UNTANGLING A PHYLOGENETIC KNOT OF SERPENTS:

INTEGRATIVE SYSTEMATICS ON NEOTROPICAL SNAKES

by

ROBERT C. JADIN

B.Sc. – Northeastern State University, Tahlequah, OK, 2005

M.Sc. – University of Texas at Tyler, Tyler, TX, 2007

A dissertation submitted to the Faculty of the Graduate School

of the University of Colorado in partial fulfillment

of the requirement for the degree of

Doctor of Philosophy

Department of Ecology and Evolutionary Biology

2013 This thesis entitled: UNTANGLING A PHYLOGENETIC KNOT OF SERPENTS: INTEGRATIVE SYSTEMATICS ON NEOTROPICAL SNAKES Written by Robert Curtis Jadin has been approved for the Department of Ecology and Evolutionary Biology

______Robert P. Guralnick

______Jeffry B. Mitton

______Andrew P. Martin

______Patrick Kociolek

______Herbert H. Covert

Date: 6 May 2013

The final copy of this thesis has been examined by the signatories and we find that both the content and the form meet acceptable presentation standards of scholarly work in the above mentioned discipline.

Jadin, Robert C. (Ph.D., Ecology and Evolutionary Biology)

UNTANGLING A PHYLOGENETIC KNOT OF SERPENTS:

INTEGRATIVE SYSTEMATICS ON NEOTROPICAL SNAKES

Dissertation directed by Robert P. Guralnick

ABSTRACT

Taxonomy based on evolutionary relationships plays a critical role in science and society, as properly named units of biological diversity allows for immediate recognition of the general appearance and basic natural history of a particular lineage. Consequently, it could be argued that systematics, a field that investigates not only the classification of organisms but also their diversity and origins, is among the most important with relevance to all other disciplines within biology. The challenge is documenting species in nature; these units are there to find but doing so requires compiling a variety of different sources of data and methods to test hypotheses about diversity. Through this dissertation, I combine my background in classical taxonomy, natural history, and descriptive morphology with modern molecular phylogenetics, implementing

Bayesian, Likelihood, and Parsimony optimality criteria, to stabilize taxonomy, detect cryptic diversity, and understand snake phylogeny. In particular, I incorporate important guidelines recently defined for integrative taxonomy, that assure accurate classification of biodiversity; these guidelines should become the new standard for all systematic revisions. To assess these guidelines, I have focused on a group and region where taxonomic and systematic knowledge is still very limited. Central and South American snakes are not only a point of continued

iii systematic confusion but also understanding the evolutionary histories of venomous species may provide applications to medical treatment and human health. Using these methods of modern systematics I investigate two lineages of Neotropical snakes whose systematic history begins with classical Linnean taxonomy and progresses over the centuries with the advancement of systematic techniques. Cryptic lineages, in combination with inadequate sampling, obscure the complete diversity of this group. This dissertation incorporates several important sub-sections of systematics including an integrative systematic revision (chapter 2), detection of cryptic species

(chapter 3), and placement of an enigmatic lineage within the tree of life (chapter 4). Overall, this research highlights how these guidelines for integrative taxonomy are paramount to achieving an informative and stable classification system. Systematic studies like this are becoming increasingly important in our current age of rapid biodiversity loss.

iv ACKNOWLEDGMENTS

I thank my advisor, Robert P. Guralnick, for his positive attitude and guidance during my time in

Boulder. Additionally, I thank Jeffry B. Mitton, Andy P. Martin, Patrick Kociolek, and Herbert

H. Covert for serving on my dissertation committee and assisting in my development. Other faculty in the department of ecology and evolutionary biology who were influential include John

Basey, Alexander Cruz, Samuel M. Flaxman, Pieter T.J. Johnson, Nolan C. Kane, and David W.

Stock. I greatly appreciate all of their guidance, wisdom, and advice.

By far, the most important person during my dissertation work has been my wife, Sarah

A. Orlofske. Her guidance, positive influence, and inspiration have greatly enhanced my development as a scientist.

Two people who have been motivating and shaped my view of evolution and science are

Patrik Nosil and Dan Brooks. Patrik taught me a lot about evolutionary biology and how it should be taught. I enjoyed and learned so much in his speciation course as well as being his

T.A. for evolutionary biology. Dan always challenges and improves my perspectives and I greatly appreciate our long conversations philosophically discussing academia, science, and life.

Throughout my time at CU I have learned a lot from my fellow classmates. In particular,

I would like to thank the following for their support through this time: S. Beals, I. Buller, E.

Cheng, A. Comeault, K. Dosch Richgels, L.P. Erb, P. Erb, A. Faist, T. Farkas, S.M. Ferrenberg,

M.B. Joseph, J.R. Mihaljevic, S. Hellmuth Paull, E.L. Paulson, A.C. Peterson, D.L. Preston, J.

Prevey, M.S. Robeson III, L. Sackett, S.L. Stowell, M.R. Wilkins, A.C. Williams, and the post- doc B.A. Goodman.

I thank the numerous collaborators that I have worked with over the years for their tremendous help in making me a better writer, biologist, and collaborator. At the top of this list

v are Eric N. Smith, Gilson A. Rivas, Josiah H. Townsend, Jonathan A. Campbell, and Todd A.

Castoe.

During my time at CU Boulder, parts of my research and travel were funded through several institutions and societies. I sincerely thank the following for the funding they provided:

American Museum of Natural History, American Society of Ichthyologists and Herpetologists,

American Society of Parasitologists, California Academy of Sciences, CU EBIO Department,

CU Graduate School, CU Museum of Natural History, East Texas Herpetological Society,

Ecology and Evolution of Infectious Diseases conference, Society for the Study of Amphibians and Reptiles, and Society of Systematic Biologists. For the last year of my dissertation I was supported through an NSF GK-12 teaching fellowship. The faculty and staff of this program were very supportive and I thank W.D. Bowman, J.A. Feld, B. Graham, G. Ross, and L. Smith for such a great opportunity that has enhanced my ability to teach and conduct scientific outreach.

vi CONTENTS

Abstract …………………………………………………………………………………………iii

Acknowledgments ……………………………………………………………………………….v

List of tables……………………………………………………………………………………..ix

List of figures……………………………………………………………………………………..x

Chapter

1 SLITHERING INTO NEOTROPICAL EVOLUTION: INTEGRATIVE SYSTEMATICS OF NEW WORLD SNAKES

Introduction to systematics………………………………………..………………………………1

Integrative systematics……………………………………………………………...………..……2

Detecting cryptic diversity………………………………………………………………………...4

Placing taxa in the tree of life……………………………………………...... 5

2 UNRAVELLING A TANGLE OF MEXICAN SERPENTS: A SYSTEMATIC REVISION OF HIGHLAND PITVIPERS

Abstract….…………….……………………………………………..……………………………7

Introduction….………………………………….…………………....……………………………8

Materials and methods……………………………….…………………….………..…………….9

Results.….……………………………….…………………….………..………………………..17

Discussion….……………………………….…………………….………..…………………….26

Appendix 2.1...…………………………………………………………………………………...39

vii

3 CRYPTIC DIVERSITY IN DISJUNCT POPULATIONS OF MIDDLE AMERICAN MONTANE PITVIPERS: A SYSTEMATIC REASSESSMENT OF CERROPHIDION GODMANI

Abstract….…………….…………………....……………………………………………………40

Introduction……………………………….…………………….………..………………………41

Methods….……………………………….…………………….………..……………………….42

Results….……………………………….…………………….………..………………………...46

Discussion….…………….…………………....…………………………………………………65

Appendix 3.1…….……………………………………………………………………………….71

4 FINDING ARBOREAL SNAKES IN AN EVOLUTIONARY TREE: PHYLOGENETIC PLACEMENT AND SYSTEMATIC REVISION OF THE NEOTROPICAL BIRDSNAKES

Abstract….…………….…………………....……………………………………………………74

Introduction….…………….…………………....………….…………………………………….75

Methods….……………………………….……………….…….………..………………………76

Results….……………………………….…………………….………..………………………...82

Discussion.……………………………….…………………….………..……………………….85

5 CONCLUSIONS

Introduction……………………………………………………………………………………….89

Future directions….…………….…………………....…………………………………………..90

Integrative taxonomy in modern biology….…………………....………………………………..92

References.…………………….…………………….………..…………………………………94

viii

LIST OF TABLES

Table

2.1 Taxa, vouchers, locality data, and GenBank accession numbers for sequences used in this study. Sequences newly added specifically for this study are in bold…………………………………………………………………13

2.2 Morphological comparisons between Cerrophidion barbouri and Agkistrodon browni………………………………………………………………………18

2.3 Measurements and counts of the type series of Cerrophidion barbouri and Agkistrodon browni………………………………………………………..19

2.4 Pairwise sequence divergences among (below diagonal) and within (diagonal) all New World genera as defined in this study using only cyt b gene………………………………………………………………………………...25

3.1 Taxa, vouchers, locality data, and GenBank accession numbers for sequences used in this study. Sequences newly added specifically for this study are in bold…………………………………………………………………44

3.2 Results from a priori model selections based on Akaike information criterion (AIC) conducted in MrModeltest 2.2 (Nylander, 2004) for partitions of the dataset…………………………………………………………………..45

4.1 GenBank numbers for DNA sequences analyzed in this study not including sequences listed in Table 4.2………………………………………………….77

4.2 Genbank numbers for DNA sequences generated in this study. Abbrevations of institutions and individuals for voucher specimens are as follows: EBRG (Museo de Biologia de la Estacion Biologica Rancho Grande, Maracay, Venezuela), JMR (Julie M. Ray field series), LSUMZ (Louisiana State Museum of Natural Science), USNM (Smithsonian Institution, National Museum of Natural History), UTA (Amphibian and Reptile Diversity Research Center, University of Texas, Arlington)………………………………………………………….80

ix LIST OF FIGURES

Figure

2.1 Agkistrodon browni (A, B; UTA R-56265) and Cerrophidion barbouri (C, D; MZFC 21432) in life, showing differences in head scalation and colour pattern……………………………………………………………...20

2.2 Dorsal view (A, B) and left side view (C, D) of Agkistrodon browni (holotype, MCZ R-42678; left) and Cerrophidion barbouri (holotype, USMN R-46347; right)……………………………………………………….20

2.3 Photo in life of Agkistrodon browni (UTA R-56264) showing its prehensile tail…………………………………………………………………………….21

2.4 Sulcate (left) and asulcate (right) views of the left hemipenis of: A, Agkistrodon browni (UTA R-4450) and B, Cerrophidion barbouri (MZFC 2881). Scale bars = 3 mm……………………………………………..22

2.5 One of three equally parsimonious trees (8165 steps) recovered from heuristic maximum parsimony analysis of 2235 bp of four mitochondrial gene fragments (12S, 16S, cytochrome b, and NADH dehydrogenase subunit 4). Nodal support of posterior probability distributions from a separate Bayesian Markov chain Monte Carlo analysis. Owing to on-going taxonomic change and to help comparing phylogenies the unique ID and locality data for each operational taxonomic unit are provided next to the species names……………………..24

2.6 Distribution map of Mixcoatlus barbouri and Mixcoatlus browni……………………....29

3.1 Bayesian phylogenetic estimate of relationships within Cerrophidion showing three distinct species-level clades within the nonmonophyletic Cerrophidion godmani. The tree represents the Bayesian 50% majority-rule consensus phylogram from a partitioned analysis of sequences from four mitochondrial gene fragments (ND4, cyt b, 12S and 16S; total of 2307 bp). Sample names follow those given in Table 3.1…………………………………………………..48

3.2 Head of Cerrophidion sasai, male holotype, UTA R-51399, showing arrangement of scales. —A. Dorsal, —B. lateral and —C. ventral views……………………………………………………………………….49

3.3 Cerrophidion sasai, male paratype in life, UTA R-51403. UTA slide no. 27134. Photo by Eric N. Smith………………………………………………....50

x 3.4 Head of Cerrophidion wilsoni, holotype, UTA R-52953, showing arrangement of scales. —A. Dorsal, —B. lateral and —C. ventral views……………………………………………………………………….58

3.5 Cerrophidion wilsoni, female holotype in life, UTA R-52953, 648 mm in total length. Photo by Eric N. Smith………………………………………....59

3.6 —A. Distant habitat photograph showing Cerro Azul in Parque Nacional Montaña de Botaderos—the type-locality of Cerrophidion wilsoni. —B. Closer photograph of the mountain where the holotype was collected, showing cloud forest that is occupied with Liquidambar. Photographs by Eric N. Smith…………………………………………………………...59

3.7 Locality map of Cerrophidion samples throughout Middle America used for the molecular analyses. Symbols correspond to the origin of the samples (taken from Castoe et al., 2009 and new localities)……………………...67

3.8 Sulcate (A, C) and asulcate (B, D) views of the left hemipenes of Cerrophidion sasai (holotype, UTA R-51399) and C. wilsoni (paratype, UTA R-59478), respectively………………………………………………….69

4.1 Phylogenetic estimate of relationships among genera and species within the Colubrinae. The tree was estimated from a Bayesian 50% majority-rule consensus composed from a concatenated multigene dataset (ND4, cyt b, and c-mos; total of 2340 bp). Numbers at nodes represent values of Bayesian posterior probabilities (PP, above) and Maximum Likelihood Bootstraps and SHL tests (BS/SHL, below). Nodal support values ≥ 95% PP and ≥ 70 BS/SHL are illustrated and considered highly supported……………………………………………………………..84

4.2 Locality map showing where tissue samples of Pseustes and Spilotes used in this study were collected throughout Central and South America. Symbols represent estimations of localities examined for molecular sampling. Pseustes poecilonotus from Nicaragua (upper insert). Phylogenetic estimate of relationships within the Pseustes and Spilotes clade, resulting from the Bayesian 50% majority-rule consensus phylogram from figure 1 (lower insert). ** = 100 posterior probability and 95–100 bootstrap/SHL support, respectively…………………………...86

xi CHAPTER 1

SLITHERING INTO NEOTROPICAL EVOLUTION: INTEGRATIVE SYSTEMATICS OF NEW WORLD SNAKES

Introduction to systematics

Taxonomy based on evolutionary relationships plays a critical role in science and society, as properly named units of biological diversity allows for immediate recognition of the general appearance and basic natural history of a particular lineage. Consequently, species as units of biodiversity have been referred to as the “backbone of biology” (Uetz, 2010) and it could be argued that systematics, a field that investigates not only the classification of organisms but also their diversity and origins, is among the most important with relevance to all other disciplines within biology (Daly et al., 2012). Therefore, it is encouraging that some studies have shown the number of taxonomists across animal groups have been steadily increasing over the past century

(Joppa et al., 2011; but see Pearson et al., 2011). This potential increase of taxonomic expertise, along with rapidly advancing evolutionary methodologies (see Losos et al., 2013), is providing a more complete view of Darwin's great tree of life.

New World snakes have been incorporated into our understanding of global biodiversity since the beginning of such endeavors. Prior to Linneaus, many individuals had illustrations made of unique plants and animals as an initial means to quantify biological diversity and New

World snakes were included in these early works (e.g., Albertus Seba). New World snakes, such as the Timber Rattlesnake (Crotalus horridus), were incorporated into the first formal species descriptions by Linneaus (1758). This trend continued in the 18th and 19th centuries through the works of Laurenti (1768), Duméril and Bibron (1835), Schlegel (1837), Günther (1858),

Boulenger (1893), and many others. And long before many other groups of biologists (e.g., entomologists; see Mickel, 1930) were arguing about the utility of type-specimens for species

1 descriptions, this material was already incorporated into common practice for snake species that

Linneaus had described (see Andersson, 1899). Over the years as scientific rigor and knowledge of species diversity has improved, biologists have recognized the need for more advanced and sophisticated methodologies to understand this immense complexity of life. As a result, the demand for taxonomy to become more integrative, less subjective, and abide by strict guidelines has greatly increased since these foundational taxonomic works (Dayrat, 2005). Taxonomic descriptions have therefore become multidisciplinary often utilizing ecology, geography, molecular data, natural history, and phylogenetic theory to understand the natural and evolutionary histories of species beyond gross morphology. As a result, taxonomy based on phylogeny is providing more information about evolutionary relationships in classification. One example of how these relationships are important to incorporate into taxonomy is the medical treatment of venomous snakebites where the evolution of venom is often strongly associated with the diversification of the venomous snakes. Applying appropriate taxonomy to venomous snakes based on their phylogeny may provide better treatment by providing knowledge about relationships through classification. In this chapter, I discuss how I have incorporated modern systematic methodologies to the study of New World snakes throughout the chapters of this dissertation.

Integrative systematics

Since the incorporation of phylogenetic information into taxonomy (e.g. Nelson, 1971), our abilities to effectively delimit species based on evolutionary histories have been greatly enhanced. Although some studies have tried to delimit species using solely molecular and phylogenetic methods (e.g. Leaché & Fujita, 2010), these methods alone violate the strict

2 demands of proper species descriptions according to the rules of the International Code of

Zoological Nomenclature (Bauer et al., 2010) as well as the guidelines for integrative taxonomy

(Dayrat, 2005). It is therefore paramount to integrate both molecular phylogenetic methods, for rigorous statistical support of species delimitation, with more traditional morphological evaluations typical of classical species descriptions. Beyond just naming groups that fit a genealogical concordance, these studies should include examination of museum specimens, including type-series, carefully addressing previous taxonomic work and synonyms, and examination of all species in the group being revised. This inclusion of these procedures not only enhances our abilities to fully understand species diversity and evolutionary history but also refrains us from further complicating taxonomy by making unnecessary errors.

In the second chapter of this dissertation, I adopt these procedures to conduct an integrative systematic revision of Cerrophidion barbouri, a polymorphic species endemic to the highlands of the Sierra Madre del Sur in Guerrero, Mexico. This species is rarely collected and is known from only a couple dozen specimens. As part of a research expedition in 2007, I found several specimens of C. barbouri allowing for the first molecular assessment of this species. I carefully re-examine nearly all known specimens of C. barbouri and the type-series of Agkistrodon browni, a junior synonym, which reveals that both names represent valid species and I therefore resurrected A. browni. I provide phylogenetic analyses using both Bayesian and maximum parsimony criteria of New World pitvipers to investigate the phylogenetic position of A. browni and C. barbouri. These results strongly support a clade consisting of A. browni, C. barbouri, and

Ophryacus melanurus, which has a distant sister relationship to O. undulatus. Based on the deep phylogenetic divergences among these species and distinctive morphology, I erect a new genus,

Mixcoatlus, for A. browni, C. barbouri, and O. melanurus and I revise the genera Cerrophidion

3 and Ophryacus in accordance with this new classification. I believe this type of rigorous methodology to delimit species is paramount to conducting systematic revisions.

Detecting cryptic diversity

The study of cryptic diversity has been increased across the “tree of life” at an exponential rate during the past two decades (Bickford et al., 2007). The process of detecting cryptic diversity has led to significant findings from the distinction between the African Forest and Savanna

Elephants (Roca et al., 2001) and the detection of more than 95% undescribed taxa within 313 provisional species of parasitoid wasps detected in northwestern Costa Rica (Smith et al., 2008).

Part of this detection of cryptic diversity is facilitated through the advancements of molecular phylogenetic studies that investigate lineage diversification across the geographic range of a species. These molecular tools are often the strongest evidence for detecting species diversity within taxa that may be morphologically or ecologically conserved, particularly within species with large distributions. These data are critical for conservation efforts that attempt to manage or protect viable populations of distinct lineages of biodiversity.

In the third chapter of this dissertation, I detect cryptic diversity within the species

Cerrophidion godmani, one of the few pitviper species incorporated into phylogeographic studies (i.e., Castoe et al., 2005, 2009; Daza et al., 2010). Historically, C. godmani has been considered a wide-ranging pitviper distributed throughout the highlands of Central America, including each country except Belize and Nicaragua. My work expanded on the the phylogeography of C. godmani by adding novel samples from isolated montane regions of

Honduras and combining these data with sampled individuals across its range. The molecular phylogenetic evidence suggested that C. godmani represents three species-level lineages that do

4 not form a clade. These three lineages are relatively conserved in their morphology but, through the examination of hundreds of C. godmani sensu lato specimens, I subsequently identified sufficient morphological characters to diagnose them as distinct. Based on the molecular and morphological evidence, I formally name and describe two of these lineages as new species and together with known geographic distributions of these populations; I inferred the geographic boundaries of these species.

Placing taxa in the tree of life

Since Darwin (1859) systematists have attempted to place all extant and extinct taxa onto a single dated phylogenetic tree of life (Benton & Ayala, 2003) and more recently to make this tree readily accessible to all on the web (e.g., Maddison et al., 2007; Rosindell & Harmon, 2012).

There are currently 1,438 species of snakes described from the New World (Uetz, 2013) and a large number of New World snake species are becoming incorporated into an evolutionary framework using multi-locus datasets (e.g., Pyron et al., 2011). These studies are increasing our understanding of the phylogenetic position of all taxa and to date only a few genera have not been incorporated into a phylogenetic framework. Those genera that have not been placed into a phylogenetic framework are often enigmatic with little predictive power to assess which higher taxonomic group (e.g., subfamily) they may correspond to, let alone where it is positioned within that clade.

In the third chapter of this dissertation, I conduct an initial systematic revision of the

Neotropical Birdsnakes, genus Pseustes. This genus is composed of two to five recognized species, depending on the authority, that occur throughout tropical Central and South America.

The genus Pseustes is one of the few genera that has not been included in a molecular phylogeny

5 to assess its phylogenetic position and untangle its complex evolutionary history using modern systematic approaches. One of these species, P. sulphureus, is among the largest snakes in the

New World, approaching three meters in length. In this chapter, I incorporate both nuclear and mitochondrial markers to conduct the first molecular-based phylogeny of Pseustes in order to assess its phylogenetic position among New World snakes and understand historical diversification of lineages within this genus. I found strong support for the paraphyly of the genus Pseustes with respect to the monotypic genus Spilotes, both of which are nested within a clade of at least 23 other New World Colubrinae genera. Based on my results, I formally stabilize the generic taxonomy of Pseustes by allocating P. sulphureus to the genus Spilotes and thereby keeping both genera monophyletic. Finally, I identify two lineages within P. poecilonotus that are currently unrecognized species.

6 CHAPTER 2

UNRAVELLING A TANGLE OF MEXICAN SERPENTS: A SYSTEMATIC REVISION OF 1 HIGHLAND PITVIPERS

Abstract

As most recently recognized, the name Cerrophidion barbouri Dunn, 1919 refers to a highland species of pitviper endemic to Guerrero, Mexico, of which Agkistrodon browni Shreve, 1938 is considered a junior synonym. This species is rarely collected and prior to recent decades it was known from only a few specimens. A careful re-examination of nearly all known specimens of

C. barbouri and the type-series of A. browni reveals that both names represent valid species and we therefore resurrect A. browni. Both species are extremely variable with respect to cephalic scalation and color pattern, which has previously confounded efforts to identify them. We provide phylogenetic analyses using both Bayesian and maximum parsimony criteria of New

World pitvipers to investigate the phylogenetic position of A. browni and C. barbouri. Our phylogenetic tree, based on 2235 bp of mitochondrial data (12S, 16S, cyt b, ND4), strongly supports a clade consisting of A. browni, C. barbouri, and Ophryacus melanurus, which has a distant sister relationship to O. undulatus. Based on the deep phylogenetic divergences among these species and distinctive morphology we recommend a new genus be recognized for A. browni, C. barbouri, and O. melanurus. Finally, we revise the genera Cerrophidion and

Ophryacus in accordance with our new classification.

1 Adapted from: R.C. Jadin, E.N. Smith, and J.A. Campbell. 2011. Unravelling a tangle of Mexican serpents: a systematic revision of highland pitvipers. Zoological Journal of the Linnean Society, 163, 943–958.

7 Introduction

Mexico is well known for its herpetofaunal diversity of approximately 1200 known amphibian and reptile species (Flores-Villela & Canseco-Márquez, 2004), including many endemic genera

(e.g., Barisia, Charadrahyla, Chiropterotriton, Rhadinophanes). In particular, pitvipers are extremely diverse in Mexico, with at least 56 species and 9 genera of the total 115 species and 14 genera recognized in the New World (Campbell & Lamar, 2004; Campbell & Flores-Villela,

2008; Fenwick et al., 2009). These numbers of New World pitvipers have increased greatly from the 90 species in 9 genera recognized only two decades ago (Campbell & Lamar, 1989). Our knowledge of pitviper diversity and relationships is constantly being refined as independent geographic lineages are distinguished and new species are discovered. With rapidly advancing phylogenetic methodologies, we are proceeding toward a more thorough understanding of the evolutionary histories of this remarkable group (see Gutberlet & Harvey, 2004; Castoe &

Parkinson, 2006; Fenwick et al., 2009).

Cerrophidion barbouri is a pitviper restricted to the Sierra Madre del Sur in southern Mexico at elevations above 2000 meters. Prior to some recent survey work, this rare species was known from only a few individuals and its evolutionary and natural history remains poorly known.

Dunn (1919) described Lachesis barbouri from Omilteme, Mexico, and distinguished it from other vipers by its undivided subcaudals, 17 dorsal scale rows, and enlarged frontal plate.

Specimens were obtained in the state of Guerrero for North American collections by Mr. Wilmot

W. Brown during the early decades of the 1900s (Campbell & Flores-Villela, 2008), and Shreve

(1938) named Agkistrodon browni from Omilteme on the basis of two specimens collected by

Brown. Shreve (1938) diagnosed the species only from Old World members of the genus

Agkistrodon (sensu lato), with which he assumed it was allied, presumably based on its large

8 cephalic plates. After observing variation in head scales and number of ventrals in Cerrophidion godmani, but lacking examination of C. barbouri specimens, Smith (1941) concluded that similar trends present in A. browni could be attributed to sexual dimorphism, stating “there is no reasonable doubt that browni and barbouri are synonymous.” Agkistrodon browni and C. barbouri possess external features, such as numbers of ventral and subcaudal scales, which overlap considerably. The most comprehensive works on C. barbouri, with A. browni considered as a synonym, have been those of Campbell (1977, 1988) and Campbell & Lamar (1989, 2004), who noted considerable variation among the morphology of C. barbouri but did not suggest the possibility of multiple species.

In this study we examine type-material and all but three known specimens of A. browni and

C. barbouri. We demonstrate the distinctiveness of A. browni from C. barbouri and provide detailed descriptions for both species. Additionally, we conduct a phylogenetic analysis using four mitochondrial gene fragments under Bayesian and parsimony criteria in order to assess the phylogenetic position of both A. browni and C. barbouri among the New World pitvipers. Our findings render the endemic Mexican pitviper genus Ophryacus paraphyletic and we therefore propose a new genus and systematic revisions of Cerrophidion and Ophryacus. Finally, we review the published natural history information attributed to C. barbouri (and by implication A. browni) and segregate this information for the two species.

Materials and Methods

External morphological data

Based on the type-series and original descriptions of Agkistrodon browni (Shreve, 1938) and

Cerrophidion barbouri (Dunn, 1919), we separated the 27 C. barbouri (sensu lato) specimens

9 into two morphotypes, ‘Cerrophidion barbouri’ and ‘Agkistrodon browni’. External morphology of 14 C. barbouri and 13 A. browni was examined (Appendix 2.1; Table 2.1). Institutional abbreviations of specimens follow Leviton et al. (1985). Descriptions and nomenclature for characters are mostly from Klauber (1972), Campbell (1977, 1988), and Campbell & Lamar

(1989, 2004) but particular methods for counting these characters have been described in the following morphological phylogenetic studies: Werman (1992), Wüster et al. (1996), Gutberlet

(1998), Gutberlet & Harvey (2002), Jadin (2010), and Jadin et al. (2010). Scale count abbreviations follow citations: scales contacting third supralabial (Jadin et al. no. 14, modified

Wüster et al. no. 28; C3SL), counted as scales directly contacting third supralabial from rostral; scales contacting supraoculars (Wüster et al. no. 27; CSupOc); gulars (Gutberlet & Harvey no. 8;

GLR); infralabials (IL); interoculabials (Gutberlet & Harvey no. 1; IOL); interrictals (Gutberlet

& Harvey no. 7; IR); intersupraoculars (Werman no. 25; ISO); dorsal scale rows at midbody

(Gutberlet & Harvey no. 10; NMSR); subcaudals (Gutberlet & Harvey no. 62; NSC); ventral scales (Gutberlet & Harvey no. 9; NVEN); prefoveals (Gutberlet & Harvey no. 2, Werman no.

37, in part; PF); postoculars (Jadin no. 19, Jadin et al. no. 16; PO); subfoveals (Gutberlet &

Harvey no. 16; SF); supralabials (Werman no. 26; SL); suboculars (Gutberlet & Harvey no. 3;

SO).

Additional characters were examined in the type-specimens of A. browni and C. barbouri

(Table 2.2) to firmly establish species allocation. Meristic characters and their abbreviations are as follows: canthals (Werman no. 32; CAN); dentary teeth (Gutberlet & Harvey no. 30; DNT); dorsal scale arrangement (DSA), number of dorsal scale rows one head length behind the head, at midbody, and one head length anterior to the vent; intercanthals (Jadin no. 21, Jadin et al. no.

18; IC); internasals (Jadin no. 20, Jadin et al. no. 15; IN); palatine teeth (Gutberlet & Harvey no.

10 28; PAL); prefrontal scales (PFR); pterygoid teeth (Gutberlet & Harvey no. 29, Werman no. 51, in part; PTY); preventral scales (PV), as defined in Dowling (1951); scales forward of frontal scale (SFF), number of dorsal head scales between frontal and rostral scales (i.e., prefrontals, canthals, and internasals); snout shape (SS), curvature of snout defined as being either pointed or round. The following mensural characteristics were obtained from the type-series using a digital caliper or dissecting microscope with an optical micrometer, and were taken to the nearest 0.1 mm. Descriptions and abbreviations follow Grismer et al. (2006) and Vogel et al. (2004): distance between nostrils (DBN); distance from lower eye margin to bottom edge of the fourth supralabial, directly below (modified from Grismer et al. (2006); DEL); distance from anterior margin of eye to the posterior margin of the nostril (DETN); distance from the anterior edge of the eye to the posterior edge of the pit cavity (DETP); distance from the anterior edge of the eye to the rostral scale (DER); second supralabial height (H2SL); third supralabial height (H3SL); horizontal eye diameter (HED); head length (HL); head width (HW); second supralabial length

(L2SL); third supralabial length (L3SL); length of frontal scale (LFS); loreal scale height (LH); loreal scale length (LL); length of supraocular scale (LSupOc); parietal scale length (PL); parietal scale width (PW); vertical eye diameter (VED); width of frontal scale (WFS); width of supraocular scale (WSupOc). The snout-to-vent length (SVL), tail length (TaL), and total body length (TL) were taken using a meter stick to the nearest millimeter.

Hemipenial preparations

We dissected and examined the left hemipenes from specimens deposited at MZFC and UTA

(MZFC 2881 & UTA R-4450). We removed by dissection at the base. We fully everted hemipenes by filling them with warm water using a blunt-tipped syringe needle. We removed

11 water and then injected hot liquid petroleum jelly with blue wax-dye until maximum expansion was achieved. Finally, we tied the organs at the base and stored them in 70% ethanol. This procedure is modified from that of Myers & Cadle (2003), Zaher & Prudente (2003), and Smith

& Ferrari-Castro (2008). Hemipenial terminology follows Dowling & Savage (1960), Keogh

(1999), and Savage (2002).

Molecular data

Genomic DNA from muscle tissue or ventral scale clips from three A. browni and one C. barbouri was isolated using a Qiagen DNeasy extraction kit and protocol. Four mitochondrial gene fragments—16S rRNA, NADH dehydrogenase subunit 4 (ND4), 12S rRNA, and cytochrome b (cyt b)—were independently PCR amplified as described in (Knight & Mindell,

1993; Arévalo et al., 1994; Parkinson et al., 1997, 2002) using Promega GoTaq® Green master mix, the primer pairs: 16SF + 16SR, ND4 + LEU, L1091 + 12E, and Gludg + AtrCB3, and annealing temperatures 45°C, 48°C, 50°C, and 48°C, respectively. Either AMPure magnetic beads (Agencourt®, Bioscience, Beverly, Massachusetts, USA) or ExoSap It (USB Corporation,

Cleveland, Ohio, USA) were used to clean amplified fragments. Post PCR cleanup sequencing protocols were performed by SeqWright Inc. (Houston, Texas, USA; http://www.seqwright.com) or the University of Texas at Arlington genomics core facility (Arlington, Texas, USA; http://gcf.uta.edu). Sequencing was performed in both forward and reverse directions and sequence chromatographs were edited together using Sequencher 4.2. Novel sequences from this study were deposited in GenBank (HM363639–HM363653). Previously published sequences of ingroup—50 additional New World pitviper taxa—and outgroup taxa—Deinagkistrodon acutus,

Gloydius halys, and Protobothrops jerdonii—were downloaded from GenBank (Table 2.1).

12 Table 2.1. Taxa, vouchers, locality data, and GenBank accession numbers for sequences used in this study. Sequences newly added specifically for this study are in bold. GenBank accession numbers per gene fragment Taxon Voucher Locality 12S 16S Cyt-b ND4 Agkistrodon bilineatus WWL Costa Rica: Guanacaste AF156593 AF156572 AY223613 AF156585 Agkistrodon browni MZFC 21429 Mexico: Guerrero: El Balcón: El HM363650 HM363651 HM363652 HM363653 Filo MZFC 21431 Mexico: Guerrero: Road from El HM363643 HM363644 HM363645 HM363646 Jilguero to Puerto del Gallo, 3192 m UTA R-56265 Mexico: Guerrero: El Balcón: La HM363647 HM363648 HM363649 ― ― ― Llave Agkistrodon contortrix Moody 338 USA: Ohio: Athens Co. AF057229 AF057276 AY223612 AF156576 Agkistrodon piscivorus CLP 30 USA: South Carolina AF057231 AF057278 AY223615 AF156578 Agkistrodon taylori CLP 140 Mexico: Tamaulipas AF057230 AF057277 AY223614 AF156580 Atropoides mexicanus CLP 168 Costa Rica: San José AF057207 AF057254 AY223584 U41871 Atropoides nummifer ENS 10515 Mexico: Puebla: San Andres DQ305422 DQ305445 DQ061195 DQ061220 Tziaulan Atropoides occiduus UTA R-29680 Guatemala: Escuintla: S. slope DQ305423 DQ305446 AY220315 AY220338 Volcán de Agua Atropoides olmec UTA R-25113 Mexico: Veracruz: Sierra de los AY223656 AY223669 AY220321 AY220344 Tuxtlas Atropoides picadoi CLP 45 Costa Rica: Alajuela: Varablanca AF057208 AF057255 AY223583 U41872 Bothriechis aurifer UTA R-35031 Guatemala: Baja Verapaz: DQ305425 DQ305448 DQ305466 DQ305483 Vicinity of La Unión Barrios Bothriechis bicolor UTA R-34156 Unknown DQ305426 DQ305449 DQ305467 DQ305484 Bothriechis lateralis MZUC R-11155 Costa Rica: Acosta AF057211 AF057258 AY223588 U41873 Bothriechis marchi UTA R-52959 Guatemala: Zacapa: Cerro del DQ305428 DQ305451 DQ305469 DQ305486 Mono Bothriechis MZUC R-11151 Costa Rica: San Gerondo de AF057212 AF057259 AY223589 AY223635 nigroviridis Dota Bothriechis rowleyi JAC 13295 Mexico: Oaxaca: Cerro Baúl DQ305427 DQ305450 DQ305468 DQ305485 Bothriechis schlegelii MZUC R-11149 Costa Rica: Cariblanco de AF057213 AF057260 AY223590 AY223636 Sarapiquí Bothriechis Costa Rica: San Vito DQ305429 DQ305452 DQ305470 DQ305487 supraciliaris Bothriechis thalassinus UTA R-52958 Guatemala: Zacapa DQ305424 DQ305447 DQ305465 DQ305482 Bothriopsis bilineata Colombia: Letícia AF057214 AF057261 AY223591 U41875 Bothriopsis LSUMZ-41037 Peru: Pasco Dept. DQ305430 DQ305453 DQ305471 DQ305488 chloromelas Bothriopsis taeniata Suriname AF057215 AF057262 AY223592 AY223637 Bothrocophias Colombia: Letícia AF057206 AF057253 AY223593 U41886 hyoprora Bothropoides diporus PT 3404 Argentina: La Rioja: Castro DQ305431 DQ305454 DQ305472 DQ305489 Barros Bothropoides RG 829 Brazil: Alagóas: Piranhas AF057219 AF057266 AY223600 U41877 erythromelas Bothropoides insularis WWW Brazil: São Paulo: Ilha Queimada AF057216 AF057263 AY223596 AF188705 Grande Bothrops asper MZUC R-11152 Costa Rica AF057218 AF057265 AY223599 U41876 Bothrops atrox WWW 743 AY223659 AY223672 AY223598 AY223641 Bothrops jararacussu DPL 104 Brazil AY223661 AY223674 AY223602 AY223643 Cerrophidion barbouri MZFC 21432 Mexico: Guerrero: El Balcón: El HM363639 HM363640 HM363641 HM363642 Moreno Cerrophidion godmani MZUC R-11153 Costa Rica: San José AF057203 AF057250 AY223578 U41879 UTA R-40008 Guatemala: Baja Verapaz: La DQ305419 DQ305442 AY220325 AY220348 Unión Barrios Cerrophidion ENS-10528 Mexico: Veracruz: Orizaba DQ305420 DQ305443 DQ061202 DQ061227 petlalcalensis Crotalus adamanteus CLP 4 USA: Florida: St. Johns Co. AF057222 AF057269 AY223605 U41880 Crotalus atrox CLP 64 USA: Texas: Jeff Davis Co. AF057225 AF057272 AY223608 AY223646 Crotalus molossus CLP 66 USA: Texas: El Paso Co. AF057224 AF057271 AY223607 AY223645 Crotalus ravus UTA live Mexico: Puebla: Zapotitlán AF057226 AF057273 AY223609 AY223647 Crotalus ruber ROM 18197–98, USA: California: Riverside Co. AF259261 AF259153 AF259191 ― ― ― ROM 18207 Crotalus tigris CLP169 USA: Arizona: Pima Co. AF057223 AF057270 AY223606 AF156574

13 Deinagkistrodon CLP 28 China AF057188 AF057235 AY223560 U41883 acutus Gloydius halys Kazakhstan AF057191 AF057238 AY223564 AY223621 Lachesis muta Cadle 135 Peru AF057221 AF057268 AY223604 AY223644 Lachesis stenophrys Costa Rica: Limón AF057220 AF057267 AY223603 U41885 Ophryacus melanurus UTA R-34605 Mexico AF057210 AF057257 AY223587 AY223634 Ophryacus undulatus CLP 73 Mexico AF057209 AF057256 AY223586 AY223633 Porthidium arcosae WWW 750 Ecuador: Manabí: Salango AY223655 AY223668 AY223582 AY223631 Porthidium dunni ENS-9705 Mexico: Oaxaca: near San Pedro AY223654 AY223667 AY223581 AY223630 Pochutla Porthidium nasutum MZUC R-11150 Costa Rica AF057204 AF057251 AY223579 U41887 Porthidium UMMZ 210276 Costa Rica: Guanacaste AF057205 AF057252 AY223580 U41888 ophryomegas Porthidium porrasi MSM Costa Rica: Puntarenas DQ305421 DQ305444 DQ061214 DQ061239 Protobothrops jerdonii CAS 215051 China: Yunnan: Nu Jiang AY294278 AY294269 AY294274 AY294264 Rhinocerophis DPL 2879 Brazil: Rio Grande do Sul: São AY223660 AY223673 AY223601 AY223642 alternatus Sebastiao do Tai Rhinocerophis MVZ 223514 Argentina: Neuguén AY223658 AY223671 AY223595 AY223639 ammodytoides Rhinocerophis cotiara WWW Brazil AF057217 AF057264 AY223597 AY223640 Sistrurus catenatus Moody 502 USA: Texas: Haskel Co. AF057227 AF057274 AY223610 AY223648 Sistrurus miliarius UTA-live USA: Florida: Lee Co. AF057228 AF057275 AY223611 U41889

Deinagkistrodon acutus was used to root every analysis. This taxonomic sampling includes all

14 previously recognized New World genera. Sequences for each gene were aligned separately, first automatically using the program MUSCLE (Edgar, 2004) and, then manually using Se-Al v2.0a11. This dataset was further edited manually or transformed using GeneDoc (Nicholas and

Nicholas, 1997). The entire 16S fragment was trimmed to 100% representation; the ND4 fragment contained 9 taxa with 15 bp missing at one end and had one complete missing operational taxonomic unit (OTU) of A. browni, though this same taxon was represented by two other sequences; 12S was allowed to have only one 25 bp stretch of sequence missing from one taxon; and cyt-b was allowed one sequence with missing data, less than 100 bp. In general there was never more than 3% of the total data per gene missing. More than 80% of the OTUs were always represented by the data. Gaps in alignments were treated as missing data and internal stop codons were not found in the two protein-coding gene fragments.

14 Phylogenetic analyses

Bayesian inference and maximum parsimony (MP) were implemented to reconstruct phylogenies. Model likelihoods for each gene fragment were independently calculated and models were chosen using the Akaike information criterion (AIC) in MrModelTestv2.2

(Nylander, 2004) and PAUP* v4.0b10 (Swofford, 2002). AIC scores for each gene fragment were found to best fit the general time reversal + invariant sites + gamma distribution rate variation model of evolution. The four gene fragments were concatenated into one NEXUS file

(2235 total bp) and protein coding genes cyt b and ND4 were partitioned into three codon positions and ribosomal RNA loci 12S and 16S were partitioned into stems and loops, resulting in a total of ten partitions (model 10x in Castoe & Parkinson 2006), and implemented for phylogenetic analyses.

Bayesian Markov chain Monte Carlo (MCMC) phylogenetic analyses were conducted using

MrBayes v3.0b4 (Ronquist & Huelsenbeck, 2003). Two simultaneous runs of four MCMC analyses, consisting of one cold and three incrementally heated chains, were initiated with random trees for a total of 5.0 × 106 generations (sampling every 100 generations). The first 1.5

× 106 generations from each run were discarded as burn-in. We used Tracer v1.5 (Rambaut &

Drummond, 2009) to detect stationarity in the Markov chain within the burn-in period.

Parsimony-based analysis of molecular data was conducted using PAUP* v4.0b10

(Swofford, 2002) under a heuristic search criterion using tree bisection-reconnection branch swapping and 10 random addition sequence replicates with all characters weighted equally. A weighted parsimony (WP) analysis was conducted utilizing a tri-level weighting scheme

(Benabib et al., 1997; Flores-Villela et al., 2000) with gaps coded as a fifth base or 21st amino acid. Tri-level weighting incorporates three different levels of information on the structure and

15 inferred function of nucleotide substitutions. Under this WP scheme, transitions have a weight of

1, transversions are weighted 2, and any nucleotide substitution that is inferred to cause an amino-acid substitution is weighted +1 more. For our parsimony analysis, the extremes of the gene fragments were further trimmed to exclude additional missing data that could potentially affect analysis when using gaps as an extra character state. This dataset of 2235 bp—16S (486 bp), ND4 (684 bp), 12S (411 bp), and cyt b (654 bp)—was coded for transition/transversion analysis (doubling the number of original characters), and the protein coding genes were transformed to amino acid (aa) sequence, ND4—227 aa, and cyt b—217 aa in length. The coded sequence comprised a total of 4914 characters, 1272 parsimony informative. All raw DNA characters were independent but with different weights according to their biochemical properties.

Weighted parsimony employed accelerated transformation optimization of character state changes. Bootstrap analysis (Felsenstein, 1985) involved 1 × 104 pseudoreplicates obtained via random addition sequence. The parsimony tri-level weighting approach of the combined gene datasets is justified by the comparative study of Kjer et al. (2007), who showed that this method outperforms all other methods, including MP, unpartitioned maximum likelihood, and Bayesian likelihood analyses.

Finally, to obtain an estimate of genetic distances we computed pairwise comparisons of the cyt b gene fragment between and within the various genera according to our classification. We calculated these distances with MEGA v4.1 (Tamura et al., 2007) and in accord with previous studies (e.g., Fenwick et al., 2009) incorporated the Kimura two-parameter model with Γ– distributed rate variation.

16 Natural history

Description of the habitat and natural history of Cerrophidion barbouri and Agkistrodon browni is from published accounts (i.e., Davis & Dixon, 1959; Campbell, 1977, 1988; Campbell &

Lamar, 1989, 2004) and personal observations.

Results

Morphological analyses

While examining 27 specimens of Cerrophidion barbouri (sensu lato), including type-material, we found distinct morphological differences that may be attributable to two species. Agkistrodon browni is readily distinguished from C. barbouri in having a greater numbers of middorsal scale rows, and fewer interoculabials, intersupraoculars, interrictals, prefoveals, postoculars, scales contacting the third supralabial and supraocular, and subfoveal rows (Figs 2.1, 2.2, Table 2.2), and a prehensile tail (Fig. 2.3). Additionally, the type-series of these two species differs greatly in these and additional characters providing further distinctions (see Table 2.3).

17

Table 2.2. Morphological comparisons between Cerrophidion barbouri and Agkistrodon browni. C. barbouri A. browni n = 14 n = 13 C3SL 5.15 4 (5 & 6) (4) CSubOc 10.25 8 (9–11) (7–9) GLR 3.44 3.85 (2–5) (3–5) IL 9.36 9.15 (8–10) (9 & 10) IOL 1 0 (1) (0) IR 24.53 20.09 (22–26) (19–22) ISO 4.36 1 (3–5) (1) NMSR 17.29 19 (17 & 19) (19) NSC 30.14 30.31 (27–32) (27–35) NVEN 140.14 138.91 (130–148) (134–145) PF 2.78 1.1 (1–6) (0–2) PO 3.3 1.92 (2–4) (1 & 2) SF 1 0 (1) (0) SL 8.61 8.04 (8–10) (7–10) SO 3.63 2.58 (3–5) (1–3) See Materials and Methods for character abbreviations. Counts of bilateral characters were taken from each side and averaged. Means are reported with ranges in parentheses.

18 Table 2.3. Measurements and counts of the type series of Cerrophidion barbouri and Agkistrodon browni. C. barbouri A. browni A. browni USMN R-46347 MCZ R-42678 MCZ R-42679 female male Holotype female Holotype Paratype CAN 2/2 2/2 2/2 CSupOc 10/10 7/7 7/7 DBN 4.65 5.1 4.9 DEL 2.7/2.7 2.8/2.7 2.7/2.8 DETN 3.6/3.7 4.6/4.5 3.8/3.7 DETP 1.2/1.0 1.1/1.1 0.8/0.8 DER 5.0/5.2 6.3/6.2 5.0/5.0 DNT a 10 10 12 DSA 19-17-15 19-19-15 17-19-15 GLR 2/2 4/4 5/4 H2SL 1.2/1.1 1.5/1.7 1.4/1.4 H3SL 2.4/2.4 2.1/1.9 2.0/2.1 HED 2.9/2.9 3.1/3.2 2.8/2.8 HL 19.45 24.8 20.77 HW 13.11 18.0 14.95 IC 4 2 2 IL 9/9 9/9 9/9 IN 4 6 4 IR 24 19 19 ISO 3 1 1 L2SL 1.3/1.7 1.1/1.2 1.3/1.4 L3SL 2.0/2.3 2.4/2.4 2.0/1.8 LFS 2.6 4.6 4.5 LH 1.3/1.3 1.5/1.5 1.4/1.5 LL 1.8/1.9 2.3/2.3 2.0/2.0 LSupOc 4.6/4.7 6.0/6.0 5.1/5.1 PAL a 3 3 3 PF 3/5 0/0 0/0 PFR ? 2 2 PL ?/? 4.0/5.5 3.4/4.3 PO 3/3 2/2 2/2 PTY* 12 11 10 PV 4 3 2 PW ?/? 3.5/3.4 2.9/3.0 SC 31 31 27 SFF 16 9 8? SL 8/9 8/8 8/8 SO 4/4 2/1 2/2 SS Pointed Round Round SVL 360 425 355 TL 404 480 391 TaL 44 55 36 VED 1.8/1.7 2.0/2.0 1.8/1.7 VEN 148 134? 141 WFS 2.9 3.8 3.0 WSupOc 2.4/2.3 3.7/3.5 2.9/2.9 *Right side only See Materials and Methods for character abbreviations. Counts and measurements are written as right/left side, measurements taken in mm.

19

Figure 2.1. Agkistrodon browni (A, B; UTA R-56265) and Cerrophidion barbouri (C, D; MZFC 21432) in life, showing differences in head scalation and colour pattern.

Figure 2.2. Dorsal view (A, B) and left side view (C, D) of Agkistrodon browni (holotype, MCZ R-42678; left) and Cerrophidion barbouri (holotype, USMN R-46347; right).

20

Figure 2.3. Photo in life of Agkistrodon browni (UTA R-56264) showing its prehensile tail.

Furthermore, hemipenial features of A. browni and C. barbouri differ greatly (Fig. 2.4). A thorough description of the hemipenis of A. browni and C. barbouri is lacking and therefore included here. The everted left hemipenes of A. browni (UTA R-4450; SVL 397 mm, TaL 53 mm, subcaudals 31 Fig. 2.4A) and C. barbouri (MZFC 2881; SVL 445 mm, TaL 60 mm, subcaudals 32, Fig. 2.4B) are, respectively, 14 and 16 mm in length and 8 and 9.5 mm in maximum width at point of bilobation; on sulcate side base with several rows of small spines (<

0.6 mm) for 2 and 1.5 mm, then rows of larger spines and hooks extending for 3 and 4.5 mm, largest protruding 3.5 and 2.5 mm; asulcate side with naked base up to 3 and 2 mm before level of bilobation and then with 3.5 and 2 mm section of small spines (< 0.5 mm) arranged in rows followed by 2.5 and 3 mm section of larger spines; each lobe with 60 and >35 spines and 20 and

~8 hooks; 15 and 12 spines and hooks around each lobe at the lower rim of calyces; calyces follow spines and hooks distally; calyces scalloped and spinous or slightly scalloped, 14 rows extending > 4 and 6 mm to apex of the hemipenis on the asulcate side; sulcus spermaticus bifurcates ca. 2 mm before site of bilobation and extending upwards through spines and calyces

21 to tip of each lobe; border of sulcus spermaticus naked to point of bilobation where small spines occur for 2 and 4 mm to level of calyces and forming the border to the apex of the lobe.

The most prominent feature of the hemipenes in A. browni are dramatically enlarged hooks, the largest being more than one-fourth the length of the entire organ. The hooks of A. browni are not located at the base of the hemipenis, the condition characterizing most viperids, but rather are on the proximal portion of the hemipenial lobes, with smaller spines at the base of the organ. We are not aware of another pitviper species in which the spines are so disproportionally large or that have such large spines on the lobes. The hemipenes of A. browni further differ from those of C. barbouri by having much larger and nearly twice the number of spines and hooks. Agkistrodon browni has a smaller relative area of calyces covering the lobes, and the calyces are more scalloped and spinous than in C. barbouri.

Figure 2.4. Sulcate (left) and asulcate (right) views of the left hemipenis of: A, Agkistrodon browni (UTA R-4450) and B, Cerrophidion barbouri (MZFC 2881). Scale bars = 3 mm.

22 Phylogenetic analyses

Our Bayesian and parsimony phylogenetic hypotheses are congruent with each other and no strongly supported conflicts exist. The weighted parsimony analysis recovered three optimal trees of 8165 steps each (Fig. 2.5). We chose the first of these trees presented by PAUP* with no apriori preference since the three optimal trees differed only in the arrangement of the closely related samples of ‘Agkistrodon browni’, one of them lacking ND4. These hypotheses are mostly congruent with that preferred by Castoe & Parkinson (2006), showing strong nodal support for the monophyly of rattlesnakes (i.e., Crotalus, Sistrurus), the Porthidium group (i.e., Atropoides,

Cerrophidion, Porthidium), the South American group (i.e., Bothriopsis, Bothrocophias,

Bothropoides, Bothrops, and Rhinocerophis), and the genera Agkistrodon (excluding A. browni),

Bothriechis, and Lachesis; although obtaining little or no support for both the backbone of the phylogeny and the monophyly of Atropoides. Agkistrodon browni, C. barbouri, and O. melanurus form a very strongly supported clade sister to O. undulatus. This phylogeny renders the endemic Mexican pitviper genus Ophryacus paraphyletic.

Genetic distances within New World pitviper genera range from 7.2 to 17.1% while distances between genera range from 11.6 to 25.5% (Table 2.4).

23

Deinagkistrodon acutus, CLP 28, China Gloydius halys, Kazakhstan Protobothrops jerdonii, CAS 215051, China, Yunnan, Nu Jiang Ophryacus undulatus, CLP 73 Mexico 98 Ophryacus melanurus, UTA R-34605, Mexico 98 Cerrophidion barbouri, MZFC 21432, Mexico, Guerrero 100 “Agkistrodon browni”, MZFC 21429, Mexico, Guerrero “Agkistrodon browni”, MZFC 21431, Mexico, Guerrero “Agkistrodon browni”, UTA R-56265, Mexico, Guerrero Lachesis muta, Cadle 135, Peru 100 Lachesis stenophrys, Costa Rica, Limon 100 Agkistrodon contortrix, Moody 338 USA, Ohio 99 Agkistrodon piscivorus, CLP 30, USA, South Carolina Agkistrodon bilineatus, WWL, Costa Rica, Guanacaste 100 Agkistrodon taylori, CLP 140, Mexico, Tamaulipas Sistrurus catenatus, Moody 502, USA, Texas 94 99 Sistrurus miliarius, UTA-live, USA, Florida Crotalus ravus, UTA-live, Mexico, Puebla Crotalus molossus, CLP 66, USA, Texas 79 Crotalus adamanteus, CLP 4, USA, Florida 100 70 Crotalus tigris, CLP169, USA, Arizona Crotalus atrox, CLP 64, USA, Texas 94 Crotalus ruber, Unknown Bothriechis schlegelii, MZUC R-11149, Costa Rica, Cariblanco de Sarapiqui 100 Bothriechis supraciliaris, Costa Rica, San Vito 77 Bothriechis lateralis, MZUC R-11155, Costa Rica, Acosta 100 Bothriechis nigroviridis, MZUC R-11151, Costa Rica, San Gerondo de Dota Bothriechis bicolor, UTA R-34156, Unknown 100 Bothriechis aurifer, UTA R-35031, Guatemala Baja, Verapaz 80 Bothriechis rowleyi, JAC 13295, Mexico, Oaxaca Bothriechis marchi, UTA R-52959, Guatemala, Zacapa 99 Bothriechis thalassinus, UTA R-52958, Guatemala, Zacapa Bothrocophias hyoprora, Colombia, Leticia Rhinocerophis ammodytoides, MVZ 223514, Argentina, Neuguen 100 Rhinocerophis alternatus, DPL 2879, Brazil, Rio Grande do Sul 71 Rhinocerophis cotiara, WWW, Brazil 79 97 Bothropoides insularis, WWW, Brazil, Sao Palo Bothropoides diporus, PT 3404, Argentina, La Rioja 100 Bothropoides erythromelas, RG 829, Brazil, Alagoas 78 100 Bothriopsis bilineata, Colombia, Leticia Bothriopsis chloromelas, LSUMZ 41037, Peru, Pasco 97 Bothriopsis taeniata, Suriname 99 Bothrops jararacussu, DPL 104, Brazil Bothrops asper, MZUC R-11152, Costa Rica 100 Bothrops atrox, WWW 743, Unknown Atropoides occiduus, UTA R-29680, Guatemala Escuintla Bayesian support 100 Atropoides olmec, UTA R-25113, Mexico, Veracruz 1.00 Atropoides mexicanus, CLP 168, Costa Rica, San Jose 99 85 0.97–0.99 87 Atropoides nummifer, ENS 10515, Mexico, Puebla Atropoides picadoi, CLP 45, Costa Rica, Alajuela Parsimony bootstrap support 96 Cerrophidion petlalcalensis, ENS 10528, Mexico, Veracruz Cerrophidion godmani, MZUC R-11153, Costa Rica, San Jose above 70% on branches Cerrophidion godmani, UTAR-40008, Guatemala, Baja Verapaz 91 Porthidium ophryomegas, UMMZ 210276, Costa Rica, Guanacaste New taxa analyzed 100 Porthidium dunni, ENS 9705, Mexico, Oaxaca Porthidium nasutum, MZUC R-11150, Costa Rica 100 Porthidium porrasi, MSM, Costa Rica, Puntarenas 92 50 changes Porthidium arcosae, WWW 750, Ecuador, Manabi

Figure 2.5. One of three equally parsimonious trees (8165 steps) recovered from heuristic maximum parsimony analysis of 2235 bp of four mitochondrial gene fragments (12S, 16S, cytochrome b, and NADH dehydrogenase subunit 4). Nodal support of posterior probability distributions from a separate Bayesian Markov chain Monte Carlo analysis. Owing to on-going taxonomic change and to help comparing phylogenies the unique ID and locality data for each operational taxonomic unit are provided next to the species names.

24

25 Discussion

Taxonomic status of Agkistrodon browni

Because of the distinctive morphological features (see results; Tables 2.2, 2.3) and genetic differentiation (Fig. 2.5), we conclude that Agkistrodon browni is a distinct species separate from

Cerrophidion barbouri. The large, flat head plates in A. browni readily distinguish it from the more moderately sized and usually keeled scales in C. barbouri (Figs 2.1, 2.2). This feature alone serves to distinguish A. browni from all other New World pitvipers except Agkistrodon,

Sistrurus, and Crotalus ravus. Because the description that Campbell & Lamar (2004: 431) provided was a composite of A. browni and C. barbouri, they were moved to state that “The number and arrangement of the scales on top of the head appear to be more variable than those reported for any other snake.” With the recognition of A. browni as a taxon separate from C. barbouri, the misconception of this statement is now apparent. This discovery prompts a number of questions, the first that of relationship. Only two phylogenetic analyses have been conducted on C. barbouri (Campbell, 1988; Jadin, 2010), both were morphological analyses of

Cerrophidion and include specimens of A. browni as C. barbouri. These two analyses lacked taxonomic sampling beyond Cerrophidion species and therefore it is not surprising that they found a basal split between C. barbouri and the other Cerrophidion. In the original description of

A. browni, Shreve (1938) mentioned that the species was probably most closely related to members of New World Agkistrodon, but was allocated to the genus due to its similarity to the

Asian pitviper Hypnale [Agkistrodon] hypnale. Therefore, its relationships were unclear prior to our study.

Our phylogenetic analyses reveal that A. browni and C. barbouri form a clade with

Ophryacus melanurus, a sister-group to O. undulatus. Additionally, A. browni, C. barbouri, and

26 O. melanurus share several seemingly derived features from O. undulatus (e.g., entire subcaudals, long and curved tail spine, flat canthals, presence of palatine teeth, and separated splenial and angular bones). The paraphyly of the genus Ophryacus with respect to A. browni and C. barbouri (Fig. 2.5) was not anticipated but provides a feasible biogeographic scenario, since all of these taxa are highland endemics to southern Mexico. We believe that the ancestor of these four species inhabited the area between the Isthmus of Tehuantepec and the Balsas River drainage. An O. undulatus ancestor appears to have separated first during the mid-Miocene within the more mesic forests to the south and eventually invaded all other mesic areas of

Oaxaca, Guerrero, Veracruz, and Puebla. The ancestor of A. browni and C. barbouri probably speciated within the Sierra Madre del Sur of Guerrero, west of Chilpancingo, giving an offshoot into the dry Valleys of Oaxaca, Morelos, and Puebla, O. melanurus, what could be considered the high southern Balsas area and associated valleys and drainages.

Basis for systematic revision

When Gutberlet (1998) removed Ophryacus melanurus from the genus Porthidium and placed it in Ophryacus, based on careful morphological analyses, he recognized it differed greatly from O. undulatus in several highly conserved characteristics (e.g., respectively, terrestrial vs. semi- arboreal habits, typically 3 vs. 0 palatine teeth, entire vs. divided subcaudals). Initially, Gutberlet sought to describe O. melanurus as a monotypic genus due to these numerous divergent features

(R. Gutberlet, pers. comm.). Now, more than a decade later, relative divergence estimates by

Castoe et al. (2009) and Daza et al. (2010) suggests that the O. melanurus lineage (now including Agkistrodon browni and Cerrophidion barbouri) and the O. undulatus lineage diverged from each other during the mid-Miocene. Although the confidence estimates overlap, their mean

27 estimates of divergences for this separation predates the splitting of the Porthidium group into three genera, the rattlesnakes into Sistrurus and Crotalus, and the Bothriopsis–Bothrocophias–

Bothropoides–Bothrops–Rhinocerophis clade.

Additionally, our pairwise comparisons show a divergence of 12.3% within these three species, which falls within the range of intrageneric divergence (Table 2.4). Divergence among these three species and O. undulatus is 14.6%, which falls within the range of intergeneric divergence. Moreover, these three species share several seemingly derived features from O. undulatus (e.g., terrestrial habits, entire subcaudals, long and curved tail spine, flat canthals, presence of palatine teeth, fewer intersupraoculars, and separated splenial and angular bones).

Therefore, on the basis of genetic distance and distinctive morphology there appears little doubt that A. browni, C. barbouri, and O. melanurus warrant allocation to their own new genus, rendering Ophryacus monophyletic. We hereby propose a new genus and summarize the morphological features for each of the species placed in this genus. Finally, the removal of C. barbouri, and thus A. browni, and O. melanurus from the genera Cerrophidion and Ophryacus, respectively, requires a revision of our concepts of these two genera. We therefore revise the genera Cerrophidion and Ophryacus.

Systematic account

Mixcoatlus

TYPE SPECIES: Agkistrodon browni Shreve, 1938.

ETYMOLOGY: The generic name is derived from the Náhuatl word Mixcoatl, meaning “cloud serpent,” a god of the Aztecs and several Mesoamerican civilizations. The name alludes to the restriction of this clade to high elevations. The gender of this name is masculine.

28 CONTENT: The genus Mixcoatlus contains M. barbouri, M. browni, and M. melanurus. Similar to

Ophryacus, Mixcoatlus is a pitviper genus endemic to the highlands of southern Mexico.

Mixcoatlus barbouri and M. browni are restricted to highland humid pine-oak and cloudforest habitats of the Sierra Madre del Sur in Guerrero, Mexico (Fig. 2.6), whereas M. melanurus occurs in highland arid tropical scrub, high deciduous forest, and seasonally dry pine-oak forest in southern Puebla and northern Oaxaca (Campbell & Lamar, 2004, map 83). This limited distribution of southern Mexico makes this genus the most restricted of New World pitvipers.

COMMON NAME: Mexican Montane Pitvipers

Figure 2.6. Distribution map of Mixcoatlus barbouri and Mixcoatlus browni.

DEFINITION AND DIAGNOSIS: Rostral broader than high, front surface flat to moderately concave

(M. melanurus); preoculars 2 (M. barbouri and M. browni) or 3 (M. melanurus), upper preocular largest and squarish, in M. melanurus middle preocular separate from supralacunal, lower

29 forming posterior border of pit and excluded from orbit; single, large, flat, platelike supraocular above eye (M. barbouri and M. browni) or 2–3 supraoculars along dorsal margin of eye including supraocular horn (single scale above eye forming flattened horn, dorsoventrally compressed in cross section, occupying most of dorsal margin of orbit, tip broadly rounded; adjacent scales along dorsal ocular margin slightly modified, projecting slightly or not); 7–14 supralabials (usually 8 in M. barbouri and M. browni and 11 in M. melanurus); lip margin strongly scalloped in M. melanurus; 8–13 infralabials; canthals and internasals relatively large, flat to rounded; crown of head covered with relatively large, flat scales with keeling beginning in parietal area (M. barbouri, M. browni) or covered by small keeled scales (M. melanurus); intersupraoculars 1 (M. browni), 3–4 (M. barbouri), or 9–13 (M. melanurus); second supralabial discrete from prelacunal (these scales may be separated by two rows of small subfoveals in M. melanurus); supralabial and subocular series in contact (M. barbouri, M. browni) or separated by

2–4 rows of small, roundish scales (M. melanurus); 1–2 postoculars; 17–21 middorsal scale rows; middorsal scales at midbody moderately slender and pointed in M. barbouri and M. browni and broad and obtusely rounded in M. melanurus; keel generally extending to tip of scale or nearly so, apical pits not apparent; free portion of apex of dorsal scales moderate in extent; 129–

148 ventrals in M. barbouri and M. browni, 137–169 in M. melanurus; subcaudals undivided,

26–35 in M. barbouri and M. browni and 42–64 in M. melanurus; tail spine straight or distally curved upwards, moderately long. In M. barbouri and M. browni dorsum usually with ill-defined zigzag stripe bordered narrowly with black, sometimes broken into discrete blotches; 25–28 dark brown lateral body blotches; dorsal ground color reddish brown. In M. melanurus dorsum with zig-zag pattern; ground color reddish brown, olive brown, or gray; dorsal scales usually finely

30 mottled or speckled with black, although this pattern may be apparent only under microscopic examination.

In M. barbouri and M. browni lateral edge of nasal bone expanded into roughly triangular shape; frontal bones mostly flat, dorsal surface with slightly elevated margins, longer than wide; postfrontal moderate in size, reaching frontal; transverse distance of postfrontal about equal to its distance along parietal bone; posterolateral edges of dorsal surface of parietals forming moderately distinct raised ridge continuing posteriorly on parietal to about level posterior to quadrate; junction between parietal and pro-otic rounded; squamosal extending posteriorly to level about equal to posterior edge of exoccipital; ectopterygoid much shorter than expanded, flattened base of pterygoid (posterior to the articulation with ectopterygoid), with flat shaft gradually tapering posteriorly; dorsal edge of palatine rounded. Three palatine teeth; 10–12 pterygoid teeth; 8–12 dentary teeth; pterygoid teeth not extending posterior to level of articulation of pterygoid with ectopterygoid; maxillary fang relatively short, being about equal in length to height of maxilla; fang at rest extending to level of about middle of supralabial 5.

In M. melanurus frontal bones with concave dorsal surface, strongly elevated margins, moderately longer than wide; postfrontals relatively small, not contacting frontal, comprising considerably less of dorsal perimeter of orbit than parietals; posterolateral edges of dorsal surface of parietals forming distinct flat shelf not continuing onto the parietal as a raised ridge; junction between parietal and pro-otic irregular, not particularly angular; anterior portion of ectopterygoid possessing shallow depression on medial side accomodating attachment of ectopterygoid retractor muscle; ectopterygoid noticeably longer than expanded, flattened base of pterygoid

(posterior to articulation with ectopterygoid) with flat shaft tapering posteriorly; apex of choanal process positioned at about midlength on palatine, process greatly reduced in height, apex

31 broadly rounded; dorsal surface of parietal roughly triangular; 3 palatine teeth, 7–10 pterygoid teeth, 7–9 dentary teeth; pterygoid teeth extending to level of articulation of pterygoid with ectopterygoid; maxillary fang relatively short, only slightly longer than height of maxilla, at rest extending to level of suture between supralabials 6–7 or supralabial 7; splenial and angular bones separate; haemapophyses separate distally.

The highland isolation of Mixcoatlus results in its allopatry to most species of pitvipers.

However, these three species are sympatric with Ophryacus undulatus throughout parts of their range but are distinguished by morphological features listed above. Additionally, M. barbouri and M. browni may be broadly sympatric with Crotalus intermedius and C. ravus but are distinguished from these species by not having a rattle at the end of their tail.

Cerrophidion Campbell & Lamar, 1992

TYPE-SPECIES: Bothriechis godmanni Günther, 1863, by subsequent designation of Campbell &

Lamar (1992).

ETYMOLOGY: The generic name comes from the Spanish cerro, meaning mountain, an allusion to the habitat, and the Greek ophidion, meaning small snake (Campbell & Lamar, 1992).

CONTENT: The genus Cerrophidion contains three species: C. godmani, C. petlalcalensis, and C. tzotzilorum. These species occur in pine-oak and cloud forests from Veracruz (Mexico) southward through the highlands of Central America to Panama (Campbell, 1985; Campbell &

Lamar, 2004: maps 79, 80) with a vertical distribution from ca. 1400–3491 m.

COMMON NAME: Middle American Montane Pitvipers.

DEFINITION AND DIAGNOSIS: Rostral wider than high, front surface flat; three preoculars, upper largest, entire, and squarish, lower forming posterior border of pit and excluded from orbit;

32 single, large, flat, platelike supraocular above eye; 7–11 supralabials; 8–12 infralabials; canthals and internasals relatively large and flat; 2–7 intersupraoculars; crown of head covered with variably sized, flat or keeled scales; keeling prominent in parietal area; second supralabial discrete from prelacunal; supralabial and subocular series in contact or separated by single row of scales; 19–23 (mode 21) middorsal dorsal scale rows; middorsal scales at midbody moderately slender and pointed; 120–150 ventrals; 22–36 undivided subcaudals; tail spine straight, moderately long.

Lateral edge of nasal broadly expanded, bone roughly quadrangular; frontal bones mostly flat, dorsal surface with slightly elevated margins, longer than wide; postfrontal large, not reaching frontal; transverse distance of postfrontal greater than its distance along parietal bone; posterolateral edges of dorsal surface of parietals forming low to moderately distinct raised ridge continuing posteriorly on parietal as low ridge; junction between parietal and pro-otic rounded to almost flat; squamosal extending to level posterior to posterior edge of exoccipital; ectopterygoid about same length as expanded, flattened base of pterygoid (posterior to the articulation with ectopterygoid) with flat shaft gradually tapering posteriorly; dorsal surface of parietal roughly triangular to sometimes rounded; 3–5 palatine teeth; 7–18 pterygoid teeth; 8–16 dentary teeth; pterygoid teeth extending just posterior to level of articulation of pterygoid with ectopterygoid in

C. godmani, but not reaching this far back in congeners; maxillary fang relatively short, being about equal in length to height of maxilla; fang at rest extending to level of about middle of supralabial 5 or suture between supralabials 5–6 (mostly after Campbell & Lamar, 2004).

Ophryacus Cope, 1887

TYPE-SPECIES: Trigonocephalus [Atropos] undulatus Jan (1859), by monotypy.

33 ETYMOLOGY: The generic name is derived from the Greek ophrys, meaning brow, and the Latin acus, meaning pointed, obviously in reference to the distinctive supraocular spine-like scale.

CONTENT: The genus Ophryacus contains only O. undulatus confined to the highlands of the

Sierra Madre Oriental (Hidalgo, Veracruz, Puebla), the Mesa del Sur (Oaxaca), and the Sierra

Madre del Sur (Oaxaca, Guerrero), where it occurs in pine-oak and cloud forest (Campbell &

Lamar, 2004: map 84).

COMMON NAME: Mexican Horned Pitviper.

DEFINITION AND DIAGNOSIS: Rostral broader than high, moderately to distinctly concave; 3 preoculars, upper largest and undivided, middle not fused with supralacunal, lower small, somewhat excluded from margin of orbit; 3–4 supraoculars along dorsal margin of eye including supraocular spine; 10–13 supralabials; lip margin not scalloped; 9–14 infralabials; single scale above eye forming long, relatively slender spine, slightly compressed to subcircular in cross section, not occupying most of dorsal margin of orbit, tip pointed; adjacent scales along dorsal ocular margin often also modified, projecting slightly; canthals and internasals often raised into short spines or with especially high keels; scales in the supraocular region small and keeled; 10–

20 (usually 12–18) intersupraoculars; top of head covered with small scales, most having tubercular keels; second supralabial usually separated from prelacunal by single small subfoveal; subocular and supralabial series separated by 2–4 rows of small, roundish scales; 21 middorsal scale rows; middorsals at midbody not noticeably broad, obtusely rounded; keel generally extending to tip of scale or nearly so, apical pits not apparent; free portion of apex of dorsal scales moderate in extent, barely overlapping contiguous scale; interstitial epidermal fold at cranial end of scale well developed; 157–178 ventrals; 37–57 subcaudals, divided; tail spine straight, about as long as preceding 2–3 subcaudals, pointed or obtusely rounded.

34 Frontal bones with concave dorsal surface, strongly elevated margins, moderately longer than wide; postfrontals moderate in size, not contacting frontal, comprising about equal amount of dorsal perimeter of orbit as parietals; posterolateral edges of dorsal surface of parietals forming distinct flat shelf continuing onto parietal as a raised ridge; junction between parietal and pro-otic irregular, not particularly angular; anterior portion of ectopterygoid possessing a shallow depression on medial side accommodating attachment of ectopterygoid retractor muscle; ectopterygoid noticeably longer than expanded, flattened base of pterygoid (posterior to articulation with ectopterygoid) with flat shaft tapering posteriorly; apex of choanal process positioned at about midlength on palatine, process moderately reduced in height, apex broadly rounded; dorsal surface of parietal roughly triangular; 0–1 (usually 0) palatine teeth, 7–10 pterygoid teeth, 7–9 dentary teeth; pterygoid teeth extending to level of articulation of pterygoid with ectopterygoid; maxillary fang relatively short, only slightly longer than height of maxilla; fang at rest extending to level of suture between supralabials 7 and 8; splenial and angular bones fused; haemapophyses in contact distally.

Dorsum with zig-zag pattern; ground color olive-brown, green, or gray, sometimes orange or yellow pigment present; dorsal scales usually finely mottled or speckled with black.

Natural history of Mixcoatlus barbouri and M. browni

Similar to Ophryacus, Mixcoatlus is a pitviper genus endemic to the highlands of southern

Mexico. Mixcoatlus appears to be found only in the western portion of the Sierra Madre del Sur of Guerrero (M. barbouri and M. browni) and northwestern Oaxaca and southeastern Puebla (M. melanurus), making it the most restricted genus of New World pitviper (Fig. 2.6 and map 83 in

Campbell & Lamar, 2004, respectively). Mixcoatlus barbouri and M. browni are probably

35 sympatric throughout much of their ranges, possess the same type-locality (Omilteme, Guerrero), and have been found near each other in the western part of their ranges. Additionally, Crotalus intermedius omiltemanus, Crotalus ravus, and Ophryacus undulatus have also been found near

Omilteme and other highland areas in Guerrero, making their sympatry with M. barbouri and M. browni likely (Davis & Dixon, 1959; Campbell & Lamar, 2004). The highest confirmed elevation records for M. barbouri and M. browni are 2608 m (MZFC 21432) and 3296 m (KU

182762), respectively.

Campbell (1988) provides a detailed description of M. barbouri and M. browni habitat in the

Sierra Madre del Sur and states that the higher elevations are dominated by pine-oak forest and cloud forest. Although it has been suggested that M. barbouri and M. browni inhabited cloud forest almost exclusively, observations of this species at lower elevations (Davis & Dixon, 1959;

Campbell, 1988; this study) suggest that these species also occurs in upper pine-oak forest where it interdigitates with cloudforest. Additionally, two individuals of M. browni, one found in fir- pine-oak forest (UTA R-4450) and the other “in bunchgrass on the sparsely wooded southern slope of Cerro Teotepec (KU R-182762)”, suggests that this species “at least inhabits several recognizable vegetation associations” (Campbell, 1988:8). Campbell (1988) also provides an investigation of the locality records for most of the specimens of M. barbouri and M. browni.

Both M. barbouri and M. browni are diurnal and are usually found basking, under cover, or moving during the day. While photographing a live M. browni, RCJ and ENS observed an individual wrapping its tail around the hook to prevent falling (e.g., Fig. 2.4). This same behavior was also observed in the field in two other individuals. Although its tail is more prehensile than such terrestrial genera as Cerrophidion and Atropoides, we have no compelling evidence that M. browni is highly arboreal. It does ascend into low vegetation; JAC observed one specimen coiled

36 on top of a stump about 1.5 m above the ground and another in a low, woody shrub about 1.0 m high.

Few M. barbouri and M. browni have been kept in captivity. One specimen of M. barbouri

(UTA R-15558) was kept for more than ten years (Campbell, 1988). An adult M. barbouri

(MZFC 21432) and two M. browni (MZFC 21431 & UTA R-56265) were kept in captivity for several months. The M. barbouri was quite active and readily ate domestic white mice, while one adult (MZFC 21431) and one juvenile (UTA R-56265) M. browni required force feeding (A.

Carbajal, pers. comm.). Specimens and scats of both M. barbouri and M. browni have contained rodent hair as well as the lizard Mesaspis gadovii (Campbell, 1988). Mesaspis gadovii probably constitutes a large portion of the diet for these two pitvipers because of the great abundance of this lizard (Campbell, 1988). Two specimens of M. browni (MZFC 21431 & UTA R-56264) were captured within a few meters of M. gadovii. Although only a few diet items have been identified, the diet of M. barbouri and M. browni probably includes lizards, orthopterans, and mammals, similar to that of Cerrophidion species, with less of their diet consisting of birds and amphibians (Campbell, 1988; Campbell & Solórzano, 1992; Campbell & Lamar, 2004; Jadin,

2007, 2010). Sceloporus adleri is abundant at these high elevations, representing another potential prey item. Many specimens of Thorius and Pseudoeurycea were collected in the vicinity of M. browni. A plethodontid salamander was found in the stomach of Cerrophidion petlalcalensis (Lόpez-Luna et al., 1999) and M. barbouri and M. browni may also consume them.

Of the 32 specimens of Mixcoatlus [Agkistrodon] browni and M. [Cerrophidion] barbouri examined in this study, 15 were M. barbouri and the other 17 specimens allocatable to M. browni (see Appendix 2.1 for details). We are aware of only three museum specimens not

37 examined by us: MZFC 2880 and 2882 and a recently collected M. browni (field number JAC

27714).

Future of pitviper discoveries

Pitvipers have received abundant attention from many scientists involved in molecular and morphological phylogenetics and represent one of the more studied reptile clades. Nonetheless, many new species have been discovered in recent decades and more is continually being revealed about their intriguing evolutionary and natural histories. Pitviper research has been conducted in Mexico by many individuals over the past century, with much of the early groundwork laid by scientists such as Dugès (1896), Cuesta-Terrón (1921) and Martín del

Campo (1935). The recognition of this new genus and continual endemic pitviper discoveries like Porthidium hespere (Campbell, 1976), Cerrophidion tzotzilorum (Campbell, 1985), C. petlalcalensis Lopez-Luna et al. (1999), and Crotalus ericsmithi Campbell & Flores-Villela

(2008) in Mexico underscores the importance of natural history collections already established and additional collecting needed in biotically rich regions of the world. In our current age of biodiversity decline, it is paramount that systematics, ecology, and natural history research in these regions proceed rapidly. Current rates of extinctions, habitat loss, and other anthropogenic changes undoubtedly will make investigations much less rewarding to future generations.

38 Appendix 2.1

Specimens examined. Museum acronyms follow Leviton et al. (1985).

Mixcoatlus barbouri

MEXICO: Guerrero: Omilteme (MCZ R-43283, USNM 46347 [holotype]); 3.2 km W of

Omilteme, 2377 m (TCWC 9455); 4.0 km SW of Omilteme, 2591 m (TCWC 10803); 2.0 km

NW of Omilteme, 2500 m (UTA R-6231); 2.5 km SW of Omilteme, 2490 m (UTA R-15558

[skin and skeleton]); captive bred from UTA R-6231 and UTA R-15558 (UTA R-6181–83,

6237–39); Omiltemi, camino a El Chayotillo (MZFC 2879); Omiltemi, 2300 m (MZFC 2881);

El Balcón: El Moreno, 2608 m (MZFC 21432).

Mixcoatlus browni

MEXICO: Guerrero: Omilteme, (MCZ R-42678 [holotype] & MCZ R-42679 [paratype]); vicinity of Chilpancingo (AMNH R-72478); “Chilpancingo” (FMNH 38503 [skin and skeleton],

38504); near “Chilpancingo” (MVZ 45253); 9.4 km NE of Puerto del Gallo, 3296 m (KU R-

182762); Road from El Jilguero to Puerto del Gallo, 3192 m (MZFC 21431) and 3103 m (UTA

R-56264); 0.8 km N of Puerto del Gallo, 2896 m (UTA R-4450); Carrizal de Bravo (UTA R-

53010); 3.4 km SSE Carrizal de Bravo, 2607 m (MZFC 16677); Asoleadero, road to Cerro

Teotepec from Milpillas (CAS 134466, 135274); El Balcón: El Filo, 2250 m (MZFC 21429); El

Balcón: La Llave, 1826 m (MZFC 21430 & UTA R-56265).

39 CHAPTER 3

CRYPTIC DIVERSITY IN DISJUNCT POPULATIONS OF MIDDLE AMERICAN MONTANE 2 PITVIPERS: A SYSTEMATIC REASSESSMENT OF CERROPHIDION GODMANI

Abstract

The discovery and taxonomic recognition of cryptic species has become increasingly frequent with the application of molecular phylogenetic analyses, particularly for species that have broad geographic distributions. Cerrophidion godmani is a wide-ranging pitviper distributed throughout the highlands of Central America. In this study we provide evidence based on both molecular phylogenetic and morphological data that C. godmani represents multiple species. We further utilize these data, together with known geographic distributions of populations, to infer boundaries of these putative species. Our phylogenetic analysis of four mitochondrial gene fragments (12S, 16S, cyt b, and ND4) reveal that C. godmani, as currently recognized, is not monophyletic, and is composed of three distinct lineages that are comparable in divergence to other recognized species in related pitviper genera. These three lineages are relatively conserved in their morphology, but we find sufficient morphological characters to diagnose them as distinct. Based on the body of evidence, we formally name and describe two new species of

Cerrophidion.

2 Adapted from: R.C. Jadin, J.H. Townsend, T.A. Castoe, and J.A. Campbell. 2012. Cryptic diversity in disjunct populations of Middle American Montane Pitvipers: a systematic reassessment of Cerrophidion godmani. Zoologica Scripta, 41, 455–470.

40 Introduction

Neotropical pitvipers have been the subject of extensive taxonomic and phylogenetic review during the last several decades (see review in Campbell & Lamar, 1992; Gutberlet & Harvey,

2004; Fenwick et al., 2009; Jadin et al., 2011). One particular group that has received a notable amount of attention has been what Burger (1971) recognized as Porthidium, a group at that time considered to consist of eight species. Although Burger’s (1971) dissertation was never formally published, his generic arrangements were adopted by Perez-Higareda et al. (1985), Campbell &

Lamar (1989), and subsequent workers. Campbell & Lamar (1989) recognized three distinct groups within the genus Porthidium: hognosed pitvipers, jumping pitvipers, and montane pitvipers. Werman (1992) removed the jumping pitvipers (P. nummifer, P. olmec, and P. picadoi) from Porthidium, placing them in a new genus, Atropoides, while Campbell & Lamar

(1992) removed the montane pitvipers (P. barbouri, P. godmani, and P. tzotzilorum) and allocated them to a new genus Cerrophidion. Cerrophidion petlalcalensis was later described by

López-Luna et al. (1999), and most recently Jadin et al. (2011) found that the taxon C. barbouri was composed of two species and allocated both, along with Ophryacus melanurus, to a new genus Mixcoatlus. Therefore, as currently understood the genus Cerrophidion contains three recognized species: C. petlalcalensis, C. tzotzilorum, and C. godmani.

Species of Cerrophidion occur in Neotropical montane habitats between approximately 1220 and 3491 m elevation (Campbell & Lamar, 2004; Köhler et al., 2006). Two of these species, C. petlalcalensis and C. tzotzilorum, are endemic to Mexico and are restricted to geographically small ranges (Campbell, 1985; López-Luna et al., 1999; Campbell & Lamar, 2004). The third putative species, C. godmani, is considered to be widely-distributed and occurs throughout highland regions from southern Mexico to western Panama (Campbell & Solórzano, 1992;

41 Campbell & Lamar, 2004). This species occurs in disjunct highland habitats (pine-oak forest, cloud forest, and alpine meadow) both north and south of Nicaragua, with the lowland

Nicaraguan Depression representing a major hiatus in distribution (Campbell & Lamar, 1989;

Campbell & Solórzano, 1992).

Campbell & Solórzano (1992) provided evidence for ecological distinction between populations from Nuclear Central America and Lower Central America. They, as well as Jadin

(2010), further detailed several aspects of morphology that differentiate between the populations in these two main geographical areas. Also during the last several years, a number of molecular phylogenetic and phylogeographic studies have including sampling of multiple C. godmani populations from throughout their range, and have shown evidence that deeply divergent and discrete evolutionary lineages exist within this wide-ranging montane species (Castoe et al.,

2003, 2005, 2009; Daza et al., 2010). Therefore, based on a growing body of evidence, it has become increasingly apparent that systematic revision of Cerrophidion was necessary to establish a species-level taxonomy that reflects the evolutionary history of these lineages.

In this study, we provide morphological and molecular evidence corroborating the hypothesis that C. godmani consists of three species, each restricted to one of three distinct physiographical regions of Middle America. Based on this evidence, we revise the species-level taxonomy of this complex.

Materials and methods

Molecular sampling

Genomic DNA was isolated from muscle tissue removed from twelve specimens of

Cerrophidion using a Qiagen DNeasy extraction kit and protocol. Four mitochondrial gene

42 fragments (NADH dehydrogenase subunit 4 (ND4), cytochrome b (cyt b), 12S rRNA, and 16S rRNA) were independently PCR amplified as described in (Knight & Mindell, 1993; Arévalo et al., 1994; Parkinson et al., 1997, 2002) using Promega GoTaq® Green master mix, the primer pairs: ND4 + LEU, Gludg + AtrCB3, L1091 + 12E, and 16SF + 16SR, and annealing temperatures 48ºC, 48ºC, 50ºC, and 45ºC, respectively. Sequencing was performed in both forward and reverse directions using the PCR primers on a Beckman Coulter automated capillary sequencer, and sequence chromatographs were edited using Sequencher 4.2. Sequences for each gene were aligned separately, first automatically using the program MUSCLE (Edgar, 2004) and, then manually rechecked using Se-Al v2.0a11. Gaps in alignments were treated as missing data and no internal stop codons were found in the two protein-coding gene fragments. Novel sequences from this study were deposited in GenBank (JQ627129–42 and JQ724143–81).

All previously published sequences of Cerrophidion were downloaded from GenBank and were combined with new sequence data generated in this study (Table 3.1). One set of fragments of

Atropoides nummifer, A. olmec, and Porthidium nasutum were used as outgroup taxa to root our

Cerrophidion phylogenetic tree.

Phylogenetic analyses

Bayesian Metropolis-Hastings coupled Markov chain Monte Carlo (BMCMC) methods were implemented to reconstruct phylogenies. To identify appropriate models of nucleotide substitution for BMCMC analyses, we used the program MrModeltest v2.2 (Nylander, 2004), run in PAUP* v4.0b10 (Swofford, 2002). We used Akaike information criterion (AIC) to select the best-fit models, as estimated by MrModeltest (Table 3.2). The four gene fragments were concatenated (2307 total bp), and this combined dataset was partitioned by gene and codon

43 position (for cyt b and ND4), resulting in a total of eight partitions (model 8x in Castoe &

Parkinson, 2006).

Table 3.1. Taxa, vouchers, locality data, and GenBank accession numbers for sequences used in this study. Sequences newly added specifically for this study are in bold. GenBank accession numbers per gene fragment Taxon alt. name voucher Locality 12S 16S Cyt-b ND4 Atropoides nummifer ENS 10515 Mexico: Puebla: San DQ305422 DQ305445 DQ061195 DQ061220 Andres Tziaulan Atropoides olmec UTA R-25113 Mexico: Veracruz: AY223656 AY223669 AY220321 AY220344 Sierra de los Tuxtlas Cerrophidion godmani CR 1 MZUCR 11153 Costa Rica: San José AF057203 AF057250 AY223578 U41879 CR 2 MSM Costa Rica: San José ― ― ― ― ― ― EU684276 EU684293 CR 3 MSM Costa Rica: San José: ― ― ― ― ― ― DQ061200 DQ061225 Goicochea CR 4 MSM Costa Rica: San José ― ― ― ― ― ― AY220328 AY220351 CR 5 MSM Costa Rica: San José ― ― ― ― ― ― EU684277 EU684294 CR 6 MSM Costa Rica: San José: ― ― ― ― ― ― DQ061199 DQ061224 Goicochea CR 7 MSM Costa Rica: San José ― ― ― ― ― ― EU684275 EU684292 ES 1 KU 291242 El Salvador: JQ724143 JQ627129 JQ724156 JQ724170 Chalatenango: Cerro El Pital ES 2 KU 289801 El Salvador: Santa JQ724144 JQ627130 JQ724157 JQ724171 Ana: Montecristo ES 3 SMF 81323 El Salvador: Santa ― ― ― ― ― ― EU693494 ― ― ― Ana: Montecristo GUAT 1 Guatemala: Guatemala ― ― ― ― ― ― EU684278 EU684295 GUAT 2 ENS 8350 & Guatemala: ― ― ― ― ― ― EU684283 EU684300 UTA R-42230 Quetzaltenango GUAT 3 Guatemala: Guatemala ― ― ― ― ― ― EU684279 EU684296 GUAT 4 UTA R-42237 Guatemala: ― ― ― ― ― ― EU684282 EU684299 Huehuetenango: La Democracia GUAT 5 ENS 8195 & Guatemala: Quiché ― ― ― ― ― ― DQ061198 DQ061223 UTA R-42262 GUAT 6 JAC 10458 & Guatemala EU684303 EU684304 EU684280 EU684297 UTA R-14161 GUAT 7 UTA R-40008 Guatemala: Baja DQ305419 DQ305442 AY220325 AY220348 Verapaz: La Unión Barrios GUAT 8 UTA R-32421 Guatemala: Baja ― ― ― ― ― ― EU684281 EU684298 Verapaz GUAT 9 Guatemala: San ― ― ― ― ― ― AY220327 AY220350 Marcos: Esquipulas Palo Gordo HON 1 ENS 10632 Honduras: Francisco JQ724145 JQ627131 JQ724158 EU684301 Morazán: La Tigra HON 2 ENS 10860 & Honduras: Olancho: JQ724146 JQ627132 JQ724159 JQ724172 UTA R-52953 Montaña de Botaderos

HON 3 USNM 578908 Honduras: Olancho: JQ724147 JQ627133 JQ724160 JQ724173 Sierra de Agalta HON 4 UTA R-59480 Honduras: Yoro: JQ724148 JQ627134 JQ724161 JQ724174 Texiguat HON 5 UTA R-59479 Honduras: Francisco JQ724149 JQ627135 JQ724162 JQ724175 Morazán: Montaña de Yoro HON 6 Honduras: Ocotepéque: ― ― ― ― ― ― EU684284 ― ― ― San Antonio de las Ojas HON 7 ENS 10631 Honduras: Ocotepéque: JQ724150 JQ627136 JQ724163 JQ724176 El Güisayote HON 8 UTA R-59478 Honduras: Ocotepeque: JQ724151 JQ627137 JQ724164 JQ724177 El Güisayote

44 HON 9 Honduras: Ocotepéque: ― ― ― ― ― ― EU684285 ― ― ― El Pital HON 10 UF 147608 Honduras: Cortés: El ― ― ― JQ627138 JQ724165 JQ724178 Cusuco HON 11 UF 147612 Honduras: Cortés: El JQ724152 JQ627139 JQ724166 JQ724179 Cusuco HON 12 UF 147610 Honduras: Cortés: El JQ724153 JQ627140 JQ724167 JQ724180 Cusuco HON 13 UF 147613 Honduras: Cortés: El JQ724154 JQ627141 JQ724168 ― ― ― Cusuco HON 14 UF 147611 Honduras: Cortés: El JQ724155 JQ627142 JQ724169 JQ724181 Cusuco MEX 1 JAC 15708 Mexico: Oaxaca: Cerro ― ― ― ― ― ― EU684287 EU684302 Baúl MEX 2 JAC 15709 Mexico: Oaxaca: Cerro ― ― ― ― ― ― AY220326 AY220349 Baúl Cerrophidion petlalcalensis ENS 10528 Mexico: Veracruz: DQ305420 DQ305443 DQ061202 DQ061227 Orizaba Cerrophidion tzotzilorum MEX 1 ENS 10529 Mexico: Chiapas: Las ― ― ― ― ― ― DQ061203 DQ061228 Rosas MEX 2 ENS 10530 Mexico: Chiapas: ― ― ― ― ― ― DQ061204 DQ061229 Zinacantán Porthidium nasutum Pnas1 MZUC R-11150 Costa Rica AF057204 AF057251 AY223579 U41887

Table 3.2. Results from a priori model selections based on Akaike information criterion (AIC) conducted in MrModeltest 2.2 (Nylander, 2004) for partitions of the dataset. Cerrophidion Total characters Parsimony-informative AIC Model phylogeny characters ND4 1st pos 231 27 GTR+Γ ND4 2nd pos 231 11 HKY ND4 3rd pos 231 117 GTR+Γ Cyt b 1st pos 237 23 HKY+I Cyt b 2nd pos 237 10 HKY+I Cyt b 3rd pos 237 111 GTR+Γ 16S 496 14 GTR+I+Γ 12S 407 23 HKY+Γ

BMCMC phylogenetic analyses were conducted using MrBayes v3.0b4 (Ronquist &

Huelsenbeck, 2003). Two simultaneous BMCMC runs were conducted (with the default MCMC settings), and run for a total of 5.0 × 106 generations per run, sampling trees and parameters every 100 generations. We used PSRF values (output by MrBayes), together with plots of cold chain likelihood values and parameter estimates visualized in Tracer v1.5.4 (Rambaut &

Drummond, 2009) to confirm stationarity and convergence of BMCMC runs. Based on this evaluation, the first 1.5 × 105 generations from each run were discarded as burn-in.

45 To obtain estimates of genetic distance between and within the various Cerrophidion species- level lineages, we computed pairwise comparisons of the cyt b gene fragment. We calculated these distances using MEGA v5.05 (Tamura et al., 2011) and in accord with previous studies

(e.g., Fenwick et al., 2009; Jadin et al., 2011) incorporated the Kimura two-parameter model with Γ–distributed rate variation.

Morphological analysis

We examined 201 preserved specimens of Cerrophidion godmani (sensu lato), for this study,

Campbell & Solórzano (1992), and Jadin (2010) (Appendix 3.1). Museum acronyms follow

Leviton et al. (1985). Definitions of scale counts and morphological features follow Campbell &

Lamar (2004). Bilateral characters are measured right/left.

We dissected and examined the left hemipenes from specimens deposited at the Amphibians and Reptile Diversity Research Center at the University of Texas at Arlington (UTA R-51399 and UTA R-59478). Hemipenes were dissected and removed at the base. We fully everted hemipenes by filling them with warm water using a blunt-tipped syringe needle. We removed water and then injected hot petroleum jelly with blue wax-dye until maximum expansion was achieved. Finally, we tied the hemipenes and stored them in 70% ethanol. This procedure is modified from that of Myers & Cadle (2003) and Zaher & Prudente (2003). Hemipenial terminology follows Dowling & Savage (1960), Keogh (1999), and Savage (2002).

Results

Our four mitochondrial gene fragment Bayesian phylogenetic tree suggests that Cerrophidion godmani (sensu lato) is represented by three distinct, well-supported putative species-level

46 clades (pp = 100), which are paraphyletic with respect to the C. petlalcalensis–C. tzotzilorum clade (Fig. 3.1). We find strong support for a sister relationship between C. godmani populations from Guatemala and Mexico (Clade 1) and a C. petlalcalensis–C. tzotzilorum clade. Together, these two lineages are inferred to be the sister group to a clade comprised of two additional distinct lineages of C. godmani populations, one comprising samples from Honduras/El Salvador populations (Clade 2) and a second comprised of Costa Rica/Panama populations (Clade 3).

Cerrophidion godmani from Guatemala and Mexico (Clade 1), as well as C. godmani from

Honduras/El Salvador (Clade 2) exhibit considerable genetic structure across the range of specimens sampled, compared to the relatively low haplotype diversity found in C. godmani populations from Costa Rica.

Description of two new species

Based on the results of the molecular phylogenetic analyses in this study (see Fig. 3.1), C. godmani appears to be a composite of three evolutionarily distinct, deeply divergent lineages meriting recognition as distinct species. Additionally, we find morphological distinctions between these allopatric lineages, further supporting this conclusion, and we consequently describe two of these lineages as new species.

The history of discovery of this species was reviewed by Campbell & Solórzano (1992).

Cerrophidion godmani was described by Günther (1863) from Guatemalan specimens

(speculated type-locality: Totonicapán: Departmento de Sacatepéquez: Guatemala; Boulenger,

1896; Campbell & Solórzano, 1992; Campbell & Lamar, 2004) and we therefore retain the name

C. godmani for populations in Guatemala and Mexico (Clade 1) and describe populations from

Honduras and El Salvador (Clade 2) and Costa Rica and Panama (Clade 3) as new species.

47

Figure 3.1. Bayesian phylogenetic estimate of relationships within Cerrophidion showing three distinct species-level clades within the nonmonophyletic Cerrophidion godmani. The tree represents the Bayesian 50% majority-rule consensus phylogram from a partitioned analysis of sequences from four mitochondrial gene fragments (ND4, cyt b, 12S and 16S; total of 2307 bp). Sample names follow those given in Table 3.1.

48 Genus Cerrophidion Campbell & Lamar, 1992

Cerrophidion sasai (Figs 3.2, 3.3)

HOLOTYPE: Amphibian and Reptile Diversity Research Center, University of Texas at Arlington

(UTA R-51399); an adult male (Fig. 3.2) from San Ramos de Tres Ríos, Departamento de San

José, Costa Rica, on 8 November 2001 by M. Sasa.

A

B

C

10mm

Figure 3.2. Head of Cerrophidion sasai, male holotype, UTA R-51399, showing arrangement of scales. —A. Dorsal, —B. lateral and —C. ventral views.

49

Figure 3.3. Cerrophidion sasai, male paratype in life, UTA R-51403. UTA slide no. 27134. Photo by Eric N. Smith.

PARATYPES: Departamento de Alajuela, Costa Rica: UTA R-35039, male, from Cariblanco, collected in August 1975 by B. Rojas; Departamento de San José, Costa Rica: UTA R-44463–

66, all juveniles, from Las Nubes de Coronado, collected in the spring of 1995 by A. Solórzano;

UTA R-51400, adult female, from Hacienda la Holanda, Nubes de Coronado, collected on 9

November 2001 by M. Sequeira; UTA R-51401–02, adult female and male, respectively, from

Vista del Mar, Guadalupe, Goicochea, collected on 9 November 2001 by J. Aguilar; UTA R-

51403, adult male, from Rancho Redondo, Nubes de Coronado, collected on 30 November 2001 by M. Sequeira.

ETYMOLOGY: The specific epithet is a patronym recognizing Mahmood Sasa Marín, an accomplished Costa Rican herpetologist. Among his many accolades, he is a recent recipient of Award for Young Scientists in 2009 given by the Academy of Sciences for the

Developing World (TWAS) and the Consejo Nacional de Investigaciones Científicas y

Tecnologicas, Costa Rica (CONICIT). He has also published numerous works on what was then

50 C. godmani in Costa Rica such as variation in allozymes across Costa Rican populations (Sasa,

1997), comparative phylogenetic and biogeographic studies of pitvipers in the region (Castoe et al., 2005, 2009) and characterizations of the venom (Durban et al., 2011; Lomonte et al., 2012).

SUGGESTED ENGLISH COMMON NAME: Costa Rica Montane Pitviper

DEFINITION AND DIAGNOSIS: Similar to all other Cerrophidion species, C. sasai is a medium- sized, blotched terrestrial pitviper; head relatively long; distinct and raised canthal ridge, typically two canthals; a large, median frontal plate occupying between 53 and 90% (x =

63.53%) of the distance between the supraoculars; broad supraoculars; nasal divided; prefoveals

0–3; prelacunal single; lacunolabials absent; loreal single; subfoveals 0–2; three preoculars; supralabials 8–10, typically 9 or 10; infralabials 9–12, typically 10 or 11; ventrals are 134–146 (x

= 138.89), undivided subcaudals 25–34 (x = 29.62), no significant sexual dimorphism; cloacal scute undivided; tail relatively short and non-prehensile; and typically 21 middorsal scale rows.

Cerrophidion sasai averages fewer scales in the frontal region, has fewer prefoveals, and tends to have a larger median scale in between the oculars than other Central American Cerrophidion species (Campbell & Solórzano, 1992; Jadin, 2010).

SIMILAR SPECIES: Several other species of pitvipers may occur sympatrically with C. sasai in certain portions of their range. Bothriechis nigroviridis has been reported to co-occur with C. sasai, but likely only in more humid riparian areas (Campbell & Solórzano, 1992). At lower portions of its elevational range C. sasai may occur sympatrically or nearly sympatrically with

Atropoides mexicanus and B. lateralis. These species are easily distinguished from C. sasai by the following characteristics: Bothriechis species have prehensile tails, are typically slender, and are mostly green. The species most closely resembling C. sasai are certain populations of A. mexicanus and A. picadoi but they have broader heads with more numerous and tuberculate

51 supracephalic scales, a narrower and longer dark post-ocular, nasorostrals (except A. picadoi), and are much stockier.

DESCRIPTION OF HOLOTYPE: Rostral broader than high (5.33 x 4.70 mm); 4 internasals anteriorly;

2/2 canthals, 4 posterior intercanthals; supraoculars more than twice as long as broad; 4 intersupraoculars, large median frontal scale occupying 54% of intersupraocular distance; large, flat scales in parietal area; interrictals 23; single loreal bounded by upper preocular and postnasal above, prelacunal, supralacunal, and prefoveals below; prefoveals 1/2 (0/1 tiny, granular), 1 subfoveal, postfoveals 0/0; prelacunal contacting second supralabial; 3/3 preoculars, upper largest, middle small with vertical suture separating it from supralacunal; suboculars 1/1, elongate, crescent-shaped; 2/3 postoculars; supralabials 10/10; mental broader than long (5.19 x

2.92 mm); infralabials 11/10; chin shields contacting first five pairs of infralabials; 3 pairs of gulars between chin shields and first preventral; dorsal scale rows 21-21-19; preventrals 3; ventrals 135; cloacal scute undivided; 33 undivided subcaudals; tail spine as long as preceding 3 subcaudals, straight, tip rounded.

MEASUREMENTS OF THE HOLOTYPE: Total length 71.3 cm; tail length 7.8 cm, comprising 10.9% of total; head 33.84 mm from front face of rostral to posterior end of mandible; head 21.7 mm at broadest point near the rictus; neck 14.1 mm directly behind jaws.

HEMIPENIS DESCRIPTION OF HOLOTYPE: Everted left hemipenis ca.19 mm in total length and 9.5 mm in maximum width at level of the crotch; on sulcate side base with several rows of small spines (< 0.4 mm) for ca. 2 mm, then rows of larger spines and hooks extending for 4 mm, largest protruding ca. 3.5 mm; asulcate side with naked base up to 5 mm before level of bilobation, then with 4 to 5 mm section of small spines (< 1.2 mm) arranged in rows followed by

4.5 mm section of larger spines; each lobe with ca. 60 spines and hooks; ca. 14 spines and hooks

52 around each lobe at the lower rim of calyces; calyces follow spines and hooks distally; calyces scalloped, ca. 19 rows extending 10 mm to apex of the hemipenis on the asulcate side; sulcus spermaticus is deep and bifurcating ca. 3 mm before site of bilobation, extending upwards through spines and calyces to tip of each lobe; border of sulcus spermaticus naked to point of bifurcation where small spines occur on outer border until bilobation at which point both the inner and outer borders of the sulcus spermaticus with small spines for 5 mm to beginning of calyces, which then form border to apex of lobe.

COLOR PATTERN IN LIFE: Dorsal ground color mostly mauve to grayish brown, usually paler on anterior third of body; 28–39 (mean = 32.4) dorsal body blotches, these frequently fused, forming zig-zag dorsal stripe; dorsal blotches dark brown to blackish with dark chestnut brown centers, darker on anterior and posterior portions of the body and uniformly dark; dorsum of tail mostly black; lateral blotches mostly opposite of lateral extensions of dorsal blotches, blackish brown to black, most longer than high, sometimes considerably so, especially on anterior of body where they may extend 10–20 scale lengths; lateral blotches at about midbody cover areas equal to 5–8 scales; dorsolateral interspaces between dorsal and lateral blotches inconspicuously paler than adjacent ground color; top of head medium brown, not as dark as body blotches; side of head pale brown with dark brown to black postocular stripe extending from lower posterior edge of eye to past angle of jaw; postocular stripe usually bordered with a narrowpale line ventrally; anterior portion of venter heavily pigmented with gray black or black mottling, particularly toward lateral part of ventrals, becoming even darker posteriorly; subcaudals, including distalmost, uniformly dark. Rust-colored or orange specimens are known, but do not appear to be common (Campbell & Solórzano, 1992; Campbell & Lamar, 2004).

53 DISTRIBUTION AND HABITAT: The presence of C. godmani (sensu lato) in the highlands of lower

Central America was not reported by Picado (1931) in his early review of the venomous snakes of Costa Rica. Later, however, Picado (1936) reported collecting this species northeast of the city of San José, Costa Rica in the Cordillera Central. Since then, its presence has been well documented in both the Costa Rican and western Panamanian highlands, where it appears to be locally abundant (Campbell & Solórzano, 1992; Savage, 2002; Solórzano, 2004).

The known range of C. sasai includes part of two mountain ranges which together cover portions of Costa Rica and Panama. These highland masses, the Cordillera Central and the

Cordillera de Talamanca, continue unbroken below the 1500 m contour from the vicinity of San

José, Costa Rica, down the primary axis of Isthmian Central America into western Panama. The

Cordillera Central of Costa Rica includes four major volcanoes exceeding 2500 m that are connected by highland ridges above 1500 m. The Cordillera Central is connected to the

Cordillera de Talamanca by the Ochomogo Pass, at approximately 1500 m (Campbell &

Solórzano, 1992). Following its abutment with the Cordillera Central, the Cordillera de

Talamanca maintains a ridgeline of greater than 2500 m west-southwestward through the

Panamanian border into the province of Chiriquí, Panama. Together, these cordilleras form the most extensive highlands in lower Central America and include the highest peaks in the region.

Cerro Chirripó in the Costa Rican extension of Cordillera de Talamanca exceeds 3800 m in elevation.

The Cordillera de Talamanca highland complex receives its moisture via the Northeastern

Tradewinds conveying moisture laden winds off the Caribbean Sea onto the Atlantic versant of

Costa Rica and western Panama, thus forming tracts of wet and moist lower montane forest on these eastern slopes (Campbell, 1999). The leeward sides of these highlands experience a

54 moderate rain shadow creating drier highland habitats, including evergreen and semi-evergreen seasonal forest, which appear to be favored by C. sasai (Campbell & Solórzano, 1992). Scott

(1969) reported the vertical distribution of C. godmani in Costa Rica (C. sasai) as being between

1420 and 2450 m.

In addition to lower montane and montane forest habitats inhabited by C. sasai, this species also inhabits disturbed highland habitats, at least those contiguous with forest fragments. The holotype and one of the paratypes were found on the edge of a forest fragment that was bordered by grassland pasture, and the three other paratypes were collected in grassland. All were found under logs during the day.

Natural History. Gravid females of C. sasai range from 341 to 512 mm in SVL and the number of young produced by a female varies between 2 and 8 (mean = 5.5 ± 1.37) (Campbell &

Solórzano, 1992). The snout-to-vent length, total length, and weight of neonates were all found to be higher in C. sasai than in C. godmani (Campbell & Solórzano, 1992). Reproduction in C. sasai is seasonal with females giving birth between April and June (mode = May), during the latter portion of the dry season and onset of the rainy season (Campbell & Solórzano 1992).

Contrary to our knowledge of reproductive cycles in many Neotropical pitvipers, C. sasai appears to follow a biennial cycle, which may be the result of the constraints posed by its highland distribution (Campbell & Solórzano, 1992).

VENOM: Reported instances of snakebite by this species are rare (Bolaños, 1984). Of a survey including 477 reported cases of snakebite in Costa Rica in 1979, only two were from C. sasai

(Bolaños, 1982). Bolaños (1972) estimated the average venom yield per individual to be 15 mg, with a relatively high LD50 (intravenous) of 76.0 ± 10 µg for 16–18 g mice. Gutiérrez & Chavez

(1980) showed that the hemorrhagic effects of venom of C. sasai were the second most severe of

55 the 10 Costa Rican species examined, requiring only 0.8 µg of venom to result in a hemorrhagic area of 10 mm in diameter in laboratory rats.

Genus Cerrophidion Campbell & Lamar, 1992

Cerrophidion wilsoni (Figs 3.4, 3.5)

HOLOTYPE: Amphibian and Reptile Diversity Research Center, University of Texas at Arlington

(UTA R-52953); an adult female (Figs 3.4, 3.5), taken from Cerro Azul, Parque Nacional

Montaña de Botaderos, Departamento de Olancho, Honduras, 15.37831°N/86.14200°W, elevation 1420 m, on 1 February 2005 at 1530 h by E. N. Smith (original number ENS 10860).

The field party included E. N. Smith, C. Chavez, J. Ferrari-Castro, J. H. Malone, J. L. Murillo, S.

Solis, and A. Sosa. Type-locality (Fig. 3.6A, B).

PARATYPES: (8). UTA R-59478, adult male, from Reserva Biologíca Güisayote, Departamento de Ocotepeque, Honduras, 14.4386°N/89.0624°W, 2195 m, collected on 18 June 2008 by I.

Luque-Montes, M. Medina-Flores, J. Townsend, and L. Wilson. UTA R-59479, subadult female, from Cataguana, Parque Nacional Montaña de Yoro, Departamento de Francisco Morazán,

Honduras, 15.0198°N/87.1221°W, 1910 m, collected on 11 March 2007 by J. Butler, L. Ketzler,

J. Townsend, S. Travers, and L. Wilson. UTA R-59480, a juvenile, from 3.45 km NNE of La

Fortuna, Reserva de Vida Silvestre Texíguat, Departamento de Yoro, Honduras,

15.4451°N/87.3046°W, 1885 m, collected on 10 April 2008 by J. Butler, L. Ketzler, J.

Slapcinsky, N. Stewart, J. Townsend, and L.D. Wilson. USNM 578908, adult female from Los

Tres Cerritos, Parque Nacional Sierra de Agalta, Departamento de Olancho, Honduras,

14.9106°N/86.0126°W, 1690 m, collected 2 January 2011 by M. Medina-Flores and O. Reyes-

Calderón. UF 147610–11, adult males from Quebrada Cantiles, Parque Nacional Cusuco,

56 Departamento de Cortés, Honduras, 15.5111°N/ 88.2448°W, 1760 m, collected 13 and 14 March

2006 by J. Townsend and L.D. Wilson. KU 291242, adult female from Cerro El Pital,

Departamento de Chalatenango, El Salvador, 14.3878°N/89.1167°W, 2485 m, collected on 16

May 2001 by R. Gregorio, R. Bolanos, O. Komar, and K. Gomez. KU 289801, adult gravid female from Cerro Montecristo, Parque Nacional Montecristo, Departamento de Santa Ana, El

Salvador, 14.4011°N/89.3617°W, 2200 m, collected on 15 July 2000 by E. Greenbaum and J.

Porras.

57 A

B

C

10 mm

Figure 3.4. Head of Cerrophidion wilsoni, holotype, UTA R-52953, showing arrangement of scales. —A. Dorsal, —B. lateral and —C. ventral views.

58 Figure 3.5. Cerrophidion wilsoni, female holotype in life, UTA R-52953, 648 mm in total length. Photo by Eric N. Smith.

A B

Figure 3.6. —A. Distant habitat photograph showing Cerro Azul in Parque Nacional Montaña de Botaderos—the type-locality of Cerrophidion wilsoni. —B. Closer photograph of the mountain where the holotype was collected, showing cloud forest that is occupied with Liquidambar. Photographs by Eric N. Smith.

ETYMOLOGY: The specific epithet is a patronym honouring Larry David Wilson. We are pleased to name this species in honour of him, in recognition of his career-long contributions to

Mesoamerican herpetology. His work in Central America began with his first trip to Mexico in

1966 and has resulted in numerous contributions, including two volumes of the book Snakes of

59 Honduras (Wilson & Meyer, 1982, 1985), numerous other books on regional herpetology (e.g.,

McCranie & Wilson, 2002; Townsend & Wilson, 2008), and most recently leading efforts to compile the seminal volume Conservation of Mesoamerican Amphibians and Reptiles (Wilson et al., 2010). He has described over 70 species of Middle American amphibians and reptiles and worked on a broad range of questions in snake systematics.

SUGGESTED ENGLISH COMMON NAME: Honduras Montane Pitviper

DEFINITION AND DIAGNOSIS: Similar to other Cerrophidion species, C. wilsoni is a medium-sized, blotched terrestrial pitviper; head relatively long; distinct and raised canthal ridge, typically two canthals; a frontal plate; broad supraoculars; nasal divided; prefoveals 1–4; prelacunal single; lacunolabials absent; loreal single; subfoveals 0–2; three preoculars; supralabials 7–11, typically

9; infralabials 10–12, typically 11; ventrals are 137–151 (x = 142.21), undivided subcaudals 23–

36 (x = 30.12), no significant sexual dimorphism; cloacal scute undivided; tail relatively short and non-prehensile; and typically 21 middorsal scale rows. Cerrophidion wilsoni differs from other Cerrophidion species in normally having fewer scales that contact the supraoculars, averaging 9 compared to 10–12 in other Cerrophidion species (Jadin, 2010). Additionally, the median frontal scale of C. wilsoni is often quite small, distinguishing it from other Cerrophidion species, and in many individuals the frontal scale is greatly reduced to the size of a normal head scale (Campbell & Solórzano, 1992). In those individuals that contain a small frontal plate, this plate occupies an average of 36% of the distance between the supraoculars (Campbell &

Solórzano, 1992).

SIMILAR SPECIES: Several other species of pitviper are sympatric with C. wilsoni at various localities in Honduras. The palm-pitviper Bothriechis marchi is sympatric with C. wilsoni in

Parque Nacional Cusuco in the Sierra de Omoa in northwestern Honduras, and C. wilsoni is

60 sympatric or nearly sympatric with Atropoides mexicanus and Bothrops asper along the lower edges of the cloud forest (Townsend & Wilson, 2008), a situation mirrored with the same four species in Reserva de Vida Silvestre Texíguat in north-central Honduras (Townsend et al., submitted). There may also be sympatric or near-sympatric populations of Bothriechis thalassinus and C. wilsoni in the vicinity of the shared border between Guatemala, Honduras, and El Salvador. These species are easily distinguished from C. wilsoni by the following characteristics: Bothriechis species have prehensile tails, are typically slender, and are mostly green, except certain color variations in B. schlegelii. The species most closely resembling C. wilsoni are certain populations of A. indomitus and A. mexicanus but they have broader heads, more numerous distinctly tuberculate supracephalic scales, a narrower and longer dark post- ocular stripe, nasorostrals present, and a much stockier habitus.

DESCRIPTION OF HOLOTYPE: Rostral broader than high (4.73 x 4.10 mm); 2 internasals anteriorly;

2/2 canthals, 4 posterior intercanthals; supraoculars more than twice as long as broad; 7 intersupraoculars; scales on head of medium size, lacking the large median frontal scale and large flat scales in parietal area usually associated with Cerrophidion species; interrictals 24; single loreal bounded by canthal above, prelacunal and prefoveals below; prefoveals 5/5 (2/2 tiny, granular), 1 subfoveal, postfoveals 4/4; prelacunal contacting second and third supralabials;

3/3 preoculars, upper largest, middle small with vertical suture separating it from supralacunal; suboculars 2/2, elongate, crescent-shaped; 3/3 postoculars; supralabials 10/10; mental broader than long (4.68 x 2.72 mm); infralabials 12/12; chin shields contacting first four pairs of infralabials; 3 pairs of gulars between chin shields and first preventral; dorsal scale rows 23-21-

17; preventrals 3; ventrals 148; cloacal scute undivided; 33 subcaudals numbers 1, 22–26, 31–33 are divided the rest undivided; tail spine as long as preceding 3 subcaudals, straight, tip rounded.

61 MEASUREMENT OF HOLOTYPE: Total length 64.8 cm; tail length 7.3 cm, comprising 11.3% of total; head 31.95 mm from front face of rostral to posterior end of mandible near the rictus; head

23.0 mm at broadest point; neck 12.3 mm directly behind jaws.

HEMIPENIS DESCRIPTION OF PARATYPE: The holotype for C. wilsoni is a female and therefore we describe the everted left hemipenis of a male paratype, UTA R-59478 (SVL 477 mm, TAL 61 mm, subcaudals 32). Hemipenis ca.17.5 mm in total length and 10 mm in maximum width at level of crotch; on sulcate side base with several rows of small spines (< 0.5 mm) followed by rows of larger spines and hooks extending for 4 mm, largest protruding ca. 3 mm; asulcate side with naked base up to 5 mm before level of bilobation, then with 4 mm section of small spines (<

0.5 mm) arranged in rows followed by 4 mm section of larger spines; each lobe with ca. 70 spines and hooks; ca. 15 spines and hooks around each lobe at lower rim of calyces; calyces follow spines and hooks distally; calyces scalloped, ca. 17 rows extending 10 mm to apex of hemipenis on asulcate side; sulcus spermaticus deep and bifurcating ca. 2.5 mm before site of bilobation and extending upwards through spines and calyces to tip of each lobe; border of sulcus spermaticus naked to point of bifurcation where small spines occur on outer border until bilobation at which point both the inner and outer borders of the sulcus spermaticus have small spines for 3 mm to the beginning of calyces which form border to apex of lobe.

COLOR PATTERN IN LIFE: Dorsal ground color mostly medium or coffee brown, usually paler on anterior third of body; 32–44 (mean = 34.8) dorsal body blotches, these usually fused, forming zig-zag dorsal stripe; dorsal blotches very dark brown or chestnut brown with centers that may be only slightly and inconspicuously paler in centers; dorsum of tail mostly dark brown; lateral blotches mostly opposite of lateral extensions of dorsal blotches, blackish brown to almost black, most subcircular, but may be horizontally elongate on anterior of body where they may extend

62 5–10 scale lengths; most lateral blotches at about midbody cover areas equal to 5–7 scales; dorsolateral interspaces between dorsal and lateral blotches paler than adjacent ground color; top of head medium to dark brown, not as dark as body blotches; side of head below postocular stripe pale brown with dark brown to black postocular stripe extending from lower posterior edge of eye to past angle of jaw, sometimes coalescing with first lateral blotch on neck; postocular stripe with a conspicuous narrow pale border ventrally; labial scales without mottling; gular area pale yellow or orange; anterior ventrals pale with little dark stippling or mottling, venter becoming darker posteriorly; proximal subcaudals dark, distalmost subcaudals invariably pale.

No distinct color phases have been reported but considerable variation has been described

(Campbell & Solórzano, 1992; Campbell & Lamar, 2004). An additional description of C. wilsoni (as C. godmani) in life and in alcohol is found in McCranie (2011:505–506).

DISTRIBUTION AND HABITAT: Cerrophidion wilsoni occurs primarily in lower montane rainforest between 1400 m and 3491 m (Campbell & Solórzano, 1992; Campbell & Lamar, 2004), and may occur peripherally in premontane rainforest and pine-oak forest as low as 1220 m (Wilson &

McCranie, 2004; Köhler et al., 2006; McCranie, 2011). Within lower montane rainforest, this species has been observed in undisturbed forest, in areas of disturbed forest, in and around coffee and other agricultural clearings, and in mountaintop elfin forest and wind scrub habitat. This species occurs in at least 13 isolated highland forest areas across Eastern Nuclear Central

America (as defined by Campbell, 1999), and all known populations of C. wilsoni are found within the borders of Honduras and El Salvador (Wilson & Meyer, 1985; Köhler et al., 2006;

McCranie, 2011). Highland areas that support populations of C. wilsoni in Honduras and El

Salvador also extend into eastern Guatemala and the species very likely occurs in that country.

63 Although Villa (1962, 1984) reported C. godmani s.l. from northern Nicaragua, neither voucher specimens nor precise locality data exist to verify its presence in this country (Campbell

& Solórzano, 1992; Campbell & Lamar, 2004; Jadin, 2010), and repeated visits to highland sites that would be the most likely to support Cerrophidion have not produced any physical evidence of its occurrence in Nicaragua (J. Sunyer and S. Travers, pers. comm.). However, Javier Sunyer

(pers. comm.) shared with us that residents of the Nicaraguan highlands are familiar with a snake they call “toboba de altura”, a highland version of the “toboba” (Porthidium ophryomegas), which they also clearly differentiate from the “mano de piedra” (Atropoides mexicanus).

Furthermore, Mesaspis moreletii—a member of a genus of lizard which is sympatric throughout most of the range of Cerrophidion (sensu stricto) and Mixcoatlus barbouri and M. browni with likely similar fundamental environmental niche space—has recently been collected at Reserva

Natural Cerro Kilambé in the northern Nicaraguan highlands (Sunyer & Köhler, 2007), indicating there is still a possibility that Nicaraguan populations of Cerrophidion may yet be discovered. Furthermore, the highest mountain range in Nicaragua, which includes the country’s tallest peak at Cerro Mogotón (2107 m), straddles the border with Honduras and biological sampling has been limited by the presence of land mines left over from the Contra-Sandinista

War of the 1980’s.

Natural History. This terrestrial species is active during the day. In Parque Nacional Cusuco, it has been repeatedly observed while coiled in direct sunlight at the entrances to small holes in root-masses of fallen trees, as well as basking in and around a large landslide. In the same forest,

C. wilsoni has been observed both coiled and active around foot trails during the day and night, under ground cover during the day, and a large female was active at night crawling along a large log. A male paratype (UTA R-59478) was collected at dusk as it crawled along a road through

64 disturbed cloud forest accessing a mountaintop communications tower in Reserva Biologíca

Güisayote. In Reserva de Vida Silvestre Texíguat in north-central Honduras, a juvenile paratype

(UTA R-59480) was found coiled in direct sunlight on top of a large log at the edge of a recently cleared forest patch. In Parque Nacional Montaña de Yoro, one juvenile was collected while active in leaf litter near a log during the day, and a subadult female paratype (UTA R-59479) was coiled at the edge of a small agricultural clearing near a house. Additional ecological notes for populations herein assigned to C. wilsoni are summarized in Wilson & Meyer (1985), Leenders

& Watkins-Colwell (2004), Köhler et al. (2006), and McCranie (2011).

Reproductive data that can be attributed to C. wilsoni is limited. However, Mertens (1952) reported one gravid female from Cerro Montecristo, El Salvador, that contained nine embryos, and McCranie (2011) reported a female collected in May near the Honduran side of Cerro El

Pital contained 14 embryos and another collected in August from the same locality was gravid.

Discussion

Systematics and evolutionary morphology

Previous phylogenetic studies (Castoe et al., 2005, 2009; Daza et al., 2010) have found weak support for the monophyly of Cerrophidion godmani sensu lato to the exclusion of a C. petlalcalensis–C. tzotzilorum clade. This relationship is not supported in our analyses, and we recovered strong support for C. godmani s.s. as the sister lineage of a C. petlalcalensis–C. tzotzilorum clade, and this cluster together as the sister group to a C. sasai–C. wilsoni clade. Our phylogenetic analysis does, however, agree with these previous studies in recovering strong support for three distinct and deeply divergent species-level clades of C. godmani s.l., which are allopatrically distributed in the highlands of Middle America (see Fig. 3.7). Previous estimates of

65 the divergence time of these clades of C. godmani s.l. infer a shared common ancestor 7.7–11.5 mya, with the most recent divergence between C. sasai and C. wilsoni occurring 3.1–6.0 mya

(Castoe et al., 2009; Daza et al., 2010).

Based on phylogenetic analysis of morphological characters, Jadin (2010) also inferred that

Guatemalan populations formed a clade and El Salvador and Honduras populations form a clade, similar to findings in this study. However, based on this morphological dataset, Jadin (2010) inferred a sister relationship between C. godmani s.l. populations from Costa Rica and Mexico.

Campbell & Solórzano (1992: 236), on the other hand, did not suggest any close relationship between these populations but were the first to point out that populations at the northern and southern extremes of the range retained certain characters that might be considered pleisiomorphic. Thus, the morphology of C. godmani s.l. apparently is more similar at the periphery of its range than populations in the center. While this relationship has never been supported by molecular data (Castoe et al., 2005, 2009; this study), it is intriguing as it suggests that either the geographically peripheral lineages are morphologically convergent, or that these peripheral lineages (C. godmani and C. sasai) have retained ancestral morphological features that the centrally distributed clade (C. wilsoni) has lost (Campbell and Solórzano, 1992).

66 20°

Caribbean Sea 18°

16°

14°

C. godmani C. petlalcalensis 12° C. sasai Pacific Ocean C. tzotzilorum C. wilsoni 10° Above 0 m Above 300 m Above 900 m Above 2000 m

- 98° - 96° - 94° - 92° - 90° - 88° - 86° - 84° - 82°

Figure 3.7. Locality map of Cerrophidion samples throughout Middle America used for the molecular analyses. Symbols correspond to the origin of the samples (taken from Castoe et al., 2009 and new localities).

It is reasonable to wonder why such deeply divergent lineages of obviously allopatric snakes have not been previously recognized as distinct species. One unique characteristic associated with snakes of the genus Cerrophidion is a tremendous amount of morphological variation among individuals within a single population, including numbers of particular scales, as well as substantial asymmetry of scales within individuals, and scale fusion, particularly on the head

(Campbell & Solórzano, 1992; Jadin, 2010). Thus, despite a tremendous amount of morphological variation observed, the extensive variation among individuals within populations has left few obviously diagnostic characters by which populations could be readily distinguished.

Often when external morphology presents such challenges, hemipenial characters may be useful to differentiate and diagnose species, but in Cerrophidion there is only slight distinctions in

67 hemipenes among the five species (see Fig. 3.8A–D in this study and Campbell & Lamar, 2004:

Figs 148–150). Thus, only with the added perspective of molecular data, together with morphological evidence and better-defined geographical distributions of populations based on many years of field collections, have distinctions of the various lineages become obvious, leading to the present systematic arrangement.

Conservation

The recognition of Cerrophidion sasai and C. wilsoni adds to the rapidly growing list of new and cryptic taxa being discovered and described from highland forests in Central America

(Townsend et al., 2011). In Honduras, much of the known and potential habitat for Cerrophidion wilsoni is found within protected areas, a fortuitous result of the 1987 law that established the

Honduran Protected Areas System (SINAPH), which ostensibly protects the sources of potable water (i.e. highland forests above 1800 m) for people living in the lowlands (Townsend &

68 A B

C D

5 mm

Figure 3.8. Sulcate (A, C) and asulcate (B, D) views of the left hemipenes of Cerrophidion sasai (holotype, UTA R-51399) and C. wilsoni (paratype, UTA R-59478), respectively.

69 Wilson, 2010). In 2011, the newest protected area in Honduras was established, Parque Nacional

Montaña de Botaderos, which includes within its limits the type locality of C. wilsoni. Including

Parque Nacional Montaña de Botaderos, C. wilsoni is known to occur within the boundaries of at least 15 protected areas in Honduras, encompassing the entire known range of the species in that country: Parque Nacional Celaque, Parque Nacional Cusuco, Parque Nacional La Tigra, Parque

Nacional Montaña de Botaderos, Parque Nacional Montaña de Yoro, Parque Nacional Sierra de

Agalta, Parque Nacional Trifinio Montecristo, Refugio de Vida Silvestre La Muralla, Refugio de

Vida Silvestre Mixicure, Refugio de Vida Silvestre Texíguat, Reserva Biologíca Cordillera de

Opalaca, Reserva Biologíca Guajiquiro, Reserva Biologíca Güisayote, Reserva Biologíca El

Pital, and Reserva Biologíca Yerba Buena. Honduras has the highest rate of herpetofaunal endemism of any Central American country (Wilson & Johnson, 2010), due largely to having over 30 isolated highland forest areas that each support their own unique suite of endemic reptiles and amphibians (Wilson & McCranie, 2004). A similar situation is found in El Salvador, where known populations of C. wilsoni occur within the boundaries of Parque Nacional

Montecristo (the Salvadoran side of Parque Nacional Trifinio Montecristo). Given that C. wilsoni occurs in at least half of the declared highland forest protected areas in Honduras and El

Salvador, and likely is simply undocumented as occurring in others, we would propose this emblematic representative of highland biodiversity as a “flagship” species for cloud forest conservation.

70 Appendix 3.1

Specimens examined. Museum acronyms follow Leviton et al. (1985).

Cerrophidion godmani

GUATEMALA: Alta Verapaz: Finca Chichén, 1410 m (UMMZ 91079–80); Baja Verapaz:

7.7 km S Purulhá, 1615 m (UTA R-5755–56 UTA R-5904–06); Chimaltenango: Acaenango,

Vuelta Siete Pecados, 2134 m (UMMZ 107034); Chichavac, 2600 m (FMNH 20264–71, 20645–

46); 14.0 km N Patzún, 2750 m (MVZ 109421); Quisaché, 1750 m (CM 41784–96, CM 41797a– c); 8.0 km N Tecpán (CAS 67026–33); near Tecpán, near ruins of Ixmiche (UMMZ 131105); near Yepocapa (UMMZ 100512); Guatemala: Los Ocale, 2000 m (UMMZ 106745);

Huehuetenango: Finca El Injerto [ca. 3.2 km N Casa Grande], 1981–2134 m (UMMZ 126433–

34); Paraíso, 1625–2200 m (UMMZ 127209–19, 127220-a–d); 2.0 km E San Juan Ixcoy, 2100 m

(UMMZ 120033); Todos Santos, 2438–2450 m (UMMZ 89214-a–f, 89215, 120520); Quiché:

Uspantán, Aldea El Chimle, ca. 2000–2400 m (UTA R-42266, 45597, 45599–45602, 45606,

45608, 45610); Nebaj, 1920 m (MCZ R-25213, 49650); Sacatepéquez: path up Volcán de Agua from Santa María de Jesús, 2580 m (MVZ 131726); San Marcos: Aldea La Fraternidad, Finca

La Esperanza, 1825 m (UTA R-38259–60, 38265–67, 38269–73); Sololá: near San José, ca.

2200 m (UMMZ 128804); Zacapa: Sierra de las Minas, Santa Clara (FMNH 64713, 64713-a–f,

167115); MEXICO: Chiapas: 13.7 km N Unión Juárez, Volcán Tacaná, 2000 m (KU 94139–40);

UMMZ 94641–42, 95151–53); Oaxaca: Cerro Baúl, 19 km NW of Rizo de Oro, Chiapas (UTA

R-5917, 6635, 6642–43, 7693, 7706, 8782).

Cerrophidion sasai

COSTA RICA: Cartago: Tres, Ríos, San Ramón (MZUCR 6279–84); 2.5 km W Sanatario

Durán, Volcán Irazú, 2374 m (USC-CRE-158); Cerro de la Muerte, Vicinity of La Trinidad,

71 2560 m (USC-CRE-1209); Turrialba (UMMZ 131325). San José: Hacienda La Holanda, Nubes de Coronado (UTA R-51400); Rancho Redondo, Coronado (UTA R-51403); San Ramos de Tres

Rios (UTA R-51399); Vista del Mar, Guadelupe, Goicochea (UTA R-51401–02); Cascajal, zona alta de San Isidro de Coronado (MZUCR 1122); Dota, alto El Jardin (MZUCR 5312); Dota,

Ladera S de Valle de Copey (MZUCR 3334); Dota, Llano San Lucas (MZUCR 6093); Dota,

Valle de Quebradas (MZUCR 5311); Dota, Quebradillas (MZUCR 5476); Moravia (UMMZ

131326); Rancho Redondo (MZUCR 2521); Between San José and Cartago, above Tres, Ríos hydroelectric plant, ca. 1829 m (UF 10439). No specific locality: (ANSP 22428, UF 30704–07).

PANAMA: Chiriqui: Cerro Punta (KU 112573); El Volcán, 1829 m (ANSP 21344, 22572–73,

22579–80); No specific locality: (TCWC 33465–66).

Cerrophidion wilsoni

EL SALVADOR: Chalatenango: Cerro El Pital, 2485 m (KU 291242); E slope Los

Esesmiles, 2134–2438 m (MVZ 40453–71); Los Esesmiles (FMNH 11000); Santa Ana: Cerro

Montecristo, Parque Nacional Montecristo, 2200 m (KU 289801); Miramundo Mountain, above

Hacienda Los Planes, 1950 m (KU 62136); Hacienda Montecristo, Metapán Mountain, 2200 m

(KU 62134–35, 63918, UMMZ 117285–86); HONDURAS: Cortes: Sierra de Omoa,

Chamelecon [possibly in error, see McCranie 2011] (MCZ R-27220); Mountains W of San Pedro

Sula (FMNH 11370); San Pedro Sula (MCZ R-27222, 27561, 33337, 33558, 33559); San Pedro,

La Cumbre [possibly in error, see McCranie 2011] (MCZ R-29410); Francisco Morazán:

Cataguana, Parque Nacional Montaña de Yoro, 1850–1910 m (JHT 2123, UTA R-59479); San

Juancito Mountains (AMNH 70165–67, ANSP 26666); Ocotepeque: Reserva Biologíca

Güisayote, 2170 m (UTA R-59478); Olancho: Cerro Azul, Parque Nacional Montaña de

Botaderos, 1420 m (UTA R-52953); Los Tres Cerritos, Parque Nacional Sierra de Agalta, 1690

72 m (USNM 578908); Yoro: 3.45 km NNE of La Fortuna, Reserva de Vida Silvestre Texíguat,

1885 m (UTA R-59480); No specific locality: (ANSP 27697, UTA R-32138).

73 CHAPTER 4

FINDING ARBOREAL SNAKES IN AN EVOLUTIONARY TREE: PHYLOGENETIC PLACEMENT AND SYSTEMATIC REVISION OF THE NEOTROPICAL BIRDSNAKES

Abstract

The genus Pseustes Fitzinger, 1843 is composed of two to five recognized species depending on the authority, including what may be the largest-sized colubrid snake in the New World. The group has a complex systematic history that has yet to be untangled using modern molecular phylogenetic approaches. The systematic position, within-group diversity, and distribution are therefore uncertain. We obtained samples of four species from multiple specimens across their distribution and analyzed one nuclear and two mitochondrial genes to determine the phylogenetic placement of the genus and infer relationships among Pseustes lineages. We find strong support for the paraphyly of the genus Pseustes with respect to the monotypic genus Spilotes, both of which are nested within a clade of at least 23 other New World Colubrinae genera. Finally, we formally revise the taxonomy of P. poecilonotus and P. sulphureus based on our results and identify two lineages that are putatively new and currently unrecognized species.

74 Introduction

Origin and evolution of the Neotropical fauna remains of great interest, especially given the complex geological history, habitat heterogeneity and high levels of endemicity of the region

(Cracraft & Prum, 1988; Duellman, 1999; Daza et al., 2009). However, well-resolved phylogenies are still lacking for many organisms and thus the evolutionary histories as well as spatial and ecological distributions remain poorly known for a majority of taxa (Daza et al.,

2009; Schargel et al., 2010). Despite recent advancements in our understanding of evolutionary relationships of Neotropical colubrid snakes, which represent a majority of snake taxa in the

New World (e.g., Pyron et al., 2011), most taxa, including several genera, have yet to be examined using molecular phylogenetic approaches. Here we begin to assemble one piece of that puzzle by focusing on the Neotropical snake genus Pseustes, commonly referred to as

“birdsnakes” or “puffing snakes.” Between two and five species are known, depending on the authority (e.g. Savage, 2002; Köhler, 2008; Uetz, 2013), including what may be the largest sized

New World colubrid snake, Pseustes sulphureus (Pérez-Santos & Moreno, 1988).

To date, the taxonomy and systematics within this genus have been based on morphological descriptions; However, polymorphisms in this clade render taxonomic decisions based on this kind of information questionable (see figures in Porras & Solórzano, 2006). Therefore, evolutionary relationships and species boundaries within Pseustes remain poorly understood.

Species of Pseustes have been classified as belonging to numerous other genera, including

Ahaetulla, Chironius, Coluber, Dipsas, Herpetodryas, Natrix, Phrynonax, Spilotes, Synchalinus,

Thamnobius, and Tropidodipsas (Uetz, 2013), as well as several subfamilies within the

Colubridae.

75 We utilize samples collected from multiple putative species across Central and South

America in order to conduct the first molecular phylogenetic analysis of this genus to (i) infer the phylogenetic position of Pseustes within the family Colubridae, (ii) determine whether species of the genus Pseustes form a monophyletic group, (iii) infer phylogenetic relationships within

Pseustes, and (iv) assess species-level diversity to resolve historical taxonomic debates. We provide initial results and inferences on the evolutionary relationships within Pseustes, thus providing a basis for further work on this understudied group of Neotropical snakes.

Methods

Molecular sampling

Genomic DNA was isolated from muscle tissue of 17 specimens of Pseustes and Spilotes using a

Qiagen DNeasy extraction kit and protocol. Two mitochondrial (NADH dehydrogenase subunit

4 (ND4) and cytochrome b (cyt b)) and one nuclear oocyte maturation factor Mos (c-mos) gene fragments were independently PCR amplified using GoTaq® Green master mix by Promega,

Madison, WI, USA. Protocols for amplification were carried out as described in Arévalo et al.

(1994), Burbrink et al. (2000), de Queiroz et al. (2002), and Lawson et al. (2005) incorporating the primer pairs ND4 + LEU, L14910 + H16064, and S77 + S78 and annealing temperatures

48ºC, 46ºC, and 55ºC, respectively. Sequencing was performed in both forward and reverse directions using the PCR primers on a Beckman Coulter automated capillary sequencer, and sequence chromatographs were edited using Sequencher 4.2, Gene Codes Corporation, Ann

Arbor, MI, USA. Sequences for each gene fragment were aligned separately, first automatically using the program MUSCLE (Edgar, 2004), and then manually rechecked using Se-Al v2.0a11

(Rambaut, 2002). No internal stop codons were found in these protein-coding gene fragments.

76 Previously published sequences of snakes within the family Colubridae were downloaded from

GenBank (see Table 4.1) and were combined with new sequence data generated in this study

(Table 4.2).

Table 4.1. GenBank numbers for DNA sequences analyzed in this study not including sequences listed in Table 4.2. Species ND4 cyt b c-mos Ahaetulla fronticincta — — — AF471072 AF471161 Arizona elegans DQ902279 DQ902101 DQ902058 Bogertophis rosaliae DQ902280 DQ902102 DQ902059 Bogertophis subocularis DQ902281 DQ902103 DQ902060 Boiga dendrophila U49303 AF471089 AF471128 Cemophora coccinea DQ902282 AF471091 AF471132 Chilomeniscus stramineus U49305 GQ895856 GQ895800 Chionactis occipitalis — — — GQ895857 GQ895801 Chironius carinatus — — — HQ529280 HQ529281 Chrysopelea paradisi — — — GQ895858 GQ895802 Coelognathus flavolineatus U49301 DQ902128 DQ902090 Coelognathus helena DQ902292 DQ902112 DQ902071 Coelognathus radiata DQ902317 DQ902121 DQ902079 Coluber constrictor AY487041 AY486914 AY486938 Coluber dorri AY487042 AY188040 AY188001 Coluber zebrinus AY487058 AY188043 AY188004 Conopsis biserialis — — — GQ895860 GQ895804 Conopsis nasus — — — GQ895861 GQ895805 Coronella austriaca AY487065 AY486930 AY486954 Coronella girondica AY487066 AF471088 AF471113 Crotaphopeltis tornieri — — — AF471093 AF471112 Dasypeltis atra — — — AF471065 AF471136 Dendrelaphis caudolineatus — — — GQ895864 GQ895808 Dendrophidion dendrophis — — — GQ895865 GQ895809 Dinodon rufozonatum — — — AF471063 AF471163 Dipsadoboa unicolor — — — AF471062 AF471139 Dispholidus typus U49302 AY188012 AY187973 Dolichophis caspius AY487039 AY376739 AY376797 Dolichophis jugularis AY487046 AY486917 AY486941 Drymarchon corais DQ902314 AF471064 AF471137 Drymobius rhombifer — — — GQ927320 GQ927313 Drymoluber dichrous — — — GQ895869 GQ895812 Eirenis eiselti AY487069 AY376747 AY376805

77 Eirenis levantinus AY487071 AY376765 AY376823 Eirenis modestus AY487072 AY486933 AY486957 Eirenis punctatolineatus AY487073 AY376755 AY376813 Elaphe bimaculata DQ902283 DQ902104 DQ902062 Elaphe climacophora DQ902285 DQ902105 DQ902064 Elaphe quadrivirgata DQ902300 DQ902120 DQ902078 Elaphe quatuorlineata AY487067 AY486931 AY486955 Elaphe rufodorsata DQ902301 DQ902123 DQ902081 Euprepiophis conspicillata DQ902286 DQ902106 DQ902065 Euprepiophis mandarina DQ902294 DQ902115 DQ902073 Gonyosoma oxycephalum DQ902309 AF471084 AF471105 Grayia tholloni DQ486326 DQ486351 DQ486175 Hemerophis socotrae AY487055 AY188042 AY188003 Hemorrhois hippocrepis AY487045 DQ451987 AY486940 Hemorrhois nummifer AY487049 AY376742 AY376800 Hierophis spinalis AY487056 AY486924 AY486948 Hierophis viridiflavus AY487057 AY486925 AY486949 Lampropeltis alterna AY497307 AF337130 FJ627799 Lampropeltis mexicana AY497310 AF337146 FJ627800 Leptophis ahaetulla — — — GQ927321 GQ927316 Lycodon zawi — — — AF471040 AF471111 Lytorhynchus diadema — — — AY188025 AY187986 Maculophis bella DQ902316 DQ902134 DQ902097 Masticophis flagellum AY487060 AY486928 AY234228 Mastigodryas boddaerti — — — GQ895867 GQ895811 Mastigodryas melanolomus — — — GQ895868 — — — Oligodon cinereus — — — AF471033 AF471101 Opheodrys aestivus — — — AF471057 AF471147 Oreocryptophis porphyracea DQ902298 DQ902118 DQ902076 Orthriophis hodgsoni DQ902318 DQ902136 DQ902096 Orthriophis moellendorffi DQ902295 DQ902116 DQ902074 Oxybelis aeneus — — — AF471056 AF471148 Pantherophis guttatus DQ902291 DQ902111 DQ902070 Philothamnus heterodermus — — — AF471055 AF471149 Phyllorhynchus decurtatus — — — AF471083 AF471098 Pituophis catenifer AF138764 AF337112 FJ627790 Pituophis deppei AF141096 FJ627818 FJ627801 Platyceps florulentus AY487043 AY486915 AY486939 Platyceps najadum AY487038 AY486912 AY486936 Platyceps rhodorachis AY487051 AY486921 AY486945 Pseudelaphe flavirufa DQ902289 DQ902109 DQ902068

78 Pseudocyclophis persicus — — — AY376757 AY376815 Pseudoficimia frontalis — — — GQ895886 GQ895827 Pseudorabdion oxycephalum — — — AF471073 DQ112083 Pseustes sulphureus Ptyas korros AY487062 AY486929 AY486953 Rhadinophis frenatum DQ902290 DQ902110 DQ902069 Rhadinophis prasina DQ902299 DQ902119 DQ902077 Rhinechis scalaris AY487068 AY486932 AY486956 Salvadora mexicana AY487075 AY486934 AY486958 Scaphiodontophis annulatus — — — GQ927323 GQ927318 Senticolis triaspis AF138775 DQ902127 DQ902086 Sonora semiannulata — — — AF471048 AF471164 Spalerosophis diadema AY487059 AF471049 AF471155 Spilotes pullatus — — — AF471041 AF471110 Stenorrhina freminvillei — — — GQ895889 GQ895830 Storeria dekayi EF417365 AF471050 AF471154 Sympholis lippiens — — — GQ895890 GQ895831 Tantilla relicta — — — AF471045 AF471107 Telescopus fallax — — — AF471043 AY188000 Thelotornis capensis — — — AF471042 AF471109 Thrasops jacksonii — — — AF471044 DQ112084 Trimorphodon biscutatus DQ497506 GQ927324 GQ927319 Zamenis hohenackeri DQ902320 DQ902137 DQ902098 Zamenis lineata DQ902319 AJ277674 DQ902099

Our preliminary analyses based on taxa throughout the family Colubridae found Pseustes to have a close affinity to members the subfamily Colubrinae and we therefore made our final analyses presented here incorporating taxa in that subfamily. Grayia tholloni, Pseudorabdion oxycephalum, Scaphiodontophis annulatus, and Storeria dekayi were used as outgroup taxa to root our Colubrinae phylogenetic tree of 95 species.

79 Table 4.2. Genbank numbers for DNA sequences generated in this study. Abbrevations of institutions and individuals for voucher specimens are as follows: EBRG (Museo de Biologia de la Estacion Biologica Rancho Grande, Maracay, Venezuela), JMR (Julie M. Ray field series), LSUMZ (Louisiana State Museum of Natural Science), USNM (Smithsonian Institution, National Museum of Natural History), UTA (Amphibian and Reptile Diversity Research Center, University of Texas, Arlington). Species Voucher Locality ND4 cyt b c-mos Pseustes USNM 564157 Gracias a Dios, poecilonotus Tapalwás, Honduras (14°51'N, 84°32'W) Pseustes LSUMZ H-14673 Río San Juan, poecilonotus Nicaragua: Ca. 15 km S. El Castillo on north bank Rio San Juan at Isla el Diamante. Pseustes JMR 725 Near La Mica poecilonotus Biological Station, El Copé, Coclé, Panama (8°37′12″N, 80°36′0″W) Pseustes JMR 744 Parque Nacional G. poecilonotus D. Omar Torrijos Herrera, Coclé, Panama (8° 40’N, 80° 37’W) Pseustes LSUMZ H-17739 polylepis Rondônia, Brazil Pseustes UTA R-55965 Morona-Santiago, polylepis Ecuador: Road to Mendez (2.65590°S; 78.20707°W) Pseustes LSUMZ 42718 polylepis “Suriname” Pseustes sp. UTA R-46140 Alta Verapaz, Guatemala: Cobán, Parque Nacional. Laguna Lachuá. ca. 175 MSNM Pseustes sp. LSUMZ 36746 “Honduras” Pseustes sp. LSUMZ 39592 North coast of Honduras

80 Pseustes sp. LSUMZ H-7806 North coast of Honduras Spilotes sp. LSUMZ H-14026 Amazonas, Brazil: Rio Ituxi at the Madeirera Scheffer (8º 20’ 47”N, 65º 42’ 57.9”W) Spilotes UTA R-52006 Petén, Guatemala: pullatus La Libertad, Parque Nacional Sierra Lacandón, Distrito Guayacán Spilotes LSUMZ 36738 pullatus “Honduras” Spilotes LSUMZ H-14023 sulphureus Amazonas, Brazil Spilotes LSUMZ 43274 Pasco Department, sulphureus Peru; 41km Villa Rica Puerto. Bermudez Hwy, 750 m elev. Spilotes LSUMZ 42645 sulphureus “Suriname” Spilotes EBRG 5107 Bolivar, Venezuela: sulphureus 25 km W Santa Elena de Uairen via El Paun

Phylogenetic analyses

We conducted mixed-model analyses on a concatenated dataset (2340 total bp), partitioned by gene as well as by codon, resulting in a total of nine partitions. We used Akaike information criterion (AIC) to identify the best-fit models of nucleotide substitution for both Bayesian inference (BI) and maximum likelihood (ML) analyses. For this we implemented the program

MrModeltest v2.2 (Nylander, 2004), run in PAUP* v4.0b10 (Swofford, 2002), which recovered the GTR + I + Γ model for all three codon positions of ND4 and cyt b while finding models

HKY + Γ for the first and second position of c-mos and GTR + Γ for the third position.

81 We inferred phylogenetic relationships using Bayesian Inference criterion implementing

MrBayes v3.0b4 (Ronquist & Huelsenbeck, 2003). Two simultaneous runs were conducted (with the default Markov chain Monte Carlo [MCMC] settings), for a total of 10.0 × 106 generations per run, sampling trees and parameters every 100 generations. We used potential scale reduction factor values (output by MrBayes), together with plots of cold chain likelihood values and parameter estimates visualized in Tracer v1.5.4 (Rambaut & Drummond, 2009) to confirm stationarity and convergence of MCMC runs. Based on this evaluation, the first 2.5 × 106 generations from each run were discarded as burn-in.

Using the same partitioning scheme described above, we inferred the ML tree using RAxML

7.2.8 and assessed tree support with the rapid-bootstrapping algorithm using 1000 non- parametric bootstraps (Stamatakis, 2006; Stamatakis et al., 2008). Additionally, we performed the SHL test (Anisimova & Gascuel, 2006; Shimodaira & Hasegawa, 1999) in RAxML 7.2.8 to provide another maximum likelihood measure of support. Support for the SHL test is measured as 1 - P, where P is equivalent to the probability of obtaining a particular test statistic under the null hypothesis that the maximum likelihood estimate of the branch is not significantly more likely than any nearest-neighbor rearrangements of that branch. All ML estimates and tests were run under the GTRCAT model.

Results

Both phylogenetic analyses recovered the genus Pseustes nested within a clade of 23 other New

World Colubrinae genera (Fig. 4.1), and paraphyletic with respect to the closely related genus

Spilotes (Fig. 4.2). Unfortunately, our phylogenetic analyses do not resolve the sister clade to

82 Pseustes + Spilotes; multiple other lineages of New World Colubrinae snakes form a basal polytomy at this node.

Within the Pseustes + Spilotes clade, we identify six divergent lineages, three within each of two larger clades (Fig. 4.2). Our results strongly support Pseustes sulphureus being more closely related to Spilotes pullatus lineages than other species of Pseustes. Pseustes poecilonotus as currently recognized is composed of two distinct, geographically separated lineages based on current sampling, one of which is found in Central America and the other in South America.

Sister to this P. poecilonotus clade is a lineage that occurs in northern Central America and may be sympatric with Honduran populations of P. poecilonotus. Finally, our findings show a deep split between Spilotes pullatus from Brazil and other S. pullatus populations in Central America, potentially being indicative of cryptic diversity.

83    3VHXVWHV    6SLORWHV      

       

     

 

      

 

   

         

     

     

Figure 4.1. Phylogenetic estimate of relationships among genera and species within the Colubrinae. The tree was estimated from a Bayesian 50% majority-rule consensus composed from a concatenated multigene dataset (ND4, cyt b, and c-mos; total of 2340 bp). Numbers at nodes represent values of Bayesian posterior probabilities (PP, above) and Maximum Likelihood Bootstraps and SHL tests (BS/SHL, below). Nodal support values ≥ 95% PP and ≥ 70 BS/SHL are illustrated and considered highly supported.

84        

    

          

    

          

   

    

   

Figure 4.1. Continued…

Discussion

Phylogeny, species boundaries, and taxonomic recommendations within Pseustes

We demonstrate that the genus Pseustes is paraphyletic with respect to Spilotes. In particular, the species Pseustes sulphureus is the sister taxon of Spilotes pullatus. Given that the genus Spilotes

(Linneaus, 1758) is older than Pseustes (Fitzinger, 1843), we recommend changing Pseustes sulphureus to Spilotes sulphureus. Although not strongly supported, S. pullatus and S. sulphureus were also found to form a clade in parsimony analyses by Hollis (2006), who analyzed the two species as part of their outgroup taxa for a morphological phylogenetic study of

Chironius. Additionally, this relationship is supported by unique defensive behavioral characteristics that include displaying their neck posture and inflating the gular region. These two species also possess the most well-developed tracheal left lung with respect to the SVL

85 among colubrid snakes (Rossman & Williams, 1966). These traits distinguish both species from the “puffing” behavior described for P. poecilonotus (as P. p. shropshirei), which is more similar to typical colubrid snakes that inflate their neck as a defensive behavior (Rand & Ortleb,

1969). Finally, Spilotes pullatus and S. sulphureus are notable for being among the largest known New World colubrid snakes (along with Clelia clelia and Drymarchon corais), being capable of reaching lengths upwards of three meters.

Figure 4.2. Locality map showing where tissue samples of Pseustes and Spilotes used in this study were collected throughout Central and South America. Symbols represent estimations of localities examined for molecular sampling. Pseustes poecilonotus from Nicaragua (upper insert). Phylogenetic estimate of relationships within the Pseustes and Spilotes clade, resulting from the Bayesian 50% majority-rule consensus phylogram from figure 1 (lower insert). ** = 100 posterior probability and 95–100 bootstrap/SHL support, respectively.

In this study, we include samples of Pseustes poecilonotus from several localities throughout

Central and South America. Günther (1858) described P. poecilonotus from Honduras and

Mexico but Boulenger (1894) later restricted this species to Honduras. Peters (1867) described P. polylepis (as Ahaetulla polylepis) from Suriname and currently these taxa are considered two of

86 four subspecies of P. poecilonotus. Our sampling included specimens from both Honduras and

Suriname and we find that these two lineages are separated geographically somewhere between western Panama and northern South America. Therefore, the name P. poecilonotus must be applied only to the Mesoamerican populations while South American populations should be treated as the separate lineage P. polylepis (Amazon and the Guayanas).

Taxonomic recommendations for the genus Pseustes

Because Spilotes sulphureus is considered the type species of Pseustes (Fitzinger, 1843) it could be recommended that the remaining species in the genus Pseustes be allocated to Phrynonax

Cope 1862 as a replacement. According to the International Code of Zoological Nomenclature, article 23.9.9.1. on the principle of priority, when the usage of a senior synonyn, in this case

Phrynonax, has been used after 1899 (Amaral (1929) was the last author who used Phrynonax to refer to species of Pseustes), then the genus should be returned to its senior name. However, the junior synonym or homonym Pseustes has been used as a valid name in at least 25 works, published by at least 10 authors in the immediately preceding 50 years, and encompassed a span of not less than 10 years (article 23.9.1.2). We therefore recommend the continued use of the genus Pseustes. The type species of the genus Phrynonax, is Tropidodipsas lunulata Cope 1860, which is a junior synomyn of P. poecilonotus. As such, we recommend that P. poecilonotus become the new type species for the genus Pseustes.

Recovering cryptic diversity within Pseustes and Spilotes

With these results we are beginning to understand the diversity and evolutionary history of this remarkable group of snakes. Our study reveals that the genus Pseustes can formally be placed

87 inside a New World colubrinae clade, along with 23 other genera, and is paraphyletic with respect to Spilotes (Fig. 4.1). Additionally, our data and analyses provide a greater understanding of the species composition and relationships within this clade (Fig. 4.2). In particular, our analyses detect two putative species-level lineages that will likely need to be resurrected from previously described species or elevated from subspecies status based on literature and type material. The first lineage is a Pseustes that occurs in both Honduras and Guatemala, which may be broadly sympatric with P. poecilonotus throughout much of its range (Fig. 4.2, labeled

Pseustes sp.). The second putative new lineage we detected suggests that populations of Spilotes pullatus from Central America are divergent from populations in South America, which may correspond to subspecies that are occasionally recognized for this taxon (Fig. 4.2, labeled

Spilotes sp.). At this time, however, this inference is based on only one divergent sample found in Brazil. A rigorous morphological assessment along with an increased sampling of molecular data for lineages within these genera are necessary for a more complete systematic assessment of diversity that ultimately should include estimations of biogeographic patterns with respect to timing of origin and diversification.

88 CHAPTER 5

CONCLUSIONS

Introduction

The guidelines for integrative taxonomy proposed by Dayrat (2005) state that taxonomic revisions should: examine numerous specimens deemed appropriate by the taxonomists of that particular group; address the infra- and inter-specific character variation; include the totality of the names available for the group; decide appropriate taxonomic changes only when supported by broad biological evidence (morphology, genealogical concordance, ecology, behavior, etc.); and deposit type specimens in a museum collection that are preserved in a way that allows for further molecular study. This dissertation has incorporated all of these guidelines as part of an integrative approach of conducting systematic revisions that should ultimately produce a well- supported and stable classification for several lineages of New World snakes. 1) In chapters two and three I examined the majority of specimens known for both Cerrophidion barbouri and C. godmani, including the type series. 2) The incorporation of these numerous specimens allowed me to examine the majority of morphological characters commonly used for pitviper systematics in order to assess the infra- and inter-specific variation of Cerrophidion taxa being revised. 3) In chapters two, three, and four I included an extensive literature review discussing synonyms and previous works on the systematics of these groups over hundreds of years before taking the final step of revising taxonomy. 4) In chapter three, I deposited type specimens and tissue samples into museums that allowed for future molecular work on all type material. Additionally, I expand on this guideline by depositing type specimens in multiple museums whenever possible. I believe this is particularly important since many type specimens of European museums across all taxa became lost during and after World War II; therefore, depositing type material across museums

89 should now be the normal procedure whenever possible. 5) Finally, chapters two, three, and four included broad biological evidence, particularly molecular phylogenetics of multiple gene fragements and optimality criteria, in order to lend unequivocal support for the detected evolutionary relationships, species diversity, and need for taxonomic revision.

By applying these guidelines to my research, I have ventured beyond classical species descriptions, which all too often distinguish between taxa using only a few characters of a few specimens, into the realm of integrative taxonomy. Using these quidelines allowed me to quantify morphological variation within and among species and provide broad biological evidence to strongly support my classification schemes built around my delimitation of species and genera. Finally, my revisions include geographic and natural history components to update the range maps of these taxa and provide information on their diet and habits. This work puts several New World snake species and genera into a phylogenetic framework that should prove valuable for future studies. In particular, this dissertation improves our understanding of New

World snake diversity by increasing our knowledge of pitvipers by three species and an additional genus endemic to the Mexican highlands as well as uncovering cryptic diversity within some of the largest non-venomous New World snakes and stabilizing the genera Pseustes and Spilotes.

Future directions

Along with morphology, the four mitochondrial genes used in this work have assisted in strongly supporting phylogenetic relationships within most of the New World pitviper genera. For example, there is strong support for relationships within the Porthidium group (genera

Atropoides, Cerrophidion, and Porthidium; Castoe et al., 2005; Jadin et al., 2010), the

90 rattlesnakes (genera Crotalus and Sistrurus; Castoe & Parkinson, 2006), the South American lanceheads (genera Bothriopsis, Bothrocophias, Bothropoides, Bothrops, and Rhinocerophis;

Fenwick et al., 2009), and the Mexican highland endemic pitvipers (genera Mixcoatlus and

Ophryacus; Jadin et al., 2011). Nevertheless, morphological and molecular datasets do not agree or show strong support for many of the relationships among these clades, and singular genera such as Agkistrodon (Cantils, Copperheads, and Cottonmouths), Bothriechis (Palm-pitvipers), or

Lachesis (Bushmasters). This lack of complete phylogenetic understanding about how the genera are related constitutes a large knowledge gap that hinders medical treatment of snakebites, as venom composition of snakes often follows the diversification of that snake lineage (Fry et al.,

2003). Additionally, knowing the phylogenetic position of the bushmasters (Lachesis), the largest and only New World pitvipers that lay eggs, is important for reproductive biology and testing of Dollo’s law to understand how they re-obtained this trait (Fenwick et al., 2012).

One possible method to resolve these evolutionary questions among pitvipers is through the use of entire mitochondrial genomes in phylogenetic analyses. Mitochondrial genomes often provide a significant amount of data that help resolve deep phylogenetic questions compared to other markers. For example, Mueller et al. (2004) was able to strongly support many evolutionary relationships among plethodontid salamanders by conducting phylogenetic analyses of 24 complete mitochondrial genomes. This work resolved many phylogenetic questions within a group that was troubled with extensive homoplastic characters (e.g., Wake, 1966; Chippindale et al., 2004; Wiens et al., 2005). Additionally, Mueller (2006) was able to answer questions about rates of molecular evolution within the mtDNA genomes of these taxa once this phylogeny was better understood. Most recently, mitochondrial genomes, along with codon and region partitioning schemes, have been initiated to resolve avian phylogenetic inquiries beyond that of

91 the traditional gene fragment analyses (Powell et al., 2013). Based on this knowledge, I am currently collecting sequence data of the entire mitochondrial genome for each major clade of pitvipers. I plan to use these mitogenomes to conduct a phylogenetic analysis of New World pitviper clades, which will encompass each genus. Conducting this phylogenetic analysis should contribute to resolving these fundamental questions about relationships of the major New World pitviper clades and the evolution of their mitochondria.

Integrative taxonomy in modern biology

In 1989, E.O. Wilson wrote that there is coming a “pluralization” of biology in which the expert naturalists and taxon-oriented scientist will regain ground previously lost to scientists focusing only on studying levels of organization (e.g., cellular biology). Under this paradigm systematists who have both an expertise in organismal biology as well as modern biological techniques (e.g., molecular biology, bioinformatics) should lead the way in scientific investigations. I believe this future has arrived with integrative taxonomy and that systematists, who become taxonomic experts with their respective organisms and incorporate multiple lines of evidence to resolve phylogenetic relationships, will uncover evolutionary patterns and processes of morphological and ecological diversification, answer questions of historical biogeography, and develop a strong and stable taxonomy. However, Wilson states in the following passage that “Because of the largely unknown nature of diversity, systematics remains a fountainhead of discoveries and new ideas in biology. If a biologist is well trained in the classification of the organisms encountered, the known facts of natural history are an open book, and new phenomena come more quickly into focus. The irony of the situation is that successful research then gets labeled as ecology, physiology, or almost anything else but its true source, the study of diversity” (Wilson,

92 1985:1227). In the current paradigm of academia it is often considered paramount to define one self under the title of a broad biological discipline such as “population geneticist” or “disease ecologist” and not a more narrowly focused taxon-specific discipline like “herpetologist”.

However, defining my research areas has never suited me since my interests span many aspects of biology and I view systematics as a broad field that allows me to work on any aspect of ecology and evolutionary biology that I find fascinating, regardless of taxa. I share these points of view with Wilson and whether I am investigating evolutionary relationships of pitvipers, documenting the range expansion of Hemidactylus geckos, or conducting experiments to assess the influence of invertebrate predators on parasite transmission in larval amphibians, I am an integrative systematist and naturalist studying biological diversity in many forms using a variety of tools for investigation. More than two decades later I am not sure if this “pluralization” has occurred across biology but I hope this dissertation plays a small role in this ambitious goal.

93 REFERENCES

Amaral, A. D. (1929). Estudo sobre ophidios neotropicos XX. Revisao do genero Phrynonax Cope, 1862. Memórias do Instituto Butantan, 4, 301–320.

Amaral, A. D. (1944). Notas sôbre a ofiologia neotropica e brasílica: sôbre a aplicacão do nome genérico Trimeresurus, em vez de Bothrops, a serpentes neotrópicas. Papéis Avulsos do Departamento de Zoologia, São Paulo, 5, 13–18.

Andersson, L. G. (1899). Catalogue of museum type species of snakes in the royal museum in Stockholm.

Anisimova, M. & Gascuel, O. (2006). Approximate likelihood ratio test for branches: a fast, accurate and powerful alternative. Systematic Biology, 55, 539–552.

Arévalo, E. S., Davies, S. K. & Sites, J. W. Jr. (1994). Mitochondrial DNA sequence divergence and phylogenetic relationships among eight chromosome races of the Sceloporus grammicus complex (Phrynosomatidae) in central Mexico. Systematic Biology, 43, 387–418.

Barbour, T. & Loveridge, A. (1929). On some Honduranian and Guatemalan snakes with the description of a new arboreal pit viper of the genus Bothrops. Bulletin of the Antivenin Institute of America, 3, 1–3.

Bauer, A. M., Parham, J. F., Brown, R. M., Stuart, B. L., Grismer, L., Papenfuss, T. J., Böhme, W., Savage, J. M., Carranza, S., Grismer, J. L., Wagner, P., Schmitz, A., Ananjeva, N. B. & Inger, R. F. (2010). Availability of new Bayesian-delimited gecko names and the importance of character-based species descriptions. Proceedings of the Royal Society Series B, 278, 490– 492.

Benabib, M., Kjer, K. M. & Sites, J. W. Jr. (1997). Mitochondrial DNA sequence-based phylogeny and the evolution of viviparity in the Sceloporus scalaris group Reptilia, Squamata). Evolution, 51, 1262–1275.

Benton, M. J. & Ayala, F. J. (2003). Dating the tree of life. Science, 300, 1698–1700.

Bickford, D., Lohman, D. J., Sodhi, N. S., Ng, P. K. L., Meier, R., Winker, K., Ingram, K. K. & Das, I. (2007). Cryptic species as a window on diversity and conservation. Trends in Ecology and Evolution, 22, 148–155.

Bocourt, M. F. (1868). Descriptions de quelques crotaliens nouveaux appartenant au genre Bothrops, recuellis dans le Guatemala. Annales des Sciences Naturelles—Zoologie et Biologie Animale, 5th ser., 10, 201–202.

Bolaños, R. (1972). Toxicity of Costa Rican snake venoms for the white mouse. American Journal of Tropical Medicine and Hygiene, 21, 360–363.

94 Bolaños, R. (1982). Las serpientes venenosas de Centroamerica y el problema del ofidismo. Primera Parte. Aspectos zoologicos, epidemiologicos y biomedicos. Revista Costarricense Ciencia Medica, 3, 165–184.

Bolaños, R. (1984). Serpientes Venenosos y Ofidismo en Centroamerica. Editorial Universidad de Costa Rica, San Jose, Costa Rica. 136pp.

Boulenger, G. A. (1893). Catalogue of the Snakes in the British Museum (Natural History). Volume I. British Mus. (Nat. Hist.), London.

Boulenger, G. A. (1894). Catalogue of the Snakes in the British Museum (Natural History). Volume II. Containing the Conclusion of the Colubridae Aglyphae. British Mus. (Nat. Hist.), London, xi, 382 pp.

Boulenger, G. A. (1896). Catalogue of the Snakes in the British Museum (Natural History). Vol III. London: Taylor and Francis. 727pp.

Burbrink, F. T., Lawson, R. & Slowinksi, J. B. (2000). MtDNA phylogeography of the North American rat snake (Elaphe obsoleta): a critique of the subspecies concept. Evolution, 54, 2107–2118.

Burger, W. L. (1971). Genera of Pitvipers. Ph.D. Dissertation. University of Kansas. Lawrence, Kansas.

Campbell, J. A. (1976). A new terrestrial pit viper of the genus Bothrops (Reptilia, Serpentes, Crotalidae) from western Mexico. Journal of Herpetology, 10, 151–160.

Campbell, J. A. (1977). The distribution, variation, and natural history of the Middle American highland pitvipers, Bothrops barbouri and Bothrops godmani. Master’s thesis, University of Texas at Arlington.

Campbell, J. A. (1985). A new species of highland pitviper of the genus Bothrops from southern Mexico. Journal of Herpetology, 19, 48–54.

Campbell, J. A. (1988). The distribution, variation, natural history, and relationships of Porthidium barbouri. Acta Zoologica Mexicana, 26, 1–32.

Campbell, J. A. (1999). Distribution patterns of amphibians in Middle America. In W. E. Duellman (Ed) Patterns of Distribution of Amphibians: a global perspective (pp. 111–210). Baltimore, MD: Johns Hopkins University Press.

Campbell, J. A. & Flores-Villela, O. (2008). A new long-tailed rattlesnake (Viperidae) from Guerrero, Mexico. Herpetologica, 64, 246–257.

Campbell, J. A. & Lamar, W. W. (1989). The Venomous Reptiles of Latin America. Ithaca, NY: Cornell University Press.

95

Campbell, J. A. & Lamar, W. W. (1992). Taxonomic status of miscellaneous Neotropical viperids with the description of a new genus. Occasional Papers of Texas Technological University, 153, 1–31.

Campbell, J. A. & Lamar, W. W. (2004). Venomous Reptiles of the Western Hemisphere. 2 Vols. Ithaca, NY: Cornell University Press.

Campbell, J. A. & Solórzano, A. (1992). The distribution, variation, and natural history of the Middle American montane pitviper, Porthidium godmani. In J. A. Campbell & E. D. Brodie, Jr. (Eds) Biology of the Pitvipers (pp. 223–250). Tyler, TX: Selva.

Campbell, J. A., Smith, E. N., Streicher, J., Acevedo, M. E. & Brodie Jr., E. D. (2010). New salamanders (Caudata: Plethodontidae) from Guatemala, with miscellaneous notes on known species. Miscellaneous Publications of the Museum of Zoology, University of Michigan, 200, i–v + 1–60.

Castoe, T. A., Chippindale, P. T., Campbell, J. A., Ammerman, L. K. & Parkinson, C. L. (2003). Molecular systematics of the Middle American jumping pitvipers (genus Atropoides) and phylogeography of the A. nummifer complex. Herpetologica, 59, 420–431.

Castoe, T. A., Sasa, M. M. & Parkinson, C. L. (2005). Modeling nuclear evolution at the mesoscale: the phylogeny of the Neotropical pitvipers of the Porthidium group (Viperidae: Crotalinae). Molecular Phylogenetics and Evolution, 37, 881–898.

Castoe, T. A. & Parkinson, C. L. (2006). Bayesian mixed models and the phylogeny of pitvipers (Viperidae: Serpentes). Molecular Phylogenetics and Evolution, 39, 91–110.

Castoe, T. A., Daza, J. M., Smith, E. N., Sasa, M. M., Kuch, U., Campbell, J. A., Chippindale, P. T. & Parkinson, C. L. (2009). Comparative phylogeography of pitvipers suggests a consensus of ancient Middle American highland biogeography. Journal of Biogeography, 36, 88–103.

Chippindale, P. T., Bonett, R. M., Baldwin, A. S. & Wiens, J. J. (2004). Phylogenetic evidence for a major reversal in life history evolution in plethodontid salamanders. Evolution, 58, 2809–2822.

Cope, E. D. (1860). Descriptions of new species of the reptilian genera Hyperolius, Liuperus, and Tropidodipsas. Proceedings of the Academy of Natural Sciences of Philadelphia, 1860, 517–518.

Cope, E. D. (1862). Catalogues of the reptiles obtained during the exploration of the Parana, Paraguay, Vermejo and Uruguay Rivers, by Capt. Thos. J. Page, U.S.N.; and of those procured by Lieut. N. Michler, U.S. Top. Eng., commander of the expedition conducting the survey of the Atrato River. Proceedings of the Academy of Natural Sciences of Philadelphia, 14, 346–359.

96 Cope, E. D. (1871). Ninth contribution to the herpetology of tropical America. Proceedings of the Academy of Natural Sciences of Philadelphia, 23, 200–224.

Cope, E. D. (1887). Catalogue of Batrachians reptiles of Central America and Mexico. Bulletin of the United States National Museum, 32, 1–89.

Cracraft, J. & Prum, R.O. (1988). Patterns and processes of diversification: speciation and historical congruence in some neotropical birds. Evolution, 42, 603–620.

Cuesta-Terrón C. (1921). Los crotalinos Mexicanos. Memorias de la Sociedad Científica Antonio Alzate, 39, 173–194.

Daly, M., Herendeen, P. S., Guralnick. R. P., Westneat, M. W. & McDade, L. (2012). Systematics agenda 2020: the mission evolves. Systematic Biology, 61, 549–552.

Darwin, C. (1859). On the Origin of Species by Means of Natural Selection, or the Preservation of Favoured Races in the Struggle for Life. 1st edn. John Murray, London.

Davis, W. B. & Dixon J. R. (1959). Snakes of the Chilpancingo region, Mexico. Proceedings of the Biological Society of Washington, 72, 79–92.

Dayrat, B. (2005). Towards integrative taxonomy. Biological Journal of the Linnean Society, 85, 407–415.

Daza, J. M., Smith, E. N., Páez, V. P. & Parkinson, C. L. (2009). Complex evolution in the Neotropics: The origin and diversification of the widespread genus Leptodeira (Serpentes: Colubridae). Molecular Phylogenetcs and Evolution, 53, 653–667.

Daza, J. M., Castoe, T. A. & Parkinson, C. L. (2010). Using regional comparative phylogeographic data from snake lineages to infer historical processes in Middle America. Ecography, 33, 343–354. de Queiroz, A., Lawson, R. & Lemos-Espinal, J. A. (2002). Phylogenetic relationships of North American garter snakes (Thamnophis) based on four mitochondrial genes: how much DNA is enough? Molecular Phylogenetcs and Evolution, 22, 315–329.

Dowling, H. G. (1951). A proposed standard system of counting ventrals in snakes. British Journal of Herpetology, 1, 97–99.

Dowling, H. G. & Savage, J. M. (1960). A guide to the snake hemipenis: a survey of basic structure and systematic characteristics. Zoologica, 45, 17–30.

Duellman, W. E. (1999). Patterns of the distribution of amphibians: a global perspective. Johns Hopkins University Press, Baltimore, Maryland.

97 Dugès, A. A. D. (1896). Reptiles y batracios de los Estados Unidos Mexicanos. La Naturaleza, 2, 479–485.

Duméril, A. M. C. & Bibron, G. (1835). Erpétologie Générale ou Histoire Naturelle Complète des Reptiles. Tome second. Librairie Encyclopédique de Roret, Paris.

Dunn, E. R. (1919). Two new crotaline snakes from western Mexico. Proceedings of the Biological Society of Washington, 32, 213–216.

Durban, J., Juárez, P., Angulo, Y., Lomonte, B., Flores-Diaz, M., Alape-Girón, A., Sasa, M., Sanz, L., Gutiérrez, J. M., Dopazo, J., Conesa, A. & Calvete, J. J. (2011). Profiling the venom gland transcriptomes of Costa Rican snakes by 454 pyrosequencing. BMC Genomics, 12, 259.

Edgar, R. C. (2004). MUSCLE: multiple sequence alignment with high accuracy and high throughput. Nucleic Acids Research, 32, 1792–1797.

Fenwick, A. M., Gutberlet, R. L. Jr., Evans, J. A. & Parkinson, C. L. (2009). Morphological and molecular evidence for phylogeny and classification of South American pitvipers, genera Bothrops, Bothriopsis, and Bothrocophias (Serpentes: Viperidae). Zoological Journal of the Linnean Society, 156, 617–640.

Fenwick, A. M., Greene, H. W. & Parkinson, C. L. (2012). The serpent and the egg: unidirectional evolution of reproductive mode in vipers? Journal of Zoological Systematics and Evolutionary Research, 50, 59–66.

Felsenstein J. 1985. Confidence limits on phylogenies: an approach using the bootstrap. Evolution 39: 783–791.

Fischer, J. G. (1880). Neue Amphibien und Reptilien. Archiv für Naturgeschichte, 46, 215–227.

Fischer, J. G. (1883). Beschreibung neuer Reptilien—Jahresbericht über das Naturhistorische Museum zu Hamburg für das Jahr 1882. Aus dem Osterprogramm des Akademischen Gymnasiums, 8, 1–16.

Fitzinger, L. (1843). Systema Reptilium, fasciculus primus, Amblyglossae. Braumüller et Seidel, Wien, 106 pp.

Flores-Villela, O. & Canseco-Márquez, L. (2004). Nuevas especies y cambios taxanómicos para la herpetofauna de México. Acta Zoológica Mexicana, nueva serie, 20, 115–144.

Flores-Villela, O., Kjer, K. M., Benabib, M. & Sites, J. W. Jr. (2000). Multiple data sets, congruence, and hypothesis testing for the phylogeny of basal groups of the lizard genus Sceloporus (Squamata, Phrynosomatidae). Systematic Biology, 49, 713–739.

98 Fry, B. G., Wüster, W., Kini, R. M., Brusic, V., Khan, A., Venkataraman, D. & Rooney, A. P. (2003). Molecular Evolution and Phylogeny of Elapid Snake Venom Three-Finger Toxins. Journal of Molecular Evolution, 57, 110–129.

Grismer, L. L., Grismer, J. L. & McGuire, J. A. (2006). A new species of pitviper of the genus Popeia (Squamata: Viperidae) from Pulau Tioman, Pahang, West Malaysia. Zootaxa, 1305, 1–19.

Gutberlet, R. L. Jr. (1993). Descriptive morphology and phylogenetic position of the Mexican black-tailed pitviper. Master’s thesis, University of Texas at Arlington.

Gutberlet, R. L. Jr. (1998). Phylogenetic position of the Mexican black-tailed pitviper (Squamata: Viperidae: Crotalinae). Herpetologica, 54, 184–206.

Gutberlet, R. L. Jr. & Harvey, M. B. (2004). The evolution of New World venomous snakes. In J. A. Campbell & W. W. Lamar (Eds) Venomous Reptiles of the Western Hemisphere, 2 Vols, (pp. 634–682). Ithaca, NY: Comstock Publishing Associates.

Gutiérrez, J. M. & Chavez, F. (1980). Efectos proteolítico, hemorrágico, y mionecrótico de los venenos de serpientes costarricenses de los géneros Bothrops, Crotalus, y Lachesis. Toxicon, 18, 315–321.

Günther, A. (1858). Catalogue of Colubrine snakes of the British Museum. Trustees of the British Museum, London, XVI, 1–281.

Günther, A. C. L. G. (1895–1902). Biologia Centrali-Americana. Reptilia and Batrachia. London, U.K.: Porter.

Günther, A. C. L. G. (1863). Third account of the snakes in the collection of the British Museum. Annals and Magazine of Natural History, 12, 348–365.

Hoge, A. R. (1965). Preliminary account on Neotropical Crotalinae (Serpentes, Viperidae). Memórias do Instituto Butantan, 32, 109–184.

Hollis, J. L. (2006). Phylogenetics of the genus Chironius Fitzinger, 1826 (Serpentes, Colubridae) based on morphology. Herpetologica, 62, 435–453.

Jadin, R. C. (2007). Prey items of the Tzotzil montane pitviper (Cerrophidion tzotzilorum). Southwestern Naturalist, 52, 437–438.

Jadin, R. C. (2010). Morphological phylogeny of the Montane pitvipers (Crotalinae: Cerrophidion), with comments on the conservation of Mesoamerican highlands. In L. D. Wilson, J. H. Townsend & J. D. Johnson (Eds) Conservation of Mesoamerican Amphibians and Reptiles (pp. 674–685). Eagle Mountain, UT: Eagle Mountain Publishing.

99 Jadin, R. C., Gutberlet, R. L. Jr. & Smith, E. N. (2010). Phylogeny, evolutionary morphology, and hemipenis descriptions of the Middle American jumping pitvipers (Serpentes: Crotalinae: Atropoides). Journal of Zoological Systematics and Evolutionary Research, 48, 360–365.

Jadin, R. C. & Smith, E. N. (2010). The Hemipenis of Trimorphodon quadruplex. Herpetological Review, 41, 157–159.

Jadin, R. C., Smith, E. N. & Campbell, J. A. (2011). Unraveling a tangle of Mexican serpents: a systematic revision of highland pitvipers. Zoological Journal of the Linnean Society, 163, 943–958.

Joppa, L. N., Roberts, D. L. & Pimm, S. L. (2011). The population ecology and social behaviour of taxonomists. Trends in Ecology and Evolution, 26, 551–553.

Keogh, J. S. (1999). Evolutionary implications of hemipenial morphology in the terrestrial Australian elapid snakes. Zoological Journal of the Linnean Society, 125, 239–278.

Kjer, K. M., Swigonova, Z., LaPolla, J. S. & Broughton, R. E. (2007). Why weight? Molecular Phylogenetics and Evolution, 43, 999–1004.

Klauber, L. M. (1972). Rattlesnakes: their habits, life histories, and influence on mankind. 2nd ed. 2 vol. Berkeley, CA: University of California Press.

Knight, A. & Mindell, D. P. (1993). Substitution bias, weighting of DNA sequence evolution, and the phylogenetic position of Fea’s viper. Systematic Biology, 42, 18–31.

Köhler, G., Veselý, M. & Greenbaum, E. (2006). The Amphibians and Reptiles of El Salvador. Malabar, FL: Krieger Publishing, 238 pp.

Köhler, G. (2008). Reptiles of Central America, 2nd edition. Herpeton.

Laurenti, J. N. (1768). Specimen medicum, exhibens synopsin reptilium emendatam cum experimentis circa venena et antidota reptilium austracorum, quod authoritate et consensu. Joan. Thomae, Vindobonae [Vienna].

Lawson, R., Slowinski, J.B., Crother, B.I., Burbrink, F.T., 2005. Phylogeny of the Colubroidea (Serpentes): New evidence from mitochondrial and nuclear genes. Molecular Phylogenetics and Evolution, 37, 581–601.

Leaché, A. D. & Fujita, M. K. (2010). Bayesian species delimitation in West African forest geckos (Hemidactylus fasciatus). Proceedings of the Royal Society Series B, 277, 3071– 3077.

Leenders, T. A. A. M. & Watkins-Colwell, G. J. (2004). Notes on a collection of amphibians and reptiles from El Salvador. Postilla, 231, 1–31.

100

Leviton, A. E., Gibbs, J. R. H., Heal, E. & Dawson, C. E. (1985). Standards in herpetology and ichthyology. Part I. Standard symbolic codes for institutional resource collections in herpetology and ichthyology. Copeia, 1985, 802–832.

Linnaeus, C. (1758). Systema naturæ per regna tria naturæ, secundum classes, ordines, genera, species, cum characteribus, differentiis, synonymis, locis. Tomus I. Editio decima, reformata. [10th ed.]. Laurentii Salvii, Holmiae [Stockholm].

Lomonte, B., Rey-Suárez, P., Tsaia, W.-C., Angulo, Y., Sasa, M., Gutiérrez, J. M. & Calvete, J. J. (2012). Snake venomics of the pit vipers Porthidium nasutum, Porthidium ophryomegas, and Cerrophidion godmani from Costa Rica: Toxicological and taxonomical insights. Journal of Proteomics, 75, 1675–1689.

López-Luna, M. A., Vogt, R. C. & de la Torree-Loranca, M. A. (1999). A new species of montane pitviper from Veracruz, Mexico. Herpetologica, 55, 382–389.

Losos, J. B., Arnold, S. J., Bejerano, G., Brodie, III, E. D., Hibbett, D., Hoekstra, H. E., Mindell, D. P., Monteiro, A., Moritz, C., Orr, H. A., Petrov, D. A.. Renner, S. S., Ricklefs, R. E., Soltis, P. S. & Turner, T. L. (2013). Evolutionary biology for the 21st century. PLoS Biology, 11, e1001466.

Maddison, D. R., Schulz K-S. & Maddision, W. P. (2007). The tree of life project. Zootaxa, 1668, 19–40.

Martín del Campo, R. (1935). Nota acerca de la distribución geográfica de los reptiles ponzoñosos en México. Anales del Instituto de Biología, Universidad Nacional Autónoma México, 6, 291–300.

McCranie, J. R. (2011). The Snakes of Honduras: Systematics, Distribution, and Conservation. Ithaca, NY: Society for the Study of Amphibians and Reptiles, x + 714 pp.

McCranie, J. R. & Wilson, L. D. (2002). The Amphibians of Honduras. Ithaca, NY: Society for the Study of Amphibians and Reptiles, x + 625 pp.

Mertens, R. (1952). Die amphibien und reptilien von El Salvador. Abhandlungen der Senckenbergischen Naturforschenden Gesellschaft, 487, 1–120.

Mickel, C. E. (1930). Description plus types vs. descriptions alone. Psyche, 37, 118–131.

Mocquard, M. F. (1908–1909). Etudes sur les reptiles. Mission scientifique au Mexique et dans l’Amérique central—Recherches zoologiques. Part 3. Impérimerie Impriale, Paris.

Mueller, R. L. (2006). Evolutionary rates, divergence dates, and the performance of mitochondrial genes in Bayesian phylogenetic analysis. Systematic Biology, 55, 289–300.

101 Mueller, R. L., Macey, J. R., Jaekel, M., Wake, D. B. & Boore, J. L. (2004). Morphological homoplasy, life history evolution, and historical biogeography of plethodontid salamanders inferred from complete mitochondrial genomes. Proceedings of the National Academy of Sciences, 101, 13820–13825.

Müller, F. (1877). Mittheilungen aus der herpetologischen. Sammlung des Basler Museums, 39 pp.

Müller, F. (1878). Katalog der im Museum und Universitätskabinet zu Baselaufgestellen Amphibien und Reptilien nebst Anmerkungen. Verhandlungen der Naturforschenden Gesellschaft in Basel, 6, 559–709.

Myers, C. W. & Cadle, J. E. (2003). On the snake hemipenis, with notes on Psomophis and techniques of eversion: a response to Dowling. Herpetological Review, 34, 295–302.

Nelson, G. J. (1971). ‘Cladism’ as a philosophy of classification. Systematic Zoology, 20, 373– 376.

Nicholas, K. B. & Nicholas, H. B. (1997). GeneDoc: a tool for editing and annotating multiple sequence alignments. www.cris.com/~Ketchup/genedoc.shtml.

Nylander, J. A. A. (2004). MrModeltest v2.2. Program distributed by the author. Evolutionary Biology Centre, Uppsala University.

Parkinson, C. L., Moody, S. M. & Ahlquist, J. E. (1997). Phylogenetic relationships of the “Agkistrodon complex” based on mitochondrial DNA sequence data. In R. S. Thorpe, W. Wüster & A. Malhotra (Eds) Snakes: Ecology, Evolution, and Snakebite (pp. 63–78). Oxford, UK: The Zoological Society of London, Clarendon Press.

Parkinson, C. L., Campbell, J. A. & Chippindale, P. T. (2002). Multigene phylogenetic analyses of pitvipers, with comments on their biogeography. In G. W. Schuett, M. Höggren, M. E. Douglas & H. W. Greene (Eds) Biology of the vipers (pp. 93–110). Eagle Mountain, UT: Eagle Mountain Publishing.

Pearson, D. L., Hamilton, A. L. & Erwin, T. L. (2011). Recovery plan for the endangered taxonomy profession. BioScience, 61, 58–63.

Perez-Higareda, G., Smith, H. M. & Julia-Zertuche, J. (1985). A new jumping pitviper, Porthidium olmec, from Veracruz, Mexico (Serpentes: Viperidae). Bulletin of the Maryland Herpetological Society, 21, 97–106.

Pérez-Santos, C. & Moreno, A. (1988). Ofidios de Colombia. Museo Regionale di Scienze Naturali Torino, Monographie VI, 517 pp.

Peters, W. C. H. (1867). Über Flederthiere (Pteropus gouldii, Rhinolophus deckenii, Vespertilio lobipes, Vesperugo temminckii) und Amphibien (Hypsilurus godeffroyi, Lygosoma scutatum,

102 Stenostoma narirostre, Onychocephalus unguirostris, Ahaetulla poylepis, Pseudechis scutella. Monatsberichte der Königlich preussischen Akademie der Wissenschaften zu Berlin, 1867, 703–712.

Picado, T. C. (1931). Serpientes Venenosas de Costa Rica, sus Venenos, Seroterápia Anti- ofidica. Imprenta Alsina, San José, Costa Rica. 222pp.

Picado, T. C. (1936). [dated 1933–1934]. Serpientes venenosas occorêntes en Costa Rica. Memorial Instututo Butantan, 8, 389–397.

Porras, L. W. & Solórzano, A. (2006). A diamond in the rough: The Northern Birdsnake Pseustes poecilonotus. Reptilia, 46, 24–29.

Posada, D. & Crandall, K. A. (1998). Model-Test: testing the model of DNA substitution. Bioinformatics, 14, 817–818.

Powell, A. F. L. A., Barker, F. K. & Lanyon, S. M. (2013). Empirical evaluation of partitioning schemes for phylogenetic analyses of mitogenomic data: An avian case study. Molecular Phylogenetics and Evolution, 66, 69–79.

Pyron, R. A., Burbrink, F. T., Colli, G. R., Montes de Oca, A. N., Vitt, L. J., Kuczynski, C. A. & Wiens, J. J. (2011). The phylogeny of advanced snakes (Colubroidea), with discovery of a new subfamily and comparison of support methods for likelihood trees. Molecular Phylogenetics Evolution, 58, 329–342.

Rambaut, A. (2002). Se-Al: Sequence alignment editor, Version 2.0a11. Available at http://tree.bio.ed.ac.uk/software/seal/

Rambaut A. & Drummond, A. J. (2009). Tracer v1.5. Available at http://tree.bio.ed.ac.uk/software/tracer/

Rand, A. S. & Ortleb, E. P. (1969). Defensive display in the colubrid snake Pseustes poecilonotus shropshirei. Herpetologica, 25, 46–48.

Roca, A. L., Georgiadis, N., Pecon-Slattery, J. & O'Brien, S. J. (2001). Genetic evidence for two species of elephant in Africa. Science, 293, 1473–1477.

Ronquist, F. & Huelsenbeck, J. P. (2003). MrBayes 3: Bayesian phylogenetic inference under mixed models. Bioinformatics, 19, 1572–1574.

Rosindell, J. & Harmon L. J. (2012). OneZoom: A fractal explorer for the tree of life. PLoS Biology, 10, e1001406.

Rossman, D. A. & Williams, K. L. (1966). Defensive behavior of the South American colubrid snakes Pseustes sulphureus (Wagler) and Spilotes pullatus (Linneaus). The Proceedings of the Louisiana Academy of Sciences, 29, 152–156.

103

Sasa, M. (1997). Cerrophidion godmani in Costa Rica: a case of extremely low allozyme variation? Journal of Herpetology, 31, 569–572.

Savage, J. M. (2002). Amphibians and Reptiles of Costa Rica. Ithaca, NY: Cornell University Press.

Schargel, W. E., Rivas, G. A., Makowsky, R., Señaris, J. C., Natera, M. A., Barros, T. R., Molina, C. R. & Barrio-Amorós, C. L. (2010). Phylogenetic systematics of the genus Gonatodes (Squamata: Sphaerodactylidae) in the Guayana region, with description of a new species from Venezuela. Systematics and Biodiversity, 8, 321–339.

Schlegel, H. (1837). Essai sur la physionomie des serpens. Partie Générale + Partie Descriptive. J. Kips, J.H.Z. & W.P. van Stockum, La Haye.

Scott, N. J. (1969). A zoographic analysis of the snakes of Costa Rica. Ph.D. Dissertation. University of Southern California. 390 pp. University Microfilms, Ann Arbor, Michigan (Dissertation Abstracts 69:17891).

Shimodaira, H. & Hasegawa, M. (1999). Multiple comparisons of Log-Likelihoods with applications to phylogenetic inference. Molecular Biology and Evolution, 16, 1114–1116.

Shreve, B. (1938). A new Agkistrodon from Mexico. Copeia, 1938, 9.

Smith, E. N. & Ferrari-Castro, J. A. (2008). A new species of jumping pitviper of the genus Atropoides (Serpentes: Viperidae: Crotalinae) from the Sierra de Botaderos and the Sierra La Muralla, Honduras. Zootaxa, 1948, 57–68.

Smith, H. M. (1941). Notes on Mexican snakes of the genus Trimeresurus. Zoologica, 26, 61–64.

Smith, M. A., Rodriguez, J. J., Whitfield, J. B., Deans, A. R., Janzen, D. H., Hallwachs, W. & Hebert, P. D. N. (2008). Extreme diversity of tropical parasitoid wasps exposed by iterative integration of natural history, DNA barcoding, morphology, and collections. Proceedings of the National Academy of Sciences, 105, 12359–12364.

Solórzano, A. (2004). Serpientes de Costa Rica: Distribución, taxonomía e historia natural/Snakes of Costa Rica: Distribution, Taxonomy, and Natural History. 791 pp. Instituto Nacional de Biodiversidad (INBio).

Stamatakis, A. (2006). RAxML-VI-HPC: maximum likelihood-based phylogenetic analyses with thousands of taxa and mixed models. Bioinformatics, 22, 2688–2690.

Stamatakis, A., Hoover, P. & Rougemont, J. (2008). A rapid bootstrap algorithm for the RAxML web-servers. Systematic Biology, 75, 758–771.

104 Sunyer, J. & Köhler, G. (2007). New country and departmental records of herpetofauna in Nicaragua. Salamandra, 43, 57–62.

Swofford, D. L. (2002). PAUP*: Phylogenetic analyses using parsimony (*and other methods), version 4.0b10. Sunderland, MA: Sinauer Associates.

Tamura, K., Dudley, J., Nei, M. & Kumar, S. (2007). MEGA4: Molecular Evolutionary Genetics Analysis (MEGA) software version 4.0. Molecular Biology and Evolution, 24, 1596–1599.

Tamura, K., Peterson, D., Peterson, N., Stecher, G., Nei, M. & Kumar, S. (2011). MEGA5: Molecular Evolutionary Genetics Analysis using Maximum Likelihood, Evolutionary Distance, and Maximum Parsimony Methods. Molecular Biology and Evolution, 28, 2731– 2739.

Townsend, J. H. & Wilson, L. D. (2008). Guide to the Amphibians and Reptiles of Cusuco National Park, Honduras/Guía de los Anfibios y Reptiles de Parque Nacional Cusuco, Honduras. Salt Lake City, UT: Bibliomania!. xiv + 322 pp.

Townsend, J. H. & Wilson, L. D. (2010). Conservation of the Honduran herpetofauna: issues and imperative. In Wilson, L. D., J. H. Townsend, & J. D. Johnson, (Eds) Conservation of Mesoamerican Amphibians and Reptiles (pp. 460–487). Eagle Mountain, Utah: Eagle Mountain Publications LC.

Townsend, J. H., Medina-Flores, M., Murillo, J. L., & Austin, J. D. (2011). Cryptic diversity in Chortís Highland moss salamanders (Caudata: Plethodontidae: Nototriton) revealed using mtDNA barcodes and phylogenetics, with a new species from eastern Honduras. Systematics and Biodiversity, 9, 275–287.

Townsend, J. H., Wilson, L. D., Medina-Flores, M., Aguilar-Urbina, E., Atkinson, B. K., Cerrato-Mendoza, C. A., Contreras-Castro, A., Gray, L. N., Herrera-B, L. A., Luque-Montes, I. R., McKewy-Mejía, M., Portillo-Avilez, A., Stubbs, A. L., & Austin, J. D. (2012). A premontane hotspot for herpetological endemism on the windward side of Refugio de Vida Silvestre Texíguat, Honduras. Salamandra, 48, 92–114.

Uetz, P. (2010). The original descriptions of reptiles. Zootaxa, 2334, 59–68.

Uetz, P. (2013). The reptile database. accessed online on 25 March 2013.

Villa, J. (1962). Las Serpientes Venenosas de Nicaragua. Editorial Novedades, Managua, Nicaragua. 93pp.

Villa, J. (1984). The venomous snakes of Nicaragua: a synopsis. Milwaukee Public Museum Contributions in Biology and Geology, 59, 1–41.

105 Vogel, G., David, P. & Pauwels, O. S. G. (2004). A review of morphological variation in Trimeresurus popeiorum (Serpentes: Viperidae: Crotalinae), with the description of two new species. Zootaxa, 727, 1–63.

Wake, D. B. (1966). Comparative osteology and evolution of the lungless salamanders, family Plethodontidae. Memoir of Southern California Academy of Sciences, 4, 1–111.

Werman, S. (1992). Phylogenetic relationships of Central and South American pitvipers of the genus Bothrops (sensu lato): cladistic analyses of biochemical and anatomical characters. In J. A. Campbell & E. D. Brodie, Jr. (Eds) Biology of the Pitvipers (pp. 21–40). Tyler, Texas: Selva.

Wiens, J. J., Bonett, R. M. & Chippindale, P. T. (2005). Ontogeny discombobulates phylogeny: paedomorphosis and higher-level salamander phylogeny. Systematic Biology, 54, 91–110.

Wilson, E. O. (1985). Time to revive systematics. Science, 230, 1227.

Wilson, E. O. (1989). The coming pluralization of biology and the stewardship of systematics. BioScience, 39, 242–245.

Wilson, L. D. & Johnson, J. D. (2010). Distributional patterns of the herpetofauna of Mesoamerica, a biodiversity hotspot. In L. D. Wilson, J. H. Townsend, & J. D. Johnson (Eds) Conservation of Mesoamerican Amphibians and Reptiles (pp. 30–235). Eagle Mountain, Utah: Eagle Mountain Publishing LC.

Wilson, L. D. & Meyer, J. R. (1982). The Snakes of Honduras. Milwaukee, WI: Milwaukee Public Museum, 159 p.

Wilson, L. D. & Meyer, J. R. (1985). The Snakes of Honduras. 2nd Edition. Milwaukee, WI: Milwaukee Public Museum, x + 150 pp.

Wilson, L. D. & McCranie, J. R. (2004). The herpetofauna of the cloud forests of Honduras. Amphibian and Reptile Conservation, 3, 34–48.

Wilson, L. D., Townsend, J. H. & Johnson, J. D. (eds.). (2010). Conservation of Mesoamerican Amphibians and Reptiles. Eagle Mountain, UT: Eagle Mountain Publishing LC. xviii + 812 pp.

Wüster W., Thorpe, R., Puorto G. & BBBSP. (1996). Systematics of the Bothrops atrox complex (Reptilia: Serpentes: Viperidae) in Brazil: a multivariate analysis. Herpetologica, 52, 263– 271.

Zaher, H. & Prudente, A. L. C. (2003). Hemipenes of Siphlophis (Serpentes, Xenodontinae) and techniques of hemipenial preparation in snakes: a response to Dowling. Herpetological Review, 34, 302–307.

106