<%BANNER%>

Freshwater River Turtle Populations Influenced by a Naturally Generated Thermal Gradient

Permanent Link: http://ufdc.ufl.edu/UFE0022756/00001

Material Information

Title: Freshwater River Turtle Populations Influenced by a Naturally Generated Thermal Gradient
Physical Description: 1 online resource (150 p.)
Language: english
Creator: Pitt, Amber
Publisher: University of Florida
Place of Publication: Gainesville, Fla.
Publication Date: 2008

Subjects

Subjects / Keywords: climate, community, geographica, graptemys, growth, population, sternotherus, temperature, trachemys, turtle, variation
Interdisciplinary Ecology -- Dissertations, Academic -- UF
Genre: Interdisciplinary Ecology thesis, Ph.D.
bibliography   ( marcgt )
theses   ( marcgt )
government publication (state, provincial, terriorial, dependent)   ( marcgt )
born-digital   ( sobekcm )
Electronic Thesis or Dissertation

Notes

Abstract: The North Fork of White River (NFWR), Ozark County, Missouri, receives a large volume of water from several major springs. These springs create a temperature gradient along the stream's length. Historical studies dating back to 1969 of a 4.6 km section of NFWR reveal that the North Fork supports a diverse turtle community, predominated by the northern map turtle (Graptemys geographica). The 4.6 km research section from which historical data are available is located downstream of the major spring-flows and thus represents a relatively cooler portion of the stream for much of the year, but especially during the summer months. Because rivers, especially those that are spring-fed, such as NFWR, are spatially and temporally dynamic, it was hypothesized that turtle communities may differ within various sections of the same river. Additionally, because temperature can influence turtles' physiology, it was hypothesized that turtles in the upstream, and therefore relatively warmer, sections of NFWR, may grow faster than those in areas downstream of the major spring-flows. Assessment of turtle populations and communities located in thermally distinct areas of NFWR indicated that community composition differed within relatively close sections of the same river. Furthermore, the turtle community located in a less impacted (by humans) section of NFWR more closely resembled the community described in 1969 surveys in an area that is now degraded. These results indicated that thermal regimes, habitat quality, and microhabitat availability influence chelonian community composition. Significant temperature differences within NFWR did not appear to influence turtle growth rate, though body temperature and behavior of G. geographica varied with thermal habitat. Growth rate did vary with time and length of growing season. These patterns indicate that global climate change may influence turtle growth rate. Von Bertalanffy growth curves were generated for male and female G. geographica inhabiting NFWR, thus providing information relevant to life history characteristics that can be compared with data from other populations or taxa.
General Note: In the series University of Florida Digital Collections.
General Note: Includes vita.
Bibliography: Includes bibliographical references.
Source of Description: Description based on online resource; title from PDF title page.
Source of Description: This bibliographic record is available under the Creative Commons CC0 public domain dedication. The University of Florida Libraries, as creator of this bibliographic record, has waived all rights to it worldwide under copyright law, including all related and neighboring rights, to the extent allowed by law.
Statement of Responsibility: by Amber Pitt.
Thesis: Thesis (Ph.D.)--University of Florida, 2008.
Local: Adviser: Nickerson, Max A.
Electronic Access: RESTRICTED TO UF STUDENTS, STAFF, FACULTY, AND ON-CAMPUS USE UNTIL 2010-12-31

Record Information

Source Institution: UFRGP
Rights Management: Applicable rights reserved.
Classification: lcc - LD1780 2008
System ID: UFE0022756:00001

Permanent Link: http://ufdc.ufl.edu/UFE0022756/00001

Material Information

Title: Freshwater River Turtle Populations Influenced by a Naturally Generated Thermal Gradient
Physical Description: 1 online resource (150 p.)
Language: english
Creator: Pitt, Amber
Publisher: University of Florida
Place of Publication: Gainesville, Fla.
Publication Date: 2008

Subjects

Subjects / Keywords: climate, community, geographica, graptemys, growth, population, sternotherus, temperature, trachemys, turtle, variation
Interdisciplinary Ecology -- Dissertations, Academic -- UF
Genre: Interdisciplinary Ecology thesis, Ph.D.
bibliography   ( marcgt )
theses   ( marcgt )
government publication (state, provincial, terriorial, dependent)   ( marcgt )
born-digital   ( sobekcm )
Electronic Thesis or Dissertation

Notes

Abstract: The North Fork of White River (NFWR), Ozark County, Missouri, receives a large volume of water from several major springs. These springs create a temperature gradient along the stream's length. Historical studies dating back to 1969 of a 4.6 km section of NFWR reveal that the North Fork supports a diverse turtle community, predominated by the northern map turtle (Graptemys geographica). The 4.6 km research section from which historical data are available is located downstream of the major spring-flows and thus represents a relatively cooler portion of the stream for much of the year, but especially during the summer months. Because rivers, especially those that are spring-fed, such as NFWR, are spatially and temporally dynamic, it was hypothesized that turtle communities may differ within various sections of the same river. Additionally, because temperature can influence turtles' physiology, it was hypothesized that turtles in the upstream, and therefore relatively warmer, sections of NFWR, may grow faster than those in areas downstream of the major spring-flows. Assessment of turtle populations and communities located in thermally distinct areas of NFWR indicated that community composition differed within relatively close sections of the same river. Furthermore, the turtle community located in a less impacted (by humans) section of NFWR more closely resembled the community described in 1969 surveys in an area that is now degraded. These results indicated that thermal regimes, habitat quality, and microhabitat availability influence chelonian community composition. Significant temperature differences within NFWR did not appear to influence turtle growth rate, though body temperature and behavior of G. geographica varied with thermal habitat. Growth rate did vary with time and length of growing season. These patterns indicate that global climate change may influence turtle growth rate. Von Bertalanffy growth curves were generated for male and female G. geographica inhabiting NFWR, thus providing information relevant to life history characteristics that can be compared with data from other populations or taxa.
General Note: In the series University of Florida Digital Collections.
General Note: Includes vita.
Bibliography: Includes bibliographical references.
Source of Description: Description based on online resource; title from PDF title page.
Source of Description: This bibliographic record is available under the Creative Commons CC0 public domain dedication. The University of Florida Libraries, as creator of this bibliographic record, has waived all rights to it worldwide under copyright law, including all related and neighboring rights, to the extent allowed by law.
Statement of Responsibility: by Amber Pitt.
Thesis: Thesis (Ph.D.)--University of Florida, 2008.
Local: Adviser: Nickerson, Max A.
Electronic Access: RESTRICTED TO UF STUDENTS, STAFF, FACULTY, AND ON-CAMPUS USE UNTIL 2010-12-31

Record Information

Source Institution: UFRGP
Rights Management: Applicable rights reserved.
Classification: lcc - LD1780 2008
System ID: UFE0022756:00001


This item has the following downloads:


Full Text

PAGE 1

FRESHWATER RIVER TURTLE POPULATI ONS INFLUENCED BY A NATURALLY GENERATED THERMAL GRADIENT By AMBER L. PITT 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 2008 1

PAGE 2

2008 Amber L. Pitt 2

PAGE 3

To my father, Albert L. Pitt, Sr., for showing me the wonders of rivers 3

PAGE 4

ACKNOWLEDGMENTS I would like to extend my gratitude to the vari ous people and organizatio ns that have aided my research. I thank my committee chair, Max A. Nickerson, for his support in completing this research. Without his guidance and generosity, this project would not have been possible. I am grateful to my committee members, Don Moll, J. Perran Ross, Mel E. Sunquist and Kent A. Vliet for their guidance, support and valuable advice. I feel extremely lucky and privileged to have had the opportunity to learn from each of these people. I extend my gratitude to the Saint Louis Z oological Park and the Reptile and Amphibian Conservation Corps (RACC) for providing funding, Sunburst Canoe Ranch for providing a campsite and use of its facilities, and the Wild Branch Fly Shop for providing a boat. Without the generosity and support of these institutions an d businesses, this research would not have been possible. I thank Joseph J. Tavano for his invaluab le field assistance, understanding, and good humor throughout this process. I could not have completed this process without his support. I thank Matthew J. Collins for his assistance with computer programming. I am grateful to the Missouri Department of Conservation (MDC) for assistance with permits. I thank Stephen Humphrey, Cathy Ritchie, and Meisha Wade for their assistance and support, especially during my last year of the program. I am grateful and lucky to have had a supportive department chair and staff during good and not-so-good times. I am grateful to Ron and Karen Goellner, Je ff Briggler, Justin and Amy Spencer, Connie Morgan, Jeff Ettling, Eric Miller, Mark Wanne r, and Chawna Schuette for their support and hospitality. Each of these peopl e assisted in turning my field site into my home away from home as they each became part of my Missouri family. 4

PAGE 5

Lastly, I would like to thank my parents, Albert and Loretta Pitt; my grandparents, Lillian and Harold Farr and Virginia Pitt; and other loved ones, especially Joseph Tavano, Matthew Collins, Amy Mahar, Jennifer Lehneman, Rachel Schwab, Jamie Rowder, and Kathryn Clements for the support and encouragement that they have given me throughout this experience. I could not have accomplished any of this without thes e people and I am foreve r grateful to them. 5

PAGE 6

TABLE OF CONTENTS page ACKNOWLEDGMENTS ...............................................................................................................4 LIST OF TABLES ...........................................................................................................................8 LIST OF FIGURES .......................................................................................................................10 LIST OF ABBREVIATIONS ........................................................................................................12 ABSTRACT ...................................................................................................................................13 CHAPTER 1 INTRODUCTION................................................................................................................. .15 Background .............................................................................................................................15 The Northern Map Turtle, Graptemys geographica (Le Sueur) .............................................17 Graptemys geographica in Missouri ......................................................................................23 North Fork of White River, Ozark County, Missouri .............................................................24 Objectives ...............................................................................................................................26 2 FRESHWATER TURTLE POPULATIONS AND COMMUNITY COMPOSITION IN TWO THERMALLY DISTINCT SECTIONS OF A RIVER...............................................31 Introduction .............................................................................................................................31 Materials and Methods ...........................................................................................................34 Results .....................................................................................................................................39 Discussion ...............................................................................................................................47 Conclusions .............................................................................................................................57 3 COMPARISON OF INSTANTANEOUS GROWTH RATES OF TEMPORALLY AND SPATIALLY DISTINCT Graptemys geographica POPULATIONS INHABITING A RIVER WITH A NATURALLY GENERATED THERMAL GRADIENT....................................................................................................................... .....75 Introduction .............................................................................................................................75 Materials and Methods ...........................................................................................................77 Results .....................................................................................................................................82 Discussion ...............................................................................................................................86 Conclusions .............................................................................................................................90 4 GROWTH CURVE ESTIMATIONS FOR Graptemys geographica INHABITING THE NORTH FORK OF WHITE RIVE R, OZARK COUNTY, MISSOURI.....................101 Introduction ...........................................................................................................................101 6

PAGE 7

Materials and Methods .........................................................................................................102 Results ...................................................................................................................................108 Discussion .............................................................................................................................110 Conclusions ...........................................................................................................................113 5 CONCLUSIONS AND RECOMMENDATIONS...............................................................123 Conclusions ...........................................................................................................................123 Temporal and Spatial Variation in Abio tic Factors Influences River Turtle Community Composition, Species Richness, and Heterogeneity. .............................123 The North Fork of White River, Ozark County, Missouri and Associated Springs (i.e., Recharge Areas) Are Contaminated with Coliform Bacteria. ...........................126 Global Climate Change May Be Enhanc ing Growth Rates of River Turtles Inhabiting Temperate Climates by Exte nding the Growing Season and May Be Leading to Earlier Age at Maturi ty and Age at Size Benefits. ..................................127 Recommendations.................................................................................................................129 Assess Turtle Communities for Shifts from Predominately Specialist Species to More Generalist Species. ...........................................................................................129 Assess Competitive Interactions and Ab ility between Generalist and Specialist Turtle Species. ............................................................................................................130 Conduct a Thorough Assessment of Water Quality in NFWR. ....................................131 Conduct a Comprehensive Assessment of Global Warmings Affects on River Turtles. .......................................................................................................................132 Assess the Relative Importance of Temper ature versus Length of Growing Season on Turtle Growth. .......................................................................................................133 Continue Monitoring the Turtle Comm unity in NFWR to Provide Insight Regarding the Continued Impacts of Clim atic Variation and Habitat Alterations on Turtle Communities. .............................................................................................134 Final Remarks .......................................................................................................................135 LIST OF REFERENCES .............................................................................................................136 BIOGRAPHICAL SKETCH .......................................................................................................150 7

PAGE 8

LIST OF TABLES Table page 2-1 Heterogeneity measures for turtle communities in the North Fork of White River, Ozark County, Missouri. ....................................................................................................59 2-2 Summary of turtle behavior and basking substrate use prior to capture in the North Fork of White River, Ozark County, Missouri. .................................................................60 2-3 Basking substrate niche breadth for turtle species in the North Fork of White River, Ozark County, Missouri. ....................................................................................................61 2-4 Piankas niche overlap index for turtle communities in the North Fork of White River, Ozark County, Missouri. .........................................................................................62 2-5 Mean water depth in which the turtles or their basking s ubstrates were located prior to capture in the North Fork of White River, Ozark County, Missouri. ............................63 2-6 Mean water depth in which the turtles or basking substrates were located prior to capture in the North Fork of Wh ite River, Ozark County, Missouri. ................................64 2-7 Tukeys post-hoc procedur e results for difference in mean water depth use by Graptemys geographica Sternotherus odoratus and Trachemys scripta elegans, partitioned by sex, in the downstream secti on of the North Fork of White River, Ozark County, Missouri in 2005 to 2007. ..........................................................................65 2-8 Schumacher-Eschmeyer population size and corresponding density estimates of Graptemys geographica in 2005, 2006, and 2007 in the North Fork of White River, Ozark County, Missouri. ....................................................................................................66 2-9 Mean plastron lengths of Graptemys geographica in 2005, 2006, and 2007 in the North Fork of White River, Ozark County, Missouri. .......................................................67 3-1 Temperatures of Graptemys geographica surface water, and air observed in the two research sections of the North Fork of White River, Ozark County, Missouri. ................92 3-2 Instantaneous growth rates of Graptemys geographica in the downstream section of North Fork of White River, Ozark County, Missouri for years 1969 1972 and 2004 2007.................................................................................................................................93 3-3 Instantaneous growth rate of Graptemys geographica in the downstream and upstream sections of the North Fork of White River, Ozark County, Missouri for years 2005 2007. .............................................................................................................94 3-4 Behavioral observations of Graptemys geographica in the downstream section of the North Fork of White River, Ozark C ounty, Missouri for years 1969 1972 and 2004 2007.................................................................................................................................95 8

PAGE 9

3-5 Behavioral observations of Graptemys geographica in the downstream and upstream sections of the North Fork of White Ri ver, Ozark County, Missouri for years 2005 2007....................................................................................................................................96 3-6 Degree and growth days for south-central Missouri. .........................................................97 3-7 Growth days of months in which Graptemys geographica are active in south-central Missouri. ............................................................................................................................98 3-8 Deviation from normal monthly mean temperature for south-central Missouri. ...............99 4-1 Von Bertalanffy growth curve model variables for Graptemys geographica estimated using nonlinear regression techniques. ............................................................................115 4-2 Summary of von Bertalanffy growth curve model variables available for Graptemys species. .............................................................................................................................116 9

PAGE 10

LIST OF FIGURES Figure page 1-1 Distribution of the northern map turtle, Graptemys geographica in North America Range map for Graptemys geographica ............................................................................29 1-2 Distribution of the northern map turtle, Graptemys geographica in Missouri.................30 2-1 Total coliform bacteria c ontent observed in the North Fork of White River, Ozark County, Missouri................................................................................................................68 2-2 Escherichia coli content observed in the North Fo rk of White River, Ozark County, Missouri.............................................................................................................................69 2-3 Turtle community structure in the down stream and upstream sections of the North Fork of White River, Ozark County, Missouri..................................................................70 2-4 Rarefaction curves for the North Fork of White River, Ozark County, Missouri, turtle communities.......................................................................................................................71 2-5 Size distribution of Graptemys geographica in the downstream and upstream sections of the North Fork of Wh ite River, Ozark County, Missouri................................73 2-6 Sex ratios of mature Graptemys geographica in the North Fork of White River, Ozark County, Missouri.....................................................................................................74 3-1 Linear regression of individual Graptemys geographica instantaneous growth rate and growth days for 1969 1972 and 2004 2007.........................................................100 4-1 Relationship between tail length and plastron length observed in the Graptemys geographica population in the North Fork of White River, Ozark County, Missouri for the sampling period 2005 2007...............................................................................117 4-2 Relationship between plastr on length and tail length/plastron length ratio observed in the Graptemys geographica population in the North Fork of White River, Ozark County, Missouri for the sampling period 2005 2007..................................................118 4-3 General growth pattern observed in the Graptemys geographica population in the North Fork of White River, Ozark County, Missouri for the sampling period 2005 2007..................................................................................................................................119 4-4 Von Bertalanffy growth curves for the Graptemys geographica population in the North Fork of White River, Ozark County, Missouri for the sampling period 2005 2007..................................................................................................................................120 10

PAGE 11

4-5 Von Bertalanffy growth curves and estimated-age data for Graptemys geographica inhabiting the North Fork of White Rive r, Ozark County, Missouri for the sampling period 2005 2007..........................................................................................................121 4-6 Von Bertalanffy growth curves for male Graptemys geographica inhabiting the North Fork of White River, Ozark County, Missouri for the sampling periods 1969 1972 and 2005 2007......................................................................................................122 11

PAGE 12

LIST OF ABBREVIATIONS CITES Convention on International Trade in Endangered Species of Wild Fauna and Flora DF Degrees of freedom IUCN International Union for the Conser vation of Nature and Natural Resources MDNR Missouri Department of Natural Resources MPN Most probable number of colony-forming units per 100 ml of water NFWR North Fork of White River NOAA National Oceanic and A tmospheric Administration USFWS United States Fish and Wildlife Service 12

PAGE 13

Abstract of Dissertation Pres ented to the Graduate School of the University of Florida in Partial Fulfillment of the Requirements for the Degree of Doctor of Philosophy FRESHWATER RIVER TURTLE POPULATI ONS INFLUENCED BY A NATURALLY GENERATED THERMAL GRADIENT By Amber L. Pitt December 2008 Chair: Max A. Nickerson Major: Interdisciplinary Ecology The North Fork of White River (NFWR), Ozar k County, Missouri, receives a large volume of water from several major springs. These sp rings create a temperature gradient along the streams length. Historical studies dating back to 1969 of a 4.6 km section of NFWR reveal that the North Fork supports a divers e turtle community, predominat ed by the northern map turtle ( Graptemys geographica ). The 4.6 km research section from which historical data are available is located downstream of the major spring-flows a nd thus represents a relatively cooler portion of the stream for much of the y ear, but especially during the su mmer months. Because rivers, especially those that are spring-fed, such as NF WR, are spatially and te mporally dynamic, it was hypothesized that turtle communitie s may differ within various sections of the same river. Additionally, because temperature can influenc e turtles physiology, it was hypothesized that turtles in the upstream, and therefore relatively wa rmer, sections of NFWR, may grow faster than those in areas downstream of the major spring-flows. Assessment of turtle populations and communitie s located in thermally distinct areas of NFWR indicated that community composition differed within rela tively close sections of the same river. Furthermore, the turtle community located in a less impacted (by humans) section of NFWR more closely resembled the community descri bed in 1969 surveys in an area that is now 13

PAGE 14

degraded. These results indicated that ther mal regimes, habitat qua lity, and microhabitat availability influence chelonian community composition. Significant temperature differences within NFWR did not appear to influence turtle growth rate, though body temperatur e and behavior of G. geographica varied with thermal habitat. Growth rate did vary with time and length of grow ing season. These patterns indicate that global climate change may influence turtle growth rate. Von Bertalanffy growth curves were generated for male and female G. geographica inhabiting NFWR, thus providi ng information relevant to life history characteristics that can be compared with data fr om other populations or taxa. 14

PAGE 15

CHAPTER 1 INTRODUCTION Background There is a general consensus that many turtle populations are declin ing worldwide (Smith 1979, Moll 1986, Kuchling 1988, Ernst et al 1994, Buhlmann and Gibbons 1997, Gibbons 1997, Haitao 2000, van Dijk et al 2000) although many species, especially those considered common (Dodd and Franz 1993:47), lack suffici ent population data to support this claim (Congdon et al. 1993, Dodd and Franz 1993, Dreslik 1998, Moll and Moll 2004). River turtles in particular are largely ignored by researchers a nd receive little public support for conservation (Moll and Moll 2004) though numerous threats to populations ex ist, including exploitation, pollution, disease, and habitat degradation, frag mentation and destruction (Ernst et al. 1994, Buhlmann and Gibbons 1997, Jacobson 1997, Ka nnan et al. 2000, Moll and Moll 2004). Turtle conservation is essential for maintain ing river ecosystems. Turtles are a major biomass component of river comm unities and connect the aquatic and terrestrial environments, bolstering nutrient cycling and energy flow (Congdon et al. 1986, Moll and Moll 2004). They maintain critical links in the food web as pr edators and scavengers (Moll and Moll 2004). Turtles at all life stages are important prey for terrestrial and semi-aquatic vertebrates such as foxes, raccoons, and otters, and for invertebrates such as fly larvae (Vogt 1981b, Cochran 1987, Moll and Moll 2004). Turtles may also be important for their roles in in terspecific mutualisms (Vogt 1979). For example, basking map turtles ( Graptemys spp.) benefit from removal of leeches by common grackles ( Quiscalus quiscula ) (Vogt 1979). In tur n, grackles benefit from the sedentary, protein-rich food source (Vogt 1979). The ecological ramificat ions of declines of certain species are unknown (Sutherland 1974, Iver son 1992), yet the impending loss of biomass, 15

PAGE 16

diminished nutrient cycling and energy flow, and simplification of food we bs associated with turtle declines have the potential to create substan tial changes in an ecosystem. Conservation of turtle species is impede d by the dearth of essential population and community data (Ernst et al. 1994, Moll and Moll 2004) This shortage of information is largely due to the sheer nature of colle cting it. Accurate assessment of turtle population status requires long-term studies on the scale of decades to acco unt for life history charac teristics such as longlife and delayed sexual maturation (Congdon et al 1993, Congdon et al 1994, Foscarini and Brooks 1997). These same life history charact eristics do not permit re duced populations to rebound from chronic disturbances, furthering the propensity for decline (Congdon and Gibbons 1990, Congdon et al 1993, Congdon et al 1994, Buhlmann and Gibbons 1997, Foscarini and Brooks 1997, Klemens 2000, Moll and Moll 2004). Ther efore, it is vital to not only conduct long-term population studies, but al so to understand the life histor y characteristics of the species of interest in order to predict how populations or species are likely to re spond to disturbances and various management actions (Frazer et al. 1991, 1993, Gibbons 1997, Heppell 1998, Hellgren et al. 2000). Despite the importance of such data, few long-term studies have been conducted or address these issues (Meylan et al 1992, Congdon et al 1993, Congdon et al 1994, Foscarini and Brooks 1997, Dreslik 1998). The studies that have been conducted have primarily focused on a single chelonian species rather than on the entire turtle assemblage and studies of turtle communities remain relatively rare (Moll 1990, Moll and Moll 2004). Without the resultant population and life history characteristics data, co nservation and management will continue to be sparse and inadequate and many species may become imperiled or more threatened (Trauth et al. 1998, Moll and Moll 2004). 16

PAGE 17

The paucity of turtle community studies may, in part, be an artifact of turtle community species composition (Bury 1979, Mo ll and Moll 2004). In a review Moll and Moll (2004) found that turtle communities were often skewed toward one or two numerous species and a smattering of rare species. Lotic habitats, for ex ample, are often dominated by map turtles (Graptemys spp.) (Bury 1979, Moll and Moll 2004). The Northern Map Turtle, Graptemys geographica (Le Sueur) The northern map turtle, Graptemys geographica is a relatively widespread, largely unstudied, diurnal sp ecies (Daigle et al 1994, Ernst et al. 1994, Fu selier and Edds 1994, Moll and Moll 2004). Graptemys geographica occur from southern Queb ec, Canada and northwestern Vermont, west of the Appalachia n Mountains to southeastern Mi nnesota and eastern Kansas, to southern Arkansas and northern Alabama (Ernst et al. 1994; Figure 1-1). Isolated populations exist in Maryland, eastern New York, and eastern and central Pennsylvani a (Ernst et al. 1994; Figure 1-1). Many accounts of this species are solely distributional (Arndt and Potter 1973, Kiviat and Buso 1977, Congdon and Gibbons 1996, Casper 1997, King et al. 1997, Walley 2000) and few historical population estimates are avai lable to reassess populati on status (Roche 2002; but see Pitt 2005). Considerable information gaps exist in the lif e history of this sexually dimorphic species (Gordon and MacCulloch 1980, Fuselier and Ed ds 1994, Roche 2002, Pitt 2005) and much of what is known of G. geographica is based on data from studies conducted in Ontario, Wisconsin, Pennsylvania, and Indiana (i.e., northern portion s of the species dist ribution) (Gordon and MacCulloch 1980, Vogt 1980a, 1981a, Iv erson 1988, Nagle et al. 2004). Graptemys geographica is a sexually dimorphic species in which females achieve much larger body proportions than males (Gordon and MacCulloch 1980, Lindeman and Sharkey 2001). Females are believed to reach sexual matu rity at approximately 10 to 13 years of age, 17

PAGE 18

while males mature in approximately four years (Vogt 1980a, Iverson 1988). The minimum plastron lengths of sexually mature females and males are estimated to be 17.5 cm and 7.5 cm, respectively (Gordon and MacCulloch 1980), tho ugh variation in minimum size of sexually mature G. geographica may occur throughout its distri bution (Gordon and MacCulloch 1980, Iverson 1988). It has been estimated that G. geographica live for at least 20 years in the wild (Ernst et al. 1994). Graptemys geographica occur in lakes, rivers, and streams with mud, rock, or gravel substrata, abundant vegetation, a nd ample sunlit basking sites of rocks and fallen trees (Gordon and MacCulloch 1980, Ernst et al. 1994, Fuselier and Edds 1994, Roche 2002). Within a habitat, G. geographica distribution may be associated with phys iological constraints, predation risk, intraand interspecific competiti on, social interactions, prey av ailability and distribution, and habitat structure (Boyer 1965, Vogt 1979, 1981a, Flaherty and Bider 1984, Pluto and Bellis 1986, Graham and Graham 1992, Daigle et al. 1 994, Ernst et al. 1994, Saumure and Livingston 1994, Fuselier and Edds 1994, Vogt and Benit ez 1997, Roche 2002). For example, Pluto and Bellis (1986) observed differential habitat utilization within a river between size classes of G. geographica Large turtles (i.e., females) occupied deep, slow-moving sections with emergent rocks and logs (Pluto and Bellis 1986). Small turtles, which include both sexes, occupied shallow, slow-moving sections with ample emergent vegetation (Pluto and Bellis 1986). In this case, habitat partitioning reflected the maximu m swimming speed associated with body size (Pluto and Bellis 1986). Larger turtles have a higher maximum swimming speed than smaller turtles and are able to navigate and rapidly move through the deeper water (Pluto and Bellis 1986). Correspondingly, habitat selection may have also reflected the relative predation risks associated with body size (Pluto and Bellis 1986). Juvenile G. geographica may use shallower 18

PAGE 19

areas to avoid large predaceous fish and are ab le to avoid terrestrial predators by concealing themselves between rocks and small logs (Pluto and Bellis 1986) or am ong emergent vegetation (personal observation). Large tu rtles are relatively immune to aquatic predation during their active season and may avoid terrest rial predators, such as raccoons, by remaining in deep water, far from shore (Pluto and Bellis 1986). Habitat partitioning may also be a mechanism for reducing competition between size classes for basking sites (Pluto and Bellis 1986), and social factors may be more important than habitat structure for basking site selection (Flaherty and Bider 1984). Basking is important to G. geographica for thermoregulation and parasite removal (Boyer 1965, Vogt 1979, Saumure and Livingston 1994) and may facilitate in terspecific mutualisms (Vogt 1979). Graptemys geographica often cluster around basking sites and ba sk communally (Flaherty and Bider 1984). Communal basking may aid in predator avoidance as G. geographica are extremely wary baskers and readily retreat into water if even one of the group is disturbe d (Daigle et al. 1994, Ernst et al. 1994, Roche 2002). In situations where basking habitat is limited, larg e turtles will displace smaller turtles (Pluto and Bellis 1986). Habitat is eventually partitioned by small turtles basking close to the shore and medium and large turt les basking midstream (Pluto and Bellis 1986). Medium and large turtles further partition bask ing sites by basking on smaller branches and the main trunk of trees, respectively, when space is limited (Pluto and Bellis 1986). Ultimately, for G. geographica to even occur in an area, basic habitat requirements must be met. In addition to containing areas for bask ing, the aquatic habitat must also have a deep pool for G. geographica to congregate in hibernacula where courtship and mating occur in spring and fall (Pluto and Bellis 1988, Graham and Graham 1992, Roche 2002). These pools must 19

PAGE 20

remain unfrozen and be deep e nough to avoid ice scouring as G. geographica are mobile on the bottom of the hibernacula (Graham and Graham 1992). Males leave the hibernacula in late spring and early summer in search of food (Pluto and Bellis 1988). Therefore, the aquatic system must provide ample prey to support them as well as the rest of the G. geographica population. Stomach content anal ysis and feeding observations have revealed a strictly aquatic diet of mollusk s, insects, crayfish, fish carrion, and trace amounts of vegetation, which was probably consumed incident ally with prey and therefore not considered a food source (Vogt 1980a, 1981a, Ernst et al. 1994, Lindeman and Sharkey 2001, Moll 1976, Roche 2002). Graptemys geographica s diet varies with prey av ailability, although this species is not abundant in areas with low mollusk densi ties (Vogt 1981a). Adult females broad heads and jaws are physically adapted for a specia lized mollusk diet (Vogt 1981a, Lindeman and Sharkey 2001), while adult males and juveniles are generally insecti vorous (Vogt 1980a, 1981a, Lindeman and Sharkey 2001). In areas of high mo llusk abundance, both genders and all size classes are primarily molluski vorous (Vogt 1981a, White and Mo ll 1992). In less prey-abundant locales, it appears that habitat and diet partitioning among genders and size classes and increased dispersal of males may be mechanisms employed to avoid competition (Pluto and Bellis 1986, 1988, Roche 2002). It has been hypot hesized that distribution of G. geographica may be more limited by prey availability than habitat structur e due to diet specialization (Vogt 1981a, Fuselier and Edds 1994, Vogt and Benitez 1997). Gravid females tend to move upstream from hibernacula toward ne sting sites (Gordon and MacCulloch 1980, Pluto and Bellis 1988), which may be a mechanism to offset the general downstream dispersal of hatchlings that occurs w ith the rivers current (Pluto and Bellis 1988). Therefore, the aquatic habitat must be located pr oximate to terrestrial ar eas that contain ample 20

PAGE 21

suitable nesting sites. Suitable nesting sites are generally located above the floodplain and are composed of soft sand or soil with good drainage and sun exposure, though sites with less desirable substrates have been used (Flahe rty and Bider 1984, Roche 2002). Sun exposure and surrounding vegetation are important elements of nesting sites as G. geographica have temperature-dependent sex determination, wi th mean incubation temperatures of 25 o C and 30.5 o C producing males and females, respectively (Gordon and MacCulloch 1980). Additionally, low temperature nest sites may be associated with fly larvae (Metoposarcophaga importuna) infestation (Vogt 1981b). If suitable nesti ng habitat is in abundance, nesting site selection may be dictated more by social factors than physical attri butes of a given area (Flaherty and Bider 1984). Graptemys geographica populations are primarily thre atened by pollution, habitat alteration and degradation, and exploitation (Ernst et al. 1994, Buhlmann and Gibbons 1997, Roche 2002). Pollution and alterations to the aqua tic environment can cause a reduction of prey and basking site abundance (Gibbons 1997, Li ndeman 1998, Moll and Moll 2004). Mollusks, the preferred prey of G. geographica are extremely pollution-sensi tive and can be suffocated by increased siltation associated with upland farming, urban runoff, and development (Ellis 1936, Dodd 1978, Ernst et al. 1994, Burkhead et al. 1997, Moll and Moll 2004). Chemical pollutants that do not eradicate prey may threaten G. geographica due to prey contamination and subsequent bioaccumulation (Kannan et al. 2000 Roche 2002). Prey abundance may also be compromised by hydrological changes created by impoundments or channelization (Moll and Moll 2004). Graptemys geographica overall health and reproductive health in particular may be compromised by snag removal that is often associ ated with river clean-u p activities as fewer 21

PAGE 22

basking sites will be availa ble to facilitate thermoregul ation (Buhlmann and Gibbons 1997, Roche 2002). Increased recreational use of an area may also effectively alter habitat by deterring use of structurally suitable basking sites (Gordon and MacCulloch 1980, Roche 2002). Nesting sites can be destroyed by alterations to both the aquatic and terrestrial habitats. Impoundments may submerge upstream nesting areas (Roche 2002) forcing females to nest in suboptimal habitats. Suboptimal habitats includ e those with low sun exposure, an abundance of vegetation, highly compacted soil, improper water drainage, or location within a floodplain (Flaherty and Bider 1984, Roche 2002). Low in cubation temperatures associated with inadequate sun exposure and abunda nt vegetation may result in skewed sex ratios (Vogt and Bull 1984). Improper water drainage and occasiona l flooding can drown or cause fly larvae infestation of eggs (Vogt 1981b), decreasing nest survival and subsequent recruitment. Terrestrial development may direc tly or indirectly destroy nesting sites (Roche 2002). Increased edge associated with development can increase p opulations of predators such as foxes, coyotes, raccoons, skunks, and opossums (Cochran 1987). Road s that must be crossed to reach nesting sites increase mortality of gravid females (Casper 1997, Roche 2002). Exploitation is increasingly threatening G. geographica populations (Roche 2002). Historically, G. geographica meat was used to dilute terrapin meat (Arndt and Potter 1973). Currently, G. geographica are harvested for subsistence us e (Moll and Moll 2004) and they are traded domestically to biological supply compan ies and internationally to pet and food markets (Moll and Moll 2004). From 1989 to 1993, total reported exports of Graptemys species increased from 73 to 37,233 individuals per y ear (Roche 2002). From 1998 to 2002, total reported exports of wild-caught Graptemys species consisted of 95,069 individuals (Schlaepfer et 22

PAGE 23

al. 2005). Though these data are not specific to G. geographica prevalence of G. geographica in the pet trade has been documented (Buhlmann and Gibbons 1997, Moll and Moll 2004). In 2006, G. geographica was listed as an Appendix III species in Convention on International Trade in Endangere d Species of wild fauna and flora (CITES) (U.S. Fish and Wildlife Service (USFWS) 2006), which includes sp ecies that are protec ted in at least one country, which has asked other CITES Parties fo r assistance in controlling the trade (CITES 2005). Graptemys geographica are currently protect ed throughout part of their range by state and provincial law (Buhlmann and Gibbons 1997, Ro che 2002), but has never been evaluated for International Union for the Conservation of Nature and Natural Resources (IUCN) listing (Moll and Moll 2004). Graptemys geographica in Missouri Graptemys geographica are protected by state law th roughout Missouri (Moll and Moll 2004), a state in which it is wide ly distributed (Johnson 2000). Graptemys geographica s range throughout Missouri includes all but the northern a nd southeastern portions of the state (Johnson 2000; Figure 1-2). Habitats incl ude rivers, sloughs, and oxbow lake s with gravel bottoms, some aquatic vegetation, and ample basking sites (Johnson 2000). Graptemys geographica are typically active from late March to October, but has been observed basking in December and February (Johnson 2000). Mating is limited to sp ring months, with nesting occurring from late April to early July and hatchi ng occurring in late su mmer, early autumn, or the following spring (White and Moll 1991, Johnson 2000). Graptemys geographica s reproductive potential and diet have been studied in the Niangua River, Missouri (White and Moll 1991, 1992). In Missouri, G. geographica oviposits two to three clutches yielding a mean of 23.3 eggs per female per year (White and Moll 1991). Elimia potosiensis a small gastropod, was the most abundant prey species consumed by both sexes and 23

PAGE 24

all life stages of G. geographica, comprising 94.1% % of their diet by volume, the remainder consisting of crayfish and insects (W hite and Moll 1992). Th is specialized diet resulted in failure to catch G geographica with fish-baited traps at the Niangua River site and may indicate a distribution limited by the availab ility of gastropods (White and Moll 1991). Threats to G. geographica in Missouri are similar to those in other parts of their range. In addition, the Missouri Department of Conservati on (MDC) found that shooting basking turtles for target practice was so common that they ha ve posted signs and web pages discouraging this behavior (Moll and Moll 2004). A population of G. geographica in the North Fork of White River (NFWR), Ozark County, Missouri, has been studied since 1969 (Pitt 2005), which represents the only long-term study of a northern map turtle population throughout their broad range. Pitt (2005) determined that the population had declined since 1969 and remained at this reduced size through 2004. The reduction in population size was attributed to a harv esting event or events that occurred between 1969 and 1980 (Pitt 2005). The inability of the population to rebound was reflective of factors associated with the initial loss of the large, a nd therefore the mo st fecund, females; habitat and water quality degradation; reduction of basking hab itats; increased recreational use of the river; and the establishment of red-eared sliders ( Trachemys scripta elegans Wied-Neuwied) (Pitt 2005). In association with the G. geographica population declin e and changes in aquatic habitat, the chelonian community composition changed, supporting a larger re lative proportion of generalist species (Pitt 2005). North Fork of White River, Ozark County, Missouri The North Fork of White River is a 99.8 km long river with a 1478.9 km 2 watershed (Bryant Watershed Project, Inc. 2008). The ma jority of the watershed is characterized by grassland/cropland and forest/woodland with ab out 13% consisting of Mark Twain National 24

PAGE 25

Forest lands (Bryant Watershed Project, Inc. 2008 ). The North Fork of White River receives a large volume of water from several major springs, which creates a substantial temperature gradient along the rivers le ngth (Nickerson and Mays 1973). In general, rivers are spa tially and temporally dynamic (Pluto and Bellis 1988, Moll and Moll 2004). A spring-fed river, such as NFWR, is particularly spatially variable because of the thermal gradients produced by the spring effluents. Due to the sometime s specialized habitat requirements of turtles (Ernst et al. 1994), it may be expected that turtle communities and populations differ within various s ections of the same river. Th e turtle community within the North Fork of White River, O zark County, Missouri, has been studied periodically since 1969 (Pitt 2005). Studies have largely been confin ed to a 4.6 km section of NFWR (Pitt 2005) and few data are available outside of that area. No thorough investiga tions of turtle communities in different sections of the river have been conducted. Due to the unique thermal characteristics of NFWR, it is possible to evaluate how natural thermal regimes affect turtle populations. Temperature has been implicated as a cause of variation in digestion, ingestion, and assimilation ra tes, which may lead to differences in growth rate among individuals and populations (Ellis 1936, Cagle 1946, Gatten 1974, Ernst 1975, Parmenter 1980, Thornhill 1982, Spotila et al. 1984, Brown and Brooks 1991, Frazer et al. 1993, Brown et al. 1994, Cadi and Joly 2003). Increased growth rates can lead to larger body sizes at various life stages and earlier maturation. Both factors may influence fitness via survival and reproductive enhancements (Bury 1979, Thornhill 1982, Congdon and Gibbons 1983, 1990, Cox et al. 1991, Congdon and Gibbons 1996, Lovich et al. 1998, Tucker et al. 1999, Tucker 2000). By identifying variations in growth rate, it is po ssible to identify where variation in other life history traits can occur (Frazer et al. 1991, 1993). To date, studies of variation in population 25

PAGE 26

growth rates within a river affected by a natura lly-induced thermal gradient are not available in peer-reviewed literature. NFWR presents a un ique opportunity to assess temporal and spatial variation in growth rate and associated parameters in naturally thermally-distinct areas. Objectives Determine whether turtle communities and populations located within a river with a natural thermal gradient vary temporally and spatially. Rationale: Theory has shifted focus from cla ssic ideas of stable a nd relatively static communities and populations to c oncepts describing temporally and spatially dynamic entities (Chesson and Case 1986, Pimm 1991). NFWR provides a unique oppor tunity to explore spatial and temporal variability in river turtle comm unities and populations. To date, few studies explore spatial and temporal variation of river turtle communities and populations (but see Moll 1977, Shively and Jackson 1985, Moll 1990, Huestis and Meylan 2004, Pitt 2005). Even fewer studies present data dating back several d ecades (but see Moll 1977, Huestis and Meylan 2004, Pitt 2005). In Chapter 2, I used the rarefaction me thod to examine species richness in each community. I evaluated heterogeneity of th e communities, which encompasses both species richness and evenness, using Simpsons index modified for a finite population. Population analyses were restricted to G. geographica the numerically-dominant species in each community. I calculated Schumache r-Eschmeyer population estimates for G. geographica and compared the population estimates us ing the Chapman and Overton method. Determine whether species distribution is associated with particular microhabitats. Rationale: Community composition is determin ed by abiotic and biotic factors which can be difficult to measure in natural settings (Diamond 1986, R oughgarden and Diamond 1986). In situ turtle studies in particular often rely on traps for colle ction (see Lagler 1943, Plummer 1979, 26

PAGE 27

Vogt 1980b for trap descriptions) which do not allow for behavioral or microhabitat observations. In this study, capture methodol ogy allowed for direct observation of turtle behavior and microhabitat use prio r to disturbance by surveyors. Capture techniques allowed for behavior and microhabitat elements to be categorized. I compared broad-scale river characteristics to elucidate f actors that differed among the turtle community and populations habitats. Fo r each species, I summarized the activity, water depth, and associated microhabitat structures. For species observed bask ing, I calculated Levins standardized niche breadth measure and Pia nkas niche overlap index based on basking substrate. These data are presented in Chapter 2. Determine whether instanta neous growth rates of Graptemys geographica vary temporally and spatially in a riv er with a natural thermal gradient. Rationale: Phenotypic plasticity in traits that can directly influence life history parameters is important for organisms that live in variab le ecosystems (Caswell 1983, Stearns and Koella 1986). Growth rate is a trait that can vary with the thermal envi ronment of ectotherms, such as turtles (Laudien 1973, Lillywhite et al. 1973, Gibbons et al. 1981, Bronikowski et al. 2001). Plasticity in growth rate may have important imp lications for turtles as global climate change is resulting in overall warming trends (Schle singer and Jiang 1991, Manabe and Stouffer 1993, Intergovernmental Panel on Climate Change 2007). Va riations in growth ra te can lead to sizerelated alterations in survival a nd fecundity that can last throughout a turtles life (Gibbons et al. 1981, Stearns and Koella 1986). Differences in growth rate have been observed in turtle species subjected to thermal effluents associated with power plants (Parmenter 1980, Gibbons et al. 1981, Thornhill 1982). No investigations of turtle growth rates in a river with a natural ther mal gradient have been published. Temporal variation in growth rate has been identified for northern turtle populations 27

PAGE 28

and attributed to global climate change (Frazer et al. 1991, 1993). No investigations of turtle growth rates have been reported for popul ations in the central United States. NFWR provided a unique opportunity to examine whether temporal and spatial differences in growth rates of river turtles occur. In Chapter 3, I compared instantaneous growth rates and body temperature of G. geographica inhabiting two thermally distinct sections of NFWR. I also examined differences in instantaneous growth rates for temporally distinct G. geographica populations. I investigated whether observed gr owth rate trends corre sponded with climate trends. Determine the age at maturity of G. geographica using a statistical method. Rationale: Statistically-generated growth curves based on field observations provide estimates of age at sexual maturity and highlig ht differences in life history characteristics between sexes and among populations and taxa (Frazer et al. 1990a, Lindeman 1997). In Chapter 4, I used data from previous and current studies of G. geographica in NFWR to evaluate whether the temporal di fferences in instantaneous growth rates observed translated into differing growth curves. I used the results to discuss general observe d trends and potential implications of global warming on turtle growth. In Chapter 5, I integrated information presente d in previous chapters to draw conclusions and recommendations applicable to future turtle studies both within and outside of NFWR. 28

PAGE 29

Figure 1-1. Distribution of the northern map turtle, Graptemys geographica in North America Range map for Graptemys geographica reproduced from Ernst, C. H., J. E. Lovich, and R. W. Barbour. 1994. Turtles of th e United States and Canada. Smithsonian Institution Press, Washington D.C. and London. Pp. 369. Used with permission of the author. 29

PAGE 30

Figure 1-2. Distribution of the northern map turtle, Graptemys geographica in Missouri (Johnson 2000). Shading indica tes the suspected range of G. geographica Black dots signify areas for which county records exist. Range map for Graptemys geographica reproduced from "The Amphibians a nd Reptiles of Missouri." Pp. 194. Copyright 2000 by the Conservati on Commission of the State of Missouri. Used with permission. 30

PAGE 31

CHAPTER 2 FRESHWATER TURTLE POPULATIONS A ND COMMUNITY COMPOSITION IN TWO THERMALLY DISTINCT SECTIONS OF A RIVER Introduction Theory has shifted focus from classic ideas of stable and relatively static communities and populations (i.e. equilibrium theories) to concep ts describing temporally and spatially dynamic entities (i.e., nonequilibrium th eories) (Chesson and Case 1986, Pimm 1991). This shift can largely be attributed to the idea that the environment (i.e., abiotic factors) is perpetually changing (Chesson and Case 1986). Changes in abiotic conditions can alter biotic conditions as organisms physiological tolerance thresholds ar e met and species composition turnover occurs (Davis 1986, Cody 1996). The resultant fluctuations in identity a nd relative abundance of species can alter community dyna mics and composition (Cody 1996). Abiotic and biotic factors can vary in space and time (Wiens 1984, Wiens et al. 1986, Pimm 1991, Cody 1996). Variation can occur over la rge or small scales (Wiens 1984, Wiens et al. 1986, Cody 1996). Variation can be associated w ith natural processes, such as those related to hydrology, geology, climate, competition, pred ation, or disease (Wiens 1984, Cody 1996), but can also be a result of what are typically de emed unnatural processes associated with human activity (Cody 1996). Rivers are the epitome of sp atially and temporally dynami c systems (Moll and Moll 2004, Pluto and Bellis 1988). Geologic composition ca n differ along a rivers length (Schumm 2005). Hydrological regimes can swiftly change (Sch umm 2005). Rivers can span great distances (Schumm 2005) and flow through different climatic zones. Most rivers have undergone massive alterations by humans (Benke 1990, Riccardi and Rasmussen 1999). Damming, riparian zone alterations, and pollution only sc ratch the surface of human-base d river modifications (Lydeard and Mayden 1995, Ricciardi and Rasmussen 1999). Du e to the inherent variability of rivers, 31

PAGE 32

populations and communities can change considerably along the le ngth of a river and within a small area of a river throughout time (Kinso lving and Bain 1993, Moll and Moll 2004). The North Fork of White River (NFWR), O zark County, Missouri is a 99.8 km long river with a 1478.9 km 2 watershed (Bryant Watershed Project, Inc. 2008). The majority of the watershed is characterized by grassland/cropland and fore st/woodland (Bryant Watershed Project, Inc. 2008). Approximately 13% of th e watershed consists of Mark Twain National Forest lands (Bryant Watershed Project, Inc. 2008). NFWR receives a large volume of water from several major springs (Nickerson and Mays 1973). The springs effluents create a temperature gradient along the stream s length (Nickerson and Mays 1973). Within Ozark County, the land bordering the section of NFWR upstream of the major springs is largely national fore st land and remains predominantly forested (Bryant Watershed Project, Inc. 2008). The land bordering the section of NFWR downstream of the major springs has been progressively developed for resort s and private residen ces (Pitt 2005, personal observation). The turtle community within a 4.6 km secti on of NFWR located downstream of the major springs has been studied periodically since 1969 (Pitt 2005). These studies indicate that NFWR supports a diverse turtle community pr edominated by the northern map turtle, Graptemys geographica (Le Sueur) (Pitt 2005). A decline in the G. geographica population occurred between 1969 and 1980 (Pitt 2005). A comparis on of population estimates indicated the G. geographica population observed in 2004 had not re bounded from the reduced 1980 population size (Pitt 2005). Between 1969 and 1980, a new species, the red-eared slider ( Trachemys scripta elegans Wied-Neuwied) joined the community within the research section (Pitt 2005). Changes in community composition were attributed mainly to harvesting of G. geographica (Pitt 2005). 32

PAGE 33

Confounding factors included otter reintroductions, fibrous algae blooms, a reduction in basking habitats, an increase in recreati onal use of the river, changes in the distribution of stream substrates, siltation, sedimentation, nutrient-load ing, fecal coliform contamination, pesticide inputs, and loss of large boulders (Pitt 2005). River turtle populations and commu nities, like all organisms, are influenced by a suite of abiotic and biotic factors (Moll 1990, Ernst et al. 1994, Bodie and Semlitsch 2000, Bodie et al. 2000, Moll and Moll 2004). Abiotic factors in clude hydric regimes, thermal regimes, permanence, habitat complexity, substrate co mposition, chemistry of the water body, and land use of the adjacent land (Moll 1990, Bodie et al. 2000, Moll and Moll 2004). Biotic factors encompass each species phylogeny, physical ad aptations and capabilities, life history characteristics, and habitat preferences, as well as interand intraspecific competition, predation, and resource (e.g., prey) avai lability (Moll 1990, Bodie et al. 2000, Moll and Moll 2004). Because turtle species can have specialized habi tat requirements (Ernst et al. 1994) and because rivers can vary greatly along their length (Schumm 2005), I predict that turtle communities differ among various sections of the same river. Spri ng-fed rivers, such as NFWR, are particularly spatially variable because of the thermal gradient s produced by the spring effluents. Therefore, I predict that differences in turtle communities can occur over short distances. NFWR provides a unique opportunity to explore spatial and temporal variability in river turtle communities and populations This type of data is in creasingly important as turtle populations are declining wo rldwide (Smith 1979, Moll 1986, Kuchling 1988, Ernst et al. 1994, Buhlmann and Gibbons 1997, Gibbons 1997, Haitao 2000, van Dijk et al. 2000, Moll and Moll 2004) and there is little information about the ecology, population st atus, and community composition of many turtle species and asse mblages (Moll 1990, Ernst et al. 1994, Moll and 33

PAGE 34

Moll 2004). To date, few studies explore spatia l and temporal variat ion of river turtle communities and populations (but see Moll 1977, Shively and Jackson 1985, Moll 1990, Huestis and Meylan 2004, Pitt 2005). Even fewer studies pr esent data dating back several decades (but see Moll 1977, Huestis and Meylan 2004, Pitt 2005). Without these data, it may be impossible to garner political, social, and financial support for conservation of imperiled chelonians (Gibbons 1997). It was my objective to: 1. Determine whethe r turtle communities a nd populations located within a river with a na tural thermal gradient vary in time and space and, 2. Determine if species distribution is associated with particular microhabita ts or abiotic factors. Materials and Methods I selected two research sectio ns within the NFWR based on th eir physical similarities, ease of access, and locations relative to the major spri ng-flows. The first area is the 4.6 km section studied by Nickerson and Mays (1973) and Pitt (2 005). This section is located downstream of the major springs and is relatively cool for mu ch of the year, especially during the summer months. A relatively large amount of historical data exists for this section. The second area is a 4.6 km section located upstream of the springs and is relatively warm. The upstream area is located within Mark Twain National Forest. Th e only development adjacent to the upstream study section is the boat ramp and campground operated by the Nati onal Forest ranger district. This clearing is located at the upstream end of the upstream research section. The two research sections are separated by appr oximately 16 river kilometers. I divided each of the two 4.6 km study sections into fifty 92-m-long stations, following the protocol of Nickerson and Mays (1973). In 200 5, I recorded stream width, water depth, substrate composition, and the presence or absence of aqua tic vegetation for each of the two research sections. I measured stream widths as the di stance between banks at each station marker along 34

PAGE 35

each of the 4.6 km research sections. I reco rded water depth, substrate composition, and vegetation for midstream and one meter from each bank at each station marker. I compared stream widths and depths using independent samp le t-tests to determine if mean values were significantly different between th e two sites. I summarized and compared substrate composition and vegetation presence using chi-squared test s of homogeneity. In 2007, I collected water samples within each research section and the local springs. I analyzed water samples for total coliform bacteria and Escherichia coli content using the Coliplate TM test (Bluewater Bioscience, Inc. Mississauga, ON, Canada). In 2005, 2006, and 2007, a research assistant and I surveyed the research sections on alternating days throughout the summer (J une 15 to August 20) between 0900 and 1800 h, weather permitting. The downstream section was surveyed for a total of 415 person hours. The upstream section was surveyed for a total of 351 pe rson hours. I recorded water temperatures at the beginning of each sampling day. I used an in dependent sample t-test to determine whether water temperatures were significantly different. Surveys were conducted primarily by snorkeli ng paired with mark-recapture techniques. Snorkeling is useful for capturing basking turtles, such as G. geographica that are otherwise difficult to trap. Methods that rely on tr aps (see Lagler 1943, Chaney and Smith 1950, Braid 1974, Vogt 1980b) are less effective for basking turtles (Cagle 1942, Fr azer et al. 1990b), especially G. geographica which may not respond to bait (L agler 1943) and are wary of traps (Pluto and Bellis 1986). Additionally, traps (see Lagler 1943, Chaney and Smith 1950, Vogt 1980b) are time consuming to assemble, cumbersome and difficult to use in a heavily-trafficked stream like the North Fork (Pluto and Bellis 1986). Trapping also does not allow for direct observations of behavior. Despite the difficulties and shortcomings associated with traps, I set 35

PAGE 36

two 0.75 m and two 1.0 m diameter hoop nets in ar eas in which many turtles were routinely observed once a week between 1800 and 0900 h in 2005 and 2006. Traps were set in each research section for several weeks to capture nonbasking and nocturnal species. Traps were not set during the day or during the weekends due to the heavy recreational use of the stream (Pitt 2005). I baited traps with sardines. I discontinue d the use of traps as they were ineffective. All captured turtles were weighed, measure d, and marked using nail polish, following the protocol of Pitt (2005). La rger turtles (plastron length > 8.6 cm) were also marked with a passive integrated transponder (P IT) tag (Destron-Fearing Corpor ation, So. St. Paul, MN, USA) injected into the anterior inguina l region parallel to the bridge of the shell following the protocol of Buhlmann and Tuberville (1998). I disinf ected injection sites and needles using 70% isopropyl alcohol and antibiotic ointment. I applied New-Skin Liquid Bandage (Medtech Laboratories, Inc., Jackson, WY, US A) to cover the injection hole. Previous studies from the downstream section of NFWR indi cated that both paint marks a nd PIT tag retention were high (Pitt 2005) and therefore appropriate for the purpose s of this study. I determined turtles sex visually when possible based on morphology. Di agnostic characteristics used included tail length and thickness, relative position of the an al opening, and foreclaw characteristics. I categorized behavior as either ba sking or submerged. If turtles were basking prior to capture, I recorded the basking substrate. I categorized basking substrates as branch, log, live tree, rock, vegetation, freely-floating vegetativ e debris (e.g., leaves, small twig s), or riverbank. I recorded water depth at each capture location. I noted location of turtles within the research section in order to detect movement patterns. I measured species richness of each turtle community sampled in NFWR using the rarefaction method. The raref action method allows comparison of communities from which 36

PAGE 37

unequal sample sizes were collected (Krebs 1989). I used Program RAREFACT Version 1.0 (Krebs 1989) to complete the calc ulations for the rarefa ction method. No spiny softshell turtles, Apalone spinifera (Le Sueur), were captured from th e upstream section of NFWR in 2007. However, I observed at least five A. spinifera regularly within the upstream section in 2007. Therefore, I calculated an additional rarefaction curve for the tu rtle community present in the upstream section of NFWR in 2007 corrected for the failure to capture any of the regularly observed A. spinifera (i.e., I added five A. spinifera to the turtle community composition of the upstream section captured in 2007). I assessed heterogeneity, which accounts for species richness and evenness (Krebs 1989), by using Simpsons Index of Diversity modified for a finite population a nd the reciprocal of Simpsons Index of Diversity. Simpsons I ndex of Diversity, also known as Hurlberts Probability of Interspecific En counter (PIE) (Hurlbert 1971) is the complement of Simpsons original index (Krebs 1989). Simpsons Index of Diversity is the mo st meaningful species diversity index as it equals the slope of the rarefaction curve at its base (Olezewski 2004). The reciprocal of Simpsons Index of Diversity was also calculated as it is easily interpreted as the number of equally common species required to obtain the calculated va lue of heterogeneity (Krebs 1989). I used Program DIVERS Version 1.1 (Krebs 1989) to calculate Simpsons Index and the reciprocal of Simpsons Index. I used Program NICHE Version 1.1 (Krebs 1989) to ca lculate niche breadth using Levins standardized niche breadth measure and the num ber of frequently used basking substrates. To run programs RAREFACT Version 1.0, DI VERS Version 1.1, and NICHE Version 1.1 (Fortran program codes available in Krebs 1989) I used the GNU Fortran compiler, gfortran (Free Software Foundation, Inc., Boston, MA, USA), to create a Microsoft Windows XP 37

PAGE 38

(Microsoft Corporation, Redmond, WA, USA) executable program. I updated the program source codes for the compiler by changing the specifi er (i.e., IN, IO) for standard input in READ statements and standard output in WRITE statemen ts to (asterisk). I further modified the DIVERS Version 1.1 (Krebs 1989) program code by removing lines 6 7 and 125 129. Lines 6 7 and 125 129 implemented a loop that allowed for entry of multiple data sets in one run of the program. The compiler had difficulty with the data types used on those lines and their removal did not affect the calcu lations performed by the program. I also modified the NICHE Version 1.1 (Krebs 1989) program code by changing the format of line continuations from using an "*" (asterisk) to using a single digit in column 6. I summed basking substrate use by species for each sample year. To assess whether species were partitioning basking substrates, I calculated estimates of Piankas Niche Overlap Index using EcoSim: Null models software for ecology version 7 (Acquired Intelligence Inc. & Kesey-Bear, Jericho, VT, USA). I set model para meters as follows: ni che breadth: retained; zero states: reshuffled; resource states: equipr obable; random number seed: 10; iterations: 1000. I used one-way analyses of variance (ANOVA) to identify differences in water depth use among G. geographica partitioned by sex located in the downstream section in 1969 to 1972 and in the upstream section in 2005 to 2007. I used one-way ANOVA to compare differences in water depth use among G. geographica stinkpots ( Sternotherus odoratus Latreille), and T. s. elegans partitioned by sex in the downstream secti on in 2005 to 2007. I used Tukeys post-hoc procedure to determine significantly differe nt samples. Analyses were limited to G. geographica in both sections in all years and T. s. elegans and S. odoratus located in the downstream section in 2005 to 2007 due to sample size. 38

PAGE 39

Species-specific statistical analyses of turtle data were limited to G. geographica due to larger sample sizes. I calculated population si ze estimates with 95% confidence intervals for G. geographica using the Schumacher-Eschmeyer method (Krebs 1989). I compared the population estimates using the Chapman and Overton met hod (as described in Seber 1982) to identify significant differences (two-tailed test with 5% level of significan ce) between research sections. I calculated standardized density estimates using the Schumacher-Eschmeyer estimated population sizes and the areas cal culated from the products of the mean stream widths and lengths of research sections. I conducted independent sample t-tests a nd analyses of variance (ANOVA) to determine whether th e mean plastron length of the G. geographica populations, partitioned by sex, varied among sampling years when assumptions of normality and equal variance were met. I tested assumptions of normality and equal variance using the KolmogorovSmirnov and Levene analyses, respectively. I used nonparametric Mann-Whitney U and Kruskal-Wallis tests when assumptions of normality and equal variance were not met. I used binomial tests to identify if sex ratios of each population differed from 1:1 (male: female) in any sampling year. I used chi-squared ( ) tests of independence to id entify if sex ratios differed between populations or among sampling years with in each population. I used SPSS version 11.5 (SPSS Inc., Chicago, IL, USA) with = 0.05 to perform all statistical analyses. Results The downstream section was significantly wider than the upstream section (mean = 43.4 m and 28.4 m, respectively; t = 7.438, df = 93, p = 0.000). Midstream water depth was significantly deeper in the dow nstream than in the upstream section (mean = 76.41 cm and 60.92 cm, respectively; t = 2.195, df = 96, p = 0.031). A similar significant pattern was observed when comparing water depths taken one meter from the east bank (mean downstream = 38.69 cm, 39

PAGE 40

mean upstream = 25.14 cm; t = 3.369, df = 96, p = 0.001). No significant difference was observed between values measured one meter from the west bank (mean downstream = 35.06 cm, mean upstream = 35.04 cm; t = 0.005, df = 96, p = 0.996). The stream substrate characteristics in both research sections did not differ from the historical description of the downstream secti on offered by Nickerson and Mays (1973) with the exception of silt and sediment deposits (but see Pitt 2005 for differences in substrate distribution in the downstream section). Substantial silt and sediment deposits were not observed in previous studies (Nickerson and Mays 1973), but were apparent in both re search sections in similar proportions ( 2 = 0.2562, df = 1, p > 0.10) in 2005. Floa ting algal masses, submerged fibrous algal growths, and emergent vegetation stands were also apparent in both study sections, though amounts were significantly higher in the downstream section ( 2 = 0.7941, df = 1, p < 0.005). Total coliform levels exceeded the values deemed safe for full body contact by the Missouri Department of Natural Resources (MDNR) (2005; Figure 2-1) in 34 of the 46 individual water samples collected from the resear ch sections and springs. Fourteen of the 46 water samples also surpassed safe levels of E. coli content for full body contact (MDNR 2005; Figure 2-2). Water temperatures were significantly cool er in the downstream section than in the upstream section (mean = 18.7 o C and 23.7 o C, respectively; t = -16.9, df = 74, p = 0.000). Graptemys geographica was the most abundant turtle specie s in both study sections in all years (Figures 2-3A and B). Trachemys scripta elegans and S. odoratus composed a higher percentage of the turtle community in the downs tream section than in the upstream section in 2005 2007 (Figures 2-3A and B). Their presen ce in the downstream section has steadily increased since previous studies (Pitt 2005). River cooters, Pseudemys concinna (Le Conte), and 40

PAGE 41

snapping turtles, Chelydra serpentina (Linnaeus), were found in low numbers in both sections (Figures 2-3A and B). A small population of A. spinifera, was consistently observed basking in the upstream section during all sampling years, though sampling methods were not conducive to their capture and their presen ce is not accurately described by Figure 2-3B. Hand-capturing A. spinifera is difficult due to their fa st swimming speed. Furtherm ore, the hoop traps failed to catch any turtles in either section and the bait was quickly consumed by non-target species, such as crayfish. Only one A. spinifera, a hatchling, was ever observed in the downstream section in recent years (Figure 2-3A). One alligator snapping turtle, Macrochelys temminckii (Harlan), was found in the upstream section in both 2006 and 2007 (Figure 2-3B). This species was never observed in the downstream s ection in prior studies. Rarefaction curves representing the expected number of species observed in specified sample sizes drawn from the turtle communities in NFWR are presented in Figures 2-4A-H. The turtle community in the downs tream section in 1969 had a hi gher than expected number of species for sample sizes greater than 40 individu als than that predicted for the communities in the same section in recent years (Figure 2-4A). As sample size increased, the expected number of species observed in the downstream section c onverged to approximately five species in all years except 2005 (Figure 2-4A). The rarefac tion curve for the downstream community in 1969 leveled off at sample sizes greater than 60 indi viduals (Figure 2-4A and B). The rarefaction curves for the upstream community in recent year s showed little tendency of leveling off with increasing sample size (Figure 2-4B). A higher e xpected number of species was predicted for all sample sizes for the upstream section than for th e downstream section in 2005 (Figure 2-4C). A higher expected number of species was predicted fo r sample sizes greater than 50 individuals for the upstream section than for the downstream sec tion in 2006 (Figure 2-4D). A lower expected 41

PAGE 42

number of species was predicted for sample sizes less than 115 individuals for the upstream section than for the downstream section in 2007 (Figure 2-4E). However, if the turtle community in the upstream section was corrected for the failure to capture A. spinifera, the rarefaction curve was shifted upwards (Figure 2-4F). The corrected rarefaction curve resulted in a higher expected number of species in the upstrea m section than in the downstream section for sample sizes larger than 70 individuals (Figure 2-4F). Figures 2-4G and H illustrate how the rarefaction curve for the turtle community in the upstream section in 2007 shifted from being below to being almost identical to the curve for the community in the same section in 2006. Comparison of Simpsons index revealed the tu rtle community in the downstream section is more diverse than the community in the upstream section (Table 2-1). The turtle communities in the downstream section in recent years were more diverse than the community observed in 1969 (Table 2-1). All species except M. temminckii were observed basking at least once during the study period (Table 2-2). More than half of all G. geographica from any of the observed populations were basking prior to capture (T able 2-2). The downstream section in 2005 to 2007 was the only community in which species other than G. geographica were in high enough abundance to evaluate basking substrate use (Table 2-2) In the downstream community, 97.7 % of all S. odoratus captured in 2005 to 2007 were obser ved basking (Table 2-2). Of the S. odoratus observed basking, 96.9% were basking on vegetati on (Table 2-2). The majority (76.5%) of T. s. elegans captured from the downstream communi ty in 2005 to 2007 were observed basking (Table 2-2). Niche breadth analysis base d on basking substrate use indicated that G. geographica had the widest niche breadth (Table 2-3). Sternotherus odoratus observed in recent 42

PAGE 43

years had the narrowest niche breadth of species for which niche breadth was evaluated (Table 23). Results of Piankas niche overlap analyses are presented in Table 2-4. Piankas niche overlap index assigns a value of 0.0 (no overlap) to 1.0 (complete overlap) to pairs of species. The only significant results obtained from the niche overlap analyses were from the downstream section in 2005 to 2007 (mean niche overlap = 0.789, probability observed mean overlap is due to chance = 0.000; variance of niche overlap = 0.02807, probability the ob served variance in overlap is due to chance = 0.002) (Table 2-4). Comparisons of water depth at capture site were limited by sample size for each species (Table 2-5). No significant difference in water depth use wa s observed among male, female, or indeterminate sex (i.e., indistinguishable sex ba sed on external character istics, hatchlings and post-hatchlings) G. geographica in the downstream section in 1969 to 1972 (F = 1.889, p = 0.163) (Table 2-6). Graptemys geographica for which sex was visually indistinguishable used significantly shallower water than males and fema les in the upstream section in 2005 to 2007 (F = 25.244, p = 0.000; Tukeys p indeterminate sex:females = 0.000, p indeterminate sex:males = 0.000, p females:males = 0.890) (Table 2-6). Comparison of G. geographica S. odoratus and T. s. elegans partitioned by sex, in the downstream section in 2005 to 2007 revealed that water depth use was significantly different (F = 10.020, p = 0.000) (Table 2-6). Tukeys post hoc analysis indicated that G. geographica for which sex was visibly indistingu ishable used significantly shallower water depths than all other species and sexes, except T. s. elegans for which sex was visibly indistinguishable and T. s. elegans males (Table 2-7). Trachemys scripta elegans for which sex was visibly indistinguishabl e used significantly shallower water depths than male G. geographica and female T. s. elegans (Table 2-7). 43

PAGE 44

The annual G. geographica population estimates in the downstream section (Table 2-8) were not significantly different (2005 and 2006: z = 1.94, p > 0.025; 2006 and 2007: z = 0.89, p > 0.18). The annual G. geographica populations estimates for the upstream section (Table 2-8) were significantly different (2005 and 2006: z = 2.32, p < 0.02; 2006 and 2007: z = 2.14, p < 0.02). The G. geographica population estimates between sec tions (Table 2-8) were not significantly different in any sampling year (2005: z = 0.12, p > 0.45; 2006: z = 0.24, p > 0.40; 2007: z = 1.19, p > 0.11). The validity of the population estimates is dependent on the nature of the data in correspondence with the assumptions of the model. Assumptions of the SchumacherEschmeyer method are: 1. the population is closed, 2. all animals have the same probability of getting captured, 3. marks do not affect catchability, 4. marks are not lost during sampling periods, and 5. all marked individuals are reported (Krebs 1989). The assumption that the population is closed is valid if movement into a nd out of the study area are equal and natality and mortality are equal. The assumption of a clos ed model was partially supported by the limited movement observed. Despite the prevalence of hatc hlings in some years, it is likely that high hatchling mortality, typical of fr eshwater turtles, negates the pot ential recruitment effect. The assumption of equal catchability of marked and unmarked individuals ap peared to be valid, except for adult females, which may have been more wary after the initial capture. Visibility of nail polish marks compensated for this increased wariness. The assumption that marks were not lost throughout the sampling pe riod was confirmed by the retention of both painted marks and PIT tags by dually-marked individuals. All marked individuals were recorded, fulfilling the final assumption. No significant differences were observed with in or between the mean plastron lengths, separated by sex, of the downstream and upstream G geographica populations in any year 44

PAGE 45

(downstream males 2005 2007: F = 1.118, p = 0.33; downstream females 2005 2007: 2 = 2.379, df = 2, p = 0.30; upstream males 2005 2007: 2 = 3.030, df = 2, p = 0.22; upstream females 2005 2007: 2 = 2.428, df = 2, p = 0.30; upstream a nd downstream males 2005: t = 0.068, df = 36, p = 0.95; upstream and downstream males 2006: z = -0.128, p = 0.90; upstream and downstream males 2007: t = 1.337, df = 78, p = 0.19; upstream and downstream females 2005: z = -.0877, p = 0.38; upstream and downs tream females 2006: z = -0.818, p = 0.41; upstream and downstream females 2007: z = -1.160, p = 0.246) (Figures 2-5A and B, Table 2-9). Sex ratios of G. geographica for which sex was distinguisha ble did not differ significantly from 1:1 (male:female) in any population in any sampling year (Downstream: 2005: p = 0.188; 2006: p = 0.237; 2007: p = 0.054; Upstream: 2005: p = 0.652, 2006: p = 1.000, 2007: p = 0.798) (Figure 2-6). Sex ratios of G. geographica for which sex was distinguishable were not significantly different betw een sampling sites (2005: = 2.25, df = 1, p > 0.05; 2006: = 1.14, df = 1, p > 0.05; 2007: = 1.16, df = 1, p > 0.05) (Figure 2-6). Sex ratios of G. geographica for which sex was distinguishable in the downstream section were signifi cantly different among sampling years with more males observed in 2007 ( = 8.062, df = 2, p < 0.05) (Figure 2-6). Sex ratios of G. geographica for which sex was distinguishable in the upstream section were not significantly different among sampling years ( = 0.173, df = 2, p > 0.05) (Figure 2-6). Movement data from recapture histories in the downstream section suggest that 84% of the recaptured G. geographica moved less than 184 m and only 11% of recaptures had moved greater than 460 m. Movement data from recapture histories in the upstream section suggest that 76% of the recaptured G. geographica moved less than 184 m and only 11% of recaptures had moved greater than 460 m. The maximum distance moved by a G. geographica in the downstream study section was by a female tu rtle that moved 1,104 m downstream. The 45

PAGE 46

maximum distance moved by a G. geographica in the upstream study section was by another female turtle that moved 3,703 m upstream. A major assumption of this research is that the two turtle communities are distinct (i.e., no migration occurs between communities). Based on historical (Nickerson unpublished data, Pitt 2005) and recent movement estimates within NF WR, the two research sections are distant enough that natural migration is unlikely. Any natural migration attempted would be limited by a man-made boulder line placed between the two research sections. This boulder line was built to create a water retention area ad jacent to a resort. If exchan ge between populations did occur, it would be possible to identify the origin of tagged individuals. One confounding factor is human-facilitated mi gration because canoeists will occasionally capture and carry individuals for the length of thei r float. During this th ree-year study, only one tagged individual from upstream was found in the downstream section. This suggests that though canoeists may occasionally carry turtles downstream, it is either rare that this transportation happens or unlikely that turtles from the upstream section are deposited in the downstream section. In either scenario, exchange between populations remains extremely low and what does occur is likely unidirectional. Similarly, unidirectional transport may occu r with flooding. In 1970, flooding dislodged a large tree from within the downstream research section and deposited it outside of the research section (Nickerson et al. 2007). Despite this disturbance, many turtles were recaptured within the same or proximate areas in which they we re originally captured (M. Nickerson personal communication). A flood in July 2007 that was la rge enough to submerge all but the very tops of the largest boulders (i.e., water rose approximate ly 1 meter in depth) di d not appear to affect the distribution of G. geographica within NFWR as recaptures of turtles within their usual 46

PAGE 47

capture sites after the flood were common. Thes e data suggest that turtles have behavioral adaptations that counteract, diminish, or el iminate displacement associated with flooding. Discussion The results of this study indicate that tu rtle communities vary in composition, species richness, and heterogeneity in time and in diffe rent sections of a river. This variation corresponds with differences in abiotic factors between river sect ions and within each section through time. Graptemys geographica was the numerically dominant sp ecies in both research sections and all sampling periods (Figure 23A and B). This result is c onsistent with patterns observed for other lotic systems, where rive r turtle fauna is dominated by Graptemys (Bury 1979, Moll and Moll 2004). No more than five species were ever observed in the downstream section in any given year (Figure 2-3A). However, the identity of species changed as ce rtain rare species were observed in some years but not ot hers (Figure 2-3A). Six species were observed in the upstream section in all years (Figure 2-3B), though this is not accurately reflected in Figure 2-3B for 2007 due to the failure to captu re the regularly observed A. spinifera The identity of species in the upstream section also varied as certain rare speci es were observed in some years but not others (Figure 2-3B). The failure to observe all rare sp ecies in all years may be a result of species being transient and not remaining in an area. Altern atively, rare species may not be observed in a given year due to the low probability of observing a numerically rare species. Species richness, as determined by the rarefaction method, was higher in 1969 than in recent years in the downstream section (Figure 24A). The maximum difference in expected number of species among all years for any give n sample size for the downstream section was between 0.07 0.87 (Figure 2-4A). Species ri chness within the upstr eam station was highly variable among years (Figure 2-4G). The maximum difference in expected number of species 47

PAGE 48

among all years for any given sample size fo r the upstream section was between 0.97 2.24 (Figure 2-4G). If the 2007 rarefaction curve for the upstream section was corrected for the failure to capture any of the regularly observed A. spinifera, it would almost entirely overlap the 2006 curve (Figure 2-4H). The maximum differ ence in expected number of species among all years for any given sample size for the upstr eam section would then be between 0.07 1.39 (Figure 2-4H). The community in the upstrea m section typically had higher species richness than the downstream section in 2005 and 2006 (Figures 2-4C-D). Data collected in 2007 indicated that the downstream section had higher species richness for most sample sizes evaluated (Figure 2-4E). However, when th e community composition in the upstream section was corrected for the failure to capture the regularly observed A. spinifera, the rarefaction curve for the community in the upstream section shifte d upwards (Figure 2-4F). Species richness for the community in the upstream section then exce eded that of the downstream section for larger sample sizes (Figure 2-4F). These results indi cated that the upstream section contained more species than the downstream section and this patt ern was reflected especially at larger sample sizes. Comparison of the species composition of each section indicated that the primary differences in species richness was due to th e upstream section having a small population of A. spinifera a species observed only once in the downstream section in recent years (Figures 2-3A and B). In addition, one M. temminckii was captured three times in the upstream section, once in 2006 and twice in 2007 (Figure 2-3B). Macrochelys temminckii was never observed in the downstream section in previous or recent surveys (Figure 2-3A). Heterogeneity, which takes into account both sp ecies richness and evenness, was higher for the downstream section in 2005 to 2007 than for either the upstream section in 2005 to 2007 or the downstream section in 1969 (Table 2-1). These results are indicative of the increased species 48

PAGE 49

evenness associated with an increase in S. odoratus and T. s. elegans observed in the downstream section in 2005 to 2007 (Figure 2-3A). Heteroge neity varied little with in each site in 2005 to 2007 (Table 2-1). However, heterogeneity of th e turtle community in the upstream section was higher in 2005 than in subsequent years (Table 2-1). This re sult corresponds with the higher species richness (Figure 2-4G) but also with a more even species composition in 2005 in the upstream community than in the following years (Figure 2-3B). The downstream community in 2005 appeared to be more even as well, but fewer overall species were found (i.e., species richness was lower than in othe r years) (Figures 2-3A and 2-4A) so the heterogeneity measure remained fairly constant (Table 2-1). Overall, the turtle community in the upstream section had higher species richness, but lower heterogeneity than that of the downstr eam section. These results indicate that the upstream section contains more rare turtle spec ies than the downstream section. Interestingly, heterogeneity measures from the downstream sect ion in 1969 were more similar to those from the upstream section in recent years (Table 2-1). This is indicative of the lower evenness of the downstream section in 1969 as richness is more similar to that of the same section in recent years (Figures 2-4A and B). All species except M. temminckii were observed basking at least once during the study period (Table 2-2). Basking is essential fo r thermoregulation, digestion, maintenance of integumentary health, vitamin D synthesis, and parasite removal (Hutchison 1979, Vogt 1979, Hart 1983, Saumure and Livingston 1994, Vogt and Benitez 1997). Many species of turtles bask, but this behavior is most often associated with turtles within the family Emydidae (Boyer 1965), such as G. geographica and T. s. elegans Few conclusions can be drawn regarding basking behavior of rare species in NFWR, except that it was obser ved (Table 2-2). Measures of 49

PAGE 50

niche breadth indicated that G. geographica used fewer types of ba sking substrates in the downstream section in 1969 to 1972 and in the upst ream section than in the downstream section in recent years (Table 2-3). In 1969 to 1972, G. geographica were most commonly observed basking on rocks and logs (Table 2-2). In the upstream section in 2005 to 2007, G. geographica frequently used branches, logs, and rocks as basking sites (Table 2-2). In 2005 to 2007, G. geographica in the downstream section frequently us ed four types of basking substrates: branches, vegetation, logs, and rocks. Use of par ticular basking substrates is likely linked with the relative abundances of bask ing substrates. Quantifying bask ing substrates was attempted initially in the downstream section, but was quickly abandoned due to the inherent difficulty of quantifying ephemeral substrates. For exampl e, branches would get dislodged and swept downstream and vegetation increased rapidly th roughout the summer months. Despite this limitation, preliminary measurements indicated there was a significantly higher amount of aquatic vegetation in the downstream section in recent years than in previous years or in the upstream section. This difference in vegetation abundance corresponds with the increased use of vegetation as a basking substrate by G. geographica (Table 2-2). Trachemys scripta elegans were also frequently observed basking on vegetation in the downstream section in 2005 to 2007 (Table 2-2). Niche breadth analysis indicated that T. s. elegans commonly used two types of basking substrates (Table 2-3), but the use of logs (N = 11) was far surpassed by the use of vegetation (N = 35) (Table 2-2). Sternotherus odoratus had the narrowest niche breadth and used only one type of basking substrate frequently (Table 2-3): vegeta tion (Table 2-2). In summary, G. geographica had the widest niche breadth (Tab le 2-3), a trait indicative of a generalist for the characteris tic in question (Krebs 1989). Sternotherus odoratus had the narrowest niche breadth (Table 23), which indicates a basking hab itat specialist (Krebs 1989). I 50

PAGE 51

hypothesize that niche breadth and use of different basking habitats may be more indicative of physiological needs and relative competitive ab ility. Basking is most highly developed in emydid turtles (Boyer 1965), which indicates the relative importance of this behavior to this taxon. Therefore, individuals may exploit less preferred basking habitats to undergo this essential behavior if preferred basking substrates are limited. Pitt (2005) noted that fewer emergent boulders were present in 2004 than in previous years. This observation suggests that fewer basking substrates were available onto whic h turtles could fully emerge from the water. This observation supports the hypothesis that G. geographica were increasing their basking niche breadth in response to fewer preferred ba sking substrates. This hypothesis corresponds with the prediction that niche breadth should increase as resources de crease (Pianka 1976). Measures of niche overlap indicate significant overlap in bask ing habitat use among species (Table 2-4). High niche overlap can be indicative of shared res ource utilization and lack of competition or intense competition that has no t yet led to resource partitioning (Pianka 1976). In this system, I suspect that competition is responsible for the obser ved patterns of basking habitat use, with G. geographica being a better competitor than T. s. elegans and S. odoratus Trachemys scripta elegans were not observed in the resear ch section in 1969 (Figure 2-3A, Table 2-2). Subsequent surveys starting in 1980 indicated a small population of T. s. elegans had become established in the downstr eam section of NFWR (Pitt 2005). Trachemys scripta elegans appearance in the research section corresponded with a decline in the G. geographica population, suggesting that T. s. elegans were able to exploit the newly available habitat (Pitt 2005). The G. geographica population failed to rebound and basking habitat, in the form of vegetation, increased in abundance (Pitt 2005). Because T. s. elegans is a highly adaptabl e generalist species (Vogt and Benitez 1997, Webb 1961), changes in community composition and habitat allowed 51

PAGE 52

the T. s. elegans population to persist in the downstream section of NFWR through 2007. This hypothesis is supported by the dispr oportionate use of vegetation by T. s. elegans in recent years (Table 2-2). A similar pattern of niche e xploitation is suspected for S. odoratus Data from 1969 1972 indicate a small population of S. odoratus that was typically observe d basking in live trees, logs, and branches (Table 2-2). More recently, the S. odoratus population has become a major portion of the turtle community in the downstream section of NFWR (Figure 2-3A). All but four of the 130 S. odoratus observed basking were associated with vegetation (Table 2-2). I hypothesize that S. odoratus were able to expand their popula tion within NFWR due to the large expanses of emergent vegetation observed in the downstream section in recen t years. In addition to availability in basking habitat, emerge nt vegetation was associated with muddy, silty substrates, a typical habitat of S. odoratus (Kingsbury 1993). Sternotherus odoratus are typically associated with more lentic habitats with slow or no current and soft bottom substrates (Ernst et al. 1994). Sternotherus odoratus within NFWR appear to be clos ely associated with water lilies ( Nymphaea spp.) and algal mats situated on emergent aquatic vegetation growing from a soft sediment substrate. With appearance of alga l mats in the downstream section since 1980 (Pitt 2005), and the relatively large area that they cover compared to the upstream section, it is likely that the downstream section cont ains more suitable habitat for S. odoratus than it did in previous studies and when compared to the upstream section. These conclusions are based on the assumption that G. geographica prefer basking sites other than vegetation, including logs, branches, and rocks. Pitt (2005) co ncluded that G. geographica preferred highly-branched, felled trees n ear deeper water. Flaherty and Bider (1984) found that G. geographica in an area with excess baski ng habitats preferred basking 52

PAGE 53

substrates that were long, narro w, far from land and aquatic vegetation, and in deep water. Flaherty and Biders (19 84) results suggest that G. geographica avoid aquatic vegetation when other types of potential baski ng substrates are present. Graptemys geographica may prefer branches, logs, and rocks as these substrates allow the turtles to remove their entire bodies from the water. This extrication from the water allo ws for drying of the integument (Cagle 1950) and parasite removal as leeches de siccate in the sun (Boyer 1965) or are removed by birds (Vogt 1979). Warmer body temperatures within a particular range are beneficial as they are associated with enhanced physiological function incl uding digestion (Cagle 1950, Hutchison 1979). The G. geographica populations within the two research sections were not significantly different in size. This result is surprising as the downstream section encompasses a significantly larger aquatic area and volume than the upstream s ection. I predict that a larger area of a given habitat would support more resources, such as ba sking sites, than a smaller area of the same habitat. More resources should increase the car rying capacity of a hab itat so I would predict more turtles would be found in a larger area. Pitt (2005) determined that the G. geographica population in the downstream section had declined from 1969 levels and remained steady at the declined value due to life history trait limita tions, interspecific interactions, and habitat characteristics. Estimated G. geographica population sizes increased among sampling years in 2005 to 2007 in both research sections (Table 28), but only the population estimates from the upstream section were significantly diffe rent. The observed differences in G. geographica population sizes correspond with in creased recruitment in 2006 and 2007 (Figures 2-5A and B). No significant differences between G. geographica plastron lengths, partitioned by sex, were observed between populations in each research section (Table 2-9). Sex ratios of G. geographica for which sex could be distinguished visu ally did not differ from 1:1 (male:female) 53

PAGE 54

for any year or station (F igure 2-6). These data provide evidence that the G. geographica populations within the upstream and downstream sections in 2005 to 2007 were structured similarly. Typically, a larger area contains more speci es and perhaps larger populations than a smaller area (Groom 2006). Additionally, larger areas often contain a higher diversity of microhabitats than do smaller areas (Pluto a nd Bellis 1988, Dunning et al. 2006). Based on the premise that larger areas can support more species than smaller areas, I would expect that more species would be observed in the downstream sec tion as it has a larger aq uatic area and volume. However, the opposite pattern was observed. When viewed in the light of the habitat differences between the two research sections, the results may not be counterintuitive. Water temperatures were signi ficantly different between th e two research sections. Because turtles are ectothermic, water temperature can strongly influence body temperature (Boyer 1965, Schuett and Gatten 19 80, Brown et al. 1994). Turtles may move to and among areas of favorable temperatures in their aq uatic habitats (Schuett and Gatten 1980, Thornhill 1982, Moll and Moll 2004). As a result, some speci es may select habitats with higher water temperatures and avoid cooler therma l regimes. Some species, such as Graptemys and Trachemys further regulate their body temperat ures by basking (Boyer 1965, Gatten 1974, Schuett and Gatten 1980, Spotila et al. 1984, Cadi and Joly 2003) and can inhabit a broader range of temperatures as long as sufficient basking habitat is av ailable. Therefore, basking allows these species to increase their distribution into cooler water bodies that would otherwise not have favorable thermal regimes. The mean water temperature in the downstream section observed in the summer months (i.e ., June, July, and August) was 18.7 o C. Low temperatures may deter or prohibit the use of the downstream section by some species because temperature 54

PAGE 55

strongly affects physiological function in turtles (C agle 1946, Boyer 1965, Gatten 1974, Parmenter 1980, Thornhill 1982, Brown and Brooks 1991, Frazer et al. 1993, Brown et al. 1994, Cadi and Joly 2003, Cadi and Joly 2004). The upstream research section is less impacted by humans. If turtles prefer less impacted habitats, the differences observed between the tu rtle communities and populations are expected. The upstream section is surrounded by Mark Twain National Forest land so the majority of the riparian zone and beyond is forested and undeveloped (Bryant Watershed Project, Inc. 2008). Along the entire upstream research section, the on ly clearing is associated with the national forest campground and boat ramp located at the ups tream boundary of the research section. In contrast, there are five cleared areas surroundi ng portions of the downstream section (Pitt 2005). The intact forest and riparian zones, such as those characterizing the upstream section, are known to prevent excess sediment runoff from surroundi ng areas during rain ev ents (Gilliam 1994). Additionally, the lack of de velopment surrounding the upstream section prevents common practices associated with development, such as applications of fertilizers and pesticides, which occurs in cleared and developed areas along ot her portions of the rive r. Lastly, private residences along NFWR are often lo cated close to the rivers edge which increases the risk of septic systems seeping their contents into the ri ver. This hypothesis is supported by the observed patterns of E. coli contamination within NFWR (Figure 2-2). Only the first station within the upstream research section had E. coli levels that exceeded concentrations deemed safe for full body contact by MDNR (Figure 2-2) In contrast, all areas sa mpled within the downstream section had E. coli levels that exceeded concentrations deemed safe for full body contact in at least one sampling event duri ng the summer (Figure 2-2). 55

PAGE 56

In addition to more riparian development and higher bacterial loads, the downstream section also contains larger amounts of aquatic vegetation, including exte nsive algal mats which were not observed in surveys preceding 2004 (Pitt 2005). Long, fibrous submerged algae ( Cladophora spp.) were observed in surveys preceding 2004, but were less plentiful and did not bloom for as long a time period (Nickerson pe rsonal communication). These observations suggest that nutrient-loading is occurring in the downstream section as algal blooms are often associated with nutrient-load ing (Groom and Vynne 2006). Previous studies confirm that nutrient-loading has occurred in NFWR (M. Solis personal communication). Further evidence of degraded habitat quality in the downs tream research section is illustrated by the comparison of the upstream sections community composition with the community composition of the downstream s ection observed in 1969 (Figure 2-3A). Comparison of these data indicates that the present day community composition of the upstream section most closely resembles that of the dow nstream section in 1969. These similarities may arise because the upstream and downstream sect ions were relatively un-impacted by people in 2005 to 2007 and 1969, respectively. Community composition changes observed within an impacted river in Illinois illustrated a change from more specialized to more generalized turtle species (Moll 1977, 2006). The hi gher relative proportion of T. s. elegans and S. odoratus (i.e., generalist species) observed in the downstream section may be symptomatic of a more heavily impacted habitat. Overall, both river sections and the associated springs are impacted to some extent. This conclusion is well illustrated by the bacteria l analyses conducted within NFWR and its associated springs (Figures 2-1 and 2-2). All sites sampled exceeded total coliform limits deemed safe for full body contact by the MDNR in at least one sampling event during the 56

PAGE 57

summer (Figure 2-1). Furthermore, all springs downstream sampling sites, and one upstream sampling site had E. coli values that exceeded the limits deemed safe for full body contact (Figure 2-2). Because the Ozark Plateau is composed of karst geology (The Nature Conservancy 2003), these results indicate that not only are stormwat er runoff and septic systems potentially directly contaminating water bodie s, but the groundwater itself is contaminated. These results do not bode well for wildlife or humans inhabiting th ese areas, especially if these results are foreshadowing high concentrations of other cont aminants, such as pesticides and metals. Previous studies have confirmed the pres ence of pesticides in NFWR (MDC 2005). Conclusions The North Fork of White River, Ozark C ounty, Missouri, supports a diverse turtle assemblage predominated by Graptemys geographica Turtle communities in NFWR varied in composition, species richness, and heterogeneity thr ough time and in close sections of a river. The downstream turtle community shifted from an assemblage of one common species and several rare species to a grouping of several common species and a fe w rare species. The turtle community in previous years was most similar to that of the upstream section in recent years. The observed variation among turtle communitie s corresponded with spatial and temporal differences in abiotic factors. Based on these results, I conclude that factors such as spring effluents, riparian development, water quality, and algal blooms are disproportionately affecting different areas and associated turtle communities within this river. These results support the hypothesis that tu rtle communities can vary in composition, species richness, and heterogeneity through time and in relatively close sections of a river with a natural thermal gradient. Variation in turt le communities and populations corresponded with differences in abiotic factors between river sec tions and within a section through time. Changes in community composition can result in increased biotic interactions, including competition. The 57

PAGE 58

observed variation among communities and the su spected overwhelming influence of abiotic factors on the turtle communities lends suppor t to nonequilibrium community theories. Nonequilibrium community theories integrate the id ea that the environment (i.e., abiotic factors) strongly influences organisms and is perpetually changing (Chesson and Ca se 1986). As a result of these changes, the populations and communities that inhabit the environment vary (Davis 1986, Cody 1996). The human population is increasing rapidly (C ohen 1995) and most ri vers have undergone massive alterations due to the water and la nd needs of humans (Benke 1990, Riccardi and Rasmussen 1999). As development alters land and water bodies, it is imperative to understand how these alterations will impact wildlife. With long-term studies that illustrate spatial and temporal variation in communities and populations, it will be possible to discern the cause of the variability (Cody 1996). With such data, conser vation and management decisions will be based on sound science rather than guesswork. Turtle populations are dec lining worldwide (Smith 1979, Moll 1986, Kuchling 1988, Ernst et al. 1994, Buhlmann and Gibbons 1997, Gibbons 1997, Haitao 2000, van Dijk et al. 2000, Moll and Moll 2004), but studies, such as this one, th at increase knowledge of the ecology, population status, and community compositi on of turtle species and asse mblages (Moll 1990, Ernst et al. 1994, Moll and Moll 2004) increase the possibility of garnering political, social, and financial support for conservation of impe riled chelonians (Gibbons 1997). This study provides new insights on spatial and temporal variati on of river turtle communities and populations and provides necess ary information for future conservation actions. 58

PAGE 59

59 Table 2-1. Heterogeneity measures for turtle communities in the North Fork of White River, Ozark County, Missouri. Sampling year Simpsons Index of Diversity (modified for a finite population) Reciprocal of Simpsons Index of Diversity Downstream section 1969 0.286 1.398 2005 0.596 2.445 2006 0.597 2.456 2007 0.537 2.143 Upstream section 2005 0.379 1.599 2006 0.167 1.198 2007 0.094 1.102 2007 + 5 Apalone spinifera 0.161 1.190

PAGE 60

60 Table 2-2. Summary of turtle be havior and basking substrate use prior to capture in the North Fo rk of White River, Ozark Count y, Missouri. All capture and recapture events for which data are available are included. # Submerged # Basking Live tree Log Branch Rock Vegetation Freely-floating veg. debris Bank Unspecified substrate Downstream section 1969 1972 Apalone spinifera 1 0 0 0 0 0 0 0 0 Chelydra serpentina 1 0 1 0 0 0 1 0 0 Graptemys geographica 40 1 28 6 36 3 3 3 17 Macrochelys temminckii None captured Pseudemys concinna None captured Sternotherus odoratus 3 4 5 3 0 0 0 0 2 Trachemys scripta elegans None captured 2005 2007 Apalone spinifera 0 0 0 0 0 1 0 0 0 Chelydra serpentina 7 0 1 0 0 2 0 0 0 Graptemys geographica 49 0 56 92 52 67 0 0 1 Macrochelys temminckii None captured Pseudemys concinna 0 0 1 0 0 1 0 0 0 Sternotherus odoratus 3 0 2 2 0 126 0 0 0 Trachemys scripta elegans 20 0 11 9 5 35 2 1 2 Upstream section 2005 2007 Apalone spinifera 2 0 1 0 0 0 0 0 0 Chelydra serpentina 2 0 0 0 0 0 0 0 0 Graptemys geographica 169 3 51 83 30 13 2 0 1 Macrochelys temminckii 3 0 0 0 0 0 0 0 0 Pseudemys concinna 1 0 0 0 1 0 0 0 0 Sternotherus odoratus 2 0 0 1 0 1 0 0 0 Trachemys scripta elegans 16 0 2 0 2 1 0 0 0

PAGE 61

Table 2-3. Basking substrate nich e breadth for turtle species in the North Fork of White River, Ozark County, Missouri. Levins Standardized Measure # of frequently used basking substrates Downstream section 1969 1972 Graptemys geographica 0.331 2 Sternotherus odoratus 0.313 3 2005 2007 Graptemys geographica 0.466 4 Sternotherus odoratus 0.011 1 Trachemys scripta elegans 0.212 2 Upstream section 2005 2007 Graptemys geographica 0.356 3 61

PAGE 62

Table 2-4. Piankas niche overlap index for tu rtle communities in the North Fork of White River, Ozark County, Missouri. Data fo r the downstream section (DS) in 1969 1972 and the upstream section (US) are repr esentative of the number of species observed basking. Dashes indi cate lack of data for species or overlap values that were represented elsewhere in the table. Apalone spinifera Chelydra serpentina Graptemys geographica Pseudemys concinna Sternotherus odoratus Trachemys scripta elegans DS 1969 72 A. spinifera ------------C. serpentina ----0.4734 --0.500 --G. geographica --------0.4948 --P. concinna ------------S. odoratus ------------T. s. elegans ------------DS 2005 07 A. spinifera --0.8944 0.4887 0.7071 0.9997 0.9169 C. serpentina ----0.6198 0.9487 0.9013 0.949 G. geographica ------0.6344 0.5057 0.7738 P. concinna --------0.7181 0.8521 S. odoratus ----------0.9250 T. s. elegans ------------US 2005 07 A. spinifera ----0.4961 0 0 0.6667 C. serpentina ------------G. geographica ------0.2918 0.6603 0.5674 P. concinna --------0 0.6667 S. odoratus ----------0.2357 T. s. elegans ------------62

PAGE 63

Table 2-5. Mean water depth in which the turtles or their baski ng substrates were located prior to capture in the North Fork of White River, Ozark C ounty, Missouri. All capture and recapture events for which da ta were available were included. N Mean water depth (m) Range (m) Standard deviation Downstream section 1969 1972 Apalone spinifera 1 0.61 ----Chelydra serpentina No data Graptemys geographica 49 0.67 0.30 1.30 0.2256 Macrochelys temminckii None captured Pseudemys concinna None captured Sternotherus odoratus 3 1.07 0.91 1.22 0.1550 Trachemys scripta elegans None captured 2005 2007 Apalone spinifera 1 0.30 ----Chelydra serpentina 9 1.03 0.45 2.00 0.5483 Graptemys geographica 257 0.74 0.10 2.00 0.3714 Macrochelys temminckii None captured Pseudemys concinna 1 1.10 ----Sternotherus odoratus 77 0.76 0.27 1.50 0.2310 Trachemys scripta elegans 66 0.71 0.05 1.75 0.3224 Upstream section 2005 2007 Apalone spinifera 1 0.35 ----Chelydra serpentina No data Graptemys geographica 278 0.69 0.06 2.00 0.3288 Macrochelys temminckii 3 1.20 0.80 1.50 0.3606 Pseudemys concinna 2 0.85 0.40 1.30 0.6364 Sternotherus odoratus 4 0.56 0.40 0.75 0.1493 Trachemys scripta elegans 11 0.80 0.05 1.50 0.4860 63

PAGE 64

Table 2-6. Mean water depth in which the turtles or basking subs trates were located prior to capture in the North Fork of White River, Ozark County, Missouri. All capture and recapture events for which data were available were included. Sex N Mean water depth (m) Range (m) Standard deviation Downstream section 1969 1972 Graptemys geographica Indeterminate 14 0.74 0.30 1.30 0.3051 Female 16 0.58 0.30 0.91 0.1729 Male 19 0.66 0.30 1.07 0.1855 2005 2007 Graptemys geographica Indeterminate 81 0.59 0.10 2.00 0.3238 Female 83 0.72 0.15 2.00 0.3654 Male 95 0.88 0.20 2.00 0.3625 Sternotherus odoratus Indeterminate Not included in analysis due to very small N (N = 5) Female 45 0.78 0.40 1.50 0.2247 Male 26 0.75 0.40 1.20 0.2278 Trachemys scripta elegans Indeterminate 32 0.60 0.05 1.00 0.2685 Female 22 0.87 0.10 1.75 0.3670 Male 12 0.71 0.40 1.20 0.2577 Upstream section 2005 2007 Graptemys geographica Indeterminate 116 0.54 0.60 1.25 0.2555 Female 83 0.79 0.10 2.00 0.3648 Male 87 0.81 0.15 1.50 0.3337 64

PAGE 65

Table 2-7. Tukeys post-hoc pr ocedure results for difference in mean water depth use by Graptemys geographica ( Gg ), Sternotherus odoratus ( So ), and Trachemys scripta elegans ( Tse ) partitioned by sex [I = indeterminate sex (i.e., hatchlings, posthatchlings), F = females, M = males] in the downstream secti on of the North Fork of White River, Ozark County, Missouri in 2005 to 2007. Significant differences are indicated by an asterisk (*). All capture and recapture events for which data were available were included. Gg I Gg F Gg M So F So M Tse I Tse F Tse M Gg I --0.000* 0.000* 0.000* 0.019* 0.949 0.000* 0.506 Gg F ----1.000 1.000 0.999 0.055 0.926 0.991 Gg M ------0.999 0.987 0.017* 0.984 0.962 So F --------1.000 0.157 0.925 0.997 So M ----------0.547 0.839 1.000 Tse I ------------0.022* 0.958 Tse F --------------0.790 Tse M ----------------65

PAGE 66

Table 2-8. Schumacher-Eschmeyer populati on size and corresponding density estimates of Graptemys geographica in 2005, 2006, and 2007 in the North Fork of White River, Ozark County, Missouri. Density estimates were based on area calculated from mean stream width and total sample length for each section. Sampling year Estimated population size 95% confidence interval Estimated density Downstream section 2005 114 80-198 1 turtle/ 1751 m 2 2006 270 172-623 1 turtle/ 739 m 2 2007 308 204-621 1 turtle/ 648 m 2 Upstream section 2005 115 92-154 1 turtle/ 1136 m 2 2006 304 189-779 1 turtle/ 430 m 2 2007 578 317-3231 1 turtle/ 226 m 2 66

PAGE 67

Table 2-9. Mean plas tron lengths (PL) of Graptemys geographica in 2005, 2006, and 2007 in the North Fork of White River, Ozark C ounty, Missouri. N represents the sample size. Sampling year Sex N Mean PL (cm) Range (cm) Standard deviation Downstream Male 14 7.8 6.7 9.4 0.8565 2005 Female 23 13.0 6.0 19.6 5.7112 Male 24 7.3 5.5 8.8 1.0336 2006 Female 34 12.7 5.0 19.6 5.9300 Male 48 7.5 5.3 9.6 0.9492 2007 Female 30 14.3 5.1 19.5 5.5940 Upstream Male 24 7.8 5.1 10.2 1.3189 2005 Female 20 11.2 5.6 19.5 4.7520 Male 39 7.5 6.0 10.2 0.8901 2006 Female 38 13.3 5.1 20.2 5.4520 Male 32 7.2 5.4 9.2 1.0488 2007 Female 29 12.6 4.3 20.5 5.0531 67

PAGE 68

0 500 1000 1500 2000 2500Stn. 1 Stn. 17 Stn. 25 Stn. 50 Blue Spring Rainbow Spring 1 Rainbow Spring 2 Spring Creek Althea Spring Stn. 1 Stn. 9-10 Stn. 17 Stn. 28 Stn. 35Sample siteMPN total coliform Sample 1 Sample 2 Sample 3 Sample 4 Sample 5 Upstream Downstream Figure 2-1. Total coliform bacter ia content observed in the North Fork of White River, Ozark County, Missouri. Upstream and downstream indicate stations within the upstream and downstream research sections respectively. Samples 1, 2, 3, 4, and 5 were collected during the third and forth week of June, first and last week of July, and first week of August 2007, respectively. Samp les were not collected at each site during each sampling period due to processing constraints. Sites for which no data were available lack values. MPN repr esents the most probable number of colonyforming units per 100 mL of water. The da shed line indicates the threshold total coliform bacteria concentra tion (200 MPN) deemed safe for full body contact by the Missouri Department of Natural Resources (MDNR) (2005). Values depicted equal to 2500 MPN represent MPN values > 2,424 rather than real values. 68

PAGE 69

0 200 400 600 800Stn. 1 Stn. 17 Stn. 25 Stn. 50 Blue Spring Rainbow Spring 1 Rainbow Spring 2 Spring Creek Althea Spring Stn. 1 Stn. 9-10 Stn. 17 Stn. 28 Stn. 35Sample siteMPN E. coli Sample 1 Sample 2 Sample 3 Sample 4 Sample 5 Upstream Downstream Figure 2-2. Escherichia coli content observed in the North Fo rk of White River, Ozark County, Missouri. Upstream and downstream indicate stations within the upstream and downstream research sections, respectively. Samples 1, 2, 3, 4, and 5 were collected during the third and forth week of June, first and last week of July, and first week of August 2007, respectively. Samples were not collected at each site during each sampling period due to processi ng constraints. Sites for which no data were available lack values. MPN represents the most probable number of colony-forming units per 100 mL of water. The dashed line indicates the threshold E. coli concentration (126 MPN) deemed safe for full body contact by the Missouri Department of Natural Resources (MDNR) (2005). Values depicted equal to 1 MPN represent MPN values < 3 rather than real values. 69

PAGE 70

70 Figure 2-3. Turtle community structure in the (A) downstream a nd (B) upstream sections of the North Fork of White River, Ozark County, Mi ssouri. No data were available for the upstream section in 1969. B 0% 20% 40% 60% 80% 100% 1969 2005 2006 2007Percent composition A 0% 20% 40% 60% 80% 100% 1969 2005Percent composition 2006 2007 Trachemys scripta elegans Macrochelys temminckii Apalone spinifera Pseudemys concinna Apalone spinifera Sternotherus odoratus Chelydra serpentina Pseudemys concinna Chelydra serpentina Trachemys scripta elegans Graptemys geographica Sternotherus odoratus Graptemys geographica

PAGE 71

A0 1 2 3 4 5 6 0 20406080100120140160Expected # of species Downstream 1969 Downstream 2005 Downsteam 2006 Downstream 2007 B0 1 2 3 4 5 6 7 0 20406080100120140160 Sl i Downstream 1969 Upstream 2005 Upstream 2006 Upstream 2007 C0 1 2 3 4 5 6 7 0 20 40 60 80 100 120Sample sizeExpected # of species Downstream 2005 Upstream 2005 D0 1 2 3 4 5 6 7 0 20406080100120140160Sample size Downsteam 2006 Upstream 2006 71 Figure 2-4. Rarefaction curves for the North Fork of White River, Ozark County, Missour i, turtle communities in the (A) downstream section ( DS ) in years 1969 and 2005, 2006, and 2007, (B) DS in 1969 and the upstream section( US ) in 2005, 2006, and 2007, (C) DS and US in 2005, (D) DS and US in 2006, (E) DS and US in 2007, (F) DS and US in 2007 with an estimate of 5 Apalone spinifera added to the US s species count, (G) US in 2005, 2006, and 2007, and (H) US in 2005, 2006, and 2007 with an estimate of 5 A. spinifera added to the US s species count.

PAGE 72

72 E0 1 2 3 4 5 6 0 20406080100120140160Expected # of species Downstream 2007 Upstream 2007 F0 1 2 3 4 5 6 7 0 20406080100120140160 Downstream 2007 Corrected Upstream 2007 G0 1 2 3 4 5 6 7 0 20 40 60 80 100 120 140Sample sizeExpected # of species Upstream 2005 Upstream 2006 Upstream 2007 H0 1 2 3 4 5 6 7 0 20 40 60 80 100 120 140Sample size Upstream 2005 Upstream 2006 Corrected Upstream 2007 Figure 2-4. Continued.

PAGE 73

A0 5 10 15 20 25 30 35 40 0-2.02.1-4.04.1-6.06.1-8.08.110.0 10.112.0 12.114.0 14.116.0 16.118.0 18.120.0 20.122.0Plastron length (cm)Number of individuals 2005 2006 2007 B0 5 10 15 20 25 30 35 40 45 0-2.02.1-4.04.1-6.06.1-8.08.110.0 10.112.0 12.114.0 14.116.0 16.118.0 18.120.0 20.122.0Plastron length (cm)Number of individuals 2005 2006 2007 Figure 2-5. Size distribution of Graptemys geographica in the (A) downstream and (B) upstream sections of the North Fork of White River, Ozark County, Missouri (Downstream: 2005: n = 62, 2006: n = 84, 2007: n = 97; Upstream: 2005: n = 63, 2006: n = 111, 2007: n = 118). Note: Individu als with PL > 12 cm are all females. 73

PAGE 74

0% 20% 40% 60% 80% 100% 200520062007200520062007 Downstream Upstream Females Males Figure 2-6. Sex ratios of mature Graptemys geographica in the North Fork of White River, Ozark County, Missouri (Downstream: 2005: n = 37, 2006: n = 58, 2007: n = 78; Upstream: 2005: n = 44, 2006: n = 77, 2007: n = 61). 74

PAGE 75

CHAPTER 3 COMPARISON OF INSTANTANEOUS GROWTH RATES OF TEMPORALLY AND SPATIALLY DISTINCT Graptemys geographica POPULATIONS INHABITING A RIVER WITH A NATURALLY GENERA TED THERMAL GRADIENT Introduction Phenotypic plasticity in traits that can directly influen ce life history parameters is important for organisms that live in variable environments (Caswell 1983, Stearns and Koella 1986). Phenotypic plasticity in lifehistory traits allows organisms to adapt to a wider range of abiotic factors than could otherwise be tole rated or inhabited with an environmentallyindependent (i.e., fixed) phenotype (Via and Lande 1985). For ectotherms, such as turtles, an important abiotic factor that infl uences phenotypic traits that can directly influence life history parameters is temperature (Laudien 1973, Lillywhite et al. 1973, Gibbons et al. 1981, Bronikowski et al. 2001). Temperature is important for chelonian physiolo gy and life history char acteristics (Bull et al. 1982, Spotila et al. 1984, Rhen and Lang 1995, Mullins and Janzen 2006). Incubation temperature can influence the de velopment rate and se x of turtle embryos (Vogt and Bull 1984) and the growth rate of hatchlings (Rhen and Lang 1995, Roosenburg and Kelley 1996). In hatchling, juvenile, and adult tu rtles, temperature is positively correlated with metabolism through thermally-induced changes in digestion, ingestion, and assimilation rates (Ellis 1936, Cagle 1946, Gatten 1974, Ernst 1975, Parmenter 1980, Thornhill 1982, Spotila et al. 1984, Brown and Brooks 1991, Frazer et al. 1993, Brown et al. 1994, Cadi and Joly 2003). Increased temperatures leading to increased physiological rates have been im plicated in increasing juvenile turtle growth rate (Cagle 1946, Gibbons et al. 1 981), leading to increased survivorship, earlier sexual maturation, and larger body size at va rious stages, including maturation (Bury 1979, Thornhill 1982, Cox et al. 1991, Tucker 2000). La rger body size and earlier sexual maturation of 75

PAGE 76

turtles can lead to reproductiv e enhancements such as earlier age at reproduction, increased clutch size, and increased size or qu ality of eggs (Congdon and Gibbons 1983, 1990, 1996; Lovich et al. 1998). Larger body sizes of sexuall y mature female turtles have been linked to lower mortality rates associated with ne sting excursions (Tucker et al. 1999). Because temperatures effects on growth rate can greatly influence life history characteristics (Cox et al. 1991), it is important to understand how spatial and temporal variation in temperature affects turtle growth rate. Addi tionally, plasticity in growth rate may have important implications for turtles as global climate change is resulting in overall warming trends (Schlesinger and Jiang 1991, Manabe and Stouf fer 1993, Intergovernmental Panel on Climate Change 2007). Field studies investigating spatial differences in growth rate have been largely limited to studies of populations that are geograph ically distant or inhabiting artificially-warmed ponds (Parmenter 1980, Gibbons et al. 1981, Thornhi ll 1982, Spotila et al. 1984, Frazer et al. 1993). Studies of variations in growth rates among turtle populati ons within a river affected by a naturally-induced temperature grad ient are not present in peer-re viewed literature. Temporal variation in growth rate has been identified for northern fre shwater turtle populations and attributed to global climate change (Frazer et al. 1991, 1993). No i nvestigations of turtle growth rates have been reported for populat ions in the central United States. The North Fork of White River (NFWR), Ozar k County, Missouri receiv es a large volume of water from several major springs, which crea tes a temperature gradient along the streams length (Nickerson and Mays 1973). NFWR supports a diverse turtle community, predominated by the Northern Map Turtle, Graptemys geographica (Pitt 2005). The turtle community within a 4.6 km section of NFWR has been studied pe riodically since 1969 (Nickerson and Mays 1973, Pitt 2005). Due to the availability of data datin g back to 1969 and the natural thermal gradient 76

PAGE 77

created by the spring effluents, NFWR provided a unique opportunity to examine if temporal and spatial differences in growth rates of river turtle s occur. My objective was to determine whether instantaneous growth rates of G. geographica vary temporally and spat ially in a river with a natural thermal gradient. Materials and Methods I selected two research sectio ns within the NFWR based on th eir physical similarities, ease of access, locations relative to the major springflows, and suspected difference in mean water temperatures. The first area is the 4.6 km sectio n studied by Nickerson and Mays (1973) and Pitt (2005). This section is located downstream of the major springs and is relatively cool for much of the year, but especially duri ng the summer months. This secti on was originally surveyed in 1969 with follow-up surveys occurring in 1970, 1971, 1972, 1980 (Nickerson unpublished data), and 2004 (Pitt 2005). The second area is a 4.6 km section located upstream of the major springs and is warmer than the downstream section fo r much of the year, but especially during the summer months. The two research sections are separated by approximately 16 river kilometers. I divided each of the two 4.6 km study sections into fifty 92-m-long stations, following the protocol of Nickerson and Mays (1973). In 2005, 2006, and 2007, a research assistant and I surveyed the research sections on alternating days throughout the summer (June 15 August 20) between 0900 and 1800 h, weather permitting. The downstream section was surveyed for a total of 415 person hours. The upstream section was su rveyed for a total of 351 person hours. I recorded air and water temperatures at the begi nning of each sampling day. I compared values obtained for each site using independent sample t-tests to determine whether significant differences in air or water temperatures existed between upstream and downstream research sections. 77

PAGE 78

Surveys were conducted by snorkeling paired w ith mark-recapture techniques. All species of turtles observed in NFWR were indiscriminate ly captured as this study was part of a broader investigation. However, only techniques and data regarding G. geographica will be discussed as the small sample sizes of other species preclude d their inclusion in this study. Snorkeling is useful for capturing G. geographica which are difficult to trap. Methods that rely on traps (see Lagler 1943, Chaney and Smith 1950, Braid 1974, Vogt 1980b) are less effective for G. geographica which may not respond to bait (Lagler 1943) and are wary of traps (Pluto and Bellis 1986). Additionally, suggested traps (see Lagler 1943, Chaney and Smith 1950, Vogt 1980b) are time consuming to assemble, cumbersome and difficult to use in a heavily-trafficked stream like the North Fork (Pluto and Bellis 1986). Hand-capturing methods allowed for behavioral observations of the turtles. All captured turtles were weighed, measure d, and marked using nail polish following the protocol of Pitt (2005). La rger turtles (plastron length > 8.6 cm) were also marked with a passive integrated transponder (P IT) tag (Destron-Fearing Corpor ation, So. St. Paul, MN, USA) injected into the anterior inguina l region parallel to the bridge of the shell following the protocol of Buhlmann and Tuberville (1998). I disinf ected injection sites and needles using 70% isopropyl alcohol and antibiotic ointment. I applied New-Skin Liquid Bandage (Medtech Laboratories, Inc., Jackson, WY, US A) to cover the injection hole. Previous studies from the downstream section of NFWR indicated that both nail polish ma rk and PIT tag retention were high (Pitt 2005) and appropriate for the purposes of this study. I measured the body temperature of larger G. geographica (PL > 5.4 cm) upon capture by inserting a T-6300 Quick-Reading Thermometer (Miller and Weber, Inc., Queens, NY, USA) into the cloaca. The thermometer was inserted a depth of approximately one fifth of th e length of the plastron length into the cloaca. 78

PAGE 79

Body temperature was recorded for male and female turtles with plastron le ngths greater than 5.6 cm and 6.2 cm, respectively, due to size constr aints. I compared body temperatures of G. geographica partitioned by sex, using independent sample t-tests when assumptions of normality and equal variance were met. I conducted nonparametric Mann-Whitney U tests when samples exhibited non-normal distributions and/ or unequal variances. I evaluated assumptions of normality and equal variance using the Kolmogorov-Smirnov and Levene analyses, respectively. I visually de termined turtles sex when possible based on morphological characteristics, including relative tail le ngth and thickness. I used chi-squared ( ) tests of independence to identif y if sex ratios of G. geographica differed between sampling sites or periods. I conducted binomial test s to identify if sex ratios differed from 1:1 (male:female). I calculated indivi dual instantaneous growth rates ( GR) using the equation (Cox et al. 1991, Brown et al. 1994): GR = (log e X 2 log e X 1 ) / (t 2 t 1 ), (3-1) where X 1 and X 2 represent the measurements of th e initial and fina l plastron lengths, respectively, and t 2 t 1 represents the time interval in ye ars that passed between the measuring events. This equation was modified from Br ody (1945). Individuals for which sex could be determined must have had a minimum of 30 days between initial capture and recapture events to be included in growth rate an alyses. Individuals for which se x could not be determined (i.e., indeterminate sex = hatchlings and post-hatchling juveniles) must have had a minimum of 7 days between initial capture and recaptu re events to be included in gr owth rate analyses. Only the final and initial measurements for each turtle we re used to ensure that each individual was included only once in the analysis. 79

PAGE 80

I compared instantaneous growth rates observed for G. geographica in the downstream section during previous (1969 to 1972) and recent (2004 to 2007) sa mple years to assess whether changes in growth rate occurred over time. I co mpared instantaneous growth rates observed for G. geographica in recent years (2005 to 2007) in the dow nstream and upstream sections to assess whether differences in growth rate are appare nt in relatively close locations with differing thermal regimes. I evaluated differences in inst antaneous growth rates, partitioned by sex, using the nonparametric Mann-Whitney U test as data sets exhibited non-normal distributions and/or unequal variances. Different time intervals introduce varying de grees of error into the calculation of instantaneous growth rate (Parchevsky 2000). Because the time period between capture events could range from seven days to four years for in dividuals of indeterminate sex and 30 days to four years for larger juveniles a nd adults, I investigated if analys es would yield different results if mean time periods were similar or significan tly different. When a significant difference in mean time between capture events was detected using either independe nt sample t-tests or nonparametric Mann-Whitney U tests, as appropriate I reevaluated sub-samp les of the original data sets using a nonparametric Mann-Whitney U test. These sub-samples included only individuals that were captured and recaptured within one samp ling season (i.e., less than 3 months). I documented the behavior of G. geographica prior to disturbance and capture. I categorized behavior as either basking or submerged. Basking was used to describe the behavior of any turtle that ha d emerged onto any substrate that allowed the majority of the turtles body to be outside of the water. Typical basking substrates included rocks, fallen trees or branches, emergent vegetation or algal mats, and the riverbank. Submerged was used to 80

PAGE 81

describe any behavior of any tur tle that maintained the majority of its body in the water. Typical submerged behaviors included swimming, feed ing, and bottom-walking. I compared the behavior of G. geographica partitioned by sex, of the 1969 to 1972 and 2004 to 2007 populations within the downstream section using a chi-squared test of independence. I also compared behavior of G. geographica partitioned by sex, of the upstream and downstream populations using a chi-square d test of independence. I gathered climatological data for sout h central Missouri for 1969 to 1972 and 2004 to 2007 from the National Oceanic and Atmospheri c Administration (NOA A) database. The annual number of degree days and the deviation of monthly temperatures from the normal mean temperatures observed in 1971 to 2000 were compared to assess the relative warmth and growing seasons of given years. Degree days, as incl uded in NOAA datasets were calculated by adding one degree day unit for every degree that the m ean daily temperature for that day exceeded 65 o F (approximately 18.3 o C) for each day of the year. Koper and Brooks (2000) presented a modification of the degree day heat index which may be more relevant to temperate species of freshwater turtles inhabiting centr al and northern United States a nd southern Canada. Koper and Brooks (2000) heat index was cal culated by adding one degree day unit for every degree that the mean daily temperature for that day exceeded 15 o C for each day of the year. Fifteen degrees Celsius was selected as the critical temper ature for Koper and Brooks (2000) calculation because it was the minimum air temper ature at which painted turtles, Chrysemys picta (Schneider), were observed feeding. Unfortunatel y, critical temperature data associated with feeding are not available for G. geographica However, the use of 15 o C as an air temperature threshold for feeding may be ju stified as a rough estimate for G. geographica Chrysemys picta and G. geographica belong to the same fam ily of turtles, Emydidae (Ernst et al. 1994). The 81

PAGE 82

geographic distribution of C. picta almost entirely overlaps that of G. geographica (Ernst et al. 1994). In Missouri, C. picta and G. geographica are active during the same time period (late March October) (Johnson 2000). In Missouri, the time period of late March through October corresponds with daily temperature means and maximums that equal or exceed 15 o C (National Oceanic and Atmospheric Administration (NOA A) 2004b). Active periods, which include feeding, for turtles are related to temperatur e (Hutchison 1979, Mahmoud and Klicka 1979) and active periods of C. picta and G. geographica in Missouri correspond with months when mean and maximum daily temperatures equal or exceed 15 o C. Therefore, I used 15 o C as an estimate for assessing active and growth periods for G. geographica I will refer to Koper and Brooks (2000) heat index as growth days from this point forward in this manuscript. I used simple linear regression to determine if instantaneous gr owth rate was correlated with the number of growth days within a year for data sets that yielded significant differenc es in instantaneous growth rates. I performed statistical anal yses using SPSS version 11.5 (SPSS Inc., Chicago, IL, USA) with = 0.05. Results Air temperatures were not significantly differe nt between research se ctions (t = 0.452, df = 72, p = 0.653; Table 3-1). Water temperatures were significantly cooler in the downstream section than in the upstream section (t = 16.9, df = 74, p = 0.000; Table 3-1). Correspondingly, turtle body temperatures observed in the downstream section were significan tly cooler than those in the upstream section (females: t = -3.1, df = 51, p = 0.003; males: z = -4.2, p = 0.000; Table 31). Sex ratios of G. geographica for which sex could be visually determined were not significantly different between sampling sites ( 2 = 0.2195, df = 1, p > 0.05) or sampling periods 82

PAGE 83

( 2 = 0.2620, df = 1, p > 0.05). Sex ratios of G. geographica for which sex could be visually determined were not significantly different from a 1:1 (male:female) ratio in either sampling site (Downstream 2005-2007 : p = 1.00; Upstream 2005-2007 : p = 0.604) or sampling period (Downstream 1969-1972 : p = 1.00; Downstream 2004-2007 : p = 0.501). Comparison of previous (1969 to 1972) and recent (2004 to 2007) instantaneous growth rates observed in the downstream section indicated that growth rates of males and females were not significantly different between sampling pe riods (males: z = -0.578, p = 0.563; females: z = 0.738, p = 0.461; Table 3-2). No significant differences in mean days between capture events were detected for male or female G. geographica between sampling years (males: z = -1.602, p = 0.109; females: z = -0.665, p = 0.506; Table 3-2). Individuals too sm all for sex to be determined based on external diagnostic char acteristics (i.e., indeterminate sex) had a significantly greater mean instantaneous growth rate in 2004 to 2007 than in 1969 to 1972 (z = -3.100, p = 0.002; Table 3-2). A signifi cant difference in days between capture events was observed for G. geographica of indeterminate sex (z = -2.643, p = 0.008; Table 3-2). This variation was due to differences in permanence of marking techni ques as carapace notching was not an approved technique by the University of Floridas Instit utional Animal Care and Use Committee for recent study years. A subset of the 1969 to 1972 data that included only the individuals for which initial and final capture events occurred in the same sampling season was compared to data from recent years. No significant differences in da ys between capture events were observed (t = 1.955, df = 15, p = 0.069; Table 3-2). Individuals for which sex could not be determined still had a significantly greater mean instantaneous gr owth rate in recent year s than in 1969 to1972 (z = -2.148, p = 0.032; Table 3-2) 83

PAGE 84

No significant differences in instantaneous growth rates of G. geographica populations in the downstream and upstream sections were obser ved (males: z = -0.919, p = 0.363; females: z = -1.209, p = 0.227; indeterminate sex: z = -0.619, p = 0.536; Table 3-3) No significant differences in mean days between capture ev ents were detected (males: z = -0.991, p = 0.322; females: z = -0.410, p = 0.682; indeterminate sex: t = 1.924, df = 16, p = 0.072; Table 3-3). Behavioral differences were not observed between sampling periods for the downstream G. geographica population (indeterminate sex: 2 = 0.3259, df = 1, p > 0.05; female: 2 = 3.1044, df = 1, p > 0.05; male: 2 = 1.2447, df = 1, p > 0.05; Table 3-4). Analyses revealed that behavior of all sexe s was dependent on the research section in which the turtles were found (indeterminate sex: 2 = 7.8466, df = 1, p < 0.01; female: 2 = 28.0507, df = 1, p < 0.001; male: 2 = 22.8975, df = 1, p < 0.001; Table 3-5). Graptemys geographica inhabiting the downstream research secti on were observed basking more often than was expected by chance, while those inhabiti ng the upstream section were more often found submerged (Table 3-5). There were more degree days and growth days in 2004 to 2007 than in 1969 to 1972 (Table 3-6). Temperatures in March, whic h mark the beginning of the active season of G. geographica in Missouri (Johnson 2000), were warmer in 2004 to 2007 than in 1969 to 1972 (Tables 3-7 and 3-8). March, April, May, August, and October had more growth days in 2004 to 2007 than in 1969 to 1972 (Table 3-7). Simple linear regression analysis indicated that instantaneous growth ra tes of individuals in which sex was indeterminate were significantly correlated with the annual number of growth days for the year in which the turtles were first captured (Pearsons correlation = 0.542, p = 0.005). Variation in growth days explained 29.4% of variation in instantaneous growth rates (R 2 84

PAGE 85

= 0.294) (Figure 3-1A). Regressi on analysis yielded similar results for samples in which only individuals that were captured within a single sampling season were included (Pearsons correlation = 0.576, p = 0.008; R 2 = 0.332; Figure 3-1B). A major assumption of this research is that the turtle populations co mpared are distinct (i.e., individuals observed in 1969 to 1972 differ from those observed in 2004 to 2007 and no migration between downstream and upstream populations occurred in 2005 to 2007). Graptemys geographica captured in 1969 to 1972 were marked by carapace notching and inspected for notches in 2004 to 2007. No distinct notches were observed, but notches can become obscured with age (Plummer 1979). Ernst et al. (1994) suggest that G. geographica may live to at least 20 years of age. If twenty years is an accurate esti mate of longevity, it is unlikely that many, if any, turtles captured in 1969 to 1972 would be present in 2004 to 2007, 32 years later. Based on historical (Nickerson unpublished da ta, Pitt 2005) and recent (see Chapter 2) G. geographica movement estimates within NFWR, the two research sectio ns are distant enough that natural migration between G. geographica populations is unlikely. Any natural migration attempted would be limited by a man-made boulder line located between the two sections. The boulder line was placed to create a water retention ar ea adjacent to a resort. If exchange between populations did occur, it would be possible to identify the or igin of tagged individuals. One confounding factor is human-facilitated mi gration because canoeists will occasionally capture and carry individuals for the length of their float (personal observation). During this three-year study, only one tagged individual from upstream was found in the downstream section. This suggests that though canoeists may occasionally carry turtles downstream, it is either rare that this transporta tion happens or unlikely that turtles from the upstream section are 85

PAGE 86

deposited in the downstream section. In either scenario, exchange between populations remains extremely low and what does o ccur is likely unidirectional. Similarly, unidirectional transport may o ccur with significant flooding. In 1970, flooding dislodged a large tree from within the downstream research section and deposited it outside of the research section (N ickerson et al. 2007). Despite this disturbance, many turtles were recaptured within the same or proximate areas in which they were originally captured (M. Nickerson personal communication). A flood in Ju ly 2007 that was large enough to submerge all but the very tops of the largest boulders (i.e., water rose approxima tely 1 meter in depth) did not appear to affect the distribution of G. geographica within NFWR as recaptures of turtles within their usual capture sites followi ng the flood were common. These data suggest that turtles have behavioral adaptations that counteract, diminish, or eliminate any potential displacement associated with flooding. Discussion Temporal variation in climatic temper ature regimes has a stronger influence on G. geographica growth rate than does sp atial variation in the aquatic thermal environment. Differences in growth rate are more detectable among turtles for which sex cannot be determined visually (i.e., indeterminate se x, hatchlings, post-hatchlings). Graptemys geographica for which sex was visibly indis tinguishable (i.e., indeterminate sex) in NFWR in 2004 to 2007 grew approximately 2.8 times faster than similarly-sized turtles in the same section in 1969 to 1972 (Table 3-2). This pattern was significantly correlated with an increased number of growth days in 2004 to 2007 than in 1969 to 1972 (Tables 3-6 and 3-7, Figure 3-1). These results suppor t the hypothesis that climatic va riation may affect the growth rate of turtles. 86

PAGE 87

However, growth rate, partitioned by sex, of G. geographica for which sex was distinguishable (i.e., older juven iles and adults) was not signifi cantly different between sampling periods (Table 3-2). Growth rate declines as turtles approach sexual maturity and adult turtle growth rates show little variation compared to that of juveniles (Cagle 1946, Chen and Lue 2002). Because of these patterns, di fferences in growth rates of adults should be minimal. These results should not be interpreted as differences in number of growth days does not affect turtle populations because the enhanced growth rates of smaller turtles may result in higher survival, earlier sexual maturation, or maturation at a larger body size (Bury 1979, Gibbons et al. 1981, Thornhill 1982, Cox et al. 1991, Tucker 2000). Earlier maturation would allow turtles to increase lifetime reproductive output (Gibbons et al. 1981). Similarly, larger body size at maturation is linked to larger clutch size for females and enhanced survival for both sexes (Gibbons et al. 1981, Thornhill 1982, Congdon and Gibbons 1983, Tucker et al. 1999). It is possible that, due to the enhanced growth rates of the juvenile turtles, individuals in 2004 to 2007 may be maturing at an earlier age or larger si ze than those in 1969 to 1972. Because data for the downstream population in 1969 to 1972 does not incl ude age or size at maturity estimates, comparison of these parameters betw een sampling periods is impossible. No significant differences in instantaneous growth rate were observed between the upstream and downstream G. geographica populations, partitioned by sex, in 2005 to 2007 (Table 3-3). This pattern is une xpected as differences in water te mperature have been implicated as contributors to variation in turtle growth rates (Cagle 1946, Brown et al. 1994, Chen and Lue 2002). Mean water temperature of the upstream section was 5 o C higher than mean water temperature of the downstream section, despite no significant difference in air temperatures (Table 3-1). This result was expected due to distribution of springs along NFWR. 87

PAGE 88

Correspondingly, G. geographica body temperatures, partitio ned by sex, were significantly higher by approximately 2 o C for turtles captured in the upstream section than in the downstream section. Lower body temperatures among turtles located in a cooler water body are expected because turtle body temperatures are often correla ted with water temperatures (Brown et al. 1994). However, turtles were ab le to maintain their body temp eratures on average only 2 o C different when mean water temperatures were 5 o C different. This discrepancy is explained by behavioral differences between th e two populations (Table 3-5). Graptemys geographica in the upstream section are submerged more often than those in the downstream section (Table 3-5). Therefore, G. geographica in the colder portion of the stream (i.e., downstream section) are partially overcoming the thermal regime s of their environments by basking. Slider turtle ( Trachemys scripta ) populations inhabiting water bodies subjected to thermal effluents of power plants grew faster than t hose inhabiting similar wa ter bodies with natural thermal regimes (Parmenter 1980, Thornhill 1982). The mean difference in water temperature between the natural and thermally-impacted water bodies in both studies was 5 o C (Parmenter 1980, Thornhill 1982). Both authors suggested that the difference in water temperature accounted, in part, for differences in turtle growth rate (Parmenter 1980, Thornhill 1982). Despite a 5 o C difference in water temperature between the two sections of NFWR, no difference in growth rate was observed. The discrepancy in results between this study and those conducted by Parmenter (1980) and Thornhill (1982) suggest that difference in temperature alone cannot increase turtle growth rate. Spotilla et al. (1984) determined that turt les in the same thermally-impacted pond studied by Parmenter (1980) did not aerially bask, but turtles in the na tural pond were often observed 88

PAGE 89

aerially basking. This behavioral pattern is sim ilar to what I observed in NFWR. Therefore, behavioral patterns are not responsible for the discrepancy in results. Parmenter (1980) suggested that the difference in f ood availability and quality in the study ponds he compared were partially responsible for the difference in growth rate. Thornhill (1982) did not find a difference in diet quantity or quality among turtle populations he investigated. These results suggest that diet was not infl uential enough to create a disparity between the observations of Parmenter (1980) and Thornhill (1982). Therefore, any potential dietary inconsistencies that may exist between the populations in NFWR are probably not responsible for the observed lack of growth rate differences. Thornhill (1982) suggested that the length of growing season was longer in thermallyimpacted ponds and lakes. The longer growing s easons were responsible for the differences in growth rates among populations (Thornhill 1982). In NFWR, the upstream section is significantly warmer throughout the summer months, but because the water temperature varies naturally with air temperature, the length of th e growing season of the upstream section will not be enhanced relative to the downstream section. The water in the downstream section may also be buffered by spring effluents so th at the temperature is more consta nt or stable than that of the upstream section. Nickerson and Mays (1973) f ound that water temperatures in December and January could be upwards of 10 o C in the downstream section. This proposed buffering effect may increase the growing season of the downstream section relative to the upstream section, yielding turtles with similar growth rates. The hypothesis that the growing s eason is responsible for differen ces in turtle growth rates is further supported by the differences in turtle growth rate observed between G. geographica in 1969 to 1972 and 2004 to 2007. More growth days and degree days were observed for 2004 to 89

PAGE 90

2007 than 1969 to 1972 (Table 3-6). When mont hly growth days are analyzed separately, March, April, May, August, and October all had more growth days in 2004 to 2007 than in 1969 to 1972 (Table 3-7). March and October mark the start and finish of the growing season, respectively (Johnson 2000) and th e increased number of growth days in 2004 to 2007 (Table 37) suggests that the growing season in recent year s was extended. These results are in accord with those found by Frazer et al. (1993) for a population of painted turtles ( Chrysemys picta ) in Michigan. Conclusions This study indicates that an ex tension of the growing season can increase turtle growth rate. Graptemys geographica growth rate in NFWR did not vary spatially (Table 3-3) despite a significant difference in water temperature created by spring effluents (Table 3-1). Graptemys geographica growth rate in NFWR did vary temporal ly (Table 3-2). Temporal variation was significantly correlated with the nu mber of growth days in the year of capture (Figure 3-1). Global climate change models predict earlier on set of spring and summer climate patterns (Gates 1993, Intergovernmental Panel on Climate Change 2007). Earlier onset of spring and summer thermal and hydrological regimes is pred icted to alter phenology (McCarty 2001), such as the onset of organisms active and growing se asons. If the observed ex tension of the growing season observed in this study is associated with global climate change, then global climate change, by extension of the growing season, may be altering the growth rate of freshwater turtles inhabiting temperate climates. Vari ations in growth rate can lead to size-related alterations in survival and fecundity (i.e., life history traits) th at can last throughout a tu rtles life (Stearns and Koella 1986, Gibbons et al. 1981). Therefore, global climate change, by extension of the 90

PAGE 91

growing season, may influence not only individual turtles growth rate and other life history traits, but also popul ation parameters. Understanding phenotypic plasticity in growth rate is essentia l for predicting the effects of thermal and climatic variances, such as thos e predicted by global climate change models, on turtle life history traits and populations. As globa l climate change is resulting in overall warming trends (Schlesinger and Jiang 1991, Manabe and Stouffer 1993, Intergovernmental Panel on Climate Change 2007), phenotypic plasticity in grow th rate and other life history traits will be essential for turtles to adjust to shifting abiotic conditions. 91

PAGE 92

Table 3-1. Temperatures of Graptemys geographica surface water, and air observed in the two research sections of the North Fork of White River, Ozark County, Missouri. N represents the sample size. Downstream section Upstream section N Mean temperature ( o C) Temperature range ( o C) N Mean temperature ( o C) Temperature range ( o C) Female G. geographica body temperature 27 24.4 20.6 29.1 26 26.2 22.8 29.8 Male G. geographica body temperature 45 24.8 19.9 29.8 29 26.8 22.9 29.4 Water temperature 45 18.7 16.4 21.4 31 23.7 21.3 27.1 Air temperature 45 29.3 20 37 29 28.9 18 39 92

PAGE 93

Table 3-2. Instantaneous growth rates of Graptemys geographica in the downstream section of North Fork of White River, Ozark County, Missouri for years 1969 1972 and 2004 2007. N represents the sample size. C.I. denotes the confidence interval. Indeterminate sex* indicates a sub-sample in which mean days between capture events has been limited to one sampling season. N Mean instantaneous growth rate (95% C.I.) Instantaneous growth rate range Mean # of days between capture events # of Days between capture events range 1969 1972 Indeterminate sex 12 0.225 (0.113 0.337) 0.019 0.493 253 14 1071 Indeterminate sex* 7 0.340 (0.216 0.465) 0.150 0.493 28 14 42 Female 16 0.054 (-0.018 0.126) 0 0.540 419 33 764 Male 25 0.009 (0.005 0.014) 0 0.041 598 61 1572 2004 2007 Indeterminate sex 10 0.594 (0.435 0.753) 0.166 0.817 18 7 28 Female 20 0.013 (-0.002 0.028) 0 0.133 535 31 1104 Male 17 0.033 (0.006 0.607) 0 0.194 443 31 1114 93

PAGE 94

Table 3-3. Instantaneous growth rate of Graptemys geographica in the downstream and upstream sections of the North Fork of White River, Ozark County, Missouri for years 2005 2007. N represents the sample size. N Mean instantaneous growth rate (95% C.I.) Instantaneous growth rate range Mean # of days between capture events # of Days between capture events range Downstream section Indeterminate sex 9 0.642 (0.508 0.775) 0.381 0.817 19 7 28 Female 13 0.019 (-0.006 0.430) 0 0.133 403 31 742 Male 11 0.049 (0.008 0.091) 0 0.194 259 31 757 Upstream section Indeterminate sex 9 0.759 (0.549 0.970) 0.391 1.41 13 7 21 Female 13 0.023 (0.010 0.362) 0 0.063 456 309 756 Male 8 0.019 (-0.002 0.040) 0 0.073 442 39 725 94

PAGE 95

Table 3-4. Behavior al observations of Graptemys geographica in the downstream section of the North Fork of White River, Ozark C ounty, Missouri for years 1969 1972 and 2004 2007. N represents the sample size. N # submerged # basking 1969 1972 Indeterminate sex 40 9 31 Female 27 10 17 Male 24 10 14 2004 2007 Indeterminate sex 125 23 102 Female 97 20 77 Male 107 32 75 95

PAGE 96

Table 3-5. Behavior al observations of Graptemys geographica in the downstream and upstream sections of the North Fork of White Ri ver, Ozark County, Missouri for years 2005 2007. N represents the sample size. N # submerged # basking Downstream section Indeterminate sex 84 6 78 Female 70 12 58 Male 88 26 62 Upstream section Indeterminate sex 110 24 86 Female 65 40 25 Male 85 56 29 96

PAGE 97

Table 3-6. Degree and growth days for s outh-central Missouri (N ational Oceanic and Atmospheric Administration 1969, 1970, 1971, 1972, 2004a, 2005, 2006, 2007). Degree days were calculated by adding one degree day unit for every degree that the mean daily temperature for that day exceeded 65 o F (approximately 18.3 o C) for each day of the year. Growth days were ca lculated by adding one degree day unit for every degree that the mean daily temperature for that day exceeded 15 o C for each day of the year. 1969 1972 2004 2007 Year Degree days Growth days Year Degree days Growth days 1969 1315 1253.3 2004 1065 1117.2 1970 1305 1252.2 2005 1538 1363.3 1971 1249 1248.3 2006 1243 1195.6 1972 1267 1217.5 2007 1474 1390.3 Total 5136 4971.3 Total 5320 5066.4 97

PAGE 98

Table 3-7. Growth days of months in which Graptemys geographica are active in south-central Missouri. Growth days were calculated by adding one degree day unit for every degree that the mean daily temp erature for that day exceeded 15 o C for each day of the year. Year March April May June July August September October 1969 0 15.6 125.6 228.1 375.6 273.9 174.7 60.0 1970 0 55.6 152.8 211.1 288.6 299.4 214.2 23.1 1971 6.1 26.4 78.9 288.3 279.2 262.5 221.7 79.2 1972 6.2 45.6 107.8 230.6 274.2 295.6 222.8 36.4 Total 10.3 143.2 465.1 958.1 1217.6 1131.4 833.4 198.7 2004 12.2 31.4 155.8 195.6 270.8 224.7 172.2 51.7 2005 0 27.8 90.0 273.6 320.8 351.9 236.7 51.7 2006 14.2 78.1 116.7 225.8 306.1 321.4 101.7 30.6 2007 41.1 23.9 148.3 229.4 268.6 396.1 199.2 75.8 Total 67.5 161.2 510.8 924.4 1166.3 1294.1 709.8 209.8 98

PAGE 99

Table 3-8. Deviation from norma l monthly mean temperature ( o F) for south-central Missouri. Normal mean temperature is based on the period from 1971 2000 (National Oceanic and Atmospheric Administra tion 1969, 1970, 1971, 1972, 2004a, 2005, 2006, 2007). Year March April May June July August September October 1969 -5.8 1.8 2 0.8 3.7 -1 1.6 0.1 1970 -4.4 4 3.6 -0.2 -1.3 0.7 4 -1 1971 -2.1 -0.2 -1.8 4.4 -1.9 -1.7 4 6.2 1972 2 1.8 1.3 0.8 -2.1 0.3 4.1 -1 2004 5.1 1.3 3.2 -1.1 -2.4 -3.8 1.3 2.8 2005 -1.8 1.1 -1.6 3.5 0.5 3.6 5.1 -0.5 2006 1.2 6.3 0.9 0.7 -0.3 1.8 -3.7 -3.3 2007 8.8 -3.0 3.3 0.9 -2.5 6.1 2.8 2.6 99

PAGE 100

1400 1300 1200 1100Instantaneous growth rate1.0 .8 .6 .4 .2 0.0 Growth days1400 1300 1200 1100Instantaneous growth rate.9 .8 .7 .6 .5 .4 .3 .2 .1 A Y = 0.0021X 2.3438 R 2 = 0.294 B Y = 0.0017X 1.6875 R 2 = 0.332 Figure 3-1. Linear re gression of individual Graptemys geographica instantaneous growth rate and growth days for 1969 1972 and 2004 2007. (A) G. geographica for which sex could not be determined based on visi ble characteristics. (B) subset of A in which data were restricted to individuals that were captured within the same sampling year. Curved lines represent 95% confidence intervals. 100

PAGE 101

CHAPTER 4 GROWTH CURVE ESTIMATIONS FOR GRAPTEMYS GEOGRAPHICA INHABITING THE NORTH FORK OF WHITE RIVER, OZARK COUNTY, MISSOURI Introduction Growth curve estimates obtained by genera tion of growth models based on field observations can provide estimates of age at sexu al maturity and highlight differences in life history characteristics between sexes and among populations and taxa (Frazer et al. 1990a, Lindeman 1997). Growth models provide a statistic al approach for estimating age of individuals for which a measurement of interest, such as pl astron length (PL), is known (Cox et al. 1991). Growth models characterize the mean growth trend for the populat ion in question (Frazer et al. 1990a). Among freshwater turtle populations, the von Bertalanffy growth model has been determined to be the best descri ptor of growth (Lindeman 1999). Von Bertalanffy growth curves have been generated for half of the 12 recognized Graptemys species (Ernst et al. 1994, Jones and Hartfield 1995, Lindeman 1999, Lindeman 2005). Lindeman (1999, 2005) created and comp ared von Bertalanffy growth curves for G. ouachitensis (Cagle), G. pseudogeographica (Gray), G. ernsti (Lovich and McCoy), G. caglei (Haynes and McKown), and G. versa (Stejneger). Jones and Ha rtfield (1995) constructed a von Bertalanffy growth curve for G. oculifera (Baur). Despite the preval ence of growth modeling for Graptemys, no von Bertalanffy growth curves have been estimated for G. geographica one of the more, if not the most, widespread Graptemys species (Ernst et al. 1994). Gordon and MacCulloch (1980) present growth data of G. geographica in Canada by plotting changes in plastral length of turtles of unknown age versus year of capture. Iverson (1988) increased the knowledge of G. geographica growth trends by comparing plastral lengths with age for individuals in which ages were known or estimated. Though these studies provide 101

PAGE 102

insight into G. geographica growth trends, they are limited in their use for comparison with other populations and taxa and do not provide in formation related to age at maturity. My first objective was to determine the mean age at maturity of G. geographica inhabiting the North Fork of White River (NFWR), O zark County, Missouri, in 2005 to 2007 using a statistical method. My second objective was to compare the von Bertalanffy growth model variables obtained for G. geographica to variables generated for other species of Graptemys. The G. geographica population within a 4.6 km section of NFWR has been studied periodically since 1969 (Nickerson and Mays 1973, Pitt 2005). Subsequent studies occurred in 1970, 1971, 1972, 1980 (Nickerson unpublished data), a nd 2004 (Pitt 2005). Previous analyses (see Chapter 3) suggested that turtles in the recen t years (2005 to 2007) were growing faster than turtles in previous years (1969 to 1972). The diffe rence in growth rate was attributed to an extension of growing season (see Chapter 3). Th ese results suggest the question, are differences in growth rates between sample periods significant enough to yiel d different growth curves and ages at maturity for G. geographica populations inhabiting the sa me section of NFWR during different time periods? My third objective was to determine if von Bertalanffy growth curves and ages at maturity of G. geographica populations inhabiting the sa me area in different time periods varied. Materials and Methods I selected two research sec tions within the NFWR based on their locations, physical similarities, and ease of access. The first area is the 4.6 km section studied by Nickerson and Mays (1973) and Pitt (2005). This section is located downstream of th e major springs and is relatively cool for much of the year, but especi ally during the summer months. This section was originally surveyed in 1969 with follow-up surveys occurring in 1970, 1971, 1972, 1980 (Nickerson unpublished data), and 2004 (Pitt 2005). The second area is a 4.6 km section located 102

PAGE 103

upstream of the major springs and is relatively wa rm for much of the year but especially during the summer months. The two research sect ions are separated by approximately 16 river kilometers. I divided each of the two 4.6 km study sections into fifty 92-m-long stations, following the protocol of Nickerson and Mays (1973). In 2005, 2006, and 2007, a research assistant and I surveyed the research sections on alternating days throughout the summer (June 15 to August 20) between 0900 and 1800 h, weather permitting. The downstream section was surveyed for a total of 415 person hours. The upstream section was surveyed for a total of 351 person hours. Surveys were conducted by snorkeling paired w ith mark-recapture techniques. All species of turtles observed in NFWR were indiscriminate ly captured as this study was part of a broader investigation. However, only techniques and data regarding G. geographica will be discussed as the small sample sizes of other species preclude d their inclusion in this study. Snorkeling is useful for capturing G. geographica which are difficult to trap. Methods that rely on traps (see Braid 1974, Chaney and Smith 1950, Lagler 1943, Vogt 1980b) are less effective for G. geographica which may not respond to bait (Lagler 1943) and are wary of traps (Pluto and Bellis 1986). Additionally, suggested traps (see Chaney and Smith 1950, Lagler 1943, Vogt 1980b) are time consuming to assemble, cumbersome and difficult to use in a heavily-trafficked stream like the North Fork (Pluto and Bellis 1986). All captured turtles were weighed, measure d, and marked using nail polish following the protocol of Pitt (2005). La rger turtles (plastron length > 8.6 cm) were also marked with a passive integrated transponder (P IT) tag (Destron-Fearing Corpor ation, So. St. Paul, MN, USA) injected into the anterior inguina l region parallel to the bridge of the shell following the protocol of Buhlmann and Tuberville (1998). I disinf ected injection sites and needles using 70% 103

PAGE 104

isopropyl alcohol and antibiotic ointment. I applied New-Skin Liquid Bandage (Medtech Laboratories, Inc., Jackson, WY, US A) to cover the injection hole. Previous studies from the downstream section of NFWR indicated that both nail polish ma rk and PIT tag retention were high (Pitt 2005) and thus appropriate for the purpose s of this study. I estima ted the age of turtles when possible by counting the annuli on the abdominal lamina. Due to the erosion of annuli with age (Graham 1979), estimates were only possi ble for very small (i.e., young) turtles. I visually determined turtles sex when possible based on morphological characteristics, including relative tail length and thickness. I recorded ta il length, which is the most apparent and easily measured diagnostic characteristic useful for determining sex in G. geographica for each turtle. I plotted tail length against plas tron length to illustrate differences in the relationship of these parameters between sexes. I then calculated th e ratio of tail length to plastron length. I plotted plastron length against this ratio to elucidate at what plastron length and ratio of tail length to plastron length differences among sexes become visible. I calculated indivi dual instantaneous growth rates ( GR) using the equation (Brown et al. 1994, Cox et al. 1991): GR = (log e X 2 log e X 1 ) / (t 2 t 1 ) (4-1) where X 1 and X 2 represent the measurements of th e initial and fina l plastron lengths, respectively, and t 2 t 1 represents the time interval in ye ars that passed between the measuring events. This equation was modified from Br ody (1945). Individuals for which sex could be determined must have had a minimum of 30 days between initial capture and recapture events to be included in growth rate an alyses. Individuals for which se x could not be determined (i.e., indeterminate sex = hatchlings and post-hatchling juveniles) must have had a minimum of seven days between initial capture and r ecapture events to be included in growth rate analyses. Only 104

PAGE 105

the final and initial measurements for each turtle were used to ensure that each individual was included only once in the analysis. I evaluated differences in instantaneous growth rates of turtles in the two research sections in 2005 to 2007, partitioned by sex, using the nonpa rametric Mann-Whitney U test as samples exhibited non-normal distributions and/or unequa l variances. I evaluated normality and equal variance using the Kolmogorov-Smirnov and Leve ne analyses, respectively. Because no significant difference in instantaneous growth rate was observed, I pooled data from the two G. geographica populations obtained from the two research s ections to create a larger sample size for analysis. I plotted instantaneous growth rate s against median plastron length to elucidate the general growth patterns observed for G. geographica in NFWR. Median plastron length indicates the median size of a turtle calculated from the plastron length measurements from the initial and final captures. I compared instantaneous growth rates of male and female G. geographica located in NFWR in 2005 to 2007 using a nonparametric Mann-Wh itney U test to determine if differences in growth rate of larg e juvenile and adu lt turtles could be discerne d. A nonparametric test was conducted as data sets had non-normal distributi ons and/or unequal variances as determined by the Kolmogorov-Smirnov and Levene analyses, respectively. I compared instantaneous growth rates observed for G. geographica in the downstream section during previous (1969 to 1972) sample y ears with instantaneous growth rates observed for G. geographica in the upstream and downstream sec tions, combined, during recent (2005 to 2007) sample years. I conducted th is analysis to assess whether ch anges in growth rate occurred over time when estimates from recent years from both sections were combined. I evaluated 105

PAGE 106

differences in instantaneous growth rates, partitioned by sex, using the nonparametric MannWhitney U test as samples had non-normal distributions and/or unequal variances. I used recapture data to estimate von Bertalan ffy growth curves using the Fabens method, which accommodates for the lack of age data (Frazer et al. 1990a). Age is difficult to discern in all but the youngest of G. geographica because annuli, the typical characteristic used to determine age of turtles, fade with ag e (Graham 1979, Lindeman 1999). The general von Bertalanffy equation is: L t = a(1 be -kt ), (4-2) where L t represents the length at age t, a is the asym ptotic length, b is a variable related to the size at hatching, e is Eulers constant (i.e., the ba se of the natural logar ithm), k represents the intrinsic growth rate variable, a nd t is the age (Frazer et al. 19 90a). The Fabens method allowed for the estimation of growth curves when age was unknown by providing the following rearrangement of the von Bertalanffy equation (Frazer et al. 1990a): L r = a (a L c )e -kd (4-3) where L r and L c represent the lengths at recapture and initial capture, respectively, and d is the time interval between L r and L c Variables a, e, and k are the same as those found in Equation 42. I estimated variables a and k by fitting mark-re capture data to Equation 4-3 using a nonlinear regression analysis conducted with SPSS version 11.5 (SPSS Inc., Chicago, IL, USA). I conducted analyses for each sex with individu als of indeterminate sex (i.e., hatchlings; posthatchlings) included in datasets for both sexes to provide data for the sm allest size classes. By solving Equation 4-2 for age zero, a simp lified equation is generated as follows: L 0 = a(1 b), (4-4) 106

PAGE 107

and can be rearranged to estimate the value of variable b: b = 1 (L 0 /a) (4-5) (Frazer et al. 1990a). I estimated L 0 the size at hatching, as the size of the smallest individuals observed (plastron length = 2.4, N = 2) I selected this value for L 0 because all turtles were collected from their aquatic ha bitat so some indeterminate time had passed since hatching. Because growth occurs rapidly in hatchlings, it was likely that growth ha d occurred in many of the immature turtles observed. My decision to use the minimum values observed is supported by evidence collected by D. Moll (personal communi cation). D. Moll (per sonal communication) found a mean PL of 2.3 cm from 36 G. geographica hatchlings (range = 1.9 cm 3.1 cm) from the Niangua River, Dallas County, Missouri incubated in a laboratory. I inserted values obtained for a, k, and b into Equation 4-2 and solved for the length at age t for each year of age. I estimated da ta sets for both sexes individually as G. geographica are sexually dimorphic in size and have differing grow th curves (Iverson 1988). I plotted data for turtles which had estimated known ages based on annuli counts along with von Bertalanffy growth curves to assess the relative fit of the model estimates. I also estimated a data set for the male G. geographica population located in the downstream section of NFWR in 1969 to 1972. Small sample size precluded the gene ration of a data set for female G. geographica present in the same population. A major assumption of this research is th at the von Bertalanffy growth model is representative of the unde rlying growth function of Graptemys geographica in NFWR. The von Bertalanffy growth model has been evaluated in comparison with other models for a diversity of turtle species with mixed results (Chen and Lue 2002, Cox et al. 1991, Iver son et al. 1991, Jones and Hartfield 1995, Lindeman 1997, Lindeman 1999). Lindeman (1997) reviewed the 107

PAGE 108

effectiveness of the von Bertalanffy model for describing freshwater turtle growth and found that it was the most accurate descriptor. A dditionally, Lindeman (1999, 2005) and Jones and Hartfield (1995) found that the von Bertalanffy growth model was appropriate for members of the genus Graptemys. Frazer et al. (1990a) demonstrated that Fa bens method was accurate for constructing growth curves of unknown-age individuals of turtles when the von Bertalanffy model represents the underlying growth function. Results Relative tail length differences between sexes b ecome apparent as pl astron lengths reach approximately 5 cm and tail length to plastron length ratio equals approximately 0.45 (Figures 41 and 4-2). Following achievemen t of plastron lengths of 5 cm, the tail length of males exceeds that of females with corresponding plastron lengths (Figure 4-1). No significant differences in instantaneous growth rates of G. geographica populations in the downstream and upstream sections were observed (males: mean GR downstream = 0.049, GR upstream = 0.019, z = -0.919, p = 0.363; females: mean GR downstream = 0.019, GR upstream = 0.023, z = -1.209, p = 0.227; indeterminate sex: mean GR downstream = 0.642, GR upstream = 0.759, z = -0.619, p = 0.536). Plotting instantaneous growth rate versus median plastron length revealed that the growth rate of G. geographica decreases with size and is close or equal to zero at larger sizes (Figure 4-3). No significant difference in instantaneous gr owth rate was observed between male and female G. geographica inhabiting NFWR in 2005 2007 (mean male GR = 0.0368, mean female GR = 0.0208, z = -1.164, p = 0.245). 108

PAGE 109

Comparison of instantaneous growth rates obse rved in the downstream section in 1969 to 1972 with values for the combined upstream and downstream populations in 2005 to 2007 indicated that growth rates of males and fema les were not significantly different between sampling years (males: mean GR 1969-1972 = 0.0095, GR 2005-2007 = 0.0626, z = -1.607, p = 0.108; females: mean GR 1969-1972 = 0.0539, GR 2005-2007 = 0.0208, z = -0.222, p = 0.825). Individuals too small for sex to be determined based on external diagnostic characteristics (i.e., indeterminate sex) had a signifi cantly greater mean instantane ous growth rate in 2005 to 2007 than in 1969 to 1972 (mean GR 1969-1972 = 0.2250, GR 2005-2007 = 0.7335, z = -4.061, p = 0.000). Estimates of variables a and k produced by th e nonlinear regression for Equation 4-3 are presented in Table 4-1. Vari able b was estimated as 0.714 and 0.879 for male and female G. geographica in NFWR in 2005 to 2007, respectively. Variable b was estimated as 0.717 for male G. geographica in NFWR in 1969 to 1972. Table 4-2 summarizes von Bert alanffy variable estimates generated in this study with thos e generated for other members of the genus Graptemys by Jones and Hartfield ( 1995) and Lindeman (1999, 2005). Based on the growth curves generated for populations in NFWR in 2005 to 2007 (Figure 44), male and female G. geographica grow at approximately the same rate through their first year. Following the first year, females c ontinue to grow at a high rate while male growth rate slows (Figure 4-4). Females achieve substantially larg er body sizes than males (Table 4-1, Figure 4-4). Using Gordon and MacCullochs (1980) estimates of plastron length at sexual maturity, females in this population mature betw een 11 and 12 years of age while males mature between three and four years of age. Comparison of age-estimated turtles relative size and age-size values obtained from the von Bertalanffy growth models (Fig ure 4-5) revealed that the von Be rtalanffy models were fairly 109

PAGE 110

representative of G. geographica growth in NFWR assuming age es timates of turtles captured in the field were reasonably accurate. Comparison of the von Bertalanffy growth cu rves calculated for ma les in 1969 to 1972 and 2005 to 2007 revealed that male G. geographica in recent years were growing faster than those observed in previous years (Figure 4-6). Base d on Gordon and MacCullo chs (1980) estimate of plastron length at sexual maturity, males in 1969 to 1972 were maturing between seven and eight years of age. Discussion Graptemys geographica grow rapidly as hatchlings and j uveniles but growth rate declines as turtles attain sexual maturity (Fig ures 4-3 and 4-4). Male and female G. geographica grow at different rates following their fi rst year, with females growing faster and attaining larger body sizes than males (Figure 4-4). These results ar e in accord with those found by Iverson (1988) for a G. geographica population in Indiana. It should be not ed that Iverson (1988) stated the growth rates of male and female G. geographica in an Indiana population dive rged at approximately two years of age, but the tr end-lines fitted to the growth data in his study (see Figure 1 in Iverson 1988) suggested that the deviation occurred following one year of age. At one year of age, G. geographica in NFWR had a mean plas tron length of 5 cm (Figure 4-4). This was the same size at which the relative tail length became longer for male G. geographica (Figures 4-1 and 4-2). This change in growth rate of males in correlation with enhanced tail length relative to females, suggest that a portion of the energy available for growth in males is diverted to the attainment of sec ondary sexual characteristics, such as increased tail length. Growth rate decreased sharply for males and fe males at approximately three to four and 11 to 12 years old, respectively (Figure 4-4). These declines in growth rate likely correspond with 110

PAGE 111

ages at sexual maturity. This hypothesis is supported by estimate s of age at maturity generated using plastron lengths at sexual maturity repor ted by Gordon and MacCulloch (1980). Based on Gordon and MacCullochs (1980) estimates of plastr on lengths at sexual matu rity, females in this population mature between 11 and 12 years of ag e while males mature between three and four years of age. These values ar e in accord with those found in other studies (Iverson 1988, Vogt 1980a). Iverson (1988) estimated age at maturity for female and male G. geographica in Indiana at nine to 10 and three to fi ve years, respectively. Vogt ( 1980a) estimated that males became sexually mature at four years of age and females after 10 to 12 years of age in a G. geographica population in Wisconsin. Values for asymptotic plastron length (variabl e a) were just above the mean adult size plastron lengths observed in the NFWR G. geographica population (mean PL male = 8.4 cm; mean PL female = 19.9 cm; Table 4-1). According to Frazer et al. (1990a), the ge nerated von Bertalanffy growth curve is a sufficient representation of th e actual growth pattern if the mean adult size observed lies just below the estimated values of asymptotic plastron length. Estimated age-at-size based on annuli c ounts corresponded with the von Bertalanffy growth curves, but agreement of count data and growth curves was greater for juveniles and males than for females (Figure 4-5). I suspect that age estimates based on annuli counts in this study may be somewhat low for the largest (i.e., olde r) females in which annuli were still visible. Alternatively, the von Bertalanffy growth curve for female G. geographica could be skewed towards a larger size at age. This hypothesis se ems unlikely as a reduction in size at age would represent a much slower juven ile growth rate and sexual matu ration would be significantly delayed. 111

PAGE 112

Despite the variation in growth rate between sexes as depicted by the von Bertalanffy growth curves, no significant differences in inst antaneous growth rates were observed for male and female G. geographica These results suggest that differe nces in growth rate between sexes are largely attributed to variati on in growth rate among juvenile G. geographica The comparisons of instantaneous growth rate data between sexes are incor porating all individuals for which sex was distinguishable. As adult turt le growth rate approaches zero, any differences in growth rate between sexes will be dampened by these near-zero and zero values. If a large enough sample size were available to partition sexe s into smaller size classes based on plastron length, I predict that a significan t difference in instantaneous gr owth rates would be observed between smaller and younger male and female G. geographica Von Bertalanffy parameters estimated in this study for G. geographica (Table 4-1) are comparable to those found for other Graptemys species of similar size (Table 4-2). These results are expected due to the close phylogen etic relationship of these species. The von Bertalanffy growth curve estimated for male G. geographica inhabiting NFWR in 1969 to 1972 approached asymptotic size at a slower rate than for males in 2005 to 2007 (Figure 4-6). This difference in growth rate is apparent when the intrin sic growth rates (k) of the two populations are examined (Table 4-1). A lower intrinsic growth rate (k) corresponds with a slower approach to asymptotic size (a) (Linde man 1999). These data suggest that males in 1969 to 1972 matured at seven to eight year s of age, twice the age of male G. geographica in NFWR in 2005 to 2007. Gibbons et al. (1981) discovered that male slider turtles ( Trachemys scripta ) were maturing at a younger age in a pond receiving thermal effluents from a nuclear power plant. Growth curves presented by Gibbons et al. (1981), suggest that male T. scripta were also maturing at a slightly larger size (approximate ly 1.5 cm larger) in the impacted pond than did 112

PAGE 113

those in the natural pond (see Figu re 2 in Gibbons et al. 1981). If a similar size discrepancy of 1.5 cm were observed for G. geographica inhabiting NFWR, male G. geographica would mature at approximately three to four years in 1969 to 1972, the same age at maturity of those in 2005 to 2007. Mature male G. geographica in 1969 to 1972 would have been smaller than those present in 2005 to 2007 (Figure 4-6). In either scenario, male G. geographica in NFWR in recent years (2005 to 2007) would benefit from earlier age at maturity or larger size-at-age as their lifetime reproductive output and/or age-specific survival rate s should be enhanced. These results suggest that males in recent years are benefiting from an extension of growing seas on in terms of age at sexual maturity and/or size-at-age enhancements. Conclusions This study revealed patterns in growth rate between sexes that account for the extreme sexual size dimorphism observed in G. geographica Juvenile G. geographica grow at a much faster rate than adults (Figures 4-3 and 4-4). A secondary sexual characteristic, relative tail length, becomes apparent when plastron lengths reach 5 cm (Figure 4-1), which corresponds with approximately one year of age (Figure 4-4). Divergence in growth rates between male and female juveniles occu rs after one year of age and male G. geographica mature at approximately three to four years of age (Figure 4-4). Female G. geographica have higher growth rates than males and attain much larger body sizes (Table 4-1, Figure 4-4). This study provides evidence that an extension of the growing season can reduce the age of maturity of G. geographica Comparison of von Bertalanffy growth curves indicated that male G. geographica were growing more slowly in 1969 to 1972 than those in 2005 to 2007 (Table 4-1, Figure 4-6). The reduction in growth rate corresponds with an estimated older age at maturity of male G. 113

PAGE 114

geographica in 1969 to 1972 relative to those in 2 005 to 2007 (Figure 4-6). Additionally, for ages less than 16 years, male G. geographica in 1969 to 1972 are predicted to be smaller than those in 2005 to 2007 (Figure 4-6). Younger age at maturity is predicted to lead to increased lifetime reproductive output (Gibbons et al. 1981). Additionally, turtles with larger body sizes at any given age are expected to have a greater chance of surviv al (Gibbons et al. 1981, Janzen 1993, Bodie and Semlitsch 2000, Janzen et al. 2000, Tucker 2000). This study increases the breadth of knowledge regarding the genera Graptemys and the depth of knowledge specific to G. geographica It also provides insigh t into potential changes in life history characteristics that may occur with gl obal climate change as global climate change is predicted to increase temperatures and result in earlier onset of spri ng and summer thermal and hydrological regimes (Gates 1993, Intergovernment al Panel on Climate Change 2007). This knowledge will be useful for predicting how populations of G. geographica and perhaps other temperate freshwater turtle species, will adjust to a rapidly changing climate or climatic variation. 114

PAGE 115

Table 4-1. Von Bertalanffy grow th curve model variables for Graptemys geographica estimated using nonlinear regression techniques. Sex Asymptotic length (a) (cm) (95% confidence interval) Intrinsic growth rate (k) (95% confidence interval) R 2 1969 1972 Male 8.5 (7.8 9.1) 0.2548 (0.1618 0.3477) 0.939 2005 2007 Male 8.4 (8.1 8.8) 0.5062 (0.2449 0.7676) 0.981 Female 19.9 (18.7 21.1) 0.1674 (0.1326 0.2022) 0.997 115

PAGE 116

Table 4-2. Summary of von Be rtalanffy growth curve mode l variables available for Graptemys species. PL represents plastron lengt h. C.I. denotes confidence interval. Species/sex Instrinsic growth rate (k) (95% C.I.) Variable b (95% C.I.) Asymptotic PL (a) (cm) (95% C.I.) Data source G. geographica Male 0.506 (0.2449-0.7676) 0.717 8.4 (8.1.8) This study Female 0.167 (0.1326-0.2022) 0.864 19.9 (18.7-21.1) This study G. oculifera Male 0.388 (0.3155-0.4613) 0.578 7.82 Jones and Hartfield 1995 Female 0.176 (0.1262-0.2258) 0.754 13.41 Jones and Hartfield 1995 G. versa Male No data No data No data Lindeman 2005 Female 0.194 (0.017-0.371) 0.862 (0.622-1.102) 15.25 Lindeman 2005 G. caglei Male 0.445 (0.397-0.493) 0.631 (0.607-0.656) 9.15 Lindeman 1999 Female No data No data 16.70 Lindeman 1999 G. ouachitensis Male 0.459 (0.417-0.500) 0.740 (0.719-0.761) 9.63 Lindeman 1999 Female 0.182 (0.169-0.194) 0.854 (0.841-0.866) 17.53 Lindeman 1999 G. pseudogeographica Male 0.498 (0.438-0.559) 0.779 (0.753-0.805) 11.45 Lindeman 1999 Female 0.161 (0.149-0.174) 0.860 (0.849-0.872) 21.07 Lindeman 1999 G. ernsti Male 0.264 (0.248-0.279) 0.672 (0.662-0.681) 9.87 Lindeman 1999 Female 0.110 (0.104-0.116) 0.866 (0.857-0.875) 22.02 Lindeman 1999 116

PAGE 117

0 1 2 3 4 5 6 7 8 9 0 5 10 15 20 25Plastron length (cm)Tail length (cm) Unidentified Female Male Figure 4-1. Relationship between tail le ngth and plastron length observed in the Graptemys geographica population in the North Fork of White River, Ozark County, Missouri for the sampling period 2005 2007. 117

PAGE 118

0 5 10 15 20 25 00.10.20.30.40.50.60.70.8Tail length/plastron length ratioPlastron length (cm) Unidentified Female Male TL/PL = 0.45 PL = 5 cm Figure 4-2. Relationship between pl astron length and tail length/pla stron length ratio observed in the Graptemys geographica population in the North Fork of White River, Ozark County, Missouri for the sampling period 2005 2007. 118

PAGE 119

0 0.002 0.004 0.006 0.008 0.01 0.012 0.014 0.016 0 5 10 15 20 25Median plastron length (cm)Instantaneous growth rate Indeterminant sex Female Male Figure 4-3. General growth pattern observed in the Graptemys geographica population in the North Fork of White River, Ozark County, Missouri for the sampling period 2005 2007. Median plastron length indicates the median size of a turtle based on the plastron length measurements from the initial and final capture. 119

PAGE 120

0 5 10 15 20 25 0 5 10 15 20 25Age (years)Plastron length (cm) Female Male Figure 4-4. Von Bertalanffy growth curves for the Graptemys geographica population in the North Fork of White River, Ozark County, Missouri for the sampling period 2005 2007. 120

PAGE 121

0 5 10 15 20 25 0 5 10 15 20 25Age (years)Plastron length (cm) Unidentified Female Male von Bertalanffy Female von Bertalanffy Male Figure 4-5. Von Bertalanffy growth curves and estimated-age data for Graptemys geographica inhabiting the North Fork of White Rive r, Ozark County, Missouri for the sampling period 2005 2007. 121

PAGE 122

0 1 2 3 4 5 6 7 8 9 0 5 10 15 20 25Age (years)Plastron length (cm) 2005-2007 1969-1972 Figure 4-6. Von Bertalanffy growth curves for male Graptemys geographica inhabiting the North Fork of White River, Ozark County, Missouri for the sampling periods 1969 1972 and 2005 2007. 122

PAGE 123

CHAPTER 5 CONCLUSIONS AND RECOMMENDATIONS In this chapter, I integrated information presen ted in previous chapters to draw conclusions and make recommendations applicable to future studies both within and outside of NFWR. Conclusions Temporal and Spatial Variation in Abiotic Factors Influences River Turtle Community Composition, Species Ric hness, and Heterogeneity Nonequilibrium community theories integr ate the idea that a perpetually-changing environment (i.e., abiotic factor s) can be a major determini ng force influencing community composition (Chesson and Case 1986). A continua lly-shifting and hetero geneous environment can cause populations and communities to vary temporally and spatially (Davis 1986, Cody 1996). These ideas are in contrast to classica l competition theory which suggests competitive interactions primarily shape communities (Chesson and Case 1986). Analyses of temporally and spatially distinct turtle communities and G. geographica populations in NFWR support nonequilibrium commun ity theories that suggest variation in abiotic conditions can drive community composition. In Chapter 2, I elucid ated a temporal shift in composition of a turtle community located within a 4.6 km section of NFWR that had been studied periodically since 1969. The turtle comm unity shifted from an assemblage of one common species and several rare species to a gr ouping of several common species and a few rare species (Figure 2-3A). Analysis of species richness suggested the tu rtle community in 1969 typically had more species than the community inhabiting the same section in recent years (2005, 2006, and 2007) (Figure 2-4A). Comparisons of hete rogeneity measures indi cated that the turtle community was more heterogeneous in 2005, 2006, and 2007 than was the turtle community in the same section in 1969 (Table 2-1). These re sults were indicative of the increased species 123

PAGE 124

evenness associated with an increase in S. odoratus and T. s. elegans observed in the downstream section in 2005, 2006, and 2007 (Figure 2-3A). The observed community changes corresponded with changes in the abiotic environment. Riparian zone development, degradation of wate r quality, siltation, and sedimentation occurred between 1969 and recent years (Figures 2-1 and 2-2; Pitt 2005). Development of riparian zones is associated with increased siltation, sedimentation, sun expos ure, and bacteria and nutrientloading (Gilliam 1994, Pusey and Arthington 2003). These factors create ideal conditions for aquatic vegetation growth and algal blooms (Sm ith et al. 1999, Groom and Vynne 2006), such as that observed in recent years in NFWR. Anal ysis of basking substrate use indicated that T. s. elegans and S. odoratus basked overwhelmingly on vegetation in recent years (Table 2-2). Trachemys scripta elegans is a highly-adaptable, generalist spec ies that will exploit most habitats (Webb 1961, Vogt and Benitez 1997). The establis hment of aquatic vegetation stands and the presence of long-lived, extensive algal blooms provide ample amounts of basking habitat that T. s. elegans are able to exploit. Similarly, the increase in abundance of the S. odoratus population in NFWR in 2005 corresponded with the growth of emergent vegetation stands and associated algal mats. All but four of the 130 S. odoratus observed basking were associated with vegetation (Table 2-2). These emergent vegetation stands were associated with muddy, silty substrates, a typical habitat of S. odoratus (Kingsbury 1993). Prior to th e establishment of extensive emergent vegetation stands, few S. odoratus were observed in NFWR (Pitt 2005), but S. odoratus comprised a major portion of the turtle community in 2005 to 2007 (Figure 2-3A). Therefore, I conclude that abio tic changes, including siltation and sedimentation, resulted in substantial vegetation growth and subsequent temporal shifts in tu rtle populations and communities in NFWR. 124

PAGE 125

Spatial variation among communities was dete cted when the above described turtle communities, which are located in a 4.6 km section located down stream of major spring flows entering NFWR, were compared with the turt le communities inhabiting a 4.6 km section of NFWR located upstream of the major spring flow s. The community in the downstream section typically had lower species richness, indicating fewer species, than the turtle community in upstream section in recent years (Figures 2-4C D, and F). The turtle community in the downstream section was more heterogeneous than the upstream turtle community in recent years (Table 2-1). These results indicate that the upstream section cont ained more rare turtle species than the downstream section in 2005 to 2007. Differences in turtle communities inhabiting the two sections corresponded with differences in habitat. Water temperature wa s warmer in the upstream section than in the downstream section. More species are expected to inhabit areas with warmer water as turtles tend to move to and among areas of favorable thermal regimes (Schuett and Gatten 1980, Thornhill 1982, Moll and Moll 2004) and the dist ribution of some species may be limited by temperature. The upstream section exhibited fe wer symptoms of human impact than did the downstream section. Riparian zone development was less prevalent in th e upstream section than in the downstream section. The upstream section had fewer stations that tested positive for excessive E. coli levels (Figure 2-2). The upstream section had less aquatic vegetation, indicating less nutrient-loading. If turtles prefer less-impacted habitats, more species would be expected in the less-impacted habitat. The community composition, species richness, a nd heterogeneity of the turtle communities located in the upstream section in 2005 to 2007 and in the downstream section in 1969 were relatively similar (Figures 2-3 A and 2-4A and B, Table 2-1). These similarities may arise from 125

PAGE 126

the comparison of two equally unimpacted (by humans) areas. Moll (1 977, 2006) found that the community composition within an impacted river changed from being predominately specialized turtle species to consisting of ma inly generalized turtle species. The higher relative proportion of the generalist T. s. elegans and S. odoratus observed in the downstream section may be symptomatic of a more heavily-impacted habitat. Understanding how communities will change in the face of a continuously-shifting environment is ever important as the huma n population continues to grow (Cohen 1995) and development and alteration of land and wate rways are pervasive (Benke 1990, Riccardi and Rasmussen 1999). The relevance of this statement is enhanced in the face of global climate change. Without understanding temporal and spatial variation in communities, it will be difficult to implement effective and effi cient conservation and management actions. The need for such data is enhanced for conserving long-liv ed, imperiled taxa such as turtles. The North Fork of White River, Ozark Co unty, Missouri and Associated Springs (i.e., Recharge Areas) Are Contamin ated with Coliform Bacteria Contamination of surface water poses threats to ecosystems, wildlife, and human health (Buhlmann and Gibbons 1997, The Nature Conserva ncy 2003). In areas with porous karst geology, such as the Ozark Plateau, surface water and soil contaminants in recharge areas can flow and seep, respectively, into the groundwater polluting underground aquifers (The Nature Conservancy 2003). Drinking water obtained from contaminated underground aquifers poses a risk to human health (The Nature Conservanc y 2003). In Chapter 2, I provided evidence that NFWR and its associated springs are contaminated with coliform bacteria. Levels of total coliform bacteria exceeded limits deemed safe for full body contact by th e Missouri Department of Natural Resources (MDNR) in at least one sampling event for all sites sampled during the summer of 2007 (Figure 2-1). Escherichia coli levels in all of the springs, downstream sampling 126

PAGE 127

sites, and one upstream sampling site exceeded limits deemed safe for full body contact by MDNR (Figure 2-2). NFWR is a heavily-used river for recreation, in cluding canoeing, tubing, and swimming (Pitt 2005). The excessive levels of coliform bacteria within NFWR pose a risk to human health and may foreshadow high concentrations of other contaminants, such as pesticides and metals. The pres ence of pesticides has been c onfirmed in previous studies of NFWR (MDC 2005). Global Climate Change May Be Enhancing Gr owth Rates of River Turtles Inhabiting Temperate Climates by Extending the Growin g Season and May Be Leading to Earlier Age at Maturity and Age at Size Benefits Global climate change has been implicated as detrimental to a suite of organisms, including turtles (Root and Schneider 1993, Gl en and Mrosovsky 2004, Willette et al. 2005, Parmesan 2006, Hawkes et al. 2007). For turtle s, global warming may cause skewed sex ratios of species with temperature-dependent sex dete rmination (Janzen 1994) and decrease the energy reserves of overwintering neonates (Willette et al. 2005). However, turtle physiology is temperature-dependent and increases in environmental temperatures have been linked with increases in growth rates in turtles (C agle 1946, Gibbons et al. 1981, Rhen and Lang 1995, Roosenburg and Kelley 1996). Increases in growth rate can lead to increased survivorship, earlier sexual maturation, and larger body size at various stag es, including maturation (Bury 1979, Thornhill 1982, Cox et al. 1991, Tucker 2000) Larger body size and earlier sexual maturation of turtles can lead to reproductive enhancements such as earlier age at reproduction, increased clutch size, and increased size or quality of eggs (Congdon and Gibbons 1983, 1990, 1996; Lovich et al. 1998). Analyses of G. geographica populations inhabiting the sa me area during different time periods suggested that the length of the grow ing season was significantly positively-correlated with instantaneous growth rates of individuals. In Chapter 3, I determined that G. geographica 127

PAGE 128

for which sex was visibly indistinguishable (i.e ., indeterminate sex) in NFWR in 2004 to 2007 grew approximately 2.8 times faster than similarl y-sized turtles in the same section in 1969 to 1972 (Table 3-2). This pattern was significantly correlated with an increased number of growth days in 2004 to 2007 than in 1969 to 1972 (Tables 3-6 and 3-7, Figure 3-1). A comparison of the number of growth days per month revealed that March, April, May, August, and October all had more growth days in 2004 to 2007 than in 1969 to 1972 (Table 3-7). The increased number of growth days in March and Oc tober in 2004 to 2007 corresponds with an extension of the growing season as these m onth mark the boundaries of G. geographica s active season in Missouri (Johnson 2000). Similar results were fo und by Frazer et al. (1993) for a population of painted turtles ( Chrysemys picta ) in Michigan. These resu lts support the hypothesis that extended growing seasons are positively affecting the growth rate of small juvenile turtles. In Chapter 4, I demonstrated that male G. geographica in NFWR in 2005 to 2007 were growing faster and attaining asymptotic size more rapidly than those in 1969 to 1972 (Figure 46). Differences in size-at-age suggest that male G. geographica in 1969 to 1972 took twice as long to mature as those in 2005 to 2007 in NFWR. However, male G. geographica may have matured at smaller sizes and the same age in 1969 to 1972 than in 2005 to 2007. Regardless, male G. geographica in NFWR in 2005 to 2007 would benefit from augmented growth rate leading to earlier age at maturity or larger size-at-age. Earlier maturation is linked with increased lifetime reproductive output (Gibbons et al. 1981) and larger body sizes are associated with greater survival (Gibbons et al. 1981, Janzen 1993, Bodie a nd Semlitsch 2000, Janzen et al. 2000, Tucker 2000). Therefore, male G. geographica in NFWR are benefiting from extended growing seasons in terms of enhanced growth rate. Male G. geographica may also be benefiting 128

PAGE 129

from reproductive and/or age-specific survival enhancements. Not enough data were available for female G. geographica from 1969 to 1972 to evaluate if patterns were similar between sexes. Global climate change is resulting in overa ll warming trends (Schlesinger and Jiang 1991, Manabe and Stouffer 1993, Intergovernmental Pa nel on Climate Change 2007) and the impacts of global warming on ecosystems and their comp onent communities and populations must be elucidated. Global climate change models predic t earlier onset of spring and summer thermal regimes and warmer temperatures (Gates 1993, Intergovernmental Panel on Climate Change 2007). These predictions are consistent with clim atic patterns observed in this study. If global climate change is responsible for the observed extended growing seasons, global climate change may be enhancing the growth rate of G. geographica within NFWR and I expect similar patterns may be occurring for other temperate river turtle species as global warming is pervasive and turtle physiology is relatively pr edictable. However, the observed extension in growing season may be due to natural variation rather than gl obal climate change. Regardless of the cause of climatic variation observed within this study, the knowledge gained will be useful for predicting how populations of G. geographica and perhaps other temperate fr eshwater turtle species, will respond to a rapidly changing climate. Recommendations Assess Turtle Communities for Shifts from Predominately Specialist Species to More Generalist Species North America supports diverse freshwater turtle communities, yet few have received more than a passing account of species presence (E rnst et al. 1994, Moll and Moll 2004). This study revealed that turtle communities can be dyna mic and changes in community composition and heterogeneity can coincide with habitat degradat ion. Symptoms of declining water quality, such as excessive bacterial concentrations and vege tation growth, coincided with an increase in 129

PAGE 130

generalist turtle specie s. Moll (1977, 2006) observed similar patterns in the turtle community inhabiting the Illinois River. Typically, small, short-lived organisms, such as aquatic macroinvertebrates, are used as indicator species of environmental degradation and biomonitoring (Lenat and Barbour 1994), but the us e of long-lived species, such as turtles, may provide information more relevant to other longlived species, such as humans. Additionally, turtles are imperiled worldwide and the conseque nces of species losses are poorly understood. Monitoring turtle communities can provide insi ghts into what changes are happening and the effects of these changes. A major limitation in monitoring lo ng-lived species is that studi es must be long-term. To confound issues further, turtle populations are declining worldwide (Smith 1979, Moll 1986, Kuchling 1988, Ernst et al. 1994, Buhlma nn and Gibbons 1997, Gibbons 1997, Haitao 2000, van Dijk et al. 2000) and most rivers are excessi vely modified (Lydeard and Mayden 1995, Ricciardi and Rasmussen 1999). However, information exis ts in the form of museum collections and historical species accounts, so some informati on can be gleaned from careful examination of such records (Moll and Moll 2004). Additionally, data that has ne ver been published often exists in the form of field notebooks hidden away in o ffices or storage (personal observation). In a time of monumental species declines and losses, it is imperative that collaborative efforts be made to exhume these data sets, reassess commun ities previously studied, and publish the results of these efforts. Assess Competitive Interactions and Ability between Generalist and Specialist Turtle Species The turtle community in NFWR changed in composition and heterogeneity between 1969 and 2005, yielding an increased re lative and absolute abundan ce of the generalist species T. s. elegans and S. odoratus (Figure 2-3A). Between 1980 and 2004, the number of permanent 130

PAGE 131

basking structures (i.e., rocks) declined in NFWR (Pitt 2005). Basking is necessary for thermoregulation, digestion, maintenance of inte gumentary health, vitamin D synthesis, and parasite removal (Hutchison 1979, Vogt 1979, Ha rt 1983, Saumure and Livingston 1994, Vogt and Benitez 1997). Many species of turtles ba sk and I observed all present species except M. temminckii basking at least once during the study period (Table 2-2). Niche overlap measurements indicated a signi ficant overlap in bask ing habitat use among turtle species in NFWR (Table 2-4). High ni che overlap can result from shared resource utilization instigated by lack of competition or intense competition that has not yet resulted in resource partitioning (Pianka 1976). Niche breadth analysis revealed that G. geographica frequently used rocks and logs as basking substrates in 1969 to 1972, but in recent years had increased its niche breadth (Tables 22 and 2-3). Trachemys scripta elegans and S. odoratus were most commonly observed basking on vegetation (Table 2-2). Niche breadth analysis revealed that S. odoratus was a basking habitat specialist using vegetati on almost exclusively in the do wnstream section in 2005 to 2007 (Tables 2-2 and 2-3). I suspect that competition was responsible for the observed patterns of basking habitat use, with G. geographica being a better competitor than T. s. elegans and S. odoratus However, no aggressive interactions were obs erved among these species in this study. With the potential for increased numbers of generalist species in sy stems with degraded ha bitats, I suggest that competition among turtle species be thoroughly asse ssed in order to predict subsequent impacts of these community changes. Conduct a Thorough Assessment of Water Quality in NFWR The majority of rivers worldwide have undergone massive alterations (Lydeard and Mayden 1995, Ricciardi and Rasmus sen 1999) and declines and extin ctions of aquatic species 131

PAGE 132

are common (Ricciardi and Rasmussen 1999). Coliform bacteria analyses indicated that NFWR and its associated springs were highly contaminated, posing a risk to human health. The establishment of silt and sediment deposits in NFWR indicated that development and associated storm runoff altered the rivers substrate and a llowed emergent vegetation stands to grow. Pesticides have also been detected in NFWR (MDC 2005). I recommend a thorough examination of water quality in NFWR and its asso ciated springs. I also recommend that turtles be examined for bioaccumulation of toxins to rev eal long-term patterns of contaminant exposure. In addition, I suggest that development along NFWR be evaluated and regulated by state officials to prevent unnecessary r unoff and septic contamination of the river. At the same time, landowners should be encouraged to restore and maintain native ve getation in riparian zones to reduce runoff. Conduct a Comprehensive Assessment of Glo bal Warmings Affects on River Turtles Results of this study suggested that global warmi ng or climatic variation that is consistent with predictions of global climate change models may be positively affecting turtle growth rates which may have additional benefits including earlier age or larger size at maturation. However, previous studies of global warmings affects on temperature-dependent sex determination in turtles suggested that global warming was femi nizing turtles which would reduce recruitment in subsequent generations as male turtles would be rare (Janzen 1994). Additional studies have also suggested that reptiles, such as turtle s, are not mobile enough to undergo range shifts through fragmented habitats in the face of a rapidly changing climate (Araujo et al. 2006). Global warming may also reduce energy stores of overwintering hatchling turtles that remain in the nest (Willette et al. 2005). Reduced energy reserves may negatively impact survival rates (Willette et al. 2005) I suggest th at comprehensive studies of global warmings affects on turtles 132

PAGE 133

be conducted in many parts of the world and with diverse species as climat e change is predicted to affect various locations differently, especia lly in terms of water resources (Arnell 1999). Assess the Relative Importance of Temperatu re versus Length of Growing Season on Turtle Growth Phenotypic plasticity in traits, such as growth rate, that can directly influence life history parameters is important for organisms inhabiti ng variable environments (Caswell 1983, Stearns and Koella 1986). For turtles and other ectotherms temperature is an influential abiotic factor that can alter phenotypic traits that directly modify lif e history parameters (Laudien 1973, Lillywhite et al. 1973, Gibbons et al. 1981, Bronikowski et al. 2001). Temperature has been positively correlated with growth rate of turt les (Cagle 1946, Gibbons et al. 1981, Rhen and Lang 1995, Roosenburg and Kelley 1996). Enhanced growth rate has been correlated with increased survivorship, earlier sexual maturation, and larger body size at va rious stages, including maturation (Bury 1979, Thornhill 1982, Cox et al. 1991, Tucker 2000). Earlier sexual maturation and increased body size has been linked with increased lifetime reproductive output (Congdon and Gibbons 1983, 1990, 1996; Lovich et al. 1998). Comparisons of turtle populations inhabiti ng thermally-distinct habitats have been conducted to assess the impact of power-plant effluents on turtles (Parmenter 1980, Thornhill 1982, Spotila et al. 1984). Studies have implicated temperature as a factor influencing turtle growth rate and size and age at maturity (Parmenter 1980, Gibbons et al. 1981, Thornhill 1982). In Chapter 3, I determined that G. geographica inhabiting two sections of a spring-fed river with significantly different thermal re gimes did not exhibit differences in mean instantaneous growth rate (Table 3-3). This result was unexpected as similar differences in water temperature at other sites have been implicated as contributors to variation in turtle growth rates 133

PAGE 134

(Parmenter 1980, Thornhill 1982). The discrepanc y in results between this study and others suggest that difference in temperature alone cannot incr ease turtle growth rate. Studies in which differences in growth rate were observed among turtles had the commonality of increased length of the growing season (Thornhill 1982). I hypothesized that the discrepancy in temperature between the two sites may be countera cted by the downstream section having an extended growi ng season relative to the upstream section. I proposed that the spring effluents buffer the water temperature of the downstream section while the water temperature of the upstream section varies naturall y with air temperature. This buffering effect could increase the growing season of the downs tream section relative to the upstream section, especially for basking species that overcome the constraints of their aquatic thermal environment. I suggest that controlled experi ments be conducted to assess whether differences in water temperature alone can influence turtle grow th rate if all other variables are held constant and food is not limiting. Not only would this pro posed experiment clarify factors that influence turtle physiology, but it would further elucidate the mechanism behind how global warming may affect turtle growth rate. I also recommend evaluating the ac tive season of turtles inhabiting different sections of NFWR to confirm or refu te whether growing seasons vary in length among areas with differing thermal regimes. Continue Monitoring the Turtle Community in NFWR to Pr ovide Insight Regarding the Continued Impacts of Climatic Variatio n and Habitat Alterations on Turtle Communities Few studies provide data from several decades and long-term monitoring is necessary to observe temporal variations in populations and communities. As development, pollution, and climate change are omnipresent, it is imperativ e to determine how these factors will impact communities and ecosystems over both short and longterm timescales. The dearth of long-term studies must be overcome to adequately predict how various influences will impact communities. 134

PAGE 135

The few communities and populations from which long-term data are available provide unique insights and opportunities for assess ing continuous changes to the environment. Therefore, I recommend that the turtle community in NFWR continue to be monitored for many more decades. Final Remarks The turtle community in NFWR presents a uniq ue opportunity to assess changes in turtle populations, communities, and habitats in the f ace of riparian development and climatic variation. As few long-term data are availabl e to assess turtle comm unities and populations and turtles are imperiled worldwide, we must take advantage of the few opportunities we do have to study these communities and populati ons if we wish to document changes that are occurring in association with human-borne envi ronmental changes. Without this documentation, it is unlikely that financial, political, and public su pport of turtle conservation will improve. 135

PAGE 136

LIST OF REFERENCES Araujo, M. B., W. Thuiller, and R. G. Pearson. 2006. Climate warming and the decline of amphibians and reptiles in Eur ope. Journal of Biogeography 33:1712-1728. Arndt, R. G. and W. A. Potter. 1973. A population of the map turtle, Graptemys geographica in the Delaware River, Pennsylva nia. Journal of Herpetology 7:375377. Arnell, N. W. 1999. Climate change and gl obal water resources. Global Environmental Change 9:S31-S49. Benke, A. C. 1990. A perspective on America s vanishing streams. Journal of the North American Benthological Society 9:77-88. Bodie, J. R., and R. D. Semlitsch. 2000. Size-specific mortality and natural selection in freshwater turtles. Copeia 2000:732-739. Bodie, J. R., R. D. Semlitsch, and R. B. Re nken. 2000. Diversity and structure of turtle assemblages: associations with wetland characters across a floodplain landscape. Ecography 23:444-456. Boyer, D. R. 1965. Ecology of the bask ing habit in turtles. Ecology 46:99-118. Braid, M. R. 1974. A bal-chatri trap for basking turtles. Copeia 1974:539-540. Brody, S. 1945. Bioenergetics and growth with special reference to the efficiency complex in domestic animals. Reinhol d Publishing Corporation, New York, New York, USA. Bronikowski, A. M., A. F. Bennett, and R. E. Lenski. 2001. Evolutionary adaptation to temperature. VIII. Effects of temperatur e on growth rate in natural isolates of Eschericia coli and Salmonella enterica from different thermal environments. Evolution 55:33-40. Brown, G. P., C. A. Bishop, and R. J. Brooks 1994. Growth rate, reproductive output, and temperature selection of snapping turtles in habitats of different productivities. Journal of Herpetology 28:405-410. Brown, G. P., and R. J. Brooks. 1991. Therma l and behavioral responses to feeding in free-ranging turtles, Chelydra serpentina Journal of Herpetology 25:273-278. Bryant Watershed Project, Inc. 2008. Wa tersheds: North Fork Watershed. Bryant Watershed Project, Inc., West Plains, Missouri, USA. http://www.watersheds.org/places /nf.htm. Accessed January 2008. 136

PAGE 137

Buhlmann, K. A. and J. W. Gibbons. 1997. Imperiled aquati c reptiles of the southeastern United States: historical revi ew and current conservation status. Pages 201-232 in G. W. Benz and D. E. Collins, editors. Aquatic fauna in peril: the southeastern perspective. Special publication 1. Sout heastern Aquatic Research Institute, Lenz Design and Communications, Decatur, Georgia, USA. Buhlmann, K. A. and T. D. Tuberville. 1998. Use of passive integrated transponder (PIT) tags for marking small freshwater turtles. Chelonian Conservation and Biology 3:102-104. Bull, J. J., R. C. Vogt, and C. J. McCoy. 1982. Sex determining temperatures in turtles: a geographic comparison. Evolution 36:326-332. Burkhead, N. M., S. J. Walsh, B. J. Freem an, and J. D. Williams. 1997. Status and restoration of the Etowah River, an imperiled southern Appalachian ecosystem. Pages 375-444 in G. W. Benz and D. E. Collins, editors. Aquatic fauna in peril: the southeastern perspec tive. Special publication 1. Southeastern Aquatic Research Institute, Lenz Design and Co mmunications, Decatur, Georgia, USA. Bury, R. B. 1979. Population ecology of freshwater turtles. Pages 571-602 in M. Harless and H. Morlock, edito rs. Turtles: perspectives and research. John Wiley & Sons, Inc., New York, New York, USA. Cadi, A. and P. Joly. 2003. Competition fo r basking places between the endangered European pond turtle (Emys orbicularis galloitalica ) and the introduced red-eared slider ( Trachemys scripta elegans ). Canadian Journal of Zoology 81:1392-1398. Cadi, A. and P. Joly. 2004. Impact of the introduction of the red-eared slider ( Trachemys scripta elegans) on survival rates of th e European pond turtle (Emys orbicularis ). Biodiversity and Conservation 13:2511-2518. Cagle, F. R. 1942. Turtle populations in southern Illinois. Copeia 1942:155-162. Cagle, F. R. 1946. The growth of the slider turtle, Pseudemys scripta elegans The American Midland Naturalist 36:685-729. Cagle, F. R. 1950. The life history of the slider turtle, Pseudemys scripta troostii (Holbrook). Ecologica l Monographs 20:31-54. Casper, G. S. 1997. Geographic distribution: Graptemys geographica (common map turtle). USA: Wisconsin. Herpetological Review 28:49. Caswell, H. 1983. Phenotypic plasticity in lif e-history traits: de mographic effects and evolutionary consequences. American Zoologist 23:35-46. Chaney, A. and C. L. Smith. 1950. Met hods for collecting mapturtles. Copeia 1950:323-324. 137

PAGE 138

Chen, T., and K. Lue. 2002. Growth patterns of the yellow-marginated box turtle ( Cuora flavomarginata) in northern Taiwan. Jour nal of Herpetology 36:201-208. Chesson, P. L., and T. J. Case. 1986. Over view: nonequilibrium co mmunity theories: change, variability, history, and coexistence. Pages 229-239 in J. Diamond and T. J. Case, editors. Community ecology. Harper and Row, New York, New York, USA. Cochran, P. A. 1987. Graptemys geographica (map turtle). Adult mortality. Herpetological Review 18:37. Cody, M. L. 1996. Introduction to long-term community ecological studies. Pages 1-15 in M. L. Cody and J. Smallwood, editors. Long-term studies of vertebrate communities. Academic Press, Inc., San Diego, California, USA. Cohen, J. E. 1995. Population growth and Earths human carrying capacity. Science 269:341-346. Congdon, J. D., A. E. Dunham, and R. C. van Loben Sels. 1993. Delayed sexual maturity and demographics of Blandings turtles ( Emydoidea blandingii): implications for conservation and mana gement of long-lived organisms. Conservation Biology 7:826-833. Congdon, J. D., A. E. Dunham, and R. C. van Loben Sels. 1994. Demographics of common snapping turtles ( Chelydra serpentina ): implications for conservation and management of long-lived organism s. American Zoologist 34:397-408. Congdon, J. D., and J. W. Gibbons. 1983. Relati onships of reproducti ve characteristics to body size in Pseudemys scripta Herpetologica 39:147-151. Congdon, J. D., and J. W. Gibbons. 1990. The evol ution of turtle life histories. Pages 45-54 in J. W. Gibbons, editor. Life histor y and ecology of the slider turtle. Smithsonian Institution Pre ss, Washington, D.C., USA. Congdon, J. D., and J. W. Gibbons. 1996. Structure and dynamics of a turtle community over two decades. Pages 137-159 in M. L. Cody and J. Smallwood, editors. Longterm studies of vertebra te communities. Academic Press, Inc., San Diego, California, USA. Congdon, J.D., J.L. Greene, and J.W. Gibbons. 1986. Biomass of freshwater turtles: a geographic comparison. The American Midland Naturalist 115:165-173. Convention on International Trade in Enda ngered Species of Wild Fauna and Flora (CITES). 2005. How CITES works. CITES, Geneva, Switzerland. http://www.cites.org/eng/disc/ how.shtml. Accessed April 2005. 138

PAGE 139

Cox, W. A., J. B. Hazelrig, M. E. Turner, R. A. Angus, and K. R. Marion. 1991. A model for growth in the musk turtle, Sternotherus minor in a north Florida Spring. Copeia 1991:954-968. Daigle, C., A. Desrosiers, and J. Bonin. 1994. Distribution and abundance of common map turtles, Graptemys geographica in the Ottawa River, Quebec. The Canadian Field-Naturalist 108:84-86. Davis, M. B. 1986. Climatic instability, time lags, and community disequilibrium. Pages 269-284 in J. Diamond and T. J. Case, editors. Community ecology. Harper and Row, New York, New York, USA. Diamond, J. 1986. Overview: laboratory expe riments, field experiments, and natural experiments. Pages 3-22 in J. Diamond and T. J. Case, editors. Community ecology. Harper and Row, New York, New York, USA. Dodd, C. K., Jr. 1978. Amphibians and reptile s: the declining species. Water Spectrum 10:24-32. Dodd, C. K., Jr. and R. Franz. 1993. Th e need for status information on common herpetofaunal species. He rpetological Review 24:47-50. Dreslik, M. J. 1998. Curre nt status and conservati on of the river cooter ( Pseudemys concinna ) in southern Illinois. Chelon ian Conservation and Biology 3:135-137. Dunning, J. B., Jr., M. J. Groom, and H. R. Pulliam. 2006. Species and landscape approaches to conservation. Pages 419-465 in M. J. Groom, G. K. Meffe, and C. R. Carroll, editors. Principles of conservation biology. 3 rd edition. Sinauer Associates, Inc. Sunderland, Massachusetts, USA. Ellis, M. M. 1936. Erosion silt as a factor in aquatic environments. Ecology 17:29-42. Ernst, C. H. 1975. Growth of the spotted turtle, Clemmys guttata Journal of Herpetology 9:313-318. Ernst, C. H., J. E. Lovich, and R. W. Bar bour. 1994. Turtles of the United States and Canada. Smithsonian Institution Press, Washington D.C., USA and London, UK. Flaherty, N. and J. R. Bider. 1984. Physi cal structures and th e social factor as determinants of habitat use by Graptemys geographica in southwestern Quebec. The American Midland Naturalist 111:259-266. Foscarini, D. A. and R. J. Brooks. 1997. A proposal to standardize data collection and implications for management of the wood turtle, Clemmys insculpta and other freshwater turtles in On tario, Canada. Pages 203-209 in Proceedings: conservation, restoration, and management of tortoises a nd turtles An intern ational conference. New York Turtle and Tortoise So ciety, New York, New York, USA. 139

PAGE 140

Frazer, N. B., J. W. Gibbons, and J. L. Gr eene. 1990a. Exploring Fabens growth interval model with data on a long-lived vertebrate, Trachemys scripta (Reptilia: Testudinata). Copeia 1990:112-118. Frazer, N. B., J. W. Gibbons, and J. L. Green e. 1991. Life history and demography of the common mud turtle Kinosternon subr ubrum in South Carolina, USA. Ecology 72:2218-2231. Frazer, N. B., J. W. Gibbons, and T. J. Owens 1990b. Turtle trapping: preliminary tests of conventional wisdom Copeia 1990:1150-1152. Frazer, N. B., J. L. Greene, and J. W. Gibbons. 1993. Temporal variation in growth rate and age at maturity of male painted turt les, Chrysemys picta. American Midland Naturalist 130:314-324. Fuselier, L. and D. Edds. 1994. Habitat pa rtitioning among three sy mpatric species of map turtles, genus Graptemys. Journal of Herpetology 28:154-158. Gates, D. M. 1993. Climate change a nd its biological consequences. Sinauer Associates, Inc. Sunderland, Massachusetts, USA. Gatten, R. E., Jr. 1974. Effect of nutritiona l status on the prefe rred body temperature of the turtles Pseudemys scripta and Terrapene ornata Copeia 1974:912-917. Gibbons, J. W. 1997. Measuring declines a nd natural variation in turtle populations: spatial lessons from long-term studies. Pages 243-246 in Proceedings: Conservation Restoration, and Management of Tortoises and Turtles An International Conference. New York Tur tle and Tortoise So ciety, New York, New York, USA. Gibbons, J. W., R. D. Semlitsch, J. L. Greene, and J. P. Schubauer. 1981. Variation in age and size at maturity of the slider turtle ( Pseudemys scripta ). The American Naturalist 117:841-845. Gilliam, J. W. 1994. Riparian wetlands a nd water quality. Journal of Environmental Quality 23:896-900. Glen, F., and N. Mrosovsky. 2004. Antigua revi sited: the impact of climate change on sand and nest temperatures at a hawksbill turtle ( Eretmochelys imbricata ) nesting beach. Global Change Biology 10:2036-2045. Gordon, D. M., and R. D. MacCulloch. 1980. An investigation of the ecology of the map turtle, Graptemys geographica (La Sueur), in the northern part of its range. Canadian Journal of Zoology 58:2210-2219. Graham, T. E. 1979. Life hi story techniques. Pages 73-95 in M. Harless and H. Morlock, editors. Turtles: perspectives and research. John Wiley & Sons, Inc., New York, New York, USA. 140

PAGE 141

Graham, R. E. and A. A. Gr aham. 1992. Metabolism and behavior of wintering common map turtles, Graptemys geographica in Vermont. Canadian Field-Naturalist 106:517-519. Groom, M. J. 2006. Threats to biodiversity. Pages 63-109 in M. J. Groom, G. K. Meffe, and C. R. Carroll, editors. Princi ples of conservation biology. 3 rd edition. Sinauer Associates, Inc., Sunderland, Massachusetts, USA. Groom, M. J., and C. H. Vynne. 2006. Ha bitat degradation and loss. Pages 173-212 in M. J. Groom, G. K. Meffe, and C. R. Carro ll, editors. Principles of conservation biology. 3 rd edition. Sinauer Associates, In c., Sunderland, Massachusetts, USA. Haitao, S. 2000. Results of turtle market surveys in Chengdu and Kunming. Turtle and Tortoise Newsletter 6:15b-16. Hart, D. R. 1983. Dietary and habitat shift with size of red-eared turtles ( Pseudemys scripta ) in a southern Louisiana popu lation. Herpetologica 39:285-290. Hawkes, L. A., A. C. Broderick, M. H. Godf rey, and B. J. Godley. 2007. Investigating the potential impacts of clim ate change on a marine turtle population. Global Change Biology 13:923-932. Hellgren, E. C., R. T. Kazmaier, D. C. Ru thven III, and D. R. Synatzske. 2000. Variation in tortoise life history: demography of Gopherus berlandieri Ecology 81:1297-1310. Heppell, S. S. 1998. Application of life-hist ory theory and population model analysis to turtle conservation. Copeia 1998:367-375. Huestis, D. L., and P. A. Meylan. 2004. The turtles of Rainbow Run (Marion County, Florida): observations on the genus Pseudemys. Southeastern Naturalist 3:595612. Hurlbert, S. H. 1971. The nonconcept of speci es diversity: a critique and alternative parameters. Ecology 52:577-586. Hutchison, V. H. 1979. Th ermoregulation. Pages 207-228 in M. Harless and H. Morlock, editors. Turtles: perspectives and research. John Wiley & Sons, Inc., New York, New York, USA. Intergovernmental Panel on Climate Change. 2007. Intergovernmental Panel on Climate Change fourth assessment report: climate change 2007: synthesi s report. Summary for policy makers. Intergovernmental Panel on Climate Change, Geneva, Switzerland. Iverson, J. B. 1988. Growth in the common map turtle, Graptemys geographica Transactions of the Kansas Academy of Science 91:153-157. 141

PAGE 142

Iverson, J. B. 1992. Global correlates of species ric hness in turtles. Herpetological Journal 2:77-81. Iverson, J. B., E. L. Barthelmess, G. R. Sm ith, and C. E. DeRivera. 1991. Growth and reproduction in the mud turtle Kinosternon hirtipes in Chihuahua, Mexico. Journal of Herpetology 25:64-72. Jacobson, E. R. 1997. Diseases in wild populati ons of turtles and tortoises: the chelonian charisma vs. coincidence conundrum. Pages 87-90 in Proceedings: conservation, restoration, and management of tortoises a nd turtles an intern ational conference. New York Turtle and Tortoise So ciety, New York, New York, USA. Janzen, F. J. 1993. An experimental analysis of natural sele ction on body size of hatchling turtles. Ecology 74:332-341. Janzen, F. J. 1994. Climate change and temperature-dependent sex determination in reptiles. Proceedings of the National Academy of Sciences 91:7487-7490. Janzen, F. J., J. K. Tucker, and G. L. Paukstis. 2000. Experimental analysis of an early life-history stage: selec tion on size of hatchling turtles. Ecology 81:2290-2304. Johnson, T. R. 2000. The Amphibians and reptiles of Missouri. 2 nd edition. Conservation Commission of Missouri, Jefferson City, Missouri, USA. Jones, R. L., and P. D. Hartfield. 1995. Population size and gr owth in the turtle Graptemys oculifera Journal of herpetology 29:426-436. Kannan, K., M. Ueda, J. A. Shelby, M. T. Mendonca, M. Kawano, M. Matsuda, T. Wakimoto, and J. P. Giesy. 2000. Polychlorinated dibenzo-p-dioxins (PCDDs), dibenzofurans (PCDFs), biphenyls (PCBs), and organochlorine pesticides in yellow-blotched map turtle from the Pas cagoula River Basin, Mississippi, USA. Archives of Environmental Contaminant Toxicology 38:362-370. King, R. B., M. J. Oldham, W. F. Weller, and D. Wynn. 1997. Historic and current amphibian and reptile distributions in the island region of western Lake Erie. The American Midland Naturalist 138:153-173. Kingsbury, B. A. 1993. Habitat use by the common musk turtle, Sternotherus odoratus in Crooked Lake, Indiana. Annual meetings of the Indiana Academy of Sciences, Purdue University, West Lafayette, Indiana. Kinsolving, A. D., and M. B. Bain. 1993. Fi sh assemblage recovery along a riverine disturbance gradient. Ecological Applications 3:531-544. Kiviat, E. and D. Buso. 1977. Geographic distribution: Graptemys geographica (map turtle). USA. New York. Herpetological Review 8:84. 142

PAGE 143

Klemens, M. W. 2000. Turtle conservation. Smithsonian Institution Press, Washington D.C., USA and London, UK. Koper, N., and R. J. Brooks. 2000. Envir onmental constraints on growth of painted turtles ( Chrysemys picta ) in northern climates. Herpetologica 56:421-432. Krebs, C. J. 1989. Ecological methodology. Harper Collins, New York, New York, USA. Kuchling, G. 1988. Population structure, reproductive potential, and increasing exploitation of the freshwater turtle Erymnochelys madagascariensis. Biological Conservation 43:107-113. Lagler, K. F. 1943. Methods of collecti ng freshwater turtles. Copeia 1943:21-25. Laudien, H. 1973. Changing reaction systems. Pages 355-399 in H. Precht, J. Christophersen, H. Hensel, and W. Larche r, editors. Temperature and life. Springer-Verlag, New York, New York, USA. Lenat, D. R., and M. T. Barbour. 1994. Using benthic macroinvertebrate community structure for rapid, cost-effective water qua lity monitoring: rapi d bioassessment. Pages 187-216 in S. L. Loeb and A. Spacie, edit ors. Biological monitoring of aquatic systems. Lewis Publishers, Boca Raton, Florida, USA. Lillywhite, H. B., P. Licht, and P. Ch elgren. 1973. The role of behavioral thermoregulation in the grow th energetics of the toad, Bufo boreas. Ecology 54:375-383. Lindeman, P. V. 1997. Contributions toward improvement of model fit in nonlinear regression modeling of turtle growth. Herpetologica 53:179-191. Lindeman, P. V. 1998. Of deadwood and map turtles ( Graptemys ): an analysis of species status for five species in three river drainages using replicated spottingscope counts of basking turtles. Linnaeus Fund research report. Chelonian Conservation and Biology 3:137-141. Lindeman, P. V. 1999. Growth curves for Graptemys, with a comparison to other Emydid turtles. The American Midland Naturalist 142:141-151. Lindeman, P. V. 2005. Aspects of the life history of the Texas map turtle ( Graptemys versa ). The American Midland Naturalist 153:378-388. Lindeman, P. V. and M. J. Sharkey. 2001. Comparative analyses of functional relationships in the evolu tion of trophic morphology in the map turtles (Emydidae: Graptemys). Herpetologica 57:313-318. 143

PAGE 144

Lovich, J. E., C. H. Ernst, R. T. Zappalorti, and D. W. Herman. 1998. Geographic variation in growth and sexual size dimorphism of bog turtles ( Clemmys muhlenbergii ). The American Midland Naturalist 139:69-78. Lydeard, C., and R. L. Mayden. 1995. A dive rse and endangered aquatic ecosystem of the southeast United States. Conservation Biology 9:800-805. Mahmoud, I. Y., and J. Klicka. 1979. Feeding, drinking, and ex cretion. Pages 229-243 in M. Harless and H. Morlock, editors. Tu rtles: perspectives and research. John Wil ey & Sons, Inc., New York, New York, USA. Manabe, S., and R. J. Stouffer. 1993. Cent ury-scale effects of increased atmospheric CO 2 on the ocean-atmosphere system. Nature 364:215-218. McCarty, J. P. 2001. Ecological consequences of recent climate change. Conservation Biology 15:320-331. Meylan, P. A., C. A. Stevens, M. E. Barnwell, and E. D. Dohm. 1992. Observations on the turtle community of Rainbow Run, Marion Co., Flor ida. Florida Scientist 55:219-228. Missouri Department of Conservation (MDC). 2005. Water Quality. Conservation Commission of Missouri, Jeffe rson City, Missouri, USA. http://www.conservation.state.mo.us/fish/watershed/northfrk/watqual/260wqtxt.ht m. Accessed June 2004. Missouri Department of Natural Resources (MDNR). 2005. Rules of Department of Natural Resources: Divisions 20 Clean Water Commission: Chapter 7 Water quality. Environmental Protection Agen cy (EPA), Washington, D.C., USA. http://www.epa.gov/waterscience/sta ndards/wqslibrary/mo/mo_7_wqs.pdf. Accessed March 2008. Moll, D. 1976. A review of supposed insect catching by basking Graptemys geographica Transactions of the Illinois St ate Academy of Science 69:302-303. Moll, D. L. 1977. Ecological investigations of turtles in a polluted ecosystem: the central Illinois River and adjacent flood plai n lakes. Dissertation. Illinois State University, Normal, Illinois, USA. Moll, D. 1986. The distribution, status, and le vel of exploitation of the freshwater turtle Dermatemys mawei in Belize, central. Biological Conservation 35:87-96. Moll, D. 1990. Population sizes and foraging ecology in a tropical freshwater stream turtle community. Journal of Herpetology 24:48-53. Moll, D. 2006. The Illinois River revisited: changes in composition and structure of the resident turtle community during the twentieth cent ury and beyond. 2006 Powdermill Meeting, Portal, Arizona, USA. 144

PAGE 145

Moll, D. and E. O. Moll. 2004. The Ecology, exploitation, and conservation of river turtles. Oxford University Pr ess, New York, New York, USA. Mullins, M. A., and F. J. Janzen. 2006. Phenotypic effects of thermal means and variances on smooth softshell turtle ( Apalone mutica ) embryos and hatchlings. Herpetologica 62:27-36. Nagle, R. D., C. L. Lutz, and A. L. Pyle. 2004. Overwintering in the nest by hatchling map turtles ( Graptemys geographica ). Canadian Journal of Zoology 82:12111218. National Oceanic and Atmospheric Administration (NOAA). 1969. Annual climatological summary (1969): West Plai ns, Missouri. U.S. Department of Commerce, National Climatic Data Cent er, Asheville, North Carolina, USA. National Oceanic and Atmospheric Administration (NOAA). 1970. Annual climatological summary (1970): West Plai ns, Missouri. U.S. Department of Commerce, National Climatic Data Cent er, Asheville, North Carolina, USA. National Oceanic and Atmospheric Administration (NOAA). 1971. Annual climatological summary (1971): West Plai ns, Missouri. U.S. Department of Commerce, National Climatic Data Cent er, Asheville, North Carolina, USA. National Oceanic and Atmospheric Administration (NOAA). 1972. Annual climatological summary (1972): West Plai ns, Missouri. U.S. Department of Commerce, National Climatic Data Cent er, Asheville, North Carolina, USA. National Oceanic and Atmospheric Ad ministration (NOAA). 2004a. Annual climatological summary (2004): West Plai ns, Missouri. U.S. Department of Commerce, National Climatic Data Cent er, Asheville, North Carolina, USA. National Oceanic and Atmospheric Admini stration (NOAA). 2004b. Climatography of the United States No. 20 (1971-2000): West Plains, Missouri. U.S. Department of Commerce, National Climatic Data Cent er, Asheville, North Carolina, USA. National Oceanic and Atmospheric Administration (NOAA). 2005. Annual climatological summary (2005): West Plai ns, Missouri. U.S. Department of Commerce, National Climatic Data Cent er, Asheville, North Carolina, USA. National Oceanic and Atmospheric Administration (NOAA). 2006. Annual climatological summary (2006): West Plai ns, Missouri. U.S. Department of Commerce, National Climatic Data Cent er, Asheville, North Carolina, USA. National Oceanic and Atmospheric Administration (NOAA). 2007. Annual climatological summary (2007): West Plai ns, Missouri. U.S. Department of Commerce, National Climatic Data Cent er, Asheville, North Carolina, USA. 145

PAGE 146

Nature Conservancy, The. 2003. Living on kars t in south central Mi ssouri: A reference guide for Ozark landowners. The Nature Conservancy, Van Buren, Missouri, USA. Nickerson, M. A. and C. E. Mays. 1973. The hellbenders: North American giant salamanders. Milwaukee Public Museum publications in biology and geology no. 1, Milwaukee, Wisconsin, USA. Nickerson, M. A., A. L. Pitt, and M. D. Prysby. 2007. The effects of flooding on hellbender salamander, Cryptobranchus alleganiensis Daudin, 1803, populations. Salamandra 43:111-117. Olszewski, T. D. 2004. A unified mathem atical framework for the measurement of richness and evenness within and am ong multiple communities. Oikos 104:377387. Parchevsky, K. V. 2000. Usi ng regularizing algorithms for the reconstruction of growth rate from the experimental da ta. Ecological Modelling 133:107-115. Parmenter, R. R. 1980. Effects of food avai lability and water temperature on the feeding ecology of pond sliders ( Chrysemys s. scripta ). Copeia 1980:503-514. Parmesan, C. 2006. Ecological and evolutiona ry responses to recent climate change. The Annual Review of Ecology, Evolution, and Systematics 37:637-669. Pianka, E. R. 1976. Competition and niche theory. Pages 114-141 in R. M. May, editor. Theoretical ecology: princi ples and applications. W. B. Saunders Company, Philadelphia, Pennsylvania, US A and Toronto, Ontario, Canada. Pimm, S. L. 1991. The balance of nature?: ecological issues in the conservation of species and communities. The University of Chicago Press, Chicago, Illinois, USA and London, UK. Pitt, A. L. 2005. Reassessment of the tur tle community in the North Fork of White River, Ozark County, Missouri. M.Sc. Thesis University of Florida, Gainesville, Florida, USA. Plummer, M. V. 1979. Collecting and marking. Pages 45-72 in M. Harless and H. Morlock, editors. Turtles: perspectives and research. John Wiley & Sons, Inc., New York, New York, USA. Pluto, T. G. and E. D. Bellis. 1986. Habitat utilization by the turtle, Graptemys geographica along a river. Journal of Herpetology 20:22-31. Pluto, T. G. and E. D. Bellis. 1988. Seas onal and annual movements of riverine map turtles, Graptemys geographica Journal of Herpetology 22:152-158. 146

PAGE 147

Pusey, B. J., and A. H. Arthington. 2003. Importance of the riparian zone to the conservation and management of freshwater fish: a review. Marine and Freshwater Research 54:1-16. Rhen, T., and J. W. Lang. 1995. Phenot ypic plasticity for growth in the common snapping turtle: eff ects of incubation temperature, clutch, and their interaction. The American Naturalist 146:726-747. Riccardi, A., and J. B. Rasmussen. 1999. Extin ction rates of North American freshwater fauna. Conservation Biology 13:1220-1222. Roche, B. 2002. COSEWIC assessment and st atus report on the northern map turtle Graptemys geographica in Canada. Committee on the Status of Endangered Wildlife in Canada, Ottawa, Ontario, Canada. Roosenburg, W. M., and K. C. Kelley. 1996. The effect of egg size and incubation temperature on growth in the turtle, Malaclemys terrapin Journal of Herpetology 30:198-204. Root, T. L., and S. H. Schneider. 1993. Can large-scale climatic models be linked with multiscale ecological studies? Conservation Biology 7:256-270. Roughgarden, J., and J. Diamond. 1986. Overview: the role of species interactions in community ecology. Pages 333-343 in J. Diamond and T. J. Case, editors. Community ecology. Harper and Row, New York, New York, USA. Saumure, R. A. and P. J. Livingston. 1994. Graptemys geographica (common map turtle). Parasites. He rpetological Review 25:121. Schlaepfer, M. A., C. Hoover, and C. K. Dodd, Jr. 2005. Challenges in evaluating the impact of the trade in amphibians and reptiles on wild populations. Bioscience 55:256-264. Schlesinger, M. E., and X. Jiang. 1991. Revised projection of future greenhouse warming. Nature 350:219-221. Schuett, G. W., and R. E. Gatten, Jr. 1980. Thermal preference in snapping turtles ( Chelydra serpentina ). Copeia 1980:149-152. Schumm, S. A. 2005. River variability and complexity. Cambridge University Press, Cambridge, UK and New York, New York, USA. Seber, G. A. F. 1982. The Estimation of animal abundance and related parameters. Macmillan Publishing Co., Inc., New York, New York, USA. Shively, S. H., and J. F. Jackson. 1985. F actors limiting the upstream distribution of the Sabine map turtle. American Midland Naturalist 114:292-303. 147

PAGE 148

Smith, N. J. H. 1979. Aquatic turtles of Amazo nia: An endangered resource. Biological Conservation 16:165-176. Smith, V. H., G. D. Tilman, and J. C. Nekol a. 1999. Eutrophication: impacts of excess nutrient inputs on freshwater, marine, and terrestrial ecosystems. Environmental Pollution 10:179-196. Spotila, J. R., R. E. Foley, J. P. Schubauer, R. D. Semlitsch, K. M. Crawford, E. A. Standora, and J. W. Gibbons. 1984. Opport unistic behavioral thermoregulation of turtles, Pseudemys scripta in response to microclimatology of a nuclear reactor cooling reservoir. Herpetologica 40:299-308. Stearns, S. C., and J. C. Koella. 1986. The evolution of phenotypic plasticity in lifehistory traits: predic tions of reaction norms for age a nd size at maturity. Evolution 40:893-913. Sutherland, J. P. 1974. Multiple stable points in natural communities. The American Naturalist 108:859-873. Thornhill, G. M. 1982. Comparative reproduction of the turtle, Chrysemys scripta elegans, in heated and natura l lakes. Journal of Herpetology 16(4):347-353. Trauth, S. E., J. D. Wilhide, and A. Ho lt. 1998. Population structure and movement patterns of alligator snapping turtles ( Macroclemys temminckii ) in northeastern Arkansas. Chelonian Conservation and Biology 3:64-70. Tucker, J. K. 2000. Body size and migration of hatchling turtles: in terand intraspecific comparisons. Journal of Herpetology 34:541-546. Tucker, J. K., N. L. Filoramo, and F. J. Janzen. 1999. Size-biased mortality due to predation in a nesti ng freshwater turtle, Trachemys scripta. The American Midland Naturalist 141:198-203. U. S. Fish and Wildlife Service (USFWS). 200 6. International affairs: U.S. AppendixIII turtles. USFWS, Washington, D.C., USA. http://www.fws.gov/international/Appendix_I II/usturtles.html. Accessed March 2007. van Dijk, P. P., B. L. Stuart, and A. G. J. Rhodin. 2000. Asian turtle trade: Proceedings of a workshop on conservation and trade of fr eshwater turtles and tortoises in Asia. Chelonian research monographs no. 2. Ne w York Turtle and Tortoise Society, New York, New York, USA. Via, S., and R. Lande. 1985. Genotype-enviro nment interaction and the evolution of phenotypic plasticity. Evolution 39:505-522. Vogt, R. C. 1979. Cleaning/feed ing symbiosis between grackles ( Quiscalus: Icteridae) and map turtles (Graptemys: Emydidae). Auk 96:609. 148

PAGE 149

Vogt, R. C. 1980a. Natural history of the map turtles Graptemys pseudogeographica and G. ouachitensis in Wisconsin. Tulane Stud ies in Zoology and Botany 22:1748. Vogt, R. C. 1980b. New methods for trap ping aquatic turtles. Copeia 1980:3. Vogt, R. C. 1981a. Food partitioning in th ree sympatric species of map turtle, genus Graptemys (Testudinata, Emydidae). The Amer ican Midland Naturalist 105:102-111. Vogt, R. C. 1981b. Turtle egg ( Graptemys : Emydidae) infestation by fly larvae. Copeia 1981:457-459. Vogt, R. C. and J. V. Benitez. 1997. Sp ecies abundance and biomass distributions in freshwater turtles. Pages 210-218 in Proceedings: conservation, restoration, and management of tortoises and turtles an international conference. New York Turtle and Tortoise Societ y, New York, New York, USA. Vogt, R. C., and J. J. Bull. 1984. Ecology of hatchling sex ratio in map turtles. Ecology 65:582-587. Walley, H. D. 2000. Geographic distribution: Graptemys geographica (common map turtle). USA. Wisconsin. Herpetological Review 31:184. Webb, R. G. 1961. Observations on th e life histories of turtles (Genus Pseudemys and Graptemys) in Lake Texoma, Oklahoma. The American Midland Naturalist 65:193-214. White, D. and D. Moll. 1991. Clutch size and annual reproductive pote ntial of the turtle Graptemys geographica in a Missouri stream. Journal of Herpetology 25:439-494. White, D. and D. Moll. 1992. Restrict ed diet of the common map turtle Graptemys geographica in a Missouri stream. The Southwestern Naturalist 37:317-318. Wiens, J. A. 1984. On understanding a non-equilibrium world: myth and reality in community patterns and pr ocesses. Pages 439-457 in D. R. Strong, Jr., D. Simberloff, L. G. Abele, and A. B. Thistle, editors. Ecological communities: conceptual issues and the evidence. Prin ceton University Press, Princeton, New Jersey, USA. Wiens, J. A., J. F. Addicott, T. J. Case, and J. Diamond. 1986. Overview: the importance of spatial and temporal scale in ecological investigations. Pages 145153 in J. Diamond and T. J. Case, editors. Community ecology. Harper and Row, New York, New York, USA. Willette, D. A. S., J. K. Tucker, and F. J. Janzen. 2005. Linking climate and physiology at the population level for a key life-history stage of turtles. Canadian Journal of Zoology 83:845-850. 149

PAGE 150

BIOGRAPHICAL SKETCH Amber L. Pitt was born in 1975, in Berlin, Vermont, and grew up in rural northern Vermont. She earned her B.A. in zoology from th e University of Vermont in 1997. In the years between undergraduate and graduate school, Ambe r worked as an environmental educator and aquatic taxonomist and chemist. She earned her M.Sc. in interdisci plinary ecology from the University of Florida in 2005. 150