U; 92 1-0* . Bahndwivici ammoskius: 22 1-0 U*; 55 2-1*; 64 5-8 U; 97 1-3*; 124 1-0 U; 151 2-3*; 154 1-0*; 240 0-1* . Shinisaurus crocodilurus: 5 0-1*; 64 7-5 U*; 94 1-0*; 134 0-1 U*; 138 0-1*; 147 6-9*; 148 4-3*; 152 0-1 U [32 1-0 U; 46 0-2; 93 0-1; 182 1-0; 188 0-1; 227 2-0; 230 1-0; 231 1-0; 253 1-2; 256 1-2; 258 2-1; 266 1-0] . Xenosauridae: 3 0-1*; 6 0-1*:; 7 0-1; 22 0-1*; 24 0-1*; 47 0-1; 52 0-2*; 59 0-1*; 66 0-1*; 69 0-1*; 83 0-1; 93 1-0; 94 1-2; 104 1-0; 105 0-1; 113 0-1; 117 0-1*; 124 0-1; 136 0-1; 137 0-1; 139 0-1; 147 4-5; 148 5-4*; 159 0-1*; 162 0-1*; 169 0-3*; 178 0-1; 181 0-1; 182 2-1*; 186 0-1*; 196 0-1*; 213 0-1*; 250 0-1; 253 0-1; 256 0- 1; 260 0-2; 261 0-1; 263 0-1; 265 0-2; 268 0-1; 269 0-1; 270 0-2*; 273 0-1 DISCUSSION Relation to Previous Studies and Taxonomic Issues Raised Under both starting hypotheses, xeno- saurs formed a clade within Anguimorpha when allowed to vary across the entire tree, supporting the monophyly of the ingroup of six extant species of Xenosaurus and three extinct taxa relative to the other included taxa. The relationship of the extinct taxa to Xenosaurus is thus consistent with historical descriptions of the fossils suggesting xeno- saur affinities. Of the two prior studies presenting explicitly phylogenetic hypothe- ses of relationships among extinct and extant xenosaurs, the position of R. rugosus outside of a clade including Xenosaurus and E. serratus is consistent with those of both Gauthier (1982) and Conrad (2005, 2008). Restes rugosus as sister to all other xeno- saurs is specifically consistent with Conrad (2005, 2008). However, E. serratus as the immediate sister to Xenosaurus was sug- ested by Gauthier (1982) but not Conrad 2005, 2008), who recovered a monophyletic Exostinus. Note that the relationships of xenosaurs were a primary focus of Gau- thier’s but not Conrad’s work (J. Conrad, 2008; personal communication). A single recent study suggests an alliance of the Mongolian Cretaceous taxon C. intermedia with xenosaurs (Conrad, 2008). A full analysis using the characters identi- fied in the present work will have to proceed after examination of fossil material, and in particular CT scans, of C. intermedia. More recent work has indicated that C. intermedia might not be an anguimorph (J. A. Gauthier, personal communication); thus, a broader spread of scleroglossan characters and taxa than used in this study might be required to establish its phyloge- netic position. As predicted by previous studies (Gau- thier, 1982: Estes, 1983; Conrad, 2008), I recovered Exostinus as a paraphyletic group consisting of two successive sister taxa to Xenosaurus. Notwithstanding concerns about the assignment of specimens to E. lancensis, that taxon would then require a new genus name, and Exostinus would become monotypic, including only E. serra- hiss propose to resurrect ‘he name Harpagosaurus, applied by Gilmore (1928) to a maxilla now referred to E. lancensis (Estes, 1964, 1983). A more formal defini- tion will require further study of known material of E. lancensis, which may repre- sent several taxa (Gao and Fox, 1996). As already discussed, the recovery of a monophyletic Xenosauridae in generating Analysis 2 is consistent with most morpho- logical phylogenetic hypotheses, but not with studies based on molecular structure or with hypotheses proposed. by Conrad (2005, 2008). Aside from this result, the close relationship of B. ammoskius to S. crocodilurus was again confirmed (Conrad, 2006). Moreover, M. ornatus, B. ammoskius + Shinisaurus as recovered by Conrad et al. (2011). That result provides support for the shinisaur affinities of the taxon (Klembara, 2008). Temporal Implications Exostinus serratus, sister to Xenosaurus, is nearly identical to the crown clade in most aspects of its known anatomy. Many of the character states appearing in the crown clade had thus arisen by the age of the Orellan sediments from which E. serratus was collected (Swisher and Prothero, 1990). Thus far, no extinct taxa that fall within the crown clade Xenosaurus have been identi- fied. This “genus” may be very ancient indeed, like some anguid “genera” (Estes, 1983; personal observation). Exostinus lan- censis, from the Late Cretaceous, is the oldest extinct taxon on the stem of Xeno- saurus according to this study. The known fossil record of Anguidae and Helodermati- dae, the putative sister groups to the xenosaur lineage in Analysis 1, are consis- tent with a Mesozoic split. Odaxosaurus piger is a primitive glyptosaur from the Cretaceous (Meszoely, 1970; Gauthier, 1982; Estes, 1983), and ae primitive eledeeeraride G. pulchra and P. nessovi XENOSAUR PHYLOGENY ¢ Bhullar ire are from the Upper Cretaceous of Mongolia and the Albian-Cenomanian of Utah, re- spectively. (If P. nessovi is indeed a helodermatid, its presence in the Early Cretaceous would suggest significant ghost lineages for xenosaurs and anguids.) The fossil record of shinisaurs extends back to the late Paleocene/early Eocene (Smith, 2006b), leaving a longer ghost lineage for the shinisaur branch of the Xenosauridae in Analysis 2. However, a number of North American Cretaceous “platynotan” taxa known from fragmentary remains have yet to be fully examined in a phylogenetic context, and among these rare be found art of the missing ‘shinisaur lineage (Estes, 1983; K. T. Smith, personal communication; personal Observation). Some notable stratigraphic incongruities are present in the phylogenetic hypotheses recovered here. The Paleocene R. rugosus is sister to all remaining xenosaurs, including the Cretaceous E. lancensis. The R. rugosus lineage has yet to be recovered from the Mesozoic. However, R. rugosus itself ap- ears to be exceedingly rare in Paleogene eee (Smith, 2006b; K. T. Smith, personal communication), and this may apply to its predecessors, as well. The secon ‘nconge ity is the closer relationship of B. ammoskius from the North American Eocene to the extant Asian S. crocodilurus than Merkur- osaurus ornatus from the European Mio- cene (Conrad, 2006; Klembara, 2008). This might suggest an early Paleogene transat- lantic dispersal of the Merkurosaurus line- age, much like that which has been suggested to have resulted in the appear- ance of the primitive helodermatid E. gallicum in the Eocene of France (Hoff- stetter, 1957) and the iguanian Geiseltaliel- lus in the Eocene of Germany (K. T. Smith, personal communication). One would then expect to find the Merkurosaurus lineage in earlier European deposits and in Early Paleogene or Mesozoic North American deposits. The former has not yet been reported, but the latter expectation may be fulfilled by the host of shinisaur-related taxa being identified from screenwashed early Eocene and late Paleocene North American microfaunas (Smith, 2006b). Alternatively, in the case of both shinisaurs and heloder- matids, the North American and European representatives could have been derived from an unknown Asian stock. Primitive helodermatids were present in present-day Mongolia during the Late Cretaceous (Gao and Norell 2000). Finally, taxa farther down the stem of Xenosaurus than R. rugosus are notably lacking. As noted already, “relatively primitive helodermatids and shinisaurs have been found, but another frustrating absence exists along the stem of Anguidae. The highly dearca glyptosaurs are the only putative stem anguids known, and even that placement is not strongly supported, for they already possess fest of the derived features of anguids (Conrad, 2005, 2008: Conrad et al., 2011; a A. Gauthier, personal communication). Biogeography of Xenosaurs There is a striking difference in latitude between the locations of collections of the stem xenosaurs in this study, all of which were found in Colorado, Wyoming, and farther north still, and the present distribu- tion of Xenosaurus in central and southern Mexico. That pattern obtains for other squamate taxa, as well—notably diploglos- sine anguids (Gauthier, 1982; Smith, 2006b; personal observation) and polychrotine iguanians (Smith, 2006a, 2006b). The ap- parent contraction of the ranges of parts of clades has been convincingly attributed to the contraction of megathermal climate zones during the global cooling following the Paleocene/Eocene thermal maximum by Smith (2006b). Xenosaurus are, with few exceptions, crevice-dwelling lizards and are distributed along the great north—south-extending mountain ranges of Mexico. The initial split within Xenosaurus is a division between a northern clade, consisting of X. newma- norum and X. platyceps in the Sierra Madre Oriental, and a southern clade, consisting of the remaining taxa (King and Thompson, Bulletin of the Museum of Comparative Zoology, Vol. 160, No. 3 1968; Canseco Marquez, 2005), of which the X. agrenon + X. rectocollaris clade is more to the west and the X. grandis + X. rackhami clade is more to the east, extend- ing into Central America. Xenosaurus rack- hami has a particularly wide distribution, and I observed more intraspecies variation in that species of Xenosaurus than in the others for which I had sample sizes larger than one or two. Further details of the biogeography of Xenosaurus were provided by Canseco Marquez (2005). Comments on Ingroup Clades Xenosauridae. Analysis 2, the uncon- strained analysis, strongly recovered a monophyletic Xenosauridae in the classical sense, as also suggested, but not cladistically tested, by Smith (2006b). Many of the characters supporting the clade were not used by Conrad (2005, 2008). Two charac- ters, the subpalpebral fossa and the medio- laterally expanded facet on the maxilla, are present in association with slightly different elements in the two taxa, but this does not invalidate their potential homology accord- ing to the tree. Several other characters, including the upfolded tab of the facial process of the maxilla, are unique within Squamata. Part of the difficulty of resolving the degree of relatedness between xeno- saurs and shinisaurs is the poor fossil record along the stem of Anguimorpha. It is quite possible that the character states that seem so peculiar to the classical Xenosauridae are in fact ancestral for Anguimorpha as a whole. Xenosaurus + Restes rugosus. The most inclusive clade of the xenosaur lineage already has the characteristic dentition of the group. Among the bones surrounding the nasal capsule, the maxilla shows a combination of plesiomorphic and derived features, notably a primitively platey, con- tinuous osteodermal covering. The palpe- bral and prefrontal already display typical xenosaur morphologies. Likewise, the fron- tal and jugal bear several synapomorphies with the other xenosaurs, but they are primitive in various respects, as well. What is preserved of the palate is intermediate between Xenosaurus and the remainder of Anguimorpha, in particular the persistence of remnants of the pterygoid dentition. Restes rugosus is the only fossil xenosaur to preserve any of the palate. The primi- tively long, narrow postorbital suggests that the suprate mporal arch was not expanded and heavily ornamented as in Xenosaurus. It is thus likely that the general flattening of the head and possibly ithe body ident in the crown clade was not as deve ‘loped i in the ancestor of this more inclusive clade. Differences in synapomorphies support- ing the clade between Analysis 1 and Analysis 2 have largely to do with the issue of xenosaurid monophyly. In Analysis 1, several features shared by xenosaurs and shinisaurs optimize as convergent and are added to the apomorphy list for Xenosaurus + R. rugosus. Restes rugosus. Restes rugosus displays a host of plesiomorphic characters for xeno- saurs, exhibiting few obvious autapomor- phies in its iown anatomy. The most obvious of its autapomorphies is the unusu- ally large angle of divergence of the medial and lateral edges of the palatine process of the pterygoid. It is unclear whether that morphology indicates a similarly peculiar morphology for the remainder of the palate, which is not preserved. In Xenosaurus, despite the widening of the head, the palatine process of the pterygoid is not particularly expanded. Xenosaurus + Exostinus lancensis. Exo- stinus lancensis is not a well-known taxon, and this clade is supported largely by the more broken-up osteoderms on the maxilla and the Xenosaurus-like domed osteoderms of the frontal. The frontal is still uncon- stricted interorbitally compared with E. serratus and Xenosaurus, but the cristae cranii approach each other more closely than in R. rugosus. Exostinus lancensis is the only fossil xenosaur, unfortunately, to pre- serve the parietal. The supratemporal pro- cesses are broken and so could not be XENOSAUR PHYLOGENY ¢ Bhullar 173 scored, but they appear to have been short as in Xenosaurus. Exostinus lancensis. This is a difficult taxon for reasons already stated, relating to incomple teness and difficulty in the assign- ment of spe cimens. Its recovered position closer to the crown clade than R. rugosus requires a reversal in tooth form from slightly bicuspid to unicuspid. However, de spite the assertion of the most recent description that all teeth in E. lancensis are unicuspid (Gao and Fox, 1996), some AMNH specimens I examined for fie study show an apical, longitudinal groove like that which extends basally font the division between the smaller mesial and the larger distal cusps in those xenosaurs that have a bicuspid morphology (e.g., AMNH 15366). The apices of the teat of all of heck specimens are damaged. Furthermore, al- though the skull nook osteoderms of E. lancensis are domed like those of E. serratus and Xenosaurus, their form is unusual. They are oval or obovate and on the parietal show a concentric distribution unique to E. lancensis. Xenosaurus + Exostinus serratus. These two taxa are nearly identical in many aspects of their osteology, although only the Sosy portions of the skull are known for serratus, and these incompletely. The os- teoderms in general are of ie characteristic Xenosaurus Hoe and distribution. The dentition also has the form seen in primitive parts of the crown clade. However, the nares are primitively large and elongate; concomitantly, the ees process of the premaxilla, although expanded, is not so expanded as in Onsite us, and the slope of the narial margin of the maxilla is gentler and longer. lmforknmaton: premaxillae are Galaneven from other fossil xenosaurs. Exostinus serratus. This is essentially a short-faced xenosaur, its rostrum autapo- morphically short, reflected in features such its reduced tooth count. Additionally, the unusually wide palatal shelf of the maxilla suggests a particularly wide skull with the iaeoallas diverging from each other at a high angle. The tok of a posterior expansion of 174 Bulletin of the Museum of Comparative Zoology, Vol. 160, No. 3 the jugal is another strange autapomorphy. The various unique features of this taxon raise the question of whether xenosaurs were relatively diversified earlier than the crown radiation. In turn, it is unclear where less autapomorphic lineages more similar to the crown might be found. Xenosaurus. Xenosaurus, in part because constituents are extant and thus much more completely known, is supported by a large number of synapomorphies. Notable are the eases ecu of the nasal process of the premaxilla and the squamosal and the relative mediolateral widening of the parie- tal. Several bones, such as the septomaxilla and coronoid, are expanded and _ enclose more nerves and vasculature than the primitive state. The body is flattened and the neural spines relatively low; various other axial and appendicular synapomorphies also exist, such that the postcranium of Xeno- saurus is fairly distinct among anguimorphs. Northern clade of Xenosaurus. The north- ern clade of Xenosaurus, widely separated geographically from the remaining species, retains a number of ancestral characters that the southern clade has lost, notably the proportions of the skull roof bones. How- ever, the taxa within the northern clade are united by a consistent suite of synapomor- phies—more than the southern taxa. Among these synapomorphies are several features of the anterior maxilla and premaxilla. Furthermore, the heavy osteodermal armor of the northern clade optimizes as primitive if the heavily pene helodermatids and anguids are the immediate sister taxon to xenosaurs. However, it optimizes as derived or ambiguous if the more lightly armored shinisaurs are used. Xenosaurus newmanorum. This is a large- bodied species (King and Thompson, 1968). Other than its large size, it is in general the less autapomorphic of the two examined species from the northern clade. It has a particularly tall head for Xenosaurus, possi- bly a primitive feature (Herrel et al., 2001). As the less autapomorphic part of the northern clade, which itself retains a num- ber of ancestral characters, X. newmanorum might be a better taxon to include in phylogenetic analyses than the commonly used X. grandis and X. platyceps (e.¢., Conrad, 2008, who noted the shortcomings of scoring a composite “Xenosaurus” from X. grandis and X. platyceps—although these two taxa do bracket the clade). Xenosaurus platyceps. This is the flattest species of Xenosaurus, in head and body. Several autapomorphies of the maxilla and skull roof relate to the particularly flat, wide head of the taxon, | unusual morpholo- gies of the axial and appendicular skeletons might also relate to this marked dorsoven- il compression. Southern clade of Xenosaurus. The south- ern clade of Xenosaurus is united by a large number of synapomorphies, most of them relatively subtle, such as the proportions of osteodermal sculpturing and proportions of articular parts of the dermal cranial roof and sidewall. The ranges of some of the taxa within are considerably greater than those of the species within the northern clade (i.e., X. rackhami; King and Thompson, 1968), and the individual species as a whole are more HE Cue divergent than the relatively similar X. newmanorum and X. platyceps. Xenosaurus agrenon + Xenosaurus recto- collaris. This clade, consisting of two little- known taxa for which I had but one specimen each, is supported by the fewest synapomorphies of the clades within Xeno- saurus. Nevertheless, some of these syn- apomorphies (e.g., the unusual notch in the posterior region of the parietal, the mor- phology of the central region of the basisphenoid, and the bizarre flattened neural spines of the lumbar region) are striking and unique, and the clade appears robust. In general, this clade shares several ancestral character states with the northern clade that are no longer present in the well- supported X. rackhami + X. grandis clade. In contrast to the meager osteodermal armor of the latter clade, the X. agrenon + X. rectocollaris clade shares relatively heavy armor with the northern clade. Xenosaurus agrenon. This species is particularly little known, appearing exter- nally very similar to X. grandis, but inter- nally sharing a number of synapomorphies with X. rectocollaris. A distinct form of the supratemporal arch is one of the few autapomorphies distinguishing it from its common ancestor with that taxon. Xenosaurus rectocollaris. This is the most unusual of the species of Xenosaurus at first glance. Its ae is particularly short and stout, although the shortening appears to involve the postorbital dermal bones of the adductor/otic region, instead of the bones surrounding the nasal capsule as in E. serratus. Externally, the taxon is distin- guished by a bold dark-on-light color pattern different from the light-on-dark patterns of the other species of Xenosaurus, and it bears a strange cuff of puffy tissue around its neck. It would be interesting to investigate oe ecological correlations of the singular anatomy of this animal. Xenosaurus rackhami + Xenosaurus grandis. These two species share a number of synapomorphies, most strikingly a dra- matic reduction of cranial osteoderms (post- cranial osteoderms are absent). Certain other features distinct to Xenosaurus, such as the anteroposterior expansion of the tip of the jugal and the mediolateral expansion of the supratemporal arch, are at their most extreme in these two taxa, which are also perhaps the best represented in U.S. museum collections. Xenosaurus rackhami. This species is slender of head and body compared with other Xenosaurus, and some. individuals have particularly flat heads. In X. rackhami, the reduction of osteoderms is the extreme among examined Xenosaurus, and several peculiar autapomorphies of the skull roof and supratemporal bar, as well as the dentition, also obtain. Xenosaurus grandis. This is a relatively unusual species of Xenosaurus, stout and robust where its sister taxon is slim and slight. The reduction of osteoderms is less extreme in X. grandis, and some of the autapomorphies of the taxon relate to its generally stout form. Xenosaurus grandis is fhe most common Species of Rene auirus in XENOSAUR PHYLOGENY © Bhullar 75 U.S. collections, and the most heavily figured and described (notably by Barrows and Smith, 1947). It is generally used as the exemplar for Xenosaurus, and sometimes for the classical Xenosauridae as a whole (Wever, 1978; Estes et al., 1988). Consid- ering ihe large number of apomorphies between X. grandis and the Xenosaurus ancestor, it is not the ideal choice. Only one recent study (Conrad, 2008; extended Conrad et al., 2011) included a prudent combination of X. grandis and X. platyceps, bracketing Xenosaurus. Morphological Characters in Phylogenetic Analysis and Specific Issues of Character Evolution My analysis includes an unusual number of characters for such a restricted group of lizards (compared with, e.g., Rieppel and Zaher [2002] for uropeltid snakes, a group of similar size and high specialization). The surfeit of characters owes in part to my discovery of a number of informative fea- tures in previously unappreciated elements, such as the septomaxilla and the palpebral— these discoveries in turn owing to the availability of disarticulated skeletal material. Characters such as small foramina in these diminutive bones were remarkably invariant within taxa, and the fact that this seems counterintuitive suggests that a large number of characters dealing with subtle dicenenees in anatomy are arbitrarily neglected in hes anatomical phylogenetics. Panher work, i particular, on subtle features of ieee ed bones, could be immensely profitable in increasing the number of gross-scale charac- ters to ee greater parity with the size of molecular- sbaler datasets (Chippindale and Wiens, 1994). Already, work on subtle and tiny foramina dotting the dermatocranial elements of certain mammals has shown that the clustering of these foramina are consis- tent within taxa and appear to be phyloge- netically informative (Wible, 2003 . Kearney et al., 2005). In addition to hypotheses of relationships, phylogenetic analyses provide a thorough 176 catalogue of anatomical changes in the structures utilized, whether at “the TOSS, histological, or molecular scale—they are in some ways “shorthand” descriptions of anatomy. Several characters or character systems here examined stand out as poten- tially fruitful for additional study. The enclosure of the ethmoid nerves within the premaxilla (character 6) is unusual within Squamata. Ness interesting that xenosaurs show both this character Fand very heavy osteodermal sculpturing—as do very old individuals of E. multicarinata. The appear- ance of xenosaur-like features in very late stage E. multicarinata might suggest a heterochronic relationship, wherein some characters of xenosaurs are peramorphic with respect to those of anguids and perhaps other anguimorphs. The increased number of foramina in the premaxilla (characters 7 and 8) and maxilla (characters 53 and 54) in xenosaurs might suggest an increased acuity of integumen- tary sensation, perhaps ‘related to the decreased utility of sight in the dark environments frequented at least by indi- viduals of Xenosaurus. Increased numbers of foramina on the face have been suggested to imply great tactile acuity in amphishae- nians, as Sal (Kearney et al., 2005). The increased diameter of the infraerbial canal within xenosaurs (character 51) indicates an increase in the size of the contained neurovascular bundle, which includes the V2 division of the trigeminal nerve. An enlarged canal for V3 is associated with the enormously elaborated sensory capabilities of the platypus snout (Rowe et al., 2008). The various septomaxillary characters identified herein are remarkably phyloge- netically useful and consistent with regard to variation, considering the general neglect of this skeletal olnient in ale non- arake squamate literature. Clearly the septomax- illa is complex and ev olutionarily labile, and I predict that it will prove a rich source of characters for other clades, as well. The morphology in Xenosaurus, with its fully enclosed medial and lateral canals, is especially remarkable, and one wonders at Bulletin of the Museum of Comparative Zoology, Vol. 160, No. 3 the possible soft tissue and sensory corre- lates. The complete enclosure of the lateral canal in Xenosaurus and Heloderma recalls the sister taxon relationship between those taxa suggested by some analyses in molec- ular-scale studies (Townsend et al., 2004). Several transformations in the temporal region, notably the lateral and ventral expansion of the postorbital and squamosal and the anteroposterior expansion of the jugal, occur along the stem of Xenosaurus. Seemingly accompanying these changes are the strong surangular crest and the subcor- onoid_ fossa. Poco these features are variously related to the enormously elabo- rated adductor musculature of Xenosaurus (Haas, 1960). If they are so related, they may together represent a complex of characters that are not fully independent from each other. Finally, Xenosaurus shows a lack of depression in the posterior end of the quadrate for the tympanic cavity (character 173). Varanidae also display this character, but the quadrates of those taxa are heavily modified. Possibly the lack of depression is related to a reduction of the tympanic cavity also associated with the scaling over (and thus apparent functional impedance) of the tympanum (character 270). However, the depression remains in Shinisaurus, ane: tympanum is also scaled over in adulthood. Character nonindependence is a recur- ring problem in phylogenetic analysis (Martins and Garland, 1991; Schaffer et al.. 1991: Huelsenbeck and Nielsen, 1999: McCracken et al., 1999). Given the level of integration among the parts of a multicel- lular (or unicellular) organism, it is unlikely that characters can ever be fully indepen- dent. Nevertheless, studies such as mine could profit from careful analysis of character correlation, taking into account ontogeny and the soft Bese relations of hance The character list I provided is arranged by sensory capsule, for example, Leeaece most dermatocranial elements are involved in sheathing one of the three capsular regions (with he otic region being more complicated because of the “additional presence of the adductor chamber). Signif- icant changes in the basic structure of one of the sensory capsules surely would produce comple mentary changes in multi- ple sheathing bones. C pcile ring the ingroup of ‘ie study, a few other pote ntial sources of nonindependence arise. The head of Xenosaurus is unusually wide, and this general change may be related to the mediolate: ‘al expansion of several der- matocranial elements and even the capture of previously external vascular and nervous structures by the septomaxilla. In E. serratus, the shortness and width of the snout could be related to several autapo- morphies of the taxon. Osteodermal devel- opment obviously shows general trends, with most regions of peicodcuine reduced, for example, in X. grandis + X. rackhami. However, these trends are not always uniform, preventing the scoring of a single combined character for the various poten- tially related transformations. For instance, cranial osteoderms are best developed in X. newmanorum and X. platyceps, especially in the former. Limb osteoderms, however, are more prominent and numerous in X. platyceps, whereas caudal osteoderms are more developed in X. newmanorum. Thus, neither taxon can been scored overall as having the most “highly developed” osteo- derms, even. if eal and postcranial categories are established. Finally. an initial for ay into studies of the intraspecies variation of characters is made herein, although the attempt is basically limited to an effort to justify character selection and scoring. A additional work needs to be done on intraspecies variation in all vertebrates, for this variation is the evolution (Darwin, 1859: Bever, 2006 and references therein). In particular, com- pared with the attempts made here, further studies must incorporate larger sample sizes, greater ontogenetic spreads, and careful control of “localities/populations and the temporal aspect of collection. Xenosaurus provides an interesting case study for population studies hecause pop- great deal of raw material of ~~ XENOSAUR PHYLOGENY ¢ Bhullar ee ulations of the species are spatially restrict- ed, often occupying a_ single rock cliff (Ballinger et al., 2000). That spatial restric- tion may account for the relatively large number of species of Xenosaurus, and it is not clear whether significant gene flow occurs among isolated populations (Lemos- Espinal et al., 2004). On a larger scale, aspects of variation can provide additional characters for phylogenetic analysis. In short, morphological characters for phylo- genetic analysis are by no means exhaust- ad. even in the most heavily studied clades A push for careful, thorough anatanes! analysis of as much of the body as possible, ranging from the microscopic scale to aspects of variation and behavior, will yield a vast number of additional characters. The limited exercise provided by this work demonstrates that even a thorough, bone- by-bone analysis of the skeleton Alone can yield hundreds of novel characters and produce a fully resolved phylogeny with high support values. In this case, congru- ence with DNA-based analyses allows further confidence in the results, but considering the number of the characters and the high support for the morphological tree, incongruence would necessitate care- ful aeeone derauon not only of the mor- phological but also of the molecular data and analyses. ACKNOWLEDGMENTS This work is dedicated to my first advisor and colleague, my AP Biology teacher Eric Kessler, faculty of science, Blue Valley North High School, Overland Park, Kansas. Also to my mother, Amarjit K. Bhullar, for showing me the way, in the words of Harvard’s alma mater, “for Right ever bravely to live.” For financial support, I am grateful to the Jackson School of Geosciences at The University of Texas at Austin, the Donald D. Harrington Foundation, National Sci- ence Foundation (Graduate Research Fel- lowship), and Harvard University (James Mills Peirce Graduate Fellowship). For 178 smaller grants, I thank Yale University, the Jackson School of Geosciences, Sigma Xi, the Society of Systematic Biologists, the American Society of Ichthyologists and Herpetologists, and the American Museum of Natural History Roosevelt Fund. For allowing me to borrow unusual amounts of material pertaining to my thesis and to other projects, I thank Drs. Matt Carrano (USNM); Mark Norell and Carl Mehling (AMNH-VP); Kenney Krysko (UF); Jens Vindum (CAS); Jim Hanken, Jonathan Losos, and José Rosado (MCZ); Travis LaDuc and Dave Cannatella (TNHC); Jim Mead (NAUQSP-JIM); Jimmy McGuire (MVZ); Greg Schneider (UMMZ); Maureen Kearney and Alan Resetar (FMNH); Jacques Gauthier (YPM-VZ and YPM-VP); Greg Watkins-Colwell (YPM-VZ); Walter Joyce (YPM-VP); Jon Campbell and Carl Franklin (UTA); Gal Lieb and Bob Webb (UTEP); Gregg Gunnell and Jeffrey Wilson (UMMP); Amy Henrici (CM); and Darryl Frost (AMNH-H). I would also like to thank Jacques Gauthier, Jessie Maisano, Maureen Kearney, and Olivier Rieppel for allowing me to use their CT scans of the heads of X. grandis and X. platyceps. For help in specimen preparation and scanning, I thank Jessie Maisano, Matt Colbert, Rich Ketcham, Marilyn Fox, and the late Bob Rainey. In addition, I thank the following fellow students for discussion and support: Krister Smith, Gabe Bever, Rachel Dunn, Tyler Lyson, Taka Tsuihiji, Walter Joyce, Jason Downs, Brian Andres, John VandenBrooks, Matt Benoit, Alana Kawa- kami, Ted Macrini, Murat Maga, Heather Ahrens, Alicia Kennedy, Agustin Scanferla, Jack C aed and Yi Hongyn. My colleagues in Mexico, Luis Canseco Marquez, Juli io Lemos-Espinal, and Adrian Nieto Montes de Oca discussed xenosaur phylogeny and ecology with me extensively. The master’s desis of which this work is the majority was enabled by a series of excellent advisors. The first of these was a Kessler, my high school biology teach- - who fostered my interest in or ganismic bile and herpetology. Second came Bulletin of the Museum of Comparative Zoology, Vol. 160, No. 3 Jacques Gauthier at Yale, who trained me in vertebrate anatomy aoe squamate sys- tematics, as well as phylogenetic methodol- ogy. Te Rowe at The ances | of Texas at Austin further assisted in my development as an anatomist and a phylogenetic thinker. My primary advisor, Chris Bell, supported me in numerous ways, not least by offering the use of his extensive collection of squamate skeletons. Finally, Arhat Abzha- nov is my current advisor at Harvard. His flexibility and patience have allowed me to begin work in a new field without sacrificing the old. Jack Conrad and Keqin Gao_ provided enormously helpful reviews of this work, for which they have my thanks and my sympa- thy. LITERATURE CITED ANANJEVA, N. B., E. M. Smirina, AND N. G. NIKITINA. 2003. Dentition of Phrynocephalus melanurus. Does tooth number depend on body size and/or age? Russian Journal of Herpetology 10: 1-6. ARISTOTLE. 350 BCEa. Categories. The Internet Classics Archive; c1995-2001 [cited 2008 August 25]. Available from: http://classics.mit.edu/Aristotle/ categories. html. ARISTOTLE. 350 BCEb. Physics. The Internet Classics Archive; c1995-2001 [cited 2008 August 25]. Available from http://classics.mit.edu/Aristotle/ physics. html. Ast, J. C. 2001. Mitochondrial DNA evidence and evolution in Varanoidea (Squamata). Cladistics 17: 2226: BALLINGER, R. E., G. R. SmirH, AND J. A. LEMos- Espinau. 2000. Xenosaurts. Catalogue of American Amphibians and Reptiles 712: 1- 3, BaraHona, F., AND L. J. BarBADILLO, 1998. Inter- and intraspecific variation in the post-natal skull of some lacertid lizards. Journal of Zoology of the Zoological Society of London 245: 393-405. Barrows, Ss. AND H. M. Smiru. 1947. The skeleton of the lizard Xenosaurus grandis (Gray). The Univer- sity of Kansas Science Bulletin 31(2): 227-281. Beuuairs, A. p’A. 1949. Observations on the snout of Varanus, and a comparison with that of other lizards and snakes. Journal of Anatomy 83: 116- 146 + 1 plate. BERNSTEIN, P. 1999. Morphology of the nasal capsule of Heloderma suspectum with comments on the systematic position of helodermatids (Squamata:). Acta Zoologica (Stockholm) 80: 219-230. Bever, G. S. 2005. Variation in the ilium of North American Bufo (Lissamphibia; Anura) and _ its implications for species-level identification of fragmentary anuran fossils. Journal of Vertebrate Paleontology 25: 548-560. Bever, G. S$. 2006. Studies on post-natal variation and variability in the vertebrate skeleton and_ its paleontological implications [unpublished Ph.D. dissertation]. Austin: The Univ. of Texas. Bever, G. S. 2008. Comparative growth in the postnatal skull of the extant North American turtle Pseud- emys texana (Testudinoidea: Emydidae) Acta Zoologica 89: 107-131. BEVER, G. S., C. J. BELL, AND J. A. Matsano. 2005. The ossified braincase and cephalic osteoderms of Shinisaurus crocodilurus (Squamata, Shinisauri- dae). Palaeontologia Electronica 8: 1-36. Buutiar, B.-A. S. 2006, Postnatal ontogenetic changes in the cranium of Varanus exanthematicus with compar- isons to other varanoids and applications to the fossil record. Journal of Vertebrate Paleontology 26: 42A. BuuLiar, B.-A. S. 2007. The enigmatic fossils Exostinus and Restes: resolving the stem and the crown of Xenosaurus, the knob-scaled lizards. Journal of Vertebrate Paleontology 27: 48A. Buuiar, B.-A. S. 2008. Anatomy and phylogeny of Xenosaurus and its extinct relatives [unpublished M.S. thesis]. Austin: The Univ. of Texas. Buutiar, B.-A. S. 2010. Cranial osteology of Exostinus serratus (Squamata: Anguimorpha), fossil sister taxon to the enigmatic clade Xenosaurus. Zoolog- ical Journal of the Linnean Society 159: 921-953. BuHutar, B.-A. S, AND K. T. Smitru. 2008. Helodermatid lizard from the Miocene of Florida, the evolution of the dentary in, and comments on dentary morphology in, Varanoidea. Journal of Herpetology 42: 286-302. Borsuk-Biatynicka, M. 1984. Anguimorphans and related lizards from the Late Cretaceous of the Gobi Desert, Mongolia. Palaeontologia Polonica 46: 1-105 + 13 plates. Canseco MArguez, L. 2005. Filogenia de las Lagartijas del Genero Xenosaurus Peters (Sauria: Xenosaur- idae) Basada en Morfologia Externa [unpublished M.S. thesis]. Mexico City: Univ. Nacional Auton- oma de Mexico. CHIPPINDALE, P. T., AND J. J. Wiens. 1994. Weighting, partitioning, and combining characters in phyloge- netic analysis. Systematic Biology 43: 278-287. Conrap, J. L. 2004. Skull, mandible, and hyoid of Shinisaurus crocodilurus Ahl (Squamata, Angui- morpha). Zoological Journal of the Linnean Sacre ty 141: 399-434. Conrap, J. L. 2005. Shinisaur osteology and the evolution of Squamata [unpublished Ph. D. disser- tation]. Chicago: Univ. of Chicago. Conrapb, J. L. 2006. An Eocene’ shinisaurid (Reptilia, Squamata) from Wyoming, U.S.A. Jounral of Vertebrate Paleontology 26: 113-126. Conrab, J. L. 2008. Phylogeny and systematics of Squamata (Reptilia) bated on morphology. Bulletin of the American Museum of Natural History, 310: T1822; XENOSAUR PHYLOGENY ¢ Bhullar 179 ConraD, J. L., J. C. Ast, S. MONTANARI, AND M. A. Nore. 2011. A combined evidence phylogenetic analysis of oe eae ot (Reptilia: Squamata). Cladistics De 2021 ic Darwin, C. 1859. On the Origin of Species by Means of Natural Selection. Loader: John Murray. Deep Scary Project, THe. 2007. “Elgaria multicari- nata,’ Digital Morphology [Internet]. Austin, Texas: DigiMorph; ¢2002 [cited 2008 August 28]. Available from: http://digimorph.org/specimens/ Elgaria_multicarinata. Epmunb, A. G. 1969. Dentition, pp. 117-200. In C. Gans, A. D’A. Bellairs, and T. S. Parsons (eds.), Biology of the Reptilia. Vol. 1. London and New York: Academic Press. Estes, R. 1964. Fossil vertebrates from the Late Cretaceous Lance Formation eastern Wyoming. University of California Publications in Geological Sciences 49; 1-180 + 5 plates. Estes, R. 1965. Notes on some Copeia 1965: 104—106. Estes, R. 1983. Sauria Terrestria, Amphisbaenia. Stuttgart; New York: Gustav Fischer Verlag. Estes, R., K. DE QUEIROZ, AND J. GAUTHIER. 1988. Phylogenetic relationships within Squamata, pp. 119-281. In R. Estes and G. Pregill (eds.), Phylogenetic Relationships of the Lizard Families: Essays Commemorating Charles L. Camp. Stan- ford, California: Stanford Univ. Press. Evans, S. E. 1980. The skull of a new eosuchian reptile from the Lower Jurassic of South Wales. Zoological Journal of the Linnean Society 70: 203-264. — Evans, S. E., AND L. J. BARBADILLO. 1998. An unusual lizard (Reptilia: Squamata) from the Early Creta- ceous of Las Hoyas, Spain. Zoological Journal of the Linnean Society 124: 235-265. Evans, S. E., Y. WaANc, AND C. Li. 2005. The Early Cretaceous lizard genus Yabeinosaurus from Chi- na: resolving an enigma. Journal of Systematic Palacontology 3: 319-335. Fraser, N. C. 1982. A new rhynchocephalian from the British Upper Trias. Palaeontology 25: 709-725 + plates 769-771. FuLLER, S., P. BaAversTOCcK, AND D. Kinc. 1998. Biogeographic origins of goannas (Varanidae): a molecular perspectiv e. Molecular Phylogenetics and Evolution 9: 294-307. Gao, K., AND R. C. Fox. 1996. Taxonomy and evolution of Late Cretaceous lizards (Reptilia: Squamata) from western Canada. Bulletin of Carnegie Muse- um of Natural History 33: 1-107. K., aNnD M. A. Nore. 1998. Taxonomic revision of Carusia (Reptilia: Squamata) from the Late Cretaceous of the Gobi Desert and phylogenetic relationships of anguimorphan lizards. American Museum Novitates 3230: 1-51. Gao, K., ann M. A. Norewtu. 2000. Taxonomic composition and systematics of Late Cretaceous lizard assemblages from Ukhaa Tolgod and adja- cent localities, Mongolis in Gobi Desert. Bulletin of Paleocene lizards. GAO, 180 the American Museum of Natural History 249: Lis: GauTuierR, J. A. 1982. Fossil xenosaurid and anguid lizards from the Early Eocene Wasatch Formation, southeast Wyoming, and a revision of the Anguioi- dea. Contributions to Geology, University of Wyoming 21: 7-54. GauTHIER, J., A. G. KLUGE, AND T. Rowe. 1988. Amniote phylogeny and the importance of fossils. Cladistics 4: 105-209. Giumore, C. W. 1928. Fossil lizards of North America. Memoirs of the National Academy of Sciences 22: 1-169 + 27 plates. Gitmore, C. W. 1942. Paleocene faunas of the Polecat Bench Formation, Park County, Wyoming, part II. lizards. Proceedings of the American Philosophic al Society 85: 159- 167. Gray, J. E. 1856. Notice of a new species of nocturnal lizard from Mexico. Annals and Magazine of Natural History 2: 270. Haas, G. 1960. On the trigeminus muscles of the lizards Xenosaurus grandis and Shinisaurus crocodilurus. American Museum Novitates 2017: 1—54. HALLeRMAN, J. 1994. Zur Morphologie der Ethmoidal- region der Iguania (Squamata)—Eine verglei- chend-anatomische Untersuchung. Bonner Zoolo- gische Monographien 35: 1-133. Harvey, M. B. 1991. The microstructure and evolution of scale surfaces in xenosaurid lizards [unpublished Master's thesis]. Arlington: The Univ. of Texas. Harvey, M. B. 1993. Microstructure, ontogeny, and evolution of scale surfaces in xenosaurid lizards. Journal of Morphology 216: 161-177. HERREL, A., E. DE Grauw, AND J. A. Lemos-EsPINAL. 2001. Head shape and bite performance in xenosaurid lizards. Journal of Experimental Zoolo- ey 290: 101-107. Horrstetrer, R. 1957. Un saurien helodermatidé (Eurheloderma gallicum nov. gen. et ap.) dans la faune fossile des phosphorites du Quercy. Bulletin de la Société Géologique de la France 7: 775-786. HUELSENBECK, J. P., AND R. NIELSEN. 1999. Effect of nonindependent substitution on phylogenetic ac- curacy. ee Biology 48: 317-328. KEARNEY, J. A. Matsano, AND T. Rowe. 2005. pa qnatouy of the extinct amphisbaenian Rhineura hatcherii (Squamata, Amphisbaenia) based on high-resolution x-ray computed tomog- raphy. Journal of Morphology 264: 1-33. Kinc, W., F. G. THompson. 1968. A review of the Aenecan lizards of the genus Xenosaurus Peters. Bulletin of the Florida State Museum 12: 93-123. Kiemsara, J. 2008. A new anguimorph lizard from the Lower Miocene of north-west Bohemia, Czech Republic. Palaeontology 51: §1-94. Larson, A. 1994. The comparison of morphological and molecular data in phylogenetic systematics, pp. 3/1390. In’ B. Schierwater B- Streit. GP: Wagner, and R. DeSalle (eds.), Molecular Ecology ene Evolution: Gee and Applications. Basel: Birkhauser Verlag. Bulletin of the Museum of Comparative Zoology, Vol. 160, No. 3 Ler, M. S. Y. 1998. Convergent evolution and character correlation in burrowing reptiles: towards a reso- lution of squamate relationships. Biological Journal of the Linnean Society 65: 369-453. M. S. Y., anp J. D. Scanton. 2002. Snake phylogeny baccd on osteology, soft anatomy and ecology. Biological Reviews 77: 305-40: LEMos- ESPINAL, |. A., “G. R. SMITE, AND gg. SEE BALLINGER. 2004. Aspects of the ecology of a distinct population of Xenosaurus platyceps from LEE, Queretaro, Mexico. Amphibia-Reptilia 25: 204- 210. Macey, J. BR... Ja-A. Scout TEA. Larsonseems: TunIvEV, N. OrLov, aNnD T. J. Papenruss. 1999. Molecular phylogenetics, tRNA evolution, and historical biogeography in anguid lizards and related taxonomic families. Molecular Phyloge- netics and Evolution 12: 250-272. Mappison, W. P., anpD D. R. Mappison. 2008. Mesquite: a modular system for evolutionary analysis. Version 2.5 [Internet] Available from: http:// mesquiteproject.org. MAcpEFRAU, H. 1997. Biologie, Haltung und Zucht der Krokodilschwanz-H6ckerechse (Shimiscumae Croco- dilurus). Zeitschrift des Kilner Zoo 40: 55-60. Maisano, J. A.,,C. J. Benr., J. Ay GAvroIERS AND eae Rowe. 2002. The osteoderms and palpebral in Lanthanotus borneensis (Squamata: Anguimor- pha). Journal of Herpetology 36: 675-682. Martins, E. P., AND T. GARLAND, Jr.. 1991. Phylogenetic analyses of the correlated evolution of Conemuons characters: a simulation study. Evolution 45: 534— Sie McCracken, K. G., J. HarsHMAN, D. A. MCCLELLAN, AND D. A. Arron. 1999. Data set incongruence and correlated character evolution: an example of functional convergence in the hind-limbs of Stifftail Diving Ducks. Systematic Biology 48: 683-714. McDoweE Lt, S. B., JR., AND C. M. Bocert. 1954. The systematic position of Lanthanotus and the affin- ities of the anguimorphan lizards. Bulletin of the American Museum of Natural History 105: 1-142 + 16 plates. Meszoety, C. A. M. 1970. North American fossil anguid lizards. Bulletin of the Museum of Comparative Zoology 139: 87-149. NoreLt, M. A., AND K. Gao. 1997. Braincase and phylogenetic relationships of Estesia mongoliensis from the Late Cretaceous of the Gobi Deon and the recognition of a new clade of lizards. American Museum Novitates 3211: 1- Nore.i, M. A., M. C. McKENNA, AND M. J. NOVACEK. 1992. Estesia mongoliensis, a new fossil varanoid from the Late Cretaceous Barun Goyot Formation of Mongolia. American Museum Novitates 3045: An ed Nypam, R. L. 2000. A new taxon of helodermatid-like lizard from the Albian-Cenomanian of Utah. Journal of Vertebrate Paleontology 20: 285-294. PrankA, E. R., AND D. R. Kino. 2004. Varanoid Lizards of the World. Bloomington and Indianapolis: Indiana University Press. RiepreL, O. 1980. The phylogeny of anguinomorph lizards. Denkschriften der Schweizerischen Natur- forschenden Gesellschaft 94: 1—-S6. RIEPPEL, O., AND H. Zanuer. 2002. The skull of the Uropeltinae (Reptilia, Serpentes), with special reference to the otico-occipital region. Bulletin of the Natural History Museum of London (Zoology) 68: 123-130. Rowe, T. 1986. Osteological diagnosis of Mammalia, L.1758, and its relationship to extinct Synapsida [unpublished Ph.D. dissertation]. Berkely: Univ. of California. Rowe, T., T. H. Ricu, P. VickErs-Ricu, M. SPRINGER, AND M. O. Woopsurne. 2008. The oldest platypus and its bearing on divergence timing of the platypus and echidna clades. Proceedings of the National Academy of Sciences 105: 1235-1242. SCHAFFER, H. B., J. M. Ciark, anp F. Pas 1991. When molecules and morphology clash: a phylo- genetic analysis of the North American ambysto- maud salamanders (Caudata: Ambystomatidae). Systematic Zoology 40: 254-303. SCOTLAND, R. W., R. G. OLMSTEAD, AND J. R. BENNETT. 2003. Phylogeny reconstruction: the role of mor- phology. Systematic Biology 52: 539-548. SMITH, H. M., AND J. B. Iverson. 1993. A new species of imopccale lizard (Reptilia: Xenosauridae) from Mexico. Bulletin of the Maryland Herpetological Society 29: 51-66. SmitH, K. T, 2006a. A diverse new assemblage of Late Eocene squamates (Reptilia) from the Chadron Formation of North Dakota, U.S.A. Palaeontologia Electronica 9: 1-44. Smitu, K. T. 2006b. Horizontal and vertical aspects of species diversity in the fossil record: alpha, beta, and the temporal nature of the richness-tempera- ture relation [unpublished Ph.D. dissertation]. New Haven, Connecticut: Yale Univ. SPRACKLAND, R. G. 1993. The biology of the Chinese crocodile lizard in captivity. The Reptilian Maga- zine 4(2): 9-15. STEBBINS, R. C. 1948. Nasal structure in lizards. American Journal of Anatomy 83: 183-221. XENOSAUR PHYLOGENY ¢ Bhullar 1S] STRONG, E. E., AND D. Lipscoms. 1999. Character rar ee inapplicable data. Cladistics 15: 363-37 1. Swisner, C. C., IL, anp D. R. Prornero. 1990. Single as CAP Ar dating of the Eocene—Oligocene transition in North America. Science 249: 760-762. SworrorD, D. L. 2001. PAUP*: Phylogenetic Analysis Using Parsimony (*and Other Methods), Ob10. Sunderland, Massachusetts: Sinauer Associates. Taytor, E. H. 1949. A preliminary account of the he rpe tology of the state of San Luis Potosi, Mexico. The University of Kansas Science Bulletin 33(Part 1): 169-215. TEMPLETON, A. R. 1983. Phylogenetic inference from restriction endonuclease cleavage site maps with particular reference to the evolution of humans and the apes. Evolution 37: 221-244. TOWNSEND, T., A. Larson, E. Louis, AND J. MACEY. 2004. Molecular phylogenetics of Squamata: the position of snake 3S, amphisbae nians, and dibamids, and the root of the squamate tree. Systematic Biology Dot (oD—1OL. VipaL, N., AND S. B. Hepces. 2005. The phylogeny of squamate reptiles (lizards, snakes, and amphisbae- nians) inferred from nine nuclear protein-coding genes. Comptes Rendus Biologies 328: 1000— LOOS. Wever, E. G. 1978. The Reptile Ear: Function. Princeton, New Jersey: Press. Wuirtesipe, D. I. 1986. The head skeleton of the Rhaetian sphenodontid Diphydontosaurus avonis gen. et sp. nov. and the modernizing of a living fossil. Philosophical Transactions of the Royal Society of London B 312: 379-430. WIBLE, J. R. 2003. On the Si osteology of the short- tailed opossum Monodelphis brevic evans (Didel- phidae, Marsupialia). Annals of Carnegie Museum 72: 137-202. Wiens, J. J. 2001. Character analysis in morphological phylogenetics: problems and solutions. Systematic Biology 530: 689-699. WIENS, J. J., AND J. L. Suincurr. 2001. How lizards turn into snakes: a phylogenetic analysis of body-form evolution in anguid feards Evolution 55: 2303- 9318. Version 4 Its Structure and Princeton Univ. ay US ISSN 0027-4100 MCZ Publications Museum of Comparative Zoology Harvard University 26 Oxford Street Cambridge, MA 02138 mczpublications@mcz.harvard.edu © The President and Fellows of Harvard College 2011. Bulletin of the Museum of Comparative Zoology Volume 160, Number 4 14 May 2012 Cryptic species within the Dendrophidion vinitor complex in Middle America (Serpentes: Colubridae) JOHN E. CADLE HARVARD UNIVERSITY | CAMBRIDGE, MASSACHUSETTS, U.S.A. BULLETIN OF THE Museum of Comparative Zoology BOARD OF EDITORS Editor: Jonathan Losos Managing Editor: Adam Baldinger Editorial Assistant: Samantha Edelheit Associate Editors: Andrew Biewener, Scott Edwards, Brian Farrell, Farish Jenkins, George Lauder, Gonzalo Giribet, Hopi Hoekstra, Jim Hanken, Jim McCarthy, Naomi Pierce, and Robert Woollacott Publications Issued or Distributed by the Museum of Comparative Zoology Harvard University Bulletin 1863- Breviora 1952- Memoirs 1865-1938 Johnsonia, Department of Mollusks, 1941-1974 Occasional Papers on Mollusks, 1945- General queries, questions about author guidelines, or permissions for MCZ Publications should be directed to the editorial assistant: MCZ Publications Museum of Comparative Zoology Harvard University 26 Oxford Street Cambridge, MA 02138 mczpublications@mcz.harvard.edu EXCHANGES AND REPRINTS All of our publications are offered for free on our website: http://www.mcz.harvard.edu/Publications/index.html | To purchase individual reprints or to join our exchange program, please contact Susan DeSanctis at the Ernst Mayr Library: mayrlib@oeb.harvard.edu. This publication has been printed on acid-free permanent paper stock. © The President and Fellows of Harvard College 2012. CRYPTIC SPECIES WITHIN THE DENDROPHIDION VINITOR COMPLEX IN MIDDLE AMERICA (SERPENTES: COLUBRIDAE) JOHN E. CADLE' CONTENTS Abstract 183 Resumen 183 Introduction 184 Materials and Methods 184 Redescription of Dendrophidion vinitor Smith, 1941 188 Dendrophidion apharocybe New Species 198 Dendrophidion crybelum New Species 209 Hemipenial Morphology vA An Introduction to Dendrophidion Hemipenes 217 Dendrophidion vinitor 220 Dendrophidion apharocybe 999 Dendrophidion crybelum 225 Discussion 228 Species Groups and Relationships 228 Biogeography 229 Acknowledgments 230 Appendix 1. Specimens Examined and Literature Records for Dendrophidion vinitor Smith 231 Appendix 2. Gazetteer 932, Literature Cited 935 Asstract. Snakes previously referred to Dendrophid- ion vinitor from southern Mexico to eastern Panama comprise three sibling species primarily distinguishable by substantial dicterences in hemipenial morphology and some subtle aspects of color pattern. Dendrophid- ion vinitor Smith is here restricted to populations from southern Mexico to Belize. Dendrophidion apharocybe new species is distributed from Honduras to Panama, primarily on the Atlantic versant. Dendrophidion crybelum new species is known from middle elevations of the Rio Coto Brus valley in southwestern Costa Rica (Pacific versant). Hemipenes of D. vinitor and D. apharocybe are similar in overall shape (short, bulbous) but the former has a highly ornate apex with ‘Research Associate, Department of Herpetology, California Academy of Sciences, 55 Music Concourse Drive, Golden Gate Park, San Francisco, California 94118 ( jcadle@calacademy.org). Bull. Mus. Comp. Zool., 160(4): 183-240, May, membranous ridges and an unusual apical boss, whereas D. apharocybe has a largely nude apex strongly inclined toward the sulcate de. Dendrophid- ion crybelum has an elongate cylindrical hemipenis with a large number of spines. In general, these species are not distinguishable by standard scutellation char- acters. Hemipenial and other characters suggest that these species are a monophyletic group within Dendrophidion and have the following relationships: (vinitor (apharocybe, crybelum)). Some aspects of the systematics and biogeography of De ndrophidion are discussed. Divergence among the three species is associated with two geological features important to speciation in Middle Anieeee: the northern Motagua— Polochic fault zone (Guatemala—Belize) and the southern Cordillera Talamanca (Costa Rica—Panama). Key words: Snakes, New Species, Central America, Mexico, Systematics, Costa Rica, Hemipenis, Morphology RESUMEN. Las serpientes anteriormente referidas a Dendrophidion vinitor desde el sur de México hasta Panama oriental se componen tres especies hermanas que se distinguen por differencias importantes en la morfologia de los hemipenes y aspectos sutiles de coloracion. Dendrophidion vinitor Sinith se limita a las poblaciones del sur de México, Guatemala, y Belice. Dendrophidion apharocybe, nueva especie, se encuen- tra desde Honduras hasta Panama primariamente en la vertiente Atlantica. Dendrophidion crybelum, nueva especie, se concoce solamente de elevaciones medias del valle del Rio Coto Brus en el suroeste-de Costa Rica (vertiente Pacifica). Los hemipenes de D. vinitor y D. apharocybe son similares en forma (corte, bulboso) pero el primero tiene un apice muy ornamentado con crestas membranosas y una protuber: ancia_ apical, mientras D. apharoc ybe tiene un apice mayormente nudo y fuertamente inclinado al lado suleado. Den- drophidion crybelum tiene un hemipene alargado y cilmdrico con muchas espinas. Generalmente, estas especies no se distinguen por characteres estandardi- zados de escutelaci6n. Los caracteristicos de los hemipenes y otros caracteres sugeren que _ estas especies se componen un grupo monofilético dentro Dendrophidion con las relaciones siguientes: (vinitor 2012 183 184 (apharocybe, crybelum)). Se discuten algunos aspectos de la sistematica y biogeografia de Dendrophidion. La divergencia entre las tres especies es asociada con dos rasgos geologicos importantes para la especiaci6n en América Central, la falla Motagua—Polochic (Guate- mala—Belice) en el norte y la Cordillera Talamanca (Costa Rica~Panama) en el sur. INTRODUCTION The Neotropical snake genus Dendrophid- ion Schlegel currently comprises eight or nine recognized species (Lieb, 1988; McCranie, 2011). Despite considerable progress in clarifying the taxonomy and species limits over the last three-quarters of a century (Smith, 1941; Peters and Orejas-Miranda, 1970; Lieb, 1988; Savage, 2002), it is clear that additional work is needed. Smith (1941) described D. vinitor (type locality: Piedras Negras, Guatemala) to accommodate a Mexican and Middle Amer- ican species with a ae anal plate, a relatively low number of subcaudals, and strongly keeled scales. His primary objective was to distinguish this species from the catch-all earlier name D. dendrophis (Du- méril, Bibron, and Duméril) (type locality: Cayenne, French Guiana), which had been widely applied to Middle American forms of Dendrophidion with an undivided anal plate. In doing so, Smith helped clarify the presence and characters of four species of Dendrophidion in Central America: D. clarkii Dunn, D. paucicarinatum (Cope), D. percarinatum (Cope), and D. vinitor. Lieb (1988) synonymized D. clarkii with D. nuchale (W. Peters; type locality: Caracas, Venezuela). However, McCranie (2011) resurrected the name clarkii (type locality: El Valle de Anton, Panama) for application to Middle American and western Colom- bian/Ecuadorian populations of this group based on patterns of geographic variation described by Lieb (1988). Nonetheless, further revision of D. nuchale/clarkii is necessary (see comments on name usage in Materials and Methods). Apart from con- tinuing ambiguity concerning D. nuchale/ clarkii, the taxonomy of Central American species has been stable since Smith’s (1941) revisions. Bulletin of the Museum of Comparative Zoology, Vol. 160, No. 4 Smith (1941) had referred specimens from southern Mexico to central Panama to D. vinitor and thought that the distribu- tion was continuous throughout this area. However, Lieb (1988, 1991) thought the distribution was highly fragmented, with disjunct segments in southern Veracruz, Mexico; Oaxaca, Mexico, to northern Gua- temala; Nicaragua to western Panama on the Atlantic versant: southwestern Costa Rica; and Darién, Panama, to northwestern Colombia. Researchers since have consid- ered D. vinitor a widespread, if discontin- uously distributed, species of the Central American herpetofauna. Lieb (1988) con- sidered D. vinitor a member of his “Den- drophidion dendrophis species group,” which is defined on the basis of hemipenial characters, strong keeling on the dorsal scales, and the point of reduction of the dorsocaudal scales. During my examination of Costa Rican material referred to D. vinitor, it was apparent that specimens from the Atlantic and Pacific versants differed substantially in hemipenial morphology. Furthermore, both of the hemipenial morphs of the Costa Rican specimens were quite distinct from hemi- penes of specimens from the northern part of the range of D. vinitor (Mexico). Further study indicated that “Dendrophidion vinitor” comprised three very similar species distin- guishable primarily by strong differences in hemipenial morphology and by a few subtle external characters. The three species occu- py geographically discrete distributions, with no contiguity or overlap currently known. The purpose of this paper is to redefine and diagnose D. vinitor and to describe the two new species within this complex. Available data on the natural history of these species are summarized, and detailed descriptions of their hemipenes are presented. MATERIALS AND METHODS General methodology and methods of recording meristic data and measurements follow procedures previously described (e.g., Cadle, 2005, 2007), but I here amplify SPECIES IN THE some of the characters particularly useful for the three species covered in this paper. Dorsocaudal reductions were meotden as the subcaudal at which the reduction from eight to six dorsal scale rows on the tail Aetna (Lieb, 1988). The point of poste- rior reduction of the dorsal scales was scored on each side of selected specimens as the ventral scute number at which the reduction occurred and the dorsal rows involved. For purposes of analyzing intra- specific differences in mean snout—vent length (SVL) of adult males and females, specimens with SVL > 450 mm were considered adults (Goldberg, 2003; Staf- ford, 2003). Similarly, because relative tail length increases proportionally with SVL, ‘hee range of adult relative tail length oe ea was Pececed for individuals with Sv; = 300 mm because analyses showed that RTL approaches an asymptote at approximately this size. When measurements or meristic data for particular specimens are referred to in the text, these data are based on my examinations (data encountered in the literature sometimes differ). Intraspecific mean differences between male a female sizes and scale counts were tested for significance using ¢ tests after testing for homogeneity of variances. P- values reported for intraspecific compari- sons are two-tailed pairwise comparisons; in the few cases in which there was a priori expectation for one sex or the other to have a greater value for a character (e.g., males having a longer tail or more subcaudals than females), p- Bales for one-tailed tests did not differ from the two-tailed comparison. Similar procedures and two-tailed tests were used for interspecific comparisons, which were analyzed separately for each sex except in cases in which intraspecific sexual differences were nonsignificant (sexes pooled in these cases). Means, standard deviations, and results of intraspecific sta- tistical comparisons for most meristic counts are presented in Table 1, and only summa- ries are given in the text. For determining tail breakage frequen- cies, I counted as “broken” only tails with a DENDROPHIDION VINITOR COMPLEX ¢ Cadle 185 clearly healed cap on the stump; thus, I record the fre ‘quency of “broken/healed” tails. In my survey of literature, it was apparent that some authors included any specimens with a tail fracture in their tail breakage frequency calculations (e.g., at least one paper recorded multiple fracture points in a high percentage of specimens). However, this method artificially inflates estimates of tail breakage frequencies be- cause of the inclusion of snakes whose tails were broken during or after capture, or even subsequent to storage in a museum jar. Although these specimens may offer clues as to the fragility of the tail in a particular species, they are not especially useful for comparative purposes. I scored the number of keeled dorsal scale rows on the neck, at midbody, and just anterior to the vent. Keels in all species of Dendrophidion are best developed (i.e., encompassing more dorsal rows) on the posterior body, but the number of keeled rows on the neck or at midbody often show interspecific differences that provide Gis= criminating characters. The three species covered in this paper are similar in their patterns of keeling. The basic pattern of temporal scales in D. vinitor and the new species here described is 2+2 (two primary, two secondary). However, temporal scales were often divided by a vertical suture (usually dividing the scale asymmetrically), or, less doreinouly. two temporal scales were fised or a temporal was fused with a supralabial. I recorded these divisions or fusions separately from the basic pattern. For example, a specimen might be recorded as having 2+2 temporals but with the upper primary and upper secondary fused on one side. Because of frequent asymmetry, tem- poral scales and supra- and infralabials were scored on each side of a specimen, and each side was treated as an independent obser- vation; the total count of a for these scale characters (Table 1) is thus about twice the number of specimens examined (damage sometimes prevented scoring on one or both sides of a given specimen). 186 Bulletin of the Museum of Comparative Zoology, Vol. 160, No. 4 TaBLe 1. SCALE COUNTS, MEASUREMENTS, AND OTHER DATA FOR THE THREE SPECIES OF THE DENDROPHIDION VINITOR COMPLEX. BODY PROPORTIONS, VENTRAL AND SUBCAUDAL COUNTS, AND NUMBER OF PALE BANDS ARE GIVEN AS RANGE FOLLOWED BY MEAN + SD. BILATERAL COUNTS ARE SEPARATED BY A SLASH (/). FOR PRIMARY AND SECONDARY TEMPORALS AND SUPRA- AND INFRALABIAL SCALES, EACH SIDE OF EACH SPECIMEN WAS COUNTED AS AN INDEPENDENT OBSERVATION. SVL, SNOUT—VENT LENGTH; MEASUREMENTS IN MILLIMETERS. SAMPLE SIZES IN PARENTHESES. ASTERISKS INDICATE STATISTICAL SIGNIFICANCE OF INTRASPECIFIC DIFFERENCES BETWEEN MEANS OF MALE AND FEMALE SIZE, PROPORTIONS, OR MERISTIC COUNTS (* P < 0.05; ** p < 0.01; *** p < 0.001); NO ASTERISK INDICATES NONSIGNIFICANCE. Dendrophidion Dendrophidion Dendrophidion vinitor Smith, 1941 —apharocybe New Species crybelum New Species Largest specimens: total length, SVL 997, 636 & 1040, 653 985, 625 Oo 847+, 595 9 1045, 672 9 893+, 631.9 Tail length/total length 0.35-0:337- 6 0:35=-0:38'C 0.34—0.36 oO 0.36 + 0.006 (8) 0.37 + 0.008 (15) 0:35 20.0007 KOK KKK * 0.33-0.36 9 0.33-0.36 9 0).34—0.35 9 0.34 + 0.012 (8) 0.35 + 0.010 (21) 0.34 + 0.003 (3) Tail length/SVL 0.53-0.59 o 0.53-0.61 & 051-0580; 0.56 + 0.019 (8) 0.5170 == 00210 GlS)) 0:55, 0102147) *K KKK * 0).49-0.56 O 0.49-0.57 9 0.50-0.54 9 0:53 22 0:025 (8) 054° 0.025; (21) 0:52: 0023) Maxillary teeth 38-45 33-44 38-44 Dorsal scales 40.8 + 1.96 (13) 17-17-15 (29) 39.1 + 2.39 (29) 1711560) 40:8 = LOLS) 17-17-15 (16) Ventrals 147-156 So 149-160 o 150215316 152.0 + 3.09 (12) 153.9 + 2.84 (31) 151.6 + 0.92 (8) KKK *KK KKK 153-162 9 152-168 9 156-162 9 V5 Oa 2293) (17) 160.8 = 3.79 (34) 160:0)2227(8) Subcaudals i= 125-0: ee ear > iUWAIR xe: ES G82=53253-( 11) 12) tease 923) 116.9 + 2.29 (8) 109-122 9 111-129 9 115-119 9 16:35 3258013) 119.9 + 4.47 (29) 117.2 = SSB) Dorsocaudal reduction, 8 to 6 40-65 Oo 32-63 O 43-58 oO (subcaudal number) 52 Gis 7-49 (02) AT8. =7e35 (29) 49.0 + 4.69 (8) ** *KKK 34-59 Q 26-52 0 42-56 9 AD Sree Oley ) Aue et iA Saolls) 48.6 + 5.53 (8) Dorsal scales, posterior reduction 85-96 Oo 85-102 o 93-99 Oo (ventral number) 92) 80275018) 94.6 + 3.89 (36) 96.0 = 1-65 (12) *KKK KKK KK 93-106 9 93-105 9 99-104 9 99.0 + 3.64 (24) 99.3 + 3.83 (36) 102.1 = 1:96:(9) Preoculars 1/1 (28) 1/1 (65) 1/1 (16) Postoculars 2/2, (28) 2/2, (63) 226) DSi) Primary temporals 2 (58) 1 (5) 2 (32) Daas) Secondary temporals 2 (58) 1S) DS) DAL) Supralabials, supralabials 83-512) Sea (2) 8, 3-5 (1) touching eye 9, 4-6 (54) 8, 4-6 (2) 9, 4-6 (31) LO =e) 9, 4-6 (119) 9) 5=6 (1) 9 57-01) SPECIES IN THE DENDROPHIDION VINITOR COMPLEX ¢ Cadle 187 TABLE 1. Dendrophidion vinitor Smith, 194] CONTINUED. Dendrophidion Dendrophidion Infralabials 9 (52) 10 (3) Eh) 51-70 60.3 + 4.69 No. of pale bands on body The width of the pale neck bands has been important in distinguishing “Dendro- phidion vinitor” auctorum from other spe- cies. However, neck bands, particularly the first two behind the head (which are often broader than other bands), can be very irregular in outline. Bands on the posterior two-thirds of the body are more uniform and narrower than the anterior bands (and less easily discriminate species). Thus, in scoring the width of neck bands, I used the third and fourth band behind the head as references and counted the scales or fractions thereof encompassed by the pale portion of the band (not including the dark bordering stipple) in the dorsolateral region at the level of dorsal scale rows 6 and 7. I counted scales in a horizontal line, not along the diagonal, as is often counted while doing dorsal scale counts; this was to minimize the effect of irregularities in band shape (e.g., zigzag or not) on scoring width. Thus, my summaries of band widths are somewhat less than some of those encountered in the literature. For example, Costa. Rican. ~“D: vinitor” are sometimes said to have pale bands 2-3 scale rows wide (Savage, 2002: 655), whereas in my scorings, most speci- mens have bands 1—2.5 rows wide. Maxillary dentition is similar in the three species covered in this paper. Teeth grad- ually enlarge anterior to posterior, but typically, four posterior teeth are abruptly enlarged (and nongrooved). The enlarged teeth are not offset, and a diastema is Abeeat (e.g., see Fig. 8). However, there is some variation within all three species in the abruptness with which the enlarged poste- rior teeth transition to the smaller anterior apharocybe New Species crybelum New Species paral) 9 (32) 8 (7) 9 (116) LO (5) 46-69 36-62 28) 54.4 + 4.48 (54) ASO bo CLS) series. I assessed some specimens as having either three or five posterior enlarged teeth (recognizing some subjectivity as to what constitutes an “enlarged” tooth). My im- pression is that posterior teeth in D. vinitor, as redefined herein, are not enlarged to the same degree or as abruptly as in the new species, D. apharocybe and D. crybelum. However, this is a subjective impression only—something that is difficult to quantify with wet preparations given the apparent variation. I have not attempted to assess this more fully, although I comment on some of the noticeable tion in the species accounts. Tooth counts are the total number of maxillary teeth, including empty tooth sockets and the enlarged posterior teeth. Everted hemipenes described in detail and illustrated herein were fully everted in the field at the time of collection. For detailed study they were removed from the specimen and inflated with colored jelly (Myers and Cadle, 2003); manual eversion was used for a few specimens for compar- ative purposes. Retracted hemipenes were slit midventrally and pinned flat for study. In addition to the hemipenes described in detail, I made reference to others that were everted to varying degrees and studied in situ. Hemipenial measurements were taken with dial vernier calipers to the nearest 0.1 mm. In the species accounts, I give brief characterizations of the hemi- penes, emphasizing salient features only. For ease of comparison, detailed descrip- tions of everted and retracted organs of all three species are deferred to a separate section at the end. However, because these species are most notably distinguished by 188 details of hemipenial morphology, the detailed descriptions should be considered integral parts of the species accounts. Hemipenial terminology is explained in the detailed accounts. “Dendrophidion vinitor” is discussed in many regional faunistic works for Central America, some of which are cited in the synonymies in the species accounts. How- ever, with few exceptions, the variational data (e.g., scale count ranges or color notes) in these works seem to derive from other sources (e.g., Lieb, 1988). Thus, it is usually not clear whether a particular account is based on specimens from the focal region. For example, Taylor (1954: 729-730) gave data on size, meristics, and color in life “taken from field notes” for “D. vinitor’ in his account of Costa Rican snakes. Howev- er, the color notes he quoted are those for the holotype, which is from Guatemala (Smith, 1941) and the Costa Rican form is a different species described herein. The size and meristic data of Taylor (1954) are a summary from throughout the range given in Smith (1941: Mexico to Panama), with no indication that that is the case. Unless it seems clear that descriptive comments apply to a specific geographic area, I do not cite them here Ipeneniige three species have been confused in previous literature. Similar caveats apply to some natural history data. Despite progress in understanding the systematics of Dendrophidion reviewed in the introduction, my own preliminary study makes it clear that further revisions are necessary. For purposes of this work, some taxonomic conventions, which will ultimate- ly be modified as revisionary work proceeds, are necessary. For convenience in the diagnoses, I use the two species groups of Dendrophidion recognized by Lieb (1988): the D. percarinatum group comprising D. bivittatum, D. brunneum, D. paucicarina- tum, and D. percarinatum; and the D. dendrophis group comprising D. dendro- phis, “D. nuchale/clarkii,” and D. vinitor (and by extension, the two new species described here). Lieb did not assign the Bulletin of the Museum of Comparative Zoology, Vol. 160, No. 4 Colombian endemic D. boshelli to a species group. Species in the D. percarinatum group differ from those in the D. dendro- phis group in having a dorsocaudal reduc- tion from § to 6 anterior to subcaudal 25 (posterior to subcaudal 25 in the D. dendrophis group, further refined for three species herein), by less strongly keeled dorsal scales, and often by smaller hemi- penial spines, among other characters (Lieb, 1988). In addition, I recognize as the “Dendrophidion vinitor complex” the two new species described herein along with D. vinitor as restricted here. McCranie (2011) resurrected the name D. clarkii for part of “D. nuchale” sensu Lieb (1988), but further revision will be necessary before the appli- cation of names in this complex becomes clear; see comments in Savage (2002: 655) and McCranie (2011: 106—107). Thus, in this paper, I refer to the complex of species represented by the names D. nuchale and D. clarkii as “D. nuchale auctorum.” Addi- tional comments on taxonomy and character distributions are given in the discussion. Specimens of D. vinitor as restricted herein are listed in Appendix 1, which also includes museum abbreviations used throughout this paper. Specimens of the new species are listed in their species accounts since this material comprises the type series. Notes on the localities and coordinates for specimens examined and for selected localities from literature are in Appendix 2. Redescription of Dendrophidion vinitor Smith, 1941 Figures 1, 2A, 3-4, 6, 18-19 Drymobius dendrophis. Gunther, 1885— 1902: 127 (part; Guatemala); Boulenger, 1894: 15-16 (part; specimens Db, d, ffrom “Vera Paz’ and “Guatemala’). See discussion of historical records under Distribution. Dendrophidion dendrophis. Dumeril et al., 1870-1909: 730-732 (part; specimens from “Peten” and “Vera Paz’). The SPECIES IN THE “Peten” specimen (MNHN 7353), illus- trated by Lieb (1988: fig. 1), is a syntype of Herpetodryas poitei Dumeril, Bibron, and Dumeéril, 1854. See discussion of historical records under Distribution. Dendrophidium dendrophis. Duges, 1892: 100-101 + pl. (Motzorongo [Veracruz], Mexico). See Smith (1943: 416) and discussion of historical records under Distribution herein. Dendrophidion vinitor. Smith, 1941: 74-75 (type locality: Piedras Negras, Guate- mala) (part; holotype and paratypes from Guatemala and Mexico). Smith, 1943: 415-416 + fig. 13. Taylor, 1944: 184 (EHT-HMS 27496-98). Smith and Tay- lor, 1945: 46. Stuart, 1948: 63. Smith and Taylor, 1950: 318 (holotype). Stuart, 1950: 23. Darling and Smith; 1954: 191 (UIMNH 33862). Taylor, 1954: 729-730 (part; color description of holotype and some meristic data quoted from Smith, 1941). Cochran, 1961: 172 (part; holo- type and paratypes, USNM 110662, 7099, 46589). Duellman, 1963: 246. Stuart, 1963: 94 (part; Mexico and Guatemala). Peters and Orejas-Miranda, 1970: 79 (part). Alvarez del Toro, 1972: 142, 144. Johnson et al. “1976” [1977]: 134, 136-137. Perez-Higareda, 1978: 69, 72. Alvarez del Toro, 1982: 190- 191. Perez-Higareda et al., 1987: 16. Smith, 1987: xxxvii (part; based on Gunther's “Drymobius dendrophis’’). Flores and Gerez, 1988: 218-219, 261. Lieb, 1988: 171 (part). Villa et al., 1988: 63 (part). Campbell and Vannini, 1989: 11. Johnson, 1989: 64. Lieb, 1991: 522.1-522.2 (part). Péerez-Higareda and Smith, 1991: 31, pl. 4. Flores-Villela, 1993: 30. Lee, 1996: 310-311 (Yucatan, base). Campbell, 1998: 207, fig. 127 (part; specimen from Veracruz, Mexico). Lee, 2000: 282-283, fig. 315 (part; Belize). Stafford and Meyer, 2000: 199- 200 (Belize). Savage, 2002: 655-656 (part). Kohler, 2003: 200 (part). Meerman DENDROPHIDION VINITOR COMPLEX ¢ Cadle 189 and Lee, 2003: 67, 70. Stafford, 2003: 111 (part; specimens from Mexico, Gua- temala). Guyer and Donnelly, 2005: 185 (part). McCranie et al., 2006: 147-148 (part). Kohler, 2008: 215 (part). Savage and Bolanos, 2009: 14 (part). McCranie, 2011: 111 (part). Holotype (Fig. 1). USNM 110662, from Piedras Negras [El Petén], Guatemala. Collected 21 Mas 1939 by Hobart M. eae and Rozella B. Smith (field number 7280) a part of collections assembled during an of the Walter Rathbone Bacon Traveling Scholarship (Smith, 1943: 416). The holo- type is presently in fair condition. I did not examine it directly but inspected dorsal and ventral photographs provided by USNM and had selected characters verified by USNM personnel. According to ones (1941), it is a subadult female 510+ mm total length, 169+ mm anes ie length (341 mm SVL). The holotype has several irregular longitudinal ventral inci- sions from the anterior body nearly to the vent and on the anterior ventral part of the tail. The tail tip is missing. Smith (1941) described the holotype in detail, but one character in his description is apparently in error: “nine supralabials, 3rd, 4th and 5th entering orbit’—a character found in no other specimen (Table 1). Both sides of the holotype i in fact have nine supralabials but it is the 4th, 5th, and 6th that touch the eye on both sides (verified by Steve Gotte, March 2011), which is the near-universal condition in the specimens I examined (Table 1). Etymology. The specific name vinitor is a Latin noun meaning ‘vine cultivator” or “vine dresser.” Latin: Texionus show that the word is derived from the noun vinum (wine). The complete derivation therefore is the stem vin- + connective -i- + the suffix - tor. The termination is a classic noun suffix meaning an agent or doer of something (vine care and pruning in this case). The name was not a good choice for a forest- dwelling snake. | Hobart Smith gave no etymological infor- mation in the original description, but 190 Bulletin of the Museum of Comparative 7 Zoology, Vol. 160, No. 4 Figure 1. = 5cm. provided different intended meanings to other taxonomists: Lee (1996: 311) quoted Smith as stating that the name means “a dweller in vines’; McCranie (2011: 113) quoted Smith as stating that he could not know for sure what he had in mind but that he had intended “vine-climber.” In- any case, Lee (1996: 311) was incorrect in stating that vinitor is derived from “vinea,” which is both an adjective pertaining to wine and a substantive meaning vineyard or vines. Diagnosis. Dendrophidion vinitor is char- acterized by (1) Dorsocaudal reduction from § to 6 occurring posterior to subcaudal 30 (range, 34-65); (2) single anal plate (rarely divided); (3) relatively low subcaudal counts (<130 in males and females); (4) pale dorsal crossbands usually less than one dorsal row wide and bordered posteriorly (often anteriorly as well) by dark brown or black; (5) immaculate ventrals and subcau- dals exce a for lateral dark pigment com- mon to all species of Dendrophidion; and Holotype of Dendrophidion vinitor Smith (OSNM 110662) from Piedras Negras, Guatemala (Petén Department). Scale (6) a bulbous hemipenis with a highly ornate apex, including: asulcate flounces; a series of free-standing membranous ridges, the me- dian one of cahich is taller than others and bisects the apex; and a raised rounded boss or protuberance on the sulcate edge of the apex (the sulcus spermaticus ends beneath its free edge). The combination of few subcaudals and (in most individuals) a single anal plate will distinguish D. vinitor from all other species of Dendrophidion except D. apharocybe, D. crybelum, and D. paucicar- inatum. Dendrophidion vinitor differs from the four species of the D. percarinatum group in having the dorsocaudal reduction from 8 to 6 posterior to subcaudal 30 and more strongly keeled dorsal scales, and from all species of the D. percar inatum group except some individuals of D. paucicarinatum in having a single anal scale. Dendrophidion paucicarinatum usually has a more uniform- ly colored dorsum lacking distinct pale crossbands, has narrow dark lines across SPECIES IN THE DENDROPHIDION VINITOR COMPLEX ¢ Cadle 19] Figure 2. Comparison of pale bands on the anterior body of representative specimens of the Dendrophidion vinitor complex. Each pair (left, right) representative of one species. (A) Dendrophidion vinitor as restricted herein (UMMZ 121145, UMMZ 122767; both Veracruz, Mexico). (B) Dendrophidion apharocybe new species (KU 112974, Nicaragua; LACM 148593, Costa Rica [holotype]). (C) Dendrophidion crybelum new species (LACM 114108, LACM 148599 [holotype]; both Costa Rica). the venter in adults and many juveniles, and has a higher number of ventrals (>175 compared with <170 in D. vinitor). Den- drophidion vinitor differs from D. boshelli in having 17 midbody scale rows (15 in D. oaeie Dendrophidion vinitor has fewer subcaudals (<130) and usually a shorter adult relative tail length (<60% of SVL) than D. nuchale auctorum and D. dendro- phis (>130 and usually >60% of SVL respectively);-the anal plate may be anther single or divided in these last two species. Dendrophidion vinitor differs from the new species described herein, D. apharo- cybe and D. crybelum, in having narrower pale bands on the neelentodar body (Fig. 2). Although “Dendrophidion vinitor” is often stated to have broader bands than other congeners, such as D. percarinatum (e.g., Savage, 2002: 655), those statements are based on a comparison of the two new species described herein, D. apharocybe and/or crybelum, with other species. Dendrophidion vinitor as redefined here has narrow pale crossbands, generally one scale row or less in width (as pointed out by Smith [1941] for the holotype), compared with crossbands more than one row wide in D. apharocybe and D. crybelum (Fig. 2). In D. vinitor, the crossbands become indistinct posteriorly or restricted to the dorsolateral/ vertebral region, whereas in D. apharocybe and D. crybelum, the posterior crossbands usually become invested with dark pigment so that each crossband appears as a transverse dark band with embedded pale ocelli. These patterns of D. apharoc ybe and D. crybelum are easily seen in the color plates of Savage (2002: pl. 416), Solérzano (2004, fig. 60), Figure 9 herein for D. apharoc ybe, and Figures 15 and 16 herein for D. crybelum. De ndrophidion vinitor has a significantly greater mean number of pale hance on 7 body than either D. aphar- ocybe (p < 001) or D. crybelum (p < 0.001), aa the ranges of band num- bers overlap aa among the three species (Table | Additionally, ee ae vinitor dif- fers from D. apharocybe and D. crybelum in hemipenial morphology (see detailed de- scriptions). The everted hemipenis of D. vinitor is rather short and stout and has a highly ornate bulbous apex ornamented ith largely nonanastomosing (free-stand- ing) membranous ridges (reduced remnants of calyces primitively present). The apex has a prominent rounded protuberance, or apical boss, on the sulcate side; the cen- trolineal sulcus spermaticus ends under- neath its free edge. One of the free-standing ridges bisects the apex in a line from the fone edge of the apical boss directly across the apex to the asulcate side. The other ridges extend obliquely outward toward the asulcate side from this median ridge. 192 Bulletin of the Museum of Comparative Zoology, Vol. 160, No. 4 Dendrophidion apharocybe has a hemipenis similar in shape to that of D. vinitor, but the apex is strongly inclined toward the sulcate side, lacks an apical boss, and, in contrast to D. vinitor, is nude except for low rounded ridges. Dendrophidion crybelum has a relatively long cylindrical hemipenis with a small, nonbulbous apex bearing short ridg- es, and a large number of enlarged spines (=70) on the hemipenial body compared with <50 in D. vinitor. Description (12 males, 17 females). Ta- ble 1 summarizes size, body proportions, and meristic data for D. vinitor. Largest specimen (UIMNH 37165) a male 997 mm total length, 636 mm SVL. Largest female (UIMNH 35546) 847+ mm total length, 595 mm SVL. Tail 35-37% of total length (53-59% of SVL) in males; 33-36% of total length (49-56% of SVL) in females. Dorsal scales in 17-17-15 scale rows, the posterior reduction usually by fusion of rows 2+3 at the level of ventrals 85-106 (see sexual dimorphism below). Ventrals 147-156 (av- eraging 152) in males, 153-162 (averaging 158.1) in females; usually one preventral anterior to ventrals (about 33% of speci- mens have two preventrals; rarely, preven- trals were absent). Anal plate nearly always single (divided in two of 29 specimens). Subcaudals 112-125 (averaging 118.6) in males, 109-122 (averaging 116.8) in fe- males. Dorsocaudal reduction at subcaudals 40-65 in males (mean 52.6), 34-59 in females (mean 45.3). Preoculars 1, posto- culars 2, primary temporals 2, secondary temporals 2, supralabials usually 9 with 4-6 bordering the eye (occasionally 8 with 3-5 bordering the eye or 10 with 5-7 bordering the eye), infralabials usually 9 (low frequen- cy of 10 or 11). Maxillary teeth 38-45 (averaging 41), typically with four posterior teeth abruptly enlarged; sometimes the enlargement appears more gradual and with three or five somewhat enlarged. Smith (1943: 416) illustrated head scalation for EHT-HMS 27496 (now UIMNH 17632). Two apical pits present on dorsal scales. About 85% of specimens have keels on all dorsal rows of the neck except row 1; the others lack keels on row 2 additionally. Virtually all specimens have all dorsal rows except row 1 keeled at mid- and posterior body; an exception is a small juvenile female (294 mm SVL) in which all rows are keeled at mid- and posterior body (keels very weak on row | in both positions). Fusions or divisions of temporal scales were moderate- ly common, a the following frequencies: upper primary divided (10), upper second- ary divided (10), upper or lower primary + secondary fused (3), lower primary divided (2), lower secondary need with ultimate supralabial (1), upper + lower secondary fused (1, partial fusion). Hemipenis unilobed with a somewhat bulbous apex; spinose region followed distally by flounces, poorly developed caly- ces, and a highly ornate apex, including free-standing spinulate ridges and an apical boss on the sulcate side. Sulcus spermaticus simple, centrolineal, with a slightly flared tip in everted organs. Geographic and Other Variation. Tails are proportionally shorter in small individ- uals. Specimens less than 300 mm SVL have tail lengths 29-34% of total length, 40-51% of SVL (N =18, males and females combined). No strong geographic trends were evident among the characters exam- ined. Johnson et al. “1976” [1977] reported a male from near Ocozocuautla (Chiapas, Mexico) with 157 ventrals and 127 sub- caudals, which are slightly greater than counts for specimens I examined (Table 1). The posterior reduction of the dorsal scales always involves lateral rows, but in addition to the rows most commonly involved (2+3), reduction occasionally occurs by fusion of rows 3+4 and rarely by loss of row 3 or fusion of rows 1+2. Two specimens from Guatemala (UTACV 22155, 22755), both from the same general locality, have divided anal plates and are the only specimens examined having this con- dition (the only known specimen from nearby in Belize, KU 300784, has a single anal plate). Ordinarily, a divided anal plate in northern specimens of Dendrophidion suggests an identity with D. nuchale auc- SPECIES IN THE DENDROPHIDION VINITOR COMPLEX © Cadle torum but the color pattern and low subcaudal counts (115-116) of these spec- imens confirms their identity as D. vinitor. Whether this is a peculiarity of this popu- lation or whether divided anal plates occur with low frequency in others is unknown (but see comments on historical records in Distribution). No other characteristic of these specimens is Pay unusual, although in UTACV 22755, the last ventral plate is also divided. At least three other species of Dendrophidion (D. dendrophis, D. nuchale auctorum, D. paucicarinatum) are variable in the single/divided nature of the anal plate. Sexual Dimorphism. A few characters showed significant sexual dimorphism (Ta- ble 1). Although the largest specimen is a male, mean sizes of adult ie (542 mm SVL) and females (535 mm SVL) are not significantly different (with the caveat that sample sizes are small and sample variances high). Females have significantly greater ventral counts than males and the posterior reduction of the dorsal scales occurs farther posteriorly in females (mean ventral 99.0, N = 24) than in males (mean ventral 92.8, N = 18; p < 0.001). In males, the dorsocaudal reduction (8 to 6) occurs at a significantly more posterior position than in females. Males also have a proportionally longer tail than females, but males and females do not differ in subcaudal counts. Coloration in Life. Color photographs of D. vinitor from Veracruz, Mexico, were published by Pérez-Higareda and Smith (1991, pl. 4) and Campbell (1998, fig. 127) and from Belize by Lee (2000, fig. 315; KU 300784 fide Julian C. Lee, pe reonal communication). Alvarez del Toro (1972, fig. 133; 1982, fig. 143) illustrated a specimen from Chiapas, Mexico, in black and white, and Dugés (1892) provided a somewhat stylized colorized drawing of a specimen from Veracruz. Smith (1941) gave detailed color notes for the holotype in life (repeated virtually verbatim by Taylor [1954: 729-730]), here paraphrased: Head and temporal region down to upper edges of two postocular 193 labials and all of the last labial brownish gray, the sutures darker and with a slightly reddish tinge (lower edge of head cap dark brown, mixed with dull brownish ee red): Be parts of anterior supralabials with a reddish tinge; supralabials below the head cap pure \ white; 59 bands on body, 54 on tail; bands on neck covering one scale length, brownish gray laterally, yellow oe sally and bordere a anterior ly, poste riorly, « both by narrow irregular areas of black; size of yellow dorsal area in light bands decreas- ing and eventually disappearing posteriorly: light bands gr adually disappearing posteri- orly; tail bands and those on posterior part of body black; fe borders of light bands interspersed or themselves bordered by brick-red, ae re especially prominent medially; central ground color between bands brownish gray anteriorly, becoming light brown tinged with red on middle ad posterior body: tail with a stripe of dark brown (black) interspersed with brick-red, involving edges of subcaudals and lower half of first dorsocaudal row; gular region white; belly yellow; subcaudals ‘yellow, ‘paler pos- teriorly. Alvarez del Toro (1972: 144) described a specimen from Chiapas, Mexico: “general coloraticn reddish brown, a little ‘grayish toward the sides. On the neck are three yellow bands, and an additional 62 yellowish ones on the body; all the bands or bars bordered posteriorly by narrow black bands. The lips and throat are white, the venter yellow orange” (description repeated by Alvarez del Toro [1982: 191] except the neck bands are said to be orange). Coloration in Preservative. Dorsal ground color of preserved specimens yellowish brown, brown, or gray (smaller specimens usually paler than larger ones). Pale cross- bands yellowish brown to gray. In adults, posterior pale crossbands become indistinct, marked only by dark transverse stippling with occasional pale flecks dorsolaterally and indistinct or absent on the flanks (Fig. 3). The number of pale bands/ocelli on the body ranged from 51 to 70 with a mean and median of 60 and a mode of 59 194 Bulletin of the Museum of Comparative Zoology, Vol. 160, No. 4 Figure 3. Representative specimens of Dendrophidion vinitor from Mexico. (A) AMNH R-66845 (Oaxaca). (B) UMMZ 122767 (Veracruz). (C) UMMZ 121145 (Veracruz). (D) UCM 39912 (Oaxaca; juvenile, 218 mm SVL). (posterior indistinct bands included; no significant difference between males and females). In about 85% of the specimens neck bands were one scale row or less in width; in the remainder, the neck bands were |—-1.5 rows wide. Vertebral region usually retains whitish spots that sometimes form a more or less continuous vertebral streak punctuated by larger irregular pale spots (like beads on a string) at he position of crossbands (Figs. 3B, 3C, 4). The dark ventrolateral tail stripe mentioned by Smith (1941; see Fig. 1) varies from distinct to indistinct. On the posterior body most specimens have a series of small, widely spaced dark spots on scale row 2 or the suture line between rows 2-3 (occasionally forming a more or less broken line). This line is “Olen highlighted by paler scales in rows 3-4 on She: posterior body (e.g., SPECIES IN THE Figure 4. A typical posterior body pattern of Beer elon vinitor (AMNH R-66845). Section is from the posterior third of the body. The vertebral row is marked by a whitish line punctuated with larger blotches at the position of pale crossbands which, on the posterior body, are mainly restricted to a dorsolateral position and heavily invested with dark pigment. Fig. 3B). Venter yellowish to whitish and immaculate except for lateral dark pigment. Most juveniles have more distinct pale crossbands on the posterior body than do adults, and in general, their patterns are — more contrasting than those in adults (Fig. 3D). Distribution (Fig. 5). Central Veracruz state, Mexico, eastward to southeastern Guatemala and southern Belize. Dendro- phidion vinitor occurs on the Atlantic and Pacific versants of the Isthmus of Tehuan- tepec but otherwise appears to be restricted to the Atlantic versant. The northernmost record is from Las Minas, Veracruz (Pérez- Higareda and Smith, 1991), assuming I have identified the locality correctly. Recorded elevations for specimens I examined are DENDROPHIDION VINITOR COMPLEX ¢ Cadle 195 lowland (<100 m) to about 800 m on the slopes of Volcan San Martin in southern Veracruz (most localities 400-600 m), with the single Belize specimen from. slightly higher, "he tween 940 and 1,035 m ese Appendix 2, Little Quartz Ridge). Several historical records of D. vinitor deserve comment because they document localities from which no recent specimens are available. Dugés (1892; “Dendrophidium dendrophis”) gave a detailed description and color dleaton of a specimen from Motzor- ongo, Veracruz, which leaves little doubt about its identity as D. vinitor. Specimens from Guatemala are scarce, and I am aware of only two specimens obtained since the holotype was collected in 1939 (Appendix lk despite considerable biological inventory of that country (e.g i uellearai. 1963: Stuart, 1963: Campbell, 1998S). There has been controversy about the origin and identity of several specimens obtained in Guatemala during the 19th century (see synonymy ander the names Drymobius dendrophis and Dendrophidion dendrophis). Duméril et al. (1870-1909: 730-732) reported a specimen obtained by Arthur Morelet from “Peten” and two others (“seen alive,” collector not indicated) from “Vera 95° \ \) @ Dendrophidion vinitor Figure 5. Distribution of Dendrophidion vinitor in southern Mexico, Guatemala, and Belize. A few symbols represent multiple contiguous localities. Arrow indicates the type locality. 196 Paz”; their composite “Dendrophidion den- drophis” also included specimens from northern South America. Some details in their account (e.g., single anal plate and low subcaudal counts [119-127]), suggest D. vinitor, although Lieb (1988: 164-165) pointed out some confusion in their sub- caudal counts and tail lengths. The “Peten” specimen is one of three syntypes of H. poitet Duméril et al. (1854: 208), a name that Lieb (1988: 165) made an objective junior syno- nym of D. dendrophis (Schlegel) by desig- nating MNHN 41 from French Guiana as the lectotype of both names, thus preserving Smith’s name D. vinitor for the Central American species. Lieb (1988, fig. 1) illus- trated the © Peten' “symtypoe™ ot H. potter, whose banding pattern is consistent with D. vinitor (pale bands becoming. restricted dorsolaterally on the posterior body). Lieb (1991: 522.1) thought the Petén specimen of Duméril et al. (1870-1909) “almost certainly originated from ... Flores [Guatemala], where |Morelet’s] collectors were most active.” I think it’s far from certain that Morelet’s specimen came from Flores, although he spent a long sojourn there. Morelet traveled extensively in the Mexican states of Tabasco and northern Chiapas, ascending the Rio Usumacinta to near the present Guatemalan border (all within the known range of D. vinitor) before traveling overland to Flores (Morelet, 1871). The specimen could have come from anywhere along this route, and I discount the “Flores” locality based on this historical material. Owing to its accessibility and location along a major route, perhaps more biologists have passed through Flores than any other part of Petén. Duellman (1963: 246) listed D. vinitor as “hypothetical” in southern Petén but obtained no material during a 2-week survey of rainforests there, nor did Stuart (1934, 1935, 1958) working in the same region. Thus, apart from western Petén along the Rio Usumacinta (type locality), the presence of D. vinitor in Petén is unsubstantiated. Giinther (1885-1902) reported four spec- imens from “Vera Paz, between Coban and Lanquin” (in present-day Alta Verapaz Bulletin of the Museum of Comparative Zoology, Vol. 160, No. 4 province) and another from “Guatemala” collected by Osbert Salvin. Their identity has been questioned mainly on the basis of scale counts and a divided anal plate reported by Giinther. It should be noted that Boulenger (1894: 16, specimens b-f) gave ventral and subcaudal counts for the same specimens, in some cases substantially different from Giinther’s. Based on Boulen- ger’s subcaudal counts, it seems likely that only specimens b, d, and f (subcaudals 113, 116, 119; respectively) are D2 vemnton because it is the only Dendrophidion from that area with such low subcaudal counts (counts of the other two specimens are consistent with D. nuchale auctorum, which is known from the region). The only character seeming to disallow the identity of the three Salvin specimens as D. vinitor is the divided anal plate reported by Giinther (not recorded by Boulenger). However, the only two recently collected Guatemalan specimens of D. vinitor are from near Salvin’s Alta Verapaz locale and have divided anal plates (see above Description), so this character is no longer an obstacle. Thus, I accept this locality and include it in Figure 5. I am aware of only one published pee based on a re-examination of Salvin’s material (Stafford, 2003; see Ap- pendix 1), but Stafford cited only one of the specimens in his study. I confirmed several erroneous reports of “Dendrophidion vinitor” from Belize and the outer Yucatan Peninsula first pointed out by Lieb (1991): Wilson (1966; based on LSUMZ 8901-03: = D. nuchale auctorum): McCoy (1970: UGEM 25708) 25794. 2580-06 9584647, 25874. = D. nuchale auctorum): and Lee (1980: 34, 65: UCM 28122, = Mastigodryas melanolomus). However, a recently collected specimen documents D. vinitor in southern Belize (KU 300784; Fig. 6). KU 300784 is a female, and _ its identification as D. vinitor is based on the narrow pale crossbands on the neck and other features of coloration, 119 subcaudals, and dorsocaudal reduction at subcaudal 47. This specimen is the basis for the record reported as a personal communication by SPECIES IN THE DENDROPHIDION VINITOR COMPLEX ¢ Cadle 197 Figure 6. in Lee (2000: fig. 315). Stafford and Meyer (2000: 200) and for the color photograph in Lee (2000, fig. 315; Julian C. Lee, personal communication, January 2011); see Meerman and Lee (2003). Natural History. The few specimens of D. vinitor accompanied by habitat information suggest that it inhabits ae ely intact forests, var iously described < “rainforest” “lower montane ies (e.g., Duell- man, 1960: 34, for Donaji, Oaxaca: Tol amson, 1989: 64, for Chiapas). Goodwin (1969: 259) gave brief habitat notes on Oaxacan localities for specimens collected by Thomas MacDou- gall, as follows: La Gloria (coffee plantations, milpas, rainforest; elevation about 1,500 ft.): Cerro Azul (“cloud forest,” high north-facing slopes swept by gale-force wines: elevation of collections from 7,000 ft.); Cerro Atravesado (open peaks to about 8.000. ft. 4 000 to pine stands, grass and rocks, patches of “cloud forest” at north end, some ranches on lower slopes. Elevation about 4,750 ft.). Few natural history or behavioral data seem to have been recorded for D. Lasik Darling and Smith (1954: 191) found <« juvenile coiled by a trail during the day on Voledén San Martin, Veracruz, Merce: Two specimens (FMNH 126554-55) are accom- panied by notes indicating they were “in heavy shade, in daytime, on ground.” John- son et al. “1976” [197 al reported a male from Dendrophidion vinitor (KU 300784), the only known specimen from Belize. A color photograph of this specimen in life is near Ocozocuautla (Chiapas, Mexico) active- ly foraging on the forest floor at 3:30 p.m. on 28 December 1974. Alvarez del Toro (1972 1982) repor ted D. vinitor as “relatively common” in the hills surrounding Presa Malpaso, a large hydroelectric reservoir in the Rio Grijalva basin of northern Chiapas, Mexico: he stated that it inhabits the shady parts of the forest and rapidly hides in the leaf litter upon disturbance. At Los Tuxtlas Biological Station Perez-Higareda (1978) spequnicred D. vinitor much more fr equent- ly in the dry season and first half of the rainy season (Apr il-October), with a peak encoun- ter rate in October, than in the period of more consistent heavier rains (November— March). Stomach contents of two specimens I examined (576 and 248 mm SVL) each comprised a single small frog (Pristimantis or Craugastor; 20 and 15 mm SVL, respective- ly). in contrast to some species of Dendro- phidion. the frequency of broken/healed tails in D. vinitor is relatively low—13.7% of the specimens examined compared with 30% or more in some species (Stafford, 2003; Cadle, 2010); this may indicate a less easily pseud- autotomic tail than other species or differ- ential predation intensities. Stafford (2003) included specimens of D. vinitor as redefined herein in a study of morphology, diet, and reproduction of 198 Bulletin of the Museum of Comparative Zoology, Vol. 160, No. 4 Dendrophidion spp.. but most of his spec- imens actually are the two new species described in this paper. Among specimens I examined, a gravid female from Veracruz (491 mm SV L: KU 27564) was collected 30 March 1949. UCM 41162 from Oaxaca (567 mm SVL) had vitellogenic ova 3- 5 mm in length and was collected 5 May 1967. UTAGV 22755. from: Guatemala (427 mm SVL), a female (? adult) with thin, nonconvoluted oviducts, was collected 16 October 1987. The two smallest individuals were 199 mm SVL (Veracruz: collected 19 September 1965) and 218 mm SVL (Oaxa- ca; collected July to September 1968). Vernacular names for D. vinitor given in various regional works include zuwmbadora aquillada ae barras (Guatemala), culebra barrada (Guatemala and Mexico), kuyun kan (Lacand6n Maya), and sabanerita (Be- lize) (Alvarez del Toro, 1972: Lee, 1996: 310; Campbell, 1998: 207; Stafford and Meyer, 2000). Dendrophidion apharocybe New Species Figures 2B, 7-10, 13B, 13D, 14C, 14D, ZO 2n Dendrophidion dendrophis (part). Gaige et al., 1937: 12 (UMMZ 79765-66). Dunn and Bailey, 1939: 15 (two specimens from Pequeni—Esperanza ridge = MCZ R-42782-83). Dendrophidion vinitor. Smith, 1941: 74-75 (part; paratypes from Nicaragua, Costa Rica, Panama). Barbour and Loveridge, 1946: 99-100 (paratypes from Nicara- gua, Costa Rica, Panama). Peters, 1952: 43 (paratypes from Nicaragua). Taylor, 1954: 729-730 (part; Costa Rica). Smith, 1958: 223 (Panama). Cochran, 1961: 172 (part; paratypes from Nicaragua). Peters and Orejas-Miranda, 1970: 79 (part). Savage, 1973: 14 (part), 1980: 92 (part). Scott et al., 1983: 372 (part; “La Selva’). Kluge, 1984: 54 (Nicaragua: UMMZ 79766). Wilson and Meyer, 1985: 41 (key). Savage and Villa, 1986: 17, 148, 169 (part). Lieb, 1988: 171 (part). Villa et al., 1988: 63 (part). Pérez-Santos and Moreno, 1989: 4 (part; Colombia listed but no details given). Ibanez and Solis, “1991” [1993]: 30, 33 (Panama). Lieb, 1991: 522.1-522.2 (part). Perez- Santos et al, 1993: 116. Guyer, 1994: 382 (Costa Rica). Auth, 1994: 16 (part). Ibanez et al., “1994” [1995]: 26 (Pana- ma). Lee, 1996: 310-311 (part). Camp- bell, 1998: 207 (part). Stafford, 1998: 16— 17. Lee, 2000: 282-283 (part). Stafford and Meyer, 2000 (part; specimen from Costa Rica), pl. 113. Savage, 2002 (part): 655-656, pl. 416. Goldberg, 2003: 298-300 (part). Kohler, 2003: 200, figs. 478 (Panama) and 479 (Costa Rica) (part). Stafford, 2003: 111 (part; specimens from Nicaragua, Costa Rica, Panama). Wilson et al., 2003: 18. Soldr- zano, 2004: 236-239, fig. 60 (part). Guyer and Donnelly, 2005: 185, pl. 149 (part). McCranie et al., 2006: 147-148 (part). Wilson and Townsend, 2006, tables 1, 3. Kohler, 2008: 215, figs. 580-581 (Panama, Nicaragua) (part). McCranie, 2009: 12. Savage and Bola- nos, 2009: 14 (part). Bolanos et al., 2010: 12 (part). McCranie, 2011: 111 (part). Holotype (Figs. 2B, 7, 14C). LACM 148593, an adult male from Finca La Selva, 40 m elevation, Heredia Province, Costa Rica. Collected 9 December 1974 by C. Dock, Carl Lieb, and Catherine Toft. Formerly CRE (Costa Rica Expeditions) 8598. The holotype is 908 mm total length; 574 mm SVL; 334 mm tail length (com- plete). It has 152 ventrals "and 123 subcaudals. The left hemipenis is nearly fully everted (small portion of apex un- everted); right hemipenis everted to the base of the apex just beyond the flounces. Most of the stratum corneum is missing from the dorsal scales. The middle half of the venter has a long midventral slit exposing the viscera; a shorter irregular slit (about 100 mm long) begins about 35 mm in front of the vent. SPECIES IN THE DENDROPHIDION VINITOR COMPLEX ¢ Cadle 199 Figure 7. Paratopotypes. LACM 148591, 148595, 148597-98, 148603—05, 148609-11, 148616, 148619, 148621: CM 53948. Other Paratypes. Honduras: Gracias a Dios: Bodega de Rio Tapalwas, 190 m, USNM 559615, 561032. Cano Awalwas (camp), 100 m, USNM 559616. Crique Ibantara, 70.m, USNM 559617. Warunta Tingni Kiamp, 150 m, USNM_ 561920. Hiltara Kiamp, 150 m, USNM 562874. Sachin Tingni, 150 m, USNM 562875. Near Crique Wahatingni, USNM 562876. Near Crique Yulpruan, 140 m, USNM 563303. Olancho: Planes de San Esteban, 1,100 m, USNM 565534. Nicaragua: Eastern Nica- ragua, ANSP 22863-64 (paratypes of D. vinitor Smith). [Atlantico Norte]: Musawas, Waspuc River, AMNH R-75223. [Atlantico Sur]: Cara de Mono, KU 112974. Recreo, Rio Mico, UMMZ 79766 (paratype of D. vinitor Smith; see Appendix 2 for locality clarification). Rio Mico, 10 mi above Re- creo, UMMZ 79765 (see Appendix 2 for locality clarification). Matagalpa: Hacienda La Cumplida, 19 km N of Matagalpa, 2500 tt. 62 mile -UMMZ 115259, Mata- galpa, MCZ R-9561 (paratype of D. vinitor Smith), UMMZ 90670 (paratype of D. vinitor Smith, formerly MCZ R-17117). Costa Rica: Alajuela: Poco Sol de La Tigre, Holotype of Dendrophidion apharocybe (LACM 148593) from La Selva, Costa Rica (Heredia Province). 540 m, LACM 148601. Cartago: ca. 2.5 km N Pavones, 700 m, LACM 148594. Pavones, near Turrialba, KU 140055. Guanacaste: Cacao Biological Station, 729-1,528 m, LACM 148589. Silencio, 10 km SES [P SSE] from La Casa, 875-940 m, LACM 148607. Heredia: Rio Puerto Viejo near junction with Rio Sarapiqui, KU 35639. 10 km WSW Puerto Viejo de Sarapiqui, MVZ 217610. Zona Protectora, La Selva, trail from 1,000-m camp to 1,500-m camp, 990 m, LACM 148600. Limon: Pandora, 50 m, LACM 148618. Suretka, MCZ R- 19342 (paratype of D. vinitor Smith). Near Suretka, Mt. Mirador, KU 35638. Panama: Bocas del Toro: Almirante, 10 m, KU 80223. 11 km NW Almirante, 600 ft. [183 ml, FMNH 153653, 154038-39. South end of Isla Popa, | km E of Sumwood Channel, LUSNIM 34.7352, da, Loma, (VW ..! Panama, MCZ R-19344 (paratype of D. vinitor Smith). Peninsula Valiente, Bluefields, 70 m, KU 107646. Peninsula Valiente, Quebrada Hido, USNM 338624. Cocleé: Continental divide N El Copé, 600-700 m, AMNH R-115922. Darién: North slope of Cerro Mali, 700—-1,200 m elev., AMNH R- 119377. Laguna, 820 m, KU 75680. [Pana- ma]: Pequeni—Esperanza ridge, near head of Rio Pequeni, 2,000 ft. [610 m], MCZ 200 R-42782 (paratype of D. vinitor Smith). Pequeni—Esperanza ridge, junction main divide, 1,200 ft. [366 m], MCZ R-42783 (paratype of D. vinitor Smith). Panama: Cerro Azul region, Rio Piedra, AMNH R- 119878. Cerro Campana, 3,000 ft. [915 ml], UMMZ 155745. S te of Cerro Campana, 900-950 m, AMNH _ B-108693. San Blas: Border of Darién, summit site, 320 m, 08°55'N, 77°51’W, FMNH 170138. Vera- guas: Cerro Delgadito, 2-4 mi. W Santa Fe, AMNH BR-147806-07. Cerro Arizona above Alto de Piedra, North of Santa Fe, 4,700 ft. [1,433 m], UMMZ 155725. Other Referred Specimens and Locality Records from Literature (specimens not seen except LACM 148622). Costa Rica: Heredia: Finca La Selva (type locality), LACM 148622 (fragmentary skin), LACM 163285-87 (skeletons). Limoén: Guapiles, UCR 6327 (Lieb, 1988). 7 km NW Phe San Clemente, UCR 7545 (Lieb, 1988). Honduras: Gracias a Dios: Las Marias (McCranie, 2011: 114). Nicaragua: No specific locality, USNM 14215 (paratype of D. vinitor Smith) (another paratype of D. vinitor from an unspecified locality in Nica- ragua, USNM 14220, was reidentified as “D. nuchale” by Stafford [2002, 2003], but that identity seems not to have been independent- ly verified). [Chontales]: Santo Domingo, Chontales Mines, 2,000 ft. [610 m], BMNH 94.10.1.19-20 (Stafford, 2003). [Rio San Juan]: Rio San Juan (KGhler, 2008, fig. 581). Panama: Bocas del Toro: South end of Isla Popa, 1 km E Sumwood Channel, USNM 319234 (skin + skeleton). Savage (2002: 656) plotted additional Costa Rican localities for D. apharocybe based on specimens that I did not see (mainly at UCR), but see comments under Distribution on two erroneous lowland Pacif- ic localities indicated by Savage. Etymology. The species name is a femi- nine noun in apposition derived from the Greek words aphares (apaupys), meaning naked or unclad, and kybe (KkbBn), meaning head. The “naked head” refers to the distinctive unadorned apex of the hemipenis of D. apharocybe compared with its hla species. Bulletin of the Museum of Comparative Zoology, Vol. 160, No. 4 Diagnosis. Dendrophidion apharocybe is characterized by (1) dorsocaudal reduction from 8 to 6 occurring posterior to subcaudal 25 (range, 26-63); (2) single anal plate; (3) relatively low subcaudal counts (<130 in males and females); (4) black-edged pale crossbands on the neck nearly always more than one scale row wide; (5) immaculate ventrals and subcaudals except for lateral dark pigment; (6) a relatively short hemi- penis with a bulbous apex strongly inclined toward the sulcate side (asulcate edge of apex higher than sulcate edge) and largely devoid of ornamentation (apex nude). The combination of few subcaudals and a single anal plate will distinguish D. apharocybe from all other species of Deneronhiee except D. vinitor, D. crybelum, and D. paucicarinatum. Dendrophidion apharocybe differs from species of the D. percarinatum group (D. bivittatum, D. brunneum, D. paucicarina- tum, D. percarinatum) in having the dorso- caudal reduction from 8 to 6 usually posterior to subcaudal 30 (26-30 in some specimens from Costa Rica and Panama; see Sexual Dimorphism and Geographic Trends). A single anal plate will distinguish D. apharocybe from all of these except some individuals of D. paucicarinatum (anal plate variable in this species). Dendrophidion paucicarinatum usually has a more uniform- ly colored dorsum lacking distinct cross- bands, has narrow dark lines across the venter in adults and many juveniles, has a higher number of ventrals (>175 compared with <170 in D. apharocybe), and has more weakly keeled dorsal scales. Dendrophidion apharocybe differs from D. boshelli in having 17 midbody scale rows (15 in D. boshelli). Dendrophidion apharocybe has fewer subcaudals (<130) and usually a shorter adult relative tail length (<60% of SVL) than D. nuchale auctorum and D. dendrophis (>130 and usually >60% of SVL, respectively); the anal plate may be either single or divided in these last two species, and their venters are often heavily marked with dark pigment (immaculate in D. apharocybe). SPECIES IN THE Dendrophidion apharocybe previously has been confused with another new species, D. crybelum, and with D. vinitor as redefined herein. Dendrophidion aphar- ocybe differs from D. erybelum (characters in parentheses) in the following hemipenial characters: hemipenis rather short and with a bulbous apex comprising well over one- third the length of the organ (longer and cylindrical, without a distinctly oe anded apex that comprises one-fourth or less the length of the hemipenis); apex strongly inclined toward the sulcate de and nude (not inclined and ornamented with many free-standing membranous ridges having embedded spinules); hemipenis with rela- tively few moderately enlarged spines, total enlarged spines <45 ( (many ed enlarged spines, total enlarged spines >70). eA. phidion apharocybe and D. crybelum are very similar in color patterns, but D. apharocybe has immaculate ventrals and subcaudals, whereas adult D. crybelum have small dark spots on the posterior ventrals and the subcaudals (juveniles sometimes have only dark suffusion on the subcaudals); see species account for D. crybelum for details. Dendrophidion apharocybe averag- es more pale body bands (Table 1) than D. crybelum (p < 0.001) and fewer than D. vinitor (p < 0.001), but the ranges overlap greatly in each case. Hemipenes of D. apharocybe and D. vinitor are similar in overall shape, but the apex of the former is nude and strongly inclined toward the sulcate side, and it lacks an apical boss. The hemipenis of D. vinitor has a highly ornate apex, including an apical boss, and it is not strongly mee (see detailed hemipenial descriptions). Addition- ally, these two species gitfer 4 in aspects of color pattern (see species account for D. vinitor). Description (31 males, 34 females). Ta- ble 1 summarizes size, body proportions, and meristic data fot D. apharocybe; McCranie (2011) summarized data for 10 Honduran specimens, most of which are also included in this summary. Largest specimen (KU 35638) a female 1,045 mm DENDROPHIDION VINITOR COMPLEX © Cadle 20] total length, 672 mm SVL. Largest male (KU 80223) 1,040 mm total length, 653 mm SVL (another male, LACM 148601, was also 653 mm SVL but had an incomplete tail). Tail 35-38% of total length (53-61% of SVL) in males; 33-36% of total length (49- 57% of SVL) in females. Dorsal scales in 17-17-15 scale rows, the posterior reduc- tion by fusion of rows 2+3 (40%) or 3+4 (54%) or loss or row 3 (6%) at the level of ventrals 85-105 (see Sexual Dimorphism below). Ventrals 149-160 (averaging 153.9) in males, 152-168 (averaging 160.8) in females; 1 or 2 preventrals anterior to ventrals (preventrals rarely absent). Anal plate single. Subcaudals 115-127 (averaging 121.1) in males, 111-129 (averaging 119.9) in females. Dorsocaidal reduction at sub- caudals 32-63 in males (mean 47.8), 26-52 in females (mean 41.5). Preoculars 1, postoculars nearly always 2 (rarely 3), primary temporals usually 2 (rarely 1), secondary temporals usually 2 (rarely 1), supralabials usually 9 with 4-6 bordering the eye (range 8-10 with other combina- tions bordering the eye; Table 1), infrala- bials usually 9 (range 7-10). Maxillary teeth 33-44 (averaging 39), typically with four posterior teeth abruptly enlarged; some- times the enlargement ap eared more gradual and with three or five somewhat enlarged (Fig. 8). Two apical pits present on dorsal scales. Nearly 60% of specimens have all dorsal rows except row | keeled on the neck (in about half of those, keels on row 2 were scored as weak); mother 34% of specimens lacked keels on rows | and 2 on the neck (in one-quarter of those keels were weak on row 3); the remaining specimens lacked keels on rows 1-3 on the neck (these proportions differ from D. crybelum; see species account). In all except five speci- mens, keels were present on all dorsal rows except row 1 at mid- and posterior body; in the five exceptions (four from Panama, one from Costa Rica) keels were present addi- tionally on scale row 1 posteriorly (some- times weak). Fusions or divisions of tempo- ral scales were common, with the following 202 Z Bulletin of the Museum of Comparative Zoology, Vol. 160, No. 4 Figure 8. Maxillary dentition of Dendrophidion apharocybe. Right maxilla of LACM 148597 showing four enlarged posterior teeth (anterior to the right). The gap anterior to the enlarged posterior teeth is an empty tooth position, not a diastema. frequencies: upper primary divided (5), upper secondary divided (5), upper primary + secondary fused (5), lower primary divided (2), lower secondary fused with the last supralabial (4), upper + lower secondary fused (2), upper + lower ae fused (2), lower secondary divided (1). On eight sides, a dorsal projection of the lower primary temporal extended between the upper primary and secondary temporals, separating them and contacting the parietal scale. In one specimen supralabials 6-8 were fused on one side. In LACM 148616 subcaudals 3-16 are entire or only partially divided. Hemipenis unilobed with a bulbous apex. Sulcus spermaticus simple, centrolineal, with a distinctly flared tip. Central part of hemipenis ornamented with spines, distal to which is a series of flounces (about four on the sulcate side to about seven on the asulcate side). The broad apex is strongly inclined toward the sulcate side and entirely nude except for some low rounded ridges (noticeable only with magnification). Ge ographic Variation, "Ontosenctic Vari- ation, and Sexual Dimorphism. No strong geographic trends were evident except in the point of dorsocaudal reduction, which is more distal in northern compared with southern specimens. The mean subcaudal number at the point of reduction for Honduran—Nicar aguan specimens is 45.4 and 50.8 (females aad males, respectively). In Panamanian specimens dhece means are 35.6 and 43.8 (the means for all non- Panamanian specimens combined are 43.5 and 49.3 for females and males. respective- ly). Dorsocaudal reductions anterior to subcaudal 30 occurred only in females from Panama and Costa Rica (uni- or bilateral reductions at subcaudals 26-29 in five specimens). Relative tail length increases with body size in subadults. Specimens <300 mm SVL have tail lengths 30-35% of total length, 44-53% of SVL (N = 15, males and females combined): see Table 1 for adult proportions. Patterns of sexual dimorphism in D. apharocybe are similar to those in D. vinitor (Table 1). Mean sizes of adult males (571 mm '‘\SVL) and females (575 mm SVL) are not significantly different (N = 19 males, 23 females). Females have a signif- icantly greater ventral count than males and the posterior reduction of the dorsal scales occurs farther posteriorly in females (mean ventral 99.3) than in males (mean ventral 94.6; p < 0.001). The dorsocaudal reduction occurs at a significantly more posterior position in males than in females. Males have a proportionally longer tail than females but the sexes do not differ in subcaudal counts. Coloration in Life. Color photographs of D. apharocybe from the type locality are published in Savage (2002, pl. 416), Solor- zano (2004, fig. 60), and Guyer and Donnelly (2005, pl 149): from other Costa Rican localities in Stafford and Meyer (2000, pl. 113) and KGhler (2003, fig. 479); from Honduras in Wilson et al. ( 2003. fig. 4; same photograph but with distinctly red/ orange tones in McCranie et al. [2006, pl. 120]) and McCranie (2011, pl. 6D); from Nicaragua in KGhler (2008, fig. 581); and from Panama in Kohler (2003, fig. 478; 2008, fig. 580). A black and white photo- graph of the head/neck (La Selva, Costa Rica) is in Lieb (1991). Salient characteristics of adult coloration in life, as described by Stafford (1998), SPECIES IN THE DENDROPHIDION VINITOR COMPLEX ¢ Cadle 203 Figure 9. Dendrophidion apharocybe in life. AMNH R-115922 from continental divide N of El Copé, Panama (Cocle Province). Adult female, 666 mm SVL. From color slide by Charles W. Myers. A reverse image of the same photo was published in Kohler (2003: fig. 478, 2008: fig. 580). Savage (2002: 656), Guyer and Donnelly (2005: 185), McCranie et al. (2006: 148), Mc@ranie-- (2001: 112) “and. Charles” W. Myers (field notes for KU 107646; AMNH R-108693, R-115922. R-119377) include a gray to brown dorsum with a series of dark- edged pale grayish to pale brown crossbands (ground Color or crossbands often with reddish or orange tones, especially on head, anterior and midbody). Posteriorly, pale bands tend to become invested with dark pigment, forming transverse series of pale ocelli set within darker pigment. Posterior ocelli are not mentioned in all color descriptions, but they are visible in virtually every published photo from throughout the range. Often a pale vertebral line posterior- ly. Usually a narrow lateral dark line on dorsal rows 2 and/or 3 on the posterior body (sometimes indistinct) and a similar dark line at the subcaudal/dorsocaudal border. Skin between the anterior dorsal scales pale blue or bluish white. Top of head brown, sometimes invested with reddish or green- ish tones. Supralabials and throat pale to bright yellow. Venter in adults whitish to yellow, ‘often with an orangish or greenish wash. Juveniles similar to adults but reddish or orange tones in dorsal ground color reduced or absent; venter usually whitish, greenish white, or orange/yellow. Guyer and Dennelly (2005: 185) de- scribed specimens from the type locality as follows: Dorsum gray brown anteriorly, shading to brown ‘posteriorly. Crossbands pale grayish tan bordered by dark. Anterior- most interspace between pale bands rusty red, the rest gray brown. Skin between the anterior doieal scales pale blue, creating a bluish tint to the light bands or, when fhe lung is expanded, a blue band. Posterior body middorsal tan stripe interrupted by thin, dark bands that shade to gray laterally. Venter immaculate white to light vellow. Head gray brown. Supralabials brownish anteriorly, white to pale yellow posteriorly. Figure 9 illustrates an adult female D. apharoc ybe from Panama, whose coloration in life was described thus (Charles W. Myers, field notes): Head brown, turning gray-brown on neck and then greenish brown over rest of body. Pale crossbands gray anteriorly, pale orange at midbody, pale brown posteriorly. Slen within cross- 204 Bulletin of the Museum of Comparative Zoology, Vol. 160, No. 4 Figure 10. (Nicaragua). (C) UMMZ 115259 (Nicaragua; primary banding pattern obscured by formation of secondary bands). (D) FMNH 153653 (Panama; juvenile, 275 mm SVL). bands bright yellow except on anterior one- third of body, where it is orange for a short distance just behind the head and bluish white thereafter. Labials, underside of head, and neck golden yellow, turning golden orange over rest of venter. Upper one- Sanit of iris tan. lower three- quarters gray brown with a few darker brown spots. Tongue, including fork, black. Coloration in Preservative. Dorsum gray to brown with pale gr ay or brown cross- Representative specimens of Dendrophidion apharocybe. (A) USNM 562874 (Honduras). (B) KU 112974 bands (sometimes offset), broader on the neck than more posteriorly (specimens with intact stratum corneum tend to be brown, those without, gray) (Fig. 10). Each cross- band bordered posteriorly by narrow irreg- ular blackish border: a less distinct border on anterior edge. The number of pale bands/ocelli on the body ranged from 46 to 69 with a mean and cde of 54 and a mode of 53. Neck bands in about 80% of the specimens were | to 1.5 scale rows in width: SPECIES IN THE DENDROPHIDION VINITOR COMPLEX ¢ Cadle in another 18% the bands were up to 2.5 rows wide; in one specimen, the neck bands were three rows wide and in another they were less than one row wide. In some preserved specimens (e.g., Fig. 10C) the pale dorsal bands are indistinct posteriorly, seemingly because of secondary lightening of the interspaces between the pale bands. Crossbands extend down to lateral edges of ventrals and merge with dark pigment on the lateral edges of the ventral scales. Posterior crossbands become invested with dark pigment so that each crossband is broken into a transverse series of vertebral and lateral pale spots (ocelli) separated by dark pigment (lateral ocelli on dorsal rows 3-5 ae comprising adjacent parts of three or four scales). Narrow broken blackish lateral stripe on suture line between dorsal rows 2-3 on posterior third or more of body; distinctness and extent of interruption of this stripe varies. Ventrolateral blackish tail stripe on suture between subcaudals and dorsocaudal row 1. The dark posterolateral and tail stripes are indistinct in some specimens. Top of head uniform brown or gray down to upper edges of supralabials. Ill-defined postocular stripe extending across top edge of penultimate supralabial. Rest of suprala- bials and gular region immaculate whitish. Venter immaculate except for fairly pale brown/gray speckling (often containing some larger dark spots) on extreme lateral edges. Subcaudals immaculate; no investing dark pigment or spots on subcaudals or posterior ventrals (compare D. crybelum). One specimen from Costa Rica (KU 35638) has an unusual ventral pattern: in addition to lateral dark blotches, the anterior third of the body has short dark lateral lines across the anterior edges of the ventral plates. These lines are never complete (as in some other Dendrophidion, e.g., D. paucicarina- tum), the extensions occupying only the lateral portions of the plates. Juveniles are similar to adults but color tones are lighter. Anterior ground color medium to pale brown (unlike juvenile D. crybelum). Ocelli on the posterior body are 205 poorly defined because the investment of pale bands with dark pigment is much less in small juveniles than in adults. Distribution (Fig. 11). Atlantic versant of Central America from extreme eastern Honduras to eastern Panama at the Colom- bia border; upland Pacific drainages in northwestern Costa Rica (Cordillera de Tilaran and Cordillera de Guanacaste) and in Panama. The elevational range derived from data associated with specimens is 10 to 1,433 m (most records <200 m). Records from the Pacific lowlands of Costa Rica (Savage, 2002: 656; Laurencio and Malone, 2009) are erroneous, as discussed below. The distributions of D. vinitor and D. apharocybe are separated by about 400 km at their closest localities in Guatemala/ Belize and Honduras, respectively (Fig. 11). Savage (2002: 656, map 11.79) indicated two lowland localities of “Dendrophidion vinitor’ on the Pacific side of Costa Rica (denoted by “X” in Fig. 17). I conclude that both are based on mistaken identities (I am indebted to Jay M. Savage for pointing me toward the source of the records and to Gerardo Chaves for information and pho- tographs of UCR specimens). The first, due east of the tip of the Nicoya Peninsula (Savage, 2002: 656, map 11.79), is based on UCR 14406 and 14620, which were ob- tained during a survey of Carara National Park (Laurencio and Malone, 2009; David Laurencio, personal communication, March 2011). These specimens were initially iden- tified as “D. vinitor,” but both have divided anal plates and pattern characters of D. percarinatum rather than “D. vinitor” (characters confirmed from photos provided by Gerardo Chaves, who also examined the specimens at my request, May 2011). The second erroneous record is from the Pacific lowlands due north of the Osa Peninsula (Savage, 2002: 656, map 11.79), purportedly based on UCR 7235 cited by Lieb (1988) from Cajon (north bank of the Rio Térraba, about 80 m, Puntarenas Province). Howev- er, that UCR number is seemingly in error, and the UCR collection currently has no specimens of “D. vinitor’ from Puntarenas 206 Bulletin of the Museum of Comparative Zoology, Vol. 160, No. 4 Figure 11. M@ Dendrophidion vinitor @ Dendrophidion apharocybe Distribution of Dendrophidion apharocybe, with adjacent localities for D. vinitor in Guatemala and Belize shown (see Fig. 5 for complete distribution of D. vinitor). A few symbols represent multiple contiguous localities. A detail of the distribution of D. apharocybe in Costa Rica is shown in Figure 17 for comparison with that of D. crybelum. Arrow indicates the type locality of D. apharocybe. Province other than the above-cited mis- identified specimens from Carara National Park. Another specimen from a locality near Cajon, LACM 148592, was previously iden- tified as D. vinitor in the LACM catalogs, but I identify this specimen as D. percar- inatum (again based on a divided anal plate, pattern chaeee and a more proximal dorsocaudal EES) Thus, there are no confirmed records of “D. vinitor’” from the Pacific lowlands of Costa Rica. Dendrophidion apharocybe undoubtedly occurs in northwestern Colombia (Chocoan region) given the proximity of definite Boneraian localities (Fig. 11). However, I have been unable to doeuiaGie a definitive Colombian record, although several authors have listed the species for Colombia (as “D. vinitor’). Pérez-Santos and Moreno (1989) included it in “addenda and corrigenda” to Pérez-Santos and Moreno (1988) but listed its distribution in Colombia as “descono- cida.” Lieb (1988) included Colombia as part of the distribution but listed no Colombian material: he later (Lieb, 1991) plotted a locality well into western Colom- bia on a distribution map for “D. vinitor” but queries to identify the source of the record went unanswered. Stafford (2003: 111) identified LACM 45443 (from Chocé, Colombia) as “D. vinitor,” but this speci- men is a member of the D. percarinatum SPECIES IN THE DENDROPHIDION VINITOR COMPLEX © Cadle group (personal observation). Several other works have cited Colombia as part of the distribution apparently based on these sources. Likewise, Lieb (1991) indicated a locality on the tip of the Azuero Peninsula (Cerro Hoya area) on the Pacific side of western Panama, which is far from any locality from which I have seen specimens (Fig. 11). According to Jay M. Savage (personal communication, March 2011), this locality is based on a misidentified speci- men. Natural History. Dendrophidion aphar- ocybe is a diurnal snake found in humid forests (lowland to montane rainforests, cloud forests) and is the best known species of the D. vinitor complex owing to inciden- tal or focused attention by students of the Costa Rican herpetofauna. Most natural history data for D. apharocybe come from the relatively well-studied La Selva Biolog- ical Station, the type locality, operated by the Organization for Tropical Studies (all literature references to “D. vinitor’): the La Selva ecosystem is described in McDade et al. (1994) and Holdridge et al. (1971, under the name “Sarapiqui’). Guyer (1994: 382) and Guyer and Donnelly (2005) reported D. apharocybe as common and semiarboreal, and Guyer and Donnelly (1990) summa- rized data on size, mass, and body propor- tions for this population. Guyer and Don- nelly (2005: 185) further observed that D. apharocybe is “typically observed crossing trails in primary and secondary forest. It is wary usually races away when ap- proached ... [and is] an adept climber... found at night coiled in understory shrubs and trees. At La Selva this snake consumes ... [Craugastor| bransfordit and ... [Den- dropsophus | ebraccata.” In the Cordillera de Guanacaste, Costa Rica (750-850 m elevation), Stafford (1998) found two juveniles (about 250 mm total length) active at 10:00 a.m. and “late morning on sunny June days after rain. Stafford reported that D. apharocybe is wary and swift and, when alarmed, raises the head and forebody high off the ground (as shown by photographs in Stafford and 207 Meyer [2000, pl. 113] and Kohler [2003, fig. 479]). Goldberg (2003) reported reproductive data for “Dendrophidion vinitor” in Costa Rica based on a mix of specimens of D. Ua Eh and D. crybelum. Because he reported data for individual specimens, it is possible to extract data for D. apharocybe, all of which are from the type locality. A gravid female (LACM 148598, 567 mm SVL) collected 11 December 1974 had a clutch size of five estimated from yolked ovarian follicles >12 mm length. A female (LACM 148609, 657 mm SVL) collected in June 1983, although “not undergoing yolk deposition” (Goldberg, 2003), has swollen convoluted oviducts (personal observation), suggesting that it had perhaps recently laid a clutch. Males undergoing spermiogenesis were collected in April, May, November, and December (N = 4): the smallest actively spermiogenic male, collected in May, was 420 mm SVL but a smaller individual (402 mm SVL) also showed some evidence of spermatid transformation. Goldberg (2003) inferred that sperm pro- duction may proceed year-round. Stafford (2003) included specimens of D. apharo- cybe in an ecological study of Dendrophid- ion spp. but presented summary data only, making it impossible to disentangle data specific to D. apharocybe from the other two species of this complex. Other natural history data for D. apharocybe and D. crybelum combined were summarized by Savage (2002) and Solérzano (2004). I made incidental observations of repro- ductive condition in several other specimens (eggs not counted unless previous incisions in a given specimen permitted thorough study): a female with enlarged (20 mm) oviductal eggs collected 21 June 1967 (KU 140055, Costa Rica; 612 mm SVL): a female with three large shelled oviductal eggs collected 23 February-6 March 1967 (FMNH 170138, Panama; 583 mm SVL); a female with five large nonoviductal eggs about 25 mm long collected in June 1963 (KU 75680, Panama; 534 mm SVL). The smallest individual of D. apharocybe (from 208 Panama) was 172 mm SVL and collected 4 April 1980; two others from Panama were 200 mm SVL (collected 23 January 1953 and os January 1975), and one individual was 204 mm SVL (collected 4 April 1980). The Saito specimens from La_ Selva, Costa Rica, 209-219 mm SVL, were col- lected between 30 June and 26 August. The smallest specimen from the northern part of the range (Honduras) was 228 mm SVL, collected in February 2006. Reports from other parts of the range give portraits of D. apharocybe similar to Costa Rican populations. In Honduras this species is infrequently encountered from 30 to 1,100 m elevation in Lowland Moist and Premontane Wet forests (Wilson et al., 2003: 18: McCranie et al., 2006: 148; Wilson and Townsend, 2006. tables 1, 3; McCranie, 2011: 113). Honduran habitats are discussed and illustrated by McCranie (2011: 22-25, pl. IA—B). Two specimens were active about midday and late afternoon on forest floor; two others were sleeping at night on understory vegetation in forest. ie Spec- imens inflated the anterior body when disturbed, exposing the pale blue skin between the dorsal scales (McCranie, 20d 3): Ibafiez et al. (“1994” [1995]: 26) reported Dendrophidion apharocybe as “infrequent, unpredictable” in eastern Panama, and it has not been recorded for the well-known herpetofauna of Barro Colorado Island, Panama (Rand and Myers, 1990). A juvenile from Cerro Campana (AMNH_ R-108693, 203 mm SVL) regurgitated the remains of an adult Eleuthei ‘odactylus [now Diasporus | diastema (Charles W. Myers, field notes). Myers (1969: 24-25; see also Anthony [1916: 358-363]) described the environ- ment at Cerro Mali (Darién Province, Panama), from which he obtained a speci- men of D. apharocybe (AMNH R-119377): “Cerro Malt is high (about 1,410 meters [AMNH R-119377 from 700 to 1,200 mJ) and wet, but is affected by the dry season and much of its forest is possibly no more than transition to cloud forest. ... the forest contains many small trees, including palms, Bulletin of the Museum of Comparative Zoology, Vol. 160, No. 4 reaching heights of + 12 meters. Larger trees up to + 30 meters are scattered through the forest. There is an understory of bushes and ferns. Most of the trees have a thin covering of moss on the trunks. There is a thick mulch layer on the ground and many rotting logs. Many bromeliads and a few orchids present; few tree ferns.” Like the other species of the D. vinitor complex (see species accounts), snakes with broken/healed tails in D. apharocybe were of low frequency (13.3%) compared with some other species of Dendrophidion (>30%, e.g., Stafford, 2003; Cadle, 2010). Dendrophidion apharocybe is sympatric with at least two other species of Dendro- phidion in some parts of its range. The distributions of D. apharocybe and D. percarinatum broadly overlap from Hon- duras through Panama, and both species occur at some localities (e.g., near Recreo, Nicaragua, and La Selva, Costa Rica). Dendrophidion apharocybe is also broadly sympatric with nuchale auctorum in Costa Rica and Panama, and both species have been collected together at some localities (e.g., Laguna, Darién, Panama, and La Selva, Costa Rica, where both occur with D. percarinatum). Dendrophidion nu- chale auctorum has not been formally reported from La Selva (Guyer, 1994; Guyer and Donnelly, 2005), but I identify as this species a color photograph in Guyer and Donnelly (2005, pl. 148, “Brown Forest Racer” = D. percarinatum according to the authors); the specimen was photographed and released at La Selva (Craig Guyer, personal communication, December 2010). Sol6rzano (2004) used the vernacular name corredora quillada for D. apharocybe in Costa hica: Because its diet comprises small terres- trial frogs almost exclusively (e.g., Stafford, 2003), populations of D. apharocybe in Costa Rica and Panama are undoubtedly affected to some extent by well-documented declines in amphibian populations in lower Central America. For example, populations of prey species of frogs at La Selva, the type locality, collectively have declined by about SPECIES IN THE DENDROPHIDION VINITOR COMPLEX © Cadle 209 Figure 12. Holotype of Dendrophidion crybelum (LACM 148599) from Finca Las Cruces, Costa Rica (Puntarenas Province). 75% since 1970, probably due to climate- driven changes in the amount of surface leaf litter evthittield et al., 2007). Several localities in western Panama from which D. apharocybe is documented (e.g., El Copé and Santa Fe) have experienced precipitous declines in amphibian populations due to disease. Lips et al. (2006) documented >90% decline in abundance and >60% decline in amphibian species richness at E] Copé. All of these declines involved species of Pristimantis, Craugastor, and Diasporus, which are the predominant prey for Den- drophidion (Stafford, declines and associated climate changes have unknown effects on snake predators such as D. apharocybe, but they deserve study. Dendrophidion crybelum New Species Figures 2C, 12, 13A, 13C, 14A, 14B, 15, 16,22, 23 ? Drymobius dendrophis, part. Boulenger, 1894: 15-16 (? specimen g from “Chir- iqui’). See comments under Distribution. 2003). These prey Dendrophidion vinitor, part (Southwestern Costa Rica mentioned explicitly or im- plicitly as part of the distribution, or listing of specimens in the type series). Savage, 1973: 14. Savage, 1980: 92. Scott et al., 1983: 372 (“Las Cruces”). Savage and Villa, 1986: 17, 148, 169. Lieb, 1988: 171. Villa et al., 1988: 63. Lieb, 1991: 522.1-522.2. Auth, 1994: 16. Lee, 1996, fig. 164 (illustration of head scales of CRE 5099; = LACM 148612). Savage, 2002: 655-656. Goldberg, 2003: 298- 300. Stafford, 2003: 111. Solorzano, 2004: 236-239. Guyer and Donnelly, 2005: 185. McCranie et al., 2006: 147-— 148. Kohler, 2008: 215. Savage and Bolanos, 2009: 14. McCranie, 2011: 111 (part). Holotype (Figs. 2C, 12, 14B). LACM 148599, an adult male from Finca Las Cruces, near San Vito de Java, 4 km S San Vito, 1.200 m elevation, Puntarenas Prov- ince, Costa Rica. Collected in September 1972 by James E. DeWeese and Ron T. Harris. Formerly CRE 3182. The holotype is 850 mm_ total length, 552 mm SVL, 210 298 mm tail length (complete), and has fully everted hemipenes (the left one removed for description and illustration). It has 152 ventrals and 115 subcaudals. Supralabial 2 on the left side is partially divided by a suture on its upper edge. The left and right upper primary temporal and the right lower primary temporal are divided by a vertical suture. Three long midventral slits through the body wall (one on anterior body, two posterior to attached collection tags). Most of the stratum corneum is missing from the dorsal scales. A histological study of the reproductive system indicated active sper- miogenesis (Goldberg, 2003). Paratopotypes. LACM — 114106—08, 148596, 148602, 148606, 148608, 148613- 15, 148617. The stated localities for the topotypes vary somewhat among the spec- imens according to the LACM catalogues (different collectors over a period of years). All have the basic locality “Finca Las Cruces” with little more specific informa- tion. Elevations associated with paratopo- types range from 1,100 to 1,300 m. Other Paratypes. LACM 148590, 148620 from Finca Loma Linda, 2 km SSW Canas Gordas, 1,170 m, Puntarenas Province, Costa Rica. LACM 148612 from Finca Las Alturas, vicinity of main plaza and surround- ing streams and forests, 1,330 m, Puntar- enas Province, Costa Rica. UF 16425 from Finca Mellizas, 14 km ENE La Union, near the Panama border [approximately 1,310 m], Puntarenas Province, Costa Rica. Etymology. The species name is derived from the Greek adjective krybelos (kpvB- nAoc) meaning “hidden” or “secret.” Trans- literation to Latin yields crybelum (with neuter gender ending to agree with the neuter generic name). It recognizes the cryptic nature of this species and the fact that it remained unrecognized for so long in a well-studied herpetofauna. Diagnosis. Dendrophidion crybelum is characterized by (1) dorsocaudal reduction from 8 to 6 occurring posterior to subcaudal 40 (range, 42-58); (2) single anal plate; (3) low subcaudal counts (<120 in males and females): (4) black-edged pale crossbands Bulletin of the Museum of Comparative Zoology, Vol. 160, No. 4 on the neck nearly always more than one scale row wide; (5) pale crossbands con- tinuing to the vent but posteriorly tending to become invested with dark pigment, form- ing a lateral and vertebral series of ocelli within dark bands; (6) in adults, a tendency for small dark spots and flecks on subcau- dals and posterior ventrals; (7) a relatively long cylindrical hemipenis with a large number of spines (>70) and a short, nonbulbous apex ornamented with free- standing membranous ridges. The combi- nation of few subcaudals and a single anal plate will Aeon D. crybelum Pom all other species of Dendrophidion except D. vinitor, D. apharocybe, and D. paucicar- inatum. Additional distinguishing characters and comparisons include the following. Dendro- phidion crybelum difters from species of the D. percarinatum group (D. bivittatum, D. brunneum, D. paucicarinatum, D. percar- inatum) in having the dorsocaudal reduction from 8 to 6 occurring posterior to subcaudal 40. A single anal Bits will distinguish D. crybelum from all of these except some individuals of D. paucicarinatum (anal plate variable in this species), but D. paucicar- inatum lacks distinct pale crossbands in adults; has narrow dark lines across the venter; and a large number of ventrals (>175 compared with <165 in D. crybe- lum). In addition to having a divided anal scale, D. percarinatum has narrow pale crossbands (<1.5 scale rows wide) on the neck. Dendrophidion bivittatum and_D. brunneum have divided anal scales and different color patterns (posterior blackish dorsolateral all lateral stripes in D. bivitta- tum and uniform greenish or with a combination of paravertebral stripes or spots, or indistinct crossbands in D. brun- neum; see Cadle, 2010). Dendrophidion crybelum differs from D. boshelli in having 17 midbody scale rows (15 in D. boshelli). Dendrophidion crybelum differs from two other members of the D. dendrophis group, D. dendrophis and D. nuchale auctorum (characteristics in parentheses), in having fewer subcaudal scales (2130); a SPECIES IN THI Figure 13. Comparison of closely size-matched males of (A, C) * DENDROPHIDION VINITOR COMPLEX * Cadle A Dendrophidion crybelum (LACM 148590, 625 mm SVL) and (B, D) D. apharocybe (LACM 148601, 653 mm SVL). (A) and (B) are to the same scale, (C) and (D) to the same scale. Relative to body size, Dendrophidion crybelum has a more robust body and a broader head than D. apharocybe. venter without extensive dark pigment, usually only scattered spots on the most posterior ventrals (venter often heavily marked with dark pigment, especially pos- teriorly); smaller body size (adults common- ly more than | m in total length); a shorter relative tail length in adults, <60% of SVL (tail usually more than 60% of SVL); and in hemipenial morphology (hemipenes_ bul- bous rather than long and cylindrical; a spinose battery followed distally by several transverse flounce-like structures). In D. dendrophis and D. nuchale auctorum the anal plate may be either single (as in D. crybelum) or divided. Dendrophidion crybelum difters from D. apharocybe and D. vinitor most notably in hemipenial morphology, including overall shape (elongate and cylindrical vs. shorter and bulbous in the last two species), number of spines (>70 vs. <45), and apical morphology (narrow and with reduced calycular structures vs. nude and strongly inclined in D. apharocybe, or with well developed membranous ridges and an apical boss in D. vinitor). See complete descrip- tions for details. Other characteristics distinguishing D. crybelum from D. apharocybe and D. vinitor are subtle. Dendrophidion crybelum is a more robust animal than either D. apharocybe or D. vinitor, which is most easily seen in side-by-side comparisons of individuals of comparable body length, as shown for two males illustrated at the same scale in Figure 13. The body of D. erybelum is more massive, and the head is larger and more angular, than similar sized specimens of D. apharocybe or D. vinitor. Adults of D. crybelum have fine dark speckling on the subcaudal scales, often concentrated along suture lines, and small dark flecks and spots on the posterior ventral plates (Fig. 14). However, these features vary among adults (very distinct and numer- ous to only scattered flecks), and they appear to develop ontogenetically. Juveniles often have only a fine peppering on the distal subcaudals forming a dark suffusion easily visible only with magnification. On the other hand, apart from lateral dark pigment on the ventrals and subcaudals common to all species of Dendrophidion, the ventral plates and subcaudals of D. apharocybe and D. vinitor are immaculate (Fig. 14; see Fig. 6 for D. vinitor). Dendrophidion crybelum averages fewer pale bands on the body than D. apharocybe or D. vinitor (Table 1; p < 0.01 and p < 0.001, respectively), although the ranges overlap greatly. The pale neck bands in D. crybelum are typically broader than those of D. vinitor (Fig. 2). Description (8 males, 8 females). Table 1 presents standard meristic and mensural Bulletin of the Museum of Comparative Zoology, Vol. 160, No. 4 Figure 14. Comparison of posterior ventral and subcaudal patterns of Dendrophidion crybelum and D. apharocybe. Left series: Ventral plates immediately anterior to the vent. Right series: Tail base immediately posterior to the vent. Vent toward the center in each series. (A, B) D. crybelum (LACM 114107 and LACM 148599 [holotype], respectively). (C, D) D. apharocybe (LACM 148593 [holotype] and LACM 148609, respectively). Posterior ventrals and subcaudals of D. crybelum have fine dark spots and speckling compared with the immaculate ventrals and subcaudals of D. apharocybe. The specimen in panel A is atypical for adult D. crybelum in having only a few widely scattered spots on the ventral scutes, but its subcaudals were extensively speckled; the ventral spotting of the specimen in panel B is more typical. Dendrophidion vinitor is similar to D. apharocybe in having immaculate posterior ventral and subcaudal patterns (see Fig. 6 for an example). data for D. crybelum. Largest male (LACM 148590) 985 mm total length, 625 mm SVL: largest female (LACM 114107) 631 mm SVL, 893+ mm total length (tail incomplete; a female 621 mm SVL was 956 mm total length). Tail 34-36% of total length (51- 58% of SVL) in males; 34-35% tot total length (50-54% of SVL) in females. Dorsal scales in 17-17-15 scale rows, the posterior reduction usually by fusion of rows 3+4 (90%: remainder 2+3) at the ventrals 93-104 (see Sexual Dimorphism below). Ventrals 150-153 (averaging 151.6) in Sales, - 1562162 (averaging 16( res females; 1 or 2 preventrals anterior to ventrals. Anal plate divided. Subcaudals 112=119: (mean 1916/9) ineamalese dat5=119 (mean 117.2) in females. Dorsocaudal reduction at subcaudals 43-58 in males (mean 49), 42-56 in females (mean 48.6). Preoculars 1, postoculars 2, primary tempo- rals 2, secondary temporals 2. Supralabials nearly always 9 with supralabials 4-6 level of bordering the eye (rarely 8 with 3-5 bor dering eye). Infralabials 9. Maxillary teeth 3844. usually with the last four ‘patna three) enlar ged. Two apical pits present on dorsal scales. In most specimens (including juveniles) all dorsal rows except row | are keeled from the neck to the vent. In occasional speci- mens (juveniles and adults) all rows are keeled at midbody and posteriorly (in these cases scales in row | usually have very weak keels and sometimes not every scale in row 1 has a detectable keel); one adult lacked keels on rows 1-3 on the neck only. Divisions (but no fusions) of temporal scales were recorded as follows: upper primary divided (5), upper secondary divided (6), lower primary divided (4). Hemipenis afndee much longer than wide, lacking a bulbous apex. Sulene sper- maticus simple centrolineal, and with a flared tip. Hemipenial body dominated by a great number of enlarged spines. Spines SPECIES IN THE Figure 15. Dendrophidion pus in life from the type locality (paratypes). (A) LACM 114107 (adult female, 631 mm SVL). (B) LACM 114108 (juvenile female, 298 mm SVL). From color slides by Roy W. McDiarmid. followed distally by a very short apex ornamented with a crowded series of flounces, calyces, and (on the apex proper) free- standing ridges. Small apical region essentially eiencd with these ornaments (ie., nude areas very small). Sexual Dimorphism and Ontogenetic Variation. Small sample sizes hamper a full exploration of sexual dimorphism, and the only statistically significant standard scale count differences are the greater number of ventrals in females compared with males (p <« 0.001; Table 1) and a different point of dorsal scale reduction (mean ventral num- bers 96 and 102.1 for males and females, respectively; p < 0.001). Males and females do not differ significantly in mean adult SVL (581 mm and ‘BTA mm SVL, respectively), subcaudal number, or the point of dorso- caudal reduction. The difference between male and female relative tail lengths is marginally significant (p = 0.045). The bo ww DENDROPHIDION VINITOR COMPLEX ¢ Cadle mean points of dorsocaudal reduction in male and female D. crybelum differ by less than one scale, and this character presents a departure from the pattern of sexual dimor- phism in D. vinitor and D. apharocybe, which are sexually dimorphic for this character (Table 1). Specimens <300 mm SVL have relatively shorter tails (31-34% of total length, 46-52% of SVL: N = 4, males and females ee than larger speci- mens (Table 1) Coloration In Life. | am unaware of previously published Ea Orr wy DD crybe lum. Figure 15 presents photographs of D. crybe Jum in life from color slides by Roy W. McDiarmid. An adult female (LACM 114107, Fig. 15A; the same speci- men is illustr ed in Fig. oe has a deep chocolate brown gr aun color; pale bands washed with orange on the neck and the posterior two- dhs of the body, grayish in between, and invested irregular ly ah black (tendency for the bands to form transverse series of ocelli posteriorly); a pale vertebral line washed with orange; and whitish Hee The juvenile (LACM 114108, Fig. 15B) i patterned similarly but has a pale ae ground color (somewhat tan): pale bands/ ocelli and pale vertebral line grayish ante- riorly, invested with yellowish aoe m poste- riorly and on the first two or three bands on the neck. McDiarmid’s slide collection also has a photo of LACM 114106 in life, an adult female with clouded eyes (preparing to shed). Its coloration appears similar to that of LACM 114107 except perhaps a paler brown ground color (not as pale as the juvenile described above) and more orang- ish wash in the pale bands. The venter A LACM 148614 (339 mm SVL) was de- scribed as “yellowish” (Roy W. McDiarmid, field note). Coloration in Preservative. Overall dorsal coloration predominantly grays and browns (tendency for brown with stratum corneum me grays without). A series of pale g oray to pale beaut crossbands (sometimes offset) on the body from the neck to the vent, continuing onto the tail as a variably distinct series of pale dorsolateral spots (Fig. 16). 214 Bulletin of the Museum of Comparative Zoology, Vol. 160, No. 4 Figure 16. Representative specimens of Dendrophidion crybelum. (A) LACM 114107 (primary banding pattern obscured partly by formation of secondary bands and darkening of primary bands during preservation; compare Fig. 15A). (B) LACM 148596. (C) LACM 148613 (juvenile, 304 mm SVL). (D) LACM 148620 (juvenile, 185 mm SVL). The number of pale bands/ocelli on the body ranged from 36 to 62 (mean 48.2). In about 58% of the specimens neck bands were 1.5—2 scale rows in width; in another 25% the bands were 1—1.5 rows wide; in the remainder, the neck bands were more than two scale rows wide or less than one row wide. Crossbands bordered anteriorly and pos- teriorly by irregular black pigment forming narrow jagged border (usually better devel- oped on posterior edges). Posterior cross- bands become invested with greater amount of dark pigment, usually becoming frag- mented into a lateral and vertebral series of ocelli (lateral ocelli occupy parts of four adjacent scales on rows 3-5); less distinct ventrolateral ocelli sometimes on row l. Crossbands extend to lateral edges of ventrals, which are marked with dark gray SPECIES IN THE DENDROPHIDION VINITOR COMPLEX ¢ Cadle and brown or black irregular spots. The first few bands on the neck and anterior body are broader than more posterior bands. Narrow broken blackish lateral es (or series of dashes/spots) occurs on the suture line between rows 2 and 3 on the posterior one-third of body. Black ventrolateral tail SHIPS on suture line between subcaudals and dorsocaudal row | extends to tail tip. Two specimens (LACM 148596, 148599) have more vivid banding pattern than others; these two also tend to be grayer than others, which are browner. Crossbands obscure in other preserved adults (LACM 148590, 114106—07) seemingly because they are invested with brown pigment (except for neck bands) nearly the same color as the ground color and the crossband borders are not distinctly marked with pigment darker than aan color. Specimens with a dark brown ground color and contrasting pale bands in life may have obscure bands when preserved (compare Figs. 15A and 16A). Top of head more or less uniform gray or brown down to superior edge of supralabials (last two supralabials mostly gray or brown). Faint blackish postocular stripe extends diagonally down across penultimate supra- labial. Gular region immaculate. Lateral portions of ventrals with dense grayish black pigment, within which are darker irregular spots. Remainder of most ventrals immaculate. However, posterior ventrals (up to about the 20th ventral anterior to the vent) have small scattered irregular black spots that vary in number from relatively many (LACM 148596, 148599) to almost none (LACM 114106- 07) (Fig. 14). Subcaudals with ventrolateral stripe (described above) and generally with dense blackish/gray pigment investing sub- caudals, especially along suture lines, and scattered irregular Sat round black spots on many subcaudals (Fig. 14). Juveniles similar to adults but pale bands/ ocelli are much better defined. Spotting or ee on posterior ventrals a proximal subcaudals very faint, but even relatively small specimens have fine speckling of dark pigment on posterior portion of tail (e.g., 215 @ Dendrophidion apharocybe A Dendrophidion crybelum Figure 17. Distribution of Dendrophidion crybelum and of D. apharocybe in Costa Rica; see Savage (2002) for a more complete representation of localities. Arrows indicate type localities. A few symbols represent multiple contiguous localities. Two localities indicated by “x” within dots in western (Pacific) Costa Rica are erroneous localities for “D. vinitor’ indicated by Savage (2002: 656); see species account for D. apharocybe. LACM 114108, 298 mm SVL), especially on suture lines and laterally. Anterior ground color dark brown to blackish (contrasting greatly with pale crossbands), becoming lighter posteriorly (pale brown to tan on posterior half to two-thirds of body). Poste- rior lateral stripe broken; tail stripe quite distinct. Secondary pale bands between the rimary crossbands already evident in LACM 148613-14. Width of pale dorsal bands in LACM 148614, 148617 is no more than about one scale wide. Juvenile Den- drophidion crybelum have a more contrast- ing pattern and well-developed ocelli on the posterior body compared with juvenile D. apharocybe. Distribution (Fig. 17). Definitely known only from middle elevations (1,100—1,330 m) at the eastern end of the Fila Costena and the south slope of the Cordillera Talamanca in southwestern Costa Rica. All known localities are in the uplands of the upper Rio Coto Brus, separated from the lowlands of the Osa Peninsula—Golfo Dulce by the Fila Costefia. Despite relatively intensive surveys, “D. vinitor” has not been recorded 216 Bulletin of the Museum of Comparative Zoology, Vol. 160, No. 4 from lowland sites on the Osa Peninsula (Scott et al., 1983; McDiarmid and Savage, 2005). See the species account for D. apharocybe for discussion of two lowland Pacific localities erroneously attributed to “D. oimnitor- Given the proximity of Costa Rican localities (Fig. 17), D. crybelum undoubt- edly occurs on the adjacent Pacific versant of western Panama (Chiriqui province), where similar environments occur. Sav age (2002) included “adjacent western [eaetee) eee in his distribution summary for D. vinitor, the basis being a specimen from “Chiriqut” listed in E. R. Dunn’s notes as having 121 subcaudals and a single anal plate (Jay M. Savage, yersonal Communica- tion, March 2011). This may be the same specimen listed by Boulenger (1894: 16; specimen g collected by Godman in Chir- iqui, now BMNH 94.5.17.8), although Boulenger’s subcaudal count is 128. No one seems to have re-examined the speci- men recently, so the presence of iD): crybelum in Panama remains likely, though unsubstantiated. Nonetheless, D. crt ybelum probably has a very restricted distribution in southwestern Costa Rica and adjacent Panama. Southwestern Costa Rica (Golfo Dulce region) has long been known for both high species os and high herpetofaunal endemism (e. E Aiiaellncant 1966: 712, 716; Savage, 1966. Tao. 20027 S5. -Si3=Siet): anes the presence of a narrowly endemic species of Dendrophidion in this region comes as little surprise. In terms of ov All species composition (Duellman, 1966) and historical biogeographic origins (Savage, 1966, 2002), the herpetofauna of the Golfo Dulce region shows ties to Atlantic lowland herpetofaunas of Costa Rica and Panama. Many closely related species pairs or conspecific populations of amphibians and reptiles show disjunct distributions in these two areas. Moreover, paleoenvironmental modeling (Chan et al., 2011: 528-531) shows that the Rio Coto Brus valley has had a long history of environmental stability. The Baie populations are isolated by the presence of subhumid habitats north and south of this area (Savage, 2002). The biogeographic events responsible for these disjunctions are explored in the discussion. Natural History. Most confirmed speci- mens of D. crybelum come from sites on the north side of the Fila Costefia (separating the Rio Coto Brus Valley from the Golfo Dulce) that have been intensiv ely studied by investigators associated with the Organiza- tion for Tropical Studies (OTS) and others (e.g., Janzen, 1973; Scott, 1976; Fauth et al., 1989: Santos-Barrera et al., 2008). Other close-by localities are on the southern slope of the Cordillera Talamanca. The type locality (Las Cruces) is classified as Pre- montane Wet Forest in the Holdridge system; other localities span a range of forest types, including Premontane and Lower Montane Wet Forest and Rainforest (Holdridge, 1967; Holdridge et al. 1971). Scott (1976: 44, 53, table 1) described the site of quantitative herpetofaunal sampling at Finca Las Cruces (1,200 m) as having a 12-year average rainfall of 4,000 mm/yr, a brief “dry” period January—March, and moderate to steep slopes. Compared with the lowlands, Las Cruces has shorter, smaller trees, more understory, and deeper litter (5-7 cm deep depending on the season). Janzen (1973: 675) added further details: “The overstory canopy is about 30 m high and densely intertwined with vines and heavily laden with epiphytes. The understory is rich in palms, Cyclanthaceae, ferns, Marantaceae, and cycads, and ap- pears in general very similar to that of the Osa [etiam lowland] primary forest sites except that the [Las Cruces] understory appears to have a much heavier epiphyllae load, the tree trunks have heavy layers of lower plants, and there seems to be a slightly reduced species richness of plants in the 0.3 to 2m height range. ... Quercus, Prunus, and Alnus are more prominent than in nearby lowland sites.” The discovery of specimens of D. crybe- lum in Costa Rica came relativ ely late, the first specimens from the type locality collected by Jay M. Savage md colleagues SPECIES IN THE DENDROPHIDION VINITOR COMPLEX ¢ Cadle in 1964. (A paratype from another locality, UF 16425, was obtained in 1962 and is the earliest known specimen.) Even in the early 1970s Janzen (1973: 675) reported consid- erable deforestation of lower montane forest in the vicinity of the type locality, which is now a patchwork of severely human-altered habitats (Santos-Barrera et al., 2008). Ac- cording to Savage (personal communication, March 2011) “good primary forest and additional second growth parcels still re- main at the type locality (266 ha: now an OTS field station and Wilson Botanical Garden) and nearby indigenous reserves. Forested habitat at Hack. Loma Linda is considerably altered, whereas Las Alturas is on the margin of La Amistad Biosphere Reserve, ahene substantial montane forest still exists (the habitat patchwork of this area is especially well shown on Google Earth). Unfortunately, the region where D. cry- belum occurs has also experienced consid- erable amphibian population declines (Lips, 1998, 1999; Lips et al., 2003, 2006; Santos- Barrera et al., 2008), a fact relevant to sustainability of Dendrophidion since leaf litter frogs comprise a major portion of the diet in all species (Stafford, 2003). The declining species documented by Lips and coworkers included species of Pristimantis, Craugastor, and Diasporus, which are major prey tor Dendrophidion. The effects of extirpation of such prey species on the population biology of their snake predators are, as yet, litle: understood (Whitfield et al., 2007). In contrast to the population declines at La Selva (see species account for D. apharocybe), the amphibian declines at Las Cruces and nearby sites were precipi- tous and due primarily to the amphibian pathogen Batrachochytrium. A recent her- petofaunal inventory of Las Cruces and surrounding areas (Santos-Barrera et al., 2008) recovered two specimens of. percarinatum but none of the three other species of es recorded there historically (D. crybelum, D. nuchale auc- torum, and D. paucicarinatum; Scott et al., 1983, and personal observations of speci- mens in LACM). The last specimen of D. bo a | crybelum held in U.S. collections was obtained in 1987. The holotype and most par itypes of D. crybelum were included (as “D. vinitor’) in a study of reproductive cycles in Dendro- phidion from Costa Rica (Goldberg, 2003) and range-wide, Mexico to Panama (Staf- ford, 2003; as summary statistics only, along with specimens of D. vinitor and D. apharocybe). Four males in Goldberg's study (552-625 mm SVL) were undergoing spermiogenesis in April, June, August, and September. Two fem: ales (621-631 mm SVL) collected in May and July each had four oviductal eggs. Of the specimens | examined, a female collected 19 June from Las Alturas (LACM 148612, 480 mm SVL), has swollen and convoluted oviducts but no vitellogenic ova, suggesting that it had recently laid a clutch. The two smallest individuals (185 and 206 mm SVL) were collected 6 June and in September, respec- tively, and their umbilical scars were nearly completely closed. The umbilical scar of the next larger (251 mm SVL) was completely closed, aad in two individuals of 304 and 339 mm SVL the umbilical scars were no longer evident. Like the other two species of the D. vinitor complex, there is a low frequency of snakes with broken/healed tails in D. crybelum (6.7% of the specimens examined), HEMIPENIAL MORPHOLOGY An Introduction to Dendrophidion Hemipenes The initial recognition of the new species described herein depended to a great extent on differences in hemipenial morphology. However, Dendrophidion hemipenes are peculiar in ways that have never been fully described: thus, it seems pertinent to provide an overview of their general struc- ture and characteristics. I fac examined retracted and everted hemipenes of all currently described species of Dendrophid- ion except D. boshelli, for which no material has been available. Hemipenes of other species will be described elsewhere. Stuart 218 (1932) long ago recognized some unusual olrictoristics of Dendrophidion hemi- penes, but a detailed description of only one species, D. brunneum, is as yet available (G Gaile, 2010). Hemipenial terminology fol- lows Dowling and Savage (1960), Myers (1974), Myers and Campbell (1981), Myers and Cadle (1994, 2003), Savage (1997), and Zaher (1999), but the terminology useful in describing certain unusual structures in Dendrophidion is further discussed here. Except where noted, these comments apply equally to the D. dendrophis and D. percarinatum groups (sensu Lieb, 1988). The hemipenis of Dendrophidion is either unilobed or slightly bilobed but the overall shape varies considerably among species—trom short, bulbous, and cylindri- cal forms described here to longer, slender and clavate shapes. A short basal nude section reel in the terminology of Savage [2002: 539]) is followed by a broader section usually ornamented with minute spines; there are no nude pockets or basal lobes (some organs have round basal bulges when everted). The relatively unadorned base is followed by an array of closely se spines that are enlarged to varying degrees and in patterns that are species specific. In general, hemipenial spines in the D. dendrophis species group are larger than those in the D D. percarinatum group. Although hemipenes of the D. dendrophis group are sometimes said to have enlarged “basal hooks” or “basal spines” (e.g., Dunn, 1933: 78; Lieb, 1988; Savage, 2002: 654). their enlarged spines are not basal in the same sense as in some snakes, in which enlarged spines are truly positioned at the base of the organ. Rather, they are attached along the central portion of the hemipenial body (truncus in the terminology of Savage [2002: 539]). In some species of both the D. dendrophis and D. percarinatum groups, a pair of spines much larger than any others is positioned near the sulcus spermaticus at the proximal edge of the array of spines (one on each side of the sulcus); this pair is relatively much larger in D. endnotes and D. nuchale auctorum than in other species Bulletin of the Museum of Comparative Zoology, Vol. 160, No. 4 in which they occur (D. dendrophis and D. nuchale auctorum also have an additional pair of enlarged spines positioned toward the asulcate side). Species of the D. vinitor complex lack such conspicuously enlarged spines. Distal to the spines are structures that range from more or less definitive calyces (cuplike structures with both longitudinal and transverse walls) to flounces (transverse ridges lacking longitudinal connecting walls), with an array of intermediate struc- tures that are neither definitive calyces nor flounces. For example, the hemipenial apex in two species of the D. vinitor complex ae long, relatively free-standing ridges that don’t conform to strict definitions of either calyces (because they do not form reticulat- ing networks) or flounces (because they are not transverse in orientation). Nonetheless, the morphological similarity among all of these structures is clear and they undoubt- edly have similar developmental origins. Further reductions of calyxlike structures result in low fleshy ridges or entirely nude apical areas. Proximal calyces/flounces have broader walls than more distal ones. These descriptors refer to structures visible in everted or retracted hemipenes. In addition, retracted hemipenes may have pseudoca- lyces, which are calyxlike structures visible in retracted organs that disappear upon eversion and full inflation (Myers and Cadle, 1994: 13-14; Cadle, 2010: 19-20). The tip of the apex in Dendrophidion has a combination of reduced calyces, free-stand- ing ridges, and/or nude areas in patterns that are species-specific. Flounces, calyces, and other similar structures are ornamented with mineralized spinules, at least proximally, but spinules are usually reduced or absent in distal calyxlike structures. Most spinules in Dendrophidion are atypical in that they lack a tip projecting from atop the calycular walls. Instead, they consist of a mineralized rod completely enclosed by the wall tissue. I refer to these as embedded spinules (see also Cadle, 2010: 16 [fig. 6], 19); they are relatively straight and more or less the same_ thickness throughout. SPECIES IN THE A general pattern in the arrangement of calyxlike structures seems common to. all species of Dendrophidion, regardless of the overall pattern in a given species. Calyces are most fully developed on the asulcate side in comparison to he sulcate side, which has few or no full-fledged calyces. On the asulcate side in most species, a calyculate region extends to the tip of the apex, sometimes as far as its ailcontt Addition- ally, at the proximal edge of the calyculate region, at least one pair of transverse flounces encircles the hemipenis in all species (some species have more than one pair). Transitions between flounces and calyces or calyxlike structures occur abrupt- ly within a single organ. The sulcus spermaticus in Dendrophidion is centrolineal and usually has a slightly flared tip with divergent sulcus lips, but the sulcus is terminally divided in an undescribed species of the D. percarinatum group and in D. dendrophis (a detailed discussion of this morphology will be presented elsewhere). Savage (2002: 539) introduced the term semicentripetal for sulcus conditions such as those Dendrophidion with a simple sulcus on a unilobed organ in which the sulcus extends to the tip of the hemipenis with minimal deviation from the midline of the sulcate surface (e.g., hemipenes described herein). This term is unnecessary because it embodies several aspects of hemipenial morphology for which vocabulary already exists, namely the overall hemipenial form (unilobed vs. bi- lobed), sulcus morphology (simple vs. bifur- cate), and sulcus orientation (centrifugal, centrolineal, or centripetal). Because these three aspects of hemipenial morphology can be combined in various ways, I prefer to employ terms that keep the descriptive concepts separate. Thus, I use a fined) to refer to simple or bifurcate sulci that pass distally with little deviation from the middle of the sulcate side of a hemipenis, whether unilobed or bilobed. Other authors (e.g., Zaher [1999] and Myers [2011]) have also used centrolineal in this broader sense. The use of the term semicentripetal has other awkward consequences. First, semi- DENDROPHIDION VINITOR COMPLEX ¢ Cadle 219 centripetal suggests a relationship to cen- tripetal sulci seen on many bifurcate hemi- penes. This relationship is unclear given that simple sulci can be derived in several ways from distinct bifurcate morphologies (centripetal, centrolineal, or centrifugal), and hemipenes in colubrids may have sulci in any of these orientations (Cadle, 2010: 18-19; see also Myers, 2011: 22-24). Secondly, some genera (e.g., Dendrophid- ion, Leptodeira, Taeniophallus) have species with both divided and simple sulci sperma- tici on unilobed to slightly bilobed organs; see Schargel et al. (2005, fig. 8) for an example from Taeniophallus and Myers (2011: 22-24) for Leptodeira. Using semi- centripetal for those species with a simple sulcus and centrolineal for those with a divided sulcus has the undesirable conse- quence of applying different names to sulcus orientations that are basically the same, the only difference being the simple or divided nature of the sulcus overall. The different terms obscure the clear relation between the simple and divided sulcus conditions within such genera. The use of centrolineal for forked or simple sulci on either bilobed or unilobed hemipenes means that the term applies to a Be te array of sulcus topologies than would be the case if it were used exclusively for forked sulci on uni- or bilobed organs (its original definition, used in conjunction with describing the morphology of some dipsadids; Myers and Campbell, 1981: 16). I believe this is a nonissue inasmuch as I use the terms centrolineal, centripetal, and centrifugal to refer only to the position of the sulcus on the hemipenis overall, regard- less of variations in other aspects of hemi- penial morphology such as lobation or whether the sulcus is bifurcate or simple. Other variations, such as deflections of a simple (centrolineal) sulcus to the right (as in Colubridae) or left (in Natricidae) lobe, can simply be described or accommodated by terms already in use (e.g., dextral and sinistral, respectively; Rossman and Eberle, 1977; Myers, 2011: 14). I believe that using terms such as centrolineal for discrete 220 Bulletin of the Museum of Comparative Zoology, Vol. 160, No. 4 aspects of morphology (in this case sulcus position only), rather than combinations of several aspects, is more straightforward and more directly communicates morphological details. The retractor penis magnus in Dendro- phidion may or may not have a_ short division at the insertion. In species of the D. vinitor complex, some specimens seemed to have a very short separation of muscle fibers at the insertion, but I failed to detect such a separation in others. Whether this reflects true morphological variation or simple difficulty of determination in very short divisions is unclear. On everted hemipenes, two internal points of insertion of the retractor can often be discerned through the apical tissue; the spacing between these points may reflect the degree of division of the muscle. Dendrophidion vinitor Everted (UMMZ 121145, Veracruz state, Mexico; Figs. 18, 19). Hemipenis unilobed and with a somewhat bulbous apex. Total length about 17 mm; about 7.5 mm across the widest point (apex). Sulcus spermaticus simple, centrolineal, with a very slightly expanded tip. The tip of the sulcus is near the sulcate side of the apex and entirely hidden beneath the free edge of a raised knob of tissue, the apical boss described below. Short proximal portion of the hemipenial body mostly nude, having only a band of scattered minute spines on the asulcate side and a few adjacent to the sulcus spermaticus just proximal to the array of enlarged spines. Central region of hemipenial body with short, robust, strongly hooked spines ar- ranged in four to five more or less regular transverse rows all around. Total spines in the array 46. Spines are shorter on the sulcate side and longer on the asulcate side but there is little proximal-to-distal size differentiation. Distal to the spines four or five flounces completely encircle the hemipenis. On the apex, these flounces grade into calyxlike structures and free-standing ridges that are more fully described below. The flounces have a short, thick, fleshy base and a wider outer membranous portion (within which are embedded spinules). Flounces some- what wider on the asulcate side than on the sulcate side, and they gradually narrow in width distally. There is an abrupt transition between the array of spines and the flounces, but in some places, particularly adjacent to the sulcus spermaticus, spines in the distalmost row are incorporated into the proximal pair of flounces (most evident in the proximal flounce) and appear as espe- cially robust spinules (Figs. 18A, 19B). These incorporated spines have a more strongly projecting and hooked distal tip compared with other spinules. On the distal portion of the asulcate side are eight to 10 poorly developed calyces between the stl three flounces (i.e., four to five between the distal two flounces, and another four to five between the penulti- mate and antepenultimate flounce). Longi- tudinal walls of calyces much lower than transverse walls (which make up the flounc- es). Among the proximal three flounces on the asulcate side are a few other poorly developed longitudinal walls, which form very large irregular calyces (obscured by the overhanging flounces in Fig. 18A). All flounces, calyces, and similar struc- tures (including those on the apex) have embedded spinules. Adjacent to the sulcus spermaticus the spinules have a short projecting tip, but away from the sulcus the spinules are completely embedded. Spinules near the sulcus are also hooked (probably represents transitional structure to spines in the array), but they are relatively straight away from the sulcus. Tip of the apex with a complex series of free-standing ridges similar to flounces except that they extend mostly obliquely across the apex from a median axis (Figs. 1SA, 19A, B). Around the periphery of the apex, especially on the asulcate side, a few poorly developed connections among the apical ridges form rudimentary calyces. One of the apical ridges (the median bisecting ridge) is taller than the others SPECIES IN THE DENDROPHIDION VINITOR COMPLEX ¢ Cadle Pa Figure 18. Hemipenes of Dendrophidion vinitor. (A) Everted organ in sulcate and asulcate views (UMMZ 121145; Veracruz, Mexico; right hemipenis). (B) Retracted organ slit midventrally and spread flat (UIMNH 35547; Oaxaca, Mexico; left hemipenis). See Figure 19 for details of apical morphology. and bisects the apex in a line extending from above the tip of the sulcus spermaticus directly across the middle of the apex; it is not straight, but undulates, and has well- developed embedded spinules (Fig. 19B). A series of shorter, lower ridges runs obliquely outward from the median bisecting ridge toward the asulcate side. These alco have embedded spinules but the spinules are less regularly developed than on the median ridge. On. its sulcate end the median bisecting ridge continues across a promi- nent bullous knob (the apical boss), bisect- ing it as well (seen in Figs. 1SA and 19A as a line of denser whitish tissue extending d o across the middle of the boss). The borders Figure 19. Details of apical morphology of the ee of Daharconaion vinitor. (A) on view of UMMZ 121145 eyo, sulcate edge of apex toward bottom). (B) Lateral view of UMMZ 121145 (everted, sulcate side to the right). (C) Apical region of a retracted organ (UIMNH 35547; distal toward the top). Abbreviations: B, apical boss; F, flounces; MBR, median bisecting ridge; R, retractor penis magnus; ss, sulcus spermaticus. Compare Figure 18. a el of the boss are well-defined and raised. The sulcate edge of the boss is elevated above the apical tissue and the sulcus spermaticus ends beneath it. Retracted (UIMNH 35547, Oaxaca state, Mexico; Figs. 1SB, 19B). Hemipenis ex- tends to the middle of subcaudal 8. There appeared to be a slight separation of muscle ce at the insertion of the retractor penis magnus. Extreme proximal portion of hemi- penial body nude, followed by a band of scattered minute spines proximal to the array of enlarged spines (minute spines perhaps a bit once around the sulcus and on the asulcate side). Ornamentation of the body dominated by an array of enlarged spines (45 total); enlarged spines robust, hooked at the tip, larger ~ proximally and on the asulcate nae oradually decreasing in size distally. E ere Sides One the eae spermaticus bordered by a line of seven to nine much smaller spines (small spines one- fourth or one-fifth the size of the enlarged spines in the array). Three flounces on sulcate side of organ, increasing to five or more on the agi eat side: ehese grade into the apical freestand- ing ridges. All flounces/ridges with embed- Hee spinules. Proximal ‘ounce broader than the more distal ones and mainly membranous rather than fleshy. Some low and weak connections extend between adjacent distal flounces, especially on the asulcate side, and form poorly developed calyces. At the tip of the organ is a prominent tissue ridge (equivalent to the median bisecting ridge and associated oblique ridges in the everted organ) extending from the asulcate side across the apex and bisecting the apical boss, which lies at the tip of the sulcus spermaticus (Fig. 19C). The boss has a prominent ridge delimiting its border but is nude apart from this homies and the bisecting ridge; its morphology in the retracted organ is Seales: to that in the everted state. The bordering ridge has a few crenulations, which seem to Tye short embedded spinules. A few loose folds (presumably forming the oblique ridges in Bulletin of the Museum of Comparative Zoology, Vol. 160, No. 4 the everted organ) lie lateral and distal to the boss and bisecting ridge; tissue mainly lateral and proximal to the boss is nude. Sulcus spermaticus simple, in the dorso- lateral wall of organ, ending distally beneath the sulcate edge of the apical boss. Tip of the sulcus not expanded, the groove actually appearing to narrow slightly at its terminus. The sulcus lips also do not appear to diverge distally. There is some variation in the length of retracted hemipenes. Of nine retracted hemipenes examined superficially in_ situ, four ended between the suture of subcau- dals 7/8 and the suture of subcaudals 8/9: four ended between the middle of subcau- dal 9 and the middle of subcaudal 10; and one extended to the middle of subcaudal 11. The major retractor muscle appeared undi- vided in four specimens, but three others appeared to have a slight separation of muscle fibers at the insertion. Dendrophidion apharocybe Everted (LACM 148600, Heredia provy- ince, Costa Rica; Fig. 20). Hemipenis short, stout, with a slightly bulbous apex having an asulcate indentation, giving a somewhat cordate shape to the apex when viewed from the sulcate side. Total length about 21 mm. Length of base proximal to spine array on maleate side about 3.5 mm. Length of apex from proximalmost flounce to tip on sulcate side about 13 mm. Maximum width of organ 11.5 mm (across middle of apex). Sulcus spermaticus simple, centrolineal, with a distinctly flared tip. Base of organ proximal to the array of enlarged spines very short, ornamented with minute spines. The relatively unornamented base is followed by a central section with enlarged spines. Thirty-six spines form the array ‘and a line of very small spines lies on each side of the sulcus just proximal to the apex. Spines arranged in somewhat irregular oblique rows on the sulcate side but no particular arrangement on the asulcate side (appear scattered, irregular). Spines on the asulcate side are much larger than those on SPECIES IN THE DENDROPHIDION VINITOR COMPLEX ¢ Cadle bo bo Te) Figure 20. Everted hemipenis of Dendrophidion apharocybe (LACM 148600 from near the type locality, left hemipenis) in sulcate, asulcate, and lateral views. Note flared tip of the sulcus spermaticus and the nude apical tip strongly inclined toward the sulcate side (sulcate to the right in lateral view). the sulcate side. Some small spines are incorporated into the first flounce on the es). Flounces have a thick fleshy base and an outer membranous part. All flounces have sulcate side. On the sulcate side all spines are more or less the same size (perhaps slightly larger proximally). On the asulcate ace the Hata spines are slightly larger than the more proximal ones. Four flounces on the sulcate side broaden to about seven on the asulcate side. The proximal flounce on the sulcate side be- comes the 3rd flounce on the asulcate side (two proximal flounces added on_ the asulcate side). Flounces curve distad toward the asulcate side, reflecting the inclination of the apex toward the sulc ate side (F (Fig. 20, lateral view). No distinct calyces exce yt Sioa a couple of irregular ones distally on the right asulcate side (i.e., the right side aoaed looking toward the apnloake side; see Fig. 20, asulcate view). These calyces are asymmetrical (no comparable ones on the left side). Several other weak calyces present between the first pair of flounces on the asulcate side (weakly developed longitudinal walls between these two flounc- embedded spinules, the tips of which occupy weak scallops on their edges; spinules occupy mainly the membranous part but enter the fleshy part slightly. Scalloping becomes progressively lexe dis- tally and medially. Apex strongly inclined so that its distal surface fees toward the sulcate side (flounces extending distad much farther on the asuleate than the sulcate side). Central part of the apex occupied by a prominent bulge, which slopes gr adually to meet the aculcate edge of the apex but oe off sharply on he sulcate side (Fig. 20, lateral view). The sulcus ends near i a a side of the organ just beneath the bulge. Apex nude except for low rounded ridges that occupy the central bulge. These ridges have the same general pattern as the raentbiee nous ridges on the hemipenis of D. vinitor (see Aeue description). That is, they extend obliquely outward toward the asulcate side from a median axis; toward the sulcus the 294 Bulletin of the ridges converge on a point on the bulge deel to the tip of the sulcus. An indication of these ridges can be seen in Figure 20 (sulcate view) as oblique darker streaks alternating with wider whitish streaks just lateral to the central part of the apex. Small areas lateral to the ae tip are smooth and without ridges. Toward the asulcate edge of the apex some of the oblique ridges are connected by very low additional ridges, forming a series of indistinct reticulating structures ( (highly reduced calyces). Retracted (MVZ 217610, Heredia prov- ince, Costa Rica; Fig. 21). Hemipenis ex- tends to about the middle of subcaudal 7 Sulcus spermaticus simple, extending dis- tally in E REa Ree wall of the organ and ending short of the tip. At the tip of the sulcus are fine loose folds of tissue within the sulcus groove that presumably expand upon eversion to form the expanded tip of the sulcus in the everted organ. Extreme basal portion of hemipenis under subcaudal 1 appears to be nude. About three rows of ae rged spines begin at level of the proximal edge of subcaudal 2 and extend to the distal edge of subcaudal 3. About 34 total spines on the organ, with 10— 12 of these very large. These spines have a somewhat unusual flattened form with a tiny point at the tip. About seven or eight rows of calyces/ flounces located primarily on the asulcate side (medial and ventral sides of the retracted organ) begin at the level of the proximal edge of subcaudal 4. Only the proximal three flounces extend all the way to the sulcus spermaticus (more distal flounces stop short of the sulcus). Thus, the asulcate surface of the hemipenis is flounced almost to the tip of the organ, whereas on the sulcate side the flounces stop well short of the tip. Distal to the three proximal flounces adjacent to the medial side of the sulcus, the tissue of the hemipenis is smooth and nude and formed into low longitudinal folds. A few calyces are nestled within the distal longitudinal folds, mostly on the asulcate side. On the eel side of the sulcus are enlarged spines up to Museum of Comparative Zoology, Vol. 160, No. 4 Figure 21. Retracted left hemipenis of Dendrophidion aphar- ocybe (MVZ 217610 from near the type locality). Apex toward the top. Abbreviations: C, thick cords composed of closely appressed flounces on each side of the apical nude area; F, asulcate flounces; N, expanse of nude tissue on each side of the sulcus spermaticus that forms the nude sulcate face of the apex in the everted organ; ss, sulcus spermaticus. Note the lack of an apical boss and large area of apical nude tissue (compare Figs. 18B, 19C, D. vinitor). about the middle of subcaudal 5, distal to which the hemipenis is nude, smooth, and in low longitudinal folds. The calyces/flounces are tightly bound, pleated tissue, with adjacent flounces barely overlapping. ihe flounces are gathered into a pair of thick “cords,” one on each side of the apical nude tissue in the intact retracted organ (Fig. 21; in the figure, both cords are on one side of the nude tissue because of the position at which the hemipenis was slit). Between the cords is thinner tissue in which the flounces are individually discrete (the thinner tissue makes up the asulcate SPECIES flounces in the everted organ). A few weakly developed longitudinal connections are resent between the distal two or three ence The flounces seem tightly bound to one another because of oe longitudi- nal connections that occur periodically between adjacent ones. Their free edges are very slightly crenulate and have embed- ded spinules. The spinules extend to the edge of the flounces but do not seem to extend entirely to their bases. In the most proximal flounce, bunches of four to five spinules are separated by thinner, narrower tissue, such that the edge of this flounce is undulating. The bunches of four or five spinules grade almost imperceptibly into the enlarged spines more proximally. Variation and Remarks. A consistent and unusual hemipenial morphology is primary evidence that the populations from Hon- duras to Panama here referred to D. apharocybe comprise a single species. Sample sizes from different parts of this range of either fully eal organs or hemipenes everted sufficiently to see the configuration of the apex are 4 (Honduras), 2 (Nicaragua), 7 (Costa Rica), 1 (Panama). In addition, I studied the internal morphol- ogy of one retracted organ each from Nicaragua, Costa Rica, and Panama. There is little variation in basic morphology or ornamentation from throughout the range (particularly the distinctive inclination of the apex, asulcate flounces, and large apical nude expanse). The satel of enlarged spines in the spine array is variable (23-40) but shows no particular geographic trend. For example, two organs with 23 spines were from Honduras and Panama and two hemipenes with 28 and 40 spines were both from Nicaragua. In large part, the strong consistency of hemipenial morphology throughout the geographic range of D. apharocybe lends integrity to the concept of this species as distinct from D. vinitor and D. crybelum. Eight retracted hemipenes of Dendrophid- ion apharocybe from Honduras to Panama were superficially examined. Their distal endpoints were between the end of sub- SIN THE DENDROPHIDION VINITOR COMPLEX © Cadle 2295 caudal 6 and the proximal edge of sub- caudal 10, with four ending at or proximal to the middle of subcaudal 7. The major retractor muscle appeared WERE in two specimens and with a slight division in two others. Dendrophidion crybelum Everted (LACM 148599, holotype, Pun- tarenas province, Costa Rica; Fig. 22). Hemipenes of the holotype were the only fully everted organs studied in detail. One other was more or less fully everted (LACM 148590); it is essentially identical to the holotype but a few differences are noted. Hemipenis cylindrical, much longer than wide, and lacking a distinctly bullones apex. Total length about 21.5 mm; width about 6 mm (about 3.6X longer ‘han wide). Sulcus spermaticus, simple, centrolineal, with an expanded tip that broadens and_ then narrows, forming a tear drop—shaped ex- panse of nude tissue. The distal narr owing may be an artifact of the field eversion preparation, perhaps slight desiccation (it seems as if some slight folds between the distal lips might expand to form an openly expanded eelcns tip as in other Dendrophid- ion hemipenes). In LACM 148590 the sulcus tip appears to expand in normal fashion and then has a narrow distal nude extension. The distally divergent lips of the sulcus seemingly end short of the nude extension. Extreme base nude but scattered minute spines occur in a band around the organ proximal to the enlarged spines. A great number of enlarged spines dominate the ornamentation of the hemipenial body (about 90-95 in LACM 148599; about 70— 75 spines in LACM 148590). About 12 much smaller spines form a line on each side of the sulcus spermaticus for most of its length. Spines are larger proximally, gradu- ally decreasing in size distally, and some- what larger on the asulcate compared with the sulcate side. Distally, spines are followed by a short, nonbulbous apex ornamented with flounces 226 Bulletin of the Museum of Comparative Zoology, Vol. 160, No. 4 Figure 22. Everted hemipenis of Dendrophidion crybelum (LACM 148599, holotype), sulcate and asulcate views. (A) Detail of apical region showing closely packed apical ridges and flared tip of the sulcus spermaticus. (B) Asulcate side of apex showing flounces and rudimentary calyces (arrows). and calyces. Two flounces adjacent to the sulcus, broadening to four on the asulcate side (the most proximal on asulcate side is short, poorly developed, and asymmetrically placed in about the middle of the asulcate side). The two additional flounces on the asulcate side are proximal to the two on the sulcate side and form from connections among spines in a transverse row (the spines then becoming relatively robust spinules within the flauncent Flounces contain em- bedded spinules, which span their membra- nous parts and extend into the outer portion of their fleshy parts (these flounces mostly consist of membranous part, very little fleshy base). On the asuleate side between the distal pair of flounces are several poorly developed calyces with very low, underde- veloped longitudinal walls: these are mostly not visible except by lifting and separating the flounces forming their transverse alles These calyces are more fully developed on the right side of the apex than the left (Fig. 22B; similar to the pattern of devel- opment in D. apharocybe). Apex with many freestanding ridges containing embedded spinules and very little ade tissue (a small amount on the sulcate side around the tip of the sulcus spermaticus). The ridges generally have the same pattern as ‘he. apical ridges in D. vinitor (i.e., a median bisecting, ‘somewhat taller ridge from which less prominent ones extend obliquely toward the asulcate side). These ridges are much lower and seem fleshier aan the more membranous ridges in D. vinitor. They have slightly scalloped edges, and, because the apex is much narrower in D. crybelum, the ridges are more tightly packed. Toward the sulcate side, the ridges converge toward a point just distal to the: tip of the Slows spermaticus. No definitive apical boss such as that in D. vinitor is apparent. However, the sulcate tip of the median bisecting ridge has a thickened nodule and, on each side, a short Figure 23. same specimen. segment of calycular tissue with somewhat chickeied peripheral edges; the segment on the right side of the sulens is par tially fused with fie median bisecting ridge, whereas the one on the left is detached: Toward their asulcate ends, the segments are attached to one of the oblique apical ridges by a constriction in the tissue, which separates them somewhat from the main part of the oblique ridge. The general configuration of this area is reminiscent of the apical boss in D. vinitor and may be a less fully developed, but homologous structure. Both hemipenes OLUD: crybelum examined, LACM 148590 and LACM 148599, have similar configura- tions in this area. Retracted (LACM 148608, Topotype, Puntarenas province, Costa Rica; Fig. 23). Hemipenis extends to the suture between subcaudals 9 and 10. No division of the major retractor was detected. The base of the organ is nude and followed by a very short ornamented with minute spines. If egion (A) Retracted right hemipenis of Dendrophidion crybelum (LACM 148608). Hemipenial body dominated by %, 4 4 a (B) Detail of the apex of the approximately 92 enlarged spines; a series of much smaller spines lines each side of sulcus spermaticus. Individual spines are robust, very long, with a small hooked spike at tip. The spinose region is followed distally by flounces/calyces; they are more calyxlike on the asulcate side (very deep calyces here), where they extend onto the apical region. All flounces/calyces have embedded spinule Si Sulcus spermaticus simple and in the dorsolateral wall of organ; seemingly not flared at tip, but there is a long very fine ridge of tissue extending along the midline etude the sulcus lips at the tip of the organ (may expand upon eversion to form frre d tip). Sulcus ends beneath the tips of a pair of membranous ridges (reduced calyx walls) at the sulcate edge of the bisecting apical ridge (the bisecting ridge extends toward the sulcate side and dw ides forming a pair of flaps resembling the apical boss in D. vinitor; these two flaps are connected by a low transverse ridge extending between bo bo them). The entire sulcus spermaticus be- tween the origin of the enlarged spines and the flounces/calyces is bordered on each side by a line of closely spaced small spines. Three other retracted hemipenes of D. crybelum extended to the middle and end of subcaudal 9 and to the proximal edge of subcaudal 10. DISCUSSION Species Groups and Relationships The recognition of three cryptic species within the D. vinitor complex, a detailed understanding of their hemipenial morphol- ogy, and comparative data on other species cee ee, provide a framework for understanding some aspects of Dendrophid- ion systematics and biogeography. I will elsewhere present detailed hemipenial de- scriptions of other species of Dendrophid- ion and, for present purposes, mention only a few salient features necessary for under- standing hemipenes of the D. vinitor complex. Characters suggested here as unifying groups of species are provisional apomorphies pending broader comparative work. One hemipenial character has been used to define the D. dendrophis species group (D. dendrophis, D. nuchale auctorum, and the three species of the D. vinitor complex as defined herein): the presence of “basal spines or elongate hooked villi” or “large basal hooks” (Lieb, 1988; Savage, 2002: 654, 656). As pointed out in the introduction to Dendrophidion hemipenes, none of the enlarged spines in Dendrophidion hemi- penes are truly basal. However, details of hemipenial morphology in the D. vinitor complex allow some refinement of hemi- penial characterizations of this group. With- in the D. dendrophis group, only D. dendrophis and D. nuchale auctorum have a pair of sulcate spines at the proximal edge of the spine array that are enormously enlarged beyond the size of most spines in the array (these two species also have up to two other moderately enlarged spines). These greatly enlarged spines correspond 8 Bulletin of the Museum of Comparative Zoology, Vol. 160, No. 4 to the “basal spines” of other authors and are much larger than any other spines on the hemipenis of these two species (unpub- lished data; see Stuart, 1932). In contrast, other than a size asymmetry noted below, species of the D. vinitor complex have no spines enormously en- larged beyond the majority of “enlarged” spines on the hemipenial body, a detail noted by Smith (1941) in his description of D. vinitor. Thus, only part of the D. dendrophis species group is characterized by greatly enlarged hemipenial spines. Moreover, although hemipenial spines in the D. dendrophis species group are gener- ally larger than those in the D. percarinatum group, this generalization does not hold for all species. For example, hemipenial spines in some specimens of D. percarindatum are approximately the same relative size as those in D. vinitor. While there is great interspecific variation in the absolute sizes of spines in Dendrophidion, all species have an array containing “enlarged” spines but the variation in spine size does not clearly distinguish the two species groups. Until more thorough comparative studies of hemipenial morphology within Dendrophid- ion demonstrate consistent differences between the D. dendrophis and D. percar- inatum groups, statements about differenc- es in relative spine enlargement need qualification to account for intragroup variation. The point of dorsocaudal reduc- tion and the prominence of keels on the dorsal scales still provide convenient char- acters for distinguishing the two species groups of Dendrophidion, although whether this convenience reflects phylogeny ulti- mately should be reexamined. Is the D. vinitor complex monophyletic? Within the D. dendrophis group, species of the D. vinitor complex share several hemi- penial characters compared with D. den- drophis and D. nuchale auctorum (charac- ters in parentheses): (1) Calyxlike structures reduced to flounces, especially on the asulcate side (Figs. 18, 20, 22) (calyces reduced, but fully oe cuplike structures present on both the sulcate and asulcate SPECIES IN THE sides); (2) increased number of flounces on the asulcate side compared with the sulcate side (flounces/calyces not increased on asulcate side, perhaps reduced); (3) enlarge- ment of spines ie the battery conspic- uously asymmetrical, with asulcate spines noticeably lar ger than sulcate spines, seen especially we alt in Figure 20, lateral view (no conspicuous general asymmetry in spine enlargement ES pair of enormously enlarged spines on the sulcate side and often another Be toward the asulcate side]); (4) series of ie eestanding apical ridges with a primarily oblique orientation from a central axis, the median bisecting ridge (freestanding ridges absent; calyces palo low reticulating ridges may be present). As indicated in the detailed hemipenial descriptions, the free- standing ridges in D. apharocybe are reduced to only low rounded ridges, but the pattern of oblique orientation is evident even in such a reduced form (Fig. 20, sulcate view). These characters, plus the absence of several characters uniquely shared by D. dendrophis and D. nuchale auctorum (e.g., greatly enlarged sulcate spines, a very regular distal row of enlarged spines within the spine array) can be taken as provisional evidence for the monophyly of the D. vinitor complex. Among the three species of the D. vinitor complex, the hemipenis of D. crybelum stands out because of its unique cylindrical form and great number of enlarged spines. My current assessment is that both of these characters are autapomorphies of D. crybe- lum because the short, bulbous hemipenial form and fewer spines shared by D. vinitor and D. apharocybe are more widespread within Dendrophidion, including the other members of the D. dendrophis group, D. dendrophis and D. nuchale auctorum. On the other hand, D. apharocybe and D. crybelum share several hemipenial charac- ters relative to D. vinitor: (1) reductions in the apical free-standing ridges (more com- pletely reduced in D. apharoc ybe than in D. crybelum); (2) poorly developed asulcate calyces that are asymmetrically placed on the right asulcate side; (3) a more strongly DENDROPHIDION VINITOR COMPLEX ¢ Cadle 229 flared tip to the sulcus spermaticus. In addition, D. apharocybe and D. crybelum are more similar in having wider pale bands that are usually dicumer the entire body length and which tend to form pale ocelli on ie, posterior body. These characters pro- vide evidence suggesting that D. apharo- cybe and D. crybelum are more closely related within the D. vinitor complex. The apical boss of D. vinitor and the nude and strongly inclined apex of D. apharocybe are then seen as autapomorphies of these two species. Apart from hemipenial differences, the three species of the D. vinitor complex are exceedingly similar in scutellation and other features cenerally useful for distinguishing snake species (T Table 1). Few of élid inter- specific comparisons of ventrals, subcau- dals, relative tail lengths, maxillary tooth counts, or the point af dorsocaudal reduc- tion were statistically significant. Even in cases in which means were significantly different, the absolute differences were minimal and character ranges overlap sub- stantially, sometimes completely, For exam- ple, mean subcaudal number is significantly different between D. apharoc ybe. and each of the other two species, but the magnitude of the difference between means in each comparison was only 2.7 or 3.4 (sexes combined). In general, the means for intraspecific differences between males and females were more substantial and of greater statistical significance than interspe- eine comparisons by sex. The lack of substantive differences in these systematic characters emphasizes the cryptic nature of these species and lends credence to their hypothesized close relationship within Den- drophidion. Biogeography Considerable progress has been made in understanding biogeographic patterns with- in Middle America in the last decade. Many of the patterns have been elucidated by comprehensive understanding of phy logeo- graphic patterns shown by Central Ameri- 230 can snakes, which in many cases have confirmed or refined prior assessments based on patterns of endemism and diver- sity (reviewed by Daza et al., 2010). Daza et al. (2010) correlated phylogeographic pat- terns with the history of major tectonic units and concluded that vicariance associated with tectonic events was a predominant agent causing speciation in Middle Ameri- can snake lineages. Two critical geologic events neat ened with divergences in snakes are the Motagua—Polochic fault zone (suture of the Maya ane Chortis blocks) in present- day Guatemala (3-8 million years ago [MYA]) and the uplift of the Cordillera ae Talamanca in Costa Rica and Panama (2.5— 3.9 MYA). Other important events are associated with the Isthmus of Tehuante- pec, the Nicaraguan Depression, and the primary Middle— Sauk American transition; however, the temporal span of divergences in these areas are broader than the firs two, perhaps because geological events in these areas had differential effects in different lineages. The distributions and relationships of species in the D. vinitor complex proposed here fit nicely into this paradigm. Dendro- phidion vinitor, the proposed eee taxon to D. apharocybe-D. crybelum, is distributed entirely north of the Motagua—Polochic fault zone, and its isolation hice most likely reflects the estimated late Miocene—Plio- cene divergence in other snake taxa across this zone (Daza et al., 2010). The more recent divergence between D. apharoc ybe and D. cr ybelum accompanied the vicari- ance of Atlantic and Pacific slope faunas resulting from elevation of the Cordillera Ratanyriea in the late Pliocene. Chan et al. (2011: 328-331), in an elegant analysis of relationships among populations of the hylid frog Dendrosophus ebraccatus across the Talamanca disjunction in Costa Rica and Panama, found the most str ongly supported model to be that suggested Shere For 1D apharocybe—crybelum. “At the level of the species concepts developed here, no differ- entiation has resulted from the other major Central American tectonic events outlined Bulletin of the Museum of Comparative Zoology, Vol. 160, No. 4 by Daza et al. (2010). Nonetheless, finer scale genetic studies could show some intraspecific differentiation due to these last events. ACKNOWLEDGMENTS I thank the following collection personnel for their help with loans and other assis- tance: Darrel Frost, David Kizirian, and Charles W. Myers (AMNH); Edward B. Daeschler and Edward Gilmore (ANSP): Colin J. McCarthy (BMNH); Lauren Scheinberg and Jens Vindum (CAS); Ste- phen P. Rogers (CM); Alan Resetar and Harold K. Voris (FMNH): Rafe Brown, Andrew Campbell, and Linda Trueb (KU); Neftali Camacho and Christine Thacker (LACM); etn a Austin and Eric Rittmeyer (LSUMZ); James Hanken, Jona- than Losos, Joe Martinez, and José P. Rosado (MCZ); Christopher Conroy, Mi- chelle Koo, Carol Spencer, and David B. Wake (MVZ); Toby Hibbitts (TCWC); George Bradley and Peter Reinthal (UAZ): Mariko Kageyama and Christy McCain (UCM): Gerardo Chaves (UCR): Kenneth Krysko (UF); Christopher Phillips and Dan Wylie (UIMNH): Ronald Nussbaum and Gregory E. Schneider (UMMZ): Steve Gotte, Roy W. McDiarmid, James Poindex- ter, Kenneth Tighe, and George R. Zu (USN M); and Jonathan Campbell and Carl J. Franklin (UTACV). Several individuals provided special help that enhanced the quality of this work. Julian Lee provided information on the Belize specimen of Dendrophidion vinitor. Kenneth Tighe provided photographs of the holotype of D. vinitor, and Steve Gotte checked some details on it. The herpetology department of the California Academy of Sciences (CAS) provided work space during completion of this work; thanks to Jens Vindum for patience and tolerance. Re- becca Morin of the CAS library was helpful in tracking down references. Jay M. Savage shared He knowledge of Costa Rican Dendrophidion, anewered many queries in detail, and provided critical information and SPECIES IN THE insights. Craig Guyer, Toby Hibbitts, and Bean ean ‘exchanged information, photographs, and obse srvations on Dendro- phidion they collected. Gerardo Chaves provided crucial data on, and photographs of, specimens in the UCR; these helped resolve some enigmatic localities and iden- tities. I am gr atcrul to Roy W. McDiarmid for aiternmation on D. crybelum and for copies of his color slides, which were scanned and checked by James Poindexter. Charles W. Myers provided copies of his field notes and color slides of Panamanian Dendrophidion and saved me from a nomen deforme and other calamities. Myers and McDiarmid or aciously permitted the use of their photographs in Figures 9 and 15. Colin J. McCarthy provided copies of BMNH ledger pages relevant to Dendro- phidion. Robert M. Timm (KU) verified data for specimens from the original field notes of Walter Dalquest. I am grateful to all for their help. For comments on the manuscript I thank Roy W. McDiarmid, Charles W. Myers, and Jay M. Savage; they didn’t always agree with my presentation but their comments sharpened my thinking and improved the manuscript substantially. APPENDIX 1. SPECIMENS EXAMINED AND LITERATURE RECORDS FOR DENDROPHIDION VINITOR SMITH Museum abbreviations used throughout are the following: AMNH—American Museum of Natural History (New York). ANSP—Academy of Natural Sciences of Philadelphia. BMNH— British Museum (Natural History) (London). CM—Carnegie Museum (Pittsburgh). Ku= University of Kansas “Museum of Natural History (Lawrence). LACM— Natural History Museum of Los Angeles County (California). Museum of Comparative Zoology (Cambridge). MNHN—Muséum National d'Histoire Naturelle (Par- is). TCWC—Texas Cooperative Wildlife Collection, Texas A&M University (College Station). UAZ— University of Arizona Museum of Natural History (Tucson). UCM—University of Colorado Museum of Natural History (Boulder). UCR—Universidad de Costa Rica Museo de Zoologia (San José). UF—Florida Museum of Natural History, University of Florida (Gainesville). UIMNH— University of Illinois Museum of Natural History (Urbana). UMMZ—University of Michigan Museum of Zoology (Ann Arbor). USNM— National Museum of Natural History, Smithsonian Institution (Washington, DC). UTACV—University of DENDROPHIDION VINITOR COMPLEX ¢ Cadle 931 Texas at Arlington Collection of Vertebrates (Arling- ton). UTE P—U Iniversity of Texas at E] Paso Centen- nial Museum (EI Paso). Bracketed data associated with localities here and elsewhere in the text are inferences derived from sources other than the original data associated with specimens as recorded in literature, museum or collectors’ catalogues, or specimen labels. Some museum databases (e.g., USNM) unfortunately do not distinguish “original” and “inferred” locality information in this era of databasing and georeferen- cing, and I have of necessity in most cases taken these sources at face value. Belize: Toledo: Little Quartz Ridge [summit, 940— 1,035 m], KU 300784 (Fig. 6). Guatemala: “Vera Paz” (Duméril et al. 1870-1909: 732: see comments under Distribution). [Alta Verapaz|: between Coban and Lanquin, BMNH 64.1.26.21 (Stafford, 2003; specimen not seen: see comments under Distribution). [zabal: Sierra de Santa Cruz, Semococh, 8 km W Finca Semuc headquarters, UTACV 22155. Sierra de Santa Cruz, Finca Semuc, Serujijé Mountain, 15°37'09°N, 89°22'05"W, 600 m, UTACV 22755. Petén: “Peten,” no specific locality (Duméril et al. 1854: 208, = MNHN 7353 fide Lieb, 1988: 164, Fig. 1; specimen not seen); Duméril et al. 1870-1909: 732; see comments under Distribution). Piedras Negras, USNM 110662 (holotype; Fig. 1). Mexico: No specific locality, USNM 7099 (paratype of D. vinitor Smith, specimen not seen; Smith, 1941). Chiapas: Presa Malpaso (Alvarez del Toro, 1972: 142, 1982: 191). 26 km N Ocozocuautla (Johnson et al. “1976” [1977], presum- ably UTEP 6038 cited by Lieb, 1985; specimen not seen). Approximately 8km S$ Solosuchiapa, camp on Rio Teapa, ca. 400 ft. [120 m], UAZ 24161. Oaxaca: Rio Chicapa near El Atravesado, 1,600 ft. [485 m], AMNH R-66845. Donaji, Mije [province], UCM 39911-12, 44481. La Gloria (north of Niltepec), 1,500 ft. [457 m], FMNH 126554—-55. La Gloria, UIMNH 17632, 35546-— 47, 37094-95, 37165; TCWC 12797. La Gloria-Cerro Azul, UIMNH 35548. Rio Negro (Grijalva), Juchitan [province], UCM 41162. Bapisee: Teapa, USNM 46589 (paratype of D. vinitor Smith, 1941; specimen not seen). Veracruz: Bastonal (S mi E of Cuatzelapa, Lago Catemaco), CM 4147S. Forest at Cascapel, upper Uzpanapa river, Isthmus of Tehuantepec, BMNH 1936.6.6.8 (Stafford, 2003; specimen not seen); see Appendix 2 for comments. Near Coyame, 1,400 ft. [427 m], UMMZ 111450. Coyame, 9-10 mi E Catemaco, UIMNH 39154, 82444. Isla (Perez-Higar- eda and Smith, 1991: 31). 60 km SE Jesus Carranza, 450 ft. [137 m], KU 23965. 25 km SE Jesus Carranza, 250 ft. [76 m], KU 27564. Las Minas (Perez-Higareda and Smith, 1991: 31). Los Tuxtlas (Perez-Higar eda and Smith, 1991: 31). Motzorongo (Dugés 1892). Tese- choacan (Perez-Higareda and Smith, 1991: 31). Uxpa- napa (Perez-Higareda and Smith, 1991: 31). SE slope Volcan San Martin, approximately 2,600 ft. [793 ml], UMMZ 121145. Volcan San Martin near base, UMMZ 122767. Voleén San Martin, UIMNH 33862. Volcan San Martin, El Tular Station, UIMNH 35472. bo W bo APPENDIX 2. GAZETTEER Except where otherwise stated, coordinates for place names are from the National Geospatial-Intelligence Agency (NGA) online gazetteer (GEOnet Names Server): http://earth-info.nga.mil/gns/html/. A few co- ordinates were taken from Google Earth (GE). Many specimens of Dendrophidion vinitor were obtained by Thomas MacDougall, field biologist extraordinaire, whose collections significantly advanced many fields of Mexican zoology and botany (Stix, 1975). Goodwin (1969) reported MacDougall’s mammal collections, and I quote from his brief characterizations of various Oaxacan localities before human alteration in the last half century; Goodwin had the opportunity to work directly with MacDougall’s notes (Root, 1975). The three northern countries harboring D. vinitor are listed first, followed by an alphabetical listing of other Middle American countries with the other two members of the D. vinitor complex. BELIZE AND GUATEMALA (Dendrophidion vinitor) Coban to Lanquin (Guatemala: Alta Verapaz). Approximately 15°34’N, 90°09’W (coordinates arbitrarily about midway along the route between the two places). Coban (approximately 1,320 m elevation) and Lanquin (approximately 335 m) lie about 43 airline kilometers apart in the Rio Cahabon basin (north side of the Sierra de Santa Cruz and near Finca Semuc, whence come the only recent Guatemalan records of D. vinitor). Finca Semuc, Serujija Mountain, Sierra de Santa Cruz, 600 m (Guatemala: Izabal). 15°37'09’N, 89°22’05"W (coordinates from 1:50,000 topo- graphic map, Instituto Geografico Nacional, Guatemala; provided by the collector of UTACV 22755, Eric N= Smith): Little Quartz Ridge (Belize: Toledo). 16°24’N, 89°05'’W. The single known Belize specimen of D. vinitor was collected at the summit of the ridge, given as “940-1035 m elevation” (Meer- man and Lee, 2003, table 1). The environment of the area is described in Meerman and Matola (2003). Piedras Negras (Guatemala: Petén). 17°11’N, 91°15’W. Type locality of D. vinitor Smith. Semococh, 8 km W Finca Semuc headquarters, Sierra de Santa Cruz (Guatemala: Izabal). This locality is seemingly none of several “Semococh” indexed by NGA, most of which are in Alta Verapaz department. Presumably close to coor- dinates for Finca Semuc listed above. MEXICO (Dendrophidion vinitor) Bastonal (8 mi E of Cuatzelapa, Lago Catemaco) (Veracruz). 18°19’'N, 94°54’W. Cascapel, upper Uzpanapa river, Isthmus of Te- huantepec (Veracruz). Not located. The spelling “Cascapel” seems clear in BMNH ledgers, as listed by Stafford (2003). This may be a Bulletin of the Museum of Comparative Zoology, Vol. 160, No. 4 transcription error for “Cascajal” (or El Cascajal), a village on the upper Rio Uxpanapa in extreme southeastern Veracruz near the Oaxaca border (17°37'N, 94°08'W). Cerro Azul (Oaxaca). See La Gloria. Coyame, 9-10 mi E Catemaco (Veracruz). 18°26’N, 95°00’W. (and Near Coyame, 1,400 ft. [427 m]). Donaji, Mije [province] (Oaxaca; MacDougall). 17°13'53"N, 95°03'40”"W. 90 m elevation in “rainforest” fide Duellman (1960: 34). El Atravesado (Oaxaca). See Rio Chicapa. El Tular (Veracruz). See Volcan San Martin. Isla (Veracruz). At least three localities with this name are in Veracruz: 18°36’N, 96°09’W and 18°02'’N, 95°31'30”"W (NGA); and “La Isla”: 17°25'N, 94°01'W (Esparza-Torres, undated). Jesus Carranza. 60 km SE, 450 ft. [137 m] (KU 23965); and 25 km SE, 250 ft. [76 m] (KU 27564) (both Veracruz). These are localities of Walter W. Dalquest (see Hall and Dalquest, 1963). Original KU catalogue data for the “60 km” locality gave the compass direction as “SW” from Jesus Carranza, which would place the locality in Oaxaca rather than Veracruz. However, accord- ing to Robert M. Timm, Dalquest’s field catalogue in the Mammal Division at KU has the locality corrected by hand (presumably by Dalquest himself) to “60 km SE Jesus Carranza.” Accord- ing to the itinerary given in Hall and Dalquest (1963: 177), Dalquest worked out of a village (Zapotal) on the Rio Coatzacoalcos during the period when KU 23965 was collected. His field notes for mammals collected at the same time indicate that he traveled upriver via the Rio Chalchijapa and then Rio Solosuchil. Coordi- nates from the gazetteer of Hall and Dalquest (1963: 184) are: junction of the Rio Chalchijapa with the Rio Coatzacoalcos (approximately 17°27'N, 94°50’W); junction of the Rio Solosuchil with the Rio Chalchijapa (approximately 17°23'N, 94°47'W); and “Rio Solosuchil” (17°14'N, 94°28'W). Coordinates for the river junctions are quite accurate, as verifiable using Google Earth. The “Rio Solosuchil” locale is in southern Veracruz at its headwaters near the Oaxaca border, and close to 60 airline kilometers from Jesus Carranza in an ESE direction (perusal of Hall and Dalquest [1963] suggests they used a fairly loose interpretation of compass directions). This is perhaps the approximate location of KU 23965 (it is not clear whether Dalquest used airline or river distances). In any case, KU 23965 was apparently collected along the Rio Solosu- chil in southern Veracruz. La Gloria (north of Niltepec), 1,500 ft. [457 ml] (Oaxaca; MacDougall). 16°47'17"N, 94°36’40"W. Located 7 mi S of Santa Maria Chimalapa in Juchitan district (Goodwin, 1969: 259). Usually stated simply “La Gloria,” this is the origin of more specimens of D. vinitorin U.S. collections than any other single locality (rivaled only by Volcan San SPECIES IN THE DENDROPHIDION VINITOR COMPLEX ¢ Cadle 22. Martin in Veracruz). Cerro Azul, associated with La Gloria in one specimen locality, is a local name for high parts of the Sierra Madre of Oaxaca— Chiapas, 25 mi NW of Santo Domingo Zanatepec, also in Juchitan district (Goodwin, 1969: 257); this is not the same “Cerro Azul” indexed by NGA. All of the La Gloria specimens were obtained by Thomas MacDougall. Habitats at La Gloria include “coffee plantations, milpas, rainforest,” whereas Cerro Azul harbors “cloud forest” (Goodwin, 1969: PAST AN Las Minas (Veracruz). 19°42’N, 97°07’W. Los Tuxtlas (Veracruz). 18°30'N, 95°10’W. Motzorongo (Veracruz) 18°39’N, 96°44’W. Gold- man (1951: 277) gives the elevation as 800 ft. (244 m) but in steeply dissected country with hills rising locally to 1,500 ft. (457 m). Goldman (1951: 277-278, 316) briefly describes the vegetation (“fairly uniformly covered with evergreen forest’), which he classified as “Humid Lower Tropical Zone.” Ocozocuautla, 26 km N (Chiapas). About 16°55’N, 93°27'W. Presa Malpaso (Chiapas). 17°08’N, 93°30’W. Also known as Presa Netzahualcoyotl. Rio Chicapa near El Atravesado, 1,600 ft. [488 m] (Oaxaca; MacDougall). Pacific drainage. El Atra- vesado (16°41'N, 94°35’W) is a village near a mountain of the same name. This is an inland locality on the Rio Chicapa, not the coastal locality between Union Hidalgo and the river's mouth in Laguna Superior indexed by Goodwin (1969: 262). Cerro Atravesado is a “large flat-topped hill 18 mi S of Santa Maria Chimalapa [with] open pine stands, grass and rocks, patches of “cloud forest” at the north end” (Goodwin, 1969: 257). Rio Negro (Grijalva), Juchitan [district] (Oaxaca; MacDougall). Atlantic drainage. Goodwin (1969: 262) gives two MacDougall localities with this designation: “[Rio Grijalva system at] Junction of Rio Negro with Rio ‘Porta Monedo’ [sic] near Chiapas border 29 mi north of Tapanatepec,” and 5 mi west of this junction. The “Rio Porta Moneda” is in the headwaters of the Rio Encajonado system (Goodwin, 1969: 262). Co- ordinates from NGA: Rio Negro, 16°49’36’N, 94°01'25”"W; Rio Portamonedas, 16°41'15’N, 94°08'22"W. Goodwin (1969) notes “heavy shade of trees in river flat” for the river junction and “many tree ferns ... rainforest [and] tapir trails” for the nearby camp. Solosuchiapa, approximately 8 km S; camp on Rio Teapa, ca. 400 ft. [122 m] (Chiapas). 17°24'N, 93°01’W. Teapa (Tabasco). 17°33’N, 92°57’W. 800 ft. [244 m] in the Humid Lower Tropical Zone fide Goldman (1951: 257-259). Tesechoacan (Veracruz). 18°08’N, 95°40’W. Uxpanapa (Veracruz). A name associated with several locations in southeastern Veracruz, including a major river and at least three towns. A OO Volcan San Martin (Veracruz). 18°33’N, 95'12’W. A number of localities on the slopes of this volcano (near the base; SE slope, approximate- ly 2,600 ft. [793 m]). The volcano is now part of the Los Tuxtlas Biosphere Reserve. Several places in southern Veracruz have the name El Tular, El Tular station being on the southwest- ern flank of the Volcan San Martin within the boundaries of the reserve (approximately 18°30’N, 95°13’W, 600 m). Goldman (1951: 283) stated that virgin [rainforest] covered the mountain and that “from the sloping plain the heavy forest, full of small palms, vines, and other undergrowth up to about 4,800 feet changed but little.” COSTA RICA (Dendrophidion apharocybe and D. crybelum) Cacao Biological Station, 729-1,528 m (Guana- caste). 10°56’N, 85 28’W. Part of the Area de Conservacion Guanacaste, the station is on the southwestern slope of Volcan Cacao at about 1,000 m. Montane rainforest to cloud forest on the upper slopes, transitioning to dry forest on the western lower slopes. Carara National Park (Puntarenas and San José). 09°46'30"N, 84°36'25’W. Near the Pacific coast in northern Puntarenas and extreme western San Jose provinces. Origin of two specimens errone- ously referred to “Dendrophidion vinitor’ (Laur- encio and Malone, 2009) indicated on the distri- bution map of Savage (2002: 656). These specimens are D. percarinatum. See Distribution in the D. apharocybe species account and Figure 17. Cajon, N bank of Rio Térraba (Puntarenas). 08°56'30"N, 83°20’W, about 80 m. Source of an apparently erroneous lowland Pacific locality for “D. vinitor’ indicated by Savage (2002: 656). See text discussion under Distribution in species account for D. apharocybe. Finca La Selva (Heredia). 10°26’N, 83°59’W, 35- 137 m (McDade and Hartshorn, 1994). Now the La Selva Biological Station of the Organization for Tropical Studies. Type locality of D. apharocybe. Finca Las Alturas, 1,330 m (Puntarenas). 08°57’N, 82°50'W. Presently the Las Alturas Biological Station operated by the Organization for Tropical Studies. Finca Las Cruces, near San Vito de Java, 4 km S San Vito, 1,200 m (Puntarenas). 08°47'35’N, 82°57'30"W (Wake et al., 2007: 557). Presently the Las Cruces Biological Station operated by the Organization for Tropical Studies. Type locality of D. crybelum. Finca Mellizas, 14 km ENE La Union near Panama border (Puntarenas). 08°53’08"N, 82°46’42’”W, approximately 1,310 m (GE). Finca Loma Linda, 2 km SSW Canas Gordas, 1,170 m (Puntarenas). 08°43.3’N, 82°54.3'W (Wake et al., 2007). 234 Guapiles (Limon). 10°17'N, 83°46’W. La Selva (Heredia). See Finca La Selva. Mt. [Cerro] Mirador near Suretka (Limon). 09°36’N, 82°57'W. Pandora, 50 m (Limon). 09°44’N, 82°58'W. Pavones, ca. 2.5 km N, 700 m (Cartago). Near Turrialba. 09°57'N, 083°37'W. Poco Sol de La Tigre, 540 m (Alajuela). 10°22’'N, 84°37'W. Puerto Viejo de Sarapiqui, 10 km WSW (Heredia). 10°28'N, 84°01’W. San Clemente, 7 km NW Penshurst (Limon). 09°50’N, 82°56’W. Silencio, 875-940 m (Guanacaste). 84°54'W. Suretka (Limon). 9°34’N, 82°56’W. See also Mt. Mirador. Rio Puerto Viejo near junction with Rio Sarapiqui (Heredia). 10°28’N, 84°02’W. Zona Protectora, La Selva, trail from 1,000 m camp to 1,500 m camp, 990 m (Heredia). Approximate- ly 10°17'N, 84°04'W. See Pringle et al. (1984). HONDURAS (Dendrophidion apharocybe) 10°28'N, Bodega de Rio Tapalwas, 190 m (Gracias a Dios). 14°55'39"N, 84°32’02”W. About 20 km NW Rus Rus. Wilson et al. (2003: 18); McCranie et al. (2006: 265). Cano Awalwas (camp), 100 m (Gracias a Dios). 14°49’N, 84°52’W. Wilson et al. (2003: 18). McCranie et al. (2006: 261). Crique Ibantara, 70 m (Gracias a Dios). 14°47’'N, 84°27'W. A tributary of the Rio Rus Rus. Wilson et al. (2003: 18); McCranie et al. (2006: 263). Crique Wahatingni, near (Gracias a Dios). Tributary of Rio Tapalwas, 200 m. McCranie et al. (2006: 265). Crique Yulpruan, near; 140 m (Gracias a Dios). 14°54'N, 84°31’W. Tributary of Rio Tapalwas, 200 m. McCranie et al. (2006: 266). Hiltara Kiamp, 150 m (Gracias a Dios). 14°57’N, 84°40’W. Along the upper portion of the Rio Warunta (McCranie, 2011: 615). Las Marias, 20 m (Gracias a Dios). 15°40’N, 84°53'’W (GE). A village along the Rio Platano (McCranie, 2011: 620). Figures transposed in latitude given by McCranie (2011: 620; 15°04’N). Planes de San Esteban, 1,100 m (Olancho) 15°05’N, 85°42’W. Northern slope of the Sierra de Agalta (McCranie, 2011: 609). Sachin Tingni, 150 m (Gracias a Dios). 14°57’N, 84°40’W. Tributary of Rio Warunta (McCranie, 20112 633). Warunta Tingni Kiamp, 150 m (Gracias a Dios). 14°55'20"N, 84°41'28”W. Campsite along upper portion of Rio Warunta (McCranie et al., 2006: 266). NICARAGUA (Dendrophidion apharocybe) Cara de Mono (Atlantico Sur), ca. 120 m. 12°07'N, 84°28'W. Bulletin of the Museum of Comparative Zoology, Vol. 160, No. 4 Hacienda La Cumplida, 19 km N of Matagalpa, 2,500 ft. [762 m] (Matagalpa). 13°00'N, 85°51’W. Matagalpa (Matagalpa). 12°55’N, 85°55’W. Musawas, Waspuc River (Atlantico Norte). 14°08'59.6"N, 84°42'18.4’W. Recero. See Recreo. Recreo, Rio Mico (Atlantico Sur). About 50 m. 12°10’N, 84°19’W. Gaige et al. (1937; see especially pp. 2-3) referred to this place as “Recero” in text and an accompanying map from information provided by Morrow J. Allen, the collector of UMMZ specimens, based in turn ona “U.S. Marine Corps survey map ca. 1930.” The field notebooks of Allen at UMMZ also refer to the place as Recero (Gregory E. Schneider, personal communication, January 2011). No independent sources | have seen refer to the place as “Recero,” which seems to be in the same location as “Recreo” or “El Recreo” in gazetteers and on modern maps (e.g., NGA and Google Earth). “Recero” appears in no geographic reference on Nicaragua that | consulted, including maps and gazetteers contemporary with and earlier than Gaige et al. (1937). However, the place is listed as “Recero” in several other taxonomic works (e.g., Smith [1941: 74, 76], Smith and Taylor [1950: 320], Dunn and Stuart [1951: 58]). All of these seem ultimately to trace back to Allen’s material or to Gaige et al. (1937). Rio Mico, 10 mi above Recreo (Atlantico Sur). Approximately 12°07'N, 84°28’W. See locality notes for Recreo above. Rio San Juan (Rio San Juan). River along the frontier between eastern Nicaragua and Costa Rica. Santo Domingo, Chontales Mines, 2,000 ft. (610 m) (Chontales). 12°16’N, 85°05’W. The locale was made famous by Belt (1874), who described the environment and geology of the area. PANAMA (Dendrophidion apharocybe) Almirante, 10 m (Bocas del Toro). 09°18’N, 82°24'W. 11 km NW Almirante, 600 ft. [183 m] (Bocas del Toro); about 09°21’'N, 82°28’W. Cerro Arizona above Alto de Piedra, North of Santa Fe (Veraguas). Not located. Alto de Piedra is a small village about 3 airline kilometers NW (not strictly N as in the original locality data) of the town of Santa Fe (8°31'N, 81°04’W). Cerro Azul region, Rio Piedra (Panama). 09°13’N, 79°18'W (Fairchild and Handley, 1966). Cerro Campana, 900-950 m (Panama). 08°41'N, 79°56’W (Fairchild and Handley, 1966). Myers (1969: 28) described the area, paraphrased here: Small area of cloud forest above 870 m. Forest of moderate height, with few large trees, many small trees, a scattering of tree ferns and small palms, a few stilt palms. Dense cover of bushes, herbs, ferns. Once abundant tree and ground bromeliads and other epiphytes have been reduced. Disturbed vegetation on the top. SPECIES IN THE DENDROPHIDION VINITOR COMPLEX * Cadle Cerro Delgadito, 2-4 mi W Santa Fe (Veraguas). Approximately 08°30’N, 81°07’W (GE). Cerro Mali (Darién). 08°07'N, 77°14'W (Fairchild and Handley, 1966). According to Myers (1969: 25) Cerro Mali is about 1,410 m elevation and southeast of Cerro Tacarcuna and the headwa- ters of the Rio Pucro (= Rio Pucuro; see Myers and Lynch, 1997, figs. 1 and 2). It lies at the southeastern end of the Serrania del Darién, a ridge separating the Pacific-draining Rio Tuira system from the Atlantic lowlands of eastern Panama and northern Colombia (the internation- al border follows the continental divide along this ridge). Myers (1969: 24-25) described the general topography and environment. Additional perspectives and details are in Anthony (1916, 1923), Gentry (1983), and Myers and Lynch (1997). El Cope, continental divide north of, 600-700 m (Cocle). The village of El Cope is at 08°37’N, 80°35'W (Fairchild and Handley, 1966). El Cope (Omar Torrijos) National Park now encompasses the continental divide. Isla Popa, south end of, Channel (Bocas del 09°09’N, 82°08'W. Laguna, 820 m (Darién). 08°04’N, 77°19'W (Fair- child and Handley, 1966). According to Charles W. Myers (personal communication), this locality is on a “ridge south of the Rio Tacarcuna (upper tributary of Rio Pucuro, Tuira drainage).” Also referred to as “La Laguna,” it is not the village of the same name near the coast in southwestern Darién Province (e.g., aS indexed in the NGA), with which it is sometimes confused. La Loma, W Panama (Bocas del Toro). 08°50'N, 82°12’W. Also known as Buenavista. 1,200 ft. elevation (366 m) on the Atlantic slope along a trail from Chiriqui Lagoon to David (Dunn, 1942: 478). Peninsula Valiente, Bluefields, 70 m (Bocas del Toro). 09°11'N, 81°55’W. Peninsula Valiente delimits the eastern side of the Laguna de Chiriqui. Peninsula Valiente, Quebrada Hido (Bocas del Toro). Not located. Pequeni—Esperanza ridge, near head of Rio Pe- queni, 2,000 ft. [610 m] (Panama). Approximately 09°29'N, 79 24’W. Now within Parque Nacional Chagres. A ridge northeast of Lago Alajuela (Madden Lake), running roughly northeast to southwest and separating the Chagres—Esper- anza river system from the Pequeni—Boquerén system (Dunn and Bailey, 1939: 4, 15). Cadle and Myers (2003: 15-17) described some of the geography of the area as altered by damming associated with construction of the Panama Canal. Ibanez et al. (“1994” [1995], fig. 1) provided an outline map of the major rivers of the region. 1 km E of Sumwood Toro). Approximately 235 Pequeni—Esperanza ridge, junction main divide, 1,200 ft. [366 m] (Panama). Approximately 09°20’N, 79 20’W. Location not precisely indi- cated by Dunn and Bailey (1939) but presumably at a lower elevation along the ridge closer to Lago Alajuela than the above-listed locality near the head of the Rio Pequeni (see above cited references). Summit site, border of Darién, 320 m (San Blas). 08°55'N, 77 '51'W (coordinates from the collector of FMNH 170138, Michael Duever). Also known as Camp Summit, a canal survey camp in the Serrania del Darién discussed by Myers (1969: 26-27, fig. 1), who gave the elevation as 358 m. LITERATURE CITED ALVAREZ DEL Toro, M. 1972. Los Reptiles de Chiapas. Segunda edition. Tuxtla Gutierrez, Chiapas, Mex- ico: Gobierno del Estado de Chiapas. ALVAREZ DEL Toro, M. 1982. Los Reptiles de Chiapas. Tercera edition. Tuxtla Gutierrez, Chiapas, Mexi- co: Instituto de Historia Natural. Antuony, H. E. 1916. Panama mammals collected in 1914-1915. Bulletin of the American Museum of Natural History 35: 357-376 + foldout ee AntHony, H. E. 1923. In the footsteps of Natural History 23: 312-324. Autu, D. L. 1994. Checklist and bibliography of the amphibians and reptiles of Panama. Smithsonian Herpetological Information Service 98: 1-59. Barpour, T., AND A. Lovertnce. 1946. First supplement to typical reptiles and amphibians. Bulletin of the Museum of Comparative Zoology 96: 59-214. Bett, T. 1874. The Naturalist in Nicaragua, a Narrative of a Residence at the Gold Mines of Chontales; Journeys in the Savannahs and Forests; with Observations on Animals and Plants in Reference to the Theory of Evolution of Living Forms. ee John Murray. BOLANOs, J. M. Savace, AND G. CHAves. 2010. vee y reptiles de C eee Rica (version: 18 de Agosto del 2010). Listas Zoolégicas Actualizadas UCR. San Pedro: Museo de Zoologia, Universidad de Costa Rica. Available from: http://museo.biolo- gia.ucr.ac.cr/ Listas/LZAPublicaciones.htm. Ac- cessed | January 2011. BouLencer, G. A. 1894. Catalogue of the snakes in the British Museum (Natural History). Volume 2. London: British Museum (Natural History). Cab e, J. E. 2005. Systematics of snakes of the Dipsas oreas complex ( (Colubridae: Dipsadinae) in western Ecuador and Peru, with revalidation of D. elegans (Boulenger) and D. ellipsifera (Boulenger). Bulle- tin of the Museum of Comparative Zoology 158: 67=136. Cape, J. E. 2007. The snake genus Sibynomorphus (Colubridae: Dipsadinae: Dipsadini) in Peru and Ecuador, with comments on the systematics of Balboa. 236 Dipsadini. Bulletin of the Museum of Comparative Zoology 158: 183-254. Cape, J. E. 2010. Systematics, natural history, and hemipenial morphology of Dendrophidion brun- neum (Giinther), a poorly known snake from the Andes of Ecuador and Peru. Zootaxa 2433: 1-24. Cap_e, J. E., anpD C. W. Myers. 2003. Systematics of snakes ncleared to Dipsas variegata in Panama and western South Aaaerionl with Tevalidaton of two species and notes on defensive behaviors in the Dipsadini Colubridae). American Museum Novi- tates 3409: 1-47. CampsELL, J. A. 1998. Amphibians and Reptiles of Northern Guatemala, the Yucatan, and _ Belize. ae University of Oklahoma Press. CamppeLt, J. A., AND J. P. VANNINI. 1989. Distribution of amphibians and reptiles in Guatemala and Belize. Proceedings of the Western Foundation of Verte- brate Zoology 2 wal Cuan, L. M., J. L. Brown, anp A. D. Yoprr. 2011. Integrating statistical genetic and geospatial meth- ods brings new power to phylogeography. Molec- ular Phylogenetics and Evolution 59: 523-537. Cocnran, D. M. 1961. Type specimens of reptiles and amphibians in the U.S. National Museum. Bulletin of the United States National Museum 220: 1-291. Daruinc, D. M., AND H. M. Smiru. 1954. A collection of reptiles and amphibians from eastern Mexico. Transactions of the Kansas Academy of Science 57: 180-195. Daza, J. M., T. A. CasTog, AND C. L. Parkinson. 2010. Using regional comparative phylogeographic data from snake lineages to infer historical processes in Middle America. “Ecography 33: 343-354. DowLinc, H. G., AND J. M. Savace. 1960. A guide to the snake hemipenis: a survey of basic structure and systematic characteristics. ‘Zoologica 455 17—-31h DUELLMAN, W. E. 1960. A distabutional study of the amphibians of the Isthmus of Tehuantepec, Mexico. University of Kansas Publications, Muse- um of Natural History 13: 19-72. DueuiMan, W. E. 1963. Amphibians and reptiles of the rainforests of southern El Petén, Guatemala. University of Kansas Publications, Museum of Natural History 15: 205-249. DuELLMAN, W. E. 1966. The Central American herpetofauna: an ecological perspective. Copeia 1966: 700-719. Duces, A. 1892. El] Dendrophidium dendrophis Schl. Fitz. Syst. Rept. 1843. La Naturaleza ser. 2, 2: 100- 101 + 1 pl. DumeErIL, A. H. A., M.-F. Bocourt, AND F. MOCQUARD. 1870-1909. Mission Scientifique au Mexique et dans Amérique Centrale. Recherches Zoologiques, Troisieme Partie, Premicre Section: Etudes sur les Reptiles. Paris: Imprimerie Impériale. DuMERIL, A. M. C., G. Brpron, AND A. H. A. DUMERIL. 1854. Erpétologie Générale ou Histoire Naturelle Complete des Reptiles. Tome 7. Paris: Librairie Encyclopédique de Roret. Bulletin of the Museum of Comparative Zoology, Vol. 160, No. 4 Dunn, E. R. 1933. Amphibians and reptiles from El Valle de Anton, Panama. Occasional Papers of the Boston Society of Natural History 8: 65-79. UNN, E. R. 1942. The American caecilians. Bulletin of the Museum of Comparative Zoology 91: 437-540. Dunn, E. R., AnD J. R. Barttey. 1939. Snakes from the uplands of the Canal Zone and of Darien. nee of the Museum of Comparative Zoology 86: 1-2 DUNN, E.R. AND EA @: STuArm WSbile Coumieniene on some recent restrictions of type localities of certain South and Central American amphibians and reptiles. Copeia 1951: 55-61. EsparzA-Torres, H. F. Undated [ca. 1977]. Estado de Veracruz |Map of the state of Veracruz]. Mexico City: Libreria Patria. FAIRCHILD, G. B., anD C. O. HANDLEY, Jr. 1966. Gazetteer of collecting localities in Panama, pp. 9-20 + map. In R. L. Wenzel and V. J. Tipton (eds.), Ectoparasites of Panama. Chicago: Field Museum of Natural History. FauTu, J. E., B. I. CROTHER, AND J. B. SLowinskI. 1989. Elevational patterns of species richness, evenness, and abundance of the Costa Rican leaf-litter herpetofauna. Biotropica 21: 178-185. FLORES-VILLELA, O. 1993. Herpetofauna Mexicana. Lista anotada de las especies de anfibios y reptiles de México, cambios taxonomicos recientes, y nuevas especies. Carnegie Museum of Natural History, Special Publication 17: iv + 73 pp. FLORES VILLELA, O., AND P. GrerREz. 1988. Conservacion en México: Sintesis sobre Vertebrados Terrestres, Vegetacion, y Uso del Suelo. Xalapa, Veracruz, Mexico: Instituto Nacional de Investigaciones sobre Recursos Bidticos. Gaicr. H. T., N. HArtwec, AnD L. @2-Srusni ar: Notes on a collection of amphibians and reptiles from eastern Nicaragua. Occasional Papers of the Museum of Zoology, University of Michigan 357: Wish. Gentry, A. H. 1983. A biological exploration of Cerro Tacarcuna. National Geographic Research Reports 53 233-235) Go.pBerc, S. R. 2003. Reproduction in four species of Dendrophidion from Costa Rica (Serpentes: Colu- bridae). Transactions of the Illinois State Academy of Science 96: 295-300. GoLtpMAN, E. A. 1951. Biological Investigations in México (Smithsonian Miscellaneous Collections 115). Washington, DC: Smithsonian Institution. Goopwin, G. G. 1969. Mammals from the state of Oaxaca, Mexico, in the American Museum of Natural History. Bulletin of the American Museum of Natural History, 141: 1-270 + pls. 1-40. Guntuer, A. C. L. G. 1885-1902. Biologia Centrali- Americana, Reptilia and Batrachia [SSAR reprint 1987]. Ithaca, New York: Society for the Study of Amphibians and Reptiles. Guyer, C. 1994. The reptile fauna: diversity and ecology, pp. 210-216 + + appendix 6. In K. S. Bawa, G. S. Hartshorn, H. A. Hespenheide, and L. A. Macdade (eds.), La Selva: Ecology and Natural SPECIES IN THE DENDROPHIDION VINITOR COMPLEX ¢ Cadle History of a Neotropical Rainforest. Chicago: The University of C nie Press. Guyer, C., AND M. A. DonnELty. 1990. Length-mass relationships ee an assemblage of tropical snakes in Costa Rica. Journal of Tropical Ecology 6: 65-76. Guyer, C., AND M. A. DonNELLY. 2005. Amphibians and Reptiles of La Selva, Costa Rica, and the Caribbean Slope: A Comprehensive Guide. Berkeley: Univer- sity of California Press. E. R., AND W. W. DatouEst. 1963. The mammals of Veracruz. University of Kansas Publications, Museum of Natural History, 14: 165-362. Houprince, L. R. 1967. Life Zone Ecology. Revised edition. San José, Costa Rica: Tropical Science Center. HowpripcE, L. R., W. C. GRENKE, W. H. HaTHeway, T. LIANG, AND J. A. Tosi, JR. 1971. Forest Environ- ments in Tropical Life Zones, a Pilot Study. Oxtord, United Kingdom: Pergamon Press. IBANEZ; ID? R:, F. A. AROSEMENA, F’. A. SOLfs, AND C. A. JaraMILLo. “1994” [1995]. Anfibios y reptiles de la Serrania Piedras-Pacora, Parque Nacional Chagres. Scientia (Panama) 9: 17-31. IsANEz, D. R., anp F. A. Sotis. “1991” [1993]. Las serpientes de Panama: lista de especies, comentar- ios taxonomicos y bibliografia. Scientia (Panama) 6: 9759 Janzen, D. H. 1973. Sweep samples of tropical foliage insects: description of study sites, with data on species abundances and size distributions. Ecology D4: 659-686. Jounson, J. D. 1989. A biogeographic analysis of the herpetofauna of northwestern Nuclear Central America. Milwaukee Public Museum Contributions in Biology and Geology 76: 1-66. JOHNSON; |. D:,’C. A: Eiy, AND R:iG. Wess, ~1976" [1977]. Biogeographical and taxonomic notes on some herpetozoa from the northern highlands of Chiapas, Mexico. Transactions of the Kansas Academy of Science 79: 131-139. Kiuce, A. G. 1984. Type-specimens of reptiles in the University of Michigan Museum of Zoology. Miscellaneous Publications of the Museum of Zoology, University of Michigan 167: 1-85. Kouter, G. 2003. Reptiles of Central America. English edition. Offenbach, Germany: Herpeton. KonLer, G. 2008. Reptiles of Central America. 2nd English edition. Offenbach, Germany: Herpeton. LaurENcIO, D., AND J. H. MALone. 2009. The amphib- ians and reptiles of Parque Nacional Carara, a transitional herpetofaunal assemblage in Costa Rica. Herpetological Conservation and Biology 4: 120-131. Ler, J. C. 1980. An ecogeographic analysis of the herpetofauna of the Yucatan Peninsula. University of Kansas Museum of Natural History, Miscella- neous Publications, 67: 1-75. Lee, J. C. 1996. The Amphibians and Reptiles of the Yucatan Peninsula. Ithaca: Comstock Publishing Associates. HALL, 231, Ler, J. C. 2000. A Field Guide to the Amphibians and ‘Reptiles of the Maya World, the Lowlands of Mexico, Northern Guatemala, and Belize. Ithaca: Cornell University Press. Lirs, C. S. 1988. Systematic status of the Neotropical snakes De ndrophidion dendrophis and D. nuchalis (Colubridae). Herpetologica 44: 162-175. Lins, C. S. 1991. Dendrophidion vinitor. Catalogue of American Amphibians and Reptiles, 522: 522.1- 599.9. Lirs, K. R. 1998. Decline of a tropical montane amphibian fauna. Conservation Biology 12: 106— LIT. Lips, K. R. 1999. Mass mortality and population declines of anurans at an upland site in western Panama. Conservation Biology 13: 117-125 Lips, K. R., F. Brem, R. Brenegs, J. D. Reeve, R. A. ALFORD, J. VoyLES, C. Carey, L. Livo, A. P. PESSIER, AND J. P. Contins. 2006. Emerging infectious disease and the loss of biodiversity in a Neotropical amphibian community. Procee dings of the National Academy of Sciences, USA 103: 3165-3170. Kk. R., D. E. GREEN, AND R. PApENpDick. 2003. C hytridiomy cosis in wild frogs from southern Costa Rica. Journal of Herpe tolozy 37: 215-218. McCoy, C. J. 1970. The snake fauna of Middlesex, British Honduras. Journal of Herpetology 4: 135- 140. McCranig, J. R. 2009. Amphibians and reptiles of Lips, Honduras (version 12 November 2009). Listas Zoologicas Actualizadas UCR. San Pedro: Museo de Zoologia, Universidad de Costa Rica. Avail- able foun: http://museo. biologia.ucr.ac.cr/Listas/ LZAPublicaciones.htm. Accessed 1 January 2011. McCranig, J. R. 2011. Snakes of Honduras: Systemat- ics, Distribution, and Conservation. Ithaca: Society for the Study of Amphibians and Reptiles. McCranlg, J. R., J. H. TOWNSEND, AnD L. D. WILson. 2006. The Amphibians and Reptiles of the Hon- duran Mosquitia. Malabar, Florida: Krieger Pub- lishing Co. McDape, L. A., K. S. Bawa, H. A. HESPENHEIDE, AND G. S. Hartsuorn. 1994. La Selva: Ecology and Natural History of a Neotropical Rain Forest. Chicago: University of Chicago Press. McDape, i: A., ann CG. S: MiGn ons 1994. La Selva Biological Station, pp. 6-14. In K. S. Bawa, G. S. Fentstiorn: H. A. Hespenheide, and L. A. Macdade (eds.), La Selva: Ecology and Natural History of a Neotropical Rainforest. Chicago: The University of Chicago Press. McDiarMip, R. W., ano J. M. Savace. 2005. The herpetofauna of the Rincon area, Peninsula de Osa, Costa Rica, a Central ees lowland ev ee forest site, pp. 366-427. In M. A. Donnelly, Crother, C. Guyer, M. TL Wake, and M. E. fee (eds.), Ecology and Evolution in the Tropics, a Herpetological Perspective. Chicago: University of Chicago Bree MEERMAN, J. C., AND J. C. Lee. Undated [published 2003]. Amphibians and reptiles of the Columbia 238 River Forest Reserve, pp. 66-70. In J. C. Meerman and S. Matola (eds.), The Columbia River Forest Reserve Little Quartz Ridge Expedition, A Biolog- ical Assessment. Belize City: Wildlife Trust. MEERMAN, J. C., AND S. Martota (eds.). Undated [published 2003]. The Columbia River Forest Reserve Little Quartz Ridge Expedition, A Biolog- ical Assessment. Belize City: Wildlife Trust. MoreELET, A. 1871. Travels in Central America, Including Accounts of Some Regions Unexplored since The Conquest [translated from the French by M. F. Squier]. London: Triibner and Co [French edition, 2 vols., 1857]. Myers, C. W. 1969. The ecological geography of cloud forest in Panama. American Museum Novitates 2396: 1-52. Myers, C. W. 1974. The systematics of Rhadinaea (Colubridae), a genus of New World snakes. Bulletin of the American Museum of Natural History 153: 1-262. Myers, C. W. 2011. A new genus and new tribe for Enicognathus melanauchen Jan, 1863, a neglected Southi American snake (Colubridae: Xenodonti- nae), with taxonomic notes on some Dipsadinae. American Museum Novitates 3715: 1-33. Myers, C. W., AND J. E. Cape. 1994. A new genus for South American snakes related to Bienes obtusa Cope (Colubridae) and resurrection of Taeniophallus Cope for the “Rhadinaea” breviros- tris group. American Museum Novitates 3102: 1— 30: Myers, C. W., aNd J. E. Cape. 2003. On the snake hemipenis, with notes on Psomophis and tech- niques of eversion: a response to Dowling. Herpetological Review 34: 295-302. Myers, C. W., AnD J. A. CAMPBELL. 1981. A new genus and species of colubrid snake from the Sierra Madre del Sur of Guerrero, Mexico. American Museum Novitates 2708: 1—20. Myers, C. W., AND J. D. Lyncu. 1997. Eleutherodacty- lus laticorpus, a peculiar new frog from the Cerro Tacarcuna Area, Panamanian- Colombian frontier. American Museum Novitates 3196: 1—12. Perez-HicarEDA, G. 1978. Reptiles and amphibians from the Estacion de Biologia Tropical “Los Tuxtlas” (U.N.A.M.), Veracruz, México. Bulle- tin of the Maryland Herpetological Society 14: O74) PEREZ-HiGaREDA, G., AND H. M. Situ. 1991. Ofidio- fauna de Veracruz, Analisis Taxonomico y Zoogeo- grafico! Ophidiofauna of Veracruz, Taxonomical and Zoogeographical Analysis [Publicaciones Especiales del Instituto de Biologia no. 7]. Los Tuxtlas, México: Universidad Nacional Aut6noma de México. PEREZ-HIGAREDA, G., R. C. VocT, AND O. A. FLORES- VILLELA. 1987. Lista anotada de los anfibios y reptiles de la region de Los Tuxtlas, Veracruz. Los Tuxtlas. Veracruz, México: Estacién de Biologia Tropical Los Tuxtlas, Instituto de Biologia, Uni- versidad Nacional Aut6noma de México. Bulletin of the Museum of Comparative Zoology, Vol. 160, No. 4 PEREZ-SANTOS, C., AND A. G. Moreno. 1988. Ofidios de Colombia. Torino: Museo Regionale di Scienze Naturali. PEREZ-SANTOS, C., AND A. G. Moreno. 1989. Addenda y corrigenda al libro “Ofidios de Colombia.” Bollet- tino del Museo Regionale di Scienze Naturali, Horinowi allie ‘ PEREZ-SANTOS, C., A. G. MORENO, AND A. GEARHART. 1993. Checklist of the snakes of Panama. Revista Espanola de Herpetologia 7: 113-122. Peters, J. A. 1952. Catalogue of type specimens in the herpetological collections of the University of Michigan Museum of Zoology. Occasional Papers of the Museum of Zoology, University of Michigan TWP Sis), Peters, J. A., AND B. OreyAs-Miranpa. 1970. Catalogue of the Neotropical Squamata: Part I. Snakes. Bulletin of the United States National Museum 297: 1-347. PRINGLE, C. M., I. CHAcon, M.Grayum, H. W. GREENE, G. HartsHorn, G. Scuatz, G. Stites, C. Gomez, AND M. Ropricuez. 1984. Natural history observa- tions and ecological evaluation of the La Selva Protection Zone, Costa Rica. Brenesia 22: 189- 206. Ranb, A. S., anD C. W. Myers. 1990. The herpetofauna of Barro Colorado Island, Panama: an ecological summary, pp. 386-409. In A. H. Gentry (ed.), Four Neotropical Rainforests. New Haven, Connecticut: Yale University Press. Root, N. J. 1975. The Thomas Baillie MacDougall collection in the American Museum of Natural History Library. Curator 18: 270-276. RossMaN, D. A., AND W. G. EBERLE. 1977. Partition of the genus Natrix, with preliminary observations on evolutionary trends in natricine snakes. Herpetolo- gica 33: 34-43, Ginros buna. © x., J. PacHeco, F. MENDozA-QUI- JANO, F. BoLaNos, G. CuAves, G. C. Dairy, P. R. EHRLICH, AND G. CEBALLOS. 2008. Diversity, natural history and conservation of amphibians and reptiles from the San Vito region, southwest- ern Costa Rica. Revista de Biologia Tropical 56: Lao = ior SavacE, J. M. 1966. The origins and history of the € eae American herpetofauna. Copeia 1966: 719-766. SavaGE, J. M. 1973. A Preliminary Handlist of the Herpetofauna of Costa Rica. Lawrence, Kansas: American Society of Ichthyologists and Herpetol- ogists. SavacE, J. M. 1980. A Handlist with Preliminary Keys to the Herpetofauna of Costa Rica. Los Angeles: Allan Hancock Foundation. SavacE, J. M. 1997. On terminology for the description of the hemipenes of squamate reptiles. Herpeto- logical Journal 7: 23-25. SavacE, J. M. 2002. The Amphibians and Reptiles of Costa Rica, a Herpetofauna between Two Conti- nents, between Two Seas. Chicago: The University of Chicago Press. SPECIES IN THE DENDROPHIDION VINITOR COMPLEX ¢ Cadle SAVAGE, J. M., AND F. Botanos. 2009. A checklist of the amphibians and reptiles of Costa Rica: additions and nomenclatural revisions. Zootaxa 2005: 1-23. SavaGE, J. M., AND J. Vinita. 1986. Introduction to the Herpetofauna of Costa Rica/Introduccion a la Her- petofauna de Costa Rica. Ithaca, New York: Society for the Study of Amphibians and Reptiles. SCHARGEL, W. E., G. Rivas FUENMAYOR, AND C. W. Myers. 2005. An enigmatic new snake from cloud forest of the Peninsula de Paria, Venezuela (Colubridae: genus Taeniophallus?). American Museum Novitates 3484; 1-22. Scott, N. J., JR. 1976. The abundance and diversity of the herpetofaunas of tropical forest litters. Biotro- pica 8: 41-58. ScoTT, N. J., JR., J. M. Savacg, AND D. C. Rosinson. 1983. Checklist of reptiles and amphibians, pp. 367-374. In D. H. Janzen (ed.), Costa Rican Natural History. Chicago: University of Chicago Press. SmituH, H. M. 1941. A new name for the Mexican snakes of the genus Dendrophidion. Proceedings of the Biological Society of Washington, 54: 73-76. SmitH, H. M. 1943. Summary of the collections of snakes and crocodilians made in Mexico under the Walter Rathbone Bacon traveling scholarship. Proceedings of the United States National Museum, 93: 393-504 + pl. 32. Smitu, H. M. 1958. Handlist of the snakes of Panama. Herpetologica 14; 222-224. SmitH, H. M. 1987. Current nomenclature for the names and material cited in Giinther’s Reptilia and Batrachia volume of the Biologia Centrali-Ameri- cana, pp. xv-li. In K. Adler (ed.), Albert C. A. G. Giinther, Biologia Centrali-Americana, Reptilia and Batrachia. Ithaca, New York: Society for the Study of Amphibians and Reptiles. SmiTH, H. M., AnD E. H. Taytor. 1945. An annotated checklist and key to the snakes of Mexico. Bulletin of the United States Natiénal Museum 187: 1-239. SmiTH, H. M., anp E. H. Taytor. 1950. Type localities of Mexican reptiles and amphibians. University of Kansas Science Bulletin 33, pt. 2: 313-380. SOLORZANO, A. 2004. Serpientes de Costa Rica: Dis- tribucion, Taxonomia e Historia Natural/ Snakes of Costa Rica: Distribution, Taxonomy and Natural History. Eagle Mountain, Utah: Eagle Mountain Publishing, LC. STAFFORD, P. J. 1998. Amphibians and reptiles of the Cordillera de Guanacaste, Costa Rica: a field list with notes on colour patterns and other observa- tions. British Herpetological Society Bulletin 62: 9- 19. STAFFORD, P. J. 2002. Record of the colubrid snake Dendrophidion nuchale (Peters) from Nicaragua. Southwestern Naturalist 47: 615-616. STAFFORD, P. J. 2003. Trophic ecology and reproduction in three species of Neotropical forest racer (Dendrophidion; Colubridae). He rpetological Jour- na V3. 01-1. 239 STAFFORD, P. J., AND J. R. Meyer. 2000. A Guide to the Reptiles of Belize. San Diego: Academic Press. Stix, J. S. 1975. Thomas Baillie MacDougall—naturalist and collector. Curator 18: 270-276. Stuart, L. C, 1932. Studies on Neotropical Colubrinae. The taxonomic status of the genus Drymobius Fitzinger. Occasional Papers of the Museum of Zoology, University of Michigan 236: 1-16. StTuarT, L. C. 1934. A eonabutien to a knowledge of the he rpe ‘tological fauna of E] Petén, Guate mala. Occasional Papers of the Museum of Zoology, U ee of Michigan 292: 1-18. STUART, L. C. 1935. A contribution to a knowledge of the he “pe tology of a portion of the savanna re gion of Central Petén, Guatemala. Miscellaneous Pub- lications of the Museum of Zoology, University of Michigan 29: 1-56. STU: a , L. C. 1948. The amphibians and reptiles of Alta Ver nee, Guatemala. Miscellaneous Publications of the Museum of Zoology, University of Michigan 69: 1-109. STuakT, L. C. 1950. A geographic study of the herpetofauna of Alta Verapaz, Guatemala. Contri- butions from the Laboratory of Vertebrate Biology, University of Michigan 45: 1-77. Stuart, L. C. 1958. A study of the herpetofauna of the Uaxactiin—Tikal area of northern El Petén, Guate- mala. Contributions from the Laboratory of Verte- brate Biology, University of Michigan 75: 1-30. StuaRrT, L. C. 1963. A checklist of the ‘herpetofauna of Guatemala. Miscellaneous Publications of the Museum of Zoology, University of Michigan 122: 1-150. Taytor, E. H. 1944. Present location of certain herpetological and other type specimens. Univer- sity of Kansas Science Bulletin 30, part 1: 117- TST; Taytor, E. H. 1954. Further studies on the serpents of Costa Rica. University of Kansas Science Bulletin 36, part 2: 673-801. VitLa, J.,.L. D. Witson, anp J. D. era Byes: Middle American Herpetolog se ra. Columbia: University a Missouri Press. WAKE, J. M. Savace, AND J. HANKEN. 2007. ae ‘salamanders from the Costa Rica—Pana- ma border region, with descriptions of two new species of Bolitoglossa. Copeia 2007: 556-565. WHITFIELD, S. M., K. E. BELL, T. Puiippi, M. Sasa, F. BoLaNos, G. CHAvEs, J. M. SAvAGE, AND M. A. Donnetty. 2007. Amphibian and reptile declines over 35 years at La Selva, Costa Rica. Proceedings of the National Academy of Sciences, USA 104: 8352-8356. Witson, L. D. 1966. Dendrophidion vinitor, an addition to the snake fauna of British Honduras. Journal of the Ohio Herpetological Society 5: 103. Witson, fi. 1D: J. Re McCranie, S. GOTTE, AND J. H. TOWNSEND. 2003. Distributional comments on some members of the herpetofauna of the Mosquitia, Honduras. Herpetological Bulletin 84: 15-19. 240 Bulletin of the Museum of Comparative Zoology, Vol. 160, No. 4 Witson, L. D., AND J. R. Meyer. 1985. The Snakes of | ZAHER, H. 1999. Hemipenial morphology of the Honduras, 2nd edition. Milwaukee: Milwaukee South American xenodontine snakes, with a Public Museum. proposal for a monophyletic Xenodontinae and Witson, L. D., AND J. H. Townsenp. 2006. The a reappraisal of colubroid hemipenes. Bulletin of herpetofauna of the rainforests of Honduras. the American Museum of Natural History 240: Caribbean Journal of Science 42: 88-113. 1-168. US ISSN 0027-4100 MCZ Publications Museum of Comparative Zoology Harvard University 26 Oxford Street Cambridge, MA 02138 mczpublications@mcz.harvard.edu © The President and Fellows of Harvard College 2012. Volume 160, Number 5 30 November 2012 On the Cyphophthalmi (Arachnida, Opiliones) types from the Museo Civico di Storia Naturale “Giacomo Doria” RONALD CLOUSE AND GONZALO GIRIBET HARVARD UNIVERSITY | CAMBRIDGE, MASSACHUSETTS, U.S.A. BULLETIN OF THE Museum of Comparative Zoology BOARD OF EDITORS Editor: Jonathan Losos Managing Editor: Deborah Smiley Associate Editors: Andrew Biewener, Scott Edwards, Brian Farrell, Farish Jenkins, George Lauder, Gonzalo Giribet, Hopi Hoekstra, Jim Hanken, Jim McCarthy, Naomi Pierce, and Robert Woollacott Publications Issued or Distributed by the Museum of Comparative Zoology Harvard University Bulletin 1863- Breviora 1952- Memoirs 1865-1938 Johnsonia, Department of Mollusks, 1941-1974 Occasional Papers on Mollusks, 1945- General queries, questions about author guidelines, or permissions for MCZ Publications should be directed to the editorial assistant: MCZ Publications Museum of Comparative Zoology Harvard University 26 Oxford Street Cambridge, MA 02138 mczpublications@mcz.harvard.edu EXCHANGES AND REPRINTS All of our publications are offered for free on our website: | http://www.mcz.harvard.edu/Publications/index.html | To purchase individual reprints or to join our exchange program, | please contact April Mullins at the Ernst Mayr Library: | mayrlib@oeb.harvard.edu. This publication has been printed on acid-free permanent paper stock. © The President and Fellows of Harvard College 2012. ON THE CYPHOPHTHALMI (ARACHNIDA, OPILIONES) TYPES FROM THE MUSEO CIVICO DI STORIA NATURALE “GIACOMO DORIA” RONALD CLOUSE' AND GONZALO GIRIBET* AsstracTt. The Museo Civico di Storia Naturale “Giacomo Doria,’ Genoa, hosts one of the most important historical collections of the Opiliones subor- der Cyphophthalmi, including all the known specimens for the type species of the genera Leptopsalis (Stylo- cellidae), Miopsalis (Styloce lid ie), and Parogovia (Neo- goveidae), as well as several other types in the families Ogoveidae and Stylocellidae (Git is unclear whether specimens in Petté ilidae and Sironidae constitute types). These specimens were recently made available to us for study, and given their importance, we discuss and illustrate them here. Study of this collection allows confirmation of the validity of Le ~ptopsalis, considered a synonym of Stylocellus for more than a century, and of Miopsalis, considered a nomen dubium in the most recent catalogue of the group. It furthermore helps to clarify the identity of several other species in the family Stylocellidae. Here we formally resurrect the genera Miopsalis T Thorell, 1890, and Leptopsalis Thorell, 1882, and transfer several species to these genera: M. collinsi (Shear, 1993) comb. nov.; M. gryllospeca (Shear, 1993) comb. nov.; M. lionota (Pocock, 1897) comb. nov.; M. sabah (Shear, 1993) comb. nov.; M. silhavyi (Rambla, 1991) comb. nov.; M. tarumpitao (Shear, 1993) comb. nov.; L. dumoga (Shear 1993) comb. nov.; L. hillyardi (Shear, 1993) comb. nov.; L. javana Thorell, 1882; L. lydekkeri (Clouse & Giribet, 2007) comb. nov.; L. modesta (Hansen & Sgrensen, 1904) comb. nov.; L. novaguinea (Clouse & Giribet, 2007) comb. nov.; L. ramblae (Giribet, 2002) comb. nov.; L. sulcata (Hansen & Sorensen, 1904) comb. nov.; L. tambusisi (Shear, 1993) comb. nov.: L. thorellii (Hansen & Sorensen, 1904) comb. nov.; L. weberii (Hansen & Serensen, 1904) comb. nov. Key words: Neogoveidae, Ogoveidae, Sironidae, Stylo- cellidae, Museum Pacanous ' Division of Invertebrate Zoology, American Muse- um of Natural History, Central Par ke West at 79th Street, New York, New York 10024 (rclouse@amnh.org). > Museum of Comparative Zoology, Department of Organismic and Evolutionary Biology, Harvard Univer- sity, 26 Oxford Street, Cambridge, Massachusetts 02138. Author for correspondence (ggiribet@oeb. harvard.edu). Bull. Mus. Comp. Zool., 160(5): INTRODUCTION One of the worst impediments of taxonomy is, along with the loss of expertise (e.g., Rodman and Cody, 2003), the dependence on type specimens that are not properly described or documented; hence the impor- tance of well-funded museums to ensure that collections are properly maintained and specimens are made available to the scientific community (Suarez and Tsutsui, 2004). But proper cataloguing and revisionary taxonomy often require examinations of large numbers of type specimens scattered in institutions around the world. These are usually in museums and university collections but are sometimes in private collections that may not be up to curatorial standards, especially ‘after the researcher passes away and the collection burdens surviving family members. Many important wollecnene have been lost this way, including some of the Opiliones collections of interest to us, most prominently the collec- tion of Julio A. Rosas Costa, with his Cyphophthalmi types (Rosas Costa, 1950) becoming lost to science for the time being. Untor eam ate le as exemplified by the ar achaid order Opiliones, many types are often simply unavailable. Here we study a collection of the harvestman suborder Cyphophthalmi and document the specimens deposited at the Museo Civico di Storia Naturale “Giacomo Doria,” Genoa, Italy (MCSN), which we had not examined until recently. The importance of this collection is without parallel in many respects, as it includes types and nontypes of many of the earliest cyphophthalmid species. JAI=257,. November, 2012 9A] 242 Ty es s of the suborder Cyphophthalmi are deposited in approximately 30 museums around the world (Giribet, 2000). A visit to half of these museums would allow one to study around 85% of the known species (a large number of recently described species are deposited in the collection of Ivo Karaman, University of Novi Sad, Serbia, but they are not av ‘ailable for study to the broader community). In terms of historical collections, many ‘of the oldest types are deposited at The Natural History Museum, London, which includes 15 primary types, although their collections are not especially large. The Muséum national d Histoire naturelle, Paris, includes numerous types (mostly from Christian Juberthie and most New Caledonian species), as well as a large general collection. The American Museum of Natural History, New York, includes 15 types and one of the lar gest general collections of Cyphophthalmi, with near ly all families represented. Currently, the largest collection in terms of diversity and including represen- tatives of all families is that of the Museum of Comparative Zoology, Harvard University, with ca. S00 lots, most of it recently collected, DNA-gr ade specimens. It also contains more types than any other institution, although few are “historical” types. Te Papa Tongarewa, Wellington, with 26 types, and the Canter- bury Museum, Christchurch, with 11 types, are indispensable visits for those studying the New Zealand fauna. The size of the MCSN collection is small compared with these other collections; however, its importance lies in the species available there, including the primary (and only) types of six species, in addition to a few other specimens, including additional type specimens. Among them are the type species for two of the five currently recognized genera in the family Stylocellidae, including dhe only described specimens in the yet- -monotypic genus Miopsalis Thorell, 1890 (Clouse et al. 30)()9: Clouse and Giribet, 2010). The foanined species are: Parogovia sironoides Hansen, 1921: Ogovea nase (Hansen, 1921); Miopsalis pulicaria Thorell, 1890; Leptopsalis beccarii Thorell, 1882; Leptopsalis javana Thorell, 1882; Leptopsalis Bulletin of the Museum of Comparative Zoology, Vol. 160, No. 5 thorellii (Hansen and Segrensen, 1904): and Leptopsalis weberii (Hansen and S¢rensen, 1904). The MCSN collection is of exceptional importance due to several factors. First is the fact that the prolific Swedish arachnol- ogist Tord Tamerlan Teodor Thorell studied arachnology with Giacomo Doria at the MCSN. In addition to describing more than 1,000 spider species from 1850 to 1900, Thorell described two stylocellid genera (Leptopsalis Thorell, 1882 and Miopsalis Thorell, 1890) and three stylocellid species, types of which are deposited at MCSN. Although Leptopsalis was synonymized with Stylocellus Westwood, 1874 by Thorell himself (1890b), and has remained in synonymy for more than a century, a clade containing the type species was recently found to be sufficiently distinct from the type species of Sealoodllis for them to be kept in separate genera (Clouse and Giribet, 2010; Clouse et al., 2009, 2011). Leptopsalis was thus informally revalidated by Clouse et al. (2009: 525), and the genus has been used in subsequent papers on stylocellid (Clouse et al., 2011) and cyphophthalmid (Giribet et al., 2011) phylogenetics. Miop- salis is currently a nomen dubium (Giribet, 2000), although it has recently been postu- lated to be fhe identity of a distinct clade that diversified mostly in Borneo (Clouse and Giribet, 2010; Clouse et al., 2009, 2011: Giribet et al., 2011). However, the identity of this clade as the genus Miopsalis was based on its inclusion of small species that closely match the description of M. puli- caria, but it could not be confirmed without examination of the type (and sole known specimen) of this species. Second, the great Italian explorer Leo- nardo Fea became an assistant at the MCSN, and his expeditions to the Gulf of Guinea yielded important specimens, in- cluding the type species of the genus Parogovia Hansen, 1921 and the second species of the genus Ogovea Hansen & Segrensen, 1921. Finally, tie collection also received specimens from Eugene Simon and Gustav Joseph, including specimens of CYPHOPHTHALMI TYPES FROM MUSEO CIvIco DiI STORIA NATURALE the type species of the genera Cyphophthal- mus Joseph, 1868 and Parasiro Hansen & Sgrensen, 1904, and it acquired some other specimens of Hansen and S@rensen’s spe- cies, including a juvenile of Purcellia illustrans Hansen & Sorensen, 1904. Few museums had, by the beginning of the 20th century, a collection as comple te as the one of the MCSN. By 1921, Cyphophthalmi had been described, and the MCSN had 10 of these represented in its collection, including eight or more types. By that time the British Museum of Natural History, London (later to become The Natural History Museum), the largest mu- seum of natural history in the oad had only types of seven cyphophthalmid species, some of these being type series shared with the MCSN. In spite of this importance, the MCSN collection of Cyphophthalmi has not been documented using modern imaging techniques, so here we provide high-reso- lution images of the most important speci- mens and ndiseuss long-standing questions about their morphology and taxonomy. MATERIALS AND METHODS Specimens were cleaned in a Branson 200 Ultrasonic cleaner and imaged under a Leica MX12.5 stereomicroscope, PLAN 0.5x, with images taken at different planes with a JVC Re CCD digital camera KY-F75U and inte- grated with the software Auto-Montage Pro version 5.02.0096 from Syncroscopy. Data provided for each specimen are a transcription of the specimen label. Modern localities and additional data are provided in the discussion of the specimens. Abbreviations: AMNH: American Museum of Natural History, New York, United States. BMNH: The Natural History Museum, London, United Kingdom. MCSN: Museo Civico di Storia Naturale “Giacomo Doria,” Genoa. Italy. 19 species ot “GIACOMO Dorta” ¢ Clouse and Giribet 943 MCGZ;: Museum of Comparative Zoology, eed University, Cambridge, Massachusetts, United States. Muséum national d’ Histoire naturelle, Paris, France. Zoologisk Museum, Statens Naturhistoriske Museum, Kobenhavns Universitet, Copenhagen, Denmark. MNHN: ZMUC: ANNOTATED TAXONOMIC SECTION Family Neogoveidae: Parogovia sironoides Hansen, 1921 Figures 1-3 Specimen in good condition with some appendages detached (Figs 1): Parogovia sironoides Hansen, 1921 Type: Male Island of Fernando Poo: Punta Frailes leg. L. Fea, X—Xl 1901 Specimen in poor condition, with white residue on exterior (Fig. 2): Parogovia sironoides Hansen, 1921 Type: Female Island of Fernando Poo: Basile (400-— 600 m) leg. L. Fea, VIII-IX 1901 The latter specimen above originally labeled incorrectly as male. Detached appendages: basichelicerite, distal ey segments, palp, leg I, leg IV (Fig. 3): Parogovia sironoides Hansen, 1921 Type: Male Five detached parts Genitalia (spermatopositor and oviposi- tor), not photographed: Parogovia sironoides Hansen, 1921 Type: Female, male Penis and ovopositor slide Mounted by G. Legg, 1986 244 Bulletin of the Museum of Comparative Zoology, Vol. 160, No. 5 Figure 1. lectotype in ventral (A), dorsal (B), and lateral (C) views. Parogovia sironoides Hansen, 1921, male para- Parogovia sironoides, the type species of the genus Parogovia Hansen, 1921, was described on the basis of two specimens from Bioko (formerly Fernando Poo), Equa- torial Guinea, one male from Punta Frailes collected by Leonardo Fea in October-— November 1901, and one female from Basilé, collected by Leonardo Fea in August-— September 1901, at an altitude of 400— 600 m. The male was eee illustrated by Hansen (1921: pl. IV, fig. 2a—l). The type material, including a preparation of the Figure 2. lectotype in ventral view. Parogovia sironoides Hansen, 1921, female para- spermatopositor and Ovipositor, was studied by Gerald Legg in 1986 (Legg, 1990). Punta Frailes, the locality of one of the syntypes, is of unknown identity; it is not found on current maps of Bioko, it is not known to local biologists interviewed by G.G., and it was not seen in a study of old maps of the former oes Colony of Fernando Poo. This is, however, one of the localities surveyed ke the Italian explor- er Leonardo Fea during his collecting trip to the Gulf of Guinea nid Portuguese West Africa, and it is the type loca of several other species of arthropods and vertebrates. Basilé refers to Pico Basilé (Mt. Basilé), formerly Pico de Santa Isabel, the highest mountain on the island of Bioko, auth an altitude of 3,011 m. It is the summit of the largest and highest of three overlapping Peconic shield volcanoes that form the detached Figure 3. appendages of lectotype male. From top to bottom, left to right: chelicera, distal articles; chelicera, basal article; palp; leg I; leg IV. Parogovia sironoides Hansen, 1921, CYPHOPHTHALMI TYPES FROM MUSEO CIVvICO DI STORIA NATURALE island. From the summit, Mt. Cameroon can be seen to the northeast. Bioko was formed along the Cameroon line, a major northeast-trending geologic fault that runs from the Atlantic Ocean into Cameroon. This line includes other volcanic islands in the Gulf of Guinea such as the island territory of Annobon. the island nation of Sao Tomé and Principe, and the massive stratovoleano of Mt. Cameroon, the latter of which also has one or more undescribed species of Parogo- via (authors’ unpublished data). Specimens identified as Parogovia siro- noides (erroneously spelled Paragovia sir- onoides in earlier publications) from Rio Campo, Continental Region of Equatorial Guinea, have been used in several taxon- omic and phylogenetic studies of Cy- phophthalmi and Opiliones (Giribet and Boyer, 2002; Giribet and Prieto, 2003; de Bivort and (iiber 2004; Boyer et al., 2005 Schwendinger and Giribet, 2005). Collec- tions of Parogovia specimens in the Conti- nental Region and on Pico Basilé in 2003 led us to correct this possible misidentifica- tion and to refer to the continental speci- mens as P. cf. sironoides, while we apply the name P. sironoides for the specimens collected on Pico Basilé (Boyer et al., 2007; Boyer and Giribet, he Clouse and Ginbet. 2007: Ginbet et al. 2010, 2011): Additional collecting in Cameroon in 2009 and subsequent phyi logenetic analysis show that these are two distinct species, and that they are not even sister spe cies (Giribet et al., 2011). This can now be oe eae by the study of the type material from MCSN. Family Ogoveidae: Ogovea nasuta (Hansen, 1921) Figures 4—6 Three specimens in optimal condition (Figs. 4-6): Ogovia nasuta Hansen, 1921 Types: Female, male, juvenile Island of Fernando Poo: Musola (400- 500 m) leg. L. Fea, I-1902 Ul “GiAcoMO Doria” ¢ Clouse and Giribet 9A! Figure 4. Ogovea nasuta (Hansen, 1921), male lectotype in dorsal (A), ventral (B), and lateral (C) views. Disarticulated graphed: specimen, not photo- Ogovia nasuta Hansen, 1921 Types Detached parts for figures Ogovea nasuta, originally spelle 1 Figure 11. Leptopsalis beccarii Thorell, 1882, first male para- lectotype, showing dorsal (A), ventral (B), and lateral (C) views. synonymized with Stylocellus sumatranus Westwood, 1874 by Thorell (1890b). Later, Hansen and Sgrensen (1904) recognized both species as valid, although they main- tained the synonymy of Leptopsalis with Stylocellus, as did Roewer (1923) and Rosas Costa (1950). Giribet’s (2000) catalogue, before examining either type, also left L. beccarii. in synonymy with S. suwmatranus, following Thorell (1890b), but the distinction between the species was recognized soon thereafter (Giribet, 2002). A molecular and continuous morphological phylogenetic anal- ysis including S. swmatranus (but without CYPHOPHTHALMI TYPES FROM MUSEO Civico DI STORIA NATURALE — Leptopsalis beccarii Thorell, 1882, second male Figure 12. paralectotype: disarticulated right chelicera (A), right legs I-III (B, left to right, respectively), and right leg IV (C). molecular data for this species) suggests that it branches out before the split between Miopsalis and Leptopsalis, although its exact position is uncertain, and _ it aiences key characters with members of the genera ‘& Giribet. Ul Giacomo Dorta” ¢ Clouse and Giribet 9! Meghalaya and Fangensis (Clouse et al., 2009). It is possible that Styloc ellus remains restricted to the type species of the genus. In contrast, our phylogenetic analysis shows that a large clade of stylocellids originating on Borneo and diversifying on the Thai-Malay Peninsula, Sumatra, Java, Sulawesi, and New Guinea includes several large Sumatran species closely resembling the type of Leptopsalis (C louse and Ge ibet, 2007: Clouse, 2010; Clouse and Giribet, 2010: Clouse et a 2009, 2011), and we formally resurrect this genus here and transfer the most Sele. members of this clade to Leptopsalis. These include L. beccarii Thorell, 1882, the type species; . dumoga (Shear, 1993) ae nov.; L: hillyardi. (Shear, 1993) comb. nov.; L. javana Thorell, 1882; L. lydekkeri (Clouse 2007) comb. nov.; L. modesta (Hansen & Sorensen, 1904) comb. nov.;: L. novaguinea (Clouse & Giribet, 2007) comb. nov.: L. ramblae (Giribet, 2002) comb. nov.; L. sulcata (Hansen & S@ren- sen, 1904) comb. nov.: L. tambusisi (Shear, 1993) comb. nov.: L. thorellii (Hansen & Sgrensen, 1904) comb. nov.; and L. weberii (Hansen & Sorensen, 1904) comb. nov. One male and one female specimens photographed (Figs. 9, 10): Leptopsalis beccarii Thorell, 1882 Probably type series Sumatra, Mt. Singalang leg. O. Beccari VII-1878 Two male specimens whole, one photo- graphed (Fig. 11). Two male specimens dissected, one with legs and _ chelicerae available, photographed (Fig. 12): Leptopsalis beccarii Thorell, 1882 Cotypes Sumatra leg. O. Beccari A male specimen collected by Elio Modigliani in Sipor a Island (off the coast of Camas) in 1891 was erroneously identified by Carl Friedrich Roewer as L. beccarii in 1935 (at that time recombined in a : Figure 13. An undescribed male from Sipora Island (near Sumatra), collected in 1891, showing dorsal (A), ventral (B), and lateral (C) views. Stylocellus). It is clearly affiliated with other large Sumatran Leptopsalis, of which we have seen several undescribed species, but it is distinctly larger than the lectotype Ole, ea beccarii. Specimen photographed ( Fie: 13): Stylocellus beccarii Hansen & Sor. Mentawei: Sereinu Leg. E. Modigliani, 1891 Det. Roewer, 1935-N. 10202 Bulletin of the Museum of Comparative Zoology, Vol. 160, No. 5 Figure 14. Leptopsalis javana Thorell, 1882, male holotype in dorsal (A), ventral (B), and lateral (C) views. Leptopsalis javana Thorell, 1882 Figures 14, 15 Leptopsalis javana was described by Thorell (1882) along with L. beccarii on the basis of a single male specimen from Tcibodas (Cibodas), Java, collected by Odoardo Beccari. We collected what appears to be L. javana at the Cibodas Botanical Garden, at the base of Gunung Gedé-Pangrango N.P., in Java. This was included in recent studies of stylocellid phylogenetics as “Java sp. LO” CyYPHOPHTHALMI TYPES FROM Museo Civico pi StorIA NATURALE “GIACOMO Dorit” ¢ Clouse and Giribet 253 Figure 15. Leptopsalis javana Thorell, 1882, male holotype: detailed views of the anal plate and surrounding ventral posterior region (A), distal parts of chelicerae, lateral view (B), and gonostome (C). (Clouse et al., 2009: Clouse and Géiribet, 2010). A BMNH specimen from Tcibodas that we have examined is identified as Stylocellus javanus but appears to be misidentified; it is actually more similar to other large species found in western Java and recently included in our molecular phylogenetic analyses. Male specimen photographed (Figs. 14, 15): Figure 16. Leptopsalis thorellii (Hansen & Sorensen, 1904), female holotype in dorsal (A), ventral (B), and lateral (C) views. Leptopsalis javana Thorell, 1882 Type Giava, Tcibodas leg. O. Beccari, X-1874 Leptopsalis thorellii (Hansen & Sorensen, 1904) Figures 16, 17 Stylocellus thorellii was described by Hansen and Sorensen (1904) on the basis of a single female specimen from Mt. Singalang (current spelling: Singgalang), Sumatra, Indonesia. It was also originally 254 Bulletin of the Museum of Comparative Zoology, Vol. 160, No. 5 epee thorellii (Hansen & Sorensen, 1904), female holotype: detailed views of the anterior lateral region (A), gonostome (B), and chelicerae (C). Figure 17. included in the type series that was used for the description of the very similar L. beccarii (1882), although phylogenetically these two species have been recovered in different Leptopsalis clades (Clouse et al., 2009), highlighting the uncertainties of morpholog- cal characters in this lar ge genus. Stylocellus thorellii Hansen & Sor., Type: Female holotype 1904 Figure 18. Leptopsalis cf. weberii (Hansen & Sorensen, 1904), male in dorsal (A), ventral (B), and lateral (C) views. Sumatra, Mt. Singalang leg. O. Beccari, VIII-1878 (Stylocellus beccarii Thor. partim) Leptopsalis cf. weberii (Hansen & So@ren- sen, 1904) Figures 18, 19 Stylocellus weberii was described by Hansen and S¢rensen (1904) on the basis of a single male See aa collected by Max Weber at Manindjau (Lake Maninjau), Sumatera Barat, Sumatra, Indonesia. The MCSN contains one male and two females from a different locality on Sumatra, iden- tified as S. weberii, although these speci- mens have not been contrasted with the CyYPHOPHTHALMI TYPES FROM MUSEO Crvico bI STORIA NATU bo UI Ul RALE “GIACOMO Doria” ¢ Clouse and Giribet Figure 19. Leptopsalis cf. weberii (Hansen & Sorensen, 1904), male showing detailed views of the anal plate (A), tarsus IV (B), chelicerae (C), gonostome (D), and lateral anterior part of body (E). type specimen from Maninjau, supposedly deposited in the Zoological Museum Am- sterdam (Giribet, 2000). Specimen photographed (Figs. 18, 19): Stylocellus weberi Hansen & Sor. One male, two females leg. E. Modigliani, 1891 Sumatra: Pangherang Det: Roewer, 1935-—N. 10201 FINAL REMARKS We provide here bibliogr aphic and taxo- nomic details as well as high-definition photographs of important type specimens in the Opiliones suborder Cyphophthalmi deposited in the MCSN. These specimens are of special importance due to their constituting the type species of three genera (in the (aaailies Neogoveidae and Stylocelli- dae), having primary types of five species, and being among the oldest cyphophthal- mid specimens neon: The study of this material allows us to confirm the validity of the genus Miopsalis, considered a nomen Abn in the most recent catalogue of Cyphophthalmi, and confidently re-erect the genus Leptopsalis, in synonymy with Stylocellus for over a century. Several 256 described species are here assigned to both genera. Finally, this article aims at showing dhe importance of natural history collections for continuing the painstaking task of describing and documenting the biological diversity of our planet and how this work cannot progress in many taxonomic groups until type specimens are made available to the community. ACKNOWLEDGMENTS Maria Tavano kindly sent us the MCSN specimens and gr aciously extended the loan a couple of times. Peter Schwendinger ne in securing the loan of the Genoa specimens, and his and an anonymous reviewers comments greatly improved an earlier version of this paper. LITERATURE CITED Boyer, 9. Li, Ry Me Grouse= Es Ro BeNAvIpES. PB: SHARMA, P. J. SCHWENDINGER, I. KARUNARATHNA, and G. GiriBet. 2007. Biogeography of the world: a case study from cyphophthalmid Opiliones, a globally distributed group of arachnids. Journal of Biogeography 34; 2070-2085. Boyer, S. L., and G. Giripet. 2007. A new model Gondwanan taxon: systematics and biogeography of the harvestman family Pettalidae (Arachnida, Opiliones, Cyphophthalmi), with a taxonomic revision of genera from Australia and New Zealand. Cladistics 23: 337-361. Boyer, S. L., I. KARAMAN, and G. Giripet. 2005. The genus Cyphophthalmus (Arachnida, Opiliones, Cyphophthalmi) in Europe: a phylogenetic ap- proach to Balkan Peninsula biogeography. Molec- ular Phylogenetics and Evolution 36: 554-567. Ciouse, R. M. 2010. Molecular and Morphometric Phylogenetics and Biogeography of a Southeast Asian Arachnid Family (Opiliones, Cyphophthalmi, Stylocellidae). Ph.D. Thesis. Pp. 121. Cambridge, Massachusetts: Department of Organismic antl Evolutionary Biology. Harvard University. Cuouse, R. M., B. L. DE Bivort, and G. Giriset. 2009. A phylogenetic analysis for the Southeast Asian mite harvestman family Stylocellidae (Opiliones : Cyphophthalmi)—a combined analysis using mor- phometric and molecular data. Invertebrate Sys- tematics 23: 515-529. CLousE, R. M., D. M. GENERAL, A. C. DiEsmos, and G. GiriBeT. 2011. An old lineage of Cyphophthalmi (Opiliones) discovered on Mindanao highlights the need for biogeographical research in Hye Philip- pines. Journal of Arachnology 39: 147-153. Crouse, R. M., and G. GiriBer. 2007. Across Lydekker’s Line—first report of mite harvestmen (Opiliones : Bulletin of the Museum of Comparative Zoology, Vol. 160, No. 5 Cyphophthalmi : Stylocellidae) from New Guinea. Invertebrate Systematics 2 207-2277 Ciousg, R. M., and G. Grripet. 2010. When Thailand was an ag nndailhe phylogeny and biogeography of mite harvestmen (Opiliones, Cyphophthalmi, Sty- locellidae) in Southeast Asia. Journal of Biogeog- raphy 37: 1114-1130. DE Brvort, B. L., and G. Giriper. 2004. A new genus of cyphophthalmid from the Iberian Peninsula with a phylogenetic analysis of the Sironidae (Arachnida : Opiliones : Cyphophthalmi) ) and a SEM database of external morphology. Invertebrate Systematics Sia GiriBET, G. 2000. Catalogue of the Cyphophthalmi of the World (Arachnida, Opiliones). Revista Ibérica de Aracnologta 2: 49-76. GirIBET, G. 2002. Stylocellus ramblae, a new stylocellid (Opiliones, Cyphophthalmi) from Singapore, with a discussion of the family Stylocellidae. Journal of Arachnology 30: 1-9. GiriBET, G., and S. L. Boyer. 2002. A cladistic analysis of the cyphophthalmid genera (Opiliones, Cy- phophthalmi). Jowrnal of Arachnology 30: 110- 128: GirIBET, G., and C. E. Prieto. 2003. A new Afrotropical Ogovea (Opiliones, Cyphophthalmi) from Camer- oon, with a discussion on the taxonomic characters in the family Ogoveidae. Zootaxa 329: 1-15. GIRIBET, G., L. Voct, A. PEREZ GONZALEZ, P. SHARMA, and A. B. Kury. 2010. A multilocus approach to harvestman (Arachnida: Opiliones) phylogeny with emphasis on biogeography and the systematics of Laniatores. Cladistics 26: 408-437. GIRIBET, G., P. P. SHARMA, L. R. BENAVIDES, S. L. BOYER, R. M. Crouse, B. L. DE Bivort, D. Dimitrov, G. Y. Kawaucul, J. Y. MuRIENNE, and P. J. SCHWENDIN- GER. 2011. Evolutionary and biogeographic history of an ancient and global group of ar achnids (Arachnida, Opiliones, Cyphophthalmi) with a new taxonomic arrangement. Biological Journal of the Linnean Society 105: 92-130. Hansen, H. J. 1921. Studies on Arthropoda I. The Pedipalpi, Ricinulei, and Opiliones (exc. Op. Laniatores) collected by Mr. Leonardo Fea in tropical West Africa and adjacent Islands. Copen- hagen: Gyldendalske Boghandel. Hansen, H. J., and W. SORENSEN. 1904. On Two Orders of Arachnida: Opiliones, gaan the Suborder Cyphophthalmi, and Ricinulei, Namely the Family Cryptostemmatoidae. Cambridge, United King- dom: Cambridge University Press. JUBERTHIE, C. 1970. Sur Suzukielus sauteri. (Roewer, 1916) opilion cyphophthalme du Japon. Revue d ‘Ecologie et de Biologie du Sol 7: 563-569. Kury, A. B. 2010. Opilionological Record—a chronicle of harvestman taxonomy. Part 1: 1758-1804. Journal of Arachnology 38: 521-529. Lecc, G. 1990. Parogovia pabsgarnoni, sp. n. (Arach- nida, Opiliones, Cyphophthalmi) from Sierra Leone, with notes on other African species of CYPHOPHTHALMI TYPES FROM MUSEO Civico DiI STORIA NATURALE ~ Parogovia. Bulletin of the British Arachnological Society 8: 113-121. RopMan, J. E., and J. H. Copy. 2003. The taxonomic impediment overcome: NSF's partnerships for enhancing expertise in taxonomy (PEET) as a model. Syste matic Biology 52: 428-435. Roewer, C. F. 1916. 7 neue Opilioniden des Zoolog. Museums in Berlin. Archiv ftir Naturgeschichte 81: 6-13. RoEewer, C. F. 1923. Die Weberknechte der Erde. Systematische Bearbeitung der bisher bekannten Opiliones. Jena: Verlag von Gustav Fisher. Rosas Costa, J. A. 1950. Sinopsis de los géneros de Sironinae, con la descripcion de dos géneros y una especie nuevos (Opiliones, Cyphophthalmi), Ar- thropoda 1: 127-151. SCHWENDINGER, P. J., and G. Giripet. 2005. The systematics of the Southeast Asian genus Fangensis Rambla (Opiliones : Cyphophthalmi_: Stylocelli- dae). Invertebrate Systematics 19: 297-323. Suear, W. A. 1980. A review of the Cyphophthalmi of the United States and Mexico, with a proposed GiAcoMo Doria” ¢ Clouse and Giribet OD reclassification of the suborder (Arachnida, Opi- liones). American Museum Novitates 2705; 1-34. SHEAR, W. x 1993. New species in the opilionid genus Stylocellus from Malaysia, Indonesia and the Phili )- pines (Opiliones, ¢ ‘yphophthalmi, Stylocellidae). Bul- letin of the British Arachnological Society 9: 174-188. SIMON, E. 1872. Notice sur les Arachiide 5 cavernicoles et hypogés. Annales de la Société entomologique de France, 5e série 2: 215-244. SuarEz, A. V., and N. D. Tsutsur. 2004. The value of museum collections for research and_ society. BioScience 54: 66-74. THORELL, T. 1882. Descrizione di alcuni Aracnidi inferiori dell’ Arcipelago Malese. Annali del Museo civico di Storia naturale di Genova 18: 21-69. THorELL, T. 1890a. Aracnidi di Pinang raccolti nel 1889 dai Sig.ri L. Loria e L. Fea. Annali del Museo civico di Storia naturale di Genova (Ser. 2a) 10: 269-383. THoRELL, T. 1890b. Aracnidi di Nias e di Sumatra raccolti nel 1886 dal Sig. E. Mondigliani. Annali del Museo civico di Storia naturale di Genova (Ser. 2a) 10: 5-106. re <) ide es ty £10’ oly : e@ ad 0 anes - Vee ao] Yaa ey 7 ie , ice @ @) — L 4 2 * 4 4 , ~y od . é ‘. US ISSN 0027-4100 MCZ Publications Museum of Comparative Zoology Harvard University 26 Oxford Street Cambridge, MA 02138 mczpublications@mcz.harvard.edu © The President and Fellows of Harvard College 2012. Bulletin of the Museum of Comparative Zoology Volume 160, Number 6 30 November 2012 Systematics of the Neotropical Snake Dendrophidion percarinatum (Serpentes: Colubridae), with Descriptions of Two New Species From Western Colombia and Ecuador and Supplementary Data on D. brunneum JOHN E. CADLE HARVARD UNIVERSITY | CAMBRIDGE, MASSACHUSETTS, U.S.A. BULLETIN OF THE Museum of Comparative Zoology BOARD OF EDITORS Editor: Jonathan Losos Managing Editor: Deborah Smiley Associate Editors: Andrew Biewener, Scott Edwards, Brian Farrell, Farish Jenkins, George Lauder, Gonzalo Giribet, Hopi Hoekstra, Jim Hanken, Jim McCarthy, Naomi Pierce, and Robert Woollacott Publications Issued or Distributed by the Museum of Comparative Zoology Harvard University Bulietin 1863- Breviora 1952- Memoirs 1865-1938 Johnsonia, Department of Mollusks, 1941-1974 Occasional Papers on Mollusks, 1945- General queries, questions about author guidelines, or permissions for MCZ Publications should be directed to the editorial assistant: MCZ Publications Museum of Comparative Zoology Harvard University 26 Oxford Street Cambridge, MA 02138 mczpublications@mcz.harvard.edu EXCHANGES AND REPRINTS All of our publications are offered for free on our website: http://www.mcz.harvard.edu/Publications/index.html To purchase individual reprints or to join our exchange program, | please contact April Mullins at the Ernst Mayr Library: mayrlib@oeb.harvard.edu. | I This publication has been printed on acid-free permanent paper stock. © The President and Fellows of Harvard College 2012. SYSTEMATICS OF THE NEOTROPICAL SNAKE DENDROPHIDION PERCARINATUM (SERPENTES: COLUBRIDAE), WITH DESCRIPTIONS OF TWO NEW SPECIES FROM WESTERN COLOMBIA AND ECUADOR AND SUPPLEMENTARY DATA ON D. BRUNNEUM JOHN E. CADLE' CONTENTS Abstract 259 Resumen 260 Introduction 260 Material and Methods 261 Taxonomic Characters of Special Relevance to the Dendrophidion percarinatum Complex 262 Color pattern 262 Supralabial/temporal pattern 265 Hemipenial characters 266 Redefinition and Description of Dendrophidion percarinatum (Cope) 266 Two New Species from Western Colombia and Ecuador 282 Dendrophidion prolixum new species 282 Dendrophidion graciliverpa new species 296 Application of the Name Dendrophidion brunneum (Giinther) and New Data from Western Ecuador 305 Hemipenial Morphology in the Dendrophidion percarinatum Complex oL7 Dendrophidion percarinatum 319 Dendrophidion prolixum 325 Dendrophidion graciliverpa 328 Comparisons of Hemipenial Morphology of Species in the Dendrophidion percarinatum Complex 331 Concluding Remarks 334 Acknowledgments 335 Appendix 1. Specimens Examined and Literature Records of Dendrophidion percarinatum and New Records of D. brunneum from Ecuador Boo Appendix 2. Gazetteer (Dendrophidion prolixum and D. graciliverpa localities) 338 Literature Cited 340 ' Department of Herpetology, California Academy of Sciences, 55 Music Concourse Drive, Golden Gate Park, San Francisco, California 94118 (jcadle@calac ademy.org). Bull. Mus. Comp. Zool., Asstract. Dendrophidion percarinatum (Cope) is redefined on the basis of standard and new characters to distinguish it from two new South American species with ahh it has previously been confused. The redefined D. percarinatum is distributed from Hon- duras through Central America to western Colombia, with a seemingly outlying locality in extreme western Venezuela. One new species, D. prolixum, is sympatric with D. percarinatum at a few localities in central western Colombia and the distribution of the new species continues southward into northwestern Ecua- dor. A second new species, D. graciliverpa, occurs throughout western Ecuador, where its distribution extensive overlaps that of D. brunneum (Giinther) Hemipenes of the two new species are unusually long and slender (gracile morphotype), a morphology distinct from other described Dendrophidion hemi- penes, which are shorter and more robust (robust morphotype). Additionally, the new species differ from D. percarinatum in color patterns but not in standard scutellation characters such as segmental counts. Similarly, the two new species differ from one another in coloration but not in scutellation or hemipenial morphology. Hemipenes of D. percarinatum and the new species are described in detail. The holotype of D. brunneum is redescribed to ensure the proper application of that name. New specimens document the widespread occurrence of D. brunneum in the lowlands of western Ecuador and apparent extensive pattern polymorphism, including unicolor, striped, crossbanded, and punctate forms; more data on coloration in life are needed. Some previous records of “D. percarinatum” from interandean valleys of the Rio Cauca/Magdalena system are from mistaken identities. However, several specimens from the Rio Magdalena resemble D. percarinatum in scutellation but differ in color pattern; their status needs further study. Key words: Colombia, Ecuador, systematics, Choco, Central America, South America, Dendrophidion, snakes, new species, hemipenis, morphology 160(6): 259-344, November, 2012 259 260 ResuMEN. Dendrophidion percarinatum (Cope) se redefine sobre la base de caracteres estandarizados y nuevos para distinguirlo de dos nuevas especies sudamericanos con que se confundieron anterior- mente. El redefinido D. percarinatum se encuentra desde Honduras por América Central hasta el oeste de Colombia, con una localidad aparentamente alejada en el extremo occidental de Venezuela. Una nueva especie, D. prolixum, es simpatrica con D. percarina- tum a pocas localidades en el centro-occidental de Colombia; la distribucion de la nueva especie continua hacia el sur hasta el noroeste de Ecuador. Una segunda nueva especie, D. graciliverpa, occurre en todo del Ecuador occidental, donde su distribuci6n traslapa la distribuci6n de D. brunneum (Giinther). Los hemi- penes de ambas nuevas especies son exceptionalmente alargados y delgados (morfotipo esbelto). Es una morfologia distinta de los otros hemipenes de Den- drophidion, que son mas corto y robusto (morfotipo robusto). Ademas, las nuevas especies se distinguen de D. percarinatum por la coloraci6n pero no se distinguen en caracteres estandarizados de escutella- cidn como cuentas segmentales. De modo parecido, las dos nuevas especies se distinguen por coloracion pero no por la escutellacién ni la morfologia de los hemipenes. Se describen los hemipenes de D. percar- inatum y las nuevas especies. Para asegurar la aplicacion apropiada del nombre D. brunneum se redescribe el holotipo de esta especie. Nuevas especi- menes documentan la ocurrencia amplia de D. brunneum en las tierras bajas del Ecuador occidental y también polimorfismo extenso de patrones, in- cluyendo patrones unicolor, rayado, con bandas, y con manchas:; se necesitan mas datos sobre la colora- cidn de vida. Algunos registros anteriores de “D. percarinatum” desde los valles interandinos de la sistema Rios Cauca/Magdalena son de identidades equivocadas. Sin embargo, existen especimenes del Rio Magdalena que son parecidos a D. percarinatum en la escutellacion pero son distintos en coloracién; su estatus merece mas estudio. INTRODUCTION At the time of his death in 1972, James A. Peters had been accumulating data in preparation for a revision of the Neotropical snake genus Dendrophidion Fitzinger and he indicated that the account given by Peters and Orejas-Miranda (1970: 79) would “be rather thoroughly changed upon completion of the review.” Forty years hence there still has been no comprehensive review of Dendrophidion, although various workers have chipped away at particular geographic or taxonomic segments of the genus. Lieb (1988), working in part from Peters’ unpublished data, resurrected D. Bulletin of the Museum of Comparative Zoology, Vol. 160, No. 6 nuchale (W. Peters) from the synonymy of D. dendrophis (Schlegel), stabilized the application of the last name by designating a lectotype for it, and clarified the applica- tion of other names in the much-confused literature on these snakes. Lieb (1988) also briefly summarized characters and distribu- tions for the nine species of Dendrophidion he recognized and apportioned eight of them between two species groups. McCra- nie (2011) resurrected the name D. clarkii Dunn for application to Central American members of the complex including clarkii and nuchale. Cadle (2010) reviewed the systematics, natural history, and hemipenial morphology of D. brunneum based on new material and field observations, and Cadle (2012) partitioned D. vinitor by describin two new cryptic species previously conti with it. Freire et al. (2010) described another new species of the D. dendrophis group from northeastern Brazil. Other taxonomic issues still need resolution. This paper focuses on Dendrophidion percarinatum, which according to current taxonomy is distributed from Honduras to Ecuador and Venezuela. During my review of D. brunneum (Cadle, 2010), I had examined material referred to “D. percar- inatum” from western Colombia and Ecua- dor and indicated that revision of D. percarinatum appeared necessary (Cadle, 2010: 2, 24). Further investigation has confirmed that suggestion and has caused me to revise some of my earlier assessments concerning D. brunneum. In particular, I came to realize that most specimens previ- ously referred to “D. percarinatum” in western Colombia and Ecuador were, in fact, distinct species. Dendrophidion per- carinatum is widespread in Central Amer- ica, but other than a few localities around the Golfo de Uraba in northern Colombia, it is definitely known from only a handful of localities in the Choco region and one in extreme western Venezuela. In this paper I review the systematics of Dendrophidion percarinatum with special reference to populations in western Colom- bia and Ecuador. Two species previously SYSTEMATICS OF DENDROPHIDION PERCARINATUM (COLUBRIDAE) confused with D. percarinatum from the last two areas are described as new. Hemi- penes of the three species are described. The review of western Ecuadorian speci- mens of Dendrophidion also resulted in the discovery that the distribution of D. brun- neum includes previously unrecognized lowland localities and broadly overlaps the distribution of one of the new species in western Ecuador. The holotype and new material of D. brunneum are reviewed to supplement the earlier account (Cadle, 2010) and to ensure proper application of this name. MATERIALS AND METHODS Methods follow procedures described in Cadle (2005, 2007, 2012). Methods for scoring dorsocaudal reductions, dorsal keel- ing, aspects of color pattern, and statistical methods are described by Cadle (2012) for other Dendrophidion. Additional comments on characters especially relevant to D. percarinatum and the new species de- scribed herein are detailed below. Measure- ments (in mm) of hemipenes and for a loreal scale character in the section on Dendro- phidion brunneum were made with Helios dial calipers to the nearest 0.01 mm under a dissecting microscope. For purposes of analyzing intraspecific differences in mean SVL of adult males and females, specimens >400 mm snout to vent length (SVL) were considered adults because sexual matura- tion occurs at 400-500 mm SVL (Goldberg, 2003; Stafford, 2003). Similarly, because relative tail length increases proportionally with SVL, the range of adult relative tail length (RTL) was assessed for individuals with SVL > 300 mm because RTL ap- pr oaches an asymptote at approximately this size. Methods of hemipenial study are covered in the introduction to that section. My sampling of Dendrophidion percar- inatum has not been even from throughout the range. Sample sizes of D. percarinatum as redefined here from throughout the range are as follows (number of males, number of females): Honduras (10, 12), ° Cadle 261 Nicaragua: ,(3;, 3), Costa. Rica (29, 26), Panama (56, 37), Colombia (7, 4). I also examined the lectotype of D. percarinatum and the holotype and additional eee of D. brunneum. Hemipenial morphology offered some early insights into the system- atics of these species, and | examined everted hemipenes from throughout the range of D. percarinatum and the internal morphology of retracted organs of selected specimens. I also examined all available everted hemipenes of the two new species and the internal morphology of other retracted organs. Except where specifically qualified (e.g., “Dendrophidion percarinatum sensu Lieb 1988”), I use the name Dendrophidion percarinatum to refer to this snake as redefined herein, which excludes certain populations in western Colombia and Ecua- dor that have been included in other literature (e.g., Lieb, 1988, 1996; Savage, 2002: Stafford, 2003: Cadle. 2010). eee to Dy percarinatum plus the two new species described herein as the “Dendrophidion percarinatum complex” with no assumption that the complex is monophyletic within Dendrophidion. These species plus D. brunneum and D. paucicarinatum comprise the D. percarinatum species group of Lieb (1988). Unresolved systematic issues pertain to other species names within Dendrophi- dion, and I use the name “D. nuchale auctorum” for the complex of species represented by this name, as explained in Cadle (2012: 188; see also Savage, 2002: 655: McCranie, 2011: LO6—107). Appendix 1 lists specimens examined of Dendrophidion percarinatum, D. brunneum (Ecuadorian specimens identified subse- quent to Cadle [2010] only), and several specimens of uncertain status (D. species inquirendum). Specimens of the new spe- cies are listed in the text. The accounts of new species include some specimens listed as “other referred specimens” 1 rather than paratypes. These specimens are all juvenile or adult females, which are sometimes difficult to identify from preserved speci- mens (depending on state of preservation 262 and/or characters of individual specimens). These difficulties are discussed in the species accounts. Scutellation data from the referred specimens were included in the data summaries except as noted (e.g., Table 1). Appendix 2 contains coordinates and notes on the localities for the new Species. Coordinates, some elevations, and other information on localities were derived primarily from Brown (1941; Ecuador), Fairchild and Handley (1966; Panama), Paynter (1993, 1997: Ecuador and Colom- bia), Lynch and Duellman (1997; Ecuador), McCranie (2011: Honduras), and the National Geospatial-Intelligence Agency (NGA, 2010-2012) online gazetteer (GEO- net). There have been changes in Ecuador- ian administrative divisions associated with some well-known localities in western Ecuador (e.g., Santo Domingo de _ los Colorados). Two new provinces, Santo Domingo de los Tsachilas and Santa Elena, were created in 2007 from portions of Pichincha and Guayas provinces, respec- tively. I record both provinces for these localities to prevent confusion when cross- referencing with other taxonomic literature on this area. Table 1 presents summary taxonomic and morphological data for species of the Dendrophidion percarinatum complex. Da- ta summarized therein for D. percarinatum are for specimens from throughout the geographic range of this species (Honduras to western Colombia). Geographic variation for some characters of D. percarinatum is discussed in the species account (see also Tables 2 and 3). Hemipenial characters are discussed in the section devoted to them after the species accounts. TAXONOMIC CHARACTERS OF SPECIAL RELEVANCE TO THE DENDROPHIDION PERCARINATUM COMPLEX At the outset it is appropriate to discuss the primary character systems used in my assessments of species boundaries within the Dendrophidion percarinatum complex. Bulletin of the Museum of Comparative Zoology, Vol. 160, No. 6 Although many of the same character systems relevant to the systematics of the D. vinitor complex (Cadle, 2012) are pertinent to D, percarinatum, the informa- tive aspects of variation are sometimes a bit different. For example, color patterns are relevant to differentiating species in both groups, but the aspects of coloration of particular value in each group differ. The Piloane sections discuss variation in sev- eral character systems as they pertain to the D. percarinatum complex. Color Pattern. Dendrophidion percarina- tum as conceived in previous works is highly variable in coloration and pattern (Lieb, 1988, 1996; Savage, 2002). However, little attention has been given to the relationship of this variation to geography or its potential systematic implications. During my review it became clear that, although pattern varia- tion within D. percarinatum in Central America is comparatively minor, in northern South America several distinct color pat- terns were present. Further consideration suggested that these distinct color patterns had both geographic integrity (i.e., discrete geographic Tae and were correlat- ed in some cases with other characters (e.¢., hemipenial morphology) suggestive of sys- tematic distinction. Thus, color pattern variation played a significant role in my assessment of species boundaries within the D. percarinatum complex, particularly in the absence of strong differentiation using standard scale characters (Lieb, 1988; Cadle, 2010, 2012: documented herein). Unfortu- nately, coloration in life is available for relatively few specimens. Changes induced by preservation and perhaps differences in preservation techniques make it impossible to correlate preserved coloration to colors in life with any ease. Some pattern elements seemingly become more prominent upon preservation (e.g., pale crossbands) even as others become more obscure. Other factors, such as the integrity of the stratum comeum on a given specimen, also affect the appear- ance of preserved specimens. Narrow pale crossbands are expressed in some individuals of all species of the SYSTEMATICS OF DENDROPHIDION PERCARINATUM (COLUBRIDAE) * Cadle 263 TABLE 1. TAXONOMIC DATA FOR DENDROPHIDION PERCARINATUM (RANGEWIDE SUMMARY) AND TWO NEW SPECIES. BODY PROPORTIONS, SEGMENTAL COUNTS, SCALE REDUCTIONS, MAXILLARY TEETH, AND NUMBER OF PALE BANDS ARE GIVEN AS RANGE FOLLOWED BY MEAN * SD. BILATERAL COUNTS ARI SEPARATED BY A SLASH (/). FOR TEMPORALS AND LABIAL SCALES, EACH SIDE OF A SPECIMEN WAS COUNTED AS AN INDEPENDENT OBSERVATION. SAMPLE SIZES IN PARENTHESES. ASTERISKS INDICATE STATISTICAL SIGNIFICANCE OF INTRASPECIFIC DIFFERENCES BETWEEN MEANS OF MALE AND FEMALE SIZE, PROPORTIONS, OR MERISTIC COUNTS (* P < 0.05; ** p < 0.01; *** p < 0.001); NO ASTERISK INDICATES NONSIGNIFICANCE. Dendrophidion percarinatum (Cope) Rangewide Summary Largest specimens: total length, SVL (mm) Male 1,358+, 852 [1,236, 747]* Female 1,300, 778 Mean adult SVL (mm) Male 401-852 573 2 98.18 (73) Female 437-718 604.4 + 74.63 (51) Tail length/total length Male 0.38—0.45 0:42 = 0.019 (52)" Female 0.37-0.43 0.40 + 0.014 (24) Tail length/SVL Male ().62—0.82 0:72 = 0,054 62)" Female 0.59-0.75 0.68 + 0.039 (24) Maxillary teeth 35-42 37.1 2 1.84 (59) ii =l7—-15.( 1380) other patterns (4) Dorsal scales Ventrals Male 147-170 155.6 S503 * Female 156-167 160.2, = 2.42) (80) Subcaudals Male IS7Z163 L504 632K Female 133157 145.6 + 5.26 (46) Total segmental counts (ventrals + subcaudals) Male 993-323 306.6 + 6.19 (72) Female 993-321 305.9 + 5.60 (46) Dorsocaudal reduction, 8 to 6 (subcaudal number) Male 8-26 G08: *3.60:(103)" Female OA LO.4vss 3.02) (SI) Dorsal scales, posterior reduction (ventral number) Male 84-102 92.6 + 3.85 (88) Female 91-106 96.8 = 3.30 (54) Dendrophidion prolixum New Species 1,003+, 650 [1,037, 642]* 976+, 675 [1,116, 662]* 455-650 582.6 + 61.95 (10) 495-675 604.3 + 76.15 (15) 0.38—0.40 0.39 = 0.009 (5) 0.35—0.41 0.40 + 0.012 (6) 0.62—0.67 0.65 + 0.021 (3) 0.63—0.69 0.66 + 0.026 (6) 36-42 36.0 = 1:65 2) 17-17-15 (37) 150-164 Love? 2 A683. (17) 152-164 160.9 = 34270017) 134-150 140.5: 5.15 (11) 133-150 143.3 = 3.07 (9) 284-310 297.0 = 8:49:11) 285-313 303.4 + 7.89 (9) 16-26 19.8 + 3.05 (17)** 8-24 12.4 + 4.50 (17) 84-99 02, Gus AO, (32) 91-110 98.7 = 3.93 (32) Dendrophidion graciliverpa New Species 1,054+, 676 [964, 605]' 922+, 663 [1,027, 631]! 409-605 517.5 2.58.30 (6) 543-663 615.5 + 37.92 °C) 0.37-0.42 0.39 + 0.0002 (9)* 0.36—0.39 0.37 + 0.014 (6) 0.59-0.72 0.65 + 0.038 (9)* 0.56—0.64 0.59 + 0.035 (6) 33-44 38.0 = 240.(23) Vele15 27) (2 17-19-17 (1 i ) 153-163 157.5 + 2.15 (22) 152-166 160.7 + 3.37 (23) 132-153 142.32 6AL (12)" 120-143 133.0 2. 7.32 (13) 290-309 299.3 = 7.01 (12) 2'75=304 293.0 + 9.42 (12) —19 12.02 3.56123) ~l 78-95 90:38. 3.73: (30)"" 90-101 94.9 + 9.59 (24) 264 Bulletin of the Museum of Comparative Zoology, Vol. 160, No. 6 TABLE 1. CONTINUED. Dendrophidion percarinatum Dendrophidion prolixum Dendrophidion graciliverpa (Cope) Rangewide Summary New Species New Species Preoculars 1/1 (186) 1/1 (34) 1/1 (43) 1/2 (2) Postoculars IVAuah) DZS?) PQ) W222) Wifey (4) 2/3) (2) APA cy”) 3/3 (4) Do (2) o/o UL) Primary temporals 1 (2) 1, (G0) 2, (84) 2 (370) D6) 3 (2) 3) (UL) Secondary Gb) PEGs) 2 (85) temporals 2 (362) onc) Gy) 3 (3) Supralabials, Siro—o aly) 9, 4—6 (67) oy, C=» (D) supralabials 8 4=51(0) WO, 4D) oy 25) (0) touching eye 8, 4-6 (2) OVA (2) 9, 4-6 (363) 9, 4-6 (76) [0 5=77Q9) Infralabials 8 (1) 8 (3) 8 (3) 9 (17) 9 (16) 9x3) 10) (PAIL) 10 (28) 10 (54) L332) 11 (20) a) RPA) (as) IPA) Supralabial/ G, 6.8% G, 76% G, 66.7% temporal P, 82.8% PaAD% P, 2.9% pattern irregular/ambiguous, 10.4% irregular/ambiguous, 19.4% irregular/ambiguous, 30.4% (221) (67) (69) No. of pale bands 71-96 49-57 Som on body S12 = 6:87 (45) 53.0 tee Or 12) HOD 22 8.26) Hemipenial robust morphotype gracile morphotype gracile morphotype characters < 40 enlarged spines > 60 enlarged spines > 60 enlarged spines + The largest specimens in these cases had incomplete tails. Measurements of the largest specimens with complete tails in brackets. Dendrophidion percarinatum complex. Ju- helpful when the bands are evident on the veniles have more distinct crossbands than neck. In these cases, the counts were made adults, in which the bands may be either — as described in Cadle (2012) for scoring the distinct or indistinct. The number of pale width of pale bands in the D. vinitor crossbands on the body (neck to vent) is complex (except that rows between bands, useful in distinguishing one of the new rather than rows encompassed by bands, species described herein compared with the | were counted). Neck bands can sometimes other two members of the complex. How- be discerned under magnification and good ever, because crossbands are not always lighting even on excessively darkened pre- evident over the entire body in some — served specimens (the dark borders to the specimens of all three species, this character bands are often better clues than the pale cannot always be scored. The width of the band itself); such observations are best pale bands is not useful for distinguishing made with the specimen submerged in the two new species from one another or — alcohol. These comments apply to preserved from D. percarinatum (the bands are museum specimens, but a caveat is that typically one dorsal row or less in width in observations presented herein for the two all three species), but the number of scale — new species suggest that the appearance rows separating pale bands on the neck is of pale crossbands can be enhanced in SYSTEMATICS OF DENDROPHIDION PERCARINATUM (COLUBRIDAE) preserved specimens compared with the live snakes. Other elements of pattern are highly variable within and between spe cies. The venter of the two new species fa the Dendrophidion percarinatum complex de- velops narrow dark transverse lines across the anterior edge of each ventral scale (a character also found in some other species of Dendrophidion). There is seemingly an ontogenetic component to this variation: larger individuals typically have more full developed lines in those species in act they occur. Nonetheless, some juveniles have ventral lines, and considerable varia- tion in the prominence of the lines exists in adults as well. These lines seem to develop first on the posterior venter and eventually can encompass nearly all the ventral scutes. Superticially, these lines might appear to be on the posterior edges of the ventral scutes, but close examination shows that they are on the anterior edges of the scales and merely show through the nearly transparent posterior edge of the adjacent anterior scute. I also examined other aspects of color pattern, such as the dark dorsal longitudinal stripes in the dorsolateral region and flanks seen in many specimens of Dendrophidion percarinatum. There is considerable varia- tion in these features, and I did not find them useful for discriminating the three species of the D. percarinatum complex. I refer to some other coloration features, particularly with reference to D. brunneum. Supralabial/Temporal Pattern. The ar- rangement of scales on the lateral surfaces of the head behind the eye is a useful aid in distinguishing species of the Dendrophi- dion percarinatum complex (perhaps more broadly within Dendrophidion, but this needs further study). I stress useful aid because it is not an infallible character. Some cases are ambiguous because the configuration of these sence is influenced by irregularities such as scale fusions, divisions, or other anomalous patterns. In other cases the patterns are downright misleading (i.e., some specimens have the © Cadle 965 atypical arrangement for the taxon to which I refer them). For this reason I do not include the supralabial patterns in the list of characters at the beginning of the diagnoses, although I do use them for comparative purposes when they seem useful. The relevant scales are the lower primary and secondary temporals and their relationship to the penultimate and ultimate supralabials. I distinguish two basic configurations (Fig. 1), which were scored on both sides of a sampling of each species. Frequencies of the patterns are given in Table 1. Early in my aa on the percarinatum complex I desig- nated these patterns as P and G for D. percarinatum and D. graciliverpa, the first pair of species that I distinguished by this character. Atypical or ambiguous configura- tions were coded separ ately from the basic configurations. P-Pattern or percarinatum Pattern (Fig. 1A). (As a mnemonic, P could also refer to the Bore shape of the penultimate supralabial in this pattern.) Penultimate supralabial (nearly always the eighth) broad- ly contacts the lower secondary temporal, separating the lower primary temporal from the ultimate (ninth) supralabial. The penul- timate supralabial is relatively Al and usually in the shape of an irregular penta- gon, with the vertical dimension of the posterior part greater than that of the anterior part. The posterior border of the penultimate supralabial is usually even with, or posterior to, the posterior border of the lower primary temporal. G-Pattern or gr aciliverpa ee (Fig. 1B). (As a mnemonic, the: first “yin graciliverpa could also refer to the eerie shape of the penultimate supralabial in this pattern.) Penultimate supralabial (nearly always the eighth) separated from the lower secondary tempor al by contact between the lower primary temporal and the ultimate (ninth) supralabial. The penultimate supralabial is usually relatively narrow and homogeneous in vertical dimension and in the shape of an elongate rectangle (sometimes squarish). The posterior hordes of the penultimate 266 supralabial is usually anterior to the posterior border of the lower primary temporal. Hemipenial Characters. Detailed descrip- tions of hemipenes of the species described herein are presented in a separate section at the end. However, hemipenial characters play a fundamental role in the recognition of the new species described here and in the application of the name Dendrophidion brunneum (Giinther) discussed later. Thus, a few comments on hemipenial length and overall shape are pertinent here. Cadle (2012: 217-220) provided a general overview of Dendrophidion hemipenes. Detailed de- scriptions and discussion of morphological variation of hemipenes in other species are in Cadle (2010: 14-20; 2012: 220-228). Most Dendrophidion hemipenes are of what may be termed the “robust (or compact) morphotype” ( ae 2). Hemipenes of this form are relatively short. The narrow proximal portion of the hemipenial body bearing minute spines is short, and the distal section bearing enlarged spines, calyces and/or flounces, and other apical ornamentation is greatly expanded. The distal bulbous section comprises half or more of the length of ee organs. In contrast, the two new species de- scribed herein are characterized a hemi- penial morphology that may be called “gracile morphotype” (Fig. 2). Compared with hemipenes of other species of Den- drophidion with the possible exception of D. bivittatum, gracile hemipenes are exception- ally long. A long, slender proximal section of the hemipenial body bears minute spines, distal to which is a moderately bulbous region bearing enlarged spines, calyces and/or flounces, and other apical ornamentation. The bulbous region of gracile hemipenes is not expanded to the extent that it is in robust organs, and it comprises much less than half of the total length of everted organs. Den- drophidion bivittatum seemingly has gracile hemipenial morphology, but I have seen only one somewhat desiccated everted organ of this species; its proportions in retracted organs is similar to the gracile hemipenes described here, though perhaps not quite as Bulletin of the Museum of Comparative Zoology, Vol. 160, No. 6 long as in the two species described herein (see Stuart, 1932: pl. I, fig. 2 for an illustration of a retracted organ of D. bivittatum). The compact and gracile morphologies are easily discerned in everted organs (Fig. 2). However, in retracted hemipenes, the overall length of the organ ne of subcaudals subtended) is a reliable indicator of the morphology, even when the internal morphology is not examined and despite some variation in the length of retracted organs (detailed in the section on hemi- penial morphology). “Robust” organs gen- erally subtend fewer than 10 subcaudals, whereas “gracile” organs usually subtend 10 or more subcaudals (up to 15 in the specimens I examined). If the internal morphology of retracted hemipenes is examined, the relative proportions of the relatively unadorned base compared with the distal section differentiate the two morphotypes (Fig. 3). For this purpose the distal section comprises the section of enlarged spines (using the most proximal sulcate enlarged spine as a landmark) + flounces/calyces + apical region. The bul- bous distal section comprises a much greater proportion of the length of retracted robust organs compared with gracile organs. REDEFINITION AND DESCRIPTION OF DENDROPHIDION PERCARINATUM (COPE) Figures 1A, 1C, 2, 4-8, 12A, 13A, 19, 38-41 Drymobius percarinatus Cope “1893” (1894): 344 (two syntypes from Boruca and Buenos Ayres, Costa Rica); “1894” [1895]: 427; 1895: 205. Drymobius dendrophis: Boulenger, 1894: 16 (part). Gunther, 1885-1902: 127 (? part; Nicaragua, Costa Rica). Amaral “41929” [1930]: 154 (part). Dunn and Emlen, 1932: 31 (Honduras; ANSP 20817). Schmidt “1933” [1935]: 15 (Pan- ama; USNM 54080). Cacocalyx percarinatus: Cope “1894” (1895): 427; 1895: 205: pl-19), figu2: 1900: 778, 781 + pl. 17, fig. 2. SYSTEMATICS OF DENDROPHIDION PERCARINATUM (COLUBRIDAE) ¢ Cadle 267 A Percarinatum pattern (P) B Graciliverpa pattern (G) Figure 1. Contrasting configurations of scales in the posterior supralabial and lower temporal regions in the Dendrophidion percarinatum complex. (A) Stereotypical percarinatum (P) configuration (D. percarinatum, LACM 148558); (B) Stereotypical graciliverpa (G) configuration (D. graciliverpa, AMNH R-110585). Bottom panel shows other representations of the typical patterns: (C) D. percarinatum, AMNH R-17374, P pattern with lower secondary temporal divided vertically; (D) D. prolixum, AMNH R-109724, Gpattern; (E) D. graciliverpa, UIMNH 92244, G pattern. Panels C, D, and E are left sides reversed. Abbreviations and symbols: X, penultimate supralabial (nearly always the eighth); LPT, lower primary temporal. Arrows indicate the lack of contact between the lower primary temporal and ultimate supralabial in the P pattern, and contact between these two scales in the G pattern (arrows lie diagonally across the ultimate labials). Note contrasting shapes of the penultimate supralabial in the P (pentagonal) and G (rectangular) patterns. See text for further discussion and variation. Dendrophidion dendrophis: Gaige et al., 636.1 (part). Perez-Santos and Moreno, 1937: 12 (part). Nicéforo Maria, 1942: 87 (part; ?specimen from Sasaima; see com- ments under Distribution). Taylor, 1951: 92 (Costa Rica). ?Pérez-Santos and Moreno, 1988: 134 (part; Pacific lowlands of Co- lombia); 1991b: 135 (part; “lower Central America and northern South America’). Dendrophidion percarinatus. Stuart, 1932: 6. Smith, 1941; 1958: 223. Taylor, 1954: 727 (lectotype designated: “the adult specimen from Boruca” [= AMNH R-17366]). Smith and Grant, 1958: 210. Sexton and Heat- wole, 1965: 41. Peters and Orejas-Miran- da, 1970: 80. Pérez-Santos and Moreno, 1988: 135 (Colombia). Dendrophidion percarinatum Dunn, 1944: 477 (? part). Savage, 1973: 14; 1980: 92; 2002: 657 + pl. 418 (part). Wilson and Meyer, 1985: 41. Scott et al., 1983: 372. Savage and Villa, 1986: 148, 169. Villa et al., 1988: 63. Lieb, 1988: 172 (part); 1996: 1988: 135 (part); 1989: 3 (part); 1991a: 138 (part). Rand and Myers, 1990: 395. Ibanez and Solis, “1991” [1993]: 30, 33. Pérez-Santos et al., 1993: 116. Auth, 1994: 16. Guyer, 1994: 382. Ibanez et al, “1994” [1995]: 26. Pérez-Santos, 1999: 99, 102. Nicholson et al., 2000: 29. McCranie et al., 2002: 27; 2006: 147, 229, 237 (part). Goldberg, 2003. Kohler, 2003: 200; 2008: 215. Stafford, 2003 (part). Solorzano, 2004: 234-236 + fig. 59 (part). Guyer and Donnelly, 2005: 184 (part). McCranie and Castaneda, 2005: 8 (part). McDiarmid and Savage, 2005: 391, 421 (part). Wilson and Townsend, 2006: 96, 106; 2007: 136, 139. Rojas-Runjaic and Rivero, 2008. Santos-Barrera et al., 2008: 766. Laurencio and Malone, 2009: 126. McCranie, 2009: 12; 2011: 108. Savage and Bolanos, 2009: 14. Bolanos et al., 2010. Cadle, 2010 (part). 268 Bulletin of the Museum of Comparative Zoology, Vol. 160, No. 6 Figure 2. Comparison showing proportional differences between the “robust” (left) and “gracile” (right) hemipenial morphotypes within Dendrophidion. The robust hemipenis photograph was scaled to the same length as the gracile organ. Horizontal lines are placed at the level of the proximal enlarged sulcate spine and the distal flounce in each organ (assumed to be homologous morphological markers). The apical region so delimited is a much greater proportion of the hemipenial body and relatively broader in the robust hemipenis than in the gracile one. Reverse silhouettes at bottom center show actual sizes of the organs. The representative hemipenes are D. percarinatum (robust, USNM 559611) and D. graciliverpa (gracile, AMNH R-110584 [holotype)). Lectotype of Dendrophidion percarina- tum (Figs. 4-6). Cope (“1893” [1894]) described Drymobius percarinatus from two Costa Rican specimens, an adult from “Boruca’ collected 13 December 1891 (now AMNH R-17366; Figs. 4-5) and a young specimen from “Buenos Ayres” (originally AMNH R-9561 but now appar- ently missing). Both were sent to Cope by George K. Cherrie, a resident of San José who swioulee d for a time for the national museum of Costa Rica. Myers (1982: 23) explained the history of f these collections and their acquisition by the American Museum of Natural History. Recent liter- ature (e.g,, Lieb, 1996; McCranie, 2011) continues to refer to “syntypes oLeD: percarinatus but Taylor (1954: 727) had designated AMNH R-17366 as the lecto- type by referring to this specimen as “the ty pe, in Ae ancenee with requirements for lectotype designations before 2000 (Inter- national Commission on Zoological No- menclature, 1999: Article 74.5). The type locality, Boruca, is a small town in the foothills of the Fila Costefia on the north side of the Rio Grande de Térraba where that river divides the coast range in southern Puntarenas Province (southwest- ern Pacific versant of Costa Rica). SYSTEMATICS OF “Robust” Figure 3. morphotypes as they appear in retracted organs (scaled to the (left) and “gracile” (right) hemipenial same length, distal toward the top). Horizontal solid lines placed as in Figure 2. Dashed lines delimit the extent of apical tissue distal to the flounces. The hemipenes are Dendrophi- dion brunneum (robust, USNM 237060) and D. graciliverpa (gracile, UIMNH 77347). Abbreviations and labels: ss, sulcus spermaticus; spine array, compact tissue bearing the enlarged spines. The flounces appear as zigzag transverse lines of tissue distal to the spine array. Dendrophidion brunneum has an extensive area of nude apical tissue in the everted hemipenis, which accounts for the large area of pleated apical tissue in the retracted organ (left; Cadle [2010] illustrated the everted hemipenis of this species). AMNH R-17366 is an adult male in good condition; the tail tip is missing and approx- imately the distal one-third of ie remaining tail is broken off and tied to the specimen (Fig. 4). There is a midventral incision in the DENDROPHIDION PERCARINATUM (COLUBRIDAE) ¢ Cadle 269 the neck) narrowly bordered by dark brown flecks, which often tend to invest the pale portion of the bands. Dorsal crossbands peter out on the anterior portion of the tail. Narrow dark brown lateral stripe along the suture line of dorsal rows 2-3 on the posterior half of body (Fig. 5B). Ventrolat- eral tail stripe at the dorsocaudal/subcaudal junction (Fig. 5C) + dusky median four rows of dor socaudals on the anterior part of the tail (dusky continues onto posterior body, where the three paravertebral rows on each side are dusky). Venter immaculate except for narrow triangular encroachment of dorsal pigment onto lateral edges of ventral scutes. Head cap extends down to top of supralabials except the last (covers two-thirds of this scale) and the penultimate (covers somewhat less than half of this scale). Hemipenis of the Lectotype. Because of the critical importance of hemipenial characters to the systematics of Dendrophi- dion, the left hemipenis of the lectotype of D. percarinatum was manually everted using methods of Myers and Cadle (2003) (the right hemipenis had been damaged by a previous incision into the tail base). The manual eversion was successful (Fig. 6), although, as is typical in many manually everted hemipenes, it is not maximally expanded (Myers and Cadle, 2003). In particular, the apex is much narrower than in everted organs described later herein, result- ing in a different shape from field-everted organs. Nonetheless, the ornamentation of the bere of the tail. Total length 834+ mm; tail length 363+ mm; SVL 471 mm. 154 ventrals (2 preventrals); 143+ subcaudals; anal plate divided; 37 maxillary teeth with the last 3 somewhat enlarged; dorsocaudal reduction from 8 to 6 at subcaudal 23: dorsal scale reduction from 17 to 15 at ventral 93; 9/9 supralabials (2-4 touching the loreal; 4-6 touching the eye); 2/2 postoculars; 2+2 temporals; 10/10 infralabials. The suprala- bial/temporal pattern is the P pattern on both sides, as described above (Fig. 5A). Narrow (<1 scale row wide) pale cross- bands over the entire body (less distinct on entire everted organ except for the very tip of the apex can be studied. Before excision, the retracted left hemipenis extended to the proximal portion of subcaudal 7. The retractor penis magnus was proximally undivided. Approximate measurements of the man- ually everted organ: total length, 15.3 mm. Length from base to the right enlarged sulcate spine, 6.8 mm. Length of apex (spinose part to tip), 8.5-9 mm. Sulcus spermaticus simple, centrolineal. Just distal to the distal flounce is a single calyx on each side of the sulcus (distal flounce forms proximal border of each calyx). Sulcus continues to center of apex; could not really 270 determine well whether the tip of the sulcus was expanded (flared) or not. Hemipenis with narrow base, distally expanded (but shape is not as bulbous as pdidiereed organs). One pair of spines at proximal edge of spine array (one on each side of sulcus, right one somewhat more proximal than the left one), each some- what larger than other spines in the array. Spine array three rows across adjacent to sulcus and on the asulcate side, narrowing slightly to two to three rows on lateral sides in between. Thirty-two spines in the array including the two owas sulcate spines. Individual spines are of typical form (see later hemipenial section for relevance), strongly mineralized, hooked at the tip. A narrow circumferential section of the hemi- penial body immediately proximal to the spine array is ornamented with minute spines/spinules; hemipenial body proximal to minute spines is nude. Distal to the spine array the hemipenis is encircled by two flounces having scalloped edges, between which are a few low longitu- dinal ridges (more concentrated on the asuleate side). On the asulcate side distal to the distalmost flounce and extending toward the apical tip are several calyces in about two rows (transverse walls of these calyces are more strongly developed than the longitudi- nal walls); from these calyces a_ raised triangular area extends to the center of the apical tip. The tip of the apex is not fully everted, and I did not attempt full eversion because of the fragility of the specimen. Apart from the sulcus and its associated pair of calyces, and the asulcate triangular extension of raised tissue, the apex is nude. Diagnosis. Dendrophidion percarinatum is characterized by (1) dorsocaudal reduc- tion from 8 to 6 occurring anterior to subcaudal 27 (range, 5-26 ); (2) divided anal plate; (3) subcaudal counts >130 in males and females: (4) subadults with narrow pale bands or transverse rows of ocelli (<1 dorsal row wide throughout the body) separated by fewer than three dorsal rows on the neck (bands retained or become obscure in adults, often heavily invested Bulletin of the Museum of Comparative Zoology, Vol. 160, No. 6 with dark pigment); total number of pale bands on the body >70 (range, 71-96); (5) ventrals immaculate except for lateral dark blotches or mottling; (6) in life, dorsal coloration various shades of brown or gray and usually including dark-bordered pale crossbands anteriorly (sometimes indistinct) and dark brown or blackish stripes posteri- orly (often a broader pair of paravertebral stripes and a narrow lateral stripe on dorsal rows 2 and/or 3); venter without extensive dark spots or transverse lines (scattered small spots may be present), and (7) everted hemipenis of the “robust” morphology, with a relatively short, narrow hemipenial body proximal to a bulbous region bearing spines, calyces, and other apical ornamentation (retracted hemipenis usually extending to between subcaudals 6 to 9, rarely aenie subcaudal 10); total number of enlarged spines on hemipenis <45 (range, 26-40). Dendrophidion percarinatum differs from species of the D. dendrophis species group (D. dendrophis, D. atlantica, D. nuchale auctorum, D. apharocybe, D. cry- belum, D. vinitor) in having a more proximal reduction in the dorsocaudal scales (nearly always >30 in the D. dendrophis group). A high number of subcaudals and divided anal plate will distinguish it from D. apharocybe, D. crybelum, and D. vinitor (<130. sub- caudals and anal plate nearly always single in these species). Dendrophidion dendro- phis, D. atlantica, and D. nuchale aactorum have different color patterns (often with extensive dark ventral spots and flecks; see Duellman, 1978: 236-237, 2005: pl. 175; Savage, 2002: 654-655, fig. 11.39c, pls. 413— AAS): taktain greater sizes than D. percar- inatum, and have several enormously en- larged hemipenial spines (not so enlarged in D. percarinatum). Dendrophidion percari- natum differs from D. boshelli in having 17 midbody scale rows (15 in D. boshelli). Dendrophidion percarinatum _ differs from D. paucicarinatum in having pale dorsal crossbands (variably distinct), often has dark longitudinal stripes on the poste- rior body, and has an immaculate venter. Dendrophidion paucicarinatum usually has SYSTEMATICS OF DENDROPHIDION PERCARINATI bo ~l v (COLUBRIDAE) ¢ Cadle Figure 4. a more uniformly colored dorsum lacking distinct pale crossbands, has narrow dark lines across the venter in adults and many juveniles, and has a_ higher ventrals (21175). than “DP. percarinatum (nearly always <170 except occasional individuals from Panama and Colombia: see discussion of geographic variation). Dendrophidion paucicarinatum may have either a single or divided anal plate. Dendrophidion bivittatum differs from D. percarinatum in having a color pattern number of Lectotype of Dendrophidion percarinatum (Cope), AMNH R-17366, from Boruca, Puntarenas Province, Costa Rica. consisting of prominent blackish dorsal stripes on the posterior body and a greenish dorsal ground color. De ndrophidion bivitta- tum alse has a shorter tail and fewer subcaudals (<60% of SVL and_ usually <130, respectively) than D. percarinatum. Dendrophidion brunneum has a greenish to brownish dorsum generally ithout pale crossbands in adults (often with dark stripes or paravertebral punctations and often with dark transverse lines and other markings on the venter). bo ~ bo Figure 5. Lectotype of Dendrophidion percarinatum (Cope), AMNH R-17366. (A) Head, left side. Note the typical P pattern of the supralabials/temporals (see Fig. 1). (B) Posterior lateral stripe on dorsal rows 2-3. (C) Ventrolateral tail stripe. Dendrophidion percarinatum _ differs from the two new species described herein (Ce): prolixum and D. graciliverpa) in having a “robust” hemipenial morphotype as char- acterized herein (“gracile” in the last two species). Dendrophidion percarinatum also differs from these species in coloration. Dendrophidion prolixum and D. graciliverpa are green on the anterior body (sometimes restricted to the head) and often have narrow dark transverse lines on the anterior edges of ventral scutes, especially on posterior Boe Additional differentiating characters and comparisons are given in “the diagnoses for the new species. Description (104 males, 85 females). Table 1 summarizes size, body proportions, and meristic data for Dendrophidion per- carinatum throughout its geographic range; geographic variation and cesial dimorphism in some characters are summarized in the next sections. Largest specimen (AMNH R- 119376 from Panama) a male 1,358+ mm total length, 852 mm SVL (largest male with a comple te tail, 1,236 mm total length, Bulletin of the Museum of Comparative Zoology, Vol. 160, No. 6 Figure 6. Manually everted hemipenis of the lectotype of Dendrophidion percarinatum, AMNH R-17366, in sulcate and asulcate views. The spinose region and apex are not fully inflated even though the hemipenis is virtually fully everted. 747 mm SVL). Largest female (MCZ R- 20552 from Panama) 1,300 mm total length, 778 mm SVL. Tail 38-45% of total length (62-82% of SVL) in males; 37-43% of total length (59-75% of SVL) in females. Dorsal scales in 17-17-15 scale rows, the posterior reduction usually by fusion of rows 2 + 3 (N = 92) 3+4(N = 19), or loss of row 3 (N = 11) at the level of vonttale 84-106. Ventrals 147-170 (averaging 155.8) in males, 156— 167 (averaging 160.2) in females; usually 2 preventrals anterior to ventrals (about 9% of specimens have | preventral; rarely, 3 preventrals were present). Anal plate divid- ed. Subcaudals 137-163 (averaging 150.7) in males, 133-157 (averaging 145.6) in fe- males. Dorsocaudal reduction at subcaudals 8—26 in males (mean 16.0), 5—24 in females (mean 10.4). Preoculars 1, postoculars ae primary temporals 2, secondary temporals 2 supralabials usually 9 with A-6 bordering the eye (occasionally 8 with 3=5 bordering SYSTEMATICS OF DENDROPHIDION PERCARINATUM the eye or 10 with 5-7 bordering the eye), infralabials usually 9 (low fre quency of 10 or 11). Maxillary teeth 33-42 (averaging 37), typically w ith 3 or4 posterior tee th e nnlargec | (occasionally only 2 teeth or up to 5 posterior teeth were enlar ged). Enlarged teeth are ungrooved, not afiset and there is no diastema. Two apical pits present on dorsal scales. About 39% of specimens have Bee only on the vertebral or vertebral + 1 or 2 paraver- tebral dorsal rows on the neck (usually at least 4-6 rows lack keels on the neck and occasionally keels are absent). At midbody, 52% of specimens lack keels only on dorsal row |: another 38% lack keels on rows | and 2. and the remainder lack keels on the first 3 or 4 dorsal rows. On the posterior body, 91% of specimens lack keels only on row 1 (sometimes weak on row 2): the remainder lack keels on rows 1 and 2. Fusions or divisions of temporal scales were moderate- ly common, with the following frequencies (counting each side separately): upper or lower primary or secondary divided (30%), irregular fusions or divisions or fragmenta- tion (8.7%), other divisions or pagane (0.8%). Eighty-three percent of scorings for the supralabial/temporal pattern were P, whereas only 6.8% were G (t fhe remaining were ambiguous or irregular). Hemipenis unilobed with a bulbous apex; overall morphology “robust” as characterized herein. Spinose region fol- lowed distally by two flounces aad poorly developed calyces. Apex delimited by the distal flounce. The apex has an asulcate roughly triangular raised area bearing calyces on the asulcate side, and a thick pad of raised tissue on each side of the distal portion of the sulcus spermaticus; lateral to the raised sulcate and asulcate areas the apex is nude. Sulcus spermaticus simple, centrolineal, extending to the center of the apex, and having a slightly flared tip in everted organs. There is considerable varia- tion in the development of the calycular structures (from fully formed to much more rudimentary) and spines (see details in the hemipenial ‘descriptions). Retracted hemipe- © Cadle OF. ~l we) (CCOLUBRIDAE) nis usually extending to subcaudals 7-9 and only 1 rarely extending to subcaudal 10° or beyond. Sexual Dimorphism, Geographic Differ- entiation, and Other Variation. Tails are proportionally shorter in small individuals. Specimens <300 mm SVL have tail lengths 36-42% of total length, 57-74% of SV L (N = 43, males and females combined). Rare individuals have 15, 16, or 18 dorsal scale rows on the neck (one individual each), and one specimen had 15 dorsal rows at mid- body. Considering the rangewide s sample, males and females differ significantly in relative tail lengths (male longer -) ventral counts (female greater), subcaudal counts (male greater), the point of dorsocaudal reduction (male more telly and the point of dorsal scale reduction (female more posterior) (Table 1). The sexes do not differ in adult body size. These are common patterns in other species of Dendrophidion (Cadle, 2012) and are found widely among other snakes. These patterns hold when samples are analyzed by geographic origin, except that samples from Panama- Colombia are not Satan dimorphic in relative tail lengths (Table 2). This nonsignificance is due to the fact that males Gon Panama- Colombia average much shorter relative tail lengths than males farther north; females from throughout the range are similar in relative tail lengths. There is minor geographic variation in size, body proportions, and segmental eounts in Dendrophidion percarinatum (Table 2). Mean body size increases from north to south in both males and females, whereas relative tail length and subcaudal counts decrease in the same pattern. The point of dorsocaudal reduction in males from Panama and Colombia is more prox- imal than in specimens from seen of Panama. For most characters the greatest quantitative change in mean phacacted values occurs heaween Costa Rica and Panama-Colombia rather than farther north (e.g., between Honduras-Nicaragua com- pared with Costa Rican specimens). For 274 TABLE 2. Bulletin of the Museum of Comparative Zoology, Vol. 160, No. 6 GEOGRAPHIC VARIATION IN SELECTED CHARACTERS OF DENDROPHIDION PERCARINATUM. DATA PRESENTATION FOLLOWS THE FORMAT IN TABLE |. Honduras-Nicaragua, 13 Males, 15 Females Adult body size (mm, SVL) Male 407-595 Syed SS iss (13) Female 437-635 593.0 = 65.37 (9) Tail/total length Male 0.42-—0.45 0.44 + 0.008 (10)** Female 0.41 OA == OONS) Tail/SVL Male 0.740.82 0.78 ~ 0.024 (10) Female 0.69-0.71 0.70 = 0.010 (3) Ventrals Male 151-156 (53:7 S=4L GOuIS Fs Female 157-160 stows 22 ESI) Subcaudals Male 153-163 157.4: 2:76) (10) Female sen 147.8 = 2.76 (8) Dorsocaudal reduction Male (4293) 176 = S74 (13) > Female 515 Oe ==12187 (lp) example, for both sexes, mean RTL of peda! Nicaraguan specimens is similar to that of Costa Rican specimens, but mean RTL for Panamanian-Colombian specimens is less. In general, males show stronger geographic divieroneauon than females ioe the same character. Ventral and subcaudal counts and relative tail lengths reported by Rojas-Runjaic and Rivero (2008) for the male from western as) are similar to Panama-Colom- bian males (Table 2): 157 ventrals, 146 subcaudals (ventrals + subcaudals, 303), and relative tail length 41% of total length, 69% of SVL. This: specimen has soueral uncommon scutellation features (1/2 pre- Costa Rica, 29 Males, 26 Females Panama-Colombia, 62 Males, 41 Females 401-652 419-852 549.5 + 65.91 (20)* 613.5 + 100.95 (39) 457-695 465-778 999.7 + 61.58 (21) 631.0 + 80.40 (21) 0.42-—0.44 0.38—0.43 0.44 + 0.002 (12)** 0.40 + 0.011 (30) 0.39-0.43 0.37-0.43 0.41 + 0.014 (8) 0.40 + 0.015 (13) 0.72-0.80 0.62-0.77 O77 = 0.025. (12). - 0.68 = 0.031 (30) 0.64—0.75 0.59-0.75 0.70 + 0.039 (8) 0.67 + 0.040 (13) 147-158 152-170 1534-= 2.95. (28) ~~" 157. 157-164 159.8 + 2.06 (26) 160.9 + 2.70 (40) 148-161 137-160 15022 42)3:544(15)= 147.8 + 5.66 (46)* 133-157 135-156 T4722 Oe) 144.3 * 4.57 (27) 10-26 §—22 IMoh apeane 7. (227 0) es 14.5 + 3.09 (62) 794 6-17 Da Sco0 10:2 + 2.71 (41) oculars, 2+3 temporals, and 9 supralabials with only supralabials 5 and 6 touching the eye); its dorsocaudal reduction at subcau- dals 24-26 is greater than the range I observed in specimens from Panama- Colombia (Table 2). The Cordillera de Talamanca of Costa Rica and western Panama is a strong biogeographic barrier to many groups of organisms and separates differentiated pop- aliweds or closely related species on the Atlantic and Pacific versants in lower Middle America (Daza et al., 2010; Chan et al., 2011). Within Dendrophidion, this mountain range separates sibling species within the D. vinitor complex and presum- SYSTEMATICS OF DENDROPHIDION PERCARINATUM (COLUBRIDAE) * Cadle bo ~l TABLE 3. GEOGRAPHIC DIFFERENCES IN SELECTED CHARACTERS OF ATLANTIC VS. PACIFIC POPULATIONS OF DENDROPHIDION PERCARINATUM IN Costa Rica. DATA PRESENTATION FOLLOWS THE FORMAT IN TABLE |. NONE OF THE ATLANTIC/PACIFIC CHARACTER DIFFERENCES WERE SIGNIFICANT IN Adult body size, SVL (mm) Male Female Tail/total length Male Female Tail/SVL Male Female Ventrals Male Female Subcaudals Male Female Dorsocaudal reduction Male Female COMPARISONS BY SEX. Atlantic Versant, (10 Males, 8 Females) 531-596 556.8 + 21.38 (8) 567-670 607.9 + 33.88 (8) 0.43-0.44 0.44 + 0.006 (3) 0.41 0.41 + 0.0 (3) 0.75—-0.80 O75 4 0.0293) 0.69-0.71 0570: =, 0.018 (3) 149-156 152.3 + 2.45 (10) 157-162 1o9:6 = 1e7 (8) 152-158 154.0-+ 2.83 (4) 143-153 147.7 + 5.03 (3) Pacific Versant, (19 Males, 18 Females) 401-652 534.3 + 80.14 (12) 457-695 594.7 + 74.70 (13) 0).42—0.44 0.44 + 0.007 (10) 0.39-0.43 0.41 + 0.019 (5) 0)..72—0.80 O77 = 10.022 (10) 0.64-0.75 0.69 + 0.050 (5) 147-158 154.3 + 2.93 (19) 157-164 159.9 + 2.22 (18) 148-161 156252 3.56 (13) 133-157 147.3 + 8.26 (8) 14-26 19.2 + 3.49 (19) 8-24 LL 23:86 (17) ably played a role in their speciation (Cadle, 2012). Thus, it was of interest to compare character differentiation between popula- tions of D. percarinatum inhabiting the Atlantic and Pacific versants in Costa Rica (Table 3). In contrast to the D. vinitor complex, in which sibling species on the Atlantic and Pacific versants are differenti- ated by color pattern and hemipenial morphology but not scutellation (Cadle, 2012), no such differentiation is apparent in D. percarinatum. None of the standard scutellation characters are significantly dif- ferent between snakes of the Atlantic and Pacific versants, and I detected no consis- tent differences in coloration or hemipenial morphology between these population segments. Thus, at least as reflected by standard external characters, Atlantic and Pacific populations of D. percarinatum in lower Middle America show no population divergence, even though the southwestern Costa Rican populations may be isolated from the Atlantic versant populations (see Distribution). Coloration in Life. A sampling of color photographs of Dendrophidion percarina- tum from Costa Rica includes Savage (2002: pl. 418), Solorzano (2004: 236, pl. 59), and Kohler, 2008 (fig. 582). Photographs of Rance: cerearnatin in life. LACM 114102 from Finca Las Cruces, Puntarenas province, Costa Rica. From a color slide by Roy W. McDiarmid. Figure 7. Honduran specimens include K6hler (2003: fig. 480), McCranie et al. (2006, pl. 119), and McCranie (2011, pls. 6B, C). Guyer and Donnelly (2005) identified their plate 148 as D. percarinatum (“Brown Forest Racer”), but I identify this photograph as D. nuchale auctorum based on its color pattern (the specimen was photographed and released at the La Selva Biological Station, Costa Rica; Craig Guyer, personal communication). Black and adie photogr aphs of D. percar- inatum from Costa Rica are in Tay lor (1954: 728) and Lieb (1996). A color photograph of a specimen from southwestern Costa Rica (near the type locality) is shown in Figure 7. Specimens from Honduras and Costa Rica are brown to yellowish brown or grayish brown with narrow dark-bordered pale brown crossbands anteriorly grading to dark crossbands posteriorly and Ofte dark stripes or alternating dark and pale stripes posteriorly and on aie tail (Savage, 2002: 658; Solérzano, 2004: 234; Guyer and Donnelly, 2005: 184; McCranie, 2011: 109). the venter is immaculate, except for lateral dark pigment common to all Den- drophidion, and white or with a yellowish to orange wash. Some specimens apparently Ree a more uniformly colored dorsum (see also Coloration in Preservative). Over- whelmingly, the predominant dorsal colors of Dendr “ophidion percarinatum are shades Bulletin of the Museum of Comparative Zoology, Vol. 160, No. 6 of brown and with a general absence of extensive green colors (compare the two new species described herein). Color notes for specimens from the southern part of the range (given below) are similar to that just Heccriped For many specimens it appears that the dark borders to the pale crossbands are more prominent than the pale portions of the bands, which are sometimes not mentioned in individual descriptions. The following color notes are extracted from the field notes of Charles W. Myers for specimens from Panama and Colombia (AMNH_ R-108468 only). Specimens are listed roughly in order of i increasing SVL: AMNH R-109643 (female, 199 mm SVL): Lateral light bars pale yellow on neck, pale grayish brown on body. Venter white, with a yellowish tinge on throat. Iris pale tan upper quarter sector, dark brown below. Tongue black with an orangish tinge near base of fork. AMNHE R-129757 (male, 213 mur SVE}: Upper quarter sector of iris tan, lower three quarters dark brown. Tongue dark brown with black fork. KU 107656 (male, 229 mm SVL): Color like [KU 107652], except venter white instead of yellow, light dorsal spots pale brown rather than gray, and yellow of scale bases is light, not bright. KU 107647 (male, 236 mm SVL): Brown above with tan light areas. Venter white. Iris pale bronze with red-brown half moons either side of pupil. KU 107645 (male, 258 mm SVIlL)2 Amte= rior part of body with pink cast to the brown background. Trace of yellow on supralabials and on neck along lower scales and first 6 ventrals. Venter immaculate white. Iris pale tan with red brown half moons. KU 107659 (male, 329 mm SVL): Brown with yellowish dark-bordered crossbars an- teriorly and black crossbars posteriorly, where there are also lateral dark lines and a vertebral light vellowish area. A yellow tinge in neck region followed by an orangish east to entire dorsum (brown) to a bit past midbody. Labials and anterior venter ae except for a yellow tinge on venter, turning pale grayish near anus and under tail. Iris SYSTEMATICS OF DENDROPHIDION PERCARINATUM rich brown, except upper one-third, which is pale orangish tan. Tongue orangish brown with gray tips. KU 107651 (female, 536 mm SVL): Red- dish brown above, turning gray-brown on posterior one-third of body ad tail: scales on anterior two-thirds of body with bright yellow anterior margins (tip of base has a raail black spot). Ene body with black and_ lighter brown crossbanding of an odd pattern. Labials and chin hike turning yellow-green on ventrals and yellow on Siecadale Upper one-fifth of iris light tan; lower parts brown. KU 107650 feaale! 604 mm SVL): Brown above, white below. KU 75678 (female, 641 mm SVL): Dorsum brown with transverse black mark- ings; bases of scales on anterior two-thirds of body yellow. head white, changing to greenish white on anterior three-fourths of belly and to pale yellow on posterior one- fourth. Subcaudals bright yellow. Iris tan above, brown below. KU 107652 (female, 651 mm SVL): Brown with grayish spotting and black crossbars. Antonios bases of scales bright yellow on anterior two-thirds of body, especially notice- able on neck where the skin does not have to be stretched to show a yellow cast. There seems to be no special behavioral display associated with this hidden color. Labials and ventrals greenish white, turning light yellow on posterior one-third of body sal bright orange-yellow under. tail. Upper quarter section of iris light brown and lower part dark brown. No dork stripes, even on tail. KU 107654—55 (male and female, 543 and 720 mm SVL): Color like [KU 107652] except stripes are present and greenish white of anterior ventrals not extending under head and on labials, which are white. Tongue red with black tips in -54 [?], all black in other. KU 107653 (male, 703 mm SVL): Color much like [KU 107652] except the dorsal ground is darker and conspicuous. stripes are present posteriorly. AMNH R-119376 (male, 852 mm SVL): Head greenish gray, neck yellowish brown Labials and underside of (COLUBRIDAE) ¢ Cadle ara (anterolateral scale bases bright yellow) turning brown on body. Unde Bide of he ad and most of belly white, turning pale yellow toward tail and yellow under tail Iris brown except that upper quarter sector is pale tan. Tongue deep red with black fork. AMNH_ R-108468 (Colombia; male 615 mm SVL): Dull brown with indistinct black stripes. Anterior bases of scales are yellow in the neck region and then pale green to slightly past adbody. after which the: scale bases are not differently colored. Suprala- bials and underside of head white: first 20 ventrals are yellow, then venter turns green- ish white and, on last few dozen ventrals and on subcaudals, yellowish orange. Tongue dark. Iris pale brown on top quarter sector, dark brown on lower three-quarters. Coloration in Preservative. Generally, the color pattern described above for living specimens is maintained in well- preserved specimens, but color tones become dull brown, olive, or gray. There is considerable variability in the distinctness of the pale crossbands and their dark edging (from virtually absent to very distinct) (Fig. 8). The extent to which this reflects color pattern in life or differences in preservation is unknown. The vertebral scale row is often distinctly paler than paravertebral rows, especially on the posterior body. In some specimens this results in the appearance of longitudinal paravertebral str ipes. There is usually a lateral dark brown stripe posteri- orly on dorsal rows 2 and/or 3: this varies from very prominent to indistinct. Occa- sional specimens from Costa Rica and Panama (and perhaps other parts of the range) have a very subdued pattern in preservative (Fig. SB) and presumably in life as well. In these individuals pale anterior bands are absent, and the dark stripes on the body and tail nearly match the dorsal ground color, rendering a pale grayish or grayish brown doxsuin with aisoue darle lines. These individuals are otherwise typical of Dendrophidion percar- inatum in other characters (two examples I have seen are females), and I assume they represent color variants within the species. 278 Bulletin of the Museum of Comparative Zoology, Vol. 160, No. 6 Figure 8. Representative specimens of Dendrophidion percarinatum. (A) UMMZ 79764 (Nicaragua). (B) UMMZ 63762 (Panama). (C) and (D) AMNH R-119376, dorsal and ventral (Panama). The venter sometimes has scattered small dark spots. Distribution. Northern Honduras (Atlan- tida Province; McCranie, 2011) eastward and south throughout Central America to northern Colombia and northwestern Venezuela (Rojas- Runjaic and Rivero, 2008), and western Colombia (Chocoan region) to the vicinity of the Bahia de Buenaventura. Dendrophidion percarinatum occurs on the Pacific versant only in Costa Rica, Panama, and Colombia. Most localities are <1,000 m elevation but D. percarinatum occurs up to 12200: eareonn southwestern Costa Rica (Rio Coto Brus valley). In Honduras the maximum elevation attained is 685 m (McCranie, 2011: 110), and other elevational records derived from speci- mens I examined are 930 m (Panama), 520 m (Nicaragua), and 200 m (Colombia; but see below for a potentially much higher record). McCranie (2011: 109) and Savage (2002: 657) mapped localities for Honduras and Costa Rica, respectively. Lieb (1996) mapped the rangewide distribution, but the localities and range he gave for D. percarinatum in South America need correcting with the taxonomic SYSTEMATICS OF DENDROPHIDION PERCARINATUM (COLUBRIDAE) 81° 78 12s @ Dendrophidion percarinatum © Cadle 279 ? D. percarinatum, questionable A Dendrophidion sp. Playa de Oro 2 wl ( Quebrada Pangala 7 oN Rio Raposo — ey Figure 9. Distribution of Dendrophidion percarinatum in Panama and northwestern South America. All known South American localities are plotted. Three labeled localities in western Colombia (arrows) are sites where D. percarinatum is sympatric with D. prolixum. Locality in western Venezuela (about 10°40’N, 72°30’W) is from Rojas-Runjaic and Rivero (2008). ? in the Rio Cauca valley is Medellin, a questioned locality for D. percarinatum (see text). Triangles in central Colombia (Rio Magdalena valley) are localities for specimens similar to D. percarinatum in segmental counts, but which may represent differentiated populations or a distinct species; see Distribution in the D. percarinatum account and Appendix 1 (Dendrophidion species inquirendum). revisions herein, about which more is said shortly. Figure 9 shows the distribution of D. percarinatum in Panama and South America as I presently understand it. The southernmost specimen in western Colombia that I refer to Dendrophidion percarinatum is USNM 151658, a juvenile from the vicinity of the Bahia de Buena- ventura (approximately 3°45"N; see Figs. 12A, 13A, and discussion in the account for D. prolixum). However, few specimens of any species of Dendrophidion in U.S. collections seem to have been collected between this point and the Ecua- dorian frontier. Some difficulties in distin- guishing juveniles of D. percarinatum and the two new species are discussed in the account for D. prolixum. Sympatry between Dendrophidion per- carinatum and the new species D. prolixum is documented at three localities in western Colombia (Fig. 9): Playa de Oro, Quebrada Pangala, and ‘the Rio Raposo just south of Buenaventura. At each locality the speci- mens documenting each species have the distinctive color pattern characters of each: exemplars are illustrated in the account for D. prolixum (see Distribution). At Playa de Oro the documenting specimens are both males with everted hemipenes, which are described and illustrated later (see Figs. 39A, 42). 280 Bulletin of the Museum of Comparative Zoology, Vol. 160, No. 6 Populations of Dendrophidion percarina- tum in southwestern Costa Rica (Rio Coto Brus valley and Golfo Dulce/Osa Peninsula, including ‘the type locality) are seemingly isolated Scam populations of the Atlantic versant and uplands of northwestern Costa Rica (Savage, 2002: 657). Carara National Park is the only documented lowland locality on the Pacific versant north of the Osa Peninsula region. Despite this apparent disjunction, Atlantic and Pacific populations in Costa Rica do not differ substantively in standard external morphological characters (see above discussion of geographic varia- tion). Similarly, D. percarinatum seemingly has a somewhat spotty distribution in Panamaa not-infrequent pattern for Pana- manian snakes (Myers, 2003; Myers et all. 2007: 12-14). The identities of D. percar- inatum and D. prolixum have previously been confused in western Colombia (see discussion in the next section), where their distributions overlap. Areas of sympatry of these two species are discussed in the species account tow 1D: prolixum. Previous Records of Dendrophidion per- carinatum in Colombia and Venezuela. Be- cause of the confused identity of South American specimens previously referred to Dendrophidion percarinatum, I briefly comment on a few South American refer- ences to this species (Ecuadorian specimens are referred to in the species account for D. graciliverpa). Lieb (1988: 166) correctly inferred that records for “D. percarinatum” reported by Aleman (1953; “D. dendro- phis”) from western Venezuela and by Roze (1966) from the Cordillera de la Costa of northern Venezuela were misidentified specimens of D. nuchale auctorum. Many specimens of D. nuchale auctorum are available from the Cordillera de la Costa. The four specimens reported by Aleman (cited by Roze) are from Zulia state in western Venezuela near the recently report- ed “first record” of D. percarinatum from Venezuela (Rojas-Runjaic and Rivero, 2008). The last authors did not mention the specimens listed by Aleman (1953), who reported segmental counts and body pro- portions consistent with either D. percar- inatum or D. nuchale auctorum. These four specimens were apparently examined, and their identity confirmed as the last species, by James R. Dixon in 1981 (Fernando Rojas-Runjaic, personal communication). The scalation data and color details given by Rojas-Runjaic and Rivero (2008) for the “first valid record” of D. percarinatum from western Venezuela are consistent with that species as redefined herein, even though some of the reported head scutellation is rare in my sample (see above section on variation). These authors also described the anterior dorsum of the specimen as “grayish green uniform,” which is seemingly unlike most descriptions of Central American specimens (brown, yellowish brown, reddish brown; see Coloration in Life). The speci- men is a male with everted hemipenes according to the authors, and hemipenial characters could confirm its identity. Other Venezuelan records of “D. percarinatum (e.g., Vest et al., 11966; Lancing igs) undoubtedly refer to D. nuchale auctorum, as recognized by Lieb (1988). Lieb (1996) included as part of the distribution of Dendrophidion percarina- tum two localities in the interandean valleys of the Rio Cauca and Rio Magdalena of northern Colombia, stating some equivoca- tion as to their identity: “Isolated popula- tions tentatively referred to [D. percarina- tum] occur in the Departments of Antioquia and Cundinamarca in north-central Colom- bia; these snakes are somewhat divergent in the anterior body color pattern from D. percarinatum in other parts of the range” (Lieb, 1996: 636.1-636.2). These records are apparently based in part on MCZ R- 21984 (Sonsén, Antioquia department) and MCZ R-42185 (Villeta, Cundinamarca de- partment), which Lieb (1988: 174) cited as specimens of D. percarinatum; he indicated at least the first locality (Sons6n) on an accompanying map (Lieb, 1988: fig. 5). However, both of these specimens are unequivocally D. bivittatum (personal observations of both specimens, and Stuart [1932] for MCZ R-21984). SYSTEMATICS OF I am aware of only one specimen of De ndrophidion percarinatum pote ntially from. the deeper interandean portion of the. nro Cauca’ BMNH. 1897.11.12.10: collected by A. E. Pratt and said to be from “Medellin.” If the locality is the well- known Andean city of that name, and is truly the point of origin of the specimen (rather than a shipping point), then this a be an elevational record for the species (1,440-1,540 m). As a cautionary note, the NGA (2010-2012) online gazet- teer (GEONet) lists six other place names “Medellin” in the northern Colombian departments of Cordoba, Sucre, Magda- lena, and Bolivar—any of which are at lower elevations and would bridge the lowland distributional “gap” between the northern Colombian localities for D. per- carinatum around the Golfo de cara (about S°N) and the westernmost Vene- zuelan locality (Fig. 9). In the absence of other documented specimens from. this area, I am hesitant to include the inter- andean city “Medellin” as a documented locality for D. percarinatum. Several specimens with segmental counts similar to Dendrophidion percarinatum are known from 150-1,242 m elevation in the Rio Magdalena valley (Fig. 9: Appendix 1, “Dendrophidion species inquirendum” - see also Dunn, 1944: 477). Lieb (1988) had examined one of these, MCZ R-42186, from Boyaca department in central Colombia but, perhaps as a lapsus, did not include Boyaca in his above-cited quotation. A specimen reported as ee dendrophis” (Nicéforo Maria, 1942: 87) oe Sasaima a eeges ETO0= 1.200 1 1, upper Rio Magdalena) may also pertain “5 ii group (or to D. nuchale auctorum). The three specimens I have seen from the Rio Magdalena do have somewhat peculiar color patterns compared with typical D. percar- inatum and, although their scale counts are Similareto, WD, pel rearinatum, segmental counts by themselves often are unhelpful in distinguishing species of Dendrophidion (Cadle, 2012; this paper). Unfortunately, the only Fale among the three is a small DENDROPHIDION PERCARINATUM (COLUBRIDAE) « Cadle 28] juvenile (217 mm SVL) with retracted he mipenes previously exposed by a some- what mangled dissection. Some aspects of their morphology (e.g., a proportionally short spinose + apical region) seem unlike other D. percarinatum hemipenes | have examined but better preparations would be needed for confirmation. Study of additional specimens from this area will be needed to resolve the taxonomic status of these populations, but the available specimens are, in any case, seemingly geographically isolated from other own populations referable to D. percarinatum (Fig. 9). A summary of the Colombian specimens referred to Dendrophidion percarinatum by Lieb (1988) and my re-assessments of their identities are as follows: MCZ R-21984, R- 492185 (= D. bivittatum); FMNH 54949, FMNH 54958-64, FMNH 54965, LACM 36782, LACM 45443, USNM 151659 (= D. prolixum): and MCZ R-42186 ‘(= species inquirendum). 1 concur with the identity of the other Colombian specimens cited by Lieb (1988) as D. percarinatum as redefined here: PMNH 63761, FMNH 63772-73, FMNH 78118, USNM 151658. Stafford (2003: 111) referred LACM 45443 from Choco depart- ment to D. vinitor but there are no docu- mented occurrences of that species complex in Colombia (Cadle, 2012: 206-207), and I refer LACM 45443 to D. prolixum. Natural History. General overviews of the natural history of Dendrophidion percari- natum include Guyer and Donnelly (1990, 2005), Savage (2002: 657-658), Solérzano (2004: 234-236), and McCranie (2011: 108-111). Diet and reproductive parame- ters are summarized by Goldberg (2003), Stafford (2003), and Sexton and Pearwole (1965), Dendrophidion percarinatum is a diurnal, terrestrial to semiarboreal snake of lowland and premontane moist tropical forests. At night it has been found sleep- ing on low vegetation. The diet consists primarily of feincsivial leaf-litter frogs (e.¢., Pristimantis, Craugastor), with a lesser component comprising lizards (Anolis, Cne- midophorus; Stafford, 2003). Dendrophidion percarinatum is oviparous with recorded bo bo 8 clutch sizes of three to six and perhaps with multiple clutches per year in the southern portion of the range (Goldberg, 2003; Stafford, 2003). Brief field notes recorded for Panama- nian specimens by Charles W. Myers include the following (observations by day except where indicated). Six specimens were on the forest floor (AMNH R- 129757: KU 107647, 10765153) 107656): Four were associated with water courses: on a riverbank, along a forest stream, in leaves of a dry stream bed, in a river (AMNH B-119376, 109643; KU 107650, 107659). One was in an open grassy situation (KU 107645), and two were sleeping at night on a palm leaf 3 ft. (~0.9 m) above a stream bank and in a Heliconia 5 ft. (~1.5 m) above ground (KU 107654-55). Myers’ notes indicate that a light quickly awakens sleeping snakes, which “then behave as if flying snakes.” One specimen encountered by day re- mained motionless rather than fleeing (KU 107652). Myers forced the hind legs of a large Craugastor (C. fitzingeri group) from either KU i or 107655 (adults, 543 and 720 mm SVL, respectively). Dendrophidion percarinatum is sympat- ric with several other species of Dendro- phidion within its range. The range of D. percarinatum overlaps with D. apharocybe and/or D. nuchale auctorum from Hon- duras to Panama, and all three occur together at some localities (e.g., La Selva Biological Station, Costa Rica). At the Las Cruces Biological Station in southwestern Costa Rica D. percarinatum is sympatric with three other species—D. crybelum, D. nuchale auctorum, and D. paucicarina- tum—yielding perhaps the highest species density of Dendrophidion anywhere. In western Colombia the distributions of D. percarinatum, D. prolixum, and D. nuchale auctorum overlap broadly, with two of the three documented sympatrically at several localities; it would be unsurprising to find the three species occurring together. Bulletin of the Museum of Comparative Zoology, Vol. 160, No. 6 TWO NEW SPECIES FROM WESTERN COLOMBIA AND ECUADOR Dendrophidion percarinatum as rede- fined here is relatively homogeneous in color pattern and hemipenial morphology from eastern Honduras to northwestern Colombia. However, in western Colombia and Ecuador are snakes similar to D. percarinatum in standard scutellation char- acters and body proportions, but they differ strongly from that species in coloration and hemipenial morphology. Lieb (1988, 1996) included these snakes in his concept of D. percarinatum, and his concept was followed by others (e.g., Savage, 2002; Stafford, 2003; Cadle, 2010). However, they comprise diag- nosable units that I consider two distinct new species. Both new species have an excep- tionally long, slender hemipenis (gracile morphology) and color patterns different from D. percarinatum. The taxonomic dis- tinction of the new species is also supported by the fact that the distributions of D. percarinatum and the first of the new species to be described overlap in western Colombia, including the three localities of documented sympatry mentioned above. In the area of distributional overlap, specimens of D. per- carinatum maintain the typical color pattern and hemipenial morphology expressed throughout the rest of its geographic range, whereas the sympatric new species is dis- tinctive in both features. The situation in western Ecuador proved more confusing because not only do the two new species occur there but D. brunneum does as well. Preserved specimens of the three species can be a challenge to distinguish, and in fact, I was unsuccessful in allocating some speci- mens to any of the three with certainty. Dendrophidion prolixum New Species Figures 1D, 10-11, 12B, 13B, 14-17, 19, 42, 43A Drymobius dendrophis. Boulenger, 1913: 1034 (specimen from Pena Lisa, Colom- bia; = BMNH 1913.11.12.40). SYSTEMATICS OF DENDROPHIDION PERCARINATUM AGE (aan (COLUBRIDAE) ¢ Cadle 283 Figure 10. Holotype of Dendrophidion prolixum (AMNH R-109721). Male from Quebrada Guangui, Choco department, Colombia. ?Dendrophidion dendrophis (part). Peters and Orejas-Miranda, 1970: 80 (“southern part of Central America and northern South America”). Perez-Santos and Mor- eno, 1988: 134 (Pacific lowlands of Colombia). Dendrophidion percarinatum (part). Dunn, 1944: 477 (? part). Savage and Villa, 1986: 148, 169. Lieb, 1988: 172; 1996: 636.1. Pérez-Santos and Moreno, 1991b: 138. Savage, 2002: 657- 658. Stafford, 2003: 111 (BMNH 1913.11.12.40). Solérzano, 2004: 234— 236. Guyer and Donnelly, 2005. McCra- nie and Castaneda, 2005: 8. McDiarmid and Savage, 2005: 391, 421. Cadle, 2010: 24. Holotype (Figs. 10-11). AMNH R-109721 from Quebrada Guanguf, 0.5 km above Rio Patia (upper Saija drainage), 100-200 m, Cauca ene Colombia [about 02°50'N, 77°25'W; Myers, 1991: 8]. Col- lected 9 pees 1973 by Charles W. Myers and John W. Daly (field number C. W. Myers 11618). The holotype is a male, presumed adult, 754 mm total length, 307 mm tail length (447 mm SVL): relanee tail length 41% of total length, 67% of SVL; dorsocaudal reduction from 8 to 6 at the level of subcaudal 27; 153 ventrals, 2 preventrals, 142 subcaudals; 12 left and 11 right infra- labials. An unusual temporal scale configu- ration: 2+3+2 on each side (Fig. 11A). Supralabial/temporal pattern G (irregular because of divisions in temporal scales). Other head scales are typical of the species (Table 1 and description below). Both hemipenes are retracted but exposed by a ventral incision in the tail base; the right hemipenis extends to the middle of ene caudal 13, the left to the middle of subcaudal 15. The retractor penis magnus is slightly divided proximally. The type retains elements of the juvenile color pattern: approximately 51 pale crossbands on the body (tending to form ocelli and somewhat indistinct posteriorly; Fig. 11B), an indistinct broken line on the suture line of dorsal rows 2 3 posteriorly and 3 (Fig. 11C), and a venter with only scattered small dark spots in addition to dark trans- verse lines indicated only at the lateral edges of the ventral scutes. Paratypes. Colombia: Cauca: Quebrada Guangut, 0.5 km above Rio Patia (upper Saija drainage), 100-200 m, AMNH R-109722-28 284 Figure 11. 109721). view. (C) Posterior lateral pattern. Holotype of Dendrophidion prolixum (AMNH R- (A) Head, right side. (B) Midbody pattern, lateral (topotypes). Choco: Pefia Lisa, Condoto, 300 ft. [90 m], BMNH 1913.11.12.40. Playa de Oro, Rio San Juan, 400 m, FMNH 54965, AMNH_ 108469. Quebrada Bochorama, Loma de Encarnacion on right bank, LACM 45443. Quebrada Docordo, “middle Rio San Juan (about 17 km airline SSW Noanama), AMNH_ R-123749-51. Quebrada Pangala, lower Rio San Juan (about 17 km airline NE Palestina), AMNH R-123746. Quebrada Taparal, lower Rio San Juan (about 7 km airline NE Palestina), AMNH R-123744, R- 123753. Upper Rio Buey, 110-160 m, LACM 36782. Sierra [Serrania] de Baudo, 00 ft. [915 m], Pacific side, ANSP 25609. Serrania de Baudo, north slope of Alto del Bulletin of the Museum of Comparative Zoology, Vol. 160, No. 6 Buey, 900 m, AMNH R-119801. Risaralda: [Between] Pueblo Rico [and] Santa Cecilia, Pacific side, S00 m, FMNH 54949, 54958-64. Valle del Cauca: Rio Raposo, Virology Field Station near Buenaventura, USNM 151659. Referred Specimens. Colombia: Choco: Quebrada Pangala, lower Rio San Juan (about 17 km airline NE Palestina), AMNH R-123747. Narino: Riquarte [= Ricaurte], 3,900 ft. [1,189 ml], Pacific side, ANSP 25608. Ecuador: Esmeraldas: Immediate environs of Cachavi, 20 m, USNM 237064. Rio Cachavi, USNM 237065. Imbabura: Paramba, northwestern Ecuador |= Haci- enda Paramba; SO0—1,000 m], FMNH 4055, 4056(?) (the last two specimens not includ- ed in data summaries). The referred spec- imens are small juveniles except FMNH A056 (adult female in fair condition). Etymology. The species name is the neuter form of the Latin adjective prolixus meaning “stretched far out” or “long,” used especially in reference to parts of the body. The reference is to the unusually long hemipenis of this species compared with most other Dendrophidion. Diagnosis. Dendrophidion prolixum is char peice’ by (1) dorsocaudal reduction from 8 to 6 occurring anterior to subcaudal 27 (range, 8-26); (2) divided anal plate; (3) subcaudal counts >130 in males and females and adult tail length >60% of SVL; (4) subadults with narrow pale cross- bands or transverse rows of ocelli separated by >3 dorsal rows on the neck (adults retain bands or become predominantly brown or green without distinct pale bands); total Annee of pale bands on the body fewer than 60 (range, 49-57) when they are distinct: (5) ventrals immaculate or (in some adults) with narrow transverse dark lines across the anterior border of each ventral plate; (6) in life, head reddish brown and dorsum mainly green (brownish green in juveniles); and (7) everted hemipenis of “gracile” morphology, with an exceptionally long, slender hemipenial body proximal to an expanded tip bearing spines, calyces, and other apical ornamentation (retracted hemi- penis nearly always to subcaudal 10 or SYSTEMATICS OF DENDROPHIDION PERCARINATUM (COLUBRIDAE) greater); total number of enlarged spines on the hemipenis >60 (65-89 in four organs studied). “Gracile” hemipenial morphology will distinguish D. prolixum from all other species of Dendrophidion except D. graci- liverpa described herein and perhaps D. bivittatum (see above comments where the gracile morphology is described). Dendro- phidion bivittatum has a different color pattern (greenish dorsum with prominent blackish longitudinal stripes), a tail <60% of SVL, and fewer than 130 subcaudals. Dendrophidion prolixum differs from species of the D. dendrophis species group (D. dendrophis, D. atlantica, D. nuchale auctorum, D. apharocybe, D. crybelum, D. vinitor) in having a reduction in the dorsocaudal scales anterior to subcaudal 30 (posterior to subcaudal 30 in the D. dendrophis group except occasional fe- males). A high number of subcaudals and divided anal plate will distinguish it from D. apharocybe, D. crybelum, and D. vinitor (<130 subcaudals and anal plate nearly always single in these species). Dendi ophi- dion dendrophis and D. nuchale auctorum may have either single or divided anal plates, but these species have different color patterns, usually involving numerous narrow pale bands and/or ocelli (see Savage, 2002: 6594-655. for discussion of PD. Rucholey, attain greater body sizes, and have different hemipenial morphologies (robust morphol- ogy and enormously enlarged spines in D. dendrophis and D. nuchale). Dendrophi- dion prolixum differs from D. boshelli in having 17 midbody scale rows (15 in D. boshelli). Dendrophidion paucicarinatum lacks dis- tinct pale crossbands and has a_ higher number of ventrals than D. prolixum (>175 compared with <165 in D. pro- lixum). eee ophidion paucicarinatum may have either a single or divided anal plate. Dendrophidion prolixum differs from D. brunneum in color pattern (adult D. brun- neum generally ae pale bands) and in hemipenial morphology (robust in D. brun- neum; see Fig. 3 and Cadle, 2010). © Cadle 285 Dendrophidion prolixum is distinguished from D. graciliverpa by the wide spacing of the pale “dorsal bands on the neck (ae generally separated by >3 dorsal scale rows im DD: prolixum, <3 dorsal rows in D. graciliverpa). Consequently, D. prolixum ha fewer pale bands on the body when these can be discerned: 49-57 in D. prolixum compared with 57-87 in D. graciliverpa. In life D. prolixum has a reddish brown head and greenish body, compared with a green he dl and brown to gray body in D. “eraciliverpa. These two species are exceedingly similar in most characteristics (Table 1), and I discovered no consistent differences in hemipenial morphology between them in the few everted hemipenes examined when intra- specific variation is considered. Preserved specimens without discernible pale cross- bands are problematic to identify, and several specimens from the borderlands of northern Ecuador and southern Colombia are of questionable referral to either D. prolixum or D. graciliverpa. Dendrophidion prolixum has previously been confused with D. percarinatum, and these species cannot be distinguished by traditional scutellation features other than a few mean character differences (Table 1). These two species differ in (1) color pattern: reddish brown head with a greenish brown to green body, and venter anne immaculate or swath dark transverse lines (D. prolixum) vs. head and body primarily browns to grays, ane venter immaculate (D2) percarinatum): (2) number of pale crossbands on the body: 49-57 and separated by 23 dorsal scale rows on the neck (prolixum) vs. 71-96 and separated by <3 dorsal scale rows on the neck (percarinatum) (pale crossbands can be indistinct in either species, but especially adult D. prolixum); (3) relationship between the posterior supralabials and temporals (see Materials and Methods and Table 1): G pattern most commonly (D. prolixum) VS. P pattern most commonly (D. percarina- tum); (4) hemipenial morphology: gracile (prolixum) vs. robust (percarinatum); re- tracted hemipenes of D. percarinatum 286 TABLE 4. Bulletin of the Museum of Comparative Zoology, Vol. 160, No. 6 COMPARISON OF PRESERVED SPECIMENS OF THREE SPECIES OF THE PERCARINATUM COMPLEX IN WESTERN COLOMBIA AND ECUADOR. PATTERN ELEMENTS ARE BASED ON JUVENILES WITH DISTINCT BANDING PATTERNS, WHICH ARE RETAINED ONLY IN SOME ADULTS OF EACH OF THE THREE SPECIES. D. percarinatum, Figures 12A, 13A Supralabial/temporal usually P pattern No. of dorsal rows <8 separating neck bands No. of pale bands on 71-96 the body Head cap brown Form of the neck bands usually bandlike Distinctness of neck bands — usually distinct Contact between neck bands and ventral pale coloration pale portion of bands usually confluent with pale ventral color rarely extend to subcaudal 10, whereas retracted organs of D. prolixum nearly always extend beyond subcaudal 10. The strong differences between D. prolixum and D. percarinatum in color pattern and hemipenial morphology are maintained in the area of western Colombia where their geographic ranges overlap, including several localities of sympatry discussed later. Distinguishing Juvenile Preserved Speci- mens of the D. percarinatum Complex. Preserved juveniles of Dendrophidion pro- lixum, D. percarinatum, and D. graciliverpa present some challenges to identify, which is relevant not only to proper species recognition but also to discerning distribu- tions, range overlap, and sympatry of the three species. The pattern of pale cross- bands on the neck and a few other characters offer differentiating, if subtle, characters (Table 4, Figs. 12-13), although no single character will necessarily be decisive for a given specimen. These characters will also work with those adults that retain distinct bands. These characters usually ocellate (rounded usually distinct pale portion of bands D. prolixum, Figures 12B, 13B usually G D. graciliverpa, Figures 12C, 13C usually G 3-4 <3 49-57 i om brown dark gray bandlike (usually) or ocellate pale spots surrounded by dark pigment); often heavily invested with dark pigment often obscured by gray head cap extending onto the neck pale portion of bands confluent or not (extension of dark gray head color often interrupts bands on lower dorsal rows) usually cut off from pale ventral color by two or three dorsal rows; often a distinct dark brown lower border on neck ocelli were instrumental in identifying the south- ernmost specimen of D. percarinatum in my sample (Figs. 12A; 13A, left) and in dem- onstrating sympatry between D. percarina- tum and D. prolixum in western Colombia discussed later in this species account. There is also potential confusion of juvenile D. graciliverpa and D. brunneum in western Ecuador, but I have seen too few of the last species to be confident of differentiating juvenile characters; this problem is dis- cussed in the later section on D. brunneum. The two most consistent characters (Table 4) are the total number of pale crossbands on the body and their separation on the neck. These two in combination will usually easily separate D. prolixum (Figs. 12B, 13B) from the other two species (pale bands are distinct in juveniles of all three species, unlike in adults). Dendrophi- dion percarinatum and D. graciliverpa have a greater total number of bands, which are more narrowly separated on the neck. Side by side comparison of either of these with D. prolixum immediately shows a greater SYSTEMATICS OF Figure 12. Juvenile specimens of (A) Dendrophidion percar- inatum (USNM 151658, Choco department, Colombia). (B) D. prolixum (USNM 237065, Esmeraldas province, Ecuador). (C) D. graciliverpa (KU 179501, Pichincha province, Ecuador). band density on the neck in the first two compared with D. prolixum (Fig. 13). Most of the other characters in Table 4 are subject to some intraspecific variation (supralabial/tempor: al pattern) or to condi- tions of preservation interacting with the color pattern of a given specimen. None- DENDROPHIDION PERCARINATUM (COLUBRIDAE) ¢ Cadle 287 theless, careful comparisons usually make it possible to identify preserve d_ specimens confidently (coloration in life would prove diagnostic, if available). For example, in Dendrophidion prolixum the neck bands are distinctly ocellate, having the form of pale spots that are usually at least partly surrounded by dark pigment and are not confluent with the pale ventral color (Fig. 13B). Neck bands in D. percarinatum and D. graciliverpa are usually more band- like a relatively straight) and usually confluent with the pale ventral color. This pattern can be disrupted if the bands are disrupted on the side of the neck, as occurs with some frequency in D. graciliverpa (e.g., Fig. 13C, right side). In ahese cases ihe upper portions of the crossbands can appear as ocelli, but they are not generally set otf by dark pigment as in D. prolixum (Fig. 13B). In a few specimens of D. graciliverpa the neck bands are distinctly more ocellate, and other characters must be used. Clearly, preserved juveniles of Dendro- phidion percarinatum and D. graciliverpa will cause the most difficulty because these two can have similar total number of bands (Table 1). Three features in combination usually permit separation: the supralabial/ temporal pattern (P vs. G, subject to intraspecific variation documented in Table 1); the shading of the head cap (usually brown and erie to the rest of the dorsum in well- preserved D. percarina- tum vs. dark gray in D. graciliverpa, which presumably ronleets the green coloration of the head/neck in life): and the distinctness of the anterior three or four pale crossbands on the neck (distinct in D. percarinatum vs. obscured by dark gray [green in life] extension of the head cap coloration onto the neck in D. graciliverpa; compare Figs. 13A, C). In D. graciliverpa the dark gray extension onto the neck sometimes (seemingly mostly in smaller juveniles) occurs only on the lower dorsal scale rows. This obscures the ventral portions of the neck bands, resulting in the appearance of ocelli in this species. This potentially creates some confusion with D. prolixum, but 288 Bulletin of the Museum of Comparative Zoology, Vol. 160, No. 6 Figure 13. Comparison of pale bands on the anterior body of juvenile specimens. Each pair (left, right) representative of one species: (A) Dendrophidion percarinatum (USNM 151658, 191 mm SVL and AMNH R-123748, 272 mm SVL; both Colombia). (B) D. prolixum (AMNH R-109725, 208 mm SVL, Colombia; USNM 237065, 210 mm SVL, Ecuador). (C) D. graciliverpa (KU 179501, 231 mm SVL and KU 291237, 265 mm SVL; both Ecuador). distinct dark brown flecks or spots generally edge the ocelli in the last species. De scription | (17 males, 17 females). Table 1 summarizes size, body proportions, and me- ristic data for Dendrophidion prolixum. Larg- est specimen (ANSP 25609) a female 675 mm SVL (total length 966+ mm, tail incomplete): the largest female with complete tail (USNM 151659) was 662 mm SVL, 1,116 mm we length. Largest male (AMNH R-123750 650 mm SVL (1,003+ mm total length, me incomplete); ; largest male with complete tail (AMNH R-123751) 642 mm SVL, 1,037 mm total length. Tail 38-40% of total length (62- 67% of SVL) in males: 38-42% of for length (63-72% of SVL) in females. Dorsal scales in 17-17-15 scale rows, the posterior reduction by fusion of rows 243 (N = 95) or 344 (N = 35) or by loss of row 3 (N = §) at the level of ventrals 84-110 esa dimorphism discussed below). Ventrals 150- 163 (averaging 157.4) in males, 152-164 (averaging dee) in females; ventrals pre- ceded by specimens | (2.4% have only 1 preventral and rarely 3 preventrals were present or pre- ventrals were absent). Anal plate divided. Subcaudals 134-150 (averaging 140.5) in males, 133-150 (averaging 142.9) in females. Preoculars 1, pesthonlanss 2 (rarely 3), prima- ry temporals usually 2 3 (rarelylh ore): secondary temporals 2 | arelge 3) _ supralabials usually 9 with 4-6 mide ‘ring the eye (rarely 2 preventrals in about 70% of 10 with 4-7 bordering the eye), infralabials usually 10 (range S— 12 een high frequencies of 9 and 11). Dorsocuicel rochenon from 8 to 6 occurs at subcaudals 16-26 in males, 8— 24 in females. Maxillary teeth 36-42 (aver- aging 38), usually with 3 or 4 posterior teeth enlarged and ungrooved; rarely, only 2 or as many as 5 oh were enl: arged. Enlarged teeth are not offset and there is no dastoun Two apical pits present on dorsal scales. Most specimens lack keels on the lower 4, 5, or 6 dorsal scale rows on the neck (occasionally lacking on higher rows); nearly all specimens lack eal only on row | at poe (sometimes keels are weak on row 2 and one specimen lacked keels on row 2); on the posterior body keels are almost always present on all except dorsal row 1 (one specimen lacked keels on rows 1-2, and another had weak or partial keels on some scales in row 1). Fusions or divisions of temporal scales occurred with the fol- lowing frequencies: ee or lower tempo- ral headed vertically upper temporal divided horizontally es primary + secondary temporal acta: ), upper + lower primary fused (1, partial fusion). Seventy-six percent of scorings of the supralabial/ temporal pattern were G, whereas only 4.5% were the P pattern; the remaining were irregular or ambiguous patterns. Hemipenis unilobed, with an exception- ally long hemipenial body proximal to a | SYSTEMATICS OF DENDROPHIDION PERCARINATUM (COLUBRIDAE) ¢ Cadle 289 ana a rn ct Figure 14. Dendrophidion prolixum in life from the type locality (AMNH R-109726). Adult male, 628 mm SVL. From a color slide by Charles W. Myers. somewhat bulbous apex. Overall morpholo- gy “gracile.” Spinose region followed distally bee 2 ribunees and poorhk ly developed calyces. Enlar ged spines >60. Apex nude except for poorly developed calyxlike structures on the asulcate side and thickened tissue immedi- ately adjacent to the tip of the sulcus spermaticus. Sulcus spermaticus simple, centrolineal, with a slightly flared tip in everted organs. Retcted hemipenis nearly always to sabcaudal 10 or greater (to 15 or more). Variation and Sexual Dimorphism. T Tails are proportionally shorter in small individ- uals. Specimens <300 mm SVL have tail lengths 35-39% of total length, 53-64% of SV te (N = 9, males and (antes combined). No strong geographic trends were evident among the characters examined. Males and formales differ significantly in ventral counts (female greater), the point of dorsocaudal reduction (male more distal), and the point of dorsal scale reduction (female more posterior) (Table 1). The sexes do not differ in adult body size, relative tail lengths, or subcaudal counts. These are common pat- terns in other species of Dendrophidion (Cadle, 2012) and are found widely among other snakes. Coloration in Life. The following color description for Dendrophidion prolixum is extracted from color notes of Charles W. Myers for specimens from western Colom- bia: the type locality (two juveniles, four adults), Serranfa de Baud6 (one juvenile), and Playa de Oro (one adult). Notes for individual specimens are presented after the summary. A color photogr aph of a specimen from the type locality is in Figure 14 (AMNH R- 109726). The color pattern changes ontogenetically from banded juveniles to (usually) more uniformly patterned adults (see variation below). Three juveniles have pale brown or tan crossbands (pale blue on the neck in one specimen) on a brown to grayish brown 290 dorsum. Adults are seemingly polymorphic in dorsal ground color (greenish to brown- ish or reddee brown); the largest adults for which color descriptions are available ($550 mm SVL) are green to dark green but may have a brown or reddish brown suffusion. The head of adults is reddish brown, contrasting with the general dorsal coloration: reddish brown dorsolateral and lateral stripes may be present on the anterior half of the body. The venter is white to grayish white anteriorly in juveniles and ee ‘and there may be a yellowish wash on the posterior venter and tail. The venter in smal] juveniles is immaculate, but larger individuals develop indistinct to prom- inent transverse grayish lines across the anterior edges of the ventral scales (superfi- cially, the lines seem to be on the posterior edges of the scutes, but that is an illusion yet by the lines showing through the nearly- -transparent posterior edge of the next anterior ventral scute). Although Myers’ notes do not indicate pale dorsal tee in individuals >447 mm SVL (see below), most preserv ed specimens retain some trace of bands (appearing as scattered whitish dorsal flecks or transverse rows of pale spots); one large adult female is strongly banded (see below) so retention of distinct crossbands may vary. The following notes on coloration in life from field notes of Charles W. Myers are arranged in order of increasing SVL so as to highlight the relation of color pattern to size (starting with AMNH R- 109721 and following are considered adults): AMNH R-119801 (juvenile, Serrania de Baudo, 211 mm SVL): Body grayish brown with pale blue lateral spots on neck, these turning pale brown on rest of body. Reddish bowl lateral stripe. Underside of head white, turning grayish white over rest of oa eee “Upper one-fifth of iris tan, lower four-fifths reddish brown. Tongue black. AMNH R-109722-23 (juveniles, Queb- rada Guangui, 235 mm and 283 mm SVL): Tan interspaces on brown dorsum. Venter white anteriorly, grading to yellowish pos- teriorly. Iris tan in upper quarter, brown below. Tongue black. Bulletin of the Museum of Comparative Zoology, Vol. 160, No. 6 AMNH R-109721 (holotype, Quebrada Guangui; male, 447 mm SVL): Reddish brown with gray interspaces; touch of green on lower sides of neck. Underside of head and anterior venter white with some blotch- es of yellow on throat and supralabials, turning light yellow on posterior belly and under tail. Iris brown, ee black. AMNH R-109727 (Quebrada Guangui; male, 564 mm SVL): Top and sides of head red-brown, body green. Labials and most of under-head bright yellow with only a few small white areas. Color otherwise like [AMNH R-109726, below]. AMNH R-109724 (Quebrada Guangui; female, 594 mm SVL): Head red-brown: body overall dark green, with some dark reddish suffusions anteriorly. Supra- and infralabials bright golden yellow. Venter whitish anteriorly, turning “golden yellow (like labials) under posterior belly and tail. Iris red-brown, ales in upper quarter sector. Tongue black. AMNH_ R-109726 (Quebrada Guangui; male, 628 mm SVL; Fig. 14): Green on snout, turning red-brown atop most of head, with the red color extending caudad on neck in the form of vague dorsal and _ lateral stripes. Supra- and infralabials bright golden yellow, genials mostly white. Venter grayish white Hin gray crosslines, turning light yellow under tail. Iris brown. Tongue black. AMNH R-108469 (Playa de Oro; male, 644 mm SVL): Green above and on outer quarters of ventrals, being brightest on lower sides and ventral tips. Head deep red-brown, this color extending on body as a pair oF vague dorsolateral stripes (parts of scale rows 6-8) and as a vague lateral stripe (rows 2-3). These stripes fe isappear at midbody, and posteriorly the dorsum acquires a slight brownish suffusion in the green. Labials and middle of venter white, turning pale yellow under tail. Iris red- pian turning tan in upper quarter. Tongue bee gray. A color photograph of the head and neck of a snake from northwestern Ecuador (Ortega- -Andrade et al., 2010: appendix 3, “Dendrophidion Enea ‘) may be a photograph of Dendrophidion prolixum. SYSTEMATICS OF DENDROPHIDION PERCARINATUM (COLUBRIDAE) * Cadle 29] Figure 15. Dendrophidion prolixum adult dorsal and ventral patterns. (A) AMNH R-123746 (532 mm SVL; Quebrada Pangala). (B) AMNH R-109727 (564 mm SVL; Quebrada Guangul). (C) AMNH R-123753 (598 mm SVL; Quebrada Taparal). 292 Bulletin of the Museum of Comparative Zoology, Vol. 160, No. 6 : ‘ee ” ge =. ae te Spain = ~ Wie a = Figure 16. Variation in expression of transverse rows of pale spots/ocelli at midbody in preserved adults of Dendrophidion prolixum. (A) FMNH 54960 (587 mm SVL). (B) FMNH 54958 (648 mm SVL). (C) FMNH 54964 (661 mm SVL). (D) USNM 151659 (662 mm SVL). Arrows in panel C indicate traces of ocelli reduced to smali pale flecks; compare panel D, which is of similar size. The photo shows the top of the head brown or reddish brown, extending onto neck with a narrow brownish lateral stripe on rows 2— 3, a dorsolateral stripe on rows 6-7 or more; remainder of neck scales bright green; upper labials and visible portions of gular/ anterior ventrals bright yellow. The color pattern is similar to sev eral of the above- described specimens (compare especially AMNH R-109726, R-108469). This photo- graph potentially documents one Ecuador- ian locality for D. prolixum (see later discussion of problematic localities). Coloration in Preservative. Ground color of adults brown, grayish brown, or dark gray, usually with some sindcacon of pale (cream to Figure wii 109725 (Choco department, Colombia; 208 mm SVL). Dendrophidion prolixum juvenile. AMNH R- whitish) crossbands or transverse rows of spots (Figs. 15-16). The variation in dorsal ground colors is perhaps due to preservation caterer Pale crossbands are more prom- inent in smaller individuals (Fig. 17) than larger ones; crossbands in the largest speci- mens are sometimes so reduced that only a trace is evident (Fig. 16C); these specimens appear dark blue gray, sometimes blackish (Figs. 15B, C). The venter is immaculate in juveniles, but most adults have indistinct to prominent transverse narrow dark gray Or blackish lines across the anterior edges of the ne scutes (sometimes more continuous r prominent posteriorly than anteriorly) (Fig 15). The whole range of color patterns in preserved specimens is seen in a series from the type locality (AMNH R-109721-28). Distribution (Fig. 18). Lowlands and premontane foothills of western Colombia (Choc6, Risaralda, Valle del Cauca, and Narifo departments) and northwestern Ecuador (Esmeraldas province). Latitudinal range from about 6°6’N south nearly to the equator (Fig. 18). Elevational distribution from about 100 m up to 930 m in the Serranfa de Baudé (Choco department, Colombia) and 1,189 m (Ricaurte, Narifo department, Colombia). The distribution in southern Colombia and northwestern SYSTEMATICS OF Ecuador presents some interpretive prob- lems taken up in the next section. Sympatry between Dendrophidion pro- lixum and D. percarinatum is documented at three localities in western Colombia: two in im Rio San Juan drainage (Playa de Oro and Quebrada Pangala) ’ and another at the Rio Raposo just south of Buenaventura (Fig. 9). At these localities the two species maintain their distinguishing characteristics as given in the bos diagnoses and in Table 4. Documentation for meee localities is provided by the following specimens: Playa de Oro (AMNH R- LOS468. percar- inatum: AMNH_ R-108469, prolixum), Quebrada Pangala (AMNH_ R-123745 and ReI2374S; percarinatum:; AMNH _ R- 123746-47, prolixum), and the Rio Raposo just south of Buenaventura (USNM 151658, percarinatum; USNM_ 151659, prolixum). Two examples are presented in Figure 19. At Playa de Oro the documenting speci- mens are both males with Pouce hemi- penes (Fig. 19A), which are described and illustrated later (see Figs. 39, 42). Interpretive Problems Associated with the Southern Portion of the Distribution of Dendrophidion prolixum. The southern portion of the distribution of Dendrophi- dion prolixum (localities 1-4 in Fig. 18) and its overlap with that of D. graciliver pa described later is problematic in several respects. The documentation for these localities is entirely based on small juveniles that I refer to D. prolixum (and one of the records is based on the identification of a »hotograph in the literature). The only adult Seales the four localities cannot with any certainty be attributed to any species. Moreover, a few adult specimens from other parts of northwestern Ecuador could be referred to either D. prolixum or D. graciliverpa. These interpretive problems are discussed in the following paragraphs, beginning with numbered ilecakives in Figure is for which concrete evidence for thie occurrence of D. prolixum exists. Reference to the above discussion on the identification of juveniles (see Diagnosis and Table 4) is pertinent here. DENDROPHIDION PERCARINATUM and USNM © Cadle (CCOLUBRIDAE) 293 Golfo de Panama Type locality Yq Figure 18. Distribution of Dendrophidion prolixum. Arrow indicates the type locality. Numbered localities are the following discussed in the text: 7, Ricaurte (Narino department, Colombia). 2, Paramba (Imbabura province, Ecuador). 3, Rio Cachavi (Esmeraldas province, Ecuador). 4, Bilsa Biological Station (Esmeraldas province, Ecuador). Localities 1 and 3 (Fig. 18; Ricaurte, Colombia and Rio Cachavi, Ecuador). These localities are documented by three juvenile specimens: ANSP 25608 (Ricaurte) 237064-65 (Rio Cachavi: Fig. 12B). Diagnostic characters of the three specimens are the wide separation, distinctness, and somewhat ocellate form of the pale bands on the neck, low number of pale body bands (52, 49, and 57, 294 Bulletin of the Museum of Comparative Zoology, Vol. 160, No. 6 Figure 19. Examples of sympatry between Dendrophidion percarinatum and D. prolixum in western Colombia. Specimens on the left are D. percarinatum, on the right are specimens of D. prolixum from the same localities: (A) Playa de Oro (AMNH R-108468 and AMNH R-108469, respectively); (B) Quebrada Pangala (AMNH R-123748 and AMNH R-123746, respectively). See Figure 9 for locations. Everted hemipenes of the specimens in panel A are illustrated in Figures 39A and 42, respectively. respectively), and brownish head/neck ground color. These characters are consid- ered diagnostic of D. prolixum, so the identity of these specimens is reasonably secure. Locality 2 (Paramba, Ecuador). Two females are available from this locality (FMNH 4055-56). FMNH 4055 is a small juvenile (234 mm SVL) with an incomplete tail, 158 ventrals, a dorsocaudal reduction at subcaudal 11, and supralabial/temporal pat- tern G on both sides. Its scale counts and the G_ supralabial/temporal pattern are consistent with either D. prolixum, PD. graciliverpa, or D. brunneum, the three species known from the region. The supra- labial/temporal pattern makes one of the first two more likely than D. brunneum simply based on frequency of occurrence (71.6% of brunneum scores were P, only 4% G). The pattern of FMNH 4055 is sugges- tive of D. prolixum (dark brown ground color with somewhat ocellate pale spots on the neck, wide spacing of the neck spots (2.5-4 dorsal rows), and 59 total bands on the body. FMNH 4056 is an adult (517 mm SVL) with 161 ventrals, 144 subcaudals, dorsocaudal reduction at subcaudal 9, a SYSTEMATICS OF DENDROPHIDION PERCARINATUM (COLUBRIDAE) relative tail length of 42% of total length (72% of SVL), and supralabial/te mporal pattern G on both sides. The specimen is overall very dark gray, almost blackish, with no. discernible eral pattern and an im- maculate venter. All of these characters are consistent with D. prolixum, D. graciliverpa, or D. brunneum (with the last less likely based on the supralabial/temporal pace). Based on FMNH 4055, I aan locality 2 as a documented locality for D. prolixum; either of the other two species could be aa by FMNH 4056. Locality 4 (Bilsa Biological Reserve). I have seen two juveniles from this locali USNM 541964 and KU 291237 (Fig. 13¢, right), which I refer to the new species D. eraciliverpa based on color pattern and scutellation characteristics (Table 4). Both have the G_ supralabial/temporal pattern, and the retracted hemipenis of KU 291237 extends to the middle of subcaudal 10 (not examined for USNM 541964). A herpeto- faunal survey report illustrates in color three species of Dendrophidion from Bilsa Bio- logical Reserve (Ortega-Andrade et al., 2010: 148). One photograph is correctly identified as “D. nuchale” (auctorum). A second, “D. percarinatus,” appears to be the species here described as D. graciliverpa. The third species, “D. brunneus,” has a color pattern very similar to one described above for D. prolixum: a reddish brown head with the brown color extending onto the neck as dorsolateral and lateral stripes on a green ground color; labials, throat, and aneenror vente yellow. Thus, I tentatively include locality 4 in the distribution of D. prolixum, but it needs verification. The photographs of “D. percarinatus” and “D. brunneus” in Ortega-Andrade et al. (2010) appear to be juveniles. If I have identified the photogr aph correctly, it would corrob- orate the only documented sympatry be- tween D. prolixum and D. graciliverpa. The fact that all of the southern records of Dendrophidion prolixum are based on juveniles and the lack of adults from this region is disconcerting. Moreover, several adults from other localities in western ¢ Cadle 295 Ecuador are without discernible pattern elements and are dark gray or brownish, similar to some adult D. orolixum. I refer these to D. graciliverpa but that is based mainly on the fact that the only specimens from the same or nearby localities are referable to that species based on color pattern characters visible in preserved specimens—certainly a less than desirable situation. Additional data on coloration in life would help resolve these problems because D. prolixum and D. graciliverpa are otherwise quite similar in characters observable on preserved specimens, includ- ing hemipenial morphology discussed later. Natural History. The type locality of Dendrophidion prolixum is also the type locality of Phyllobates terribilis and Colos- tethus lacrimosus (Myers et al., 1978, Myers, 1991). Myers et al. (1978: 321-324, figs. 4-5) described and illustrated the area as of 9-21 February 1973, when Myers and Daly collected the topotypic series of D. prolixum. The following is quoted from their account (references to figures and notes omitted): [Quebrada Guangui is set] in rough hilly country at the western foot of a northerly inclined spur of the Cordillera Occidental. . Slopes are more often steep than gentle, and perpendicular surfaces are not uncom- mon. Hillside soils are gravelly in places. Drainage is by clear-water streams flowing over rock, gravel, and sand. The principal stream, a tributary of the Rio Saija, is the Rio Patia, which originates along the western base of Cerro Tambor. The Quebrada Guangut is a southward flowing tributary that empties into the Rio Patia at an elevation of about 90 m. above sea level. Hilltops in the immediate vicinity are about 200 m. above sea level. The region has a decidedly tropical wet climate (Afin the Koppen system [Koppen, 1931]) [and] receives a yearly rainfall probably in excess of 5 m. It seems certain that relative humidity is always very high, especially inside the forest. 296 There is no undisturbed forest along the larger streams, where small terraces and adjacent hillsides are either under cultiva- tion ... or in dense second growth. Inland, the native lowland rain forest is relatively undisturbed but only of moderate height, probably due to the precipitous slopes. There are occasional tall emergents that break the uniformity of the forest canopy. Most of the larger trees have buttressed roots, and _ tall palms with stilt roots are common. Tree-trunk moss is sparse. Small bromeliads commonly grow low on the trunks, but the bromeliad population is not dense and they rarely occur on the ground. The understory and ground vegetation of saplings and treelets, small palms, and herbaceous plants and ferns, varies from dense to moderately open. The forest tends to be most open on gravelly slopes, some of which are quite wet due to seepage. Leaf litter is sparse. (Myers et al., 1978: 321-322) Three specimens are accompanied by notes indicating they were active by day on the forest floor (AMNH R-119801; juvenile), on a stream bank (AMNH R-109727: adult), or on the ground in a brushy part of a ridge top forest (but with large trees, shaded at ground level) (AMNH R-109726; adult). In addition to sympatry between Dendro- phidion prolixum and D. percarinatum dis- cussed above, D. prolixum is broadly sym- patric with D. nuchale auctorum throughout western Colombia; sympatry is documented at the type locality (AMNH _ R-109718-20) and Quebrada Docordé (CAS 119591, 119604; AMNH R-123752). The distributions of D. prolixum and D. graciliverpa seemingly overlap in northwestern Ecuador and docu- mentation of actual sympatry is poor, but may occur at the Bilsa Biological Station (Esmer- aldas province); see above discussion. Dendrophidion graciliverpa New Species Figures: b Bed Be2s 3.0lQG 1306. 39B, 43B, 44-46 ? Herpetodryas dendrophis (part). Jan 1863: 81 (Cayenne, Popayan, Guayaquil, Ecua- Bulletin of the Museum of Comparative Zoology, Vol. 160, No. 6 dor). Jan and Sordelli, 1869: Livr. 31, pl. 3, fig. 2 (young individual from “Equateur’). The young specimen illustrated by Jan and Sordelli has a pattern and head scalation consistent with D. graciliverpa; the adult (their fig. 1), a blotched snake, appears to be Drymobius rhombifer (a conclusion reached independently by Savage and McDiarmid; personal com- munication from Jay M. Savage). Drymobius dendrophis (part). Boulenger, 1894: 16 (a male from western Ecuador collected by Louis Fraser, BMNH 1860.6.16.59) ?Dendrophidion dendrophis (part). Perez- Santos and Moreno, 1988: 134 (Pacific lowlands of Colombia). Pérez-Santos and Moreno, 1991b: 135 (“southern part of Central America and northern South America”). Dendrophidion percarinatum (part; western Ecuador included implicitly or explicitly as part of the distribution). Dunn, 1944: 477 (? part). Savage and Villa, 1986: 148, 169. Lieb, 1988: 172; 1996: 636.1. Almendariz, 1991: 143. Pérez-Santos and Moreno, 1991a: 134. Savage, 2002: 657. Kohler, 2003: 200; 2008: 215. Solorzano, 2004: 234-236. Guyer and Donnelly, 2005. McCranie and Cas- taneda, 2005: 8. McDiarmid and Savage, 2005: 391, 421. Rojas-Runjaic and Riv- ero, 2008: 129. Cadle, 2010: 24. Dendrophidion percarinatus. Ortega-Andrade et al., 2010: 148. Holotype (Figs. 20, 21, 23, 26A). AMNH R-110584 from 3 km E Pasaje, 30 m elevation, El] Oro province, Ecuador [03°20'S, 79°49’W]. Collected 11 February 1974 by Charles W. Myers and John W. Daly (field number C. W. Myers 12250). The holotype is an adult male, 964 mm total length, 359 mm tail length (605 mm SVL); relative tail length 37% of total length, 59% of SVL; dorsocaudal reduction from 8 to 6 at the level of subcaudals 26-27: 157 ventrals, 2 preventrals, 137 subcaudals. Head scales are typical of the species SYSTEMATICS OF DENDROPHIDION PERCARINATUM (COLUBRIDAE) ¢ Cadle 297 Figure 20. Holotype of Dendrophidion graciliverpa (AMNH R-110584). Adult male from near Pasaje, El Oro province, Ecuador. (Table 1 and description below) except that three postoculars are present on each side; supralabial/temporal pattern G (Fig. 21). Both hemipenes are fully everted. The right hemipenis was fully inflated for ‘leEcatien and description but was not removed from the specimen; at full inflation it is about 38.6 mm in length (Fig. 44). The preserved type has pale craéshamds (87 on the body) and a pale vertebral line (Fig. 26A); the venter has dense gray pigment laterally, indistinct narrow ‘orayish lines on hie anterior edges of the ventral scutes (more prominent posteriorly), and irregular small dark spots on the posterior ventrals. The head and neck are dark gray. In life the head was bright green ae the dorsum brown. to orangish brown but without distinct crossbands (detailed color notes below; see Fig. 23). Paratopotypes. AMNEH > R2110585—86, AMNH R-119835. Other Paratypes: “Peru,” no specific meaty [probably Ecuador: see Cadle, 1989: 422-423], ANSP 5519. Ecua- dor: “western Ecuador.” no other data. BMNH. 1860.6.16.59.” be seetesak Cha- guarapata, 2,000 ft. AMNH R-23032. El Oro: Hualtaco, USNM 237085. Rosa Delia Plantation, USNM 60523. Esmeraldas: Bilsa Biological Reserve, KU 291237. [Guayas|: Guayaquil, USNM 12268. Rio Pescado [about 488 m], AMNH R-23438. [Guayas? |: Headwaters of the Rio Congo, USNM 237063. Imbabura: Lita, USNM 237084. Los Rios: Finca Playa Grande 53 m|) UIMNH 77347. Playas de Montalvo, 15 m, UMMZ 83949. |Los Rios]: Centro Cientifico Rio Palenque, 47 km S. Santo Domingo de los Colorados on rd to Quevedo [220 ml], MCR R-156328—29, -R-156955. [Loja]: Ala- mor, AMNH 22232. Pichincha: Puerto Quito, MCZ R-166539. [Santo Domingo de los Tsachilas| {ex Pichincha]: Canoas near Santo Domingo de Los Colorados, USNM 237067. Finca La Esperanza near ‘Santo Domingo de los Colorados, USNM 237072. Joe Ramsey Farm, km 19 on Chone Road. 1S km W of Santo Domingo de_ los Colorados, USNM 237069. Meme, km 96 on road to Saloya at crossing of Rio Toachi, * J arbitrarily assigned this BMNH catalogue number to one specimen of a series of four (see footnote 3, Appendix 1, for details). By this scheme, BMNH 1860.6.16.59 is a male with an incomplete tail, and the largest specimen of the series (1,054+ mm total length, 676 mm SVL). These on should enable the identification of the correct specimen if the catalogue numbers ultimately become dissociated. The other three specimens of the series are Dendrophidion brunneum. 298 Bulletin of the Museum of Comparative Holotype of Dendrophiaion graciliverpa (AMNH R- 110584), head in lateral view. Figure 21. USNM 237074-75. Mulaute, on tributary of Rio Blanco, USNM 237073. Rancho Santa Teresita, km 25 on route to Chone from Santo Domingo de Los Colorados, USNM 283531-32. Rio Baba, 24 km S Santo Domingo de los Colorados, UIMNH 77345, 92243. Rio Baba, 19 km S and 5 km E Santo Domingo de los Colorados, UIMNH 92244. Santo Domingo de los Colorados, 550-660 m, KU 179500-01: USNM 237068. 2 km E of Santo Domingo de Los alenainc USNM 237070. 5 km W of Santo Domingo de Los Colorados, USNM 237071. Other Referred Specimens (Ecuador). Co- topaxi: Las Pampas [1,750 mJ], MCZ R- 163968—69. Esmeraldas: Quininde [about 100 mJ], USNM 237066. 41 km WSW of Quininde, Bilsa Biological Reserve, Black Trail, 300 m, USNM 541964. [Los Rios]: 1 km N Buena Fe, MCZ R-156327. [Santo Domingo de los Tsdchilas| \ex Pichincha): below Rio Toachi, USNM 237076. Etymology. The specific name is a femi- nine noun in apposition derived from a Latin words gracilis (slender or gracile) verpa (penis). The name refers to ‘the oe slender hemipenis of this species in compar- ison specifically to Dendrophidion percar- inatum, with which it has been confused, but also more generally to hemipenes of most other species of Dend ophidion. Diagnosis. Dende graciliverpa is characterized by (1) dorsocaudal reduction from 8 to 6 occurring anterior to subcaudal 28 (range, 7-27); (2 ‘divided anal plate; (3) subcaudal counts =120 in males and females: (4) subadults with narrow pale bands (<1 dorsal scale width on the neck) Zoology, Vol. 160, No. 6 or transverse rows of ocelli; adults retain bands or become predominantly brown or green (pale bands usually separated by fewer oes three dorsal scale rows on the neck: total number of pale bands on the body >55); (5) ventrals immaculate or with narrow trans- verse dark lines across the anterior border of each ventral plate; (6) in life, head greenish brown to green and body brownish, cao or grayish: and (7) everted hemipenis of the “gracile” morphology, with an exceptionally long. slender hemipenial body proximal to an expanded distal portion, w hich bears spines, calvces, and other apical ornamentation (retracted hemipenis nearly always to sub- caudal 10 or greater); total number of enlarged spines on hemipenis >80 (81, $4, and 116 in three studied organs). “Gracile” hemipenial morphology will distinguish D. graciliverpa from all other species of Dendrophidion except D. pro- lixum described herein and perhaps D. bivittatum (see above comments where the gracile morphology is described). Dendro- phidion bivittatum has a different color pattern (greenish dorsum with blackish longitudinal stripes). Dendrophidion graciliverpa differs from species of the D. dendrophis species group (D. dendrophis, D. atlantica, D. nuchale auctorum, D. apharocybe, D. crybelum, D vinitor) in having a reduction in the dorsocaudal scales anterior to subcaudal 30 (posterior to subcaudal 30 in the D. dendrophis group except occasional fe- males). A divided anal plate will distinguish it from D. apharocybe, D. crybelum, and D. vinitor (anal sais nearly always single in these species). Dendrophidion dendrophis and D. nuchale auctorum may have either single or divided anal plates but have lore color patterns (see Duellman, 1978: 236-237, 2005: pl. 175; Savage, 2002: 654, pls. 413-415), attain greater body sizes, and have different hemipenial mor- phologies (robust morphology and enor- mously enlarged spines in D. dendrophis and D. nuchale). Dend ophidion cracili- verpa differs from D. boshelli in having 17 midbody scale rows (15 in D. boshelli). SYSTEMATICS OF DENDROPHIDION PERCARINATUM (COLUBRIDAE) * Cadle Dendrophidion graciliverpa has previous- ly been confused with D. percarinatum, and these species cannot be distinguished by traditional scutellation features other than a few mean character differences (Table 1). These two species differ in (1) color pattern: greenish head and anterior body, and venter either immaculate or with dark transverse lines (D. graciliverpa) vs. head and body primarily browns to grays, and venter immaculate (D. percarinatum); (2) hemi- penial morphology: gracile (graciliverpa) vs. robust (percarinatum); retracted hemipenis nearly always to subcaudal 10 or greater (D. graciliverpa) vs. <10 subcaudals (D. per- carinatum); total number of enlarged spines on the hemipenis >80 (D. graciliverpa) vs. fewer than 45 (D. percarinatum); (3) relationship between the posterior suprala- bials and temporals as discussed above: G pattern most commonly (D. graciliverpa) vs. P pattern most commonly (D. percarina- tum). Dendrophidion graciliverpa has a greater number of pale bands on the body and different overall coloration than D. prolixum (see account for the last species and Table 4 for details). Description (22 males, 21 females). Table 1 summarizes size, body proportions, and meristic data for Dendrophidion gra- ciliverpa. Largest specimen (BMNH 1860.6.16.59) a male 676 mm SVL (1,054+ mm total length, tail incomplete; largest male with a complete tail (AMNH_ R-110584) 605 mm SVL (964 mm total length). Largest female (KU 179500) 663 mm SVL (9224+ mm total length, tail incomplete); largest female with complete tail (USNM 237085) 631 mm SVL, 1,027 mm total length. Tail 37-42% of total length (59-72% of SVL) in males; 36- 39% of total length (56-64% of SVL) in females. Dorsal scales in 17-17-15 scale rows, the posterior reduction by fusion of rows 2+3 (N = 13), 3+4 (N = 34), or loss of row 3 (N = 7) at the level of ventrals 78-101 (see sexual dimorphism below). Ventrals 153-163 (averaging 157.5) in males, 152- 166 (averaging 160.7) in females; ventrals preceded by 2 (rarely 1 or 3) preventrals. Anal plate divided. Subcaudals 132-153 299 (averaging 142.3) in males, 120-143 (averag- ing 133.5) in females. Preoculars 1, post- oculars 2 (occasionally 3), primary temporals usually 2 (rarely 3), secondary temporals 2 (rarely 1), supralabials usually 9 with 4—6 bordering the eye (rarely § and with other combinations bordering the eye), infralabials usually 10 (range 8-12 with high frequencies of 9 and 11). Dorsocaudal reduction from 8 to 6 occurs at subcaudals 12—27 in males, 7— 19 in females. Maxillary teeth 33-44 (aver- aging 39), usually with the 3 or 4 posterior teeth enlarged, ungrooved, and not offset; rarely, 5 teeth were enlarged. Maxillary diastema absent. Two apical pits present on dorsal scales. Usually the lower 5-7 dorsal rows on the neck lack keels; keels are usually lacking on the lower 2 rows at midbody (or row | only but with weak keels on row 2); all except the first dorsal row are keeled on the posterior body (keels sometimes weak or absent on row 2 as well). Nearly 67% of scorings for the supra- labial/temporal pattern were G, and only about 3% were P (the remainder ambiguous or irregular). Fusions or divisions of temporal scales were moderately common, with the following frequencies: upper or lower primary or secondary divided vertically (5.8%), upper primary divided horizontally (1%), upper or lower primary + secondary fused (1%). One specimen had 17-19-17 dorsal scale rows. Tails are proportionally shorter in small individuals. Specimens <300 mm SVL have tail lengths 34-38% of total length, 52-62% of SVL (N = 11, males and fouales combined). No strong geographic trends were evident among the characters examined. Hemipenis unilobed and of the gracile morphotype. Proximal portion of hemipe- nial body exceptionally long, slender, and ornamented with minute spines. Distal region slightly bulbous, containing a spinose region followed distally by 2 flounces and a largely nude apex. Usually an exceptionally large irregular calyx, and sometimes addi- tional poorly developed calyces, on the asulcate side of the apex. Sulcus spermati- cus simple, centrolineal, with the appear- ance of a terminal division in everted organs 300 Bulletin of the Museum of Comparative Zoology, Vol. 160, No. 6 Figure 22. Holotype of Dendrophidion graciliverpa in life (AMNH R-110584). Adult male, 605 mm SVL. From a color slide by Charles W. Myers. (divergent sulcus lips separated by a trian- gular tissue wedge). Sexual Dimorphism. Females average a significantly greater mean adult body size ad ventral counts and a more posterior point of dorsal scale reduction than males, whereas males have a greater mean relative tail length and snbenndal counts and a more posterior point of dorsocaudal reduction than females (Table 1). Coloration in Life. Color notes for four specimens from the type locality (field notes of Charles W. Myers hr AMNH R-110584- 86, R-119835) indicate that the head and neck are greenish brown in juveniles to bright green in adults. Ground color of the rest on the dorsum is brown to orangish brown or olive. Concealed anterolateral edges of the dorsal scales are bright yellow, especially on the anterior body. Supralabials yellowish white to golden yellow. Gular region white. Venter pale green to bright golden yellow. Myers’ coloration notes on individual specimens are here quoted in full: AMNH R-110584—86 (holotype and two topotypes; Fig. 22): Head and neck green- ish brown in smallest specimen [AMNH R- 110586; 409 mm SVL], bright green in [AMNH R-110584—85; 605 and 453 mm SVL, respectively]. Body color brown to orangish brown with concealed (anterolat- eral) scale edges bright yellow, especially on anterior body. Supralabials yellowish white (two smaller specimens) to light golden yellow (largest; holotype). Under head white; slight yellowish tinge under neck; anterior venter pale green, turning bright golden yellow under posterior body and tail. Upper one-third of iris tan, lower two-thirds dark brown. Tongue black, including tips. AMNH R-119835 (topotype, 565 mm SVL): Head and front of neck region bright green, turning olive with grayish black markings on body. Genial plates white, but labials and entire venter bright golden yellow. Top quarter-sector of iris tan, lower three-quarters dark brown. Tongue, including tips, black. Color notes for specimens from Pichincha province are the following: KU 179501 (231 mm SVL; field notes of John D. Lynch): Brown above with creamy brown bars edged with black (neck and head with green wash). Lips and throat lemon yellow. Venter somewhat dirty SYSTEMATICS OF DENDROPHIDION PERCARINATUM (COLUBRIDAE) Figure 23. Dendrophidion graciliverpa in life (MCZ R-156328, a juvenile, 308 mm SVL, from Pichincha province). From a color slide by Kenneth Miyata (courtesy of MCZ Department of Herpetology). yellow. Upper edge of iris pale copper. Most of iris reddish brown. MCZ 156328 (308 mm SVL; notes from a color slide by Kenneth Miyata reproduced here as Fig. 23): Similar to the above-described AMNH specimens in having a greenish brown head and neck, with the rest of the dorsum various shades of brown, including pale brown crossbands or dorsolateral ocelli on a medium brown ground color; on the posterior half to two-thirds of the body, a narrow dusky to dark brown lateral line along the suture line of dorsal rows 2 and 3, ending at the vent; and a dark brown line along the subcaudal/dorsocaudal border the entire length of the tail. The labials, throat, and visible portions of the anterior ventrals are pale yellow. KU 179500 (663 mm SVL; field notes of John D. Lynch): Brown above with pea green on skin between scales. Lips, throat, and anterior venter lemon yellow. Middle one-third of venter dull gray-green with some yellow. Posterior belly and underside of tail yellow. Iris brown. Based on these descriptions it appears that the green coloration of the head and neck may aoe less intense or greenish brown in smaller specimens, becoming bright green in adults, at least in El] Oro province. The presence of green on the head was not mentioned for KU 179500, an adult female, but was present in a juvenile from the same locality (KU 179501) and in MCZ 156328 (Fig. 23) from a nearby locality. This may indicate some variation in the presence of anterior green coloration. Oddly, the prom- inent pale crossbands and vertebral line in the preserved holotype (Fig. 20) are not © Cadle 30] evident in the live specimen (Fig. 22). Indistinct indications of crossbands can be discerned by subtle transverse alignment of dark pigme nt middorsally but this sould not be clear except by comparison of the photos of the live and preserved specimens. In fact, Myers’ notes mention no pale bands on any of the four specimens from the type locality and yet all of them have distinct pale bands in preserved specimens (Fig. 24D shows another example). This surely gives one pause in the interpretation of the Consider able variation in pattern evident among the preserved specimens referred to this spe- cies. The possibility that more than one taxon is represented should be reevaluated as material with associated color notes becomes available. A photograph of a specimen from Esmer- aldas province (Ortega-Andrade et al., 2010: 148, ‘ Dendrophidion percarinatus: ’) may be D. graciliverpa. Its head is greenish gray anteriorly, more greenish posteriorly. Su- pralabials whitish. Anterior dorsal ground color pale green. The pale bands on the neck are more ocellate (invested with blackish pigment tending to form round yellow spots partly surrounded with black) than is typical in D. graciliverpa, but their spacing appears too close for D. prolixum. Anterior ventrals pale yellow with whitish patches. The disparity in appearance between pale-banded preserved specimens of Den- drophidion graciliverpa compared with more uniformly colored live specimens (Figs. 20, 22) is considerable, but there is precedent in other snakes for similar alterations in color pattern due to preserva- tion. Smith (1955) reported a situation in which transverse series of pale spots ap- peared in preserved specimens of Thamno- phis rufipunctatus in the positions occupied by dark spots in life. The difference in spotting resulted in dramatically different appearances of the preserv ed vs. live specimens. Smith attributed the coloration differences, in part, to the differential solubility of various skin pigments in alco- hol. Realization of these preservational 302 Bulletin of the Museum of Comparative po ag, | fj big! i f THY fs Ks o f At : Figure 24. Zoology, Vol. 160, No. 6 Kan C4 —— ae Representative preserved specimens of Dendrophidion graciliverpa with prominent bands. (A) AMNH R-23438 (Guayas province). (B) AMNH R-22232 (Loja province). (C) UIMNH 77347 (Los Rios province). (D) AMNH R-110585 (El Oro province, topotype). effects on color pattern helped Smith resolve some nomenclatural issues related to original descriptions of color pattern in this group of Thamnophis. What remains unclear is why some preserved specimens I refer to D. graciliverpa have prominent pale bands in preservative whereas others do not (see the next section). The differences among preserved specimens suggest some- thing more than simple differential solubil- ity of pigments. This will probably only be resolved when more data on coloration in life and preservative for individual speci- mens are available. Coloration in Preservative. Preserved adult specimens exhibit two basic patterns, one with pale crossbands (similar to the juvenile pattern) and the other more SYSTEMATICS OF DENDROPHIDION PERCARINATUM (COLUBRIDAE) ¢ Cadle 303 Figure 25. Representative preserved specimens of Dendrophidion graciliverpa without strongly banded patterns (dorsal and ventral). (A) USNM 283532. (B) KU 179500. Both specimens from Pichincha province. unicolor or with indistinct crossbands (Figs. 24-26). The extent of gradation between these pattern forms is unclear, although there is considerable variation in the expression of distinct bands. The two forms are described separately here. Crossbanded pattern (Figs. VA. -QGAS C, F): Dorsal ground color brown, grayish brown, reddicti brown, or gray. Individual scales finely peppered waiehh blackish/dark brown, often more concentrated on poste- rior scale edges (keels often paler than rest of scale). Head and neck often darker (often dark gray) than remainder of dorsum. Many specimens have narrow pale crossbands (one dorsal scale or less) bordered with irregular black or dark brown flecks on anterior and posterior edges; these are present the entire body length and usually become indistinct on the proximal portion of the tail. Crossbands often less distinct on neck of adults (more evident in juveniles overall), restricted middorsally (lower sides of neck dark gray or gray). Vertebral row often paler or Bf contrasting color (e.g., gray compared with brown) than adjacent dorsal rows. On the posterior one-third of the body, dorsal rows 3 and 4 or just row 3 often with pale centers or paler overall than adjacent scale rows. Venter immaculate in small juveniles. Larger individuals (some juveniles but mostly adults) have varying expression of narrow dark transverse lines across anterior edges of ventral scales (expression varies from fine peppering of dark pigment across scales to dense 304 Bulletin of the Museum of Comparative Zoology, Vol. 160, No. 6 Figure 26. Variation in dorsal patterns of preserved specimens of Dendrophidion graciliverpa. Dorsal views at approximately midbody except (F) (lateral view); adults except (B) (juvenile). (A) AMNH R-110584 (holotype, El Oro province). (B) KU 179501 (juvenile, Pichincha province). (C) AMNH R-23438 (Guayas province). (D) USNM 237070 (Pichincha province). (E) USNM 237085 (El Oro province). (F) AMNH R-23438 lateral view (Guayas province). complete lines). These lines are often more distinct on posterior venter than anteriorly, where they are often incomplete across the middle of the scales. Subcaudal suture lines often bordered similarly. In addition to transverse lines, the venter may have scattered dark spots or speckling, especially posteriorly | (but nowhere dense), sometimes giving the venter a “dirty” white appear- ance. Sometimes a poorly defined ventro- lateral dark brown stripe on the anterior part of the tail at the subcaudal/caudodorsal border. Unicolor form (Figs. 25, 26D—-E): Dor- sum generally gray to aa gray (sometimes more brow ai with stratum corneum intact). Crossbands nowhere distinct (some- times a middorsal indication, such as USNM 283532, in which they are indicated on posterior half of body as pale punctua- tions along vertebral line and dark flecking indicating the dark borders for a greater portion of the body (Fig. 25A). Dark transverse ventral lines acts to the crossbanded form. with similar variable expression in larger individuals. Comments on Referral of Specimens. Dendrophidion graciliverpa as here con- ceived is highly variable in the color patterns of preserv ed specimens. Specimens with prominent crossbands (Fig. 24) in preserved adults are mainly fron the southern portion of the distribution (south- ern Guayas, E] Oro, and Loja provinces). Preserved specimens from the north (Pi- chincha, Chimborazo, and Imbabura prov- inces) generally lack prominent crossbands in dale. indistinct indications of bands are usually present) (Fig. 25). But there are exceptions to this generalization. For ex- ample, USNM 237085, a specimen from southern neces near the type locality and where nearly all other specimens have distinct pale crossbands, has only trans- versely oriented middorsal irregular brown spots or stipple indicating Re Unfortu- nately, the extent of color variation in life is not known and pale crossbands seemingly can be enhanced upon preservation, since several specimens with color notes or photographs in life do not mention prom- inent pale crossbands. This is particularly well shown by photographs of the holotype in life and preserved (Figs. 20, 22, 26A). Nonetheless, a color photograph of juvenile from the northern populations SYSTEMATICS OF DENDROPHIDION PERCARINATUM (COLUBRIDAE) ¢ Cadle 305 (Fig. 23) is very similar to the life colors of the holotype. | detected no scutellation differences among northern vs. southern specimens, although scale differences often seem not to be good predictors of species limits within De ndrophidion. Hemipenial morphology discussed later likewise suggests no cryptic species within my concept OF DER hichior be graciliverpa. However, I have examined relatively few hemipenes from different parts of the range, and some hemipenial variation seems not to correlate well with geography (see account of hemipenial morphology). The only evert- ed organ available from northern Ecuador (USNM 237069 described and illustrated later) is virtually identical to the everted hemipenis of the holotype from southern Ecuador, despite strong differences in preserv ed color pattern Dentcon the two specimens. Until more thorough data are available on geographic difexenees in col- oration, sul Reieision of D. graciliverpa is unwarranted. There is the added complication that Dendrophidion prolixum also occurs in northwestern Ecuador if several juveniles are correctly referred to that taxon (see account for that species). Resolving the extent of distributional overlap and sy mpa- try-ot 2) graciliverpa and D. prolixum in this area will require more specimens than have been available. As suggested above, the two species seemingly occur together at the Bilsa Biological Station! but a few adult specimens Sathout distinct patterns (dor- sum uniformly dark gray or brown) could be referred to either species. The referral of these Ecuadorian specimens to D. gracili- verpa is based to a great extent on the fact that other specimens (especially juveniles) with distinctive graciliverpa-like patterns have been collected at proximate localities, but no specimens with prolixum-like pat- terns are known from Ecuador except the few juveniles cited earlier. More definitive evidence is certainly desirable. Distribution (Fig. 27). Dendrophidion graciliverpa occurs in the lowlands of western Ecuador from Esmeraldas and Imbabura provinces in the north to Loja Soke ae in the south. The upper e ‘levational record, 1,750 m, is based on a juvenile male and fe fale from Cotopaxi province ( (MCZ 163968—69, referred specimens). These specimens have typical graciliverpa banding patterns, and the retracted hemipenis of the male extends to the proximal suture of subcaudal 15 (potential confusion with D. brunneum is possible in this area, but gracile hemipenial morphology is decisive for identification). Natural History. The holotype and three topotypes were collected while active in the late afternoon in dry leaf litter in a cacao plantation. AMNH R-119835 ate an Epipedobates anthonyi from the same site kept in a common collecting bag and later ate another from a nearby population about 1,000 m higher. Myers and Daly saw five specimens of D. graciliverpa during four man-hours on two visits to the type locality (one escaped after its tail broke off while pinned down under a boot). KU 179500-0O1 were collected by day in banana groves; the tail of the first broke during capture (John D. Lynch, field notes at Kw) Despite its wide distribution in western Ecuador, virtually nothing has been reported concerning the natural history of this species. The specimens I refer to D. graciliverpa seemingly occupy several distinct forest types in western Ecuador (Dodson and Gentry, 1978), but detailed habitat data are needed because of microgeographic variation in the habitat mosaic Of hie area. The distribution of D. graciliverpa narrowly overlaps those of D. Hchale auctorum and D. prolixum in northwestern Ecuador and more broadly overlaps that of D. brunneum. APPLICATION OF THE NAME DENDROPHIDION BRUNNEUM (GUNTHER) AND NEW DATA FROM WESTERN ECUADOR When I began work on the Dendrophi- dion percarinatum group in western Co- lombia and Ecuador it quickly became apparent that at least four species-group 306 Bulletin of the Museum of Comparative Zoology, Vol. 160, No. 6 Figure 27. taxa of the percarinatum group are present in this area (D. nuchale auctorum of the dendrophis group also occurs there, but it is distinguished from species of the percarinatum group: see Lieb, 1988: Sav- age. 2002). Two are the previously named taxa D. percarinatum (Cope), as redefined herein, and D. brunneum (Giinther, 1858). an Ecuadorian and Peruvian species for which Lieb (1988) and Cadle (2010) sum- marized taxonomic and natural history data. Dendrophidion percarinatum and D. brun- neum (sensu Lieb, 1988: Cadle, 2010: hereafter D. brunneum sensu Lieb/Cadle have hemipenes of the “robust” morphology easily 600 m contour above 3000 m 50 100 km Distribution of Dendrophidion graciliverpa. Arrow indicates the type locality. characterized herein. The other two taxa in western Colombia and Ecuador are charac- terized by long, slender hemipenes (gracile morphotype and aspects of coloration. These are the two new taxa described herein: D. prolixum and D. graciliverpa. All four western South American taxa of the per rcarinatum group are exceedingly similar in standard taxonomic characters (e.g., scale counts and body proportions) and are most easily distinguished by color patterns and hemipenial ‘morphology. However, color patterns can be difficult to interpret in some preserved specimens and some spec- imens of all four species appear uncannily SYSTEMATICS OF similar. The situation is not helped by the fact that color patterns within D. brunneum and D. graciliverpa vary considerably as judged by preserved specimens. At least D. brunneum as prese tly understood is poly- morphic in life colors (Cadle, 2010: data herein). Cadle (2010) had followed Lieb (1988) in recognizing Dendrophidion brunneum as a species primarily of the uplands in Ecuador and Peru, with the lowland type locality (Guayaquil, Ecuador) assumed to be a shipping point for the holotype. Cadle (2010: 10-11) concluded that no lowland records of D. brunneum from Ecuador could be confirmed, although several from northern Peru seemed valid. That assertion is overturned by some reinterpretations and the new material uncovered in this study. Lieb (1988) had examined only three specimens of D. brunneum but summarized data on other specimens from the notes of James A. Peters. Cadle (2010) examined the three specimens seen by Lieb and many others of the same taxon, primarily from Peru. However, neither Cadle (2010) nor Lieb (1988) had examined the holotype of D. brunneum (BMNH 1946.1.12.98). My continued study of specimens from western Ecuador (mainly labeled “D. per- carinatum” or “D. dendrophis” in their respective collections) led to the realization that D. brunneum sensu Lieb/Cadle was much more widely distributed in western Ecuador than previously recognized. More- over, it occurs broadly in a lowlands, where its distribution overlaps thatror WD, graciliverpa. Both species occur at Guaya- quil, if the locality as a point of origin is to be believed, and have been taken at some other closely contiguous localities. Thus, two species, one with a robust hemipenis (putative D. brunneum) and the other with a gracile hemipenis (D. graciliverpa), have been confused. (D. prolixum seemingly occurs only in far northwestern Ecuador and seems to pose no problem concerning the name D. brunneum.) It was only through examination of hemipenial mor- phology (mainly of retracted organs, as few DENDROPHIDION PERCARINATUM (COLUBRIDAE) © Cadle 307 everted hemipenes were available) that the spec ies were sorted out. Preserved females, juveniles, and poorly preserved specimens of these taxa can be difficult to distinguish— scutellation is of little help, and subtle aspects of dorsal pattern sometimes provide the best clues. The existence of two easily confused taxa in western Ecuador made it imperative to determine to which taxon—the robust or gracile hemipenis form—the name Dendro- phidion brunneum (Giinther) applies. This presented a problem inasmuch as_ the holotype is a female with an incomplete tail, and few coloration or pattern characters are discernible on the 150+ year old specimen. Nonetheless, a detailed consid- eration of the holotype suggests that Lieb (1988) and Cadle (2010) applied the name correctly to the taxon having a hemipenis of the robust morphotype. To mae this case, I redescribe the holotype of D. brunneum and then compare it to specimens of D. brunneum sensu Lieb/Cadle and to D. graciliverpa in greater detail. I then sum- marize my current understanding of the apparent considerable color pattern poly- morphism in D. brunneum and its distribu- tion in Ecuador. Redescription of the Holotype of Den- drophidion brunneum. The — holotype of Dendrophidion brunneum (BMNH 1946.1.12.98; Figs. 28-29) is an adult fe- male said to be from Guayaquil. It is in fair condition, somewhat soft, and with most of the stratum corneum missing. There is a long midventral slit in the base of the tail. Total length 867+ mm; tail length 224+ mm (with a healed cap on the stump); SVL 643 mm. Ventrals 154 (2 preventrals); 62+ subcaudals; anal plate divided; dorsocaudal reduction from 8 to 6 at subcaudal 10: Dorsals in 17-17-15 rows, the posterior reduction at ventrals 91/93 by fusion of rows 3+ 4 (left) and 2 + 3 (right); 9/9 supralabials (2-3 touching the loreal: 4—6 touching the eye); 2/2 postoculars; 2 + 2 temporals each side (upper primary temporal divided ver- tically on both sides); 10/10 infralabials; 34/ 36 maxillary teeth (left/right) with 4 or 5 308 Bulletin of the Museum of Comparative Zoology, Vol. 160, No. 6 Figure 28. Figure 29. (BMNH 1946.1.12.98). (A) Head, right side, showing the P pattern of supralabial/temporal relationship (left side virtually identical). (B) Posterior body in lateral view showing subtle paling of centers of dorsal scales in row 2 (arrows). (C) Approximately midbody, middorsal view showing three pale paravertebral rows. V indicates the vertebral row. Holotype of Dendrophidion brunneum (Gunther) Holotype of Dendrophidion brunneum (Gunther) (BMNH 1946.1.12.98) in dorsal and ventral views. posterior teeth enlarged, ungrooved, not offset, and without a eee The supra- labial/temporal pattern, as described in the section on systematic characters, is the P pattern on both sides (Fig. 29A). It should be noted that Giinther’s (1858) measurements of the holotype—‘length of tail 12”; total length 39”’—are not in modern inches, as the symbol ” (double prime) is interpreted nowadays and as I thought before examining the specimen ican 2010: 4). Either ee measurements are in error or, more likely, are in an archaic measure of inch or some other unit. I have not thoroughly investigated this detail, which seems unimportant for present pur- poses; the length of an inch has varied through its many centuries of use.’ In ede inches the holotype is about 34.25 inches total length, 8.75 inches in- complete tail length. Crater 1858) did not mention the incomplete tail, ee Bou- lenger (1894) and Parker (1938) ke Other specimens of D. oe have 4 or 5 enlarged posterior maxillary teeth like the >In a nearly contemporaneous paper Cope (1863) also used the double prime symbol for measurements of several specimens. Myers and Cadle (1994: 25) concluded that Cope’s svmbol did not er: inches but perhaps represented the metric system (em), which seemed to be the case for the holotype of Tae niophal- lus poecilopogon. However, metric units make no more sense than inches for the type of Dendrophidion brunneum using Giinther’s numbers. SYSTEMATICS OF DENDROPHIDION PERCARINATUM (COLUBRIDAE) ¢ Cadle 309 Figure 30. Maxillary dentition of Dendrophidion brunneum, UF 66199. holotype. which is illustrated by another specimen (Fig. 30). The posterior teeth in D. brunneum are relatively more robust and not enlarged to the same extent (compared with more anterior teeth) as in some other Dendrophidion species (e.g., D. apharo- cybe) (Cadle, 2012: fig. 8). The dorsum of the holotype is blue gray without stratum corneum, olive brown in yatches where the stratum corneum is intact. On the posterior half of the body, scales in dorsal row 2 have somewhat pale centers (Fig. 29B). In a few places the 3 paravertebral rows appear paler compared with a and vertebral row and lower flanks (Fig. 29C), but this is very subtle. Venter een laterally with blue gray patches similar to the dorsum. Beginning about one-quarter of the body length behiad the head are dark gray ventral fies extending across the anterior edges of each ventral scute. These are at first restricted laterally and little more than dense stippling, but by one-third the body length they are complete across the venter; they become more solid and dense posteriorly, tending to refragment a short distance anterior to the vent. ieateral subcaudal sutures lined with dark gray, and a few scattered subcaudal spots, but otherwise the subcaudals are immaculate. Application of the Name Dendrophidion brunneum. The conclusion that Lieb (1988) and Cadle (2010) applied the name brun- neum correctly to the robust-hemipenis taxon of western Ecuador is suggested mainly by two characters of the holot Ie. First, the holotype has the P supralabial/ temporal pattern on both sides (Fig. 29A). This is by far the most prevalent pattern in D. brunneum sensu Lieb/Cadle: 71.6% of scorings have the P pattern compared with 4% with the G ee (No= Jr in D. ee ‘rpa the G pattern is predominant (66.7 1% of ee G compared with only 2.9% B Table Second, the Hiss of the loreal in the holotype is more consistent with its shape and size in Dendrophidion brunneum sensu Lieb/Cadle than with D. graciliverpa. Cadle (2010: 4) indicated that the shape of the loreal seemed useful in distinguishing Den- drophidion brunneum from “D. percarina- tum” (= D. graciliverpa described here): rectangular and longer than tall in D. baianoieni, an irregular polygon as tall as or taller than long in D. graciliverpa. | quantified this char acter by measuring the loreal length (along the base at the supralabial horder) and height (greatest height, usually at the posterior prefrontal suture) for a sample of D. brunneum sensu Lieb/Cadle and D. graciliverpa (both loreal scales were measured for each specimen in the sample). The results (Fig. 31) show that the subjective impression at loreal shape difference is substantially true—in D. brun- neum the loreal scales are longer than tall, in contrast to D. gracilive rpa, W those loreal height for a given length is nearly always greater than that of D. Drneune Further- more, loreals in D. graciliverpa do not attain the lengths, even in large specimens, as the loreals of D. mmo (the longest mea- sured loreal of D. graciliverpa, 2.34 mm, was in a specimen 658 mm SVL, just shy of the maximum recorded length of 676 mm SVL). The holotype of D. benniciin (aster- isks indicated by arrows in Fig. 31) clearly falls in line with the specimens of D. brunneum sensu Lieb/Cadle. Based on the supralabial/temporal pat- tern and the shape of the loreal, the female 310 Bulletin of the Museum of Comparative Zoology, Vol. 160, No. 6 S a nee 7 as rg °° = a mee + = 7 oo Bote oF eet is gud Sa > ee cape 6g =o cae: 2 > 2 = O D. graciliverpa @ D. brunneum * D. brunneum holotype 0) 1 2 3 Loreal length (mm) Figure 31. Bivariate plot of loreal length vs. loreal height in a sample of Dendrophidion brunneum and D. graciliverpa. The points for the holotype of D. brunneum (BMNH 1946.1.12.98) are shown by asterisks indicated by the arrows. holotype of Dendrophidion brunneum is reliably associated with the snakes of “robust” hemipenial morphology in western Ecuador. Apart from these two characters, virtually all other characters of the holotype are consistent with either D. brunneum or D. 21 aciliverpa. A few pattern characters are similar to other specimens referred to D. brunneum (e-8.. pale centers on dorsal scales of row 2 and indistinct pale paraver- tebral rows; see Figs. 29B, 34B and Cadle, 2010: 7=8). but thes are subtle and with substantial interpretation in the old holo- type. Both species can have dark transverse lines on the venter, as in BMNH 1946.1.12.98 (e.g., Figs. 25, 32). Moreover, the specimens newly srefened to D. brun- neum in this paper significantly broaden my concept of color pattern variation in this species. In the following sections I address the color polymorphism and new distribu- tional data. Color Pattern Polymorphism in Adult Dendrophidion brunneum. Cadle (2010: 6-8) discussed polymorphism in dorsal coloration in life within a Peruvian popula- tion of Dendrophidion brunneum. This polymorphism involved mainly transitions between a primarily creenish to more brownish, olive, or coppery coloration, and some individual specimens showed anterior to posterior transitions in this shading. Nonetheless, significant variation in dorsal pattern was suggested by two preserved Ecuadorian specimens with dorsolateral and lateral stripes and a Peruvian specimen that had indistinct pale anterior crossbands (Cadle, 2010: 7-8). Other specimens had dark paravertebral spots usually more prev- alent on the posterior body. The new specimens show that these color variants are more common and widespread among Ecuadorian populations. With a few ApDIE exceptions, specimens from south- ern Ecuador and (especially) ) northern Peru appear more uniform in color pattern. The emerging picture is that D. brunneum (assuming only one taxon is involved) exhibits Sabee are color polymorphism, such as that documented for many other snake species; see Wolf and Werner (1994) for a brief review and Brodie (1990, 1992) for an elegant analysis of a case of extensive polymorphism in Thamnophis. Major pat- tern characteristics of adults are taken in turn in the following paragraphs (juvenile patterns are disenaced separately below). Except for the first two color morphs, my interpretations are based entirely on pre- served specimens (with the same caveats about translating preserved color patterns to life colors or vice versa mentioned connection with D. graciliverpa). Unicolor pattern (green to brown in life without distinct dark stripes, yunctations, or pale crossbands; Fig. 32A): Cadle (2010) described and illustrated this color pattern in detail. The dorsum is green to brown, sometimes with yellow, bronze, or coppery highlights. All specimens from northern Peru I have seen in life, including juveniles 277-300 mm SVL, are unicolor. This is the predominant coloration of preserved spec- imens from both Peru and southern Ecua- dor, assuming that “green in life” becomes, in preserv sdk specimens, brown with intact stratum corneum or slate blue or dark gray without stratum corneum (a common rela- tionship in snakes). Black or dark brown flecks or spots are present in the paraver- tebral region of some live and preserved specimens, usually more prevalent on the SYSTEMATICS OF DENDROPHIDION PERCARINATUM (COLUBRIDAE) ¢ Cadle ol] Figure 32. posterior body (Cadle, 2010: 6—7); these are comparable to similar spots in the “punc- tate” pattern described below but seemingly less prominent. Some specimens ea central western Ecuador (e.g., UF 85105 from Los Rios province) have nearly uni- formly brownish dorsums with only hints of paravertebral spots (see punctate pattern below). Whether these are unicolor or appear more punctate in life would be interesting to know. Striped pattern (dorsal ground color green in life, stripes dark brown to blackish; Figs. 32B, 33A): Snakes with this pattern have broad dark dorsolateral stripes (2-3 paravertebral rows wide) and a narrow dark lateral stripe on dorsal row 2, both usually more distinct on the posterior body. In life the dorsum of KU 142802 (Fig. : 32B) was olive green with dark brown stripes; venter creamy white with a tinge of yellow laterally on the neck and faint hue on the anterior Unicolor and striped color morphs of Dendrophidion brunneum from near Loja city in southern Ecuador (their localities are about 30-35 km apart). (A) USNM 237081. (B) KU 142802. margins of the ventrals; top of head olive brown, iris brown, tongue reddish brown with a black tip (field notes at KU). A color photograph perhaps of this form is in Yanez- Munoz et al. (2009: fig. 61, labeled “Dry- moluber sp.”); this photogr aph shows a bright green snake (head somewhat darker green) and dark brown stripes that are less distnct on the anterior body. In preserved specimens the stripes are usually rusty reddish brown and can show less contrast with the dorsal ground color than in live specimens; in excessively darkened speci- mens (ground color dark gray) the rusty stripes appear paler than fhe background. Some striped specimens have dark paraver- tebral punctations on some part of the body (e.g., KU 142802, Fig. 34A) ) and the striped ae punctate color morphs may_ totally grade into one another (see be low): colors in life would help interpret these patterns. I suspect that older specimens (e.g., the Figure 33. 88467). (C) Banded pattern (USNM 237079). holotype and MCZ 8393; Figs. 29B, 34B) with pale paravertebral dorsal rows and pale dorsal row 2 on the posterior body may represent the striped morph as it ages in preservative. The striped color morph is known from montane areas in the vicinity of Loja city (KU 142802) and Azuay province (USNM 237042) and from the lowlands farther north in Guayas province (UF 87940 Dorsal pattern polymorphism in Dendrophidion brunneum. (A) Striped pattern (UF 87940. (B) Punctate pattern (UF from Daule, Fig. 33A; USNM 237059 from Guayaquil). Punctate pattern set within dorsolateral rusty or pale brown stripes (colors in life unknown; Figs. 33B, 34B, C): In the low- lands of Guayas province (Rio Daule/ Babahoyo system) and in northern Ecuador (Imbabura province) are specimens having paravertebral stripes of varying distinctness SYSTEMATICS OF DENDROPHIDION PERCARINATUM (COLUBRIDAE) * Cadle WwW WwW Figure 34. Details of pattern polymorphism in Dendrophidion brunneum. (A) Striped pattern. KU 142802 anterior body (left) and midbody (right), dorsal perspectives. (B) Punctate pattern. MCZ R-8393 midbody in dorsal (left) and lateral (right) perspectives. Arrows in lateral perspective indicate the pale centers of dorsal scales in row 2 (compare Fig. 29B). (C) Punctate pattern. BMNH 1860.6.16.58 posterior body in dorsal perspective. (D) Banded pattern. USNM 237079 midbody in dorsal perspective. (usually rusty red in preserved specimens) but also with extensive dark punctations within the stripes. In different specimens either the punctations or the stripes may have greater prominence, and either ele- ment can be quite indistinct. The puncta- tions are often more distinct on one part of the body than another. Examples are UF 66199, 85105, SS8466—67 (Fig. 33B): MCZ R-8393 (Fig. 34B): BMNH_ 1860.6.16.58 (Fig. 34C); USNM 237061 (Fig. 35A), USNM 237083. Note that the punctations are restricted to the paravertebral areas and do not cross the vertebral line to form dark transverse lines. This subtlety is often a clue in distinguishing some preserved specimens of D. brunneum from some confusingly similar D. graciliverpa. In the last species, dark borders to the pale crossbands superficially have the appear- ance of the paravertebral spots in D. brunneum, but in D. graciliverpa the dark borders tend to cross the vertebral region to form a broken or more continuous middorsal transverse dark line (Fig. 35). However, this difference can be confused when the crossbands are offset, as occurs frequently in D. graciliverpa, or when the dark borders to the pale bands are frag- mented. Crossbanded pattern (colors in life un- known; Figs. 33C, 34D): Other than small juveniles discussed below, crossbanded specimens occur together with unicolor specimens in the highlands of Loja province and in northern Peru. Only three examples are known to me: USNM 237078-79 (567 and 328 mm SVL, male and female, respectively; presumed adult and subadult) and FMNH 232578 (adult male, 514 mm SVL: Cadle, 2010: 8). These three are from the region where the unicolor morph 314 Bulletin of the Museum of Comparative Zoology, Vol. 160, No. 6 Figure 35. drophidion brunneum and D. graciliverpa. (A) D. brunneum, punctate pattern (USNM 237061). (B) D. graciliverpa (USNM 12268). Both from Guayas province. V indicates the vertebral row. In D. brunneum the punctations are paravertebral in Comparison of midbody pattern details of Den- position and do not cross the vertebral scale row. In D. graciliverpa the dark borders to the pale bands form irregular lines crossing the vertebral scale row. See also Figures 24 and 26. predominates, and unicolor specimens are known from the same localities as the three crossbanded specimens (unicolor USNM 237077, 237080-S1 from the same locality as USNM 237078-79, and many unicolor specimens eon the same locality as FMNH 232578: Cadle, 2010). Crossbanded speci- mens presumab ly represent individuals that retain the juvenile banded pattern to a greater size than most individuals. The bands on the juvenile (USNM 237079) are much more aoe than those on the adults (USNM 237078, FMNH 232578). Ventral Sane in Dendrophidion brun- neum vary from immaculate to having dense dark gray transverse lines across the anterior edges of the ventral scutes (Fig. 32). Addi- onl dark spotting may be present, and in some specimens there is extensive expan- sion of the lateral dark pigment common to all specimens so that only a central part of each ventral scute is relativ ely clear (this pattern seems rare). There seems to be little correlation of these patterns with size (most juveniles have immaculate venters but several have well developed transverse lines). Adults with relatively immaculate venters are mainly from the Rio Daule/ Babahoyo system in central western Ecua- dor, whereas specimens from southern Ecuador and Peru usually have extensive ventral markings; this parallels the situation in juveniles (see below). But there are exceptions in both regions. For example, the holotype from central western Ecuador has a densely lined venter (Fig. 28). Juvenile Color Patterns in Dendrophidion brunneum. Cadle (2010) referred only three juveniles (277-300 mm SVL) from northern Peru to Dendrophidion brunneum. These specimens had the uniformly green/brown dorsal pattern characteristic oe southern populations. The absence of small juveniles attributable to D. brunneum was puzzling until it became clearer how to distinguish juveniles of D. prolixum, D. graciliverpa, and D. brunneum brought about by work on the percarinatum complex. Consequently, I now believe that Cadle (2010) erred in referring several small specimens from western Ecuador to “D. percarinatum (= Py graciliverpa). These specimens are D. brunneum having distinctly crossbanded patterns characteristic of juveniles of most species of Dendrophidion. The insight into the identity of these specimens was provided by ANSP 18122 (Chimborazo_ province: Figs. 36A, B); a male of 194 mm SVL having distinct pale crossbands on the anterior body (present posteriorly as well but there is a general fading on the posterior body); the posterior body has dark paravertebral punctations virtually identical to the above-described adult punctate pattern. Although I had ue the specimen fas the 2010 study, I subsequently examined the internal mor- phology of one of its retracted hemipenes and confirmed that its morphology SYSTEMATICS OF Figure 36. ANSP 18122 dorsal and anterior body detail (Chimborazo province, Ecuador). (C) and (D) BMNH 1930.10.12.22 dorsal and ventral (Loja province, Ecuador). Juveniles of Dendrophidion brunneum. (A) and (B) DENDROPHIDION PERCARINATUM OT (COLUBRIDAE) ¢ Cadle 31) conforms to other retracted hemipenes of D. brunneum (robust morphotype and extensive apical nude area; see Fig. 3). | now believe that two females having similar patterns that previously ide tified as “2: percarinatum” (= D. grac iliverpa), ANSP 5709 and FMNH 16942 | 317 and 284 mm SVL, respectively; Cadle, 2010: figs. 3A, C), also are juvenile 7 brunneum. Two other small banded juveniles that I identify as D. brunneum are BMNH 1930.10.12.21-99 (239° and: 219 “mm SVL, respectively; the first is illustrated in Figs; 36C, D). These were collected along with two adults of the unicolor pattern Grom the same locality (BMNH_ 1930.10.12.20, 1931.11.3.10). Bands on the neck extend ventrally to dorsal row | or to the edges of the ventrals, but on the posterior body the bands are seemingly restricted to the middorsal area tthe. posterior dorsum has the appearance of squarish blotches sepa- rated by pale interspaces; Fig. 36C) . The pale dorsal crossbands of these two speci- mens are much more distinct in areas without stratum corneum than when the stratum corneum is present (compare anterior and posterior body in Fig. 36C). No typical adult D. grac iliverpa are known from the highlands of southern Loja prov- ince (Fig. 27), whereas D. brunneum adults were eolletied at the same locality as the two juveniles. In BMNH 1930.10.12.21-22 there is the appearance of transverse dark ventral lines (Fig. 36D). In part, this is an illusion because the poor state of preserva- tion of these specimens resulted in clearing along sutures between adjacent ventral plates: the cleared areas appear dark gray, giving the illusion of distinct transverse ae Ss. Nonet heless, on the posterior body these specimens have somewhat heavier dark stippling along the anterior edges of the ventral plates but nowhere distinct crosslines. The juveniles I refer to Dendrophidion brunneum may be distinguishable from juvenile D. graciliverpa in having slightly broader pale bands on the necks In D. graciliverpa the neck bands are less than one dorsal scale row wide, whereas in D. 316 Bulletin of the Museum of Comparative Zoology, Vol. 160, No. 6 brunneum these bands are 1—-1.5 rows wide. Larger samples would be needed to confirm whether this apparent difference holds. Similarly, the supralabial/temporal pattern can provide a clue, but this character is ambiguous in some specimens—BMNH 1930.10.12.21-22 and USNM 237079 have the P pattern typical of D. brunneum, but USNM 237080 has irregular (fragmented) temporals, and ANSP 18122 has the G pattern on both sides. Covariation of Other Characters with Coloration Morphs. Given the apparent color polymorphism in Tension brunneum as I now conceive it, the question of how other characters covary with the patterns should be addressed, something I have not undertaken in any detail. Cae tion and other external characters of the “new” specimens are similar to my previous summary (Cadle, 2010) except that several specimens extend the lower end of the distribution of subcaudal counts: for males to 125 subcaudals from 139 (BMNH 1860.6.16.60 from “western Ecuador’), for females to 118 subcaudals from 135 (BMNH 1860.6.16.58, also “western Ecua- dor”). Everted hemipenes very similar to those I described earlier for the unicolor morph (Cadle, 2010) are confirmed for the striped morph (KU 142802) and for the punctate morph (UF 88467). The internal morphology of retracted hemipenes (with characteristic extensive apical nude region: Cadle, 2010; Fig. 3) was coubeucd in additional specimens, and hemipenial length to verify “robust” me eae was ded in still others (see Table 5) Critical to thorough investigation would be more extensive knowledge of color variation in life and whether there is any geographic segregation of the color morphs, as appears to be the case for the unicolor morph. This should also help interpret patterns of specimens already on museum shelves. Of particular interest is the vicinity of Loja city in southern Ecuador where three pattern morphs occur: striped (KU 142802, which also has elements of the punctate morph; see Fig. 34A), unicolor (USNM 237077, 237080-S8S1: BMNH 1930.10.12.20), and banded (USNM 237078-79) (Figs. 32, 33C, 34A, D). The last two color morphs occur together at La Argelia, Loja. A more detailed eae sis of color variation in this region should give some insights into wheth- er and how the color morphs intergrade. Since the possibility remains that more than one taxon still resides under the name brunneum, the Loja region might provide critical insights into this systematic question. The realization that four species of the percarinatum ¢ group occur in western Ecua- dor Eee advances comprehension of the systematics of Dend rophidion in this area but is only a first step toward full under- standing. Supplementart Notes on Hemipenial Morphology. Cadl (2010) described evert- ed hemipenes of Dendrophidion brunneum in detail. Little can be added to that report, but some minor differences were noted in the hemipenes of specimens examined in this study. Just as in tones of the three species ‘described in detail in the final section of this paper, there is apparent variation in the extent of development of asulcate apical calyces and the distal flounce- like structures in D. brunneum. KU 142802 (Loja eee) and UF 88467 (Guayas province ) have everted hemipenes. In KU 142802 the fully everted, but not maximally inflated, hemipenis shows fairly well dev el- oped asulcate apical calyces that extend nearly to the center of the apex: otherwise, the extensive nude apical region is similar to organs previously desea (Cadle, 2010). Spine numbers were not reported by Cadle (2010) but KU 142802 has ap roximately 150-160 spines on each organ. The enlarged sulcate spines are only marginally larger than other spines in the array, which is about five to six rows across on the sulcate and asulcate sides, three to four rows on the lateral sides. The tip of the sulcus spermaticus is slightly eee es in both KU 142802 and UF 88467, but without a distinct tissue ridge separating the divergent lips; the expansion is more difficult to see when the tissue is stretched upon full inflation. Two retracted organs SYSTEMATICS OF DENDROPHIDION PERCARINATUM (COLUBRIDAE) from Guayaquil (USNM 237059-60) show differential development of the asulcate apical calyces and the third flounce (both more fully dev elope din USNM 237059). Just as in the he smipenes described later herein, there may be significant variation in the deve lopment of ornamentation in D. brun- neum but more everted hemipenes would be needed to evaluate this possibility. Retracted hemipenes of D. brunneum are S—10. sub- caudals long, with most between 8 and 9 subcaudals in length; these data are present- ed in Table 5 where comparisons with the other species are made. Distributional Notes and Sympatry of Dendrophidion brunneum and D. gracili- verpa. Cadle (2010) found no veabable ay (<1,000 m) records of Dendrophi- dion brunneum in Ecuador, although he considered several lowland iscainese from northern Peru valid. The new specimens (Appendix 1) show that this species as presently conceived is distributed widely in western Ecuador, particularly from the Rio Daule/Babahoyo system and southward (Fig. 37). Contrary to what I earlier claimed (Cadle, 2010: 10), I now believe that the e locality “Guayaquil” is a valid locality D. brunneum since other specimens from Guayaquil and nearby can now be documented: two juveniles reidentified above (ANSP 5709, FMNH 16942), MCZ R-8393, and USNM 237059-60. MCZ R- 8393 was collected by Edward Whymper and is apparently the specimen fenoned by Boulenger (1882: 462; 1891: 132), oho pointed out the accuracy of Whymper’s localities. Cadle (2010: 11) had pondered the whereabouts of this specimen because it had disappeared from the BMNH by the time Boulenger’s Catalogue appear ed (Bou- lenger, 1894); for years the specimen had been identified in the MCZ as D. dendro- phis. Likewise, USNM 237059-60 were obtained by Gustavo Orcés-V., and there seems no reason to question dhe locality. Sympatry of D. brunneum and D. gracili- verpa at Guayaquil is confirmed Re three males: USNM_ 237059-60 (brunneum., hemipenes to the middle of subcaudal 8 ¢ Cadle ke ag and the end of subcaudal 9, respectively) and USNM 12268 (graciliverpa, hemipenis to the proximal edge of subcaudal 13). In addition to length, the internal hemipenial morphol- ogy of these he mipe nes was confirmed. The re 1S a curious gap between the two northern localities of Dendrophidion brun- neum (Fig. 37) and those farther south. The northern Tecatiae ‘s are represented by males so. that hemipenial morphology could be confirmed (UMMZ 83706, USNM 237083). In between the two northern localities and the next one to the south, many specimens of D. graciliverpa have been obtained from the intensely worked region in the vicinity of Santo Domingo de he Colorados and the Rio Palenque Science Center (Fig, 22 Yeu-1): brunneum has not turned up in this area judging by specimens in U.S. collections. The reason for ane gap is not obvious, but gaps seem more common in the distributions of some species of Dendrophidion than in other codistributed snakes (e.g., D. clarkii in Costa Rica and western ae oe J. E. Cadle and J. M. Savage, unpublished data). HEMIPENIAL MORPHOLOGY IN THE DENDROPHIDION PERCARINATUM COMPLEX The distinction between the “robust” and “gracile” hemipenial morphotypes was out- lined at the outset. In this section I describe in detail everted and retracted hemipenes of Dendrophidion percarinatum (robust mor- photype) and D. prolixum and D. gracili- verpa (gracile morphotype). Cadle "(2012: 217-220) gave an overview of the structure of Dendi cophidion hemipenes and terminol- ogy as applied here and detailed descrip- tions of three species in the D. vinitor complex. Cadle (2010) described the hemi- penis of D. brunneum. Hemipenes were studied in both retracted and everted conditions using methods outlined by Myers and Cadle (2003). Descriptions of ‘everted hemipenes are based on organs that were fully everted in the field at the time of preservation. Some mineralized hemipenial structures were visualized by staining with Alizarin Red S (Cadle, 1996: 35). 318 Bulletin of the Museum of Comparative Zoology, Vol. 160, No. 6 Figure 37. referenced by Cadle (2010) are also included. The length of retracted hemipenes was scored as the number of subcaudals (or fraction) subtended by the organ. Most specimens for which retracted organs were examined were large enough to score fractional subcaudals converted to a num- ber. For example, a hemipenis extending to the suture between subcaudals 7 and 8 was scored “8” (i.e., extending to the proximal suture of subcaudal §). A hemi- penis extending to the middle of subcaudal S was scored as “8.5.” Hemipenes extend- ing to points between the proximal suture and midpoint, or between the midpoint 600 m contour above 3000 m Distribution of Dendrophidion brunneum in Ecuador based on the new material examined. Ecuadorian localities and the distal suture of subcaudal 8, were scored as “8.25” and “8.75,” respectively. This system clearly involves some subjec- tivity for pliable tissue and is more difficult for small specimens, but the resulting difficulties are ameliorated by the summa- ry score groupings used for comparisons (Table 5). Because I was comparing spec- imens in some cases of a considerable size range, I was conscious of a possible relationship between snake size and re- tracted hemipenial length. However, there seemed to be no substantial relationship between size and hemipenial length— SYSTEMATICS OF DENDROPHIDION PERCARINATUM (COLUBRIDAE) TABLE 5. PERCARINATUM ARE SORTED BY GEOGRAPHIC ORIGIN. NV N 5.5-5.75 6.25-7 D. percarinatum Honduras y} i} Nicaragua 3 Costa Rica 12 ] Panama Bir i) 6 Colombia 4 D. prolixum é D. graciliverpa 20 4 S D. brunneum 4 2 juveniles had hemipenes within the range of adult hemipenial lengths as measured by subcaudals. Dendrophidion percarinatum Hemipenes of Dendrophidion percarina- tum as redefined here are similar in basic structure from Honduras to Colombia (particularly when compared with the two new species described herein). Nonetheless, there is variation in virtually all aspects of hemipenial morphology, including the form of the spines and the degree to which fully formed apical calyces are developed. De- spite this var iation, hemipenial morphology does not suggest cryptic species within D. percarinatum, as was the case in the D. vinitor complex (Cadle, 2012). The basic structure of hemipenes of D. percarinatum is illustrated in Figure 38 with hemipenes from Honduras and Costa Rica. Some of the variation in structure is shown in Figures 39-40 using specimens from Costa Rica, Panama, and Colombia. However, the structural variation described is not restrict- ed to any particular geographic region, although individual hemipenes may have unique features, as exemplified by the seemingly unique shape of UMMZ 124061 (Barro Colorado Island, Panama; Figs. 39C, 40). Sample sizes for everted hemipenes © Cadle 319 LENGTH IN SUBCAUDALS OF RETRACTED HEMIPENES OF FOUR SPECIES OF THE DENDROPHIDION PERCARINATUM SPECIES GROUP. DATA FOR D. SAMPLE SIZE. Length (No. of Subcaudals) 125-8 §.25-9 95-10 10.5 | | | 5 5 | 1] 14 3 | 3 | IR is es ig [295-13 1325-14 142515 Hayes) | 3 i | 2 6 4 4 §.25-9 9.5-9.75 10.25 1] 3 | examined are: Honduras (9), Nicaragua (1), Costa Rica (14), Panama (6), Colowibin (2. The following description is a composite derived from several specimens. Everted. Typical everted hemipenes from throughout the range of Dendrophidion percarinatum have a narrow, relatively short hemipenial body proximally and a bulbous distal section comprising the spinose region and apex (Fig. 38). The hemipenial body proximal to die enlarged spines comprises one half or less of the length of the organ overall: it is ornamented math minute spines in a broad band proximal to the enlarged spines. The base of the hemipenial body is nude. In adults everted hemipenes are about 25 mm or less in length. Spines are arranged in about three loosely arranged rows all Sout (perhaps slightly fren on the lateral sides). The spinose section is followed distally by two closely spaced flounces that completely encircle the organ except where interrupted by the sulcus spermaticus. The proximal flounce is broader than the distal flounce and has a more extensive fleshy base. The distal flounce forms a definitive border around the periphery of the apex. Both flounces have embedded spinules (Cadle, 2010: 19). Between the flounces are a few low, longitudinal connections, which are concen- 320 Bulletin of the Museum of Comparative Zoology, Vol. 160, No. 6 flounces parasulcus pad Figure 38. Hemipenial morphology of Dendrophidion percarinatum. (A) and (B) Sulcate and asulcate views of USNM 559611 (Honduras). (C) Sulcate view of LSUMZ 34112 (Costa Rica). (D) Apical view of USNM 559611, sulcate edge at bottom. Abbreviations: N, nude patches on lateral sides of apex; ss, sulcus spermaticus. Figure 39. Variations in hemipenial morphology of Dendrophidion percarinatum. (A) AMNH R-108468, sulcate view (Colombia). (B) USNM 259130, sulcate view of distal region (spines + incompletely everted apex) (Costa Rica). Arrows indicate sheathed spines. (C) UMMZ 124061, asulcate view, alizarin stained (Panama). All specimens have unusual formations of spines and/or flounces and other ornamentation compared with typical organs (Fig. 38). See text for full explanation. SYSTEMATICS OF DENDROPHIDION PERCARINATUM Figure 40. Hemipenis of Dendrophidion percarinatum with an unusual morphology. UMMZ 124061 from Barro Colorado Island, Panama. (A) Sulcate and asulcate views. (B) Apical view, sulcate side toward the bottom. Stained with alizarin red so that embedded spinules are visible within the flounces. Apical view shows the near absence of ornamentation. See text for full explanation. trated on the asulcate side and often incompletely connect the flounces, forming poorly dev eloped calyces. On the Anca side of the apex directly opposite the tip of the sulcus spermaticus are several calyces distal to the distal flounce. These calyces are within a triangular t thickened expanse of © Cadle AY bo ((COLUBRIDAE) tissue that narrows distally and ends Oppo- site the tip of the sulcus spermaticus just short of the apice al center (Fig. 38D). These calyces are subje ct to some variation in development described below, but when fully He veloped, there are about three calyces adjacent to the distal flounce, several more distally, and usually a low ridge of tissue bisecting the triangular raised tissue within which he calyces sit. Periph- eral to the asulcate triangular calycular tissue and structures associated with the tip of the sulcus spermaticus, the apex is nude. The sulcus spermaticus is simple and centrolineal and ends with a slightly flared tip just short of the center of f the apex; the flared tip can only be seen by parting and lifting the bulbous tissue on each side. After traversing both flounces, the sulcus sperma- ticus is bordered on each side by a small, roughly triangular pad of raised tissue. Within each pad is usually a single depres- sion just distal to the distal flounce; 1 interpret these as rudimentary calyces (parasulcus calyces or parasulcus pads when no pense calyx is present). Variation in Hemipenial Morphology of Be ary hieien percarinatum. With a clear exception described below, hemipenes of Dendrophidion percarinatum from throughout the range are similar in form. The shape of the distal portion (spines + apical region) is sometimes more elongate, sometimes more rounded, but it is unclear how much of this variation is simply due to differences in the preservation of this pliable tissue. The length of everted hemi- penes of D: percarinatum is 16.6-25 mm in several adults (compare D. prolixum and D. gracilive rpa). There is minor variation in the aaner of spines in the array, as shown by these counts from everted hemipenes given as (range, mean, sample size); counts include the two enlarged sulcate spines: Honduras (31-37, 33.4, 9), Nicaragua (34, N =), Costa: Rica.(2640132;6; 14)r Panama (30=36,.31.8. 6), Colombia (30=36;33,.2)In most hemipenes the enlarged sulcate spines at the proximal edge of the array (see Cadle, 309. Bulletin of the Museum of Comparative Zoology, Vol. 160, No. 6 2012: 218) are distinctly larger than any other spines, but in some organs the size distinction is much more subtle. There is much more variation in the development of apical ornamentation and in the morphology of individual spines. Figure 38D illustrates a hemipenis with fully developed apical ornamentation (asul- cate triangular raised tissue with calyces, and thick pads bordering the apical portion of the sulcus spermaticus). The asulcate calyces have well dete walls and fully formed cuplike structure. The triangular expanse of tissue within which they sit abuts the distal flounce on the asulcate side of the apex (the distal flounce forms the proximal transverse walls of the most proximal calyces; these proximal calyces are visible at the distal tip of the organ in asulcate view, Fig. 38B). Typically, the asulcate series of calyces comprises three proximal calyces followed by two more distal and then several more irregular calyces toward the center tip of the apex. On the opposite side of the apex each of the sulcate pads has a shallow depression, which I interpret as an incom- pletely formed calyx; fully formed calyces within these pads were not observed in the hemipenes I examined. Hemipenes with this fully developed dees: (Fig. 38D) are found through- out the range of Dendrophidion percarina- tum. However, also throughout the range are hemipenes that vary in the extent to which the ornamentation is developed. In yarticular, substantial differences in the ae of individual spines and the develop- ment of the asulcate apical calyces can result in organs that appear strongly differ- entiated from those with fully developed ornamentation. The strong delimitation of the apical nude areas by raised ridges observed in USNM 559611 (Fig. 38D) is attained in only a few hemipenes I exam- ined, although these nude areas are present in all organs. I resist the temptation to state what is “normal” development in D. percar- inatum because hemipenes with rudimen- tary ornamentation were common among the organs I examined. | Some of the variation is illustrated in Figures 39-40 (variation in spine morphol- ogy is discussed below). AMNH R-108468 (Fig. 39A) from Colombia shows consider- able reduction of the spines to mere nubbins, as well as more subtle reductions in other ornamentation. The asulcate caly- ces are reduced to irregular calyxlike structures and ridges (only a single defini tive asulcate calyx is present, located adja- cent to the distal flounce), and its flounces are narrower and less projecting than in many other hemipenes. The only other available everted hemipenis from Colombia (AMNH R-123745) has typical spines and more definitive apical calyces. Other hemi- penes show reductions similar to AMNH R-108468: when such reductions occur, the resulting structures are usually more irreg- ular than when the full complement of ornamentation is present. Reduction in one aspect of satenolee? (e.g., spines) usually entails reductions in others (e.g., calyces). But there are exceptions—USNM 259130 from Costa Rica has reduced spines but unreduced apical calycular — structures (Fig. 39B). The most extreme modifications were observed in a single specimen of Dendrophidion percarinatum from Panama, whose characteristics are described in detail. The asulcate apical calycular area can sometimes appear as a third incomplete (noncircumferential) flounce on the asul- cate edge of the apex, with a few incomplete longitudinal calycular walls. For example, AMNH R-123745 has a single small meee and mostly complete calyx just distal to the circumferential flounces (proximal wall formed by the distal dontenl Distal to the median calyx is an undulating wall (incom- plete flounce) centered on the asulcate side and extending about one-third of the distance around the apex. Underneath this incomplete flounce are about four low, incomplete longitudinal walls, partitioning the space between the incomplete flounce and the small median asulcate calyx. A Most Unusual Hemipenis of Dendro- phidion percarinatum. Hemipenes — of UMMZ 124061 from Barro Colorado SYSTEMATICS OF DENDROPHIDION PERCARINATUM (COLUBRIDAE) Panama, were the most peculiar organs I studied, and both are modified eailarls (Fig. 40). They have an unusual overall nape (a longer proximal portion of the hemipenial body ae, less bulbous distal region than is typical), and nearly all aspects of their ornamentation are ee d (Figs. 39C, 40). These hemipenes for a time misled me as to what the typical morphology for Dendrophidion percarinatum hemi- penes actually was—hemipenes of UMMZ 124061 were among the first of this species I studied in detail ond I entertained ae idea that cryptic species were involved. Exami- nation of other organs convinced me otherwise, and I now view the morphology of UMMZ 124061 as one extreme within a species in which variation in hemipenial morphology seems unusually great. Other specimens from BCI or the adjacent Canal Zone (e.g., CM 6869, KU 80589, UMMZ 297811) have hemipenes more typical in shape and ornamentation than UMMZ 124061, although their ornamentation varies within the limits described above. The unusual shape of UMMZ 124061 does not appear to be due to preservation artifact, which in any case, would not affect the peculiar ornamentation of this specimen. I stained the right hemipenis of UMMZ 124061 with alizarin red to visualize miner- alized structures more fully. Mineralized embedded spinules are seen in Figures 39C and 40 as dark parallel streaks wathin the flounces and other calycular structures. The mineralized portion of the spines is con- cealed by thick fleshy tissue. The narrow proximal portion of the hemipenial body has a sparse covering of minute spines on its distal half, extending farther proximally on the asulcate side than on the sulcate side (Fig. 40A; the alizarin-stained minute spines appear as tiny dark spots on the proximal portion). There are about four rows of enlarged spines adjacent to the sulcus spermaticus, narrowing to two rows on the lateral surfaces and continuing in two rows to the asulcate side. The ‘ ‘emia ged” sulcate spines are scarcely larger than other spines in the array. The Rieu spermaticus is Island, ¢ Cadle CO bo WwW centrolineal and, after transecting the flounces, ends somewhat short of the middle of the apex. Its tip is distinctly flared and even appears some what divided by a low wedge of tissue separating the divergent lips. Individual spine s have a very strange structure. Each spine is envelope ‘d by thick fleshy tissue. Close inspection shows that the tissue envelops the tip of the spine like a hood, covering ae spine surface toward the apex (i.e., the Gis surface) and the hooked tip. The result is that the tips of the spines are not visible except by lifting up the fleshy covering to expose them: without lifting the sheath, ahs area of the hemipenial body appears to be ornamented with blunt fleshy projections, as they appear in Figures 39C and 40A. The alizarin staining showed that the spines were indeed painerdlived under- neath and within the fleshy covering (the proximal portion of the spines extended into the fleshy sheath, even though the spine tips were free), This peculiar structure is re- seated to a greater or lesser extent in many aes of Dendi ophidion percarinatum (discussed below). Distal to the enlarged spines are two circumferential flounces similar to those on other D. percarinatum organs. On the asulcate side distal to the flounces is a single very large, asymmetrically positioned calyx (Fig. 39C). Two longitudinal, poorly developed calycular walls faxther subdivide this calyx. The distal wall of the asulcate calyx forms a sharp angle on the edge of the apex, from which a love fleshy ridge extends nearly to the middle of the apex. On the sulcate side distal to the flounces, the sulcus spermaticus is bordered on each side by a single incomplete calyx with somewhat thickened walls (the wall adjacent to the sulcus is not clearly defined). Alizarin staining showed that poorly formed embed- ded spinules were present at irregular intervals in both the sulcate and asulcate calyces. Morphology of Hemipenial Spines in Dendrophidion. The peculiar structure of the spines just described for UMMZ 324 124061 is not an isolated case either within D. percarinatum or more broadly in the genus. In many Dendrophidion hemipenes the spines are of the typical morphology seen in most snakes. In D. percarinatum aici species of this species group typical spines are relatively short and strongly hooked at the tip (e.g., Fig. 38). However, in other hemipenes the ae are partially or entirely enclosed by a fleshy sheath and, in some cases, seem not only enclosed but also much reduced in size (Fig. 39A). These sheathed (or celate, from celatus, Latin for “hidden*) spines have a blunt, fleshy appearance, but a mineralized rod within the fleshy tissue can be demonstrated by ne by shining a strong light through the translucent tissue, or by staining with alizarin red. In some sheathed spines the mineralized portion does not project at all from the fleshy portion; in others a small point or hook protrudes through the tip (Fig. 39B, left arrow). The enveloping fleshy tissue was, in some cases such as UMMZ 124061 described above, hoodlike, whereas in others it seemed to form a more complete sheath surrounding the entire spine. Whether spines are typical or sheathed varies intraspecifically in species of both the TER aed dbisdse and D. percar- inatum species groups. Dendrophidion per- carinatum seems especially prone to this type of variation, but this could be because more everted organs have been available for this species than for any other. I detected no geographic trends in spine morphology (the variation occurs throughout the geographic range of the species). Some hemipenes I examined had a mix of sheathed ee ical spines, whereas others were entirely of one type or another. In occasional specimens (e.g, USNM 259130), the spines of one hemipenis were predominantly of typical morphology, whereas the other had celate spines. Considering the entire array of spine morphologies in Dendrophidion, there is probably a continual gradation between the celate and typical forms. What I refer to as celate spines may be equivalent to similar structures reported in Bulletin of the Museum of Comparative Zoology, Vol. 160, No. 6 some other snakes, for example “fleshy ae age ase that form the swollen ases of spines” on some elapid hemipenes (Keogh, 1999: 250). Keogh (1999) scored this as a separate ornamental category than spines for Australian elapids and reported no intraspecific variation (fleshy protuber- ances were either present or absent for a given species). Smith (1943: 151) stated that the short spines on the hemipenis of Elaphe taeniura were “enclosed in a voluminous sheath.” Myers (1974: 33) reported an aberrant specimen of Rhadinella godmani in which hemipenial spines were absent, but their locations were indicated by “enlarged tissue bases.” Given the prevalence and seemingly continuous and intraspecific var- iation between celate and typical spines in Dendrophidion, it seems best to recognize celate spines as a common variant of spine STOR OWGlOEe rather than anomalies or an entirely different kind of hemipenial orna- mentation for this group of snakes. Possibly, the variable development of spines is related to other reduced aspects of hemi- penial morphology in Dendrophidion, such as the variable development and manifesta- tion of calyxlike structures. Data for several snake species indicate that spines, calyces, and flounces derive from common anlagen early in development (Clark, 1944). Retracted Hemipenes of Dendrophidion percarinatum. Retracted hemipenes of Den- drophidion percarinatum extend from the middle of subcaudal 5 up to the middle of subcaudal 10 (Table 5), with a clear modal length between 7 and 9 subcaudals. The retractor eet magnus has a very short proximal division. The length of retracted hemipenes observed within D. percarina- tum (a 5-subcaudal span) is comparable to the span observed in some other snakes (e.g., Rhadinaea decorata |9-subcaudal span]; Myers, 1974: 75), Dipsadoboa uni- color and D. weileri (5-subcaudal spans; Rasmussen, 1993: 146), and Tantilla spp. (several species with 5—7-subcaudal spans; Cole and Hardy, 1981: 225-234). The morphology of the retracted hemi- penis corresponds in a_ straightforward SYSTEMATICS OF Figure 41. tum (AMNH R-17374, Costa Rica). Abbreviations: ESS, enlarged sulcate spines; F, flounces; N, nude apical areas; ss, Sulcus spermaticus. Retracted hemipenis of Dendrophidion percarina- Distal toward the top. manner to the everted organ (Fig. 41): proximal portion ornamented with minute spines; a spine array including a pair of spines slightly larger than others adjacent to sulcus spermaticus at the proximal edge of the array (enlarged sulcate spines); two flounces followed distally by an apex with asulcate calyxlike structures and_ lateral nude areas; and a simple sulcus spermaticus in its dorsolateral wall, ending with a slightly DENDROPHIDION PERCARINATUM A bo Ul (COLUBRIDAE) ¢ Cadle flared tip at the distal end of the retracted organ. A few weak longitudinal connections are present between the flounces. Dendrophidion prolixum Everted (AMNH_ R-108469, Choco de- partment, Colombia; left hemipenis, Fig. 42). Overall morphology is the “gracile” form. Total length of organ approximately oS— 30 mm. Le neth of the bulbous distal region from the base of the e nlarged sulcate spines to the Up of the apex is 8.7 mm (30% of total length). Length from proximal flounce to tip of apex approximately 4.5 mm. The hemi- penis is unilobed, somewhat clavate, and without basal pockets or lobes. Sulcus spermaticus is centrolineal and ends just short of the center of the apex. The portion of the sulcus distal to the flounces is bordered by a thick pad of tissue on each side: no de pression or other indication of rudimentary calyces were evident within these pads. The tip of the sulcus is slightly expanded and with a narrow we .dge of tissue between the diver- gent lips, resulting in the appearance of a terminal division. The edge of the tissue bordering the short branches of the sulcus is seemingly somewhat thickened (denser white compared with adjacent wedge tis- sue), perhaps indicative of lip tissue. How- ever, the wedge can be manipulated and flattened unlike in a truly divided sulcus. Each short branch of the “divided” tip is slightly expanded into a teardrop-shape at its tip. Hemipem ial body proximal to the spine array is ornamented with minute spines all around except for a nude patch comprising about the basal one-third of the organ on the left side; otherwise, the minute spines go all the way to the base of the organ and are 1 rather densely arrayed. These minute spines are seemingly somewhat more densely arrayed toward he spine array than more basally. The enlarged sulc ‘ate spines at the prox- imal edge of the spine array are about two or thvee times the size of others in the battery (enlarged spine on the right side 326 Bulletin of the Museum of Comparative Zoology, Vol. 160, No. 6 Figure 42. larger than the one on the left side of the sulcus). Total number of spines in the battery approximately $7+2 enlarged sulcate spines. Spines are in four to five loosely arranged rows on the sulcate and asulcate sides, three rows on the lateral sides. Distal to the spines are a circumferential flounces. The proximal flounce is broader than the distal flounce and has an outer more membranous portion and a somewhat fleshy base. The distal flounce has only a very narrow outer membranous portion; most of this flounce is fleshy. Embedded spinules are in the me ‘mbranous portions of both flounces. The borders of the flounces are smooth (non- crenulate) but somewhat undulating. On the asulcate side of the apex is a triangular is of tissue extending toward the center of the apex from the distal flounce (point tow ard the apex). Within this pad is an irregular, poorly differentiated depression (rudimentary calyx) with fleshy edges. Except for this asulcate pad and pair of Everted hemipenis of Dendrophidion prolixum (AMNH R-108469) in sulcate and asulcate views. those bordering the distal portion of the sulcus spermaticus, the apex is nude. A pair of shallow dimples, prob yably the points of internal attachment of the retractor penis magnus, is in the center of the apex. Retracted (FMNH 54960, Risaralda de- partment, Colombia; Figs. 43A, C). The left hemipenis had been previously slit some- what irregularly on its ventral/medial sur- fee Left and right hemipenes extend to middle of pabeandal 11. Retractor penis magnus appears proximally undivided, but there may be some separation of the muscle fibers. Sulcus spermaticus simple, but its tip is flared, and it ends at the tip of the organ. The entire portion of the sulcus distal to oe flounces is bordered by a thick pad of tissue on each side: each of these contains a single, very shallow calyx. The long proximal portion of the hemi- penial body has long longitudinal folds and minute spines. On the sulcate side these spines extend nearly to the base of the organ, but toward the asulcate side a long SYSTEMATICS OF DENDROPHIDION PERCARINATUM B P| eS. | BO ~ (COLUBRIDAE) ¢ Cadle asulcate Figure 43. Retracted hemipenes of Dendrophidion prolixum and D. graciliverpa. Distal toward the top. (A) D. prolixum (FMNH 54960; slit medially). (B) D. graciliverpa (UIMNH 77347; slit midventrally). (C) Apical region of FMNH 54960 with landmarks indicated. Abbreviations and labels: ESS, enlarged sulcate spines; F, flounces; N, apical nude areas; ss, sulcus spermaticus. Asterisks (*), thick pads of apical tissue bordering the distal part of the sulcus. Asulcate pad, thickened asulcate apical tissue bearing rudimentary calyces. section of the organ is nude, and_ the minute spines are restricted to the distal third to half of the basal section. The array of enlarged spines is broader on the suleate and ee sides (about four to five rows across), and narrower on the intervening sides (lateral sides of the everted organ; about three rows across). A pair of enlarged sulcate spines at the proximal edge of the array is much larger than other spines. The spines are arrayed along the longitudinal folds occupying this area of the organ. Distal to the spine array are two ne which are connected at intervals with poorly developed longitudinal ridges. Each flounce comprises a basal fleshy portion and an outer membranous portion with embedded spinules. The embedded spinules are short straight splints about the same width throughout. Distal to the flounces is a very short apex, nude except for a thick pad of tissue extending from the distal flounce to the tip of the organ and the pair of pads bordering the sulcus spermaticus. The asulcate apical pad seems to have two or more poorly Teele calyces (these are difficult to probe apart), Variation in Hemipe nial Morphology. ‘verted hemipenes of only four specimens of Dendrophidion prolixum were available ee) bo (AMNH R-1L08469: Re12375025 Re 109726). There is little variation in mor- phology among these except for the asulcate apical calyx and associated structures, the calyx being more definitive in some organs than in others (similar to the variation summarized above for D. percarinatum and for D. graciliverpa described below). The “enlarged” sulcate spines are some- times scarcely larger than other spines in array. The approximate number of spines in these organs is as follows: AMNH R- 123750, 89+2 (left organ); AMNH R- 109726: 654+2 (right), 70 + 2 (left); AMNH R-123751: 87+2 (right and left). The right hemipenis AMNH _ R-109726 had a total length of 42 mm, and the length from the enlar ged sulcate spines to the tip of the apex was 7.9 mm (19% of the total length). Retracted hemipenes vary in length from nine subcaudals (one organ only) to 15.5 subcaudals (Table 5). Dendrophidion graciliverpa Everted (AMNH R-110584; holotype, E] Oro province, Ecuador; right hemipenis) (Figs. 44-45). Total length about 39 mm. Length of apex from the proximal enlarged sulcate spine to tip of apex, 9.6 mm (distal section about 25% of the length of the organ). Maximum width of expanded apical region, §.4 mm. Hemipenial body proximal to spines compressed laterally (i.e., broader when viewed from lateral side fia when viewed from sulcate or asulcate side). Overall form of “gracile” morphology. The hemipenis is unilobed, with a gradually expanded distal section and without basal pockets or lobes. The long section of the hemipenial body proximal to the enlarged spines is ornamented with minute spines except for a nude basal patch on the right lateral and asulcate sides. Sulcus spermaticus centrolineal, extend- ing to the center of the apex, terminally da ided with equal branches about 1.1 mm long (Fig. 45). The divergent branches are separated by a thick triangular wedge of tissue that may have distal “lip tissue” 8 Bulletin of the Museum of Comparative Zoology, Vol. 160, No. 6 (hence, truly divided although this would have to be verified histologically). The enlarged spine array contains about 116+2 proximal sulcate spines about three times larger than any other spines in array. Spines in about five to six loosely arranged rows on the sulcate and asulcate sides, about three rows on each lateral side. Spines in the array somewhat larger proximally than distally. Spines followed distally by two circumferential flounces bearing embedded spinules, complete except where transected by the sulcus spermaticus; proximal flounce broader than distal one. Flounces have a thick fleshy inner portion and outer mem- branous portion; the spinules span the width of the membranous portion but barely enter the fleshy portions. Entire sulcus spermaticus distal to the flounces is bordered by thick triangular pads of tissue (* in Fig. 45). These are broader adjacent to the distal flounce and each has a calyxlike depression. On the asulcate side of the apex, the distal flounce splits to form a large irregular calyx at the sulcate edge of the apical region, the asulcate calyx (Fig. 44B); from the distal wall of this calyx an irregular raised triangular area of tissue extends nearly to the center of the apex just opposite the tip of the sulcus spermaticus. On the apical tip lateral to the sulcus and the asulcate triangular tissue, the apex is nude, very smooth, and strongly demarcated peripherally by the distal flounce (Fig. 45). Variation in Morphology of Everted Hemipenes. Two other everted hemipenes of Dendrophidion graciliverpa were studied in detail: AMNH R-119835 ( (topotype) and USNM 237069 (Pichincha province, Ecua- dor) (Fig. 46). These hemipenes are similar in basic structure to that of the holotype, but there is some variation in the shape of the apex and calyces. Some features are essentially similar to the holotype: the arrangement and relative sizes of spines, the sulcate pads distal to the flounces, the terminal division of the sulcus, the presence (but not shape) of an irregular asulcate calyx, and apical ornamentation. Basic data on the two organs are: SYSTEMATICS OF DENDROPHIDION PERCARINATUM (COLUBRIDAE) ¢ Cadle 329 asulcate calyx Figure 44. Everted hemipenis of Dendrophidion graciliverpa (AMNH R-110584, holotype) in sulcate and asulcate views. (A) and (B) Details of apical region in sulcate and asulcate views, respectively. USNM 237069 (Figs. 46A, C): Total length approximately 26 mm. apex from the proximal enlarged suileate spine to the tip of the apex, 7. 3 mm (distal section of hemipenis 28% of the total length). Tip of the sulcus spermaticus with div ergent lips but not clearly divided as in AMNH R-110584, R-119835. About 84 small spines + 2 enlarged sulcate spines. On the asulcate side of the apex is a thick triangular pad of tissue, within which is a large “irregular calyx whose proximal wall is formed by the distal flounce. The lar ge calyx has several small depressions eaehin it (or rudimentary partitions), which makes it appear that the lar ge calyx is formed by Length of fusion and obliteration of several smaller calyces. AMNH B-119835 (Figs. 46B, D): Total length 29.4 mm. Length of narrow part of body 21 mm igieasared to the base of the pair of enlarged spines at the proximal edge of the apex). Diameter of narrow part 4 x 5 mm. Length of apex from the proximal enlar ged suloate spine to aD of apex, S mm (lisa section of hemipenis 27% of the total length). Diameter of globose part 7 X fe mim, Tip of the suleus spermaticus seemingly divided about 1 mm because a short ridge separates the divergent lips. About 81 + 2 enlar ged sulcate spines. The asulcate calyx is a large oval formed by 330 Bulletin of the Museum of Comparative Zoology, Vol. 160, No. 6 Figure 45. Apical view of tne everted hemipenis of Dendro- phidion graciliverpa (AMNH R-110584, holotype). The contrast is exaggerated so as to emphasize the divided tip of the sulcus spermaticus. Abbreviations and labels: Lower arrow, point of division of the sulcus spermaticus; upper arrows, tips of the branches of the divided sulcus; asterisks (*), thick pads of apical tissue bordering the distal part of the sulcus (parasulcus pads); DF, distal flounce; PF, proximal flounce; ss, sulcus spermaticus. robust, well-defined walls created by divi- sion of the distal flounce (Fig. 46D). Within the asulcate calyx are about six pockets, seemingly formed by thinning of tissue of the hemipenial wall: these pockets are irregular in size and position, although most are Sed under the distal fleshy ee der of the calyx. Immediately distal to the asulcate calyx is “4 “small: deep triangular hole surrounded by thick fleshy ridges (one of which forms part of the distal wall of the asulcate calyx); the ridges surrounding the pore fuse distally, forming a groove that extends distally and ends short Fi: the center of the apex. This hole is presumably a rudimentary calyx. Although all three everted organs of Dendrophidion graciliverpa are basically similar the shapes of their apices differ. The apex and apical AMNH R-110284 (holotype, El Oro prov- ince; Fig. 44) are more similar to USNM 237069 (Pichincha province; Figs. 46A, C) than they are to the topotypic specimen (AMNH R-119835: Figs. 46B, D). The apex ornamentation of of the last is more rounded than in the other two hemipenes, and it has more rudimen- tary dev aien of the asulcate ve apical tissue. On the other hand, the large oval asulcate calyx of AMNH_ R-119835 is a very well defined structure with thick walls (Fig. 46D), whereas in the other two specimens its) amore rudimentary trian- gular structure without hypertrophied walls (Figs. 44B, C). The strong differences among the few everted hemipenes of D. graciliverpa examined suggest that consid- erable variation in hemipenial morphology might characterize this species, just as suggested above for D. percarinatum. How- ever, based on the present small sample it is unclear that this variation is taxonomically significant. It should be noted that hemi- penes depicted in Figures 44 and 46B and D are from “crossbanded” specimens of D. graciliverpa, whereas Figures 46A and C are from the “unicolor” pattern morph, as discussed in the section on coloration of preserved specimens. Retracted. 1 examined the internal mor- phology of five retracted hemipenes from specimens I refer to De ndrophidieae sracili- verpa (ANSP 5519: BMNH 1860.6.16. 59: UIMNH 77347: USNM 12268, 237084). All are very similar to one another and _ the following description is a composite account. UIMNH 77347 is illustrated in Figure 43B. These hemipenes extended posteriorly to the suture between subcaudals 10 and 11 up to the proximal portion of subcaudal 14; when hemipenes that were examined only superficially are included, the lengths were 9-15 subcaudals (Table 5). Measured lengths in adults were 27-40.7 mm. The retractor penis magnus appeared distinctly divided in UIMNH 77347 but in other organs the separation was not so distinct. The sulcus spermaticus is simple and centrolineal (in the dorsolateral wall of the organ). It ends just short of the distal tip of the organ and has a flared tip (a couple of the specimens had a low wedge of tissue between the divergent lips but ‘thas was most likely a simple fold resulting from the packing of apical tissue in the retr acied condition). SYSTEMATICS OF DENDROPHIDION PERCARINATUM WW W (COLUBRIDAE) ¢ Cadle Figure 46. Variation in hemipenial morphology of Dendrophidion graciliverpa. (A) USNM 237069 from Pichincha province, Ecuador (sulcate view). (B) AMNH R-119835 from El Oro province, Ecuador (topotype, sulcate view). (C) Detail of apex, USNM 237069, asulcate side. (D) Detail of apex, AMNH R-119835, asulcate side. Arrows in panel C indicate the poorly formed walls of the triangular asulcate calyx; compare panel D, in which the calyx is oval and has prominent walls. The asulcate calyx in the holotype (Fig. 44B) is more similar to that in panel C than to the topotype in panel D. The proximal part of the hemipenis has many longitudinal ridges or folds, which are pnemicated with minute spines except for the extreme base of the organ. enlarged sulcate spines is positioned at the proximal edge of the spine array, which is about threes or four rows across on the sulcate and asulcate sides, slightly fewer in between. The spine array is follawed distally by a pair of flounces, between which are some rudimentary longitudinal connections. Distal to the flomnces. on the asulcate side (ventromedial side of the in situ organ) is a single large irregular calyx, the proximal wall of Pennelia is foamed by the distal flounce. From this calyx a ridge or pad of tissue extends nearly to the tip of the sulcus spermaticus on the apex. Each side of the A pair of distal portion of the sulcus spermaticus is bordered by a fleshy pad, each with a slight depression. The apex distal to the floumees is very short (about one subcaudal in length) and has thin longitudinal folds (nude tissue in the everted he ‘mipenis), ee are largely obscured by the pleats of the feunces covering it from ventral perspective. COMPARISONS OF HEMIPENIAL MORPHOLOGY OF SPECIES IN THE DENDROPHIDION PERCARINATUM COMPLEX Apart from robust vs. gr acile morphology, hemipenes of the dee species of the Dendrophidion ie ae complex are similar in basic details: (1) proximal portion of hemipenial body ornamented at least in 3 ey) bo part with minute spines; (2) enlarged spine array three to six rows across on the cae and asulcate sides and with a somewhat enlarged pair of sulcate spines at its oroximal edge; (3) two circumferential Hane with embedded spinules; (4) roughly triangular asulcate apical calycular tissue that varies in morphology, but includ- ing at least one large proximal calyx bordering the distal flounce and a more or less defined ridge or other calyces extending nearly to the center of the apex; (5) a fleshy pad (sometimes with a more or less evident calyx within) bordering each side of the apical part of the sulcus spermaticus distal to the flounces; and (6) tip of the sulcus spermaticus terminally flea in everted organs, and sometimes seemingly divided by a tissue wedge in D. prolixum and D. graciliverpa. Some of these characteristics are more widely distributed within Dendro- phidion (Cadle, 2012: 217-220). For exam- ple, two (and no more) complete circumfer- ential flounces are characteristic of all species! im the D. percarinatum species group, and enlarged sulcate spines are characteristic of most species of Dendrophi- dion, although details vary among species. Dendrophidion prolixum and D. gracili- verpa have nearly identical hemipenes and share several putatively derived hemipenial characters compared with other Dendrophi- dion, especially D. percarinatum. Foremost is “gracile” hemipenial morphology, which is unique among known Dendrophidion hemipenes, with the possible exception of D. bivittatum (see the introductory section on gracile and robust hemipenial morpho- types). Gracile morphology actually com- prises several associated, but distinct, char- acteristics: the hemipenial body proximal to the enlarged spines is exceptionally long relative to the spinose + apical regions; the ues + apical region is relatively and absolutely less bulbous than in robust hemipenes; and gracile hemipenes are longer in absolute terms than robust organs (measured length or length in subcaudals for retracted organs). Second, D. prolixum and D. graciliverpa have a large number of Bulletin of the Museum of Comparative Zoology, Vol. 160, No. 6 spines (>60 compared with =40 in D. percarinatum). Such a large number of spines is found elsewhere in Dendrophidion only in D. crybelum (Cadle, 2012: 225), D. bivittatum (about 65 spines in the single organ counted), D. ete auctorum (>85 in one species of the complex; Cadle and Savage, unpublished data), and D. brun- neum, in which spines are very small and can number upward of 150. Hemipenes of D. crybelum and D. brunneum are other- wise dissimilar to D. prolixum and_ D. graciliverpa apart from characters common tom thew: percarinatum species group shared by D. brunneum, D. prolixum, and D. graciliverpa. Third, in D. prolixum and ID) graciliverpa the spines are greatly reduced in size and closely packed together (shared with D. brunneum and perhaps with D. bivittatum). Finally, in everted hemi- penes of D. prolixum and D. graciliverpa the apex is strongly protruding distal to the flounces as a rounded or more angular convexity (compare Figs. 44A, 46C, D). This apical morphology may characterize D. bivittatum (the only everted organ examined is problematic to interpret in this regard), but it also appears rarely in D. percarinatum, as shown by the peculiar everted hemipenes of UMMZ 124061 (Fig. 40). These unusual and _ putatively derived hemipenial characters provide ot evidence that D. prolixum and D. graciliverpa are closely related within Den- drophidion. Further study is needed to evaluate the potential volationsti of WD) bivittatum to this pair based on several shared hemipenial characters. Given the variability that seems to char- acterize hemipenes of the Dendrophidion percarinatum complex, I am hesitant to make much of small differences among the few available everted hemipenial preparations of D. prolixum and D. graciliverpa. The pro- found differences evident between two specimens of D. graciliverpa from the type locality (Figs. 44, 46B, D) give one pause in making too much of some differences— especially given the greater similarity of one of these (Fig. 44) to a Colombian specimen SYSTEMATICS OF DENDROPHIDION PERCARINATUM (COLUBRIDAE) ¢ Cadle of D. prolixum (Fig. 42)! Preparations differ in the shape of the apex, the prominence of the asulcate triangular apical tissue, and the shape and definitiveness of the irregular melts calyx and the parasulcus calyces. Because of the variability in these organs I fail to find consistent differences in hemi- penial morphology between D. prolixum and D. graciliverpa. The strongly differentiated asulcate calyx of AMNH_ R-119835 (Fig. 46D) has no equivalent in any other hemipenis of either species. Similarly, a few other differences, including the degree of compression of the long proximal portion of the hemipenial body, the presence and size of nude areas among the minute spines, the width and morphology of the flounces, and so on cannot be ascribed taxonomic signifi- cance without additional sampling. The Sulcus Spermaticus in the Dendro- phidion percarinatum Complex. In the three species of the Dendrophidion percarinatum complex the sulcus spermaticus in retracted organs ends at the distal tip of the organ, where its lips diverge slightly as the thick parasulcus pads give way to the nude apical tissue (Figs. 41, 43C). This condition is ee in everted hemipenes of D. percarinatum, in which the tip of the sulcus is entirely confluent with the nude apical area and the lateral lips of the sulcus curve sharply around the distal end of the parasulcus pads (Fig. 38D). Everted hemi- penes of Dendrophidion prolixum and D. graciliverpa differ in that a narrow wedge of raised tissue occupies the space between the moderately divergent sulcus lips, resulting in the appearance of a terminal division of the sulcus (Fig. 45). With the small number of preparations of these species available, it is not clear how consistent this apparent division is or whether, in fact, medial “lip” tissue is present on the raised wedge within the divergent lateral lips. In D. prolixum the tissue wedge can be easily flattened with manipulation, suggesting that this morphol- ogy is not a truly divided sulcus. In two hemipenes of D. graciliverpa the sulcus appeared to have a short terminal division, whereas a third seemed to have only B30 divergent sulcus lips. This suggests that there may be some variation in the expres- sion of this feature. Nearly complete evolu- tionary loss of the sulcus bifurcation may result in such variation or configurations that are not easily interpreted with respect to a more fully expressed bifurcation; see Myers (2011: 23-24) for possible routes of evolutionary loss of bifurcation of the colubrid sulcus spermaticus. The apparent terminal division of the sulcus spermaticus in Dendrophidion pro- lixum and D. graciliverpa is unusual within Dendrophidion but not unique. Dendro- phidion dendrophis and D._ atlantica (Freire et al., 2010) have an unambiguously divided sulcus, with fully developed medial and lateral lip tissue along both short branches after the bifurcation. The sulcus is much more deeply forked in D. dendro- phis than in D. prolixum or D. graciliverpa. Moreover, the fork is apparent in both the retracted and everted hemipenes in D. dendrophis (personal observation), as is typical for divided sulci in general and in contrast to D. prolixum and D. graciliverpa. Thus, the sulcus spermaticus in rape i phidion shows a variety of conditions: truly bifurcate in D. dendrophis and D. atlantica, marginally bifurcate or not in D. prolixum and D. graciliverpa, or with an expanded tip (divergent lips) in the other species. The terminus of the sulcus spermaticus in the Dendrophidion percarinatum complex shows similar ambiguity in interpretation that Myers (2011: 22-24) found in Lepto- deira. In some species of Leptodeira the sulcus is “not clearly forked or divided, [and] terms such as ‘simple’, ‘single’, or ‘unforked’ are not adequate descriptors” of the sulcus condition (Myers, 2011: 23). The situation in the D. percarinatum complex iS subtly different from what Myers found in Leptodeira. In_ that genus, some species have a truly simple sulcus, but more commonly a small terminal fork is present in retracted hemipenes but lost during eversion, or the terminus presents divergent lips in everted organs. In the D. percar- inatum complex retracted hemipenes do not 334 Bulletin of the Museum of Comparative Zoology, Vol. 160, No. 6 exhibit a bifurcate sulcus, only an expanded sulcus tip, whereas everted organs of two species (D. prolixum and D. graciliverpa) can have an apparent terminal division to the sulcus. CONCLUDING REMARKS My review of two species complexes in Dendrophidion shows that speciation within this genus is frequently accompanied by very little differentiation in scutellation characters that typically distinguish snake species (Cadle, 2012; the present work; Table 1). Snake taxonomists rightfully use these characters because of their proven utility in studies of snake systematics, quantitative genetics, and geogr aphic varla- tion, not to mention the ease with which they are scored. However, Dendrophidion offers cautionary examples that closely related snake species may not be distin- guishable by the usual scale characters. We currently have no measures of how wide- spread or representative this situation might be among other snakes (a few other examples are well known, e.g., some North American species of Thamnophis). Subtle differences in coloration and rather more profound differences in male genital morphology accompanied speciation within the D. vinitor complex (Cadle, 2012), such that distinguishing these species on the basis of external characteristics is problem- atic (fortunately for nontaxonomists, distri- butions of those three species are mutually exclusive). A contrasting situation is pre- sented by the D. percarinatum complex (this work). Although D. percarinatum is distinguished from the other two species in both color pattern characters and hemipe- nial morphology, D. prolixum and_ D. graciliverpa show differences from one another only in coloration. Their hemipenes appear identical when intraspecific variation in hemipenial characters is taken into account, and the two species are not easily distinguished on the basis of scale charac- ters. onthe scenario of mosaic character evolution or disparate evolutionary rates among characters is probably more common among snakes than is presently recognized. Close attention to subtle morphological differences among populations offers the best clues to species recognition in these cases. Many putatively widespread snake species may represent complexes of species distinguishable in only subtle ways. The existence of cryptic species in snakes such as Dendrophidion has more than taxonomic interest because it bears on one of the most pressing biodiversity issues of our time: potential declines or extirpations of these snakes in much of their range. Although amphibian biologists have become well attuned to declines of Neotr opical frogs and salamanders, reptile biologists have a potentially more difficult job of recognizing declines because absence can be more difficult to detect, especially for snakes (Pounds, 2000: 158-159). Nonetheless, be- cause many tropical forest snakes (including all species of Dendrophidion) prey upon anurans, their populations are surely affect- ed by amphibian declines and ‘general environmental degradation. Cadle (2012: 208-209, 216-217) pointed out that popu- lations of two species of the D. vinitor complex are probably affected to some degree by alekea amphibian declines and climate-mediated changes in rainforest leaf litter cover (Lips et al., 2006, Whitfield et al., 2007). But there are. almostemo quantitative or systematic measures of these effects. Dendrophidion crybelum, known only from a very localized area in southwestern Costa Rica, has not been seen since 1987 despite resurvey of its type locality (Santos- Barrera et al., 2008: Cadle, 2012: 217). At Monteverde, Costa Rica, populations of 11 snake species, including at least one species of Dendrophidion (D. paucicarinatum), have declined since 1987, and two Species are possibly locally extinct (Pounds, 2000). Similarly, D. prolixum, described herein, possibly has not been seen since the type series was collected in 1973 (admittedly, few collections have been made from within its known range in the intervening years). The SYSTEMATICS OF DENDROPHIDION PERCARINATUM (COLUBRIDAE) problem of decline and possible extinction is especially acute when unre ‘cognized cryptic species with restricted ranges exist within what is thought to be a single widespread species, such as the Dendrophidion species covered here and in Cadle (2012). Loss of these species, in some cases before their formal descriptions, is cause for heightened awareness among tropical biologists of the potential collapse of populations of snake predators whose biology is intimately tied to declining prey species. Such ramifying ecosystem aed ts call for redoubled efforts at biodiversity documentation (including sys- tematic studies to reveal cryptic biodiversity) and development of effective methods of ecosystem preservation and restoration. ACKNOWLEDGMENTS Many collection personnel offered indis- pensable assistance and loans during this study: Darrel Frost, David Kizirian, and Charles W. Myers (AMNH); Edward B. Daeschler and Edward (Ned) Gilmore (ANSP); Patrick Campbell, Tracy Heath, and Colin McCarthy (BMNH); Lauren Scheinberg and Jens Vindum (CAS); Ste- phen P. Rogers (CM): Maureen Kearney, Alan Resetar, Sarah Rieboldt, and Harold K. Voris (FMNH); Rafe Brown, Andrew Campbell, and Linda Trueb (KU); Neftali Camacho, Jeff Seigel, and Christine Thacker (LACM); Christopher Austin, Jeff Boundy, and Alison Hamilton (LSUMZ); James Hanken, Jonathan Losos, and José P. Rosado (MCZ); Christopher Conroy, Mi- chelle Koo, Jimmy McGuire, Carol Spencer, and David B. Wake (MVZ); Toby Hibbitts (TCWC); Michael Granatosky and Kenneth L. Krysko (UF); Christopher Phillips and Dan Wylie (UIMNH); Ronald Nussbaum and Greg E. Schneider (UMMZ); Steve Gotte, Roy W. McDiarmid, James Poindex- ter, Robert Wilson, and Céorge fh. Zuc (USNM); and Jonathan Campbell and Carl J. Franklin (UTACV). Charles W. Myers was particularly helpful and generous in sharing his photographs and field notes on Dendrophidion; his fine series of D. pro- © Cadle 335 lixum helped in understanding ontogenetic patterns in that species. Other special assistance was provided by Colin — J. McCarthy for preliminary data and photo- graphs of the holotype of D. brunneum: José P. Rosado for a scan of a color slide by Kenneth Miyata; Roy W. McDiarmid for his photograph of D. percarinatum; Patrick Campbell for photographs of BMNH spec- imens; Fernando Rojas-Runjaic for informa- tion on specimens in the Museo de Historia Natural La Salle: and Neil Duncan and David Kine for delving into AMNH archives to help resolve certain Ecuadorian localities. Thanks to Roy W. McDiarmid, Charles W. Myers, and the MCZ Herpetol- ogy Department for permission to reproduce dhcix color photogr aphs. and to Darrel Frost for generous permissions to prepare hemi- penes from AMNH specimens. For com- ments and discussions that greatly improved the manuscript, I am deeply indebted especially to Jay M. Savage, Jonathan Losos, and two anonymous reviewers. APPENDIX 1. SPECIMENS EXAMINED AND LITERATURE RECORDS OF DENDROPHIDION PERCARINATUM AND NEW RECORDS OF D. BRUNNEUM FROM ECUADOR Museum abbreviations used throughout are the following: AMNH—American Museum of Natural History (New York). ANSP—Academy of Natural Sciences of Philadelphia. BMNH—The Natural History Museum (London). CM—Carne- gie Museum (Pittsburgh). KU—University of Kansas Museum of Natural History (Lawrence). LACM—Natural History Museum of Los Angeles County (Los Angeles). MCZ—Museum of Com: parative Zoology (Cambridge). MHNLS—Museo de Historia Natural La Salle (C (Caracas) “TCW C= Texas Cooperative Wildlife Collection, Texas A and M University (College Station). UF—Florida Museum of Natural History, University of Florida (Gainesville). UIMNH— University of Illinois Museum of Natural History (Urbana). UMMZ— University of Michigan Museum of Zoology (Ann Arbor). USNM—National Museum. of Néeucil History, Smithsonian Institution (Washington, DC). UTACV—University Texas at Arlington Collection of Vertebrates ( (Arlington). 336 Bulletin of the Museum of Comparative Zoology, Vol. 160, No. 6 Bracketed data associated with localities here and elsewhere in the text are inferences derived from sources other than original data associated with specimens as ncoodcdk in literature, muse- um or collectors’ catalogues, or specimen labels. Countries are listed no to south Honduras to Colombia, followed by Venezuela. Dendrophidion percarinatum Honduras: Atlantida: CURLA Forestry Sta- tion, 120-500 m [15°42’N, 86°51’W fide USNM database], USNM 559613-14. Lancetilla [15°42’N, 87°26’W], MCZ R-29677. Colon: Los Andes [15°50’N, 85°08’W], ANSP 20817. Gracias a Dios: Bodega de Rio Tapalwas, 190 m [14°56’N, 84°32’W], USNM 561031, 563301. Hiltara Kiamp, 150 m [on the upper Rio Warunta; 14°57'N, 84°40’W], USNM 563300, 565532. Between Hiltara Kiamp and Sachin Tingni Kiamp, 150 m [14°57'N, 84°40’], USNM 563489. Kipla Tingni Kiamp, 160 m [on the upper Rio Warunta; 14°56'N, 84°40’W], USNM 565533. Oscana, 60 m [14°42’N, 84°27'W], USNM 562872-73. Rus Rus, 60 m [14°43’N, 84°27'W], USNM 559612, 561918. Sachin Tingni, 150 m [tributary of upper Rio Warunta: 14°57'N, 84°40’W], USNM 563302, 563488. San San Hil Kiamp, 190 m [hill E of Rio Tapalwas; 14°57'N, 84°31'W], USNM 563490, 564078. Olancho: Los Chorritos near Campamento, 950 m [14°33’N, 86°39’W; McCranie (2011) gives 685 m as the elevation], USNM 337504. Confluence of Rio W ampu and Rio Yanguay, 110 m [15°03’N, 85°08’W], USNM 391734, Quebrada E] Guasimo, 140 m [tributary of Rio Patuca; 14°35’N, 85°18’W], USNM 559611. Warunta Tingni Kiamp, 150 m [14°55’N, 84°41'W: McCranie, 2011: 639], USNM 561919. Nicaragua: [Atldntico Norte]: Bonanza, 850 ft. [260 m; 14°02’N, 84°35’W], KU 86183. Atlantico Norte: Musawas, Huaspuc_ River [14°09’N, 84°42’W], AMNH R-75428. Atldntico Sur: El Recreo, 10 mi. W Rama [12°10’N, 84°19’W], LACM 74148. Seven miles above Rama, Rio Siquia [approximately 12°10'N, 84°18’W], UMMZ 79767. 10 mi. above Recero [= iRecreo]), Rio Siquia) Mico: |12°074N; 84°26'W; see Cadle, 2012: 234], UMMZ 79764. Boaco: Comoapa [= Camoapa; 12°23'N, 85°31’W, about 520 m], MCZ R-9550. Grom tales: nto Domingo, Chontales Mines, 2,000 ft. [610 m; 12°16’N, 85°05’W], BMNH 94.10.1.18 (specimen not seen; Stafford, 2003). Costa Rica: No additional data, AMNH RB- 17374, USNM 259130. Alajuela: Venado, 9 km N Arenal, 252 m [10°33’N, 84°45’W], LACM 148580. Cartago: Pavones de _ Turrialba [(09°54'N, 83°37'W], UTACV 12898. Turrialba, 605 m [09°54’N, 83°41’W], LACM 148577, 148579. Guanacaste: Tilaran, 1,300 ft. [400 m; 10°28’N, 84°58’W], USNM 70663. Heredia: Finca La Selva, 2.4 km SE Puerto Viejo [35- 137 m: 10°26'N, 83°59’W: various locations within the La Selva Biological Station], KU 305559: LACM 148558, 148560, 148583—86. Limon: Barra del Colorado, 4 m [10°42’N, 83°36’W], LACM 148587. Batan [10°05’N, §3°20'W], KU 30998. Vicinity of Cahuita, about 4 m [09°44'N, 82°50’W], LACM 148582. Los Diamantes [about 300 m; 10°12’N, 83°47’'W experimental station about 1 km E Guapiles], KU 30997. La Lola, 39 m [10°06’N, 83°23'’W], LACM 148578. Pandora, 17 m [09°44’N, 82°58’W], LACM 148581. 2.3 km E Siquirres, 280 m [10°06’N, 83°31’W], UMMZ 137389. Puntarenas: Boruca [09°00'N, 83°19’W], AMNH B-17366 (lectotype). Parque Nacional Carara, 1.9 rd mi. S Rio Tarcoles on Hwy 34 [09°46’N, 84°36’W], TCWC 84024, 84083. Palmar [a Taylor locality = Palmar Sur fide Jay M. Savage, personal communication], KU 31948. 15 km E Palmar Norte, N Lagarto at Quebrada Yan, 70 m [08°57’'N, 83°23’W], LACM 148574. 2 km S entrada! Palmar Sur, 15 m [8°58’N, 83°27'W], LACM 148592. Vicinity of Rincon de Osa [08°42'N, 83°29’W: various localities within 7.5 km W to SW of the settlement, 5-60 m elevation; see McDiarmid and Savage, 2005], ANSP 27900; KU 102505-06; LACM 114100- Ol, ~148561—148563, 148565, 148570=7a: 148575_-76: LSUMZ 3411213. Fines East@ra- ces, about 6 km S San Vito de Java [approxi- mately 1,200 m; 08°47'35’N, 82°57'30"W], LACM 114102-04, 148566. Golfite? 12m [(08°36’N, 83°09’W], LACM 148568. 0.4 km W of Mctel Bella Vista, Golfito, 15 m [08°36'N, 83°09’W], LACM 148564. 6.3 km S of Pan American Hwy on Golfito Rd., 7 m [08°37’N, §3°04'W], LACM 148567. Gromaco, at juncture of Rio Cotén and Rio Coto Brus, 480 m [(08°55'N, 93°06’W], LACM 148569. Vicinity of Rio Disciplina, 80 m [08°58’N, 83°20’W], LACM 148588. San Luis River at footbridge, about 740 m [10°16’N, 84°49’W], LACM 148559. San José: 1.4 mi. N and 0.6 mi. NNE (by road) Bijagual; Montafias Jamaica, Parque Nacional Carara [09°43’N, 84°34'W], TCWC SYSTEMATICS OF DENDROPHIDION PERCARINATUM (COLUBRIDAE) ¢ Cadle Bal 83370. “Los Cusingos” [09°43'N, 84°34"W; in Quizarra on the lower slopes of Volcan Chirrip6 near Santa Elena], Kohler (2008: fig. 582). Panama: No additional data, FMNH 31214— 16 (heads only). “Canal Zone,” no additional data, FMNH 6118. Bocas del Toro: Almirante, 10 m [09°18’N, 82°24’W], KU 80224, UMMZ 142638. 1.5 mi. W Almirante, Nigua Creek, 10 m [09°17’N, 82°24’W], KU 107644. 4 km W Almirante, 10 m [09°17'N, 82°24’W], KU 107645. Hill above Miramar, 180 m [08°59'N, 82°15'W], KU107647. Torres, western Panama [09°25'N, 82°31'W], MCZ R-19343. [Chiriqut]: “Chiriqui,” BMNH 94.5.17.8-9 (specimens not seen: Stafford, 2003). Coclé: El Valle de Anton, 2,000 ft. [610 m: 08°37’N, 80°08’W], AMNH R-71681. Coldn: Canal Zone, Acliote [= Achiote] Rd., 5.1 km NW Escobal Rd., N of road [approximately 09°15’N, 80°02’W], UMMZ 155740. Agua Clara, Chagres River [09°11'N, 79°41’W], ANSP 25144. Near Buena Vista on Trans-Isthmian Highway, 200 ft. [61 m; 09°16’N, 79°41’W], FMNH 154474. Canal Zone, Buena Vista Penin- sula, 1.75 km NNW of Frijoles [09°10'34’N, 79°48'W], USNM 196306. Canal Zone, Camp Chaeres, 120 ft [40 m; 09°21’N, 79°57'W], KU 75676. Canal Zone, Fort Randolph [09°23’N, 79°53'W], MCZ R-20552. Canal Zone, Salamanca Hydrographic Station, Rio Pequeni [09°20'N, 79°36'W], MCZ R-39978. Gamboa [09°07'N, 79°42'W], FMNH 154510, USNM 297811. Gattin [09°16’N, 79°55’W], FMNH 16760, USNM 54080. Canal Zone, Atlantic side, Gattn, Fort Davis} [09°17 N, 79°54’W], MICZ R-22255. Carti Rd. [not located], USNM 266157. Darién: Rio Tuira, Boca de Cupe, 30 m [08°03'N, 77°35’W], AMNH R-119376. Cana, 2,000 ft. [610 m; O74 N, 77-42) W KU! 107651, USNM 50123. [Cerro] Tacarcuna, 550 m [08°10'N, 77°18’W], KU fo67 79, Origa. site, 8°46’ 78°00 (30 m), FMNH_ 170152. Along Rio Canclén [= Rio Canglon] near mouth of Rio Chucunaque [ap- proximately 08°20'N, 77°46’W], UMMZ 124063. Along Rio Canclon [= Rio Canglén] near crossing of Inter-American Highway [approximately 08°20'20"N, 77°49'50’W], UMMZ 124064. Near mouth of Rio Canclon |= Rio Canglon; 08°19'N, 77°46'W], UMMZ 124199. Rio Tuira at Rio Mono, 130 m, 107-43’ N, 77°33’ W], KU 107652—- 0, Yarrssa. |= Yavizal [O8°1l’N, 77°41'Wi, UMMZ 83144. [Herrera]: Cerro Mangillo, 2,800 ft., Veragua [854 m; = Cerro Manglillo, 07°34'N, 80°47'W; the Manglillo Massif straddles the border between Veraguas and Herrera province; see Dunn (1943) for a brief recountin of his route], ANSP 22446. Los Santos: E slopes a Cerro Hoya, 930 m [07°19’N, 80°40’W], KU 107659. Panamd: Canal Zone, Alhajuela [O9°1L1’N, 79°33’W], UMMZ 76019. Altos de Majé [(08°48’N, 78°31’W], AMNH_ R-109643. Barro Colorado Island: AMNH R-77573, 8997 1-— 72; ANSP 22560, 22878: CM S 6869, S 7711; KU 75674—75, 80589-90: MCZ R-18902; UMMZ 63762; 124061-62, 124065-66; USNM 120815. Canal Zone, Chiva Chiva [09°02’N, 79°35'W], MCZ R-24002. Cocoli [08°59’N, 79°35'50’"W], USNM_ 193447. Canal Zone, Contractor's Hill [09°02'’N, 79°39'W], CAS 98388. Canal Zone, Curunditi [08°59’N, 79°33’W], KU 80255. Canal Zone, Fort Clayton [09°00'N, 79°34’W], KU 107649, 110290, MCZ R-25124. Canal Zone near Fort Clayton, [09°00'N, 79°34'W], UIMNH 41705-19. Fort Clayton, Cardenas River [09°00’N, 79°34’W], KU 110291. Gamboa or Pedro Miguel [approximately 09°03'N, 79°40'W], FMNH 154516. Canal Zone, Juan Mine [09°10’N, 79°39'W], MCZ R-26646. Canal Zone, Madden Forest, Rio Pedro Miguel [09°06'N, 79°37'W], KU 107650. Canal Zone, [Madden| Forest Preserve [09°06'N, 79°37'W], AMNH R- 89970. Canal Zone, Red Tank [09°00'N, 79°36'W], MCZ R-24000. Cerro Campana, 2,500 ft. [762 m; 08°41'N, 79°56’W], AMNH R- 76000. Cerro Campana, 800 m [08°41'N, 79°56'W], AMNH_ R-129757. Cerro Jefe [(09°14’N, 79°21’W], UMMZ 155732. San Blas: Armila, Quebrada Venado [08°40'N, 77°28'W], USNM 150139. San Ignacio de Tupile, mainland, 2:5’ mu. inland, ca. 250 tt. [75 m; 09°15'N, 78°09'W], USNM 241656. Veraguas: Isla Gober- nadora [07°33'N, 81°12’W], KU 107648. Colombia: Antioquia: Medellin [06°15'N, 75°35'W], BMNH_ 1897.11.12.10 (questionable record; see Distribution in the D. percarinatum account). Uraba, Rio Currulao, 50 m [OS°O1'N, 76°44’W], FMNH_ 63772-73. Uraba, Turbo [(08°06’N, 76°44’W], FMNH 63761. Villa Arteaga [135 m; 07°22’N, 76°29'W], FMNH 78118, USNM 267273. Choc6é: Quebrada Pangala, lower Rio San Juan (about 17 km airline NE Palestina), 04°15’N, 77°0'W, AMNH R-123745, R-123748. Vicinity of Playa de Oro, upper Rio San Juan, ca. 200 m [05°19'N, 76°24’W], AMNH _ R-108468. Sierra [Serranta] del Darién, Chocé, 600 ft., Pacific side [183 m], ANSP 25606. Valle del Cauca: Rio Raposo, Virology Field Station near Buenaventura [(03°43’N, 77°08’W], USNM 151658. Venezuela: Zulia: Sierra de Perija, Finca El Progreso, 840 m [10°43'13.30"N, 72°29'16.60"W], MHNLS 17932 (specimen not seen; Rojas-Runjaic and Rivero, 2008). 338 Dendrophidion brunneum (Ecuador only; see Cadle, 2010: 24 for other records) Ecuador: No other data, AMNH_ R-18324. “Western Ecuador,’ BMNH_ 1860.6.16.58, 1860.6.16.60, 1860.6.16.67.4 Chimborazo: Huigra to Rio Chiguancay [02°13’S, 79°03’W], ANSP 18122. [El Oro]: Portovelo [610 m; 03°42’48”S, 79°36'51"W; Lynch and Duellman, 1997: 215], AMNH 18322. Zaruma [03°41’S, 79°37'W], BMNH_ 1894.5.29.1. Guayas: 5-10 km S Daule [about 01°55’S, 80°00’W], UF 87940, 88466-67. 18 km FE, Duran- [0-50 m; about 02°14’S: 79°38'’W], UF 88468. El Milagro [13 m; 02°07'S, 79°36’W], USNM 237061-62. Guaya- quil [02°10’S, 79°50’W], BMNH 1946.1.12.98 (holotype); MCZ R-8393; USNM 237059. Near Guayaquil, USNM 237060. 3 km E Olan [POlon], Crespo Hacienda, 450 ft. [137 m; Olon is on the coast at 01°48'20"S, 8045'28”W; now in the newly created province of Santa Elena], UF 66199. Imbabura: Ibarra [about 2,200 m; 00°21'N, 78°07'W], USNM 237083. Intag [about 1,200 m; 00°20'N, 78°32’W], UMMZ 83706. Loja: La Argelia, Malacatos, 2,100 m [04°14'S, WO Ti NMS WSNMe = 23707 7-38 Loja: Loja 25200" “am: © 04°00'S. 79213 Wile “BMINE 1930.10,12.20=22* 1931. VL3i0! Masaca (=: Ha= cienda Masaca, near Loja) [03°53’S, 79°14’W], USNM 237082. [Loja]: Alamor [1,325 m; 04°02'S, 80°02’W], AMNH 22232. Los Rios: Finca Playa Grande [53 m; 01°01’30"S, 79°27'39"W, 1.6 km N Quevedo; Lynch and Duellman, 1997: 216], UIMNH 77347. 20 mi. NE Quevedo [about 00°53’S, 79°17'W], UF 85105. Dendrophidion species inquirendum Colombia: Boyacd: Muzo {1,242 m; 05°32'N, 74°06'W], MCZ R-42186. Santander: El Centro, “Four specimens were associated with the five BMNH catalog numbers 1860.6.16.58—60, 67-68 when specimens were received on loan. The specimens were not individually tagged with unique catalog numbers but each specimen had previously been individually identified by linen thread tied around the neck: 1, 2, or 3 threads + the smallest of the four with no thread. For purposes of reference and data collection I assigned the three with threads (1, 2, and 3, respectively) to numbers 1860.6.16.58-60, the specimen without thread to 1860.6.16.67 (—68 being therefore unused). Three of the specimens are Dendrophidion brunneum (1860.6.16.58, —60, —67). BMNH 1860.6.16.59 is a specimen of D. graciliverpa (see footnote 2). Bulletin of the Museum of Comparative Zoology, Vol. 160, No. 6 150 m [06°55'N, 73°44°W]|, USNM > 26724 Landazuri, 900 m [06°13’N, 73°49’W], USNM DA PAP: APPENDIX 2. GAZETTEER (DENDROPHIDION PROLIXUM AND D. GRACILIVERPA LOCALITIES) Dendrophidion prolixum (Colombia except as noted [Ecuador]) Cachavi, 20 m (Esmeraldas | Ecuador]). About 00°58’N, 78°48’W. Also known as Cachabé and San Javier de Cachabé (Paynter, 1993; Duellman and Lynch, 1'997: 217). Lita, Rio Mira (Imbabura [Ecuador]). 00°50'24"N, 78°27'18”"W; Duellman and Lynch, L992 2135 Paramba, northwestern Ecuador (Imbabura). 1,067 m. 00°49'N, = Hacienda Paramba. 18 21 Ws Pena Lisa, Condoto, 300 ft. (Choc6) 90 m; 05°04'N, 76°38'W. Playa de Oro, Rio San Juan, 400 m (Choco) 05°19'N, 076°24'’W. [Between] Pueblo Rico [and] Santa Cecilia, Pacific side, SOO m (Risaralda). Santa Cecilia: 5°20’N, 76°08'W Pueblo Ricos ale Ne 76°02’ W. Quebrada Bochorama, Loma de Encarnacion on right bank (Choc6) about 51 km SE of Quibdo at approximately 5°20’N, 76°23’W, about 400 m elevation: Brame and Wake, 1972: Tt Quebrada Docord6, middle Rio San Juan about 17 km airline SSW Noanama (Choc6). 04°33’N, 77°0'W (coordinates from AMNH database). Quebrada Guangui, 0.5 km above Rio Patia (upper Saija drainage), 100-200 my o@anea department, Colombia [about 02°50’N, 77°25'W; Myers, 1991: 8]. Quebrada Pangala, lower Rio San Juan about 17 km airline NE Palestina (Choc6). 04°15’N, 77°00'W (coordinates from AMNH database). Quebrada Taparal, lower Rio San Juan about 7 km airline NE Palestina (Choco). 4°12'N, 77°07'W (coordinates from AMNH database). Quininde (Esmeraldas). 00°18'50"N, 79°27'40"W About 100 m (Duellman and Lynch, 1997: 216). SYSTEMATICS OF DENDROPHIDION PERCARINATUM (COLUBRIDAE) Upper Rio Buey, 110-160 m (Choco). proximately 06°06’ N, 77°05'W. (Esmeraldas Ap- Rio Cachavi [Ecuador]). 01°03'N, cannes Rio Raposo, Virology Field Station near Bue- naventura (Valle del Cauca). 03°43’00"N, 77°08'00"W. Riquarte [= Ricaurte], 3,900 ft. [1,189 ml, Pacific side (Narino). 01°13'N, 77°59'W. Sierra [Serranfa] de Baud6, 3,000 ft. [915 mJ, Pacific side (Choc6). Approximately 06°00'N, 077°05'W. Serrania de Baud6, N slope of Alto del Buey, 900 m (Choc6). 06°06'N, 077°13'W. Dendrophidion graciliverpa (Ecuador) Alamor (Loja). 04°02'S, 80°02'W. Bilsa Biological Reserve (Esmeraldas). 00°15'N, 79°45'W. Ortega-Andrade et al. (2010: 1) give coordinates: 00°21'33"N, 79°42'02”W: 300-750 m elevation. Buena Fe, 1 km N of (Los Rios). 02°16’S, 79°37'W Canoas near Santo Domingo De Los Colora- dos. (Santo Domingo de los Tsachilas/Pichincha) Not located. ; Centro Cientifico Rio Palenque (Los Rios). 00°35'11"S, 79°22'W, 220 m elevation. Along the road heoveen Santo Domingo de los Colomiles and Quevedo. Dodson and Gentry (1978) describe the environment and geography of the area. Chaguarapata, 2,000 ft. (Chimborazo). Ap- proximately 02°07’ S, 78°59'W. This locality goes by several spellings in the literature (mostly ornithological): Chaguarpata (Paynter, 1993), Charguarpata (Chapman, 1926), Chahuarpata (various). The elevations given are not consistent among sources. AMNH catalogue data for AMNH R-23032 give 2,000 ft., which may be the actual elevation the specimen was collected by G. H. H. Tate. Chapman (1926: 705) gives the elevation of Charguarpata as 5,800 ft., whereas Tate’s typed itinerary in the AMNH mammal department gives 2,300 ft. for the elevation. Chapman (1926: 706) states that Chaguarpata is “in the forest above. Cayandeled,” which is a hacienda north of Bucay in the Rio Chimbo basin (see Chapman, 1926: map, pl. 30). Finca La Esperanza near Santo Domingo de los Colorados. (Santo Domingo de los Tsachilas/ Pichincha). 00°15’S, 79°09'W, 500 m; a farm just NW of Santo Domingo de los Colorados ( Lynch and Duellman, 1997: 212). © Cadle 339 Finca Playa Grande (Los Rios). 53 m; 01°01'30"S, 79°27'39"W;: near Quevedo; Lynch and Duellman, 1997: 216. Guayaquil (Guayas). 02°10'S, 79°54'W. Hualtaco (El Oro). 03°26'S, 80°15’W. Joe Ramsey Farm, km 19 on Chone Road, 18 km W of Santo Domingo de los Colorados (Santo Domingo de los Tsachilas/Pichincha). 00°14'S, 79°20’W (USNM electronic database). Las Pampas (Cotopaxi). 1,750 m,; 00°40'S, 78°50'W. Also referred to as San Francisco de las Pampas (Lynch and Duellman, 1997). Meme, km 96 on road to Saloya at crossing of Rio Toachi (Pichincha). 00°06'S, 79°08'W. Mulaute, on tributary of Rio Blanco. (Pi- chincha). 00°05'S, 79°09’W (USNM electronic database). Playas De Montalvo 15 m (Los Rios). 01°48'S, 79°20'W (Paynter, 1993). Also referred to as “Playas” (Chapman, 1926: 732; Brown, 1941: 836). Puerto Quito (Pichincha). 00°07'N, 79°16'W. Quininde (Esmeraldas). 00°18'50"N, 79°27'40"W, 40 m elevation (Lynch and Duell- man, 1997: 216). Also known as Rosa Zarate. Rancho Santa Teresita, km 25 on route to Chone from Santo Domingo de Los Colorados (Santo Domingo de_ los Tsachilas/Pichincha). 00°15’S, 79°23'W (USNM electronic database). Rio Baba (Santo Domingo de los Tsachilas/ Pichincha). Approximately 00°25'S, 79°17'W, Lynch and Duellman, 1997: 216. Humid tropi- cal. Southward flowing river on Pacific coastal plain; riparian rainforest along river. Rio Congo, headwaters ar USNM database places this locality in Guayas province, which is the location of the main part of the Rio Congo. The “headwaters,” depending on interpretation, are potentially farther north (Manabi or Los Rios provinces). Rio Palenque Science Center. Cientifico Rio Palenque. Rio Pescado (Guayas). About 02°41'S, 79°32, W.. This is a collecting site of G. H. H. Tate, a mammalogist who participated in the AMNH ornithological expeditions in the early 1920s (Chapman, 1926). Its location seems to be of some confusion, as various published papers on insects, mammals, and frogs that Tate collected there place the locality in at least four different Ecuadorian provinces (Manabi, Guayas, Azuay, Chimborazo). I am_ grateful to Neil Duncan of the AMNH mammalogy department See Centro 340 for checking Tate’s field notes and a typed summary of the trip in the department archives. He provided the details given here. Tate worked at “Rio Pescado” from May 14 to June 3, 1922 (e.g, AMNH_ R-23438, a specimen of Dendro- phidion graciliverpa, was collected 19 May 1922). The camp on the Rio Pescado was three hours by trail east from Naranjal (02°40'22”S, 79°36'54”"W) into the foothills at 1,600 ft [488 m]. The camp was located about 0.5 mi. above the junction of the Rio Pescado and the Rio Chacayacu and near the Guayas-Azuay provincial border. Below Rio Toachi (Santo Domingo de los Tsachilas/Pichincha). Lat/Long from USNM database: 00°11'S, 79°11’W (USNM electronic database). Lynch and Duellman (1997: 217) for Rio Toachi: 00°23'S, 78°56’W, 800 m elevation. Rosa. Delia Plantahon (El Oro). 03 mS 79°55'W. Santo Domingo de los Colorados, 550-660 m (Santo Domingo de los Tsachilas/Pichincha). OOSLSS 279-10" W. LITERATURE CITED ALEMAN, G. C. 1953. Contribucién al estudio de los reptiles y batracios de la Sierra de Perija. Memorias de la Sociedad de Ciencias Naturales de La Salle 13205-2225; ALMENDARIZ, A. 1991. Anfibios y reptiles. Politécnica 16: 89-162. Amaral, A. po. “1929” [1930]. Estudos sobre ophidios neotropicos. XVHI—Lista remissiva dos ophidios da Regiao Neotropica. Memorias do Instituto Butantan 4: 127-128 + i-viii + 129-271. Autu, D. L. 1994. Checklist and bibliography of the amphibians and reptiles of Panama. Smithsonian Herpetological Information Service 98: 1-59. Bo.aNos, F., J. M. Savace, AND G. CuaAves. 2010. Anfibios y reptiles de Costa Rica, pp. 1-19. In O. Breedy, R. Vargas, and F. Bolanos (eds.), Listas Zoologicas Actualizadas UCR. San Pedro: Museo de Zoologia, Univ. de Costa Rica. [modified 18 August 2010; cited 1 January 2012]. Available from: http://museo. biologia.ucr.ac.cr. BouLENGER, G. A. 1882. Account of the reptiles and batrachians collected by Mr. Edward Whymper in Ecuador in 1879-1880. Annals and Magazine of Natural History 9: 457-467. BouLENGER, G. A. 1891. Reptilia & Batrachia, pp. 128- 136. In E. Whymper (ed.), Supplementary Appen- dix to Travels amongst the Great Andes of the Equator. London: John Murray. BouLencer, G. A. 1894. Catalogue of the Snakes in the British Museum (Natural History). Vol. 2. London: British Museum (Natural History). BouLeNGcER, G. A. 1913. A collection of Batrachians and reptiles made by Dr. H.G.F. Spurrell, F.Z.S., in the Bulletin of the Museum of Comparative Zoology, Vol. 160, No. 6 Choco, Colombia. Proceedings of the Zoological Society of London 68: 1019-1038. BraME, A. H., JR., AND D. B. WaKE. 1972. New species of salamanders (genus Bolitoglossa) from Colom- bia, Ecuador, and Panama. Contributions in Science, Natural History Museum of Los Angeles County 219: 1-34. Bropig, Epmunpb, D. III. 1990. Genetics of the garter’s getaway. Natural History 1990: 45-51. BropizE, Epmunp, D. III. 1992. Correlational selection for color pattern and antipredator behavior in the garter snake Thamnophis ordinoides. Evolution 46: 1284-1298. Brown, F. M. 1941. A gazetteer of entomological stations in Ecuador. Annals of the Entomological Society of America 34: 809-851. Cape, J. E. 1989. A new species of Coniophanes (Serpentes: Colubridae) from northwestern Peru. Herpetologica 45: 411-424. Cab.e, J. E. 1996. Systematics of snakes of the genus Geodipsas (Colubridae) from Madagascar, with descriptions of new species and observations on natural history. Bulletin of the Museum of Com- parative Zoology 155: 33-87. Cape, J. E. 2005. Systematics of snakes of the Dipsas oreas complex (Colubridae: Dipsadinae) in western Ecuador and Peru, with revalidation of D. elegans (Boulenger) and D. ellipsifera (Boulenger). Bulle- tin of the Museum of Comparative Zoology 158: 67-136. Cape, J. E. 2007. The snake genus Sibynomorphus (Colubridae: Dipsadinae: Dipsadini) in Peru and Ecuador, with comments on the systematics of Dipsadini. Bulletin of the Museum of Comparative Zoology 158: 183-284. Cap.e, J. E. 2010. Systematics, natural history, and hemipenial morphology of Dendrophidion brun- neum (Giinther), a poorly known snake from the Andes of Ecuador and Peru. Zootaxa 2433: 1—24. Cape, J. E. 2012. Cryptic species within the Den- drophidion vinitor complex in Middle America (Serpentes: Colubridae). Bulletin of the Museum of Comparative Zoology 160: 183-240. Cuan, L. M., J. L. Brown, anp A. D. Yoper. 2011. Integrating statistical genetic and geospatial meth- ods brings new power to phylogeography. Molec- ular Phylogenetics and Evolution 59: 523-537. CuapMaNn, F. M. 1926. The distribution of bird-life in Ecuador; a contribution to a study of the origin of Andean bird-life. Bulletin of the American Museum of Natural History 55. Ciark, H. 1944. The anatomy and embryology of the hemipenis of Lampropeltis, Diadophis, and Tham- nophis and their value as criteria of relationship in the family Colubridae. Proceedings of the Iowa Academy of Sciences 51: 411-445. CoLe, C. J. anp L. M. Harpy. 1981. Systematics of North American colubrid snakes related to Tantilla planiceps (Blainville). Bulletin of the American Museum of Natural History 171: 199-284. SYSTEMATICS OF DENDROPHIDION PERCARINATUM (COLUBRIDAE) Corr, E. D. 1863. Descriptions of new American Squamata in the museum of the Smithsonian Institution, Washington. Proceedings of the Acad- emy of Natural Sciences of Philadel lphia 15: 100- 106. Corr, E. D. “1893” [1894]. Second addition to the knowledge of the Batrachia and Reptilia of Costa Rica. Proceedings of the American Philosophical Society 31: 333-347. Corr, E. D. “1894” [1895]. On a collection of Batrachia and Reptilia from the island of Hainan. Proceed- ings of the Academy of Natural Sciences of Philade [phia 46: 423-428. Corr, E. D. 1895. The classification of the Ophidia. Transactions of the American Philosophical Society 18: 186-2 19 + pls. 14-33. Corer, E. D. 1900. The crocodilians, lizards, and snakes of North America. Report of the U.S. National Museum, under the Direction of the Smithsonian Institution 1898: 155-1294 Daza, J. M., T. A. CasTor, AND C. L. Parkinson. 2010. Using regional comparative phylogeographic data from snake lineages to infer historical processes in Middle America. Ecography 33: 343-354. Dopson, C. H., anp A. H. Gentry. 1978. Flora of the Rio Palenque Science Center, Los Rios Province, Ecuador. Selbyana 4: 1-628. DuELLMAN, W. E. 1978. The biology of an equatorial herpetofauna in Amazonian Bevador Miscella- neous Publications, Museum of Natural History, University of Kansas 65: 1-352. DuELLMAN, W. E. 2005. Cusco Amazonico: The Lives of Amphibians and Reptiles in an Amazonian Rain- forest. Ithaca, New York: Comell Univ. Press. Dunn, E. R. 1943. Zoological results of the Azuero Peninsula Panama Expedition of 1940. Part 1—A new species of Peripatus. Notulae Naturae 123: 1-3. Dunn, E. R. 1944. The snake genus Dendrophidion in Colombia. Caldasia 2: 474-477. Dunn, E. R., anp J. T. EMLEN, JR. 1932. Reptiles and amphibians from Honduras. Proceedings of the Academy of Natural Sciences of Philadel ‘Iphia 84: 21-32. FAIRCHILD, G. B., AND C. O. HANDLEY, JR. 1966. Gazetteer of collecting localities in Panama, pp. 9-20 + map. In R. L. Wenzel and V. J. Tipton (eds.), Ectoparasites of Panama. Chicago: Field Museum of Natural History. FREIRE, E. M. J., U. CaraMascui, AND U. GONGALVES. 2010. A new species of Dendrophidion (Serpentes: Colubridae) from the Atlantic rain northeastern Brazil. Zootaxa 2719: 62-68. Gaice, oH. Te N. Harrwec,: anp L. ‘C; Stuart, 1937. Notes on a collection of amphibians and reptiles from eastern Nicaragua. Occasional Papers of the Museum of Zoology, University of Michigan 357: 1-18. GoLpBERG, S. R. 2003. Reproduction in four species of Dendrophidion from Costa Rica (Serpentes: Colu- forest of ¢ Cadle 34] bridae). Transactions of the Illinois State Academy of Science 96: 295-300. Guntner, A. 1858. Catalogue of Colubrine Snakes in the Collection of the British Museum. London: British Museum (Natural History). Gunner, A. 1885-1902. Biologia Centrali-Americana, Reptilia and Batrachia |SSAR_ reprint 1987]. Ithaca, New York: Society for the Study of Amphibians and Reptiles. Guyer, C. 1994. The reptile fauna: diversity and ecology, pp. 210-216 + appendix 6. In McDade, L. A, K. S. Bawa, H. A. Hespenheide, and G. S. Hartshorn (eds.), La Selva: Ecology and Natural History of a Neotropical Rainforest. Chicago: The Univ. of Chicago Press. Guyer, C., AND M. A. DonneL_y. 1990. Length-mass relationships among an assemblage of tropical snakes in Costa Rica. Journal of Tropical Ecology 6: 65-76. Guyer, C., AND M. A. DONNELLY. 2005. Amphibians and Reptiles of La Selva, Costa Rica, and the Caribbean Slope: a Comprehensive Guide. Berkeley: Univ. of California Press. IBANEZ, D. R., F. A. AROSEMENA, F.. A. SOLIS, AND C. A. JaraMILLo, “1994” [1995]. Anfibios y reptiles de la Serrania Piedras-Pacora, Parque Nacional Chi agres. Scientia (Panama) 9: 17-31. IBANEZ, D. R., AND F. A. Sotis. “1991” [1993]. Las serpientes de Panama: lista de especies, comentar- ios taxonomicos y bibliografia. Scientia (Panama) 6: Di=o2. INTERNATIONAL COMMISSION ON ZOOLOGICAL NOMENCLA- TURE. 1999. International Code of Zoological Nomenclature. 4th ed. London: International rust for Zoological Nomenclature in association with the British Museum (Natural History). Jan, G. 1863. Elenco sistematico degli Ofidi descritti e disegnati per UIconographia Ge merale. Milan: A. Lémibar di. Jan, G., and F. Sorpeuur. 1860-1881. Iconographie générale des ophidiens. Vols. 1-3, livrs. 1-50. Milan: Jan and Sordelli. Keocu, J. S. 1999. Evolutionary implications of hemi- penial morphology in the terrestrial Australian elapid snakes. Zoological Journal of the Linnean Society 125: 239-278, Kouter, G. 2003. Reptiles of Central America (English edition). Offenbach, Germany: Herpeton. Konter, G. 2008. Reptiles of Central America (2nd English edition). Offenbach, Germany: Herpeton. Kopren, W. P. 1931. Grundrif der Klimakunde. Berlin: Walter de Gruyter und Co. LancInI, A. R. 1979. Serpientes de Venezuela. Caracas: Ernesto Armitano. Laurencio, D., AND J. H. MALone. 2009. The amphibians and reptiles of Parque Nacional Carara, a transitional herpetofaunal assemblage in Costa Rica. Herpeto- logical Conservation ane Biology 4: 120-131. LIEB, C. S. 1988. Systematic status of the Neotropical snakes De ndrophidion dendrophis and D. nuchalis (Colubridae). Herpetologica 44: 162-175. 342 Bulletin of the Museum of Comparative Zoology, Vol. 160, No. 6 Lies, C. S. 1996. Dendrophidion percarinatum. Cata- logue of American Amphibians and Reptiles 636: 636. 1—636.2. Lips, K. R., F. Brem, R. BRENES, J. D. REEVE, R. A. ALFORD, J. VOYLES, C. Carey, L. Livo, A. P. PEssIER, AND J. P. Co tins. 2006. Emerging infectious disease and the loss of biodiversity in a ‘Neotropical amphibian community. Proceedings of the National Academy of Sciences, USA 103: 3165-3170. Lyncu, J. D., AND W. E. DuELLMan. 1997. Frogs of the genus Eleutherodac tylus (Leptodactylidae) in west- ern Ecuador: systematics, ecology, and biogeogra- phy. University of Kansas Museum of Natural History, Special Publications 23: iv + 236 + 8 pls. McCranig, J. R. 2009. Amphibians and reptiles of Honduras [Internet], pp. 1-20. In O. Breedy, R Vargas, and F. Bolafios (eds.), Listas Zoologicas Actualizadas UCR. San Pedro: Museo de Zoologia, Univ. de Costa Rica | modified 12 November 2009]. Available from: http://museo.biologia.ucr.ac.cr. McCranig, J. R. 2011. Snakes of Honduras: Systemat- ics, Distribution, and Conservation. Ithaca, New York: Society for the Study of Amphibians and Reptiles. McCranlg, J. R., anD F. E. CasTaNepa. 2005. The herpetofauna of Parque Nacional Pico Bonito, Honduras. Phyllomedusa 4: 3-16. McCRANIE, RR. Ee ICASTANEDA. AND “Ko EP NICHOLSON. 2002. Preliminary results of herpeto- faunal survey work in the Rus Rus region, Honduras: a proposed biological reserve. Herpeto- logical Bulletin 81: 22-29. Wy telCrasts J. R., J. H. TowNsEnp, anp L. D. WILson. 2006. The Amphibians and Reptiles of the Hon- duran Mosquitia. Malabar, Florida: Krieger Pub- lishing Co. McDiarMip, R. W., and J. M. Savace. 2005. The herpetofauna of the Rincon area, Peninsula de Osa, Costa Rica, a Central American lowland evergreen forest site, pp. 366-427. In M. A. Donnelly, B. I. Crother, C. Guyer, M. H. Wake, and M. E. White (eds.), Ecology. and Evolution in the Tropics, a Herpetological Perspective. Chicago: Chicago Press. Myers, C. W. 1974. The systematics of Rhadinaca (Colubridae), a genus of New World snakes. Bulletin of the American Museum of Natural History 153: 1-262. Myers, C. W. 1982. Blunt-headed vine snakes (Im- antodes) in Panama, including a new species and other revisionary notes. American Museum Novi- tates 2738: 1-50. Myers, C. W. 1991. Distribution of the dendrobatid frog Colostethus chocoensis and description of a related species occurring macrosympatrically. American Museum Novitates 3010: 1—15. Myers, C. W. 2003. Rare snakes—five new species from eastern Panama: reviews of northern Atractus and southern Geophis (Colubridae: Dipsadinae ). Amer- ican Museum Novitates 3391: 1—47. Univ. of Myers, C. W. 2011. A new genus and new tribe for Enicognathus melanauchen Jan, 1863, a neglected South American snake (Colubridae: Xenodonti- nae), with taxonomic notes on some Dipsadinae. American Museum Novitates 3715: 1-33. Myers, C. W., AND J. E. Capie. 1994. A new genus for South erica snakes related to Rhadinaea obtusa Cope (Colubridae) and resurrection of Taeniophallus Cope for the “Rhadinaea” breviros- tris group. American Museum Novitates 3102: 133. Myers, C. W., anp J. E. Capie. 2003. On the snake hemipenis, with notes on Psomophis and tech- niques of eversion: a response to Dowling. Herpetological Review 34: 295-302. Myers, C. W., J. W. Daty, anp B. Matkin. 1978. A dangerously toxic new frog (Phyllobates) used by Embera Indians of western Colombia, with discus- sion of blowgun fabrication and dart poisoning. Bulletin of the American Museum of Natural History 161: 307-366. Myers, C. W., D. R. IBANEZ, AND J. E. CaDLE. 2007. On the uniquely fragmented distribution of a rare Panamanian snake, Dipsas nicholsi (Colubridae: Dipsadinae). American Museum Novitates 3554: Mic, NATIONAL GEOSPATIAL-INTELLIGENCE AGENCY (NGA) [INTERNET]. Springfield (VA): NGA GEOnet Names Server (GNS) (US); 2010-2012. Available from: http://earth-info.nga.mil/gns/htm/l/. NicEForRO Maria, H. 1942. as ofidios de Colombia. Revista de la Academia Colombiana de Ciencias Exactas, Fisicas y Naturales 5: 84-101. NICHOLSON, K.E., J. R. McCraniz, AND G. KOHLER. 2000. Herpetofaunal expedition to Parque Nacio- nal Patuca: a newly established park in Honduras. Herpetological Bulletin 72: 26-31. OrTEGA-ANDRADE, H. M., J. BERMINGHAM, C. AULESTIA, AND C. Paucar. 2010. Herpetofauna of the Bilsa Biological Station, province of Esmeraldas, Ecua- dor. Check List 6: 119-154. Parker, H. W. 1938. The vertical distribution of some reptiles and amphibians in southern Ecuador. Annals and Magazine of Natural History series AS64050) Paynter, R. A., Jr., 1993. Ornithological Gazetteer of Ecuador. 2nd ed. Cambridge: Museum of Com- parative Zoology. Paynter, R. A., Jr.. 1997. Ornithological Gazetteer of Colombia. 2nd ed. Cambridge: Museu of Com- parative Zoology. PEREZ-SANTOS, C. 1999. Se rpientes de Panama/Snakes of Panama. Seville: Comité Espanol del Programa MaB y de la Red IberoMaB de la UNESCO. PéREzZ-SANTos, C., AND A. G. Moreno. 1988. Ofidios de Colombia. Torino: Museo Regionale di Scienze Naturali. PérEz-Santos, C., AND A. G. Moreno. 1989. Addenda y corrigenda al libro “Ofidios de Colombia.” Bollet- tino del Museo Regionale di Scienze Naturali, Torino 7: 1-17. SYSTEMATICS OF DENDROPHIDION PERCARINATUM (COLUBRII DAE ) PEREZ-SANTOS, C., AND A. G. Moreno. 199 1a. Distribucién y amplitud altitudinal de las serpientes en Ecuador. Sey Espanola de Herpetologta 5: 125-140. PEREZ-SANTOS, C., AND A. G. Moreno. 1991b. Ser- pientes de Ecuador. Torino: Museo Regionale di Scienze Naturali. PEREZ-SANTOS, C., A. G. MORENO, AND A. GEARHART. 1993. Checklist of the snakes of Panama. Revista Espanola de Herpetologta 7: 113-122 PETERS, J. A., AND B. OREJAS-MIRANDA. 1970, Catalogue of the Neotropical Squamata: part I. Syalkas: Bulletin of the United States National Museum 29 7-e l= 347. Pounps, J. A. 2000. Amphibians and reptiles, pp. 149- 177. In N. M. Nadkarni and N. T. Wheelwright (eds.), Monteverde, Ecology and Conservation of a Tropical Cloud Forest. New York: Oxford Univ. Press. Ranp, A. S., AND C. W. Myers. 1990. The herpetofauna of Barro Colorado Island, Panama: an ecological summary, pp. 356—409. In A. H. Gentry (ed.), Four Neotropical Rainforests. New Haven, Connecticut: Yale Univ. Press. RasMUSSEN, J. B. 1993. A taxonomic review of the Dipsadoboa unicolor complex, including a phylo- genetic analysis of the genus (Serpentes, Dipsadi- Ae Boiginae). Steenstrupia 19: 129-196. Royas- RUNJAIC, F. J. M., anp E. E. INFANTE RIVERO. 2008. First record of the forest racer snake Dendrophidion percarinatum (Cope, 1893) (Ser- pentes: Colubridae) from Venezuela. Caribbean Journal of Science 44; 128-130. Roze, J. A. 1966. La Taxonomia y Zoogeografia de los Ofidios en Venezuela, Caracas: Univ. Central de Venezuela. SANTOS-BARRERA, G., J. PACHECO, F. MENDOZA-QUIJANO, Pa BoLaNnos.G. CHaves. G.-C, Day P..R: EARLICH, AND G. CEBaLLos. 2008. Diversity, natural history and conservation of amphibians and reptiles from the San Vito region, southwestern Costa Rica. Revista de Biologta Tropical 56: 755-778. SAVAGE, J. M. 1973. A preliminary handlist of the Herpetofauna of Costa Rica. Lawrence, Kansas: American Society of Ichthyologists and Herpetol- ogists. SavacE, J. M. 1980. A Handlist with Preliminary Keys to the Herpetofauna of Costa Rica. Los Angeles: Allan Hancock Foundation. SavacE, J. M. 2002. The Amphibians and Reptiles of Costa Rica, a Herpetofauna between Two Conti- nents, between Two Seas. Chicago: The Univ. of Chicago Press. SAVAGE, J. M., AND F. BotaNos. 2009. A checklist of the amphibians and reptiles of Costa Rica: additions and nomenclatural revisions. Zootaxa 2005: 1-23. SAVAGE, J. M., AND J. Vita. 1986. Introduction to the Herpetofauna of Costa Rica/Introduccién a la Herpetofauna de Costa Rica. Ithaca, New York: Society for the Study of Amphibians and Reptiles. Scumut, K. P. “1933” [1935]. Amphibians and reptiles collected by the Smithsonian Biological Survey of © Cadle D435 the Panama Canal Zone. Smithsonian Miscella- neous el §9: 1-20. SCOT NN... J. M. SavacE, anp D. C. Rosinson. 1983. cae of reptiles and amphibians, pp. 367-374. In D. H. Janzen (ed.), Costa Rican Natural History. Chicago: Univ. of Chicago Press. SEXTON, O. J., AND H. HEatwo.e. 1965. Life notes on some Panamanian snakes. Journal of Science 5: 39-43. SmitH, H. M. 1941. A new name for the Mexican snakes of the genus Dendrophidion. Proceedings of the Biolog- ical Society of Washington 54: 73-76. SmitH, H. M. 1955. Effect of preservatives upon pattern in a Mexican garter snake. Herpetologica 14: 165-168. Smitu, H. M. 1958. Handlist of the snakes of Psanama. Herpetologica 11: 222-224. SMITH, H. M., anp C. Grant. 1958. New and noteworthy snakes from Panama. Herpetologica 14: 207215. Suiru, M. A. 1943. The Fauna of British India, Ceylon, and Burma, Including the Whole of the Indo-Chinese Sub-region. Reptilia oa Amphibia. Vol. 11]—Ser- pentes. "London: Taylor and Francis Ltd. SOLORZANO, A. 2004. Serpientes de Costa Rica: Dis- tribucion, Taxonomia e Historia Natural/ Snakes of Costa Rica: Distribution, Taxonomy and Natural History. Eagle Mountain, Utah: Eagle Mountain Publishing LC. STAFFORD, P. ip 2003. Trophic ecology and reproduction in three species of Neotropical forest racer (Dendrophidion; Colubridae). Herpetological Jour- nal 13: 101-111. Stuart, L. C. 1932. Studies on Neotropical Colubrinae. I. The taxonomic status of the genus Drymobius Fitzinger. Occasional Papers of the Museum of Zoology, University of Michigan 236: 1-16. Taytor, E. H. 1951. A brief review of the snakes of Costa Rica. University of Kansas Science Bulletin 34, part 1: 3-188. Taytor, E. H. 1954. Further studies on the serpents of Costa Rica. University of Kansas Science Bulletin 36, part 2: 673-801. Test, F. H., O. J. Sexton, AND H. HEATWoLe. 1966. Reptiles of Rancho Grande and vicinity, Estado Aragua, Venezuela. Miscellaneous Publications of the Museum of Zoology, University of Michigan 128: 1-63. VILLA, J., L. D. Witson, anp J. D. JoHNson. 1988. Middle American Her; A_ Bibliographic Checklist. Columbia: Univ. of Missouri Press. WHITFIELD, S. M., K. E. BELL, T. Puivippi, M. Sasa, F. BotaNnos, G. CHAves, J. M. Savace, AND M. A. DonNELLY. 2007. Amphibian and reptile declines over 35 years at La Selva, Costa Rica. Proceedings of the National Academy of Sciences, USA 104: 8359-8356. Witson, L. D., AND J. R. Meyer. 1985. The Snakes of Honduras. 2nd ed. Milwaukee: Milwaukee Public Museum. Witson, L. D., anp J. H. Townsenp. 2006. The herpetofauna of the rainforests of Honduras. Caribbean Journal of Science 42: 88-113. history Caribbean 344 Bulletin of the Museum of Comparative Zoology, Vol. 160, No. 6 Wison, L. D., AND J. H. TOWNSEND. 2007. Biogeography and conservation of the herpetofauna of the upland pine-oak forests of Honduras. Biota Neotropica 7: 131-142. Wo tr, M., and Y. L. WerNeER. 1994. The striped colour pattern and striped/non-striped polymor- phism in snakes (Reptilia: Ophidia). Biological Reviews of the Cambridge Philosophical Society 69: 599-610. YANEZ-MuNoz, M., P. Mrza-Ramos, C. C. MARTINEZ, AND M. Reyes P. 2009. Anfibios y reptiles del sur occidente de Ecuador (version 1). Rapid Color Guide (The Field Museum) 237: 1-4. US ISSN 0027-4100 MCZ Publications Museum of Comparative Zoology Harvard University 26 Oxford Street Cambridge, MA 02138 mczpublications@mcz.harvard.edu © The President and Fellows of Harvard College 2012. Bulletin of the Museum of Comparative Zoology Volume 160, Number 7 18 February 2013 Phylogeny of the Dactyloa Clade of Anolis Lizards: New Insights From Combining Morphological and Molecular Data MARIA DEL ROSARIO CASTANEDA AND KEVIN DE QUEIROZ HARVARD UNIVERSITY | CAMBRIDGE, MASSACHUSETTS, U.S.A. BULLETIN OF THE Museum of Comparative Zoology BOARD OF EDITORS Editor: Jonathan Losos Managing Editor: Deborah Smiley Associate Editors: Andrew Biewener, Scott Edwards, Brian Farrell, Farish Jenkins, George Lauder, Gonzalo Giribet, Hopi Hoekstra, Jim Hanken, Jim McCarthy, Naomi Pierce, and Robert Woollacott Publications Issued or Distributed by the Museum of Comparative Zoology Harvard University Bulletin 1863- Breviora 1952- Memoirs 1865-1938 Johnsonia, Department of Mollusks, 1941-1974 Occasional Papers on Mollusks, 1945- General queries, questions about author guidelines, or permissions for MCZ Publications should be directed to the editorial assistant: MCZ Publications Museum of Comparative Zoology Harvard University 26 Oxford Street Cambridge, MA 02138 mczpublications@mcz.harvard.edu EXCHANGES AND REPRINTS All of our publications are offered for free on our website: http://www.mcz.harvard.edu/Publications/index.html To purchase individual reprints or to join our exchange program, please contact April Mullins at the Ernst Mayr Library: mayrlib@oeb.harvard.edu. This publication has been printed on acid-free permanent paper stock. © The President and Fellows of Harvard College 2013. PHYLOGENY OF THE DACTYLOA CLADE OF ANOLIS LIZARDS: NEW INSIGHTS FROM COMBINING MORPHOLOGICAL AND MOLECULAR DATA MARIA DEL ROSARIO CASTANEDA'?? AND KEVIN DE QUEIROZ? CONTENTS Note Added in Proof 345 Abstract 346 Introduction 347 Current Taxonomy within Dactyloa 349 Y Materials and Methods Sil Taxon and Character Sampling 351 Character Coding 352 Continuous characters 352 Polymorphic characters Boe Comparison of coding methods 353 Morphological Data Sets and Phylogenetic Analyses 353 Combined Data Sets and Phylogenetic Analyses 354 Tests of Phylogenetic Hypotheses 300 Results 356 Comparisons Between Coding Methods 356 Phylogenetic Analyses 356 Morphology -only data sets 356 Combined data sets 357 Tests of Phylogenetic Hypotheses 363 Discussion 364 Differences Among Coding Methods 365 Phylogeny of Dac tyloa 366 Previously Recognized Taxa 370 Proposed Taxonomy oul Dactyloa 374 aequatorialis series 3716 latifrons series SW Megaloa 378 punctatus series 379 ' Department of Biological Sciences, The George Washington University, 2023 G Street NW, Washing- ton, DC 20052. Department of Vertebrate Zoology, National Museum of Natural History, Smithsonian Institution, MRC 162, Washington, DC 20560. Author for correspondence (mcastanedaprada@fas.harvard.edu). Address through April 2014: Museum of Compar- ative Zoology, Harvard University, 26 Oxford Street, Cambridge, Massachusetts 02138. Bull. Mus. Comp. Zool., 160(7): 345-398, February, roquet series 38] heterodermus series 38] Phenacosaurus 382 Incertae sedis 384 Acknowledgments 385 Appendix I. Morphological character descriptions 385 Literature Cited 394 NOTE ADDED IN PROOF Shortly after our paper was accepted, Nicholson and colle: agues published a phylogenetic analysis of anoles seh a proposal to divide Anolis into eight genera (Nicholson, K. E., B. I. Crother, C. Guyer, anid If M. Savage. 2012. It is time for a new classification of anoles (Squamata: Dactyloidae). Zootaxa 3477: 1- 108). Here, we comment briefly on their study as it pertains to the phylogeny and taxonomy of the Dactyloa clade. Despite not inferring Dactyloa to be monophyletic in the tree used for ther proposed taxonomy (i.e., the consensus tree from the combined morphological and molecular parsimony analysis; their fig. 5A, note positions of ee bonairensis, chloris, A. per- accae, and A. apollinaris), Nicholson et al. (2012) recognized Deciulon as one of their eight genera without making reference to this inconsistency (although Dactyloa was inferred to be monophyletic in sieir molecular tree, fig. 4A). By contrast, our combined data set supported the monophyly of Dactyloa (Figs. 3, 4), and we have chosen to treat Dactyloa as a subclade of Anolis rather than as a separate genus in the interest of avoiding disruptive and unnecessary name changes. Some of our informally named series correspond, with some differences in species composition, to the species groups proposed by Nicholson et al. (2012). We describe the differences below. Our latifrons series corresponds to their latifrons species group, except that in the tree purportedly used for their taxonomy (fig. 5A), A. aequatorialis and A. ventrimaculatus were inferred to be part of this species 2013 345 346 group (both species are absent from their molecular tree, fig. 4A), although their classification (appendix III) places both species in their punctata species group with no explanation for this inconsistency. We inferred these two species with strong support to be part of a monophyletic aequatorialis series that is mutually exclusive with respect to both the latifrons and punctatus series. Additionally, we have tentatively placed A. mirus and A. parilis in the aequatorialis series based on their previous inclusion in the traditional aequatorialis series (Williams, 1975; Ayala- Varela and Velasco, 2010); the tentative assignment reflects the current absence of these species from explicit phylogenetic analyses. By contrast, Nicholson et al. (2012) assigned A. mirus and A. parilis, neither of W ak was included in any of their analyses, to their latifrons species gr oup without explanation. Finally, we placed A. propinquus in the latifrons series based on its hypothesized close relationship to A. apollinaris (Williams, 1988). By contrast, Nicholson et al. (2012) placed this species, which was not included in any of their phylogenetic analyses, in their punctata species group without explanation. The combination of our aequatorialis and puncta- tus series corresponds roughly to the punctata species group in the classification of Nicholson et al. (2012, appendix III). We inferred these two series to be mutually exclusive clades (results further supported by molecular data alone: Geen and de Queiroz, 2011). Contradicting their own taxonomy, the tree of Nicholson et al. (2012. fig. 5A) supports the separation of the aequatorialis series, in that A. aequatorialis, A. ventrimaculatus, A. chloris, and A. peraccae are not inferred to be part of their punctata species group, despite being referred to that group in their classifi- cation (appendix III). Their tree does place A. fasciatus in their punctata species group, whereas our results indicate that this species is part of the aequatorialis series. We have treated A. calimae and A. cuscoensis as incertae sedis within Dactyloa based on conflicting results for A. calimae (also found by Castafieda and de Queiroz, 2011) and the inferred inclusion of A. cuscoensis by Poe et al. (2008) in clades not inferred in our study. By contrast, Nicholson et al. (2012) referred these two species to their punctata species group, although neither species was included in any of their phylogenetic analyses. Similarly, we have treated A. laevis and A. phyllorhinus, species formerly placed in the laevis series, as incertae sedis based on their current absence from explicit phyloge- netic analyses (although we consider it likely that A. phyllorhinus belongs to the punctatus series). By contrast, Nicholson et al. (2012) assigned both of these species to the punctata species group, although neither was included in any of their phylogenetic analyses. Our Phenacosaurus and our heterodermus series both correspond approximately to the heterodema species group of Nicholson et al. (2012), with the exception that they included A. carlostoddi, A. bellipeniculus, and A. neblininus. We consider these Bulletin of the Museum of Comparative Zoology, Vol. 160, No. 7 three species as incertae sedis within Dactyloa based on conflicting results in our analyses for A. carlostoddi and A. neblininus and the absence from explicit phyloge- netic analyses of A. bellipeniculus, as well as_ its previously ‘inferred close relationship to A. neblininus (Myers and Donnelly, 1996). Our roquet series corresponds approximately to their roquet species group. However, in the tree purportedly used for their taxonomy (their fig. 5A), their roquet species group is not monophyletic: A. bonairensis is inferred as sister to A. occultus outside of Dactyloa (A. bonairensis is not included in their molecular-only tree; fig. 4A). By contrast, we inferred A. bonairensis to be part of a monophyletic roquet series (Figs. 3, 4). Our combined analyses are based on a sample of 60 of the 83 currently recognized species in the Dactyloa clade, 40 of which were sampled for molecular data, whereas the combined analysis of Nicholson et al. (2012) is based on a sample of 31 Dactyloa species, 16 of which were sampled for molecular data (three others were sampled for molecular data only). Additionally, our molecular data consists of ~4,950 base positions representing three gene regions and both mitochon- drial and nuclear DNA, whereas theirs consists of ~1,500 base positions representing one of the two mitochondrial gene regions used in our study. Because our results are based on larger samples of Dactyloa species (for both molecular and morphological data), as well as larger samples of molecular data (with respect to both numbers of bases and numbers of gene fragments, and including both mitochondrial and nuclear genes), and because many of their taxonomic conclusions that differ from ours are either contradict- ed by their own results or unsubstantiated, we do not consider any of the differences between our phyloge- netic results and taxonomic conclusions compared with those in the study by Nicholson et al. (2012) to warrant changes to our proposed taxonomy. In contrast to Nicholson et al. (2012), we refrain from assigning some species to series and treat some taxonomic assignments as tentative because of contradictory results or poorly supported inferences, and we present justifications for all taxonomic decisions pertaining to species not included in our analyses. Apstract. We present a phylogenetic analysis of the Dactyloa clade of Anolis lizards, based on morpholog- ical (66 characters of external morphology and osteology) and molecular (~4,700 bases of mitochon- drial and nuclear DNA) data. Our set of morphological characters includes some that exhibit continuous variation and others that exhibit polymorphism within species; we explored different coding methods for these classes of characters. We performed parsimony and Bayesian analyses on morphology-only and com- bined data _ sets. Additionally, we explicitly tested hypotheses of monophyly of: 1) Dact yloa including Phenacosaurus, 2) Dactyloa excluding Phenacosaurus (as _ traditionally circumscribed), 3) taxa previously ranked as series or species groups described based on PHYLOGENY OF THE DacryLoa ¢ Castaneda and de Queiroz 347 morphological characters, and 4) clades inferred from molecular data. The morphological data alone did not yield Dactyloa or any of the previously recognized series described based on ieee al characters: only the Phenacosaurus clade (as de limited based on molecular data) was inferred with the morphological data, and only in the parsimony analysis. In contrast, Dactyloa was inferred as monophyletic with the combined data set, although topology tests failed to reject the hypothesis of non-monophyly. Additionally, five clades inferred based on molecular data (eastern, latifrons, Phenacosaurus, roquet, and western) were inferred with the combined data sets with variable support and including additional species for which molecular data were not available and which have geographic distributions that conform to those of the clades in which they were included. Of the previously recognized taxa based on morphological characters, only athe roquet series, which corresponds in species composition to the roquet clade, was inferred with the combined data. Topology tests with the combined data set ee the monophyly of the aequatorialis, latifrons (as traditionally circumscribed), and punctatus series but not that of the tigrinus series and Phenacosaurus (as traditionally Careumseribed). Our phylogenetic analyses and topology tests indicate that a new taxonomy for Dac tyloa is warranted; we therefore present a revised taxonomy based on the results our phylogenetic analyses and employing phylogenetic definitions of taxon names. Key words: Anolis, ; coding, Phylogeny, Taxonomy Character Dactyloa, INTRODUCTION The Anolis clade, one of the most diverse groups of vertebrates traditionally ranked as a genus, is composed of 384 currently recognized species (Uetz, 2012). This group of lizards is primarily Neotropical in distri- bution. Its members are characterized (with a few exceptions) by the possession of adhesive toe pads formed by laterally expanded subdigital scales, called lamellae, that are Boexe by microscopic setae, and of extensible and often brightly colored throat fans, called dewlaps, that are sup- ported by elongated second ceratobran- chials and occur in males and often in females (Etheridge, 1959). Based on Etheridge’s (1959) seminal work on the phylogeny and taxonomy of anoles, two large groups, traditionally ranked as sections, were informally recog- nized within Anolis based on the absence (alpha section) or presence (beta section) of transverse processes on the anterior auto- tomic aia vertebrae. Each section was further subdivided into series and species groups based on morphological characters (Etheridge, 1959; Galant 1976a,b). Sub- sequent to the morphological studies of Etheridge (1959) and Williams (1976a,b), a wide variety of data have been brought to bear on the phylogeny and taxonomy of Anolis, including albumin immunology (eg Gorman et al., 1980b, 1984; Shochat fe Dessauer, 1981), allozymes (e.g., Gorman and Kim, 1976; Gorman et al., 1980a; Burnell and Hedges, pee, behavior (e.¢., Gorman, 1965S), karyotypes (e.g., Gorman et al., 1968, 1983: Gorman and Samar 1975), and DNA sequences (e. g. Jackman et al., 1999: Schneider et al., 2001: Glor et al., 2003). Analyses of Picse data have pr avided support for the monophyly of the beta section and of several series and species Brolps (e.g., Creer et al., foe oe et al., 2001; Jackman et al., ; Nicholson, 2002). However, they ie Ke indicated that other groups, including the alpha section, are not monophyletic. Additionally, the phylogenetic relationships within and among some groups are still disputed (e.g., Jackman et ale 1999: Nicholson, 2002; Poe, 2004). Within the alpha section, Etheridge (1959) recognized the latifrons series for species with at least four postxiphisternal chevrons attached to the bony dorsal ribs and an arrow-shaped interclavicle (in which the lateral processes of the interclavicle are divergent from the Sasa parts of the clavicles). Etheridge’s (1959) latifrons series was composed of all mainland alpha Anolis (excluding Phenacosaurus; see below) along with the species in the roquet series from the southern Lesser Antilles, as well as Anolis agassizi and A. gorgonae from the Pacific islands of Malpelo and Gorgona, respectively. The latifrons series of Ether- idge (1959) corresponds to the genus Dactyloa, one of five genera recognized by Guyer and Savage (1987 [1986]) based on a sroposal to “divide” Anolis taxonomically. Although the recognition of those genera is 348 controversial (Cannatella and de Queiroz, 1989: Williams, 1989: Poe and Ibanez, 2007), some recent authors apply some of the same names to clades within Anolis regardless of rank and not necessarily with en composition (e.g., Nicholson, 2002: Brandley and de Queiroz, 2004: de Queiroz and Reeder, 2008). In the present study, we use the name Dactyloa for the clade originat- ing in the most recent common ancestor of fhe species included in the genus Dactyloa by Savage and Guyer (1989), nich also includes the ariles formerly assigned to the genus Phenacosaurus according to the ees of gees phylogenetic analyses (e. g., Jackman et 1999: Boe 2004: Nickolcont et al., 2005: Gistact, and de Queiroz, 2011). Currently, there are 83 recognized species in the Dactyloa clade, distributed among seven subgroups (based on morphological charac- ters) commonly assigned to the rank of series: aequatorialis, laevis, ‘Tatifrons, punctatus, Phe- nacosaurus, roquet, and tigrinus (see “Cur- rent Taxonomy within Dact Fyloa,” below). Three phy logenetic analyses have includ- ed more than 20% of the currently recog- nized Dactyloa species: Poe (2004) ) included 28 species, Nicholson et al. (2005) included 17 species (13 of which were included in Poe |2004]), and Castaneda and de Queiroz (2011) included 42 species (including 22 of 28 of Poe [2004] and 15 of 17 of Nicholson et al. [2005]). In Poe’s (2004) combined analysis of allozyme, karyotype, morpholog- ical, and molecular data. Dactyloa (as defined in the previous paragraph, the name was not used by Poe) was inferred to be monophyletic based on the arrow shape of the interclavicle and the presence of a splenial in the mandible. However, boot- strap support for the clade (Poe, 2004, fig. 2, node 352) was less than 50% and both morphological characters supporting it are Bulletin of the Museum of Comparative Zoology, Vol. 160, No. 7 Of the seven subgroups described within Dactyloa based on morphological charac- ters, only the roquet series has been consistently inferred by several phylogenetic analyses and has passed explicit statistical tests of monophyly (Jackman et al., 1999: Poe, 2004: Nicholson et al., 2005: Castafieda and de Queiroz, 2011). Phenacosaurus, as traditionally circumscribed, was iabomed as monophy letic by Poe (2004) and Nicholson et al. (2005), although only two and three species of this group, respectively, were included in their phylogenetic analyses. In the analyses of Castafieda and de Queiroz (2011), which included six species of na nerenrs. the group was inferred as monophyletic with the exception of A. neblininus. The remaining subgroups have not been inferred in phylogenetic analyses or passed explicit statistical tests of mono- phyly, although the laevis series has not been tested fPoe 2004: Nicholson et al., 2005; Castaneda and de Queiroz, 2011). In contrast to the poor support for the traditional series, Castaneda and de Queiroz (2011) inferred five strongly supported subclades within Dactyloa, ‘which they recognized informally as eastern, latifrons, Phenacosaurus, roquet, and western clades. Although some of these clades bear the same names as species gr oups recognized by Williams (1976b) and series recognized by Savage and Guyer (1989), their composition is not necessarily the same. In this study, we describe and score 66 morphological characters (external and os- teological) for 60 species of Dactyloa and 6 outgroup species (including non-Dactyloa Anolis and non-Anolis Polvelneness to resolve the phylogenetic relationships with- in the Dact yloa clade. We analyze the morphological characters alone ‘and in combination with ~4,720 bases of DNA reversals to ancestral conditions. In the analyses of Nicholson et al. (2005, fig. 1) and Castafieda and de Queiroz (2011, fig. 1), based on molecular data, Dactyloa was inferred with moderate to strong bootstrap support (280%) and Bayesian posterior probabilities (=0.90). sequence data presented by Castafieda and de Queiroz (2011). We perform parsimony and Bavesian analyses and examine differ- ent coding methods for continuous and poly morphic characters. We use tree topol- ogy tests to test hypotheses of monophyly of: 1) Dactyloa including Phenacosaurus, 2) PHYLOGENY OF THE DacTyLoa * Castaneda and de Queiroz 349 Dactyloa excluding Phenacosaurus (as tra- ditionally Daeaenecrihe d), 3) the traditional- ly recognized series delimited based on morphological characters (Williams, 1976b; Savage and Guyer, 1989), and 4) the clades saferred based on molecular data (Casta- neda and de Queiroz, 2011). Based on the results of our analy ses, we present a revised taxonomy that is consistent with the current knowlec lge of the phylogenetic relationships within the Dactyloa dade! Current Taxonomy within Dactyloa Based on morphological characters, six subgroups ranked as species ques by Williams (1976b) and as series by Savage and Guyer (1989) have been recognized within Dactyloa: aequatorialis, laevis, lati- frons, punctatus, roquet, and tigrinus. In agreement with recent phylogenetic analyses (e.g., Jackman et al., 1999; Poe, 2004; Nicholson et al., 2005: Castaneda and de Queiroz, 2011), we recognize the group of species previously identified as the genus Phenacosaurus as an additional subgroup of Dactyloa. Some Dactyloa species Have not been assigned to any of these subgroups; for example, "Anolis agassizi, A. anchicayae, A. cuscoensis, and A. ee were referred to what is here recognized as the Dactyloa clade, but with no series assignment (Ether- idge, 1959; Poe et al., 2008, 2009a wb). Other species have been assigned to subgroups, but assignment was inconsistent. For example, Anolis kunayalae was described as morpho- logically similar to A. mirus and A. parilis, both members of the aequatorialis series, but assigned by the describing authors (Hulebak etal. 200 Q).to:.the latifrons group sensu stricto (= latifrons species group of Williams, 1976b) which is equivalent to the latifrons series of Savage and Guyer (1989); we therefore consider the series assignment of this species uncertain. 7 Aequatorialis series. The aequatorialis series is currently composed of 13 species: A. aequatorialis, A. anoriensis, A. antio- quiae, A. eulaemus, A. fitchi, A. gemmosus, A. maculigula, A. megalopithecus, A. mirus, A. otongae, A. parilis, A. podocarpus, and A. ventrimaculatus, which are characterized by moderate to large body size (adult male snout-to-vent length [SVL] 66-101 mm), small head series: smooth ventral scales, uniform dorsal scalation, and in some species narrow toe lamellae (Williams, 1976b: Williams and Acosta, ie Ayala- Varela and Torres-Carvajal, 2010; Ayala- Varela and Velasco, 2010: Velasco et al. 2010). The species in the acquatorialis series are distributed between 1,300 and 2.500 m above sea level in the Andes of Colombia (western and central cordilleras) and Ecuador (eastern and western slopes) (Williams and Duellman, 1984; Ayala-Varela and Torres-Carvajal, 2010; Velasco et al., 2010; Ayala and Castro, unpublished). Laevis series. The laevis series, composed of A. laevis, A. phyllorhinus, and A. proboscis, is characterized by the presence of a soft, median protuberance from the snout, called a proboscis (Williams, 1976b, 1979) or nose leaf (Peters and Orces, 1956). Members of this series have .a disjunct geographic distribution: A. laevis is distri- buted in the eastern foothills of the Peruvian Andes, A. proboscis is found at mid-elevations on the western slopes of the Ecuadorian Andes, and A. phyllorhinus is found in central Amazonia (Williams, 1979: Rodrigues et al., 2002). Latifrons series. The latifrons series is composed of 12 species: A. apollinaris, A. ie Bee A. danieli, A. fraseri, A. frenatus, A. insignis, A. latifrons, A. microtus, A. prin- ceps, A. propinquus, A. purpurescens, and A. squamulatus, which are characterized by adult SVL > 100 mm, large dewlaps in adult males (>500 mm‘), expanded toe lamellae, small head scales, smooth to weakly keeled ventral scales, and uniform dorsal scalation (Williams, 1976b; Savage and Talbot, 1978). These species, ails called the giant mainland anoles (Dunn, 1937), are distributed in the lowlands and premontane forests of Costa Rica, western Panama, Colombia (western cordillera), and Ecuador; in the northern central lowlands of Venezuela; and in the inter-Andean valleys of Colombia (Savage 390 and Talbot, 1978: Arosemena et al., 1991: Ayala and Castro, unpublished). Punctatus series. The punctatus series is composed of 21 species: A. anatoloros, A. boettgeri, A. calimae, A. caquetae, A. chloris, A. shoconinne A. deltae, A. dissimilis, A fasciatus, A. festae, A. gorgonae, A. huilae, AOCaTe TN. nigrolineatus, A. peraccae, A. philopunctatus, “A. punctatus, A. santamar- tae, A. soinii, A. transversalis, and A. vaupesianus. The characters used to diag- nose this series include adult SVL < 100 mm, wide toe lamellae (compared with the narrow lamellae observed in the aequatorialis se- ries), small head scales, smooth to weakly keeled ventral scales (except in A. punctatus boulengeri, which has strongly keeled ven- trals), uniform dorsal scalation, and in some species a protuberant snout in males (Wil- liams, 1976b, 1982). Species in the punctatus series are distributed in the western lowlands of Panama, Colombia, and Ecuador; the mid- to high elevations of the Andes of Colombia (including the Sierra Nevada de Santa Marta), Venezuela, and Peru (eastern slope); the Amazon region and the Orinoco delta (Williams, 1982; Rodrigues, 1988; Poe and Yanez-Miranda, 2008: Poe et al., 2008, 2009a,b; Ayala and Castro, unpublished). Roquet series. The roquet series is composed of 9 species: A. aeneus, A. blanquillanus, A. bonairensis, A. extremus, A. griseus, A. luciae, A. richardii, A. roquet, an A. trinitatis. The monophyly of this series is supported by karyological (Gorman and Atkins, 1969), morphological (Lazell, 1972; Poe, 2004), and molecular data (cytochrome b sequences, Giannassi et al., 2000; ND2 sequences, Creer et al., 2001). Six morphological synapomorphies support the monophyly of this series: 1) greater sexual size dimorphism, 2) an increase in interparietal scale size relative to surround- ing scales, 3) an increase in mean number of postmental scales, 4) a straight (as opposed to concave) posterior border of the mental scale, 5) supraorbital semicircles in contact, and 6) interparietal scale in contact with the supraorbital semicircles (Poe, 2004). The roquet series is distributed in the southern Bulletin of the Museum of Comparative Zoology, Vol. 160, No. 7 Lesser Antilles, from Martinique south to Grenada, and on the islands of La Blan- quilla, Bonaire, Tobago, and Trinidad (where A. aeneus and A. trinitatis have been introduced; Gorman and Dessauer, 1965, 1966) and Guyana (where A. aeneus has been introduced; Gorman and Des- sauer, 1965; Gorman et al., 1971). Tigrinus series. The tigrinus series is composed of 9 species: A. lamari, A. menta, A. nasofrontalis, A. paravertebralis, A. pseu- dotigrinus, A. ruizii, A. solitarius, A. tigrinus, aaile A umbrivagus, and is chavacteneed by small body size (adult male SVL = 40- 60 mm), large smooth head scales, a large interparietal scale bordered by large scales and usually in contact with the supraorbital semicircles, and ventral scales smooth and larger than dorsal scales. Some species exhibit a parietal knob (a small projection of the posteriormost end of the central ridge of the Y-shaped parietal crests), externally visible in some species on the occipital area between the post-interparietal scales and nape scales (Williams, 1976b, 1992). Species in the tigrinus series are distributed in high elevations of the Sierra Nevada de Santa Marta (Colombia), the Andes of Colombia (eastern cordillera) and Venezuela, and the Atlantic forest of southeastern Brazil (Wil- liams, 1992: Bernal Carlo and Roze, 2005). Phenacosaurus. Phenacosaurus is com- posed of 11 species: A. bellipeniculus, A carlostoddi, A. euskalerriari, A. heteroder- mus, A. inderenae, A. neblininus, A. nice- fori, A. orcesi, A. tetarii, A. vanzolinii, and A. williamsmittermeierorum. Earlier, Phe- nacosaurus was considered a separate genus from Anolis (Barbour, 1920) based on the heterogeneous dorsal scalation (enlarged round flat scales surrounded by smaller scales and granules), the tail structure (probably prehensile), an elevated rim of head plates (casque), digits widely and evenly dilated (such that their sides are nee and a “feebly developed” dorso- nuchal crest (Lazell, 1969). However, recent phylogenetic analyses (Poe, 1998, 2004; Jackman et al, 1999; Nicholson et aly 2005: Gastaacda and de Queiroz, 2011) PHYLOGENY OF THE DactyLoa * Castaneda and de Queiroz 3 inferred these species to be nested within the clades composed of the species assigned to both Anolis and Dactyloa; chavetore: we here consider Phenacosaurus another sub- group of Dactyloa. Phenacosaurus species are distributed in the Andean ipninl (between 1,300 and 3,000 m) of Colombia, northern Ecuador, central Peru, and west- ern Venezuela and the isolated tepuis of southeastern Venezuela (Lazell, 1969: Myers: et als, 1993; Barros et al., 1996, Myers and Donnelly, 1996: Williatas et al., 1996: Poe and Yanez-Miranda, 2007). MATERIALS AND METHODS Taxon and Character Sampling Morphological data were collected for 60 species of Dactyloa, representing the sub- groups aequatorialis, latifrons, laevis, Phe- nNacosaurus , punctatus, roquet, and tigrinus. Anolis anoriensis (a recently described species formerly considered part of A. eulaemus) was treated as conspecific with A. eulaemus given that the description of the former was published after our data analy- Ul specimens examined is given in the Supple- mentary Appe ndix 1.’ Sixty-six morphological characters were examined, including both continuous char- acters (those that can be represented by real numbers, tail length) and discrete characters (those that can only be repre- sented by integer values, including meristic and presence/absence, number of elongated superciliary scales). This data set includes characters of external morphology and osteology that have been previously used in Anolis phylogenetic analyses, have been regarded as diagnostic for Anolis subgroups. or have been used historically for species identification (Etheridge, 1959: Williams, 1976b, 1989; de Queiroz, 1987; Etheridge and de Queiroz, 1988; Frost and Etheridge, 1989; Williams et al., 1995; Poe, 1998, 20004: Jackman et al., 1999; Brandley and de Queiroz, 2004). Given that sexual dimorphism occurs in many species of Anolis (e.g., Schoener, 1969; Butler et al., 2000, 2007), characters were scored for both males and females and combined only when t tests (for continuous characters) or sale tests (for discrete characters) Po c.2., Po €.2., ses were performed. Six species were included as outgroups: one non-Anolis Polychrotinae Bel ehit marmoratus) and five species representing different series of non-Dactyloa Anolis (Anolis bimaculatus, A. cuvieri, A. equestris, A. occultus, A. sagrei). A total of 643 alcohol-preserved (66 species: 393 males, 250 females), 123 dry (49 species), and 10 cleared and stained (9 species) specimens were examined. Addi- tional data were collected from radiographs of 394 specimens (60 species). External characters were scored for all 66 species, and osteological characters were scored for 63 species (14 of which were only scored from radiographs and thus lack data for all cranial characters). The largest specimens available were examined as a proxy for including adult specimens only. All speci- mens measured at least 70% of the maxi- mum SVL reported in the literature for the same sex and species (Williams and Acosta, 1996; Savage, 2002). A complete list of revealed no significant difference between the sexes or when tests could not be performed because sample sizes were too small. When significant differences were found, only data from males were used. However, given the small number of specimens available as dry skeletons and the absence of information on sex for roughly one-third of them, data for charac- ters examined on dry specimens were combined without evaluating whether some of the characters exhibit sexual dimorphism. To ensure character independence, we performed correlation tests between char- acters. To remove the effects of correlation, we estimated residuals by regressing each variable against the correlated variable: residuals were used in subsequent analyses. In cases where a character was correlated with several others (e.g., SVL, head length, ' Supplementary material referenced in this paper is available online at www.mcz.harvard.edu/Publications/. 352 Bulletin of the Museum of Comparative Zoology, Vol. 160, No. 7 and head width), after residual estimation between two of the variables a second correlation test was performed to ensure that the resulting residuals were not still correlated with the other characters. A list of the characters analyzed, including mea- surement and coding details, is given in Appendix I. Character Coding Continuous Characters. Continuous char- acters were coded using two methods: the gap-weighting method of Thiele (1993), and Torres-Carv ajal’s (2007) modification of Wiens (2001) modification of Thiele’s meth- od. In Thiele’s (1993) method, continuous characters are coded into discrete values while retaining information about order and relative distance between states. Average values per species were standardized by calculating the natural logarithm (In, or In + 1 when zero average values were present) to ensure equal variances; then each standard- ized average value (x) was range-standard- ized by applying the formula x, = [(x — min)/(max — iin) 123 mm: Williams and Acosta, 1996) compared with most Anolis species, and as a result, a decrease in maximum male SVL is inferred as a syni ipomorphy of Dactyloa. In the parsimony analysis, the major clades inferred by Casta- fieda and de Queiroz (2011)—that is, eastern, latifrons, Phenacosaurus, roquet, and west- erm—were inferred with weak to strong nodal support (BS 6-93%). Similarly, in the Bayesian analysis, all five clades were in- ferred with weak to strong nodal support (0.15 < PP < 0.97). For the purpose of assigning species to these clades (eastern, latifrons, roquet, Phenacosaurus, and west- ern), the clades were delimited using nodes bounded by species for which molecule ge were available (e.g., the eastern clade i defined as the clade originating with the aa common ancestor of a particular set of species inferred from molecular data [Cas- taneda and de Queiroz, 2011], thus excluding species outside that node that are more closely related to the eastern clade than to any of the other four mutually exclusive clades). In both parsimony and Bayesian analyses, the same sets of additional species, for which only morphological data were available, were included in the western and Phenacosaurus clades. In the case of the latifrons and eastern clades, different sets of additional species (for which only morpho- logical characters were av ailable) were in- formed in the parsimony and Bayesian analyses. No additional species were includ- ed in the roquet clade in either analysis. In the following paragraphs, the species com- position of Genel clade is detailed, with daggers (+) indicating species lacking molec- abe Sars. The sy napomorphies that support each clade, inferred based on the parsimony analysis, are given in Supplementary Appen- dix 5. The western clade was inferred with weak to moderate support in the parsimony and Bayesian analyses (BS = 25%, PP = 0.82; Figs. 3, 4) and is supported by nine 398 Bulletin of the Museum of Comparative Zoology, Vol. 160, No. 7 Polychrus marmoratus A aequatorialis § outgroups “fi eu aemus Bi aequatorialis a AC inesel hecus ™ laevis A. frenate Wi latifrons 82 Ae rons rinceps HI Phenacosaurus y a ‘Sq A Wee's @ punctatus a podocarptis maculigula ® roquet Agemimosus Ss Re, tigrinus a ‘bigeitgen Mi unassigned 22 oris A. peraccae A. fasciatus A. festae re A. eggs! : aeneus A. trinitatis 60 -—A. extremus A. luciae ag ae . GISCUS .. A nchardii 81 A. bimaculatus llinaris Oa ii oabidae am eae urpurescens A i Bs A. cuvieri A. equestris A. insignis A. chocorum . transversalis A. huilae A. punctatus vaupesianus A. anatoloros A. caquetae A. dissimilis santamartae A. philopunctatus A. jacare A. ruiZil A.calimae KX bonairensis jie mente solitarius A. carlostoddi . orcesi te A. cosndena 14 61 A. etgrod rmus A wanzolinie’ | Phenacosaurus A. tetarii A. nicefor A. microtus A. proboscis A. neblininus A. tigrinus A. occultus 7.0 Figure 1. Most parsimonious tree inferred with the Torres-freq morphology-only data set (TL = 466.63, Cl = 0.22, RI = 0.53). Bootstrap support (BS) values are shown above branches; missing values above branches indicate BS = 0%. The traditional species groups/series based on morphological characters (see text for details) are differentiated by color. One major Dactyloa subclade, of those described based on molecular data (see text for details), is indicated on the right. PHYLOGENY OF THE DactyLoa * Castaneda and de Queiroz 359 Polychrus marmoratus ‘sas A. aequatorialis ®@ outgroups 034 — A’ euiaemus ®@ aequatorialis 0.43), 9 A. A ohio yjae egalopithecus % laevis 0.44 ar ventnnaculanus WS Jatifrons 0.15, 0-30—A, gemmosus i Phenacosaurus 14g A. Ma taetee .25 ca peictanis ms chlaris @ roquet ots} O40. e A jmaculigula aa oettgeri a figrinus ‘ILA fasciatus g Mi unassigned A chocorum | 0 apeen 1 carpus st A fichh squamulatus 0.50 0.4 A. frenatus 0.20 0.73L_,sA. /atifrons 0.93-— A princeps V6 A. fraserl A. insignis 0.631_0.79 A. cuvieri A. microtus 0. A. equestris sidan ate A. ata 0.81 AP carlostoddi _ calimae 0.65 A. occultus A. proboscis .ta_p A. punctatus 0.92 A. vaupesianus A. jacare A. bonairensis 3 0.88— A. richardil p78 a A. griseus : A. extremus 0.76 A. lyciae goassi2i o3qA- roguet 0.45 ~"l 0.1K A aeneus 0.12 0.90 -A. sagrel 0.2¢ A. bimaculatus A. trinitatis A. anatoloros A. huilae asA., PIZII A.. transversalis . caquetae 0.40 A. santamartae wart ante A. dissimilis A. solitarius 0.67 A. neblininus 0.64 | A. menta 0.41 A. tigrinus , 6 ee ees rl ve A. ike ee ee A. heterodermus A. tetarii A. inderenae Se Sette Bad A. vanzolinii 0.3 Figure 2. Bayesian maximum clade credibility tree inferred with the Thiele6-mode morphology-only data set using the Mkv + rv model. Bayesian posterior probabilities are shown above branches. The traditional species groups/series based on morphological characters (see text for details) are differentiated by color. 360 Bulletin of the Museum of Comparative Zoology, Vol. 160, No. 7 Polychrus marmoratus a? 69 A. aequatorialis B® outgroups : jae fers Wi aequatorialis i Aj antioquia le regalopitnecust : emmosu chloris : 7 - ® /atifrons 1 nt festae M@ Phenacosaurus Ast 47 peraccae A. fasciatust ®@ punctatus A. boettgertt A. fitchi & roquet 50 A. squamulatust ® tigrinus A. podocarpus A. huilae @ unassigned 3 A. ag agassizi 6: A. oe nis A. mi Ss Il O inanst : A fee A. Piacoa : A. frenatus latifrons A. latifronst rinceps A. coailece “ A. maculigula A. danieli A. purpurescenst Ai aquetae! Imilis antamartaet+ . calimae 14 19 A. mentat 12 A. neblininus 7: A. solitariusy A. ruiziit 69 A. anatoloras . jacare 2 : 30 yy Carlostoddit é A. tigrinus eastern ils A. punctatus A. vaupesianust* A. transversalis 99 A. aeneus 99 A. extremus o7 | A praree ichardii 90 A. griseus roquet A. trinitatis aw) ‘bonairensis A. luciae 20 ’ SSSs S238 SAQsS sa Ashpterode 8 i inderenae / renae _ A. ‘vanzolinil Phenacosaurus “le “Aiat A. nicefori A. proboscist 04 A. bimaculatus ——— A. sagrel A. cuvieri A. equestris A. occultus 200.0 Figure 3. Most parsimonious tree inferred with the CombTorres-freq combined data set (TL = 10,431.86, Cl = 0.30, RI = 0.43). Bootstrap support (BS) values are shown above branches; missing values above branches indicate BS = 0%. Daggers (7) following species names indicate the species for which only morphological data were available. The traditional species groups/ series based on morphological characters (see text for details) are differentiated by color. Major Dactyloa subclades described based on molecular data (see text for details) are indicated on the right. The Dactyloa clade is indicated with a black dot on the corresponding node. PHYLOGENY OF THE DacTyLoa ¢ Castaneda and de Queiroz 361 morphological characters (Supplementary Appendix 5). In both analyses, this clade is composed “of the same 10 species: A. aequatorialis, A. antioquiaet, A. chloris, A. ae mus, A. fasciatust, A. festae, A. gem- mosus, A. megalopithecus*, A. peraccae, and A. ventrimaculatus. Within this clade, the topologies are largely congruent, with the exception of the position of A. gemmosus and the internal relationships within the clade composed of A. chloris, A. fasciatus, A. festae, and A. peraccae. Additionally, the parsimony analysis, A. boettgeri is inferred with weak support as Hie sister group of the western clade (BS = 18%), whereas in the Bayesian analysis, the sister taxon of the western clade is the clade (A boettgeri, A. huilae) (PP = 0.49). The latifrons clade was inferred in the yarsimony analysis with low nodal support a = 20%; Fig. 3) and is supported by five morphological eee acters (Supplementary Appendix 5); it is composed of 13 species: A. agassizi, A. apollinarist, A. ee A. chocorum, A. danieli, A. fraseri, A. frenatus, A. insignis, A. maculigula, A. microtus, A. latifrons*. A. princeps, and A. purpures- censt. In the Bayesian analysis, the latifrons clade was inferred with low support (PP = 0.46; Fig. 4) and is composed of the same set of species as the parsimony analysis with the addition of A. squamulatus. Three mutually exclusive subclades were inferred by both analyses: (A. agassizi (A. microtus, Asis) Vi Bo = 8270 PP = O61), (A: casildae, A. maculigula) (BS = 92%, PP = 0.62), and (A. frenatus ( (A. latifrons, A princeps) ge = 80%, PP = 0.72). In both analyses, A. philopunctatus was inferred as the sister taxon of the latifrons clade. The eastern clade was inferred in the parsimony analysis (Fig. 3) with low nodal support (BS = 6%) and is supported by seven morphological characters (Supple- mentary Appendix 5); it is composed of eight species: A. anatoloros, A. ae es ueyacane sia orcesi|, A, punctaius, A. tigrinus, A. transversalis, and A. vaupesia- nust. In the Bayesian analysis (Fig. 4), the eastern clade was inferred with low nodal support (PP = 0:15) with 1] species: A. anatoloros, A. dissimilist, A. jacare, A. mentat, A. punctatus, A. ruiz uit, A. santa- martaet, A. solitariust, A. tigrinus, A. transversalis, and A. vaupesianus?. Despite the differences in species composition, two subclades were ‘aterred in both analyses (Figs. 3, 4): (A. transversalis (A. punctatus, A. vaupesianus)) (BS = 61%, PP = 0.71) and (A. anatoloros, A. jacare) (BS = 69%, PP = 0.94). The roquet clade was inferred with strong support in both parsimony and Bayesian analyses (BS = 93%, PP = 0.97; Figs. 3, 4) and is supported by 16 morphological characters (Supplementary Appendix 5). In both analyses, this clade is composed of the same eight species: A. aeneus, A. bo- nairensis, A. extremus, A. griseus, A. luciae, A. richardii, A. roquet, and A. trinitatis. The topology within this clade is identical for both phylogenetic analyses, with all nodes moderately to strongly supported (BS = 72%, PP = 0.96). The Phenacosaurus clade was inferred with moderate support in both ey and Bayesian analyses (BS = 87%; = 0.84; Figs. 3, 4) and is supported by i mor- phological characters (Supplementary Ap- pendix 5). In both analyses, this clade is composed of the same six species, A. euskalerriari, A. heterodermus, A. inderenae, A. nicefori, A. tetariit, and A. vanzolinii, all previously placed in the genus Phenaco- saurus. The relationships aathin this clade are identical between parsimony and Bayes- ian analyses; however, in the parsimony analysis, A. proboscis is inferred as its sister group (BS = 58%), whereas in the Bayesian analysis, A. orcesi is inferred as its sister group with moderate support (PP = 0.87), and A. proboscis is the sister group to that (A orcesi, Phenacosaurus) clade (PP = 0.43). In both parsimony and Bayesian analyses, nine species, A. boettgerit, A. calimae, A. caquetaet, A. fitchi, A. Syae A. neblininus, A. philopunctatust, A: podocarpus, and A. proboscis¥, were not included in any of the five major clades within Dactyloa when those clades are treated as originating in the last lod 362 Bulletin of the Museum of Comparative Zoology, Vol. 160, No. 7 a Polychrus marmoratus 0,72 A, aequatorialis ne A. eulaemu itonure ).9 , A: Pedatpitectiet A. oe jyentrimaculatus 0.39 eracc 0.41 K fasciatust 0,49 | A. festae A. chloris 0.24 0.38 -A. boettgerit western 0.27 oes 0.62 . Maculigula A. casildaeé 0.16 A,,a20 er sguamulatust ; 63 _f A, tenatus latifrons ve rinceps ge A. We atifrons 0.14 0.48 ocorum 1.72 A. puipurescenst o A ences A. philopunctafust 0.29 A. anatoloros 0.21-— A. ligrinus A. mé nta 0.24 A. solitariust 12 0.26 A. dissimilist eastern — . santamartaet A. ruiziiy 0.33 0.96 A. punctatus 0.71 A. vaupesianust A. transversalis A. caquetaet (un ; A. aeneus AS Oh v 2—A, extremus ji i A. mguee richardil griseus roquet A. trinitatis 0.96 A. bonairensis ciae 0.2 sKelemar | ae terod ermus Mi outgroups inderen 9.99 | a aye aa Phenacosaurus [J unassigned 4. tetariiy . D8 4 _nicefori E 4 aequatorialis . orcesit © laevis A. proboscist A. ne fininus Wi latifrons A. carlostoddit - 4 Wi punctatus 0.9 . occultus 0.93 A. equestris i roquet 0,95 A. cuvierl ® tigrinus A. sagrei Seat A. bimacufatus 0.08 W Phenacosaurus Figure 4. Bayesian maximum clade credibility tree inferred with the CombThiele6-mode combined data set. Bayesian posterior probabilities (PP) are shown above branches; asterisks (*) indicate PP = 1.0. Daggers (+) following species names indicate the species for which only morphological data were available. The traditional species groups/series based on morphological characters (see text for details) are differentiated by color. Major Dactyloa subclades described based on molecular data (see text for details) are indicated on the right. The Dactyloa clade is indicated with a black dot on the corresponding node. PHYLOGENY OF THE DacryLoa ¢ Castatieda and de Queiroz 363 TABLE |. RESULTS OF THE WILCOXON SIGNED RANKS (WSR) AND BAYESIAN (B) TESTS OF PHYLOGENETIC HYPOTHESES OF PREVIOUSLY RECOGNIZED PTAXA BASED ON MORPHOLOGICAL CHARACTERS (TOP) AND OF CLADES INFERRED BASED ON MOLECULAR DATA (BOTTOM) ON THE BASIS OF MORPHOLOGICAI DATA ONLY (TORRES-FREQ, THIELE6-MODE). FOR THE TORRES-FREQ DATA SET, DIFFERENCES BETWEEN TREE LENGTHS OF UNCONSTRAINED ANALYSES AND THOSE CONSTRAINED TO CORRESPOND TO EACH TESTED HYPOTHESIS (ATL) AND WSR P-VALUES ARE GIVEN. FOR THE BAYESIAN TESTS OF THE THIELE6- MODE DATA SET, THE PRESENCE (+) OR ABSENCE ( —~) OF THE ALTERNATIVE TOPOLOGY IN THE 95° 0 CREDIBLE SET OF TREES IS SHOWN. SIGNIFICANT RESULTS ARE INDICATED WITH AN ASTERISK (*). Test Hypothesis ATL Traditional groups Dactyloa 6.86 Dactyloa excluding Phenacosaurus* 13.44 aequatorialis series ().99 latifrons series? 1.00 punctatus series 9.32 roquet series 0.84 tigrinus series 2.89 Phenacosaurus* 2.33 Groups based on molecular data Eastern clade 2.58 latifrons clade 8.42 Phenacosaurus clade n/a‘ Phenacosaurus clade not monophyletic (0.44 roquet clade 0.84 Western clade 1.05 Dataset Torres-fre q Thiele6-mode WSR P-Value B ().2S0 —* 0.00 |* =F ().697 = 0.131 =i On 332, __* 0.851 + 0.175 =4 0.629 = 0.468 aa 0.145 a n/a‘ + 0.827 n/a‘ 0.851 + ().969 a * As traditionally circumscribed, which includes (of the species sampled) A. carlostoddi, A. euskalerriari, A. heterodermus, A. inderenae, A. neblininus, A. nicefori, A. oe A. tetarii, and A. vanzolinit. > As traditionally circumscribed, which includes (of the species sampled) microtus, A. princeps, A. purpurescens, and A. squamulatus. A. apollinaris, A. casildae, A. danieli, A. fraseri, A. frenatus, A. insignis, A. latifrons, A. ~ Not applicable: the hypothesis in question was present in the optimal (unconstrained) tree(s), so the alternative hypothesis (monophyly or non-monophyly) was tested instead. common ancestors of the species for which molecular data were a calible (see above). However, A. boettgerit, A. fitchi, A. huilae, and A. podocarpus were consistently placed closer to the western clade than to any of the four other major clades; A. philopunctatust was consistently placed closer to the latifrons clade, and A. proboscist was consistenty placed closer to Phenacosaurus. Additionally, the positions of A. carlostoddit, A. dissi- milist, A. mentat, A. orcesit, A, re A. santamartae?*, A. solitariust, and A. squamu- latust were MeOnsistent ( yarticularly relative to the five major clades) Rewvest parsimony and Bayesian analyses. Tests of Phylogenetic Hypotheses The WSR and Bayesian tests performed on the morphology -only data sets (Torres- freq and Thiele6-mode, respectively) vielded very different results (Table 1): the WSR test failed to reject the monophyly of Dactyloa and each of the subgroups previously described based on morpholog- ical characters: aequatorialis, latifrons, punctatus, roquet, tigrinus, and Phenaco- saurus: in contrast, the Bayesian test rejected the monophyly of Dactyloa and all of the previously recognized subgroups except the roguet series. ‘Of the hypotheses tested, only the monophyly of Dactyloa excluding Phenacosaurus was rejected by both dhe WSR and Bayesian tests. When testing the clades inferred based on molecu- lar data (Castaneda and de Queiroz, 2011) with the morphological data, contradictory results between the parsimony and Bayesian approaches were found again (T Table 1): monophyly of the eastern, latifrons, roquet, and western clades was not rejected with the 364 Bulletin of the Museum of Comparative Zoology, Vol. 160, No. 7 TABLE 2. RESULTS OF THE WILCOXON SIGN RANKS (WSR) AND BAYESIAN (B) TESTS OF PHYLOGENETIC HYPOTHESES OF PREVIOUSLY RECOGNIZED TAXA BASED ON MORPHOLOGICAL CHARACTERS (TOP) AND OF CLADES INFERRED BASED ON MOLECULAR DATA (BOTTOM) ON THE BASIS OF COMBINED MORPHOLOGICAL AND MOLECULAR DATA (COMBTORRES-FREQ, COMBTHIELE6-MODE). FOR THE COMBTORRES-FREQ DATA SET, DIFFERENCES BETWEEN TREE LENGTHS OF UNCONSTRAINED ANALYSES AND THOSE CONSTRAINED TO CORRESPOND TO EACH TESTED HYPOTHESIS (ATL) AND WSR P-VALUES ARE GIVEN. FoR THE BAYESIAN TESTS OF THE COMBTHIELE6-MODE DATA SET, THE PRESENCE (+) OR ABSENCE (—) OF THE ALTERNATIVE TOPOLOGY IN THE 95% OF CREDIBLE SET OF TREES IS SHOWN. SIGNIFICANT RESULTS ARE INDICATED WITH AN ASTERISK (*). Dataset CombTorres-freq CombThiele6-mode Test Hypothesis Adie WSR P-Value B Groups based on morphological data Dactyloa not monophyletic 1.47 O12 a Dactyloa excluding Phenacosaurus* 44.83 ().003* —* aequatorialis series 135.33 <(0.001* —* latifrons series? 83.91 <0.001* =i punctatus series 170.43 <(0.001* = roquet series not monophyletic 35.09 0.016% + tigrinus series 4.82 0.336 + Phenacosaurus group* B15) 0.253 4 Groups based on molecular data Eastern clade not monophyletic 17.63 (0.366 —* latifrons clade not monophyletic 25.56 0.054 = Phenacosaurus clade not monophyletic 59.81 <= () OOF —* roquet clade not monophyletic 25 al 0.077 —* Western clade not monophyletic 9.90 0.687 —* * As traditionally circumscribed, which includes (of the species sampled) A. carlostoddi, A. euskalerriari, A. heterodermus, A. inderenae, A. neblininus, A. nicefori, A. Ges. A. tetarii, and A. vanzolinit. b As traditionally circumscribed, which includes (of the species sampled) A. apollinaris, A. casildae, A. danieli, A. fraseri, A. frenatus, A. insignis, A. latifrons, A. microtus, A. princeps, A. purpurescens, and A. squamulatus. WSR test, but it was rejected—except in the case of the roquet series—with the Bayesian test. The parsimony analysis indicated mono- phyly of the Phenacosaurus clade, but the WSR test failed to reject its non-monophyly; conversely, the Bayesian analysis indicated non- monophyly of the group, but the Bayes- ian test failed to reject its monophyly. With the combined data sets, the WSR and Bayesian tests yielded mostly congruent results concerning taxa recognized previously on the basis of morphological characters (Table 2). The non-monophyly of Dactyloa (given that this hypothesis was inferred in the optimal tree) and the monophyly of the tigrinus series and Phenacosaurus were not rejected by either test. In contrast, both tests rejected the monophyly of the aequatorialis. latifrons, and punctatus series. The non- monophyly of the roquet series (a clade inferred in both parsimony and Bayesian optimal trees), was rejected by the WSR test, but not by the Bayesian test. The monophyly of Dactyloa excluding Phenacosaurus was rejected by both the WSR and Bayesian tests. The clades inferred based on molecular data (Castaneda and de Queiroz, 2011) were also present in the parsimony and Bayesian optimal trees of the combined data sets; therefore, the non-monophyly of these groups was tested. Results obtained with the WSR and Bayesian tests differed in most cases (Table 2): the WSR test rejected the hypoth- esis of non-monophyly of the Phenacosaurus clade and failed to reject the non-monophyly of the eastern, latifrons, roquet, and western clades. In contrast, the Bayesian tests rejected the non-monophyly of all these clades. DISCUSSION The objectives of this study were to reconstruct the phylogeny of the Dactyloa clade based on morphological characters alone and in combination with molecular data, to explore different coding methods for continuous and polymorphic characters that PHYLOGENY OF THE DacTyLoa ¢ Castaneda and de Queiroz 365 were part of our data set, and to test hypotheses of monophyly of previously described taxa. In the following paragraphs, we discuss the advantages and disadk antages of the different coding methods used, the phylogenetic re lationships inferred, and their implications regarding previously recognized taxa. Finally, based on our findings, we propose a new taxonomy. that recognizes only monophyletic taxa and in which names are defined following the rules of PhyloCode. Differences Among Coding Methods For continuous characters, the coding method of Torres-Carvajal (2007) resulted in characters containing the lar gest amount of phylogenetic signal, Tollowed: by Thiele’s (1993) method using 101 character states. The main disadv antage of Torres-Carvajal’s (2007) method is that it results in a significant increase of computation time compared with Thiele’s (1993) method (MRC, personal observation), presumably because it uses step matrices. Although Thiele’s (1993) method discretizes continuous characters, it maintains information on order and magni- tude of change between states. ies its implementation using 101 character states (i.e., allowing a 0. 0] resolution between states) might Te sufficient to approximate a continuous distribution (particularly if the values from the continuous distribution are estimated at a similar level of precision). Despite significant differences in g, values, reciprocal W SR tests (Larson, 1998) indicate that phylogenetic inferences between the two methods (at least for the Dactyloa data set) are not strongly in conflict: tests for differences ReRUCen the optimal trees result- ing from only continuous characters under each of the two coding methods (i.e., Thiele’s [1993] gap- -weighting method eich 101 character states or behes: Carvajal’s [2007] step matrix modification of it) using the data sets produced by each of the coding methods were not statistically significant pe = using the Thiele- coded Acai seus —. 0.3911 using the Torres-coded data set). In contrast, reciprocal tests between the optimal trees resulting from only continuous characters coded with Thiele’s (1993) method using 101] states compare d with using 6 states were statistically significant (P = 0.036 using the data set with 101 states; P = 0.023 using the data set with 6 states). Similarly, reciprocal tests between the optimal trees resulting from only continuous characters coded witli Torres-Carvajal’s method versus Thiele’s (1993) method using six states were also statistically different {p = ().040 using the Torres-coded data set; P = 0.046 using the Thiele-coded data set). These results com- bined with the results based on the statistic indicate that a larger number of character states significantly increases the amount of phylogenetic information record- ed by Thiele’s coding method. Moreover, these findings suggest that Thiele’s (1993) method, when implemented using a large number of character states, may be an effective alternative to fully continuous coding methods. Despite per for ming more poorly according to g; values, it did not yield a significantly Aiftereit tree according to reciprocal tests. At least in this case, the ress of information appears to be small and is compensated by lesser computational re- quirements. The polymorphic characters coded as frequency arrays using Manhattan distance step matrices were found, based on ea values, to contain significantly more phylo- genetic signal than hose coded as standard binary or “anelestate characters and scored using modal conditions. Reciprocal tests indicate that there are significant differenc- es between the optimal trees resulting from data sets including only poly morphic char- acters coded using these two methods (P < 0.001 using the frequency arrays— veal data set: P < (0.001 using the “standard coding with modes data set). This result Farther supports prev ious studies on empir- ical (Wiens, 1995, 1998) and simulated (Wiens and Servedio, 1997) data, showing that methods that incorporate frequency information outperform, based on accuracy measurements, other methods for analyzing polymorphic characters (including scoring 366 Bulletin of the Museum of Comparative Zoology, Vol. 160, No. 7 modal conditions). The advantages of fre- quency methods include making use of more phylogenetic information and reducing the effects of sampling errors, in that the probability of being mislead by the presence or absence of states occurring at low frequen- cies is reduced, which is particularly impor- tant with small sample sizes (Swofford and Berlocher, 1987: Wiens, 1995; Wiens and Servedio, 1997). The main disadvantage of some of the methods incorporating frequency information is that they significantly increase the computation time required because of the use of step matrices (e.g., Wiens, 2000; MRC, personal observation). Phylogeny of Dactyloa This study presents the phylogenetic relationships of Dactyloa based on molecu- lar data for 40 species and morphological data for the same 40 species and 20 additional ones (for a total of 60 species). This represents a substantial improvement upon previous studies, which included a maximum of 42 species (Castafieda and de Queiroz, 2011), 2 of which were not included in the current study (see “Materi- als and Methods”) for molecular data only, or a maximum of 28 species (Poe, 2004) for multiple data sources. For the combined data sets, Dactyloa was inferred to be monophyletic, provided that it includes the Phenacosaurus species, in agreement with previous studies (Poe, 1998, 2004; Jackman et al., 1999: Nicholson et al., 2005; Casta- fieda and de Queiroz, 2011), although its monophyly was not strongly supported according to topology tests employing con- straint trees (the hypothesis of non-mono- phyly was not rejected, though monophyly of Dact yloa excluding Phenacosaurus was rejected). Previous analyses of the entire Anolis clade based on combined data, with a similar set of characters and coding meth- ods, inferred a fully resolved but poorly supported Dact yloa (Poe}: 2004) fig. »2); although the analysis of only the morpho- logical component of this data set did not infer Dactyloa (Poe, 2004, fig. 5). In contrast, previous analyses based solely on molecular data strongly supported the monophyly of Dactyloa (Nicholson et al., 2005; Castafieda and de Queiroz, 2011). The considerable difference in nodal support between the combined (morpho- logical and molecular) and molecular-only analy ses could derive from intrinsic charac- teristics of the morphological characters. For example, the morphological characters might be highly homoplastic, introducing support for conflicting groupings within the morphological data set, or they might have low phylogenetic pforeeoe content, allow- ing multiple placements of taxa for which only morphological data are avail- able. Additionally, it is possible that conflicts between the phylogenetic signal of the morphological and molecular data sets result in a reduction in nodal support. In our analyses, we excluded characters commonly used in studies of morphological conver- gence (e.g., limb and tail length) to avoid this potential bias (but see de Queiroz |1996, 2000] and Poe [2005] for the advantages of including these characters in phylogenetic reconstruction). Trees inferred with the morphology-only data set showed a general lack of support for any particular pe (regardless of the coding method used), which suggests that the morphological data might not have sufficient phylogenetic signal to produce any strong conflict with the molecular data (and therefore strongly affect nodal support). However, reciprocal WSR topology tests comparing the tree inferred from morphological data (Fig. 1) with that inferred based on molecular data (Castaneda and de Queiroz, 2011, fig. 1A) indicated strong disagreement between the two data sets (P < 0.0001 for each case). Therefore, the difference in nodal support between the molecular and combined analyses would seem to result, in this case, from multiple placements of at least some of the species that were scored for morphological charac- ters only, as well as the larger total number of species withia the Dactyloa clade (so that support is distributed among a larger num- ber of nodes). PHYLOGENY OF THE DactyLoa * Castaneda and de Queiroz 367 Castafieda and de Queiroz (2011) in- ferred, based on molecular data, five strongly supported clades (informally named eastern, ee Phenacosaurus, roquet, and west- ern) with coherent geographic distributions. Based on the combined data, we inferred those same five clades with the inclusion of additional species for which only morpho- logical data are available. In agr eement with the inferred relationships of theese species, in all cases, their geographic distributions lie within or on the periphery of the clades in which they were placed. The discussion that follows concerns the composition and inter- nal relationships of the five clades and is based exclusively on the results obtained with the combined analyses. Following Castaneda and de Queiroz (2011), we also adopt (in the following discussion) delimita- tions of the clades based on the species for which molecular data were ails The western clade, as delimited by Castaneda and de Queiroz (2011), included seven species (A. aequatorialis, A. anorien- is, A. chloris, A. festae, A. gemmosus, A. peraccae, and A. ventrimaculatus) distribut- ed in the western and central cordilleras of Colombia, the western slopes of the Ecua- dorian Andes, and the pacific lowlands of Colombia and Ecuador. In this study, the western clade was inferred to include three additional species (considering that we treated A. anoriensis as conspecific with A. eulaemus; see “Materials and Methods”): A antioquiae, A. fasciatus, and A. megalopithe- cus, distributed in the northernmost part of the western cordillera in Colombia (A. antioquiae and A. megalopithecus) and in the Pacific lowlands of central Ecuador (A fasciatus). Two primary subclades were inferred within the western clade. The first includes four species, A. chloris, A. fascia- tus, A. festae, and A. peraccae, all previously placed in the punctatus series (Savage and Guyer, 1989) or species group (Williams, 1976b), that have a humid forest distribu- tion below 1,000 m above sea level and Ae to moderate body size (max SVL = 62 59, and 52 mm, respectively [Williams oe Acosta, 1996]). The second subclade includes six species, A. aequatorialis, A. antioquiae, A. eulaemus, A. gemmosus, A. a a and A. ventrimaculatus, all previously placed in the aequatorialis series (Savage and Guyer, 1989) or species group (Walliams: 1976b, 1985: Williams and Duell- man, 1984: Rueda Almonacid, 1989), dis- tributed from 1,500 to 2,000 m ae sea level and with moderate to large body size (max SVL = 92, 72 [MRC, personal observation], 101, 66, Sl, and SO mm, respectively [Williams and Acosta, 1996]). In both parsimony and Bayesian optimal trees (Figs. 3, 4), Anolis boettger ri, A. fitchi, A. huilae, and A. podocarpus were inferred closer to the western clade than to any of the other five major clades. Castafieda and de Queiroz (2011) also inferred this close relationship for the last three of those species based on molecular data. Anolis boettgeri and A. huilae were previously > included in the punctatus series (Williams, 1976b; Poe et al., 2008), whereas A. fitchi and A. podocarpus were included in the aequatorialis series (Williams, 1976b; Ayala- Varela and Torres-Carvajal, 2010). The geographic distributions of these four spe- cies at mid to high elevations in the eastern slopes of the Andes of Colombia (A. huilae), Ecuador (A. fitchi, A. podocarpus), and Peru (A. boettgeri), do not correspond with the Pacific lowland and Colombian inter- Andean valley distribution of the western clade and suggest that a dispersal or vicariance event was associated with the branch separating the eastern and western species (i.e., the one at the base of the western clade). The latifrons clade, as delimited by Castaneda and de Queiroz (2011), included 12 species (A. agassizi, A. casildae, A. chocorum, A. danieli, A. fraseri, A. frenatus, A. insignis, A. maculigula, A. microtus, A. princeps, A. sp], ae ee sp2) distributed in the Pacific lowlands of Costa Rica, Panama, Colombia (including Malpelo island), and Ecuador and in the Colombian inter- Andean valleys below 1,000 m above sea level (except A. danieli, which ranges from 1700 to 2,200 mi): Most of these species 368 were previously placed in the latifrons species group (Williams, 1976b) or series (Savage and Guyer, 1989), and in all except A. oh aieieices adult males reach a SVL greater than 100 mm (large size was consid- ered a diagnostic feature of the traditional latifrons series {Williams, 1976b], also called the giant mainland anoles |[Dunn, 1937]). We inferred the latifrons clade (excluding A. sp1 and A. sp2. which were not included in this study, see “Materials and Methods”), with the inclusion of three additional species in the parsimony analysis, A. apollinaris, A. latifrons, and A. purpurescens, and further including A. squamulatus in the Bayesian analysis. All four potentially additional species were previously included in the latifrons series or species group (Williams, 1976b; Savage and Guyer, 1989) and have Pacific low aan A. latifrons and A. purpurescens) or inter ree (A. apollinaris) distributions, except A. squamulatus (see below). In A. apollinaris, A. latifrons, and A. squamulatus adult male maximum SVL ex- ceeds 100 mm (106, 133, and 122 mm, respectively; Williams and Acosta, 1996; Ugueto et al., 2009): A. purpurescens, which is noe from a small number of specimens, appears to be smaller (max SVL = 78 mm: Williams and Acosta, 1996). In both analyses, A. philopunctatus is inferred as the sister group of the latifrons clade. However, this species was not considered as part of the latifrons clade based on the delimitation of the clade using species for which molecular data were ay Ailible (see above). This species was previously included in the punctatus series and_ is distributed in the Brazilian Amazon, and its adult maximum SVL = 73 mm (Rodrigues, 1988). Anolis squamulatus was previously placed in the latifrons species group of Williams (1976b) and series of Savage and Guyer (1989) based on its large dew lap. small head scales, uniform dorsal Pe and large body size (max adult male SVL > 100 mm: Williams and Acosta, 1996). Although inferred as part of the latifrons clade in the Bayesian analysis (Fig. 4), it was not in the par simony analysis (Fig, 3). Moreover, A. squamulatus is the only species, of the sampled taxa, previ- ously placed in the latifrons species group or Bulletin of the Museum of Comparative Zoology, Vol. 160, No. 7 series that was not inferred as part of the latifrons clade in the parsimony analysis. However, the geographic distribution of A. squamulatus, in the cloud forests of the northern Venezuelan Andes, does not corre- spond to the Pacific lowland and Colombian inter-Andean valley distribution of the latifrons clade or to the Pacific mid and low elevations in Colombia and Ecuador distribution of the western clade (to which it was inferred as bein closer in the parsimony analysis). Instead, it corresponds more closely to the geographic distribution of the eastern clade. Either of these alternative relationships of A. squamula- tus (within or closer to either the western or the eastern clades) would require the conver- gent evolution of large body size with members of the latifrons clade. In agreement with previous studies suggesting ihe close relation- ship between A. frenatus, A. latifrons, and A. princeps and including the possibility that those three taxa represent a single species (Savage and Talbot, 1975; Williams, 1988: Castafieda and de Queiroz, 2011), we inferred the clade ((A. frenatus (A. latifrons, princeps)) in both parsimony and Bayesian analyses. The eastern clade, as delimited by Castafieda and de Queiroz (2011), included five species (A. anatoloros, A. jacare, A. punctatus, A. transversalis, and A. tigrinus ) distributed in the northern portion nes the eastern cordillera of Colombia into the Venezuelan Andes and the Amazon region. In our parsimony results, the eastern clade was inferred to include three additional species: A. carlostoddi, A. orcesi, and A. vaupesianus, whereas in the Bayesian anal- ysis, it was inferred to include six additional species: A: dissimilis, A. menta, A. ruizii, A: santamartae, A. solitarius, and A. vaupesia- nus. All of the potential additional species have an eastern Andean and Amazonian distribution. They occur in Amazonia (A. vaupesianus, A. dissimilis), the eastern slopes of the northern Andes of Ecuador (A. orcesi), the eastern cordillera of Colom- bia (A. ruizii), the Sierra Nevada de Santa Marta in Galea (A. menta, A. santamar- tae, A. solitarius), and the Chimanta tepui in PHYLOGENY OF THE DacryLoa * Castatieda and de Queiroz 369 Venezuela (A. carlostoddi). Within the eastern clade, two subclades were inferred in both analyses: the first includes si species, A. anatoloros, A. jacare, and . tigrinus, with mostly Andean, high-ele a ar dicuabetions and smaller body size (max male SVL °=" 55-68 mm: Williams and Acosta, 1996; Ugueto et al., 2007); the second subclade includes three species, A. punctatus, A. transversalis, and A vaupe- sianus, with Amazonian, low-elevation dis- tributions, and larger size (max male SVL = 76-82 mm: Williams and Acosta, 1996). It is not surprising that A. vaupesianus (distrib- uted in the Vaupes and Amazonas depart- ments in Colombia and known only ae the type series) was consistently inferred as the sister taxon of A. punctatus (with a broad Amazonian distribution). These two species were described as close relatives that differ primarily in the size and degree of keeling of the ventral scales (smaller ‘and weakly keeled in A. vaupesianus versus larger and strongly keeled in A. punctatus), the dewlap coloration in preservative (black skin with white scales in A. vaupesianus versus light skin with small dark spots and purplish scales in A. punctatus) and the dorsal color pattern of preserved specimens (in A. vaupesianus, “brown, strongly blotched with darker, dorsal blotches tending to form transverse series across the back” versus an unpatterned dark dorsum in A. punctatus) (Williams, 1982: 8). The unique color pattern of A. vaupesianus is only found in one of the paratypes (UTA 6850), whereas the colora- tion of the other specimens in the type series is not particularly different from that of A. punctatus (Williams, 1982: 8-9). Given the small differences separating these two spe- cies, a more comprehensive sampling of A. punctatus in Colombia (currently <10 spec- imens are known) and the collection of additional molecular data (particularly for A. vaupesianus) could clarify whether these two taxa are conspecific as well as whether the characters used to distinguish them represent extremes imva eodunuans distribu- tion or are based on an atypical specimen. In contrast to the likely close relationship between A. punctatus and A baupestanus, the relationships of A. carlostoddi and A orcesi (parsimony) or A. dissimilis, A. menta, A. santamartae, and A. solitarius (Bayesian) to A. tigrinus are less clear. The former two species were previously placed in Phenaco- and thus not considered closely saurus related to A. tigrinus. In contrast, two of the latter species, A. menta and A. solitarius, were placed in the tigrinus series (Williams, 1976b; Ayala et al., 1984) (the other two, A. dissimilis and A. santamartae, were placed in the punctatus series). In both cases, the putative clade formed by all three or all six species has low support, and given the low resolving power of the morphological data set and the Tact that molecular data are available for neither A. carlostoddi and A. orcesi nor A. dissimilis, menta, A. santamartae, A. ruizii, and A. solitarius, the inferred rela- tionships are questionable. The roquet clade, as delimited by Casta- fieda and de Queiroz (2011), included eight Species: WAL eaeneus) 4A: bonairensis, "A. extremus, A. griseus, A. luciae, A. richardii, A. roquet, and A. trinitatis), distributed in the southern Lesser Antilles, from Martini- que to Grenada, as well as the islands of Bonaire and Tobago (with introduced pop- ulations in Trinidad and Guyana {Gorman and Dessauer, 1965, 1966; Gorman et al., 1971]). This clade corresponds to the previously described roquet species group or series (Underwood, 1959: Gorman and Atkins, 1967, 1969: Lazell, 1972: Williams, 1976a; Savage and Guyer, 1989; Creer et al., 2001). One additional species, A. blanquilla- nus, from the island of La Blanquilla, was previously referred to the roquet series (Williams, 1976a) but was not included in our analyses; however, its inclusion in the roquet clade is supported by allozyme data (Yang et al., 1974; Creer et al., 2001). Poe (2004) imfonted the roquet clade (BS = 74%, fig. 2), supported by six morphological characters (see “Current Taxonomy within Dactyloa”); two of those, greater sexual size dimorphism and an increase in the number oS postmental scales, were also inferred < synapomorphies for the clade in this ait 370 The Phenacosaurus clade, as delimited by Castafieda and de Queiroz (2011), included five species (A. euskalerriari, A. heteroder- mus, A. inderenae, A. nicefori, and A. vanzolinii) distributed in the Andean re- ions of Colombia, Ecuador, and Venezuela, all of which were previously placed in the genus Phenacosaurus (Lazell, 1969; Barros et al., 1996). In this study, the Phenaco- saurus clade was inferred with the addition of A. tetarii, a species that was previously placed in the genus Phenacosaurus (Barros et al., 1996) and whose geographic distribu- tion (Venezuelan Andes) conforms to that of the clade. Three other species that were previously referred to Phenacosaurus, A. carlostoddi, A. orcesi, and A. pee were not inferred to be part of this clade in the parsimony analysis. However, in the Bayesian analysis, A. orcesi was inferred as the sister group of the Phenacosaurus clade with moderate support Cae SMe Fig. 4). In contrast, in the parsimony analysis, A. orcesi was placed in the eastern clade (as was A. carlostoddi), although with weak support (BS = 6%, 24%, and 19%). In both parsimony and Bayesian analyses, A. pro- boscis was inferred as sister group of the Phenacosaurus clade (or of the Phenaco- saurs clade plus A. orcesi) with weak support (BS =~ 58%,.-PP =. 0:43)) a see OnSeP also eed by Poe (2004, fig. 2). A close relationship between A. probos- cis and species traditionally referred to Phenacosaurus was also inferred by Poe et al. (2009b, 2012). The geographic distribu- tions of both species lie on the periphery of that of the Phenacosaurus clade: A. orcesi is distributed along the eastern slopes of the northern Andes of Ecuador, whereas A. proboscis is distributed along the western slopes of the northern Andes of Ecuador. The more deeply nested species within the Phenacosaurus clade (A. heterodermus, A. inderenae, A. tetarii, and A. vanzolinii) are larger in size (max male SVL=76, 98, 86, and 104 mm, respectively [Williams and Acosta, 1996]) and correspond to the heterodermus group of Williams et al. (1996); the two earlier diverging lineages Bulletin of the Museum of Comparative Zoology, Vol. 160, No. 7 (A. euskalerriari and A. nicefori) are smaller in size (max male SVL = 53 and 63 mm, respectively [Williams and Acosta, 1996]). Except for A. euskalerriari, all of the species in the Phenacosaurus dace have heteroge- neous dorsal scales, suggesting that this condition originated in the ancestral lineage of A. heterodermus and A. nicefori after it diverged from that of A. euskalerriari. Anolis carlostoddi, A. orcesi, and A. nebli- ninus, which were previously placed in the genus Phenacosaurus but were not inferred in this study to be part of the Phenacosaurus clade (although A. orcesi was inferred to be closely related in the Bayesian analysis), also lack heterogeneous dorsal scales. Seventeen species (A. boettgeri, A. cali- mae, A. caquetae, A. carlostoddi, A. dissim- ilis, A. jfitchi, A! huilaey A menten neblininus, A. orcesi, A. philopunctatus, A. podocarpus, A. proboscis, A. ruizti, A. santamartae, A. solitarius, and A. squamula- tus) were not consistently placed in any of the five mutually exclusive clades just discussed, either because their positions did not satisfy the criterion based on the last common ancestor of the species for which molecular data were available or because they differed across phylogenetic methods (and were commonly poorly supported). However, 6 of the 17 species (A. boettgeri, A. fitchi, A. huilae, A. philopunctatus, A. podocarpus, and A. proboscis) were each consistently placed closer to one of the five clades than to the others. For those species whose relationships differed among analyses (A. calimae, A. caquetae, A. carlostoddi, A dissimilis, A. menta, A. neblininus, A. orcesi, A. ruizii, A. santamartae, A. solitarius, and A. squamulatus), 9 out of 11 of which currently lack molecular data (all but A. calimae and A. neblininus), more data will be necessary to clarify their relationships within Dactyloa. Previously Recognized Taxa None of the traditionally recognized sub- groups of Dactyloa based on morphological characters (aequatorialis, latifrons, Phenaco- sdurus, punctatus, roquet, and tigrinus ) were PHYLOGENY OF THE DacryLoa * Castatieda and de Queiroz 37] inferred in the optimal trees inferred from either the morphology-only or the combined data sets except the roquet series in the combined analyses. The parsimony-based topology tests (WSR) using the morpholo- gy-only data set failed to reject the mono- phy ly of Dac tyloa or any of the previously described subgroups (Table 1); therefore, despite the morphological data not support- ing any of these groups when analyzed under parsimony, it is also unable to reject any of them using parsimony-based tests. In con- trast, the Bayesian tests using the morphol- ogy-only data set rejected the hypotheses of monophyly of Dactyloa and all traditionally recognized series except the roquet series. With the combined data sets and both parsimony and Bayesian tests (2), the aequa- torialis, latifrons, and punctatus series were also strongly rejected, but the tigrinus series and Phenacosaurus were not rejected. Contradictory evidence and absence of support for the series described based on morphological characters is consistent with previous suggestions that morphological characters tieed for series delimitation might show a high degree of convergence and parallelism (Williams, 1976b: 260) and that some series were described only for conve- nience (Williams, 1979: 10). Furthermore, traditional series delimitation did not dis- tinguish clearly between ancestral and derived conditions, and the former are not indicative of close phylogenetic relation- ships. Differences between the results of WSR and Bayesian tests might reflect the conservativeness of the WSR—the require- ment of a stronger signal to reject a given hypothesis (Lee, 200) =hecense Bayesian tests often rejected hypotheses when WSR tests did not, but WSR tests rarely rejected hypotheses not rejected by Bayesian tests. Differences between the data sets (i.e., resulting from alternative coding mietiods do not appear to be the reason for the different results between the two tests. For one thing, the Bayesian tests rejected more of the hypotheses despite using the data set (Thiele6-mode) that contained the least phylogenetic signal. For another, WSR tests performed on the Thiele6-mode data set (the data set used for the Bayesian tests) yielded the same Net a results as with the Torres- freq data set (results not shown). Proposed Taxonomy Our results indicate that a revised taxon- omy for Dactyloa is warranted. Optimal phylogenetic trees and topology tests indi- cate that most of the previously recognized taxa within Dactyloa based on morphological characters and traditionally ranked as species groups or series are not monophyletic. Moreover, our previously published results based on molecular data (Castaieda and de Queiroz, 2011) indicate the existence of five well-supported major subclades, and the results ‘ the combined analyses of morpho- logical and molecular data in the present study both corroborate the monophyly and clarify the composition of those subclades. Here, we propose a revised taxonomy based on the results of our el a analyses, including names that are defined explicitly in terms of phylogenetic relationships (de Queiroz and Gauthier, 1990, 1992). The optimal topologies obtained from the different (parsimony versus Bayesian) meth- ods are in substantial but not complete agreement. We considered it inappropriate to select one topology over the other because each ane has advantages and disadvantages: The parsimony tree is based on a character coding method that incorpo- rates more phylogenetic information, whereas the Bayesian tree is based on more realistic evolutionary models. Therefore, we used a consensus tree as the basis for our proposed taxonomy (Fig. 5). Specifically, we used a pruned and regrafted* consensus tree (Gordon, 1980; Finden and Gordon, 1985: Bryant, 2003) derived from the most parsimonious tree (Fig. 3) and the maxi- mum clade credibility tree (Fig. 4) for the combined morphological and molecular >The term “grafted” seems more appropriate here, given that the aprach has not been grafted before; however, we use “regrafted” becinte it has been commonly used in the literature. ww ~l bo Dactyloa Phenacosauru Bulletin of the Museum of Comparative Zoology, Vol. 160, No. 7 A. aequatorialis A. eulaemus -——_— A. gemmosus A. plblaieh belor Al A. megal +— A. festae A. peraccae -— A. chloris A. fasciatust . boettgeri* ep . podocarpus $$ m$m$ maa . agassizi . insignis . microtus . casildae . maculigula . apollinaris* . frenatus . latifronst : ea A. fraseri A. chocorum A. purpurescens* een AA, Canieli ee AV Ohilopunctatust ee en SEN UNS ES A. vaupesianust A. punctatus A. transversalis A. tigrinus A. jacare A. anatoloros b>b Megaloa PDSEEEDEDED _—™ $$ A. santamartaet oer es ay dissimilist L faa A. solitarius* A. mentat . aeneus . extremus . roquet _ . richardit . griseus . trinitatis | . bonairensis . luciae } . euskalerriari . heterodermus . inderenae . vanzolinii . tetarif; . hicefori . proboscist DELDDDDSDDDSESSEBED ces AA Orcesit opithecus+ A. ventrimaculatus aequatorialis series latifrons series punctatus series roquet series heterodermus series incertae sedis Figure 5. Pruned and regrafted consensus tree based on the most parsimonious tree (inferred with the CombTorres-freq data set; Fig. 3) and the Bayesian Maximum Clade Credibility tree (inferred with the CombThiele6-mode data set; Fig. 4) with the taxonomy proposed in this study. Daggers (+) following species names indicate the species for which only morphological data were available. Regrafted branches are indicated by a break near their bases. See text for details concerning alternative prunings and regraftings 1) within the clade composed of A. chloris, A. fasciatus, A. festae, and A. peraccae; 2) of A. purpurescens versus A. daniel’; and 3) of members of the punctatus series. The Dactyloa clade is indicated with a black dot on the corresponding node. Five mutually exclusive, informally named clades (“series”) within Dactyloa are distinguished by color, with lighter versions of the PHYLOGENY OF THE DacTYLoa ¢ Castaneda and de Queiroz oT data sets (i.e, CombTorres-freq and CombThiele6-mode matrices). Because 2 few of the species (e.g., A. carlostoddi, A. orcesi, A. squamulatus) have very een relationships on the primary trees, the strict and semistrict consensus trees had little resolution, and we therefore originally intended to use an Adams consen- sus tree (for a review of consensus methods, see Swofford |1991]). However, the Adams consensus tree contained unexpected groups that we felt could not be justified on the basis of the primary trees (e.g., A. squamulatus closer to the western clade than to the latifrons clade, which is contradicted by the Bayesian tree; Fig. 4), and it turned out that the pruned aide regratted consensus tree exhibited the desied properties we had incorrectly attributed to the Adams consen- sus method (i.e., placement of species with conflicting relationships within the smallest possible Cede in the consensus tree for which there is complete agreement concern- ing their higher level relationships in the primary trees). To generate the pruned and _ regratted consensus tree, we first produced agree- ment subtrees (also called common pruned trees; Finden and Gordon, 1985) using PAUP* with identical topologies that ventult from removing (pruning) the same set of taxa from the primary trees (Finden and Gordon, 1985). We obtained 12 largest agreement subtrees for the 60 Dactyloa species (including only one outgroup spe- cies, Polychrus marmoratus, to root the trees), which differed only in the inclusion of all possible combinations of two species from the clade composed of A. festae, A. chloris, A. fasciatus, and A. peraccae (six possible combinations) and the inclusion of either A. danieli or A. purpurescens. We arbitrarily selected one of the 12 largest agreement subtrees (the one including A. fasciatus, A. peraccae, and A. purpurescens) <_ Sw) as the base tree for the regrafting process. In the second step, we manually reattached each previously pruned species or set of species to the node representing the most recent common ancestor of its alternative placements on the primary trees. Because the position of the entire « ‘astern clade and its possible relatives differed between the two primary trees (closer to the latifrons clade in the parsimony tree versus closer to the roquet clade in the Bayesian tree), all of those species were excluded (oni the largest agreement subtrees. To determine whether those species should be reattached singly or in sets, we determined the largest agreement subtree for those species (A. anatoloros, A. caquetae, A. carlostoddi, A. dissimilis, A. jacare, A. menta, A. orcesi, A. punctatus, A. ruizii, A. santamartae, A. solitarius, A. tigrinus, A. transversalis, A. vaupesianus), plus one representative each of the western, latiforns, Phenacosaurus, and roquet clades; we also included Poly- chrus marmoratus to root the trees. We obtained four largest agreement subtrees representing two ditforent topologies for the set of species making up the eastern clade and its possible relatives (the other two topologies differed only in whether the representative of the western or the roquet clade was included). Both topologies in- cluded (A. transversalis (A. punctatus, A. vaupesianus)), but one included (A. tigrinus (A. anatoloros, A. jacare)) and A. caquetae, whereas the other included ((A. dissimilis, A. santamartae) (A. menta, A. solitarius)). We selected the former for grafting because those groups were consistent, Aiea yruning, with ehe primary trees, whereas the lager depended on pruning the representative of the latifrons clade, which we intend to be a fixed point of reference (i.e., not a candidate a pruning) in this secondary analysis (given that several members of the latifr “ons chide were present in all of the primary same hue indicating tentative assignment to the clade represented by that hue. These informally named clades have the same name as some of the groups traditionally ranked as series within Anolis; however, speciés composition is not necessarily identical. A black bar across a branch indicates an apomorphy used to define the clade name above the bar. 374 agreement subtrees). The resulting pruned and regrafted consensus tree (Fig. 5) serves as the basis for our taxonomy. In all but two cases, we have selected preexisting names that have been applied traditionally to groups of species approxi- mating, to one degree or another, the clades to which we apply them. Three of the selected names are similar in appearance to genus names; however, they are here tied to élades rather than to the rank of genus and are implicitly ranked below the genus level (given that we use the name Anolis in the binomina of all of the species included in the named clades). Five of the selected names combine the species name (epithet) of the first described species (e.g., “roquet’) with the name of a rank (i.e., “series”). Because those names violate the rule stating that clade names must be single words beginning with a capital letter (ICPN, Article 17.1), they are treated here as informal names. They are nevertheless given explicit phylogenetic definitions as guides for applying the names in the context of future phylogenetic hypotheses. Despite being given explicit phylogenetic defini- tions, the names in question are compatible with traditional nomenclature, in that they have been defined so as to ensure that they will always refer to mutually exclusive taxa (see de Queiroz and Donoghue, 2013), as would names associated with the rank of series under traditional nomenclature. As a consequence, the “series” names are ap- plied to clades that are more inclusive than the five clades based on the species for which molecular data were available that formed the basis of our discussion in the “Phylogeny of Dactyloa” section (above), though those five less inclusive clades form the cores of the “series” clades. The number of Dactyloa subclades named in our taxono- my exceeds five, the number of well- supported clades based on molecular data (Castafieda and de Queiroz, 2011) and corroborated in this study, because in two cases, we considered it useful to name additional clades associated with the origins of distinctive apomorphies. Bulletin of the Museum of Comparative Zoology, Vol. 160, No. 7 Dactyloa Wagler 1830, converted clade name Definition (branch-modified node- based): The crown clade originating in the most recent common ancestor of Anolis punctatus Daudin 1802 and all extant species that share a more recent common ancestor with A. punctatus than with A bimaculatus (Sparrman 1784), A. cuvieri Merrem 1820, A. equestris Merrem 1820, A. occultus Williams and Rivero 1965, and A. sagrei Duméril and Bibron 1837. Refer- ence phylogeny: Figure 5, this study (see also Poe, 2004, figs. 2-4). Inferred com- position: Anolis aeneus Gray 1840, A. aequatorialis Werner 1894, A. agassizi Stejneger 1900, A. anatoloros Ugueto, Rivas, Barros, Sanne Pacheco and Gar- cfa-Pérez 2007, A. anchicayae Poe, Velasco, Miyata and Williams 2009 (inclusion based on inferred close relationship to A. peraccae following Poe et al., 2009b), A. anoriensis Velasco, Gutiérrez-Cardenas and Quintero- Angel 2010 (inclusion based on inferred close relationship to A. aequatorialis follow- ing Castafieda and de Queiroz, 2011), A. antioquiae Williams 1985, A. apollinaris Boulenger 1919, A. bellipeniculus (Myers and Donnelly 1996) (inclusion based on inferred close relationship to A. neblininus following Myers and Donnelly, 1996), A. Blanton Hummelinck 1940 (inclusion based on inferred close relationship to A. bonairensis following Yang et al., 1974), A. boettgeri Boulenger 1911, A. bonairensis Reche en 1923, A. iealnae Ayala, Harris and Williams 1983, A. caquetae W ieee 1974, A. carlostoddi (Williams, Praderio and Gorzula 1996), A. casildae Arosemena, Ibanez, and de Sousa 1991, A. chloris Boulenger 1898, A. chocorum Williams and Duellman 1967, A. cuscoensis Poe, Yanez-Miranda and Lehr 2008 (inclusion based on the results of Poe et al., 2008), A. danieli Williams 1988, A. deltae Williams 1974 (inclusion based on inferred close relationship to A. dissimilis following Wil- liams, 1974), A. dissimilis Williams 1965, A. eulaemus Bae 1908, A. euskalerriari (Barros, Williams, and Viloria 1996), A. PHYLOGENY OF THE DacTyLoa ¢ Castaneda and de Queiroz O10 extremus Garman 1887, A. fasciatus Bou- lenger 1885, A. festae Peracca 1904, A. fitchi Williams and ‘Duellman 1984, A. fraseri Giinther 1859, A. frenatus Cope 1899, A. gemmosus O'Shaughnessy 1875, A. gorgo- nae Barbour 1905 (inclusion based on inferred close relationship to A. andianus [a synonym of A. gemmosus according to Williams and ioweliman (1984) | following Barbour, 1905), A. griseus Garman 1887, A. heterodermus Duméril 1851, A. huilae Williams 1982, A. ibanezi Poe, Latella, Ryan and Schaad 2009 (inclusion based on inferred close relationship to A. chocorum following Poe et al., 2009a), A. inderenae (Rueda and Hernandez-Camacho 1988), A. insignis Cope 1871, A. jacare Boulenger 1903. A. kunayalae Hulebak, Poe, Ibanez and Williams 2007 (inclusion based on inferred close relationship to A. mirus following Hulebak et al., 2007), A. laevis (Cope 1876) (inclusion based on inferred close relationship to A. proboscis following Williams, 1979), A. lamari Williams 1992 (inclusion based on inferred close relation- ship to A. tigrinus following Williams, 1992), A. latifrons Berthold 1846, A. luciae Garman 1887, A. maculigula Williams 1984, A. megalopithecus Rueda Almonacid 1989, A. menta Ayala, Harris and Williams 1984, A. microtus Cope 1871, A. mirus Williams 1963 (inclusion based on inferred close relationship to A. fraseri following Williams, 1963), A. nasofrontalis Amaral 1933 (inclu- sion based on inferred close relationship to A. tigrinus following Amaral, 1933), A. neblininus (Myers, Williams and McDiar- mid 1993), A. nicefori (Dunn 1944), A. nigrolineatus Williams 1965, A. orcesi (La- zell 1969), A. otongae Ayala- Varela and Velasco 2010 (inclusion based on inferred close relationship to A. gemmosus followin Ayala-Varela and Welnede, 2010), A. para- vertebralis Bernal Carlo and Roze 2005 (inclusion based on inferred close relation- ship to A. solitarius following Bernal Carlo and Roze, 2005), A. parilis Williams 1975 (inclusion based on inferred close relation- ship to A. mirus following Williams, 1975), A. peraccae Boulenger 1898, A. philopunc- tatus Rodrigues 1988, A. phyllorhinus Myers and Carvalho 1945 (inclusion based on inferred close relationship to A. puncta- tus following Myers and Carvalho, 1945), A. podocarpus “Ayala- Varela and Torres-Carva- jal 2010, A. princeps Boulenger 1902, A. proboscl Peters and Orcés 1956, A. pro- pinquus Williams 1984 (inclusion based on inferred close relationship to A. apollinaris following Williams, 1988), A. pseudotigrinus Amaral 1933 (inclusion based on inferred close relationship to A. tigrinus following Amaral, 1933), A. punctatus Daudin 1802, A. purpurescens Cope 1899, A. richardii Duméril and Bibron 1837, A. roquet (Bon- naterre 1789), A. rwizii Rueda and Williams 1986, A. santamartae Williams 1982, A. soinii Poe and Yanez-Miranda 2008 (inclu- sion based on inferred close relationship to A. transversalis following Poe and Yanez- Miranda, 2008), A. ROliarils Ruthven 1916, A.squamulatus Peters 1863, A. tetarii (Bar- ros, Williams, and Viloria 1996), A. tigrinus Peters 1863, A. transversalis Duméril 1851, A. trinitatis Reinhardt and Liitken 1862, A. umbrivagus Bernal Carlo and Roze 2005 (inclusion based on inferred close relation- ship to A. solitarius following Bernal Carlo ae Roze, 2005), A. vanz velai (Williams, Orcés, Matheus, and Bleiweiss 1996), A. vaupesianus Williams 1982, A. ventrimacu- latus Boulenger 1911, and A. williamsmit- termeierorum Poe and Yanez-Miranda 2007 (inclusion based on inferred close relation- ship to A. orcesi following Poe and Yafiez- Miranda, 2007). Comments: The name Dactyloa was previously used by Savage and Guyer (1989) for a taxon ranked as a genus containing all species referred to the Dactyloa clade in this study except those species formerly assigned to the genus Phenacosaurus Sale 1920 (A. bellipeni- culus, A. carlostoddi, aaa A. heterodermus, A. ie renae, A. neblininus, A. nicefori, A. orcesi, A. ic A. vanzoli- nii). Here, the name Dactyloa is not associated with the rank of genus (it is implicitly associated with a lancer rank) so that the binomina of the included species retain the prenomen (genus name) Anolis. 376 The following named clades are subclades of Dactyloa. aequatorialis series Savage and Guyer 1989, informal clade name Definition (branch-modified node- based): The crown clade originating in the most recent common ancestor of Anolis aequatorialis Werner 1894 and all extant species that share a more recent common ancestor with A. aequatorialis than with A. latifrons Berthold 1846, A. punctatus Dau- din 1802, A. roquet (Bonnaterre 1789), and A. heterodermus Duméril 1851. Reference phylogeny: Figure 5, this study. Inferred composition: Anolis aequatorialis Werner 1894, A. anoriensis Velasco, Gutiérrez- Cardenas and Quintero-Angel 2010 (see “Comments”), A. antioquiae Williams 1985, A. boettgeri Boulenger 1911 (see “Comments”), A. chloris Boulenger 1898, A. eulaemus Boulenger 1908, A. fasciatus Boulenger 1885, A. festae Peracca 1904, A. fitchi Williams and Duellman 1984 (see “Comments” ), A. gemmosus O'Shaughnessy 1875, A. huilae Williams 1982 (see “Com- ments”), A. megalopithecus Rueda Almona- cid 1989, A. peraccae Boulenger 1898, A. podocarpus Ayala-Varela and Torres-Carva- jal 2010 (see “Comments”), and A. ventri- maculatus Boulenger 1911. Other species that may belong to the aequatorialis series are A. anchicayae Poe, Velasco, Miyata and Williams 2009, A. mirus Williams 1963, A. otongae Ayala-Varela and Velasco 2010, and A. parilis Williams 1975 (see “Comments”). Comments: Referral of A. anoriensis to this clade is based on the results of Castafieda and de Queiroz (2011). In the context of the reference phylogeny, six species not traditionally eS: in the aequatorialis series (e.g., Savage and Guyer, 1989) are included in the aequatorialis series as conceptualized here: A. boettgeri, A. chloris, A. fasciatus, A. festae, A. huilae, and A. peraccae. There is strong evidence Leg le inclusion of A. chloris, A. festae, and A. peraccae (Castafieda and de Queiroz, 2011), and A. fasciatus is placed consistently in a subclade with those three Bulletin of the Museum of Comparative Zoology, Vol. 160, No. 7 species (Figs. 3, 4). In contrast, because of inconsistent placement of A. huilae be- tween analyses (Fig. 3 versus Fig. 4), be- cause of weak support for the inclusion of both A. boettgeri and A. huilae and because molecular data are currently lacking for A. boettgeri, inclusion of those two species in the aequatorialis series should be consid- ered tentative. Additionally, A. maculigula, a species included in the traditional cir- cumscription of the aequatorialis series, is excluded based on strong evidence sup- porting its inclusion in the latifrons series (see below). The inclusion of A. fitchi and A. podocarpus, traditionally included in the aequatorialis series (Savage and Guyer, 1989), should also be considered tentative given the weak support for the relevant relationships (Figs. 3, 4). Although the parsimony analysis (Fig. 3) places A. squa- mulatus in the aequatorialis series, we have tentatively retained that species in the latifrons series based on the results of the Bayesian analysis (Fig. 4) and its large body size (see “Comments” on the latifrons series). The inclusion of A. mirus, A. otongae, and A. parilis, previously consid- ered members of the aequatorialis series (Williams, 1975; Ayala-Varela and Velasco, 2010), should also be considered tentative iven the current absence of these species a4 explicit phylogenetic analyses. Anolis anchicayae, a oaed as Closely related to A. peraccae in a recent phylogenetic analysis (Poe et al., 2009b), should also be consid- ered tentatively included in the aequator- ialis series given that in that analysis these two species were not inferred as close relatives of A. aequatorialis. The aequatorialis series as conceptualized here corresponds approximately to the western clade of Castafieda and de Queiroz (2011). However, the aequatorialis series is more inclusive than the western clade of Castafieda and de Queiroz (2011) in that it appears to include A. boettgeri, A. fitchi, A. huilae, and A. podocarpus and might also include some species currently considered incertae sedis within Dactyloa or absent from explicit phylogenetic analyses if they 9 PHYLOGENY OF THE DacryLoa ¢ Castarieda and de Queiroz Ort are found to be more closely related to A. aequatorialis than to A. latifrons, A. punc- tatus, A. roquet, and A. heterodermus. Additionally, species in the western clade of Castafieda and de Queiroz (2011) are characterized by having a cohesive western Andean geographic distabunen: but A. boettgeri, A. fitchi, A. huilae, and A. nodocarpus, and possibly other species in the more inclusive aequatorialis series, do not conform to this geogr aphic pattern (see “Phylogeny of Dactyloa” above for more details about the geogr aphic distributions of the four species mentored)! latifrons series Gorman and Dessauer 1966, informal clade name Definition (branch-modified node- based): The crown clade originating in the most recent common ancestor of Anolis latifrons Berthold 1846 and all extant species that share a more recent common ancestor with A. latifrons than with Anolis aequatorialis Werner 1894, A. punctatus Daudin 1802, A. roquet (Bonnaterre 1789), and A. heterodermus Duméril 1851. Refer- ence phylogeny: Figure 5, this study. Inferred composition: Anolis agassizi Stejneger 1900, A. apollinaris Boulenger 1919, A. casildae Arosemena, Ibanez, and de Sousa 1991, A. chocorum Williams and Duellman 1967, A. danieli Williams 1988, A. fraseri Giinther 1859, A. frenatus Cope 1899, A. insignis Cope 1871, A. fia ile Hulebak, Boe Ibanez and Williams 2007, A. latifrons Berthold 1846, A. maculigula Williams 1984, A. microtus Cope 1871, A. princeps Boulenger 1902, A. philopunctatus Rodrigues 1988 (see “Comments”), A. purpurescens Cope 1899, and A. squamula- tus Peters 1863 (see “Comments’). Other species that may belong to the latifrons series are A. ibanezi Poe, SLatclia Ryan and Schaad 2009, and A. propinquus Williams 1984 (see “Comments’). Comments: Re- ferral of A. kunayalae to this clade is based on the results of Nicholson et al. (2005, fig. 1. wivere —A: kunayalae corresponds to Nicholson et al.’s “New Species 1” [Hulebak et al., 2007]). In the context of the reference phylogeny, four species not traditionally included in the latifrons series (e.g., Savage and Guyer, 1989) are included in the latifrons series as conceptualized here: A. agassizi, A. chocorum, A. maculigula, and A. philopunctatus. There is strong evidence supporting inclusion of the former three species (Castafieda and de Queiroz, 2011). In contrast, because of weak support for the relationship of A. philopunctatus (Figs. 3, 4) and because of a current lack of molecular data, its inclusion in the latifrons series should be considered tentative. Similarly, the inclusion of A. squamulatus, tradition- ally included in the latifrons series (e.g., Savage and Guyer, 1989) should be consid- ered tentative given the weak and inconsis- tent support oe the relevant relationships (Fig. 3 versus Fig. 4) and the current lack of See data. We have tentatively includ- ed A. squamulatus in the latifrons series, where it was placed in the Bayesian tree (Fig. 4), rather than in the aequatorialis series, where is was placed in the parsimony tree (Fig. 3) or in the punctatus series, with which it agrees best in terms of geographic distribution (see “Phylogeny of Dactyloa, above) because it shares the derived char- acter of large body size with members of the latifrons series. Anolis propinquus has traditionally been considered part of the latifrons series (Williams, 1988); however, the inclusion of this species slioala be considered tentative given its current ab- sence from explicit phylogenetic analyses. Inclusion of Anolis ibanezi, inferred as closely related to A. chocorum in a recent phylogenetic analysis (Poe et al., 2009a), should also be considered tentative given that the phylogenetic tree was not on n by Poe et al. (2009a): thus, the placement of these two species with respect to the latifrons series as conceptualized here is uncertain. The first use of the name “latifrons series” appears to have been in Etheridge’s (1959) dissertation; however, that use dee not qualify as published according to the ICPN (Article 4.2). Moreover, Etheridge used the name for a more inclusive taxon 378 approximating the clade to which the name Dactyloa is applied here. The oldest pub- lished use of the name “latifrons series” appears to be that of Gorman and Dessauer (1966), who also used the name for the more alee clade. As delimited here, the latifrons series more closely approximates the taxon called the laticeps group by Cope (1899; which appears to be a lapsus because the species is eee [p. 7] referred to by the correct name A. latifrons), the giant mainland anoles or quid eae group by Dunn (1937), the latifrons species group by Williams (1988), the latifrons series by Savage and Guyer (1989), and the latifrons Aece of Castafieda and de Queiroz (2011). However, the latifrons series as conceptualized here is potentially more inclusive than the latifrons clade of Castaneda and de Queiroz (2011), in that it appears to include A. philopunctatus and might also include some species currently MS ed incertae sedis within Dactyloa or absent from explicit phylogenetic analyses if they are found to be more closely related to the A. latifrons series than to ie aequator- ialis, A. punctatus, A. roquet, and A. heterodermus. Megaloa Castaneda and de Queiroz, new clade name Definition (apomorphy-based): The clade originating in the ancestor in which a maximum SVL > 100 mm in males, synapomorphic with that of Anolis latifrons Berthold 1546, originated. Reference phy- logeny: Figure 5, this study. Inferred composition: Anolis agassizi Stejneger 1900, A. apollinaris Boulenger 1919, A. casildae Arosemena, Ibanez, aon de Sousa 1991, A. chocorum Williams and Duellman 1967 (see “Comments’’), A. danieli Williams 1988, A. fraseri Giinther 1859, A. frenatus Cope 1899, A. insignis Cope 1871, A. latifrons Berthold 1846, A. maculigula Williams 1984, A. microtus Cope 1871, A. princeps Boulenger 1902, and A. purpur- escens Cope 1899 (see ya eM Another apenis that might belong to Megaloa iS Z A. squamulatus Peters 1863. (see Bulletin of the Museum of Comparative Zoology, Vol. 160, No. 7 “Comments’). Etymology: Derived from the Greek Mega (large) + loa (the last part of the name Dock yloa) in reference to the large body size of the members of this subclade of Dactyloa.’. Comments: Two supraspecific names are based on species in this clade: Diaphoranolis Barbour 1923 (type Species: = D. brooksi = Anolis insignis according to Etheridge [1959] and Savage and Talbot [1978]) and Mariguana Dunn 1939 (type species = Anolis agassizi). These names were applied to taxa sariked as genera and separated from Anolis based on differ- ences in dorsal scalation (juxtaposed pave- ment-like scales in A. insignis |Barbour, 1923], and tiny non- imbricating granules interspersed with larger, single, obtusely keeled scales in A. agassizi [Dunn, 1939: Etheridge, 1959]) and dewlap morphology (supposedly nonextensible in A. insignis |Barbour, 1923] and poorly developed in A. agassizi |Dunn, 1939]). Given that we are emphasizing the associations of names with clades, rather than with categorical ranks, and that neither of these names has been associated with the clade of mainland anoles with large body size (if they have been enced with clades at all, those clades are subclades of the large size clade), it is more appropriate to create a new name for this clade than to use either Diaphoranolis or Mariguana (which remain available for smaller clades including their type species). Therefore, we created a name that refers etymologically to the large size character (see “Etymology’ "i In the context of the reference phylogeny, A. chocorum and A. purpurescens are included in Megaloa despite not being known to possess the synapomorphy of the clade. In the case of A. chocorum, smaller size is parsimoniously interpreted as a reversal. In the case of A. purpurescens, the only two known male specimens have >The component loa is not intended to have any other meaning beyond reference to Dactyloa because it contains parts of both of the Greek words on which the name Dactyloa is based (daktylos, finger + oa, hem, border; in reference to the toe pads). PHYLOGENY OF THE DacryLoa ¢ Castaneda and de Queiroz SVLs of 74 and 78 mm and have been considered juveniles (Williams 1985; MRC, personal observation), which suggests that adults may reach a body size larger than 100 mm. Megaloa corresponds closely to the lati- ceps group of Cope (1899; which appears to be a lapsus Eee ee the species is elsewhere [Cope, 1899: 7] referred to by the correct name, A. latifrons), the giant mainland anoles or squamulatus-latifrons group of Dunn (1937), the latifrons species group of Williams (1988), the latifrons series of Savage and Guyer (1989), and the latifrons clade of Castaneda and de Queiroz (2011). However, it should be noted that Megaloa as conceptualized here is less inclusive than the latifrons series as con- ceptualized here, in excluding species that are more Closely related to A. latifrons than to A. aequatorialis, A. punctatus, A. roquet, and A. heterodermus but branched from the lineage leading to A. latifrons before large size evolved (currently, there is only one known species, A. philopunctatus, that is considered to belong to the latifrons series but not to Megaloa |Fig. 5]). If A. squamu- latus (which exhibits large body size) is part of the latifrons series, then it is also likely part of Megaloa, although it might not be art of either clade (see “Comments” on the ae series). punctatus series Guyer and Savage 1987 (“1986”), informal clade name Definition (branch-modified node- based): The crown clade originating in the most recent common ancestor of A. punc- tatus Daudin 1802 and all extant species that share a more recent common ancestor with A. punctatus than with Anolis aequa- torialis Werner 1894, A. latifrons Berthold 1846, A. roquet (Bonnaterre 1789), and A. heterodermus Duméril 1851. Reference phylogeny: Figure 5, this study. Inferred composition: Anolis anatoloros Ugueto, Rivas, Barros, Sanchez-Pacheco and Gar- cia-Pérez 2007, A. caquetae Williams 1974 (see “Comments’), A. dissimilis Williams 1965 (see “Comments”), A. jacare Boulenger 379 1903, A. menta Ayala, Harris and Williams 1984 (see “Comments’”), A. punctatus Dau- din 1802, A. ruizii Rueda and Williams 1986 (see “Comments’), A. santamartae Williams 1982 (see “Comments’’), A. solitarius Ruth- ven 1916 (see “Comments”), A. tigrinus Peters 1863, A. transversalis Duméril 1851, and A. vaupesianus Williams 1982. Other species that might belong to the punctatus series are A. deltae Williams 1974, A. gorgonae Barbour 1905, A. lamari Williams 1992, A. nasofrontalis Amaral 1933, A. paravertebralis Bernal Carlo and Roze 2005, A. pseudotigrinus Amaral 1933, A. soinii Poe and Yanez-Miranda 2008, and A. umbrivagus Bernal Carlo and Roze 2005 (see “Comments” ). Comments: In the context of the reference phylogeny, eight species traditionally associated with the punctatus series are excluded (A. boettgeri, A. chloris, A. chocorum, A. fasciatus, A. festae, A huilae, A. peraccae, A. philopuncta- tus), and four species not traditionally associated with the punctatus series (all placed in the tigrinus series, see below) are included (A. menta, A. ruizii, A. solitarius, A. tigrinus). Anolis tigrinus and its previously hypothesized relatives have traditionally been included in the tigrinus series (e.g., Williams, 1976b, 1992: Savage and Guyer, 1989): however, strong evidence supports A. tigrinus as nested within the punctatus series (Castaneda and de Queiroz, 2011). Williams (1992) noted problems in distinguishing the punctatus and tigrinus series (referred to by him as species groups) and raised the possibility that the tigrinus series is an ecomorphic subgroup of the punctatus series (he considered members of the tigrinus series to be representatives of the twig ecomorph, whereas he classified at least some members of the punctatus series as trunk-crown anoles). Our results support this hypothesis and we therefore consider A. tigrinus and its relatives part of the punctatus series rather than a separate tigrinus series, although a clade containing A. tigrinus and all species closer to it than to A. punctatus could be recognized as a sub-series or a species group within the punctatus series. 380 Bulletin of the Museum of Comparative Zoology, Vol. 160, No. 7 Strong evidence also supports the inclusion of A. chloris, A. fasciatus, A. festae, and A. peraccae in the aequatorialis series and A. chocorum in the latifrons series (see “Com- ments” on the aequatorialis and the latifrons series, above) and therefore the exclusion of those species from the punctatus series. In contrast, because of inconsistent placement or weak support for the relevant relationships between analyses (Fig. 3 versus Fig. 4), and because of a current ‘lnek of molecilen data for most of the species (all except A. huilae), exclusion of A. boettgeri, A. huilae, and A. philopunctatus from the punctatus series and inclusion of A. menta, A. ruizii, and A. solitarius in that series should be considered tentative. For similar reasons, inclusion of A. caquetae, A. dissimilis, and A. santamartae, all traditionally included in the punctatus series (Williams, 1965, 1974, 1982), should also be considered tentative. Anolis calimae, another species traditionally referred to the punctatus series (Ayala et al., 1983), was inferred as closely related to species that we tentatively refer to the punctatus series in the parsimony trees (Eig. 73) but not im the Bayesian tree (Fig. 4): because of this and additional contr: adictory results regarding the relationship between A. pane soap the punctatus series (Castafieda and de Queiroz, 2011), A. calimae is here considered incertae sedis (which does not rule out inclusion in the punctatus ser ies). The long branch leading to this species and its ambiguous relationships are consistent with W ‘kere: (1983) conclu- sion that this species has no evident close relatives. The geographic distribution of A. calimae in the central portion of the western Cordillera of Colombia between 1,300 and 1,800 m (Ayala et al., 1983) does not correspond to the eastern distribution of the punctatus series but instead corresponds more closely to the distribution of the aequatorialis or the heterodermus series. Although A. carlostoddi and A. orcesi were placed in the punctatus series in the parsimony tree (Fig. 3), they were not placed there in the Bay esian tree (Fig. 4). We have considered A. carlostoddi incertae sedis within Dactyloa based on the deep level of disagreement concerning its place- ment between analyses (Figs. 3, 4), as indicated by its basal regr: ‘on position on the pruned and regrafted consensus tree (Fig. 5). The eastern distribution of A. carlostoddi suggests that it may be part of the punctatus series, as does that of A. neblininus, another species that we consider incertae sedis but which is grouped in the parsimony tree with species that we tenta- tively refer to the punctatus series (A. menta, A. solitarius, and A. ruizii). By contrast, we have tentatively assigned A. orcesi to the heterodermus series (and Phenacosaurus) based on moderate support from the Bayesian analysis for its inclusion (Fig. 4) as well as its possession of charac- ters of the twig ecomorph (Losos, 2009). The distribution of A. orcesi on the eastern slopes of the northern Andes of Ecuador is consistent with referral to the heterodermus series, although it is also compatible with referral to the punctatus series (see “Com- ments” section on the heterodermus series below). The inclusion of A. deltae, A. gorgonae, and A. soinii, traditionally con- aera members of the punctatus series (Williams and Duellman, 1967: Williams, 1974: Poe and Yanez-Miranda, 2008), and A. lamari, A. nasofrontalis, A. paraverteb- ralis, A. pseudotigrinus, and A. umbrivagus, traditionally considered members of the tigrinus series (Williams, 1992; Bernal Carlo and Roze, 2005), should also be considered tentative given their current absence from explicit phylogenetic analyses. The punctatus series as conceptualized here is more inclusive than the eastern clade of Castaneda and de Queiroz (2011) in that it includes species that are more closely related to A. punctatus than to Anolis aequatorialis, A. latifrons, A. roquet, and A. heterodermus, but that div erged before the last common ancestor of the acmbes s of the eastern clade and might also include some species cur- rently considered incertae sedis within Dac- tyloa or absent from explicit phylogenetic analyses. Additionally, species in the eastern clade of Castafieda and de Queiroz (2011) are characterized by having a cohesive PHYLOGENY OF THE DactyLoa ¢ Castaneda and de Queiroz 38] eastern Andean and Amazonian geographic distribution, and it is possible that some species in the punctatus series do not conform to this geographic pattern. All of the species here tentatively referred to the punctatus series (A. caquetae, A. dissimilis, A. menta, A. ruizii, A. santamartae, A. solitarius) are outside of the eastern clade in the parsimony tree (Fig. 3), and one of them (A. caquetae) is outside of the eastern clade in the Bayesian tree (Fig. 4), although all have eastern geographic distributions (see “Phylogeny of Dactyloa” above for details). roquet series Williams 1976a, informal clade name Definition (branch-modified node- based): The crown clade originating in the most recent common ancestor of Anolis roquet (Bonnaterre 1789) and all extant species that share a more recent common ancestor with A. roquet than with A. aequatorialis Werner 1894, A. latifrons Berthold 1846, A. punctatus Daudin 1802, and A. heterodermus Duméril 1851. Refer- ence phylogeny: Figure 5, this study. Inferred composition: Anolis aeneus Gra 1840, A. blanquillanus Hummelinck 1940 (see “Comments” ), A. bonairensis Ruthven 1923, A. extremus Garman 1887, A. griseus Garman 1887, A. luciae Garman 1887, A. richardii Duméril and Bibron 1837, A. roquet (Bonnaterre 1789), and A. trinitatis Reinhardt and Lutken 1862. Comments: Referral of A. blanquillanus to this clade is based on the results of Yang et al. (1974) and Creer et al. (2001). The roquet series as conceptualized here corresponds exactly in known composition to the roquet group, species group, series, and clade of previous authors (Underwood, 1959: Gorman and Atkins, 1967, 1969: Lazell, 1972: Williams. 1976a; Savage and Guyer, 1989; Creer et al., 2001; Castafieda and de Queiroz, 2011). Nevertheless, the roquet series as concep- tualized here is potentially more inclusive than the roquet clade of Castafieda and de Queiroz (2011) in that it could include some species currently considered incertae sedis within Dactyloa or absent from explicit phylogenetic analyses if they are found to be more closely related to the roquet clade than to the four other mutually exclusive Dactyloa subclades inferred by Castaneda and de Queiroz (2011). heterodermus series Castaneda and de Queiroz, new informal clade name Definition (branch-modified node- based): The crown clade originating in the most recent common ancestor of Anolis heterodermus Duméril 1851 and all extant species that share a more recent common ancestor with A. heterodermus than with Anolis aequatorialis Werner 1894, A. lati- frons Berthold 1846, A. punctatus Daudin 1802, and A. roquet (Bonnaterre 1789). Reference phylogeny: Figure 5, this study. Inferred composition: Anolis eu- skalerriari. (Barros, Williams, and Viloria 1996), A. heterodermus Duméril 1851, A. inderenae (Rueda and Hernandez-Camacho 1988), A. nicefori (Dunn 1944), A. orcesi (Lazell 1969) (see “Comments”), A. probos- cis Peters and Orcés 1956 (see “Com- ments’), A. tetarii (Barros, Williams, and Viloria 1996), and A. vanzolinii (Williams, Orcés, Matheus, and Bleiweiss 1996). An- other species that might belong to the heterodermus series is A. williamsmitter- meierorum Poe and Yanez-Miranda 2007 (see “Comments’). Comments: Inclusion of A. orcesi, a species traditionally included in Phenacosaurus (Lazell, 1969), should be considered tentative given the inconsistent placement of this species between analyses (Fig. 3 versus Fig. 4) and because molecu- lar data are currently lacking. We have included A. orcesi in the heterodermus series, where it was placed in the Bayesian tree (Fig. 4), rather than in the punctatus series, where it was placed in the parsimony tree (Fig. 3), because of the stronger sup- port obtained for that relationship as well as its sharing of characters of the twig eco- morph with other species traditionally re- ferred to Phenacosaurus (Losos, 2009). The geographic distribution of A. orcesi along the eastern slopes of the northern Ecuadorean Andes corresponds with the distribution of 382 Bulletin of the Museum of Comparative Zoology, Vol. 160, No. 7 the heterodermus series, but also with that of the punctatus series. Anolis williamsmitter- meierorum., previously considered closely al to A. orcesi (Williams and Bae - 1991: Poe and Yaniez-Miranda, 2007), i snetes ely referred to Phenacosaurus given the current absence of this species ‘Grom explicit phylogenetic analyses. Anolis proboscis, A. laevis, and A. phyl- lorhinus were previously placed in the laevis species group or series, a group character- ized by the presence of a nose leaf (Williams, 1979). Here we consider the relationships of A. laevis and A. phyllorhi- nus to be uncertain (see “Comments” section on Phenacosaurus below). The geographic distribution of A. laevis in the eastern foothills of the Peruvian Andes does not suggest a close relationship with proboscis but is consistent with referral to the heterodermus series (as well as the aequatorialis and punctatus series). In contrast, the geographic distribution of A. phyllorhinus in central Amazonia (Williams, 1979; Rodrigues et al., 2002) suggests neither a dese relationship to A. proboscis nor inclusion in the heterodermus series but instead suggests inclusion in the punctatus series as ‘proposed by Rodrigues et al. (2002), Yanez-Munoz et al. (2010). and Poe et al. (2012). If A. laevis and one or both other species form a clade within the heterodermus series, then that clade could be recognized as the laevis species group (see also comments on Scytomycterus, below). If A. laevis and one or both other species are closely related to members of the punctatus series, as has been hypothe- sized previously (Williams, 1965, 1979), then they should be included within the punctatus series (perhaps as the laevis species group). However, if A. laevis and one or both of the other species lie outside of the five clades whose names incorporate the term “series” as defined here, then it would be appropriate to include them in a separate laevis series (defined as the most inclusive crown clade containing A. laevis but not Anolis aequatorialis, A. latifrons, A. punctatus, A. roquet, and A. heterodermus). Regardless of whether a separate laevis series is to be recognized, if A. laevis forms a clade with either or both A. phyllorhinus and A. proboscis that can be diagnosed by the nose- leaf synapomorphy (but see Ydnez-Mufioz et al., 2010), the name Scytomycterus Cope 1876 (derived from the Greek Skytos, skin or leather, + mykteros, nose; type species = A. laevis) areulle be an appropriate name for that clade. However, if A. phyllorhinus or A. proboscis form a clade but are not closely related to A. laevis, which differs from the other two species in having only a rudimen- tary nose leaf (Williams, “1979). the name Scytomycterus is not appropriate for that clade (given that the type is A. laevis); therefore, if that clade is to be named, a new name would be appropriate. The heterodermus series as conceptual- ized here corresponds closely to the P ee cosaurus Clade of Castafieda and de Queiroz (2011) and Phenacosaurus as conceptual- ized here. However, it should be noted that the heterodermus series as conceptualized here is potentially more inclusive than Phenacosaurus as conceptualized here (see below), in that it might include some species currently considered incertae sedis within Dactyloa or absent from explicit phyloge- netic analyses if they are found to be more closely related to A. heterodermus than to members of the other four well-supported clades but branched from the lineage leading to A. heterodermus before the twig morphology evolved (currently, all species assigned to the heterodermus series are also relcmed to Phenacosaurus). Phenacosaurus Barbour 1920, converted clade name Definition (apomorphy-based): The clade originating in the ancestor in which the combination of morphological charac- ters of the twig ecomorph (long pointed snout; forelimbs, hindlimbs, and eal short in proportion to body size), synapomorphic with that in Anolis heterodermus Duméril 1851, originated. Reference phylogeny: Figure 5, cae study. Inferred composition: Anolis euskalerriari (Barros, Williams, and PHYLOGENY OF THE DacryLoa ¢ Castaneda and de Queiroz 38. Viloria 1996), A. heterodermus Duméril 1851, A. inderenae (Rueda and Hernandez- Camacho 1988), A. nicefori (Dunn 1944), A. orcesi (Lazell 1969) (see “Comments”), A. proboscis Peters and Orcés 1956 (see “Com- ments’), A. tetarii (Barros, Williams, and Viloria 1996), and A. vanzolinii (Williams. Orcés, Matheus, and Bleiweiss 1996). An- other species that might belong to Phenaco- saurus is A. williamsmittermeierorum Poe and Yanez-Miranda 2007 (see “Comments’ ). Comments: Phenacosaurus was originally proposed (Barbour, 1920) as the name of a genus separate from Anolis. However, the aacicon of subsequently discovered species (e.g., Dunn, 1944; Lazell, 1969; Rueda and Hernandez-Camacho, 1988; Myers et al., 1993: Barros et al., 1996; Williams et al., 1996) has decreased the morphological gap between the two taxa, and several phy loge- netic studies (e.g., Jackman et al., 1999: Poe, 2004: Nicholson: et al., 2005; Castaneda and de Queiroz, 2011; this study) have inferred Phenacosaurus to be nested within Anolis, so that recognizing Phenacosaurus as a genus would fonder Agols paraphyletic. We here fore use the name Phenacosaurus for a subclade of Anolis that is not associated with the rank of genus (it is implicitly associated with a lower rank). All species in the Phenacosaurus clade have been considered twig anoles (Losos, 2009), an ecomorpholo- gical category characterized by long pointed snouts, fo toepad lamellae, short abe: and short, often prehensile, tails. Because several of those characters were ae in the original diagnosis of Phenacosaurus (Barbour, 1920), we haw e defined that name as referring to the clade of twig anoles that includes its type species (A. heterodermus). In the context of the reference phylogeny, one species not traditionally efamed to Phenacosaurus is included (A. proboscis). Molecular data are currently lacking for A proboscis, but this species was goneistenty placed with other species referred to Phena- cosaurus (see also Poe et al., 2009b, 2012). Anolis proboscis possesses the morphological features characteristic of the twig ecomorph: long pointed snout, forelimbs, hindlimbs, WwW and tail short in proportion to body size. Moreover, ecological data indicates A. pro- boscis should be classified as a twig anole (Losos et al., 2012; Poe et al., 2012). Two species traditionally referred to Phenaco- saurus, Anolis carlostoddi and A. neblininus. were pk ced etna ‘ntly between analyses (Fig. < 3 versus Fig. 4 but in neither case were they pli need ae Phenacosaurus. Because molecular data are curre ntly lacking for A. carlostoddi and because some analyses based on molecular data suggest inclusion of A. neblininus in Phenacosaurus (Castaneda and de Queiroz, 2011, fig. 2C), neither species can be confidently excluded. Both species are here considered incertae sedis within Dactyloa. Inclusion of A. orcesi, a species traditionally included in Phenaco- saurus (Lazell, 1969), should be considered tentative given the inconsistent placement of this species between analyses (Fig. 3 versus Fig. 4) and because molecular data are currently lacking. We have included A. orcesi in Phenacosaurus, where it was placed in the Bayesian tree (Fig. 4), rather than in the punctatus series, W ae “ Was placed in the parsimony tree (Fig. 3), because of the stronger support bened for that relation- ship as well as its sharing of characters of the twig ecomorph with ae species tradition- ally referred to Phenacosaurus (Losos, 2009). Anolis williamsmittermeierorum, previously considered closely related to A. orcesi (Williams and Mittermeier, 1991: Poe and Yafiez-Miranda, 2007), is tentatively referred to Phenacosaurus given the current absence of this species from explicit phylogenetic analyses. Anolis bellipeniculus is tentatively excluded from Phenacosaurus based on its pent hypothesized close relationship to A. neblininus (Myers and Donnelly, 1996) and is here considered to be of uncertain position within Dactyloa (see “Incertae sedis,” below). Williams (1979) hypothesized that A. laevis and A. phyllorhinus are closely related to A. proboscis, which is here included in Phenacosaurus; however, that relationship has been questioned by Yanez-Munoz et al. (2010) and Poe et al. (2012), and therefore 384 Bulletin of the Museum of Comparative Zoology, Vol. 160, No. 7 we consider A. laevis and A. phyllorhinus to be incertae sedis within Dactyloa. Little is known about A. laevis, which is known only from the type specimen, now in poor condition (Williams, 1979). However, as illustrated in Williams (1979, fig. 1), this specimen does not possess a nose leaf but only a protruding rostral scale, which is questionably homologous with the ample appendages of the oie: species. Moreover, the geographic distribution of A. [aevis in the eastern ‘foothills of the Peruvian Andes does not suggest a close relationship with A proboscis, and although it is consistent with referral to the heterodermus series, it is also consistent with referral to the aequatorialis and punctatus series. Anolis phyllorhinus is better known, and although it possesses a true nose leaf, which is both similar to and different from that of A. proboscis, the information in Williams (1979) and Rodri- gues et al. (2002) suggest that A. phyllorhi- nus is a trunk-crown rather than a twig anole (e.g., green coloration, moderate snout and limb lengths, long tail, high lamella counts, relativ oe large perch Aen upward flight behavior, and high degree of similarity to “A. punctatus, which has been classified as a trunk-crown anole [Williams, 1992]). Moreover, the geographic distribution of A. phyllorhinus in central Amazonia (Williams, 1979; Rodrigues et al., 2002) suggests neither a ee relationship to A. proboscis nor inclusion in the heterodermus series but instead suggests inclusion in the punctatus series. Although we think that Yanez-Mufioz et al. (2010) are likely correct in assigning A. phyllorhinus in the punctatus series, the inclusion of neither A. phyllorhinus nor A. laevis in an explicit phylogenetic analysis leads us to treat both species as incertae sedis within Dactyloa. Phenacosaurus as conceptualized here is more inclusive than the Phenacosaurus clade of Castafieda and de Queiroz (2011) in that it contains species (e.g., A. proboscis and possibly A. orcesi, see below) that share the twig ecomorph synapomorphy with A. hetero- ie mus but lie outside of the smallest clade containing A. heterodermus and A. euskalerriari. Phenacosaurus as conceptual- ized here is less inclusive than the hetero- dermus series as conceptualized, in excluding species that are more closely related to A. heterodermus than to members of the other four clades recognized here whose names include the term “series,” but branched from the lineage leading to A. heterodermus before the twig morphology evolved (although cur- rently all known species referred to Phenaco- saurus are also referred to the heterodermus series). Incertae sedis The placement of the following species could not be resolved because of lack of data or because of conflicting results between analyses, and we therefore defer assigning them to any of the above de- scribed clades until more definitive evi- dence is available: A. calimae Ayala, Harris and Williams 1983, A. carlostoddi (Williams, Emon? and Gorzula 1996), A. laevis (Cope 1876), A. neblininus (Myers, Williams and Mabiaaaid 1993), and A. phyllorhinus Myers and Carvalho 1945. Possible rela- tionships of these species have been dis- cussed under “Comments” on the punctatus series (A. calimae, A. carlostoddi, and A. neblininus), the heterodermus series (A. laevis and A. phyllorhinus), and Phenaco- saurus (A. carlostoddi, A. laevis, A. neblini- nus and A. phyllorhinus). We also consider the position of A. bellipeniculus (Myers and Donnelly 1996) to be uncertain within Dactyloa given its hypothesized close rela- tionship to A. neblininus (Myers and Donnelly, 1996) and the uncertain relation- ships of that species. The eastern distribu- tion of A. bellipeniculus on the isolated Cerro Yavi tepui of southeastern Venezuela (Myers and Donnelly, 1996) suggests that it may be part of ie punctatus series. Similarly, the placement of A. cuscoensis Poe, ieee Miranda and Lehr 2008 is considered unresolved, because although this species has been included in an explicit phylogenetic analysis (Poe et al., 2008), its hypothesized relationships are incongruent PHYLOGENY OF THE DacryLoa ¢ Castaneda and de Queiroz 385 with the clades recognized here. The geographic distribution of this species along the eastern slopes of the southern Peruvian Andes (Poe et al., 2008) is congruent with the distributions of the Tere ini. punctatus, and aequatorialis series (if ex- tended to the south). Although many species in the aequatorialis series have western distributions, the earliest branching species within the clade (including boettgeri, which was considered closely related to A. cuscoensis) inhabit the eastern slopes of the Andes. According to the definitions presented above, some species might not belong to any of the five clades whose names incorporate the term “series”; specifically, any species or clade that is sister to a clade composed of two or more of the five clades whose names include the term “series” would not be a member of any of those clades. If strong support were to be found for such relation- ships, new “series” names could be pro- posed for the corresponding species or clades, although such names might be judged unnecessary for “series” ’ composed of single species. Currently, however, most known species of Dact yloa are at least tentatively referable to one of the five mutually exclusive “series” clades, and even those species that are the best candidates for not being members of those clades (i.e., the species Ghat we consider incertae Reais) might belong to them. ACKNOWLEDGMENTS This research was partially funded by The George Washington University and the Ernst Mayr Travel ee ant in Animal Systematics. For access to herpetological collections, the first author thanks (in Colombia) Mauricio Alvarez and Diego Perico (Instituto Hum- boldt), Hermano Roque Casallas and Arturo Rodriguez (Museo La Salle), Fernando Castro (Universidad del Valle), John Lynch and John Jairo Mueses- N@isncies (Instutito de Ciencias Naturales, Universidad Nacional). Vivian Paez and Paul D. Gutiérrez (Museo de Herpetologia—Universidad de Antioquia); (in Venezuela) Tito Barros and Gilson Rivas (Museo de Biologia—Universidad del Zulia), Celsa Cenaris (Museo de Historia Natural La Salle); (in Ecuador) Ana Almendariz (Uni- versidad Politécnica del Ecuador). Luis Coloma (formerly at Universidad Catélica de Quito), Mario Yaénez-Mufioz (Museo Ecuatoriano de Ciencias); and (in the United States) James Hanken, Jonathan Losos, and Jose Rosado (Museum of Comparative Zool- ogy) and Jeff Seigel (Los Angeles County Museuri): Steve Gotte, Ken Tighe, Rob Wilson, and Addison Wynn (U.S. National Museum of Natural History) provided help with the clearing and staining of specimens, radiographs, specimen (pane and other collection-related issues. James Clark pro- vided comments on earlier versions that resulted in significant improvements. Omar Torres-Carvajal provided FREQPARS files and help with the coding methods. APPENDIX | Morphological Character Descriptions Description of morphological characters used in phylogenetic analyses. Ranges of species means (for continuous characters) correspond to values before data transformation and coding. Results of correlation tests (R? and P values) are shown. External Characters, Examined on Alcohol-Pre- served Specimens 1. Maximum male snout-to-vent length (SVL; Wil- liams et al., 1995, character 35) ) WMieasured with a 1-mm precision ruler from the tip of the snout to the anterior lip of the cloacal opening. Continu- ous character. Range: 41-170 mm. Ratio of maximum female SVL to maximum male SVL (Poe, 1998, character 11), both measured with a l-mm precision ruler. This character was not correlated with SVL (R? = 0.03, P = 0.25). Continuous character. Range: 0.64-1.35. 3. Length of head (Poe, 2004, character 4), measured with 0.01-mm precision calipers from the tip of the snout to the anterior edge of the ear ope ning. This character was cor netted: with SVL (R? = 0.94, P< 0.001) and head width (R? = 0.97, P < 0.001). To correct for size, head length mean values were natural log transformed anid regressed on natural log-transformed SVL mean values. Residuals were subsequently used. Continuous character. Range: 10.68-47.16 mm. bo 386 6. ~I Bulletin of the Museum of Comparative Zoology, Vol. 160, No. 7 Width of head (Poe, 2004, character 5), measured with 0.01-mm precision calipers at the widest part of the head—usually the corners of the mouth. This character was correlated with SVL (R° = 0.94, P < 0.001) and head length (R? = 0.97, P < 0.001). To correct for size, head width mean values were natural log transformed and regressed on natural log-transformed SVL mean alice Residuals were “subsequently used. Con- tinuous character. Range: 5. 00- 29.46 mm. Height of ear (Poe, 2004, character 6), measured between the internal borders with 0.01-mm precision calipers. This character was correlated with SVL (R* = 0.58, P < 0.001), head length (R = 0.46, P < 0.001), and head width (R? = 0.56, P < (0.001). To correct for size, ear height mean values were natural log transformed and _re- gressed on natural log-transformed SVL mean values. Residuals were subsequently used. Con- tinuous character. Range: 0.60-4.84 mm. Interparietal scale length (modified from Poe, 2004, character 7), measured with 0.01-mm_ precision calipers from the anterior to posterior edges of the scale. The interparietal scale is defined as the scale overlying the parietal foramen (Peters, 1964). Located in the parietal area, this scale is typically of larger size than surrounding scales and exhibits an area of clear skin above the parietal eye. In some species, no clear skin area is observed, ‘but a scale appears to be homologous to the interparietal based on position, shape, and size. These scales were measured as interparietals. When scale edges were not parallel to each other, the distance between the most anterior to the most posterior points on the scale were measured. This character was correlated with SVL (R° = 0.07, P = 0.03), head length (R° = 0:09:42 — 0.002), anchhead width (R= — 0105 2 = 0.01). To correct for size, interparietal length mean values were natural log transformed and regressed on natural log-transformed SVL mean values. Residuals were subsequently used. Continuous character. Range: 0.55—3.86 mm. Mean number of dorsal scales in 5% of SVL (Poe, 2004, character 19). The equivalent of 5% of SVL was set on 0.01-mm precision calipers, and the number of scales contained in this length was counted three times (using the average as the final count) lateral to the dorsal midline at the level of the forelimbs. This character is an estimate of dorsal scale size. Continuous charac- ter. Range: 4.20-18.27 Mean number of ventral scales in 5% of SVL (Poe, 2004, character 20). The equivalent of 5% of SVL was set on 0.01-mm precision calipers, and the number of scales contained in this length was counted three times (using the average as final count) lateral to the ventral midline in middle and posterior areas of the body. This character is an estimate of ventral scale size. Continuous character. Range: 5.08—13.80. oO 10. isle 13. Mean number of scales between the second canthals (Williams et al., 1995, character 2). Minimum count between left and right second canthals, excluding canthal scales. This character was not correlated with SVL (R° — 0\0ieps— 0.38), head length (R* < 0.001, P = 0.93), o1 head width (R2 = 0.003, P = 0.64), thus no correction for size was applied. Continuous character. Range: 2.00-17.50. Mean number of postrostral scales (Williams et al., 1995, character 3). Postrostrals are all scales in contact with (posterior to) the rostral scale, between supralabials. This character was corre- lated with SVL (R? = 0.08, P = 0.02) and head width (R° = 0.07, P = 0.03), but not head length (R? = 0.05, P = 0.06). To correct for size, mean numbers of postrostral scales were natural log transformed and regressed on natural log—trans- formed SVL mean values. Residuals were subse- quenly used. Continuous character. Range: 2.88-8.60. Mean number of scales between supraorbital semicircles (Williams et al., 1995, character 6). Minimum count between left and right supraor- bital seimicircles. This character was correlated with SVL (R° = 0.20, P < 0.001), head length (R° = 0.16, P < 0.001), and head width (R? = 0.18, P < 0.001). To correct for size, the mean numbers of scales between supraorbital semicircles were In(x + 1) transformed and regressed on natural log— transformed SVL mean values. Residuals were subsequently used. The In(x + 1) transformation was used because this character contains mean zero values. Continuous character. Range: 0-5.50. Mean number of loreal rows (Williams et al., 1995, character 10). Loreal scales cover the area between canthals, supralabials, and subocular scales. Rows were counted as the minimum number of scales, in a straight line, from the first or second canthal to the sublabial scales on the right side of the head, unless the area was damaged, and then the left side was scored. This character was correlated with SVL (R? = 0.12, P = 0.01), head length (R* = 0.06, P = 0.05), and head width (R2 = 0.10, P = 0.01), To comecttowsize: mean numbers of loreal rows were natural log transformed and regressed on natural log-trans. formed SVL mean malice Residuals were subse- quently used. Continuous character. Range: 1.00— 10.20. Mean number of supralabial scales to below the center of the eye (Williams et al., 1995, character 16), counted from the rostral (not included) to the midpoint of the eye. More than half the scale had to be anterior to the center of the eye to be included in the count. This character was correlated with SVL (R? = 0.12, P = 0.01), head length (R° = 0.15, P < 0.001), and head width (R = 0:09. P 0.02). To correct for size, mean numbers of supralabial scales were natural log 15. IG; PHYLOGENY OF THE DacTyLoa * Castaneda and de Queiroz 387 transformed and regressed on natural log—trans- formed SVL mean values. Residuals were subse- que tly used. Continuous character. Range: 5.00—11.00. Mean number of postmental scales al., 1995, character 17) (Williams et . Postmentals are all scales in contact with (posterior to) the mental scale between the infralabials (i.e., including the anteriormost sublabial scale on left and_ right sides). This character was correlated with SVL (R? = 0.06, P = 0.04) and head width (R? = 0.07, P = 0.03), but not head length (R? = 0.03, P = 0.21). To correct for size, mean number of postmental scales were natural log transformed and regressed on natural log-transformed SVL mean values. Residuals were “subsequently used. Continuous character. Range: 2.63—10.13. Mean number of sublabial scales (Williams et al., 1995, character 18: Poe, 2004, character 44). Sublabial scales are abruptly enlarged scales (more than twice the size) located medial and parallel to the infralabials and posterior to the mental. This character was correlated with SVL (R203, P= 0.001); head length (R* = 0.07, P = ().03), and head width (R? = = lly P= 0,01); To correct for size, the mean number of sublabial scales were In(x + 1) transformed and regressed on natural log-transformed SVL mean. values. Residuals were subsequently used. The In(x + 1) transformation was used because this character contains mean zero values. Continuous character. Range: 0-7.00. Mean number of scales between the interparietal scale and the supraorbital semicircles (Williams et al., 1995, character 13; Poe, 2004, character 46). The minimum number of scales between the interparietal scale and the supraorbital semicir- cles was counted. This character was correlated with SVL (R° = 0.11, P = 0.01), head length (R? — (07. P = 0.03),.and head widthi(kR? = 0. 09, P = ().02). To correct for size, the mean number of scales between interparietal and supraorbital semicircles were In(x + 1) transformed and regressed on natural log-transformed SVL mean valies. Residuals were subsequently used. The In(x + 1) transformation was used because this character contains mean zero values. Continuous character. Range: 0-7.25. Number of elongated superciliary scales (Wil- liams et al., 1995, character 8). Superciliaries are scales along the dorsal rim of the orbit, and elongation occurs toward the posterior end of the orbit. Left and right sides were scored separately. States: (0) 0, (1) Lees (38. Polymorphic character. Ordered. Number of scales between subocular and supra- labial scales (Williams et al., 1995, character 15: Poe, 2004, character. 28). The minimum number of scales was recorded for each specimen. States: (0) 0, (1) 1, (2) 2. Polymorphic character. Ordered. 19. bo to Number of ventral scales posteriorly bordering one scale (modified from Poe, 2004, character 14). Middle and posterior ventral areas were examined. Ventral scales may be bordered posteriorly by two scales (0), by two and three scales (1), or by three Seales (2). Polymorphic character. Ordered. Shape of the base of the tail (modified from Williams et al., 1995, character 30; Poe, 2004, character 15). On each specimen, at the point where the knee would reach the tail if the leg were folded back, the height and width of the tail was measured, and then the ratio of width/height was calculated. States: (0) tail round, for ratios larger than 1; (1) tail laterally compressed, for ratios smaller than 1. Poly morphic character. aes overlap ( (Williams - al., 1995, character Poe, 2004, character 9). The toepad under ae III and IV may ace caer pe phalanx II (0) or not project distally (1), or the toepad may be completely absent (2 a Nonae phic character. Ordered. Male dewlap extension (Williams et al., 1995, character 33; Poe, 2004, character 16). On the ventral side, the posterior extension of the unfolded dewlap is examined. Four states were considered: posterior extension past the arm insertion (QO), posterior Seenon . arm insertion (1), shorter than arm extension (2), dewlap absent (3). Polymorphic character. ae Female dewlap extension (Williams et al., 1995, character 34; Poe, 2004, character 17). Measure- ment and coding as in male dewlap extension. This character is not correlated with male dewlap extension (R? = 0.210, P = 0.136); therefore, it was considered a separate character. Polymor- phic character. Ordered. Size of scales in supraocular discs (modified from Poe, 2004, character 41). Three different states were considered: (0) scales vary continuously in size, in which a few scales are slightly larger (less than twice the size) than the others, showing gradual reduction in size; (1) one to three abruptly enlarged scales (more than twice oe size) with all other scales of smaller size; and ( all scales about equal in size. ey character. Unordered. Dewlap scales (modified from Poe, 2004, charac- ter 21). The scales on the dewlap may be in rows of single scales (0); in double rows (1) or have scattered scales covering the entire a (2). In some specimens, most rows were either single or double, with a few rows exhibiting the alternative condition. In such cases, the most common condition was scored for the specimen. Polymor- phic character. Unordered. Width of mental relative to rostral (modified from Poe, 2004, character 27). In ventral view, the mental scale may be broader than the rostral (0), the rostral scale may be broader than the mental (1), or both 388 bo oa 30. Sl. Bulletin of the Museum of Comparative Zoology, Vol. 160, No. 7 scales may show the same width (2). character. Unordered. Enlarged postanal scales in males (Williams et al., 1995, character 32). Postanal scales may be: (0) absent, (1) present, as a pair of significantly oe es than four times the surrounding scales), 2) present, as a series of more than two scales qa enlarged (less than twice the size of surrounding scales). ). Polymorphic character. Un- ordered. Presence or ). Polymorphic absence of tail crest in males (Williams et al., 1995, character 31). The tail crest in males may be: (0) absent, (1) present as a series of enlarged, but not elevated, serrated scales, or (2) present as the result of enlarged neural spines. The presence of a crest is associated with sex and age of the specimen; therefore, when intraspecific variation was observed (presence and absence) the species was coded as present. However, states | and 2 were never observed in the same species. Unordered. Heterogeneous flank scales (modified from Wil- liams et al., ae character 23). Heterogeneous scales may be (0) absent, (1) very large and separated ‘from one another by many eeales of much smaller size, (2) a mosaic of scales of different sizes but not very different in size from one another, or (3) of average size surrounded by granular-minute scales. Polymorphic character. Unordered. Mental scale (Poe, 2004, character 26). The mental scale may be partially divided (0), in which a longitudinal split begins from the posterior edge of ane mental but Hoes: not reach the anterior edge, or completely divided (1), in which the split is complete. Polymorphic character. Frontal depression (Poe, 2004, character 45). A depression around the frontal area may be absent (0), in which case the dorsal surface of the snout is flat, or present (1). Polymorphie character. Presence or absence of an externally visible parietal eye (Estes et al., 1988, character 26). The parietal eye, when visible externally, is located within the interparietal scale (see char- acter 6). States: absent (0), present (1). Polymor- phic character. Keeling of dorsal, ventral, supradigital and head scales (Williams et al., 1995, characters 20, 25, 29, 1: Poe 2004, character 40). Dorsal, ventral, and Be scales may be smooth (S) or keeled K); head scales may in addition be rugose (R) or on pustules (P). The four apparently indepen- dent characters were combined as one. after correlation was found between ventral, supradigi- tal, and head keeling with dorsal keeling (R* = 0.201. P= 00001. Rh? = 016385) P= 0.0001. R2 = 0.456, P < 0.0001, respectively). The condition present in the majority of the scales was reported. Dorsal scale keeling was scored excluding mid- dorsal scales heeause these often differ fom the remaining dorsals (e. g., some species exhibit smooth dorsal scales, but a double row of keeled middorsal scales). Weakly keeled specimens were coded as keeled. Rugose refers to multiple, less pronounced keels or bent ridges (these two conditions were commonly foul combined in one scale); with pustules refers to multiple granular projections scattered on the scale. States (tar dorsals, ventrals, supradigitals, head scales): KKKK (0), KKKR (1), KKKS (2), KKSP (3), KSKK (4))-KSKS (5), SSKRe(6).iSSKSs@m=sSss (S), SSSR (9). Modal condition coded. Unordered. Osteological Characters Examined on Dry, Cleared, and Stained Specimens and/or Radiographs. 34. OD: B10) Shape of parietal crests (Etheridge, 1959, fig. 9; Cannatella and de Queiroz, 1989, characters 6, 7: Williams, 1989, character 7, modified from Poe, 1998, character 87). Three different states were considered: (0) trapezoid—shape: lateral borders of the crest reach the occipital crest directly (i.e., do not touch each other before occipital crest contact); (1) V-shape: lateral borders of the crest join at the point of contact with the occipital crest, and there is no extension beyond the point of contact; (2) Y-shape: lateral borders of the crest join before occipital crest contact and extend posteriorly beyond the point of contact (i.e., a unified crest extends toward or beyond the occipital). Etheridge (1959) showed that this character exhibits ontogenetic variation, from a U/trapezoid shape seen in early stages to an intermediate V-shape, to a Y- shaped crest seen in adult stages. To compensate for the absence of sex and SVL information to confirm adulthood in some specimens, the most devel- oped state observed (following Etheridge’s ontogenetic sequence) was scored for each species. Ordered. Presence or absence of crenulation along lateral edges of parietal (Poe, 1998, character 88). The parietal may exhibit irregular (crenulated) or smooth lateral edges. States: absent (O), present (1). Polymorphic chia acter. Extension of the parietal roof (modified from Poe, 2004, character 59). In anoles, a parietal casque has been defined as the shelf-like posterolateral extension of the parietal roof over the supratem- poral processes of the parietal. Poe (2004) coded the presence or absence of the casque, but we found variation in the length of the extension. The roof extension may be large, almost completely covering the supratemporal processes and some- times extending bey ‘ond the posterior most mar gin, or the extension could be small, leaving more of the supratemporal process uncoy ered ands not reaching the posterior margin of the parietal. States: not extended (0; e.g., A. chocorum, MCZ 115732); present and small, not reaching posteriormost margin of supratemporal processes (1; e.g., A. fitchi, MCZ 178084); present and large, reaching or PHYLOGENY OF THE DacTYLoA ¢ Castaneda and de Queiroz 389 pm-par (0) Figure 6. Dorsal views of the skulls of three anoles illustrating differences in the extension of the parietal roof (character 36). (a) Anolis chocorum, MCZ 115732 (state 0); (b) Anolis fitchi, MCZ 178084 (state 1); (c) Anolis heterodermus, MCZ 110138 (state 2). Scale bar = 5 mm. Abbreviations: par, parietal; st-par, Supratemporal processes of the parietal; pm-par, posterior margin of parietal. extending beyond the posteriormost margin of supratemporal processes (2; e.g., A. heterodermus, MCZ 110138) (Fig. 6). The largest extension of the parietal roof observed among all individuals was scored for each species. Ordered. Parietal foramen (Etheridge, 1959; Williams, 1989, character 5). The parietal foramen may be located completely nn the parietal (0) or may be at the fronto-pé arietal suture (1). Cases in which the foramen is located within the parietal but connect- ed to the fronto-parietal by a suture were coded as 0. Absence of the parietal foramen was coded as ?, instead of as a third state, given that the information on the presence or Absences of an externally visible parietal eye was coded as a separate character 39. (ch: Wc icte Yr 32) from alcohol- -pre SCIFVE Yel S} YE cime VS. Polymorphic character. Fronto- parietal suture (this study). The fronto- pi ee suture may form a straight transverse line (ie.,. perpe dicular to the longitudinal axis of the ee AO; C.8., 2. \. danieli, MCZ 164894) or it may exhibit a posteriorly convex curve in the center because of extension of the frontal bone into the parietal bone (1; e.g., A. eulaemus, MCZ 158390 (Fig. 7). Significant variation was observed in the latter state (from a slight to a substantial protuberance), but this variation was not quan- tified. Polymorphic character. Presence or absence of postfrontal and de Queiroz, 1988, character 6; Poe, (Etheridge 1998, Figure 7. Dorsal views of the skulls of two anoles illustrating differences in the fronto-parietal suture (character 38). (a) Anolis danieli, MCZ 164894 (state 0); (b) Anolis eulaemus, MCZ 158390 (state 1). Scale bar = 5 mm. Abbreviations: fr, frontal; par, parietal. 390 40. 41. 43. 44, Bulletin of the Museum of Comparative Zoology, Vol. 160, No. 7 character 92; 2004, character 62). The postfrontal bone is located in the posterodorsal margin of the orbit between (or overlapping) the frontal and the postorbital. States: absent (0), present (1). Polymorphic character. Frontal (Poe, 1998, character 94: 2004, character 64). The anterior suture of the frontal may be in contact only with nasals (0), may be separated from nasals by an open gap (1 ), or may be in contact with both the premaxilla and nasal (2). State 1 includes cases where the gap was along the entire suture or, most commonly, in tie center only, allowing partial lateral contact between frontal and nasals. Posterior extension of the premaxilla, sufficient to potentially contact the frontal (ie., if the gap were absent), was never observed along with the open gap. Polymorphic character. Unordered. Prefrontals (Poe, 1998, character 93, fig. 4). Prefrontals may be in contact with nasals (0) or may be separated from nasals by the contact between frontal and maxilla (1). Any contact between prefrontal and nasal was scored as 0. Differences between left and right sides were observed in some specimens; therefore, each side was treated separately for frequency calculations. Polymorphic character. Posterior extension of maxilla (Poe, 1998, charac- ter 103, fig. 8). Different landmarks have been used as boundaries to quantify the posterior extension of the maxilla (e.g., Estes et al., 1988, character 27; Frost and Etheridge, 1989, dravactsn 3). Following Poe (1998), the posterior edge of the ectopterygoid was used to delimit two different states: (0) maxilla does not extend posteriorly beyond the posterior edge of ectopterygoid (including cases in which it extends to that level) or (1) maxilla extends beyond the posterior edge of ectopterygoid. Differences between left and right sides were observed in some specimens; therefore, each side was treated independently for frequency calculations. Polymorphic character. Mean number of premaxillary teeth (de Queiroz, 1987, characters 43, 44). This ae was not correlated with SVL (R? = 0.01, P = 0.46), head length (R° = 0.013, P = 0.39), or head width (R? = (0.009, P = 0.47); thus, no correction for size was applied. Range: 6-13. Continuous character. Presence or absence of pterygoid teeth (Ether- idge, 1959; Poe, 1998, character 101). Pterygoid teeth are found along the edge facing the pyriform recess, either clumped or in a single row. States: absent (0 ), present (1). Polymorphic character. Presence or absence of contact between jugal and ee (Frost and Etheridge, 1989, character 8). The jugal and squamosal bones may be in contact along the ventral edge of the temporal bar, or they may be separated by the postorbital bone. In some specimens, differences between the left and right sides were found; therefore, each side 46. 48. 49. was treated separately for frequency calculations. States: absence (0), presence (1). Polymorphic character. Shape of posteroventral corner of jugal (modified from Poe, 2004, character 69). Poe (2004) recognized two states of this character: postero- ventral corner of jugal i is anterior to the posterior edge of jugal (in species where the posterior edge of the jugal shows a straight or convex border) or is posterior to the posterior edge of the jugal (in species where the posterior edge of jugal shows a concave border). However, we found these two character states not to be mutually exclusive; therefore, the states were modified as follows: posterior border of the jugal concave, with a sharp (pointed) posteroventral corner (0), or posterior border straight or convex, with a rounded posteroventral corner (1). Differences between left and right sides were observed in some specimens; therefore, each side was treated independently for frequency calculations. Poly- morphic character. Presence or absence of contact between parietal and epipterygoid (Poe, 1998, character 99). The epipterygoid extends from the palate toward the skull roof and may or may not reach the parietal. In some species, the most distal portion of the epipterygoid is cartilaginous and often lost during skull preparation, rendering the structure not in contact with the parietal. Cases in which the absence of contact is an artifact of preparation could not be distinguished from those in which the epipterygoid (with or without cartilaginous ortion) is short enough not to be in contact with the parietal. All cases with no contact were coded as absence. No intraspecific variation was ob- served. States: absent (0), presenti 1): Supraoccipital cresting (Poe, 1998, character 105, fig. 9; 2004, character 55). The supraoccipital may show: (0) a single medial process (called processus ascendens; e.g., A. heterodermus, MCZ 110133); (1) a medial process in addition to two distinct and smaller lateral processes (not always ossified; e.g., A. chloris, MCZ 101290) or (2) a continuous (e.¢., A. agassizi, MCZ 18088) or partially continuous crest - (showing two lateral processes with a distinct crest between them) running along the edge of the osseus labyrinth (Fig. 8). Significant ontogenetic variation was observed within each one of the states, but a sequence linking all three states was not observed; therefore, the modal condition was scored for each species. Unordered. Contact between parietal and supraoccipital (this study). The parietal may be widely separated from the supraoccipital, leaving free space between the two on either side of the processus ascendens (0; e.g., A. princeps, MCZ 147444), or may be in contact (or almost in contact) with the supraoc- cipital, leaving no open space in between (1; e.g., A. ventrimaculatus, MCZ 127711) (Fig. 9). Poly- morphic character. PHYLOGENY OF THE DactyLoa ¢ Castaneda and de Queiroz 39] Figure 8. Posterior views of the skulls of three anoles illustrating differences in the supraoccipital cresting (character 48). (a) Anolis heterodermus, MCZ 110133 (state 0); (b) Anolis chloris, MCZ 101290 (state 1); (c) Anolis agassizi, MCZ 18088 (state 2). Scale bar = 5 mm. Abbreviations: pa, processus ascendens; soc, supraoccipital. 50. Extension of the supratemporal processes of the observed within some species: thus, the devel- parietal (this study). In some species (e.g., A oped state (1.e., presence of quadrate lateral agassizi, MCZ 27120), the supratemporal pro- shelf) was scored for the species when observed. cesses of the parietal extend dorsally forming a States: absent (0), present ( i: . vertical flange (1); in others (e.g., A. heteroder- 22. Presence or absence of angular process of mus, MCZ 110133), the supratemporal processes prearticular (de Queiroz, 1987, character 41, fig. of the parietal do not extend (0) (Fig. 10). 28; Poe, 1998, character 110, fig. 11). This process Significant variation was observed in the height is located on the medial side of the retroarticular of the extension, but it was not quantified. process of the prearticular and has a fin-like or Additionally, ontogenetic variation was observed rounded shape. Presence was coded as a signifi- within some species; therefore, the most devel- cant extension beyond an imaginary line along the oped state (i.e., supratemporal processes extend- medial edge (in dorsal view) of the prearticular. ed) was scored for the species if it was observed Absent and rudimentary processes were coded as in any specimens. States: supratemporal process- absent. Differences in size of the process were es of the parietal not extended (0), extended (1). observed, but were not quantified. Poe (1998, 51. Presence or absence of the quadrate lateral shelf 2004) called this structure angular process of the (Poe, 1998, character 106, fig. 10). The quadrate articular, but the articular is an endochondral lateral shelf is the lateral extension of the external (rather than dermal) bone that results from the edge of the quadrate. Ontogenetic variation was ossification of the posterior end of Meckel’s Figure 9. Posterior views of the skulls of two anoles illustrating differences in the contact between the parietal and supraoccipital (character 49). (a) Anolis princeps, MCZ 147444 (state 0); (b) A. ventrimaculatus, MCZ 127711 (state 1). Scale bar = 5 mm. Abbreviations: par, parietal; soc, supraoccipital. 392 Bulletin of the Museum of Comparative Zoology, Vol. 160, No. 7 Figure 10. Posterior views of the skulls of two anoles illustrating differences in the extension of the supratemporal processes of the parietal (character 50). (a) Anolis heterodermus, MCZ 110133 (state 0); (b) Anolis agassizi, MCZ 27120 (state 1). Scale bar = 5 mm. Abbreviations: st-par, supratemporal processes of the parietal. cartilage and forms the articular condyle, but posterior border is the point used for comparison. neither the retroarticular nor the angular process States: posterior border of dentary is anterior to (de Queiroz, 1987). No intraspecific variation was mandibular fossa (0) or within mandibular fossa observed. States: absent (0), present (1). (1). Polymorphic character. 53. Position of posteriormost tooth with respect to 56. Position of surangular foramen (Frost and the combined alveolar-mylohyoid foramen (camf; Etheridge, 1988, character 19, fig. 3; Poe, 1998, modified from de Queiroz, 1987, characters 34, character 115, fig. 13). The surangular foramen 35; Poe, 1998, character 109; 2004, character 81). (on the lateral surface of the manelibles same as Etheridge (1959) reported that in some iguanids Poe’s [2004] supra-angular foramen) may be (e.g., anoles) the anterior mylohyoid foramen located entirely within the surangular (0) or be (amf, usually located within the splenial) is united partially bordered by the dentary (1). Differences with the anterior inferior alveolar foramen (aiaf, between left and right sides were observed in located between the dentary and _ splenial), some specimens; therefore, each side was treated resulting in a single foramen. In the present separately for frequency calculations. Polymor- study, the single foramen is called the combined phic character. alveolar-my ohiy oid foramen (camf). Poe (1998, 57. Presence or absence of splenial bone (Etheridge, 2004) compared the position of the posteriormost 1959: Poe, 2004, character 85). States: absent (0), tooth to the amf, which is the same as this present as anteromedial sliver (1), or present and character. We compared the position of the large, as in Polychrus and other non-anole posterior edge of the posteriormost tooth to the iguanids (2). No intraspecific variation was camf, and considered three states: posteriormost anon ed. Ordered. tooth is anterior to camf (0), overlaps with camf 58. Presence or absence of angular bone (Etheridge, (1), posteriormost tooth is posterior to camf (2). 1959). States: absent (0), present (1). No Left and right mandibles were coded separately. intraspecific variation was observed. Polymorphic character. Unordered. 59. Overlap between clavicles and lateral processes of 54. Shape of the posterior suture of dentary (Poe, interclavicle (modified from Etheridge, 1959). 1998, character 111, fig. 12). In lateral view, the Etheridge (1959) described two different types of suture of the dentary with the surangular may interclavicles in anoles: arrow-shaped (in which the have a distinctly pronged (i.e., with two process- lateral processes of the interclavicle are caudolat- es) or a blunt, undifferentiated shape. No erally directed and only medially overlapped by the intraspecific variation was observed. States: clavicle) or T- shaped (in which the lateral process- pronged (0), blunt (1). es are laterally directed and broadly overlapped by 55. Position of posterior suture of dentary, relative to the clavicle). The two components of the inter- mandibular fossa (Poe, 1998, character 112). clavicle shape, as described by Etheridge (1959), Given the possible shape of this suture (blunt can vary independently; therefore, this character or pronged), the anteriormost aspect of the was divided into two. The first was quantified as PHYLOGENY OF THE DacryLoa ¢ Castaneda and de Queiroz 393 Figure 11. overlapped distance; tl, Ventral view of the pectoral girdle illustrating details on measurements of the overlap between clavicles and the lateral processes of the interclavicle (character 59). Scale bar = 5 mm. Abbreviations: cl, clavicle; icl, interclavicle; od, total length of lateral process of interclavicle. 60. 61. the fraction of the length of the lateral process of the interclavicle in direct contact with (i. = overlapped by) the clavicle. The total length o the lateral process was measured as a straight 7 from the midline of the interclavicle (an imaginary line along the long axis of the median [posterior] process) to the tip of the lateral process (Fig. 11). The overlapped distance was measured along the same straight line. Length measurements were made on photogr aphs of “dry or clear and stained interclavicles using the software Mac Morph (Spen- cer and Spencer, 1993). Two measurements were made on each side (left and right sides separately) and used to calculate the average per species. Continuous character. Range: 0.36—0.96. Angle between the median (posterior) process and the lateral process of the interclavicle (this is the second character derived from the arrow- shaped and T-shaped conditions of Etheridge [1959] described in the previous character). The angle was measured between the long axis of the median process and that of the lateral process (as described in the previous character; ie 12) on photographs of dry or cleared and stained interclavicles, using the software MacMorph (Spencer and Spencer, 1993). Two measure- ments were made on each side (left and right sides separately) and used to calculate the average per species. Continuous character. Range: 44.9-64.5. Postxiphisternal inscriptional rib formula (Ether- idge, 1959). The postxiphisternal inscriptional ribs are the cartilaginous ventral rib elements located Figure 12. Ventral view of the pectoral girdle illustrating details on measurements of the angle between the median (posterior) process and the lateral process of the interclavicle (character 60). Abbreviations: Ip-icl, lateral process of inter- clavicle; mp-icl, mediai process of interclavicle. 62. 64. 65: caudal to the xiphisternum. The first number in the formula refers to the number of such ribs attached to the (ossified) dorsal ribs; the second refers to the number of floating (unattached) postxiphisternal inscriptional ribs caudal to the attached ones. The modal condition was scored for each ee ae CO) 222 (O21 2 40 (Saal: (4) 5:0, (5) 5:1, (6) 5:2, (7) 8:1. This character was ordered oe. a step matrix (modified from Jackman et al., 1999), in which the gain or loss of a rib or a connection (from attached to floating or vice versa) costs one step. Number of presacral vertebrae (Etheridge, 1959). The presacral vertebrae are a vertebrate anterior to the sacrum. States: (0) Cy Dee 12) DA. Wo) Lo. Ne) 2a: Ce es character. Ordered. Number of lumbar vertebrae (Etheridge, 1959). The lumbar vertebrae are post-thoracic vertebrae (i.e., those that are not attached directly or indirectly to the sternum) that bear no _ ribs. States: (0) 1, (1) 2, (2) 3, (3) 4, (4) 5. Polymorphic character. Ordered. Type of caudal vertebrae (Etheridge, 1959). Caudal vertebrae mi wy be of the alpha type (0), in which the transverse processes are caudolat- erally or laterally directed and present only on the most anterior vertebrae (7-15), or the beta type (1), in which transverse ee sses are present much farther posteriorly in the c audal sequence, where they are directed craniolater- ally. No instraspecific variation was observed. Caudal autotomy septa (Etheridge, 1959). Autot- omy septa are observed in radiographs as unossi- 394 Bulletin of the Museum of Comparative Zoology, Vol. 160, No. 7 fied areas in the vertebrae anterior, posterior, or through the transverse process. The anteriormost autotomy septum usually coincides with a change in the condition of the transverse process (e.g., disappearance, change in direction, or appearance of a second pair; Etheridge, 1959). This character exhibits ontogenetic var iation, as in some species, progressive fasion of the septa occurs from caudal to cranial with age (Etheridge, 1959). To account for this variation, three states were recognized: septa absent in all specimens, representing those species that do not have (at any life stage) autotomy (O), septa present in some specimens and absent in others, representing those species that progres- sively lose autotomy with age (1), or septa present in all specimens, representing those species that retain autotomy throughout life Nprouded that large individuals were Serial | 2). This coding approach is strongly biased by eae size but was the best found fal incorpor ate information on the gradual change of the character. Ordered. 66. Mean number of caudal vertebrae bearing transverse processes (Etheridge, 1959). Trans- verse processes are always present on the anteriormost caudal vertebrae and progressively decrease in size posteriorly until absent. Contin- uous character. Range: 6. 832.9. LITERATURE CITED ALFARO, M. E., S. ZOLLER, AND F. LuTzoni. 2003. Bayes or bootstrap? A simulation study comparing the performance of Bayesian Markov chain Monte Carlo sampling and bootstrapping in assessing phylogenetic confidence. Molecular Biology ana Evolution 20: 255-266. AROSEMENA, F. A., R. TRARE D., AND F. DE Sousa. 1991. Una nueva especies de Anolis (Squamata: Iguani- dae) del grupo latifrons de Fortuna, Panama. Revista de Biologia Tropical 39: 255-262. AYALA, S., D. Harris, AND E.. E. WILLIAMS. 1983. New or problematic Anolis from Colombia. I. Anolis calimae, new species, from the cloud forest of western Colombia. Breviora 475: 1-11. AyaLa, S. C., AND F. Castro, Familia Iguanidae—II Anolinos. In S. C. Ayala and F. Gace (eds) Saurios de Colombia. Unpublished manuscript. AYALA, S. C., D. M. Harris, AND E.. E. WiLtiaMs. 1984. Anolis menta, sp. n. (Sauria, Iguanidae), a new tigrinus group anole from the west side of the Cane Marta Mountains, Colombia. Papéis Avulsos de Zoologia 35: 135-145. AYALA-VARELA, F. P., AND O. TorRES-CarvajAL. 2010. A new species of dactyloid anole (Iguanidae, Poly- chrotinae, Anolis) from the southeastern slopes of the Andes of Ecuador. ZooKeys 53: 59-73. AYALA-VARELA, F’. P., AND J. A. VELAsco. 2010. A new species of dactyloid anole (Squamata: Iguanidae) from the western Andes of Ecuador. Zootaxa 2577: 46—56. Barsour, T. 1920. A note on Xiphocercus. Proceedings of the New England Zoological Club 7: 61-63. Barsour, T. 1923. Notes on reptiles and amphibians fom Panama. Occasional Papers of the Museum of Zoology, University of Michigan 129: 1-16. Barros, T., E. E. WILLIAMS, AND A. Vitoria. 1996. The genus Phenacosaurus (Squamata: Iguania) in western Venezuela: Phenacosaurus tetarii, new species, Phenacosaurus euskalerriari, new species, and Phe- nacosaurus nicefori Dunn, 1944. Breviora 504: 1-30. BERLOCHER, S. H., anp D. L. Sworrorp. 1997. Searching for phylogenetic trees under the fre- quency parsimony criterion: an approximation using generalized parsimony. Systematic Biology 46: 211-215. BERNAL Caro, A., AND J. A. Roze. 2005. Lizards of the genus Anolis (Reptilia: Polychrotidae) from Sierra Nevada de Santa Marta, Colombia, with descrip- tion of two new species. Novedades Colombianas 8: 9-26. BraNDLEY, M. C., AND K. DE QueEtRoz. 2004. Phylogeny, ecomorphological evolution, and historical bioge- ography of the Anolis cristatellus group. Herpeto- logical Monographs 18: 90-126. Bryant, D. 2003. A classification of consensus methods for phylogenetics, pp. 1-21. In M. Janowitz, F. J. Lapointe, F. McMorris, B. Mirkin, and F. Roberts (eds.), BioConsensus. Piscataway, New Jersey: Aenean Mathematical Society Publications—DI- MACS Series in Discrete Mathematics and Theo- retical Computer Science. BuRNELL, K. L., anpD S. B. HEpcEs. 1990. Relationships of West Indian Anolis (Sauria: Iguanidae): an approach using slow-evolving protein loci. Carib- bean Journal of Sciences 26: 7-30. BuTuer, M. A., S. A. SAWYER, AND J. B. Losos. 2007. Sexual dimorphism and adaptive radiation in Anolis lizards. Nature 447: 202-205. Buter, M. A., T. W. SCHOENER, AND J. B. Losos. 2000. The relationship between sexual size dimorphism and habitat use in Greater Antillean Anolis lizards. Evolution 54: 259-272. CANNATELLA, D., AND K. DE QuEtRoz. 1989. Phyloge- netic sy cremate of the Anoles: is a new taxonomy warranted? Systematic Zoology 38: 57-69. CasTANEDA, M. d. R., AND K. DE Queiroz. 2011. Phylogenetic relationships of the Dactyloa clade of Anolis lizards based on nuclear and mitochon- drial DNA sequence data. Molecular Phylogenetics and Evolution 61: 784-800. Corr, E. D. 1899. Contributions to the herpetology of New Grenada and Argentina, with descriptions of new forms. Philadelphia Museums Science Bulletin jie vil. CreER, D. A., K. DE Quetroz, T. R. Jackman, J. B. Losos, AND A. Larson. 2001. Systematics of the Anolis roquet series of the southern Lesser Antilles. Journal of Herpetology 35: 428-441. DE Quetroz, K. 1987. Phylogenetic systematics of Iguanine lizards. A comparative osteological study. PHYLOGENY OF THE DactyLoa ¢ Castaneda and de Queiroz 395 University of California Publications in Zoology 118: 1-203. DE Quetroz, K. 1996. Including the characters of interest during tree reconstruction and the problems of circularity and bias in studies of character evolution. American Naturalist 148: 700-708. yf QuerRoz, K. 2000. Logical problems associated with including and excluding characters during tree reconstruction and their ‘implications for the: study of morphological character evolution, pp. 192— 212. In J. J. Wiens (ed.), Phylogenetic Morphological Data. Washington, DC: Smithso- nian Institution Press. E Quetroz, K., AND M. DonocuueE. 2013. Phylogenetic nomenclature, hierarchical information, and test- ability. Systematic Biology. doi: 10.1093/sysbio/ sys054 bE QuerRoz, K., AND J. Gauruier. 1990. Phylogeny as a central principle in taxonomy: phylogenetic defini- tions of taxon names. Systematic Zoology 39: 307— Say DE QuEoz, K., aND J. Gautuier. 1992. Phylogenetic taxonomy. Annual Review of Ecology and System- atics 23: 449-480. DE QUEIROZ, K., AND T. REEDER. 2008. Squamata— Lizards, pp. 24-45. In B. I. Crother (ed.), Scientific and Standard English Names of Amphibians and Reptiles of North “America North of Mexico, with Comments Regarding Confidence in Our Under- standing. Society for qhic Study of Amphibians and Reptiles, Herpetological Circular No. 37. DRUMMOND, A. J., AND A. RampauT. 2007. BEAST: Bayesian evolutionary analysis by sampling trees. BMC Evolutionary Biology 7: 2 Dunn, E. R. 1937. The giant mainland Anoles. Proceedings of the New England Zoological Club 16: 5-9. Dunn, E. R. 1939. Zoological results of the George Vanderbilt South Pacific Expedition of 1937. Part II—The lizards of Malpelo Island, Colombia. Notulae Naturae 4: 1-3. Dunn, E. R. 1944. The lizard genus Phenacosaurus. Caldasia 3: 57-62 Estes, R., K. DE QUEIROZ, AND J. GAUTHIER. 1988. Phylogenetic relationships within Squamata, pp. 119- 981. In R. Estes and G. Pregill (eds.), Phylogenetic Relationships of the Lizard Fami- lies —Essays Commemorating Charles. L. Camp. Stanford, California: Stanford Univ. Press. Erueripce, R. 1959. The relationships of the Anoles (Reptilia: Sauria: Iguanidae): an interpretation based on skeletal morphology. Ph.D. Thesis. Ann Arbor, Michigan: Univ. of Michigan. ETHERIDGE, R., AND K. DE QuErIROz. 1988. A phylogeny of Iguanidae, pp. 283-367. In R. Estes and G. Pregill (eds.), Phylogenetic Relationships of the Lizard Families—Essays Commemorating Charles L. Camp. Stanford, California: Stanford Univ. Press. FELSENSTEIN, J. 1985. Confidence limits on phylogenies: an approach using the bootstrap. Evolution 39: 783-791. —s Analysis of FINDEN, C. R., anD A. D. Gorpon. 1985. Obtaining common pruned trees. Journal of Classification 2: 955=2,76. Fisuer, R. A. 1930. The moments of the distribution for normal samples of measures of departure from normality. Proceedings of the Royal Society A, 130: 16-28. Frost, D. R., AND R. Ernertpce. 1989. A phylogenetic analysis aiid taxonomy of iguanian lizards (Reptilia: Squamata). Miscellaneous “Publications Univ ersity of Kansas Natural History Museum 81: 1-65. GIANNASI, N., R. S. THORPE, AND A. MALHOTRA. 2000. A phylogenetic analysis on body size evolution in the Anolis roquet group ( Sauria: Iguanidae): character displacement or size assortment? Molecular Ecol- ogy 9: 193-202. Gtor, R. E., J. J. KOLBE, R. PowELL, A. LARSON, AND J. B. Losos. 2003. Phylogenetic analysis of ecological and morphological diversification in Hispaniolan trunk-ground anoles (Anolis cybotes group). Evo- lution 57: 2383-2397. xORDON, A. D. 1980. On the assessment and compar- ison of classifications, pp. 149-160. In R. Tomas- sone (ed.), Analyse de Donnees et Informatique. Le Chesnay, France: INRIA. Gorman, G. C. 1968. The relationships of Anolis of the roquet species group (Sauria: Iguanidae). III. Comparative study of display Relwion Breviora 284: 1-31. Gorman, G. C., anp L. Atkins. 1967. The relationships of Anolis of the roquet species group (Sauria: Iguanidae). II. Comparative chromosome cytology. Systematic Zoology 16: 137-143. Gorman, G. C., AND L. Atkins. 1969. The zoogeography of Lesser Antillean Anolis lizards—an analysis based upon chromosomes and lactic dehydroge- nases. Bulletin of the Museum of Comparative Zoology 138: 53-80. GormaN, G. C., D. G. Buty, M. SOULE, AND S. Y. YANG. 1980a. The relationships of the Anolis cristatellus group: electrophoretic analysis. Journal of Herpe- tology 14: 269-278. Gorman, G. C., D. G. Buty, M. SouLg, AND S. Y. YANG. 1983. The relationships of the Puerto Rican Anolis: electrophoretic and karyotypic studies, pp. 626-— 642. In A. G. J. Rhodin and K. Miyata (eds.), Advances in Herpetology and Evolutionary Biols: gy. Cambridge, Massachusetts: Museum of Com- parative Zoology. Gorman, G. C., D. G. ButH, AND J. S. WyLEs. 1980b. Anolis lizards of the eastern Caribbean: a case study in evolution. HI. A cladistic analysis of albumin immunological data, and the definitions of species groups. Systematic Zoology 29: 143-158. Gorman, G. C., AND H. C. Dessauer. 1965. Hemoglobin and transferring electrophoresis and relationships of island populations of Anolis lizards. Science 150: 1454-1455. Gorman, G. C., and H. C. Dessauver. 1966. The relationships of Anolis of the roquet species group (Sauria: Iguanidae). I. Electrophoretic comparison 396 of blood proteins. Comparative Biochemistry and Physiology 19: 845-853. Gorman, G. C., anD Y. J. Kim. 1976. Anolis lizards of the eastern Caribbean: a case study in evolution. II. Genetic relationships and genetic variation of the bimaculatus group. Systematic Zoology 20: 167-185. Gorman, G. C., P. Licut, H. C. DEssAuER, AND J. O. Boos. 1971. Reproductive failure among the hybridizing Anolis lizards of Trinidad. Systematic Zoology 20: 1-18. Gorman, G. C., C. S. Lies, AnD R. H. Harwoop. 1984. The relationships of Anolis gadovi: albumin immunological evidence. Caribbean Journal of Science 20: 145-152. Gorman, G. C., AND R. B. Stamm. 1975. The Anolis lizards of Mona, Redonda, and La Blanquilla: chromosomes, relationships, and natural history notes. Journal of Herpetology 9: 197-205. Gorman, G. C., R. THomas, AND L. ATKINS. 1968. Intra- and interspecific chromosome variation in Anolis cristatellus and its closest relatives. Breviora 293: 1-12. Guyer, C., aND J. M. Savace. 1987 [1986]. Cladistic relationships among anoles (Sauria: Iguanidae). Systematic Zoology 335: 509-531. Hitus, D. M. 1991. Discriminating between phyloge- netic signal and random noise in DNA sequences, pp. 278-294. In M. M. Miyamoto and J. Cracraft (eds.), Phylogenetic Analyses of DNA Sequences. New York: Oxford Univ. Press. Hitus, D. M., anp J. P. HuUELSENBECK. 1992. Signal, noise and reliability in molecular phylogenetic analyses. Journal of Heredity 83: 189-195. HUELSENBECK, J. P. 1991. Tree-length distribution skewness: an indicator of phylogenetic information. Systematic Zoology 40: 257-270. HUELSENBECK, J. P., F. Ronquist, R. NIELSEN, AND J. P. Bo.iBack. 2001. Bayesian inference of phylogeny and its impact on evolutionary biology. Science 294: 2310-2314. HuLesak, E., S. Por, R. IBANEz, AND E.. E. WILLIAMS. 2007. A striking new species of Anolis lizard (Squamata, Iguania) from Panama. Phyllomedusa 6: 5—10. Jackman, T. R., D. J. Inscuicx, K. DE QueEtRoz, J. B. Losos, aND A. Larson. 2002. Molecular phyloge- netic perspective on evolution of lizards of the Anolis grahami series. Journal of Experimental Zoology 294: 1-16. JACKMAN, T. R., A. Larson, K. DE QUEIROZ, AND J. B. Losos. 1999. Phylogenetic relationships and tempo of early diversification in Anolis lizards. Systematic Biology 48: 254-285. Kass, R. E., anp A. E. Rarrery. 1995. Bayes factors. Journal of the American Statistical Association 90: T13=195. LarGeT, B., anpD D. L. Simon. 1999. Markov chain Monte Carlo algorithms for the Bayesian analysis of phylogenetic trees. Molecular Biology and Evolution 16: 750-759. Bulletin of the Museum of Comparative Zoology, Vol. 160, No. 7 Larson, A. 1998. The comparison of morphological and molecular data in phylogenetic systematics, pp. 275-296. In R. DeSalle and B. Schierwater (eds.), Molecular Approaches to Ecology and Evolution. Basel, Switzerland: Birkhauser Verlag. LazeLL, J. D., Jr. 1969. The genus Phenacosaurus (Sauria: Iguanidae). Breviora 325; 1-24. LazeLt, J. D., Jk. 1972. The anoles (Sauria: Iguanidae) of the Lesser Antilles. Bulletin of the Museum of Comparative Zoology 143: 1-115. Ler, M. S. Y. 2000. Tree robustness and clade significance. Systematic Biology 49: 29-836. Lewis, P. O. 2001. A likelihood approach to estimating phylogeny from discrete morphological character data. Systematic Biology 50: 913-925. Lewis, P. O., M. T. HOLDER, AND K. E. HOLsINGER. 2005. Polytomies and Bayesian phylogenetic infer- ence. Systematic Biology 54: 241-253. Losos, J. B. 2009. Lizards in an Evolutionary Tree: Ecology and Adaptive Radiation of Anoles. Berke- ley and Los Angeles, California: Univ. of California Press. Losos, J. B., M. L. Woottey, D. L. MaHLer, O. TorrEs- CarvajJAL, K. E. CRANDELL, E. W. ScHAAD, A. E. NARVAEZ, F. AYALA-VARELA, AND A. HERREL. 2012. Notes on the natural history of the little-known Ecuadorian homed anole, Anolis proboscis. Bre- viora 531: 1-17. Mappison, D. R., anD W. P. Mappison. 2001. MacClade: Analysis of Phylogeny and Character Evolution. Version 4.03. Sunderland, Massachu- setts: Sinauer Associates. Mappison, W. P., aND D. R. Mappison. 2011. Mesquite: A Modular System for Evolutionary Analysis. Version 2.75. Available from: http:// mesquiteproject.org. Myers, C. W., aNnD M. A. DONNELLY. 1996. A new herpetofauna from Cerro Yavi, Venezuela: first results of the Robert G. Goelet American Muse- um—TERRAMAR expedition to the northwestern Tepuis. American Museum Novitates 3172: 1-56. Myers, C. W., E. E. WILtiaMs, AND R. W. McDIARMID. 1993. A new anoline lizard (Phenacosaurus) from the highland of Cerro de la Neblina, southern Vene- zuela. American Museum Novitates 3070: 1-15. NICHOLSON, K. E. 2002. Phylogenetic analysis and a test of the current infrageneric classification of Norops (Beta Anolis). Herpetological Monographs 16: 93- 120. NICHOLSON, K. E., R. E. Gior, J. J. KotBe, A. Larson, S. B. HepcEs, AND J. B. Losos. 2005. Mainland colonization by island lizards. Journal of Biogeog- raphy 32: 929-938. PaceL, M., aNnp A. Meape. 2005. Mixture models in phylogenetic inference, pp. 121-139. In O. Gascuel (ed.), Mathematics of Evolution and Phylogeny. Oxford: Oxford Univ. Press. Peters, J. A. 1964. Dictionary of Herpetology: A Brief and Meaningful Definition of Words and Terms Used in Herpetology. New York: Hafner VII. PHYLOGENY OF THE DacryLoa * Castarieda and de Queiroz 397 PETERS, J. A., AND V. G. Orcs. 1956. A third leaf-nosed species of the lizard genus Anolis from South America. Breviora 62: 8: Por, S. 1998. Skull characters and the cladistic relationships of the Hispaniolan dwarf twig Anolis. Herpetological Monographs 12: 192-236. Por, S. 2004. Phylogeny of anoles. Herpetological Monographs 18: 37-89. Por, S. 2005. A study of the utility of convergent characters for phylogeny reconstruction: do eco- morphological characters track evolutionary history in Anolis lizards? Zoology 108: 337-343. Pog, S., F. AyaLa, I. M. LaTetia, T. L. KENNEpy, J. A. CHRISTENSEN, L. N. Gray, N. J. BLEa, B. M. Armyo, ‘ND E. W. Scuaab. 2012. Morphology, phylogeny, and behavior of Anolis proboscis. Breviora 530: 1-11. Por, S., AND R. IBANEz. 2007. A new species of Anolis lizard from the cordillera de Talamanca of western Panama. Journal of Herpetology 41: 263-270. Por, S., I. M. LaTEtua, M. J. RYAN, AND E. W. SCHAAD. 2009a. A new species of Anolis lizard (Squamata, Iguania) from Panama. Phyllomedusa 8: 81-87. Por, S., J. VeLAsco, K. Miyata, AND E. E. WILLIAMS. 2009b. Descriptions of two nomen nudum species of Anolis lizard from northwestern South America. Breviora 516: 1-16. PoE, S., AND C. YANEZ-MIRANDA. 2007. A new species of phenacosaur Anolis from Peru. Herpetologica 63: ZN9G= 223: S., AND C. YANEZ-MiraAnpDA. 2008. Another new species of green Anolis (Squamata: Iguania) from the eastern Andes of Peru. Journal oft Herpetology AD 564-571) Pog, S., C. YANEZ-MIRANDA, AND E. LEHR. 2008. Notes on variation in Anolis boettgeri Boulenger 1911, assess- ment of the status of ae Alena Henle and Ehlr 1991, and description of a new species of Anolis (Squamata: Iguania) similar to Anolis boettgeri. Journal of Herpetology 42: 251-259. Posapa, J. L., AND K. A. CraNpDALL. 1998. Modeltest: testing the model of DNA substitution. Version 3.7 (June 2005). Bioinformatics 14: 817-818. Ropricues, M. T. 1988. A new anole of the punctatus group from central Amazonia (Sauria: Iguanidae). Papeis Avulsos de Zoologia 36: 333-336. Ropricues, M. T., V. Xavier, G. Skuk, AND D. Pavan. 2002. New specimens of Anolis phyllorhinus (Squamata, Polychrotidae): the first female of the species and of proboscid anoles. Papéis Avulsos De Zoologia 42: 363-380. RONQUIST, im AND J. P. HUELSENBECK. 2003. MrBayes 3: Bayesian phylogenetic inference under mixed models. Bioinformatics 19: 1572-1574. Ruepa, J. V., AND J. I. HERNANDEZ Camacuo. 1988. Bp enecosanms inderenae (Sauria: Iguanidae), nueva especie gigante, proveniente de la cordillera oriental de Colombia. Trianea 2: 339-350. RUuEDA ALMONACID, J. V. 1989. Un nuevo y extraordi- nario saurio de color rojo (Iguanidae: Anolis) para la Cordillera Occidental de ‘Colombia. Trianea 3: 85-92. PoE, SavaGE, J. M. 2002. The Amphibians and Reptiles of Costa Rica. Herpetofauna between Two Conti- nents, between Two Seas. Chicago: Univ. of Chicago Press. SAVAGE, J. M., and C. Guyer. 1989. Infrageneric classification and species composition of fhe anole genera, Anolis, Ctenonotus, Dactyloa, Norops, and Semiurus (Sauria: Iguanidae). Amphibia-Reptilia 10: 105-116. SAVAGE, J. M., AND J. J. TaBor. 1978. The giant anoline lizards of Costa Rica and western Panama. Copeia 1978: 480-492. SCHNEIDER, C. J., J. B. Losos, AND K. DE QuEtRoz. 2001. Evolutionary relationships of ie yeah bimacula- tus group ‘from the northern Lesser Antilles. Journal of Herpetology 35: 1-12. SCHOENER, T. W. 1969. Size patterns in West Indian Anolis lizards: I. Size and species diversity. Systematic Zoology 18: 386-401. SHOCHAT, D., AND H. C. DessauErR. 1981. Comparative study of albumins of Anolis lizards of the Caribbean islands. Comparative Biochemistry and Physiology 68A: 67-73. SOKAL, R. R., anpD F. J. Routr. 1995. Biometry: The Principles and Practice of Statistics in Biological Research. 3rd ed. San Francisco: W.H. Freeman. SPENCER, M. A., AND G. S. SPENCER. 1993. MacMorph Image Analy sis Software, Version 2.1. Stony Brook, New York. Suzuki, Y., G. V. GLazKo, AND M. NE. 2002. Over- credibility of molecular phylogenies obtained by Bayesian phylogenetics. Proceedings of the Nation- al Academy of Science 99: 16138-16143. SworrorD, D. L. 1991. When are phylogeny estimates from molecular and morphological data incongru- ent? pp. 295-333. In M. M. Miyamoto and J. Cracraft (eds.), Phylogenetic Analysis of DNA Sequences. New York: Oxford Univ. Press. SworrorbD, D. L. 2002. PAUP*. Phylogenetic Analysis Using Parsimony (* and Other Methods). Version 4, Sianderland: Massachusetts: Sinauer Associates. SWOFFORD, D. L., AND S. H. BERLOCHER. 1987. Inferring evolutionary trees from gene frequency data under the principle of maximum parsimony. Systematic Zoology 36; 293-325. TEMPLETON, A. R. 1983. Phylogenetic inference from restriction endonuclease cleavage site maps with particular reference to the ev lugar of humans and apes. Evolution 37: 221-244. THIELE, K. 1993. The holy grail of the perfect character: the cladistic treatment of morphometric data. Cladistics 9: 275-304. TorRES-CARVAJAL, O. 2007. Phylogeny and biogeogra- phy of a large radiation of dean lizards ( Touania, Stenocercus). Zoologica Scripta 36: 311- 326. Uetz, P. (Ep.) [Internet]. The Reptile Database [cited 2012 Mar 10]. Available from: http:/Avww.reptile- database.org/. UcuETo, G. N., G. Rivas, T. Barros, AND E. N. SMITH. 2009. A revision of the Venezuelan anoles II: redescription of Anolis squamulatus Peters 1863 398 and Anolis tigrinus Peters 1863 (Reptilia: Poly- chrotidae). Caribbean Journal of Science 45: 30- bis Ucueto, G. N., G. Rivas-FuENmayor, T. Barros, S. J. SANCHEZ-PACHECO, AND J. E. Garcia-PEREZ. 2007. A revision of the Venezuelan Anoles I: a new Anolis species from the Andes of Venezuela with the redescription of Anolis jacare Boulenger 1903 (Reptilia: Polychrotidae) and the eben of the status of Anolis nigropunctatus Williams 1974. Zootaxa 1501: 1-30. UnperRwoop, G. 1959. The anoles of the eastern Caribbean (Sauria, Iguanidae): revisionary notes. Bulletin of the Museum of Comparative Zoology 121(part III): 191-226. VELasco, J. A., P. D. A. GUTIERREZ-CARDENAS, AND A. QUINTERO-ANGEL. 2010. A new species of Anolis of the aequatorialis group (Squamata: Iguania) from the central Andes of Colombia. Herpetological Journal 20: 231-236. ; Wiens, J. J. 1995. Polymorphic characters in phyloge- netic systematics. Systematic Biology 44; 482-500. Wiens, J. J. 1998. Testing phylogenetic methods with tree congruence: phy logenetic analysis of polymor- phic morphological elnractets in phrynosomatid lizards. Systematic Biology 47: 427-444. Wiens, J. J. 2000. Coding morphological variation within species and higher taxa for Ponce cae analysis, pp. 115- 145. In J. J. Wiens (ed.), Phylogenetic Analysis of Morphological Data. Washington, DC: Smithsonian Institution Press. WIENS, J. J. 2001. Character analysis in morphological phylogenetics: problems and ‘solutions. St ystematic Biology 50: 689-699. Wiens, J. J.. AND M. R. Servepio. 1997. Accuracy of phylogenetic analysis including and excluding polymorphic characters. Systematic Biology, AG: Soo os. Wits, E. E. 1965. South American Anolis (Sauria, Iguanidae): two new species of the punctatus group. Breviora 233: 1-15. Witurams, E. E. 1974. South American Anolis: three new species related to Anolis nigrolineatus and A. dissimilis. Breviora 422: 1-15. Witurams, E. E. 1975. South American Anolis: Anolis parilis, new species, near A. mirus Williams. Breviora 434: 1-S. Wiis, E. E. 1976a. West Indian anoles: a taxonomic and evolutionary summary. 1. Introduction and a species list. Breviora 440: 1-21. Witurams, E. E. 1976b. South American anoles: the species groups. Papéis Avulsos de Zoologia 29: 259-268. Wituiams, E. E. 1979. South American anoles: the species groups. 2. The proboscis anoles (Anolis laevis group). Breviora 449: 1-19. Wituias, E. E. 1982. Three new species of the Anolis punctatus complex from Amazonian and _inter- Andean Colombia, with comments on the eastern Bulletin of the Museum of Comparative Zoology, Vol. 160, No. 7 members of the punctatus species group. Breviora 467: 1-38. Wituiams, E. E. 1985. New or problematic Anolis from Colombia. IV. Anolis antioquiae, new species of the Anolis eulaemus subgroup from western Colombia. Breviora 482: 1-9. Wituias, E. E. 1988. New or problematic Anolis from Colombia. V. Anolis danieli, a new species of the latifrons species group and a reassessment Anolis apollinaris Boulengeri, 1919. Breviora 489: 1-25. Wituas, E. E. 1989. A critique of Guyer and Savage (1986): cladistic relationships among anoles (Sauria: Iguanidae): are the data available to reclassify anoles? pp. 434-478. In C. A. Woods (ed.), Biogeography of the West Indies: Past, Present and Future. Gainsville, Florida: Sandhill Crane Press. Wituiams, E. E. 1992. New or problematic Anolis from Colombia. VII. Anolis lamari, a new anole from the cordillera oriental of Colombia, with a discussion of tigrinus and punctatus species groups boundaries. Breviora 495: 1-24. Wiuuams, E. E., anp N. Acosta B. 1996. The Anolis Handlist. Version 1.0b0. Cambridge, Massachu- setts: Museum of Comparative Zoology. WituraMs, E. E., anp W. E. DuELiMaANn. 1967. Anolis chocorum, a new punctatus-like anole from Darien, Panama (Sauria, Iguanidae). Breviora 256: 1-12. WituraMs, E. E., anp W. E. DugELLMan. 1984. Anolis fitchi, a new species of the Anolis aequatorialis group from Ecuador and Colombia, pp. 257-266. In R. A. Siegler, L. E. Hunt, J. L. Knight EL. Rae de - bs Bae | (eds.), Vertebrate g A Tribute 6 Henry S. Fitch. pide Kansas University of Kansas, Museum of Natural History Special Publication No. 10. WILuiAMsS, E. E., AND R. MITTERMEIER. 1991. A Peruvian phenacosaur (Squamata: Iguania). Breviora 492: i=1G: WituiaMs, E. E., M. J. PRADERIO, AND S. GorzuLa. 1996. A phenacosaur from Chimanta Tepui, Venezuela. Breviora 506: 1-15. WituiaMs, E. E., H. Ranp, A. S. RaNp, anp R. J. O'Hara. 1995. A computer approach to the comparison and identification of species in difficult taxonomic groups. Breviora 502: 1-47. YANEZ-Munoz, M. H., M. A. UrcILEs, M. ALTAMIRANO- BENAVIDES, AND S. R. CAcERES. 2010. Redescrip- cin de Anolis proboscis Peters & Orcés (Reptilia: Polychrotidae), con el descubrimiento de las hembras de la especie y comentarios sobre su distribuci6n y taxonomia. Avances 2: B7—B15. YANG, S. Y., M. SOULE, AND G. C. Gorman. 1974. Anolis lizards of the eastern Caribbean: a case study in evolution. I. Genetic relationships, phylogeny, and colonization sequence of the roquet group. Sys- tematic Zoology 23: 387-399. Zar, J. H. 1999. Biostatistical Analysis. 4th Ed. Upper Saddle River, New Jersey: Prentice Hall. / F f = } = | 1 iS .s ” : - * Me ; ne may P ay (ty Mine GO e , ae Ae het, i Species names and credits for the photographs included on the front cover Clockwise, starting from upper right corner: Anolis agassizi, Maria Margarita Womack Anolis punctatus, Roy W. McDiarmid Anolis aequatorialis, D. Luke Mahler Anolis trinitatis, Kevin de Queiroz Anolis heterodermus, Juan L. Parra US ISSN 0027-4100 MCZ Publications Museum of Comparative Zoology Harvard University 26 Oxford Street Cambridge, MA 02138 mczpublications@mcz.harvard.edu © The President and Fellows of Harvard College 2013