Phylogeography of the Desert Horned Lizard (Phrynosoma Platyrhinos) and the Short-Horned Lizard (Phrynosoma Douglassi): Patterns of Divergence and Diversity

UNLV Retrospective Theses & Dissertations

1-1-1994

Phylogeography of the desert horned lizard (Phrynosoma platyrhinos) and the short-horned lizard (Phrynosoma douglassi): Patterns of divergence and diversity

Kenneth Bruce Jones University of Nevada, Las Vegas

Follow this and additional works at: https://digitalscholarship.unlv.edu/rtds

Repository Citation Jones, Kenneth Bruce, "Phylogeography of the desert horned lizard (Phrynosoma platyrhinos) and the short-horned lizard (Phrynosoma douglassi): Patterns of divergence and diversity" (1994). UNLV Retrospective Theses & Dissertations. 2987. http://dx.doi.org/10.25669/hcai-vpf6

This Dissertation is protected by copyright and/or related rights. It has been brought to you by Digital Scholarship@UNLV with permission from the rights-holder(s). You are free to use this Dissertation in any way that is permitted by the copyright and related rights legislation that applies to your use. For other uses you need to obtain permission from the rights-holder(s) directly, unless additional rights are indicated by a Creative Commons license in the record and/or on the work itself.

This Dissertation has been accepted for inclusion in UNLV Retrospective Theses & Dissertations by an authorized administrator of Digital Scholarship@UNLV. For more information, please contact [email protected]. INFORMATION TO USERS

This manuscript has been reproduced from the microfilm master. UMI films the text directly from the original or copy submitted. Thus, some thesis and dissertation copies are in typewriter face, while others may be from any type of computer printer.

The quality of this reproduction is dependent upon the qualify of the copy submitted. Broken or indistinct print, colored or poor quality illustrations and photographs, print bleedthrough, substandard margins, and improper alignment can adversely affect reproduction.

In the unlikely event that the author did not send UMI a complete manuscript and there are missing pages, these will be noted. Also, if unauthorized copyright material had to be removed, a note will indicate the deletion.

Oversize materials (e.g., maps, drawings, charts) are reproduced by sectioning the original, beginning at the upper left-hand comer and continuing from left to right in equal sections with small overlaps. Each original is also photographed in one exposure and is included in reduced form at the back of the boot

Photographs included in the original manuscript have been reproduced xerographically in this copy. Higher quality 6" x 9" black and white photographic prints are available for any photographs or illustrations appearing in this copy for an additional charge. Contact UMI directly to order.

A Bell & Howell Information Company 300 North Z eeb Road. Ann Arbor. Ml 48106-1346 USA 313/761-4700 800/521-0600

PHYLOGEOGRAPHY OF THE DESERT HORNED LIZARD

(PHRYNOSOMA PLATYRHINOS1 AND THE SHORT-HORNED

LIZARD (PHRYNOSOMA DOUGLASSD: PATTERNS

OF DIVERGENCE AND DIVERSITY

Kenneth Bruce Jones

A thesis submitted in partial fulfillment of the requirements for the degree of

Doctor of Philosophy

Environmental Biology

Department of Biological Studies University of Nevada, Las Vegas December 1995 UMI Number: 9614365

This microform edition is protected against unauthorized copying under Title 17, United States Code.

Chairperson, Brett R. Riddle, Ph.D.

Examining Committee Member, Daniel B. Thompson, Ph.D.

Examining Committee Member, Stanley D. Smith, Ph.D.

Examining Committee Member, James H. Brown, Ph.D.

Graduate Faculty Representative, Frederick Bachhuber, Ph.D.

Interim Dean, Graduate College, Cheryl L. Bowles, Ed.D ABSTRACT

Historical environmental change is thought to have played an important role in the diversification of the biota of western North America. Many patterns of diversification have been associated with glacial-interglacial cycles of the latest Pleistocene. Recent evidence on diversification patterns in small mammals suggests a response to older environmental change, especially the uplift of the western North American cordillera.

To evaluate the relative influence of old and recent historical environmental change on lineage diversification, mitochondrial DNA molecular phylogeographies of

Phrvnosoma douelassi (the short-homed lizard) and Phrvnosoma platvrhinos (the desert homed lizard) were analyzed. Both species are widespread and have relatively old histories in western North America.

P. douelassi and R platvrhinos demonstrated multiple scales of lineage diversification representating responses to relatively old and recent historical environmental change. Deep levels of divergence in P. douelassi were structured geographically among Kuchler Physiographic regions and were temporally concordant with the late Tertiary uplift of the western North American cordillera. Significant phylogenetic gaps between regions suggest long-term barriers to gene flow among regions. P. platvrhinos demonstrated a deep divergence between the southern Sonoran

Desert and the remainder of the species’ range. With the assumption of 2% divergence per million years, this divergence roughly corresponds to the Bouse Embayment, a large

lake incursion that divided areas east and west of the Colorado River during the Pliocene.

Assuming a roughly consistent rate of molecular evolution, both species exhibit

divergence depths within regions indicative of response to Pleistocene glacial-interglacial

cycles.

The depth of lineage diversity in P. douelassi and P. platvrhinos was concordant

with a model of habitat change during the last glacial maximum. Regions that maintained

large patches of suitable habitat during the last glacial maximum had deep levels of

divergence whereas regions that lost large areas of suitable habitat had shallow levels of

within-region divergence.

The phylogeographic patterns of P. douelassi and P. platvrhinos. and results of

simulated habitat change, do not support a model of mass habitat shift to the south in response to glaciation. Rather, results of this study suggest that suitable habitats and populations persisted in many areas within the current ranges of both species ranges during the last glacial maximum, although the distribution of P. platvrhinos was more limited. TABLE OF CONTENTS

ABSTRACT...... iii

TABLE OF CONTENTS...... v

LIST OF FIGURES...... vii

LIST OF TABLES...... ix

ACKNOWLEDGEMENTS ...... x

CHAPTER 1 INTRODUCTION...... 1 Statement of the Problem ...... 1 Molecular Phylogeography ...... 5 The Western United States as a Study Area ...... 9 Woodland and Desert Taxa as Indicators of Concordance between Environmental Change and Lineage Divergence ...... 11 Phrvnosoma platvrhinos and Phrvnosoma douelassi as Indicators of Lineage Response to Historical Environmental Change ...... 13 Hypotheses ...... 15

CHAPTER 2 PHYLOGEOGRAPHY OF PHRYNOSOMA DOUGLASSI...... 18 Introduction ...... 18 Methods and Materials ...... 22 DNA Extraction ...... 24 PCR Amplification ...... 25 RFLP Digestion of Amplified mtDNA Fragments ...... 27 Sequencing Procedure ...... 28 Data Analysis ...... 29 Restriction-Site Data Analysis ...... 29 Sequence Data Analysis ...... 32 R esults...... 36 Distribution and Ecology ...... 36 Phylogenetic Relationships and Phylogeography - Restriction-site Analysis ...... 39 Discussion ...... 59 Phylogeographic Patterns ...... 63 Estimating Times of Divergence and Concordance with Environmental Change ...... 68 Distribution and Ecology ...... 69 Phylogenetic Relationships ...... 71 Summary and Conclusions ...... 72

v CHAPTER 3 PHYLOGEOGRAPHY OF PHRYNOSOMA PLATYRHINOS...... 74 Introduction ...... 74 Methods and Materials ...... 77 DNA Extraction ...... 77 PCR Amplification ...... 79 Restriction Enzyme Digestion of Amplified mtDNA Fragments ...... 79 Data Analysis ...... 79 R esults...... 81 Distribution and Ecology ...... 81 Phylogenetic Relationships and Phylogeography ...... 83 Discussion ...... 93 Phylogeographic Relationships ...... 93 Distribution and Ecology ...... 98 Summary and Conclusions ...... 100

CHAPTER 4 DISTRIBUTION OF PHRYNOSOMA DOUGLASSI AND PHRYNOSOMA PLATYRHINOS HABITAT DURING THE LAST GLACIAL MAXIMUM AND PRESENT...... 101 Introduction ...... 101 Methods and Materials...... 105 R esults...... 109 Habitat Suitability of Phrvnosoma douelassi ...... 109 Habitat Suitability of Phrvnosoma platvrhinos ...... 112 Discussion ...... 117 Habitat Persistence and Phylogeographic Patterns in P. douelassi .... 117 Habitat Persistence and Phylogeographic Patterns in P. platvrhinos . . . 119 Summary and Conclusions ...... 120

CHAPTER 5 DISSERTATION SUMMARY AND CONCLUDING REMARKS .. 122

LITERATURE CITED ...... 129

APPENDIX I ...... 140

APPENDIX H ...... 144

APPENDIX H I...... 146

APPENDIX IV ...... 148

APPENDIX V ...... 166 LIST OF FIGURES

Figure 1 Locations of Phrvnosoma douelassi sample sites ...... 23 Figure 2 Geographic distribution of Kuchler (1964) Physiographic Regions in the western United States ...... 31 Figure 3 Estimated base-pair saturation in transitions (ta) and transversions (tv) by codon position ...... 35 Figure 4 Phylogenetic relationships of composite P. douelassi haplotypes estimated from a Neighbor-joining analysis of restriction-site data...... 42 Figure 5 Phylogenetic relationships of composite P. douelassi haplotypes estimated from a Maximum Parsimony analysis of restriction-site data...... 46 Figure 6 Geographic distribution of composite restriction-site haplotypes of P. douelassi ...... 48 Figure 7 Phylogenetic relationships of composite Phrvnosoma douelassi haplotypes estimated from a Maximum Likelihood analysis of sequence data ...... 50 Figure 8 Simplified regional phylogeny of Phrvnosoma douelassi based on the Maximum Likelihood analysis ...... 52 Figure 9 Phylogenetic relationships of composite Phrvnosoma douelassi haplotypes estimated from a Maximum Parsimoney analysis of sequence data ...... 56 Figure 10 Geographic distribution of composite sequence haplotypes of Phrvnosoma douelassi...... 58 Figure 11 Habitat preferences of individual species of Phrvnosoma mapped on to the phylogeny of the genus ...... 60 Figure 12 Estimated phylogeny of the genus Phrvnosoma based on morphological and molecular data ...... 61 Figure 13 Geographic distribution of sample sites for Phrvnosoma platvrhinos ...... 78 Figure 14 Phylogenetic relationships of composite P. platvrhinos haplotypes estimated from a Neighbor-joining analysis of restriction-site distance data ...... 87 Figure 15 Phylogenetic relationships of composite P. platvrhinos haplotypes estimated from a Maximum Parsimony analysis based on restriction-site data ...... 89

vii Figure 16 Geographic distribution of composite P. platvrhinos haplotypes ...... 92 Figure 17 Modeled distribution of current R douelassi suitable habitat ...... I ll Figure 18 Modeled distribution of suitable R douelassi habitat 18.000 years ago ...... 113 Figure 19 Modeled distribution of current P. Platvrhinos habitat ...... 115 Figure 20 Modeled distribution of P. platvrhinos suitable habitat 18.000 years ago ...... 116 LIST OF TABLES

Table 1 Hierarchical arrangement of sample sites into populations and regions for an AMOVA analysis of P. douelassi restriction-site distance data ...... 33 Table 2 Hierarchical arrangement of sample sites into populations and regions for an AMOVA analysis of P. douelassi sequence distance data...... 37 Table 3 Summary of P. douelassi mtDNA haplotype restriction-site patterns ...... 40 Table 4 Nucleotide divergence of P. douelassi among regions and nucleotide diversity within regions (diagonal) based on restriction-site analysis ...... 44 Table 5 Restriction-site haplotype diversity of P. douelassi by Kuchler Physiographic Region ...... 47 Table 6 Nucleotide divergence distances among and nucleotide diversity within (diagonal) Kuchler Physiographic regions for P. douelassi ...... 54 Table 7 Sequence haplotype diversity of R douelassi by Kuchler Physiographic Region ...... 57 Table 8 Hierarchical arrangement of sample sites into populations and regions for an AMOVA analysis of P. platvrhinos restriction-site distance data ...... 82 Table 9 Summary of P. platvrhinos mtDNA haplotype restriction- site patterns ...... 84 Table 10 Sequence divergence estimates for P. platvrhinos ...... 86 Table 11 Nucleotide divergence of P. platvrhinos among and nucleotide diversity within (diagonal) regions based on restriction-site analysis ...... 91 Table 12 mtDNA diversity of P. platvrhinos among regions based on haplotypes generated from restriction-site analysis in R E A P ...... 94 Table 13 Habitat models used to estimate current and historic (18,000 years ago) suitable habitat of P. douelassi ...... 107 Table 14 Habitat models used to estimate current and historic (18,000 years ago) suitable habitat of P. platvrhinos ...... 108 Table 15 Estimates of suitable habitat for P. douelassi at present and 18.000 years ago ...... 110 Table 16 Estimates of suitable habitat for P. platvrhinos at present and 18.000 years ago ...... 114 ACKNOWLEDGEMENTS

I am indebted to my Major Professor, Dr. Brett Riddle, for guidance and

suggestions throughout the entire project, and to other members of my committee,

Dr. Stan Smith, Dr. Daniel Thompson, Dr. Fred Bachhuber, and Dr. James Brown, for

thorough and thoughtful comments on research presented in this manuscript. I thank

Dale Tribby (U.S. BLM, Miles City, MT), Terry Rucci (U.S. BLM, Bishop, CA), Jeff

Harrick (U.S. ARS, Las Cruces, NM), Walt Whitford (U.S. EPA, Las Cruces, NM),

Gaylon Holmes (Casper, WY), Cecil Schwabe (Univ of Arizona, Tucson, AZ), Charlie

Painter (NM Game and Fish, Sante Fe, NM), George Sheppard (U.S. Forest Service,

Williams, AZ), Chuck Peterson (Idaho State Univ., Idaho Falls, ID), Mel Wilhelm (U.S.

Forest Service, Springerville, AZ), Judith Lanchaster (DRI, Reno, NV), Ron Marlowe

(UNLV), Bill Kepner (U.S. EPA, Las Vegas), Nancy Olsen (U.S. National Park Service,

Mendora, ND), Steve Thompson (SD Game and Fish, Pierre, SD), David Orange

(UNLV), my sons Justin, Luke , and Mick, and their friends, Kyle Ingerson, Billy

Osgood, and Jon Strong, for collecting or assisting in the collection of specimens. I am

extremely grateful to Kelly Zamudio for conducting analyses of transition to transversion ratios in Phrvnosoma douglassi (Figure 3), providing advice on sequence analysis of P.

douelassi. and providing sequence data on P. douelassi and additional specimens for restriction-site analysis from Arizona, California, Colorado, New Mexico, Texas, Utah,

x and Washington. I am also grateful to Todd Reeder (National Mus. Nat. History,

Washington, D.C.) for providing phylogenetic information on the genus Phrvnosoma. and to David Nickle (NIH, Phoenix, AZ) and David Orange for advice in implementing lab procedures and phylogenetic analyses. I am extremely indebted to Jim Wickham (TVA,

Norris, TN), Kurt Riitters (TVA, Norris, TN), Bob O’Neill (ORNL, Oak Ridge, TN), Bill

Kepner, Deb Chaloud (EPA), and Walt Whitford, who in addition to being close working associates and friends, influenced my thinking about habitat and landscape ecology concepts applied to the research. I am also indebted to several other people at the EPA

Laboratory in Las Vegas, but especially Gareth Pearson, Llewellyn Williams, Bob

Snelling, Bob Schonbrod, Terry Grady, Richard Gamas, and Wayne Marchant, for making this entire project possible. I am grateful to Gail Corbin, Paula Fry, Aggie

Henning, Lori Pearson, and Pam Grossmann of Applied Technology Associates for preparation of the dissertation. I am grateful to my mother, Pattye Rice, my father,

Kenneth Jones, and my stepfather, Joseph Rice, for creating a creative and exciting childhood environment which led to my passion for discovery. Finally, I am indebted to my wife, Barbara, and sons, Justin, Luke, and Mick, for their unwavering support over the entire project. CHAPTER 1

INTRODUCTION

Statement of the Problem

Considerable debate has centered around the degree to which historical environmental change has influenced the composition, structure, and function of ecosystems (Clements 1916, Gleason 1926, MacArthur and Wilson 1967, Graham 1986,

Ricklefs 1987, Graham and Mead 1987, Mayden 1988, O'Neill et al. 1991, DiMichele

1994). Some of this discussion has centered around whether species in ecosystems respond in concert, or in an individualistic manner to environmental change (Vrba 1985,

Graham 1986, Webb 1987, Graham and Mead 1987, Huntley and Webb 1987, Mayden

1988, Hewitt 1993, DiMichele 1994).

Central to this debate is the degree to which historical environmental change influences patterns of diversification in animals and plants (Rosen 1978, Brooks 1985,

Vrba 1985, Cracraft 1988, Riddle 1995). Addressing this question requires an understanding of concordance between the sequence of environmental changes and the sequence of diversification within and across lineages (e.g., branching sequences within and among lineages, Brown 1995, Riddle 1995). However, because of the complexity of historical changes in the environment and biotic responses to those changes, establishing

1 concordance between diversification and historical environmental change has proven

difficult (Brown 1995).

A variety of historical environmental changes are thought to have played a

significant role in shaping divergence patterns within and among taxa, including

mountain building (Riddle 1995), retreat and advancement of inland seas (Lamb et al.

1989), and glacial-interglacial cycles (Betancourt et al. 1990). The potential

macroevolutionary responses to environmental change include vicariance,

migration/colonization, and extinction, whereas microevolutionary responses could

include a combination of natural selection, genetic drift, and gene flow (Endler 1986,

Templeton 1991). A number of models have been invoked to explain changes in lineage

divergence and diversity, including vicariant speciation, peripheral isolation speciation, and shifts in diversity due to changing habitat locations (Brown and Gibson 1983, Brooks and McLennan 1991, Avise 1994). Of these, vicariance models frequently have been used to explain congruence between divergence and historical environmental change

(Brown 1971, Patterson 1980, Cracraft 1983, Patterson and Atmar 1986, Lomolino et al.

1989, Riddle and Honeycutt 1990).

Historical biogeographic and paleoecological studies have produced mixed results in establishing correlations between patterns of diversification and environmental change

(Brooks and McLennan 1991, Cracraft 1989, Avise 1992,1994), even when patterns of species diversity and historical environmental change are well known. The primary problem is that species or taxonomic lineages (and hence the history of divergence) are not well characterized. Issues of scale, in part, may be responsible for lack of concordance between

patterns of environmental change and patterns of diversification (Avise 1994). Many

studies associating environmental change with patterns of diversity have dealt with areas

of endemism, species with limited ranges, or environmental change at fine scales (Brooks

and McLennan 1991, Brown 1995). Reconstruction of the sequence of historical events,

and establishment of concordance between lineage evolution and environmental change

at relatively fine scales is difficult due to the complicated sequences of environmental and

biotic change that occur at these scales.

Over the past 10 years, greater emphasis has been placed on comparing species

and lineages with widespread distributions (Avise et al. 1987, Cracraft 1988, Mayden

1988, Riddle 1995). Evolutionary responses of animals to historical environmental change in western North America have been documented from the fossil record (Graham

1986), and more recently, from phylogenetic studies (Riddle and Honeycutt 1990, Echelle and Dowling 1992, Lamb et al. 1989,1992, Riddle et al. 1993, Riddle 1995).

Phylogenetic studies of mammals have revealed relatively ancient divergence of species in response to mountain building and associated increases in aridity (Riddle 1995), and more recent divergence within and among closely related taxa in response to glacial and interglacial cycles of the Pleistocene (Riddle and Honeycutt 1990, Riddle et al. 1993).

Phylogenetic studies of fish and reptiles have revealed divergence patterns concordant with Pleistocene environmental change (Echelle and Dowling 1992, Lamb et al. 1992).

Moreover, within lineages, there appear to be nested hierarchies of divergence at regional scales (Avise 1989, Riddle 1995). Hierarchy theory (Allen and Starr 1982, O'Neill et al.

1986) predicts that the response of a nested system or entity (e.g., various related taxa within a region) is constrained by conditions at the broader scale (e.g., environmental conditions and regionally-defined lineages).

In many cases, multiple depths of divergence have been thought to be difficult to demonstrate because more recent environmental change events are assumed to have erased evidence of deeper divergence at any given site (Brown 1995). However, when evaluated over a region (as opposed to local area), long-term divergence patterns are often apparent (Riddle 1995).

Finally, most studies of environmental change and lineage diversification have emphasized concordance between relatively recent environmental change (e.g., glacial cycles associated with the mid to late Quaternary) and divergence patterns (Riddle 1995).

Yet in North America some of the most dramatic changes in biome distribution and connectivity occurred between approximately 3 and 28 million years ago, primarily in response to regional scale climate change associated with mountain building and changes in ocean currents (Axelrod 1958,1979,1985, Christiansen and McKee 1978, Lucchitta

1979, Bamoski 1984, Leopold and Denton 1987). Only within the past few years have we begun to understand the influence of deep history (e.g. geologic events) on lineage divergence (Riddle 1995). Further research is needed to determine how many taxa exhibit nested patterns of divergence, whether taxa from the same geographic regions have similar scales of nested hierarchy, and whether branching points within lineages correspond to old as well as relatively recent environmental change. 5

Molecular Phylogeography: A Method of Explaining Relationships Between Lineage and Environmental Change Histories

Application of molecular markers to phylogenetic studies of lineages representing broad geographic areas has increased significantly over the past 10 years (see Avise 1994 for a thorough review). Molecular phylogeography represents a remarkable integration of micro (e.g., DNA sequence analysis) and macro (e.g., continental-scale biogeography, macroecology, and paleoecology) techniques, and provides a framework for understanding hierarchical temporal and spatial relationships within and among lineages

(Avise 1994).

Phylogenetic approaches have their roots in cladistic methods (Hennig 1966), and their primary aim has been to determine evolutionary relationships among a wide range of taxa (Cracraft 1986,1988). The goal of phylogenetic analysis is to produce a tree that best represents evolutionary relationships among taxa (Hillis and Moritz 1990, Avise

1994), given the data at hand. Trees are generated by comparing characteristics

(characters) of taxa through a series of statistical approaches, of which the most commonly employed is parsimony (Cracraft 1989, Brooks and McLennan 1991, Harvey and Pagel 1991, Avise 1994). Under a parsimony framework, characters that provide the most resolution to tree topology are those that are uniquely shared among taxa relative to outgroups, and that don't frequently switch between ancestral and derived states (Avise

1994). By obtaining samples from different geographic areas, it is possible to look at the geographic context of phylogenetic relationships (Avise 1994). Large phylogenetic gaps between two geographically adjacent species would suggest they diverged in the distant past, whereas small gaps would imply a more recent divergence. The mapping of phylogeny onto geography (e.g., area cladograms, Platnick and Nelson 1978, Page 1988) to evaluate the geographic context of lineage divergence is a fundamental objective of phylogeography (Avise et al. 1987,1994).

Over the last 10 years, molecular markers (characters), including those derived from specific regions of nuclear and mitochondrial DNA (mtDNA), have become the primary data of phylogenetic and phylogeographic analyses (Avise 1994). The application of molecular characters in discerning historical relationships within and among taxa representing different geographic areas defines the field of molecular phylogeography (Avise et al. 1987, Avise 1989,1994). The shift toward molecular characters has resulted from significant breakthroughs in molecular techniques (e.g.

Polymerase Chain Reaction or PCR techniques, Avise 1994), reduced costs associated with techniques, and distinctive advantages of molecular approaches (Avise 1994). Of the molecular markers used to evaluate phylogenies in animals, the mtDNA molecule has been assayed the most (Avise 1994), primarily because of its intrinsically high rate of fixation of new mutations and high rate of lineage sorting among isolated populations, which creates a highly sensitive marker. Genotypes from mtDNA represent nonrecombining characters that are transmitted predominantly through material lines

(Avise 1994). For simplicity, these genotypes are referred to as haplotypes.

Some of the advantages of molecular characters include (Avise 1994): (1) an ability to select different gene regions to address different scales of divergence due to tremendous variation in rates of evolution across different coding and non-coding regions of DNA, (2) an ability to apply the data to questions of phylogenetics and population genetics (e.g., the history of population attributes, including effective population size and 7

gene flow), (3) an ability to compare phylogeographic patterns across taxa and hence

examine degrees of concordance in patterns of divergence, and (4) the ability to

determine the degree to which environmental change is concordant with diversification

patterns in one to several lineages across multiple geographic scales. Conversely,

morphological characters have some distinctive disadvantages in reconstructing historical

relationships within and among lineages (Avise 1994). First, variation in morphological

characters may result from the influence of the environment and not as a result of

evolutionary history. Second, selection of morphological characters can vary across

different lineages, making cross-lineage comparisons difficult.

Molecular markers are not without their problems. Molecular characters do not

guarantee the correct phylogenetic tree because of problems associated with shared

ancestral polymorphisms and multiple base pair substitutions at nucleotide sites. These

problems can be mitigated to a certain degree through understanding of evolutionary

processes that affect mutation rates and molecular sequences (Simon et al. 1994).

In the molecular phylogenetic approach, sequence divergence between populations is used to reconstruct historical relationships among populations.

Populations with smaller differences in sequence divergence are thought to be more closely related than populations with greater sequence divergence. This relationship assumes that sequence divergence is a function of mutation rate and that the mutation rate is relatively constant across the taxa being examined within the particular gene region being assayed.

Unequal rates of sequence divergence among lineages can result from large differences in effective population sizes, and when a particular gene region is linked to a trait (behavioral, physiological, morphological) that has different levels of fitness in

different environments. In small populations, genetic drift results in a relatively fast rate

of fixation and loss of mutations, whereas in large populations fixation or loss of mutations occurs more slowly. Therefore, lineages comprized of small populations are likely to diverge more rapidly than lineages consisting of large populations. However, over relatively long time periods (millions of years), effective population sizes of widely distributed species are likely to be roughly equal.

When a large portion of a gene region is linked to a trait that has differential fitness in different environments, natural selection could result in different rates of sequence divergence among lineages. One solution to this problem is to select gene regions that are relatively neutral with regard to natural selection (e.g., non-coding regions). Empirical data on mutation rates of different gene regions in animal mtDNA suggest that non-coding regions mutate at greater rates than coding regions (Avise 1994).

Mutation rates of coding regions are lower because they are tied more directly to functional aspects of the organisms (Avise 1994) which in turn are more likely to affect the fitness of the organism. Additionally, mutation rates have been shown to vary for a particular gene region across major taxomic groups (Martin et al. 1992). However, mutation rates of particular gene regions are less likley to vary among closely related taxa

(Simon et al. 1994).

One of the greatest challenges in relating patterns of molecular phylogeography to historical environmental change is calibration of the approximate dates of environmental change to branching events in the phylogeny (Riddle 1995). The assumption is that 9

concordance between lineage branching and environmental change at one to several

scales indicates a response of the lineage to environmental change.

Calibration is difficult because of a paucity in fossil records, and because

estimates of mutation rates used to construct molecular clocks can vary tremendously

across organisms and molecular markers (Martin et al. 1992, Avise 1994). However, it

may be possible to adjust clock estimates through increased knowledge of the particular

molecular markers being assayed, and knowledge of the organism. For example, it

appears that mutation rates of molecular markers in endotherms are 3-8 times faster than

those of ecotherms (Martin et al. 1992). Additionally, calibration can be improved by

incorporation of additional fossil records (Avise 1994). Finally, the type and scale of

questions being addressed may require only coarse calibration (Riddle 1995). For

example, if a hypothesis addresses whether splits in a phylogeny are associated with

events of the late Pleistocene (i.e., last 100 thousand years), or pre-Pleistocene (before the

last 2 million years), then one only needs to place divergence events into chronological

order representing approximate time frames (see Riddle 1995).

The Western United States as a Study Area

The western United States offers excellent opportunities to evaluate the association between divergence patterns in biota and historical environmental changes, primarily because of the tremendous depth and magnitude of historical environmental change (Riddle 1995), and the strong relationships between biophysical attributes (e.g., elevation) and the pattern and distribution of biomes (Wickham et al. 1995).

Paleoecological records suggest that the western United States has undergone numerous 10

shifts in the distribution and relative connectivity of major biomes over the latest 40

million years (e.g., Axelrod 1958,1979,1983,1985, Graham 1986, VanDevender et al.

1987, Betancourt et al. 1990), although many species exhibited individualistic responses

to environmental change (Betancourt et al. 1990). The primary driver of these

transformations has been climate change associated with changes in ocean currents,

mountain building, and glacial-interglacial cycles (Betancourt et al. 1990, Haffer 1982,

Webb and Bartlein 1992, Riddle 1995). In the late Oligocene to mid-Miocene there was a

gradual but consistent trend towards intraannual wet and dry periods, and this resulted in

evolution and spread of the more xeric adapted Madro-Tertiary Geoflora, especially in the

southwestern United States and northern Mexico (Axelrod 1958). In the late Miocene to

Pliocene, there was significant mountain building throughout the western United States,

and significant rainshadow effects resulted from the uplift of the Sierra Nevadas, the

Cascades, and the Rocky Mountains (Axelrod 1958,1979,1985, Axelrod and Bailey

1976, Lucchitta 1979, Leopold and Denton 1987). The result was establishment of major desert formations throughout the western United States. Evidence from fossil records suggest that arid adapted plants found in deserts of today occurred along the more xeric edges of the Madro-Tertiary geoflora during the Miocene (Axelrod 1958). Therefore, it appears that some desert plants evolved gradually along the edges of more mesic plant communities and spread and diversified during subsequent drying trends (Chaney and

Elias 1938, Axelrod 1958).

During the Pleistocene, western North America was characterized by large-scale cyclic changes in biome pattern and distribution due to glacial and interglacial cycles

(Betancourt et al. 1990). Since approximately 750,000 years ago, these cycles have 11

occurred approximately every 100,000 years (Imbrie et al. 1984). Generally, during

glacial periods, woodland and grassland communities were widespread in what is today

desert (e.g., valleys between mountain ranges); during glacial periods, deserts had more

limited distributions than today and are hypothesized to have persisted in pockets or

refugia (Cole 1986, Betancourt et al. 1990). Conversely, interglacial periods (including

current conditions) have been characterized by widespread and relatively continuous

desert communities and woodlands have been limited to higher elevations (Betancourt et

al. 1990).

Woodland and Desert Taxa as Indicators of Concordance between Environmental Change and Lineage Divergence

To test hypotheses of multi-scaled lineage divergence and environmental change,

it is necessary to select species that have significant historical depth and broad ranges.

Moreover, it is important to select species with high habitat fidelity such that changes in

the distribution, sizes and connectivity of habitats would result in lineage divergence or lineage extinctions. Finally, selection of species with different habitat requirements may provide independent tests of congruence between environmental change and lineage divergence because different habitats (e.g., woodlands versus deserts) should respond differently to environmental change. Species with high desert or woodland habitat fidelity should exhibit divergence patterns that reflect historical changes in the distribution and connectivity of preferred habitats. It is likely that the size and connectivity of woodlands and deserts have differed considerably in any given period of time since deserts are more widespread during warm, dry periods, and woodlands during 12

cooler, more mesic periods (Betancourt et al. 1990). For example, because deserts have

experienced longer periods of isolation than woodlands during glacial-interglacial cycles

of the last 750,000 years (90,000 versus 10,000 years per 100,000 year cycle; Imbrie et al.

1984), they may possess species with greater genetic diversity than similar species

occupying woodlands. However, because desert species are anticipated to have had

considerable post-Pleistocene gene flow across broad geographic regions of the western

United States (primarily due to high habitat connectivity over the past 8,000 years),

populations have likely experienced considerable mixing; hence, there may be

considerable genetic mixing among regions. Woodland species may exhibit a similar

pattern, but with lower overall genetic diversity.

Although desert and woodland species are influenced by historical environmental

change, depths and frequencies of divergence should differ. This conjecture assumes that

deserts are strong barriers to migration by woodland species and woodlands are strong

barriers to migration by desert species. These patterns may not exist for those species that

can disperse across both deserts and woodlands (e.g., birds).

The relative depths of divergence that are uncovered by phylogeographic analysis

and the overall genetic richness for a given area may be a function of the degree to which suitable habitats were maintained during the last glacial maximum, approximately 18,000 years ago (Graham 1986). Areas where suitable habitats were lost (i.e., during the last glacial maximum) and then became reestablished during the current interglacial cycle, may possess populations with genetic composition reflecting recent divergence and dispersal. The ephemeral nature of these populations might preclude an ancient phylogenetic signal because the more ancient signal is lost when populations are 13

extirpated during glacial periods. The ephemeral nature of these areas may also result in

lower overall genetic diversity.

Phrvnosoma platvrhinos and Phrvnosoma douglassi as Indicators of Lineage Response to Historical Environmental Change

Analysis of Phrvnosoma platvrhinos and Phrvnosoma douglassi provides an

opportunity to evaluate multiple scales of divergence and diversification and hypotheses

of environmental change. Both species are widespread and exhibit different habitat

fidelities. P. platvrhinos is an arid adapted species occurring in the Sonoran, Mojave, and

southern portion of the Great Basin Desert, whereas P. douglassi occurs across most of

western North America in grasslands, woodlands, short-grass prairies, and the Great

Basin, generally above 2000 m. These two species allow for independent tests of

concordance between diversification and environmental change within the same broad

area (the western U.S.). Furthermore, both species have limited dispersal capabilities (as

compared to birds and other mobile taxa) and, therefore, should exhibit strong

diversification signals relative to environmental change.

P. douglassi and P. platvrhinos may exhibit different levels of divergence as a result of historical changes in their preferred habitats, and because of their relative times of divergence and geographic dispersal. Montanucci (1987) has surmised that P. douglassi went through a major radiation over much of western North America in the

Miocene in concordance with the spread of the Madro-Tertiary Geoflora, whereas P. platvrhinos diverged more recently from an ancestral desert form in the Pliocene.

Therefore, patterns of divergence and diversification in P. platvrhinos may correspond to 14

environmental changes between the late Pliocene and post-Pleistocene whereas

divergence patterns in P. douglassi may reflect a series of environmental changes between

the Miocene and post-Pleistocene.

Finally, each species may exhibit nested spatial hierarchies of divergence and

diversification within and among regions as a function of historical dispersal and

environmental change. Distinct regional trajectories are thought to result from relatively

deep environmental change and subsequent divergence (Riddle 1995). The region's

relatively unique environmental characteristics then shape the subsequent lineage

response of nested populations (e.g., regional conditions determine and constrain

potential responses of imbedded populations, Avise 1994). Populations would have had

to persist over relatively long periods to produce a multiple-depth hierarchy. When a set

of populations from a region are lost (e.g., due to glaciation), the lineage history of that

area is lost. It is possible that both species possess distinct regional lineages, primarily

because of their broad ranges and history of changes in preferred habitats (see earlier

discussion).

Avise (1989) identified four general phylogeographic patterns and associated hypotheses that relate to historical relationships within monophyletic lineages:

Category I High overall haplotype divergence with high geographic localization of

haplotypes;

Category n High overall haplotype divergence but low geographic localization;

Category in Low overall haplotype divergence but high geographic localization;

Category IV Low overall haplotype divergence and local differentiation. 15

Each category implies different historical relationships among taxa in a

monophyletic group. Category I suggests a relatively deep divergence history with strong

spatial hierarchy (e.g. regional integrity). Category II suggests a repeated series of

divergence and diversification, followed by widespread dispersal and population mixing

among regions (e.g., lacks regional historical integrity). This pattern would suggest that

populations become isolated (enough to cause a signal of diversification), and

subsequently, dispersed over large areas due to high mobility and/or low habitat fidelity

(e.g., habitat generalist). Category HI suggests that recent environmental change has

structured the taxa into spatial units. Category IV suggests that a series of populations act

as a metapopulation; little divergence and diversity result from frequent interchange

among populations. Patterns of haplotype diversity and divergence in P. platvrhinos and

P. douglassi can be assessed within the context of Avise's (1989) phylogeographic hypotheses to interpret lineage histories in the two lizard species.

Hypotheses

The following general hypotheses were evaluated relative to historical environmental change and patterns of divergence and diversification in P. platvrhinos and

P. douglassi:

Hypothesis 1 - P. douglassi and P. platvrhinos exhibit multiple levels of

divergence that roughly correspond to large-scale historical patterns of

environmental change prior to, during, and following the Pleistocene due to large-

scale environmental change that restricted gene flow among regions, and because 16 both species have relatively old histories in western North America (pre-

Pleistocene). Relatively deep branching patterns on the two species’ phylogenies are roughly concordant with time frames of vicariance associated with mountain building and other geologic activities within and among regions; more recent branching patterns are roughly concordant with time frames of vicariance or dispersal events associated with glacial-interglacial cycles. An alternative hypothesis is that one or both species exhibit little difference in divergence among regions due to lack of persistent barriers to gene flow.

Hypothesis 2 - P. platvrhinos and P. douglassi lineages are arranged in regional geographic hierarchies as a function of habitat persistence within a region and long-term persistence of barriers to gene flow among regions. An altnemative hypothesis is that one or more regions do not consist of distinct lineages due to lack of persistent barriers to gene flow with other regions.

Hypothesis 3 - Diversification patterns of P. platvrhinos and P. douglassi exhibit north-south gradients: a greater number of divergence levels and haplotypes exist in southern areas of the western United States than northern areas due to greater persistence of suitable habitats in southern regions since the Miocene. An alternative hypothesis is that a north-south gradient does not exist in one or both species because suitable habitat persisted in northern regions, or because most southern haplotypes dispersed into northern regions since the last glacial maximum. 17

Hypothesis 4 - P. douglassi and P. platvrhinos exhibit different patterns and levels

of lineage divergence as a function of differences in habitat preference, history of changes in preferred habitat, lineage history, and geographic range. An alternative hypothesis is that there is no difference in the overall pattern or depth of divergence between the two species. CHAPTER 2

PHYLOGEOGRAPHY OF PHRYNOSOMA DOUGLASSI

Introduction

The western United States has undergone signficant changes in the distribution

and evolution of biomes since approximately 28 million years ago (Axelrod 1958,1979,

1985, Van Devender 1986, Betancourt et al. 1990). Between approximately 3 to 28

million years ago, large scale changes in regional climates associated with mountain

building and changes in ocean currents and temperatures resulted in signficant changes in

the distribution of biomes over most of western North America. Since approximately 3

million years ago, glacial-interglacial cycles have been the primary drivers of large scale

changes in biome distributions.

Although historical environmental change has long been recognized for its role in

plant and animal evolution in western North America, only recently have we recognized

the importance of older environmental change in structuring lineage diversification

(Riddle et al. 1993, Riddle 1995).

Wide ranging species that have relatively deep lineage histories could demonstrate multiple levels of divergence, depending on the species' habitat fidelity and historical changes in the distribution, size, and connectivity of habitats within and among different regions. In particular, phylogeographic analysis of widespread woodland species with 19 relatively deep histories may provide a means to evaluate lineage divergence in response to relatively old and more recent environmental change. The woodland biome, including pine-oak, oak, chaparral, and pinyon-juniper, have a deep history of evolution in western

North America, and its distribution, size, and connectivity has varied greatly over the region from the Oligocene to the present in response to climate change associated with mountain building and glacial-interglacial cycles (Axelrod 1958, Betancourt et al. 1990).

P. douglassi is a wide ranging species that occurs in woodlands, coniferous forests, broken or badlands topographies within grasslands and prairies, and at relatively high elevations and latitudes within the Great Basin Desert (Montanucci 1987, Jones, personal observations). The species appears to have evolved in the southwestern United

States or northern Mexico during or prior to the Miocene (Montannuci 1987). A fossil of the species from the Split Rock Formation in Wyoming representing the middle Miocene

(approximately 15 million years ago, Robinson and Van Devender 1973) suggests that the species had evolved by the middle Miocene. Homed lizards have relatively low mobility

(as compared to birds and large mammals) and P. douglassi is generally absent in desert habitats of less than 1500 meters in elevation anywhere within its range. Even within

Great Basin Desert habitats, P. douglassi is often associated with woodland in mountainous terrain (Pianka and Parker 1975, Jones, personal observations). Therefore, even though the species is wide-ranging, it is generally absent from low to mid elevation deserts and demonstrates high fidelity to habitats with moderate to cool temperature regimes (Pianka and Parker 1975, Jones personal observations). Viviparity in P. douglassi is apparently an adaption to inhabiting colder environments (Dumas 1964).

Based on these factors, P. douglassi appears to be an appropriate species for investigating 20 the association between phylogenetic diversification and relatively old and recent historical environmental change.

Phylogenetic depth and diversity of lineages may vary among different regions as a function of habitat persistence. Persistence of suitable habitat influences effective pop ulation sizes (Gilpin 1987) which in turn influence extinction probability and genetic diversity of populations over time (Allendorf and Leary 1986, Lande and Barrowclough

1987). Therefore, persistence of suitable habitats within a region can have a dramatic effect on lineage diversity and persistence (Vrba 1985, Avise 1994). The phylogeography of P. douglassi may provide insight into long-term habitat persistence within and among regions.

The geographic pattern of population structure within and among regions could be used to infer mass tracking of habitat by P. douglassi. If habitats have experienced large- scale shifts (e.g., across regions), then little phylogeographic structure in P. douglassi within and among regions is expected (Avise 1994). Lack of phylogeographic structure would result from dispersal and mixing of populations over relatively large areas as habitats contracted and expanded. However, some geographic structuring within regions might result from lineage sorting (Avise 1994) and dispersal of peripheral haplotypes into expanding habitat areas.

Mitochondrial DNA Restriction Fragment Length Polymorphism (RFLP) and sequence variation were analyzed in P. douglassi from across its range to address the following hypotheses:

Hypothesis 1 - P. douglassi exhibits multiple levels of divergence that roughly

correspond to large-scale historical patterns of environmental change prior to, 21 during, and following the Pleistocene due to environmental change that restricted gene flow among regions. Relatively deep branching patterns on the P. douglassi phylogeny are roughly concordant with time frames of vicariance associated with mountain building and other geologic activities within and among regions; more recent branching patterns are roughly concordant with time frames of vicariance or dispersal events associated with glacial-interglacial cycles. An alternative hypothesis is that P. douglassi exhibits little difference in divergence among regions due to lack of persistent barriers to gene flow.

Hypothesis 2 - P. douglassi diversification is arranged in regional geographic hierarchies as a function of habitat persistence within a region and long-term persistence of barriers to gene flow among regions. An altnemative hypothesis is that one or more regions do not consist of distinct lineages due to lack of persistent barriers to gene flow with other regions.

Hypothesis 3 - Diversification patterns of P. douglassi exhibit north-south gradients: a greater number of divergence levels and haplotypes exist in southern areas of the western United States than northern areas due to greater persistence of suitable habitats in southern regions since the Miocene. An alternative hypothesis is that a north-south gradient does not exist in P. douglassi because suitable habitat persisted in northern regions, or because most southern haplotypes dispersed into northern regions since the last glacial maximum. 22

Methods and Materials

One hundred Phrvnosoma douglassi were collected from throughout the species' range (except in northern Mexico) between August 1991 and August 1994 (Figure 1,

Appendix I). Despite several attempts to obtain Mexican collecting permits, no permits were obtained; therefore, collections where limited to the species' range in the United

States. This represents the majority of the range of P. douglassi. Historical records from museums and literature, as well as interviews with biologists from throughout the west, were used to identify potential habitats and areas to sample. Specimens were obtained during short as well as extended field trips, and from biologists in various locations (via mail). Lizards were captured (by hand) by walking slowly through suitable habitat, but especially along disturbed areas traversing natural habitat, including secondary dirt roads and their shoulders where the species was abundant (see results). Specimens were obtained throughout daylight hours, although mid-summer sampling was mostly in the mid-morning or one-two hours before dark (Appendix I). In addition to noting the time of capture, habitat, approximate temperature, specific location, sex, and size, the number of individuals seen were noted (those collected and released). This was done to get a rough estimate of population size for individual sites. When possible, priority was given to collecting young animals, as opposed to mature animals - this reduced the probability of impacting the population. The general abundance of ants (a primary food item of both species) in areas where lizards were collected was also noted.

Lizards were transported live from the collection site and euthanized by freezing or injection of Nembutal (a dilution of sodium pentobarbital), upon return to Las Vegas.

Soft tissues, including the liver and heart, were extracted in the Molecular Systematics 23

.3 9 '

28 is ' 14.-15'

26 •11 * 29

34 .•S3. 32 35

>’>•1 A * • 20

23'.i

Figure 1 Locations of Phrvnosoma douglassi sample sites. The outer boundary indicates the approximate range of the species. Numbers correspond to collection data given in Appendix I. 24

Laboratory (UNLV Biology Building) and stored in Nunc tubes at -80C until tissues were

processed. For road-kill specimens, soft tissues were extracted at the site and placed into

a container of liquid nitrogen. These specimens were transferred from the container and

stored in Nunc tubes at -80C. Carcasses of lizards were preserved as fluid-preserved

museum research specimens, and deposited in the National Museum of Natural History,

Washington, D.C.

Collecting permits were obtained from each State (Appendix II) prior to collection

of lizards. Additionally, an animal handling protocol was submitted and approved by the

UNLV Animal Care Center prior to collection and transportation of lizards.

All mtDNA processing was conducted in the Molecular Systematics Laboratory,

Room 205, White Hall, University of Nevada, Las Vegas, with the exception of 65 P.

douglassi sequences, which were contributed by Kelly Zamudio, Department of Zoology,

University of Washington.

DNA Extraction

Tissues were removed from cold storage, thawed at room temperature

(approximately 22°C), and placed on ice in Nunc tubes. Total genomic DNA was

extracted from thawed liver and heart tissues according the protocol of Hillis et al.

(1990). Approximately 100 mg of either liver or heart tissue was placed into 500 ul of

STE (0.1 M NaCl, 0.05 Tris-HCL pH 7.5,0.001 EDTA) in a 1.5 ml plastic centrifuge tube. Then, 25 ul of a 20 mg mL'1 solution of proteinase K was added to the tube. The tissue was ground using a tissue homogenizer. Twenty-five ul of 20% sodium docecyl 25 sulfate was added, and the solution was incubated at 55°C for 2 - 2 Vz hour, with occasional mixing.

After incubation, an equal volume (550 ul) of PCI (phenol, chloroform, and isoamyl alcohol in the volumetric ratio 25:24:1) was added to the solution and incubated at room temperature for 5 minutes. The mixture was then centrifuged for 5 minutes at

7000 g. The aqueous layer was removed using a micropipette and transferred to a clean tube. The aqueous phase was then extracted with PCI a second time. Then, an equal volume of Cl (chloroform and isoamyl alcohol in the volumetric ratio 24:1) was added and incubated at room temperature for 5 minutes. The mixture was then centrifuged for 3 minutes at 7000 g. The aqueous layer was removed and extracted with Cl a second time.

After transferring the aqueous solution to a clean tube, one-tenth volume (about

45 ul) of 2 M NaCl and 1 ml of cold absolute ethanol were added to precipitate the DNA.

This solution was incubated overnight at -20°C. The DNA precipitate was then centrifuged for 5 minutes at 7000 g to pellet the DNA. The ethanol was decanted and the pellet dried for 10 minutes in a vacuum centrifuge. The pellet was resuspended in 250 ul double-distilled water (ddH20) overnight at 5°C. Samples of resuspended total genomic

DNA were stored in a freezer at -20°C.

PCR Amplification

Three distinct double-stranded fragments from protein-coding regions of the mtDNA genome (ND2-COI, ND4, Cyt-B) were amplified by Polymerase Chain Reaction (PCR)(Innis and Gelfand 1990). These regions were selected because of their relatively rapid rate of evolution. The ND2-COI primers amplified a fragment of approximately 2150 base pairs that 26 included the NADH subunit 2 gene, five intervening tRNAs, and 705 base pairs of the COI gene. The ND2-COI primer sequences are: L3880, 5'-TAAGCTATCGGGC CCATACC-3'; and H6033, 5'- ACTTCAGGGTGCCCAAAGAATCA-3'. The ND2-COIPCR fragment was used to produce Restriction-Fragment Length Polymorphism (RFLP) data.

The ND4 and Cyt-B primers were acquired from the University of Utah, Human

Genetics Laboratory in Salt Lake City, Utah. Fragments amplified for these two regions were used in the mtDNA sequence analysis. The ND4 primers allowed for sequencing of an approximately 350 base pair stretch, whereas the Cyt-b primers created a stretch of approximately 400 base pairs. Primer sequences for the ND4 fragment are: ND4, 5' -

ACCTATGACTACCAAAAGCTCATGTAGAAGC - 3'; and LEU, 5’ -

CATTACnTTACTTGGATTTGCACCA - 3. The Cyt-b primer sequences are: MVZ14,

5’ - GGTCTTCATCTCC GGTTTACAAGAC; and MVZ49, 5’ -

ATAAAAACAATGACAATCATACGAA - 3'.

The amplification procedure followed that described by Hillis et al. (1990). First,

3 ul of total genomic DNA, 10 ul of 10X Vent buffer (500 mM KCL, 100 mM Tris-HCL pH 9.0,1 .0% Triton X-100), 5 ul of each of the two appropriate primers (10 pm u l1), 16 ul 5mM dNTP mix, 1-2 ul 25 mM MgCl, 1 ul Vent polymerase, and enough ddH20 to make a volume of 100 ul. the solution was then vortexed, centrifuged, and overlaid with mineral oil. The DNA was amplified in a Perkin Elmer thermal cycler. Conditions for amplification were as follows: denaturation at 95°C for 1 minute, annealing at 50-57°C

(this varied depending on responses of individual samples), and extension at 72°C for 2 minutes with 4 seconds of extension added to each successive cycle. The amplification was run for 25-30 cycles (this varied depending on responses of individual samples). 27

Each sample was checked by running a 4 ul aliquot of each amplification through

a 0.8 % agarose gel by electrophoresis. After electrophoresis, gels were stained in

ethidium bromide for 20 minutes and visualized by ultraviolet fluorescence. The sizes of

the fluorescing DNA fragments were determined by their position relative to a standard

(latter) of fragments with known molecular weight, which was placed in the gel along

with samples. Successful amplifications were stored at -20 C°for later use.

RFLP Digestion of Amplified mtDNA Fragments

The ND2-COI gene region of a subset of samples representing the entire

geographic range of P. douglassi was digested with 16 restriction endonucleases. Three

of the 16 restriction enzymes showed no variation. The remaining 13 used in the analysis

of P. douglassi included Avail, BstNI, Ddel. HaelH Hhal. Hindi. Hinfl. HinPI, Hpall,

Mbol. MspI. Rsal. and Tag I. A total of 100 P. douglassi representing 48 distinct sample

sites were analyzed for restriction site variation, following the protocol of Dowling et al.

(1990). For each individual, 5 ul of amplified mtDNA product were digested according to manufacturers' protocols.

Fragments produced by each digestion were separated according to size by agarose gel electrophoresis. Each sample was mixed with 5 ul of 5% blucose and then placed into individual gel lanes. In addition to lizard samples, 25 ul of ladder (standards with known fragment sizes) were added to estimate fragment sizes of individual digests.

Gels were run in a IX TAE buffer at 140-145 V for approximately 2 hours. After electrophoresis, gels were stained in ethidium bromide for 20 minutes, then visualized by 28

ultraviolet fluorescence. The gels were photographed using a Polaroid camera for

subsequent data analysis.

Because mtDNA is a closed loop, the number of restriction sites is equal to the

number of fragments minus one (Avise 1994). This relationship allows for establishing

the relative position and number of restriction sites for each enzyme across the entire set

of specimens. Using this assumption, a linear representation of the relative position of

each restriction site for each enzyme for all specimens of P. douglassi was created. A

binary code was used to score the presence (1) or absence (0) of restriction sites per

enzyme per specimen. The binary matrix represents inferred restriction sites, not RFLP

fragments. A set of binary codes from all 13 enzymes represents a composite mtDNA

haplotype. Variation in this matrix provides the primary data used in phylogenetic

analysis (see below).

Sequencing Procedure

A 350 base-pair stretch of the mtDNA ND4 region, and the two ends of the Cyt-b

fragment (approximately 400 base pairs) for both species were sequenced using a Sanger

dideoxy procedure (Hillis et al. 1990). Double-stranded mtDNA fragments representing

these two regions were PCR-amplified using the ND4 and LEU (ND4 fragment) and the

MVZ49 and MVZ14 (Cyt-b) primers. Single-strand DNA was produced through a second round of PCR using 2.5 ul of the double-strand PCR product by diluting the concentration of one primer to 1/50 of the first-round concentration (MVZ49). Single strand fragments were then cleaned using an ultra-free millipore filter. Single-strand mtDNA was then subjected to the dideoxy sequencing procedure which was performed 29

with Sequenase 2.0 polymerase enzyme, reagents, and nucleotide triphosphates and [35S]-

dCTP. The MVZ49 and a separate ND4 internal primer (5' - CGTCAAACAGACCTA

AAATCA -3') were used as the sequencing primers for Cyt-b and ND4 fragments,

respectively. Samples representing the A C G and T reactions (from sequencing) were

loaded into a polyacrylamide gel and subjected to gel electrophoresis. Samples were

visualized by autoradiography; films with gels were exposed for 2-3 days. A total of 65

specimens of P. douglassi representing 38 different geographic locations were sequenced.

Additionally, two specimens of P. ditmarsi. and one specimen each of P. coronatum. P.

platvrhinos. and P. orbiculare were sequenced for outgroup analysis. Sequences of

specimens were then read (from exposed film) and loaded into a database. Sequences

were aligned by eye to each other and to the published sequences of Xenopus (Roe 1985).

Data Analysis

Restriction-Site Data Analysis

A restriction-site binary data matrix was constructed and stored in MacClade

(Version 3, Maddison and Maddison 1992). The P. douglassi phylogeny was estimated

using the maximum parsimony approach of PAUP 3.1.1 (Swofford 1993). The heuristic

search option was used with 10 replicate searches and random addition of taxa. Before

submitting the data matrix to parsimony analysis, redundant haplotypes and

uninformative characters were removed. However, all samples were used in constructing haplotype distribution maps (see results section). A bootstrap analysis of 1000 replications was performed in PAUP, each with a single stepwise addition, to evaluate 30 relative levels of character support for internal nodes of the estimated P. douglassi phylogeny.

The restriction-site binary data matrix was converted into haplotype and enzyme files and were then submitted to distance, haplotype, and Monte Carlo Analysis in REAP

(Version 4.0, McElroy et al. 1992). Genetic distances among composite haplotypes were estimated as the number of nucleotide substitutions per site from Nei (1987).

Phylogenetic relationships, based on the distance matrix, were evaluated using a

Neighbor-joining algorithm (Saitou and Nei 1987).

Restriction-site haplotype diversity of P. douglassi was evaluated in several ways.

First, a geographic map of haplotype distribution was compiled to provide a simplified representation of the spatial relationships among haplotypes. Second, haplotypes were assigned to the seven Kuchler physiographic regions in which they occurred and haplotype diversity and nucleotide divergence and diversity were calculated for regions according to Nei (1987, equations 8.4, 8.5, 8.12) as implemented in REAP. Nucleotide diversity is an indicator of the diversity of phylogenetic depth within a region. The Black

Hills Uplift physiographic region and the Upper Missouri Basin and Broken Lands region, and the Southern Rocky Mountain region and the Rocky Mountain Piedmont regions were combined for these analyses. Figure 2 illustrates the distribution of Kuchler

Physiographic Regions in the western United States. Third, the extent of geographic heterogeneity in the frequency distributions of haplotypes were estimated from a Monte

Carlo simulation (1000 replications) as implemented in REAP (Roff and Bentzen 1989).

The Monte Carlo simulation evaluates the probability of obtaining a X2 value from the distance data matrix that is larger than that obtained from a series of randomizations of 31

Northern Pacific Border Northern Rocky Upper Missouri Mtns. Basin & Broken Lands Cascade Mountains Black Middle Hills Columbia Rocky Uplift Sierra Plateau Mtns. Si Mountains' Wyoming Basin Upper Basin & Range Great Southern Plains Pacific Colorado Border Plateau

Southern L o w erH R_ocky Mtns Rocky Basin Mountain & Range Piedmont

Figure 2 Geographic distribution of Kuchler (1964) Physiographic Regions in the western United States. 32 that matrix. Finally, the genetic distance matrix and haplotype distributions were submitted to an AMOVA analysis (Excoffier et al. 1992) to determine the degree to which variation in the genetic distance matrix was apportioned within populations, among populations within regions, and among regions. Table 1 summarizes assignment of sample sites into populations and regions for the AMOVA analysis of restriction-site data.

Sequence Data Analysis

Two methods were used to estimate the P. douglassi phylogeny: Maximum likelihood from a genetic distance matrix (Felsenstein 1981,1993) and Maximum

Parsimony Analysis (Swofford 1993), in combination with various character weighting schemes. A data matrix of aligned sequences was created in MacClade (Version 3.0,

Maddison and Maddison 1992). Each base position was treated as an unordered character with four possible character states (ACGT). Ancestral character states were determined using outgroup comparisons (Maddison et al. 1984). In both analyses, trees were rooted with the most distant outgroup (P. platvrhinos). and other outgroups (P. orbiculare. and P. ditmarsi) were included in the ingroup, as a test of P. douglassi monophyly.

Maximum likelihood trees were constructed and evaluated using the DNAML program in Phylip 3.5 (Felsenstein 1993). Two weighting schemes were used in the analysis: a 1:5 transition to transversion weighting scheme and a codon position weighting scheme in an inverse proportion to the observed substitution rates at each position. For the second weighting, nucleotides were grouped into domains corresponding to first, second, and third codon positions. The relative observed rates of 33

Table 1 Hierarchical arrangement of sample sites into populations and regions for an AMOVA analysis of P. douglassi restriction-site distance data. Sample site numbers correspond to those illustrated in Figure 1. Region Population No. Sample Sites Columbia Basin 1 9,26,27 2 11,12 3 28,37,38 Wyoming Basin 4 34,39-42,44,45 5 43 Upper Missouri Basin 6 19 7 13-16,25,29,46-48 Southern Rocky Mountain 8 10 9 20-23 Upper Basin and Range 10 36 11 17,18 Colorado Plateau 12 1,2,3 13 4,5,24 14 35 15 31-33 Lower Basin and Range 16 6-8 17 30 34 positional variability were calculated by dividing the total number of substitutions in each domain of each gene region by the total number of sites sequenced in that domain, thus calculating a rate at which 1st, 2nd, and 3rd codon position substitutions occur in the ingroup. These rates where then used in the "categories" option of Phylip 3.5. This scheme effectively downweighted all 3rd position changes relative to the 1st and 2nd positions. Alternative topologies were tested for significance (p < .05) using Kishino and

Hasegawa's test for Maximum Likelihood (Hasegawa and Kishino 1989).

A pairwise sequence comparison was performed in the Molecular Evolutionary

Genetics Program (MEGA, Version 1.01, Kumar et al. 1993) to determine the distribution and amount of variation, and levels of saturation by codon position, in the sequence database. To test for saturation of base substitutions at each codon position, percent sequence divergences (based on two substitution classes: transitions and transversions) were plotted against Tamura-Nei estimates for relative divergence for the same substitution classes (Figure 3). The Tamura-Nei divergences are analogous to the uncorrected percent divergences; however, they consider deviations from equal base compositions and differences in substitution rates among bases. Non-isometric plots indicate increasing saturation of substitutions for transitions or transversions at each codon position. The analysis shown in Figure 3 indicates that two classes of transitions are potentially saturated: those at the 1st and 3rd codon positions. Therefore, the two weighting schemes that reduce the weight of these codon positions (as described above) were used.

Maximum parsimony phylogenies were estimated using PAUP 3.1.1 (Swofford

1993). Similar to the restriction-site parsimony analysis, a heuristic search option with 10 35

o CM

o CO o 0 CO o CO c 0 k—O) a> >

CO ro D. « 0 0 E CM m LU

CM 2 0 Z1 i_Cti © 3 E |2

a. CM ratio is the observed proportion of pair-wise divergence/the Tamura-Nei estimated pair-wise estimated Tamura-Nei divergence/the pair-wise of proportion the observed is ratio divergence in the sequence data. sequence the in divergence o CO <0 CM CO 3 3 position. codon by (tv) transversions T! and (ta) transitions in saturation base-pair Estimated (%) ssoua6j3Aja asjAAjjed leuopjodcud Figure Figure 36

replicate searches and random addition of taxa was used. Similar to the Maximum

Likelihood Analysis, two weighting schemes were used: one assuming equal weights for every codon domain but assigning different weights to transitions (1) and transversions

(5) by constructing a step matrix, and one using the inverse of the observed rates of change at each codon position in the "weights" option in PAUP. A bootstrap analysis

(1000 replications, each with a single, random stepwise addition of taxa) was used to evaluate relative levels of character support for internal nodes of the estimated P. douglassi phylogeny.

Haplotype divergence and diversity were estimated for each region from the sequence database in an identical manner to that listed for the restriction-site analysis (see above). Table 2 summaries the assignment of sample sites to populations and regions for the AMOVA analysis based on sequence data.

Results

Distribution and Ecology

One hundred P. douglassi were collected from 48 sites throughout the species range. During the first 2 years of the study (1991-1992), lizards were difficult to find.

On an extended field trip in 1992, lizards were found on only 2 of 25 sites. However, after discovering where to search for P. douglassi. lizards became easier to find (see below).

Lizards were commonly observed on sites with annual forbs and grasses and exposed rock or bare ground throughout their range, regardless of habitat type. In the western Great Plains and eastern Montana, P. douglassi was observed on broken, open 37

Table 2 Hierarchical arrangement of sample sites into populations and regions for an AMOVA analysis of P. douglassi sequence distance data. Sample site numbers correspond to those illustrated in Figure 1. Region Population No. Sample Site Columbia Basin 1 9

2 26

3 28,37,38

4 11,12

5 27

Wyoming Basin 6 34,39-42,44,45

7 43

Upper Missouri Basin 8 13-16,19,25,29

9 46-48

Southern Rocky Mountain 10 10,23

11 20-22

Upper Basin and Range 12 36

13 17,18

Colorado Plateau 14 1,2,3

15 4,5,24

16 35

17 31-33

Lower Basin and Range 18 6-8

19 30 38

terrain adjacent to rivers, in patches of badlands, or on broken hillsides within expansive

prairies. At relatively high elevation sites within coniferous forest, lizards were observed

along the edges between meadows or early successional stage forests, and developed

forests where soil was interspersed with annual forbs, grasses, and rocks. On sites within

the Wyoming Basin, lizards were observed in low-height desert scrub with mixtures of

annual grasses and forbs, perennial shrubs, and open ground. In the Great Basin of

Nevada and western Utah where the species is sympatric with P. platvrhinos. P. douglassi

was found in areas of sagebrush, but only in areas adjacent to pinyon-juniper woodland or

grassland. In southeastern Arizona, P. douglassi occurred in a variety of habitats,

ranging from grassland to pine-oak forests and woodlands. On all of these sites, ants and

insects were abundant.

Where P. douglassi were found, they appeared to be quite abundant. For example, on the Fossil Agate site in western Nebraska, the Bitter Creek, Wyoming site, the Mount Shasta site in northern California, and the Duck Creek site in Southern Utah,

20 or more lizards were counted representing a number of age classes in less than one hour of search time. Of the sites searched, P. douglassi was found to be most abundant along human disturbed sites, especially lightly used dirt roads and hiking trails. This was especially evident among sites in the western Great Plains and eastern Montana.

Frequently, P. douglassi was found on dirt roads that extended several kilometers into extensive prairies. The prairie habitat consisted of nearly 100% ground cover and never yielded any P. douglassi. but the roads that emanated from hills along rivers or badlands and traversed prairies always yielded lizards. This phenomenon was true of roads, 39

powerlines, and trails throughout the species' range. Similar to natural sites where lizards

were abundant, these sites harbored large numbers of ants and insects.

Phylogenetic Relationships and Phylogeography - Restriction-site Analysis

Restriction digests of the ND2-COI mtDNA fragment resulted in a total of 31

unique haplotypes (Table 3). Estimated sequence divergence between haplotypes ranged

from 0.5 percent to 25.3% (Appendix HI). A Neighbor-joining tree based on sequence

divergence revealed several levels of divergence and associations (Figure 4). Haplotypes

from the Pacific Northwest, from eastern and southeastern Arizona, and from the

remainder of the species range formed three deep branches (divergence >12.3%, Figure

4). The large deep clade comprising the majority of the species range consisted of several

other relatively deep splits, including a branch with one haplotype from central Arizona

(9.9% divergence), and branches separating a mostly western clade consisting of northern

and southern Utah, northern Arizona, and northeastern Colorado (the exception) from a

more eastern and northern clade consisting of eastern Utah, the Wyoming Basin, the

Upper Missouri Basin and Broken Lands, northern and eastern New Mexico, west Texas

and southeastern Arizona (Figure 4). The large, mostly western clade had three other

relatively deep branches (divergence from 6.5 to 8.2%): a clade consisting of the

Grantsville, Utah population (8.6% divergence), a clade consisting of the Pawnee

National Grassland population in northeastern Colorado (8.2% divergence), and the clade consisting of northern Arizona and southern Utah (6.5% divergence, Figure 4). Each of three clades have further differentiation, ranging from 0.7 to 3.9% divergence (Figure 4).

The mostly northern and eastern clade exhibited four relatively deep branches of Table 3 Summary of P. douglassi mtDNA haplotype restriction-site patterns. Haplotype num bers correspond to geographic locations

given in Figure 6. Restriction-site characters are give in binary form where present = 1 and absent = 0. Appendix I for specific

inform ation regarding specim ens. pi •a 1 x a m •a 1 «/) co CO O o O o o o o o O o o o o o o o o ’ © o O O o O o O O o O o o o o iH - - 1-H 1-H 1-H o 1-H © r-H o i-H i-H 1-H 1“H 1-H -H iH iH iH iH iH rH £5 5 £ r-H i-H i o 1—H i-H H ^ i-H i—H H i i—H i-H i—H i—H 1—H H i- r-H o o i-H i-H o o o o o o O o s O - iH 1-H rH E ^H i—Hr-H © i—< r-H 0 0 0 ^ ^ - 0 0 0 0 0 0 OOOOOO-—'OOOOOO 0 0 H i o o o o o 1 H iH i-H -H 0 2 o o o § 1—H o co o o i—H rH H1H1-H 1-H rH © iH 1-H H ? 1—H o o o o o o o o s o o H r ° o 1 - o 1 o o o o o o o o o o o O 8 s - rH O r-H 1-H 1-H © © 1—H o o o i-H o O - 1-H 1-H 1-H 1—H 1-H o H i o H r o H r o 1 - - - - i-H i-H r-H i-H H - 8 o 8 o o o 1—H o rH © 1-H 1—H in 1-H o o o o o

o I 0 9 o W v • / o © 1-H © rH 1-H i-H o 1—H i-H 1-H o © 1-H 1-H o i-H © o o o o o HrH rH o o o i-H o rH O o o o o o o o o o o o t— o o o © 8 o o o © © o 1-H o r-H r-H r-H H i o 1-H i H i o o oo o o o o - i-H 1-H 1-H 1-H 1-H H - i-H o r-»n © O o o o o o 1-H Hr rH rH rH i-H o - 1—H r-H o i-H o - 1H - i"H o o o i-H o 1-H o 1-H o 1 © 1-H © i-H O i-H © i-H o l-H o o 1 ^ 0 o 0 1-H - O i-H © r-H i-H i-H O i-H i-H H - 0 9 o o i-H o H o © © o 1 1 o 1-H - - i-H i-H H - - -H iH i-H H i- i-H H - o o o o o o o o © g © H o o O o O 1-H H o H1-H rH i-H © i—H H r o O O © 8 H o o o H © 1-H 1-H CN 1-H o o o o o o 8 1 j v o H1-H rH 1—H ,-H © © o rH 1—H© i-H H r CO rH o — 1—H r-H 1—H i-H o o o o o o o o rH o o i-H H - o Q o 8 o o o o o o i-H rH 1 o H -

40 41 1— o oO orH o o

o o o o H iH CO o o «

•a o o G. o o 8 o o o o o

OOOOOOOOOOQO © © O oooooooooooo o o © o

o 0 © © T—' 1-H 1—1 © ctiD 1-1 o ffi s © 2 © © o © © o © ©

a o © © 8 8 8 © 8 © ©

©OOOOOO© © © © © rHiHrHrHlHrHiHrH i-H i-H i-H i-H iH © rH rH H H rH rH rH rH 1-H © O O O O O O i-h ©

ao © © © © © © © © 1-H © © © © © © 1-H © © 1-H © O O

©©oooooi-i^; O O rH rH rH rH rH rH H r H ^ H i—H i—H t_~J l ~J 1—H ^ H r H O O O O ^ ^ O ’—Ii i-H rH rH C) i™H < iH rH i-H i-H i-H i-H i-H rH H O O ©©©©©©©OOOO—iO

1 (NWAZ) 13 (N AZ) 14 (N AZ) 18 (N AZ) 5 (S UT) 16 (S UT) • 15 (S UT) • 17 (S UT/E NV) • 9 (NE CO) ■ 29 (NE CO) ■ 19 (N UT) ■ 20 (N UT) . 2 (E UT) . 4 (N NM) ■ 6 (N NM) • 3 (S NM) ■ 30 (W TX) 10 (NB/WY/MT/Canada/ ND/SD) • 21 (NE WY) • 11 (NB) ■ 31 (SE AZ) ■ 26 (Cent. AZ) 8 (ID) 22 (SE OR) 23 (NE OR) 24 (NE WA) 25 (NE WA) 28 (S WA)

■ 7 (W NM) 12 (E AZ) 27 (SE AZ) + + 15.0 12.5 10.0 5.0 1.0 % Sequence Divergence

Figure 4 Phylogenetic relationships of composite R douglassi haplotypes estimated from a Neighbor-joining analysis of restriction-site data. Numbers correspond to haplotypes listed in Table 1 and illustrated in Figure 6. 43

divergence: a branch representing a single haplotype in southeastern Arizona (7.4%

divergence), a branch representing a single haplotype from western Texas (7.3% divergence),

a branch representing all haplotypes in the Wyoming Basin and Upper Missouri Basin and

Broken Lands (7.4% divergence), and a branch representing eastern Utah and northern and

southern New Mexico (6.6% divergence, Figure 4). Each of these four clades had further

differentiation representing 0.8 to 3.5% divergence (Figure 4). Other than the Idaho Falls

population (4.0% divergence), Pacific Northwest populations had relatively low divergence

within the clade (Figure 4). Finally, the three haplotype clade representing eastern and

southeastern Arizona had considerable divergence among its members (3.3 to 9.6%

divergence, Figure 4).

A comparison of nucleotide divergence distances among the seven Kuchler

physiograhic regions reveal the greatest distance between the Columbia Basin and the

Wyoming Basin (14.3%, Table 4), despite the fact that these two regions are adjacent to each

other near the intersection of Wyoming, Utah, and Idaho (Figure 2). The Columbia Basin

was consistently the most different from any other region (range 8.6 to 14.3%). The

Wyoming Basin and Upper Missouri Basin and Broken Lands were the most similar regions

(0.01%). A Monte Carlo simulation (1000 replications) of haplotype heterogeneity by physiographic region revealed a highly significant, non-random pattern (probability of exceeding the original X2 value was 0).

An AMOVA analysis of three scales of haplotype divergence (nested analysis, including within population, among populations within individual physiographic regions, and among region variance) revealed that among populations within regions and among region scales explained the greatest percentage of overall variance in the restriction-site 44

Table 4 Nucleotide divergence of P. douglassi among regions and nucleotide diversity within regions (diagonal) based on restriction-site analysis. Regions are based on Kuchler physiographic regions within the species' geographic range. Rocky Colub WyomB UpBasnR LwBasR ColPlat UpMiso Mtn. 0.015 14.3 0.002 9.7 10.6 0.025 9.4 2.7 5.5 0.081 8.6 4.0 3.8 0.2 0.056 14.3 0.1 10.6 2.7 4.0 0.003 7.4 6.8 3.5 3.0 2.1 6.5 0.006 45 distance matrix (51% and 38%, respectively) and the probability of a random group of heterozygosity being greater than the observed heterozygosity (1000 reps) was 0.010.

The Maximum Parsimony analysis supports deep divergence of a Pacific Northwest

Clade, and a large clade consisting of all other haplotypes (Bootstrap value of 93%, Figure 5).

Additionally, within this large clade, there is a relatively deep branch separating the Pawnee

National Grassland and the remainder of haplotypes (Bootstrap value of 91%). Beyond these, no other deep branches were supported by bootstrap values of over 30% (Figure 5).

A comparison of haplotype diversity by Kuchler physiographic regions revealed that the Lower Basin and Range Physiographic region had the greatest haplotype diversity (0.96), followed by the Colorado Plateau (0.89, Table 5). The Columbia Basin had relatively high haplotype diversity, whereas the Wyoming Basin, Upper Missouri Basin and Broken Lands, and Southern Rocky Mountains had relatively low diversity (Table 5). The Lower Basin and

Range and Colorado Plateau regions had the greatest nucleotide diversity (0.81 and 0.56, respectively) and the Wyoming Basin and Missouri Basin and Broken Lands regions the lowest (0.002 and 0.003, respectively, Table 4).

Most restriction-site haplotypes were locally distributed (Figure 6). However, haplotype #10 had a very broad distribution across the entire Wyoming Basin and Upper

Missouri Basin and Broken Lands physiographic regions. Additionally, haplotype #22 was widespread across central Oregon and northern California. Finally, haplotype #17 ranged from the northeastern portion of the Colorado Plateau in southern Utah through eastern

Nevada. 46

-1 (NW AZ) 35 -13 (N AZ) 30 -18 (N AZ) 61 34 -14 (N AZ) - 26 (Cent. AZ)

45 - 7 (W NM) 24 -12 (E AZ) ■ 27 (SE AZ) 26 - 2 (E UT) 39 - 4 (N NM) 18 - 6 (N NM) -19 (N UT) 15 ’ 17 (S UT) •31 (SE AZ) ■ 5 (S UT) ■ 15 (S UT) . 16 (S UT) ■ 10 (NB/WY/MT/SD/ND/Canada) 34 ■ 20 (N UT) 70 • 21 (NE WY) ■ 11 (NB) 91 ■ 3 (S NM) • 30 (W TX)

91 . 9 (NE CO) . 29 (NE CO)

35 22 (SE OR) 23 (NE OR)

65 78 25 (NE WA) 24 (NE WA) 28 (S WA) 8 (ID) Figure 5 Phylogenetic relationships of composite P. douglassi haplotypes estimated from a Maximum Parsimony analysis of restriction-site data. Numbers on internal branches are bootstrap values based on 1000 replications. Numbers on terminal branches correspond to those given in Figure 6. 47

Table 5 Restriction-site haplotype diversity of P. douglassi by Kuchler Physiographic Region. Haplotype diversity was generated from the DA program in REAP. Region Haplotype Diversity Columbia Basin 0.73 +/- 0.09 Wyoming Basin 0.28+/-0.11 Upper Basin and Range 0.61 +/-0.16 Lower Basin and Range 0.96 +/- 0.06 Southern Rocky Mountains 0.89 +/- 0.05 Upper Missouri Basin 0.15+/-0.13 Colorado Plateau 0.29+/-0.19 48

L* •

.26 ©

Figure 6 Geographic distribution of composite restriction-site haplotypes of P. douglassi. Haplotype numbers correspond to those given in Table 1. 49

Sequence Analysis

Sequence analysis of 783 base pairs of the ND4 and two ends of the Cyt-b mtDNA fragments of 65 P. douglassi (Appendix IV) yielded a total of 43 composite haplotypes

(Appendix V). Alignment was straightforward because no insertions or deletions were found in the genes sequenced. Sequence divergence between outgroup and ingroup lineages ranged from 12.9 % (between P. orbiculare and P. douglassi haplotype #26 from Grand County,

Utah) to 18.7% (between P. platvrhinos and P. douglassi haplotype # 17 from Williams,

Arizona, Appendix V).

Both weighting schemes (ta/tv ratio and codon position weighting) yielded two iden tical trees in the Maximum Likelihood analysis. Similar to the restriction-site analysis of P. douglassi. the Maximum Likelihood analysis revealed several scales of divergence (Figure 7).

However, the analysis yielded some different phylogeographic relationships among haplo types. However, both analyses showed a deep divergence of the Columbia Basin lineage.

Furthermore, both analyses demonstrated a geographic disjunction among haplotypes in a clade consisting of a population from northeastern Colorado and populations of the Upper

Basin and Range physiographic region (Figures 4 and 7). Finally, both analyses demon strated similar relationships among many closely related terminal nodes (Figures 4 and 7).

The major patterns indicated by the Maximum Likelihood analysis of sequence variation were: (1) paraphyly of P. douglassi as a whole with regard to P. ditmarsi. (2) a geographic division of P. douglassi into three main clades: a group of populations in the

Columbia Basin, a group of populations inhabiting mostly the Upper Basin and Range and

Colorado Plateau regions (with notable exceptions from northeast Colorado and southern 50

P. ditmarsi Cloudcroft, NM (3) - Pawnee Nat. Grassl., Co (24) m Hualapal Mtns, AZ (1) Williams, AZ (17) -E;Flagstaff, AZ (11) Currie, NV (28) rl Grantsville, UT (18) Webster Flat, UT (5) Duck Ck, UT (12) Cochise -DuckCk, UT(13) vn Co., AZ — Mammoth Ck, UT (14) '(32) — Mammoth Ck, UT (15) - Grants, NM (25) Rio de Las Vacas, NM (6) Moab, UT (2) Taos, NM (4) _ Otero Mesa, TX (27) VI — Cochise Co., AZ (22) _n— Cochise Co., AZ (30) 1 i— Cochise Co., AZ (31) •— Cochise Co., AZ (29) Thompson, UT (26) Canada (33) II Canada (34) rE!Canada (35) — Medora, ND(16) — Forsyth, MT (42) — Crow Buttes, SD (41) NE, WY (43) Tonto, N.F., AZ (21) Pietown, NM (7) Springerville, AZ (10) Springerville, AZ (9) Idaho Falls, ID (8) Santiam Junct, OR (36) Mtn Shasta, CA(19) Mtn Shasta, CA (20) Bums, OR (39) Bums, OR (40) Wilbur, WA (37) Boardman, OR (38) t E Ellenburg, WA (23)

Figure 7 Phylogenetic relationships of composite Phrvnosoma douglassi haplotypes estimated from a Maximum Likelihood analysis of sequence data. Haplotype numbers correspond to distributions illustrated in Figure 10. Roman numerals correspond to clades illustrated in Figure 8. 51

New Mexico), and a large clade including the remaining populations, and (3) basal position

of the Columbia Basin clade relative to P. ditmarsi and all other lineages of P. douglassi

(Figure 7). A simplified tree representing broad phylogeographic relationships is illustrated

in Figure 8.

Several levels of divergence and diversification are evident within each of the three

deep branches described above. Within the Upper Basin and Range/Colorado Plateau some

distinct phylogeographic relationships were evident, including a group representing northern

Arizona (3.7% divergence), a group representing southern and northern Utah and eastern

Nevada (4.2% divergence), and a small group consisting of the Pawnee National Grassland

and Cloudcroft, New Mexico populations (5.3% divergence, Figure 7). Within the large

eastern and northern clade, five general groups were evident, including a group of

populations representing the entire Wyoming Basin and Upper Missouri Basin and Broken

Lands regions (5.2% divergence), a group of populations representing northern New Mexico

and eastern Utah (4.4% divergent), a group representing eastern and southeastern Arizona and western New Mexico (4.0% divergence), a group representing southeastern Arizona and west Texas (5.2% divergence), and one population from southeastern Arizona (7.8% divergence, Figure 7). The Columbia Plateau had multiple-levels of divergence, with a multi-level group representing Oregon, Washington, and northern California, and a single haplotype representing the Idaho population (4.7% divergence, Figure 7).

Geographic relationships of P. douglassi at different depths of divergence provide insight into historical relationships within the species. The deepest level of divergence represents a break between the Columbia Basin region and the remaining populations Figure 8 Simplified regional phylogeny of Phrynosoma douglassi based on the Maximum Likelihood analysis (Figure 7). 53

(Figure 7). The next level of divergence represents differentiation of the large eastern clade into three regions, including the disjunct region noted in both the restriction-site and sequence analyses, and P. ditmarsi (Figure 7). The third and fourth levels of divergence represent further differentiation in the two original clades (Figure 7).

Nucleotide diversity generally increased from north to south, with the exception of the

Columbia Basin which was moderately diverse (0.029, Table 6). The Colorado Plateau and

Lower Basin and Range regions had the greatest nucleotide diversity (0.051 and 0.050, respectively) and the Upper Missouri Basin and Broken Lands and Wyoming Basin the lowest diversity (0.003 and 0.006, respectively, Table 6).

Similar to the restriction-site analysis, P. douglassi populations were assigned to physiographic regions (see previous results) and then analyzed for nucleotide divergence in

REAP. This analysis differed from the restriction-site analysis in that it used sequence- derived haplotypes and the Maximum likelihood matrix generated in PHYLIP 3.5 (Appendix

V). Similar to the analysis on restriction-site haplotypes, the Columbia Basin consistently exhibited the greatest divergence among any of the regions, the Columbia and Wyoming

Basins were highly divergent, and the Wyoming Basin and Upper Missouri Basin and Broken

Lands regions were the most similar (Table 6). However, divergence values among regions were considerably lower than those generated from the restriction-site analysis (Table 6). A

Monte Carlo simulation (1000 replications) of haplotype heterogeneity by physiographic region revealed a highly significant, non-random pattern (probability of exceeding the original X2 value was 0).

An AMOVA analysis of three geographic scales (nested analysis, including within population, among populations within individual physiographic regions, and among 54

Table 6 Nucleotide divergence distances among and nucleotide diversity within (diagonal) Kuchler Physiographic regions for P. douglassi as calculated from the sequence distance matrix in REAP. ColubB WyomB UpperBR LowerBR ColPl UMissou SRockMtn 0.029 8.4 0.006 9.3 6.3 0.023 6.5 2.2 5.2 0.050 7.0 3.6 1.6 2.2 0.051 8.7 0.1 6.6 2.5 3.9 0.003 5.9 2.6 3.0 1.3 0.5 2.9 0.010 55

region variance) yielded results similar to the restriction-site AMOVA analysis; scales

representing variance among populations within regions and among region explained the

greatest percentage of overall variance in the distance matrix (48% and 38%, respectively)

and the probability of a random group of heterozygosity being > the observed heterozygosity

(1000 reps) was 0.010.

Maximum parsimony analysis of the 783 nucleotide positions and a 1:5 scheme in

favor of transitions resulted in 10 most parsimonious shortest trees (1086 steps in length,

CI-0.523 and RI=0.789). A strict consensus tree of these 10 minimum-length trees yielded a

very similar phylogeny to that generated from a Maximum likelihood approach (). Weighting

by codon position in parsimony yielded congruent results. Bootstrap values for different

nodes in the tree vary, but unlike the restriction-site tree, many nodes exceed 80% ().

Similar to the restriction-site analysis, haplotype diversity was greatest in the Lower

Basin and Range and Colorado Plateau physiographic regions. However, unlike the restriction-site analysis, the Columbia Basin and Upper Missouri Basin had relatively high haplotype diversity (Table 7), primarily because sequence analysis can detect finer levels of genetic differentiation than restriction-site analysis. All of the haplotypes in the Upper

Missouri and Broken Lands region represented less than 1.0% divergence from their closest relative (Appendix V).

Similar to the restriction-site analyses, geographic distributions of haplotypes derived from sequence analysis are highly localized, but to a greater extent (Figure 10). Only haplotype #16 within the Wyoming Basin and Upper Missouri Basin and Broken Lands regions exhibits a regional distribution (Figure 10). 56

100 P. ditmarsi 100 Hualapai Mtns, AZ (1) 76 Williams, AZ (17) 64 Flagstaff, AZ (11) 96 Mammoth Ck, UT (15) 80 64 Duck Ck, UT (12) 97 Duck Ck, UT (13) Webster Flat, UT (5) 65 99 Mammoth Ck, UT (14) m~ 57 Grantsville, UT (18) Currie, NV (28) 99 Cloudcroft, NM (3) Pawnee Nat. Grassl., Co (24) Moab, UT (2) 61 Grants, NM (25) IV Rio de Las Vacas, NM (6) Taos, NM (4) Otero Mesa, TX (27) Medora, ND (16) 53 Canada (33) 70 Canada (34) Canada (35) 100 Crow Buttes, SD (41) 86 99 Forsyth, MT (42) NE, WY (43) II Thompson, UT (26) Cochise Co., AZ (32) vn Pietown, NM (7) 68 Cochise Co., AZ (22) Springerville, AZ (10) Springerville, AZ (9) 92 Tonto, N.F., AZ (21) 73 yy i Cochise Co., AZ (29) VI Cochise Co., AZ (30) t=EE Cochise Co., AZ (31) Idaho Falls, ID (8) Mtn Shasta, CA (19) Mtn Shasta, CA (20) 100 92 Bums, OR (39) Bums, OR (40) Santiam Junct, OR (36) Wilbur, WA (37) 96 Boardman, OR (38) Ellenburg, WA (23) P. orbiculare P. platyrhinos

Figure 9 Phylogenetic relationships of composite Phrvnosoma douglassi haplotypes estimated from a Maximum Parsimony analysis of sequence data. Haplotype numbers correspond to distributions illustrated in Figure 10. Roman numerals correspond to clades illustrated in Figure 8. 57

Table 7 Sequence haplotype diversity of P. douglassi by Kuchler Physiographic Region. Haplotype diversity was generated from the DA program in REAP. Region Haplotype Diversity Columbia Basin 0.95 +/- 0.04 Wyoming Basin 0.71 +/- 0.18 Upper Basin and Range 0.60+/-0.17 Lower Basin and Range 1.00+/-0.05 Southern Rocky Mountains 0.50 +/- 0.27 Upper Missouri Basin 0.82+/-0.10 Colorado Plateau 0.99 +/- 0.03 58

Figure 10 Geographic distribution of composite sequence haplotypes of Phrvnosoma douglassi. Haplotype numbers correspond to those given in Figures 7 and 9 and Appendix V. Roman numerals correspond to clades illustrated in Figure 8. 59

Discussion

Multiple-levels of divergence in P. douglassi suggest a relatively deep and complex history. The genus Phrvnosoma is thought to have diverged from sceloporines during the late

Oligocene-early Miocene (Montanucci 1987, Etheridge and de Queiroz 1988). Habitat types mapped on to the phylogeny of the genus (from Reeder et al., unpublished manuscript) suggests that the ancestral Phrvnosoma inhabited deciduous forest or woodland (Figure 11).

Deciduous forests were widespread in the late Oligocene through the early Miocene (Axelrod

1958). Sometime between the early and mid-Miocene, the ancestral Phrvnosoma differentiated into some lineages that were adapted to more xeric forest and woodland forms

(Figure 11). One of these lines included the ancestor that gave rise to the P. orbiculare. P. douglassi, and P. ditmarsi woodland clade (Figure 12).

Unlike rapid environmental changes in other regions of the western United States during the Miocene, changes in the Columbia Basin were more gradual (Chaney and Axelrod

1959, Leopold and Denton 1987). During the Miocene, this area maintained deciduous hardwood and montane conifer forests. Although patches of open ground and grassland existed on volcanically disturbed areas throughout the Miocene (Taggart et al. 1982), large stands of mesic forest dominated the regional landscape until the Pliocene (4-5 million years ago). At this time, grassland and steppe habitats became more dominant, especially in the easternmost portion of the Columbia Basin (Leopold and Denton 1987). The drying trend within the Columbia Basin was associated with rainshadow effects caused by the uplift of the

Cascade and Olympia Mountains (Chaney and Axelrod 1959). This pattern of habitat stability and change corresponds to relative depths of P. douglassi divergence in the 60

Tropical Dec. Forest (Mexico)

Chihuahuan Des./Prairie — Sonoran Desert Sonoran/Mojave — Deserts Sonoran Desert

Chihuahuan Desert Coastal Chaparral/ Oak Woodland Woodland/Forest/ Great Basin Desert/ Prairie Montane Woodland/ Forest Mexico

Montane Woodland/ Forest Mexico Tropical Arid Forest/Scrub Mexico

Tropical Arid Forest/Scrub Mexico

Tropical Dec. Forest?

Figure 11 Habitat preferences of individual species of Phrvnosoma mapped on to the phylogeny of the genus (Figure 12). 61

P. asio

P. cornutum I— P. mcalli

P. platyrhinos

P. solare

P. modestum

P. coronatum

P. douglassi

P. orbiculare

P. ditmarsi

P. taurus

P. braconnieri

Ances

Figure 12 Estimated phytogeny of the genus Phrvnosoma based on morphological and molecular data (from Reeder, unpublished manuscript). 62

region (as estimated from sequence analysis): there is a deep level of divergence (>11%) of

the region followed by divergence of the Idaho Falls population on the east side of the

Columbia Basin (4.1%, based on sequence data analysis). The divergence of the region from

all other regions may reflect the relative stability of Columbia Basin habitats during the

Miocene, as compared to more rapid environmental changes outside of the Columbia Basin

(Chaney and Axelrod 1959, Leopold and Denton 1987). Lack of haplotype mixing among

and depth of divergence between haplotypes of the Columbia Basin and Upper Basin and

Range regions suggests a strong barrier to gene flow between these two adjacent regions

(Avise et al. 1987). An extensive volcanic formation (The High Lava Plains) that was

formed in the mid-Miocene separates these two regions along their border (Christiansen and

McKee 1978), and this may have acted as a strong zoogeographic barrier to gene flow. The northern Rocky Mountains of western Wyoming and northern Utah may be a strong barrier to gene flow among these two regions.

The major disjunction of populations (western group and two eastern groups, Figure

8), representing a relatively deep level of divergence, may reflect a haplotype that was widespread across the region prior to the formation of the Rocky Mountains and uplift of the

Colorado Plateau. Isolation due to uplift of the Rocky Mountains and Colorado Plateau and subsequent genetic drift may have eliminated populations with this haplotype from intermediate areas.

Overlap of haplotypes representing several scales of pre-Pleistocene divergence in southeastern Arizona suggest that large blocks of habitat have persisted in this area from the

Miocene to present (see discussion below). 63

Although pre-late Pleistocene events formed the basic phylogeographic structure differentiating large regions (as in mammals, Riddle 1995), a signal of late Pleistocene divergence and diversification is evident in P. douelassi in all regions. This signal suggests that changes in the distributions of habitats during glacial/interglacial cycles of the late

Pleistocene influenced P. douelassi diversity. Lack of older levels of divergence and the widespread nature of very closely related haplotypes in the Wyoming Basin and Upper

Missouri Basin and Broken Lands suggest that habitats and P. douelassi populations were lost in these regions during glacial-interglacial cycles of the Pleistocene. Following the last glacial period, P. douelassi redispersed into suitable habitat throughout the region. Sequence distances suggest that closely related, but unique haplotypes have evolved in these two regions since the last glacial maximum. Alternatively, these haplotypes may have diverged in the late Pleistocene and become geographically structured due to lineage sorting (Avise

1994). Unfortunately, sequence distance estimates are not sensitive enough to relate lineage divergence to specific environment change events of the Pleistocene.

Phylogeographic Patterns

The phylogeographic pattern of P. douelassi represents Avise's (1989) Category I phylogeographic pattern, where haplotype divergence is great across the species, and clades are geographically structured. This phylogeographic pattern has been demonstrated for many different taxa (Avise et al. 1987,1989) and implies long-term, barriers to gene flow such that different regions occur along different clades on the intraspecific phylogeny (Avise 1989).

However, at the regional scale, P. douelassi demonstrated other phylogeographic patterns.

The Wyoming Basin and Upper Missouri Basin and Broken Lands regions exhibited an 64

Avise (1987) Category El phylogeographic pattern, where haplotypes are geographically

structured but closely related (e.g, shallow depth). This type of phylogeographic pattern at

the regional scale is common within taxa that exhibit a Category I phylogeographic pattern at

the broadest scale (Avise et al. 1987). Avise (1987) has postulated that geographic

structuring among closely related haplotypes results from limited gene flow between

populations not subdivided by long-term zoogeographic barriers. However, reduced gene

flow may result in new mutations being confined to subsets of the species' range within the

region (Avise et al. 1987).

The southeastern Arizona population of P. douelassi may represent an example of an

Avise (1987) Category E phylogeographic pattern. Haplotypes in this area exhibit large

differences and every lizard collected in the area had a different haplotype (based on analysis

of sequence data). Two samples (haplotype #'s 39 and 41) from the Chiricahua Mountains

were 6.9% divergent. Generally, nucleotide sequence distances in mtDNA between

conspecific individuals rarely exceed 1-2% (Avise et al. 1987). Moreover, no geographic

structuring among distantly related haplotypes was evident in the southeastern Arizona

population. Avise (1987,1989) provides no explanation for this type of phylogeographic

pattern, although an attempt to explain this pattern is given below.

The relative number and depths of divergence in populations should reflect historical

change in population size (i.e., effective population size, Avise 1994). Effective population

sizes are largely a function of the persistence of habitat (Gilpin 1987, Bowers and Dooley

1991, Pulliam et al. 1992). Areas that maintain large tracts of suitable habitat over time

should exhibit deep and shaUow phylogenetic divergence (Avise et al. 1987,1989). In areas where suitable habitats are lost, or in areas where habitat size is reduced such that genetic 65 bottlenecks are created, the phylogenetic history is lost or reduced (Avise 1994). This model assumes that populations and their genetic variability are lost when suitable habitat is lost.

The model also assumes that, on a regional scale, factors other than habitat loss, including competition and disease, have little impact on population persistence. The model does not assume a static pattern of habitat and populations within a region; certain levels of habitat tracking and population movement are expected. However, divergence depth and richness should indicate the relative history of habitat persistence over the region. The shallow phylogenetic depth of the Wyoming Basin and Upper Missouri Basin and Broken Land regions exemplify areas with ephemeral habitat.

Five out of seven physiographic regions, including the Columbia Basin, Colorado

Plateau, Lower Basin and Range (especially southeastern Arizona), southern Rocky

Mountain, and Upper Basin and Range physiographic regions had deep (pre-Pleistocene) phylogenetic divergence; the first three regions have three or more scales of divergence estimated to have preceded the Pleistocene (see Figure 7). These phylogeographic patterns suggest widespread persistence of habitat across the species' range, except for the northern interior regions. These findings reject the general model of mass shifts of populations to southern regions in response to glacial cycles (Graham 1986, Van Devender 1986, Betancourt et al. 1990). However, divergence richness and depth generally increases from north to south, excluding the Columbia Basin. Long-term persistence of habitat in the Columbia

Basin, despite its northerly geographic position, likely reflects the region's proximity to the

Pacific Ocean (e.g., climate moderating effect).

Haplotype diversity may reflect a combination of persistence of suitable habitat and moderate levels of habitat change within the region. Persistence of large habitat blocks 66

allows for accumulation of large populations and genetic variability (Lande and

Barrowclough 1987), and moderate changes in habitat quality and distribution can increase

diversity and the rate of evolution (Smith 1989, Avise 1994). Habitats that undergo frequent

and severe disturbance often have lower species diversity than those that experience moderate

to low levels of disturbance (Slobodkin and Sanders 1969). Southeastern Arizona may

exemplify this process. P. douelassi haplotypes are relatively divergent and diverse in the

area, and there appears to be no spatial structuring (e.g., Avise 1987 Category n

phylogeographic type). Moreover, the area has tremendous topographical relief, ranging

from 600 to 3700 m (USGS 1987b) and lies within the southern region of the U.S. where

glaciation may have had less impact on suitable habitat than in more northerly regions.

Finally, the area possesses a number of woodland, forest, and grassland types that are suitable

habitat for P. douelassi. These conditions make the area extremely resilient to

environmental change - the area as a whole maintains large blocks of suitable habitats under most climate change scenarios, but undergoes shifts in habitat mosaics across the regional landscape. The diversity of habitats and elevations may result in uneven gene flow among populations and foster diversification. Habitat mosaic changes during climate change may foster further isolation or mixing. Further research is needed to understand mechanisms that account for a pattern of multi-scaled haplotype divergence with little or no spatial structure.

The phylogeography of P. douelassi suggests at least two general modes of divergence: a deep vicariance mode and a dispersal and divergence mode (Avise et al. 1987,

1994). Vicariance involves fragmentation of wide ranging species into isolated populations, with subsequent divergence of isolated populations (Bush 1975, Brooks and McLennan

1991). Dispersal and divergence with gene flow involves dispersal of a taxon into an area, 67 followed by divergence associated with reduced or disrupted gene flow (Bush 1975, Avise

1994). The former mode generally involves a strong extrinsic barrier to gene flow (Bush

1994), whereas the later mode generally involves weak intrinsic barriers that reduce gene flow (Bush 1975, Avise 1994).

Large geographic separation of the P. douelassi clade representing northeast

Colorado, southern New Mexico, and western Utah and Nevada may represent regional-scale vicariance associated with uplift of the Rocky Mountains and Colorado Plateau, and development of the Rio Grande Depression. Alternatively, the Wyoming Basin and Upper

Missouri Basin and Broken Lands represent a divergence model involving dispersal and divergence with gene flow. P. douelassi populations within these two regions represent a recent (<10,000 years ago), large-scale dispersal event. Certain populations have diverged since the last glacial maximum and maintain geographic structure despite a relatively continuous distribution of habitat (i.e., the population in northeastern Wyoming). Non- random haplotype distribution within relatively continuous habitats may reflect non-random, geographic sorting of haplotypes, and subsequent restriction of gene flow (Avise 1994).

Populations in southern Utah also exhibit a high diversity of recently evolved haplotypes, but they apparently lack geographic structure. As mentioned early, P. douelassi occupies patchy forest habitats in this region. Patchy habitats can reduce effective population sizes which reduces diversity within populations. However, if a large number of patches remain, then each patch will "fix" different combinations of ancestral haplotypes and thus the overall regional population will maintain high diversity (Bush 1975, Brooks and MacLennan

1991, Lacy and Lindenmayer 1995). This process operating in the past might account for relatively high haplotype diversity within relatively large, continuous populations. 68

Additionally, habitat mosaics representing different successional stages shift frequently in forest ecosystems (Romme and Knight 1982, Hall et al. 1991); this may result in mixing of haplotypes and account for lack of geographic structure.

Estimating Times of Divergence and Concordance with Environmental Change

Estimates of approximate time of divergence and comparison of these estimates to large-scale environmental change require estimation of the rate of evolution in the taxa (e.g., a molecular clock, Wilson et al. 1985, Avise 1994, Riddle 1995). The overall rate of mtDNA sequence divergence has been estimated at 1-2% per million years (Wilson et al. 1985, Avise

1994), although there may be considerable variation among taxa with different metabolic rates, body sizes, and reproductive fecundity (Martin et al. 1992). Generally, ectothermic taxa are thought to have slower rates of mtDNA evolution (Martin et al. 1992).

The degree of precision required in relating sequence divergence to historical events depends on the question, since some questions require only approximate estimates of divergence (Riddle et al. 1993, Riddle 1995). For example, small mammals diverged prior to a late Pleistocene timeframe. Divergence estimates were sufficiently deep that either a 1 or

2% divergence estimate per million years would have put the divergence prior to the

Pleistocene. Similarly, questions addressed in this study require only approximate estimates of sequence evolution rates.

Additionally, transition substitutions are more common than transversion substitutions and this can cause non-linear estimates in rates of evolution due to saturation

(Simon et al. 1994). However, transversion rates of substitution in some mammals are approximately linear with time divergences of < 80 million years (Irwin et al. 1991). In this 69

study, transitions were down-weighted to reduce the influence of transition saturation on

divergence estimates.

Fossil records can be used to calibrate relative times of divergence. Although the

fossil record for P. douelassi is incomplete, a record from the Split Rock Formation of

Wyoming confirms the presence of P. douelassi in northern latitudes of the western United

States by the mid-Miocene (Robinson and Van Devender 1973).

Distribution and Ecology

P. douelassi has one of the largest ranges of any reptile in western North America

and is found in a variety of habitats (Stebbins 1985, Monanucci 1987). Unlike published

accounts (Pianka and Parker 1975), P. douelassi was found to be abundant. Data collected in

this study confirm a widespread distribution in a number of habitats. However, P. douelassi

populations appear to be discontinuous at different scales throughout many portions of its

range. P. douelassi observed in this study within prairie and grassland communities, especially those of eastern Montana, northeastern Wyoming, and western Colorado,

Nebraska, South Dakota, and North Dakota, were restricted to broken and badlands sites associated with river drainages and specific geologic features (Jones, personal observations).

Broken and badlands sites are largely discontinuous within this region, except for those following major river courses (e.g., the Yellowstone River in eastern Montana, USGS

1987a). Extensive searches for P. douelassi in continuous grassland and prairie in the eastern portion of the species' range yielded no lizards. These data suggest that P. douelassi populations may be discontinuous in the northeastern portion of its range, and that extensive grasslands are barriers to gene flow. Sequence divergence between haplotypes of the Pawnee 70

National Grassland site in northeastern Colorado and Fossil Agate site in western Nebraska

support this conclusion. These sites are 7.9% divergent and share no closely related

haplotypes, yet they are only 175 km apart. The Pawnee site is most closely related to

Southern Rocky Mountain populations in New Mexico and the Fossil Agate site shares a

haplotype broadly distributed throughout the Wyoming Basin and Upper Missouri Basin and

Broken Lands regions. Drainages at the Pawnee site emanate from the Southern Rocky

Mountain Province, whereas drainages at the Fossil Agate site emanate from the Wyoming

Basin (USGS 1987a). Apparently, P. douelassi dispersed along broken lands associated with

river drainages into the western Great Plains, but from different sources. It is unknown

whether Pawnee haplotypes diverged in northeast Colorado (indicating long-term habitat

persistence), or whether the haplotypes present represent a dispersal from the Southern Rocky

Mountains (no samples were obtained from the Denver, Colorado area).

Discontinuous distribution of P. douelassi is also evident at finer scales (e.g., forest

stand, personal observations). For example, P. douelassi is found in early successional stage

sites within forests, and along edges between early and late successional stages. However, P.

douelassi is not found within closed forest canopies (personal observations). Because most

forests are comprised of landscapes with mixed successional stages (Romme and Knight

1982, Hall et al. 1991), it follows that P. douelassi populations would be patchy. Patchy

distributions may lead to divergence among populations that appear to be relatively continuous at the broader scale (see later discussion).

The association of P. douelassi with disturbed sites (bluffs, badlands, early successional stage forests, open shrubland or woodland), may correspond to prey abundance and suitable escape habitat (Montanucci 1981). Disturbed sites have high abundances of 71 annual grasses and forbs, which tend to result in high abundances of ants (DeMers 1993,

Whitford 1994), a primary food item of P. douelassi. Disturbed sites also tend to have bare ground, rock, or loose soil on the surface which may provide burrowing, thermal, and escape substrate (Whiting et al. 1993). Sites with closed forest canopies or grass cover generally lack these characteristics. Great abundance of P. douelassi along lightly traveled dirt roads and trails within extensive prairies support this general conclusion; roads and trails mimic natural disturbance within these vegetation communities.

Relatively discontinuous distribution of P. douelassi at different scales due to habitat heterogeneity may, in part, account for phylogeographic patterns observed in the species (see previous discussion).

Phylogenetic Relationships

High levels of geographic variation in morphology have been recognized in this species and these have been used to delineate a number of subspecies (Reeve 1952).

However, no phylogenetic hypothesis has been proposed for the subspecies or races within

P. douelassi. Phylogenetic analysis of mtDNA sequences generally does not support subspecific delineations of Reeve (1952), nor phylogenetic relationships between P. ditmarsi and P. douelassi (Montanucci 1987). However, the Columbia Basin clade (see discussion below) closely overlaps the distribution of the subspecies P. douelassi douelassi (Reeve

1952).

The phylogeny generated in this study places P. ditmarsi within the P. douelassi lineage. This contradicts the existing morphology-based phylogeny for the genus

Phrvnosoma in which P. douelassi and P. ditmarsi are not closely related (Montanucci 1987). 72

However, recent work based on morphology and molecular data (Reeder et al., unpublished

manuscript) indicates that P. ditmarsi and P. douelassi are closely related taxa (Figure 12).

P. ditmarsi is restricted to montane woodlands and grassland in central Mexico where it is

roughly sympatric with P. douelassi (Reeve 1952). These two taxa and P. orbiculare form a

distinct woodland/montane grassland clade within the overall Phrvnosoma phylogeny

(Reeder et al., unpublished manuscript, Figures 11 and 12).

Although P. ditmarsi appears to be paraphyletic to P. douelassi. the mtDNA

phylogeny produced in this study represents only one gene lineage and does not necessarily

represent the overall organismal phylogenetic relationship between P. douelassi and P.

ditmarsi (Avise 1994). However, the phylogeny generated in this study represents the best estimate of phylogenetic relationships among these homed lizards from an mtDNA lineage standpoint.

Summary and Conclusions

The mtDNA phylogeny of P. douelassi suggests that the species has had a long and complex history of lineage response to historical environmental change. Multiple-levels of mtDNA sequence divergence suggest that the species diversified in response to both deep and recent environmental change, but deep history appears to have played a key role in structuring

P. douelassi lineages within regions. Ancient levels of divergence were associated with broad regional boundaries, with relatively recent levels of divergence geographically structured within regions. Significant phylogenetic gaps between regions suggest persistence of a barrier to gene flow among regions. This pattern constitutes a Category I phylogeographic patten described by Avise (1987). Deep splits in the P. douelassi phylogeny 73 appear to be temporally concordant with changes in regional climates associated with the late

Tertiary uplift of the western North American cordillera.

Pleistocene glacial-interglacial cycles also appear to have influenced diversification in

P. douelassi. Shallow levels of sequence divergence are evident in all regions and these are temporally concordant with environmental change of Pleistocene glacial-interglacial cycles.

Relatively deep levels of haplotype divergence within and large phylogenetic gaps among regions suggest that habitats and populations persisted in 5 of the 7 Kuchler

Physiographic regions during the last glacial maximum. Phylogeographic analysis of other woodland species are needed to evaluate whether the regional refugia model applies to other woodland species.

Shallow phylogenetic depth of P. douelassi populations in the Wyoming and Upper

Missouri Basin regions suggests that the species expanded into these areas subsequent to the last glacial period. The fossil record from the Split Rock Formation from approximately 15 million years ago suggests that the species has a deep history in the area. Apparently, the phylogenetic history of P. douelassi in these two regions was lost with the extirpation or bottlenecking of populations during Pleistocene glacial periods. CHAPTER 3

PHYLOGEOGRAPHY OF PHRYNOSOMA PLATYRHINOS

Introduction

Historical environmental change in western North America has had a dramatic impact on the evolution of desert lineages since at least the Miocene (Axelrod 1958,

1979,1985, Van Devender 1986, Lamb et al. 1989,1992, Betancourt et al. 1990, Riddle and Honeycutt 1990, Riddle et al. 1993, Riddle 1995). Diversification of desert taxa has been postulated to coincide with geologic events preceding the Pleistocene (Lamb et al.

1989, Riddle 1995) and glacial-interglacial cycles between the late Pliocene through the

Pleistocene (Riddle and Honeycutt 1990, Echelle and Dowling 1992, Lamb et al. 1992,

Riddle et al. 1993). Concordance between the depth of lineage divergence and historical environmental change within a region depend on (1) when the ancestor evolved within or dispersed to the region, (2) the degree of habitat fidelity of taxa within the lineage, (3) the extent and magnitude of historical habitat change, and (4) demographic, behavioral, and physiological characteristics of the taxa that influence effective population sizes (Avise

1994).

The desert homed lizard (Phrvnosoma platvrhinos) is a wide ranging species that may possess sufficient evolutionary depth to demonstrate both deep (pre-Pleistocene) and

74 75 more recent (Pleistocene) divergence in response to historical environmental change.

Although no fossil records are available for P. platvrhinos prior to the Pleistocene, a recent molecular and morphological study of the genus Phrvnosoma places the evolution of the desert lineage in the late Miocene (Reeder, unpublished manuscript). Presch

(1969) concluded that P. platvrhinos was derived from an ancestral P. coronatum stock, perhaps as early as the late Miocene. However, based on a phylogeny derived from skeletal morphology, Montanucci (1987) concluded that P. platvrhinos evolved as a result of Pleistocene glacial-interglacial cycles. Montanucci’s (1987) model of Pleistocene evolution of P. platvrhinos fits the general model of arid lineage divergence in response to Pleistocene glacial-interglacial cycles (Betancourt et al. 1990). A phylogenetic study of P. platvrhinos based on mtDNA molecular analysis could reveal whether this species evolved prior to or during the Pleistocene.

P. platvrhinos is ubiquitously distributed throughout the Mojave, Sonoran, and

Great Basin Deserts (Reeve 1952, Stebbins 1985). Populations of P. platvrhinos may have diverged in response to either late Tertiary events or glacially-induced isolation and contraction of deserts in the Pleistocene, but subsequently dispersed over large areas during interglacial periods. The result of this scenario might be multiple levels of mtDNA diversity with limited geographic structuring of haplotypes (e.g.,

Phylogeographic Category n, Avise 1987). The desert tortoise (Gopherus agassizi) exhibits levels of mtDNA divergence indicative of the Pliocene (Lamb et al. 1989).

Lamb et al. (1989) concluded that the relatively old divergence between tortoise populations east and west of the Colorado River resulted from the Bouse Embayment

(Smith 1970), a large marine incursion along the Colorado River that divided east and 76 west tortoise populations. Other phylogenetic studies of Mojave Desert reptiles, including the chuckwalla (Sauromalus obesus) and the desert iguana (Dipsosaurus doralis), have shown divergence patterns indicative of the Pleistocene (Lamb et al. 1992).

A molecular phylogeographic study could reveal the role of pre-Pleistocene and

Pleistocene environmental change on divergence patterns in P. platvrhinos.

During the Pleistocene, geographic distributions of desert taxa are thought to have shifted in response to mass geographic displacement of suitable desert habitats (Van

Devender 1986, Betancourt et al. 1990), although, recently, this hypotheses has been challenged (Riddle et al. 1993). Phylogenetic patterns in P. platvrhinos should reflect the historical stability of suitable habitats over the species' range. Areas that have not maintained suitable habitats, especially during the last glacial period, should be comprised of recently evolved haplotypes with little divergence depth. Conversely, areas with long-standing habitat stability should have multiple-levels of diversity.

The phylogeography of P. platvrhinos was analyzed using an mtDNA restriction site analysis to address the following hypotheses:

Hypothesis 1 - P. platvrhinos exhibits a deep phylogeographic break between

Mojave/Great Basin and Sonoran populations attributable to isolation and

divergence during the Bouse Embayment of the Pliocene;

Hypothesis 2 - Distribution of P. platvrhinos haplotypes represent a combination

of broad-scale mixing and geographic localization attributable to contraction and

expansion of deserts from the late Pliocene through the late Pleistocene; 77

Hypothesis 3 - Populations in southern areas of the species' range have greater

depths of divergence and haplotype diversity than more northerly areas of the

species' range because of greater habitat persistence in southern areas from the late

Pliocene to the present.

Methods and Materials

Eighty-three Phrvnosoma platvrhinos were obtained from 34 sites throughout the species' range (except in northern Mexico) between August 1991 and August 1994

(Figure 13, Appendix I). Methods of obtaining and transporting P. platvrhinos were similar to those described for P. douelassi (Chapter 2), except that the primary habitats searched were Sonoran, Mojave, and Great Basin Desertscrub. As in P. douelassi. no specimens of P. platvrhinos were obtained from Mexico. Approximately 85% of P. platvrhinos's range is in the United States.

All mtDNA processing and analysis of P. platvrhinos were conducted in the

Molecular Systematics Laboratory, Room 205, White Hall, University of Nevada, Las

Vegas.

DNA Extraction

An extraction procedure identical to that used in P. douelassi (see Chapter 2) was used to extract P. platvrhinos mtDNA. 78

•7 12* °28 29 • 27 J 14*

Figure 13 Geographic distribution of sample sites for Phrvnosoma platyrhinos. Sample site numbers correspond to collection data given in Appendix I. 79

PCR Amplification

Similar to P. douelassi. the ND2-COI mtDNA region was amplified for restriction-site analysis. An amplification procedure identical to that used in P. douelassi was used to amplify P. platvrhinos mtDNA fragments (Chapter 2).

Restriction Enzyme Digestion of Amplified mtDNA Fragments

The restriction enzyme digestion procedure used for P. platvrhinos was identical to that described for P. douelassi (Chapter 2), except that 12 of the enzymes demonstrated variation in restriction site presence or absence. Enzymes used to assay 83 P. platvrhinos from 34 sites included: Avail. BstNI. Hae HI, Hindi. Hinfl. HinPI. Hpall. Kpnl, Mbol.

Mspl. Rsal and Taal

Data Analysis

The restriction-site binary data matrix was constructed and stored in MacClade

(Version 3, Maddison and Maddison 1992). The phylogeny of P. platvrhinos was estimated using the maximum parsimony approach of PAUP 3.1.1 (Swofford 1993). The heuristic search option in PAUP was used with 10 replicate searches and random addition of taxa. Before submitting the data matrix to parsimony analysis, redundant haplotypes and uninformative characters were eliminated. However, all samples were used in constructing haplotype distribution maps (see results section). A bootstrap analysis of

1000 replicates was performed in PAUP, each with a single stepwise addition, to evaluate relative levels of character support for internal nodes of the estimated P. platvrhinos phylogeny. 80

The restriction-site binary data matrix was converted into haplotype and enzyme files and were then submitted to distance, haplotype, and Monte Carlo Analysis in REAP

(Version 4.0, McElroy et al. 1992). Estimates of evolutionary distance among composite haplotypes were estimated as the number of nucleotide substitutions per site from Nei

(1987). Phylogenetic relationships, based on the distance matrix, were evaluated using a

Neighbor-joining algorithm (Saitou and Nei 1987).

Restriction-site haplotype diveresity and nucleotide divergence and diversity of P. platvrhinos were evaluated in several ways. First, a geographic map of haplotype distribution was compiled to provide a simplified representation of the spatial relationships among haplotypes. To evaluate haplotype and nucleotide divergence and diversity, haplotypes were assigned to seven general geographic regions: (1) the northern

Great Basin, (2) the southern Great Basin, (3) the eastern Great Basin, (4) the southeastern Mojave Desert, (5) the northwestern Mojave Desert, (6) the southern

Sonoran Desert, and (7) the remainder of the Sonoran Desert. Geographic delineations were based on a combination of desert boundaries and the phylogeny derived for P. platvrhinos. Nucleotide divergence among the seven areas was calculated from Nei

(1987) as implemented in REAP. Additionally, nucleotide and haplotype diversity were calculated within each region according to Nei (1987; equations 8.4, 8.5, and 8.12) as implemented in REAP to compare overall haplotype diversity and the relative depth of diversity among regions.

The extent of geographic heterogeneity in the frequency distributions of haplotypes was determined across the seven regions listed above by a Monte Carlo simulation (1000 replications) as implemented in REAP (Roff and Bentzen 1989). The 81

Monte Carlo simulation evaluates the probability of obtaining a X2 value from the

distance data matrix that is larger than that obtained from a series of randomizations of

that matrix. Finally, the genetic distance matrix and haplotype distributions were

analyzed with AMOVA (Excoffier et al. 1992) to determine the degree to which variation in the genetic distance matrix was apportioned within populations, among populations within regions, and among regions. For this analysis, sample sites were assigned into populations and regions as given in Table 8.

Results

Distribution and Ecology

Lizards were observed in a variety of desert habitats, including low elevation sagebrush communities within the northern, eastern, and southern Great Basin; blackbrush and creosote communities in the Mojave Desert; and creosote and Saguaro- palo verde communities in the Sonoran Desert. Additionally, P. platvrhinos was observed in transition zones between desertscrub and juniper woodland, primarily within the

Sonoran and Mojave Deserts. Within desert communities, P. platvrhinos was found on a variety of substrates, ranging from deep sand to volcanic (basalt) rock. Within the northern Great Basin and southern portion of the Columbia Basin, P. platvrhinos were found in areas dominated by volcanic formation (e.g., Black Rock Desert of Northwestern

Nevada; Fields, Oregon; Murphy, Idaho).

P. platvrhinos were very common on sites where they were found. On the Fields,

Oregon site, over 30 adult P. platvrhinos were observed in less than 0.5 hours; all of these animals were basking on basalt rock along a paved road. On the Round Mountain, 82

Table 8 Hierarchical arrangement of sample sites into populations and regions for an AMOVA analysis of P. platvrhinos restriction-site distance data. Sample site numbers correspond to those illustrated in Figure 13. Region Population No. Sample Sites Southern Sonoran 1 24

2 23

Western/Central Sonoran 3 21

4 20,22

Northwestern Mojave 5 15

Southern/Southwestern Mojave 6 1-4

7 5,25

8 17-19

Northern Great Basin 9 32-34

10 10,11

11 8,9

12 7,12

Southern Great Basin 13 13

14 14

Eastern Great Basin 15 6,28,30,31

16 26,27 83

Nevada site, over 40 adult P. platvrhinos were observed in less than 0.5 hours; similar to the Field, Oregon site, they were basking on basalt rock. Finally, on a site in Clark

County, Nevada, over 50 adult P. platvrhinos were observed in less than one hour. These animals were seen moving across a lightly-traveled dirt road. On all of these sites, P. platvrhinos was easily the most common lizard.

Lizards were most commonly observed along disturbed sites, including lightly to heavily traveled dirt roads, along shoulders of paved roads traversing desert habitats, underneath power lines, and at dump sites located outside of rural communities.

Common to all of these sites were rocks located along the edge of the disturbance. It is unknown whether populations are actually larger along these areas, or whether animals were simply easier to observe. However, all of these sites possessed large numbers of ant nests, as compared to areas within continuous desert scrub, and this might account for the large number of lizards observed on these sites.

Phylogenetic Relationships and Phylogeography

Restriction digests of the ND2-COI mtDNA fragment resulted in a total of 25 composite haplotypes (Table 9). Estimated sequence divergence between haplotypes ranged from 0.3 to 6.8% (Table 10). A Neighbor-joining tree based on sequence divergence revealed multiple scales of divergence (Figure 14). Haplotypes from the southern Sonoran Desert and a clade representing all other haplotypes formed the deepest split in the tree (5% divergence, Figure 14). The next deepest divergence (3%) involved a split of the large clade into a lineage consisting of one haplotype from the western

Sonoran with the remainder of the large clade (Figure 14). Three other splits representing 84

O O O i—i <—i o o o o o o o o o o o r-H i-H r-H i-H i-H i-H r-H 1—H i - H o1-H o 1-H o i-H o i-H o T—H o 1-Ho i-H o o o o o o o o o

o o o O O O O O O O O ' —1>-,00 © s i—H i- H 1 -H I-H i-H 1—( 1-H i-H r-H i-H i-H 1-H 1-H i-H i-H 1—H x>C/5 o o o ooooooooooo>—< o a 1 -H i- H f***! i - H t -H 1-H f*"*3 i-H 1-H i-H o y - * '-Hi—H a o o oooooooo o o o o o o oooooooooooo o I T-H i -H 1-H i-H i-H i-H i-H i-H i-H i-H i-H 1 -H o i - H 1 -H o o ooooooooooooo o i - H r H 1 -H 1- H i —H 1-H 1-H 1-H 1-H 1-H 1-H 1-H 1—H i - H i —« 1 -H XJ o o ooooooooooooo o

S' oooooooooooo1-H i-H I-H i-H i-H i-H r-H i-H O O i-“l "—*

ooooooooooooo o o o i-H i-H i-H i-H i-H i-H t-H i-H i-H i-H 1-H 1“ H i- H 1-H o1-H oooooooooooo; o o o i-H i-H i-H 1-H i-H 1-H i-H i-H T—H i-H i-H ^5 & i - H i - H O © T"HT “Hr -“,|,“Hr -, r “HT—lr "H a o

oo o 1-° H o£oo o

o oo o o o o oo o o o 1—H o o o x> oo oo o 8 S o a oo oo o o

0^0 1 -H h H

©©©©©©©©©©©©©© i-H i-H i-H 1-H i-H i-H i-H i-H i—H i-H 1-H 1-H i-H i-H i-h O'-|'-<0000000000

=#:

H 1 -H O J cnH- »nvor'00CT\O’—icscn-^-invo HI o oQOr^OOOQi-Hi-Hi-Hi-Hi-Hi-HT-H

Table 10 Sequence divergence estimates for P. platvrhinos.

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23

0 1.0 0 1.4 0.4 0 3.0 2.7 2.3 0 2.4 1.4 1.8 2.6 0 0.3 1.4 1.8 3.5 2.8 0 0.7 1.8 2.2 3.1 3.3 0.3 0 1.9 1.5 1.1 1.2 3.1 2.3 1.9 0 1.8 0.7 1.1 2.7 2.1 2.2 1.8 1.5 0 1.4 0.4 0.7 2.3 1.8 1.8 1.5 1.1 0.4 0 0.3 1.4 1.8 2.7 2.8 0.7 0.3 1.5 1.5 1.1 0 4.1 4.7 4.3 4.4 6.5 4.7 4.3 3.0 4.7 4.3 3.7 0 4.5 5.0 4.7 4.8 6.8 5.0 4.6 3.5 5.0 4.7 4.1 1.3 0 0.7 1.8 2.2 3.1 3.3 1.1 0.7 1.9 1.8 1.5 0.3 4.3 4.6 0 1.8 0.7 1.1 2.7 2.1 2.2 1.8 1.5 0.7 0.4 1.4 4.7 5.0 1.8 0 1.9 2.3 2.8 3.8 3.9 2.3 1.9 2.5 2.3 1.9 1.5 3.0 4.4 1.1 2.3 0 2.6 3.1 3.5 3.7 3.8 3.0 2.7 3.3 3.0 2.7 2.2 5.9 6.2 1.9 3.1 2.4 0 2.2 1.1 1.5 3.2 2.6 2.6 2.2 1.9 1.1 0.7 1.8 5.3 5.6 2.2 1.1 2.8 3.5 0 2.6 1.5 1.9 2.8 3.0 3.0 2.7 2.4 1.5 1.1 2.2 5.9 6.2 2.7 1.5 3.3 4.0 0.4 0 1.7 0.7 1.0 3.4 2.0 2.1 2.5 2.2 0.7 1.0 2.1 5.4 5.8 2.5 1.4 3.0 2.9 1.8 2.2 0 1.4 0.4 0.7 3.0 1.7 1.8 2.2 1.9 0.3 0.7 1.8 5.0 5.4 2.2 1.0 2.7 3.4 1.4 1.8 0.3 0 2.1 1.0 1.4 3.1 2.5 2.5 2.2 1.9 1.0 0.7 1.8 5.1 5.4 2.2 1.0 2.7 3.4 1.4 1.8 1.7 1.4 0 2.2 1.1 1.5 3.3 2.6 2.7 2.3 2.0 1.1 0.7 1.9 3.5 4.8 2.3 1.1 1.2 3.7 1.5 1.9 1.8 1.5 1.5 0 1.8 0.7 1.1 2.7 2.1 2.2 1.8 1.5 0.7 0.4 1.4 4.7 5.0 1.8 0.7 2.3 3.1 1.1 1.5 1.4 1.0 0.3 1.1 1.8 0.7 1.1 2.7 2.1 2.2 1.8 1.5 0.7 0.4 1.4 4.7 5.0 1.8 0.7 2.3 3.1 1.1 1.5 1.4 1.0 1.0 1.1 87

Clark Co., NV (1) Clark Co., NV (6) Clark Co., NV (7) Wikieup, AZ (11) Whitman, AZ (14) Clark Co., NV: Beaver Dam Slope, UT: Blythe, CA (2) Clark Co., NV (3) Clark Co., NV: Elko.NV: Western, UT (9) Round Mtn, NV (20) Western UT (21) , Large Range Haplo (10) Joshua Tree N.P., CA (15) Malheur Co., ID (25) Round Mtn, NV (22) Round Mtn, NV (24) Bishop, CA (18) Bishop, CA (19) Bouse, AZ (16) Round Mtn, NV (23) Clark Co., NV (4) Clark Co., NV (8) Clark Co., NV (5) Bouse, AZ (17) Ajo.AZ (12) Tucson, AZ (13)

5.0 4.0 3.0 2.0 1.0 % Sequence Divergence

Figure 14 Phylogenetic relationships of composite P. platvrhinos haplotypes estimated from a Neighbor-joining analysis of restriction-site distance data. Numbers in parentheses correspond to haplotypes given in Table 9 and Figure 16. 88 greater than 2% divergence each included haplotypes occurring in Clark County, Nevada

(Figure 14).

Maximum parsimony analysis supports the relatively deep split of the southern

Sonoran clade from other populations (Figure 15), and a relatively deep split of a western

Sonoran Desert haplotype as indicated in the distance analysis (Figure 15). However, unlike the distance analysis, it does not yield a deep split of Mojave Desert haplotypes

(numbers 4 and 8, Figure 15). Other than the deep branches representing the southern and western Sonoran clades, and the terminal nodes differentiating two haplotypes in each of the northwestern Mojave and southern Sonoran Deserts, no other branches were supported by bootstrap values of greater than 50% (Figure 15).

Divergence patterns of haplotypes within clades suggest that lizards with ancestral haplotypes have undergone a series of range expansions and contractions, especially within the Mojave and southern Great Basin desert scrub. For example, haplotypes from the western Sonoran Desert (Bouse, Arizona) and the Southern Great Basin (Round

Mountain, Nevada) form a relatively deep clade (2.4% divergence, Figure 14); these haplotypes occur only in these areas and the two sites are approximately 600 km apart.

The two haplotypes within this clade are 1.2% divergent (Figure 14). This pattern suggests that the ancestor of the two haplotypes was widespread, and subsequently, became fragmented. The distance tree also suggests multiple depths of divergence and considerable mixing of haplotypes in the mid-Mojave region (e.g. Clark County,

Nevada). Haplotypes within the mid-Mojave region represent divergences ranging from

0.3 to 2.5% (Figure 14). 89

1) 6 ) Mid-Mojave 7) } 11) Northern Sonoran 4) Sonoran •34 7) Western Sonoran 2) Mid-Mojave 5) } r-15 9) Eastern Great Basin/Mojave r 30 20) Eastern Great Basin I—40 21) Eastern Great Basin 10) Northern Great Basin, Mid-Mojave, S.W. Mojave __ 4 15) Mid-Mojave 22) ■57 J Southern Great Basin 24) 25) Northern Great Basin 3) 4) } Mid-Mojave 8)

r-58 8) | Northwest Mojave 9) 6) Western Sonoran 23) Southern Great Basin 2) .90 Southern Sonoran 3) }

Figure 15 Phylogenetic relationships of composite P. platvrhinos haplotypes estimated from a Maximum Parsimony analysis based on restriction-site data. Numbers on branches represent bootstrap values based on 1000 replications. Numbers in parentheses correspond to haplotype numbers given in Table 9 and Figure 16. 90

The southern Sonoran desert was consistently most divergent from all other

regions (3.09 - 4.55%, Table 11). The remainder of the Sonoran Desert had the next

deepest range of divergence levels (0.96 - 3.09), but these were on average 3 times less

divergent that the southern Sonoran area (Table 11). The northern Great Basin area

demonstrated the least amount of divergence from other areas, and was only 0.02 and

0.08% divergent from the southern Great Basin and Mojave areas (excluding the

northwestern Mojave). However, the northern Great Basin is 0.56% divergent from the

eastern Great Basin (Table 11), and the eastern Great Basin is most similar to the

southern Great Basin (Table 11). This suggests limited gene flow between the northern

and eastern Great Basin areas. A Monte Carlo simulation (1000 replications) of

haplotype heterogeneity by the seven geographic areas indicated in Table 11 revealed a

highly significant, non-random pattern (probability of exceeding the original X2 value was

0).

An AMOVA analysis of three scales of haplotype divergence (nested analysis of variability within populations, among populations, and among regions) revealed that among region variance explained the greatest portion of the total variance (63%). Among population and within population variance explained 6% and 31% of the overall variance, respectively. The probability of a random group of heterozygosity being > the observed eterozygosity (1000 reps) was 0.009.

Haplotype 10 was common across a large geographic area, ranging from the northern Great Basin to the southwestern Mojave Desert (Figure 16). However, it was absent from areas south and east of the Colorado River. Although haplotype 10 has a wide distribution and there is considerable mixing of haplotype distributions in the 91

Table 11 Nucleotide divergence of P. platvrhinos among and nucleotide diversity within (diagonal) regions based on restriction-site analysis. Regions were determined from a combination of desert boundaries and the phylogenetic tree generated from genetic divergence. NGB = Northern Great Basin, EGB = Eastern Great Basin, SGB = Southern Great Basin, MOJ = Mojave, NWM = Northwest Mohave, SON = Sonoran (other than Southern Sonoran), and SSON = Southern Sonoran NGB EGB SGB MOJ NWM SON SSON 0.001

0.56 0.004

0.02 0.58 0.004

0.08 0.34 0.11 0.010

0.79 1.35 0.81 0.88 0.002

1.19 1.68 1.15 0.96 2.11 0.013

4.08 4.55 3.99 3.76 5.25 3.09 0.005 92

!• 25

21/

I?.*

Figure 16 Geographic distribution of composite P. platvrhinos haplotypes. Haplotype numbers correspond to those given in Table 9. 93

Mojave Desert, geographic structuring of haplotype distribution is evident throughout most of the species’ range (Figure 16).

Nucleotide diversity among areas reveals a general pattern of increasing diversity from north to south, with the exception of the southern Sonoran Desert and the northwestern Mojave Desert (Table 11). The Sonoran and southern Mojave Deserts had the greatest nucleotide diversity (0.013 and 0.010, respectively) and the northern Great

Basin Desert the lowest diversity (0.001, Table 11). Low nucleotide diversity in the southern Sonoran and northwestern Mojave Deserts may reflect small sample sizes of these areas.

With the exception of the southern Sonoran Desert, haplotype diversity increased clinally from north to south (Table 12). The northern Great Basin was by far the least diverse (0.12, Table 12). It consisted of only two, closely related haplotypes (10 and 25), of which one represented all but one of the samples. The Mojave (0.845) and Sonoran

(0.800) Deserts were the most diverse (Table 12). The lack of diversity in the southern

Sonoran Desert may reflect a relatively small sample size (5 specimens from 2 localities).

Discussion

Phylogeographic Relationships

Phylogeographic patterns of P. platvrhinos suggest the species diversified in response to relatively deep and recent environmental change. Based on a 2% sequence divergence per million years (Wilson et al. 1985), the southern Sonoran populations of

P. platvrhinos diverged from other populations between 2.5 and 3.0 million years ago.

However, Martin et al. (1992) suggest that sequence divergence may be 5-10 times slower 94

Table 12 mtDNA diversity of P. platvrhinos among regions based on haplotypes generated from restriction-site analysis in REAP. I .11 . I L II...... -■ --- -...... Northern Great Basin Desert 0.12 +/- 0.10

Eastern Great Basin Desert 0.64 +/- 0.10

Southern Great Basin Desert 0.64 +/- 0.15

Mojave Desert 0.84 +/- 0.05

Northwestern Mojave Desert 0.67 +/- 0.02

Northern and Central Sonoran Desert 0.80 +/- 0.17

Southern Sonoran Desert 0.40 +/- 0.24 95

in ectotherms than endotherms. The estimated divergence rate of 2% per million years is

based on mammal studies (Wilson et al. 1985). Therefore, the southern Sonoran

population of P. platvrhinos may have diverged considerably earlier than the beginning of

the Pleistocene. The divergence depth and pattern east and west of the Colorado River

are nearly identical to those observed in the desert tortoise (Gooherus aeassizi. Lamb et

al. 1989). Lamb et al. (1989) concluded that the east-west divergence in the desert

tortoise was associated with the Bouse Embayment of the Pliocene. The formation of the

embayment, running from the Sea of Cortez along the Colorado to the present location of

Mojave Lake, dissected the tortoise range into an east and west half (Lamb et al. 1989).

Similarly, the Bouse Embayment may have fragmented P. platvrhinos into an east and west half, resulting in divergence levels indicated in the mtDNA analysis.

Although parsimony and distance analyses yielded different relationships among populations below the relatively deep nodes, both suggest a mode of population dispersal and contraction during the Pleistocene. Both analyses indicated a clade consisting of populations from the southern Great Basin and western Sonoran, yet these two areas are separated by nearly 600 km. Moreover, this observation, in addition to observations of relatively high haplotype and nucleotide diversity in the southern Great Basin, suggest that P. platvrhinos populations persisted in relatively northern areas throughout the

Pleistocene. Alternatively, haplotypes occurring in the southern Great Basin could represent more recent (e.g. post-Pleistocene) migration of relatively deep and shallow haplotypes into the region. However, this pattern would have required the extinction of the deeper haplotype from the approximately 600 km area between the western Sonoran and southern Great Basin populations. Therefore, this scenario is unlikely. 96

Although phylogeographic patterns suggest that populations and habitat persisted across a large portion of the species range, haplotype and nucleotide diversity generally increase from north to south (if the Sonoran Desert is considered as one and not two regions). This north-south gradient may reflect overall effective population sizes within a given region. Because of persistence of relatively large desert areas at low elevations in western Arizona, southern Nevada, and southeastern California (Cole 1986, Lomolino et al. 1989, Betancourt et al. 1990), large populations of P. platvrhinos may have existed during glacial periods in the southern range of the species. Greater population size plus partial habitat isolation within the larger habitat area during glacial periods might account for greater diversity and depth of divergence in the south. Partial isolation of the larger habitat during glaciation was likely due to the region’s varied topographical relief (USGS

1987b).

Persistence of populations in the Mojave Desert and southern Great Basin during glacial periods may reflect the distribution of volcanic formations within these two regions. The current distribution of P. platvrhinos in the northern Great Basin and

Columbia Basin suggest that populations can persist in cold areas where basalt rock is present. The southern-most Great Basin population sampled in this study occurred on a volcanic formation located in the lowest basin within the area. Volcanic formations may have acted as refugia during glacial periods and allowed populations to persist within the current Mojave Desert and southern Great Basin areas that were occupied by pinyon- juniper woodlands. A fossil record of P. platvrhinos associated with a juniper woodland

(approximately 13,000 years BP) supports this general conclusion (Van Devender et al.

1991). 97

High haplotype and nucleotide diversity in the mid-Mojave Desert (e.g., Clark

County, Nevada) may reflect the area’s relative position between large, persistent habitat to the south, and pinyon-juniper woodlands to the north. If the reconstructed maps of

Lomolino et al. (1989) and Betancourt et al. (1990) are approximately correct, then the southern Mojave Region came into contact with the larger desert habitat to the south and woodland to the north during glacial periods. This scenario may have lead to a series of peripheral isolations of populations representative of the southern region. Peripheral isolation of populations and subsequent divergence is thought to be a primary mode of speciation (Bush 1975). This scenario may have been repeated over a number of glacial periods within the Pleistocene, resulting in the great richness of haplotype diversity and divergence observed within the area today.

Haplotype diversity and divergence within the northern Great Basin are appreciably lower than elsewhere, indicating that this area did not maintain P. platvrhinos populations or suitable habitat during the last glacial period. Only two haplotypes exist over the entire northern one-third of the species’ range, and one of them (southeastern

Idaho) differs by only one restriction-site. A similar pattern of post-glacial northern expansion of ranges has been observed in lodgepole pine in western North America

(Cwynar and MacDonald 1987) and in the grasshopper species (Chorthippus parallelus) in northern Europe (Cooper et al. 1995). Reduced haplotype diversity may reflect repeated bottlenecking of pioneer populations within northern regions (Hewitt 1993), and lack of time necessary to diversify (e.g., over the past 10,000 years).

Nucleotide distances and patterns in haplotype distributions suggest limited gene flow between the eastern and northern Great Basin. Haplotype divergence between the 98 two regions is 0.6% and populations within each region appear to emanate from different sources. Haplotypes in the eastern Great Basin are closely associated with eastern

Mojave Desert haplotypes whereas haplotypes in the northern Great Basin are closely associated with the southern Great Basin and western Mojave Desert haplotypes.

Mountainous areas separate the eastern and western Great Basin and these may act as barriers to gene flow between the two regions.

Absence of haplotype 10 in the Sonoran Desert suggests that the Colorado River has been a significant barrier to gene flow between west and east populations during at least the current interglacial period. However, relatively shallow divergence between haplotypes in the northern Sonoran desert and Mojave Desert northwest of the Colorado

River suggest that the river was not a consistent barrier to gene flow between the two deserts in the Pleistocene.

Finally, although there are some regionally-distributed haplotypes, including a recently evolved haplotype that extends over 80% of the species north-south range

(haplotype 10), P. platvrhinos haplotypes generally exhibit geographic structuring. This suggests that the origin and spatial structuring of P. platvrhinos haplotypes within geographic areas were constrained by the broader and deeper history delineating lineages among areas (Avise 1994, Riddle 1995).

Distribution and Ecology

Contrary to published accounts of low population density (Pianka and Parker

1975), P. platvrhinos was found to be extremely abundant throughout its range. In fact, in 99 most areas of the Mojave and Great Basin Deserts, it appeared to be the most abundant lizard.

Great abundance of these lizards on disturbed sites may be a function of prey availability and abundance, and the presence of loose soil and rock. Ant nest density appeared to be greatest at or adjacent to disturbed sites, and these sites were often lined with rock or dirt embankments (Jones, unpublished data). Rock substrate may provide a thermal environment necessary to digest ants (Whitford, personal communications) and to maintain preferred body temperatures (Pianka and Parker 1975). Moreover, dirt embankments may provide ideal nesting substrate.

Apparently, moderate frequency of road maintenance (e.g. shoulders of dirt roads) and use in these areas maintains a relatively continuous state of disturbance. Most desert environments, by nature, have patches of surface disturbance (e.g., along ephemeral washes and arroyos); dirt roads appear to mimic natural disturbance. However, a high incidence of P. platvrhinos observations along disturbed areas, as compared with less disturbed habitat areas, may reflect differences in detectability of the lizard between the two types of habitats rather than real differences in abundance.

The extension of the species’ range in the northern Great Basin may reflect the distribution of volcanic geologic formations to the north. P. platvrhinos was found almost exclusively in areas with volcanic rock. These areas may provide thermal conditions necessary for the species to survive in colder, northern regions. 100

Summary and Conclusions

Phylogeographic patterns of P. platvrhinos suggest the species diversified in

response to relatively deep and recent environmental change. The Bouse Embayment

appears to have divided the geographic range of P. platvrhinos into populations east and

west of the Lower Colorado River between the Pliocene to late Pleistocene; this resulted

in divergence between the two regions. This deep divergence pattern is similar to that

observed in the desert tortoise. Further differentiation of P. platvrhinos within regions is

roughly congruent with Pleistocene glacial-interglacial cycles.

Phylogeographic patterns suggest that habitats and populations persisted in what

is today the southern Great Basin, Mojave Desert, and Sonoran Desert during the last

glacial maximum. Persistence of populations in the southern Great Basin and the Mojave

Desert may reflect the presence of abiotic factors, including basalt rock formations, which

may have moderated the impact of glaciation. Northern Great Basin populations of P.

platvrhinos appear to represent a recent expansion of the species northward during the current interglacial period. CHAPTER 4

DISTRIBUTION OF PHRYNOSOMA DOUGLASSI AND

PHRYNOSOMA PLATYRHINOS HABITAT

DURING THE LAST GLACIAL MAXIMUM

AND PRESENT

Introduction

The distribution, pattern (size and connectivity), and persistence of suitable

habitat influence effective population sizes (Gilpin 1987, Bowers and Dooley 1991,

Saunders et al. 1991, Verboom et al. 1991, Danielson 1992, Pulliam et al. 1992, Taylor et

al. 1993, Thomas 1994, McIntyre 1995). Effective population sizes in turn influence the probability of local extinction and the maintanence of genetic diversity over time

(Allendorf and Leary 1986, Lande and Barrowclough 1987, Lacy and Lindenmayer 1995).

Therefore, persistence in the size and connectivity of suitable habitats within a region can have a dramatic effect on lineage diversity and persistence (Avise 1994).

The distribution and connectivity of suitable habitats may be as important in determining effective population size and persistence as the total area of suitable habitat within a region (Taylor et al. 1993). Connectivity influences the effective habitat size available to a species (Lande and Barrowclough 1987), and can influence the extinction

101 102 probability of a local population (Wilcove et al. 1986, Gilpin 1987). Habitat connectivity may be especially important in animals with relatively weak dispersal capability and high habitat fidelity. Therefore, an assessment of habitat persistence within a region should include metrics of connectivity as well the mean and variance in habitat patch size.

Climate change associated with glacial-interglacial cycles within the late Pliocene through the Pleistocene had a dramatic effect on the spatial distribution of major biomes in the western United States (Van Devender et al. 1986, Lomolino et al. 1989, Betancourt et al. 1990). During glacial periods, most regions were dominated by woodlands, forests, and alpine tundra (Betancourt et al. 1990). In western North America at the height of the last glacial period (approximately 18,000 years ago), a large glacier occupied areas of

Montana and the Dakotas (Graham 1986). Desert biomes were confined to low elevations in the extreme southwestern United States (Cole 1986, Lomolino et al. 1989,

Betancourt et al. 1990). During interglacial periods, desert biomes were presumably widespread across the southwestern United States (Betancourt et al. 1990). Similar to today, woodlands were widespread at elevations of 1220-2300 m (Lomolino et al. 1989,

Betancourt et al. 1990). Montane and boreal forests existed on isolated mountain ranges of the southwestern United States and at higher elevations in the Rocky Mountains,

Intermountain regions of Wyoming, Utah and Nevada, and in the Sierra and Cascade ranges (Thompson 1990).

Numerous studies, especially those involving packrat middens and pollen records, have estimated changes in the elevation of vegetation zones during glacial periods (Cole

1985, Betancourt 1990, Spaulding 1990, Thompson 1990, Van Devender 1990).

Vegetation zones are loosely defined and they roughly correspond to biome-level 103 classification. During the last glacial period, vegetation zones were depressed 1000 to

1300 m below current elevations in the Great Basin (Thompson 1990), 900 to 1300 m in the Mojave Desert (Spaulding 1990), 500 to 1200 m on the Colorado Plateau (Betancourt

1990), including 800 to 1200 m in the Grand Canyon (Cole 1986), and 800 to 1200 m in the Sonoran Desert (Van Devender 1990). Finer-scale changes in plant community composition occurred in response to glacial-interglacial cycles that were not predictably associated with elevation (Betancourt et al. 1990).

Considerable heterogeneity in topographical relief across the western U.S. played a large role in structuring the response of plant communities and organisms to glaciation

(Lomolino et al. 1989, Betancourt et al. 1990). Lomolino et al. (1989) and Betancourt et al. (1990) estimated the distribution of biomes in the southwestern U.S. during the last glacial maximum from a combination of packrat midden, pollen, and topography data.

Lomolino et al. (1989) applied their model to the existing topography of the southwestern

U.S.

Spatial variation in genetic divergence and diversity in Phrvnosoma douglassi and

Phrvnosoma platvrhinos (Chapters 2 and 3, respectively) suggest that populations of both species were influenced by shifts in the distribution of suitable habitats associated with glacial-interglacial cycles. The deep levels of divergence and high haplotype diversity observed in southern portions of both species ranges are consistent with hypotheses of long-term persistence of large populations during the last glacial maximum. Conversely, shallower levels of divergence and low haplotype diversity observed in the northern Great

Basin in P. platvrhinos and the Wyoming Basin and Upper Missouri Basin and Broken 104

Lands regions in P. douglassi are consistent with the hypothesis that population sizes

were small during the last glacial maximum.

P. douglassi occupies a range of habitats across much of the western United States

(Stebbins 1985). However, P. douglassi is largely confined to regions occupied by

woodlands, forests, and grasslands in the southern and central portions of its range. Only

at relatively high elevation within the Wyoming and Columbia Basin does the species

commonly inhabit deserts. Conversely, P. platvrhinos is widely distributed in the Great

Basin, Mojave, and Sonoran Deserts (Stebbins 1985). It also inhabits transition zones

between deserts and woodlands and deserts and grassland (Jones 1988). Patterns of

habitat fidelity suggest that models of habitat distribution, size, and connectivity could be

employed to assist in understanding the north-south trends in increasing mtDNA lineage

divergence and haplotype diversity demonstrated in this study.

Habitat conditions were simulated across the western United States during the last

glacial maximum by developing a model that incorporated historical data on habitat-

elevational gradients. This model was then applied to a comprehensive elevational

database covering the entire western United States. Existing data on the distribution of P.

douglassi and P. platvrhinos were also used to develop the habitat models. The primary objective of the simulation was to test the hypothesis that divergence and diversity patterns in P. douglassi and P. platvrhinos correspond to a specific model of habitat transformation from the height of the last glacial maximum to modem times. 105

Methods and Materials

Current and historic (last glacial maximum) distribution and pattern of suitable P. douglassi and P. platvrhinos habitats were simulated in the Grid program within the

Arc/Info Geographic Information System (Version 7.0.2, ESRI1994). Models of suitable habitat were constructed for both species from existing literature and observations made during field trips. The models included a range of elevations inhabited by the species within specified geographic regions. Models were constructed for different regions because of great variation in habitat use and elevation associations in both species.

Regions, based on Kuchler (1964), for P. douglassi models included: (1) the Columbia

Basin, (2) the Upper Basin and Range, (3) the Colorado Plateau, (4) the Lower Basin and

Range, (5) the Upper Missouri and Wyoming Basins, and (6) the Southern Rocky

Mountain and Piedmont region. Regions for P. platvrhinos models included: (1) the

Lower Basin and Range, (2) the Upper Basin and Range, and (3) the Colorado Plateau.

These regional boundaries are similar to those used in the phylogeographic analyses of the two species (Chapters 2 and 3).

Digital Kuchler coverages were clipped out for each of the six regions in

Arc/Grid. For each region and species, Kuchler vegetation classes were combined into two classes: suitable and unsuitable. Suitable Kuchler types are those that support P. douglassi and P. platvrhinos populations, whereas unsuitable Kuchler types are those that lack populations. Existing literature and field data collected in this study were used to determine which Kuchler types were suitable and unsuitable. Regional coverages of the

1:2,500,000 scale Digital Elevation Model (DEM) map (USGS 1987b) were also clipped out and merged with the regional Kuchler map. From the combined grid, elevation 106

ranges of suitable and unsuitable habitat were calculated. This approach produced a

model of suitable/unsuitable habitat constructed for each species based on elevation (see

Tables 13 and 14). To apply the model to the DEM map, it was necessary to make

elevational values for suitable and unsuitable habitats mutually exclusive. The

elevational model was then applied to the DEM map to illustrate the distribution of

present-day suitable habitat.

Data from existing studies (Cole 1986, Lomolino et al. 1989, Betancourt 1990,

Betancourt et al. 1990, Spaulding 1990, Thompson 1990) were used to determine

elevational shifts in suitable habitats during the last glacial maximum. Tables 13 and 14

summarize elevational ranges used in the models to simulate P. douglassi and P.

platvrhinos habitat distribution during the last glacial maximum, respectively. These

models were then applied to the DEM map in each of the six regions to determine the

distribution of suitable habitats during the last glacial maximum. To illustrate conditions

of suitable habitat for each species across their entire range, the six regions were merged into composite coverages.

Because patch size and connectivity affect habitat suitability, averages of and ranges in patch size and connectivity were calculated for each simulation. Formulas for patch size (Equation PSIZ) and connectivity (Equation IEDG) are given in Riitters et al.

(1995) as implemented in the LANDSTAT statistical program (Kepner et al. 1995). The minimum resolution (cell size) of habitat was 100 hectares.

Finally, results of habitat simulations were compared and contrasted with divergence and diversity within P. platvrhinos and P. douglassi (Chapters 2 and 3) to 107

Table 13 Habitat models used to estimate current and historic (18,000 years ago) suitable habitat of P. douglassi. The models are based on a range of elevations in which the species is found. Elevations are given in meters above sea level. Region Current 18,000 years ago Lower Basin and Range Suitable >1372 <3200 >610 <2135 Unsuitable <1372 >3200 <610 >2135 Upper Basin and Range Suitable >1220 <2440 <1372 Unsuitable <1220 >2440 >1372 Colorado Plateau/Wyoming Basin Suitable >1220 <2745 <1680 Unsuitable <1220 >2745 >1680 Columbia Basin Suitable >185 <2135 <1220 Unsuitable <185 >2135 >1220 Upper Missouri Basin and Broken Lands Suitable >460 <1525 <460 Unsuitable <460 >1525 >460 Southern Rocky Mountains and Piedmont Suitable >460 <2745 <1675 Unsuitable <460 >2745 >1675 108

Table 14 Habitat models used to estimate current and historic (18,000 years ago) suitable habitat of P. platvrhinos. The models are based on a range of elevations in which the species is found. Elevations are given in meters above sea level. Region Current 18,000 years ago Lower Basin and Range Suitable <1220 <460 Unsuitable >1220 >460 Upper Basin and Range Suitable >1650 <460 Unsuitable <1650 >460 Colorado Plateau Suitable <1220 <460 Unsuitable >1220 >460 109 determine the level of geographic concordance between modeled habitat persistence and patterns of diversity and divergence in the two lizard species.

Results

Habitat Suitability of Phrvnosoma douglassi

Simulation of current habitat conditions revealed large amounts of continuous habitat throughout all six regions (Table 15). Similar to distribution maps given by

Stebbins (1985), the species was mostly absent from high elevations in the southern and northern Rocky Mountains, and within the Sonoran, Mojave, and Chihuahuan Deserts

(Figure 17). The Upper Missouri Basin had the highest percentage of suitable habitat, the largest average habitat patch size, and the highest index of connectivity (Table 15). The

Lower Basin and Range region had the smallest average patch size and lowest value of connectivity, although relatively large, continuous habitat existed in southeastern Arizona

(Figure 17). Although large continuous habitat is estimated within the Upper Missouri

Basin region, suitable habitat is likely far less continuous in this region because P. douglassi is restricted to badlands and river drainage. A geologic map combined with the

Digital Line Graph spatial database might provide a more accurate estimate of suitable habitat within this region. Additionally, certain areas identified as suitable may not have existing P. douglassi populations. An example of this is the Spring Mountains northwest of Las Vegas, Nevada. The area possesses a fairly large patch of suitable habitat (as validated by field visit), but no populations of P. douglassi are known to occur in the area.

Modeled distribution and amount of suitable habitat of P. douglassi during the last glacial maximum are considerably different than estimates of current habitat distribution 110

Table 15 Estimates of suitable habitat for P. douglassi at present and 18,000 years ago. Estimates are based on application of the model to 1:2,500,000 digital elevation models for the western United States. Habitat sizes are per 100 hectares. 1 = percent of total area suitable, 2 = average patch size, 3 = maximum patch size, 4 = relative connectivity of suitable habitat (IEDG equation in Riitters et al. 1995). Region abbreviations are as follows: CLP = Colorado Plateau; UBR = Upper Basin and Range; LBR = Lower Basin and Range; SRM = Southern Rocky Mountains; UMB = Upper Missouri and Wyoming Basins; CMB = Columbia Basin

Current 18,000 years ago

Region 1 2 3 4 1 2 3 4

CLP 84.80% 3655 577360 0.98 13.40% 631 61172 0.90

UBR 89.00% 6449 344419 0.97 18.30% 586 23520 0.90

LBR 16.80% 201 9824 0.89 68.80% 1213 132763 0.98

SRM 94.80% 50482 655968 0.99 87.00% 15095 603675 0.99

UMB 96.40% 110612 442448 0.99 0.99% 649 4301 0.92

CMB 86.40% 3958 655259 0.97 29.00% 2261 61950 0.94 Figure 17 Modeled distribution of current P. douelassi suitable habitat. Cross- hatched areas are suitable and white areas unsuitable. 112 and amount (Table 15, Figure 18). Based on the model, considerably smaller amounts of suitable habitat remained on the Columbia Basin, Colorado Plateau, Upper Basin and

Range, and the Upper Missouri Basin and Broken Lands regions, although the Columbia

Basin, Colorado Plateau, and Upper Basin and Range maintained relatively large patches of continuous habitat (Table 15, Figure 18). The amount of suitable habitat increased considerably in the Lower Basin and Range (16.8% to 68.8%, Table 15) and the Southern

Rocky Mountain region remained approximately the same (Table 15).

Habitat Suitability of Phrvnosoma platvrhinos

Simulations of the current distribution of suitable P. platvrhinos habitat revealed large, continuous habitats across the Upper and Lower Basin and Range regions, and relatively small amounts of suitable habitat on the Colorado Plateau (Table 16). Habitat on the Colorado Plateau follows the Colorado River and, therefore, is highly connected

(Figure 19, Table 16). However, even though suitable habitat exists in the eastern portion of the Lower Basin and Range region, this area is not occupied by P. platvrhinos. The distribution of suitable habitat given in Figure 19 (excluding the eastern portion of the map) matches the distribution of the species given by Stebbins (1985), but correctly excludes the species from relatively high elevation areas within its general distribution

(Figure 19). Of the regions, the Lower Basin and Range had the largest average patch size and greatest connectivity (Table 16).

Similar to P. douelassi. estimates of suitable habitat of P. platvrhinos during the last glacial maximum are different than those for current habitat suitability (Figure 20). With the exception of small habitat patches in the southern and southeastern Great Basin, the Figure 18 Modeled distribution of suitable R douelassi habitat 18,000 years ago. Cross-hatched areas are suitable and white areas unsuitable. 114

Table 16 Estimates of suitable habitat for P. platvrhinos at present and 18,000 years ago. Estimates are based on application of models (Table 14) to 1:2,500,000 digital elevation models for the western United States. Habitat sizes are per 100 hectares. 1 = percent of total area suitable, 2 = average patch size, 3 = maximum patch size, 4 = relative connectivity of suitable habitat (IEDG equation in Riitters et al. 1995). Region abbreviations are the same as listed in Table 15. Current 18,000 years ago Region 1 2 3 4 1 2 3 4 CLP 4.00% 263 16635 0.87 0.03% 15 166 0.58

UBR 43.00% 1414 102011 0.92 0.17% 55 276 0.80

LBR 68.80% 4409 189658 0.97 21.60% 2261 61950 0.94 Denver

Los Angeles

El Paso

Figure 19 Modeled distribution of current P. Platvrhinos habitat. Cross-hatched areas are suitable and white areas are unsuitable. 116

Denver

■as

Los Angeles

El Paso

Figure 20 Modeled distribution of P. platvrhinos suitable habitat 18,000 years ago. Cross-hatached areas are suitable and white areas are unsuitable. 117

Great Basin is largely devoid of suitable habitat. Suitable habitat within the Lower Basin and Range shrinks dramatically to areas adjacent to the Lower Colorado River and in and around Death Valley (Figure 20). This pattern of suitable habitat is approximately equal to that illustrated by Lomino et al. (1989) for the Lower Colorado River area. However, some patches of suitable habitat persist along the Colorado River within the Mojave

Desert, and in isolated patches within other areas of the Mojave Desert (Figure 20).

Discussion

Results of the simulations suggest considerable change in the distribution and connectivity of habitats of P. douelassi and P. platvrhinos over the last 18,000 years. The simulations also suggest that certain areas in many regions have maintained suitable habitats since the last glacial maximum. However, the results of the simulation do not imply that plant community associations remained constant within regions that maintain suitable habitat. Rather, the results suggest that these regions maintained general physical conditions suitable for the two lizard species.

Habitat Persistence and Phylogeographic Patterns in P. douelassi

Simulated distributions and pattern of suitable habitat at present and 18,000 years ago suggest suitable habitats have persisted in the Columbia Basin, the Upper Basin and

Range, the Lower Basin and Range, the Southern Rocky Mountain Piedmont, and the

Colorado Plateau regions since the last glacial maximum.

Modeled distributions of suitable habitats 18,000 years ago approximate the patterns of geographic distribution of mtDNA haplotypes in P. douelassi in these regions 118

(Chapter 2). Each of these regions had relatively deep levels of mtDNA divergence and haplotype diversity. Deep levels of divergence and haplotype diversity in these regions may reflect relatively large effective population sizes associated with large habitat patches that persisted across glacial-interglacial periods; therefore, the phylogenetic history of the populations are maintained.

An alternative model explaining relatively deep phylogenetic history in a region involves large-scale geographic shifts in suitable habitat to the south during glacial periods, and subsequent expansion to the north during interglacial periods (after Van

Devender 1986, Betancourt et al. 1990). However, this alternative model is not generally supported (with exceptions of the Upper Missouri and Wyoming Basins) by the simulation of habitat change or phylogeography of P. douelassi derived in this study.

Although there is considerable change in the distribution and patch sizes of suitable habitat at the regional scale, the habitat simulation suggests that relatively large habitat refugia remained in 5 of the 7 regions during the last glacial maximum. Additionally, large gaps in phylogenetic relationships of P. douelassi between regions suggest long term barriers to gene flow among regions (see Chapter 2) and persistence of suitable habitats within regions across glacial-interglacial cycles. If suitable P. douelassi habitat had persisted only in southern regions during the last glacial maximum and, subsequently, expanded into northern regions during the current interglacial period, one would expect phylogenies among haplotypes to not reflect regional boundaries. This is not the case in this study.

The model suggests that the Upper Missouri and Wyoming Basins maintained little suitable habitat 18,000 years ago and populations in these areas exhibited shallow 119

levels of mtDNA divergence and relatively low haplotype diversity. These patterns

suggest that suitable habitats and populations were lost during the last glacial period, and

that populations became reestablished with the northern expansion of suitable habitats

during the current interglacial period.

Habitat Persistence and Phylogeographic Patterns in P. platvrhinos

Simulations of P. platvrhinos habitat distribution suggest relatively severe

contraction of suitable habitat during the glacial maximum to low elevation areas surrounding the Lower Colorado River Drainage, and at low elevation sites in the Death

Valley region. Habitat contraction during the last glacial period is followed by subsequent widespread habitat expansion. The contraction and subsequent expansion of suitable habitat closely resembles the contraction and expansion of desert habitats illustrated by Lomolino et al. (1989) and Betancourt et al. (1990). Concordant patterns are not surprising since P. platvrhinos is a desert species and since both studies simulated conditions from changes in elevation associations. Mitochondrial DNA diversity in P. p latvrhinos is concordant with these models of habitat persistence. Populations in southern areas of the Mojave and Sonoran Deserts along the Lower Colorado River exhibit high mtDNA haplotype diversity and deep divergence levels. Conversely, populations in the Upper Basin and Range region are considerably less diverse.

However, populations could have persisted in the southern Great Basin during the latest

Pleistocene glacial period. The habitat suitability map did not include basalt volcanic formations, which may have moderated environmental conditions during the last glacial period and allowed populations to persist farther to the north (see Chapter 3). 120

The widespread expansion of desert habitats during the latest post-glacial period

may account for considerable mixing of P. platvrhinos haplotypes, especially in the

Mojave Desert. Furthermore, expansion of P. platvrhinos into the eastern and western

Great Basin from different sources reflects the distribution of suitable habitat in the

models. Eastern Great Basin populations are most closely related to Mojave populations,

and continuous habitats connect these two regions in the model. Conversely, populations

in the northwestern Great Basin are most closely related to populations in the southern

Great Basin, and continuous suitable habitats connect these two regions in the model.

Finally, divergence depths between the eastern and western Great Basin may be related to

a large barrier of unsuitable habitat (relatively high elevation desert and woodland)

dividing the two regions.

Summary and Conclusions

The habitat change model used in this study suggest that large amounts of suitable

habitat persisted during the last glacial maximum, but especially that of P. douelassi.

Large, relatively continuous patches of P. douelassi habitat peristed in the Columbia

Basin, the Upper Basin and Range, the Colorado Plateau (along the Colorado River

Basin), the Southern Rocky Mountains, and Lower Basin and Range regions. These are the same regions in which deep phylogenetic diversity was observed (Chapter 2).

Conservsely, suitable habitat in the Upper Missouri and Wyoming Basins did not persist and these regions demonstrated only recent phylogenetic divergence and little diversity among haplotypes. Results of the model independently corroborate conclusions derived 121 from phylogenetic analysis: that habitats persisted in some, but not all northern regions of the species’ range.

The habitat change model suggests that a relatively large, continuous patch of suitable P. platvrhinos habitat persisted along the Lower Colorado River Basin during the last glacial maximum, and that desert habitats expanded northward during the current interglacial period. This result is consistent with other simulations of change for this region. Habitat persistence roughly corresponded to phylogenetic depth within specific geographic areas; areas that exhibited shallow phylogenetic diversity did not maintain suitable habitat, those with relatively deep divergence generally did.

The habitat simulation appears to provide an estimate of the amount and pattern of suitable habitat, and concordance with phylogenetic diversity suggests that the model may provide a reasonable estimate of change at a regional scale.

The model is not sensitive to fmer-scale environmental change because the data used to conduct the simulation are relatively coarse-grained, and because vegetation patterns at finer scales are constrained by variables other than elevation, such as soils.

The current model is also insensitive to abiotic factors (e.g., rock formations) that might account for persistent of populations in areas with otherwise unsuitable habitat. It may be possible to improve the model by adding a spatial geologic component. CHAPTER V

DISSERTATION SUMMARY AND CONCLUDING REMARKS

The complex history of environmental change in western North America has made it difficult to determine the contributions of relatively old (e.g., uplift of the western

North American cordillera) versus more recent (e.g., glacial-interglacial cycles) events to lineage diversification patterns. Recently, however, molecular phylogeographic analyses of extant taxa have been used to unravel lineage responses to deep versus more recent environmental change (Riddle 1995).

A molecular phylogeographic analysis of two geographically widespread homed lizards, Phrvnosoma douelassi and Phrvnosoma platvrhinos. provide evidence of lineage response to environmental change associated with deep and more recent history.

P. douelassi and P. platvrhinos exhibit two patterns of phylogenetic gaps indicative of lineage responses to relatively deep historical environmental change. First, both species exhibit large phylogenetic gaps between adjacent geographic areas. In P. douelassi. large phylogenetic gaps exist between the Columbia Basin and all other extant populations of the lizard. This phylogeographic pattern is consistent with relatively deep

(8-15 million years ago) historical environmental change in western North America, and the historical biogeography and evolution of the species. Fossil and molecular evidence

122 123

suggest that P. douelassi. a lizard that prefers woodlands, grasslands, and coniferous

forests, probably dispersed in conjunction with the spread of these biome-types across

much of western North America in the late Oligocene to mid-Miocene.

During the Miocene to early Pliocene, it appears that a large and persistent barrier

to gene flow between the Columbia Basin and other populations was established. During

this time, environmental change (steppe and desert establishment, woodland and forest

contraction) in the Columbia Basin lagged behind that of regions to the south and east.

These differences may have led to divergence between Columbia Basin populations and

populations in all other regions. Furthermore, an extensive volcanic ridge separating the

Columbia Basin from the Upper Basin and Range Physiographic region was formed

approximately 15 million years ago and may have been a barrier to gene flow between

these regions. Additionally, uplift of the Rocky Mountains may have isolated the

Columbia Basin populations from those of the Wyoming Basin. The phylogeny of P.

douelassi suggests that all populations in the Columbia Basin were derived from a

common Columbia Basin ancestor; no populations within the Basin are more closely

related to populations within other regions. This suggests that the barrier isolating

Columbia Basin populations has persisted from approximately the Miocene to the

present.

In P. platvrhinos. relatively large phylogenetic gaps exist between the southern

Sonoran Desert populations and all other populations. This relatively deep split corresponds to the Bouse Embayment, a Pliocene marine incursion extending from the

Sea of Cortez along the Colorado River to Mohave Lake. The Bouse Embayment may 124

have isolated populations east and west of the Colorado River between the Pliocene to

late Pleistocene.

Major geographic disjunctions of populations representing deeply evolved clades

may be a second indicator of lineage response to relatively deep historical environmental

change. This type of phylogeographic pattern suggests that the common ancestor of the

two geographically separated regions was once widespread, but subsequently became

isolated by a large-scale vicariance event. A clade of P. douelassi populations from

northeast Colorado, southern New Mexico, and eastern Nevada and southern Utah

represent a relatively deep level of divergence (5-10 millions years Ago), corresponding

to the uplift of the southern Rocky Mountains and the Colorado Plateau. The uplift of these regions may have resulted in a large-scale vicariance event dividing an ancestral population into eastern and western clades. P. platvrhinos exhibits a similar geographic disjunction between a clade consisting of haplotypes from the southern Great Basin and western Sonoran Desert. Although the level of divergence of this clade is relatively deep in relationship to the overall divergence within P. platvrhinos. it likely represents a vicariance event associated with Pleistocene glacial-interglacial cycles.

Smaller phylogenetic gaps between certain regions and between populations within regions suggest lineage response to more recent historical environmental change, especially Pleistocene glacial-interglacial cycles. P. douelassi and P. platvrhinos exhibited a number of scales of nucleotide divergence that could be associated with

Pleistocene glacial-interglacial cycles.

The phylogeographic structure of P. douelassi and P. platvrhinos provides insight into how diversification within any given region may have been structured by deep versus 125

more recent environmental change. The overall phylogeography of P. douelassi suggests

that deep history (geologic activities) established a set of regional lineages. Subsequent

changes in the geographic structure of regional lineages resulted from environmental change particular to the region. This nested spatial and temporal pattern follows the general principals of ecological hierarchy theory. Geologic activities representing relatively deep history set the spatial domain for regional climates. Changes in regional climates resulted in variation in biome patterns within and among regions, and this led to divergence of P. douelassi at broad regional scales. Subsequent environmental change associated with Pleistocene glacial-interglacial cycles resulted in changes in the distribution of biomes within regions, which in turn resulted in further diversification within regional lineages.

Differences in regional phylogenies of P. douelassi exemplify how different histories of environmental change within regions can influence lineage diversity. In northern regions, including the Missouri Basin and Broken Lands and the Wyoming

Basin, latest Pleistocene glaciation appears to have eliminated suitable habitats and extirpated populations. Therefore, the phylogeographic structure of P. douelassi in these regions probably represents a reestablishment of suitable habitat and expansion of populations from adjacent regions during the current interglacial period. Conversely, relatively deep phylogeographic structure in other regions suggests persistence of habitats throughout the series of Pleistocene glacial-interglacial cycles. A model of habitat change associated with the last glacial maximum is consistent with these results. These results support a model of regional refugia, rather than a single southern refugia. 126

The phylogeographic structure of P. platvrhinos suggests that relatively deep historical environmental change (the Bouse Embayment) separated the species into two regional lineages; the southern Sonoran Desert and the remainder of the species range.

However, P. platvrhinos exhibited considerable variation in phylogeographic structure across most of its range. Haplotypes within the Sonoran Desert were geographically structured, whereas considerable haplotype mixing occurred within the southern Mojave

Desert Region. Haplotype mixing in the southern Mojave Desert may reflect numerous expansions and contractions of desert and woodland habitats in the region during

Pleistocene glacial-interglacial cycles. Finally, the phylogeographic pattern of populations in the northern Great Basin reflect a shallow level of divergence. Similar to

P. douelassi. this likely reflects loss of suitable habitat and extirpation of populations in this region during glacial periods.

The phylogeographic patterns derived from extant populations of P. platvrhinos and P. douelassi cannot be used to infer the exact history of either species within any given area. Phylogeographic structure of P. douelassi in Wyoming exemplifies this problem. A fossil record of P. douelassi from the Split Rock Formation suggests that the species occurred in Wyoming during the Miocene, yet the phylogenetic pattern of extant populations in Wyoming suggest a shallow history (latest Pleistocene). In this case, the relatively deep phylogenetic signal was lost with the extirpation of populations from the area, probably during one of the Pleistocene glacial periods.

Phylogenetic diversity of P. platvrhinos and P. douelassi exhibited a general trend of increasing diversity from north to south across their ranges, with the exception of P. douelassi in the Columbia Basin Physiographic region. A similar north-south gradient in 127

species richness exists across Phrvnosoma - southern regions, including Mexico have 4-7

species, mid-latitude regions generally have two species, and northern regions only one

(P. douelassi). The same general north-south gradient exists across reptiles in western

North America (Currie 1991).

Gradients of increasing diversity from north to south have often been associated

with the relative degree of environmental harshness, including low average temperatures

and extreme temperatures, reduced growing seasons length, which generally increases

from north to south. However, changes in the stability and persistence of suitable habitats

across glacial-interglacial cycles may contribute to north-south diversity gradients,

primarily because persistence of habitat influences net effective population size which in

turn can influence genetic diversity. Under this model, northernmost areas lose

significant amounts of diversity because a wide range of suitable habitats are not

available during glacial periods. Regions immediately south of northern regions maintain

fragmented or isolated areas of suitable habitats, but fragmentation and reduction in the

size of suitable habitats results in decreased net effective population sizes. Areas in

southern regions maintain substantial areas of suitable habitat and, therefore, maintain

relatively large net effective population sizes.

In conclusion, the molecular phylogeographic approach used in this study

provided evidence of a hierarchical response of two widespread lizard species to first,

older geologic events, and second, more recent glacial-interglacial cycles. The method

also uncovered a tremendous amount of divergence and diversity in the two lizard species that was not apparent from analysis of morphological variability. Furthermore, morphological traits provide no meaningful way to construct the phylogeographies of the 128 two lizard species, nor in determining intraspecific responses to deep versus relatively recent historical environmental change. Finally, a key to demonstrating biotic responses to deep and more recent history is the selection of widespread taxa with relatively deep lineage history and habitat fidelity related to specific biomes-types. LITERATURE CITED

Allen, T.F.H., and T.B. Starr. 1982. Hierarchy: perspectives for ecological complexity. Univ. Chicago Press, Chicago.

Allendorf, F.W., and R.F. Leary. 1986. Heterozygosity and fitness in natural populations of animals. Pp. 57-76, in M.E. Soule (ed.), Conservation biology: the science of scarcity and diversity. Sinauer Assoc, Inc., Sunderland, Massachusetts.

Avise, J.C., J. Arnold, R.M. Ball, E. Bermingham, T. Lamb, J.E. Neigel, C.A. Reeb, and N.C. Saunders. 1987. Intraspecific phylogeography: the mitochondrial DNA bridge between population genetics and systematics. Annu. Rev. Ecol. Syst. 18:489-522.

Avise, J.C. 1989. Gene trees and organismal histories: a phylogenetic approach to population biology. Evolution 43:1192-1208.

Avise, J.C. 1992. Molecular population structure and the biogeographic history of a regional fauna: a case history with lessons for conservation biology. Oikos 63:62-76.

Avise, J.C. 1994. Molecular markers, natural history and evolution. Chapman and Hall, Inc., New York. 511pp.

Axelrod, D.I. 1958. Evolution of the Madro-Tertiary Geoflora. Bot. Rev. 24:433-509.

Axelrod, D.I. 1979. Age and origin of Sonoran Desert vegetation. Occ. Papers California Acad. Sci. 132:1-69.

Axelrod, D.I. 1983. Paleobotanical history of the western deserts. Pp. 113-129, in S.G. Wells and D.R. Haragan (eds.), Origin and evolution of Deserts. Univ. New Mexico Press, Albuquerque.

Axelrod, D.I. 1985. Rise of the grassland biome, central North America. Botanical Rev. 51:163-201.

Axelrod, D.I, and H.P. Bailey. 1976. Terrtiary vegetation, climate and altitude of the Rio Grande depression, New Mexico-Colorado. Paleobiology 2:235-254.

129 130

Bamowski, C.W. 1984. Late Miocene vegetational and climatic variations inferred from a pollen record in northwest Wyoming. Science 223:49-51.

Betancourt, J.L., T.R. Van Devender, and P.S. Martin (eds.). 1990. Packrat middens: the last 40,000 years of biotic change. Univ. Arizona Press, Tucson, Arizona.

Bowers, M.A., and J.L. Dooley, Jr. 1991. Landscape composition and the intensity and outcome of two-species competition. Oikos 60:180-186.

Brooks, D.R. 1985. Historical ecology: a new approach to studying the evolution of ecological associations. Ann. Missouri Bot. Garden 72:660-680.

Brooks, D.R., and D. A. McLennan. 1991. Phylogeny, ecology, and behavior: a research program in comparative biology. Univ. Chicago Press, Chicago.

Brown, J.H. 1971. Mammals on moutaintops: nonequilibrium insular biogeography. Am. Nat. 105:467-478.

Brown, J.H. 1995. Macroecology. Univ. Chicago Press, Chicago.

Brown, J.H., and A.C. Gibson. 1983. Biogeography. Mosby, St. Louis, Missouri.

Bush, G.L. 1975. Modes of animal speciation. Ann. Rev. Ecology and Syst. Zool. 3:99- 104.

Chaney, R.W., and M.K. Elias. 1938. Miocene and Pliocene floras of western North America. Contrib. Paleotol, Carnegie Inst, of Washington, Washington, D.C. 272 pp.

Chaney, R.W., and D.I. Axelrod. 1959. Miocene floras of the Columbia Plateau. Carnegie Inst, of Washington Publ. Number 617, Washington, D.C.

Christiansen, R.L, and E.H. McKee. 1978. Late Cenozoic volcanic and tectonic evolution of the Great Basin and Columbia intermountain regions. Geol. Soc. America Memoir 152: 645-673.

Clements, F.E. 1916. Plant succession: an analysis of the development of vegetation. Publ. No. 242. Carnegie Inst., Washington, D.C.

Cole, K.L. 1986. The lower Colorado River valley: a Pleistocene desert. Quaternary Res. 25:392-400.

Cooper, S.J.B., K.M. Ibrahim, and G.M. Hewitt. 1995. Postglacial expansion and subdivision of the grasshopper Chorthippus parallelus in Europe. Mol. Ecol. 4:49-60. 131

Cracraft, J. 1983. Cladistic analysis and vicariance biogeography. Amer. Sci. 7:273- 281.

Cracraft, J. 1986. Origin and evolution of continental biotas: speciation and historical congruence within the Australian avifauna. Evolution 40:977-996.

Cracraft, J. 1988. Deep-history biogeography: retrieving the historical pattern of evolving continental biotas. Syst. Zool. 37:221-236.

Cracraft, J. 1989. Speciation and its ontology: the empirical consequences of alternative species concepts for understanding patterns and processes of differentiation. Pp., 28-59, in Otte, D., and J.A. Endler (eds.), Speciation and its consquences. Sinauer Assoc., Sunderland, Massachusetts.

Currie, D.J. 1991. Energy and large-scale patterns of animal- and plant-species richness. American Nat. 137:27-49.

Cwynar, C.L, and G.M. MacDonald. 1987. Geographic variation of lodgepole pine in relation to population history. Am. Nat. 129:463-469.

Danielson, B.J. 1992. Habitat selection, interspecific interactions and landscape composition. Evol. Ecol. 6:399-411.

DeMers, M.N. 1993. Roadside ditches as corridors for range expansion of the western harvestor ant (Pogonomvrmex occidentalis Cresson). Landscape Ecol. 8:93-102.

DiMichelle, W.A. 1994. Ecological patterns in time and space. Paleobiology 20:89-92.

Dowling, T.E., C. Moritz, and J.D. Palmer. 1990. Nucleic acids II: restriction-site analysis. Pp. 250-317, in D.M.Hillis and C. Moritz (eds.), Molecular systematics. Sinauer, Sunderland, Massachusetts.

Dumas, P.C. 1964. Species-pair allopatry in the genus Rana and Phrvnosoma. Ecology 45:178-181.

Echelle, A.A., and T.E. Dowling. 1992. Mitochondrial DNA variation and evolution of the Death Valley pupfishes (Cyprindon, Cyprinodontidae). Evolution 46:193-206.

Endler, J.A. 1986. Natural selection in the wild. Princeton Univ Press, Princeton, New Jersey.

Environmental Systems Research Institute (ESRI). 1994. Understanding GIS: the Arc/Info method, version 7.0.2. John Wiley and Sons, Inc., New York. 132

Etheridge, R., and K. de Queiroz. 1988. A phylogeny of Iguanidae. pp. 83-368, in Estes, R. and G. Pregill (eds.), Phylogenetic relationships of lizard families: essays commemorating Charles L. Camp. Stanford Univ. Press, Palo Alto, California.

Excoffier, L., P.E. Smouse, and J. M. Quattro. 1992. Analysis of molecular variance inferred from metric distances among DNA haplotypes: applications to human mitochondrial DNA restriction data. Genetics 131:479-491.

Felsenstein, J. 1981. Evolutionary trees from gene frequencies and quantitative characters: finding maximum likelihood estimates. Evolution 35:1229-1242.

Felsenstein, J. 1993. PHYLIP (Phylogenetic Inference Package, version 3.5). Distributed by author, University of Washington, Seattle, Washington.

Findley, J.S. 1969. Biogeography of southwestern boreal and desert animals. Pp. 113- 128, in J.K. Jones, Jr. (ed.), Contributions in mammalogy. Univ. Kansas Misc. Publ. Mus. Nat. History. 51:1-428.

Gilpin, M.E. 1987. Spatial structure and population vulnerability. Pp. 125-139, in M.E. Soule (ed.), Viable populations for conservation. Cambridge Univ. Press, Cambridge, UK.

Gleason, H.A. 1926. The individualistic concept of plant association. Bull. Torrey Bot. Club 53:7-26.

Graham, R.W. 1986. Response of mammalian communities to environmental changes during the late Quaternary. Pp. 300-313, in J. Diamond, and T.J. Case (eds.), Community ecology. Harper and Row Publ., New York.

Graham, R.W., and J.I. Mead. 1987. Environmental fluctuations and evolution of mammalian faunas during the last deglaciation in North America. Pp. 371-402, in W.F. Ruddiman, and H.E. Wright, Jr. (eds.), North America and adjacent oceans during the last deglaciation. Geological Soc. America, The Geology of North America, v.K-3, Boulder, Colorado.

Haffer, J. 1982. General aspects of the refuge theory. Pp. 6-24, in G.T. Prance (ed.), Biological differentiation in the tropics. Columbia Univ Press, New York.

Hall, F.G., D.B. Botkin, D.e. Strebel, K.D. Woods, and S.J. Goetz. 1991. Large-scale patterns of forest succession as determined by remote sensing. Ecology 72:628- 640.

Harvey, P.H., and M.d. Pagel. 1991. The comparative method in evolutionary biology. Oxford Univ. Press, Oxford. 133

Hasegawa, M., and H. Kishino. 1989. Confidence limits on the maximum likelihood estimate of the hominoid tree from mitochondrial-DNA sequences. Evolution 43:672-677.

Hennig, W. 1966. Phylogenetic systematics. Univ. Illinois Press, Champaign-Urbana.

Hewitt, G.M. 1993. Postglacial distribution and species substructure: lessons from pollen, insects and hybrid zones. Pp 98-123, in Lees, D.R., and D. Edwards (eds.), Evolutionary patterns and processes. Academic Press.

Hillis and C. Moritz (eds.), Molecular systematics. Sinauer, Sunderland, Mass.

Huntley, B., and T. Webb, HI. 1987. Migration: species' response to climatic variations caused by changes in the earth's orbit. J. Biogeography 16:5-19.

Imbrie, J., J.D. Hays, D.G. Martinson, A. McIntyre, A.C. Mix et al. 1984. The orbital theory of Pleistocene climate: support from a revised chronology of the marine 1180 record. Pp. 269-305, in A Berger, J. Imbrie, J. Hays, G. Kukia, and B. Saltzman (eds.), Milankovitch and Climate. Dordrecht, Netherlands.

Innis, M.A., and D.H. Gelfand. 1990. Optimization of PCRs. Pp. 3-12, in M.A. Innis, D.H. Gelfand, J J. Sninsky, and T.J. White (eds.), PCR protocols: a guide to methods and applications. Academic Press, San Diego, California.

Irwin, D.M, T.D. Kocher, and A.C. Wilson. 1991. Evolution of the cytochrome b gene of mammals. J. Mol. Evol. 32:128-144.

Jones, K.B., L.P. Kepner, and T.E. Martin. 1985. Species of reptiles occupying habitat islands in western Arizona: a deterministic assemblage. Oecologia 66:595-601.

Jones, K.B. 1988. Distributions and habitat associations of herpetofauna in Arizona: comparisons by habitat type. Pp. 109-128, in R.C. Szaro, K.E. Severson, and D.R. Patton (eds.), Management of amphibians, reptiles, and small mammals in North America. USDA Forest Service Gen Tech. Rpt. RM-166, Fort Collins, Colorado.

Kepner, W.G., K.B. Jones, D.J. Chaloud, J.D. Wickham, K.H. Riitters, and R.V. O’Neill. 1995. Mid-atlantic landscape indicators project plan. U.S. Environ. Prot. Agency No. EPA/620/R-95/003, Washington, D.C.

Kiester, A.R. 1971. Species density of North American amphibians and reptiles. Syst. Zool. 20:127-137.

Kuchler, A.H. 1964. Potential natural vegetation of the conterminous United States. American Geogr. Soc. Spec. Publ. No. 36, Washington, D.C. 134

Kumar, S., K. Tamura, and M. Nei. 1993. MEGA: molecular evolutionary genetic analysis, version 1.0. Pennsylvania State Univ., University Park.

Lacy, R.C., and D.B. Lindenmayer. 1995. A simulation study of the impacts of population subsdivision on the moutain brushtail possum, Trichosurus caninus Ogilby (Phalangeridae: Marsupialia), in south-eastern Australia: n Loss of genetic variation within and between subpopulations. Biol. Conserv. 73:131-142.

Lamb, T.C., J.C. Avise, and J.W. Gibbons. 1989. Phylogeographic patterns in mitochondrial DNA of the desert tortoise (Xerobates agassizi), and evolutionary relationships among the North American gopher tortoises. Evolution 43:76-87.

Lamb, T.C., T.R. Jones, and J.C. Avise. 1992. Phylogeographic histories of representative herpetofauna of the southwestern United States: mitochondrial DNA variation in the desert iguana (Diososaurus dorsalis) and the Chuckwall (Sauromalus obesus). J. Evol. Biol. 5:465-480.

Lande, R., and G.F. Barrowclough. 1987. Effective population size, genetic variation, and their use in population management. Pp. 87-123, in M.E. Soule (ed.), Viable populations for conservation. Cambridge Univ. Press, Cambridge, Great Britain.

Leopold, E.B., and M.F. Denton. 1987. Comparative age of grassland and steppe east and west of the northern rocky mountains. Ann. Missouri Bot. Gard. 74:841-867.

Lomolino, M.V., J.H. Brown, and R. Davis. 1989. Island biogeography of montane forest mammals in the American Southwest. Ecology 70:180-194.

Lucchitta, I. 1979. Late Cenozoic uplift of the southwestern Colorado Plateau andadjacent lower Colorado River region. Tectonophys. 61:63-95.

MacArthur, R.H., and E.O. Wilson. 1967. The theory of island biogeography. Princeton Univ. Press, Princeton, New Jersey.

Maddison, W.P., and D.R. Maddison. 1992. MacClade, version 3.0 manual. Sinauer, Sunderland, Massachucetts.

Maddison, W.P., M.J. Donoghue, and D.R. Maddison. 1984. Outgroup analysis and parsimony. Syst. Zool. 33:83-103.

Martin, A.P., G.J.P. Naylor, and S.R. Palumbi. 1992. Rates of mitochondrial DNA evolution in sharks are slow compared with mammals. Nature 357:153-155.

Mayden, R.L. 1988. Biogeography, parsimony, and evolution in North American freshwater fishes. Syst. Zool 37:329-355. 135

McElroy, D., P. Moran, E. Bermingham, and I. Kornfield. 1992. REAP: The Restriction Enzyme Analysis Package, version 4.0. Univ. of Maine, Orono.

McIntyre, N.E. 1995. Effects of forest patch size on avian diversity. Landscape Ecol. 10:85-99.

Montanucci, R.R. 1981. Habitat separation between Phrvnosoma douelassi and P. orbiculare (Lacertilia: Iguanidae) in Mexico. Copeia (1981): 147-153.

Montanucci, R.R. 1987. A phylogenetic study of the homed lizards, eenus Phrvnosoma. based on skeletal and external morphology. Contii. in Sci, Los Angeles County Mus, Number 390, 36 pp.

Murphy, R.W. 1983. Paleobiogeography and genetic differentiation of the Baja California herpetofauna. Occas. Papers Calif. Aca. Sci. 137:1-48.

Nei, M. 1987. Molecular evolutionary genetics. Columbia University Press, New York.

O'Neill, R.V., D.L. DeAngelis, J.b. Waide, and T.H.F. Allen. 1986. A hierarchical concept of ecosystems. Princeton Univ. Press, Princeton, New Jersey.

O'Neill, R.V., R.H. Gardner, B.T. Milne, M.G. Turner, and B. Jackson. 1991. Heterogeneity and spatial hierarchies. Pp. 85-89, in Kolasa, J., and S.T.A. Pickett (eds.), Ecological heterogeneity. Springer-Verlag, New York.

Otte, D., and J.A. Endler (eds.). 1989. Speciation and its consequences. Sinauer Assoc., Inc., Sunderland, Massachusetts.

Page, R.D.M. 1988. Quantitative cladistic biogeography: constructing and comparing area cladograms. Syst. Zool. 37:254-270.

Paterson, B.D. 1980. Montane mammalian biogeography in New Mexico. Southwest. Nat. 25:33-40.

Paterson, B.D., and W. Atmar. 1986. Nested subsets and the structure of insular mammalian faunas and archipelagos. Bio. J. Linnean Soc. 28:65-82.

Patton, J.L., and M.F. Smith. 1989. Population structure and the genetic and morphologic divergence among pocket gopher species (genus Thomomys). Pp. 284-304, in Otte, D. and J.A. Endler (eds.), Speciation and its consequences. Sinauer Assoc., Sunderland, Massachecetts.

Pianka, E.R., and W.S. Parker. 1975. Ecology of homed lizards: a review with special reference to Phrvnosoma platvrhinos. Copeia 1975: 141-162. 136

Platnick, N.I., and G. Nelson. 1978. A method of analysis for historical biogeography. Syst. Zool. 27:1-16

Presch, W. 1969. Evolutionary osteology and relationships of the homed lizard genus Phrvnosoma (Family Iguanidae). Copeia (1969):250-275.

Pulliam, H.R., J.B. Dunning, Jr., and J. Liu. 1992. Population dynamics in complex landscapes: a case study. Ecol. Appl. 2:165-177.

Reeve, W.L. 1952. Taxonomy and distribution of the homed lizard genus Phrynosoma. Univ. Kansa Sci. Bull. 34:817-960.

Rice, W.R., and E.E. Hostert 1993. Laboratory experiments on speciation: what have we learned in forty years? Evolution 47:1637-1653.

Ricklefs, R.E. 1987. Community diversity: relative roles of local and regional processes. Science 235:167-171.

Riddle, B.R., and R.L. Honeycutt 1990. Historical biogeography in North American arid regions: an approach using mitochondrial-DNA phylogeny in grasshopper mice (genus Onvchomvs). Evolution 44:1-15.

Riddle, B.R., R.L. Honeycutt, and P.L. Lee. 1993. Mitochondrial DNA phylogeography in northern grasshopper mice (Onvchomvs leucogaster): the influence of Quaternary climatic oscillations on population dispersion and divergence. Molec. Ecol. 2:183-193.

Riddle, B.R. 1995. Molecular biogeography in the pocket mice (Peroenathus and Chaetodipus). and grasshopper mice (Onvchomvs): the late cenozoic development of a North American aridlands rodent guild. J. Mamm. 76:283-301.

Riitters, K.H., R.V. O’Neill, C.T. Hunsaker, J.D. Wickham, D.H. Yankee, S.P. Timmins, K.B. Jones, and B.L. Jackson. 1995. A factoral analysis of landscape pattern and structure metrics. Landscape Ecol. 10:23-39.

Robinson, M., and T. Van Devender. 1973. Miocene lizards from Wyoming and Nebraska. Copeia (1973):698-704.

Roff, D.A., and P. Bentzen. 1989. The statistical analysis of mitochondrial DNA polymorphisms: X2 and the problem of small samples. Mol. Biol. Evol. 6:539- 545.

Romme, W.H., and D.H. Knight 1982. Landscape diversity: the concept applied to Yellowstone Nationa Park. BioScience 32:664-670 137

Rosen, D.E. 1978. Vicariant patterns and historical explanation in biogeography. Syst. Zool. 27:159-188.

Saitou, N, and M. Nei. 1987. The neighbor-joining method: a new method for reconstructing phylogenetic trees. Mol. Biol. Evol. 4:406-425.

Saunders, D.A., RJ. Hobbs, and C.r. Margules. 1991. Biological consequences of ecosystem fragmentation: a review. Conserv. Biol. 5:18-32.

Schruben, P.G., R.E. Arndt, and W J. Bawiec. 1994. Geology of the conterminous united states at 1:2,500,000 scale - a digital representation of the 1974 P.B. King and H.M. Beikman map. U.S. Geol. Surv. Dig. Data Ser. DDS-11, Reston, Virginia.

Shaffer, M. 1987. Minimum viable populations: coping with uncertainty. Pp. 69-86, in M.E. Soule (ed.), Viable populations for conservation. Cambridge Univ. Press, Cambridge, UK.

Simon, C., F. Frati, A. Beckenbach, B. Crespi, H. Liu, and P. Flook. 1994. Evolution, weighting, and phylogenetic utility of mitochondrial gene sequences and a compilation of conserved polymerase chain reaction primers. Ann. Entomol. Soc. America 87:651-701.

Slobodkin, L.G., and H.L. Sanders. 1969. On the contribution of environmental predictability to species diversity. Brookhaven Symp. Biol. 22:82-95.

Smith, P.B. 1970. New marine evidence for a Pliocene marine embayment along the lower Colorado River area, California and Arizona. Geol. Soc. Amer. Bull. 81:1411-1420.

Smith, J.M. 1989. Evolutionary genetics. Oxford University Press.

Spaulding, W.G. 1990. Vegetational and climate development of the Mojave Desert: the last glacial maximum. Pp. 166-199, in Betancourt, J.L., T.R. Van Devender, and P.S. Martin (eds.), Packrat middens: the last 40,000 years of biotic change. Univ. Arizona Press, Tucson, Arizona.

Stebbins, R.C. 1985. Field guide to western reptiles and amphibians. Houghton Mifflin Co., Boston.

Swofford, D.L. 1993. PAUP: Phylogenetic analysis using parsimony, version 3.1.1. Illinois Nat. Hist. Mus., Champaign, Illinois. 138

Taggert, R.E., A.T. Cross, and L. Satchell. 1982. Effects of periodic volcanism on Miocene vegetation distribution in eastern Oregon and western Idaho. 3rd Proc. N. American Paleontol. Conv. 3:535-540.

Taylor, P.D., L. Fahrig, K. Henein, and G. Merriam. 1993. Connectivity is a vital element of landscape structure. Oikos 68:571-573.

Templeton, A.R. 1989. The meaning of species and speciation: a genetic perspective. Pp. 3-27, in D. Otte and A.R. Endler (eds.), Speciation and its consequences. Sinauer, Sunderland, Massachusetts.

Thomas, C.D. 1994. Extinction, colonization, and metapopulations: environmental tracking by rare species. Conserv. Biol. 8:373-378.

Thompson, R.S. 1990. Late Quaternary vegetation and climate in the Great Basin. Pp. 200-39, in J.L. Betancourt. T.R. Van Devender, and P.S. Martin (eds.), Packrat middens, the last 40,000 years of biotic change. Univ. Arizona Press, Tucson.

United States Geological Survey (USGS). 1987a. Digital line graphs from 1:2,000,000- scale maps. U.S. Geol. Surv. Tech. Guide #3, Reston, Virginia.

United States Geological Survey (USGS). 1987b. Digital elevation models. U.S. Geol. Surv. Tech. Guide #5, Reston, Virginia.

Van Devender, T.R. 1986. Climatic cadences and the composition of Chihuanhuan Desert communities: the late Pleistocene packrat midden record. Pp. 285-299, in J. Diamond and T. J. Case (eds.), Community ecology. Harper and Row, New York.

Van Devender, T.R. 1990. Late quaternary vegetation and climate of the Sonoran Desert, United States and Mexico. Pp. 134-163, in J.L. Betancourt. T.R. Van Devender, and P.S. Martin (eds.), Packrat middens, the last 40,000 years of biotic change. Univ. Arizona Press, Tucson.

Van Devender, T.R., and J.I. Mead. 1978. Early Holocene and late Pleistocene amphibians and reptiles in Sonoran Desert packrat middens. Copeia 1978:464-475.

Van Devender, T.R., R.S. Thompson, and J. L. Betancourt. 1987. Vegetation history of the deserts of southwestern North America: the nature and timing of the Late Wisconsin-Holocene transition. In W.F. Ruddiman and H.E. Wright, Jr. (eds.), North America and adjacent oceans during the last glaciation. The Geol. Soc. America. Vol K-3, Boulder, Colorado. 139

Van Devender, T.R., J.I. Mead, and A.M Rea. 1991. Late Quaternary plants and vertebrates from Picacho Peak, Arizona. Southwest. Nat. 36:302-314.

Verboom, J., A. Schotman, P. Opdam, and J.A.J. Metz. 1991. European nuthatch metapopulations in a fragmented agricultural landscape. Oikos 61:149-156.

Vrba, E.S. 1985. Environment and evolution: alternative causes of the temporal distribution of evolutionary events. S. African J. Sci. 81:229-236.

Webb, T., LEI. 1987. The appearance and disappearance of major vegetational assemblages: long-term vegetational dynamics in eastern North America. Vegetatio 69:177-187.

Webb, T., HI., and P.J. Bartlein. 1992. Global changes during the last 3 million years: climatic controls and biotic responses. Ann. Rev. Ecol. Syst. 23:141-173.

Whitford, W.G. 1994. Harvester ants as indicators of rangeland quality and a description of a technique for rapid estimation of harverstor ant colony density. J. Arid Environ., in press.

Whiting, M.J., J.R. Dixon, and R.C. Murray. 1993. Spatial distribution of a population of Texas homed lizards (Phrvnomosa comutum: Phrynosomatidae) relative to habitat and prey. Southwest. Nat. 38:150-154.

Wickham, J.D., T.G. Wade, K.B. Jones, K.H. Riitters, and R.V. O’Neill. 1995. Diversity of ecological complexes of the United States. Vegetatio in press.

Wilcove, D.S., C.H. McLellan, and A.P. Dobson. 1986. Habitat fragmentation in the temperate zone. Pp. 237-256, in M.E. Soule (ed.), Conservation biology. Sinauer Assoc, Inc., Sunderland, Massachusetts.

Wilson, A.C., R.I. Cann, S.M. Carr et al. 1985. Mitochondrial DNA and two December 12, 1995perspectives on evolutionary genetics. Biol. J. Linnean Soc. 26:375-400. APPENDIX I

Collecting records for P. douglassi and P. platvrhinos, and outgroups. Each specimen is identified by species name, locality, date of collection, habitat type, and LVT

(Las Vegas Tissue) number. Numbers in parens correspond to locality records illustrated in Figure 13 for P. platvrhinos. and Figure 1 for P. douglassi.

140 141

Phrvnosoma coronatum: New Mexico, Jomado Exp. Range, Dona Ana County, Chihuahuan Desertscrub, 20-Jun-94, LVT980

Phrvnosoma douglassi: USA: Arizona, (1) Hualapai Moutains, Mohave Co., Mixed Conifer, 8-Aug-91, LVT806-808, (2) Williams, Coconino Co, Ponderosa Pine, 15-Aug- 92, LVT889,901, (3) Flagstaff, Coconino Co, Ponderosa Pine, 15-Aug-92, LVT890, (4) Springerville, Apache Co, Ponderosa Pine, 9-Oct-93, LVT887-888, (5) Near Young (Tonto National Forest), Gila County, Date Unknown, LVT985, (6) Chircichua Mtns, Cochise County, Pine-oak woodland, Date Unknown, LVT999-1000, (7) Huachuca Mtns near Scotia Canyon, Cochise County, Pine-oak woodland, Date Unknown, LVT989, (8) Chirichua National Monument, Cochise County, Pine-oak woodland, Date Unknown, LVT1002, (8) Krantz Ranch, Cochise County, Pine-oak woodland, Date Unknown, LVT1001; California, (9) Mt. Shasta, Grasshopper Flat, Siskiyou County, Mixed conifer, Date Unknown, LVT994-995; Colorado, (10) Pawnee National Grasslands (Two Forks Creek), Weld County, Broken river bluffs through grassland, 16-Jun-93, LVT876-880; Idaho, (11) INEL Long-term Study Site, 40 mi NW of Idaho Falls, Butte County, Sagebrush desertscrub, 24-Aug-93, LVT874-875, (12), INEL Facility, Idaho Falls, Bonneville County, Developed/grassland, LVT982; Montana, (13) 40 miles NW of Custer, Musselshell County, Grassland, 7-Jun-94, LVT970, (14) 6 miles N. of Cluster, Yellowstone County, grassland, 4-Jul-94, LVT947, (15) 10 miles N. of Forsyth, Rosebud County, grassland, 5-Jul-94, LVT948, (16) 45 miles NW of Forsyth, Rosebud County, grassland, 5-Jul-94, LVT949; Nevada, (17) 30 miles N. of Ely, White Pine County, Pinyon-juniper/grassland ecotone, 28-May-94, LVT923, (18) Currie, Elko County, Pinyon-Juniper/sagebrush, Date Unknown, LVT990; Nebraska, (19) Fossil Agate National Monument, Sioux County, Broken hilltop above river through grassland, 16- Jun-93, LVT882-886; New Mexico, (20) Thoreau, McKinley County, Pinyon- juniper/grassland, Date Unknown, LVT988, (21) Rio de Las Vacas, Sandoval County, Mixed conifer, 26-Jul-93, LVT871-872, (22) 17.4 miles NW Taos, Taos County, Sagebrush desertscrub, 26-Jul-93, LVT869, (23) 25 Mi N. of Cloucroft, Otero County, Mixed Conifer, 28-Jun-92, LVT866, (24) 1 mi W. of Pietown, Catron County, Pinyon- juniper, 25-Jul-93, LVT873, North Dakota, (25) Theodore Roosevelt National Park (Petrified Forest), Billings County, Grassland, 21-Sep-93, LVT900; Oregon, (26) 15 mi NW Bums, Harney County, Pinyon-juniper, 9-Aug-94, LVT955-957, (27) Santiam Junction (air strip), Linn County, Mixed Conifer, 13-Aug-94, LVT964-966, (28) 6 mi SW of Boardman, Morrow County, Desertscrub, ll-Aug-94, LVT958-960; South Dakota, (29) Crow Buttes, Harding County, Broken Hillside within grassland, 7/8-Jul-94, LVT950-951; Texas, (30) Otero Mesa, Hudspeth County, Habitat unknown, Date Unknown, LVT987; Utah, (31) Duck Creek, Kane County, Mixed Conifer, 3-Jul-93, LVT891-895,914, (32) Webster Flat, Kane County, Meadow witin Mixed Conifer, 3-Jul- 93, LVT870, (33) Mammoth Creek, Garfield County, Mixed Conifer, 3-Jul-93, LVT896- 899, (34) Thompson, grassland, Grand County, Date Unknown, LVT2341, (35) La Sal Mtns, San Juan County, Pinyon-juniper, 24-Jun-92, LVT863-865, (36) Grantsville (historic Pianka and Parker site), Tooele County, Pinyon-juniper/grassland ecotone, 12- Aug-92, LVT908-913; Washington, (37) Clockum Pass, between Wenatchee and Ellenburg, Kittitas County, habitat unknown, Date Unknown, LVT993, (38) 6 mi W. of 142

Wilbur, Lincoln County, grassland/desertscrub, 12-Aug-94, LVT961-963; Wyoming, (39) 18 mi S. of Bittercreek, Sweetwater County, Sagebrush Desertscrub/grassland ecotone, 12-Aug-92, LVT902-907, (40) 6 mi. N. of Fort Washakie, Fremont County, grassland, 2- Jul-94, LYT943-945, (41) Worland (at town dump site), Washakie County, broken bluff within grassland, 2-M-94, LVT946, (45) Medicine Bow Nat. Forest, Albany County, grassland, 12-Aug-94, LVT969, (43) 40 mi NE of Caspar, Johnson County, on road extending up from bluffs through grassland, 8-Jul-94, LVT952-954, (44) Green River, Sweetwater County, on bluffs 1 mile S. of town, sagebrush/grassland, 7-Aug-94, LVT968; (42) 15 miles N. Caspar, Natrona County, sagebrush/grassland, 14-Aug-94, LVT967. CANADA: Alberta, (46) Bow Island, Date Unknown, LVT1004-1005, (47) Coulee, Date Unknown, LVT1006, (48) Comrey West, Date Unknown, LVT1007.

Phrvnosoma ditmarsi: Mexico; Sonora - (tissues from Montanucci collection, RRM 2459), LVT996,1003

Phrvnosoma mcalli: Arizona, Marine Corp. Aux. Road #2, Luke Air Force Range, Yuma County, Creosote Bush Desert, 30-Sept-93, LVT971

Phrvnosoma platvrhinos: Arizona, (24) 9 miles N. of Ajo, Maricopa County, Sonoran Desertscrub, Arizona Upland, 16-Apr-92, LVT810,817,819-820, (23) Tucson Mtn. Park, West Side of Tucson Mtns, Pima County, Sonoran Desert, Arizona Upland, 16-Apr-92, LVT818, (21) Bouse Dunes, Sonoran Desert, Arizona Upland, La Paz County, 7-May-92, LVT829-832, (22) Whitman, Maricopa County, Sonoran Desert, Arizona Upland, 7-Apr- 92, LVT361, (20) Wikieup, Mohave County, Sonoran Desert, Arizona Upland, 14-May- 93, LVT733; California, (17) 1 miles SW of Amboy, San Bernardino County, Mojave Desertscrub, 12-Apr-92, LVT812-814, (18) 0.5 miles NW of Eagle Mtn., Riverside County, Mojave Desertscrub, 12-Apr-92, LVT815, (19) 10 miles WSW of Blythe, Riverside County, Mojave Desertscrub, 12-Apr-92, LVT816, (15) 17 mi SE of Bishop, Inyo County, Mojave Desertscrub, 31-May-92, LVT851-854; Idaho, (32) 6 miles S. of Murphy, Owyhee County, saltbush-greasewood desertscrub, 23-Jun-94, LVT934; Nevada,(3)1 mile W. of Overton, Clark County, Mojave Desertscrub, 2-May-92, LVT465, (1) El Dorado Valley, Clark County, Mojave Desertscrub, 25-Apr-92, LVT473, (2) Mile Post #24 -115, Clark County, Mojave Desertscrub, 16-Oct-91, LVT801-802, (2) 1515 Right-of-way, between arrowhead and college, Clark County, Mojave Desertscrub, 2-Oct-91, LVT803-805, (2) Junct Spring Mtn. and Durango Rds, Las Vegas, Cark County, Mojave Desertscrub, 7-Apr-92, LVT811, (2) Charleston Blvd, W. of Peccole Range, Las Vegas, Clark County, Mojave Desertscrub, 16-Apr-92, LVT823, (4) 10 miles N. of Apex site off Hwy 93, Clark County, Mojave Desertscrub, 26-Apr-92, LVT824- 826, (14) 10 miles S. of Hwy 377 and Hwy 376 Junct., just E. of Hwy 376, Nye County, Great Basin Desertscrub w/basalt rock, 24-May-92, LVT833, 834, 838-840, 843-845, (13) Round Mtn dump site, Nye County, Great Basin Desertscrub, 24-May-92, LVT841- 842, (12) Stillwater Mtns., Churchhill County, Great Basin Desertscrub, 3-Jul-92, LVT867, (5) 39 miles N. of Hiko, Lincoln County, Great Basin Desertscrub, 31-May-93, LVT915, (7) 40 miles S. of Battle Mtn, Lander County, Great Basin Desertscrub, 1-Jun- 93, LVT916-918, (10) 25 miles E. of Black Rock Desert, Pershing County, Great Basin 143

Desertscrub w/basalt substrate, 2-Jun-93, LVT919, (11) Black Rock Desert, Great Basin Desertscrub, Pershing County, Great Basin Desertscrub w/ sandy substrate, 2-Jun-93, LVT920-922, (6) 0.5 miles S. of Currie, Elko County, Great Basin Desertscrub w/ basalt substrate, 28-May-94, LVT924, (9) 35 miles NNE of Winnimuca, Humbolt County, Great Basin Desertscrub w/ sand substrate, 25-Jun-94, LVT941, (8) 30 miles NNE of Winnimuca, Humbolt County, Great Basin Desertscrub, 25-Jun-94, LVT942; Oregon, (33) 1 mile W. Hwy 95 on Jordan Range Rd., Malheur County, Great Basin Desertscrub, 24-Jun-94, LVT935, (34) 2 miles S. of Fields on rocks, Harney County, Great Basin Desertscrub, 24-Jun-94; Utah, (25) Beaver Dam Slope, Washington County, Mojave Desertscrub, l-Jun-92, LVT855-857, (26) 2 miles NE of Lund, Iron County, Great Basin Desertscrub, 13-Jun-92, LVT858-861, (31) 15 miles S. of Etna, Fox Elder County, saltbush-greasewood desertscrub, w/ basalt substrate, 29-May-94, LVT925-926, (30) 35 miles N. of Wendover, Tooele County, saltbush-greasewood desertscrub, 29-May-94, LVT927-928, (29) 15 miles N. of Wendover, Tooele County, saltbush-greasewood desertscrub, 29-May-94, LVT929, (28) 20 miles N. of Calleo, Tooele County, saltbush- greasewood desertscrub, 29-May-94, LVT930, (27) Desert Exp. Range off Hwy21, Millard County, Great Basin Desertscrub, 30-May-94, LVT931-933

Phrvnosoma orbiculare: Mexico; Nuevo Leon, Exact Location Unknown, MYZ137792, LVT997 APPENDIX H

State and special collecting permits under which specimens for the project were obtained.

144 Arizona (1991-1995): #JNS00000177

California (1992- 1993): #8017

Colorado (1992-1994): #92-0454, #93-0454, #94-0454

Idaho (1992-1994): #SC0212921

Montana (1992-1994): #1319, #1339

Nevada (1991-1995): #S6118, #S7538, #S9496

Nebraska (1992-1993): #92-88, #93-21

New Mexico (1992-1995): #1972

North Dakota (1992-1994): #0067, #0134

Theodore Roosevelt National Park Special Permit (North Dakota, 1992-1994): #A9015

Oregon (1992-1994): #411-92, #411-93, #411-94

South Dakota (1992-1994): #7

Texas (1992): #SPR-0592-529

Utah (1992-1994): #1COLL00962

Washington (1992-1994): #070

Wyoming (1991-1994): #035, #036, #091 APPENDIX m

Restriction-site distance matrix for Phrvnosoma douglassi haplotypes. See Figure 6 for haplotype distributions.

146 147

O O' o O OO *-<

CS CS Os r l m f '

r- o Tf 0 0 0 0 VO d Tf O' O'

0 0 ft cs r- co

0.0 T|'• d 17.0

s*o •*t vn

0.0 vo o 16.4 9*0 0 VO VO

0 sO 0 ^ to VI d d O' ^ O' 11.1 13.6 11.7 22.1

0 0 0 VO 0 Os vi

d CS O' d 5.6 to to 13.0 13.5 1S.S

0 vo CS O' *fr 00 CS O vO

d d d d 00 9.3 9.8 d d 11.0 to VO VO cs to o v> 8.3 9.4 cs o2 0 0 d d 9.9 00 Tt" d d O' O' CO O' r- o

cs ■tj2 0 0 d Tj2 6.5 d d d d d 10.5 11.1 13.8

VO 0 0 Vi 0 V) OO O'

d cs Os 0 5.6 d d os 13.0 11.7 13.5 O' r- O' 00 cs 'tf

d 00 O' d 4.8 13.0 14.9

tj- 0 «fr V) ft vt o v>

o d d 00 d d to cs 9.7 to ’■t 'O r** 0 0 14.1 d 15.5 (S CN VI VO cs cs 0 0 c** o> sf vi d O' ' t 0 d d 12.6 13.1 15.9

0 0 cs 00 O' NO Vl CO Vi 00 8.7 7.7 0 0 00 Os d d d 9.1 d co cs CO 0 0 r~t O' rj- 0 0 co 3.3 3.8 'vf l-H O' to d 4.9 c4 cs to

£ £ 21.5

o co vo

~~d Tt= c d d os to d d d d 9.3 9.9 d d os~~

~~o cs \T) O) 0 0 »n to Os Vi 00 CS O VO d d 01 -4 tf Tf to d d 00 9.3 d d d 11.0 601 o to 00 ^ ^ r Tf VO OS cs 00 O' CO O' © Os~~

0 0 c O' d d d d d d d 9.3 9.9 d os o d d 11.7 O rf 00 in m tn »o 00 0 0 VO 0 0 -tj- 00 8.8 9.8 © cn d d d d d d d d Os d d 9.4 d O f** O' O' O' ^ ^ ^ ^ VO 0 VD Vi CO ,-H f * SO d d d d d to d 0 0 r- 0 0 d d d 0 0 9.9 d © 00' d 10.5 12.4

~~•—< T—< VO Tf n 'J- 0 cs Vi O 00 O~~

~~=1 to* CO d d 00 d 7.9 § 11.1 f> cs tn v> vo cs tn r» 00 O' 24 26 EC 20 cs 23 a APPENDIX IV~~

~~Sequence data matrix for 65 Phvmosoma douglassi. Numbers correspond to locality data listed in Appendix I.~~

148 CD Eh U U Eh < < U < U Eh Eh < Eh U < Eh Eh Eh Eh Eh U < U Eh U Eh Eh U < CJ Ej Eh < Eh < U < < < E H E H U <2 d 0EHEH0dEHd«! Eh Eh Eh d d Eh Eh. 0 d EHEh d d d d O 0rf< 0 0 d 0 d d d > 0 0 d 0 d d 0 d Eh 0 Eh Eh 0 d Eh Eh Eh d d iJ 0 d Eh EH Eh d d J 0 0 Eh Eh Eh d 151

0 < 0 EH0 < Eh < O Eh O < Eh 0 < 0 Eh 0 < Eh 0 < 0 Eh 0 < Eh < 0 < Eh < Eh Ej 0 EH < < < EH Eh O <0 0 0 < E h 0 < < < 0 0< Eh 0 . 0 0 < EH 0 < < < 0 0 < 0 < < < J O < E h 0 E h Eh < < 0 0EH 0 Eh Eh 0 < Eh 0 Eh Eh < < Eh Eh < < 152

~~U < H H U H H H U H H u < U Hu < H < U H U < H U O < H U H < < H H < < g U < H JUOHUH J U O HU H H < 153~~

a «d u eh u < Eh < Ej CJ fid Eh eh q fid eh q fd q eh q fd eh d EH CJ Ej q EH q EH q EH E‘h q ' c <; cj fid fid q eh eh q eh fd d Eh fid U Eh . q _ 2 . EH q q q fid Eh q q eh q fid . q fid eh q (C < q < , _ fid 0 fid . q - >d fid q fid 3 q id id q fid q .q _ _eh q q q q EH q q q q eh eh q q q q EH EH q q q q q cj q q Ej Ej q Ej q q q_ Ej q Ej q q q eh q Ej q q q EH q EH q < q rij rij EH d eh id q q q UEHOU< _ . q q q eh q q *d fd q eh q q fd _ _«dqEHqq< rtjq^Enqqqq eh q q q q fd eh q q q q fdqfidEnqqqq ^ Eh U Eh 0 < 1 rtl rti eh q fid fid fd u. q eh q fd fid fid idEnqEnqfidfidfid EH<d q q eh q q EHfdfidqqEHqq qEnqqqcjfidfid q eh q q q q q eh q q- q q ' < eh q q q q fid fi=d q eh eh eh q q eh eh ~eh q“ eh eh eh q fid fid Eh qEHEHEnqfidfidEn EH q EH r ) EH eh q fid fid »d ■ ■ EH Eh q fid fid fid ^ fi=d Ej _ q . ■ q q eh q fd eh q q eh q fd fid q fid fid Eh q q ' eh -d q q fid fid Eh fid q q fid eh fid fid eh q dEiEjU Eh EH q u u q q q q q q q q q q q q q q q q q d s s s eh q < q eh q <=d _ eh q eh q fd q fid Eh EH 0 eh q fid q Eh Eh q ^ eh q q fid eh q q . q Eh fd eh q q fid Ej q Ej fd eh q q Eh q Eh Eh fiC U Eh Eh EH fid q EH >d q fid _■ fd , q eh eh >d q >d Eh fid q Eh fd q >d Eh Eh q E. h Eh - - eh q h ■ eh q qeh EHEnqEHEnqqEH Eh Eh q Eh q q EH Eh fill fill 0 Eh fid fid Eh < «d q Eh q fidfidEHfidfidC’ EHEH fid fid EH fid fid q eh %%-Eh q Eh q q fid fid eh q eh q q d fidfidEnqEiqqfid fid fid eh q Eh q q Eh q U E-h q fid q eh q q ~ Ej q fid q fid EnqqEnqfidqid eh q__ qeh q fid q eh q q <1 EHq------q eh q q fid eh q q >d EHqqfidEnqqfid eh q q fid eh q q . q EH EH EH Ej q EH EH EH EH EH fid q qEHEHEHEjEjfidCJ qEHEHEHEHEHfidq q q eh eh fid fid eh q q eh eh fid rtj Eh fid qqEHEHfidfidEHfid q q eh eh -d fd ~ Eh Ej CJ fill q Eh < Eh Ej CJ fid q EH fid EH eh eh q < U Eh Eh Eh q fid q Eh q fill fill EH q q id q fd fid eh q q fid fid q fd fd eh q q q id id eh q q q q eh EH CD fid Ej eh q EH Eh y < ^ q q eh eh <3 < Ej <5 q q EH EH ^ ' d q eh eh fid >d q Eh Eh q q <1 eh H Ej Ej q q fid eh q q fd eh q q q fd eh “ Eh eh q q eh fid fid - eh q q Eh Eh q q fid EH Eh q q eh q q < Ej Ej eh q q fid eh q q fd eh q q >d q q q q fid fid Eh q q q q . . q q q q id fd q q q q q . . q << q q q q q q q q q q q q q q q q q q q q q q q q q q q q q Eh Eh q q q q eh eh q q q q eh eh q q q q eh EHEnqqfidqqEH Eh q q fill cj q fid eh q q fid CJ CD fid eh q q >d CJ CD fid Enqq-dfidycDfid Eh Eh Eh q fid fid q EH EH Ej q fid -d q EH Eh eh q fid fd q EnEHEnqfidfidfidq q q fill, .eh . eh q q q q »d. _ . eh q q q q » d E H . ,- . q q qqidEHfidEnqq eh q q q q Eh eh ■ EH q q q q q Eh Eh »d eh q q q q EH Eh fidEnqqqqEHEn fid fid q fid fid q q fid fid fid - q j fid «d q q fd fid -d cd fid fd q q fd'd'd fid eh Ej CJ CJ EH fid Eh EH q q EH fid fid q eh y ^ fid fid q q fid fid q q - - fid -d q q . . _ Eh Ej CJ fid . fid eh eh eh q fid , . - eh eh q id criqqqEHfidfidq m q q q Eh fid fid q fid m q q q fd fid q m q q q Eh id fd q ooEnqEHqEHEHEH oo Eh fid Eh q Eh Eh Eh fid oo EH fid Eh q EH q q EH Eh EH ' Eh q q Ej Eh EH fid fid Ej Eh q q Eh Eh > q q fid eh q > q q fid eh q fid fd fid > q q (d _■ ~ q q >d Eh J q fid EH q EH EH j q q eh q eh eh fd fid (J q q EH q EH i-3 q q q 154 u e u U C Eh u e u Eh U C Eh CJ C CJ U C Eh U C U Eh U C Eh C EH e C Eh Eh U C a e EH C Eh EH U c u e C Eh Eh U c. u_ e , Ej C Eh Ej U u u u CJ CJ C CS UUCDCUUCCD U u CD U U C CD UUCDCUUCCD h h h CEjCEHCCJEjCD C Eh C CE CE CUE CD ' y % 9 Eh CJ U 3 ^ 3 Eh U U Eh C C C C E-i U U c c c Eh U U Eh C C C C Eh U Eh Eh CS Eh CS U U Eh U Eh CD EH CD U U Eh U eh CD U U Eh U Eh CD Eh CD C U Eh C C Eh U C CD Eh Eh C C Eh U C CD ' C CD Eh C C Eh U CJ U Eh Eh CJ C CJ U U Ei Eh Eh U C U U U U U Eh Eh U C U C CS U(=5 U CDEh CD Eh rf! EH C _ _ UUUUDU< Eh U UCJ U CD U 3 U U U U CD U C U U U U CD U C EHCDUCDCEh UCCC > U U C eh U . >UCDCEh UCCC U U C Eh U . JUUE h UEh Eh CC (J) U CD EH U EH EH iJUCDEh UEh Eh CC U EhEh 155

~~U < O Eh U < Eh 0 rt! 0 E 0 rt! E 0 < 0 E 0 < E < U rt| Ej < E-i Ej <0 0 0 < E 0 . > 0 0 < 0 EE rtj *5 0 E E < < 0 0 E 0 E E hJ 0 0 E 156~~

0 2 0 Eh 0 2 Eh 0 2 O Eh 0 2 Eh 0 2 0 0 2 Eh 0 2 0 Ej 0 2 Eh rtJUrtJEjrtJEHEHU 2 y <1 Eh 2 Eh Eh 0 2 Eh 2 2 Eh Eh 0 2 Eh 2 fit 2 Eh Eh 0 0 0 0 2 0 0 2 0 0 0 0 2 0 0 2 0 0 0 0 0 0 2 0 0 0 0 2 0 0 2 0 rtj Ehrtj “Eh ^ U Eh q 2 E h 2 E h 2 c-, E h 0 2 ' Eh 2 2 0 Eh 0 fit EH 2 EH 2 y Ej CD Eh 00 Eh 2 2 2 2 " " " ■ Eh 2 2 2 2 Eh 0 0 2 2 2 2 EH 0 0 EH fit 2 2 2 U UE h 0 Eh 0 0 EH 0 EH 0 Eh 0 Eh Eh 0 Eh 0 EH 0 Eh 0 Ej CD Eh 0 2 0E h Eh 2 2 Eh 0 2 0 2 2 Eh 0 2 0 Eh EH 2 2 Eh 0 0 0Ej 0 0 Eh Eh 0 2 0 0 Eh Eh 0 2 0 0 0 Eh Ej 0 0 2 0 rt! 02 2 0 ‘ 2 0 ‘ 2 Eh 0 2 0 fit fit 0 2 0 0 Eh O ^ Eh 2 Eh 0 2 0 0 EH0 2 EH 2 2 0 y rt! U Eh EH Eh 0 0 2 0 2 0 2 fit 0 EH0 Eh 2 0 2 U 5 0 Eh Eh Eh Eh 2 2 2 0 0 0 rt! 0 0 0 2 0 0 0 2 0 2 Eh 0 Eh 0 0 2 0 2 Eh 0 EH 0 fit 0 Ej 0 2 2 Eh 0 2 Eh 2 0 0 2 Eh 0 fit Eh Eh rt! 0 rt! 0 Eh 2 0 2 , Eh 2 0 2 0 2 0 Eh Eh 2 0 2 0 0 Eh 0 0 rt! 0 Ej 0 0 ■,2 Eh _ 0 0 0 0 2 Ej CD 0 0 0 0 0 2 rt! rt! 0 Eh Eh 2 2 0 "Eh Eh 2 2 Eh 2 2 0 Eh Eh 2 2 Eh 2 2 0 0 Eh 0 0 rt! Eh 0 0 O 2 Eh 0 Eh Ej Ej 0 0 2 Eh 0 Eh Eh Eh 0 0 2 Eh 0 Eh 0 0 0 ' Eh 0 0 0 2 Eh 2 2 Eh Eh 0 0 Eh 2 2 Eh Eh 0 0 2 0 0 0 rt! 0- -0 -0 2 2 Ej 0 EH Eh 0 0 2 Fj Eh 0 0 0 2 2 0 0 0 0', . . _ 0 0 0 0 2 2 2 E H 0 0 0 0 2 2 Eh 0 0 0 0 2 . 0rt!00000rtj 0 2 0 0 0 0 0 2 0 2 0 0 0 0 0 0 0 0 0 0 0 0 0 EhEh022CJ0Eh - - - 0 0 Eh EH Eh 0 0 0 0 Eh EhEh002CJ0Eh EH0022EjCD2 Eh 0 2 Eh 0 0 2 y <2 ^ EH0022CJCD2 EhEhEh02220 0 2 2 2 0 - EH Eh 0 2 2 0 EhEhEh02220 00202EHEH0 . 0 2 EH Eh 0 0 0 2 Eh 0 0 002Eh2EhEh0 Eh 0 0 0 0 Eh Eh Eh 0 0 0 0 Eh Eh 2Eh0000EhEh 2rt!EH0000EHEH 2 2 0 2 2 2 0 2 2 0 2 2 2 0 2 2 0 2 2 0 0 2 2 2 0 2 2 0 0 <00rtjEnEn00 0 0 2 EH EH 0 0 0 0 2 Eh Eh Eh 0 2002EHEHEH0 0Enrt!0000EH Eh 2 0 0 0 0 Eh 2 0 2 0 0 Eh 0EH20200EH “ “ Eh 0 0Eh EH 0 Eh Eh“ Eh " 0 0 EH Eh 0 Eh 0 0 Eh Eh 0 EhEhEh00EhEh0 “ Eh 00 0 0 2 Eh Eh Eh 0 0 0 0 0 Eh EH 0 0 0 2EH0EHEjCJ00 Eh rt! 3 0 2 0 Eh 0 0 0- Ej 0 0 E h 0 E h 2 2 u 2 ' 02 Eh rt! Eh 0 2 Eh , fit 0 Eh Eh 2 2 u 2 Eh<' . 0 2 Eh EH 0 0 Eh Ej ^ 0 0 Eh . 0 - 0 0 EH 2 0 2 eh 0 02 2 0 0 0 Eh 2 0 0 0 0 0 0 0000022 E H 2 0 E H 0 0 0 Eh _ 0 EH 0 0 0 Eh Ej 0 0 Eh Eh 2 0 Eh 2 0 2 2 E H 0 0 0 2 0 2 2 0 0 0 2 0 2 2 0 0 0 2 0 2 fit " 0 " 0 0 0 E H 0 0 E H 0 0 0 0 Eh 0 0 _ 0 0 0 0 Eh 0 0 0 0 0 0 Eh 0 0 0 0 0 Eh 2 0 2 0 U 0 E h Eh 2 0 rt! 0 0 0 Ej Eh 2 0 2 0 0 0 Eh 2 0 2 0 0 E h 2 E h 2 0 2 0 0 Eh Eh 2 0 2 0 Eh Eh 2 2 0 2 CJ Eh Eh 2 2 0 2 EH00EHrt*EHEjCJ Eh 0 0 ^ B E H C J Eh 2 0 EH Ej Eh Eh Eh 2 0 E h Eh 2 eh 2 2 2 0 Eh Eh 2 EH 2 2 2 Eh E-h E-h . Eh 2 2 Eh Eh Eh ' ' " 2 Eh 00 2 0 EH 0 0 2 0 CJ Ej 2 : 0 0 0 2 . Eh 2 00 0 . EH Eh 00 Eh 0 Eh 0 0 Eh 0 2 Eh Eh 0 0 Eh 2 FH Eh 0 0 E h Eh 2 00 Eh Eh 2 0 0 Ej 0 EH‘ 2 0 0 Eh 2 y- Eh " 2 2 0 0 Eh Eh Eh 0 2 Eh Eh Eh 0 2 . 2 Eh Eh Eh 0 2 2 2 Eh Eh Eh 0 2 ■rt1 0 0 0" Eh 2 2 0 2 CJ 0 0 Eh 2 2 0 0 0 0 EH 2 2 0 0 0 0 2 2 0 2 0 0E h Eh 0 0 Eh Eh Eh 0 Eh 0 Eh Ej Ej Eh 0 Eh Eh Eh Eh 2 0 0 2 Eh 0 0 2 Eh ' 2 0 Ej Eh Eh 0 0 Ej Eh 2 2 Eh Eh 0 Eh 0 . 0 0 2 Eh 0 . 0 0 2 0 2 2 2 0 Eh Eh 0 Eh Eh 0 2 EH 0 Eh Eh 0 0 Eh Eh Eh 2 2 0 < Eh Eh 0 Eh Eh EH 0 < Eh y ^ y U rtj EH 0 rtj Eh Eh 0 Eh Eh < EH < rt! Eh Eh U rtj0rt!0rt!EH00 < 0 < 0 < E h 0 0 _ _ , Eh < Eh Eh 0 u u a O O < 0 0 0 0 < 0 0 0 0 0 0 0 < 0 0 0 0 0 0 0 < 0 0 < 0 < Eh < O Eh 0 Eh i=5 s , a g E ! E h 0 Eh 0 Eh Eh Eh 0 Eh 0 0 0 Eh Eh . 0 _ 0 _ Eh Eh 0 Eh Eh 0 0 Eh Eh Eh 0 Eh Eh 0 0 rtj rtj Eh O 0 < Eh 0 > 0 0 0 Eh 0 rtj sS rtj 0 0 0 > 0 0 < Eh 0 < < a 0 0 eh 0 Eh Eh 0 0 < Eh < E< Eh < < J 0 0 EH 0 Eh Eh < < 158

U rt! O Eh U rt! Eh 0 rt! Ej Eh 0 Eh Eh 0 2 0 -eh 0 2r-n Ec-1h 0 2 0 Eh 0 2 Eh 2 0 2 E h 2 E h Eh 0 rt!0rt!0rt!EH00 2 Eh 2 2 ' 2 ' 'E h ‘Eh 0 2 0 2 E h 2E h Eh 0 0 0 0 2 0 0 2 0 000rt!00rt!0 0 0 0 2 0 0 2 0 0 0 0 2 0 0 2 0 2 Eh 2 CJ Ej 0 rt! Eh rt! Eh < 2 EH 2 Eh 0 0 Eh 0 2 Eh 2 Eh CJ Eh CD Eh U U . 2 2 2 EH 0 0 Eh -2 ,--' 0 — 0 ^ 2 2 2 2 Eh 0 0 0 2 2 2 U CJ Eh Eh 0 U CD 0 0 EH Eh 0 EH 0 Eh 0Eh 0Eh 0E h 0 h 0 _ Eh 0E Eh 0 0 0 2 CD " rtj rtj Eh O rt! 0 Eh rtj rtj Eh 0 2 0 E h Eh 2 2 E h 0 2 0 2 2 Eh 0 CJ U Eh Eh O 2 O 0 Eh Ej - . Eh O 2 O - Ej Ej Eh 0 2 0 y O Eh 0 2 0 2 CD O 2 Eh O 0 0 2 rtj 0 irtj Eh 0 2 0 2 2 0 2 Eh 0 2 0 0 2 Eh 0 Eh CD Ej 0 *2 Eh fC rt 0 0 Ej 0 2 Eh 2 0 0 0 Eh 0 2 2 u 0 2 U ^ Eh Eh 2 O O rt! 0 EHE n O r t0 0 2 0 E H 0 E H 2 0 2 2 0 Eh 2 0 0 CJ Eh 2 Ej O 2 Eh rt!rtj tjEh 0 rt! Eh ~ J ~EH 2 J ' Eh 0 “ 2 " Eh Eh 0 2 Eh a O U rt] Eh CD Eh 0 2 Eh 0 0 , 0 0 Eh 0 Eh 0 0 Eh 0 Eh CD Eh EH _ EH (=2 CD CD Eh 0 Eh 2 0 0 0 0-| £-4 0_ 0_ Eh 0 Eh Eh Eh , _0 0 Eh O O O CD CJ 2 Eh 0 0 0 0 0 0 2 E H 0 0 0 0 c5 0 2 Eh 0 0 0 0 0 0 2 Eh CD 2 2 Eh Eh _ 2 eh 2 2 u a u u O O O CJ 2 0 0 0 0 0 0 0 0 CJ CD CJ 2 s s s Eh 0 0 rt -2 0 0 0 2 0 0 0 2 Eh CD 2 CD Eh Eh CD 0 0 rt0 2 Eh 0 2 0 Eh Eh 0 Eh 0 2 0 3 Eh 0 EH CJrtEjrt “ 0 EH EH0 2 Eh 2 e h o u Eh 0 2 Eh . rt! CJ CDrt] 2 2 y 2 0 0 2 2 0 Eh 2 0 0 2 2 2 0 Eh 2 0 0 rt! CD CD 2 U 2 0 2 0 0 2 0 .,0% 2 2 0 0 2 0 2 0 2 2 0 0 2 0 CJ U Eh Eh hE CD 0 0 2 - Eh 0 Eh 0 0020 EH 0 Eh 0 0o2-0 Eh Eh Eh 0 CJ 0 CJ ' Eh CD «< 0 0 0 0 EH 0 2 0 0 0 0 Eh 0 2 0 Eh 0 0 Eh 0 2 rt! 0 cj a CJ Eh Eh 0 0 0 0 0 Eh Eh 0 0 0 EH 0 Eh Eh 2 0 0 0 Eh Eh Eh < rt! CJ Eh Eh Eh Eh 2 2 0 Eh Eh 0 Eh Eh Eh Eh 2 ‘ Eh u c j i 2 a Eh CJ Eh 0 0 2 0 Eh 0 Eh Eh 0 Eh 0 0 2 Eh rt! Ej u J . rt! CJ CD 2 Ej 0 2 ^ 0 0 Ej 0 2 Eh O 2 Ej 0 0 Eh rt! U CJ CD CD Eh Eh 2 0 2 0 0 Eh , 20,- 2 0 Eh Eh 2 0 _ Eh U Eh 0 0 rtj ' Eh CD U 0 Eh 0 0 0 - ____0 0 0 0 0 0 2 Ej CD O 0 Eh 0 0 2 Ej 0 0 rt! rt! 0 0 Eh rt! CD Eh 2 2 0 E h Eh 2 2 E h 0 2 0 Eh Eh 2 2 Eh 2 2 0 Eh Eh 2 2 Eh 0 0 rt! Eh 0 ^ Ej ^ 002EH0EnEnEj 002EH0E 0 2 E H 0 0 2 2 0 0 2 Eh 0 Eh , y q r t rt! rtj u Eh Eh Eh 0 2 2 2 0 Eh Eh Eh o 2 2 2 0 ‘. Eh_ Eh_ 0 2 < < u 0 0 , _ . Eh Eh CJ 0 0 2 E h 2Eh Eh 0 0 0 2 E h 2 eh 0 0 00202EhEh0 CJ CD CD CJ Eh Eh 2E h 0 0 0 0 E h Eh 2 E 0 0 0 0 E E 2Eh0000EhEh 2 2 . - O 2 . -E h 0- O - 2 2 0 2 2 0 2 2 2 2 0 2 2 2 0 2 20 2 2 2 0 0 0 0 Eh Eh 0 0 2 0■ 00 2 EH EH EH 0 2002EHEHEH0 0 0 2 EH EH 0 0 Eh rt! 0 0 0 0 Eh 0 EH2 0 0 0 0 Eh 0 E H 0 0 2 0 0 E H 0 Eh 00 0 0 0 Eh Eh Eh 0 0 Eh Eh 0 Eh Eh E_ h 0 _ 0 Eh Eh 0 Eh Eh Eh 0 0 E h Eh 0 Eh Eh Eh 0 0 Eh Eh 0 ^ 0------Eh CJ CJ 0 CJ 0 Eh 0E h 0 0 0 0 2Eh Eh Eh 0 0 0 0 2 Eh 0E h 0 0 0 0 EH 0 EH < 0 Eh 0E h 0E h 0 E h 2 2 u 2 0 E_ h _ 0 Eh 0 2 0 2 0 2 2 0 Eh 0 2 E E 2 2 0 r t E 2 Eh 0 2 ‘ 2 Eh 2 Eh ‘ • EH rtj 0 0 0 Eh . , _ 0 0 Eh 0 0 2 E h Eh 0 0 Eh 2 0 2 Eh Eh 0 CD , 0 0 0 0 0 -, 0 0 0 0 0 2 2 0 0 0 Eh O CD 2 2 _ _ Eh Eh rt! 0 0 0 EHE h 0 0 0 0 E h Eh 2 0 E h 0 0 0 Eh .E h 2 0 Eh 2 0 2 2 " 0 _ 0 0 2 0 2 2 Eh Eh 2 0 2 2 E H 0 0 0 2 0 2 2 0 0 0 0 Eh 0 0 0 0 0 0 EH 00 2 Eh _ 0 Eh 0 0 Eh 0 0 ‘ 0 ^ 0 Eh 0 0 0 0 0 i 2‘ 0 2 Eh 0 0 Eh 0 0 2 Eh 0 0 Ej Eh 2 0 2 rt! 0 rt! EH 2 0 2 EH 0 Eh Eh 2 Eh 2 Eh 0 0 Eh 2 192 0 Eh 2 Eh Ej Ej 2 0 Eh 2 E -h ■ -Eh j EH• Eh Eh 2 Eh 2 2 Eh Eh0 2 EH 2 0 0 0 0 Eh 20 0 0 2 0 Eh 20 0 0 2 0 0 0 Eh 0 0 2 2 Eh Eh 0 0 Eh 2 2 Eh Eh 0 0 Eh 0 0 rtj 0 0 2 0 Eh 2 2 0 0 2 0 Eh 2 0 0 0 2 2 0 Eh Eh n22EHEHEH02 cn 2 2 E h Eh Eh 0 2 oo 0 0 0 C M 0 0 0 E H 2 2 0 cm0 0 0 E h 2 2 0 oo000Eh220 CM Eh 0 Eh CM Eh 0 -• Eh 0 Eh Eh Eh CMEH0EH0EH0EH c^Eh OEh OOE h Eh Eh 0 0 i 0 0 Ej 2 Eh 0 0 Eh Eh Eh CD Eh 0 0 Eh 2 Eh ' ‘ > 0 0 - 002EnOrtlrtjrtj > 0 0 2 Eh 0 2 > 0 0 2 Eh 0 a 0 0 J 0 2 Eh 2E h Eh 2 2 0 2 Eh 0 Eh Eh (J 0 0 Eh 0 0 Eh 159

0 2 0 Eh 0 2 Eh CJ 2 0 Eh CJ 2 Eh y 2 (j eh 0 0 Eh Cj 2 0 Ej 0 0 Eh 2 0 2 E h 2 E h Eh 0 2 CJ 2 Ej 2 Eh Eh CJ 2 Eh 2 2 2 Eh 2 2 2 Eh Ej CJ 0 0 0 2 0 0 2 0 CJ CJ CD 2 CJ CJ 2 CD 0002 S B 3 8 0 0 0 2 0 0 2 0 2 Eh 2 Eh 2 CJ Eh CD 2 Eh 0 Eh 2 Eh 0 Eh 2 0 Eh 0 0 0 - U 0 2 2 2 2 Eh 0 0 0 3333 Eh 0 0 0 2 2 2 2 ' Eh Eh CD O CD Eh 0 Eh 0 EH 0 EH 0 Eh 0 Eh 0 Eh 0 Eh 0 2 2 Eh CJ 2 0 Eh 2 2 Eh 0 2 2 Eh 0 CJ CJ Eh CJ 2 CJ 0 0 Ej .Eh 0 0 0 Eh 0 0 0 " 2 ^ 0 2 0 2 0" 2 E-i 0 - ' Eh 0 2 CJ CD Eh 0 Eh 2 2 0 0 0 CD 2 CD CD 2 0 2 0 2 - CD 2 CJ 2 Eh 0 J 2 Ej 0 2 Eh 0 2 Eh Eh 0 E ' h 0 Eh 2 Eh 0 Eh Eh 0 Eh CD CD Eh 0_ .... Eh 2 _0 ^ 0 Eh 0 0 Eh 000 0 0 < CJ CJ CJ 2 Eh 0 0 0 0 0 0 2 0 0 0 0 0 0 2 Eh 0 2 Eh 0 2 CJ Eh 0 2 Eh 0 2 CJ 000 CJ CJ CD 2 rt! 0 Eh 0 Eh 0 Eh 2 ■ " ■" Eh 0 Eh 2 _ _ _ _ Eh Eh Eh Eh 2 Eh 0 Eh 0 Eh CJ Ej 0 "Eh Eh 2 2 0 Eh 0 00220 Eh CJ Eh 0 Eh 2 2 0 rt! 2 0 Eh 0 Eh Eh CD Eh U Eh CJ Eh Eh Eh Eh Eh 0 Eh Eh 0 Eh 0 ' Eh Eh CJ 0 Eh CJ a Eh Ej Ej 0 Ej "EH 0 Eh Eh 0 0 2 0 CD 2 a cd 2 Eh 2 2 Ej 2 Eh Ej 2 ^ EH Eh 2 2 Eh 2 2 0 Ej Ej 2 2 - 2 2 Ej § 5 2 2 cd 2 c$ 2 2 0 2 2 0 2 2 0 2 0 0 2 CD u EH 0 0 0 0 0 Eh 0 0 0 0 U Eh CJ CD CJ CD CD CJ eh CJ o Ej 0 0 0 EH 0 Ej 0 Eh 2 CD 2 CD Eh 2 CD 2 0 EH 2 CD Eh 2 Eh 2 0 . 0 0 2 CJ CJ _ _ a cj 2 0 0 0 Eh 0 0 2 0 0 , 0 Eh 0 0 2 U CD CJ CJ Eh CJ CD U CJ Eh 0 0 0 0 2 0 2 Eh 0 0 0 0 - — - CD 2 2 2 _ _ Eh CD 2 2 2 Eh 0 2 2 2 2 E._.0h 0 Eh 2 2 2 r=< 2 cj cd cj cj a 2 2 0 C D C J U C J . _ V Eh 0 0 2 2 0 0 EH 00 Eh U CD Eh 0 < CD Eh CJ CD Eh CJ 2 2 0 Eh 0 0 0 0 2 2 EH 0 00 0 2 2 0 EH- Eh Eh CD2 2 0 U "Eh Eh ' Eh CD J ■' 0 Eh 0 Eh Eh Eh 0 2 2 ^ Eh Eh CD < < Eh CJ Eh Eh CD Eh E■ h 0 2 2 ij y i j y ij y Ej Ej 2 Eh 0 2 2 i j y ^ Ej O 2 2 2 Eh CJ_ CD Eh CJ . Ej 0 . 2 2 (2 CD rt! CJ CD 2 CD 2 2' EH 2 CJ CD 2 0 2 0 0 2 0 EH 2 ' 0 0 2 EH S 3 61 EH eh CJ 2 EH 2 2 Eh Eh CJ 2 eh ^ i j ^ 2 EH 33 ^ ^ y 0000 cj a cj cj Eh 2 2 0 0 0 0 0 0 0 0 CJ CD CJ CD S 3 3 cj a cj 2 0 2 2 0 0 0 2 0 0 0 2 %%% Eh CD < CD Eh Eh CD EH CD 2 CD Eh Eh 0 0 2 0 Eh Eh 0 Eh 0 2 0 Eh Eh 0 2 Eh 0 0 Eh CJ Eh 2 EH CD CD rt! Eh 0 Ej 2 Eh 0 0 Ej 0 Ej 2 Eh 0 0 Eh 0 Eh Eh 2 0 Eh rt! CJ rt! Eh 2 CJ CJ Eh 2 0 2 Eh 2 0 Eh 2 0 2 Eh 2 0 Eh 2 0 2 EH eh CJ "_ _ CJ CD Eh Eh Eh CJ ‘. CJ O 0 Eh E.h Eh 0 . . 0 0 eh Eh Eh 0 ". . 0 0 Eh 2 2 Eh rtj rtj CD Eh Eh ^ 2 2 0 Eh Ej 2 2 Eh 2 2 0Eh 0 2 2 ^ 2 2 0 Eh 0 2 2 eh .2 >2 Eh CJ Eh CD CJ rt! 2 2 EH CJ Eh 0 U 2 2 2 Eh 0 Eh 00 2 2 2 0 EH 0 0 2 Eh 0 0 Eh a a Eh CJ 2 EJ 2 Eh CJ CJ -, CJ 2 0 2 Eh 0 0 _E h _ 0 20 2 0 2 0 2 Eh Eh CJ rt! EH CD CD 2 Eh CJ CJ 2 Eh 0 0 2 Eh 0 0 2 Eh 0 0 2 Eh 0 0 2 Eh 0 0 2 ° Ej Eh Eh Eh E - h 2 . CJ _ CJ Eh Ej Eh 0 Eh Eh E “h Eh Ej2 0 0 EH EH EH i j i j 2 0 CD CD rt!, Eh rt! rtj Ej rtj CD CD 2 EH “ 2ij ij2 E3h 2y 0 0 _ E h Eh 2 2 Eh . 0 0 EH EH 2 2 Ej 2 E B 0 2 O E 2 y Eh Eh 0 2 0 Eh Eh Ej CJ " 2 CJ Eh 2 Ej * ^j y 0 Eh 2 Ej CJ rt! U22 g : 00 2 2 0 2 2 Eh 0 0 2 2 0 2 2 Eh 0 0 . . 2 Eh 0 0 2 2 0 0 0 Eh 0 2 Eh EnCDCJEHCDrt!EH2 Eh CD U Eh 0 2 Eh 2 0 0 0 Eh CD2 Ej 2rtJCDEH CJ CD 2 Eh O > 0 0 2 Eh Eh 0 0_ 0 _ 2 . Eh 0 > 0 0 2 Eh 0 ^ U CD Eh CJ Eh Eh 2 2 i-D O 2 Eh 0 Eh Eh 0 0 Eh 0 Eh EH 1-3 0 0 Eh 0 Eh EH 160

U rfj U Eh U 0 Eh CJ rt! O EH U rt! EH CJ rt eh a rt a Eh U rt eh rtJCJrtJCJrtJEnEnCJ rt Eh Ej CJ rt a rt CJ rt eh Eh U u u 0 r t u u « ! 0 U U 0 C U U r f ! 0 CJ a rt CD a a 0 rt a a rt 0 rtjEnUE-i^UEHCg Eh rt! EH CJ CD rt! Eh CJrt! rt! rt] > CJ CD CD Eh U 0 rt EH U a 0 0 Eh tJ CJ CD Eh a EHE h r t rt! iJ CJ CD Eh CJ a 0 Eh a EH eh 1 a 0 Eh O 161

O < O Eh O O Eh 0 <0 Eh 0 < Eh 0 < 0 Eh 0 < Eh 0 < 0 Eh 0 < Eh < U < CJ < Eh Ej < 0< Eh < Eh Ej 0 <0 0 0 Eh Eh Eh 0 < < 0 0 E h Eh Eh 0 0 0 Eh Eh 0 < < 0 0 Eh EH0 0 < < E j C D < EH 0 0 . Ej CD < . 0 0 < < 0 0 < , Eh 0 0 . y *3 < Eh Eh El 0 < < 0 EHEHEH0<<<0 Eh Eh Eh 0 < < 0 EHEHEH0<<<0 0 0 < Eh _ .EH _EH 0 _ 0 0 < 0 < E h Eh 0 0 0 < 0 Eh Eh 0 0 0 < 0 < E h Eh 0 Eh 0 0 0 0 Eh Eh < E h 0 0 0 0 E h Eh 0 0 0 0 Eh Eh < E h 0 0 0 0 E h Eh < < 0 Eh < 0 < < < < 0 < < < 0 < 0 < < < 0 < < < 0 < < < 0 0 0 < Eh Eh Eh 0 < 0 0 < E h Eh 0 0 0 < Eh H 0 0 < O O < E < E H 0 0 " < 0 0 0 0 Eh 0 E h < 0 0 0 0 E h < 0 0 0 0 Eh Eh < 0 0 0 0 Eh Eh 0 0 Eh Eh 0Eh Eh Eh 0 0 E h Eh 0 Eh 0 0 Eh Eh 0 Eh ‘ Eh 0 0 Eh Eh 0 0 Eh 0 0 00 < E h Eh Eh 0 0 0 0 ‘ Eh 0 0 0 0 < Eh Eh Eh 0 0 0 0 EH Eh < < 0 <0 E H 0 E H < < 0 < < 0 < 0 Eh 0Eh < < 0 < ' ' ‘ Eh < E H 0 < < 0 < E H < < Eh < Eh 0 << 0 < Eh < 0 0 0 0 E h < 0 < E h Eh 0 0 < Ej Eh 0 0E h < 0 < Eh Eh 0 0 E h 0 0 < < 0 0 0 < < 0 0 0E h 0 0 < < 0 0 0 E h Eh < 0 E h 0 0 Eh 0 0 0 Eh _Eh < 0 Eh 0 < 0 < < 0 0 0 < 0 < < E h 0 0 0 < 0 < < 0 0 0 < 0 < < -000 Eh 0 EH 0 0 0 E h 0 0 E h 0 0 0 . 0 0 0 0 Eh 0 0 Eh 0 0 0 Eh < En<0<000Ej 0 0 0 Ej E<<0<000EH 0 0 00EH 0 0 0 Eh Eh Eh 0 . „ Eh 0 < < < Eh 0 < < < i-3 0 0 Eh 0 Eh Eh 0 Eh Eh 0 0 E h 0 E h Eh < < 0 Eh Eh < < U i=5 U eh U r t j EH U =5ehi =5cd rtj CD Eh rtj CD i=5 CD 2j rt! i=5 _ rt! . CD_ CJ U U Eh U CJ < UU i=5UEh UCJ i=5 a a _ a a rtj UUrtOEHUUrt! w Eh Eh CD U CJ f5EH < (< p5 eh a a cd Ej < a a cd CJ CD UEh rtj vd U CD U rt! 1=5 a EH UE h U U Eh Eh a\ EH a EH Eh Eh Eh , Eh CJ CJ Eh rt! Eh ' a a eh a a Eh a a Ej Eh rtj i=5 Eh > CJ CD *5 Eh U . a CD CD a CD CD > a cd i=5 U *5 < rtj J U CD Eh U Eh Eh i=5 eh 1-5 a rt! EH Eh Eh i=5 i=5 163

0 d EH eh 0 EH eh 0 rtj Eh Eh 0 Eh Eh 0 d Eh Eh 0 Eh Eh 0 d Eh Eh 0 Eh Eh d 0 d d d E n 0 0 d 0 d 0 d E H 0 0 d 0 d 0 d Eh 0 0 d 0 d 0 d E n 0 0 UUO 0 0 d Eh 0 . 0 0 d Ej ° 0 0 d > 0 0 0 Eh 0 . hd 0 rtj Eh rtj Eh Eh J 0" d Eh d Eh Eh J 0 d Eh d Eh Eh 164

U rt Eh Eh U Eh Eh U rtj U Eh U rt! Eh U rt! CJ Eh CJ rtj EH cj rt a U rt Eh COrfJCIrt! Eh U U rtJCJrtjEHrtjEHEnCJ rtJCJrtJEnrtjEHEHCJ rt CJ rt rt Eh U a u u c x i u o o u uuOrtiuariicD UUCDrtUUrtCD CJ CJ CD CJ CJ rt CD rt Eh rt J EH rt! EH rt U Ej CD rt! Eh rt! Eh “ rt - CJ - - Ej - -O rt' Eh rt rt CJ Ej CD e-h •=*; i=*! >=*; <=t! Eh CJ O Eh rtj rtj rtj rtj Eh cj a rtj rfj rij rfj CD Eh CD CJ Eh CD Eh CD U CJ Eh CJ Ej CD Eh CD CJ CJ Eh _ Eh CD Eh CD rti Eh U rt! CD Eh Eh rtj rt! Eh U rt! CD Eh EH rt) rtj EH CJ rt CD Eh rt rt Eh U U rt U Eh Eh U rt! CJ CJ CJ Ej Eh EH q rt! a O CJ EH eh CJ r t CJ J eh a Eh CJ rt! CD rt rtj CJ Eh U rt CD rt' q " rt Eh CJ u cd O CD Ej CD rt Eh rt) CJ CD Eh r t rt CJ CD . cd cd CD CD rt! CJ Eh CD CD E j Eh rt CD CD U ^ Eh rt! Eh U rt! CJ < Eh U rt! Eh r t Ej U rt Eh Eh CD U U rti CJ Eh CD Eh Eh r t Eh CD Eh - s w cd cd u CDEh Eh CD CD Eh Eh r t _CD CD Eh UUUUCDUrtEn U U U U O U r t ! CJ CJ CJ CJ CD CJ rt! U U U U O U r t EnCDrtEnUrtUrt Eh CDrt! Eh CJ rtj CJ Eh CD rt! Eh CJ rt) CJ Eh CD rt Eh CJ rt CJ CJ CJ CD Eh CJ E~h rt! rti CJ. . .U . .CD . Eh cj -E-i rt! a cj cd Eh CJ ■E-i rt U ^ E h Eh Eh UE h Eh Eh rt!E h CJ Ej E .h . . Eh Eh Eh < Eh U Eh Eh Eh Eh Eh r t Eh CJ Eh Eh Eh Eh EUEH CJ CD CD U rt rt rt U CD r t EH U r t r t r t CJCDrtEHUrtrtrt >UCDrtEHUrtrtrt J U rt! Eh Eh rt! rtj U CD Eh U Eh Eh r t r t hDCJCDEHCJEHEHrfirt JUOEHUEHEnrtrt 165

~~U < U Eh U <1 Eh~~

U U U U U■ O rt! U O rt) r t) Eh O O Eh rt) U Eh EH Eh U Eh rt) rt) O rtj rtj Eh U Eh U U Eh U U Eh U rt) U Eh Eh U rt) Eh O O U Eh Eh Eh Eh Ej O O rt) Eh rt) rt) Eh Eh U < U Eh U rt) rt) Eh U Eh Eh O U Eh O rt) _ rt) O Eh rt) U O rt) O O rt) U O U . U Eh Eh Eh O O U O U“ ‘ Eh O rt) O O U U U Eh Eh EH Eh rt) rt) U EH EH U U rt! U Eh U Eh U Eh U *2 EH U O Eh rt) U rt) U O Eh U Ej O O rt) rt) rt) O Eh Eh rt) rt)<2 u Eh U U rt) Eh U E-1 Eh E) Eh U O O Eh rt) rt) U U U rt) U U U U , rt) . rt) . - OrtJOUUUUrt) EHEHOrt)rt)UUEH EHUUrtJrfEHUC EHEHEHUf5 rt)rt)U UUrtJUrtjEnEnU rtJEHUOOUEHEn ^rt)rt)Ort)rt)rt)U i^UUrtJEHEHUU OEnrtJOUUUEn EhEhEhOUEhEhU Eh U Eh U U U U Eh U Eh rt) ' ' ' Eh U rt) U U U Eh . U U O Eh O O O U Eh Eh _ rt) U Eh U rt) U rt) rt) . u u o U EH U U Eh U U O Eh rt) O tCi U U O Ej U U Eh Eh rt) U rt) Eh O U 3 Eh Eh “ rt! U Eh rti Eh *2 U Eh rt) u u rt! rtl rt) Eh U U Eh ^ E j Eh O U Eh rt) Eh Ej <2 < in U O U ft rt) U cn Eh U Eh U Eh Eh u ^ Eh Eh > U O rt) U . t-3 U O Eh Eh Eh APPENDIX V

mtDNA distance matrix for P. douelassi haplotypes. Distances were generated from ND5 and Cyt-b sequence matrix with transitions/tranversions weighted 1:5. Haplotypes numbers correspond to locations given in Figure 10. Outgroups distances also are given (PHDI =

~~Phrvnosoma ditmarsi. PHOR = Phrvnosoma orbiculare. PHPL = Phrvnosoma platvrhinos).~~

~~166 167 rinHHOnOlNinHOI 4 N -<4 VIIO rl Ol (1 (1 O oain**o3h'>in*inr>'«'«*ooo H O 0 0 0~~

~~tH4bh(ie)MnO<«a)OCDOI<><«400lCDAU~~

~~lOhtniAhM/inuihin'iooorionnn_i _j _i _j _J .1~~

~~tnui ^<00pHiO000HH >rtrt0rtrtoo00ortortrt rl H HrlrlrlrlnrlrlHHH~~

0 in i«Hnnnci*ooHAAinti> in p. o 0 p» rt o co co in <1 <1hCBHWCOMMnOkACOCDOhirt rt > rt rt rt i h* co 4V)(044(on(B040(iu)4' ' 0 W Oco to n cn o i 0 0pHOrt00P>00O000rt0'rt rt rt '000rt rt rt rt rt co co co

~~ppHrirtO0000000rt00 rl0tPrtrtrt0rtrtrt I1 0 0 rtOPH0 TH0 0 O 0 0 0 0 O~~

~~<« rl (1 < Ifl nn^owMANrtootwoeo 10000000001 hOrtiAOhoi^oiaacpon w4 *4~~

co rt co 0 o rt okrt^inntootrortcototocortrt o rt to oi f* 00000 o d pi rt hC0rlinehC 0nOlCDC»OOO CO CO rt rt rl o co o 33 _ l —I 33' 0 <0 O rt in P* o o CO <0 * Ol o rt Ol Ok00 COo Cl rt Cl o ci cc p. 0 0 0 rt rt o rt CO* rt CCCOCOrt in CD p. Ol 0 00 COCD00 rt COCO00 cc cc CC 0 0 O 3 rt rt rt rt 3 3 rt rt rt n rt Ok rt o o CC<0 o Cl in p* 0 Oko COCC Ol COCOO Cl rt Cl P»O Ol Cl CC0 0 rt o O o CO0 rt cc CCCOCOrt in co pHOl Cl Ol 00 COCOrt 00 00 CO rt CCCC0 0 O 11. rt rt rt rt rt 12. rt rt rt 0 0 rt to COCO COo CCrt Cl CC rt CDin o in Cl o pH Ok CCrt o rt in 0 Cl rt o O o P* <*> n rt rt pHO rt cc P- COrt in p* p* Ol * Ok Ol COCOrt CO00 a rl rt rt CCcc 0 0 P» a rt rt rt rt rt rt rt rt

O COrl rl o p. o rt P»to Ph 00 Cl eo in cc *0 00 00 P* o 00 COO cc rt CC rt rt rt 0 0 rt rt 0 p* CD 00 COp* * CO P* o o rt Cl oi p* in Cl * pH p* in cc 0 P*0 0 0 o o o O o 0 0 0 rt rt rt rt rt rt rt in in rt rt o o o CO P»Ok rt Cl to CO CC Ol pH COeo O CCCOrt Cl cc rt CC 0 rt rt P»0 rt rt 0 o p* CO 00 CO00 * p* p* o rt rt c> Ol p* * Cl <0p* cd in cc in pH 0 0 o e O O 0 0 0 rt rt * rt rt rt rt rt n p* to o * Ol Ol rt rt CCcc rt to cc Cl Ok *0 COrt o in ci 0 rt pH Cl CCCl <00 O 0 P»0 0 0 o Ok rt rt o in in o o 0 rt rt Okrt Cl rt rt rt o o O 0 0 0 0 0 O O O rl 3 3 3 rt rt rt a d rt rt rt rt rt rt 3 3 rt rt rt rt rt rt rl rt rt

Cl O P»rt in m 0 <0 <0 in rt « * CCcc n o p p. to cc Ok O CC00 P*pH 0 O COin in rt O 0 0 0 Ol rt rt is i- o P*P» rt p* P* o o cc Cl 01 P» 0 Cl 0 pH p» in CC0 P*Cl Cl 0 Ok Ok Ol o O 0 0 0 rt rt rt rt pH to 0 p* P»rt ci rt rl rl rt rt COOl <0 *0 rl P» rt 00 pHCl Ok o pHrt rt 0 cc 0 o 0 0 0 0 O 0 O rt o n 0 P* CD 00 00 00 0 00 0» o o Cl 0 O P 0 Cl <0 CDp* in 0 in CO0 0 0 o O O 0 0 0

rt rt rt rt 10. 10. rt rt rt in rl rt rt >0 rt in rt Ol rt o o CC0 o Cl in p* in Ol o 00 CC Ol 00 00 O Cl cc Cl P* O Ok 0 0 0 0 0 07 r* rt CDCO0 o o o a 00 0 rt CC CDCDrt in co pH Ol Cl Ol 07 CO07 rt CO07 09 rt CC rt CC0 0 0 0 rt a rt rt rt rt rt rt rt rt

to rl in in pH rt 0 rt m P»Ol cc * co p. * COCC 0 COcc 0 p* in 0 rt O cc 0 0 Ol Ok0 3 0 pH COCO01 p* p* COCOCOOk Ol Ok pH Ok Ol CO00 to rt CO CO00 P* pH p* O 0 0 0 P^ P. 18. 19. 19. 20. 18. 20. 17, A rt rl rt rt rt rt rt rt rt rt rl rl rl rt rt rl rt rt rt rt rt rt rt 18. rt rt rt rt rl rt rt rt rt rt rt rt OS r< 0 <0 to Cl in in P 0 0 0 0 0 0 0 0 Cl Cl 0 0 0 0 0

Pi rt rt rl rt rt rl rt rt rt rt rt rl rt rt rt rt rl rt rt rt rt rt rt 15. 15. 16. rt rt rt rl rt rt rt rl rt rl rt rt rt rt rl rl rt rt rt m 0 rt rt u> Ol o to * OO C* rt Ok in 0 rt CCCC rt rt CCin P» Ok O in rt CCin Cl CO0 to cc in 0 rt Ok co e 0 0 O 0 0 0 0 h CO<71 COo 00 Ol 00 COrt 00 OkOk Ol OkOl OkO Ol rt Cl P COCC O 01 COOkOkO COOk00 rt Oka a ci CC CC0 0 Ok 0 0 rt rt rt rt rt rt rt rt rt rt rt rt rt rt rt Pii 0 Cl OS a H a 3 rt rt rt * in to pH 00 Ol o rt *0 in to Ol o rt cc to pH 07 3 oi o rt CCCl 0 0 o rt 0 0 12 13 25 17 18 Qi D» rt rt rt r t rt r t cc CCcc 23 24 cc cc cc P i CC Cl Cl Cl Cl Cl 35 36 37 38 Cl 0 0 0 0 Appandix v (continued) ) c N H O i o o c h * * o * i i* f n i i ie o i i i io c c i i ic c ic c H i ic O ic i ic o p n in c o i c c i c c i ( m n i i i C l C O O O O O O O I C I C O O H I ioH(inciciiorioiinHoto O O O - t O I I M l l H O H l f H h H O A I O H i i C O O l t O O O l C O O l H C O C O C O i coi oi r* i ICO I ^ c> C- « I 01 oi n i

~~h («U)IO«I~~

~~1 l N N O D C O H O iC O O inoc)* HU^OO OlCl H * « * hHHUI^COCON nin i O O O O N l r I M i i oi iniQO ' rl I ’ lO H H H iH O H ItO h 1A in I I * O 40 uioooooiiiiiii m m dHlAriOlUICIOHHCO~~

~~u m H 4 4 I / M I / C I O O H W C 4 0 OI C OHrl icoP» ■ H COIS cl HCO I OI Ol 10 O < I Cl I O l C I < Cl O I rl 4 h O O Ol|S •~~

II N O i n i i in H < HHH HOO h r* c* *

~~168~~