1 ANTHROPOGENIC COMMUNITY DISTURBANCE: GENERAL PATTERNS, QUALIFICATIONS AND ANIM AL MEDIATED NUTRIENT TRANSPORT By A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL OF THE UNIVERSITY OF FLORIDA IN PART IAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY UNIVERSITY OF FLORIDA 2012
4 ACKNOWLEDGMENTS I thank Julie Allen, Leslie Babonis, Nicole Botteri, Scott Clayman, Girard Cua, Rachel Heuer, Umar Karaman, Greg Keller, Tristan Kimbrell, Gustavo Londoo, Ryan McCleary, Yaneke Paulay, and Vanessa Trujillo for their indis pensable and usually cheerful hel p with field work under extremely uncomfortable and sometimes dangerous conditions I could not have done the work without their physical and moral support. Matthew Salomon, Gigi Ostrow, Chikako Matsuba, Gypsy Price, and Ann Laffey assisted immensely with laboratory techniques advice, and support Charlie Baer, John Krigbaum and Andrew Zimmerman kindly allowed me to work in their labs. I also thank Jeong Sun Young and Michael Ritter who assisted with translating the Ko rean and German names. M ichael Barfield, S m r iti Bhotika, N icole Botteri, C olin A. Chapman, R obert D. Holt, G reg Keller, A ndy Lepp, K risten Sauby, and R obert Ulanowicz provided helpful comments on various chapters My committee, Robert Holt, Colin Chapman John Krigbaum, Brian Silliman, and George Tanner, was extremely supportive and patient with the unconventionality of this project and deserve many thanks for their expertise and input. Funding was provided by an American Society of Mammalogists Grant In Aid, a National Science Foundation Doctoral Dissertation Improvement Grant, a University of Florida Alumni Fellowship, and the Department of Zoology. The University of Florida Herbarium provided assistance with plant identification. The Florida Natural H istory Museum kindly provided access to specimens. The Lower Suwannee National Wildlife Refuge cooperated with the study.
5 TABLE OF CONTENTS page ACKNOWLEDGMENTS ................................ ................................ ................................ .. 4 LIST OF TABLES ................................ ................................ ................................ ............ 8 LIST OF FIGURES ................................ ................................ ................................ .......... 9 LIST OF ABBREVIATIONS ................................ ................................ ........................... 13 ABSTRACT ................................ ................................ ................................ ................... 14 CHAPTER 1 OVERVIEW ................................ ................................ ................................ ............ 16 2 CHANGES IN NUTRIENT TRANSPORT PROCESSES OVER GRADIENTS OF ANTHROPOGENIC COMMUNITY MODIFI CATION ................................ .............. 26 Introduction ................................ ................................ ................................ ............. 26 Predictable Human Impacts ................................ ................................ .................... 28 Trophic P rocesses ................................ ................................ ............................ 30 Novel Resources ................................ ................................ .............................. 32 Generalist Omnivores ................................ ................................ ....................... 33 Community Change Rate ................................ ................................ ................. 33 Home Range Size ................................ ................................ ............................ 34 Body Size ................................ ................................ ................................ ......... 34 Effects on Nutri ent Transport ................................ ................................ .................. 35 Conclusion ................................ ................................ ................................ .............. 38 3 TOP DOWN EFFECTS OF A GENERALIST OMNIVORE MEDIATED THROUGH LANDSCAPE PATTERNS ................................ ................................ ... 40 Introduction ................................ ................................ ................................ ............. 40 Suiod Traits as Drivers of Ecological Patterns ................................ .................. 41 Impacts in C oastal Florida ................................ ................................ ................ 44 Methods ................................ ................................ ................................ .................. 45 Study Site ................................ ................................ ................................ ......... 45 Study Design ................................ ................................ ................................ .... 48 Feral Hog rooting ................................ ................................ ....................... 48 Exclosures ................................ ................................ ................................ 49 Landscape composition ................................ ................................ ............. 52 Relative densities ................................ ................................ ....................... 55 Results ................................ ................................ ................................ .................... 55 Discussion ................................ ................................ ................................ .............. 64 Island vs. Mainland Hog Numbers ................................ ................................ .... 64
6 Rooting Activity ................................ ................................ ................................ 67 Hog Effects on Herbaceous Communiti es ................................ ........................ 68 4 A LTERATION OF NUTRIENT FLOWS BY AN EXOTIC SPECIES ........................ 7 5 Introduction ................................ ................................ ................................ ............. 75 Methods ................................ ................................ ................................ .................. 78 Study site ................................ ................................ ................................ .......... 78 Study Design ................................ ................................ ................................ .... 79 Fecal/tiss ue samples ................................ ................................ ................. 79 Diet ................................ ................................ ................................ ............ 81 Nutrient deposition ................................ ................................ ..................... 87 Soil and vegetation nutrients ................................ ................................ ...... 88 Population structure ................................ ................................ ................... 90 Results ................................ ................................ ................................ .................... 94 Diet ................................ ................................ ................................ ................... 94 Nutrient deposition ................................ ................................ ......................... 103 Soil and Vegetation Nutrients ................................ ................................ ........... 99 Population str ucture ................................ ................................ ....................... 105 Discussion ................................ ................................ ................................ ............ 112 Diet ................................ ................................ ................................ ................. 112 Nutrient deposition ................................ ................................ ......................... 116 Soil and plant nutrients ................................ ................................ ................... 116 Population structure ................................ ................................ ....................... 119 The Ecology of Feral Hogs ................................ ................................ ............. 122 Conclusion ................................ ................................ ................................ ...... 126 5 DEFINING WILDERNESS ................................ ................................ .................... 127 Introduc tion ................................ ................................ ................................ ........... 127 Defining Wilderness American Style ................................ ................................ ..... 128 A Basal Definition: The Not So Great Divide ................................ ........................ 132 Universality of the Basal Definition ................................ ................................ ....... 137 Caveats and Conclusions ................................ ................................ ..................... 144 6 TROOPING FAIRIES, TROLLS, AND TALKING TIGERS: THE INFLUENCE OF TRADITIONAL WILDERNESS ARCHETYPES ON CURRENT LAND USE PATTERNS ................................ ................................ ................................ ........... 148 Introduction ................................ ................................ ................................ ........... 148 Method s ................................ ................................ ................................ ................ 153 Wilderness Archetypes ................................ ................................ ................... 153 Folk tales ................................ ................................ ................................ .. 154 Protected areas ................................ ................................ ........................ 154 Analysis ................................ ................................ ................................ .... 156 Land Use Patterns ................................ ................................ .......................... 156 Ruggedness index ................................ ................................ ................... 157
7 Border transects ................................ ................................ ....................... 159 Results ................................ ................................ ................................ .................. 161 Wilderness Archetypes ................................ ................................ ................... 161 Land Use Patterns ................................ ................................ .......................... 161 Ruggedness index ................................ ................................ ................... 161 Border transects ................................ ................................ ....................... 166 Discussion ................................ ................................ ................................ ............ 168 Wilderness Archetypes ................................ ................................ ................... 168 Land Use Patterns ................................ ................................ .......................... 175 Ruggedness index ................................ ................................ ................... 175 Border transects ................................ ................................ ....................... 178 Conclusions and Conservation Impl ications ................................ ......................... 179 7 CONCLUSION ................................ ................................ ................................ ...... 185 APPENDIX A Supplementary information for Chapter 2 ................................ ............................. 193 Plants ................................ ................................ ................................ .................... 193 Invertebrates ................................ ................................ ................................ ......... 193 Insects ................................ ................................ ................................ .................. 193 Cr ustaceans ................................ ................................ ................................ .......... 194 Fish ................................ ................................ ................................ ....................... 194 Reptiles ................................ ................................ ................................ ................. 194 Mammals ................................ ................................ ................................ .............. 194 Birds ................................ ................................ ................................ ...................... 195 B Fecal decomposition study from Chapter 4 ................................ ........................... 199 Methods ................................ ................................ ................................ ................ 199 Results ................................ ................................ ................................ .................. 200 C DELINEATIONS OF SAMPLING AREAS FOR CHAPTER 6 ............................... 202 LIST OF REFE RENCES ................................ ................................ ............................. 209 BIOGRAPHICAL SKETCH ................................ ................................ .......................... 231
8 LIST OF TABLES Table page 3 1 The herbaceous spec ies contained in the study plots, their occurrences in the plots on the islands and mainland, and which area, if any, had significantly more occurrences. ................................ ................................ .......... 61 3 2 A comparison of Feral Hog relat ive abundance indices for the island coastal hammock and marsh area compared with the mainland coastal hammock area ................................ ................................ ................................ .................... 63 4 1 The microsatellite loci used in this study. ................................ ........................... 93 4 2 Coastal Feral Hog diet as estimated through fecal analysis. .............................. 95 4 3 Estimated (N, P, K) deposition rates on the island and mainland coastal areas b y Feral Hogs in the Lower Suwannee National Wildlife Refuge. ........... 105 5 1 Examples of cultural wilderness categories ................................ ...................... 145 6 1 The tota l number of folk tales and protected areas used from each nation with the number of wildernesses occurring in the folktales and terrain/habitat references in the protected area names. ................................ .......................... 155 6 2 Logistic regression results for the ruggedness index (RI) plots for Korea and Germany (at three scales) and the Korean and non Korean large urban areas. ................................ ................................ ................................ ............... 163 AB 1 Calculated nutrient (N, P, K) deposition rates on the island and mainland coastal areas by Feral Hogs in the Lower Suwannee National Wildlife Refuge. ................................ ................................ ................................ ............. 201
9 LIST OF FIGURES Figure page 2 1 Gross trends in community characteristics along the gradie nt of anthropogenic influence ................................ ................................ .................... 31 2 2 Relationships between the direct anthropogenic impacts on biotic c ommunities indirect impacts on biotic communities, and indirect impacts on animal mediated nutri ent transport discussed in the text ................................ ... 36 3 1 An overview of the study site in the Lower Suwannee Nati onal Wildlife Re fuge, Florida ................................ ................................ ................................ .. 46 3 2 A time series of a bulrush area in the tidal marsh, Lower Suwa nnee Wildlife Refuge, Florida. ................................ ................................ ................................ .. 47 3 3 The ma jor habitats surrounding the sampled isla nd and mainland hammock areas ................................ ................................ ................................ ................. 54 3 4 Feral Hog seasonal rooting activity in a coastal hammock community on the mainland (M) as well as isolated hammoc k islands (I) in the tidal marsh in the Lower Suwannee Nati onal Wildlife Refuge, Florida ................................ ........... 56 3 5 The effects of Feral H og exclosures and artificial rooting relative to paired controls on the h erbaceous species richness in a coastal hammock community on the mainland as well as isolated islands in the marsh in the Lower Suwannee Nati onal Wildlife Refuge, Florida. ................................ ........... 58 3 6 The effects o f Feral H og exclosures and artificial rooting relative to paired controls on the herbaceous cover in a coastal hammock community on the mainland as well as isolated islands in the marsh in the Lower Suwannee Nati onal Wildlife Refuge, Florida. ................................ ................................ ....... 58 3 7 The effects of Feral H og exclosures and artificial rooting on the herbaceous biomass in a coastal hammock community on the mainland as well as isolated islands in the marsh in the Lower Suwannee Nat ional Wildlife Refuge, Florid a relative to paired controls ................................ ......................... 59 3 8 The ratio of the island and mainland cumulative bulrush area indices. ............... 62 3 9 Relative abundance indices for Feral Hogs in the coastal mainland hammock and the isolated hammock islands in the marsh in the Lower Suwannee Nati onal Wildlife Refuge, Florida. ................................ ................................ ....... 63 4 1 An overview of the study site in the Lower Suwannee Nati onal Wildlife Refuge, Florida ................................ ................................ ................................ .. 80
10 4 2 Diet of coastal Feral Hogs in the Lower Suwannee National Wildlife Refuge, Florida based o n a nalysis of fecal contents. ................................ ...................... 96 4 3 Stable isotope signatures of potential food resources and the diet of coastal Feral Hogs in the Lower Suwannee Nati onal Wildlife Refuge, Florida. ............... 97 4 4 A comparison of the stable carbon isotope signatures from bone apatite and bone collagen for the indiv idual hogs used in this study. ................................ .... 98 4 5 The relationship between relative age ran k based on tooth eruption and wear nnee National Wildlife Refuge. ................................ ................................ ................................ ............... 99 4 6 Stable isotope signatures of coastal Feral Hog and inland Feral Hog ind ividuals. ................................ ................................ ................................ ........ 100 4 7 Stable isotope signatures of Feral Hogs compared to those of other vertebrates ................................ ................................ ................................ ....... 101 4 8 Depletion of 15 N with distance from the ocean acros s different species of plants. ................................ ................................ ................................ .............. 102 4 9 The r 15 N of the vegetation and distance from the ocean in the Lower Suwannee National Wildlife Refuge, Florida. ............................... 103 4 10 Feral Hog fecal load in a coastal hammock community on the mainland (M) as well as isolated hammock islands (I) in the tidal marsh in the Lower Suwannee Nati onal Wildlife Refuge, Florida. ................................ ................... 104 4 11 Pre treatment nutrient richness of coastal hammock soils in the Lower Suwannee National Wil dlife Refuge from summer 2007. ................................ 106 4 12 Changes in phosphorus, potassium, and nitrate/nitrite nitrogen in the soil in the tilled and exclosure plots co mpared to the paired controls in island and mainland coastal hammock forests in the Lower Suwannee Nati onal Wildlife Refuge, Florida. ................................ ................................ ................................ 10 7 4 13 The treatment effect relative to paired control p lots of 3.5 years of hog exclosures and periodic soil disturbance on phosphorus, potassium, and total Kjeldahl nitrogen (TKN) in herbaceous community biomass ........................... 108 4 14 Nutrient concentrations in herbaceous community biomass in control, hog exclosure, and soil disturbance (tilled) plots in the coastal hammock on the mainland and isolated islands in the tidal marsh after 3.5 years of treatment and 21 months after the last tilling ................................ ................................ ... 109 4 15 Coat color categories of Feral Hogs in and near the Lower Suwa nnee National Wildlife Refuge. ................................ ................................ .................. 110
11 4 16 Locations of observation s of Fer al Hogs by coat color ................................ .... 111 4 17 Genetic groupings of Feral Hogs in the Lower Suwannee National Wildlife Refuge as determined by S TRUCTURE ................................ ............................... 111 4 18 Locations and genetic groupings of Feral Hog samples from the Lower Suwannee National Wildlife Refuge and environs.. ................................ .......... 113 4 19 Mean pair wise genetic distances between indivi duals of the 3 inferred population groupings for Feral Hogs in or near the Lower Suwa nnee National Wildlife Refuge ................................ ................................ ................................ 114 4 20 Aerial photographs of a portion of the field site showing the spati al stability of the marsh community over time. ................................ ................................ ...... 125 6 1 The functional relationships among a wilderness archetype and the cultural and natural qualities that influ ence and are influenced by it ............................. 149 6 2 Comparisons between the wilderness archetypes reflected in folk tales and the geographical references in protected area names for A ) Germany, B ) Ireland, an d C ) Korea ................................ ................................ ...................... 162 6 3 Ruggedness indices of civilized and wilderness plots in Korea and German mountain areas at three scales ................................ ................................ ........ 164 6 4 Ruggedness indices of civilized and wilderness plots in Korean and non Korean large urban areas. ................................ ................................ ................ 165 6 5 Relative elevational transects perpendicular to and centered on the civilization wilderness boundaries in Kor ea and German mountain areas at three different scales. ................................ ................................ ....................... 167 6 6 Coefficients of Variation for the civilization wilderness boundaries in Korea and German mountains for three different transect lengths. ............................. 168 6 7 Elevational transects perpendicular to and centered on the civilization wilderness boundaries in Korean and non Korean large urban areas ............. 169 6 8 Coefficients of Variation for the relative elevational transects centered on the civilization wilderness boundaries in Korean and non Korean large urba n areas ................................ ................................ ................................ ............... 170 6 9 Google Earth images of characteristically Korean A ) and German mountain B ) d evelopment patterns. ................................ ................................ ................. 183 AB 1 An example of the exclosures used in the study ................................ .............. 196 AB 2 Collecti on of the herbaceous biomass. ................................ ............................. 197
12 AB 3 The effects of hog exclusion on the herbaceous community of the coastal hammock in the Lower Suwann ee National Wildlife Refuge after 3.5 years were frequently visually striking. ................................ ................................ ....... 198 AC 1 Feral Hog fecal decay in the coastal hammock of the Lower Suwannee Nati onal Wildlife Refuge ................................ ................................ .................. 200 AD 1 The mountainous areas in Germany (outlined in white) used for the ruggedness indices and civilization wilderness border transects. ..................... 202 AD 2 Urban areas used for comparisons of ruggedness indices and civilizat ion wilderness border transects ................................ ................................ ............. 207
13 LIST OF ABBREVIATION S CI Confidence Interval CV Coefficient of Variation K Potassium N Nitrogen P Phosphorus PCL ppt P arts per thousand RI Ruggedness Index ROC Receiver Operating Characteristic LSNWR Lower Suwannee National Wildlife Refuge TKN Total Kjeldahl Nitrogen
14 Abstract of Dissertation Presented to the Graduate School of the Unive rsity of Florida in Partial Fulfillment of the Requirements for the Degree of Doctor of Philosophy ANTHROPOGENIC COMMUNITY DISTURBANCE: GENERAL PATTERNS, QUALIFICATIONS, AND ANIMAL MEDIATED NUTRIENT TRANSPORT By M ay 2012 Chair: Robert D. Holt Major: Zo ology Ecology has tended to ignore or avoid human impacts for many reasons; yet they are pervasive and must be incorporated into general ecology. Broad patterns of anthropogenic community modification (decreases in maximum body size and home range of consumers ; increases in community change rate, novel resources, and relative importance of generalists; and shifts in the importance of top down vs. bottom up control) should generate predictable changes in animal me diated nutrient transport such as: altered importance, shortened distances, increased cross habitat transfers, and increased flow into refuges. A case study of an exotic generalist, Feral Hogs ( Sus scrofa ) is used to examine these expected changes and t heir landscape dependencies H ogs affected herbaceous species richness and percent cover but not biomass, in landscape specific ways reflecting the juxtaposition of refuge and feeding habitats. The hogs influenced nutrient flows in the system. The sour ce of the nutrients deposited on the islands was the marsh : both fecal and stable isotope analysis revealed a diet largely of the marsh plant American Bulrush ( Schoenoplectus americanus ). B ased on dietary and phenotypic differences between coastal and inl and hogs transport distances were short Thus the role the hogs play in the system are consistent with the hypothesized
15 effects that generalists have on nutrient transport in human impacted systems: short distance inter habitat flow into refuge areas fro m resource areas. The patterns of human impacts hypothesized here and examined with the case study of Feral Hogs in Florida coastal ecosystems are placed in a broader context by examining cultural perceptions of wilderness A basal definition of wilderne ss is presented that can act as a starting point when attempting to understand differing wilderness concepts among cultures L and use patterns should differ among cultures in ways that reflect concepts of wilderness Wilderness archetypes were establishe d from Germany (forests) and Korea (mountains) Land use patterns were found to differ with wilderness archetype These results indicate the important, but largely unrecognized, influence cultural details have on land use patterns and thus anthropogenic disturbance of biotic communities
16 CHAPTER 1 OVERVIEW The science of ecology has largely focused on natural systems with presumed minimal human disturbance ( Collins et al. 2000 ) However, a nthropogenic impacts have a long history of ecological study: indeed one of the earliest published ecological observations was by the Greek philosopher Plato in c. 360 BCE who in a dialogue describing the history of the peninsula of Attica (where Athens is located) has Critias sta te: In its great fertility our land far surpassed any other, for it was then capable of supporting a great army of men who did not work the land [ ] What has now survived of this land can rival any other land in the variety and quality of its crops and th e pasture it offers all species of animals. But, at that time, our land produced all this not only of high quality but in great abundance. You might ask how this is credible and how our present land could possibly be cal led a vestige of our earlier land. [ ] Many and great were the floods [ ] and during this succession of natural disasters the soil was washed down from the high places. It did not form any considerable alluvial deposits, as in other regions, but it disappeared into the deep, as in flood a fter flood it was continuously washed into the sea from all sides. What actually remains is like our small and barren islands, and, compared to the land it once was, Attica of today is like the skeleton revealed by a wasting disease, once all the rich top soil has been eroded and only the thin body of the land remains. But in that age our land was undiminished and had high hills with soil upon them; what we now call the Rocky Barrens were covered with deep rich soil. And in the mountains there were dense forests of which there still survives clear evidence. Some of our mountains can now grow just barely enough for bees, but it was not so long ago that [lofty trees grew there]. There can still be found intact rafters cut from trees that were felled and br ought down to be used for the greatest building projects. And there were many trees that were cultivated for their fruit and they provided limitless fodder for flocks of sheep and goats. Every year there was a harvest of Zeus sent rain. It was not lost, as it is now, as it flows off the hard surface of the ground into the sea, but the deep soil absorbed the rain and it stored it away as it created a reservoir with a covering of clay soil above it; and, as it distributed the water it had absorbed from the high places into its hollows, it produced an abundant flow of water to feed springs and rivers throughout every region of the country
17 [ ] The land was cultivated with great skill [ ] by farmers [ ] who [ ] possessed a most fertile land and water in abunda nce, and above this land a climate and seasons that were most temperate. [ ] [T]he acropolis was very different then than it is now. A single night of torrential rain stripped the acropolis of its soil and reduced it to bare limestone in a storm that was accompanied by earthquakes. Before the destructive flood of Deucalion, this was the third such cataclysmic storm [ ] It was entirely covered by soil and, except for some small outcroppings, level on the top [ ] There was a single spring in the location o f the present acropolis, but it has been choked by the debris of earthquakes [of that night], and its waters now flow only in a trickle [ ] But it provided men of that age with an abundant supply of water. ( Clay 1997, pp. 1296 1298 ) Plato was clearly awa re of the devastating ecological changes that had occurred in Greece due to human activities and intriguingly he made the connection s between loss of vegetative cover soil erosion, increased rain runoff, and lowered water tables. However, he attributed these changed to cataclysmic natural events great floods and earthquakes rather than chronic mundane human activities such as cutting timber, fire, and livestock grazing (Hughes & Thirgood 1982) But the themes Plato identified still resonate as environm ental concerns today. More recently, the study of ecology in a particularly anthropogenic habitat abandoned agricultural fields, leading to what is called old field succession has been a font of ecological theories of competition facilitation, and communi ty assembly ( Billings 1938 ; Clements 1916 ; De Steven 1991; Inouye et al. 1987; Odum 1960; Oostings 1942 ; Oostings & Humphreys 1940 ) Also, t here has been a recent increase in interest in studying ecological patterns and processes in specifically anthro pogenic landscapes and h abitats (such as urban ecology) ( McDonnell & Pickett 1993 ; Pickett et al. 2001 ; Wetterer 1999 ) and a newly minted NSF center (U. of Maryland) is focused on the
18 interface between human and natural processes However, the realization s that rather few ecosystems were functioning in a completely pristine state even at the time of first scientific observation and that essentially no ecosystems are operating free of human influence under the present regime of anthropogenic climate change nitrogen deposition, and other impacts are being but slowly incorporated into ecological paradigms and study designs ( Barlow et al. in press ; Ellis & Ramankut ty 2008 ; Vitousek 1994 ; Vitousek et al. 1997b ) There are many reasons why anthropogenic influences might be ignored by ecologists. Rational re asons, including the desire to study a system in equilibrium or with a complete species comple ment and patterns of landscape and food web connectivity, the minimization of noise in the data, establish ment of a baseline, etc ., are numerous and valid ( Noss & Cooperrider 1994 ; Papworth et al. 2009 ) An untouched pristine system is likely to be closer to some kind of long term equilibrium relative to a system subjected to continually shifting anthropogenic pressures T herefore popul ation demographics, interspecific interactions food web characteristics ecosystem feedbacks, and other measurable qualities are more likely to have reached potential equilibrium states ( be they dynamic or sta ble point attractor ). In contrast, a system w ith anthropogenic influences might be undergoing relatively fast changes in effect existing in an unbalanced transient state that would make interpretation and generalization of the data difficult. For example, predator:prey ratios likely differ tremendo usly between natural systems and anthropogenic systems in which the predators are artificially maintained at high levels due to human supplementation of their diets, as occurs with
19 mesopredators in suburban and urban settings (Robinson & Wilcove 1994) Th is could set prey into long term trajectories towards extinction. Pristine ecosystems gain and lose species fairly slowly compared to systems under anthropogenic influence particularly because dispersal is usually constrained to quite local scales. I t is reasonable to assume that when one observes a community in a natural state complements of species and functional groups present are essentially complete and operating largely in the evolutionary contexts to which their traits are adapted Studying sp ecies relationships and adaptations is much more difficult when important species been extirpated from a system or behaviors are maladaptive under the anthropogenic regime. The traits of species still present in effect are adapted to a community context t hat no longer exists. For example, the fragmentation of the eastern forests of North America has resulted in an increase in Brown headed Cowbirds, an important nest parasite on song birds that prefers open areas and therefore most strongly impacts forest b irds near the edge of the forest (Robinson & Wilcove 1994) Under the se new conditions of fragmentation, augmented by food supplementation in agricultural systems that can sustain high cowbird abundances, the acceptance of cowbird eggs by the host parents is a doomed strategy Yet some host species seem to show no discrimination. This maladaptive behavior seems difficult to understand, until one recalls that under past conditions there was little danger of parasitism by the majority of forest song birds and therefore no need for egg discrimination. So the current host behaviors are maladaptive in the recently changed, current environment. The establishment of baseline ecological data when possible in systems that have not been greatly perturbed by human s, is of course extremely important to be able to
20 understand how systems change and respond to perturbations. Tracking anthropogenic impacts today is of fundamental importance and assessing these impacts require baseline data, either temporal, to track ho w a particular system changes through time, or spatial, to determine how one system is being affected by comparing it to a similar system with less anthropogenic influence. In addition to the plethora of rational reasons to study systems with little anthr opogenic presence, it is likely that emotional reasons also play large role s In spite of the much purported objectivity of scientists, it is no secret that ecologists tend to share an emotional (and some might go so far as to say spiritual) bond with Nat ure ( Wilson 1984 ) This may lead to a kind of bias in the foci of ecological studies: anthropogenic impacts, which tend to be destructive and unsightly are distressing to dwell on, and so may be much less studied Moreover, anthropogenic landscapes almost by definition have humans present, and their imprints, which means that humans must be included among the objects of study. Anthropology is the study of humankind. Yet there has traditionally been a gap between ecology and anthropology in terms of study subjects and frames of reference that likely implicitly reflects the importance of the Great Divide (the Human/Nature dichot omy) in human perceptions of reality (Callicott 2007; Oelschlaeger 2007) As human populations and resource use burgeons, t is becoming increasingly both more important to study anthropogenic impacts and more difficult to find systems that are uninfluenc ed by humans h uman impacts in effect are becom ing pervasive across the globe
21 This study is an attempt to address this issue, including conceptual reviews, a detailed case study, and analyses of cultural perceptions of wilderness. One issue is assessing o verall patterns of biotic community change over the entire gradient of anthropogenic influence There is increasing attention being given to the interplay of population and community processes and ecosystem processes such as nutrient pools and dynamics (t he meta ecosystem process; Leibold et al. 2010; Loreau et al. 2003) Therefore, one theme that is a particular focus of this dissertation is anthropogenic changes in animal mediated nutrient flows To this end in Chapter 2 a series of conceptual models of community traits that are vulnerable to anthropogenic influence are explored and discussed. Although mainly focusing on vertebrates, some of the conceptual models presented there involve community traits that directly involve or touch on other taxa as well. The emphasis on vertebrates derives from my focus on the influences anthropogenic impacts have on animal mediated nutrient flows, which are likely to be particularly mediated by vertebrates. This issue is explored conceptually in this chapter, espe cially the increased role of generalist omnivores in anthropogenic landscapes Hypotheses are developed as to the types of changes likely to be seen in animal mediated nutrient flow along the gradient of human influence. In Chapter s 3 and 4 these hypothe ses are examined in a case study exemplifying the type of generalist omnivore likely to be becoming dominant in animal mediated nutrient flows in human dominated landscapes. This case study involves Feral Hogs ( Sus scrofa ) and their impacts in Florida coa stal ecosystems. Feral Hogs are an exotic species of global concern (Oliver & Brisbin 1993) From domestication centers in Eurasia, they were taken and subsequently released or escaped from captivity
22 throughout the world, becoming pests first in Melanesi a and throughout the Pacific islands and later in Australia, New Zealand, North and South America, and more oceanic islands (Oliver & Brisbin 1993) Because of their feeding strategy of rooting and breeding potential, Feral Hogs have often proved devastati ng to native ecosystems (Hone 2002; Nogueira Filho et al. 2009; Vitousek et al. 1997a) As generalists, they successfully colonize ecosystems as diverse as scrub deserts to tropical rainforests Saunders & McLeod 1999) Not only are they themselves a majo r component in human impacts on many ecosystems, but they also represent the type of generalist (whether exotic or native) that is expected to increase in communities under human influence. Chapter 3 examines how Feral Hogs influence local plant communit ies. Feral Hog impacts on characteristics of the herbaceous community of the coastal hammock system on the gulf coast of Florida are examined, specifically species richness, cover, and biomass The impacts of the hogs were experimentally examined through the use of hog exclosures and artificially tilled plots (mimicking rooting) compared to paired control plots. The presence of the coastal hammock both lining the coast and as islands dotting the tidal marsh enabled an examination of the importance of la ndscape patterns of resource juxtaposition (marsh feeding areas and hammock resting areas) because treatment effects could be compared between the islands and the mainland In Chapter 4 the hypotheses introduced in Chapter 2 regarding the nutrient transpor t syndrome of generalists in a human influenced area are examined in detail for the case study of the Feral Hogs. The source of the nutrients transported by the hogs is
23 established through fecal dietary analysis and stable isotope analysis of bone tissues and foods. The destination of the transported nutrients is determined through fecal surveys (an index of nutrient inputs), and the importance of the nutrients to the herbaceous community is examined through nutrient analyses of the soil and herbaceous bi omass. As in Chapter 3, the influence of landscape patterns and habitat juxtaposition is addressed. The potential for long distance nutrient transport by Feral H ogs is gauged through stable isotope analysis of hogs at different distances from the coast a nd through studies of genetic population structure, both with opportunistically gathered data on hog phenotypes and with a more formal genetic analysis using microsatellites gleaned from fecal samples. The case study contained within Chapters 3 and 4 ill ustrates in a concrete fashion many of the concepts introduced in Chapter 2 This cas study shows that impacts of introduced vertebrates on community composition and nutrient flows can be profound and apply even to a largely natural system on the lower ha lf of the gradient of anthropogenic influence proposed in Chapter 2 However, human impacts differ by more than just the position on an anthropogenic gradient. The latter section of this study recognizes the need to sometimes qualify the generalizations t hat can be made about anthropomorphic changes with cultural details that influence the way people interact with their environment. As cultural views on the position of humans in relation to the natural world are likely important in determining how people will choose to use a landscape, Chapter 5 delves into the Great Divide and addresses the differences and core similarities of differing cultural definitions of wilderness. Problems arising from differences among these definitions are especially
24 germane to conservation efforts, which often incorporate people, strategies, and science from multiple nations and thus lead to a lack of understanding across cultural divisions It is hoped that some of these problems can be more easily understood and overcome wit h an acknowledgement of the universality in human cultures of a basal wilderness concept upon which the diversity of perspectives is built. Chapter 6 takes this basal wilderness concept one step further U sing a comparison between current cultures (most ly Korea and Germany), I examine how differing wilderness archetypes can lead to divergent land use patterns. Wilderness archetypes are examined in traditional folktales as well as in the names of current protected areas. These two sources are likely to reflect the ideals of the peoples rather than just the realities of their environments. The land use patterns are examined in light of the distinction between cultural areas and wilderness made within particular cultures The influence of wilderness arch etypes (and whether they are based on topography or land cover ) on land use decision rules is examined by comparing the ruggedness of cultural areas and wilderness areas between different cultures and with elevational transects perpendicular to the cultura l wilderness borders. Differences in the resistance of a wilderness archetype to development is likely to alter conservation issues and priorities and thus different nations will have unique challenges that will be difficult to understand without an exami nation of the cultural perspective Thus this work is a multifaceted attempt to integrate broad trends in human ecology with animal mediated nutrient transport and landscape patterns of habitat juxtaposition while recognizing the influences of differing cultural norms on the development of landscape patterns and fragmentation. It is broad and ambitious in
25 scope with the hopes of synthesizing information from divergent fields to provide a better perspective on the issues surrounding human interactions wi th their world.
26 CHAPTER 2 CHANGES IN NUTRIENT TRANSPORT PROCESSES OVER GRADIENTS OF ANTHROPOGENIC COMMUNITY MODIFICATION Introduction Before being driven extinct by human actions in the early 20 th century, the Passenger Pigeon ( Ectopistes migratorius ) was the most numerous bird in the world, forming enormous migrating flocks of hundreds of millions that periodically blackened the skies across most of eastern North America. Early observers noted that the depth of the pigeon guano under roosting sites was ty pically several centimeters thick, sometimes reaching over 50 cm deep. This amounted to an estimated deposition of 8 kg of nitrogen (N) and 1 kg each of phosphorus (P) and potassium (K) per hectare in nesting areas over the 30 day breeding cycle. The muc h denser roosting sites would have received an estimated 45 kg N and 7 kg each of K and P per hectare in a single night ( Ellsworth & McComb 2003 ) concentrating nutrients gleaned from a much larger feeding area. Because the area impacted annually by nesting was likely ~10 million hectares and b y roosting was likely ~5 million hectares, the extinction of the Passenger Pigeon must have caused massive but undescribed changes in ecosystem nutrient cycles, flows, and budgets across eastern North America ( Ellsworth & McComb 2003 ) The extinction of the Passenger Pigeon ended the concentrati on of nutrients derived from foraging areas of 20,000 to 80,000 km 2 into roosting areas that were only 0.02 to 10 km 2 and nesting areas that were typically 80 km 2 by 3 5 billion birds over the eastern portion of North America. The devastation to woody v egetation caused by the shear biomass of roosting birds and the elimination of undergrowth by guano deposition is well documented, but the impacts of the elimination of pigeon mediated
27 nutrient transport and concentration on the eastern forests are unknown ( Ellsworth & McComb 2003 ) Other exampl es of the importance of animal mediated nutrient transport include the introduction of foxes to previously predator free islands in the Aleutian archipelago that reduced nesting seabird populations and the marine derived nutrients they transported to such an extent that the plant community shifted from nutrient rich grasslands to tundra ( Croll et al. 2005 ) Sea turtle nesting introduces a significant amount of nutrients into the beach dune systems. Along a 21 km stretch of beach, 14,305 loggerhead turtle nests deposited over 1000 kg of nitrogen and 93 kg of phosph orus, leaving over 700 kg N and 57 kg P in the system after the successful hatchlings returned to the ocean. Because sand is so nutrient poor, this input is likely extremely important for the plant growth that stabilizes coastal dunes around the world and has likely been significantly reduced by anthropogenic reduction in global sea turtle numbers ( Bouchard & Bjorndal 2000 ) While Mohr et al. ( 2005 ) found that the exclusion of Wild Boar and Red Deer and thus their nutrient donations reduced soil fertility in a European forest, the impacts of animals on the nutrient budgets of an ecosystem are often much more complex than simply moving things around. As found in the wildebeest migration system of the Serengeti, the nutrient transport of large migrating herds of ungulates can be overshadowed by their indirect effects such as increasing the speed of nutrient cycling (though digestion of plant matter) and reducing t he frequency of fire (through reduction in fuel loads), a significant route of nitrogen loss in many systems ( Holdo et al. 2007 )
28 In spite of the fact that p atterns of nutrient and energy flow within ecosystems have long been central to ecology and understanding the functioning of ecosystems ( Eva ns 1956 ) it is only recently that t he importance of animal mediated nutrient flows between different ecosystems such as th e examples above has become recognized ( Polis et al. 1997 ; Vanni & Headworth 2004 ) This recognition rests on two pillars of ecology: nutrient/energy flow and the distributions and abundances of species Nutrient flows be tween ecosystems can be disrupted by population reductions and extinctions of some nutrient transporters ( Sekergioglu 2006 ) or by the increase in importance of others, including exotic spec ies ( Ehrenfeld 2010 ) Many of these population changes leading to altered nutrient fluxes are caused or exacerbated by anthropogenic community changes, habi tat reduction, degradation, or fragmentation. In turn, c hanges in patterns of nutrient flow can alter competitive interactions within communities ( Keddy et al. 2000 ) vulnerability to exotics ( Thompson & Leishman 2005 ) and even the physical structure of communities ( Croll et al. 2005 ) likely leading to further community a nd nutrient flow changes It is therefore necessary to understand how human activities in general tend to affect biotic communities and how those changes translate into changes in nutrient flow patterns across landscapes to fully understand how ecosystem s respond to anthropogenic influences. Predictable Human Impacts In spite of the wide range of human cultures, technologies, and subsistence systems, there appear to be predictable progressions of community alteration along the gradient of degree of human influence. While the gradient is multi dimensional and involves many direct and indirect anthropogenic impacts, it is strongly associated with
29 population density ( Ellis & Ramankutty 2008 ) The gradient may be temporal, spatial, or both. A previously untrammeled region may be subjected to greater levels of use and settlement, with anthropogenic impacts increasi ng through time. Alternatively, a system with stable anthropogenic presence can have spatial gradients from rarely exploited hinterlands (or protected areas) to urban areas. Of course, anthropogenic presence is unlikely to be stable so a combination of b oth spatial and temporal gradients is probably most common as anthropogenic impacts wax and wane with population densities, use intensities, and the rise and fall of civilizations. Here I unite the historical and spatial perspectives of anthropogenic cha nge as two expressions of the same process. My goal is to present a synthesis of many ideas and perspectives that are familiar in conservation biology and ecology but to my knowledge have not been integrated together with nutrient transfer processes befor e. Along the gradient of increasing anthropogenic impact, there appears to be a relatively predictable progression of general community changes (Figure 2 1). (Although specific ecosystem or cultural characteristics in some cases can lead to unpredictable types of changes such as potential alternative stable states or differing land use patterns ( Worman 2010 ) these can be considered as overlays and refinements to the general pattern.) I focus on six general ecosystem qualities that are likely to be particularly germane to anthropogenic alterations of animal mediated nutrient transfer: gross trop hic processes (bottom up vs. top down forcing), appearance of novel resources, the relative importance of specialists vs. generalist consumers, rate of community change (in terms of basic parameters such as species identity, diversity, richness, evenness, and connectivity as well as physical traits like structure, temperature regime, landscape
30 diversity, etc.), maximum possible home range size, and maximum body size. (See Appendix A for a selection of specific examples across a range of taxa.) Trophic Proce sses The first notable impact following human colonization of new areas has tended to be the extirpation of the megafauna ( Barnosky et al. 2004 ; Terborgh & Estes 2010a ) Typically the next community response to human impacts has been the loss of large predators ( Terborgh & Estes 2010a ) through habitat loss, fragmentation, and direct persecution for the protection of human life and livestock, although reduction in prey numbers through overhunting and competition from domestic livestock can also play roles. With the loss of top predators, herbivores and mesopredators experience a release (also encouraged by human feeding of free ranging domestic animals), often with devastating effects on their food supplies ( Brashares et al. 2010 ; Ripple et al. 2010 ) However, the larger and herd forming herbivores onl y have a small window of release before habitat loss and fragmentation first reduce opportunities for migration and then reduce blocks of contiguous habitat below the size required for sustainable populations; extirpating them as well. At this point, the absence of the dominant herbivores may allow vegetation to reach a less exploited state superficially similar to conditions in the presence of top carnivores; especially if increases in mesopredators suppress increases in smaller herbivores. However, lan dscapes with adequate refuge areas may experience an explosion of small to medium herbivorous species able to exploit edges and multiple fragmented habitat types (such as the White tailed Deer that have become a problem in many
31 Figure 2 1 Gross trends in community characteristics along the gradient of anthropogenic influence. Arrows indicate influences between community characteristics and influences of events that tend to be concentrated alon g certain parts of the gradient White arrows are positive influences; black arrows are negative influences ; cross hatched arrows have different effects in different parts of the gradient The Top down:Bottom up ratio is the relative strength experienced by herbivores. LMHL = Large Migrating Herd Loss LHL = La rge Herbivore Loss. MPD = Mesopredator Dominance. SMPC = Sequential Mesopredator Collapse.
32 suburban areas in North America; ( Ripple et al. 2010 ) maintaining top down suppression of the vegetation where they are not squeezed out by denser development. Wi th further urbanization, the larger of the remaining herbivores are lost and mesopredators come to dominate the food web until most of them are lost from the system as well ( McKinney 2002 ) Thus the gradient of human impact may be characterized by a shifting of dominance between top down and bottom up processes (Figure 2 1 E ). While the shifting between top down and bottom up forces is likely to be a general character of anthropogenic influence, the exact placement of the shifts for a given area depends on cultural qualities such as prevalence of hunting and tolerance for wildlife, which affect the viability of the remnant wildlife populations ( Terborgh & Estes 2010a ) Novel Resources Through the gradient of anthropogenic i nfluence, novel resources and opportunities increase: from minor scavenging opportunities around small settlement middens, to domesticated livestock, crops, and introduced exotics providing new foods and new concentrations of food, to surges of productivit y caused by increased edge or secondary growth, to the tons of waste produced by industrial societies (Figure 2 1 F ). The anthropogenic concentration and production of food and waste is tied with habitats being altered or replaced largely through livestock production and agriculture. As biotic communities respond to habitats being fragmented and reduced in area, even resources within fragments change as interior species disappear and edge and exotic species increase.
33 Generalist Omnivores While the increasi ng appearance of novel resources and other community changes are likely to be the death knell for specialists dependent on specific foods, predictable patterns, or stable and unfragmented habitats, they provide vast opportunities for generalists able to ta ke advantage of increased landscape diversity with new resources and habitats (Figure 2 1 D ; DeStefano & DeGraaf 2003 ; Fraterrigo & Wiens 2005 ; Marzluff and Neatherlin 2006 ; Prange et al. 2004 ) and the subsequent relaxation of competition from the now ext irpated specialist species ( Fou fopoulos & Ives 1999 ; Henle et al. 2004 ) New sources of food are most likely to be exploited by dietary generalists, in particular omnivores. The mesopredators so infamous for being beneficiaries of human disturbance are largely omnivores whose influence on prey numbers is not only increased by lack of top down suppression by larger predators but also by the avoidance (conferred by omnivory and novel anthropogenic foods) of bottom up control by prey a vailability ( Bra shares et al. 2010 ) Community Change Rate As initial anthropogenic changes lead to community losses and restructuring, those changes themselves likely result in further losses of specialists and relative gains in generalists and exotics (often providing novel resources and/or being generalists themselves; Ruesink 2005 ) thereby initiating a positive feedback period of rapid increases in community change rate (Figu re 2 1 B ). At some point, the community change rate (however defined) must stop increasing, but it is probably maintained at a higher level in rural anthropogenic landscapes through the rapid seasonal community changes characteristic of agriculture and cha nges in land use. In more urban areas, biotic community change rate is likely suppressed compared to agricultural areas due to
34 the permanence of urban structures and land cover compared to the seasonal changes of agriculture. Home Range S ize As remaining habitat patches become smaller, degraded, and more fragmented, the maximum home range size shrinks and large daily movements are likely to be avoided as they can result in greater contact with humans, both of which reduce maximum supportable body size (Fig ure 2 1 C ). This effect is likely to be somewhat mitigated in generalists, which, because of the wide range of resources available to them, can have smaller home ranges than equally sized carnivores ( Harestad & Bunnel 1979 ) or even some herbivores ( e.g. frugivores; Waser 1975 ) This mitigation is likely to be especially important when human altered landscapes that are not conducive to wide ranging (because of roads, fencing, quickly changing land cover, and a general intolerance for many types of wild life) have concentrated novel resources ( Bozek et al. 2007 ; Prange et al. 2004 ; Tigas et al. 2002 ) Body S ize In many anthropogenic landscapes and for many species the existence of refuges is needed for continued survival ( Crooks & Soul 1999 ; Livingston et al. 2002 ) Refuges can serve as areas to avoid persecution or disturbance, area s for breeding, or simply provide space for den sites ( Tigas et al. 2 002 ) What can be considered a refuge depends on the body size of the animal. When human dominated areas are interwoven with extensive natural areas, large bodied generalist omnivores such as wild and feral pigs ( Mayer & Brisbin 1991 ) and baboo ns ( Brashares et al. 2010 ) can thrive. This occurs where there is small scale agriculture and small settlements due to low human
35 population or because larger scale agriculture and settlement are confined, for example, by topography ( Tigas et al. 2002 ; Worman 2010 ) However w hen human dominance is more complete these larger species disappear because they are often viewed as pests and cannot persist in the absence of large refuge s (Figure 2 1 A ) Thus in suburban areas the size of the refuges determines the type of smaller generalist omnivores that can flourish, such as foxes, R accoons O possums ( Prange et al. 2004 ) Brushtail Possums ( Hill et al. 2007 ) and C oyotes ( Connolly 1978 ; Tigas et al. 2002 ) ; with the larger bodied species like coyotes requiring correspondingly larger refuges ( Crooks & Soul 1999 ; Randa & Yunger 2006 ) However, usually only the smallest synanthropic generalis t omnivores, such as rats small birds, insects, etc. can thrive in completely urban ized areas ( McKinney 2002 ) This pattern can be seen as a continuation o f the anthropogenic sequential collapse in body sizes that begins with the extirpation of the megafauna and ends in series of smaller and smaller mesopredators. The process resembles a Russian matryoshka doll that starts as a Wooly Mammoth and ends as a S ewer Rat. Effects on Nutrient T ransport The predictable effects of human activities on community characteristics are likely to have further effects on nutrient cycling within and flow between ecosystems (Figure 2 2). There appear to be four main ways anim al mediated nutrient transport may be being affected by anthropogenic changes: alteration of its relative importance, a shortening of the average transport distance, a relative increase in nutrient transfers among different habitat types, and a particular directionality of flow. Sequential anthropogenic changes in tropic structure (by removal of megaherbivores, then top predators, then large herbivores, etc. ) shift the relative
36 Figure 2 2. Relationships between the direct anthropogenic impacts on biot ic communities (solid outline boxes), indirect impacts on biotic communities (dashed outline boxes), and indirect impacts on animal mediated nutrient transport (dotted outline boxes) discussed in the text. Arrows indicate directionality of influence. For instance, the addition of exotics changes the community and community changes tend to increase vulnerability to exotics. importance of top down and bottom up forces ( Terborgh & Estes 2010a ) Under conditions when dominant herbivore numbers are largely restricted by predation, there will be a larger standing crop of pr oducer biomass and a correspondingly larger flow of
37 nutrients through the detrital food web. However, when herbivores are uncontrolled by predation (with predator proof megaherbivores, a lack of top predators, or near the end of the sequential mesopredato r collapse) a greater fraction of the primary production will be shunted through consumers into the grazing portion of the food web. Under conditions of bottom up forcing, when consumer numbers are limited more strongly by their food supply than by their predators, it follows that animal mediated nutrient transfer may become correspondingly more important. Nutrient transport can be important over both large and small distances ( Dur bin et al. 1979 ) for instance, when the nutrients are retrieved from a sink ( Vanni & Headworth 2004 ) or transported from resource rich zones ( Johnston & Naiman 1987 ) both of which are characteristic of human impacted areas ( Prange et al. 2004 ) With anth ropogenic reduction in home range size, fragmentation of habitats, and hampering of migrations, it is likely that short range nutrient transport has become relatively more important in human dominated areas. This shift is probably exaggerated by the impor tance of refuges in matrices that may provide rich foraging opportunities but discourage wide ranging. The increased complexity of the landscape (more fragmentation and frequent juxtaposition of differing habitat types), the sharpness of edges, the presenc e of concentrated resources in otherwise undesirable habitats and small refuge areas, and the increased relative importance of generalists less limited by habitat type found in human dominated areas all suggest that nutrient transfers across habitat bounda ries should become relatively more important in disturbed compared to undisturbed areas. This could amplify nutrient ties between systems especially in severely fragmented
38 areas Small patches are unlikely to contain all the requirements for life thus i nterhabitat movement goes from possible (for generalists) to necessary. Because of this necessity, foraging often involve s ranging outside refuges and nutrients sequestered from feeding habitats may be concentrated in refuges. This will be especially tr ue in areas where novel resources are concentrated in areas (such as an agricultural field or garbage dump) that are unsuitable for other uses. It seems likely that the overall animal mediated flow of nutrients in anthropogenic landscapes is primarily int o refuge areas largely from novel resources outside of the refuges. This flow may eutrophy the soil or water in the refuges making invasion by exotic species more likely, encouraging further community change in the patches that remain. Interestingly, whil e generalists are likely to be favored in multiple ways by anthropogenic habitat alteration, it seems that their direct and indirect roles in altering nutrient flow patterns in anthropogenic landscapes are probably largely because they are an increasingly dominant guild in the community rather than because of inherent generalist qualities. The primary way inherent generalist qualities influence nutrient transfer is through the ability to use a diversity of habitats and thus transport nutrients from one sys tem to another (Figure 2 2 ). Otherwise, their importance to nutrient regime changes may be simply because they are the primary type of animal able to survive in human modified landscapes and thus become the guild primarily responsible for nutrient flows t hat would be directly or indirectly modified anthropogenically, even if a community were composed entirely of specialist species. Conclusion It is likely that the altered animal mediated nutrient flows created by anthropogenic influence are characterized b y short range, inter habitat transfers with a net flow into
39 refuge areas. Generalist omnivores are likely principal players primarily because they benefit from or are resistant to many anthropogenic changes far more than specialists and thus increase in d ominance in disturbed landscapes. Inter habitat transfers, however, do seem likely to be increased by generalists specifically, as generalists are more likely to incorporate multiple habitats in their daily routines. Of course, the overall pattern of nu trient flow in an anthropogenic landscape will often be dominated by human inputs and concentrations. Eutrophication through fertilization and pollution and the concentration of nutrients from agricultural areas to urban areas have replaced long distance animal mediated nutrient transfer over large swaths of the globe. However, within the human dominated landscape, it is likely that animal mediated transport is important in redistributing nutrients from anthropogenic sources into nearby refuge areas, espe cially when bottom up forces dominate. These sorts of c hanges in nutrient flow patterns may make habitat management and restoration even more challenging as habitat remnants receive influxes of anthropogenically derived nutrients from high densities of ge neralist species supplemented with novel resources. This eutrophication of habitat patches will likely result in upsetting the competitive interactions of the flora and encourage invasion by exotic species, putting native species and communities at furthe r risk The ways in which anthropogenic community changes alter animal mediated nutrient flows have been neither much studied nor quantified, yet they have potentially broad impacts on the functions of ecosystems worldwide and thus need to be further addr essed.
40 CHAPTER 3 TOP DOWN EFFECTS OF A GENERALIST OMNIVORE MEDIATED THROUGH LANDSCAPE PATTERNS Introduction The relative importance of top down and bottom up forces in structuring communities has a long history of description and debate ( Ehrlich & Birch 1967 ; Hairston et al. 1960 ; Oksanen & Oksanen 2000 ; Polis et al. 2000 ; Power 1992 ; Slobodkin et al. 1967 ; Terborgh 2010 ; Terborgh & Estes 2010b ; Terborgh et al. 2001 ) Increasingly it is becoming clear that the loss of large predators and t he subsequent increases in the numbers and boldness of the consumers that are their prey have ramifications that reach much further than merely reducing the standing crop of producer biomass. Increases in the abundance of consumers at intermediate trophic levels can impact ecosystem properties as diverse as disease dynamics, nutrient cycles, vulnerability to invasives, fire regime, species richness, and species diversity ( Daskalov et al. 2007 ; Estes et al. 2011 ; Holdo et al. 2009 ; Maron et al. 2006 ; Pace et al. 1999 ; Paine 1980 ; Terborgh & Estes 2010b ; Ter borgh et al. 2001 ) Moreover, many such consumers are mobile, implying that such shifts can ramify across landscapes. With the loss of predators and a subsequent increase in bottom up control, landscape processes, such as landscape supplementation (whe n additional resources occur in different habitat) and complementation (when needed resources occur in different habitats) ( Dunning et al. 1992 ) are likely to be altered and often become more important as th e impacts of consumers released from predation spread across landscapes. Comparable effects can arise when consumers are introduced into environments lacking predators and parasites.
41 Some of the more globally successful vertebrate consumers in the face of anthropogenic change are members of the Superfamily Suiodea, pigs and their relatives ( Graves 1984 ) These taxa can be key drivers mediating anthropogenic effects on ecosystem properties over the areas of the Earth in which they occur ( Bratton 19 75 ; Nogueira Filho et al. 2009 ; Roemer et al. 2001 ) Their size and defense capabilities limit their vulnerability to mesopredators (except for a few instances of mesopredators acting in packs, e.g. dingoes; Newsome 1990 ) This group of consumers is not only often subjected to strong landscape complementation because of its requirement for permanent water sources (see below) but, because its members are mobil e opportunistic generalists, they also likely often benefit strongly from landscape supplementation, magnifying their impacts on local habitats and patches (a neighborhood effect). In this paper, after sketching some key traits of suiods that enhance thei r ecological effects, we present an observational and experimental study of their impacts on herbaceous plant communities in coastal Florida. Suiod Traits as Drivers of Ecological Patterns Swine and peccaries (Superfamily: Suoidea) have an unusual foraging strategy of rooting for animal and vegetable matter, as well as foraging on aboveground resources. This ability to access hidden resources as part of an omnivorous diet has likely contributed to their successes both as native faunal components on most co ntinents and as exotic invasives of global importance ( Graves 1984 ; Oliver & Brisbin 1993 ) Native species of pig (Family: Suidae) are found in scrub and non boreal forest habitats over most of Eurasia and North Africa ( Oliver et al. 1993 ) The tropical forests of the Malay Peninsula and islands of Southeast Asia hold nine species ( Duff 2004 ) and were likely a center of pig evolution ( Larson et al. 2005 ) Native pig species are also present
42 in most of sub Saharan Africa in forests, bush, and open habitats ( Kingdon 1979 ) Peccaries (Family: Tayassuidae) are found mostly in tropical forests and scrub of South and Central America, with one species extending up to the deserts and scrublands of southwestern North America ( Sowls 1997 ) Australia, most of North America, and most oceanic islands were devoid of suoids for most of the Holocene until pigs (mostly S. scrof a ) were widely introduced by humans ( Oliver & Brisbin 1993 ) These intr oductions were often domesticated pigs that either escaped or were deliberately managed as free ranging livestock. However, in some cases, exotic populations were established either by the introduction of true Wild Boar or the transportation of Feral Hog s from one area to another to facilitate hunting ( Cuevas et al. 2010 ; Graves 1984 ) a process which still occurs ( Hampton et al. 2004a ) (Feral Hogs are defined to be animals whose ancestors were domesticated pigs that escaped captivity and subsequently readapted to the wild. Morphologically they are intermediate between wild and domestic S. scrofa ; Mayer & Brisbin 1991 .) Wild Boar have also in s everal instances reestablished previously extirpated populations by escaping from captivity ( Goulding 2001 ; Welander 2000 ) If suoids as an agent of disturbance have long been part of the selective environment of a plant community, one would expect species in that community to have evolved various adaptations to help cope with such disturbance; all species without such adaptations will have been filtered out of the communit y. Although their characteristic rooting behavior has important impacts on the vegetation in areas where suiods are native ( Ickes et al. 2001 ) rooting in nave ecosystems can be truly
43 devastating ( Hone 2002 ; Nogueira Filho et al. 2009 ; Vitousek et al. 1997a ) Rooting a llows access to subterranean plant storage organs, fungus, and animal foods not readily available to most animals, which contributes to the flexible omnivorous diets of these consumers. These extra resources doubtlessly bolster the extremely high hog dens ities in some areas (from 2 pigs/km 2 in Australia to 37 pigs/km 2 in Hawaii, resulting in ~130 and 1870 kg of pig biomass/km 2 respectively; Saunders & McLeod 1999 ) The destructive potential of suoid rooting on pl ant communities is magnified by these high population densities and by the fact that suoids often forage in large groups, systematically devastating swaths of turf. The effects of exotic hogs as major disturbance agents on plant communities have been desc ribed in diverse ecosystems from deserts ( Cuevas et al. 2010 ) to Mediterranean climate grasslands ( Kotanen 1 995 ) to tropical rainforests ( Nogueira Filho et al. 2009 ) to sub alpine grasslands ( Hone 2002 ) Another suoid char acteristic that affects, and at times constrains, their ecological impacts is their physiological dependence on wallowing for cooling and the maintenance of a healthy skin condition, including the removal of ectoparasites ( Graves 1984 ) Suoid wallowing involves bodies of water or mud, which is thus quite different from the dust bathing o f many other species (which is often also called wallowing). Because the need for wallowing ties most suoids to permanent access to water and shelter from the sun is necessary for heat balance ( Huynh et al. 2007 ; Kingdon 1979 ) landscape complementation is likely to be particularly important to understanding the magnitude of suoid impacts ( Adkin s & Harveson 2007 ; Baber & Coblentz 1986 ; Choquenot & Ruscoe 2003 ; Ickes et al. 2005 ) Even in ar eas (or times; Baber & Coblentz 1986 ) with widely available water, landscape supplementation of resources from neighboring habitats may
44 need to be taken into account to understand the impacts suoids have on their habitat. Most prior studies have concentrated on measuring the physical damage caused by hogs and ignored the potential influence of landscape structure (including the interactive effects of supplementation and complementation) on the location and magnitude of hog impacts. Impacts in Coastal Florida This study investigates the importance of landscape patterns in determining Feral Hog impacts in coastal Florida. This region today lacks a top predator of Fer al Hogs (except for occasional human hunters). Hogs were first released in Florida by the Spaniards, and are abundant with effects that are obvious to even casual observers in many communities (pers. obs.; Abrahamson & Hartnett 1990 ; Engeman et al. 2003 ; Mayer & Brisbin 1991 ) Our focus is on how consumption and rooting disturbance by Feral Hogs influences the structure of understory herbaceous plant communities. Our techniques in particular use exclosures and simulated rooting to mimic hog disturbance. To address the importance of landscape context on this interaction, we compare herbaceous communities on islands of coastal hammock forest embedded in an important feeding habitat (tidal marsh), with herbaceous communities in comparable coastal forest on mainland peninsulas adjacent to the marsh. Field obser vations have shown that Feral Hogs forage considerably in coastal marshes, yet these hogs also utilize forest habitats for nesting and wallowing as well as foraging (unpub. data). The effects of landscape complementation and supplementation should be part icularly strong on the islands, completely surrounded by the marsh foraging habitat. We hypothesized that hogs overall decrease herbaceous species richness, cover, and biomass, in both landscapes, and moreover that hog impacts would be greater on the
45 isla nds because of their increased integration with the marsh feeding areas. This integration should result in the islands providing a more productive foraging landscape while still providing nesting and shade opportunities and therefore should be more heavil y used and impacted by the hogs. Methods Study Site The Lower Suwannee National Wildlife Refuge (214 km 2 ) is on the west coast of centered on the mouth of the Suwannee River on the west coast of northern peninsular Florida The study site is located in a portion of the refuge south of the river in the coastal region (Figure 3 1). The region of the site was chosen for this study because of a combination of accessibility (by road and boat), and because the landscape contained both a number of hammock islands in the marsh and nearby hammock mainland areas. The coastline is rimmed by a slightly brackish tidal marsh (1 2 ppt) fed by the Suwannee River and California Swamp. The marsh is dominated by the fresh /brackish water S aw grass ( Cladium jamaicense ) and B lack N eedle R ush ( Juncus roemerianus ). A somewhat salt tolerant species J. roemerianus extends more seaward than C. jamaicense which lines creeks and hugs the coastal hammocks American Bulrush ( Schoenoplectus ( Scirpus ) americanus ) is found in less densely vegetated areas with a disturbed and looser substrate. The bulrush areas are heavily used by the pigs and are likely maintained by and may even result from pig rooting (Figure 3 2; CW pers. obs.) Preliminary dietary analyses (presented in more detail elsewhere) suggest that the pigs garner considerable food resources from the marsh environment. The complex boundary between the coastal hammock and the marsh allows these three marsh plant species to form a mosaic of m onodominant
46 patches. There are also a few small (~ 10 m 2 ) evaporation pools bordering coastal hammocks that are dominated by Spike Grass ( Distichlis spicata ). Figure 3 1. An overview of the study site in the Lower Suwannee National Wildlife Refuge, Flo locations for this study are indicated. Locations of plot sets are indicated by squares. The coastal hammock bordering the tidal marsh is characterized by E astern R ed C edar ( Juniperus virginiania ) a nd C abbage P alm ( Sabal palmetto ), with oaks ( Quercus
47 sp. ) and planted native pines ( Pinus spp. ) also being common The subcanopy is dominated by Y aupon H olly ( Ilex vomitoria ), Dahoon Holly ( Ilex cassine ), bay ( Persea spp. ) and, in the drier areas, S aw P al metto ( Serenoa repens ). In addition to the mainland, the coastal hammock is also found on small (~1000 12,000 m 2 ) forested islands that are more integrated into the tidal marsh than are the mainland areas (Figure 3 1) The amount of herbaceous undergrowt h varies greatly from none in the wettest areas to complete cover in the areas least subjected to flooding. Figure 3 2. A time series of a bulrush area in the tidal marsh, Lower Suwannee Wildlife Refuge, Florida. Note the same palm snag in the d istance. The foreground is an area of American Bulrush, the background is Black Needlerush, the tip of a Sawgrass area is seen on the right, and islands of coastal hammock are in the far background. A ) The winter of 2005. B ) The area was freshly rooted in the spring of 2 010. C) The spring of 2011. D) A portrait of the porcine perpetrator actively rooting in a different bulrush area in the spring of 2010. B C D A
48 A number of mid sized mammals can be found in these habitats. In particular, the Nine banded Ar madillo ( Dasypus novemcinctus ), Virginia Opossum ( Didelphis virginiana ), and Raccoon ( Procyon lotor ) are common in the coastal hammock including the islands. Less common are the Coyote ( Canis latrans ) and Bobcat ( Felis rufus ), the latter of which is known to travel through the marsh to visit the coastal hammock islands. White tailed Deer ( Odocoileus virginianus ) are rare but present throughout the coastal hammock including the islands. Feral Hogs ( Sus scrofa ), however, are extremely common, compared to t he other species, and moreover are long term residents, having been introduced to Florida almost 500 years ago by Hernando de Soto ( Mayer & Brisbin 1991 ) The study area comprised sampling areas within 3 coastal hammock areas of the mainland and 8 10 island s (depending on the analysis) on either side of Barnett Creek. All of these areas were within 1.25 km of each other (Figure 3 1). These sampling areas were selected based on accessibility, proximity, and habitat. Study Design Feral Hog rooting An index o f the extent of Feral Hog rooting was measured at the end of each season (spring, summer, fall, and winter) from the summer of 2005 to the fall of 2006 by observing rooting within 1 m of either side of 50 m line transects. The rooting adjacent distances w ere summed for each transect and those sums were averaged for the islands and the mainland to form the index. A total of 25 mainland transects were measured each time: 5 for the western portion of the mainland and 10 each for the middle and eastern portio ns. The numbers of mainland transects were based roughly on the length of marsh boundary in each area. A total of 30 transects were measured
49 on the islands each sampling period. The number of transects measured on each island varied from 1 to 5, dependi ng on island size. The percent of the areas rooted was estimated by measuring the length of the line transects that traversed rooted areas and dividing by the total of the transect lengths. Due to the non normal distribution of the data, the distances ro oted for the mainland and islands were compared using a Wilcoxon sum rank test for each season. Exclosures Rooting is only the most visually obvious impact of hogs on the herbaceous community. They also trample, browse/graze, and deposit nutrients; all of these processes could affect the herbaceous community through direct and indirect mechanisms. To a ssess the cumulative impact of Feral H ogs on herbaceous plants, experimental manipulations were carried out within 30 sets of 4 m x 4 m plots: 15 sets on the mainland and 15 on the islands. Five sets were placed in each of the three mainland areas. The number of plot sets per island varied from 1 3, depending on the size of the island. The plot sets were placed to best maximize the spacing between the pl ot sets while still being reasonably representative of the herbaceous community in the area. This semi systematic approach was chosen due to the difficulties of purely random placement of multiple plot sets in small areas. These difficulties include plo t sets overlapping with each other, extending into the marsh, not representing the typical herbaceous community types in an area, or not placed so all three plots could share similar communities. Each set consisted of three plots: a control, an exclosure, and a tilled plot (to mimic increased hog rooting). The exclosures were built to encompass square 4 m x 4 posts at the corners,
50 and a wooden stake in the middle of each side (Figure AB 1) After an exclosure was built, out of the four adjacent 4 x 4 m areas, the two were chosen that qualitatively most resembled the exclosure in terms of the herbaceous communities. Of the two adjacent plots, one was assigned randomly for till treatment and the other served as the control. A till treatment was included to explore the effect of increased physical disturbance on the herbaceous community. The external plots were given a 1 m buffer zone from the exclosure. Tilled plots were lightly tilled with a hand hoe every six months starting in December 2007 (three months after the exclosures were established) until December 2008 to a depth of ~5 10 cm to mimic the physical disturbance of hog rooting. The presence of fresh hog rooting in the non exclosure plots was noted opportunistically throughout the study. Hog rooting and armadillo rooting were clearly distinguishable based on size and geometry of the disturbance. Herbaceous vegetation surveys were done in each plot every six months (3 months offset from the tilling of the tilled plots) from fall 2007 to spring 2009 and one final time in the fall of 2010. Herbaceous species richness (including morphospecies) in the plots was recorded in each sampling period. Specimens were taken to the University o f Florida Herbarium for identification by herbarium staff. In some cases, it was impossible to identify specimens to species because only non reproductive individuals were available for collection. Herbaceous cover was estimated using the point method ( Goodall 1952 ) with five evenly spaced points in four evenly spaced rows in each plot. Vegetation was scored separately in 25 cm height increments up to a height of 1 m for each point.
51 In the spring of 2011, at the end of the field study, aboveground plant samples were harvested to estimate living aboveground biomass of the herbaceous community in the plots (Figure AB 2) Due to its destructive nature, biomass sampling could only be done once over the course of the study. For plants with partly woody stems ( e.g. Smilax spp. and Rubus spp. ) only the herbaceous portions were taken to prevent the dead woody mass from influencing the results. For safety reasons, Poison Ivy ( Rhus radicans ) was no t sampled. Though somewhat widely spread among the plots ( Smilax and R. radicans especially) these taxa were always at very low density and it is reasonable to presume that they contributed little to the overall biomass of the plots. For plots in which the biomass was sparse, the entire aboveground living herbaceous biomass was taken. In more thickly vegetated plots, a subsample of the biomass was taken sequentially from non overlapping randomly placed 1 m x 1 m quadrats until enough biomass had been co llected for laboratory analysis of nutritional content (not presented here). In plots with extremely rank growth, 0.25 m x 0.25 m quadrats were sampled, as harvesting everything from a 1 m x 1 m plot would have been impractical. The samples were dried at 50 C and weighed to estimate above ground living herbaceous biomass. For species richness, cover, and biomass, the effect size of the treatments were calculated by subtracting the cumulative changes in the control from the cumulative changes in the paire d treatment plots. 95% confidence intervals were established by bootstrapping 10,000 times. In cases that were consistent with predictions, the confidence intervals are one tailed, otherwise, they are two tailed. If the confidence intervals did not over lap zero, the effect was deemed significant.
52 The species composition of the herbaceous communities of the islands and mainland were compared by summing the total number of times a species appeared during the vegetation surveys for island plots and also for mainland plots. The frequencies of each species in the two different location types were compared with X 2 Issues of multiple comparisons apply if the existence of at least some significant difference is of interest (which is not the case here) but not if the comparisons are used to be informative about the occurrence pattern of any single species. However, if the overall difference between the island and mainland communities is of interest (as is the case), multiple comparisons do not present a concept ual difficulty, as Type I errors should not be biased in producing false positives for either location type. False positives should reduce the detected difference between the islands and the mainland, making uncorrected comparisons more conservative. The refore, species occurrence ( ) Landscape composition The landscape context for the mainla nd and island study areas was determined by measuring the percent covered respectively by hammock, bulrush areas of the marsh, and the marsh excluding bulrush areas in 50 m segments along 1 km transects using aerial photography provided by Google Earth and ground truthed. (Nearest neighbor distances between islands within a 1 km radius of the study site were determined as well.) Transects into the marsh were placed in the four cardinal directions around each island. As there were 3 mainland areas and 9 i slands, each mainland area was sampled three times to equalize sample sizes with each cardinal direction sampled by three evenly spaced transects per mainland area.
53 For the directions in which the mainland hammock areas were contiguous with the coastal for est, a pseudo boundary of the sampling area, and transect starting point, was set 25 m away from the furthest out exclosure. This corresponds to the overall average distance from exclosures to their nearest marsh edge. Otherwise, the boundary between ham mock and marsh was used as the transect starting point. The transects were 1 km in length so as to enclose the length of the major axis (0.70 0.75 km) of daily ranges of Feral H ogs in a similar tidal marsh system ( Wood & Brenneman 1980 ) This distance should therefore include the habitat available in a typical foraging day for a hog based in each hammock area. The average percentage of each habitat in 50 m segm ents along the transects (given in Figure 3 3) was converted to an index of area for the bulrush feeding areas by squaring the average distance from the hammock for each transect segment and subtracting the square of the next shorter distance to give the r elative area represented by each 50 m segment. Then the relative area was multiplied by the average percentage of bulrush found in each segment to give the relative areas covered by bulrush at each distance from the hammocks. The area index was summed wi th those of shorter distances to give a cumulative area index. This index gives the relative availability of bulrush within a given distance of the island and mainland hammock areas. The ratio of the island and mainland cumulative area indices for the bu lrush feeding areas is comparable to the ratios of the island to mainland hog density indices (see below). 95% confidence intervals for the ratio were established by bootstrapping to determine at which distances the ratio deviated from 1.
54 Figure 3 3 T he major habitats surrounding the sampled island and mainland hammock areas. The bulrush areas used for feeding by the Feral H ogs are indicated separately from the rest of the marsh. Note that the marsh area is constant with distance from the mainland ar eas but increases closer to the islands giving the islands greater access to the marsh. Note also that the peak of bulrush extent is closer to the island hammocks than the mainland hammock.
55 Relative densities Relative hog densities for the mainland area a nd the island areas were established by compiling opportunistically gathered data into three indices: direct encounters, hog remains, and hog nests. Encounters in the marsh were assigned to the closest hammock area. Reencounters with the same individual s in the same day were not counted. Hog remains were never found more than 10 m from a hammock. Remains were considered to be from the same individual if they were found within ~20 30 m of each other, were consistent in size and age, and did not repeat skeletal elements. Nests were constructed by the hogs gathering pine needles, palm fronds, or other detritus into oval rings averaging 2 m x 1.5 m x 0.2 m. Nests were counted if they appeared to be fresh and had not been counted before. Because the is lands were accessed by passing through the mainland areas, the three opportunistic indices are likely biased towards the mainland. Therefore, an unbiased index was calculated based on systematic hog fecal counts made in 50 x 2 m plots (26 30 island plots and 25 mainland plots). The counts were done every season from the summer of 2005 until the fall of 2006 with an additional sampling in the winter of 2008 2009. New plots were used for each season. Piles of feces, defined as a cluster of fecal pellets o f similar size and apparent age within 0.5 m of each other, were counted in the plots. Because the fecal decay rate was extremely low (50% survival to 471 days), changes over time could not be examined. Results The amount of rooting varied widely among lo cations within a season and also varied widely among seasons for a single location. Transects on the mainland had
56 significantly more hog rooting than did those on the islands in the 2005 seasons was estimated to be 6, 9, and 3% rooted in the summer, fall, and winter of 2005, respectively. In the same time periods, the islands were estimated to be 1, 1, and 0.3% rooted. However, in 2006, rooting in the hammocks plummeted (mainland spring, summer, and fall: 0.44, 0.12, and 0.08%; island: 0.04, 0.39, and 0.46%) and no significant differences were found between the islands and the mainland (Figure 3 4). Figure 3 4. Feral Hog seasonal rooting activity in a coastal hammock community on the mainlan d (M) as well as isolated hammock islands (I) in the tidal marsh in the Lower Suwannee National Wildlife Refuge, Florida. Seasons in which there medians, the boxes indicate the 25 th and 75 th percentiles, and the whiskers indicate the last sample within 1.5 times the interquartile distance One unexpected result of the study was that experimental manipulation could positively influence hog behavior. Tilling increased the chan ce of hog rooting compared to control plots. In 74 instances of hog rooting in plots, 50 were observed in tilled plots 2 = 9.1, df = 1, p = 0.002). Additionally, tilling increased the
57 average extent of rooting in the plots that were rooted (tilled: 23% rooted; control: 11% rooted ; Wilcoxon sum rank test: S = 666.5, one tailed p = 0.007). In a paired comparison in which at least one of the two plots for each location was rooted, tilled plots had on average 16% more of their area rooted than did their respective controls (Wilcoxon signed rank test: S = 563.0, p < 0.0001). The overall average herbaceous species richness in controls was greater in the mainland controls ( = 7.0 species/16 m 2 plot, SD = 3.8, range = 0 17 sp./plot) than in the island controls ( = 3.2 species/16 m 2 plot, SD = 2.5, range = 0 13 sp./plot; Wilcoxon sum rank test: S = 7361, p < 0.001). The response of species richness to the treatments was different on the islands and the mainland. There was no detectable effe ct of the exclosures compared to controls but an immediate and temporary effect of the tilling on herbaceous species richness in the island plots. However, on the mainland, there was a significant and asymptotic increase in species richness in the exclosu res (~1.6 species/16m 2 plot) compared to controls and a lagged decrease in the tilled plots ( 2 species/16m 2 plot) and subsequent recovery 15 months after tilling was ceased (bootstrapped 95% confidence intervals are shown in Figure 3 5). Herbaceous cove r showed similar patterns to species richness. Overall, cover in controls was higher on the mainland ( = 30%, SD = 0.30, range = 0 95%) than on the islands ( = 19%, SD = 0.24, range = 0 75%). The exclosures had no detectable increase in herbaceous cover on the islands but a slight possibly asymptotic increase on the mainland (~6%) compared to controls ; however this was only significant in one month. Tilling had a negative effect in both areas compared to co ntrols but a stronger
58 impact on the mainland (island: 5%; mainland: 21%) with little or no apparent recovery a full 15 months after tilling ceased (island: 6%; mainland: 14%; Figure 3 6). Figure 3 5. The effects of Feral H og exclosures and artific ial rooting relative to paired controls on the herbaceous species richness in a coastal hammock community on the mainland as well as isolated islands in the marsh in the Lower Suwannee National Wildlife Refuge, Florida. Tilling treatment times are indicat ed with arrows. 95% confidence intervals are indicated. Figure 3 6. The effects of Feral H og exclosures and artificial rooting relative to paired controls on the herbaceous cover in a coastal hammock community on the mainland as well as isolated islan ds in the marsh in the Lower Suwannee National Wildlife Refuge, Florida. Tilling treatment times are indicated with arrows. 95% confidence intervals are indicated.
59 Herbaceous biomass was higher in the mainland controls ( = 7.7 g/m 2 SD = 9.4, range = 0.1 28.7 g/m 2 ) than in the island controls ( = 3.9 g/m 2 SD = 5.0, range = 0 16.4 g/m 2 ). Tilled plots on the islands had reduced biomass compared to their controls ( 1.4 g/m 2 ) while tilled plots on the mainla nd were not significantly different from their controls 15 months after tilling had ceased. Exclosures on both the islands and mainland had similar increases in biomass compared to their controls (island: 13.5 g/m 2 ; mainland: 12.9 g/m 2 ; Figure 3 7 Figure AB 3 ). Figure 3 7. The effects of Feral H og exclosures and artificial rooting on the herbaceous biomass in a coastal hammock community on the mainland as well as isolated islands in the marsh in the Lower Suwannee National Wildlife Refuge, Florida rel ative to paired controls The biomass was collected after 3.5 years of exclosure treatment and 21 months after tilling was ceased. 95% confidence intervals are indicated.
6 0 Out of 56 herbaceous species, 54 were found on the mainland, while only 39 were fo und on the islands (Table 1). Note that the area sampled is equal for the mainland and islands. Of these 56 species, 32 were found more frequently in the mainland plots 2 < 68.2, all p < 0.04), 22 were found with equal frequency in both areas, and only 2, Spanish Dagger ( Yucca aloifolia ) and a vetch ( Vicia sp. ), were found more frequently in the island plo ts (X 2 = 5.0, p = 0.025 and X 2 = 8.0, p were found more frequently on the mainland, 40 were found with equal frequencies, and none were found more frequently on the isl ands. In either case, significantly more plant species had higher frequencies on the mainland than on the islands (X 2 = 26.5, p < 0.0001; X 2 = 16, p < 0.0001; respectively). The 95% confidence intervals of the cumulative bulrush area index ratio did not i nclude 1 for distances of 100 to 300 m from the hammock edge; indicating that greater bulrush area is accessible from the islands than from the mainland in that distance range. In that range, the islands provided access to an average of 2.8 times (range: 2.3 3.4) the area of bulrush habitat as the mainland did. At no distance did the mainland provide more access than the islands to bulrush feeding areas (Figure 3 8). Nearest neighbor distances between islands averaged 55 m and ranged from 10 to 273 m, w ithin the zone of bulrush advantage when foraging from the islands compared to foraging from the mainland. Relative hog densities averaged 3.4 times higher in the island/marsh than in the coastal mainland (6.1 < X 2 < 420, all p < 0.01; Table 3 2). The unb iased fecal count index was 4.4 times higher in the island/marsh than on the mainland (Figure 3 9).
61 Table 3 1. The herbaceous species contained in the study plots, their occurrences in the plots on the islands and mainland, and which area, if any, had si gnificantly more occurrences. Occurrence (225 max) Common name Latin name Family Island Mainland More? Beaked Panicgrass Panicum anceps Poaceae 132 126 St. Augustine grass Stenotaphrum secundatum Poaceae 79 153 M** Poison Ivy Toxicodendron radican s Anacardiaceae 67 124 M** green briar Smilax sp. Smilacaceae 82 103 fimbry Fimbristylis sp. Cyperaceae 72 107 M marsh pennywort Hydrocotyle umbellata or verticillata Apiaceae 16 108 M** Wand Goldenrod Solidago stricta Asteraceae 36 81 M** Basketgras s Oplismenus hirtellus Poaceae 19 92 M** Carolina Wild Petunia Ruellia caroliniensis Acanthaceae 15 59 M** Sawgrass Cladium jamaicense Cyperaceae 22 33 Straw colored Flatsedge Cyperus cf. strigosus Cyperaceae 18 36 M Erect Centella Centella erecta Ara liaceae 1 48 M Frogfruit Phyla nodiflora Verbenaceae 4 37 M** Dichanthelium sp. Poaceae 13 26 M unidentified grass* Poaceae 9 26 M Trumpet Creeper Campsis radicans Bignoniaceae 3 32 M** black/dewberry Rubus sp. Rosaceae 1 33 M** rosette grass Dicha nthelium sp. Poaceae 10 21 M Mikana sp. Asteraceae 7 20 M Shiny Woodoats Chasmanthium nitidum Poaceae 7 19 M unidentified herb* 4 22 M** Saltbush Baccharis halimifolia Asteraceae 7 17 M Saltmarsh Fingergrass Eustachys glauca Poaceae 10 13 Florida Paspalum Paspalum floridanum Poaceae 7 14 Rice Button Aster Aster dumosus Asteraceae 1 20 M** unidentified legume* Fabaceae 0 21 M** vetch Vicia sp. Fabaceae 17 4 I White topped Sedge Dichromena colorata Cyperaceae 0 20 M** Giant Ironweed Vernonia gigantean Asteraceae 9 10 Climbing Aster Symphyotrichum carolinianum Asteraceae 1 17 M** Saltmarsh Morning glory cf. Ipomoea sagittata Convolvulaceae 12 5 Physalis walteri Solanaceae 9 5 Seaside Brookweed Samolus valerandi parv iflorus Samolaceae 4 8 jessamine Gelsemium sp. Gelsemiaceae 0 12 M** Leafy Marsh pink Sabatia calycina Gentianaceae 0 11 M** unidentified herb* 0 8 M violet Viola sp. Violaceae 0 8 M Virginia Creeper Parthenocissus quinquefolia Vitaceae 0 8 M unid entified herb* 0 7 M unidentified grass* Poaceae 2 5 Spanish Dagger Yucca aloifolia Asparagaceae 5 0 I Saururus cernuus Saururaceae 0 4 M hibiscus Hibiscus sp. Malvaceae 0 4 M Black Needlerush Juncus roemerianus Juncaceae 3 1 unide ntified herb* 1 2 unidentified herb* 1 1 Roundleaf Bluet Houstonia procumbens Rubiaceae 1 1 waterhyssop Bacopa sp. Veronicaceae 0 2 Ponysfoot Dichondra sp. Convolvulaceae 1 1 Blue flag Iris versicolor Iridaceae 0 1 Jewelweed Impatiens capen sis Balsaminaceae 0 1 unidentified herb* 1 0 bedstraw Galium sp. Rubiaceae 0 1
62 Table 3 1 Continued Occurrence (225 max) Common name Latin name Family Island Mainland More? thistle Cynareae 0 1 unidentified herb* Asteraceae 0 1 unidentif ied herb* 0 1 *Unidentified plants were unable to be identified by the University of Florida Herbarium because the samples were either vegetative or unavailable in the case of species with a single occurance. ** The difference is still significant when Figure 3 8 The ratio of the island and mainland cumulative bulrush area indices. Bars indicate 95% confidence intervals. Numbers above bars indicate upper 97.5% confidence limits too high fo r the vertical axis. Within 300 m islands are surrounded by greater areas of bulrush used as feeding grounds by the Feral H ogs than is the mainland. The highest peak (3.6) approximates and the confidence intervals between 100 and 300 m enclose the estima te for the ratio between island and mainland hog densities. At greater distances, the difference disappears and the ratio approximates 1.
63 Table 3 2. A comparison of Feral Hog relative abundance indices for the island coastal hammock and marsh area compar ed with the mainland coastal hammock area. All indices show significantly lower hog densities on the mainland. The bias towards the mainland in the three opportunistic sample based indices results from the mainland being used to access the islands. Obse rvations Index Island Mainland Island/Mainland Bias Sampling strategy Heard/Seen 90 52 1.7 Mainland Opportunistic Remains 23 9 2.6 Mainland Opportunistic Scat 1134 189 4.4* None Systematic Nests 29 6 4.8 Mainland Opportunistic *Corrected for sampl ing area. Figure 3 9. Relative abundance indices for Feral Hogs in the coastal mainland hammock and the isolated hammock islands in the marsh in the Lower Suwannee National Wildlife Refuge, Florida. The scat count index is the only one for which time s pent or area covered is controlled for and is therefore unbiased. The others are probably slightly biased towards the mainland. All indices show significantly higher hog densities in the island/marsh area than on the mainland.
64 Discussion The impact of mo bile, generalist consumers upon local communities of their resources is likely to reflect processes operating over broader spatial scales ( Polis et al. 1997 ) Understanding such impacts can be crucial for a wide range of issues in conservation biology and environmental management. In our study, landscape context appeared to have a strong influence on the impact of Feral Hogs, however not always in the ways initially predicted. We first discuss the factors that control differences between hog numbers on islands and the mainland, and the seemingly contrary patterns in rooting magnitude, then turn to differences in im pacts of hogs on herbaceous plant communities as revealed by the exclosure studies. Island vs. Mainland Hog Numbers Hog density as estimated by several indices was 3 4 times higher on the island hammocks than on the mainland, reflecting, we suggest, the fact that the islands are embedded within an important feeding habitat. It is worth discussing in more detail several processes that would be operating to produce this density effect. The hogs seem to spend a great deal of time foraging in the bulrush are as of the marsh, as judged by the extent of damage they cause there, by the cumulative impression of incidental observations, and by unpublished dietary data. Though peninsulas from the mainland do abut and extend into the marsh, an animal living on one o f the islands would on average have much more access to greater areas of marsh than an animal living on the mainland (seen in Figure 3 3). Although hogs are often observed foraging in the marsh completely exposed to the sun on cloudless and hot days, hamm ock areas do provide shelter from the sun as well as dry land for rest and respite from flooding in the marsh and are therefore likely to be heavily used for shelter
65 even if lacking in food resources. Wallows are most commonly found in the shaded edges be tween hammock and marsh (although they do also occur at times in the marsh itself; CW, pers. obs.), which suggests that the marsh is used primarily for feeding, while the hammocks are preferred at other times for other activities. This landscape complement ation/supplementation we suggest is likely the main reason why hog densities are so much higher in the islands. This pattern is consistent poor hammock) have much greater dens ities of consumers sustained in effect by more productive habitats (the feeding areas in the marsh) than would be found on greater ( Oksanen 1990 ) Yet note that the heavy use of the islands is likely to also be r elated to distinct conditions they provide (viz., shade and dry land) not found in the tidal marsh and therefore this kind of landscape effect is more complex than just a simple spill over effect. However, other mechanisms could also contribute to the obse rved island effect. Another reason for the greatly elevated abundance of pigs on islands could be that the islands provide better security for piglets from mesopredators. It should be noted that Bobcats have been tracked onto the islands, as have Coyotes (whose scat has been observed to contain hog hair and small hooves, consistent with predation on piglets); mesopredators thus are unlikely to be excluded from the islands entirely. However, the islands are small and completely surrounded by periodically flooded marsh, so predator density may be lower and the ability to detect predators may be higher. In any case, hogs do not make up a large part of the diets of Bobcats and Coyotes; Florida Panthers, currently extirpated from the area, are the only effect ive adult hog predators that exist
66 anywhere in Florida ( Maehr 1997 ) It should be noted that hunting is practiced in the refuge so human hunters are in effect an other top predator on adult hogs. Many hunters know that the islands have high hog densities and so specifically hunt on the islands (C.W., pers. obs.) but hunting season is restricted to the fall/winter months. The overall combined predation risk, exper ienced mostly by piglets, may thus be slightly lower on the islands at least outside of hunting season. Although this is plausible, other lines of evidence suggest that hog abundance is actually regulated by resources, rather than predation. Judging from the emaciated hogs (especially lactating females) observed during the course of this study (C.W.), top down control (predation) is likely to be much less important than bottom up control (nutritional stress) in determining hog density. The area of bulrush feeding grounds surrounding the islands can be up to 3.4 times the area surrounding the mainland. The maximum access to resources conferred by island living is found by foraging within 150 meters of islands with a significant advantage extending to foragi ng within 300 m. Because the average distance between islands is 55 m with even the longest nearest neighbor distance between islands being shorter than 300 m it is clearly more efficient to feed in bulrush areas by moving between islands rather than fora ging from a mainland base; a neighborhood effect ( Dunning et al. 1992 ) The fact that the islands appear to have similar magnitudes of higher foraging area and hog densities compared the mainland supports the conclusion that the islands are far more important for providing access to resources than they are for providing protection for predators.
67 This difference in consumer abundance, however, is not necessarily matched by corresponding hog impacts either as disturbance agent gauged by rooting activity or by net effects on plant communities. Rooting Activity When there was a difference in observed rooting extent, the mainland had more evidence of rooting, despite by all signs having fewer hogs per unit area. This suggests that there is a difference in individual hog foraging behavior, depending on whether or not the hog is on an island or on the mainland. The percents of the hammock areas on the mainland rooted in 2005 (3 9%) straddle estimates made in other systems (8%, Baron 1982 ; 7.4%, Kotanen 1995 ; 4%, Mitchell & Mayer 1997 ; 3%, Welander 2000 ; 3.8%, Engeman et al. 2003 ) However, the level of rooting in 2 below estimates in these other systems. It must be kept in mind that this rooting estimate is limited to the hammock areas and that rooting activity in the marsh is not included. However, the area of rooting in the marsh (i.e., the bulrush areas) can be as high as 18% in the landward portion (Figure 3 3 ) similar to the reported 19% rooted in a Florida freshwater marsh ( Engeman et al. 2004 ) and the marsh likely strongly supplements if not overshadows the food resources available in the hammock. The islands appear to be used less as feeding areas, even during the times the hammock habitat is heavily rooted. Instead, the islands seem to se rve more as staging areas for access to more profitable marsh feeding areas. The greater extent of rooting on the mainland than on the islands could have been due to greater herbaceous biomass there, providing more plant and possibly correlated invertebrat e resources available on the mainland than found on the islands. However, this difference was not consistent through time and rooting activity on the mainland
68 almost disappeared in 2006. Given the notorious behavioral flexibility of Feral Hogs ( Graves 1984 ; Oliver & Brisbin 1993 ) intense rooting on the mainland should be seen as an opportunistic feeding strategy either because the mainland only periodically provides worthwhile amounts of food or because it provides an important fallback feeding area when normal marsh resources fail. One result that we did not anticipate was that tilling appeared to foster further rooting by pigs. The observation that tilling increased the chances of a plot being subsequently rooted rai ses some interesting possibilities. First, tilling could simply break up the soil and thus make a softer and more attractive rooting substrate for the hogs. Second, tilling could be exposing food and making its detection by hogs more likely, encouraging them to root and feed in an area. Third, tilling could initiate an information cascade by being interpreted by the hogs as evidence of other hogs rooting and finding food in an area, encouraging still more rooting of the same area. These possibilities ar e not mutually exclusive. Although the second possibility depends on the artifact of food remaining in a tilled area (there would presumably be less food after real rooting), the positive feedback of hogs being attracted to previously rooted areas could h ave a strong influence on the spatial pattern of hog damage across a landscape, increasing its spatial variance and the spatial variability of effects on the communities. This positive feeding feedback may also be at the root of the bulrush dominated area s in the marsh that seem to be important foraging areas for the hogs. The impact of hogs specifically as ecological engineers would be an interesting focus for further study. Hog Effects on Herbaceous Communities In spite of the fact that there appear to be much greater hog densities in the island than mainland hammocks, effects of excluding the hogs on herbaceous species
69 richness and cover were only seen on the mainland with island plots showing no increase in either assay of plant community structure. T he decreases in species richness and cover due to the presence of hogs and increased tilling observed in the mainland areas were consistent with the known substantial impacts of hogs on herbaceous vegetation in other habitat types around the globe ( Bratton 1975 ; Cuevas et al. 2010 ; Hone 2002 ; Kotanen 1995 ; Nogueira Filho et al. 2009 ) as was the rapid recovery of species richness after tilling was ceased ( Kotanen 1995 ) However, the l ack of effect of exclosures and tilling on herbaceous species richness and cover on the islands was completely unexpected. We suggest the following scenario as a plausible explanation for this effect. This scenario involves the confluence of several factors, some reflecting the likely nature of hog responses to habitat spatial heterogeneity, and some to differences between island and mainland plant communities in this environment. Even though hog densities appear to be 3 to 4 times higher o n the islands, it was the mainland herbaceous community that responded to the exclusion of hogs. This seems consistent with the observation that rooting itself was more intense on the mainland at the beginning of the study, however the response to hog exc lusion occurred within a year, thus the impacts of higher hog rooting on the mainland early in the study would have disappeared in the control plots as the reduction in mainland rooting occurred long before the end of the study. Thus the final effect of t he exclosures would have been zero if it had been only the impact of hog rooting that was being excluded. Additionally, tilling had an effect on the mainland but not on the islands, the opposite of
70 what would be expected if the initial higher levels of r ooting on the mainland were driving the treatment effect of the exclosures on the herbaceous community. The unresponsiveness of the island herbaceous community arguably reflects how pigs, which have been present in this ecosystem for centuries, can act a s environmental filters for plant communities. If one is to observe an increase in species richness in a prey community after a predator or other consumer is removed, then species which are being strongly impacted must still be present in low numbers or s ufficiently close in space to provide dispersing propagules. If hogs differentially and strongly affect certain species, sustained higher hog densities on islands may have simply extirpated long ago disturbance vulnerable plants, thus precluding the obser vation of an effect of hog removal on island plant communities, at least over the time scale of this study. There are many examples in other systems of generalist predators eliminating vulnerable prey species from small islands ( Holt 2010 ) The islands in the study are obviously smaller in area than the mainland habitat, and so a give n intensity of suppression via any means total population sizes for affected species, and thus higher likely extinction rates. The fact that the island community is ess entially a subset of the mainland community and most species occurred with greater frequency on the mainland support this thought. Moreover, the spatial patchiness of hog disturbance means that in larger mainland areas, some plants may escape destruction by chance and provide foci for re colonization. The islands are less likely to harbor such refuges simply because of their constricted size. Thus, hog vulnerable species could potentially have populations on the mainland because either the land area is l arge enough to support stable populations
71 at low densities overall or there are permanent refuge areas of lower disturbance, or because areas lie undisturbed long enough as transient refuges for vulnerable plants to colonize, grow, reproduce, and colonize new patches before hogs use the area again. While the exclosures on the mainland could have been colonized by hog vulnerable plants that were able to survive either in low densities or in spatial or temporal refuges, the islands are too small for any of these mechanisms to function and thus we suggest all the vulnerable plants in the regional species pool were likely to have been extirpated from the islands long ago. The observation that the remaining island community was more resistant to impacts of Fer ( 1982 ) study that showed that Feral Hogs had no effect on the flora of an island that constantly experienced disturbance (in that case from inclement weather). However, in the current study, the mainland areas and the islands are extremely c lose and likely experience similar overall climatic physical disturbance regimes but the much higher densities of hogs on the islands increase biotic disturbance rates. Thus the hogs appear to have produced a disturbance resistant herbaceous community on the islands by restricting the species in the regional pool that are able to survive there. This idea is also supported by the results on herbaceous cover. Hog exclusion may have slightly increased the herbaceous cover in the mainland, but it had no imp act on the islands. Tilling did have an effect in both areas, but it was much greater on the mainland. Thus not only were the species richnesses of the islands affected less by both treatments, but the cover of the entire community was less impacted by t he treatments on the islands. These results again imply that the subset of herbaceous
72 species found on the islands is there largely because it is composed of plant species that survive disturbance by Feral Hogs. Oddly, however, exclusion of hogs had a pos itive effect of similar magnitude on the herbaceous biomass on both the islands and the mainland. This result shows that the hogs are having an impact on the herbaceous community viewed in aggregate. This result is not in contrast with the species richne ss and cover results, which suggest that the species found on the islands are the species better able to deal with the impacts of Feral Hogs. The marsh likely forms a barrier to dispersal for many plants making the island communities more dispersal limited than communities on areas of similar sizes on the mainland. Wind dispersed seeds that manage to escape a sheltered hammock environment and travel through the marsh are more likely to be stopped by the longer edge of the larger mainland hammock. Seeds di spersed by gravity or autochory are very unlikely to make it across the marsh barrier at all, as are seeds dispersed by small mammals. However, seed dispersal by birds or large mammals (especially by hogs) is unlikely to find the marsh a firm barrier. Th e fact that the hog exclosures were 4 m x 4 m makes them essentially very small targets and therefore unlikely destinations for rare propagules. So over the course of this study, long distance colonization ( e.g. from mainland sources) seems exceedingly un likely. This study demonstrates the importance that landscape patterns can have in determining the impact of a consumer with little top down control. In this case, the mechanisms involved appear to be landscape complementation and a neighborhood effect, i n which the spatial juxtaposition of distinct required resources alters the effective
73 density and movement of the actor and therefore its impacts on the system, and an emergent island effect, in which isolated and small local communities are vulnerable to species loss and have fewer opportunities for recolonization. Both of these mechanisms can be useful for prioritizing conservation responses to exotic species like Feral Hogs. Controlling Feral H og populations often proves difficult ( Cowled et al. 2006 ; Cruz et al. 2005 ; Hampton et al. 2004 b ; McIlroy & Gifford 1997 ) and is more likely to be economical and effective if their vital resources are understood and incorporated. For example, Feral Hogs are tied to water so moist areas not only have the highest dens ities of hogs ( Adkins & Harveson 2007 ; Baber & Coblentz 1986 ) but in dry areas each waterway may have a relatively independent po pulation making it more effective to concentrate elimination efforts on a single waterway at a time rather than in multiple areas simultaneously ( Hampton et al. 2004 b ) In the coast al hammock system, access to marsh resources appears to be important in maintaining high densities of hogs, so control efforts focusing on the marsh are likely to be more effective at reducing the hog population than would be diffusing efforts over a wider area. In addition to providing guideline strategies for more effective control of exotic species, an understanding of landscape patterns and how they influence the impacts of the exotic may be able to help focus efforts to conserve endangered native speci es by indicating which areas are most likely to be heavily impacted and, furthermore, which species in those areas are especially likely to be vulnerable ( Jolley et al. 2010 ) This is especially the case in recently invaded areas where impacts on the ecosystem are at an
74 early stage and species numbers and population levels have not yet been fully altered by the exotic species. The hogs have a clear top down impact on the herbaceous community of the hammock. Not only do they reduce the plant biomass on both the island and the mainland areas, they also were shown to have more complex interactions in that they reduced the species richness and cover in only the mainland hammock. They are likely to have had even greater impacts on the islands that are too long term to have been detected by this study. T esting this idea might require either much longer term or larger exclosures or directed colonization of species present on the mainland but absent on the islands. The lack of effective predation on the adults probably allows for much greater densities and impacts of hogs than would be seen if the only effective predator of hogs in Florida, the Florida Panther ( Maehr 1997 ) were still in the area. The dominance of Feral Hog in the diet of the Florida Panther, the high densities of hogs across much of Florida, and the importance of top down impacts in this community suggest an avenue through which the extirpation of a top predator is having continued influence on co mmunity characteristics.
75 CHAPTER 4 ALTERATION OF NUTRIE NT FLOWS BY AN EXOTIC SPECIES Introduction Patterns of nutrient and energy flow within ecosystems have long been considered to be central to ecology and the functioning of ecosystems (Evans 1956). T he importance of anima l mediated nutrient flows between different ecosystems and habitats is becoming increasingly recognized in many different contexts (Polis et al. 1997, Vanni et al. 2004). These n utrient flows can be disrupted by population reductions and extinctions of so me nutrient transporters (Sekerg ioglu 2006) or by the increase in importance of others, including exotic species. Changes in patterns of nutrient flow can alter competitive interactions within communities (Keddy et al. 2000), vulnera bility to exotics ( Thompson & Leishman 2005 ) and even the physical structure of communities (Croll et al. 2005) making conservation more challenging These changes may be further exac erbated by anthropogenic habitat reduction, degradation, and fragmentation and the introduction of exotic species Human actions often benefit medium and small sized mammalian generalists (DeStefano & DeGraaf 2003, Fraterrigo & Wiens 2005, Prange et al. 2004 see Chapter 2) due to declines in large predators (Gipson 1978), reduction s in specialist competitors (Foufopoulos & Ives 1999, Henle et al. 2004), increases in landscape diversity (Marzluff & Neatherlin 2006), appearance s of novel resources (Marzluf f & Neatherlin 2006), and introductions to new areas (Ruesink 2005). All these changes typify human landscape interactions. Because generalists tend to increase in human impacted landscapes, their importance to nutrient flows is probably increasing. Thi s could amplify nutrient ties
76 between systems as generalist omnivores can use multiple habitat types for different resources In many anthropogenic landscapes, the existence of refuges is needed for continued survival of larger vertebrates (Livingston et al. 2002). Refuges serve as areas to avoid persecution or disturbance, breeding habitat or simply provide space for den sites (Tigas et al. 2002 ). Because foraging may involve ranging to different habitats when resources are found concentrated in non r efuge areas (agricultural lands, middens, etc. ) nutrients sequestered from feeding habitats are likely to be concentrated in refuges. Nutrient transport can be important over both large and small distances (Durbin et al 1979), for instance, when the nu trients are retrieved from a sink (Vanni & Headworth 2004) or transported from resource rich zones (Johnston & Naiman 1987), which are characteristic of human impacted areas (Prange et al. 2004). However most previous work has focused on stable, usually natural, conditions (Vanni et al. 2004) or on the extinction of an important nutrient transporter (Croll et al. 2005). Examinations of nutrient flow patterns under conditions of species additions or increases are sparse. Because global anthropogenic chan ges favor generalist omnivores, their roles in mediating nutrient flow patterns need to be explored (Chapter 2) In this study, the nutrient transport function of a globally important generalist omnivore, Feral Hogs ( Olive r & Brisbin 1993 ) is examined in a system that has been shaped by human activities largely though the introduction of exotics (specifically hogs) and the elimination of effective predators ( Maehr 1997 ) This system, the coastal marsh/hammock complex in Florida is an intermingling of refuge areas for hogs (the
77 hammocks) and feeding areas (the marsh) (Chapter 3) This provides a n opportunity to examine how Feral Hogs shape the n utrient regime of a complex landscape Feral Hogs make excellent subjects for a study of nutrient transfer on a landscape level because of their relatively large size, omnivorous diet, high population densities, flexible foraging strategies ( Mayer & Brisbin 1991 ) and high fecundity and growth rates ( Oliver & Brisbin 1993 ) The effects of feral pigs have bee n studied in many different ecosystems in numerous countries ( Anderson & Stone 1993 ; Baron 1982 ; Cruz et al. 2005 ; Engeman et al. 2003 ; Ilse & Hellgren 1995 ; Kotane n 1995 ; McIlroy & Gifford 1997 ; Singer et al. 1984 ) However, most studies on environmental impacts of pigs have been limited to measuring rooting damage, erosion, and the direct effects of feeding and other activities on native communities ( e.g. Bratton 1975 ; Ickes et al. 2001 ) while very few examine their impact on ecosystem attributes such as nutrient regimes ( but see Mohr et al. 2005 ) I here use Feral H ogs as a model organism to investigate t he characteristics and extent of nutrient flow mediated by a generalist omnivore in a human altered system. I hypothesize d that a source of the hog transported nutrients is found in the marsh and that inter habitat nutrient transfer will be unbalanced wit h a net flow of nutrients into refuge areas. First, the source of the nutrients will be established through fecal dietary analysis and verified through stable isotope analysis. Then, the relative nutrient deposition in different refuge areas will be esti mated through fecal count surveys and estimates of absolute nutrient deposition rates will be made based on fecal nutrient analysis and estimated fecal production rates of Feral Hog populations. The impact of nutrients transported to the refuge areas will be examined though soil and plant tissue
78 nutrient levels inside and outside of hog exclosures Finally, t he potential importance of long distance transport in this system will be examined through the use of stable isotopes and genetic techniques to explo re population structure and size and the potential for movement of individuals between the coast and interior Methods Study site The Lower Suwannee National Wildlife Refuge (LSNWR; 214 km 2 ) is on the west and is c entered on the mouth of the Suwannee River. Inland, t he habitat is composed primarily o f planted upland pine woodlands and secondary growth lowland hardwood swamp s. The study site is located in a portion of the refuge south of the river on the coast (Fig ure 4 1). The coastline is rimmed by tidal marsh a slightly brackish marsh (1 2 ppt unpub. data ) fed by the Suwannee River and California Swamp. The marsh is dominated by the fresh /brackish water S aw grass ( Cladium jamaicense ) and B lack N eedle R ush ( Ju ncus roemerianus ). A somewhat salt tolerant species J. roemerianus extends more seaward than C. jamaicense which lines creeks and hugs the coastal hammocks The American Bulrush ( Schoenoplectus ( Scirpus ) americanus ) is found in areas that are less de nsely vegetated and have a disturbed looser substrate. At th e complex boundary between the coastal hammock and the marsh these three species form a mosaic of patches dominated by a single species. There are also a few small (~ 10 m 2 ) evaporation pools b ordering coastal hammocks that are dominated by Spike Grass ( Distichlis spicata ). The coastal hammock bordering the tidal marsh is characterized by E astern R ed C edar ( Juniperus virginiania ) and C abbage P alm ( Sabal palmetto ), with oaks ( Quercus sp. ) and pla nted native pines ( Pinus spp. ) also being common The subcanopy is
79 dominated by Y aupon H olly ( Ilex vomitoria ), bay ( Persea spp. ) and, in the drier areas, S aw P almetto ( Serenoa repens ). In addition to the mainland, the coastal hammock is found on small (~ 1000 12,000 m 2 ) forested islands embedded in the tidal marsh. The amount of herbaceous undergrowth varies greatly from essentially none in the wettest areas to high cover in the areas least subjected to flooding. Feral Hogs are a conspicuous element in t he local mammal community. The Nine banded Armadillo ( Dasypus novemcinctus ), Virginia Opossum ( Didelphis virginiana ), and Raccoon ( Procyon lotor ) are common in the coastal hammock including the islands. Less common are the Coyote ( Canis latrans ) and Bobc at ( Felis rufus ), the latter of which travels through the marsh to the coastal hammock islands. White tailed Deer ( Odocoileus virginianus ) are rare but present throughout the coastal hammock including the islands. Feral Hogs ( Sus scrofa ), however, are ex tremely common (pers. obs.; Mayer & Brisbin, 1991) The s ampling area of the study contained of three coastal hammock areas of the mainland and 8 10 islands on either side of Barnett Creek. All of these areas were within 1.25 km of each other (Figure 4 1) Study Design Fecal/tissue samples Fresh (< 1 week old, usually < 3 days old) hog fecal samples were gathered either opportunistically throughout the study or specifically sought by intensively searching areas with fresh hog sign or by stalking and backtr acking observed animals. Fecal samples were stored in 95% ethanol at 40 C. Each sample consisted of a single fecal
80 or road killed animals and stored at 40 C. Fec al samples were used to for dietary, nutrient, and genetic information (below). Figure 4 1. An overview of the study site in the Lower Suwannee National Wildlife locations for this stud y are indicated. Locations of plot sets are indicated by squares.
81 Diet Fecal dietary analysis: Fecal samples were washed over a 2 mm sieve and the retained material identified under a dissecting microscope. If necessary, identification was assisted by co mparison with ground or whole specimens from known food species. Each sample was systematically examined and all components recorded. Percent composition was estimated with a point sampling method by spreading the samples evenly over a 5 x 5 grid of poin ts and recording the material over each point. If more than one component occurred over the same point, the score for that point was divided that appeared to have been d ead when consumed and was likely ingested incidentally during feeding. Examples included pine and juniper needles, oak leaves, palm wood, palm leaves, and twigs. A food was considered to have been intentionally consumed if the amount in a sample exceeded the maximum amount of detritus found in all the samples. Unmasticated maggots, dung beetles, and fungus were ignored as obvious post excretion additions. The only fraction that proved impossible to identify was masses of white slimy mucus in one specimen Stable isotopes: Bone samples of hogs and other omnivores (Virginia Opossum, Didelphis virginiana ; Raccoon, Procyon lotor ; Marsh Rice Rat, Oryzomys palustris ; and Hispid Cotton Rat Sigmodon hispidus ), strict herbivores (White tailed Deer Odocoileus vir ginianus and Domestic Cattle Bos primigenius taurus ) and carnivores (American Alligator Alligator mississippiensis and River Otter Lontra canadensis ) were gathered opportunistically during the study to clarify the trophic position of hogs in the coas tal food web ( Minagawa et al. 2005 ) using isotopic analysis Though no cattle are currently found in the LSNWR, they were previously run in the area and bone specimens of
82 young animal s were infrequently f ound. Hog skeletal remains were common. H owever other species proved to be rare and sample sizes are therefore limited For comparison with coastal hogs, samples of inland hogs were also opportunistically gathered or were taken fr om museum specimens kept in the Florida Museum of Natural History (catalogue numbers UF2048, UF3282, and UF5975). The two fractions of bone typically used in stable isotope analyses are the mineral fraction, apatite, and the protein fraction, collagen. T he collagen fraction contains both nitrogen and carbon. Stable isotope values are reported in the notation ( 15 N or 13 C). For example, 15 N sample = ((( 15 N/ 14 N) sample /( 15 N/ 14 N) standard ) 1) x 1000, and thus the units are per mil ( ). The standard for nitrogen is atmospheric nitrogen. The standard for carbon is Peedee Belemite limestone (Kelley 20 00). The nitrogen reflects the protein in the diet that is retained and incorporated into body proteins. However, as some dietary proteins are burned for energy and this process selectively metabolizes the lighter isotopes, the nitrogen signal of the coll agen is enriched in the heavier isotopes compared to the signal of the diet (this is generally 2 5 ; hog specific fractionation is discussed below) (Ambrose et al. 2003) The collagen carbon is largely derived from the carbon in the protein source, both because essential amino acids by definition retain their original composition and because it is more metabolically efficient to use non essential amino acids from the diet when possible (Ambrose et al. 2003) As discussed below, fractionation of carbon be tween trophic levels is relatively small but there is fractionation among different tissue types of the same individual that must be accounted for.
83 The apatite fraction contains carbon containing carbonate derived from carbon ic acid in the blood. This car bonic acid is in turn derived from metabolizing organic molecules to release energy and CO 2 which dissolves in the blood forming the carbonic acid. Thus the carbon signal of the apatite fraction of the bone reflects foods that are metabolized for energy. In an omnivore, carbohydrates are often the main energy source but amino acids that are not incorporated into proteinaceous ti ssues are also metabolized for energy. Therefore, the carbon signals in collagen and apatite often tell different stories with the collagen signal reflecting protein sources in the diet (meat) and the apatite signal reflecting the whole diet (Ambrose et al. 2003) The differences between the two signals should be the greatest in individuals who have moderate amounts of meat in th eir diet and the differences should be less when individuals have very little (so are getting most of their protein from the same plant foods as their carbohydrates) or an over abundance of meat (so are burning a great deal of protein as energy). C ollagen and apatite fractions of the bone samples were isolated with standard techniques ( Ambrose 1993 ) in the Bone Chemist ry Laboratory in the University of Florida Department of Anthropology Bone samples were scraped clean of visible dirt and sonicated in deionized distilled water (DI H 2 O) in 10 minute treatments until the water was no longer cloudy. L ipids were removed u nder pressure with Accelerated Solvent Extraction (model ASE300; 2:1 dichloromethanol:methanol; two cycles of heating for 5 min to 100C, static for 5 min, 60% volume flushed, and purged for 100 sec), and the bone was ground into powder with a mortar and p estle or Spex mill The
84 resulting powder was sieve partitioned into two size categories: particles 0.50 mm to 0.25 mm in size (for collagen) and particles < 0.25 mm in size (for apatite) Apatite was purified from the < 0.25 mm fraction by treatment in a 2% NaOCl solution for 16 hours to oxidize the sample and remove the organic portion of the bone. The samples were then rinsed with distilled water to neutrality and treated with a 0.2 M solution of acetic acid for 16 hours (to remove any diagenic carbonat es). The samples were rinsed to neutrality and l yophi lized. Collagen was purified from the 0.25 0.50 mm fraction by treatment in 0.2 M HCl (replaced every 18 hours) to demineralize the bone until purified collagen was visually apparent (collagen is lowe r density than bone and sinks slowly after agitation). The samples were then rinsed to neutrality with distilled water and treated with 0.0625 M NaOH for 16 hours. The samples were again rinsed to neutrality, placed in 10 ml of 10 3 M HCl, and heated at 95 C for 5 hours to dissolve the collagen. To completely dissolve the collagen, 35 l 1 M HCl were added to the samples, which were heated at 95 C for another 5 hours. The evaporated liquid was replaced with 10 3 M HCl. The solution was separated from the solid fraction with centrifugation and evaporated at 65 C to ~2 ml before being ly ophi lized. In addition to vertebrate bones, the isotopic signatures of putative foods/basal resources were also measured. Plant samples were gathered from the marsh, t he islands, the coastal hammock mainland, and from three inland locations in the refuge. The samples were sonicated in distilled water until the water no longer became cloudy. They were then dried at 50 C and ground in a Wiley mill. Invertebrate sample s were removed from the shell and dried at 50 C, before being ground in a Wiley mill. Gastric
85 contents were retained as they would be consumed by the hogs along with the invertebrate tissue. In 7 out of 11 cases, the invertebrate samples comprised one i ndividual. In the other cases, the samples contained more than one individual. The samples were weighted accordingly during statistical analysis. Invertebrates were divided into two groups based on diet and stable isotope signatures: omnivores/detritivo res (the Carolina Marsh Clam, Polymesoda caroliniana ; Gulf Mud Fiddler Crab, Uca virens ; Heavy Marsh Crab, Sesarma reticulata ; and the Olive Nerite, Vitta usnea syn. Neritina reclivata ) and herbivores (Florida Marsh Snail, Detracia floridana ) Stable isoto 15 13 C) were measured in the Light Stable Isotope Lab oratory in the Department of Geolog ical Sciences at the University of Florida. Plant and invertebrate samples were measured on a Finnigan MAT 252 isotope ratio mass spectrometer w ith a ConFlow II interface linked to an elemental analyzer. The results were corrected against the USGS40 (L glutamic acid) standard (n 13 C 15 N. Two samples (a Sm i lax and a bulrush) had odd chromato grams and 15 N values were enriched by more than an order of magnitude over the other samples These were rerun, with similar but numerically inconsistent results so they were considered contaminated and not used. The apatite fraction of the bone sampl es was measured on the Finnigan MAT 252 with a Keil III carbonate preparation system. The results were corrected against the NBS The collagen fraction of the bone samples was mea sured by first determining the C:N atomic ratio of each sample with a Carlo Erba NA1500 CNS elemental analyzer and
86 corrected against an atropine standard (70.56% C, 4.84% N, n = 6) with measurement precisions of 0.07% N and 0.75% C. All bone collagen sa mples had acceptable C:N ratios (2.9 3.6) ( DeNiro 1985 ) S table isotope ratio s were measured on the Finnigan MAT 252 with the ConFlow II interface and corrected with the USGS40 standard (n = 9) 13 C and 0.11 15 N. 13 13 C values produced by differing methods of photosynthesis (average for C 3 = 4 = ( Kelly 2000 ) However there is a fractionation of 13 C among the tissues within an animal. The isotopic signature of diet of the hogs was calculated by accounting for the average carbon fractionation between whole hog diets and hog apatite (average = 13 Howland et al. 2003 ) and the average nitrogen fractionation between hog diets and hog collagen (average = 2.2 15 Hare et al. 19 91 ) Apatite carbon signatures were used to estimate the signature of the hog diet because the fractionation from diet to bone collagen is more variable (CV = 72.4%) than that of the apatite (CV = 12.7%) ( Howland et al. 2003 ) Because 15 N is enriched with increasing trophic level, suckling young are often enriched compared to their mothers when the females are catabolizing tissues to produce milk ( Newsome et al. 2006 ; Polischuk et al. 2001 ; Sare et al. 2005 ) although this does not occur in all species or situations ( Jenkins et al. 2001 ) Thi s phenomenon 15 N values of the coastal hogs 15 N values. Specimens that included tooth rows were assigned a relative ranking based on t ooth eruption. Adults with completely erupted third molars were ranked based on tooth wear.
87 Interior hogs were not included in this analysis because basal resource 15 N values in different areas can vary dramatically ( Ambrose 1993 ; Kelly 2000 ; Peterson & Fry 1987 ) The amount of C 4 derived material in the d iet of the hogs was estimated using a standard linear mixing model ( Phillips & Gregg 2003 ) 13 C signature of bulrush and the average 13 C signature of C 4 plants from this study for the coastal h ogs and the average 13 C signatures of C 3 and C 4 plants from this study for the inland hogs Because feeding maize to wildlife for hunting purposes appears to be common (pers. obs.), t he potential contribution of maize to the hog diet was estimated in th e same way but with the C 4 signature replaced by 13 C signature from Hare et al. ( 1991 ) which is intermediate between and similar to values given in Norr ( 2004 ) and Howland et al. ( 2003 ) Nutrient deposition The standing crop of Feral H og feces in the coastal hammock was estimated by counting the number of piles of feces in 2 m x 50 m transects at the end of each season when they were within 0.5 m of each other and appeared to be the same age and similar in size and composition. Due to the non normal distribution of the data, the number of fecal piles for the mainland and islands were compared using a Wilcoxon sum rank for each season. Subsamples of the hog fecal samples were sent to the Institute for Food and Agricultural Sciences Analytical Research Laboratory (IFAS ARL) at the University of Florida for manure analysis of potassium, phosphorus, and total Kjeldahl nitrogen. For K and P, the samples were combusted at 550 C for 4 hours and the ash digested in HCl. The resulting solution was diluted with distilled water and the K and P levels
88 measured with an inductively coupled plasma emission spectrometer ( Peters et al. 2003 ) For total Kjeldahl nitrogen, samples were digested overnight at room temperatur e in an H 3 SeO 3 /H 2 SO 4 digestion mixture and then further digested for 80 minutes at 370 400C. The resulting solution was measured using an Alpkem colorimetric autoanalyzer ( Peters et al. 2003 ) Estimated fecal deposition rates were estimated by combining hog densities and body masses from similar systems (27 pigs/km 2 ; 44.9 kg; Saunders & McLeod 1999 ) and solid waste production of Domestic Hogs ( Chastain et al. 1999 ) which resulted in an estimated deposition rate of 800 340 SD kg of feces/km 2 /year (see also Appendix B) This was adjusted with the fecal count relative abundances of the mainland and islands and the measured fecal nutrient content from this study to estimate the nutrient deposition rate from feces in the different areas. Because nutrient deposition occurs through urine in addition to feces, N, P, and K deposition through urine was estimated based on calculated fecal nutrient deposition and the proportions of N, P, and K in Domestic Hog nutrient budgets excreted in the feces and in urine ( Kephart & Sherritt 1990 ; Sands et al. 2001 ) Soil and vegetation nutrients Exclosures: 30 sets of 4 m x 4 m plots were set up: 15 sets on the mainland and 15 on the islands. Five plot sets were placed in each of the 3 mainland areas. The number of plot sets on the islands varied from 1 3 based on the size of the islands. The plot sets were placed t o maximize the diversity of the herbaceous community sampled while spacing out the plot sets as much as possible. Each set consisted of three plots: a control, an exclosure, and a tilled plot (to mimic increased hog rooting). Once the 4 m x 4 m exclosure
89 the sides of each exclosure that most resembled the exclosure in herbaceous community were chosen for the other two plots and randomly assigned to either till treatment or control. The external plots were given a 1 m buffer zone from the exclosure. Tilled plots were lightly tilled every six months to a depth of ~5 10 cm to mimic the physical disturbance of hog rooting. The presence of fresh hog rooting in the external plots was no ted opportunistically throughout the study. Nutrient analysis: Th ree sub samples of soil were taken from near the center of each plot every six months (3 months offset from the tilling of the tilled plots) from summer 2007 to spring 2009 and one final time in the winter of 2010 2011. The samples were dried, sifted twice through 2 mm screen, and sent to IFAS ARL for analysis of nitrate/nitrite, potassium, and phosphorus concentrations. These analyses were done as nitrogen, potassium, and phosphorus are imp ortant macronutrients and often limiting for plant growth. Nitrate and nitrite are the forms of nitrogen most readily available for plant uptake. Samples used for NO x analysis were water extracted as nitrate and nitrite are very water soluble. Because P and K are not as water soluble, samples for potassium and phosphorus analysis were extracted with the Merlich 1 technique, the standard for non alkaline Florida soils ( Mylavarapu & Kennelley 2002 ) Merlich 1 extraction uses a mixture of hydrochloric and sulfuric acids to extract soil nutrients. The nutrient concentration of the resulting solution is analyzed with spe ctrophotometry and, in the case of phosphorus, colorimetry ( Mylavarapu & Kennelley 2002 ) In the spring of 2011, p lant samples were taken to estimate the nutrient content of the living aboveground biomass of the herbaceous community in the plots. For plants
90 with woody stems ( e.g. Smilax spp. and Rubus spp. ) only the leaves were taken. Poison ivy ( Rhus radicans ) w as not sampled at all. Though somewhat widely spread, Smilax and R. radicans especially, these genera were always at very low density and would have added little to the overall biomass. For plots in which the biomass was sparse, the entire aboveground l iving herbaceous biomass was taken. In more thickly populated plots, a subsample of the overlapping randomly placed 1 m x 1 m quadrats until the 1 liter sample bag had been filled. The samples were then dried at 50 C, g round, and sent to IFAS ARL at the University of Florida for potassium, phosphorus, and total Kjeldahl nitrogen analyses. Potassium and phosphorus were determined by combusting the samples at 500C followed by digestion with 6M HCl. After dilution, the n utrient concentrations were measured with an Inductively Coupled Plasma Spectrometer ( Mylavarapu & Moon 2007 ) Total Kjeldahl nitrogen was measured by combining the samples with Kjeldahl digestion mixture (94% K 2 SO 4 3% CuSO 4 and 3% TiO 2 ) and H 2 SO 4 The samples we re then digested at 250C for 1 hour followed by digesting at 365C for 2.5 3 hours. The solutions were diluted with distilled water and the TKN was measured with an air segmented continuous flow Alpkem colorimetric autoanalyzer ( Mylavarapu & Moon 2007 ) Population structure Coat coloration : The coloration and location of hogs observed were recorded opportunistically throughout the study. The color categories were based on the genetics of swine coloration with each category representing a distinct genotype ( Fang et al. 2009 ; Kijas et al. 1998 ; Pielberg et al. 2002 ) and on field distinguishability. The exception was white belted, which, although genetically distinct from white spotted
91 ( Kijas et al. 1998 ) was lumped with white spotted due to its low frequency and the possibility of mistaking white spotted with belted when the hog is only partially or momentarily visible. Hog sightings were put into three categories: coastal, refuge, and inland. The refuge category included hogs seen in the refuge or along the boundaries, except those seen in the coastal hammock. The inland category included hogs seen farther inland than the refuge. During statistical analysis, categories with no observations we sizes in some categories ( Yates 1934 ) Genetics: DNA was extracted from fecal samples with a Qiagen QIAamp DNA Stool Mini Kit and from muscle samples with a Qiagen QIAamp DNA Tissue Mini Kit. The concentration of extracted DNA was determined with a Nanodrop ND 1000 Spectrophotometer. That hog DNA was present in the extractions was verified by amplifying an S. scrofa specific mitochondrial D loop marker (SsD; Haunshi et al. 2009 ) and running a gel. Because SsD was long (835 bp) compared to the microsatelli tes used, it was possible that the DNA from a sample had degraded into lengths shorter than SsD but still long enough to maintain the microsatellite information. Therefore, samples with relatively high concentrations of DNA (greater than ~20 l/ml) where checked against the longest microsatellite locus if SsD could not be amplified. Fourteen hog dinucleotide repeat microsatellite loci (Table 4 1) were amplified for each sample with successful DNA extraction. These loci are well characterized and informati ve in S. scrofa ( Lowden et al. 2002 ; Martinez et al. 2000 ; Vernesi et al. 2003 ) and have been useful in Feral H og population genetics in other parts of the world ( Cowled et al. 2006 ; Hampton et al. 2004b ; Spencer et al. 2005 )
92 CAC GAC GTT GTA AAA CGA C a fluorescent dye. Two of the 14 microsatellites were never found to work, even on the tissue samples and were therefore dropped from use during the study. For multiplexing, microsatellites were assigned to groups of 4 or less based on non overlapping f ragment lengths. Each microsatellite in a group was assigned one of four fluorescent dyes: 6 the fragments with an M13 universal primer ( Schuelke 2000 ) PCR reactions contained 2.5 l bovine serum albumin, 12.5 l GoTaq colorless Master Mix, 0.1 l forward primer, 1 l reverse primer, 1 l fluorescent dye attached to M13 universal primer, and 50 ng DNA in a final volume of 25 l. Reaction cycles consisted of 94C for 3 min for one cycle; 95C for 40 s, XC for 40 s, and 72C for 40 s for 10 cycles; 95C for 40s, 48C for 40 s, and 72C for 40 s for 30 cycles; and 72C for nts the specific temperature for each primer (Table 4 1). The second set of cycles allows the universal primer with dye to attach and replicate into the microsatellite copies. An adjustment to this format was required by the primers with a 58C annealing temperature, which needed a 50C annealing temperature in the second cycle to function properly. The microsatellite PCR products from fecal samples were diluted to two different concentrations for fragment analysis: 50x and 100x. Those from tissue sample s were diluted 100x and 200x. The samples were then sent to the Interdisciplinary Center for Biotechnology Research at the University of Florida where the fragment lengths were th
93 the computer program GeneMarker and alleles based on microsatellite length were identified for each sample. Samples with fewer than 9 successfully characterized loci were excluded from further analysis. Table 4 1. The microsatellite loci used in this study. Name Temp n A n E H O H E Success Length S0002 63 12 7.2 0.692 0.877 26 (33) 211 246 S0005 59 11 8.3 0.636 0.922 11 (12) 238 260 S0026 55 10 4.5 0.651 0.760 27 (43) 111 123 S0068 62 0 S0090 55 5 3.6 0.600 0.800 5 (5) 260 267 S0155 55 6 2.8 0.556 0.655 27 (32) 166 180 S0226 55 7 2.6 0.423 0.624 26 (43) 194 225 SW122 58 0 SW240 58 4 2.9 0.333 0.681 12 (12) 110 136 SW632 58 9 4.2 0.667 0.775 27 (30) 163 194 SW857 58 9 2.6 0.593 0.632 27 (46) 143 173 SW911 56 6 3.6 0.444 0.73 4 27 (42) 172 184 SW936 58 9 4.5 1.000 0.792 27 (46) 109 131 SW951 58 8 3.2 0.704 0.700 27 (49) 137 152 Temp = annealing temperature (C) n A = actual number of alleles n E = effective number of alleles H O = observed heterozygosity H E = unbiased estima te of heterozygosity ( number of individuals with that locus characterized out of the original 49 successful DNA extractions) = observed ra nge in base pair number The allele data were formatted, summarized, and analyzed with the application GenAlEx 6.1 ( Peakall & Smouse 2006 ) and the program G ENEPOP 4.0.10 ( Raymond & Rousset 1995 ; Rousset 2008 ) and the population structure analyzed with a Markov chain Monte Carlo algorithm in the program S TRUCTURE 2.3.3 ( Falush et al. 2007 ; Hubisz et al. 2009 ; Pritchard et al. 2000 ) S TRUCTURE was run with a 50,000 burn in period followed by 1,000,000 MCMC repetitions in 10 runs per putative K ( i.e. the number of populations) ( Cowled et al. 2006 ; Spencer et al. 2005 ) The simulations ran from K = 1 to 5 based on the possibility that the 4 general geographic areas from which samples wer e drawn (Barnett Creek, Trout Creek, Salt Creek, and Levy County Road
94 347) could be one to several distinct populations depending on (unknown) patterns of hog movement The computer program S TRUCTURE H ARVESTER ( Earl & vonHoldt 2011 ) ( Evanno et al. 2005 ) Effective population size was estimated using the formula N e = (n E 1)/4; where N e is the effective population size, n E is the effective number of alleles, and is the mutation rate ( Frankham 1996 ; Kimura & Crow 1964 ) A n assumed mutation rate of 1.0 x 10 3 was used following Hampton et al. ( 2004b ) who used the same loci as this study. Actual population size was estimated based on an average N e /N ratio of 0.11 in wild populations ( Frankham 1995 1996 ; McKay 2007 ) Results Diet Of the 19 substances identified from 39 fecal s amples only 9 were considered selected foods (i.e., appeared at least once as > 12% of any sample, the maximum amount of detritus found in a single sample) (Table 4 2, Figure 4 2). The diet of the coastal hogs in the Lower Suwannee National Wildlife Refug e was surprisingly specialized 81% 5% (mean SE) American Bulrush rhizome based on fecal analysis. The next highest dietary fraction was Cabbage Palm fruit/seeds but only at 4% 3%. Of the total number of fecal samples examined, 90% contained some amou nt of American Bulrush rhizome and 36% contained Cabbage Palm fruit/seeds. American Bulrush rhizome was the most frequently selected food and was a selected food in 90% of the samples. The next most frequently selected food, palm fruit/seeds, was a selec ted food in 10% of the samples (Figure 4 2).
95 Table 4 2. Coastal Feral Hog diet as estimated through fecal analysis. Diet component Common name Scientific name T ype % Diet Bulrush Rhizome American Bulrush Schoenoplectus ( Scirpus ) americanus S 81.1 Palm F ruit/Seed Cabbage Palm Sabal palmetto S 4.3 Acorn Oak Quercus spp. S 2.5 Saw grass Saw grass Cladium jamaicense S 2.5 Crab Shell various various S 2.2 Juniper Berry Eastern Red Cedar Juniperus virginiana S 2.2 Smilax Rhizome Greenbriar Smilax spp. S 2 .2 Detritus N/A N/A I 0.8 Grass N/A N/A S 0.7 Insect N/A N/A I 0.6 Snail Shell Florida Marsh Snail Olive Nerite Detracia floridana Vitta usnea ( Neritina reclivata ) I* 0.3 White Slime ? ? I 0.3 Hair rodent and canine N/A I* 0.2 Mussel Shell Atlantic Ribbed Mussel Geukensia demissa I 0.2 Clam Shell Carolina Marsh Clam Polymesoda caroliniana I T Bulrush Shoot American Bulrush Schoenoplectus ( Scripus ) americanus I T Sugarberry (seed) Sugarberry Celtis laevigata I T Herb Leaf ? ? I T S = Selected fo od ( found as >12% in a single sample, the highest level attained by detritus) I = Incidental food (always < 12%) T = Trace present but not detected in the point sampling The stable isotope data were consistent with the fecal data. The isotopic signature of the coastal hog diet matched that of bulrush rhizomes with no deviation caused by a significant dietary fraction missed in the fecal analysis (Figure 4 3). The carbon sig nals in the apatite and collagen fractions were highly correlated (r s = 0.86, p << 0.001; Figure 4 4) with no indication of a source of protein different from the major source of energy, consistent with a diet dominated by a single food. There was a signi ficant negative 15 N (r s = 0.7, one tailed p = 0.018, n = 9; Figure 4 15 N values is due to an enrichment from nursing and the subsequent life long remodeling of bone.
96 Figure 4 2 Diet of coastal Feral Hogs in the Lower Suwannee National Wildlife Refuge, Florida based on analysis of fecal contents. As 12% was the highest amount of detritus found in any sample, foods found at greater amounts were considered to have been selected for during feeding. Standard errors are indicated. The stable isotope signatures of coastal hogs (n = 12) are similar to the signatures of hogs from inland Florida (n = 10) for 15 N (t = 0.3, two tailed p = 0.79) but distinct for 13 C (apatite: t = 4.3, two tailed p = 0.002; collagen: Wilcoxon sum rank S = 173, p < 0.001) (Figure 4 6). The inland hogs that were the closest to the coast had among the most divergent signatures, im plying little movement perpendicular to the coast by 13 C signatures imply that the diet of the coastal hogs averages 0% C 4 plants (SD = 7%, range: 10 to 14%) and the diet of the inland hogs averages 57% C 4 plants (SD = 38%, range: 7 to 101%) or 46% maize (SD = 31%, range: 6 to 83%).
97 Figure 4 3. Stable isotope signatures of potential food resources and the diet of coastal Feral Hogs in the Lower Suwannee National Wildlife Refuge, Florida. Standard deviations are indicated. Compared to other vertebrates, Feral Hogs appear to be slightly more enriched in 15 N than herbivores (deer and cattle) and fairly depleted compared to pure carnivores (the otter and alligator) and even other omnivores (rats, raccoons, and opossums) (Figure 4 7 ). Thi s seems at first glance to indicate an intermediate trophic position; however, it is more likely due to the enriched 15 N values of bulrush (the primary hog food) compared to some of the other plants, including the grasses. As expected, the browser herbiv ore (deer) seems to have an entirely C 3 derived diet (C 4 derived average = 4%, SD = 5%, n = 3) while the grazer herbivore (cattle) has a large dietary fraction made from C 4 grasses (average = 66%, SD = 25%, n = 3).
98 Figure 4 4. A comparison of the stabl e carbon isotope signatures from bone apatite and bone collagen for the individual hogs used in this study. The two measures are highly correlated indicating that differing sources of protein and energy are not detected. There was a significant pattern o f 15 N depletion in vegetation with increasing distance from the ocean (Figure s 4 8 and 4 9 ) (n = 34, r s = 0.79, p << 0.001) after 15 N values of the different species. Nutrient deposition Fecal counts were significantly higher on the islands than on the mainland during each sampling period (summer 2005, fall 2005, winter 2005 6, and spring 2006; Wilcoxon sum rank tests, all p << 0.001) (Figure 4 10 ). Overall, the island fecal load averaged 7.9 fecal piles/100 m 2 0.0004 SE and the mainland fecal load averaged 1.8 fecal piles/100 m 2 0.0002 SE; a fecal load 4.4 times higher on the islands than the
99 mainland. Thus the islands are likely receiving 81% of the nutrients being deposited by the hogs in the hammock and the mainland is rece iving 19%. Figure 4 5 The relationship between relative age rank based on tooth eruption and 15 N in the Lower Suwannee National Wildlife 15 N signal from suckling that diminishes gradually with age (rs = 0.7, one tailed p = 0.018). The average N, P, and K content of hog feces, and estimated nutrient deposition rates for the islands and coastal mainland are given in Table 4 3. Soil and Vegetation Nutrients The pre treatment levels of nitrate/nitrite nitrogen, phosphorus, and potassium were significantly higher in the coastal hammock islands in the marsh than on the mainland (Wilcoxon sum rank: NO x N: S = 1714, p = 0.007; P: S = 1650, p = 0.001; K: S
100 Figure 4 6 Stable isotope signatures of coastal Feral Hog and inland Feral Hog individuals. A 15 13 C signature from apatite. B 15 13 hogs were from Levy County Road 3 47 that forms an eastern border of the refuge, ~3 Gilcrist hogs were from ~ 90 km from the coast, although the provenance of the more 13 C depleted specimen was not specified beyond the county level. Note that the inland hogs geograp hically closest to the coastal hogs have isotopic signatures that are among the furthest from those of the coastal hogs. A B
101 Figure 4 7 Stable isotope signatures of Feral Hogs compared to those of other vertebrates. A 15 N signature from collagen a 13 C signature from apatite. B 15 13 C signature from collagen. All samples are from the coastal hammock or marsh except for the inland hogs and the otter specimen, which was from CR 347 on the eastern border of the refuge, 9 km from the study site and 0.5 km from the Suwannee River. The hog data are collapsed from Figure 4 6 Standard deviations are indicated. A B
102 Figure 4 8 Depletion of 15 N with distance from the ocean across different species of plants. The distance from the ocean is ~4.2 km for the islands, ~4.8 km for the mainland, ~5.1 for inland point 1, ~ 5.8 for inland point 2, and ~6.7 for inland point 3. Standard deviations are indicated. = 1418, p < 0.001; n = 90) (Figure 4 11 ). Unexpectedly, there were no clear treatment effects of hog exclosure or tilling on the soil richness of any of the three nutrients examined (Figure 4 1 2 ). Contrary to the soil nutrient results, the nutrient content of the biomass was affected by hog exclusion. The phosphorus content of t he herbaceous community biomass in the plots overall averaged 1.1 mg/g and ranged from 0.4 3.2 mg/g. The island controls had higher P levels than the mainland controls (island: 1.4 0.18 SE; mainland: 1.0 0.08; Wilcoxon sum rank S = 239, one tailed p = 0.01). Tilling had no
103 long term observable effect compared to the controls. Exclusion of hogs lowered P in the herbaceous biomass by 0 .3 mg/g on the islands but had no ef fect on the mainland (Figure 4 1 3 ). Figure 4 9 15 N of the vegetation and distance from the ocean in the Lower Suwannee National Wildlife Refuge, Florida. The potassium content of the herbaceous community biomass in the plots overall averaged 12.9 mg/g and ranged from 4.1 25.6 mg/g. The island and mainland control K levels were not significantly different (island: 14.4 1.8 SE; mainland: 12.0 0.9; t = 1.2, one tailed p = 0.12). Tilling had no long term observable effect compared to the controls. Exclusion of hogs lowered K in the herbaceous biomass by 3 .0 mg/g on the islands but had no ef fect on the mainland (Figure 4 13 ).
104 Figure 4 10 Feral Hog fecal load in a coastal hammock community on the mainland (M) as well as isolated hammock islands (I) in the tidal marsh in the Lower Suwannee National Wildlife Refuge, Florida. The lines in the boxes indicate the medians, the boxes indicate the 25 th and 75 th percentiles, and the whiskers indicate the last sample within 1.5 times the interquartile distance. Samples outside that range are indicated with dots. M eans are indicated by the horizontal gray lines. The fecal loads in each season are significantly higher on the islands. The total Kjeldahl nitrogen (TKN) of the herbaceous community biomass in the plots overall averaged 12.7 mg/g and ranged from 5.0 26 .2 mg/g. The island controls had a trend towards higher TKN than the mainland controls (island : 14.9 1.3 SE ; mainland : 12.4 0.9 SE; t = 1.55, one tailed p = 0.07 ). Tilling had no long term observable effect compared to the controls. Exclusion of hogs lowered TKN in the herbaceous biomass by 3.3 mg/g on the islands but had no ef fect on the mainland (Figure 4 13 ).
105 Table 4 3. Estimated (N, P, K) deposition rates on the island and mainland coastal areas by Feral Hogs in the Lower Suwannee National Wildli fe Refuge. % of Fecal deposition (kg/km 2 /year) Urine:Feces Urine deposition (kg/ km 2 /year) Total deposition (kg/ km 2 /year) Nutrient feces 1 Island Mainland ratio 2 Island Mainland Island Mainland N 0.42 0.04 3 2.71 0.71 0.64 0.17 2.75 0.45 7.45 4.94 1 .75 1.25 10.16 4.99 2.38 1.26 P 0.17 0.02 1.10 0.24 0.26 0.06 0.54 0.09 0.59 0.28 0.14 0.07 1.69 0.37 0.40 0.09 K 0.14 0.01 0.90 0.19 0.21 0.05 3.25 0.16 2.94 0.69 0.69 0.22 3.84 0.72 0.90 0.22 1 By wet weight 2 The ratios of the amount s of the nutrients excreted through urine and feces based on hog nutrient budgets ( Kephart & Sherritt 1990 ; Sands et al. 2001 ) 3 Standard error Taken as a whole, the nutrient content in the herbaceous biomass showed the same patterns for all three nutrients: the mainland biomass was generally lower than island biomass. Additionally hog exclosures on the islands resulted in nutrient levels lowered to the levels found on the mainland making two fairly distinct groups: 1) the island controls an d island tilled plots and 2) the island exclosures, mainland controls, mainland tilled plots, and mainland exclosures (Figure 4 14 ). Population structure A total of 162 sightings of hogs allowed classification according to coat color: 102 on the coast, 25 in the non coastal portion of the refuge, and 35 inland from the refuge. 2 = 37.2, p << 0.001), with the coastal and refuge hogs different from the inland hogs but not from ea 2 2 = 10.9, p = 0.012; 2 = 0.87, p = 0.83) (Figure 4 15). Black was by far the most common color on the coast and in the refuge but the color distribution was more equitable inland,
106 F igure 4 11 Pre treatment nutrient richness of coastal hammock soils in the Lower Suwannee National Wildlife Refuge from summer 2007. Levels of all three nutrients, nitrate/nitrite nitrogen, phosphorus, and potassium, are significantly higher on the isla nds in the tidal marsh than on the mainland. Box plots are as in Figure 4 10.
107 Figure 4 12 Changes in phosphorus, potassium, and nitrate/nitrite nitrogen in the soil in the tilled and exclosure plots compared to the paired controls in island and mainland coastal hammock forests in the Lower Suwannee National Wildlife Refuge, Florida. Arrows ( E and F ) indicate tilling events to mimic increase hog rooting. 95% confidence intervals are indicated. A B C D E F
108 Figure 4 13 The treatment effect relative to pai red control plots of 3.5 years of hog exclosures and periodic soil disturbance on phosphorus, potassium, and total Kjeldahl nitrogen (TKN) in herbaceous community biomass. 95% bootstrapping confidence intervals are indicated. Note that the effect is the same for each of the nutrients: hog exclusion on the islands reduces the nutrient content of the herbaceous biomass.
109 Figure 4 14 Nutrient concentrations in herbaceous community biomass in control, hog exclosure, and soil disturbance (tilled) plots i n the coastal hammock on the mainland and isolated islands in the tidal marsh after 3.5 years of treatment and 21 months after the last tilling. The same basic pattern is apparent in all three nutrients (A) phosphorus, B) potassium, and C) total Kjeldahl nitrogen (TKN)): island controls have higher concentrations than mainland controls and hog exclusion on the islands reduces nutrient concentrations to mainland levels but has no effect on the mainland. Tilling appears to have no long term impact in eithe r area. The main groupings are easily distinguished in the circle comparisons of the same data on the right sides of each panel. The boxes indicate the median and the 25% and 75% quartiles, and the whiskers indicate the last sample within 1.5 times the in terquartile distance. Samples outside that range are indicated with dots. Means are indicated by the horizontal gray lines with the total mean indicated by the dotted line. Significant groups are indicated with letters above the boxes. A B C
110 indicating that the refuge/coastal hogs may be genetically distinct from hogs farther inland. When individual observations were plotted by location, the few samples north of the Suwannee appear to have a lower frequency of black phenotype than the rest of the coast (Fig ure 4 1 6 ). Figure 4 15 Coat color categories of Feral Hogs in and near the Lower Suwannee National Wildlife Refuge. Significant groups are indicated by letters. Out of 101 fecal samples gathered, microsatellites were amplified from 49 individuals. 27 of the 49 had 9 or more loci characterized. All loci were polymorphic with an average of 8 alleles/locus (2.45 SD) and a range of 4 12. Heterozygosity (H E ) of the loci averaged 7.46 (0.09 SD) and ranged from 6.24 9.22 (Table 4 1). Three genetic grou pings were inf from the study location (n = 12 for Barnett Creek 1 (BC1); n = 3 for Barnett Creek 2 (BC2)) while the last group was represented by individuals within the study location and all the peripheral indivi duals (n = 12 for General (G)) (Figures 4 1 7 and 4 1 8 ). The
111 Figure 4 16 Locations of observations of Feral Hogs by coat color. Individual sightings that overlapped due to the scale of the map were spread out for clarity. Most of the coastal observa tions are clustered around the study site ( Figure 4 1) and islands 2 km south of the study site. Figure 4 17 Genetic groupings of Feral Hogs in the Lower Suwannee National Wildlife Refuge as determined by S TRUCTURE Barnett Creek runs through the stu dy location. Peripheral samples were taken 2.8 km south southeast of the study location in the coastal hammock, 6.5 km southeast on the border of the refuge, 6 km north northeast on the border, 14 km north northeast, and 8.5 km northwest in the coastal ha mmock on the other side of the Suwannee River.
112 fixation index (F ST ), a measure of genetic differentiation between groups ( Wright 1965 ) was higher between BC1 and BC2 (0.3261) than between BC1 and G (0.1285) and between BC2 and G (0.1326). The genetic differentiation between populations was PT wise genetic distance between individuals was significantly greater within G than expected and significantly less than expected within B C1 and within BC2 (Figure 4 1 9 ). The numbers of migrants/generation between populations were estimated to be 1.3 (G BC1), 1.2 (G BC2), and 0.3 (BC1 BC2).The overall effective population size (for all three genetic groups together) was calculated to be 792 131 SE, which likely reflects an actual adult population size of 7197 2422 SE. Given the estimate of 27 pigs/km 2 used for nutrient deposition rates, this population may cover 200 300 km 2 essentially the size (though not necessarily the shape) of LSNWR (214 km 2 ). Discussion Diet Both the fecal analysis and the stable isotope signatures revealed that American Bulrush dominates the diet of Feral Hogs in the coastal hammock. In spite of this extreme and unexpected specialization on a single food sou rce, the hogs were clearly selectively feeding on other foods from time to time. One sample contained over a dozen Olive Nerite opercula and others contained a great deal of hair so these were not likely to have been incidentally ingested. Also, it was o bserved during the study that skeletal remains were often surrounded by hog trampled ground with the smaller skeletal components broken or missing, implying cannibalism. These observations as well as the fact that minor components in the overall diet oft en appeared as the major component in a few samples, indicates that the hogs will
113 Figure 4 1 8 Locations and genetic groupings of Feral Hog samples from the Lower Suwannee National Wildlife Refuge and environs. a) Locations of the study and individual peripheral samples. b) Locations of individual samples from the study location. Note that the genetic divisions are not related to geographic divisions. The color coding is the same as in Figure 4 1 7
114 Figure 4 19 Mean pair wise genetic distances bet ween individuals of the 3 inferred population groupings for Feral Hogs in or near the Lower Suwannee National Wildlife Refuge. 95% confidence intervals determined by bootstrapping are indicated. Error bars are the 95% confidence intervals for the means. opportunistically forage on other resources periodically (e.g., meat), which may not be detected in fecal analysis. However, the stable isotope data do not indicate that a major dietary component is being missed by the fecal analysis. Because the isotope signature of the hog diet almost perfectly overlays the bulrush signature, any other major fractions would have to be improbably balanced to remain undetected. Therefore, the vast majority of nutrients flowing through the coastal hogs are being sourced f rom the marsh generally and from American Bulrush specifically. 15 N signature in the collagen of younger animals is an expected result of lactating females catabolizing their own tissues to produce milk consumed by their offspring ( Newsome et al. 2006 ; Polischuk et al. 2001 ; Sare et al. 2005 ) Females able to lactate using increased food intake rather than body reserves, should not confer an enriched 15 N signature to their young ( Jenkins et al. 2001 ) The observation that lactating females seen during the st udy were boney and in poor condition (indicating
115 that lactation required the use of body reserves) is consistent with the negative correlation between 15 N and estimated biological age. 15 N in plant tissues with increasing distance from the coast is likely due to 15 N values in precipitation deposited N (NH 4 + : 3 : Shearer & Kohl 1992 ) that is countered based on distance from the coast by the enriched values found in coastal plants c aused by marine Heaton 1987 ; Virginia & Delwiche 1982 ) or found in saline soil; Heaton 1987 ) The importance of the marsh as a source of resource s likely results in a neighborhood effect that lea ds to their heavy use of the islands and relative avoidance of the mainland (Figure 4 10 and Chapter 3). The close juxtaposition of food and shelter results in fairly short transport distances, likely on the order of 100 200 m (Chapter 3). A third requ ired resource for pigs, wallows ( Fernndez Llario 2005 ) seems to be concentrated in open mud along trails through the marsh or on the shaded border with the hammock where the marsh vegetation cannot grow (pers. obs.) and are thus conveniently accessed from both habitats. The incorporation of C 4 plant derive d foods in the diets of non coastal hogs does not seem to be a result of participation in a C 4 grass food chain that is absent from the coast. All the grasses measured were C 4 grasses including two of the most common species in the coastal hammock. Rathe r it is likely that the C 4 resource being consumed is feed corn. The land bordering the refuge and farther inland is largely privately owned hunting and agricultural land including cattle range (pers. obs.). Hogs
116 outside the refuge are possibly feeding on corn pilfered from cattle operations and, probably more importantly, wildlife feeding stations meant to attract game. Based on observations of discarded feed corn bags, the feeding of wildlife peaks just before hunting season but may also occur year ro und to ensure that game animals are permanent residents. The fact that coastal pigs have very distinct stable isotope signatures from pigs on the border of LSNWF indicates that movement between the coastal hammock and inland areas is not typical and nutrie nt transport distances from the marsh are short and only go as far as the coastal hammock. This observation suggests that there might be little migration of individuals between the coastal hammock and other areas, which may lead to genetic population stru cture (discussed below). Nutrient deposition The amount of nutrients being transported by the hogs appears to be a fraction of the atmospheric deposition recorded for this region of Florida: N islands: 3%, N mainland: 1% ( Hendry et al. 1981 ) ; P islands: 7%, P mainland: 2% ( Hendry et al. 1981 ) ; and K islands: 11%, mainland: 3% ( Grimshaw & Dolske 2002 ) However, because of the centuries long history of hogs in Florida ( Mayer & Brisbin 1991 ) these small inputs appear to have resulted in an uneven distribution of nutrients in the coastal hammock, with the island soils being richer in nut rients than those on the mainland (Figure 4 1 1 ). Additionally, the high rates of denitrification that occur in waterlogged soils ( Morris 1991 ) likely increase the importa nce of nitrogen deposited by the hogs. Soil and plant nutrients The exclosures and tilling treatments did not have clear patterns of effects on the soil nutrients. This may be because the scale of the exclosures was smaller than the
117 scale at which some soil processes operate or that the treatments were too short in time The exclosures were built to address questions about the herbaceous community and were far too small to incorporate the root system of mature trees. Tree roots comprised an average of 8% of the dry weight of the soil cores and much greater (unmeasured) weight and volume fractions while fresh. Thus trees with roots both inside and outside an exclosure could have been supplementing tissues inside the exclosures with nutrients brought fro m richer soils on the outside. When the soil samples were dried and ground, root fragments could have biased the soil results towards the nutrient content of the trees. Another difficulty with the soil samples is the fact that they were gathered at single points in time. Because the nutrients (especially N) are often mobile in the soils, patterns of rainfall and flooding can add to the unpredictability of nutrient availability caused by animal depositions and other periodic pulses of nutrient flux. This variation likely overwhelmed the once per 6 month soil sampling schedule and confounded the soil richness portion of the experiment. The exclosures did show that the hogs do have a measurable effect on the nutrient content of the herbaceous biomass, howeve r ; even over the four year time scale of this study A soil sample is representative of one instant in time so any given sample is not likely to be an accurate measure of the changing nutrient levels available for plant growth, but plant tissue gives much better estimates of the available nutrients as it provides an average of nutrient levels experienced over time ( Sartain 2008 ) Thus while tilling do es not have a significant long term impact on the nitrogen environment
118 experienced by herbaceous plants, exclusion of hogs does have an impact on the islands but not on the mainland (Figures 4 1 3 and 4 1 4 ). When hogs are excluded from an island hammock, th e nutrient levels in herbaceous biomass drop to levels found on the mainland. This indicates that hogs are important in transporting nutrients to the islands but that the density of hogs on the mainland is small enough that the amounts of N, P, and K depo sited there are negligible. It also indicates that the effect is seen fairly quickly in the herbaceous vegetation (within 4 years). Interestingly, one of the most common herbaceous plants, St. Augustine Grass, is a common turf grass and therefore has fair ly well described nutrient requirements. It grows best with access to enough nutrients for 2 3% of its tissues to be nitrogen 0.15 0.5% phosphorus and 1 3% potassium ( Sartain 2008 ) To put this in perspective, plant tissues are typically 1.5% N (range: 1 5% in non woody tissues; Mattson 1980 ) 0.2% P and 1.0% K ( Billings 1970 ) It is apparent that in this system some nutrients may be limiting for plant growth (Figure 4 1 4 ). The island non excl osure biomasses approach typical plant levels of nitrogen content but none reach ideal levels for St. Augustine Grass, while the mainland and island exclosure levels are lower than typical. Because denitrification proceeds apace in waterlogged soils ( Morris 1991 ) in this c onstantly moist system nitrogen loss is probably severe so additions by hogs are important in maintaining current levels. Phosphorus levels also seem to be insufficient for ideal growth by St. Augustine Grass. The nutrient content of the plant tissues is below typical for plants in general and below ideal on both the mainland and the island. However, on the islands
119 phosphorus levels approach the lower e nd of the range for non exclosure plots (Figure 4 1 4 A ). The soil on the islands is within the lower end of the range required for good growth (Figure 4 1 1 ; 16 30 mg/kg; Sartain 2008 ) but the mainland is even lower. Potassium levels appear to be adequate for St Augustine Grass and above what is typical for plants in general, as might be expected given that the richness of the soil in potassium (Figure 4 1 1 60 mg/kg; Sartain 2008 ) by one order of magnitude. The observed increase in K levels in p lant tissue on the islands likely has more to do with release from the limitations of other nutrients rather than the higher K levels in the island soils. Population structure The stable isotope data suggested that there might be different populations of pigs using the coastal hammock and the rest of LSNWR, at least the inland border of the refuge. The observed hog color frequencies (Figures 4 1 5 and 4 1 6 ), however, suggest that the coastal hammock and refuge hogs might be in a single genetically mixed p opulation (even if individuals rarely move to different habitats during their lifetimes) but the inland hogs might be distinct. It was expected that coastal hogs should be predominantly black as that color confers protection from the fierce Florida sun w hile the pigs are foraging in the marsh, but hogs living in fully forested systems, such as the upland pine/hardwood swamp system of the rest of the refuge, should show much more color variation as the selection for sun protection is relaxed under a canopy ( Mayer et al. 1989 ) This expectation is clearly met when considering the coastal hammock and the inland areas. However the refuge hogs are predominantly black as well. This observation does not necessarily conflict with the expectation for increased selection for black in exposed habitats
120 because 1) black is likely to be beneficial in a largely forested habitat as it is less visible to predators/hunters and 2) if phenotypes are not or weakly selected in the rest of the refuge, phenotype frequencies there could still be influenced by strong selection on the coast and migration between the two areas. Interestingly, the black phenotype does not appear to be dominant in frequency in the few observations north of the Suwannee River (Figure 4 1 6 ), contrary to expectations. This might be due to an influx of new migrants in the recent past but is more likely to be simply a random effect of a small sample size fro m that area. The impact of small sample size is exaggerated in this case because the social system of hogs is based on groups of related individuals (sounders) and samples of phenotype from related individuals are not independent of each other. The o bse rvations from that location were collected during only two days, so it is possible that all the pigs belonged to only one to three sounders and were highly related. Of the three genetic groups found in the genetic data, one contained all the peripheral pig s and some pigs from the study location while the other two were found only in the study location but intermingled geographically (Figure 4 1 8 ). The two groups from the study location, BC1 and BC2, are probably sounders composed of closely related individ uals. In other populations of Feral Hogs, sounders are distinct enough genetically to be detected with similar methods to those used here ( Spenc er et al. 2005 ) The General group (G) is very unlikely to be a single thoroughly well traveled sounder. Rather, G is probably composed of individuals from different sounders. Because S TRUCTURE was not able to group any of those hogs with any others f rom the same sounder, it created a general group that encompassed all hogs from sounders
121 with only one representative and two groups with hogs from sounders with multiple representatives. This interpretation is supported by the F ST values among groups: th e sounders are more distinct from each other than they are from a group that includes members of multiple sounders from the general population. Because each sounder probably has its own distinct familial genetic profiles, it is less similar to another soun der than to a group made of genetically diverse members of the general population. The intra group pair wise genetic distances also support the two sounders within a larger population conclusion. Any two individuals from the general population are more g enetically different than expected based on the entire population structure while any two individuals from the sounders are more similar than expected (Figure 4 1 9 ). These results must be treated with caution, however, due to the small number of individual s in BC2 and the lack of resolution in the general population. The scale of genetic differentiation detected was much finer in resolution (between sounders) than required in this study (between populations in geographic areas). In order to fully tease ou t the inter sounder population structure hinted at by the stable isotope and color frequency data, a much more intensive and complete sampling regime will be needed with multiple samples from each sounder. The population estimate, 7197 2422 SE pigs, see ms reasonable considering that this region of Florida has long been known as having an extremely high density of Feral Hogs ( Mayer & Brisbin 1991 ) However, given the assumptions of estimating an effective population size, it is likely that this populati on estimate should be taken as an upper limit. The method used to estimate effective population size assumes that the
122 equilibrium has been reached between addition of new alleles through mutation and extinction of alleles through genetic drift ( Frankham 1996 ) This process takes a great deal of time. For example, the effective population of humans is estimated from molecular data to be ~10,000. This excessively low estimate is a signature from past bottlenecks in human populations and not an accurate gauge of modern N e ( Charlesworth 2009 ) In the case of Feral Hogs along the Gulf Coast, they are likely derived from multiple lines of domestic hogs introduced to Florida multiple times since Her nando de Soto first introduced them in 1539 ( Towne & Wentworth 1950 ) Until the 1960 s Florida had an open range law that allowed livestock owners to let their animals roam free throughout the landscape and a continual source of novel alleles for the feral population ( Mayer & Brisbin 1991 ) Even if each breed ha s low genetic diversity, a population derived from multiple breeds from around the world has the potential to hold much greater genetic variation than would be expected based on population size ( Laval et al. 2000 ) The Ecology of Feral Hogs Feral Hogs have strong impacts on the biomass and species richness of the coastal hammock herbaceous community (Chapter 3) and, as evidenced by this chapter, they also influence the nutrient status of the herbaceous community through deposition of marsh derived nutrients in the coastal hammock. Their influence on biomass an d perhaps species richness is likely due to an increase in disturbance, which in itself can increase the susceptibility of plants to herbivory by insects ( Spiller & Agrawal 2003 ) However, because of the waterlogged condition of the coastal hammock, oxygen limitation is likely one of the most important limitations on plant
123 growth. Slight differences in elevation (amounting to centimeters) appear to lead to differences in flooding frequency and exposure to moisture wicked up from lower soil layers and may lead to extreme differences in herbaceous cover when other conditions seem comparable (pers. obs.). Through rooting, hogs may partially mitigate this limitation by aerating the so il, but the direct effect the disturbance has on biomass likely eliminates this benefit in te rms of standing crop of biomass ( though it might lead to faster recovery from rooting than from a less d isruptive form of biomass loss, such as grazing). This may change competitive interactions e.g. benefiting plants not specifically adapted to waterlogged soils by giving them areas of temporarily increased O 2 availability. The herbaceous community in this habitat was likely constrained to a great extent by the moist anoxic soil conditions, yet nutrient deposition still had a significant effect on the plants. In less anoxic terrestrial systems nutrient limitation might play a larger role in restricting plant growth and determining species composition. Therefor e nutrient deposition might be even more important in terrestrial systems than it is in the coastal hammock Nutrient enrichment can also shift competitive interactions among plants ( Gaucherand et al. 2006 ) and favors ruderal species, which are also benefited by disturbance and reduced competition (reduced biomass and reduced species richness). These sorts of changes are typical of anth ropogenic influences and favor many exotic species. Thus it is likely that Feral Hogs in this and other systems diminish the resistance of the native communities to compositional changes and invasion by exotic species, not just by the physical disturbance that tends to be the focus of most studies, but also by depositing nutrients derived from rich feeding grounds (wetlands, agricultural
124 lands, middens, etc.) in refuge habitats that may already be fragmented and vulnerable to biotic invasion and degradatio n. Another important aspect of Feral Hog ecology in this system appears to be occurring in the marsh. American Bulrush was observed to grow only in th e areas of the marsh rooted by Feral H ogs. Even when the bulrush had been left unrooted long enough to m ature, the soil remained more loosely packed and the vegetation much less crowded than in neighboring Black Needlerush or Sawgrass patches. This observation leads to three possible conclusions: 1) either hogs feed only in the bulrush patches, 2) bulrush g rows only in hog rooted areas, 3) both 1 and 2. If hogs only fed in the bulrush patches, it would be unlikely for the hogs to root the entirety of every patch and the edges of the patches should be unstable as feeding by the hogs should tend to reduce the competitive ability of bulrush compared to other marsh plants. However, no unrooted patches of bulrush were ever found and the bulrush patches are amazingly stable. A comparison of aerial photographs taken over 50 years shows stable patches of rooted bul rush in the same areas (Figure 4 20 ) In fact, small enclosed patches of needlerush within the bulrush patches are surprisingly stable as well. Because Saw grass is a fresh/brackish water species ( Wunderlin 1998 ) its patches are usually found along the creek or next to the hammocks. Black Needlerush is a brackish water species ( Pennings et al. 2005 ) and is found (in this area) from the hammock all the way to the ocean. American Bulrush is a brackish/salin e water species ( Wunderlin 1998 ) but is only found in hog rooted area s near the coast and islands.
125 If bulrush only grew in hog rooted areas, the patches would similarly be unstable as hogs would be creating new bulrush patches as they rooted through the marsh. However, not only are the patches stable but, based on bulrush important dietary component, it is unclear what would drive hogs to root outside bulrush patches. Figure 4 20 Aerial photographs of a portion of the field site showing the spatial stability of the marsh community over time. A ) 19 61, B ) 8/30/2006, C) 1/30/2008, D ) 1/18/2012. br = American Bulrush, nr = Black Needlerush, sg = Saw grass. Patch composition can be identified by coloration and the sharper shadowed border usually found around bulrush patches due to the difference in he ight and density of the bulrush patches. Color changes with season but needlerush patches are normally a dark green, Saw grass is usually a light green or yellow green, and bulrush is normally a light green or brown green although patches are often very d ark if freshly rooted. Courtesy of USGS and Google.
126 The most likely scenario is that the hogs and the bulrush are locked in an ecological mutualism: bulrush provides the hogs with an abundant food source while the hogs maintain the bulrush patches by per iodic rooting that discourages invasion by needlerush and Saw grass (which are probably competitively dominant to bulrush in their preferred salinities). Thus both species are regulating the other, which may explain the stability of the bulrush patches. At the edges of the bulrush patches, the needlerush is sometimes pushed up as the hogs root the entire patch, likely checking the advance of the needlerush (pers. obs.). Additionally, during this study it was noted that most bulrush rhizomes tended to be in a horizon below 10 cm into the substrate. This may be due to the hogs consuming most of the rhizomes above this level, explaining how bulrush is able to flourish despite heavy feeding by hogs. More work is needed to elucidate the exact nature and exte nt of this mutualism between an omnivore and its (locally) dominant food Conclusion Feral H ogs are global pests that are extremely destructive to native plants, animals, and ecosystems in general. Much research has been done describing the physical impac ts they have through trampling, rooting, and consumption, but the indirect effects they have on the ecosystems through nutrient transport may also be important and should not be ignored. Additionally, when hogs have been present in a system for a long per iod of time, they may have established novel ecological links that change the structure of the system. While the hog induced changes may or may not be desirable, studies beyond cataloguing physical damage will give a deeper understanding of the types of i mpacts that exotics and omnivores have on native communitie s, and of ecological interactions as a whole.
127 CHAPTER 5 DEFINING WILDERNESS Introduction The division between Human and Nature is one of the most fundamental distinctions made about the essence of reality ( Brown 2007 ) The existence, exact placement, and c haracter of this Great Divide have implications for the proper role and nature of humanity itself and are therefore matters of longstanding and often heated debate ( Callicott 2007 ; Oelschlaeger 2007 ) This debate was previously meant to define an objective, naturalistic morality but more recently the primary importance of the Human/Nature division has been to conservation biology ( Cronon 1995 ; Ereshefsky 2007 ; Oelschlaeger 1991 ) With this perspective, the focus sh ifts to the division between human engineered landscapes (cultural areas) and natural landscapes (wilderness). Thus the discussion revolves around the meanings of the word With the globalization of the largely American invention of nationa l parks, came the ( Harmon 1987 ) Because the American ideal of wilderness is a vast uninhabited, roadless, purely natural landscape ( Nash 1967 ) there has been a backlash against the concept of wilderness in general in countries that may have n o areas that could qualify as wilderness under this particular definition (due to space or extent of anthropogenic influence) and in countries with populations dependant on the natural landscape ( Bayet 1994 ; Guha 1989 ) These populations have historicall y been and are currently still susceptible to being dispossessed of their lands, livelihoods, and eventually social and cultural cohesion when protected areas are gazetted ( Goldman 2011 ) Thus hostili ty towards past
128 concept of wilderness and sometimes even to conservation in general ( Guha 1998 ) By examining the core philosophical problems with universally applying the dominant Western definition of wilderness, I hope to uncover a basal definition of wilderness that is both philosophically sound and useful in conceptually uniting the co nflicting Defining Wilderness American Style The standard legal definition found in the U. S. Wilderness Act ( 1964 ) reflects the American cultural ideal, guides land use and the establishment of protected areas within the United States, and, acting as a model for other parts of the world, has ramifications for conservation far beyond its jurisdiction. It s tates: A wilderness, in contrast with those areas where man and his own works dominate the landscape, is hereby recognized as an area where the earth and its community of life are untrammeled by man, where man himself is a visitor who does not remain. Th is is the core wilderness definition of the Act which further qualifies a legal is natural with little human influence, and 3) has no permanent human presence. These three points are consistent with and often appear in the laws of other countries with large areas of sparsely inhabited wildlands e.g. Canada : Wilderness and Ecological Reserves Act ( 1980 ) and Provincial P arks and Conservation Reserves Act ( 2006 ) Australia : NSW Wilderness Act ( 1987 ) and Iceland : Nature Conservation Act ( 1999 ) and in the International Union for the Conservation of Nature Wilderness Area classification of protected areas (category Ib ; Dudley 2008 )
129 These three points resonate with many cultural definitions of wilderness. However, they do not necessarily transfer well into cultures that may have very different economic, spatial, and philoso phical relationships with nature and thus very different ideas of what wilderness is. Given that wilderness acts as a sort of Holy Grail (an ideal goal combined with a romantic vision with long history and deep spiritual and emotional meanings) for the co nservation movement, American style wilderness ideals were naturally exported at the same time as American style conservation science. Unfortunately, the mismatch between ideas of wilderness and practical circumstances on the ground can lead to misunderst andings, mismanagement, and/or the inappropriate establishment of fortress style conservation and the resulting displacement and disenfranchisement of indigenous populations ( Goldman 2011 ; Guha 1989 ) Worse, the outright hostility to the American ideal of wilder ness and the fear of new colonialism can develop into an aversion to conservation in general and hamstring international and local conservation efforts ( Spinage 1998 ) In part to address these problems, many legal definitions of wilderness arbitrarily focus on t echnological level, traditional use, race/ethnicity, etc For example, the South Australia Wilderness Protection Act (SAWPA) ( 1992 ) ecosystems must not have been affected, or must have been affected to only a minor However, definitions of wilderness that allow traditional use are plagued with the philosophical problems justifying why certain uses, technologies, or people are compatible with wilderness while others are not. While new technologies tend to be more effective than traditional ones and thus increase the potential ecological impact of each individual, even the most traditional
130 technologies are not in and of themselves necessarily sustainable, as witnessed b y numerous ancient anthropogenic extinctions ( Barnosky et al. 2004 ) and landscape transformations ( Diamond 1994 ) Even if a particular traditional system has been in place for long periods of time, changing environmental conditions may demand corresponding changes in use to maintain sustainability. Traditional use by indigenous peoples in the eyes of colonizers and their descendents tends to be whichever indigenous group happened to be in an area doing whatever they were doing at the time of colonization. This view does not allow for change (cultural, spatial, or env ironmental), and naturally leads to a dogmatic assumption of stability and sustainability. This view is encoded in the SAWPA ( 1992 ) condition befor rests on assumptions of general stability of the number of traditional users and their individual impacts. With environmental conditions, numbers of traditional users, and access to more effective technologies changing, it is clear that any traditional use must be managed in some way; however limiting access to subsistence users presents greater moral and practical difficulties than limiting access to recreational or facultative user s. Adding to the difficulties involved in defining traditional use is the fact that the issue of traditional use often does not revolve around traditional use, or even traditional technologies, it revolves around the ethnicity or race of the traditional us ers As examples, the SAWPA allows for hunting and observation of traditions by Aboriginals in wilderness areas, the Finnish Act on Wilderness Reserves ( 1991 ) states th at wilderness
131 reserves are established in part to preserve Lapp culture, and Dear and Myers ( 2005 ) found that the race of the hunter determined whether or not subsis tence hunting in Alaskan wilderness was supported for 33% of recreational visitors. Traditional use, then, is often a compromise between having a human free wilderness area and social justice for indigenous peoples, which can lead to the shaky conclusion that a blond hunter in Alaska is somehow incompatible with the wildness of a place but an Inuit hunter is not. Likewise, an ethnic German hunter would not be incompatible with Alpine wilderness but a naturalized German hunter born in Lesotho would. The mental contortions involved in philosophically allowing for certain ethnicities or races in wilderness, defined as a place without humans, at best puts a barrier between two groups and at worst belittles a culture and dehumanizes a people. If humans do n ot live in wilderness, it follows that anyone living in a wilderness is, ipso facto subhuman. This is essentially the line of thought that led to the colonial (and current ; ( Diegues 1998 ) idea of terra nullius (un or underdeveloped/u ncultivated land is not owned and is therefore free for the taking), which gave colonists a moral excuse, or even imperative, for dispossessing indigenous people of their land ( Bayet 1994 ) T o avoid misunderstandings, defuse conflict, correct prejudices, and facilitate communication, it is necessary to unify in the vast array of wilderness definitions. This cannot be done by replacing a diversity of definitions with uniformity because even th e best single definition would be inappropriate in many cultures and contexts. Rather than replacing diversity, it would be better to uncover the fundamental unity upon which the diversity of definitions and opinions are built. If a fundamental and logic ally sound definition of wilderness unbiased by any particular culture can be established, it can
132 help make people aware of their own deeply held perceptions and serve as a basis for intercultural communication and understanding about conservation, allowin g the concept of conservation to be locally adapted and appreciated more readily. A Basal Definition: The Not So Great Divide ( Ereshefsky 2007 ) and run the gamut from excluding all human influence whatsoever, t o allowing some types of human influence, to allowing all human influence (because humans are just anoth er natural product of evolution ; Henberg 1994 ) Contrasting with the extreme definitions, which are clear in theory but uselessly encompass either no thing or everything, the practical definitions of wilderness allow a confusing variety of types and extents of human influence ( Ereshefsky 2007 ) it inherently requires a negative definition ; not a definition of what it is, but of what it is not Just as dark, cold, a nd quiet do not require the total lack of light, heat, and sound, wilderness can be defined as the relative lack of cultural influence. The exact placement and width of the boundary between cultural space and wilderness and how much time and effort is nee ded to transform an area of wilderness into a cultural space will depend on the culture of the observer. However, it is fairly easy to construct a usable and flexible definition of cultural space. Cultural space is an area in which the biotic community is purposefully and fundamentally altered and that alteration is constantly maintained by humans, for the use of humans and/or their domesticated symbionts. Generally this involves a
133 significant conscious effort to control and transform land cover and exc lude undesirable flora and fauna. Because significant effort is involved, the boundaries of cultural landscapes tend to be surprisingly sharp with an area either maintained or not ( Worman 2010 ) For example, settlements and agricultural land (including improved pasture planted with desirable grasses) are obviously within the cultural sphere. However, a river, polluted beyond recognition, though significantly and continuously altered, is not within the cultural sphere as the alteration was not an intentional endpoint. A grassland, burned to attract game or provide better grazing, is purposefully altered but not maintained so a lso does not enter into the cultural sphere. A coppiced woodland is also altered but with little maintenance or effort to control. Instead, what effort is expended is focused on management of a few key characteristics of a small subset of species, not pu rposeful maintenance of a fundamentally different community. Forest clearing, whether done to procure wood or produce rangeland, does not necessarily move an area into the cultural sphere if the subsequent succession proceeds without undue human direction Low intensity management of certain species or morphotypes, such as periodic burning or clearing of woody species does not transfer an area into the cultural sphere. High intensity management to purposefully control the grass and herb species and acces s by wildlife, however, would. frequent, and consistent activities meant to preserve a fundame ntal anthropogenic infrequent, and often contingent activities meant to preserve or increase some value of
134 the biotic community. Clearly there is a continuum between thes e two terms and, thus, a zone of transition (however narrow or broad) between cultural space and wilderness. Defining wilderness as any area outside human cultural space avoids several of the problems inherent to other definitions ( Ereshefsky 2007 ) It avoids the futility of the extreme definitions and the arbitrariness of the others. This definition allows for human use of wilderness (whether recreational or subsistence) and, perhaps most importantly, is largely decoupled from spatial considerations. Many definitions of wildern ess explicitly or implicitly focus on size as a major criterion, and for good reasons: larger areas tend to be more sublime, more biologically diverse, more ecologically complete and stable, more dangerous, and with more opportunities for solitude; thus, m ore wild. Leopold ( 1921 ) defines a wilderness as being large enough for a two week packing trip and many official definitions include minimum sizes for wilderness areas (2023 ha Wilderness Act 1964 ; 2500 ha Nature Conservation Act 1999 ) state that wilderness es are large areas (Wilderness and Ecological Reserves Act 1980 ; Provincial Parks and Conservation Reserves Act 2006 ; Dudley 2008 ) or imply large areas by the scale of established wildernesses (15,000 294,000 ha ; Act on Wilderness Reserves 1991 ) Implicit to scale is the generally unspoken issue of continuity. Wildernesses are seen as areas that are entire landscapes uninterrupted and unfragmented by human works. Of course, there is a multitude of practical reasons to stress continuity, including ecosystem sensitivity to fragmentation and ease of management. However, by focusing on continuity and entire landscapes, the official and unofficial interpretations of globe where only smal l areas remain wild or human populations are integrated into a
135 largely natural landscape. Focusing on landscape scale wilderness conservation can lead to lost opportunities for gazetting protected areas, the barring of local people from accessing importan t resources, or the relocation of entire communities. Fears of these externally applied social upheavals can lead to an anti conservation backlash from people who would likely otherwise support conservation ( Goldman 2011 ) By cultural area, the concept of wilderness becomes independent of size and continuity. Wildness then relates to wilderness as water relates to a body of water regardless of size and shape. Just as bo dies of water vary from huge and contiguous (oceans) to small (prairie potholes) with complex boundaries (rivers and streams) wilderness is not limited to vast uninterrupted wild tracts but is instead an area of wildness whether large and contiguous with s imple boundaries or small with complex boundaries enclosing or interdigitated with cultural areas. Thus, under this definition, a landscape can contain both wilderness and cultural lands and small cultural areas do not need to be forced under the spatial and philosophical umbrella of wilderness. Yet the question remains: why are the cultural spaces of humans considered so fundamentally different from the living spaces engineered by other organisms like beaver ponds or termite mounds? Implicit to all defin itions of wilderness is the tenet that humans are the only force able to destroy wilderness ( Ereshefsky 2007 ) Even those that focus on ecosystem qualities, such as intact ecosystem functions or diversity, are still based on the assumption that relevant disruptions of the ecosystem qualities are anthropogenic. A volcano covering a forest with lava is seen as fundamentally different from humans paving a parking lot.
136 Callicott ( 2007 ) makes the case for the Great Divide being based in the differing temporal scales of evolution and culture. Cultural changes are orders of magn itude faster than geological, climactic, or evolutionary changes, and land transformations initiated by human cultures cannot be adapted to as quickly as they change. This perspective is largely true, and certainly important when considering anthropogenic impacts, however, even non anthropogenic changes can occur at super evolutionary speeds. Quick geologic changes, such as volcanic eruptions, landslides, meteors, unusual floods or draining events, are thankfully rare and one time events. On the other ha nd, non anthropogenic biological invasions can have rapid and persistant impacts on the native communities just like human mediated invasions and anthropogenic landscape transformation. Additionally, the temporal mismatch explanation is a post hoc justif ication requiring modern science to explain a concept that seems to have been shared by all human cultures (see below). Modern biology, while obliterating past beliefs about the special creation of Human as apart, above, and fundamentally different from N ature, seems to tacitly support such a division by focusing much of ecology and conservation biology on the biotic world sans humans ( i.e. wilderness). The answer to this apparent conundrum lies in the matter of perception. Scientists, and every other h uman, view the world from an anthropocentric perspective. This statement is not meant to be pejorative and is certainly made with the human capability of empathy in mind, but just as individual humans have a sense of self and others, humans collectively l ook outwards from themselves and their cultural space at the Other, the Wilderness. If perspective is shifted from anthropocentric to, for example, castoracentric, it is easy to see how the
137 concept of wilderness also shifts. The lodge, dam canals and d am created pond are this area is the castoracentric wilderness. The wooded land around the pond is utilized for food and materiel but is not transformed and maintained in the way that the engineered space is and remains a place of relative vulnerability and danger. The concept of wilderness need not divide Human from Nature because every species that creates engineered space, however small and temporary, separates an a rea from the whole and in doing so, creates its own version of wilderness. Of course, humans are primarily concerned with the anthropocentric version of wilderness and anthropogenic impacts on wilderness This is why deforestation caused by beaver ponds i s seen as fundamentally different from anthropogenic deforestation. Simply put, humans are concerned with human presence in and impacts on anthropocentric wilderness both through a sense of responsibility and because it would be difficult if not unnatural to view the world any other way. The extent and magnitude of anthropogenic alteration of the biosphere are just two of the qualities that make especially in terms of preservin g biodiversity, but they in no way separate or remove humans from that community. Thus the definition of wilderness as the space outside the engineered space of a species solves an age old philosophical conundrum by recognizing that the Great Divide is si mply a identity boundary possessed by every species dividing it and its engineered space (if any) from the rest of the biosphere. Universality of the Basal Definition For this definition of wilderness to be of pancultural use for discussions of conservat ion, the distinction between human engineered (or cultural) space and extra
138 most if not all peoples. This position is likely to be a controversial one as the idea that eve ry human culture contains some sort of wilderness concept defies long and fondly held paradigms that focus on differences in amount of material culture between peoples. These paradigms include both the outdated that attempt to place certain other culture s on the other side of the Great Divide (such as the paradigms of colonial Europe, which often viewed conquered peoples as sub human or different biological species altogether) and the more current paradigms that view present ecological problems as symptom s of too much emphasis on the Great Divide and loss of connection with the wilderness. However true or not the latter may be, denying that other peoples recognize cultural space as distinct from everything else (wilderness) is a variant of the Noble Savag e idea that civilization is the cause of all human woes and somewhere there is a group of people who live happy fulfilling lives of peace and harmony because they are one with Nature, or at the very least that it is possible to cure human ills by living as one with Nature. It is of course impossible to positively establish that all human cultures have had a wilderness concept, but any creation and maintenance of cultural space necessitates a recognition of the boundaries of the cultural space and thus a rec ognition of the division between the cultural space and everything else. Whether the cultural space is a permanent cement clad city surrounded by intensively used and clearly demarcated agricultural fields, or a group of temporary grass huts in a clearing the area must be recognized as something different, something more humanized than everything else. Otherwise, there would be no reason to create and maintain it in the first place. Both an
139 awareness of self and self works (in terms of individuals and a s a species) and an evolutionary advantage in being able to recognize the works (and therefore cultural areas) of unknown humans ( Dawkins 2006 ) imply a universal human notion of wilderness. While it is not feasible to review the wilderness concept in every human culture, a review of some will help illustrate the diversity of its expression among cultures. Some authors reviewe d here explicitly discuss cultural concepts of wilderness or the Great Divide, others imply a wilderness concept, and one denies the existence of a wilderness concept in his culture. At titudes and emotions towards wilderness can vary widely among culture s, individuals, and time periods, and even the distinctness of the division between Human and Nature can vary from being culturally explicit and celebrated to a division that is purposefully minimized (see below). Even the belief that humans are one with Nature or part of the natural world does not eliminate the distinction between self and other and cultural space and wilderness. This phenomenon can be seen quite clearly in modern Western cultures where the belief in a deep and fundamentally important Gr eat Divide exists in spite of the knowledge that humans are simply a naked, bipedal, super intelligent chimpanzee with genetic and evolutionary kinship with the rest of life and completely dependent on it for survival. In spite of the differences in att itudes and emotions towards wilderness and differences in emphasis placed on its boundary with cultural space, concepts of wilderness in diverse cultures are unified by certain basic perceptions and spatial organizations. Because wilderness is the constan t antithesis to cultural space a concept of wilderness can be recognized in other cultures by perceived qualities of
140 danger, chaos, isolation, and /or foreignness that oppose the qualities of relative safety, control, community, and /or belonging that chara cterize cultural spaces ( Worman 2010 ) Additionally, cultural areas, especially small ones, tend to be centralized with rela tively simple, often distinctly demarcated, boundaries with the wilderness (see below). ( 1993 ) San group of hunt ers and gatherers in southern Africa, explicitly illustrates a detailed civilization/wilderness boundary (p. 35) The village ( chu/ o ) is the wilderne ss (p. 31) described as a series of concentric zones with unique characters and uses: the central plaza, the ring of huts and firepits, a zone of middens behind the huts, a larger zone for butchering and cooking of larger animals a buffer zone, and an area for defecation. Beyond these is the (bush), the wilderness. The foundational definition of wilderness presented in the current paper would count only the plaza and huts as cultural space as these spaces are maintained cl vision of cultural space seems to add the middens, butchering zone, buffer zone and knowledge of and are comfortable in the bu t they have no great love or spiritual reverence for the wilderness (p. 31) ( 1961 ) study of another group of hunters and gatherers, the Mbuti Pygmies of central Africa, he does not specifically address the existence of a Great Divide and many comments made both by Turnbull and the Bambuti seem to indicate a lack of a wilderness concept. For instance, Turnbull mentions many times that the Bambuti view the forest as their home and are completely comfortable in it (pp. 13 14,
141 55, 172 ) He stresses their faith in the fundamental goodness of the forest (pp. 92 9 3) danger [in the forest] (p. 14) us everything we need food, clothing, shelter, warmth (p. 92) Yet the conclusion that the Bambuti lack a wilderness concept would be (p. 45) and example (p. 92) both of which indicate the solidarity of cultural punishm ent for incest: alone. Nobody will accept him into their group after what he has done. And he will die, because one cannot live alone in the forest. The forest will kill him, (p. 112) Thus, while the Bambuti have affection and reverence for the forest they also recognize taken in the context of his contrasting the Bambuti forest relations hip with the distaste the Bambuti hold for the Bantu villages (pp. 172, 227) and the horrible fear and hatred the neighbouring Bantu tribes have for the forest they view as hostile and evil (p. 55) The Bantu agriculturalists are described as having perce ptions of the forest that are easily recognizable as a wilderness concept. The forest is seen as foreign, rarely visited, constantly fought, and filled with dangerous spirits and animals (pp. 13, 171) Thus while both peoples view the forest as wildernes s, their emotions regarding and perceptions of the wilderness are polar opposites.
142 Chagnon ( 1992 ) people of South America, have a very explicit and prideful Great Divide based on the possession of culture by humans. They divide the world into urihi t rim (things of the forest) and yahi t r im (things of the village) (p. 105) the latter of which also includes gardens (p. 55) children have a safety limit around the villages beyond which they are not allowed ( Plotkin 1993 p. 252) and the adults have an abject fear of being caught in the forest at night without fire due to the dange rs presented by malevolent ghosts ( bore ) and jaguars ( Chagnon 1992 pp 106 &113) The Naxi people from southwestern China describe the Great Divide as having occurred when the Naxi progenitor and his brother Shu, the Nature governing spirit, went their separate ways with Naxi taking the valleys, crops, and domesti c animals and Shu taking the mountains, rivers, forests, birds, and wild animals ( Xu et al. 2005 ) So me cultures physically delineate the boundary between cultural space and wilderness. For example, the Tibetans divide the world into internal space (the villages and fields) and external space (the mountains and forests) with the boundary marked with ston e piles ( mani ) and incense burners ( Litzinger 2004 ) Likewise, the Hani from Southeast Asia build a gate ( lanlkang ) to separate the village ( pucang ; pu = village, cang = place or forest) from the nonhuman world ( baolcang ). I nterestingly, the villages are surrounded by protective forests hundreds of meters wide that are also called pucang ( Xu et al. 2005 ) which seems to indicate that the protective forest is incorporated into the cultural space or at least considered transitional between the cultural space and wilderness.
143 ( 1933 ) autoethnography he states that wilderness was a European idea that was completely foreign to the S ioux. His claim has been repeated and used as a criticism of the concept of wilderness ( Callicott & Nelson 1998 ; Nash 1967 4 th ed. p. xiii) however it is more correctly interpreted as a criticism of the contemporaneous Euro Am of his childhood in relatively intact Siouan culture is both highly idealized, elegiac, and purposefully contrasted with that of the Euro Americans. It must be understood that his writing is more literary than literal and with a strong social agenda. (This is made clear one ; e.g. Hassrick 1964 .) Did Standing Bear gloss over the Lakota perceptio n of wilderness? While the unity of all things is stressed in Lakota cosmology, the division between human and non human is important ( DeMallie 1987 ) Indeed, the Lakota do have the concept of cohesive cultural space as separate from wilderness. The solidarity of cultural space is evidenced by the prototypical layout of a village (the actual layout of which was often dependent on topography and obs tacles ; Hassrick 1964 p.173 174 ) : the tipis arranged in a circle around a central plaza ( hocoka ) with an opening in the circle for the official entrance ( oinapa ) with sweat lodges ( ini tipi ) outside the circle ( Walker 1982 pp. 21 23) The limits of cultural space were sociall y significant as people setting up tipis away from the main camp were establishing a different camp with different governance ( Walker 1982 pp. 23 25) More telling is that Lakota does have words for wilderness: makoskan (a place ing in
144 tiwokitahena (being away from houses, homeless) ( Buechel & Manhart 2002 ) ( 1933 ) apparent rejection of the wilderness concept was actually only a rejection of the dominant Euro American perception of wilderness at the (p. 38). Furthermore, numerous native examples of the wilderness concept from North, Central, and South America ( Bierhorst 1994 pp. 140 155) should be enough to dispel the myth that American Indians lacked a concept of wilderness ( Rolston 1991 ) Caveats and Conclusions One strength of the definition of wilderness presented here is that it is flexible enough to act as a uniform foundation through which cultural variations can begin to be understood. It does not define temporal or spatial limits of wilderness. Thus it leaves many question s for different cultures to answer differently. It may be clear that land ploughed and farmed every year is maintained in cultural space, but if abandoned, how long before it leaves cultural space and becomes wilderness? By this basic definition, abandon ed land would immediately leave cultural space and enter into wilderness, but different cultures likely have different time frames. In some cultures the transition may be immediate and in others take a few years or even a few centuries. In the interim, t he cemeteries (which are cultural areas but not the realm of the living ; ( Bierhorst 1994 p. 145) Compounded difficulty comes from not abandoning the land but letting it lie fallow. Do future plans to reuse land affect the perception of it s time to return to wilderness? If three years (for example) is enough for a return to wilderness once abandoned, is a field left fallow for four years also considered wilderness? While the answer will vary
145 from culture to culture, i t does however seem l ikely that typical fallow periods would be shorter than the required return time to wilderness for any particular culture. In other cases, a tree crop may be planted but not maintained or destroyed until harvest if at all. If five years (for example) is e nough for an abandoned plot to return to wilderness, does a 20 year rotation in pine harvest mean that an area of silviculture becomes a wilderness at some point before harvest if not maintained in any way? For example, a forest full of even aged monocult ures is considered a type of wilderness by Germans ( Imort 2005 ) but a tree farm by others ( O'Hara 2001 ) It is clear that under this definition of wilderness, which does not consider size, age, fragmentation, and disturbance, that all wilderness areas are not equiva lent. They may differ in biome/ecosystem type, level of wildness, biotic and conservation value, and cultural categorization (Table 5 1), all of which must be taken into consideration when establishing conservation plans Table 5 1 Examples of cultural wilderness categorie s Category Explanation Example Culture/Location Sources Archetypical One habitat type acting as an ideal cultural image of wilderness Forests Germany ( Worman 2010 ) Mountains Korea ( Worman 2010 ) No lands Buffer zones be tween hostile peoples Great Prussian Wilderness Teutonic Knights, Lithuanians, Poles ( van Vuure 2005 ) Demilitarized Zone Korea ( Kim 1997 ) Sacred lands Places with religious or spiritual meaning Rwenzori peaks Bukonzo, Uganda ( Stacey 2003 ) Sacred groves Western Ghats, India ( Gadgil & Vartak 1976 ) Holy hills Dai ( Xu et al. 2005 ) Wastelands Areas with little or no value esp. for development Badlands National Park South Dakota Savannas Tiris ( Plotkin 1993 )
146 This basal definition of wilderness can be constructively used even in the most disturbed regions because it embraces most of the landscapes in which conservationists are generally interested: everything from pristine natural to highly disturbed quasi natural ecosystems, from cand idates for preservation to candidates for restoration. However, like other wilderness definitions, the one proposed here cannot define most conservation goals. Goals such as ecosystem representation, function, productivity, native diversity, stability, e volvability, or preservation of a scientific baseline are not necessarily informed by any wilderness concept and some can be pursued at least partly in cultural areas. However, given that land use decisions appear to be influenced by cultural definitions of wilderness ( Worman 2010 ) it would be surprising if conservation decisions were not likewise influenced, even if ostensibl y purely scientific cross cultural conservationists put them aside when they are out of place and use more culturally appropriate wilderness definitions. Using local in terpretations of wilderness may help communicate the more easily verbalized scientific and practical benefits on a deeper emotional level. Additionally, conservation goals that explicitly include aesthetics and recreation do need to be informed by local a nd/or external wilderness ideals (depending on the targeted group). A common foundational starting point may facilitate the creation and explanation of both scientific and aesthetic goals that are acceptable to multiple groups, allowing conservation to b e tailored to local cultures and conditions. Locally adapting conservation initiatives to give more cultural and intellectual ownership to the people directly affected by them will likely encourage local support for and the ultimate success
147 of the initiat ives and should not be minimized or overlooked ( Lepp & Holland 2006 ) Additionally, the concept of wilderness appea ring to be culturally universal signals that wilderness is an assumed part of human world views. Conservation is simply an active extension and defence of that assumpti on.
148 CHAPTER 6 TROOPING FAIRIES, TROLLS, AND TALKING TIGERS: THE INFLUENCE OF TRADITIONAL WILDERNESS ARCHETYPES ON CURRENT LAND USE PATTERNS Introduction Wilderness is the landscape on the other side of the Great Divide that separates Human from Nature ( Nash 1967 ; Oelschlaeger 1991 ) It is the polar opposite of civilization, yet as the ever present Otherness, it plays a diversity of important roles within human cultures, including spiritual, psychological, mythical, and economic ( Nash 1967 ) subject of freq uent debates ( Callicott 2007 ; Callicott & Nelson 1998 ; Ereshefsky 2007 ; Nelson & Callicott 2008 ) made more contentious by the cultural, subcultural, and individual subjectivity of the rational and emotional perceptions and interpretations of wildernes s ( Nash 1967 ) Here I explore the idea that different cultures have different wilderness ar chetypes (a habitat that represents the ultimate or ideal wilderness to a culture) due to different external and internal forces, which may result in differing landscape use patterns and differing conservation priorities, concerns, and opportunities (Fig ur e 6 1). An inspection of the spiritual meanings of wilderness can provide an example of how diverse and flexible views of wilderness can be. In some cultures ( e.g. East Asian, Nash 1967 ; Zaramo Bantus, Sunseri 2005 ; and Mbuti Pygmies, Turnbull 1961 ) wilderness has traditionally been viewed as being a place of spirituality and holiness (though not necessarily free of spiritual danger), whereas in others, ( e.g. medieval European, ancient Hebrew, and early American; Nash 1967 ) wilderness was seen as
149 Figure 6 1. The functional relationships among a wilder ness archetype and the cultural and natural qualities that influence and are influenced by it. Only relationships discussed in the paper are included for simplicity. The color of the arrows indicates the typical nature of the interactions: white = positi ve, black = negative, cross hatched = positive and/or negative or more complex depending on the situation. For example, conservation has a limiting influence on development, but development has a complex influence on conservation (development destroys hab itats but by increasing the rarity of a habitat, makes conservation efforts more likely). Thick arrows indicate connections that are major themes dealt with in this paper. Thin arrows indicate connections that have minor mentions in this paper. The thic knesses of the arrows do not indicate strength or importance of the relationships. example, the dominant European/American view of wilderness shifted from largely fearful and hostile (pre Romantic), to idealized and sublime (Romantic), to recreational
150 and ecological (Modern) ( Imort 2005 ; Nash 1967 ) A diversity of views can also exist within a single culture at one point in time because of dif ferences among subcultures and individuals ( Nash 1967 ; Ross Bryant 2005 ) However, because of the coexistence of both negative qualities (dangers and isolation) and positive qualities (economic value and relief from the evils of society) inherent in wildernesses, any perception of wilderness is likely to be comp lex. In addition to the varied emotional and intellectual responses to wildernesses, the habitat that represents the physical ideal, or archetype, of wilderness also varies among cultures. For example, the ancient Hebrew desert wilderness archetype contr asts with the central European forest archetype ( Nash 1967 ) and the bush wilderness archetyp es found among various peoples of the savanna/scrub regions of Southern and East Africa ( Lee 1993 ; Burnett & wa Kang'ethe 1994 ) These wilderness archetypes are clearly based upon the dominant natural land cover in each area, but can likely be even the simplest landscape has a diversity of habitats, and cultures come into contact with multiple landscapes, interpretations of wilderness archetypes are probably also influenced by the negative qualities of each habitat type such as relative hostility (inclu ding dangers like large predators, diseases, and other people), resistance to development, and difficulty of travel. Religion, history, and amount of contact with and arche type. Additionally, habitat complexity promotes a multiplicity of interpretations: is a forested mountain a mountain or a forest? Is a swamp a forest or a wetland?
151 In spite of a diversity of views and archetypes, some generalizations can be made about wi lderness. Wilderness can be defined as any area that is outside the cultural sphere and therefore relatively free from the conscious high intensity maintenance to which the biotic communities within the cultural sphere are subjected (Chapter 5) Thusly d efined, wilderness is a very broad category that includes areas from the large, pristine, primordial ideal of wilderness to small scale, heavily disturbed secondary growth. Although the range of economic activities that take place in wilderness and the im portance of each naturally varies among economies and cultures, wilderness generally provides direct material economic value through wood/fuel cutting, hunting, fishing, gathering of medicinal, edible, or useful plants and fungi, foraging by domesticated a nimals, gathering wild honey, etc. Thus the relationships between wilderness and human societies tend to be extractive and even at times managerial but typically with limited input and control by the people. The positive values provided by wildernesses are mitigated by their real and perceived dangers (including wild animals, dangerous people, malevolent spirits, lack of assistance in an emergency, disease, etc. ; Blackbourn 2004 ; Gudermann 2005 ; Imort 2005 ; Nash 1967 ) which of course vary with location a nd culture. Additionally, because wildernesses are outside the cultural sphere and contain relatively few and mostly transient users, physical and psychological social controls are weak or absent ( Nash 1967 ) Hence wildernesses assume another role as havens for people rejected by, ejected from, or otherwise escaping from society ( e.g. thieves, murderers, witches, persecuted groups, vagabonds, etc. ; Blackbourn 2004 ; Radkau 2005 ; Sunseri 2005 ) So, although uses of wilderness are many and normal, individuals who spend too much
152 time in the wilderness away from respectable society are potentially viewed with suspicion. This situation presents a se rious risk of ostracism or scapegoating and likely synergizes with the more obvious dangers of wilderness in discouraging people from venturing too far from civilization when settling. The perception of dangers (both real and imagined), and thus the avoid ance of wilderness, is likely to be especially strong for archetypical wildernesses lending them a heightened level of protection from and resistance to development. This reluctance to develop archetypes on the part of individuals or governments need not be conscious or entirely rational given the strong emotions elicited by wildernesses and the importance of emotions in ostensibly rational cost benefit decisions ( Bechara 2004 ) If cultural archetypes of wilderness vary and archetypical wilderness areas are somewhat protected by the perception of their wildness, cultures with different wilderness archetypes might create culturally distinctive land use patterns reflecting their wi lderness archetypes (operating in concert with topographical, geological, and ecological considerations). Differences in patterns are likely most easily seen between cultures with wilderness archetypes defined by topography and those defined by land cover because terrain is much more resistant to change than land cover and therefore more likely to successfully check development. The hypothesis that the land use patterns of cultures vary predictably with their respective wilderness archetypes is here exam ined by first attempting to distill wilderness archetypes from the folktales of three cultures. Then, for the cultures with identifiable wilderness archetypes, the ruggedness of the cultural areas was compared with that of the wilderness areas within and between cultures. Additionally, elevational cross sections of the cultural wilderness boundary
153 were compared between cultures. Cultures with topographically defined wilderness archetypes should have stricter tolerances for ruggedness in developed areas a s well as civilization/wilderness transitions that have more predictable and consistent elevational cross sections. Methods Wilderness Archetypes To establish the cultural archetypes of wilderness, wilderness descriptions were gleaned from the folklore of three cultures: German, Irish, and Korean (see next section). These archetypes were then compared with wilderness types contained in the names of modern protected areas in each nation to establish the continuity and influence of the traditional archetype s. Though influenced by many other considerations, the naming of protected areas likely gives insight into how the areas are perceived by the managers and how they are intended to be perceived by the public, and thus should be biased towards cultural arch etypes of wilderness. The three cultures were chosen due to the availability of folklore, as well as an a priori expectation that the wilderness archetypes of these cultures differed. The possible terrain/habitat categories (cliff/canyon, forest, island, marine/coastal, meadow, mountain, river/lake, rock/cave, scrub, stream/pond, valley, and wetland) were defined by the data. moraines, and other highlands, as well as more distinctly montane te rrain such as summits, peaks, etc. san (mountain), which is used more inclusively than its English counterpart and is applied to relatively small hills as well as rocky summits (pers. obs.). Th is linguistic difference likely reflects a difference in cultural perceptions. Thus the mountain category of Ireland
154 and Germany is more inclusive than it would be otherwise, biasing their data towards mountain wilderness archetypes. Some geographical re ferences were counted as multiple geographical types for both the folktales and the protected areas ( e.g. is both a forest and a wetland). Folk tales The Zipes ( 2003 ) traditional German wilderness archetype. Data for the Irish archetype were taken from Yeats ( 1892 ) and Curtin ( Curtin 1890 ) ( 1982 ) and Ha ( 1969 ) Wilder nesses were defined as non anthropogenic places where something dangerous, magical, or supernatural happened to a human protagonist. In the cases that involve a structure that is clearly anthropogenic (whether magical: a candy house, or not: an abandoned cabin) but embedded in and isolated by a natural landscape, the natural landscape was considered a wilderness. To control for story length, each story only counted once for each wilderness type found therein. Stories that were clearly duplicates were onl y represented by the most complete version. Protected areas Wilderness types were gathered from protected areas that were intra national ( e.g. Biosphere Reserves were excluded), ecologically nonspecific ( e.g. Marine Protected Areas in South Korea were ex cluded), and not primarily for single taxon conservation ( e.g. North Korean National Monuments and Protected Sites in Ireland were excluded). Protected areas were limited to the national/federal and provincial/state levels to produce data sets comparable in size to those of the folktales (Table 6 1).
155 Table 6 1. The total number of folk tales and protected areas used from each nation with the number of wildernesses occurring in the folktales and terrain/habitat references in the protected area names. Ge rmany Ireland Korea Folk tales Total stories 279 104 212 Wildernesses 125 61 63 Protected areas Total names 113 120 68 Terrain/Habitat 88 117 62 Protected area names in Germany were taken from the Bundesamt fr Naturschutz website (the German Feder al Nature Conservation Agency; www.bfn.de ; accessed May, 2009). National Parks and Nature Parks were used for Germany. Protected area names in Ireland were taken from the websites of tSeirbhs Pirceanna Nisinta agus Fi adhlra (the Irish National Parks and Wildlife Service; www.npws.ie ; accessed May, 2009) and the Northern Ireland Environment Agency ( www.ni environment.gov.uk ; accessed May, 2 009). National Parks and Nature Reserves were used for Ireland and National Nature Reserves for Northern Ireland. Protected area names for both North and South Korea were taken from the World Database on Protected Areas ( w ww.wdpa.org ; accessed May, 2009) augmented with South Korean Provincial Parks from Park et al. ( 2000 ) National Park names were checked against the Guknipgong weon Seobiseu website (Korean National Park Service; http://english.knps.or.kr/ ; accessed September, 2009). National Parks, Nature Reserves, and Provincial Parks were used for South Korea. National Parks, Natural Reserves, Protected Areas, and Scenery Reserves were used f or North Korea. Both embedded and isolated geographical references in protected area names were counted ( e.g. Glen dalough and Correl Glen ) due to the synthetic structure of the names (especially in German and Korean). In contrast to the methods used for the
156 folktales, protected area names with repeated terrain/habitat references were allowed to count more than once in the same category ( e.g. Strom berg Heuchel berg or Slie ve Bloom Mountains sliabh uous geographical references ( e.g. lough using descriptions of the protected areas given by the respective government website and satellite imagery ( www.wpda.org ). Analy sis To determine if dominant wilderness archetypes exist in each of the three cultures ( i.e. if mentions of a particular wilderness type were unusually common), the data sets ( Ross 2003 ) If none were found, the wilderness mentions in folk tales were compared with those in protected area names rmally distributed. Land Use Patterns To address the hypothesis that patterns of land use co vary with wilderness archetypes, elevational data gathered from Google Earth 5.1 (available from http://earth.google.com/ ) w ere transformed into a ruggedness index (RI) to compare the irregularity of the civilized and wilderness landscapes of the nations with clear wilderness archetypes (Germany and Korea). To reduce the differences in overall topography of the two countries, only mountainous areas in Germany were used. Six mountainous German areas were delineated: Rothaargebirge, Harz, the Odenwald Spassart Rhon Thurgingian complex, the Black Forest, the Bavarian Forest Mountains, and the Northern Limestone Alps (Figure A C 1) Elevational transects perpendicular to wilderness civilization boundaries were also compared.
157 Government influences land use patterns and intranational cultural differences exist in both Korea and Germany ( Ch'oe 2000 ; Chaney 2005 ) This is simpler in Korea where traditional cultural differences and current political differences occur between the north and the south; however, preliminary analysis showed there to be no significa nt differences between either the civilized or wilderness ruggedness indices between North and South Korea. Thus, although different or more detailed methods may show differences between the two, for the purposes of this study, North and South Korea were unified. To avoid potential biases stemming from current agricultural differences, the same metrics were applied to city areas hemmed in by mountains. However, because there are no large German cities in mountainous areas, the five largest urban areas in Korea (Seoul Incheon, Busan, Daegu, Gwangju, and Daejeon; Cox 200 8 ) were compared with Sao Paulo, Mexico City, Los Angeles, Istanbul, and Rio de Janeiro, the five largest urban areas hemmed in by mountains in the rest of the world ( Cox 2008 ) excluding Japan and China as they are culturally similar to Korea. The sampling of the urban areas included the areas of contiguous urban development and the surrounding and included areas of wilderness (Figure A C 2). The sampling area delineation generally followed the ridgeline nearest to the urban development while excluding agricultural land and non contiguous urban areas and keeping t he delineation as simple as possible. Small areas of agricultural land still included were not sampled. Ruggedness index The RI was based on a hexagonal plot with 7 elevation sampling points (the center and the 6 corners), creating 12 equidistant neighbor relationships. The absolute values
158 of the differences in elevation between neighboring points were summed to create the index. Random locations were sampled until 50 civilized plots and 50 wilderness plots were accumulated. Plots were excluded if the l and use was not easily defined ( e.g. obscured by cloud cover or low photographic resolution). Plots were considered to be crop fields, improved pasture, orchards, etc they were not intensively maintained (natural and managed forests, meadows, grasslands, wetlands, etc. ). Korean tombs, which are generally scattered individually in forested mountains and consciously placed in wilderness, were allowed in wilderness plots. Though Korean Buddhist temples are also purposely placed in the wilderness, they are significant human structures (often with accompanying gardens) and, like hunting cabins, can be hard to distinguish from ho mes. Thus wilderness plots with any buildings or mines (which can alter the topography) were excluded. Roads and watercourses, along with their rights of ways and narrow riparian zones, were allowed in either type of plot because of their linear nature a nd their presence in both types of and not used for analysis. Three spatial scales were examined: 250 m, 125 m, and 25 m hexagons (opposite corner to corner distances). These scales run from the scale of villages down to the scale of individual houses. The same center points were used for each of the scales. However as the plot size decreased, mixed plots shifted into the civilized or wilderness category and sampling was cea sed once 50 civilized and 50 wilderness plots were measured. For the urban comparisons, only the 125 m scale was used. This scale was
159 small enough to sample the limited islands of wilderness in the urban areas and also most clearly showed land use differ ences. Logistic regressions were applied to the RI profiles to examine the application of land use decision rules. Inverse predictions were made to determine the ruggedness indices corresponding to 5%, 50%, and 95% probabilities of being civilized. 95% ( 1954 ) method and compared between the German and Korean predictions at each scale. However, some civilization inverse predictions could not be established. In these cases, the inverse predictions parisons, the 50% inverse predictions and the ranges between the 5% and 95% inverse predictions were compared with t tests. It is expected that cultures with topographic based wilderness archetypes will have better fitted logistical regressions (higher l og likelihood differences from the null, higher r 2 and greater areas under the Receiver Operating Characteristic (ROC) curves), lower 50% inverse predictions, and a smaller spread between the 95% and 5% inverse predictions than cultures with land cover ba sed wilderness archetypes. However, as rugged topography has practical ramifications for all human activities, ruggedness should remain an important predictor of civilization across all cultures. Border transects To establish the characteristic topography of civilization wilderness boundaries, transects perpendicular to the boundaries were established at the nearest clear, simple boundary to the first 50 plots in each area. Here again, three spatial scales were used: 1500 m, 750 m, and 150 m. Transects o f these lengths sample the equivalent of six
160 ruggedness plots at each of the three scales, measuring three in civilization before was sampled at 31 points (the boundary and 15 points in each landscape category). For the urban comparisons, only the 750 m scale was used and transect centers were required to be at least 100 m apart. Elevational data were taken relative to the mid point of the transect (the boundary betwe en wilderness and cultural landscapes). Boundaries created by cultures whose archetypical wilderness is based on topography rather than plant community likely run more perpendicular to the slope of the land, thus the transects should show maximized elevat ional gradients. In contrast, the boundaries created by cultures with land cover based wilderness archetypes are likely to be less perpendicular to the topographic slope, meaning that the elevational gradient measured along a transect perpendicular to the boundary may be somewhat lower than the true slope. Additionally, plant community based wilderness archetypes should lead to greater variation in relative elevation among transects with transect segments on either side of the civilization wilderness bord er both above and below the border elevation. The presence of positive and negative relative elevations should further depress the elevational means in these cultures. The absolute values of the coefficients of variation ( |CV| ) were used to combine the v ariability and means of the sampling positions along the transects. Using boundary elevation regardless of whether they were above or below. Thus the |CV| topographical ly defined border transects were expected to be lower (lower standard deviation and higher mean) than the |CV| of vegetation defined transects. To detect
161 differences between the |CV| s were established for the |CV| boundary transects using bootstrapping with 1,000 replicates and the percentile method ( Efron 1981 ) The relativ e elevational transect |CV| profiles of Korean and non Korean cities were compared using repeated measures MANOVA after the data were log transformed using bootstrapping w ith 1,000 replicates and the percentile method ( Efron 1981 ) Results Wilderness Archetypes German and Korean folk tales each showed a clear wilderness archetype. German folk tales had (PCL) of 59.9. Korean folk tales had 54 mentions of mountains, which exceeded the PCL of 40.0. Both German and Korean protected area names exhibited the same wilderness archetype as thei r folk tales (German forest = 31, PCL = 29.6; Korean mountains = 48, PCL = 27.2; Fig ure 6 2 A,C ). No wilderness type exceeded the PCL for either Irish folk tales or Irish protected area names (Fig ure 6 2 B ). However, wilderness types mentioned in Irish folk tales and protected area names were significantly correlated ( n = 12, r = 0.71, p = 0.01). Land Use Patterns Ruggedness index All the logistic regressions for transitions between civilization and wilderness were significant (Table 6 2). As expected, the l ogistic regressions for Korea (at all scales and including the urban sampling) tended to have greater log likelihood
162 Figure 6 2 Comparisons between the wilderness archetypes reflected in folk tales and the geographical references in protect ed area names for A ) Germany, B ) Ireland, and C Geographical categories exceeding their limit are outliers and considered categ ories are exaggerated in Germany and Ireland by using the Korean perception, lumping hilly terrain with truly mountainous terrain. differences from the null, higher r 2 values, and greater areas under ROC curve s than the non Korean data sets (Table 6 2) i ndicating that the transition from civilization to wilderness is more predictable based on ruggedness in Korea than in German mountains (Fig ure 6 3) and in Korean urban areas than in non Korean urban areas C B A Protected area PCL Folk tale PCL Korea Ireland Germany
163 (Fig ure 6 4). The exceptions to this general patt ern were Los Angeles and Seoul Incheon and Busan, which overlapped in some metrics. Table 6 2. Logistic regression results for the ruggedness index (RI) plots for Korea and Germany (at three scales) and the Korean and non Korean large urban areas. Neg. Log Area under Inverse prediction (m) d (95% CI e ) likelihood a r 2b ROC curve c 95% 50% 5% Korea 250 m 69.3 1.00 1.00 78 80 82 Germany M 250 m 32.3 0.47 0.90 ( 72) 7 (45) (110) 135 (168) (216) 263 (367) Korea 125 m 69.3 1.00 1.00 42 43 44 Germany M 125 m 29.4 0.42 0.89 ( 44) 1 (23) (60) 74 (91) (122) 147 (201) Korea 25 m 44.7 0.64 0.97 ( 3) 4 (7) (12) 14 (18) (21) 25 (34) Germany M 25 m 21.5 0.31 0.85 ( 14) 2 (4) (13) 16 (20) (28) 35 (48) S eoul Incheon 48.3 0.70 0.97 ( 7) 16 (28) (41) 50 (64) (69) 84 (121) Busan 50.2 0.72 0.98 ( 3) 36 (54) (71) 86 (102) (116) 136 (178) Daegu 69.3 1.00 1.00 42 43 43 Gwangju 69.3 1.00 1.00 40 41 41 Daejeon 58.0 0.84 0.99 ( 1) 2 6 (36) (41) 50 (66) (60) 74 (119) Sao Paulo 13.3 0.19 0.76 ( 119) 36 ( 6) (37) 51 (68) (105) 137 (229) Mexico City 26.7 0.39 0.91 ( 77) 27 ( 2) (41) 56 (75) (109) 138 (201) Los Angeles 50.4 0.73 0.97 ( 27) 13 (33) (5 5) 72 (97) (104) 132 (195) Istanbul 11.8 0.17 0.78 ( 131) 40 ( 8) (38) 53 (71) (112) 147 (246) Rio de Janeiro 22.3 0.32 0.80 ( 131) 52 ( 19) (36) 55 (80) (121) 161 (264) a The difference between the log likelihood of the logistic models and the null model (no relationship) b p < 0.0001 for all regressions c The area under the ROC (Receiver Operating Characteristic) curve indicates the ability of the model to correctly categorize a RI as civilized or wilderness. Perfect predictive ability = 1.00, no predictive ability = 0.50 d The inverse prediction gives the RI that results in a 95%, 50%, and 5% probability of being civilized. e Confidence Interval The inverse predictions from the logistic regressions followed expectations in the two larger scales (250 and 125 m plots) with the Korean 95%, 50%, and 5% civilization
164 Figure 6 3. Ruggedness indices of civilized and wilderness plots in Korea and German mountain areas at three scales (n = 50 for each category): A) 250 m plots, B) 125 m plots. C ) 25 m plots. Note the lack of overlap between the Korean plots but wide overlap between the German plots in A and B At the smallest scale ( C ), the pattern begins to break down. The data have been jittered on the horizontal axis to show overlapping points. Gray circles represent the 5%, 50%, and 95% logistic regression inverse predictions with 95% confidence intervals. C = Civilized plots; W = Wilderness plots probabilities occurring within a very narrow range (essentially at the same point) and those of the German mountains widely spread with non a very distinct transition from civilization to wilderness based on ruggedness in Korea that is lacking in Germany. Also, the Korean 50% probabilities w of the German mountain 50% probabilities (Fig ure 6 3 A and B ), indicating a higher tolerance for ruggedness in civilized areas in Germany. However, at the smallest scale (25 m), A B C
165 Figure 6 4 Ruggedness indices of civilized and w ilderness plots in Korean and non Korean large urban areas (125 m plots, n = 50 for each category). The data have been jittered on the horizontal axis to show overlapping points. Gray circles represent the 5%, 50%, and 95% logistic regression inverse pred ictions with 95% confidence intervals. C = Civilized plots; W = Wilderness plots; S I = Seoul Incheon; B = Busan; Dg = Daegu; G = Gwangju; Dj = Daejeon; SP = Sao Paulo; MC = Mexico City; LA = Los Angeles; I = Istanbul; RdJ = Rio de Janeiro the pattern di their counterparts in the German mountains (Fig ure 6 3 C ).
166 The ranges from the 95% to the 5% civilization probabilities were significantly smaller in Korean urban areas than in n on Korean urban areas ( K = 43.6, n K = 171.2, t = 5.16, p < 0.001) indicating that Korean urban areas have a sharper transition (Fig ure 6 4). Surprisingly, however, the 50% civilization probabilities were not different ( K = 53.8, n K = 57.3, t = 0.39, p = 0.355). Border transects Borders between wild and civilized areas were generally quite sharp and obvious. At all three scales (1500, 750, and 150 m), the Korean relative elevational border transects s howed S shaped profiles (Fig ure 6 5 A C ) while the German mountain ure 6 5 D F However, not all samp ling locations had non overlapping sampling locations; 750 m scale: 30/30; 150 m scale: 29/30; Fig ure 6 6). The Korean urban area border transects also showed S shaped profiles (Fig ure 6 7 A E ) while the non Korean urban area border transects were more symmetrical with unkempt bowtie profiles similar to those in Germany (Fig ure 6 7 F J ). The profile of the Rio de Janeiro transects is the most similar to that of the Korean transects; however, Rio de Janeiro differs from the Korean urban areas in having a concentration of flat transects, especially on the wilderness side (Fig ure 6 7 J ). The means of the Korean Korean urban areas (F = 15.1, p = 0.005). All of the Korean urban relative elevational border Korean urban areas and the 95% the means did not overlap (Figure 6 8).
167 Figure 6 5 Relative elevational transect s perpendicular to and centered on the civilization wilderness boundaries in Korea and German mountain areas at three different scales (n = 50 for each set): 1500m, 750 m, and 150 m. Note that Korean transects exhibit an S shaped profile while transects in the German mountains have a more symmetrical unkempt bowtie like profile. Transect position Transect position A D B E C F
168 Figure 6 6 Coefficients of Variation for the civilization wilderness boundaries in Korea and German mountains for three different transect lengths (1500 m, 750 m, and 15 0 m; n = 50 for each set). Symbols at the bottom of the graph indicate overlapping ( ) and non overlapping (*) confidence intervals at each sampling location for each scale. Discussion Wilderness Archetypes Both German and Korean folktales showed clear wilderness archetypes (forest and mountain respectively), but Irish folktales did not (see below). The forest is so recognized as the German archetype of wilderness that German culture has often been considered a product of the forests ( Chaney 2004 2005 ; Imort 2005 ; Mauch 2004 ; Radkau 2005 ; Williams 2005 ) In Germany, mountainous areas are limited while forests 750 1500 150
169 Figure 6 7. Elevational tran sects perpendicular to and centered on the civilization wilderness boundaries in Korean and non Korean large urban areas (750 m, n = 50 for each set). Note that transects from Korean urban areas exhibit an S shaped profile while those from non Korean urba n areas tend to have more symmetrical unkempt bowtie shaped profiles. A B C D F G H E I J
170 Figure 6 8. Coefficients of Variation for the relative elevational transects centered on the civilization wilderness boundaries in Korean and non Korean large urban areas (n = 50 for each set). The means and 95% confidence intervals of the two groups are also indicated. were widespread. Thus mountains would have been less important to the average German than the thick, often swampy forests which are considerable obstacles for farmin g, grazing livestock, and travel. These conditions help explain the growth of a landcover based wilderness archetype in Germany. On the other hand, Korea, while originally heavily forested (South Korea is still ~66% forested; ( Park et al. 2000 ) is also a largely mountainous peninsula, thus providing opportunities for both mountains and forest to serve as the pan Korean wilderness archetype.
171 There are several probable reasons for the clear dominance of mountains as the K orean wilderness archetype. Mountains and forests both present obstacles to development. However, while forests can be cleared, mountains are more resistant. Even labor intensive terracing of mountainsides does not eliminate the logistical difficulties caused by the overall slope. In addition to the practical characteristics of rugged terrain, mountains present very different psychological impacts than forests. Mountains visually dominate the surrounding landscape and confer commanding and awe inspiri ng views in ways forests cannot. The vistas provided by mountains are likely ( Oelschlaeger 1991 ) It seems likely that these qualities and the relative closeness of the summits to the sky (often associated with or home to important gods) contributes to the association of mountains with the supernatural in many cultures ( e.g. Burnett & wa Kang'ethe 1994 ) In addition to the general physical and psychological impacts of mounta ins, the agricultural economy of Korea has probably played an important part in the development of the mountain wilderness archetype. Rice has been grown in Korea since 2000 1500 B.C.E. ( Choe & Bale 2002 ) and has had a far greater cultural importance than would be predicted by the area of land actually under rice cultivation ( Lee 1994 ) This bias has likely been caused by a general preference for rice over other dietary staples, the greater profitability of rice as a crop, and the greater productivity and dependability of paddies over dry fields ( Lee 1994 ) Because of the need to flood rice paddies, topography presents a larger obstacle to rice agriculture than forests as carving rice paddies out of a mountainside necessitates not only the clearing
172 of the forest but the major engineeri ng feat of terracing. Even though other staple and minor crops can be grown in upland fields, the cultural dominance of rice probably influenced the rise of mountains as the Korean wilderness archetype. The development of a mountain wilderness archetype was plainly not influenced by any technological or logistical inability to build in the mountains as evidenced by the preferential construction of temples (historically and currently) and large fortresses (historically) in the mountains. These development s are isolated from the rest of the cultural sphere by wilderness, which can be desirable for defensive or spiritual reasons and are clear exceptions to normal land use patterns. Small farms and settlements can also be found in the mountains. However, th e discovery of derelict fields and homes only in the mountains and not in the lowlands (through the sampling during this study and pers. obs.) indicates that such locations are likely considered marginal and not rapidly reclaimed once abandoned. Unlike G ermany and Korea, Ireland did not show any clear wilderness archetypes. Based on the historic importance of cattle and dry field agriculture in Ireland, as well as the original heavily forested landscape with large wetland areas ( Raftery 1994 ) it was expected that an Irish wilderness archetype would be forest, wetland, or both. Because Irish wetlands are common and associated with obstacles to travel, ancient religious significance, and extraordinary finds (Irish elk antlers, Gould 1974 ; ancient artifacts, Raftery 1994 ) they were expected to have particular significance. However, wetlands were ranked 9 th out of 12 wilderness types in fo lktales. Another unexpected observation in the Irish tales was the seemingly high number of anthropogenic places around which the magical or dangerous parts of the story centered without reference to a surrounding
173 natural landscape. Interestingly, most o f these locations were ancient raths and Neolithic tumuli, which were likely to have been viewed by the storytellers as having been built by the old gods or fairies, not by humans. The occurrences of habitat types in the protected area names of the three n ations reflected the wildernesses gleaned from their folktales. Even Ireland, which lacked a clearly defined wilderness archetype, had a significant correlation between the habitat types mentioned in folktales and in the protected area names. While these results suggest that traditional wilderness archetypes are persistent and pervasive and are reflected in the names of protected areas, other factors likely influence the naming of protected areas. First of all, it could be posited that a correlation sho uld exist between mentions of habitat in folktales and in protected area names under null conditions: there is no development of a wilderness archetype and habitat types appear in both the folktales and in protected area names simply in the proportion in which they exist within a given that traditional wilderness archetypes probably develop over much longer time scales than protected areas are gazetted and renamed, it is unlikely that null derived distributions would be well matched as they should reflect different time periods with different habitat proportions. Additionally, in the case of Germany, presence of a forest wilderness archetype is consciously recognized, wh ile in Korea the null model would predict forests and mountains to be almost equally represented as mountains are generally forested, which is clearly not the case.
174 Secondly, conservation status may impact naming decisions. For example, especially rare or endangered habitat types are likely to be disproportionately targeted for conservation and disproportionately mentioned in protected area names. This bias probably explains the high occurrence of wetlands in Irish protected area names as well as similar high occurrences of wetlands in the protected areas of some other nations ( e.g. France and Serbia; Worman unpub. data). Thirdly, areas may be purposefully excluded from protected area status predicated on the assumption that their inaccessibility is prote ction enough (and also hampers tourism). In this case, wilderness archetypes would be underrepresented due to the sharpened perception of their inaccessibility. For example, the Rwenzori Mountains of Uganda were passed over for national park status in th ( Willock 1964 ) due to their inacce ssibility. In spite of this, they, and other mountainous areas in Uganda, have subsequently been gazetted as national parks and are threatened by encroachment. different goals, both of which vary from country to country. For example, in the United States, Wilderness Areas and Wildlife Refuges have very different dominant habitats (mountains and rivers/lakes, respectively; Worman unpub. data) related to their different goals. Thus, different types of protected areas can bias the data if they have divergent goals. With these caveats in mind, the suggestion that protected area names reflect cultural wilderness archetypes is supported by the examination of some specific examples. In K orea, Naksan Provincial Park is a typical beach/coastal park with the
175 border running parallel to and enclosing the shoreline but not running far inland. This park, though, is named after a small nearby hill ( Naksan temple. Anothe r Korean example is Mun gyeongsaejae Provincial Park, a valley park surrounded by a high ridge. Though the ridgeline is not actually within the borders of the park, the name of the park focuses on the ridge and its height ( saejae reportedly implies that the ridge is an obstacle even to birds; Jeong S. Y., pers. com.). Such is the strength of the Korean wilderness archetype and its spiritual and recreational significance (as evidenced by the placement of temples and tombs in the mountains an d the throngs of weekend mountaineers; pers. obs.) that protected areas are sometimes branded with montane monikers even when they protect entirely different landscapes. Similar cultural importance is evident in the frequent use of the German wilderness a rchetype in attempts to unify a frequently divided people with a common binding forest mythology ( Chaney 2004 ; Imort 2005 ) Land Use Patterns Ruggedness index The importance of topography on human settlement patterns in general is clearly shown by the significance of all the RI l ogistic regressions. It is interesting, however, that finer differences are evident between cultures. At the largest two scales, the RI shows that Korea has very clear and sharp transitions between civilization and wilderness, while Germany has ranges of RI shared by both civilization and wilderness. There are also indications that the 50% transition point in Germany is higher than in Korea. It appears that Korea is less likely to develop mountains than is Germany. Both of these results were expected a s a deforested mountain is still a mountain but no longer a forest and thus still fits into the Korean, but not the German, archetype of
176 wilderness. If wilderness is perceived to be inappropriate for settlement by people in general, it would explain the a pparent Korean reluctance to develop rugged areas. The loss of differences between the Korean and German profiles at the smallest scale (25 m) appears to be due to plots that contained both civilization and wilderness (mixed plots) at the two larger scales shifting into the civilization category as the plot size was reduced. This is evidenced by the fact that these plots in Korea, ranging from RI 0 m to 24 m, were the only civilized plots at that scale above RI 8 m (Fig ure 6 3). In comparison, the same ty pe of plot in Germany ranged from RI 7 m to 24 m with plots that had been civilized at both larger scales ranging from RI 0 m to 24 m and with plots that had been mixed at the largest scale but civilized at the medium scale ranging from RI 3 m to 28 m. Th at, in Korea, these more rugged plots only appear in the civilized column at the smallest scale indicates that development in rugged areas is smaller and patchier than in Germany. The comparisons between Korean urban areas and large non East Asian urban ar eas show patterns similar to the Korean German comparisons. Again, Korea shows much sharper transitions between the ruggedness of civilization and wilderness with smaller ranges of overlap than non Korean urban areas (Fig ure 6 4). The Korean urban areas (like Korea in general) showed an almost total lack of any relatively flat wilderness plots (a single coastal wetland plot near Seoul Incheon was the only exception and the only wetland sampled in Korea). The scarcity of wetland wildernesses is probably d ue to a combination of natural Korean topography and the value of wetlands as areas for rice farming, as well as the intensive use of low lying areas due to the reluctance to develop mountains.
177 The typical Korean ruggedness profile at the medium scale (b oth for Korea in general and the urban areas) shows an RI cutoff of around 40 m, below which there are almost entirely civilized plots and above which there are almost entirely wilderness plots (Fig ure 6 3 B and 6 4). Busan appears to be the most divergent with civilized plots continuing up to an RI of 145 m and the wilderness plots beginning at an RI of 59 m. port city in a fairly rugged area, tied spatially and econom ically to its harbor but without much flatland to develop. Thus, as Busan grows, the extreme scarcity of flatland likely causes development to gradually push up the sides of the mountains. The generality of this pattern is supported by the observation th at the outlier from the civilized plots in Seoul Incheon (at RI 91 m) is a new apartment complex built on the side of a mountain, apparently on previously undeveloped land and not an older established neighborhood or farmland judging from the surrounding a nd encompassed vegetation. For the non Korean urban areas, the RI profile of Los Angeles appears to be most similar to those of the Korean urban areas (Table 6 2, Fig ure 6 4). This similarity is most likely not due to any unwillingness on the part of Los Angelinos to develop rugged areas. On the contrary, hill and mountainside homes are in high demand, in spite of the dangers of earthquakes, fires, and debris flows in the area ( McPhee 1989 ) Rather, the similarity between the Los Angeles ruggedness profile and those of the Korean urban areas seems to be an artifact of sampling. The rugged areas in Los Angeles are extremel y limited compared to the extent of the total urban area, which is largely flat. Thus, while development in rugged areas may be preferred (though costly) for some
178 uses, few of the plots fell in civilized mountainous areas because of the limited proportion of mountainous terrain within Los Angeles itself. Interestingly, though the ranges encompassed by the 95% and 5% civilization probabilities are larger in the non Korean urban areas, the 50% civilization probabilities are not different. These results indi cate that there may be similar decision rules across cultures with regards to ruggedness and development. However, Korea, with its mountain wilderness archetype, appears to follow the rule more strictly than other cultures. Border transects The relative e levational civilization wilderness border transects for Korea at all three scales and for Korean urban areas showed consistent S shaped profiles with the civilized half of the transects flat and slightly below the elevation of the border and the wilderness half curving above the border elevation (Fig ures 6 5 and 6 7). The German and non Korean urban areas, on the other hand, showed more symmetrical profiles on both axes, giving the impression of more or less unkempt bowties. However, here too the civiliza tion side tends to appear flatter and the wilderness side more positive, indicating, like the ruggedness indices, the universal impact of topography on settlement patterns. ) lower in Korea and Korean urban areas than in German mountains and non Korean urban areas, respectively (Fig ures 6 6 and 6 8). It is notable that this metric apparently discriminates better than the RI, with Los Angeles, Busan, and Seoul well contained w ithin their respective groups. Additionally, Busan, which had high ruggedness limits
179 elevational transect profile (Fig ure 6 7 B ), it can be seen that the civilized half of the transects often dip farther below the civilization wilderness border than typical for Korean cities, suggesting again that development in Busan has pushed up the mountainsides, probably due to severe crowding of the flatlands. There is much more varia tion among the |CV| profiles of the non Korean urban areas (Fig ure 6 8). This is not surprising as they are drawn from different geographic and cultural areas around the world. It is interesting to note that the Rio de Janeiro |CV| profile dips closer to the Korean urban areas than its own grouping, although distinctly different from the Korean urban areas based on the ruggedness indices (Table 6 2, Figure 6 4). This similarity is due to the lack of developed hills in the Rio de Janeiro transects. Howev er, the differences between Rio de Janeiro and the Korean urban areas (the slightly higher |CV| profile and ruggedness indices) are due to the flat wetland wildernesses in Rio de Janeiro (Fig ure 6 7 J ). Conclusions and Conservation Implications The differen ces between the Korean and German land use patterns are consistent with the hypothesis that they are influenced by wilderness archetypes. The null expectation that apparent differences in land use patterns are caused by nations with higher population densi ties being forced to develop topographically undesirable areas predicts the opposite of what was observed here. Korea has a higher population density (South: 478 people/km 2 North: 195 people/km 2 total: 322 people/km 2 ) than does Germany (231 people/km 2 ) ( UN 2008 ) yet its land use is guided more by topography, not less, indicating the potential importance of cultural details like wilderness archetypes.
180 This study raises se veral questions about the nature of wilderness archetypes: how common, specific, strong, persistent, and intraculturally variable are wilderness archetypes? There are likely other types beyond the two broad categories examined here (topographic and land cover), such as aquatic or even seasonally or daily variable. It may appear that topographic wilderness archetypes would be more common than land cover archetypes because of the difficulties presented by rugged terrain; however, judging by the comparison s of the urban areas, topographic wilderness archetypes may actually be rarer than land cover archetypes. It seems likely that certain characteristics would increase the chances of the development of topographic wilderness archetypes in a culture: fairly rugged native terrain, an economic system strongly affected by terrain (such as rice farming), and ties with another culture having a topographic archetype. This is not to say that mountains are necessarily the only topographical wilderness archetype poss home terrain, perhaps providing safety from marauding enemies or flooding rivers or by providing better environmental conditions ( e.g. better for crops or livestock or fewer problem s with disease). In these cases, a wilderness archetype might develop in which the valleys or lowlands were the archetypical wilderness. When a wilderness archetype is present in a culture, the attitudes towards the wilderness and those in it (both negati ve and positive) are likely to have the effect of keeping the wilderness undeveloped and outside the cultural sphere. Negative attitudes ( e.g. fear, revulsion, and suspicion) will tend to decrease the probability of development in an archetypical wildern ess, while positive attitudes ( e.g. appreciation of beauty, recreation, ecosystem functions, and material value) will tend to increase the probability
181 of protection of an archetypical wilderness. A clear example of the latter occurred when the German emp ire gazetted forest reserves in Tanganyika. The areas protected and managed as forests were almost completely closed canopy montane forests but the much more widespread open miombo woodlands were not recognized as true forests and were thus excluded from forestry reserves ( Sunseri 2005 ) While a wildern ess archetype may somewhat protect a certain habitat or landscape type, it is likely at the expense of others as development is then concentrated in non archetypical wilderness areas. Decreased development of archetypical wildernesses is probably most ext reme in the cases of topographically defined wildernesses as it is much easier to alter land cover than elevation. For example, in Korea the land use pattern that has left the mountains largely untouched has concentrated development in valleys and along r ivers, and continues more recently in the so called reclamation of marine mudflats. In addition to the increased representation in the non anthropogenic landscape of the archetypical wilderness, the reluctance to develop an archetypical wilderness likely r esults in less fragmentation. Again, this is probably more pronounced with topographically defined wildernesses. An archetypical wilderness is likely to be devoured from the edges more so than other habitat types as people may generally be reluctant to s ettle deeply within it. However, the rate of development on any boundary is likely to be uneven (especially for land cover defined wildernesses), altering the degree of fragmentation. Korea and German mountains exhibit these patterns clearly: in Korea, f orest is rarely found isolated in patches as it follows the ridges of the mountains, but in Germany, forest patches occur more frequently as forest clearing
182 does not follow the terrain as closely (Fig ure 6 9). Furthermore, wilderness archetypes are like ly to be more spatially and temporally stable habitats as human use shifts through the landscape preferentially disturbing other habitat types. Because land cover wildernesses can be cleared but recover, the interplay between recovery and consumption may create shifting spatial patterns of wilderness/civilization interaction that cannot occur with topographical wildernesses. More stable and thus older archetypical wilderness communities could in turn strengthen the cultural archetype. Another layer of c omplexity is added by considering the cultural attitude towards wilderness. Cultures with positive attitudes towards wilderness should be more likely to allow or actively preserve patches of their archetypical wilderness in close proximity to settlements while cultures with negative attitudes towards wilderness should be less likely to tolerate patches of their wilderness archetype near settlements. Thus, on a larger scale, the existence of a wilderness archetype likely discourages development embedded in the wilderness, but on a smaller scale the attitudes towards the wilderness likely influence tolerance of archetype patches embedded in cultural landscapes. Incorporating local cultures can be an effective way of engaging communities in conservation progr ams through the celebration of positive cultural attitudes towards wildlife ( Kuriyan 2002 ) and the use of traditional methods of conflict resolution ( Lepp & Holland 2006 ) Wilderness archetypes provide another motivation to tailor conservation strategies to the local cultur e. It may be easier to introduce the concept and practice of protected area conservation if an archetypical wilderness is used as a flagship. Also, conservation is most likely to be successful when more traditional valuations of wilderness ( e.g. extract ive, aesthetic, national, and spiritual) are allowed to prosper
183 Figure 6 9: Google Earth images o f characteristically A) Korean and B) German mountain development patterns. Note how the Korean development is focused in the lowlands leaving the upland s largely forested (the exceptions being generally due to forest management for timber extraction and not development per se), whereas the German development is less bounded by topography. (Korean: Image NSPO 2009/Spot Image, 2009 ZENRIN, 2009 SK En ergy, Image 2009 DigitalGlobe; German: Image 2009 GeoContent, Image 2009 Aero West, 2009 Tele Atlas, 2009 Europa Technologies.) B A
184 alongside more recent valuations ( e.g. tourism, ecosystem services, and biodiversity), thus maintaining and strengthe ning the idea and worth of wilderness in the local culture. However, as suggested by this study, non archetypical wilderness areas may be in greater, more immediate need for protection as they are likely to be more threatened by development. Thus while w ilderness archetypes may be useful for introducing conservation, subsequent education about the benefits of conservation should specifically address the different values of different habitats emphasizing those in most danger. Unfortunately, increasing popu lation pressure and globalization are likely to speed cultural change and eventually break down traditional cultural and psychological barriers to development in archetypical wildernesses. These shifts could result in increasing development of previously avoided wilderness areas, necessitating a reevaluation of conservation priorities. Thus, wilderness archetypes are important to conservation not only because of their influence on past and current land use patterns and their worth in promoting conservatio n, but also because of their potential for change.
185 CHAPTER 7 CONCLUSION Understanding anthropogenic impacts on the biosphere is arguably among the most important and urgent tasks for biology today. S ynthesizing the impacts of thousands of different cu ltures with vastly different extractive strategies on countless complex, interconnected, and poorly understood ecological systems is daunting This study has suggested that one step towards such a synthetic understanding is provided by examining anthropog enic impacts as a broad and unified gradient and applying this general gradient and its effects to considering how a specific and understudied ecosystem process, animal mediated nutrient transfer, changes along this gradient. This permits one to conceptua lly examine the potential ramifications of anthropogenic change on one key aspect of nutrient flows. T hen t hrough the use of a model organism in a field system the real world impact of predicted anthropogenic changes was studied experimentally and observ ationally under field conditions Finally modifications to the general gradient of anthropogenic effects were suggested to arise based on cultural perceptions of wilderness and resulting land use differences This hypothesis was examined through a combi nation of approaches taken in anthropology and landscape ecology T he first part of t his study (Chapter 2) suggested that t here is a spatio temporal gradient of anthropogenic influence on biotic systems ranging from low density hunting, through pastoralism and the n the wide range of agricultural practices and densities, to higher density habitation from suburban to urban settlements. It was suggested that along t he gradient of anthropogenic influence there are impacts on animal mediated nutrient transfer t hrough s everal main routes: maximum body size and home range
186 decrease s of animal consumers alterations in community change rate, appearances of novel resources, increases in the relative importance of generalists and combinations of these produce complex changes in the relative importance of top down vs. bottom up control. These changes likely influence animal mediated nutrient transport processes by altering their importance depending on trophic structure, shortening transport distances, increasing the importance of cross habitat nutrient transfers, and encouraging flow of nutrients into refuge habitat s. The increased eutrophication of small refuge fragments is likely to present another challenge to native species in human dominated landscapes. Feral H o gs were used as a model generalist omnivore species that has benefited tremendously from anthropogenic disturbance. In a multi habitat system located on the Florida Gulf Coast experimental manipulations (Chapter 3) revealed that hogs had differing impac ts on the herbaceous community in the hammock refuge habitat based on landscape patterns, specifically the juxtaposition of marsh feeding habitat. Hog densities were ~4 times higher in the island hammocks embedded in the marsh than in the mainland hammock s along the edge of the marsh likely due to greater access to feeding areas in the marsh conferred by the islands. Y et exclusion of hogs resulted in increased species richness and percent cover on the mainland but not the islands This result is suggest ed to likely be due to a combination of the extirpation of hog vulnerable plant s pecies from islands with sustained high hog densities combined with the difficulty of dispersing to the islands from mainland populations. Thus the heavy consumptive and dis turbance pressure by hogs has likely created a disturbance resistant herbaceous community on the islands by filtering out vulnerable species. Herbaceous biomass, on
187 the other hand, increased in similar amounts in bot h areas when hogs were excluded, which indicates that even the disturbance resistant communities are under constant pressure. It is unlikely that the influence of the hogs will be reduced without control of their populations and movement patterns by a top predator, such as the Florida Panther, a controlling presence that is currently absent in the system. This study thus suggests that Feral Hogs in Florida are likely having strong impacts on the structure and species composition of at least some plant communities The nutrient study (Chapter 4) supported the hypotheses that generalist consumer s can mov e nutrients short distances from feeding habitats and concentrat e them in refuge habitats. The only important food e stablished by fecal analysis and confirmed by stable isotope analysis is bulrush from well defined patches in the marsh, which is thus the principal source of nutrients transported by the Feral Hogs. The fact that the coastal hog diet is composed almost entirely of bulrush and the observation that the bulrush areas are the only marsh community the hogs root and the rooted patches of the marsh are the only place where bulrush grows implies a tight ecological mutualism between the Feral Hogs and the American Bulrush P opulations of both species likely benefit and reach higher densities than they would alone. Because of the impact of the hogs on the hammock community and the unique nature of the bulrush patches, this mutualism is likely central to the functioning of this entire system. The potential existence of this mutualism was an un anticipated finding of the study. The islands, with the greater density of hogs than the nearby mainland are thus the destination of the majority of hog mediated nutrient deposition The islands have
188 soils richer in P, K and N than the mainland soils p otentially indicating a sustained movement of nutrients by hogs from marshes into the terrestrial island soils Artificial tilling to mimic rooting disturbance did not have a detectable lasting effect on plant tissue nutrient content but exclusion of the hogs quickly reduced the nutrient content of the island biomass to mainland levels. Exclusion of hogs on the mainland did not affect levels of nutrients in the plant tissue. Even though the hog nutrient deposition is by some measures modest ( 1 11% of th e atmospheric deposition in this area ) the se experiments demonstrate that hog deposition is biologically significant because the concentrations of nutrients in the herbaceous biomass depend on differences in hog abundance and access Comparisons with typ ical plant nutrient levels reveal that these N and P are likely to be limiting in this system but K is abundant in both the soils and the plant tissues. Thus, the study reveals that hogs are able to indirectly alter plant nutrient content. The nutrient transport by the hogs appears to be fairly short range, less than 300 m with movement of nutrients to the closer islands surrounded by marsh far outweighing movement to the slightly more distant mainland (400 500 m). Based on the favoritism the hogs show the islands compared to the mainland, and given the stable isotope result s and color phenotypes, there seems to be very little long distance movement of the hogs from the coast inland and therefore nutrient transport distances are short. Indeed, the hog s ubpopulation that occupies the coast seems quite distinct from that found even a short distance inland. Therefore, t he Feral Hogs acting as exotic generalists move nutrients as predicted based on general anthropogenic patterns of disturbance: short range, inter habitat transport with net flows into refuge areas from
189 feeding areas but with little reason to expect large scale, long distance nutrient transport All of these ecological processes play out in landscapes modified by human impacts. In the Florid a system, Feral Hogs are an introduced exotic freed from the potential regulation of their numbers by an almost extirpated top predator, the Florida Panther. However, it is important to recognize that there is a diversity of human cultures and this divers ity impacts their relationships with ecosystems and influences to a greater or lesser extent the overall patterns of anthropogenic impacts on biotic systems. The final portion of this dissertation has explored the cultural context of anthropogenic impacts A root of the diversity of human interactions with their division between H uman and N ature and therefore between wilderness and cultural space is fundamental to the way people perceive r eality. Though the character of that division and its universality in human cultures has been long debated, the division is inherent to self awareness as a species and the creation of cultural space. In spite of differing emotions towards, emphasis on, a nd valuations of wilderness among cultures, diverse examples suggest that there is a basal universal wilderness concept but with substantial and consequential variation among its many incarnations This concept of course varies among cultures and one ob vious aspect that differs is the archetypical wilderness envisioned by a people An archetypical wilderness is one viewed in a given culture as the ideal or ultimate wilderness. Readings of German, Irish, and Korean folktales show that, traditionally, Ge rman culture view ed the forest as the archetypical wilderness and Korean culture view ed mountains as the archetypical
190 wilderness while no wilderness archetype could be detected for Irish culture. An analysis of protected area names in the same three cul tures showed similar patterns with German protected areas having largely forest derived names, Korean protected areas having mountain derived names, and the prevalence of habitat types in Irish protected area names being correlated with those in Irish folk tales. The fact that the German archetype was based on land cover and the Korean archetype was based on topography led to very specific expectations for the landscape use patterns for each culture. Because deep wilderness should be avoided for developme nt, an archetypical wilderness should be especially avoided and developed from the edges. Because land cover is easily changed compared to topography, topography should more severely delimit Korean cultural areas than German ones. Landscape level analyses were carried out to test this hypothesis. The expectation was satisfied in the l and use patterns in Korea and Germany with respect to ruggedness and relative elevation: Korea had a more conservative and stricter decision rule with regards to development of rugged areas and more predictable elevational transitions from civilization to wilderness, consistent with a reluctance to develop a mountain wilderness archetype. Of course, Korea and Germany have very different agricultural systems, with topography being much more important in the type of wet paddy agriculture favored in Korea. Therefore, urban development was examined with respect to the same ruggedness index and elevational transects. Because Germany does not have large cities in mountainous area s, the largest non East Asian cities were used from around the world.
191 K orean cities had similar patterns to all of Korea while non East Asian cities hemmed in by mountains were variable but similar to Germany, implying that c ulture, not agriculture plays a major role in patterns of land development. This is particularly powerful in showing the impact of wilderness archetypes as a culture that needs flat agricultural land can be expected to be less likely to build cities in the valuable lowlands and instead use uplands for urban development. This is not what is observed in Korea. Anthropology has a long history of inclusion of environmental variables (such as habitat type/biome) in understanding human cultures and their development. Ecology, for its part, while often recognizing the profound influences that humans often have on biotic communities and ecosystems, has tended to either ignore anthropogenic influences altogether or treat them as givens without examining humanity as an ecological force in and of itself with broad rules and patterns that are followed in spite of the vast diversity of human cultures and impacts that can alter the details and thus some of the practical ramifications of anthropogenic influences on the biosphere. Unfortunately, remain ing ignorant of human behavior and patterns of impacts is no longer a realistic option for ecologists. On the one hand, there are practically no ecosystems available for study that are complet ely free from human influence. O n the other hand anthropogeni preservation of biodiversity and ecosystem functioning increasingly demands that anthropogenic impacts be understood and mitigated. While the Great Divide may be a primary character of human pe rception, in many ways it is being unintentionally broken down by the growing dominance of anthropogenic impacts and must be crossed
192 conceptually in the ecological and conservation sciences in order to deal with the effects of increasing human population d ensity and global biotic degradation.
193 APPENDIX A SUPPLEMENTARY INFORM ATION FOR CHAPTER 2 Although larger ve rtebrates form the focus of Chapter 2 many of the anthropogenic community effects discussed in the paper have been shown in a diverse range of taxa For example: Plants Generalist plants have increased over time in British upland habitats because of anthropomorphic nutrient enrichment and disturbance ( Smart et al. 2005 ) Generalist competitive (i.e., those that are dominant in high nutrient areas) plant species have increased with time in a Belgium urban area at the expense of specialist stress tolerators ( Van der Veken et al. 2004 ) Invertebrates Anthropogenic simplification of seafloor habitats impacts habitat specialists more than generalists ( Thrush et al. 2006 ) Insects Fragmentation had negligible effects on insect habitat generalists ( Bender et al. 1998 ) Narrow habitat specialist and dietary specialist insects are especially prone to anthropogenic extinction ( Dunn 2005 ) Fragmentation and urbanization favor generalist insects and especially reduce predator and parasitoid insects ( Gibb & Hochuli 2002 ) Omnivorous insects are the least sensitive to a desert urban gradient ( McIntyre et al. 2001 ) Carabid beetle specialists get smaller and the generalists more common moving towards the urban end of an urban rural gradient ( Ishitani et al. 2003 ) Butterfly generalists do better with the novel foods and changes in urban areas than specialists ( New & Sands 2003 ) Ants are the dominant predator in cities but vertebrates are the dominant predators in rural areas ( Wetterer 1999 ) Spillover of insect generalist predators from agricultural lands to wildlands can ( Rand & Louda 2006 )
194 Generalist insect predators get more abundant and more important in impacted landscapes but specialists are reduced ( Rand & Tscharntke 2007 ) The number of generalist butterfly species actually increased in smaller fragments (perhaps because they were drawn i n from the matrix) while specialist species increased with habitat patch size ( Steffan Dewenter & Tscharntke 2000 ) A simplification of insect food web s with fewer interactions dominating the webs in modified habitats even though there was a retention of species diversity, however the most common parasitoid was more specialized in modified habitats (maybe because its preferred prey is more abundant so wa s more focused on than expected) ( Tylianakis et al. 2007 ) Smaller fragments h ave fewer monophagous insects ( Zabel & Tscharntke 1998 ) Generalist bee and wasp species were not sensitive to urbanization ( Zanette et al. 2005 ) Crustaceans Generalist exotic crustaceans dominate streams in developed areas ( Brasher et al. 2006 ) Fish Generalist exotic freshwater fish dominate streams in developed areas ( Brasher et al. 2006 ) Reef fish generalists were more successful after coral disturbance than specialists ( Feary 2007 ) Reptiles Reptile habitat specialists were more likely to go extinct on islands ( Foufopoulos & Ives 1999 ) But see Attum et al. ( 2006 ) : lizard desert specialists are favored compared to generalist with human desertification becaus e they are adapted to the desert extremes that are exacerbated by desertification. Mammals Fragmentation had negligible effects on mammal habitat generalists ( Bender et al. 1998 ) There is an increase in generalist mammalian mesopredators, far more exotics in cities, and ants are the dominant predator in cities but vertebrates are the dominant predators in rural areas ( Wetterer 1999 )
195 Raccoon Dogs in agricultural areas have lower activity levels than mountain individuals and individuals using concentrated resources like garbage and shoreline have s mall home ranges ( Saeki et al. 2007 ) Birds Fragmenta tion had negligible effects on bird habitat generalists ( Bender et al. 1998 ) Exotic and generalist birds are more common in disturbed areas ( Acevedo & Restrepo 20 08 ) Urban birds have broader environmental tolerances ( Bonier et al. 2007 ) High human density rural areas have more habitat generalist bird species because of less native habitat, more roads, and more fragmentation ( Boren et al. 1999 ) In France, birds responded more negatively to fragmentation and disturbance the more habitat specialized they were ( Devictor et al. 2008 ) Increase in exurban density favored some generalist and reduced specialist birds ( Fraterrigo & Wiens 2005 ) Generalist avian nest predators are more common in urban areas ( Jokimki 2000 ) Specialist birds are negatively as sociated and generalists positively associated with forest edges, showing the effect of fragmentation ( Jones et al. 2000 ) Crows and ravens have smaller home ranges and higher reprod uction with access ( Marzluff & Neatherlin 2006 ) Generalist birds increase at the expense of specialists through time in an urban park ( Bergen et al. 1998 )
196 APPENDIX B SUPPLEMENTARY FIGURE S FOR CHAPTER 3 Figure AB 1. An example of the exclosures used in the study. The exclosures are 4 m fencing and supported with metal U posts, wooden stakes, and convenient trees.
197 Figure AB 2 Collection of the herbaceous biomass. A 1 m x 1 m quadrat was placed randomly in one of the 16 square meters contained within each plot ( A ). All livi ng herbaceous biomass was collected within the quadrats ( B ) with the exception of woody stems and Poison Ivy. A B
198 Figure AB 3 The effects of hog exclusion on the herbaceous community of the coastal hammock in the Lower Suwannee National Wildlife Refuge af ter 3.5 years were frequently visually striking.
199 APPENDIX C FECAL DECOMPOSITION STUDY FROM CHAPTER 4 An attempt was made to estimate nutrient deposition based on fecal deposition rate based on an observed fecal decay rate. Unfortunately, this method prod uced unreasonably low estimates of hog density (on the order of 1/100 km 2 ) and nutrient deposition (below). This failure could have been due to a bias in fresh feces found and monitored. While one specimen was noted to decay beyond recognizability in a s ingle day, most specimens were likely found after several days, biasing the subsequent results by excluding quickly decomposing feces. Additionally, the sample size was necessarily small as fresh feces were in high demand for genetic work. The details ar e included here for completeness. Methods The rate of fecal decay was established by marking fresh piles and noting when they had decayed beyond recognition. The date of decay was defined to be the midpoint between the last positive observation and the ob servation of disappearance. Because there was no evidence of change in fecal load among seasons, the deposition rate was considered to be equal to the decay rate. Estimated fecal nutrient deposition rates were estimated using the observed fecal decay rat mainland, the measured fecal nutrient content, and the measured weight of the fecal samples. Because nutrient deposition occurs through urine in addition to feces, N, P, and K depositio n through urine was estimated based on calculated fecal nutrient deposition and the proportions of N, P, and K in domestic hog nutrient budgets excreted in the feces and in urine ( Kephart & Sherritt 1990 ; Sands et al. 2001 )
200 Results Fecal dec ay had a linear function with 50% decay at 471 days (Figure A C 1), which equates to a deposition rate of 3.4 fecal piles/100 m 2 /year on the islands and 0.5 fecal piles/100 m 2 /year on the mainland. The average weight of a fecal pile was 122 g 12 SE (n = 3 9). The average N, P, and K content of hog feces, and estimated nutrient deposition rates for the islands and coastal mainland are given in Table A C 1. Figure A C 1. Feral Hog fecal decay in the coastal hammock of the Lower Suwannee National Wildlife Re fuge (n = 14). The trend line is dashed.
201 Table A C 1 Calculated nutrient (N, P, K) deposition rates on the island and mainland coastal areas by Feral Hogs in the Lower Suwannee National Wildlife Refuge. % of Fecal deposition (g/100 m 2 /year) Urine:Fece s Urine deposition (g/100 m 2 /year) Total deposition (g/100 m 2 /year) Nutrient feces 1 Island Mainland ratio 2 Island Mainland Island Mainland N 0.42 0.04 3 1.75 0.73 0.28 0.12 2.75 0.45 4.80 2.75 0.77 0.45 6.55 2.85 1.05 0.46 P 0.17 0.02 0.70 0.3 0 0.11 0.05 0.54 0.09 0.38 0.19 0.06 0.03 1.08 0.35 0.17 0.06 K 0.14 0.01 0.56 0.23 0.09 0.04 3.25 0.16 1.83 0.76 0.30 0.12 2.40 0.80 0.39 0.13 1 By wet weight 2 The ratios of the amounts of the nutrients excreted through urine and feces base d on hog nutrient budgets ( Kephart & She rritt 1990 ; Sands et al. 2001 ) 3 Standard error
202 APPENDIX D DELINEATIONS OF SAMPLING AREAS FOR CHAPTER 6 Figure A D 1. The mountainous areas in Germany (out lined in white) used for the rug gedness indices and civilization wilderness border transects. ( 2009 Google, 2009 Tele Atlas, 2009 Europa Technologies, 2009 Transnavicom, Ltd.) Bavarian Forest Mountains Northern Limestone Alps Black Forest Odenwald Spassart Rhon Thurgingian Harz Rothaargebirge
203 A: Seoul Incheon B: Busan
204 C: Daegu D: Gwangju
205 E: Daejeon F: Sao Paulo
206 G: Mexico Ci ty H. Los Angeles
207 Figure A D 2: Urban areas used for comparisons of ruggedness indices and civilization wi lderness border transects. A) Seoul Incheon ( 2009 Cnes/Spot Image, Image 2009 TerraMetrics, Image 2009 DigitalGlobe), B) Busan ( 2009 Cnes/Spot Image, Image 2009 GeoEye, Image 2009 DigitalGlobe, Data 2009 MIRC/JHA), C) Daegu ( 2009 Cnes/Spot Image, Image 2009 DigitalGlobe), D) Gwangju ( 2009 Cnes/Spot Image, Image 2009 DigitalGlobe), E) Daejeon ( 2009 Cnes/Spot Image, Image 2009 DigitalGlobe), F) Sao Paulo (Image 2009 TerraMetrics, Image 2009 I: Istanbul J: Rio de Janeiro
208 DigitalGlobe, Image 2009 GeoEye), G) Mexico City (Image 2009 GeoEye, 2009 Cnes/Spot Image, Image 2009 DigitalGlobe), H) Los Angeles (Image U.S. Geological Survey, Image 2009 DigitalGlobe, Data U.S. Navy), I) Istanbul (Image 2009 DigitalGlobe, Data U.S. Navy, 2009 Cnes/Spot Image), and J) Rio de Janeiro (Image 2009 TerraMetrics, Image 2009 GeoEye, Image 2009 DigitalGlobe Data SIO, Data NOAA, Data U.S. Navy, Data NGA, Data GEBCO).
209 LIST OF REFERENCES Act on Wilderness Reserves. 1991. Finland Abrahamson, W G., and D. C. Hartnett. 1990. Pine flatwoods and dry prairies. Pages 103 149 in R. L. Myers, and J. J. Ewel, editors. Ecosystems of Florida. University of Central Florida Press, Orlando. Acevedo, M. A., and C. Restrepo. 2008. Land cover and land use chan ge and its contribution to the large scale organization of Puerto Rico's bird assemblages. Diversity and Distributions 14 :114 122. Adkins, R. N., and L. A. Harveson. 2007. Demographic and spatial characteristics of feral hogs in the Chihuahuan Desert, Texa s. Human -Wildlife Conflicts 1 :152 160. Ambrose, S. H. 1993. Isotopic analysis of paleodiets: Methodological and interpretive considerations. Pages 59 130 in M. K. Sandford, editor. Investigations of Ancient Human Tissue. Gordon and Breach Science Publishe rs, New York. Ambrose, S. H., J. Buikstra, and H. W. Krueger. 2003. Status and gender differences in diet at Mound 72, Cahokia, revealed by isotopic analysis of bone. Journal of Anthropological Archaeology 22 :217 226. Anderson, S. J., and C. P. Stone. 1993 Snaring to control Feral Pigs Sus scrofa in a remote Hawaiian rain forest. Biological Conservation 63 :195 201. Attum, O., P. Eason, G. Cobbs, and S. M. Baha El Din. 2006. Response of a desert lizard community to habitat degradation: do ideas about habita t specialists/generalists hold? Biological Conservation 133 :52 62. Baber, D. W., and B. E. Coblentz. 1986. Density, home range, habitat use, and reproduction in Feral Pigs on Santa Catalina Island. Journal of Mammalogy 67 :512 525. Barlow, J., T. A. Gardner A. C. Lees, L. Parry, and C. A. Peres. in press. How pristine are tropical forests? An ecological perspective on the pre Colombian human footprint in Amazonia and implications for contemporary conservation. Biological Conservation. Barnosky, A. D., P. L. Koch, R. S. Feranec, S. L. Wing, and A. B. Shabel. 2004. Assessing the causes of late Pleistocene extinctions on the continents. Science 306 :70 75. Baron, J. 1982. Effects of Feral Hogs ( Sus scrofa ) on the vegetation of Horn Island, Mississippi. The Ameri can Midland Naturalist 107 :202 205.
210 Bayet, F. 1994. Overturning the doctrine: Indigenous people and wilderness -being Aboriginal in the environmental movement. Social Alternatives 13 :27 32. Bechara, A. 2004. The role of emotion in decision making: Evidence from neurological patients with orbitofrontal damage. Brain and Cognition 55 :30 40. Bender, D. J., T. A. Contreras, and L. Fahrig. 1998. Habitat loss and population decline: A meta analysis of the patch size effect. Ecology 79 :517 533. Bergen, F., A. Schw erk, and M. Abs. 1998. Long term observation of the fauna of two man made nature habitats in cities of the Ruhr Valley area. Pages 618 622 in J. Brueste, H. Feldman, and O. Uhlmann, editors. Urban Ecology. Springer, Berlin. Bierhorst, J. 1994. The Way of t he Earth: Native America and the Environment. William Morrow and Co., New York. Billings, W. D. 1938. The structure and developement of old field shortleaf pine stands and certain associated physical properties of the soil. Ecological Monographs 8 :437 500. Billings, W. D. 1970. Plants, Man and the Ecosystem, 2nd ed. Wadsworth Publishing, Belmont, CA. Blackbourn, D. 2004. "Conquests from Barbarism": Taming Nature in Frederick the Great's Prussia. Pages 10 30 in C. Mauch, editor. Nature in German History. Ber ghahn Books, New York. Bonier, F., P. R. Martin, and J. C. Wingfield. 2007. Urban birds have broader environmental tolerance. Biology Letters 3 :670 673. Boren, J. C., D. M. Engle, M. W. Palmer, R. E. Masters, and T. Criner. 1999. Land use change effects on breeding bird community composition. Journal of Range Management 52 :420 430. Bouchard, S. S., and K. A. Bjorndal. 2000. Sea turtles as biological transporters of nutrients and energy from marine to terrestrial ecosystems. Ecology 81 :2305 2313. Bozek, C. K ., S. Prange, and S. D. Gehrt. 2007. The influence of anthropogenic resources on multi scale habitat selection by raccoons. Urban Ecosystems 10 :413 425. Brashares, J. S., L. R. Prugh, C. J. Stoner, and C. Epps, W. 2010. Ecological and conservation implicat ions of mesopredator release in J. Terborgh, and J. A. Estes, editors. Trophic Cascades: Predators, Prey, and the Changing Dynamics of Nature. Island Press, Washington, D.C.
211 Brasher, A. M. D., C. D. Luton, S. L. Goodbred, and R. H. Wolff. 2006. Invasion pa tterns along elevation and urbanization gradients in Hawaiian streams. Transactions of the American Fisheries Society 135 :1109 1129. Bratton, S. P. 1975. The effect of the European wild boar, Sus scrofa on gray beech forest in the Great Smoky Mountains. E cology 56 :1356 1366. Brown, C. S. 2007. An introduction to the problem of boundaries in ecological theory and practice. Pages ix xvi in C. S. Brown, and T. Toadvine, editors. Nature's Edge: Boundary Explorations in Theory and Practice. State University of New York Press, Albany. Buechel, E., and P. Manhart. 2002. wilderness. Page 530. Lakota Dictionary. University of Nebraska Press, Lincoln. Burnett, G. W., and K. wa Kang'ethe. 1994. Wilderness and the Bantu mind. Environmental Ethics 16 :145 160. Callicott, J. B. 2007. Lamarck redux: Temporal scale as the key to the boundary between the human and natural worlds. Pages 19 39 in C. S. Brown, and T. Toadvine, editors. Nature's Edge: Boundary Explorations in Theory and Practice. State University of New York Pres s, Albany. Callicott, J. B., and M. P. Nelson. 1998. Introduction. Pages 1 20 in J. B. Callicott, and M. P. Nelson, editors. The Great New Wilderness Debate. The University of Georgia Press, Athens, GA. Ch'oe, Y. h. 2000. Korea since 1945. Pages 367 451 in Y. h. Ch'oe, P. H. Lee, and W. T. De Bary, editors. Sources of Korean Tradition: Volume Two: From the Sixteenth to the Twentieth Centuries. Colombia University Press, New York. Chaney, S. 2004. For n ation and prosperity, health and a green environment: Protecting Nature in West Germany, 1945 1970. Pages 93 118 in C. Mauch, editor. Nature in German History. Berghahn Books, New York. Chaney, S. 2005. Protecting Nature in a Divided Nation: Conservation i n the two Germanys, 1945 1972. Pages 207 243 in T. Lekan and T. Zeller editors. Germany's Nature: Cultural Landscapes and Environmental History. Rutgers University Press, New Brunswick, NJ. Charlesworth, B. 2009. Effective population size and patterns of molecular evolution and variation. Nature Reviews 10 :195 205. Chastain, J. P., J. J. Camberato, J. E. Albrecht, and J. Adams. 1999. Swine manure production and nutrient content. Pages 1 17. South Carolina Confined Animal Manure Managers Certification Progr am. Clemson University, Clemson, SC.
212 Choe, C. P., and M. T. Bale. 2002. Current perspectives on settlement, subsistence, and cultivation in prehistoric Korea. Arctic Anthropology 39 :95 121. Choquenot, D., and W. A. Ruscoe. 2003. Landscape complementation a nd food limitation of large herbivores: habitat related constraints on the foraging efficiency of wild pigs. Journal of Animal Ecology 72 :14 26. Clay, D. 1997. Critias. Pages 1292 1306. in J. M. Cooper and D. S. Hutchinson, editors. Plato Complete Works. H ackett Publishing, Indianapolis, IN. Clements, F. E. 1916. Plant succession, an analysis of the development of vegetation. Pages 1 512. Carnegie Institution of Washington Publication 242. Collins, J., A. Kinzig, N. Grimm, W. F. Fagan, D. Hope, J. Wu, and E Borer. 2000. A new urban ecology: Modeling human communities as integral parts of ecosystems poses special problems for the development and testing of ecological theory. American Scientist 88 :416 425. Connolly, G. E. 1978. Predator conntrol and coyote po pulations: A review of simulation models. Pages 327 345 in M. Bekoff, editor. Coyotes: Biology, Behavior, and Management. Academic Press, New York. Cowled, B. D., S. J. Lapidge, J. O. Hampton, and P. B. Spencer. 2006. Measured the demographic and genetic e ffects of pest control in a highly persecuted Feral Pig population. Journal of Wildlife Management 70 :1690 1697. Cox, W. 2008. Demographia World Urban Areas: Population & Density. Pages 1 89. Demographia http://www.demographia.com/. Croll, D. A., J. L. Mar on, J. A. Estes, E. M. Danner, and G. V. Byrd. 2005. Introduced predators transform subarchtic islands from grassland to tundra. Science 307 :1959 1961. Cronon, W. 1995. The Trouble with Wilderness; or, Getting Back to the Wrong Nature. Pages 69 90 in W. Cr onon, editor. Uncommon Ground: Rethinking the Human Place in Nature. W.W. Norton, New York. Crooks, K. R., and M. Soul. 1999. Mesopredator release and avifaunal extinctions in a fragmented system. Nature 400 :563 566. Cruz, F., C. J. Donlan, K. Campbell, and V. Carrion. 2005. Conservation action in the Galpagos: feral pig ( Sus scrofa ) eradication from Santiago Island. Biological Conservation 121 :473 478. Cuevas, M. F., A. Novillo, C. Campos, M. A. Dacar, and R. A. Ojeda. 2010. Food habits and impacts of r ooting behavior of the invasive wild boar, Sus scrofa in a protected area of the Monte Desert, Argentina. Journal of Arid Environments 74 :1582 1585.
213 Curtin, J. 1890. Myths and Folk Tales of Ireland. Dover Publications, Inc., New York. Daskalov, G. M., A. N. Grishin, S. Rodionov, and V. Mihneva. 2007. Trophic cascades triggered by overfishing reveal possible mechanisms of ecosystem regime shifts. PNAS 104 :10518 10523. Dawkins, R. 2006. The God Delusion. Houghton Mifflin Harcourt. De Steven, D. 1991. Experim ents on mechanisms of tree establishment in old field succession: Seedling emergence. Ecology 72 : 1066 1075. Dear, C. E., and O. E. Myers. 2005. Conflicting understandings of wilderness and subsistence in Alaskan National Parks. Society & Natural Resources 18 :821 837. DeMallie, R. J. 1987. Lakota belief and ritual in the nineteenth century. Pages 25 43 in R. J. DeMallie, and D. R. Parks, editors. Sioux Indian Religion. University of Oklahoma Press, Norman. DeNiro, M. J. 1985. Post mortem preservation and al teration of in vivo bone collagen isotope ratios in relation ot paleodietary reconstruction. Nature 317 :806 809. DeStefano, S., and R. M. DeGraaf. 2003. Exploring the ecology of suburban wildlife. Frontiers in Ecology and the Environment 1 :95 101. Devictor V., R. Julliard, and F. Jiguet. 2008. Distribution of specialist and generalist species along spatial gradients of habitat disturbance and fragmentation. Oikos 117 :507 514. Diamond, J. 1994. Ecological collapses of past civilizations. Proceedings of the American Philosophical Society 138 :363 370. Diegues, A. C. 1998. Recycled rain forest myths. Terra Nova: Nature and Culture 3 :113 127. Dudley, N. 2008. Guidelines for Applying Protected Area Management Categories. IUCN, Gland, Switzerland. Duff, A. 2004. M ammals of the World: A Checklist. Yale University Press, New Haven, CT. Dunn, R. R. 2005. Modern insect extinctions, the neglected majority. Conservation Biology 19 :1030 1036. Dunning, J. B., B. J. Danielson, and H. R. Pulliam. 1992. Ecological processes t hat affect populations in complex landscapes. Oikos 65 :169 175. Durbin, A. S., S. W. Nixon, and C. A. Oviatt. 1979. Effects of the spawning migration of the alewife, Alosa pseudoharengus on freshwater ecosystems. Ecology 60 :8 17.
214 Earl, D. A., and B. M. vo nHoldt. 2011. STRUCTURE HARVESTER: a website and program for visualizing STRUCTURE output and implementing the Evanno method. Conservation Genetics Resources. Efron, B. 1981. Nonparametric Standard Errors and Confidence Intervals. The Canadian Journal of S tatistics 9 :138 158. Ehrenfeld, J. G. 2010. Ecosystem consequences of biological invasions. Annual Review of Ecology Evolution and Systematics 41 :59 80. Ehrlich, P. R., and L. C. Birch. 1967. The "balance of nature" and "population control". The American N aturalist 101 :97. Ellis, E. C., and N. Ramankutty. 2008. Putting people on the map: anthropogenic biomes of the world. Frontiers in Ecology and the Environment 6 :439 447. Ellsworth, J. W., and B. C. McComb. 2003. Potential effects of Passenger Pigeon flock s on the structure and composition of presettlement forests of eastern North America. Conservation Biology 17 :1548 1558. Engeman, R. M., H. T. Smith, R. Severson, M. A. Severson, J. Woolard, S. A. Shwiff, B. Constantin, and D. Griffin. 2004. Damage reducti on estimates and benefit cost ratios for feral swine control from the last remnant of a basin marsh system in Florida. Environmental Conservation 31 :207 211. Engeman, R. M., H. T. Smith, S. A. Shwiff, B. Constantin, J. Woolard, M. Nelson, and D. Griffin. 2 003. Prevelence and economic value of feral swine damage to native habitat in three Florida state parks. Environmental Conservation 30 :319 324. Ereshefsky, M. 2007. Where the wild things are: environmental preservation and human nature. Biology & Philosoph y 22 :57 72. Estes, J. A., J. Terborgh, J. S. Brashares, M. E. Power, J. Berger, W. J. Bond, S. R. Carpenter, T. E. Essington, R. D. Holt, J. B. C. Jackson, R. J. Marquis, L. Oksanen, T. Oksanen, R. T. Paine, E. K. Pikitch, W. J. Ripple, S. A. Sandin, M. Sc heffer, T. W. Schoener, J. B. Shurin, A. R. E. Sinclair, M. E. Soul, R. Virtanen, and D. Wardle. 2011. Trophic downgrading of planet Earth. Science 33 :301 306. Evanno, G., S. Regnaut, and J. Goudet. 2005. Detecting the number of clusters of individuals us ing the software STRUCTURE: a simulation study. Molecular Ecology 14 :2611 2620. Evans, C. 1956. Ecosystem as the basic unit in ecology. Science 123 :1127 1128. Falush, D., M. Stephens, and J. K. Pritchard. 2007. Inference of population structure using multi locus genotype data: dominant markers and null alleles. Molecular Ecology Notes 7 :574 578.
215 Fang, M., G. Larson, H. S. Ribeiro, N. Li, and L. Andersson. 2009. Contrasting mode of evolution at a coat color locus in wild and domestic pigs. PLoS Genetics 5 :1 6 Feary, D. A. 2007. The influence of resource specialization on the response of reef fish to coral disturbance. Marine Biology 153 :153 161. Fernndez Llario, P. 2005. The sexual function of wallowing in male Wild Boar ( Sus scrofa ). Journal of Ethology 23 : 9 14. Fieller, E. C. 1954. Some problems in interval estimation. Journal of the Royal Statistical Society. B. 16 :175 185. Foufopoulos, J., and A. R. Ives. 1999. Reptile extinctions on land bridge islands: life history attributes and vulnerability to extinc tion. American Naturalist 153 :1 25. Frankham, R. 1995. Effective population size/adult population size ratios in wildlife: a review. Genetical Research 66 :95 107. Frankham, R. 1996. Relationship of genetic variation to population size in wildlife. Conserva tion Biology 10 :1500 1508. Fraterrigo, J. M., and J. A. Wiens. 2005. Bird communities of the Colorado Rocky Mountains along a gradient of exurban development. Landscape and Urban Planning 71 :263 275. Gadgil, M., and V. D. Vartak. 1976. The sacred groves of Western Ghats in India. Economic Botany 30 :152 160. Gaucherand, S., P. Laincourt, and S. Lavorel. 2006. Importance and intensity of competition along a fertility gradient and across species. Journal of Vegetation Science 17 :455 464. Gibb, H., and D. F. Ho chuli. 2002. Habitat fragmentation in an urban environment: large and small fragments support different arthropod assemblages. Biological Conservation 106 :91 100. Gipson, P. S. 1978. Coyotes and related Canis in the Southeastern United States with a commen t on Mexican and Central American Canis In Coyotes: Biology, Behavior, and Management M. Bekoff ed. Academic Press, New York. 191 208. Goldman, M. J. 2011. Strangers in their own land: Maasai and wildlife conservation in northern Tanzania. Conservation a nd Society 9 :65 79. Goodall, D. W. 1952. Some considerations in the use of point quadrat methods for the analysis of vegetation. Australian Journal of Biological Science 5 :1 41. Gould, S. J. 1974. The origin and function of 'bizarre' structures: Antler siz e and skull size in the 'Irish Elk', Megaloceros giganteus Evolution 28 :191 220.
216 Goulding, M. J. 2001. Possible genetic sources of free living Wild Boar ( Sus scrofa ) in southern England. Mammal Review 31 :245 248. Graves, H. B. 1984. Behavior and ecology o f wild and feral swine ( Sus scrofa ). Journal of Animal Science 58 :482 492. Grimshaw, H. J., and D. A. Dolske. 2002. Rainfall concentrations and wet atmosphereic deposition of phosphorus and other constituents in Florida, USA. Water, Air, and Soil Pollution 137 :117 140. Gudermann, R. 2005. Conviction and Constraint: Hydraulic Engineers and Agricultural Amelioration Projects in Nineteenth Century Prussia. Pages 33 54 in T. L. T. Zeller, editor. Germany's Nature: Cultural Landscape and Environmental History. R utgers University Press, New Brunswick, NJ. Guha, R. 1989. Radical American environmentalism and wilderness preservation. Environmental Ethics 11 :71 83. Guha, R. 1998. Deep ecology revisited. Pages 271 279 in J. B. Callicott, and M. P. Nelson, editors. The Great New Wilderness Debate. University of Georgia Press, Athens, GA. Ha, T. H. 1969. The Korean Nights Entertainments. Yonsei University Press, Seoul, Korea. Hairston, N. G., F. E. Smith, and L. B. Slobodkin. 1960. Community structure, population control and competition. The American Naturalist 94 :421 425. Hampton, J. O., J. Pluske, and P. B. S. Spencer. 2004a. A preliminary genetic study of the social biology of feral pigs in south western Australia and the implications for management. Wildlife Research 31 :375 381. Hampton, J. O., P. B. Spencer, D. L. Alpers, L. E. Twigg, A. P. Woolnough, J. Doust, T Higgs, and J. R. Pluske. 2004b Molecular techniques, wildlife management and the importance of genetic population structure and dispersal: a case study wi th feral pigs. Journal of Applied Ecology 41 :735 743. Hare, P. E., M. L. Fogel, T. W. J. Stafford, A. D. Mitchell, and T. C. Hoering. 1991. The isotopic composition of carbon and nitrogen in individual amino acids isolated from modern and fossil proteins. Journal of Archaeological Science 18 :277 292. Harestad, A. S., and F. L. Bunnel. 1979. Home range and body weight -a reevaluation. Ecology 60 :389 402. Harmon, D. 1987. Cultural diversity, human subsistence, and the national park ideal. Environmental Ethics 9 :147 158.
217 Hassrick, R. B. 1964. The Sioux: Life and Customs of a Warrior Society. University of Oklahoma Press, Norman. Haunshi, S., R. Basumatary, P. S. Girish, S. Koley, R. K. Bardoloi, and A. Kumar. 2009. Identification of chicken, duck, pigeon and pi g meat by species specific markers of mitochondrial origin. Meat Science 83 :454 459. Heaton, T. H. E. 1987. The 15 N/ 14 N ratios of plants in South Africa and Namibia: relationship to climate and coastal/saline environments. Oecologia 74 :236 246. Henberg, M. 1994. Wilderness, Myth, and the American Character. The Key Reporter 59 :7 11. Hendry, C. D., P. L. Brezonik, and E. S. Edgerton. 1981. Atmospheric deposition of nitrogen and phosphorus in Florida. Pages 199 215 in S. J. Eisenreich, editor. Atmospheric Pol lutants in Natural Waters. Ann Arbor Science, Ann Arbor, Michigan. Henle, K., K. F. Davies, M. Kleyer, C. Margules, and J. Settele. 2004. Predictors of species sensitivity to fragmentation. Biodiversity and Conservation 13 :207 251. Hill, N. J., K. A. Carbe ry, and E. M. Deane. 2007. Human possum conflict in urban Sydney, Australia: Public perceptions and implications for species management. Human Dimensions of Wildlife 12 :101 113. Holdo, R. M., R. D. Holt, M. B. Coughenour, and M. E. Ritchie. 2007. Plant pro ductivity and soil nitrogen as a function of grazing, migration and fire in an African savanna. Journal of Ecology 95 :115 128. Holdo, R. M., A. R. E. Sinclair, A. P. Dobson, K. L. Metzger, B. B. M., M. E. Ritchie, and R. D. Holt. 2009. A disease mediated t rophic cascade in the Serengeti and its implications for ecosystem C. PLOS 7 :1 12. Holt, R. D. 2010. Towards a trophic island biogeography: reflections on the interface of island biogeography and food web ecology. Pages 143 185 in J. B. Losos, and R. E. Ri cklefs, editors. The Theory of Island Biogeography Revisited. Princeton University Press, Princeton. Hone, J. 2002. Feral pigs in Namadgi National Park, Australia: dynamics, impacts and management. Biological Conservation 105 :231 242. Howland, M. R., L. T. Corr, S. M. M. Young, V. Jones, S. Jim, N. J. van der Merwe, A. D. Mitchell, and R. P. Evershed. 2003. Expression of the dietary isotope signal in the compound 13 C values of pig bone lipids and amino acids. International Journal of Osteoarchaeol ogy 13 :54 65.
218 Hubisz, M. J., D. Falush, M. Stephens, and J. K. Pritchard. 2009. Inferring weak population structure with the assistance of sample group information. Molecular Ecology Resources 9 :1322 1332. Hughes, J. D. and J. V. Thirgood. 1982. Deforestat ion, erosion, and forest management in ancient Greece and Rome. Journal of Forest History April : 60 75. Huynh, T. T. T., A. J. A. Aarnink, M. J. W. Heetkamp, M. W. A. Verstegen, and B. Kemp. 2007. Evaporative heat loss from group housed growing pigs at hi gh ambient temperatures. Journal of Thermal Biology 32 :293 299. Ickes, K., S. J. DeWalt, and S. Appanah. 2001. Effects of native pigs ( Sus scrofa ) on woody understorey vegetation in a Malaysian lowland rain forest. Journal of Tropical Ecology 17 :191 206. I ckes, K., C. J. Paciorek, and S. C. Thomas. 2005. Impacts of nest construction by native pigs ( Sus scrofa ) on lowland Malaysian rain forest saplings. Ecology 86 :1540 1547. Ilse, L. M., and E. C. Hellgren. 1995. Resource partitioning in sympatric population s of collared peccaries and feral hogs in southern Texas. Journal of Mammology 76 :784 799. Imort, M. 2005. A Sylvan People: Wilhelmine Forestry and the Forest as a Symbol of Germandom. Pages 55 80 in T. Lekan, and T. Zeller, editors. Germany's Nature: Cult ural Landscape and Environmental History. Rutgers University Press, New Brunswick, NJ. Inouye, R. S., N. J. Huntly, D. Tilman, J. R. Tester, M. Stillwell, K. C. Zinnel. 1987. Old field succession on a Minnesota sand plain. Ecology 68 : 12 26. Ishitani, M., D. J. Kotze, and J. Niemel. 2003. Changes in carabid beetle assemblages across an urban rural gradient in Japan. Ecography 26 :481 489. Jenkins, S. G., S. T. Partridge, T. R. Stephenson, S. D. Farley, and C. T. Robbins. 2001. Nitrogen and carbon isotope fr actionation between mothers, neonates, and nursing offspring. Oecologia 129 :336 341. Johnston, C. A., and R. J. Naiman. 1987. Boundary dynamics at the aquatic terrestrial interface: the influence of beaver and geomorphology Landscape Ecology 1 :47 57. Joki mki, J. 2000. Artificial nest predation and abundance of birds along an urban gradient. The Condor 102 :838 847. Jolley, D. B., S. S. Ditchkoff, B. D. Sparklin, L. B. Hanson, M. S. Mitchell, and J. Grand, B. 2010. Estimate of herpetofauna depredation by a population of wild pigs. Journal of Mammalogy 91 :519 524.
219 Jones, K. B., A. C. Neale, M. S. Nash, K. H. Riitters, J. D. Wickham, R. V. O'Neill, and R. D. Van Remortel. 2000. Landscape correlates of breeding bird richness across the United States Mid Atlanti c region. Environmental Monitoring and Assessment 63 :159 174. Keddy, P., C. Gaudet, and L. H. Fraser. 2000. Effects of low and high nutrients on the competitive hierarchy of 26 shoreline plants. Journal of Ecology 88 :413 423. Kelly, J. F. 2000. Stable isot opes of carbon and nitrogen in the study of avian and mammalian trophic ecology. Canadian Journal of Zoology 78 :1 27. Kephart, K. B., and G. W. Sherritt. 1990. Performance and nutrient balance in growing swine fed low protein diets supplemented with amino acids and potassium. Journal of Animal Science 68 :1999 2008. Kijas, J. M. H., R. Wales, A. Trnsten, P. Chardon, M. Moller, and L. Andersson. 1998. Melanocortin receptor 1 (MC1R) mutations and coat color in pigs. Genetics 150 :1177 1185. Kim, K. C. 1997. Pr eserving biodiversity in Korea's Demilitarized Zone. Science 278 :242 243. Kimura, M., and J. F. Crow. 1964. The number of alleles that can be maintained in a finite population. Genetics 49 :725 738. Kingdon, J. 1979. East African Mammals: an Atlas of Evolut ion in Africa IIIB Large Mammals. University of Chicago Press, Chicago. Kotanen, P. M. 1995. Responses of vegetation to a changing regime of disturbance: effects of feral pigs in a Californian coastal prairie. Ecography 18 :190 199. Kuriyan, R. 2002. Linkin g local perceptions of elephants and conservation: Samburu pastoralists in northern Kenya. Society & Natural Resources 15 :949 957. Larson, G., K. Dobney, U. Albarella, M. Fang, E. Matisoo Smith, J. Robins, S. Lowden, H. Finlayson, T. Brand, E. Willerslev, P. Rowley Conwy, L. Andersson, and A. Cooper. 2005. Worldwid phylogeography of Wild Boar reveals multiple centers of pig domestication. Science 307 :1618 1621. Laval, G., N. Iannuccelli, C. Legault, D. Milan, M. A. M. Groenen, E. Giuffra, L. Andersson, P. H Nissen, C. B. Jrgensen, P. Beeckmann, H. Geldermann, J. L. Foulley, C. Chevalet, and L. Ollivier. 2000. Genetic diversity of eleven European pig breeds. Genetics Selection Evolution 32 :187 203. Lee, H. C. 1994. Korean rice culture and characteristics of agricultural genre painting. Journal of Rural Development 17 :57 72.
220 Lee, R. B. 1993. The Dobe Ju/'hoansi. Harcourt Brace College Publishers, Fort Worth, TX. Leibold, M. A., E. P. Economo, and P. Peres Neto. 2010. Metacommunity phylogenetics: separating th e roles of environmental filters and historical biogeography. Ecology Letters 13 : 1290 1299. Leopold, A. 1921. The wilderness and its place in forest recreational policy. Journal of Forestry 19 :718 721. Lepp, A., and S. Holland. 2006. A comparison of attit udes toward state led conservation and community based conservation in the village of Bigodi, Uganda. Society & Natural Resources 19 :609 623. Litzinger, R. 2004. The mobilization of "Nature": Perspectives from north west Yunnan. The China Quarterly 178 :488 504. Livingston, M., W. W. Shaw, and L. K. Harris. 2002. A model for assessing wildlife habitats in urban landscapes of eastern Pima County, Arizona (USA). Landscape and Urban Planning 64 :131 144. Loreau, M., N. Mouquet, R. D. Holt. 2003. Meta ecosystems: a theoretical framework for spatial ecosystem ecology. Ecology Letters 6 : 673 679. Lowden, S., H. A. Finlayson, A. A. Macdonald, A. C. Downing, S. J. Goodman, K. Leus, L. Kaspe, E. Wahyuni, and A. L. Archibald. 2002. Application of Sus scrofa microsatell ite markers to wild suiformes. Conservation Genetics 3 :347 350. Maehr, D. S. 1997. The comparative ecology of bobcat, black bear and Florida panther in south Florida. Bulletin of the Florida Museum of Natural History 40 :1 176. Maron, J. L., J. A. Estes, D. A. Croll, E. M. Danner, S. C. Elmendorf, and S. L. Buckelew. 2006. An introduced predator alters Aleutian island plant communities by thwarting nutrient subsidies. Ecological Monographs 76 :3 24. Martinez, A. M., J. V. Delgado, A. Rodero, and J. L. Vega Pl a. 2000. Genetic structure of the Iberian pig breed using microsatellites. Animal Genetics 31 :295 301. Marzluff, J. M., and E. Neatherlin. 2006. Corvid response to human settlements and campgrounds: Causes, consequences, and challenges for conservation. Bi ological Conservation 130 :279 288. Mattson, W. J., Jr. 1980. Herbivory in relation to plant nitrogen content. Annual Review of Ecology and Systematics 11 :119 1161. Mauch, C. 2004. Nature and Nation in Transatlantic Perspective. Pages 1 9 in C. Mauch, edito r. Nature in German History. Berghahn Books, New York.
221 Mayer, J. J., and I. L. Brisbin 1991. Wild pigs in the United States: their history, comparative morphology, and current status. University of Georgia Press, Athens. Mayer, J. J., I. L. Brisbin, and J. M. Sweeney. 1989. Temporal dynamics of color phenotypes in an isolated population of Feral Swine. Acta Theriologica 34 :247 252. McDonnell, M. J., and S. T. A. Pickett, editors. 1993. Humans as Components of Ecosystems: The Ecology of Subtle Human Effects and Populated Areas. Springer, New York. McIlroy, J. C., and E. J. Gifford. 1997. The 'Judas' pig technique: a method that could enhance control programmes against Feral Pigs, Sus scrofa Wildlife Research 24 :483 491. McIntyre, N. E., J. Rango, W. F. Fagan and S. H. Faeth. 2001. Ground arthropod community structure in a heterogeneous urban environment. Landscape and Urban Planning 52 :257 274. McKay, T. F. 2007. Wild populations are smaller than we think: a commentary on 'Effective population size/adult pop ulation size ratios in wildlife: a review' by Richard Frankham. Genetical Research 89 :489. McKinney, M. L. 2002. Urbanization, biodiversity, and conservation. BioScience 52 :883 890. McPhee, J. 1989. The Control of Nature. The Noonday Press, New York. Minag awa, M., A. Matsui, and N. Ishigiro. 2005. Patterns of prehistoric boar Sus scrofa domestication and inter island pig trading across the East China Sea, as determined by carbon and nitrogen isotope analysis. Chemical Geology 218 :91 102. Mitchell, J., and R Mayer. 1997. Diggings by Feral Pigs within the Wet Tropics World Heritage Area of North Queensland. Wildlife Research 24 :591 601. Mohr, D., L. W. Cohnstaedt, and W. Topp. 2005. Wild Boar and Red Deer affect soil nutrients and soil biota iin steep oak sta nds of the Eifel. Soil Biology & Biochemistry 37 :693 700. Morris, J. T. 1991. Effects of nitrogen loading on wetland ecosystems with particular reference to atmospheric deposition. Annual Review of Ecology and Systematics 22 :257 279. Mylavarapu, R. S., and E. D. Kennelley. 2002. UF/IFAS Extension Soil Testing Laboratory (ESTL) analytical procedures and training manual in S. a. W. S. D. Circular 1248, Florida Cooperative Extension Service, Institute of Food and Agricultural Sciences, University of Florida, e ditor.
222 Mylavarapu, R. S., and D. L. Moon. 2007. UF/IFAS Extension Soil Testing Laboratory (ESTL) Analytical Procedures and Training Manual. Page 25 in I. f. F. a. A. Sciences, editor. University of Florida, Gainesville. Nash, R. 1967. Wilderness and the Am erican Mind. Yale University Press, New Haven, Connecticut, USA. Nature Conservation Act. 1999. Iceland Nelson, M. P., and J. B. Callicott. 2008. Introduction. Pages 1 17 in M. P. Nelson, and J. B. Callicott, editors. The Wilderness Debate Rages On. The Un iversity of Georgia Press, Athens, GA. New, T. R., and D. P. A. Sands. 2003. Conservation concerns for butterflies in urban areas of Australia. Journal of Insect Conservation 6 :207 215. Newsome, A. 1990. The control of vertebrate pests by vertebrate predat ors. TREE 5 :187 191. Newsome, S. D., P. L. Koch, M. A. Etnier, and D. Aurioles Gamboa. 2006. Using carbon and nitrogen isotope values to investigate maternal strategies in northeast Pacific otariids. Marine Mammal Science 22 :556 572. Nogueira Filho, S. L. G., S. S. C. Nogueira, and J. M. V. Fragoso. 2009. Ecological impacts of feral pigs in the Hawaiian Islands. Biodiversity and Conservation 18 :3677 3683. Norr, L. 2004. Stable isotope analysis and dietary inference. Pages 169 184 in D. L. Hutchinson, editor Bioarchaeology of the Florida Gulf Coast: Adaptation, Conflict, and Change. University Press of Florida, Gainesville. Noss, R. F., and A. Y. Cooperrider 1994. Saving Nature's Legacy: Protecting and Restoring Biodiversity. Island Press, Washington, D.C. O 'Hara, K. L. 2001. The silviculture of transformation -a commentary. Forest Ecology and Management 151 :81 86. Odum, E. P. 1960. Organic production and turnover in old field succession. Ecology 41 : 34 49. Oelschlaeger, M. 1991. The Idea of Wilderness. Yal e University Press, New Haven, CT. Oelschlaeger, M. 2007. Boundaries and Darwin: Bridging the Great Divide. Pages 3 17 in C. S. Brown, and T. Toadvine, editors. Nature's Edge: Boundary Explorations in Ecological Theory and Practice. State University of New York Press, Albany, New York, USA.
223 Oksanen, L., and T. Oksanen. 2000. The logic and realism of the Hypothesis of Exploitation Ecosystems. The American Naturalist 155 :703 723. Oksanen, T. 1990. Exploitation ecosystems in heterogeneous habitat complexes. Ev olutionary Ecology 4 :220 234. Oliver, W. L. R., and I. L. Brisbin. 1993 Introduced and feral pigs: problems, policy, and priorities. Pages 179 191 in W. L. R. Oliver, editor. Pigs, Peccaries, and Hippos. Status Survey and Conservation Action Plan of the I UCN World Conservation Union, Gland, Switzerland. Oliver, W. L. R., I. L. Brisbin, and S. Takahashi. 1993. The Eurasian Wild Pig ( Sus scrofa ). Pages 112 121 in W. L. R. Oliver, editor. Pigs, Peccaries and Hippos. Status Survey and Conservation Action Plan of the IUCN World Conservation Union, Gland, Switzerland. Oostings, H. J. 1942. An ecological analysis of the plant communities of Piedmont, North Carolina. American Midland Naturalist 28 :1 126. Oostings, H. J., and M. E. Humphreys. 1940. Buried viable see ds in a successional series of old field and forest soils. Bulletin of the Torrey Botany Club 67 :253 273. Pace, M. L., J. J. Cole, S. R. Carpenter, and J. F. Kitchell. 1999. Trophic cascades revealed in diverse ecosystems. TREE 14 :483 488. Paine, R. T. 198 0. Food webs: linkage, interaction strength and community infrastructure. The Journal of Animal Ecology 49 :666 685. Papworth, S. K., J. Rist, L. Coad, and E. J. Milner Gulland. 2009. Evidence for shifting baseline syndrome in conservation. Conservation Let ters 2 :93 100. Park, Y. H., K. S. Lee, H. Y. Lee, I. SON, and J. R. Lee, editors. 2000. Atlas of Korea. Sung Ji Mun Hwa Co., Ltd., Seoul, Korea. Peakall, R., and P. E. Smouse. 2006. GENALEX 6: genetic analysis in Excel. Population genetic software for tea ching and research. Molecular Ecology Notes 6 Pennings, S. C., M. B. Grant, and M. D. Bertness. 2005. Plant zonation in low latitude salt marshes: disentangling the roles of flooding, salinity and competition. Journal of Ecology 93 :159 167. Peters, J., S. M. Combs, B. Hoskins, J. Jarman, M. E. Watson, A. M. Wolf, and N. Wolf. 2003. Recommended Methods of Manure Analysis. Page 57 in U. o. Wsconsin Extension, editor. Cooperative Extension Publishing, Madison. Peterson, B. J., and B. Fry. 1987. Stable isot opes in ecosystem studies. Annual Review of Ecology and Systematics 18 :293 320.
224 Phillips, D. L., and J. W. Gregg. 2003. Source partitioning using stable isotopes: coping with too many sources. Oecologia 136 :261 269. Pickett, S. T. A., M. L. Cadenasso, J. M Grove, C. H. Nilon, R. V. Pouyat, W. C. Zipperer, and R. Costanza. 2001. Urban ecological systems: Linking terrestrial ecological, physical, and socioeconomic components of metropolitan areas. Annual Review of Ecology and Systematics 32 :127 157. Pielberg G., C. Olsson, A. C. Syvnen, and L. Andersson. 2002. Unexpectedly high allelic diversity at the KIT locus causing dominant white color in the Domestic Pig. Genetics 160 :305 311. Plotkin, M. J. 1993. Tales of a Shaman's Apprentice. Viking, New York. Poli s, G. A., W. B. Anderson, and R. D. Holt. 1997. Toward an integration of landscape and food web ecology: the dynamics of spatially subsidized food webs. Annual Review of Ecology and Systematics 28 :289 316. Polis, G. A., A. L. W. Sears, G. r. Huxel, D. R. S trong, and J. Maron. 2000. When is a trophic cascade a trophic cascade? TREE 15 :473 475. Polischuk, S. C., K. A. Hobson, and M. A. Ramsay. 2001. Use of stable carbon and nitrogen isotopes to assess weaning and fasting in female polar bears and their cubs. Canadian Journal of Zoology 79 :499 511. Power, M. E. 1992. Top down and bottom up forces in food webs: do plants have primacy? Ecology 73 :733 746. Prange, S., S. D. Gehrt, and E. P. Wiggers. 2004. Influences of anthropogenic resources on raccoon ( Procyon lotor ) movements and spatial distribution. Journal of Mammalogy 85 :483 490. Pritchard, J. K., M. Stephens, and P. Donnelly. 2000. Inference of population structure using multilocus genotype data. Genetics 155 :945 959. Provincial Parks and Conservation Rese rves Act. 2006. Ontario Radkau, J. 2005. Germany as a Focus of European "Particularities" in Environmental History. Pages 17 32 in T. Lekan, and T. Zeller, editors. Germany's Nature: Cultural Landscapes and Environmental History. Rutgers University Press, New Brunswick, NJ. Raftery, B. 1994. Pagan Celtic Ireland: The Enigma of the Irish Iron Age. Thames and Hudson, New York. Rand, T. A., and S. M. Louda. 2006. Spillover of agriculturally supsidized predators as a potential threat to native insect herbivores in fragmented landscapes. Conservation Biology 20 :1720 1729.
225 Rand, T. A., and T. Tscharntke. 2007. Contrasting effects of natural habitat loss on generalist and specialist aphid natural enemies. Oikos 116 :1353 1362. Randa, L. A., and J. A. Yunger. 2006. C arnivore occurrence along an urban rural gradient: a landscape level analysis. Journal of Mammalogy 87 :1154 1164. Raymond, M., and F. Rousset. 1995. GENEPOP (version 1.2): population genetics software for exacts tests and ecumenicism. Journal of Heredity 8 6 :248 249. Ripple, W. J., T. P. Rooney, and R. L. Beschta. 2010. Large predators, deer, and trophic cascades in boreal and temperate ecosystems. Pages 141 177 in J. Terborgh, and J. A. Estes, editors. Trophic Cascades: Predators, prey, and the changing dyn amics of nature. Island Press, Washington, D.C. Robinson, S. K and D. S. Wilcove. 1994. Forest fragmentation in the temperate zone and its effects on migratory songbirds. Bird Conservation International 4 :233 249. Roemer, G. W., C. J. Donlan, and F. Courch amp. 2001. Golden eagles, feral pigs, and insular carnivores: How exotic species turn native predators into prey. PNAS 99 :791 796. Rolston, H. 1991. The wilderness idea reaffirmed. The Environmental Professional 13 :370 377. Ross Bryant, L. 2005. Sacred sit es: Nature and nation in the US National Parks. Religion and American Culture a Journal of Interpretation 15 :31 62. Ross, S. M. 2003. Peirce's criterion for the elimination of suspect experimental data. Journal of Engineering Technology 20 :38 41. Rousset, F. 2008. Genepop '007: a complete reimplementation of the Genepop software for Windows and Linux. Molecular Ecology Resources 8 :103 106. Ruesink, J. 2005. Global analysis of factors affecting the outcomme of freshwater fish introductions. Conservation Biol ogy 19 :1883 1893. Saeki, M., P. J. Johnson, and D. W. MacDonald. 2007. Movements and habitat selection of Raccoon Dogs ( Nyctereutes procyonoides ) in a mosaic landscape. Journal of Mammalogy 88 :1098 1111. Sands, J. S., D. Ragland, C. Baxter, B. C. Joern, T. E. Sauber, and O. Adeola. 2001. Phosphorus bioavailability, growth performance, and nutrient balance in pigs fed high available phosphorus corn and phytase. Journal of Animal Science 79 :2134 2142. Sare, D. T. J., J. S. Millar, and F. J. Longstaffe. 2005. Nitrogen and carbon isotope fractionation between mothers and offspring in red backed voles ( Clethrionomys gapperi ). Canadian Journal of Zoology 83 :712 716.
226 Sartain, J. B. 2008. Soil and tissue testing and interpretation for Florida turfgrasses, SL181 in F. C. E. S. Soil and Water Science, Institute for Food and Agricultural Sciences, editor. University of Florida, Gainesville, FL. Saunders, G., and S. McLeod. 1999. Predicting home range size from the body mass or population densities of feral pigs, Sus sc rofa (Artiodactyla: Suidae). Australian Journal of Ecology 24 :538 543. Schuelke, M. 2000. An economic method for the flourescent labeling of PCR fragments: A poor man's approach to genotyping for research and high throughput diagnostics. Nature Biotechnolo gy 18 :233 234. Sekergioglu, C. H. 2006. Increasing awareness of avain ecological function. Trends in Ecology and Evolution 21 :464 471. Shearer, G., and D. H. Kohl. 1992. Information derived from variation in the natural abundance of 15 N in complex biologic al systems. Pages 133 189 in E. Buncel, and W. H. Saunders, Jr., editors. Isotopes in Organic Chemistry. Elsevier, New York. distributions. Journal of the American Statistic al Association 62 :626 633. Singer, F. J., W. T. Swank, and E. E. Clebsch. 1984. Effects of wild pig rooting in a deciduous forest. Journal of Wildlife Management 48 :465 473. Slobodkin, L. B., F. E. Smith, and N. G. Hairston. 1967. Regulation in terrestrial ecosystems, and the implied balance of nature. The American Naturalist 101 :109 124. Smart, S. M., R. G. H. Bunce, R. Marrs, M. LeDuc, L. G. Firbank, L. C. Maskell, W. A. Scott, K. Thompson, and K. J. Walker. 2005. Large scale changes in the abundance of c ommon higher plant species across Britain between 1978, 1990 and 1998 as a consequence of human activity: Tests of hypothesised changes in trait representation. Biological Conservation 124 :355 371. Sowls, L. K. 1997. Javelinas and Other Peccaries: Their Bi ology, Management, and Use. Texas A&M University Press, College Station. Spencer, P. B., S. J. Lapidge, J. O. Hampton, and J. R. Pluske. 2005. The sociogenetic structure of a controlled feral pig population. Wildlife Research 32 :297 304. Spiller, D. A., an d A. A. Agrawal. 2003. Intense disturbance enhances plant susceptibility to herbivory: natural and experimental evidence. Ecology 84 :890 897. Spinage, C. 1998. Social change and conservation misrepresentation in Africa. Oryx 32 :265 276.
227 Stacey, T. 2003. Tr ibe: The Hidden History of the Mountains of the Moon. Stacey International, London. Standing Bear, L. 1933. Land of the Spotted Eagle. University of Nebraska Press, Lincoln. Steffan Dewenter, I., and T. Tscharntke. 2000. Butterfly communty structure in fra gmented habitats. Ecology Letters 3 :449 456. Sunseri, T. 2005. Forestry and the German Imperial Imagination: Conflicts over Forest Use in German East Africa. Pages 81 107 in T. Lekan, and T. Zeller, editors. Germany's Nature: Cultural Landscapes and Enviro nmental History. Rutgers University Press, New Brunswick, NJ. Terborgh, J. 2010. The trophic cascade on islands. Pages 116 142 in J. B. Losos, and R. E. Ricklefs, editors. The Theory of Island Biogeography Revisited. Princeton University Press, Princeton, NJ. Terborgh, J., and J. A. Estes. 2010a. Conclusion: Our trophically degraded planet. Pages 353 367 in J. Terborgh, and J. A. Estes, editors. Trophic Cascades: Predators, Prey, and the Changing Dynamics of Nature. Island Press, Washington, D. C. Terborgh, J., and J. A. Estes, editors. 2010b. Trophic Cascades: Predators, Prey, and the Changing Dynamics of Nature. Island Press, Washington D.C. Terborgh, J., L. Lopez, P. Nuez V., M. Rao, G. Shahabuddin, G. Orihuela, M. Riveros, R. Ascanio, G. H. Adler, T. D. Lambert, and L. Balbas. 2001. Ecological meltdown in predator free forest fragments. Science 294 :1923 1926. Thompson, V. P., and M. R. Leishman. 2005. Post fire vegetation dynamics in nutrient enriched and non enriched sclerophyll woodland. Austral Ecolog y 30 :250 260. Thrush, S. F., J. S. Gray, J. E. Hewitt, and K. I. Ugland. 2006. Predicting the effects of habitat homogenization on marine biodiversity. Ecological Applications 16 :1636 1642. Tigas, L. A., D. H. van Vuren, and R. M. Sauvajot. 2002. Behaviora l responses of bobcats and coyotes to habitat fragmentation and corridors in and urban environment. Biological Conservation 108 :299 306. Towne, C. W., and E. N. Wentworth 1950. Pigs from Cave to Cornbelt. University of Oklahoma Press, Norman, OK. Turnbull, C. M. 1961. The Forest People: A Study of the Pygmies of the Congo. Simon and Schuster, New York.
228 Tylianakis, J. M., T. Tscharntke, and O. T. Lewis. 2007. Habitat modification alters the structure of tropical host parasitoid food webs. Nature 445 :202 205 UN. 2008. World Population Prospects: The 2008 Revision Population Database http://esa.un.org/unpp. Van der Veken, S., K. Verheyen, and M. Hermy. 2004. Plant species loss in an urban area (Turnhout, Belgium) from 1880 to 1999 and its environmental determ inants. Flora 199 :516 523. van Vuure, C. 2005. Retracing the Aurochs: history, morphology and ecology of an extinct wild ox. Pensoft, Moscow. Vanni, M. J., D. L. Deangelis, D. E. Schindler, and G. R. Huxel. 2004. Overview: Crosshhabitat flux of nutrients a nd detritus. Pages 3 11 in G. A. Polis, M. E. Power, and G. R. Huxel, editors. Food webs at the landscape level. The University of Chicago Press, Chicago, IL. Vanni, M. J., and J. Headworth. 2004. Cross habitat transport of nutrients by omnivorous fish alo ng a productivity gradient: Integrating watersheds and reservoir food webs. Pages 43 61 in G. A. Polis, M. E. Power, and G. R. Huxel, editors. Food Webs at the Landscape Level. University of Chicago Press, Chicago, IL. Vernesi, C., B. Crestanello, E. Pecch oili, D. Tartari, D. Caramelli, H. Hauffe, and G. Bertorelle. 2003. The genetic impact of demographic decline and reintrodution in the wild boar ( Sus scrofa ): a microsatellite analysis. Molecular Ecology 12 :585 595. Virginia, R. A., and C. C. Delwiche. 198 2. Natural 15 N abundance of presumed N 2 fixing and non N 2 fixing plants from selected ecosystems. Oecologia 54 :317 325. Vitousek, P. M. 1994. Beyond global warming: Ecology and global change. Ecology 75 :1861 1876. Vitousek, P. M., C. M. D'Antonio, L. L. Lo ope, M. Rejmnek, and R. Westbrooks. 1997a. Introduced species: a significant component of human caused global change. New Zealand Journal of Ecology 21 :1 16. Vitousek, P. M., H. A. Mooney, J. Lubchenco, and J. M. Melillo. 1997b. Human domination of the Ea rth's ecosystems. Science 277 :494 499. Walker, J. R. 1982. Lakota Society. University of Nebraska Press, Lincoln. Waser, P. 1975. Monthly variations in feeding and activity patterns of the mangabey, Cercopithicus albigena (Lydekker). East African Wildlife Journal 13 :249 263. Welander, J. 2000. Spatial and temporal dynamics of wild boar ( Sus scrofa ) rooting in a mosaic landscape. Journal of the Zoological Society of London 252 :263 271.
229 Wetterer, J. K. 1999. Urban Ecology. Pages 664 647 in D. E. Alexander, an d R. W. Fairbridge, editors. Encyclopedia of Environmental Science. Kluwer. Wilderness Act. 1964. United States of America Wilderness Act. 1987. New South Wales Wilderness and Ecological Reserves Act. 1980. Newfoundland and Labrador Wilderness Protection A ct. 1992. South Australia Williams, J. A. 2005. Protecting Nature between Democracy and Dictatorship: The Changing Ideology of the Bourgeois Conservationist Movement, 1925 1935. Pages 183 206 in T. Lekan, and T. Zeller, editors. Germany's Nature: Cultural Landscapes and Environmental History. Rutgers University Press, New Brunswick, NJ. Willock, C. 1964. The Enormous Zoo: A Profile of the Uganda National Parks. Harcout, Brace & World, New York. Wilson, E. O. 1984. Biophilia. Harvard University Press, Cambri dge. Wood, G. W., and R. E. Brenneman. 1980. Feral Hog movements and habitat use in coastal South Carolina. The journal of Wildlife Management 44 :420 427. Worman C. O. 2010 Trooping fairies, trolls, and talking tigers: the influence of traditional wilder ness archetypes on current land use patterns. Biodiversity and Conservation 19 :3171 3193. Wright, S. 1965. Interpretation of population structure by F Statistics with special regard to systems of mating. Evolution 19 :395 420. Wunderlin, R. P. 1998. Guide t o the Vascular Plants of Florida. University Press of Florida, Gainesville, FL. Xu, J., E. T. Ma, D. Tashi, Y. Fu, Z. Lu, and D. Melick. 2005. Integrating sacred knowledge for conservation: cultures and landscapes in southwest China. Ecology and Society 10 :7 31. Yates, F. 1934. Contingency tables involving small numbers and the X 2 test. Supplement to the Journal of the Royal Statistical Society 1 :217 235. Yeats, W. B. 1892. Irish Fairy and Folk Tales. Barnes and Noble Books, New York, New York, USA. Zabel, J., and T. Tscharntke. 1998. Does fragmentation of Urtica habitats affect phytophagous and predatory insects differently? Oecologia 116 :419 425.
230 Zanette, L. R. S., R. P. Martins, and S. P. Ribeiro. 2005. Effects of urbanization on Neotropical wasp and bee assemblages in a Brazilian metropolis. Landscape and Urban Planning 71 :105 121. Zipes, J. 2003. The Complete Fairy Tales of the Brothers Grimm. Bantam Books, New York. S. 1982. Folk Tales from Korea. Hollym Corp., Seoul, Korea.
231 BIOGRAPHICAL SKETCH the Minnesota public school system (wandering the forests, mountains, and deserts of North America with his family did not hurt either) and attended North Hollywood High School Magnet for Biological and Mathematical Sciences. He earned his Bachelor of Science at Iowa State University with specializations in restoration ecology and mammal ian behavior. After graduation he joined the Peace Corps and was sent to Uganda where he learned more than he really wanted to. Before entering graduate school, he taught in and explored Korea. He procured his MSc. in z oology at the University of Florid a by studying frugivorous monkeys and their habitat use and nutritional ecology. In the spring of 2012, he received his Ph.D. in b iology from the University of Florida for the work described here.