Carabid Beetle Ecology As a Function of Disturbance In

Home , Abaris (beetle), Agonum punctiforme, Amara aenea, Amara cupreolata, Amara familiaris, Amara pallipes

CARABID BEETLE ECOLOGY AS A FUNCTION OF DISTURBANCE IN

TEMPERATE AND TROPICAL FORESTS

KATHRYN NICOLE RILEY

A Dissertation Submitted to the Graduate Faculty of

WAKE FOREST UNIVERSITY GRADUATE SCHOOL OF ARTS AND SCIENCES

in Partial Fulfillment of the Requirements

for the Degree of

DOCTOR OF PHILOSOPHY

Biology

December 2016

Winston-Salem, North Carolina

Approved By:

Robert A. Browne, Ph.D., Advisor

Terry L. Erwin, Ph.D., Chair

T. Michael Anderson, Ph.D.

William E. Conner, Ph.D.

Miles R. Silman, Ph.D.

ACKNOWLEDGMENTS

First, I want to thank my Advisor Dr. Robert Browne. His guidance, insight and support over the past eight years, not only made this dissertation a success, but profoundly influenced my life. I will be forever grateful for the opportunities in London,

Australia and New Zealand which were truly amazing, life enriching experiences. I have been fortunate to work with you all these years and your kindness, patience and advice is sincerely appreciated. I am forever in your debt.

Thank you to my committee members, Drs. Miles Silman, T. Michael Anderson,

William Conner and Terry Erwin for all of their advice, suggestions and insight during my time as a graduate student, as well as their feedback for this dissertation. I am grateful to Drs. Silman and Anderson from whom I have learned much and who have driven me to be a better ecologist. Thank you, Dr. C for your positive influences over the years and instilling my interest in insects. Terry, to whom I owe so much, I am thankful for the invitations to countless meetings and incorporation into the Carabidologists circle. Thank you for the introduction to Tiputini, one of my favorite places in the world. I cannot thank you enough for the opportunities, insight, knowledge and experiences.

Without the support and encouragement of my parents, this degree would not have been possible. Their unconditional support, faith and confidence in my abilities have always encouraged me to pursue my ambitions. To my brother, Nick, for always being there and believing in both me and my goals. Thank you to my family for all the years of stress-filled phone calls, seven hour drives and so much more. To my fiancé, Kevin, thank you for supporting me in all ways possible over the past three and a half years. For

ii understanding the many long nights spent in front of the computer. I am forever thankful to have you by my side and look forward to seeing what life has in store for us.

I want to mention the professors at Francis Marion University whose influences have led me to this point in my career. Thank you, Travis Knowles for your enthusiasm and influence over the many years since I was an undergraduate in your Organismal

Biology class. I appreciate your continued interest and support in my career. To Drs.

Tamatha Barbeau and Gregory Pryor, thanks for your support with my undergraduate research project which sparked my interest in field work, diversity and disturbance.

I’d like to thank my friends and colleagues, whose support was vital to my success as a graduate student. First thanks to wonderful friends in Maryland, Katie Miller

Johnson and Meagan Converse Deal for their continued friendships despite the long distance. Katie, thank you for all long emails, texts and the many, many visits – I am truly lucky to have you as a friend. Thanks to my forever lab-mate, Sarah Maveety, for all the support throughout graduate school as both a lab-mate and friend. Couldn’t have done it without you. To my many friends in the WFU Biology Department (both former and present), thank you for keeping me sane, listening to worries/concerns and, most importantly, all the good times over our many years here.

I owe thanks to the undergraduate students who have contributed to these projects over the years. There were many field assistants both in the Amazon and in North

Carolina who made the collection of the thousands of carabid beetles possible. I am grateful to the Department of Biology of Wake Forest University for the many years of support, the experiences and the positive environment it provided.

iii

TABLE OF CONTENTS ACKNOWLEDGEMENTS…………………………………………………………...... ii

LIST OF TABLES………………………………………………………….……………vii

LIST OF FIGURES……………………………………………………………………...xii

LIST OF ABBREVIATIONS…………………………………………………………xviii

ABSTRACT..…………………………………………………………………………...xix

CHAPTER I……………………………………………………………………………….1

General introduction

Ecological disturbance and forests……………………………………..…1

Insects and Carabidae…………………………………...………………..2

Objectives……………………..…………………………………………...4

LITERATURE CITED……………………………………………………7

CHAPTER II……………………………………………………………………………..16

Changes in carabid beetle (Coleoptera: Carabidae) assemblages with forest age

and sampling year

ABSTRACT……………………………………………………………...17

INTRODUCTION……………………………………………………….18

METHODS………………………………………………………………22

RESULTS………………………………………………………………..30

DISCUSSION……………………………………………………………40

ACKNOWLEDGMENTS…..…………………………………………...56

LITERATURE CITED…………………………………………………..57

TABLES…………………………………………………………………79

FIGURES………………………………………………………………...82

APPENDIX A……………………………………………………………91

APPENDIX B……………………………………………………………94

APPENDIX C……………………………………………………………98

APPENDIX D…………………………………………………………..107

APPENDIX E…………………………………………………………..109

APPENDIX F…………………………………………………………..111

APPENDIX G…………………………………………………………..117

APPENDIX H…………………………………………………………..120

CHAPTER III…………………………………………………………………………..122

Changes in species traits of carabid beetles (Coleoptera: Carabidae)

with forest age

ABSTRACT…………………………………………………..………...123

INTRODUCTION…………………………………………………..….124

METHODS……………………………………………………………..128

RESULTS..……………………………………………………………..134

DISCUSSION…………………………………………………..………137

ACKNOWLEDGMENTS…..…………………………………..……...144

LITERATURE CITED…………………………………………………145

TABLES………………………………………………………………..163

FIGURES…………………………………………………………...…..166

APPENDIX A…………………………………………………………..172

APPENDIX B…………………………………………………………..178

APPENDIX C…………………………………………………………..186

APPENDIX D…………………………………………………………..187

CHAPTER IV…………………………………………………………………………..188

Carabid beetle (Coleoptera: Carabidae) species richness, diversity and community

structure in the understory of temporarily flooded and non-flooded Amazonian

forests

ABSTRACT………………………………………………………….....189

INTRODUCTION……………………………………………………...190

METHODS……………………………………………………………..195

RESULTS………………………………………………………………203

DISCUSSION…………………………………………………………..209

ACKNOWLEDGMENTS…………..………………………………….220

LITERATURE CITED…………………………………………………221

TABLES………………………………………………………………..246

FIGURES……………………………………………………………….248

APPENDIX A…………………………………………………………..257

APPENDIX B…………………………………………………………..259

APPENDIX C…………………………………………………………..265

APPENDIX D…………………………………………………..………270

APPENDIX E………………………………………..…………………272

CHAPTER V…………………………………………………………………………...273

Results from two sampling techniques for carabid beetles (Coleoptera: Carabidae)

in temporarily flooded and terra firme rainforest of western Amazonia

Published in Studies on Neotropical Fauna and the Environment (2016)

ABSTRACT………………………………………………………….....274

INTRODUCTION……………………………………………………...275

MATERIALS AND METHODS…...…………………………………..279

RESULTS………………………………………………………………284

DISCUSSION…………………………………………………………..287

CONCLUSIONS AND RECOMMENDATIONS……………………..292

ACKNOWLEDGMENTS……………..……………………………….294

REFERENCES…………………………………………………………295

TABLES………………………………………………………………..310

FIGURES……………………………………………………………….312

APPENDIX 1…………………………………………………………...319

APPENDIX 2…………………………………………………………...322

CHAPTER VI…………………………………………………………………………..324

CONCLUSIONS………………………………...……………………………..324

SCHOLASTIC VITA…………………………………………………………………..326

vii

LIST OF TABLES

II Table – 1. Measures of abundance, species richness and diversity of Carabidae among

the four forest age classes and the summed totals for the dataset…………..………79

II Table – 2. Mean Bray-Curtis dissimilarity values for the forest age classes with all

sampling years combined…………………………………………………………...80

II Table – 3. Carabid species with significant associations (p < 0.05) for one of the four

forest age classes or groups of forest age classes. Associations were based on the

point-biserial group-equalized phi coefficient…………………………….………..81

II Table – 4. Numbers of individuals, species richness and diversity of carabid beetles by

sampling year…………………………………………………………………….…82

II Table – B1. Additional site information for the 34 sampling sites including tree

circumferences and site elevation…………..……………………………………….94

II Table – B2. Longitude and latitude for the 34 study sites…………………………….97

II Table – C1. Scientific names, species codes and tribes for the 76 Carabidae species

collected …………………………….……………………..………………………..98

II Table – C2. Number of individuals of each species collected from the four forest age

classes……………………………………………………..……………………….101

II Table – C3. Numbers of individuals for each species collected from the four forest age

classes for each sampling year…………………………………………………….103

II Table – D1. Mean Bray-Curtis dissimilarity values by sampling year for community

analyses of year-to-year variations among forest age classes……………………..107

II Table – D2. Mean Bray-Curtis dissimilarity by sampling year for community analyses

of year-to-year variations among forest age classes ………………………………108

viii

II Table – E1. Mean Bray-Curtis dissimilarity values within and among forest age with

the first three sampling years combined (extra year of sampling omitted)………..109

II Table – F1. Mean Jaccard dissimilarity values (presence/absence) for the forest age

classes with all sampling years combined…………………………………………114

II Table – F2. Mean Jaccard dissimilarity values (presence/absence) within and among

sampling years for the forest age classes…………………………………………..115

II Table – F3. Mean Jaccard dissimilarity values (presence/absence) for within and

among forest age classes by sampling year………………………………………..116

II Table – H1. Carabidae tribes with higher than expected numbers of abundant

species……………………………………………………………………………..120

III Table – 1. The seven ecological species traits for Carabidae used in analyses of

functional diversity indices and individual traits……………………………….…163

III Table – 2. Mean (± s.e.) numbers of individuals (N), species richness (S) and

Simpson’s concentration index (1/D) for four forest age classes…………………165

III Table – 3. Analysis of deviance results (frequentist approach) of each response

variable for forest age and year (N = 94, df = 3)…………………………….…….164

III Table – A1. Scientific names, species codes and tribes for the 48 species in this

study..…...... 172

III Table – A2. Species classifications for the 18 species trait categories……………...174

III Table – A3. Number of individuals of each species collected from four forest age

classes...……………………………………………………………………………176

III Tables – B1-B5. Supplementary tables: Model sets with AICc comparisons of linear

model parameters for numbers of individuals, species richness, diversity and

functional diversity indices (FD)…………………………………………………..179

III Tables – B6-B21. Supplementary tables: Model sets with AICc comparisons of linear

model parameters for ecological species traits……..…….…………………..……181

III Table – C1. Analysis of deviance results (frequentist approach) for each trait category

for the two fixed effects: forest age and year………………………………….…..186

III Table – D1. The means ± standard error for the numbers of individuals (N), species

richness (S), the inverse Simpson’s concentration index (1/D) and functional

diversity indices for the four forest age classes and sampling years………………187

IV Table – 1. Measures of abundance, species richness and diversity of Carabidae

collected for floodplain (FP) and terra firme (TF) forests……………………..…..246

IV Table – 2. List of the characteristic Carabidae species which had significant indicator

values for floodplain (FP) and terra firme (TF) forests……………………………247

IV Table – B1. List of the 143 Carabidae morphospecies collected, including the tribe,

genera and identification code for each morphospeices…………………………...259

IV Table – C1. Carabidae tribes with higher numbers of common species than

expected.…………………………………………………………………………..265

V Table – 1. Numbers of individuals, richness, diversity and nonparametric estimators for

Carabidae collected by flight intercept traps (FITs) and hand sampling in floodplain

(FP), terra firme (TF) forests and with all data combined……………………..…..310

V Table – 2. Bray-Curtis mean dissimilarity measures within collection techniques

(diagonal) and between collection techniques (shaded boxes) for Carabidae

community assemblages………………………………………………...…..……..311

V Table – A1. Tribal and generic composition of numbers of carabid individuals by

collection technique and forest type as well as the combined totals……...……….319

V Table – A2. Tribal and generic composition of Carabidae numbers of species by

collection technique and forest type as well as the combined totals…………..…..322

LIST OF FIGURES

I Figure – 1. A hypothesis for the phylogenetic relationships of the four suborders of

Coleoptera…………………………………………………………………………..13

I Figure – 2. A hypothesis for the phylogenetic relationships of the extant families within

Adephaga……………………………………………………………………………14

I Figure – 3. A hypothesis for the phylogenetic relationships of higher taxa within

Carabidae……………………………………………………………………………15

II Figure – 1. Map of the study area and the 34 sampling sites with forest age classes as

depicted by different point markers………………………………………………....83

II Figure – 2. Rarefaction interpolation (solid lines) and extrapolation (dashed lines)

curves for zero, ten, fifty and ninety year forests with all four sampling years

combined…………………………………………………………………………....84

II Figure – 3. Boxplots by forest age and sampling year for the abundance-based

nonparametric diversity estimator, Chao1.…………...... 85

II Figure – 4. Rank abundance distribution curves for the four forest age classes with all

sampling years combined……….…………………….…………………………….86

II Figure – 5. Non-metric multidimensional scaling (NMDS) ordination of Carabidae

species assemblages for the four forest age classes with all sampling years combined.

All species (S = 76) were included in the analysis………………………………….87

II Figure – 6. Numbers of individuals and species known to prefer open habitats or

shaded habitats changed with forest age……………………………………………88

II Figure – 7. Numbers of individuals and number of species for the first three sampling

years……………………………..…………………………………………………..89

xii

II Figure – 8. Ordination analyses depicting year-to-year variation in species composition

within zero, ten, fifty and ninety year old forests.………………………………….90

II Figure – 9. Ordination analyses depicting year-to-year variation among forest age

classes. Data includes all species from the first three years of sampling…………...91

II Figure – A1. Normal monthly averages for precipitation, max temperature (Tmax),

minimum temperature (Tmin), and mean temperature (Tmean)...... 92

II Figure – A2. Annual summed precipitation (inches) and average temperatures (°F)

from 2006 – 2013 for the northern Piedmont of North Carolina……...…….……...93

II Figure – E1. Non-metric multidimensional scaling (NMDS) ordination of Carabidae

species assemblages for the four forest age classes with all the first three sampling

years combined. Extra sampling year excluded.…………………………….…….110

II Figure – F1. Non-metric multidimensional scaling (NMDS) ordination of Carabidae

species assemblages for the four forest age classes using the Jaccard dissimilarity

index (presence/absence data)….…………………………………………..……...111

II Figure – F2. Ordination analyses depicting year-to-year variation in species

composition within zero, ten, fifty and ninety year old forests using the Jaccard

dissimilarity index (presence/absence data) ………………………………………112

II Figure – F3. Ordination analyses depicting year-to-year variation among forest age

classes using the Jaccard dissimilarity index (presence/absence data) ……..……113

II Figure – G1. Carabid tribes with significant differences of number of individuals

among forest age classes. …………………………………………………………117

II Figure – G2. Numbers of individuals for five Carabidae tribes among the three

sampling years…………………………………………………………………….119

xiii

II Figure – H1. For the rare species category, boxplots depict the number of individuals

(A) and number of species (B) among forest age classes………………….…….121

III Figure – 1. Map of the 34 sampling sites in the North Carolina Piedmont with forest

age classes depicted by different point markers..…………….……………….…..166

III Figure – 2. Functional diversity indices (mean ± s.e.) among the forest age classes:

functional richness (FRic), functional evenness (FEve) and functional dispersion

(FDis)…………………….…………………………………………..………..….167

III Figure – 3. Number of Carabidae species (mean ± s.e.) for six ecological species trait

categories among four forest age classes………………………………………….168

III Figure – 4. Relative frequency histograms of Carabidae body length for four forest age

classes……………………………………………………………………………...169

III Figure – 5. Relative frequency histograms of Carabidae body length with all forest age

classes combined…………………………………………………………………..170

III Figure – 6. Numbers of Carabidae species (mean ± s.e.) for two species traits which

were close statistical significance (p < 0.1)…………….……………………….…171

IV Figure – 1. Map of the study area near Yasuní National Park in eastern Ecuador….248

IV Figure – 2. Rarefaction interpolation (solid lines) and extrapolation curves (dashed

lines) with point markers represent the sampling extent of the current study and

shaded areas depicting unconditional 95% confidence intervals……………….…249

IV Figure – 3. Inverse Simpson’s concentration (1/D) for floodplain (FP) forest and terra

firme (TF) forests.……………………………………………………………..…..251

IV Figure – 4. Mean beetle body length (mm) of carabid individuals collected in

floodplain forest (FP) and terra firme forest (TF), excluding Cicindelini………...252

xiv

IV Figure – 5. Number of Cincidelini (tiger beetles) collected from terra firme (TF) and

floodplain (FP) forests. ……………………………………………………………253

IV Figure – 6. Simpson’s evenness index (E1/D) indicated floodplain (FP) species

assemblages were significantly more even than terra firme (TF) forests…………254

IV Figure – 7. Rank abundance distribution curves for temporarily flooded (FP) and non-

flooded terra firme forests (TF)……………………………...…………………….255

IV Figure – 8. Non-metric multidimensional scaling (NMDS) ordination using Bray-

Curtis dissimilarity for Carabidae species assemblages from floodplain (FP) and

terra firme (TF) forests………………………………………………………….…256

IV Figure – A1. Water height for the Tiputini River at Tiputini Biodiversity Station,

Ecuador………………………………………………………………………….…257

IV Figure – C1. Number of individuals for the morphospecies Pentb (genus Pentacomia)

for floodplain (FP) and terra firme (TF) forests…………………..……………….266

IV Figure – C2. Number of rare species for the floodplain (FP) and terra firme (TF) forest

(p = 0.04).………………………………………………………………………….267

IV Figure – C3. Species relative abundance by the number of sampling sites at which

they were present ……………………………………………………….…………268

IV Figure – C4. Rank abundance distribution curves for terra firme rainforest (TF) in

Ecuador, and mature, mixed hardwood deciduous temperate forest in North Carolina,

USA……………………………………………………………………..…………267

IV Figure – D1. Numbers of individuals for Carabidae tribes with significant differences

between floodplain (FP) and terra firme (TF) forests………..…………………....270

IV Figure – D2. Numbers of species for Carabidae tribes for floodplain (FP) and terra

firme (TF) forests………………………………………………………………….271

IV Figure – E1. Non-metric multidimensional scaling (NMDS) ordination based on

Jaccard dissimilarity values (presence/absence data) for Carabidae species

assemblages for floodplain (FP) and terra firme (TF) forests……………….….…272

V Figure – 1. a, South America with the Amazon Basin outlined; b, Ecuador with the

locations of Tiputini Biodiversity Station (TBS) and Yasuní National Park; and c,

the 24 study sites at TBS………………………………………………………..…312

V Figure – 2. Flight-intercept trap (FIT) utilized in this study. FITs consisted of a frame

made of PVC tubing, screening stretched between PVC tubing, with a support tube

at the top of the trap, to create an approximately 1 m2 flight barrier……….…….313

V Figure – 3. Boxplot showing numbers of individuals among total flight intercept traps

(FITs) collections and hand sampling techniques between the two sampling

techniques in floodplain (FP) and terra firme (TF) forests…………………..……314

V Figure – 4. Sample-based rarefaction curves, corrected by the number of incidences,

for flight intercept traps (FITs), for hand sampling and for both methods

combined…………………………………………………………………………..315

V Figure – 5. Sample-based rarefaction curves, corrected by the number of incidences,

for flight intercept traps (FITs) and hand sampling within floodplain (FP) and terra

firme (TF) forests……………………………………………………………….....316

V Figure – 6. The three non-parametric species richness estimators and observed

numbers of uniques (species represented in only one sample) and duplicates (species

represented in only two samples) ………………………………………………...317

xvi

V Figure – 7. Non-metric multidimensional scaling (NMDS) ordination based on 48 data

points, representing each of the 24 study sites in terms each sampling methods (FITs

and hand) in floodplain (FP) and terra firme (TF) forests……………...………….318

xvii

ABBREVIATIONS

Abbreviations are noted within chapters. Those listed here in alphabetical order were more frequently used or more commonly used.

 AICc – Akaike information criterion corrected for small sample size

 carabid – beetle within the family Carabidae

 df – degrees of freedom

 GLM – generalized linear model

 GLMM – generalized linear mixed model

 FP – floodplain forest

 ha – hectare

 km – kilometer

 m – meter

 mm – millimeter

 n – number of individuals of a subsample

 N – total number of individuals

 NMDS – non-metric multidimensional scaling

 oz. - ounce

 PMSP – Pilot Mountain State Park

 S – number of species

 s.e. – standard error of the mean

 TBS – Tiputini Biodiversity Station

 TF – terra firme forest

xviii

ABSTRACT

Riley, Kathryn N.

CARABID BEETLE ECOLOGY AS A FUNCTION OF DISTURBANCE IN

TEMPERATE AND TROPICAL FORESTS

Dissertation under the direction of

Robert A. Browne, Ph.D., Professor of Biology

Ecological disturbance is an episodic event which causes temporary disruption to components integral to structure and function of an ecosystem. With the majority of today’s forests recovering from past disturbance, knowledge of the changes and patterns of forest species in relation to disturbance events are important to our understanding of ecosystem change. The focus of this research is to examine the effects of forest disturbance on species distributions and species assemblages, with carabid beetles

(Coleoptera: Carabidae) as the model organism group.

Chapter I of this dissertation provides a general overview of disturbance and its role in diversity and community-level processes, as well as a brief introduction to forest disturbances and the high diversity of arthropods including Carabidae. In the remaining chapters, I examine the responses of Carabidae for two types of disturbances: the clear- cutting of North American temperate forests and the temporary flooding of a Neotropical forest. My first objective was to examine variations in carabid beetle abundance, species

xix richness, diversity, and species assemblages with changes in forest age and for sampling year in the temperate forests of North Carolina (Chapter II). Next, I quantified changes in the ecological species traits of carabid beetles with temperate forest age (Chapter III). In

Chapter IV, I investigated the distributions of abundance, species richness, diversity, and species assemblages for carabid beetles from temporarily flooded and non-flooded terra firme forests of the Ecuadorian Amazon. Lastly in Chapter V, I evaluated differences in

Amazonian carabid abundance, diversity and species assemblages collected by hand sampling and flight intercept traps (FITs).

The results show that there is a significant relationship between forest age and carabid beetles for the temperate forests of North Carolina. Carabid species richness was significantly higher in recently clear-cut areas compared to more mature forests. Species composition was significantly different among forest age classes, with the most distinct carabid assemblages from recently clear-cut areas. Sampling year can also affect the number of individuals, species richness, diversity and communities of Carabidae both within and among forest age classes. For three functional diversity indices measuring richness, evenness and dispersion, there were no significant differences among forest age classes. The characterization of carabid species assemblages based on ecological species traits demonstrated that there were significant shifts for six traits with forest age.

In Amazonian temporarily flooded forests (FP) and non-flooded terra firme forest

(TF) the species richness, diversity and community evenness were higher in FP compared to TF forests. Community analyses indicated that there are distinct differences in

Carabidae species assemblages between FP and TF forests with few morphospecies shared between the two forest types. In terms of the number of individuals collected,

xx hand sampling was not as efficient for TF as it was for FP forests. Community-level analyses indicated that FITs and hand sampling target different components of the

Carabidae community with high dissimilarity values between groups signifying the high level of variation among sampling techniques within FP and TF forests.

In conclusion, the diversity and community assemblages of Carabidae for temperate and tropical forests responded to the disturbances studied. Both showed an increase in species richness and diversity in more recently disturbed habitat. There also were distinct carabid beetle species assemblages for recently disturbed habitats compared to habitats not recently influenced by these disturbance factors.

xxi

CHAPTER I

INTRODUCTION

Ecological disturbance and forests

Ecological disturbance is an episodic event that causes temporary disruption to components integral to the structure and function of an ecosystem (Pickett & White 1985;

Hobbs 2009), which has a vital role in the dynamics of diversity and community structure within ecosystems (Connell 1978; McLeod 1980; Petraitis et al. 1989; Rosenzweig 1995).

Disturbances can vary in extent, frequency and magnitude, with the effects depending on the type of disturbance, as well as site characteristics such as climate, soil, biota and vegetation (Pickett & White 1985; Rykiel 1985; Hobbs 2009). Ecosystems are influenced by an array of naturally occurring disturbance factors such as fires, wind, flooding and hurricanes (White 1979; Pickett & White 1985; Attiwill 1994; White & Jentsch 2001).

Human activities, such as forestry, pollution and fire suppression, frequently modify natural disturbances regimes or initiate novel disturbances to ecological systems (Attiwill

1994; White & Jentsch 2001; Hobbs 2009; Solórano & Páez-Acosta 2009). Disturbance events can change the conditions of the physical environment, the demographic rates, as well as the distribution and availability of resources (Tilman 1982; Petraitis et al. 1989;

Hobbs 2009; Mouillot et al. 2013), thus potentially altering the competitive balance of a system that can affect mechanisms of coexistence and dominance patterns of inhabiting species (Levin & Paine 1974; Connell 1978; Tilman 1982; Southwood 1988; Chaneton &

Facelli 1991). Natural disturbance has been postulated as an important mechanism in the maintenance of biological diversity in communities including the immense diversity of tropical forests (Connell 1978; Petraitis et al. 1989; Rosenzweig 1995).

Natural disturbances play a fundamental role in the structure and function of forest ecosystems (Pickett & White 1985; Oliver & Larson 1990; Foster et al. 1998). In contemporary times, human activities, both indirectly and directly, are another major source of change within forests (Oliver & Larson 1990; Attiwill 1994; Linke et al. 2007;

Solórano & Páez-Acosta 2009). Deforestation and forest use have increased immensely over the past century and are predicted to continue to increase (Solórano & Páez-Acosta

2009). Forests and their responses to disturbance are also vulnerable, both directly and indirectly, to the effects of global climate change (Dale et al. 2001; Bush et al. 2004;

Bonan 2008). In a rapidly changing world, knowledge of the changes and patterns of forest species in relation to disturbance events are of high importance. This is especially true for the smaller forest inhabitants (“the little things”), arthropods and their relatives which are the dominate taxa in terms of biomass, numbers and diversity within forests

(Wilson 1987; Erwin 1996).

Insects and Carabidae

Insects and their relatives dominate global biodiversity (Hammond 1992;

Samways 1993) with approximately 1.24 million arthropod species described by science, over 80% of which are insects (Zhang 2011). The recent global estimate of 6.1 million arthropod species (Hamilton et al. 2013), suggests that around 60 – 80% of species await description (Erwin & Johnson 2000; Dunn 2005; Hamilton et al. 2010, 2013; May 2010).

The majority of undescribed taxa occur in tropical regions where global insect diversity is

2 the highest (Hamilton et al. 2013). Based on the estimates of species awaiting description, we are in the early stages of a comprehensive understanding of these animals which make up an overwhelming portion of the Earth’s species.

Beetles (Coleoptera), the largest order within Insecta, constitute a fourth of all animal species with approximately 400,000+ described species (Hammond 1992; Erwin

1996; Hunt et al. 2007). The immense diversity of Coleoptera makes research of these animals particularly important to ecological processes and the distribution of global diversity across time and space (Erwin 1996). The beetle family Carabidae is in the second largest suborder of Coleoptera, the Adephaga (Figure 1), and is the most species rich family within this suborder (Maddison 2000; Bousqet 2012). Adephaga can be divided into the terrestrial fauna, the Geadephaga, and the aquatic fauna, the

Hydradephaga (Figure 2; Crowson 1960; Beutel and Haas 2000; Beutel et al. 2008;

Lawrence et al. 2011; Bousquet 2012). Monophyly is currently not well supported for the

Carabidae (see Maddison et al. 2009) and although molecular sequence data has been used for specific subsets of Carabidae, much of the current classification is predominantly based on morphological data (Maddison 2006; Bousquet 2012). Figure 3 provides a higher level phylogeny for Carabidae; however, more data are necessary to reach a consensus for higher classifications (Maddison 2006; Maddison et al. 2009;

Bousquet 2012). With approximately 40,000 species globally, Carabidae is the third most species rich beetle family (Erwin 1985) and they are found on all continents except

Antarctica. Carabids are holometabolous (i.e. have complete metamorphosis), with the full life cycle for the majority of species taking around one year to complete and one reproductive event (Lövei & Sunderland 1996). Carabid beetles are relatively well-

3 known taxonomically, easily sampled in high numbers and respond to abiotic and biotic variations including disturbances (Thiele 1977; Lövei & Sunderland 1996; Rainio &

Niemela 2003), making them effective model organisms for ecological studies (Niemelä et al. 2000; Rainio & Niemela 2003; Pearce & Venier 2006; Koivula 2011; Kotze et al.

2011).

Objectives

The goal of this research is to gain a greater understanding of patterns in species abundance, diversity and community assemblages in relation to forest disturbance, with carabid beetles serving as the model organism group. I examine the responses of

Carabidae for two types of disturbances: the clear-cutting of North American temperate forests, an anthropogenic disturbance, and the temporary flooding of a Neotropical forest, a natural disturbance factor. Collectively, the results of this dissertation are based on

7,803 carabid beetles, collected from two continents over a four year period (2009 – 2013) through 168 hand collecting events and the repeated sampling of 432 pitfall traps and 24 flight interception traps. Chapters II and III discuss results from temperate forests while

Chapters IV and V discuss results from a tropical rainforest.

In Chapter II, the objectives are to examine variations in carabid beetle abundance, speices richness, diversity and species assemblages among four forest age classes and sampling year in North Carolina. Four forest age classes, ranging from recent clear-cuts to ninety year old forests, were sampled from 2009 – 2013, with predictions that (1)

Carabidae abundance will be lowest in recent clear-cuts and highest in ninety year forests;

(2) richness and diversity will increase after clear-cutting due to an influx of species well adapted to open habitats; (3) forest age classes will be characterized by distinct carabid

4 assemblages with the most distinct species assemblages occurring between recent clear- cuts and ninety year old forests; (4) year-to-year changes in species assemblages within the forest age classes will be most rapid in the clear-cut areas due to the recent disturbance; and (5) there will likely be year-to-year variations in species assemblages among forest age classes; however, overall community-level patterns should remain consistent among sampling years.

Changes in carabid functional diversity and ecological species traits in response to temperate forest age are examined in Chapter III. Classification of diversity and communities based on traits relating to species ecology rather than on taxonomy alone can reveal ecologically important aspects of biological communities and serve broader conservation purposes (Whittaker 1975). Functional diversity and species trait patterns are predicted to change with forest age with the carabid communities of recent clear-cuts will likely contain a different suite of traits compared to those of more mature forests. I hypothesize that functional richness will be the highest for recent clear-cuts but functional evenness and diversity will be higher in the fifty and ninety year forests compared to less mature forests.

The differences in Carabidae species richness, diversity and community composition within the understory of temporarily flooded forests (FP) and non-flooded terra firme forest (TF) of eastern Ecuador are examined in Chapter IV. I predict that FP forests will have higher species richness and diversity than TF forests. Since there are species specific adaptions and survival strategies for flooding regimes, species assemblages may differ between the two forest types with greater variation among sampling sites within FP communities.

In Chapter V, I examine differences in the estimates of carabid diversity and species assemblages from two collection techniques, hand sampling and flight intercept traps (FITs). The objectives of this study were to investigate differences between FITs and hand sampling in terms of numbers of individuals, richness, diversity and species composition for Carabidae within the understory of two Amazonian forest types. The two sampling techniques are expected to target different portions of the carabid beetle community, but the extent of these differences and the relative sampling efficiency differs between FP and TF forests, have not previously been explored.

LITERATURE CITED

Attiwill, P. M. 1994. The disturbance of forest ecosystems: the ecological basis for

conservative management. Forest Ecology and Management 63:247–300.

Beutel, R. G., and F. Haas. 2010. Phylogenetic relationships of the suborders of

Coleoptera (Insecta). Cladistics 16:103–141.

Beutel, R. G., I. Ribera, O. R. P. Bininda-Emonds. 2008. A genus-level supertree of

Adephaga (Coleoptera). Organisms, Diversity & Evolution 7:255–269.

Bonan, G. B. 2008. Forests and climate change: forcings, feedbacks, and the climate

benefits of forests. Science 320:1444–1449.

Bousquet, Y. 2012. Catalogue of Geadephaga (Coleoptera, Adephaga) of America, north

of Mexico. ZooKeys 1722:1–1722.

Bush, M. B., M. R. Silman, and D. H. Urrego. 2004. 48,000 years of climate and forest

change in a biodiversity hot spot. Science 303:827–829.

Chaneton, E. J., and J. M. Facelli. 1991. Disturbance effects on plant community

diversity: spatial scales and dominance hierarchies. Vegetatio 93:143–155.

Connell, J. H. 1978. Diversity in tropical rain forests and coral reefs. Science 199:1302–

1310.

Crowson, R. A. 1960. The phylogeny of Coleoptera. Annual Review of Entomology, 5:

111-134.

Dale, V. H., L. A. Joyce, S. McNulty, R. P. Neilson, M. P. Ayres, M. D. Flannigan, and B.

Michael Wotton. 2001. Climate change and forest disturbances. BioScience 51:723–

734.

Dunn, R. R. 2005. Modern insect extinctions, the neglected majority. Conservation

Biology 19:1030–1036.

Erwin, T. L. 1985. The taxon pulse: a general pattern of lineage radiation and extinction

among carabid beetles. Pages 438–472 in G. E. Ball, editor. Taxonomy, Phylogeny

and Zoography of Beetles and Ants. Dr. W. Junk Publishers, Dordrecht.

Erwin, T. L. 1996. Biodiversity at its utmost: tropical forest beetles. Pages 27–40 in M. L.

Reaka-Kudla, D. E. Wilson, and E. O. Wilson, editors. Biodiversity II:

Understanding and Protecting our Biological Resources. Joseph Henry Press,

Washington, DC.

Erwin, T. L., and P. J. Johnson. 2000. Naming species, a new paradigm for crisis

management in taxonomy: rapid journal validation of scientific names enhanced

with more complete descriptions on the internet. Coleopterists Bulletin 54:269–278.

Foster, D. R., D. H. Knight, and J. F. Franklin. 1998. Landscape patterns and legacies

resulting from large, infrequent forest disturbances. Ecosystems 1:497-510.

Hamilton, A. J. et al. 2013. Estimating global arthropod species richness: refining

probabilistic models using probability bounds analysis. Oecologia 171:357–365.

Hamilton, A. J., Y. Basset, K. K. Benke, P. S. Grimbacher, S. E. Miller, V. Novotný, G.

A. Samuelson, N. E. Stork, G. D. Weiblen, and J. D. L. Yen. 2010. Quantifying

uncertainty in estimation of tropical arthropod species richness. American Naturalist

176:90–95.

Hammond, P. 1992. Species inventory. Pages 17–39 in B. Groombridge, editor. Global

Biodiversity: Status of the Earth’s Living Resources. Chapman and Hall, London.

Hobbs, R. J. 2009. Restoration Ecology. Pages 566–572 in S. Levin, editor. The

Princeton Guide to Ecology. Princeton University Press, Princeton, New Jersey.

Hunt et al. 2007. A comprehensive phylogeny of beetles reveals the evolutionary origins

of a superradiation. Science 318:1913–1916.

Koivula, M. J. 2011. Useful model organisms, indicators, or both? Ground beetles

(Coleoptera, Carabidae) reflecting environmental conditions. Pages 287–317 in D. J.

Kotze, T. Assmann, J. Noordijk, H. Turin, and R. Vermeulen, editors. Carabid

Beetles as Bioindicators: Biogeographical, Ecological and Environmental Studies.

ZooKeys100:287–317.

Kotze, J. D. et al. 2011. Forty years of carabid beetle research in Europe - from taxonomy,

biology, ecology and population studies to bioindication, habitat assessment and

conservation. ZooKeys 100:55–148.

Lawrence, J. F., A. Ślipiński, A. E. Seago, M. K. Thayer, A. F. Newton, and A. E.

Marvaldi. 2011. Phylogeny of the Coleoptera based on morphological characters of

adults and larvae. Annales Zoologici 61:1-217.

Levin, S. A., and R. T. Paine. 1974. Disturbance, patch formation, and community

structure. Proceedings of the National Academy of Sciences 71:2744–2747.

Linke, J., M. G. Betts, M. B. Lavigne, and S. Franklin. 2007. Introduction: Structure,

function, and change of forest landscapes. Pages 1–29 in M. Wulder and S. E.

Franklin, editors. Understanding Forest Disturbance and Spatial Pattern: Remote

Sensing and GIS Approaches. CRC Press, Taylor & Francis Group, LLC. Boca

Raton, Florida.

Lövei, G. L., and K. D. Sunderland. 1996. Ecology and behavior of ground beetles

(Coleoptera: Carabidae). Annual Review of Entomology 41:231–256.

Maddison, David R. 1995. Adephaga. Version 01 January 1995 (under

construction). Available from: http://tolweb.org/Adephaga/8875/1995.01.01 in The

Tree of Life Web Project, http://tolweb.org/

Maddison, D. R. 2000. Coleoptera. Beetles. Version 11 September 2000 (under

construction). Available from: http://tolweb.org/Coleoptera/8221/2000.09.11 in The

Tree of Life Web Project, http://tolweb.org/

Maddison, D. R. 2006. Carabidae. Ground beetles and tiger beetles. Version 11 April

2006 (under construction). Available from:

http://tolweb.org/Carabidae/8895/2006.04.11 in The Tree of Life Web Project,

http://tolweb.org/

Maddison, D. R. W. Moore, M. D. Baker, T. M. Ellis, K. A. Ober, J. J. Cannone, and R.

R. Gutell. 2009. Monophyly of terrestrial adephagan beetles as indicated by three

nuclear genes (Coleoptera: Carabidae and Trachypachidae). Zoologica Scripta

38:43–62.

May, R. M. 2010. Tropical arthropod species, more or less? Science 329:41–42.

McLeod, J. M. 1980. Forests, disturbances, and insects. Canadian Entomologist

112:1185–1192.

Mouillot, D., N. A. J. Graham, S. Villéger, N. W. H. Mason, and D. R. Bellwood. 2013.

A functional approach reveals community responses to disturbances. Trends in

Ecology and Evolution 28:167–177.

Niemelä, J., J. Kotze, A. Ashworth, P. Brandmayr, K. Desender, T. New, L. Penev, M.

Samways, and J. Spence. 2000. The search for common anthropogenic impacts on

biodiversity: A global network. Journal of Insect Conservation 4:3–9.

Oliver, C., and B. Larson, editors. 1990. Forest Stand Dyanmics. McGraw-Hill, Inc. New

York, New York.

Pearce, J. L., and L. A. Venier. 2006. The use of ground beetles (Coleoptera: Carabidae)

and spiders (Araneae) as bioindicators of sustainable forest management: A review.

Ecological Indicators 6:780–793.

Petraitis, P. S., R. E. Latham, and R. A. Niesenbaum. 1989. The maintenance of species

diversity by disturbance. Quarterly Review of Biology 64:393–418.

Pickett, S. T. A., and P. S. White, editors. 1985. The Ecology of Natural Disturbance and

Patch Dynamics. Academic Press, Orlando, FL.

Rainio, J., and J. Niemela. 2003. Ground beetles (Coleoptera : Carabidae) as bioindicators.

Biodiversity and Conservation 12:487–506.

Rosenzweig, M. L. 1995. Species diversity in space and time. Cambridge University

Press, Cambridge, UK.

Rykiel, E. J. 1985. Towards a definition of ecological disturbance. Australian Journal of

Ecology 10:361–365.

Samways, M. J. 1993. Insects in biodiversity conservation - some perspectives and

directives. Biodiversity and Conservation 2:258–282.

Solórano, L. A., and G. Páez-Acosta. 2009. Forests. Pages 606–613 in S. Levin, editor.

The Princeton Guide to Ecology. Princeton University Press, Princeton, New Jersey.

Southwood, T. R. E. 1988. Tactics, strategies and templets. Oikos 52:3–18.

Thiele, H. U. 1977. Carabid beetles in Their Environments: A Study on Habitat Selection

by Adaptation in Physiology and Behaviour. Springer-Verlag, Berlin.

Tilman, D. 1982. Resource competition and community structure. Princeton University

Press, Princeton, New Jersey.

White, P. S. 1979. Pattern, process, and natural disturbance in vegetation. Botanical

Review 45:229–299.

White, P. S., and A. Jentsch. 2001. The search for generality in studies of disturbance and

ecosystem dynamics. Progress in Botany 62: 399–450.

Whittaker, R. H. 1975. Communities and Ecosystems. MacMillan, New York, New York.

Wilson, E. O. 1987. The little things that run the world (The importance and conservation

of invertebrates). Conservation Biology 1:344–346.

Zhang, Z. 2011. Phylum Arthropoda von Siebold, 1848. Z. Q. Zhang, editor. Animal

Biodiversity: An Outline of Higher-level Classification and Survey of Taxonomic

Richness. Zootaxa 3148: 99–103.

I Figure – 1. A hypothesis for the phylogenetic relationships of the four suborders of

Coleoptera. The phylogenetic tree (Beutel and Haas 2000) is from The Tree of Life Web

Project (Maddison 2000).

I Figure – 2. A hypothesis for the phylogenetic relationships of the extant families within Adephaga. The rectangle highlights Carabidae, the study group for this dissertation. The phylogenetic tree from The Tree of Life Web Project (Maddison

1995).

I Figure – 3. A hypothesis for the phylogenetic relationships of higher taxa within

Carabidae. Question marks denote uncertain placement for tribes which have been classified within Carabidae, as well as raised to the rank of family or included in another family. Cicindelitae has been considered a distinct family by several coleopterists.

Adapted from rom The Tree of Life Web Project (Maddison 2006).

CHAPTER II

CHANGES IN CARABID BEETLE (COLEOPTERA: CARABIDAE)

ASSEMBLAGES WITH FOREST AGE AND SAMPLING YEAR

ABSTRACT

The objectives of this study were to assess changes in Carabidae among four forest age forest classes for (1) numbers of individuals, richness and diversity; (2) common and rare species; (3) species composition; (4) year-to-year variations in numbers of individuals, richness and diversity; and (5) changes in species composition among and within sampling years in relation to forest age. Carabid beetles were sampled by pitfall traps between 2009 and 2013 at 34 sites representing recently clear-cut areas (‘zero year’), ten year old forest, fifty year old forest and ninety year old forest. There were 6,548

Carabidae collected, from 16 tribes, 33 genera and 76 species. There were no differences in mean values among forest age classes for number of individuals, species number or for mean inverse Simpson concentration indices. Rarefaction curves indicate that the rarified richness for recent clear-cuts is higher than those for older forests. When all sampling years are combined, non-metric multidimensional scaling (NMDS) followed by permutational multivariate analysis showed that there were significant differences in species composition among forest age classes. The numbers of individuals and species collected in sampling year three were significantly lower than the previous two sampling years and Simpson’s evenness values indicated species assemblages were significantly more even in year three. There were significantly higher numbers of rare individuals and species collected from zero year age class compared to older forest age classes. NMDS analysis indicated that there were significant variations in species composition within zero, fifty and ninety year forests among sampling years. Year-to-year variations in species composition among the forest age classes were also detected and there were significant differences among the forest age classes in all three sampling years. Distinct

17 differences in species assemblages among forest age classes indicate that conservation decisions should be considered at the landscape level.

INTRODUCTION

Temperate forests, which cover approximately 50% of the eastern United States, have been highly modified in the past 300 years, with much of the area recovering from past disturbance events (Ricketts & Imhoff 2003; Bonan 2008; Drummond & Loveland

2010; Pan et al. 2011). Agricultural abandonment, commercial forestry, agriculture and urbanization have created a mosaic of various aged forest stands and land use cover types

(Oosting 1942; Christensen & Peet 1984; Henderson & Walsh 1995; Napton et al. 2010;

Pan et al. 2011; Brown 2015). With the need to support a rapidly growing population,

U.S. Southeastern forests are particularly vulnerable to anthropogenic pressures of urbanization and other land uses (Ricketts & Imhoff 2003; Bonan 2008; Drummond &

Loveland 2010). Southeastern forests were included in the “Global 200” list of conservation priority areas across the globe (Olson & Dinerstein 2002) and have been identified as a top ecoregion priority for conservation based on richness, endemism and risks associated with urbanization (Ricketts & Imhoff 2003).

Of the 18.6 million acres of forest in North Carolina in 2013, only 4% are reserved from commercial timber production, with the greatest timberland proportion occurring in the Piedmont region (Brown 2015). Clear-cutting, a dominant method of timber harvesting, causes abrupt change from forested to open habitat, radically altering the ground conditions and soil of the given habitat (Covington 1981; Swank & Vose

1988). Removal of the forest canopy leads to increased soil and air temperatures, larger fluctuations of temperature minima and maxima, decreased surface moisture, increased

18 wind flow and, perhaps most importantly, increased solar radiation. Changes associated with clear-cutting also cause changes in leaf litter composition and depth, soil pH, organic and nutrient matter, nutrient cycling and biological processes (Bazzaz 1979;

Covington 1981; Christensen & Peet 1984; Swank & Vose 1988; Chen et al. 1993).

Changes associated with clear-cutting also cause changes in the ground dwelling invertebrate communities, many of which are crucial to ecosystem processes and constitute a large portion of forest biodiversity (Lenski 1982; Heliövaara & Väisänen

1984; Stork & Eggleton 1992; Niemelä 1997).

In temperate forests, the Carabidae (Coleoptera) are dominant among forest floor invertebrates, both in numbers of individuals and species (Thiele 1977; Erwin 1985;

Werner & Raffa 2000). With approximately 40,000 carabid species globally, Carabidae is the third most species rich beetle family in North America, with an estimated 2,450 species north of Mexico (Erwin 1985; Ciegler 2000; Bousquet 2012). Many carabid species are ecologically important as generalist predators, while others are predominantly phytophagous, omnivorous or specialist feeders (Thiele 1977; Erwin 1996; Lövei &

Sunderland 1996). Carabid beetles are relatively well-known taxonomically, easily sampled in high numbers and respond to abiotic and biotic variations including disturbances and other environmental changes (Thiele 1977; Lövei & Sunderland 1996;

Rainio & Niemela 2003), making them effective model organisms for ecological studies

(Niemelä et al. 2000; Rainio & Niemela 2003; Pearce & Venier 2006; Koivula 2011;

Kotze et al. 2011). The environmental parameters and habitat characteristics associated with forest age generate species specific responses which are manifested in variations at the community level (Niemelä et al. 1996).

Carabidae richness and diversity are typically higher in the early phases following a clear-cutting event compared to mature forests (Niemelä et al. 1993; Spence et al. 1996;

Butterfield 1997; Jukes et al. 2001; Koivula et al. 2002; Pearce et al. 2003; de Warnaffe

& Lebrun 2004; Magura et al. 2015). This increased richness is usually associated with an increase in habitat generalist and open habitat species which have higher tolerances for temperature and light compared to forest species (Thiele 1977; Ings and Hartley 1999).

Disturbance can alter the competitive balance within an ecosystem, affecting coexistence mechanisms and dominance patterns of the inhabiting species and cause an increase in species diversity (Levin & Paine 1974; Connell 1978). I hypothesize that species richness and diversity will be highest in recently clear-cuts, with intermediate values for the most mature forests, and the lowest values for intermediately aged forests.

Previous studies, primarily from high latitude boreal forests, have found that forest age classes are characterized by distinct carabid assemblages, with the most extreme differences between early and mature forest stages (Szyszko 1990; Niemelä et al.

1993, 1996; Haila et al. 1994; Butterfield 1997; Ings & Hartley 1999; Koivula et al. 2002;

Magura et al. 2006; Taboada et al. 2008). The relationship between forest age and

Carabidae community structure is likely an indirect one, with species responding to changes in microclimatic differences within recently disturbed and forested areas rather than particular tree species (Ings and Hartley 1999). Forest canopy closure initiates changes on the forest floor, which is typically a pivotal point for ground-dwelling arthropods communities, including carabid beetles (Swank & Vose 1988; Chen et al.

1993; Niemelä et al. 1993, 1996; Pearce & Venier 2006; Koivula 2011). Closed canopies reduce the amount of sunlight and decrease wind at ground level, lowering temperature

20 and increasing humidity, creating more moderate conditions (Swank & Vose 1988; Chen et al. 1993). I hypothesize that changes associated with forest age in southeastern U.S. forests will be reflected in Carabidae species assemblages, and that the carabid communities occurring in recently clear-cut areas and mature forests will be particularly distinct compared to each other.

Although year-to-year variation in carabid species compositions can be substantial, it has not often been considered in the majority of analyses. Inter-annual studies are necessary to understand population and community dynamics (den Boer 1981,

1985; Luff 1982; Desender 1996; Kotze & Niemela 2002). Year-to-year changes in plant species richness, vegetation structure and habitat heterogeneity have been shown to be important factors for temporal variations in carabid abundance (Brose 2003; Harvey et al.

2008). Previous studies have found low fluctuation in numbers of individuals among years (Luff 1982; Niemelä et al 1988) while other studies have found marked differences in abundance of common species (den Boer 1985; French and Elliott 1999). I predict that more commonly collected species will be present in all sampling years but their abundances will vary significantly among years. Year-to-year changes in species assemblages will be most rapid in the recently clear-cut areas due to the recent disturbance and regenerative processes. More mature forests will have lower turnover in species assemblages among sampling years but the abundances of species present will vary due to stochasticity, variations in weather and microhabitat structure (Kotze and

Niemelä 2002; Harvey et al. 2008).

The majority of carabid diversity and community studies have been conducted in higher latitude boreal forests (Szyszko 1983; Niemelä et al. 1993, 1996; Spence et al.

1996; Butterfield 1997; Koivula et al. 2002; Jacobs et al. 2008) or in temperate forests of

Canada and northern Europe (Thiele 1977; Pearce et al. 2003; Latty et al. 2006; Jelaska et al. 2011; Kwiatkowski 2011; Magura et al. 2015). Studies from mid-latitudinal North

American temperate forests are underrepresented in the literature (Lenski 1982; Jennings et al. 1986; Riley & Browne 2011), particularly for the southeastern U.S. (Ciegler 2000).

The objectives of this study are to assess changes among four forest age classes in carabid beetle (1) numbers of individuals, richness and diversity; (2) common and rare species; (3) species composition; (4) year-to-year variation in numbers of individuals, richness, diversity; and (5) species composition both among and within sampling years.

METHODS

Study area

The Piedmont of North Carolina is a transitional region between the Blue Ridge

Mountains and Southeastern Plains with rolling topography and forests accounting for more than half of the land cover (Oosting 1942; Napton et al. 2010). Normal monthly mean temperatures for the study area are 35.0 – 76.0° F, with mean annual precipitation of 46.9 – 47.0 in. relatively evenly spread throughout the year (State Climate Office 2016;

Figure A1). The early stages of forest regeneration in the Piedmont are dominated by herbaceous plants (e.g. several Asteraceae, Panicum) and tree seedlings (Oosting 1942).

Rapidly growing pines (Pinus) form a closed canopy a few years after disturbance and typically persist to dominate the mid-successional stages while, at approximately the same time, early shade-intolerant hardwoods (e.g. Liriodendron tulipifera, Acer rubrum) become established (Peet & Christensen 1980; Christensen & Peet 1981). Decades after initial disturbance, hardwoods outnumber pines until the community transitions to more

22 shade-tolerant deciduous trees (e.g. Quercus spp., Carya spp.). Mature forests of the

Piedmont are uneven aged, hardwood deciduous forest, characterized by Quercus (oak),

Carya (hickory), and Liriodendron (poplar) species and on drier areas mixed oak-pine

(Pinus spp.). The overwhelming majority of southeastern U.S. forests are secondary forests regrown on abandoned agricultural land (Christensen & Peet 1984; Peet &

Christensen 1987; Henderson & Walsh 1995; Drummond & Loveland 2010; Napton et al.

2010). The classical descriptions of Piedmont succession are based on forests in the central Piedmont, approximately 120 km east of the current study area. Forests in the study area appear to reach the mature uneven-aged deciduous stage slightly earlier than what has been reported for the classical studies of North Carolina forest succession

(Oosting 1942; Peet & Christensen 1980, 1987, Christensen & Peet 1981, 1984).

Collection of Carabidae

Sampling occurred at 34 sites representing four forest age classes within Stokes and Surry counties of North Carolina (80° N, 36°W) (Figure 1; Appendix B, Table B2).

Four forest age classes are represented: recently clear-cut (n = 9), 10 – 20 year forest (n =

8), 40 – 60 year forest (n = 7) and 80 – 95 year forest (n =10), which are referred to as zero, ten, fifty and ninety year age groups respectively (see Appendix B for additional site information). Forest ages were estimated using historical records, forest structure and tree girth. Of the 34 sampling sites, 29 sites were repeatedly sampled from May 2009 through April 2012. Twenty-two sampling sites were located within Pilot Mountain State

Park (PMSP) either in the two larger continuous sections of the park or within the

Corridor/Bridle Trail (length = 10.6 km) which connects the larger sections. Six zero year sites were located within a 65 ha plot which was logged in December 2008. In order to

23 improve the spatial distribution of sample sites for the two younger age classes, an additional year of sampling, from November 2012 through October 2013, was implemented at seven sites (four zero year sites and three ten year sites). Two of these extra year sites (one zero year and one ten year site) were also sampled during the original three year period. Newly added zero year sites were clear-cut between Nov. 2011 and Nov. 2012. Site selection was affected by the availability of forest sites of the proper age class, urban surroundings, and collection permits. Sampling sites are similar in altitude (258 - 371 m) (Appendix B, Table B1) and soil classification (i.e.,

Metagraywacke and Muscovite – Biotite Schist (CZma2) (Brown 1985).

Pitfall trapping has shown to be useful for comparing species presence and relative activity/abundance levels of larger ground dwelling beetles across habitat types

(Thiele 1977; Spence & Niemelä 1994; Woodcock 2005). Although pitfall trap sampling likely has inherent biases relating to activity (i.e. ‘activity density’) (Greenslade 1964;

Luff 1975; Thiele 1977; Adis 1979; Spence & Niemelä 1994), sampling under standard conditions over the entire active season provides reliable information on activity densities and allows for quantitative comparison among species (Baars 1979; Luff 1982; Loreau

1992; Southwood & Henderson 2000). Passive pitfall trapping samples continuously, minimizing investigator associated biases (Melbourne 1999; Woodcock 2005). In this study, pitfall traps consisted of 16 oz. plastic drinking cups (Solo® with a 12 cm diameter opening) that were positioned flush with the ground. To reduce the amount of water and debris in the trap, a flat, dense foam plastic cover was placed over each cup and fixed approximately 5 cm above the ground with nails. The cups contained non-toxic antifreeze as a preservative. There were a total of 12 traps at each sampling site, divided into two

24 groups of six each with groups located approximately 10 m apart (see Riley 2010, Figure

1b). Carabids were collected monthly by straining the trap contents through a fine mesh screen and subsequently preserved in 95% ethanol. Individuals were identified to species using a morphological key (Ciegler 2000), with confirmation made as needed with specimens at the Smithsonian National Museum of Natural History in Washington, DC.

Statistical analysis

Numbers of individuals, richness and diversity

For each sampling site, individual pitfall trap collections were combined for each month, with the raw numbers of individuals adjusted for differences in number of trap nights between collection periods and for number of intact traps (i.e. versus those overturned). Total number of individuals collected at a site was standardized to 365 trap nights per month (12 traps x 30.42 days). Inter-annual comparisons were made between sampling years one, two and three, with the extra sampling year excluded from these analyses. For each sampling site and year, the number of individuals (N), raw species richness (S) and a Hill diversity number, the inverse Simpson’s concentration index (1/D)

(Simpson 1949; Jost 2006; Chao et al. 2014) were determined. To detect changes at a higher taxonomic level, the number of individuals and species of each carabid tribe were calculated for the forest age classes. These values were used as response variables in linear modeling, with forest age and sampling year as predictors (fixed effects). To assess changes among forest age classes and sampling years, I used generalized linear mixed models (GLMMs) with Poisson error distributions, log links and sampling site as a random factor to account for repeated sampling. If over-dispersion was detected (> 1.5), standard errors were corrected through a GLMM with a negative-binomial error

25 distribution, log link and sampling site as a random factor (Zuur et al. 2009; O’Hara &

Kotze 2010). The appropriate type of linear model (e.g. Poisson GLMM) was selected based on the data and the Akaike information criterion corrected for small sample size

(AICc) (Sugiura 1978; Hurvich & Tsai 1989; Bolker et al. 2009; Zuur et al. 2009;

Burnham et al. 2011). Significance was determined through analysis of deviance using

Wald chi-square test statistics (Zuur et al. 2009) and Tukey tests for multiple comparisons of means. Linear modeling was completed using R packages ‘lme4’,

‘MASS’, ‘car’, and ‘multcomp’ (Venables & Ripley 2002; Hothorn et al. 2008; Fox &

Weisberg 2010; Bates et al. 2015; R Core Team 2016).

Since species richness is sensitive to sample size, it was standardized by rarefaction which determines the expected number of species from a random subsample of individuals from the overall data set (Gotelli & Colwell 2001, 2011; Colwell et al.

2012). Rarefaction curves with 95% unconditional confidence intervals were constructed using 500 randomizations in R package ‘iNEXT’ (Chao et al. 2014; Hsieh et al. 2015).

To correct for differences in sample density among sampling sites, the x-axis was rescaled to individuals (Gotelli & Colwell 2001, 2011, Colwell et al. 2004, 2012). To estimate the number of undetected species, rarefaction curves were extrapolated by a factor of two based on the smallest sample size among age classes (Colwell et al. 2012).

Nonparametric species richness estimators were calculated to provide the potential true richness of the forest age classes. The nonparametric species richness estimator, Chao1, is an abundance-based estimator and was used for extrapolation of rarefaction. Chao1is usually recommended as the lower bound of asymptotic richness (Chao 1984; Colwell et al. 2012). To account for estimator bias, two additional nonparametric diversity

26 estimators and associated standard error (randomizations = 1000) were calculated through

R packages ‘vegan’ (Oksanen et al. 2016): the abundance-based coverage estimator

(ACE) (Chao & Lee 1992; Chao & Yang 1993; Lee & Chao 1994) and sample-based first order jackknife, Jack1 (Burnham & Overton 1979; Heltshe & Forrester 1983). Estimators were selected based on the abundance, distribution and evenness of the data, in addition to high performance in previous studies (Chazdon et al. 1998; Brose & Martinez 2004;

Walther & Moore 2005; Hortal et al. 2006; Gotelli & Colwell 2011). Since ACE accounts for the frequency of rare species, it has been shown to be appropriate for datasets with a high proportion of rare species as is common for studies examining Carabidae assemblages (Chao and Lee 1992; Chazdon et al. 1998; Belaoussoff et al. 2003). Jack1 has performed well in a variety of studies (Palmer 1990; Colwell & Coddington 1994;

Walther & Moore 2005; Hortal et al. 2006; Basualdo 2011) and it shows low bias for data with higher evenness and sample coverage (Brose & Martinez 2004; Brose et al. 2003).

The three estimators were averaged together to evaluate to sample completeness.

Species assemblages

Simpson’s evenness (E1/D) was analyzed though a beta-regression model with logit link function since values were continuous on the interval (0-1) using R package

‘betareg’ (Ferrari & Cribari-Neto 2004; Cribari-Neto & Zeileis 2010). This evenness index is known to have high precision and low bias (Smith & Wilson 1996; Payne et al.

2005). Differences in species abundance distributions among the forest age classes were examined through rank abundance curves using logged relative abundance versus species rank.

To determine if there were differences between more commonly collected species compared to less frequently collected species, each species was classified into rarity categories based on the proportion of their abundance to the overall carabid beetle catch

(relative abundance) for the entire sampling period (all sampling years combined). The activity-abundance of Carabidae can be measured at several scales and a threshold between commonly collected and rare species can be set based on the data. Species were classified as ‘dominant’ if their relative abundance contributed more than 10.0% of the total number of individuals with all forest ages combined. Species collected in higher numbers (‘common’) contributed 1.0 – 9.99% while less frequently sampled species

(‘rare’) contributed less than 1.0% to the total number of individuals collected. Changes in the number of individuals and the number of species among forest age classes for these three rarity categories were analyzed using GLMMs. To determine if the distribution of common and rare species was random among Carabidae tribes, expected and observed values were determined and compared by Chi-square test. Dominant and common species were combined for this analysis and referred to as ‘abundant’. The expected number of commonly collected and rare species was determined by multiplying the total number of species collected for each tribe by the proportion of species in each rarity category (e.g.,

58 of 76 species were rare, 76.32%).

Differences in species compositions among the forest age classes and among forest age classes both within and among sampling years were analyzed using Bray-

Curtis dissimilarity and non-metric multidimensional scaling (NMDS). NMDS is a robust unconstrained ordination method which does not make assumptions about the underlying nature of species distributions (Minchin 1987; Clarke 1993; McCune & Grace 2002;

Oksanen 2015). All species were included in the analyses, with raw abundances standardized through relative abundance by sampling site (Faith et al. 1987; Minchin

1987) as recommended for Carabidae pitfall trap studies (Luff et al. 1992; Dufrene &

Legendre 1997; Kotze et al. 2011). To ensure solution stability, NMDS was performed on the dissimilarity matrix (Bray-Curtis or Jaccard) using 100 runs with random start points

(Minchin 1987). Statistical significances were determined with adonis, a robust permutational multivariate analysis of variance (using 999 permutations) (Legendre &

Anderson 1999; Anderson 2001; McArdle & Anderson 2001; Oksanen 2015). To test for multivariate homogeneity of dispersions within forest types we used R function betadisper and post-hoc test, TukeyHSD (Anderson 2006; Anderson et al. 2006; Oksanen

2015). All community-based analyses were completed using R package ‘vegan’ and

‘MASS’ (Venables & Ripley 2002; R Core Team 2016; Oksanen et al. 2016).

To provide additional insight into differences in species assemblages among the four forest age classes and determine significant species associations with forest age classes, point-biserial group-equalized phi coefficients (De Cáceres & Legendre 2009) were calculated. Correlation indices were used to determine the species ecological preferences among a set of alternative site groups for each species individually. This statistic considers absences and presences in the determination of site groups or site group combinations (De Cáceres & Legendre 2009; De Cáceres et al. 2010, 2012;

Legendre & Legendre 2012). Species included in the analysis had an association index value of at least 0.4 (De Cáceres 2013). Analysis was completed using 999 permutations with function multipatt in R package ‘Indicspecies’ (De Cáceres et al. 2010). Holm correction for multiple testing was used for experiment-wide conclusions (Holm 1979;

De Cáceres & Legendre 2009). Species were classified by habitat preferences based on ground conditions: open, shaded or either, as described in the literature (Lindroth 1961-

1969; Erwin 1981; Larochelle & Larivière 2003). Numbers of individuals and species were compared among forest age classes using linear models.

In order to examine spatial autocorrelation a multivariate Mantel correlogram was constructed (Mantel 1967; Oden & Sokal 1986; Sokal 1986; Borcard & Legendre 2012;

Legendre & Legendre 2012). Bray-Curtis dissimilarity values for species assemblages and the Cartesian coordinates of sampling sites were used in the analysis. Species data was Hellinger-transformed and detrended prior to analysis to satisfy normality and the second-order stationarity condition (Borcard et al. 2011). Sturge’s rule was used to determine the number of distance classes (Borcard et al. 2011). Following 5000 permutations the significance of the normalized Mantel statistic for the distance classes were evaluated using Holm correction for multiple testing (Holm 1979; Borcard et al.

2011; Goslee & Urban 2013). Autocorrelation analyses were completed in ‘vegan’

(Oksanen et al. 2016) and Cartesian coordinates using R package ‘SoDA’ (Chambers

2013; R Core Team 2016).

RESULTS

Numbers of individuals, richness and diversity

A total of 6,548 Carabidae representing 16 tribes, 33 genera and 76 species were collected (Table 1 & Appendix C, Table C1). There were no differences in mean values among forest age classes for number of individuals, species number or the inverse

Simpson concentration but there was significant interaction between forest age and

30 sampling year (χ2 = 16.9, df = 7, p = 0.018). At the rarefied sample size (n = 1,352) the lack of overlap in the unconditional 95% confidence intervals indicated that the zero year forests were significantly more species rich than the older age classes (Figure 2).

Extrapolated rarefaction curves suggest more species will be collected with increased sampling for all forest age classes except for ten year forests. Averaging the three nonparametric diversity estimators yielded the following values for sample completeness among forest age classes: zero (65.9%), ten (90.8%), fifty (80.6%), and ninety (71.9%).

The nonparametric estimators also indicate that the zero and ninety age classes have the highest number of undetected species (lowest sample completeness) and the highest estimated true species richness. In comparison, relatively few species remain undetected for ten and fifty year forests (Table 1). The three estimators have not completely stabilized and most estimates have relatively large standard errors. The ten and ninety year forests had significantly lower mean Chao1 values than the zero age class (χ2 = 16.1, df = 3, p = 0.001; Figure 3).

The abundance of five Carabidae tribes changed among forest age classes

(Appendix G, Figure G1). All significant differences were between the zero year age class and a combination of the older three forest age classes. In the zero year age class, there were fewer Cicindelini compared to ninety year forests (χ2 = 8.8, df = 3, p = 0.032), fewer Cychrini compared to the ten and ninety year age class (χ2 = 21.3, df = 3, p <

0.001), fewer Licinini versus fifty year forests (χ2 = 12.3, df = 3, p = 0.006), more

Harpalini compared to the three older forest age classes (H = 23.6, df = 3, p < 0.001) and more Scaritini compared to ten year old forests (H = 8.02, df = 3, p < 0.046). There were no significant differences in number of species among forest age classes for Carabidae

31 tribes, except for Harpalini which had more species in the zero year class compared to the three older forest age classes (H = 25.6, df = 3, p < 0.001).

Species assemblages with forest age

Of the 76 species collected, 34% (S = 26) were present in all four age classes and

49% (S = 37) were present in three of four forest age classes (Appendix C, Table C2).

The three most abundant species, Cyclotrachelus sigillatus (16.3%), Sphaeroderus stenostomus (14.9%), and Pterostichus sculptus (13.4%), accounted for 44.6% of the total number of individuals. These three species, which were classified as dominant species

(3.9% of the total S), occurred in all forest age classes, with at least two of the three species among the three most commonly collected species for each age class (Appendix

C, Table C2). Fifteen species (19.7% of the total S) were classified as common and 58 species were classified as rare (76.3% of the total S), with common species accounting for 44.0% of the total number of individuals collected whereas rare species accounted for only 11.3%. Although Agonum punctiforme was classified as common based on relative abundance, it only occurred at four sampling sites over the entire sampling period and all four sites were in the zero age class. In contrast, all other dominant and common species occurred at 13 or more sample sites. Four rare species occurred at 13 sampling sites or more with their overall relative abundances ranging between 0.43% – 0.67%. Twenty one species were represented by one individual (singletons), with the greatest number of singletons occurring in the zero year age class (n = 10) and ninety year forests (n = 7)

(see Appendix C, Table C2). There were no differences in mean numbers of individuals or numbers of species among forest age classes for dominant or common species.

However, there was a significantly higher mean number of rare individuals (χ2 = 46.9, p

< 0.001) collected from the zero year age class compared to the three older forest age classes (Appendix H, Figure H1A) and a significantly higher mean number of species (χ2

= 15.0, p = 0.002) in the zero year class compared to ten and ninety year forests

(Appendix H, Figure H1B).

Eighteen species from nine Carabidae tribes were considered abundant (i.e. consisting of dominant and common species). The distribution of abundant species within the nine tribes was significantly different than the expected numbers of species (chi- square χ2 = 18.0, df = 8, p < 0.025; Appendix H, Table H1). The distributions of the 58 rare species within 14 tribes were not different than expected.

The highest proportion of rare species occurred in the zero year age class (70.7%), followed by the ninety (53.4%), fifty (43.1%) and ten (39.7) year old forest age classes.

Of the total numbers of species collected in each forest age class, 53.9% were classified as rare for the zero year age class compared to 40.8% in ninety, 32.9% in fifty and 30.3% in ten year old age classes. The rank abundance curves (Figure 4) highlight the high abundance of rare species (ranks 15 – 43) for the zero year age class compared to older forest age classes, as well as the higher overall species richness, as indicated by the increased length of the zero year rank abundance curve. The rank abundance curves also suggest that species abundance distributions were similar among the three older forest age classes. The slope of the zero year age class was lower, indicating higher evenness for the zero year age class (Figure 4), which is primarily due to higher abundance of less frequently collected species. However, the mean Simpson’s evenness index was lowest for the zero year sites (0.39) and highest for fifty year forests (0.45), but the differences among forest age classes were not significant.

Ordination analysis with all sampling years combined (Figure 5) showed the greatest separation along axis 1, with no overlap of the zero year polygon with any of the other forest age polygons and with the highest degree of overlap for the ten year forest polygon with the fifty and ninety year polygons (k = 2; stress = 17.0). Zero year forest sites were negatively associated with axis 1 and had the most distinct species composition compared to other forest age classes. The minimal overlap of the fifty and ninety year polygons suggests that these forests maintain different species assemblages, with neither of these species assemblages different from that of ten year forests. Statistical analysis confirmed there were significant differences in species composition among forest age classes (F = 4.0, df = 3, p < 0.001). Ordination analysis based on presence/absence data, showed that the zero year polygon had little overlap with the polygons of the older three forest age classes (Appendix F, Table F1 & Figure F1). There is a higher degree of overlap in the polygons for the fifty and ninety year forests in the presence/absence

NMDS, with approximately half of the sites from each of the two age classes occurring close together in the ordination biplot. The multivariate homogeneity of group dispersions test among forest age classes was not significantly different (F = 2.6, df = 3, p

= 0.07). Bray-Curtis dissimilarity index values were relatively high, with no value below

0.49. The mean within forest age Bray-Curtis dissimilarity value (0.57) was lower than the mean value for among age classes (0.68). The zero age class had the highest within and among group dissimilarity (Table 2) as reflected in Figure 5. The zero and ten year polygons had the largest areas (area: zero = 1.47, ten = 0.85) while fifty and ninety year polygons were more compact (fifty = 0.27, ninety = 0.32). Using Jaccard dissimilarity index (presence/absence), there were minimal differences for among and within

34 dissimilarity values (mean = 0.027) compared to the abundance-based Bray-Curtis dissimilarity values (Appendix F, Table F1). Abundance had the largest effect on within group dissimilarity (0.076) for ninety year forest age class and between group dissimilarity (0.078) for the zero and ten year forest age classes.

Ordination analysis with all years combined were conducted without the extra year sites (only the first three sampling years included) and the species assemblages for the zero year sites were still the most distinct (Appendix E, Figure E1). Although the area of the zero year polygon without the extra year sites is smaller than the zero year polygon with the extra year sites included, it still has the largest area compared to other age class polygons (Figure E1) and has the highest within group Bray-Curtis dissimilarity value

(Table E1). For the ten year old forest polygon, the area is greatly reduced compared to when extra year sites are included in the analysis due to the removal of the outlying site

(TenBB). When the extra year sites are excluded, the within class Bray-Curtis dissimilarity value is lower (0.48) than when extra year sites were included, and the ten year old age class has the lowest within class Bray-Curtis dissimilarity value for any forest age class (Table E1). The among group dissimilarity values for ten versus fifty and ninety year forests are almost equal to when the extra year sites were included (Table E1).

Additionally, the ten year old forest polygon only overlaps with the fifty year forest polygon rather than overlapping with both the ninety and fifty year old forest polygons

(Figure E1).

Thirty-five species were classified as open habitat species, 36 species as shaded habitat species with the remaining having more general habitat preferences (based on the literature). Significantly more individuals preferring open habitats were collected from

35 zero year forests compared to ten and ninety year forests and there were significantly more open habitat individuals collected in fifty year forests compared to ninety (χ2 = 34.4, df = 3, p < 0.001; Figure 6A). A significantly higher number of species preferring open conditions were collected from zero year sites compared to older forest age classes and the number of open habitat species in fifty year forests was also higher than in ninety year forests (χ2 = 47.6, df = 3, p < 0.001; Figure 6B). There were significantly fewer individuals which preferred shaded conditions collected in zero year forests compared to fifty year forests (χ2 = 10.3, df = 3, p = 0.016; Figure 6A). Zero year forests also had significantly lower number of species which preferred shaded habitat species than ten, fifty and ninety year forests (χ2 = 20.2, df = 3, p < 0.001; Figure 6B). All of the dominant species in this study were shaded habitat species. For the 35 open habitat species, five were common species (14.3%) and 30 were rare (85.7%). For the 36 shaded habitat species, three were dominant species (8.3%), nine were common species (37.5%) and 24 were rare species (66.7%).

Several species with significant associations for a particular forest age class or a combination of forest age classes were detected in the habitat association analysis (Table

3). The highest number of species showed associations with the zero year forests. No species showed a significant association between zero year forest and any combination of older forest age classes (e.g. zero + ten + fifty). The majority of species showing preference for zero year forest were members of the Harpalini and Zabrini tribes, whereas species significantly associated with ten, fifty and ninety year forests were members of the Pterostichini, Licinini, Galertini, Platynini and Cychrini tribes (Appendix C, Table

C1). A few species were significantly associated with a combination of ten, fifty, and

36 ninety year forest but there was no significant association with the combination of fifty + ninety year forests only.

Year-to-year variation

The mean number of individuals in year three was significantly less than in years one or two (χ2 = 29.1, df = 2, p < 0.001; Figure 7A). In year three, the number of individuals collected was approximately half that collected during previous years (Table

4). Mean species number was also significantly lower in year three compared to years one and two (χ2 = 22.1, df = 2, p < 0.001; Figure 7B) and there was a significant interaction between forest age and year (χ2 = 15.2, df = 6, p = 0.019). The mean Chao1 value was significantly lower for year three than for years one and two (χ2 = 14.3, df = 3, p = 0.003; Figure 3B). However, there were no significant differences in 1/D values among sampling years and rarefaction analysis of sampling years showed no significant differences in cumulative richness at the rarefied number of individuals (n = 1,207).

Numbers of individuals for five carabid tribes varied significantly with sampling year (Appendix G, Figure G2). Fewer Cychrini were collected in year three compared to year one (χ2 = 6.6, df = 2, p = 0.036), more Licinini in year two compared to one and three (χ2 = 8.6, df = 2, p = 0.013). Platynini differed among all years with the highest number collected in year one (χ2 = 28.4, df = 2, p < 0.001), while fewer Pterostichini were collected in year three (χ2 = 37.2, df = 2, < 0.001) and there were fewer Zabrini individuals collected in year two compared to years one and three (χ2 = 10.1, df = 2, p =

0.006). There were no differences in number of species for any Carabidae tribes with sampling year.

Species assemblages with sampling year

Three species, C. sigillatus, S. stenostomus and P. sculptus, were dominant in all three years, although their dominance rank differed by sampling year (Appendix C, Table

C3). The number individuals collected for the three dominant species was significantly higher in years one and two compared to year three (χ2 = 42.4, df = 2, p < 0.001), and for common species, a lower number of individuals were collected in year three (χ2 = 16.1, df

= 2, p < 0.001). The number of common species collected in year three was also lower than years one and two (χ2 = 19.5, df = 2, p < 0.001) but there were no differences in species number for rare species. Number of dominant species was not analyzed since there were only three species in this category.

Mean Simpson’s evenness values indicated that species assemblages were significantly more even in year three compared to years one and two (χ2 = 18.6, df = 2, p

< 0.001; Table 4). Community analyses of annual variations within forest age classes indicated the zero year sites had the highest mean Bray-Curtis dissimilarities values within sampling year (0.67) as well as among sampling years (0.66). Mean Bray-Curtis dissimilarity values (abundance-based) within sampling year for the three older forest age classes ranged from 0.56 to 0.58 and from 0.55 to 0.60 among sampling years (Appendix

D, Table D1). Jaccard dissimilarity values (presence/absence data) and Bray-Curtis values were similar (mean differences between Jaccard and Bray-Curtis values: zero =

0.07, ten = 0.11, fifty = 0.06, ninety = 0.06), although dissimilarities were higher overall using presence/absence data (Appendix F, Table F2). Ordination analyses of annual variations within forest age classes indicated that there were differences in species composition among years (Figure 8A-D). Changes in the shape and position of sampling

38 year polygons suggest that highest year-to-year variation occurred in the zero and fifty year forests, and that the area of the ninety year forest polygon for sampling year three was more than double (1.79) that of year one (0.76) and year two (0.61). Ten year forest polygon had the most consistent position in the NMDS biplot among sampling years.

Permutational multivariate analysis, with sampling site as a blocked factor, indicated that there were significant variations in yearly species composition for zero (F = 0.86, df = 2, p = 0.018), fifty (F = 1.1, df = 2, p = 0.004) and ninety year forests (F = 1.37, df = 2, p =

0.002). The homogeneity of dispersion values was insignificant for all year-to-year ordination analyses. Ordination analyses using presence/absence data for within age yearly variation show similar patterns to abundance-based data (Appendix F, Figure F2).

The ordination figures and differences in the dissimilarity indices suggest that using abundance-based data had the strongest effect on the ten year old forest age class followed by the zero year age class although overall differences were minimal.

Ordination analyses depicting year-to-year variation among forest age classes

(Figure 9A-C) showed that a general pattern is present among sampling years. There were significant variations in species compositions among forest age classes for all three sample years: year one (stress = 17.8, k =2; F = 4.0, df = 3, p = 0.001), year two (stress =

17.4, k = 2; F = 3.6, df = 3 p = 0.001) and year three (stress = 20.4, k = 2; F = 3.0, df = 2, p = 0.001). For all years, the zero year site polygons are more negatively correlated with axis 1 and axis 2 compared to older forest age polygons. Among sampling years, the center of the ninety year forest polygon changed its position in ordination space the most relative to other age polygons. For the first sampling year, the zero year polygon did not overlap with the polygons of the three older forest age classes (Figure 9A). This is

39 reflected in the high mean Bray-Curtis dissimilarity value (among groups: 0.81) for the zero age class compared to the three older forest classes (Appendix D, Table D2). For sampling years two and three, the zero age polygon slightly overlapped the polygons of older forest age classes (Figures 9B & 9C) and the mean dissimilarity values among age classes were somewhat lower than it was for year one, especially for year two (see

Appendix D, Table D2). The Jaccard dissimilarity indices among forest age classes is lower (except in one instance) in all three sampling years compared to Bray-Curtis dissimilarity (Appendix F, Table F3) although the values were relatively similar (mean differences between Jaccard and Bray-Curtis values: year one = 0.12, year two = 0.13, year three = 0.09). Ordination analyses using presence/absence data show similar patterns to abundance-based data (Appendix F, Figure F3). The sizes of some polygons decrease slightly, especially for sampling year three, which has the highest overlap of the ten, fifty and ninety year polygons. The zero year sites show overlap in each sampling year for presence/absence data but remain distinct in each sampling year.

Spatial autocorrelation analysis indicated that there was a significant positive correlation at the smallest distance class (0 - 0.0169 km) and also a significant negative correlation for a middle distance class (9.217 - 13.818 km). Since no two study sites were within the smallest distance class, these sites can be considered spatially independent.

DISCUSSION

Numbers of individuals, richness and diversity

Previous studies in boreal forests (Niemelä et al. 1993; Spence et al. 1996;

Butterfield 1997) and European temperate forests (Magura et al. 2015) have reported

40 higher Carabidae abundance for recently clear-cut areas. In this study, abundance was highest for clear-cuts sites although the differences were not statistically significant, as was also reported by Koivula et al. (2002). There were also no significant differences in mean species richness among forest age classes for this study, as was also reported by

Lenski (1982) for sites in western North Carolina. However, rarefaction curves for the current study indicate that rarefied cumulative richness values of the zero year sites (i.e. clear-cuts) were significantly higher than the older age classes (Chao et al. 2014), supporting my prediction of higher richness in the early phases after a clear-cutting event

(Niemelä et al. 1993; Spence et al. 1996; Butterfield 1997; Jukes et al. 2001; Koivula et al. 2002; Pearce et al. 2003; de Warnaffe & Lebrun 2004; Magura et al. 2015). In addition, all three of nonparametric estimators indicate that the zero year age class has higher true richness than the three older forests classes. In other studies, forest floor invertebrates, such as hunting spiders (Pajunen et al. 1995; Niemelä et al. 1996; Buddle et al. 2000; Gallé et al. 2014) and ants (Punttila et al. 1991) have shown higher richness after clear-cutting or during the earlier stages of forest recovery. In contrast, greater species richness has been reported in more mature forests for saproxylic beetles

(Lassauce et al. 2013), other Coleoptera (Southwood et al. 1979; Johansson et al. 2016), web-building spiders (Coyle 1981; Pajunen et al. 1995; Buddle et al. 2000) and millipedes (Magura et al. 2015). Abundance and species richness of small mammals and breeding birds have increased after clear-cutting in North American temperate forests

(Kirkland 1990; Keller et al. 2003), and the species richness of trees and herbaceous plants typically increases through succession in the study region (Southwood et al. 1979;

Elliott et al. 1997; Peet et al. 2014).

The results of this study suggest that the higher richness of zero age forests is due to the presence of carabid species preferring open conditions, combined with the temporary persistence of shaded habitat species. Elevated richness of recent clear-cuts owing to the combined presence of open habitat and forest species has been reported in other studies (Niemelä et al. 1993; Buddle et al. 2006). Forest carabid species can persist in an area for up to two to three years after clear-cutting (Niemelä et al. 1993; Spence et al. 1996; Koivula et al. 2002; Jacobs et al. 2008). Open habitat species tended to be rare with over 85% of the open habitat species were classified as rare. The higher cumulative richness of the zero year age class was increased by the high number of rare species sampled from this age class. Recently disturbed habitats typically have an increased chance of stochastic events and higher frequency of vagrant species which may increase the likelihood of rarer species (Niemelä et al. 1996). The higher abundance and species number of open habitat species in the zero year sites compared to the three older age classes is consistent with results from other studies (Szyszko 1983; Niemelä et al. 1993;

Koivula et al. 2002; Pearce et al. 2003; Karen et al. 2008; Taboada et al. 2008; Magura et al. 2015). The numbers of open habitat carabid individuals and species drop drastically by year ten and remain low in fifty and ninety year forests. These carabid species, the majority of which are macropterous (flight capable), are able to colonize clear-cut areas rapidly and decline soon after forest canopy closure (Haila et al. 1994; Spence et al. 1996;

Butterfield 1997; Jukes et al. 2001). In this study there were slightly more open habitat individuals and species in fifty year forests compared to ten and ninety year forest sites.

This may be due to the fact that the majority for the fifty year forest age sites occur within the PMSP Corridor trail, adjacent to agricultural fields/residential areas and are

42 therefore subjected to the effects of fragmentation. Open habitat species are able to invade several meters into the forest (Spence et al. 1996; Magura et al. 2001). Clear-cut areas typically have higher abundances of herbaceous plants, providing seeds and other vegetal matter which are important for diets of many open habitat carabid species

(Covington 1981; Niemelä et al. 1996; Pearce et al. 2003; Karen et al. 2008). In addition, natural or anthropogenic disturbance and increased habitat heterogeneity can decrease dominance within Carabidae communities (Levin & Paine 1974; Connell 1978; Liebherr

& Mahar 1979; Lenski 1982; Karen et al. 2008).

Species composition among forest age classes

In temperate and boreal forests, carabid communities typically have a small number of dominant species which account for a large number of the total individuals while the majority of species are rare (Belaoussoff et al. 2003; Karen et al. 2008;

Schirmel et al. 2011; Wang et al. 2014). The rarity classifications and rank abundance distributions of this study concur with these findings. The rank abundance curves suggest the species assemblages of the zero year age class were more even than the older forest age classes while Simpson’s evenness index indicated the opposite trend, with recent clear-cuts having the lowest evenness among forest age classes. This is most likely due to the omission of variation among study sites and sampling years in the rank abundance curves as they were created using cumulative data (all sampling years and study sites combined). When these variations are incorporated into the analysis the evenness of recent clear-cuts is lower than it appears in the rank abundance curve and the differences among forest age classes are not significant.

Species classified as abundant were statistically more likely to be in the Licinini and Pterostichini while fewer Harpalini species than expected were abundant. Although

Harpalini was by far the most species rich tribe, only one of the 22 Harpalini species collected was common while the remainder were rare, with the majority of these rare species collected from recent clear-cuts. Rare species play a role in the distinctness of community assemblages among the forest age classes, particularly for the zero year age class, where more than 50% of the species are rare. Higher numbers of rare species in recent clear-cuts and ninety year old forests indicate that these forest age classes are important for maintaining Carabidae diversity at the landscape level. Abundant species were also non-randomly distributed into the aforementioned tribes, suggesting a predictable oligarchy exists for carabid beetles within temperate Piedmont forests, such as that found for Amazonian trees by Pitman et al. (2001). These results suggest that pitfall sampling in Piedmont forests will likely uncover a predictable set of species from the aforementioned tribes.

Community analyses indicated there were differences in carabid species assemblages among the four forest age classes, with the most distinct differences between the zero year forests and the older age classes, likely relating to canopy closure. These patterns are similar to those reported for northern higher latitude forests (Szyszko 1990;

Niemelä et al. 1993, 1996; Haila et al. 1994; Butterfield 1997; Ings & Hartley 1999;

Koivula et al. 2002; Magura et al. 2006). Other ground dwelling invertebrates that have shown distinct species assemblages between recent clear-cuts and more mature forests include saproxylic beetles (Kaila et al. 1997; Lassauce et al. 2013), spiders (Niemelä et al.

1996), millipedes (Magura et al. 2015) and ants (Niemelä et al. 1996; Gallé et al. 2016).

In U.S. temperate forests, small mammals and birds also show distinct community differences between recently clear-cut areas and mature aged forests (Kirkland 1990;

Keller et al. 2003). Plant communities in this region also form distinct species assemblages with changes in forest age (Oosting 1942; Peet & Christensen 1980;

Christensen & Peet 1981).

Based on rank abundance curves and the ordination analysis, the most pronounced difference in species assemblages was between the zero year forests versus older forest age classes, primarily due to the greater number of less frequently collected rare species.

In the ordination analysis, the zero year polygon had little overlap with other age polygons when using abundance-based or presence/absence dissimilarly indices, suggesting that the uniqueness of the zero year age polygon is chiefly due to differences in species presence. Both ordination methods suggest that differences between the fifty and ninety year forest age classes are due to a combination of differences in species presence as well as species abundances for fifty and ninety year forest sites. Based only on species presence data, approximately one half of the ninety year polygon and three quarters of the fifty year age polygons overlap with each other compared to the more distinct separation between polygons when abundance is incorporated. When all data is combined, fifty and ninety year forests share 52% of the 59 species occurring in one or both of these two forest age classes. Overall, for all age classes the relatively similar values for the Bray-Curtis and Jaccard indices indicate that the strongest effect on community compositions within and among age classes are the presence/absence of species rather than differences in abundance. However, differences in species abundances also contributed to the distinctiveness of each of the forest age classes. The disparities

45 between Bray-Curtis and Jaccard dissimilarities were greatest for zero and ten year old forest age classes which may relate to the varying abundances of the Cychrini, Harpalini,

Cicindelini and Scaritini tribes.

The ten year forest ordination polygon had the second largest area, primarily due to the negative score on axis 2 for one study site (TenBB) in the ordination analysis. The ten year forest polygon encompassed the majority of the fifty and ninety year forest sites, suggesting that ten year forests harbored different species assemblages from the zero year sites and that the ten year forests’ species assemblages overlapped with those of the fifty and ninety year forests. Based on presence/absence data, most polygons were similar in area to their abundance-based counterparts, with the exception of the ten year age polygon which is approximately half the area based on presence/absence data. This difference is mainly due to the placement of one site, TenBB, suggesting a large role for abundance at this site.

The large variation among zero year sites for the ordination analysis suggests that the high cumulative richness of this forest age class is not primarily due to higher alpha diversity at individual study sites but is also influenced by higher levels of beta diversity among sites. This is supported by the high cumulative carabid richness of the zero year age classes when the study sites were pooled by forest age and may also at least partially explain the lack of significance in mean richness among forest age classes. Compared to mature forests, previous studies in boreal forests measuring carabid richness have also reported higher levels of beta diversity in recently clear-cut areas (Haila et al. 1994;

Niemelä et al. 1996), as well as for younger forest age classes in pine plantations and boreal forests (Niemelä et al. 1996; Taboada et al. 2008; Johansson et al. 2016). Higher

46 variation in carabid assemblages among sites may be correlated with increased heterogeneity in habitat cover for recent clear-cut sites, due to factors such as newly established localized vegetation, woody debris, and greater spatial variation in abiotic conditions including temperature, light, wind and humidity (Covington 1981; Holt et al.

1999; Pearce et al. 2003). The extra year of sampling for the zero and ten year age classes increased the geographic distance among sites compared to fifty and ninety year sites, as well as increased the temporal variability, both of which added to differences in species composition among sites within these age classes. When ordination analysis with the first three sampling years combined were conducted without the extra year sites, community patterns were similar for the zero year sites, including the unique species assemblages and the high beta diversity among sites compared to the older forest age classes. In comparison, removing the extra year sites does affect community-level interpretations for the ten year old forest. The polygon area is greatly reduced compared to when extra year sites are included, reflecting the lower within age group dissimilarity value. The relationship with ninety year old forests changed somewhat in the ordination analysis but the main influence of the extra year sites for ten year old forests was increased within group variability which was mainly owing to an outlying sampling site (TenBB). When the extra year sites are excluded, interpretation of the community analyses for the zero year sites remains mostly unchanged; however, interpretation for the ten year old forest sites does change, but predominantly due to the exclusion of the aforementioned outlying site (TenBB). Another extra year site for ten year old forest, TenPZ, produced similar results as the original ten year old sites, suggesting that site conditions of TenBB may differ from other ten year old sites.

Habitat association analysis indicated that the highest number of significant species associations were for zero year sites compared to other age classes. All but one of the species with a significant association with the zero year class have been reported to prefer open habitats, the exception being Myas coracinus, which has previously been reported to prefer forested habitats in eastern North America (Erwin 1981; Duchesne et al.

1999; Larochelle & Larivière 2003; Jennings & Tallamy 2006; Pearce & Venier 2006).

In this study, the majority of M. coracinus were collected from zero year forests, with only a few individuals collected from ten year forest sites and it was absent from fifty and ninety year forests. The results from this study are more consistent with those from western Canada and South Dakota where M. coracinus has been reported to be restricted to open habitats (Barlow 1970; Niemelä et al. 1992b). This species likely inhabits a variety of habitats which may differ with geographic location as habitats preferences for specific species can differ among regions (Erwin 1981; Koivula 2011). In this study, the majority of species associated with the zero year age class are from genera known to include open habitat species (e.g. Amara, Notiobia, and Harpalus). Many species in these genera feed on plant material, are good colonizers (Lindroth 1961-1969; Thiele 1977) and have been reported to occur in high numbers in clear-cuts areas (Lenski 1982;

Niemelä et al. 1993; Spence et al. 1996; Butterfield 1997; Koivula et al. 2002).

According to tribal-level analyses, the higher abundance and species number for

Harpalini in the zero year age class had an effect on the difference between this age class and the older three age classes. Although previous studies have reported Pterostichus adoxus typically occurs in mature forests (MacLean & Usis 1992; Pearce & Venier 2006;

Chavez & Browne 2011), it was significantly associated with ten year forests in this

48 study. Lenski (1982) reported that it was the second most abundant species for clear-cut sites, providing additional evidence that P. adoxus is a generalist species.

The smaller area of the ordination polygons for the fifty and ninety year forest sites and the lower dissimilarity values suggest lower site to site variation within these age classes relative to zero and ten year forest sites. The minimal overlap of abundance- based ordination polygons and a 58% – 59% dissimilarity values suggests that although fifty and ninety year forests share over half of their species they are relatively distinct assemblages. After the forest canopy has closed, open habitat carabid species start to decline and forest species requiring environmentally specific ground conditions become established (Niemelä et al 1996; Ings and Hartley 1999). This transition from open habitat to forest species can lower site-to-site variation for Carabidae species assemblages of older forests compared to younger forests (Niemelä et al. 1996; Spence et al. 1996;

Butterfield 1997; Koivula et al. 2002; Taboada et al. 2008). The lower site-to-site variation of the more mature forests relative to zero year forest sites may be due to the decline of vagile open habitat species, combined with the increased heterogeneity of microhabitats within clear-cuts and the decreased chance of stochastic events in older forests (Niemelä et al. 1996).

The number of shaded habitat individuals, as well as number of shaded habitat species, increased from zero year forests to fifty and ninety year forests. There is a pronounced community-level change between the zero year forest age class and the three older forest age classes which is likely due to the drop in open habitat individuals and species. Carabid beetle communities from older forests have been found to be more predictable than those of clear-cuts. This is typically attributed to the increased stability

49 of more mature forest habitat relative to recently disturbed habitats, resulting in viable populations across multiple generations (Darlington 1943; den Boer et al. 1980; Gutierrez

& Menendez 1997). Carabids preferring older forest habitats are typically flightless, which has been attributed to habitat stability (Darlington 1943; den Boer et al. 1980;

Riley & Browne 2011).

Differences in species assemblages between the zero year class and the older forest age classes were affected by the higher number of individuals from the Cychrini,

Harpalini and Cicindelini. Tribal analyses indicated that there are significant differences in the abundance of Cychrini among forest age classes, primarily due to the dominant species S. stenostomus. Cychrini abundance also plays a role in the differences between the zero year age class and both the ten and ninety year forests and, to a lesser extent, between the fifty year forests and ten and ninety year forests. Differences in abundance of

Licinini also likely affected the species composition between zero year age class and fifty year forests and to a lesser degree between ten and ninety year forests.

In this study, carabid beetles demonstrated species specific responses to forest age, which may be linked to variations in specific habitat requirements for survival and reproduction (Niemelä et al. 1996). Species which demonstrated significant associations with ten, fifty and ninety year forest classes are known to prefer shaded ground and/or classified as forest species (Lindroth 1961-1969; Thiele 1977; Erwin 1981; Lenski 1982;

Purrington et al. 1989; Larochelle & Larivière 2003; Pearce et al. 2003). For example,

Dicaelus spp., Galerita bicolor, Platynus decentis, S. stenostomus, Pterostichus spp. showed significant associations with fifty and ninety year forests or a combination of these older forest age classes. More carabid species were significantly associated with

50 fifty year forests than ninety year forests, indicating that a higher number of species preferred middle aged forest than the more mature forests in this study. Pines are the dominant trees in most fifty year forests in North Carolina Piedmont, which can affect ground conditions considerably by lowering soil pH, altering litter composition, litter dimension and structure (Paje & Mossakowski 1984; Peet & Christensen 1987; Pearce et al. 2003). These results suggest that middle aged Piedmont forests maintained relatively distinct carabid assemblages compared to more mature forests which are usually dominated by deciduous trees. Due to extensive timber commercial operations in southeastern U.S. forests, areas are routinely harvested and subsequently replanted or regenerate, resulting in over 80% of forests being less than 60 years old (Pan et al. 2011).

S. stenostomus and Dicealus ambiguus were significantly associated with ten, fifty and ninety year forests, indicating that they may be forest generalists requiring a closed canopy but not a specific forest age. S. stenostomus has been listed as a forest species in other studies, with its occurrence associated with the presence of slugs and snails since it is a gastropod specialist feeder (Lindroth 1961-1969; Thiele 1977; Lenski 1982; Pearce &

Venier 2006; Browne et al. 2014).

Year-to-year variation

Sampling year had more of an effect on abundance and species richness than forest age; however, these differences are predominantly the result of the low number of individuals collected in year three (May 2011 – Apr. 2012). In most cases, years one and two had relatively similar species numbers and numbers of individuals. The decrease in abundance in sample year three is unlikely to be related to localized population depletions due to continuous sampling at the same location since other pitfall trap studies

51 which have repeatedly sampled for three or more years (Luff 1982; Niemelä et al. 1993;

Haila et al. 1994; Jukes et al. 2001; Günther & Assmann 2004; Harvey et al. 2008;

Kwiatkowski 2011) or even more than a decade (den Boer 1981) have not found long- term decreases in carabid abundance. The lower abundances could relate to below average annual precipitation and above average annual temperatures in 2010 – 2012

(State Climate Office 2016; Appendix A, Figure A2). Temperature and precipitation can affect carabid populations, either directly and/or indirectly (Haila et al. 1994; Desender

1996; Kotze & Niemela 2002). Weather conditions were near normal in the first year of sampling but the second and third years were interspersed with moderate drought conditions inlcuding record high temperatures and lower precipitation associated with the

La Niña ENSO cycle (NC DMAC 2010; NC DMAC 2011; NC DENR 2012). Many carabid species are sensitive to decreased soil moisture, which has been proposed as a dominant factor affecting species distributions (Thiele 1977; Luff et al. 1989; Epstein &

Kulman 1990; Gardner 1991; Niemelä et al. 1992a; Rykken et al. 1997; Karen et al.

2008). Egg, larva and pupa developmental stages are particularly susceptible to drier conditions (Thiele 1977; Lundgren et al. 2005).

Annual Carabidae pitfall trap catches have been reported to exhibit large year-to- year variations in abundance and species composition. Haila et al. (1994) found that over a six year sampling period carabid yearly abundances fluctuated six fold. Dominant species usually remain dominant across years, but the abundance of these species can vary widely among sampling years (Haila et al. 1994; Desender 1996; French & Elliott

1999). In this study, the same three species were dominant in all years; however, from year two to year three the number of individuals for these dominant species decreased

52 significantly between 34% – 68%. In addition to the three dominant species, the numbers of individuals collected from several tribes were also lower in sampling year three compared to years one and two, including Cychrini, Licinini and Pterostichini. Year-to- year changes in plant species richness, vegetation structure and habitat heterogeneity have been shown to be important factors in temporal variations in carabid abundance

(Brose 2003; Harvey et al. 2008). Relatively few long term studies have documented fluctuations and patterns of carabid populations (den Boer 1981; Luff 1982) therefore additional studies are needed, particularly for the U.S. (den Boer 1990),

Species assemblages with sampling year

The similarity in patterns between ordination analyses using abundance-based data and presence/absence data suggest that species presence among sampling sites is the main driver of the observed differences among forest age classes within sampling years.

Removing abundance from the ordination analyses had minimal effects on the community analyses. In this study, the zero age class showed the greatest within age year-to-year variation in carabid species composition and ten year forests showed the least. Previous studies have suggested that there is rapid community change for carabids during the first few years after clear-cutting (Szyszko 1983; Niemelä et al. 1993, 1996;

Haila et al. 1994; Butterfield 1997). At clear-cut sites, the ground cover varied spatially and changed markedly from year to year (pers. obs). Changes in habitat structure and plant communities are relatively rapid in the first years after a clear-cutting event, which in turn, are likely reflected in the carabid community composition (Oosting 1942; Horn

1980; Covington 1981; Seastedt & Crossley 1981; Niemelä et al. 1996; Elliott et al. 1997;

Buddle et al. 2006). Approximately a year after clear-cutting at the zero year sites in this

53 study, numerous species of tall, weedy vegetation, densely clustered shrubs and tree seedlings became established, as well as tree regrowth from stumps. This newly established vegetation was patchily distributed (pers. obs.), as is typical for regenerating forests in this region (Horn 1980; Seastedt & Crossley 1981; Peet & Christensen 1987).

Ground conditions were partially to mostly shade two to three years after clear-cutting, similar to the regeneration of southern Appalachian forests (Horn 1980; Seastedt &

Crossley 1981; Elliott et al. 1997). As ground conditions including canopy closure are pivotal for Carabidae species assemblages (Niemelä et al. 1993, 1996; Pearce & Venier

2006; Koivula 2011), these changes likely influenced the year-to-year variations in species compositions observed for the zero year age class. The ground of ten year old forests in this study was mostly if not completely shaded by dense juvenile tree cover during the growing season.

Ninety year forests had the second highest variation in species assemblages among sampling years, primarily due to the increased variability of sampling sites in year three, as reflected in the increased area of the year three ordination polygon. Carabids typically aggregate within forest microhabitats, causing them to be patchily distributed within forests (Greenslade 1964; Thiele 1977; Niemelä et al. 1992a, 1996). Changes in the year-to-year dominance structure as well as patterns and localities of these microhabitats may contribute to annual changes in carabid populations.

Many carabid population studies consist of a single year of sampling which may be further limited to only periods when carabids are active. In this study, pitfall trapping indicates that carabid activity was highest from May through October (79% of total catch) but individuals were collected throughout the year, with 21% of the catch between

November through April. In order to provide an accurate representation of species composition and abundance for this study area, sampling should occur over the entire year. The results from this study indicate that the amount of time before sampling is initiated following clear-cutting will affect estimates of community-level differences. The variation among individual years is captured in the overall ordination analysis (i.e.

Figures 5 & E1), with the forest age polygons at intermediate positions and area sizes for ordination figures when all sampling years are combined. The community structure of

Carabidae show a rapid response to clear-cutting, which is supported by the year one ordination analysis among forest age classes. In contrast, Niemelä et al. (1993) found more modest changes in the first collection season after a boreal forest clear-cutting event, with accelerated changes in the second year as forest specialist species disappear. There were significant differences among the forest age classes in all three sampling years, indicating that single year studies will likely capture the main patterns in carabid beetle species assemblages among different aged forests but overlook variations in species composition among years. Studies examining carabid responses to environmental changes will be more informative if they span multiple years since Carabidae responses will be time dependent and species assemblages will likely vary among years (Kotze and

Niemelä 2002).

In conclusion, this study confirmed there is a relationship between forest age and carabid beetles for temperate forests in the Southeastern U.S. Species richness and diversity were less variable with forest age compared to species assemblages. Initially, clear-cutting positively influenced species richness but as trees recolonize an area closing the forest canopy, species richness decreased and remains stable as the forest continues to

55 mature. Characteristic community assemblages for different aged forests are likely due to species-specific responses to clear-cutting and age specific environmental changes.

Variations among sampling year can affect the number of carabid individuals, species richness, diversity and community assemblage. Distinct differences in species assemblages among forest age classes indicate that conservation decisions need to be considered at the landscape scale.

ACKNOWLEDGMENTS

I thank Wake Forest University and the Environmental Program of Wake Forest

University for the financial support, Matt Windsor of Pilot Mountain State Park for permission to sample Carabidae, Terry L. Erwin, Ph.D. of the Smithsonian Institution for access to the Carabidae collection at the National Museum of Natural History, and the multitude of undergraduates who assisted with the monthly field collections. I especially thank Nick Riley for driving two hours to assist me in the field and enduring the discomfort of having wet feet for the majority of it.

LITERATURE CITED

Adis, J. 1979. Problems of interpreting arthropod sampling with pitfall traps.

Zoologischer Anzeiger 202:177–184.

Anderson, M. J. 2001. A new method for non-parametric multivariate analysis of

variance. Austral Ecology 26:32–46.

Anderson, M. J. 2006. Distance-based tests for homogeneity of multivariate dispersions.

Biometrics 62:245–53.

Anderson, M. J., K. E. Ellingsen, and B. H. McArdle. 2006. Multivariate dispersion as a

measure of beta diversity. Ecology Letters 9:683–693.

Baars, M. A. 1979. Catches in pitfall traps in relation to mean densities of carabid beetles.

Oecologia 41:25–46.

Barlow, C. A. 1970. Phenology and distribution of some Pterostichus (Coleoptera:

Carabidae) of eastern Canada. New York Entomological Society 78: 215-236.

Basualdo, C. V. 2011. Choosing the best non-parametric richness estimator for benthic

macroinvertebrates databases. Revista de al Sociedad Entomológica Argentina

70:27–38.

Bates, D., M. Maechler, B. M. Bolker, and S. Walker. 2015. Fitting linear mixed-effects

models using lme4. Journal of Statistical Software 67:1–48.

Bazzaz, F. A. 1979. The physiological ecology of plant succession. Annual Review of

Ecology and Systematics 10:351–371.

Belaoussoff, S., P. G. Kevan, S. Murphy, C. Swanton. 2003. Assessing tillage

disturbance on assemblages of ground beetles (Coleoptera: Carabidae) by using a

range of ecological indices. Biodiversity and Conservation 12:851–882.

Bolker, B. M., M. E. Brooks, C. J. Clark, S. W. Geange, J. R. Poulsen, M. H. H. Stevens,

and J. S. S. White. 2009. Generalized linear mixed models: a practical guide for

ecology and evolution. Trends in Ecology and Evolution 24:127–135.

Bonan, G. B. 2008. Forests and climate change: forcings, feedbacks, and the climate

benefits of forests. Science 320:1444–1449.

Borcard, D., F. Gillet, and P. Legendre. 2011. Numerical ecology in R. R. Gentleman, K.

Hornik, and G. Parmigiani, editors. Use R! Series. Springer Science, New York,

New York, USA.

Borcard, D., and P. Legendre. 2012. Is the Mantel correlogram powerful enough to be

useful in ecological analysis? A simulation study. Ecology 93:1473–1481.

Bousquet, Y. 2012. Catalogue of Geadephaga (Coleoptera, Adephaga) of America, north

of Mexico. ZooKeys 1722:1–1722.

Brose, U. 2003. Bottom-up control of carabid beetle communities in early successional

wetlands: mediated by vegetation structure or plant diversity? Oecologia 135:407–

413.

Brose, U., and N. D. Martinez. 2004. Estimating the richness of species with variable

mobility. Oikos 105:292–300.

Brose, U., N. Martinez, and R. J. Williams. 2003. Estimating species richness: Sensitivity

to sample coverage and insensitivity to spatial patterns. Ecology. 84:2364-2377.

Brown, M. J. 2015. Forests of North Carolina, 2013. Resource Update FS-47. U.S.

Department of Agriculture Forest Service, Southern Research Station, Asheville, NC.

Brown, P. M., and J. M. Parker. 1985. Geologic Map of North Carolina. North Carolina

Geological Society.

Browne, R. A., S. A. Maveety, L. Cooper, and K. N. Riley. 2014. Ground beetle

(Coleoptera: Carabidae) species composition in the Southern Appalachian

Mountains. Southeastern Naturalist 13:407–422.

Buddle, C. M., J. R. Spence, and D. W. Langor. 2000. Succession of boreal forest spider

assemblages following wildfire and harvesting. Ecography 23:424–436.

Buddle, C. M., D. W. Langor, G. R. Pohl, and J. R. Spence. 2006. Arthropod responses to

harvesting and wildfire: Implications for emulation of natural disturbance in forest

management. Biological Conservation 128:346–357.

Burnham, K. P., D. R. Anderson, and K. P. Huyvaert. 2011. AIC model selection and

multimodel inference in behavioral ecology: Some background, observations, and

comparisons. Behavioral Ecology and Sociobiology 65:23–35.

Burnham, K. P., and W. S. Overton. 1979. Robust estimation of population size when

capture probabilities vary among animals. Ecology 60:927–936.

Butterfield, J. 1997. Carabid community succession during the forestry cycle in conifer

plantations. Ecography 20:614–625.

Chambers, J. M. 2013. SoDA: functions and examples for “software for data analysis”. R

package version 1.0-6. Available from: https://CRAN.R-project.org/package=SoDA

Chao, A. 1984. Nonparametric Estimation of the number of classes in a population.

Scandinavian Journal of Statistics 11:265–270.

Chao, A., N. J. Gotelli, T. C. Hsieh, E. L. Sander, K. H. Ma, R. K. Colwell, and A. M.

Ellison. 2014. Rarefaction and extrapolation with Hill numbers: A framework for

sampling and estimation in species diversity studies. Ecological Monographs 84:45–

67.

Chao, A., and S. M. Lee. 1992. Estimating the number of classes via sample coverage.

Journal of the American Statistical Association 87:210–217.

Chao, A., and M. C. K. Yang. 1993. Stopping rules and estimation for recapture

debugging with unequal failure rates. Biometrika 80:193–201.

Chavez, C., and R. A. Browne. 2011. Carabidae (ground beetle) species composition of

Southern Appalachian spruce-fir forests. Southeastern Naturalist 10:591–608.

Chazdon, R. L., R. K. Colwell, J. S. Denslow, and M. R. Guariguata. 1998. Statistical

methods for estimating species richness of woody regeneration in primary and

secondary rain forests of Northeastern Costa Rica. Pages 285–309 in F. Dallmeier

and J. A. Comiskey, editors. Forest Biodiversity Research, Monitoring and

Modeling: Conceptual Background and Old World Case Studies Man and the

Biosphere Series. Vol. 20. Parthenon Publishing, Paris.

Chen, J., J. F. Franklin, and T. A. Spies. 1993. Contrasting microclimates among clearcut,

edge, and interior of old-growth Douglas-fir forest. Agricultural and Forest

Meteorology 63:219–237.

Christensen, N. L., and R. K. Peet. 1981. Secondary forest succession on the North

Carolina Piedmont. Pages 230–245 in D. West, H. Shugart, and D. Botkin, editors.

Forest Succession: Concepts and Application. Springer-Verlag, New York.

Christensen, N. L., and R. K. Peet. 1984. Convergence during secondary forest

succession. Journal of Ecology 72:25–36.

Ciegler, J. C. 2000. Ground beetles and wrinkled bark beetles of South Carolina

(Coleoptera: Geadephaga: Carabidae and Rhysodidae) Clemson University,

Clemson, SC.

Clarke, K. R. 1993. Non-parametric multivariate analyses of changes in community

structure. Australian Journal of Ecology 18:117–143.

Colwell, R. K. et al. 2012. Models and estimators linking individual-based and sample-

based rarefaction, extrapolation and comparison of assemblages. Journal of Plant

Ecology 5:3–21.

Colwell, R. K., C. X. Mao, and J. Chang. 2004. Interpolating, extrapolating, and

comparing incidence-based species accumulation curves. Ecology 85:2717–2727.

Connell, J. H. 1978. Diversity in tropical rain forests and coral reefs. Science 199:1302–

1310.

Covington, W. W. 1981. Changes in forest floor organic matter and nutrient content

following clear cutting in northern hardwoods. Ecology 62:41–48.

Coyle, F. A. 1981. Effects of clearcutting on the spider community of a southern

Appalachian forest. The Journal of Arachnology 9:285–298.

Cribari-Neto, F., and A. Zeileis. 2010. Beta regression in R. Journal of Statistical

Software 34:1-24.

Darlington, P. J. 1943. Carabidae of mountains and islands: data on the evolution of

isolated faunas, and on atrophy of wings. Ecological Monographs 13:37–61.

De Cáceres, M. 2013. How to use the indicspecies package (ver. 1.7.1). Available from:

https://cran.r-project.org/web/packages/indicspecies/vignettes/

indicspeciesTutorial.pdf

De Cáceres, M., and P. Legendre. 2009. Associations between species and groups of sites:

Indices and statistical inference. Ecology 90:3566–3574.

De Cáceres, M., P. Legendre, and M. Moretti. 2010. Improving indicator species analysis

by combining groups of sites. Oikos 119:1674–1684.

De Cáceres, M., P. Legendre, S. K. Wiser, and L. Brotons. 2012. Using species

combinations in indicator value analyses. Methods in Ecology and Evolution 3:973–

982. de Warnaffe, G. B., and P. Lebrun. 2004. Effects of forest management on carabid

beetles in Belgium: Implications for biodiversity conservation. Biological

Conservation 118:219–234. den Boer, P. J. 1981. On the survival of populations in a heterogeneous and variable

environment. Oecologia 50:39–53. den Boer, P. J. 1985. Fluctuations of density and survival of carabid populations.

Oecologia 67:322–330. den Boer, P. J. 1990. Density limits and survival of local populations in 64 carabid

species with different powers of dispersal. Journal of Evolutionary Biology 3:19–48. den Boer, P. J., T. H. P. van Huizen, W. den Boer-Daanje, B. Aukema, and C. F. M. den

Bieman. 1980. Wing polymorphphism and dimorphism in ground beetles as stages

in an evolutionary process (Coleoptera, Carabidae). Entomologia Generalis 6:107–

134.

Desender, K. R. C. 1996. Diversity and dynamics of coastal dune carabids. Annales

Zoologici Fennici 33:65–75.

Drummond, M. A., and T. R. Loveland. 2010. Land-use pressure and a transition to

forest-cover loss in the eastern United States. BioScience 60:286–298.

Dufrene, M., and P. Legendre. 1997. Species assemblages and indicator species: The

need for a flexible asymmetrical approach. Ecological Monographs 67:345–366.

Duchesne, L. C., R. A. Lautenschlager, F. W. Bell. 1999. Effects of clear-cutting and

plant competition control methods on carabid (Coleoptera: Carabidae) assemblages

in northwestern Ontario. Environmental Monitoring and Assessment 56:87–96.

Elliott, K. J., L. R. Boring, W. T. Swank, and B. R. Haines. 1997. Successional changes

in plant species diversity and composition after clearcutting a Southern Appalachian

watershed. Forest Ecology and Management 92:67–85.

Epstein, M. E., and H. M. Kulman. 1990. Habitat distribution and seasonal occurrence of

carabid beetles in east-central Minnesota. American Midland Naturalist 123:209–

225.

Erwin, T. L. 1981. Natural history of Plummers Island, Maryland. XXVI. The ground

beetles of a temperate forest site (Coleoptera: Carabidae): An analysis of fauna in

relation to size, habitat selection, vagility, seasonality, and extinction. Bulletin of the

Biological Society of Washington 5:105–224.

Erwin, T. L. 1985. The taxon pulse: A general pattern of lineage radiation and extinction

among carabid beetles. Pages 438–472 in G. E. Ball, editor. Taxonomy, Phylogeny

and Zoography of Beetles and Ants. Dr. W. Junk Publishers, Dordrecht.

Erwin, T. L. 1996. Biodiversity at its utmost: Tropical forest beetles. Pages 27–40 in M.

L. Reaka-kudla, D. E. Wilson, and E. O. Wilson, editors. Biodiversity II:

Understanding and Protecting our Biological Resources. Joseph Henry Press,

Washington, DC.

Faith, D. P., P. R. Minchin, and L. Belbin. 1987. Compositional dissimilarity as a robust

measure of ecological distance. Vegetatio 69:57–68.

Ferrari, S. L. P., and F. Cribari-Neto. 2004. Beta Regression for modelling rates and

proportions. Journal of Applied Statistics 31:799–815.

Fox, J., and S. Weisberg. 2010. An {R} Companion to Applied Regression. 2nd edition.

Sage, Thousand Oaks, CA. Available from:

http://socserv.socsci.mcmaster.ca/jfox/Books/Companion

French, B. W., and N. C. Elliott. 1999. Temporal and spatial distribution of ground beetle

(Coleoptera: Carabidae) assemblages in grasslands and adjacent wheat fields.

Pedobiologia 43:73–84.

Gallé, R., I. Maák, and N. Szpisjak. 2014. The effects of habitat parameters and forest

age on the ground dwelling spiders of lowland poplar forests (Hungary). Journal of

Insect Conservation 18:791–799.

Gallé, R., A. Torma, and I. Maák. 2016. The effect of forest age and habitat structure on

the ground-dwelling ant assemblages of lowland poplar plantations. Agricultural and

Forest Entomology 18:151–156.

Gardner, S. M. 1991. Ground beetle (Coleoptera: Carabidae) communities on upland

heath and their association with heathland flora. Journal of Biogeography 18:281–

289.

Goslee, S. C., and D. L. Urban. 2013. The ecodist package for dissimilarity-based

analysis of ecological data. Journal of Statistical Software 22:1-19.

Gotelli, N. J., and R. K. Colwell. 2001. Quantifying biodiversity: procedures and pitfalls

in the measurement and comparison of species richness. Ecology Letters 4:379–391.

Gotelli, N. J., and R. K. Colwell. 2011. Estimating species richness. Pages 39–54 in A. E.

Magurran and B. J. McGill, editors. Biological Diversity: Frontiers in Measuring

and Assessment. Oxford University Press, New York.

Greenslade, P. J. M. 1964. Pitfall trapping as a method for studying populations of

Carabidae (Coleoptera). Journal of Animal Ecology 33:301–310.

Günther, J., and T. Assmann. 2004. Fluctuations of carabid populations inhabiting an

ancient woodland (Coleoptera, Carabidae). Pedobiologia 48:159–164.

Gutierrez, D., and R. Menendez. 1997. Patterns in the distribution, abundance and body

size of carabid beetles (Coleoptera: Caraboidea) in relation to dispersal ability.

Journal of Biogeography 24:903–914.

Haila, Y., I. K. Hanski, J. Niemelä, P. Punttila, S. Raivio, and H. Tukia. 1994. Forestry

and the boreal fauna: Matching management with natural forest dynamics. Annales

Zoologici Fennici 31:187–202.

Harvey, J. A., W. H. van der Putten, H. Turin, R. Wagenaar, and T. M. Bezemer. 2008.

Effects of changes in plant species richness and community traits on carabid

assemblages and feeding guilds. Agriculture Ecosystems and Environment 127:100–

106.

Heliövaara, K., and R. Väisänen. 1984. Effects of modern forestry on northwestern

European forests invertebrates: A synthesis. Acta Forestalia Fennica 189:1–32.

Heltshe, J. F., and N. E. Forrester. 1983. Estimating species richness using the Jackknife

procedure. Biometrics 39:1–11.

Henderson, B. M., and S. J. Walsh. 1995. “Plowed, paved, or in succession”: Land-cover

change on the North Carolina Piedmont. Southeastern Geographer 35:132–149.

Holm, S. 1979. A simple sequentially rejective multiple test procedure. Scandinavian

Journal of Statistics 6:65–70.

Holt, R. D., J. H. Lawton, G. A. Polis, and N. D. Martinez. 1999. Trophic rank and the

species-area relationship. Ecology 80:1495–1504.

Horn, J. C. 1980. Short-term changes in vegetation after clearcutting in the southern

Appalachians. Southern Appalachian Botanical Society 45:88–96.

Hortal, J., P. A. V. Borges, and C. Gaspar. 2006. Evaluating the performance of species

richness estimators: sensitivity to sample grain size. Journal of Animal Ecology

75:274–287.

Hothorn, T., F. Bretz, and P. Westfall. 2008. Simultaneous inference in general

parametric models. Biometrical Journal 50:346–363.

Hsieh, T. C., K. H. Ma, and A. Chao. 2015. iNEXT: An R package for interpolation and

extrapolation of species diversity (Hill numbers). R package version 2.0.8. Available

from: http://chao.stat.nthu.edu.tw/blog/software-download

Hurvich, C. M., and C.-L. Tsai. 1989. Regression and time series model selection in

small samples. Biometrika 76:297–307.

Ings, T. C., and S. E. Hartley. 1999. The effect of habitat structure on carabid

communities during the regeneration of a native Scottish forest. Forest Ecology and

Management 119:123–136.

Jacobs, J. M., T. T. Work, and J. R. Spence. 2008. Influences of succession and harvest

intensity on ground beetle (Coleoptera: Carabidae) populations in the boreal

mixedwood forests of Alberta, Canada: species matter. Pages 437–462 in Back to

the Roots and Back to the Future. Towards a New Synthesis amongst Taxonomic,

Ecological, and Biogeographical Approaches in Carabidology Proceedings of the

XIII European Carabidologists Meeting. Blagoevgrad, Bulgaria.

Jelaska, L. Š., V. Dumbović, and M. Kučinić. 2011. Carabid beetle diversity and mean

individual biomass in beech forests of various ages. ZooKeys100:393–405.

Jennings, D. T., M. W. Houseweart, and G. A. Dunn. 1986. Carabid beetles (Coleoptera:

Carabidae) associated with strip clearcut and dense spruce-fir forests of Maine.

Coleopterists Bulletin 40:251–263.

Jennings, V. H., and D. W. Tallamy. 2006. Composition and abundance of ground-

dwelling Coleoptera in a fragmented and continuous forest. Environmental

Entomology 35:1550–1560.

Johansson, T., J. Hjältén, J. Olsson, M. Dynesius, and J-M. Roberge. 2016. Long-term

effects of clear-cutting on epigaeic beetle assemblages in boreal forests. Forest

Ecology and Management 359:65–73.

Jost, L. 2006. Entropy and diversity. Oikos 113:363–375.

Jukes, M. R., A. J. Peace, and R. Ferris. 2001. Carabid beetle communities associated

with coniferous plantations in Britain: The influence of site, ground vegetation and

stand structure. Forest Ecology and Management 148:271–286.

Kaila, L., P. Martikainen, and P. Punttila. 1997. Dead trees left in clear-cuts benefit

saproxylic Coleoptera adapted to natural disturbances in boreal forest. Biodiversity

and Conservation 6:1–18.

Karen, M., J. O’Halloran, J. Breen, P. Giller, J. Pithon, and T. Kelly. 2008. Distribution

and composition of carabid beetle (Coleoptera, Carabidae) communities across the

plantation forest cycle – Implications for management. Forest Ecology and

Management 256:624–632.

Keller, J. K., M. E. Richmond, and C. R. Smith. 2003. An explanation of patterns of

breeding bird species richness and density following clearcutting in northeastern

USA forests. Forest Ecology and Management 174:541–564.

Kirkland, G. L. 1990. Patterns of initial small mammal community change after

clearcutting of temperate North American forests. Oikos 59:313–320.

Koivula, M. J. 2011. Useful model organisms, indicators, or both? Ground beetles

(Coleoptera, Carabidae) reflecting environmental conditions. Pages 287–317 in D. J.

Kotze, T. Assmann, J. Noordijk, H. Turin, and R. Vermeulen, editors. Carabid

Beetles as Bioindicators: Biogeographical, Ecological and Environmental Studies.

ZooKeys 100:287–317.

Koivula, M., J. Kukkonen, and J. Niemela. 2002. Boreal carabid-beetle (Coleoptera,

Carabidae) assemblages along the clear-cut originated succession gradient.

Biodiversity and Conservation 11:1269–1288.

Kotze, D. J., and J. Niemela. 2002. Year-to-year variation in carabid beetle (Coleoptera,

Carabidae) assemblages on the Åland lslands, southwest Finland. Journal of

Biogeography 55:375–386.

Kotze, J. D. et al. 2011. Forty years of carabid beetle research in Europe – From

taxonomy, biology, ecology and population studies to bioindication, habitat

assessment and conservation. ZooKeys 100:55–148.

Kwiatkowski, A. 2011. Assemblages of carabid beetles (Coleoptera, Carabidae) in humid

forest habitats of different stages of succession in the Puszcza Knyszyńska Forest

(northeastern Poland). ZooKeys 459:447–59.

Larochelle, A., and M.-C. Larivière. 2003. A Natural History of the Ground-Beetles

(Coleoptera: Carabidae) of America North of Mexico. Pensoft Publishers, Sofia.

Lassauce, A., L. Larrieu, Y. Paillet, F. Lieutier, and C. Bouget. 2013. The effects of

forest age on saproxylic beetle biodiversity: Implications of shortened and extended

rotation lengths in a French oak high forest. Insect Conservation and Diversity

6:396–410.

Latty, E. F., S. M. Werner, D. J. Mladenoff, K. F. Raffa, and T. A. Sickley. 2006.

Response of ground beetle (Carabidae) assemblages to logging history in northern

hardwood–hemlock forests. Forest Ecology and Management 222:335–347.

Lee, S. M., and A. Chao. 1994. Estimating population size via coverage for closed

capture-recapture models. Biometrics 50:88–97.

Legendre, P., and M. J. Anderson. 1999. Distance-based redundancy analysis: Testing

multispecies responses in multifactorial ecological experiments. Ecological

Monographs 69:1–24.

Legendre, P., and L. Legendre, editors. 2012. Numerial Ecology. Vol. 24, 3rd edition.

Elsevier, Amsterdam.

Lenski, R. E. 1982. The impact of forest cutting on the diversity of ground beetles

(Coleoptera, Carabidae) in the southern Appalachians. Ecological Entomology

7:385–390.

Levin, S. A., and R. T. Paine. 1974. Disturbance, patch formation, and community

structure. Proceedings of the National Academy of Sciences71:2744–2747.

Liebherr, J., and J. Mahar. 1979. The carabid fauna of the upland oak forest in Michigan:

survey and analysis. Coleopterists Bulletin 33:183–197.

Lindroth, C. H. 1961-1969. The ground-beetles (Carabidae, excl. Cicindelinae) of Canada

and Alaska, Parts 1-6. Opuscula Entomologica. Suppl. 20, 24, 29, 33, 34, 35.

Loreau, M. 1992. Species abundance patterns and the structure of ground-beetle

communities. Annales Zoologici Fennici 28:49–56.

Lövei, G. L., and K. D. Sunderland. 1996. Ecology and behavior of ground beetles

(Coleoptera: Carabidae). Annual Review of Entomology 41:231–256.

Luff, M. L. 1975. Some features influencing the efficiency of pitfall traps. Oecologia

19:345–357.

Luff, M. L. 1982. Population dynamics of Carabidae. Annals of Applied Biology

101:164–170.

Luff, M. L., M. D. Eyre, and S. P. Rushton. 1989. Classification and ordination of

habitats of ground beetles (Coleoptera , Carabidae) in North- East England. Journal

of Biogeography 16:121–130.

Luff, M. L., M. D. Eyre, and S. P. Rushton. 1992. Classification and prediction of

grassland habitats using ground beetles (Coleoptera, Carabidae). Journal of

Environmental Management 35:301–315.

Lundgren, J. G., J. J. Duan, M. S. Paradise, and R. N. Wiedenmann. 2005. Rearing

protocol and life history traits for Poecilus chalcites (Coleoptera: Carabidae) in the

laboratory. Journal of Entomological Science 40:126–135.

MacLean, D. B., and J. D. Usis. 1992. Ground Beetles (Coleoptera: Carabidae) of eastern

Ohio forests threatened by the gypsy moth, Lymantria dispar (L.) (Lepidoptera:

Lymantriidae). Ohio Journal of Science 92:46–50.

Magura, T., B. Tóthmérész, and T. Molnár. 2001. Forest edge and diversity: Carabids

along forest-grassland transects. Biodiversity and Conservation 10:287–300.

Magura, T., B. Tóthmérész, and Z. Elek. 2006. Changes in carabid beetle assemblages as

Norway spruce plantations age. Community Ecology 7:1–12.

Magura, T., D. Bogyó, S. Mizser, D. D. Nagy, and B. Tóthmérész. 2015. Recovery of

ground-dwelling assemblages during reforestation with native oak depends on the

mobility and feeding habits of the species. Forest Ecology and Management

339:117–126.

Mantel, N. 1967. The detection of disease clustering and a generalized regression

approach. Cancer Research 27:209–220.

McArdle, B. H., and M. J. Anderson. 2001. Fitting multivariate models to community

data: A comment on distance-based redundancy analysis. Ecology 82:290–297.

McCune, B., and J. B. Grace. 2002. Analysis of Ecological Communities. MjM software

design, Gleneden Beach, Oregon.

Melbourne, B. A. 1999. Bias in the effect of habitat structure on pitfall traps: An

experimental evaluation. Austral Ecology 24:228–239.

Minchin, P. R. 1987. An evaluation of the relative robustness of techniques for ecological

ordination. Vegetatio 69:89–107.

Napton, D. E., R. F. Auch, R. Headley, and J. L. Taylor. 2010. Land changes and their

driving forces in the Southeastern United States. Regional Environmental Change

10:37–53.

Niemelä, J. 1997. Invertebrates and boreal forest management. Conservation Biology

11:601–610.

Niemelä, J., Y. Haila, E. Halme, T. Lahti, T. Pajunen, and P. Punttila. 1988. The

distribution of carabid beetles in fragments of old coniferous taiga and adjacent

managed forest. Annales Zoologici Fennici 25:107–119.

Niemelä, J., Y. Haila, E. Halme, T. Pajunen, and P. Punttila. 1992a. Small-scale

heterogeneity in the spatial distribution carabid beetles in the southern Finnish taiga.

Journal of Biogeography 19:173–181.

Niemelä, J., J. Spence, D. Spence. 1992b. Habitat assocations and seasonal activity of

ground-beetles (Coleoptera, Carabidae) in Central Alberta. Canadian Entomologist

124: 521-540.

Niemelä, J. K., Y. Haila, and P. Punttila. 1996. The importance of small-scale

heterogeneity in boreal forests: Variation in diversity in forest-floor invertebrates

across the succession gradient. Ecography 19:352–368.

Niemelä, J. K., D. Langor, and J. R. Spence. 1993. Effects of clear-cut harvesting on

Boreal ground-beetle assemblages (Coleoptera, Carabidae) in Western Canada.

Conservation Biology 7:551–561.

Niemelä, J., J. Kotze, A. Ashworth, P. Brandmayr, K. Desender, T. New, L. Penev, M.

Samways, and J. Spence. 2000. The search for common anthropogenic impacts on

biodiversity: A global network. Journal of Insect Conservation 4:3–9.

North Carolina Department of Environment and Natural Resources Division of Water

Resources. 2012. Status report to the general assembly on the North Carolina

Drought Management Advisory Council: July 1, 2011 - June 30, 2012.

North Carolina Drought Management Advisory Council. 2010. Annual Activities Report

- 2010. North Carolina Division of Water Resources, Department of Environment

and Natural Resources.

North Carolina Drought Management Advisory Council. 2011. Annual Activities Report

- 2011. North Carolina Division of Water Resources, Department of Environment

and Natural Resources.

O’Hara, R. B., and D. J. Kotze. 2010. Do not log-transform count data. Methods in

Ecology and Evolution 1:118–122.

Oden, N. L., and R. R. Sokal. 1986. Directional autocorrelation: An extension of spatial

correlograms to two dimensions. Systematic Zoology 35:608–617.

Oksanen, J. 2015. Multivariate analysis of ecological communities in R: vegan tutorial. R

package version 3.0.1. Available from:

http://cc.oulu.fi/~jarioksa/opetus/metodi/vegantutor.pdf

Oksanen, J., F. G. Blanchet, R. Kindt, P. Legendre, P. R. Minchin, R. B. O’Hara, G. L.

Simpson, P. Solymos, M. H. H. Stevens, and H. Wagner. 2016. vegan: Community

Ecology Package. Available from: https://CRAN.R-project.org/package=vegan

Olson, D. M., and E. Dinerstein. 2002. The global 200: Priority ecoregions for global

conservation. Annals of the Missouri Botanical Garden 89:199–224.

Oosting, H. J. 1942. An ecological analysis of the plant communities of Piedmont, North

Carolina. American Midland Naturalist 28:1–126.

Paje, F., and D. Mossakowski. 1984. pH-preferences and habitat selection in carabid

beetles. Oecologia 64:41–46.

Pajunen, T., Y. Haila, E. Halme, J. Niemelä, and P. Punttila. 1995. Ground-dwelling

spiders (Arachnida, Araneae) in fragmented old forests and surrounding managed

forests in Southern Finland. Ecography 18:62–72.

Palmer M. W. 1990. The estimation of species richness by extrapolation. Ecology

71:1195–1198.

Pan, Y., J. M. Chen, R. Birdsey, K. McCullough, L. He, and F. Deng. 2011. Age

structure and disturbance legacy of North American forests. Biogeosciences 8:715–

732.

Payne, L. X., D. E. Schindler, J. K. Parrish, and S. A. Temple. 2005. Quantifying spatial

pattern with evenness indices. Ecological Applications 15:507–520.

Pearce, J. L., and L. A. Venier. 2006. The use of ground beetles (Coleoptera: Carabidae)

and spiders (Araneae) as bioindicators of sustainable forest management: A review.

Ecological Indicators 6:780–793.

Pearce, J. L., L. A. Venier, J. McKee, J. Pedlar, and D. McKenney. 2003. Influence of

habitat and microhabitat on carabid (Coleoptera: Carabidae) assemblages in four

stand types. Canadian Entomologist 135:337–357.

Peet, R. K., and N. L. Christensen. 1980. Succession: A population process. Vegetatio

43:131–140.

Peet, R. K., and N. L. Christensen. 1987. Competition and tree death. Bioscience 37:586–

595.

Peet, R. K., N. L. Christensen, and F. S. Gilliam. 2014. Temporal patterns in herbaceous

layer communities of the North Carolina Piedmont. F. S. Gilliam, editor. The

Herbaceous Layer in Forests of Eastern North America. Oxford University Press.

Pitman, N. C. A., J. W. Terborgh, M. R. Silman, P. Núñez, D. A. Neill, C. E. Cerón, W.

A. Palacios, and M. Aulestia. 2001. Dominance and distribution of tree species in

upper Amazonian terra firme forests. Ecology 82:2101–2117.

Punttila, P., Y. Haila, T. Pajunen, and H. Tbkia. 1991. Colonisation of clearcut forests by

ants in the Southern Finnish taiga: A quantitative survey. Oikos 61:250–262.

Purrington, F. F., J. E. Bater, M. G. Paoletti, and B. R. Stinner. 1989. Ground beetles

from a remnant oak-maple-beech forest and its surroundings in northeastern Ohio

(Coleoptera: Carabidae). Great Lakes Entomologist 22:105–110.

R Core Team. 2016. R: A Language and Environment for Statistical Computing. R

Foundation for Statistical Computing, Vienna, Austria. Available from:

https://www.R-project.org/

Rainio, J., and J. Niemela. 2003. Ground beetles (Coleoptera: Carabidae) as bioindicators.

Biodiversity and Conservation 12:487–506.

Ricketts, T., and M. Imhoff. 2003. Biodiversity, urban areas, and agriculture: Locating

priority ecoregions for conservation. Conservation Ecology 8:1.

Riley, K. N. 2010. Ground beetles (Coleoptera: Carabidae) as biodiversity indicators for

age structure in Piedmont Forests? Master’s Thesis,Wake Forest University,

Winston-Salem, NC

Riley, K. N., and R. A. Browne. 2011. Changes in ground beetle diversity and

community composition in age structured forests (Coleoptera, Carabidae). ZooKeys

147:601–621.

Rykken, J. J., D. E. Capen, and S. P. Mahabir. 1997. Ground Beetles as indicators of land

type diversity in the Green Mountains of Vermont. Conservation Biology 11:522–

530.

Schirmel, J., L. Timler, and S. Buchholz. 2011. Impact of the invasive moss Campylopus

introflexus on carabid beetles (Coleoptera: Carabidae) and spiders (Araneae) in

acidic coastal dunes at the southern Baltic Sea. Biological Invasions 13:605–620.

Seastedt, T. R., and D. A. Crossley. 1981. Microarthropod response following cable

logging and clear-cutting in the southern Appalachians. Ecology 62:126–135.

Simpson, E. H. 1949. Measurement of diversity. Nature 163:688.

Smith, B., and J. B. Wilson. 1996. A consumer’s guide to evenness indices. Oikos 76:70–

82.

Sokal, R. R. 1986. Spatial data analysis and historical processes. Pages 29–43 in E. Diday

and et al., editors. Data Analysis and Informatics, IV. North-Holland, Amsterdam.

Southwood, T. R. E., V. K. Brown, and P. M. Reader. 1979. The relationships of plant

and insect diversities in succession. Biological Journal of the Linnaean Society

12:327–348.

Southwood, T. R. E., and P. A. Henderson. 2000. Ecological Methods. 3rd edition.

Blackwell Science Ltd, Oxford.

Spence, J. R., and J. Niemelä. 1994. Sampling carabid assemblages with pitfall traps: The

maddness and the method. Canadian Entomologist 126:881–894.

Spence, J. R., D. W. Langor, J. Niemelä, H. A. Carcamo, and C. R. Currie. 1996.

Northern forestry and carabids: The case for concern about old-growth species.

Annales Zoologici Fennici 33:173–184.

State Climate Office of North Carolina, NC State University. CRONOS [internet

database] available at http://climate.ncsu.edu/cronos/. Accessed August 25, 2016.

Stork, N. E., and P. Eggleton. 1992. Invertebrates as determinants and indicators of soil

quality. American Journal of Alternative Agriculture 7:38–47.

Sugiura, N. 1978. Further analysis of the data by Akaike’s Information Criterion and the

finite corrections. Communications in Statistics - Theory and Methods 7:13–26.

Swank, W. T., and J. M. Vose. 1988. Effects of cutting practices on microenvironment in

relation to hardwood regeneration. Pages 71–88 in H. C. Clya, A. W. Perkey, and W.

E. Kidd, editors. Guidelines for Regenerating Appalachian Hardwood Stands: Proc.

Workshop, Morgantown, WV, 24-26 May 1988. Society of American Foresters

Publication 88-03, West Virginia University Books, Morgantown, WV.

Szyszko, J. 1983. State of Carabidae (Col.) fauna in fresh pine forest and tentative

valorisation of this environment. Warsaw Agricultural University Press, Warsaw.

Szyszko, J. 1990. Planning of prophylaxis in threatened pine forest biocenoses based on

an analysis of the fauna of epigeic Carabidae. Agricultural University Press, Warsaw.

Taboada, A., D. J. Kotze, R. Tárrega, and J. M. Salgado. 2008. Carabids of differently

aged reforested pinewoods and a natural pine forest in a historically modified

landscape. Basic and Applied Ecology 9:161–171.

Thiele, H. U. 1977. Carabid Beetles in Their Environments. A Study on Habitat Selection

by Adaptation in Physiology and Behaviour. Springer-Verlag, Berlin.

Venables, W. N., and B. D. Ripley. 2002. Modern Applied Statistics with S. 4th edition.

Springer, New York.

Walther, B. A., and J. L. Moore. 2005. The concept of bias, precision and accuracy, and

their use in testing the performance of species richness estimators, with a literature

review of estimator performance. Ecography 28:815–829.

Werner, S. M., and K. F. Raffa. 2000. Effects of forest management practices on the

diversity of ground-occurring beetles in mixed northern hardwood forests of the

Great Lakes Region. Forest Ecology and Management 139:135–155.

Woodcock, B. A. 2005. Pitfall trapping in ecological studies. Pages 37–57 in S. R.

Leather, editor. Insect Sampling in Forest Ecosystems. Blackwell Publishing,

Oxford.

Zuur, A. F., E. N. Ieno, N. J. Walker, A. A. Saveliev, and G. M. Smith. 2009. Mixed

Effects Models and Extensions in Ecology. R. M. Gail, K. Krickeberg, J. M. Samet,

A. Tsiatis, and W. Wong, editors. Springer Science+Business Media, LLC, New

York, USA.

TABLES

II Table – 1. Measures of numbers of individuals, species richness and diversity of

Carabidae among the four forest age classes and the summed totals for the dataset. The values include all sampling years. Note: 1/D = inverse Simpson’s concentration index,

E1/D = Simpson’s evenness.

Measure Zero Ten Fifty Ninety Total No. individuals 1726 1352 1570 1901 6548 Avg. individuals (± s.e.) 78.4 ± 14.6 64.4 ± 8.1 74.8 ± 11.0 63.4 ± 5.6 -- No. tribes 14 12 14 14 16 No. genera 27 22 26 27 33 No. species 59 40 42 48 76 Avg. species (± s.e.) 13.8 ± 1.1 10.7 ± 1.0 12.7 ± 1.0 11.5 ± 0.5 -- Avg. 1/D 5.26 ± 0.74 4.17 ± 0.23 5.29 ± 0.32 4.76 ± 0.26 4.87 ± 0.22

Avg. E1/D (± s.e.) 0.39 ± 0.03 0.44 ± 0.04 0.45 ± 0.03 0.43 ± 0.03 -- Richness estimators Chao1 (± s.e.) 107.1 ± 36.9 41.1 ± 1.6 55.5 ± 12.4 76.2 ± 23.2 186.2 ± 92.1 ACE 80.6 42.4 49.3 60.8 120.0 Jack1 (± s.e.) 80.0 ± 6.7 48.6 ±4.2 51.5 ± 4.1 63.5 ± 4.6 97.8 ± 5.5

II Table – 2. Mean Bray-Curtis dissimilarity values for forest age classes when all sampling years are combined. Shaded values along the diagonal are the mean within- group dissimilarities, with non-shaded values representing among-group dissimilarities.

Dissimilarity values range between 0 – 1, with a value of 1 the maximum dissimilarity possible (no shared species).

Zero Ten Fifty Ninety Zero 0.713 Ten 0.794 0.560 Fifty 0.724 0.611 0.495 Ninety 0.750 0.583 0.577 0.490

II Table – 3. Carabid species with significant associations (p < 0.05) for one of the four forest age classes or groups of forest age classes. Associations were based on the point- biserial group-equalized phi coefficient (De Cáceres & Legendre 2009) which measures the strength of species ecological preferences with the associated habitat. Adjusted p- values (p.adj holm) for multiple testing correspond to experiment-level conclusions.

p.adj Species Code Tribe:Genus Coefficient p (holm) Zero Harpen Harpalini: Harpalus 0.602 0.003** 0.036* Selopa Harpalini: Selenophorus 0.588 0.003** 0.036* Amacup Zabrini: Amara 0.540 0.009** 0.063 Amacra Zabrini: Amara 0.474 0.002** 0.030* Anirus Harpalini: Anisodactylus 0.462 0.025* 0.084 Notter Harpalini: Notiobia 0.447 0.004** 0.036* Myacor Pterostichini: Myas 0.445 0.033* 0.084 Anihap Harpalini: Anisodactylus 0.414 0.017* 0.084

Ten

Pteado Pterostichini: Pterostichus 0.691 0.002** 0.030*

Fifty Dicdil Licinini: Dicaelus 0.644 0.002** 0.030* Pladec Platynini: Platynus 0.572 0.001** 0.017* Ptescu Pterostichini: Pterostichus 0.511 0.016* 0.084

Ninety Ptemoe Pterostichini: Pterostichus 0.597 0.004** 0.036* Galbic Galertini: Galerita 0.439 0.039* 0.084

Ten + Ninety Dicpol Licinini: Dicaelus 0.632 0.001 0.017*

Ten + Fifty + Ninety Sphste Cychrini: Sphaeroderus 0.569 0.003 0.036* Dicamb Licinini: Dicaelus 0.509 0.014 0.084

II Table – 4. Numbers of individuals, species richness and diversity of carabid beetles by sampling year. The additional year of sampling was omitted to facilitate inter-annual comparisons. Note: 1/D = inverse Simpson’s concentration index, E1/D = Simpson’s evenness.

Measure One Two Three N total 2310 2135 1210 No. tribes 15 13 14 No. genera 31 27 26 S 61 53 49 1/D (± s.e.) 4.76 ± 0.31 5.12 ± 0.33 4.72 ± 0.47

Avg. E1/D (± s.e.) 0.38 ± 0.02 0.40 ± 0.03 0.52 ± 0.03

FIGURES

II Figure – 1. Map of the study area and the 34 sampling sites with forest age classes depicted by different point markers. Sites with an asterisk next to the point marker are where sampling occurred for an additional year.

II Figure – 2. Rarefaction interpolation (solid lines) and extrapolation (dashed lines) curves for zero, ten, fifty and ninety year forests with all four sampling years combined.

Point markers represent the sampling extent of the current study, with shaded areas depicting unconditional 95% confidence intervals. The zero age class differs significantly from the other three age classes at the rarefied sample size (n = 1,352).

II Figure – 3. Boxplots by forest age and sampling year for the abundance-based nonparametric diversity estimator, Chao1. A: Chao1 was significantly lower for ten and ninety year forests compared to the zero age class (p = 0.001). B: Chao1 was significantly lower for year three than years one and two (p = 0.003).

II Figure – 4. Rank abundance distribution curves for the four forest age classes with all sampling years combined.

II Figure – 5. Non-metric multidimensional scaling (NMDS) ordination of Carabidae species assemblages for the four forest age classes with all sampling years combined (k =

2; stress = 17.0). All species (S = 76) were included in the analysis with each data point

(denoted by site name) representing one of the 34 sampling sites based on relative abundance using the Bray-Curtis dissimilarity index. Significant differences (p < 0.001) in species assemblages occurred among forest age classes, with the most distinct species assemblages between the zero age class and the three older forest age classes.

Just for

II Figure – 6. Numbers of individuals (N) and numbers of species (S) known to prefer open habitats or shaded habitats changed with forest age. The higher richness for zero year sites was primarily due to an influx of species preferring open habitats. Forest age classes with different letters are significantly different from those with the same letter. A:

There were significant differences for mean numbers of individuals preferring open conditions among the forest age classes (p < 0.001) and for mean numbers of individuals preferring shaded conditions (p = 0.016). B: There were significant differences for mean number of open habitat species with forest age (p < 0.001) and for mean number of shaded habitat species with forest age (p < 0.001).

II Figure – 7. Numbers of individuals and number of species for the first three sampling years. A: Significantly fewer individuals were collected in year three compared to years one and two (p < 0.001). B: Mean species number was significantly lower in year three compared to years one and two (p < 0.001).

II Figure – 8. Ordination analyses depicting year-to-year variation in species composition within zero, ten, fifty and ninety year old forests. Results include all species sampled in the first three years and are derived using relative abundance with the Bray-Curtis dissimilarity index. Forest age categories, with the final solution NMDS stress in parentheses, are A: zero year (13.9), B: ten year (6.1), C: fifty year (12.6), D: ninety year

(18.6). The largest year to year changes in species composition occurred for the zero year sites (p = 0.018) and the lowest inter-annual changes for ten year forests. There also were significant differences in year-to-year species composition for fifty (p = 0.004) and ninety year forests (p = 0.002).

II Figure – 9. Ordination analyses depicting year-to-year variation among forest age classes. These results include all species sampled in the first three years and are derived using relative abundance with the Bray-Curtis dissimilarity index. Forest age categories, with the final solution NMDS stress in parentheses, are A: year one (17.8), B: year two

(17.4), C: year three (20.4). There were significant variations among forest age classes in each sampling year (p < 0.001).

APPENDIX A

II Figure – A1. Normal monthly averages for precipitation, max temperature (Tmax), minimum temperature (Tmin), mean temperature (Tmean) for the study area. Climate normals are based on data from 1971 – 2000 and collected in Danbury, NC (36.4 N, -80.2

W) in Stokes County (State Climate Office 2016).

* Data from the Climate Division provided by the National Centers for Environmental Information (NOAA) obtained from State Climate Office of North Carolina, NC State University.

II Figure – A2. Annual summed precipitation (inches) and average temperatures (°F) from 2006 – 2013 for the northern Piedmont of North Carolina (State Climate Office

2016). Normal average temperature and precipitation (dashed lines) were calculated using data from 1971 – 2013.

* Data from the Climate Division provided by the National Centers for Environmental Information (NOAA) obtained from State Climate Office of North Carolina, NC State University.

APPENDIX B

Table II – B1. Additional information for the 34 sampling sites, including circumferences for the two largest trees within a ten meter radius of the center of each study site and site elevation (a.s.l.). The recently clear-cut areas (‘Zero’), 10 – 20 year old forest (‘Ten’), 40

– 60 year old forest (‘Fifty’), and 80 – 95 year forest (‘Ninety’). Va = Virginia.

Circum- Circum- Elev Site Tree Species ference Tree Species ference (m) (in.) (in.) ZeroA 309 ZeroB 307 ZeroC 304 ZeroD 295 ZeroE 302 ZeroF 298 ZeroPT 333 ZeroSP 381 ZeroWL 315 Liriodendron Liriodendron 23.2, 22.8, TenA 324 tulipifera (Yellow 21.2, 23.7 tulipifera (Yellow 24.2 poplar) poplar) Pinus virginiana Pinus virginiana TenB 334 12.1 9.8 (Va oldfield) (Va oldfield) Liriodendron Prunus TenC 335 tulipifera (Yellow 6.5 pensylvanica 8.9 poplar) (Cherry) Quercus alba Pinus virginiana TenD 320 11.7 11.7 (White oak) (Va oldfield) Liriodendron Pinus virginiana TenE 318 15.6 tulipifera (Yellow 14.6 (Va oldfield) poplar) Pinus virginiana Pinus virginiana TenF 323 13.5 13.2 (Va oldfield) (Va oldfield) Liriodendron Pinus taeda TenBB 368 18 tulipifera (Yellow 15 (Loblolly) poplar) Pinus taeda Pinus taeda TenPZ 270 25 16 (Loblolly) (Loblolly) Pinus virginiana Pinus virginiana FiftyA 325 34.6 25.9 (Va oldfield) (Va oldfield) Pinus echinata Pinus echinata FiftyB 318 26.5 25.9 (shortleaf) (shortleaf) Quercus alba Quercus alba FiftyC 347 31.3 33.7 (White oak) (White oak) Pinus virginiana Pinus virginiana FiftyD 303 51.4 43.9 (Va oldfield) (Va oldfield) Pinus virginiana Pinus virginiana FiftyE 290 41.1 40.6 (Va oldfield) (Va oldfield) Pinus virginiana Pinus virginiana FiftyF 297 43.3 39.8 (Va oldfield) (Va oldfield) Pinus virginiana Quercus alba 36.6, 22.7, FiftyG 285 36.5 (Va oldfield) (White oak) 31.7 Liriodendron Liriodendron 49.5, 56.1, 40.3, 44.3, NinetyA 342 tulipifera (Yellow tulipifera (Yellow 54.5 48.1 poplar) poplar) Liriodendron Quercus alba NinetyB 328 81.1 tulipifera (Yellow 70.5 (White oak) poplar)

Quercus prinus Quercus alba NinetyC 314 82.3 87.6 (Chesnut Oak) (White oak) Quercus prinus Quercus prinus NinetyD 314 72.5 68 (Chesnut Oak) (Chesnut Oak) Quercus alba NinetyE 295 64.9 Quercus (Oak) 49.6 (White oak) Quercus alba Quercus velutina NinetyF 290 55.5 51.8 (White oak) (Black oak) Quercus alba Quercus alba NinetyG 258 93.9 53.1 (White oak) (White oak) Quercus alba Quercus alba NinetyH 271 48.2 44.3 (White oak) (White oak) Quercus alba Quercus prinus NinetyI 265 45.2 50.1 (White oak) (Chesnut Oak) Quercus alba Quercus alba NinetyJ 267 58.4 46.5 (White oak) (White oak)

Table II – B2. Longitude and latitude data for the 34 study sites.

AGE SITE LON LAT Zero ZeroA -80.2504 36.33590 Zero ZeroB -80.2505 36.33600 Zero ZeroC -80.2509 36.33582 Zero ZeroD -80.2525 36.33225 Zero ZeroE -80.2524 36.33250 Zero ZeroF -80.2527 36.33257 Zero ZeroPT -80.3957 36.34834 Zero ZeroSP -80.6946 36.52844 Zero ZeroWL -80.1855 36.36258 Ten TenA -80.4642 36.32737 Ten TenB -80.4578 36.34438 Ten TenBB -80.3914 36.35029 Ten TenC -80.4583 36.34443 Ten TenD -80.4579 36.34242 Ten TenE -80.4578 36.34258 Ten TenF -80.4578 36.34305 Ten TenPZ -80.1496 36.41340 Fifty FiftyA -80.4639 36.32713 Fifty FiftyB -80.4672 36.31923 Fifty FiftyC -80.4678 36.32970 Fifty FiftyD -80.4828 36.30270 Fifty FiftyE -80.4829 36.30035 Fifty FiftyF -80.4899 36.28443 Fifty FiftyG -80.4890 36.27047 Ninety NinetyA -80.4624 36.34068 Ninety NinetyB -80.4608 36.34152 Ninety NinetyC -80.4625 36.32888 Ninety NinetyD -80.4659 36.32115 Ninety NinetyE -80.4739 36.31360 Ninety NinetyF -80.4724 36.31458 Ninety NinetyG -80.4905 36.26818 Ninety NinetyH -80.4874 36.26237 Ninety NinetyI -80.4862 36.26133 Ninety NinetyJ -80.4956 36.27113

APPENDIX C

II Table – C1. Scientific names, species codes and tribes for the 76 Carabidae species collected.

Species Species code Tribe Agonum pallipes Fabricius Agopal Platynini Agonum punctiforme Fabricius Agopun Platynini Amara aenea (De Geer) Amaaen Zabrini Amara crassispina LeConte Amacra Zabrini Amara cupreolata Putzeys Amacup Zabrini Amara familiaris (Duftschmid) Amafam Zabrini Amara impuncticollis (Say) Amaimp Zabrini Amara musculis (Say) Amamus Zabrini Amphasia interstitialis (Say) Ampint Harpalini Anisodactylus carbonarius (Say) Anicar Harpalini Anisodactylus dulcicollis (LaFerté-Sénecteré) Anidul Harpalini Anisodactylus furvus LeConte Anifur Harpalini Anisodactylus haplomus Chaudior Anihap Harpalini Anisodactylus harrisii LeConte Anihar Harpalini Anisodactylus melanopus (Haldeman) Animel Harpalini Anisodactylus nigerrimus (Dejean) Aninig Harpalini Anisodactylus rusticus (Say) Anirus Harpalini Anisodactylus verticalis (LeConte) Aniver Harpalini Apenes lucidulus (Dejean) Apeluc Lebiini Brachinus americanus (LeConte) Braame Brachinini Brachinus fumans (Fabricius) Brafum Brachinini Calathus opaculus LeConte Calopa Platynini Carabus goryi Dejean Cargor Carabini Carabus sylvosus Say Carsyl Carabini Chlaenius aestivus Say Chlaes Chlaenini Chlaenius amoenus Dejean Chlamo Chlaenini Chlaenius emarginatus Say Chlema Chlaenini Cicindela sexguttata Fabricius Cicsex Cicindelini Cicindela unipunctata (Fabricius) Cicuni Cicindelini Colliuris pensylvanica (Linnaeus) Colpen Ctenodactylini Cyclotrachelus freitaga Bousquet Cycfre Pterostichini

Cyclotrachelus sigillatus (Say) Cycsig Pterostichini Cymindis americanus (Dejean) Cymame Lebiini Cymindis limbatus (Dejean) Cymlim Lebiini Cymindis neglectus (Haldeman) Cymneg Lebiini Dicaelus ambiguus LaFerté-Sénecteré Dicamb Licinini Dicaelus dilatatus Say Dicdil Licinini Dicaelus elongatus Bonelli Dicelo Licinini Dicaelus furvus Dejean Dicfur Licinini Dicaelus politus Dejean Dicpol Licinini Dicaelus purpuratus Bonelli Dicpur Licinini Galerita bicolor Dry Galbic Galeritini Harpalus compar Leconte Harcom Harpalini Harpalus fulgens Csiki Harful Harpalini Harpalus herbivagus Say Harher Harpalini Harpalus katiae Battoni Harkat Harpalini Harpalus pensylvanica De Geer Harpen Harpalini Harpalus spadiceus (Dejean) Harspa Harpalini Myas coracinus (Say) Myacor Pterostichini Notiobia terminata (Say) Notter Harpalini Notiophilus aeneus (Herbst) Notaen Notiophilini Oodes fluvialis LeConte Oodflu Oodini Pasimachus punctulatus Haldeman Paspun Scaritini Platynus decentis (Say) Pladec Platynini Poecilus chalcites (Say) Poecha Pterostichini Poecilus lucublandus Bonelli Poeluc Pterostichini Poecilus morphospecies wide Poelucw Pterostichini Poecilus morphospecies a Poemora Pterostichini Poecilus morphospecies b Poemorb Pterostichini Pterostichus adoxus (Say) Pteado Pterostichini Pterostichus coracinus (Newman) Ptecor Pterostichini Pterostichus moestus (Say) Ptemoe Pterostichini Pterostichus sculptus LeConte Ptescu Pterostichini Rhadine caudata (LeConte) Rhacau Platynini Scaphinotus andrewsii Valentine Scaand Cychrini Scaphinotus viduus (Dejean) Scavid Cychrini Scarites subterraneus Fabricius Scasub Scaritini Selenophorus ellipticus Dejean Selell Harpalini Selenophorus opalinus (LeConte) Selopa Harpalini Sphaeroderus bicarinatus (LeConte) Sphbic Cychrini Sphaeroderus stenostomus (Weber) Sphste Cychrini

Stenolophus ochropezus (Say) Steoch Harpalini Synuchus impunctatus (Say) Synimp Platynini Tetraca virginica (Linnaeus) Tetvir Cicindelini Trichotichnus autumnalis (Say) Triaut Harpalini Trichotichnus dichrous (Dejean) Tridic Harpalini

100

II Table – C2. Number of individuals for each species collected from the four forest age classes.

Species code Zero Ten Fifty Ninety Total Agopal 6 6 Agopun 77 77 Amaaen 16 1 17 Amacra 24 1 25 Amacup 18 2 20 Amafam 13 3 6 1 22 Amaimp 30 7 6 44 Amamus 1 1 Ampint 5 2 3 6 16 Anicar 1 1 Anidul 1 1 2 Anifur 16 16 Anihap 16 1 1 18 Anihar 32 1 5 38 Animel 1 1 Aninig 1 1 Anirus 34 2 1 38 Aniver 1 1 Apeluc 3 4 4 11 Braame 1 1 Brafum 1 1 Calopa 113 38 138 27 316 Cargor 1 4 19 54 78 Carsyl 8 2 11 3 24 Chlaes 73 60 31 94 258 Chlamo 2 2 3 7 Chlema 3 11 1 7 22 Cicsex 1 9 6 14 30 Cicuni 1 3 4 Colpen 1 1 Cycfre 2 24 105 125 256 Cycsig 300 108 193 467 1068 Cymame 1 3 6 6 16 Cymlim 2 1 3 Cymneg 1 1 Dicamb 5 88 62 89 243

101

Dicdil 8 35 142 30 214 Dicelo 64 9 24 2 99 Dicfur 1 1 Dicpol 4 30 10 42 85 Dicpur 6 2 4 12 Galbic 7 27 17 113 164 Harcom 1 1 Harful 1 1 Harher 1 1 Harkat 37 1 38 Harpen 270 11 57 13 352 Harspa 1 1 Myacor 23 5 28 Notter 42 42 Notaen 2 6 8 Oodflu 1 1 Paspun 45 11 9 38 103 Pladec 1 21 11 33 Poecha 1 1 Poeluc 4 4 1 9 Poelucw 1 1 2 Poemora 9 3 10 2 24 Poemorb 2 4 1 8 Pteado 14 311 7 35 366 Ptecor 49 27 21 110 207 Ptemoe 1 27 27 Ptescu 199 154 425 102 880 Rhacau 5 2 5 12 Scaand 13 15 3 11 43 Scavid 4 2 6 Scasub 41 9 10 4 64 Selell 1 1 Selopa 18 2 19 Sphbic 1 1 Sphste 48 312 194 421 975 Steoch 1 1 Synimp 1 1 Tetvir 8 6 7 7 28 Triaut 3 3 Tridic 1 1 Column totals 1726 1352 1570 1901 6548

102

II Table – C3. Numbers of individuals for each species collected from the four forest age classes for each sampling year. ‘Ex’ represents the additional year of sampling for zero and ten year old age classes.

Species Code Zero Ten Fifty Ninety Year 1 2 3 Ex 1 2 3 Ex 1 2 3 1 2 3 Total Agopal 6 6 Agopun 6 1 71 77 Amaaen 16 1 17 Amacra 4 4 8 8 1 25 Amacup 7 6 2 3 2 1 20 Amafam 9 5 3 5 1 1 22 Amaimp 3 2 21 4 7 5 2 44 Amamus 1 1 Ampint 5 2 1 2 1 5 16 Anicar 1 1 Anidul 1 1 2 Anifur 2 5 7 2 16 Anihap 1 1 4 10 1 1 18 Anihar 9 5 5 13 1 3 2 38 Animel 1 1 Aninig 1 1 Anirus 10 7 7 10 1 1 1 38 Aniver 1 1 Apeluc 2 1 1 1 2 1 3 11 Braame 1 1

103

Brafum 1 1 Calopa 5 6 10 92 31 5 1 83 45 9 14 11 2 316 Cargor 1 2 2 3 7 8 7 16 32 78 Carsyl 5 1 2 2 2 8 1 2 1 24 Chlaes 20 17 7 29 28 9 15 8 2 15 13 20 35 40 258 Chlamo 2 1 1 2 1 7 Chlema 1 2 8 2 1 1 1 2 4 22 Cicsex 1 3 4 3 2 2 2 4 2 8 30 Cicuni 1 1 2 4 Colpen 1 1 Cycfre 2 7 10 2 5 33 25 47 26 73 26 256 Cycsig 98 127 69 6 25 27 14 42 70 95 28 160 257 50 1068 Cymame 1 3 1 5 4 2 16 Cymlim 2 1 3 Cymneg 1 1 Dicamb 2 2 1 27 37 12 12 14 23 25 31 35 22 243 Dicdil 2 3 3 12 13 2 7 27 54 60 7 17 6 214 Dicelo 7 19 10 28 3 3 3 5 14 5 2 99 Dicfur 1 1 Dicpol 1 3 12 7 5 6 2 8 16 17 8 85 Dicpur 2 1 3 1 1 1 3 12 Galbic 2 1 1 3 21 1 5 6 11 60 37 16 164 Harcom 1 1 Harful 1 1 Harher 1 1 Harkat 1 1 35 1 38 Harpen 60 25 14 171 3 4 1 4 24 19 14 5 6 2 352

104

Harspa 1 1 Myacor 15 4 2 2 5 28 Notter 37 1 4 42 Notaen 1 1 1 5 8 Oodflu 1 1 Paspun 23 19 3 6 1 3 1 1 5 3 5 11 22 103 Pladec 1 9 8 3 4 5 1 33 Poecha 1 1 Poeluc 1 3 1 1 1 1 9 Poelucw 1 1 2 Poemora 2 3 4 1 2 5 3 1 2 24 Poemorb 2 1 3 1 8 Pteado 2 7 2 3 183 75 47 7 6 1 27 5 3 366 Ptecor 9 10 27 3 4 17 6 4 14 3 34 50 26 207 Ptemoe 1 10 17 27 Ptescu 22 26 21 130 37 78 8 31 231 119 75 71 22 10 880 Rhacau 2 3 2 4 2 12 Scaand 13 6 2 2 5 1 1 1 5 3 4 43 Scavid 2 2 2 6 Scasub 6 11 8 16 4 5 7 3 2 1 1 64 Selell 1 1 Selopa 9 6 3 1 1 19 Sphbic 1 1 Sphste 6 12 24 6 136 102 34 41 54 80 61 195 131 95 975 Steoch 1 1 Synimp 1 1 Tetvir 1 1 2 4 1 4 1 7 7 28

105

Triaut 2 1 3 Tridic 1 1 Column totals 407 354 285 680 575 408 156 213 599 594 377 730 780 391 6548

106

APPENDIX D

II Table – D1. Mean Bray-Curtis dissimilarity values by sampling year for community analyses of year-to-year variations among forest age classes (A: zero, B: ten, C: fifty, D: ninety). The mean within-group dissimilarities are the shaded boxes along the diagonal and the non-shaded values are the among group dissimilarities. The dissimilarity values range between 0 – 1, with a value of 1 the maximum dissimilarity possible (no shared species).

A One Two Three One 0.694

Two 0.628 0.611

Three 0.702 0.637 0.700

B One Two Three One 0.515

Two 0.545 0.605

Three 0.529 0.576 0.564

C One Two Three One 0.599

Two 0.577 0.555

Three 0.580 0.573 0.564

D One Two Three One 0.552

Two 0.554 0.524

Three 0.621 0.613 0.677

107

II Table – D2. Mean Bray-Curtis dissimilarity values by sampling year (A: one, B: two,

C: three) for community analyses of year-to-year variations among forest age classes.

The mean within-group dissimilarities are the shaded boxes along the diagonal and the non-shaded values are the among group dissimilarities. The dissimilarity values range between 0 – 1, with a value of 1 being the maximum dissimilarity possible (no shared species).

A Zero Ten Fifty Ninety Zero 0.694

Ten 0.861 0.515

Fifty 0.786 0.699 0.599

Ninety 0.780 0.613 0.637 0.552

B Zero Ten Fifty Ninety Zero 0.611

Ten 0.790 0.605

Fifty 0.693 0.659 0.555

Ninety 0.676 0.682 0.592 0.524

C Zero Ten Fifty Ninety Zero 0.700

Ten 0.839 0.564

Fifty 0.759 0.726 0.564

Ninety 0.759 0.719 0.693 0.677

108

APPENDIX E

II Table – E1. Mean Bray-Curtis dissimilarity values within and among forest age classes, with the first three sampling years combined. Mean within-group dissimilarities are shaded boxes along the diagonal and the non-shaded values are the among group dissimilarities. The dissimilarity values range between 0 – 1, with a value of 1 being the maximum dissimilarity possible (no shared species).

Zero Ten Fifty Ninety Zero 0.602 Ten 0.792 0.483 Fifty 0.709 0.621 0.495 Ninety 0.677 0.593 0.577 0.490

109

II Figure – E1. Non-metric multidimensional scaling (NMDS) ordination of Carabidae species assemblages for the four forest age classes with the first three sampling years combined. Each data point (denoted by site name) represents one of the 29 original sampling sites. Significant differences (p < 0.001) in species assemblages occurred among forest age classes (k = 2; stress = 16.7), with the most distinct species assemblages between the zero age class compared to older forest age classes. Polygon areas are as follows: zero (0.67), ten (0.21), fifty (0.28) and ninety (0.40).

110

APPENDIX F

II Table – F1. Mean Jaccard dissimilarity values (presence/absence data) for the forest age classes with all sampling years combined. Shaded values along the diagonal are the mean within-group dissimilarities and the non-shaded values represent the among-group dissimilarities. The dissimilarity values range between 0 – 1, with a value of 1 the maximum dissimilarity possible (no shared species).

Zero Ten Fifty Ninety Zero 0.707 Ten 0.716 0.579 Fifty 0.703 0.594 0.518 Ninety 0.750 0.599 0.593 0.567

111

II Table – F2. Mean Jaccard dissimilarity values (presence/absence data) by sampling year for community analyses of year-to-year variations within forest age classes (A: zero,

B: ten, C: fifty, D: ninety). The mean within-group dissimilarities are the shaded boxes along the diagonal and the non-shaded values are the among group dissimilarities. The dissimilarity values range between 0 – 1, with a value of 1 being the maximum dissimilarity possible (no shared species).

A One Two Three One 0.737 Two 0.708 0.707 Three 0.742 0.724 0.748

B One Two Three One 0.621

Two 0.616 0.637

Three 0.744 0.694 0.679

C One Two Three One 0.621

Two 0.619 0.574

Three 0.644 0.661 0.660

D One Two Three One 0.649

Two 0.603 0.581

Three 0.686 0.657 0.713

112

II Table – F3. Mean Jaccard dissimilarity values (presence/absence data) by sampling year (A: one, B: two, C: three) for community analyses of year-to-year variations among forest age classes. The mean within-group dissimilarities are the shaded boxes along the diagonal and the non-shaded values are the among group dissimilarities. The dissimilarity values range between 0 – 1, with a value of 1 being the maximum dissimilarity possible

(no shared species).

A Zero Ten Fifty Ninety Zero 0.594

Ten 0.653 0.465

Fifty 0.624 0.559 0.459

Ninety 0.665 0.521 0.519 0.488

B Zero Ten Fifty Ninety Zero 0.554

Ten 0.580 0.476

Fifty 0.572 0.517 0.417

Ninety 0.614 0.487 0.492 0.415

C Zero Ten Fifty Ninety Zero 0.605

Ten 0.765 0.568

Fifty 0.628 0.636 0.503

Ninety 0.683 0.617 0.564 0.567

113

II Figure – F1. Non-metric multidimensional scaling (NMDS) ordination of Carabidae species assemblages for the four forest age classes with all species included (k = 2; stress

= 18.8). Results are based on presence/absence data using Jaccard dissimilarity and for all four sampling years combined. Each data point (denoted by site name) represents one of the 34 sampling sites. Significant differences in species assemblages occurred among forest age classes (p < 0.001). The dispersion of the fifty year polygon was significantly less compared to the zero year polygon (p = 0.018). The most distinct species assemblages occurred between zero age class and the three older forest age classes.

114

II Figure – F2. Ordination analyses depicting year-to-year variation in species composition within zero, ten, fifty and ninety year old forests. Results include all species collected in the first three years, based on presence/absence data using Jaccard dissimilarity values. Forest age categories, with the final solution NMDS stress in parentheses, are A: zero year (18.9), B: ten year (10.9), C: fifty year (18.0), D: ninety year (20.8). There were significant differences in year-to-year species composition for ten

(p = 0.011), fifty (p = 0.003) and ninety year forests (p = 0.014).

115

II Figure – F3. Ordination analyses depicting year-to-year variation among forest age classes. Analyses included all species in sampling years one, two and three and are based on presence/absence data using Jaccard dissimilarity values. Forest age categories, with the final solution NMDS stress in parentheses, are A: year one (22.0), B: year two (19.0),

C: year three (20.7). There were significant variations among forest age classes in each sampling year (p < 0.001).

116

APPENDIX G

II Figure – G1. Carabid tribes with significant differences of number of individuals among forest age classes. Only tribes with significant differences are shown and forest age classes with different letters are significantly different from those with the same letter.

117

118

II Figure – G2. Numbers of individuals for Carabidae tribes among sampling years one, two and three. Only tribes with significant differences are shown (p < 0.05) and years with different letters are significantly different from those with the same letter.

119

APPENDIX H

II Table – H1. Carabidae tribes with higher than expected numbers of abundant species.

The number of species collected in this study are listed under ‘observed’, and the expected numbers of species were determined by multiplying the total number of species collected in that family by the proportion of species in each rarity category. Analysis includes all forest ages and sampling years combined. The relative abundance of commonly collected species contributed > 1% to the total number of individuals collected, whereas less frequently sampled species contributed < 1%. The number of abundant species among carabid tribes were significantly different than the expected number of species (Chi-square χ2 = 18.0, p < 0.025). Differences in the distribution of rare species among tribes were not significantly different than expected.

Abundant No. of S Direction of difference Tribe Observed Expected difference More than expected Licinini 4 1.42 2.58 Pterostichini 5 2.84 2.16 Scaritini 2 0.47 1.53 Galeritini 1 0.24 0.76 Platynini 2 1.42 0.58 Carabini 1 0.47 0.53

Approximately equal Chlaen ini 1 0.71 0.29 Cychrini 1 0.95 0.05

Less than expected Harpalini 1 5.21 -4.21

120

II Figure – H1. For the rare species category, boxplots depict (A) the number of individuals and (B) number of species among forest age classes. Significantly more rare individuals occurred for the zero age class compared to other age classes (p < 0.001) and significantly more rare species occurred in the zero age class compared to ten and ninety year forests (p = 0.002).

121

CHAPTER III

CHANGES IN SPECIES TRAITS OF CARABID BEETLES (COLEOPTERA:

CARABIDAE) WITH FOREST AGE

122

ABSTRACT

Assessing the effects of disturbance on community assemblages through changes in species traits provides an alternative approach to assist with the identification of ecological patterns and increase our ability to predict responses to disturbance regimes.

The environmental changes associated with clear-cutting and forest regeneration may affect not only the taxonomic species assemblages of carabid beetles (Coleoptera:

Carabidae), but also their ecological traits. The aims of this study were to investigate changes in ecological species traits and functional diversity for carabids among four forest age classes in the Piedmont of North Carolina. Sampling occurred over four years at 34 sites (408 pitfall traps) in four forest age classes: recently clear-cut (zero year), ten year old, fifty year old and ninety year old. A total of 6,489 carabid beetles representing

13 tribes, 27 genera and 48 species are included in the analysis. No significant differences among the forest age classes were detected for mean number of individuals, species richness, inverse Simpson’s diversity concentration, or for three functional diversity indices: functional richness, functional evenness and functional dispersion. Significant trends with forest age occurred for six of 18 species trait categories. Significantly higher numbers of macropterous (long-winged), omnivorous and pale legged species were collected from zero year forests compared to older forests. The number of brachypterous

(flightless) species increased significantly with forest age. The results from this study can be compared to carabid species traits in temperate forests worldwide and may provide insight to the responses of species traits (i.e. dispersal ability, body size) to clear-cutting and other forest disturbances for other ground-dwelling invertebrate taxa.

123

INTRODUCTION

Classification of communities using traits relating to species ecology, rather than taxonomy alone, can reveal ecologically important aspects of biological communities and serve broader conservation purposes (Whittaker 1975; Díaz & Cabido 2001; de Bello et al. 2010; Gagic et al. 2015). Analyses based on species traits and functional diversity indices convey information on the abundance and richness of traits, as well as the redundancy and complementarity of co-occurring species (Díaz & Cabido 2001; Petchey

& Gaston 2006; de Bello et al. 2010). Trait-based studies provide a less frequently reported dimension of animal biodiversity and, unlike taxonomic analyses, data are comparable across taxa, geographic locations, ecosystems and an array of disturbance types (Willby et al. 2000; Violle et al. 2007; Mouillot et al. 2013). Assessing changes in the composition, diversity and richness of ecological species traits along gradients can assist in the identification of ecological patterns and increase our ability to predict responses to disturbance regimes (Mouillot et al. 2013) which is valuable for policy implementation and practices (Díaz & Cabido 2001; Moretti et al. 2009). The use of species traits and functional attributes in analyses have been widely used in plant ecology and are increasingly being used in ecological studies for animal taxa, including insects and other invertebrates (Willby et al. 2000; Moretti & Legg 2009; Vandewalle et al. 2010;

Simons et al. 2016).

Forests support a large portion of global biodiversity, with the majority of forests recovering from past disturbance events (World Commission on Forests and Sustainable

Development 1999; Lindenmayer et al. 2006; Bonan 2008; Pan et al. 2011). Disturbances vary in type and severity and can drastically alter species traits assemblages in a system

124

(Pickett & White 1985; Mouillot et al. 2013). One type of anthropogenic disturbance, clear-cutting, causes an abrupt change from a forested to an open habitat. Removal of the forest canopy leads to substantial changes in ground conditions including increased insolation, increased soil and air temperatures, larger fluctuation between minima and maxima temperatures, aridity and increased wind flow at ground level (Chen et al. 1993,

1999). After a clear-cutting event, changes in leaf litter, soil pH, organic and nutrient matter also affect ground dwelling invertebrate communities (Thiele 1977; Bazzaz 1979;

Covington 1981; Christensen & Peet 1984; Heliövaara & Väisänen 1984; Swank & Vose

1988; Niemelä 1997).

With approximately 40,000 species, Carabidae is one of the most diverse families of beetles (Coleoptera) (Erwin 1985). Carabids play ecologically important roles in communities and are relatively well known biologically and ecologically (Thiele 1977;

Erwin 1996; Lövei & Sunderland 1996). Their ground-dwelling lifestyle and sensitivity to surroundings make them effective model organisms for ecological and environmental changes (Lövei & Sunderland 1996; Niemelä et al. 2000; Rainio & Niemela 2003; Pearce

& Venier 2006; Koivula 2011; Kotze et al. 2011). Numerous studies have documented the relationships between carabid beetle physical attributes (morphology), habit and ecology (Erwin 1979, 1981; Evans 1985, 1994; Forsythe 1987; Bauer et al. 1998; Ribera et al. 1999a, 1999b, 2001; Konuma & Chiba 2007). Changes in Carabidae species traits including physical attributes, trophic level and life history characteristics have been incorporated into investigations of species diversity and assemblages as a function of disturbance or across habitat types (Erwin 1981; Blake et al. 1994; Ribera et al. 2001;

Cole et al. 2002; Sadler et al. 2006; Elek & Lövei 2007; Lambeets et al. 2008, 2009;

125

Pizzolotto 2009; Vandewalle et al. 2010; Schirmel et al. 2011, 2012; Bargmann et al.

2016), but few studies have examined the effects of disturbance on carabid species traits and functional diversity in temperate forests (Erwin 1981).

The environmental changes associated with clear-cutting and forest regeneration affect species richness, diversity and assemblage patterns of carabid beetles. Since forest age classes have been found to be characterized by distinct taxonomic carabid assemblages (Szyszko 1990; Niemelä et al. 1993; Spence et al. 1996; Szyszko et al. 1996;

Butterfield 1997; Koivula et al. 2002; Magura et al. 2006; Riley & Browne 2011), changes in species trait patterns with forest age would also be expected. The objective of this study is to test this assumption, identify characteristic species traits, and detect shifts in trait assemblages among four forest age classes.

Since mature forests typically have higher habitat stability (e.g. more stable abiotic ground conditions) compared to recent clear-cuts (Covington 1981; Swank &

Vose 1988; Chen et al. 1993), I predict that there will be a higher proportion of flight- capable species than flight-restricted species in younger forests (Southwood 1977, 1988;

Thiele 1977; Niemelä et al. 1993; Koivula et al. 2002; Jelaska et al. 2011). The occurrence of winged and non-winged carabid species is not random; it can be linked with locality as well as habitat stability (Darlington 1943; Southwood 1977; Erwin 1979; den Boer et al. 1980; Roff 1990). Temporary habitats typically contain a higher number of macropterous species (fully wings), with higher vagility, while brachypterous species

(reduced wings), with lower dispersal power, have been positively correlated with persistent, stable habitats (Darlington 1943; Southwood 1977, 1988; Kavanaugh 1985;

Roff 1990; Wagner & Liebherr 1992; Lövei & Sunderland 1996). I also predict an

126

inverse relationship between Carabidae body size and habitat disturbance, as has been reported in several studies (Šustek 1987; Blake et al. 1994; Niemelä et al. 2002; Magura et al. 2006; Sadler et al. 2006; Elek & Lövei 2007; Vandewalle et al. 2010; Jelaska et al.

2011; Birkhofer et al. 2015), including habitat succession specifically (Szyszko 1990;

Szyszko et al. 2000; Rainio & Niemela 2003; Schwerk & Szyszko 2007; Schirmel et al.

2012). Invertebrate body size often reflects physiological capabilities such as resistance to desiccation and starvation and dispersal ability (Šustek 1987; Cushman et al. 1993;

Entling et al. 2010). Older forests typically have more stable environments and resource availability, habitat characteristics which I predict will be associated with larger carabids since they typically have longer developmental periods (Southwood 1977; Blake et al.

1994; Lövei & Sunderland 1996). Since the diversity of herbaceous plants and grasses should be higher in recent clear-cuts compared to more mature forests, increasing the availability of seeds and plant material (Peet & Christensen 1987), I predict there will be a higher number of carabid species which rely on these as food sources (i.e. Harpalini,

Zabrini tribes) in recently clear-cut areas (Lindroth 1961-1969; Thiele 1977; Ribera et al.

2001). In addition, since animals with specialized diets are typically found in more stable habitat conditions, species with specialized modes of nutrition may be higher in mature forest age classes. Lastly, since Vandewalle et al. (2010) reported a positive correlation with carabid body pubescence and habitat disturbance, I hypothesize there will be a higher number of pubescent species in clear-cut habitats compared to more forested habitats. In summary, the objectives of this study are to examine the changes in

Carabidae among four forest age classes for (1) taxonomic richness and diversity; (2)

127

functional richness and diversity as indicated by multivariate indices and (3) shifts in ecological species traits.

METHODS

Collection of Carabidae

Carabid beetles were sampled from 34 sites within Stokes and Surry counties in the northern Piedmont region of North Carolina (Figure 1). Twenty-two sampling sites were located within Pilot Mountain State Park (PMSP) either in the two larger continuous sections of the park (Mountain or Yadkin River sections), or along the approximately 10 km section of the Corridor/Bridle Trail (length = 10.6 km) which connects the larger sections. Four forest age classes are represented: recently clear-cut (n = 9), 10 – 20 year old forest (n = 8), 40 – 60 year old forest (n = 7), and 80 – 95 year forest (n = 10), which are referred to as zero, ten, fifty and ninety age groups, respectively. Forest ages were estimated with historical records, tree girth and forest composition and structure. Of the

34 sampling sites, 29 were sampled for three years from May 2009 through April 2012.

Six zero year sites were located within a 65 ha plot which was logged in December 2008

(see Figure 1). An additional year of sampling, from November 2012 through October

2013, was implemented at seven sites in order to improve the spatial distribution for the younger age classes. Two of these extra year sites (one zero year and one ten year site) were sampled during the original three year period. Newly added zero year sites were clear-cut between Nov. 2011 and Nov. 2012. Site selection was affected by availability of forest sites of the proper age class, urban surroundings and collection permits. Sampling

128

sites were similar in altitude (258 – 371 m) and had the same soil classification (i.e.,

Metagraywacke and Muscovite – Biotite Schist (Brown 1985).

Pitfall trapping is useful to compare species presence and the relative activity/abundance levels of larger ground dwelling beetles across habitat types (Thiele

1977; Spence & Niemelä 1994; Woodcock 2005). Although pitfall trap catches have inherent biases relating to activity (‘activity density’) (Greenslade 1964; Luff 1975;

Thiele 1977; Adis 1979; Spence & Niemelä 1994), sampling under standard conditions over the entire active season provides reliable information on activity densities and allows for quantitative comparison among species (Baars 1979; Luff 1982; Loreau 1992;

Southwood & Henderson 2000). This passive sampling approach samples continuously which minimizes investigator associated biases (Melbourne 1999; Woodcock 2005).

Pitfall traps consisted of a plastic drinking cup (16 oz. Solo® cup with 12 cm diameter opening) that were positioned flush with the ground. A flat, dense foam plastic cover to keep rain and other debris out of the trap was loosely placed over each trap and fixed approximately 5 cm above the ground using nails and the traps contained a non-toxic antifreeze preservative. Each sampling site had a total of 12 traps, divided into two groups of six traps (arranged in two small triangles) with the two groups located approximately 10 m apart (see Riley 2010). The trap contents were strained through a fine mesh screen monthly, with Carabidae preserved in 95% ethanol. Species identifications were made using the morphological key by Ciegler (2000), with confirmation made as needed with specimens at the Smithsonian National Museum of

Natural History in Washington, DC.

129

Species traits

For more commonly collected species, seven ecological species traits (Table 1) were selected to reflect important features of Carabidae physical attributes, life history and ecology and have been used in previous trait-based studies (Lindroth 1960-1969;

Thiele 1977; Erwin 1979, 1981; Ribera et al. 2001; Cole et al. 2002; Schirmel et al. 2012).

Species were classified into 18 nominal categories for the seven species traits by specimen examination, or based on information provided in the Carabidae literature

(Lindroth 1960-1969; Thiele 1977; Erwin 1981; Larochelle & Larivière 2003). For each collection year, 12 individuals were assessed or measured (if n < 12 all available specimens were used), with approximately equal representation of males and females and forest age classes when possible. When species data was unavailable or inconclusive for a trait, the species was excluded from the analysis. For every fully intact specimen in the study, Carabidae apparent body length (ABL) was measured to the nearest tenth of a millimeter. ABL is the length from the extreme point of the mandible to apex of elytra. It is a standard carabid beetle measure for ecological studies which provides a reliable estimate of overall size (Erwin & Kavanaugh 1981; Szyszko 1983). Since all other traits were categorical, average length was used to classify species into four size class categories, although numeric body length values were also analyzed. For wing morphology, individuals with hind-wings extending the length of the abdomen were classified as macropterous (with the potential for flight) and species with reduced hind- wings (shorter than the length of the abdomen), or fused elytra and/or no obvious wings were classified as brachypterous (incapable of flight). While the presence of long wings does not guarantee the ability to fly because there may be atrophy of thoracic flight

130

muscles (Thiele 1977), it is commonly used as evidence of dispersal capability (den Boer

1970, 1990). The flight capability of long-winged species was reinforced using the available literature (Lindroth 1961-1969; Erwin 1981; Larochelle & Larivière 2003). For dimorphic-winged species, all specimens were examined and the species was classified according to the dominant wing state. A chi-square test was used to determine if species were significantly more abundant in the spring and summer months (March – August), or the autumn and winter months (September – February). For species where less than ten individuals were collected, the literature was consulted for typical annual activity periods, or excluded from the analysis if species specific information was lacking in the literature.

The adult dietary preference for carabid species was based on the literature (Lindroth

1961-1969; Thiele 1977; Erwin 1981; Larochelle & Larivière 2003). See Table 1 and

Appendix A (Tables A1 & A2) for the full list of ecological species traits and the number of carabid species in each category.

Statistical analysis

To reduce the effect of vagrant and/or rare species and to focus on the traits of species more abundant in the community, only species captured at three or more of the 34 sampling sites (based on all years combined) were included in the analysis. Pitfall trap collections were combined by site each month, with the raw number of individuals adjusted for differences in number of trap nights between collection periods and number of intact traps (versus traps overturned from a disturbance). Total number of individuals collected at a site was standardized to 365 trap nights per month (12 traps x 30.42 days).

For each sampling site and year, the number of individuals (N), species richness (S) and the inverse Simpson’s concentration index (1/D), a Hill diversity number (Simpson 1949;

131

Jost 2006; Chao et al. 2014) were determined. Three multivariate functional diversity indices (FD) were used to examine Carabidae abundance and distribution in functional space across the forest age classes: functional richness (FRic), functional evenness (FEve) and functional dispersion (FDis) (Mason et al. 2005; Villéger 2008; Laliberté & Legendre

2010). For categorical data, FRic is simply the number of unique trait combinations.

FEve weighs the number of unique functional types by abundance to determine the functional evenness across communities and decreases when abundance is lower among species or functional distances are less regular among species (Villéger 2008; Schleuter et al. 2010). FDis expresses functional diversity which decreases as the dominance of single species increases. It is the mean distance in multidimensional space (dispersion) of a given species to the centroid of all other species, with species weighted by relative abundance. Unlike FRic, it is unaffected by species richness and not strongly subject to the effects of outliers. High values correspond to a high number of functionally different species with approximately even abundances (Laliberté & Legendre 2010).

Generalized linear mixed models (GLMMs) with Poisson error distributions were used to assess changes among the forest age classes for species richness, diversity, functional diversity indices and species trait categories. For each species trait category, the number of species per site by year was used as the response variable. Forest age and sampling year were used as predictors (fixed effects) with sampling site as a random factor to account for repeated sampling (Zuur et al. 2009). Although forest age was the main interest, year was included as a fixed effect due to the low number of factor levels

(i.e. 3-4 years) (Bolker et al. 2009; Gotelli & Ellison 2013). If GLMM models failed to meet assumptions (mainly heteroscedasticity) then a non-parametric Kruskal-Wallis rank

132

sum test was used. A GLMM with a negative-binomial error distribution, log link and sampling site as a random factor was more appropriate to assess changes in numbers of individuals among forest age classes (O’Hara & Kotze 2010). FEve values were continuous on the interval (0 – 1), so a beta-regression model with logit link function was used using package ‘betareg’ (Ferrari & Cribari-Neto 2004; Cribari-Neto & Zeileis 2010).

Lastly, FDis across forest age classes was assessed using a GLMM with Gaussian error structure and identity link function and sampling site as a random factor. Numeric body length values were used to construct relative frequency histograms using abundance data in order to examine the body size distribution for each forest age class. Normality of body size distributions were tested using Shapiro-Wilk tests. For the body size distribution with all age classes combined an Anderson-Darling normality test was used (since N >

5000). A relative frequency histogram of the number of species for body length with all age classes combined was also constructed with normality tested using a Shapiro-Wilk test. Mean body length among forest age classes was analyzed using a linear mixed model (LMM) with sampling site as random factor.

Statistical inferences were based on the frequentist (i.e. based on p-values) and information-theoretic approaches. This combined the strengths of two statistical paradigms including the familiarity with p-value based conclusions and the multitude of information provided by information-theoretic methods (Bolker et al. 2009; Burnham et al. 2011; Murtaugh 2014). Model selection was inferred by Akaike information criterion corrected for small sample size (AICc) (Sugiura 1978; Hurvich & Tsai 1989; Burnham et al. 2011). Linear model sets were evaluated, ranked and compared using the following criteria: Δ AICc (difference in the lowest AICc and another model in the set), model

133

weight (normalized probability of a model in the data set) and an evidence ratio (measure of the probability of a given model relative to other models in the set) (Burnham et al.

2011). Significance and p-values were determined through analysis of deviance using

Wald chi-square test statistics (Zuur et al. 2009). All statistical analyses were conducted using R 3.3.0 (R Core Team 2016) including packages ‘lme4’ (Bates et al. 2015), ‘car’

(Fox & Weisberg 2010), ‘MuMIn’ (Barton 2016) and ‘multcomp’ (Hothorn et al. 2008).

RESULTS

Number of individuals and diversity

For species collected from at least three sites over the entire study, there were a total of 6,489 Carabidae individuals representing 13 tribes, 27 genera and 48 species

(Appendix A, Table A1). The removal of rare species resulted in the exclusion of 29 species (62 individuals), the majority of which were from zero year forests (S = 15) and ninety year old forest (S = 10). See Appendix B for parameter model sets for the information-theoretic approach and Appendix C for frequentist results. There were no significant differences in mean number of individuals, richness or the inverse Simpson’s concentration values (1/D) among the four forest age classes (Tables 2 & 3; Appendix A,

Table A3) nor was age included in the top model for any of the candidate model sets.

However, forest age was included in the second most likely model of the set for all three measures. Mean number of individuals was significantly lower for year three compared to all other years (Tables 2 & 3), and species richness was also significantly lower for year three compared to the previous two years (Tables 2 & 3). See Appendix D (Table

D1) for differences among sampling years; however, since inter-annual differences are

134

addressed in Chapter II of this dissertation I will focus on differences among forest age classes in this chapter. There were no significant differences for any of the functional diversity (FD) indices among forest age classes (Figure 2). Forest age was not included in the top model for any indices but it was in the second best model for FRic and FEve

(Appendix B, Tables B3 & B4).

Ecological species traits

Forest age was included in the top model and was statistically significant (p <

0.05) for six of 18 species trait categories (Figure 3; see Appendices B & C for full statistical results). A significantly higher mean number of large species was collected from zero year areas compared to ten year forests. Mean species numbers for the large size category were lower for ten and ninety year forest compared to zero and fifty year forests. Carabid beetle body lengths ranged from 5.0 mm to 35.0 mm. Body length frequency histograms showed that relative frequencies of body lengths varied with forest age (Figure 4) but mean body lengths (log10) were not significantly different among forest age classes. Body length distributions for all forest age classes were non-normal

(zero: W = 0.90, p < 0.001; ten: W = 0.91, p < 0.001; fifty: W = 1.00, p < 0.001; ninety:

W = 0.94, p < 0.001), as was the relative frequency distribution for all forest age classes combined (A = 42.35, p < 0.001). For the zero year age class, > 75% of individuals were between 8.1 – 10.0 mm and 14.1 – 20.0 mm, and compared to all other forest age classes, the zero age class had the highest proportion of individuals between 8.1 – 10.0 mm. More than 75% of individuals had body lengths between 12.1 – 20.0 mm for ten and ninety year old forests, and for fifty year old forests > 75% of individuals between 10.1 – 20.0 mm (Figure 4). The zero year age class showed a bimodal distribution curve, with lower

135

frequency of body lengths between 10.1 – 14.0 mm. For all forest age classes combined, body lengths of 75% of individuals were between 12.1 – 20.0 mm (Figure 5A) and 75% of species were between 8.1 – 18.0 mm and 20.1 – 22.0 mm (Figure 5B).

Both of the wing morphology trait categories demonstrated distinct trends with forest age. Macropterous species number was significantly higher for the zero year age class compared to all other forest age classes (Figure 3). Conversely there was a significantly higher number of brachypterous species for ninety year forests compared to zero year forests (Figure 3). The number of species with a mixed diet was significantly higher for zero year forests than the other age categories and forest age was the only fixed effect in the top model, predicted to be 22.1 times more likely than the next best model

(Figure 3; Appendix B; Table B14). The number of species with pale legs was significantly higher for zero year forests compared to ninety year forests and close to significantly lower for ten and fifty year forests (Figure 3). For the pale leg trait category, forest age was the only fixed effect in the highest ranked model (approximately 2.7 times more likely than the next best model; Appendix B, Table B19). For the pubescence trait, the number of glabrous species collected in ninety year forest was significantly lower than zero year forests (Figure 3). However, there was not a clear trend with forest age as zero year and fifty year forests had higher mean numbers of glabrous species compared to ten year and fifty year forests. For the glabrous species, forest age and year were in the top model, with an evidence ratio suggesting it was approximately 1.6 times more likely than the next best model, which only included year (Appendix B, Table B22).

Two trait categories which have been reported to be correlated with forest succession in the literature, demonstrated trends with forest age in this study and were

136

close to statistically significant (p < 0.1). As forest age increased, the number of species with their main period of annual activity in the autumn and winter months decreased

(Figure 6). The mean number of gastropod specialist (snails and slugs) species was lowest in zero forests compared to the other three age groups (Figure 6). For both autumn/winter seasonal activity and specialist diet, the second best model in the candidate model set included forest age with the Δ AICc (0.26) and evidence ratio both suggesting the differences between the two best models is low (Appendix B; Table B11

& B15).

DISCUSSION

Numbers of individuals, species richness and diversity

Changes in the number of individuals, species richness and the Simpson’s concentration index (1/D) were similar across the four forest age classes with slightly higher average values in year zero and fifty year forest classes. The similarity in these three measures indicates any changes in functional diversity indices and ecological species traits among forest age classes are likely related to shifts in the carabid beetle community composition rather than differences in numbers, richness and diversity.

Functional richness and diversity

Previous carabid beetle studies have found significant differences in multidimensional functional diversity indices across habitat types or disturbance classes

(Vandewalle et al. 2010; Woodcock et al. 2010; Schirmel et al. 2012; Schirmel &

Buchholz 2013; Birkhofer et al. 2015) as have other invertebrate studies across various disturbance gradients (Moretti et al. 2009, 2010). In contrast, the functional diversity

137

indices in this study were not significantly different among age classes, nor were those reported by Gossner et al. (2013) for a forest management study in European beech forests. For this study, the fact that all age classes were forest habitats dominated by trees, with the trees removed only recently from the recent clear-cuts, may account for the insignificant differences among the functional diversity indices. Previous studies which have found greater disparity among functional diversity indices typically examined gradients with more distinct land-use types (e.g. agricultural fields versus mature woodland) (Vandewalle et al. 2010; Schirmel et al. 2012; Birkhofer et al. 2015) rather than chronosequential stages of a habitat type. Species replacement of flora and fauna occurs with forest maturity and can include shifts in species life traits without substantially altering the overall species number, diversity, functional richness and functional diversity of a given area (Roscher et al. 2012; Schirmel et al. 2012; Gossner et al. 2013). Since 12 of the 18 trait categories were not statistically different among forest age classes, multidimensional functional diversity indices which combine several traits into one value were correspondingly not appreciably different. Subtle changes or opposing trends in individual single traits can be masked by FD indices (Roscher et al.

2012; Gossner et al. 2013). As with richness and diversity indices, removal of rare species may also have caused greater convergence for FD indices compared to results utilizing the full data set.

Ecological species traits

The most pronounced differences in species traits were between the zero year forests and the older forests, with carabid species from the zero age class more likely to be macropterous (long-winged), have pale legs and a mixed diet. Brachypterous (reduced

138

wings, incapable of flight) species number increased significantly with forest age. Other trends which were close to statistical significance (p < 0.1) suggest that species in recent clear-cuts are more likely to have higher activity in autumn and winter months (Sept. –

Feb.) and not to be gastropod specialist feeders.

Changes in body length (size) categories among forest age classes did not support predicted hypotheses. Large body size in less disturbed habitats has been one of the most distinct traits reported in other carabid beetle studies (Blake et al. 1994; Spence et al.

1996; Szyszko et al. 2000; Ribera et al. 2001; Cole et al. 2002; Karen et al. 2008; Jelaska et al. 2011; Schirmel et al. 2012). In this study, the large body length category was significantly different among forest age classes but the trend did not clearly shift in one direction with forest age. Ten year forests had the lowest mean number of large species which was affected by the low number of individuals and species collected during sampling year three, particularly for the most abundant ‘large’ species in the ten year age class, Pterostichus sculptus, which was considerable lower in sampling year three. Only eight species were included in the large size category, three of which were among the four most commonly collected species in the study and were present in all forest age classes. The lack of statistical significances between beetle size categories and forest age in certain age classes may be related to the low species numbers for some of the categories resulting in low statistical power in some cases.

Differences in numeric mean body length among forest age classes were also not statistically significant. The bimodal distribution in the zero year age class was due to a lower frequency of individuals between 10.1 – 14.0 mm. Calathus opaculus was the only species with a relatively high number of individuals within this size range and almost half

139

of individuals were in the next smaller size bin (8.1 – 10.0 mm). All other species with body lengths between 10.1 – 14.0 mm occurred in relatively low numbers (relative abundance < 2%), particularly true for individuals between 12.1 – 14.0 mm. The higher frequency of individuals between 8.1 – 10.0 mm is due to the abundances of two Amara species, Notobia terminata and approximately half of C. opaculus individuals. Increase in relative frequencies for individuals between 14.1 – 20.0 mm were due to high abundances of Cyclotrachelus sigillatus, Harpalus pensylvanicus and P. sculptus within the zero year forest class (relative abundances > 10%).

Comparison of the relative frequency histogram for the number of species at

Plummers Island in Maryland by Erwin (1981) and Figure 5B suggests there were fewer species less than 9 mm (especially those < 5 mm) in the current study. In this study there were no species less than 5 mm and only 6.25% of the species had body lengths between

5 – 8 mm. This may be due to differences in sampling techniques as pitfall traps are more efficient in collecting larger carabids (Greenslade 1964; Luff 1975; Thiele 1977). Several of the carabid body size histograms for carabid communities across the globe in Erwin

(1981), including the one for Plummers Island, demonstrated bimodalities in carabid body size distributions which has been hypothesized to be a result of the influence of ants.

Ants have been shown to negatively affect carabids in many ways, particularly through interference competition (Hawes et al. 2002; Reznikova & Dorosheva 2004), leading to negative correlations between abundances of carabid beetles and ants (Heliölä, et al. 2001;

Koivula et al. 2002). For the Chesapeake Bay area in Maryland, the size distribution for the majority of ants is between 3 – 7 mm and there is a corresponding decrease in the frequency of carabid fauna for those size classes (see Erwin 1981, Figure 16). Carabidae

140

body sizes of the zero year class in this study showed a bimodal distribution but the body sizes causing the ‘dip’ in the histogram are from 10.1 – 14.0 mm, a larger size category than the body sizes for the majority of ants from the Chesapeake Bay area reported in

Erwin (1981). Additional information is needed for body sizes of ants occurring in recently clear-cut habitats within the Piedmont of North Carolina to determine if ants play a role in the carabid fauna body size distributions. Since clear-cutting has been shown to lower ant abundance and species richness, as well as alter the community assemblages in deciduous forests of South Carolina (Zettler et al. 2004), the influence of ants may change with forest age. Other taxa with similar lifestyles and habitat preferences

(i.e. spiders, rove beetles) and larger predators (e.g. small mammals and birds) may also influence the size distribution of carabid beetles (Erwin 1981).

As expected, there were significantly higher numbers of macropterous species in clear-cuts and a significant increase in the number of brachypterous species with forest age, which agrees with previous Carabidae studies (Niemelä et al. 1993; Butterfield 1997;

Koivula et al. 2002; Riley & Browne 2011; Birkhofer et al. 2015), and other terrestrial insects along disturbance gradients (Birkhofer et al. 2015; Simons et al. 2016). In general there are more flightless beetles in forested areas than open habitats (Southwood 1962;

Roff 1990; Desender et al. 1999; Karen et al. 2008). In more unstable, early succession habitats macropterous species have superior dispersal power and higher recolonization capabilities, increasing their ability to locate patchy resources and relocate if habitat suitability changes (Šustek 1987; den Boer 1970; Thiele 1977; Roff 1990). These species are often collected in clear-cuts with many adapted to open habitat conditions (Thiele

1977; Niemelä et al. 1993; Koivula et al. 2002).

141

The significantly higher number of species with a mixed (omnivorous) diet in zero year forests may relate to increased habitat heterogeneity which could increase the number of food sources available (Covington 1981; Niemelä et al. 1996; Holt et al. 1999;

Pearce et al. 2003; Rooney et al. 2008). Newly established herbaceous plants and grasses provide an influx of plant-based material (e.g. seeds) for carabids with a mixed diet

(Thiele 1977; Peet & Christensen 1987; Pearce et al. 2003). Species able to consume a wider variety of food items and/or with opportunistic feeding strategies would have an advantage in a more unpredictable, changing environment (e.g. a recent clear-cut).

In the eastern United States, adults of carabid species have characteristic seasonal activity peaks that are consistent among years (Erwin 1981; Luff 1982; Lövei &

Sunderland 1996). In this study, most species (S = 26) exhibited higher activity in the spring/summer months (Mar. – Aug.) while eight species had peak abundance in the autumn/winter months (Sept. – Feb.) which is consistent with other North American studies (Niemelä et al. 1992; French & Elliott 1999; Werner & Raffa 2003; Browne et al.

2014). Annual activity peaks have been associated with reproductive periods, breeding seasons, prey availability and abiotic factors such as rainfall, temperature, etc. (Thiele

1977; Loreau 1985, 1988; Wolda 1988; Niemelä et al. 1992; Lövei & Sunderland 1996;

Ribera et al. 2001). In this study, the number of species active in the autumn/winter decreased with forest age. Previous studies have reported lower numbers of individuals collected in the hotter, drier summer months for some Carabidae species (Erwin 1981;

Niemelä et al. 1992; Spence & Niemelä 1994). This period of lower activity-abundance could possibly be more intense for zero year sites which are characterized by direct sunlight, higher ground temperatures and lower humidity compared to older aged forests

142

with more canopy cover which can have a buffering effect to create more moderate ground conditions (Swank & Vose 1988; Chen et al. 1993). This could be a contributing factor for the higher number of autumn/winter active species in the zero year age class.

The mean number of individuals collected by month varied among forest age classes

(unpubl. data) with the zero age class exhibiting a more distinct biannual abundance peak

(in May – Jun. and Sept. – Oct.), compared to other age classes. In contrast, in Ribera et al. (2001) more disturbed sites had a higher number of summer active species which they attributed to a higher number of spring breeding species in disturbed areas.

Body color and pattern have also been associated with different aged forests.

Erwin (1981) found pale legged carabids were more abundant in riparian and open habitats while species with darker legs dominated forested habitats. Body coloration and color patterns, including legs and metallic sheen, may be related to visibility by predators within a given habitat (Erwin 1979). As with this study, previous research has shown positive associations with pale species (body and legs) and more open and/or disturbed landscapes (Ribera et al. 2001; Vandewalle et al. 2010).

Although Vandewalle et al. (2010) reported a positive correlation with pubescence and disturbance, there was no discernible relationship between pubescence and forest age in this study. Since pubescence was a dichotomous trait, the significant finding for glabrous species may be a reflection of changes in the number of pubescent species with forest age. However, with only seven species classified as pubescent the analysis of this trait has relatively low statistical power.

Trait-based studies have inherent difficulties, including the determination of trait effects on fitness, selection of traits for analysis, inadequate knowledge of species

143

ecology and biology and satisfactory classifications of species into trait categories

(Petchey & Gaston 2006; Villéger 2008; Gagic et al. 2015). A future direction is to incorporate a phylogenetic component to this study to determine which traits are products of relatedness and which occur independently of phylogeny. For a more in depth analysis of adult carabid diet, gut analyses of specimens may provide direct evidence of dietary intake for each species. This study identified significant shifts in select Carabidae species traits among four forest age classes. Characterization of carabid assemblages based on ecological species traits provided insight into functional diversity and shifts in species traits of carabids with forest age, with species traits for recent clear-cuts especially distinct. The results of this study can be compared to Carabidae species traits in temperate forests worldwide. Additionally, some carabid traits (e.g. dispersal ability, body size) may provide insight to the species trait responses of other ground-dwelling invertebrate taxa to the effects of clear-cutting and forest regeneration (Birkhofer et al.

2015; Simons et al. 2016).

ACKNOWLEDGMENTS

Wake Forest University and the Environmental Program of Wake Forest University provided financial support. Many thanks to Matt Windsor for permission to sample

Carabidae within Pilot Mountain State Park, to Terry L. Erwin of the Smithsonian

Institution for access to the Carabidae collection at the National Museum of Natural

History, to the many undergraduates who assisted with field collections, to Erinn Allred,

Bridget Bauman and Jordan Stefko for assisting with species trait classifications, and to

Emily Tompkins for her guidance on information-theoretic analyses.

144

LITERATURE CITED

Adis, J. 1979. Problems of interpreting arthropod sampling with pitfall traps.

Zoologischer Anzeiger 202:177–184.

Baars, M. A. 1979. Catches in pitfall traps in relation to mean densities of carabid beetles.

Oecologia 41:25–46.

Bargmann, T., E. Heegaard, B. A. Hatteland, J. D. Chipperfield, and J. A. Grytnes. 2016.

Species trait selection along a prescribed fire chronosequence. Insect

Conservation and Diversity 9:446-455.

Barton, K. 2016. MuMIn: Multi-Model Inference. R package version 1.15.6. Available

from: https://CRAN.R-project.org/package=MuMIn

Bates D., M. Maechler, B. Bolker, and S. Walker. 2015. Fitting linear mixed-effects

models using lme4. Journal of Statistical Software 67:1 1–48.

Bauer, T., K. Desender, T. Morwinsky, and O. Betz. 1998. Eye morphology reflects

habitat demands in three closely related ground beetle species (Coleoptera:

Carabidae). Journal of Zoology 245:467–472.

Bazzaz, F. A. 1979. The physiological ecology of plant succession. Annual Review of

Ecology and Systematics 10:351–371.

Birkhofer, K., H. G. Smith, W. W. Weisser, V. Wolters, and M. M. Gossner. 2015. Land-

use effects on the functional distinctness of arthropod communities. Ecography 38:

889-900.

Blake, S., G. N. Foster, M. D. Eyre, and M. L. Luff. 1994. Effects of habitat type and

grassland management-practices on the body-size distribution of carabid beetles. 145

Pedobiologia 38:502–512.

Bolker, B. M., M. E. Brooks, C. J. Clark, S. W. Geange, J. R. Poulsen, M. H. H. Stevens,

and J.-S. S. White. 2009. Generalized linear mixed models: a practical guide for

ecology and evolution. Trends in Ecology and Evolution 24:127–135.

Bonan, G. B. 2008. Forests and climate change: forcings, feedbacks, and the climate

benefits of forests. Science 320:1444–1449.

Brown, P. M., and J. M. Parker. 1985. Geologic Map of North Carolina. North Carolina

Geological Society.

Browne, R. A., S. A. Maveety, L. Cooper, and K. N. Riley. 2014. Ground beetle

(Coleoptera: Carabidae) species composition in the southern Appalachian

Mountains. Southeastern Naturalist 13:407–422.

Burnham, K. P., D. R. Anderson, and K. P. Huyvaert. 2011. AIC model selection and

multimodel inference in behavioral ecology: Some background, observations, and

comparisons. Behavioral Ecology and Sociobiology 65:23–35.

Butterfield, J. 1997. Carabid community succession during the forestry cycle in conifer

plantations. Ecography 20:614–625.

Chao, A., N. J. Gotelli, T. C. Hsieh, E. L. Sander, K. H. Ma, R. K. Colwell, and A. M.

Ellison. 2014. Rarefaction and extrapolation with Hill numbers: A framework for

sampling and estimation in species diversity studies. Ecological Monographs

84:45–67.

Chen, J., J. F. Franklin, and T. A. Spies. 1993. Contrasting microclimates among clearcut,

edge, and interior of old-growth Douglas-fir forest. Agricultural and Forest

146

Meteorology 63:219–237.

Chen, J., S. C. Saunders, T. R. Crow, R. J. Naiman, K. D. Brosofske, G. D. Mroz, B. L.

Brookshire, and J. F. Franklin. 1999. Microclimate in forest ecosystem and

landscape ecology. BioScience 49:288–297.

Christensen, N. L., and R. K. Peet. 1984. Convergence during secondary forest

succession. Journal of Ecology 72:25–36.

Ciegler, J. C. 2000. Ground beetles and wrinkled bark beetles of South Carolina

(Coleoptera: Geadephaga: Carabidae and Rhysodidae). Clemson University,

Clemson, SC.

Cole, L. J., D. I. McCracken, P. Dennis, I. S. Downie, A. L. Griffin, G. N. Foster, K. J.

Murphy, and T. Waterhouse. 2002. Relationships between agricultural

management and ecological groups of ground beetles (Coleoptera: Carabidae) on

Scottish farmland. Agriculture, Ecosystems & Environment 93:323–336.

Covington, W. W. 1981. Changes in forest floor organic matter and nutrient content

following clear cutting in northern hardwoods. Ecology 62:41–48.

Cribari-Neto, F., and A. Zeileis. 2010. Beta regression in R. Journal of Statistical

Software 34:1-24.

Cushman, J. H., J. H. Lawton, and B. F. J. Manly. 1993. Latitudinal patterns in European

ant assemblages: Variation in species richness and body size. Oecologia 95:30–37.

Darlington, P. J. 1943. Carabidae of mountains and islands: Data on the evolution of

isolated faunas, and on atrophy of wings. Ecological Monographs 13:37–61. de Bello, F. et al. 2010. Towards an assessment of multiple ecosystem processes and 147

services via functional traits. Biodiversity and Conservation 19:2873–2893. den Boer, P. J. 1970. On the significance of dispersal power for populations of carabid-

beetles (Coleoptera, Carabidae). Oecologia 4:1–28. den Boer, P. J. 1990. Density limits and survival of local populations in 64 carabid

species with different powers of dispersal. Journal of Evolutionary Biology 3:19–

48. den Boer, P. J., T. H. P. van Huizen, W. den Boer-Daanje, B. Aukema, and C. F. M. den

Bieman. 1980. Wing polymorphphism and dimorphism in ground beetles as

stages in an evolutionary process (Coleoptera, Carabidae). Entomologia Generalis

6:107–134.

Desender, K., A. Ervynck, and G. Tack. 1999. Beetle diversity and historical ecology of

woodlands in Flanders. Belgian Journal of Zoology 129:139–155.

Díaz, S., and M. Cabido. 2001. Vive la différence: Plant functional diversity matters to

ecosystem processes. Trends in Ecology and Evolution 16:646–655.

Elek, Z., and G. L. Lövei. 2007. Patterns in ground beetle (Coleoptera: Carabidae)

assemblages along an urbanisation gradient in Denmark. Acta Oecologica

32:104–111.

Entling, W., M. H. Schmidt-Entling, S. Bacher, R. Brandl, and W. Nentwig. 2010. Body

size-climate relationships of European spiders. Journal of Biogeography 37:477–

485.

Erwin, T. L. 1979. Thoughts on the evolutionary history of ground beetles: Hypotheses

generated from comparative faunal analyses of lowland forest sites in temperate

148

and tropical regions. Pages 539–592 in T. L. Erwin, G. E. Ball, and D. R.

Whitehead editors. Carabid Beetles: Their Evolution, Natural History, and

Classification. Dr W. Junk Publishers, Dordrecht.

Erwin, T. L. 1981. Natural history of Plummers Island, Maryland. XXVI. The ground

beetles of a temperate forest site (Coleoptera: Carabidae): An analysis of fauna in

relation to size, habitat selection, vagility, seasonality, and extinction. Bulletin of

the Biological Society of Washington 5:105–224.

Erwin, T. L. 1985. The taxon pulse: A general pattern of lineage radiation and extinction

among carabid beetles. Pages 438–472 in G. E. Ball, editor. Taxonomy,

Phylogeny and Zoography of Beetles and Ants. Dr. W. Junk Publishers,

Dordrecht.

Erwin, T. L. 1996. Biodiversity at its utmost: tropical forest beetles. Pages 27–40 in M. L.

Reaka-Kudla, D. E. Wilson, and E. O. Wilson, editors. Biodiversity II:

Understanding and Protecting our Biological Resources. Joseph Henry Press,

Washington, DC.

Erwin, T. L., and D. H. Kavanaugh. 1981. Systematics and zoogeography of Bembidion

Latreille: I. The carlhi and erasum groups of western North America (Coleoptera:

Carabidae, Bembidiini). Entomologica Scandinavica Supplement 15:33–72.

Evans, M. E. G. 1985. Feeding mechanisms, and their variation in form, of some adult

ground beetles (Coleoptera; Caraboidea). Journal of Zoology 206:113–143.

Evans, M. E. G. 1994. The carabid body plan: A functional interpretation. Pages 25–31 in

K. Desender, M. Dufrene, M. Loreau, M. L. Luff, and J. P. Maelfait, editors.

149

Carabid Beetles: Ecology and Evolution. Kluwer Academic Publishers, the

Netherlands.

Ferrari, S. L. P., and F. Cribari-Neto. 2004. Beta regression for modelling rates and

proportions. Journal of Applied Statistics 31:799–815.

Forsythe, T. G. 1987. The relationship between body form and habit in some Carabidae

(Coleoptera). Journal of Zoology 211:643–666.

Fox, J., and S. Weisberg. 2010. An {R} Companion to Applied Regression. 2nd edition.

Sage, Thousand Oaks, CA. Avialable from:

http://socserv.socsci.mcmaster.ca/jfox/Books/Companion

French, B. W., and N. C. Elliott. 1999. Temporal and spatial distribution of ground beetle

(Coleoptera: Carabidae) assemblages in grasslands and adjacent wheat fields.

Pedobiologia 43:73–84.

Gagic, V. et al. 2015. Functional identity and diversity of animals predict ecosystem

functioning better than species-based indices. Proceedings of the Royal Society B

282:20142620.

Gossner, M. M., T. Lachat, J. Brunet, G. Isacsson, C. Bouget, H. Brustel, R. Brandl, W.

W. Weisser, and J. Müller. 2013. Current near-to-nature forest management

effects on functional trait composition of saproxylic beetles in beech forests.

Conservation Biology 27:605–614.

Gotelli, N. J., and A. Ellison. 2013. A Primer of Ecological Statistics. 2nd edition.

Sinauer Associates, Inc. Sutherland, Massachusetts.

Greenslade, P. J. M. 1964. Pitfall trapping as a method for studying populations of

150

Carabidae (Coleoptera). Journal of Animal Ecology 33:301–310.

Hawes, C., A. J. A. Stewart, and H. F. Evans. 2002. The impact of wood ants (Formica

rufa) on the distribution and abundance of ground beetles (Coleoptera: Carabidae)

in a Scots pine plantation. Oecologia 131:612–619.

Heliölä, J., M. Koivula, and J. Niemelä. 2001. Distribution of carabid beetles (Coleoptera,

Carabidae) across a boreal forest-clearcut ecotone. Conservation Biology 15:370–

377.

Heliövaara, K., and R. Väisänen. 1984. Effects of modern forestry on northwestern

European forests invertebrates: A synthesis. Acta Forestalia Fennica 189:1–32.

Holt, R. D., J. H. Lawton, G. A. Polis, and N. D. Martinez. 1999. Trophic rank and the

species-area relationship. Ecology 80:1495–1504.

Hothorn T., F. Bretz and P. Westfall. 2008. Simultaneous inference in general parametric

models. Biometrical Journal 50:346–363.

Hurvich, C. M., and C. L. Tsai. 1989. Regression and time series model selection in small

samples. Biometrika 76:297–307.

Jelaska, L. Š., V. Dumbović, and M. Kučinić. 2011. Carabid beetle diversity and mean

individual biomass in beech forests of various ages. Zookeys 100:393–405.

Jost, L. 2006. Entropy and diversity. Oikos 113:363–375.

Karen, M., J. O’Halloran, J. Breen, P. Giller, J. Pithon, and T. Kelly. 2008. Distribution

and composition of carabid beetle (Coleoptera, Carabidae) communities across

the plantation forest cycle - Implications for management. Forest Ecology and

Management 256:624–632.

151

Kavanaugh, D. H. 1985. On wing atrophy in carabid beetles (Coleoptera: Carabidae),

with special reference to Nearctic Nebria. Pages 408–431 in G. E. Ball, editor.

Taxonomy, Phylogeny and Zoography of Beetles and Ants. Dr W. Junk

Publishers. Dordrecht.

Koivula, M. J. 2011. Useful model organisms, indicators, or both? Ground beetles

(Coleoptera, Carabidae) reflecting environmental conditions. Pages 287–317 in D.

J. Kotze, T. Assmann, J. Noordijk, H. Turin, and R. Vermeulen, editors. Carabid

Beetles as Bioindicators: Biogeographical, Ecological and Environmental Studies.

ZooKeys 100:287–317..

Koivula, M., J. Kukkonen, and J. Niemela. 2002. Boreal carabid-beetle (Coleoptera,

Carabidae) assemblages along the clear-cut originated succession gradient.

Biodiversity and Conservation 11:1269–1288.

Konuma, J., and S. Chiba. 2007. Trade-offs between force and fit: Extreme morphologies

associated with feeding behavior in carabid beetles. American Naturalist 170:90–

100.

Kotze, J. D. et al. 2011. Forty years of carabid beetle research in Europe - From

taxonomy, biology, ecology and population studies to bioindication, habitat

assessment and conservation. ZooKeys 100:55–148.

Laliberté, E., and P. Legendre. 2010. A distance-based framework for measuring

functional diversity from multiple traits. Ecology 91:299–305.

Lambeets, K., M. L. Vandegehuchte, J. P. Maelfait, and D. Bonte. 2009. Integrating

environmental conditions and functional life-history traits for riparian arthropod

152

conservation planning. Biological Conservation 142:625–637.

Lambeets, K., M. L. Vandegehuchte, J. P. Maelfait, and D. Bonte. 2008. Understanding

the impact of flooding on trait-displacement and shifts in assemblage structure of

predatory arthropods on river banks. Journal of Animal Ecology 77:1162–1174.

Larochelle, A., and M. C. Larivière. 2003. A natural history of the ground-beetles

(Coleoptera: Carabidae) of America north of Mexico. Pensoft Publishers, Sofia.

Lindenmayer, D. B., J. F. Franklin, and J. Fischer. 2006. General management principles

and a checklist of strategies to guide forest biodiversity conservation. Biological

Conservation 131:433–445.

Lindroth, C. H. 1961-1969. The ground-beetles (Carabidae, excl. Cicindelinae) of Canada

and Alaska, Parts 1-6. Opuscula Entomologica. Suppl. 20, 24, 29, 33, 34, 35.

Loreau, M. 1985. Annual activity and life cycles of carabid beetles in two forest

communities. Holarctic Ecology 8:228–235.

Loreau, M. 1988. Determinants of the seasonal pattern in the niche structure of a forest

carabid community. Pedobiologia 31:75–87.

Loreau, M. 1992. Species abundance patterns and the structure of ground-beetle

communities. Annales Zoologici Fennici 28:49–56.

Lövei, G. L., and K. D. Sunderland. 1996. Ecology and behavior of ground beetles

(Coleoptera: Carabidae). Annual Review of Entomology 41:231–256.

Luff, M. L. 1975. Some features influencing the efficiency of pitfall traps. Oecologia

19:345–357.

153

Luff, M. L. 1982. Population dynamics of Carabidae. Annals of Applied Biology

101:164–170.

Magura, T., B. Tóthmérész, and G. L. Lövei. 2006. Body size inequality of carabids

along an urbanisation gradient. Basic and Applied Ecology 7:472–482.

Mason, N., D. Mouillot, W. Lee, and J. Wilson. 2005. Functional richness, functional

evenness and functional divergence: the primary components of functional

diversity. Oikos 1:112–118.

Melbourne, B. A. 1999. Bias in the effect of habitat structure on pitfall traps: An

experimental evaluation. Austral Ecology 24:228–239.

Moretti, M., F. De Bello, S. P. M. Roberts, and S. G. Potts. 2009. Taxonomical vs.

functional responses of bee communities to fire in two contrasting climatic

regions. Journal of Animal Ecology 78:98–108.

Moretti, M., M. De Cáceres, C. Pradella, M. K. Obrist, B. Wermelinger, P. Legendre, and

P. Duelli. 2010. Fire-induced taxonomic and functional changes in saproxylic

beetle communities in fire sensitive regions. Ecography 33:760–771.

Moretti, M., and C. Legg. 2009. Combining plant and animal traits to assess community

functional responses to disturbance. Ecography 32:299–309.

Mouillot, D., N. A. J. Graham, S. Villéger, N. W. H. Mason, and D. R. Bellwood. 2013.

A functional approach reveals community responses to disturbances. Trends in

Ecology and Evolution 28:167–177.

Murtaugh, P. A. 2014. In defense of P values. Ecology 95:611–617.

Niemelä, J. 1997. Invertebrates and boreal forest management. Conservation Biology 154

11:601–610.

Niemelä, J. K., Y. Haila, and P. Punttila. 1996. The importance of small-scale

heterogeneity in boreal forests: Variation in diversity in forest-floor invertebrates

across the succession gradient. Ecography 19:352–368.

Niemelä, J., J. R. Spence, and D. Spence. 1992. Habitat assocations and seasonal activity

of ground-beetles (Coleoptera, Carabidae) in Central Alberta. Canadian

Entomologist 124:521–540.

Niemelä, J. K., D. Langor, and J. R. Spence. 1993. Effects of clear-cut harvesting on

Boreal ground-beetle assemblages (Coleoptera, Carabidae) in Western Canada.

Conservation Biology 7:551–561.

Niemelä, J., D. J. Kotze, S. Venn, L. Penev, I. Stoyanov, D. Hartley, and E. M. de Oca.

2002. Carabid beetle assemblages (Coleoptera, Carabidae) across urban-rural

gradients: an international comparison. Landscape Ecology 17:387–401.

Niemelä, J., J. Kotze, A. Ashworth, P. Brandmayr, K. Desender, T. New, L. Penev, M.

Samways, and J. Spence. 2000. The search for common anthropogenic impacts on

biodiversity: A global network. Journal of Insect Conservation 4:3–9.

O’Hara, R. B., and D. J. Kotze. 2010. Do not log-transform count data. Methods in

Ecology and Evolution 1:118–122.

Pan, Y., J. M. Chen, R. Birdsey, K. McCullough, L. He, and F. Deng. 2011. Age

structure and disturbance legacy of North American forests. Biogeosciences

8:715–732.

Pearce, J. L., and L. A. Venier. 2006. The use of ground beetles (Coleoptera: Carabidae)

155

and spiders (Araneae) as bioindicators of sustainable forest management: A

review. Ecological Indicators 6:780–793.

Pearce, J. L., L. A. Venier, J. McKee, J. Pedlar, and D. McKenney. 2003. Influence of

habitat and microhabitat on carabid (Coleoptera: Carabidae) assemblages in four

stand types. Canadian Entomologist 135:337–357.

Peet, R. K., and N. L. Christensen. 1987. Competition and tree death. Bioscience 37:586–

595.

Petchey, O. L., and K. J. Gaston. 2006. Functional diversity: Back to basics and looking

forward. Ecology Letters 9:741–758.

Pickett, S. T. A., and P. S. White, editors. 1985. The Ecology of Natural Disturbance and

Patch Dynamics. Academic Press, Orlando.

Pizzolotto, R. 2009. Characterization of different habitats on the basis of the species traits

and eco-field approach. Acta Oecologica 35:142–148.

R Core Team. 2016. R: Language and Environment for Statistical Computing. R

Foundation for Statistical Computing, Vienna, Austria.

Rainio, J., and J. Niemela. 2003. Ground beetles (Coleoptera: Carabidae) as bioindicators.

Biodiversity and Conservation 12:487–506.

Reznikova, Z., and H. Dorosheva. 2004. Impacts of red wood ants Formica polyctena on

the spatial distribution and behavioural patterns of ground beetles (Carabidae).

Pedobiologia 48:15–21.

Ribera, I., S. Doledec, I. Downie, and G. N. Foster. 2001. Effect of land disturbance and

stress on species traits of ground beetle assemblages. Ecology 82:1112–1129. 156

Ribera, I., G. N. Foster, I. S. Downie, D. I. McCracken, and V. J. Abernethy. 1999a. A

comparative study of the morphology and life traits of Scottish ground beetles

(Coleoptera, Carabidae). Annales Zoologici Fennici 36:21–37.

Ribera, I., D. I. McCracken, G. N. Foster, I. S. Downie, and V. J. Abernethy. 1999b.

Morphological diversity of ground beetles (Coleoptera: Carabidae) in Scottish

agricultural land. Journal of Zoology 247:1–18.

Riley, K. N. 2010. Ground beetles (Coleoptera: Carabidae) as biodiversity indicators for

age structure in Piedmont Forests? Master’s Thesis, Wake Forest University.

Riley, K. N., and R. A. Browne. 2011. Changes in ground beetle diversity and

community composition in age structured forests (Coleoptera, Carabidae).

ZooKeys 621:601–21.

Roff, D. A. 1990. The evolution of flightlessness in insects. Ecological Monographs

60:389–421.

Rooney, N., K. S. McCann, and J. C. Moore. 2008. A landscape theory for food web

architecture. Ecology Letters 11:867–881.

Roscher, C., J. Schumacher, M. Gubsch, A. Lipowsky, A. Weigelt, N. Buchmann, B.

Schmid, and E. D. Schulze. 2012. Using plant functional traits to explain

diversity-productivity relationships. PLoS ONE 7: e36760.

Sadler, J. P., E. C. Small, H. Fiszpan, M. G. Telfer, and J. Niemela. 2006. Investigating

environmental variation and landscape characteristics of an urban-rural gradient

using woodland carabid assemblages. Journal of Biogeography 33:1126–1138.

Schirmel, J., and S. Buchholz. 2013. Invasive moss alters patterns in life-history traits

157

and functional diversity of spiders and carabids. Biological Invasions 15:1089–

1100.

Schirmel, J., L. Timler, and S. Buchholz. 2011. Impact of the invasive moss Campylopus

introflexus on carabid beetles (Coleoptera: Carabidae) and spiders (Araneae) in

acidic coastal dunes at the southern Baltic Sea. Biological Invasions 13:605–620.

Schirmel, J., I. Blindow, and S. Buchholz. 2012. Life-history trait and functional diversity

patterns of ground beetles and spiders along a coastal heathland successional

gradient. Basic and Applied Ecology 13:606–614.

Schirmel, J., and S. Buchholz. 2013. Invasive moss alters patterns in life-history traits

and functional diversity of spiders and carabids. Biological Invasions 15:1089–

1100.

Schleuter, D., M. Daufresne, F. Massol, and C. Argillier. 2010. A user’s guide to

functional diversity indices. Ecological Monographs 80:469–484.

Schwerk, A., and J. Szyszko. 2007. Increase of Mean Individual Biomass (MIB) of

Carabidae (Coleoptera) in relation to succession in forest habitats. Wiadomości

Entomologiczne 26:195–206.

Simons, N. K., W. W. Weisser, and M. M. Gossner. 2016. Multi-taxa approach shows

consistent shifts in arthropod functional traits along grassland land-use intensity

gradient. Ecology 97:754–764.

Simpson, E. H. 1949. Measurement of diversity. Nature 163:688.

Southwood, T. R. E. 1962. Migration of terrestrial arthropods in relation to habitat.

Biological Reviews 37:171–214.

158

Southwood, T. R. E. 1977. Habitat, the templet for ecological strategies? Journal of

Animal Ecology 46:337–365.

Southwood, T. R. E. 1988. Tactics, strategies and templets. Oikos 52:3–18.

Southwood, T. R. E., and P. A. Henderson. 2000. Ecological Methods. 3rd edition.

Blackwell Science Ltd, Oxford, UK.

Spence, J. R., and J. Niemelä. 1994. Sampling carabid assemblages with pitfall traps: The

maddness and the method. Canadian Entomologist 126:881–894.

Spence, J. R., D. W. Langor, J. Niemelä, H. A. Carcamo, and C. R. Currie. 1996.

Northern forestry and carabids: The case for concern about old-growth species.

Annales Zoologici Fennici 33:173–184.

Sugiura, N. 1978. Further analysis of the data by Akaike’s Information Criterion and the

finite corrections. Communications in Statistics - Theory and Methods 7:13–26.

Šustek, Z. 1987. Changes in body size structure of carabid communities (Coleoptera,

Carabidae) along an urbanisation gradient. Biológia (Bratislava) 42:145–156.

Swank, W. T., and J. M. Vose. 1988. Effects of cutting practices on microenvironment in

relation to hardwood regeneration. Pages 71–88 in H. C. Clya, A. W. Perkey, and

W. E. Kidd, editors. Guidelines for Regenerating Appalachian Hardwood Stands:

Proc. Workshop, Morgantown, WV, 24-26 May 1988. Society of American

Foresters Publication 88-03, West Virginia University Books, Morgantown, WV.

Szyszko, J. 1983. Methods of macrofauna investigation. Pages 10–16 in J. Szyszko and A.

Szujecki, editors. The Process of Forest Soil Macrofauna Formation after

Afforestation of Farmland. Warsaw Agricultural University Press, Warsaw,

159

Poland.

Szyszko, J. 1990. Planning of Prophylaxis in Threatened Pine Forest Biocenoses Based

on an Analysis of the Fauna of Epigeic Carabidae. Agricultural University Press,

Warsaw.

Szyszko, J., H. J. W. Vermeulen, and P. J. den Boer. 1996. Survival and reproduction in

relation to habitat quality and food availability for Pterostichus oblongopunctatus

F. (Carabidae, Col.). Acta Jutlandica 71:25–40.

Szyszko, J., H. J. W. Vermeulen, K. Klimaszewski, M. Abs, and A. Schwerk. 2000.

Mean Individual Biomass (MIB) of ground beetles as an indicator of the state of

the environment. Pages 289–294 in P. Brandmayr, editor. Natural History and

Applied Ecology of Carabid Beetles. Pensoft Publishers, Sofia, Moscow.

Thiele, H. U. 1977. Carabid Beetles in Their Environments. A Study on Habitat Selection

by Adaptation in Physiology and Behaviour. Springer-Verlag, Berlin.

Vandewalle, M. et al. 2010. Functional traits as indicators of biodiversity response to

land use changes across ecosystems and organisms. Biodiversity and

Conservation 19:2921–2947.

Villéger, S. 2008. New multidimensional functional diversity indices for a multifaceted

framwork in functional ecology. Ecology 89:2290–2301.

Violle, C., M. L. Navas, D. Vile, E. Kazakou, C. Fortunel, I. Hummel, and E. Garnier.

2007. Let the concept of trait be functional! Oikos 116:882–892.

Wagner, D. L., and J. K. Liebherr. 1992. Flightlessness in insects. Trends in Ecology &

Evolution 7:216–220.

160

Werner, S. M., and K. F. Raffa. 2003. Seasonal activity of adult, ground-occurring

beetles (Coleoptera) in forests of Northeastern Wisconsin and the upper peninsula

of Michigan. American Midland Naturalist 149:121–133.

Whittaker, R. H. 1975. Communities and Ecosystems. MacMillan Publishing, Inc., New

York.

Willby, N. J., V. J. Abernethy, and B. O. L. Demars. 2000. Attribute-based classification

of European hydrophytes and its relationship to habitat utilization. Freshwater

Biology 43:43–74.

Wolda, H. 1988. Insect seasonality: Why? Annual Review of Ecology and Systematics

19:1–18.

Woodcock, B. A. 2005. Pitfall trapping in ecological studies. Pages 37–57 in S. R.

Leather, editor. Insect Sampling in Forest Ecosystems. Blackwell Publishing,

Oxford.

Woodcock, B. A., J. Redhead, A. J. Vanbergen, L. Hulmes, S. Hulmes, J. Peyton, M.

Nowakowski, R. F. Pywell, and M. S. Heard. 2010. Impact of habitat type and

landscape structure on biomass, species richness and functional diversity of

ground beetles. Agriculture, Ecosystems and Environment 139:181–186..

World Commission on Forests and Sustainable Development. 1999. Our forests our

future. Summary Report of the World Commisson on Forests and Sustainable

Development. Winnipeg, Manitoba, Canada.

Zettler, J. A., M. D. Taylor, C. R. Allen, and T. P. Spira. 2004. Consequences of forest

clear-cuts for native and nonindigenous ants (Hymenoptera: Formicidae). Annals

161

of the Entomological Society of America 97:513–518.

Zuur, A. F., E. N. Ieno, N. J. Walker, A. A. Saveliev, and G. M. Smith. 2009. Mixed

Effects Models and Extensions in Ecology with R. M. Gail, K. Krickeberg, J. M.

Samet, A. Tsiatis, and W. Wong, editors. Springer Science+Business Media, LLC,

New York, USA.

162

TABLES

III Table – 1. Seven ecological species traits for Carabidae included in the analyses of functional diversity indices and individual traits. ’Code’ is the trait identification representing the 18 species trait categories. ‘No. of S’ is the number of Carabidae species included for each trait category.

Species trait Code Category No. of S Length (size) LEN small (5-9 mm) 6 medium (9-15 mm) 23

large (15-20 mm) 8

very large (>20 mm) 11

Wing morphology WIN brachypterous 22 macropterous 26

Seasonality SEA spring/summer (SprSum) 26 autumn/winter (AutWin) 9

Diet of adult* DIET specialist 8 carnivore 23 omnivore (mix) 10 herbivore 5 Color of body BOD non-metallic 15 metallic 33

Color of leg LEG dark 29 pale 18

Pubescence of PUB glabrous 41 body pubescent 7 * indicates data based on literature. Main sources (Lindroth 1961-1969; Erwin 1981; Larochelle A, Larivière 2003).

163

III Table – 2. Mean (± s.e.) numbers of individuals (N), species richness (S) and

Simpson’s concentration index (1/D) for four forest age classes. Means for zero and ten year old forests include an additional year of sampling. Values are based on sampling site per year (zero, n = 22; ten, n = 21; fifty, n = 21; ninety, n = 30).

Measure Zero Ten Fifty Ninety N 76.9 ± 14.4 64.3 ± 8.0 74.5 ± 11.0 62.8 ± 5.6 S 13.1 ± 1.0 10.6 ± 1.0 12.4 ± 0.9 11.0 ± 0.5 1/D 5.25 ± 0.73 4.15 ± 0.23 5.25 ± 0.31 4.67 ± 0.25

164

III Table – 3. Analysis of deviance results (frequentist approach) of each response variable for forest age and year (N = 94, df = 3). Response variables include number of individuals (N), raw species number (S) and inverse Simpson’s concentration diversity index (1/D). Significance was determined through analysis of deviance using Wald chi- square test statistics (χ2) or Kruskal-Wallis rank sum test (H).

Fixed effect Response Age Year N 1.29, p = 0.731 49.5, p < 0.001*** S 3.5, p = 0.316 21.9, p < 0.001*** 1/D H = 6.5, p = 0.09 H = 3.1, p = 0.380

165

FIGURES

III Figure – 1. Map of the 34 sampling sites in the North Carolina Piedmont with forest age classes depicted by different point markers. Sites with * indicate where sampling occurred for an additional year.

166

III Figure – 2. Functional diversity indices (mean ± s.e.) for the forest age classes: functional richness (FRic), functional evenness (FEve) and functional dispersion (FDis).

All species traits are included in the analyses. No significant differences occurred for any of the functional diversity indices with forest age.

167

III Figure – 3. Number of Carabidae species (mean ± s.e.) for six ecological species trait categories among the four forest age classes. Significance levels among forest age classes are indicated by *p < 0.05, **p < 0.01, ***p < 0.001.

168

III Figure – 4. Relative frequency histograms of Carabidae body length for the four forest age classes. Distributions for all forest age classes were non-normal (p < 0.001).

169

III Figure – 5. Relative frequency histograms of Carabidae body length with all forest age classes combined. (A) Body length distributions for both the abundance-based curve (p <

0.001) and (B) the species-based curve (p = 0.04) were non-normal.

A B

170

III Figure – 6. Numbers of Carabidae species (mean ± s.e.) for two species traits which were close to statistical significance (p < 0.1). These two trait categories have shown significant trends with forest disturbance in previous studies.

171

APPENDIX A

III Table – A1. Scientific names, species codes and tribes for the 48 species in this study.

Species Species Tribe code Agonum punctiforme Fabricius Agopun Platynini Amara crassispina LeConte Amacra Zabrini Amara cupreolata Putzeys Amacup Zabrini Amara familiaris (Duftschmid) Amafam Zabrini Amara impuncticollis (Say) Amaimp Zabrini Amphasia interstitialis (Say) Ampint Harpalini Anisodactylus furvus LeConte Anifur Harpalini Anisodactylus haplomus Chaudior Anihap Harpalini Anisodactylus harrisii LeConte Anihar Harpalini Anisodactylus rusticus (Say) Anirus Harpalini Apenes lucidulus (Dejean) Apeluc Lebiini Calathus opaculus LeConte Calopa Platynini Carabus goryi Dejean Cargor Carabini Carabus sylvosus Say Carsyl Carabini Chlaenius aestivus Say Chlaes Chlaenini Chlaenius amoenus Dejean Chlamo Chlaenini Chlaenius emarginatus Say Chlema Chlaenini Cicindela sexguttata Fabricius Cicsex Cicindelini Cyclotrachelus freitaga Bousquet Cycfre Pterostichini Cyclotrachelus sigillatus (Say) Cycsig Pterostichini Cymindis americanus (Dejean) Cymame Lebiini Dicaelus ambiguus LaFerté-Sénecteré Dicamb Licinini Dicaelus dilatatus Say Dicdil Licinini Dicaelus elongatus Bonelli Dicelo Licinini Dicaelus politus Dejean Dicpol Licinini Dicaelus purpuratus Bonelli Dicpur Licinini Galerita bicolor Dry Galbic Galeritini Harpalus katiae Battoni Harkat Harpalini Harpalus pensylvanica De Geer Harpen Harpalini Myas coracinus (Say) Myacor Pterostichini Notiobia terminata (Say) Notter Harpalini Notiophilus aeneus (Herbst) Notaen Notiophilini Pasimachus punctulatus Haldeman Paspun Scaritini

172

Platynus decentis (Say) Pladec Platynini Poecilus lucublandus Bonelli Poeluc Pterostichini Poecilus morphospecies a Poemora Pterostichini Poecilus morphospecies b Poemorb Pterostichini Pterostichus adoxus (Say) Pteado Pterostichini Pterostichus coracinus (Newman) Ptecor Pterostichini Pterostichus moestus (Say) Ptemoe Pterostichini Pterostichus sculptus LeConte Ptescu Pterostichini Rhadine caudata (LeConte) Rhacau Platynini Scaphinotus andrewsii Valentine Scaand Cychrini Scaphinotus viduus (Dejean) Scavid Cychrini Scarites subterraneus Fabricius Scasub Scaritini Selenophorus opalinus (LeConte) Selopa Harpalini Sphaeroderus stenostomus (Weber) Sphste Cychrini Tetraca virginica (Linnaeus) Tetvir Cicindelini

173

III Table – A2. Species classifications for the 18 species trait categories. ‘NA’ means data was not available or could not be determined based on the literature. Abbreviations are as follows: length (LEN); seasonality (SEA), AutWin = main annual active period between

Sep. – Feb. SprSum = main annual active period between Mar. – Aug.; wing morphology

(WIN), macro = macropterous, long winged, brachy = brachypterous, short winged; diet

(DIET), carn = carnivore, herb = herbivore, mix = omnivore, spec = specialist; body

(BODY), metal = metallic, nonmetal = non-metallic; pubescence (PUB), glab = glabrous body, pub = pubescent body.

Species code LEN SEA WIN DIET BODY LEG PUB Agopun small AutWin macro carn nonmetal pale glab Amacra small SprSum macro herb metal pale glab Amacup small NA macro herb metal pale glab Amafam small SprSum macro herb nonmetal pale glab Amaimp small SprSum macro herb metal pale glab Ampint med SprSum macro herb nonmetal pale pub Anifur med SprSum macro mix nonmetal dark glab Anihap med NA macro mix nonmetal dark glab Anihar med SprSum macro mix nonmetal dark glab Anirus med SprSum macro mix nonmetal dark glab Apeluc med NA macro NA nonmetal pale glab Calopa med AutWin macro carn nonmetal pale glab Cargor verylarge SprSum brachy carn nonmetal dark glab Carsyl verylarge AutWin brachy carn nonmetal dark glab Chlaes large SprSum brachy mix metal pale pub Chlamo med NA macro mix metal pale pub Chlema med SprSum macro mix metal pale pub Cicsex med SprSum macro carn metal dark pub Cycfre med SprSum brachy carn nonmetal dark glab Cycsig large SprSum brachy carn nonmetal dark glab Cymame med NA brachy carn nonmetal pale pub Dicamb verylarge SprSum brachy spec nonmetal dark glab Dicdil verylarge SprSum brachy spec nonmetal dark glab Dicelo large SprSum brachy spec nonmetal dark glab Dicpol med SprSum brachy spec nonmetal dark glab 174

Dicpur verylarge SprSum brachy spec nonmetal dark glab Galbic verylarge SprSum macro carn nonmetal pale pub Harkat verylarge SprSum macro mix nonmetal dark glab Harpen large AutWin macro mix nonmetal pale glab Myacor med AutWin brachy NA metal dark glab Notaen small NA macro carn metal pale glab Notter med AutWin macro mix nonmetal pale glab Paspun verylarge SprSum brachy carn nonmetal dark glab Pladec med SprSum brachy carn nonmetal dark glab Poeluc med NA macro carn metal dark glab Poemora med SprSum macro carn nonmetal dark glab Poemorb med NA macro carn metal dark glab Pteado med AutWin brachy carn nonmetal dark glab Ptecor large NA brachy carn nonmetal dark glab Ptemoe large SprSum brachy carn nonmetal dark glab Ptescu large AutWin brachy carn nonmetal dark glab Rhacau med NA brachy carn nonmetal NA glab Scaand verylarge AutWin brachy spec metal dark glab Scasub large SprSum macro carn nonmetal dark glab Scavid verylarge NA brachy spec metal dark glab Selopa med NA macro carn nonmetal pale glab Sphste med NA brachy spec metal dark glab Tetvir verylarge SprSum macro carn metal pale glab

175

III Table – A3. Number of individuals for each species collected from the four forest age classes. Only species collected at three or more of the 34 sampling sites (based on all years combined) were included in the analysis.

Species code Zero Ten Fifty Ninety Total Agopun 77 77 Amacra 24 1 25 Amacup 18 2 20 Amafam 13 3 6 1 22 Amaimp 30 7 6 44 Ampint 5 2 3 6 16 Anifur 16 16 Anihap 16 1 1 18 Anihar 32 1 5 38 Anirus 34 2 1 38 Apeluc 3 4 4 11 Calopa 113 38 138 27 316 Cargor 1 4 19 54 78 Carsyl 8 2 11 3 24 Chlaes 73 60 31 94 258 Chlamo 2 2 3 7 Chlema 3 11 1 7 22 Cicsex 1 9 6 14 30 Cycfre 2 24 105 125 256 Cycsig 300 108 193 467 1068 Cymame 1 3 6 6 16 Dicamb 5 88 62 89 243 Dicdil 8 35 142 30 214 Dicelo 64 9 24 2 99 Dicpol 4 30 10 42 85 Dicpur 6 2 4 12 Galbic 7 27 17 113 164 Harkat 37 1 38 Harpen 270 11 57 13 352 Myacor 23 5 28 Notter 42 42 Notaen 2 6 8 Paspun 45 11 9 38 103

176

Pladec 1 21 11 33 Poeluc 4 4 1 9 Poemora 9 3 10 2 24 Poemorb 2 4 1 8 Pteado 14 311 7 35 366 Ptecor 49 27 21 110 207 Ptemoe 1 27 27 Ptescu 199 154 425 102 880 Rhacau 5 2 5 12 Scaand 13 15 3 11 43 Scavid 4 2 6 Scasub 41 9 10 4 64 Selopa 18 2 19 Sphste 48 312 194 421 975 Tetvir 8 6 7 7 28 Column totals 1691 1349 1563 1885 6489

177

APPENDIX B

Model sets with AICc comparisons of linear model parameters for numbers of individuals, species richness, diversity, functional diversity (FD) and ecological species traits. The Δ

AICc is the difference between the specified model and the model with the lowest AICc; model weight is a normalized measure providing the relative strength of a model relative to other candidates determined through the model likelihood divided by the total likelihood of all models; evidence ratio (calculated by dividing the weight of the given model by the weight of the top model) is a measure of the probability of a given model within the dataset. Models with a Δ of 9-11 have low empirical support and those over 20 have little or no support. Evidence ratio can be interpreted as a measure of the probability of a given model relative to other models in the set. Note: A generalized mixed linear model (GLMM) with Poisson error distribution and sampling site as a random factor was used unless otherwise stated. The notation (~1) refers to the null model (intercept only).

178

Numbers of individuals, species richness, diversity and functional diversity indices (FD)

III Table – B1. Parameter model set for numbers of individuals (N). A negative binomial

GLMM with log link function was used with a dispersion parameter of 6.35 (N = 94).

Model Residual deviance AICc Δ AICc Weight Evidence Ratio df

N ~ YEAR 900.8 913.8 0 0.951 1 6 N ~ AGE + YEAR 899.6 919.7 5.93 0.049 19.41 9 N ~ 1 946.6 952.9 39.1 0 3 N ~ AGE 945.7 958.7 44.95 0 6

III Table – B2. Parameter model set for species richness (S) (N = 94).

Model Residual deviance AICc Δ AICc Weight Evidence Ratio df

S ~ YEAR 499.5 510.2 0 0.871 1 5 S ~ AGE + YEAR 496.3 514 3.83 0.128 6.80 8 S ~ 1 522.6 526.7 16.49 0 2 S ~ AGE 519.3 530 19.83 0 5

III Table – B3. Parameter model set for functional richness (FRic) (N = 94).

Model Residual deviance AICc Δ AICc Weight Evidence Ratio df

FRic ~ YEAR 485.4 496.1 0 0.884 1 5 FRic ~ AGE + YEAR 482.5 500.2 4.08 0.115 7.69 8 FRic ~ 1 505.1 509.3 13.15 0.001 884.00 2 FRic ~ AGE 502.1 512.8 16.68 0 5

179

III Table – B4. Parameter model set for functional evenness (FEve). Since FEve values were continuous on the interval (0-1), a beta-regression model with logit link function (N

= 94).

Model logLik AICc Δ AICc Weight Evidence Ratio df FEve ~ YEAR 94.15 -177.6 0 0.958 1 5 FEve ~ AGE + YEAR 94.401 -171.1 6.51 0.037 25.89 8 FEve~1 85.483 -166.8 10.78 0.004 239.50 2 FEve~ AGE 86.005 -161.3 16.29 0 5

III Table – B5. Parameter model set for functional dispersion (FDis). GLMM with

Gaussian error structure, identity link function and study site as a random factor was used

(N = 94).

Model Residual deviance AICc Δ AICc Weight Evidence Ratio df FDis ~ 1 0.3865 -245.5 0 0.832 1 2 FDis ~ YEAR 0.3786 -240.9 4.59 0.084 9.90 5 FDis ~ AGE 0.3792 -240.8 4.74 0.078 10.67 5 FDis ~ AGE + YEAR 0.371 -235.8 9.7 0.006 138.67 8

180

Ecological species trait categories

III Table – B6. Parameter model set for small body length (N = 29).

Residual Evidence Model AICc Δ AICc Weight df deviance Ratio small ~ 1 83.66 88.1 0 0.861 1.00 2 small ~ AGE 79.87 92.5 4.36 0.097 8.88 5 small ~ YEAR 81.63 94.2 6.12 0.04 21.53 5 small ~ AGE + YEAR 77.48 100.7 12.56 0.002 430.50 8

III Table – B7. Parameter model set for medium (‘med’) body length (N = 93).

Model Residual deviance AICc Δ AICc Weight Evidence Ratio df med ~ YEAR 365.7 376.4 0 0.916 1.00 5 med ~ 1 378.3 382.4 6.07 0.044 20.82 2 med ~ AGE + YEAR 365.1 382.8 6.42 0.037 24.76 8 med ~ AGE 377.2 387.9 11.54 0.003 305.33 5

III Table – B8. Parameter model set for large body length (N = 94).

Residual Δ Evidence Model AICc Weight df deviance AICc Ratio large ~ AGE + YEAR 353.3 371 0 0.429 1.00 8 large ~ AGE 361.5 372.2 1.21 0.234 1.83 5 large ~ YEAR 361.5 372.2 1.21 0.234 1.83 5 large ~ 1 369.7 373.8 2.86 0.103 4.17 2

III Table – B9. Parameter model set for very large (‘v.large’) body length (N = 90).

Residual Δ Evidence Model AICc Weight df deviance AICc Ratio v.large ~ 1 320.3 324.5 0 0.549 1.00 2 v.large ~ YEAR 315.1 325.8 1.34 0.281 1.95 5 v.large ~ AGE 316.8 327.5 3.01 0.122 4.50 5 v.large ~ AGE + YEAR 311.6 329.4 4.89 0.048 11.44 8

181

III Table – B10. Parameter model set for macropterous (‘mac’) wing morphology (N =

89).

Model Residual deviance AICc Δ AICc Weight Evidence Ratio df mac ~ AGE 363.1 373.8 0 0.83 1.00 5 mac ~ AGE + YEAR 359.3 377.1 3.21 0.167 4.97 8 mac ~ 1 381.5 385.7 11.84 0.002 415.00 2 mac ~ YEAR 375.8 386.5 12.68 0.001 830.00 5

III Table – B11. Parameter model set of values for autumn/winter (‘AutWin’) seasonality

(main activity period) (N = 89).

Residual Δ Evidence Model AICc Weight df deviance AICc Ratio AutWin ~ YEAR 293.6 304.3 0 0.316 1.00 5 AutWin ~ AGE 293.8 304.6 0.26 0.278 1.14 5 AutWin ~ AGE + YEAR 287 304.8 0.45 0.252 1.25 8 AutWin ~ 1 301.6 305.7 1.43 0.155 2.04 2

III Table – B12. Parameter model set for carnivorous (‘carn’) diet (N = 94).

Residual Evidence Model AICc Δ AICc Weight df deviance Ratio carn ~ YEAR 408.2 418.8 0 0.751 1 5 carn ~ AGE + YEAR 403.6 421.2 2.4 0.226 3.32 8 carn ~ 1 422.4 426.5 7.65 0.016 49.94 2 carn ~ AGE 417.7 428.4 9.54 0.006 125.57 5

182

III Table – B13. Parameter model set for herbivore (‘herb’) diet (N = 28).

Model Residual deviance AICc Δ AICc Weight Evidence Ratio df herb ~ 1 80.2 84.7 0 0.881 1.00 2 herb ~ AGE 76.87 89.6 4.92 0.075 11.75 5 herb ~ YEAR 78 90.7 6.05 0.043 20.49 5 herb ~ AGE + YEAR 74.61 98.2 13.52 0.001 881.00 8

III Table – B14. Parameter model set for the omnivore (‘mix’) diet (N = 79).

Model Residual deviance AICc Δ AICc Weight Evidence Ratio df mix ~ AGE 236.2 247.1 0 0.951 1 5 mix ~ AGE + YEAR 235.2 253.2 6.19 0.043 22.12 8 mix ~ 1 253.3 257.5 10.42 0.005 2 mix ~ YEAR 250.9 261.8 14.71 0.001 5

III Table – B15. Parameter model set for specialist (‘spec’) diet (N = 91).

Residual Evidence Model AICc Δ AICc Weight df deviance Ratio spec ~ YEAR 319.9 330.6 0 0.307 1 5 spec ~ AGE + YEAR 313.1 330.8 0.26 0.27 1.14 8 spec ~ AGE 320.2 330.9 0.36 0.256 1.20 5 spec ~ 1 327.7 331.8 1.23 0.166 1.85 2

III Table – B16. Parameter model set for non-metallic (‘nonmet’) body color (N = 94).

Residual Δ Evidence Model AICc Weight df deviance AICc Ratio nonmet ~ YEAR 461.4 472.1 0 0.798 1.00 5 nonmet ~ AGE+YEAR 457.1 474.8 2.75 0.202 3.95 8 nonmet ~ 1 491.5 495.6 23.59 0 2 nonmet ~ AGE 487.3 497.9 25.89 0 5

183

III Table – B17. Parameter model set for the metallic (‘met’) body color (N = 92).

Residual Evidence Model AICc Δ AICc Weight df deviance Ratio met ~ 1 326.8 330.9 0 0.539 1.00 2 met ~ AGE 321 331.7 0.74 0.373 1.45 5 met ~ YEAR 324.3 335 4.1 0.069 5.41 5 met ~ AGE + YEAR 319.8 337.6 6.65 0.019 3.63 8

III Table – B18. Parameter model set for dark leg color (N = 94).

Residual Evidence Model AICc Δ AICc Weight df deviance Ratio dark ~ YEAR 440.6 451.3 0 0.929 1.00 5 dark ~ AGE + YEAR 439 456.7 5.37 0.063 14.75 8 dark ~ 1 456.8 461 9.66 0.007 9.00 2 dark ~ AGE 454.9 465.6 14.32 0.001 7.00 5

III Table – B19. Parameter model set for pale leg color (N = 88).

Residual Evidence Model AICc Δ AICc Weight df deviance Ratio pale ~ AGE 340.8 351.5 0 0.62 1 5 pale ~ 1 349.3 353.5 1.95 0.233 2.66 2 pale ~ YEAR 337.9 355.7 4.22 0.075 3.11 8 pale ~ AGE + YEAR 345.1 355.8 4.32 0.071 1.06 5

III Table – B20. Parameter model set for pubescence (‘pub’) body (N = 73).

Model Residual Evidence AICc Δ AICc Weight df deviance Ratio pub ~ 1 221.5 225.7 0 0.835 1.00 2 pub ~ AGE 218.6 229.5 3.83 0.123 6.79 5 pub ~ YEAR 221 231.9 6.22 0.037 3.32 5 pub ~ AGE + YEAR 217.9 236.2 10.44 0.005 7.40 8

184

III Table – B21. Parameter model set for glabrous (‘glab’) body (N = 94).

Evidence Model Residual deviance AICc Δ AICc Weight df Ratio glab ~ AGE + YEAR 466.3 484 0 0.617 1 8 glab ~ YEAR 474.3 485 0.96 0.383 1.61 5 glab ~ AGE 489.5 500.2 16.15 0 5 glab ~ 1 497.7 501.8 17.76 0 2

185

APPENDIX C

III Table – C1. Analysis of deviance results (frequentist approach) for each trait category for two fixed effects: forest age and year (df = 3). Significance was determined through analysis of deviance using Wald chi-square test statistics (χ2). If models failed to meet assumptions (e.g. homogeneity), non-parametric Kruskal-Wallis rank sum test was used

(denoted with H as the critical statistic).

Fixed effect Trait category n Age Year LEN small 29 3.8, p = 0.281 2.4, p = 0.522 medium 93 0.6, p = 0.819 11.7, p = 0.008 large 94 9.2, p = 0.026* 8.1, p = 0.046* very large 90 3.3, p = 0.345 5.0, p = 0.172 WIN macropterous 89 21.1, p < 0.001*** 3.9, p = 0.276 brachypterous 94 H = 8.4, p = 0.039* H = 19.4, p < 0.001*** SEA autumn/winter 89 6.6, p = 0.084 6.5, p = 0.088 spring/summer 94 H = 4.0, p = 0.266 H = 16.5, p = 0.001*** DIET specialist 91 6.8 , p = 0.079 7.2, p = 0.067 carnivore 94 4.6 , p = 0.204 13.4, p = 0.004** herbivore 28 3.0, p = 0.396 2.3, p = 0.519 mix 79 19.9, p < 0.001*** 1.0, p = 0.790 BODY non-metallic 94 4.76, p = 0.191 28. 3, p < 0.001*** metallic 92 4.9, p = 0.178 1.2, p = 0.764 LEG dark 94 1.7, p = 0.636 15.4, p = 0.001*** pale 88 8.5, p = 0.036* 2.8, p = 0.412 PUB pubescent 73 3.0, p = 0.387 0.8, p = 0.860 glabrous 94 9.6, p = 0.023* 22.1, p < 0.001***

186

APPENDIX D

III Table – D1. Means (± s.e.) for numbers of individuals (N), species richness (S), inverse Simpson’s concentration index (1/D) and functional diversity indices for forest age classes and sampling years. Means for sampling years one, two and three are based on the original 29 sampling sites (see Figure 1) and means for the extra year are based on the additional year of sampling for zero year and ten year age classes.

Measure Year Zero Ten Fifty Ninety

One 64.2 ± 8.4 95.5 ± 11.8 85.4 ± 25.7 72.2 ± 8.5 Two 58.6 ± 9.6 68.0 ± 9.2 84.3 ± 15.5 77.4 ± 9.1 N Three 47.0 ± 11.6 26.0 ± 3.7 53.7 ± 13.9 38.7 ± 6.8 Extra 168.0 ± 55.2 70.63 ± 29.7 ------One 14.3 ± 1.0 13.8 ± 1.6 11.4 ± 1.2 11.6 ± 0.7 Two 13.7 ± 1.7 10.7 ± 1.1 15.6 ± 1.7 12.5 ± 0.6 S Three 11.8 ± 2.6 5.8 ± 0.7 10.3 ± 1.1 9.0 ± 0.9 Extra 12.5 ± 3.5 13.7 ± 2.4 ------

One 5.73 ± 1.18 4.33 ± 0.39 4.38 ± 0.63 4.53 ± 0.35 Two 5.12 ± 1.02 4.02 ± 0.50 6.21 ± 0.37 4.91 ± 0.51 1/D Three 5.35 ± 2.03 3.57 ± 0.28 5.16 ± 0.40 4.58 ± 0.49 Extra 5.21 ± 2.05 5.26 ± 0.44 ------

One 13.83 ± 0.87 13.17 ± 1.58 11.00 ± 1.16 10.44 ± 0.74 Two 12.33 ± 1.36 10.00 ± 1.10 14.14 ± 1.45 11.30 ± 0.62 FRic Three 11.00 ± 2.22 5.67 ± 0.80 9.29 ± 1.06 8.60 ± 0.91 Extra 11.50 ± 2.87 12.33 ± 2.60 ------

One 0.64 ± 0.03 0.57 ± 0.04 0.60 ± 0.03 0.60 ± 0.02 Two 0.62 ± 0.04 0.56 ± 0.04 0.63 ± 0.02 0.57 ± 0.03 FEve Three 0.64 ± 0.02 0.69 ± 0.05 0.66 ± 0.04 0.69 ± 0.03 Extra 0.49 ± 0.07 0.63 ± 0.01 ------

One 0.34 ± 0.03 0.30 ± 0.02 0.26 ± 0.03 0.32 ± 0.01 Two 0.30 ± 0.03 0.25 ± 0.02 0.30 ± 0.01 0.28 ± 0.03 FDis Three 0.28 ± 0.03 0.28 ± 0.03 0.31 ± 0.02 0.32 ± 0.02 Extra 0.28 ± 0.06 0.32 ± 0.03 ------

187

CHAPTER IV

CARABID BEETLE (COLEOPTERA: CARABIDAE) RICHNESS, DIVERSITY AND

COMMUNITY STRUCTURE IN THE UNDERSTORY OF TEMPORARILY

FLOODED AND NON-FLOODED AMAZONIAN FORESTS

188

ABSTRACT

Although tropical regions harbor the highest level of arthropod diversity on Earth, the majority of these taxa are taxonomically and scientifically unknown. Few studies have compared understory arthropod diversity and species assemblages for different forest types within Amazonia. This study examines the differences in species richness, diversity and community composition of carabid beetles (Coleoptera: Carabidae) in temporarily flooded (FP) forest and non-flooded terra firme (TF) forest in the Yasuní area of Ecuador.

The forest understory was sampled by flight intercept traps (FITs) and systematic nightly hand collections in June and July 2011 and 2012, and in October and November 2011. A total of 1,255 Carabidae individuals representing 20 tribes, 54 genera and 143 morphospecies were collected. Mean number of individuals and mean species richness did not differ between FP and TF; however, numbers of Cicindelini (tiger beetles) and

Pentagonicini were higher in TF forest while numbers of Lachnophorini and Scaritini were higher in FP forest. Interpolated rarefaction curves show FP had significantly higher rarefied richness but extrapolation curves and nonparametric diversity estimators suggest these differences may decrease with additional sampling. Means for the inverse

Simpson’s concentration index and a nonparametric diversity estimator, ACE, were significantly higher for FP than TF forest. Nonmetric multidimensional scaling ordination and dissimilarity coefficient values suggest that FP and TF forests maintain unique assemblages with minimal overlap in community composition. Greater understanding of the richness, diversity and community assemblages of Yasuní rainforests are needed as anthropogenic pressures, particularly petroleum extraction, continue to threaten this megadiverse area.

189

INTRODUCTION

Insects and their relatives dominate global biodiversity (Erwin 1982; Hammond

1992; Samways 1993; Basset et al. 2012). Although approximately 1.24 million arthropod species have been described, over 80% which are insects (Zhang 2011), a recent estimate by Hamilton et al. (2013) suggests that 60 – 80% of 6.1 million arthropod species await description (Erwin & Johnson 2000; Dunn 2005; Hamilton et al. 2010,

2013; May 2010). As major components of highly diverse ecosystems, arthropods play vital roles in ecological processes (Samways 1993; Kim 1993). With current rates of deforestation, degradation, many tropical forest species, including arthropods will remain unknown indefinitely (Kellert 1993; Samways 1993; Kim 1993; Godfray et al. 1999;

Dunn 2005; Finer et al. 2008).

A greater comprehension of tropical arthropod diversity and communities is needed for realistic global diversity estimates, to discern ecological patterns including species distributions and to detect and mitigate disturbance and conservation strategies

(Erwin 1991a; Oliver & Beattie 1996; Carlton et al. 2004; Rohr et al. 2007; Erwin &

Geraci 2009; May 2010; Mora et al. 2011). Although species turnover across habitats is an important aspect of Amazonian diversity (Pitman et al. 2001; Condit et al. 2002), we have only a limited understanding of the structure, general patterns and assembly rules for arthropods communities in tropical forests, even across broad scale environmental gradients. Arthropods are expected to have high spatial turnover across lowland tropical forests owing, at least in part, to high ecological specialization compared to temperate forests (Erwin 1983; Lucky et al. 2002; Dyer et al. 2007; Fine et al. 2013). Fine et al.

190

(2013) found high turnover of herbivorous insects for a single tree species across two

Amazonian forest types, as well as high turnover within the same habitat.

More research on species turnover is needed, particularly for non-herbivore arthropods, as the majority of current estimates and knowledge are based on herbivorous arthropods (Kitching 2006; Fine et al. 2013). Amazonian forests have been classified into broad types based on forest structure, drainage, floristic composition, geology and soils

(Prance 1979; Duivenvoorden & Lips 1995; Herrera-MacBryde & Neill 1997; Baraloto et al. 2011). Two major forest types in Amazonia, including the Yasuní area of Ecuador, are floodplain forests and upland non-flooded terra firme forests (Pitman 2000; Myster 2014).

Terra firme forests are the most widespread forest type in the Amazon Basin while floodplain forests are riparian forests subject to episodic inundations by adjacent rivers during periods of high precipitation or upland Andean precipitation and/or snowmelt

(Prance 1979; Junk et al. 1989; Worbes 1997; ter Steege et al. 2000; Parolin et al. 2004;

Junk & Piedade 2010; Myster 2014).

Inventories of tree communities across Amazonia have shown that although both forest types are rich in species, terra firme forests are generally more species rich than floodplain forests (Campbell et al. 1986; Balslev et al. 1987; Gentry 1988, 1992;

Duivenvoorden & Lips 1995; ter Steege et al. 2000; Parolin et al. 2004; Pitman et al.

2014). Arthropod investigations have also shown similar increases in diversity for terra firme forests relative to flooded forests for spiders (Höffer 1997), ants (Wilson 1987;

Majer & Delabie 1994; Mertl et al. 2009), canopy arthropods (Erwin 1983; Adis et al.

1984; Adis et al. 2010) and terrestrial arthropods in general (Adis 1997; Adis & Junk

2002; Lamarre et al. 2016). However only a relatively small amount of comparative data

191

are currently available, particularly for understory Coleoptera (Adis & Junk 2002). The reasons for the reported richness patterns are typically attributed to stress due to flooding events (Junk et al. 1989; Duivenvoorden 1995; ter Steege et al. 2000; Wittmann et al.

2004; Adis et al. 2010) and differences in overall area coverage (Rosenzweig 1995; ter

Steege et al. 2000; Tuomisto et al. 2003; Myster 2014). In addition to species number, floodplain and terra firme forests can have prominent differences in vegetation structure, dynamics, stem density, soil moisture and soil chemical composition (Campbell et al.

1986; Balslev et al. 1987; Duivenvoorden & Lips 1995; Worbes 1997; Pitman 2000;

Baraloto et al. 2011). Differences in environmental conditions, resource availability and disturbance regimes between the two forest types may select for particular ecological strategies leading to differences in community structure (Kubitzki 1989; Duivenvoorden

& Lips 1995; Parolin et al. 2004; Wittmann et al. 2010; Fine et al. 2013), including arthropod community compositions (Erwin 1983; Adis 1997; Ellis et al. 2001; Mertl et al.

2009; Baccaro et al. 2013; Lamarre et al. 2016), particularily for Coleoptera (Adis et al.

2010).

Within Insecta, the beetles (Coleoptera) are functionally, ecologically and morphologically diverse and extremely species rich, with beetles comprising up to a quarter of the planet’s species (Erwin 1996; Ødegaard 2000; Hunt et al. 2007). The beetle family Carabidae has a global distribution, with many predaceous species while other taxa are predominately phytophagous, omnivorous or have specialized modes of nutrition

(Thiele 1977; Lövei & Sunderland 1996). With approximately 40,000 described species,

Carabidae are one of the most diverse families of beetles, with the highest diversity in

192

tropical regions (Erwin 1985; Paarmann et al. 2002), including the Ecuadorian Amazon

(Lucky et al. 2002; Erwin et al. 2005).

Only a few ecological studies have compared understory arthropod diversity and species assemblages from different forest types within Amazonia (Adis 1981; Pearson &

Derr 1986; Majer & Delabie 1994; Zerm et al. 2001; Mertl et al. 2009; Lamarre et al.

2016). Much of the current knowledge is based on research within Central Amazonia where floodplain forests are annually inundated by several meters of water for several months a year (i.e. Irmler 1979a, 1979b; Adis 1981; Paarmann et al. 1982; Zerm et al.

2001). This study focuses on forests within the highly diverse western Amazon which experiences less severe inundation events compared to Central Amazonia, both in duration and frequency (Burnham et al. 2001; Hoorn et al. 2010). The goals of this study are to (1) investigate differences in species abundance, richness and diversity between flooded and non-flooded terra firme forests; (2) identify patterns in the abundance and species distributions of common and rare species; and (3) examine carabid species assemblages for each forest type, including identification of taxa characteristic of floodplain and terra firme forests. In contrast to the majority of results reported for other taxa, I predict that temporarily flooded (FP) forests of lowland Ecuador will have greater species richness and diversity for Carabidae compared to adjacent terra firme (TF) forests.

Since the taxon pulse concept (Erwin 1979, 1985) suggests that early carabid lineages were associated with waterside habitats, higher diversity is predicted for FP forest which experience temporary flooding events. Higher Carabidae richness has been previously reported for flooded forests relative to terra firme forests in some studies (Erwin & Adis

1982; Pearson 1984; Pearson & Derr 1986). Additionally, since higher species richness

193

has been reported for carabid beetles response to forest disturbances (e.g. clear-cutting)

(Erwin 1979; Riley & Browne 2011), disturbance associated with flooding events may also increase carabid species richness through reduction in competitve mechanisms

(Connell 1978; Erwin & Adis 1982; Zerm & Adis 2001a, 2001b). Disturbance through flooding has deterministic properties as it has been occurring for millions of years (Junk

1997), but also stochastic components resulting in unpredictable effects (Junk 2000;

Zerm and Adis 2001a). Floodplain communities consist of not only of residents, completing their full life cycle in floodedplain habitats, but also migrants completing part of their life cycle in flooded forests. Also, there are species, either regular or occasional, which visit flooded forests from other habitats (Adis 1997; Junk 1997; Adis & Junk

2002). I hypothesize that the majority of species will be rare, in terms of the overall carabid beetle catch, and FP forest will have a higher proportion of rare species than TF forest, owing to the higher frequency of migrant and visiting species. Since common species will likely occur in both forest types (Hanski 1982; Pitman et al. 2001), I do not expect a significant difference in numbers of common species or individuals between FP and TF forests. However, I expect carabid species composition to differ between the forest types, because of the aforementioned differences between FP and TF forests partiularily plant composition, forest structure and resource availability (Campbell et al.

1986; Balslev et al. 1987; Worbes 1997). Species specific adaptions and survival strategies for flooding regimes have been reported for tropical carabid beetles (Adis 1982;

Adis 1997; Anorim et al. 1997; Adis et al. 1998; Paarmann et al. 1982), suggesting that distribution of certain species may be specific to flooded areas. As a result of flooding

194

events, I expect higher variability within species assemblages for FP forests compared to

TF forest.

METHODS

Study site

This research was conducted at Tiputini Biodiversity Station (TBS) (0°37’55” S,

76°08’39” W; 190 – 270 m elevation) (Figures 1A & 1B) located in the Ecuadorian

Amazon, within the province of Orellana. TBS consists of 650 ha of moist tropical rainforest and borders the megadiverse Yasuní National Park (Bass et al. 2010). The forest surrounding TBS has experienced relatively minimal anthropogenic disturbance.

The area has average annual temperatures of 24 – 27°C and high yearly rainfall

(approximately 3200 mm). The climate is considered aseasonal compared to other areas in Amazonia, as no month receives < 100 mm of rain. However there are seasonal patterns in precipitation with rainier months (‘wetter season’) from April through June and September – October and drier periods (‘drier season’) in August and November –

January (Pitman 2000; Blake et al. 2011).

The area is relatively flat but several ridges, approximately 25 – 50 m higher than rivers and streams, add slight topographic variation. Approximately 90% of the landscape surrounding TBS is non-flooded terra firme forest (TF) which occurs along the slopes and tops of ridges. Rivers dissect these large blocks of terra firme forest creating narrow bands of floodplain (FP) forest which comprise around 2% of the study area (Pitman

2000; Pitman et al. 2001, 2014; Mapa de Vegetación del Ecuador 2012). The palm,

Iriartea deltoidea is a dominant tree species in both FP and TF forests near TBS (Pitman

195

et al. 2001; Myster 2014). (For more information on forest structure and tree assemblages in the Yasuní area see Burnham et al. 2001, Pitman et al. 2001 and Myster 2014). FP forest occurs within a relatively narrow area along the Tiputini River, an intermediately sized, mixed-water river (contributions of white and black water) with a relatively high sediment load (Pitman 2000). Tiputini River water height mainly depends on rainfall and snowmelt in the Andes Mountains (Prance 1979; Pitman 2000; Burnham 2002). See

Appendix A for more information on mean annual river heights. Forests are inundated by the river for various lengths of time depending on elevation and distance from the river

(Burnham et al. 2001; Pitman et al. 2001; Mertl et al. 2009). Previous studies have estimated that the floodplains around TBS are inundated (up to two meters) one to three times a year with inundations lasting less than a week (Pitman 2000; Burnham et al. 2001;

Leimbeck & Balslev 2001; Burnham 2002; Mertl et al. 2009).

TBS is located on a geologically young landform consisting of fluvial deposits

(red clays and alluvium) originating from the Andes Mountains and are rich in exchangeable bases (Hoorn et al. 2010). TF soils have been classified as Ultisols which are clayey, acidic and high in iron and aluminum (Pitman 2000; Sombroek 2000;

Tuomisto et al. 2003). FP soils are less acidic, lower in iron and aluminum but higher in other nutrients (Ca, Mg, K, Na, P) and organic carbon compared to TF soils (Kapos et al.

1990; Duivenvoorden & Lips 1995; Pitman 2000)

Collection of Carabidae

Carabid beetles were sampled during the rainier months of June and July in 2011 and 2012 and during the transition to the drier season in October and November of 2011.

Two sampling methods, flight intercept traps (FITs) and hand sampling, were utilized at

196

24 sites in the forest understory with half of the sites located in FP forest and half in TF forest (Figure 1C). Both methods were employed in the rainier seasons but for logistical reasons only FITs were employed in the transitional season in 2011. The number of FIT trapping days varied slightly among forest types due to flooding events in the 2011 wet season, ranging from 33 – 49 days in the 2011 wetter season, 30 – 31 days in the 2012 wetter season and 29 days in the 2011 transitional season.

Each FIT consisted of PVC tubes, screen netting, a plastic awning (to prevent rain and debris from entering the trays) and plastic collection trays. Screening was stretched between two 1 m PVC tube, with a 1 m support PVC tube across the top of the trap to create an approximately 1 m2 flight barrier (See Riley et al. 2016 for additional details on trapping techniques). Plastic trays were placed underneath the flight barrier to collect specimens and filled with water and biodegradable soap. Sampling trays were emptied every other day by filtering the contents through a metal strainer with mesh openings < 1 mm.

Since the majority of Carabidae are active at night (Lövei & Sunderland 1996), hand sampling occurred between 8:30 pm and 10:30 pm. This included active searching of the ground (leaf litter, tree trunks, etc.), tree trunks and vegetation (up to eye level) and decaying logs. Seven hand sampling events were performed at each of the 24 sampling sites, four in 2011 and three in 2012. Each sampling event consisted of 0.5 hour of search effort within 100 meters of the FIT for each site. I participated in every collection event with the assistance of 1 – 2 amateur collectors. Immediately after capture, carabids were preserved in 95% ethanol.

197

Carabid beetles were sorted using a dissecting microscope and identified to genera and morphospecies. Since a species-level taxonomic key is not available for the study region, a genera level key for French Guyana Carabidae (Erwin et al. 2012) was used, with subsequent identification to morphospecies and taxonomic confirmation by T.

Erwin. The use of morphospecies for Neotropical arthropods has been successful in ecological studies and is the most feasible option for an extraordinarily speciose region, with the majority of species not described (Oliver & Beattie 1996; Stork et al. 2008;

Grimbacher & Stork 2009).

Statistical analysis

Numbers of individuals, richness and diversity

For each sampling site, the number of individuals (N), raw morphospecies richness (S) (hereafter referred to as species richness) and the inverse Simpson’s concentration index (1/D), a Hill diversity number (Simpson 1949; Jost 2006; Chao et al.

2014) were determined. To detect changes at higher taxonomic levels, the numbers of individuals and species for each Carabidae tribe found in FP and TF forests were also analyzed. To compare mean Carabidae body length between FP and TF forests, apparent body length (ABL) was determined by measuring at least 80% of all specimens with the exception of the tiger beetles (Cicinidelini) in which 23% were measured. ABL is the length from the extreme point of the mandible to apex of elytra and is considered a standard carabid beetle measure for ecological studies which provides a reliable estimate of overall size for Carabidae (Erwin & Kavanaugh 1981). To compare FP and TF forests, a two-sample t-test with equal variance was used. If the assumption of homogeneity was violated then a two-sample t-test with a Welch correction was used. Data failing to meet

198

the assumption of normality and nonparametric estimators were analyzed with a

Wilcoxon rank sum test.

To account for differences in numbers of individuals collected between forest types and under-sampling bias, sample size was standardized by rarefaction which determines the expected number of species from a random subsample of individuals from the overall data set (Gotelli & Colwell 2001, 2011). Rarefaction curves for each forest type were constructed in R package, ‘iNEXT’ (Hsieh et al. 2015), using 500 randomizations to determine 95% unconditional confidence intervals (Chao et al. 2014).

The x-axis was rescaled to individuals to correct for differences in sample density among sampling sites within each forest type (Gotelli & Colwell 2001, 2011, Colwell et al. 2004,

2012). To estimate the number of undetected species rarefaction curves were extrapolated by a factor of two from the smallest sample size between forest types (Colwell et al.

2012). The abundance-based non-parametric species richness estimator, Chao1 was calculated to estimate potential true richness and to extrapolate the rarefaction curves.

Chao1 has been recommended as the lower bound of asymptotic richness (Chao 1984;

Hortal et al. 2006; Colwell et al. 2012). To account for estimator bias, two additional non-parametric diversity estimators and standard errors (randomizations = 1000) were calculated through R package ‘vegan’ (Oksanen et al. 2016): the first order jackknife,

Jack1 (Burnham & Overton 1979; Heltshe & Forrester 1983) and the abundance-based coverage estimator, ACE (Chao & Lee 1992; Chao & Yang 1993; Lee & Chao 1994).

Estimators were selected based on the species abundance distribution and evenness of the data, all of which affect estimator biases, in addition to the high performance reported in previous studies (Chazdon et al. 1998; Brose & Martinez 2004; Walther & Moore 2005;

199

Hortal et al. 2006; Gotelli & Colwell 2011). Since ACE accounts for the frequency of rare species, it is more appropriate for datasets with a high proportion of rare species as is common for studies examining Carabidae assemblages (Chao and Lee 1992; Chazdon et al. 1998; Belaoussoff et al. 2003). Jack1 has performed well in a variety of studies

(Colwell & Coddington 1994; Walther & Moore 2005; Hortal et al. 2006; Basualdo 2011) with low bias for data with higher evenness and more mobile organisms (Palmer 1990;

Brose & Martinez 2004). Results from the three estimators were averaged to provide values relating to sample completeness. Estimators were selected based on the aforementioned data attributes and high performance in previous studies (Colwell &

Coddington 1994; Brose & Martinez 2004; Walther & Moore 2005; Hortal et al. 2006;

Gotelli & Colwell 2011).

Species assemblages

Simpson’s evenness (E1/D) values were calculated for each sampling site and significance determined by a Wilcoxon rank sum test. This evenness index has high precision and low bias (Smith & Wilson 1996; Payne et al. 2005). Rank abundance curves were used to examine species abundance distributions for FP and TF forests using logged relative abundance versus species rank. To determine if there were differences between more commonly collected species and less frequently collected species, species were assigned to rarity categories based on the proportion of their abundance to the overall carabid beetle catch (relative abundance) over the entire sampling period with both forest types combined. Species were classified as ‘dominant’ if their relative abundance contributed more than 10.0% of the total number of individuals collected.

Species collected in higher numbers contributing 1.0 – 9.99% were classified as

200

‘common’ while less frequently sampled species were classified as ‘rare’ if they contributed less than 1.0% to the total number of individuals collected. Changes in the number of individuals and the number of species between FP and TF forests for these three rarity categories were analyzed through two-sample t-tests or, when t-test assumptions were not met, nonparametric Wilcoxon tests. To determine if rarity category was related to the number of sampling sites at which a species was present, a generalized linear model (GLM) was used with Poisson error distributions and a log link, with significance determined by analysis of deviance using the Wald chi-square test statistic.

To determine if the distributions of common and rare species were random among

Carabidae tribes, expected and observed values were determined and compared using a chi-square test. Dominant and common species were combined for this analysis and referred to as ‘abundant’. The expected numbers of abundant and rare species were determined by multiplying the total number of species collected in the tribe by the proportion of species in each rarity category (e.g. 125 of 143 species were rare, 87.4%).

Species assemblages for FP and TF were compared using Bray-Curtis and Jaccard dissimilarity indices and ordination analysis with non-metric multidimensional scaling

(NMDS) methods (Minchin 1987; Oksanen 2015). Bray-Curtis (abundance-based) and

Jaccard (presence/absence) dissimilarity values range between 0 (identical species assemblages) and 1 (completely different). NMDS is a robust unconstrained ordination method which does not make assumptions about the underlying nature of species distributions. Since the mean-variance relationships were similar between FP and TF forests, raw abundance values of all species were used as the input for community-level analyses (Minchin 1987; Clarke 1993; McCune & Grace 2002; Nekola et al. 2008). To

201

ensure solution stability, all NMDS analyses were performed using a dissimilarity matrix

(Bray-Curtis or Jaccard), using 100 runs with random start points (Minchin 1987).

Statistical significance was determined with adonis, a robust permutational multivariate analysis of variance (permutations = 999) (Legendre & Anderson 1999; Anderson 2001;

McArdle & Anderson 2001; Oksanen 2015). To test for multivariate homogeneity of dispersions within forest types I used function betadisper and TukeyHSD for the post-hoc test (Anderson 2006; Anderson et al. 2006; Oksanen 2015). All community-based analyses completed using R package ‘vegan’ and ‘MASS’ (Venables & Ripley 2002; R

Core Team 2016; Oksanen et al. 2016). Linear modeling was completed using R packages ‘lme4’, ‘car’, and ‘multcomp’ (Hothorn et al. 2008; Fox & Weisberg 2010;

Bates et al. 2015; R Core Team 2016).

Characteristic species

To identify species driven differences in community level analyses, the species indicator statistic (IndVal) was calculated. This statistic, developed by Dufrêne and

Legendre (1997), is insensitive to community-level beta diversity and does not consider absences as negative preferences (Dufrene & Legendre 1997; De Cáceres & Legendre

2009). IndVal is calculated independently for each species and is the product of two components: specificity (A: uniqueness to a group using relative abundance) and fidelity

(B: relative frequency of occurrence within a group). The highest value is 1, which indicates a species was present only in the given habitat type (specificity, i.e. only in TF forests) and present in all samples (fidelity, i.e. at all TF sampling sites) (Dufrene &

Legendre 1997; De Cáceres & Legendre 2009). Only species which occurred at three sites (i.e. 25% of the number of sites in a forest type) over the entire sampling period

202

were included in the analysis. Indicator value and statistical significance were determined using function multipatt in R package ‘indicspecies’ (permutations = 999). Holm correction for multiple testing used in experiment wide conclusions (Holm 1979; De

Cáceres & Legendre 2009).

Spatial autocorrelation

In order to examine spatial autocorrelation, a multivariate Mantel correlogram was constructed (Mantel 1967; Oden & Sokal 1986; Sokal 1986; Borcard & Legendre

2012; Legendre & Legendre 2012). The analysis inputs were the Bray-Curtis dissimilarity for species assemblages and the Cartesian coordinates of the sampling sites determined using R package ‘SoDA’ (Chambers 2013). If a linear trend was detected, species data was Hellinger-transformed and detrended prior analysis to satisfy normality and the second-order stationarity condition (Borcard et al. 2011). Sturge’s rule was used to determine the number of distance classes (Borcard et al. 2011). Following 5000 permutations the significance of the normalized Mantel statistic for the distance classes were evaluated using Holm correction for multiple testing (Holm 1979; Borcard et al.

2011; Goslee & Urban 2013). Autocorrelation analyses were completed using R package

‘vegan’ (Oksanen et al. 2016; R Core Team 2016).

RESULTS

Numbers of individuals, richness and diversity

A total of 1,255 Carabidae individuals were collected, representing 20 tribes, 54 genera and 143 species (Table 1). There were no significant differences between forest types for mean number of individuals or mean species number, although raw richness

203

was close to significant (p = 0.060; Table 1). The rarefaction curve for the overall dataset is not near an asymptote and the extrapolation curve estimates a total species number of

189.0 ± 22.3 if sample effort was increased by a factor of two (Figure 2A). The rarefied richness of FP was significantly higher than TF, which is supported by the lack of overlap between the 95% unconditional confidence intervals for the rarefaction curves at the rarefied sample size (n = 564; Figure 2B). The rarefied numbers of species (± 95% CI) for FP and TF were 96.0 ± 8.0 and 71.7 ± 8.1, respectively. The extrapolated rarefaction curves suggest that as sample size increases the difference in cumulative richness between FP and TF will decrease and the differences may not be significant once sample size is increased by a factor of two. For the nonparametric estimators, two of the three predict TF will have a higher true richness than FP (Table 1). Averaging the three estimators together yielded the following approximations of sample completeness among forest types and for the overall dataset completeness: FP (74.2%), TF (55.4%) and overall

(62.0%). None of the estimators have stabilized and most have relatively large standard errors. The inverse Simpson’s concentration diversity index was significantly higher for

FP than TF (t = 3.3, df = 22, p = 0.003; Figure 3).

Carabidae body length (ABL) ranged from 1.3 to 24.7 mm, with an overall mean

(± s.e.) of 6.68 ± 0.19 mm for both forest types combined. There were no significant differences in mean body size between FP (6.78 ± 0.19) and TF individuals (6.59 ± 0.34).

Since tiger beetle abundance differed between forest types and they have a relatively large body size (mean range: 9.9 – 16.1 mm), body length was also analyzed without

Cincidelini. Excluding tiger beetles, mean body length of individuals collected TF (5.47

204

± 0.15 mm) was significantly smaller than FP individuals (6.45 ± 0.16 mm; t = 4.4, df =

22, p < 0.001; Figure 4).

Species assemblages

There were tribal-level differences between the two forest types with six tribes

(Callistini, Collyridini, Galertini, Hiletini, Oodini, Perigonini) collected from FP forest only and no tribes were unique to TF forest (Appendix B, Table B1). Two of the six tribes unique to FP forest were represented by a single individual. FP and TF forests shared 14 tribes, with 20 tribes represented in FP forest and 14 in TF. The most species rich tribes for FP forest were Lebiini (14.6% of total S), Scaritini (14.6% of total S) and

Lachnophorini (12.5% of total S), followed by Harpalini and Bembidiini (both with 11.5% of total S). The most species rich tribe for TF forest by a large margin was Lebiini (35.4% of total S), followed by Harpalini (10.1% of total S), Bembidiini (10.1% of total S) and

Pterostichini (8.9% of total S). When both forest types are combined, four species of

Cicindelini accounted for 35% (n = 438) of the total number of individuals sampled. The mean number of Cicindelini individuals was significantly greater for TF forest than FP forest (W = 24, p = 0.011; Figure 5). For carabid tribes occurring in both forest types, the mean number of individuals for Scaritini was significantly higher in FP forest (t = 6.5, df

= 11, p < 0.001) but significantly lower for Pentagonicini (W = 14, p = 0.040; Appendix

D, Figure D1). The mean species number was significantly higher in FP forest for

Lachnophorini (W = 64, p = 0.008) and Scaritini (t = 7.3, df = 11, p < 0.001), but significantly lower for Pentagonicini in FP forest (W = 15, p = 0.036; Appendix D,

Figure D2).

205

Pentacomia (Pentb) within the Cicindelini was the only species to be classified as dominant, accounting for 23.3% of the overall number of individuals sampled (n = 292).

There were 18 species classified as common (12.6% of the total S) and 125 species as rare (87.4% of total S). In relation to the total numbers of individuals collected, individuals classified as common made up 48.9% (n = 614) while rare individuals contributed 27.8% (n = 349). Within the rare species, there were 63 represented by a single individual (singletons) and 22 species by two individuals (doubletons). Of the total number of species collected, 43.8% were singletons and 15.3% doubletons. Pentb made up 11.7% of the total individuals in FP forest and 32.7% of the individuals in TF forest, with the number of Pentb individuals in TF significantly higher than FP forest (W = 20.5,

N = 21, p = 0.02; Appendix C, Figure C1). There were no differences in numbers of individuals for common or rare species between the forest types.

Rare species contributed a large proportion of the total species richness for both forest types (80.2% for FP forest and 78.5% for TF forest). The number of species classified as rare was significantly higher for FP forest than TF forests (t = 2.2, df = 22, p

= 0.04; Appendix C, Figure C2) but there were no differences in the number of species classified as common. All but two common species (90% of total common S) occurred in both forest types while only 15 rare species (12% of total rare S) occurred in both forest types. Dominant species occurred at a higher number of sampling sites than common species and common species occurred at a higher number of sampling sites than rare species (χ2 = 291.2, df = 2, p < 0.001; Appendix C, Figure C3).

There were 19 abundant species (abundant = dominant + common), contributing

72.2% of the total number of individuals collected and representing seven Carabidae

206

tribes. The distribution of these commonly collected species within the seven tribes was significantly different than expected species number calculated based on the total number of species in each tribe (chi-square χ2 = 33.6, df = 6, p < 0.001; Appendix C, Table C1).

The distribution of 124 rare species within 19 tribes was not significantly different than expected.

Simpson’s evenness index (E1/D) indicated that species assemblages had significantly higher evenness for FP than TF forests (t = 2.9, df = 22, p = 0.008; Figure 6).

The rank abundance distribution curves (Figure 7) show that both forest communities have long ‘tails’ of rare species, the FP forest curve is less steep than the TF forest curve, further supporting the increased evenness of species assemblages for FP forests. The high dominance of the first ranked species (Pentb) for TF forest is evident in the TF abundance curve. The abundance distribution curve for FP forest is longer, indicating the increased species richness of this forest, and the lower slope of the curve for ranks 25 –

62. The two abundance curves diverge after the threshold between common and rare species, approximately at species rank 25. This divergence is due an increased number of species represented by 3 – 4 individuals in FP forests (n = 18) compared to TF (n = 6), and due to a higher number of doubleton species in FP (n = 19) compared to TF forests (n

= 9). There were also a higher number of singleton species collected for TF forest (n = 33) compared to FP forest (n = 30). To compare a tropical and temperate forest assemblage, rank abundance distribution curves for the TF forest at TBS and an approximately 90 year old forest from the Piedmont of North Carolina are shown in Figure C4 (Appendix

C). The greater contribution of rare species in the tropical Carabidae assemblage compared to the temperate curve is clear from the ‘tails’ for the curves. The temperate

207

forest curve was created using data from Chapter II of this dissertation and it should be noted that the temperate forest curve was sampled for a longer period of time using different sampling methods.

NMDS ordination indicated that there is clear separation between FP and TF species assemblages with no overlap in their respective polygons (Figure 8; stress = 13.7).

The greatest divergence was along axis 1, with FP sites aggregated on the positive side and TF sites on the negative side of the axis. The mean within group Bray-Curtis dissimilarity was 0.68 for FP forest and 0.65 for TF forest while the between group dissimilarity value was 0.81, with significant differences between FP and TF species assemblages (F = 6.64, df = 1, p < 0.001). The area of the FP polygon (1.18) was larger than the TF polygon (0.72) but there was no difference in dispersion within multivariate space between FP and TF sites. The two forest types shared 32 out of the 143 total species (22%) and 21 of 54 total genera (38%). When singletons were eliminated from the comparison, FP and TF share 40% of species and 51% of genera. The low number of shared species and high between group mean Bray-Curtis dissimilarity value is reflected in the ordination biplot (Figure 8). To determine the influence of abundance on NMDS, ordination analysis was also completed with presence/absence data using Jaccard dissimilarity values. The resulting patterns for FP and TF forest are highly similar to the results based on abundance data, with the species assemblages for FP and TF forest significantly different (F = 3.86, df = 1, p < 0.001; Appendix E; Figure E1). The mean

Jaccard dissimilarly value between FP and TF forest was 0.87 while mean within forest type values were 0.80 for FP forest and 0.76 for TF forest. Using presence/absence data

208

affected the within forest type dissimilarity values more than the dissimilarity value between FP and TF forests.

Characteristic species

There were several species that had significant indicator values for sites within FP or TF forests (Table 2). Most of these species had high specificity values (A) for a forest type (mean = 0.96) while fidelity values were relatively lower (mean = 0.61). Paratachys

(Bembidiini) and Peruphorticus (Lachnophorini) had representative morphospecies for each forest type. There were several taxa with significantly associations for FP forest within the Scaritini (Nyctosyles and Clivina) as well as a Eucaerus (Lachnophorini) species while morphospecies from within Pentagonicini (Pentagonica) were significantly associated with TF sites.

Spatial autocorrelation

Since a linear trend was detected (F = 3.08, df = 2, p < 0.01), data were Hellinger- transformed and detrended prior to construction of the Mantel correlogram. There were six distance classes, with the smallest ranging from 0 to 32.5 meters; however, no significant spatial correlations were detected.

DISCUSSION

Numbers of individuals, richness and diversity

The results from this study indicate that understory Carabidae richness, diversity and species composition from eastern Ecuadorian lowland rainforests are influenced by forest type, but there was no significant difference in the number of Carabidae individuals by forest type. In contrast, Lamarre et al. (2016) found higher numbers of individuals for

209

Carabidae and Cicindelidae in seasonally flooded forests than terra firme forest in French

Guiana and Peru, with approximately 75% of carabid and 91% of cicindelid individuals collected in flooded forests. This disparity may be due to several factors, including differences in species compositions, sampling protocol, sample completeness and the time period and duration of sampling periods. In this study, all Cicindelini species occurred in both FP and TF forests but significantly more Cicindelini individuals were collected from TF forest. In other parts of the Amazon, Cicindelini has been found to be more abundant in terra firme forests (Zerm et al. 2001; Adis & Junk 2002), but higher abundance has also been reported for floodplain forests (Pearson & Derr 1986; Lamarre et al. 2016), suggesting that there is not a clear abundance pattern for Cicindelini between flooded and terra firme forests in Amazonia. Studies examining invertebrate assemblages have found lower abundances within the understory of flooded forests compared to terra firme forests (Pearson & Derr 1986; Mertl et al. 2009; Lamarre et al. 2016). Since the flooding cycle causes seasonal movements and shifts in the biotic communities of riparian areas, the time of year will likely affect the relative number of individuals occurring in the understory of floodplain and nearby terra firme forests (Irmler 1979a;

Adis 1981; Erwin & Adis 1982; Adis & Junk 2002). Abundance patterns for Carabidae vary in space and time, depending upon factors such as flooding duration, frequency and intensity.

Although mean species richness was not statistically different (but close), all other measures of richness and diversity were significantly higher in FP forest compared to TF forest which agrees with some previous studies (Erwin & Adis 1982; Pearson 1984;

Pearson & Derr 1986). The higher species richness and diversity of FP forest is likely

210

related to the higher number of rare species sampled in FP forest compared to TF forest.

An increase in Carabidae richness after a flooding event has been recorded for Central

Amazonian forests (Adis 1981), as well as for temperate forests (Bonn 2000; Ellis et al.

2001; Bonn et al. 2002; Lambeets et al. 2008, 2009). After an inundation event in the FP forests near TBS the litter layer and ground vegetation are covered by a thin layer of mud and sediments (pers. obs.; Mertl et al. 2009). The higher nutrient content in the soils of flooded forests could be attractive to carabids or the higher richness could be related to an increase in the abundance of food resources after an inundation event (Irmler 1979b;

Duivenvoorden & Lips 1995; Bonn 2000; Adis & Junk 2002; Haugaasen & Peres 2005;

Adis et al. 2010). Flooding could also increase the number of microhabitats available of a given area, potentially increasing the alpha diversity (Salo et al. 1986; Erwin 1991b;

Kalliola et al. 1991; Zerm & Adis 2001a; Wittmann et al. 2006). Flooding events could also play a role in the moderation of competitive exclusion, resulting in higher diversity for floodplains through higher levels of coexistence (Levin & Paine 1974; Connell 1978;

Erwin & Adis 1982; Zerm & Adis 2001a, 2001b).

Similar to the results of this study, studies in Central Amazonia have shown higher tiger beetle richness in terra firme forests (Adis et al. 1998; Zerm et al. 2001; Adis

& Junk 2002). Differences in carabid beetle richness and diversity patterns among localities within Amazonia may be due to local variations among sites, particularly for flooded forests as spatial and temporal changes are continuously occurring through erosion, deposition and changes in the river channel (Kalliola et al. 1991; Junk 2000).

Richness and diversity have been strongly tied to length, frequency and severity of flooding events (Bonn et al. 2002), which will vary among Amazonian rivers depending

211

on factors such as flooding cycles, topography and weather patterns (Junk et al. 1989;

Junk 2000; Tuomisto et al. 2003; Adis et al. 2010; Junk & Piedade 2010). Temporarily flooded forests with less severe and frequent flooding regimes, such as those near TBS, will likely show different patterns in richness and diversity compared to more intense flooding regimes in Central Amazonia.

Previous studies have suggested that a large proportion of Amazonian carabid beetles species inhabit riverine and wetland habitats (Erwin 1979; Erwin & Adis 1982).

Since Carabidae may have evolved and radiated from tropical wetland habitats (Erwin

1979, 1985), they may be well adapted to the changing environmental conditions along riverbanks (Adis 1982). Species specific traits within Carabidae assist in enduring unfavorable periods during the flooding cycle such as: synchronized life cycles, reproductive dormancy, submersion tolerance (Erwin 1979, 1985; Adis 1982; Erwin &

Adis 1982; Paarmann et al. 1982; Adis et al. 1986; Amorim et al. 1997; Zerm & Adis

2001b, 2003). The higher observed Carabidae richness and diversity values in FP compared to TF may relate to historical and evolutionary factors. Although Amazonian wetland and floodplain habitats may have experienced repeated reductions and expansions through changes in paleoclimate, it has been suggested they continued to be available possibly serving as refuges and therefore allowing species with adaptions to flooding cycles to persist (Erwin & Adis 1982; Kubitzki 1989; Irion et al. 1995, 1997,

2010; Müller et al. 1995; Haffer & Prance 2001; Adis & Junk 2002; Hoorn et al. 2010;

Wittmann et al. 2010).

The results from this study suggest that Carabidae species richness per unit area or search effort is higher in FP forests than TF forests, which may help explain the lower

212

observed richness and sample completeness in TF forest. Based on extrapolation curves, a substantial proportion of TF species remain to be collected, and the observed difference in species richness between FP and TF sites will decrease with greater sampling effort.

The nonparametric diversity estimator, ACE, also suggests TF has a higher true richness than FP. Only the Jack1 estimator indicates that FP forests will maintain higher overall richness with increased sampling effort. The geographical spatial coverage of TF sites was slightly greater than that of FP sites. Since terra firme forests are more extensive than flooded forests within Amazonia, richness and diversity estimates for terra firme forest in this study should increase relative to floodplain forest as sample area increases

(Rosenzweig 1995; ter Steege et al. 2000; Pitman et al. 2001; Burnham 2002; Tuomisto et al. 2003; Myster 2014). Compared to terra firme forests, flooded forests are limited to narrower bands adjacent to rivers and lakes and are more fragmented at the landscape level (ter Steege et al. 2000; Pitman et al. 2001). Larger more contiguous areas generally contain a greater number of species, which has been postulated to be reason for the higher richness of many taxa in terra firme compared to other Amazonian forest types (ter

Steege et al. 2000).

In terms of the number of individuals collected, hand sampling was not as efficient at TF sites as it was at FP sites (Riley et al. 2016). The smaller overall body size of carabids (excluding Cicindelini) from the TF forests may have contributed to hand collecting inefficiency for TF sites and be reflected in the lower species number for TF forests. However, other explanations exist including greater vertical stratification of

Carabidae in TF forest and more aggregated carabid distributions in FP forests (see

Chapter V for further discussion). Carabids in tropical regions are known to have patchy

213

distributions, making their abundance appear very low and higher sampling effort needed to sample a representative portion of the fauna (Erwin 1979; Paarmann et al. 2002; Erwin et al. 2005). Since the analysis of sample completeness suggests that increased sampling effort will likely uncover many more species, particularly for TF forest, additional sampling is needed for broader conclusions regarding Carabidae species richness and diversity in the Yasuní area.

There is little comparable data on how arthropod body size varies among

Amazonian forest types and few studies on tropical Carabidae body size at the community-level. A body size histogram for Carabidae in Erwin (1981) shows that the majority of species at Barro Colorado Island (BCI) Panama are 2 mm and between 4 – 8 mm, with a gap between 2 and 4 mm. Erwin (1983) found smaller body lengths for terra firme compared to floodplain forests for lowland tropical canopy Coleoptera and reported the overall mean body length for Carabidae was 3.23 mm, which is smaller than the averages for both FP and TF forests in this study. However, canopy carabid species assemblages are likely different from the ground fauna (Erwin 1982; Erwin et al. 2005;

Charles & Basset 2005; Stork et al. 2016).

Species assemblages

Rank abundance curves for tropical Carabidae communities have rarely if ever been reported. In temperate and boreal forests, carabid communities typically have a few dominant species with the majority of species being rare (Belaoussoff et al. 2003), similar to the assemblages for FP and TF forests in this study. Dominant and common species comprised 72.2% of the total number of individuals but only 13.3% of the total observed species. As with many tropical insect assemblages, the contribution of rare species was

214

high (Novotný & Basset 2000), with rarer species typically having a greater contribution to species assemblages compared to temperate communities. Singleton species can make up half of the observed species in tropical insect assemblages, even with a relatively high amount of sampling effort (Novotný & Basset 2000). In this study, singletons contributed slightly less than half of the overall species richness (43.8%) but the results from a canopy fogging in a nearby Ecuadorian terra firme forest found 28.6% were singletons

(Lucky et al. 2002). Methodological effects can also cause species to appear to be rare.

For example, a species may be abundant within a given area but undersampled due to inefficiency of the sampling methods used (Longino et al. 2002). Additional sampling is needed to provide additional information concerning rarity and determine if rare species in this study remain rare at higher sampling effort.

Abundant species were significantly more likely to be within the Cicindelini,

Lachnophorini, Loxandrini, Pentagonicini and Bembidiini tribes, with fewer Lebiini species collected than expected. Although Lebiini was by far the most species rich tribe

(S = 35), only two species were classified as common while the remaining 33 species occurred at low numbers. In contrast, all four of the Cicindelini species were classified as abundant. More commonly sampled species were sampled from a higher number of sampling sites within the study area, suggesting that they have wider spatial distributions than less frequently encountered species. This agrees with previously reported positive relationships between local population abundance and the occurrence at a greater proportion of sample sites (Gaston & Lawton 1988; Hanski 1982; Brown 1984; Niemelä

& Spence 1994). In this study, the presence or absence of rare species strongly influenced estimates of Carabidae richness, diversity and community-level differences between FP

215

and TF, with a small proportion of abundant species (13%) contributing 72% of the total number of individuals. The percentage of species considered common (15%) and rare

(85%) for Ecuadorian trees by Pitman et al. (2001) are similar to the percentages for this study (common = 13%, rare = 87%), suggesting that an oligarchy for carabid beetles exists within the study area. Additional sampling is needed in other localities to determine the geographic limits of this oligarchy. These results suggest that sampling by hand and FITs in the forests near TBS will likely uncover a predictable set of species from the aforementioned tribes.

Differences between FP and TF forests occurred not only at the species level but also at the tribal level. There were higher numbers of Scaritini and Lachnophorini individuals and/or species in FP forest and higher numbers of Pentagonicini and

Cicindelini individuals and/or species in TF forest. With the exception of species within

Lebia (Lebiini) and Paratachys (Bembidiini), all the characteristic species with significant IndVal values belonged to one of the tribes with significant differences in abundance or species number between forest types. Although life histories are unknown for most Amazonian carabid species, general life history information is known for many genera and tribes. For example, the majority of the Lebiini genera sampled in this study are not typically collected from the ground but more frequently are observed on fallen or standing tree trunks, vegetation, in suspended vegetation or associated with fig falls (i.e.

Apenes). Genera within Scaritini and Lachnophorini are typically more ground-dwelling whereas Pentagonicini and Cicindelini occur on vegetation (per. obs.; Erwin et al. 2012).

Many Lachnophorini prefer wetter habitats and many species within the Scaritini have been collected from wetter habitats including those within inundation forests (Erwin et al.

216

2012; Erwin & Zamorano 2014). The genus Peruphorticus (Lachnophorini) has been found in dry or wet leaf litter within Amazonia (Erwin & Zamorano 2014) therefore its distribution is likely less associated with wetter habitats. This supports the results of this study where Peruphorticus was sampled at several sites from both FP and TF forests, with one species found to be characteristic of FP while another species was characteristic of TF forest. Paratachys (Bembidiini) are a diverse group within Central and South

America and have been sampled from many different ground microhabitats (Erwin et al.

2012). This genus was also sampled from both forest types, with characteristic species for both FP and TF forests. Differences in microhabitat preferences among tribes and genera may be the reason for the observed differences in tribal species richness and numbers of individuals, as well as variations in species composition between forest types.

Community-level analyses indicate that distinct carabid species assemblages occurred in FP and TF forests, with this finding further supported by the relatively high number of characteristic species for each forest type. The high level of dissimilarity in species assemblages between the two forest types is reinforced by previous tropical studies of vegetation (ter Steege et al. 2000), birds (Remsen & Parker 1983), understory arthropods (Adis & Junk 2002; Lamarre et al. 2016), canopy arthropods (Erwin 1983;

Adis et al. 2010), spiders (Höffer 1997) and understory tiger beetles (Zerm et al. 2001).

In this study, only 32 of 143 species were collected in both forest types, suggesting floodplain forests maintain unique assemblages compared to neighboring non-flooded forests. The distinct assemblages suggest species within these forest communities have likely adapted to their respective forest type and/or exhibit strong habitat preference between flooded and terra firme forests. Differences in species assemblages may be

217

related to factors such as flooding regime adaptions, differences in plant composition, forest structure and resource availability between FP and TF forests. Near Manaus, Brazil,

Zerm et al. (2001) reported similar results for 25 tiger beetle species, with only two species occurring in both floodplain and terra firme forests. The spatial and temporal distribution of tiger beetle guilds have been shown to be determined by specific microhabitats within a particular habitat type (such as open areas, forest type) (Zerm and

Adis 2001a; Zerm et al. 2001; Adis & Junk 2002). Lamarre et al. (2016) reported significant associations for seasonally flooded forest or terra firme forests for several arthropod families, with carabid beetles demonstrating a particularly strong association with seasonally flooded forests. Temperate studies have also shown Carabidae communities to be sensitive to the effects of flooding (Ellis et al. 2001; Rothenbücher &

Schaefer 2006; Lambeets et al. 2008).

Since flooded forests oscillate between aquatic and terrestrial phases and there is high variability in flooding events among rivers, it is difficult to adequately measure the overall biodiversity and community composition of flooded forests (Junk et al. 1989;

Junk 2000). Floodplain communities consist of residents, migrants and visiting species

(Adis 1997; Junk 1997; Adis & Junk 2002). Species collected from flooded forests may be briefly inhabiting opportunists or long-term inhabitants which are highly adapted organisms (Adis 1997). The presence of both resident and migrant species in FP forest may have contributed to the higher observed richness and diversity in this study, as well as the higher number of rare species in FP forest. During flooding events, Carabidae may move into the canopy or sub-canopy strata via tree trunks (Irmler 1973, 1979a; Adis 1981;

Adis et al. 1997), may move horizontally following the waterline (Irmler 1979a; Adis et

218

al. 1997; Zerm & Adis 2003) or commonly fly to non-flooded uplands (Irmler 1973; Adis

1982; Erwin & Adis 1982; Adis et al. 1986) as the majority of carabid beetles living in tropical forests are winged and presumably capable of flight (Erwin 1979). Other specializations to flooding regimes have also been reported, including retreat to subsoil layers (e.g. a flightless Stratiotes species), aestivation and submersion tolerance (Adis

1982, 1997; Adis et al. 1986; Rothenbücher & Schaefer 2006; Lambeets et al. 2009).

Terborgh and Andresen (1998) suggest that vegetation occurring in floodplain forests may be more closely allied taxonomically to adjacent non-flooded forests rather than to other flooded forests. Conversely, specific adaptions to life in flooded forests could cause a convergent effect on floodplain community composition, causing them to be more similar to one another, even at broad geographic scales, than to adjacent terra firme forests (e.g. Lamarre et al., 2016). In this study, two genera (Paratachys and

Peruphorticus) had species with significant associations for FP as well as TF forests, and there were 14 shared tribes between the two forest types. A paired site design at a larger geographic scale could help determine if Carabidae richness, diversity and species composition are more similar within geographically distant forests of the same forest type or to an adjacent forest type.

This is one of the few ecological-based studies of Carabidae which compares two major Amazonian forest types but the conclusions are limited by the relatively small area sampled over a relatively short timeframe. Since many Amazonian Carabidae are seasonal, particularly in flooded habitats (Paarmann 1986; Pearson & Derr 1986; Adis &

Junk 2002; Erwin et al. 2005), additional sampling is needed within the drier season at

TBS. For a more comprehensive representation of Carabidae richness, diversity and

219

species assemblages, sampling should extend throughout the year (Pearson & Derr 1986;

Lucky et al. 2002; Erwin et al. 2005). For a detailed assessment of the overall richness and diversity of flooded forests and for measuring the effects of flooding events, sampling should occur before and after several flooding events (Junk 2000). In this study the ground and understory fauna were sampled, but to determine overall patterns in diversity and species assemblages for Carabidae within Yasuní forests, all forest strata should be included in community-level analyses (Basset et al. 2001, 2003; Chung 2004;

Charles & Basset 2005), especially since tropical arthropods in the forest canopy may be more species rich than the understory (Erwin 1982; Erwin et al. 2005; Charles & Basset

2005). Greater knowledge and understanding of Yasuní rainforests are of are imperative as anthropogenic factors, particularly oil exploration and extraction, continue to threaten these megadiverse areas (Finer et al. 2009, 2010; Bass et al. 2010).

ACKNOWLEDGMENTS

I am thankful to the Ecuadorian Ministry of the Environment for collection and exportation permits. Thank you to the staff and directors of Tiputini Biodiversity Station and the Universidad San Francisco de Quito for enabling this research by providing the necessary assistance and support to make this project possible. I am also grateful to K.

Swing, D. Mosquera, and H. Alverez for assisting with materials and providing vital information for this research project. C. Kliesen, K. Trujillo and J. Macanilla were fundamental to nightly hand sampling. Financial support was provided by Wake Forest

University, the Richter Scholarship Program and the Smithsonian Institution’s National

Museum of Natural History.

220

LITERATURE CITED

Adis, J. 1981. Comparative ecological studies of the terrestrial arthropod fauna in Central

Amazonian inundation-forests. Amazoniana 7:87–173.

Adis, J. 1982. Eco-entomological observations from the Amazon: II. Carabids are

adapted to inundation-forests! Coleopterists Bulletin 36:439–440.

Adis, J. 1997. Terrestrial invertebrates: Survival strategies, group spectrum, dominance

and activity patterns. Pages 299–317 in W. J. Junk, editor. The Central Amazon

Floodplain, Ecology of a Pulsing System. Springer-Verlag, Heidelberg.

Adis, J., and W. J. Junk. 2002. Terrestrial invertebrates inhabiting lowland river

floodplains of Central Amazonia and Central Europe: A review. Freshwater Biology

47:711–731.

Adis, J., Y. D. Lubin, and G. Montgomery. 1984. Arthropods from the canopy of

inundated and terra firme forests near Manaus, Brazil, with critical considerations on

the pyrethrum-fogging technique. Studies on Neotropical Fauna and Environment

19:223–236.

Adis, J., W. Paarmann, and T. L. Erwin. 1986. On the natural history and ecology of

small terrestrial ground-beetles (Col.: Bembidiini: Tachynia: Polyderis from an

Amazonian black-water inundation forest. Pages 413–427 in P. J. den Boer, M. L.

Luff, D. Mossakowski, and F. Weber, editors. Carabid Beetles: Their Adaptions and

Dynamics: XVIIth International Congress of Entomology, Hamburg, 1984. Gustav

Fischer, Stuttgart.

Adis, J., M. A. Amorim, T. L. Erwin, and T. Bauer. 1997. On ecology, life history and

221

survival strategies of a wing-dimorphic ground beetle (Col.: Carabidae: Odacanthini:

Colliuris) inhabiting Central Amazonian inundation forests. Studies on Neotropical

Fauna and Environment 32:174–192.

Adis, J., T. L. Erwin, L. D. Battirola, and S. M. Ketelhut. 2010. The importance of

Amazonian floodplain forests for animal biodiversity: Beetles in canopies of

floodplain and upland forests. Pages 313–325 in W. J. Junk, M. T. F. Piedade, F.

Wittmann, J. Schöngart, and P. Parolin, editors. Amazonian Floodplain Forests:

Ecophysiology, Biodiversity and Sustainable Management. Springer

Science+Business Media B.V., New York, New York.

Adis, J., W. Paarmann, M. A. Amorim, E. Arndt, and C. R. V Fonseca. 1998. On

occurrence, habitat specificity and natural history of adult tiger beetles near Manaus,

Central Amazonia, and key to the larvae of tiger beetle genera. Acta Amazonia

28:247–272.

Amorim, M. A., J. Adis, and W. Paarmann. 1997. Ecology and adaptions of Pentacomia

egregia (Chaudoir) (Cicindelinae: Carabidae) to Central Amazonian floodplains.

Ecotropica 3:71–82.

Anderson, M. J. 2001. A new method for non-parametric multivariate analysis of

variance. Austral Ecology 26:32–46.

Anderson, M. J. 2006. Distance-based tests for homogeneity of multivariate dispersions.

Biometrics 62:245–53.

Anderson, M. J., K. E. Ellingsen, and B. H. McArdle. 2006. Multivariate dispersion as a

measure of beta diversity. Ecology Letters 9:683–693.

222

Baccaro, F. B., I. F. Rocha, B. E. G. del Aguila, J. Schietti, T. Emilio, J. L. P. D. V. Pinto,

A. P. Lima, and W. E. Magnusson. 2013. Changes in ground-dwelling ant functional

diversity are correlated with water-table level in an Amazonian terra firme forest.

Biotropica 45:755–763.

Balslev, H., J. Luteyn, B. Øllgaard, and L. B. Holm-Nielsen. 1987. Composition and

structure of adjacent unflooded and floodplain forest in Amazonian Ecuador. Opera

Botanica 92:37–57.

Baraloto, C. et al. 2011. Disentangling stand and environmental correlates of

aboveground biomass in Amazonian forests. Global Change Biology 17:2677–2688.

Bass, M. S. et al. 2010. Global conservation significance of Ecuador’s Yasuní National

Park. PLoS ONE 5:e8767.

Basset, Y. et al. 2001. Stratification and diel activity of arthropods in a lowland rainforest

in Gabon. Biological Journal of the Linnean Society 72:585–607.

Basset, Y. et al. 2012. Arthropod diversity in a tropical forest. Science 338:1481–1484.

Basset, Y., P. Hammond, H. Barrios, J. D. Holloway, and S. Miller. 2003. Vertical

stratification of arthropod assemblages. Pages 17–27 in Y. Basset, V. Novotny, S. E.

Miller, and R. L. Kitching, editors. Arthropods of Tropical Forests: Spatio-Temporal

Dynamics and Resource Use in the Canopy. Cambridge University Press,

Cambridge, UK.

Basualdo, C. V. 2011. Choosing the best non-parametric richness estimator for benthic

macroinvertebrates databases. Revista de al Sociedad Entomológica Argentina

70:27–38.

223

Bates, D., M. Maechler, B. M. Bolker, and S. Walker. 2015. Fitting linear mixed-effects

models using lme4. Journal of Statistical Software 67:1–48.

Becker, R. A., A. R. Wilks, R. Brownrigg, and T. P. Minka. 2015. maps: Draw

Geographical Maps.

Belaoussoff, S., P. G. Kevan, S. Murphy, and C. Swanton. 2003. Assessing tillage

disturbance on assemblages of ground beetles (Coleoptera: Carabidae) by using a

range of ecological indices. Biodiversity and Conservation 12:851–882.

Bivand, R. S., E. J. Pebesma, and V. Gomez-Rubio. 2013. Applied Spatial Data Analysis

with R. 2nd edition. Springer, New York, New York.

Blake, J. G., D. Mosquera, J. Guerra, B. A. Loiselle, D. Romo, and K. Swing. 2011.

Mineral licks as diversity hotspots in lowland forest of eastern Ecuador. Diversity

3:217–234.

Bonn, A. 2000. Flight activity of carabid beetles on a river margin in relation to

fluctuating water levels. Pages 147–160 in P. Brandmayer, G. L. Lovei, T. Z.

Brandmayr, A. Casale, and A. Vigna Taglianti, editors. Natural History and Applied

Ecology of Carabid Beetles. Pensoft Publishers, Moscow.

Bonn, A., K. Hagen, and D. Wohlgemuth-Von Reiche. 2002. The significance of flood

regimes for carabid beetle and spider communities in riparian habitats - A

comparison of three major rivers in Germany. River Research and Applications

18:43–64.

Borcard, D., F. Gillet, and P. Legendre. 2011. Numerical Ecology in R. R. Gentleman, K.

Hornik, and G. Parmigiani, editors. Use R! Series. Springer Science, New York,

224

New York.

Borcard, D., and P. Legendre. 2012. Is the Mantel correlogram powerful enough to be

useful in ecological analysis? A simulation study. Ecology 93:1473–1481.

Brose, U., and N. D. Martinez. 2004. Estimating the richness of species with variable

mobility. Oikos 105:292–300.

Brown, J. H. 1984. On the relationship between abundance and distribution of species.

American Naturalist 124:255–279.

Brunsdon, C., and H. Chen. 2014. GISTools: Some further GIS capabilities for R. R

package version 0.7-4. Available from: https://CRAN.R-

project.org/package=GISTools

Burnham, K. P., and W. S. Overton. 1979. Robust estimation of population size when

capture probabilities vary among animals. Ecology 60:927–936.

Burnham, R. J. 2002. Dominance, diversity and distribution of lianas in Yasuní, Ecuador:

who is on top? Journal of Tropical Ecology 18:845–864.

Burnham, R. J., N. C. A. Pitman, K. R. Johnson, and P. Wilf. 2001. Habitat-related error

in estimating temperatures from leaf margins in a humid tropical forest. American

Journal of Botany 88:1096–1102.

Campbell, D. G., D. C. Daly, G. T. Prance, and U. N. Maciel. 1986. Quantitative

ecological inventory of terra firme and várzea tropical forest on the Rio Xingu,

Brazilian Amazon. Brittonia 38:369–393.

Carlton, C., M. Dean, and A. Tishechkin. 2004. Diversity of two beetle taxa at a western

Amazonian locality (Coleoptera: Histeridae; Staphylinidae, Pselaphinae).

225

Coleopterists Bulletin 58:163–170.

Cassola, F., and D. L. Pearson. 2001. Neotropical tiger beetles (Coleoptera: Cicindelidae):

checklist and biogeography. Biota Colombiana 2:3–24.

Chambers, J. M. 2013. SoDA: Functions and examples for “software for data analysis.” R

package version 1.0-6. Available from: https://CRAN.R-project.org/package=SoDA

Chao, A. 1984. Nonparametric estimation of the number of classes in a population.

Scandinavian Journal of Statistics 11:265–270.

Chao, A., N. J. Gotelli, T. C. C. Hsieh, E. L. Sander, K. H. Ma, R. K. Colwell, and A. M.

Ellison. 2014. Rarefaction and extrapolation with Hill numbers: A framework for

sampling and estimation in species diversity studies. Ecological Monographs 84:45–

67.

Chao, A., and S. M. Lee. 1992. Estimating the number of classes via sample coverage.

Journal of the American Statistical Association 87:210–217.

Chao, A., and M. C. K. Yang. 1993. Stopping rules and estimation for recapture

debugging with unequal failure rates. Biometrika 80:193–201.

Charles, E., and Y. Basset. 2005. Vertical stratification of leaf-beetle assemblages

(Coleoptera: Chrysomelidae) in two forest types in Panama. Journal of Tropical

Ecology 21:329–336.

Chung, A. Y. C. 2004. Vertical stratification of beetles (Coleoptera) using flight intercept

traps in a lowland rainforest of Sabah, Malaysia. Sepilok Bulletin 1:29–41.

Chazdon, R. L., R. K. Colwell, J. S. Denslow, and M. R. Guariguata. 1998. Statistical

methods for estimating species richness of woody regeneration in primary and

226

secondary rain forests of Northeastern Costa Rica. Pages 285–309 in F. Dallmeier

and J. A. Comiskey, editors. Forest Biodiversity Research, Monitoring and

Modeling: Conceptual Background and Old World Case Studies Man and the

Biosphere Series. Vol 20. Parthenon Publishing, Paris.

Clarke, K. R. 1993. Non-parametric multivariate analyses of changes in community

structure. Australian Journal of Ecology 18:117–143.

Colwell, R. K. et al. 2012. Models and estimators linking individual-based and sample-

based rarefaction, extrapolation and comparison of assemblages. Journal of Plant

Ecology 5:3–21.

Colwell, R. K., and J. A. Coddington. 1994. Estimating terrestrial biodiversity through

extrapolation. Philosophical Transactions of the Royal Society of London Series B

345:101–18.

Colwell, R. K., C. X. Mao, and J. Chang. 2004. Interpolating, extrapolating, and

comparing incidence-based species accumulation curves. Ecology 85:2717–2727.

Condit, R. et al. 2002. Beta-diversity in tropical forest trees. Science 295:666–9.

Connell, J. H. 1978. Diversity in tropical rain forests and coral reefs. Science 199:1302–

1310.

De Cáceres, M., and P. Legendre. 2009. Associations between species and groups of sites:

Indices and statistical inference. Ecology 90:3566–3574.

Dufrene, M., and P. Legendre. 1997. Species assemblages and indicator species: The

need for a flexible asymmetrical approach. Ecological Monographs 67:345–366.

Duivenvoorden, J. E. 1995. Tree species composition and rain forest-environment

227

relationships in the middle Caquetá area, Colombia, NW Amazonia. Vegetatio

120:91–113.

Duivenvoorden, J. F., and J. M. Lips. 1995. A Land-Ecological Study of Soils,

Vegetataion, and Plant Diversity in Colombian Amazonia. Tropenbos Foundation,

Leiden.

Dunn, R. R. 2005. Modern insect extinctions, the neglected majority. Conservation

Biology 19:1030–1036.

Dyer, L. A. et al. 2007. Host specificity of Lepidoptera in tropical and temperate forests.

Nature 448:696–9.

Ellis, L. M., C. S. Crawford, and M. C. M. Jr. 2001. Influence of annual flooding on

terrestrial arthropod assemblages of a Rio Grande riparian forest. Regulated Rivers:

Research and Management 20:1–20.

Erwin, T. L. 1979. Thoughts on the evolutionary history of ground beetles: Hypotheses

generated from comparative faunal analyses of lowland forest sites in temperate and

tropical regions. Pages 539–592 in T. L. Erwin, G. E. Ball, and D. R. Whitehead

editors. Carabid Beetles: Their Evolution, Natural History, and Classification. Dr W.

Junk Publishers, Dordrecht.

Erwin, T. L. 1982. Tropical forests: Their richness in Coleoptera and other arthropod

species. Coleopterists Bulletin 36:74–75.

Erwin, T. L. 1981. Natural history of Plummers Island, Maryland. XXVI. The ground

beetles of a temperate forest site (Coleoptera: Carabidae): An analysis of fauna in

relation to size, habitat selection, vagility, seasonality, and extinction. Bulletin of the

228

Biological Society of Washington 5:105–224.

Erwin, T. L. 1983. Beetles and other arthropods of the tropical forest canopies at Manaus,

Brazil, sampled by insecticidal fogging. Pages 59–75 in S. Sutton, T. Whitmore, and

A. Chadwick, editors. Tropical Rainforest: Ecology and Management. Blackwell

Scientific, Oxford.

Erwin, T. L. 1985. The taxon pulse: A general pattern of lineage radiation and extinction

among carabid beetles. Pages 438–472 in G. E. Ball, editor. Taxonomy, Phylogeny

and Zoography of Beetles and Ants. Dr. W. Junk Publishers, Dordrecht.

Erwin, T. L. 1991a. How many species are there?: Revisited. Conservation Biology

5:330–333.

Erwin, T. L. 1991b. Natural history carabid beetles BIOLAT Biological Station, Rio

Manu, Pakitza, Peru. Revista Peruana de Entomologia 33:1–85.

Erwin, T. L. 1996. Biodiversity at its utmost: tropical forest beetles. Pages 27–40 in M. L.

Reaka-kudla, D. E. Wilson, and E. O. Wilson, editors. Biodiversity II:

Understanding and Protecting our Biological Resources. Joseph Henry Press,

Washington, DC.

Erwin, T. L., and D. H. Kavanaugh. 1981. Systematics and zoogeography of Bembidion

Latreille: I. The carlhi and erasum groups of western North America (Coleoptera:

Carabidae, Bembidiini). Entomologica Scandinavica Supplement 15:33–72.

Erwin, T. L., and J. Adis. 1982. Amazonian inundation forests their role as short-term

refuges and generators of speices richness and taxon pulses. Pages 358–371 in G. T.

Prance, editor. Biological Diversification in the Tropics. Proc. 5th International

229

Symposium of Association of Tropical Biology, Venezuela, 1979, Caracas.

Erwin, T. L., and C. J. Geraci. 2009. Amazonian rainforests and their richness of

Coleoptera, a dominant life form in the critical zone of the Neotropics. Pages 49–67

in R. Foottit and P. B. Adler, editors. Insect Biodiversity: Science and Society, 1st

edition. Blackwell Publishing, Oxford.

Erwin, T. L., and P. J. Johnson. 2000. Naming species, a new paradigm for crisis

management in taxonomy: Rapid journal validation of scientific names enhanced

with more complete descriptions on the internet. Coleopterists Bulletin 54:269–278.

Erwin, T. L., and L. Zamorano. 2014. A synopsis of the tribe Lachnophorini, with a new

genus of Neotropical distribution and a revision of the Neotropical genus Asklepia

Liebke, 1938 (Insecta, Coleoptera, Carabidae). ZooKeys 430:1–108.

Erwin, T. L., C. Micheli, and G. Hevel. 2012. The beetle family Carabidae of Guyane

(French Guyana): The tribes, genera, and number of species known, with notes on

their ways of life. (Insecta: Coleoptera: Adephaga). Le Coléoptériste 5:1–92.

Erwin, T. L., M. C. Pimienta, O. E. Murillo, and V. Aschero. 2005. Mapping patterns of

β-diversity for beetles across the western Amazon Basin: A preliminary case for

improving inventory methods and conservation strategies. Proceedings of the

California Academy of Sciences 56:72–85.

Fine, P. V. A., M. R. Metz, J. Lokvam, I. Mesones, J. M. A. Zuniga, G. P. A. Lamarre, M.

V. Pilco, and C. Baraloto. 2013. Insect herbivores, chemical innovation, and the

evolution of habitat specialization in Amazonian trees. Ecology 94:1764–1775.

Finer, M., C. N. Jenkins, S. L. Pimm, B. Keane, and C. Ross. 2008. Oil and gas projects

230

in the Western Amazon: Threats to wilderness, biodiversity, and indigenous peoples.

PLoS ONE 3:e2932.

Finer, M., R. Moncel, and C. N. Jenkins. 2010. Leaving the oil under the Amazon:

Ecuador’s Yasuní-ITT Initiative. Biotropica 42:63–66.

Finer, M., V. Vijay, F. Ponce, C. N. Jenkins, and T. R. Kahn. 2009. Ecuador’s Yasuní

Biosphere Reserve: A brief modern history and conservation challenges.

Environmental Research Letters 4:34005.

Fox, J., and S. Weisberg. 2010. An {R} Companion to Applied Regression. 2nd edition.

Sage, Thousand Oaks, CA. Avialable from:

http://socserv.socsci.mcmaster.ca/jfox/Books/Companion

Gentry, A. H. 1988. Changes in plant community diversity and floristic composition on environmental and geographical gradients. Annals of the Missouri Botanical Garden

75:1–34.

Gentry, A. H. 1992. Tropical forest biodiversity: distributional and their patterns

conservational significance. Oikos 63:19–28.

Godfray, H. C. J., O. T. Lewis, and J. Memmott. 1999. Studying insect diversity in the

tropics. Philosophical Transactions of the Royal Society of London. Series B

354:1811–1824.

Goslee, S. C., and D. L. Urban. 2013. The ecodist package for dissimilarity-based

analysis of ecological data. Journal of Statistical Software 22:1-19.

Gotelli, N. J., and R. K. Colwell. 2001. Quantifying biodiversity: Procedures and pitfalls

in the measurement and comparison of species richness. Ecology Letters 4:379–391.

231

Gotelli, N. J., and R. K. Colwell. 2011. Estimating species richness. Pages 39–54 in A. E.

Magurran and B. J. McGill, editors. Biological Diversity: Frontiers in Measuring

and Assessment. Oxford University Press, New York.

Grimbacher, P. S., and N. E. Stork. 2009. Seasonality of a diverse beetle assemblage

inhabiting lowland tropical rain forest in Australia. Biotropica 41:328–337.

Haffer, J., and G. T. Prance. 2001. Climatic forcing of evolution in Amazonia during the

Cenozoic: On the refuge theory of biotic differentiation. Amazoniana 16:579–607.

Hamilton, A. J. et al. 2013. Estimating global arthropod species richness: Refining

probabilistic models using probability bounds analysis. Oecologia 171:357–365.

Hamilton, A. J., Y. Basset, K. K. Benke, P. S. Grimbacher, S. E. Miller, V. Novotný, G.

A. Samuelson, N. E. Stork, G. D. Weiblen, and J. D. L. Yen. 2010. Quantifying

uncertainty in estimation of tropical arthropod species richness. American Naturalist

176:90–95.

Hammond, P. 1992. Species inventory. Pages 17–39 in B. Groombridge, editor. Global

Biodiversity: Status of the Earth’s Living Resources. Chapman and Hall, London.

Hanski, I. 1982. Dynamics of regional distribution: The core and satellite species

hypothesis. Oikos 38:210–221.

Haugaasen, T., and C. A. Peres. 2005. Mammal assemblage structure in Amazonian

flooded and unflooded forests. Journal of Tropical Ecology 21:133–145.

Heltshe, J. F., and N. E. Forrester. 1983. Estimating species richness using the Jackknife

procedure. Biometrics 39:1–11.

Herrera-MacBryde, O., and D. A. Neill. 1997. Yasuní National Park and the Waorani

232

Ethnic Reserve, Ecuador. S. D. Davis, V. H. Heywood, O. Herrera-MacBryde, J.

Villa-Lobos, and A. C. Hamilton, editors. Centres of plant diversity: A Guide and

Strategy for their Conservation. Vol. 3. WWF and IUCN, Cambridge, UK.

Hijmans, R. J. 2015. raster: Geographic data analysis and modeling. R package version

2.5-2. Available from: https://CRAN.R-project.org/package=raster

Höffer, H. 1997. The spider communities. Pages 373–383 in W. J. Junk, editor. The

Central Amazon Floodplain, Ecology of a Pulsing System. Springer-Verlag, Berlin

Heidelberg.

Holm, S. 1979. A simple sequentially rejective multiple test procedure. Scandinavian

Journal of Statistics 6:65–70.

Hoorn, C. et al. 2010. Amazonia through time: Andean uplift, climate change, landscape

evolution, and biodiversity. Science 330:927.

Hortal, J., P. A. V. Borges, and C. Gaspar. 2006. Evaluating the performance of species

richness estimators: Sensitivity to sample grain size. Journal of Animal Ecology

75:274–287.

Hothorn, T., F. Bretz, and P. Westfall. 2008. Simultaneous inference in general

parametric models. Biometrical Journal 50:346–363.

Hsieh, T. C., K. H. Ma, and A. Chao. 2015. iNEXT: An R package for interpolation and

extrapolation of species diversity (Hill numbers). R package version 2.0.8, Available

from: http://chao.stat.nthu.edu.tw/blog/software-download

Hunt, T. et al. 2007. A comphrensive phylogeny of beetles reveals the evolutionary

origins of a superradiation. Science 318:1913–1916.

233

Irion, G., J. A. S. N. de Mello, J. Morais, M. T. F. Piedade, W. J. Junk, and L. Garming.

2010. Development of Amazon valley during the middle to late Quaternary:

sedimentological and climatological observations. Pages 27–42 in W. J. Junk, M. T.

F. Piedade, P. Parolin, F. Wittmann, and J. Schöngart, editors. Amazonian

Floodplain Forests: Ecophysiology, Biodiversity and Sustainable Management.

Springer Science+Business Media B.V., Dordrecht.

Irion, G., W. J. Junk, and J. A. S. N. de Mello. 1997. The large Central Amazonian river

floodplains near Manaus: Geological, climatological, hydrological, and

geomorphical aspects. Pages 23–46 in W. J. Junk, editor. The Central Amazon

Floodplain, Ecology of a Pulsing System. Springer-Verlag Berlin Heidelberg.

Irion, G., J. Müller, J. Nunes de Mello, and W. J. Junk. 1995. Quaternary geology of the

Amazonian Lowland. Geo-Marine Letters 15:172–178.

Irmler, U. 1973. Population-dynamic and physiological adaption of Pentacomia egregia

Chaud. (Col. Cicindelidae) to the Amazonian inundation forest. Amazoniana 4:292–

227.

Irmler, U. 1979a. Abundance fluctuations and habitat changes of soil beetles in Central

Amazonian inundation forests (Coleoptera: Carabidae, Staphylinidae). Studies on

Neotropical Fauna and Environment 14:1–16.

Irmler, U. 1979b. Considerations on structure and function of the “Central-Amazonian

Inundation Forest Ecosystem” with particular emphasis on selected soil animals.

Oecologia 43:1–18.

Jost, L. 2006. Entropy and diversity. Oikos 113:363–375.

234

Junk, W. J. 1997. General aspects of floodplain ecology with special reference to

Amazonian floodplains. Pages 3–18 in W. J. Junk, editor. The Central Amazon

Floodplain, Ecology of a Pulsing System. Vol. 126. Springer-Verlag Berlin

Heidelberg.

Junk, W. J. 2000. Mechanisms for development and maintenance of biodiversity

Neotropical. Pages 119–139 in B. Gopal, W. J. Junk, and J. A. Davis, editors.

Biodiversity in Wetlands: Assessment, Function and Conservation. Vol 1. Backhuys

Publishers, Leiden.

Junk, W. J., P. B. Bayley, and R. E. Sparks. 1989. The flood pulse concept in river-

floodplain systems. Pages 110–127 in D. P. Dodge, editor. Proceedings of the

International Large River Symposium. Canadian Special Publication of Fisheries

and Aquatic Sciences 106.

Junk, W. J., and M. T. F. Piedade. 2010. An introduction to South American wetland

forests: Distribution, definitions and general characterization. Pages 3–25 in W. J.

Junk, M. T. F. Piedade, F. Wittmann, J. Schöngart, and P. Parolin, editors.

Amazonian Floodplain Forests: Ecophysiology, Biodiversity and Sustainable

Management. Springer Science+Business Media B.V., New York, New York.

Kalliola, R., J. Salo, M. Puhakka, and M. Rajasilta. 1991. New site formation and

colonizing vegetation in primary succession on the western Amazon floodplains.

Journal of Ecology 79:877–901.

Kapos, V., E. Pallant, A. Bien, and S. Freskos. 1990. Gap frequencies in lowland rain

forest sites on contrasting soils in Amazonian Ecuador. Biotropica 22:218–225.

235

Kellert, S. R. 1993. Values and perceptions of invertebrates. Conservation Biology

7:845–855.

Kim, K. C. 1993. Biodiversity, conservation and inventory: Why insects matter.

Biodiversity and Conservation 2:191–214.

Kitching, R. L. 2006. Crafting the pieces of the diversity jigsaw puzzle. Science

313:1055–1057.

Kremen, C., R. K. Colwell, T. L. Erwin, D. D. Murphy, R. F. Noss, and M. A. Sanjayan.

1993. Terrestrial arthropod assemblages: Their use in conservation planning.

Conservation Biology 7:796–808.

Kubitzki, K. 1989. The ecogeographical differentiation of Amazonian inundation forests.

Plant Systematics and Evolution 162:285–304.

Lamarre, G. P. A., B. Hérault, P. V. A. Fine, V. Vedel, R. Lupoli, I. Mesones, and C.

Baraloto. 2016. Taxonomic and functional composition of arthropod assemblages

across contrasting Amazonian forests. Journal of Animal Ecology 85:227–239.

Lambeets, K., M. L. Vandegehuchte, J. P. Maelfait, and D. Bonte. 2009. Integrating

environmental conditions and functional life-history traits for riparian arthropod

conservation planning. Biological Conservation 142:625–637.

Lambeets, K., M. L. Vandegehuchte, J. P. Maelfait, and D. Bonte. 2008. Understanding

the impact of flooding on trait-displacement and shifts in assemblage structure of

predatory arthropods on river banks. Journal of Animal Ecology 77:1162–1174.

Lee, S. M., and A. Chao. 1994. Estimating population size via coverage for closed

capture-recapture models. Biometrics 50:88–97.

236

Legendre, P., and M. J. Anderson. 1999. Distance-based redundancy analysis: Testing

multispecies responses in multifactorial ecological experiments. Ecological

Monographs 69:1–24.

Legendre, P., and L. Legendre, editors. 2012. Numerial Ecology. Vol 24, 3rd edition.

Elsevier, Amsterdam.

Leimbeck, R. M., and H. Balslev. 2001. Species richness and abundance of epiphytic

Araceae on adjacent floodplain and upland forest in Amazonian. Biodiversity and

Conservation 10:1579–1593.

Levin, S. A., and R. T. Paine. 1974. Disturbance, patch formation, and community

structure. Proceedings of the National Academy of Sciences 71:2744–2747.

Longino, J. T., J. Coddington, and R. Colewell. 2002. The ant fauna of a tropical rain

forest: Estimating species richness three different ways. Ecology 83:689–702.

Lövei, G. L., and K. D. Sunderland. 1996. Ecology and behavior of ground beetles

(Coleoptera: Carabidae). Annual Review of Entomology 41:231–256.

Lucky, A., T. L. Erwin, and J. Witman. 2002. Temporal and spatial diversity and

distribution of arboreal Carabidae (Coleoptera) in a western Amazonian rain forest.

Biotropica 34:376–386.

Majer, J. D., and J. H. C. Delabie. 1994. Comparison of the ant communities of annually

inundated and terra firme forests at Trombetas in the Brazilian Amazon. Insectes

Sociaux 41:343–359.

Mantel, N. 1967. The detection of disease clustering and a generalized regression

approach. Cancer Research 27:209–220.

237

Mapa de Vegetación del Ecuador. 2012. Sistema de clasificación de los ecosistemas del

Ecuador continental. Subsecretaría de Patrimonio Natural, Ministerio de Ambiente

del Ecuador.

May, R. M. 1988. Many species are there on Earth? Science 241:1441–1449.

May, R. M. 2010. Tropical arthropod species, more or less? Science 329:41–42.

McArdle, B. H., and M. J. Anderson. 2001. Fitting multivariate models to community

data: A comment on distance-based redundancy analysis. Ecology 82:290–297.

McCune, B., and J. B. Grace. 2002. Analysis of Ecological Communities. MjM software

design, Gleneden Beach, Oregon.

Mertl, A. L., K. T. R. Wilkie, J. F. A. Traniello, K. T. Ryder Wilkie, and J. F. A.

Traniello. 2009. Impact of flooding on the species richness, density and composition

of Amazonian litter-nesting ants. Biotropica 41:633–641.

Minchin, P. R. 1987. An evaluation of the relative robustness of techniques for ecological

ordination. Vegetatio 69:89–107.

Mora, C., D. P. Tittensor, S. Adl, A. G. B. Simpson, and B. Worm. 2011. How many

species are there on earth and in the ocean? PLoS Biology 9:1–8.

Müller, J., G. Irion, J. Nunes de Mello, and W. J. Junk. 1995. Hydrological changes of

the Amazon during the last glacial-interglacial cycle in Central Amazonia (Brazil).

Naturwissenschaften 82:232–235.

Myster, R. W. 2014. Várzea forest vs. terra firme floristics and physical structure in the

Ecuadorean Amazon. Ecotropica 20:35–44.

Nekola, J. C., A. L. Šizling, A. G. Boyer, and D. Storch. 2008. Artifactions in the log-

238

transformation of species abundance distributions. Folia Geobotanica 43:259–268.

Niemelä, J., and J. Spence. 1994. Distribution of forest dwelling carabids (Coleoptera):

spatial scale and the concept of communities. Ecography 17:166–175.

Novotný, V., and Y. Basset. 2000. Rare species in communities of tropical insect

herbivores: Pondering the mystery of singletons. Oikos 89:564–572.

Ødegaard, F. 2000. How many species of arthropods? Erwin’s estimate revised.

Biological Journal of the Linnean Society 71:583–597.

Oden, N. L., and R. R. Sokal. 1986. Directional autocorrelation: An extension of spatial

correlograms to two dimensions. Systematic Zoology 35:608–617.

Oksanen, J. 2015. Multivariate analysis of ecological communities in R: vegan tutorial. R

package version 3.0.1. Available from:

http://cc.oulu.fi/~jarioksa/opetus/metodi/vegantutor.pdf

Oksanen, J., F. G. Blanchet, R. Kindt, P. Legendre, P. R. Minchin, R. B. O’Hara, G. L.

Simpson, P. Solymos, M. H. H. Stevens, and H. Wagner. 2016. vegan: Community

Ecology Package. Available from: https://CRAN.R-project.org/package=vegan

Oliver, I., and A. J. Beattie. 1996. Designing a cost-effective invertebrate survey: A test

of methods for rapid assessment of biodiversity. Ecological Applications 6:594–607.

Paarmann, W. 1986. Seasonality and its control by enviromental factors in tropical

ground beetles. Pages 157–171 in P. den Boer, editor. Carabid Beetles: Their

Adaptations and Dynamics: XVIIth International Congress of Entomology,

Hamburg, 1984. Gustav Fischer, New York.

Paarmann, W., U. Irmler, and J. Adis. 1982. Pentacomia egregia Chaud. (Carabidae,

239

Cicindelinae), a univoltine species in the Amazonian inundation forest. Coleopterists

Bulletin 36:183–188.

Paarmann, W. et al. 2002. The structure of ground beetle assemblages (Coleoptera:

Carabidae) at fruit falls of Melastomataceae trees in a Brazilian terra firme rain

forest. Biotropica 34:368–375.

Palmer M. W. 1990. The estimation of species richness by extrapolation. Ecology

71:1195–1198.

Parolin, P. et al. 2004. Central Amazonian floodplain forests: Tree adaptations in a

pulsing system. The Botanical Review 70:357–380.

Payne, L. X., D. E. Schindler, J. K. Parrish, and S. A. Temple. 2005. Quantifying spatial

pattern with evenness indices. Ecological Applications 15:507–520.

Pearson, D. L. 1984. The tiger beetles (Coleoptera: Cicindelidae) of the Tambopata

reserved zone, Madre de Dios, Peru. Revista Peruana de Entomologia 27:15–24.

Pearson, D. L. 1988. Biology of tiger beetles. Annual Review Entomology 33:123–147.

Pearson, D. L., and J. A. Derr. 1986. Seasonal patterns of lowland forest floor arthropod

abundance in southeastern Peru. Biotropica 18:244–256.

Pebesma, E. J., and R. S. Bivand. 2005. Classes and methods for spatial data in R. R

News 5. Available from: http://cran.r-project.org/doc/Rnews/

Pitman, N. C. A. 2000. A large-scale inventory of two Amazonian tree communities. PhD

thesis. Duke University, Durham, North Carolina, USA.

Pitman, N. C. A., J. W. Terborgh, M. R. Silman, P. Núñez, D. A. Neill, C. E. Cerón, W.

A. Palacios, and M. Aulestia. 2001. Dominance and distribution of tree species in

240

upper Amazonian terra firme forests. Ecology 82:2101–2117.

Pitman, N. C. A., J. E. G. Andino, M. Aulestia, C. E. Cerón, D. A. Neill, W. Palacios, G.

Rivas-Torres, M. R. Silman, and J. W. Terborgh. 2014. Distribution and abundance

of tree species in swamp forests of Amazonian Ecuador. Ecography 37:902–915.

Prance, G. T. 1979. Notes on the Vegetation of Amazonia III. The terminology of

Amazonian forest types subject to inundation. Brittonia 31:26–38.

Primack, R. B., and R. T. Corlett. 2005. Insects: Diverse, abundant, and ecologically

important. Page 198 in R. B. Primack and R. T. Corlett, editors. Tropical Rain

Forests: An Ecological and Biogeographical Comparison. Blackwell Publishing,

Oxford, UK.

R Core Team. 2016. R: A Language and Environment for Statistical Computing. R

Foundation for Statistical Computing, Vienna, Austria. Available from:

https://www.R-project.org/

Remsen, J. V, and T. A. Parker. 1983. Contribution of river-created habitats to bird

species richness in Amazonia. Biotropica 15:223–231.

Riley, K. N., and R. A. Browne. 2011. Changes in ground beetle diversity and

community composition in age structured forests (Coleoptera, Carabidae). ZooKeys

621:601–21.

Riley, K. N., R. A. Browne, and T. L. Erwin. 2016. Results from two sampling

techniques for carabid beetles (Coleoptera: Carabidae) in temporarily flooded and

terra firme rainforest of western Amazonia. Studies on Neotropical Fauna and

Environment 51:78–95.

241

Rohr, J. R., C. G. Mahan, and K. C. Kim. 2007. Developing a monitoring program for

invertebrates: Guidelines and a case study. Conservation Biology 21:422–433.

Rosenzweig, M. L. 1995. Species Diversity in Space and Time. Cambridge University

Press.

Rothenbücher, J., and M. Schaefer. 2006. Submersion tolerance in floodplain arthropod

communities. Basic and Applied Ecology 7:398–408.

Salo, J., R. Kalliola, I. Hakkinen, Y. Makinen, P. Niemela, M. Puhakka, and P. D. Coley.

1986. River dynamics and the diversity of Amazon lowland forest. Nature 322:254–

258.

Samways, M. J. 1993. Insects in biodiversity conservation - Some perspectives and

directives. Biodiversity and Conservation 2:258–282.

Simpson, E. H. 1949. Measurement of diversity. Nature 163:688.

Smith, B., and J. B. Wilson. 1996. A consumer’s guide to evenness indices. Oikos 76:70–

82.

Sokal, R. R. 1986. Spatial data analysis and historical processes. Pages 29–43 in E. Diday

and et al., editors. Data Analysis and Informatics, IV. North-Holland, Amsterdam.

Sombroek, W. 2000. Amazon landforms and soils in relation to biological diversity. Acta

Amazonia 30:81-100..

Stork, N. E. 1988. Insect diversity: Facts, fiction and speculation. Biological Journal of

the Linnean Society 35:321–337.

Stork, N. E., M. Stone, L. Sam. 2016. Vertical stratification of beetles in tropical

rainforests as sampled by light traps in North Queensland, Australia. Austral

242

Ecology 41:168–178.

Stork, N. E., P. S. Grimbacher, R. Storey, R. G. Oberprieler, C. Reid, and S. A. Slipinski.

2008. What determines whether a species of insect is described? Evidence from a

study of tropical forest beetles. Insect Conservation and Diversity 1:114–119. ter Steege, H., D. Sabatier, H. Castellanos, T. Van Andel, J. Duivenvoorden, A. A. de

Oliveira, R. Ek, R. Lilwah, P. Maas, and S. Mori. 2000. An analysis of the floristic

composition and diversity of Amazonian forests including those of the Guiana

Shield. Journal of Tropical Ecology 16:801–828.

Terborgh, J., and E. Andresen. 1998. The composition of Amazonian forests: Patterns at

local and regional scales. Journal of Tropical Ecology 14:645–664.

Thiele, H. U. 1977. Carabid Beetles in Their Environments. A Study on Habitat Selection

by Adaptation in Physiology and Behaviour. Springer-Verlag, Berlin.

Tuomisto, H., A. D. Poulsen, K. Ruokolainen, R. C. Moran, C. Quintana, J. Celi, and G.

Canas. 2003. Linking floristic patterns with soil heterogeneity and satellite imagery

in Ecuadorian Amazonia. Ecological Applications 13:352–371.

Venables, W. N., and B. D. Ripley. 2002. Modern Applied Statistics with S. 4th edition.

Springer, New York.

Walther, B. A., and J. L. Moore. 2005. The concept of bias, precision and accuracy, and

their use in testing the performance of species richness estimators, with a literature

review of estimator performance. Ecography 28:815–829.

Wilson, E. O. 1987. The arboreal ant fauna of Peruvian Amazon forests: A first

assessment. Biotropica 19:245–251.

243

Wittmann, F., W. J. Junk, and M. T. F. Piedade. 2004. The várzea forests in Amazonia:

Flooding and the highly dynamic geomorphology interact with natural forest

succession. Forest Ecology and Management 196:199–212.

Wittmann, F., J. Schöngart, and W. J. Junk. 2010. Phytogeography, species diversity,

community structure and dynamics of central Amazonian floodplain forests. Pages

61–102 in W. J. Junk, M. T. F. Piedade, P. Parolin, F. Wittmann, and J. Schöngart,

editors. Amazonian Floodplain Forests: Ecophysiology, Biodiversity and

Sustainable Management. Springer Science+Business Media B.V., Dordrecht.

Wittmann, F., J. Schöngart, J. C. Montero, T. Motzer, W. J. Junk, M. T. F. Piedade, H. L.

Queiroz, and M. Worbes. 2006. Tree species composition and diversity gradients in

white-water forests across the Amazon Basin. Journal of Biogeography 33:1334–

1347.

Worbes, M. 1997. The forest ecosystem of the floodplains. Pages 223–265 in W. J. Junk,

editor. The Central Amazon Floodplain, Ecology of a Pulsing System. Springer-

Verlag, Berlin Heidelberg.

Zerm, M., and J. Adis. 2000. On the life cycles of Phaeoxandtha species (Coleoptera:

Cicindelinae) from Central Amazon Floodplains (Brazil). Ecotropica 6:141–155.

Zerm, M., and J. Adis. 2001a. Spatio-temporal distribution of larval and adult tiger

beetles (Coleoptera: Cicindelidae) from open areas in Central Amazonian

floodplains (Brazil). Studies on Neotropical Fauna and Environment 36:185–198.

Zerm, M., and J. Adis. 2001b. Further observation on natural history and survival

strategies of riverine tiger beetles (Coleoptera: Cicindelidae) from open habitats in

244

central Amazonian floodplains (Brazil). Ecotropica 7:115–137.

Zerm, M., and J. Adis. 2003. Survival strategy of bombardier beetle, Pheropsophus

rivieri, (Col.: Carabidae) in central Amazonian blackwater floodplain (Brazil).

Amazoniana XVII:503–508.

Zerm, M., J. Adis, W. Paarmann, M. A. Amorim, and C. R. V. da Fonseca. 2001. On

habitat specificity, life cycles and guild structure in tiger beetles of Central

Amazonia (Brazil) (Coleoptera: Cicindelinae). Entomologia Generalis 25:141–154.

Zhang, Z. 2011. Phylum Arthropoda von Siebold, 1848. Z. Q. Zhang, editor. Animal

Biodiversity: An Outline of Higher-level Classification and Survey of Taxonomic

Richness. Zootaxa 3148: 99–103.

245

TABLES

IV Table – 1. Measures of abundance, species richness and diversity of Carabidae collected for floodplain (FP) and terra firme (TF) forests.

Measure FP TF Total No. individuals 564 691 1255 Avg. individuals (± s.e.) 47.0 ± 5.0 57.6 ± 7.1 52.3 ± 4.4 No. morphospecies 96 79 143 Avg. morphospecies (± s.e.) 21.6 ± 2.0 16.9 ± 1.6 19.3 ± 1.3 No. tribes 20 14 20 No. genera 44 31 54 No. singletons 34 38 63 No. doubletons 15 8 22

Richness estimators

Chao1 (± s.e.) 126.4 ± 13.7 159.1 ± 38.3 237.4 ± 32.9

Jack1 (± s.e.) 130.8 ± 12.9 122.1 ± 9.8 210.1 ± 17.8 ACE 131.4 147.0 248.6

246

IV Table – 2. List of the characteristic Carabidae species which had significant indicator values for floodplain (FP) and terra firme (TF) forests. Indicator values were calculated using the IndVal statistic developed by Dufrêne & Legendre (1997). ‘A’ measures the specificity of a given species using relative abundance within and among study sites for a forest type and ‘B’ measures fidelity though relative frequency within the study sites of each forest type. Adjusted p-values (p.adj holm) for multiple testing correspond to experiment-level conclusions.

IndVal p.adj Morphospecies Tribe:Genus A B p index (holm) FP Clivb Scaritini: Clivina 0.969 0.917 0.942 0.001 0.025* Perua Lachnophorini: Peruphorticus 0.94 0.917 0.93 0.001 0.025* Nyctb Scaritini: Nyctosyles 1 0.583 0.764 0.005 0.110 Nycta Scaritini: Nyctosyles 0.917 0.583 0.731 0.021 0.378 Eucaa Lachnophorini: Eucaerus 1 0.500 0.707 0.013 0.247 Parac Bembidiini: Paratachys 1 0.500 0.707 0.012 0.24

Pentgd Pentagonicini: Pentagonica 0.895 0.833 0.863 0.001 0.025* Lebih Lebiini: Lebia 1 0.500 0.707 0.01 0.21 Perua_2 Lachnophorini: Peruphorticus 0.971 0.500 0.697 0.031 0.527 Pentga Pentagonicini: Pentagonica 0.917 0.500 0.677 0.039 0.72 Parae Bembidiini: Paratachys 1 0.417 0.645 0.045 0.72

247

FIGURES

IV Figure – 1. Map of the study area. A: South America with the Amazon Basin outlined,

B: Ecuador with the locations of Tiputini Biodiversity Station (TBS) and Yasuní National

Park and C: Shows a digitial evelation map of TBS and the 24 sampling sites are shown with floodplain (FP) forest sites represented as blue squares and terra firme (TF) forest as green circles.

* Maps created using R packages: ‘raster’ (Hijmans 2015), ‘sp’ (Pebesma & Bivand 2005; Bivand et al. 2013), ‘GISTools’ (Brunsdon & Chen 2014), and ‘maps’ (Becker et al. 2015).

248

IV Figure – 2. Rarefaction interpolation (solid lines) and extrapolation curves (dashed lines) with point markers representing the sampling extent of the study and shaded areas depicting unconditional 95% confidence intervals. A: The dataset with FP and TF forests types combined with richness extrapolated to n = 2,510 (twice the number of individuals collected). B: Rarefaction curves for floodplain (FP) forests (blue square) and terra firme

(TF) forests (green circle) with sample size extrapolated to n=1,128 (twice the number of individuals collected in FP forests). FP forests were significantly more rich than TF forests at the rarefied sample size (n = 564).

249

250

IV Figure – 3. Inverse Simpson’s concentration (1/D) for floodplain (FP) forest and terra firme (TF) forest. FP forest had significantly higher 1/D values than TF (p = 0.003).

251

IV Figure – 4. Mean beetle body length (mm) of carabid individuals collected in floodplain forest (FP) and terra firme forest (TF), excluding Cicindelini. Body length for

FP forest was significantly larger than TF forest (p < 0.001).

252

IV Figure – 5. Number of Cincidelini (tiger beetles) collected from terra firme (TF) and floodplain (FP) forests. There were significantly more Cicindelini individuals collected in

TF forest (p = 0.011).

253

IV Figure – 6. Simpson’s evenness index (E1/D) for species assemblages were significantly more even from floodplain (FP) than terra firme (TF) forests (p = 0.008).

254

IV Figure – 7. Rank abundance distribution curves for temporarily flooded (FP) and non- flooded terra firme forest (TF).

255

IV Figure – 8. Non-metric multidimensional scaling (NMDS) ordination using Bray-

Curtis dissimilarity for Carabidae species assemblages from floodplain (FP) and terra firme (TF) forests (stress = 13.7, k = 2). Each data point represents one of 24 sampling sites, with square point markers representing FP forest sites and circles representing TF forest sites. Species assemblages are significantly different between FP and TF (p <

0.001).

256

APPENDIX A

IV Figure – A1. Water height for the Tiputini River at Tiputini Biodiversity Station,

Ecuador. (A) Monthly values represent overall mean river height and the mean maxima and minima based on data from 2009 – 2014, (B) three years before and after the sampling period for this study. Monthly river height values during the sampling period,

2011(B) and 2012, (C) for mean river height in addition to water height maxima and minima.

257

258

APPENDIX B

IV Table – B1. List of the 143 Carabidae morphospecies collected, including the tribe, genera and identification code for each morphospeices. Numbers of individuals are provided for floodplain (FP) and terra firme (TF) forests, as well as combined totals.

Carabidae taxa FP TF Total Bembidiini 65 42 107 Erwinana 1 1 2 Erwia 1 1 Erwib 1 1 Geballusa 2 6 8 Gebasp 2 6 8 Meotachys 2 2 Meotsp 2 2 Micratopus 4 4 Micrsp 4 4 Mioptachys 1 1 Miopsp 1 1 New Genus 1 3 1 4 NG1a 3 3 NG1b 1 1 New Genus 2 2 2 NG2sp 2 2 Paratachys 50 33 83 Paraa 33 11 44 Parab 2 13 15 Parac 14 14 Parad 1 1 Parae 9 9 Polyderis 1 1 Polysp 1 1 Callistini 6 6 Dercylus 6 6 Derca 4 4 Dercb 2 2 Calophaenini 2 4 6

259

Calophaena 2 4 6 Caloa 2 2 Calob 3 3 Caloc 1 1 Cicindelini 109 329 438 Odontocheila 30 94 124 Odona 14 46 60 Odonb 16 48 64 Pentacomia 79 235 314 Penta 13 9 22 Pentb 66 226 292 Collyridini 1 1 Ctenostoma 1 1 Ctensp 1 1 Galeritini 6 6 Galerita 6 6 Galea 2 2 Galeb 2 2 Galec 2 2 Harpalini 36 24 60 Notiobia 23 14 37 Notia 7 3 10 Notia_2 2 1 3 Notia_3 2 7 9 Notib 8 3 11 Notib_2 1 1 Notib_3 1 1 Notib_4 1 1 Notic 1 1 Selenophorus 13 10 23 Selee 9 2 11 Selea 3 5 8 Seleb 2 2 Selef 1 1 Selee_6 1 1 Helluonini 6 3 9 Helluobrochus 2 2 Hellub_a 1 1 Hellub_b 1 1 Helluomorpha 1 1

260

Hellum 1 1 Helluomorphoides 5 1 6 Hella 3 3 Hellb 1 1 Hellc 2 2 Hiletini 5 5 Ecuamaragnathus 5 5 Ecuabat 5 5 Lachnophorini 129 41 170 Amphithasus 3 3 Ampha 2 2 Amphb 1 1 Eucaerus 10 10 Eucaa 6 6 Eucaa_2 3 3 Eucab 1 1 Peruphorticus 106 41 147 Perua 84 5 89 Perua_2 1 33 34 Perua_3 13 3 16 Perub 7 7 Perub_2 1 1 Pseudophorticus 10 10 Pseua 6 6 Pseub 4 4 Lebiini 30 90 120 Apenes 5 15 20 Apena 1 1 Apena_2 1 1 2 Apena_3 2 2 Apenb 1 2 3 Apenb_2 1 1 Apenb_3a 1 1 Apenb_3b 1 1 Apenb_3c 2 2 Apenb_4 1 1 Apenb_5 1 1 Apenbb 1 1 Apenc 1 1 Apenc_2 1 1

261

Apenc_3 1 1 Apene 1 1 Coptodera 2 2 Coptb 1 1 Copta 1 1 Eucheila 1 1 Euchsp 1 1 Hyboptera 2 2 Hybosp 2 2 Lebia 21 69 90 Lebia 3 5 8 Lebib 1 1 Lebic 1 1 Lebid 2 32 34 Lebid_5 1 1 Lebie 7 8 15 Lebif 1 1 Lebig 2 2 Lebih 10 10 Lebij 1 1 Lebik 4 1 5 Lebil 2 8 10 Lebio 1 1 Negrea 1 1 Negrsp 1 1 Nemotarsus 1 1 Nemosp 1 1 Stenognathus 3 3 Stensp 3 3 Loxandrini 45 47 92 Adrimus 1 1 Adrisp 1 1 Loxandrus 41 47 88 Loxaa 1 1 Loxab 20 19 39 Loxac 3 3 Loxae 7 6 13 Loxae_2 2 2 Loxae_3 10 9 19 Loxae_4 7 7

262

Loxae_6 2 2 Loxaf 1 1 Loxag 1 1 Stolonis 3 3 Stolsp 3 3 Odacanthini 8 2 10 Colliuris 8 2 10 Colla 3 3 Collb 1 1 Collc 4 4 Collc_2 1 1 Colld 1 1 Oodini 6 6 Macroprotus 5 5 Macrsp 5 5 Oodinus 1 1 Oodisp 1 1 Pentagonicini 17 78 95 Pentagonica 17 78 95 Pentga 1 11 12 Pentgb 12 33 45 Pentgd 4 34 38 Perigonini 1 1 Perigona 1 1 Perisp 1 1 Platynini 6 3 9 Glyptolenus 6 3 9 Glypsp 6 3 9 Pterostichini 1 23 24 Abaris 17 17 Abara 12 12 Abarb 5 5 Haplobothynus 1 2 3 Hapla 2 2 Haplb 1 1 Pseudabarys 1 1 Pseuabsp 1 1 Trichonillia 3 3 Trica 2 2 Tricb 1 1

263

Scaritini 84 3 87 Ardistomis 1 1 Ardisp 1 1 Camptidius 1 1 Camptisp 1 1 Camptodontus 1 1 Camptosp 1 1 Clivina 42 1 43 Cliva 2 2 Clivb 31 1 32 Clivc 1 1 Clivc_2 3 3 Clivd 1 1 Clivd_2 4 4 Nyctosyles 32 1 33 Nycta 11 1 12 Nyctb 21 21 Oxydrepanus 5 5 Oxyda 1 1 Oxydb 3 3 Oxydc 1 1 Stratiotes 3 3 Strasp 3 3 Zuphiini 1 2 3 Pseudaptinus 1 2 3 Pseuapa 1 1 Pseuapb 2 2 Column totals 564 691 1255

264

APPENDIX C

IV Table – C1. Carabidae tribes with higher numbers of common species than expected.

The number of species collected in this study are listed under ‘observed’, whereas the expected number of species was calculated (see methods). Analysis conducted on the combined values for both floodplain and terra firme forests. The number of abundant species among carabid tribes was significantly different than the expected number of species (chi-square χ2 = 33.6, p < 0.001). Differences in the distribution of ‘rare’ species among tribes were not significantly different.

Abundant No. of S Direction of difference Tribe Observed Expected Difference More than expected Bembidiini 3 1.99 1.01 Cicindelini 4 0.53 3.47 Lachnophorini 3 1.59 1.41 Loxandrini 3 1.59 1.41 Pentagonicini 2 0.40 1.60

Approximately equal Scaritini 2 1.99 0.01

Less than expected Lebiini 2 4.65 -2.65

265

IV Figure – C1. Number of individuals for the morphospecies Pentb (genus Pentacomia), for floodplain (FP) and terra firme (TF) forests (p = 0.02).

266

IV Figure – C2. Number of rare species for the floodplain (FP) and terra firme (TF) forest

(p = 0.04).

267

IV Figure – C3. Species relative abundance by the number of sampling sites at which they were present. Point markers represent the 143 morphospecies coded according the three rarity categories: ‘dominant’ (circle), ‘common’ (triangles) and ‘rare’ (diamonds).

Species classified as ‘dominant’ occurred at a higher number of sampling sites than

‘common’ species and ‘common’ species occurred at a higher number of sampling sites than ‘rare’ species (p < 0.001).

268

IV Figure – C4. Rank abundance distribution curves for terra firme rainforest (TF) in

Ecuador, and mature (approximate 90 year old), mixed hardwood deciduous temperate forest (Ninety) in North Carolina, USA.

269

APPENDIX D

IV Figure – D1. Numbers of individuals for Carabidae tribes with significant differences

(p < 0.05) between floodplain (FP) and terra firme (TF) forests.

270

IV Figure – D2. Numbers of species for Carabidae tribes with significant differences (p <

0.05) between floodplain (FP) and terra firme (TF) forests.

271

APPENDIX E

IV Figure – E1. Non-metric multidimensional scaling (NMDS) ordination based on

Jaccard dissimilarity values (presence/absence data) of Carabidae species assemblages for floodplain (FP) and terra firme (TF) forests (stress = 13.7, k = 2). Each data point represents one of the 24 sampling sites. Significant differences (p < 0.001) in species assemblages occurred between FP and TF.

272

CHAPTER V

RESULTS FROM TWO SAMPLING TECHNIQUES FOR CARABID BEETLES

(COLEOPTERA: CARABIDAE) IN TEMPORARILY FLOODED AND TERRA

FIRME RAINFOREST OF WESTERN AMAZONIA

Kathryn N. Rileya*, Robert A. Brownea, Terry L. Erwinb

a Department of Biology, Wake Forest University, 226 Winston Hall, Box 7325 Reynolda

Station Winston-Salem, NC 27109, NC USA. phone: 336-758-4731, cell phone: 443-

744-3527. bDepartment of Entomology, National Museum of Natural History, Smithsonian

Institution, Washington, DC, USA.

The following manuscript has been published in Studies on Neotropical Fauna and the Environment (2016, 51:78–95). Stylistic variations are due to the requirements of the journal.

273

Abstract

Carabidae beetles (Coleoptera) were sampled by flight intercept traps (FITs) and hand sampling in the understory of a lowland rainforest in eastern Ecuador. Sampling occurred in the June and July of 2011 and 2012 at 24 sites within temporarily flooded forest (FP) and non-flooded terra firme (TF) forest. A total of 674 Carabidae (excluding Cicindelinae) individuals, representing 18 tribes, 50 genera and 133 morphospecies, were collected.

Significantly more individuals were collected by hand compared to FITs, particularly by hand in FP forests. After correction through sample-based rarefaction there are no significant differences in species richness between sampling methods and forest types.

Ordination and dissimilarity analysis show FITs and hand sampling target different portions of the Carabidae community. Understanding sampling biases for various sampling methods and how it may be affected by forest type will result in more accurate and efficient surveys in tropical forests.

Keywords: flight intercept trap (FIT); Neotropics; Ecuador; hand sampling; trapping methods; Tiputini Biodiversity Station (TBS); inundation forests; Insecta; species assemblages; diversity.

274

Introduction

Insects and other arthropods dominate biodiversity in terms of individuals, biomass and species richness and play a vital role in ecological processes and ecosystem function

(Erwin 1982; May 1988; Kremen et al. 1993; Samways 1993; Kim 1993). While the tropics harbor the greatest insect diversity, the majority of taxa have not been collected and/or described (Erwin 1982; Stork 1988; Godfray et al. 1999; Primack & Corlett 2005;

Fonseca 2009). With current rates of tropical deforestation, degradation and conversion, many tropical forest insect species will remain unknown indefinitely (Samways 1993;

Kim 1993; Dunn 2005). A greater comprehension of tropical insect diversity and communities is needed to discern ecological patterns, detect and mitigate disturbance and develop conservation strategies (Erwin 1991a; Carlton et al. 2004; Rohr et al. 2007;

Erwin & Geraci 2009).

Ecological entomology studies in Amazonian forests are challenging for several reasons, including the sheer diversity of the fauna, incompleteness of taxonomic descriptions, adequate spatio-temporal replication and logistical challenges associated with sampling complex ecosystems (Lawton et al. 1998; Godfray et al. 1999; Wagner

2000; Kitching et al. 2001; Brehm et al. 2008). Ideally, tropical entomological surveys should utilize a range of sampling techniques with the scale varied in space and time

(Gadagkar et al. 1990; Longino & Colwell 1997; Kitching et al. 2001; Ozanne 2005;

Missa et al. 2009). However, field work is not always ‘ideal’; time and resources typically limit the number of methods researchers employ (Kitching et al. 2001). A variety of collection methods have been employed in tropical forests, including insecticidal fogging, flight interception, beating and sweeping vegetation, hand samples

275

and leaf litter sifting (Erwin 1982; Southwood & Henderson 2000; Leather 2005;

Lamarre et al. 2012); however, there are few studies that provide quantitative differences in insect abundance, richness, diversity and species composition among sampling methods in the Amazonian forest understory (Missa et al. 2009; Lamarre et al. 2012).

It is well known that sampling methods target different portions of insect communities (Morse et al. 1988; Southwood & Henderson 2000; Kitching et al. 2001;

Hyvärinen et al. 2006; Missa et al. 2009). Trapping method efficiency can be affected by faunistic differences (taxonomic and behavior) between two areas and/or habitats and the trappability of taxa can change with habitat structure (Gadagkar et al. 1990; Melbourne

1999; Missa et al. 2009). Tropical researchers have to make decisions integral to the success of a study before departure to the field site. Beforehand knowledge of performance, effectiveness and short-comings of various sampling methods would be immensely beneficial, especially for ‘hyperdiverse’ groups of organisms, such as insects

(Colwell & Coddington 1994; Hyvärinen et al. 2006; Bouget et al. 2008). Planning for, employing and maintaining plots at remote field sites is difficult and requires time and money (Gadagkar et al. 1990; Coddington et al. 1991) with many common sampling methods in temperate regions being less successful or failing completely for tropical fauna (Erwin 1979; Paarmann et al. 2001).

Flight intercept traps (FITs) have a high success rate for collecting flying arthropods within forested habitats (Peck & Davies 1980; Hill & Cermak 1997; Stork &

Grimbacher 2006, 2009; Missa et al. 2009; Lamarre et al. 2012). FITs are reported to be more effective at collecting insects with relatively greater mass, that tend to drop after hitting a surface (like many Coleoptera), and those which fly in a direct path (unlike the

276

‘fluttering’ of butterflies) (Peck & Davies 1980; Southwood & Henderson 2000; Bouget et al. 2008; Missa et al. 2009; Lamarre et al. 2012). FITs are a commonly used passive sampling method which provides representative baseline data, repeatable results and consistent function among study sites (Chapman & Kinghorn 1955; Bouget et al. 2008;

Missa et al. 2009). In addition, passive intercept traps also have the ability to collect individuals with relatively low labor intensity; however, they can add complications including bias towards more actively flying species, distinction of tourist species, installation difficulties, material necessities, trap disturbances, and lack of information regarding micro habitat preferences (Southwood & Henderson 2000; Hyvärinen et al.

2006; Bouget et al. 2008; Missa et al. 2009). FITs have successfully collected Coleoptera in previous tropical studies (Chung et al. 2000; Stork & Grimbacher 2006, 2009; Missa et al. 2009; Erwin et al. 2012; Lamarre et al. 2012), but their effectiveness for collecting carabid beetles is not well known in the rainforest understory of Amazonia so an objective of this study was to evaluate this. Hand sampling has been employed with success in biodiversity assessments of tropical arthropods including spiders (Coddington et al. 1991), ants (Majer & Delabie 1994; Brühl et al. 1998; Mertl et al. 2009; Ryder

Wilkie et al. 2010) and beetles (Carlton et al. 2004), including Carabidae (Erwin 1991b;

Maveety et al. 2011, 2014; Maveety & Browne 2014). It has the benefit of collecting an individual directly from the habitat in which it occurs. This method is more labor intensive; it can lead to low numbers of individuals per species and collectors can individually vary in their effectiveness depending on expertise and experience (Bouget et al. 2008). These two methods were chosen because they were predicted to be efficient, complementary and have minimal overlap in species composition (Coddington et al. 1991;

277

Carlton et al. 2004). FITs target individuals flying through the forest understory (Peck &

Davies 1980; Hyvärinen et al. 2006; Stork & Grimbacher 2006) while hand sampling focuses on the flight restricted and/or less mobile species occurring on vegetation, trees and the forest floor.

Coleoptera, and other arthropods, are extremely diverse and of high interest for biodiversity related studies. Information regarding efficiency, species richness and defining community characterization for different trapping methods is needed, particularly for lowland Neotropical forests (Colwell & Coddington 1994; Longino &

Colwell 1997; Hyvärinen et al. 2006; Rohr et al. 2007). Carabidae are of interest as subjects for biodiversity and bioindicator studies (McGeoch 1998; Niemelä et al. 2000;

Koivula 2011). There are approximately 40,000 described species of carabid beetles (also known as ground beetles), making it one of the most diverse beetle families (Lövei &

Sunderland 1996; Niemelä 1996). Previous studies (Lucky et al. 2002; Erwin et al. 2005) have documented the immense diversity of Carabidae within the Ecuadorian Amazon.

The objectives of this study were to (1) investigate differences among FITs and hand sampling in terms of numbers of individuals, richness and diversity within Carabidae

(excluding Cicindelinae); (2) compare species composition among sampling techniques; and (3) evaluate results of both methods among temporally flooded and non-flooded forest in the understory of lowland tropical rainforest in eastern Ecuador.

278

Materials and methods

Study site

The study area is located in western Amazonia in Ecuador, within the province of

Orellana, at Tiputini Biodiversity Station (TBS) (0°37’55” S, 76°08’39” W; 190-270 m elevation) (Figure 1). The station, consisting of 650 ha of moist tropical rainforest, borders the megadiverse Yasuní National Park (Bass et al. 2010). The forest surrounding

TBS has experienced relatively minimal anthropogenic disturbance. The area has warm average annual temperatures (24-27°C) and high yearly rainfall (approximately 3200 mm). There is subtle seasonal variation, with the wet season extending from April to

August and the dry season from October to February (Pitman 2000). Approximately 90% of the landscape around TBS is terra firme forest, which does not flood. Standing water in terra firme forests may last for only a few hours after heavy rains (Burnham 2002).

Floodplain forests in this study occur along the Tiputini River, a mixed water river with the largest contribution from white water (Pitman 2000) and are inundated by the adjacent Tiputini for various time periods, typically after heavy rains in the Andes

Mountains (Prance 1979; Mertl et al. 2009). Maps created using R packages: ‘raster’

(Hijmans 2015), ‘sp’ (Pebesma & Bivand 2005; Bivand et al. 2013), ‘GISTools’

(Brunsdon & Chen 2014), and ‘maps’ (Becker et al. 2015).

Collection of Carabidae

Carabid beetles were collected over two field seasons in 2011 and 2012 with sampling occurring in June and July of each year. Two sampling methods, FITs and hand, were employed at 24 sites in the forest understory, with half of the sites located in temporarily

279

flooded forest (FP) and half in non-flooded terra firme (TF) forest (Figure 1). Trapping days for FITs ranged from 65 to 76 days, varying slightly due to flooding events, logistical limitations and trap disturbance. However, a repeated measures t-test showed the difference between the two forest types for numbers of carabids collected per trapping day to be insignificant for 65 collection days (t63=1.99, p=0.91).

Each FIT consisted of PVC tubing, screen netting, a plastic awning (to prevent rain and debris from entering the trays) and plastic collection trays (Figure 2). Screening was stretched between two 1 m PVC tubing, with 1 m support tubing across the top of the trap to create an approximately 1 m2 flight barrier. Due to limited material availability, the color of the screen netting was not identical for all traps. Screens were black or green, with the exception of one white screen, and there were equal numbers of FITs with green and black screening in floodplain and terra firme forests. Raw numbers of individuals and species composition were not significantly different among screen colors. Plastic trays were placed side by side underneath the flight barrier to collect specimens and filled with water mixed with biodegradable soap. There was the possibility individuals may have fallen into small gaps between plastic trays but this error was likely small and any effect equally applied across all FITs. Sampling trays were emptied every other day by filtering the contents through a metal strainer with mesh openings less than <1 mm.

Since many Amazonian Carabidae are active at night (T. Erwin, pers. obs.), hand sampling occurred between 20:30 and 22:30 h. This method included searching the ground (leaf litter, rocks, etc.), tree trunks and vegetation (up to eye level) and decaying logs. Seven hand sampling events were performed at each of the 24 study sites, four in

2011 and three in 2012. Each sampling event consisted of 0.5 hour of search effort and

280

occurred within 100 meters of the FIT for each study site. The senior author participated in every collection event with 1-2 other amateur collectors assisting. Immediately after capture, carabids were preserved in 95% ethanol. In addition to FITs and hand samples, during the 2011 field season three pitfall traps were employed at each study site.

However, due to the low number of individuals collected in 2011 (n=22) pitfall traps were not continued in 2012 and were not used in subsequent analyses.

Carabid beetles were sorted using a stereo-microscope, identified to genera and subsequently to morphospecies (hereafter referred to as species). Since species-level taxonomic keys are not available for the region, the generic key for French Guyana

Carabidae (Erwin et al. 2012) was used followed by identification to morphospecies based on external morphology with confirmation by T. Erwin, Curator of Coleoptera,

Smithsonian Institution, National Museum of National History. The use of morphospecies for Neotropical arthropods has been successful in ecological studies and is the most feasible option for an extraordinarily speciose region, with the majority of species not described (Oliver & Beattie 1996; Stork et al. 2008; Grimbacher & Stork

2009).

Statistical analysis

Number of individuals, richness and diversity

Raw numbers of individuals and species collected by FITs and hand sampling were compared using a Wilcoxon signed-ranked test. Samples of mobile organisms are affected by abundance and species richness, in addition to activity patterns, therefore sample size was standardized by rarefaction for quantitative comparison (Gotelli &

281

Colwell 2001; Gotelli & Colwell 2011), using R 3.0.1, package vegan (Oksanen et al.

2013; R Core Team 2013). Rarefaction provides the expected number of species/incidences from a random subsample of the overall data set and allows trapping methods that are used simultaneously in the same area to be compared at a standardized sampling effort (Longino et al. 2002; Ellison et al. 2007; Rohr et al. 2007; Gotelli &

Colwell 2011). Since all sampling techniques have inherent and typically unknown sampling bias, species detection may vary, creating the potential for some species to be over- or under-represented, so incidence-based data was used in rarefaction analysis

(Gotelli & Colwell 2001, 2011; Longino et al. 2002; King & Porter 2005; Ellison et al.

2007). Sample-based rarefaction curves were constructed in EstimateS v. 9.1.0. using

1000 randomizations (resamples) and depicted with 95% unconditional confidence intervals (Colwell et al. 2012; Colwell 2013) and the x-axis for the rarefaction curves were rescaled to species occurrences (incidence data) (Gotelli & Colwell 2001, 2011;

Longino et al. 2002; King & Porter 2005; Ellison et al. 2007). Fisher’s alpha was computed using abundance with EstimateS (Colwell 2013) with individuals from samples added to the sampling pool at random and subjected to 1000 randomizations without replacement to calculate a mean and a bootstrapped standard deviation (Colwell 2013).

Three incidence-based non-parametric diversity estimators, Chao2 (Chao 1984),

Jackknife2 (Jack2) (Burnham & Overton 1979) and Incidence Coverage Estimator (ICE)

(Lee & Chao 1994), were calculated to estimate potential true richness (Colwell &

Coddington 1994; Longino et al. 2002; Brose et al. 2003; Chao 2005). EstimateS was used to calculate log-linear 95% confidence intervals for Chao2 and Jack2. Sample-based estimators were chosen based on their performance in similar studies and because they

282

represent the upper (Jack2), intermediate and lower (Chao2) limits for nonparametric richness estimators (Chao 1984, Colwell & Coddington 1994; Chiarucci et al. 2003;

Brose et al. 2003; King & Porter 2005). Uniques, species represented in a single sample only, and duplicates, species known from two samples, were calculated. They are often used as indicators of sampling completeness and are two main components in calculations of nonparametric estimators (Longino et al. 2002; Brehm et al. 2008).

Species assemblages

Species assemblages collected by the two sampling methods in FP and TF forest were analyzed through non-metric multidimensional scaling (NMDS) ordination. Function metaMDS in R, vegan package, was used to perform ordination analysis with vegdist to compute dissimilarity using the Bray-Curtis measure (Oksanen et al. 2013; R Core Team

2013). Since there were large differences among species abundances, abundance data was standardized through relative abundance and 100 random starts were utilized in the

NMDS (Legendre & Gallagher 2001; Kotze et al. 2011; Oksanen 2013). To determine if carabid species assemblages differed among sampling techniques and forest types, mean

Bray-Curtis dissimilarity measures were calculated. Dissimilarity values were scaled from 0 (identical) to 1 (no overlap). All analyses of dissimilarity matrices were carried out in R (R Core Team 2013) using the vegan package (Oksanen et al. 2013). Prior to dissimilarity analysis, data were also standardized through relative abundance to reduce large differences in abundances in order to improve detection of dissimilarities (Oksanen

2013). Dissimilarity data was grouped by collection technique (FIT or hand) and then by forest type (FP or TF). The function meandist was used to calculate mean dissimilarity

283

values among study groups. Function adonis was used to compare dissimilarity measures among technique and forest type groups (Anderson 2001; Oksanen et al. 2013).

Homogeneity of group variances was tested using betadisper post-hoc test, TukeyHSD

(Anderson 2006; Anderson et al. 2006; Oksanen 2013).

Results

The majority of individuals collected were understory species, which are commonly found in the leaf litter, on foliage, on tree trunks and fallen trees and associated bracket fungi, and in areas of highly concentrated organic matter (i.e. rotting foliage, fruit and blossom falls). During hand sampling, the majority of carabids were collected from the ground (> 97%) but were also collected from vegetation leaves, tree trunks and decaying woody or other organic matter.

Number of individuals

With both collection techniques combined, 674 Carabidae individuals were collected, representing 18 tribes, 50 genera and 133 species. Significantly more individuals were collected by hand (median=20) than by FITs (median = 9; W23 = 34.5, p < 0.003; Table 1).

Significantly more individuals were collected by hand in FP forests (median = 27) than in

TF forests (median = 6; W12 = 5, p < 0.008) but there was no difference in FITs collections in number of individuals between the two forest types (medians: FP = 7, TF =

10; W12 = 25, p = 0.803; Figure 3). Each of the two sampling methods selected,

284

exclusively or by higher numbers, for particular Carabidae higher taxa and species (see

Appendix 1).

Richness and diversity

Raw species number was significantly greater for hand than for FITs collections (W12 =

53.5, p < 0.04; Table 1), with 50% (S = 67) of the total number of species collected by hand only, 39% (S = 52) by FIT only and 10% (S = 14) collected by both techniques.

Raw species number was significantly greater for hand sampling in FP forests than for TF

(W12 = 3, p < 0.05), but after the number of incidences were corrected through rarefaction there were no significant differences in richness between FITs and hand sampling in either forest type (Figures 4, 5). There is a high degree of overlap in the 95% confidence intervals at the rarified incidence number for both FITs and hand rarefaction curves and each technique in the two forest types. The relatively steep slopes of all sample-based rarefaction accumulation curves, particularly for FITs, indicate that only a portion of the overall Carabidae diversity was sampled in this study (Figures 4, 5). The high proportion of uniques (39-78% of incidence total) suggests that many species have not been sampled by either technique (Table 1; Figure 6). Even when data for both techniques are combined the slope of the sample-based rarefaction curve becomes less steep but still is not asymptotic (Figure 4).

All diversity estimators predict more species will be added with greater sampling effort for both techniques (Table 1; Figure 6). For the entire dataset, all three estimators predict that 55.4-59.1% of the potential true species number was collected in the study period. FITs sampled 50.1-55.5% and hand samples collected approximately 54.7-60.5%

285

of the potential richness of the area. The Chao2 estimators stabilized for FIT and hand sampling and ICE for hand sampling. For the entire dataset all diversity estimators predict similar richness, within a 7% range. Fisher’s alpha was not significantly different between FIT and hand collections, nor for either collection method in FP or TF forests

(Table 1).

Species assemblages

In the NMDS ordination analysis, most of the hand sampling study sites were more negatively correlated with axis 1 compared to the FIT sites (Figure 7; stress = 18.5, k = 2).

The separation of polygons for collection technique in each forest type along axis 1 suggests that FITs and hand sampling target different portions of the Carabidae fauna

(see Appendix 2). The ordination figure showed no overlap between the FIT and hand sampling polygons for FP forest sites. Polygons representing FP forest for FITs was entirely nested within the respective TF polygon, indicating that collection technique more strongly affected species composition than forest type. FITs in the TF had the highest compositional variation while hand sampling in the FP had the least (Figure 7).

The Bray-Curtis dissimilarity values corroborate the high compositional differences both within and between FITs and hand sampling, as well as when each is analyzed by forest type (Table 2). Average species composition within sampling techniques were slightly more similar than between collection techniques (within: 0.90, between: 0.96). Composition analyzed by technique within FP and TF forests followed the same pattern (Table 2). The lowest within group similarity was hand sampling the floodplains (0.66). The groups with the greatest compositional dissimilarity were the TF

286

hand and TF FIT samples (0.978) and also between the FP hand and TF FIT samples

(0.987). No dissimilarity values between groups were below 0.89 signifying a high level of variation. Dissimilarity matrices between FIT and hand collections were significantly different (F1, 44=5.86, p<0.001, permutations=999) although there was no difference in the homogeneities of variance between overall FITs and hand dissimilarity matrices

(F1,46=2.15, p=0.15). There were significant differences between hand and FITs dissimilarity matrices for floodplain and terra firme forests (F1, 44=3.54, p<0.001, permutations=999) and a significant interaction between collection technique and forest type (F1, 44=5.12, p<0.001, permutations=999). For homogeneities among dissimilarity matrices, there were significant differences between collection technique and forest type

(F3, 44=7.44, p=0.004). This significance is due to pairwise differences between the two techniques in FP forests (p=0.03), FP forest hand samples versus TF forest FIT samples

(p<0.001) and hand samples between TF and FP forests (p<0.001).

Discussion

Effectiveness of sampling methods in terms of numbers of individuals, richness and diversity

Our results indicate that for absolute numbers of carabid beetles, hand sampling was more successful than flight intercept trapping, particularly in floodplain forests. However, in TF forests the number of individuals collected by hand is only slightly higher than

FITs. Although more individuals were collected by hand than by FITs, differences in

287

species richness between sampling techniques were not significant. In studies examining differences among sampling methods, hand sampling has been shown to be an efficient method in terms of number of species for ants (King & Porter 2005; Ellison et al. 2007) and Carabidae in the Peruvian Andes (Maveety et al. 2011). Rarefaction curves and richness estimators indicate that approximately 50% of the Carabidae species that may occur in the understory of FP and TF forests at TBS were sampled, with the majority of remaining species likely to be rare (Colwell & Coddington 1994; Lucky et al. 2002;

Colwell et al. 2012). The stabilizing values and similarity in overall values for nonparametric diversity estimators add confidence to the estimates (Longino et al. 2002).

Incomplete sampling is common for tropical forest arthropod studies due to logistical issues of sampling, the high number of rare species, high beta diversity both spatially and temporally, and the sheer diversity of the tropics (Longino et al. 2002; Lucky et al. 2002;

Coddington et al. 2009; Magurran & McGill 2011). The steep slopes of accumulation curves suggest a high number carabid species are yet to be collected, particularly by FITs.

Additionally, Cicindelinae (Coleoptera: Carabidae) individuals were collected during this study but not included in analysis and results as all individuals (n = 315) were collected by FITs.

Comparison of species assemblages between sampling methods

The distinct species assemblages between FIT and hand sampling are a key finding of this study. Some carabid genera that were rarely collected or missed entirely by one technique were sampled by the other (Appendices 1, 2). Individuals collected by FITs were possibly more cryptic species and those actively flying through the lower understory,

288

while those not sampled by FITs consisted of flight-restricted and/or more sedentary species (Chatzimanolis et al. 2004; Bouget et al. 2008; Missa et al. 2009). Carabid species assemblages sampled by the two methods have a minimal number of shared species. In the NMDS analysis, the limited area of overlap between the hand sampling polygons and the TF FITs polygon is almost entirely due to an outlying FIT site in the TF.

Only three individuals were collected from this study site and a goodness-of-fit test revealed it had the second poorest fit of any point in the analysis. Therefore, we have low confidence in its current placement in NMDS and it is likely to change with greater sample size. With the exception of hand sampling in FP forests (66.3%), species composition by sampling method alone or by method and forest type did not fall below

81.8% dissimilarity, indicating the high species turnover within and between sampling methods in FP and TF understory forests in eastern Ecuador. Multiple collection techniques and sampling of different forest types will collect a greater proportion of

Amazonian carabid communities (Longino & Colwell 1997; Kitching et al. 2001;

Hyvärinen et al. 2006; Missa et al. 2009).

Evaluation of sampling methods among forest types

The lower capture rate for hand sampling in TF sites compared to FP sites may be due to several factors including the vertical stratification of niches and the size of the area sampled relative to the overall coverage of this forest type. Terra firme is the most widespread forest type in Amazonia, including forested land around TBS (Pitman 2000;

Burnham 2002). Therefore, the comparable relative area of TF forest sampled in this study is lower than that of FP forests. We expect additional sampling in TF forests to

289

yield many more species related to proportion of area compared to area sampled in this study. Larger overall area coverage of TF forests may contribute to higher diversity due to several ecological processes including but not limited to metapopulation dynamics and rates of speciation and extinction (Rosenzweig 1995; ter Steege et al., 2000). In addition to restricted area coverage of FP forests, it has been hypothesized that the density and biomass of organisms may be greater than those in TF forests (Haugaasen & Peres 2005,

2009) because of the addition of nutrient-rich sediments to the soils of flooded areas during flooding events (Prance, 1979; Junk et al., 1989). Previous studies have recorded an increase in abundance of terrestrial beetle species (including Carabidae), especially immediately following the reduction in water levels, after a flooding event (Irmler, 1979b;

Adis, 1981). The higher nutrient content in the soils of flooded forests could possibly attract carabids or post-inundated habitats may offer abundant food resources for

Carabidae (Bonn, 2000).

Upland TF forests also typically have a taller and more complex forest canopy than FP forests, resulting in possibly greater vertical stratification of the biota and greater niche availability (Campbell et al. 1986; Richards 1996; Basset et al. 2003). A tropical forest study (Charles & Basset 2005) suggested that forest physiognomy and tree architecture affected vertical stratification of another beetle group, Chrysomelidae.

Additionally, previous tropical forest research indicates that distinct communities of

Coleoptera occur in the understory and canopy (Erwin 1982; Basset et al. 2003; Erwin et al. 2005; Charles & Basset 2005; Stork & Grimbacher 2006). In this study, a few canopy species were collected which occasionally fly through the understory (Erwin et al. 2012).

The taller, more complex canopy of TF forests may be the reason hand sampling in the

290

understory was less effective per time unit compared to FP forests since carabid taxa may be more vertically partitioned and/or there is greater diversity in the canopy (Erwin 1982).

Groups leading an arboreal or semi-arboreal lifestyle are less likely to be sampled in understory collections (Brühl et al. 1998; Basset et al. 2003; Charles & Basset 2005).

Pitfall trapping is the most commonly used collection technique for carabid beetles (Woodcock 2005). It has been shown to be an effective collection technique for a great number of ecological carabid studies in boreal (e.g. Spence & Niemelä 1994;

Niemelä et al. 1996; Koivula et al. 2002; Hyvärinen et al. 2006), temperate (Thiele 1977;

Werner & Raffa 2000; Jennings & Tallamy 2006; Riley & Browne 2011) and some higher latitude or altitude tropical forests (Moraes et al. 2013), but it has not been particularly successful in tropical forests (Paarmann et al. 2001; Paarmann et al. 2002;

Maveety et al. 2011). Most of the hand captures were individuals active on the forest floor; therefore, we do not believe the low catches are due to lower carabid activity rates.

Rather the lower success rate of pitfall traps in lowland tropical forests may be due to the lower abundance of large-bodied carabids (>7 mm) inhabiting the forest floor (Erwin, pers. obs.). Larger Carabidae are more likely to fall into pitfall traps due to higher weight and running speed (Adis 1979) and smaller carabids may be able to more readily detect the traps (Work et al. 2002; Erwin pers. comm). Another possibility is the increased complexity and dimensionality of the understory vegetation, leaf litter and other forest floor components which may create an ineffective catching environment for pitfall trapping (Maveety et al. 2011). Finally, the low success may be due to the hypothesized patchy distribution of Carabidae in tropical forests, with individuals gravitating toward

291

areas of high nutrient availability (e.g. fallen fruit, fermenting flowers and/or leaves, animal waste, fallen trees, etc.) (Erwin 1979; Paarmann et al. 2002).

Conclusions and recommendations

The results of this study demonstrate that FITs and hand sampling are two successful methods for surveying Carabidae in the understory of Amazonian forests. Both sampling techniques resulted in a relatively high number of individuals in a short time period and selected for different portions of the carabid community, representing a wide range of tribes, families and species. In terms of numbers of individuals, richness and diversity, the combined use of FITs and hand sampling was clearly more effective than either method alone. FIT sampling may improve if the area of the flight barrier screen was increased, particularly in width (T. Erwin pers. obs.). Bouget et al. (2008) found no significant differences in abundance, diversity or composition in collection of saproxylic beetles for black versus transparent barriers. Changing the flight barrier to a transparent material, as in a windowpane trap, has shown to be successful for Carabidae and other

Coleoptera in lowland tropical forests of French Guyana (Erwin et al. 2012; Lamarre et al.

2012). Several studies utilizing flight intercept traps, similar to the ones used in this study, or windowpane traps, have reported success with a wider barrier (Peck & Davies 1980;

Chung et al. 2000; Carlton et al. 2004; Chatzimanolis et al. 2004; Lamarre et al. 2012).

Eliminating the PVC pipes might also make the FITs easier to transport to study site locations. For forest studies, the barrier could be secured and drawn taught using strings tied to trees and other vegetation and to stakes in the ground (Martikainen et al. 2000;

Carlton et al. 2004; Chatzimanolis et al. 2004). Lastly, FITs have been effectively

292

employed suspended from one to several meters above the ground (Stork & Grimbacher

2006; Missa et al. 2009; Lamarre et al 2012).

The relatively short timeframe, consisting of a few weeks of collections over two years during the wetter season, limit the conclusions of this study. Results from FITs and hand sampling may vary with season. Although no uniform pattern may exist for the wet tropics (Wolda 1988), studies have shown activity and diversity can peak just prior to or in the wet season for a high number of tropical insect taxa (Wolda 1978; Erwin & Scott

1980; Novotny & Basset 1998; Devries & Walla 2001; Grimbacher & Stork 2009).

Additionally, tropical insects have shown high turnover among seasons (Lucky et al.

2002; Grimbacher & Stork 2009). Carabidae can have species specific responses to seasonal variation depending on life-history factors (e.g. body size, ecology, and trophic guild etc.) and therefore should be collected in the dry and wet seasons as well as the transitional period between the two (Lucky et al. 2002). Sampling should also occur in as many biotopes as possible since carabid species are known to be intimately tied to attributes of particular habitats resulting in patchy distributions (Erwin 1979; Lövei &

Sunderland 1996; Paarmann et al. 2002; Koivula 2011). For a complete assessment of

Carabidae in these forests and to account for vertical stratification, the canopy and various middle forest strata should be sampled in addition to the understory (Erwin 1982;

Lövei & Sunderland 1996; Stork & Grimbacher 2006). This study provides a baseline for future tropical insect studies as the results have wide application across terra firme and floodplain forests of Amazonia. It provides information on effectiveness and complementarity of hand sampling and FITs in two major forest types in lowland

Amazonia for carabid beetles. Additionally, these results may provide guidance for

293

sampling Coleoptera or other insect groups with similar life styles and behavior.

Scientific guidance is necessary in the selection of sampling methods and the relative affinity of particular target taxa (Kitching et al. 2001; Missa et al. 2009). It provides valuable information relating to study design and implementation for future entomofauna studies in Amazonia and other similar forest sites.

Acknowledgments

The Ecuadorian Ministry of the Environment provided collection and exportation permits.

We thank the staff and directors of Tiputini Biodiversity Station and the Universidad San

Francisco de Quito for permission for this research to be conducted. We are grateful to K.

Swing, D. Mosquera, and H. Alverez for assisting with materials and providing vital information for this research project. C. Kliesen, K. Trujillo and J. Macanilla helped to collect specimens. Financial support was provided by Wake Forest University, the

Richter Scholarship Program and the National Museum of Natural History Smithsonian

Institution.

294

References

Adis J. 1979. Problems of interpreting arthropod sampling with pitfall traps. Zool Anz.

202(3/4):177–184.

Adis, J. 1981. Comparative ecological studies of the terrestrial arthropod fauna in Central

Amazonian inundation-forests. Amazoniana. 7(2): 87–173.

Anderson MJ, Ellingsen KE, McArdle BH. 2006. Multivariate dispersion as a measure of

beta diversity. Ecol Lett. 9:683–693.

Anderson MJ. 2001. A new method for non-parametric multivariate analysis of variance.

Austral Ecol. 26(1):32–46.

Anderson MJ. 2006. Distance-based tests for homogeneity of multivariate dispersions.

Biometrics. 62:245–53.

Bass MS, Finer M, Jenkins CN, Kreft H, Cisneros-Heredia DF, McCracken SF, Pitman

NCA, English PH, Swing K, Villa G, et al. 2010. Global conservation significance

of Ecuador’s Yasuní National Park. PLoS One. 5:e8767.

Basset Y, Hammond P, Barrios H, Holloway JD, Miller S. 2003. Vertical stratification of

arthropod assemblages. In: Basset Y, Novotny V, Miller SE, Kitching RL, editors.

Arthropods of tropical Forests: spatio-temporal dynamics and resource use in the

canopy. Cambridge (UK): Cambridge University Press; p. 17–27.

Becker RA, Wilks AR, Brownrigg R, Minka TP. 2015. maps: Draw geographical maps

[Internet]. R package version 2.3-11. [cited 2016 March 2]. Available from:

http://CRAN.R-project.org/package=maps

295

Bivand RS, Pebesma E, Gomez-Rubio V. 2013. Applied spatial data analysis with R

[Internet]. 2nd edition. New York: Springer; [cited 2016 March 2]. Available from:

http://www.asdar-book.org/

Bonn A. 2000. Flight activity of carabid beetles on a river margin in relation to

fluctuating water levels. In Brandmayer P, Lövei GL, Brandmayr TZ, Casale A,

Vigna Taglianti A, editors. Natural history and applied ecology of carabid beetles.

Moscow (Russia): Pensoft Publishers; p. 147–160.

Bouget C, Brustel H, Brin A, Noblecourt T. 2008. Sampling saproxylic beetles with

window flight traps: methodological insights. Rev. Écol. (Terre Vie). 10:21–32.

Brehm G, Fiedler K, Hauser CL, Dalitz H. 2008. Methodological challenges of a

megadiverse ecosystem. In: Beck E, Bendix J, Kottke I, Makeschin F, Mosandl R,

editors. Gradients in a tropical mountain ecosystem of Ecuador. Berlin(Germany):

Springer-Verlag Berlin Heidelberg; p. 41–47.

Brose U, Martinez ND, Williams RJ. 2003. Estimating species richness: sensitivity to

sample coverage and insensitivity to spatial patterns. Ecology. 84(9):2364–2377.

Brühl C, Gunsalam G, Linsenmair KE. 1998. Stratification of ants (Hymenoptera,

Formicidae) in a primary rain forest in Sabah, Borneo. J Trop Ecol. 14(3):285–297.

Brunsdon C, Chen H. 2014. GISTools: Some further GIS capabilities for R [Internet]. R

package version 0.7-4. [cited 2016 March 2]. Available from: http://CRAN.R-

project.org/package=GISTools

Burnham KP, Overton WS. 1979. Robust estimation of population size when capture

probabilities vary among animals. Ecology. 60:927–936.

296

Burnham RJ. 2002. Dominance, diversity and distribution of lianas in Yasuní, Ecuador:

who is on top? J Trop Ecol. 18(6):845–864.

Campbell DG, Daly DC, Prance GT, Maciel UN. 1986. Quantitative ecological inventory

of terra firme and Várzea tropical forest on the Rio Xingu, Brazilian Amazon.

Brittonia. 38(4):369–393.

Carlton C, Dean M, Tishechkin A. 2004. Diversity of two beetle taxa at a western

Amazonian locality (Coleoptera: Histeridae; Staphylinidae, Pselaphinae). Coleopt

Bull. 58(2):163–170.

Chao A. 1984. Nonparametric estimation of the number of classes in a population. Scand

J Stat. 11(4):265–270.

Chao A. 2005. Species richness estimation. In: Balakrishnan N, Read CB, Vidakovic B,

editors. Encyclopedia of statisical sciences. New York (NY): Wiley; p. 7909–7916.

Chapman JA, Kinghorn JM. 1955. Window flight traps for insects. Can Entomol.

87(1):46–47.

Charles E, Basset Y. 2005. Vertical stratification of leaf-beetle assemblages (Coleoptera:

Chrysomelidae) in two forest types in Panama. J Trop Ecol. 21(3):329–336.

Chatzimanolis S, Ashe JS, Hanley RS. 2004. Diurnal/nocturnal activity of rove beetles

(Coleoptera: Staphylinidae) on Barro Colorado Island, Panama assayed by flight

intercept trap. Coleopt Bull. 58(4):569–577.

Chiarucci A, Enright NJ, Perry GLW, Miller BP, Lamont BB. 2003. Performance of

nonparametric species richness estimators in a high diversity plant community.

Divers Distrib. 9(4):283–295.

297

Chung AYC, Eggleton P, Speight MR, Hammond PM, Chey VK. 2000. The diversity of

beetle assemblages in different habitat types in Sabah, Malaysia. Bull Entomol Res.

90(6):475–496.

Coddington JA, Agnarsson I, Miller JA, Kuntner M, Hormiga G. 2009. Undersampling

bias: the null hypothesis for singleton species in tropical arthropod surveys. J Anim

Ecol. 78(3):573–584.

Coddington JA, Griswold CE, Davila DS, Penaranda E, Larcher SF. 1991. Designing and

testing sampling protocols to estimate biodiversity in tropical ecosystems. In:

Dudley EC, editor. The unity of evolutionary biology: Proceedings of the Fourth

International Congress of Systematic and Evolutionary Biology. Portland (OR).

Dioscorides Press; p. 44–60.

Colwell RK. 2013. EstimateS: Statistical estimation of species richness and shared

species from samples [computer program]. Version 9.1.0. [cited 2015 May 22].

Available from: (http://purl.oclc.org/estimates.

Colwell RK, Chao A, Gotelli NJ, Shang-Lin S-Y, Mao CX, Chazdon RL, Longino JT.

2012. Models and estimators linking individual-based and sample-based rarefaction,

extrapolation and comparison of assemblages. J Plant Ecol. 5(1):3–21.

Colwell RK, Coddington JA. 1994. Estimating terrestrial biodiversity through

extrapolation. Philos Trans R Soc London Ser B. 345(1311):101–118.

Devries P, Walla TR. 2001. Species diversity and community structure in neotropical

fruit-feeding butterflies. Biol J Linn Soc. 74(1):1–15.

Dunn RR. 2005. Modern insect extinctions, the neglected majority. Conserv Biol.

19(4):1030–1036.

298

Ellison AM, Record S, Arguello A, Gotelli NJ. 2007. Rapid inventory of the ant

assemblage in a temperate hardwood forest: species composition and assessment of

sampling methods. Environ Entomol. 36(4):766–775.

Erwin TL. 1979. Thoughts on the evolutionary history of ground beetles: hypotheses

generated from comparative faunal analyses of lowland forest sites in temperate

and tropical regions. In: Erwin TL, Ball GE, Whitehead DR, editors.Carabid beetles

their evolution, natural history, and classification. Vol. 1. Dordrecht (Netherlands):

Dr W. Junk by Publishers; p. 539–592.

Erwin T. 1982. Tropical Forests: Their richness in Coleoptera and other arthropod species.

Coleopt Bull. 36:74–75.

Erwin TL. 1991a. How many species are there?: revisited. Conserv Biol. 5(3):330–333.

Erwin TL. 1991b. Natural history carabid beetles BIOLAT Biological Station, Rio Manu,

Pakitza, Peru. Rev Peru Entomol. 33:1–85.

Erwin TL, Geraci CJ. 2009. Amazonian rainforests and their richness of Coleoptera, a

dominant life form in the critical zone of the Neotropics. In: Foottit R, Adler PB,

editors. Insect biodiversity: science and society. 1st ed. Oxford (UK): Blackwell

Publishing; p. 49–67.

Erwin T, Micheli C, Hevel G. 2012. The beetle family Carabidae of Guyane (French

Guyana): the tribes, genera, and number of species known, with notes on their ways

of life. Bull Le Coleopt. 5:1–88.

Erwin TL, Pimienta MC, Murillo OE, Aschero V. 2005. Mapping patterns of β-diversity

for beetles across the western Amazon Basin: a preliminary case for improving

299

inventory methods and conservation strategies. Proc Calif Acad Sci.

56(Suppl1):72–85.

Erwin TL, Scott JC. 1980. Seasonal and size patterns, trophic structure and richness of

Coleoptera in the tropical arboreal ecosystem: the fauna of the tree Luehea

seemannii Triana and Planch in the canal zone of Panama. Coleopt Bull.

34(3):305–322.

Fonseca CR. 2009. The silent mass extinction of insect herbivores in biodiversity

hotspots. Conserv Biol. 23(6):1507–1515.

Gadagkar R, Chandrashekara K, Nair P. 1990. Insect species diversity in the tropics:

sampling methods and a case study. J Bombay Nat Hist Soc. 87(3):337–353.

Global Administrative Areas [Internet]. GADM database of Global Administrative Areas.

Version 2.0. ). University of Berkeley, Museum of Vertebrate Zoology and the

International Rice Research Institute (CA) (2012). [cited 2016 March 2]. Available

from: www.gadm.org

Godfray HCJ, Lewis OT, Memmott J. 1999. Studying insect diversity in the tropics.

Philos Trans R Soc London Ser B. 354:1811–1824.

Gotelli NJ, Colwell RK. 2001. Quantifying biodiversity: procedures and pitfalls in the

measurement and comparison of species richness. Ecol Lett. 4:379–391.

Gotelli NJ, Colwell RK. 2011. Estimating species richness. In: Magurran AE, McGill BJ,

editors. Biological diversity: frontiers in measuring and assessment. New

York(NY): Oxford University Press; p. 39–54.

Grimbacher PS, Stork NE. 2009. Seasonality of a diverse beetle assemblage inhabiting

lowland tropical rain forest in Australia. Biotropica. 41(3):328–337.

300

Haugaasen T, Peres CA. 2005. Mammal assemblage structure in Amazonian flooded and

unflooded forests. J Trop Ecol. 21(2):133–145.

Haugaasen T, Peres CA. 2009. Interspecific primate associations in Amazonian flooded

and unflooded forests. Primates. 50(3):239-251.

Hijmans RJ. 2015. raster: Geographic data analysis and modeling [Internet]. R package

version 2.4-20. [cited 2016 March 2]. Available from: http://CRAN.R-

project.org/package=raster

Hill CJ, Cermak M. 1997. A new design and some preliminary results for a flight

intercept trap to sample forest canopy arthropods. Aust J Entomol. 36:51–55.

Hyvärinen E, Kouki J, Martikainen P. 2006. A comparison of three trapping methods

used to survey forest-dwelling Coleoptera. Eur J Entomol. 103:397–407.

Irmler U. 1979. Abundance fluctuations and habitat changes of soil beetles in Central

Amazonian inundation forests (Coleoptera: Carabidae, Staphylinidae). Stud.

Neotrop. Fauna E. 14(1), 1–16.

IUCN and UNEP-WCMC. 2014-2015. The World Database on Protected Areas (WDPA)

[Internet]. Cambridge (UK): UNEP- WCMC. 2014 October – [cited 2016 March 2].

Available from: www.protectedplanet.net

Jennings VH, Tallamy DW. 2006. Composition and abundance of ground-dwelling

Coleoptera in a fragmented and continuous forest. Environ Entomol. 35(6):1550–

1560.

Junk WJ, Bayley PB, Sparks RE. 1989. The flood pulse concept in river-floodplain

systems. In: Dodge, DP, editor. Proceedings of the International Large River

Symposium. Can Spec Publ Fish Aquat Sci. 106:110–127.

301

Kim KC. 1993. Biodiversity, conservation and inventory: why insects matter. Biodivers

Conserv. 2:191–214.

King JR, Porter SD. 2005. Evaluation of sampling methods and species richness

estimators for ants in upland ecosystems in Florida. Environ Conserv. 34(6):1566–

1578.

Kitching RL, Li D, Stork NE. 2001. Assessing biodiversity “sampling packages”: how

similar are arthropod assemblages in different tropical rainforests? Biodivers

Conserv. 10(5):793–813.

Koivula M, Kukkonen J, Niemelä J. 2002. Boreal carabid-beetle (Coleoptera, Carabidae)

assemblages along the clear-cut originated succession gradient. Biodivers Conserv.

11:1269–1288.

Koivula MJ. 2011. Useful model organisms, indicators, or both? Ground beetles

(Coleoptera, Carabidae) reflecting environmental conditions. Zookeys. 100:287–

317.

Kotze DJ, Brandmayr P, Casale A, Dauffy-Richard E, Dekoninck W, Koivula MJ, Lövei

GL, Mossakowski D, Noordijk J, Paarmann W et al. 2011. Forty years of carabid

beetle research in Europe-from taxonomy, biology, ecology and population studies

to bioindication, habitat assessment and conservation. ZooKeys: 100:55-148.

Kremen C, Colwell RK, Erwin TL, Murphy DD, Noss RF, Sanjayan MA. 1993.

Terrestrial arthropod assemblages: their use in conservation planning. Conserv Biol.

7(4):796-808.

Lamarre GPA, Molto Q, Fine PVA, Baraloto C. 2012. A comparison of two common

flight interception traps to survey tropical arthropods. Zookeys. 55:43–55.

302

Lawton JH, Bignell DE, Bolton B, Bloemers GF, Eggleton P, Hammond PM, Hodda M,

Holt RD, Larsen TB, Mawdsley NA, et al. 1998. Biodiversity inventories, indicator

taxa and effects on habitat modification in tropical forest. Nature. 391(Jan

1998):72-76.

Leather SR, editor. 2005. Insect sampling in forest ecosystems. Oxford (UK): Blackwell

Publishing. 303 p.

Lee S-M, Chao A. 1994. Estimating population size via coverage for closed capture-

recapture models. Biometrics. 50(1):88–97.

Legendre P, Gallagher E. 2001. Ecologically meaningful transformations for ordination

of species data. Oecologia. 129:271–280.

Longino JT, Coddington J, Colwell RK. 2002. The ant fauna of a tropical rain forest:

estimating species richness three different ways. Ecology. 83(3):689–702.

Longino JT, Colwell RK. 1997. Biodiversity assessment using structured inventory:

capturing the ant fauna of a tropical rain forest. Ecol Appl. 7(4):1263–1277.

Lövei GL, Sunderland KD. 1996. Ecology and behavior of ground beetles (Coleoptera:

Carabidae). Annu Rev Entomol. 41:231–256.

Lucky A, Erwin T, Witman J. 2002. Temporal and spatial diversity and distribution of

arboreal Carabidae (Coleoptera) in a western Amazonian rain forest. Biotropica.

34(3):376–386.

Magurran AE, McGill BJ, editors. 2011. Biological diversity: frontiers in measurement

and assessment. Oxford (UK): Oxford University Press. 345 p.

303

Majer JD, Delabie JHC. 1994. Comparison of the ant communities of annually inundated

and terra firme forests at Trombetas in the Brazilian Amazon. Insectes Soc.

41(4):343–359.

Martikainen P, Siitonen J, Punttila P, Kaila L, Rauh J. 2000. Species richness of

Coleoptera in mature managed and old-growth boreal forests in southern Finland.

Biol Conserv. 94(2):199–209.

Maveety SA, Browne RA, Erwin TL. 2011. Carabidae diversity along an altitudinal

gradient in a Peruvian cloud forest (Coleoptera). Zookeys. 147:651–666.

Maveety SA, Browne RA, Erwin TL. 2014. Carabid beetle diversity and community

composition as related to altitude and seasonality in Andean forests. Stud Neotrop

Fauna E. 48(3):165–174.

Maveety SA, Browne RA. 2014. Patterns of morphology in carabid beetles (Coleoptera:

Carabidae) along a Neotropical altitudinal gradient. Int J Trop Insect Sci.

34(3):157–171.

May RM. 1988. Many species are there on Earth? Science (4872). 241:1441–1449.

McGeoch MA. 1998. The selection, testing and application of terrestrial insects as

bioindicators. Biol Rev. 73(2):181–201.

Melbourne BA. 1999. Bias in the effect of habitat structure on pitfall traps: an

experimental evaluation. Austral Ecol. 24(3):228–239.

Mertl AL, Wilkie KTR, Traniello JFA. 2009. Impact of flooding on the species richness,

density and composition of Amazonian litter-nesting ants. Biotropica. 41(5):633–

641.

304

Missa O, Basset Y, Alonso A, Miller SE, Curletti G, Meyer M, Eardley C, Mansell MW,

Wagner T. 2009. Monitoring arthropods in a tropical landscape: relative effects of

sampling methods and habitat types on trap catches. J Insect Conserv. 13:103–118.

Moraes RM, de Souza Mendonça MJ, Ott R. 2013. Carabid beetle assemblages in three

environments in the Araucaria humid forest of southern Brazil. Rev Bras Entomol.

57(1):67–74.

Morse DR, Stork NE, Lawton JH. 1988. Species number, species abundance and body

length relationships of arboreal beetles in Bornean lowland rain forest trees. Ecol

Entomol. 13:25–37.

Niemelä J, Haila Y, Punttila P. 1996. The importance of small-scale heterogeneity in

boreal forests: variation in diversity in forest-floor invertebrates across the

succession gradient. Ecography. 19(3):352–368.

Niemelä J, Kotze J, Ashworth A, Brandmayr P, Desender K, New T, Penev L, Samways

M, Spence J. 2000. The search for common anthropogenic impacts on biodiversity:

a global network. J Insect Conserv. 4:3–9.

Niemelä J. 1996. From systematics to conservation – carabidologists do it all. Ann Zool

Fennici. 33(1):1–4.

Novotny V, Basset Y. 1998. Seasonality of sap-sucking insects (Auchenorrhyncha,

Hemiptera) feeding on Ficus (Moraceae) in a lowland rain forest in New Guinea.

Oecologia. 115(4):514–522.

Oksanen J, Blanchet FG, Kindt R, Legendre P, Minchin PR, O’Hara RB, Simpson GL,

Solymos P, Stevens MHH, Wagner H. 2013. vegan: Community Ecology Package

305

[Internet]. [cited 2015 May 10]. Available from: https://cran.r-

project.org/web/packages/vegan/index.html

Oksanen J. 2013. Multivariate analysis of ecological communities in R: vegan tutorial

[Internet]. p. 1–43. [cited 2015 May 10]. Available from:

http://cc.oulu.fi/~jarioksa/softhelp/vegan.html

Oliver I, Beattie AJ. 1996. Designing a cost-effective invertebrate survey: a test of

methods for rapid assessment of biodiversity. Ecol Appl. 6(2):594–607.

Ozanne CMP. 2005. Sampling methods for forest understory vegetation. In: Leather SR,

Lawton JH, Likens GE, editors. Insect sampling in forest ecosystems. Oxford (UK):

Blackwell Publishing; p. 58–76.

Paarmann W, Adis J, Stork N, Gutzmann B, Stumpe P, Staritz B, Bolte H, Kuppers S,

Holzkamp K, Niers C, et al. 2001. The structure of ground beetle assemblages

(Coleoptera : Carabidae) at fig fruit falls (Moraceae) in a terra firme rain forest near

Manaus (Brazil). J Trop Ecol. 17(4):549–561.

Paarmann W, Gutzmann B, Stumpe P, Bolte H, Kuippers S, Holzkamp K, Niers C, Adis J,

Stork NE, da Fonseca CRV. 2002. The structure of ground beetle assemblages

(Coleoptera: Carabidae) at fruit falls of Melastomataceae trees in a Brazilian terra

firme rain forest. Biotropica. 34(3):368–375.

Pebesma EJ, Bivand RS. 2005. Classes and methods for spatial data in R. R News. 5(2):

9-13.

Peck SB, Davies AE. 1980. Collecting small beetles with large-area “window” traps.

Coleopt Bull. 34(2):237–239.

306

Pitman NCA. 2000. A large-scale inventory of two Amazonian tree communities

[dissertation]. [Durham, (NC)]: Duke University.

Prance GT. 1979. Notes on the vegetation of Amazonia III. The terminology of

Amazonian forest types subject to inundation. Brittonia. 31(1):26–38.

Primack RB, Corlett RT. 2005. Insects: diverse, abundant, and ecologically important. In:

Primack RB, Corlett RT, editors.Tropical rain forests: an ecological and

biogeographical comparison . Oxford(UK): Blackwell Publishing; 198 p.

R Core Team. 2013. R: A language and environment for statistical computing [Internet].

Available from: https://www.r-project.org/

Richards PW. 1996. The tropical rain forest an ecological study. 2nd ed. Cambridge (UK):

Cambridge University Press. 575 p.

Riley KN, Browne RA. 2011. Changes in ground beetle diversity and community

composition in age structured forests (Coleoptera, Carabidae). Zookeys. 621:601–

21.

Rohr JR, Mahan CG, Kim KC. 2007. Developing a monitoring program for invertebrates:

guidelines and a case study. Conserv Biol. 21(2):422–433.

Rosenzweig ML. 1995. Species diversity in space and time. Cambridge (UK): Cambridge

University Press. 436 p.

Ryder Wilkie KT, Mertl AL, Traniello JFA. 2010. Species diversity and distribution

patterns of the ants of Amazonian Ecuador. PLoS One. 5:e13146.

Samways MJ. 1993. Insects in biodiversity conservation - some perspectives and

directives. Biodivers Conserv. 2(3):258–282.

307

Southwood TRE, Henderson PA. 2000. Ecological Methods. 3rd ed. Southwood TRE,

Henderson PA, editors. Oxford(UK): Blackwell Science Ltd. 575 p.

Spence JR, Niemelä J. 1994. Sampling carabid assemblages with pitfall traps: the

maddness and the method. Can Entomol. 126(3):881-894.

Stork NE. 1988. Insect diversity: facts , fiction and speculation. Biol J Linn Soc.

35(4):321–337.

Stork NE, Grimbacher PS. 2006. Beetle assemblages from an Australian tropical

rainforest show that the canopy and the ground strata contribute equally to

biodiversity. Proc R Soc B. 273(1596):1969–1975.

Stork NE, Grimbacher PS. 2009. Seasonality of a diverse beetle assemblage inhabiting

lowland tropical rain forest in Australia. Biotropica. 41(3). 328-337.

Stork NE, Grimbacher PS, Storey R, Oberprieler RG, Reid C, Slipinski SA. 2008. What

determines whether a species of insect is described? Evidence from a study of

tropical forest beetles. Insect Conserv Divers. 1(2):114–119. ter Steege H, Sabatier D, Castellanos H, Van Andel T, Duivenvoorden J, de Oliveira AA,

Ek R, Lilwah R, Maas P, Mori S. 2000. An analysis of the floristic composition and

diversity of Amazonian forests including those of the Guiana Shield. J Trop Ecol.

16(6):801–828.

Thiele HU. 1977. Carabid Beetles in Their Environments. A Study on Habitat Selection

by Adaptation in Physiology and Behaviour. Berlin (Germany): Springer-Verlag.

369 p.

Wagner T. 2000. Influence of forest type and tree species on canopy-dwelling beetles in

Budongo Forest, Uganda. Biotropica. 32(3):502–514.

308

Werner SM, Raffa KF. 2000. Effects of forest management practices on the diversity of

ground-occurring beetles in mixed northern hardwood forests of the Great Lakes

Region. For Ecol Manage. 139(1-3):135–155.

Wolda H. 1978. Seasonal fluctuations in rainfall, food and abundance of tropical insects.

J Anim Ecol. 47(2):369–381.

Wolda H. 1988. Insect seasonality: why? Annu Rev Ecol Syst. 19:1–18.

Woodcock BA. 2005. Pitfall trapping in ecological studies. In: Leather SR, editor. Insect

sampling in forest ecosystems. Oxford (UK): Blackwell Publishing; p. 37–57.

Work TT, Buddle CM, Korinus LM, Spence JR. 2002. Pitfall trap size and capture of

three taxa of litter-dwelling arthropods: implications for biodiversity studies.

Environ Entomol. 31(3):438–448.

309

TABLES

Table V – 1. Numbers of individuals, richness, diversity and nonparametric estimators for Carabidae collected by flight intercept traps (FITs) and hand sampling in floodplain (FP), terra firme (TF) forests and with all data combined.

Variable FIT FP FIT TF All FIT Hand FP Hand TF All Hand Overall No. of individuals* 112 115 227 310 137 447 674 No. of morphospecies 43 37 66 54 43 81 133 No. of uniques (% total S) 24 (56) 20 (54) 34 (52) 23 (43) 33 (76) 42 (48) 66 (50) No. of duplicates (% total S) 11 (14) 7 (10) 14 (21) 12 (8) 5 (7) 16 (20) 27 (20) No. of incidences 81 70 151 144 69 213 364 Fisher’s alpha* (± sd) 25.5 ± 3.9 18.9 ± 2.8 31.3±3.3 18.9 ± 1.8 21.5 ± 2.9 28.9 ± 2.3 49.6 ± 3.1

Richness estimators

Chao2 (± sd) 67.0 ± 13.1 63.2 ± 16.2 105.6 ± 18.3 74.20 ± 11.2 142.8 ± 57.5 133.8 ± 22.2 212.0 ± 26.3

Jack2 76.7 66.7 117.5 85.16 96.96 145.72 235.6

ICE 85.2 60.6 120.0 80.9 179.22 146.4 226.6

Higher Taxa

No. of Tribes 13 11 14 16 9 16 18

No. of Genera 26 18 31 26 17 34 50 * based abundance-based data

310

Table V – 2. Bray-Curtis mean dissimilarity measures within collection techniques

(diagonal) and between collection techniques (shaded boxes) for Carabidae community assemblages. Dissimilarity values range from 0 (identical) to 1 (no overlap). (a) Overall dissimilarity values for flight intercept traps (FITs) and hand sampling; (b) FITs and hand sampling in floodplain (FP) and terra firme (TF) forests.

a) FIT Hand FIT 0.900 Hand 0.955 0.847

b) FIT FP FIT TF Hand FP Hand TF FIT FP 0.818 FIT TF 0.938 0.899 Hand FP 0.960 0.987 0.663 Hand TF 0.894 0.978 0.922 0.867

311

FIGURES

Figure V – 1. a, South America with the Amazon Basin outlined; b, Ecuador with the locations of Tiputini Biodiversity Station (TBS) and Yasuní National Park; and c, the 24 study sites at TBS with floodplain (FP) forest sites represented as filled circles and terra firme (TF) forest as triangles.

312

Figure V – 2. Flight-intercept trap (FIT) utilized in this study. FITs consisted of a frame made of PVC tubing, screening stretched between PVC tubing, with a support tube at the top of the trap, to create an approximately 1 m2 flight barrier, a plastic rain hood and plastic trays.

313

Figure V – 3. Boxplot showing numbers of individuals among total flight intercept traps

(FITs) collections and hand sampling techniques between the two sampling techniques in floodplain (FP) and terra firme (TF). For each box the bold central line is the median, the box indicates 25-75% quartile ranges and whiskers are the minima and maxima values, and outliers are indicated by open circles. The star denotes statistical significance (p < 0.003).

314

Figure V – 4. Sample-based rarefaction curves, corrected by the number of incidences, for flight intercept traps (FITs), for hand sampling and for both methods combined. To compare FITs and hand sampling, the rarefaction curves were corrected to 151 incidences.

Error bars represent 95% confidence intervals.

315

Figure V – 5. Sample-based rarefaction curves, corrected by the number of incidences, for flight intercept traps (FITs) and hand sampling within floodplain (FP) and terra firme

(TF) forests. To compare rarefaction curves, the number of incidences were rarefied at 69 for each of the four curves. Error bars represent 95% confidence intervals.

316

Figure V – 6. The three non-parametric species richness estimators and observed numbers of uniques (species represented in only one sample) and duplicates (species represented in only two samples) by the number of incidences for the complete dataset

(Ni=364).

317

Figure V – 7. Non-metric multidimensional scaling (NMDS) ordination based on 48 data points, representing each of the 24 study sites in terms each sampling methods (FITs and hand) in floodplain (FP) and terra firme (TF) forest (stress = 18.5, k = 2). The square and reverse triangle represents the calculated centroid point for each sampling method.

318

APPENDIX 1

Table V – A1. Tribal and generic composition of numbers of carabid individuals by collection technique and forest type as well as the combined totals. The overall numbers of individuals are shown as grand total. FIT: flight intercept trap, FP: floodplain forest.

TF: terra firme forest. Note: Specimens named ‘New genus’ could not be placed in existing genera. They are likely genera new to Carabidae.

319

FIT Hand Grand Total Taxa FP TF Combined FP TF Combined Bembidiini 42 31 73 23 10 33 106 Erwinana 1 1 2 2 Geballusa 2 5 7 7 Meotachys 2 2 2 Micratopus 4 4 4 Mioptachys 1 1 1 New genus 1 3 1 4 4 New genus 2 2 2 2 Paratachys 27 23 50 23 10 33 83 Polyderis 1 1 1 Callistini 6 6 6 Dercylus 6 6 6 Calophaenini 1 4 5 5 Calophaena 1 4 5 5 Galeritini 3 3 1 1 4 Galerita 3 3 1 1 4 Harpalini 1 3 4 32 20 52 56 Notiobia 1 1 22 14 36 37 Selenophorus 3 3 10 6 16 19 Helluonini 3 1 4 2 2 4 8 Helluobronchus 2 2 2 Helluomorphoides 3 1 4 2 2 6 Hiletini 5 5 5 Ecuamaragnathus 2 2 2 Hiletus 3 3 3 Lachnophorini 2 1 3 126 40 166 169 Amphithasus 3 3 3 Eucaerus 10 10 10 Peruphorticus 2 1 3 104 40 144 147 Pseudophorticus 9 9 9 Lebiini 16 35 51 4 14 18 69 Apenes 1 3 4 4 11 15 19 Coptodera 1 1 1 1 2 Eucheila 1 1 1 Hyboptera 2 2 2 Lebia 12 29 41 41 Negrea 1 1 1 Stenognathus 3 3 3 Loxandrini 18 16 34 25 24 49 83

320

Adrimus 1 1 1 Loxandrus 17 16 33 22 25 47 80 Stolonis 3 3 3 Odacanthini 3 2 5 1 1 6 Colliuris 3 2 5 1 1 6 Oodini 6 6 6 Macroprotus 5 5 5 Oodinus 1 1 1 Pentagonicini 9 20 29 1 1 30 Pentagonica 9 20 29 1 1 30 Perigonini 1 1 1 Perigona 1 1 1 Platynini 4 1 5 2 2 7 Glyptolenus 4 1 5 2 2 7 Pterostichini 18 17 35 26 46 72 107 Abaris 17 17 17 Haplobothynus 1 1 1 1 2 Pseudabarys 1 1 1 Trichonillia 3 3 3 Scaritini 9 9 74 3 77 86 Ardistomis 1 1 1 Camptidius 1 1 1 Camptodontus 1 1 1 Clivina 3 3 38 1 39 42 Nyctosyles 32 1 33 33 Oxydrepanus 4 4 1 1 5 Stratiotes 3 3 3 Zuphiini 1 2 3 3 Pseudaptinus 1 2 3 3 Grand Total 112 115 227 310 137 447 674

321

APPENDIX 2

Table V – A2. Tribal and generic composition of Carabidae numbers of species by collection technique and forest type as well as the combined totals. Overall species numbers are shown as grand total. FIT: flight intercept trap, FP: floodplain forest. TF: terra firme forest.

FIT Hand Grand Taxa FP TF Combined FP TF Combined Total Bembidiini 11 7 15 2 1 2 15 Erwinana 1 1 2 2 Geballusa 1 1 1 1 Meotachys 1 1 1 Micratopus 1 1 1 Mioptachys 1 1 1 New genus 1 1 1 2 2 New genus 2 1 1 1 Paratachys 4 3 5 2 1 2 5 Polyderis 1 1 1 Callistini 2 2 2 Dercylus 2 2 2 Calophaenini 1 2 3 3 Calophaena 1 2 3 3 Galerit ini 2 2 1 1 2 Galerita 2 2 1 1 2 Harpalini 1 2 3 9 8 12 13 Notiobia 1 1 7 4 7 8 Selenophorus 2 2 2 4 5 5 Helluonini 1 1 2 1 2 3 5 Helluobronchus 2 2 2 Helluomorphoides 1 1 2 1 1 3 Hiletini 2 2 2 Ecuamaragnathus 1 1 1 Hiletus 1 1 1 Lachnophorini 1 1 2 12 3 12 12 Amphithasus 2 2 2 Eucaerus 3 3 3 Peruphorticus 1 1 2 5 3 5 5

322

Pseudophorticus 2 2 2 Lebiini 9 13 18 3 13 14 31 Apenes 1 3 4 3 10 11 14 Coptodera 1 1 1 1 2 Eucheila 1 1 1 Hyboptera 1 1 1 Lebia 7 8 11 11 Negrea 1 1 1 Stenognathus 1 1 1 Loxandrini 4 4 5 6 7 10 12 Adrimus 1 1 1 Loxandrus 3 4 4 5 7 9 10 Stolonis 1 1 1 Odacanthini 2 2 4 1 1 5 Colliuris 2 2 4 1 1 5 Oodini 2 2 2 Macroprotus 1 1 1 Oodinus 1 1 1 Pentagonicini 3 3 3 1 1 3 Pentagonica 3 3 3 1 1 3 Perigonini 1 1 1 Perigona 1 1 1 Platynini 1 1 1 1 1 1 Glyptolenus 1 1 1 1 1 1 1 1 1 5 6 7 Pterostichini Abaris 2 2 2 Haplobothynus 1 1 1 1 2 Pseudabarys 1 1 1 Trichonillia 2 2 2 Scaritini 6 6 9 3 10 15 Ardistomis 1 1 1 Camptidius 1 1 1 Camptodontus 1 1 1 Clivina 2 2 5 1 5 6 Nyctosyles 2 1 2 2 Oxydrepanus 2 2 1 1 3 Stratiotes 1 1 1 Zuphiini 1 1 2 2 Pseudaptinus 1 1 2 2 Grand Total 43 37 66 54 43 81 133 Species named ‘New genus’ could not be placed in existing genera and are likely genera new to Carabidae.

323

CHAPTER VI

CONCLUSIONS

The goal of these studies was to better understanding patterns of species abundance, diversity and community assemblages following ecological disturbance, with carabid beetles serving as the model organism. Temperate and tropical forests showed an increase in species richness and diversity in more recently disturbed habitat. Carabid beetle species assemblages were significantly different between recently disturbed habitats compared to habitats not recently influenced by these disturbance factors.

Chapter II demonstrated the relationship between forest age and carabid beetles of

North Carolina temperate forests. Species richness and diversity were less variable with forest age compared to species assemblages. Variations among sampling year affected the number of carabid individuals, species richness, diversity and community assemblage.

Distinct differences in species assemblages among forest age classes suggest that conservation decisions need to be considered at the landscape scale.

In Chapter III, I characterize carabid species assemblages, based on ecological species traits, demonstrating that there were significant shifts in carabid species traits with forest age. However, most trait categories (12 out of 18) did not change significantly

324

with forest age, likely affecting the results of the functional diversity indices for forest age classes. The traits and number of traits selected for multivariate analyses may have strongly affected the outcomes of the functional diversity analyses as they may mask changes in individual traits.

Species richness, diversity and community evenness were higher in Amazonian temporarily flooded forests (FP) than in non-flooded terra firme forests (TF). However, the conclusions are limited by the relatively small area sampled over a relatively short timeframe. The results of Chapter V demonstrate that FITs and hand sampling are two successful methods for surveying Carabidae in the understory of Amazonian forests. Both sampling techniques resulted in a relatively high number of individuals in a short time period and selected for different portions of the carabid community. In terms of numbers of individuals, hand sampling was particularly successful in floodplain forests compared to non-flooded terra firme forests.

325

SCHOLASTIC VITA

KATHRYN NICOLE RILEY

BORN: July 1, 1984, Baltimore, Maryland

EDUCATION: Francis Marion University Florence, South Carolina, 2002-2006 B.S., Biology, 2006

Wake Forest University M.S., Biology, 2010 Thesis Title: Ground beetles (Coleoptera: Carabidae) as biodiversity indicators for age structure in Piedmont forests?

SCHOLASTIC AND PROFESSIONAL EXPERIENCE:

Graduate Teaching Assistantship, Wake Forest University Sept. 2010 – May 2016, Courses: Ecology and Evolution, Genetics and Molecular Biology, Ecology Teaching Assistant, Global Program Studies, Wake Forest University May – June 2016, Ecology and Resource Management of Australia and New Zealand Resident Assistant, International Studies, Wake Forest University Aug. 2014 – Dec. 2014, WFU Semester Abroad Graduate Teaching Assistantship, Wake Forest University Aug. 2008 – May 2010, Courses: Biology and the Human Condition Contract Specialist, Environmental Protection Agency, Washington, DC Nov. 2006 – Aug. 2008 Project Manager, Brook Environmental & Engineering May 2006 – Nov. 2006

326

SCHOLARSHIPS AND GRANTS:

Vecellio Award, Wake Forest University, 2012 Richter Scholar Award, Wake Forest University, 2011 Elton C. Cocke Travel Award, Wake Forest University, 2010, 2011 Alumni Travel Award, Wake Forest University, 2010, 2011, 2012, 2013, 2015 Honorarium, Smithsonian Institution, 2010 Environmental Program financial support, Wake Forest University, 2008, 2009, 2010

PROFESSIONAL SOCIETIES:

Beta Beta Beta, Biological Society, 2005 Phi Kappa Phi, 2005 Chi Alpha Sigma, 2006 Entomological Society of America (ESA), 2008-Present Ecological Society of America (ESA); 2012 Research Partner, Department of Entomology, Smithsonian Institution, National Museum of Natural History, 2015-present Student Collaborator, Department of Entomology, National Museum of Natural History, Smithsonian Institution; 2012-2015 The Coleopterists Society; 2010-Present

PRESENTATIONS:

Riley, KN and RA Browne. 2016. Patterns in ecological species traits of carabid beetles (Coleoptera: Carabidae) with forest age in the Piedmont, North Carolina. Poster Presentation at The 4th international symposium of Carabidology at the University of Georgia, Athens, Georgia.

Riley, KN and RA Browne. 2015. Species traits and functional diversity patterns of carabid beetles (Coleoptera: Carabidae) along a forest age gradient in the Piedmont, North Carolina. Poster Presentation at the Entomological Society of America Annual Meeting, Minneapolis Conference Center, Minneapolis, Minnesota.

Riley, KN, RA. Browne and TL Erwin. 2013. Carabid beetles (Coleoptera: Carabidae) of temporally inundated and non-inundated lowland forest in the Ecuadorian Amazon. Graduate Student Ten-Minute Paper Competition: SysEB, Entomological Society of America Annual Meeting, Austin, Texas.

327

Riley, KN and RA Browne. 2011. Changes in carabid (Coleoptera: Carabidae) diversity and community structure in age structured forests. Oral ten-minute paper presentation at the Entomological Society of America Annual Meeting, Reno-Sparks Convention Center, Reno, Nevada.

Riley, KN and RA Browne. 2011. Changes in carabid (Coleoptera: Carabidae) diversity and community structure in age structured forests. Poster presentation at the Graduate Research Day, Wake Forest University, Winston-Salem, North Carolina.

Riley, KN and RA Browne. 2010. Ground beetles (Coleoptera: Carabidae) as biodiversity indicators for age structure in Piedmont forests. Poster Presentation at the Entomological Society of America Annual Meeting, Town and Country Inn, San Diego, California.

Riley, KN and RA Browne. 2010. Ground beetles (Coleoptera: Carabidae) as biodiversity indicators for age structure in Piedmont forests. M.S. Departmental Seminar. Wake Forest University, Winston-Salem, North Carolina.

PUBLICATIONS:

Riley KN, RA Browne, TL Erwin. 2016. Results from two sampling techniques for carabid beetles (Coleoptera: Carabidae) in temporarily flooded and terra firme rainforest of western Amazonia. Studies on Neotropical Fauna and the Environment. 51(1), 78-95.

Riley KN, RA Browne. 2014. First Documented Record for a Species of Concern, Rhadine caudata (Leconte) (Coleoptera: Carabidae: Platynini), from North Carolina, USA. The Coleopterists Bulletin. 68 (2), 239-240.

Browne RA, SA Maveety, L Cooper, KN Riley. 2014. Ground Beetle (Coleoptera: Carabidae) Species Composition in the Southern Appalachian Mountains. Southeastern Naturalist. 13 (2), 407-422.

Riley KN, RA Browne. 2011. Changes in ground beetle diversity and community composition in age structured forests (Coleoptera, Carabidae) In: Erwin T (Ed) Proceedings of a symposium honoring the careers of Ross and Joyce Bell and their contributions to scientific work, Burlington, Vermont, 12–15 June 2010. ZooKeys

328

Riley KN. 2010. Ground beetles as biodiversity indicators for age structure in Piedmont Forests? MS thesis, Winston-Salem, North Carolina: Wake Forest University.

329