MIAMI UNIVERSITY The Graduate School

Certificate for Approving the Dissertation

We hereby approve the Dissertation

of

Richard A. Seidel

Candidate for the Degree:

Doctor of Philosophy

Director Dr. David J. Berg

Reader Dr. Brian Keane

Reader Dr. Nancy G. Solomon

Reader Dr. Bruce A. Steinly Jr.

Graduate School Representative Dr. A. John Bailer ABSTRACT

CONSERVATION BIOLOGY OF THE GAMMARUS PECOS SPECIES COMPLEX: ECOLOGICAL PATTERNS ACROSS AQUATIC HABITATS IN AN ARID ECOSYSTEM

by Richard A. Seidel

This dissertation consists of three chapters, each of which addresses a topic in one of three related categories of research as required by the Ph.D. program in ecology as directed through the Department of Zoology at Miami University. Chapter 1, Phylogeographic analysis reveals multiple cryptic species of amphipods (Crustacea: Amphipoda) in Chihuahuan Desert springs, investigates how biodiversity conservation and the identification of conservation units among invertebrates are complicated by low levels of morphological difference, particularly among aquatic taxa. Accordingly, biodiversity is often underestimated in communities of aquatic invertebrates, as revealed by high genetic divergence between cryptic species. I analyzed PCR-amplified portions of the mitochondrial cytochrome c oxidase I (COI) gene and 16S rRNA gene for amphipods in the Gammarus pecos species complex endemic to springs in the Chihuahuan Desert of southeast New Mexico and west Texas. My analyses uncover the presence of seven separate species in this complex, of which only three nominal taxa are formally described. The distribution of these species is highly correlated with geography, with many present only in one spring or one spatially-restricted cluster of springs, indicating that each species likely merits protection under the U.S. Endangered Species Act. I present evidence suggesting that habitat fragmentation, long- distance colonization, and isolation-by-distance have occurred at different temporal and spatial scales within this system to produce the lineages that I report. Chapter 2, Detecting conservation units using morphological versus molecular criteria: evaluating the Gammarus pecos species complex as a test case , compares the results of morphological versus molecular biodiversity assessments within the G. pecos species complex. I compared results from an earlier morphology-based study to my results from screening 166 COI gene sequences according to Moritz’ Evolutionarily Significant Unit (ESU) concept and a DNA barcode-based Species Screening Threshold (SST) concept. I found strong concordance between the two molecular screening methods, but these two molecular methods separated populations as distinct from one another whereas the morphological method alone failed to separate the same populations. Overall, I found that morphological and molecular techniques for biodiversity estimation should be combined, when possible, to produce a powerful tool for addressing taxonomic and conservation issues. Chapter 3 , Salinity tolerance as a potential driver of ecological speciation in amphipods (Gammarus spp.) from the northern Chihuahuan Desert , examines population level responses to salinity. My results suggest barriers to gene flow between populations as a result of ecologically- based divergent selection, and that tolerance to habitat salinity has structured biodiversity across springs in the northern Chihauhuan Desert. Furthermore, salinity tolerance is correlated with environmental salinity. This work shows a role for both putatively neutral processes (e.g. isolation and genetic drift) and natural selection (acting on population-level physiological responses to habitat salinities). Taken together, these results provide clues important for future biodiversity investigations in geographically isolated aquatic habitats, and shed light on the understudied and underestimated levels of biodiversity present in desert spring systems. CONSERVATION BIOLOGY OF THE GAMMARUS PECOS SPECIES COMPLEX: ECOLOGICAL PATTERNS ACROSS AQUATIC HABITATS IN AN ARID ECOSYSTEM

A DISSERTATION

Submitted to the Faculty of

Miami University

in partial

fulfillment of the requirements

for the degree of

Doctor of Philosophy

Department of Zoology

by

Richard A. Seidel

Miami University

Oxford, Ohio

2009

Dissertation Director: Dr. David J. Berg Table of Contents Table of Contents ...... ii List of Tables...... iv List of Figures ...... v Dedication ...... vii Acknowledgement...... viii General Introduction...... 1 Literature Cited...... 5 Chapter 1 ...... 7 Phylogeographic analysis reveals multiple cryptic species of amphipods (Crustacea: Amphipoda) in Chihuahuan Desert springs Abstract ...... 7 Introduction ...... 8 Methods...... 11 Results ...... 13 Discussion ...... 15 Literature Cited...... 22 Acknowledgments ...... 32 Chapter 2 ...... 42 Detecting conservation units using morphological versus molecular criteria: evaluating the Gammarus pecos species complex as a test case Abstract ...... 42 Introduction ...... 43 Methods...... 45 Results ...... 48 Discussion ...... 50 Literature Cited ...... 54 Acknowledgments ...... 58 Chapter 3 ...... 65 Salinity tolerance as a potential driver of ecological speciation in amphipods ( Gammarus spp.) from the northern Chihuahuan Desert Abstract ...... 65 Introduction ...... 66 Methods...... 67 Results ...... 69 Discussion ...... 70

ii Literature Cited...... 74 Acknowledgments ...... 79 General Conclusions...... 86 Literature Cited...... 87 Appendix I: Mismatch distributions...... 88 Appendix II: Probit plots...... 90 Appendix III: Raw mortality data from salinity tolerance experiment...... 95

iii List of Tables Chapter 1

Table 1. Sample sites and population codes for the Gammarus pecos species complex, as identified by Cole (1985). All individuals analyzed during this study were collected within 300 m of the coordinates shown. The last three sites either, 1) harbor no amphipods presently, or 2) cannot be located based upon earlier descriptions given. Extant populations which were not considered by Cole (1985) are indicated in the second column by the character “-”...... 34

Table 2. Summary of descriptive genetic data for combined sequences, with corresponding GenBank accession numbers for each fragment...... 35

Table 3. Pairwise F ST for all populations (* = not significantly different from 0, at α = 0.05) ...... 36

Chapter 2

Table 1. Sample sites and population codes for the G. pecos species complex. All individuals genetically analyzed during this study are identified by population code in the far left column, and were collected within 300 m of the coordinates shown. The last three sites either, 1) harbor no amphipods presently, or 2) cannot be located based upon earlier descriptions given. Extant populations which were not included in Cole (1985) are indicated in the second column by the character “-” ...... 64

Chapter 3

Table 1. Sample sites and population codes for the Gammarus pecos species complex, as identified by Cole (1985). All individuals analyzed during this study were collected within 300 m of the coordinates shown...... 83

Table 2. Physiological response parameters estimated using probit analysis, displayed by population code and ambient salinity. P-values included in this table are associated with Pearson P2 goodness-of-fit tests ...... 84

Table 3. Above diagonal: Pairwise river distances (km) between respective springs, measured along spring runs and the Pecos River. Malpais Spring (MS) is not hydrologically connected to other springs (the symbol “—” means no river distance estimated). Below diagonal: Average pairwise nucleotide differences (percentage difference) between populations, with corresponding F ST values in paretheses (adapted from Seidel et al. 2009). F ST theoretical minimum of 0 means no genetic divergence due to high gene flow, and theoretical maximum of 1 means complete fixation for different alleles in population due to absence of gene flow ...... 85

iv List of Figures

Chapter 1

Figure 1. Collection sites for G. pecos complex populations from the Chihuahuan Desert. Mountains are indicated by diagonal lines. See Table 1 for population codes...... 37

Figure 2. Pairwise geographic distance along rivers vs. pairwise F ST ...... 38

Figure 3. A haplotype network constructed from 134 combination sequences for COI and 16S. Each unique haplotype is displayed as an oval, and the size of each oval corresponds to the haplotype frequency, with frequencies higher than 1 denoted by number (e.g. n = 3). Small yellow circles correspond to hypothesized intermediate haplotypes not detected in the sample. Relatively large gaps have number of intermediates indicated inside lightning bolt symbol...... 39

Figure 4. Maximum likelihood (ML) tree constructed from the 91 unique concatenated haplotypes, with bootstrap values (≥ 50%) shown on branches and nodes...... 40

Chapter 2

Figure 1. Body plan of the amphipod, modified from Hillewaert (2006)...... 60

Figure 2. Collection sites for 12 extant populations of the G. pecos complex from the Chihuahuan Desert. Two sites with populations now extirpated (eIC and eNS) are also marked for reference. Population codes follow Table 1...... 61

Figure 3. Results are shown for the three methods of analyses: A) morphological distance, B) evolutionarily significant units, and C) 10× species screening threshold. Colored boxes delineate nominally conspecific entities, although molecular results reveal problems with these currently accepted species names. Population codes follow Table 1...... 62

Figure 4. A) Neighbor-joining (NJ), and B) maximum parsimony (MP) trees, each using COI haplotype data and 1000 pseudoreplicates to obtain bootstrap values. Haplotype groups shown at tips are collapsed to reciprocally monophyletic clades, with population codes as in Figure 1. Triangles shown at tips on NJ tree are sized according to degree of sequence divergence for haplotypes within collapsed clades. Numerical values at nodes are the percentages of bootstrap replicates containing each node. Tip labels indicate the number of haplotypes included from each population...... 63

Chapter 3

Figure 1. Locations for 9 extant populations of the Gammarus pecos species complex from the Chihuahuan Desert used in the field experiments (populations codes identified in Table 1)...... 81

v Figure 2. Tree showing the physiological distances, PDs, between pairs of populations in the Gammarus pecos species complex, as assessed by salinity tolerance assay. Clades are labeled at right with corresponding ambient habitat salinities, which clustered according to low, medium, and high levels...... 82

vi Dedication

To my wife, Irma: thank you so much for your unconditional love and support. You have moved across the world with me, you have endured a typhoon and tight living quarters with me, and you have worked so hard at your job to help our family afford many needed things. You have truly earned your PhT ( Put him Through) degree, just like my grandmother June did for Grandpa. For these things and many others, I will be forever grateful. Maraming salamat po. Mahal kita, sobra!

To my son, Hayden: thank you for your laughter, your joy, and your boundless curiosity and energy. Even as I complete this dissertation for my doctorate, I can already see in your four-year- old face the excitement and interest in nature that I remember having early in life. I love to watch you explore the world and hear you ask questions—some of which I am still able to answer. Soon it will be your turn to venture out and explore, learn, and discover, and I have every confidence that you will bear a great gift in your life doing whatever you choose. I cannot wait for the days to come when you will return home to visit your mom and me, and you will teach us things. We can already see your desire to help others and make a difference, and your mom and I love you so much.

To my parents, Dave and Lynda: thank you for supporting me and believing in me. Thank you for telling me that I could do it, and for believing it. Thank you for training me up in the way I should go, and while I do not feel that old yet, I plan not to depart from that training. Thank you for sacrificing and taking the proverbial smaller portion for yourselves, when you knew that would allow larger portions for Lesley and me. I love you.

To my grandfather, Don: thank you for showing me nature. Thank you for explaining details to me and pointing out things that I never would have thought to look at or think about. I wish you could be at Miami University in August of 2009, because without you I am quite sure that I would be someplace else doing something far different than receiving my doctoral diploma in zoology, just as you received at Cornell University in 1960. Thank you for encouraging my curiosity and showing me it was (and is) okay to be fascinated by natural things around us— subjects so interesting that a life could be spent understanding those things better. I hope I’ll be able to communicate to others what I learn, just as you did to me.

To my master’s advisor, Terry Donaldson: thank you for agreeing to mentor me back at the University of Guam Marine Laboratory. While I had years of science teaching under my belt when I came to you, I had not done a minute of formal biological research at that time. You agreed to help a young man who graduated from a university you had never heard of, and you gave considerable time and attention to mentoring a young scientist who needed your early help to learn the ropes. Thank you for teaching me persistence, thank you for criticizing my writing so I could make it better with your help, and thank you for encouraging and supporting me even when it meant risking your own reputation and position on an unknown quantity like me. Si yu’us ma’ase!

vii Acknowledgements

I am forever indebted to a great team of people who helped me bring this body of work forward. First I thank my doctoral advisor, Dave Berg, for his support, his patience, and his belief that various data collection challenges could ultimately be overcome and light would be shed on interesting questions. Dave always encouraged me to think more broadly than just the taxa I study, but instead to think about the big questions in biology and to address the critical gaps in scientific knowledge. I also thank my doctoral committee: John Bailer, Brian Keane, Nancy Solomon, and Bruce Steinly. I greatly appreciate their guidance and expert advice during my doctoral research process. Thank you to Brian Lang, who was not officially on my doctoral committee, but was still central to teaching me how to conduct field research. Without Brian, I would probably be someplace in the Chihuahuan Desert dead from a rattlesnake bite, slumped over a cactus, clutching an empty water bottle. When a scientist ventures out of the DNA lab and into the field, the real fun begins, but that enthusiasm must be tempered with a keen eye for the ecosystem one enters, because without proper respect for and understanding of that system, things can go very wrong. I thank Chris Wood, who always offered patience and excellent advice when I was using the Center for Bioinformatics and Functional Genomics (CBFG) facility here at Miami University. I thank the other members of the Berg lab: Jorge Lassús, Todd Levine, Angie Matthews, Emy Monroe, and Makiri Sei for reading earlier versions of this work and for collaborating with me to solve various problems. I thank Jeff Howland, David Riskind, Tom Johnson, Larry Paul, Patricia Griffin, Junior Kearns, Robert Myers, and John Karges for assistance through their agencies and offices related to access to field locations. I also thank Miami undergraduates Neil Bruce, Vybhav Jetty, and Kirk Webber, who labored long and hard on repetitive but important tasks: I greatly appreciated their help. Finally, I thank the entire Department of Zoology faculty, students, and staff, who were all great to work with and to know.

Funding for this work was provided by the New Mexico Department of Game and Fish, The United States Fish and Wildlife Service Region 2, The Crustacean Society, and the Miami University Summer Field Research Workshop. This dissertation has been approved for public release by White Sands Missile Range for unlimited distribution; the White Sands Missile Range Operational Security review was completed 25 August 2009.

viii General Introduction

The modern American southwest presents a landscape with geologic and topographic features sculpted by a long history of tectonic plate movement and erosion. To understand the modern habitats in which imperiled desert spring biota occur, it is necessary to consider the geological events which both shaped the physical and chemical characteristics of the modern environment, and facilitated the stranding of progenitors that have given rise to the modern lineages now present in these springs. Various saltwater transgressions have occurred onto the North American continent, most recently with the Western Interior Seaway (WIS). The earliest phase of the WIS began in the mid-Cretaceous, ca. 100 Ma (Stanley 1999), during which the Zuni transgression brought seawater from the Arctic Ocean south across western North America towards the proto-Gulf of Mexico (Baldridge 2004). This narrow seaway formed without connecting to the Pacific Ocean to the west, due to the higher elevation mountain range separating the craton (stable continental shelf) from the much more expansive saline waters of the Pacific (Baldridge 2004). Around 90 Ma, the WIS reached a maximum width of 1000 km, and was up to 900 m deep (Monroe & Wicander 1997; Stanley 1999). The marine waters accumulated in bays, lagoons, and channeled estuaries, resulting in erosion and deposition of marine mud and sand (Baldridge 2004). These waters were filled with diverse marine organisms, including predaceous marine reptiles, sharks, advanced bony fish, and various invertebrates including mollusks, ammonites, belemnites, foraminiferans, and radiolarians (Monroe & Wicander 1997). Then in the late Cretaceous, the marine waters began to recede in a gradual but intermittent manner, characterized by punctuated and frequent reversals in relative sea level (Baldridge 2004). The eventual recession of one of these bodies, the Lewis Sea, marked the final retreat of the WIS from North America, ca. 74 Ma (Baldridge 2004), leaving more of the continent exposed than at most times in the Phanerozoic (Algeo & Seslavinsky 1995). Despite the final withdrawal of saltwater from the continent, the American southwest still differed markedly from its current condition. Fossil floras of the Eocene (ca. 58 – 35 Ma) suggest that the American Southwest was once dominated by wet summers and relatively stable temperatures (Malusa 1992). During the Pleistocene (ca.

1 1.6 – 0.01 Ma), numerous freshwater lakes formed, including Lake Bonneville (a pluvial lake covering much of the Great Basin region) and Lake Lahontan (an endorheic lake covering much of northwestern Nevada) as several of the larger and longer lasting examples (Hostetler & Giorgi 1992). Based on studies involving fossil pollen, as recently as 10 Ka the climate of the American southwest was considerably cooler and moister than today (Brune 1981). Then 7 Ka, the climate became drier; grasses, oaks and hickories came to predominate across the savannah (Brune 1981). Around 4.5 Ka, more drying had occurred, and mesquite, agave, and acacia appeared (Brune 1981). In recent centuries, the arrival of Spanish missionaries to the southwest resulted in trees being cut to provide lumber for barracks and homes, followed by the conversion of newly deforested land into pasture or croplands (Brune 1981). With the new system of grazing introduced, in which domesticated animals (primarily cattle) were confined to limited areas of pasture, grasses were overgrazed and many native species became extinct (Brune 1981). Analysis of long term data sets for plant demography and climatology, and the Holocene (10 Ka – present) fossil history of the northern Chihuahuan desert have confirmed the general shift from semidesert grassland to desert shrubland, based on climate change, overgrazing, or both (Buffington & Herbel 1965, Neilson 1986). During the last century, the desertification process has been accelerated as an unintended consequence of agricultural endeavors that were never well-suited to the modern soil composition and precipitation regime (Reisner 1993). Despite the arid surroundings, expanding cities with ever-increasing human populations have become common, which has required the use of groundwater to provide for human usage. Apart from the rivers still flowing through the modern desert southwest (and occasional lakes and ponds), the only aquatic habitats remaining are springs (in Spanish, cienegas). Springs are the spillways through which the overflow of surplus ground water passes (Stevens & Meretsky 2008). These spillways are typically natural openings in the rock or soil, although springs can also form from the coalescence of a large numbers of seeps (Brune 1981). These springs form aquatic habitats for faunas that include species derived from marine progenitors, which were stranded upon the recession of the WIS. Modern spring communities include flatworms, amphipods, isopods, snails, crayfish, insects, salamanders, frogs, turtles, and fish (Brune 1981). Various terrestrial vertebrates,

2 including bears, raccoons, snakes, and birds feed on these aquatic organisms, in addition to relying on springs for drinking water (Brune 1981). Consideration of this historical context helps explain the modern distribution of biota associated with these desert springs. Unfortunately, the loss of spring habitats by groundwater mining and habitat alterations (e.g. diversions, damming, dewatering, channelization) is a major threat to aquatic biodiversity in arid regions of the modern West (Glennon 2002), where isolated spring systems often harbor unique assemblages of narrowly endemic biota (Minckley & Unmack 2000; Hershler et al. 2002; Hershler & Sada 2002; Sada & Vinyard 2002; Sada et al. 2005). Due to the high endemism encountered at desert springs (Hershler et al. 1999; Echelle et al. 2005), we would expect the widespread occurrence of undescribed, endemic taxa of conservation concern in desert springs. For this dissertation, I have conducted a series of studies involving amphipods in the Gammarus pecos species complex, a group endemic to springs associated with the Pecos River in southeastern New Mexico and western Texas. My studies evaluated morphological, genetic, and ecophysiological relationships in this species complex and identified the mechanisms by which these entities evolved. The molecular genetic investigations undertaken in Chapter 1 are useful for detecting neutral genetic variation, although the prevailing assumption that mitochondrial markers are selectively neutral is not without some empirical and theoretical refutation (Ballard & Kreitman 1995; Nachman et al. 1996; Weinreich & Rand 2000). Chapter 2 focuses on the detection of species boundaries according to three methods which evaluate fragments of cytochrome c oxidase I (COI), the “DNA barcode” gene. The mechanism under study in Chapter 3 is selection for salinity tolerance, which has driven local adaptation and may have led to ecological speciation. In my dissertation, I investigated genetically neutral variation using mtDNA sequencing, and also a trait (tolerance to salinity) that is under selection. Studying both the neutral variation and a trait under selection among these amphipods helped elucidate the evolutionary relationships and affinities between the species detected within this species complex. The results of my investigations will be of great interest to conservation organizations such as the New Mexico Department of Game and Fish and the United

3 States Fish and Wildlife Service. Due to conservation concerns about these organisms, my conclusions will help to inform conservation planning for numerous other organisms endemic to the same habitats.

4 Literature Cited

Algeo TJ, Seslavinsky KB. 1995. The paleozoic world: continental flooding, hypsometry, and sealevel. American Journal of Science 295: 787-822.

Baldridge WS. 2004. Geology of the American Southwest: A Journey Through Two Billion Years of Plate-tectonic History. Cambridge University Press, UK.

Ballard JWO, Kreitman M. 1995. Is mitochondrial DNA a strictly neutral marker? Trends in Ecology & Evolution 12: 485-488.

Brune G. 1981. Springs of Texas, vol. 1. Branch-Smith Inc., Fort Worth, Texas.

Buffington LC, Herbel CH. 1965. Vegetational changes on a semidesert grassland range from 1853 to 1963. Ecological Monographs 35: 139-164.

Echelle AA, Carson EW, Echelle AF, Van Den Bussche RA, Dowling TE, Meyer A. 2005. Historical biogeography of the new-world pupfish genus Cyprinodon (Teleostei: Cyprinodontidae). Copeia 2005: 320-339.

Glennon R. 2002. Water Follies: Groundwater Pumping and the Fate of America’s Fresh Waters. Island Press, Washington, D.C.

Hershler R, Liu H-P, Mulvey M. 1999. Phylogenetic relationships within the aquatic snail genus Tryonia : implications for biogeography of the North American Southwest. Molecular Phylogenetics and Evolution 13: 377-391.

Hershler R, Liu H-P, Stockwell CA. 2002. A new genus and species of aquatic gastropods (Rissooidea: Hydrobiidae) from the North American Southwest: phylogenetic relationships and biogeography. Proceedings of the Biological Society of Washington 115: 171-188.

Hershler R, Sada DW. 2002. Biogeography of Great Basin aquatic snails of the genus Pyrgulopsis . In: Great Basin Aquatic Systems History (eds., Hershler R, Madsen DB & Currey), pp. 255-276, Smithsonian Contributions to Earth Sciences, 33.

Hostetler SB, Giorgi F. 1992. Use of a regional atmospheric model to simulate lake- atmosphere feedbacks associated with pleistocene lakes Lahontan and Bonneville. Climate Dynamics 7: 39-44.

Malusa J. 1992. Phylogeny and biogeography of the pinyon pines ( Pinus subsect. Cembroides). Systematic Botany 17: 42-66.

Minckley WL, Unmack PJ. 2000. Western springs: their faunas and threats to their existence. In: Freshwater Ecoregions of North America: A Conservation Assessment (Abell RA et al., eds.), pp. 52-53. Island Press, Washington, D.C.

5 Monroe JS, Wicander R. 1997. The Changing Earth: Exploring Geology and Evolution, 2nd Ed. West Publishing Company, Belmont.

Nachman MW, Brown WM, Stoneking M, Aquadro CR. 1996. Nonneutral mitochondrial DNA variation in humans and chimpanzees. Genetics 142: 953-963.

Neilson RP. 1986. High-resolution climatic analysis and southwest biogeography. Science 232: 27-34.

Reisner M. 1993. Cadillac Desert: The American West and its Disappearing Water, Revised Ed. Penguin Books, New York.

Sada DW, Vinyard GL. 2002. Anthropogenic changes in biogeography of Great Basin aquatic biota. In: Great Basin Aquatic Systems History (eds., Hershler R, Madsen DB and Currey DR), pp. 255-276. Smithsonian Contributions to Earth Sciences, 33.

Sada DW, Fleishman E, Murphy DD. 2005. Associations among spring-dependent aquatic assemblages and environmental and land use gradients in a Mojave Desert mountain range. Diversity and Distributions 11: 91-99.

Stanley SM. 1999. Earth System History. W.H. Freeman and Company, New York.

Stevens LE, Meretsky VJ. (eds) 2008. Aridland Springs in North America: Ecology and Conservation. University of Arizona Press, Tucson, AZ.

Weinreich DM, Rand DM. 2000. Contrasting patterns of nonneutral evolution in proteins encoded in nuclear and mitochondrial genomes. Genetics 156: 385-399.

6 Chapter 1 1

Phylogeographic analysis reveals multiple cryptic species of amphipods (Crustacea: Amphipoda) in Chihuahuan Desert springs

ABSTRACT

Biodiversity conservation and the identification of conservation units among invertebrates are complicated by low levels of morphological difference, particularly among aquatic taxa. Accordingly, biodiversity is often underestimated in communities of aquatic invertebrates, as revealed by high genetic divergence between cryptic species. I analyzed PCR-amplified portions of the mitochondrial cytochrome c oxidase I (COI) gene and 16S rRNA gene for amphipods in the Gammarus pecos species complex endemic to springs in the Chihuahuan Desert of southeast New Mexico and west Texas. My analyses uncover the presence of seven separate species in this complex, of which only three nominal taxa are formally described. The distribution of these species is highly correlated with geography, with many present only in one spring or one spatially- restricted cluster of springs, indicating that each species likely merits protection under the U.S. Endangered Species Act. I present evidence suggesting that habitat fragmentation, long-distance colonization, and isolation-by-distance have occurred at different temporal and spatial scales within this system to produce the lineages that we report. I show that patterns detected in the G. pecos species complex also correlate with endemic fishes (Gambusia spp., pupfish) and hydrobiid snails. My results provide clues important for future biodiversity investigations in geographically isolated aquatic habitats, and shed light on the understudied and underestimated levels of biodiversity present in desert spring systems.

1 Manuscript published : Seidel RA , BK Lang, and DJ Berg. 2009. Phylogeographic analysis reveals multiple cryptic species of amphipods (Crustacea: Amphipoda) in Chihuahuan Desert springs. Biological Conservation 142: 2303-2313.

7 INTRODUCTION

Biodiversity conservation relies heavily on the identification and description of units of conservation. The determination of these minimal units brings the field of conservation biology into close association with systematic biology, a discipline that seeks to discover monophyletic groups at higher taxonomic levels, and to delineate distinct lineages at lower levels (Dimmick et al., 1999; Wheeler and Meier, 2000). The identification of discrete species as units of conservation continues to present challenges for investigators, but recent approaches represent improvement in analytical detection of discontinuous evolutionary groups. While natural processes including speciation and extinction produce relatively deep divergences between species (depicted by longer branch lengths between species on phylogenetic trees), relatively recent coalescence processes produce a multitude of shallow branches at the population level (Pons et al., 2006). Recently Pons et al. (2006) emphasized the logic and practical application of combining the approaches of phylogenetics and population genetics, the two programs which investigate biological relationships above and below the species boundary (Brower et al., 1996). In conservation practice, genetic investigations have helped identify appropriate conservation units, including geographical locations where appropriate management actions would be most effectively implemented in organisms as varied as Komodo dragons (Ciofi et al., 1999) and southwest Australian plants (Coates, 2000). One practical challenge in assessing biodiversity involves the difficulty of identifying the units of diversity in the field, which undermines the reliability of estimates of distribution (McNeely et al., 1990). The challenge becomes even more acute when the task focuses on aquatic invertebrates, which often display low levels of morphological distinctiveness (Müller, 2000; Pfenninger et al., 2003; Witt et al., 2003). This might be because the actual cues used by aquatic taxa for conspecific recognition may not involve the same morphological characters used by taxonomists for species determination (Knowlton, 1993). Given such low levels of morphological difference, we would expect traditional taxonomy to underestimate marine and freshwater biodiversity (Thorpe and Solé-Cava, 1994; Gómez et al., 2002). With the advent of molecular techniques, conservation biologists now have additional means for discovering diagnostic characters in organisms that are indistinguishable based on morphology alone. Molecular genetic

8 techniques have revealed substantial hidden diversity within morphologically delimited species (Remerie et al., 2006), and unusually high levels of genetic divergence between cryptic species (Bucklin et al., 1995; Knowlton and Weigt, 1998; Lee, 2000). Identification of species boundaries is particularly crucial in situations involving endangered species assessments. For 38 recent endangered species petitions in the United States, 81% of those showing genetic distinction were granted protection status (Fallon, 2007), which underscores the importance of using genetic markers to reveal important differences among morphologically similar taxa, particularly among aquatic invertebrates. Amphipods comprising the Gammarus pecos species complex (Cole, 1985) are endemic to spring systems associated with the Pecos River of New Mexico and Texas (Figure 1). It has been hypothesized that these freshwater amphipods are derived from a broadly distributed marine progenitor that became isolated inland upon the recession of the Western Interior Seaway from the North American continent during the Late Cretaceous (Bousfield, 1958; Holsinger, 1976; Baldridge, 2004). Members of this complex likely speciated in response to diverse ecological conditions that developed in various aquatic environments differing in elevation, substrate mineral composition, drainage patterns, and local hydrochemical conditions. This complex consists of three nominal species ( Gammarus pecos Cole and Bousfield, 1970; Gammarus desperatus Cole, 1981; and Gammarus hyalelloides Cole, 1976) differentiated by morphology, at least six populations of undetermined taxonomic affinity that may represent several undescribed species, and at least two other populations presumed to be extirpated (Cole and Bousfield, 1970; Cole, 1976; Cole, 1981; Cole, 1985). All current species designations (Table 1) based on earlier morphological findings are treated as hypotheses which will be evaluated based upon the genetic results we present. Presently, this group of endemic amphipods is confronted with a high rate of imperilment related to habitat modification and groundwater withdrawal (Lang et al., 2003). Loss of spring habitat by groundwater mining and habitat alterations (e.g. diversions, damming, dewatering, channelization) is a major threat to aquatic biodiversity in arid regions of the western United States (Glennon, 2002), where isolated spring systems often harbor unique assemblages of narrowly endemic biota (Minckley and Unmack, 2000; Hershler and

9 Sada, 2002; Sada and Vinyard, 2002; Sada et al., 2005). In this study, I combined a model-based tree estimation method with a population genetic analysis to investigate the phylogeography and species composition of amphipods comprising the Gammarus pecos species complex. I employed the Wiens and Penkrot (WP) method to detect species using a general tree-based inference protocol (Wiens and Penkrot, 2002: fig. 1) by constructing a phylogenetic tree based upon the mitochondrial cytochrome c oxidase subunit I (COI) and the 16S rRNA (16S) gene fragments. My objective in this study was to evaluate these high resolution genetic data to clarify the number of species present in a faunal group of conservation concern. While a previous investigation of this species complex using allozymes allowed some degree of resolution in detecting differences among amphipod populations (Gervasio et al., 2004), our sequencing of the partial mitochondrial genome provides greatly increased resolution, revealing the presence of previously undetected species that may merit conservation action. Because similar processes of geographic isolation following diversification are likely to operate on other Chihuahuan Desert aquatic biota (Sei et al., 2009), our results are likely to predict biogeographic patterns for many other taxa. A key motivation for identifying species in the G. pecos species complex relates to the protections available under the United States Endangered Species Act (ESA) of 1973. The ESA is a federal law that was originally designed to prevent the extinction and to assist in the recovery of the rarest creatures on Earth and particularly those in the United States (Vaughan, 1994). Under this federal statute, species are listed as “endangered” or “threatened.” Listing occurs solely on the basis of the best scientific and commercial data available (16 U.S.C. § 1533(b)(1)(A) ). Although the U.S. Congress stated that species should be protected based upon extinction risk without regard to taxonomic classification (U.S. Congress 1982), there is strong evidence that certain types of species are favored over others in the ESA listing process (Kareiva et al., 2006). For example, vertebrates comprise only 2 percent of U.S. biodiversity, but represent almost 30 percent of federally listed species (Kareiva et al., 2006). In contrast, invertebrates comprise 84 percent of the species composition of overall U.S. biodiversity, but only 14 percent of the listed species (Kareiva et al., 2006). The proportion of invertebrates protected under the ESA is also noticeably low considering that invertebrates eclipse all other biota in terms

10 of sheer numbers, species richness, and biomass, but also represent faunal elements important to functioning ecosystems (Black et al., 2001). Additionally, there is a striking contrast between relatively low U.S. expenditures for invertebrates compared to greater funds allocated for vertebrates (Black et al., 2001; Czech and Krausman, 2001). Finally, invertebrates may be more vulnerable than listed vertebrates because their smaller body size and shorter individual lifespans may make them more vulnerable to environmental fluctuations (Murphy et al., 1990). In the case of the Gammarus pecos complex, G. desperatus is listed as “endangered,” while G. pecos and G. hyalleloides are candidates for possible listing. MATERIALS AND METHODS I obtained amphipods from extant populations in the G. pecos species complex from 12 spring sites (Figure 1; Table 1) associated with the Pecos River basin in southeastern New Mexico and western Texas. Eleven of these spring sites are located in the Pecos River watershed of the Permian Basin (Cartwright, 1930), while Malpais Spring is located within the endorheic Tularosa Basin, west of the Permian Basin. Two populations sampled by Gerald Cole in the 1970s (Cole, 1981; Cole, 1985) are presumed extirpated, while we were unable to reconcile the locality of Cole’s population “M.” Hand nets were used to collect amphipods (30-200 animals per site) from the water column, macrophytes, or substrata. All samples were preserved in 95% ethanol. Complete genomic DNA was extracted from 10 - 40 individuals per spring site. DNA extraction involved dissecting either intact pleon or pereopods from each individual and followed a standard extraction protocol using proteinase K, ribonuclease, and several Promega ® reagents (Nuclei Lysis Solution, cat.# A7943; Protein Precipitation Solution, cat.# A7953). DNA extraction from limbs might in the future be done nondestructively, given that crustaceans can autotomize and regrow appendages, some relatively rapidly (Seidel, 2005). A 680-bp region of the COI gene was amplified using the primers LCO1490: GGTCAACAAATCATAAAGATATTGG and a shortened version of HCO2198: TCAGGGTGACCAAAAAATCA (Folmer et al., 1994). Also for COI, the custom internal primers CBL4f: GTGAAGAGAGAAAATAGCTA and CBL4r: ATYATAATTGGGGGGTTC were developed for use in amplifying shorter COI gene fragments when the Folmer et al. (1994) “universal” primers failed to produce

11 amplification of the full 680-bp fragment. Each 50-µL polymerase chain reaction (PCR) contained 25 µL Taq PCR Master Mix (3.7 U Taq DNA polymerase, 1.5 mM MgCl 2 and 200 µM each dNTP; Qiagen ® cat.# 201443), 0.05 nmol each primer, ca. 2 mM additional

MgCl 2, 5 ng DNA template and 6.5 µL molecular water. PCR conditions consisted of 3 min at 94 oC followed by 5 cycles of 1 min at 94 oC, 1 min at 45 oC, 1 min at 72 oC; followed by 35 cycles of 1 min at 94 oC, 1 min at 50 oC, 1 min at 72 oC; followed by 5 min at 72 oC. COI gene products were isolated using electrophoresis in a 2% agarose gel and extracted using Qiagen ® QIAquick Spin Columns and buffers (cat.# 28106). Cycle sequencing reactions were performed using ABI BigDye terminator v3.1 sequencing kits (25 cycles, annealing temperature: 50 oC). PCR products were sequenced in both directions using the PCR primers mentioned above and an ABI 3130 automated sequencer (Applied Biosystems). Forward and reverse sequences were aligned using BioEdit (Hall, 1999) to verify basecall accuracy. While the final COI alignment for nine of the focal populations contained sequences 620-bp in length, the populations BLBC, BLSS, and BLU6 (those requiring the custom internal primers) produced COI sequences 236-bp in length. For the 16S rRNA gene, a 480-bp region was amplified using the primers 16STf: GGTAWHYTRACYGTGCTAAG (Macdonald et al., 2005) and 16Sbr: CCGGTTTGAACTCAGATCATGT (Palumbi and Benzei, 1991). The 50-µL PCR reactions contained the same reagent quantities as listed above for the COI gene. PCR conditions for 16S consisted of 4 min at 95 oC followed by 40 cycles of 1 min at 95 oC, 1 min at 42 oC, 2.5 min at 72 oC; followed by 7 min at 72 oC. PCR products for 16S were sequenced in both directions and aligned as for COI. The final 16S alignment for all individuals contained sequences 476-bp in length. GenBank BLAST searches were used to verify the homology of COI and 16S gene sequences used in this study. Combination mitochondrial sequences were assembled by joining the COI and 16S gene sequences end-to-end for all animals that sequenced successfully for both genes. I calculated descriptive statistics of genetic diversity (number of unique and shared haplotypes, mean number of pairwise differences) within each population using Arlequin v2.000 (Schneider et al., 2000). Divergence of populations was measured by

calculating F ST for pairwise combinations of populations, using the distance method

12 rather than haplotype frequencies. I measured geographic distances, distances along permanent and ephemeral streams connecting springs sites, for all pairs of populations and tested for isolation-by-distance by examining the correlation of pairwise geographic distance and pairwise F ST using a Mantel test (Sokal, 1979; Sharbel et al., 2000). I used the program TCS v1.18 (Clement et al., 2000) to generate networks for clusters of haplotypes separated by 19 mutational steps or fewer. Given the high sequence divergence between some of the populations, a single unified haplotype network was not possible using only TCS. To determine the fewest number of hypothesized intermediates necessary for connecting the various subnetworks into a single grand network, the most common haplotypes from each subnetwork were identified and analyzed separately using Arlequin v2.000 (Schneider et al., 2000). The shortest calculated paths of hypothesized intermediates were then added manually to the grand network. I performed a nested clade analysis using ANeCA (Panchal, 2007), a fully automated implementation of Nested Clade Phylogeographic Analysis (NCPA) described by Templeton et al. (1995), which incorporates both TCS v1.18 (Clement et al., 2000) and GeoDis v2.2 (Posada et al., 2000) as part of the automation. To detect distinct species using the WP method, I constructed a maximum- likelihood (ML) tree using Modeltest 3.7 coupled with PAUP* 4.0b (Swofford, 2000; Posada, 2003). I applied the hierarchicial-likelihood-ratio test (hLRT) to calculate the best-fit model of nucleotide substitution for the concatenated mitochondrial sequences in the complete alignment. The ML analyses were performed using PAUP 4.0b with the appropriate substitution model (HKY + G) selected by Modeltest 3.7, based on heuristic searches with the tree-bisection-reconnection (TBR) branch-swapping algorithm. After 274,993 rearrangements, the single best ML tree was saved by PAUP* (-ln L = 4606.23843), which served as the starting tree for the node bootstrap analysis (100 replicates) performed in TreeFinder (Jobb et al., 2004). The tree was rooted using a corresponding concatenated sequence from Gammarus oceanicus obtained from GenBank (COI: AY926674; 16S: AY926728). RESULTS I recovered combination sequences for 134 individuals from 12 populations of Gammarus . A total of 91 haplotypes were identified; the number of haplotypes per site

13 ranged from 3 to 34, with all but two of these being confined to a single site (Table 2). Where haplotypes were shared between spring sites (BLBC and BLSS; ESS and GS), the shared haplotype was also the most common one in both populations. Haplotype richness was correlated with sample size ( r = 0.964, n = 12, p < 0.001). The Mantel test between geographic distance (km) and pairwise population FST showed a strong positive correlation between the two matrices ( r = 0.834, P = 0.001, for log-transformed data;

Figure 2). The calculated F ST values (most significantly > 0) for all possible pairs of populations ranged from 0.001 to 0.989 (arithmetic mean = 0.882; Table 3). After analyzing separately the COI and 16S fragments comprising the combined haplotypes, we report COI within-species diversities of 1.2% (mean) and among-species divergences of 15.5% (mean). For 16S, we report within-species diversities of 0.8% (mean) and among species divergences of 12.4% (mean). I estimated the genealogical relationships among the 134 combined sequences following Templeton et al. (1992) and displayed the structure of those relationships using a haplotype network (Figure 3). This network visually depicts the relationships among all combined sequences described on a spring-by-spring basis and shows a strong association of diversity and divergence based on spring location (Table 2). The distinctive nature of the haplotype clusters found in each spring is visually apparent in the haplotype network (Figure 3). Of the 12 springs, eight contain genetic groups that share

no haplotypes with other springs in the study area. The pairwise F ST estimates among these eight springs are all well above F ST = 0.25 (Table 3), indicating very great differentiation (Hartl and Clark, 2005). The two remaining pairs of springs (BLBC and

BLSS; ESS and GS) display much lower F ST values. Analysis of Molecular Variance, AMOVA (Excoffier et al., 1992), performed on these groups using Arlequin (Schneider et al., 2000), indicated that 93.9% of total genetic variation was among groups, 1.4% was among populations within groups (species 1 and 5), and 4.7% was within populations. Nested clade analysis testing the null hypothesis of no geographical association of haplotypes revealed the most likely historical processes that have resulted in the geographical distribution of the haplotypes described in our study. Most population comparisons resulted in the failure to reject the null based on the absence of statistically significant distances within clades. However, based on the GeoDis v2.2 (Posada et al.,

14 2000) inference key, allopatric fragmentation likely accounts for the present genetic structure at ESS, GS, and SSS. For MS and SB, the inference key indicated long- distance colonization or past fragmentation, although these two processes are not mutually exclusive. The ML tree used to detect species according to the dichotomous key for the WP method showed two major clades, indicated as A and B in Figure 4. Within these major clades, we detected at least seven separate species in this complex: 1) BLBC, BLSS, and BLU6; 2) BLHM; 3) CS; 4) DY; 5) ESS, GS, and SSS; 6) MS and SB; 7) PL; all were consistent with the WP definition of species as exclusive clades without gene flow (Wiens and Penkrot, 2002: fig. 1c). DISCUSSION Speciation is not always reflected by morphological change and therefore, it is very important to detect cryptic species for accurate biodiversity estimation and effective conservation planning (Bickford et al., 2007). The intraspecific diversity and interspecific divergence levels we report are similar in magnitude to corresponding genetic differences reported for amphipod species from other systems (COI: Meyran et al., 1997; Meyran and Taberlet, 1998; Müller, 2000; Witt and Hebert, 2000; Kelly et al., 2006; Rock et al., 2007; 16S: France and Kocher, 1996; Quan et al., 2001). Both COI and 16S genes have demonstrated utility in phylogeographic investigations of other crustaceans as well (Stillman and Reeb, 2001; Quan et al., 2004, MacDonald et al., 2005).

The strong positive correlation between F ST and river distance is consistent with allopatric speciation of the G. pecos complex via isolation-by-distance (Vrijenhoek,

1998). The shape of the F ST vs River Distance plot indicates fixation even at relatively modest distances of population separation. This is not a surprise considering the relatively weak passive dispersal of amphipods (Thomas et al. 1998; Bilton et al., 2001) and the extreme geographic isolation of most springs (Hershler et al., 1999). Weak dispersal and extreme habitat isolation both contribute to reduced gene flow, leading to population divergence based upon the effects of isolation, genetic drift, and selection (Lesica and Allendorf, 1995). Under these conditions, high fixation is expected even at relatively close geographic locations due to dispersal difficulty across hot desert terrain.

15 Furthermore, the general shape of the plot in Figure 2 is most consistent with Case IV described by Hutchison and Templeton (1999), in which gene flow is more effective at shorter distances of geographic separation and drift is more influential at greater distances among populations. The majority of F ST values are greater than 0.9, indicating a high degree of differentiation (Hartl and Clark, 2005) and the strong likelihood of continued

divergence over time (Lowe et al., 2004). Furthermore, high F ST values indicate reduced or absent gene flow and often reveal the presence of distinct species or subspecies (Hogg et al., 2000). Thus, it is likely that most of these populations have been isolated for substantial periods of time. The genetic data for populations from ESS and GS indicate these amphipods are very similar to SSS, even though SSS and DY are referable to Gammarus pecos (Cole

1985). The F ST values for SSS vs. ESS and SSS vs. GS are much lower than DY vs. ESS and DY vs. GS. Based on these data and evidence from the haplotype network, SSS is far more genetically similar to ESS and GS than to “conspecific” DY. These data are consistent with the spatial clustering of ESS, GS and SSS in the Toyah Basin, all of which are considerably closer to one another than to DY, which is located approximately 86 km east of the Toyah Basin springs. We conclude that Gammarus pecos is restricted to the Diamond Y Spring system (the type locality), whereas the Toyah Basin harbors an undescribed gammarid in ESS, SSS, GS and Gammarus hyalleloides from Phantom Lake Spring. The undescribed gammarid at CS also shows distinctiveness from all other populations in the study region. All of these affinities are consistent with the nested- clade results, which implicated the processes of allopatric fragmentation for ESS, GS, and SSS, and long-distance colonization or past fragmentation for MS and SB. The affinities of BLBC, BLSS, and BLU6 are consistent with recent gene flow, which makes sense given the close spatial proximity of these springs. The results of our nested clade analysis indicate biological processes completely consistent with the topology of the ML tree. While all seven clades we identified as distinct species showed exclusivity and lack of gene flow, the case of species 6 (MS and SB) warrants further evaluation. While incipient speciation could be occurring between MS and SB based upon the weak separation visible at the bottom of the ML tree in Figure 4, we interpret this separation as

16 insufficient evidence for separate species status using the WP method. The pairwise F ST for MS and SB was 0.332. While indicative of partial differentiation, this F ST is further evidence that incipient speciation is likely occurring, but insufficient time has passed for complete differentiation as detected between the other six species in our study. The major division on the ML tree between clade A and clade B is consistent with the geography of southeastern New Mexico and west Texas. The Sacramento Mountains (elevation: ca. 1400 – 2955 m) begin a mountain range which passes south between Malpais Spring to the west and the Bitter Lake springs to the east. This range continues southeast into the Guadeloupe Mountains (elevation: ca. 1300 – 2667 m), in which Sitting Bull Spring is located at an elevation of 1540 m. These mountains form a distinct biogeographic boundary separating SB and MS from the lower elevation springs at Bitter Lake (elevation: ca. 1065 m), the Toyah Basin (GS, ESS, PL and SSS, elevation: ca. 1050 m), DY (elevation: ca. 840 m), and CS (elevation: ca. 746). While MS is at an elevation of ca. 1263 m, it is located in the endorheic Tularosa Basin, west of the mountain range, with no surface water flow to any other water body. Thus, a deep phylogenetic division separating SB and MS (collectively species 6) from the other species is consistent with isolation. Gammarus desperatus , a state and federal endangered species, is present on Bitter Lake National Wildlife Refuge at BLBC and BLSS. We interpret the BLU6 population to be conspecific with BLBC and BLSS, due to lack of exclusivity on the ML tree and relatively small separation (3 hypothesized intermediates) between BLU6 and the pair

BLBC & BLSS on the network. Pairwise F ST values for BLU6 vs. BLBC and BLU6 vs. BLSS indicate far less differentiation and more recent gene flow relative to all pairs among the remaining populations. The amphipod present at BLHM is nominally considered G. desperatus (Federal Register, 2005), but given the F ST values of BLHM compared to other Bitter Lake populations (all ≥ 0.889), we interpret these data to support the exclusivity of BLHM, indicating that it is a different species than those present at the other Bitter Lake sites. This interpretation is supported by the haplotype network, which shows ≥ 48 hypothesized intermediates between BLHM and other populations. While current management of Bitter Lake National Wildlife Refuge amphipods assumes only the presence of G. desperatus , our results indicate the presence of at least two species

17 within the relatively close confines of the refuge. The presence of multiple distinct lineages in such a confined geographical area (Bitter Lake sites are all less than 11 km apart) is consistent with multiple colonizations of this landscape. The results from the nested clade analysis which indicate fragmentation, long distance colonization, or both, are consistent with the conclusions presented above, and also the geological history of the region during which seawater (late Cretaceous) and Pleistocene pluvial lakes provided the requisite dispersal routes between present day spring locations (Baldridge, 2004). The lower portion of the Pecos River dates from the early Tertiary, while the modern upper Pecos originated much more recently (Echelle and Echelle, 1978). When the San Juan Mountains (southwestern Colorado) uplifted in the early Tertiary, the ancestral Rio Grande-Pecos River was formed, which flowed southeast across New Mexico until it emptied into the present lower Pecos (Thomas, 1972; Belcher, 1975). During the ensuing Miocene, uplift of the Sangre de Cristo Mountains shifted the northern range of the drainage, causing the headwaters of the Rio Grande to flow west, possibly into the Tularosa Basin (Thomas, 1972; Belcher, 1975). At this time, Belcher (1975) suggests that the Rio Grande might have flowed across the present day range of the Guadalupe Mountains, emptying into the lower Pecos. The similarity between MS (Tularosa Basin) and SB (Guadalupe Mountains) could be explained by long-distance colonization along this water route. However, the overland connection from the Tularosa Basin through to the lower Pecos would have ended upon the Pliocene uplift of the Guadalupe Mountains (Belcher, 1975). Moritz (1994) proposed the combination of sampling nuclear genes and mitochondrial genes to identify significant genetic differences among populations, an approach particularly useful in clarifying relationships that lie close to the boundary between conspecific populations and distinct species. All of our provisional species are exclusive given our mitochondrial data, and most were shown to exhibit significant divergence in nuclear allele frequencies at allozyme loci (Gervasio et al., 2004). We recommend that the species present at these seven locations (or location sets, in the cases of species 1, 5, and 6) be managed as separate units of conservation, due to their discreteness and distinctness, both genetically and geographically. While each of these species is locally abundant, each is also endemic to a single spring system or a network of

18 geographically proximate spring systems. While several species within this complex were originally described morphologically (Cole and Bousfield, 1970; Cole, 1976; Cole, 1981), a variety of crustacean studies have used molecular evidence as verification or refutation of earlier morphology-based conclusions about species identity (Sotela et al., 2008). Recent molecular studies have both verified morphological taxonomies in crustaceans (Geller et al., 1997, Mathews et al., 2002, Sotela et al., 2008), and refuted them (Tsoi et al., 2005; Reuschel and Schubart, 2006; Cook et al., 2006). Even though we assembled combined sequences by linking fragments for the generally more conservative 16S gene with COI fragments, the provisional species we propose show divergences in the ranges of those found in these other studies. The use of a concatenated analysis, in which sequences from multiple regions of a genome are combined, has shown utility in investigations of both animal and plant phylogenies (Pitra et al., 2000; Zhang et al., 2005). To test for congruence between the data partitions, we constructed separate networks for the complete COI data set (n = 163 COI sequences) and 16S data set (n = 172 16S sequences), before making the decision to perform the concatenated analysis and display results for that. Networks constructed from mtDNA haplotypes have shown considerable and increasing utility in revealing the geographical distribution of geneological lineages (Emerson and Hewitt, 2005), and have been used to elucidate the relationships among congeneric plants (Londo et al. 2006) and animals (Shaw 1999; Larson et al. 2005). Our analysis of separate COI and 16S networks (not shown) were largely congruent, with only the minor exception of the BLHM haplotypes, which clustered together with BLBC, BLSS, and BLU6 on the 16S network, but resolved as separate from the other three Bitter Lake populations on the COI network. Accordingly, we present the combined analyses as the closest representation of the actual species phylogeny, based on data from the partial mitochondrial genome. My detection of relatively high levels of cryptic diversity informs the larger discussion among conservation biologists and aquatic ecologists about the proportion of biodiversity harbored in freshwater systems. It has been suggested that freshwater biodiversity has been underestimated due to several factors, including relatively less research effort directed towards aquatic invertebrates (Strayer, 2006), unexplored and

19 poorly known groundwater habitats (Strayer, 2006), and the common presence of morphologically cryptic species in aquatic ecosystems (Lee and Frost, 2002; Witt et al., 2003; Lefébure et al., 2006). Furthermore, freshwater spring habitats occurring in hot deserts support a disproportionate level of species diversity (Minckley and Unmack, 2000) and high genetic diversity (Thomas et al., 1997; Thomas et al., 1998; Witt et al., 2003; Witt et al., 2006) due to extreme habitat patchiness, stable environmental conditions, and long-term isolation (Thomas et al., 1998; Minckley and Unmack, 2000). Even with reports pointing towards high diversity in desert springs, most of the diversity found in these geographically isolated habitats is not well documented. In the case of native fauna in the western United States, species dwelling in spring and spring-brook habitats are also in greater danger of extinction than organisms associated with more hydrologically integrated habitat types (Sada and Vinyard, 2002). Despite insufficient research effort directed at the arid landscapes typical of hot deserts, we know that spring systems within these regions are geographically isolated and that a single system can often be the only remaining habitat for endemics like hydrobiid snails (Hershler et al., 1999; Hershler and Sada, 2002) and pupfishes (Echelle et al., 2005). Our findings are consistent with those of Stevens and Meretsky (2008), who conclude that aridland springs rank among those habitats with the greatest structural complexity, biodiversity, productivity, and evolutionary potential, and at the same time are among the most threatened systems. I report very distinct genetic patterns which are likely to be shared among diverse aquatic taxa distributed across desert landscapes. These results further underscore the increasing role of phylogeographic analysis in studies of regional biogeography (Crews and Hedin, 2006). Considerable diversity, distributed in discrete habitats, has been reported for odonates (K. Gaines, pers. comm.), ostracodes (A. Smith, pers. comm.), and hydrobiid snails (Hershler et al., 2002) from the northern Chihuahuan Desert. Farther south in the Chihuahuan Desert, at Cuatro Ciénegas in Coahuila, México, a spring complex supports at least 70 endemic aquatic vertebrates, diverse microbes (Souza et al., 2006), endemic snails (Moline et al., 2004), and rare living stromatolites (Dinger et al., 2006). Given the cryptic diversity reported among aquatic invertebrates (Müller, 2000; Pfenninger et al., 2003; Cook et al., 2008), coupled with the high endemism encountered

20 at desert springs (Hershler et al., 1999; Echelle et al., 2005), we would expect the widespread occurrence of undescribed, endemic taxa of conservation concern in desert springs. The patterns we present can guide future biodiversity assessment efforts, based on the expectation of similar trends among diverse taxa. One such example is the correlation of our mtDNA results and the allozyme data (Gervasio et al., 2004) for the G. pecos complex, with genetic data for the endemic fish Gambusia pecosensis (Echelle et al., 1989), showing that in habitats where these amphipods and fish co-occur, they are similarly diverse, yet distinct from other lineages occurring nearby. We also note similar patterns comparing our genetic data to those for cyprinid pupfishes (Echelle et al., 1987) and hydrobiid snails (Hershler et al., 1999), considering their similar patches of high genetic diversity restricted to discrete and isolated habitats. Future biodiversity investigations in the Chihuahuan Desert will almost certainly uncover species meriting protection among the region’s highly endemic fauna, given previously undetected variation and high degree of crypsis among invertebrates (Remerie et al., 2006), particularly within aquatic habitats (Müller, 2000; Pfenninger et al., 2003; Witt et al., 2003) and those specifically located in xeric regions (Thomas et al., 1998). Given the heightened extinction risk driven by global climate change (Thomas et al., 2004), the planet’s desert biota rank among those members expected to suffer in the coming century (Sala et al., 2000). Thus on a global scale, deserts deserve biodiversity reevaluations, based on the patterns revealed in the northern Chihuahuan Desert.

21 LITERATURE CITED

Baldridge, W.S., 2004. Geology of the American Southwest: A Journey Through Two Billion Years of Plate-tectonic History. Cambridge University Press, UK.

Belcher, R.C., 1975. The geomorphic evolution of the Rio Grande. Baylor Geological Studies Bulletin 29, 1-64.

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

Bilton, D.T., Freeland, J.R., Okamura, B., 2001. Dispersal in freshwater invertebrates. Annual Review of Ecology and Systematics 32, 159-181.

Black, S.H., Shepard, M., Allen, M.M., 2001. Endangered invertebrates: the case for greater attention to invertebrate conservation. Endangered Species UDPATE 18, 42-50.

Bousfield, E.L., 1958. Fresh-water amphipod crustaceans of glaciated North America. The Canadian Field-Naturalist 72, 55-113.

Brower, A.V.Z., DeSalle, R., Vogler, A., 1996. Gene trees, species trees, and systematics. Annual Review of Ecology and Systematics 27, 423-450.

Bucklin, A., Frost, B.W., Kocker, T.D., 1995. Molecular systematics of seven species of Calanus and three species of Metridia (Calanoida; Copepoda). Marine Biology 121, 655- 664.

Cartwright, L.D., 1930. Transverse section of Permian Basin, west Texas and southeast New Mexico. Bulletin of the American Association of Petroleum Geologists 14, 969-981.

Ciofi, C., Beaumont, M.A., Swingland, I.R., Bruford, M.W., 1999. Genetic divergence and units for conservation in the Komodo dragon Varanus komodoensis . Proceeding of the Royal Society of London B 266, 2269-2274.

Clement, M., Posada, D., Crandall, K.A., 2000. TCS: a computer program to estimate gene genealogies. Molecular Ecology 9, 1657-1660.

Coates, D.J., 2000. Defining conservation units in a rich and fragmented flora: implications for the management of genetic resources and evolutionary processes in south-west Australian plants. Australian Journal of Botany 48, 329-339.

Cole, G.A., Bousfield, E.L., 1970. A new freshwater Gammarus (Crustacea: Amphipoda) from western Texas. American Midland Naturalist 83, 89-95.

Cole, G.A., 1976. A new amphipod crustacean, Gammarus hyalelloides n. sp., from

22 Texas. Transactions of the American Microscopical Society 89, 80-85.

Cole, G.A., 1981. Gammarus desperatus , a new species from New Mexico (Crustacea: Amphipoda). Hydrobiologia 76, 27-32.

Cole, G.A., 1985. Analysis of the Gammarus-pecos complex (Crustacea: Amphipoda) in Texas and New Mexico, USA. Journal of the Arizona-Nevada Academy of Science 20, 93-103.

Cook, B.D., Baker, A.M., Page, T.J., Grant, S.C., Fawcett, J.H., Hurwood, D.A., Hughes, J.M., 2006. Biogeographic history of an Australian freshwater shrimp, Paratya australiensis (Atyidae): the role life history transition in phylogeographic diversification. Molecular Ecology 15, 1083-1093.

Cook, B.D., Page, T.J., Hughes, J.M., 2008. Importance of cryptic species for identifying 'representative' units of biodiversity for freshwater conservation. Biological Conservation 141, 2821-2831.

Crews, S.C., Hedin, M., 2006. Studies of morphological and molecular phylogenetic divergence in spiders (Araneae: Homalonychus ) from the American southwest, including divergence along the Baja California Peninsula. Molecular Phylogenetics and Evolution 38, 470-487.

Czech, B., Krausman P.R., 2001. The Endangered Species Act: History, Conservation Biology, and Public Policy. The Johns Hopkins University Press. Baltimore, Maryland.

Dimmick, W.W., Ghedotti, M.J., Grose, M.J., Maglia, A.M., Meinhardt, D.J., Pennock, D.S., 1999. The importance of systematic biology in defining units of conservation. Conservation Biology 13, 653-660.

Dinger, E.C., Hendrickson, D.A., Winsborough, B.M., Marks, J.C., 2006. Role of fish in structuring invertebrates on stromatolites in Cuatro Ciénegas, México. Hydrobiologia 563, 407-420.

Echelle, A.A., Echelle, A.F., 1978. The Pecos River pupfish, Cyprinodon pecosensus n. sp. (Cyprinodontidae), with comments on its evolutionary origin. Copeia 1978, 569-582.

Echelle, A.A., Echelle, A.F., Edds, D.R., 1987. Population structure of four pupfish species (Cyprinodontidae: Cyprinodon ) fromt the Chihuahuan Desert region of New Mexico and Texas: allozymic variation. Copeia 1987, 668-681.

Echelle, A.A., Echelle, A.F., Edds, D.R., 1989. Conservation genetics of a spring- dwelling desert fish, the Pecos gambusia ( Gambusia nobilis , Poeciliidae). Conservation Biology 3, 159-169.

Echelle, A.A., Carson, E.W., Echelle, A.F., Van Den Bussche, R.A., Dowling, T.E.,

23 Meyer, A., 2005. Historical biogeography of the new-world pupfish genus Cyprinodon (Teleostei: Cyprinodontidae). Copeia 2005, 320-339.

Emerson, B.C., Hewitt, G.M., 2005. Phylogeography. Current Biology 15, R367-R371.

Excoffier L, Smouse P, Quattro J. 1992. Analysis of molecular variance inferred from metric distances among DNA haplotypes: application to human mitochondrial DNA restriction data. Genetics 131: 479–491.

Fallon, S.M., 2007. Genetic data and the listing of species under the U.S. Endangered Species Act. Conservation Biology 21, 1186-1195.

Federal Register, 2005. Endangered and Threatened Wildlife and Plants; Listing Roswell springsnail, Koster's tryonia, Pecos assiminea, and Noel's amphipod as Endangered With Critical Habitat. 50 CFR Part 17, 70(152):46304-46333.

Folmer, O., Black, M., Hoeh, W., Lutz, R., Vrijenhoek, R., 1994. DNA primers for amplification of mitochondrial cytochrome c oxidase subunit I from diverse metazoan invertebrates. Molecular Marine Biology and Biotechnology 3, 294-299.

France, S.C., Kocher, T.D., 1996. Geographic and bathymetric patterns of mitochondrial 16S rRNA sequence divergence among deep-sea amphipods, Eurythenes gryllus . Marine Biology 126, 633-643.

Geller, J.B., Walton, E.D., Grosholtz, E.D., Ruiz, G.M., 1997. Cryptic invasions of the crab Carcinus detected by molecular phylogeography. Molecular Ecology 6, 901-906.

Gervasio, V, Berg, D.J., Lang, B.K., Allan, N.L., Guttman, S.I., 2004. Genetic diversity in the Gammarus pecos species complex: implications for conservation and regional biogeography in the Chihuahuan Desert. Limnology and Oceanography 49, 520-531.

Glennon, R., 2002. Water Follies: Groundwater Pumping and the Fate of America’s Fresh Waters. Island Press, Washington, D.C.

Gómez, A., Serra, M., Carvalho, G.R., Lunt, D.H., 2002. Speciation in ancient cryptic species complexes: evidence from the molecular phylogeny of Brachionus plicatilis (Rotifera). Evolution 56, 1431-1444.

Hall, T.A., 1999. BioEdit: a user friendly biological sequence alignment editor and analysis program for Windows 95/98/NT. Nucleic Acids Symposium Series 41, 95-98.

Hartl, D.L., Clark, A.G., 2005. Principles of Population Genetics, 4th Ed. Sinauer Associates, Inc. Sunderland, MA.

Hershler, R., Liu, H.-P., Mulvey, M., 1999. Phylogenetic relationships within the aquatic snail genus Tryonia : implications for biogeography of the North American Southwest.

24 Molecular Phylogenetics and Evolution 13, 377-391.

Hershler, R, Liu H.-P., Stockwell, C.A., 2002. A new genus and species of aquatic gastropods (Rissooidea: Hydrobiidae) from the North American Southwest: phylogenetic relationships and biogeography. Proceedings of the Biological Society of Washington 115, 171-188.

Hershler, R., Sada, D.W., 2002. Biogeography of Great Basin aquatic snails of the genus Pyrgulopsis . In: Great Basin Aquatic Systems History (eds., Hershler R, Madsen DB, Currey DR), pp. 255-276. Smithsonian Contributions to Earth Sciences, 33.

Hogg, I.D., de Lafontaine, Y., Eadie, J.M., 2000. Genotypic variation among Gammarus fasciatus (Crustacea: Amphipoda) from the Great Lakes--St. Lawrence River: Implications for the conservation of widespread freshwater invertebrates. Canadian Journal of Fisheries and Aquatic Sciences 57, 1843-1853.

Holsinger, J.R., 1976. The freshwater amphipod crustaceans (Gammaridae) of North America. U. S. Environmental Protection Agency, Water Pollution Control Research Series, 18050 ELD04/72.

Hutchison D.W., Templeton A.R., 1999. Correlation of pairwise genetic and geographic distance measures: inferring the relative influences of gene flow and drift on the distribution of genetic variability. Evolution 53, 1898-1914.

Jobb, G., von Haeseler, A., Strimmer, K., 2004. TREEFINDER: a powerful graphical analysis environment for molecular phylogenetics. BMC Evolutionary Biology 4:18.

Kareiva, P., Tear, T.H., Solie, S., Brown, M.L., Sotomayor, L., Yuan-Farrell, C. 2006. Nongovernmental Organizations. In: The Endangered Species Act at Thirty: Renewing the Conservation Promise, Vol. 1 (eds., Goble DD, Scott JM, Davis FW), p. 178. Island Press. Washington, DC.

Kelly, D.W., MacIsaac, H.J., Heath, D.D., 2006. Vicariance and dispersal effects on phylogeographic structure and speciation in a widespread estuarine invertebrate. Evolution 60, 257-267.

Knowlton, N., 1993. Sibling species in the sea. Annual Review of Ecology and Systematics 24, 189-216.

Knowlton, N., Weigt, L.A., 1998. New dates and new rates for divergence across the Isthmus of Panama. Proceedings of the Royal Society of London B 265, 2257-2263.

Lang, B.K., Gervasio, V., Berg, D.J., Guttman, S.I., Allan, N.J., Gordon, M.E., Warrick, G., 2003. Gammarid amphipods of northern Chihuahuan Desert spring systems: an imperiled fauna. Museum of Texas Tech University, Special Publications 46, 47-57.

25 Larson, G., Dobney, K., Albarella, U., Fang, M., Matisoo-Smith, E., Robins, J., Lowden, S., Finlayson, H., Brand, T., Willerslev, E., Rowley-Conwy, P., Andersson, L., Cooper, A., 2005. Worldwide phylogeography of wild boar reveals multiple centers of pig domestication. Science 307, 1618-1621.

Lee, C.E., 2000. Global phylogeography of a cryptic copepod species complex and reproductive isolation between genetically proximate "populations." Evolution 54, 2014- 2027.

Lee, C.E., Frost, B.W., 2002. Morphological stasis in the Eurytemora affinis species complex (Copepoda: Temoridae). Hydrobiologia 480, 111-128.

Lefébure, T., Douady, C.J., Gouy, M., Trontelj, P., Briolay, J., Gibert, J., 2006. Phylogeography of a subterranean amphipod reveals cryptic diversity and dynamic evolution in extreme environments. Molecular Ecology 15, 1797-1806.

Lesica, P., Allendorf, F.W., 1995. When are peripheral populations valuable for conservation? Conservation Biology 9, 753-760.

Londo, J.P., Chiang, Y.-C., Hung K.-H., Chiang T.-Y., Schaal, B.A., 2006. Phylogeography of Asian wild rice, Oryza rufipogon , reveals multiple independent domestications of cultivated rice, Oryza sativa . Proceedings of the National Academy of Sciences 103, 9578-9583.

Lowe, A., Harris, S., Ashton, P., 2004. Ecological Genetics: Design, Analysis, and Application. Blackwell Publishing Ltd, Malden, MA.

Macdonald, K.S., Yampolsky, L., Duffy, J.E., 2005. Molecular and morphological evolution of the amphipod radiation of Lake Baikal. Molecular Phylogenetics and Evolution 35, 323-343.

Mathews, L.M., Schubart, C.D., Neigel, J.E., Felder, D.L., 2002. Genetic, ecological, and behavioural divergence between two sibling snapping shrimp species (Crustacea: Decapoda: Alpheus ). Molecular Ecology 11, 1427-1437.

McNeely, J.A., Miller, K.R., Reid, W.V., Mittermeier, R.A., Werner, T.B., 1990. Conserving the world's biological diversity. International Union for Conservation of Nature and Natural Resources/World Resources Institute/Conservation International/World Wildlife Fund/US World Bank, Gland.

Meyran, J.C., Monnerot, M., Taberlet, P., 1997. Taxonomic status and phylogenetic relationships of some species of the genus Gammarus (Crustacea, Amphipoda) deduced from mitochondrial DNA sequences. Molecular Phylogenetic and Evolution 8, 1-10.

Meyran, J.C., Taberlet, P., 1998. Mitochondrial DNA polymorphism among alpine populations of Gammarus lacustris (Crustacea, Amphipoda). Freshwater Biology, 39,

26 259-265.

Minckley, W.L., Unmack, P.J., 2000. Western springs: their faunas and threats to their existence. In Freshwater Ecoregions of North America: A Conservation Assessment (Abell RA et al., eds.), pp. 52-53. Island Press, Washington, D.C.

Moline, A.B., Shuster, S.M., Hendrickson, D.A., Marks, J.C., 2004. Genetic variation in a desert aquatic snail ( Nymphophilus minckleyi ) from Cuatro Ciénegas, Coahuila, México. Hydrobiologia 522, 179-192.

Moritz, C., 1994. Defining 'evolutionarily significant units' for conservation. Trends in Ecology and Evolution 9, 373-375.

Müller, J., 2000. Mitochondrial DNA variation and the evolutionary history of cryptic Gammarus fossarum types. Molecular Phylogenetics and Evolution 15, 260-268.

Murphy, D.D., Freas K.E., Weiss, S.B., 1990. An environment-metapopulation approach to population viability analysis for a threatened invertebrate. Conservation Biology 4, 41- 51.

Palumbi, S.R., Benzei, J., 1991. Large mitochondrial DNA differences between morphologically similar Penaeid shrimp. Molecular Marine Biology and Biotechnology 1, 27-34.

Palumbi, S.R., Martin, A., Romano, S., McMillan, W.V., Stice, L., Grabowski, G., 1991. The Simple Fool's Guide to PCR. v2.0. University of Hawaii.

Panchal, M., 2007. The automation of Nested Clade Phylogeographic Analysis. Bioinformatics 23, 509-510.

Pfenninger, M., Staubach, S., Albrecht, C., Streit, B., Schwenk, K., 2003, Ecological and morphological differentiation among cryptic evolutionary lineages in freshwater limpets of the nominal form-group Ancyclus fluviatilis (O.F. Muller, 1774). Molecular Ecology 12, 2731-2745.

Pitra, C., Lieckfeldt, D., Alonso J.C., 2000. Population subdivision in Europe’s great bustard inferred from mitochondrial and nuclear DNA sequence variation. Molecular Ecology 9, 1165-1170.

Pons, J., Barraclough, T.G., Gomez-Zurita, J., Cardoso, A., Duran, D.P., Hazell, S., Kamoun, S., Sumlin, W.D., Vogler, A.P., 2006. Sequence-based species delimitation for the DNA taxonomy of undescribed insects. Systematic Biology 55, 595-609.

Posada, D., 2003. Using MODELTEST and PAUP* to select a model of nucleotide substitution. In: Current Protocols in Bioinformatics (eds Baxenavis, D.B., Page, R.D.M., Petsko, G.A., Stein, L.D., Storma, G.D.), pp. 6.5.1-6.5.14. Wiley, New York.

27 Posada, D., Crandall, K.A., Templeton, A.R., 2000. GeoDis: A program for the cladistic nested analysis of the geographical distribution of genetic haplotypes. Molecular Ecology 9, 487-488.

Quan, J., Lü, X.-M., Zhuang, Z., Dai, J., Deng, J., Zhang, Y.-P., 2001. Low genetic variation of Penaeus chinensis as revealed by mitochondrial COI and 16S rRNA gene sequences. Biochemical Genetics 39, 279-284.

Quan, J., Zhuang, Z., Deng, J., Dai, J., Zhang, Y.-P., 2004. Phylogenetic relationships of 12 Penaeoidea shrimp species deduced from mitochondrial DNA sequences. Biochemical Genetics 42, 331-345.

Remerie, T., Bourgois, T., Peelaers, D., Vierstraete, A., Vanfleteren, J., Vanreusel, A., 2006. Phylogeographic patterns of the mysid Mesopodopsis slabberi (Crustacean: Mysida) in Western Europe: evidence for high molecular diversity and cryptic speciation. Marine Biology 149, 465-481.

Reuschel, S., Schubart, C.D., 2006. Phylogeny and geographic differentiation of Atlanto- Mediterranean species of the genus Xantho (Crustacea: Brachyura: Xanthidae) based on genetic and morphometric analyses. Marine Biology 148, 853-866.

Rock, J., Ironside, J., Potter, T., Whitely, N.M., Lunt, D.H., 2007. Phylogeography and environmental diversification of a highly adaptable marine amphipod, Gammarus duebeni . Heredity 99, 102-111.

Sada, D.W., Vinyard, G.L., 2002. Anthropogenic changes in biogeography of Great Basin aquatic biota. In: Great Basin Aquatic Systems History (eds., Hershler R, Madsen DB and Currey DR), pp. 255-276. Smithsonian Contributions to Earth Sciences, 33.

Sada, D.W., Fleishman, E., Murphy, D.D., 2005. Associations among spring-dependent aquatic assemblages and environmental and land use gradients in a Mojave Desert mountain range. Diversity and Distributions 11, 91-99.

Sala, O.E., Chapin III, F.S., Armesto, J.J., Berlow, E., Bloomfield, J., Dirzo, R., Huber- Sanwald, E., Huenneke, L.F., Jackson, R.B., Kinzig, A., Leemans, R., Lodge, D.M., Mooney, H.A., Oesterheld, M., Poff, N.L., Sykes, M.T., Walker, B.H., Walker, M., Wall, D.H., 2000. Global biodiversity scenarios for the year 2100. Science 287, 1770-1774.

Schneider, S., Kueffer, J.M., Roesslie, D., Excoffier, L., 2000. Arlequin ver. 2.0: A software for population genetic data analysis. Genetics and Biometry Laboratory, University of Geneva, Geneva.

Sei, M., Lang, B.K., Berg, D.J., 2009. Genetic and community similarities are correlated in endemic-rich springs of the northern Chihuahuan Desert. Global Ecology and Biogeography 18, 192-201.

28 Seidel, R., 2005. Rapid hydrostatic limb inflation in the prawn Macrobrachium lar . Journal of Crustacean Biology 25, 460-461.

Sharbel TF, Haubold B, Mitchell-Olds T. 2000. Genetic isolation by distance in Arabidopsis thaliana : biogeography and postglacial colonization of Europe. Molecular Ecology 9: 2109-2118.

Shaw, K.L., 1999. A nested analysis of song groups and species boundaries in the Hawaiian cricket genus Laupala . Molecular Phylogenetics and Evolution 11, 332-341.

Sokal RR. 1979. Testing statistical significance of geographic variation patterns. Systematic Zoology 28:227-232.

Sotelo, G., Morán, P., Posada, D., 2008. Genetic identification of the northeastern Atlantic spiny spider crab as Maja brachydactyla Balss, 1922. Journal of Crustacean Biology 28, 76-81.

Souza, V., Espinoza-Asuar, L., Escalante, A.E., Eguiarte, L.E., Farmer, J., Forney, L., Lloret, L., Rodríguez-Martínez, J.M., Soberón, X., Dirzo, R., Elser, J.J., 2006. An endangered oasis of aquatic microbial biodiversity in the Chihuahuan desert. Proceedings of the National Academy of Sciences 103, 6565-6570.

Stevens, L.E., Meretsky, V.J. (eds) 2008. Aridland Springs in North America: Ecology and Conservation. University of Arizona Press, Tucson, AZ.

Stillman, J.H., Reeb, C.A., 2001. Molecular phylogeny of eastern Pacific porcelain crabs, genera Petrolisthes and Pachycheles , based on the mtDNA 16S rDNA sequence: phylogeographic and systematic implications. Molecular Phylogenetics and Evolution 19, 236-245.

Strayer, D.L., 2006. Challenges for freshwater invertebrates conservation. Journal of North American Benthological Society 25, 271-287.

Swofford, D.L., 2000. PAUP* Phylogenetic analysis using parsimony (* and other methods). Version 4. Sinauer Assoc., Sunderland, MA.

Templeton, A.R., Crandall, K.A., Sing, D.F., 1992. A cladistic analysis of phenotypic associations with haplotypes inferred from restriction endonuclease mapping and DNA sequence data. III. Cladogram estimation. Genetics 132, 619-633.

Templeton, A.R., Routman, E., Phillips, C.A., 1995. Separating population structure from population history: a cladistic analysis of the geographical distribution of mitochondrial DNA haplotypes in the tiger salamander, Ambystoma tigrinum . Genetics 140, 767-782.

Thomas, C.D., Cameron, A., Green, R.E., Bakkenes, M., Beaumont, L.J., Collingham,

29 Y.C., Erasmus, B.F.N., de Siqueira, M.F., Grainger, A., Hannah, L., Hughes, L., Huntley, B., van Jaarsveld, A.S., Midgley, G.F., Miles, L., Ortega-Huerta, M.A., Peterson, A.T., Phillips, O.L., Williams, S.E., 2004. Extinction risk from climate change. Nature 427, 145-148.

Thomas, E.P., Blinn, D.W., Keim, P., 1997. Genetic and behavioural divergence among desert spring amphipod populations. Freshwater Biology, 38, 137-143.

Thomas, E.P., Blinn, D.W., Keim, P., 1998. Do xeric landscapes increase genetic divergence in aquatic ecosystems? Freshwater Biology 40, 587-593.

Thomas, T.G., 1972. The geomorphic evolution of the Pecos River system. Baylor Geological Studies Bulletin 22, 1-40.

Thorpe, J.P., Solé-Cava, A.M., 1994. The use of allozyme electrophoresis in invertebrate systematics. Zoologica Scripta 23, 3-18.

Tsoi, K.H., Wang, Z.Y., Chu, K.H., 2005. Genetic divergence between two morphologically similar varieties of the Kumura shrimp Penaeus japonicus . Marine Biology 147, 367-379.

U.S. Congress. 1982. Senate Report no. 418, 97th Congress, 2d Session.

Vaughan, R., 1994. Endangered Species Act Handbook. Government Institutes, Inc. Rockville, Maryland.

Vrijenhoek, R.C., 1998. Conservation genetics of freshwater fish. Journal of Fish Biology 53, 394-412.

Wheeler, Q.D., Meier, R., eds., 2000. Species concepts and phylogenetic theory: a debate. 256 pp. Columbia University Press, New York.

Wiens, J.J., Penkrot, T.A., 2002. Delimiting species using DNA and morphological variation and discordant species limits in spiny lizards ( Sceloporus ). Systematic Biology 51, 69-91.

Witt, J.D.S., Blinn, D.W., Hebert, P.D.N., 2003. The recent evolutionary origin of the phenotypically novel amphipod Hyalella montezuma offers and ecological explanation for morphological stasis in a closely allied species complex. Molecular Ecology 12, 405- 413.

Witt, J.D.S., Hebert, P.D.N., 2000. Cryptic species diversity and evolution in the amphipod genus Hyalella within central glaciated North America: a molecular phylogenetic approach. Canadian Journal of Fisheries and Aquatic Sciences 57, 687-698.

Witt, J.D.S., Threloff, D.L., Hebert, P.D.N., 2006. DNA barcoding reveals extraordinary

30 cryptic diversity in an amphipod genus: implications for desert spring conservation. Molecular Ecology 15, 3073-3082.

Zhang, Q., Chiang, Y., George, M., Liu, J.Q., Abbott, R.J., 2005. Phylogeography of the Qinghai-Tibetan endemic Juniperus przewalskii (Cupressaceae) inferred from chloroplast DNA sequence variation. Molecular Ecology 14, 3513-3524.

16 U.S.C. § 1533(b)(1)(A). The Endangered Species Act of 1973.

31 ACKNOWLEDGEMENTS

I thank Todd Levine, Emy Monroe, and Makiri Sei (Department of Zoology, Miami University) for their valuable input and suggestions during the development of this project. Access and permission to collect on public and private lands was granted by: Jeff Howland (BLNWR), David Riskind and Tom Johnson (TPWD), Larry Paul (Lincoln National Forest), Patricia Griffin, Junior Kearns and Robert Myers (WSMR), and John Karges (TNC-TX). I thank Chris Wood (Center for Bioinformatics and Functional Genomics, Miami University) for assistance during the sequencing phase of this work. I also thank Miami University undergraduates Neil Bruce and Kirk Weber for their assistance. Comments from Andrew Gill and two anonymous reviewers greatly improved our manuscript. Funding was provided by the New Mexico Department of Game and Fish (Contract No. 04-516.0000-0093), U. S. Fish and Wildlife Service (Region 2), and the Miami University Field Research Workshop. This report has been approved for public release by White Sands Missile Range for unlimited distribution; the White Sands Missile Range Operational Security review was completed 12 May 2009.

32 Figure Legends:

Figure 1. Collection sites for G. pecos complex populations from the Chihuahuan Desert. Mountains are indicated by diagonal lines. See Table 1 for population codes.

Figure 2. Pairwise geographic distance along rivers vs. pairwise F ST .

Figure 3. A haplotype network constructed from 134 combination sequences for COI and 16S. Each unique haplotype is displayed as an oval, and the size of each oval corresponds to the haplotype frequency, with frequencies higher than 1 denoted by number (e.g. n = 3). Small yellow circles correspond to hypothesized intermediate haplotypes not detected in the sample. Relatively large gaps have number of intermediates indicated inside lightning bolt symbol.

Figure 4. Maximum likelihood (ML) tree constructed from the 91 unique concatenated haplotypes, with bootstrap values ( ≥ 50%) shown on branches and nodes.

33 Table 1. Sample sites and population codes for the Gammarus pecos species complex, as identified by Cole (1985). All individuals analyzed during this study were collected within 300 m of the coordinates shown. The last three sites either, 1) harbor no amphipods presently, or 2) cannot be located based upon earlier descriptions given. Extant populations which were not considered by Cole (1985) are indicated in the second column by the character “-”.

Population Cole (1985) Sample Site Location Latitude / Longitude Nominal Species Code Designation BLBC “D” Bitter Creek, Bitter Lake National Wildlife Refuge, Chaves 33 o 28 ΄ 46˝ N / 104 o 25 ΄ 39˝ W Gammarus desperatus County, NM BLHM “D” Hunter Marsh, Bitter Lake National Wildlife Refuge, Chaves 33 o 24 ΄ 52˝ N / 104 o 25 ΄ 16˝ W Gammarus desperatus County, NM BLSS “D” Sago Spring, Bitter Lake National Wildlife Refuge, Chaves 33 o 28 ΄ 41˝ N / 104 o 25 ΄ 11˝ W Gammarus desperatus County, NM BLU6 “D” Unit 6, Bitter Lake National Wildlife Refuge, Chaves County, 33 o 26 ΄ 46˝ N / 104 o 24 ΄ 16˝ W Gammarus desperatus NM CS - Caroline Spring (also called “T5”), Independence Creek, 30 o 26 ΄ 40˝ N / 101 o 43 ΄ 13˝ W Gammarus sp. Terrell County, TX DY “P” Diamond Y Spring (also called “Willbank Spring”), Diamond 31 o 02 ΄ 12˝ N / 102 o 53 ΄ 27˝ W Gammarus pecos Y Draw, Pecos County, TX ESS - East Sandia Spring, Toyah Creek, Jeff Davis County, TX 30 o 59 ΄ 28˝ N / 103 o 43 ΄ 44˝ W Gammarus sp. GS - Giffin Spring, Toyah Creek, Jeff Davis County, TX 30 o 56 ΄ 45˝ N / 103 o 47 ΄ 23˝ W Gammarus sp. MS - Malpais Spring, White Sands Missile Range, Otero County, 33 o 17 ΄ 18˝ N / 106 o 18 ΄ 33˝ W Gammarus sp. NM PL “H” Phantom Lake Spring, Toyah Creek, Jeff Davis County, TX 30 o 56 ΄ 05˝ N / 103 o 50 ΄ 58˝ W Gammarus hyalelloides SB “E” Sitting Bull Spring, Lincoln National Forest, Eddy County, 32 o 14 ΄ 12˝ N / 104 o 42 ΄ 08˝ W Gammarus sp. NM SSS “S” San Solomon Spring, Toyah Creek, Jeff Davis County, TX 30 o 56 ΄ 41˝ N / 103 o 47 ΄ 11˝ W Gammarus pecos (extirpated) “D” North Spring, Roswell, Chaves County, NM 33 o 25 ΄ 30˝ N / 104 o 29 ΄ 20˝ W Gammarus desperatus (extirpated) “C” Irrigation Canal, Jeff Davis County, TX 30 o 56 ΄ 00˝ N / 103 o 50 ΄ 40˝ W Gammarus sp. ? “M” Reeves County, TX unknown Gammarus hyalelloides

34 Table 2. Summary of descriptive genetic data for combined sequences, with corresponding GenBank accession numbers for each fragment.

No. of combined No. of unique No. of haplotypes Mean population Accession Nos. Accession Nos. sequences per haplotypes per shared with other diversity ( 0 no. of population spring spring(s) pairwise differences) COI 16S BLBC 6 3 1 1.000 FJ948613, FJ948631 – FJ948632 FJ948704, FJ948722 – FJ948723 BLHM 3 3 0 8.000 FJ948609 – FJ948611 FJ948700 – FJ948702 BLSS 9 7 1 1.944 FJ948612 – FJ948618 FJ948703 – FJ948709 BLU6 13 12 0 5.385 FJ948619 – FJ948630 FJ948710 – FJ948721 CS 10 8 0 3.533 FJ948633 – FJ948640 FJ948724 – FJ948731 DY 11 3 0 1.091 FJ948641 – FJ948643 FJ948732 – FJ949734 ESS 9 3 1 0.444 FJ948644 – FJ948646 FJ948735 – FJ948737 GS 8 4 1 1.429 FJ948645, FJ948647 – FJ948649 FJ948736, FJ948738 – FJ948740 MS 8 5 0 2.500 FJ948650 – FJ948654 FJ948741 – FJ948745 PL 12 7 0 1.955 FJ948655 – FJ948661 FJ948746 – FJ948752 SB 36 34 0 8.884 FJ948575 – FJ948608 FJ948666 – FJ948699 SSS 9 4 0 2.444 FJ948662 – FJ948665 FJ948753 – FJ948756

35 Table 3. Pairwise F ST for all populations (* = not significantly different from 0, at α = 0.05).

BLBC BLHM BLSS BLU6 CS DY ESS GS MS PL SB SSS BLBC - BLHM 0.944 - BLSS 0.045* 0.941 - BLU6 0.412 0.889 0.436 - CS 0.948 0.938 0.945 0.915 - DY 0.983 0.974 0.976 0.946 0.961 - ESS 0.989 0.977 0.981 0.947 0.957 0.988 - GS 0.980 0.966 0.973 0.940 0.947 0.982 0.001* - MS 0.985 0.956 0.983 0.967 0.972 0.986 0.988 0.983 - PL 0.974 0.965 0.969 0.942 0.944 0.978 0.929 0.909 0.982 - SB 0.941 0.901 0.944 0.940 0.933 0.944 0.941 0.939 0.332 0.943 - SSS 0.970 0.959 0.966 0.936 0.939 0.975 0.346 0.287 0.979 0.895 0.939 -

36

37 38 39 40 41 Chapter 2 2

Detecting conservation units using morphological versus molecular criteria: evaluating the Gammarus pecos species complex as a test case

ABSTRACT

Recent studies have revealed the high frequency of morphologically cryptic taxa among invertebrates, particularly among aquatic taxa. In light of a sharp global decline in terrestrial, marine and freshwater ecosystems, rapid biodiversity screening methods are urgently needed. I suggest that morphological and molecular characters are both valuable and serve complementary roles in the assessment of biodiversity, but conclusions based on the analysis of these different classes of characters should be compared for degree of concordance. Amphipods in the Gammarus pecos species complex of the northern Chihuahuan Desert represented an ideal model system for comparing the detected number of discrete ecological entities according to morphological and two molecular methodologies, based on a cytochrome c oxidase I (COI) dataset. I compared results from an earlier morphology-based assessment to my results from screening 166 COI gene sequences according to Moritz’ Evolutionarily Significant Unit (ESU) concept and a DNA barcode-based Species Screening Threshold (SST) concept. Overall, my results showed strong concordance between the two molecular screening methods, with the main difference being the phylogenetic placement of the Phantom Lake amphipod population. The molecular methods showed that two other populations are distinct from one another, whereas the morphological method alone failed to separate them. My results underscore the need to look beyond morphology in screening for conservation units among aquatic invertebrates in desert ecosystems and elsewhere. Furthermore, morphological and molecular techniques are shown to be complementary, and when used in conjunction can be a powerful tool for addressing taxonomic and conservation issues dependent on the accuracy of biodiversity assessment.

2 Manscript to be submitted to the journal Conservation Genetics

42 INTRODUCTION

In light of documented declines in biodiversity within terrestrial, marine and freshwater ecosystems (Jenkins 2003), rapid screening methods are urgently needed to assess biodiversity while populations are still extant. Accurate identification of species has been described as crucial in all areas of biology, but is especially needed in biodiversity conservation (Balakrishnan 2005; Witt et al. 2006). The oldest taxonomic tradition is rooted in morphological analysis, involving the identification, description, and enumeration of morphometric and meristic characters useful in grouping organisms based on degree of external similarity. Such characters and calculated morphological distances can be used to delimit species. However, there are abundant examples of morphologically cryptic taxa reported in the literature (Avise 2004), with a disproportionately high number of these examples present within invertebrate groups, particularly within aquatic invertebrates (Thomas et al. 1997; Thomas et al. 1998; Lee & Frost 2002; Lefébure et al. 2006; Witt et al. 2006). While the older and more traditional taxonomic techniques have been valuable for estimating the biodiversity present in specific habitats and ecoregions, there remains a lack of close association between morphological and molecular genetic rates of change (Bromham & Hendy 2000). Most estimates of biodiversity based exclusively on morphological data are expected to under- represent the full range of diversity present at the focal level of analysis (Seidel et al. 2009). Various other methods have been employed for assessing the number and identity of discrete entities that contribute to the biodiversity of ecosystems, to avoid morphological ambiguities. For example, Moritz (1994) updated the concept of Evolutionarily Significant Units (ESUs) to focus on the single criterion of genetic distinctiveness. Whereas earlier definitions of ESUs focused on reproductive isolation and ecological distinctness (Ryder 1986; Waples 1991), Moritz (1994) emphasized that ESUs should show genetic distinctiveness as demonstrated by reciprocal monophyly for mtDNA haplotypes, and significant divergence of allele frequencies at nuclear loci. While this definition omits the ecological distinctness component present in earlier ESU definitions, there is the implicit assumption that genetic divergence will be accompanied by ecological divergence.

43 In addition to morphological distance and the phylogeny-based ESU approach, a third approach, based on DNA barcoding (Hajibabaei et al. 2007), has been used to identify distinct lineages among groups with high morphological similarity. Comparisons of mtDNA haplotypes for the cytochrome c oxidase subunit I (COI) gene are used in DNA barcoding studies to search for divergences that are sufficiently large as to indicate species differences with high degree of certainty (Witt et al. 2006). A species screening threshold (SST), set at 10 times the average COI haplotype divergence within recognized species, has been used to detect new provisional species among a group of neotropical skipper butterflies (Hebert et al. 2004). A similar 10× SST helped detect a relatively high number of cryptic species within Hyalella , a taxonomically challenging genus of amphipods (Witt et al. 2006). Hebert et al. (2004) and Witt et al. (2006) considered the 10× threshold to be conservative and therefore useful in detecting descrete species groups. These studies did not attempt to formally describe novel species based on patterns of sequence divergence; instead they aimed to flag lineages as provisional species that potentially merit formal recognition. Despite many recent advances in the molecular techniques that have become available for biodiversity assessment, taxonomy and systematics (Avise 2004), the difficult problem of species identification still persists for investigators in these disciplines (Balakrishnan 2005). While many biologists view the inclusion of molecular techniques as an improvement over the more traditional reliance on morphological techniques alone, Will and Rubinoff (2004) strongly caution against the replacement of the latter by the former. They assert that if a molecular identification system (e.g. DNA barcoding) replaces morphological studies, the limitations of the molecular methods may ultimately impede our understanding of biodiversity. I suggest that morphological and molecular characters are both valuable and serve complementary roles in the assessment of biodiversity, but conclusions based on the analysis of these different classes of characters should be compared for degree of concordance. I conducted such a comparison by examining the taxonomic relationships of a group of amphipods (members of the Gammarus pecos species complex) endemic to spring systems associated with the Pecos River flowing through New Mexico and Texas, an arid region where aquatic habitats are disappearing because of groundwater

44 withdrawal (Lang et al. 2003). Development and implementation of effective conservation measures is premature until ambiguous taxonomic relationships within the focal system have been clarified. The G. pecos complex represents an ideal model system for comparing the detected number of discrete ecological entities based on morphological distance, ESU and SST frameworks, because of previous morphological work on this complex (Cole 1985) and the genetic diversity uncovered in this study and by Gervasio et al. (2004). Whereas this study investigates mitochondrial markers, Gervasio et al. (2004) evaluated allozyme (nuclear) diversity for most of the same populations, allowing for a more comprehensive genetic survey. Results from these comparisons address the issue of concordance between analyses of morphological versus molecular biodiversity. The earlier morphological results produced by Cole (1985) will serve as a prediction about the number of discrete entities present in the G. pecos complex, which can be tested using molecular genetic data interpreted using the ESU and SST frameworks. Since current species descriptions are based on Cole’s (1985) morphology-based determinations, my molecular results will have important taxonomic and conservation implications for populations in which genetic distinctions are detected, but no previous analysis found corresponding morphological differences. METHODS Cole’s Morphological Methodology The morphological distance results used for comparison in this study were calculated by Cole (1985) from meristic and morphometric characters for the seven Gammarus populations in Texas and New Mexico known in 1985 (Cole 1985). Various features of antenna 1, antenna 2, coxal plates 1-4, pereopods 1-7, uronites 1-3, and uropod 3 (Fig. 1) were used to assess character states. Mann-Whitney U tests were used by Cole (1985) to differentiate various character states based on rejecting the null hypothesis of no significant differences between populations at α = 0.05. Population pairs for which the null was not rejected were interpreted as being conspecific, while population pairs with significant differences were interpreted as being different species. My analysis of Cole’s results (1985) involved the examination of the geographic distribution of morphologically discrete entities, as identified in Cole’s study. I overlaid these entity-by-location combinations with current species names, and created a

45 representation of those morphological groups to be compared with my results from molecular genetic analysis. Molecular Methodology Fresh amphipod speciments for genetic analysis were collected from extant populations in the G. pecos species complex at 12 spring sites (Fig. 2, Table 1) associated with the Pecos River in southeastern New Mexico and western Texas in 2003, 2004 and 2005. Eleven of these spring sites are located within the Pecos River basin of the Permian Basin (Cartwright 1930) and one site, Malpais Spring (MS), is located within the endorheic Tularosa Basin, NM, to the west of the Permian Basin. I used fine mesh hand nets to collect amphipods (30-200 animals per site) from the water column, macrophytes or rocky bottom of each spring. I stored all animals in glass collection jars filled with 95% ethanol until DNA extraction could be performed. I extracted complete genomic DNA from 10 - 40 individuals per site. DNA extraction involved dissecting either intact pleon or pereopods from each individual in a standard extraction protocol involving proteinase K, ribonuclease, and several Promega ® reagents (Nuclei Lysis Solution, cat.# A7943; Protein Precipitation Solution, cat.# A7953). DNA extraction from limbs might in the future be done nondestructively, given that most crustaceans can autotomize and regrow appendages, some relatively rapidly (Seidel 2005). A 680-bp region of the COI gene was amplified using the primers LCO1490: GGTCAACAAATCATAAAGATATTGG (Folmer et al. 1994) and a shortened version of HCO2198: TCAGGGTGACCAAAAAATCA (Folmer et al. 1994). Also for COI, I developed custom internal primers CBL4f: GTGAAGAGAGAAAATAGCTA and CBL4r: ATYATAATTGGGGGGTTC for use in amplifying shorter COI gene fragments when the Folmer et al. (1994) “universal” primers failed to produce amplification. Each 50-µL polymerase chain reaction (PCR) contained 25 µL Taq PCR Master Mix (3.7 U Taq DNA polymerase, 1.5 mM MgCl 2 and 200 µM each dNTP; Qiagen ® cat.# 201443), 0.05 nmol of each primer, ca. 2 mM additional MgCl 2, 5 ng DNA template and 6.5 µL molecular water. PCR conditions consisted of 3 min at 94 oC followed by 5 cycles of 1 min at 94 oC, 1 min at 45 oC, 1 min at 72 oC; followed by 35 cycles of 1 min at 94 oC, 1 min at 50 oC, 1 min at 72 oC; followed by 5 min at 72 oC. COI gene products were isolated using electrophoresis in a

46 2% agarose gel and were gel-extracted using Qiagen ® QIAquick Spin Columns and buffers (cat.# 28106). Cycle sequencing reactions were performed using ABI BigDye terminator v3.1 sequencing kits (25 cycles, annealing temperature: 50 oC) and the primers mentioned above. I sequenced PCR products (COI fragments) in both directions with an ABI 3130 automated sequencer (Applied Biosystems). I aligned forward and reverse sequences using BioEdit (Hall 1999) to verify basecall accuracy; the final alignment for nine of the focal populations contained sequences 620-bp in length, whereas the populations BLBC, BLSS, and BLU6 which required the custom internal primers, produced sequences 240-bp in length. Data Analysis For investigation within the ESU framework, I employed neighbor-joining (NJ) and maximum parsimony (MP) phylogenetic analyses of our COI sequence data using MEGA3.1 (Kumar et al. 2004) with default settings. Although both NJ and MP methods produce structures which visually present the similarity relationships of focal taxa, their tree-building algorithms differ. In the NJ method, a distance-matrix is produced using genetic distances which are calculated from multiple sequence alignments, but without invoking any particular evolutionary model (Avise 2004). In contrast, the MP method implicitly makes use of a specific model of evolution (parsimony) to construct a tree according to character states (Avise 2004). Next, multiple trees (1 per sample) were estimated by the automated procedure and a point was recorded every time each grouping from the original tree occurred in the sample trees (Avise 2004). I used bootstrapping to derive confidence values for groupings in the trees, which involved taking 1000 random samples of sites from the alignment. Bootstrap support numbers were percentages of sample trees which produced the same grouping of taxa as the original tree. These values appear at nodes and branches on the final tree. To generate NJ and MP trees using the gene sequences obtained in this study, I used COI sequences representing unique haplotypes, with haplotype count and geographic source noted in each haplotype name (as shown on trees). As an outgroup, I added one COI sequence for Echinogammarus ischnus obtained from GenBank (accession no. AY326115). I then assessed both tree types for reciprocal monophyly among tip clades, following the definition given in Moritz (1994).

47 Lastly, for investigation within the SST framework, I used Arlequin v2.000 (Schneider et al. 2000) to identify the number of unique haplotypes within the G. pecos species complex. From the population average pairwise differences calculated using Arlequin v2.000 (Schneider et al. 2000), I divided the number of differences by the number of nucleotides in the sequences compared (620 or 240 bp), then multiplied by 100% to obtain percent divergences for each population (constant rate of nucleotide mutation across sites assumed). I set the SST at 10 times the average of these within- population estimates, and then employed this 10× SST to identify provisional species according to Witt et al. (2006). I interpreted pairwise interpopulation comparisons that did not rise to this level as members of the same provisional species. RESULTS Morphological Overall, morphological assessment of the G. pecos species complex revealed five distinct species. Two of these are from locations where populations have been extirpated. The Gammarus from Diamond Y Spring (DY) and those from San Solomon Spring (SSS) display insufficient distinctiveness to conclude that these populations are different species, according to Cole (1985), as shown by a solid line beneath these two in Figure 3A. Amphipods at DY and SSS were designated G. pecos . Cole (1985) determined the “SSS” population to be related to Gammarus “eIC,” a canal-dwelling form now extirpated, although not closely enough to be considered the same species. Gammarus “SB” from Eddy County, NM, was found by Cole (1985) to be a distinct species, although it was most similar to Gammarus “eNS,” a population now extirpated from North Spring in Roswell, NM. The latter and nearby organisms are considered G.

desperatus . Gammarus “ ?RC” (symbol “?” used to reflect ambiguous locality) from Reeves County, TX, and Gammarus “PL,” from Toyah Creek, TX, were the same and designated G. hyalelloides Cole 1976. The Gammarus currently extant at other sites, shown in Table 1, were unknown to Cole (1985). Molecular I examined 166 COI gene sequences (no. of individuals per site shown in Table 1), which contained 90 haplotypes. Each spring contained on average 7.5 COI haplotypes (range: 1 – 34 haplotypes). The mean intrapopulation sequence divergence was 1.2%

48 (range: 0.1 – 3.7%) and the mean interpopulation sequence divergence was 14.2% (range: 0.5 – 23.4%). For the ESU framework, I examined for reciprocal monophyly among tip clades on the Neighbor-joining (NJ) and Maximum Parsimony (MP) trees (Fig. 4) following the definition given in Moritz (1994), and identified the following reciprocally monophyletic clades: 1) BLHM, MS, SB; 2) BLBC, BLSS, BLU6; 3) ESS, GS, PL, SSS; 4) DY; and 5) CS. All tips showed strong bootstrap support, ranging from 80 – 100% on the NJ tree, and 62 – 99% on the MP tree. Clade 5 (from CS) was reciprocally monophyletic on both the NJ and MP trees, but a difference in tree topology was apparent based on the placement of this clade. On the NJ distance-based tree (Fig. 4A), clade 5 appeared basal to clades 2 – 4, but clade 5 appeared as a tip most similar to clade 3 on the MP character- based analysis (Fig. 4B). Both NJ and MP analyses produced the same five reciprocally monophyletic clades (Fig. 3B). Results for the same molecular data obtained in this study, but analyzed in the SST framework, are depicted in Figure 3C. The SST analysis identified the following groups as distinct: BLBC, BLSS, and BLU6 (group 1); BLHM, MS, and SB (group 2); and GS, SSS, and ESS (group 3). PL was not different from CS under the 10 × SST, and CS was not different from DY when this same threshold was applied, but PL was different from DY. This analysis showed that SSS and DY, which are both nominally consider to be G. pecos , are two separate species. No molecular data are available for the populations Cole (1985) designated “C,” “D,” and “M,” because “C” and “D” have been extirpated and the location of “M” was ambiguous. Fortunately, Gammarus “D” ( G. desperatus ) is currently extant at five spring locations on the Bitter Lake National Wildlife Refuge (BLNWR) in Roswell, NM, according to New Mexico wildlife officials (Brian Lang, personal communication). All BLNWR populations are relatively near the extinct North Spring population ( ≤ 25 km). Therefore, our molecular results for the four BLNWR populations (BLBC, BLHM, BLSS, and BLU6) warrant comparison with Cole’s (1985) results for Gammarus “D.” Overall results for the morphological and molecular based analyses are summarized in Figure 3.

49 DISCUSSION The three nominal species currently included within the G. pecos complex ( G. pecos Cole and Bousfield, 1970; G. desperatus Cole, 1981; and G. hyalelloides Cole, 1976) were originally differentiated by morphology, and another 6 populations of undetermined taxonomic affinity were identified and may represent several undescribed species (Cole and Bousfield 1970; Cole 1976, 1981, 1985). At this time, at least two other populations are presumed to have been extirpated. All current species designations (outlined by location in Table 1) based on these earlier morphological descriptions are treated as hypotheses which are compared based upon the genetic results I present. Overall, my results show strong concordance between the ESU framework groupings (Fig. 3B) and the SST framework groupings (Fig. 3C), with the exception being the position of the Phantom Lake (PL) population. Despite this single difference, both the ESU and SST results show that the SSS and DY populations are distinct, whereas the morphological analysis did not separate SSS and DY (Fig. 3A shows that both are G. pecos ). The molecular data show that G. pecos is present only at DY, the type locality for this species. Even though SSS is morphologically indistinguishable from G. pecos , the ESU analysis shows that the SSS population (along with the other Toyah Basin springs GS and ESS) is probably G. hyalelloides based on genetic similarity with G. hyalelloides at PL, the type locality for this species. However, the SST analysis shows that GS, SSS and ESS residents are distinct from both G. pecos and G. hyalelloides , and is evidence that suggests these three Toyah Basin springs harbor an undescribed species. In further support of the ESUs detected here, all provisional species are reciprocally monophyletic ( sensu Moritz 1994), and most were shown to exhibit significant divergence in nuclear allele frequencies (at allozyme loci; Gervasio et al. 2004), a second criterion suggested for genetic determination of ESU (Moritz 1994). While the ESU and SST frameworks for screening biodiversity uncovered mostly the same groups, the high threshold of detection imposed by the SST (10× the intrapopulation genetic diversity to identify boundaries between provisional species) likely accounts for the SST results showing DY and CS are conspecific. The ESU method (dependant on reciprocal monophyly) separated these two as distinct entities.

50 Despite small differences apparent between the ESU and SST screening methods, implications for current species designations are clear. The state and federally listed G. desperatus , while apparently homogenous from a morphological standpoint, has a distinct type at BLBC, BLSS and BLU6, different from the one present at BLHM. The latter appears, based on the genetic analyses (ESU and SST), to be more similar to geographically distant MS and SB. Additionally, G. pecos at DY and SSS is composed of two distinct provisional species. This is a condition concealed by the indistinguishable morphology of DY and SSS. Conservation could be critical for the long term survival of the entities detected via our molecular analyses. NatureServe ( www.natureserve.org , accessed July 2009) reports the current status for the three nominal species in this complex, with G. desperatus listed as endangered (U.S. Endangered Species Act), imperiled (NatureServe Global Status), and critically endangered (IUCN). While not yet federally listed endangered species in the United States, G. pecos and G. hyalelloides are both critically imperiled (NatureServe Global Status) and G. pecos is categorized as vulnerable (IUCN). If GS, SSS, and ESS are to be considered distinct from both G. hyalelloides at PL and G. pecos at DY, then the ranges for both of these nominal species is even smaller then currently thought. In addition, the new provisional species identified are themselves likely to be placed into categories of conservation concern. While many terrestrial ecosystems are imperiled (Olson et al. 2002), aquatic ecosystems are also in danger and in need of research effort (Abell 2002; Jenkins 2003). Over the last 30 years, freshwater ecosystems specifically have been identified as locations with faster loss of biodiversity compared to both terrestrial and marine ecosystems (Loh 2002). Siltation, canalization, water abstraction, dam construction, and overfishing are recognized anthropogenic agents that act to decrease biodiversity in freshwater ecosystems (Jenkins 2003). The degradation and loss of aquatic habitats in the northern Chihuahuan Desert is driven by ongoing and human-induced groundwater withdrawal (Lang et al. 2003). The biotic communities present at such springs can include flatworms, amphipods, isopods, snails, crayfish, insects, salamanders, frogs, turtles, and fishes (Brune 1981). Various terrestrial animals, including bears, raccoons, snakes, and birds feed on these aquatic organisms, in addition to relying on springs for

51 drinking water (Brune 1981). Amphipods in the G. pecos species complex are widely distributed across the springs in this system associated with the Pecos River, and represent an ideal group with which to compare the performance of different methods of conservation unit detection, methods relevant across diverse taxonomic groups and habitat types. The patterns revealed by both morphological and molecular analyses are likely to be shared among diverse aquatic taxa distributed across desert landscapes. One example of this expected trend is the correlation of our mtDNA results and the allozyme data (Gervasio et al. 2004) for the G. pecos complex, with allozyme data for the endemic fish Gambusia pecosensis (Echelle et al. 1989), showing that in habitats where these amphipods and fish co-occur, they are similarly diverse, yet distinct from other lineages occurring nearby. I note similar patterns when comparing my genetic data to those for cyprinid pupfishes (Echelle et al. 1987) and hydrobiid snails (Hershler et al. 1999); these groups also show high genetic diversity restricted to discrete and isolated habitats. Similar biogeographic trends are recognized among aquatic invertebrate communities (Sei et al. 2009) and terrestrial invertebrate taxa (i.e., land snails, see Bequaert & Miller 1973; Metcalf 1997; Metcalf & Smartt, 1997) of this geographic region. Even terrestrial vertebrates like the ridge-nose rattlesnake mirror the biogeographic pattern found among the amphipods we investigated; i.e. distinct geographic areas harboring distinct genetic entities (Holycross & Douglas 2007). I present my results in the context of rapidly disappearing spring habitat and the urgent need for reliable biological data relevant to conservation management. In the introduction I commented on the history of, and ambiguities associated with, morphology-based screening methods. Given all that is known about the contribution of genetic factors to extinction risk (Frankham 2005), it is encouraging to note that for 38 recent endangered species petitions, 81% of those showing genetic distinction were granted protection status (Fallon 2007). Since identification of species boundaries is so crucial in situations involving endangered species assessments, I emphasize the utility of using genetic markers to reveal important differences among morphologically similar taxa, particularly among aquatic invertebrates. I agree with Dunn (2003), who acknowledged the need for taxonomy based on

52 morphological and genetic data (see also Lipscomb et al. 2003), because in some cases important morphological differences may be present that are not accompanied by comparable divergence in neutral molecular markers. However, Dunn (2003) went on to predict the high likelihood that field taxonomy will continue to be based on consideration of morphological traits, primarily because many workers lack both the resources and expertise to obtain DNA sequences or to interpret them. While true in some cases, particularly in developing countries, Avise (2004) contends that molecular data will continue to become more reliable, less expensive, and more readily obtainable. Despite the accessibility of morphology, a primary morphological taxonomy challenge (particularly among aquatic invertebrates) relates to the considerable plasticity of morphological characters, e.g. the shell morphology of marine mussels, which has contributed to a confused and largely error-prone taxonomy (Koehn 1991). Fortunately the complementary use of molecular data has improved biodiversity assessments and taxonomies for mollusks (Koehn 1991), rotifers (Ciros-Pérez et al. 2001), and brine shrimp (Browne & Bowen 1991). Despite occasional species-level inconsistencies, morphological and genetic data have been combined to successfully delimit species of invertebrates (e.g. snails; Ponder et al. 1991), vertebrates (e.g. birds; Leisler et al. 1997), and plants (e.g. buckwheat; Ohnishi & Matsuoka 1996). My results comparing genetic variability interpreted within the frameworks of ESU and SST, to the earlier studies of morphological variability for the G. pecos species complex, underscore the need to look beyond morphology in screening for conservation units among aquatic invertebrates in desert ecosystems and elsewhere. Given the reliability, speed and decreasing cost of modern genetic methods (Avise 2004), conservation efforts should continue to assess relevant genetic markers to accurately detect hidden biodiversity present even in groups that are lumped based on preliminary morphological assessments. Biota distinguished by genetic evaluation can then be examined more closely for distinctiveness according to morphological and geographical criteria, as some level of distinctiveness could be present at these levels also, but might have been overlooked. In this way, a “total evidence approach” can best inform conservation management and the best science, as called for in the United States Endangered Species Act, can be brought to bear on the biodiversity crisis which confronts us.

53 Literature Cited:

Abell, R. 2002. Conservation biology for the biodiversity crisis: a freshwater follow-up. Conservation Biology 16: 1435-1437.

Avise, J.C. 2004. Molecular Markers, Natural History, and Evolution, 2nd Ed. Sinauer Associates. Sunderland, Massachusetts.

Balakrishnan, R. 2005. Species concepts, species boundaries and species identification: a view from the tropics. Systematic Biology 54: 689-693

Bequaert, J.C. and W. B. Miller, 1973. The Mollusks of the Arid Southwest with an Arizona checklist. University of Arizona Press, Tuscon.

Bromham, L.D., and M.D. Hendy. 2000. Can fast early rates reconcile molecular dates with the Cambrian explosion? Proceedings of the Royal Society of London B 267: 1041-1047.

Browne, R.A. and S.T. Bowen. 1991. Taxonomy and population genetics of Artemia . Pages 221-235, in R.A. Browne, P. Sorgeloos, and C.N.A. Trotman, editors. Artemia Biology. CRC Press, Boca Raton, Florida.

Brune, G. 1981. Springs of Texas, vol. 1. Branch-Smith Inc., Fort Worth, Texas.

Cartwright, L.D. 1930. Transverse section of Permian basin, west Texas and southeast New Mexico. Bulletin of the American Association of Petroleum Geologists 14: 969-981.

Ciros-Pérez, J., A. Gómez, and M. Serra. 2001. On the taxonomy of three sympatric sibling species of the Brachionus plicatilis (Rotifera) complex from Spain, with the description of B. ibericus n. sp. Journal of Plankton Research 23: 1311-1328.

Cole, G.A. 1976. A new amphipod crustacean, Gammarus hyalelloides n. sp., from Texas. Transactions of the American Microscopical Society 89: 80-85.

Cole, G.A. 1981. Gammarus desperatus , a new species from New Mexico (Crustacea: Amphipoda). Hydrobiologia 76: 27-32.

Cole, G.A. 1985. Analysis of the Gammarus-pecos complex (Crustacea: Amphipoda) in Texas and New Mexico, USA. Journal of the Arizona-Nevada Academy of Science 20: 93-103.

Cole, G.A., and E.L. Bousfield. 1970. A new freshwater Gammarus (Crustacea: Amphipoda) from Western Texas. American Midland Naturalist 83: 89-95.

Dunn, C.P. 2003. Keeping taxonomy based in morphology. TRENDS in Ecology and

54 Evolution 18: 270-271.

Echelle, A.A., A.F. Echelle, and D. R. Edds. 1987. Population structure of four pupfish species (Cyprinodontidae: Cyprinodon ) from the Chihuahuan Desert region of New Mexico and Texas: allozymic variation. Copeia 1987: 668-681.

Echelle, A.A., A.F. Echelle, and D. R. Edds. 1989. Conservation genetics of a spring- dwelling desert fish, the Pecos gambusia ( Gambusia nobilis , Poeciliidae). Conservation Biology 3: 159-169.

Fallon, S.M. 2007. Genetic data and the listing of species under the U.S. Endangered Species Act. Conservation Biology 21: 1186-1195.

Folmer, O., M. Black, W. Hoeh, R. Lutz, and R. Vrijenhoek. 1994. DNA primers for amplification of mitochondrial cytochrome c oxidase subunit I from diverse metazoan invertebrates. Molecular Marine Biology and Biotechnology 3: 294-299.

Frankham, R. 2005. Genetics and extinction. Biological Conservation 126: 131-140.

Gervasio, V., D.J. Berg, B.K. Lang, N.L. Allan, and S.I. Guttman. 2004. Genetic diversity in the Gammarus pecos species complex: implications for conservation and regional biogeography in the Chihuahuan Desert. Limnology and Oceanography 49: 520-531.

Hajibabaei, M., G.A.C. Singer, P.D.N. Hebert, and D.A. Hickey. 2007. DNA barcoding: how it complements taxonomy, molecular phylogenetics and population genetics. Trends in Genetics 23: 167-172.

Hall, T.A. 1999. BioEdit: a user friendly biological sequence alignment editor and analysis program for Windows 95/98/NT. Nucleic Acids Symposium Series 41: 95- 98.

Hebert, P.D.N., E.H. Penton, J.M. Burns, D.H. Janzen, and W. Hallwachs. 2004. Ten species in one: DNA barcoding reveals cryptic species in the neotropical skipper butterfly Astraptes fulgerator . Proceedings of the National Academy of Sciences, USA 101: 14812-14817.

Hershler, R., H.-P. Liu, and M. Mulvey. 1999. Phylogenetic relationships within the aquatic snail genus Tryonia : implications for biogeography of the North American Southwest. Molecular Phylogenetics and Evolution 13: 377-391.

Hillewaert, H. 2006. Body plan of an amphipod. Original image licensed uder Creative Commons Attribute ShareAlike 2.5.

Holycross, A.T., and M.E. Douglas. 2007. Geographic isolation, genetic divergence, and ecological non-exchangeability define ESUs in a threatened sky-island rattlesnake.

55 Biological Conservation 134: 142-154.

Jenkins, M. 2003. Prospects for biodiversity. Science 302 : 1175-1177.

Koehn, R.K. 1991. The genetics and taxonomy of species in the genus Mytilus . Aquaculture 94: 125-145.

Kumar, S., K. Tamura, and M. Nei. 2004. MEGA3: Integrated software for Molecular Evolutionary Genetics Analysis and sequence alignment. Briefings in Bioinformatics 5: 150-163.

Lang, B.K., V. Gervasio, D.J. Berg, S.I. Guttman, N.J. Allan, M.E. Gordon, and G. Warrick. 2003. Gammarid amphipods of Northern Chihuahuan Desert spring systems: an imperiled fauna. Museum of Texas Tech University, Special Publications 46: 47-57.

Lee, C.E., and B.W. Frost. 2002. Morphological stasis in the Eurytemora affinis species complex (Copepoda: Temoridae). Hydrobiologia 480: 111-128.

Lefébure, T., C.J. Douady, M. Gouy, P. Trontelj, J. Briolay, and J. Gibert. 2006. Phylogeography of a subterranean amphipod reveals cryptic diversity and dynamic evolution in extreme environments. Molecular Ecology 15: 1797-1806.

Leisler, B., P. Heidrich, K. Schulze-Hagen, and M. Wink. 1997. Taxonomy and phylogeny of reed warblers (genus Acrocephalus ) based on mtDNA sequences and morphology. Journal of Ornithology 138: 469-496.

Lipscomb, D., N. Platnick, and Q. Wheeler. 2003. The intellectual content of taxonomy: a comment on DNA in taxonomy. TRENDS in Ecology and Evolution 18: 65-67.

Loh, J. 2002. Living Planet Report: 2002. WWF, World Wide Fund for Nature, Gland, Switzerland.

Metcalf, A.L., 1997. Land snails of New Mexico from a historical zoogeographic point of view. Pages 71 – 108 in A.L. Metcalf and R.A. Smartt, editors. Land Snails of New Mexico. New Mexico Museum of Natural History and Science Bulletin 10, Albuquerque.

Metcalf, A.L., and R.A. Smartt. 1997. Land snails of New Mexico. Pages 1 – 69 in A.L. Metcalf and R.A. Smartt, editors. Land Snails of New Mexico. New Mexico Museum of Natural History and Science Bulletin 10, Albuquerque.

Moritz, C. 1994. Defining ‘Evolutionarily Significant Units’ for conservation. Trends in Ecology and Evolution 9: 373-375.

Ohnishi, O., Y. Matsuoka. 1996. Search for the wild ancestor of buckwheat II. Taxonomy

56 of Fagopyrum (Polygonaceae) species based on morphology, isozymes and cpDNA variability. Genes and Genetic Systems 71: 383-390.

Olson, D.M., E. Dinerstein, G.V.N. Powell, and E.D. Wikramamanayake. 2002. Conservation biology for the biodiversity crisis. Conservation Biology 16: 1-3.

Ponder,W.F., D.J. Colgan, and G.A. Clark. 1991. The morphology, taxonomy and genetic-structure of Tatea (Mullusca, Gastropoda, Hydrobiidae), estaurine snails from temperate Australia. Australian Journal of Zoology 39: 447-497.

Ryder, O.A. 1986. Species conservation and systematics: the dilemma of subspecies. Trends in Ecology and Evolution 1: 9-10.

Schneider, S, J.M. Kueffer, D. Roesslie, and L. Excoffier. 2000. Arlequin ver. 2.0: A software for population genetic data analysis. Genetics and Biometry Laboratory, University of Geneva, Geneva.

Sei, M., B.K. Lang, and D.J. Berg. 2009. Genetic and community similarities are correlated in endemic-rich springs of the northern Chihuahuan Desert. Global Ecology and Biogeography 18: 192-201.

Seidel, R.A. 2005. Rapid hydrostatic limb inflation in the prawn Macrobrachium lar . Journal of Crustacean Biology 25 : 460-461.

Seidel, R.A., B.K. Lang, and D.J. Berg. 2009. Phylogeographic analysis reveals multiple cryptic species of amphipods (Crustacea: Amphipoda) in Chihuahuan Desert springs. Biological Conservation 142 : 2303-2313.

Thomas, E.P., D.W. Blinn, and P. Keim. 1997. Genetic and behavioural divergence among desert spring amphipod populations. Freshwater Biology 38: 137-143.

Thomas, E.P., D.W. Blinn, and P. Keim. 1998. Do xeric landscapes increase genetic divergence in aquatic ecosystems? Freshwater Biology 40: 587-593.

Waples, R.S. 1991. Pacific salmon, Oncorhynchus spp., and the definition of "species" under the Endangered Species Act. Marine Fisheries Review 53: 11-22.

Will, K.W., and D. Rubinoff. 2004. Myth of the molecule: DNA barcodes for species cannot replace morphology for identification and classification. Cladistics 20: 47- 55.

Witt, J.D.S., D.L. Threloff, and P.D.N. Hebert. 2006. DNA barcoding reveals extraordinary cryptic diversity in an amphipod genus: implications for desert spring conservation. Molecular Ecology 15: 3073-3082.

57 ACKNOWLEDGEMENTS

We thank T. Levine, E. Monroe, and M. Sei (Department of Zoology, Miami University) for their valuable input and suggestions during the development of this project. Access and permission to collect on public and private lands was granted by: J. Howland (Bitter Lake National Wildlife Refuge), D. Riskind and T. Johnson (Texas Parks and Wildlife Department), L. Paul (Lincoln National Forest), P. Griffin, J. Kerns and R. Myers (White Sands Missile Range), and J. Karges (The Nature Conservancy). We thank C. Wood (Center for Bioinformatics and Functional Genomics, Miami University) for assistance during the sequencing phase of this work. We also thank Miami University undergraduates N. Bruce and K. Weber for their assistance. B. Keane, N. Solomon, and B. Steinly provided comments that improved an earlier draft of this manuscript. Funding was provided by the New Mexico Department of Game and Fish (Contract No. 04- 516.0000-0093), U.S. Fish and Wildlife Agency (Region 2), and the Miami University Field Research Workshop. This report has been approved for public release by White Sands Missile Range for unlimited distribution; the White Sands Missile Range Operational Security review was completed 25 August 2009.

58 Figure Legends

Fig. 1: Body plan of the amphipod, modified from Hillewaert (2006)

Fig. 2: Collection sites for 12 extant populations of the G. pecos complex from the Chihuahuan Desert. Two sites with populations now extirpated (eIC and eNS) are also marked for reference. Population codes follow Table 1.

Fig. 3: Results are shown for the three methods of analyses: A) morphological distance, B) evolutionarily significant units, and C) 10× species screening threshold. Colored boxes delineate nominally conspecific entities, although molecular results reveal problems with these currently accepted species names. Population codes follow Table 1.

Fig. 4: A) Neighbor-joining (NJ), and B) maximum parsimony (MP) trees, each using COI haplotype data and 1000 pseudoreplicates to obtain bootstrap values. Haplotype groups shown at tips are collapsed to reciprocally monophyletic clades, with population codes as in Figure 1. Triangles shown at tips on NJ tree (A) are sized according to degree of sequence divergence for haplotypes within collapsed clades. Numerical values at nodes are the percentages of bootstrap pseudoreplicates containing each node. Tip labels indicate the number of haplotypes included from each population.

59 60 61 62 63 Table 1.

Sample sites and population codes for the G. pecos species complex. All individuals genetically analyzed during this study are identified by population code in the far left column, and were collected within 300 m of the coordinates shown. The last three sites either, 1) harbor no amphipods presently, or 2) cannot be located based upon earlier descriptions given. Extant populations which were not included in Cole (1985) are indicated in the second column by the character “-”.

Population Cole (1985) Sample Site Location Latitude / Nominal Species Individuals Code Designation Longitude samples per site: BLBC “D” Bitter Creek, Bitter Lake 33 o 28 ΄ 46˝ N / Gammarus 10 National Wildlife Refuge, 104 o 25 ΄ 39˝ W desperatus NM, USA BLHM “D” Hunter Marsh, Bitter Lake 33 o 24 ΄ 52˝ N / Gammarus 15 National Wildlife Refuge, 104 o 25 ΄ 16˝ W desperatus NM, USA BLSS “D” Sago Spring, Bitter Lake 33 o 28 ΄ 41˝ N / Gammarus 10 National Wildlife Refuge, 104 o 25 ΄ 11˝ W desperatus NM, USA BLU6 “D” Unit 6, Bitter Lake National 33 o 26 ΄ 46˝ N / Gammarus 14 Wildlife Refuge, NM, USA 104 o 24 ΄ 16˝ W desperatus CS - Caroline Spring (also called 30 o 26 ΄ 40˝ N / Gammarus sp. 12 “T5”), Independence Creek, 101 o 43 ΄ 13˝ W Terrell Co., TX, USA DY “P” Diamond Y Spring (also 31 o 02 ΄ 12˝ N / Gammarus pecos 11 called “Willbank Spring”), 102 o 53 ΄ 27˝ W Diamond Y Draw, TX, USA ESS - East Sandia Spring, Toyah 30 o 59 ΄ 28˝ N / Gammarus sp. 10 Creek, TX, USA 103 o 43 ΄ 44˝ W GS - Giffin Spring, Toyah Creek, 30 o 56 ΄ 45˝ N / Gammarus sp. 10 TX, USA 103 o 47 ΄ 23˝ W MS - Malpais Spring, White Sands 33 o 17 ΄ 18˝ N / Gammarus sp. 10 Missile Range, NM, USA 106 o 18 ΄ 33˝ W PL “H” Phantom Lake Spring, Toyah 30 o 56 ΄ 05˝ N / Gammarus 13 Creek, TX, USA 103 o 50 ΄ 58˝ W hyalelloides SB “E” Sitting Bull Falls, Lincoln 32 o 14 ΄ 12˝ N / Gammarus sp. 40 National Forest, NM, USA 104 o 42 ΄ 08˝ W SSS “S” San Solomon Spring, Toyah 30 o 56 ΄ 41˝ N / Gammarus pecos 11 Creek, TX, USA 103 o 47 ΄ 11˝ W eNS “D” North Spring, Roswell, NM, 33 o 25 ΄ 30˝ N / Gammarus -- (extirpated) USA 104 o 29 ΄ 20˝ W desperatus eIC “C” Irrigation Canal, Jeff Davis 30 o 56 ΄ 00˝ N / Gammarus sp. -- (extirpated) County, TX, USA 103 o 50 ΄ 40˝ W ?RC “M” Reeves County, TX, USA unknown Gammarus -- hyalelloides

64 Chapter 3 3

Salinity tolerance as a potential driver of ecological speciation in amphipods (Gammarus spp.) from the northern Chihuahuan Desert

ABSTRACT

Ecological speciation is the process by which barriers to gene flow evolve between populations as a result of ecologically-based divergent selection. Environmental salinity has been identified as one of the most important ecological drivers influencing distribution, abundance, and species richness of aquatic organisms. Springs of the northern Chihuahuan Desert vary in salinity and are home to the Gammarus pecos (Crustacea: Amphipoda) species complex. I used field experiments to compare salinity tolerance among nine amphipod populations from geographically discrete habitats differing in ambient salinities, and to calculate physiological distances among populations. Cluster analysis placed populations into three groups corresponding to low, medium, and high ambient salinities. Partial Mantel tests revealed significant positive correlations between salinity tolerance and habitat salinity, after controlling for other variables such as geographic distances and neutral genetic differences. This provides evidence that ecological speciation could be occurring among amphipod populations at different springs, as indicated by dissimilar physiological responses which follow differences in ambient spring salinities. Reinforced by dispersal barriers between springs, gene flow is restricted between populations and selection preserves variants that most effectively tolerate local salinity levels. Gammarus diversification in the northern Chihuahuan Desert is driven by vicariance and isolation, along with local selection and adaptation.

3 Manuscript invited and in review at the Journal of the North American Benthological Society

65 INTRODUCTION

Ecological speciation has recently been defined as the set of processes by which barriers to gene flow evolve between populations as a result of ecologically-based divergent selection (Rundle and Nosil 2005). Such ecological processes are the interactions between organisms and environments that cause natural selection and that facilitate population establishment, growth and decline (Schluter 1996). The ecological processes that drive speciation are incompletely understood and the topic has only received attention in the literature investigating speciation in sympatry (Rice and Hostert 1993; Bush 1994; Schluter 1996). Schluter (1996) noted that field studies of selection on traits resulting in reproductive isolation are sorely needed, and would complement traditional genetic approaches for investigating the process of speciation. When ecological speciation occurs in aquatic habitats, it is necessary to understand which hydrochemical parameters are most important in this process. Aquatic biota occur over a wide range of salt concentrations (Oren 2002), and environmental salinity has been shown to be a key determinant of life history traits for taxa living where salt levels vary (Neupart et al. 2002). In aquatic habitats, salinity directly effects the metabolism, oxygen consumption, growth and survival of organisms (Jian et al. 2003). While salinity is only one environmental factor shaping biodiversity in aquatic habitats, it has also been identified as one of the most important drivers influencing the distributions, abundances, and species richness among aquatic invertebrates (Colburn 1988; Galat et al. 1988; Williams et al. 1990; Cervetto et al. 1999). For example, increasing salinity was correlated with decreasing species richness of macroinvertebrate communities in the aquatic habitats of Death Valley, California (Colburn 1988). Salinity also affects the distribution of amphipods in the genus Gammarus across North American (Beadle and Cragg 1940) and European (Kolding and Fenchel 1979) waters. In addition to determining the presence or absence of species, salinity also influences genetic diversity at the population level. Nuclear genetic polymorphisms have been reported in pupfish in the genus Cyprinodon , in which distinct clades are associated with salinity variation across spring habitats in the Chihuahuan Desert (Echelle et al. 1987; Stockwell and Mulvey 1998). Brine shrimp ( Artemia ) genotypes face rapid

66 selection along salinity gradients in both field and laboratory settings (Browne and Hoopes 1990). Prokaryote genetic diversity has been shown to change along salinity gradients in coastal ecosystems (Benlloch et al. 2002; Bovier and del Giorgio 2002). Thus, salinity can influence both community composition and population genetic diversity. My goal in the present study was to determine whether salinity may influence ecological speciation of Gammarus amphipods occurring in springs of the northern Chihuahuan Desert. Spring habitats in arid ecosystems are among the most structurally complex, ecologically diverse, productive, “evolutionarily provocative,” and threatened ecosystems on earth (Stevens and Meretsky 2008). Spring waters across the northern Chihuahuan Desert display ambient salinities as low as freshwater ( ≤ 0.5 ppt salt) and as high as 9 ppt salt, with recorded spikes as high as 21 ppt salt (Cole 1985; Stockwell and Mulvey 1998)—a level approaching the normal ocean salinity of 35 ppt salt (Glen et al. 1998). Using field experiments, I show how such differences in ambient spring salinities are associated with dissimilar physiological responses that are consistent with the ecological speciation hypothesis. MATERIALS & METHODS To investigate salinity tolerance of amphipods comprising the Gammarus pecos species complex (Cole 1985), I conducted salinity response field experiments on nine populations located in southeastern New Mexico and western Texas spring systems (Figure 1) during June-September 2008. The G. pecos complex consists of three described species ( G. pecos Cole and Bousfield, 1970; Gammarus desperatus Cole, 1981; Gammarus hyalelloides Cole, 1976), and there may be at least four undescribed species (Seidel et al. 2009). Experiments were conducted in eight springs located in the Pecos River watershed of the Permian Basin (Cartwright 1930), and a ninth spring (Malpais Spring) located within the endorheic Tularosa Basin—located west of the Permian Basin in New Mexico. Mean ambient salinities of these springs ranged from 0.32 ppt to 7.80 ppt but varied little throughout the study period (within-spring salinity range ≤ 3 ppt). Water temperatures at the time of experiments were 17-26 oC. I used hand sieves to collect amphipods from the water column, macrophytes, or substrata. For each experiment, I confined 10 animals in each of 12 large plastic fishbags (Jack’s Aquarium

67 & Pets), each containing 6 liters of water drawn from the respective spring. I added synthetic seasalt (Instant Ocean ®) to each bag to obtain the following nominal concentrations: ambient (A), A + 2.5 ppt, A + 5 ppt, A + 10 ppt, A + 20 ppt, and A + 40 ppt, with two replicates for each treatment concentration. I sealed all 12 bags, then placed them into the spring and covered them with a tarp, to maintain water temperatures within each bag close to ambient thermal conditions. At 0 hours and at 12-hour intervals up to 72 hours, I used a Hydrolab Surveyor 4a TM water quality instrument (Hydrolab Corporation) to record temperature ( oC) and dissolved oxygen (mg/L) to monitor for artificial deviations from ambient levels that could later confound analyses. Also at the same intervals, I visually assessed for and recorded mortality. Finally, I measured salinity (ppt) in each bag at 0 hours, and assumed salinity to be unchanging during the 72-hour experiment. For each amphipod population, I analyzed cumulative mortality data for each salinity treatment effect with probit analysis (Finney 1971) using NCSS software (Hintze 2001). For probit regressions on each population dataset, I estimated parameters including the 72-hr LC 50 (the concentration killing 50% of the target organism after 72 2 hours), Β0 (intercept) and Β1 (slope), and p-values for Pearson χ goodness-of-fit tests. I tested for equivalence of concentration-response relationships by assessing for parallelism of probit regressions among all nine populations (Oris and Bailer 1997). A

single toxicity endpoint (e.g. 72-hr LC 50 ) adequately describes the relative sensitivity to a single toxin (e.g. salt) only when the two response relationships are parallel (constant)

over the entire concentration axis, as indicated by equivalent Β1 values (Oris and Bailer

1997: fig 1a). Because no Β1 values for Gammarus population probit regressions were

observed to be equivalent, the 72-hr LC 50 values could not be used to fully characterize and compare the salinity-mortality response patterns of focal amphipod populations.

Instead, the parameters Β0 and Β1 for two populations (e.g. a and b below) were plotted on a coordinate system, from which I calculated a “physiological distance” (PD or Euclidean distance) using the formula:

I then used a single linkage version of hierarchical agglomerative clustering (Day and Edelsbrunner 1984, Zhao and Karypis 2003) to construct a tree where the length of the

68 branches reflected the PD between the pair of clusters (nearest neighbors) joined in each step of the procedure. To determine the relationship of PDs to other ecologically relevant parameters, I used a partial Mantel test (Smouse et al. 1986) to compute correlations between two 9 × 9 distance matrices, while controlling for a third 9 × 9 distance matrix. The partial Mantel test computes this correlation while assessing the extent of any autocorrelation among subjects, a property which violates the assumption of independence of observations (Cressie 1993, Fortin and Gurevitch 1993). The presence of autocorrelation can cause two variables to appear correlated when actually both are linked to a third common variable; thus, the effect of the third variable must be removed to determine if the original two variables are actually correlated (Reynolds and Houle 2002). I first used a partial Mantel test to assess for correlation between PD and genetic distance, while removing the effect of the third variable, river distance. I followed with another partial Mantel test to assess for correlation between PD and ambient salinity difference between springs, while controlling for genetic distance (F ST and pairwise nucleotide mismatch) or spatial separation (river distance and elevation difference) between population pairs. I conducted all Partial Mantel tests using the program R, v. 2.8.1 (R Development Core Team 2008).

For measures of genetic distance, I used both pairwise F ST and pairwise nucleotide mismatch values for concatenated mitochondrial gene sequences obtained from the same populations (Seidel et al. 2009). RESULTS Mortality in the salinity tolerance assays increased with increasing salinity in the treatments in all nine populations tested, and ranged from 0 (of 10) individuals dead at 72 hours (most often, but not exclusively in the ambient salinity control), to 10 (of 10) individuals dead at 72 hours (almost exclusively in the A + 40 ppt salt treatment).

Estimated LC 50 ’s ranged from 13.09 ppt to 20.79 ppt across the nine sites, with a mean of 17.39 ± 2.91 ppt ( ± SD) (Table 2). However, because of the nonparallel salinity-response probit response relationships, the majority of analyses focused on the parameters associated with the probit regressions. Among populations and observed ambient spring salinities, these physiological response parameters ranged from –4.52 to 3.34 for

69 2 intercepts (B 0), 1.34 to 7.81 for slopes (B 1), and 0.00 to 0.16 for Pearson χ p-values (associated with goodness-of-fit tests) (Table 2). Pairwise river distances between springs ranged from 4 – 580 km and average genetic distance ranged from < 1 to 19

nucleotide differences, with F ST of 0.33 – 0.99 (Table 3; adapted from Seidel et al. 2009). Partial Mantel tests indicated that neither measure of genetic distance was found to be associated with PD (Mantel r ≤ 0.023; p ≥ 0.466) after adjusting for geographic distance. Significant positive correlation was found between PD and ambient salinity difference between spring pairs when the effects of geographic (river) distance and elevation difference were removed (Mantel r = 0.493, p-value = 0.039; Mantel r = 0.503, p-value = 0.043, respectively). Similarly, significant positive correlation was also found between PD and ambient salinity difference between spring pairs when the effects of genetic distance were removed (Mantel r = 0.484, p-value = 0.049; Mantel r = 0.505, p- value = 0.041, respectively). When data from the PD distance matrix are displayed as a tree, three distinct groups of populations were visible; these groups are related by ambient salinity (Figure 2). DISCUSSION The results of this study show that the salinity-mortality response patterns of each population differ from one another in varying degrees and that populations group by ambient salinity, after accounting for geographic distance and mitochondrial similarity. Probit analyses for population-level salinity tolerance data (calculated intercept, slope and Pearson Chi Square p-value) have allowed me to characterize and compare nine populations of amphipods across the Gammarus pecos species complex in southeastern New Mexico and western Texas. This study provides field-based evidence of the kind that Schluter (1996) suggested was necessary to better understand speciation. Spring systems of southeastern New Mexico and western Texas are known to vary in salinity, both temporally and geographically (Brune 1981; Hoagstrom 2009). Natural salinity may overshadow anthropogenic salinization in the Lower Pecos River and associated water sources, although the relative contributions of naturally saline springs versus human disturbances to river salinization can be uncertain (Hoagstrom 2009). Given that the amount of salt in the Pecos River and its source springs and springbrooks is likely a historical feature (and not anthropogenic), this would have

70 permitted selection for salinity tolerance in populations of the G. pecos complex, according to local hydrochemical conditions. Several other recent studies have suggested salinity as a mechanism for selection leading to speciation of aquatic organisms. Gammarus duebeni can be found in both marine and landlocked freshwater environments (Rock et al. 2007). The taxonomic distinction between freshwater and brachish-marine forms was recognized when G. duebeni was split into two subspecies: G.d. celticus and G.d. duebeni , based on the significance of their distinctive adaptations to salt (Stock and Pinkster 1970; Rock et al. 2007). Morphological analysis confirmed the differentiation of marine and freshwater forms based on dissimilar leg anatomy (Pinkster et al. 1970; Stock and Pinkster 1970), despite evidence that both G. duebeni subspecies are capable of some salinity acclimation (Sutcliffe 1978; Sutcliffe 2000). The distinct structure shown in the physiological distance tree (Figure 2) is consistent with results of Smurov and Fokin (2001) who investigated salinity tolerance as a way to understand the phylogeny of various species in the genus Paramecium . Species of Paramecium have been analyzed using both morphometric and genetic techniques, but results have been inconsistent; however the work of Smurov and Fokin (2001) allowed the ecological grouping of these species. In both protozoans and amphipods, salinity tolerance has been used to identify and formally recognize taxonomic differences between allied groups. It is notable that clusters on the tree break into groups according to ambient habitat salinity, an outcome underscored by the significant positive correlation detected between PD and ambient salinity difference between spring pairs, even when the effects of additional variables were removed in the partial Mantel tests. Ecological speciation requires an ecological source of divergent selection (Rundle and Nosil 2005). I propose that variable salinity of spring habitats is one possible ecological source influencing divergent selection in the G. pecos species complex. Ecological speciation also requires reproductive isolation and a genetic mechanism linking the reproductive isolation to the ecological source of divergent selection (Rundle and Nosil 2005). Since an organism’s physiological capacity to respond to environmental salinity depends on cellular ion channels (quantity and activity level; Ahearn et al. 1999), and these channels have a

71 genetic basis (Csonka and Hanson 1991; Bilton et al. 2002), future research might thoroughly establish this connection for the G. pecos species complex. My results showing physiological differentiation according to ambient salinity, along with the high degree of isolation and genetic divergence among these gammarid populations (Seidel et al. 2009), are consistent with a hypothesis of ecological speciation in this group. While the tree presented here shows patterns not entirely consistent with the mitochondrial gene tree presented in Seidel et al. (2009), such deviations could be because genes for salinity tolerance were not analyzed. It seems likely that a tree based on genes for salinity tolerance would likely appear more similar to my PD tree, but no gene sequences associated with salinity tolerance are currently available. The genus Gammarus is known to be a highly successful group of amphipod crustaceans, with representative species found in marine, brackish, and freshwater habitats (Hou et al. 2007). When dispersal routes are present between habitats with different salinities, osmotic stress increases with dispersal distance, and results in a greater allocation of metabolic energy for osmotic regulation (Normant et al. 2004). Gammarus species that tolerate variable and/or extreme salinities are more successful in invading new habitats (Herkül and Kotta 2007), while other gammarids that thrive only within narrow salinity ranges remain geographically restricted according to respective osmotic requirements (Beadle and Cragg 1940, Kolding and Fenchel 1979). I suggest salinity variation across the Gammarus pecos habitat range, coupled with extreme geographic isolation, is a barrier to gene flow, and therefore, a driver of ecological speciation. Presumably, all populations within the G. pecos complex descended from a common marine ancestor (Bousfield 1958; Holsinger 1976; Baldridge 2004), and my results provide ecological context and a plausable mechanism consistent with the features of the extant G. pecos complex. Reinforced by dry desert dispersal barriers between respective springs, gene flow is likely restricted between populations of the G. pecos complex based upon incompatible hydrochemical conditions encountered by potential dispersers; selection then preserves variants that most effectively tolerate local salinity levels. Variation and speciation in the G. pecos complex seems to be driven by a combination of geographic isolation combined with genetic drift (Seidel et al. 2009) and local selection, as shown here. Patterns of

72 genetic variation are also related among a variety of other desert aquatic taxa from these spring systems (Sei et al. 2009). The results from this study are important both for the understanding of speciation and because the combined effects of isolation/drift acting in concert with local selection are likely to be important for many aquatic organisms in the Chihuahuan Desert and potentially other deserts around the world. Other models of speciation such as the reinforcement model, the divergence-with-gene-flow model, and bottleneck model have received attention along with speciation models in which polyploidization and hybridization play a central role (Rice and Hostert 1993; Coyne and Orr 2004). For prezygotic isolation by selection alone, 100 separate mathematical models have been published (Kirkpatrick and Ravigné 2002). Ecological speciation stands apart from these other models in which reproductive isolation involves key processes other than ecologically-based divergent selection (Rundle and Nosil 2005). I suggest that the results from this study not only support a hypothesis of ecological speciation within the G. pecos complex, but will also facilitate conservation planning and habitat management efforts for imperilled members of the G. pecos complex and potentially other threatened and endangered aquatic taxa sharing the same or similar habitats.

73 Literature Cited:

Ahearn GA, Duerr JM, Zhuang Z, Brown RJ, Aslamkhan A, and Killebrew DA. 1999. Ion transport processes of crustacean epithelial cells. Physiological and Biochemical Zoology 72:1-18.

Baldridge WS. 2004. Geology of the American Southwest: A Journey Through Two Billion Years of Plate-tectonic History. Cambridge University Press, UK.

Beadle LC, Cragg JB. 1940. Studies on adaptation to salinity in Gammarus spp. Journal of Experimental Biology 17:153-163.

Benlloch S, López-López A, Casamayor EO, Øvreås L, Goddard V, Daae FL, Smerdon G, Massana R, Joint I, Thingstad F, Pedrós-Alió C, Rodríguez-Valera F. 2002. Prokaryotic genetic diversity throughout the salinity gradient of a coastal solar saltern. Environmental Microbiology 4:349-360.

Bilton DT, Paula J, Bishop JDD. 2002. Dispersal, genetic differentiation and speciation in estuarine organisms. Estaurine, Coastal and Shelf Science 55:937-952.

Bousfield EL. 1958. Fresh-water amphipod crustaceans of glaciated North America. The Canadian Field-Naturalist 72: 55-113.

Bouvier TC, del Giorgio PA. 2002. Compositional changes in free-living bacterial communities along a salinity gradient in two temperate estuaries. Limnology and Oceanography 47:453-470.

Browne RA, Hoopes CW. 1990. Genotype diversity and selection in asexual brine shrimp (Artemia ). Evolution 44:1035-1051.

Brune G. 1981. Springs of Texas, vol. 1. Branch-Smith Inc., Fort Worth, Texas.

Bush GL. 1994. Sympatric speciation in animals - new wine in old bottles. Trends in Ecology and Evolution 9:285-288.

Cartwright LD. 1930. Transverse section of Permian Basin, west Texas and southeast New Mexico. Bulletin of the American Association of Petroleum Geologists 14:969-981.

Cervetto G, Guady R, Pagano M. 1999. Influence of salinity on the distribution of Acartia tonsa (Copepoda, Calanoida). Journal of Experimental Marine Biology and Ecology 239:33-45.

Colburn EA. 1988. Factors influencing species diversity in saline waters of Death Valley, USA. Hydrobiologia 158:215-226.

Cole GA. 1976. A new amphipod crustacean, Gammarus hyalelloides n. sp., from Texas.

74 Transactions of the American Microscopical Society 89:80-85.

Cole GA. 1981. Gammarus desperatus , a new species from New Mexico (Crustacea: Amphipoda). Hydrobiologia 76:27-32.

Cole GA. 1985. Analysis of the Gammarus-pecos complex (Crustacea: Amphipoda) in Texas and New Mexico, USA. Journal of the Arizona-Nevada Academy of Science 20:93-103.

Cole GA, Bousfield EL. 1970. A new freshwater Gammarus (Crustacea: Amphipoda) from western Texas. American Midland Naturalist 83:89-95.

Coyne JA, Orr HA. 2004. Speciation. Sinauer Associates, Inc., Sunderland, MA.

Cressie NAC. 1993. Statistics for Spatial Data. John Wiley and Sons, Inc.

Csonka LN, Hanson AD. 1991. Prokaryotic osmoregulation: genetics and physiology. Annual Review of Microbiology 45:569-606.

Day WHE, Edelsbrunner H. 1984. Efficient algorithms for agglomerative hierarchical clustering methods. Journal of Classification 1:7-24.

Echelle AA, Echelle AF, Edds DR. 1987. Population structure of four pupfish species (Cyprinodontidae: Cyprinodon ) from the Chihuahuan Desert region of New Mexico and Texas: allozyme variation. Copeia 1987:668-681.

Feder ME, Bennett AF, Huey RB. 2000. Evolutionary physiology. Annual Review of Ecology and Systematics 31:315-341.

Finney DJ. 1971. Probit Analysis. Cambridge University Press, Cambridge, England, 333 pp.

Fortin M, Gurevitch J. 1993. Mantel tests: spatial structure in field experiments. In: Scheiner SM, Gurevitch J. (Eds.), Design and Analysis of Ecological Experiments. Chapman & Hall.

Glenn EP, Brown JJ, O'Leary JW. 1998. Irrigating crops with seawater. Scientific American 279:76-81.

Herkül K, Kotta J. 2007. New records of the amphipods Chelicorophium curvispinum , Gammarus tigrinus , G. duebeni , and G. lacustris in the Estonian coastal sea. Proceedings of the Estonian Academy of Sciences. Biology, Ecology 56:290-296.

Hintze J. 2001. NCSS and PASS. Number Cruncher Statistical Systems. Kaysville, Utah.

Hoagstrom CW. 2009. Causes and impacts of salinization in the Lower Pecos River.

75 Great Plains Research 19:27-44.

Holsinger JR. 1976. The freshwater amphipod crustaceans (Gammaridae) of North America. U. S. Environmental Protection Agency, Water Pollution Control Research Series, 18050 ELD04/72.

Hou Z, Fu J, Li S. 2007. A molecular phylogeny of the genus Gammarus (Crustacea: Amphipoda) based on mitochondrial and nuclear gene sequences. Molecular Phylogenetics and Evolution 45:596-611.

Jian C-Y, Cheng S-Y, Chen J-C. 2003. Temperature and salinity tolerance of yellowfin sea bream, Acanthopagrus latus , at different salinity and temperature levels. Aquaculture Research 34:175-185.

Kirkpatrick M, Ravigné V. 2002. Speciation by natural and sexual selection: models and experiments. The American Naturalist 159:S22-S35.

Kolding S, Fenchel TM. 1979. Coexistence and life cycle characteristics of five species of the amphipod genus Gammarus . Oikos 33:323-327.

Lauder GV. 1988. Phylogeny and physiology. Evolution 42:1113-1114.

Moritz C, Hillis DM. 1996. Molecular systematics: context and controversies. In: Hillis M, Moritz C, Mable BK (Eds.), Molecular Systematics. Sinauer, Sunderland, MA., pp. 1- 13.

Neuparth T, Costa FO, Costa MH. 2002. Effects of temperature and salinity on life history of the marine amphipod Gammarus locusta : implications for ecotoxicological testing. Ecotoxicology 11:61-73.

Normant M, Schmolz E, Lamprecht I. 2004. Heat production rate of the Baltic amphipod Gammarus oceanicus at varying salinities. Thermochimica Acta 415:135-139.

Oren A. 2002. Diversity of halophilic microorganisms: environments, phylogeny, physiology, and applications. Journal of Industrial Microbiology & Biotechnology 28: 56-63.

Oris JT, Bailer AJ. 1997. Equivalence of concentration-response relationships in aquatic toxicology studies: testing and implications for potency estimation. Environmental Toxicology and Chemistry 16:2204-2209.

Pinkster S, Dennert AL, Stock B, Stock JH. 1970. The problem of European freshwater populations of Gammarus duebeni Liljeborg, 1852. Bijdr Dierk 40:116-147.

R Development Core Team. 2008. R: A language and environment for statistical computing, reference index version 2.8.1. R Foundation for Statistical Computing,

76 Vienna, Austria. ISBN 3-900051-07-0.

Reynolds CE, Houle G. 2002. Mantel and partial Mantel tests suggest some factors that may control the local distribution of Aster laurentianus at Îles de la Madeleine, Québec. Plant Ecology 164:19-27.

Rice WR, Hostert EE. 1993. Laboratory experiments on speciation: what have we learned in 40 years? Evolution 47:1637-1653.

Rock J, Ironside J, Potter T, Whiteley NM, Lunt DH. 2007. Phylogeography and environmental diversification of a highly adaptable marine amphipod, Gammarus duebeni . Heredity 99:102-111.

Rundle HD, Nosil P. 2005. Ecological speciation. Ecology Letters 8:336-352.

Schluter D. 1996. Ecological speciation in postglacial fishes. Philosophical Transactions of the Royal Society of London B 351:807-814.

Sei M, Lang BK, Berg DJ. 2009. Genetic and community similarities are correlated in endemic-rich springs of the northern Chihuahuan Desert. Global Ecology and Biogeography 18:192-201.

Seidel RA, BK Lang, and DJ Berg. 2009. Phylogeographic analysis reveals multiple cryptic species of amphipods (Crustacea: Amphipoda) in Chihuahuan Desert springs. Biological Conservation 142:2303-2313.

Smouse PE, Long JC, Sokal RR. 1986. Multiple regression and correlation extensions of the Mantel test of matrix correspondence. Systematic Zoology 35:627-632.

Smurov AO, Fokin SI. 2001. Use of salinity tolerance data for investigation of phylogeny of Paramecium (Ciliophora, Peniculia). Protistology 2:130-138.

Stevens LE, Meretsky VJ. 2008. Aridland Springs in North America: Ecology and Conservation (Arizona-Sonora Desert Museum Studies in Natural History). University of Arizona Press.

Stock JH, Pinkster S. 1970. Irish and French fresh water populations of Gammarus duebeni subspecifically different from brackish water populations. Nature 228:874-875.

Sutcliffe DW. 1978. Water chemistry and osmoregulation in some arthropods, especially Malacostraca. Freshwater Biology Association Annual Report 46:57-69.

Sutcliffe DW. 2000. Subspecies, morphs and clines in the amphipods Gammarus duebeni from fresh and saline waters. Freshwater Forum 13: 60-75.

Williams WD, Boulton AJ, Taaffe RG. 1990. Salinity as a determinant of salt lake fauna:

77 a question of scale. Hydrobiologia 197:257-266.

Zhao Y, Karypis G. 2003. Clustering in life sciences. In: Khodursky A, Brownstein M (Eds.), Functional Genomics: Methods and Protocols. Humana Press. Pp. 183-218.

78 Acknowledgments

We thank Todd Levine, Emy Monroe, and Makiri Sei (Department of Zoology, Miami University) for their valuable input and suggestions during the development of this project. Access and permission to collect and conduct field experiments on public and private lands was granted by: Jeff Howland (BLNWR), David Riskind and Tom Johnson (TPWD), Larry Paul (Lincoln National Forest), Patricia Griffin, Junior Kerns and Robert Myers (WSMR), and John Karges and Jason Wrinkle (TNC-TX). We also thank Miami University undergraduates Allison Fralic, Vybhav Jetty, Kate Moran, and Kirk Weber for their assistance. Funding was provided by the New Mexico Department of Game and Fish (Contract No. 04-516.0000-0093), U. S. Fish and Wildlife Service (Region 2), the Crustacean Society, and the Miami University Field Research Workshop. This report has been been approved for public release by White Sands Missile Range for unlimited distribution; the White Sands Missile Range Operational Security review was completed 25 August 2009.

79 Figure Legends

Figure 1. Locations for 9 extant populations of the Gammarus pecos species complex from the Chihuahuan Desert used in the field experiments (populations codes identified in Table 1).

Figure 2. Tree showing the physiological distances, PDs, between pairs of populations in the Gammarus pecos species complex, as assessed by salinity tolerance assay. Clades are labeled at right with corresponding ambient habitat salinities, which clustered according to low, medium, and high levels.

80 81 82 Table 1. Sample sites and population codes for the Gammarus pecos species complex, as identified by Cole (1985). All individuals analyzed during this study were collected within 300 m of the coordinates shown.

Population Cole (1985) Sample Site Location Latitude / Longitude Nominal Species Code Designation BLHM “D” Hunter Marsh, Bitter Lake National Wildlife Refuge, Chaves 33 o 24 ΄ 52˝ N / 104 o 25 ΄ 16˝ W Gammarus desperatus County, NM BLSS “D” Sago Spring, Bitter Lake National Wildlife Refuge, Chaves 33 o 28 ΄ 41˝ N / 104 o 25 ΄ 11˝ W Gammarus desperatus County, NM BLU6 “D” Unit 6, Bitter Lake National Wildlife Refuge, Chaves County, 33 o 26 ΄ 46˝ N / 104 o 24 ΄ 16˝ W Gammarus desperatus NM CS - Caroline Spring (also called “T5”), Independence Creek, 30 o 26 ΄ 40˝ N / 101 o 43 ΄ 13˝ W Gammarus sp. Terrell County, TX DY “P” Diamond Y Spring (also called “Willbank Spring”), Diamond 31 o 02 ΄ 12˝ N / 102 o 53 ΄ 27˝ W Gammarus pecos Y Draw, Pecos County, TX ESS - East Sandia Spring, Toyah Creek, Jeff Davis County, TX 30 o 59 ΄ 28˝ N / 103 o 43 ΄ 44˝ W Gammarus sp. MS - Malpais Spring, White Sands Missile Range, Otero County, 33 o 17 ΄ 18˝ N / 106 o 18 ΄ 33˝ W Gammarus sp. NM SB “E” Sitting Bull Spring, Lincoln National Forest, Eddy County, 32 o 14 ΄ 12˝ N / 104 o 42 ΄ 08˝ W Gammarus sp. NM SSS “S” San Solomon Spring, Toyah Creek, Jeff Davis County, TX 30 o 56 ΄ 41˝ N / 103 o 47 ΄ 11˝ W Gammarus pecos

83 Table 2: Physiological response parameters estimated using probit analysis, displayed by population code and ambient salinity. P-values included in this table are associated with Pearson P2 goodness-of-fit tests.

Population LC 50 Intercept (B 0) Slope (B 1) p-value Ambient Salinity

BLHM 15.63 0.02 4.17 0.16 4.72 ppt

BLSS 20.79 - 1.68 5.07 0.04 5.88 ppt

BLU6 20.65 1.84 2.16 0.07 2.70 ppt

CS 13.09 2.28 2.43 0.00 0.45 ppt

DY 16.54 - 4.52 7.81 0.05 7.80 ppt

ESS 18.62 1.57 2.70 0.00 2.54 ppt

MS 20.29 - 0.28 4.04 0.00 3.98 ppt

SB 17.14 3.34 1.34 0.00 0.32 ppt

SSS 13.78 0.78 3.71 0.00 2.12 ppt

84 Table 3. Above diagonal: Pairwise river distances (km) between respective springs, measured along spring runs and the Pecos River. Malpais Spring (MS) is not hydrologically connected to other springs (the symbol “—” means no river distance estimated). Below diagonal: Average pairwise nucleotide differences (percentage difference) between populations, with corresponding F ST values in paretheses (adapted from Seidel et al. 2009). F ST theoretical minimum of 0 means no genetic divergence due to high gene flow, and theoretical maximum of 1 means complete fixation for different alleles in population due to absence of gene flow.

BLHM BLSS BLU6 CS DY ESS MS SB SSS BLHM − 10.16 4.52 567.44 494.58 403.39 − 229.82 413.93 BLSS 7.64 (0.94) − 5.64 577.01 504.15 412.96 − 239.39 423.50 BLU6 7.35 (0.89) 0.96 (0.44) − 571.37 498.51 407.32 − 233.75 417.86 CS 6.52 (0.94) 7.11 (0.95) 7.55 (0.91) − 204.02 326.61 − 458.44 337.15 DY 7.90 (0.97) 8.61 (0.98) 8.96 (0.95) 5.20 (0.96) − 253.75 − 385.58 264.29 ESS 7.81 (0.98) 8.82 (0.98) 9.04 (0.95) 4.43 (0.96) 6.09 (0.99) − − 294.39 12.42 MS 7.81 (0.96) 18.04 (0.98) 18.10 (0.97) 10.01 (0.97) 11.09 (0.99) 10.59 (0.99) − − − SB 8.14 (0.90) 18.71 (0.94) 18.75 (0.94) 10.46 (0.93) 11.52 (0.94) 11.07 (0.94) 0.90 (0.33) − 304.93 SSS 7.92 (0.96) 8.97 (0.97) 9.20 (0.94) 4.55 (0.94) 6.21 (0.98) 0.20 (0.35) 10.79 (0.98) 11.26 (0.94) −

85 General Conclusion

The three chapters presented in this dissertation underscore the cryptic diversity present in aridland spring systems in the American southwest. In Chapter 1, I identify seven novel species of spring amphipods in the Gammarus pecos species complex of the northern Chihuahan Desert, a group in which only three species were previously described. Chapter 2 investigates species boundaries and units of conservation detectable using molecular genetic and morphological comparisons, and reveals that results for both categories of characters should still be pursued and examined for congruence. Considering Chapter 1 and Chapter 2 results together, I show that multiple markers should be used when possible, to increase accuracy by sampling a larger portion of the genome. Finally, the results of Chapter 3 show that selection based on salinity tolerance can drive ecological speciation, while at the same time the neutral process of genetic drift enhances the distinctiveness of taxa separated by hot desert dispersal barriers. Conservation planning by relevant governmental agencies can now proceed using the results of my investigations, with the ultimate goal of managing and protecting spring animals with unique genetic and ecological characteristics. Future biodiversity investigations in the Chihuahuan Desert will almost certainly uncover even more species meriting protection among the region’s highly endemic fauna, given previously undetected variation and the high degree of crypsis among invertebrates (Remerie et al., 2006), particularly within aquatic habitats (Müller, 2000; Pfenninger et al., 2003; Witt et al., 2003) and those specifically located in xeric regions (Thomas et al., 1998; Sei et al., 2009). Future discoveries made in the aquatic habitats of arid regions will likely be made based upon trends uncovered by this dissertation research. My discoveries presented here will increase our understanding of, 1) the general nature of biodiversity, 2) the distribution of its components across landscapes, and 3) how to detect and quantify units of biodiversity.

86 Literature Cited: Müller J. 2000. Mitochondrial DNA variation and the evolutionary history of cryptic Gammarus fossarum types. Molecular Phylogenetics and Evolution 15:260-268.

Pfenninger M, Staubach S, Albrecht C, Streit B, Schwenk K. 2003. Ecological and morphological differentiation among cryptic evolutionary lineages in freshwater limpets of the nominal form-group Ancyclus fluviatilis (O.F. Muller, 1774). Molecular Ecology 12:2731-2745.

Remerie T, Bourgois T, Peelaers D, Vierstraete A, Vanfleteren J, Vanreusel A. 2006. Phylogeographic patterns of the mysid Mesopodopsis slabberi (Crustacean: Mysida) in Western Europe: evidence for high molecular diversity and cryptic speciation. Marine Biology 149:465-481.

Sei M, Lang BK, Berg DJ. 2009. Genetic and community similarities are correlated in endemic-rich springs of the northern Chihuahuan Desert. Global Ecology and Biogeography 18:192-201.

Thomas EP, Blinn DW, Keim P. 1998. Do xeric landscapes increase genetic divergence in aquatic ecosystems? Freshwater Biology 40:587-593.

Witt JDS, Blinn DW, Hebert PDN. 2003. The recent evolutionary origin of the phenotypically novel amphipod Hyalella montezuma offers an ecological explanation for morphological stasis in a closely allied species complex. Molecular Ecology 12:405-413.

87 Appendix I

88 89 Appendix II Probit plots for all nine sites at which salinity tolerance was assessed (using cummulative mortality at 72-hour interval). “Empirical probit” displayed on the y-axis refers to estimated probit values obtained in the anlysis.

2a. Bitter Lake, Hunter Marsh (BLHM)

2b. Bitter Lake, Sago Spring (BLSS)

90 2c. Bitter Lake, Unit 6 (BLU6)

2d. Caroline Spring (CS)

91 2e. Diamond Y Spring (DY)

2f. East Sandia Spring (ESS)

92 2g. Malpais Spring (MS)

2h. Sitting Bull Spring (SB)

93 2i. San Solomon Spring (SSS)

94 Appendix III Raw mortality data for salinity tolerance treatments, by site, for all time intervals.

3a. Bitter Lake, Hunter Marsh (BLHM)

3b. Bitter Lake, Sago Spring (BLSS)

3c. Bitter Lake, Unit 6 (BLU6)

95 3d. Caroline Spring (CS)

3e. Diamond Y Spring (DY)

3f. East Sandia Spring (ESS)

3g. Malpais Spring (MS)

96 3h. Sitting Bull Spring (SB)

3i. San Solomon Spring (SSS)

97