Florida State University Libraries

Electronic Theses, Treatises and Dissertations The Graduate School

2012

Population Genetics of Pseudacris FeriarumMoses J. Michelsohn

Follow this and additional works at the FSU Digital Library. For more information, please contact lib-ir@fsu.edu

THE FLORIDA STATE UNIVERSITY

COLLEGE OF ARTS AND SCIENCES

POPULATION GENETICS OF PSEUDACRIS FERIARUM

By

MOSES J. MICHELSOHN

A Thesis submitted to the

Department of Biological Science

in partial fulfillment of the

requirements for the degree of

Master of Science

Degree Awarded:

Spring Semester, 2012

ii

Moses J. Michelsohn defended this thesis on April 2, 2012.

The members of the supervisory committee were:

Emily Moriarty Lemmon

Professor Directing Thesis

Scott J. Steppan

Committee Member

Peter Beerli

Committee Member

The Graduate School has verified and approved the above-named committee members, and

certifies that the thesis has been approved in accordance with university requirements.

iii

This thesis is dedicated to all the Eckerd College folks that helped me get here:

Steven Denison, Elizabeth Forys, Yvonne King, William Szelistowski, & Peter Meylan

iv

ACKNOWLEDGEMENTS

I am very grateful for the guidance of my supervisory committee: Emily Lemmon, Scott

Steppan, and Peter Beerli. They have provided me with countless conversations, meaningful

suggestions, and the encouragement that was necessary for me to complete this project. This

work was supported by a Theodore Roosevelt Memorial Grant from the American Museum of

Natural History and by the Florida State University Department of Biological Science.

I am indebted to the many people who assisted me with specimen and tissue collection

including: L. N. Barrow, A. M. Heupel, A. R. Lemmon, E. M. Lemmon, D. B. Means, S. O.

Sunderman, A. R. Warwick. I greatly appreciate the laboratory help and training I received from

M. E. Bedwell, S. A. Emme, H. Ralicki, and S. Miller; as well as the help with analyses I

received from P. Beerli, K. E. Lotterhos, and M. Palczewski. I am especially thankful for the

encouragement, numerous manuscript revisions, and field assistance (including far too many

nights driving endlessly around Dothan) provided by Lisa Barrow.

v

TABLE OF CONTENTS

List of Tables ................................................................................................................................. vi

List of Figures ............................................................................................................................... vii

Abstract ........................................................................................................................................ viii

1. A TEST OF WALLACE'S RIVERINE BARRIER HYPOTHESIS: MODELING THE

EFFECTS OF THE APALACHICOLA-CHATTAHOOCHEE RIVER ON GENE MIGRATION

IN PSEUDACRIS FERIARUM ........................................................................................................1

1.1 Introduction ....................................................................................................................1

1.1.1 Wallace's Riverine Barrier Hypotheses .............................................................2

1.1.2 Oxbow Models ...................................................................................................5

1.1.3 Flood Model .......................................................................................................5

1.1.4 Additional Models .............................................................................................6

1.1.5 Study System .....................................................................................................6

1.2 Methods..........................................................................................................................7

1.3 Results ............................................................................................................................9

1.4 Discussion ....................................................................................................................10

1.4.1 Methods Used in Evaluating Riverine Effects .................................................12

1.4.2 Future Directions .............................................................................................15

2. CASCADING SPECIATION: EVIDENCE FOR POPULATION GENETIC EFFECTS OF

REPRODUCTIVE CHARACTER DISPLACEMENT IN PSEUDACRIS FERIARUM ..............17

2.1 Introduction ..................................................................................................................17

2.1.1 Study System ...................................................................................................19

2.2 Methods........................................................................................................................20

2.3 Results ..........................................................................................................................22

2.4 Discussion ....................................................................................................................22

TABLES ........................................................................................................................................25

FIGURES .......................................................................................................................................40

APPENDIX A SPECIMEN COLLECTION INFORMATION ...................................................44

APPENDIX B PROGRAM μSCOPE ...........................................................................................47

APPENDIX C 454-GENERATED PRIMER DEVELOPEMENT ..............................................49

APPENDIX D MULTIPLEX PCR PRIMER SETS & PROTOCOLS ........................................51

REFERENCES ..............................................................................................................................53

BIOGRAPHICAL SKETCH .........................................................................................................63

vi

LIST OF TABLES

1 A comparison of riverine barrier hypotheses ........................................................................25

2 Collection data for Pseudacris feriarum tissues ....................................................................26

3 Results of Hardy-Weinberg exact tests (p-values) and linkage disequilibrium probability

tests (p-values) conducted in Genepop v4.1 ..........................................................................31

4 Distance matrix of all inter-population distances used for creating geofiles ........................31

5 Bezier approximated marginal likelihoods (ln mL) of three models testing the effect of a

river on gene migration .........................................................................................................32

6 Parameter estimates for models of cross-river gene flow .....................................................32

7 Bezier approximated marginal likelihoods (ln mL) of three models testing riverine effects

on gene migration along a river .............................................................................................33

8 Parameter estimates for models of riverine gene flow ..........................................................34

9 Bezier approximated marginal likelihood (ln mL) results for tests of gene flow across a

region of reproductive character displacement ......................................................................35

10 Parameter estimates for models of gene flow across a region of reproductive character

displacement ..........................................................................................................................35

11 Microsatellite primer sequences developed from 454 reads .................................................36

12 Results of microsatellite primer amplification in a panel of Pseudacris feriarum selected

from across the range of the species ......................................................................................37

13 Primer sequences for sixteen microsatellite loci in Pseudacris feriarum ............................38

14 Multiplex composition used to genotype individuals for sixteen loci ...................................39

vii

LIST OF FIGURES

1 Models of cross-river gene migration ....................................................................................40

2 Models of riverine effects on gene migration ........................................................................40

3 Map of sampling localities ....................................................................................................42

4 Models of gene migration across a region of reproductive character displacement .............43

viii

ABSTRACT

Many genetic patterns observed within and between species are often attributed to

processes that affect interpopulation genetic exchange. These patterns are often taken as

evidence of the genetic processes without explicit tests of the population genetic dynamics

operating within species. The first chapter of this thesis uses a population genetic approach to

test Wallace's riverine barrier hypothesis, a 150-year-old theory that has largely been based on

interpretation of broad scale patterns rather than focused studies of the process. This work helps

clarify the definitions of many riverine hypotheses and uses a Bayesian model comparison

approach to test these hypotheses in Pseudacris feriarum, the upland chorus frog, along the

Apalachicola-Chattahoochee River using eleven microsatellite loci. A flood model of gene

migration best explains riverine effects in this species. This model is proposed as an alternative

way to think about riverine effects on gene flow that should be tested more broadly. The second

chapter of this thesis builds on previous observations of reproductive character displacement and

its effects on speciation. A Bayesian model testing approach is used to determine if predictions

based on female choice experiments lead to population differentiation. Eleven microsatellite loci

are used to model gene migration across a region of reproductive character displacement. This

approach provides evidence for the genetic consequences expected under a speciation cascade

model of taxon diversification.

1

CHAPTER ONE

A TEST OF WALLACE'S RIVERINE BARRIER HYPOTHESIS:

MODELING THE EFFECTS OF THE APALACHICOLA-

CHATTAHOOCHEE RIVER ON GENE MIGRATION IN

PSEUDACRIS FERIARUM

Introduction

Many geographic features are believed to act as genetic barriers, potentially capable of

leading to vicariant speciation (Mayr 1942; Cf. Coyne and Orr 2004). Rivers have often been

noted as barriers between lineages (reviewed in Colwell 2000), although the literature on this

subject often confuses observed patterns with the processes that create them. The interpretation

of Wallace's riverine barrier hypothesis is also muddled because the name has been applied to

many hypotheses concerning riverine effects on populations. Additionally, there has been a lack

of statistical rigor in the treatment of riverine effects (Knowles & Maddison 2002). This study

used a Bayesian model comparison approach (Beerli & Palczewski 2010) to investigate the

effects of the Apalachicola-Chattahoochee River on the population genetics of Pseudacris

feriarum, the upland chorus frog, as a model for evaluating population genetic processes that

lead to observed genetic patterns.

The phrase 'Wallace's riverine barrier hypothesis' has been used to describe a number of

related models describing the effects rivers have on their adjacent floras and faunas. These

models have often been poorly defined in the scientific literature, leading to inadequate tests of

their effects (but see Patton et al. 2004; Funk et al. 2007). In the following sections I will

summarize and define the tenets of each of these theories and present a few additional models

predicting the possible influence of rivers on gene movement.

2

Wallace's Riverine Barrier Hypotheses

Alfred Russel Wallace became the eponym of riverine barrier hypotheses due to

comments made in an 1852 address to the Zoological Society of London about Primates

observed in South America:

During my residence in the Amazon district I took every opportunity of determining the

limits of species, and I soon found that the Amazon, the Rio Negro and the Madeira

formed the limits beyond which certain species never passed.… On approaching the

sources of the rivers they cease to be a boundary, and most of the species are found on

both sides of them.

Wallace's description indicates a view that species' range limits are often bounded by major

rivers (Wallace 1876), but that the rivers' sources offer little impediment to movement, although

Wallace does not speculate whether this is an effect of river width or contiguous populations

above the headwaters. Wallace's confidence in this effect of rivers was so great that he proposed

the division of Amazonian fauna into four districts defined by major rivers (Wallace 1852). This

pattern of rivers acting as barriers in the Amazon was also noted to occur in Lepidoptera by

Wallace's fellow naturalist Henry Walter Bates (1863).

Over a century later, Hershkovitz (1977) expanded on Wallace's range demarcations of

primates noting that although not all species complexes share the same internal boundaries (e.g..,

Saguinus fuscicollis complex v. S. mystax complex), rivers divide "well-differentiated races" in

each case. A more recent study (Ayres & Clutton-Brock 1992) demonstrated a negative

correlation between annual river discharge and primate community similarity on opposite banks

of major Amazonian rivers. This first hypothesis addressing the effects of rivers represents a

view that rivers act as barriers that impede invasion by adjacent populations. These patterns of

subtaxa division along rivers are consistent with Wallace's (1876) statements and form the tenets

of what I will refer to as Wallace's Riverine Barrier Hypothesis (Table 1, WRBH).

Patton et al. (1994) reinterpreted the Riverine Barrier Hypothesis sensu Wallace (WRBH)

as a theory about gene movement. This hypothesis, which I will refer to as the Genetic Riverine

Barrier Hypothesis (Table 1, GRBH), states that genetic divergence between populations

spanning a river should be positively correlated with increasing river width. In this hypothesis,

rivers act as barriers to interpopulation dispersal within a species complex. Rivers may act as

3

either complete or incomplete barriers with the expectation that genetic similarity should

increase upstream where successful cross-river dispersal events are more common.

Funk et al. (2007) modified the GRBH with the additional condition that rivers must act

as regions of primary diversification within a lineage, a hypothesis I will refer to as the Primary

Diversification Riverine Barrier Hypothesis (Table 1, PDRBH). This definition is more

restrictive in that it assigns a river the role of initiating allopatric diversification rather than

simply acting to reduce gene migration between populations which may have already diversified

in isolation. Funk et al. (2007) outlines four testable predictions of the PDRBH:

(1) populations [that are] reciprocally monophyletic… will occur on opposite sides of

major… rivers after sufficient time has elapsed for lineage sorting

(2) genetic divergence between populations will be positively correlated with the

presence of intervening rivers after removing the effects of geographic distance

(3) genetic divergence between populations separated by a river will be greater in the

river's lower section than in the headwaters

(4) rivers will act as areas of primary differentiation rather than secondary contact of

lineages

Predictions (2) and (3) are consistent with the GRBH, while prediction (4) is unique to the

PDRBH. Prediction (1) describes a pattern (reciprocal monophyly of lineages on opposite

banks), which could only occur if all cross-river gene migration has ceased long enough to yield

reciprocally monophyletic lineages. In a coalescent population genetics framework, prediction

(1) precludes prediction (3) that genetic similarity will be greater upstream, although the

prediction could potentially be maintained if only testing statistical relationships rather than

using a genealogical approach. Prediction (1) made in Funk et al. (2007) referred specifically to

mtDNA haplotypes, a situation which could rectify predictions (1) and (3) if only male

individuals disperse and nuclear genes are included in analyses. In this situation mtDNA

lineages could diverge to reciprocal monophyly while the nuDNA continues to be exchanged

between cross-river populations. This situation does not represent a general case amongst all

organisms so I do not include prediction (1) in my definition of PDRBH (Table 1).

Haffer (1969) suggested a role of rivers as regions that help stabilize regions of secondary

contact. If rivers act to impede gene migration then secondary contact zones may drift to match

4

these regions of migration paucity (Barton & Hewitt 1985). Whether rivers act as primary or

secondary sites of speciation has no effect on the process of riverine barriers impeding gene

migration in proportion to river width. For this reason I focus on measuring the contemporary

effects of a riverine barrier using the GRBH as a model for gene migration. Under the GRBH

gene migration rates are expected to be (A) reduced in cross-river routes when compared to

routes that do not cross rivers after removing the effects of geographic distance; and (B) higher

near a river's headwaters than further downstream.

In this study I used a Bayesian model comparison approach to test the expectations of the

GRBH. Two sets of models were used to test this theory, one for each expectation. In addition I

tested other models of gene migration to investigate the gene flow processes that lead to the

patterns of genetic diversity observed along and across rivers. This model testing approach

emphasizes modeling of process rather than detection of pattern attempting to more accurately

reconstruct the effects on a river.

A series of models related to expectation (A) are presented in Figure 1. This first series

of models represents populations spanning a river exchanging migrants with one another. Each

arrow in the figure represents a mutation-scaled migration rate, M, while each circle represents a

population with a mutation-scaled population size, Θ. The Full Migration model (Fig. 1 A)

represents a parameter rich model that would be needed to explain a complicated history of gene

flow amongst these populations. The Equivalent Linear Migration model (Fig. 1 B) is a direct

violation of expectation (A) in that it reduces all M parameters to one shared rate. This model

represents a situation where rivers do not impede inter-population gene flow at any different rate

than regions without rivers spanning them. The Linear River Effect model (Fig. 1 C) restricts the

migration rate between populations without rivers between them to one M parameter and those

that do span a river to a separate set of M rates. This essentially models a situation where all

overland migration occurs at a constant rate, but cross-river migration rates are unrestricted by

the model. This first set of models will be used to determine if a river acts to reduce gene

migration between populations.

Figure 2 presents a series of models of riverine effects on population genetics. These

models are created based on the assumption that a river acts to reduce gene flow between

populations. The independent models presented each represent a different view on the specific

directional effects a river might have on cross-river genetic connectivity. The figure contains

5

two models of the GRBH that include expectation (B). The first model (Fig. 2 A) assumes that a

river acts as a complete barrier to gene migration and that cross-river dispersal events occur only

at the river's headwaters. The second model (Fig. 2 B) assumes that the river acts as an

incomplete barrier allowing gene flow along its length inversely proportional to river width.

Additional gene migration models that are inconsistent with the GRBH are described below.

Oxbow Models

River meandering and subsequent oxbow formation can act as a means of passive transfer

of biota across a river. A population that is passively transferred may form a genetic enclave on

an opposite bank. This population may interbreed with local populations leading to a genetic

pattern identical to that generated by individuals actively crossing the river (Hershkovitz 1983).

Rivers meander over low gradients which increase in frequency lower in a stream course. This

increased sinuosity downstream causes a directional permeability of a river that is counter to

river barrier hypotheses, i.e. cross-river genetic connectivity is expected to increase downstream.

An oxbow model of riverine effects still requires that a river must act as a barrier to gene

migration, an expectation that can be tested using cross-river population models (Fig. 1). Two

oxbow models are presented in Figure 2. The first model (Fig. 2 C) assumes that passive transfer

occurs only at the lowest regions of the river while the other model (Fig. 2 D) allows for either

passive or active gene migration along the length of the river with increased frequency

downstream.

Flood Model

Another alternative model of gene migration across rivers has been proposed for plants

by Jacquemyn et al. (2006). This model is based on periodic river flooding events causing long

range dispersal of progeny. Individuals that raft downstream have increased chances of crossing

to the other bank the further they travel. This model (Fig. 2 E) predicts genetic similarity in

cross-river downstream populations as a result of immigration from populations on both sides of

a river further upstream. This model has been used to explain genetic structuring in plants

(Jacquemyn et al. 2006; Van Looy et al. 2009) and should be considered as an alternative to

oxbow models in creating downstream genetic connectivity patterns.

6

Additional Models

An equivalent migration model (Fig. 2 F) was specifically designed as a competitor to

both riverine barrier and oxbow models in that it restricts cross-river migration rates to be

equivalent along the entire river thus removing any directional effects. A nearest neighbor

model (Fig. 2 G) represents a complex alternative to all other models. This model is parameter

rich and should only be favored if true migration rates do not match any other hypothesized

model due to irregular patterns of gene flow in a system.

Study System

The genetic effects of rivers were tested in Pseudacris feriarum, the upland chorus frog,

along the Apalachicola-Chattahoochee River. P. feriarum is a wide-ranging species distributed

across the Eastern United States. P. feriarum is generally restricted to regions above the Fall

Line in the Southeastern United States, but its range extends deep into the Gulf Coastal Plain

along the Apalachicola River (Fig. 3). In this region P. feriarum breeds in ephemeral cypress or

gum swamps along the floodplain of the Apalachicola, while further north along the

Chattahoochee River it breeds in ditches, flooded fields, and open woods (Carr & Goin 1959;

Crenshaw & Blair, 1959; Mount 1975; Jensen et al. 2008). Pseudacris do not move far from a

breeding locality during the breeding season, less than 250 meters in P. triseriata (Kramer 1973),

and have generally small home ranges, less than two acres (Kramer 1974).

The region around the Apalachicola-Chattahoochee River system has been shown to be a

contact zone or range limit for numerous organisms (Blaney 1971; Means 1977; Engle &

Summers 2000; Marshall et al. 2000; Gonzales & Hamrick 2005; Swenson & Howard 2005;

Soltis et al. 2006; Pauly et al. 2007; Walker et al. 2009). The Apalachicola-Chattahoochee

River is one of a few Southeastern rivers originating above the coastal plain, completely

bisecting the coastal plain, thus presenting a potentially complete barrier to coastal plain species

(Pauly et al. 2007). The Apalachicola-Chattahoochee River flows through what was once an

extensive embayment that possibly extended all the way to the Fall Line (Neill 1957), thus the

lower Apalachicola-Chattahoochee River may act as a secondary contact zone between lineages

separated by a once greater saltwater barrier (Ellsworth et al. 1994; Burbrink et al. 2000; Church

et al. 2003). More recently, a large passive transfer of populations across the lower Apalachicola-

Chattahoochee River may have occurred during the Apalachicola stream capture events

7

(Randazzo & Jones 1997). These genetic consequences of these events may be teased apart by

other methods, but they are considered to be too distant in the past to leave signatures in high

mutation-rate markers such as microsatellites. More recent meandering of the Apalachicola-

Chattahoochee River may have led to passive transfer of populations across the river through

oxbowing events. Oxbowing and dispersal (either active or passive) are the only processes

expected to have affected gene migration during the contemporary history of the Apalachicola-

Chattahoochee River.

Many previous studies on the effects of rivers have suffered from inappropriate river or

organism choice (See Chapter 1 Discussion for a review of past studies). Pseudacris feriarum

along the Apalachicola-Chattahoochee River provides an ideal test of the genetic effects of rivers

because the history of the river has been relatively well characterized (Neill 1957; Randazzo &

Jones 1997) and the study organism is likely to experience the river as a major barrier to

dispersal due to its size and dispersal capabilities (Kramer 1973).

Methods

Pseudacris feriarum were collected from eight populations along the Apalachicola-

Chattahoochee River in Florida, Georgia and Alabama (Fig. 3). Tissue samples were obtained

from twenty individuals per population using either toe-clipping or dissection (Table 2;

Appendix A). Voucher specimens from each site were preserved in formalin. Populations were

sampled in two transects across the Apalachicola-Chattahoochee River: a Northern transect

consisting of populations in Macon and Russell Counties, Alabama, and Muscogee County,

Georgia; and a Southern transect consisting of populations in Houston County, Alabama, and

Early and Miller Counties, Georgia (Fig. 3). Two populations (Gulf and Liberty Counties, FL;

Fig. 3) were sampled from the southern extension of P. feriarum's range along the Apalachicola

River.

Genomic DNA was extracted from all 160 P. feriarum individuals using an E.Z.N.A. Gel

Extraction Kit (Omega Bio-Tek). Each individual was genotyped for a total of eleven

microsatellite loci; seven tetramer loci using published primer sets (Lemmon et al. 2011) and

primer sets developed using program μScope v0.0.6 (Appendix B) on previously generated 454

data (Lemmon & Lemmon unpublished; Appendix C), as well as for four dimer loci (Bedwell &

Lemmon unpublished; Ralicki & Lemmon unpublished). Details of primer development are

8

listed in Appendix C. Multiplex PCR reactions were set up using a Multiplex PCR Kit

(QIAGEN) and amplified using protocols described in Appendix D. Fragment analysis was

conducted on an Applied Biosystems 3730 Genetic Analyzer with Capillary Electrophoresis at

the Florida State University Core Research Facilities. Microsatellite fragment lengths were

measured using GeneMapper v4.1 (Applied Biosystems). Allele calling was done using tandem

v1.08 (Matschiner & Salzburger 2009). These allele calls were used to test for Hardy-Weinberg

equilibrium and linkage disequilibrium (Table 3) using Genepop v4.1 (Rousset 2008). Two loci

(D_C08b and A_E09d) exhibited linkage disequilibrium with the marker D_D12b, although not

with each other (Table 3). Lemmon et al. (2011) found no evidence of linkage disequilibrium

between D_D12b and D_C08b across 85 P. feriarum sampled from Macon County, AL. All

three markers were included in further analyses.

Fragment lengths were used to write a series of microsatellite data files for Migrate-n

v3.2.17 (Beerli 2009). Allele binning was performed by Migrate-n using the "#@M" command

in all data files. Migrate data files were written for the Northern transect and Southern transect

to test the effects of cross-river migration (Fig. 1). A third data file was written to test the

riverine effects on population genetics (Fig. 2). This data file included six populations along the

river: Liberty and Gulf Counties, FL; Houston and Russell Counties, AL; and Muscogee and

Early Counties, GA. (Fig. 3). Model parameters were set in Migrate-n using the user-specified

model options. All Migrate-n runs were conducted using a continuous Brownian-motion model

of microsatellite evolution (Felsenstein 2004). Locality information (Table 4) was used to scale

migration rate parameter estimates using interpopulation geographic distance. A strait line

distance matrix (geofile) was imported into Migrate-n to scale by these interpopulation

geographic distances.

Migrate-n run length was determined by setting up successive runs of the most parameter

rich model, the Nearest Neighbor model (Fig. 2 G). Run convergence and parameter prior

ranges were determined using posterior distribution plots of parameter values summed over ten

independent runs output by Migrate-n. A Uniform prior with a range of 0–400 was used for

estimating mutation-scaled migration rates, M, and a Uniform prior with a range of 0–100 was

used for estimating mutation-scaled population size, Θ. Parameter space for each Migrate-n

model was searched with a burn-in of 4,000,000 generations followed by 10,000 steps that were

9

sampled at iterations of every 100 generations to reduce autocorrelation for a grand total of

50,000,000 parameter value iterations explored per long chain. A Metropolis-coupled Markov

chain Monte Carlo (MCMCMC) approach (Geyer 1991) utilizing 4 adaptively heated chains was

implemented to better explore parameter space (Beerli 2006). Parameter values for each model

were sampled using 50 independent long chain samplings. Parameter value distributions were

summed over the fifty independent runs. The natural log marginal likelihood of each model

was estimated using Bézier quadrature (Beerli and Palczewski 2010), an alternative to the

harmonic mean which gives better estimates of the distribution when few heated chains are used.

The natural log marginal likelihood of each model was used to compute natural log Bayes factors

(BF: Kass & Raftery 1995) for model ranking and comparison.

Results

The comparison of cross-river gene migration models yielded favored support (>6,000

Bayes Factors (BF) for each transect) for models of restricted cross-river migration (Fig. 1 C

Linear River Effect) over migration models which held migration constant between populations

regardless of whether the migration route crossed the Apalachicola-Chattahoochee River (Fig. 1

B Equivalent Linear Migration). This same pattern of model support was observed in both the

Northern and Southern transects (Table 5). This supports the idea that rivers affect gene flow

differently from overland migration.

In each transect the Full Linear Migration model was favored overall by >3,300 BF

(Table 5), indicating support for a complex gene flow history in both transects. The model

parameter values estimated for this model in each transect indicate that there is directional skew

in overland migration. This skew may explain why the two models that restricted overland

migration rates (Fig. 1 B & C) received low model support values (Table 5). The cross-river M

parameter estimates in the favored Full Linear Migration model (Table 6) of both transects are

not recognizably lower than the overland M parameter estimates until they are converted to

numbers of migrants using the formula: ( ). In the Northern transect 120.9

individuals/generation migrate cross-river between populations while 2109.8

individuals/generation migrate between populations on the same side of the river. In the

Southern transect 251.5 individuals/generation migrate between cross-river populations while

10

271.5 individuals/generation migrate between same bank populations. This reduced number of

migrants per generation, especially in the Northern transect, indicates that the Apalachicola-

Chattahoochee does act to retard gene flow allowing for comparison of models testing the

directional effects of a river on gene migration.

The comparison of models explaining how rivers affect gene flow (Fig. 2) resulted in

overwhelming support (1,738 BF better than the second best model) for a flood model (Fig. 2 E)

of gene flow (Table 7). The worst riverine effects model was the Equal Migration model which

held cross-river migration constant along the entire river shed. This worst model exhibited far

less support (12,438 BF) than the most parameter rich model (Fig. 2 G). Parameter estimates for

all models are listed in Table 8.

Discussion

The first test of genetic effects of rivers matched the expectation that rivers act to retard

gene flow (Table 5; Funk et al. 2007). The Equivalent Linear Migration (Fig. 1 B) model

exhibited the lowest support in both transects (Table 5), while the Full Linear Migration (Fig. 1

A) model exhibited the highest support (Table 5). When the parameter estimates of the most

parameter rich model, the Full Linear Migration model (Fig. 1 A), are examined (Table 8) one

can see why this complicated model received high support. Asymmetry in migration rate

between populations on the same of the river is apparent (Table 6). This asymmetry is the likely

cause for preference of the Full Linear Migration model, the only model which does not restrict

overland migration to being symmetrical.

Another interesting pattern was observed from the estimated migration rates of the Full

Linear Migration models of both transects. In the Northern transect immigration East across the

river from the Russell County, AL into the Muscogee County, GA population occurs at

approximately twice the rate of the reverse (Table 6). The Southern transect shows a similar

pattern with immigration East from Houston County, AL, into Early County, GA, also at

approximately twice the rate of the reverse (Table 6). This cross-river directional gene flow

presents and interesting pattern, but this directionality is not expected to be caused by a riverine

effect.

The comparison of riverine effect models (Fig. 2) yielded higher support (greater than

1738 BF; Table 7) for the Flood model (Fig. 2 E) over all Oxbow and RBH models (Fig. 2 A–D).

11

The overwhelming support (1738.3 BF better than the second best model) for this model

conflicts with classic thinking about the influence of rivers on gene flow.

The Equal Migration model (Fig. 2 F) showed the least support in this study (5105.2 BF

worse than the best model; Table 7). This model specifically tests the assumption that a river has

a directional effect on gene migration. The asymmetric gene migration noted in the first set of

models may explain why this model faired so poorly. This cross-river asymmetry is accounted

for by the Nearest Neighbor model. In this model the cross-river migration rates are free to be

asymmetrical. The poor relative support for this model versus all oxbow or RBH models (at

least 1701.6 BF worse than any RBH or oxbow model) indicates that even when asymmetry is

allowed, the model is over parameterized by estimating six cross-river migration rates. This

lends support to the idea that rivers have a directional effect on gene migration.

Studies testing riverine barrier hypotheses have yielded mixed results on these effect

(Colwell 2000). One of the best known tests of the GRBH recovered a pattern of downstream

connectivity (Patton et al. 1994). The Flood model and both Oxbow models (Fig. 2 B, C, and E)

each describe processes that yield patterns of higher genetic connectivity downstream along a

river. There is a potential that the current framework in which riverine barrier hypotheses have

been evaluated is too restricted to test other processes. Although these studies have often been

used to cast doubt on the existence of a riverine barrier model (consistent with GRBH or

PDRBH), they have not been used to evaluate alternative models. The approach used in this

study that models process has the power to distinguish between the routes by which genetic

patterns may be created. This power to discriminate between models of process rather than

describe pattern allows for a better reconstruction of past events by directly modeling them.

The Flood model proposed in this research makes some intuitive sense; individuals of

small species often cross rivers through rafting on debris which transports individuals both

across a river and downstream. As debris floats downstream, the probability of the debris

landing on the opposite side increases with downstream distance until it has an approximately

equal chance of being deposited on either bank. Although I have termed this a Flood model

based its proposal by Jacquemyn et al. (2006), a flooding event is not necessary for this process

to operate. A direction of flow is the important part of this model, which may allow future

studies to test the effects of current speed, flow volume, turbidity, etc. on cross-river gene

migration.

12

Methods Used in Evaluating Riverine Effects

Many approaches have been used to investigate the effects of rivers on organisms. Most

of these methods suffer from modeling rivers as impermeable barriers while claiming to test a

riverine hypothesis that defines them as permeable. Tests that model rivers as permeable tend to

lack reasonable alternative models in their analyses. This problem leads to a situation where data

can reject a riverine hypothesis, but is not used to support a riverine model over competing

alternatives beyond panmixia. I have summarized many of the methods previously used to

address riverine effects on populations below and evaluate their effectiveness when compared to

the model testing framework used in this study.

Very few attempts have been made to directly measure riverine effects on dispersal. Klee

et al. (2004) directly assessed the strength of rivers as dispersal barriers through relocation

experiments. This study counted the number of marked individuals found crossing a river after

being relocated to the opposite side. All other evidence for riverine barrier processes comes

from interpreting the patterns left behind.

A few pattern based approaches have been used to test riverine hypotheses without

genetic data, a concept more in line with WRBH (Table 1). Ayres & Clutton-Brock (1992)

interpreted patterns of species similarity along rivers as evidence for riverine barrier effects.

That work used river characterizations, including size, and compared it to a species similarity

index. Jaccard's similarity index (Ludwig & Reynolds 1988) has also been used to make

pairwise comparisons of community similarity at upriver and downriver sites (Gascon et al.

2000). Hayes & Sewlal (2004) used chi-squared tests to determine if range limits within avian

families were influenced by riverine barriers. All these studies are designed to evaluate the

patterns left behind by riverine barrier effects but suffer from a lack of power to identify

causality due to the vast scope of other factors influencing species' range borders (e.g.,

interspecies dependencies).

Another series of pattern based approaches use morphology to look for intraspecies

differences along a river. One of the greatest problems with these morphological studies is that

they cannot include large datasets of homologous characters, such as genetic data, which limits

their power to detect process. Differences in morphological traits and karyotypes have been

13

observed to correspond with rivers in some taxa (Willey & Willey 1967; Lamborot 1991),

although these studies are only able to detect that rivers act as impediments to dispersal, not a

directional effect particular to rivers. Pounds & Jackson (1981) used canonical analysis (Gould

& Johnston 1972) to evaluate phenetic differences in lizard populations within interfluvia

between major rivers. ANOVA and MANOVA analyses have also been used to test if

populations separated by a river differ significantly in morphometric characters (Lamborot &

Eaton 1997; Lamborot et al. 2003; Pauly et al. 2006). These relationships indicate that rivers can

act as breaks in types, but do not directly demonstrate the processes of a river as a directionally

permeable barrier.

Some explicit tests of genetic patterns have been used to show that rivers act as

impediments to gene flow. Bates et al. (2004) compared sequence divergence between same-

bank and cross-river individuals. Slatkin's (Slatkin 1993) has been used to compare pair-wise

population genetic differentiation between same-bank and cross-river populations (Gascon et al.

1998). Additionally, Gascon et al. (1998) used randomization tests to compare distributions of

arc-cosine genetic distances (Cavalli-Sforza & Edwards 1967) generated under panmixia or river

segregation simulations to determine if rivers acted as barriers to gene flow. These studies once

again are useful in confirming a pattern, but do not explicitly model a process.

Clustering analyses have also been used to demonstrate genetic breaks at rivers. Avise et

al. (1979) utilized cluster analysis using estimates of p (Upholt 1977) to determine if natural

genetic subdivisions corresponded to river regions. Principal component analysis of

morphological characters (Lamborot & Eaton 1997) and genetic data (Mylecraine et al. 2004)

has also been used to cluster populations and look for corresponding river barriers. Population

assignment methods (Pritchard et al. 2000; Piry et al. 2004; Guillot 2008) have been used to

determine if genetic populations correspond to river features (Newman & Rissler 2011; Dίaz-

Muñoz 2012). These assignment methods are largely concerned with grouping panmictic units,

but are insufficient for measuring population processes such as gene migration (Palsbøll et al.

2007). They also lack the power to determine process from pattern.

Many studies have used summary statistics to determine if phylogenetic patterns

consistent with a riverine barrier hypothesis exist. Patton et al. (1994) compared cross-bank

sequence divergence at multiple sites along the course of a river to determine if upstream

populations had a higher genetic similarity than downstream populations. Mantel tests (Mantel

14

1967), partial Mantel tests (Smouse et al. 1986), and partial regression analysis (Smouse et al.

1986) have also been used by many authors to evaluate effects of rivers in a population genetics

context (Gascon et al. 1996; Gonzales & Hamrick 2005; Pellegrino et al. 2005; Cabanne et al.

2007; Funk et al. 2007; Lemmon et al. 2007). Gascon et al. (1996) used inbreeding coefficient

analysis to determine hierarchical differences between within-site, same-bank, and total genetic

variances in populations spanning a river. AMOVA has also been used to attribute genetic

variation to rivers by comparing genetic and geographic distances (Davis et al. 2002; Aleixo

2004; Eriksson et al. 2004; Pauly et al. 2006; Cabanne et al. 2007; Zhang et al. 2007; Jalil et al.

2008; Colombi 2010; Dίaz-Muñoz 2012). These methods evaluate whether a pattern consistent

with a riverine barrier hypothesis is observed, but they do not test the underlying process against

other processes that might yield similar patterns of cross-river differentiation.

A number of researchers have used phylogenetic methods to evaluate patterns left by

riverine barrier effects. Unweighted pair group method with arithmetic mean has been used to

construct dendrograms which are then compared to sample localities spanning a river (Gascon et

al. 1998; Mylecraine et al. 2004; Gonzales & Hamrick 2005). Hall & Harvey (2002) used a

morphological parsimony tree to determine if bifurcations corresponded to rivers. Many studies

have generated phylogenetic trees that are then used to evaluate if rivers correspond to

bifurcations (Ellsworth et al. 1994; Chesser 1999; Collins & Dubach 1999; Burbrink et al. 2000;

Donovan et al. 2000; Church et al. 2003; Aleixo 2004; Liu et al. 2005; Kozak et al. 2006; Pauly

et al. 2006; Fu & Zeng 2008, Li et al. 2009). Although these studies often yield robust trees they

lack a means to test the effects of riverine barriers beyond post-hoc attributions of taxonomic

splits to river regions.

Very few studies have used statistically rigorous tests of riverine barriers in a

phylogenetic context. Lougheed et al. (1999) used topology-dependent permutation tail

probability tests (Faith & Cranston 1991) to compare trees constrained to exhibit patterns under a

riverine barrier to trees that were unconstrained. Constraint tree filtering (Miller et al. 2002) was

used by Newman & Rissler (2011) to compare constrained and unconstrained trees. Although

these studies make explicit tests of lineage histories, they assume that a model of a riverine

barrier hypothesis is equivalent to a reciprocal monophyly constraint on each side of a river, i.e.

the river acts as an impermeable barrier. Funk et al. (2007) used parametric bootstrap tests

(Goldman et al. 2000) to evaluate the genetic effects of rivers at different sites along their

15

courses, but still used reciprocal monophyly as the constraint. These methods all use a

phylogenetic context to test a river as a barrier, but do not model the directional effects

consistent with riverine barrier theory. Reciprocal monophyly as a constraint is also a model

inconsistent with the permeable barrier model of a river.

Although many studies have attempted to find patterns consistent with a riverine barrier

hypothesis model, this study is the first to explicitly model the process rather than describe the

pattern. Although the effects of rivers are certain to leave a phylogenetic signal, this pattern can

be obscured due to the dynamic nature of rivers (Patton & da Silva 1998). Problems with

assumptions of the model versus its expected pattern have lead to many incongruous tests of

riverine hypotheses. Tests of riverine effects in this population genetics context are more

appropriate because they attempt to capture the process while it is operating in a system that still

shares genetic exchange, at least in part, across a riverine barrier. Bayesian inference in Migrate-

n (Beerli & Palczewski 2010) offers a statistically rigorous method for discriminating between

models of gene flow across permeable barriers, such as rivers. These methods also allow for

discrimination between competing generalized models of historical gene flow allowing more

confidence in a general conclusion than previous tests that only offer an opportunity to reject a

conclusion without an opportunity to measure model support. These methods are also

advantageous because they explicitly test the process that leads to observed patterns rather than

simply measuring summary statistics of current genetic distributions.

Future Directions

Although this test was only performed on one species, on one river, it provides an

example of the utility of Bayesian model testing in discovering the processes underlying

observed genetic patterns. The Flood model which was favored in this study provides a

promising alternative model of how rivers affect the gene flow of species. This model should be

tested, along with all other hypothesized models of gene flow, across a wide panel of rivers and

taxa. This study provides a framework for reevaluating previous tests of riverine barrier

hypotheses with an emphasis on the contemporary processes that determine genetic patterns

observed around rivers. Some previously collected datasets may be reinterpreted using these

methods to determine the processes generating the previously detected patterns. The powerful

16

analytical approach used in this study should be applied broadly to determine if rivers generate

phylogeographic effects consistent with the proposed Flood model.

17

CHAPTER TWO

CASCADING SPECIATION: EVIDENCE FOR POPULATION

GENETIC EFFECTS OF REPRODUCTIVE CHARACTER

DISPLACEMENT IN PSEUDACRIS FERIARUM

Introduction

Processes of speciation have historically been classified based on geographic pattern.

The majority of speciation events have been categorized as allopatric or peripatric (Futuyma &

Mayer 1980). Under these scenarios speciation occurs through the accumulation of reproductive

barriers in geographically and genetically isolated populations (reviewed in Coyne & Orr 2004).

Sympatric speciation occurs when genetic isolation evolves in the absence of geographic

isolation. This has been well documented in many systems; e.g.., African Cichlidae (Kocher

2004), Rhagoletis pomonella (Filchak et al. 2000); although it is not considered to be a prevalent

mode of speciation (Phillimore et al. 2008). Parapatric speciation, in which peripheral

populations diverge with gene flow, has long been accepted as a theoretical possibility (Fisher

1930; Balkau & Feldman 1973), although conclusive examples in nature are sparse (reviewed in

Coyne & Orr 2004 and Berner et al. 2009). Conclusive examples of parapatric speciation are

difficult to demonstrate due to the difficulties in reconstructing historical rates of gene flow

(Pinho & Hey 2010), but parapatric speciation may represent an underreported, yet frequent

mode of diversification (Gavrilets et al. 1998; Coyne & Orr 2004).

Hybrid zones are often thought of as regions of secondary contact between allopatric

lineages (Barton & Hewitt 1985), although they can be generated by geographic differences in

selective pressures along a gradient—even in the face of constant gene flow (Endler 1977;

Schilthuizen 2000). Simulation studies that model a series of populations distributed linearly

along an environmental gradient have generated evidence that such a regime can lead to

population differentiation and even 'speciation' (Gavrilets 1999; Gavrilets et al. 2000). Models

that yield peripheral population genetic isolation and 'speciation', i.e. a threshold number of

differing loci between individuals, assume that few mutations are required for reproductive

18

isolation. These models emphasize environmental selection gradients but establish a framework

that is useful for thinking about trait and population divergence along sexual selection gradients.

Character displacement (Brown & Wilson 1956) often occurs when two related species

overlap in geographic range. Reproductive character displacement is a mechanism in which

reproductive isolation between related species is strengthened through trait divergence in

sympatric regions (Brown & Wilson 1956; Crozier 1974). A species that exhibits overlap in a

reproductive trait and partial overlap in geographic range with a related species is composed of a

series of populations that exist along a sexual selection gradient analogous to the Gavrilets

(1999) model. Neural network simulation studies have demonstrated that this sexual selection

gradient can result in reproductive character displacement in multiple dimensions in separate

sympatric regions (Pfennig & Ryan 2006; Pfennig & Ryan 2007). Lemmon (2009) provided an

empirical example where a reproductive signal trait diverged in a unique fashion in at least three

regions of sympatry.

Reinforcement is one process that can exert directional selection on a reproductive trait in

sympatry to yield reproductive character displacement (Howard 1993). Under reinforcement,

hybrid offspring between related species exhibit reduced fitness. This fitness reduction leads to

selection on prezygotic isolating mechanisms in the parental species to avoid hybridization. If a

single trait acts as both a signal of interspecies identification and as a signal of intraspecies mate

quality (Rand et al. 1992), then reinforcement can generate a reproductive character

displacement that both reduces interspecies hybridization and alters population gene flow within

the species exhibiting the reproductive character displacement. Alternatively, ecological

speciation may also generate the reproductive character displacement that alters interspecies and

intraspecies gene flow. These processes have been demonstrated in studies of peripatric

speciation followed by secondary contact with the parent lineage (Hoskin et al. 2005) and in

intraspecies primary hybrid zones (Pfennig 2000; Lemmon 2009).

The downstream effects of reproductive character displacement may offer insight into the

process of parapatric speciation. Reproductive character displacement driven by sympatry with a

related species can lead to increased genetic isolation between sympatric and allopatric

populations within species (Higgie & Blows 2008). This process may be repeated independently

in multiple interspecies range overlap zones leading to multiple peripheral populations exhibiting

genetic isolation from both the allopatric range of the species and each other sympatric isolate

19

(Howard 1993). This series of isolating events stemming from range overlap between related

species is termed a 'speciation cascade' (Pfennig & Ryan 2006) due to its potential as a process

leading to rapid speciation events. Ortiz-Barrientos et al. (2009) further implicate reinforcement

as the driving force in this burst of speciation in their 'cascade reinforcement hypothesis'.

Numerous studies have now demonstrated that reproductive character displacement can

lead to divergent trait preferences in sympatric and allopatric populations (e.g.., Pfennig 2000;

Lemmon 2009). Very few studies, however, have evaluated the population genetic

consequences of these trait preferences to evaluate their potential for leading to reproductive

isolation (but see Rice & Pfennig 2010). In this work, I will test the population genetic

predictions of previous female choice experiments (Lemmon 2009) to evaluate support that

range overlap can ultimately lead to reproductive isolation and eventual speciation as in the

speciation cascade model (Pfennig & Ryan 2006).

Study System

Pseudacris feriarum, the upland chorus frog, provides an ideal case study for parapatric

speciation in progress. This species ranges across most of the Eastern United States above the

Fall Line (Fig. 3). The species' range extends into the coastal plain along a few major drainages

of the Atlantic and Gulf coasts. These range extensions bring P. feriarum into sympatry with P.

nigrita, the southern chorus frog, a coastal plains congener (Fig. 3). The acoustic signal of P.

feriarum in these sympatric extensions has undergone reproductive character displacement

(Fouquette 1975; Lemmon 2009) due to reinforcement (Lemmon & Lemmon 2010). This

selective gradient between allopatry and sympatry is repeated along numerous drainages

representing at least three independent events resulting in interspecies hybrid zones between call

types (Lemmon 2009). Each of these hybrid zones have resulted in unique displaced signal

characters (Lemmon 2009).

Lemmon (2009) demonstrated, through female preference experiments, that sympatric P.

feriarum males express a call type that is favored by adjacent allopatric P. feriarum females

when compared to local males along the Apalachicola-Chattahoochee River. Conversely,

allopatric P. feriarum males express a call type that is rejected by adjacent sympatric P. feriarum

females when compared to local males. This preference sets up a case in which a selective

20

gradient is established whereby males that disperse toward the central range of this species into

allopatry are expected to have high fitness while males dispersing into the sympatric peripheral

range are expected to have low fitness. This fitness scenario should lead to increased isolation of

the peripheral populations consistent with models of parapatric speciation (Nosil et al. 2005).

This study used a Bayesian model comparison method to test the predictions of the

female choice experiments in Lemmon (2009). The contemporary gene flow between

populations along the sympatric/allopatric selective gradient was modeled (Fig. 4) to determine

if this system exemplifies a case of increased genetic isolation in sympatric populations

potentially leading to parapatric speciation.

A series of migration models were created to be compared to determine if gene flow

reflects the predictions of Lemmon (2009). The Full Migration model (Fig. 4 A) estimates

independent migration rates in both directions between adjacent populations. This is a parameter

rich model that would explain a complex gene migration pattern. The Equivalent Migration

model (Fig. 4 B) restricts all migration rates, M, to be equivalent, representing no difference in

migration rate caused by reproductive character displacement. The Sympatric Isolation model

(Fig. 4 C) is based on the predictions of Lemmon (2009) that allopatric males dispersing into

sympatry will be less fit than local males. In this model migration from allopatry into sympatry

is not estimated as it is assumed to be a parameter that has little effect on the data and represents

over parameterization.

Methods

Pseudacris feriarum were collected from six populations spanning the region of

reproductive character displacement along the Apalachicola-Chattahoochee River in Florida,

Georgia, and Alabama (Fig. 3). Tissue samples were obtained from twenty individuals per

population using either toe-clipping or dissection (Table 2; Appendix A). Voucher specimens

from each site were preserved in formalin. Populations were sampled in two transects across the

region of reproductive character displacement: an Eastern transect composed of populations

from Russell and Houston Counties, Alabama, and Gulf County, Florida; and a Western transect

composed of populations from Muscogee and Early Counties, Georgia, and Liberty County,

Florida (Fig. 3).

21

Genomic DNA was extracted from 120 P. feriarum individuals using an E.Z.N.A. Gel

Extraction Kit (Omega Bio-Tek). Each individual was genotyped for a total of eleven

microsatellite loci; seven tetramer loci using published primer sets (Lemmon et al. 2011) and

primer sets developed using program μScope v0.0.6 (Appendix B) on previously generated 454

data (Lemmon & Lemmon unpublished; Appendix C), as well as for four dimer loci (Bedwell &

Lemmon unpublished; Ralicki & Lemmon unpublished). Details of primer development are

listed in Appendix C. Multiplex PCR reactions were set up using a Multiplex PCR Kit

(QIAGEN) and amplified using protocols described in Appendix D. Fragment analysis was

conducted on an Applied Biosystems 3730 Genetic Analyzer with Capillary Electrophoresis at

the Florida State University Core Research Facilities. Microsatellite fragment lengths were

measured using GeneMapper v4.1 (Applied Biosystems). Allele calling was done using tandem

v1.08 (Matschiner & Salzburger 2009). These allele calls were used to test for Hardy-Weinberg

equilibrium and linkage disequilibrium (Table 3) using Genepop v4.1 (Rousset 2008). Two loci

(D_C08b and A_E09d) exhibited linkage disequilibrium with the marker D_D12b, although not

with each other (Table 3). Lemmon et al. (2011) found no evidence of linkage disequilibrium

between D_D12b and D_C08b across 85 P. feriarum sampled from Macon County, AL. All

three markers were included in further analyses.

Fragment lengths were used to write a series of microsatellite data files for Migrate-n

v3.2.17 (Beerli 2009). Allele binning was performed by Migrate-n using the "#@M" command

in all data files. Migrate data files were written for the Eastern transect and Western transect to

test the effects of reproductive character displacement on gene migration (Fig. 3). Model

parameters were set in Migrate-n using the user-specified model options. All Migrate-n runs

were conducted using a continuous Brownian-motion model of microsatellite evolution

(Felsenstein 2004). Locality information (Table 4) was used to scale migration rate parameter

estimates using interpopulation geographic distance. A strait line distance matrix (geofile) was

imported into Migrate-n to scale by these interpopulation geographic distances.

A Uniform prior with a range of 0–400 was used for estimating mutation-scaled

migration rates, M, and a Uniform prior with a range of 0–100 was used for estimating mutation-

scaled population size, Θ. Parameter space for each Migrate-n model was searched with a burn-

in of 4,000,000 generations followed by 10,000 steps that were sampled at iterations of every

100 generations to reduce autocorrelation for a grand total of 50,000,000 parameter value

22

iterations explored per long chain. A Metropolis-coupled Markov chain Monte Carlo

(MCMCMC) approach (Geyer 1991) utilizing 4 adaptively heated chains was implemented to

better explore parameter space (Beerli 2006). Parameter values for each model were sampled

using 50 independent long chain samplings. Parameter value distributions were summed over

the fifty independent runs. The natural log marginal likelihood of each model was estimated

using Bézier quadrature (Beerli and Palczewski 2010), an alternative to the harmonic mean

which gives better estimates of the distribution when few heated chains are used. The natural log

marginal likelihood of each model was used to compute natural log Bayes factors (BF: Kass &

Raftery 1995) for model ranking and comparison.

Results

The comparison of gene flow across regions of reproductive character displacement

resulted in patterns that differed between the Eastern and Western transects. In the Eastern

transect the sympatric isolation model (Fig. 4 C) was favored (181 BF better than the second best

model; Table 9). In the Western transect a full migration model (Fig. 4 A) was favored (393 BF

better than the second best model; Table 9).

The individual parameter estimates for all models are listed in Table 10. Migration rate

(M) parameter estimates for migration into allopatry are the lowest of all M parameters for the

Western transect in the best supported model. The Sympatric Isolation model (Fig. 4 C) is

favored in the Eastern transect. In this transect the M parameter estimate for the Full Migration

model (Table 10) is lowest for migration from allopatry into sympatry which is consistent with

the idea that it is a nuisance parameter that is best left out of the model as in the Sympatric

Isolation model.

Discussion

This research evaluated the plausibility of the speciation cascade hypothesis and provides

some of the first genetic evidence supporting this model. In the Eastern transect the Sympatric

Isolation model (Fig. 4 C) was strongly favored over all other models (180.5 BF compared to the

second best model; Table 9). This model matches the expected outcome based on previous work

(Lemmon 2009) and suggests that female preferences may drive interpopulation gene migration.

Analysis of the Western transect, however, favors a Full Migration model (Fig. 4 A) over all

23

other models (392.9 BF compared to the second best model; Table 9), leading to the conclusion

that reproductive character displacement may not be affecting all populations in this region of

sympatry equivalently.

Reproductive character displacement is known to occur on the Western side of the

Apalachicola-Chattahoochee River (Fouquette 1975; Lemmon personal communication). The

female preference tests conducted by Lemmon (2009) used sympatric females from populations

on the Eastern side of the river where the best supported migration model matches the female

choice experiment expectations. Although the acoustic signal of P. feriarum is displaced on the

West side of the Apalachicola-Chattahoochee River, this may represent a case of parallel

evolution. Previous phylogenetic work (Lemmon & Lemmon 2008) provides some evidence of

independent invasions of the Coastal Plain on either side of this river. In this scenario, sympatry

with P. nigrita in the current study may actually represent two independently derived patterns of

reproductive character displacement. It is plausible that female choice experiment conducted

using P. feriarum from west of the Apalachicola-Chattahoochee River may yield differing results

from Lemmon (2009).

One of the problems that must be overcome in a speciation cascade model is the lack of

genetic isolation sympatric and allopatric regions. If a reproductive signal diverges in sympatry

into a less fit intraspecies trait space then isolation of peripheral populations is not likely to occur

due to dispersing allopatric individuals attaining reproductive success within the sympatric

region (e.g. Pfennig 2000). In P. feriarum the divergent signal is actually favored in allopatry

and sympatry over the allopatric signal (Lemmon 2009). This situation, represented by the

Sympatric Isolation model (Fig. 4 C), could generate the isolation necessary for speciation to

occur under a speciation cascade model (Howard 1993). Current evidence for species cascade

models largely consist of descriptions of reproductive character displacement (Howard 1993;

Pfennig & Ryan 2006). This current study provides the first evidence of documented

reproductive character displacement leading to genetic isolation due to female preferences,

which has the potential to generate parapatric speciation under a cascading speciation model.

This research provides a first step in determining the nature of parapatric speciation

within Pseudacris feriarum. This system provides the opportunity to test speciation cascade

theory in multiple independent regions of secondary contact between P. feriarum and P. nigrita

where reproductive character displacement has been observed (Lemmon 2009). Female choice

24

experiments can be used to generate models for gene flow at each of these contact zones to

evaluate a broader pattern of preference effects. This study has laid the groundwork for future

studies that can incorporate multiple contact zones and evaluate them with similar methods

resulting in replicated studies of the effects of reproductive character displacement on

contemporary gene flow.

25

Table 1. A comparison of riverine barrier hypotheses.

Hypothesis Wallace's Riverine

Barrier Hypothesis

Genetic Riverine

Barrier Hypothesis

Primary

Diversification

Riverine Barrier

Hypothesis

Abbreviation WRBH GRBH PDRBH

Process

Rivers impede species

range expansion.

Rivers impede

dispersal by

individuals.

Rivers impede

dispersal by

individuals.

Rivers initiate

allopatric

diversification.

Expected

Pattern

Species ranges are

bordered by rivers.

Communities on

opposite banks become

more dissimilar as a

river becomes wider.

Genetic divergence

between cross-bank

populations within a

species is positively

correlated with river

width.

Genetic divergence

between cross-bank

populations within a

species is positively

correlated with river

width.

Populations do not

show evidence of

recent expansion into

regions around a river

(secondary contact).

26

Table 2. Collection data for Pseudacris feriarum tissues.

Specimen

Number Collection Date Collection Locality Latitude Longitude

ECM0936 3-1-2004 Macon, AL 32.44951 -85.65438

ECM0937 3-1-2004 Macon, AL 32.44951 -85.65438

ECM0938 3-1-2004 Macon, AL 32.44951 -85.65438

ECM0939 3-1-2004 Macon, AL 32.44951 -85.65438

ECM0940 3-1-2004 Macon, AL 32.44951 -85.65438

ECM0941 3-1-2004 Macon, AL 32.44951 -85.65438

ECM0942 3-1-2004 Macon, AL 32.44951 -85.65438

ECM0944 3-1-2004 Macon, AL 32.44951 -85.65438

ECM0945 3-1-2004 Macon, AL 32.44951 -85.65438

ECM0946 3-1-2004 Macon, AL 32.44951 -85.65438

ECM0947 3-1-2004 Macon, AL 32.44951 -85.65438

ECM0948 3-1-2004 Macon, AL 32.44951 -85.65438

ECM0949 3-1-2004 Macon, AL 32.44951 -85.65438

ECM0977 3-3-2004 Macon, AL 32.44951 -85.65438

ECM0978 3-3-2004 Macon, AL 32.44951 -85.65438

ECM0979 3-3-2004 Macon, AL 32.44951 -85.65438

ECM0980 3-3-2004 Macon, AL 32.44951 -85.65438

ECM0981 3-3-2004 Macon, AL 32.44951 -85.65438

ECM0982 3-3-2004 Macon, AL 32.44951 -85.65438

ECM1244 2-3-2005 Macon, AL 32.44411 -85.65521

MJM00603 12-2-2009 Liberty, FL 30.21116 -85.06274

MJM00604 12-2-2009 Liberty, FL 30.21116 -85.06274

MJM00605 12-2-2009 Liberty, FL 30.21116 -85.06274

MJM00606 12-2-2009 Liberty, FL 30.21116 -85.06274

MJM00607 12-2-2009 Liberty, FL 30.21116 -85.06274

MJM00608 12-2-2009 Liberty, FL 30.21116 -85.06274

MJM00609 12-2-2009 Liberty, FL 30.21116 -85.06274

MJM00610 12-2-2009 Liberty, FL 30.21116 -85.06274

MJM00611 12-2-2009 Liberty, FL 30.21116 -85.06274

MJM00612 12-2-2009 Liberty, FL 30.21116 -85.06274

MJM00613 12-2-2009 Liberty, FL 30.21116 -85.06274

MJM00614 12-2-2009 Liberty, FL 30.21116 -85.06274

MJM00633 12-3-2009 Liberty, FL 30.19702 -85.07587

MJM00634 12-3-2009 Liberty, FL 30.19702 -85.07587

MJM00635 12-3-2009 Liberty, FL 30.19702 -85.07587

27

Table 2 (continued).

Specimen

Number

Collection

Date Collection Locality Latitude Longitude

MJM00636 12-3-2009 Liberty, FL 30.19702 -85.07587

MJM00637 12-3-2009 Liberty, FL 30.19702 -85.07587

MJM00638 12-3-2009 Liberty, FL 30.19702 -85.07587

MJM00647 12-3-2009 Liberty, FL 30.19702 -85.07587

MJM00648 12-3-2009 Liberty, FL 30.19702 -85.07587

MJM00685 1-16-2010 Gulf, FL 29.98787 -85.11545

MJM00689 1-16-2010 Gulf, FL 29.98253 -85.10759

MJM00692 1-16-2010 Gulf, FL 29.97876 -85.10328

MJM00693 1-16-2010 Gulf, FL 29.96974 -85.09278

MJM00694 1-16-2010 Gulf, FL 29.97025 -85.08444

MJM00697 1-16-2010 Gulf, FL 29.97023 -85.09402

MJM00745 1-21-2010 Miller, GA 31.1272 -84.69959

MJM00746 1-21-2010 Miller, GA 31.1272 -84.69959

MJM00747 1-21-2010 Miller, GA 31.1272 -84.69959

MJM00748 1-21-2010 Miller, GA 31.1272 -84.69959

MJM00749 1-21-2010 Miller, GA 31.1272 -84.69959

MJM00750 1-21-2010 Miller, GA 31.1272 -84.69959

MJM00751 1-21-2010 Miller, GA 31.13693 -84.70488

MJM00752 1-21-2010 Miller, GA 31.13693 -84.70488

MJM00756 1-21-2010 Miller, GA 31.1689 -84.75686

MJM00757 1-21-2010 Miller, GA 31.1689 -84.75686

MJM00758 1-21-2010 Miller, GA 31.1689 -84.75686

MJM00761 1-21-2010 Miller, GA 31.16536 -84.76899

MJM00776 1-22-2010 Houston, AL 31.06472 -85.04583

MJM00777 1-22-2010 Houston, AL 31.06472 -85.04583

MJM00778 1-22-2010 Houston, AL 31.06472 -85.04583

MJM00779 1-22-2010 Houston, AL 31.06472 -85.04583

MJM00780 1-22-2010 Houston, AL 31.06472 -85.04583

MJM00781 1-22-2010 Houston, AL 31.06472 -85.04583

MJM00782 1-22-2010 Houston, AL 31.06472 -85.04583

MJM00783 1-22-2010 Houston, AL 31.06472 -85.04583

MJM00784 1-22-2010 Houston, AL 31.06472 -85.04583

MJM00785 1-22-2010 Houston, AL 31.06472 -85.04583

MJM00786 1-22-2010 Houston, AL 31.06472 -85.04583

MJM00787 1-22-2010 Houston, AL 31.06472 -85.04583

28

Table 2 (continued).

Specimen

Number

Collection

Date Collection Locality Latitude Longitude

MJM00788 1-22-2010 Houston, AL 31.06472 -85.04583

MJM00789 1-22-2010 Houston, AL 31.06472 -85.04583

MJM00790 1-22-2010 Houston, AL 31.06472 -85.04583

MJM00791 1-22-2010 Houston, AL 31.06472 -85.04583

MJM00792 1-22-2010 Houston, AL 31.06472 -85.04583

MJM00793 1-22-2010 Houston, AL 31.06472 -85.04583

MJM00794 1-22-2010 Houston, AL 31.06472 -85.04583

MJM00795 1-22-2010 Houston, AL 31.06472 -85.04583

MJM00981 1-1-2011 Gulf, FL 30.034 -85.17026

MJM00983 1-1-2011 Gulf, FL 30.02963 -85.16807

MJM00985 1-1-2011 Gulf, FL 30.00083 -85.14157

MJM00986 1-1-2011 Gulf, FL 30.00083 -85.14157

MJM00987 1-1-2011 Gulf, FL 30.00083 -85.14157

MJM00988 1-1-2011 Gulf, FL 29.99604 -85.12428

MJM00989 1-1-2011 Gulf, FL 29.99604 -85.12428

MJM00994 1-1-2011 Gulf, FL 29.9969 -85.12465

MJM00995 1-1-2011 Gulf, FL 29.9969 -85.12465

MJM00998 1-1-2011 Gulf, FL 30.00093 -85.13302

MJM01000 1-1-2011 Gulf, FL 30.00087 -85.13306

MJM01001 1-1-2011 Gulf, FL 30.00087 -85.13306

MJM01002 1-1-2011 Gulf, FL 30.00087 -85.13306

MJM01006 1-1-2011 Gulf, FL 29.96949 -85.09112

MJM01017 1-31-2011 Miller, GA 31.17422 -84.74257

MJM01018 1-31-2011 Miller, GA 31.17422 -84.74257

MJM01019 1-31-2011 Miller, GA 31.17422 -84.74257

MJM01020 1-31-2011 Miller, GA 31.17422 -84.74257

MJM01021 1-31-2011 Miller, GA 31.17422 -84.74257

MJM01022 1-31-2011 Miller, GA 31.17422 -84.74257

MJM01023 1-31-2011 Miller, GA 31.17422 -84.74257

MJM01024 1-31-2011 Miller, GA 31.17422 -84.74257

MJM01028 3-1-2011 Early, GA 31.36524 -84.94428

MJM01029 3-1-2011 Early, GA 31.36524 -84.94428

MJM01030 3-1-2011 Early, GA 31.36524 -84.94428

MJM01031 3-1-2011 Early, GA 31.36524 -84.94428

MJM01032 3-1-2011 Early, GA 31.36524 -84.94428

29

Table 2 (continued).

Specimen

Number

Collection

Date Collection Locality Latitude Longitude

MJM01033 3-1-2011 Early, GA 31.36524 -84.94428

MJM01034 3-1-2011 Early, GA 31.36524 -84.94428

MJM01035 3-1-2011 Early, GA 31.36524 -84.94428

MJM01036 3-1-2011 Early, GA 31.36524 -84.94428

MJM01037 3-1-2011 Early, GA 31.36524 -84.94428

MJM01038 3-1-2011 Early, GA 31.36524 -84.94428

MJM01039 3-1-2011 Early, GA 31.36524 -84.94428

MJM01040 3-1-2011 Early, GA 31.36524 -84.94428

MJM01041 3-1-2011 Early, GA 31.36524 -84.94428

MJM01042 3-1-2011 Early, GA 31.36524 -84.94428

MJM01043 3-1-2011 Early, GA 31.36524 -84.94428

MJM01044 3-1-2011 Early, GA 31.36524 -84.94428

MJM01045 3-1-2011 Early, GA 31.36524 -84.94428

MJM01046 3-1-2011 Early, GA 31.36524 -84.94428

MJM01047 3-1-2011 Early, GA 31.36524 -84.94428

MJM01067 3-6-2011 Muscogee, GA 32.5527 -84.87294

MJM01068 3-6-2011 Muscogee, GA 32.5527 -84.87294

MJM01069 3-6-2011 Muscogee, GA 32.5527 -84.87294

MJM01070 3-6-2011 Muscogee, GA 32.5527 -84.87294

MJM01071 3-6-2011 Muscogee, GA 32.5527 -84.87294

MJM01072 3-6-2011 Muscogee, GA 32.5527 -84.87294

MJM01073 3-6-2011 Muscogee, GA 32.5527 -84.87294

MJM01074 3-6-2011 Muscogee, GA 32.5527 -84.87294

MJM01075 3-6-2011 Muscogee, GA 32.5527 -84.87294

MJM01076 3-6-2011 Muscogee, GA 32.5527 -84.87294

MJM01077 3-6-2011 Muscogee, GA 32.5527 -84.87294

MJM01078 3-6-2011 Muscogee, GA 32.5527 -84.87294

MJM01079 3-6-2011 Muscogee, GA 32.5527 -84.87294

MJM01080 3-6-2011 Muscogee, GA 32.5527 -84.87294

MJM01081 3-6-2011 Muscogee, GA 32.5527 -84.87294

MJM01082 3-6-2011 Muscogee, GA 32.5527 -84.87294

MJM01083 3-6-2011 Muscogee, GA 32.5527 -84.87294

MJM01084 3-6-2011 Muscogee, GA 32.5527 -84.87294

MJM01085 3-6-2011 Muscogee, GA 32.5527 -84.87294

MJM01086 3-6-2011 Muscogee, GA 32.5527 -84.87294

30

Table 2 (continued).

Specimen

Number

Collection

Date Collection Locality Latitude Longitude

MJM01090 3-9-2011 Russell, AL 32.45618 -85.19571

MJM01091 3-9-2011 Russell, AL 32.45618 -85.19571

MJM01092 3-9-2011 Russell, AL 32.45618 -85.19571

MJM01093 3-9-2011 Russell, AL 32.45618 -85.19571

MJM01094 3-9-2011 Russell, AL 32.45618 -85.19571

MJM01095 3-9-2011 Russell, AL 32.45618 -85.19571

MJM01096 3-9-2011 Russell, AL 32.45602 -85.19704

MJM01097 3-9-2011 Russell, AL 32.45602 -85.19704

MJM01098 3-9-2011 Russell, AL 32.45602 -85.19704

MJM01099 3-9-2011 Russell, AL 32.45602 -85.19704

MJM01100 3-9-2011 Russell, AL 32.45602 -85.19704

MJM01101 3-9-2011 Russell, AL 32.45602 -85.19704

MJM01102 3-9-2011 Russell, AL 32.45602 -85.19704

MJM01103 3-9-2011 Russell, AL 32.45602 -85.19704

MJM01104 3-10-2011 Russell, AL 32.45602 -85.19704

MJM01105 3-10-2011 Russell, AL 32.45602 -85.19704

MJM01106 3-10-2011 Russell, AL 32.45602 -85.19704

MJM01107 3-10-2011 Russell, AL 32.45602 -85.19704

MJM01108 3-10-2011 Russell, AL 32.45602 -85.19704

MJM01109 3-10-2011 Russell, AL 32.45602 -85.19704

31

Table 3. Results of Hardy-Weinberg exact tests (p-values) and linkage disequilibrium

probability tests (p-values) conducted in Genepop v4.1. Loci that significantly deviated from

Hardy-Weinberg equilibrium across all sampled populations and loci that exhibited significant

linkage disequilibrium are highlighted in blue.

D6 0.0000

D_D12b 0.0000 0.857

D_C08b 0.0413 0.691 0.004

D_D10 0.208 0.987 0.840 0.972

P_fer_57550 0.866 0.752 0.959 0.274 0.491

P_fer_46999 0.0000 0.131 0.996 0.998 0.213 0.987

P_fer_101070 0.315 0.986 0.859 0.997 0.389 0.181 0.937

A_C08d 0.0000 0.766 0.999 0.991 0.982 0.641 0.995 1.000

A_E09d 0.0014 0.291 0.002 0.500 0.935 0.989 0.995 1.000 1.000

A_B04d 0.631 0.066 0.747 0.724 1.000 0.588 0.716 0.959 0.912 0.555

P_fer_DBNXL 0.266 0.221 0.812 0.957 0.144 0.686 0.994 0.993 0.800 1.000 0.213

HW

P V

alu

e

D6

D_D

12b

D_C

08b

D_D

10

P_fe

r_57550

P_fe

r_46999

P_fe

r_101070

A_C

08d

A_E

09d

A_B

04d

Linkage Disequilibrium

Table 4. Distance matrix of all inter-population distances used for creating geofiles. All

distances are recorded in km.

Russell 42.23

Muscogee 74.50 32.16

Houston 163.48 154.94 165.80

Early 137.23 123.28 131.85 34.70

Miller 165.21 148.48 153.35 31.37 28.59

Gulf 279.45 175.80 287.10 121.42 155.32 137.67

Liberty 255.17 250.74 261.90 96.24 130.13 112.95 25.20

Macon Russell Muscogee Houston Early Miller Gulf

32

Table 5. Bezier approximated marginal likelihoods (ln mL) of three models testing the effect of

a river on gene migration. Model rankings are indicated, with the best supported model

highlighted in each transect. Bayes factors are computed for each model compared to the best

model in each transect. The Northern transect consists of populations collected in Macon Co.,

AL, Russell Co., AL, and Muscogee Co., GA; the Southern transect consists of populations

collected in Houston Co., AL, Early Co., GA, and Miller Co., GA

Northern Transect Southern Transect

Model ln mL Rank BF ln mL Rank BF

Full Linear Migration -13814.8 1 -15481.3 1

Equivalent Linear Migration -24156.9 3 10342.2 -29073.0 3 13591.7

Linear River Effect -17142.5 2 3327.7 -19145.3 2 3663.9

Table 6. Parameter estimates for models of cross-river gene flow. The mode of the probability

distribution for each parameter is reported. The model with the highest support (Table 5) is

highlighted in blue.

Parameter

Full

Linear

Migration

Equivalent

Linear

Migration

Linear

River

Effect

Nort

her

n T

ran

sect

ΘMacon 98.100 13.966 96.233

ΘRussell 4.233 7.833 3.566

ΘMuscogee 2.833 4.033 2.366

MMacon → Russell 328.133 392.667 334.530

MRussell → Macon 71.867 392.667 28.933

MRussell → Muscogee 111.867 392.667 392.667

MMuscogee → Russell 39.333 392.667 392.667

Sou

ther

n T

ran

sect

ΘMiller 8.7 4.366 8.966

ΘEarly 1.966 2.766 1.833

ΘHouston 11.566 7.833 6.7

MMiller → Early 203.6 383.867 244.4

MEarly → Miller 78.8 383.867 54

MEarly → Houston 63.867 383.867 392.667

MHouston → Early 135.867 383.867 392.667

33

Table 7. Bezier approximated marginal likelihoods (ln mL) of three models testing riverine

effects on gene migration along a river. Model rankings are indicated, with the best supported

model highlighted.

Model ln mL Rank BF

Nearest Neighbor -30174.1 6 5105.2

Equal Migration -42612.9 7 17544.1

Complete Oxbow -26907.0 3 1838.1

Incomplete Oxbow -28472.5 5 3403.7

Complete RBH -26807.2 2 1738.3

Incomplete RBH -27776.9 4 2708.0

Flood -25068.9 1

34

Table 8. Parameter estimates for models of riverine gene flow. The mode of the probability

distribution for each parameter is reported. The model with the highest support (Table 7) is

highlighted in blue.

Parameter

Nearest

Neighbor

Equal

Migration

Complete

Oxbow

Incomplete

Oxbow

Complete

RBH

Incomplete

RBH Flood

ΘRussell 17.433 6.033 10.660 20.766 17.433 13.900 14.833

ΘHouston 4.566 4.633 7.433 3.366 8.233 5.300 5.233

ΘGulf 5.500 5.900 3.900 5.500 8.233 10.433 7.100

ΘMuscogee 2.166 2.966 4.833 7.033 1.833 2.100 5.366

ΘEarly 1.966 1.966 4.033 1.766 4.100 1.966 2.633

ΘLiberty 6.700 5.433 3.966 6.033 13.900 15.700 10.100

MRussell → Houston 8.933 0.133 66.800 14.800 386.800 0.133 0.133

MHouston → Russell 17.733 48.933 74.267 99.600 24.933 59.867 45.200

MHouston → Gulf 387.867 62.533 258.267 386.000 392.133 392.133 392.133

MGulf → Houston 379.067 387.333 390.267 190.800 48.133 94.533 214.267

MMuscogee → Early 37.200 20.933 0.133 13.200 158.800 31.867 0.133

MEarly → Muscogee 11.333 6.267 81.200 95.600 23.333 4.933 50.000

MEarly → Liberty 389.467 380.933 310.000 378.267 392.133 389.200 62.267

MLiberty → Early 158.000 382.533 387.867 231.333 18.800 91.333 96.133

MRussell → Muscogee 11.600 392.667 49.467 11.867

MMuscogee → Russell 0.133 392.667 49.733 0.133

MHouston → Early 0.133 392.667 7.330 16.400

MEarly → Houston 0.133 392.667 7.867 380.400

MGulf → Liberty 9.733 392.667 381.733 45.200

MLiberty → Gulf 0.133 392.667 13.467 76.933

MRussell → Early 227.333

MMuscogee → Houston 34.800

MHouston → Liberty 392.133

MEarly → Gulf 383.600

35

Table 9. Bezier approximated marginal likelihood (ln mL) results for tests of gene flow across a

region of reproductive character displacement. The best supported model is highlighted in each

model set.

Western Transect Eastern Transect

Model ln mL Rank BF ln mL Rank BF

Full Migration -10668.9 1 -11044.7 2 180.5

Equivalent Migration -12056.4 3 1387.5 -12487.3 3 1623.1

Sympatric Isolation -11061.8 2 392.9 -10864.2 1

Table 10. Parameter estimates for models of gene flow across a region of reproductive character

displacement. The mode of the probability distribution for each parameter is reported. The

model with the highest support (Table 9) is highlighted in blue.

Parameter

Full

Migration

Equivalent

Migration

Sympatric

Isolation

Wes

tern

Tra

nse

ct

ΘRussell 13.167 12.500 10.700

ΘHouston 7.100 8.767 13.766

ΘGulf 9.967 8.233 9.166

MRussell → Houston 384.667 393.200

MHouston → Russell 56.133 393.200 248.933

MHouston → Gulf 392.400 393.200 385.733

MGulf → Houston 383.867 393.200 317.733

East

ern

Tra

nse

ct

ΘMuscogee 6.966 5.500 5.766

ΘEarly 3.567 5.766 4.566

ΘLiberty 22.900 18.366 28.966

MMuscogee → Early 15.867 392.667

MEarly → Muscogee 51.600 392.667 131.330

MEarly → Liberty 392.400 392.667 75.067

MLiberty → Early 80.400 392.667 211.067

36

Table 11. Microsatellite primer sequences developed from 454 reads (M-13 tag not included). R

ever

se P

rim

er S

equ

ence

AA

AA

CC

AG

AA

AC

CC

CA

TA

GT

AA

AT

C

AA

TG

TT

GT

TC

TA

AA

TA

GA

CG

TG

AT

GA

A

AG

TC

AC

TT

TC

AT

GT

TT

CT

CT

TG

TT

CA

T

TA

TT

AA

GT

GA

CC

AC

TG

AT

CC

TA

CC

TC

TT

AG

TC

TA

GT

TT

CC

CT

CA

GG

TT

AA

AT

AT

C

GC

TG

AC

CA

GT

AT

AC

TT

TG

TT

TA

AT

GC

TT

AT

CT

GA

GT

AG

TA

AA

TG

GA

AA

TG

TT

AA

GG

AA

GT

GG

TA

TT

TC

AA

AA

AC

AG

AC

CT

AT

C

GT

TA

GG

CT

CA

TT

CT

CA

CT

TA

CA

AT

TT

C

TG

TT

AG

TA

AC

TA

TT

TA

TT

AT

GT

GC

GA

CA

GT

AT

AA

TT

TT

CA

GC

AC

TC

TT

CT

AT

GA

CA

TC

AA

AG

TA

AT

AA

AA

AT

TT

AG

CT

CA

TG

GA

TT

CT

AA

AA

TT

TG

TG

TT

AT

GG

AT

AT

AG

AG

CA

TC

GT

AC

AT

TT

CT

TC

TA

AT

CC

TT

TT

TG

GA

A

TT

AA

TG

TT

GT

AT

CA

CA

CG

TT

CA

TA

AT

AA

AG

TT

TT

GT

CT

TC

TT

TT

TG

AA

CT

CT

GC

AG

TT

TA

CT

CA

TT

TT

GT

TC

CT

TG

GT

CT

TA

AA

TA

CT

AA

AG

GT

GG

GG

AA

TA

TC

AA

GT

AA

CT

TG

AT

TT

CA

GC

AT

CT

TG

TT

TC

T

AG

AC

TT

TG

CA

GA

TT

CA

AA

AG

TC

AT

T

GT

GT

AA

GG

TT

AA

AA

AC

TA

TA

AG

GT

GA

CA

AA

Forw

ard

Pri

mer

Seq

uen

ce

TA

TT

GC

TC

TT

CA

TT

AA

AA

AT

TT

CT

TT

CT

C

CT

TG

TC

AT

GT

AA

TT

CT

CA

TC

TT

CT

TA

AT

G

AT

TA

CT

TC

TG

AC

AA

AT

AA

CG

TA

AT

GC

TA

AG

GA

TC

TC

AG

AA

CC

AG

AT

CA

TA

CT

CA

TT

AC

CA

TA

AT

TA

TT

AG

AC

CT

GA

AC

AG

AT

TC

TC

AC

CA

GT

GT

AA

AG

CC

TT

TT

AT

TT

GA

A

AA

TT

TA

TT

AA

GT

GC

TC

AA

TG

CC

AG

T

AA

AA

TT

GA

CA

CA

AA

AA

TG

TC

TA

TA

GG

G

AA

CC

AG

AT

AT

TT

TG

CA

TT

GC

TA

TC

AC

TG

GA

AG

GC

GG

TG

TC

TG

GA

AT

AG

TA

AA

GC

AG

CA

AG

TA

TT

GA

CC

TA

AG

TA

CT

GA

GT

AT

TT

AG

TG

CA

AT

TT

CC

AA

AT

TT

GT

GG

TT

TA

GA

CA

GC

AT

GA

AA

TT

AT

TC

GT

AA

AG

AA

CT

AC

TG

GT

AG

AG

AC

AC

CC

AA

AC

AC

AT

TT

TC

TT

AG

CA

TT

GT

CA

TC

TT

CA

TG

GT

AT

AT

TT

TG

TG

AT

GT

T

GA

GA

AT

TA

CT

GA

CT

CG

CT

AA

GG

GT

A

TA

AT

CA

GT

GT

TC

AT

CT

CC

TG

AT

AA

TT

G

CA

AA

CG

AC

AA

AA

AT

AA

AA

TA

TG

TC

AC

TA

AT

AC

TG

GA

AA

GA

AG

AA

TG

CA

AA

GA

T

GA

CT

AT

GA

CA

GA

AT

AT

TA

AC

CT

AA

TG

GA

AA

Moti

f

AA

AT

AC

AT

AA

AG

AG

AT

AC

GG

AA

AT

AA

GT

AG

AT

AC

GG

AC

AT

AG

AT

AA

AT

AA

AT

AA

AT

AA

AT

AA

AT

AA

GT

AG

GG

AA

AC

AC

CG

AA

AT

Locu

s N

am

e

P_fe

r_64

64

P_fe

r_92

51

P_fe

r_26

421

P_fe

r_32

589

P_fe

r_34

263

P_fe

r_44

482

P_fe

r_46

724

P_fe

r_46

999

P_fe

r_52

033

P_fe

r_52

281

P_fe

r_57

550

P_fe

r_64

901

P_fe

r_69

274

P_fe

r_70

269

P_fe

r_73

827

P_fe

r_83

490

P_fe

r_10

0489

P_fe

r_10

1070

P_fe

r_10

5253

P_fe

r_11

3902

P_fe

r_11

6964

37

Table 12. Results of microsatellite primer amplification in a panel of Pseudacris feriarum

selected from across the range of the species. A 'Y' indicates that a sample amplified with at

least one detectable allele, while an 'N' indicates that no alleles amplified in that sample. The

total observed alleles (NA) within the 9 individuals genotyped is recorded. The three loci

accepted for use in this thesis are highlighted in blue.

Locus Name 454 R

epea

t N

um

ber

EC

M388 M

aco

n, A

L

EC

M7041 H

eard

, G

A

EC

M389 M

aco

n, A

L

EC

M7

042 H

eard

, G

A

EC

M7143 P

rin

ce E

dw

ard

, V

A

MJM

00603 L

iber

ty, F

L

MJM

00782 H

ou

ston

, A

L

MJM

00981 G

ulf

, F

L

MJM

01017 M

ille

r, G

A

NA

P_fer_6464 4 N N

P_fer_9251 4 N N

P_fer_26421 4 N N

P_fer_32589 14 Y N

P_fer_34263 5 N N

P_fer_44482 5 Y Y Y Y Y Y Y Y Y 3

P_fer_46724 4 Y Y Y Y Y Y Y Y Y 1

P_fer_46999 10 Y Y Y Y Y Y N Y Y 9

P_fer_52033 4 N N

P_fer_52281 4 N N

P_fer_57550 4 Y Y N Y Y Y N Y Y 4

P_fer_64901 5 N N

P_fer_69274 4 N N

P_fer_70269 4 Y Y Y N Y Y Y N N 2

P_fer_73827 4 N N

P_fer_83490 4 Y Y N N N N N N N

P_fer_100489 4 Y N N N N N N N N

P_fer_101070 4 Y Y Y N Y Y Y Y Y 11

P_fer_105253 5 N N

P_fer_113902 4 N Y N N N N N N N

P_fer_116964 6 N N

38

Table 13. Primer sequences for sixteen microsatellite loci in Pseudacris feriarum. Reverse

primer sequences are highlighted in blue.

Locus Name X-mer Primer Sequence

D6 Tetramer CTGCTGTGATATTTTTGTG

GGTGTCGTGAGCTAAGTGTT

D_D12b Tetramer TATAACATGTAACTGGGCTAACA

AGGAGAAGAGCCATTTCCTG

D_C08b Tetramer CTTACACAGCTCCATAGAATATGACA

ACAAACCTACAGGAGCTGATAGAAT

D_D10 Tetramer CTCTACATACATTTACCTTCTACCTTC

GCTGTCTACTGAATTTATACTGTAAGG

P_fer_57550 Tetramer GAATAGTAAAGCAGCAAGTATTGACCTA

ATAATTTTCAGCACTCTTCTATGACATC

P_fer_46999 Tetramer AAAATTGACACAAAAATGTCTATAGGG

AAGTGGTATTTCAAAAACAGACCTATC

P_fer_101070 Tetramer TAATCAGTGTTCATCTCCTGATAATTG

TAAATACTAAAGGTGGGGAATATCAAGT

A_C08d Dimer GGTAATAAACCAATCTTAAATTAGTAACACAA

CTCATGGATACTACAACTCTCGACAGTAAC

A_E09d Dimer ATTACATGTAGCTTTTAAGGATATATTGTTGC

TAAATGCTTTATCTCTGTTACATCTGTATAGG

P_fer_G79VC Dimer ACCAGGACTAGCTATATAAGAACATAGACTTT

CCGATACCTCTCTTGCATGTGT

A_B04d Dimer AGTGACTTCCAATATCCTATAGTCCTCTT

ATGGACTGTAGCAGTTACCATGTGTAGT

P_fer_DJURT Dimer AGAAGTATTCTTGCACCCTAGGAAT

GACACATTATTTCATTCTGTAACTAAAAAGTA

P_fer_A7NK3_2 Dimer GAATATTTTAGTTTTCCCTTTTCAATAAATC

TGTGATAAATACGGTATATGTATATGTGTCTG

P_fer_A3PXV Dimer TAATAAACAAAACACCTTTGTAGGTGAAAATA

ATCTGATTTAGTAGTCAGTTGTGCAGTAAAAG

P_fer_CIMKE Dimer CCGTGTTGTGTATTAACTATAAGGAAAAA

ACTAGTATCACCACACAAAACAAACAA

P_fer_DBNXL Dimer ATCAATAGAGGAGGTGGGAGCTG

ACATATAGAGCAATAACTGAGCTGTGAAT

39

Table 14. Multiplex composition used to genotype individuals for sixteen loci. Five loci

amplified poorly, inconsistently, or exhibited primer nonspecificity; these loci are highlighted in

blue and were not included in data analysis. The total observed alleles (NA) within the 160

individuals genotyped for this project, the observed shortest and longest fragment lengths, and

the fluorophore used are recorded. Recorded fragment lengths include the primers and use the

fluorophores indicated.

Mu

ltip

lex

Pri

mer

Set

Na

me

Dev

elop

er

NA

Sh

ort

est

Fra

gm

ent

Len

gth

L

on

ges

t F

ragm

ent

Len

gth

Flu

oro

ph

ore

1

D6 Lemmon et al. (2011) 25 100 372 NED

D_D12b Lemmon et al. (2011) 28 268 532 FAM

D_C08b Lemmon et al. (2011) 16 227 445 HEX

2

D_D10 Lemmon et al. (2011) 6 152 398 NED

P_fer_57550 Michelsohn (Appendix C) 7 248 267 FAM

P_fer_46999 Michelsohn (Appendix C) 23 184 270 PET

P_fer_101070 Michelsohn (Appendix C) 10 191 237 VIC

3

A_C08d Bedwell & Lemmon (unpublished) 19 300 300 NED

A_E09d Bedwell & Lemmon (unpublished) 27 216 216 FAM

P_fer_G79VC Ralicki & Lemmon (unpublished) N/A 135 168 PET

A_B04d Bedwell & Lemmon (unpublished) 24 138 138 VIC

4

P_fer_DJURT Ralicki & Lemmon (unpublished) N/A 190 202 NED

P_fer_A7NK3_2 Ralicki & Lemmon (unpublished) N/A 237 260 FAM

P_fer_A3PXV Ralicki & Lemmon (unpublished) N/A 107 165 PET

P_fer_CIMKE Ralicki & Lemmon (unpublished) N/A 155 180 VIC

N/A P_fer_DBNXL Ralicki & Lemmon (unpublished) 18 152 186 NED

40

Figure 1. Illustrations of gene flow models used to evaluate the effect of a river on gene

migration compared to overland migration. The vertical blue bar represents the Apalachicola-

Chattahoochee River; solid black arrows represent independent M parameters; dashed red arrows

represent M parameters that are restricted to be equivalent estimates within a model (1 M

parameter for all equivalent rates); circles represent independent Θ parameters. These models

were tested in both a Northern (1- Macon, AL; 2- Russell, AL; 3- Muscogee, GA) and a

Southern transect (1- Miller, GA; 2- Early, GA; 3- Houston, AL). The total number of

parameters simulated in each model is noted in parentheses after the model name.

Figure 2 (next page). Illustrations of gene flow models used to evaluate the effects of rivers on

population genetics. The vertical blue arrow represents the Apalachicola-Chattahoochee River

and its direction of flow; solid black arrows represent independent M parameters; dashed red

arrows represent M parameters that are restricted to be equivalent estimates within a model (1 M

parameter for all equivalent rates); circles represent Θ parameters with subscripts denoting the

counties where populations were sampled (R- Russell, AL; H- Houston, AL; G- Gulf, FL; M-

Muscogee, GA; E- Early, GA; L- Liberty, FL). The total number of parameters simulated in

each model is noted in parentheses after the model name.

41

A. Complete RBH (16) B. Incomplete RBH (18) C. Complete Oxbow ( 16)

@ @ @ @ @ @

ll ll ll ll ll ll

@ @ @ @ @ @

ll ll ll ll ll ll

@ @ @ @ @ @

D. Incomplete Oxbow (18) E. Flood (18) F. Equal Migration (15)

@ @ @ @ ｀ ｾＺ＠ ］ｾ ｀＠

ll ll ﾣ Ｑ ｾ＠ ll ll

@ @ ｀ ｾＺ＠ Ｚｾ Ｚ ｀＠

ll ll ll ll ll ll

@ @ @ @ ｀ ｾＺ＠ ］ｾ ｀＠

G. Nearest Neighbor (20)

42

Figure 3. Map depicting sampling localities used in this study. The Apalachicola-

Chattahoochee River is indicated by the blue line; the dashed white line represents the

approximate region of the reproductive character displacement cline. County names are

indicated on the sampling localities in the inset map.

43

Figure 4. Illustrations of gene flow models used to evaluate the effect of reproductive character

displacement (RCD) on gene migration. The vertical green bar represents the region of RCD;

solid black arrows represent independent M parameters; dashed red arrows represent M

parameters that are restricted to be equivalent estimates within a model (1 M parameter for all

equivalent rates); circles represent independent Θ parameters for sympatric populations of P.

feriarum; squares represent independent Θ parameters for allopatric populations of P. feriarum.

These models were tested in both an Eastern (1- Liberty, FL; 2- Early, GA; 3- Muscogee, GA)

and a Western transect (1- Gulf, FL; 2- Houston, AL; 3- Russell, AL). The total number of

parameters simulated in each model is noted in parentheses after the model name.

44

APPENDIX A

SPECIMEN COLLECTION INFORMATION

All specimens used in this work (Table 2) were collected under the following permits

issued to E. M. Lemmon or M. J. Michelsohn:

Florida: LSSC-09-0362

Georgia: 22378

Alabama: 2010000037468680

All live animals were collected, tissue sampled, and preserved following protocols

established in the Animal Care and Use Committee (ACUC) protocol 0905 for the E. M.

Lemmon lab at Florida State University.

45

46

47

APPENDIX B

PROGRAM μSCOPE

The program μScope is a python script that was developed in this research to search for

candidate microsatellite loci using 454 sequence reads. This program detects and classifies

sequence reads that contain microsatellites. These sequence reads are then output for use in

primer design. Individual microsatellites are classified by motif length as x-mers where x

represents the motif length (e.g.. a tetramer is represented as 4-mer).

The program requires five user selected input values to run, as well as at least one data

file in FASTA format. These input values include:

threshold_value - This input value is used to determine the number of times an individual

x-mer sequence must be repeated to be detected as a microsatellite. The program

uses the following inequality to determine if a repeated sequence is considered to

occur more often than expected by chance.

In this inequality x is the number of nucleotides in an x-mer motif, and j is an

integer value specifying a number of repeats. The integer value of j is increased

sequentially until the inequality is true. The smallest integer value of j for which

the inequality is true is set as the significance_value for all x-mers of length x.

All x-mer sequences repeated equal to or more times than the significance_value

for that x-mer are considered to be microsatellites.

max_motif_length - This input value is used to determine the largest x-mer length that

will be detected as a microsatellite by the program. In μScope v0.0.6 this input is

limited to integer values in the range 1–6.

file_extension - This input value allows the user to specify which data files will be used in

the search. All data files in the current working directory with the extension

specified will be read into the program. These data files are required to be in

FASTA format but can have any extension specified.

min_intemotif_length - This input value sets the nucleotide distance between any detected

microsatellites within a sequence read that is considered to be interacting.

48

minimum_repeats - This input value determines the minimum number of x-mer repeats

within a microsatellite region required for the containing sequence read to be

included within an output file of this program.

The program classifies detected microsatellites based on their proximity to other detected

microsatellites (specified using the input min_intemotif_length). The microsatellite regions

within a read are classified as one of four types:

Perfect - Microsatellites that contain a single motif x-mer separated from all other

microsatellites by a distance of min_intermotif_length or greater.

Imperfect Monomer - Microsatellite regions that contain a single motif x-mer but also

include non-repetitive insertions.

Imperfect Multimer - Microsatellite regions that contain multiple motifs of the same x-

mer. These may also include non-repetitive insertions.

Compound - Microsatellite regions that contain multiple motifs, some of which do not

have the same x-mer. These regions may also include non-repetitive insertions.

After searching and sorting all microsatellite regions from all the imported data, the sequence

reads containing the microsatellite regions that match the user inputs are output in a series of

FASTA formatted files in the current working directory. An output file is generated for each

microsatellite type in each x-mer.

Run times vary depending on the number and length of imported sequence reads, number

of microsatellites contained within each sequence read, as well as user input values. While the

program is running a rough estimate search completeness is displayed. When a search is

completed, summary values are displayed that list the numbers of sequence reads containing

microsatellite regions of each type and x-mer.

49

APPENDIX C

454-GENERATED PRIMER DEVELOPMENT

The program μScope (Appendix B) was used to search through 130,222 previously

generated 454 sequence reads (Lemmon & Lemmon unpublished). The search used a

threshold_value of 0.5 which requires dimers to be repeated twice (e.g.. ACACAC) and all other

x-mers to be repeated only once before they are counted as microsatellites. The search used a

max_motif_length of 6 and a min_intemotif_length of 50. Fifty-five reads containing at least four

consecutive repeats (minimum_repeats = 4) of a perfect tetramer microsatellite were imported

into BioEdit v7.0.9.0 (Hall 1999). The NCBI BLAST v2.0 (Altschul et al. 1997) plug-in was

used to identify overlapping reads. All overlapping reads were assembled into aligned contigs

by using the NCBI BLAST results. Thirty-three non-overlapping reads and aligned contigs were

imported into Primer3 v0.4.0 (Rozen & Skaletsky 2000) for primer design. Designed primers

were checked by eye to avoid placement in regions that contained homopolymers due to the risk

of 454 sequencing error (Margulies et al. 2005).

M-13 tagged primer sets (Schuelke 2000) were designed for twenty-one loci (Table 11).

These primer sets were first tested using gDNA samples from two P. feriarum, ECM388 and

ECM7041 (Table 12). Genomic DNA was extracted from tissue samples using an E.Z.N.A. Gel

Extraction Kit (Omega Bio-tek). Microsatellites were amplified in 10μL reactions of 1μL 10×

NH4 Reaction Buffer (Biolase), 0.8μL 10mM dNTPs, 0.6μL 50mM MgCl2, 6.5μL H2O, 0.05μL

Biolase Taq, 0.1μL 2.5mM M-13 tagged forward primer, 0.4μL 10mM reverse primer, 0.1μL

2.5mM FAM tagged M-13 tag, and 0.5μL 15–20ng/μL template gDNA on a DNA Engine Tetrad

2 thermal cycler using the following PCR program: (1) 94°C for 3 min, (2) 94°C denaturation

for 30 s, (3) 54°C annealing for 30 s, (4) 72°C extension for 45 s, repeat steps 2–4 for 25 cycles,

(5) 94°C denaturation for 30 s, (6) 53°C annealing for 30 s, (7) 72°C extension for 45 s, repeat

steps 5–7 for 8 cycles, (8) 72°C extension for 10 min. Fragment analysis of PCR products was

conducted on an Applied Biosystems 3730 Genetic Analyzer with Capillary Electrophoresis

using Genescan 500 ROX size standard. Loci were genotyped using GeneMapper v4.1 (Applied

Biosystems). Primers pairs that exhibited clear amplification in at least one individual were

tested across a wider panel of seven more P. feriarum individuals (Table 12).

50

Six primer sets amplified well across the test panel (Table 12). Three primers showed

low levels of allelic diversity (NA < 4 across 9 individuals) and were excluded from further use

due to budget constraints. Three primers (P_fer_46999, P_fer_57550, and P_fer_101070) were

considered viable for use in this thesis.

51

APPENDIX D

MULTIPLEX PCR PRIMER SETS & PROTOCOLS

Multiplex sets were designed to keep dimer and tetramer microsatellite primers (Table

13) together. Fragment length ranges from previous M-13 tagged genotyping (Appendix C;

Lemmon unpublished; Bedwell & Lemmon unpublished; Ralicki & Lemmon unpublished) were

used to group primers into minimally overlapping groups to minimize pull-up peaks in analysis.

Multiplex set 1 used 3 fluorophore chemistry (NED, FAM, HEX) while sets 2, 3, and 4 used 4

fluorophore chemistry (NED, FAM, PET, VIC). Primer P_fer_DBNXL was tagged with NED

but not included within a multiplex.

PCR reaction mixtures for all multiplexes were similar except for the content of the

primer mix. 10× multiplex primer mix was made for each multiplex with fluorescent tagged

forward primers and reverse primers all mixed at a concentration of 100μM. 10μL reactions of

5μL 2× Qiagen Multimix, 3μL dsH20, 1μL 15–20ng/μL template gDNA, and 1μL 10× multiplex

primer mix. This protocol yielded a 10μM concentration of each primer in the final reaction

mixtures. The primer set for P_fer_DBNXL was diluted as though it were in a multiplex;

otherwise reactions were set up the same for this locus.

The following PCR protocol was used for multiplex 1: (1) 95°C for 15 min, (2) 94°C

denaturation for 30 s, (3) 54°C annealing for 90 s, (4) 72°C extension for 90 s, repeat steps 2–4

for 35 cycles, (5) 72°C extension for 7 min. The PCR protocol used for multiplexes 2 & 4 were

the same: (1) 95°C for 15 min, (2) 94°C denaturation for 30 s, (3) 54°C annealing for 90 s, (4)

72°C extension for 60 s, repeat steps 2–4 for 35 cycles, (5) 60°C extension for 30 min. The

following PCR protocol was used for multiplex 3: (1) 95°C for 15 min, (2) 94°C denaturation

for 30 s, (3) 48°C annealing for 90 s, (4) 72°C extension for 90 s, repeat steps 2–4 for 35 cycles,

(5) 72°C extension for 7 min. The PCR protocol used for M-13 tagged primer tests in Appendix

C was used for amplification in P_fer_DBNXL.

Fragment analysis of PCR products was conducted on an Applied Biosystems 3730

Genetic Analyzer with Capillary Electrophoresis using Genescan 500 ROX size standard for

multiplex 1 and P_fer_DBNXL and Genescan 500 LIZ size standard for multiplexes 2, 3, & 4.

Loci were genotyped using GeneMapper v4.1 (Applied Biosystems). The primer set

P_fer_G79VC as well as all primers in multiplex 4 yielded poor quality genotype data. Most of

52

these five loci exhibited multiple peaks likely resulting from primer nonspecificity. Additionally

many loci amplified inconsistently across individuals when PCR and fragment analysis was

duplicated. Some of these primers were designed using a separate panel of test individuals

(Lemmon & Ralicki unpublished) which exhibited better results. All of these loci showed

geographic regional inconsistencies in amplification indicating that these loci may be useful for

fine scale studies in some areas. All resulting genotypes from these five loci were excluded from

analysis in this thesis due to their poor quality across the sampled range (Fig. 3). Fragment

length data for the eleven included loci was imported into tandem v1.08 (Matschiner &

Salzburger 2009) for automated binning to determine observed alleles for each locus (NA, Table

14).

53

REFERENCES

Aleixo, A. 2004. Historical diversification of a terra-firme forest bird superspecies: a

phylogeographic perspective on the role of different hypotheses of Amazonian

diversification. Evolution 58: 1303–1317.

Altschul, S. F., T. L. Madden, A. A. Schäffer, J. Zhang, Z. Zhang, W. Miller, & D. J. Lipman.

1997. Gapped BLAST and PSI-BLAST: a new generation of protein database search

programs. Nucleic Acids Research 25: 3389–3402.

Avise, J. C. 1992. Molecular population structure and the biogeographic history of a regional

fauna: a case history with lessons for conservation biology. Oikos 63: 62–76.

Avise, J. C., C. Giblin-Davidson, J. Laerm, J. C. Patton, & R. A. Lansman. 1979. Mitochondrial

DNA clones and matriarchal phylogeny within and among geographic populations of the

pocket gopher, Geomys pinetis. Proceedings of the National Academy of Sciences 76:

6694–6698.

Ayres, J. M. & T. H. Clutton-Brock. 1992. River boundaries and species range size in

Amazonian Primates. The American Naturalist 140 (3): 531–537.

Balkau, B. J. & M. M. Feldman. 1973. Selection for migration modification. Genetics 74:

171–174.

Barton, N. H., & G. M. Hewitt. 1985. Analysis of hybrid zones. Annual Review of Ecology

and Systematics 16: 113–148.

Bates, H. W. 1863. The naturalist on the river Amazons; a record of adventures, habits of

animals, sketches of Brazilian and Indian life, and aspects of nature under the equator,

during eleven years of travel. 2 volumes. John Murray: London.

Bates, J. M., J. Haffer, & E. Grismer. 2004. Avian mitochondrial DNA sequence divergence

across a headwater stream of the Rio Tapajós, a major Amazonian river. Journal of

Ornithology 145: 199–205.

Beerli, P. 2006. Comparison of Bayesian and maximum likelihood inference of population

genetic parameters. Bioinformatics 22: 341–345.

Beerli, P. 2009. How to use migrate or why are markov chain monte carlo programs difficult to

use? In Bertorelle, G., M. W. Bruford, H. C. Hauffe, A. Rizzoli, & C. Vernesi eds.

Population genetics for animal conservation (pp. 42–79). Cambridge University Press:

Cambridge.

Beerli, P. & M. Palczewski. 2010. Unified framework to evaluate panmixia and migration

direction among multiple sampling locations. Genetics 185: 313–326.

54

Berner, D., A. Grandchamp, & A. P. Hendry. 2009. Variable progress toward ecological

speciation in parapatry: stickleback across eight lake-stream transitions. Evolution 63:

1740–1753.

Blaney, R. M. 1971. An annotated check list and biogeographic analysis of the insular

herpetofauna of the apalachicola region, Florida. Herpetologica 27: 406–430.

Brown, W. L., & E. O. Wilson. 1956. Character displacement. Systematic Zoology 5: 49–64.

Burbrink, F. T., R. Lawson, & J. B. Slowinski. 2000. Mitochondrial DNA phylogeography of

the polytypic North American rat snake (Elaphe obsoleta): A critique of the subspecies

concept. Evolution 54: 2107–2118.

Cabanne, G. S., F. R. Santos, & C. Y. Miyaki. 2007. Phylogeography of Xiphorhynchus fuscus

(Passerirormes, Dendrocolaptidae): vicariance and recent demographic expansion in

southern Atlantic forest. Biological Journal of the Linnean Society 91: 73–84.

Carr, A. & C. J. Goin. 1959. Guide to the reptiles, amphibians, and fresh-water fishes of

Florida. University of Florida Press: Gainesville.

Chesser, R. T. 1999. Molecular systematic of the Rhinocryptid genus Pteroptochos. The

Condor 101: 439–446.

Church, S. A., J. M. Kraus, J. C. Mitchell, D. R. Church, D. R. Taylor. 2003. Evidence for

multiple Pleistocene refugia in the postglacial expansion of the eastern tiger salamander,

Ambystoma tigrinum tigrinum. Evolution 57: 372–383.

Collins, A. C. & J. M. Dubach. 1999. Biogeographic and ecological forces responsible for

speciation in Ateles. International Journal of Primatology 21: 421–444.

Colombi, V. H., S. R. Lopes, & V. Fagundes. 2010. Testing the Rio Doce as a riverine barrier

in shaping the Atlantic rainforest population divergence in the rodent Akdon cursor.

Genetics and Molecular Biology 33: 785–789.

Colwell, R. K. 2000. A barrier runs through it... or maybe just a river. Proceedings of the

National Academy of Science 97: 13470–13472.

Coyne, J. A., & H. A. Orr. 2004. Speciation. Sinauer Associates: Sunderland, MA.

Crenshaw, J. W., & W. F. Blair. 1959. Relationships in the Pseudacris nigrita complex in

Southwestern Georgia. Copeia 1959: 215–222.

Crozier, R. H. 1974. Niche shape and genetic aspects of character displacement. American

Zoologist 14: 1151–1157.

55

Davis, L. M., T. C. Glenn, D. C. Strickland, L. J. Guillette Jr., R. M. Elsey, W. E. Rhodes, H. C.

Dessauer, & R. H. Sawyer. 2002. Microsatellite DNA analyses support an East-West

phylogeographic split of American alligator populations. Journal of Experimental

Zoology 294: 352–372.

Dίaz-Muñoz, S. L. 2012. Role of recent and old riverine barriers in fine-scale population

genetic structure of Geoffroy's tamarin (Saguinus geoffroyi) in the Panama Canal

watershed. Ecology and Evolution 2: 298–309.

Donovan, M. F., R. D. Semlitsch, & E. J. Routman. 2000. Biogeography of the Southeastern

United States: a comparison of salamander phylogeographic studies. Evolution 54:

1449–1456.

Ellsworth, D. L., R. L. Honeycutt, N. J. Silvy, J. W. Bickham, & W. D. Klimstra. 1994.

Historical biogeography and contemporary patterns of mitochondrial DNA variation in

white-tailed deer from the Southeastern United States. Evolution 48: 122–136.

Endler, J. A. 1977. Geographic variation, speciation, and clines. Princeton University Press:

Princeton, NJ.

Engle, V. D. & J. K. Summers. 2000. Biogeography of benthic macroinvertebrates in estuaries

along the Gulf of Mexico and western Atlantic coasts. Hydrobiologia 436: 17–33.

Eriksson, J., G. Hohmann, C. Boesch, & L. Vigilant. 2004. Rivers influence the population

genetic structure of bonobos (Pan paniscus). Molecular Ecology 13: 3425–3435.

Faith, D. P., & P. S. Cranston. 1991. Could a cladogram this short have arisen by chance

alone?: on permutation tests for cladistic structure. Cladistics 7: 1–28.

Felsenstein, J. 2004. Inferring Phylogenies. Sinauer Associates: Sunderland, Massachusetts.

Filchak, K. E., J. B. Roethele, & J. L. Feder. 2000. Natural selection and sympatric divergence

in the apple maggot Rhagoletis pomonella. Nature 407: 739–742.

Fisher, R. A. 1930. The genetical theory of natural selection. Clarendon Press: Oxford,

England.

Fouquette, M. J. 1975. Speciation in chorus frogs. I. reproductive character displacement in the

Pseudacris nigrita complex. Systematic Zoology 24: 16–23.

Fu, J. & X. Zeng. 2008. How many species are in the genus Batrachuperus? a

phylogeographical analysis of the stream salamanders (family Hynobiidae) from

southwestern China. Molecular Ecology 17: 1469–1488.

56

Funk, W. C., J. P. Caldwell, C. E. Peden, J. M. Padial, I. De la Riva, & D. Cannatella. 2007.

Tests of biogeographic hypotheses for diversification in the Amazonian forest frog,

Physalaemus petersi. Molecular Phylogenetics and Evolution 44: 825–37.

Futuyma, D. J. & G. C. Mayer. 1980. Non-allopatric speciation in animals. Systematic Zoology

29: 254–271.

Gascon, C., J. R. Malcom, J. L. Patton, M. N. F. da Silva, J. P. Bogart, S. C. Lougheed, C. A.

Peres, S. Neckel, & P. T. Boag. 2000. Riverine barriers and the geographic distribution

of Amazonian species. Proceedings of the National Academy of Sciences 97: 13672–13677.

Gascon, C., S. C. Lougheed, & J. P. Bogart. 1996. Genetic and morphological variation in

Vanzolinius discodactylus: a test of the river hypothesis of speciation. Biotropica 28:

376–387.

Gascon, C., S. C. Lougheed, & J. P. Bogart. 1998. Patterns of genetic population differentiation

in four species of Amazonian frogs: a test of the riverine barrier hypothesis. Biotropica

30: 104–119.

Gavrilets, S. 1999. A dynamical theory of speciation on holey adaptive landscapes. The

American Naturalist 154: 1–22

Gavrilets, S., H. Li, & M. D. Vose. 1998. Rapid parapatric speciation on holey adaptive

landscapes. Proceedings of the Royal Society B 265: 1483–1489.

Gavrilets, S., H. Li, & M. D. Vose. 2000. Patterns of parapatric speciation. Evolution 54:

1126–1134.

Geyer, C. J. 1991. Markov chain Monte Carlo maximum likelihood. In Keramidas (ed.)

Computing Science and Statistics: Proceedings of the 23rd Symposium on the Interface.

Interface Foundation, Fairfax Station, pp. 156–163.

Goldman, N., Anderson, J. P., & Rodrigo, A. G. 2000. Likelihood-based tests of topologies in

phylogenetics. Systematic Biology 49: 652–670.

Gonzales, E. & J. L. Hamrick. 2005. Distribution of genetic diversity among disjunct

populations of the rare forest understory herb, Trillium reliquum. Heredity 95: 306–314.

Guillot, G. 2008. Inference of structure in subdivided populations at low levels of genetic

differentiation–the correlated allele frequencies model revisited. Bioinformatics 24:

2222–2228.

Haffer, J. 1969. Speciation in Amazonian forest birds. Science 165: 131–166

57

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.

Hayes, F. E. & J. N. Sewlal. 2004. The Amazon river as a dispersal barrier to passerine birds:

effects of river width, habitat and taxonomy. Journal of Biogeography 31: 1809–1818.

Hershkovitz, P. 1977. Living new world monkeys (Platyrrhini); with an introduction to

Primates Vol. 1. University of Chicago Press: Chicago.

Hershkovitz, P. 1983. Two species of night monkeys, genus Aotus (Cebidae, Platyrrhini): a

preliminary report on Aotus taxonomy. American Journal of Primatology 4: 209–243.

Higgie, M., & M. W. Blows. 2008. The evolution of reproductive character displacement

conflicts with how sexual selection operates within a species. Evolution 62: 1192–1203.

Hoskin, C. J., M. Higgie, K. R. McDonald, & C. Moritz. 2005. Reinforcement drives rapid

allopatric speciation. Nature 437: 1353–1356.

Howard, D. J. 1993. Reinforcement: origin, dynamics, and fate of an evolutionary hypothesis.

In Harrison, R. G. (Ed.) Hybrid Zones and the Evolutionary Process. Oxford University

Press: New York, NY.

Jacquemyn, H., O. Honnay, K. Van Looy, & P. Breyne. 2006. Spatiotemporal structure of

genetic variation of a spreading plant metapopulation on dynamic riverbanks along the

Meuse River. Heredity 96: 471–8.

Jalil, M. F., J. Cable, J. Sinyor, I. Lackman-Ancrenaz, M. Ancrenaz, M. W. Bruford, & B.

Goossens. 2008. Riverine effects on mitochondrial structure of Bornean ogang-utans

(Pongo pygmaeus) at two spatial scales. Molecular Ecology 17: 2898–2909.

Jensen, J. B., C. D. Camp, W. Gibbons, M. J. Elliott. 2008. Amphibians and reptiles of Georgia.

University of Georgia Press: Athens.

Kass, R. E. & A. E. Raftery. 1995. Bayes factors. Journal of the American Statistical

Association 90: 773–795.

Klee, R. V., A. C. Mahoney, C. C. Christopher, & G. W. Barrett. 2004. Riverine peninsulas: an

experimental approach to homing in white-footed mice (Peromyscus leucopus).

American Midland Naturalist 151: 408–413.

Knowles, L. L. 2001. Did the Pleistocene glaciations promote divergence? tests of explicit

refugial models in montane grasshoppers. Molecular Ecology 10: 691–701.

Knowles, L. L., & W. P. Maddison. 2002. Statistical phylogeography. Molecular Ecology 11:

2623–2635.

58

Kocher, T. D. 2004. Adaptive evolution and explosive speciation: the cichlid fish model.

Nature Reviews Genetics 5: 288–298.

Kozak, K. H., R. A. Blaine, & A. Larson. 2006. Gene lineages and eastern North American

paleodrainage basins: phylogeography and speciation in salamanders of the Eurycea

bislineata species complex. Molecular Ecology 15: 191–207.

Kramer, D. C. 1973. Movements of western chorus frogs Pseudacris triseriata triseriata tagged

with Co60. Journal of Herpetology 7: 231–235.

Kramer, D. C. 1974. Home range of the western chorus frog Pseudacris triseriata triseriata.

Journal of Herpetology 8: 245–246.

Lamborot, M. 1991. Karyotypic variation among populations of Liolaemus monticola

(Tropiduridae) separated by riverine barriers in the Andean range. Copeia 1991: 1044–1059.

Lamborot, M. & L. Eaton. 1997. The Maipo river as a biogeographical barrier to Liolaemus

monticola (Tropiduridae) in the mountain ranges of central Chile. Journal of Zoological

Systematics and Evolutionary Research 35: 105–111.

Lamborot, M., L. Eaton, & B. A. Carrasco. 2003. The Aconcagua river as another barrier to

Liolaemus monticola (Sauria: Iguanidae) chromosomal races of central Chile. Revista

Chilena de Historia Natural 76: 23–34.

Lemmon, E. M. 2009. Diversification of conspecific signals in sympatry: geographic overlap

drives multi-dimensional reproductive character displacement in frogs. Evolution 63:

1155–1170.

Lemmon, A. R., & E. M. Lemmon. 2008. A likelihood framework for estimating

phylogeographic history on a continuous landscape. Systematic Biology 57: 544–561.

Lemmon, E. M., & A. R. Lemmon. 2010. Reinforcement in chorus frogs: lifetime fitness

estimates including intrinsic natural selection and sexual selection against hybrids.

Evolution 64: 1748–1761.

Lemmon, E. M., A. R. Lemmon, & D. C. Cannatella. 2007. Geological and climatic forces

driving speciation in the continentally distributed trilling chorus frogs (Pseudacris).

Evolution 61: 2086–2103Lemmon, E. M., M. Murphy, & T. E. Juenger. 2011.

Identification and characterization of nuclear microsatellite loci for multiple species of

chorus frogs (Pseudacris) for population genetic analyses. Conservation Genetics

Resources 3: 233–237.

Lemmon, E. M., M. Murphy, & T. E. Juenger. 2011. Identification and characterization of

nuclear microsatellite loci for multiple species of chorus frogs (Pseudacris) for

population genetic analyses. Conservation Genetics Resources 3: 233–237.

59

Li, R., W. Chen, L. Tu, & J. Fu. 2009. Rivers as barriers for high elevation amphibians: a

phylogeographic analysis of the alpine stream frog of the Hengduan mountains. Journal

of Zoology 277: 309–316.

Liu, F. R., P. E. Moler, & M. M. Miyamoto. 2005. Phylogeography of the salamander genus

Pseudobranchus in the southeastern United States. Molecular Phylogenetics and

Evolution 39: 149–159.

Lougheed, S. C., C. Gascon, D. A. Jones, J. P. Bogart, & P. T. Boag. 1999. Ridges and rivers: a

test of competing hypotheses of Amazonian diversification using a dart-poison frog

(Epipedobates femoralis). Proceedings of the Royal Society B 266: 1829–1835.

Ludwig, J. A., & J. F. Reynolds. 1988. Statistical ecology: a primer on methods and

computing. John Wiley & Sons: USA.

Mantel, N. 1967. The detection of disease clustering and a generalized regression approach.

Cancer Research 27: 209–220.

Margulies, M., M. Egholm, W. E. Altman, S. Attiya, J. S. Bader, L. A. Bemben, J. Berka, M. S.

Braverman, Y. Chen, Z. Chen, S. B. Dewell, L. Du, J. M. Fierro, X. V. Gomes, B. C.

Godwin, W. He, S. Helgesen, C. H. Ho, G. P. Irzyk, S. C. Jando, M. L. I. Alenquer, T. P.

Jarvie, K. B. Jirage, J. Kim, J. R. Knight, J. R. Lanza, J. H. Leamon, S. M. Lefkowitz, M.

Lei, J. Li, K. L. Lohman, H. Lu, V. B. Makhijani, K. E. McDade, M. P. McKenna, E. W.

Myers, E. Nickerson, J. R. Nobile, R. Plant, B. P. Puc, M. T. Ronan, G. T. Roth, G. J.

Sarkis, J. F. Simons, J. W. Simpson, M. Srinivasan, K. R. Tartaro, A. Tomasz, K. A.

Vogt, G. A. Volkmer, S. H. Wang, Y. Wang, M. P. Weiner, P. Yu, R. F. Begley, & J. M.

Rothberg. 2005. Genome sequencing in microfabricated high-density picolitre reactors.

Nature 437: 376–380.

Marshall, S. D., W. R. Hoeh, & M. A. Deyrup. 2000. Biogeography and conservation biology

of Florida's Geolycosa wolf spiders: threatened spiders in endangered ecosystems.

Journal of Insect Conservation 4: 11–21.

Matschiner, M., & W. Salzburger. 2009. TANDEM: integrating automated allele binning into

genetics and genomics workflows. Bioinformatics 25: 1982–1983.

Mayr, E. 1942. Systematics and the origin of species from the viewpoint of a zoologist.

Columbia University Press: New York.

Means, D. B. 1977. Aspects of the significance to terrestrial vertebrates of the apalachicola

river drainage basin, Florida. Florida Marine Research Publications 26: 37–67.

Miller, R. E., T. R. Buckley, & P. S. Manos. 2002. An examination of the monophyly of

morning glory taxa using Bayesian phylogenetic inference. Systematic Biology 51:

740–753.

60

Mount, R. H. 1975. The reptiles and amphibians of Alabama. University of Alabama Press:

Tuscaloosa.

Mylecraine, K. A., J. E. Kuser, P. E. Smouse, & G. L. Zimmermann. 2004. Geographic

allozyme variation in Atlantic white-cedar, Chamaecyparis thyoides (Cupressaceae).

Canadian Journal of Forest Research 34: 2443–2454.

Neill, W. T. 1957. Historical biogeography of present-day Florida. Bulletin of the Florida State

Museum 2: 175–220

Newman, C. E. & L. J. Rissler. 2011. Phylogeographic analyses of the southern leopard frog:

the impact of geography and climate on the distribution of genetic lineages vs.

subspecies. Molecular Ecology 20: 5295–5312.

Nosil, P., T. H. Vines, & D. J. Funk. 2005. Reproductive isolation caused by natural selection

against immigrants from divergent habitats. Evolution 59: 705–719.

Ortiz-Barrientos, D., A. Grealy, & P. Nosil. 2009. The genetics and ecology of reinforcement:

implications for the evolution of prezygotic isolation in sympatry and beyond. The Year

in Evolutionary Biology 2009. Annals of the New York Academy of Sciences 1168 156–182.

Palsbøll, P. J., M. Bérubé, & F. W. Allendorf. 2006. Identification of management units using

population genetic data. Trends in Ecology and Evolution 22: 11–16.

Patton, J. L., M. N. F. da Silva, & J. R. Malcolm. 1994. Gene genealogy and differentiation

among arboreal spiny rats (Rodentia: Echimyidae) of the Amazon Basin: a test of the

riverine barrier hypothesis. Evolution 48: 1314–1323.

Pauly, G. B., O. Piskurek, & H. B. Shaffer. 2007. Phylogeographic concordance in the

southeastern United States: the flatwoods salamander, Ambystoma cingulatum, as a test

case. Molecular Ecology 16: 415–29.

Pellegrino, K. C. M., M. T. Rodrigues, A. N. Waite, M. Morando, Y. Y. Yassuda, & J. Sites Jr.

2005. Phylogeography and species limits in the Gymnodactylus darwinii complex

(Gekkonidae, Squamata): genetic structure coincides with river systems in the Brazilian

Atlantic forest. Biological Journal of the Linnean Society 85: 13–26.

Pfennig, K. S. 2000. Female spadefoot toads compromise on mate quality to ensure conspecific

matings. Behavioral Ecology 11: 220–227.

Pfennig, K. S., & M. J. Ryan. 2006. Reproductive character displacement generates

reproductive isolation among conspecific populations: an artificial neural network study.

Proceedings of the Royal Society B 273: 1361–1368.

61

Pfennig, K. S., & M. J. Ryan. 2007. Character displacement and the evolution of mate choice:

an artificial neural network approach. Philosophical Transactions of the Royal Society B

362: 411–419.

Phillimore, A. B., C. D. L. Orme, G. H. Thomas, T. M. Blackburn, P. M. Bennett, K. J. Gaston,

& I. P. F. Owens. 2008. Sympatric speciation in birds is rare: insights from range data

and simulations. The American Naturalist 171: 646–657.

Pinho, C., & J. Hey. 2010. Divergence with gene flow: models and data. Annual Review of

Ecology, Evolution, and Systematics 2010 41: 215–230.

Piry, S., A. Alapetite, J. M. Cornuet, D. Paetkau, L. Baudouin, & A. Estoup. 2004.

GENECLASS2: a software for genetic assignment and first-generation migrant

detection. Journal of Heredity 95: 536–539.

Pounds, J. A. & J. F. Jackson. 1981. Riverine barriers to gene flow and the differentiation of

fence lizard populations. Evolution 35: 516–528.

Rand, S. A., M. J. Ryan, & W. Wilczynski. 1992. Signal redundancy and receiver

permissiveness in acoustic mate recognition by the Túngara frog, Physalaemus

pustulosus. American Zoologist 32: 81–90.

Randazzo, A. F. & D. S. Jones eds. 1997. The geology of Florida. University Press of Florida:

Gainesville.

Rice, A. M., & D. W. Pfennig. 2010. Does character displacement initiate speciation? evidence

of reduced gene flow between populations experiencing divergent selection. Journal of

Evolutionary Biology 23: 854–865.

Rousset, F. 2008. Genepop'007: a complete reimplementation of the Genepop software for

Windows and Linux. Molecular Ecology Resources 8: 103–106.

Rozen, S. & H. J. Skaletsky. 2000. Primer3 on the WWW for general users and for biologist

programmers. In Krawetz S, Misener S (eds.) Bioinformatics Methods and Protocols:

Methods in Molecular Biology. Humana Press, Totowa, NJ, pp 365–386.

Schilthuizen, M. 2000. Ecotone: speciation-prone. Trends in Ecology & Evolution 15: 130–131.

Schuelke, M. 2000. An economic method for the fluorescent labeling of PCR fragments.

Nature Biotechnology 18: 233–234.

Slatkin, M. 1993. Isolation by distance in equilibrium and non-equilibrium populations.

Evolution 47: 264–279.

62

Smouse, P. E., J. C. Long, & R. R. Sokal. 1986. Multiple regression and correlation extensions

of the Mantel test of matrix correspondence. Systematic Zoology 35: 627–632.

Soltis, D. E., A. B. Morris, J. S. McLachlan, P. S. Manos, & P. S. Soltis. 2006. Comparative

phylogeography of unglaciated eastern North America. Molecular Ecology 15: 4261–4293.

Swenson, N. G. & D. J. Howard. 2005. Clustering of contact zones, hybrid zones, and

phylogeographic breaks in North America. The American Naturalist 166: 581–591.

Upholt, W. B. 1977. Estimation of DNA sequence divergence from comparison of restriction

endonuclease digests. Nucleic Acids Research 4: 1257–1265.

Van Looy, K., H. Jacquemyn, P. Breyne, & O. Honnay. 2009. Effects of flood events on the

genetic structure of riparian populations of the grassland plant Origanum vulgare.

Biological Conservation 142: 870–8.

Wallace, A. R. 1852. On the monkeys of the Amazon. Proceedings of the Zoological Society

of London 20: 107–10.

Wallace, A. R. 1876. The geographical distribution of animals; with a study of the relations of

living and extinct faunas as elucidation the past changes of the Earth's surface. 2

volumes. Harper & Brothers: New York.

Walker, M. J., A. K. Stockman, P. E. Marek, & J. E. Bond. 2009. Pleistocene glacial refugia

across the Appalachian Mountains and coastal plain in the millipede genus Narceus:

evidence from population genetic, phylogeographic, and paleoclimatic data. BMC

Evolutionary Biology 9: 25.

Willey, R. B. & R. L. Willey. 1967. Barriers to gene flow in natural populations of

grasshoppers I. the Black Canyon of the Gunnison River and Arphia conspera. Psyche

74: 42–57.

Zhang, Z., X. Zheng, & S. Ge. 2007. Population genetic structure of Vitex negundo

(Verbenaceae) in the Three-gorge area of the Yangtze river: the riverine barrier to seed

dispersal in plants. Biochemical Systematics and Ecology 35: 506–516.

63

BIOGRAPHICAL SKETCH

Moses James Michelsohn

Moses was born on August 16, 1985 in Roswell, New Mexico. He grew up watching

birds and often dragged his family to Bitter Lake National Wildlife Refuge to see the geese and

cranes. In 2003 Moses left his desert home to attend Eckerd College in St. Petersburg, Florida,

where he first learned about cladistics from Peter Meylan. He graduated with a Bachelor of

Science degree in Biology in 2007, then spent a few years working as a herpetologist across the

United States and in Costa Rica. He joined Emily Lemmon's lab at Florida State University in

2009 where he began working on Pseudacris. There he became increasingly fascinated by

coalescent theory and Bayesian methods with the encouragement of Peter Beerli. Following

graduation, he hopes to continue working in phylogenetics and one day return to a more arid

climate.