<%BANNER%>

Applying Terrestrial Landscape Ecology Principles to the Design and Management of Marine Protected Areas in Coral Reef E...


PAGE 1

APPLYING TERRESTRIAL LANDSCAPE ECOLOGY PRINCIPLES TO THE DESIGN AND MANAGEMENT OF MARI NE PROTECTED AREAS IN CORAL REEF ECOSYSTEMS By LINDA ERICA RIKKI GROBER-DUNSMORE A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL OF THE UNIVERSITY OF FLOR IDA IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY UNIVERSITY OF FLORIDA 2005

PAGE 2

Copyright 2005 by Linda Erica Rikki Grober-Dunsmore

PAGE 3

This dissertation is dedicated to my bedste far, Jacob Nielsen and grandfather, Hyman Grober. You shared your passion for and in trigue with life and knowledge; and are my models of hard work, integrity, and compassion.

PAGE 4

ACKNOWLEDGMENTS Many people provided tremendous assistance and support in the completion of this dissertation and all of them deserve special thanks. I apologize if I have inadvertently omitted anyone; it is not my intention. I extend my thanks and sincere appreciation to everyone who has helped in any way, along the way. I greatly appreciate the guidance and support of my supervisory committee: Dr. Thomas K. Frazer (Chair), Nicholas Funicelli, William J. Lindberg, and Paul Zwick. It has been a great pleasure working with my advisor Tom Frazer, who gently supported me in my research and scientific development, and served as a mentor and friend over the past few years. His hands-off style and steadfast belief in my abilities fostered my confidence as a scientist, and I truly look forward to collaborations with him throughout my career. I would not have been given this opportunity without one person (Dr. Nick Funicelli, who I will truly miss as a friend and confidante). I relied on his insights on life, science, and management and learned much from him. Always cheerful and upbeat, he picked me up many times on this journey. I appreciate the support and encouragement that Bill Lindberg always provided, and the many discussions about the philosophy of science and my research results. He was always available to discuss science and he opened my eyes to the rigors and history of the scientific method. I have learned a great deal from Dr. Paul Zwick about geographic information systems and systems ecology. His enthusiasm, honesty, and optimism were invaluable to me throughout this study. iv

PAGE 5

I thank Jim Beets for pushing me to achieve my highest standards and for giving me the impetus to reach for the impossible in myself. Through his passion for fish ecology and natural history, I became addicted to learning about corals reefs ecosystems, and applying my strengths to support coral reef conservation. My study would not have been possible without the assistance of many people. Dr. Mike Allen, Dr. Chuck Cichra, and Howard Jelks, each helped by providing statistical advice, often without much advance warning. I also thank the scientists and resource managers at the Virgin Islands National Park who facilitated my field work on St. John, US Virgin Islands, (particularly Rafe Boulon, Chief of Resource Management; and Dr. Caroline Rogers, Research Scientist of the United States Geological Survey) (USGS)). Thomas Kelley, Jeff Miller, Rob Waara, Sherrie Caseau, Jack Hopkins and Jim Petterson of Virgin Islands National Park assisted either in the field or through logistical and administrative support while on St. John. Ilsa Kuffner, of the USGS, helped in the field and has been a valuable friend. Working in the field on St. John with Alan Friedlander, Jim Beets, Nick Wolff, and Ellen Link piqued my early interest in marine ecology. Those days in the field in the Caribbean continue to be some of the most memorable, comical, and enjoyable days of my life. In the Florida Keys, many people helped to facilitate this research. Particularly helpful were Billey Causey (FKNMS), Dr. Jim Bohnsack (NOAA-NMFS), Ben Richards (FKNMS), Brian Keller (FKNMS), Joanne Delaney (FKNMS), and the people at National Undersea Research Center in Key Largo and Mote Marine Lab in Summerland Key, Florida. v

PAGE 6

To my tireless and optimistic field assistants (Jason Hale, Victor Bonito, Thomas Kelley, Luis Rocha, George Dennis, Mary Hart, Duncan Vaughan, and Doug Marcinek) I am eternally grateful. They spent many hours underwater; they counted hundreds, if not thousands, of fishes; and endured many days subjected to the vagaries of sea lice, sinking boats, and rough seas. Some even ate their words instead of their lunch, and carried on with a smile. I cannot thank them enough. I thank my labmates (Sky Notestein, Stephanie Keller, Jason Hale, Jaime Greenewalt, Dan Goodfriend, and Kate Lazar) for their constructive criticism and sympathetic ear during moments of crisis and sheer exhaustion. Kelly Jacoby is responsible for helping with all of the figures and tables, and she always offered her help generously and with a smile. I also appreciate the folks in Dr. Gustav Paulays lab, who contributed greatly to my experience, both scientific and personal, at the University of Florida. Dr. Gustav Paulay, Chris Meyer, Lisa Kirkendale, and John Starmer each helped develop my scientific thoughts, and often reviewed presentations for scientific symposia or drafts of publications. I am exceptionally grateful to Victor Bonito for his cheerful encouragement, infectious precision, timely pep talks, and child-rearing discussions. As a scientist, he is a model of keen logic and intrigue; and as a friend, he is as good as they get. Financial assistance for this research was provided by several organizations. Dr. Russ Hall and Dr. Nick Funicelli at the Biological Resources Division; and Dr. Suzette Kimball of the Eastern Regional Office of the Biological Resources Division of the USGS supported my development as a scientist both administratively and financially. Dr. Gary Brewer (of the USGS) has always believed in the USGS Coral Reef Program; vi

PAGE 7

and Dr. Jeff Keay and Dr. Lynn Lefebrve deserve my sincere appreciation. The American Academy for the Advancement of Scientists, the Canon Science Scholars Program, and the National Park Service generously supported all stages of this research. I also thank the Graduate School of the University of Florida for assistance with logistical, administrative, and financial concerns. I cannot adequately thank the administrative staff of the Department of Fisheries and Aquatic Sciences (Jennifer Hemelbracht, Susan Morgan, Melissa Altomari, and Sherri Giardina) for attending to the many intricate administrative details that are critical to a project of this magnitude. I also thank the administrative staff of the USGS (Christine Fadeley, Tracy Marinello, Brenda Turrentine). Considering my faithful cadre of friends, I thank my lucky stars (the Pleaides) for gifting with me such a loyal and fun group of people. They helped me achieve this goal, and made my life richer through their kindness and support. They each helped me find the balance in life, work, and family, and gently served as constant reminders not to take any of this too seriously. I extend my hearfelt thanks. Most important, I thank my family members for their everlasting support, for enduring hours of long distance calls, and for encouraging me to continue when it was most definitely not the easiest thing to do. They tirelessly listened to my soliloquies on science topics when they didnt necessarily understand, nor care to. My father, Ron Grober, passed on to me an incredible strength of will to live and learn, and gave me the tenacity of a bulldog. His recent support means more to me than I can express. I thank my mother, Hanne Nielsen, whose warmth and compassion (and carefree view of our journey through life), gave me the ability to laugh at myself during this endeavor. I thank vii

PAGE 8

my sister, Suzanne Rose, who suffered years of being dragged around in a rowboat as a child; and who as an adult has become my closest friend. Most of all, I extend my deepest thanks and appreciation to my son, Thatcher Kai, who has walked this path beside me from the beginning, and continues to be my model of optimism. He has had to sacrifice much in order for me to fulfill this goal, and has taught me so much along the way. I hope that my work contributes to a better world in his future. viii

PAGE 9

TABLE OF CONTENTS page ACKNOWLEDGMENTS.................................................................................................iv LIST OF TABLES.............................................................................................................xi LIST OF FIGURES.........................................................................................................xiv ABSTRACT.....................................................................................................................xvi CHAPTER 1 INTRODUCTION........................................................................................................1 2 INFLUENCE OF LANDSCAPE STRUCTURE ON REEF FISH ASSEMBLAGES.......................................................................................................11 Introduction.................................................................................................................12 Study Area..................................................................................................................14 Methods......................................................................................................................14 Reef fish Sampling..............................................................................................15 Habitat Sampling.................................................................................................16 Data Analysis.......................................................................................................17 Results.........................................................................................................................19 Landscape Structure............................................................................................19 Reef Fish Assemblage Structure.........................................................................20 Discussion...................................................................................................................22 Conclusions.................................................................................................................29 3 EVIDENCE OF FUNCTIONAL CONNECTIVITY IN A CORAL REEF ECOSYSTEM.............................................................................................................45 Introduction.................................................................................................................46 Study Area..................................................................................................................49 Methods......................................................................................................................50 Reef fish Sampling..............................................................................................50 Temporal Sampling.............................................................................................51 Habitat Sampling........................................................................................................51 Statistical Analyses..............................................................................................52 ix

PAGE 10

Results.........................................................................................................................53 Entire Assemblage Level Parameters..................................................................54 Abundances within Reef Fish Groups.................................................................54 Species Richness.................................................................................................55 Mobility...............................................................................................................55 Spatial Extent.......................................................................................................56 Temporal Consistency.........................................................................................56 Relative Influence of Fine and Landscape-scale Measures.................................56 Discussion...................................................................................................................57 Conclusions.................................................................................................................64 4 REEF FISHES RESPOND TO VARIATION IN LANDSCAPE STRUCTURE.....79 Introduction.................................................................................................................80 Methods......................................................................................................................84 Study Areas.........................................................................................................84 Habitat Sampling.................................................................................................86 Reef Fish Sampling.............................................................................................87 Statistical Analyses..............................................................................................88 Results.........................................................................................................................91 Discussion...................................................................................................................97 Conclusions...............................................................................................................105 5 SUMMARY..............................................................................................................120 APPENDIX REEF FISH DATABASE..........................................................................134 LIST OF REFERENCES.................................................................................................145 BIOGRAPHICAL SKETCH...........................................................................................163 x

PAGE 11

LIST OF TABLES Table page 2-1 Reef fish assemblage parameters (n = 30) used as dependent variables in statistical analyses....................................................................................................31 2-2 Fourteen metrics used to quantify the landscape structure of the 20 study reefs sampled in 1994 and 2001 in St. John, USVI..........................................................32 2-3 Summary statistics on reef configuration, context and rugosity and for select reef fish assemblage parameters (entire assemblage level, trophic level and mobility guilds) for 20 study reefs sampled in 1994 and 2001, St. John, USVI for metrics at the 100 m spatial extent with coefficient of variation for landscape parameters and standard error for reef fish parameters..............................................................33 2-4 Pearson product moment correlation matrix of the 14 landscape-scale habitat variables, with the resultant 9 remaining significant variables, at the 100 m spatial extent for the 20 reef sites sampled in 1994 and 2001 in St. John, USVI....34 2-5 Principal component analyses on the correlation matrix of the 8 residual landscape-scale habitat variables at the 100 m spatial extent for the 20 study reefs sampled in 1994 and 2001 in St. John, USVI..................................................35 2-6 Stepwise regression results to determine the influence of principal components on reef fish assemblage structure at the 20 study reefs sampled in 1994 and 2001 in St. John, USVI at the 100 m spatial extent..........................................................36 2-7 Stepwise multiple regression results of the influence of reef configuration on reef fish assemblage structure for the 1994 (N = 14) study reefs in St. John, USVI.........................................................................................................................37 2-8 Stepwise multiple regression results of the influence of reef context on reef fish assemblage structure for the 1994 (N = 14) study reefs in St. John, USVI.............38 2-9 Stepwise multiple regression results of the relative influence of landscape and fine-scale habitat measures on reef fish assemblage structure on the 1994 (N = 14) study reefs on St. John, USVI............................................................................39 3-1 Most abundant taxa in each reef fish group for the 22 study reefs sampled in 2002 in St. John, U.S. Virgin Islands.......................................................................71 xi

PAGE 12

3-2 Variable names, transformations, minimum, maximum and mean values for each reef fish parameter and landscape metric, with the standard error for fish parameters and the coefficient of variation for habitat measures for the 22 study reefs sampled in 2002 in St. John, US Virgin Islands..............................................72 3-3 Simple linear regression for entire assemblage level parameters of reef fish communities with the areal coverage of seagrass within 1 km of each study reef as the independent variable at the 22 study reefs, sampled in 2002 in St. John, U.S. Virgin Islands*.................................................................................................73 3-4 Simple linear regression of abundances of mobile invertebrate feeders, grunts, snappers, groupers, and seagrass-associated taxa and within mobility guilds with the areal coverage of seagrass within 1 km of each study reef at the 22 study reefs, sampled in 2002 in St. John, U.S. Virgin Islands*.........................................74 3-5 Simple linear regression of abundances of the adult and juvenile components for grunts, snappers, groupers, and seagrass-associated taxa with the areal coverage of seagrass within 1 km of each study reef at the 22 study reefs, sampled in 2002 in St. John, U.S. Virgin Islands................................................................................75 3-6 Simple linear regression analyses of cumulative species richness of MIFs, haemulids, epinephelids, lutjanids and within resident, mobile and transient mobility guilds the areal coverage of seagrass within 1 km of each study reef as the independent variable at the 22 study reefs, sampled in 2002 in St. John,US Virgin Islands...........................................................................................................76 3-7 Spearman rank correlations of relationships of each fish parameter and the areal coverage of seagrass at the 250 m spatial extent for the 8 study reefs in St. John, sampled in 2002 and 2003........................................................................................77 3-8 Influence of fine-scale (rugosity) and landscape-scale (seagrass) features in predicting reef fish parameters for the 22 study reefs sampled in 2002...................78 4-1. Reef fish assemblage parameters (n = 30) used as dependent variables in statistical analyses..................................................................................................114 4-2. Study reef name, protection status and reef fish sampling effort for FKNMS (May 2003) and St John, USVI (July-August 2002)..............................................115 4-3. Stepwise regression results indicating the relationship of landscape configuration and reef fish assemblage structure for the FKNMS and US Virgin Islands, with the R-square, associated p-value, and explanatory habitat variable.116 4-4. Stepwise multiple regression results of the relationships of reef context and reef fish assemblage structure for FKNMS and US Virgin Islands, with the R-square, p-value, and explanatory variable by location........................................................117 xii

PAGE 13

4-5. Average dissimilarity in reef fish community structure at the trophic and family level in the FKNMS and US Virgin Islands, and the relative contribution of each trophic and family category to community dissimilarities using SIMPER analyses on standardized data................................................................................118 4-6 T-test comparison results to test for significant differences in abundance of the 30 reef fish parameters in the FKNMS (SPA and reference sites n = 8 reefs) and the US Virgin Islands (MPA and reference sites n = 22 reefs)..............................119 xiii

PAGE 14

LIST OF FIGURES Figure page 2-1 Location of St. John, US Virgin Islands in the Caribbean basin..............................40 2-2 Distribution of the 20 study reefs around the island of St. John, USVI Below the name of each reef is the number of fish point counts per reef...........................41 2-3 PCA plots of the landscape structure of the coral reef environments of the 20 study reefs sampled in 1994 and 2001 in St. John, USVI at the A) 100 m, B) 250 m and C) 500 m spatial extent..................................................................................42 2-4 Effects of reef configuration on mean fish abundances of A) transient fishes, B) adult omnivores, and C) adult piscivores for the 1994 (N =14) study reefs in St. John, USVI...............................................................................................................43 2-5 Effects of reef context on mean abundance of particular fish groups for the 1994 (N =14) study reefs in St. John, USVI.....................................................................44 3-1 Location of the 22 study reefs around the island of St. John, US Virgin Islands sampled in 2002, with the eight study reefs re-sampled in 2003 indicated in bold...........................................................................................................................66 3-2 The relationship of a) cumulative richness and b) mean species richness with the areal coverage of seagrass (hectares) at 250 m for the 22 study reefs sampled in 2002 in St. John, U.S. Virgin Islands The x-axis is log10 (x +1) transformed.......67 3-3 The relationship of mean abundances of a) MIFs, b) haemulids, c) seagrass-associated taxa and d) lutjanids with the areal coverage of seagrass (hectares) within 250 m of the 22 study reefs in St. John, U.S. Virgin Islands sampled in 2002 Mean abundances and the areal coverage of seagrass are log10 (x +1) transformed...............................................................................................................68 3-4. The relationship of cumulative species richness of a) MIFs, b) haemulids and c) lutjanids with the areal coverage of seagrass (hectares) within 250 m for 22 study reefs in St. John, U.S. Virgin Islands sampled in 2002. The y-axis is log10 (x+1) transformed.....................................................................................................69 xiv

PAGE 15

3-5 Spearman rank correlations for those reef fish parameters that demonstrated a consistent relationship with the areal coverage of seagrass habitat between 2002 ( ) and 2003 ( ) at the subset of 8 study reefs in St. John, U.S. Virgin Islands.......................................................................................................................70 4-1 Location of the 17 study reefs sampled in 2003 in the Florida Keys National Marine Sanctuary and the 22 study reefs sampled in 2002 in St. John, U.S. Virgin Islands.........................................................................................................107 4-2 Simple linear regression results of the effects of reef context (the areal coverage of pavement and reef habitat log10 (x+1)) at the 100-meter spatial scale on mean abundance of various reef fish parameters at the 17 study reefs sampled in 2003 in the FKNMS........................................................................................................108 4-3 Simple linear regression results of the only relationship that remained consistent, and significant, across systems (the relationship of patch diversity and piscivore abundances) in the FKNMS and US Virgin Islands..............................................109 4-4 Comparison of the relative proportion of mapped habitat classes within 500 m of the study reefs in the FKNMS and US Virgin Islands.......................................110 4-5 Reef fish abundanceseagrass relationships for the US Virgin Islands and the FKNMS (n = 39)....................................................................................................111 4-6 Multidimensional scaling plots of the (A) trophic and (B) family structure of reef fish communities and the (C) landscape structure of the US Virgin Islands (squares) and FKNMS (triangles)..........................................................................112 4-7 MDS plots of the (A) trophic and (B) family level reef fish community structure for reefs in the US Virgin Islands and FKNMS, with reefs classified according to high (circles), moderate (open triangles) and low seagrass (upside down closed triangles) areal coverages............................................................................113 xv

PAGE 16

Abstract of Dissertation Presented to the Graduate School of the University of Florida in Partial Fulfillment of the Requirements for the Degree of Doctor of Philosophy APPLYING TERRESTRIAL LANDSCAPE ECOLOGY PRINCIPLES TO THE DESIGN AND MANAGEMENT OF MARINE PROTECTED AREAS IN CORAL REEF ECOSYSTEMS By Linda Erica Rikki Grober-Dunsmore August 2005 Chair: Thomas K. Frazer Major Department: Fisheries and Aquatic Sciences Marine protected areas (MPAs) represent a popular, but often controversial, management option for the conservation of dwindling reef fish populations worldwide. Questions concerning appropriate design criteria for MPAs lie at the center of the controversy, and reflect a need to better understand the influence of landscape structure of coral reef ecosystems (e.g., size, shape, context of habitat patches) on reef fish assemblage structure. I explored the utility of various landscape metrics in predicting reef fish assemblage structure and found that reef context explained considerable variation in the several reef fish parameters. Specifically, I found that particular groups of fishes were associated with particular types of habitat. Based on these results, I designed a new study in the US Virgin Islands to determine examine whether functional habitat linkages between reef and seagrass habitat patches were detectable at a landscape-scale. Consistent with predictions, entire assemblage level parameters and abundances and species richness of mobile invertebrate feeders, haemulids, lutjanids, and xvi

PAGE 17

epinephelids were each significantly greater at reefs with seagrass within 1 kilometer of the study reef patch. The generality of reef context as a predictor of reef fish assemblage structure was then tested in the Florida Keys National Marine Sanctuary. Though reef context was significant in both systems, the particular habitat type responsible for the reef fish habitat relationships differed between the coral reef landscapes. Seagrass was a strong predictor of abundances and species richness of mobile invertebrate feeders, haemulids, and lutjanids in the US Virgin Islands, but was not a predictor of these same fishes in Florida. Thus, the processes that structure reef fish communities appear to respond to variation in the landscape structure of these coral reef environments. These results are relevant to marine protected areas design, since they suggest that general design rules do not necessarily apply across systems. Rather, comparative studies are critical for developing the universal design principles to locate marine protected areas that meet their conservation and/or fisheries objectives. xvii

PAGE 18

CHAPTER 1 INTRODUCTION Coral reef ecosystems are degrading worldwide with losses of biodiversity, declines in coral cover, and decreases in the average size and abundances of many coral reef fishes (Wilkinson 2000), and marine protected areas (MPAs) are gaining in popularity as the best management option for dealing with these concerns (Allison et al. 1998, Murray et al. 1999). Coral reef ecosystems are heterogeneous landscapes, comprising topographically-complex, calcium carbonate skeletal structures in which stony corals provide the major framework (Hallock 1997). Coral reefs are embedded in a mosaic of different habitat patches (e.g., reef, seagrass, open water, and mangrove forest) that are connected to one another through the movements of energy, material (e.g., fecal or detrital matter) and marine organisms (Ogden 1997) such as reef fishes. Coral reef fishes are recreationally and commercially important components of coral reef ecosystems. Reef fish communities exist as spatially divided populations that reside in this mosaic. Connections among local subpopulations are maintained by the export and import of larvae from other subpopulations; or through the movement of fishes during ontogeny, foraging, or spawning (Sale 2002). Distribution of reef fish communities is likely governed by multiple biological and environmental processes that operate at a variety of spatial and temporal scales including biological processes such as predation (Hixon and Beets 1989, Hixon 1991), competition (Smith and Tyler 1973), recruitment limitation (Sale et al. 1984, Doherty and Fowler 1994), and priority effects (Almany 2003). While studies conducted at a small spatial scale associate various fish 1

PAGE 19

2 parameters (e.g., fish density and biomass) with reef substratum complexity (Luckhurst and Luckhurst 1978, Gladfelter et al. 1980, Hixon and Beets 1989) and reef surface area (Molles 1978, Gladfelter et al. 1980), little research has explored whether these relationships scale up (but see Ault and Johnson 1998a, Acosta and Roberston 2002, Christensen et al. 2003, Jeffrey 2004). Without an understanding of the distribution of reef fish communities at large spatial scales (> 10s of meters), scientists are ill-equipped to advise resource managers on decisions that require a large-scale examination (e.g., marine protected areas). Marine protected areas (MPAs), one of the most highly advocated forms of ecosystem-based management, can provide a spatial escape for intensely exploited species (Allison et al. 1998, Murray et al. 1999). MPAs constitute a broad spectrum of areas that are afforded some level of protection for the purpose of managing resources for sustainable use, and safeguarding ecosystem function and biodiversity (Plan Development Team 1990). Their potential has been demonstrated both theoretically (Plan Development Team 1990, Roberts and Polunin 1991, Carr and Reed 1993, Allison et al. 1998) and empirically (Rakitin and Kramer 1996). They can increase average size, abundance, and biomass of exploited organisms (see reviews by Rowley 1994, Halpern 2003); and networks of MPAs can insulate habitats and communities from extractive activities that lead to losses in biodiversity and changes in species interactions (Tegner and Dayton 2000, Murray et al. 1999). Because decisions about the placement of reserves are largely political, scientists have had few opportunities to understand the biological implications of reserve design (Allison et al. 1998). Moreover, the design of MPAs has typically focused on single habitats, neglecting associated habitats that may benefit reef fishes during various

PAGE 20

3 stages of their life history (Ogden and Ehrlich 1977, Helfman et al. 1982) and failed to consider potentially-important functional habitat linkages between various habitat patches. At present, few quantitative rules exists for design and management of MPAs. In fact, criteria are broad (e.g., representation, replication). Decisions about the siting, location, size, and composition of MPAs are sorely needed in many places, yet it is currently difficult, if not impossible, to predict how alternative spatial arrangement influence the ability of an MPA to meet its stated conservation and/or fisheries objectives. Various landscape elements (e.g., the amount of edge habitat, corridor placement, and landscape connectedness) have considerable influence on the distribution of terrestrial organisms (Turner et al. 2001). Therefore it is crucial to explore the relevance of landscape elements in structuring coral reef fish communities as a prerequisite for designing effective MPAs. Because improperly designed refuges may provide a false sense of protection, and thereby endanger a fishery (Carr and Reed 1993), identifying simple metrics useful in predicting reef fish assemblage structure would be extremely valuable to resource managers. Combining the disciplines of landscape and coral reef ecology provides a logical starting point for addressing important management questions relevant to habitat-based conservation of reef fishes. The discipline of landscape ecology deals with interactions and exchanges across large areas, relating the structure of an area to its function (Forman and Godron 1986). By using geo-referenced maps of vegetation, soils, and elevation, terrestrial landscape ecologists have quantified aspects of spatial patterning using a number of metrics, including (but not limited to) patch size and shape, and total area of critical habitats (Turner 1989, Forman 1995). These metrics, calculated statistics of landscape pattern

PAGE 21

4 (Frohn 1998), can be used to predict the outcomes of ecological processes such as dispersal success (Gustafson and Gardner 1996, Schumaker 1996) and population dynamics such as density (McGarigal and McComb 1995), distribution (With and Crist 1995), community structure (Noss 1983), and survival probability (Fahrig 1997). In fact, large-scale metrics of habitat diversity have been successfully used to predict total species richness and abundance of birds, butterflies, and reptiles (Rafe et al. 1985, Rosenzweig 1995, Ricklefs and Lovette 1999). The challenge for scientists is to identify those landscape-scale metrics that may serve as proxies for resource managers of areas with high species diversity and abundance in the coral reef landscape. Particular habitats such as seagrass communities may play a key role in structuring reef fish communities. Seagrass communities serve as refuge habitat for small fishes and benthic invertebrates, and may be beneficial to the settlement, survivorship, and growth for a variety of fishes and invertebrates that spend their adult life on the reef (Parrish 1989, Baelde 1990, Ogden 1997, Nagelkerken et al. 2000). Seagrass beds are some of the most productive ecosystems of the world (Zieman and Wetzel 1980, Duarte and Chiscano 1999), often forming a dense and extensive below-ground network of roots and rhizomes that support a structurally-complex system of short shoots. A diverse epibiont community attaches to the seagrass blades (Fong et al. 2000), and infaunal organisms live in the sediment of the seagrass, acting to enrich or stabilize the substrate (Suchanek 1983, Peterson and Heck 1999). Many organisms live in seagrass patches including mollusks (Irlandi et al. 1999), crustaceans (Arrivillaga and Baltz 1999), and fishes (Ogden 1997). High densities and biomass of economically-important reef fishes (such as snappers, grunts, and groupers) have been attributed to the availability of food resources in

PAGE 22

5 surrounding seagrass habitat (Randall 1963), perhaps due to movement patterns of these exploited fish (Tulevech and Recksieck 1994, Burke 1995). No studies, however, have quantified the benefits of potentially critical habitat linkages between reef and seagrass (exception: Wolff 1996) or applied recently developed technologies such as geographic information systems, to determine whether the influences of such habitat linkages are detectable at a landscape-scale. If benefits of such habitat linkages can be detected at a large spatial scale (100s of meters), identifying and then subsequently protecting these habitat linkages (which has proven extremely valuable in terrestrial conservation) (Noss 1983, Forman 1995, Turner et al. 2001) may be feasible for delineating boundaries of MPAs in coral reef ecosystems. Experience applying a landscape ecology approach in terrestrial systems provides a framework for addressing a number of relevant resource management questions that may prove valuable in marine systems. What are the relationships among coral reef landscape structure and reef fish assemblage structure?; Can functional habitat linkages be identified at a landscape-scale, and is it possible to measure and quantify the potential consequences of these habitat linkages?; What is the appropriate spatial scale for addressing these questions?; Can faunal-habitat relationships detected in one coral reef landscape be generalized to another? To address these questions, a hypothetico-deductive study (Platt 1964, Peters 1991) was designed to explore the utility of terrestrial landscape ecology principles in the US Virgin Islands and the Florida Key National Marine Sanctuary (FKNMS) coral reef landscapes. Because little prior research had been conducted at large spatial scales in coral reef systems (exceptions: Appeldoorn et al. 2003, Kendall et al. 2003, Christensen

PAGE 23

6 et al. 2003, Jeffrey 2004), it was necessary to start by exploring relationships of coral reef landscape structure with reef fish assemblage structure (Chapter 2). In the inductive stage of my study, data were collected from the US Virgin Islands, analyzed statistically, interpreted without having a prior specific hypothesis (although generally I expected landscape structure would be correlated with reef fish assemblage structure), and results were used to generate hypotheses tested later. Specifically, I explored those measures of landscape structure that have proven valuable in terrestrial ecosystems for predicting areas of high abundance and species richness. The preceding exploratory approach, however, does not eliminate the risk of detecting patterns that are not biologically relevant or the risk of missing significant relationships that are (Bissonette and Storch 2003). Because organisms with different habitat requirements, feeding behaviors, and mobility can respond to the landscape differently (Turner et al. 2001, Sisk et al. 1997), analyses must be conducted with consideration for the natural history and ecology of the organisms of interest (Bissonette and Storch 2003). Therefore, while the entire shallow-water reef fish community was of interest to me in this study, this diverse assemblage of fishes (140 different taxa) was sub-divided into groups of species that share a common set of life-history traits, morphological or behavioral attributes, or ecological functions. The response of these functional fish groups (trophic and mobility guilds, taxonomic groups, and by life history stage) to landscape features were examined later. Because pattern exists at every spatial scale (Wiens 1989), it is necessary to link the organisms, species, or the processes being considered to the scales appropriate to the specific questions of interest (Bissonette and Storch 2003). The process for finding the

PAGE 24

7 relevant scale is not well understood (Bissonette and Storch 2003), but generally researchers recommend analyzing data at multiple spatial extents to identify concordance with response variables using spatial statistics and multiple regression procedures (Pearson 1993, Pearson et al. 1995, Pedlar et al. 1997). Particularly in spatially heterogeneous systems, where little to no work has been conducted at a landscape-scale, faunalhabitat relationships should be explored at multiple spatial extents. Therefore, four spatial extents (100 m, 250 m, 500 m, and 1 km) were explored in all subsequent chapters of this dissertation. It is also valuable to understand whether the distribution of organisms at a given location is explained by characteristics of the immediate locale (fine-scale within patch characteristics) or by the attributes of the surrounding landscape (landscape characteristics), so resource managers can evaluate the trade-offs of collecting patch-level or landscape-level information. For example, Pearson (1993) found that some birds responded only to characteristics of the local habitat (vegetation characteristics such as height, density, and species composition), while other bird species responded only to landscape context (amount of each habitat surrounding study areas). Thus, detailed within-patch surveys may be required to predict the presence of some species, whereas remotely-sensed surveys may be sufficient to detect the presence of others. In coral reef ecosystems, various within-patch measures of coral reef patch quality have been shown to influence reef fish abundance and diversity. These patch-level measures include coral cover (Bell and Galzin 1984) and topographic complexity (Luckhurst and Luckhurst 1978, Hixon and Beets 1989, Friedlander and Parrish 1998), therefore, the fine-scale measure of rugosity (Luckhurst and Luckhurst 1978) and various landscape-scale metrics

PAGE 25

8 of habitat were evaluated to determine their relative influence in structuring reef fish communities for all analyses. As progress in science is ideally made by the sequential development of hypotheses and the execution of experiments designed to test these hypotheses (Platt 1964, Quinn and Dunham 1983), ecologically-meaningful, and significant findings from the exploratory analyses were used to derive testable hypotheses for further study (Chapter 3). Of particular promise were metrics of reef context, which quantify the spatial arrangement and composition of surrounding habitat patches. A landscape patch, by definition, is bounded by something else, and these adjacent or proximal habitat patches can strongly influence organisms within that patch (Turner et al. 2001). Correspondingly, scientists increasingly recognize that continental reserves are not islands surrounded by a neutral sea (Janzen 1986); rather isolating a reserve can lead to ecosystem degeneration and the extent and rapidity of this degeneration can depend upon the ecological condition of adjacent habitat patches (e.g., Kushlan 1979). Consequently, I examined how reef context influenced reef fish community structure. Specifically, I tested hypotheses that entire assemblage level parameters, and abundances and species richness within trophic and taxonomic groupings, would be higher at reefs with seagrass within 1 kilometer of the study reef patch. Although considerable attention has been dedicated to the temporal variability inherent to natural communities, surprisingly little research has addressed the temporal consistency of faunalhabitat relationships, particularly those that occur at a landscape scale (Turner et al. 2001, Bissonette and Storch 2003). Many landscape studies have been discrete sampling events, and have failed to consider temporal variation in resources

PAGE 26

9 or reproductive opportunities (Turner et al. 2001). Coral reef fish communities are notoriously dynamic, both spatially and temporally (Sale et al. 1984, Sale 2002), therefore, their temporal dynamics represent an important ecological dimension. Temporal variations in fish abundance may be induced by reproductive movements (Colin et al. 1997), ontogenetic shifts (Appeldoorn et al. 1997), feeding migrations (Ogden and Zieman 1977, Ogden and Quinn 1984), spatially segregated foraging and resting locations (Meyer et al. 2000), and spatial heterogeneity within and among habitat patches. If landscape metrics of the coral reef landscape are to prove valuable in understanding the distribution and abundance of coral reef fishes, a more thorough understanding of the temporal consistency of reef fishhabitat relationships is critical. To begin to meet this need, a portion of this study (Chapter 3) was conducted over 2 years. Detection of faunalhabitat relationships in one system does not necessarily imply that resource managers can expect organisms in another system to respond to analogous landscape features in the same manner. Because each landscape is unique, the size and distribution of habitat differs, and may putatively exert different landscape-specific constraints (Bissonette and Storch 2003) on species richness, abundance, community structure, recruitment, and movement. If the constraints of each landscape result in qualitatively different responses of the reef fishes without apparent thresholds, then we may have no hope of developing a predictive theory for forecasting the fish assemblage structure of a given reef patch. Thus, in order to determine the generality of reef fishhabitat relationships detected in the insular US Virgin Islands to other Caribbean coral reef ecosystems, the study was replicated (Chapter 4) spatially in the Florida Keys National Marine Sanctuary.

PAGE 27

10 The approach selected in this study is beneficial in that it addressed many of the short-comings of terrestrial landscape studies (Turner et al. 2001) by replicating the study both in time and space. In addition, multiple aspects of this research remain consistent throughout the dissertation, allowing insight into the generality and reliably of various measures in other systems and over time. For example, each chapter examines the influence of landscape pattern to the entire reef fish community, and then focuses on trophic and mobility guilds and taxonomic groupings of fishes. Furthermore, because there is no single correct spatial scale to describe a system (Levin 1992), a multi-scalar approach was adopted. Every study analysed relationships at multiple spatial scales including the fine-scale within-patch characteristics and explored the strength of relationships at various landscape scales (100 m 1 km). While as a discipline, terrestrial landscape ecology has progressed from purely non-quantitative descriptive studies (Wiens 1992) to increasing emphasis on spatial statistics, modeling, and experimental design (Hobbs and Norton 1996), this dissertation also progresses from the exploratory to the testing of specific hypotheses. These hypotheses are tested over time and space in a robust study design, thus insights derived from this dissertation provide a foundation for applying the principles of landscape ecology to tropical marine systems, and improve our understanding of the relationships of reef fish communities to large-scale habitat features.

PAGE 28

CHAPTER 2 INFLUENCE OF LANDSCAPE STRUCTURE ON REEF FISH ASSEMBLAGES Marine protected areas represent a popular, but often controversial management option for the conservation of dwindling reef fish populations worldwide. Questions concerning appropriate design criteria for marine protected areas lie at the center of the controversy, and reflect a need to better understand the influence of landscape structure of coral reef ecosystems (e.g., size, shape, and context of habitat patches) on reef fish assemblage structure. Herein, I investigated the relationships between landscape structure and reef fish assemblage structure at 20 study reefs around the island of St. John, US Virgin Islands. Various measures of landscape structure were calculated and transformed into a reduced set of composite indices using principal component analyses (PCA) to synthesize data on the spatial patterning of the study reefs. However, composite indices (i.e. measures of habitat diversity) were not particularly informative for predicting reef fish assemblage structure. Rather, relationships were interpreted more easily when functional groups of fishes were related to individual habitat features. In particular, reef context was strongly associated with multiple reef fish parameters (e.g., abundances within trophic guilds and taxonomic groups). Fishes responded to benthic structure at multiple spatial scales, with each fish group correlated to a unique suite of variables. Accordingly, future experiments should be designed to test functional relationships based on the ecology of the organisms of interest. My study illustrates promise in applying a landscape ecology approach to coral reef ecosystems, and provides 11

PAGE 29

12 an empirical basis to further test the influence of specific habitat features in structuring reef fish communities. Introduction The management of tropical marine environments calls for interdisciplinary studies and innovative methodologies that consider processes occurring over broad spatial scales (Allison et al. 1998). Landscape ecology is interdisciplinary by nature, with an appropriate focus on broad-scale patterns and ecological processes (Forman and Godron 1986). A landscape generally refers to a heterogeneous area composed of local interacting ecosystems (Forman 1995) made up of homogenous units, called habitat patches. Landscape structure describes the composition and spatial arrangement of habitat patches (Forman and Godron 1986), and has been quantified using a number of metrics (ONeill et al. 1988) including composite indices (e.g., habitat diversity, principal components), measures of configuration (e.g., patch size), and measures of context (composition of surrounding habitat patches) (Turner 1989). The use of such metrics, derived largely from island biogeography theory (MacArthur and Wilson 1967), metapopulation theory (Hanski 1999), and patch dynamics (Pickett and White 1985) has improved our understanding of how landscape features influence terrestrial communities (Turner 1989, Gardner and ONeill 1991). Because of its focus on broad and multiple spatial scales and entire ecosystems, landscape ecology has proven extremely valuable in addressing management problems in terrestrial systems (e.g., reserve design) (Noss 1983, Forman 1995). A landscape ecology approach to the study of coral reef fishes, however, has received little attention, until recently (Kendall et al. 2003, 2004, Jeffrey 2004). Our understanding of the dynamics of reef fish assemblages has been largely derived from studies conducted at fine spatial scales (1 m2 plots) (Williams 1980, Sale et

PAGE 30

13 al. 1994, Pittman and McAlpine 2003), limiting our ability to predict the effects of large-scale features on reef fishes (Sale 2002). Fine-scale measures of topographic complexity (Hixon and Beets 1989, Friedlander and Parrish 1998), hole size (Friedlander and Parrish 1998), and coral cover (Bell and Galzin 1984) can influence reef fish assemblage structure. The few existing large-scale studies have examined relatively gross characteristics such as latitudinal gradients (Ebeling and Hixon 1991) and coral reef zonation (Williams 1991); features that are not particularly useful for selecting specific reef areas as candidates for protection. Thus, it is unclear whether findings from fine-scale studies can be extrapolated to large-scale resource management concerns. Understanding functional relationships between landscape structure and reef fish distribution at a broad spatial scale may therefore be useful for delineating the boundaries of MPAs (Christensen et al. 2003), since coral reef ecosystems exist as a complex mosaic of habitat patches (i.e. reefs, seagrass patches, and mangrove stands), and are therefore ideally suited for a landscape ecology approach. The purpose of my study was to determine whether commonly-used terrestrial metrics can be quantified for coral reef environments, and to determine whether these metrics might be used to predict reef fish assemblage structure. Toward this end, a suite of landscape metrics was explored using multivariate statistics. Through a variety of procedures to eliminate redundancy and autocorrelation, I developed composite indices to synthesize data on the spatial patterning of the reef study sites. I then examined the utility of these composite indices to predict reefs that have relatively high reef fish species diversity and abundance. In addition, I explored individual habitat features (reef configuration and reef context) separately, and examined the relative importance of fine

PAGE 31

14 and landscape-scale habitat measures on reef fish assemblage structure. Configuration was selected because measures of mean patch size, shape, and arrangement are frequently associated with species abundance and diversity (Robinson et al. 1995, Villard et al. 1999); and protected area configuration can influence organisms within and outside their boundaries (Diamond 1975, Sisk et al. 1997, Mazerolle and Villard 1999). Context was selected because of the increasing recognition of its role in sustaining species targeted for conservation in terrestrial systems (Mladenoff et al. 1995, Robinson et al. 1995, Sisk et al. 1997). The relative influence of fine-scale (rugosity) and landscape-scale habitat characteristics on fish assemblage structure was explored because knowing the importance of features at each spatial scale can save precious resources for resource managers (Mazerolle and Villard 1999). Study Area Coral reefs around the island of St. John, US Virgin Islands (Figure 2-1) were selected for study, because benthic habitat were readily available (Kendall et al. 2001). Habitat maps (digitized from aerial photographs taken at an altitude of 5000 feet in 1999), were classified by visual interpretation by NOAA, using 26 discrete and non-overlapping habitat classes, with a minimum mapping unit of 1 acre (Kendall et al. 2001). Most sampling sites were located on the lower fore reef of fringing and patch reefs, dominated by Montastraea annularis or mixed corals (8-30% living cover; pers. obs.), although several were dominated by old Acropora palmata framework (5-10 % living cover; pers. obs.). Study reefs occurred in water depths between 5 and 15 m. Methods Twenty reefs were sampled: 14 reefs in 1994 and 6 in 2001 (Figure 2-2). Reefs were selected from an existing fish database as representative locations that varied with

PAGE 32

15 respect to landscape features, yet relatively similar depth, reef morphology, and coral cover. For exploratory analyses, the 1994 and 2001 datasets were combined. Reefs sampled in 2001 were included to expand gradients in several habitat parameters of interest. To investigate specific functional relationships (e.g., configuration, context), only the 1994 dataset was used to reduce potential temporal variability due to changes in fishing pressure and storm damage. Reef fish Sampling Fish sampling was conducted within reef habitat only. Reef-associated fishes were sampled over a 10-day period in July 1994 (Beets and Friedlander 1994), and over a 5-day period in July 2001. The number of fish point counts per reef were determined based on reef size following Monte Carlo simulation, and ranged from 8-20 point counts per reef (Figure 2-2). In 1994, a modified Bohnsack and Bannerot (1986) point count method (a reduction in the sample radius from 7.5 m to 5 m) was used, whereas the original point count method was used in 2001. Mean species richness refers to the mean number of species observed per point count per replicate reef, whereas cumulative richness refers to the total number of species observed during all point counts at a reef. Abundance refers to the mean number of reef-associated fishes observed per point count per replicate reef. Two species were eliminated from abundance analyses, since these tended to overwhelm abundance estimates and are difficult to count accurately--Jenkinsia spp. (herring) and Coryphopterus personatus (masked/glass goby). Randall (1967) and Fish Base (Froese and Pauly 2002) were used as references to classify all fishes by trophic guild: piscivore, herbivore, mobile invertebrate feeder (MIF), sessile invertebrate feeder (SIF), planktivore, or omnivore (Appendix A). To the degree that the ecology of each species is known, each fish was classified into mobility guilds: resident, mobile or

PAGE 33

16 transient (Appendix A). Resident species are sedentary and site-attached, and do not typically move from their primary reef patch. Mobile species are those that have restricted movements and may roam from the primary reef patch. Transient species are vagile, and can range on the scale of kilometers. Taxonomic groups of commercially and ecologically-important fishes were analyzed separately (e.g., haemulids, lutjanids, and scarids). Fishes were further subdivided into juvenile and adult categories, based on length of maturity where possible (Froese and Pauly 2002), to examine the influence of life-history stage on functional relationships. This resulted in 30 reef fish assemblage parameters (Table 2-1). Habitat Sampling The fine-scale measure of rugosity was obtained by running an underwater tape measure as closely as possible over the contour of the substratum. For each reef, 10 rugosity samples were collected along 10-m transects. The resultant mean value was used for subsequent analyses. Because these reefs are natural habitat patches, microhabitat variation that was not quantified, likely exists. The original map classification scheme (Kendall et al. 2001) was condensed from 26 to 9 distinct and non-overlapping habitat classes (i.e. mud, mangrove, sand, reef, pavement, bedrock, seagrass, macroalgae, and deep unknown). Deep unknown was typically a deep (> 16m) soft-bottom habitat (R. Grober-Dunsmore, unpubl. data). This reduced set of nine habitat classes was selected based on terrestrial studies (which frequently use 5-10 habitat classes for resource management purposes; Turner et al. 2001), and to simplify results for resource managers. Haphazard groundtruthing was conducted in each habitat polygon within 100 m of each study reef. Percent cover of benthic invertebrates and substrate types were estimated using 1m2 quadrats.

PAGE 34

17 Fourteen metrics, used to quantify various aspects of the configuration and context of the study reefs (Table 2-2), were calculated with ArcView 3.2 (ESRI 1996). A single value for each (reef) patch metric was calculated (n = 3). Each landscape metric was calculated at three spatial extents; 100 m, 250 m, and 500 m from the leading edge of each reef. These extents were selected to represent a range of potential importance based on the known natural history of reef fishes. Because the area of deep unknown increased considerably beyond 500 m, the 1 km spatial extent was not included, resulting in 36 metrics that were correlated to reef fish parameters (Table 2-3). Most sample reefs were determined to be subsections of larger mapped polygons, thus each reef polygon was slightly modified in Arc View 3.2 (ESRI 1996) to reflect the standardized 25,000-m2 subsections where fish data were collected. A separate heterogeneity study (Grober-Dunsmore et al. 2004) determined that for most fish parameters, no significant differences existed between these reef sections and the entire mapped polygons. These results, and the fact that slightly-modified polygons were used for all calculations, justify the use of these reefs for exploratory purposes. Data Analysis To reduce the 36 landscape metrics into a more parsimonious dataset of composite indices that capture the wealth of information contained within the original dataset, a combination of techniques was applied. (1) Pearson product-moment correlations (Ppmc) between each pair of metrics, and (2) principal component analysis (PCA) using a correlation matrix. Each spatial extent was examined separately and results were compared across extents. Ppmc was applied sequentially by examining significant pair-wise correlations (Sokal and Rohlf 1995) to reduce the number of variables to a 3:1 ratio (observations to variables), which is required for PCA (McGarigal

PAGE 35

18 et al. 2000). The choice of an index within a group of redundant metrics was determined by selecting ecologically-meaningful metrics and eliminating variables that failed normality tests. Ppmc reduced the 14 metrics to 8, and PCA was subsequently applied to further synthesize these into a smaller set of linear combinations (components) of the original variables. Loadings on original variables were used for interpretation. Principal component plots of these landscape metrics were used to organize the sampling entities (reefs) in multivariate space. Reefs were classified into one of three categories based on our pre-existing knowledge of the surrounding coral reef environments. PCA plots were then displayed at each spatial extent to determine whether characterization of study reefs using landscape metrics corresponded to our local knowledge and groundtruthing data, allowing us to assess the broad accuracy of remotely sensed benthic habitat maps. To explore the strength and nature of the relationships between landscape structure and fish assemblage parameters, stepwise multiple regression analyses using significant principal components as the independent variables, for each of the thirty reef fish parameters, was conducted. To control family-wise error rate for multiple correlations, sequential Dunn-Sidak Bonferroni corrections were applied using the number of reef fish parameters (n = 30) tested (Sokal and Rohlf 1995) for all subsequent regressions. Specific functional relationships were examined separately: 1) reef configuration, 2) reef context, and 3) the relative influence of fine and landscape-scale habitat parameters, using the 1994 reef fish dataset only (n = 14). Landscape variables were selected based on Ppmc results (Sokal and Rohlf 1995) though several variables that were

PAGE 36

19 eliminated from PCA were included to verify their potential importance. Only those that met assumptions of statistical independence were tested in a given model. Reef configuration variables were perimeter to area ratio (P: A) of each reef, reef size, and the number of habitat patches within 100 m. Reef context variables were surrounding habitat diversity and the areal coverage of reef, bedrock, seagrass and deepwater within 100 m. Fine-scale and landscape-scale variables were rugosity and the areal coverage of deepwater, seagrass and reef within 100 m. To optimize model performance and reduce potential effects due to multicollinearity, a series of diagnostic tests were used: 1) Akaikes Information Criterion (Akaike 1974), 2) leverage effects plots, 3) Durbin-Watson statistic, and 4) condition number (Belsley et al. 1980) for every stepwise regression analysis. Simple linear regressions were created to determine the stability of models using residual plots and residual normality plots (Sokal and Rohlf 1995). Where necessary, reef fish and habitat data were log10 (x + 1) transformed to improve normality and data were tested using Shapiro-Wilks statistic (Sokal and Rohlf 1995). All statistical analyses were conducted with JMP 8.01 (SAS 2003). Statistical significance was accepted at the p 0.05, unless otherwise noted. Results Landscape Structure Configuration and context of study reefs varied widely. Most metrics had coefficients of variation > 50 % of the mean, indicating that gradients in many aspects of the landscape were represented (Table 2-3). Of fourteen initial landscape variables, Ppmc resulted in 8 remaining landscape metrics (Table 2-4). PCA results were consistent across spatial extents (i.e. 100 m, 250 m, 500 m). In fact, the variance explained, the eigenvalues and the distribution of loadings were remarkably comparable. However,

PAGE 37

20 interpretation was easiest at the 100 m spatial extent since the proportion of area classified as deep unknown was minimized. Thus, a single spatial extent (100 m) was selected for further analyses, using the remaining 8 landscape variables: number of habitat patches, reef size, habitat diversity, area of deep unknown, pavement, reef, sand and seagrass habitat. PCA of these 8 landscape metrics at 100 m revealed four dominant components of variation based on retention of eigenvalues greater than the average, i.e. > 1 (Jackson 1993). These components explained approximately 80 % of the total variance of the original landscape variables (Table 2-5). However, landscape structure was not adequately represented by a single or even a few gradients. Final communalities indicated that most of the residual configuration indices were well accounted for by the four components, with no notable exceptions. Principal component plots using landscape metrics generally corresponded with my pre-existing knowledge and groundtruthing of the local environments of these study reefs (Figure 3), though there were several outliers. PCA plots were also in general concordance across spatial extents (100 m, 250 m, and 500 m), although the 100 spatial extent resulted in strongest clustering of reefs in concordance with my local knowledge. Thus, PCA plots illustrated that benthic habitat maps were capable of differentiating among reef types (Figure 2-3). Reef Fish Assemblage Structure A total of 57,002 fishes representing 171 different species were recorded during 341 censuses at the 20 study reefs. Principal components proved to be useful in explaining only a limited number of reef fish assemblage parameters. Both measures of species richness (i.e. mean richness

PAGE 38

21 and cumulative richness) were marginally correlated (21% and 26%) with PC4, a positive gradient of seagrass (Table 2-6). Fifty-three percent of the variation in herbivore abundance was explained by PC2 and PC3, positive gradients of sand and habitat diversity. Acanthurids (a major component of the herbivore guild) were also positively correlated to PC2 (Table 2-6). Forty-three percent of haemulid abundance was negatively correlated to PC1, and 46 % of lutjanid abundance was negatively correlated to PC1 and positively related to PC2 (Table 2-6). Configuration was generally a poor predictor of reef fish assemblage structure. There were a few exceptions. Seventy-four percent of the variation in the abundance of transient fishes (e.g., jacks) was explained by P: A of each reef and the number of habitat patches (Table 2-7, Figure 2-4). Abundances of two other trophic guilds (piscivores and omnivores) and three taxonomic groups (pomacentrids, acanthurids, and pomacanthids) were marginally correlated to P: A (Table 2-7, Figure 2-4). Importantly, examination of regression plots revealed the influence of single points, and residual plots revealed that several relationships exhibited heteroscedascity, thus calling into question the stability of these relationships (Sokal and Rohlf 1995). No reef fish assemblage parameter was correlated with reef size. Reef context was correlated with thirteen of thirty possible reef fish assemblage parameters (Table 2-8). Species richness was positively correlated with the areal coverage of seagrass (Table 2-8). Several ecologically-relevant relationships between specific habitat types and abundances within trophic and taxonomic groups were also evident. Adult mobile invertebrate feeders (i.e. 64 species) were positively correlated with the areal coverage of seagrass (R2 = 0.33) and adult piscivores were positively

PAGE 39

22 correlated with the areal coverage of reef within 100 m (R2 = 0.51) (Table 2-8, Figure 2-5). Several taxonomic groups were predicted, based on their life history, to be correlated with a particular habitat type, e.g., adult haemulids and lutjanids with seagrass and adult serranids with reef habitat. Simple linear regressions, based on stepwise results, were generally consistent with these predictions (Figure 2-5). Fifty-three percent of the observed variation in the mean abundance of adult haemulids, and 68 % of the variation in the mean abundance of adult lutjanids was explained by seagrass coverage (Figure 2-5). Juveniles of several groups of fishes were correlated with deep unknown habitat (Table 2-8). Subsequent examination of residuals plots and residual normality plots indicated that most relationships were stable; those that were not were eliminated from the results reported here. Discussion The coral reef landscape variables were successfully reduced into four principal components, thereby synthesizing the wealth of information contained within the benthic dataset (ONeill et al. 1988, Riitters et al. 1995). When plotted, these components were able to differentiate between reefs of varying configuration and habitat composition, thus benthic habitat maps clearly appear useful in describing and quantifying coral reef landscapes. Clustering of reefs along PCA axes remained consistent across multiple spatial extents (100, 250, and 500 m), a result that represents an important contribution to the application of landscape ecology principle to coral reef ecosystems, since it is critical to determine the relevant scale of analyses for landscape studies (Gardner and ONeill 1991). Because with increasing spatial extent, the total area of deep unknown increased, I recommend that the 100 m or 250 m spatial extent is the most informative characterization of the spatial patterning of the coral reef landscape, at least in this

PAGE 40

23 system. Selection of scale is a defining challenge since different patterns emerge at different scale (Wiens 1989), and the appropriate spatial extent may depend on the benthic habitat maps available and the question of interest. Ecologically-meaningful interpretation of the principal components proved difficult, because loadings were distributed across many variables, and no single component accounted for > 26 % of the variability contained in the original dataset. These findings differ from many terrestrial studies (McGarigal and McComb 1995, Riitters et al. 1995), where distinct elements of the landscape are often described by components that represent habitat complexity (Riitters et al. 1995), fragmentation (Andrn 1994), or patch shape (McGarigal and McComb 1995). This analysis suggests that complex composite indices are less informative than individual spatial features for characterizing the coral reef landscape at the individual reef scale. In general, principal components were poor predictors of reef fish assemblage structure since most relationships were more easily interpreted using individual habitat features. For instance, species richness exhibited a marginal association with PC1 and PC4, which represented gradients in the areal coverage of reef and seagrass area, habitats considered critical for many reef fish species (Ogden and Zieman 1977, Sale 2002). Likewise, the association of herbivores with PC2 (a gradient of the areal coverage of shallow sand and seagrass habitat), may indicate the availability of shallow foraging habitat, where sufficient light is available for photosynthesis of their primary food source, algae. When testing new approaches in a new system, it is critical to determine the appropriate measures for understanding the spatial distribution of organisms (Turner 1989, Wiens 1992), thus this negative result may help guide future coral reef studies.

PAGE 41

24 Several factors may explain the inability of principal components to predict reef fishes. These components may contain too much information to be germane to reef fishes since fish may not respond to multiple habitat parameters. Rather, interpretation suggests that specific fish groups respond to specific habitat features. Additionally, it appears that all species may not conform to the same landscape pattern, as in terrestrial systems (Mladenoff et al. 1995, Lindenmayer et al. 2003), but that each organism may respond to specific features at particular spatial scales. The other composite index, habitat diversity, was also not a good predictor of reef fish diversity and abundance, which is contrary to predictions based on terrestrial research (Rafe et al. 1985, Ricklefs and Lovette 1999). These findings may indicate that this is not an appropriate measure of habitat diversity, since relationships can heavily depend upon the specific definition (e.g., elevation, vegetation structure) of habitat diversity (Rafe et al. 1985, Turner 1989). In Palau, another habitat diversity measure also failed to predict species diversity and species richness (Donaldson 2002) of reef ifshes though Jeffrey (2004) found that measures of habitat richness and diversity were correlated (both positively and negatively) with several measures of trophic composition and occurrence of several species of fishes. I found frequent negative associations of individual reef fish parameters with habitat diversity, which may suggest that specific habitat types are likely to be better predictors of assemblage structure than habitat diversity per se. These findings lead us to concur with terrestrial studies that challenge the effectiveness of generic landscape indices (i.e., principal components and habitat diversity indices) to design protected areas (Lindenmayer et al. 2003) at the scale of individual reefs. Relationships in this study were better understood by examining

PAGE 42

25 specific habitat features, therefore future studies may need to be designed to examine specific functional relationships between particular groups of fishes and specific habitat features. Although useful in some terrestrial systems (Andrn 1994), but highly variable and weak in others (Trzcinski et al. 1999), configuration measures were generally not effective in predicting reef fish assemblage structure. These findings corroborate those of Pittman et al. (2004), which revealed that configuration explained less of the variation in the spatial distribution of fishes than habitat composition. There were a few potentially, ecologically-relevant relationships. The strong, positive association of reef P: A with abundances of transients (e.g., jacks, yellowtail snapper) may reflect their foraging behavior along reef edges. The negative association of reef P: A with adult piscivores was surprising. Perhaps larger transient predators prey on smaller piscivorous fishes along the edge, reducing their abundance. In addition, fish traps are typically set along reef edges in St. John, and other island locations, thus piscivores along reef edges may be more susceptible to fishing mortality, which may explain, in part, these findings. Surprisingly, reef size was not positively correlated with any reef fish parameter. These findings contrasts with terrestrial (Diamond 1975), small-scale patch reef (Molles 1978, Bohnsack and Talbot 1980, Sale et al. 1994) and seagrass studies (Irlandi et al. 1999, Hovel and Lipcius 2001), but may be a consequence of the limited gradient in reef size in this study (though the coefficient of variation was > 67 % of the mean). It is possible that beyond a minimum reef size (which these reefs may exceed), the structure of reef fish communities may be mediated by other factors such as reef context (see below), physical disturbance (Syms 1998), larval supply (Sale et al. 1984, Doherty and

PAGE 43

26 Fowler 1994), and/or predation (Hixon and Beets 1989). Such scale effects have been demonstrated in reef communities; e.g., the paradigm of locally-controlled recruitment from a superabundant pool of larval fish, developed at the scale of local populations (Smith and Tyler 1972) has been shown to be invalid at the scale of the whole reef population (Sale et al. 1984). Reef context appears to be an important determinant of reef fish assemblage structure, corroborating findings involving multiple taxa in terrestrial (McGarigal and McComb 1995, Mazerolle and Villard 1999, Trzcinski et al. 1999), coral reef (Kendall et al. 2003) and seagrass systems (McAlpine et al. 2004). In particular, the areal coverage of seagrass, an important nursery and larval settlement habitat (Shulman and Ogden 1987, Ogden and Zieman 1977) and foraging area for some fishes (Randall 1967) was strongly associated with entire assemblage parameters (e.g., cumulative species richness). Seagrass habitat may contribute to higher species richness as a result of nutrient transfer and movement of invertebrates and energy from highly productive seagrass to adjacent reef habitat (Duarte 2000). For instance, Tektite and Yawzi reef, structurally complex reefs with the highest mean species richness values, are located within a bay with dense Thalassia testudinum. Higher species richness was also detected in mangroves adjacent to continuous seagrass in Australia (Pittman et al. 2004), and at reefs proximal to nursery habitats (i.e. seagrass) in Colombia using large-scale habitat maps habitats (Appeldoorn et al. 2003), although in coral reef systems researchers were not able to eliminate confounding factors of near shore-offshore effects nor separate independent contributions of other soft-bottom habitats.

PAGE 44

27 Reef context was also strongly associated with abundances within specific trophic guilds and taxonomic groupings. As expected based on terrestrial (Turner 1989, Sisk 1997) and marine research (Pittman and McAlpine 2003, Pittman et al. 2004), relationships that met the model selection criteria were consistent with the ecology of each particular fish group. For example, the positive relationship of MIF abundances with seagrass is consistent with the foraging behavior of species in this trophic guild (e.g., taxa within mullidae, haemulidae, and lutjanidae). The relationship for haemulids and seagrass was even stronger, which is expected since some haemulids forage off-reef in seagrass nocturnally (Ogden and Quinn 1984). Common piscivorous fishes, which may forage preferentially in reef habitat, such as Aulostomus maculatus, Carynx sp., Scomberomorous regalis, Synodus intermedius were more abundant where there were large areas of reef habitat (e.g., Eagle Shoals and Tektite). The positive association of juvenile omnivores with deep water habitat is consistent with the functional role of deep water as a source of ichthyoplankton. Several species of omnivores (e.g., apogonids, blenniids) are fairly non-mobile as larvae, and are thought to recruit directly to reef substrate from the plankton. A direct test of this hypothesis, however, would be required to determine whether deep water enhances planktonic larval delivery at the reef scale. These findings suggest that a landscape ecology approach can be valuable for identifying functional linkages between organisms and their coral reef habitats at a scale appropriate for resource management decisions. The fine-scale measure of rugosity was of limited value in predicting reef fish assemblage structure. The exception was for highly site-attached fishes, e.g., omnivores which were primarily blenniids, gobiids, and pomacentrids. The inability of rugosity to

PAGE 45

28 predict reef fish assemblage structure, though contrary to previous small-scale research (Hixon and Beets 1989), may indicate the ineffectiveness of this measure to characterize topographic complexity within a single habitat at the scale of whole reefs. Most previous studies that detected rugosity relationships were conducted across multiple habitats (Friedlander and Parrish 1998) or used manipulated patch reefs to maximize the gradient of rugosity (Hixon and Beets 1989). Some reef fishes respond to habitat features at fine spatial scales, while other reef fishes respond to features at landscape scales. For several fish groups, the combination of fine and landscape-scale features provided the best predictive model, findings that support, in part, small-scale reef research (Walsh 1985). Thus, scale has profound effects on resultant patterns (Wiens 1989) with fine-scale measures often better predictors of one group of organisms, and landscape measures predictors of others (Mitchell et al. 2001, Mazerolle and Villard 1999). This organism-based perspective appears to be true for coral reef fishes (Pittman et al. 2004), consequently future studies should acknowledge that species perceive the landscape in different ways. The relevant scale of investigation may depend on life history attributes of individual fish species (Hovel and Lipcius 2003, Pittman et al. 2004), or biological processes such as foraging behavior (Shulman and Ogden 1987), and predation (Hixon and Beets 1989). While these results, using a single fine-scale measure of rugosity, suggest that landscape-scale measures are more valuable in predicting most reef fish parameters, fishes appear to respond to benthic structure at multiple spatial scales, with each species responding to a unique suite of variables. Future studies will require an organism-based perspective that explores multiple spatial scales.

PAGE 46

29 These results should be interpreted within the scope and limitations of this purely correlative study, and although a range in the values of different metrics was represented, I had little control on experimental units. These reefs are natural habitat patches, therefore considerable microhabitat variation exists, which is neither measure nor controlled. Reefs also do not represent a perfect gradient in landscape scale habitat features since sample units were selected from the naturally available set of reefs; rather they vary across multiple gradients. Additionally, while considerable groundtruthing was conducted, the benthic habitat maps were accepted without major modification. Since each decision in the mapping process effects the determination and analyses of spatial structure, it also effects our results of relationships between this structure and ecological pattern. Finally, the reef fish populations of the Virgin Islands have been heavily exploited (Beets and Rogers 2001), therefore future studies should examine reef fish distributions in less fished areas. Conclusions This exploratory analysis allowed me to investigate relationships between landscape structure and reef fish assemblage structure at 20 reefs to develop hypotheses and guide future coral reef landscape studies. Landscape-scale metrics proved valuable in characterizing and quantifying the landscape structure of the coral reef environment (e.g., size, shape and context). Principal components of these metrics, however, were correlated with few reef fish assemblage parameters. Interpretation of the few reef fishprincipal component relationships led to the conclusion that individual habitat features are better measures of the influence of the spatial patterning of the coral reef landscape to reef fish species diversity and abundance. Specifically, reef context was associated with reef fish diversity and abundance for many groups of reef fishes. Because results

PAGE 47

30 revealed that species responded to different scales depending on their life history attributes and habitat requirements, future studies will examine specific functional relationships (e.g., seagrass and grunts) based on the ecological requirements of the particular fish groups of interest. If the results detected in this exploratory study are replicable across systems and scales, combining the disciplines of landscape ecology and reef fish ecology offers promise in addressing important management questions relevant to habitat-based conservation of reef fishes.

PAGE 48

31 Table 2-1. Reef fish assemblage parameters (n = 30) used as dependent variables in statistical analyses Entire assemblage level parameters Trophic guilds Mobility guilds Taxonomic groupings Cumulative species richness Herbivores (J & A) Resident Acanthurids (J & A) Mean species richness Mobile invertebrate feeders (J & A) Mobile Serranids (J & A) Total abundance Omnivores (J & A) Transient Haemulids (J & A) Piscivores (J & A) Lutjanids (J & A) Planktivores Scarids (J & A) Sessile invertebrate feeders Chaetodontids Holocentrids Labrids Pomacanthids Note: Fish groups are not always mutually exclusive. Hypoplectrus species were not included in Serranid grouping. For each trophic guild and taxonomic grouping, reef fish parameters were further subdivided into juvenile and adult components, where indicated (J = juvenile, A = adult).

PAGE 49

32 Table 2-2. Fourteen metrics used to quantify the landscape structure of the 20 study reefs sampled in 1994 and 2001 in St. John, USVI. Patch metric Definition Formula Patch Size (reef)p Size of individual habitat patches. Area (m2) Polygon Size (reef)p Size of individual patch Area (m2) P:A of a Patchp Sum of the patch edge divided by patch area for patch of interest. P : A for particular patches Habitat Richnessl Number of different habitat types present in an extent. Number of different habitat types Patch Richnessl Number of patches of each habitat type in extent of interest. Number of patches Habitat Area bedrockl deepl algal plainl pavementl reefl sandl seagrassl Amount of each habitat type in landscape. Area (m2) in each habitat type Patch Diversityl Total abundance and type of different patches (pi is the proportion of habitat for every individual patch) pi ln pi; pi = proportion of area in (m2) of patch i for all patches Habitat Diversityl Same as patch diversity but boundaries of similar habitat patches (by habitat class) are dissolved so that number of patches does not influence index. p i for habitat diversity is proportion of habitat for all patches. pi ln pi pi= proportion of area in (m2) of habitat type i for all habitat types Note: For each patch metricp, a single value was calculated for every reef. Each landscape metricl was calculated at three spatial extents (100 m, 250 m, and 500 m) resulting in a total of 36 metrics for each study reef.

PAGE 50

33 Table 2-3. Summary statistics on reef configuration, context and rugosity and for select reef fish assemblage parameters (entire assemblage level, trophic level and mobility guilds) for 20 study reefs sampled in 1994 and 2001, St. John, USVI for metrics at the 100 m spatial extent with coefficient of variation for landscape parameters and standard error for reef fish parameters. Habitat parameter Measure Transform Min Max Mean CV Total area Ha None 6.53 18.3 13.02 24.89 H Patch diversity Index None 1.13 2.45 1.63 20.54 H diversity Index None 0.54 1.57 1.17 24.32 Habitat richness # habitat types None 2.00 6.00 4.15 27.39 Patch richness # patches Log10 (x +1) 5.00 19.00 8.50 43.09 Size of reef Ha None 0.54 15.74 6.59 67.62 P:A reef ratio None 0.03 0.09 0.05 34.33 Reef Ha Log10 (x +1) 0.83 17.89 4.18 91.42 Seagrass Ha Log10 (x +1) 0.00 7.80 1.32 166.26 Bedrock Ha Log10 (x +1) 0.00 3.56 0.80 142.92 Pavement Ha Log10 (x +1) 0.00 11.33 2.72 105.72 Deep water Ha Log10 (x+1) 0.00 7.30 2.18 120.10 Algal plain Ha Log10 (x +1) 0.00 4.02 0.58 222.76 Sand Ha Log10 (x +1) 0.00 6.03 1.17 160.39 Rugosity Index None 1.38 2.82 2.00 19.48 Reef fish Parameter Units Transform Min Max Mean SE Mean spp richness Number Log10(x + 1) 19.67 32.14 23.43 0.67 Cum spp richness Number Log10(x + 1) 51.00 88.00 67.85 2.25 Total Abundance Number Log10(x + 1) 24.12 88.13 55.23 0.02 A Herbivore Number Log10(x + 1) 4.89 25.92 10.22 0.04 J Herbivore Number Log10(x + 1) 14.14 73.13 41.66 0.04 A MIF Number Log10(x + 1) 1.04 18.95 4.37 0.05 J MIF Number Log10(x + 1) 7.71 32.11 16.38 0.03 A PISCI Number Log10(x + 1) 0.05 1.75 0.45 0.02 J PISCI Number Log10(x + 1) 0.58 3.27 1.63 0.03 PLANK Nu mber Log10(x + 1) 3.17 9 50.21 36.15 0.08 A OMNI Number Log10(x + 1) 0.00 1.19 0.26 0.02 J OMNI Number Log10(x + 1) 1.57 34.48 8.55 0.08 SIF Nu mber Log10x + 1) 1.69 34.48 9.23 0.03 Resident A Number Log10(x + 1) 32.88 103.7 65.07 0.05 Mobile A Number Log10x + 1) 0.48 11.59 1.95 0.03 Transient A Number Log10(x + 1) 24.12 88.13 55.23 0.04 Note: All parameters for mean abundance, except where indicated. All data are backtransformed. CV = coefficient of variation, SE = standard error.

PAGE 51

34 Table 2-4. Pearson product moment correlation matrix of the 14 landscape-scale habitat variables, with the resultant 9 remaining significant variables, at the 100 m spatial extent for the 20 reef sites sampled in 1994 and 2001 in St. John, USVI. Note: Values in bold represent those with significant pair-wise correlations. Those in shadow represent variables that were excluded based on significant pair-wise correlations. These 14 variables were thereby reduced to the remaining 8 variables (those not in shadow), which were used in subsequent principal component analyses. P:A reef # Patch Habitat richness Reef size Focal reef Patch diversity Habitat diversity Bed rock Deep Algal plain Pave-ment Reef Sand Seagrass P:A reef 1.0 -0.03 0.03 -0.47 -0.76 -0.23 -0.03 -0.31 0.50 -0.02 0.17 -0.44 -0.32 -0.24 # patches -0.03 1.00 0.44 0.43 0.33 0.84 0.29 -0.08 -0.05 0.02 0.20 0.36 0.09 -0.04 Habitat richness 0.03 0.44 1.00 0.41 0.23 0.63 0.87 0.45 0.17 0.41 -0.30 -0.19 0.25 0.04 Reef size -0.47 0.43 0.41 1.00 0.40 0.44 0.43 0.29 -0.15 0.36 0.003 0.24 0.25 -0.28 Focal reef -0.76 0.33 0.23 0.40 1.00 0.46 0.09 0.40 -0.55 -0.01 -0.24 0.46 0.23 0.38 Patch diversity -0.23 0.84 0.63 0.44 0.46 1.00 0.58 0.17 -0.10 0.17 -0.05 0.31 0.25 0.06 Habitat diversity -0.03 0.29 0.87 0.43 0.09 0.58 1.00 0.46 0.22 0.45 -0.16 -0.14 0.30 -0.09 Bedrock -0.31 -0.08 0.45 0.29 0.40 0.17 0.46 1.00 -0.20 0.21 -0.76 0.05 0.25 0.26 Deep 0.50 -0.05 0.17 -0.15 -0.55 -0.10 0.22 -0.20 1.00 0.30 0.12 -0.23 -0.43 -0.41 Algal plain -0.02 0.02 0.41 0.36 -0.01 0.17 0.45 0.21 0.30 1.00 -0.08 -0.02 -0.05 -0.27 Pavement 0.17 0.20 -0.30 0.003 -0.24 -0.05 -0.16 -0.76 0.12 -0.08 1.00 0.30 -0.23 -0.44 Reef -0.44 0.36 -0.20 0.24 0.46 0.31 -0.14 0.05 -0.23 -0.02 0.30 1.00 -0.12 -0.20 Sand -0.32 0.09 0.25 0.25 0.23 0.25 0.30 0.25 -0.43 -0.05 -0.23 -0.12 1.00 -0.14 Seagrass -0.24 -0.04 0.04 -0.28 0.38 0.06 -0.09 0.26 -0.41 -0.27 -0.44 -0.20 -0.14 1.00

PAGE 52

35 Table 2-5. Principal component analyses on the correlation matrix of the 8 residual landscape-scale habitat variables at the 100 m spatial extent for the 20 study reefs sampled in 1994 and 2001 in St. John, USVI. PC1 PC2 PC3 PC4 Eigenvalue 2.10 1.82 1.49 1.01 Percent 26.33 22.76 18.65 12.61 Cum Percent 26.33 49.09 67.74 80.35 # Patches 0.488 0.013 -0.149 0.470 Reef Size 0.549 0.173 0.033 0.049 Habitat Diversity 0.359 0.258 0.493 0.243 Deep -0.029 -0.422 0.607 0.216 Pavement 0.211 -0.537 -0.163 -0.192 Reef 0.333 -0.227 -0.516 0.084 Sand 0.234 0.490 0.043 -0.570 Seagrass -0.347 0.381 -0.265 0.550 Note: Loadings in bold represent the top thr ee variables that cont ribute the most to individual components.

PAGE 53

36 Table 2-6. Stepwise regression results to determine the influence of principal components on reef fish assemblage structure at the 20 study reefs sampled in 1994 and 2001 in St. John, USVI at the 100 m spatial extent Fish Model PC1 PC2 PC3 PC4 parameter R2 b1 R2 p b2 R2 p b3 R2 p b4 R2 p Mean richness 21% 1.38 0.21 0.040 Cumulative richness 26% 5.17 0.27 0.020 Herbivores 53% 0.05 0.22 0.01 -0.06 0.31 0.004 Omnivores 21% 0.14 0.21 0.044 Haemulids 43% -0.14 0.43 0.002 Epinephelids 23% -0.05 0.23 0.030 Acanthurids 35% 0.09 0.18 0.05 -0.10 0.17 0.040 Lutjanids 46% -0.07 0.32 0.006 0.05 0.14 0.050 Mobile 28% 0.05 0.28 0.02 Note: Each of the 30 reef fish parameters were used as dependent variables. Linear models: log abundance = b0 + b1 (PC). Data represented are mean and cumulative species richness values and mean abundances within each guild derived from a minimum of 16 samples per reef. See table 4 for definitions of individual principal components. The suite of 30 fish parameters were analyzed, however, only those reef fish parameters with statistically significant relationships are reported. P-values are Sequential Dunn-Sidak Bonferroni-corrected for the total number of comparisons (n = 30). Only those relationships with p < 0.05 are presented.

PAGE 54

37 Table 2-7. Stepwise multiple regression results of the influence of reef configuration on reef fish assemblage structure for the 1994 (N = 14) study reefs in St. John, USVI. Reef fish parameter Habitat parameter Model R 2 Partial R2 p-value A Piscivores P:A reef (-) 0.32 0.32 0.0500 A Omnivores P:A reef 0.40 0.40 0.0100 J Haemulids # patches (-) 0.39 0.39 0.0200 A Acanthurids P:A reef 0.33 0.33 0.0300 Pomacanthids P:A reef 0.37 0.37 0.0200 J Lutjanids # patches (-) 0.36 0.36 0.0300 Transients P:A # patches 0.74 0.62 0.12 0.0002 0.0500 Note: Independent variables were: P:A of each reef, the number of patches and reef size. Results for 1994 reefs (n=14), with model R2 and partial regression values for each variable with p < 0.05 level. P-values are Sequential Dunn-Sidak Bonferroni-corrected for the total number of comparisons. Model effects are all positive, except where indicated (-). The suite of 30 fish parameters were analyzed, however, only those reef fish parameters with statistically significant relationships are reported. For all other reef fish parameters, there are no significant relationships. A = adult, J = juvenile.

PAGE 55

38 Table 2-8. Stepwise multiple regression results of the influence of reef context on reef fish assemblage structure for the 1994 (N = 14) study reefs in St. John, USVI Fish parameter R2 Partial R2 p-value Habitat parameter Mean species richness 0.28 0.28 0.04 Seagrass J Herbivores 0.30 0.30 0.04 Bedrock A MIFs 0.33 0.33 0.03 Seagrass J Omnivores 0.48 0.48 <0.01 Deepwater A Piscivores 0.51 0.51 <0.01 Reef J Piscivores 0.47 0.24 0.23 0.03 0.05 Deepwater Seagrass SIFs 0.63 0.39 0.27 <0.01 <0.01 Bedrock (-) Deepwater A Haemulids 0.53 0.53 <0.01 Seagrass A Epinephelids 0.52 0.28 0.24 <0.01 0.04 Reef Seagrass J Acanthurids 0.67 0.42 0.25 <0.01 0.01 Deepwater (-) H (-) A Lutjanids 0.68 0.50 <0.01 Seagrass J Lutjanids 0.46 0.46 <0.01 Bedrock Mobile 0.39 0.39 0.02 H Note: Independent variables were: H and areal extent of reef, bedrock, seagrass and deep unknown within 100 m. Model R2 and partial regression values for each variable with p <0.05 level. P-values are Sequential Dunn-Sidak Bonferroni-corrected for the total number of comparisons. The suite of 30 fish parameters were analyzed, however, only those reef fish parameters with statistically significant relationships are reported. Model effects are positive except where indicated by (-). A = adult, J = juvenile.

PAGE 56

39 Table 2-9. Stepwise multiple regression results of the relative influence of landscape and fine-scale habitat measures on reef fish assemblage structure on the 1994 (N = 14) study reefs on St. John, USVI. Fish parameter R2 Partial R2 p-value Habitat parameter Mean species richness 0.68 0.38 0.30 0.0040 0.0008 Seagrass Rugosity A MIFs 0.53 0.33 0.20 0.0100 0.0500 Seagrass Rugosity J Omnivores 0.76 0.60 0.16 0.0079 0.0200 Rugosity Reef A Piscivores 0.53 0.53 0.0090 Reef J Piscivores 0.71 0.36 0.35 0.0030 0.0070 Rugosity Seagrass A Haemulids 0.71 0.55 0.16 0.0006 0.0300 Seagrass Rugosity A Epinephelids 0.54 0.38 0.16 0.0090 0.0400 Reef Seagrass J Acanthurids 0.46 0.46 0.0100 Deepwater A Lutjanids 0.44 0.44 0.0050 Seagrass Note: Independent variables are rugosity and the areal coverage of deep unknown, seagrass and reef within 100 m. Results for the 14 reefs sampled in 1994, with model R2 and partial regression values for each variable significant at the p = 0.05 level. P-values are Sequential Dunn-Sidak Bonferroni-corrected for the total number of comparisons. Relationships in bold are those that rugosity contributed to explanatory power. The 30 fish parameters were analyzed. Only fish parameters with significant relationships are reported. A= adult, J = juvenile.

PAGE 57

40 Figure 2-1. Location of St. John, US Virgin Islands in the Caribbean basin.

PAGE 58

41 Figure 2-2. Distribution of the 20 study reefs around the island of St. John, USVI. Below the name of each reef is the number of fish point counts per reef.

PAGE 59

42 -4-3-2-1012 -3.0-2.5-2.0-1.5-1.0-0.500.51.01.52.02.53.0 -3.0-2.5-2.0-1.5-1.0-0.500.51.01.52.02.5 -2.5-2.0-1.5-1.0-0.500.51.01.52.02.5 -3.0-2.5-2.0-1.5-1.0-0.500.51.01.52.02.53.0 -3.0-2.5-2.0-1.5-1.0-0.500.51.01.52.02.5 Figure 2-3. PCA plots of the landscape structure of the coral reef environments of the 20 study reefs sampled in 1994 and 2001 in St. John, USVI at the A) 100 m, B) 250 m and C) 500 m spatial extent. The x-axis is PC 1 and y-axis is PC 2. Symbols refer to different types of reefs based on pre-existing knowledge of the coral reef landscapes, irrespective of the benthic habitat maps. Triangles refer to shallow reefs with surrounding pavement habitat, squares refer to reefs with seagrass nearby, and upside down triangles refer to isolated patches of reef with large areas of deep water nearby. A B C

PAGE 60

43 R2 = 0.62p < 0.000200.40.81.2 R2 = 0.40p < 0.0100.10.20.3 0.4 R2 = 0.28p < 0.0500.10.20.30.40.50.020.040.060.080.1P:A reef Figure 2-4. Effects of reef configuration on mean fish abundances of A) transient fishes, B) adult omnivores, and C) adult piscivores for the 1994 (N =14) study reefs in St. John, USVI. A B C

PAGE 61

44 R2 = 0.28p=0.04420222426283032340.00.20.40.60.81.0Seagrass R2 = 0.33p=0.030.00.20.40.60.81.01.21.40.00.20.40.60.81.0Seagrass R2 = 0.53p=0.0030.80.40.60.81.0 00.10.20.30.40.50.60.790.00.2Seagrass 0. R2 = 0.51p=0.000900.10.20.30.40.50.20.40.60.81.01.2 1.4Reef R2 = 0.46p=0.000600.10.20.30.40.500.20.40.60.8Bedrock R2 = 0.48p=0.00440.00.20.40.60.81.01.21.41.61.80.00.20.40.60.8Deep unknown Figure 2-5. Effects of reef context on mean abundance of particular fish groups for the 1994 (N =14) study reefs in St. John, USVI. A) Mean species richness and areal extent of seagrass 100 m, B) mean abundance of MIFs and areal extent of seagrass 100 m, C) mean abundance of haemulids and areal extent of seagrass 100 m, D) mean abundance of piscivores and areal extent of reef 100 m, E) mean abundance of juvenile lutjanids and areal extent of bedrock 100 m, and F) mean abundance of juvenile omnivores and areal extent of deep unknown 100m. Independent variables are each log10 X +1 in hectares. MIFs refer to mobile invertebrate feeders. A = adult, J = juvenile. B D F A E C

PAGE 62

CHAPTER 3 EVIDENCE OF FUNCTIONAL CONNECTIVITY IN A CORAL REEF ECOSYSTEM Coral reef ecosystems are deteriorating worldwide, with symptoms including loss of hard corals, declines in abundances of exploited reef fishes and reduced biological diversity. Marine protected areas (MPAs) represent an important management tool for reducing this degradation; however, their effectiveness is contingent on our understanding of key ecological patterns and processes at appropriate spatial scales. MPA effectiveness may also be dependent upon maintaining critical linkages between essential habitat patches (e.g., seagrass and reef). Exploratory analyses of the relationship of reef fish assemblage structure with coral reef landscape structure in the U.S. Virgin Islandsone of the first studies that applied a landscape ecology approach to coral reef ecosystemsprovided the foundation for developing specific hypotheses (ie.e reef context influences reef fish assemblage structure at the scale of individual reefs). These hypotheses were then tested in this new study at 22 independent reefs. As expected, reef context influenced the structure of reef fish assemblages, and specific relationships were functionally consistent with the ecology of the fishes of interest. Consistent with predictions, reefs with neighboring seagrass had the highest total fish abundance, and highest abundances of fishes within the mobile invertebrate feeding guild, and within the exploited families of haemulidae (grunts) and lutjanidae (snappers). Species richness for the entire fish community and within particular fish groups were also strongly associated with the areal coverage of seagrass neighboring study reefs, suggesting the importance of habitat linkages to a diversity of species. Potential habitat 45

PAGE 63

46 linkages were detected as far away as 1 kilometer, which may indicate that reef fishes perceive the landscape at this spatial scale. These findings infer that functional habitat connectivity/juxtaposition between essential habitat patches is important in structuring reef fish assemblages, and further suggests that landscape measures of this habitat connectivity may be useful to managers in the design of MPAs. Introduction A landscape generally refers to a heterogeneous area composed of local interacting ecosystems (Forman 1995) made up of homogenous units, called habitat patches, and the sizes and spatial arrangement of these patches can exert a strong influence on the diversity, abundance, distribution, and movement patterns of organisms (Wiens 1989). Movements and flows of energy, nutrients or organisms between ecosystems can be either be facilitated or inhibited by the spatial arrangement of these habitat patches (Forman 1995, Turner et al. 2001). The degree to which this exchange occurs can depend upon the boundary (the zone composed of the edges of adjacent ecosystems), context (adjacency, neighborhood, and location within a landscape), or connectivity (how connected or spatially continuous a corridor or matrix is) of habitat patches within the landscape mosaic (Forman 1995). As we strive to manage entire ecosystems, maintaining functional habitat linkages within this landscape mosaic may be crucial to the effectiveness of protected areas (Noss 1983; Forman 1995, Robinson et al. 1995, Turner et al. 2001, Lindenmayer et al. 2002), since a failure to consider spatial elements such as edges, movement corridors, and landscape context in the design of protected areas is likely to result in undesirable changes in community characteristics and possibly the loss of key species (Noss 1983, McGarigal & McComb 1995).

PAGE 64

47 Though these principles were largely derived in terrestrial systems, they likely apply to tropical marine systems (Grober-Dunsmore et al. 2004, Kendall et al. 2003, 2004, Pittman et al. 2004, Jeffrey 2004), and may be essential for designing functional marine protected areas (MPAs) in coral reef landscapes, particularly as coral reef landscapes exist as a complex mosaic of interacting habitat patches (i.e., seagrass, reef, and mangrove). Maintaining important habitat linkages among habitat patches should be considered in the design of MPAs, which are increasingly being considered as a primary ecosystem-based tool for improving fisheries management and protecting biodiversity (Carr & Reed 1993, Allison et al. 1998). Unfortunately, there is little information on how to best design MPAs (Ballentine 1997, Sale 2005), particularly for the conservation of the complex of fishes that constitute the reef fish community. Historically, coral reef fishes have been difficult to manage, in part, because different species often have different habitat requirements (Sale 2002). Moreover, these habitat requirements frequently change with ontogeny (Lindeman et al. 2000, Appeldoorn et al. 2003). Though generally considered site-attached following settlement (Sale 1991), many regularly move hundreds and even thousands of meters from their primary resting and foraging locations (Plan Development Team 1990, Corless et al. 1997). These movement patterns are often species and life history stage-specific and occur across multiple spatial scales (Randall 1962, Lindeman et al. 2000, Appeldoorn et al. 2003). Some migrate daily to forage (Stark & Davis 1966, Ogden & Zieman 1977, Helfman et al. 1983, Holland et al. 1993, Tulevech & Recksiek 1994), others migrate annually to spawning aggregations (Colin et al. 1997) and several exhibit multiple ontogenetic shifts in habitat (Randall 1962, Appeldoorn et al. 1997). To date, the overwhelming majority of

PAGE 65

48 reef fish studies have been conducted at small spatial scales (e.g., 1 m plots) (see Sale 2002), limiting our ability to understand important habitat linkages across larger spatial scales. Recent landscape-scale (hundreds of meters to kilometers) research however, provides correlative evidence that cross-shelf location (Christensen et al. 2003), reef context (Appeldoorn et al. 2003, Kendall et al. 2003, 2004, Grober-Dunsmore et al. 2004a, 2004b, Pittman and McAlpine 2003, Mumby et al. 2004, Pittman et al. 2004) and landscape connectivity (Ault & Johnson 1998, Jeffrey 2004) influences reef fish community structure. This study builds on previous research by assessing the generality of my correlative models from the U.S. Virgin Islands (Grober-Dunsmore et al. 2004b) at an independent set of smaller candidate reefs (<1 hectare) and explicitly testing hypotheses of the importance of functional habitat linkages between two habitat types, reef and seagrass, in structuring reef fish communities. The following predictions were made (1) each reef fish parameter of interest (i.e. entire assemblage level parameters and abundances and species richness within mobile invertebrate feeders (MIFs), grunts, (Haemulidae), snappers (Lutjanidae), and groupers (Epinephelines) will be greater at reefs proximal to seagrass, (2) abundance and species richness of mobile fishes will be greater at reefs proximal to seagrass, but resident and transient guilds will not, (3) relationships of reef fish parameters and the areal coverage of seagrass will be similar in nature and strength between years and (4) the fine-scale measure of rugosity and landscape-scale measures of seagrass will both be included in the best predictive models of fish assemblage structure. These particular fish families were selected because they are important components of the subsistence fishery in the Caribbean region, and because

PAGE 66

49 they rely on a variety of habitats (i.e. seagrass) for foraging and settlement (Lindeman et al. 2000). Consequently, they are more likely to demonstrate effects of habitat linkages between reef and seagrass. Study Area This study was conducted in the shallow waters around the island of St. John, USVI, located in the Northern Antilles approximately 88 km east of Puerto Rico (Figure 2-1). St. John is part of the Puerto Rico Bank, a submerged plateau defined by the 183 m depth contour extending from eastern Puerto Rico to the island of Anegada. Study reefs (n = 22) were selected around St. John (Figure 3-1) to maximize the variation of the particular landscape parameter of interest (i.e. the areal coverage of seagrass habitat), while reducing variation due to within-reef characteristics (e.g., coral cover). Each study reef was ~ 1-2 hectares in total area, within 1 km of the shoreline and occurred at depths of 3 to 10 m. All reefs were dominated by Montastraea annularis with Agaricia agaricites, Porites porites, P. astreaoides, Siderastraea siderea and S. radians contributing to total living coral cover, estimated at approximately 5-15 % for each reef. The background matrices of study reef patches were typically hard-bottom or sand with large colonies of M. annularis providing the major structural components. In general, the neighboring seagrass communities were quite similar (e.g., typically dominated by Thalassia testudinum, in shallow areas with contributions from Syringodium filiforme and varying densities of rhizophytic algae).

PAGE 67

50 Methods Reef fish Sampling Reef fishes were censused using a standardized visual point count census method (Bohnsack & Bannerot 1986), where all reef fishes were identified within a 5 minute sampling period, and enumerated during the following 10 minute period within a sampling radius of 7.5 m2. Total lengths of fishes were estimated to the nearest centimeter. Sampling effort was standardized to reef size (one census per 1500 m2) following Monte Carlo simulation in Grober-Dunsmore et al. (2004a). A single observer (RGD) conducted all reef fish censuses to eliminate observer variability. A suite of fish parameters was estimated from the visual point count data (see Table 3-1 and Grober-Dunsmore et al. 2004 for more details). Randall (1967) and Froese and Pauly (2002) were used as references to classify all fishes by trophic guild: piscivore, herbivore, mobile invertebrate feeder (MIF), sessile invertebrate feeder (SIF), planktivore or omnivore. A seagrass-associated category of fishes was also created using Lindeman et al. (1998), Lindeman et al. (2000), Appeldoorn et al. (2003) and personal observations (for Epinephelus striatus and Cephalopholis cruentatus), and includes these grunt, grouper and snapper with explicit associations with seagrass: Lutjanus apodus, L. analis, L. griseus, L. jocu L. mahogoni, Ocyurus chrysurus, Haemulon plumieri, H. sciurus, H. flavolineatum, E. striatus and C. cruentatus. Commercially-important families of grunts, snappers, and groupers were examined separately. Given the known natural history, each species was also into mobility guilds. Resident species are site-attached, and do not typically move from their primary reef patch. Mobile species have restricted movements among adjacent habitat patches, and may roam from the primary reef patch during foraging. Transient species are vagile, ranging on the scale of kilometers. Fishes were

PAGE 68

51 also subdivided into juvenile and adult categories, based on average size at maturity (Froese & Pauly 2002), where possible. Using minimum estimated length, a fish was placed in the juvenile category if it was below the average estimated size at maturity, and placed in the adult category if it was at or above this estimate. Mean abundance values based on replicate censuses at each reef were calculated and used in subsequent analyses for all groups. Temporal Sampling In August 2003, reef fishes at a subset (n = 8) of the original 22 (2002) reefs were recensused (Figure 3-1, reefs in bold) to assess the temporally consistency of detected relationships. These reefs were selected as they represented a near maximal gradient in seagrass areal coverage, and had similar benthic community structure and depth, reducing effects of among-reef variability. Sampling effort was increased to improve the accuracy of the estimate of the mean value for each reef fish parameter, thus a minimum of 10 censuses per reef were conducted, regardless of reef size. Habitat Sampling Landscape metrics of reef context were calculated with ArcView 3.2 software (ESRI 1996), using digital benthic habitat maps. Maps were created from aerial photographs flown at an altitude of 5000 feet in 1999 and were classified by visual interpretation, using 26 discrete and non-overlapping habitat classes, with a minimum mapping unit of 1 acre (Kendall et al. 2001). The original map classification scheme was condensed from 26 to 7 distinct, non-overlapping habitat classes (pavement, sand, reef, bedrock, seagrass, macroalgae and deep unknown), so results would be more broadly applicable to resource managers and to reduce potential classification errors. Reefs served as focal units for all analyses because MPAs are frequently designed around

PAGE 69

52 individual reefs (e.g., Florida Keys National Marine Sanctuary). Landscape metrics were calculated at the 100 m, 250 m, 500 m and 1 km spatial extent, to explore the appropriate scale for each response variable. Rugosity was measured in situ along ten 10 m transects at each reef fish census location, using methods described by Luckhurst and Luckhurst (1978). A mean was calculated for each reef. Statistical Analyses All data were checked for normality using Shapiro-Wilks tests (p < 0.01) (Zar 1984), and transformed where appropriate (Table 3-1). If assumptions were met, analyses were conducted using parametric statistics, and non-parametric statistics if not. For all analyses, each reef fish parameter was used as the dependent variable and the landscape and/or rugosity measure was used as the independent variable. Residual plots were examined to assess stability of regression models (Sokal & Rohlf 1995). Because predictions were derived at different study reefs, analyses were conducted to verify the importance of reef context at these independent reefs using stepwise multiple regression methods as in Grober-Dunsmore et al. (2004b) for all reef fish parameters. Model II linear regression analysis (Sokal & Rohlf 1995) was used to test the prediction: 1) abundances of MIFs, grunts, snappers, groupers, mobile reef fishes, seagrass-associated taxa and species richness values within these fish groups are higher at reefs proximal to seagrass. Model II linear regression was also used to test the prediction: 2) mobile taxa will be influenced by reef context (measured by the areal coverage of seagrass), whereas resident and transient taxa will not. Non-parametric Spearman Rho correlations (Sokal and Rohlf 1995) were calculated for the subset of 8 reefs (2003), sampled in 2002 and 2003 to test the prediction: 3) relationships between reef fish parameters and reef context will be similar in nature (e.g., direction) and

PAGE 70

53 strength between years. Statistical comparisons of the abundances of reef fishes, using raw census data (n = 130), were conducted using the non-parametric Kruskal-Wallis to test for differences in abundance between years. Stepwise multiple regression analyses was used to test the prediction: 4) reef fishes will be structured by both landscape-scale and fine-scale measures of habitat. Results A total of 107 fish censuses were conducted at 22 reefs during July and August 2002. One hundred eighteen species were identified and a total of 14,389 individuals recorded. In August 2003, 97 reef fish censuses were conducted at the subset of 8 reefs, with one hundred twenty two species identified and a total of 14,239 individuals recorded. The ten most abundant taxa in each fish group are listed in Table 3-1. The configuration and composition of the study landscapes varied widely among reefs, with coefficients of variation for most landscape metrics > 50 % of the mean (Table 3-2). Gradients in the areal coverage of seagrass were represented adequately to test hypotheses concerning reefseagrass habitat linkages (Table 3-2). Rugosity also varied, although to a small degree than seagrass (Table 3-2). Exploratory stepwise regression analyses confirmed that reef context was the best predictor of reef fish assemblage structure, allowing me to test specific hypotheses concerning the importance of seagrass in particular. For all subsequent analyses, I verified that reefs were not consistently higher for other reef fish parameters to refute the possibility that some unmeasured factor did not result in high values for all fish parameters.

PAGE 71

54 Entire Assemblage Level Parameters The areal coverage of seagrass was a positive predictor of all entire assemblage level parameters for every spatial extent (Table 3-3), with 48-58 % of the variation in cumulative species richness explained by seagrass (Table 3-3, Fig. 3-2). Higher cumulative richness at reefs with seagrass nearby can be partially attributed to the presence of several species with specific dependencies on seagrass (e.g., Aetobatus narinari, Xyrichtys splendens, Sparisoma radians, Holocentrus rufus). Reefs surrounded by seagrass had the highest total number of species. Abundances within Reef Fish Groups The areal coverage of seagrass was a positive predictor of abundances of grunts (R2 = 0.52-0.57), seagrass-associated taxa (R2 = 0.33-0.50), snappers (R2 = 0.34-0.40), and MIFs (R2 = 0.27-0.41), though groupers were marginally associated (R2 = 0.13-0.17) with seagrass (Table 3-4, Fig. 3-3). Reefs with large expanses of adjacent seagrass had the highest abundances of MIFs, grunts and seagrass-associated taxa were Donkey, Hansens and Marys. Abundances within Life History Stages Juvenile grunts were strongly (R2 = 0.52-0.56), whereas adults were weakly associated with the areal coverage of seagrass (R2 =0.16-0.21) (Table 3-5), and the juvenile component of the seagrass-associated taxa also exhibited a stronger relationship than the adult (Table 3-5). The only reefs without juvenile yellow-tail snapper (Ocyurus chrysurus) were those without seagrass within 1 km (e.g., Peter Bay, PeterOne), and juvenile schoolmaster (Lutjanus apodus) and mahogony snapper (L. mahogoni) were only present at reefs with extensive shallow back-reef seagrass (e.g., Saba E, Marys, Rendezvous).

PAGE 72

55 Species Richness The areal coverage of seagrass was also a positive predictor of species richness within MIFs (R2 = 0.63-0.72), grunts (R2 = 0.56-0.71), snappers (R2 = 0.24-0.41) and groupers (R2 = 0.23-0.33) (Table 3-6, Figure 3-4). Increased species richness values were frequently attributed to the presence of those particular species associated with seagrass. For example, species richness of MIFs was increased by the presence of species that reside or forage in seagrass (e.g., Aetobatus narinari, Holocentrus marianus, and Calamus pennatula), and species richness of snapper was increased by the presence of small schools of juvenile schoolmaster, grey snapper and lane snapper. Grunt species richness was enhanced by the presence of H. plumieri, H. carbonarium H. chysargyreum, and H. macrostomum; while H. plumieri is a known seagrass-specialist, seagrass dependencies of the other taxa are not understood. Residual analysis suggested that caution is warranted in interpreting the relationship for snappers and groupers. Mobility The areal coverage of seagrass was a positive predictor of abundances (R2 = 0.19-0.21) and species richness (R2 = 0.55-0.69) of mobile taxa, but not for resident (except at the 250 m extent) and transient species (Table 3-4, Table 3-6). Abundances of transient taxa violated normality tests, and may have contributed to my inability to detect relationships. Species richness was enhanced at reefs with neighboring seagrass by the presence of more mobile species such as: Acanthurus chirugus, Balistes vetula, Calamus calamus, Gerres cinereus and Holocentrus rufus, species that likely to migrate into neighboring foraging patches (Kramer & Chapman 1999).

PAGE 73

56 Spatial Extent Although the R2 and p-values were comparable across all spatial extents for most reef fish parameters and/or groups, the strength of relationships for the adult life history stages was often greatest at the 500 m spatial extent, and generally occurred at the 100 m or 250 m spatial extent for the juvenile life history stages (Table 3-5). For species richness, the strength of associations was strongest at the 100 m (grunts, MIFs) and 250 m (snappers, groupers, and mobile taxa) spatial extents (Table 36). Temporal Consistency Twelve of sixteen reef fishhabitat associations were similar between years in 2002 and 2003 (Table 3-7, Figure 3-5). Of these twelve, seven relationships (cumulative richness, mean richness, total abundance, juvenile grunt abundance, juvenile grouper abundance, species richness of MIFs, and species richness of groupers) were positively correlated to the areal coverage of seagrass in 202 and 2003 at the 8 reefs (Figure 3-5). Five were consistent in that they were not correlated with seagrass in either year (adult and juvenile MIF abundance, adult and juvenile snapper abundance, and species richness of snappers). Several relationships from the 2002 dataset of 22 reefs were not evident when analysis was constrained to the subset of 8 reefs. The reduction in sample size from 22 to 8 clearly resulted in a loss of statistical power (Sokal and Rohlf 1995), and abundances of three fish groups (MIFs, adult grunts, and adult snappers) were statistically lower in 2002 compared to 2003 (Table 3-7). Relative Influence of Fine and Landscape-scale Measures The fine-scale measure of rugosity was of limited value in predicting reef fish assemblage structure; rather, landscape-scale measures of seagrass areal coverage were better predictors of most fish assemblage parameters, though there were a few exceptions

PAGE 74

57 (Table 3-8). Collinearity diagnostics suggested that although weak relationships exist between the independent variables, all model condition indices were less than 10, well below the recommended value of 30 (Belsley et al. 1980). Discussion As predicted, landscape-scale measures of the areal coverage of seagrass calculated with simple GIS tools and habitat maps were useful surrogates for reefs with high entire assemblage level parameters (e.g., cumulative species richness). These findings are relevant because earlier marine ecologists alluded to the importance of reef context (Ogden & Zieman 1977, Parrish 1989) and high fish diversity has been attributed to the presence of seagrass in the Florida Keys (Robblee & Zeiman 1984), the Virgin Islands (Quinn & Ogden 1984, Grober-Dunsmore et al. 2004a), the Red Sea (Khalaf & Kochzius 2002) and the Australia (Pittman et al. 2004) and with other habitats such as algal beds (Rossier & Kulbicki 2000) and mangroves (Birkeland 1985, Thollot 1992) in post-hoc correlative analyses, but these results contribute to the growing body of evidence that suggests that resource managers may be able to use landscape-scale measures of reef context to detect areas with high fish abundance and diversity (Appeldoorn et al. 2003, Kendall et al. 2003, Pittman and McAlpine 2003, Jeffrey 2004). Abundances and species richness of predicted reef fish groups (grunts, seagrass-associated taxa, snappers and MIFs) were higher at reefs with neighboring seagrass, and observed patterns of distribution were remarkably consistent with the foraging ecology of each fish group. For example, haemulids exhibited the strongest association to seagrass; several taxa are known to regularly move from reefs into adjacent seagrass patches (Burke 1995, Appeldoorn et al. 1997, Beets et al. 2003), sheltering on or near the reef by day and feeding on crustacean fauna in surrounding seagrass by night (Ogden & Quinn

PAGE 75

58 1984, Robblee & Zieman 1984, Burke 1995, Appeldoorn et al. 1997, Beets et al. 2003). Abundances of seagrass-associated taxa were also strongly associated with seagrass, which is expected since this guild is comprised exclusively of species that complete part of their life history in seagrass. Because many MIF and snapper species have generalist foraging requirements (Sale 2002), it is not unexpected that relationships with seagrassthough still robust are less than that observed for grunts. Remarkably, the strength of associations for abundances and species richness of the different fish groups corresponded quite strongly with their known natural history, with the strength of relationships generally following this rank order; grunts, seagrass-associated, snappers, MIFs, and groupers. This study design does allow inferences concerning the mechanisms that might explain the higher species richness and abundances of reef fishes at reefs proximal to seagrass. (1) Settlement and survivorship of some juvenile fishes may be higher in seagrass (Nagelkerken et al. 2000, 2002, Appeldoorn et al. 2003) due to its high plant diversity (e.g., epiphytes, algae, and phytoplankton) (Bell & Pollard 1989), which may facilitate coexistence of species that would otherwise compete (Keller 1983). Settlement (of individuals and different species) may be higher if seagrass intercepts larval fish more effectively than other habitat patches (Parrish 1989). (2) The structural complexity of seagrass can provide shelter from predation (Parrish 1989, Robertson & Blaber 1992), thereby increasing survivorship of fishes in seagrass communities, which then may migrate to neighboring reef patches as they mature. (3) Reef fish movement may occur more readily in highly-connected marine landscapes (Noss 1983), therefore patches that have habitat linkages to other essential habitat patches may support more fishes. (4)

PAGE 76

59 Energy, nutrients and organic matter generated within seagrass communities (Duarte 2000) may flow to nearby reefs through direct animal movement (Meyer et al. 1983, Meyer & Schultz 1985), predation, or outwelling of dissolved and particulate organic matter (Sogard 1989, Beck et al. 2001), providing important resources for reef fishes. While the mechanisms responsible for increased richness and abundances of fishes at reefs proximal to seagrass remains to be identified, some combination of these factors appears beneficial for recruitment, settlement, survivorship and/or coexistence of large numbers of fishes. Contrary to predictions, groupers were weakly associated with seagrass, although detecting relationships at this spatial scale may be inhibited by their low population densities (0-0.2 fishes per sample) and species composition. Historically, high densities of groupers at several reefs in the U.S. Virgin Islands have been attributed to surrounding seagrass communities (Randall 1962). Over time, grouper species dominance has shifted to smaller, potentially less mobile, taxa; E. guttatus, Cephalopholis cruentatus and C. fulvus are relatively more abundant than E. striatus, E. morio, and Mycteroperca tigris. These latter groupers likely ranged over comparatively large areas; E. striatus made daily movements of 400 m (Carter et al. 1994). In contrast, presently-dominant taxa (e.g., E. guttatus) demonstrated high short-term fidelity at small patch reefs in Bermuda (Bardach 1958) and the U.S. Virgin Islands (Randall 1962, Beets et al. 2003). Habitat requirements may be less stringent, abundances may be severely depressed (thereby limiting our ability to detect relationships; Osenberg et al. 2002), or shifts in the species composition may favor more site-attached species, such that detecting habitat linkages for groupers is difficult in this system.

PAGE 77

60 The juvenile component of the fish assemblage generally exhibited a stronger seagrass association compared to adults, perhaps due to nursery benefits (Nagelkerken et al. 2002) such as increased recruitment, increased habitat-mediated post-settlement survivorship (Beukers & Jones 1997), and/or direct (Meyer et al. 1983, Meyer & Schultz 1985) or indirect transfer (Duarte 2000) of nutrients and energy from adjacent seagrass. The strongest evidence that seagrass provides a nursery benefit is provided by the relationships for juvenile grunts and juvenile seagrass-associated fishes with seagrass, which was also demonstrated by Kendall et al. 2003 in St. Croix for juvenile Haemulon flavolineatum. These two groups are comprised of those taxa with the greatest dependencies on seagrass (Lindeman et al. 2000), and as expected these relationships were relatively strong (R2 = 0.56, R2 = 0.46), respectively. Fishing pressure and within-reef heterogeneity may also explain why the adult component of the reef fish assemblage generally exhibited weaker seagrass associations compared to juveniles. Reef fish populations of the U. S. Virgin Islands appear heavily over-fished (Rogers & Beets 2001), and densities of exploitable-sized fishes were typically low in this study. In addition, adult grunts and snappers feed on benthic invertebrates (Randall 1967, Rooker 1995, Nagelkerken et al. 2000) that typically reside in soft-bottom patches (Gillanders et al. 2003). Such soft bottom patches contribute to within-reef heterogeneity, and these patches may not be detectable at the resolution of these habitat maps (Kendall et al. 2001). Consequently habitat linkages are expected to be more evident in less-fished systems and for juveniles (which was the case) and habitat specialists.

PAGE 78

61 Consistent with predictions and terrestrial research (Sisk et al. 1997, Mitchell et al. 2001, Turner et al. 2001), the mobility (vagility) of fishes influenced how fishes relate to the coral reef landscape. Resident taxa are more likely associated with finer-scaled features (Hixon & Beets 1989, Sale 1998), as evidenced by the positive correlation between rugosity and abundances of resident fishes, and the lack of an association with this landscape measure of seagrass. The absence of a relationship for transient taxa is consistent with their habitat utilization patterns, (e.g., Carynx spp. are likely to transit indiscriminately over sand, reef, and seagrass habitats), as they respond to features at larger spatial scales (i.e. 100s to 1000s of meters). These findings are a strong imperative to focus on the scales that are appropriate to the organism, and indicate that these scales can often be predicted by considering the ecology of the particular species. Reef fishseagrass associations were evident up tp 1 km spatial extent from study reefs, suggesting that functional habitat linkages may operate at least at this spatial scale. Again, the appropriate spatial scale appeared remarkably consistent with the ecology of fishes; particularly with their life history stage. For example, juvenile relationships were most strong at close distances (100 m and 250 m), whereas adults were strongest at 500 m and 1 kilometer. Given the high risk of mortality for juvenile fishes, reefs that have seagrass in closer proximity may reduce predation risks, which would explain in part the strong association at closer distances. Many adult fishes travel greater distances. For example, several tagging studies revealed transit distances of 100 m 400 m for adult H. plumieri (Tulevech & Recksiek 1994), H. sciurus (Beets et al. 2003) and lutjanids (Chapman & Kramer 2000). These results are relevant for determining the geographic

PAGE 79

62 boundaries of MPAs, since failure to include habitat patches at least 1 kilometer away may result in mortality as fishes move outside MPA boundaries. While several recent studies provide evidence of the importance of reef context to reef fish assemblage structure (Nagelkerken et al. 2000, 2002, Appeldoorn et al. 2003, Kendall et al. 2003, 2004, Dorenbosch et al. 2004, Jeffrey 2004, Mumby et al. 2004), this study builds on existing research in several important ways. 1) This study presents data on abundance, rather than simple presence or absence of fishes. This distinction is important since marine resource managers are frequently interested in determining the location of reefs with high numbers of individuals and species. 2) This study differentiates between soft-bottom (e.g., seagrass, sand) and hardbottom (e.g., reef, pavement), which has often been difficult. 3) Measures of each habitat were quantified using geo-referenced, digital habitat maps and GIS, which has historically not been feasible. 4) Confounding effects of previous studies (i.e. reef size, location and other habitat types) were minimized, though not eliminated. 5) Relationships were examined for the entire fish community (e.g., within trophic and taxonomic groups), thereby providing a comprehensive functional perspective of the response of fish parameters to reef context. Previous studies have often been restricted to a subset of the reef fish community (Nagelkerken et al. 2000, 2002, Kendall et al. 2003). 6) Explicitly-stated apriori hypotheses were tested in this study, and study reefs were selected to test these hypotheses. Future studies that apply landscape ecology principles to marine systems should adopt a rigorous hypotheticodeductive approach. To investigate the generality of new findings, it is ideal to determine the temporal and spatial consistency of results (Sale et al. 1994), thus it is relevant that many reef fish

PAGE 80

63 seagrass relationships were similar between 2002 and 2003. For those relationships that were not similar, a loss of statistical power (caused by the reduction in sample size) and low fish densities may have lessened the ability to detect associations (e.g., Eggleston et al. 1998). Reef fish communities display stochastic variability in community structure at small spatial and temporal scales, and it is unclear from this limited temporal analysis whether the absence of several expected relationships in 2002 is an indication that results are not reliable. Clearly, future research is necessary to determine the influence of the complex spatial and temporal variability of reef fish communities in predicting reef fish habitat associations. While multiple sources of temporal variability induced by larval recruitment (Doherty and Fowler 1994), post-settlement mortality (Hixon 1991), fishing pressure and food availability (Williams 1980) exist, habitat relationships appeared remarkably consistent over time, providing added confidence that this evidence of strong habitat linkages is not a fluke consequence from a single sampling event. Contrary to predictions and smaller-scale studies (e.g., Luckhurst and Luckhurst 1978, Hixon and Beets 1989, Friedlander and Parrish 1998), the fine-scale measure of rugosity was not a predictor of most reef fish parameters, which may have important consequences for MPA design. Landscape parameters are increasingly easy to measure (Green et al. 2000, Kendall et al. 2001) conversely; rugosity is a time-consuming in-water measure. If such high quality benthic maps as those produced today can now accurately identify topographically complex reef habitat (Kendall et al. 2001), landscape measures may be superior to fine-scale for designing effective spatial management schemes. The failure of this measure of rugosity to predict remaining fish parameters may be because rugosity is only a positive predictor of fish assemblage structure when sampling across a

PAGE 81

64 larger range of topographic complexity (among distinct habitats). Sampling was constrained within a single high relief habitat for this study, thus the range in rugosity was lower than others (e.g., Friedlander and Parrish 1998). In addition, other measures of structural complexity (e.g., hole size) may be better predictors of fish assemblage structure (Friedlander & Parrish 1998). This multi-scalar analysis confirms that there is clearly no correct scale that can be universally applied to all organisms (Bisonnette 2003), and corroborates terrestrial (Graham and Blake 2001) and marine studies (Grober-Dunsmore et al. 2004b) that habitatfaunal relationships are inextricably scale-dependent. Clearly, scale has profound effects on resultant patterns (Wiens 1989) with fine-scale measures often better predictors of one group of organisms, and landscape measures predictors of others (Mitchell et al. 2001, Mazerolle and Villard 1999). This organism-based perspective appears to be true for coral reef fishes (Pittman and McAlpine 2003, Pittman et al. 2004); consequently future studies should acknowledge that species perceive the landscape in different ways. Conclusions Landscape measures of coral reef context appear to be valuable predictors of coral reef fish assemblage structure, which may have important implications for MPA design. Specifically, the areal coverage of seagrass may be used to successfully predict which reefs have diverse reef fish assemblages and high abundances of commercially and recreationally important species. Although this study cannot identify the processes driving relationships, it provides strong evidence that functionally-linked marine landscapes contribute to increased species richness and abundance of many fish groups. Admittedly, MPA design requires an understanding of numerous factors (i.e. location, distribution and amount of various habitats necessary for spawning, recruitment, larval

PAGE 82

65 export, settlement, growth, foraging and reproduction). However, given the urgency of MPA design decisions, the selection of areas for conservation should also consider their contribution to the whole system and how well the location of a patch relates or links to other patches within the landscape.

PAGE 83

66 Figure 3-1. Location of the 22 study reefs around the island of St. John, US Virgin Islands sampled in 2002, with the eight study reefs re-sampled in 2003 indicated in bold.

PAGE 84

67 R2 = 0.58p < 0.00013035404550556065 70a) 00.20.40.60.811.21.41.61.8Cumulative species richness R2 = 0.27p < 0.01151719212325272931333500.20.40.60.811.21.41.61.8Seagrass 250 mMean species richnessb) Figure 3-2. The relationship of a) cumulative richness and b) mean species richness with the areal coverage of seagrass (hectares) at 250 m for the 22 study reefs sampled in 2002 in St. John, U.S. Virgin Islands. The x-axis is log10 (x +1) transformed.

PAGE 85

68 R2 = 0.41p < 0.00011.11.21.31.41.51.61.71.81.922.100.20.40.60.811.21.41.61.8MIF mean abundancea) R2 = 0.57p < 0.000100.20.40.60.811.21.41.61.800.20.40.60.811.21.41.61.8Haemulid mean abundanceb) R2 = 0.50p < 0.000200.20.40.60.811.21.41.61.8200.20.40.60.811.21.41.61.8Seagrass 250 mSeagrass-associated mean abundance c) R2 = 0.40p < 0.002900.10.20.30.40.50.60.70.80.900.20.40.60.811.21.41.61.8Seagrass 250 mLutjanid mean abundanced) Figure 3-3. The relationship of mean abundances of a) MIFs, b) haemulids, c) seagrass-associated taxa and d) lutjanids with the areal coverage of seagrass (hectares) within 250 m of the 22 study reefs in St. John, U.S. Virgin Islands sampled in 2002. Mean abundances and the areal coverage of seagrass are log10 (x +1) transformed.

PAGE 86

69 R2 = 0.69p = 0.000115202530354000.20.40.60.811.21.41.61.8MIF cumulative richness a) R2 = 0.70p = 0.000101234567800.20.40.60.811.21.41.61.8Haemulid cumulative richnessb) R2 = 0.41p = 0.001300.5 11.52Lutjanid cumu 2.533.544.500.20.40.60.811.21.41.61.8Seagrass 250 mlative richnessc) Figure 3-4. The relationship of cumulative species richness of a) MIFs, b) haemulids and c) lutjanids with the areal coverage of seagrass (hectares) within 250 m for 22 study reefs in St. John, U.S. Virgin Islands sampled in 2002. The y-axis is log10 (x+1) transformed.

PAGE 87

70 303540455055606570758000.20.40.60.81Cumulative richness2003Rho = 0.57p < 0.12002Rho = 0.66p < 0.07a) 1.51.71.92.12.32.500.20.40.60.81Total abundance 2002Rho = 0.59p < 0.112003Rho = 0.57p < 0.13b) 00.511.5200.20.40.60.81Abundance of juvenile haemulids2003Rho = 0.81p < 0.022002Rho = 0.81p < 0.02c) 00.10.20.30.4 0. 500.20.40.60.81Abundance of juvenile epinephelids2003Rho = 0.76p < 0.012002Rho = 0.83p < 0.01d) 152025303500.20.40.60.81Seagrass 250 mMIF richness 2003Rho = 0.84p < 0.012002Rho = 0.94p < 0.01e) 01234 5 00.20.40.60.81Seagrass 250 mRichness of epinephelids2003Rho = 0.84p < 0.012002Rho = 0.73p < 0.03f) Figure 3-5. Spearman rank correlations for those reef fish parameters that demonstrated a consistent relationship with the areal coverage of seagrass habitat between 2002 ( ) and 2003 ( ) at the subset of 8 study reefs in St. John, U.S. Virgin Islands

PAGE 88

71 Table 3-1. Most abundant taxa in each reef fish group for the 22 study reefs sampled in 2002 in St. John, U.S. Virgin Islands. Mobile invertebrate feeders Reef fish parameter Dominant taxa Total abundance Thalassoma bifasciatum, Acanthurus coerulus, Haemulon aurolineatum, Scarus spp. (unidentified juveniles), H. flavolineatum, Stegastes planifrons, A. bahianus, Halichoeres garnoti, S. leucostictus, and Caranx ruber MIFs* H. flavolineatum, H. aurolineatum, H. garnoti, Abedufduf saxtalis, Halichoeres bivittatus, Ocyurus chrysurus, Holocentrus rufus, Halichoeres maculapinna, Hypoplectrus puella, and Mulloides martinicus Haemulids (grunts) Haemulon aurolineatum, H. flavolineatum, H. juvenile, H. sciurus, H. plumieri, H. parrai, H. carbonarium, H. macrostomum, and H. chysargyreum Lutjanids (snappers) Ocyurus chrysurus, Lutjanus apodus, L. synagris, L. mahogani, L. jocu, L. griseus, and L. analis Epinephelines (groupers) Epinephelus guttatus, Cephalopholis cruentatus, C. fulvus, E. striatus, and E. adscensionis Mobile taxa Acanthurus coerulus, Haemulon aurolineatum, Scarus spp. (unidentified juveniles) Haemulon flavolineatum, Halichoeres garnoti, Haemulon spp. (unidentified juveniles), Thalassoma bifasciatum, Halichoeres bivittatus, and Sparisoma aurofrenatum Resident taxa Stegastes planifrons, S. leucostictus, S. partitus, Abedufduf saxtalis, Chromis cyanea, C. multilineatum, Hypoplectrus puella, and S. variabilis Transient taxa Carynx ruber, Ocyurus chrysurus, Inermia vittata, Scomberomorous regalis, and Dasyatis americana

PAGE 89

72 Table 3-2. Variable names, transformations, minimum, maximum and mean values for each reef fish parameter and landscape metric, with the standard error for fish parameters and the coefficient of variation for habitat measures for the 22 study reefs sampled in 2002 in St. John, US Virgin Islands. a Divided into adult and juvenile components, also. bThose parameters that failed Shapiro-Wilks normality tests, using p < 0.01. Dependent Variable Transform Min Max Mean SE Cumulative richness Log (x +1) 30.00 66.00 44.20 2.37 Mean species richness Log (x +1) 16.00 0.66 22.42 0.97 Total abundance Log (x +1) 322.00 38.81 99.00 0.05 MIF abundance Log (x +1) 10.22 99.00 25.92 0.06 Grunt abundance a Log (x +1) 0.00 59.26 3.89 0.12 Snapper abundance a Log (x +1) 0.00 22.99 1.69 0.07 Grouper abundance a Log (x +1) 0.00 2.09 0.78 0.03 Seagrass asso. abundancea Log (x +1) b 0.50 68.50 13.2 3.92 Mobile abundance Log (x +1) 26.54 193.98 59.26 0.05 Residents abundance Log (x +1) 8.77 101.33 25.92 0.06 Transients abundances Log (x +1) b 0.00 17.20 2.31 0.08 Richness MIFs None 9.00 30.00 17.36 1.52 Richness haemulids None 0.00 6.00 3.13 0.44 Richness lutjanids None 0.00 4.00 2.18 0.28 Richness epinephelids None b 0.00 6.00 3.59 0.23 Richness mobile None 17.00 40.00 26.10 1.53 Richness residents None 8.00 24.00 15.68 0.86 Richness transients None 0.00 4.00 2.18 0.23 Independent Variable Transform Min Max Mean CV Area (ha) seagrass 100 m Log (x +1) 0.00 11.59 1.21 97.96 Area (ha) seagrass 250 m Log (x +1) 0.00 25.92 3.27 78.36 Area (ha) seagrass 500 m Log (x +1) b 0.00 41.36 6.76 65.23 Area (ha) seagrass 1 km Log (x +1) 1.01 76.62 13.79 38.97 Rugosity None 1.18 2.28 1.56 15.70

PAGE 90

73 Table 3-3. Simple linear regression for entire assemblage level parameters of reef fish communities with the areal coverage of seagrass within 1 km of each study reef as the independent variable at the 22 study reefs, sampled in 2002 in St. John, U.S. Virgin Islands*. *For each relationship, the dependent and independent variables are presented with corresponding F-ratio, entire model R2 and p-value for the final model. Relationships in bold represent the spatial extent with the strongest relationship, for the given fish parameter. All relationships are positive (+). Dependent variable Independent Variable F ratio Model R2 (x 100) p-value Mean Species Richness Seagrass 100m (+) Seagrass 250 m (+) Seagrass 500 m (+) Seagrass 1 km (+) 6.21 7.42 10.95 7.82 23.68 27.02 35.38 29.16 0.0200 0.0100 0.0035 0.0015 Cumulative Species Richness Seagrass 100 m (+) Seagrass 250 m (+) Seagrass 500 m (+) Seagrass 1 km (+) 17.62 27.29 24.02 19.59 47.84 57.77 54.57 50.77 0.0004 0.0001 0.0001 0.0003 Mean Abundance Seagrass 100 m (+) Seagrass 250 m (+) Seagrass 500 m (+) Seagrass 1km (+) 4.22 6.21 9.38 8.98 17.44 23.60 31.94 32.09 0.0500 0.0200 0.0060 0.0070

PAGE 91

74 Table 3-4. Simple linear regression of abundances of mobile invertebrate feeders, grunts, snappers, groupers, and seagrass-associated taxa and within mobility guilds with the areal coverage of seagrass within 1 km of each study reef at the 22 study reefs, sampled in 2002 in St. John, U.S. Virgin Islands*. *Models with p-values less than 0.15 are shown. a p > 0.05 level. NS means not significant at the p < 0.15. b Bold represents the spatial extent with the strongest relationship. Dependent variable Independent Variable F-ratio Model R2 (x 100) p-value Adult Haemulids Seagrass 100 m (+) Seagrass 250 m (+) Seagrass 500 m (+) Seagrass 1 km (+) 3.82 4.85 5.34 5.03 16.06 19.50 21.08 20.92 0.064* 0.0390 0.0300 0.0370 Juvenile Haemulids Seagrass 100 m (+) Seagrass 250 m (+) Seagrass 500 m (+) Seagrass 1 km (+) 21.38 23.36 19.84 21.27 54.29 56.48 52.43 55.59 0.0002 0.0001 0.0003 0.0002 Adult Lutjanids Seagrass 100 m (+) Seagrass 250 m (+) Seagrass 500 m (+) Seagrass 1 km (+) 3.52 5.29 6.06 5.56 16.37 22.72 25.20 22.63 0.076* 0.0330 0.0240 0.0290 Juvenile Lutjanids Seagrass 100 m (+) Seagrass 250 m (+) Seagrass 500 m (+) Seagrass 1 km (+) 5.34 4.48 3.20 --21.08 18.65 13.81 --0.0300 0.0400 0.080* NS Adult Epinephelids Seagrass 100 m (+) Seagrass 250 m (+) Seagrass 500 m (+) Seagrass 1 km (+) --3.10 4.73 ----14.70 20.81 --NS 0.090* 0.0400 NS Juvenile Epinephelids Seagrass 100 m (+) Seagrass 250 m (+) Seagrass 500 m (+) Seagrass 1 km (+) ----------------NS NS NS NS Adult seagrass-associated taxa Seagrass 100 m (+) Seagrass 250 m (+) Seagrass 500 m (+) Seagrass 1 km (+) 3.42 5.63 7.14 14.59 21.96 26.30 0.079* 0.0280 0.0150 Juvenile seagrass-associated taxa Seagrass 100 m (+) Seagrass 250 m (+) Seagrass 500 m (+) Seagrass 1 km (+) 17.26 13.87 9.92 46.32 40.05 33.14 0.0005 0.0013 0.0051

PAGE 92

75 Table 3-5. Simple linear regression of abundances of the adult and juvenile components for grunts, snappers, groupers, and seagrass-associated taxa with the areal coverage of seagrass within 1 km of each study reef at the 22 study reefs, sampled in 2002 in St. John, U.S. Virgin Islands. Models with p-values less than 0.15 are shown. a p > 0.05 level. NS means not significant at the p < 0.15. b Bold represents the spatial extent with the strongest relationship. Dependent variable Independent Variable F ratio Model R2 (x100) p-value MIFs Seagrass 100m (+) Seagrass 250 m (+)b Seagrass 500 m (+) Seagrass 1 km (+) 9.50 11.13 10.01 6.88 32.23 41.25 33.36 26.60 0.0058 0.0033 0.0049 0.0160 Grunts (Haemulids) Seagrass 100 m (+) Seagrass 250 m (+) Seagrass 500 m (+) Seagrass 1 km (+)b 20.38 23.47 19.27 22.84 53.10 56.59 51.66 57.32 0.0003 0.0001 0.0001 0.0002 Snappers (Lutjanids) Seagrass 100 m (+) Seagrass 250 m (+)b Seagrass 500 m (+) Seagrass 1 km (+) 10.26 11.82 9.42 8.83 36.29 39.64 34.35 34.17 0.0049 0.0029 0.0066 0.0086 Groupers (Epinephelines) Seagrass 100 m (+) Seagrass 250 m (+) Seagrass 500 m (+) Seagrass 1 km (+)b 2.57 3.71 2.67 3.57 12.51 17.09 12.93 17.36 0.120a 0.070 a 0.120 a 0.070 a Seagrass-associated taxa Seagrass 100 m (+) Seagrass 250 m (+)b Seagrass 500 m (+) Seagrass 1 km (+) 16.05 14.93 11.21 10.35 44.52 50.12 35.01 33.42 0.0007 0.0002 0.0032 0.0030 Resident taxa Seagrass 100 m (+) Seagrass 250 m (+) Seagrass 500 m (+) Seagrass 1 km (+) ---5.76 ---------22.36 ------NS 0.0260 NS NS Mobile taxa Seagrass 100 m (+) Seagrass 250 m (+)b Seagrass 500 m (+) Seagrass 1 km (+) 4.79 4.86 ---4.47 20.00 20.34 ---19.07 0.0406 0.0390 NS 0.0478 Transient taxa Seagrass 100 m (+) Seagrass 250 m (+) Seagrass 500 m (+) Seagrass 1 km (+) ------------------------NS NS NS NS

PAGE 93

76 Table 3-6. Simple linear regression analyses of cumulative species richness of MIFs, haemulids, epinephelids, lutjanids and within resident, mobile and transient mobility guilds the areal coverage of seagrass within 1 km of each study reef as the independent variable at the 22 study reefs, sampled in 2002 in St. John,US Virgin Islands. Dependent variable Independent Variable F ratio Model R2 (x 100) p-value Cumulative species richness of MIFs Seagrass 100m (+) Seagrass 250 m (+) Seagrass 500 m (+) Seagrass 1 km (+) 46.47 63.59 45.84 28.83 72.29 68.59 71.81 62.91 0.0001 0.0001 0.0001 0.0001 Cumulative species richness of Haemulids Seagrass 100 m (+) Seagrass 250 m (+) Seagrass 500 m (+) Seagrass 1 km (+) 49.89 45.46 25.36 28.59 71.38 69.44 55.91 60.07 0.0001 0.0001 0.0001 0.0001 Cumulative species richness of Lutjanids Seagrass 100 m (+) Seagrass 250 m (+) Seagrass 500 m (+) Seagrass 1 km (+) 12.63 14.09 8.22 6.06 38.73 41.33 29.12 24.19 0.0020 0.0013 0.0095 0.0235 Cumulative species richnes of Epinephelids Seagrass 100 m (+) Seagrass 250 m (+) Seagrass 500 m (+) Seagrass 1 km (+) 8.77 9.06 7.19 5.08 32.76 33.49 28.57 23.00 0.0084 0.0075 0.0152 0.0377 Cumulative species richness of Resident Taxa Seagrass 100 m (+) Seagrass 250 m (+) Seagrass 500 m (+) Seagrass 1 km (+) ----------------NS NS NS NS Cumulative species richness of Mobile Taxa Seagrass 100 m (+) Seagrass 250 m (+) Seagrass 500 m (+) Seagrass 1 km (+) 29.79 43.68 34.31 23.51 59.83 68.59 63.17 55.30 0.0001 0.0001 0.0001 0.0001 Cumulative species richness of Transient Taxa Seagrass 100 m (+) Seagrass 250 m (+) Seagrass 500 m (+) Seagrass 1 km (+) ----6.21 ------25.65 --NS NS 0.0220 NS Cumulative species richness of Transient Taxa Seagrass 100 m (+) Seagrass 250 m (+) Seagrass 500 m (+) Seagrass 1 km (+) ----6.21 ------25.65 --NS NS 0.0220 NS Models with p-values less than 0.15 are shown. a p > 0.05 level. NS means not significant at the p < 0.15. b Bold represents the spatial extent with the strongest relationship.

PAGE 94

77 Table 3-7. Spearman rank correlations of relationships of each fish parameter and the areal coverage of seagrass at the 250 m spatial extent for the 8 study reefs in St. John, sampled in 2002 and 2003. For each relationship, the Spearman Rho correlation, and probability < Rho are presented by year. Kruskal-Wallis tests for significant differences in abundance using raw census data for 2002 (n = 43) and 2003 (n = 87) between years are presented, with direction of change indicating an increase (+) or decrease (-) in abundance. Relationships in bold are those that are consistent between years. All p-values < 0.20 are presented. Reef fish parameter 2002 Rho Prob Rho 2003 Rho Prob Rho KruskalWallis Direction of change Cumulative richness 66.21 0.07 56.63 0.100 Mean species richness 59.21 0.11 57.14 0.130 Total abundance 59.21 0.11 57.14 0.130 0.186 Abundances of MIFs 61.23 0.09 --NS 0.016 Abundances of adult MIFs 52.78 0.18 --NS 0.137 Abundances of juvenile MIFs 52.78 0.18 --NS 0.386 Abundances of adult haemulids --------NS 0.022 Abundances of juvenile haemulids 88.35 0.004 80.95 0.015 0.690 Abundances of adult epinephelids 68.45 0.06 --NS 0.268 Abundances of juvenile epinephelids 82.47 0.01 76.19 0.009 0.292 Abundances of adult lutjanids ------NS 0.012 Abundances of juvenile lutjanids 50.36 0.20 --NS 0.847 Cumulative richness MIFs 84.09 0.001 93.86 0.001 Cumulative richness haemulids 75.16 0.03 --NS Cumulative richness lutjanids ------NS Cumulative richness epinephelids 73.03 0.03 84.01 0.001

PAGE 95

78 Table 3-8. Influence of fine-scale (rugosity) and landscape-scale (seagrass) features in predicting reef fish parameters for the 22 study reefs sampled in 2002. Dependent variable Independent Variable Partial R2 Model R2 p-value Cumulative richness Seagrass 250 m Rugosity 71.76 --71.76 < 0.0001 Mean species richness Seagrass 250 m Rugosity 35.67 --35.67 0.0054 Total abundance Seagrass 250 m Rugosity 9.60 33.10 42.70 0.0033 Abundances of MIFs Seagrass 250 m Rugosity 40.36 --40.36 0.0026 Abundances of Haemulids Seagrass 250 m Rugosity 56.59 --56.59 < 0.0001 Abundances of Epinephelids Seagrass 250 m Rugosity 17.09 --17.09 0.070* Abundances of Lutjanids Seagrass 250 m Rugosity 39.64 --39.64 0.0029 Abundances of Resident taxa Seagrass 250 m Rugosity --28.92 28.92 0.0200 Abundances of Mobile taxa Seagrass 250 m Rugosity 33.79 12.89 46.58 0.0072 Species richness of MIFs Seagrass 250 m Rugosity 69.35 --69.35 < 0.0001 Species richness of Haemulids Seagrass 250 m Rugosity 69.98 --69.98 < 0.0001 Species richness of Epinephelids Seagrass 250 m Rugosity 19.97 --19.97 0.0400 Species richness of Lutjanids Seagrass 250 m Rugosity 41.34 41.34 0.0013 Species richness of Resident taxa Seagrass 250 m Rugosity 25.41 --25.41 0.0200 Species richness of Mobile taxa Seagrass 250 m Rugosity 69.37 --69.37 < 0.0001 Models with p-values less than 0.15 are shown. An asterisks (*) designates where p > 0.05 level. NS means not significant at the p < 0.15. Relationships in bold indicate where rugosity was a significant predictor variable in the final model.

PAGE 96

CHAPTER 4 REEF FISHES RESPOND TO VARIATION IN LANDSCAPE STRUCTURE Reef context explained a significant amount of the variability in reef fish assemblage structure in the Florida Keys National Marine Sanctuary (FKNMS) and the US Virgin Islands. Although reef context was significant in both systems, the specific measure of context varied. Rather, the particular habitat type (e.g., seagrass, pavement) responsible for the reef fishhabitat relationships differed between systems. Although the areal coverage of seagrass positively predicted abundances and species richness of mobile invertebrate feeders, haemulids, and lutjanids in the US Virgin Islands, seagrass did not explain a significant amount of the variation of these same fishes when analyses was limited to the FKNMS only. Instead, the amount of pavement and reef habitats were positively associated with several reef fish parameters in Florida. Differences in the landscape structure of the two systems may explain this disparity. In the US Virgin Islands, seagrass comprises a relatively small proportion of the essential fish habitat. However in Florida, seagrass is the dominant habitat type, whereas pavement, which was common in the US Virgin Islands, comprises a smaller proportion of the essential fish habitat. Thus, the processes that structure reef fish communities appear to respond to variation in the landscape structure of these coral reef environments. The shape of reef fishseagrass curves for pooled data including the FKNMS and US Virgin Islands suggest that there may be critical threshold of seagrass habitat. Once this critical threshold of seagrass is exceeded, seagrass may become less important in structuring reef fish communities than other habitat types. These results are relevant to marine protected 79

PAGE 97

80 areas design, since they suggest that general design rules do not necessarily apply across systems. Comparative studies such as this are critical for developing the universal design principles required to establish marine protected areas that meet their conservation and/or fisheries objectives. Introduction Landscape-scale metrics, traditionally employed in the study of terrestrial systems, have been applied recently to the study of coral reef ecosystems to better understand fishhabitat relationships (Grober-Dunsmore et al. 2004a, 2004b, 2004c and Kendall et al. 2003 in the US Virgin Islands; Christensen et al. 2003 in Puerto Rico, Appeldoorn et al. 2003 in Columbia, Jeffrey 2004 in the Florida Keys). Coral reef ecosystems, like most terrestrial ecosystems, can be described quantitatively by the spatial pattern or arrangement of landscape elements (Forman 1995, Dramstad et al. 1996), including patches of habitat and corridors of movement. Ault and Johnson (1998a, 1998b), Appeldoorn et al. (2003), Kendall et al. (2003) and Grober-Dunsmore et al. (2004a, 2004b, 2004c) have all used a landscape-scale approach to demonstrate significant fishhabitat relationships, and generally conclude that the context of individual habitat patches may be a critical factor determining reef fish community structure. In the latter case, simple measures of reef context (calculated with GIS software) were used to predict reef locations with high reef fish abundances and reef fish diversity in the US Virgin Islands (Grober-Dunsmore et al. 2004a, 2004b). Mean abundance and species richness of several groups of exploited reef fishes, in particular, showed strong positive correlations with those reefs in close proximity to seagrass habitat (Grober-Dunsmore et al. 2004c). The generality of these findings, however, have not yet been tested.

PAGE 98

81 To assist in the development of an emerging, coherent landscape-scale theory that explains the structure of reef fish communities, it is essential to critically examine observed patterns of reef fish abundance, distribution, species richness and diversity, and determine which landscape elements, if any, best describe these patterns. Furthermore, it is necessary to explore how reliably these landscape elements predict assemblage structure in other systems, and under differing environmental conditions. Studies of landscape structure effects on fauna in terrestrial habitats (Paton 1994, Trzcinski et al. 1999, Villard et al. 1999) often yield markedly different results, and, as a consequence, generalizations regarding these relationships have been slow to evolve. Inconsistencies in relational patterns can occur due to discrepancies in the scale of observation (Hewitt et al. 1998, Wiens and Milne 1989), the natural heterogeneity of ecosystems (Kolasa and Pickett 1991), disparity in the metrics used to measure spatial patterning (Frohn 1998), differences in sampling techniques and in the spatial resolution and grain of the benthic habitat maps. In the pursuit of identifying general operational guidelines for MPA design, it will be mandatory to replicate experiments in time and space (Sale 2002), to understand how variation in the structure of underlying habitats may influence the composition, structure and distribution of reef fish communities. Faunal abundance and survivorship of specific organisms in marine systems has been associated with landscape configuration (Robbins and Bell 1994, Irlandi et al. 1999, Hovel and Lipcius 2001), landscape context (Bell et al. 2001, Appeldoorn et al. 2003, Grober-Dunsmore et al. 2004a, 2004b, 2004c), and reef connectivity (Ault and Johnson 1998), although few of these studies have been replicated in space, limiting our ability to draw general conclusions. The present study was carried out in the Florida Keys

PAGE 99

82 National Marine Sanctuary (FKNMS), and was designed to test the generality of reef fishhabitat relationships previously detected in the US Virgin Islands (Grober-Dunsmore et al. 2004c). FKNMS represents one of the few other Caribbean locations with digitally-referenced benthic habitat maps, and was therefore amenable to applying the same landscape approach and methods as that used in the previous study. FKNMS was selected for several reasons. First, the FKNMS habitat maps were created jointly by Florida Marine Research Institute (FMRI) and National Oceanographic and Atmospheric Association (NOAA). Because the habitat maps for the US Virgin Islands were also created by NOAA, I assumed some consistency in methods throughout the classification and interpretation process. Furthermore, the grain, which is the smallest resolvable unit of study (King 1991) and the extent, which is the area over which observations are made (Morrison and Hall 2001) were comparable between studies. This is important to my comparisons, since interpretation of how ecological systems are structured often depends on the spatial and temporal scale at which a study is conducted. In fact, results of studies conducted at different scales may not be comparable (Osenberg et al. 1999). Second, because reef fish populations in the US Virgin Islands are heavily exploited and several targeted fish groups (e.g., groupers) exhibit low population densities (Rogers and Beets 2001), many reef fishhabitat relationships were hypothesized to be stronger in less-fished systems, such as no take areas where abundances of reef fish populations are expected to be higher. The designation of a network of special protected areas (SPA) in the FKNMS afforded the opportunity to sample inside no fishing areas (where reef fish populations were expected to be less fished than the US Virgin Islands) and to compare reef fishhabitat relationships inside and outside of no-take areas. Finally, it is

PAGE 100

83 preferential to test the generality of landscape relationships in systems where the habitat patches are arranged differently, to determine the robustness of landscape variables as predictors among locations (Bissonette and Storch 2003). The US Virgin Islands is an insular fringing reef system with a highly variable distribution of habitat patches comprised of seagrass, reef, sand, and pavement interspersed in no particular pattern from shore. In comparison, the FKNMS is a continental bank reef system with a more homogeneous distribution of habitat patches. For example, the continuous bank system of shallow spur and groove reef is 6-10 kilometers offshore, and separated from the coast by large expanses of seagrass habitat, interspersed with patch reefs. If reef fishhabitat relationships are similar and robust across these two disparate ecosystems, i.e. the US Virgin Islands and FKNMS, patterns might be expected to hold across other Caribbean coral reef systems. The objective of this study was to determine how robust reef fishhabitat relationships are to variation in landscape structure, since the expectation that relationships will be similar across systems assumes that the processes that gave rise to these relationships are not modified by variation in landscape structure. The primary predictions, based on my findings in the US Virgin Islands, were that reef fish assemblage structure would not be correlated with reef configuration (e.g., reef size, shape), complex landscape-scale (e.g., habitat diversity) or fine-scale (rugosity) metrics of habitat heterogeneity, yet reef fishes would be positively correlated with reef context, i.e. the spatial arrangement and composition of surrounding habitat patches. Specifically, I predicted that species richness and abundances of targeted fishes (i.e. mobile invertebrate feeders, haemulids, lutjanids, and serranids) would be higher at reefs

PAGE 101

84 proximal to seagrass. If these reef fish assemblage parameters were not higher at reefs proximal to seagrass habitat, I expected that variation in the landscape structure of the coral reef environments of the US Virgin Islands and FKNMS would explain differences in the response of reef fishes to landscape structure. In addition, relationships were also explored across the largest landscape gradient possible, by pooling data from the US Virgin Islands and FKNMS. Furthermore, I expected that reef fish assemblage structure would be associated with landscape-scale measures of landscape structure to a greater extent than with the fine-scale measure of rugosity. I also tested whether there were significant differences in abundance of targeted reef fishes within protected and unprotected reference sites to assess whether reef fishhabitat relationships were stronger at those sites with higher reef fish abundances. Methods Study Areas This study was conducted in the FKNMS in May 2003, and results from the FKNMS were compared to those from a previous study in St. John, U.S. Virgin Islands (Grober-Dunsmore et al. 2004c). The FKNMS (980,000 ha), designated in 1990 to protect nationally significant biological and cultural marine resources, including critical coral reef habitats, seagrass beds, hard-bottom communities, and mangrove shorelines, encompasses all but the northernmost extent of the Florida Reef Tract (Department of Commerce 1997). A network of 24 fully protected zones, which cover approximately 6% of the Sanctuary, but protect 65% of shallow bank reef habitats and 10% of coral resources overall, were implemented in 1997 and in 2001 to preserve specific areas more completely (NOAA et al. 2002). Inside these special protected areas (SPA), most fishing activities are prohibited. In contrast, the island of St. John has 56% of the marine portion

PAGE 102

85 of its near shore waters under the jurisdiction of Virgin Islands National Park (VINP). The VINP was established in 1956 with marine portions totaling 2,287 ha added in 1962. Most types of fishing are allowed within VINP, and effects of intensive fishing are evident (Rogers and Beets 2001). While spear fishing is banned, rod and reel, line traps and fish traps are still allowed within VINP and the territorial waters of the US Virgin Islands. Study reefs in Florida were selected to facilitate comparisons between systems by reducing variability in reef type and inshore-offshore effects and to include the maximum number of SPA and reference pairs. Therefore, all SPAs (which were all spur and groove habitat), and their approximate reference (fished) reefs were sampled (Figure 4-1). To the extent feasible, I selected the reference monitoring sites used by Bohnsack et al. (1999). Following ground-truthing, only four SPA-reference pairs were used as comparative reef pairs to reduce variability based on within-reef characteristics such as coral cover and topographic complexity. Thus, all study reefs in the FKNMS were relatively shallow (2 m) and comprised of an eroding submerged reef framework, with living coral cover estimated between 2 and 20 % (Miller et al. 2000). Dominant corals were Montastraea annularis with Porites astreoides, P. porites, Diploria labyrinthiformes, D. strigosa, Siderastraea siderea and S. radians making lesser contributions to living coral cover. The background matrix consisted of sand channels in deeper areas (5-13 m) and an eroding calcium carbonate framework in shallower areas (1-5 m). In comparison, the 22 study reefs sampled in 2002 in the US Virgin Islands (shallow (1 m) fringing and patch reefs; Grober-Dunsmore et al. 2004c) were selected

PAGE 103

86 as the primary comparative reefs so that the dominant shallow water reef types for each system were used for comparisons. Several were contained in a sand matrix with large boulders of coral, primarily M. annularis, providing the major topographic relief. Other reefs had more distinct zonation with edge and platform (hard-bottom) zones. Total living coral cover in the US Virgin Islands, estimated at 5 %, was dominated by M. annularis, with lesser contributions by Porites astreoides, S. siderea, S. radians, D. labyrinthiformes and D. strigosa. Habitat Sampling Aerial photographs of the FKNMS were taken in December 1991-March 1992 and were interpreted by FMRI and NOAA personnel in 1998 (Florida Marine Research Institute et al. 1998). Habitat types were classified as one of four major categories, i.e. corals, seagrasses, hard-bottom and bare substrate, and these four broad categories were further refined to represent 24 subcategories at a nominal scale of 1:48,000 (see Florida Marine Research Institute et al. 1998 for details). Benthic habitat maps of the US Virgin Islands were created in 2001, from aerial imagery acquired in 1999 (Kendall et al. 2001). For the FKNMS benthic maps, the minimum area delineated was 0.5 ha, which contrasts slightly with the minimum area of 0.4 ha for the US Virgin Islands. The maps were created approximately 5-9 years apart, with the US Virgin Islands being more recent. To the extent possible, landscape metrics in the FKNMS were calculated the same as in the 2002 study in the US Virgin Islands (Grober-Dunsmore et al. 2004a, 2004b, 2004c). To facilitate comparisons between systems, the 24 habitat classes in the FKNMS were reduced to the same number of classes (n = 7) as in the US Virgin Islands. Habitat classes in the FKNMS were: inland water, pavement, reef, patch reef, sand, seagrass, and unknown, although only five of these habitat classes occurred within 500 m of my study

PAGE 104

87 reefs. These habitat classes contrast slightly with those in the US Virgin Islands, i.e. pavement, reef, bedrock, sand, seagrass, deep unknown, and macro algal plain. The classification schemes were created to be as similar as possible, given differences between the systems (e.g., bedrock was not classified in Florida, and thus was not included; patch reefs were classified separately in Florida to keep the total number of habitats as equitable as possible). However, there were slight differences in the definitions for unknown habitat. In the US Virgin Islands, deep unknown represents habitat that could not be identified from aerial photos, typically due to depths in excess of 20 m (pers. obs.). Groundtruthing revealed, however, that much of this was soft-bottom and to a lesser extent hard-bottom. In the Florida Keys, turbid water in the inshore areas precluded accurate classification of large portions of Hawks Channel, and was classified as unknown (FMRI 1998). Groundtruthing revealed that much of this was seagrass habitat (pers. obs.). Comparisons of the coral reef landscape structure of the FKNMS and US Virgin Islands were conducted at reef and ecosystem-scales. Comparisons at the reef-scale were conducted at 100 m, 250 m and 500 m from study reefs, because spatial extent should be two to five times larger than the spatial features of the habitat patches in the landscape (ONeill et al. 1996). To account for minor differences in the size of study reefs, I also calculated the proportion of area of each habitat type per unit area of sample reef. At the ecosystem-scale, the area of each habitat within the entire mapped spatial extent (e.g., Carysfort to Sand Key in the FKNMS; see Figure 4-1) was calculated for each system. Reef Fish Sampling Fish sampling was conducted using the point count method described by Bohnsack and Bannerot (1986). Sampling effort in the US Virgin Islands was determined based on

PAGE 105

88 reef size to achieve a minimum of 1 point count per 1500 m2 following a sample size optimization procedure using pre-existing data (Grober-Dunsmore et al. 2004a). Sampling effort was increased in FKNMS so that a minimum of 12 point counts per reef were conducted, regardless of reef size (Table 4-2) in an attempt to improve accuracy of reef fish parameter estimates within each reef. Comparisons of the mean values for reef fish parameter estimates in the FKNMS were conducted between the two levels of sampling effort. In general, there were not differences in mean abundance estimates with increased sampling, though species richness obviously increased with increased sampling. Point counts were randomly-generated in each reef polygon using a script in Arc View 3.2 (ESRI 1996). Corresponding GPS coordinates were used to approximate the location of point counts using waterproof site maps, GPS coordinates, and ranging techniques in the water. Reef fish census data were used to calculate a suite of reef fish parameters, which are the same as those in Grober-Dunsmore et al. 2004b (Table 4-1). Statistical Analyses Predictions of the relationships of reef fish assemblage structure to various landscape structure parameters were derived based on my findings from study reefs in the US Virgin Islands. To confirm that findings in the FKNMS were similar to those in the US Virgin Islands, the same stepwise multiple regression analyses were conducted in both studies. Every reef fish parameter was tested and used as the dependent variable (Table 4-1), and landscape variables were, in general, the same as those in Grober-Dunsmore et al. 2002b. The only differences were that reef context measures differed slightly between the US Virgin Islands and FKNMS, due to differences in classification (see above). In the FKNMS, reef context measures were: the areal coverage (ha) of reef, patch reef, pavement, seagrass and sand habitats. The areal coverage of reef, seagrass,

PAGE 106

89 pavement, bedrock and sand habitats were used for the US Virgin Islands. The strength and nature of reef fishhabitat relationships were explored at the 100-meter, 250-meter and 500-meter spatial extents; these spatial scales were selected based on the known ecology of the majority of the focal organisms (Kotliar and Wiens 1990), the reef fish community as a whole. To further explore the generality of findings across systems, relationships were regressed using the FKNMS (n = 17) and USVI (n = 22) reefs for a total of 39 reefs, thereby maximizing the gradient for several landscape variables. For all regressions, a p-value of 0.05 was set for variables to enter the model, to reduce Type I errors (Zar 1984). Model selection was improved using Mallows p (Mallows 1973), leverage effects plots (Sall 1990) and Akaikes Information Criterion (AIC) (Akaike 1974). Because it is not appropriate to simply select the model with the lowest AIC value, each alternative model was evaluated by determining the difference between model AIC and the minimum AIC; i < 2 was used as an acceptable value (Burnham and Anderson 1998). In addition, sequential Bonferroni Dunn-Sidak corrections were applied, using the total number of comparisons (n = 30) (Zar 1984). For every significant multiple regression, a simple linear regression was conducted to graphically display data and perform residual analyses (Sokal and Rohlf 1995). To determine whether reef fishhabitat relationships were stronger where abundances of reef fishes were higher, mean abundance within each fish group was compared between SPA and reference reef pairs (n = 4) in the FKNMS. Paired t-tests assuming equal variance were conducted, after it was verified that for most reef fish parameters, variance was not significantly different between no-take and reference sites (Zar 1984). Simple linear regressions were developed separately for SPA and reference

PAGE 107

90 reefs. Then, variance ratio tests (Zar 1984) were used to test whether the strengths of relationships were stronger where (SPA or reference sites) reef fish abundances were highest. In the US Virgin Islands, mean abundance of each fish group was compared between MPA (n = 15) and references reefs (n = 7) using t-tests assuming equal variance, after it also was determined that variance was not significantly different for most reef fish parameters between reefs inside and outside VINP. A variety of multivariate methods available in PRIMER software (Clarke and Warwick 2001) were employed to analyze differences in the reef fish communities among individual reefs and among different reef systems (FKNMS and US Virgin Islands). For reef fish community analyses, I calculated Bray-Curtis similarity on log10-transformed abundances (a satisfactory coefficient for biological data on community structure; Clarke and Warwick 2001) at the trophic and family levels, and conducted analysis of similarities (ANOSIM) to test for differences in reef fish community structure between the FKNMS and US Virgin Islands. ANOSIM calculates a global R statistic that reflects the differences in variability between groups, compared to within groups (so R values are proportional to differences between the groups) and checks for significance of R using permutation tests (Clarke and Warwick 2001). To quantify the relative contribution of trophic and family categories to dissimilarities, I conducted a similarity breakdown using the SIMPER procedure, which provides similarity percentages with respect to contribution to average similarity within a group and average dissimilarity between groups (Clarke and Warwick 2001). Non-metric multi-dimensional scaling (MDS) was used to further analyze reef fish community structure and the landscape structure in both the FKNMS and US Virgin Islands. MDS can be used to generate plots

PAGE 108

91 in which the distance between points is proportional to their degree of dissimilarity, so closer points are more similar than points farther away (Clarke and Warwick 2001). MDS plots were calculated using mean abundances for reef fish community structure, and measures of rugosity, reef size and the areal coverage of the habitat types common to the FKNMS and US Virgin Islands for every reef were used for comparing landscape structure between systems. In MDS plots, reefs were classified in two manners. First, reefs were classified according to location (1 = USVI, 2 = FKNMS) to examine differences between systems. Second, reefs were classified according to the amount (hectares) of seagrass seagrass (1 = 0.5 ha 2 = 0.5.2 ha, 3 = 1.2 ha) within 500 meters of each study reef. This classification was also used for each of the other habitat types common to both systems, to verify that other habitat types did not result in a similar spatial patterning of reefs for trophic and family structure, as that found using seagrass. Results A total of 59,239 individual fishes representing 152 species were recorded during 316 point counts at the 17 study reefs in the Florida Keys National Marine Sanctuary, which compares to the 14,389 individual fishes recorded representing 118 species during 107 point counts at 22 study reefs in the US Virgin Islands during July and August 2002. As expected based on my findings in the US Virgin Islands, reef configuration was not a good predictor of fish assemblage structure in FKNMS. In fact, configuration was weakly associated with only 5 of 30 fish parameters in the FKNMS, and explained < 30 % of the variation of any single fish parameter (Table 4-3). Though reef size was marginally correlated with total fish abundance (R2 = 0.22), abundances of mobile fishes (R2 = 0.26), serranids (R2 = 0.27) and lutjanids (R2 = 0.24) (Table 4-3) in FKNMS, examination of Mallows Cp and AIC revealed that these model estimates were fairly

PAGE 109

92 unstable (e.g., total abundance Cp = 3.84, i = 4.3). Pooled datasets for the FKNMS and USVI confirmed these findings. Likewise, rugosity was a poor predictor of reef fish assemblage structure, as no fish parameter was significantly correlated with rugosity. The coefficient of variation for rugosity was rather low, although no relationship was detected after datasets of the FKNMS and US Virgin Islands were pooled to maximize the gradient of rugosity. As in the US Virgin Islands, reef context explained the greatest amount of the variation in reef fish assemblage structure, as 16 of 30 possible fish parameters were strongly associated with measures of reef context in FKNMS (Table 4-4). However, the particular measure of reef context associated with the majority of fish parameters differed between systems (Table 4-4). Contrary to my expectations, seagrass cover was not the best predictor of reef fishes in the FKNMS; rather the amount of pavement and reef habitat was strongly associated with several reef fish parameters (Table 4-4, Figure 4-2A-F). Within trophic guilds, abundances of herbivores (R2 = 0.52) and piscivores (R2 = 0.33) were associated with pavement (Table 4-4, Figure 4-2B-C). Within mobility guilds, resident fishes were correlated (R2 = 0.53) with reef habitat (Figure 4-2D), whereas mobile fishes were strongly correlated (R2 = 0.70) with reef, pavement and seagrass habitats (Table 4-4). Within taxonomic groupings, abundances of chaetodontids (R2 = 0.31), acanthurids (R2 = 0.31), and pomacentrids (R2 = 0.48) were positively associated with reef habitat (Table 4-4, Figure 4-2E-F). The landscape-scale measure of habitat patch diversity was a predictor of one reef fish parameter only; piscivore abundance was positively associated with patch diversity at the 250 m spatial extent only in the FKNMS (R2 = 0.35) and the US Virgin Islands (R2

PAGE 110

93 = 0.51) (Figure 4-3). Model selection techniques confirmed the stability of these relationships (e.g., Cp = 2, and i < 2), yet this was the only one that remained consistent across systems. Differences in reef fish community structure were not likely to account for the lack of consistency in reef fishhabitat relationships across systems. In fact, reef fish community structure was quite similar in the US Virgin Islands and FKNMS at the entire assemblage level (cumulative species richness per reef in USVI = 44.23, FKNMS = 68.53; mean species richness per point count in USVI = 22.42, FKNMS = 22.06; mean number of fishes per point count in USVI = 96.72, FKNMS = 179.46). Higher mean abundances of reef fishes in the FKNMS can be attributed to the tendency of several species (e.g., Abeduduf saxtalis) to aggregate around divers, possibly as a result of fish feeding. Reef fish community structure was significantly different between the US Virgin Islands and FKNMS at the trophic (one-way ANOSIM, global r = 0.16, p < 0.02) and family level (one-way ANOSIM, global r = 0.22, p < 0.01), though these differences were minor. These minor differences in reef fish community structure can be attributed to a 30 % difference in mean planktivore abundance and 26 % difference in herbivore abundance, with the means higher in the FKNMS for both parameters (Table 4-5). In addition, there was a 19 % difference in mean abundance of haemulids between sites (Table 4-5), with a slightly higher mean in the FKNMS compared to the USVI. In general, though, these differences were minor, thus fundamental differences in community structure do not preclude comparisons of reef fishhabitat relationships between systems.

PAGE 111

94 The spatial arrangement and composition of habitat patches differed between the FKNMS and US Virgin Islands. There were significant differences (df = 37, p < 0.0001) in the mean areal coverage of seagrass within 500 m of focal reefs (FKNMS = 12.10 ha; US Virgin Islands = 6.76 ha), with a higher coefficient of variation of seagrass at 500 m in the US Virgin Islands (65.0) compared to the FKNMS (12.10). Furthermore, in the FKNMS, seagrass comprises 41 % of the habitat within 500 m of study reefs, which contrasts with 19 % in the US Virgin Islands (Figure 4-4). At the scale of the entire system, 69 % of the mapped extent of the FKNMS consists of seagrass, whereas 26 % is hardbottom (which includes pavement, reef, and patch reef). Differences in the amount of pavement and reef within 500 m of the study reefs were less clear, as the absolute percentages of pavement habitat are quite similar between study systems (20 % versus 17 %, Figure 4-4). However, the relative amount of pavement (17 %) and reef (30 %) in the FKNMS was low relative to the amount of seagrass (41%) in the FKNMS (Figure 4-4). In addition, the mean amount (ha) of pavement at 500 m in FKNMS was significantly less (2.3 ha) than the US Virgin Islands (12.2 ha). The coefficient of variation for pavement was comparable between systems, except at the 250-meter scale, in which case, the FKNMS was slightly higher than the US Virgin Islands. Finally, the proportion of potentially essential fish habitat (seagrass, reef and pavement habitats within 500 m of study reefs) was higher in the FKNMS (96 %) than the USVI (55 %) (Figure 4-4). Comparisons of the landscape structure of the entire mapped systems revealed important differences between the FKNMS and US Virgin Islands. In general, the landscape structure of the FKNMS is more homogeneous (Figure 4-6C), as there is a distinct pattern of inshore seagrass that extends 6 km offshore, at which point spur

PAGE 112

95 and groove reef development typically occurs. MDS plots revealed that reefs in the FKNMS clustered together, with less variability in landscape structure than the USVI (Figure 4-6C), although there were three outlier reefs (Figure 4-6C). Conversely, reefs in the US Virgin Islands exhibited greater variability in landscape structure (Figure 4-6C), with reefs dispersed across the MDS plots (Figure 4-6C). Likewise, there is a less distinct zonation of reef formation in the US Virgin Islands compared to the FKNMS. The stress value of 0.07 for the MDS plot of landscape structure corresponds to a good ordination with no real prospect of misinterpretation (Clarke and Warwick 2001). Although there were not significant differences in fish community structure among systems (USVI and FKNMS), at the individual reef level, there was greater variation in reef fish assemblage structure in the US Virgin Islands compared to the FKNMS (Figure 4-6A-B). For example, in the FKNMS, both at the trophic and taxonomic levels, reefs were clustered tightly together in MDS plots (Figure 4-6A-B). Conversely, in the US Virgin Islands, there was greater variability at the trophic and taxonomic level, in reef fish assemblage structure (Figure 4-6A-B). There was also considerably more variability in landscape parameters in the US Virgin Islands, compared to the FKNMS (Figure 4-6C). Thus, variability in reef fish community structure appears greater in the more heterogeneous coral reef landscape of the USVI (Figure 4-6C), and lowest at reefs in the more homogeneous FKNMS coral reef landscape (Figure 4-6A-C). Stress values for the MDS plots of reef fish community structure were < 0.2 for trophic structure, which gives a potentially useful 2-dimensional picture and < 0.1 for family structure, which corresponds to a good ordination (Figure 4-6A-B).

PAGE 113

96 Reef fish community structure for individual reefs corresponded to differences in the presence or absence of seagrass (Figure 7A-B). Both at the trophic and family level, there was clear segregation between reefs with and without seagrass nearby (Figure 7A-B). Reefs with high to medium amounts of seagrass clustered together, regardless of location (US Virgin Islands or FKNMS), although reefs with moderate amounts (class = 2) of seagrass were less distinguishable from one another (Figure 4-7A-B) There was a tendency for reefs with moderate amounts of seagrass to be located closer to reefs with high amounts of seagrass (Figure 4-7A-B) Notably, when this classification was applied to the other common habitat types (e.g., pavement), there was no spatial pattern evident. There were few differences in reef fish assemblage structure inside SPAs and MPAs compared to fished reference reefs in either FKNMS or the US Virgin Islands (Table 4-6). Importantly, the abundances of targeted reef fishes (e.g., haemulids, lutjanids) were not higher inside protected areas. Actually, for several fish parameters, abundances were greater at the fished reference reefs (Table 4-6). The lack of significant differences in reef fish abundance for most parameters of interest inhibited our ability to test whether reef fishhabitat relationships are higher where reef fish abundances are higher. Variance ratio tests revealed no significant differences in the strength of associations at reefs with higher abundances compared to those with lower fish abundances. The models that best explained the relationships of various reef fish parameters and seagrass area for the pooled data (FKNMS and US Virgin Islands) were curvilinear (Figure 4-5A-D). Forty-five percent of the variation in haemulid abundance, and 30 % of the variation in lutjanid abundance was explained by a quadratic relationship with

PAGE 114

97 seagrass (Figure 4-5A-D). For most relationships, the quadratic reaches an asymptote, at which point additional seagrass does not further increase each fish parameter (Figure 4-5A-D). Discussion Reef configuration measures explained little of the variation in reef fish assemblage structure in either the FKNMS or the US Virgin Islands. These findings contrast with reef fish research conducted at a small spatial scale (1 meters), where reef size (Molles 1978, Bohnsack et al. 1994), reef patchiness (Acosta and Robertson 2002), reef connectivity (contiguous versus patch reefs; Ault and Johnson 1998a, 1998b) and reef shape (Shulman 1984, Eristhee and Oxenford 2001) measures were shown to strongly influence reef fish assemblage structure (e.g., total abundance, species richness). A plausible explanation for the inability of configuration to predict fish assemblage structure in the FKNMS and US Virgin Islands is that the scale of analyses (10s to 100s of meters) is above the minimum scale where configuration effects are germane to reef fishes, and therefore these effects were not detectable. In terrestrial systems, configuration effects are only detectable and/or operable below some minimum size threshold (McGarigal and McComb 1995, Fahrig 1997, 1998, Trzcinski et al. 1999). For example, Fahrig (1997, 1998) found that when the percentage of habitat in a landscape exceeded 20%, species persistence was virtually assured, regardless of configuration. Likewise, Andrn (1994) showed that patch isolation only becomes important when the percentage of critical habitat decreases below a 20% threshold. I hypothesize that my study reefs are above the minimum size (0.2 ha) where configuration effects are easily detectable, and that below some minimum size, which my reefs exceed, the influence of reef size and shape may become more important.

PAGE 115

98 While the speciesarea relationships detected for insular habitats (Preston 1962, MacArthur and Wilson 1967) provides a firm basis for predicting that fish abundance and species richness will be greater on large reef patches compared to smaller reef patches, reef size explained little of the variation in reef fish assemblage structure in the FKNMS. Conversely, reef size was a strong predictor of cumulative species richness (R2 = 0.43, p < 0.001) in the US Virgin Islands. These contradictory findings support terrestrial researchers who concluded that speciesarea relationships were too ambiguous to indicate whether two reserves, each half the size of a larger reserve, will contain more species than the single one (Simberloff and Abele 1976, Diamond 1975). The contentious SLOSS (single large or several small) reserve design debate (Quinn and Harrison 1988), asks whether a single large reserve or several small reserves of equal total area contain more species? While there are convincing management arguments for single large reserves (Terborgh 1976), my study suggests that small reefs may provide conservation benefits comparable to a single large reef, at this spatial scale. Furthermore, some of the smallest, yet most heterogeneous reefs had the highest values for several reef fish parameters (e.g., North North Dry Rocks had highest species richness), which may indicate the importance of microhabitat influences on species assemblages (Gilpin and Diamond 1980). The issue of SLOSS may not have a simple, clear solution, and may be a function of many factors including habitat heterogeneity. Neither the fine-scale measures of habitat heterogeneity (rugosity) nor landscape-scale measures of habitat diversity were useful in explaining variation in reef fish assemblage structure, which contradicts ecological theory that predicts species diversity increases with habitat diversity (Ricklefs 1973, Roughgarden 1996), yet supports my

PAGE 116

99 earlier findings in the US Virgin Islands (Grober-Dunsmore et al. 2004a, 2004b, 2004c). The only exception was patch diversity, which was positively associated with piscivore abundance; a relationship that may be facilitated by the piscivore behavior of searching the reef edge for prey (Sweatman and Robertson 1994). The inability of habitat diversity to explain reef fish assemblage structure is also contrary to findings from a number of fine-scale coral reef studies (1s of meters), which demonstrated a positive relationship with reef fish diversity and abundance and habitat heterogeneity, measured by parameters such as spatial complexity (Hixon and Beets 1989), living coral (Reese 1981), coral reef zonation (Friedlander and Parrish 1998), and reef surface area (Gladfelter et al. 1980). Contrary to recent experimental evidence (Armbrecht et al. 2004), it does not appear that diversity per se at one level creates conditions that promote diversity at another level in coral reef systems, therefore one key finding is that habitat diversity may not scale up at the landscape level (i.e. 100s of meters scale) as a predictor of reef fish assemblage structure. Habitat diversity (at fine or landscape spatial scales) may fail to explain reef fish assemblage structure for several reasons. At the landscape scale, individual habitat types appear more important than diversity per se, as revealed in this and previous studies (Wiens 2002, Grober-Dunsmore et al. 2004a, 2004b, 2004c, Lindenmayer and Hobbs 2004, Jeffrey 2004, though see Cushman and McGarigal 2003). The fine-scale measure of rugosity may not have been capable of predicting reef fish community structure in this study because sampling was constrained to reef habitat only. Historically, rugosity was a good predictor of fish assemblage structure because it indicated topographically-complex habitat (Friedlander and Parrish 1998, Luckhurst and Luckhurst 1978), however, benthic

PAGE 117

100 habitat maps are likely to be capable of assuming this function. Because rugosity is a costly in-water measure, it will be important to explore remotely-obtained measures of rugosity (e.g., LIDAR, side-scan sonar) that may be readily available for measuring within-reef heterogeneity at a scale appropriate for marine resource managers. Therefore, landscape-scale measures of habitat diversity may not prove useful in discerning the value of individual reefs as candidates for protection. Context clearly exerts a strong influence on reef fish assemblage structure. Reef contextreef fish relationships that passed model selection criterion were consistent with our knowledge of the ecology of each particular fish group, as was recently demonstrated in the US Virgin Islands (Kendall et al. 2003, Grober-Dunsmore et al. 2004c), Puerto Rico (Christensen et al. 2003), Belize (Mumby 2004), Colombia (Appeldoorn et al. 2003) and the Florida Keys (Jeffrey 2004). For example, total fish abundance was associated with the areal coverage of pavement, perhaps because the dominant fish in the assemblage (labrids) tend to roam across this low relief habitat. Similarly, grazers (herbivores) were also associated with pavement, where they often forage. Not surprising is the fact that identified resident reef species were strongly associated with the areal coverage of reef habitat. Thus, different fish groups do not appear to perceive the landscape the same, thus it will be necessary to focus on organisms (Wiens 1989, Pearson 1993) or functional groups (as in this study) in the future to understand how landscape structure affects marine systems. This study contributes to a growing body of work that points to the importance of reef context (Appeldoorn et al. 2003, Christensen et al. 2003, Kendall et al. 2003, Mumby et al. 2004), and builds on recent research by replicating the study spatially, demonstrating the effects of reef context to the entire reef fish community

PAGE 118

101 (not just select parameters), and providing quantified measures of reef context, rather than simple presence or absence for multiple habitat types. The processes that structure reef fish communities appear modified by variation in landscape structure. While measures of reef context consistently explained the greatest variation in reef fish assemblage structure in both study systems, the individual habitat types driving the relationships varied. For example, seagrass was strongly associated with many fish parameters in the US Virgin Islands (e.g., abundances within MIFs, grunts, snappers), yet pavement and reef habitats were associated with the majority of reef fish parameters in the FKNMS. Differences in the landscape structure of the two study systems may help explain this disparity in findings. The distribution of seagrass habitat is highly variable in the US Virgin Islands; seagrass is the background matrix of several embayments, yet absent completely in others. In this landscape, seagrass is a strong predictor of many commercially and recreationally-important reef fishes in the US Virgin Islands. Conversely, the distribution of seagrass is less variable, and the dominant habitat type around study reefs in the FKNMS, but not a predictor of these same commercially and recreationally-important fishes. Thus, seagrass could be used as a reliable predictor of these fish groups only in the landscape where seagrass was less available. These findings have important consequences for MPA design, suggesting that habitat estimates for predicting reef fishes derived from one system, may not be broadly applicable in another (Lindenmayer et al. 2003). It appears that applying a particular metric (e.g., areal coverage of seagrass) simplistically may be dangerous, if the regional ecology and specific landscape is not considered explicitly (Noss 1983). Rather, MPA

PAGE 119

102 design may be more complicated, and require intimate knowledge of the spatial arrangement and composition of the habitat patches in each system. Perhaps there is some critical threshold of habitat, at which point other habitat types structure reef fish communities. Combined reef fish datasets (FKNMS and US Virgin Islands) revealed a significant curvilinear relationship; with increasing seagrass, several reef fish parameters increased until an asymptote was reached. While the critical threshold hypothesis purports that a nonlinear decrease in richness or abundance occurs below some level of habitat loss (Summerville and Crist 2001), perhaps at some threshold of seagrass habitat gain, seagrass no longer has pronounced (or detectable) influence on the reef fish community. If such a critical threshold exists, it appears to occur between 20-30 % of the total area of habitat, since at this amount of seagrass, reefs in the FKNMS and US Virgin Islands converged in simple linear regression models and became difficult to distinguish from other. The inability to detect a relationship with seagrass within the FKNMS system on its own, may be explained by the abundance of seagrass immediately beyond the 250-meter spatial scale. The critical threshold hypothesis is further supported by the MDS plots of reef fish community structure at the trophic and family level, which reveal a clear spatial segregation of reefs (FKNMS and US Virgin Islands) with and without seagrass. Regardless of location, i.e. USVI or FKNMS, reefs with large amounts of seagrass were more similar to one another. Reefs with moderate amounts of seagrass most resembled reefs with large amounts of seagrass, suggesting that modest amounts of seagrass may yield considerable influence on reef fish community structure. The absence of a spatial pattern when reefs were classified by other habitat types confirmed the importance of

PAGE 120

103 seagrass in particular, and eliminated the possibility that other habitat types were responsible for observed spatial patterns. Such thresholds in the response of organisms to landscape characteristics is widely recognized in terrestrial systems (Wiens 2002), and can be useful in forecasting threshold points (e.g., habitat loss or fragmentation) given precise rules for movement and dispersal (With and Crist 1995). While much remains to be learned about the importance of possible habitat thresholds in coral reef systems, combining datasets from the US Virgin Islands and FKNMS (though problematic for some purposes), allowed a potentially valuable insight that would not have been possible with data only from an individual system. The relative amount of an individual habitat type within a given system may be as important as the absolute amount in structuring reef fish communities. In the USVI, pavement and reef habitats were common, but were poor predictors of most fish parameters. In the FKNMS these same habitat types were relatively less common, but explained significantly more of the observed variation in the reef fish parameters of interest. These findings are congruent with those from an experimental shelter manipulation where groupers responded to shelter limitation when reefs were surrounded by sand, but failed to do so when reefs were surrounded by pavement habitat (Hart 2002). Thus, my initial assumption (stated earlier in the text) that the processes that gave rise to the relationships in the US Virgin Islands are not modified by variation in the landscape structure, appears incorrect. Rather, the critical element driving differences in community structure in one landscape may depend upon the specific landscape attributes of that landscape.

PAGE 121

104 For instance, the relative heterogeneity or homogeneity of a particular coral reef landscape may determine the importance of landscape attributes in structuring communities. Reef fish community structure was more variable in the more heterogeneous landscape of the US Virgin Islands, whereas fish community structure was less variable in the more homogeneous landscape of the FKNMS. The FKNMS reef tract (totaling 591,422 ha) has a consistent pattern of 6 kilometers of inshore seagrass, interspersed with small patch reefs, until reaching the offshore drowned spur and groove zone to the east. This pattern extends the 285-kilometer length of FKNMS. Perhaps the relative homogeneity of the FKNMS coral reef landscape inhibited our ability to evaluate the importance of landscape structure on reef associated fishes in the FKNMS. The US Virgin Islands systems is smaller, and extends from the shoreline of St. John to the British Virgin Islands to the north, and to deep water to the south; totaling 49,588 ha. The spatial patterning of habitat patches in the US Virgin Islands is more heterogeneous, with patches of reef and seagrass interspersed among other habitat patches that are similar in areal coverage. Some areas of the island have distinct enclosed embayments, whereas other areas are more exposed. Thus, the range of variability in landscape parameters may be greater within the US Virgin Islands system, and as a consequence of this landscape heterogeneity, the reef fish community structure may be more variable. In spite of the greater variability of reefs (with regard to fish community structure at the trophic and family levels), the overall fish community structure between systems was remarkably similar as demonstrated by ANOSIM. Such consistency in structure has been found across regions and reef types (Bohnsack and Talbot 1980, Bohnsack et al. 1999),

PAGE 122

105 yet served as a control by eliminating the possibility that differences in reef fish community structure precluded the ability to make comparisons across systems. Reef fish abundances of family and trophic fish groups were not higher inside protected areas relative to reference (fished) areas, which greatly inhibited the ability to test whether reef fishhabitat relationships are stronger where fish densities are high. In fact, the abundances of several fish groups were higher at open access reference reefs. While abundances of select target species (e.g., Ocyrus chyrsyus) may be increasing inside SPAs in the FKNMS (see Ault et al. 1998, Bohnsack et al. 1999), protection measures may be too recent to produce positive community-level effects. The limited time since establishment of FKNMS SPAs (six years) may partially explain the absence of protection effects, since Looe Key, which has the longest protection history (> 20 years), had the highest mean abundance of adult and juvenile groupers, adult grunts, and adult snappers. Our findings suggest that the effects of protecting heavily-exploited fishes may only become apparent after six or more years of no-take status and enforcement. In fact, some have suggested it may take as long as 50 years (Russ and Alcala 2004). With regard to the VINP, it is not surprising that reef fish abundances were not higher inside park boundaries, since the primary fishing gear (trap fishing) is still allowed. Reef fish populations may continue to decline in the US Virgin Islands, unless strict fishing regulations are instituted and enforced (Rogers and Beets 2001). Studies to determine whether reef fish density influences our ability to detect relationships with landscape features remain to be conducted. Conclusions This study provides considerable insight into the generalities of applying landscape ecology principles to coral reef ecosystems. Most importantly, reef context clearly

PAGE 123

106 influences reef fish assemblage structure, and I recommend that selection of reef patches for conservation should be based on how well a patch relates or links to other patches in the landscape. However, extrapolation of patterns observed in local studies may not hold regionally (Flather and Sauer 1996) or across systems. Furthermore, considerable risk is often associated with broad application of conservation strategies (Kareiva 1987), and as a consequence, conservation policies must consider the natural variation in landscape structure, since overlooking system differences may lead to errors arising from simple landscape perceptions of patterns (Kotliar and Wiens 1990). It is clear from our findings that the same landscape can be perceived quite differently by different species (Lindenmayer et al. 2003, Westphal et al. 2003). Thus, a functional MPA design must consider the autecology of those species targeted for management. Finally, while this inductive approach is less conclusive than the hypothetico-deductive method based on experimental manipulations (Underwood 1997), it may be the only practical means to ascertain the influence of factors acting at large spatial scales.

PAGE 124

107 Figure 4-1. Location of the 17 study reefs sampled in 2003 in the Florida Keys National Marine Sanctuary and the 22 study reefs sampled in 2002 in St. John, U.S. Virgin Islands.

PAGE 125

108 Total Abundance R2 = 0.47p < 0.011.922.12.22.32.42.500.511.522.53Abundance log (x+1)A) Resident Abundance R2 = 0.53p < 0.011.41.51.61.71.81.922.12.22.30.511.522.5 3 D) Herbivore AbundanceR2 = 0.52p < 0.011.51.61.71.81.922.12.200.511.522.53Abundance log (x+1)B) Chaetodontid Abundance R2 = 0.31p < 0.0100.10.20.30.40.50.511.522.53E) Piscivore AbundanceR2 = 0.33p < 0.020.10.20.30.40.50.600.511.522.53 Acanthurid AbundanceR2 = 0.31p < 0.020.70.80.911.10.511.522.5Reef (ha) log x +1F) 3 Abundance log (x+1)C) Pavement (ha) log x +1 Figure 4-2. Simple linear regression results of the effects of reef context (the areal coverage of pavement and reef habitat log10 (x+1)) at the 100-meter spatial scale on mean abundance of various reef fish parameters at the 17 study reefs sampled in 2003 in the FKNMS.

PAGE 126

109 USVIR2 = 0.510.91 FKNMSR2 = 0.3500.10.20.30.40.50.60.70.811.21.41.61.822.22.42.6Patch diversity 250 mPiscivore Abundance FKNMS UVSI Figure 4-3. Simple linear regression results of the only relationship that remained consistent, and significant, across systems (the relationship of patch diversity and piscivore abundances) in the FKNMS and US Virgin Islands.

PAGE 127

110 FKNMS Seagrass41%Pavement17%Sand4%Reef30%Patch reef8% USVIDeep unknown19%Pavement20%Reef12%Bedrock4%Seagrass19%Sand22%Macroalgae4% Figure 4-4. Comparison of the relative proportion of mapped habitat classes within 500 m of the study reefs in the FKNMS and US Virgin Islands.

PAGE 128

111 R2 = 0.45p < 0.000100.20.40.60.811.21.41.61.8200.511.5 Haemulid Abundance FKNMS UVSI A ) R2 = 0.30p < 0.001500.20.40.60.811.21.41.600.511.52Seagrass 500 mLutjanid Abundance FKNMS UVSIB) R2 = 0.33p < 0.000711.21.41.61.8200.511.5Seagrass 250m Mobile invert feeder abundance FKNMS UVSIC) R2 = 0.35p < 0.000115202530354000.511.5Seagrass 250 m Mobile invert feeder Richness FKNMS UVSID ) Figure 4-5. Reef fish abundanceseagrass relationships for the US Virgin Islands and the FKNMS (n = 39). The amount of seagrass (ha) log10x + 1 is on the x-axis. The spatial extent (250 m and 500 m) with the strongest relationship is shown for each fish parameter. Mean abundance values were log10x + 1-transformed.

PAGE 129

112 Stress = 0.10A) Stress = 0.14B) S Stress = 0.10C Figure 4-6. Multidimensional scaling plots of the (A) trophic and (B) family structure of reef fish communities and the (C) landscape structure of the US Virgin Islands (squares) and FKNMS (triangles).

PAGE 130

113 Stress = 0.14A) Stress = 0.08B) Figure 4-7. MDS plots of the (A) trophic and (B) family level reef fish community structure for reefs in the US Virgin Islands and FKNMS, with reefs classified according to high (circles), moderate (open triangles) and low seagrass (upside down closed triangles) areal coverages.

PAGE 131

114 Table 4-1. Reef fish assemblage parameters (n = 30) used as dependent variables in statistical analyses Entire assemblage level parameters Trophic guilds Mobility guilds Taxonomic groupings Cumulative species richness Herbivores (J & A) Resident Acanthurid (J & A) Mean species richness Mobile invertebrate feeders (J & A) Mobile Serranid (J & A) Total abundance Omnivores (J & A) Transient Haemulid (J & A) Piscivores (J & A) Lutjanid (J & A) Planktivores Pomacanthid Sessile invertebrate feeders Scarid (J & A) Holocentrid Labrid Chaetodontid Note: Fish groups are not always mutually exclusive. Hypoplectrus species were not included in Serranid grouping. For each trophic guild and taxonomic grouping, reef fish parameters were further subdivided into juvenile and adult components, where indicated (J = juvenile, A = adult).

PAGE 132

115 Table 4-2. Study reef name, protection status and reef fish sampling effort for FKNMS (May 2003) and St John, USVI (July-August 2002). FKNMS (n = 17) Protection status No. of samples USVI (n = 22) Protection status No. of samples Carysfort SPA 18 Caneel MPA 6 Davis SPA 16 Cruz1 MPA 4 Dry Rocks SPA 13 Donkey MPA 6 E Dry Rocks SPA 21 Gibneydeep MPA 4 Eastern Sambo SPA 12 Gibneylinear MPA 8 Elbow SPA 13 Hansen Reference 6 French SPA 22 Hawkshallow MPA 6 Grecian Rocks SPA 14 Henley Reference 4 Little Grecian Reference 14 LindPoint MPA 6 Looe Key SPA 31 Marys MPA 2 Molasses SPA 20 Mennebec MPA 5 North Dry Rocks Reference 16 Peterbay MPA 4 NN Dry Rocks Reference 15 Peterlinear MPA 4 Pickles Reference 15 Peterone MPA 4 Sand Key SPA 17 Peterthree MPA 4 Sombrero SPA 14 Princess Reference 4 Western Sambos SPA 20 Rendezvous Reference 10 SabaE Reference 4 SabaW Reference 2 Waterlemon MPA 8 WhistlingCay MPA 4 Windswept MPA 12 Note: In the FKNMS, SPA refers to special protected areas, which are no-take areas, and in the US Virgin Islands, MPA refers to reefs within the boundaries of Virgin Islands National Park, where fishing is permitted, except with spear guns. Reference reefs refer to areas outside the boundaries of SPAs in FKNMS and outside the boundaries of Virgin Islands National Park in the US Virgin Islands.

PAGE 133

116 Table 4-3. Stepwise regression results indicating the relationship of landscape configuration and reef fish assemblage structure for the FKNMS and US Virgin Islands, with the R-square, associated p-value, and explanatory habitat variable. Reef fish parameter FKNMS R2 FKNMSp-value Habitat variable USVIR2 USVSI p-value Habitat variable Total Abundance 22% 0.06 Reef size --None Mobile 26% 0.04 Reef size --None Serranid 27% 0.03 Reef size --None Pomacanthid 30% 0.02 No. of patches --None Lutjanid 24% 0.03 Reef size --None Notes: Relationships for the US Virgin Islands are presented, only if significant relationships were found in FKNMS, though all reef fish parameters were explored. Input habitat variables were reef size, number of habitat patches within 100 m and perimeter to area ratio of each study reef. The -symbol indicates where no relationship was detected.

PAGE 134

117 Table 4-4. Stepwise multiple regression results of the relationships of reef context and reef fish assemblage structure for FKNMS and US Virgin Islands, with the R-square, p-value, and explanatory variable by location. Fish parameter FKNMSR2 FKNMS p-value Habitat variable USVI R2 USVI p-value Habitat variable Entire Total abundance 47 % 0.0025 Pavement 17 % 0.0500 Seagrass Total richness ------47 % 0.0004 Seagrass Mean richness 57 % 0.0013 0.0220 0.0086 Pavement Seagrass Reef 24 % 0.0200 Seagrass Trophic Herbivore 52 % 0.0011 Pavement ------A MIF J MIF --24 % --0.0400 --Patch reef 32 % 0.0050 Seagrass J Omnivore 32 % 0.0100 Seagrass 14 % 0.0300 Seagrass Piscivore 33 % 0.0170 Pavement 16 % 0.0300 Reef Planktivore 50 % 0.0017 Reef ------SIF 56 % 0.0047 0.0015 Patch reefReef ------------Mobility Resident 53 % 0.0009 Reef ------Mobile 70 % 0.0004 0.0024 0.0429 Patch reefPavementSeagrass 20 % 0.0400 Seagrass Family Acanthurid 31 % 0.0200 Reef ------Chaetodon 31 % 0.0210 Reef ------Haemulid --48 % 0.0004 Seagrass Holocentrid --20 % 0.0300 Seagrass Labrid 51 % 0.0013 Pavement ------Pomacentrid 48 % 0.0020 Reef 34 % 0.0400 Pavement A Lutjanid ------39 % 0.0052 Seagrass Reef A Scarid J Scarid 62 % 70 % 0.0003 0.0028 0.0002 0.0067 Patch reef Pavement Seagrass Patch reef ------Note: Relationships that are significant (p < 0.05) following Sequential Dunn-Sidak Bonferroni-corrections are presented. The --symbol indicates where no relationship was detected.

PAGE 135

118 Table 4-5. Average dissimilarity in reef fish community structure at the trophic and family level in the FKNMS and US Virgin Islands, and the relative contribution of each trophic and family category to community dissimilarities using SIMPER analyses on standardized data. Dissimilarity Relative Contribution Trophic 29.19 Planktivore 30.79 Herbi 26.06 MIF 21.63 Omni 12.38 Pisci 6.29 SIF 1.06 Family 38.56 Haemulid 18.45 Scarid17.23 Pomacentrid 16.58 Acanthurid 14.63 Labrid 13.65 Lutjanid5.51 Note: Plankti = planktivore; Herbi = herbivore; MIF = mobile invertebrate feeders; Omni = omnivore; Pisci = piscivore; SIF = Sessile invertebrate feeders; Haem = Haemulid

PAGE 136

119 Table 4-6. T-test comparison results to test for significant differences in abundance of the 30 reef fish parameters in the FKNMS (SPA and reference sites n = 8 reefs) and the US Virgin Islands (MPA and reference sites n = 22 reefs). Fish parameter FKNMS USV Virgin Islands Cumulative species richness 0.63 0.630 Mean species richness 0.29 (Ref) 0.001 Total abundance 0.44 (Ref) 0.06 J Herbivores A Herbivores 0.59 (SPA) 0.01 0.76 (Ref) 0.06 J MIFs A MIFs 0.37 0.61 0.31 (Ref) 0.01 J Omnivores A Omnivores 0.86 0.58 0.11 (Ref) 0.07 J Piscivores A Piscivores (Ref) 0.07 (Ref) 0.09 (Ref) 0.07 (Ref) 0.07 Planktivores 0.60 (Ref) 0.07 SIFs 0.14 0.48 Resident 0.19 (Ref) 0.08 Mobile 0.27 (Ref) 0.06 Transient 0.28 (Ref) 0.07 J Acanthurids A Acanthurids 0.96 0.11 0.32 0.13 J Serranids A Serranids 0.26 0.35 (MPA) 0.05 0.21 J Haemulids A Haemulids 0.81 0.79 (Ref) 0.10 (Ref) 0.008 J Lutjanids A Lutjanids 0.58 0.34 0.19 0.15 Pomacanthids 0.59 0.48 J Scarids A Scarids 0.29 0.20 0.99 (Ref) 0.08 Holocentrids 0.99 (Ref) 0.03 Labrids 0.75 (Ref) 0.01 Chaetodontids 0.26 0.35 Note: A = adult, J = juvenile

PAGE 137

CHAPTER 5 SUMMARY Maintenance of biodiversity (the abundance, variety, and genetic makeup of native animals and plants) requires a landscape perspective (Franklin 1993) that complements population, community, and ecosystem considerations. A landscape ecology perspective is compellingly distinct from an ecosystem approach since functional landscapes typically consist of multiple ecosystems that are spatially differentiated but still interact through energy flow and ecological processes (Bissonette and Storch 2003). While an ecosystem approach is an important advance over single species management (which has dominated fisheries management approaches for decades), loss of biodiversity and sustainable populations of targeted species will continue unless conservation of functioning landscapes and their associated spatial processes is addressed (Bissonette and Storch 2003). Conservation initiatives, such as marine protected areas design, that neglect the ecological and evolutionary context of a landscape and their component ecosystems and species are likely to fail (Bissonette and Storch 2003). Though these lanscape principles were derived in terrestrial systems, it is reasonable to assume that they may similarly apply to marine systems. Marine resource managers and scientists have recently recognized the need for large-scale, ecosystem-based approaches to enhance the conservation of marine biodiversity (Allison et al. 1998, Murray et al. 1999, Carr et al. 2003) and to manage sustainable fisheries (Pikitch et al. 2004). To address this need and because so little work has been previously conducted at a landscape-scale in marine systems, a hypothetico120

PAGE 138

121 deductive study approach (Peters 1991) was developed as a starting point toward applying terrestrial landscape ecology principles in tropical marine systems. Coral reef ecosystems were used as the study system, as they exist as a mosaic of interacting habitat patches including reef, hardbottom, seagrass, and sand. The objectives of this dissertation study were to: 1) explore the ability of landscape metrics to quantify and characterize the spatial arrangement of habitat patches in coral reef ecosystems, 2) explore relationships between reef fish assemblage structure and these metrics of landscape structure to develop testable hypotheses, 3) design a new study to test these hypotheses to further understand the relationship of reef fish assemblage structure to landscape structure of coral reef environments, and 4) to test the generality of findings from previous studies by repeating the studies temporally and spatially. Three separate studies (Chapter 2, Chapter 3, and Chapter 4) were conducted over multiple years at two locations in the Caribbean (the US Virgin Islands and the Florida Keys National Marine Sanctuary). Empirical data of the reef fish community and landscape structures of 59 study reefs (representing reefs of varying depth, types, and locations) formed the basis of this dissertation. To accomplish the first objective (Chapter 2), a suite of fourteen patch and landscape-level metrics (that have proven useful in describing terrestrial environments) were calculated using geographic information systems (GIS) tools and digitized benthic habitat maps for the study reef landscapes at four spatial extents (100 m, 250 m, 500 m, 1 km). The ability of these metrics to describe and distinguish between different coral reef landscapes was explored by examining simple summary statistics. Furthermore, Pearson-product moment correlations and principal component analyses were applied to reduce

PAGE 139

122 the landscape metrics into a more parsimonious set of variables that represent the information in the original benthic datasets. Examination of principal component plots at the four spatial extents revealed that the 250 m and 500 m spatial extents were the most appropriate scales for describing individual reef landscapes, in large part because beyond this extent, the total area of deep unknown increased considerably. These complex indices, however, were difficult to interpret because loadings were distributed across multiple variables. Thus, I concluded that individual habitat features, which were divided into two logical categories (configuration and context) were most valuable in characterizing and quantifying the landscape structure of the coral reef environments (e.g., size, shape, and context). Because the creation of new hypotheses is one of the most obscure and demanding aspects of science (Peters 1991), an existing dataset of the reef fish assemblage structure and landscape structure of 20 reefs from St. John, US Virgin Islands was used to accomplish objective two (to explore relationships to generate testable hypotheses). I found, contrary to findings from large-scale studies in terrestrial systems (Rafe et al. 1985, Rosenzweig 1995, Ricklefs and Lovette 1999) and coral reef fish studies conducted at a small (110 m) spatial scale (Bell and Galzin 1984, Hixon and Beets 1989, Friedlander and Parrish 1998), that measures of habitat heterogeneity (e.g., habitat diversity) do not appear to explain much of the variation in reef fish assemblage structure (Grober-Dunsmore et al. 2004a). Furthermore, the complex principal components were similarly not useful in explaining variation in reef fish assemblage structure. These findings contrast with those from terrestrial systems, where composite measures such as habitat diversity and principal components have successfully predicted hotspots of

PAGE 140

123 diversity and abundance (Rosenzweig 1995, Ricklefs and Lovette 1999, Turner et al. 2001). Interestingly, habitat diversity was often negatively associated with various fish assemblage parameters, thus I concluded that single habitat types may be valuable in explaining variation in reef fish assemblage structure. In Chapter 2, I found that landscape-scale measures of the areal coverage of different habitat types played an important role in determining the distribution of reef fishes in the US Virgin Islands. Reef context, as measured by the areal coverage of individual habitats, clearly influenced patterns of abundance and species richness at the entire assemblage level (e.g., total fish abundance, species richness) as well as within trophic and mobility guilds and taxonomic groups of fishes. One habitat type in particular, seagrass, explained a large amount of the variation in abundances of several exploited reef fishes (e.g., haemulids, lutjanids). Given these results, I generated a set of the most biologically-relevant hypotheses. I then identified which hypotheses were robust using model selection techniques to ensure that insights into habitat associations do not continue to be ambiguous and have poor explanatory power (Bissonette and Storch 2003). Hypotheses that met model selection criteria were used to design a new experiment to accomplish objective 3. Chapter 3 was specifically designed based on exploratory findings from the previous study to test the importance of seagrass habitat surrounding study reefs on reef fish assemblage structure parameters of interest to resource managers. The reef fish parameters of interest include entire assemblage level parameters (e.g., cumulative species richness) and abundance and species richness within trophic and mobility guilds and within taxonomic groups of fishes. As expected, landscape-scale habitat associations

PAGE 141

124 for several reef fish groups were clearly evident. Consistent with predictions, entire assemblage level parameters and fish abundances and species richness within mobile invertebrate feeders, haemulids, lutjanids, and epinephelids were significantly higher at those reefs with seagrass habitat within 1 kilometer, and these fish parameters increased with increasing areal coverage of seagrass. Consequently, the areal coverage of seagrass habitat proximal to individual reef patches may be used to successfully predict which reefs have high abundances of commercially and recreationally-important species and diverse reef fish assemblages. Although the study design of Chapter 3 is not able to identify the processes driving relationships, it provides strong evidence that functionally-linked marine landscapes contribute to increased species richness and abundances of several important fish groups. While several recent studies provide evidence of the importance of reef context (Nagelkerken et al. 2000, 2002, Appeldoorn et al. 2003, Kendall et al. 2003, Dorenbosch et al. 2004, Mumby et al. 2004), the study in Chapter 3 builds on existing research in several important ways, thereby contributing to our understanding of how landscape ecology principles apply in tropical marine environments. First, this study addresses the effects of reef context on fish density, rather than on simple absence or presence of individual species. Second, this study differentiates between soft-bottom (e.g., seagrass, sand) and hardbottom (reef, pavement) habitats. Third, measures of each habitat type were quantified using digital habitat maps and GIS, rather than using estimates derived using non-digital habitat maps and a planimeter. Fourth, considerable groundtruthing of the neighboring habitat patches was conducted to confirm habitat measures. Fifth, confounding effects of previous studies (i.e. reef size, location, and other habitat types)

PAGE 142

125 were minimized. Sixth, relationships were examined at the community level, within trophic guilds, within taxonomic groups, and within mobility guilds, thereby providing a comprehensive functional perspective of the response of various dimensions of the fish community to reef context. Finally, no previous study to our knowledge was designed explicitly to test the influence of reef context (specifically seagrass habitat) to a number of reef fish parameters of importance to resource managers. Therefore, this work represents an important contribution to our understanding of how landsacpe features influence reef fishes. Because extrapolation of patterns observed in one landscape may not hold regionally or across systems (Flather and Sauer 1996), and because the coral reefs of the US Virgin Islands have been exploited from fishing pressure and suffered hurricane damage (Rogers et al. 1991), I designed a separate study (Chapter 4, objective 4) in the Florida Keys National Marine Sanctuary (FKNMS) to determine the generality of reef fishhabitat relationships detected in the US Virgin Islands. Measures of reef context consistently explained the greatest amoung of variation for most reef fish parameters compared to measures of reef configuration, complex indices (e.g., habitat diversity) and fine-scale measures of habitat (rugosity) in both the US Virgin Islands and the FKNMS. The individual habitat measure of reef context driving the relationships (e.g., areal extent of seagrass or reef), however, varied between the US Virgin Islands and FKNMS. These findings suggest that the processes that structure reef fish communities appear modified by variation in landscape structure. For example, seagrass was strongly associated with many fish parameters in the US Virgin Islands (e.g., abundances of haemulids and lutjanids), yet pavement and reef habitats were associated with the

PAGE 143

126 majority of reef fish parameters in the FKNMS. Differences in the landscape structure of the two study systems may help explain this disparity. The distribution of seagrass habitat is highly variable in the US Virgin Islands; seagrass is the background matrix of several embayments, yet it is absent completely in others. Seagrass also comprises a relatively small proportion of the total habitat within the landscape, yet seagrass is a strong predictor of many commercially and recreationally-important reef fishes. Conversely, around study reefs in the FKNMS, the distribution of seagrass is less variable, and it is the dominant habitat type, however, it is not a predictor of these same commercially and recreationally-important fishes. Thus, seagrass was a reliable predictor of these reef fishes only in the landscape where seagrass habitat was relatively less available. These findings have important consequences for MPA design, suggesting that habitat estimates for predicting reef fishes derived from one system may not be broadly applicable in another (Lindenmayer et al. 2003). It appears that applying a particular metric (e.g., areal coverage of seagrass) simplistically may be dangerous if the regional ecology and specific landscape is not considered explicitly (Noss 1983). Perhaps there is some critical threshold of habitat at which point other habitat types structure reef fish communities. When datasets from the FKNMS and US Virgin Islands were combined, the best model fit was a curvilinear relationship; as seagrass increased, several reef fish parameters increased until an asymptote was reached. While the critical threshold hypothesis purports that a nonlinear decrease in richness or abundance occurs below some level of habitat loss (Summerville and Crist 2001), perhaps at some threshold of seagrass habitat gain, seagrass no longer has pronounced (or detectable) influence on the reef fish community. If such a critical threshold exists, it appears to occur between

PAGE 144

127 20-30 % of the total area of habitat, since at this areal coverage of seagrass, reef fish parameters from the FKNMS and US Virgin Islands begin to converge for most fish groups. The inability to detect a relationship with seagrass within the FKNMS system when examined independently may be explained by the ubiquity of seagrass habitat immediately beyond the 250 m spatial scale. Beyond this spatial scale, these spur and groove reefs have vast expanses of back-reef seagrass, and seagrass becomes the dominant habitat type for most study reefs. The absence of significant relationships between reef fish parameters and seagrass under these conditions indicates that caution is warranted in applying broad design principles across systems, since the processes that structure fish communities appear to respond to variation in the landscape structure of each particular system. My findings generally concur with common terrestrial principles: when seeking specific predictive models, structuring mechanisms may always be local and empirically-based, making prediction difficult and perhaps impossible (Bissonette and Storch 2003). Several findings in this dissertation remained consistent across the different study designs, therefore these may represent general principles for applying a landscape ecology approach to tropical marine systems. First, many landscape metrics were valuable in describing and quantifying the landscape structure of the coral reef environments. However, the challenge was to identify those appropriate to the objectives of the study and to organisms or processes of interest. In the US Virgin Islands and the FKNMS, individual habitat features appeared more valuable than complex indices (e.g., principal components) in characterizing the coral reef landscapes. Second, the assumption that all species conform to the same coral reef landscape pattern is rejected by

PAGE 145

128 findings in this dissertation. Rather, the same landscape can be perceived quite differently by different species, as has been consistently reported in terrestrial findings (Lindenmayer et al. 2003, Westphal et al. 2003). Third, landscape relationships were generally consistent with the natural history of each reef fish group, since various life history stages and specific trophic and mobility guilds responded to different habitat features. These findings broadly concur with those from terrestrial systems, which reveal the importance of the natural history of each organism of interest (e.g., life history stage, mobility, dispersal, habitat generalist or specialist) (Stamps et al. 1987, Sisk et al. 1997, Mitchell et al. 2001). Clearly, the design of functional MPAs must consider the autoecology of the species of interest. Fourth, the appropriate spatial scale was also a function of the unique natural history of each reef fish group. For example, juveniles responded to spatial features at a smaller scale (meters to 250 meters) compared to adults (100 m-1 km). In addition, there were clear differences in the response of fishes to landscape characteristics based on mobility. In general, resident species were most closely associated with the fine-scale measure of rugosity within the reef patch, mobile species were strongly associated with landscape characteristics at the scale of 100 m 1km, and transient species were typically not associated with either fine-scale or landscape scale characteristics. Clearly, there is no single scale that is universally appropriate to reef fishes, thus any landscape study warrants investigation at multiple spatial scales (Storch and Bissonette 2003). Fifth, reef context clearly matters. The context of individual reef patches can exert a strong influence on the assemblage structure of reef fishes, perhaps as much or more than the characteristics within a given patch. Selection of reef patches for MPA design should therefore consider how well a

PAGE 146

129 patch relates or links to other patches in the landscape. Sixth, the processes structuring reef fish communities may be modified with variation in the structure of the landscape. Because seagrass habitat was only a valuable predictor of various reef fish groups in the landscape where seagrass was relatively limited (US Virgin Islands), and not a predictor of these same groups in the landscape where seagrass was relatively ubiquitous suggests that managers must consider the specific attributes of each coral reef landscape. Results from the FKNMS suggest that there may be strong limits and thresholds to the response of reef fishes to individual landscape features, which is the case in terrestrial systems (Wiens 2002). Lastly, landscape-scale habitat features appear more valuable than the fine-scale measure of rugosity in predicting most reef fish groups. These findings provide promise for resource managers, since landscape scale metrics are becoming more readily obtainable, and may save precious in-water resources if they can reliably predict reef fish abundance and diversity (Margules and Pressey 2000). There were also consistent negative findings in this research, and though notoriously underreported in the scientific literature, negative results can provide valuable insights for future research. First, complex indices such as habitat diversity and principal components were not predictors of reef fish assemblage structure, findings that contradict several terrestrial studies (Rafe et al. 1985, Ricklefs and Lovette 1999). Second, neither reef size nor reef shape were predictors of reef fish assemblage structure, although both small-scale coral reef studies (Molles 1978, Bohnsack et al. 1994, Gladfelter et al. 1980) and terrestrial theory (MacArthur and Wilson 1967) and empirical studies (Helzer and Jelinski 1999) indicate otherwise. Third, at this spatial scale, the number of habitat patches or habitat richness (the number of different habitat patch types)

PAGE 147

130 are not predictors of reef fish assemblage structure. Finally, the fine-scale measure of rugosity is not capable of predicting reef fish assemblage structure when sampling is limited to within topographically complex reef habitat. Consistency among these negative findings may eliminate particular landscape features as potential predictors of reef fish community structure, or help us to understand the conditions that these metrics do not apparently apply. Because sampling was conducted over a wide range of reef types and locations, and many reef fish parameters were investigated in every study, these negative results appear to be broadly consistent. The findings reported in this dissertation must be interpreted with consideration for the natural spatial and temporal variability of this dynamic system. This is particularly important when interpreting results and applying models derived from this system to another. Understanding the organization of natural communities in spatially heterogeneous environments is difficult because of the inherent complexity of natural systems (Drake 1990), the high spatial and temporal variability of reef fish communities (Sale et al. 1994), and the multiple interactions and processes that occur across scales (Drake 1990). For example, local reef fish abundance may be determined by the relative magnitude of larval recruitment, juvenile settlement, emigration by adults, or by predation and/or competition for structural refuge (Hixon and Beets 1993). Each of these attributes may vary in space and time, potentially obscuring detection of patterns in community organization, and thus contribute to a lower than expected explanatory power for any particular landscape metric. Reef fish community structure in the US Virgin Islands and the FKNMS implies that both systems are heavily overfished. Both systems are dominated by herbivores,

PAGE 148

131 with piscivores contributing less than 5 % of the mean abundance of fishes within trophic guilds. The mean abundance of exploited reef fishes such as haemulids, lutjanids, and serranids was extremely low in both systems, particularly in the exploitable size classs (adults). Low densities of exploited reef fishes may have impaired my ability to detect relationships of reef fishes with landscape-scale features, and thus results must be interpreted with some caution. Further experimentation in less-fished systems will be required to determine whether fish densitiy influences the response of reef fishes to landscape structure. However, the absence of differences in reef fish community structure within and outside protected areas in both systems highlights that current protective measures (i.e. inadequacy of existing fishing regulations, enforcement) are not resulting in significantly different reef fish communities. These findings corroborate other studies in the US Virgin Islands (Rogers and Beets 2001) and FKNMS (Ault et al. 1998c), which conclude that reef fish populations are heavily-exploited due to fishing pressure from recreational and commercial-level fishing. These marine protected areas occur within the jurisdiction of the US federal government, arguably in two of the few locations in the Caribbean that have access to adequate resources for enforcement and education. The overfished status of reef fish communities within these MPAs does not bode well for the rest of the Caribbean region (Jackson et al. 2003). Generalizations relevant to MPA design will be difficult, and may depend on the population structure of the target reef fish community, the degree of fishing within and outside the boundaries of an MPA, the degree of protection sought, recovery rates of the species of particular interest, and landscape features of the area being considered for protection. Some species invariably require very large areas to achieve reasonable

PAGE 149

132 protection, however, findings from this study indicate that the sedentary nature of the majority of fishes may indicate that relative small (1-5 km diameter) MPAs may provide conservation benefits, at least for some species. Resident fishes appear most likely to be influenced by within-patch characteristics (e.g., coral cover), and do not respond to habitat features beyond 500 m from study reefs. Mobile fishes, however, were most influenced by habitat features from 10s of meter to at least 1 km away, and thus MPA should consider the home range sizes of species of interest. Finally, it should be recognized that transient fishes may not be easily conserved within MPAs. Admittedly, this study does not address two important mechanisms that likely influence the ability of a particular MPA to meet its objectives; larval dispersal and spillover. While spillover is not directly addressed in this study, findings imply that spillover benefits to surrounding areas may depend upon the spatial arrangement of the specific coral reef landscape. Larval dispersal was not measured, but further research is clearly needed to determine the extent to which reef fish populations are open or closed (Mora and Sale 2002). Larval dispersal and settlement are likely influenced by a number of factors, including the location, distribution and amount of various habitats necessary for spawning, recruitment, larval export, settlement, growth, foraging and reproduction. I recommend several areas for future research to improve our understanding of how to apply terrestrial landscape ecology principles to tropical marine systems. First, Mechanistic studies that examine individual movements (e.g., sonic tagging) are needed to track the flow of materials and organisms between various habitat patches. Such studies should be designed to examine specific aspects of the landscape (e.g., isolation, corridors of movement) and specific species that vary in their life history strategies (e.g.,

PAGE 150

133 fishes from different mobility guilds). Second, these results must be interpreted within the scope and limitations of this study, and require testing in different coral reef landscapes for results to more generalizable. Future studies should test whether the processes that structure reef fish communities respond to variation in the landscape structure of individual reef landscapes. To do this, studies should be designed in landscapes that vary along a particular landscape gradient of interest to test whether critical thresholds of habitat exist, and if so to identify the threshold values. Finally, nuetral landscape models, which represent real landscapes as null models, have proven extremely valuable in theoretical analyses of pattern-process relationships in terrestrial systems. The utililty of neutral models should be explored in coral reef ecosystems, using a variety of reef fish species. Because terrestrial landscape ecology is only recently being applied to tropical marine systems (Appeldoorn et al. 2003, Christensen et al. 2003, Kendall et al. 2003, Jeffrey 2004), there is considerable research needed to improve our understanding of the landscape principles in marine systems. Although this area of research is in its infancy, clearly there is considerable promise in using terrestrial landscape ecology principles to help us better design and manage reef fishes in coral reef environments.

PAGE 151

APPENDIX REEF FISH DATABASE Species code for every reef fish species encountered in the US Virgin Islands and FKNMS, with the genus, species, diet, trophic guild, size at maturity and mobility guild. Classification into trophic guilds determined using Randall 1965, 1967 and Froese and Pauly 2002. If a single food item comprised greater than 50 % of the diet of a species, the species was classified into that particular trophic guild. indicates those taxa detected only in the Florida Keys. Species Genus & species Diet Trophic Maturity Mobility ABSA Abudefduf saxatilis zoobenthos 44%, zooplankton 13% mif 14.6 resident ABTA Abudefduf taurus* plants 94% herb 15.0 resident ACBA Acanthurus bahianus plants 97% herb 10.0 mobile ACCH Acanthurus chirurgus plants 94% herb 15.5 mobile ACCO Acanthurus coeruleus plants 93% herb 12.0 mobile ACMA Acanthemblemaria maria plankton plank 2.2 resident ACSP Acanthemblemaria sp. no data plank 3.9 resident AENA Aetobatus narinari zoobenthos 54% mif 141.4 transient ALSC Aluterus scriptus zoobenthos sif 58.3 transient AMPI Amblycirrhitus pinos zooplankton 46%, zoobenthos plank 6.7 resident ANSU Anisotremus surinamensis* zoobenthos 99%, crabs, urchins, gastropods mif 36.6 transient ANVI Anisotremus virginicus zoobenthos 70% mif 22.8 transient APBI Apogon binotatus zooplankton plank 8.8 resident APMA Apogon maculatus shrimps 49%, crabs 24%, polychaetes 4%, crustaceans 13% mif 7.7 resident APSP Apogon sp. zooplankton plank 5.0 resident APTO Apogon townsendi no data omni 4.8 resident 134

PAGE 152

135 Species Genus & species Diet Trophic Maturity Mobility ARRH Archosargus rhomboidalis plants 45% herb 8.0 transient ATSP Atherinomorus sp. plankton plank 7.0 transient AUMA Aulostomus maculatus nekton 74%, zoobenthos 23% pisci 53.5 resident BAVE Balistes vetula echinoderm 73%, crabs 5%, shrimp 6% mif 25.0 mobile BOLU Bothus lunatus nekton 86% pisci 27.0 mobile BORU Bodianus rufus crabs 35%, echinoderms 34%, gastropods 10% mif 23.8 mobile CABA Caranx bartholomaei nekton 97% pisci 31.0 transient CACA Calamus calamus zoobenthos 98.5%, polychaetes, mollusks sif 32.1 mobile CALA Caranx latus nekton 100% pisci 40.0 transient CAMA Cantherhines macroceros zoobenthos 87%, sponges tunicates sif 27.0 mobile CAPE Calamus pennatula crabs 24%, echinoderms 14%, bivalve 12%, gastropod 8% mif 22.3 mobile CAPU Cantherhinus pullus plants 50%, zoobenthos 50%-crustaceans omni 12.9 mobile CARO Canthigaster rostrata plants 20%, sponges 17%, crabs 13%, polychaetes 12%, amphipods 10%, gastropods 10% omni 8.2 resident CARU Caranx ruber nekton, zooplankton pisci 24.0 transient CASP Calamus sp.* based on other species mif 27.0 mobile CASU Canthidermis sufflamen* echinoderms 30%, mollusks 24%, zooplankton 42% omni 36.6 transient CHAN Chilomycterus antillarum sessile benthic invertebrates sif 18.5 mobile CHCA Chaetodon capistratus zoobenthos 38% cnidarian, 31% sif 9.0 mobile

PAGE 153

136 Species Genus & species Diet Trophic Maturity Mobility CHCY Chromis cyanea zooplankton 52.4%, planktonic crustaceans plank 10.0 resident CHMU Chromis multilineatus zooplankton 88% plank 12.9 resident CHOC Chaetodon ocellatus zoobenthos 90%, worms 35%, cnidarians 53% sif 13.6 mobile CHSC Chromis scotti* no data plank 7.0 resident CHSE Chaetodon sedentarius detritus 40%, polychaetes 16%, shrimps 16%, zoobenthos sif 10.0 mobile CHST Chaetodon striatus zoobenthos 59%, polychaete worms 60%, cnidarians sif 12.8 mobile CLPA Clepticus parrai zooplankton 90% plank 18.5 resident CLSP Clupeidae zooplanktivorous plank 10.0 transient CODI Coryphopterus dicrus no data omni 3.8 resident COEI Coryphopterus eidolon no data omni 4.5 resident COGL Coryphopterus glaucofraenum plants 50%, ostracods 12%, benthic invertebrates 30% herb 2.4 resident COLI Coryphopterus lipernes no data omni 2.4 resident COPE Coryphopterus personatus no data omni 3.1 resident DAAM Dasyatis americana nekton 21.8%, crustaceans, worms mif 98.8 transient DEMA Decapterus macarellus zooplankton 97% plank 23.8 transient DEPU Decapterus punctatus zooplankton 66%, ostracods 12% plank 11.0 transient DIAR Diplodus argenteus no data mif 18.0 mobile DIHO Diodon holocanthus* mollusks, sea urchins, hermit crabs, 70% mif 29.0 mobile DIHY Diodon hystrix echinoderms, crustaceans, mollusks 99% mif 49.31 mobile

PAGE 154

137 Species Genus & species Diet Trophic Maturity Mobility ECNA Echeneis naucrates zooplankton, detritus, isopods, crustaceans mif 58.3 mobile ELBI Elagatis bipinnulata nekton and invertebrates mif 90.0 transient ELSA Elops saurus nekton 43%, crustaceans, crabs & shrimp pisci 53.5 transient EMAT Emmelichthyops atlanticus nekton and crustaceans pisci 8.8 transient EPAD Epinephelus adscensionis zoobenthos 82%, crust, mollusks, nekton 20% mif 25.0 mobile EPCR Epinephelus cruentatus nekton 67%, zoobenthos 17% pisci 20.0 mobile EPFU Epinephelus fulvus nekton 46%, benthic crustaceans 56% pisci 16.0 mobile EPGU Epinephelus guttatus zoobenthos 70% crabs 40%, stomatopods 21%, nekton 21% mif 25.0 mobile EPIT Epinephelus itajara* zoobenthos lobster 70%, nekton 13% mif 120.0 mobile EPMO Epinephelus morio zoobenthos-nekton mif 40.0 mobile EPST Epinephelus striatus nekton 59%, crustaceans pisci 48.0 mobile EQAC Equetus acuminatus* zoobenthos 74% mif 14.6 mobile EQPU Equetus punctatus zoobenthos 99%, crustaceans, urchins mif 16.9 mobile FITA Fistularia tabacaria nekton pisci 98.8 mobile GECI Gerres cinereus benthic crust, worms, mollusks sif 18.0 mobile GICI Ginglymostoma cirratum finfish 99% pisci 227.0 mobile GNTH Gnatholepis thompsoni plants 74%, copepods 18% herb 5.9 resident GOEV Gobiosoma evelynae ectoparasites sif 3.1 resident GOGE Gobiosoma genie ectoparasites sif 3.5 resident

PAGE 155

138 Species Genus & species Diet Trophic Maturity Mobility GOSA Gobiosoma saucrum Based on other species sif 3.8 resident GRLO Gramma loreto planktonic and benthic crustaceans mif 3.29 resident GYFU Gymnothorax funebris* nekton pisci 120.4 mobile GYMO Gymnothorax moringa nekton pisci 62.9 mobile GYVI Gymnothorax vicinus nekton 63%, zoobenthos 25% pisci 63.9 mobile HAAL Haemulon album* zoobenthos 28%, echinoderms, worms mif 36.0 mobile HAAU Haemulon aurolineatum zoobenthos 64%, zooplankton 35% mif 13.0 mobile HABI Halichoeres bivittatus crabs 25%, urchins 18.7%, polychaetes 18%, gastropods 12% mif 14.1 mobile HACA Haemulon carbonarium zoobenthos 99%, crabs 38%, gastropods 15%, urchins 10%, chitons 8% mif 21.7 mobile HACH Haemulon chrysargyreum zooplankton 40%, zoobenthos 40% plank 14.6 mobile HAFL Haemulon flavolineatum zoobenthos -worms, crustaceans mif 16.0 mobile HAGA Halichoeres garnoti zoobenthos, benthic crustacean mif 12.5 mobile HAJU Haemulon sp. juvenile zooplankton plank 14.0 mobile HAMA Halichoeres maculipinna worms 49%, planktonic and benthic crustaceans mif 11.8 mobile HAMC Haemulon macrostomum zoobenthos-benthic crustaceans mif 25.4 mobile HAME Haemulon melanurum* benthic crust mif 19.0 mobile HAPA Haemulon parrai benthic crust mif 22.3 mobile HAPI Halichoeres pictus zoobenthos mif 8.8 mobile HAPL Haemulon plumieri zoobenthos, benthic crustacean mif 18.0 mobile

PAGE 156

139 Species Genus & species Diet Trophic Maturity Mobility HAPO Halichoeres poeyi zoobenthos 100% mif 12.9 mobile HARA Halichoeres radiatus zoobenthos 100% mif 25.5 mobile HASC Haemulon sciurus zoobenthos 97%, nekton 3% mif 13.0 mobile HASP Haemulon sp. based on other species mif 13.0 mobile HESI Hemiemblemaria simulus* No data sif 7.0 resident HOAD Holocentrus adscensionis zoobenthos 100%, crabs, polychaetes, gastropods, isopods mif 14.5 mobile HOBE Holacanthus bermudensis* sponges, tunicates sif 26.5 mobile HOCI Holacanthus ciliaris sponges, tunicates sif 23.0 mobile HOCO Holocentrus coruscus zoobenthos, benthic crustacean mif 10.0 resident HOMA Holocentrus marianus zoobenthos, crabs, shrimp, mif 11.8 resident HORU Holocentrus rufus zoobenthos, mollusks, echinoderms mif 13.5 mobile HOTR Holacanthus tricolor zoobenthos 97%, crabs sif 17.0 mobile HYAB Hypoplectrus aberrans shrimps, prawns mif 8.8 resident HYCH Hypoplectrus chlorurus shrimp, prawns, fish, crabs mif 8.6 resident HYGU Hypoplectrus guttavarius based on other species mif 8.8 resident HYIN Hypoplectrus indigo based on other species mif 9.6 resident HYJU Hypoplectrus sp. No data plank 8.5 resident HYNI Hypoplectrus nigricans fish, inverts mif 10.2 resident HYPU Hypoplectrus puella shrimps, crabs, fish, mysids mif 10.2 resident HYSP Hypoplectrus species based on other species mif 8.5 resident HYUN Hypoplectrus unicolor benthic crustaceans, fish mif 8.6 resident INVI Inermia vitata zooplankton plank 14.6 transient

PAGE 157

140 Species Genus & species Diet Trophic Maturity Mobility JELA Jenkinsia lamprotaenia zooplankton plank 2.9 transient JESP Jenkinsia sp. based on other species plank 2.9 transient KYSE Kyphosus sectatrix plants 100% herb 42.0 transient LABI Lactophrys bicaudalis sponges, tunicates, echinoderms 20%, algae sif 28.0 mobile LAMA Lachnolaimus maximus mollusks, crabs, urchins, amphipods mif 46.1 transient LANU Labrisomus nuchipinnis crabs, gastropods, urchins, fish mif 14.6 resident LAPO Lactophrys polygonia shrimps, crabs, gastropods, sponges, tunicates mif 29.0 mobile LAQU Lactophrys quadricornis no data mif 27.5 mobile LATR Lactophrys triqueter worms, annelids, ascidians: 50%, crabs 12%, sponges and tunicates 12% sif 27.5 mobile LIRU Liopropoma rubre related to soap fishes feed on crabs, mantis, shrimps mif 53.5 resident LUAN Lutjanus analis crabs 44%, mollusks 13%, nekton 30% mif 51.0 mobile LUAP Lutjanus apodus Nekton 100% adults, zoobenthos juv pisci 25.0 resident LUCY Lutjanus cyanopterus nekton 100% pisci 81.1 mobile LUGR Lutjanus griseus zoobenthos 40%, fish 40% mif 20.0 mobile LUJO Lutjanus jocu zoobenthos, crab, shrimp, fish 65%-61% pisci 45.0 mobile LUMA Lutjanus mahogani finfish, zoobenthos, shrimps, crabs mif 28.0 mobile

PAGE 158

141 Species Genus & species Diet Trophic Maturity Mobility LUSY Lutjanus synagris zoobenthos-crabs, stomatopods, polychaetes, mif 18.5 mobile MABO Malacoctenus boehlkei no omni 4.7 resident MAMA Malacoctenus macropus based on other species omni 4.1 resident MAPL Malacanthus plumieri echin, crabs, stomatopods, polychaetes mif 39.1 mobile MASP Malacoctenus sp. based on other species omni 4.1 resident MATR Malacoctenus triangulates benthic crustaceans miff 5.4 resident MAVE Malacoctenus versicolor benthic algae, copepods, amphipods, eggs omni 5.5 resident MEAT Megalops atlanticus fish pisci 110.0 transient MENI Melichthys niger plants 76%, zoobenthos herb 29.0 transient MICH Microspathodon chrysurus plants 93% herb 13.5 resident MOTU Monacanthus tuckeri zooplankton 42%, detritus 42%, benthic crustaceans rest omni 7.0 mobile MUMA Mulloidichthys martinicus zooplankton 35%, zoobenthos 30%, bivalves 14%, polychaetes 19%, echinoderms 9%, chitons mif 18.0 mobile MUMI Muraena miliaris fish, crabs no relative abundance data pisci 39.1 mobile MYBO Mycteroperca bonaci* nekton 100% pisci 61.5 mobile MYIN Mycteroperca interstitialis nekton 100% pisci 45.3 mobile MYJA Myripristis jacobus zoobenthos 90%, crabs, stomatopods, mysids, shrimps mif 15.7 resident

PAGE 159

142 Species Genus & species Diet Trophic Maturity Mobility MYVE Mycteroperca venenosa nekton 96% pisci 51.0 mobile NIUS Nicholsina usta plants 100% herb 18.5 mobile OCCH Ocyurus chrysurus crabs, isopods mif 24.0 transient ODDE Odontoscion dentex isopods, crabs, fish mif 18.5 OPAT Ophioblennius atlanticus detritus, algae omni 12.4 resident OPAU Opistognathus aurifrons zooplankton 97% plank 7.0 OPOG Opisthonema oglinum zooplankton 70% plank 13.0 PESC Pempheris schomburgki congeneric amphipods and mysids mif resident POAR Pomacanthus arcuatus zoobenthos sponges tunicates 81% sif 24.0 mobile POPA Pomacanthus paru zoobenthos 90%, tunicates, ascidians sif 26.7 mobile PRAR Priacanthus arenatus zooplankton 52%, worm 11%, shrimps 35% plank 26.8 resident PRCR Priacanthus cruentatus zooplankton 60%, shrimps 35%, polychaetes plank 29.4 resident PSMA Pseudupeneus maculatus crabs 31%, shrimps 22%, mysids mif 18.0 mobile RERE Remora remora* zooplankton 22%, isopods 20%, polychaetes 13.7%, detritus plank 47.1 transient RYSA Rypticus saponaceus nekton 48%, shrimps crabs rest, parasites pisci 21.2 resident SCCA Scomberomorus cavalla nekton 93% pisci 79.0 transient SCCL Scarus coelestinus* plants 98% herb 42.5 mobile SCCO Scarus coeruleus plants 100% herb 30.5 mobile SCCR Scarus croicensis plants 100% herb 21.2 mobile

PAGE 160

143 Species Genus & species Diet Trophic Maturity Mobility SCGU Scarus guacamaia* plants 100% herb 62.0 mobile SCJU Scarus sp. juvenile based on other species herb 21.2 mobile SCPL Scorpaena plumieri zooplankton 43%, crabs, shrimps, octopi 57% plank 26.5 resident SCRE Scomberomorus regalis nekton 100% pisci 36.0 transient SCSP Scarus species based on other species herb mobile SCTA Scarus taeniopterus plants 82% herb 21.2 mobile SCVE Scarus vetula plants 94% herb 30.6 mobile SEBA Serranus baldwini* nekton and shrimp, zoobenthos mif 8.2 resident SETA Serranus tabacarius nekton 100% pisci 14.1 resident SETI Serranus tigrinus shrimps 72% mif 17.9 resident SETO Serranus tortugarum plankton 72% plank 14.1 resident SPAT Sparisoma atomarium no data herb 15.7 resident SPAU Sparisoma aurofrenatum plants 98% herb 15.0 mobile SPBA Sphyraena barracuda nekton 96% pisci 66.0 mobile SPCH Sparisoma chrysopterum plants 84% herb 23.9 resident SPJU Sparisoma sp. juvenile based on other species herb 13.1 mobile SPRA Sparisoma radians plants 88% herb 12.9 resident SPRU Sparisoma rubripinne plants 93% herb 16.0 mobile

PAGE 161

144 Species Genus & species Diet Trophic Maturity Mobility SPSP Sphoeroides spengleri bivalves 28, crabs 22%, plants 9%, amphipods 8%, polychaetes 8%, urchins 7% omni 18.5 resident SPVI Sparisoma viride plants 98% herb 17.0 mobile STDI Stegastes diencaeus algae, detritus omni 8.5 resident STDO Stegastes dorsopunicans plants 64% herb 10.0 resident STLE Stegastes leucostictus plants 28%, polychaete 15%, nekton 8%, cnidarians 8% omni 7.0 resident STPA Stegastes partitus benthic algae herb 7.0 resident STPL Stegastes planifrons detritus 27%, plants 24%, cnidarians 20%, zoobenthos 11% omni 9.3 resident STVA Stegastes variabilis plants 57%, worms, tunicates, isopods herb 8.5 resident SYFO Synodus foetens fish 100% pisci 27.0 resident SYIN Synodus intermedius fish 95% pisci 27.0 resident THBI Thalassoma bifasciatum zooplankton 51%, benthic inverts 32%, fish 9.4%, isopods, plank 15.7 resident TRFA Trachinotus falcatus fish pisci 60.1 transient TYCR Tylosurus crocodilus nekton 91% pisci 76.7 transient URJA Urolophus jamaicensis* bony fish, shrimps, bivalves, annelids mif 42.0 transient XYSP Xyrichtys splendens* zooplankton 72%, amphipods 12.5%, bivalves 3%, gastropods 5% plank 11.5 resident

PAGE 162

LIST OF REFERENCES Acosta, C.A. and Roberston, D.N. 2002. Diversity in coral reef communities: the effects of habitat patchiness revisited. Marine Ecology Progress Series 227: 87-96. Akaike, H. 1974. Factor Analysis and AIC. Pschychometrika 52: 317-332. Allison, G.W., Lubchenco, J. and Carr, M.H. 1998. Marine reserves are necessary but not sufficient for marine conservation. Ecological Applications Supplement 8(1): S79-S92. Almany, G.R. 2003. Priority effects in coral reef fish communities. Ecology 84 (7): 1920-1935. Andrn, H. 1994. Effect of habitat fragmentation on birds and mammals in landscapes with different proportions of suitable habitat: a review. Oikos 71:355-366. Appeldoorn, R.S., Recksiek, C.W., Hill, R.L., Pagan, F.E. and Dennis, G.D. 1997. Marine protected areas and reef fish movements: the role of habitat in controlling ontogenetic migration. Proceedings of the 8th International Coral Reef Symposium 2: 1917-1922. Appeldoorn, R.S., Friedlander, A., Sladek Nowlis, J., Usseglio, P. and Mitchell-Chui, A. 2003. Habitat connectivity in reef fish communities and marine reserve design in Old Providence-Santa Catalina, Colombia. Gulf and Caribbean Research 14(2): 61-77. Armbrecht, I., Perfecto, I. and Vandermeer, J. 2004. Enigmatic Biodiversity Correlations: Ant diversity responds to diverse resources. Science 304(5668): 284-286. Arrivillaga, A. and Baltz, D.M. 1999. Comparison of fishes and macroinvertebrates on seagrass and bare-sand sites on Guatemala's Atlantic coast. Bulletin of Marine Science 65(2): 301-319. Ault, T.R. and Johnson, C.R. 1998a. Spatially and temporally predictable fish communities on coral reefs. Ecological Monographs 68(1): 25-50. Ault, T.R. and Johnson, C.R. 1998b. Spatial variation in fish species richness on coral reefs: habitat fragmentation and stochastic structuring processes. Oikos 82(2): 354-364. 145

PAGE 163

146 Ault, J.S. Bohnsack, J.A., and Meester, G.A. 1998c. A retrospective (1979-1996) multispecies assessment of coral reef fish stocks in the Florida Keys. Fishery Bulletin 96(3): 395-414. Baelde, P. 1990. Differences in the structure of fish assemblages in Thalassia testudinum beds in Guadeloupe, French West Indies, and their ecological significance. Marine Biology 105(1): 163-173. Bardach, J.E. 1958. On the movements of certain Bermuda reef fishes. Ecology 39: 139-146. Beck, M.W., Heck, K.L., Able, K.W., Childers D.L., Eggleston, D. B., Gillanders, B. M., Halpern, B., Hays, C.G., Hoshino, K., Minello, T. J., Orth, R.J., Sheridan, P.F. and Weinstein, M. P. 2001. The identification, conservation and management of estuarine and marine nurseries for fish and invertebrates. BioScience 51(8): 633-641. Beets, J.P. and Friedlander, A.F. 1994. Unpublished reef fish monitoring dataset for St. John, USVI monitoring program. NPS database. Virgin Islands National Park. Beets, J., Muehlstein, L., Haught, K. and Schmitges, H. 2003. Habitat connectivity in coastal environments: patterns and movements of Caribbean coral reef fishes with emphasis on bluestriped grunt, Haemulon sciurus. Gulf and Caribbean Research Volume 14(2): 29-42. Bell, J.D. and Galzin, R. 1984. Influence of live coral cover on coral-reef fish communities. Marine Ecology Progress Series 15: 265-274. Bell, J.D. and Pollard, D.A. 1989. Ecology of fish assemblages and fisheries associated with seagrasses. In Larkum, A.W.D., McComb, A.J. and Shepard S.A. (editors). Biology of the seagrasses: a treatise on the biology of seagrasses with special reference to the Australian region. Elsevier, Amsterdam, 565-609. Bell, S.S., Brooks, R.A., Robbins, B.D., Fonseca, M.S. and Hall, M.O. 2001. Faunal responses to fragmentation in seagrass habitats: implications for seagrass conservation. Biological Conservation 100: 115-123. Belsley, D.A., Kuh, E., and Welsch, R.E. 1980. Regression Diagnostics: identifying influential data and sources of collinearity. John Wiley & Sons, New York. Beukers, J.S. and Jones, G.P. 1997. Habitat complexity modifies the impact of piscivores on a coral reef fish population. Oecologia 114: 50-59. Birkeland, C. 1985. Ecological interactions between mangroves, seagrass beds and coral reefs. Pages 1-65 in Birkeland, C. (editor). Ecological interactions between tropical coastal ecosystems. UNEP regional seas reports and studies Number 73. Earthprint, Stevenage, UK.

PAGE 164

147 Bissonette, J.A. 2003. Linking landscape patterns to biological reality. Pages 15-32 in Bissonette, J.A. and Storch, I. (editors). Landscape ecology and resource management: linking theory with practice. Island Press. Bohnsack, J.A. and Talbot, F.H. 1980. Species packing by reef fishes on Australian and Caribbean reefs: An experimental approach. Bulletin of Marine Science 30: 710-723. Bohnsack, J.A. and Bannerot, S.P. 1986. A stationary visual census technique for quantitatively assessing community structure of coral reef fishes. NOAA Technical Report NMFS 41, 15 pp. Bohnsack, J.A., Harper, D.E., McClellan, D.B. and Hulsbeck, M. 1994. Effects of reef size on colonization and assemblage structure of fishes at artificial reefs off Southeastern Florida, USA. Bulletin of Marine Science 55(2-3): 796-823. Bohnsack, J.A., McClennan, D.B., Harper, D.E., Davenport, G.S., Konoval, G.J., Eklund, A.M., Contillo, J.P., Bolden, S.K., Fischel, P.C., Sandorf, G.S., Javech, J.C., White, M.W., Pickett, M.H., Hulsbesk, M.W. and Tobias, J.L. 1999. Baseline data for evaluating reef fish populations in the Florida Keys, 1979-1998. Burnham, K.P. and Anderson, D.R. 1998. Model selection and inference: a practical information-theoretic approach. Springer-Verlag, New York, New York. Burke, N. 1995. Nocturnal foraging habitats of French and bluestriped grunts, Haemulon flavolineatum and H. sciurus at Tobacco Caye, Belize. Environmental Biology of Fishes 42: 365-374. Carr, M.H. and Reed, D.C. 1993. Conceptual issues relevant to marine harvest refuges: Examples from temperate reef fishes. Canadian Journal of Fisheries and Aquatic Sciences 50: 2019-2028. Carr, M.H., Neigel, J. E., Estes, J.A., Andelman, S., Warner, R.R. and Largier, J.L. 2003. Comparing marine and terrestrial ecosystems: Implications for the design of coastal marine reserves. Ecological Applications 13(1): S90-S107. Carter, J.G., Marrow, G.J. and Pryor, C. 1994. Aspects of the ecology and reproduction of Nassau grouper, Epinephelus striatus, off the coast of Belize, Central America. Proceedings of the Gulf and Caribbean Fisheries Institute 43: 65-111. Chapman, M.R. and Kramer, D.L. 2000. Movements of fishes within and among fringing coral reefs in Barbados. Environmental Biology of Fishes 57: 11-24. Christensen, J.D., Jeffrey, C.F.G., Caldow, C. Monaco, M., Kendall, M.S. and Appeldoorn, R.S. 2003. Cross-shelf habitat utilization patterns of reef fishes in southwestern Puerto Rico. Gulf and Caribbean Research 14(2): 9-27.

PAGE 165

148 Clarke, K.R. and Warwick, R.M. 2001. Change in Marine Communities: An Approach to Statistical Analysis and Interpretation. 1st edition: Plymouth Marine Laboratory. Colin, P.L., Laroche, W.A. and Brothers, E.B. 1997. Ingress and settlement in the Nassau grouper, Epinephelus striatus (Pisces: Serranidae), with relationship to spawning occurrence. Bulletin of Marine Science 60: 656-667. Corless, M., Hatcher, B.G., Hunte, W., and Scott, S. 1997. Assessing the potential for fish migration from marine reserves to adjacent fished areas in the Soufriere Marine Management Area, St. Lucia. Proceedings of the Gulf and Caribbean Fisheries Institute 47: 71-98. Cushman, S.A. and McGarigal, K. 2003. Landscape-level patterns of avian diversity in the Oregon Coast Range. Ecological Mongraphs 73(2): 259-281. Dennis, G.D. 1992. Resource utilization by members of a guild of benthic feeding coral reef fish. Ph.D. thesis. University of Puerto Rico, Mayaguez, PR, USA, 224 pp. Department of Commerce. 1997. Code of Federal Register. 15 CFR Parts 922, 929, and 937. Florida Keys National Marine Sanctuary Final Rule. June 12, 1997. Diamond, J.M. 1975. The island dilemma: lessons of modern biogeographic studies for the design of natural preserves. Biological Conservation 7: 129-146. Doherty, P. and Fowler, T. 1994. An empirical test of recruitment limitation in a coral reef fish. Science 263: 935-939. Donaldson, T.J. 2002. High islands versus low islands: a comparison of fish faunal composition of the Palau Islands. Environmental Biology of Fishes 65(2): 241-248. Dorenbosch, M., van Riel, M.C., Nagelkerken, I. and van der Velde, G. 2004. The relationship of reef fish densities to the proximity of mangrove and seagrass nurseries. Estuarine, Coastal and Shelf Science 60: 37-48. Drake, J.A. 1990. Communities as assembled structures: do rules govern pattern? Trends in Ecology and Evolution 5(5): 159-163. Dramstad, W.E., Olson, J.D. and Forman, R.T.T. 1996. Landscape ecology principles in landscape architecture and land-use planning. Harvard University, Island Press. 80 pp. Duarte, C.M. 2000. Marine biodiversity and ecosystem services: and elusive link. Journal of Experimental Marine Biology and Ecology 250: 117-131. Duarte, C.M. and Chiscano, C.L. 1999. Seagrass biomass and production: a reassessment. Aquatic Botany 65: 159-174.

PAGE 166

149 Ebeling, A.W. and Hixon, M.A. 1991. Tropical and temperate reef fishes: comparison of community structure. Pages 509-563 in Sale, P. (editor). The ecology of fishes on coral reefs. Academic Press, San Diego, CA. Eggleston, D.B. 1995. Recruitment in Nassau grouper Epinephelus striatus: post-settlement abundance, microhabitat features, and ontogenetic habitat shifts. Marine Ecology Progress Series 124: 9-22. Eggleston, D.B., Etherington, L.L. and Elis, W.E. 1998. Organism response to habitat patchiness: species and habitat-dependent recruitment of decapod crustaceans. Journal of Experimental Marine Biology and Ecology 223: 111-132. Eristhee, N. and Oxenford, H.A. 2001. Home range size and use of space by Bermuda chub Kyphosus sectatriz (L.) in two marine reserves in the Soufriere Marine Management Area, St. Lucia, West Indies. Journal of Fish Biology 59: 129-151 Supplement A. ESRI. 1996. Environmental Systems Research Institute. ArcView Spatial Analyst. Redlands CA. Fahrig, L. 1997. Relative effects of habitat loss and fragmentation on population extinction. Journal of Wildlife Management 61: 603-610. Fahrig, L. 1998. When does fragmentation of breeding habitat affect population survival? Ecological Modeling 105: 273-292. Flather, C.H. and Sauer, J.R. 1996. Using landscape ecology to test hypotheses about large-scale abundance patterns in migratory birds. Ecology 77(1): 28-35. Florida Marine Research Institute (FMRI)and NOAA. 1998. Benthic habitats of the Florida Keys. FMRI Technical Report TR-4. 53 pp. Fong, C.W., Lee, S.Y. and Wu, R.S.S. 2000. The effects of epiphytic algae and their grazers on the intertidal seagrass Zostera japonica. Aquatic Botany 67: 251-261. Forman, R.T. 1995. Land mosaics: The ecology of landscapes and regions. Cambridge University Press, New York, USA. Forman, R.T. and Godron, M. 1986. Landscape ecology. John Wiley and Sons, New York. Franklin, J.F. 1993. Preserving biodiversity: species, ecosystems or landscapes? Ecological Applications 3: 202-205. Friedlander, A.M. and Parrish, J. 1998. Habitat characteristics affecting fish assemblages on a Hawaiian coral reef. Journal of Experimental Marine Biology and Ecology 224(1): 1-30.

PAGE 167

150 Froese R. and Pauly, D. 2002. Fish Base World Wide Web electronic publication. www.fishbase.org December 09, 2002. Frohn, R.C. 1998. Remote sensing for landscape ecology: new metric indicators for monitoring, modeling, and assessment of ecosystems. CRC Press, Boca Raton, Fl. Gardner, R.H. and ONeill, R.V. 1991. Pattern, process and predictability: the use of neutral models for landscape analysis. Pages 289-308 in Turner, M.G. and Gardner, R.H. (editors) Quantitative methods in landscape ecology: the analysis and interpretation of landscape heterogeneity. Springer-Verlag, New York, New York, USA. Gillanders, B.M., Able, K.W., Brown, J.A., Eggelston, D.B. and Sheridan, P.F. 2003. Evidence of connectivity between juvenile and adult habitats for mobile marine fauna: an important component of nurseries. Marine Ecology Progress Series 247: 281-295. Gilpin, M.E. and Diamond, J.M. 1980. Subdivision of nature reserves and the maintenance of species diversity. Nature 285: 567-568. Gladfelter, W.B., Ogden, J.C. and Gladfelter, E.H. 1980. Similarity and diversity among coral reef fish communities: a comparison between tropical western atlantic (Virgin Islands) and tropical central pacific (Marshall Islands) patch reefs. Ecology 61(5): 1156-1168. Graham, C.H. and Blake, J.G. 2001. Influence of patch and landscape-level factors on bird assemblages in a fragmented tropical landscape. Ecological Applications 11(6): 1709-1721. Green, E.P., Mumby, P.J., Edwards, A.J. and Clark, C.D. 2000. Remote sensing handbook for tropical coastal management. UNESCO publishing. Grober-Dunsmore, R., Frazer, T.K., Beets, J.P., Funicelli, N., and Zwick, P. 2004a. The significance of adjacent habitats on reef fish assemblage structure: are relationships detectable and quantifiable at a landscape-scale? Pages 713-734 in Proceedings of the 55th Gulf and Caribbean Fisheries Institute Conference. Xel Ha, Mexico, November 2002. Grober-Dunsmore, R., Frazer T., Beets, J., Lindberg, W. J., Zwick, P. and Funicelli, N. 2004b. The influence of landscape structure on reef fish assemblages. Chapter 2 of this dissertation to be submitted to Journal of Landscape Ecology. Grober-Dunsmore, R., Frazer T., Lindberg, W., Zwick, P., Beets, J.P. and Funicelli, N. 2004c. Functional habitat connectivity in coral reef ecosystems. Chapter 3 of this dissertation to be submitted to Journal of Conservation Biology.

PAGE 168

151 Grober-Dunsmore, R., Frazer, T. and Lindberg, W.J. 2004d. Reef fishes respond to variation in landscape structure. Chapter 4 of this dissertation to be submitted to Marine Ecology Progress Series. Gustafson, E.J. and Gardner, R.H. 1996. The effect of landscape heterogeneity on the probablility of patch colonization. Ecology 77: 94-107. Hallock, P. 1997. Reefs and reef limestones in earth history. Pages 13-42 in Birkeland, C. (editor) Life and death of coral reefs. Chapman and Hall, New York. Halpern, B. S. 2003. The impact of marine reserves: Do reserves work and does reserve size matter? Ecological Applications 13(1 Supplement): S117-S137. Hanski, I. 1999. Habitat connectivity, habitat continuity, and metapopulations in dynamic landscapes. Oikos 87(2): 209-219. Hart, M.K. 2002. Habitat-mediated direct and indirect effects among three serranid fishes. Masters Thesis, University of Florida. 56 pp. Hatcher, B.G., Johannes, R.E., and Robertson, A. I. 1989. Review of research relevant to the conservation of shallow tropical marine ecosystems. Oceanography and Marine Biology an Annual Review. 27: 337-414. Heck, K.L., Nadeau, D.A., and Thomas, R. 1997. The nursery role of seagrass beds. Gulf of Mexico Science 1: 50-54. Helfman, G.S., Meyer, J.L. and McFarland, W.N. 1982. The ontogeny of twilight migration patterns in grunts (Pisces Haemulidae). Animal Behavior 30(2): 317-326. Helzer, C.J. and Jelinski, D.E. 1999. The relative importance of patch area and perimeter area ratio to grassland breeding birds. Ecological Application 9(4): 1448-1458. Hewitt, J.E., Thrush, S.F., Cummings, V.J. and Turner, S.J. 1998. The effect of changing sampling scales on our ability to detect effects of large-scale processes on communities. Journal of Experimental Marine Biology and Ecology 227(2): 251-264. Hixon, M.A. and Beets, J.P. 1989. Shelter characteristics and Caribbean fish assemblages: Experiments with artificial reefs. Bulletin of Marine Science 44: 666-680. Hixon, M.A. 1991. Predation as a process structuring coral reef fish communities. Pages 475-508 in Sale, P.F. (editor). The ecology of fishes on coral reefs. Academic Press, Inc. San Diego, CA. Hixon, M.A. and Beets, J.P. 1993. Predation, prey refuges, and the structure of coral-reef fish assemblages. Ecological Monographs 63(1): 77-101.

PAGE 169

152 Hobbs, R.J. and Norton, D.A. 1996. Towards a conceptual framework for restoration ecology. Restoration Ecology 4(2): 92-110. Holland, K.N., Peterson, J.D., Lowe, C.G. and Wetherbee, B.M. 1993. Movements, distribution and growth rates of the white goatfish Mulloides flavolineatus in a fisheries conservation zone. Bulletin of Marine Science 52: 982-992. Hovel, K.A. and Lipcius, R.N. 2001. Habitat fragmentation in a seagrass landscape: patch size and complexity control blue crab survival. Ecology 82: 1814-1829. Irlandi, E.A., Orlando, B.A. and Ambrose, W.G. Jr. 1999. Influence of seagrass habitat patch size on growth and survival of juvenile bay scallops, Argopecten irradians concentricus (Say). Journal of Experimental Marine Biology and Ecology 235: 21-43. Jackson, D.A. 1993. Stopping rules in principal components analysis: a comparison of heuristical and statistical approaches. Ecology 74(8): 2204-2214. Janzen, D.H. 1986. The eternal external threat. Pages 286-303 in Soule, M.E. (editor). Conservation Biology. Sinauer Associates, Sunderland, MA. Kareiva, P. 1987. Habitat fragmentation and the stability of predator prey interactions. Nature 326(6111): 388-390. Keller, B.D. 1983. Coexistence of sea urchins in seagrass meadows: An experimental analysis of competition and predation. Ecology 64: 1581-1598. Kendall, M.S., Kruer, C.R., Buja, K.R., Christensen, J.D., Finkbeiner, M. and Monaco, M.E. 2001. NOAA Technical Memorandum NOS NCCOS CCMA 152 (On-Line) Methods used to map the benthic habitats of Puerto Rico and the US Virgin Islands. Kendall, M.S., Christensen, J.D. and Hillis-Starr, Z. 2003. Multiscale data used to analyze the spatial distribution of French grunts, Haemulon flavolineatum, relative to hard and soft bottom in a benthic landscape. Environmental Biology of Fishes 66(1): 19-26. Khalaf, M.A. and Kochzius, M. 2002. Community structure and biogeography of shore fishes in the Gulf of Aqaba, Red Sea. Helgoland Marine Research 55(4): 252-284. King, A.W. 1991. Translating models across scales in the landscape. Pages 479-517 in Turner, M.G. and Gardner, R.H. (editors). Quantitative methods in landscape ecology: the analysis and interpretation of landscape heterogeneity. Springer-Verlag, New York, New York. Koenig, C.C. and Coleman, F.C. 1998. Absolute abundance and survival of juvenile gags in sea grass beds of the northeastern Gulf of Mexico. Transactions of the American Fisheries Society 127: 44-55.

PAGE 170

153 Kolasa, J. and Pickett, S.T.A. 1991. Ecological heterogeneity. New York, Springer-Verlag, 332 pp. Kotliar, N.B. and Wiens, J.A. 1990. Multiple scales of patchiness and patch structure: a hierarchical framework for the study of heterogeneity. Oikos 59: 253-260. Kramer, D.L. and Chapman, M.R. 1999. Implications of fish home range size and relocation for marine reserve function. Environmental Biology of Fishes 55: 65-79. Kushlan, J.A. 1979. Design and management of continental wildlife reserves lessons from the Everglades. Biological Conservation 15(4) 281-290. Levin, S.A. 1992. The problem of pattern and scale in ecology. Ecology 73(6): 1943-1967. Lindeman, K.C., Diaz, G., Serafy, J.E. and Ault, J.S. 1998. A spatial framework for assessing cross-shelf habitat use among newly settled grunts and snappers. Proceedings of the Gulf and Caribbean Fisheries Institute 50: 385-416. Lindeman, K.C., Pugliese, R., Waugh, G.T. and Ault, J.S. 2000. Developmental patterns within a multi-species reef fishery: management applications for essential fish habitats and marine reserves. Bulletin of Marine Science 66: 929-956. Lindenmayer, D.B., Cunningham, R.B., Donnelly, C.F., Nix, H. and Lindenmayer, B.D. 2002. Effects of forest fragmentation on bird assemblages in novel landscape context. Ecological Monographs 72(1): 1-18. Lindenmayer, D.B., McIntyre, S. and Fischer, J. 2003. Birds in eucalypt and pine forests: a landscape alteration and its implications for research models of faunal habitat use. Biological Conservation 110: 45-53. Lindenmayer, D.B. and Hobbs, R.J. 2004. Fauna conservation in Australian plantation forests-a review. Biological Conservation 119(2): 151-168. Luckhurst, B.E. and Luckhurst, K. 1978. Analysis of the influence of substrate variables on coral reef fish communities. Marine Biology 49: 317-323. MacArthur, R.H. and Wilson, E.O. 1967. The theory of island biogeography. Princeton University Press. Princeton, New Jersey. Mallows, C.L. 1973. Some Comments on Cp. Technometrics 15: 661-675. Margules, C.R. and Pressey, R.L. 2000. Systematic conservation planning. Nature 405: 243-253. Mazerolle, M.J. and Villard, M.A. 1999. Patch characteristics and landscape context as predictors of species presence and abundance: A review. Ecoscience 6(1): 117-124.

PAGE 171

154 McGarigal, K. and McComb, W.C. 1995. Relationships between landscape structure and breeding birds in the Oregon coast range. Ecological Monographs 65: 235-260. McGarigal, K., Cushman, S. and Stafford, S. 2000. Multivariate statistics for wildlife and ecology research. Springer-Verlag, New York. Meyer, J.L. Schultz, E.T., and Helfman, G.S. 1983. Fish schools an asset to corals. Science 220(4601): 1047-1049. Meyer, J.L. and Schultz, E.T. 1985. Migrating haemulid fishes as a source of nutrients and organic matter on coral reefs. Limnology and Oceanography 30: 146-156. Meyer, C.G., Holland, K.N., Wetherbee, B.M. and Lowe, C.G. 2000. Movement patterns, habitat utilization, homerange size and fidelity of whitesaddle goatfish, Parupeneus prorphyreus, in a marine reserve. Environmental Biology of Fishes 59: 235-242. Miller, S.L., Swanson, D.W., and Chiappone, M. 2000. Rapid assessment and monitoring of fully protected zones in the FKNMS in Florida Keys National Marine Sanctuary Monitoring Report, Environmental Protection Agency, State of Florida. www.fknms.noa.noaa.gov/research monitoring/welcome FKNMS report 2002. Mitchell, M.S., Lancia, R.A. and Gerwin, J.A. 2001. Using landscape data to predict the distribution of birds on a managed forest: effects of scale. Ecological Applications 11: 1692-1708. Mladendoff, D.J., Sickley, T.A., Haight, R.G. and Wydeven, A.P. 1995. A regional landscape analysis and prediction of favorable gray wolf habitat in the northern Great Lakes region. Conservation Biology 9(2): 279-294. Molles, M.C. 1978. Fish species diversity on model and natural patch reefs: experimental insular biogeography. Ecological Monographs 48: 289-305. Mora, C. and Sale, P.F. 2002. Are populations of coral reef fish open or closed? Trends in Ecology and Evolution. 17(9): 422-428. Morrison, M.L. and Hall, L.S. 2001. Standard terminology: toward a common language to advance ecological understanding and application. In Scott, J.M., Heglund, P.J., Morrison, M., Raphael, M., Haufler, J. and Wall, B. (editors) Predicting species occurrences: issues of scale and accuracy. Island Press, Covello, CA. Mumby, P.J., Edwards, A.J., Arias-Gonzalez, J.E., Lindeman, K.C. Blackwell, P.G., Gall, A., Gorczynska, M.I., Harborne, A.R., Pescod, C.L., Renken, H., Wabnitz, C.C.C. and Llewellyn, G. 2004. Mangroves enhance the biomass of coral reef fish communities in the Caribbean. Nature 427: 533-536.

PAGE 172

155 Murray, S.N., Ambrose, R.F., Bohnsack, J.A., Botsford, L.W., Carr, M.H., Davis, G.E., Dayton, P.K., Gotshall, D., Gunderson, D.R., Hixon, M.A., Lubchenco, J., Mangel, M., MacCall, A., McArdle, D.A., Ogden, J.C., Roughgarden, J., Starr, R.M., Tegner, M.J. and Yoklavich, M.M. 1999. No-take reserve networks: Protection for fishery populations and marine ecosystems. Fisheries Management 24: 11-21. Nagelkerken, I., van der Velde, G., Gorissen, M.W., Meijer, G.J., Vant Hof, T. and den Hartog, C. 2000. Importance of mangroves, seagrass beds and the shallow coral as a nursery for important coral reef fishes, using a visual census technique. Estuarine, Coastal and Shelf Science 51: 31-44. Nagelkerken, I., Roberts, C.M., van der Velde, G., Dorenbosch, M., van Riel, M.C., de la Moriniere, E,. Cocheret and Nienhuis, P.H. 2002. How important are mangroves and seagrass beds for coral-reef fish? The nursery hypothesis tested on an island scale. Marine Ecology 244: 299-305. National Oceanographic and Atmospheric Administration, Florida Keys National Marine Sanctuary, US Environmental Protection Agency, State of Florida. 2002. Sanctuary Monitoring Report 2000. www.fknms.nos.noaa.gov/research_monitoring National Oceanographic and Atmospheric Administration, Florida Keys National Marine Sanctuary, US Environmental Protection Agency, State of Florida. 2003. Sanctuary Science Report 2001: An Ecosystem Report Card. March 21, 2003. Noss, R. 1983. A regional landscape approach to maintain diversity. Bioscience 33(11): 700-706. Ogden, J.C. and Ehrlich, P. R. 1977. The behavior of heterotypic resting schools of juvenile grunts Pomadasyidae. Marine Biology 42(3): 273-280. Ogden, J.C. and Zieman, J.C. 1977. Ecological aspects of coral reef-seagrass bed contacts in the Caribbean. Pages 377-382 in Taylor, D.L. (editor). Proceedings of the Third International Coral Reef Symposium Volume 1. Rosenstiel School of Marine and Atmospheric Science, Miami, Florida. Ogden, J.C. and Quinn, T.P. 1984. Migration in coal reef fishes: Ecological significance and orientation mechanisms. Pages 293-308 in McCleave, J.D., Arnold, G.P. Dodson, J.J. and Neill, W.H. (editors). Mechanisms of migration in fishes. Plenum Press, New York. Ogden, J.C. 1997. Ecosystem interactions in the tropical coastal seascape. Pages 288-297 in Birkeland, C. (editor). Life and Death of Coral Reefs. Chapman and Hall, NY, NY. ONeill, R.V., Krummel, J.R., Gardner, R.H., Sugihara, G., Jackson, B.L., deAngelis, D.L., Milne, B.T., Turner, M.G., Zygmunt, B., Christensen, S.W., Dale, V.H. and Graham, R.L. 1988. Indices of landscape pattern. Landscape Ecology 1(3): 153-162.

PAGE 173

156 ONeill, R.V., Hunsacker, C.T., Timmins, S.P., Jackson, B.L., Jones, K.B., Riitters, K.H. and Wickham, J.D. 1996. Scale problems in reporting landscape pattern at the regional scale. Landscape Ecology 11(6): 169-180. Osenberg, C.W., St. Mary, C.M., Schmitt, R.J., Holbrook, S.J., Chesson, P., and Byrne, B. 2002. Rethinking ecological inference: density dependence in reef fishes. Ecology Letters 5: 715-721. Osenberg, C., Sarnelle, O., Cooper, S.D. and Holt, R.D. 1999. Resolving ecological questions through meta-analysis: goals, metrics and models. Ecology 80: 1105-1117. Parrish, J.D. 1989. Fish communities of interacting shallow-water habitats in tropical oceanic regions. Marine Ecology Progress Series 58(1-2): 143-160. Paton, P.W.C. 1994. The effect of edge on avian nest success: how strong is the evidence? Conservation Biology 8: 17-26. Pearson, S.M. 1993. The spatial extent and relative influence of landscape-level factors on wintering bird populations. Landscape Ecology 8: 3-18. Pearson, S.M., Turner, M.G., Wallace, L.L. and Romme, W.H. 1995. Winter habitat use by large ungulates following fires in northern Yellowstone National Park. Ecological Applications 5: 744-755. Pedlar, J.H., Fahrig, L. and Merriam, H.G. 1997. Raccoon habitat use at two spatial scales. Journal of Wildlife Management 61: 102-112. Peters, R.H. 1991. A critique for ecology. Cambridge University Press. Peterson, B.J. and Heck, K.L. 1999. The potential for suspension feeding bivalves to increase seagrass productivity. Journal of Experimental Marine Biology and Ecology 240(1): 37-52. Pickett, S.T.A. and White, P.S. 1985. (editors). The ecology natural disturbance and patch dynamics. Academic Press, New York. Pielou, E.C. 1977. Mathematical ecology. Wiley, New York. Pikitch, E.K., Santora, C., Babcock, E.A., Bakun, A., Bonfil, R., Conover, D.O., Dayton, P., Doukakis, P., Fluharty, D., Heneman, B., Houde, E.D., Link, J., Livingston, P.A., Mangel, M. McAllister, M.K., Pope, J. and Sainsbury, K.J. 2004. Ecosystem-based fishery management. Science 305(5682): 346-347. Pittman, S.J. and McAlpine, C.A. 2003. Movement of marine fishe and decapod crustaceans: process, theory and application. Advances in Marine Biology 44: 205-294.

PAGE 174

157 Pittman, S.J., McAlpine, C.A. and Pittman, K.M. 2004. Linking fish and prawns to their environment: a hierarchical landscape approach. Marine Ecology Progress Series 283: 233-254. Plan Development Team. 1990. The potential of marine fishery reserves for reef fish management in the US southern Atlantic. NOAA Technical Memo. NMFS-SEFC261: 40 pp. Platt, J.R. 1964. Strong inference. Science 146: 347-353. Preston, F.W. 1962. The canonical distribution of commonness and rarity. Ecology 43: 185-215. Quinn, J.F. and Dunham, A.E. 1983. On hypothesis testing in ecology and evolution. American Naturalist 122: 602-617. Quinn, J.F. and Harrison, S.P. 1988. Effects of habitat fragmentation and isolation on species richnessevidence from biogeographic patterns. Oecologia 75(1): 132-140. Quinn, T.P. and Ogden, J.C. 1984. Field evidence of compass orientation in migrating juvenile grunts (Haemulidae). Journal of Experimental Marine Biology and Ecology 81: 181-192. Rafe, R.W., Usher, M.B. and Jefferson, R.G. 1985. Birds on reserves: the influence of area and habitat on species richness. Journal of Applied Ecology 22: 327-35. Rakitin, A. and Kramer, D.L. 1996. Effect of a marine reserve on the distribution of coral reef fishes in Barbados. Marine Ecology Progress series 131: 97-113. Randall, J.C. 1962. Tagging reef fishes in the Virgin Islands. Proceedings of the Gulf and Caribbean Fisheries Institute 14: 20-241. Randall, J.C. 1963. An analysis of the fish populations of artificial and natural reefs in the Virgin Islands. Caribbean Journal of Science 3(1): 31-46. Randall, J.C. 1965. Food habits of the Nassau grouper (Epinephelus striatus). Proceedings of the Association of Island Marine Laboratories 6: 13-16. Randall, J.C. 1967. Food habits of reef fishes of the West Indies. Studies in Tropical Oceanography (Miami) 5: 665-847. Reese, E.S. 1981. Predation of corals by fishes of family Chaetodontidae: implications for conservation and management of coral reef ecosystems. Bulletin of Marine Science 31:594-604. Ricklefs, R.E. 1973. Ecology. Chiron Press, Portland. 861 pp.

PAGE 175

158 Ricklefs, R.E. and Lovette, I.J. 1999. The roles of island area per se and habitat diversity in the species-area relationships of four Lesser Antillean faunal groups. Journal of Animal Ecology 68(6): 1142-1160. Riitters, K.H., ONeill, R.V., Hunsaker, C.T., Wickham, J.D., Yankee, D.H., Timmins, S.P., Jones, K.B. and Jackson, B.L. 1995. A factor analysis of landscape pattern and structure metrics. Landscape Ecology 10: 23-40. Robbins, B.D. and Bell, S.S. 1994. Seagrass landscapes: a terrestrial approach to the marine subtidal environment. Trends in Ecology and Evolution 9: 301-304. Robblee, M.B. and Zieman, J.C. 1984. Diel variation in the fish fauna of a tropical seagrass feeding ground. Bulletin of Marine science 34: 335-345. Roberts, C.M. and Polunin, N.V.C. 1991. Are marine reserves effective in management of fisheries? Reviews in Fish Biology and Fisheries 1(1): 65-91. Robertson, A.I. and Blaber, S.J.M. 1992. Plankton, epibenthos and fish communities. Coastal Estuarine Studies 41: 173-224. Robinson, S.K., Thompson, F.R., Donovan, T.M., Whitehead, D.R. and Faaborg, J. 1995. Regional forest fragmentation and the nesting success of migratory birds. Science 267: 1987-1990. Rogers, C., McLain, L. and Tobias, C. 1991. Effects of Hurricane Hugo (1989) on a coral reef in St. John, USVI. Marine Ecology Progress Series 78: 189-199. Rogers, C. and Beets, J. 2001. Degradation of marine ecosystems and decline of fishery resources in marine protected areas in the US Virgin Islands. Environmental Conservation 28: 312-322. Rooker, J.R. 1995. Feeding ecology of schoolmaster, Lutjanus apodus (Walbaum), from southwestern Puerto Rico. Bulletin of Marine Science 56: 881-894. Rosenzweig, M.L. 1995. Species Diversity in Space and Time. Cambridge University Press, New York. 436 pp. Rossier, O. and Kulbicki, M. 2000. A comparison of fish assemblages from two types of algal beds and coral reefs in the south-west lagoon of New Caledonia. Cybium 24(1): 3-26. Roughgarden, J.D. 1996. Theory of population genetics and evolutionary ecology: an introduction. Prentice-Hall, Upper Saddle River. 612 pp. Rowley, R.J. 1994. Marine reserves in fisheries management. Aquatic Conservation. 4(3): 233-254.

PAGE 176

159 Russ, G.R. and Alacala, A,C. 2004. Marine reserves: long-term protection is required for full recovery of predatory fish populations. Oecologia 138: 622-627. Sale, P.F., Doherty, P.J., Eckert, G.J., Douglas, W.A. and Ferrell, D.J. 1984. Large scale spatial and temporal variation in recruitment to fish populations on coral reefs. Oecologia 64: 191-198. Sale, P.F. 1991. Reef fish communities: open nonequilibrial systems. Pages 564-596 in Sale, P.F. (editor). The ecology of fishes on coral reefs. Academic Press, London, England. Sale, P.F., Guy, J.A. and Steel, W.J. 1994. Ecological structure of assemblages of coral reef fishes on isolated patch reefs. Oecologia 98: 83-99. Sale, P.F. 1998. Appropriate spatial scales for studies of reef-fish ecology. Australian Journal of Ecology 23(3): 202-208. Sale, P.F. 2002. The science we need to develop for more effective management. Pages 361-376 in Sale, P.F. (editor) Coral reef fishes: dynamics and diversity in a complex ecosystem. Academic Press, London. Sall, J.P. 1990. Leverage Plots for General Linear Hypotheses. American Statistician 44(4): 303-315. SAS Institute. 2003. JMP 8.01 software. Cary, North Carolina. Shulman, M.J. and Ogden, J.C. 1987. What controls tropical reef fish populations: recruitment or benthic mortality? An example in the Caribbean reef fish Haemulon flavolineatum. Marine Ecology Progress Series 39: 233-242. Schumaker, N.H. 1996. Using landscape indices to predict habitat connectivity. Ecology 77(4): 1210-1225. Shulman, M.J. 1984. Resource limitation and recruitment patterns in a coral-reef fish assemblage. Journal of Experimental Marine Biology and Ecology 74(1): 85-109. Simberloff, D. and Abele, L.G. 1976. Island biogeographic theory and conservation practice. Science 191: 285-286. Sisk, T.D., Haddad, N.M. and Ehrlich, P.R. 1997. Bird assemblages in patchy woodlands: modeling the effects of edge and matrix habitats. Ecological Applications 7(4): 1170-1180. Smith, C.L. and Tyler, J.C. 1973. Population ecology of a Bahamian suprabenthic fish assemblage. American Museum Novitates 2528: 1-38.

PAGE 177

160 Smith, C.L. and Tyler, J.C. 1972. Space resource sharing in a coral reef fish community. Pages 125-170 in Collette, B.B. and and Earle, S.A. (editors) Results of the Tektite Program: Ecology of reef fishes. Los Angeles County Museum Science Bulletin 14. Sogard, S.M. 1989. Colonisation of artificial seagrass by fishes and decapod crustaceans: importance of proximity to natural eelgrass. Journal of Experimental Marine Biology and Ecology 133: 15-37. Sokal, R.R. and Rohlf, F.J. 1995. Biometry, 3rd edition, Freeman, New York. Stamps, J.A., Buechner, M. and Krishnan, V.V. 1987. The effects of edge permeability and habitat geometry on emigration from patches of habitat. American Naturalist 129(4): 533-552. Stark, W.A. and Davis, W.P. 1996. Night habits of fishes of Alligator Reef, Florida. Ichthyoplogica 38: 313-356. Suchanek, T.H. 1983. Control of seagrass communities and sediment distribution by Callianassa (Crustacea, Thalassinidea). Journal of Marine Research 41(2): 281-298. Sugihara, G. and May, R.M. 1990. Applications of fractals in ecology. Trends in Ecology and Evolution 5: 79-86. Summerville, K.S. and Crist, T.O. 2001. Effects of experimental habitat fragmentation on patch use by butterflies and skippers (Lepidoptera). Ecology 82(5): 1360-1370. Sweatman, H. and Robertson, D.R. 1994. Grazing halos and predation on juvenile Caribbean surgeonfishes. Marine Ecology Progress Series 111(1-2): 1-6. Syms, C. 1995. Multi-scale analyses of habitat association in a guild of blennoid fishes. Marine Ecology Progress Series 125: 31-43. Syms, C. 1998. Disturbance and the structure of coral reef fish communities on the reef slope. Journal of Experimental Marine Biology and Ecology 230(2): 151-167. Tegner, M. J. and Dayton, P.K. 2000. Ecosystem effects of fishing in kelp forest communities. ICES Journal of Marine Science 57(3): 579-589. Terborgh, J. 1976. Island biogeography and conservationstrategy and limitations. Science 193(4257): 1029-1030. Thollot, P. 1992. Importance of mangroves for Pacific reef fish species, myth or reality? Proceedings of the 7th International Coral Reef Symposium 2: 934-941. Trzcinski, M.K., Fahrig, L. and Merriam, G. 1999. Independent effects of forest cover and fragmentation on the distribution of forest breeding birds. Ecological Applications 9(2): 586-593.

PAGE 178

161 Tulevech, S.M. and Recksiek, C.W. 1994. Acoustic tracking of adult white grunt, Haemulon plumieri, in Puerto Rico and Florida. Fisheries Research 19: 301-319. Turner, M.G. 1989. Landscape ecology: the effect of pattern on process. Annual Review of Ecological Systems 20: 171-197. Turner, M.G., Gardner, R.H. and ONeill, R.V. 2001. Landscape ecology in theory and practice: pattern and process. 401 pp. Underwood, A.J. 1997. Experiments in ecolkogy: their logical design and interpretation using analysis of variance. Cambridge University Press. 504 pp. van Dorp, D. and Opdam, P.F.M. 1987. Effects of patch size, isolation and regional abundance on forest bird communities. Landscape Ecology 1: 59-73. Villard, M., Trzcinski, M.K. and Merriam, G. 1999. Fragmentation effects on forest breeding birds: relative influence of woodland cover and configuration on landscape occupancy. Conservation Biology 13: 774-783. Walsh, W.J. 1985. Reef fish community dynamics on small artificial reefs-the influence of isolation, habitat structure and biogeography. Bulletin of Marine Science 36(2): 357-376. Westphal, M.I., Field, S.A., Tyre, A.J., Paton, D. and Possingham, H.P. 2003. Effects of landsacpe pattern on bird species distribution in the Mt. Lofty Ranges, South Australia. Landscape Ecology 18(4): 413-426. Wiens, J.A. 1976. Population response to patchy environments. Annual Review of Ecology and Systematics 7: 81-129. Wiens, J.A. and Milne, B.T. 1989. Scaling of landscapes in landscape ecology, or landscape ecology from a beetles perspective. Landscape Ecology 3: 87-96. Wiens, J. 1989. Spatial scaling in ecology. Functional Ecology 3: 385-397. Wiens, J.A. 1992. Ecology 2000: an essay on future directions in ecology. Bulletin of the Ecological Society of America 73(3): 165. Wiens, J.A. 2002. Riverine landscapes: taking landscape ecology into the water. Freshwater Biology 47(4): 501-515. Williams, D. McB. 1980. Dynamics of the pomacentrid community on small patch reefs in One Tree Lagoon (Great Barrier Reef). Bulletin of Marine Science 30: 159-170. Williams, D. McB. 1991. Patterns and processes in the distribution of coral reef fishes. Pages 437-474 in Sale, P.F. (editor) The ecology of fishes on coral reefs. Academic Press, San Diego, CA.

PAGE 179

162 Wilkinson, C. 2000. Status of the Coral Reefs of the World: 2000. AIMS, United States Coral Reef Task Force, Cape Ferguson, Queensland, and Dampier, Western Australia. 136 pp. With, K.A. and Crist, T.O. 1995. Critical thresholds in species responses to landscape structure. Ecology 76: 2446-2460. Wolff, N.H. 1996. The fish assemblages within four habitats found in the near shore waters of St. John, United States Virgin Islands: with some insights into the nature of trap fishing. Masters Thesis. Department of Oceanography, University of Rhode Island. 207 pp. Zar, J.H. 1984. Biostatistical analysis. (Second edition). Prentice Hall. Englewood Cliffs, N.J. 718 pp. Zieman, J.C. and Wetzel, R.G. 1980. Productivity in seagrass: methods and rates. Pages 87-116 in Phillips, R.C. and McRoy, C.P. (editors) Handbook of seagrass biology. Garland Publishing, Inc. New York.

PAGE 180

BIOGRAPHICAL SKETCH Linda Erica Rikki Grober-Dunsmore was born in Oslo, Norway, on October 15, 1964 and was raised in Ft. Pierce, Florida. She graduated from Florida State University (Tallahassee) in 1987, and completed her Masters of Environmental Management at Duke University (Durham, NC) in 1992 in coastal resource ecology, with Dr. Orin Pilkey. There she was selected as a fellow for the Organization for Tropical Studies, to attend the tropical ecology and conservation program in Costa Rica. Rikki worked for the World Wildlife Funds Biodiversity Support Program, for Dr. Ed Towle of the Island Resources Foundation in St. Kitts, British West Indies. She then worked as a marine ecologist for Virgin Islands National Park and Biosphere Reserve. There she conducted coral reef monitoring and research, under the direction of Dr. Caroline Rogers, of the US National Park Service, the National Biological Service and the US Geological Survey (USGS). In 1997, she worked with Dr. Doug Markle (at the University of Oregon in the Department of Fisheries and Wildlife) on endangered suckers in the Lower Klamath River Basin. In 1999, she started with the Hawaii Coral Reef Initiative, Hawaii, where she edited the Proceedings of the Hawaii Coral Reef Monitoring Workshop with Dr. Jim Maragos and worked with Dr. Mark Ridgley at the University of Hawaiis, Department of Geography to develop a strategic, multi-objective framework to integrate scientific knowledge with resource-use patterns to improve coral reef management. She worked 163

PAGE 181

164 with Dr. Chip Fletcher at the Department of Coastal Geology and Geophysics, University of Hawaii, researching a book Living with the Shores of Hawai'i". Rikki matriculated to the University of Florida in fall 2000, to work with Dr. Tom Frazer under a Student Cooperative Education Program with the Biological Resources Division of the USGS, in the Coral Reef Ecology program headed by Dr. Nick Funicelli. Rikki spent several months a year in the US Virgin Islands, and conducted field research in the US Virgin Islands, and Turks and Caicos Islands, and the Florida Keys. Rikki is also mother to Thatcher Kai Dunsmore, born August 4th 1997.


xml version 1.0 encoding UTF-8
REPORT xmlns http:www.fcla.edudlsmddaitss xmlns:xsi http:www.w3.org2001XMLSchema-instance xsi:schemaLocation http:www.fcla.edudlsmddaitssdaitssReport.xsd
INGEST IEID E20110115_AAAAEL INGEST_TIME 2011-01-16T00:39:50Z PACKAGE UFE0008350_00001
AGREEMENT_INFO ACCOUNT UF PROJECT UFDC
FILES
FILE SIZE 15181 DFID F20110115_AADAWE ORIGIN DEPOSITOR PATH dunsmore_l_Page_047.pro GLOBAL false PRESERVATION BIT MESSAGE_DIGEST ALGORITHM MD5
e786b6700545283e6683fdec181f1361
SHA-1
0fe4f91c99176010ea924ad8d9c521226d3faa13
64223 F20110115_AADBAW dunsmore_l_Page_171.pro
a2d674453ca651a1430ecad61435c3b0
0407c3fb750205843491a670a0ecb52e00ccea05
49736 F20110115_AADAVQ dunsmore_l_Page_033.pro
f37a155c7f9fccaa08e16cdf2662300c
333b6551477e607a084a88f8f0867c1d11776aa3
1802 F20110115_AADBBL dunsmore_l_Page_005.txt
b6169e6243ea12331a9dc7cd7e38dd29
4457b8bd1cc761371a9ec56eea484cd9470cbe08
23619 F20110115_AADAWF dunsmore_l_Page_048.pro
f4cce8ff5b8dc51009e60ad7d071d039
9526be8e606e1baa8b3ffa4cda3af58bb0888599
70406 F20110115_AADBAX dunsmore_l_Page_172.pro
64ac6ca6e9ec683d7510a36abbecf518
0386f27a1597e8f36da9105c44512bb0578277ff
51449 F20110115_AADAVR dunsmore_l_Page_034.pro
5de536e8941080d9d322e2cf30488e3b
d9d6fc7cb387f7b2151220ab693a2fa8112445ce
2044 F20110115_AADBCA dunsmore_l_Page_020.txt
fc3397f466c73552867af70c8e0a9926
74ad20d26e985725f92202de9589749408efbf6c
1951 F20110115_AADBBM dunsmore_l_Page_006.txt
8bce9f272887e9e13dac40a01ac4f8dc
01fb2d4a4dbab67537e1cd9ad6bf30c34a59a1a3
41306 F20110115_AADAWG dunsmore_l_Page_049.pro
e696a04014600bf20096211762a7fdf8
814c2e059de027e9d5e103ddae4254c40cf7de1a
59608 F20110115_AADBAY dunsmore_l_Page_173.pro
78a30cb521a6f1949a22767fe7109f8f
1539afd15514ce8068bddb32bfa5c7a7978bc3d9
48363 F20110115_AADAVS dunsmore_l_Page_035.pro
c1d048c975d0367db8d4eeecc0fe5304
700807bcaa0d048a76ee202afd69db0ec2164b51
2024 F20110115_AADBCB dunsmore_l_Page_022.txt
99c1bf3af528bc22f25f8939caf2ff9f
7b0af4c055d05f02c6da595c5f2337e57fb94400
68593 F20110115_AADAWH dunsmore_l_Page_050.pro
47ea2eeb5d2a51870f25310f98bf777d
6bf897555ce925bc7ded8948bef13b9de03be378
55841 F20110115_AADBAZ dunsmore_l_Page_174.pro
01be8f5ec5217d231ff2bde1e2704fbb
de19cb992602418bbafeb2cfdf2dc2ee9d8bcd7a
50167 F20110115_AADAVT dunsmore_l_Page_036.pro
67b736dae349b697a053f3d6eb7ea162
5c0c1d16a05e6ecd3c8a17bc2a1ffc0329227582
2026 F20110115_AADBCC dunsmore_l_Page_023.txt
6270472d663d0a4fc55bb898df7433d3
3b1394f1fd72b554661b8aa27741bebf87b281e4
1965 F20110115_AADBBN dunsmore_l_Page_007.txt
89bd472e222d3ac6000eca2afaa73318
4c96e2aecd3fa34d1909fbc6a27c7e2f046b6fd6
56454 F20110115_AADAWI dunsmore_l_Page_051.pro
e2c3de94c60c746c7caef1c5f6036704
fa53087e5b517e1578e1a017d9682f7c06d1bd9e
47119 F20110115_AADAVU dunsmore_l_Page_037.pro
fa4b8b45df8438cd7ce0d41d10ebe23a
ce199555a5c340a3757f5327b84811784d7b93b7
2008 F20110115_AADBCD dunsmore_l_Page_024.txt
50763c5d813885d92cc91ae77083d4d8
e3226bf72812a565422f1fd6ccf2d75ee5b5c654
526 F20110115_AADBBO dunsmore_l_Page_008.txt
b53b6b503f2a391c622f1f70089d18d1
a1bd213848dab12edabd7e4839c408e7a43a1631
22478 F20110115_AADAWJ dunsmore_l_Page_052.pro
883cde91adb991817a6216636a7dc867
c3285724e22f03215bc4940bea9b6b90676975ec
49957 F20110115_AADAVV dunsmore_l_Page_038.pro
69c57aefc66a0742f0d5d18f3eb09a13
b59632f03f37587a65da26f576eecc94f120722a
1963 F20110115_AADBCE dunsmore_l_Page_025.txt
c63019bfd2cee631ec63fe75bf978958
b020090dad3580f594ba813687f72d67a334826d
3075 F20110115_AADBBP dunsmore_l_Page_009.txt
2adf284619ebc6c42b48be1857decae0
5169536a596f645faf5401957b1c0b060a94ba01
43611 F20110115_AADAWK dunsmore_l_Page_053.pro
bc352e69e64f433a5715e7010acf4adb
3bef07cb21ac6d7697922d9a4f1ab6d808b5ce23
52330 F20110115_AADAVW dunsmore_l_Page_039.pro
1c3472a570007b46dc0e04209fd1949a
23308b27424d5d796cae4b498b8ca9bd0bed88bb
1994 F20110115_AADBCF dunsmore_l_Page_026.txt
aac8413d6efb9dc38509bdb18285ff22
06260c28f9dbd7a44a50dce42338a8e60e77b6e2
2659 F20110115_AADBBQ dunsmore_l_Page_010.txt
3b195fc20c4304b244bbad699ef2908e
bf774a3e44114dd8cbae2f5c05c91e131024568b
35522 F20110115_AADAWL dunsmore_l_Page_054.pro
a16cb09535dda807c9552f51cc726d8b
e1be609ec90f946e304c5f2af41f293476fb3ca4
50979 F20110115_AADAVX dunsmore_l_Page_040.pro
4a26bc565c82cdef1856b6115f26815e
966a28db72f4c00b24822cc790460c1c4babb233
1653 F20110115_AADBCG dunsmore_l_Page_027.txt
b285085b98950cfdd9f1f7d4815039e3
469751d725cce2accbc02ceb2f2aa8c2df578f79
43746 F20110115_AADAXA dunsmore_l_Page_070.pro
e7fa862ef9ee9d26af9907e893d72e3e
9abf5945ac581387a1bcf36c2ffba5331f009397
2748 F20110115_AADBBR dunsmore_l_Page_011.txt
fd269de8f1aa4eee0677bf41f61a2eba
47731d56ff7a89601821b6cc71e8804239a3f878
40051 F20110115_AADAWM dunsmore_l_Page_055.pro
9339efce8c59ad0aad316442d36c3c5d
efbb03674a53d7aae8626593120a4c05d66cfdd0
52407 F20110115_AADAVY dunsmore_l_Page_041.pro
d80ab5a21d6df619166c8fba5dc18ee3
5fce884f9a3d0d57e0d605681a0f7aa44efe72a3
1841 F20110115_AADBCH dunsmore_l_Page_028.txt
3ad1a9adb3f388c586a9e94765b012aa
427f0b75c6a6f4683a7cb8c8e0c829b850a6c185
45718 F20110115_AADAXB dunsmore_l_Page_071.pro
b653bb068d1f25adf3d1df824d043202
d044a8ff4709e4eb63d57bbdfaf787a36966e3b6
3336 F20110115_AADBBS dunsmore_l_Page_012.txt
53966d555b6606634150678161c558ea
450564223ec26ed6ccca1e28ab99a5263f6a3f97
36504 F20110115_AADAWN dunsmore_l_Page_056.pro
1db6560c4beb6fa86f0037d404dbcf76
e1e7f6d8e718a02ab581c1c91bf2702caa73ba28
51516 F20110115_AADAVZ dunsmore_l_Page_042.pro
416db91a0786ee78e3695e3598e3a36f
12624201a86be1fca76376c47905d4e385139062
2041 F20110115_AADBCI dunsmore_l_Page_029.txt
d3958db477d00a94f466f5ae5ed6dfbc
dea9b3afac83aee13107759d94554f02125d9ea0
45053 F20110115_AADAXC dunsmore_l_Page_072.pro
219a6d3cf0afd90a60039f27e9096748
2d5d0779855bbf9d0d9d2275cf8791828593f2de
660 F20110115_AADBBT dunsmore_l_Page_013.txt
c64cbabbfeb32a15c9a0a0f92be97498
42fb981600a80eb62d6e9ab6ad6de9735010daa5
4334 F20110115_AADAWO dunsmore_l_Page_057.pro
30a4a3cae7ce7eb6948a307ecd9d23d2
d756b3c557310ad084bfc800f3b047ddcdcb8a82
2047 F20110115_AADBCJ dunsmore_l_Page_030.txt
82dc9f7c8d58fcc785f56ec2df67e48a
51bb0b8fef5452debd4d7ec6482ef78a0dab8114
49686 F20110115_AADAXD dunsmore_l_Page_073.pro
fa3686d349acc3419ab6df1a24b33534
62bcfe0af9df0055f064f5c65edfc4015ad09419
2558 F20110115_AADBBU dunsmore_l_Page_014.txt
f845944ff59633ee5d861a7eb9935eb2
5ba82ab280bd8c8caa99ab77198adbb2b1869496
4411 F20110115_AADAWP dunsmore_l_Page_058.pro
14600c71d624083415d2a25054f5c137
c4b4f21d8d5768fd16904c7216cee78e7f9663d1
2021 F20110115_AADBCK dunsmore_l_Page_031.txt
76cb4b64d55cf45f4780eaf2b7ad5128
f53e9fa64ba72d896784709b0dbcbdd06a010b81
2335 F20110115_AADBBV dunsmore_l_Page_015.txt
54e4b0f1cec707587fd1e7d411c1659e
d47fa8a9409f55c560b8fcfb4f1605f7169d47ca
20730 F20110115_AADAWQ dunsmore_l_Page_059.pro
2315d9cfc561a2c4643489c58e38134d
73f9a00c7c7da8b0d94b4f51ab3cd515703f12a6
2037 F20110115_AADBCL dunsmore_l_Page_032.txt
5b5382f5748f30bb795c4c8b60438481
bb26d4df8eb365c065693d2d891b41efd3f6c62e
51893 F20110115_AADAXE dunsmore_l_Page_074.pro
ac669c8e65b243b3df70ec88f6d9e1f2
3bd81bc87a3c14e37861bbc1415992106e0ad3c1
1805 F20110115_AADBBW dunsmore_l_Page_016.txt
b706ff18de8c209258fb93e8d8834d44
541084fb092f3caa30b697d1800c7a9825857f4f
8970 F20110115_AADAWR dunsmore_l_Page_060.pro
71cbbb2ad3804afb11db1ad53bde6e7d
7be1ce0ce93ab3d42da3e69d78e68f82e1090eed
645 F20110115_AADBDA dunsmore_l_Page_047.txt
b431da3ece6335bbdd4426ace47113e0
1673b8634695eb4f1093b8f293fb33e71348be78
1958 F20110115_AADBCM dunsmore_l_Page_033.txt
886b931ffe730dff8638c5831941cdeb
9b7eb9e4f5badb9c5e9b468506c34d0dc5e9f90f
50967 F20110115_AADAXF dunsmore_l_Page_075.pro
24c61a257c50cc13cfa3427cd3332c83
a4197e586aa3ffd6e78d3dddb1f2dbc159c8b73d
1168 F20110115_AADBBX dunsmore_l_Page_017.txt
524f55b706371e2a3d7b908e338fff54
74a92dc86b8996b9c5e7cf0747eeb3f04b4c7d3c
29785 F20110115_AADAWS dunsmore_l_Page_061.pro
8589c962cf3e5153e6b468cdfcff03de
4c57f2e5aa60bcebdcc1b3f1c8c1658683b7c54d
1059 F20110115_AADBDB dunsmore_l_Page_048.txt
dbe36811dcd838cef6b82cead670e1e7
f0a3215fb6eded9e8156433ca35d75cbcf7bfad2
2030 F20110115_AADBCN dunsmore_l_Page_034.txt
0ae7edc7728d66cba99eb8267b80cd72
50e8cec9032e257c425ca0a2bef6b311d9587aed
50045 F20110115_AADAXG dunsmore_l_Page_076.pro
7c6f420569314465f2af9aa3287049f3
03c07ed16968737f9c059ad6b30545ef703019f2
1879 F20110115_AADBBY dunsmore_l_Page_018.txt
cd896950ea01ff5b4286476adf28112a
5f90584dce75780282fb07e74ed89104a9ce0ea9
47862 F20110115_AADAWT dunsmore_l_Page_062.pro
5ede4a8cd3fd948174ffbf80ea183e23
d3cf4d12db1189e55edb235709c989e141e7647a
24905 F20110115_AADAAA dunsmore_l_Page_008.jpg
26c42b8caf4d607555f6fa515ca32df8
3befbf99ef832a6acb213c32e1721d5a600d42d6
1979 F20110115_AADBDC dunsmore_l_Page_049.txt
26a4d0d2801a62833e465ef6e96ec0bc
bf72376640b1aa83ffff5d1d809d4f9d9cc04f83
47268 F20110115_AADAXH dunsmore_l_Page_077.pro
4fd2de186eb52593a4fccf4042123208
39832fc99b83b965f45fc4e26b01e1eae3d8f1c6
2015 F20110115_AADBBZ dunsmore_l_Page_019.txt
2b3c9f101d3b55cb57a251e961639b9d
a679dd8f0e8650870dead2acef92934dee4cac94
49130 F20110115_AADAWU dunsmore_l_Page_063.pro
34b1c6a51752ba00dc79f42183a01b17
c6cf25e489ff9c0a93ac9757475a1d78e82d30cb
63623 F20110115_AADAAB dunsmore_l_Page_009.jpg
11feb316202343029232a32341c5dc4d
cabdf983190b0330d11afb9d7a9a8bd607264158
3157 F20110115_AADBDD dunsmore_l_Page_050.txt
34bd28f32a18da1b45608622bbd4de51
91a262b741bd501fb529c59b807942eedd2c9990
1915 F20110115_AADBCO dunsmore_l_Page_035.txt
3f75c72b065d9558194b0d86e1402ae3
3ff2729a67fc64c17f73ff5de122c0860fb1f04e
51426 F20110115_AADAXI dunsmore_l_Page_078.pro
9c575456aa2fbd5e5d0dc59fce4d6b18
af71aa4afdc5e4f3fe185f10ed12061a4c511dc0
52034 F20110115_AADAWV dunsmore_l_Page_064.pro
c52c7473e16f3a8d48999e4ec3023547
29f3c02cbe725eb165a6a4de85202e46912814e4
56620 F20110115_AADAAC dunsmore_l_Page_010.jpg
aa6f62469c3534cc4c4a28d76af7cc77
4013c747256ab38a13080e447c810a1797b1108a
2505 F20110115_AADBDE dunsmore_l_Page_051.txt
edf2be3b4f74174dfacc4e65e6ef4f2f
a921fb03106b2634cc9aa544e0aeffdc333c0710
2012 F20110115_AADBCP dunsmore_l_Page_036.txt
3bd839762082bc3dc1a28dd2322d7652
8fc74bd36f616b67d1a0e73a031eaa2aa52e2a85
51459 F20110115_AADAXJ dunsmore_l_Page_079.pro
6c404fe7a85ff7b32cd0ebf89a2c6cdd
08f5a1833a87bdd7f2de1f29a169e0211e557546
50827 F20110115_AADAWW dunsmore_l_Page_065.pro
b2be60f26309a574fb324f3d67e8a46e
ac56729ee86df8eacecac06a1aa6a5528eb57ae6
89470 F20110115_AADAAD dunsmore_l_Page_011.jpg
79ceb460ecbc8dfc548d05bfa40714a8
4d48339df23a3c56407edafb52980ea23cab14e6
1192 F20110115_AADBDF dunsmore_l_Page_052.txt
d36a3ece5c54d9de40bbb487a4b6cde7
8120b7a2e2fdd7a916bdababe660b83f394578c4
1876 F20110115_AADBCQ dunsmore_l_Page_037.txt
29e0108d810e10b6564a1f4ea8a80d0c
502a14f902af4486a2f7a5ad2f73010daa0bec96
53427 F20110115_AADAXK dunsmore_l_Page_080.pro
aeb807710a409e49a236260866677275
63df016934b114f78c9b5f3424a6bcbc0531eb1f
41203 F20110115_AADAWX dunsmore_l_Page_066.pro
f7c97a56071bf528843477ac628687a4
092eaa418c7967d914a3f0d563d0c0d621f9effb
117974 F20110115_AADAAE dunsmore_l_Page_012.jpg
b1225aea3f269730a62b091240e1e9a4
129615b49fcac2c85ab0fe158692dd8bd8e1ff82
2046 F20110115_AADBDG dunsmore_l_Page_053.txt
bdee5cf014039de593b8ac779bfc55a1
f22236a260fe44ebe54bb424f45a099f085e9236
47978 F20110115_AADAYA dunsmore_l_Page_096.pro
cb8653d8b3ea9dbb91c084b74260993b
bd6eceaa276cd9aac27deec178448e2d0732d38e
1967 F20110115_AADBCR dunsmore_l_Page_038.txt
e27219755dd47f75b1d4f8731bcaab1d
589fbaf22d1818083a0eb8481685847683fa64fc
52004 F20110115_AADAXL dunsmore_l_Page_081.pro
a84f7f2109c070a73cf9e18db8cb4f5a
c92d3e7ce0ac7cdbc1c443376d1494accfd14c3c
30237 F20110115_AADAAF dunsmore_l_Page_013.jpg
7a79be1c73c4961bae4f39ba2150adb8
3babd1deec90d78cfd7a2a31a867eec748bbe67d
1649 F20110115_AADBDH dunsmore_l_Page_054.txt
4ca1c02032ea7b13438c815d4fc93421
e8d873485f4bfaa7704045c884e221cc888efe66
49213 F20110115_AADAYB dunsmore_l_Page_097.pro
1b6a5e9a652bf27d5bd2046f28dc051e
e68a5b92e4b00933d3d2eaa0d138f5767fbde0d7
2086 F20110115_AADBCS dunsmore_l_Page_039.txt
0bb5a52d2fb54c21a3d5058a5ac919e4
90ee5a1a5bc2274a10850eb4aec16940d1922e5b
8220 F20110115_AADAXM dunsmore_l_Page_082.pro
3def4295f85c00a66036c40cb6bac1e5
05e2df22b4df51c372669038bce1b9c8c0a5aebe
50760 F20110115_AADAWY dunsmore_l_Page_067.pro
9c44ab30e3edf08e84b703b54cf45588
68e52322df241bb097cf126c84e2f0935429b77f
83166 F20110115_AADAAG dunsmore_l_Page_014.jpg
e84483f9b4f8c0e6246dc89d09688030
323e5f005bc6c0c314a391f290dcda4ca95c6257
1953 F20110115_AADBDI dunsmore_l_Page_055.txt
c9501f7a9851930d7370b45695973430
f1b32a352551bf331c40f21fe31c39d3e36ae8a4
52572 F20110115_AADAYC dunsmore_l_Page_098.pro
9412f58ed2c8b23bd15daf5ad0e572b8
94bb4cf91d232cc14dbf258b024d3a9049404000
F20110115_AADBCT dunsmore_l_Page_040.txt
f5fd092aae64a10ee42e5f748241f3f6
8afcceaa6166433c22bf9626d217e6627f0ca1f5
4817 F20110115_AADAXN dunsmore_l_Page_083.pro
b0d212901753debb8bf4a3804aac4b2e
7fbd80f145837fff81d5c656d0890ee6dcf309c1
51872 F20110115_AADAWZ dunsmore_l_Page_069.pro
e506c6fb53e40c60068d384e90e6ec7c
f559d2e9269af2061d59da75576226827fd66621
1685 F20110115_AADBDJ dunsmore_l_Page_056.txt
1ef4ec68e0738840852f9fa1cef42539
72716acf443b7e91348a6aba761c12ebbb15bc74
80371 F20110115_AADAAH dunsmore_l_Page_015.jpg
fd7644048f6853fc9094dcb6f51ff438
0230260ce14426354e2c511d8b703372e71806b2
52528 F20110115_AADAYD dunsmore_l_Page_099.pro
b2d3fc010d06a13a4cb17a66b3cb6e0e
ab223dda8c17f348632bc1296e4e56a2999cf422
2063 F20110115_AADBCU dunsmore_l_Page_041.txt
e1630507f7fca3fb0ae02f7a848d7ed6
86ef2c4c8841b6c1b29653886e324057093b0940
12098 F20110115_AADAXO dunsmore_l_Page_084.pro
f80fb2312fd8537a53083307160461b6
a9403c9f18722acb39bd43d743e44ec51db56d74
309 F20110115_AADBDK dunsmore_l_Page_057.txt
708022ef4231268db5fb111a48187e6b
4727cb21acc79e47aaf268ecbe6dc1ed7edaab68
63956 F20110115_AADAAI dunsmore_l_Page_016.jpg
c84de677674c48d687d199fccf4fe59f
18f6eda59a0d1d5f0eb2b2e5529442a8be9dddfe
51028 F20110115_AADAYE dunsmore_l_Page_100.pro
9bba72170112f95fc25b62686ea2c4f9
55ce193f95e66df1f15b1e6a2717f54c39e58e8b
2029 F20110115_AADBCV dunsmore_l_Page_042.txt
8fc1fdbc34346b5f700d79993336ade2
6fd95b03e77ffc75b0fab257d4ec5b937b141fa1
21467 F20110115_AADAXP dunsmore_l_Page_085.pro
d60d0c1803f9a8646d6b2ba2332bf241
b90d16a2a998b8f93acd557471eef12aea5d39d0
242 F20110115_AADBDL dunsmore_l_Page_058.txt
87a8ebaeef4b95b90e2de17dd7e22aa7
e603bdd028af058af44f955dadcdbaf1ba279d39
45371 F20110115_AADAAJ dunsmore_l_Page_017.jpg
6d8aa62ead34105ff073cd9b7b01fc92
bc9c9c48b79f830e2bc9ed9ac3c2c9c0154f5b22
F20110115_AADBCW dunsmore_l_Page_043.txt
492d66b58967bc7118904dbd5f64eadb
a69bedc87f1015e054f7d9811b979934d026a514
16335 F20110115_AADAXQ dunsmore_l_Page_086.pro
d944d470d863f6771521fb26aeb2e55b
a65225b4d777a0389ad826a2277022540d188f2b
1964 F20110115_AADBEA dunsmore_l_Page_073.txt
a9e6abc2f905b3a36f8d67e2a02ee452
e9ac30d2767ac20fb0a38e4291ece8c5fc0536a9
844 F20110115_AADBDM dunsmore_l_Page_059.txt
ab78931788a0fd4f8dc249cef72ffc01
9d98823196d483ad8636325ecee0b1fc3286ffbd
66163 F20110115_AADAAK dunsmore_l_Page_018.jpg
1d6e492bf5f5f2dc69f2834582dc6366
2a5ddbaf5dace6fee6c394113557a1e3a4b77293
74209 F20110115_AACZZD dunsmore_l_Page_115.jpg
6975723864ab7897abe257cdb873ee80
a3d7466a0d3ac87ee57f50bb57523af4adc268d4
50071 F20110115_AADAYF dunsmore_l_Page_101.pro
9956aa726596d0593ba7b8db3c72f99d
93e170b1f21164ab60332e8fc744bed15f0f106f
2023 F20110115_AADBCX dunsmore_l_Page_044.txt
86e874e6d6fa3d2857390299d15c749f
4ec6be8f46de26eb4725733923525b3a7a5d79aa
30150 F20110115_AADAXR dunsmore_l_Page_087.pro
0e2b1f05f5874e5aa3f9d2fc036aed41
281a2ebef2e9c8b38ae3b05d3e353e1c248b5bf7
2073 F20110115_AADBEB dunsmore_l_Page_074.txt
8b0d0b24a0ef27ed2f334c5de4397cdc
f0ab74eeec97d2852b48786f65173d77f0e0d880
408 F20110115_AADBDN dunsmore_l_Page_060.txt
cf445a08bb2ec65f9fd8602de493a89e
259f352f753b6ee14715ca8f97e4e8659edba03c
84615 F20110115_AACZZE dunsmore_l_Page_166.jpg
d8663ccaf38934bd812356f8c8486320
d57a01a474d3f07b1eee980527d4051067cec8f9
47833 F20110115_AADAYG dunsmore_l_Page_102.pro
d377e9adb7ea7ede511a3b3fd217733a
3e7871c62de90bd48a734270bb74c69efed7878f
1983 F20110115_AADBCY dunsmore_l_Page_045.txt
fa6318480923f358ea459fa1af81d328
7c7fcb01b68f8cd081013843a708d23d1a6b1c68
41332 F20110115_AADAXS dunsmore_l_Page_088.pro
8057e54c85b8109ad90e91a061f5ef2d
73ad2e944ddfc63d0a41c081945f19317128197e
73666 F20110115_AADAAL dunsmore_l_Page_019.jpg
69a29201cbf00a07abdcbc565eb12563
be5b7673c8b12fc7f6e24afa811eec9d6f425eb2
2000 F20110115_AADBEC dunsmore_l_Page_075.txt
660095fac59e7cc1cdf307b2e7b5507f
56b805899a7e224786f7296af265640a196fa432
1391 F20110115_AADBDO dunsmore_l_Page_061.txt
e719e1ff9331afefceeb68e7a9437bdf
e578ab194a51f2dc45804d9dfb8a062f33621c05
6292 F20110115_AACZZF dunsmore_l_Page_091thm.jpg
04a803acfed401e854e128638099587d
99257812cafe067fb1dd1983de6ec1a41362190a
51936 F20110115_AADAYH dunsmore_l_Page_103.pro
a56f5b7709e6c34811ce02d9b6233083
06994fa5fba25e2424d6a709caa6169681dcbdc0
2097 F20110115_AADBCZ dunsmore_l_Page_046.txt
ce8e47fc4d648d5e2f869976eba42966
9a1398f4447bd7177b4327fb68f5aa7a7959c0fc
48715 F20110115_AADAXT dunsmore_l_Page_089.pro
545a306e675e923daf4238890b50c96c
87163283329228cbc88807fe3e4323032829cd88
73219 F20110115_AADABA dunsmore_l_Page_034.jpg
28c05ff1d7ef5642d520dae2d2ed60e2
dd7c2dace64c138e60e2cef67ccacdfcbd4e396b
73656 F20110115_AADAAM dunsmore_l_Page_020.jpg
4e4d2cf67635db963206b78fd5632793
03e3078b51a96aec52814248771e13428d566369
1972 F20110115_AADBED dunsmore_l_Page_076.txt
b0d71fe6b7d642397d5668394bcbc814
3858079b3ae037e6ca7107d79020090024211d7b
51640 F20110115_AADAYI dunsmore_l_Page_104.pro
aceaa5dc1c04b0149d38c503912ea38c
1cb4d2079cfd49c11da42db9adab9033b8854b87
33295 F20110115_AADAXU dunsmore_l_Page_090.pro
90fde4e3c7873b74be22de8adb3e8a61
395578a074f3ab0cea41f7ac3d9fc1385ca961b1
69213 F20110115_AADABB dunsmore_l_Page_035.jpg
b992831baa1ca4ee9aec2d8c036646e1
63ce8caa7b22c63653265c08aa5feb20d364f561
2054 F20110115_AACZZG dunsmore_l_Page_021.txt
61acfb6d5916a09b89ce2bd46ccee121
4b02a52787dafb392ddb502b98a73b5a15b2a657
1890 F20110115_AADBEE dunsmore_l_Page_077.txt
b31b6122c0bdca6f8d63fa788d70476d
4faf934a12242a9040caa63213e37babd6afc171
1917 F20110115_AADBDP dunsmore_l_Page_062.txt
c0efcce35c60e03d9b118fc518572542
f9a59f4e9c52f56e3e3d74533b64312d95aa5c85
52853 F20110115_AADAYJ dunsmore_l_Page_105.pro
2de079516458ac6dce73e17f1a3a2628
78db1b002851375161247eeecf9ede64f6632961
49706 F20110115_AADAXV dunsmore_l_Page_091.pro
4ef09654b659418a7c298fe48840b1bb
84efbc3c9280b23a17eb8c3e78eb9164e9355b1b
70705 F20110115_AADABC dunsmore_l_Page_036.jpg
2acd77b944d2a6b0d3b215e37944b587
28975a592651d587b118acee0586bcef01517d9f
74799 F20110115_AADAAN dunsmore_l_Page_021.jpg
f2f6ce2cdd12031fb514903eb54452ec
4cb1bc9ab6e8f43f2ae9cc63ca39add5b62ae531
1053954 F20110115_AACZZH dunsmore_l_Page_140.tif
cb2200c269a18ef1423219f6be5df56e
995b08f5ef465df8e627ca7335cb47b14038cd6d
2028 F20110115_AADBEF dunsmore_l_Page_078.txt
7517101b50dd54fe2de95c74c0755e15
37ed1cb5610a48906bec44af02db4658a5166ca6
1966 F20110115_AADBDQ dunsmore_l_Page_063.txt
eef6e9e7d1052beff7321643ea7746e2
fbc5d7a1a4bd16624a6e2fbd039528cbe3a811c1
51174 F20110115_AADAYK dunsmore_l_Page_106.pro
e232f8b8a762c883c4c7b4e7c9fa5fe7
51d847e0a8691aa9faf84c56039d74da087a3f67
48937 F20110115_AADAXW dunsmore_l_Page_092.pro
3ef5cb953b571abb28551213566da2cc
66b39b099fca615cd3724293e70a08b6b81cc04e
68518 F20110115_AADABD dunsmore_l_Page_037.jpg
f1fd03af845bb1161bb087eb02b8f0d9
c1c721e2a918bbb2ba835770b4798d6157f02a75
70491 F20110115_AADAAO dunsmore_l_Page_022.jpg
2409eca55f2ca65f8fe211d542c33943
7a025e64dafd06b8ae7637f79cb454929a31646c
22739 F20110115_AACZZI dunsmore_l_Page_102.QC.jpg
8df7f5c5ce44d18283d79543f4a123e2
748d1c0d447a8c17c1cfb62a086135208b0bc489
F20110115_AADBEG dunsmore_l_Page_079.txt
787e724edc9b8a259a2938d4ff29a3af
b75b041ecd72a5bd950f07bfadf2e65ef80228c2
F20110115_AADBDR dunsmore_l_Page_064.txt
647d5f46dddbd338a8a2c730c1c08c42
3b904d90ceebe167affb1868ddf8008b20c1630a
51033 F20110115_AADAYL dunsmore_l_Page_107.pro
85d46bd018f7596ba8bd36aa39fd9d0b
c9dd6dc4261f531a55e5d709a7bb80dd1a59ee14
64637 F20110115_AADAXX dunsmore_l_Page_093.pro
5ceb5de71ffa8c0d51c70df98009cea5
2c32f4a29a153398cbed48df30514909c6ab4b4f
71256 F20110115_AADABE dunsmore_l_Page_038.jpg
214486afe022f7c1d8e9eff02797cbe7
bcd3e1f334697221870c14dd4c84967775a40fbf
71973 F20110115_AADAAP dunsmore_l_Page_023.jpg
b7bc7e118e72b173d35f99d65e724048
27806aed19baeeb03994ae6bf736b10d1e9608a3
112954 F20110115_AACZZJ dunsmore_l_Page_107.jp2
601114d8b07691251aecec642b599f93
a82258517f977f12d3d15cd669309f952c45677d
51752 F20110115_AADAZA dunsmore_l_Page_122.pro
519a31e5de623f2882d9f918544a2211
d488ecb25418d511ba54e21896ba9b02ffef5880
2094 F20110115_AADBEH dunsmore_l_Page_080.txt
85b174d53d215013743868acf6cbb983
731be6b05b213fdb9e7072627a0b58b317d641e2
F20110115_AADBDS dunsmore_l_Page_065.txt
450cbd82f279e66acfb50b0b8eb26771
a5b0d291cf440b30024f11a7a437f8cc76c28786
52825 F20110115_AADAYM dunsmore_l_Page_108.pro
6a0a54218fb2b6677756130fc2c2e8a3
435f9d789977c27a192a7339449dc8bb77a0940c
43300 F20110115_AADAXY dunsmore_l_Page_094.pro
b1b64262058244c0b6cebff46899a447
95fbacee8811cc0c208979d674884721894a7a0c
73742 F20110115_AADABF dunsmore_l_Page_039.jpg
2305a7cc7115e90c7cd81d1984353b82
4f75ca935296f6d50217153856b173ef93ded52a
71795 F20110115_AADAAQ dunsmore_l_Page_024.jpg
71d0e1eefa0d4db7bf18e535856cdbf0
386ddab7050506ff11565f5938044a82e299ebf6
48778 F20110115_AACZZK dunsmore_l_Page_068.pro
acca8d1e18d886214b24f0b462b41fe7
cb010904563ea1ea87922fa826e42c515929c951
31555 F20110115_AADAZB dunsmore_l_Page_123.pro
9dcc6c1b949af3278efbd6a752e485db
bfc3b4d3ae931fb610a9ceaf8e615391ba03cb52
2072 F20110115_AADBEI dunsmore_l_Page_081.txt
f878803453f84f398751a97bba5972b5
27cd113e0efa12fac4cf0dfd8f96f2febdb96486
1680 F20110115_AADBDT dunsmore_l_Page_066.txt
31af658dc21faaf8ab7e0017d4ab8095
7aabbbd4796d9260fb233e54400888d1cf32c636
49927 F20110115_AADAYN dunsmore_l_Page_109.pro
ecea66c2ada92e47f5a0f4f4ee57e0c8
776eda74e073a1bab66cb01f4fadac3105e941da
43398 F20110115_AADAXZ dunsmore_l_Page_095.pro
1b73cc8404734d88ce5c21932892b5d9
08a25c2687b34842b4d8eedc6c9bae72dea89451
72622 F20110115_AADABG dunsmore_l_Page_040.jpg
6c3fe0e3ccfef0bcfd78318e18f48f42
3612de31ec39f4349d12dc77471d1010d60ba49f
70721 F20110115_AADAAR dunsmore_l_Page_025.jpg
f2f53a793f02635b20d87c958e0fa123
5365fb4ef2a2f77b8e002c943685acebe288ac44
6556 F20110115_AACZZL dunsmore_l_Page_146thm.jpg
ba4ee73bba58e207923c55ebdc3a6a2c
883e9fd769bdea40cf777a17f0c543784f16b922
5279 F20110115_AADAZC dunsmore_l_Page_124.pro
b44cf9dc81bec6f6906a2311c323bb50
870ccfb8ca3404d517335991eef092e79c1a34b8
369 F20110115_AADBEJ dunsmore_l_Page_082.txt
982f2d44551385d292633b44a45e75c8
ba8ef10263c324871ab226b266c7d17f6aad6c98
2048 F20110115_AADBDU dunsmore_l_Page_067.txt
aabc58b741bbd027e843ef077113c53d
086c86d7768332b6e588c6ab67aff84540250eb4
45379 F20110115_AADAYO dunsmore_l_Page_110.pro
c02ac98c0194dbc1d82e1b6a871ca2dc
b0241e0ea9fdcb1744a089bccc09d215e992fc8b
73640 F20110115_AADABH dunsmore_l_Page_041.jpg
2d80a18f0a1b35b6fe0707a8bd860020
52ebe1f798225ec9acf9b940f3060dd2d0ed8ce2
72640 F20110115_AADAAS dunsmore_l_Page_026.jpg
3e13c33b50e0caa0e6e2b8f16406f16d
a5a99f6f76a0925a169edad5dc53d5c4385bfd6a
23342 F20110115_AADAZD dunsmore_l_Page_125.pro
3f3cae64c1f09f8bac6717a3a58f494a
cae9c660983748624043416b6898008f988a96e8
260 F20110115_AADBEK dunsmore_l_Page_083.txt
9a50a25a0c8d4d11a2748739054c2249
828b30619cb8ac52ff955571e3cf26d26faa033c
1920 F20110115_AADBDV dunsmore_l_Page_068.txt
11cb448f8d55d9a825d3db470a5e711d
3f50a394ce768f03f5c3dc059e70868453b6a518
51698 F20110115_AADAYP dunsmore_l_Page_111.pro
c71e0a41de7ee5b44582a064aa40497a
c3f85d6aa64c9bf2f819f87b22e8e5741481cfb0
72627 F20110115_AADABI dunsmore_l_Page_042.jpg
9421654ff07e79e35b1cb8d99e8e20d6
b8cee64f8406d1b94b026b8762a9f50ef3c3b5dd
60556 F20110115_AADAAT dunsmore_l_Page_027.jpg
44160ef43fdf2e97278c0478a1eed834
c6a16167d523d3c2c37c0355d95be95b490717c1
F20110115_AACZZM dunsmore_l_Page_179.tif
33bc15f3e948040ebb716e8c9bb3e560
38af8f88990eac0b023da1b915ad0ea0a3db2bb6
10766 F20110115_AADAZE dunsmore_l_Page_126.pro
8a0d26d8cfd8541765196b2e3abb8c38
44e650f75d8e44c28ef685fe91859aca26291d58
567 F20110115_AADBEL dunsmore_l_Page_084.txt
bed7f3041cba1b6b86dd57bdebb501d5
423fa2f4a98793e75745da45da45bc331f31e3bf
F20110115_AADBDW dunsmore_l_Page_069.txt
b6aed55f50bf7ae0cd10bfdcba70e6ae
99a9a45cccace700893be43678fc1015f92f4327
48939 F20110115_AADAYQ dunsmore_l_Page_112.pro
ae8a985cf8d9b5bcf9486876e9f1855b
7d19d04ca99b56bd6c2bfc566c7ebfb168afaf2a
69187 F20110115_AADABJ dunsmore_l_Page_043.jpg
e13b6f3c87ed2899321f1bc51c6456e5
140c211d1c9e6e881e94336625ade7d8d25fe970
66383 F20110115_AADAAU dunsmore_l_Page_028.jpg
ef9ea4f1cbc0bf72acefd50defc659ff
47c43d27f0d3a7bdd3e5c229f08b28725f33a0c3
51735 F20110115_AACZZN dunsmore_l_Page_141.pro
a57011ff8b175868ac4b22623b46a171
0b5d7c3740c25e3e88cd2e7175c9d0427f2930e0
9528 F20110115_AADAZF dunsmore_l_Page_127.pro
da605ab0b6db3a19889f77baebed299c
05547e283d6f3fdd040caf60290efd3b70d8f2e4
2002 F20110115_AADBFA dunsmore_l_Page_100.txt
ae85b1cccf313460a4eaca2c78f8b869
fa20edd79c83dac54ad2cf1f6545d9026f0aabe2
1099 F20110115_AADBEM dunsmore_l_Page_085.txt
04c258fb208117b156f79fd248a36209
82f335b9c23c4648c1cb24107fb9e177e6c7af53
1783 F20110115_AADBDX dunsmore_l_Page_070.txt
7213a43ed8e6695cba068e28a7e8b172
4adc083450367ce066b8cc9a2445147e0fde71ed
50462 F20110115_AADAYR dunsmore_l_Page_113.pro
12a4f5d3fa12aa4dde8cf8210de5f7b2
a3d9f4d3c72ee955ab4c0f50daee472581d38d6b
73427 F20110115_AADABK dunsmore_l_Page_044.jpg
d1b8317ba77af9dcb1f26297c242c427
65e5dab7ff2060032b00059e4e553d5d0d30403a
73254 F20110115_AADAAV dunsmore_l_Page_029.jpg
e86da811e0542afd27cb4ca43bf8db23
b23db34b3309c5aa665da76bbdb09cb6a0715f76
25271604 F20110115_AACZZO dunsmore_l_Page_056.tif
2b8e10d918aa4936fb624b212d43f2c0
1c2adf0fd85aa2dd24f0d86017a854a110d89522
F20110115_AADBFB dunsmore_l_Page_101.txt
c791621704c16dc6417ede207f4c1791
9d07b5410480ce88b9287c1a6951b5e175f50a4b
785 F20110115_AADBEN dunsmore_l_Page_086.txt
95e4a7f8df8a36cd19affab5e1367b02
fedc1d1da87cda966a5ab487b2e020667b70a1c9
1814 F20110115_AADBDY dunsmore_l_Page_071.txt
54ba5bf9557e320dec9af57eaf1e09dc
e9a552474d926e9d19487c5dbd13ab8d8339f904
49940 F20110115_AADAYS dunsmore_l_Page_114.pro
979d3e351e720403b43001ca864fc15d
a590777e9de05ba7446f9e07520aab379353d4e0
72071 F20110115_AADABL dunsmore_l_Page_045.jpg
ac8a428095cefc25df89f173d95d01f6
3d7d945ad5289eb96d90ff694199fc7db0fa0ffd
74115 F20110115_AADAAW dunsmore_l_Page_030.jpg
d73b44dc0d7a3ebe430c19366dbc6108
4a03e05fc2b0a41cad5d62981324c61f9f221317
1970 F20110115_AACZZP dunsmore_l_Page_097.txt
a5f73bc18b7aca6b523ca86b9984a46c
2eddf792e7439830c7c8cac57a82a8d1263c7846
16683 F20110115_AADAZG dunsmore_l_Page_128.pro
31dc9195473809d47c33aaf535fa4444
c31ec32168cb485a75e8eae633a11e58c5bba024
1892 F20110115_AADBFC dunsmore_l_Page_102.txt
8ee695663304815df63fec35abe3a663
18cfac230e5cab242455b3bf6aab31cfbf6f4f12
1750 F20110115_AADBEO dunsmore_l_Page_087.txt
b918266432e0e7b3a75e4244faf11815
ea490526c606689e9c0e98f3ddef67e86e387dd9
1785 F20110115_AADBDZ dunsmore_l_Page_072.txt
26c83da98cbe2f83657fd0c76d238028
5546171c636943ec17b8830286d74358f7d1be1c
52011 F20110115_AADAYT dunsmore_l_Page_115.pro
cfae14f2baa197bbb3b941d440cfb35f
88d8d6f49b7d43ba6ffb4ff0b1cc4e47a4be62a7
27328 F20110115_AADACA dunsmore_l_Page_060.jpg
02eef0d81f8f8f78c4de4cda39c2272e
4aa2e05a5441c42d6db4af4b3bd7338a2d355d62
73257 F20110115_AADABM dunsmore_l_Page_046.jpg
d976c50ea0af025896914a838c603678
674738e62a6337510a55e8309c4792ed8e6c129e
70905 F20110115_AADAAX dunsmore_l_Page_031.jpg
1083109b6f7fc1e85672ba09e6aaf0f6
6fa6daed54c9e06f10d89243f008b97d5ebadc3a
210042 F20110115_AACZZQ UFE0008350_00001.mets FULL
8301c756d712150db8361d746cffb054
aadc53dfb642463c3b95a2aa5349c49a8c15c550
8964 F20110115_AADAZH dunsmore_l_Page_129.pro
6d4bd07d542b38ba8abeec6d81da8c81
b593611e2d4256c537507c711dbd776198e6f2ba
2039 F20110115_AADBFD dunsmore_l_Page_103.txt
560130f3cf8de35ce8d4c4f3fcd59048
30d0a03ee43c6ef5fdd0c47b4f74e2776a8eeca2
1720 F20110115_AADBEP dunsmore_l_Page_088.txt
3f7faaf304236525cff23faffff8f26d
9dfbd8e3e72a020869df07bc573c48c86ef6ae8a
52063 F20110115_AADAYU dunsmore_l_Page_116.pro
8d3472d38ea4f4c0edf08cf70e85cbea
8e8deb3968648f3b30da326c5ecb39ac44258d62
52452 F20110115_AADACB dunsmore_l_Page_061.jpg
e7bc0c4252561364a9e110a8e5345f9a
9e4a184b7a76b9cdf474570a94776c5301068b09
27063 F20110115_AADABN dunsmore_l_Page_047.jpg
5538ac57d0df3b08510cef9b3138c35c
0b01ab5cf97bfbb9ee8395b4db7c1bdc011623df
74347 F20110115_AADAAY dunsmore_l_Page_032.jpg
e534ff04a6f89c1e021b81698f0552e2
e14882b130892ede572c0d4e4b3bf76f8f490180
10680 F20110115_AADAZI dunsmore_l_Page_130.pro
f0a7e76a488a24261c97f6ab1de2327f
9043d9345213772f8177e9a91f5daf485abd8a88
2033 F20110115_AADBFE dunsmore_l_Page_104.txt
d80d40462c52d0e86355e629f71d6df9
7dc2158fef39e0ced16fa4ddb5e1e156883cf904
52600 F20110115_AADAYV dunsmore_l_Page_117.pro
61b792d9925a72958eeb23d65870241d
21a3cd644f2c66968825945b6c64c3911626d3db
69574 F20110115_AADACC dunsmore_l_Page_062.jpg
e05a61f16620195a0c204d630ad3cef2
7129f5252b232465ecf5a7539412676304bdb4c4
70813 F20110115_AADAAZ dunsmore_l_Page_033.jpg
6f91ae3af5effbb3e6c97cbedd95c80c
2a7daa3fa8e2b601ddd36e070049d11e41edb707
22790 F20110115_AADAZJ dunsmore_l_Page_131.pro
0bb8a7c48496e28cbeb94177206679eb
4beb8b8b615817075901337e605e889d4f376014
2067 F20110115_AADBFF dunsmore_l_Page_105.txt
b355de041679efcdb45530657c4b81cb
718ddaa8a7ecaac8516666f0509eff720580514c
2185 F20110115_AADBEQ dunsmore_l_Page_089.txt
547f1a4c765fefa0d7865d3b9c4a6c40
811fc4acf815205c098eecf22c3ae58828b19693
50016 F20110115_AADAYW dunsmore_l_Page_118.pro
413b72996d15d5095f2c47d3a4310c81
8a318fe66e548ff08eecf512751b911e31dc7048
69218 F20110115_AADACD dunsmore_l_Page_063.jpg
3d31c21b7a5707dd5bdfdb989993e420
40ddbdf6b81bb413e8d0b517781746a0b9c97784
44337 F20110115_AADABO dunsmore_l_Page_048.jpg
d2df369a50d98c3b58cbbfe62239ccb3
2c1d3f497fc32763a3a9ba0f3b3547dec6fa26f3
26922 F20110115_AACZZT dunsmore_l_Page_001.jpg
07bcaffb5cd3cab5f5a5284f78c86b09
ff1d3c49db862ee83ec7b99bfcd2f31d2109de03
45694 F20110115_AADAZK dunsmore_l_Page_132.pro
693dbac755966a2af4e36c1c2061af30
083ad08ce1d56752675ee59d4e414751d211f8e6
2013 F20110115_AADBFG dunsmore_l_Page_106.txt
ff3e7c0bdcae25a3f93832bb9bf6cc38
9d252490d766ad1377911943f5ec8a14dc040c01
1693 F20110115_AADBER dunsmore_l_Page_090.txt
83d7d4f9e715bd3b748fedfcaaaffc08
8fc0d7e28430bbb452df4c8a91d667d6da085a3f
50874 F20110115_AADAYX dunsmore_l_Page_119.pro
f0a3d299aa51556dd01f186f05d47050
7432bcc6b0a88153da5749163d17b48dee4ecce2
74752 F20110115_AADACE dunsmore_l_Page_064.jpg
4c0eb027604fe0d5001e520dbab54958
c86b078eee50a59c4c0cf410fb386088cda16371
66014 F20110115_AADABP dunsmore_l_Page_049.jpg
45309cb02635004e23a101bf1dd3fcce
04d9c7b87d2b25dac9b008125f42eed73c645f49
11154 F20110115_AACZZU dunsmore_l_Page_002.jpg
61c31712998a6906536e2bffc11d0e4c
94fefe9f1d8e8e5ce6cd2e456ad18bb886606816
26837 F20110115_AADAZL dunsmore_l_Page_133.pro
23aa410dc71ffd1106f7a80d303f1e14
87c60435b02021260a207009bbb7d09f6e4a60d8
2003 F20110115_AADBFH dunsmore_l_Page_107.txt
6aa4f27081a1b2cb83f14f8c7154acfa
e0cd1d59d04344fbb19fba7fd1432c2b99eaa744
2113 F20110115_AADBES dunsmore_l_Page_091.txt
02f166a542ac3f7553ad84fa2864ab31
16311f8282cb34f930b9b72e7e9958383eb26a6a
47141 F20110115_AADAYY dunsmore_l_Page_120.pro
fa286de91232c6b641c77c51829a4686
1f08f0a47d3787a71d72bfab8858ab69725d2b95
72438 F20110115_AADACF dunsmore_l_Page_065.jpg
8f036d70f1b00f385c73bc2f7a8b8314
5c7095a4999630b2656f2c89b2a47145fa8cddd2
97883 F20110115_AADABQ dunsmore_l_Page_050.jpg
3d151496d1e88383f7ccfa936baca8d8
efe4f9a4ab155fc45dcb119daa15686004193ed8
15466 F20110115_AACZZV dunsmore_l_Page_003.jpg
64fe088c99587e6212fccb8a7081e7e8
6308a62e856e487e23f1fa92fc7832c27c31ce23
41166 F20110115_AADAZM dunsmore_l_Page_134.pro
941d5fe7d21251a0acd42f385df53427
a0f807655f1529d99d4381002820f02308277b4a
F20110115_AADBFI dunsmore_l_Page_108.txt
94ec3d32e723f8d0b64db1c3e9fc1cc2
13b8e1cb95642964d566bb2ab4d8597de39cc2ae
2119 F20110115_AADBET dunsmore_l_Page_092.txt
638cebc505583b833e5a933a9c03541f
3df4950da5b5f0e97391b66e581fd6079e9a49f1
49926 F20110115_AADAYZ dunsmore_l_Page_121.pro
58e502a91866ada289efc4db593788df
22165c063124d79428994768ffe95a2e30ec210c
60529 F20110115_AADACG dunsmore_l_Page_066.jpg
28bfe547f7db6b94b0b0fffb31874bcf
27df00a45c0a1e6e7c228f030f19f5f653124f7b
57546 F20110115_AADABR dunsmore_l_Page_051.jpg
05a2b9b16354c72a030ee9327b795fce
28a5db80b53374a752dbda600d20634430e167e8
63037 F20110115_AACZZW dunsmore_l_Page_004.jpg
6c28ccb13772496af8b6e7abdf68b126
a8c98c3901f868fad547a3b56584dadf023a6212
19213 F20110115_AADAZN dunsmore_l_Page_135.pro
00d59c1690aca218cc5fb8dbaa6a0ec7
a21d3afc8e83eebb94f565d4560e546a918fc107
1968 F20110115_AADBFJ dunsmore_l_Page_109.txt
50ebb0c7f64d63867fd535edbd341929
841620209b4bf434fbcfcba358651bb850856265
3021 F20110115_AADBEU dunsmore_l_Page_093.txt
3a9da5413332994a0e5fb71afc5e65cf
41a7068aee8e02a180c3d89feb34a2a9c71030ab
72727 F20110115_AADACH dunsmore_l_Page_067.jpg
81521429c11b94e7bd5d43ea50401c2e
92f1706555dc1e86ee346d41c17dddd78872f97f
38391 F20110115_AADABS dunsmore_l_Page_052.jpg
e3aadbbb2da521951af322db5d572d34
64dfefdb1cb3b9bb745916fc5152178c33a08617
67438 F20110115_AACZZX dunsmore_l_Page_005.jpg
4b482ba019f4fee59ba212dedb00c34d
97a7af1573e40dfb09cc43e386481dc8986ed9ed
30667 F20110115_AADAZO dunsmore_l_Page_136.pro
64a7c1f0a60450485d12e5a64bf65c14
b470636aecf43bef972bd3e5770e3e528a275934
1794 F20110115_AADBFK dunsmore_l_Page_110.txt
e2081faecdea33807d0a6a37b231f05e
3137cc1417d6aec4460cc0d01eba1f294e28abf0
1935 F20110115_AADBEV dunsmore_l_Page_094.txt
528aa978068e5036b17d74f5665c9b10
8df1d34662bc2ff8247f066a6a95e9c1f442d031
68948 F20110115_AADACI dunsmore_l_Page_068.jpg
336ec973c8d0d34924aaa09f6b98f5b1
959fc4aec7a87511de7790845c9d5b0627e4d03b
42209 F20110115_AADABT dunsmore_l_Page_053.jpg
55b5b6a6affb36ebd0c878650bf2a522
d849bb55b858c5cc795df9b6b4db18870af4f923
71024 F20110115_AACZZY dunsmore_l_Page_006.jpg
d5609e41fe013dd57ca26af14454f452
b1343b74e088712f31db8f6d5737bc139d24e920
44869 F20110115_AADAZP dunsmore_l_Page_137.pro
c38f5427ba9b661054fc84b4430216bd
01570d7eb7c12ebb163fb3ebcc0c68ebc5cc52e4
2034 F20110115_AADBFL dunsmore_l_Page_111.txt
e3458899f30a9a2f3cce3bab741d5c38
63d85d06b85d5ebbdd3b43c7dc3b43379f050df4
F20110115_AADBEW dunsmore_l_Page_095.txt
8f552f34e901f8f922077a884e11733d
49fd080540ba458ca4a7aed3d2ed9dca75031dec
72819 F20110115_AADACJ dunsmore_l_Page_069.jpg
392bf209f3c642c67bed2761006263d6
423df6c957ffc09fddf6f0a47fb3b0e53c0eb6c5
54943 F20110115_AADABU dunsmore_l_Page_054.jpg
390562536d09c10335889ab67c0ef29f
e6ddbd9a08b00d6f70d4b8c1f069ef772ddb00d6
71117 F20110115_AACZZZ dunsmore_l_Page_007.jpg
99bd7404efb3ab22e6892772659c6874
7bf0a6b9af9d43465ed414f3e4c9191d96d59da0
51269 F20110115_AADAZQ dunsmore_l_Page_138.pro
cd718a34406d6d5b347f996ff5085508
bb16c596c868e75ad6f6629a3697794b04f2b2a9
543 F20110115_AADBGA dunsmore_l_Page_126.txt
8cfc799704f428ec108d1f274b904916
fc007874ff8b59722d6390a850429e0ea78de440
1929 F20110115_AADBFM dunsmore_l_Page_112.txt
880920ca266ba3850c7ca9807809fb43
1a2cf0d3958da588b7e39af8b65dab90519bfbed
1925 F20110115_AADBEX dunsmore_l_Page_096.txt
30c2bb1bfe6b205dac5b7afe56a4655c
7ccde212ddaf14ad9146773f3f5814222e2641ad
62787 F20110115_AADACK dunsmore_l_Page_070.jpg
6ba01444b9d269de86ed76c2bd6a8a16
d2f5489e9a88c62b25e08b21b3603d5eb8c051ba
60536 F20110115_AADABV dunsmore_l_Page_055.jpg
271febb5827d79b542d8c189cbe3431a
2b2d9066d971bc53b98381ce7a8177ea761d4059
51307 F20110115_AADAZR dunsmore_l_Page_139.pro
25204a3361f9211c4a09c31d83dc2959
674da2e6f76b0f15396525dbd812c4384490ac62
532 F20110115_AADBGB dunsmore_l_Page_127.txt
252ebabc9d71a272947df1cab5e9513f
1bfe3f147f72319c79e4f3cabf156182e6353ed1
1993 F20110115_AADBFN dunsmore_l_Page_113.txt
0c1967ad0bddfaff2922f886952f3740
ea1cbe9d771d1710afe17362103c1fc98831e137
2068 F20110115_AADBEY dunsmore_l_Page_098.txt
e70f566dea8b6558259045ffbcce4a98
34991e7db97e5c715ec12bd01d1890a91d638318
67658 F20110115_AADACL dunsmore_l_Page_071.jpg
616afa800a48791abd524056bdc183b3
456a96ef1272915bd724d99927e4dfc669ca84ad
60401 F20110115_AADABW dunsmore_l_Page_056.jpg
100a785806a678f26d9929998ffcb759
5bc1d066fd50e6dd1fb73fb79aeca105d6df5880
50489 F20110115_AADAZS dunsmore_l_Page_140.pro
72c2c80655a0ee2df0d9800df4a218db
2e5792bd83809b2298fcbdfb3d532d88dc7f3673
833 F20110115_AADBGC dunsmore_l_Page_128.txt
38c5dd46109a4cfd75609eb837178e75
3529dd212bf3826ff644fe2a3e42210927eb8e8e
1999 F20110115_AADBFO dunsmore_l_Page_114.txt
d33372a0b740896717cbd8545d56cddb
47439f1fcf3af3069a8ececa53d625448e1e69e8
F20110115_AADBEZ dunsmore_l_Page_099.txt
d6ccaef592a5d7a826087efeaf9b4b64
12e42354850492c6cef76fe5d1532fe98ef8da08
65639 F20110115_AADACM dunsmore_l_Page_072.jpg
2b1aca03e5b7e1ddce03dae50315b7af
caf9a37baee9369c01b2c623986f13ba02e02ca4
32150 F20110115_AADABX dunsmore_l_Page_057.jpg
58cefb7ca6ffea1e222b7a4ca627e631
14edb838cbd0f93ba9c4c13da88a195d89402815
50486 F20110115_AADAZT dunsmore_l_Page_142.pro
482cb037c9d18936130dc758802a4226
a116a39aff73290beb5c955368524053a9a18de8
32212 F20110115_AADADA dunsmore_l_Page_086.jpg
5326d080d48e9a4458e5c34f3da3f9a8
0085aae4b563aff1dc847885f1c3c27ea7bc6bd3
449 F20110115_AADBGD dunsmore_l_Page_129.txt
960b20b3ad5895e4ef0c3f6f29f49b92
2252f57d985d94176c651c3cfeaddb622dcc1dc6
F20110115_AADBFP dunsmore_l_Page_115.txt
7f541054b4f71c6ba70af0509ffad2e5
c28d2b851fbfe975077f6f64f5046697fde0a3d9
71572 F20110115_AADACN dunsmore_l_Page_073.jpg
dba2ad98cc30080dc752fd27334b175f
677100099e15031cd1e799962fcb3c2fb1d5761d
29119 F20110115_AADABY dunsmore_l_Page_058.jpg
1fff55dee9357755f8a5ef144500bd34
02c9d86fa111cf23b43b1273963c6243a529bebb
53212 F20110115_AADAZU dunsmore_l_Page_143.pro
ae5e3a4e256365f8993abc7cbed4c49f
6b4600f367259de814eb9a5d21f23e25e50f8170
44911 F20110115_AADADB dunsmore_l_Page_087.jpg
375ad93a1d5e68ca4ee5db8a74cffeb8
9a5f5857a433e99ab7aaaddd0d6bdc7d6a9b564c
462 F20110115_AADBGE dunsmore_l_Page_130.txt
cf4d4bebf4efbddd593e029c9eb5caf0
ad2add5105365d8caa69a1e6e5916fa49436dd5c
F20110115_AADBFQ dunsmore_l_Page_116.txt
03309693d9f6439495f66a873fb87873
9ed7e64bc4a91708198dc38a2e09541fdd330f31
74643 F20110115_AADACO dunsmore_l_Page_074.jpg
72652d0aba8d1f00b467b02b3ab4179f
389bd060bd9811ba0b2b8b16ff47e3ad785c6760
28860 F20110115_AADABZ dunsmore_l_Page_059.jpg
07187f889e970ebbfd1438fa9d9faff2
e619927f156d37725f59a3241d69176d986ccbd4
50507 F20110115_AADAZV dunsmore_l_Page_144.pro
9432f96cad2ae38123fee3c2255818a3
e357d76500a90940bbe8b8418360ab56c00cd534
69159 F20110115_AADADC dunsmore_l_Page_088.jpg
eba8b3db3e1f40dca1378a948aac75f4
cf3517d31bc6c251d043fccdbb1fd0c659bf6c95
1043 F20110115_AADBGF dunsmore_l_Page_131.txt
dc5811c40564ca32039e69f4024cd92f
23191c6b0573d4ce90121d84929d4eec5cad40c2
53435 F20110115_AADAZW dunsmore_l_Page_145.pro
c61fc243b5db0eb46d4eccc84b8e56b4
ccfadd5af6e3697e90dd557ebf6a19fbeeda8fb2
77638 F20110115_AADADD dunsmore_l_Page_089.jpg
419b2b4c6520aa547ffd56b2b456b4d7
df89a26aa66427602a2c64493e023c15abe041b6
2320 F20110115_AADBGG dunsmore_l_Page_132.txt
fd635dd1722a1ed78bcd1992bcd0ebc9
2756c1d0ffeee7cb527f8dcadd484db06ba816ef
2061 F20110115_AADBFR dunsmore_l_Page_117.txt
99d0c65d9c17f0fab836f5cafbddbc84
1d8402e810c5e0430aad8b6a6a276bb851de0970
73330 F20110115_AADACP dunsmore_l_Page_075.jpg
b3276aec011565fc9a635b2d6f153170
5e4a118d7b99dfa5951d8598ef00f385f89f47fd
53004 F20110115_AADAZX dunsmore_l_Page_146.pro
1f74e05e3d201bf250fe1e283b9e7a11
2a8fbae11111068b8dbb76ccce94b319d5dcebbd
54700 F20110115_AADADE dunsmore_l_Page_090.jpg
959a121b51299e992e1936e07bafe912
806f5533c9d223da3bb6cd25eda69f141d97c4ba
1250 F20110115_AADBGH dunsmore_l_Page_133.txt
d94c33244779cefeb60eda3a44606169
929ff630846bedc6d02698ee6c32d7b0697f6c34
F20110115_AADBFS dunsmore_l_Page_118.txt
b9f469a06914b18ca604bbaf3e1dbd15
b432923aff265bf7d1bd19a9b840973249dda559
71324 F20110115_AADACQ dunsmore_l_Page_076.jpg
0616791737a92a847615dfa5f8e74275
823e9ecc9c9aae0399bfbd14d28e444d493f3012
50188 F20110115_AADAZY dunsmore_l_Page_147.pro
48579e9163be8490f0f83986b7de202e
e05441ac0c245fee36d1ed5a6fbeddeea4ea9fb8
82526 F20110115_AADADF dunsmore_l_Page_091.jpg
ad1a9315a78c07011c62943586db6308
aad7962ff19a85a37017ab86e3e64e72e527fdfa
1918 F20110115_AADBGI dunsmore_l_Page_134.txt
2d824487e0388baf89fa81916f8e6763
32f99b1d33f1454192707b47652f30f787b9b5a8
F20110115_AADBFT dunsmore_l_Page_119.txt
3ee980380320e72bd029f3bf6a5bdf23
705565726377ba18fa07c8a672931c9812e696e2
68862 F20110115_AADACR dunsmore_l_Page_077.jpg
cba809fd9a2e6610cbdb8b19b1821d45
f9f41210bc47749729f45b26464df48183680626
52463 F20110115_AADAZZ dunsmore_l_Page_148.pro
61dfab9748e1baae65e57530e66c339d
5b2f6ecec88276dfc6378017816764d72886aaca
83270 F20110115_AADADG dunsmore_l_Page_092.jpg
98edf1f59792baa6719e9134527a5912
b4d5716bbf6ac1fa7b729b835b5644d42b5ef98d
864 F20110115_AADBGJ dunsmore_l_Page_135.txt
baba6d40076fa0d8c68ee43edbfd2a1b
0d2bb0c88217a4e68928b0c4afab6dd442b56483
1862 F20110115_AADBFU dunsmore_l_Page_120.txt
7e9192cbd4686cbf08cd27ea003705c6
4a2e0c0bf984a68bbb52f81a083ab024833462cc
72618 F20110115_AADACS dunsmore_l_Page_078.jpg
eff606534b7c8b768fb4fc8f698dc227
5025ae76541889d68d0321c4fc4046df6ea74be5
93339 F20110115_AADADH dunsmore_l_Page_093.jpg
3ae21edf9d4b46b120041eb6eefb1ab2
d839085c6d35151c73cbd5ecb0124a7fa1a114c7
1264 F20110115_AADBGK dunsmore_l_Page_136.txt
b738b67d0edcb7204ec78fd177bf9acc
7d60a9c30dee5df198b134d816346e6273a463fb
F20110115_AADBFV dunsmore_l_Page_121.txt
ad642f7074f25271f1babfa30abb3eef
29eaeb50a09af769f4386eaa9d287f0fee216602
72790 F20110115_AADACT dunsmore_l_Page_079.jpg
7c68dddfb738e4a90fb59f19095efa0c
cd2456c6f4a99da298652b32bf00f564f79ce9e5
73065 F20110115_AADADI dunsmore_l_Page_094.jpg
7a01e9bef11eb20d121aed675b599b10
cff538d315857e4d352ea29021724542cad45acc
1845 F20110115_AADBGL dunsmore_l_Page_137.txt
44987335f1bd81628e99ab54ccaccbfd
7427692185cc79f75f19e66d56c3d6b5b4a59576
2077 F20110115_AADBFW dunsmore_l_Page_122.txt
0192dc9604ceeb6e274db398ef2eb81b
d77d5ac612665b6d89ce9826caeb73ae7e0508ce
74937 F20110115_AADACU dunsmore_l_Page_080.jpg
627796a181011ba24ae2e8792d4d3874
0f63f0d15afbc88f727e8589a07a756977a7b794
74674 F20110115_AADADJ dunsmore_l_Page_095.jpg
56831bd9247891a11516aae532105d88
c78e2f5d7cfaba1e35246170092dd32cc2bda7b1
2891 F20110115_AADBHA dunsmore_l_Page_152.txt
f3ff3391d850ff68ec8a4050184623ad
1371ff7d3b94d08888e54f4b4c6c552c24a18590
2011 F20110115_AADBGM dunsmore_l_Page_138.txt
58b5cae42dc63e50a3fe52732aa9d3ef
7831f8814223424c97e652e33710f61b65dbb8e9
1251 F20110115_AADBFX dunsmore_l_Page_123.txt
2d7db2f89045b90b5e57dbf10b61e608
e85e585f4f862b20dd799c8806f3cd56a4e4c3f0
73064 F20110115_AADACV dunsmore_l_Page_081.jpg
a7d34ab5575f03a8b67d4fc353655d09
cdeb04d239bd0f20d955ee06535c12230effd609
69239 F20110115_AADADK dunsmore_l_Page_096.jpg
f1327da9e69d73bbbaea63a7355dd7f3
9721a5d9633f48d21e8ab48cbe67ae2a66bfe68f
2992 F20110115_AADBHB dunsmore_l_Page_153.txt
20293cfdb3e58eb44f30bf67bb925dc7
792c758f1080be4b2e086fc3ea7d26840481287a
F20110115_AADBGN dunsmore_l_Page_139.txt
a63c8d6ed0a26cd21ed5a9a50d586d41
e7efb4fa1b3f52be5d6887073cdc5d76e40995be
310 F20110115_AADBFY dunsmore_l_Page_124.txt
4aa02fa3933ed6fc269b8c9d8c689834
0134a1eb91570bb162d44bb6860825f66a29ed85
18368 F20110115_AADACW dunsmore_l_Page_082.jpg
23c03dbd3202f9ec35fdf8ee6d1d805f
0a6c8f35c01c15549bab2e1b79dc6e9ba490df80
69567 F20110115_AADADL dunsmore_l_Page_097.jpg
37243e960baa0a0b5ae0a5c17665f57c
5206927a2f6a859dad4fbab05b8e025c0a4206b9
3065 F20110115_AADBHC dunsmore_l_Page_154.txt
9c63acd42e42dd609faa999e0c4cd056
77f832f200f3d82a90634cb290df147fd99c21e0
1987 F20110115_AADBGO dunsmore_l_Page_140.txt
e2950ce87d4542fceb24fcc862fade35
144325a8dc460b17454460a1e9524cd6211790a6
1256 F20110115_AADBFZ dunsmore_l_Page_125.txt
eaf4c5f1ab0c667755334804e99011a9
907766b983eb4c65c61e43e354ec0360c67713ba
37414 F20110115_AADACX dunsmore_l_Page_083.jpg
bebd2436486ca3853ccf16d44650d635
21f5a07d567bd70e9f8005fd9e066caad10fc640
71746 F20110115_AADAEA dunsmore_l_Page_112.jpg
24d98f737ed2a9eb9aa8f30769f25c94
f925254a2db23c4da5bfa36d50f0fc4e4972cc59
74007 F20110115_AADADM dunsmore_l_Page_098.jpg
a47a2fd5d608fc97e97b677ec2b6f935
bfbad80edb72409b807ebbd79252b721ae428c05
3115 F20110115_AADBHD dunsmore_l_Page_155.txt
6919ed5ea95e88232c37562b6759501b
8e888154ea5bb00b0eb060778d4b6b21d389d420
F20110115_AADBGP dunsmore_l_Page_141.txt
ec5d0ff081ddfd06b2fe4a1a2e5cb43e
3c734161c4ef791a648ce3c81b7d451ac2850224
31535 F20110115_AADACY dunsmore_l_Page_084.jpg
cdcdf98f6c1f1dcfe9fb82bb56f9c0b8
964b0f68a0119db6e9c3d26bcdaa09f3e1640aa4
72378 F20110115_AADAEB dunsmore_l_Page_113.jpg
46eea14a57b0e63036733374c7b7f809
12bfe85e2cb4f951bee2042841c517e7b9913797
76104 F20110115_AADADN dunsmore_l_Page_099.jpg
bdc4017da6061962644b8998fee9ce13
312787ee8e28333353aa802493fddba2df9adce3
F20110115_AADBHE dunsmore_l_Page_156.txt
1766f974be167dbc864570c7acf78f71
3166cf8052c597bbf44e6d4cec29305f685a325d
1988 F20110115_AADBGQ dunsmore_l_Page_142.txt
7ee1525d94f085b7716f5061483416ca
48fc8bc6c34fc56fba1690344bb4ad7edc9f5633
36418 F20110115_AADACZ dunsmore_l_Page_085.jpg
9198e80f64bfef15d5772e8436a43115
773fa02e02da946a790bd4f3f8751bd776a90687
72137 F20110115_AADAEC dunsmore_l_Page_114.jpg
35c91ff88953c02a5ad055cd3024fa8a
5413d83227db76f01e6d5d9a6d53a4a4f67eb838
71805 F20110115_AADADO dunsmore_l_Page_100.jpg
61d3abf25e2e6b374ba8da98575fe897
55dae2988d1a64e71c43b65bcb09031ecc09286b
2608 F20110115_AADBHF dunsmore_l_Page_157.txt
e682529415942ea444ef66082fd1d6a2
0988295372eec67104c58a359bf5eeb347d76277
2085 F20110115_AADBGR dunsmore_l_Page_143.txt
f668bbc02e142a8712a085d34784c1e0
7f7f14b19d06d50f8af131eb24056a919ad3f919
74037 F20110115_AADAED dunsmore_l_Page_116.jpg
fe6504c9e8af4fb857254763a697eb45
04136fb6c08e306dcfe9c5503f08712d02255851
71695 F20110115_AADADP dunsmore_l_Page_101.jpg
423356ceb6e810e96b9dcfef206df3f1
8bff7ed94aa53c4106031f63da146688ce83a28d
2730 F20110115_AADBHG dunsmore_l_Page_158.txt
453182b9a1133bf56b1e02f8e5202e18
0212a1a37e4cb7e3a76c0d41f9dd844441f327ad
74403 F20110115_AADAEE dunsmore_l_Page_117.jpg
5c10afc8986fac3ac8d3996b5f37795b
2450efeecf08204923ee38340598a6016afa889a
2688 F20110115_AADBHH dunsmore_l_Page_159.txt
540179ea936b36f706d6a8cb49dced48
946dd37590bb84e9abffb2fe140072e3057fb4ca
1982 F20110115_AADBGS dunsmore_l_Page_144.txt
f8ba1f54eb54cf44fa86f1a413f14514
73882b16df5be619e5200e672097b9d3de8b24e5
70991 F20110115_AADAEF dunsmore_l_Page_118.jpg
2a013dd0e3cf97e01877a42cad773de7
37ae73e3429a9aac83339f871488da22db7a1d92
69156 F20110115_AADADQ dunsmore_l_Page_102.jpg
51742fd5069b944292ad9a8ee0d43520
5d461548b640e0fba37bc80490e6cd529584006a
2223 F20110115_AADBHI dunsmore_l_Page_160.txt
85f58c79ca4f2405414e1029cfe099e0
ac91f8645247bc1e85209779178a6ec10a0c5d3c
2089 F20110115_AADBGT dunsmore_l_Page_145.txt
60289f7ecae88e8c670dbc7c24c91893
dc0d8bc4b324ea6396567c26ccca959df0ba31b6
72644 F20110115_AADAEG dunsmore_l_Page_119.jpg
db49ebd97e779a9396c623e02b09b5fc
6ae3823f5944e4872183a338f05513c19800eb4e
74517 F20110115_AADADR dunsmore_l_Page_103.jpg
81042b5174f8b3732b637cad06aac695
40eaf9644a86c5bef250004a795a1e8dfcc004a7
2548 F20110115_AADBHJ dunsmore_l_Page_161.txt
b7ca68c830b7e59abf56d7efc0e4dace
b86e4881d10d21550218640bfca3578bd34ba82d
2078 F20110115_AADBGU dunsmore_l_Page_146.txt
29029365b6e10a8b645482ae61044405
e597505815d9eaabd9744083bc99d5981ff1526e
67793 F20110115_AADAEH dunsmore_l_Page_120.jpg
d82c74ee68d8375281611d458197a8c0
7e333e2853dcef5df10567ccf097b44c6fa0362e
73811 F20110115_AADADS dunsmore_l_Page_104.jpg
2c8979676ae4daeb5cdfe7912a441f59
9a10adec40bce2de15786e8bc3a8676c5ce5b768
2045 F20110115_AADBHK dunsmore_l_Page_162.txt
c409af4706743ed8495e42731f506f87
33d4e8008abf540d7691e18c90d192205d65af9c
1980 F20110115_AADBGV dunsmore_l_Page_147.txt
b0e69b7d2c7fb96fb628e429a27f2fdf
34167fd47529dc5b46e2dcc77197fcee532c0908
72340 F20110115_AADAEI dunsmore_l_Page_121.jpg
6c3d4c52035b8a66d64f325b76b12ec7
ebd80b3bdf4d6e952bb4e8df63a10898eedc7574
75192 F20110115_AADADT dunsmore_l_Page_105.jpg
ffc8adf5b596316d997e6f6d72c7609d
08a5648cac080421d1904af57007b74cf7c7e7f3
2682 F20110115_AADBHL dunsmore_l_Page_163.txt
23d6855a362a2bb8a3776c5a9b4998db
4eb2e55c39bf53ecc4128549750e65cfe9e7928e
2057 F20110115_AADBGW dunsmore_l_Page_148.txt
a153f0d161db108bc0e2152e53d41787
8a288d0febe88b1de412dea3efbeb57340cf7174
74219 F20110115_AADAEJ dunsmore_l_Page_122.jpg
7376426d3316894577c09d307c170b1c
5317fda27fce2607a7b378a009615ad4e4cd3202
73578 F20110115_AADADU dunsmore_l_Page_106.jpg
2df195e0589d9c4b752cc5d3f8f49154
5e831d4493ec5e58dd6a443aae5c078d2e2f4c3f
2423 F20110115_AADBIA dunsmore_l_Page_178.txt
7db68bcd4350d1d7dc5207908b747e4b
0f1408944d4a681ad92031251c832d878d032b6d
2697 F20110115_AADBHM dunsmore_l_Page_164.txt
131d8f6dd7ba8a562e591d0235454240
78d8e1d491dfbb16e84b0b4b3689cfef4956b869
2071 F20110115_AADBGX dunsmore_l_Page_149.txt
631160fda3c06fea55a39173278a0d7a
658bff0a44d8728377926231b3dcfeed7479f718
48160 F20110115_AADAEK dunsmore_l_Page_123.jpg
4c7d422063661ca8ee2891e32531d30e
ad888c552d2c8dbc33eb205eb0b784916b4c9d82
74109 F20110115_AADADV dunsmore_l_Page_107.jpg
08a70758833da6357037cf97162a733b
2ff05ba4e1acd1cdd1ffb812e6c00e40b385b92d
957 F20110115_AADBIB dunsmore_l_Page_179.txt
29ebc364096d8f5e39802f2bc848e921
477ebdc761295d5adc3209ed180d3308129fc28d
2513 F20110115_AADBHN dunsmore_l_Page_165.txt
eef3d428beaff3e2aa53765f3c4922fe
ebd2c02302cdd572e1ff4af42f2f6db920dcc79c
1478 F20110115_AADBGY dunsmore_l_Page_150.txt
d9e18cd7c408b588159889c1c949c52f
fc3cccf3ae14925b62b9dbf5b9585108f58368dc
39836 F20110115_AADAEL dunsmore_l_Page_124.jpg
42c955eb8c3c314c1b1a4095304d2498
69951f84c99c646ceaffa4f19dfa4d949e1926d6
75392 F20110115_AADADW dunsmore_l_Page_108.jpg
c4ec0a99a7d287df0e6d6ab6f1af39d1
4269d250d41e8db89cd2be29214612400ef89af7
1670 F20110115_AADBIC dunsmore_l_Page_180.txt
3df93521c952f2cff190f98a114ff520
8a6da7bf25c43dec88b34c235b5970da46ace9e6
2369 F20110115_AADBHO dunsmore_l_Page_166.txt
3f89e8ed6b8fc77e61cffa16df753be1
59d48c0d9e638b212e6fa63f19119a741732e9cc
2779 F20110115_AADBGZ dunsmore_l_Page_151.txt
cba170fe012eb0b842901c1d1aedc7fa
ef8a22cb55b235123bdcf30a0bf484d78c7ace96
72528 F20110115_AADAFA dunsmore_l_Page_139.jpg
c46c5ca72ef4a84d5aa7dcb15b672151
30ea57bf16e3ec1d6f1a9ede98b87f585d71aebd
46106 F20110115_AADAEM dunsmore_l_Page_125.jpg
a083cb6d0477674980730b512874ff28
68f8a98bad1fb327f528634ef61516dd56a22a9e
72157 F20110115_AADADX dunsmore_l_Page_109.jpg
93e0a6b1081703708a2c10edb9c6ae30
896daad1af7e570e912f7338754a481a5efcec3e
689 F20110115_AADBID dunsmore_l_Page_181.txt
203920223668a987f0216c2b41b5e5fc
d995bf52af29519043f5661cec53da1f751aca67
2588 F20110115_AADBHP dunsmore_l_Page_167.txt
007d440e09581126b0b2e28c3dfb5c66
7aaf97a59e8e2c02d52af1c0d78594bdf65efebe
71664 F20110115_AADAFB dunsmore_l_Page_140.jpg
b632e86dde5bc3cf09e50e3785ec556f
73357eb233e10bd911d1f18246b8c9708b668256
28172 F20110115_AADAEN dunsmore_l_Page_126.jpg
3efb6603b5d04118a1fcbe45fb93ba29
a152250c1985015a1238d8c22a2fd2382e6e01fa
67583 F20110115_AADADY dunsmore_l_Page_110.jpg
e2e9c5c074598a9479b6040708cfc979
f10dcd4b832cbd621981e40d271b6e5f88e298b6
1719612 F20110115_AADBIE dunsmore_l.pdf
4d633c8500c1af8dead35794eaeb6cf0
cb29227563a8201506cfab69e77122f4c8702ce9
2497 F20110115_AADBHQ dunsmore_l_Page_168.txt
2917842947adf0b2c2c57af6a0d09d8e
4c893efe7ffd5ddbda900ea83dbb820800ae23cb
73416 F20110115_AADAFC dunsmore_l_Page_141.jpg
59d819ef8dd6efb825278027004ee677
8abcc3cc0b9fb61becc94e2224301ff11e230b3b
36293 F20110115_AADAEO dunsmore_l_Page_127.jpg
eb5b8adbca7442d2be7ee5d5387aa38a
54cec63ece030c85c25392a080a237ab6773b418
75007 F20110115_AADADZ dunsmore_l_Page_111.jpg
7639e2536e969e953f6bfddb1e1c61b4
080a2fb5c58806dc74d746c273491d4495a95846
6578 F20110115_AADBIF dunsmore_l_Page_173thm.jpg
e578f25382f6f4419a631ae3888ddee9
59b2454472979440da0103e2f39f2a81b8bad8b8
2643 F20110115_AADBHR dunsmore_l_Page_169.txt
2594443e07e7a126848994ceed5513d2
2c2388894c745e5a54be42c4e3b07633710a7c52
72253 F20110115_AADAFD dunsmore_l_Page_142.jpg
0bd68fa70faff1504fa5f50ea24ef6c6
66935a981fe1debfda7ff3896744f6b03db2e9ed
44695 F20110115_AADAEP dunsmore_l_Page_128.jpg
f8d7fe376eb9f075e2a63ec354d20142
94a566e996a784163c332ee722b859f2ac852de4
23790 F20110115_AADBIG dunsmore_l_Page_075.QC.jpg
6e0ac87cfe279bcdf42c2899c1936ec1
eac64b65d731de471a48b22747f7acc37395fc7b
2456 F20110115_AADBHS dunsmore_l_Page_170.txt
55f4d4f75960c45759755c7e81c693a3
28584f9f0c6ce06bd5680b5ecc24cbd950f37802
75413 F20110115_AADAFE dunsmore_l_Page_143.jpg
23f41506506bebf9070440db0b19b99b
c9843b0cd5c90f9f7c4ede764df4f247ca20c472
30741 F20110115_AADAEQ dunsmore_l_Page_129.jpg
453ab1e78394473178350fce6c35467c
6cec2d342ed9feff5e5e4bf712bd57c2f365f0a3
23146 F20110115_AADBIH dunsmore_l_Page_045.QC.jpg
dba9a8082c2a37228e47a38942878821
97bbd7566f0d0178f4f14e7329b24ef0d6c26bae
73070 F20110115_AADAFF dunsmore_l_Page_144.jpg
9e1ae727a0c167bfcbdeb71649ad2707
950c86378b03288e0e2c82b3e55e45226b601c27
25293 F20110115_AADBII dunsmore_l_Page_167.QC.jpg
f175cea7e9e5ac8a5ccdac3402ea8214
2667759902183a8fdaf35edbc12d971f6ca5d2c2
2595 F20110115_AADBHT dunsmore_l_Page_171.txt
20cc2101d136926810022d6abdaa90d0
8ab44e1426527c5ddebf4d90adc88b878cb871ed
74325 F20110115_AADAFG dunsmore_l_Page_145.jpg
73f55ef1969eacb9bdb0129b603dbc72
23a5894d0803309f835de9dd026124b941aa25f2
29611 F20110115_AADAER dunsmore_l_Page_130.jpg
9af221214d75469ddf5323f985692834
47521808526698d34a25adb53f25fabc83c11e4a
6539 F20110115_AADBIJ dunsmore_l_Page_075thm.jpg
47fe906a5dc953892a1fe68d9dcf68ed
13f3c47ff21ce32ad8ca07678145bd2f4edfc425
2852 F20110115_AADBHU dunsmore_l_Page_172.txt
10aef15703c570bfee3f09281506345b
b6358196cdb816295dca406fd74843ebbc4e8626
74298 F20110115_AADAFH dunsmore_l_Page_146.jpg
b7516cadcf39eefc40bffe80bd0824f6
2ab94ec033e7098c91a0ccdb26021afebb4c9700
37053 F20110115_AADAES dunsmore_l_Page_131.jpg
416ddfe9954139a17c8f51980a00db81
cfa4c4e6ccf19e5b37ed325bd33f61b26d2a7f2b
3511 F20110115_AADBIK dunsmore_l_Page_131thm.jpg
ba8d5ab30a46fd26818fd754404a82c8
749de73662645845c1fe1f0d0de49053adca2531
2413 F20110115_AADBHV dunsmore_l_Page_173.txt
b3b30df018c0815540ceb9ac8b45f9d1
4b783af9df5f346890340566eda543fb1956ea72
59467 F20110115_AADAET dunsmore_l_Page_132.jpg
afc5e1c6f6a61019dc6d5b954793ce89
4c20520fba95ca63385e3f90d693cbd697dffd06
71363 F20110115_AADAFI dunsmore_l_Page_147.jpg
015a736550358ae345568f14b77e5639
0b19f998afa98c945600bf855a8db2bd2774e180
19812 F20110115_AADBIL dunsmore_l_Page_027.QC.jpg
76ff1657396c8b5469b8addabb96dc21
c42dba3775956284e2e97114a178ef1ef26374b9
2263 F20110115_AADBHW dunsmore_l_Page_174.txt
ff68f05e99f5b09fe96d71ab45f8e3c7
ce774726c9b5985ca8e8d9eb494ec5f72c30fd12
40416 F20110115_AADAEU dunsmore_l_Page_133.jpg
c463bf63ff736af9231ec7ef4ff35d55
0368f96099f02f0ab6322bedcf2886519f6db001
73947 F20110115_AADAFJ dunsmore_l_Page_148.jpg
289785548bd1701cfe6b71c3fcc73f4d
f10342ce124d1b45356ef47fec9de33418ee1f52
24572 F20110115_AADBJA dunsmore_l_Page_105.QC.jpg
d19d4ab4d8f07a877e8453be9fbbe356
43011da49bb8cfd17d12df0fd9179f3dfc4f302b
6566 F20110115_AADBIM dunsmore_l_Page_039thm.jpg
0a0e310a7d4348f4d5b582ae60bc1dba
b2329f7a801d2b9089a61caf4799553afbf47263
2381 F20110115_AADBHX dunsmore_l_Page_175.txt
275ffc4a84f84c815e212c534f386a58
fb94195dfd85753b447a2d49c57d0be8cf668c6b
65399 F20110115_AADAEV dunsmore_l_Page_134.jpg
c29c910aa15b093b150414309568c0ed
bdf7725ce7089cb05b4ff71a1104c17cc4bb59cb
74134 F20110115_AADAFK dunsmore_l_Page_149.jpg
fb412d782d0fec82650bfd2d8a543720
838cd00d42bff90d95fac982e749df699a4c2179
11790 F20110115_AADBJB dunsmore_l_Page_053.QC.jpg
8b2e9148350a5e22e6d0e9b2c182bb42
25f4805d732f21c04066ece3b00d7628e195f807
25834 F20110115_AADBIN dunsmore_l_Page_093.QC.jpg
a06c8962aea018aa50ac76e4237fd497
36c3bec2ff3a904d21f993857df65dd75a2225a7
2258 F20110115_AADBHY dunsmore_l_Page_176.txt
4ed5fb11801d523178f27b752a0ed673
79c56eacdffef5cd356c2a18c10867004d447a20
33903 F20110115_AADAEW dunsmore_l_Page_135.jpg
f0dff84a07b07d4d2a3e2cd60c8107a8
8299ecc5adf86ebc9359e2018707bbd2a3aa5b03
54855 F20110115_AADAFL dunsmore_l_Page_150.jpg
26014a05e48b29b6303e78c59a61ff7c
33f2569aeff421222a4a90e9974815fc4cbe848f
6028 F20110115_AADBJC dunsmore_l_Page_153thm.jpg
4ba072a0e7434aa349bfc59e21dc5fd6
f0766c1e93b5af926ad00f5633a6b8d0cb19830e
23516 F20110115_AADBIO dunsmore_l_Page_140.QC.jpg
43f662ee0b02e1fc30939216bfb354d5
49c4e09483937776e55b605bd2706853a3e5275b
2501 F20110115_AADBHZ dunsmore_l_Page_177.txt
dad2b5a76d47597c0985863bf80042f5
8192e7ae0a57f58a1608f029e86826aaccd3357a
47983 F20110115_AADAEX dunsmore_l_Page_136.jpg
9b93fed0cf62519b10de051f9a6731c0
382e62759e2cc14aa0f267e8bc67f0cb3cdd9e9a
90659 F20110115_AADAGA dunsmore_l_Page_165.jpg
8cb65f3137274b77f87272b50c529215
81e7e6db4c0e18ccf50f17cf26d506c0645e0b5f
78456 F20110115_AADAFM dunsmore_l_Page_151.jpg
f07ca409a37132117e974c07f4f94287
4d8558cd3284b08d40cd4772548e70cc45d708ba
4500 F20110115_AADBJD dunsmore_l_Page_055thm.jpg
2ffae6db0a9d325d25c7547ec46f07cf
027bb5471cf168e0af5d89832f8ba80cba1fc9cf
22160 F20110115_AADBIP dunsmore_l_Page_151.QC.jpg
69b13926beea9b858502b4da411758e3
98f012a7c2ddbeb8f2a66d3afd2afb80283e35cb
65240 F20110115_AADAEY dunsmore_l_Page_137.jpg
db3fb15c67602c6b02d2b45a05967b4d
a7aee8216cafc27af496e726b92a35f2175ae4f7
88073 F20110115_AADAGB dunsmore_l_Page_167.jpg
3e79a367cd37cb95ea5f74cee2fdbc8b
0296b3d7564eb608e0f3c7c66e2074fc52567dd8
76208 F20110115_AADAFN dunsmore_l_Page_152.jpg
0eeee7d9bfeccc754ece5dc03eee1c76
fd7651939ad65a8780f97e22147cf0091809cbb9
6595 F20110115_AADBJE dunsmore_l_Page_107thm.jpg
d1588ded6b84609ef7c184d0999891ca
3e865b65f4b35b75dc8e42acdc746b6e81c81931
6889 F20110115_AADBIQ dunsmore_l_Page_177thm.jpg
78ba997fb81f1e3e997158000922fdf0
f366587e45cb978b5cecd866b3651f30d6e8fa9c
72814 F20110115_AADAEZ dunsmore_l_Page_138.jpg
495f889be3f44209fea8a8af722ebc25
d9c42b6ba755ab9eeab3a9293edc581e6b357f04
93245 F20110115_AADAGC dunsmore_l_Page_168.jpg
dd65327f485134400faba3b15d32093f
319c7d4f36a8c38a9d75681f8dd6949e04df3278
74640 F20110115_AADAFO dunsmore_l_Page_153.jpg
0befe1ba4872e3a805dc4c7a8f386513
3dcd55e29e0090c21276c6a0d3e3235883adf57a
19958 F20110115_AADBJF dunsmore_l_Page_066.QC.jpg
3e0da07bd4f0a5c434d2005f5c43d3ad
f15d4dfc489c323165e1d28b5dbe4c91c9e9fab0
15364 F20110115_AADBIR dunsmore_l_Page_125.QC.jpg
ca11d6beb5ae2d0da73fe95d0b6f349c
8c9b8cf2daf5175ac09c60105e537d5550f2ddbb
94691 F20110115_AADAGD dunsmore_l_Page_169.jpg
54850e3f4cd3d96ebe623fd070f0a25e
91a0a81391991d1855d61fda14c6228103bd091f
75546 F20110115_AADAFP dunsmore_l_Page_154.jpg
29fbabb8cb70aa2077ea89b3ad579c48
caef74c424e6439ca2a9652fe2ae9ddd52256d4f
5802 F20110115_AADBJG dunsmore_l_Page_134thm.jpg
6fc2483cb46b869ec849d1cbe71cdec9
7baec87c22499cb7dc802471a842368480cc5851
22735 F20110115_AADBIS dunsmore_l_Page_154.QC.jpg
fd3535485c130e3c2759cf750b24256e
5a226204cee60e806c03a15da99ce8b59af91d8d
83603 F20110115_AADAGE dunsmore_l_Page_170.jpg
f6d628775ab979a8a626733787146061
628862a98a2e3423634174ed068cb3f6698da5b1
81424 F20110115_AADAFQ dunsmore_l_Page_155.jpg
14b9dc5488dedbdf8d7180e62a8bf5f4
61cab6b490bc12d5ba4ecb82acf88aed4ac48384
4721 F20110115_AADBJH dunsmore_l_Page_087thm.jpg
e1ac25be2265aeea3a151da27be34a15
6c05d181350e83cbc76db0bd2d855ed2da6c91fa
23637 F20110115_AADBIT dunsmore_l_Page_042.QC.jpg
f1ff8933e4a72ee51fce6577ea23c4fb
ec9c68bd6c7642fe1cfa4c588d00070d98962725
100972 F20110115_AADAGF dunsmore_l_Page_171.jpg
0f6ee9620e0bf4b3d57d35f3e178d9fd
948cefb028d9b4fbf5533363c9f05b219b6873d3
82207 F20110115_AADAFR dunsmore_l_Page_156.jpg
b050a548780fc74ec4491a0cfb832746
5ed500849a446d8b4475023cb2513bc5ced2f951
6355 F20110115_AADBJI dunsmore_l_Page_109thm.jpg
4813fcd643806074ae6d2c3ceeb12895
3e6152dfa40da9991651d08a46a0d371b76708c9
97118 F20110115_AADAGG dunsmore_l_Page_172.jpg
ebc644d468e1cd8f27bcc6dcc2c35040
9a582448f105debef86d1d137230e27d294e7f9b
23389 F20110115_AADBJJ dunsmore_l_Page_100.QC.jpg
b38b7767125a6c2deba97d42766ef56f
b3e14bf1ce8ef173ed67b0d667600f131b230526
20986 F20110115_AADBIU dunsmore_l_Page_137.QC.jpg
8638de1fa46c2fc4371c8e393a790143
f18ab05d663da91ccbdd20612d35ff2120806ff9
87605 F20110115_AADAGH dunsmore_l_Page_173.jpg
e715f2e95c3f23ccc220917cf7240771
061278abc42193740d5608b9428d35f4cfa4c2f0
67692 F20110115_AADAFS dunsmore_l_Page_157.jpg
87b9ea3d4448e8862ab2a8f4b9ad0597
12f956fd6c6c5886beab12501f26c9098b783fc9
3662 F20110115_AADBJK dunsmore_l_Page_127thm.jpg
c1a57ca0ace19f3056bec39e34f668eb
ac218285ff46d28132b20e8602ffbeab14b8f3af
23110 F20110115_AADBIV dunsmore_l_Page_152.QC.jpg
8d027449358ef6555c084c69e182f9b6
73bd4fa1425d394f3da118316641719af4b667cf
79644 F20110115_AADAGI dunsmore_l_Page_174.jpg
921eadd4768b298c7144206a0fc241ae
55c84be6c2a459d08c69c704538fd3b56f7d6c96
72875 F20110115_AADAFT dunsmore_l_Page_158.jpg
1dd76cf63acc87b3c1e76667e79fa35e
81d0e2798d8b6e160bb628b05351bf82a899f919
14126 F20110115_AADBJL dunsmore_l_Page_128.QC.jpg
01d7bb317033a759f9b130cebc206111
322806658feeb270b6a078aab0e6728b67b360ce
23766 F20110115_AADBIW dunsmore_l_Page_069.QC.jpg
a186e3f54d0b09058712b4334addd49a
c51f9cf4c46f4b2e774b7c03aa5ab2aa9bbdd15f
97200 F20110115_AADAGJ dunsmore_l_Page_175.jpg
d02f65d747549953d33d11913b750c5c
b9b069eea787f4082d01e364b6eae855585ca021
74783 F20110115_AADAFU dunsmore_l_Page_159.jpg
871bf22b99b15e87f8a78d61d6ec9429
34eeb42ee97d0e08e7f345b2cfd7c44e5e24510b
23906 F20110115_AADBKA dunsmore_l_Page_029.QC.jpg
2ed6d1e2b5ac59f77f060c7fc90571f1
cb77702605c126788d9cec43760553f6d2ee7215
22886 F20110115_AADBJM dunsmore_l_Page_118.QC.jpg
7bcc04f0c1ebec08fb2e46c41fa35bf0
42f6b105145760dc32f7f9993e9de91973476bd5
16375 F20110115_AADBIX dunsmore_l_Page_009.QC.jpg
4019741acf78f9d4be1ec17fca900d3b
d7737d405918b2c2780e026d15a9add48f7441a7
77363 F20110115_AADAGK dunsmore_l_Page_176.jpg
20736bb05ed67857e26076271956182f
971a1c788d41611e7981c2b94a041289efe6dbf7
65806 F20110115_AADAFV dunsmore_l_Page_160.jpg
e6c56c667fd045d458315e0c6719d0c6
bfd812cd5e5dc0791c2f09dc5febae2333ae53fe
23622 F20110115_AADBKB dunsmore_l_Page_038.QC.jpg
31492feb6a9fa7dad25e11341d39b7be
7311dfebd7278b19c7a0ac32a612a5b54d93303d
6455 F20110115_AADBJN dunsmore_l_Page_113thm.jpg
85f3dd80c3ca716f294ec9d37a6d0ad7
5d0445a16ebb55ac1381118192858cdb88460d88
24513 F20110115_AADBIY dunsmore_l_Page_111.QC.jpg
8826e80846847f54c7e72c69e56e1a5b
614edfe8145ecd1e69c816a574eb09aecb74c238
87526 F20110115_AADAGL dunsmore_l_Page_177.jpg
19bf9b3ebfad5bcbdf0d29ffec81cf78
23b57cff0559ac0510bb8b2fac7bbec467c8a8ad
64990 F20110115_AADAFW dunsmore_l_Page_161.jpg
b73a84ea09276bc39bee531130347cf4
3ba19aea1bb3d062632decc41461d81286f1077d
22067 F20110115_AADBKC dunsmore_l_Page_091.QC.jpg
b9282e7eef52f7178894e396aab08b3b
01010b26ce16cf87a5c43b4a333b6713d7013aec
21329 F20110115_AADBJO dunsmore_l_Page_162.QC.jpg
bd926e2673e0e75dc714f46169514752
1887a2b323e6f3611988ab69687c4c3b48b541da
23972 F20110115_AADBIZ dunsmore_l_Page_098.QC.jpg
c267bcda1858c37a29c4119626eb7a51
b452147c99c96573f70972c947c37ee5211f1b15
1051984 F20110115_AADAHA dunsmore_l_Page_011.jp2
5b1f9aad1a550d642a1621788574cda0
abfd66160f2cbf4968def3bfe811c30758c676c8
82750 F20110115_AADAGM dunsmore_l_Page_178.jpg
cfa62b4b768da6a757e36a9309003277
bed79e6178e4a1fba61345c23d299e0cd35a4270
74224 F20110115_AADAFX dunsmore_l_Page_162.jpg
16ccd08b59d028f2f89784886196939f
47b2590a710d2cfa40d34d66edeb5a406c67b31c
23520 F20110115_AADBKD dunsmore_l_Page_040.QC.jpg
ab7d150444b0fe7adbf7b563b706df9c
1ddb0829cc5864a285ddd48e384b485ed656a824
9956 F20110115_AADBJP dunsmore_l_Page_130.QC.jpg
7f44281a679a31da689b5b58e4893651
61563818b2c411d0415eaa4163df28210b12537e
1051968 F20110115_AADAHB dunsmore_l_Page_012.jp2
cbeffe678473710bd071a1811712484d
610f5fa81429982f65eaf0898017fa0fe80d0935
39929 F20110115_AADAGN dunsmore_l_Page_179.jpg
a141865739b4a2e0b44af8cbf4010e75
a129707fe0a1d1eeb79fbb5cd1ba1a645e5e90ec
92320 F20110115_AADAFY dunsmore_l_Page_163.jpg
2d35e7a6d4e58554f741cdbdfed18879
bab9ca76f791f661a2fb52ac2196136978603d95
23763 F20110115_AADBKE dunsmore_l_Page_138.QC.jpg
aa108580661c4d5123fe864e24fa6eb0
08c697938632158e650d4ab3e2499bcd03602efb
21990 F20110115_AADBJQ dunsmore_l_Page_028.QC.jpg
9971bbd77a34f81defcf9cc427d08f24
d6072fc0b880059765c315c1cd34a8ce24a9f559
700967 F20110115_AADAHC dunsmore_l_Page_013.jp2
070e5145990bca80f645c7f7269a1e59
a2d91ee351dd14b6ae31e54bcf91b9aac916c9cd
62434 F20110115_AADAGO dunsmore_l_Page_180.jpg
727b714126bfd15a1e8775dc827478bb
5f821512a3e5206e3bf6a55703d7421c9feed718
94252 F20110115_AADAFZ dunsmore_l_Page_164.jpg
f5ccc7605f27c0346670e0a7bb4ff0f7
e5baa34f2dd317a7ae9d1fb166df35721faab2f5
23532 F20110115_AADBKF dunsmore_l_Page_119.QC.jpg
c7cbf1a2bf297c1cf5558a127881f6f5
af39cac07db73d54a140f93b7242df32d66aa706
20637 F20110115_AADBJR dunsmore_l_Page_180.QC.jpg
d393d923b72fa6c38724236e4083d780
c023045be13887eb92d1bb5c235dd36c9c6fa908
1051970 F20110115_AADAHD dunsmore_l_Page_014.jp2
e7024b5067d936589f7fd9802d9b4ac7
a36d2fda076eb87176513f02963f0d9c73b28e97
31041 F20110115_AADAGP dunsmore_l_Page_181.jpg
cb632d76cc58cdcfbd0b1709f1146c14
9a5a830f10ae0c742218cc4c20990d57e24b9793
6529 F20110115_AADBKG dunsmore_l_Page_007thm.jpg
b9b701c4d1c7e8dfa080f6a75d96c6d7
51956c10bb4ee3f1585292f1fba605afeaa907de
24985 F20110115_AADBJS dunsmore_l_Page_155.QC.jpg
7d23e1852fc83a0fd3b424c57ead7eae
9242daf1ef91b12eb9f66ba3abd157b722c00c24
1051962 F20110115_AADAHE dunsmore_l_Page_015.jp2
44d799fc8fbbcb87f7ee723690044653
d8b2fe3febc58fa154b42701e067ea3ac6120cfb
29772 F20110115_AADAGQ dunsmore_l_Page_001.jp2
08c904c757dac45f755803d3e259371d
176bf49bea59f7c72818e78c57f594761d93f74d
22178 F20110115_AADBKH dunsmore_l_Page_071.QC.jpg
34b8cd6ecf07e117a58f9ff9027781d7
24b1b3b99a561ffade43086472420a03be20650f
6590 F20110115_AADBJT dunsmore_l_Page_033thm.jpg
d3972789fe35570d7bd6e3b6aed188f1
fea0b502b5cd7f6e46dc89642cda05251381ff47
91498 F20110115_AADAHF dunsmore_l_Page_016.jp2
ccc3d608c0515304a2a36725a20012b7
7cf83494d81e6387265a6fe2b16177042e65ff9a
6987 F20110115_AADAGR dunsmore_l_Page_002.jp2
902b463573e25a420c00439cfaeca04a
7cb4214a1964e715026bb7d63771adc648ea8270
6508 F20110115_AADBKI dunsmore_l_Page_079thm.jpg
b282a08019d607deaa1d0cb39979b39e
6de9bb7fcb44fe075f51fb684bb94bfbc8169eeb
6766 F20110115_AADBJU dunsmore_l_Page_167thm.jpg
e9492b0e0f3af03d6af6e9635fcda6fa
6f9c0c8974f68fcbf1603547365eea2f009082a0
65934 F20110115_AADAHG dunsmore_l_Page_017.jp2
11b55cb5fc49c34c9bfa2f8505703e87
4fe2d8e17e4e4da71672155ff2df4a7fdf2e7989
15558 F20110115_AADAGS dunsmore_l_Page_003.jp2
6a1fe4be883f66a4343db0e1798b919d
e4bc07a44c5a65a4e661f3a2c9137e682381d128
6822 F20110115_AADBKJ dunsmore_l_Page_163thm.jpg
2c153cab0e50adcbaa62cfc4a4dae44f
78a52bb21aebf6e6edc3aea7754085d9e1566e16
101231 F20110115_AADAHH dunsmore_l_Page_018.jp2
e58c72fd25399ab29e3eedc9055a1570
4b4a084a56ded7c4384525120bfeab2fd89e9de5
23130 F20110115_AADBKK dunsmore_l_Page_097.QC.jpg
d2100248a2081dec420170c6f9fbccbb
72175215b036e4374005636db20f5cb36a8dc7fd
6482 F20110115_AADBJV dunsmore_l_Page_147thm.jpg
1ce05d69dc3a1fed2bfe17da672d307c
40c8752ccfe7ba85665eb4bd14e613f3b1579037
113495 F20110115_AADAHI dunsmore_l_Page_019.jp2
a2480acbc332321b150b82e64858479f
06e057ae863c16de2450f33d690e96c275fe0c53
93577 F20110115_AADAGT dunsmore_l_Page_004.jp2
7b35581604829e779a08bbd61d6f4218
529d04543863fff5df8c86b872f646c62ac9582d
2805 F20110115_AADBKL dunsmore_l_Page_047thm.jpg
32d9d0542f9187088f9d745e31c48188
c214191912c00006fd391376af85c22ffe41b6d4
19693 F20110115_AADBJW dunsmore_l_Page_016.QC.jpg
102cd96c3abb7cab0146f2b9455c9dd4
7c3df1f8d414e82418abaf7962b0731db19f1564
112998 F20110115_AADAHJ dunsmore_l_Page_020.jp2
2c5d49c01211c907b81d8286a79146d5
d45f4b4a50aeb2f2c50d3b6e987629fa72830ed8
100322 F20110115_AADAGU dunsmore_l_Page_005.jp2
a28819d544f55f25d4990bd38642205b
9dcd8f6610c776d0823a96a95b61894874752b75
6844 F20110115_AADBLA dunsmore_l_Page_050thm.jpg
52c98be950418847db69718caf8398a6
fc929c55acf78300f715629830a91dd2d064edfc
22525 F20110115_AADBKM dunsmore_l_Page_035.QC.jpg
c97087470e8e52a8e2d208cb6c1d4e34
c98c094e2c7c6e514dc9ec46825fd1dbe9f98bf1
25024 F20110115_AADBJX dunsmore_l_Page_173.QC.jpg
16233901c8b5a6445a33c22f30cfc210
7eb1436e7b286b5541dd9c7ab3484c6fc421428d
115454 F20110115_AADAHK dunsmore_l_Page_021.jp2
4da07714e0dcab531bea27c8442e6a6c
997ddac291eb1d01918eeb751501bdae55c0f2d9
107082 F20110115_AADAGV dunsmore_l_Page_006.jp2
a9c7bfb3253157c774a423b3fadb092e
562f17d2d5e7daa4dec00c5ff59ddb22fdb84a2c
21815 F20110115_AADBLB dunsmore_l_Page_095.QC.jpg
33714e56191470fb5db029caa75b9325
83f4fe9d0c155d0678d6295f5e4c08d9fd71a5d6
6409 F20110115_AADBKN dunsmore_l_Page_118thm.jpg
bd95fe2a8f808a3457ca5ef8645b4a23
861b5e62b98dc3bb24b84739e7769495c2fbcb0f
22095 F20110115_AADBJY dunsmore_l_Page_077.QC.jpg
cf0d471fa6ea32fc533413d27efb0c22
772398bf4371c7f9f18c995dc98ad8ad43937595
111536 F20110115_AADAHL dunsmore_l_Page_022.jp2
e5b018d2ac5e64f845cd26c5bcdf9668
01e8224c619e8087b46e69e6774a133d1dfdf552
106941 F20110115_AADAGW dunsmore_l_Page_007.jp2
90e997be1d6214b59c594bf6a18a1658
dce65d0ececfd7eaf12b1f724593472dba945209
6543 F20110115_AADBLC dunsmore_l_Page_145thm.jpg
e291d48e1167cb3c64be9bc33d990925
fc56df751cd60f523cb63e81ad7e903622ed539e
2515 F20110115_AADBKO dunsmore_l_Page_013thm.jpg
6cb7dcd467c4a70d42f517ab23e6c836
7bf98de6eb18db538fda460c257fe2db5bdb8ec2
6644 F20110115_AADBJZ dunsmore_l_Page_042thm.jpg
7e24d03983dd85d916d0670a371cf70e
fefecd0bc1190599cf71cc51077aa2b362c80957
111910 F20110115_AADAHM dunsmore_l_Page_023.jp2
117b37a999d03efcb3fae9a6fb64c461
3fbb20f21e145a57b2784e4075a28d44b51047e8
31368 F20110115_AADAGX dunsmore_l_Page_008.jp2
6d5ff259f8f34ff9e4b7ed0f0e27ef17
76df84ffa6887271dbb1c35f27229f5a21033c58
103902 F20110115_AADAIA dunsmore_l_Page_037.jp2
996d27acaeb37b3d612cf6a254058359
d1cb92af773b0534f284e44eac0f46190267ab5b
26329 F20110115_AADBLD dunsmore_l_Page_164.QC.jpg
044536a89bffd21d9d13f9c9c2948315
66cd8c7666d83691f33e4ebf9a19d8c6b089cd1f
6279 F20110115_AADBKP dunsmore_l_Page_063thm.jpg
1aa30fc4489a3263373fb6521cae3d31
aea9613b3653f8c6ea502f113478e25c9a8031aa
111243 F20110115_AADAHN dunsmore_l_Page_024.jp2
29730264611d72fda1ebe0834d707bd4
06a9af7353b69a80717bc62ab83b13aaf78e4ca0
1051965 F20110115_AADAGY dunsmore_l_Page_009.jp2
6c58bce62997b14908c17dde44552cec
2785c4615c92484da281259147a2db9f7322699d
110281 F20110115_AADAIB dunsmore_l_Page_038.jp2
2830ba783fc9149b6a7c435e84bf63e7
c41687e0bcfe2e7b37380e23f4937fc2089c00e7
6603 F20110115_AADBLE dunsmore_l_Page_098thm.jpg
0c4453a9666a4389cd551bdcc9035760
2e975ab422cd020d8dca5d3edf410b61b9eaa5f5
24341 F20110115_AADBKQ dunsmore_l_Page_032.QC.jpg
b68be5ec4ef64ac3e0af70be9444a7db
ee8dc74cf59aade8d9e994ccab5337d731310304
108804 F20110115_AADAHO dunsmore_l_Page_025.jp2
7b79a4c63b2c126db30892dfd8ecc1b9
3ef07fa05c13a8f7502177cbdf81f75feb5d519d
1051952 F20110115_AADAGZ dunsmore_l_Page_010.jp2
b95170c599b515804994a7d8b8fdec69
6fa612dccb597ee8dc16d4fabe78532650dd23a1
112240 F20110115_AADAIC dunsmore_l_Page_039.jp2
8538b68b7a053424b33cd9e208ca3f4f
1d0545224f98cd411575778d7dae6cec7c4ba4d3
24137 F20110115_AADBLF dunsmore_l_Page_122.QC.jpg
c13fb3286acb5ff33a9bc1e91d5bf08d
f1136d1b0c341a343c92e99418636b295b447ea0
6404 F20110115_AADBKR dunsmore_l_Page_100thm.jpg
b930edc03a58b9d998e4c43967cd4a8a
609a3a555cb797829d0f3eb19efebf4cea29943c
110563 F20110115_AADAHP dunsmore_l_Page_026.jp2
706da78d44219f88d64d0230baa1a0be
8a19e73aa2683b197548d84d083d13774a6df56c
112153 F20110115_AADAID dunsmore_l_Page_040.jp2
893f615c43d2049f410c24d7d2248f64
e96cbdb39e5ffc19e0e26f9911f3c513c1337656
24136 F20110115_AADBLG dunsmore_l_Page_107.QC.jpg
70511323e1525fab566e632c74f187fd
3c2cf7331812124c3366e3e15142651a2a008447
20654 F20110115_AADBKS dunsmore_l_Page_004.QC.jpg
e8af045b93f2ce5f7f40851d82fc59f9
0d6b99eaaaa093c260f2da0acca4a7c54ddebd32
91105 F20110115_AADAHQ dunsmore_l_Page_027.jp2
142a2a8b5f182d55056cd9117f4422c1
06d75f1dd95f38d5329d2d26d0366794d5a8f7b0
112759 F20110115_AADAIE dunsmore_l_Page_041.jp2
2f111208e9dc16fe51e07d8227fd9128
ad0f53ea4e41a62f2216692d81e42041f44679f2
6266 F20110115_AADBLH dunsmore_l_Page_092thm.jpg
04696f5a50ec75c6a78ce40edbf8cc2c
d6c589d4e19ed72564cd8bbab80f79162b20900d
11869 F20110115_AADBKT dunsmore_l_Page_083.QC.jpg
0e818be4332dab6319525a036a72f752
2c9857dcc066c33afece1a83c1c99b50026274dc
99787 F20110115_AADAHR dunsmore_l_Page_028.jp2
4e78168fdb0cdf60e298e4dcce0a5508
1a51213cee867dc83ea88a31b057be10435b8103
111273 F20110115_AADAIF dunsmore_l_Page_042.jp2
4f8acc67e13d8116eb4acc55ca0a8b83
d2d651a5ad11569bc3ad0e8135decf0c3cb1a4c9
6938 F20110115_AADBLI dunsmore_l_Page_165thm.jpg
3f29419e859738de0ad9692320959e56
6a42c7553953f877baaf337b44d815b22d19e77e
20264 F20110115_AADBKU dunsmore_l_Page_089.QC.jpg
ec9b6a381f5ff528254e376be70dfeba
f8d13bd7def4639e8ee3460f2ff44dad0d662e15
112988 F20110115_AADAHS dunsmore_l_Page_029.jp2
8d22d027d9c4bc75c38644ff81b80cf0
b6930caa1d339d2737d3f736daf8f6c24a0c8719
105368 F20110115_AADAIG dunsmore_l_Page_043.jp2
dfd0b1a76cee69863f8b2d891c44f748
6e12b45245b6b9555ab10cba38993b88a97e01ae
6240 F20110115_AADBLJ dunsmore_l_Page_062thm.jpg
75ed6a77b46f67a06022f2993d4f946b
94ca14e42a5f763db1f1e6f84403f4f14ec0adf1
24154 F20110115_AADBKV dunsmore_l_Page_021.QC.jpg
0ce0a7f5cee4a670621381756c950dcf
64e79f2eaaefa432a2279fc614aae430af96fcb9
112787 F20110115_AADAHT dunsmore_l_Page_030.jp2
ab0a51644b51d28f96550fc6d0d0e334
e07e1cb06f09ad0d95e7c93ee71af844f94e9bc0
112722 F20110115_AADAIH dunsmore_l_Page_044.jp2
3c61604a0312e19a3053519c5a6dba9b
0a285d3756f152bd6f590086089654019b08e820
24170 F20110115_AADBLK dunsmore_l_Page_117.QC.jpg
f29e5c9dd3e2f2d7be498caf971417f1
33ca00263768b182c75e9b3bd65523ee51250d7c
110358 F20110115_AADAII dunsmore_l_Page_045.jp2
99aefdc75f96755d30dab0aa644e7d61
e9329258cee907e2f37d3859e3842f4c989731a9
6418 F20110115_AADBLL dunsmore_l_Page_031thm.jpg
656bcce1dc4df02790cb66206b9044c8
32b0eac78ffef26e6262c3c11ac61789ff427c78
6649 F20110115_AADBKW dunsmore_l_Page_111thm.jpg
1eed57980e7d71760f388b4061116e6d
1fa609ea85eab1ce594ac84ea72287baa04b2e36
109317 F20110115_AADAHU dunsmore_l_Page_031.jp2
9c0e7f73046ac5d53dbb2ded273498e4
69b88a8d7cc6043ad63ae0404f75ace02c4a38df
112383 F20110115_AADAIJ dunsmore_l_Page_046.jp2
8cebdbf5629947f23406e7ab975718a4
5739b887073256919ff53f24e659f282ffb2f249
9024 F20110115_AADBMA dunsmore_l_Page_058.QC.jpg
4b4a906bc8d9bc14c728f069bd0e984d
bd37fa0441342518b68fbe235586e57d03c1eaf1
23437 F20110115_AADBLM dunsmore_l_Page_014.QC.jpg
1cb17b56174cb313e574cecb0db2b593
7fbd194ba466d1a3713cc6f6fbb3c0026c1a3a0c
26409 F20110115_AADBKX dunsmore_l_Page_050.QC.jpg
c1beec8d8e34c7723e5a9631ef1e810b
d670487d03f5ed76635550c53a8ab82567d8bf41
113543 F20110115_AADAHV dunsmore_l_Page_032.jp2
f1c1743013b6eeb57d56fb47bd984d6c
bf4fa930024896c340769389a7bee3407aed60c4
36373 F20110115_AADAIK dunsmore_l_Page_047.jp2
f5edef0195e6917a58141773c3a9ff20
394e9adbb4c3a667f0db0ce3cc9a7f3ffe79a931
11140 F20110115_AADBMB dunsmore_l_Page_052.QC.jpg
9222e92ce480e4a1ab927fadc4b7c7db
285017e6ccb141f25d7e02f8bd1d4db711e8f5c1
6562 F20110115_AADBLN dunsmore_l_Page_116thm.jpg
616e66dee73b7f1366f755540e414d13
0584d2ce5104ce983331581c804a24e529e0bb77
12402 F20110115_AADBKY dunsmore_l_Page_124.QC.jpg
4ab818cb7e8c4f70961b8aad9aeb4889
9a48fbd07e9505b841f23214ab4d2acf5cf4c927
108795 F20110115_AADAHW dunsmore_l_Page_033.jp2
6a201dc48c684f802cc75846af6caba6
c21216775f9a26f5056e294b477bfa2d6b3045fe
539929 F20110115_AADAIL dunsmore_l_Page_048.jp2
2ce1643cc330bd901d9e6b305dc12f9e
d9561a6e43b02f92b75dbffa22f3bd1ae83864ac
4939 F20110115_AADBMC dunsmore_l_Page_049thm.jpg
926d34e9adb8775f73c87487185d7241
aa31142e595b32e35d3299919ef2184d5c62333e
17381 F20110115_AADBLO dunsmore_l_Page_056.QC.jpg
5a81023590ed9adf294174fb3c7eda8f
305f903ad5de1a3e138bb64b87ab2ea994fa36cd
23187 F20110115_AADBKZ dunsmore_l_Page_031.QC.jpg
b8fb5c2d178a8ab7dcdfd7ca14e0b58a
5b5562dd13777bafb35aff05a1d13e1a4ff906f0
112866 F20110115_AADAHX dunsmore_l_Page_034.jp2
c032898f6ad2bccf7529a7269cfc143e
6076ae779d9e0041e6d1f74bec47b7ad39e8eb52
109372 F20110115_AADAJA dunsmore_l_Page_063.jp2
f86654e54c24724a1d06a157dcc71972
eb0847d6565d3bc9100e59fa2a074f1ea11f53c1
879393 F20110115_AADAIM dunsmore_l_Page_049.jp2
82c6b12381d7c317deb8cae34cf2c614
db4ea114ae141e3bb45397b246ab74e5f4eaec49
10368 F20110115_AADBMD dunsmore_l_Page_129.QC.jpg
0e731ccd9d4c68da8e624d2d6437df2a
75a1e08cb5f0b5ef215591a36cd3d2aa4496afda
6066 F20110115_AADBLP dunsmore_l_Page_082.QC.jpg
6070dd8990e57c2338b1177040e6e3d3
034930b361e15935f3e4b7391504fb88dcd9475b
106016 F20110115_AADAHY dunsmore_l_Page_035.jp2
f05f571e0b81954cff6601fb919336a6
c14e2afc4db626145bcdb35a49c23af8bc40c8e6
114641 F20110115_AADAJB dunsmore_l_Page_064.jp2
b1e8cfc4fb1a6ff5e79f0454798029f0
81b907ce2bad6c7702375952e89231e15d3fd0c6
1051969 F20110115_AADAIN dunsmore_l_Page_050.jp2
8ea73bf3774d507dc99a3b3a220e7f48
418b9104be07260af6d4b8ab4cfb923e1ca19cf0
6623 F20110115_AADBME dunsmore_l_Page_067thm.jpg
1ff71dc42b97b0d361035661823078a1
adf1410872d64b2a832a0b3f19c1b785ce563cc7
6920 F20110115_AADBLQ dunsmore_l_Page_164thm.jpg
69d281c70c204f0763497d67672565e6
96bd82b19b46591de5e441844fdda772e0143c0e
109502 F20110115_AADAHZ dunsmore_l_Page_036.jp2
f37cf1e2b2c12b2c0d18d9e40e363085
4c480e2cfcdd6b79986648595e2219fd4b31ff2a
111288 F20110115_AADAJC dunsmore_l_Page_065.jp2
a88cf23252e3ec6d2e1370550d2d6677
fa1444e5a5d5b34d24088036db41e48d3c937a69
1051944 F20110115_AADAIO dunsmore_l_Page_051.jp2
82996e758bdb03d95b66777e75b5e0bc
358c80624319c836d5dfa7c47b55f5e66a965b8c
6498 F20110115_AADBMF dunsmore_l_Page_046thm.jpg
15999ca7f7e5ef08cb5e10129f4a3bd1
310942c7b8a4ea06f0f1d309e94526d9a4700ce8
3986 F20110115_AADBLR dunsmore_l_Page_054thm.jpg
812516fae28b981475810a9bc41455fc
f722067934df16ac56fb25d232f1129e445bedd4
92371 F20110115_AADAJD dunsmore_l_Page_066.jp2
e94c87ec17b78bc5e82a3434fcf091d6
fe729dc538a022e310d9f5d629f1fdfe035b9628
455242 F20110115_AADAIP dunsmore_l_Page_052.jp2
e249243ff297c64439157f103980c71b
2c5dc490e9d9ab1e1c08a9b956abb9bf40cc719c
3497 F20110115_AADBMG dunsmore_l_Page_133thm.jpg
498c3ee0b46fe0f0ab09620f6f26be99
d7872c662ccba1598697977352fcbd9d7b7e9d09
3311 F20110115_AADBLS dunsmore_l_Page_085thm.jpg
3b1a486fabfedbfa7caaa38eb62fddca
91596923fd11da2261795d97d867e33c258cf67f
111952 F20110115_AADAJE dunsmore_l_Page_067.jp2
975a08e04b63ecc4e710534dbfd6c4a6
c8693473366978cdd3831e58ae1a7a04a0a7af0e
814003 F20110115_AADAIQ dunsmore_l_Page_053.jp2
f00d05880d6334f5dab0e7149e723a3c
8d6735a13d26710bd208df86261b3aacce8ae1c3
6605 F20110115_AADBMH dunsmore_l_Page_141thm.jpg
11260c7f3821792e61a58a213e20dd2e
b3b8003e19c1986389eaf6d3e7b4047cad26d0f7
3123 F20110115_AADBLT dunsmore_l_Page_181thm.jpg
5f2413b8a7d1b971b6e96d8d991a8723
cc0036ee5eb7a9b16b5af56956173fcd2b544c63
108029 F20110115_AADAJF dunsmore_l_Page_068.jp2
6aade3f8c6d07322794188f85bddaf18
c7bdc02a8ef5d889c3c64c6c2bcd1b070ac60aa2
686575 F20110115_AADAIR dunsmore_l_Page_054.jp2
dfc3e8924cf47baf766005cd0c42dbb0
25445532a7672f25c36e3cc42d33f6cf5ea8fd05
24624 F20110115_AADBMI dunsmore_l_Page_080.QC.jpg
a4147dd986708ac423717a3d4a66a6cf
915accf5beb797ac2ae6c915991f8f24cf6b9bbb
23094 F20110115_AADBLU dunsmore_l_Page_121.QC.jpg
85f1756175f7e9ebcdb5ef48188b4d21
2ea0e88a0355b0333b17de1f35995b4d978f1ec0
113669 F20110115_AADAJG dunsmore_l_Page_069.jp2
6b279e09113671e773137a6b554b067f
98e8b80e6cd19fbd1814f086d2fe291fa8be6daf
777571 F20110115_AADAIS dunsmore_l_Page_055.jp2
f7964eae802dc31aedc95b293dab2363
ae5e41b478202288613bb3c4a9c5bb65f84b08e2
6782 F20110115_AADBMJ dunsmore_l_Page_170thm.jpg
d9d7929aae7b5bd1ec0c8d3b0a20ba41
a3722456efbce4a7ab5732f2c500218564bb1187
6584 F20110115_AADBLV dunsmore_l_Page_022thm.jpg
de0a22631921156cdc30fbb1e0abaa4e
15b6202ab0897de751855826ace2b18d03ec2373
97475 F20110115_AADAJH dunsmore_l_Page_070.jp2
5bc2be71b63060d90405bb362ae2c3c6
034898e17eb22b25e6961e29fd6a3d035197bac3
800836 F20110115_AADAIT dunsmore_l_Page_056.jp2
e207f3c6af1aa8a36cf6e4e31db0094a
c7f4a4e882df5e83991b4f0694d58c50b3d89d7f
6484 F20110115_AADBMK dunsmore_l_Page_097thm.jpg
39db629987dfd5a619c052c6cfc73860
e5acd76ed321d26c8a980795cb25aa63476d0930
24374 F20110115_AADBLW dunsmore_l_Page_104.QC.jpg
f4b2687d2b455e6a067edd571e7d9c4c
affcfca786c61d85d6874ecfcaa7e7be5a1f3279
103515 F20110115_AADAJI dunsmore_l_Page_071.jp2
bb0e44976c4cd36075dba9ff999bebe2
e0abdd7702067bb9d662d547f1b38214edc8b64a
388229 F20110115_AADAIU dunsmore_l_Page_057.jp2
8583473902b86f6b1ed00dbc2884931a
ae32edc5771ab2f30178aacc0e1385e94e5fb0b4
27069 F20110115_AADBML dunsmore_l_Page_175.QC.jpg
ca2b25124f82d6fb2fe3f57a06ccf682
aa283d7f9398181c05a71797e5a4006b6ee42fea
101918 F20110115_AADAJJ dunsmore_l_Page_072.jp2
2ca6c07c97245a92fc9f16b4aee7aba2
c948b329c6288849cbb55b50516b07ddcfa75856
6285 F20110115_AADBNA dunsmore_l_Page_144thm.jpg
db00043015b48f733e73c7a6eeb1728d
937ebd0e8c068a4d583e67b1ce461e5c50504d58
5914 F20110115_AADBMM dunsmore_l_Page_014thm.jpg
be204ea6a3bb49cbe509acb7ded4d883
9d165bcb9e7b9894d0afcd82f5db1e356145cdf6
5789 F20110115_AADBLX dunsmore_l_Page_151thm.jpg
87217c4a4c880d4016a622ebe9a5fb3a
ca39ac24d5fbf92d8b00d4166a730c3ced741300
109384 F20110115_AADAJK dunsmore_l_Page_073.jp2
c1b73d4bf35c6928d50aaf23e51d8bbd
4a638d1ef6a6ab03678b8d8b9e4ca46ac04f7384
717153 F20110115_AADAIV dunsmore_l_Page_058.jp2
a00015caf403a66df0756943600a31a0
623e0019dd68a6a5a4703f8dfd05f9f6deb273d2
29993 F20110115_AADBNB dunsmore_l_Page_012.QC.jpg
a1527d5396c92cf0e04a0f0f996d42c2
b4345ddd65c49d6937dc2153d375cd22da17feeb
6624 F20110115_AADBMN dunsmore_l_Page_081thm.jpg
46f7799583389f7bfe06015fc87f17b8
dd6dd9010113189918d00f01150f0ec98de3138c
2634 F20110115_AADBLY dunsmore_l_Page_058thm.jpg
35e7df5b2946b5bcb83452a023f4d476
e3ff2d65ff25e4ca58c4eeafe6154bfcb96a6d8f
116009 F20110115_AADAJL dunsmore_l_Page_074.jp2
66a2b1fb214b347a19d50c0b1cf021af
31c72bc41828b7a52c3dc7c62f52d9b6658671e3
558840 F20110115_AADAIW dunsmore_l_Page_059.jp2
7fd25e7f5994cb740aadcbd08ac22fcc
976137e83e29b05588cd0ff951c42b99af8bebbc
21365 F20110115_AADBNC dunsmore_l_Page_094.QC.jpg
16aeca6785a8a38dd9b29161b6af29c2
57b1d2967f63f2917988eeb0946e343839bfbfa2
14570 F20110115_AADBMO dunsmore_l_Page_054.QC.jpg
5d8adb16c7d94af27d3dc714ea2b8dbf
aa2baac0e981b5f30cb4995e61d9d5738a51f2ff
6338 F20110115_AADBLZ dunsmore_l_Page_043thm.jpg
9d35f756932068a6065e10e15e85ca08
52c226f822a7e68bfd80e53bd2d43f49c23ecd17
1042964 F20110115_AADAKA dunsmore_l_Page_089.jp2
1f16a7ec71bac7977c646dee00e18379
9bd26150672e29758805f8dff744e675efbeae7a
112967 F20110115_AADAJM dunsmore_l_Page_075.jp2
1a77fef69441dcf489543bbb962c6551
c2c4ef93e56568b62a0da8001337cdb4bd34249f
308504 F20110115_AADAIX dunsmore_l_Page_060.jp2
a84d120f45b8d90bb7d069f11da18d40
e8dcc998882c06a376fc2cc617a22616c89057d5
2620 F20110115_AADBND dunsmore_l_Page_008thm.jpg
922a54b148bdf0b8a637493b470fcf15
9c5ea270352d46c0641faaf65286ba538b95a606
10097 F20110115_AADBMP dunsmore_l_Page_181.QC.jpg
6c4333a9f76c4d91aec8be0523cdb51d
c6f1220ff5a1aa1ddfd9c0b151a5da5d7055f785
721038 F20110115_AADAKB dunsmore_l_Page_090.jp2
f8d28a2bcbb2f8de1fa5f98e016c44b7
740e64fc475a51d003bcd84e5e88d585c80ee4a5
110558 F20110115_AADAJN dunsmore_l_Page_076.jp2
af95734a83c598eb65af5de975e2812a
3cbed64fae07668cfd4d3b451989a2013a3a6819
700224 F20110115_AADAIY dunsmore_l_Page_061.jp2
447b99f8acad845aa279b70946f99de3
2ddf6a5c0746200b74ef523d4aed0f2658dec90c
5990 F20110115_AADBNE dunsmore_l_Page_018thm.jpg
07c86c45b3ceb2f4fd949340b9e212a7
3ae6c5bcb6412184ad2d971a9815bcbd1ba18175
24101 F20110115_AADBMQ dunsmore_l_Page_115.QC.jpg
37ca9540fedd42ecba8a2973a2db6eda
34b3903bd29da960a70d1f0777f3fba890eafa55
1051938 F20110115_AADAKC dunsmore_l_Page_091.jp2
7b2f9589282e96ade3d8e922cff325f9
0e9c3cef26590b33f7d9e05dfba4e5a47fa65e83
105759 F20110115_AADAJO dunsmore_l_Page_077.jp2
9225e901a010e2deca2ce047adefc401
a94c6f22873cebb0845a17643843fd69602d2dcf
105256 F20110115_AADAIZ dunsmore_l_Page_062.jp2
fc589f78b3c1ed3c101965b20cba7d27
fbad679e3b6f7e7f9733c465b7c5d7c563a9177b
23581 F20110115_AADBNF dunsmore_l_Page_144.QC.jpg
ad7759cba95f53879bfb74be7f56b32d
93ae9b5c3dc30f9cff6edb0ba4cfa1abdd71c4e0
2791 F20110115_AADBMR dunsmore_l_Page_059thm.jpg
280687b8026603639f7fba11733b7e47
59a4b6c083f094f09d756fb8de8161b4ddc9629f
1051971 F20110115_AADAKD dunsmore_l_Page_092.jp2
0c08931fcb5f22a29a4eeb6dd515bafe
a139305121559c0f82131b935db3b1ce614a3f81
110825 F20110115_AADAJP dunsmore_l_Page_078.jp2
44d1ad551c607056417396a8026a4ddb
3027cbb644a56f4cd0960d8f1570e0bb9a32ffd1
23980 F20110115_AADBNG dunsmore_l_Page_020.QC.jpg
691e929a5bb2b08e4556e70006f1426c
cf457973c0486b46936ade2f17433d6d910f1d60
3416 F20110115_AADBMS dunsmore_l_Page_084thm.jpg
6b75ce12576733472ca54109074feaa9
d6821a20a3d919e566d3f820724ca5d3564bc371
1051941 F20110115_AADAKE dunsmore_l_Page_093.jp2
7abadff71ccabb22b5f606f04808e0d6
1b4eb418b71c9288c7eae5b3058af87a539eedbc
111131 F20110115_AADAJQ dunsmore_l_Page_079.jp2
037dd60fc0c3b6f4558ef3d3473f95d9
40b26c6ff8caf2103a6e326ea57032b85759d4e0
9873 F20110115_AADBNH dunsmore_l_Page_057.QC.jpg
960ffae41172a6dfe350dee7ed24d84c
c6bd7d594ef3217d9261ed6226e495b9bd10a30d
5126 F20110115_AADBMT dunsmore_l_Page_150thm.jpg
f9ea175388e0df8385cec3615f640211
cb51d062ff8cd5b15587a8c1edf7b6bded6dc5a6
115887 F20110115_AADAJR dunsmore_l_Page_080.jp2
6c86635d326710b2715ab960cf9637a6
a4365012aae2c545db9b75fe5c1208e50e29169c
955863 F20110115_AADAKF dunsmore_l_Page_094.jp2
5c662f9e87150f9b6f98c371a17e1d53
75463aa49850b2344f2ffa6d20a8fe4b38bee920
6597 F20110115_AADBNI dunsmore_l_Page_069thm.jpg
bbfdc5ee92c302801c5297742a0c5fe7
0d219785b5db8169dbd9c583f4638649f5405676
23698 F20110115_AADBMU dunsmore_l_Page_141.QC.jpg
9eb24a64b6f31819c120269de9c2685c
e5ecbd97a4cd8c538af93a040a0905e1b6c7c558
113743 F20110115_AADAJS dunsmore_l_Page_081.jp2
0fd728c282415cb5652e977c31043fe5
14ee066d984f1052885002e7f8d6d3202b772928
1011620 F20110115_AADAKG dunsmore_l_Page_095.jp2
e5b33171d768f8cc0ed76596a2bbd3ed
a36b50f6124f165e78af0c4f02b2b9648bdf91fb
5928 F20110115_AADBNJ dunsmore_l_Page_137thm.jpg
43a9df6e0587a97aeebcf48dc1c503fa
8fc90a2f7dcfe400d0160cfb7a6dae543a7a0cab
22906 F20110115_AADBMV dunsmore_l_Page_159.QC.jpg
07eaa1ef710584d28495faa10f9f1d00
a23c18b91d4d6a47230e01c3ff2fe223c15c1f57
21154 F20110115_AADAJT dunsmore_l_Page_082.jp2
d46e6d99c1638a6e47cf9d26f5c2a238
85472c06786c1ea17739b36b65513a3ab19ef5b1
103965 F20110115_AADAKH dunsmore_l_Page_096.jp2
9d2da4b0c3d046690b3d278c86adf6d4
f42b96386a445deaf2a684cb7c4a7d6eb65f4a98
23677 F20110115_AADBNK dunsmore_l_Page_024.QC.jpg
bda8381eca2df57154fd66f289eb1413
a64ed3e111eebc31097a16eac8f5054d3c5b375c
6750 F20110115_AADBMW dunsmore_l_Page_149thm.jpg
7869ea1c34d5fe51cc41ff157dfff8e7
d8c67f68d2551f61624fb66676484e03b611562f
537016 F20110115_AADAJU dunsmore_l_Page_083.jp2
c7b8ee745e6e5af60665841941eccf9c
991ad584d3b9dc3c35ce50c91976a69683a20266
108579 F20110115_AADAKI dunsmore_l_Page_097.jp2
7adcf5ea8832a3e2226ac3f90500d9a8
a30354f6aa35e4fc5c0305e26d8b18fcf7ca1652
24832 F20110115_AADBNL dunsmore_l_Page_074.QC.jpg
6daa8c6805f9a5ad9260c2201c5af823
4daab17ba93969f1e07a7cf80796b1789bdaf363
6711 F20110115_AADBMX dunsmore_l_Page_104thm.jpg
8ae6c9a7721583cad008186f3c645879
ff166c6f8fa74dbca3a3f9cc1696bef3f725db0d
359889 F20110115_AADAJV dunsmore_l_Page_084.jp2
9099be6190354b5b25def74092e3a6f3
8b1566b8e16f9ca4fe722e02b3b0909ea60927eb
113134 F20110115_AADAKJ dunsmore_l_Page_098.jp2
c042aa14501262e1ba371a98f117186b
ee0581deaee9c53744b717173b2ecf999f1d854b
24998 F20110115_AADBOA dunsmore_l_Page_170.QC.jpg
4bf330a6c30432bb1c1f912998fae51f
7aadb96d5491e201ff543ea79893b54bf2d41ea4
6629 F20110115_AADBNM dunsmore_l_Page_166thm.jpg
d7f28a3854d434dfd0d976ce798a4cdf
1e34f94da0354d7e6467fb5e161e30519ddfd849
115568 F20110115_AADAKK dunsmore_l_Page_099.jp2
4d44f2dd451a4ec205930e44a2f2c030
addf2ff3ddce68a6241eb0f960e9b9487f534478
6841 F20110115_AADBOB dunsmore_l_Page_155thm.jpg
1a429f5ec0ea8c31e5248f330efb397b
c82743219d17e2b519f74dcea1e6ec12a2100d99
24288 F20110115_AADBNN dunsmore_l_Page_106.QC.jpg
143c1fb52cd35d3e2b14f8f2162fa5ee
94e5a319b2840624462c934772a20a4fb4ba617b
6676 F20110115_AADBMY dunsmore_l_Page_105thm.jpg
c327c1e222a0b1d9f6ff40fd00cab22b
0688b79ca12949a742027e5d6826c9eb3e451f80
448571 F20110115_AADAJW dunsmore_l_Page_085.jp2
f4cf8b8babf77a02f1103ec9c67deddd
1cbd09a2caf8c4e58e9a730590450fa28f1716ec
110552 F20110115_AADAKL dunsmore_l_Page_100.jp2
b989447535f112c139a2cd36fd53e087
5107762f22f56e7a209e238434bd483b35ecc4e1
9081 F20110115_AADBOC dunsmore_l_Page_047.QC.jpg
1f44ad29a895c51adb17e48a99f22496
a6c369a969348bcc07e222bfc50e3cafaf50bd59
23987 F20110115_AADBNO dunsmore_l_Page_166.QC.jpg
fcdf3b6892dfe2c76006c345de3ff8e2
503b2f9da20d83b0e214f577980febca045fa5e8
19878 F20110115_AADBMZ dunsmore_l_Page_161.QC.jpg
097cc6faa18d159c22c2fde239299667
a18ed67fa9348ed8a36cce1ee44f2d6e8f2eff8a
387162 F20110115_AADAJX dunsmore_l_Page_086.jp2
3a22d344f5ee9a88e3c1c910bfe44c69
41dba1518491335734dfce45556dfe2a81c65e6e
113999 F20110115_AADALA dunsmore_l_Page_116.jp2
8604cd4a087620abd6786cc482d2d6cf
b948f0ae91607b5f046f70df40909a7e74ff013c
110206 F20110115_AADAKM dunsmore_l_Page_101.jp2
43f9268e75f0245a5c58ae85185f254d
c96e9362d2acd9d56b5d393a8e929b55c52f0a46
23630 F20110115_AADBOD dunsmore_l_Page_065.QC.jpg
e13dddaf35c9a333bef0b513d16f8e65
787cfd9d684529ab95402add93044a887cb972a2
6647 F20110115_AADBNP dunsmore_l_Page_122thm.jpg
5cbeecf3a2ab51a45802c7c9b49f9d57
6b7c19212fa47e8d09b0704b52dfaea9e23ccf06
596167 F20110115_AADAJY dunsmore_l_Page_087.jp2
68d0070b5312f3278ae240b996e0ff20
c89d5e7b4a61c7264cc0aa3485ca04981bbfe090
114391 F20110115_AADALB dunsmore_l_Page_117.jp2
953bd550943c216d6738c4da65e4658b
c04decf823866c134fbf064afc0cdb244bf333d8
105878 F20110115_AADAKN dunsmore_l_Page_102.jp2
c54226ef724a8f96d4cc04dbc97d1bb9
3c9d85ccfaf7b9ffe0a6317d83736fdfeb6a3d32
8380 F20110115_AADBOE dunsmore_l_Page_001.QC.jpg
5b9b4aa2732187a5240bf35b3bb06d9b
9a57e3f614f27a8d012a8e393478de493b38f843
23483 F20110115_AADBNQ dunsmore_l_Page_147.QC.jpg
ba9c8f2a1240a8ed3900cbbe4d4239d1
03557be88d616625d5e4de6975f596bf4da421a7
1024219 F20110115_AADAJZ dunsmore_l_Page_088.jp2
8865247198b9f65a9112d7870d909d1a
758c0e07e592466b0ddaffc730429c185305330e
109209 F20110115_AADALC dunsmore_l_Page_118.jp2
1061695b64cc0547b60e691f5b461110
81cd929191fc82922287f74342b088ec4ed6c41f
113519 F20110115_AADAKO dunsmore_l_Page_103.jp2
ef0c7ad30e634c4b47ef856fc48d31e4
85b3025f8f81030d3ff17acb589cfc23270be6f4
3729 F20110115_AADBOF dunsmore_l_Page_136thm.jpg
a8b785596e967c8df0dd9333725b0749
393feded724930b693e90272281fe024b852b9e7
15037 F20110115_AADBNR dunsmore_l_Page_087.QC.jpg
596f913794f497325384970db87cc4d8
68f9cecc340f0b3a5c1206100f09090041a94cca
111418 F20110115_AADALD dunsmore_l_Page_119.jp2
91ff5c252139a041b5842eff008cba84
ab52ffd3c6e6f52a28094d7b9b8b9a936a45374c
111396 F20110115_AADAKP dunsmore_l_Page_104.jp2
dd94e837c3f155c4ef7f39e5cfd752c9
4ffe624f9b5a73353dc2d6e970f3e3c81b898e35
23715 F20110115_AADBOG dunsmore_l_Page_176.QC.jpg
bdfd88bd1ec417df65dc8267d77652b1
4da3da6de330617e9357cac6cafb86f6416a1276
F20110115_AADBNS dunsmore_l_Page_026.QC.jpg
e413446179ff81f1d059ef1c1f8f4387
6c92a6fc122790a14e709a6978147e35bdd1937b
102764 F20110115_AADALE dunsmore_l_Page_120.jp2
bc8c623fc1b00a219211e425a7c42721
f3abdeba90cd27863bfce26f1bae0b687656e6a6
115840 F20110115_AADAKQ dunsmore_l_Page_105.jp2
f103cd9833e7f6b9924ab6102800d3ea
5452eb406258500ac42a7b1b6b9cca5eb16d3dfa
10678 F20110115_AADBOH dunsmore_l_Page_127.QC.jpg
e01b0bc05736211f3f7de6eaf00199c3
f33fc75e30e1af8b95df61eb11959d157a2ec8bf
3679 F20110115_AADBNT dunsmore_l_Page_048thm.jpg
84e8efdd691c3048d5b75eb9c0c222bf
caaa9499a575916bc8da16d7001ae2dfca06d5a5
109177 F20110115_AADALF dunsmore_l_Page_121.jp2
95fc166b1fd98d910b133c791350a242
75335616ea9406407b566624ab252205af69e4dd
112454 F20110115_AADAKR dunsmore_l_Page_106.jp2
8ccf3a8ac5aafe8899b42f390ea1ab20
27d7f4397375432d7c0db884f2596acbf35a7001
20658 F20110115_AADBOI dunsmore_l_Page_160.QC.jpg
bf9dff917cd34b4b24a136f3fa60a40e
08978be07012492ce4300c45f75fc6fdebac86c6
24864 F20110115_AADBNU dunsmore_l_Page_108.QC.jpg
d4286ba6609b2a2d1ecc1b48771fa5e4
9d39ec56e0f31cd260926ca15a2fcdf0ffd7c2d3
112591 F20110115_AADALG dunsmore_l_Page_122.jp2
00ce11be1a9666210a27182c9abebeba
81fb60aa2a257260551b4265dd9c9bb90df9caed
114415 F20110115_AADAKS dunsmore_l_Page_108.jp2
35ee40385ea48f3f3cac6f46b1f19fc0
deed6cbd6cc8803782a06c5d7d72ac9aa94c56cc
6622 F20110115_AADBOJ dunsmore_l_Page_139thm.jpg
041d7e0bac4654fcee0fa6ca17aa10ad
19fc5e50472d1e553d7a61b341f0f1e70c359bf7
5483 F20110115_AADBNV dunsmore_l_Page_027thm.jpg
25725b636b52ae7a898d6110351f8557
e94fa9f00afc65810da7346d798cc9403c39bde2
71279 F20110115_AADALH dunsmore_l_Page_123.jp2
fd85c71a08ff3c2c04b1090762dce619
22c6573736e7204ad56f847ec8417db202e5ff93
109737 F20110115_AADAKT dunsmore_l_Page_109.jp2
ebe8243b1cb4e1053a1c5251b31c7438
1e60e2ff230ac444d772f4458e1d0d27d3dc8b3a
24545 F20110115_AADBOK dunsmore_l_Page_011.QC.jpg
bc826b969d0ff0119a86a6e5f329973d
400c216d8428d47b17cf7f14663f62d0e744f978
23556 F20110115_AADBNW dunsmore_l_Page_101.QC.jpg
81c965da1de960dc72353662d5c10676
84195eafc5f637c716de06f306c9a92c3febf617
647744 F20110115_AADALI dunsmore_l_Page_124.jp2
6ef86520b887e653292bbfccfa36c4af
e4165b13288e308d3dfb43bf38b134f3c0639a11
99960 F20110115_AADAKU dunsmore_l_Page_110.jp2
1f898905a6a19f20d356f2f7da33d9cb
79b0cf8daa3f6f4460a7186eb2ac616466c505d2
24231 F20110115_AADBOL dunsmore_l_Page_116.QC.jpg
fd1dd6a0cc04582d149d38a56ea4f88a
4fa6a2af058e97903e4758165be107513e70d2f1
23325 F20110115_AADBNX dunsmore_l_Page_139.QC.jpg
53f975d0feb285cbaf479277fd713185
1d2df0fb6a23192525e653a16173861021bb838e
610774 F20110115_AADALJ dunsmore_l_Page_125.jp2
ddb3eac8bc4720f49033c5c59de76f74
b93a09ea1a2d4f826d6b01ef179c958322c73993
113726 F20110115_AADAKV dunsmore_l_Page_111.jp2
f598dac16e7651ca3e6dead4c7523e0a
3d220d8b972dbf747c1c73df7e80fec2e2e5ea42
F20110115_AADBPA dunsmore_l_Page_117thm.jpg
b90085e79e6c7a0a510ab88e1c2c8b4c
3ab3452ebc5cce5eb4160ad8ef28488fafa9fed5
7928 F20110115_AADBOM dunsmore_l_Page_126.QC.jpg
ce2c3080969173d6e05d886112d6346d
f8dcf2ed07f5b130f61d79d8e269fd03dfc600e0
11378 F20110115_AADBNY dunsmore_l_Page_179.QC.jpg
83108e608d589fb70e345347a8727af7
999eb6036ac0e6541ac7e82a7c22b420a298cd19
283390 F20110115_AADALK dunsmore_l_Page_126.jp2
7644ac381bcdf32bde60922459f3cad1
202bf0f11e74a7ce222ce46438e2093678467c46
106762 F20110115_AADAKW dunsmore_l_Page_112.jp2
33de44ca7f305328d9826abf969e3c53
85b08b31560f160bd75260e9655d1a1ee870c92e
17911 F20110115_AADBPB dunsmore_l_Page_150.QC.jpg
59dd4aff4c782a7d1506f3e961390ae6
e21a14bebed50e8ed59fa9af8964b08ad6385670
6646 F20110115_AADBON dunsmore_l_Page_095thm.jpg
9fd9e3a9d1c252648bfc6ed845ce4e21
47a5891b5b5199ec8075618ae792a57dffa7bd96
408600 F20110115_AADALL dunsmore_l_Page_127.jp2
5781250b1ecac67ac3adbe83d1e8f8c3
bbc381087f73eae72959b0a5f9d8b1dad57f5300
21052 F20110115_AADBPC dunsmore_l_Page_072.QC.jpg
6a66e6dab110d395e33fec3e5588ca96
70f83824e5f7dd652fa9c648c12006d894f1967b
3081 F20110115_AADBOO dunsmore_l_Page_057thm.jpg
91842e877ab6f843f7cfbc1eabf5276b
90e90e7a73bcd5cf3364eb032f2bf6128cc1e7eb
6431 F20110115_AADBNZ dunsmore_l_Page_073thm.jpg
480f769beda9f9a62f9a587bb308f06c
4fb3c7e800db66c45e57218522c4c4c785934800
110873 F20110115_AADAMA dunsmore_l_Page_142.jp2
0442d03357940d23c027a6c72f5d7d66
dbcf28cac8cc9f6e1d642cb5b14bf46a38e7019f
528108 F20110115_AADALM dunsmore_l_Page_128.jp2
8eef39b6f5eade63843ef6e3afac8763
f2038ec9ba55f73a1638471965105b882f7b62b4
110426 F20110115_AADAKX dunsmore_l_Page_113.jp2
a305ef989b5639d5a7d21df227a98f49
d053a2bed8704a00bd7f101d009752ee9ad74f55
6685 F20110115_AADBPD dunsmore_l_Page_108thm.jpg
1de0cd308b9924595d3e817c43b13c56
5f316641474c0b41b058828bf35bdfa1a6699343
22636 F20110115_AADBOP dunsmore_l_Page_063.QC.jpg
a1f7c00ee049457af843ebe2fb61141e
b38d193a767f378122a1adb7ef9c3b480fa97a1f
115455 F20110115_AADAMB dunsmore_l_Page_143.jp2
e5069cca2f565302265db177da78ee28
09094ec52c6f58f2047130b809a5ae571184b4b9
268400 F20110115_AADALN dunsmore_l_Page_129.jp2
3a28120ce6971fda4365401509b4d16f
553d9a7d0e403a79df1cbc468c43e09c5c73c5af
110546 F20110115_AADAKY dunsmore_l_Page_114.jp2
cfd2edaa0bc86f33ad0e3940a98aeb19
7246bc7fe0f6548a70e6cf993901a908d3b6f2a7
23882 F20110115_AADBPE dunsmore_l_Page_079.QC.jpg
8554715290d3c537a4a85048ed3fbdaa
2db49ceb427d8831a6f7e24b00872a2ac5fa72ae
4620 F20110115_AADBOQ dunsmore_l_Page_128thm.jpg
f68bf39aa0886a044ad521ee18b3dda0
61603d855eb2bc78272fd1bcf0743212b07e8a7c
111181 F20110115_AADAMC dunsmore_l_Page_144.jp2
981f7fd37a7d2fd59af68cbccc29aed8
997166ab48ca246be099d76f79a7cf77f5830765
305153 F20110115_AADALO dunsmore_l_Page_130.jp2
c6ad413145d0072766d556f121fcdff9
08603a4c8e63fb92f8f6ff627b772fecd1ae9797
113930 F20110115_AADAKZ dunsmore_l_Page_115.jp2
969ef98a573f69a300556a9bb1ae5975
1d62e876979b4bf5c23e7e8e7a3692c4cf70334b
24104 F20110115_AADBPF dunsmore_l_Page_039.QC.jpg
556114b9010bca4cf44639712db1d1ca
151d412baa556c73caed544acfeade91a60af205
23642 F20110115_AADBOR dunsmore_l_Page_081.QC.jpg
46c6571d66e92404af00b07287c438cd
4b34d142c87565802d1a3445c08d8fd79a0986b6
113665 F20110115_AADAMD dunsmore_l_Page_145.jp2
2fb8116f1295fddfeb1f3536c284f244
2e17a822cf344d87237724a8ddda0eeb92c1e688
52152 F20110115_AADALP dunsmore_l_Page_131.jp2
0d363df9b2bed68c554c5925d09ee7f5
fee2ecbe217c821000505d401688100562841685
6321 F20110115_AADBPG dunsmore_l_Page_120thm.jpg
a26fd14b0e05cfe2acdfbcc48be5b714
bfbde79b135d4b36f30cc390b027cac3b14d94fc
3170 F20110115_AADBOS dunsmore_l_Page_135thm.jpg
c42add2b5d8e15f977640c9485ed1475
8e4b394ab4fbba92b7383bda6816fc6acca7c1cf
114565 F20110115_AADAME dunsmore_l_Page_146.jp2
9329bbca7309b0c9939dd4f6202cec98
d4331f2c1897b6afe75cae60d7dfa79c6b83bf52
83090 F20110115_AADALQ dunsmore_l_Page_132.jp2
2f8b930ac9cae3f626816c752169cf61
ac7ced2a0f289413094670a2f7d60d564676e7f9
6532 F20110115_AADBPH dunsmore_l_Page_106thm.jpg
98314de553f4722af58f05167f54c897
4d9d04b9cb90ae700fef1115d6cd5d30aa88aa82
6699 F20110115_AADBOT dunsmore_l_Page_178thm.jpg
6f7c28b8e0169d6a3227ceda2dd3e4e9
53e460aadb280739e503c0a70887a2b71c1e7fcf
108973 F20110115_AADAMF dunsmore_l_Page_147.jp2
e703b548d24706d70140536845d7f812
4ae42e7aa9d33d719b36e1fa64440e2228cd9a97
53618 F20110115_AADALR dunsmore_l_Page_133.jp2
1122b2c053fc589f6c012880fc591b8a
e5019560dbdcfba2eecafc3bca916b4990268dae
23826 F20110115_AADBPI dunsmore_l_Page_114.QC.jpg
7a3312fe55f297f652e4459dea15130d
47427e0e756ae8a99bdac1aea16e26f6d18cd890
6490 F20110115_AADBOU dunsmore_l_Page_030thm.jpg
717feab578e3f966b33bea84fb0c4f74
6331755d8976faa1ba30b679672e7a209def25b7
112716 F20110115_AADAMG dunsmore_l_Page_148.jp2
a87547566add4c169740d9bb9aa5b71d
7bdb2f42d7038feb686974e28fb81f92fa45776b
93412 F20110115_AADALS dunsmore_l_Page_134.jp2
5a86462e0dd883bdb40e5b65c333b432
494ef526a55b7086eff9d93b48c0af202da05d18
6472 F20110115_AADBPJ dunsmore_l_Page_140thm.jpg
f51682aa633cb487e1e8962810a5af49
f91180937547ad33f5ba5abad11eca9ae66b0524
7150 F20110115_AADBOV dunsmore_l_Page_171thm.jpg
c9b6c606f076b225a0de975fdcdd2b80
e1e24a4b79a4e2dc3f5359812153aa55e327e422
113461 F20110115_AADAMH dunsmore_l_Page_149.jp2
db9f0fead4358b6d90b7cb1aa5279afb
6432905229d5a0b62abcb66fd412d7c337a1f7fd
41495 F20110115_AADALT dunsmore_l_Page_135.jp2
8df5be660ef0321236eee2a45a309acc
55433822abca77e5d6081553f6fa283081ce8f82
6144 F20110115_AADBPK dunsmore_l_Page_154thm.jpg
15335f046a315e213dc1be331419200c
9ec6e0312464952510e210d6d6bd319e33f0a51d
6944 F20110115_AADBOW dunsmore_l_Page_175thm.jpg
611db6f0b3370ed419a84b7049d5028a
0c37968c37c059c631606ce397fc3ce00802d77a
81582 F20110115_AADAMI dunsmore_l_Page_150.jp2
5397a75b1973d70ba5f93ffc3f07a33c
90315744fef702319a6808a654146913800bb964
67365 F20110115_AADALU dunsmore_l_Page_136.jp2
0ce5d7e578eb6a689404d014c8be72b9
6e253068c823a62dc4989af0b0d4dfae792ff968
7312 F20110115_AADBPL dunsmore_l_Page_012thm.jpg
e94d268ee5263fd00af9ac5c69cbfd4f
f976e461172a68f928828342acef4946e2d56bab
19657 F20110115_AADBOX dunsmore_l_Page_088.QC.jpg
df07f41b12c69c3c9f6c70c8c4422ccd
6e74c221e4121f74e2b04a1d51cb9b33136c8d0d
1051928 F20110115_AADAMJ dunsmore_l_Page_151.jp2
2743c81cc75bd35695c4d13ef59636db
0689b0b4b297e3261c0a803e94364e17596e2efe
100609 F20110115_AADALV dunsmore_l_Page_137.jp2
9837f4031db4abca15b34b913c14a7ef
340815f713a144a13c2e20bda84a3b1f82a8c248
4929 F20110115_AADBQA dunsmore_l_Page_125thm.jpg
c20359203fc4c33601cad64682bb0ee4
a617a2a92e45ef058f23e8b38a1bcc4c27b506d8
5861 F20110115_AADBPM dunsmore_l_Page_162thm.jpg
e4eb3780228d87e0cfa2e1303fcb2ab0
0ca667675e36ef1b79e6cadc0f4fe2b443c758bd
23294 F20110115_AADBOY dunsmore_l_Page_036.QC.jpg
d7cfdc5e82dbd75cbe617d30527e0b45
56f900e8e90504f3344fcfee340fab16bb0aaf29
1051953 F20110115_AADAMK dunsmore_l_Page_152.jp2
feb00d219c07355bad989752391c2e2e
90663ecb9ca6ddce85739845b46547fd8014c5c0
111974 F20110115_AADALW dunsmore_l_Page_138.jp2
da65df80107899c50fddeadb001c0321
4d6f24ebe2f7b2cf0edd73a1e415ea4602a10e5b
6361 F20110115_AADBQB dunsmore_l_Page_102thm.jpg
0f6bc48b428ab05df3d4ef52af9e6ec2
68690b5d8fb44130eb60605b93ab1bfbcded5468
8428 F20110115_AADBPN dunsmore_l_Page_008.QC.jpg
03f4e6bd45b02a9cbd7b95642f6851e5
38d7d93695ab60c5016a61fc0c87d430dbfe6e05
16158 F20110115_AADBOZ dunsmore_l_Page_055.QC.jpg
c4a1788fc634be3d17cfbd59fda4ba58
c6177973567d5d8b63762cef27225b58ab61841b
1050971 F20110115_AADAML dunsmore_l_Page_153.jp2
ac798c57e483feaf9df8a97f6c476b3b
e2ab62d5c626fa19f22e777270364200e8233369
111737 F20110115_AADALX dunsmore_l_Page_139.jp2
a3cfec4f2ce894848819ac3a7f02a9b2
f35f2bc7b3f777ef2d13e092d561a15e64551cd5
12549 F20110115_AADBQC dunsmore_l_Page_048.QC.jpg
25cec831465ec7caa98a71244f1d7b8d
5b84cf6dd451c922c61aab8f783ebe054dcad4e7
14854 F20110115_AADBPO dunsmore_l_Page_017.QC.jpg
3201d6538fcc3a59452b9f97a13d0e1b
60c13edbc1a460134683d36a8108d570b3b85e87
F20110115_AADAMM dunsmore_l_Page_154.jp2
5bbb38097a35fb5c69e8710fea29f825
a686f423ad8e4a5adda01d7956a3d88bf57c408c
F20110115_AADANA dunsmore_l_Page_168.jp2
5dad8a1946196d1975d7b4572549066f
a4b861cc4e00928769b12007897f4a9358d49d91
3216 F20110115_AADBQD dunsmore_l_Page_130thm.jpg
dc8dfdd113ad88f1261910b774c2ffce
ad4ce007c5f6f5f7ba6452666c33e9e527b6cb2d
4754 F20110115_AADBPP dunsmore_l_Page_003.QC.jpg
159211f4c63a14dcaaaf2794e7f91ffb
b83674ecc0a69acffcfa7a99ffae0a6a1b7fa7fd
1051974 F20110115_AADAMN dunsmore_l_Page_155.jp2
80d5992e5b4eb51cc851e3ab24b8b53d
ea869741ec69f961605e0c869c923f1691d14d67
110258 F20110115_AADALY dunsmore_l_Page_140.jp2
c89b7386bd9c461768e9c28185213ae5
41fe380089e1691b1fc96975ce95cda831e0a0ed
137390 F20110115_AADANB dunsmore_l_Page_169.jp2
f7e5bfe800a5bcb4794330265c6d264f
3dd8a423300ce14e5de1ec9d9d5c9532cbe68ab5
27803 F20110115_AADBQE dunsmore_l_Page_171.QC.jpg
5cc29d9235affaa614d9335be98d8cff
14cb98c7ef89f5c59f4509b95c7b05bddae160cf
22728 F20110115_AADBPQ dunsmore_l_Page_096.QC.jpg
16681cfe36fbfef199c45b532624ea78
d5777adf8cb7f48a9cd39f69b5c720960e56374d
1051913 F20110115_AADAMO dunsmore_l_Page_156.jp2
6cd05e4f435f8a5b53c0ead474f4b5d3
bec961fc27c015073e137d8ccc5630b90037b0ad
113041 F20110115_AADALZ dunsmore_l_Page_141.jp2
9c49d3b29202958fe8553e95714323ca
36029dc5f34b24f8e56d5d385bc1f5f01b8dd71e
127877 F20110115_AADANC dunsmore_l_Page_170.jp2
11195dc6c9f34e9a6a9bf32ca578fdf5
2f5fbef7948417aca530112780e98eb74bed6251
F20110115_AADBQF dunsmore_l_Page_046.QC.jpg
42c67345972ba7e8526ce00daef8347d
03a58f36ae813ef0460b9d0b44dc085647177a66
6297 F20110115_AADBPR dunsmore_l_Page_011thm.jpg
035cdbfc79803af24cac8d792173477f
949fb4aba7ab3bfb253f2cd77aa6a07dd5625b5c
932575 F20110115_AADAMP dunsmore_l_Page_157.jp2
fbb5f04a33864036eb2a7dc6cb15c69e
7e54724aeb2c050d01eb2654cfd7d26421b33533
F20110115_AADAND dunsmore_l_Page_171.jp2
af57dd7e9092b3cf10984f04cca94eed
e1e2e8e43bf3b5500fb63ce0b199d63cf00c371b
15899 F20110115_AADBQG dunsmore_l_Page_123.QC.jpg
0d8c77f4d1e6992aec2a70e7f5492c99
867de58434ddac43aafb630b144b7858de6aa8aa
6138 F20110115_AADBPS dunsmore_l_Page_152thm.jpg
eff861e86e59b6f62965c6331f112dd0
0c05e74f16469b34c723b8dd5e0429e222246809
967548 F20110115_AADAMQ dunsmore_l_Page_158.jp2
b397fe6c6a1d728376ad069677cd0376
8a3d5dec1b995ad4310668b4a384ae141df030a4
145257 F20110115_AADANE dunsmore_l_Page_172.jp2
dfb460d30b56f1594b1de8641b34a839
6aefd26fe8edbccfeea55e1ec9cc71eb74081cca
6674 F20110115_AADBQH dunsmore_l_Page_138thm.jpg
0fd97ba050dd44bea85396d8d077e03f
c2107d3578d2fd29eb45a6c0e649d4330b4cd021
6461 F20110115_AADBPT dunsmore_l_Page_112thm.jpg
5657c28dc8ea2620d34d8c80b79d00ec
23faafae991f4faeb9673c90452c26b15d81e5d1
1027518 F20110115_AADAMR dunsmore_l_Page_159.jp2
1bf5818d1f0ee4c7dadb13c8092534c3
0f8abdb5bfb34b0a259c521b14583aed5a180451
123595 F20110115_AADANF dunsmore_l_Page_173.jp2
9f19e8f424d0a50ca43e520921e07448
6aa274d557386ac94c0d9dbb00f1cff08a6de680
6434 F20110115_AADBQI dunsmore_l_Page_041thm.jpg
010c5ef8659583e4c154278ec3e75b6e
92c2193d10990989c0db97523fac86a751b383a4
25531 F20110115_AADBPU dunsmore_l_Page_156.QC.jpg
611aaee1610be524f59d50eeee912660
8d0d6e59e5eb9513409a4581eb46ea3e2a13c4b0
877997 F20110115_AADAMS dunsmore_l_Page_160.jp2
83a7609853ef21ad80a91a93d76cd33c
9dbeeca910efc2ba0d975ab1bb848c934b922f78
117139 F20110115_AADANG dunsmore_l_Page_174.jp2
fdfd0d58a45bd99faed929334ce5daa5
78fc0dfbddddfe245f99fd39516ea7c07b699964
6638 F20110115_AADBQJ dunsmore_l_Page_080thm.jpg
9910d39cc0dbda275a43af4b78d4418b
f34d1df8cc290507cb907afeb6ece7cc9dc9f071
1824 F20110115_AADBPV dunsmore_l_Page_003thm.jpg
b61083cdeb845bb3a89629582fe6424b
9ef972cbf662a9ffcc55c945fcebbd2db7c6de02
900089 F20110115_AADAMT dunsmore_l_Page_161.jp2
001085a75bd9d8a68aaa893703d1585e
1024145389f4b84c6bdf0ffe1db44946171c1e85
1051977 F20110115_AADANH dunsmore_l_Page_175.jp2
90357a93f629b9a976a44a711f5e6e80
1b9241044a9691aa90252dd09a44fe265b383988
27491 F20110115_AADBQK dunsmore_l_Page_172.QC.jpg
e5f2d8209b6d34c61343f0b0ea857f25
8a1f894f8304d4aaf2c164c651c45c7ea31a982c
24793 F20110115_AADBPW dunsmore_l_Page_178.QC.jpg
59369fa51790bc69578ae8ae3c4e980b
b7d4bf540a9cfed749449300ce8764676e781cc7
F20110115_AADAMU dunsmore_l_Page_162.jp2
72096ee32367965899a7c8a5fb7c484f
6776accf8bc9a96a2798bc828cd9ff82a7a17ed3
117927 F20110115_AADANI dunsmore_l_Page_176.jp2
a93f000f69b5005bee1c4e2a9afef4e0
812f1e127c8a89d8d989f97ca0dcda6bc6a6e90f
6408 F20110115_AADBQL dunsmore_l_Page_076thm.jpg
c24ac2bd67bbbc2f8ff48fad9d5dc4bb
e9910c4b137ce61b9b8c8399395d1378c9b18f4b
6376 F20110115_AADBPX dunsmore_l_Page_174thm.jpg
e7520ac10ac831382ff39a2237608666
afa71830cf89e080b897fd500c306814986cb0ff
135289 F20110115_AADAMV dunsmore_l_Page_163.jp2
68bcbd4d463990727e8bb3e39cd01529
e9072edda3e3821bf1b45ac4a0bb9c79af48e18d
130121 F20110115_AADANJ dunsmore_l_Page_177.jp2
9b5e39f2a59aab0e2198da48b2addc95
5d3430f06b8c9e1d5b3d18628e6782a2332f2df4
22041 F20110115_AADBRA dunsmore_l_Page_120.QC.jpg
ec6216cc5c6bdfc0f037ccea35ea65ce
af5d959b8d73285546747dedbc16bbcfbe83cf88
23735 F20110115_AADBQM dunsmore_l_Page_142.QC.jpg
83809434550fe1880b91c4070fe56420
e453bf8dce8542a3efd67fb8849a2b3cc624fb59
24283 F20110115_AADBPY dunsmore_l_Page_149.QC.jpg
97c5c24f37a9d23571a0437cc12848d3
ddeaae3bc3d19589d845f67403cb02950607a40c
136808 F20110115_AADAMW dunsmore_l_Page_164.jp2
bca179cb86bfeaf6ec322522cdd3dcc8
69e5a18b317b92fb583cfd05ff2eb6d25d845837
127373 F20110115_AADANK dunsmore_l_Page_178.jp2
6485e41c6bac5cae0cc96b03940fd416
0d46f8eafd95f471d896bc36e474d9aa9afc3d72
4530 F20110115_AADBRB dunsmore_l_Page_009thm.jpg
37184028e81648d56260cefb1d4ff28b
f236207b117ad2716b4bca09daa762e15614c805
4525 F20110115_AADBQN dunsmore_l_Page_123thm.jpg
34baacd5565328e5468b954a0478ef8a
42f6c3b605b073e74b4ba4f54a9b30dd3f9492de
6538 F20110115_AADBPZ dunsmore_l_Page_036thm.jpg
d1043cc1b6d91e65ce821a49b9a69fbe
9132f6ec999c13a88855ab1e1753620a27240bca
130008 F20110115_AADAMX dunsmore_l_Page_165.jp2
ed809ee8ac5320ed28b5a900b68c405f
ebc2ec520178a41d646f702fc7a3ebf15cd2fd2e
51923 F20110115_AADANL dunsmore_l_Page_179.jp2
109afcea9e41103a721d9696318f5b44
7343ae62beca3f577d7f6373940f4fc8e713be71
22730 F20110115_AADBRC dunsmore_l_Page_068.QC.jpg
aef149123fe6d1d1b80ba6153f15ffff
1fd31617fd9efc9b44bb673d3c61e6ba10e5e154
6600 F20110115_AADBQO dunsmore_l_Page_032thm.jpg
6214b1ec8c518b5cf17cc514893cef6f
2ef08093ed54709f7890d5853c1b893d5e066c7d
125349 F20110115_AADAMY dunsmore_l_Page_166.jp2
4fac3273aae2f5c5cd4b63242ff95af5
68423a0bd9d60e15127c6d6035a7e498ac2fef1a
F20110115_AADAOA dunsmore_l_Page_013.tif
3967fc20b6f7e68d4aba076eaf892789
f361f1149df249660f8da2ed4647075acec93a6d
92263 F20110115_AADANM dunsmore_l_Page_180.jp2
deae489bb112a9617035252a70a2eead
32781edcc33d0dff85f9cdb34a2887591ea6848c
6412 F20110115_AADBRD dunsmore_l_Page_023thm.jpg
8beb18f88f8e787ce723a34e5e42f89b
523f0d8c3b859e9dee7cab7d5710afd0a32079df
6731 F20110115_AADBQP dunsmore_l_Page_103thm.jpg
c54b9c7933b7e8dde8a224bc8b52a757
91c3b78bcd3323ffaae0ca0e64392bfff3f4e09f
F20110115_AADAOB dunsmore_l_Page_014.tif
751790e617385616b9500ac8c641d2cb
6973010abedc90d2d49bb8dd6385de21d5b15756
40170 F20110115_AADANN dunsmore_l_Page_181.jp2
e70876df99f45004bd6cbec9aa19e527
73619332fd661807556b13d8aa9d26196bebad6a
24655 F20110115_AADBRE dunsmore_l_Page_099.QC.jpg
1514e6ecf59d6c3d8f8d61393d774ba9
f8127d2e669ca7f43de4d41bd93e91a5184a9f83
22776 F20110115_AADBQQ dunsmore_l_Page_033.QC.jpg
532f9b4faf3110c3f135d009b694a810
5ae8c85e34361ff2a12d5ddc337bf9a3fe000cad
132035 F20110115_AADAMZ dunsmore_l_Page_167.jp2
d9947a9694c2c2f3e19ce3be4ab2f7a0
1cdcccc094832ce901b5ea30adead22071a3524a
F20110115_AADAOC dunsmore_l_Page_015.tif
db017ed456c42bd1f68bce0aeef92975
9b8516d8948171991fad2ab653600bfdf4e8f749
F20110115_AADANO dunsmore_l_Page_001.tif
507fb97a297acb5a3133fad507ef0603
41018e4b9c682218adce9b37d1633a8c7b881b40
6218 F20110115_AADBRF dunsmore_l_Page_071thm.jpg
c8136898c153eb3d05000fb22d546f98
35f96a74646e64010e3fb22eb6dec8e63d2a0d59
F20110115_AADBQR dunsmore_l_Page_037thm.jpg
590934e7383de1aff0f5b6117bd08620
1ec7e5d85a4e2ba04dfc26d914a91f6a460981af
F20110115_AADAOD dunsmore_l_Page_016.tif
4e1fbb31a95ec5b8a6cf4f17f382ea8d
4f5997614937950bc9046817b2391815019e645a
F20110115_AADANP dunsmore_l_Page_002.tif
181a24431c64eaa2ba654c1eb75beade
c3f03b022fc2401befec755983929c625198a4fc
24232 F20110115_AADBRG dunsmore_l_Page_145.QC.jpg
e8f6af7a4393889c3c82f7caa3eb9d2d
02dfb2d6864c3107587ddab514e6fd9e7b4b026a
24115 F20110115_AADBQS dunsmore_l_Page_064.QC.jpg
698973cb9f9454385836d1ffe4a0b902
a47384e162c2b4a020b9a59b3fe5616e89417a33
F20110115_AADAOE dunsmore_l_Page_017.tif
b069afc4a95d316f8826e4071ce8021d
6d7575942a4c2c5b531abf88ac806d84ec4bf3d2
F20110115_AADANQ dunsmore_l_Page_003.tif
2783cc17efd18f49b5d7a74850446638
3212592abe90db2b0bdd66f0e0542f159eae6af9
24200 F20110115_AADBRH dunsmore_l_Page_146.QC.jpg
54d4c15e1fe7a61602eb89e7ebc323c2
b5ead243934c2d1523a2e388081725efd2eed317
2814 F20110115_AADBQT dunsmore_l_Page_001thm.jpg
392cea3f5ff49867aed37ded13b6047f
6d10c405ee979399f0407c97107a151a31c5136e
F20110115_AADAOF dunsmore_l_Page_018.tif
5a62bc7a534b5fa680a80d8c4136a7e2
66f9a428dadce6a893b23de7e65348eae4996685
F20110115_AADANR dunsmore_l_Page_004.tif
294c0a917abb78b0eccd3512db6e5f12
d79e1bae7123b09abe9e1d5fb89dbb71661d8f64
6475 F20110115_AADBRI dunsmore_l_Page_038thm.jpg
98b112ce09589de903d66b01b265d14e
ee2c8bea39ddd630a0964be83c49b13c15aa0dd5
6506 F20110115_AADBQU dunsmore_l_Page_101thm.jpg
0af529a9b2c505555f9e6007849e7ac7
2d495674271ce6f8ecb9c957c1c82b3327b1cb14
F20110115_AADAOG dunsmore_l_Page_019.tif
564a87b494765b8a2326330233f76070
70e51de5cfe6ef4bedc4f50ee91d1561e7497764
F20110115_AADANS dunsmore_l_Page_005.tif
32d4888597d5bfcbefe38c52f5c1b3dd
a271556822dec7033b64c8c8d26c288eaa4f5ed6
22649 F20110115_AADBRJ dunsmore_l_Page_062.QC.jpg
df19f139b9d430b2aee1ac12c50dd194
3f0ff12e45b76fe4f5dedb1ac0623fdf7bb13d29
6275 F20110115_AADBQV dunsmore_l_Page_077thm.jpg
6760f0ce4f751432ad6ab8b0ebe14da5
f4989581e55d69275af4b92421f036be1c4d2585
F20110115_AADAOH dunsmore_l_Page_020.tif
e02a73ce6406b78c32c46ea2cf9f86f5
75ae3de04935a6b1436fcb55a4fafe2aedd96e5e
F20110115_AADANT dunsmore_l_Page_006.tif
c20926bbaa0ef9dae1629eb8adc6d27b
31decc66c651d69433e21dada3c28be2a22f5495
23045 F20110115_AADBRK dunsmore_l_Page_022.QC.jpg
a88bfb2551948c88206dcebcbe59ce54
52cccde28b8dd0185e2576f30c9f814138631a38
13386 F20110115_AADBQW dunsmore_l_Page_136.QC.jpg
71d9de036c5cd3ab8001f97b94c4ed3f
e4cc766bf1c717a2232a5b74ea7b98974fd878ce
F20110115_AADAOI dunsmore_l_Page_021.tif
05985ef8af15e277becefa34b8065e2f
b44027ddea043a5641b213a76f3539856bf3151c
F20110115_AADANU dunsmore_l_Page_007.tif
a98986f8f5fc8a9f62947b519c57ecb4
f1e216439fad1ce5461b4c58f645b668ad78e764
23886 F20110115_AADBRL dunsmore_l_Page_041.QC.jpg
4087f105893cb579cc9834a3b01fad51
0c9f005afb05d7f4ac76ade605a77b8bb62f9a6d
5585 F20110115_AADBQX dunsmore_l_Page_157thm.jpg
96edc69ba9d2b912a7a3afb71813c15d
ee60f9d8033eb9c33b2fa6e422b673cfe9382cf1
F20110115_AADAOJ dunsmore_l_Page_022.tif
c65a677a55ccd8e1e2f4bafdd10e4464
a5302cdf3de39336874dd1882bf0cc45399a5fbd
F20110115_AADANV dunsmore_l_Page_008.tif
e062b6b0dbdc6bbdd5412ef45e27d0a0
2ed4b9625a8dc096e0ce5e887b381d5641c7049c
16843 F20110115_AADBSA dunsmore_l_Page_132.QC.jpg
5b2de5ac048c4bd2231348dee9c0b979
eab7aad041c9da0fab270845f931da1b7d297294
1435 F20110115_AADBRM dunsmore_l_Page_002thm.jpg
33372c7e6ffef4e02f506dce0787be2f
441e759ceee3a2fd8bbc9230a0f05d75f0041357
3365 F20110115_AADBQY dunsmore_l_Page_083thm.jpg
151f689935054e26a601f80b4297e1c2
a5f7bec89beacee4873c27a7a75c38a511fc3f7d
F20110115_AADAOK dunsmore_l_Page_023.tif
fde4af9b784866d5dd19f92b9364bd74
d67a46d5336758e72ea74e050101d6a5cf557100
F20110115_AADANW dunsmore_l_Page_009.tif
1df0298bbf27f893e1261e298d3e05c5
1099b1467dfd86d85beb2b18fe91ab004db89d82
21987 F20110115_AADBSB dunsmore_l_Page_110.QC.jpg
8acffc5ee7035c18453f2c9cb51055af
8d8a68855a1c9df7a5d6bbe442b5c8e28152ac02
11303 F20110115_AADBRN dunsmore_l_Page_131.QC.jpg
a51f87c1687e7fdb5b3f569ecc675954
1ab5cadc0b24dbe096f87198a25b03df10960469
6136 F20110115_AADBQZ dunsmore_l_Page_035thm.jpg
454bf7ecd90f44acb4f56b083cf04b24
cd0fa52607c3d89cdf7aa4b27d6f2cf6410053cf
F20110115_AADAOL dunsmore_l_Page_024.tif
15ef639aab2f827a86415a931b499d28
98ce726ba929fa89a51c85b530bf27eca3bee9c9
F20110115_AADANX dunsmore_l_Page_010.tif
3004da39c33c7203d974cf751e134fc6
8180f1b25ab2799bcad860fa8ec0952d6ed00fdb
5940 F20110115_AADBSC dunsmore_l_Page_070thm.jpg
fe94d813c5c2c8a5a135fa27068786f9
6573870dd73beaede7bc6941f4947b09191fa237
F20110115_AADBRO dunsmore_l_Page_034.QC.jpg
cbba51b7f15fc17a299991f2c1589641
4d9239e5f18e88cb0def615c18ad098cc0107f0e
F20110115_AADAPA dunsmore_l_Page_039.tif
178e1a3358ff12f2dfe619aa4ff2e4e1
dce8241caf0b505de59b7bc0049652eed8d062ce
F20110115_AADAOM dunsmore_l_Page_025.tif
d3fc20a813de8a39ff606d3308d710fa
59502d6cfc96ac088a400868be546c91b88c1ef1
F20110115_AADANY dunsmore_l_Page_011.tif
24eb684d4c08e3c0d3bf90753dc69fc5
62ecbe8dc8a228d9cfced54332eedd4dbdc66a5d
24663 F20110115_AADBSD dunsmore_l_Page_103.QC.jpg
983320fb09492a6ad1f881b4a7dc80a5
46a7eab0c1514f3e0da0c3fd9d88939565572c0d
6651 F20110115_AADBRP dunsmore_l_Page_019thm.jpg
d931c783eaec5baef3671588bc8de185
772803c18339cb701c18b19917c2914119b313c4
F20110115_AADAPB dunsmore_l_Page_040.tif
a342c9b342fceb43086995426161b96a
c0768744b7bd3ce182024ba3ca3807d7aff394be
F20110115_AADAON dunsmore_l_Page_026.tif
2ac6f912698c69bf06b1d21a4ebfe677
c2b5d743f6ba3a530bc3158f91dbce79aa8d8ef2
F20110115_AADANZ dunsmore_l_Page_012.tif
9be894dd05fdf64946eaf0b62da5dc93
1e9e8d23bbbf58592342cb7f5c8d7558c8d6ee78
23024 F20110115_AADBSE dunsmore_l_Page_006.QC.jpg
a8f317ec41aace29b5885406520a14c0
78d7ceb73d5d7fd1aca0bfa78f8be958ba3a131f
5394 F20110115_AADBRQ dunsmore_l_Page_015thm.jpg
57b0f8f9177c5a3e5fbfd5845d219cc6
6bfefee3aa6822d90866821fecceaf1a4e4d7804
F20110115_AADAOO dunsmore_l_Page_027.tif
2a4a152557595f5fd616f47ff9608549
f1d1eb44062f9548e7f9f69b86fbc35750619437
F20110115_AADAPC dunsmore_l_Page_041.tif
55b7407759e98e428f985840d51e5cfc
ecd75c1bee4c4dac02de3d78e3f4e26d915a0fb6
4110 F20110115_AADBSF dunsmore_l_Page_051thm.jpg
27082ad589ad0f5f2a05c709c82d5fe8
05062e62ddef813cd1e4bdd001a583a5d1216376
26661 F20110115_AADBRR dunsmore_l_Page_169.QC.jpg
e6bdb069689a3847a38e74b812ba662c
5f090dd5cef1ced0c63fac7cc70a582b486b1c6d
F20110115_AADAOP dunsmore_l_Page_028.tif
258f2771b9d0e0e0da2527b227a414f3
da56ce3d8b5ad6b430898e2e2894292195d04fab
F20110115_AADAPD dunsmore_l_Page_042.tif
f519bb7f6215f59e41f2dda97f4e475e
81c7c6beebf038444e9072d6f6c432a43419d8ef
6129 F20110115_AADBSG dunsmore_l_Page_093thm.jpg
e3250e5689bb3ee027248e9dd28885ce
a3f765941d7236f26ecdb96e4a023cdd594bc148
15297 F20110115_AADBRS dunsmore_l_Page_090.QC.jpg
b34d3b4b2c41edf7d18bc7879248c4fb
c913d0410cbdd3255154a336d44c578e5ce71ff9
F20110115_AADAOQ dunsmore_l_Page_029.tif
ae0001f58d4c21690232c870b0923140
efa19b4f330bcdf47f43429c8bbfeab1c23ca185
F20110115_AADAPE dunsmore_l_Page_043.tif
fcbe4e9c951ba4ca032fc3681f5fe0b9
c9e0a0d0eae4f799f8e4a2a6ff5331e0534eae23
6025 F20110115_AADBSH dunsmore_l_Page_158thm.jpg
4bcbad4a922ac1d7118562e1059ecfe1
09cc2178aa1eaf299dcd4054c362dc3c0926ad2f
20818 F20110115_AADBRT dunsmore_l_Page_157.QC.jpg
9bbf27646bad7cbe8999e8df6afd08a8
e2be42c95820befd356fc336baa2aee0a8b6e8a3
F20110115_AADAOR dunsmore_l_Page_030.tif
c8441c80776024e5a67a91f53e36b7a0
ddd32f53de315b1867c92406a1cd202f344bdfa6
F20110115_AADAPF dunsmore_l_Page_044.tif
ef77915e2cc13a8c31ef61319a816887
e10196ebcc763346c5dcb66c9e6022288011cf04
3325 F20110115_AADBSI dunsmore_l_Page_060thm.jpg
48bdfba0ac16ded2e69636c3eef9e09c
b797bfe3cf51e66307f834e934b7cbc879efe8a4
5986 F20110115_AADBRU dunsmore_l_Page_028thm.jpg
80da78fe4c9fc1945afe1fd87482b8d3
a69a4dd74faa60f696f5e01e54502e19b8fd6872
F20110115_AADAOS dunsmore_l_Page_031.tif
b957017bc645a494836b01c6ca18292f
2d143589214b299e23c959e1ebe499ecd935a28f
F20110115_AADAPG dunsmore_l_Page_045.tif
4fef5b9730cd2834fb43de6c2772ad8d
48e3c2555c9594e8b4066a77baeb39056f1cf77e
3360 F20110115_AADBSJ dunsmore_l_Page_052thm.jpg
44b6ef561139d655bc7d65df82b4fb66
28f755d560ebd7c2364b9af5146772786032cd2f
6391 F20110115_AADBRV dunsmore_l_Page_156thm.jpg
809818c85d72d95e760a3b709ea37ad8
09a63aa02edf65864458fd3a7bf11bb4a2119dc1
F20110115_AADAOT dunsmore_l_Page_032.tif
9b372dd1a4ed8db83230b59da236453b
6de0e2b7523d657ce47b97f859db390d703cc278
F20110115_AADAPH dunsmore_l_Page_046.tif
8148695dd0724776a9470dfb3d515ee0
beaf7dd0672b9bcc917cc26d48bbca4f5db5c078
3634 F20110115_AADBSK dunsmore_l_Page_002.QC.jpg
5e246cbe44cc896c240b7262941e9aec
fe4beeafbdcde5b04336e52d4ebcf100166249d7
4326 F20110115_AADBRW dunsmore_l_Page_090thm.jpg
0a09ed5961350cc2f73102b66f3b39e4
5896a2bfae33de48ef6517f291e8c83ea21704b0
F20110115_AADAOU dunsmore_l_Page_033.tif
c23377452fccb5e38cff0929cf85e643
8d6a87c95969a644924d7994b0efaa9804c0b428
F20110115_AADAPI dunsmore_l_Page_047.tif
04dabea2c5adf644c669846e243068b7
f36c2616d8543a06f72aca6b73e31fd9bfbe423d
24127 F20110115_AADBSL dunsmore_l_Page_148.QC.jpg
ae8563c4420c9ee5faa2f2bdb1395bd4
4e9dfb8759d8aa8b40cfbaa265169a616e0eee09
6512 F20110115_AADBRX dunsmore_l_Page_026thm.jpg
521d8f52f87cab1d8b9da431edf0cf73
84b5268afc329803839954a15b2d0c37f3b12412
F20110115_AADAOV dunsmore_l_Page_034.tif
bf240cb441b8bb8def37498c00129469
170301451a99b5d2255a9f0ba37329609fb75bbd
F20110115_AADAPJ dunsmore_l_Page_048.tif
407f6b5bf2451946339fadb0810d0f3d
805902a56ae597ebeb3fa716ddedd3f5a6ffad79
23929 F20110115_AADBTA dunsmore_l_Page_019.QC.jpg
f230bbf214769b85fae4670690502f4c
8bc589df8e1b80f651d33498cb441bbfe158168e
6612 F20110115_AADBSM dunsmore_l_Page_021thm.jpg
ced88526fba451f36ad4e61001f45465
67e4287a2bbcf7444dbab82b788a15e70d018834
6462 F20110115_AADBRY dunsmore_l_Page_142thm.jpg
ebd12b8810ed53eb6b715f9b12d2b0f6
08a1b186a2f53aeff911841326523332bcd1c80c
F20110115_AADAOW dunsmore_l_Page_035.tif
c86caf612616fca4dfb405f0949b8c9d
b091d641e2c42d0d3c15cf0893723157c0c36313
F20110115_AADAPK dunsmore_l_Page_049.tif
4996b7473e3dd38868391feb61078dce
9c4a2904fcae2dcd4dd193cd2e87442b875501dd
23638 F20110115_AADBTB dunsmore_l_Page_023.QC.jpg
893ae2cc90740a11c5033cbfd6a57977
ef0c3120b26f1e3b7fd4f41b51ecc297db8e5c2c
6347 F20110115_AADBSN dunsmore_l_Page_176thm.jpg
d025e260b275c78ca1d4fc087b6fba45
e4a5d1d5a5ac515a7e3907ddec56da7dd47f0439
10067 F20110115_AADBRZ dunsmore_l_Page_084.QC.jpg
a6d743afb2580725a27e1f500c6e2183
024feda0e65994155bf090c0d6d604d99a1e36e2
F20110115_AADAOX dunsmore_l_Page_036.tif
5ca38ee2c5ce1750a4f1d819b7cdb4b6
8d7932d9d2908f82bdae55d534633acb130ce812
F20110115_AADAPL dunsmore_l_Page_050.tif
1e7bc965838e63066080c28a2d8d492b
a5e6eb57b69d379d309e5e1c354618023f60d63c
9909 F20110115_AADBTC dunsmore_l_Page_135.QC.jpg
f0bbd7783fc1652ae59a121be22e76fd
f75f033837f79091b27dddee77b28887905bf69a
5511 F20110115_AADBSO dunsmore_l_Page_161thm.jpg
c5cacb07ec13b9be8c419264e70af3c7
d33522289d73992b5ce052754ebc6d91327e9a0b
F20110115_AADAOY dunsmore_l_Page_037.tif
7c38ff822e0767fd17c5425d7bf37383
4f3fce72a5acdb44b69b7b85d90eeb851b84de1f
F20110115_AADAQA dunsmore_l_Page_066.tif
870227added4c3da069f1520f754c115
bb6f7954b46b576b6d3227a08ead39ef4f36d83d
25265604 F20110115_AADAPM dunsmore_l_Page_051.tif
bb0cb5f32aa9c01ec196c67fd9ddd87d
3a6fb00e857e61f7fd992af9f93ba26a07123450
22142 F20110115_AADBTD dunsmore_l_Page_158.QC.jpg
2e0ed23446bc37b07f7f62b1404f2402
a94e74aa5241b0f52116147355e0e321e029b098
21986 F20110115_AADBSP dunsmore_l_Page_037.QC.jpg
e63aa8df522dafe15c3b3be64012c435
4663b16263930e211fe85232a059500b0c82a74a
F20110115_AADAOZ dunsmore_l_Page_038.tif
1af43693b4df27690dacf64e6efe6a00
4c45eb8dba35e10b6b2d70763f9c67e082a6d566
F20110115_AADAQB dunsmore_l_Page_067.tif
d77a9fcbd43183502e31faa2a013de1f
eb79c34eaacd2193e40cb72f05c04fe5e4da2667
F20110115_AADAPN dunsmore_l_Page_052.tif
c17d0e66effb057c0de6d22e5d34de22
a196bfdc9f99d18c1d5f9446b4fd6047ec193479
23501 F20110115_AADBTE dunsmore_l_Page_073.QC.jpg
7618502a7392f36d6185e4f8cbe16f95
fb14f128b532bc4c88bd475cfbf9c29046c1fcee
6528 F20110115_AADBSQ dunsmore_l_Page_119thm.jpg
7ab3cb87dcf9cfd6c7fa32dd57149ddb
d4d971707709383f4a9475e22e4a15875a974c38
F20110115_AADAQC dunsmore_l_Page_068.tif
8677332d17b96fa9374dacf4b4a80273
437c7dadbde83ed7d388a324e29d796a6e761e4d
F20110115_AADAPO dunsmore_l_Page_053.tif
837f8dbdf7ba92ce997cb91d4fb47f62
277e6251a065c1182ea86f89ea20d2b442177b94
3549 F20110115_AADBTF dunsmore_l_Page_124thm.jpg
482028a6700b20abda32fc552ab024d9
b95cb93826ec15e757a892e9bf5e892a3c06add5
6356 F20110115_AADBSR dunsmore_l_Page_068thm.jpg
55c218edea80fa1b06cdc07ded84f935
d7953d14ee627869a3b22b29cdebf7dd1e8c8337
F20110115_AADAQD dunsmore_l_Page_069.tif
83e19d2b2067b1e0615ee45c09ddff82
950256709fb4655531aa3b76c66934280ddc5cd8
F20110115_AADAPP dunsmore_l_Page_054.tif
100595b546dba0a83ac682447aaf8cc4
f4236dd1eadbd5a67b8fc9d6812fee3576108d23
23689 F20110115_AADBTG dunsmore_l_Page_044.QC.jpg
7d8814acd16b18336173f93cc08486fc
4b7eb3428e077f67186a7ad10412b39494f46aae
22631 F20110115_AADBSS dunsmore_l_Page_092.QC.jpg
bea819e87e0b989c9ba476850e060a0e
8a5b7ef4d2cced7329ec19a67c6e9cab662fea25
F20110115_AADAQE dunsmore_l_Page_070.tif
116cf5bbdab1e76cd6974dbc04800ab6
d272524a89d4248d52d72405fffc120f75e50e39
F20110115_AADAPQ dunsmore_l_Page_055.tif
7cc70b529848dbaf2771a82ba363cf1b
977f1ab2035158695c469a53130fd97f90e2d3c0
22235 F20110115_AADBTH dunsmore_l_Page_043.QC.jpg
9dfe80ce699b4b0bdd71d97b46c05a62
8fe2d3151fe659fbceb06392a17fe1964b265f69
6582 F20110115_AADBST dunsmore_l_Page_114thm.jpg
a6ddf4c870c300e6dfcc4c36d7f52802
c4d44f3a80334ddaf2529042b95de9eb529b1b6a
F20110115_AADAQF dunsmore_l_Page_071.tif
3a4ba7659246b3619c1f2833f115c818
7a33e43b5139ae72b51cd81482b4770b80e86392
8423998 F20110115_AADAPR dunsmore_l_Page_057.tif
1b794ace935fbf2b1a114ff6d85e457a
3e9f54cf222a510f14d88ec441a12193635f6017
6501 F20110115_AADBTI dunsmore_l_Page_115thm.jpg
33a66658fb983f86e83f30a3d63b7b55
c58eeed08c0403020932c22340e487ea1d958208
23815 F20110115_AADBSU dunsmore_l_Page_030.QC.jpg
d0949117817796f3dac2293bc2ae62d1
12ce8ee3f38ea26c644a3cb1a1ca4cb21620614e
F20110115_AADAQG dunsmore_l_Page_072.tif
85820cd779fd3910c6b4ec8da6894c2c
a25f49bca69c8880ae61d9c8bc79e72176b779ed
8425398 F20110115_AADAPS dunsmore_l_Page_058.tif
a65c3d0c115363d31ecbcdf0caa8cab2
11bd38bbb2aae51568af9930e2a3e016da9b3ab3
5922 F20110115_AADBTJ dunsmore_l_Page_072thm.jpg
9331dca47b6867d6bbeb17b1120b5b56
f1af10e5c703ae6036fd84495f2bac2f2d075cb6
9498 F20110115_AADBSV dunsmore_l_Page_060.QC.jpg
135116fb84928a471b5e5425820ac411
5772e01d56b054622fb1b5631bf1bff8eb3886d5
F20110115_AADAQH dunsmore_l_Page_073.tif
2f4f549e324ea821e1de9669021ec5c1
94c9b15e3520ab88887423115901e54ab4d99128
F20110115_AADAPT dunsmore_l_Page_059.tif
c1d2da27dbc959e0d2f31d591dc57f9c
3c7bf321af210a476e731e4a593b5db859b19d0e
6425 F20110115_AADBTK dunsmore_l_Page_045thm.jpg
d16f59d5155f96fac73233b2b5b54f4e
dfbf95c99b281cedc48875fbade84032d454174f
5740 F20110115_AADBSW dunsmore_l_Page_066thm.jpg
124fb6b3327c2447b3329f2e665d925d
6857449dbdcacc04a2ef9554be7ecf4e0707ad11
F20110115_AADAQI dunsmore_l_Page_074.tif
59289424ea77b83e91efb4711d40c3ef
ecc3533f1e0b8ef67482bc397bc9a8ea9bcca0c1
F20110115_AADAPU dunsmore_l_Page_060.tif
1e8fb883ad8bb3e9842f47e9c1725f9e
fec8758afb095874bfbca4e3d1e65339681d7c84
11266 F20110115_AADBTL dunsmore_l_Page_085.QC.jpg
ca5798f9057ca32f30c69d4fa021bd3a
512666627f8d4880d687f16873a029862a880ea9
15695 F20110115_AADBSX dunsmore_l_Page_061.QC.jpg
ad25bd629156793853221ec9e2c12e03
35a3cebddc1d1bee55a50e68471f312c95f14242
F20110115_AADAQJ dunsmore_l_Page_075.tif
152f10bb8b3a1456fcf9fc97c4021338
c9797419ae8070616b5f44c89ab77e39991e6adb
F20110115_AADAPV dunsmore_l_Page_061.tif
1be93b270f7b5d0038798b1d56ad8daa
8c2906124968f7629e48660abd11520c7cfc8f72
20811 F20110115_AADBUA dunsmore_l_Page_015.QC.jpg
6c923c7e5f7ac431665c2e4cb4d63660
4b09750edca9e8b22a5cf60f30f3b49378ad1e0f
5807 F20110115_AADBTM dunsmore_l_Page_004thm.jpg
2f99456dde6410c043aa15767a1fda25
8d6e7435f0d6dc3c99c3617e190f451dbbb75418
22998 F20110115_AADBSY dunsmore_l_Page_007.QC.jpg
5842d61f651673999b44ca9eabc88cab
52a5c28d64c2329156ab63d2de5d38de39d9181e
F20110115_AADAQK dunsmore_l_Page_076.tif
dc933545b82b97f87234663dc8dcd409
11742946f5e571dadd9a9b9818c00be13310d81d
F20110115_AADAPW dunsmore_l_Page_062.tif
b255b4aa504dbf6bd7eeb0ca395bc081
e3fea8cf8dac9272f2f82574e78e86491bcfe710
5445 F20110115_AADBUB dunsmore_l_Page_016thm.jpg
20be6278745ace687ae953d2ba7ce28b
49b0b1546928e666ca896465bed4394de0966494
F20110115_AADBTN dunsmore_l_Page_024thm.jpg
21a9bb724900f30c9aa3da9a0619dec2
e5eb149e5e518b6f5500823a926d34dce18c6706
22863 F20110115_AADBSZ dunsmore_l_Page_025.QC.jpg
a4da2dd92028a72f1554fc8840e3e583
b2b220aefd8c0e7dc8d6885bca94c573a27b28b1
F20110115_AADAQL dunsmore_l_Page_077.tif
1b0964f69487b5033b5ce96459df4174
973baa14ae6ff9dc52593dcab262b3a20fe7d31d
F20110115_AADAPX dunsmore_l_Page_063.tif
e81c8cf51b31ac843f36d36b4b5a4c7e
7629775e544cdf41138915bdef2c6da3964374c8
4277 F20110115_AADBUC dunsmore_l_Page_017thm.jpg
7f71ff9232ef1fc00984f0c7fc46729a
d0e82dcacc0bab9beab6381770a52478d302dee9
23218 F20110115_AADBTO dunsmore_l_Page_076.QC.jpg
1b24ae7d9d24527034ab67349362c2ac
59e74b18378dd39df4fe9861731c5f25215364d9
F20110115_AADARA dunsmore_l_Page_092.tif
8100a797249e9f6164ff6c177de2d03c
8ac895a5256cfdf653ce1d4107db5e045d07e743
F20110115_AADAQM dunsmore_l_Page_078.tif
5e521100291b890c1c4e53fc2dd07477
30a402e893ef900f46717d208abd0ada71174b37
F20110115_AADAPY dunsmore_l_Page_064.tif
bb082a943c472de5773e8724c226f2b3
4becef7d41e3b8657c412ebdece01cfb75230c1f
21270 F20110115_AADBUD dunsmore_l_Page_018.QC.jpg
62988143bc64b9abda3fc58c022b48cc
0c2ce05d9b6ba9c2e58b62b6318e33690a8b1fd2
6242 F20110115_AADBTP dunsmore_l_Page_110thm.jpg
1475024652141e86ac12950b049c25f9
7a4cf40246682829a2ab83abbf16cd6eab4804bc
F20110115_AADARB dunsmore_l_Page_093.tif
924420c7930e76e675b3ab9e9b2e9bd1
297c48888b13ee6ac357a08cb9b639479ec407da
F20110115_AADAQN dunsmore_l_Page_079.tif
b83d473bf5fd36c6015d4d5be47a08fa
93aa3adaaf2117ce9b03c6122750a114d201ca91
F20110115_AADAPZ dunsmore_l_Page_065.tif
0cabee027fb4149ed78c69b6beea2680
fe990e8b661e3305d600adb6662362f4250c567a
6448 F20110115_AADBUE dunsmore_l_Page_020thm.jpg
470a80bd7c14507d39dc050ceeea017e
b02b76358d38137153346dcaf5b8cd0f1171b75e
6984 F20110115_AADBTQ dunsmore_l_Page_169thm.jpg
7587c4e1e92a9379eb14c24340fb4c2a
40e1d77965f230fc4ec1b7a3c972dbe11e9b9b55
F20110115_AADARC dunsmore_l_Page_094.tif
17130c859f7d518f570fd89b07212ff2
034c1a5ebade89a8195de7062d82bd3a6d0fa732
F20110115_AADAQO dunsmore_l_Page_080.tif
8e265d3d0746cae2f852d0b2ace16289
5962e4956943dade6d2704f467031d891d89f661
6231 F20110115_AADBUF dunsmore_l_Page_025thm.jpg
fe1ba5afae45f1c0796e5e794ad5f827
87d530c8947d57e99fb51379c27c5aa573a337e8
F20110115_AADBTR dunsmore_l_Page_005thm.jpg
a563a247b2f04a3662f1e768516b959c
85d2b44491160ee4e88144c48b40cfc1da448b99
F20110115_AADARD dunsmore_l_Page_095.tif
2caa597750b2cb10b2a032c801f06dd8
2a6999eab9b1e61cbe7e988fa6cd5b872daa6a9d
F20110115_AADAQP dunsmore_l_Page_081.tif
b205a6b150be92c161c3eae338cfdfa5
a3c55ee57bc430149b3ff61948159a69eb368908
F20110115_AADBUG dunsmore_l_Page_029thm.jpg
d159d318babe8938765ca3282521c6b8
81d09468854d8f36d981d98b9b753f81ca115af1
6596 F20110115_AADBTS dunsmore_l_Page_065thm.jpg
ad7750d56221a3ad28e91201a607bf10
f5cb44e508ea14bbdcd0f558fc46bd74a99b286e
F20110115_AADARE dunsmore_l_Page_096.tif
bed46af0b2fa10b9dc6405488816ebe2
2b71dcd57b81147631bde7a6a67a438d2d75380c
F20110115_AADAQQ dunsmore_l_Page_082.tif
fa2b7ade2133cdcd2001045b6ddb51ea
72637d4444fae2a5e3911f7590608dffd2de4fe2
6695 F20110115_AADBUH dunsmore_l_Page_034thm.jpg
5d1e00c7ef81c5182cc0c3eee1e9bac4
52ef2937e9bab3d27f946821654f474257c9c880
24313 F20110115_AADBTT dunsmore_l_Page_143.QC.jpg
d4bc465971fa513b5302eac3b2e10f1d
acaf95136ee386eb5cc4790c03e2137f66f4ac30
F20110115_AADARF dunsmore_l_Page_097.tif
4e803055502c3f8888939cd38ebf33f7
b6908b730874eaa94ef27f6feaae50daba241cfa
F20110115_AADAQR dunsmore_l_Page_083.tif
5363a2a885ed5e5ff8947d117b03b7f8
1b246add4fabb7db397ed9aa99c12d5723fae356
6485 F20110115_AADBUI dunsmore_l_Page_040thm.jpg
be332faf866fde1d495482887af726eb
a587eb66ab47416520461060b6cd8f87fcea48e1
272450 F20110115_AADBTU UFE0008350_00001.xml
bb1bf0886a0115703ea0a0092cf2a4c4
cf240e06cf0c01359183ede65d87516ef55d6531
F20110115_AADARG dunsmore_l_Page_098.tif
a2177a96acd0fbfc1af619f8127a334a
21cc5bcfaee703a34f631a883ee7815567fdfd91
F20110115_AADAQS dunsmore_l_Page_084.tif
2ee972284d4861d9da941710ad38d199
9bb90f4b510d4bad6360676a06a1306f21106207
6488 F20110115_AADBUJ dunsmore_l_Page_044thm.jpg
26ae4cabf29b93a20f3441911b63fe46
ebcbead45cb9ee5bf08f93cc7f9b0012294cf43e
22320 F20110115_AADBTV dunsmore_l_Page_005.QC.jpg
22437e1305648f4f40a624c3eb2cf205
ded8b429b3529b2fb9f6f87c612c42052742394e
F20110115_AADARH dunsmore_l_Page_099.tif
fd171d991086edcf56b79516d513e5e2
51ef81d0c664671ac95ba031cbf1e3c75c1f7d38
F20110115_AADAQT dunsmore_l_Page_085.tif
257e831e312697de40d5269bcc160170
795eb584504d585ce5f1f53746654bf5f52f0fba
18307 F20110115_AADBUK dunsmore_l_Page_049.QC.jpg
98dd5b63f02b67d5178b6512609588b3
8fbd12493aa4326e2ddc7a0d4ef78f1d0db91701
6416 F20110115_AADBTW dunsmore_l_Page_006thm.jpg
3d2c587a6586b88855c6a35b2fcebf0b
918041170022390fabc8a5cfcc50578cbc310017
F20110115_AADARI dunsmore_l_Page_100.tif
9e9844c359a2d647f105a39c44e413e2
fb1365016ce0a067d10136b766d19611f6bf2753
F20110115_AADAQU dunsmore_l_Page_086.tif
5af4fcb838044bf6633fc9c177974661
8482a61bce94725ed3c1f99dfb7a840a717ed8bb
5329 F20110115_AADBVA dunsmore_l_Page_089thm.jpg
ecf38b1f681cd770f79bdcf959a51cd5
5f13abba1f4b335cee09beec3f2a45e99ebfe666
16051 F20110115_AADBUL dunsmore_l_Page_051.QC.jpg
3416bea2424ef064a5cae179f16bab32
5be431d49c73dcc3600f4c920d90b4be50258a90
14384 F20110115_AADBTX dunsmore_l_Page_010.QC.jpg
2167996b35365ffd5fba4cc2b6fe2ef6
1516ff133d53ade41238a8f2dc708b54a25d87fc
F20110115_AADARJ dunsmore_l_Page_101.tif
3aeff0cd207b136d158f4022722d2684
edcba73fc142cee27903a61dc7b6f3c976dd1915
F20110115_AADAQV dunsmore_l_Page_087.tif
ac467fc9cdd56e30b8316fca0079e339
138c0bf9d97050c2cbeb9d1c6c49c7432230f5d5
3482 F20110115_AADBUM dunsmore_l_Page_053thm.jpg
967992065fae3cb897ea9e6b7e40953c
11940a97aa2f4f9fe30e948ce4bc27e40d9b596d
3968 F20110115_AADBTY dunsmore_l_Page_010thm.jpg
f06efb801369ede20b78863f80e6c4bf
111fb912578b948d44c17b313adf32937963fb73
F20110115_AADARK dunsmore_l_Page_102.tif
97321102609d2abd8a9ebfbc7e3b7a55
eb526842f6e704ca01fdaa4665fc675beac88e69
F20110115_AADAQW dunsmore_l_Page_088.tif
8455fd821a12acfdefeeee2ef520d0e4
44439bcb4c5ec522987a17c15abac67c7cae5fc2
6276 F20110115_AADBVB dunsmore_l_Page_094thm.jpg
ea89d7514498753a8acc00dae32b8c07
1ef38fe9928fc631700485042cd882abf8ada77c
5187 F20110115_AADBUN dunsmore_l_Page_056thm.jpg
95cd4a9a548159c86eb5afb2f96da66e
7dde78342e53954dfd416cb2a03e3a6facf2b5a8
7877 F20110115_AADBTZ dunsmore_l_Page_013.QC.jpg
4c8ab4246c2a8919b5df5e4be5928135
acbe450cee4540bbde45df00c69f025fc19193f0
F20110115_AADARL dunsmore_l_Page_103.tif
84a5e68eb18aa680d1548bf9fc530ea5
b473f7d2620d1b85b6a88b48f102702e34baa6e1
F20110115_AADAQX dunsmore_l_Page_089.tif
af768efc7e8e04eac33b2a8034b11be0
3fc6038f4b40fe12fd951118833094c80f3c60a3
6277 F20110115_AADBVC dunsmore_l_Page_096thm.jpg
f7c5f7d002869ba3f7f15505e5f28465
d9c4e025481e3f5820ac7bccb4b56f34f0fa6bc3
8487 F20110115_AADBUO dunsmore_l_Page_059.QC.jpg
ef6f93f862c6bb9af53064dad0d945e9
0bc9646bdbaaa30a4d0678adf2fc1215bcc76ee3
F20110115_AADARM dunsmore_l_Page_104.tif
474acfc5ff535717327149d299852e83
a6f5a7801092704df0cc786afdfff641110a68a9
F20110115_AADAQY dunsmore_l_Page_090.tif
61fd81a2e2f2f9fcd86547667c19b0a1
a7dddbf201363827014ffbb4189bedc487ccbec7
F20110115_AADASA dunsmore_l_Page_118.tif
c38fd933621b29021553ea21ece46830
604b4574e0141a71c0a4409693c97af9537f3f49
6724 F20110115_AADBVD dunsmore_l_Page_099thm.jpg
e94b6484916354a9a6577113093b1447
f4205cda33e9d1253f161fea2d8db1c0f43bed88
4848 F20110115_AADBUP dunsmore_l_Page_061thm.jpg
f54c1c49e84af06dd25eba9c8bd45546
92bffd9df614e78aeb75fe84d026ec45df7573a4
F20110115_AADARN dunsmore_l_Page_105.tif
584cdfe51e335df846939ebbb64bddbc
cf019cca232f50bd239ee2dd24d190ebaab79147
F20110115_AADAQZ dunsmore_l_Page_091.tif
5893607eafa75eff151d220e60004295
3dad678fc4c1169f9cd3ca7ce5b0a8afeb4dbc6d
F20110115_AADASB dunsmore_l_Page_119.tif
df574ccd57b8b17a4ecdd2e552284672
20a10eef9aba7867d09e1b9a6fc1b7576e9fd91d
23558 F20110115_AADBVE dunsmore_l_Page_109.QC.jpg
01bd6647ad2f2c3c7fb517de253b5999
27c1a5a597f565b4b1df6c3246f978071f6561c6
6660 F20110115_AADBUQ dunsmore_l_Page_064thm.jpg
c33cc5406518800b4becafbb3901a28b
02a6e2266dc9865595c617a829b5d1820bd2c2c5
F20110115_AADARO dunsmore_l_Page_106.tif
fa4134bcda45eb6dd6d74c84b0200daf
765fa449025be2b4feae881aff18ac65ca463ac1
F20110115_AADASC dunsmore_l_Page_120.tif
fff6fd3c943c226cf79fc5084729b8ab
d34bf012b1994787f3d1704326dd42effccb2c7b
F20110115_AADBVF dunsmore_l_Page_112.QC.jpg
27183d0cf02c9d12b6fa97820aaabce1
4eeff3c289c8883f399a1f6e76ec5dd6f2c92136
23604 F20110115_AADBUR dunsmore_l_Page_067.QC.jpg
8928170b14f079c8d21e8b2889aa4ac3
cfd4b34e96ee0846ddf77491bd8b8c78d59cd14d
F20110115_AADARP dunsmore_l_Page_107.tif
a7bf939a513823d38f9ae766be92e12c
f96121338e844aeba7643b5b0460c821b3cc16f4
F20110115_AADASD dunsmore_l_Page_121.tif
8115de4d7e1c788dbd3f9fe0a29db550
f418e407993a9d0e8e8134118e1d789ee4a1c784
23404 F20110115_AADBVG dunsmore_l_Page_113.QC.jpg
d23e3e6d7ef9ca191d393a092d83b678
96a945ccbd16cc9d65c165adcd693ba61dc4699d
20361 F20110115_AADBUS dunsmore_l_Page_070.QC.jpg
dd150b1d42cd0e8d829f14db491ef847
5ef11768380077d341e56ad3ae94214526fbbe6e
F20110115_AADARQ dunsmore_l_Page_108.tif
c03ef6f4a8049cde6b35e18f9574950b
3f68cf6f365873573ff6088efda54d2312fc48f4
F20110115_AADASE dunsmore_l_Page_122.tif
57263a921a353b267a8ef9b473bd85ef
e62d90f2a603449e83fa50ae9b289018ba4255a8
6526 F20110115_AADBVH dunsmore_l_Page_121thm.jpg
9b6d68b3c69b0c09b459081786c7fdea
975776e6714a1a00d4f2e7c50159e77647575974
6754 F20110115_AADBUT dunsmore_l_Page_074thm.jpg
c565b2db5f1124393f8e4177a29bf9ed
0e8fba3cc83eeb7061b97c85ad6897738466ab9a
F20110115_AADARR dunsmore_l_Page_109.tif
0ed6475db62ee60069acf1931ec65d45
67edd4a768a3f1c38944f161213378a58fc3ed3a
F20110115_AADASF dunsmore_l_Page_123.tif
4ed691ded7f4cbd2128960e8ffe9f2e1
7cd6e863c09b7ffb498761e01f0c299b409eb3ed
2824 F20110115_AADBVI dunsmore_l_Page_126thm.jpg
6a18d430c352cbee21447524b7c2442a
f79db3e0398cb01b6c311a9232d37b011c38f222
23368 F20110115_AADBUU dunsmore_l_Page_078.QC.jpg
a5aeb65874670bdb97d60e9637d0599c
4d979dd70307242f9287b6a32278036564a39e76
F20110115_AADARS dunsmore_l_Page_110.tif
eefd9c94cfb5f3587da25d64c5fecce6
2ef34e5f2530139e13c6746c3d47be2f17d49c73
F20110115_AADASG dunsmore_l_Page_124.tif
8d55bacdf03aa1e4c442df6e6299bcc6
6c3f224bed6b96be85a8f580ff1cd1f4fdd79eb3
3368 F20110115_AADBVJ dunsmore_l_Page_129thm.jpg
d464b0eada01ae872b8fc331af625c61
0bb8c5cf9edb3293e7a1c9555a71ea2ad6817cab
6589 F20110115_AADBUV dunsmore_l_Page_078thm.jpg
2afd89e33f76705258927d7fd26a5f44
e9857972abce103cef2dec821410a9cbe1089874
F20110115_AADART dunsmore_l_Page_111.tif
df7caff1b7e46638e25a087ebc4e01bb
856b30d75cadd80757381922f3f336c9a6b44787
F20110115_AADASH dunsmore_l_Page_125.tif
782997575a9928d13710dcab4a8ce19a
757cd0135a8e249b9e2296c72600c1c3840c4804
4613 F20110115_AADBVK dunsmore_l_Page_132thm.jpg
6cd61a567f33f383bb4f1370a8902270
35ca85215ccc6ac6b9eb82d34987551d041cd6d3
2189 F20110115_AADBUW dunsmore_l_Page_082thm.jpg
227b90baccb0b8d1c01644a87aaf29c5
e051641e6a44d7baf7a64cb5e0f20befa2d84cd1
F20110115_AADARU dunsmore_l_Page_112.tif
d6942265af6da33c22784e04f30f04dc
9e1b92fb8f3fafdbd9352310f1052246b018ecf1
F20110115_AADASI dunsmore_l_Page_126.tif
2f90f81dfb7dc4da92de41c545e83fb9
37a2a73f74d72e7a668e868e8ea9252f14d268ff
5708 F20110115_AADBWA dunsmore_l_Page_180thm.jpg
dcd3bbe2f504649623943cbf64d79132
4dd7ca59e162a127b084de78560e6d0e1d42cd50
11548 F20110115_AADBVL dunsmore_l_Page_133.QC.jpg
9603f5504195aa3ed53a91636477ecab
876f22ee5ef426e36e1c3349579df875592adda1
10088 F20110115_AADBUX dunsmore_l_Page_086.QC.jpg
7c6e79e03b6d0766c6c26ec144a3fc9a
8314e320fb00bd4b7ab0e30255b5477e6831594b
F20110115_AADARV dunsmore_l_Page_113.tif
d84172d36927b7045112cc41940223b1
26acbfe526dc50abb68cdcd305f28faf948ec871
F20110115_AADASJ dunsmore_l_Page_127.tif
cd3198cd6c284a8fe6f2bd144b4af5fa
13c8334e7d3b9b1292b2f0d4e0c31b0c8b891bd8
19816 F20110115_AADBVM dunsmore_l_Page_134.QC.jpg
021fea0ed4d86c628b9407c587cf1db8
f8e025bde3c410872a228c765b995ac3c4af57be
3275 F20110115_AADBUY dunsmore_l_Page_086thm.jpg
3edad18f02f85f3f9cec1c67e431548a
ea6a5bc316a70d19dabe3f669cd5538f0fc4997d
F20110115_AADARW dunsmore_l_Page_114.tif
ff176e99839c264254e29d29159dd60d
2d8f425cd79e7eca09f8249bc6122e29dabb517a
F20110115_AADASK dunsmore_l_Page_128.tif
3bf57d154ebe4b153ac99f88dbccb637
501ddcd00ced894223a8c2df96bc6d4b73ba6b5c
6628 F20110115_AADBVN dunsmore_l_Page_143thm.jpg
22380d2db97508c3e8657256a838b55b
a175b7d900b6465e3f4b606f664bb51371b33221
5498 F20110115_AADBUZ dunsmore_l_Page_088thm.jpg
17e337048cf979a56a1719abc130157c
eca5412d4ebf039aec5a053f647f36d5bef20d8d
F20110115_AADARX dunsmore_l_Page_115.tif
cee19553810350907729e54e68931f1a
f22f01b819f1b58accf99e75e73b683c4d1329ef
F20110115_AADASL dunsmore_l_Page_129.tif
9de668c8d678bdbd04925129fb784d3a
dcb00322bd2edb2a26236ba3fea30697325643a3
6607 F20110115_AADBVO dunsmore_l_Page_148thm.jpg
0d543befa76097849f460328cb3d5d93
374f987716580db3967ae1f2d9fbb6322fff4d1d
F20110115_AADARY dunsmore_l_Page_116.tif
3542664fc01ffa4ce5295a0784239f6c
2674e40092ceeb1883ddfb6d16a1d4fcc3647e51
F20110115_AADATA dunsmore_l_Page_145.tif
8300aee3c56c60f69d9662d57e74b4f1
cc86f03eaa78fbec15568bbdaa28e4ab4c9b1910
F20110115_AADASM dunsmore_l_Page_130.tif
24e00d7241b8ba60e6df2d02fbe62400
e5738e339de5f497d5d7ba2b5fd9393147e6648e
22852 F20110115_AADBVP dunsmore_l_Page_153.QC.jpg
0d1a71f29d050852a8e6c6da24f2d7c2
7988cff0920cf492b18064c2cde70e5cdb7a560f
F20110115_AADARZ dunsmore_l_Page_117.tif
eb5ec44fa705435aa0934a85a27a67d2
54b874d49d87535b99c1511e9d0f15d2e06c7eb9
F20110115_AADATB dunsmore_l_Page_146.tif
6c17627c65bb3ecec4374173ec284068
69184d01855b71b4a9b423a55c918eca827fca9a
F20110115_AADASN dunsmore_l_Page_131.tif
1d924c6a67085451c83eccd2e243b275
37add7afedacd25ccd02f20f3765afce174006d0
5980 F20110115_AADBVQ dunsmore_l_Page_159thm.jpg
135aba1152730720a4c8cbf48e58585e
80112c00f45bca54a197180b811ade5cc3948ce6
F20110115_AADATC dunsmore_l_Page_147.tif
6881f3cf3e95263888cf17959f75a00e
9a99244c3b51ffeb13fbb80e95d7cff4d3a63092
F20110115_AADASO dunsmore_l_Page_132.tif
b4a6f61e6e00e654f16bb763392ace00
4743cb3eeb18f551f92d96d9cd2ff0385b8ce8da
5619 F20110115_AADBVR dunsmore_l_Page_160thm.jpg
ff4038b73b4966bbc0670ac507d4e287
7095d175241e5b00e16f665ebf671d4a81e1f3da
F20110115_AADATD dunsmore_l_Page_148.tif
9346b295500c21f9fcc7555f05a8867e
e840cdd3aa025ddefa65b5587d803fcfb1267294
F20110115_AADASP dunsmore_l_Page_133.tif
b61211cbd12b97da8238c8e9c836a477
9bfd5df9eb29fa891d75dc67a6eb8c338d4ce975
25454 F20110115_AADBVS dunsmore_l_Page_163.QC.jpg
bc6ec572994c1616ddc8f03982ef9933
8c9d720a08e147dab861a3e3f4e9f469d38686ac
F20110115_AADATE dunsmore_l_Page_149.tif
16e43d044cf7d8a348d57fc059f9cf47
7018413710a90b34fd8e46f53831e4fcbfd7068f
F20110115_AADASQ dunsmore_l_Page_134.tif
d202521b8a0e251607a3523ec1e7f13d
7dea405d0c3d32fe5fa879484ea9aa45b660f852
26368 F20110115_AADBVT dunsmore_l_Page_165.QC.jpg
dc77befb0607769117eeb751579b0e2f
b655da93ed8e6d3be3e160a68b20858e3301b1d5
F20110115_AADATF dunsmore_l_Page_150.tif
26626b27ed71224cb8cf01754a2407e7
8c5e46b846ef6a664afa5cb535de238dadfc3e64
F20110115_AADASR dunsmore_l_Page_135.tif
ab8088146fc116a403dd239175d9e4b1
3459982aab807c6d74e6436aeb83a42f3478aab7
26799 F20110115_AADBVU dunsmore_l_Page_168.QC.jpg
941a9e7dac666ddc1acf8e0dda431990
3d10813f0554b98677cda7bf928a709cb06061a9
F20110115_AADATG dunsmore_l_Page_151.tif
b51982d21cb580a70e2ee42ebf49826a
511e92c26bdb0cc71e49071e6870699f387c85f2
F20110115_AADASS dunsmore_l_Page_136.tif
5e58c7cd6a12af0f73f8aba31cb2c5f5
bb1aa416b898422866d66fa69f2a303614ceccac
7422 F20110115_AADBVV dunsmore_l_Page_168thm.jpg
ade908a765c200dcbda04343af5c8699
bf77956c00e65ccee8ba2f855232ccb9635e0c45
F20110115_AADATH dunsmore_l_Page_152.tif
7f6957da9db63bebc445144b77f0530a
f9bc1e6d23f734caae3a656adb399d5e29d0a27b
F20110115_AADAST dunsmore_l_Page_137.tif
3149c25166512859c8ee56f2e01f5ee2
281dd50d7948ae119c2dae4fc26e7bbada95813e
7370 F20110115_AADBVW dunsmore_l_Page_172thm.jpg
5d164a4925e96ff18b8688eb50d98494
ada82c8c829868fbacecf8141be0c8d926f8fa24
F20110115_AADATI dunsmore_l_Page_153.tif
0f8066efb39986cc53cabfb25ff287cb
fdda90f12cc6521dd088df5fc06f78e95a2b1e57
F20110115_AADASU dunsmore_l_Page_138.tif
6e4b7020065929b01a7e6909d8a83986
cacf4afc323fcba4ef30284f113b58d0c3418c2e
23467 F20110115_AADBVX dunsmore_l_Page_174.QC.jpg
c37f7fcd904a7c437c32b855f09a76a5
c53ce5e0a04d5d59f3c0221ec5b6efccec1eeac9
F20110115_AADATJ dunsmore_l_Page_154.tif
c9fe14b42b590beca4cbf91cf81d41a6
c20c7e78d8d8effc82c14e207c7a8702b99580ec
F20110115_AADASV dunsmore_l_Page_139.tif
c2cdd7d053fb3e7087e63f3986cc80bc
36fbed6c7ee1849173f18cdec478d4fcb190b56e
25731 F20110115_AADBVY dunsmore_l_Page_177.QC.jpg
9edbeaebfeb0206fec4c9d2b1038931e
8de73a281a8336bff65ad88028f20b98a44bc0f1
F20110115_AADATK dunsmore_l_Page_155.tif
051ed12a738d366183f07c1db65091dd
5614898c0e719ddb282e25ee4186f8742c1b6014
F20110115_AADASW dunsmore_l_Page_141.tif
160a574e63adc831a7973e4d6830aabc
185f51f605155e677f5bf4530a14b5c4124ef404
3445 F20110115_AADBVZ dunsmore_l_Page_179thm.jpg
033059d9470675f7370b3d0a1084549f
ba07687aab054359e774a25d57764f082cc8a47d
F20110115_AADATL dunsmore_l_Page_156.tif
df792bf8ac2863b69ae5a6d8deee81be
4c045f6e1a41c3a3fd77adb7742d7f7dcb9ee37f
F20110115_AADASX dunsmore_l_Page_142.tif
f05019bb55f977280d2f06968adb628b
a95c9ae41227fb515b7bc9471240b99cf6ace295
F20110115_AADAUA dunsmore_l_Page_171.tif
58a7803f15b185b77ec1462586ff3879
eccfa79d50e76fe13aea3f9ce20560869dfb3b9e
F20110115_AADATM dunsmore_l_Page_157.tif
a12a04f3b1bf9a0d4ac9462ee2c92781
795fbe69958885419e048aea211a670074f9a352
F20110115_AADASY dunsmore_l_Page_143.tif
af31b994839a7d4a00b39d9e0f5f7106
ebf6c2cc44b988ff3ef59ab7ab5a485f76376c6e
F20110115_AADATN dunsmore_l_Page_158.tif
ced07d978c8aa82e3a9b27d8db13705a
85adbbe55d9c4fc2aa8775707873fe1638cf89d3
F20110115_AADASZ dunsmore_l_Page_144.tif
8b5f95a200fd70a0c4112c4ef42cacbf
975a8e83a4ec4d531e2d4dcb0c0961b16a40e14c
F20110115_AADAUB dunsmore_l_Page_172.tif
22a7c1853c9348112c8b87eaa94311af
3a62ec89ba8a711ff789cab5e65819ef21f80f36
F20110115_AADATO dunsmore_l_Page_159.tif
cda0ff93c93163b1f67fccb5295647b0
7947e1b53b56f710b49c218c2ad00bbf2ded0453
F20110115_AADAUC dunsmore_l_Page_173.tif
8fb29eac9043d2439cd688df0ade611f
4986d4c0e5007625018f2adad9581c199601e360
F20110115_AADATP dunsmore_l_Page_160.tif
02cdcbbebc9c6454df1d633f4630d12c
8f1faf5697d6469d1285b0d030137b211d76604f
F20110115_AADAUD dunsmore_l_Page_174.tif
d73eab86dbde14d2e2e37066b80098da
5be0d53ac75ac85925477c894ed84e84926b9ca4
F20110115_AADATQ dunsmore_l_Page_161.tif
be0a27ef4735275665a2be7d7cc82de4
e8f53943ea2fe0c245432acf12fa065dc898887a
F20110115_AADAUE dunsmore_l_Page_175.tif
b9a4ebff28e77601a89783f1c34d7761
c896650d1f6223fae10b861e250f03cc10282229
F20110115_AADATR dunsmore_l_Page_162.tif
e10a674b6b172b7ec6f488a41e1e5424
3b4c3e69b458475e1e33c7eb55451e8c45b2497f
52839 F20110115_AADBAA dunsmore_l_Page_149.pro
94731666a9d42d87e63305a8aa47c164
fe20ac58b30e9909f77dcca6b83407f661b728d0
F20110115_AADAUF dunsmore_l_Page_176.tif
ccc58957248555f8417d7dc76a6e2632
3870e2b6da49d9ac7eabac04927cfaf91496d1fc
F20110115_AADATS dunsmore_l_Page_163.tif
df1b145455c928147db894bcf0f2c7ac
461e1dd3616b625309661be3ff2519f74913d8bc
37268 F20110115_AADBAB dunsmore_l_Page_150.pro
d8b26f4fcd22b2cd6951a0e908d3be76
3606615eba42084355d66ac7f66762ed07dd6c22
F20110115_AADAUG dunsmore_l_Page_177.tif
43ea033b8541da75e5327fee0f4c4876
c658a66a27b639c646240d71f605907e671d7cbe
F20110115_AADATT dunsmore_l_Page_164.tif
c97be14ccd318484be74207f03a36d79
d0f0d1be68c4a93e951f3979d37049788687fa04
52434 F20110115_AADBAC dunsmore_l_Page_151.pro
085b51775f9dd1a5401f73b035516638
7b062b28906616847963c2ff45c45addfed97f59
F20110115_AADAUH dunsmore_l_Page_178.tif
48b89f60665f684f6c1e8872a25131aa
3f686fecdf1e85c87139707dbcd5bd198004c7fd
F20110115_AADATU dunsmore_l_Page_165.tif
ba93c3b2496eefba1fee59229978d972
ff9eb722370e7d67c0e653fb758550b9f84833f3
50212 F20110115_AADBAD dunsmore_l_Page_152.pro
0560e140661f1282e9fd895d5b66e7ed
c95d34fd404edde50cd6aec2fee8767f15dc8e47
F20110115_AADAUI dunsmore_l_Page_180.tif
044de3c24fc0bc772f8bd99db08db4d9
1f2b6028b1bb31dbe67fadc92a1902af36c585c9
F20110115_AADATV dunsmore_l_Page_166.tif
fdf62d9b8d335dd38e85334310bc2eb8
05fd811fe01c493b788459ab874848882742d9ba
52055 F20110115_AADBAE dunsmore_l_Page_153.pro
5f902720ef04b41f0f6eff025c3343c5
611ab81f14c4a8d4bce89377b6fed94a2e0502b5
F20110115_AADAUJ dunsmore_l_Page_181.tif
10ad24ce7dc6b06e950049edb3224167
d5ce7bbbe5d93ddaa32c26c5675545159c719bd1
F20110115_AADATW dunsmore_l_Page_167.tif
7f32efd49d08bbcf83fe7d65920b6604
b6f92be7419194e89120f3b7d9cf98ec4b38e094
53223 F20110115_AADBAF dunsmore_l_Page_154.pro
0914747f0ca2736a0f9612617e920a82
62914495bf38313d129c22d2db6e2db1b36062ae
10153 F20110115_AADAUK dunsmore_l_Page_001.pro
56b9f6a263aa21cce5e3c4299a318348
63fadea8c98964ce3af2e8ff99e87dc3908b579b
F20110115_AADATX dunsmore_l_Page_168.tif
dcbfbf8ba8a811fb2cb600f921317aac
fa55456c1a3d71c4a5521844fc9176e78bc036a5
55537 F20110115_AADBAG dunsmore_l_Page_155.pro
7b3ad52e0108449a56de24e0ceb87cd5
ab1417679755e4d6e6a2d9e7cec6a089530a922c
29460 F20110115_AADAVA dunsmore_l_Page_017.pro
c12a890a24e99aac44b85c2eed62664a
ba10d255c1a23f4cbdf94ce1a5e606f66099c7eb
1697 F20110115_AADAUL dunsmore_l_Page_002.pro
961e923c521029a1d951c032b3bc4bad
bfb453ebbf7b9c158d872e6b98bf52379ba4fe08
F20110115_AADATY dunsmore_l_Page_169.tif
b923f55490b5114a070a49ede1ce10d9
5f44e230bf185960721f94839743f508823705ee
57899 F20110115_AADBAH dunsmore_l_Page_156.pro
5813dffa97529c7ee9bf49ff70b44ae1
5e3ea3eeb9375e6f8f9281fdbeea93f2bda2365e
45904 F20110115_AADAVB dunsmore_l_Page_018.pro
cb5f16d20c24d930b179be1414639f6f
61b2f595ef37d2792e66a0463dfe3baf6bc185c1
5961 F20110115_AADAUM dunsmore_l_Page_003.pro
00de4c420261ffcdf305b65241cb3d93
d93600853bc6979adf4975738c714447d3d38135
F20110115_AADATZ dunsmore_l_Page_170.tif
8c561e675f1baa62d6adb2015924c1b9
149ca3f5608ba15d7f2dbc89a464d2c98208097c
44574 F20110115_AADBAI dunsmore_l_Page_157.pro
2321dc11acee3c25896471cb6f215dbb
3694528cccc67ecfd38a4c80038fde394de8bcde
42930 F20110115_AADAUN dunsmore_l_Page_004.pro
69be982e37e920bd34053df0b4f032f2
fd0a1db003050efa338368f65cbe91869d2c14ab
47146 F20110115_AADBAJ dunsmore_l_Page_158.pro
fb50737d998c990fc7d2df687378d48d
bbcab4141dc6763c967e5cc4416c18e79b1285eb
51378 F20110115_AADAVC dunsmore_l_Page_019.pro
318b957d3f340211b0650a934096c963
b7f81df5d0ed6e8e943c3de7b193c2fa222593cd
45291 F20110115_AADAUO dunsmore_l_Page_005.pro
c7b870584931c5c8a4f9e7dfc2ff07cb
c13e19eca26c6cd45dc722a3b79a15f52e38d2e9
47810 F20110115_AADBAK dunsmore_l_Page_159.pro
cefe168de98f84567b78fd7ee037b1de
8bbf4555b57f4d123708df06b8a2edbcbeca44af
52126 F20110115_AADAVD dunsmore_l_Page_020.pro
13b1ce73a10f243daa37d3c91a3b4282
c0db9ab68ccc9fb2b648f8cd3cef3a80187f4816
49229 F20110115_AADAUP dunsmore_l_Page_006.pro
bc83a10a13dd664fb95785cb5146a31f
29a969ae5b576870fb7a82de9c6791418014cf62
41267 F20110115_AADBAL dunsmore_l_Page_160.pro
fb12dc4e4765b1a22904f7a0b17fc4f5
fd6b68d4a3a745e11f4e989563c46a641562f751
52371 F20110115_AADAVE dunsmore_l_Page_021.pro
8842b3f5ed1f4de7e0786c3208760dba
f7df8558ad4f46b2398bf13791f8411c135e8dc4
49758 F20110115_AADAUQ dunsmore_l_Page_007.pro
29aa7a14a67b8b46bbc717131ebc799c
7c1849e76f76d7e7e5d6e1989bbfadfc52f70d3e
50783 F20110115_AADAVF dunsmore_l_Page_022.pro
69f3c27044d1aa17fbeb1d17a0f75a7b
8865ee0cf306ffc3b4607fbaff6e77e8a56161c9
13208 F20110115_AADAUR dunsmore_l_Page_008.pro
226bed19e717105df1fd90bd890ee94a
22b6fdf68ff8a45e706ffacd0476299d4c7e99b7
58590 F20110115_AADBBA dunsmore_l_Page_175.pro
7254ce713ec0c52871518eb805a90b24
cde8766b54811f6e219e68dd13fe771d054961a8
43239 F20110115_AADBAM dunsmore_l_Page_161.pro
277b0b8dd4f80996120b403edf8fdbbb
b7cea65821cc803e906f67de329e384f13edb854
51446 F20110115_AADAVG dunsmore_l_Page_023.pro
a02adc9f2c778032011e03624f98e11f
3f2558f4fd3ae69a534ecd44db7faacff9ecc795
75408 F20110115_AADAUS dunsmore_l_Page_009.pro
49bdc8e8f120bdbe8c8df68ed12dc1ab
17cd4bf15de3fe85c71e265ef63e16c454a12e45
55545 F20110115_AADBBB dunsmore_l_Page_176.pro
73035710a2b255233065b44e05a8c4f0
260c768d92552840d111b0f8edc6913af7923fa6
49996 F20110115_AADBAN dunsmore_l_Page_162.pro
04cfd4f81ef26d238c1885d79b55149a
bdba6e97ef141a0b9cb138c4936fee659b4ff0be
51173 F20110115_AADAVH dunsmore_l_Page_024.pro
6339d074a844bdb08a435aef14ced6b7
182940466c71537d9e7cb32a7d776715fb138e9d
63673 F20110115_AADAUT dunsmore_l_Page_010.pro
e6848d564adca4eb16963451bf7ee405
8624b9c786fbbeab8839d6801fe1bb781719d4bb
61775 F20110115_AADBBC dunsmore_l_Page_177.pro
c712081e2bb7cc0fd36274196f0040e8
8a0a12ea7e9733dc64e723638251893ffcc8bd62
65708 F20110115_AADBAO dunsmore_l_Page_163.pro
e41dbfb0ae566a4ab252e1d9e8f8d758
75db3624567b9e1f94f3ef70fabfd60af6d89e0a
49876 F20110115_AADAVI dunsmore_l_Page_025.pro
775be3bca2f6bcb84ecf7f69762fdaad
4f782a7545ffab874752c7dd0944c2e5c7a42970
67210 F20110115_AADAUU dunsmore_l_Page_011.pro
e04ae69f70f32249df4556685554db01
8a9ed4dde21896c9e2cf7d59f725985abf6ff9d5
60177 F20110115_AADBBD dunsmore_l_Page_178.pro
9531fea1ae6baf54698c11afd0bad1e1
87a8d2e571a4b82819f75ba6999e7805b5e627db
66633 F20110115_AADBAP dunsmore_l_Page_164.pro
4ea2887b86e3009f94a67d917c07baab
efffbf2d29b901b483254f1a80c9cef25b7ebbea
50709 F20110115_AADAVJ dunsmore_l_Page_026.pro
903ed569e0e0d969ea3bb5eb94c62b28
fa2cab4f796ab3a015bae9fe46e48e23d36d627f
82480 F20110115_AADAUV dunsmore_l_Page_012.pro
76fbdf336e69f070bc4a30f17d8dced9
06e634f49f68bc2928280800dfa85ecfc8dd2914
23131 F20110115_AADBBE dunsmore_l_Page_179.pro
ea6eafe7634287747f6b7aedbf7e785b
efa65b4936cc668edaf0478dab2bbeab9080e20e
62009 F20110115_AADBAQ dunsmore_l_Page_165.pro
05786c5bff3ab3d5d2d98d497dcb87fe
efd9935dea89c052891e35b66d10065728f19fae
41536 F20110115_AADAVK dunsmore_l_Page_027.pro
6f06ff44ee352afe18602fd23e38de82
8ec652fd45c51b38201fc16dae8ed53046adf6fb
16127 F20110115_AADAUW dunsmore_l_Page_013.pro
0a715366f07a201fcf6c3800f2fb4370
e070f2de44686ce1cad9365246a368c267e22f45
41241 F20110115_AADBBF dunsmore_l_Page_180.pro
ca9f8f99dd20e6d5fb6d6a7f36680434
6be0f9333fb133395d8977667eee836a8e29b7ec
47587 F20110115_AADAWA dunsmore_l_Page_043.pro
76d797d4e24995adae03c4ddb0ba8606
6afb4b6d33b7a40a264178a410072f314fae8341
58306 F20110115_AADBAR dunsmore_l_Page_166.pro
e9c8d044dabfbd6eedee240666ec4551
9dc88008ddbc6d2a10695703d1edcd3b3eb770d0
45681 F20110115_AADAVL dunsmore_l_Page_028.pro
709eae7eee802a43499246b6b3619343
8bf9598e63e1dc2d91fbb36e051ef8c275035bb9
62505 F20110115_AADAUX dunsmore_l_Page_014.pro
a391ba486349a8e14b7f8a53b5b1d70f
59c9974e1a686999930f29b1e150138f247435cb
17058 F20110115_AADBBG dunsmore_l_Page_181.pro
979403e01b99efb819dbd61456b86304
09fb07c7c9aaab3945d82d1f73b339570f0b8931
51277 F20110115_AADAWB dunsmore_l_Page_044.pro
e3640fd90d3b36a18de9108633ff5b39
13c8b854814dd9b0b9c329af9df7270781842111
63711 F20110115_AADBAS dunsmore_l_Page_167.pro
aa416782bb135f621cad57fc79c9ef00
a5f5d13e2a458589ba19b4b642e6ce03ff3ab73d
51083 F20110115_AADAVM dunsmore_l_Page_029.pro
dce62b5a6f4ed6a17bc1fbd1f69f2966
8f79e1e56f1256ee65c059773dd2667140876fc3
57168 F20110115_AADAUY dunsmore_l_Page_015.pro
863dfe7dfecc0978812a4667db1952d5
4e5306bedcfe7d128d73bd6adc8bc768b13ed8a4
521 F20110115_AADBBH dunsmore_l_Page_001.txt
abfdae2a99ae86d3756e7355003c0f31
51903f187b2a343fdefb8fc85780a8069287cf96
50367 F20110115_AADAWC dunsmore_l_Page_045.pro
3a86e4cfda3cea634aeeba89f6584e97
6f92d162462070a55f9cd088ed8b391347dba685
61715 F20110115_AADBAT dunsmore_l_Page_168.pro
531430762f19be21cc7125409bb11418
b9755c196af47831c6f6702a6c5f4bd74a93ed87
52219 F20110115_AADAVN dunsmore_l_Page_030.pro
f334117cf0a1c306319d0c0be94094b6
84ec118d2769183f3d76eeb9ce23f30bf3de56be
41292 F20110115_AADAUZ dunsmore_l_Page_016.pro
ed0b9f7989ca0f6f91e976a1e070abf7
9ac09918f6589b4c4a5cb994a77bfe0f085021e3
149 F20110115_AADBBI dunsmore_l_Page_002.txt
a1d34eb82e3b75afb6e2cd430428368b
0c584f0fd157d6cf1b26db15d5ad7241d0228c04
65256 F20110115_AADBAU dunsmore_l_Page_169.pro
21a8b525ff1536dd65c9a1174317dc16
847468cd014816523f78b4770830b85d0c4b94ba
49340 F20110115_AADAVO dunsmore_l_Page_031.pro
610f8df8a2c774850b1a126e3306fb76
74c7ad10f74abcd5eef723598847dd26d8028ef3
304 F20110115_AADBBJ dunsmore_l_Page_003.txt
c2418feb978189bc85c2940c13d55955
6bf88052b75600671f5d364d44219d77aed98cf6
52479 F20110115_AADAWD dunsmore_l_Page_046.pro
5971f335eef514de3135670ac40b25c4
a352eb5dc4cda71817d49bb96f085941fdf3d507
60921 F20110115_AADBAV dunsmore_l_Page_170.pro
9e5a882b1a5cf50afc43af678c5ebba6
6f1809b46447af0b34dc9adf7bf6a60e120d8cab
52040 F20110115_AADAVP dunsmore_l_Page_032.pro
e7d6d22d2c9a4c848d83afcf85db369a
b0c667aca5cd5b9257b85ed80c3d0a23385bf2ee
1732 F20110115_AADBBK dunsmore_l_Page_004.txt
850e470e5efc4a7bd7569a70d5dd521a
6c7d5887954f8ab0cd4c7eae2f74dff7b0b16c9d


Permanent Link: http://ufdc.ufl.edu/UFE0008350/00001

Material Information

Title: Applying Terrestrial Landscape Ecology Principles to the Design and Management of Marine Protected Areas in Coral Reef Ecosystems
Physical Description: Mixed Material
Copyright Date: 2008

Record Information

Source Institution: University of Florida
Holding Location: University of Florida
Rights Management: All rights reserved by the source institution and holding location.
System ID: UFE0008350:00001

Permanent Link: http://ufdc.ufl.edu/UFE0008350/00001

Material Information

Title: Applying Terrestrial Landscape Ecology Principles to the Design and Management of Marine Protected Areas in Coral Reef Ecosystems
Physical Description: Mixed Material
Copyright Date: 2008

Record Information

Source Institution: University of Florida
Holding Location: University of Florida
Rights Management: All rights reserved by the source institution and holding location.
System ID: UFE0008350:00001


This item has the following downloads:


Full Text












APPLYING TERRESTRIAL LANDSCAPE ECOLOGY PRINCIPLES TO THE
DESIGN AND MANAGEMENT OF MARINE PROTECTED AREAS IN CORAL
REEF ECOSYSTEMS













By

LINDA ERICA "RIKKI" GROBER-DUNSMORE


A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL
OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT
OF THE REQUIREMENTS FOR THE DEGREE OF
DOCTOR OF PHILOSOPHY

UNIVERSITY OF FLORIDA


2005

































Copyright 2005

by

Linda Erica "Rikki" Grober-Dunsmore


































This dissertation is dedicated to my bedstefar, Jacob Nielsen and grandfather, Hyman
Grober. You shared your passion for and intrigue with life and knowledge; and are my
models of hard work, integrity, and compassion.















ACKNOWLEDGMENTS

Many people provided tremendous assistance and support in the completion of this

dissertation and all of them deserve special thanks. I apologize if I have inadvertently

omitted anyone; it is not my intention. I extend my thanks and sincere appreciation to

everyone who has helped in any way, along the way.

I greatly appreciate the guidance and support of my supervisory committee:

Dr. Thomas K. Frazer (Chair), Nicholas Funicelli, William J. Lindberg, and Paul Zwick.

It has been a great pleasure working with my advisor Tom Frazer, who gently supported

me in my research and scientific development, and served as a mentor and friend over the

past few years. His hands-off style and steadfast belief in my abilities fostered my

confidence as a scientist, and I truly look forward to collaborations with him throughout

my career. I would not have been given this opportunity without one person (Dr. Nick

Funicelli, who I will truly miss as a friend and confidante). I relied on his insights on

life, science, and management and learned much from him. Always cheerful and upbeat,

he picked me up many times on this journey. I appreciate the support and encouragement

that Bill Lindberg always provided, and the many discussions about the philosophy of

science and my research results. He was always available to discuss science and he

opened my eyes to the rigors and history of the scientific method. I have learned a great

deal from Dr. Paul Zwick about geographic information systems and systems ecology.

His enthusiasm, honesty, and optimism were invaluable to me throughout this study.









I thank Jim Beets for pushing me to achieve my highest standards and for giving

me the impetus to reach for the impossible in myself. Through his passion for fish

ecology and natural history, I became addicted to learning about corals reefs ecosystems,

and applying my strengths to support coral reef conservation.

My study would not have been possible without the assistance of many people. Dr.

Mike Allen, Dr. Chuck Cichra, and Howard Jelks, each helped by providing statistical

advice, often without much advance warning. I also thank the scientists and resource

managers at the Virgin Islands National Park who facilitated my field work on St. John,

US Virgin Islands, (particularly Rafe Boulon, Chief of Resource Management; and Dr.

Caroline Rogers, Research Scientist of the United States Geological Survey) (USGS)).

Thomas Kelley, Jeff Miller, Rob Waara, Sherrie Caseau, Jack Hopkins and Jim Petterson

of Virgin Islands National Park assisted either in the field or through logistical and

administrative support while on St. John. Ilsa Kuffner, of the USGS, helped in the field

and has been a valuable friend. Working in the field on St. John with Alan Friedlander,

Jim Beets, Nick Wolff, and Ellen Link piqued my early interest in marine ecology.

Those days in the field in the Caribbean continue to be some of the most memorable,

comical, and enjoyable days of my life.

In the Florida Keys, many people helped to facilitate this research. Particularly

helpful were Billey Causey (FKNMS), Dr. Jim Bohnsack (NOAA-NMFS), Ben Richards

(FKNMS), Brian Keller (FKNMS), Joanne Delaney (FKNMS), and the people at

National Undersea Research Center in Key Largo and Mote Marine Lab in Summerland

Key, Florida.









To my tireless and optimistic field assistants (Jason Hale, Victor Bonito, Thomas

Kelley, Luis Rocha, George Dennis, Mary Hart, Duncan Vaughan, and Doug Marcinek) I

am eternally grateful. They spent many hours underwater; they counted hundreds, if not

thousands, of fishes; and endured many days subjected to the vagaries of sea lice, sinking

boats, and rough seas. Some even ate their words instead of their lunch, and carried on

with a smile. I cannot thank them enough.

I thank my labmates (Sky Notestein, Stephanie Keller, Jason Hale, Jaime

Greenewalt, Dan Goodfriend, and Kate Lazar) for their constructive criticism and

sympathetic ear during moments of crisis and sheer exhaustion. Kelly Jacoby is

responsible for helping with all of the figures and tables, and she always offered her help

generously and with a smile. I also appreciate the folks in Dr. Gustav Paulay's lab, who

contributed greatly to my experience, both scientific and personal, at the University of

Florida. Dr. Gustav Paulay, Chris Meyer, Lisa Kirkendale, and John Starmer each helped

develop my scientific thoughts, and often reviewed presentations for scientific symposia

or drafts of publications.

I am exceptionally grateful to Victor Bonito for his cheerful encouragement,

infectious precision, timely pep talks, and child-rearing discussions. As a scientist, he is

a model of keen logic and intrigue; and as a friend, he is as good as they get.

Financial assistance for this research was provided by several organizations. Dr.

Russ Hall and Dr. Nick Funicelli at the Biological Resources Division; and Dr. Suzette

Kimball of the Eastern Regional Office of the Biological Resources Division of the

USGS supported my development as a scientist both administratively and financially.

Dr. Gary Brewer (of the USGS) has always believed in the USGS Coral Reef Program;









and Dr. Jeff Keay and Dr. Lynn Lefebrve deserve my sincere appreciation. The

American Academy for the Advancement of Scientists, the Canon Science Scholars

Program, and the National Park Service generously supported all stages of this research.

I also thank the Graduate School of the University of Florida for assistance with

logistical, administrative, and financial concerns.

I cannot adequately thank the administrative staff of the Department of Fisheries

and Aquatic Sciences (Jennifer Hemelbracht, Susan Morgan, Melissa Altomari, and

Sherri Giardina) for attending to the many intricate administrative details that are critical

to a project of this magnitude. I also thank the administrative staff of the USGS

(Christine Fadeley, Tracy Marinello, Brenda Turrentine).

Considering my faithful cadre of friends, I thank my lucky stars (the Pleaides) for

gifting with me such a loyal and fun group of people. They helped me achieve this goal,

and made my life richer through their kindness and support. They each helped me find

the balance in life, work, and family, and gently served as constant reminders not to take

any of this too seriously. I extend my hearfelt thanks.

Most important, I thank my family members for their everlasting support, for

enduring hours of long distance calls, and for encouraging me to continue when it was

most definitely not the easiest thing to do. They tirelessly listened to my soliloquies on

science topics when they didn't necessarily understand, nor care to. My father, Ron

Grober, passed on to me an incredible strength of will to live and learn, and gave me the

tenacity of a bulldog. His recent support means more to me than I can express. I thank

my mother, Hanne Nielsen, whose warmth and compassion (and carefree view of our

journey through life), gave me the ability to laugh at myself during this endeavor. I thank









my sister, Suzanne Rose, who suffered years of being dragged around in a rowboat as a

child; and who as an adult has become my closest friend. Most of all, I extend my

deepest thanks and appreciation to my son, Thatcher Kai, who has walked this path

beside me from the beginning, and continues to be my model of optimism. He has had to

sacrifice much in order for me to fulfill this goal, and has taught me so much along the

way. I hope that my work contributes to a better world in his future.
















TABLE OF CONTENTS

page

A C K N O W L E D G M E N T S ................................................................................................. iv

LIST OF TABLES ........................................................................ .... xi

LIST OF FIGURES ......... ....................... .......... ....... ........... xiv

ABSTRACT ........ .............. ............. ...... ...................... xvi

CHAPTER

1 IN T R O D U C T IO N ............................................................................. .............. ...

2 INFLUENCE OF LANDSCAPE STRUCTURE ON REEF FISH
A SSEM BLA GES .................. ............................ .. .......... .............. 11

Introduction .......................................................................................................12
S tu d y A re a .......................................................................................14
M e th o d s ..............................................................................14
R eef fish S am pling ...... .... ...................................... ................ ...... ................15
H ab itat S am p lin g ... .......... ................................................ ......... .... ..... 16
Data Analysis ......................................................................... ......... ................... 17
R e su lts ...............................1 9.............................
L landscape Structure ........................ .................... .. ........ .... ....... ... 19
Reef Fish A ssem blage Structure ........................................ ...... ............... 20
D iscu ssio n ...................................... ................................................. 2 2
C o n clu sio n s..................................................... ................ 2 9

3 EVIDENCE OF FUNCTIONAL CONNECTIVITY IN A CORAL REEF
ECOSYSTEM ...................................................... ........... ............... 45

In tro d u ctio n .......................................................................................4 6
S tu d y A rea ................................................................4 9
M e th o d s ..............................................................................5 0
R eef fish Sam pling .............. .................................................... ..... .................50
T em poral Sam pling ............................... ................ .. .. ........ .... ............5 1
H habitat Sam pling .......................................................................5 1
Statistical A n aly ses........... ...... .................................... .............. .. ... .... .... .. 52










R esu lts.................. ............ .. ........... ..... .................. ................ 53
Entire Assemblage Level Parameters ........................................................54
Abundances within Reef Fish Groups .............................................................54
S p e cie s R ich n e ss ............................................................................................ 5 5
M o b ility ............................................................................................................... 5 5
Spatial E xtent..................................................... 56
Tem poral Consistency ................ ............. ...... ................... ............... 56
Relative Influence of Fine and Landscape-scale Measures.............................56
D iscu ssio n ............. ......... .. .. ......... .. .. ......... ........................................5 7
C o n clu sio n s..................................................... ................ 6 4

4 REEF FISHES RESPOND TO VARIATION IN LANDSCAPE STRUCTURE .....79

In tro d u ctio n .............. ..... .......... ........................................................................... 8 0
M e th o d s ..............................................................................8 4
S tu d y A re a s .................................................................................................... 8 4
Habitat Sampling ...... .................. .......... .........86
R eef F ish S am p lin g ........................................................................................ 8 7
Statistical A nalyses.................................................. 88
Results ............. ..................... ............. ...............91
D iscu ssion ......... ...... ............ .................................... ............................97
C o n c lu sio n s.......................................................................................................... 1 0 5

5 SUM M ARY ........... .... ..... .. ... ........................ ........ 120

APPENDIX REEF FISH DATABASE .................................. 134

LIST OF REFEREN CES ............ ........................... ........................... 14

BIOGRAPHICAL SKETCH ................................. ........ ..... ..................... .. 163























x















LIST OF TABLES


Table pge

2-1 Reef fish assemblage parameters (n = 30) used as dependent variables in
statistical analyses ....................... ........................ .. .. ......... .............. 3 1

2-2 Fourteen metrics used to quantify the landscape structure of the 20 study reefs
sampled in 1994 and 2001 in St. John, USVI. ................................. ............... 32

2-3 Summary statistics on reef configuration, context and rugosity and for select reef
fish assemblage parameters (entire assemblage level, trophic level and mobility
guilds) for 20 study reefs sampled in 1994 and 2001, St. John, USVI for metrics
at the 100 m spatial extent with coefficient of variation for landscape parameters
and standard error for reef fish parameters. ............................ ....... .................33

2-4 Pearson product moment correlation matrix of the 14 landscape-scale habitat
variables, with the resultant 9 remaining significant variables, at the 100 m
spatial extent for the 20 reef sites sampled in 1994 and 2001 in St. John, USVI. ...34

2-5 Principal component analyses on the correlation matrix of the 8 residual
landscape-scale habitat variables at the 100 m spatial extent for the 20 study
reefs sampled in 1994 and 2001 in St. John, USVI...................................... 35

2-6 Stepwise regression results to determine the influence of principal components
on reef fish assemblage structure at the 20 study reefs sampled in 1994 and 2001
in St. John, U SVI at the 100 m spatial extent .................................. ............... 36

2-7 Stepwise multiple regression results of the influence of reef configuration on
reef fish assemblage structure for the 1994 (N = 14) study reefs in St. John,
U S V I .............................................................................. 3 7

2-8 Stepwise multiple regression results of the influence of reef context on reef fish
assemblage structure for the 1994 (N = 14) study reefs in St. John, USVI .............38

2-9 Stepwise multiple regression results of the relative influence of landscape and
fine-scale habitat measures on reef fish assemblage structure on the 1994 (N =
14) study reefs on St. John, U SVI. ........................................ ........................ 39

3-1 Most abundant taxa in each reef fish group for the 22 study reefs sampled in
2002 in St. John, U .S. V irgin Islands. .......................................... ............... 71









3-2 Variable names, transformations, minimum, maximum and mean values for each
reef fish parameter and landscape metric, with the standard error for fish
parameters and the coefficient of variation for habitat measures for the 22 study
reefs sampled in 2002 in St. John, US Virgin Islands...........................................72

3-3 Simple linear regression for entire assemblage level parameters of reef fish
communities with the areal coverage of seagrass within 1 km of each study reef
as the independent variable at the 22 study reefs, sampled in 2002 in St. John,
U .S. V irgin Islands*. ........................................ ................... ......... 73

3-4 Simple linear regression of abundances of mobile invertebrate feeders, grunts,
snappers, groupers, and seagrass-associated taxa and within mobility guilds with
the areal coverage of seagrass within 1 km of each study reef at the 22 study
reefs, sampled in 2002 in St. John, U.S. Virgin Islands*.............. .................74

3-5 Simple linear regression of abundances of the adult and juvenile components for
grunts, snappers, groupers, and seagrass-associated taxa with the areal coverage
of seagrass within 1 km of each study reef at the 22 study reefs, sampled in 2002
in St. John, U .S. V irgin Islands. ......................................................................... 75

3-6 Simple linear regression analyses of cumulative species richness of MIFs,
haemulids, epinephelids, lutjanids and within resident, mobile and transient
mobility guilds the areal coverage of seagrass within 1 km of each study reef as
the independent variable at the 22 study reefs, sampled in 2002 in St. John,US
V irgin Islands. ........................................................................76

3-7 Spearman rank correlations of relationships of each fish parameter and the areal
coverage of seagrass at the 250 m spatial extent for the 8 study reefs in St. John,
sam pled in 2002 and 2003 ........................................................ .............. 77

3-8 Influence of fine-scale (rugosity) and landscape-scale (seagrass) features in
predicting reef fish parameters for the 22 study reefs sampled in 2002...................78

4-1. Reef fish assemblage parameters (n = 30) used as dependent variables in
statistical an aly ses .................................................................................. 1 14

4-2. Study reef name, protection status and reef fish sampling effort for FKNMS
(M ay 2003) and St John, USVI (July-August 2002) ............................................115

4-3. Stepwise regression results indicating the relationship of landscape
configuration and reef fish assemblage structure for the FKNMS and US Virgin
Islands, with the R-square, associated p-value, and explanatory habitat variable. 116

4-4. Stepwise multiple regression results of the relationships of reef context and reef
fish assemblage structure for FKNMS and US Virgin Islands, with the R-square,
p-value, and explanatory variable by location.......................................................117









4-5. Average dissimilarity in reef fish community structure at the trophic and family
level in the FKNMS and US Virgin Islands, and the relative contribution of each
trophic and family category to community dissimilarities using SIMPER
analyses on standardized data. ................................................................... ..... ..118

4-6 T-test comparison results to test for significant differences in abundance of the
30 reef fish parameters in the FKNMS (SPA and reference sites n = 8 reefs) and
the US Virgin Islands (MPA and reference sites n = 22 reefs). ..........................119















LIST OF FIGURES


Figure p

2-1 Location of St. John, US Virgin Islands in the Caribbean basin ...........................40

2-2 Distribution of the 20 study reefs around the island of St. John, USVI. Below
the name of each reef is the number of fish point counts per reef. ..........................41

2-3 PCA plots of the landscape structure of the coral reef environments of the 20
study reefs sampled in 1994 and 2001 in St. John, USVI at the A) 100 m, B) 250
m and C) 500 m spatial extent.......................................... ............................ 42

2-4 Effects of reef configuration on mean fish abundances of A) transient fishes, B)
adult omnivores, and C) adult piscivores for the 1994 (N =14) study reefs in St.
John, U SV I. .......................................... ............................ 43

2-5 Effects of reef context on mean abundance of particular fish groups for the 1994
(N =14) study reefs in St. John, U SVI. .............. ...................... ..................... 44

3-1 Location of the 22 study reefs around the island of St. John, US Virgin Islands
sampled in 2002, with the eight study reefs re-sampled in 2003 indicated in
b o ld ......................................................................................... . 6 6

3-2 The relationship of a) cumulative richness and b) mean species richness with the
areal coverage of seagrass (hectares) at 250 m for the 22 study reefs sampled in
2002 in St. John, U.S. Virgin Islands. The x-axis is logo (x +1) transformed. ......67

3-3 The relationship of mean abundances of a) MIFs, b) haemulids, c) seagrass-
associated taxa and d) lutjanids with the areal coverage of seagrass (hectares)
within 250 m of the 22 study reefs in St. John, U.S. Virgin Islands sampled in
2002. Mean abundances and the areal coverage of seagrass are logo (x +1)
transform ed ............... ......... ..........................................................68

3-4. The relationship of cumulative species richness of a) MIFs, b) haemulids and c)
lutjanids with the areal coverage of seagrass (hectares) within 250 m for 22
study reefs in St. John, U.S. Virgin Islands sampled in 2002. The y-axis is logo
(x+ 1) transform ed ................................... .. .............. ....... ......... 69









3-5 Spearman rank correlations for those reef fish parameters that demonstrated a
consistent relationship with the areal coverage of seagrass habitat between 2002
( ) and 2003 ( ) at the subset of 8 study reefs in St. John, U.S. Virgin
Isla n d s ........................................................................ 7 0

4-1. Location of the 17 study reefs sampled in 2003 in the Florida Keys National
Marine Sanctuary and the 22 study reefs sampled in 2002 in St. John, U.S.
V irgin Islands. ..................................................................... 107

4-2 Simple linear regression results of the effects of reef context (the areal coverage
of pavement and reef habitat logo (x+1)) at the 100-meter spatial scale on mean
abundance of various reef fish parameters at the 17 study reefs sampled in 2003
in the FK N M S. ......................................................................108

4-3 Simple linear regression results of the only relationship that remained consistent,
and significant, across systems (the relationship of patch diversity and piscivore
abundances) in the FKNMS and US Virgin Islands. ...........................................109

4-4 Comparison of the relative proportion of mapped habitat classes within 500 m
of the study reefs in the FKNMS and US Virgin Islands..................................... 110

4-5 Reef fish abundance-seagrass relationships for the US Virgin Islands and the
FK N M S (n = 39). ........................................................ .. ...... ............. 111

4-6 Multidimensional scaling plots of the (A) trophic and (B) family structure of
reef fish communities and the (C) landscape structure of the US Virgin Islands
(squares) and FKNM S (triangles). ............................................... .................. 112

4-7 MDS plots of the (A) trophic and (B) family level reef fish community structure
for reefs in the US Virgin Islands and FKNMS, with reefs classified according
to high (circles), moderate (open triangles) and low seagrass (upside down
closed triangles) areal coverages ......................................................................... 113















Abstract of Dissertation Presented to the Graduate School
of the University of Florida in Partial Fulfillment of the
Requirements for the Degree of Doctor of Philosophy

APPLYING TERRESTRIAL LANDSCAPE ECOLOGY PRINCIPLES TO THE
DESIGN AND MANAGEMENT OF MARINE PROTECTED AREAS IN CORAL
REEF ECOSYSTEMS

By

Linda Erica "Rikki" Grober-Dunsmore

August 2005

Chair: Thomas K. Frazer
Major Department: Fisheries and Aquatic Sciences

Marine protected areas (MPAs) represent a popular, but often controversial,

management option for the conservation of dwindling reef fish populations worldwide.

Questions concerning appropriate design criteria for MPAs lie at the center of the

controversy, and reflect a need to better understand the influence of landscape structure

of coral reef ecosystems (e.g., size, shape, context of habitat patches) on reef fish

assemblage structure. I explored the utility of various landscape metrics in predicting

reef fish assemblage structure and found that reef context explained considerable

variation in the several reef fish parameters. Specifically, I found that particular groups

of fishes were associated with particular types of habitat. Based on these results, I

designed a new study in the US Virgin Islands to determine examine whether functional

habitat linkages between reef and seagrass habitat patches were detectable at a landscape-

scale. Consistent with predictions, entire assemblage level parameters and abundances

and species richness of mobile invertebrate feeders, haemulids, lutjanids, and









epinephelids were each significantly greater at reefs with seagrass within 1 kilometer of

the study reef patch. The generality of reef context as a predictor of reef fish assemblage

structure was then tested in the Florida Keys National Marine Sanctuary. Though reef

context was significant in both systems, the particular habitat type responsible for the reef

fish habitat relationships differed between the coral reef landscapes. Seagrass was a

strong predictor of abundances and species richness of mobile invertebrate feeders,

haemulids, and lutjanids in the US Virgin Islands, but was not a predictor of these same

fishes in Florida. Thus, the processes that structure reef fish communities appear to

respond to variation in the landscape structure of these coral reef environments. These

results are relevant to marine protected areas design, since they suggest that general

design rules do not necessarily apply across systems. Rather, comparative studies are

critical for developing the universal design principles to locate marine protected areas

that meet their conservation and/or fisheries objectives.














CHAPTER 1
INTRODUCTION

Coral reef ecosystems are degrading worldwide with losses of biodiversity,

declines in coral cover, and decreases in the average size and abundances of many coral

reef fishes (Wilkinson 2000), and marine protected areas (MPAs) are gaining in

popularity as the best management option for dealing with these concerns (Allison et al.

1998, Murray et al. 1999). Coral reef ecosystems are heterogeneous landscapes,

comprising topographically-complex, calcium carbonate skeletal structures in which

stony corals provide the major framework (Hallock 1997). Coral reefs are embedded in a

mosaic of different habitat patches (e.g., reef, seagrass, open water, and mangrove forest)

that are connected to one another through the movements of energy, material (e.g., fecal

or detrital matter) and marine organisms (Ogden 1997) such as reef fishes.

Coral reef fishes are recreationally and commercially important components of

coral reef ecosystems. Reef fish communities exist as spatially divided populations that

reside in this mosaic. Connections among local subpopulations are maintained by the

export and import of larvae from other subpopulations; or through the movement of

fishes during ontogeny, foraging, or spawning (Sale 2002). Distribution of reef fish

communities is likely governed by multiple biological and environmental processes that

operate at a variety of spatial and temporal scales including biological processes such as

predation (Hixon and Beets 1989, Hixon 1991), competition (Smith and Tyler 1973),

recruitment limitation (Sale et al. 1984, Doherty and Fowler 1994), and priority effects

(Almany 2003). While studies conducted at a small spatial scale associate various fish









parameters (e.g., fish density and biomass) with reef substratum complexity (Luckhurst

and Luckhurst 1978, Gladfelter et al. 1980, Hixon and Beets 1989) and reef surface area

(Molles 1978, Gladfelter et al. 1980), little research has explored whether these

relationships scale up (but see Ault and Johnson 1998a, Acosta and Roberston 2002,

Christensen et al. 2003, Jeffrey 2004).

Without an understanding of the distribution of reef fish communities at large

spatial scales (> 10's of meters), scientists are ill-equipped to advise resource managers

on decisions that require a large-scale examination (e.g., marine protected areas). Marine

protected areas (MPAs), one of the most highly advocated forms of ecosystem-based

management, can provide a spatial escape for intensely exploited species (Allison et al.

1998, Murray et al. 1999). MPAs constitute a broad spectrum of areas that are afforded

some level of protection for the purpose of managing resources for sustainable use, and

safeguarding ecosystem function and biodiversity (Plan Development Team 1990). Their

potential has been demonstrated both theoretically (Plan Development Team 1990,

Roberts and Polunin 1991, Carr and Reed 1993, Allison et al. 1998) and empirically

(Rakitin and Kramer 1996). They can increase average size, abundance, and biomass of

exploited organisms (see reviews by Rowley 1994, Halpem 2003); and networks of

MPAs can insulate habitats and communities from extractive activities that lead to losses

in biodiversity and changes in species interactions (Tegner and Dayton 2000, Murray et

al. 1999). Because decisions about the placement of reserves are largely political,

scientists have had few opportunities to understand the biological implications of reserve

design (Allison et al. 1998). Moreover, the design of MPAs has typically focused on

single habitats, neglecting associated habitats that may benefit reef fishes during various









stages of their life history (Ogden and Ehrlich 1977, Helfman et al. 1982) and failed to

consider potentially-important functional habitat linkages between various habitat

patches. At present, few quantitative rules exists for design and management of MPAs.

In fact, criteria are broad (e.g., representation, replication). Decisions about the siting,

location, size, and composition of MPAs are sorely needed in many places, yet it is

currently difficult, if not impossible, to predict how alternative spatial arrangement

influence the ability of an MPA to meet its stated conservation and/or fisheries

objectives. Various landscape elements (e.g., the amount of edge habitat, corridor

placement, and landscape connectedness) have considerable influence on the distribution

of terrestrial organisms (Turner et al. 2001). Therefore it is crucial to explore the

relevance of landscape elements in structuring coral reef fish communities as a

prerequisite for designing effective MPAs. Because improperly designed refuges may

provide a false sense of protection, and thereby endanger a fishery (Carr and Reed 1993),

identifying simple metrics useful in predicting reef fish assemblage structure would be

extremely valuable to resource managers. Combining the disciplines of landscape and

coral reef ecology provides a logical starting point for addressing important management

questions relevant to habitat-based conservation of reef fishes.

The discipline of landscape ecology deals with interactions and exchanges across

large areas, relating the structure of an area to its function (Forman and Godron 1986).

By using geo-referenced maps of vegetation, soils, and elevation, terrestrial landscape

ecologists have quantified aspects of spatial patterning using a number of metrics,

including (but not limited to) patch size and shape, and total area of critical habitats

(Turner 1989, Forman 1995). These metrics, calculated statistics of landscape pattern









(Frohn 1998), can be used to predict the outcomes of ecological processes such as

dispersal success (Gustafson and Gardner 1996, Schumaker 1996) and population

dynamics such as density (McGarigal and McComb 1995), distribution (With and Crist

1995), community structure (Noss 1983), and survival probability (Fahrig 1997). In fact,

large-scale metrics of habitat diversity have been successfully used to predict total

species richness and abundance of birds, butterflies, and reptiles (Rafe et al. 1985,

Rosenzweig 1995, Ricklefs and Lovette 1999). The challenge for scientists is to identify

those landscape-scale metrics that may serve as proxies for resource managers of areas

with high species diversity and abundance in the coral reef landscape.

Particular habitats such as seagrass communities may play a key role in structuring

reef fish communities. Seagrass communities serve as refuge habitat for small fishes and

benthic invertebrates, and may be beneficial to the settlement, survivorship, and growth

for a variety of fishes and invertebrates that spend their adult life on the reef (Parrish

1989, Baelde 1990, Ogden 1997, Nagelkerken et al. 2000). Seagrass beds are some of

the most productive ecosystems of the world (Zieman and Wetzel 1980, Duarte and

Chiscano 1999), often forming a dense and extensive below-ground network of roots and

rhizomes that support a structurally-complex system of short shoots. A diverse epibiont

community attaches to the seagrass blades (Fong et al. 2000), and infaunal organisms live

in the sediment of the seagrass, acting to enrich or stabilize the substrate (Suchanek 1983,

Peterson and Heck 1999). Many organisms live in seagrass patches including mollusks

(Irlandi et al. 1999), crustaceans (Arrivillaga and Baltz 1999), and fishes (Ogden 1997).

High densities and biomass of economically-important reef fishes (such as snappers,

grunts, and groupers) have been attributed to the availability of food resources in









surrounding seagrass habitat (Randall 1963), perhaps due to movement patterns of these

exploited fish (Tulevech and Recksieck 1994, Burke 1995). No studies, however, have

quantified the benefits of potentially critical habitat linkages between reef and seagrass

(exception: Wolff 1996) or applied recently developed technologies such as geographic

information systems, to determine whether the influences of such habitat linkages are

detectable at a landscape-scale. If benefits of such habitat linkages can be detected at a

large spatial scale (100's of meters), identifying and then subsequently protecting these

habitat linkages (which has proven extremely valuable in terrestrial conservation) (Noss

1983, Forman 1995, Turner et al. 2001) may be feasible for delineating boundaries of

MPAs in coral reef ecosystems.

Experience applying a landscape ecology approach in terrestrial systems provides a

framework for addressing a number of relevant resource management questions that may

prove valuable in marine systems.

* What are the relationships among coral reef landscape structure and reef fish
assemblage structure?;

* Can functional habitat linkages be identified at a landscape-scale, and is it possible
to measure and quantify the potential consequences of these habitat linkages?;

* What is the appropriate spatial scale for addressing these questions?;

* Can faunal-habitat relationships detected in one coral reef landscape be generalized
to another?

To address these questions, a hypothetico-deductive study (Platt 1964, Peters 1991)

was designed to explore the utility of terrestrial landscape ecology principles in the US

Virgin Islands and the Florida Key National Marine Sanctuary (FKNMS) coral reef

landscapes. Because little prior research had been conducted at large spatial scales in

coral reef systems (exceptions: Appeldoorn et al. 2003, Kendall et al. 2003, Christensen









et al. 2003, Jeffrey 2004), it was necessary to start by exploring relationships of coral reef

landscape structure with reef fish assemblage structure (Chapter 2). In the inductive

stage of my study, data were collected from the US Virgin Islands, analyzed statistically,

interpreted without having a prior specific hypothesis (although generally I expected

landscape structure would be correlated with reef fish assemblage structure), and results

were used to generate hypotheses tested later. Specifically, I explored those measures of

landscape structure that have proven valuable in terrestrial ecosystems for predicting

areas of high abundance and species richness.

The preceding exploratory approach, however, does not eliminate the risk of

detecting patterns that are not biologically relevant or the risk of missing significant

relationships that are (Bissonette and Storch 2003). Because organisms with different

habitat requirements, feeding behaviors, and mobility can respond to the landscape

differently (Turner et al. 2001, Sisk et al. 1997), analyses must be conducted with

consideration for the natural history and ecology of the organisms of interest (Bissonette

and Storch 2003). Therefore, while the entire shallow-water reef fish community was of

interest to me in this study, this diverse assemblage of fishes (140 different taxa) was sub-

divided into groups of species that share a common set of life-history traits,

morphological or behavioral attributes, or ecological functions. The response of these

functional fish groups trophicc and mobility guilds, taxonomic groups, and by life history

stage) to landscape features were examined later.

Because pattern exists at every spatial scale (Wiens 1989), it is necessary to link the

organisms, species, or the processes being considered to the scales appropriate to the

specific questions of interest (Bissonette and Storch 2003). The process for finding the









relevant scale is not well understood (Bissonette and Storch 2003), but generally

researchers recommend analyzing data at multiple spatial extents to identify concordance

with response variables using spatial statistics and multiple regression procedures

(Pearson 1993, Pearson et al. 1995, Pedlar et al. 1997). Particularly in spatially

heterogeneous systems, where little to no work has been conducted at a landscape-scale,

faunal-habitat relationships should be explored at multiple spatial extents. Therefore,

four spatial extents (100 m, 250 m, 500 m, and 1 km) were explored in all subsequent

chapters of this dissertation.

It is also valuable to understand whether the distribution of organisms at a given

location is explained by characteristics of the immediate locale (fine-scale within patch

characteristics) or by the attributes of the surrounding landscape (landscape

characteristics), so resource managers can evaluate the trade-offs of collecting patch-level

or landscape-level information. For example, Pearson (1993) found that some birds

responded only to characteristics of the local habitat (vegetation characteristics such as

height, density, and species composition), while other bird species responded only to

landscape context (amount of each habitat surrounding study areas). Thus, detailed

within-patch surveys may be required to predict the presence of some species, whereas

remotely-sensed surveys may be sufficient to detect the presence of others. In coral reef

ecosystems, various within-patch measures of coral reef patch quality have been shown

to influence reef fish abundance and diversity. These patch-level measures include coral

cover (Bell and Galzin 1984) and topographic complexity (Luckhurst and Luckhurst

1978, Hixon and Beets 1989, Friedlander and Parrish 1998), therefore, the fine-scale

measure of rugosity (Luckhurst and Luckhurst 1978) and various landscape-scale metrics









of habitat were evaluated to determine their relative influence in structuring reef fish

communities for all analyses.

As progress in science is ideally made by the sequential development of hypotheses

and the execution of experiments designed to test these hypotheses (Platt 1964, Quinn

and Dunham 1983), ecologically-meaningful, and significant findings from the

exploratory analyses were used to derive testable hypotheses for further study (Chapter

3). Of particular promise were metrics of reef context, which quantify the spatial

arrangement and composition of surrounding habitat patches. A landscape patch, by

definition, is bounded by something else, and these adjacent or proximal habitat patches

can strongly influence organisms within that patch (Turner et al. 2001).

Correspondingly, scientists increasingly recognize that continental reserves are not

islands surrounded by a neutral sea (Janzen 1986); rather isolating a reserve can lead to

ecosystem degeneration and the extent and rapidity of this degeneration can depend upon

the ecological condition of adjacent habitat patches (e.g., Kushlan 1979). Consequently,

I examined how reef context influenced reef fish community structure. Specifically, I

tested hypotheses that entire assemblage level parameters, and abundances and species

richness within trophic and taxonomic groupings, would be higher at reefs with seagrass

within 1 kilometer of the study reef patch.

Although considerable attention has been dedicated to the temporal variability

inherent to natural communities, surprisingly little research has addressed the temporal

consistency of faunal-habitat relationships, particularly those that occur at a landscape

scale (Turner et al. 2001, Bissonette and Storch 2003). Many landscape studies have

been discrete sampling events, and have failed to consider temporal variation in resources









or reproductive opportunities (Turner et al. 2001). Coral reef fish communities are

notoriously dynamic, both spatially and temporally (Sale et al. 1984, Sale 2002),

therefore, their temporal dynamics represent an important ecological dimension.

Temporal variations in fish abundance may be induced by reproductive movements

(Colin et al. 1997), ontogenetic shifts (Appeldoorn et al. 1997), feeding migrations

(Ogden and Zieman 1977, Ogden and Quinn 1984), spatially segregated foraging and

resting locations (Meyer et al. 2000), and spatial heterogeneity within and among habitat

patches. If landscape metrics of the coral reef landscape are to prove valuable in

understanding the distribution and abundance of coral reef fishes, a more thorough

understanding of the temporal consistency of reef fish-habitat relationships is critical. To

begin to meet this need, a portion of this study (Chapter 3) was conducted over 2 years.

Detection of faunal-habitat relationships in one system does not necessarily imply

that resource managers can expect organisms in another system to respond to analogous

landscape features in the same manner. Because each landscape is unique, the size and

distribution of habitat differs, and may putatively exert different landscape-specific

constraints (Bissonette and Storch 2003) on species richness, abundance, community

structure, recruitment, and movement. If the constraints of each landscape result in

qualitatively different responses of the reef fishes without apparent thresholds, then we

may have no hope of developing a predictive theory for forecasting the fish assemblage

structure of a given reef patch. Thus, in order to determine the generality of reef fish-

habitat relationships detected in the insular US Virgin Islands to other Caribbean coral

reef ecosystems, the study was replicated (Chapter 4) spatially in the Florida Keys

National Marine Sanctuary.









The approach selected in this study is beneficial in that it addressed many of the

short-comings of terrestrial landscape studies (Turner et al. 2001) by replicating the study

both in time and space. In addition, multiple aspects of this research remain consistent

throughout the dissertation, allowing insight into the generality and reliably of various

measures in other systems and over time. For example, each chapter examines the

influence of landscape pattern to the entire reef fish community, and then focuses on

trophic and mobility guilds and taxonomic groupings of fishes. Furthermore, because

there is no single correct spatial scale to describe a system (Levin 1992), a multi-scalar

approach was adopted. Every study analysed relationships at multiple spatial scales

including the fine-scale within-patch characteristics and explored the strength of

relationships at various landscape scales (100 m 1 km). While as a discipline,

terrestrial landscape ecology has progressed from purely non-quantitative descriptive

studies (Wiens 1992) to increasing emphasis on spatial statistics, modeling, and

experimental design (Hobbs and Norton 1996), this dissertation also progresses from the

exploratory to the testing of specific hypotheses. These hypotheses are tested over time

and space in a robust study design, thus insights derived from this dissertation provide a

foundation for applying the principles of landscape ecology to tropical marine systems,

and improve our understanding of the relationships of reef fish communities to large-

scale habitat features.














CHAPTER 2
INFLUENCE OF LANDSCAPE STRUCTURE ON REEF FISH ASSEMBLAGES

Marine protected areas represent a popular, but often controversial management

option for the conservation of dwindling reef fish populations worldwide. Questions

concerning appropriate design criteria for marine protected areas lie at the center of the

controversy, and reflect a need to better understand the influence of landscape structure

of coral reef ecosystems (e.g., size, shape, and context of habitat patches) on reef fish

assemblage structure. Herein, I investigated the relationships between landscape

structure and reef fish assemblage structure at 20 study reefs around the island of St.

John, US Virgin Islands. Various measures of landscape structure were calculated and

transformed into a reduced set of composite indices using principal component analyses

(PCA) to synthesize data on the spatial patterning of the study reefs. However,

composite indices (i.e. measures of habitat diversity) were not particularly informative

for predicting reef fish assemblage structure. Rather, relationships were interpreted more

easily when functional groups of fishes were related to individual habitat features. In

particular, reef context was strongly associated with multiple reef fish parameters (e.g.,

abundances within trophic guilds and taxonomic groups). Fishes responded to benthic

structure at multiple spatial scales, with each fish group correlated to a unique suite of

variables. Accordingly, future experiments should be designed to test functional

relationships based on the ecology of the organisms of interest. My study illustrates

promise in applying a landscape ecology approach to coral reef ecosystems, and provides









an empirical basis to further test the influence of specific habitat features in structuring

reef fish communities.

Introduction

The management of tropical marine environments calls for interdisciplinary studies

and innovative methodologies that consider processes occurring over broad spatial scales

(Allison et al. 1998). Landscape ecology is interdisciplinary by nature, with an

appropriate focus on broad-scale patterns and ecological processes (Forman and Godron

1986). A landscape generally refers to a heterogeneous area composed of local

interacting ecosystems (Forman 1995) made up of homogenous units, called habitat

patches. Landscape structure describes the composition and spatial arrangement of

habitat patches (Forman and Godron 1986), and has been quantified using a number of

metrics (O'Neill et al. 1988) including composite indices (e.g., habitat diversity, principal

components), measures of configuration (e.g., patch size), and measures of context

(composition of surrounding habitat patches) (Turner 1989). The use of such metrics,

derived largely from island biogeography theory (MacArthur and Wilson 1967),

metapopulation theory (Hanski 1999), and patch dynamics (Pickett and White 1985) has

improved our understanding of how landscape features influence terrestrial communities

(Turner 1989, Gardner and O'Neill 1991). Because of its focus on broad and multiple

spatial scales and entire ecosystems, landscape ecology has proven extremely valuable in

addressing management problems in terrestrial systems (e.g., reserve design) (Noss 1983,

Forman 1995). A landscape ecology approach to the study of coral reef fishes, however,

has received little attention, until recently (Kendall et al. 2003, 2004, Jeffrey 2004).

Our understanding of the dynamics of reef fish assemblages has been largely

derived from studies conducted at fine spatial scales (1 m2 plots) (Williams 1980, Sale et









al. 1994, Pittman and McAlpine 2003), limiting our ability to predict the effects of large-

scale features on reef fishes (Sale 2002). Fine-scale measures of topographic complexity

(Hixon and Beets 1989, Friedlander and Parrish 1998), hole size (Friedlander and Parrish

1998), and coral cover (Bell and Galzin 1984) can influence reef fish assemblage

structure. The few existing large-scale studies have examined relatively gross

characteristics such as latitudinal gradients (Ebeling and Hixon 1991) and coral reef

zonation (Williams 1991); features that are not particularly useful for selecting specific

reef areas as candidates for protection. Thus, it is unclear whether findings from fine-

scale studies can be extrapolated to large-scale resource management concerns.

Understanding functional relationships between landscape structure and reef fish

distribution at a broad spatial scale may therefore be useful for delineating the boundaries

of MPAs (Christensen et al. 2003), since coral reef ecosystems exist as a complex mosaic

of habitat patches (i.e. reefs, seagrass patches, and mangrove stands), and are therefore

ideally suited for a landscape ecology approach.

The purpose of my study was to determine whether commonly-used terrestrial

metrics can be quantified for coral reef environments, and to determine whether these

metrics might be used to predict reef fish assemblage structure. Toward this end, a suite

of landscape metrics was explored using multivariate statistics. Through a variety of

procedures to eliminate redundancy and autocorrelation, I developed composite indices to

synthesize data on the spatial patterning of the reef study sites. I then examined the

utility of these composite indices to predict reefs that have relatively high reef fish

species diversity and abundance. In addition, I explored individual habitat features (reef

configuration and reef context) separately, and examined the relative importance of fine-









and landscape-scale habitat measures on reef fish assemblage structure. Configuration

was selected because measures of mean patch size, shape, and arrangement are frequently

associated with species abundance and diversity (Robinson et al. 1995, Villard et al.

1999); and protected area configuration can influence organisms within and outside their

boundaries (Diamond 1975, Sisk et al. 1997, Mazerolle and Villard 1999). Context was

selected because of the increasing recognition of its role in sustaining species targeted for

conservation in terrestrial systems (Mladenoff et al. 1995, Robinson et al. 1995, Sisk et

al. 1997). The relative influence of fine-scale (rugosity) and landscape-scale habitat

characteristics on fish assemblage structure was explored because knowing the

importance of features at each spatial scale can save precious resources for resource

managers (Mazerolle and Villard 1999).

Study Area

Coral reefs around the island of St. John, US Virgin Islands (Figure 2-1) were

selected for study, because benthic habitat were readily available (Kendall et al. 2001).

Habitat maps (digitized from aerial photographs taken at an altitude of 5000 feet in

1999), were classified by visual interpretation by NOAA, using 26 discrete and non-

overlapping habitat classes, with a minimum mapping unit of 1 acre (Kendall et al. 2001).

Most sampling sites were located on the lower fore reef of fringing and patch reefs,

dominated by Montastraea annularis or mixed corals (8-30% living cover; pers. obs.),

although several were dominated by old Acroporapalmata framework (5-10 % living

cover; pers. obs.). Study reefs occurred in water depths between 5 and 15 m.

Methods

Twenty reefs were sampled: 14 reefs in 1994 and 6 in 2001 (Figure 2-2). Reefs

were selected from an existing fish database as representative locations that varied with









respect to landscape features, yet relatively similar depth, reef morphology, and coral

cover. For exploratory analyses, the 1994 and 2001 datasets were combined. Reefs

sampled in 2001 were included to expand gradients in several habitat parameters of

interest. To investigate specific functional relationships (e.g., configuration, context),

only the 1994 dataset was used to reduce potential temporal variability due to changes in

fishing pressure and storm damage.

Reef fish Sampling

Fish sampling was conducted within reef habitat only. Reef-associated fishes were

sampled over a 10-day period in July 1994 (Beets and Friedlander 1994), and over a

5-day period in July 2001. The number of fish point counts per reef were determined

based on reef size following Monte Carlo simulation, and ranged from 8-20 point counts

per reef (Figure 2-2). In 1994, a modified Bohnsack and Bannerot (1986) point count

method (a reduction in the sample radius from 7.5 m to 5 m) was used, whereas the

original point count method was used in 2001. Mean species richness refers to the mean

number of species observed per point count per replicate reef, whereas cumulative

richness refers to the total number of species observed during all point counts at a reef.

Abundance refers to the mean number of reef-associated fishes observed per point count

per replicate reef. Two species were eliminated from abundance analyses, since these

tended to overwhelm abundance estimates and are difficult to count accurately--Jenkinsia

spp. (herring) and Coryphopteruspersonatus (masked/glass goby). Randall (1967) and

Fish Base (Froese and Pauly 2002) were used as references to classify all fishes by

trophic guild: piscivore, herbivore, mobile invertebrate feeder (MIF), sessile invertebrate

feeder (SIF), planktivore, or omnivore (Appendix A). To the degree that the ecology of

each species is known, each fish was classified into mobility guilds: resident, mobile or









transient (Appendix A). Resident species are sedentary and site-attached, and do not

typically move from their primary reef patch. Mobile species are those that have

restricted movements and may roam from the primary reef patch. Transient species are

vagile, and can range on the scale of kilometers. Taxonomic groups of commercially and

ecologically-important fishes were analyzed separately (e.g., haemulids, lutjanids, and

scarids). Fishes were further subdivided into juvenile and adult categories, based on

length of maturity where possible (Froese and Pauly 2002), to examine the influence of

life-history stage on functional relationships. This resulted in 30 reef fish assemblage

parameters (Table 2-1).

Habitat Sampling

The fine-scale measure of rugosity was obtained by running an underwater tape

measure as closely as possible over the contour of the substratum. For each reef, 10

rugosity samples were collected along 10-m transects. The resultant mean value was

used for subsequent analyses. Because these reefs are natural habitat patches,

microhabitat variation that was not quantified, likely exists.

The original map classification scheme (Kendall et al. 2001) was condensed from

26 to 9 distinct and non-overlapping habitat classes (i.e. mud, mangrove, sand, reef,

pavement, bedrock, seagrass, macroalgae, and deep unknown). Deep unknown was

typically a deep (> 16m) soft-bottom habitat (R. Grober-Dunsmore, unpubl. data). This

reduced set of nine habitat classes was selected based on terrestrial studies (which

frequently use 5-10 habitat classes for resource management purposes; Turner et al.

2001), and to simplify results for resource managers. Haphazard groundtruthing was

conducted in each habitat polygon within 100 m of each study reef. Percent cover of

benthic invertebrates and substrate types were estimated using lm2 quadrats.









Fourteen metrics, used to quantify various aspects of the configuration and

context of the study reefs (Table 2-2), were calculated with ArcView 3.2 (ESRI 1996). A

single value for each (reef) patch metric was calculated (n = 3). Each landscape metric

was calculated at three spatial extents; 100 m, 250 m, and 500 m from the leading edge of

each reef. These extents were selected to represent a range of potential importance based

on the known natural history of reef fishes. Because the area of deep unknown increased

considerably beyond 500 m, the 1 km spatial extent was not included, resulting in 36

metrics that were correlated to reef fish parameters (Table 2-3).

Most sample reefs were determined to be subsections of larger mapped polygons,

thus each reef polygon was slightly modified in Arc View 3.2 (ESRI 1996) to reflect the

standardized 25,000-m2 subsections where fish data were collected. A separate

heterogeneity study (Grober-Dunsmore et al. 2004) determined that for most fish

parameters, no significant differences existed between these reef sections and the entire

mapped polygons. These results, and the fact that slightly-modified polygons were used

for all calculations, justify the use of these reefs for exploratory purposes.

Data Analysis

To reduce the 36 landscape metrics into a more parsimonious dataset of

composite indices that capture the wealth of information contained within the original

dataset, a combination of techniques was applied. (1) Pearson product-moment

correlations (Ppmc) between each pair of metrics, and (2) principal component analysis

(PCA) using a correlation matrix. Each spatial extent was examined separately and

results were compared across extents. Ppmc was applied sequentially by examining

significant pair-wise correlations (Sokal and Rohlf 1995) to reduce the number of

variables to a 3:1 ratio (observations to variables), which is required for PCA (McGarigal









et al. 2000). The choice of an index within a group of redundant metrics was determined

by selecting ecologically-meaningful metrics and eliminating variables that failed

normality tests. Ppmc reduced the 14 metrics to 8, and PCA was subsequently applied to

further synthesize these into a smaller set of linear combinations (components) of the

original variables. Loadings on original variables were used for interpretation.

Principal component plots of these landscape metrics were used to organize the

sampling entities (reefs) in multivariate space. Reefs were classified into one of three

categories based on our pre-existing knowledge of the surrounding coral reef

environments. PCA plots were then displayed at each spatial extent to determine whether

characterization of study reefs using landscape metrics corresponded to our local

knowledge and groundtruthing data, allowing us to assess the broad accuracy of remotely

sensed benthic habitat maps.

To explore the strength and nature of the relationships between landscape

structure and fish assemblage parameters, stepwise multiple regression analyses using

significant principal components as the independent variables, for each of the thirty reef

fish parameters, was conducted. To control family-wise error rate for multiple

correlations, sequential Dunn-Sidak Bonferroni corrections were applied using the

number of reef fish parameters (n = 30) tested (Sokal and Rohlf 1995) for all subsequent

regressions.

Specific functional relationships were examined separately: 1) reef configuration,

2) reef context, and 3) the relative influence of fine and landscape-scale habitat

parameters, using the 1994 reef fish dataset only (n = 14). Landscape variables were

selected based on Ppmc results (Sokal and Rohlf 1995) though several variables that were









eliminated from PCA were included to verify their potential importance. Only those that

met assumptions of statistical independence were tested in a given model. Reef

configuration variables were perimeter to area ratio (P: A) of each reef, reef size, and the

number of habitat patches within 100 m. Reef context variables were surrounding habitat

diversity and the areal coverage of reef, bedrock, seagrass and deepwater within 100 m.

Fine-scale and landscape-scale variables were rugosity and the areal coverage of

deepwater, seagrass and reef within 100 m. To optimize model performance and reduce

potential effects due to multicollinearity, a series of diagnostic tests were used:

1) Akaike's Information Criterion (Akaike 1974), 2) leverage effects plots, 3) Durbin-

Watson statistic, and 4) condition number (Belsley et al. 1980) for every stepwise

regression analysis. Simple linear regressions were created to determine the stability of

models using residual plots and residual normality plots (Sokal and Rohlf 1995).

Where necessary, reef fish and habitat data were logo (x + 1) transformed to

improve normality and data were tested using Shapiro-Wilks statistic (Sokal and Rohlf

1995). All statistical analyses were conducted with JMP 8.01 (SAS 2003). Statistical

significance was accepted at the p < 0.05, unless otherwise noted.

Results

Landscape Structure

Configuration and context of study reefs varied widely. Most metrics had

coefficients of variation > 50 % of the mean, indicating that gradients in many aspects of

the landscape were represented (Table 2-3). Of fourteen initial landscape variables,

Ppmc resulted in 8 remaining landscape metrics (Table 2-4). PCA results were consistent

across spatial extents (i.e. 100 m, 250 m, 500 m). In fact, the variance explained, the

eigenvalues and the distribution of loadings were remarkably comparable. However,









interpretation was easiest at the 100 m spatial extent since the proportion of area

classified as deep unknown was minimized. Thus, a single spatial extent (100 m) was

selected for further analyses, using the remaining 8 landscape variables: number of

habitat patches, reef size, habitat diversity, area of deep unknown, pavement, reef, sand

and seagrass habitat.

PCA of these 8 landscape metrics at 100 m revealed four dominant components of

variation based on retention of eigenvalues greater than the average, i.e. X > 1 (Jackson

1993). These components explained approximately 80 % of the total variance of the

original landscape variables (Table 2-5). However, landscape structure was not

adequately represented by a single or even a few gradients. Final communalities

indicated that most of the residual configuration indices were well accounted for by the

four components, with no notable exceptions.

Principal component plots using landscape metrics generally corresponded with my

pre-existing knowledge and groundtruthing of the local environments of these study reefs

(Figure 3), though there were several outliers. PCA plots were also in general

concordance across spatial extents (100 m, 250 m, and 500 m), although the 100 spatial

extent resulted in strongest clustering of reefs in concordance with my local knowledge.

Thus, PCA plots illustrated that benthic habitat maps were capable of differentiating

among reef types (Figure 2-3).

Reef Fish Assemblage Structure

A total of 57,002 fishes representing 171 different species were recorded during

341 censuses at the 20 study reefs.

Principal components proved to be useful in explaining only a limited number of

reef fish assemblage parameters. Both measures of species richness (i.e. mean richness









and cumulative richness) were marginally correlated (21% and 26%) with PC4, a positive

gradient of seagrass (Table 2-6). Fifty-three percent of the variation in herbivore

abundance was explained by PC2 and PC3, positive gradients of sand and habitat

diversity. Acanthurids (a major component of the herbivore guild) were also positively

correlated to PC2 (Table 2-6). Forty-three percent of haemulid abundance was

negatively correlated to PC 1, and 46 % of lutj anid abundance was negatively correlated

to PC1 and positively related to PC2 (Table 2-6).

Configuration was generally a poor predictor of reef fish assemblage structure.

There were a few exceptions. Seventy-four percent of the variation in the abundance of

transient fishes (e.g., jacks) was explained by P: A of each reef and the number of habitat

patches (Table 2-7, Figure 2-4). Abundances of two other trophic guilds (piscivores and

omnivores) and three taxonomic groups (pomacentrids, acanthurids, and pomacanthids)

were marginally correlated to P: A (Table 2-7, Figure 2-4). Importantly, examination of

regression plots revealed the influence of single points, and residual plots revealed that

several relationships exhibited heteroscedascity, thus calling into question the stability of

these relationships (Sokal and Rohlf 1995). No reef fish assemblage parameter was

correlated with reef size.

Reef context was correlated with thirteen of thirty possible reef fish assemblage

parameters (Table 2-8). Species richness was positively correlated with the areal

coverage of seagrass (Table 2-8). Several ecologically-relevant relationships between

specific habitat types and abundances within trophic and taxonomic groups were also

evident. Adult mobile invertebrate feeders (i.e. 64 species) were positively correlated

with the areal coverage of seagrass (R2 = 0.33) and adult piscivores were positively









correlated with the areal coverage of reef within 100 m (R2 = 0.51) (Table 2-8, Figure 2-

5). Several taxonomic groups were predicted, based on their life history, to be correlated

with a particular habitat type, e.g., adult haemulids and lutjanids with seagrass and adult

serranids with reef habitat. Simple linear regressions, based on stepwise results, were

generally consistent with these predictions (Figure 2-5). Fifty-three percent of the

observed variation in the mean abundance of adult haemulids, and 68 % of the variation

in the mean abundance of adult lutjanids was explained by seagrass coverage (Figure 2-

5). Juveniles of several groups of fishes were correlated with deep unknown habitat

(Table 2-8). Subsequent examination of residuals plots and residual normality plots

indicated that most relationships were stable; those that were not were eliminated from

the results reported here.

Discussion

The coral reef landscape variables were successfully reduced into four principal

components, thereby synthesizing the wealth of information contained within the benthic

dataset (O'Neill et al. 1988, Riitters et al. 1995). When plotted, these components were

able to differentiate between reefs of varying configuration and habitat composition, thus

benthic habitat maps clearly appear useful in describing and quantifying coral reef

landscapes. Clustering of reefs along PCA axes remained consistent across multiple

spatial extents (100, 250, and 500 m), a result that represents an important contribution to

the application of landscape ecology principle to coral reef ecosystems, since it is critical

to determine the relevant scale of analyses for landscape studies (Gardner and O'Neill

1991). Because with increasing spatial extent, the total area of deep unknown increased,

I recommend that the 100 m or 250 m spatial extent is the most informative

characterization of the spatial patterning of the coral reef landscape, at least in this









system. Selection of scale is a defining challenge since different patterns emerge at

different scale (Wiens 1989), and the appropriate spatial extent may depend on the

benthic habitat maps available and the question of interest.

Ecologically-meaningful interpretation of the principal components proved

difficult, because loadings were distributed across many variables, and no single

component accounted for > 26 % of the variability contained in the original dataset.

These findings differ from many terrestrial studies (McGarigal and McComb 1995,

Riitters et al. 1995), where distinct elements of the landscape are often described by

components that represent habitat complexity (Riitters et al. 1995), fragmentation

(Andren 1994), or patch shape (McGarigal and McComb 1995). This analysis suggests

that complex composite indices are less informative than individual spatial features for

characterizing the coral reef landscape at the individual reef scale.

In general, principal components were poor predictors of reef fish assemblage

structure since most relationships were more easily interpreted using individual habitat

features. For instance, species richness exhibited a marginal association with PC1 and

PC4, which represented gradients in the areal coverage of reef and seagrass area, habitats

considered critical for many reef fish species (Ogden and Zieman 1977, Sale 2002).

Likewise, the association of herbivores with PC2 (a gradient of the areal coverage of

shallow sand and seagrass habitat), may indicate the availability of shallow foraging

habitat, where sufficient light is available for photosynthesis of their primary food source,

algae. When testing new approaches in a new system, it is critical to determine the

appropriate measures for understanding the spatial distribution of organisms (Turner

1989, Wiens 1992), thus this negative result may help guide future coral reef studies.









Several factors may explain the inability of principal components to predict reef

fishes. These components may contain too much information to be germane to reef

fishes since fish may not respond to multiple habitat parameters. Rather, interpretation

suggests that specific fish groups respond to specific habitat features. Additionally, it

appears that all species may not conform to the same landscape pattern, as in terrestrial

systems (Mladenoff et al. 1995, Lindenmayer et al. 2003), but that each organism may

respond to specific features at particular spatial scales.

The other composite index, habitat diversity, was also not a good predictor of reef

fish diversity and abundance, which is contrary to predictions based on terrestrial

research (Rafe et al. 1985, Ricklefs and Lovette 1999). These findings may indicate that

this is not an appropriate measure of habitat diversity, since relationships can heavily

depend upon the specific definition (e.g., elevation, vegetation structure) of habitat

diversity (Rafe et al. 1985, Turner 1989). In Palau, another habitat diversity measure

also failed to predict species diversity and species richness (Donaldson 2002) of reef

ifshes though Jeffrey (2004) found that measures of habitat richness and diversity were

correlated (both positively and negatively) with several measures of trophic composition

and occurrence of several species of fishes. I found frequent negative associations of

individual reef fish parameters with habitat diversity, which may suggest that specific

habitat types are likely to be better predictors of assemblage structure than habitat

diversity per se. These findings lead us to concur with terrestrial studies that challenge

the effectiveness of generic landscape indices (i.e., principal components and habitat

diversity indices) to design protected areas (Lindenmayer et al. 2003) at the scale of

individual reefs. Relationships in this study were better understood by examining









specific habitat features, therefore future studies may need to be designed to examine

specific functional relationships between particular groups of fishes and specific habitat

features.

Although useful in some terrestrial systems (Andren 1994), but highly variable

and weak in others (Trzcinski et al. 1999), configuration measures were generally not

effective in predicting reef fish assemblage structure. These findings corroborate those of

Pittman et al. (2004), which revealed that configuration explained less of the variation in

the spatial distribution of fishes than habitat composition. There were a few potentially,

ecologically-relevant relationships. The strong, positive association of reef P: A with

abundances of transients (e.g., jacks, yellowtail snapper) may reflect their foraging

behavior along reef edges. The negative association of reef P: A with adult piscivores

was surprising. Perhaps larger transient predators prey on smaller piscivorous fishes

along the edge, reducing their abundance. In addition, fish traps are typically set along

reef edges in St. John, and other island locations, thus piscivores along reef edges may be

more susceptible to fishing mortality, which may explain, in part, these findings.

Surprisingly, reef size was not positively correlated with any reef fish parameter.

These findings contrasts with terrestrial (Diamond 1975), small-scale patch reef (Molles

1978, Bohnsack and Talbot 1980, Sale et al. 1994) and seagrass studies (Irlandi et al.

1999, Hovel and Lipcius 2001), but may be a consequence of the limited gradient in reef

size in this study (though the coefficient of variation was > 67 % of the mean). It is

possible that beyond a minimum reef size (which these reefs may exceed), the structure

of reef fish communities may be mediated by other factors such as reef context (see

below), physical disturbance (Syms 1998), larval supply (Sale et al. 1984, Doherty and









Fowler 1994), and/or predation (Hixon and Beets 1989). Such scale effects have been

demonstrated in reef communities; e.g., the paradigm of locally-controlled recruitment

from a superabundant pool of larval fish, developed at the scale of local populations

(Smith and Tyler 1972) has been shown to be invalid at the scale of the whole reef

population (Sale et al. 1984).

Reef context appears to be an important determinant of reef fish assemblage

structure, corroborating findings involving multiple taxa in terrestrial (McGarigal and

McComb 1995, Mazerolle and Villard 1999, Trzcinski et al. 1999), coral reef (Kendall et

al. 2003) and seagrass systems (McAlpine et al. 2004). In particular, the areal coverage

of seagrass, an important nursery and larval settlement habitat (Shulman and Ogden

1987, Ogden and Zieman 1977) and foraging area for some fishes (Randall 1967) was

strongly associated with entire assemblage parameters (e.g., cumulative species richness).

Seagrass habitat may contribute to higher species richness as a result of nutrient transfer

and movement of invertebrates and energy from highly productive seagrass to adjacent

reef habitat (Duarte 2000). For instance, Tektite and Yawzi reef, structurally complex

reefs with the highest mean species richness values, are located within a bay with dense

Thalassia testudinum. Higher species richness was also detected in mangroves adjacent

to continuous seagrass in Australia (Pittman et al. 2004), and at reefs proximal to nursery

habitats (i.e. seagrass) in Colombia using large-scale habitat maps habitats (Appeldoorn

et al. 2003), although in coral reef systems researchers were not able to eliminate

confounding factors of near shore-offshore effects nor separate independent contributions

of other soft-bottom habitats.









Reef context was also strongly associated with abundances within specific trophic

guilds and taxonomic groupings. As expected based on terrestrial (Turner 1989, Sisk

1997) and marine research (Pittman and McAlpine 2003, Pittman et al. 2004),

relationships that met the model selection criteria were consistent with the ecology of

each particular fish group. For example, the positive relationship of MIF abundances

with seagrass is consistent with the foraging behavior of species in this trophic guild

(e.g., taxa within mullidae, haemulidae, and lutjanidae). The relationship for haemulids

and seagrass was even stronger, which is expected since some haemulids forage off-reef

in seagrass nocturnally (Ogden and Quinn 1984). Common piscivorous fishes, which

may forage preferentially in reef habitat, such as Aulostomus maculatus, Carynx sp.,

Scomberomorous regalis, Synodus intermedius were more abundant where there were

large areas of reef habitat (e.g., Eagle Shoals and Tektite). The positive association of

juvenile omnivores with deep water habitat is consistent with the functional role of deep

water as a source of ichthyoplankton. Several species of omnivores (e.g., apogonids,

blenniids) are fairly non-mobile as larvae, and are thought to recruit directly to reef

substrate from the plankton. A direct test of this hypothesis, however, would be required

to determine whether deep water enhances planktonic larval delivery at the reef scale.

These findings suggest that a landscape ecology approach can be valuable for identifying

functional linkages between organisms and their coral reef habitats at a scale appropriate

for resource management decisions.

The fine-scale measure of rugosity was of limited value in predicting reef fish

assemblage structure. The exception was for highly site-attached fishes, e.g., omnivores

which were primarily blenniids, gobiids, and pomacentrids. The inability of rugosity to









predict reef fish assemblage structure, though contrary to previous small-scale research

(Hixon and Beets 1989), may indicate the ineffectiveness of this measure to characterize

topographic complexity within a single habitat at the scale of whole reefs. Most previous

studies that detected rugosity relationships were conducted across multiple habitats

(Friedlander and Parrish 1998) or used manipulated patch reefs to maximize the gradient

of rugosity (Hixon and Beets 1989).

Some reef fishes respond to habitat features at fine spatial scales, while other reef

fishes respond to features at landscape scales. For several fish groups, the combination of

fine and landscape-scale features provided the best predictive model, findings that

support, in part, small-scale reef research (Walsh 1985). Thus, scale has profound effects

on resultant patterns (Wiens 1989) with fine-scale measures often better predictors of one

group of organisms, and landscape measures predictors of others (Mitchell et al. 2001,

Mazerolle and Villard 1999). This organism-based perspective appears to be true for

coral reef fishes (Pittman et al. 2004), consequently future studies should acknowledge

that species perceive the landscape in different ways. The relevant scale of investigation

may depend on life history attributes of individual fish species (Hovel and Lipcius 2003,

Pittman et al. 2004), or biological processes such as foraging behavior (Shulman and

Ogden 1987), and predation (Hixon and Beets 1989). While these results, using a single

fine-scale measure of rugosity, suggest that landscape-scale measures are more valuable

in predicting most reef fish parameters, fishes appear to respond to benthic structure at

multiple spatial scales, with each species responding to a unique suite of variables.

Future studies will require an organism-based perspective that explores multiple spatial

scales.









These results should be interpreted within the scope and limitations of this purely

correlative study, and although a range in the values of different metrics was represented,

I had little control on experimental units. These reefs are natural habitat patches,

therefore considerable microhabitat variation exists, which is neither measure nor

controlled. Reefs also do not represent a perfect gradient in landscape scale habitat

features since sample units were selected from the naturally available set of reefs; rather

they vary across multiple gradients. Additionally, while considerable groundtruthing was

conducted, the benthic habitat maps were accepted without major modification. Since

each decision in the mapping process effects the determination and analyses of spatial

structure, it also effects our results of relationships between this structure and ecological

pattern. Finally, the reef fish populations of the Virgin Islands have been heavily

exploited (Beets and Rogers 2001), therefore future studies should examine reef fish

distributions in less fished areas.

Conclusions

This exploratory analysis allowed me to investigate relationships between

landscape structure and reef fish assemblage structure at 20 reefs to develop hypotheses

and guide future coral reef landscape studies. Landscape-scale metrics proved valuable

in characterizing and quantifying the landscape structure of the coral reef environment

(e.g., size, shape and context). Principal components of these metrics, however, were

correlated with few reef fish assemblage parameters. Interpretation of the few reef fish-

principal component relationships led to the conclusion that individual habitat features

are better measures of the influence of the spatial patterning of the coral reef landscape to

reef fish species diversity and abundance. Specifically, reef context was associated with

reef fish diversity and abundance for many groups of reef fishes. Because results






30


revealed that species responded to different scales depending on their life history

attributes and habitat requirements, future studies will examine specific functional

relationships (e.g., seagrass and grunts) based on the ecological requirements of the

particular fish groups of interest. If the results detected in this exploratory study are

replicable across systems and scales, combining the disciplines of landscape ecology and

reef fish ecology offers promise in addressing important management questions relevant

to habitat-based conservation of reef fishes.









Table 2-1. Reef fish assemblage parameters (n
statistical analyses
Entire assemblage Trophic
level parameters guilds
Cumulative species Herbivores (J & A)
richness
Mean species richness Mobile invertebrate


Total abundance


feeders (J & A)
Omnivores (J & A)
Piscivores (J & A)
Planktivores
Sessile invertebrate
feeders


30) used as dependent variables in


Mobility
guilds
Resident

Mobile


Taxonomic
groupings
Acanthurids (J & A)

Serranids (J & A)


Transient Haemulids (J & A)
Lutjanids (J & A)
Scarids (J & A)
Chaetodontids

Holocentrids
Labrids
Pomacanthids


Note: Fish groups are not always mutually exclusive. Hypoplectrus species were not
included in Serranid grouping. For each trophic guild and taxonomic grouping, reef fish
parameters were further subdivided into juvenile and adult components, where indicated
(J =juvenile, A = adult).









Table 2-2. Fourteen metrics used to quantify the landscape structure of the 20 study reefs
sampled in 1994 and 2001 in St. John, USVI.
Patch metric Definition Formula


Patch Size (reef)p
Polygon Size (reef)p
P:A of a Patchp

Habitat Richness1

Patch Richness1

Habitat Area

bedrock'
deep'
algal plain'
pavement'
reef1
sand'
seagrass


Size of individual habitat patches.
Size of individual patch
Sum of the patch edge divided by
patch area for patch of interest.
Number of different habitat types
present in an extent.
Number of patches of each habitat
type in extent of interest.
Amount of each habitat type in
landscape.


Area (m2)
Area (m2)
I P : A for particular
patches
Number of different
habitat types
Number of patches

Area (m2) in each
habitat type


Patch Diversity' Total abundance and type of I pi In pi;
different patches (pi is the pi = proportion of
proportion of habitat for every area in (m2) of patch
individual patch) i for all patches
Habitat Diversity' Same as patch diversity but I pi In pi
boundaries of similar habitat pi= proportion of
patches (by habitat class) are area in (m2) of
dissolved so that number of habitat type i for all
patches does not influence index, habitat types
p i for habitat diversity is
proportion of habitat for all
patches.
Note: For each patch metricp, a single value was calculated for every reef. Each
landscape metric' was calculated at three spatial extents (100 m, 250 m, and 500 m)
resulting in a total of 36 metrics for each study reef.









Table 2-3. Summary statistics on reef configuration, context and rugosity and for select
reef fish assemblage parameters (entire assemblage level, trophic level and
mobility guilds) for 20 study reefs sampled in 1994 and 2001, St. John, USVI
for metrics at the 100 m spatial extent with coefficient of variation for
landscape parameters and standard error for reef fish parameters.
Habitat
Measure Transform Min Max Mean CV
parameter
Total area Ha None 6.53 18.3 13.02 24.89
H' Patch diversity Index None 1.13 2.45 1.63 20.54
H' diversity Index None 0.54 1.57 1.17 24.32
Habitat richness # habitat types None 2.00 6.00 4.15 27.39
Patch richness # patches Logio (x +1) 5.00 19.00 8.50 43.09
Size of reef Ha None 0.54 15.74 6.59 67.62
P:A reef ratio None 0.03 0.09 0.05 34.33
Reef Ha Logio (x +1) 0.83 17.89 4.18 91.42
Seagrass Ha Logio (x +1) 0.00 7.80 1.32 166.26
Bedrock Ha Logio (x +1) 0.00 3.56 0.80 142.92
Pavement Ha Logio (x +1) 0.00 11.33 2.72 105.72
Deep water Ha Logio (x+1) 0.00 7.30 2.18 120.10
Algal plain Ha Logio (x +1) 0.00 4.02 0.58 222.76
Sand Ha Logio (x +1) 0.00 6.03 1.17 160.39
Rugosity Index None 1.38 2.82 2.00 19.48
Reef fish
ReUnits Transform Min Max Mean SE
Parameter
Mean spp richness Number Logio(x + 1) 19.67 32.14 23.43 0.67
Cum spp richness Number Logio(x + 1) 51.00 88.00 67.85 2.25
Total Abundance Number Logio(x + 1) 24.12 88.13 55.23 0.02
A Herbivore Number Logio(x + 1) 4.89 25.92 10.22 0.04
JHerbivore Number Logio(x+ 1) 14.14 73.13 41.66 0.04
AMIF Number Logio(x + 1) 1.04 18.95 4.37 0.05
JMIF Number Logio(x+ 1) 7.71 32.11 16.38 0.03
A PISCI Number Logio(x + 1) 0.05 1.75 0.45 0.02
J PISCI Number Logio(x + 1) 0.58 3.27 1.63 0.03
PLANK Number Logio(x + 1) 3.17 950.21 36.15 0.08
A OMNI Number Logio(x + 1) 0.00 1.19 0.26 0.02
J OMNI Number Logio(x + 1) 1.57 34.48 8.55 0.08
SIF Number Logiox + 1) 1.69 34.48 9.23 0.03
Resident A Number Logio(x + 1) 32.88 103.7 65.07 0.05
Mobile A Number Logiox + 1) 0.48 11.59 1.95 0.03
Transient A Number Logio(x+ 1) 24.12 88.13 55.23 0.04
Note: All parameters for mean abundance, except where indicated. All data are
backtransformed. CV = coefficient of variation, SE = standard error.













Table 2-4. Pearson product moment correlation matrix of the 14 landscape-scale habitat variables, with the resultant 9 remaining
significant variables, at the 100 m spatial extent for the 20 reef sites sampled in 1994 and 2001 in St. John, USVI.


P:A
reef
1.0
-0.03
0.03

-0.47
-0.76
-0.23


#
Patch
-0.03
1.00
0.44

0.43
0.33
0.84


-0.03 0.29


P:A reef
# patches
Habitat
richness
Reef size
Focal reef
Patch
diversity
Habitat
diversity
Bedrock
Deep
Algal
plain
Pavement
Reef
Sand
Seagrass


-0.08
-0.05
0.02

0.20
0.36
0.09
-0.04


Habitat
richness
0.03
0.44
1.00

0.41
0.23
0.63


Reef
size
-0.47
0.43
0.41

1.00
0.40
0.44


0.87 0.43


0.45
0.17
0.41

-0.30
-0.20
0.25
0.04


0.29
-0.15
0.36

.0.003
0.24
0.25
-0.28


Focal
reef
-0.76
0.33
0.23

0.40
1.00
0.46

0.09

0.40
-0.55
-0.01

-0.24
0.46
0.23
0.38


Patch
diversity
-0.23
0.84
0.63

0.44
0.46
1.00

0.58

0.17
-0.10
0.17

-0.05
0.31
0.25
0.06


Habitat
diversity
-0.03
0.29
0.87

0.43
0.09
0.58


Bed
rock
-0.31
-0.08
0.45

0.29
0.40
0.17


Deep

0.50
-0.05
0.17

-0.15
-0.55
-0.10


Algal
plain
-0.02
0.02
0.41

0.36
-0.01
0.17


Pave-
ment
0.17
0.20
-0.30

-0.003
-0.24
-0.05


1.00 0.46 0.22 0.45 -0.16


0.46
0.22
0.45

-0.16
-0.14
0.30
-0.09


1.00
-0.20
0.21

-0.76
0.05
0.25
0.26


-0.20
1.00
0.30

0.12
-0.23
-0.43
-0.41


0.21
0.30
1.00

-0.08
-0.02
-0.05
-0.27


-0.76
0.12
-0.08

1.00
0.30
-0.23
-0.44


Reef

-0.44
0.36
-0.19

0.24
0.46
0.31


Sand

-0.32
0.09
0.25

0.25
0.23
0.25


-0.14 0.30


0.05
-0.23
-0.02

0.30
1.00
-0.12
-0.20


0.25
-0.43
-0.05

-0.23
-0.12
1.00
-0.14


-0.31
0.50
-0.02

0.17
-0.44
-0.32
-0.24


Seagrass

-0.24
-0.04
0.04

-0.28
0.38
0.06

-0.09

0.26
-0.41
-0.27

-0.44
-0.20
-0.14
1.00


Note: Values in bold represent those with significant pair-wise correlations. Those in shadow represent variables that were excluded
based on significant pair-wise correlations. These 14 variables were thereby reduced to the remaining 8 variables (those not in
shadow), which were used in subsequent principal component analyses.









Table 2-5. Principal component analyses on the correlation matrix of the 8 residual
landscape-scale habitat variables at the 100 m spatial extent for the 20 study
reefs sampled in 1994 and 2001 in St. John, USVI.
PC1 PC2 PC3 PC4
Eigenvalue 2.10 1.82 1.49 1.01
Percent 26.33 22.76 18.65 12.61
Cum Percent 26.33 49.09 67.74 80.35

# Patches 0.488 0.013 -0.149 0.470
Reef Size 0.549 0.173 0.033 0.049
Habitat Diversity 0.359 0.258 0.493 0.243
Deep -0.029 -0.422 0.607 0.216
Pavement 0.211 -0.537 -0.163 -0.192
Reef 0.333 -0.227 -0.516 0.084
Sand 0.234 0.490 0.043 -0.570
Seagrass -0.347 0.381 -0.265 0.550
Note: Loadings in bold represent the top three variables that contribute the most to
individual components.












Table 2-6. Stepwise regression results to determine the influence of principal components on reef Esh assemblage structure
uts d reefs sam led in 1994 and 2001 in S t


at the 20


LL. ,. y ^'^'-i. L* ,*M .m xi x J r i ^^ .i ^\ J \j i. OUXXX. p- V JL -i J.^ m ^' iJU, m, j ^,^\.i^'m
Fish Model PC1 PC2 PC3 PC4
parameter R2 bl R2 p b2 R2 p b3 R2 p b4 R2 p
Mean richness 21% 1.38 0.21 0.040
Cumulative richness 26% 5.17 0.27 0.020
Herbivores 53% 0.05 0.22 0.01 -0.06 0.31 0.004
Omnivores 21% 0.14 0.21 0.044
Haemulids 43% -0.14 0.43 0.002
Epinephelids 23% -0.05 0.23 0.030
Acanthurids 35% 0.09 0.18 0.05 -0.10 0.17 0.040
Lutjanids 46% -0.07 0.32 0.006 0.05 0.14 0.050
Mobile 28% __0.05 0.28 0.02___


Note: Each of the 30 reef fish parameters were used as dependent variables. Linear models: log abundance = bO + bl (PC). Data
represented are mean and cumulative species richness values and mean abundances within each guild derived from a minimum of 16
samples per reef. See table 4 for definitions of individual principal components. The suite of 30 fish parameters were analyzed,
however, only those reef fish parameters with statistically significant relationships are reported. P-values are Sequential Dunn-Sidak
Bonferroni-corrected for the total number of comparisons (n = 30). Only those relationships with p < 0.05 are presented.









Table 2-7. Stepwise multiple regression results of the influence of reef configuration on
reef fish assemblage structure for the 1994 (N = 14) study reefs in St. John,
USVI.


Reef fish Habitat 2 2
Reeffish Habitat Model R2 Partial R2 p-value
parameter parameter
A Piscivores P:A reef (-) 0.32 0.32 0.0500
A Omnivores P:A reef 0.40 0.40 0.0100
J Haemulids # patches (-) 0.39 0.39 0.0200
A Acanthurids P:A reef 0.33 0.33 0.0300


Pomacanthids P:A reef 0.37 0.37 0.0200
J Lutjanids # patches (-) 0.36 0.36 0.0300
Transients P:A 0.74 0.62 0.0002
# patches 0.12 0.0500
Note: Independent variables were: P:A of each reef, the number of patches and reef size.
Results for 1994 reefs (n=14), with model R2 and partial regression values for each
variable with p < 0.05 level. P-values are Sequential Dunn-Sidak Bonferroni-corrected
for the total number of comparisons. Model effects are all positive, except where
indicated (-). The suite of 30 fish parameters were analyzed, however, only those reef
fish parameters with statistically significant relationships are reported. For all other reef
fish parameters, there are no significant relationships. A = adult, J= juvenile.









Table 2-8. Stepwise multiple regression results of the influence of reef context on reef
fish assemblage structure for the 1994 (N = 14) study reefs in St. John, USVI
Fish parameter R2 Partial R2 p-value Habitat parameter
Mean species richness 0.28 0.28 0.04 Seagrass
J Herbivores 0.30 0.30 0.04 Bedrock
A MIFs 0.33 0.33 0.03 Seagrass
J Omnivores 0.48 0.48 <0.01 Deepwater
A Piscivores 0.51 0.51 <0.01 Reef
J Piscivores 0.47 0.24 0.03 Deepwater
0.23 0.05 Seagrass
SIFs 0.63 0.39 <0.01 Bedrock (-)
0.27 <0.01 Deepwater
A Haemulids 0.53 0.53 <0.01 Seagrass
A Epinephelids 0.52 0.28 <0.01 Reef
0.24 0.04 Seagrass
J Acanthurids 0.67 0.42 <0.01 Deepwater (-)
0.25 0.01 H' (-)
A Lutjanids 0.68 0.50 <0.01 Seagrass
J Lutjanids 0.46 0.46 <0.01 Bedrock
Mobile 0.39 0.39 0.02 H'
Note: Independent variables were: H' and areal extent of reef, bedrock, seagrass and deep
unknown within 100 m. Model R2 and partial regression values for each variable with p
<0.05 level. P-values are Sequential Dunn-Sidak Bonferroni-corrected for the total
number of comparisons. The suite of 30 fish parameters were analyzed, however, only
those reef fish parameters with statistically significant relationships are reported. Model
effects are positive except where indicated by (-). A = adult, J =juvenile.









Table 2-9. Stepwise multiple regression results of the relative influence of landscape and
fine-scale habitat measures on reef fish assemblage structure on the 1994
(N = 14) study reefs on St. John, USVI.
2 Partial Habitat
Fish parameter R 2 p-value parat
R2 p-valuparameter


Mean species richness


A MIFs


J Omnivores


A Piscivores

J Piscivores


A Haemulids


A Epinephelids


J Acanthurids

A Lutjanids


0.68


0.53


0.76


0.53

0.71


0.71


0.54


0.46

0.44


0.38
0.30

0.33
0.20

0.60
0.16

0.53

0.36
0.35

0.55
0.16

0.38
0.16

0.46

0.44


0.0040
0.0008

0.0100
0.0500

0.0079
0.0200

0.0090

0.0030
0.0070

0.0006
0.0300

0.0090
0.0400

0.0100

0.0050


Seagrass
Rugosity

Seagrass
Rugosity

Rugosity
Reef

Reef

Rugosity
Seagrass

Seagrass
Rugosity

Reef
Seagrass

Deepwater

Seagrass


Note: Independent variables are rugosity and the areal coverage of deep unknown,
seagrass and reef within 100 m. Results for the 14 reefs sampled in 1994, with model R2
and partial regression values for each variable significant at the p = 0.05 level. P-values
are Sequential Dunn-Sidak Bonferroni-corrected for the total number of comparisons.
Relationships in bold are those that rugosity contributed to explanatory power. The 30
fish parameters were analyzed. Only fish parameters with significant relationships are
reported. A= adult, J =juvenile.






40



"-
w^tli.
\% *






`'A

b- ':
.--'' v
.-.-"--' "**


Figure 2-1. Location of St. John, US Virgin Islands in the Caribbean basin.





































Figure 2-2. Distribution of the 20 study reefs around the island of St. John, USVI. Below the name of each reef is the number of fish
point counts per reef.



















05
0 Vy


-30 -25 -20 -15 -10 -05 0 05 10 15 20 25


20 C


E V


-30 1 1
-30 -25 -20 -15 -10 -05 0 05 10 15 20 25 30


Figure 2-3.


PCA plots of the landscape structure of the coral reef environments of the 20 study reefs sampled in 1994 and 2001 in St.
John, USVI at the A) 100 m, B) 250 m and C) 500 m spatial extent. The x-axis is PC 1 and y-axis is PC 2. Symbols refer
to different types of reefs based on pre-existing knowledge of the coral reef landscapes, irrespective of the benthic habitat
maps. Triangles refer to shallow reefs with surrounding pavement habitat, squares refer to reefs with seagrass nearby,
and upside down triangles refer to isolated patches of reef with large areas of deep water nearby.


05+


0 V
-05




















R2 = 0.62
p < 0.0002


R = 0.40
p < 0.01


0.


0
0.02


0.04


0.06


R2 = 0.28
p < 0.05



0.08 0.1


P:A reef


Figure 2-4. Effects of reef configuration on mean fish abundances of A) transient fishes,
B) adult omnivores, and C) adult piscivores for the 1994 (N =14) study reefs
in St. John, USVI.















R2 = 0.28
32 p=0.044


1 4 R2 = 0.33

1 2 p=0.03


00 02 04 06 08 10

Seagrass


R2= 0.53
p=0.003


c


00 02 04 06 08 10

Seagrass


08

06

04

02

00
00 02 04 06 08 10
Seagrass


05
R2 = 0.51
04 p=0.0009


03


02


01

0
02 04 06 08 10 12 14

Reef


R2 = 0.46
p=0.0006


0 02 04 06 08

Bedrock


R2 = 0.48
p=0.0044 O


00 02 04 06 08

Deep unknown


Figure 2-5. Effects of reef context on mean abundance of particular fish groups for the
1994 (N =14) study reefs in St. John, USVI. A) Mean species richness and
areal extent of seagrass 100 m, B) mean abundance of MIFs and areal extent
of seagrass 100 m, C) mean abundance of haemulids and areal extent of
seagrass 100 m, D) mean abundance of piscivores and areal extent of reef 100

m, E) mean abundance of juvenile lutjanids and areal extent of bedrock 100
m, and F) mean abundance of juvenile omnivores and areal extent of deep
unknown 100m. Independent variables are each logl0 X +1 in hectares.
MIFs refer to mobile invertebrate feeders. A = adult, J = juvenile.














CHAPTER 3
EVIDENCE OF FUNCTIONAL CONNECTIVITY IN A CORAL REEF ECOSYSTEM

Coral reef ecosystems are deteriorating worldwide, with symptoms including loss

of hard corals, declines in abundances of exploited reef fishes and reduced biological

diversity. Marine protected areas (MPAs) represent an important management tool for

reducing this degradation; however, their effectiveness is contingent on our

understanding of key ecological patterns and processes at appropriate spatial scales.

MPA effectiveness may also be dependent upon maintaining critical linkages between

essential habitat patches (e.g., seagrass and reef). Exploratory analyses of the

relationship of reef fish assemblage structure with coral reef landscape structure in the

U.S. Virgin Islands- one of the first studies that applied a landscape ecology approach to

coral reef ecosystems- provided the foundation for developing specific hypotheses (ie.e

reef context influences reef fish assemblage structure at the scale of individual reefs).

These hypotheses were then tested in this new study at 22 independent reefs. As

expected, reef context influenced the structure of reef fish assemblages, and specific

relationships were functionally consistent with the ecology of the fishes of interest.

Consistent with predictions, reefs with neighboring seagrass had the highest total fish

abundance, and highest abundances of fishes within the mobile invertebrate feeding

guild, and within the exploited families of haemulidae (grunts) and lutjanidae (snappers).

Species richness for the entire fish community and within particular fish groups were also

strongly associated with the areal coverage of seagrass neighboring study reefs,

suggesting the importance of habitat linkages to a diversity of species. Potential habitat









linkages were detected as far away as 1 kilometer, which may indicate that reef fishes

perceive the landscape at this spatial scale. These findings infer that functional habitat

connectivity/juxtaposition between essential habitat patches is important in structuring

reef fish assemblages, and further suggests that landscape measures of this habitat

connectivity may be useful to managers in the design of MPAs.

Introduction

A landscape generally refers to a heterogeneous area composed of local interacting

ecosystems (Forman 1995) made up of homogenous units, called habitat patches, and the

sizes and spatial arrangement of these patches can exert a strong influence on the

diversity, abundance, distribution, and movement patterns of organisms (Wiens 1989).

Movements and flows of energy, nutrients or organisms between ecosystems can be

either be facilitated or inhibited by the spatial arrangement of these habitat patches

(Forman 1995, Turner et al. 2001). The degree to which this exchange occurs can

depend upon the boundary (the zone composed of the edges of adjacent ecosystems),

context adjacencyy, neighborhood, and location within a landscape), or connectivity (how

connected or spatially continuous a corridor or matrix is) of habitat patches within the

landscape mosaic (Forman 1995). As we strive to manage entire ecosystems,

maintaining functional habitat linkages within this landscape mosaic may be crucial to

the effectiveness of protected areas (Noss 1983; Forman 1995, Robinson et al. 1995,

Turner et al. 2001, Lindenmayer et al. 2002), since a failure to consider spatial elements

such as edges, movement corridors, and landscape context in the design of protected

areas is likely to result in undesirable changes in community characteristics and possibly

the loss of key species (Noss 1983, McGarigal & McComb 1995).









Though these principles were largely derived in terrestrial systems, they likely

apply to tropical marine systems (Grober-Dunsmore et al. 2004, Kendall et al. 2003,

2004, Pittman et al. 2004, Jeffrey 2004), and may be essential for designing functional

marine protected areas (MPAs) in coral reef landscapes, particularly as coral reef

landscapes exist as a complex mosaic of interacting habitat patches (i.e., seagrass, reef,

and mangrove). Maintaining important habitat linkages among habitat patches should be

considered in the design of MPAs, which are increasingly being considered as a primary

ecosystem-based tool for improving fisheries management and protecting biodiversity

(Carr & Reed 1993, Allison et al. 1998). Unfortunately, there is little information on how

to best design MPAs (Ballentine 1997, Sale 2005), particularly for the conservation of the

complex of fishes that constitute the reef fish community.

Historically, coral reef fishes have been difficult to manage, in part, because

different species often have different habitat requirements (Sale 2002). Moreover, these

habitat requirements frequently change with ontogeny (Lindeman et al. 2000, Appeldoorn

et al. 2003). Though generally considered site-attached following settlement (Sale 1991),

many regularly move hundreds and even thousands of meters from their primary resting

and foraging locations (Plan Development Team 1990, Corless et al. 1997). These

movement patterns are often species and life history stage-specific and occur across

multiple spatial scales (Randall 1962, Lindeman et al. 2000, Appeldoorn et al. 2003).

Some migrate daily to forage (Stark & Davis 1966, Ogden & Zieman 1977, Helfman et

al. 1983, Holland et al. 1993, Tulevech & Recksiek 1994), others migrate annually to

spawning aggregations (Colin et al. 1997) and several exhibit multiple ontogenetic shifts

in habitat (Randall 1962, Appeldoorn et al. 1997). To date, the overwhelming majority of









reef fish studies have been conducted at small spatial scales (e.g., 1 m plots) (see Sale

2002), limiting our ability to understand important habitat linkages across larger spatial

scales.

Recent landscape-scale (hundreds of meters to kilometers) research however,

provides correlative evidence that cross-shelf location (Christensen et al. 2003), reef

context (Appeldoom et al. 2003, Kendall et al. 2003, 2004, Grober-Dunsmore et al.

2004a, 2004b, Pittman and McAlpine 2003, Mumby et al. 2004, Pittman et al. 2004) and

landscape connectivity (Ault & Johnson 1998, Jeffrey 2004) influences reef fish

community structure. This study builds on previous research by assessing the generality

of my correlative models from the U.S. Virgin Islands (Grober-Dunsmore et al. 2004b) at

an independent set of smaller candidate reefs (<1 hectare) and explicitly testing

hypotheses of the importance of functional habitat linkages between two habitat types,

reef and seagrass, in structuring reef fish communities. The following predictions were

made (1) each reef fish parameter of interest (i.e. entire assemblage level parameters and

abundances and species richness within mobile invertebrate feeders (MIFs), grunts,

(Haemulidae), snappers (Lutjanidae), and groupers (Epinephelines) will be greater at

reefs proximal to seagrass, (2) abundance and species richness of mobile fishes will be

greater at reefs proximal to seagrass, but resident and transient guilds will not, (3)

relationships of reef fish parameters and the areal coverage of seagrass will be similar in

nature and strength between years and (4) the fine-scale measure of rugosity and

landscape-scale measures of seagrass will both be included in the best predictive models

of fish assemblage structure. These particular fish families were selected because they

are important components of the subsistence fishery in the Caribbean region, and because









they rely on a variety of habitats (i.e. seagrass) for foraging and settlement (Lindeman et

al. 2000). Consequently, they are more likely to demonstrate effects of habitat linkages

between reef and seagrass.

Study Area

This study was conducted in the shallow waters around the island of St. John,

USVI, located in the Northern Antilles approximately 88 km east of Puerto Rico (Figure

2-1). St. John is part of the Puerto Rico Bank, a submerged plateau defined by the 183 m

depth contour extending from eastern Puerto Rico to the island of Anegada. Study reefs

(n = 22) were selected around St. John (Figure 3-1) to maximize the variation of the

particular landscape parameter of interest (i.e. the areal coverage of seagrass habitat),

while reducing variation due to within-reef characteristics (e.g., coral cover). Each study

reef was 1-2 hectares in total area, within 1 km of the shoreline and occurred at depths

of 3 to 10 m.

All reefs were dominated by Montastraea annularis with Agaricia agaricites,

Porites porites, P. astreaoides, Siderastraea siderea and S. radians contributing to total

living coral cover, estimated at approximately 5-15 % for each reef. The background

matrices of study reef patches were typically hard-bottom or sand with large colonies of

M. annularis providing the major structural components. In general, the neighboring

seagrass communities were quite similar (e.g., typically dominated by Thalassia

testudinum, in shallow areas with contributions from Syringodiumfiliforme and varying

densities of rhizophytic algae).









Methods

Reef fish Sampling

Reef fishes were censused using a standardized visual point count census method

(Bohnsack & Bannerot 1986), where all reef fishes were identified within a 5 minute

sampling period, and enumerated during the following 10 minute period within a

sampling radius of 7.5 m2. Total lengths of fishes were estimated to the nearest

centimeter. Sampling effort was standardized to reef size (one census per 1500 m2)

following Monte Carlo simulation in Grober-Dunsmore et al. (2004a). A single observer

(RGD) conducted all reef fish censuses to eliminate observer variability.

A suite of fish parameters was estimated from the visual point count data (see

Table 3-1 and Grober-Dunsmore et al. 2004 for more details). Randall (1967) and Froese

and Pauly (2002) were used as references to classify all fishes by trophic guild: piscivore,

herbivore, mobile invertebrate feeder (MIF), sessile invertebrate feeder (SIF), planktivore

or omnivore. A 'seagrass-associated' category of fishes was also created using Lindeman

et al. (1998), Lindeman et al. (2000), Appeldoorn et al. (2003) and personal observations

(for Epinephelus striatus and Cephalopholis cruentatus), and includes these grunt,

grouper and snapper with explicit associations with seagrass: Lutjanus apodus, L. analis,

L. griseus, L. jocu, L. mahogoni, Ocyurus chrysurus, Haemulonplumieri, H. sciurus, H.

flavolineatum, E. striatus and C. cruentatus. Commercially-important families of grunts,

snappers, and groupers were examined separately. Given the known natural history, each

species was also into mobility guilds. Resident species are site-attached, and do not

typically move from their primary reef patch. Mobile species have restricted movements

among adjacent habitat patches, and may roam from the primary reef patch during

foraging. Transient species are vagile, ranging on the scale of kilometers. Fishes were









also subdivided into juvenile and adult categories, based on average size at maturity

(Froese & Pauly 2002), where possible. Using minimum estimated length, a fish was

placed in the juvenile category if it was below the average estimated size at maturity, and

placed in the adult category if it was at or above this estimate. Mean abundance values

based on replicate censuses at each reef were calculated and used in subsequent analyses

for all groups.

Temporal Sampling

In August 2003, reef fishes at a subset (n = 8) of the original 22 (2002) reefs were

recensused (Figure 3-1, reefs in bold) to assess the temporally consistency of detected

relationships. These reefs were selected as they represented a near maximal gradient in

seagrass areal coverage, and had similar benthic community structure and depth, reducing

effects of among-reef variability. Sampling effort was increased to improve the accuracy

of the estimate of the mean value for each reef fish parameter, thus a minimum of 10

censuses per reef were conducted, regardless of reef size.

Habitat Sampling

Landscape metrics of reef context were calculated with ArcView 3.2 software

(ESRI 1996), using digital benthic habitat maps. Maps were created from aerial

photographs flown at an altitude of 5000 feet in 1999 and were classified by visual

interpretation, using 26 discrete and non-overlapping habitat classes, with a minimum

mapping unit of 1 acre (Kendall et al. 2001). The original map classification scheme was

condensed from 26 to 7 distinct, non-overlapping habitat classes (pavement, sand, reef,

bedrock, seagrass, macroalgae and deep unknown), so results would be more broadly

applicable to resource managers and to reduce potential classification errors. Reefs

served as focal units for all analyses because MPAs are frequently designed around









individual reefs (e.g., Florida Keys National Marine Sanctuary). Landscape metrics were

calculated at the 100 m, 250 m, 500 m and 1 km spatial extent, to explore the appropriate

scale for each response variable. Rugosity was measured in situ along ten 10 m transects

at each reef fish census location, using methods described by Luckhurst and Luckhurst

(1978). A mean was calculated for each reef.

Statistical Analyses

All data were checked for normality using Shapiro-Wilks tests (p < 0.01) (Zar

1984), and transformed where appropriate (Table 3-1). If assumptions were met,

analyses were conducted using parametric statistics, and non-parametric statistics if not.

For all analyses, each reef fish parameter was used as the dependent variable and the

landscape and/or rugosity measure was used as the independent variable. Residual plots

were examined to assess stability of regression models (Sokal & Rohlf 1995).

Because predictions were derived at different study reefs, analyses were conducted

to verify the importance of reef context at these independent reefs using stepwise

multiple regression methods as in Grober-Dunsmore et al. (2004b) for all reef fish

parameters. Model II linear regression analysis (Sokal & Rohlf 1995) was used to test

the prediction: 1) abundances of MIFs, grunts, snappers, groupers, mobile reef fishes,

seagrass-associated taxa and species richness values within these fish groups are higher at

reefs proximal to seagrass. Model II linear regression was also used to test the

prediction: 2) mobile taxa will be influenced by reef context (measured by the areal

coverage of seagrass), whereas resident and transient taxa will not. Non-parametric

Spearman Rho correlations (Sokal and Rohlf 1995) were calculated for the subset of 8

reefs (2003), sampled in 2002 and 2003 to test the prediction: 3) relationships between

reef fish parameters and reef context will be similar in nature (e.g., direction) and









strength between years. Statistical comparisons of the abundances of reef fishes, using

raw census data (n = 130), were conducted using the non-parametric Kruskal-Wallis to

test for differences in abundance between years. Stepwise multiple regression analyses

was used to test the prediction: 4) reef fishes will be structured by both landscape-scale

and fine-scale measures of habitat.

Results

A total of 107 fish censuses were conducted at 22 reefs during July and August

2002. One hundred eighteen species were identified and a total of 14,389 individuals

recorded. In August 2003, 97 reef fish censuses were conducted at the subset of 8 reefs,

with one hundred twenty two species identified and a total of 14,239 individuals

recorded. The ten most abundant taxa in each fish group are listed in Table 3-1.

The configuration and composition of the study landscapes varied widely among

reefs, with coefficients of variation for most landscape metrics > 50 % of the mean (Table

3-2). Gradients in the areal coverage of seagrass were represented adequately to test

hypotheses concerning reef-seagrass habitat linkages (Table 3-2). Rugosity also varied,

although to a small degree than seagrass (Table 3-2).

Exploratory stepwise regression analyses confirmed that reef context was the best

predictor of reef fish assemblage structure, allowing me to test specific hypotheses

concerning the importance of seagrass in particular. For all subsequent analyses, I

verified that reefs were not consistently higher for other reef fish parameters to refute the

possibility that some unmeasured factor did not result in high values for all fish

parameters.









Entire Assemblage Level Parameters

The areal coverage of seagrass was a positive predictor of all entire assemblage

level parameters for every spatial extent (Table 3-3), with 48-58 % of the variation in

cumulative species richness explained by seagrass (Table 3-3, Fig. 3-2). Higher

cumulative richness at reefs with seagrass nearby can be partially attributed to the

presence of several species with specific dependencies on seagrass (e.g., Aetobatus

narinari, Xyrichtys splendens, Sparisoma radians, Holocentrus rufus). Reefs surrounded

by seagrass had the highest total number of species.

Abundances within Reef Fish Groups

The areal coverage of seagrass was a positive predictor of abundances of grunts

(R2 = 0.52-0.57), seagrass-associated taxa (R2 = 0.33-0.50), snappers (R2 = 0.34-0.40),

and MIFs (R2 = 0.27-0.41), though groupers were marginally associated (R2 = 0.13-0.17)

with seagrass (Table 3-4, Fig. 3-3). Reefs with large expanses of adjacent seagrass had

the highest abundances of MIFs, grunts and seagrass-associated taxa were Donkey,

Hansens and Marys.

Abundances within Life History Stages

Juvenile grunts were strongly (R2 = 0.52-0.56), whereas adults were weakly

associated with the areal coverage of seagrass (R2 =0.16-0.21) (Table 3-5), and the

juvenile component of the seagrass-associated taxa also exhibited a stronger relationship

than the adult (Table 3-5). The only reefs without juvenile yellow-tail snapper (Ocyurus

chrysurus) were those without seagrass within 1 km (e.g., Peter Bay, PeterOne), and

juvenile schoolmaster (Lutjanus apodus) and mahogony snapper (L. mahogoni) were

only present at reefs with extensive shallow back-reef seagrass (e.g., Saba E, Mary's,

Rendezvous).









Species Richness

The areal coverage of seagrass was also a positive predictor of species richness

within MIFs (R2 = 0.63-0.72), grunts (R2 = 0.56-0.71), snappers (R2 = 0.24-0.41) and

groupers (R2 = 0.23-0.33) (Table 3-6, Figure 3-4). Increased species richness values

were frequently attributed to the presence of those particular species associated with

seagrass. For example, species richness of MIFs was increased by the presence of

species that reside or forage in seagrass (e.g., Aetobatus narinari, Holocentrus marianus,

and Calamuspennatula), and species richness of snapper was increased by the presence

of small schools of juvenile schoolmaster, grey snapper and lane snapper. Grunt species

richness was enhanced by the presence ofH. plumieri, H. carbonarium H. chysargyreum,

and H. macrostomum; while H. plumieri is a known seagrass-specialist, seagrass

dependencies of the other taxa are not understood. Residual analysis suggested that

caution is warranted in interpreting the relationship for snappers and groupers.

Mobility

The areal coverage of seagrass was a positive predictor of abundances (R2 = 0.19-

0.21) and species richness (R2 = 0.55-0.69) of mobile taxa, but not for resident (except at

the 250 m extent) and transient species (Table 3-4, Table 3-6). Abundances of transient

taxa violated normality tests, and may have contributed to my inability to detect

relationships. Species richness was enhanced at reefs with neighboring seagrass by the

presence of more mobile species such as: Acanthurus chirugus, Balistes vetula, Calamus

calamus, Gerres cinereus and Holocentrus rufus, species that likely to migrate into

neighboring foraging patches (Kramer & Chapman 1999).









Spatial Extent

Although the R2 and p-values were comparable across all spatial extents for most

reef fish parameters and/or groups, the strength of relationships for the adult life history

stages was often greatest at the 500 m spatial extent, and generally occurred at the 100 m

or 250 m spatial extent for the juvenile life history stages (Table 3-5). For species

richness, the strength of associations was strongest at the 100 m (grunts, MIFs) and 250

m (snappers, groupers, and mobile taxa) spatial extents (Table 3- 6).

Temporal Consistency

Twelve of sixteen reef fish-habitat associations were similar between years in 2002

and 2003 (Table 3-7, Figure 3-5). Of these twelve, seven relationships (cumulative

richness, mean richness, total abundance, juvenile grunt abundance, juvenile grouper

abundance, species richness of MIFs, and species richness of groupers) were positively

correlated to the areal coverage of seagrass in 202 and 2003 at the 8 reefs (Figure 3-5).

Five were consistent in that they were not correlated with seagrass in either year (adult

and juvenile MIF abundance, adult and juvenile snapper abundance, and species richness

of snappers). Several relationships from the 2002 dataset of 22 reefs were not evident

when analysis was constrained to the subset of 8 reefs. The reduction in sample size from

22 to 8 clearly resulted in a loss of statistical power (Sokal and Rohlf 1995), and

abundances of three fish groups (MIFs, adult grunts, and adult snappers) were statistically

lower in 2002 compared to 2003 (Table 3-7).

Relative Influence of Fine and Landscape-scale Measures

The fine-scale measure of rugosity was of limited value in predicting reef fish

assemblage structure; rather, landscape-scale measures of seagrass areal coverage were

better predictors of most fish assemblage parameters, though there were a few exceptions









(Table 3-8). Collinearity diagnostics suggested that although weak relationships exist

between the independent variables, all model condition indices were less than 10, well

below the recommended value of 30 (Belsley et al. 1980).

Discussion

As predicted, landscape-scale measures of the areal coverage of seagrass calculated

with simple GIS tools and habitat maps were useful surrogates for reefs with high entire

assemblage level parameters (e.g., cumulative species richness). These findings are

relevant because earlier marine ecologists alluded to the importance of reef context

(Ogden & Zieman 1977, Parrish 1989) and high fish diversity has been attributed to the

presence of seagrass in the Florida Keys (Robblee & Zeiman 1984), the Virgin Islands

(Quinn & Ogden 1984, Grober-Dunsmore et al. 2004a), the Red Sea (Khalaf & Kochzius

2002) and the Australia (Pittman et al. 2004) and with other habitats such as algal beds

(Rossier & Kulbicki 2000) and mangroves (Birkeland 1985, Thollot 1992) in post-hoc

correlative analyses, but these results contribute to the growing body of evidence that

suggests that resource managers may be able to use landscape-scale measures of reef

context to detect areas with high fish abundance and diversity (Appeldoorn et al. 2003,

Kendall et al. 2003, Pittman and McAlpine 2003, Jeffrey 2004).

Abundances and species richness of predicted reef fish groups (grunts, seagrass-

associated taxa, snappers and MIFs) were higher at reefs with neighboring seagrass, and

observed patterns of distribution were remarkably consistent with the foraging ecology of

each fish group. For example, haemulids exhibited the strongest association to seagrass;

several taxa are known to regularly move from reefs into adjacent seagrass patches

(Burke 1995, Appeldoorn et al. 1997, Beets et al. 2003), sheltering on or near the reef by

day and feeding on crustacean fauna in surrounding seagrass by night (Ogden & Quinn









1984, Robblee & Zieman 1984, Burke 1995, Appeldoorn et al. 1997, Beets et al. 2003).

Abundances of seagrass-associated taxa were also strongly associated with seagrass,

which is expected since this guild is comprised exclusively of species that complete part

of their life history in seagrass. Because many MIF and snapper species have generalist

foraging requirements (Sale 2002), it is not unexpected that relationships with seagrass-

though still robust are less than that observed for grunts. Remarkably, the strength of

associations for abundances and species richness of the different fish groups

corresponded quite strongly with their known natural history, with the strength of

relationships generally following this rank order; grunts, seagrass-associated, snappers,

MIFs, and groupers.

This study design does allow inferences concerning the mechanisms that might

explain the higher species richness and abundances of reef fishes at reefs proximal to

seagrass. (1) Settlement and survivorship of some juvenile fishes may be higher in

seagrass (Nagelkerken et al. 2000, 2002, Appeldoorn et al. 2003) due to its high plant

diversity (e.g., epiphytes, algae, and phytoplankton) (Bell & Pollard 1989), which may

facilitate coexistence of species that would otherwise compete (Keller 1983). Settlement

(of individuals and different species) may be higher if seagrass intercepts larval fish more

effectively than other habitat patches (Parrish 1989). (2) The structural complexity of

seagrass can provide shelter from predation (Parrish 1989, Robertson & Blaber 1992),

thereby increasing survivorship of fishes in seagrass communities, which then may

migrate to neighboring reef patches as they mature. (3) Reef fish movement may occur

more readily in highly-connected marine landscapes (Noss 1983), therefore patches that

have habitat linkages to other essential habitat patches may support more fishes. (4)









Energy, nutrients and organic matter generated within seagrass communities (Duarte

2000) may flow to nearby reefs through direct animal movement (Meyer et al. 1983,

Meyer & Schultz 1985), predation, or outwelling of dissolved and particulate organic

matter (Sogard 1989, Beck et al. 2001), providing important resources for reef fishes.

While the mechanisms responsible for increased richness and abundances of fishes at

reefs proximal to seagrass remains to be identified, some combination of these factors

appears beneficial for recruitment, settlement, survivorship and/or coexistence of large

numbers of fishes.

Contrary to predictions, groupers were weakly associated with seagrass, although

detecting relationships at this spatial scale may be inhibited by their low population

densities (0-0.2 fishes per sample) and species composition. Historically, high densities

of groupers at several reefs in the U.S. Virgin Islands have been attributed to surrounding

seagrass communities (Randall 1962). Over time, grouper species dominance has shifted

to smaller, potentially less mobile, taxa; E. guttatus, Cephalopholis cruentatus and C.

fulvus are relatively more abundant than E. striatus, E. morio, and Mycteroperca tigris.

These latter groupers likely ranged over comparatively large areas; E. striatus made daily

movements of 400 m (Carter et al. 1994). In contrast, presently-dominant taxa (e.g., E.

guttatus) demonstrated high short-term fidelity at small patch reefs in Bermuda (Bardach

1958) and the U.S. Virgin Islands (Randall 1962, Beets et al. 2003). Habitat

requirements may be less stringent, abundances may be severely depressed (thereby

limiting our ability to detect relationships; Osenberg et al. 2002), or shifts in the species

composition may favor more site-attached species, such that detecting habitat linkages for

groupers is difficult in this system.









The juvenile component of the fish assemblage generally exhibited a stronger

seagrass association compared to adults, perhaps due to nursery benefits (Nagelkerken et

al. 2002) such as increased recruitment, increased habitat-mediated post-settlement

survivorship (Beukers & Jones 1997), and/or direct (Meyer et al. 1983, Meyer & Schultz

1985) or indirect transfer (Duarte 2000) of nutrients and energy from adjacent seagrass.

The strongest evidence that seagrass provides a nursery benefit is provided by the

relationships for juvenile grunts and juvenile seagrass-associated fishes with seagrass,

which was also demonstrated by Kendall et al. 2003 in St. Croix for juvenile Haemulon

flavolineatum. These two groups are comprised of those taxa with the greatest

dependencies on seagrass (Lindeman et al. 2000), and as expected these relationships

were relatively strong (R2 = 0.56, R2= 0.46), respectively.

Fishing pressure and within-reef heterogeneity may also explain why the adult

component of the reef fish assemblage generally exhibited weaker seagrass associations

compared to juveniles. Reef fish populations of the U. S. Virgin Islands appear heavily

over-fished (Rogers & Beets 2001), and densities of exploitable-sized fishes were

typically low in this study. In addition, adult grunts and snappers feed on benthic

invertebrates (Randall 1967, Rooker 1995, Nagelkerken et al. 2000) that typically reside

in soft-bottom patches (Gillanders et al. 2003). Such soft bottom patches contribute to

within-reef heterogeneity, and these patches may not be detectable at the resolution of

these habitat maps (Kendall et al. 2001). Consequently habitat linkages are expected to

be more evident in less-fished systems and for juveniles (which was the case) and habitat

specialists.









Consistent with predictions and terrestrial research (Sisk et al. 1997, Mitchell et al.

2001, Turner et al. 2001), the mobility (vagility) of fishes influenced how fishes relate to

the coral reef landscape. Resident taxa are more likely associated with finer-scaled

features (Hixon & Beets 1989, Sale 1998), as evidenced by the positive correlation

between rugosity and abundances of resident fishes, and the lack of an association with

this landscape measure of seagrass. The absence of a relationship for transient taxa is

consistent with their habitat utilization patterns, (e.g., Carynx spp. are likely to transit

indiscriminately over sand, reef, and seagrass habitats), as they respond to features at

larger spatial scales (i.e. 100's to 1000's of meters). These findings are a strong

imperative to focus on the scales that are appropriate to the organism, and indicate that

these scales can often be predicted by considering the ecology of the particular species.

Reef fish-seagrass associations were evident up tp 1 km spatial extent from study

reefs, suggesting that functional habitat linkages may operate at least at this spatial scale.

Again, the appropriate spatial scale appeared remarkably consistent with the ecology of

fishes; particularly with their life history stage. For example, juvenile relationships were

most strong at close distances (100 m and 250 m), whereas adults were strongest at 500

m and 1 kilometer. Given the high risk of mortality for juvenile fishes, reefs that have

seagrass in closer proximity may reduce predation risks, which would explain in part the

strong association at closer distances. Many adult fishes travel greater distances. For

example, several tagging studies revealed transit distances of 100 m 400 m for adult H.

plumieri (Tulevech & Recksiek 1994), H. sciurus (Beets et al. 2003) and lutjanids

(Chapman & Kramer 2000). These results are relevant for determining the geographic









boundaries of MPAs, since failure to include habitat patches at least 1 kilometer away

may result in mortality as fishes move outside MPA boundaries.

While several recent studies provide evidence of the importance of reef context to

reef fish assemblage structure (Nagelkerken et al. 2000, 2002, Appeldoom et al. 2003,

Kendall et al. 2003, 2004, Dorenbosch et al. 2004, Jeffrey 2004, Mumby et al. 2004), this

study builds on existing research in several important ways. 1) This study presents data

on abundance, rather than simple presence or absence of fishes. This distinction is

important since marine resource managers are frequently interested in determining the

location of reefs with high numbers of individuals and species. 2) This study

differentiates between soft-bottom (e.g., seagrass, sand) and hardbottom (e.g., reef,

pavement), which has often been difficult. 3) Measures of each habitat were quantified

using geo-referenced, digital habitat maps and GIS, which has historically not been

feasible. 4) Confounding effects of previous studies (i.e. reef size, location and other

habitat types) were minimized, though not eliminated. 5) Relationships were examined

for the entire fish community (e.g., within trophic and taxonomic groups), thereby

providing a comprehensive functional perspective of the response of fish parameters to

reef context. Previous studies have often been restricted to a subset of the reef fish

community (Nagelkerken et al. 2000, 2002, Kendall et al. 2003). 6) Explicitly-stated

apriori hypotheses were tested in this study, and study reefs were selected to test these

hypotheses. Future studies that apply landscape ecology principles to marine systems

should adopt a rigorous hypothetico- deductive approach.

To investigate the generality of new findings, it is ideal to determine the temporal

and spatial consistency of results (Sale et al. 1994), thus it is relevant that many reef fish-









seagrass relationships were similar between 2002 and 2003. For those relationships that

were not similar, a loss of statistical power (caused by the reduction in sample size) and

low fish densities may have lessened the ability to detect associations (e.g., Eggleston et

al. 1998). Reef fish communities display stochastic variability in community structure at

small spatial and temporal scales, and it is unclear from this limited temporal analysis

whether the absence of several expected relationships in 2002 is an indication that results

are not reliable. Clearly, future research is necessary to determine the influence of the

complex spatial and temporal variability of reef fish communities in predicting reef fish

habitat associations. While multiple sources of temporal variability induced by larval

recruitment (Doherty and Fowler 1994), post-settlement mortality (Hixon 1991), fishing

pressure and food availability (Williams 1980) exist, habitat relationships appeared

remarkably consistent over time, providing added confidence that this evidence of strong

habitat linkages is not a fluke consequence from a single sampling event.

Contrary to predictions and smaller-scale studies (e.g., Luckhurst and Luckhurst

1978, Hixon and Beets 1989, Friedlander and Parrish 1998), the fine-scale measure of

rugosity was not a predictor of most reef fish parameters, which may have important

consequences for MPA design. Landscape parameters are increasingly easy to measure

(Green et al. 2000, Kendall et al. 2001) conversely; rugosity is a time-consuming in-water

measure. If such high quality benthic maps as those produced today can now accurately

identify topographically complex reef habitat (Kendall et al. 2001), landscape measures

may be superior to fine-scale for designing effective spatial management schemes. The

failure of this measure of rugosity to predict remaining fish parameters may be because

rugosity is only a positive predictor of fish assemblage structure when sampling across a









larger range of topographic complexity (among distinct habitats). Sampling was

constrained within a single high relief habitat for this study, thus the range in rugosity

was lower than others (e.g., Friedlander and Parrish 1998). In addition, other measures of

structural complexity (e.g., hole size) may be better predictors of fish assemblage

structure (Friedlander & Parrish 1998). This multi-scalar analysis confirms that there is

clearly no correct scale that can be universally applied to all organisms (Bisonnette

2003), and corroborates terrestrial (Graham and Blake 2001) and marine studies (Grober-

Dunsmore et al. 2004b) that habitat-faunal relationships are inextricably scale-dependent.

Clearly, scale has profound effects on resultant patterns (Wiens 1989) with fine-scale

measures often better predictors of one group of organisms, and landscape measures

predictors of others (Mitchell et al. 2001, Mazerolle and Villard 1999). This organism-

based perspective appears to be true for coral reef fishes (Pittman and McAlpine 2003,

Pittman et al. 2004); consequently future studies should acknowledge that species

perceive the landscape in different ways.

Conclusions

Landscape measures of coral reef context appear to be valuable predictors of coral

reef fish assemblage structure, which may have important implications for MPA design.

Specifically, the areal coverage of seagrass may be used to successfully predict which

reefs have diverse reef fish assemblages and high abundances of commercially and

recreationally important species. Although this study cannot identify the processes

driving relationships, it provides strong evidence that functionally-linked marine

landscapes contribute to increased species richness and abundance of many fish groups.

Admittedly, MPA design requires an understanding of numerous factors (i.e. location,

distribution and amount of various habitats necessary for spawning, recruitment, larval






65


export, settlement, growth, foraging and reproduction). However, given the urgency of

MPA design decisions, the selection of areas for conservation should also consider their

contribution to the whole system and how well the location of a patch relates or links to

other patches within the landscape.































Figure 3-1. Location of the 22 study reefs around the island of St. John, US Virgin
Islands sampled in 2002, with the eight study reefs re-sampled in 2003
indicated in bold.












a) 70
R2 = 0.58
65 p < 0.0001


60


0 02 04 06 08 1 12 14 16 18


R2 = 0.27
33 p < 0.01

31

29

27


* .


0 02 04 06 08 1 12 14

Seagrass 250 m


16 18


Figure 3-2. The relationship of a) cumulative richness and b) mean species richness with
the areal coverage of seagrass (hectares) at 250 m for the 22 study reefs
sampled in 2002 in St. John, U.S. Virgin Islands. The x-axis is logo (x +1)
transformed.












a) R2 = 0.41
2 p< 0.0001

19

18
C
0 17

m 16

S 15
E
14

13

12 *

11
0 02 04 06 08 1 12 14 16 18


b) 1 8
R2 = 0.57
1 6 p < 0.0001

14

S12
.I
I 1

E 08

1 06
E

S04

02


0 02 04 06 08


12 14 16 18


14 16 18


0 02 04 06 08 1 12 14
Seagrass 250 m


Figure 3-3. The relationship of mean abundances of a) MIFs, b) haemulids, c) seagrass-

associated taxa and d) lutjanids with the areal coverage of seagrass

(hectares) within 250 m of the 22 study reefs in St. John, U.S. Virgin Islands

sampled in 2002. Mean abundances and the areal coverage of seagrass are

logo (x +1) transformed.


R2 = 0.50
p < 0.0002


. .


R2 = 0.40

oa p < 0.0029


c) 2

18

S16

14

12

E 1

08

06
n 06




o 04
02
S,


0 02 04 06 08 1 12
Seagrass 250 m


16 1


















a) R2 = 0.69
p = 0.0001


30 *


0 02 04 06 08 1 12 14 16 18



R2 = 0.70
7 p = 0.0001


. .


* /6.r


* *


0 02 04 06 08 1 12 14 16 18


c) 45 R2= 0.41
p = 0.0013


35

3
I 3

..
S25

E 26

u
_r


* ** *


0 02 04 06 08 1 12
Seagrass 250 m


14 16 18


Figure 3-4. The relationship of cumulative species richness of a) MIFs, b) haemulids and

c) lutjanids with the areal coverage of seagrass (hectares) within 250 m for 22

study reefs in St. John, U.S. Virgin Islands sampled in 2002. The y-axis is

logo (x+1) transformed.


*/* *




)* *























2003
Rho= 0.57
p<0.1



*.,'


23


* E~.-
---
U

A


2002
Rho = 0.66
p < 0.07


0 02 04 06 08 1


2002
Rho = 0.59
p <0.11


21 *





19 a




17 2003
Rho = 0.57
p<0.13


15
0 04 06 08


2003
Rho = 0.81
15 < 0.02

*






S 2002

A, "

05 ,'"/
/-- 2002
,' 9 Rho = 0.81
p < 0.02



0 02 04 06 08


2003
Rho = 0.84
p < 0.01


* ,


/ U


// -


V ..
.'^^'


0 02 04 06
Seagrass 250 m


2002
Rho = 0.94
p < 0.01





08 1


2003
04 Rho=0.76
p < 0.01


U
A
A
.- ,
4, -
.-* -,
-, -- U
U


- /.


U


2002
Rho = 0.83
S p < 0.01


0 02 04 06 08 1


.'


2003
Rho = 0.84
p<0.01


,/

I -- 2002
Rho = 0.73
p <0.03


0 02 04 06
Seagrass 250 m


08 1


Figure 3-5. Spearman rank correlations for those reef fish parameters that demonstrated a

consistent relationship with the areal coverage of seagrass habitat between

2002 ( ----) and 2003 ( -- --) at the subset of 8 study reefs in St. John, U.S.

Virgin Islands


a) 0s









Table 3-1. Most abundant taxa in each reef fish group for the 22 study reefs sampled in
2002 in St. John, U.S. Virgin Islands.


Reef fish


parameter
Total abundance




MIFs*




Haemulids
(grunts)


Lutjanids
(snappers)

Epinephelines
grouperss)

Mobile taxa




Resident taxa



Transient taxa


* Mobile invertebrate feeders


Dominant taxa

Thalassoma bifasciatum, Acanthurus coerulus, Haemulon
aurolineatum, Scarus spp. (unidentified juveniles), H. flavolineatum,
Stegastes planifrons, A. bahianus, Halichoeres garnoti, S.
leucostictus, and Caranx ruber

H. flavolineatum, H. aurolineatum, H. garnoti, Abedufdufsaxtalis,
Halichoeres bivittatus, Ocyurus chrysurus, Holocentrus rufus,
Halichoeres maculapinna, Hypoplectrus puella, and Mulloides
martinicus

Haemulon aurolineatum, H. flavolineatum, H. juvenile, H. sciurus,
H. plumieri, H. parrai, H. carbonarium, H. macrostomum, and H.
chysargyreum

Ocyurus chrysurus, Lutjanus apodus, L. synagris, L. mahogani, L.
jocu, L. griseus, and L. analis

Epinephelus guttatus, Cephalopholis cruentatus, C. fulvus, E.
striatus, and E. adscensionis

Acanthurus coerulus, Haemulon aurolineatum, Scarus spp.
(unidentified juveniles) Haemulon flavolineatum, Halichoeres
garnoti, Haemulon spp. (unidentified juveniles), Thalassoma
bifasciatum, Halichoeres bivittatus, and Sparisoma aurofrenatum

Stegastes planifrons, S. leucostictus, S. partitus, Abedufdufsaxtalis,
Chromis cyanea, C. multilineatum, Hypoplectruspuella, and S.
variabilis

Carynx ruber, Ocyurus chrysurus, Inermia vittata, Scomberomorous
regalis, and Dasyatis americana









Table 3-2. Variable names, transformations, minimum, maximum and mean values for
each reef fish parameter and landscape metric, with the standard error for fish
parameters and the coefficient of variation for habitat measures for the 22
study reefs sampled in 2002 in St. John, US Virgin Islands.


Transform
Log (x +1)
Log (x +1)
Log (x +1)
Log (x +1)
Log (x +1)


Snapper abundance a Log (x +1)
Grouper abundance a Log (x +1)
Seagrass asso. abundance Log (x +1) b
Mobile abundance Log (x +1)
Residents abundance Log (x +1)
Transients abundances Log (x +1) b
Richness MIFs None
Richness haemulids None
Richness lutjanids None
Richness epinephelids None b
Richness mobile None
Richness residents None
Richness transients None
Independent Variable Transform
Area (ha) seagrass 100 m Log (x +1)
Area (ha) seagrass 250 m Log (x +1)
Area (ha) seagrass 500 m Log (x +1) b
Area (ha) seagrass 1 km Log (x +1)
Rugosity None
a Divided into adult and juvenile components, also.
Shapiro-Wilks normality tests, using p < 0.01.


Dependent Variable
Cumulative richness
Mean species richness
Total abundance
MIF abundance
Grunt abundance a


Min
30.00
16.00
322.00
10.22
0.00
0.00
0.00
0.50
26.54
8.77
0.00
9.00
0.00
0.00
0.00
17.00
8.00
0.00
Min
0.00
0.00
0.00
1.01
1.18


bThose parameters that failed


Max
66.00
0.66
38.81
99.00
59.26
22.99
2.09
68.50
193.98
101.33
17.20
30.00
6.00
4.00
6.00
40.00
24.00
4.00
Max
11.59
25.92
41.36
76.62
2.28


Mean
44.20
22.42
99.00
25.92
3.89
1.69
0.78
13.2
59.26
25.92
2.31
17.36
3.13
2.18
3.59
26.10
15.68
2.18
Mean
1.21
3.27
6.76
13.79
1.56


SE
2.37
0.97
0.05
0.06
0.12
0.07
0.03
3.92
0.05
0.06
0.08
1.52
0.44
0.28
0.23
1.53
0.86
0.23
CV
97.96
78.36
65.23
38.97
15.70









Table 3-3. Simple linear regression for entire assemblage level parameters of reef fish
communities with the areal coverage of seagrass within 1 km of each study
reef as the independent variable at the 22 study reefs, sampled in 2002 in St.
John, U.S. Virgin Islands*.

Dependent variable Independent Variable F ratio Model R2 p-value
(x 100)
Mean Species Seagrass 100m (+) 6.21 23.68 0.0200
Richness Seagrass 250 m (+) 7.42 27.02 0.0100
Seagrass 500 m (+) 10.95 35.38 0.0035
Seagrass 1 km (+) 7.82 29.16 0.0015
Cumulative Species Seagrass 100 m (+) 17.62 47.84 0.0004
Richness Seagrass 250 m (+) 27.29 57.77 0.0001
Seagrass 500 m (+) 24.02 54.57 0.0001
Seagrass 1 km (+) 19.59 50.77 0.0003
Mean Abundance Seagrass 100 m (+) 4.22 17.44 0.0500
Seagrass 250 m (+) 6.21 23.60 0.0200
Seagrass 500 m (+) 9.38 31.94 0.0060
Seagrass 1km (+) 8.98 32.09 0.0070
*For each relationship, the dependent and independent variables are presented with
corresponding F-ratio, entire model R2 and p-value for the final model. Relationships in
bold represent the spatial extent with the strongest relationship, for the given fish
parameter. All relationships are positive (+).









Table 3-4. Simple linear regression of abundances of mobile invertebrate feeders, grunts,
snappers, groupers, and seagrass-associated taxa and within mobility guilds
with the areal coverage of seagrass within 1 km of each study reef at the 22
study reefs, sampled in 2002 in St. John, U.S. Virgin Islands*.


Dependent variable


Adult Haemulids




Juvenile Haemulids




Adult Lutjanids




Juvenile Lutjanids




Adult Epinephelids




Juvenile Epinephelids




Adult seagrass-
associated taxa



Juvenile seagrass-
associated taxa


Independent Variable


Seagrass 100 m (+)
Seagrass 250 m (+)
Seagrass 500 m (+)
Seagrass 1 km (+)
Seagrass 100 m (+)
Seagrass 250 m (+)
Seagrass 500 m (+)
Seagrass 1 km (+)
Seagrass 100 m (+)
Seagrass 250 m (+)
Seagrass 500 m (+)
Seagrass 1 km (+)
Seagrass 100 m (+)
Seagrass 250 m (+)
Seagrass 500 m (+)
Seagrass 1 km (+)
Seagrass 100 m (+)
Seagrass 250 m (+)
Seagrass 500 m (+)
Seagrass 1 km (+)
Seagrass 100 m (+)
Seagrass 250 m (+)
Seagrass 500 m (+)
Seagrass 1 km (+)
Seagrass 100 m (+)
Seagrass 250 m (+)
Seagrass 500 m (+)
Seagrass 1 km (+)
Seagrass 100 m (+)
Seagrass 250 m (+)
Seagrass 500 m (+)
Seagrass 1 km (+)


F-ratio Model R2
(x 100)


3.82
4.85
5.34
5.03
21.38
23.36
19.84
21.27
3.52
5.29
6.06
5.56
5.34
4.48
3.20



3.10
4.73


3.42
5.63
7.14


17.26
13.87
9.92


16.06
19.50
21.08
20.92
54.29
56.48
52.43
55.59
16.37
22.72
25.20
22.63
21.08
18.65
13.81



14.70
20.81


14.59
21.96
26.30


46.32
40.05
33.14


p-value


0.064*
0.0390
0.0300
0.0370
0.0002
0.0001
0.0003
0.0002
0.076*
0.0330
0.0240
0.0290
0.0300
0.0400
0.080*
NS
NS
0.090*
0.0400
NS
NS
NS
NS
NS
0.079*
0.0280
0.0150


0.0005
0.0013
0.0051


*Models with p-values less than 0.15 are shown. a p > 0.05 level. NS means not
significant at the p < 0.15. b Bold represents the spatial extent with the strongest
relationship.









Table 3-5. Simple linear regression of abundances of the adult and juvenile components
for grunts, snappers, groupers, and seagrass-associated taxa with the areal
coverage of seagrass within 1 km of each study reef at the 22 study reefs,
sampled in 2002 in St. John, U.S. Virgin Islands.


Dependent
variable


MIFs


Grunts
(Haemulids)


Snappers
(Lutjanids)


Groupers
(Epinephelines)


Seagrass-
associated taxa



Resident taxa




Mobile taxa




Transient taxa


Independent Variable


Seagrass 100m (+)
Seagrass 250 m (+)b
Seagrass 500 m (+)
Seagrass 1 km (+)
Seagrass 100 m (+)
Seagrass 250 m (+)
Seagrass 500 m (+)
Seagrass 1 km (+)b
Seagrass 100 m (+)
Seagrass 250 m (+)b
Seagrass 500 m (+)
Seagrass 1 km (+)
Seagrass 100 m (+)
Seagrass 250 m (+)
Seagrass 500 m (+)
Seagrass 1 km (+)b
Seagrass 100 m (+)
Seagrass 250 m (+)b
Seagrass 500 m (+)
Seagrass 1 km (+)

Seagrass 100 m (+)
Seagrass 250 m (+)
Seagrass 500 m (+)
Seagrass 1 km (+)
Seagrass 100 m (+)
Seagrass 250 m (+)b
Seagrass 500 m (+)
Seagrass 1 km (+)
Seagrass 100 m (+)
Seagrass 250 m (+)
Seagrass 500 m (+)
Seagrass 1 km (+)


F ratio


9.50
11.13
10.01
6.88
20.38
23.47
19.27
22.84
10.26
11.82
9.42
8.83
2.57
3.71
2.67
3.57
16.05
14.93
11.21
10.35


5.76



4.79
4.86

4.47


Model R2
(xlOO)
32.23
41.25
33.36
26.60
53.10
56.59
51.66
57.32
36.29
39.64
34.35
34.17
12.51
17.09
12.93
17.36
44.52
50.12
35.01
33.42


22.36



20.00
20.34

19.07


p-value


0.0058
0.0033
0.0049
0.0160
0.0003
0.0001
0.0001
0.0002
0.0049
0.0029
0.0066
0.0086
0.120a
0.070 a
0.120 a
0.070 a
0.0007
0.0002
0.0032
0.0030
NS
0.0260
NS
NS
0.0406
0.0390
NS
0.0478
NS
NS
NS
NS


* Models with p-values less than 0.15 are shown. a p > 0.05 level. NS means not
significant at the p < 0.15. b Bold represents the spatial extent with the strongest
relationship.









Table 3-6. Simple linear regression analyses of cumulative species richness of MIFs,
haemulids, epinephelids, lutjanids and within resident, mobile and transient
mobility guilds the areal coverage of seagrass within 1 km of each study reef
as the independent variable at the 22 study reefs, sampled in 2002 in St.

Dependent Independent Variable F ratio Model R2 p-value
variable (x 100)
Cumulative Seagrass 100m (+) 46.47 72.29 0.0001
species richness Seagrass 250 m (+) 63.59 68.59 0.0001
of Seagrass 500 m (+) 45.84 71.81 0.0001
MIFs Seagrass 1 km (+) 28.83 62.91 0.0001
Cumulative Seagrass 100 m (+) 49.89 71.38 0.0001
species richness Seagrass 250 m (+) 45.46 69.44 0.0001
of Seagrass 500 m (+) 25.36 55.91 0.0001
Haemulids Seagrass 1 km (+) 28.59 60.07 0.0001
Cumulative Seagrass 100 m (+) 12.63 38.73 0.0020
species richness Seagrass 250 m (+) 14.09 41.33 0.0013
of Seagrass 500 m (+) 8.22 29.12 0.0095
Lutjanids Seagrass 1 km (+) 6.06 24.19 0.0235
Cumulative Seagrass 100 m (+) 8.77 32.76 0.0084
species richnes Seagrass 250 m (+) 9.06 33.49 0.0075
of Seagrass 500 m (+) 7.19 28.57 0.0152
Epinephelids Seagrass 1 km (+) 5.08 23.00 0.0377
Cumulative Seagrass 100 m (+) --- --- NS
species richness Seagrass 250 m (+) --- --- NS
of Seagrass 500 m (+) --- --- NS
Resident Taxa Seagrass 1 km (+) --- --- NS
Cumulative Seagrass 100 m (+) 29.79 59.83 0.0001
species richness Seagrass 250 m (+) 43.68 68.59 0.0001
of Seagrass 500 m (+) 34.31 63.17 0.0001
Mobile Taxa Seagrass 1 km (+) 23.51 55.30 0.0001
Cumulative Seagrass 100 m (+) --- --- NS
species richness Seagrass 250 m (+) --- --- NS
of Seagrass 500 m (+) 6.21 25.65 0.0220
Transient Taxa Seagrass 1 km (+) --- --- NS
Cumulative Seagrass 100 m (+) --- --- NS
species richness Seagrass 250 m (+) --- --- NS
of Seagrass 500 m (+) 6.21 25.65 0.0220
Transient Taxa Seagrass 1 km (+) --- --- NS


John,US Virgin Islands.
* Models with p-values less than 0.15 are


shown. ap > 0.05 level. NS means not


significant at the p < 0.15. bBold represents the spatial extent with the strongest
relationship.









Table 3-7. Spearman rank correlations of relationships of each fish parameter and the
areal coverage of seagrass at the 250 m spatial extent for the 8 study reefs in
St. John, sampled in 2002 and 2003. For each relationship, the Spearman Rho
correlation, and probability < Rho are presented by year. Kruskal-Wallis tests
for significant differences in abundance using raw census data for 2002 (n =
43) and 2003 (n = 87) between years are presented, with direction of change
indicating an increase (+) or decrease (-) in abundance. Relationships in bold
are those that are consistent between years. All p-values < 0.20 are presented.


Reef fish parameter


Cumulative richness
Mean species richness
Total abundance
Abundances ofMIFs
Abundances of adult MIFs
Abundances of juvenile
MIFs
Abundances of adult
haemulids
Abundances of juvenile
haemulids
Abundances of adult
epinephelids
Abundances of juvenile
epinephelids
Abundances of adult
lutjanids
Abundances of juvenile
lutjanids
Cumulative richness MIFs
Cumulative richness
haemulids
Cumulative richness
lutjanids
Cumulative richness
epinephelids


2002
Rho
66.21
59.21
59.21
61.23
52.78
52.78


Prob
Rho
0.07
0.11
0.11
0.09
0.18
0.18


2003
Rho
56.63


Prob
Rho
0.100


57.14 0.130
57.14 0.130
--- NS
--- NS
--- NS


-- -- -- NS


88.35 0.004 80.95 0.015


68.45 0.06


--- NS


82.47 0.01 76.19 0.009

-- -- NS


50.36 0.20


84.09


75.16


--- NS


Kruskal-
Wallis


0.186
0.016
0.137
0.386


0.022


0.690


0.268


0.292


0.012


0.847


0.001 93.86 0.001
0.03 --- NS


NS


73.03 0.03 84.01 0.001


Direction
ofchange






78


Table 3-8. Influence of fine-scale (rugosity) and landscape-scale (seagrass) features in
predicting reef fish parameters for the 22 study reefs sampled in 2002.
Dependent variable Independent Partial Model p-value
Variable R2 R2


Cumulative richness


Mean species richness


Total abundance


Abundances of MIFs


Abundances of Haemulids


Abundances of Epinephelids


Abundances of Lutjanids


Abundances of Resident taxa


Abundances of
Mobile taxa
Species richness of MIFs


Species richness of Haemulids


Species richness of Epinephelids


Species richness of Lutjanids

Species richness of Resident
taxa
Species richness of Mobile taxa


Seagrass 250 m
Rugosity
Seagrass 250 m
Rugosity
Seagrass 250 m
Rugosity
Seagrass 250 m
Rugosity
Seagrass 250 m
Rugosity
Seagrass 250 m
Rugosity
Seagrass 250 m
Rugosity
Seagrass 250 m
Rugosity
Seagrass 250 m
Rugosity
Seagrass 250 m
Rugosity
Seagrass 250 m
Rugosity
Seagrass 250 m
Rugosity
Seagrass 250 m
Rugosity
Seagrass 250 m
Rugosity
Seagrass 250 m
Rugosity


71.76


35.67


9.60
33.10
40.36


56.59


17.09


39.64



28.92
33.79
12.89
69.35


69.98


19.97


41.34

25.41


69.37


Models with p-values less than 0.15 are shown. An asterisks (*) designates where p >
0.05 level. NS means not significant at the p < 0.15. Relationships in bold indicate where
rugosity was a significant predictor variable in the final model.


71.76


35.67


42.70


40.36


56.59


17.09


39.64


28.92


46.58


69.35


69.98


19.97


41.34

25.41


< 0.0001


0.0054


0.0033


0.0026


< 0.0001


0.070*


0.0029


0.0200


0.0072


< 0.0001


< 0.0001


0.0400


0.0013

0.0200


69.37 <0.0001














CHAPTER 4
REEF FISHES RESPOND TO VARIATION IN LANDSCAPE STRUCTURE

Reef context explained a significant amount of the variability in reef fish

assemblage structure in the Florida Keys National Marine Sanctuary (FKNMS) and the

US Virgin Islands. Although reef context was significant in both systems, the specific

measure of context varied. Rather, the particular habitat type (e.g., seagrass, pavement)

responsible for the reef fish-habitat relationships differed between systems. Although the

areal coverage of seagrass positively predicted abundances and species richness of

mobile invertebrate feeders, haemulids, and lutjanids in the US Virgin Islands, seagrass

did not explain a significant amount of the variation of these same fishes when analyses

was limited to the FKNMS only. Instead, the amount of pavement and reef habitats were

positively associated with several reef fish parameters in Florida. Differences in the

landscape structure of the two systems may explain this disparity. In the US Virgin

Islands, seagrass comprises a relatively small proportion of the essential fish habitat.

However in Florida, seagrass is the dominant habitat type, whereas pavement, which was

common in the US Virgin Islands, comprises a smaller proportion of the essential fish

habitat. Thus, the processes that structure reef fish communities appear to respond to

variation in the landscape structure of these coral reef environments. The shape of reef

fish-seagrass curves for pooled data including the FKNMS and US Virgin Islands

suggest that there may be critical threshold of seagrass habitat. Once this critical

threshold of seagrass is exceeded, seagrass may become less important in structuring reef

fish communities than other habitat types. These results are relevant to marine protected









areas design, since they suggest that general design rules do not necessarily apply across

systems. Comparative studies such as this are critical for developing the universal design

principles required to establish marine protected areas that meet their conservation and/or

fisheries objectives.

Introduction

Landscape-scale metrics, traditionally employed in the study of terrestrial systems,

have been applied recently to the study of coral reef ecosystems to better understand fish-

habitat relationships (Grober-Dunsmore et al. 2004a, 2004b, 2004c and Kendall et al.

2003 in the US Virgin Islands; Christensen et al. 2003 in Puerto Rico, Appeldoorn et al.

2003 in Columbia, Jeffrey 2004 in the Florida Keys). Coral reef ecosystems, like most

terrestrial ecosystems, can be described quantitatively by the spatial pattern or

arrangement of landscape elements (Forman 1995, Dramstad et al. 1996), including

patches of habitat and corridors of movement. Ault and Johnson (1998a, 1998b),

Appeldoorn et al. (2003), Kendall et al. (2003) and Grober-Dunsmore et al. (2004a,

2004b, 2004c) have all used a landscape-scale approach to demonstrate significant fish-

habitat relationships, and generally conclude that the context of individual habitat patches

may be a critical factor determining reef fish community structure. In the latter case,

simple measures of reef context (calculated with GIS software) were used to predict reef

locations with high reef fish abundances and reef fish diversity in the US Virgin Islands

(Grober-Dunsmore et al. 2004a, 2004b). Mean abundance and species richness of several

groups of exploited reef fishes, in particular, showed strong positive correlations with

those reefs in close proximity to seagrass habitat (Grober-Dunsmore et al. 2004c). The

generality of these findings, however, have not yet been tested.









To assist in the development of an emerging, coherent landscape-scale theory that

explains the structure of reef fish communities, it is essential to critically examine

observed patterns of reef fish abundance, distribution, species richness and diversity, and

determine which landscape elements, if any, best describe these patterns. Furthermore, it

is necessary to explore how reliably these landscape elements predict assemblage

structure in other systems, and under differing environmental conditions. Studies of

landscape structure effects on fauna in terrestrial habitats (Paton 1994, Trzcinski et al.

1999, Villard et al. 1999) often yield markedly different results, and, as a consequence,

generalizations regarding these relationships have been slow to evolve. Inconsistencies

in relational patterns can occur due to discrepancies in the scale of observation (Hewitt et

al. 1998, Wiens and Milne 1989), the natural heterogeneity of ecosystems (Kolasa and

Pickett 1991), disparity in the metrics used to measure spatial patterning (Frohn 1998),

differences in sampling techniques and in the spatial resolution and grain of the benthic

habitat maps. In the pursuit of identifying general operational guidelines for MPA

design, it will be mandatory to replicate experiments in time and space (Sale 2002), to

understand how variation in the structure of underlying habitats may influence the

composition, structure and distribution of reef fish communities.

Faunal abundance and survivorship of specific organisms in marine systems has

been associated with landscape configuration (Robbins and Bell 1994, Irlandi et al. 1999,

Hovel and Lipcius 2001), landscape context (Bell et al. 2001, Appeldoorn et al. 2003,

Grober-Dunsmore et al. 2004a, 2004b, 2004c), and reef connectivity (Ault and Johnson

1998), although few of these studies have been replicated in space, limiting our ability to

draw general conclusions. The present study was carried out in the Florida Keys









National Marine Sanctuary (FKNMS), and was designed to test the generality of reef

fish-habitat relationships previously detected in the US Virgin Islands (Grober-

Dunsmore et al. 2004c). FKNMS represents one of the few other Caribbean locations

with digitally-referenced benthic habitat maps, and was therefore amenable to applying

the same landscape approach and methods as that used in the previous study. FKNMS

was selected for several reasons. First, the FKNMS habitat maps were created jointly by

Florida Marine Research Institute (FMRI) and National Oceanographic and Atmospheric

Association (NOAA). Because the habitat maps for the US Virgin Islands were also

created by NOAA, I assumed some consistency in methods throughout the classification

and interpretation process. Furthermore, the grain, which is the smallest resolvable unit

of study (King 1991) and the extent, which is the area over which observations are made

(Morrison and Hall 2001) were comparable between studies. This is important to my

comparisons, since interpretation of how ecological systems are structured often depends

on the spatial and temporal scale at which a study is conducted. In fact, results of studies

conducted at different scales may not be comparable (Osenberg et al. 1999). Second,

because reef fish populations in the US Virgin Islands are heavily exploited and several

targeted fish groups (e.g., groupers) exhibit low population densities (Rogers and Beets

2001), many reef fish-habitat relationships were hypothesized to be stronger in less-

fished systems, such as no take areas where abundances of reef fish populations are

expected to be higher. The designation of a network of special protected areas (SPA) in

the FKNMS afforded the opportunity to sample inside no fishing areas (where reef fish

populations were expected to be less fished than the US Virgin Islands) and to compare

reef fish-habitat relationships inside and outside of no-take areas. Finally, it is









preferential to test the generality of landscape relationships in systems where the habitat

patches are arranged differently, to determine the robustness of landscape variables as

predictors among locations (Bissonette and Storch 2003). The US Virgin Islands is an

insular fringing reef system with a highly variable distribution of habitat patches

comprised of seagrass, reef, sand, and pavement interspersed in no particular pattern from

shore. In comparison, the FKNMS is a continental bank reef system with a more

homogeneous distribution of habitat patches. For example, the continuous bank system

of shallow spur and groove reef is 6-10 kilometers offshore, and separated from the coast

by large expanses of seagrass habitat, interspersed with patch reefs. If reef fish-habitat

relationships are similar and robust across these two disparate ecosystems, i.e. the US

Virgin Islands and FKNMS, patterns might be expected to hold across other Caribbean

coral reef systems.

The objective of this study was to determine how robust reef fish-habitat

relationships are to variation in landscape structure, since the expectation that

relationships will be similar across systems assumes that the processes that gave rise to

these relationships are not modified by variation in landscape structure. The primary

predictions, based on my findings in the US Virgin Islands, were that reef fish

assemblage structure would not be correlated with reef configuration (e.g., reef size,

shape), complex landscape-scale (e.g., habitat diversity) or fine-scale (rugosity) metrics

of habitat heterogeneity, yet reef fishes would be positively correlated with reef context,

i.e. the spatial arrangement and composition of surrounding habitat patches. Specifically,

I predicted that species richness and abundances of targeted fishes (i.e. mobile

invertebrate feeders, haemulids, lutjanids, and serranids) would be higher at reefs