1 BEHAVIORAL ECOLOGY AND EFFECTS OF DISTURBANCE ON THE SUWANNEE COOTER ( Pseudemys concinna suwanniensis ) IN A BLACKWATER SPRING-FED RIVER By YURII V. KORNILEV A THESIS PRESENTED TO THE GRADUATE SCHOOL OF THE UNIVERSITY OF FLOR IDA IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF MASTER OF SCIENCE UNIVERSITY OF FLORIDA 2008
2 2008 Yurii V. Kornilev
3 To the Cooters
4 ACKNOWLEDGMENTS I express my deepest gratitude to my adviso r Dr. C. Kenneth Dodd, Jr. and my committee: Dr. Gerald Johnston, Dr. Harvey Lillywhite, a nd Dr. Max Nickerson for their continuous mentorship and encouragement throughout the l ong process. Jim Wood and the Santa Fe Canoe Outpost staff offered invaluable information and logistical support. I am indebted to William Barichivich and USGS-Gainesville who provided extremely valuable suggestions and logistical support. Numerous people volunteered as great fi eld assistants and were vital in amassing the data: Julia Altman, Ben Atkinson, Jason Butle r, Christopher Cattau, Rebecca Cox, Kathleen Coyne, Michael Deidrick, Caitlin Hicks, Kris tine Hoffmann, Alexander Iliev, Jen Johnston, Gerald Johnston, Robert Lara, Anthony Lau, James Nifong, Wes Oehmeg, Alexander Pries, Elizabeth Roznik, Travis Thomas, Brett Tornwa ll. Alexander Iliev and Georgi Popgeorgiev provided important help with st atistics and data analysis. Kr istine Hoffmann, Samuel Jones, David Pike, Elizabeth Roznik, and Matt Shirley donated field equipment. I thank V. Morgan Tyrone and the staff of OLeno State Park / River Rise Preserve State Park Dr. Dale Jackson and Dr. Peter Meylan gave me helpful insights about the life of the Suwannee Cooter. Megan Wetherington graciously provided me with rive r level data. Kristine Grayson shared her knowledge about iButtons. Sarah Tobing help ed me by identifying plants. My sincere appreciation goes to the SNRE / WEC office staff for helpi ng me jump through bureaucratic hoops. I am indebted to Dr. Mike Dorcas for he lping me see the way and his mentorship during the first difficult steps. Finally, I owe my family and friends for being there with me at all times. To these and all people I met while doing this study, thank you. The PADI Foundation and the Reptile and Amphibian Conservation Corps provided financial support. The study has been a pproved by IACUC protocol E885, FWC permit WX07166, and FL DEP permit 04230712.
5 TABLE OF CONTENTS page ACKNOWLEDGMENTS ............................................................................................................... 4TABLE OF CONTENTS ............................................................................................................. ....5LIST OF TABLES ...........................................................................................................................7LIST OF FIGURES .........................................................................................................................8ABSTRACT ...................................................................................................................... .............10CHAPTER 1 GENERAL INTRODUCTION .............................................................................................. 11Project Objectives ...................................................................................................................11Review of Risk-Disturbance Hypothesis ................................................................................12Reptile Body Temperature and Thermoregulation ................................................................. 13Human Disturbance to Basking Turtles ..................................................................................14The Ecology of Suwannee Cooters .........................................................................................16Description of the Habi tat and Study Site ..............................................................................182 USE OF DATALOGGERS (IBUTTONS) TO OBTAIN THERMAL PROFILES OF FREE-RANGING TURTLES ................................................................................................ 28Introduction .................................................................................................................. ...........28Materials and Methods ...........................................................................................................29Results .....................................................................................................................................33Discussion .................................................................................................................... ...........383 EFFECTS OF LIMITED BOATIN G ON BASKING BEHAVIOR ...................................... 55Introduction .................................................................................................................. ...........55Materials and Methods ...........................................................................................................56Results .....................................................................................................................................58Discussion .................................................................................................................... ...........604 HOME RANGE, MOVEMENT, SPATIAL DISTRIBUTION, AND RELATIVE ABUNDANCE ..................................................................................................................... ..69Introduction .................................................................................................................. ...........69Materials and Methods ...........................................................................................................70Study Site .........................................................................................................................70Capture and Marking ....................................................................................................... 71
6 Observations .................................................................................................................. ..72Visual Observations .........................................................................................................73Paint Mark Retention ....................................................................................................... 74Spatial Distribution and Abundance ................................................................................ 74Habitat Assessment ......................................................................................................... 75Results .....................................................................................................................................76Visual Observations .........................................................................................................76Paint Mark Retention ....................................................................................................... 77Home Range and Long-Term Movements ...................................................................... 77Relative Abundance .........................................................................................................80Distribution across the Habitat ........................................................................................ 80Habitat Assessment ......................................................................................................... 81Discussion .................................................................................................................... ...........82Relative Abundance .........................................................................................................82Paint Mark Retention ....................................................................................................... 83Effects of Handling ..........................................................................................................83Home Range, Distribution across the Habitat, an d Long-Term Movement ....................84Conclusions .....................................................................................................................885 SUMMARY AND CONSERVAT ION RECOMMENDATIONS ........................................ 99Summary ....................................................................................................................... ..........99Conservation Recommendations .......................................................................................... 101APPENDIX A MORPHOMETRIC PROCEDURES ...................................................................................104B TUKEY MULTIPLE COMPARISON OF MEANS ............................................................ 105LIST OF REFERENCES .............................................................................................................107BIOGRAPHICAL SKETCH .......................................................................................................117
7 LIST OF TABLES Table page 2-1 Number of times a turtle initiated fi rst aerial basking, rela tive to environmental conditions .................................................................................................................... .......543-1 Percentage of turtles obse rved basking during turtle counts in relation to number of rented boats, day, and mean daily Tair. ............................................................................... 673-2 Correlation matrix between macrohabita t variables and total turtle counts. ..................... 684-1 Physical differences between the different sections used for multivariate regression analysis ...................................................................................................................... .........924-2 Home ranges based on mark/recapture and visual observations ....................................... 934-3 Mean distance between initial capture a nd final observation, based on mark/recapture and visual observations ......................................................................................................944-4 Lincoln-Petersen estimate of relative abundance of adult and subadult turtles.. ............... 954-5 Distribution across study sections and relative abundance of different life stages. ........... 964-6 Correlation coefficients between macr ohabitat variables a nd turtle counts ...................... 98B-1 Tukey multiple comparisons of mean s 95% confidence level test comparing individual turtles mean TiButton. ....................................................................................... 105
8 LIST OF FIGURES Figure page 1-1 Typical P. c. suwanniensis ma le and female courting in a spring ..................................... 221-2 Satellite image of the study site and the surrounding area ................................................. 231-3 Satellite image of the study site at the Santa Fe River ....................................................... 241-4 Santa Fe Rivers daily recorded water leve l in relation to sea level at the two closest stations to the study site, 2001 ................................................................................. 251-5 Representative study si te mid-river depths ........................................................................ 261-6 Santa Fe River mean daily Tair and Twater ...........................................................................272-1 Straight-line carapace length and mass at initial capture of 89 P. c. suwanniensis ...........422-2 iButton dimensions and attachment to zip ties .................................................................. 432-3 Individual marking scheme, paint numb er position, and data logger attachment ............. 442-4 Comparison of TiButton and Tb of two typical adult Suwannee Cooters.............................. 452-5 Representative TiButton and environmental temperature data .............................................. 462-6 Comparison of daily averag e, minimum, and maximum TiButton of three turtles ...............472-7 Median TiButton for all data, days, and nights ......................................................................482-8 Tukey multiple comparisons of mean s 95% confidence level test comparing individual turtles mean TiButton ..........................................................................................502-9 Histograms of daily/nightly TiButton for 10 free-ranging P. c. suwanniensis .....................513-1 Daily pattern of distribution of boat a nd human observations at Santa Fe River .............. 643-2 Number of boat rentals per day correlated with mean daily Tair.. ......................................653-3 Comparison between TiButton, boat traffic, and environmental temperatures ..................... 664-1 Distribution of original capt ures and subsequent recaptures ............................................. 894-2 Paint mark longevity ...................................................................................................... ....904-3 Distribution of observations of P. c. suwanniensis originally captured in the field site .... 91
9 LIST OF ABBREVIATIONS CI Confidence interval CL Straight carapace length df Degrees of freedom OSP OLeno State Park RRPSP River Rise Pr eserve State Park SD Standard deviation SFCO Santa Fe Canoe Outpost Tair Air temperature Tb Body temperature TiButton Temperature recorded by iButton Twater Water temperature
10 Abstract of Thesis Presen ted to the Graduate School of the University of Florida in Partial Fulfillment of the Requirements for the Degree of Master of Science BEHAVIORAL ECOLOGY AND EFFECTS OF DISTURBANCE ON THE SUWANNEE COOTER ( Pseudemys concinna suwanniensis ) IN A BLACKWATER SPRING-FED RIVER By Yurii V. Kornilev August 2008 Chair: C. Kenneth Dodd, Jr. Major: Interdisciplinary Ecology Riverine turtles in blackwater streams are ge nerally understudied, but their populations are declining due to numerous anth ropogenic factors. In 2007, I studied a population of the Suwannee Cooter (Pseudemys concinna suwanniensis), a turtle endemic to Florida, in the springfed tannic Santa Fe River. I found an abundant and healthy population of 32 individuals / ha, including all size classes. I obtained thermal profiles of 10 free-rang ing individuals using miniature dataloggers. I failed to detect differences in thermal preference based on sex, but gathered support for the hypothesis that even in thermally stable and warm aquatic environments, aerial basking has thermoregulatory significan ce for the Cooters. Pronounced individual variation was evident in moveme nt patterns: some individuals covered long distances quickly (more than 5 km in 7 days) while others main tained limited home ranges of less than 200 m. Limited paddle boating had no detected effect s on the population. Continued and long-term research and monitoring will provide more needed biological information to be used in better conservation management. Not excee ding the existing levels of rive r use and banning the take of freshwater turtles should provi de protection to the population.
11 CHAPTER 1 GENERAL INTRODUCTION River-inhabiting turtles are generally understudied (revi ewed by Moll & Moll 2004), but their often large population sizes suggest that they are important for the proper functioning of riverine ecosystems (Iverson 1982; D. Jackson & Walker 1997). Currently, turtle populations worldwide face numerous challenges to survival caused by humans (Rhodin 1999). Some examples include human-made physical changes of river flow for management purposes (Bodie 2001), unsustainable harvest for food, medicinal purposes, or the pet industry (Gibbons et al. 2000), and indirect effects of human recreat ion (Sajwaj & Lang 2000; Moore & Seigel 2006). These threats apply to a great ex tent for the species-rich blackw ater rivers, which are widely distributed throughout the southeastern United States. Project Objectives In this study, I tested the a pplicability of miniature temper ature dataloggers (iButtons) for obtaining ecological data and quantifying respon ses to human disturbance of free-ranging Suwannee Cooters (Pseudemys concinna suwanniensis ) during the primary activity season. I provide an estimate of relative abundance and li fe stage distribution, and examine macrohabitat use and movement patterns in an understudied ha bitat. I hypothesized th at turtles react to increased human disturbance by decreasing baski ng times, which would be reflected in thermal profiles recorded using iButton data loggers Such information will facilitate a broader understanding of the risk-disturbance hypothesis (Frid & Dill 2002) in a conservation-related context for a long-lived aquatic reptile, and provide for better planning and management of riverine turtles.
12 Review of Risk-Disturbance Hypothesis The risk-disturbance hypothesis states that anti-predator behavior has costs to other activities and that disturbance stimuli are treated in a similar manner as pr edation risk (reviewed in Frid & Dill 2002). Both costs and risks include an optimization of trade-offs that follow economic principles and try to minimize perceive d risk of mortality at the cost of missed opportunity and loss of resources. Frid & Dill (2 002) argued two important points: 1) even though predator-specific anti-pre dator behaviors exist and so me human disturbances are evolutionarily too recent to ha ve generated such specific beha viors, generalized threatening stimuli are also present (Dill 1974 a, b); 2) habituation to disturbance is only partial because animals tend to overestimate rather than undere stimate risk (i.e., if risk is underestimated, mortality may occur), and might even show in creased anti-predator behavior. Studies of responses to disturbance of various taxa have detected changes in feeding habits, habitat use, and time allocated when caring for young (waterfowl: Gill et al. 1996; Rodge rs & Schwikert 2002; marine mammals: Constantine et al. 2004; King & Heinen 2004; la rge terrestrial game species: Stockwell et al. 1991). More importantly, such changes in behavior can lead to reduced population sizes (Gill et al. 1996; Beale & Monaghan 2004). In contrast, a strong case has been made recen tly that behavioral responses to disturbance are not indicative of the consequenc es of disturbance: healthier i ndividuals with more options for retreat sites might incur lower costs for distur bance avoidance and therefore exhibit stronger responses (Gill et al. 2001; Beale & Monaghan 2004) Thus, behavioral responses such as fleeing distance per se might be misleading and their interpre tation should be treated with caution, especially if not combined w ith studies on population biology.
13 Reptile Body Temperature and Thermoregulation Only a few cases have been recorded in whic h reptiles can substant ially increase their body temperature (Tb) physiologically over the surrounding environmental temperature. Adult Leatherback Sea Turtles ( Dermochelys coriacea ) benefit from their large surface to volume ratio and store metabolic heat (Frair et al. 1972). Brooding female Burmese ( Python molurus bivittatus ) and Indian Rock Pythons ( P. m. molurus ) utilize muscle shiveri ng to heat their egg clutches (Hutchison et al. 1966). Apart from such exceptions, many reptiles ar e capable of using fl uctuating microhabitat conditions to maintain optimum Tb, different from environmental temperature (Cowles & Bogert 1944; Dorcas & Peterson 1997). Still, because of the high costs of maintaining elevated Tb, many small and/or aquatic reptiles are thermoconformers that select mi crohabitats without regards for temperature (e.g., Brisbane River Turtle, Emydura signata, Manning & Grigg 1997; Spiny Softshell, Apalone spinifera Plummer et al. 2005). Since ectotherm ecology is greatly influenced by Tb, some animals might be selecting for optimal temperatures (Gatte n 1974; but see Huey 1982). Tb directly affects activity, metabolism, and growth (Huey 1982; Cossins & Bo wler 1987; Congdon 1989). Therefore, Tb is highly correlated with many behavioral, physiological, reproductive, and developmental variables, and directly influences such fitness variables as susceptibility to predat ion, reproductive effort, energy balance, survivorship, hab itat utilization, and pa tterns of distributi on (Peterson et al. 1993; Zimmermann et al. 1994). Othe r benefits of increased Tb are rapid development of oviductal eggs, more successful el imination of external parasites, the promotion of vitamin D synthesis by the skin, and accelerated rates of digestion (Kenyon 1925; Cagle 1950; Pritchard & Greenhood 1968; Gatten 1974; Harlow et al. 1976; Vogt 1979; Parmenter 1981; Whittow & Balazs 1982; Hammond et al. 1988; Avery et al. 1993; Swimmer 1997).
14 Turtles regulate their Tb both by means of physiological and behavioral responses to achieve preferred levels (Hutchison 1979). Behavi oral responses include two forms of basking amongst emydid turtles ( sensu Moll & Legler 1971): aquatic ba sking (floating just below the surface of the water) and atmospheric basking (si tting on an object out of water, exposed to the sun). Under favorable conditions, atmo spheric basking leads to higher Tb than aquatic basking (Moll & Legler 1971; Auth 1975; Standora 1982). At least for Painted Turtles ( Chrysemys picta ) and likely other aquatic turtles, thermoregulati on seems to be the primary factor influencing basking behavior (Crawford et al. 1983). In addition to thermoregulation, basking provide s other benefits to turtles. Basking might eliminate epizoic algae (Neill & Allen 1954), help decrease leech loads (Ryan & Lambert 2005), as well as allow the turtle to rest (Cagle 1950). However, basking also has certain costs to the organism. For example, aerial basking decreases time available for foraging, while increasing certain metabolic demands and the risk of encountering certain types of predators (Huey & Slatkin 1976). Human Disturbance to Basking Turtles Humans and their activities have been demonstrated to be a major cause for herpetofaunal declines, largely through habitat loss and degrad ation (Gibbons et al. 2000). Additionally, direct mortality on roads caused by vehicular traffic, and the potential for im pacts to the population away from the road, have been well documente d (Dodd et al. 2004; Steen & Gibbs 2004; Aresco 2005; Steen et al. 2006). Sometimes ar tificial structures such as ra ilroad ties and rails lead to individual mortality and fragment habitats (Kornilev et al. 2006). Diamondback Terrapin mortality ( Malaclemys terrapin ) has been linked to drowning in unchecked crab traps (Bishop 1983; Roosenburg et al. 1997; Dorc as et al. 2007). Increases in human recreation rapidly and
15 negatively impacted two p rotected Wood Turtle (Glyptemys insculpta ) populations, presumably as people collected them for pets (Garber & Burger 1995). More recently, indirect negative effects of hum an activity on animal behavior have been recognized. As human population in Florida multiplies at an al arming rate (the 18.5 million of 2007 are predicted to double by 2050) animals, including turtles, will have to cope with increased human disturbance caused by nature ph otographers, recreational hikers, boaters, and others. Although all turtles bask occasionally, the emydi d turtles of North America (particularly the genera Chrysemys Graptemys, Pseudemys, and Trachemys ) are especially conspicuous, primarily due to lengthy basking at visible locations, often along rivers where human recreational activities are frequent. Apart from direct im pacts such as malicious shooting at turtles (plinking; Ernst et al. 1994) and consumption by humans for food (Moll & Moll 2004), insufficient information is available concerning in direct impacts of human recreational activities on riverine turtles. Basking turtles readily pl unge back into the water when approached by humans (Boyer 1965), and the risk-disturbance hypothesis suggests th at chelonians should perceive humans as direct threats. A recent study supports the hypothesis that human recreational activities, such as boating, ne gatively influence the time spent basking by Yellow-blotched Map Turtles ( Graptemys flavimaculata ; Moore & Seigel 2006). Sajwaj & Lang (2000) warned that fishing and boating caused fr equent disturbances to ba sking Blandings Turtles ( Emydoidea blandingii) Moore & Siegel (2006) suggested that th e low frequency of multiple clutches in G. flavimaculata is due to decreased Tb. Furthermore, the increase in human-oriented river use might lead to clearing of driftwood and logs (termed snag removal), t hus decreasing potential basking sites and increasing predation risk (Grayson & Dorcas 2004). A significant correlation
16 between the availability of woody debris and turtle densities has been demonstrated in multiple studies, with rapid population dec lines following debris remova l for river improvement (Dodd 1990, Fuselier & Edds 1994, Vandewalle & Chri stiansen 1996, Reese & Welsh 1998, Lindeman 1999, Bodie 2001). The Ecology of Suwannee Cooters The River Cooter (Emydidae: Pseudemys concinna ) is a widespread riverine turtle in the USA. Endemic to the Panhandle and northern Florida, the subspecies P. c. suwanniensis (the Suwannee Cooter; for taxonomic ju stification and review see D. Jackson 2006 and citations therein) is among the largest emydids in North Am erica with straight carapace lengths of up to 44 cm and masses greater than 10 kg (Pritchard 1980; D. Jackson & Walker 1997). The Suwannee Cooter is a fast and strong swimmer, inhabiting rivers and associated lakes and impoundments. The Cooter is almost a stri ct herbivore (Lagueux et al. 1995) and does not seem to be attracted to bait such as cut fish and chicken livers typically used for capture of carnivorous species. However, it is easily observed in large nu mbers since it tends to spend extensive time basking on logs. The major contributions to our knowledge of th is subspecies can be found in the following articles and citations therein: a recent exte nsive species account (D. Jackson 2006), population size, demographic structure, and growth (Mar chand 1942; C. Jackson 1970; Meylan et al. 1992, Huestis & Meylan 2004), reprodu ction (C. Jackson & M. Jack son 1968; D. Jackson 1988; D. Jackson & Walker 1997), courtship (C. Jackso n & Davis 1972), and feeding (Lagueux et al. 1995; Bjorndal et al. 1997). Although not as extreme as turtles in the genus Graptemys (Map Turtles) P. c. suwanniensis exhibits pronounced sexual dimorphism, with mature females ca. 40% longer and ca. 200% heavier than males (D. Jackso n 2006). Sexually mature males are recognized
17 by their elongated foreclaws, longer precloacal ta il lengths, and more streamlined shells. Both sexes have narrow-lined markings on the head an d neck region, almost completely black legs and carapaces, melanistic anterior plastral seam s, and individually uni que dark blotches and spotting on the yellowish underside of the marginal scutes and the plastron (Fig. 1-1). Even though the range of P. c. suwanniensis is limited to northern Florida, it can be locally abundant. The Suwannee Cooter can be an important part of the trophic web and can attain large population sizes and densities of around 390 kg per km of river or 48 kg per ha surface water of mature female turtles, and up to 434 kg of annual reproductive biomass production per female (D. Jackson & Walker 1997). Historically, the Suwannee Cooter, colloqui ally known as Suwannee chicken, has undergone intensive unsustainable harvest for human consumption at some localities (Carr 1983). As of 2008, legal take of tur tles is only restricted by a bag limit of two turtles per day per person and a closed harvest season during nestin g, from April 15 to July 31 (D. Jackson 2006). Although less common than in the past, isol ated events of heavy consumption of Pseudemys sp. are still a threat to populations. In 2004, remains of at l east 170 turtles we re found in northcentral Florida, with 95% belonging to P. c. suwanniensis ; presumably, only one family was responsible for this mass turtle consumption (G Heinrich, pers. comm.). Typical for turtles, P. c. suwanniensis is a long lived species (e stimates of more than 25 years), with unpredictable juvenile mortality compared to high adult surviv orship, late maturity (12 years for females), iteroparity, and small annual reproductive effort (D. Jackson & Walker 1997; D. Jackson 2003). It has been established for some turtles that even minimal take, especially of adult females, can lead to rapid population dec lines (Dodd 1988; Congdon et al. 1994; Tucker & Moll 1997).
18 Because of its restricted range and past expl oitation, the Suwannee Coot er is a Species of Special Concern (Wood 1996) and its conservation status is S3 (rare to uncommon, Buhlmann & Gibbons 1997). Furthermore, P. c. suwanniensis has been identified as a Species of Greatest Conservation Need in Florida (FFWCC 2005). However, currently there is inadequate attention to regulation and enforcement, which hinders the conservation of th is amiable turtle. Description of the Habitat and Study Site The field study was conducted between 15 Ma y 2007 and 13 October 2007 at the lower Santa Fe River along the border of Alachua and Columbia counties, north-central Florida, USA (ca. 29N, 82W). I recorded 51 field days specifically for this study, for approximately 400 hours of observations. Originating in the Santa Fe and Alto lakes and their associated swamps, the Santa Fe River disappears undergro und for 4.6 km at OLeno State Park (OSP). The river emerges as a 1st order magnitude spring 3 km upstream from the US Highway 441 (US 441) bridge, at the Rise in River Rise Pres erve State Park (RRPSP). The study focused on the 3.4 km reach downstream from the US 441 bridge, but included information gathered upstream to the Rise (Fig. 1-2, 1-3). Throughout the study site, the ri verfront is comprised of a mosaic of land owned by RRPSP and private individuals. The OSP and th e connected RRPSP and McCall Park are the largest public protected areas along the Santa Fe River; other protected areas include the nearby Santa Fe River Ranch upstream from OSP, a nd Poe Springs Park, Ichetucknee and Santa Fe Springs Conservation Area, Fort White Mitiga tion Park Wildlife and Environmental Area and Ichetucknee Springs State Pa rk downstream from RRPSP. RRPSP encompasses 736 hectares and, with OSP, includes 18 distinct natural communities as defined by the Florida Natural Areas Inventor y (FNAI). An excellent re source with exhaustive information on the history and physical characteris tics of the park and th e river is the current
19 OSP and RRPSP Unit Management Plan from 2003; the next plan will not be ready before 2010 (V. Tyrone, pers. comm.). During normal conditions, the Santa Fe River in the study area is a blackwater stream with almost zero visibility due to high tannin concentrations. Although the river is fed by many springs before it drains into the Suwannee River, the entire Suwannee drainage is strongly affected by seasonal precipitation pattern s. Thus, droughts during 2006 throughout south Georgia and north Florida resulted in an abnormally low river levels and greatly decreased tannin loads. Throughout the field site the water clarity varied between murky to crystal clear (visibility of 1.5 10 m) depending on proximity to springs, bottom substrate, surface vegetation, and algal blooms. These non-normal conditions allowed unde rwater observations and snorkeling as a capture technique (introduced as goggling by Marchand 1945 a). Water depth varied throughout the study site, but did not fluctuate substantially through time (Fig. 1-4, 1-5). The current was slow and co nsistent throughout the study site, except in a few narrow and shallow channels and riffles 450 m downstream from the SFCO. The river width was ca. 40 m throughout most of the river, but varied locally from 25 to 85 m. Slight elevation gradients and occasional river flooding sometimes resulted in one bank of the river being low with a gentle slope, while the opposite shore would rise abruptly for 5 m, sometimes after a narrow 0.5 2.0 m deep beach. Because of such physiographic differences within the vicinity of the study site, the Santa Fe River flow s through four Florida Natural Area Inventory (FNAI) communities: predominantly floodplain forest and floodplain swamp surrounded by upland mixed forest, and a patc h of bottomland forest. Bald-cypress (Taxodium distichum ) often grew to the water level, and several species of pines ( Pinus spp.) and oaks
20 ( Quercus spp. ) formed parts of the canopy. Saw palmetto ( Serenoa repens ) was often found in the understory. Within the river, the dominant floating vegetation consisted of water pennywort ( Hydrocotyle sp. ), small duckweed (Lemna valdiviana), water spangles (Salvinia minima), and the introduced water hyacinth ( Eichhornia crassipes). Subsurface vegetati on included eel grass (Vallisneria americana), strap-leaf sagittaria (Sagittaria kurziana), and the non-native hydrilla ( Hydrilla vericillata ) and parrot feather ( Myriophyllum aquaticum ). Turtles used this vegetation for food and refuge. Taking of any wildlife, including turtles, from the OSP/RRPSP is prohibited. Although a few visitors to the park commented on having eaten turtle meat in the past, I did not observe any instances of humans extracti ng turtles from the river. The turtle community in the study site app eared to be overwhelmingly dominated by the basking emydid P. c. suwanniensis, with the Yellowbelly Slider ( Trachemys s. scripta) also being common. Alligator Snapping ( Macrochelys temminckii ), Loggerhead Musk ( Sternotherus minor ), and Florida Softshell ( Apalone ferox) turtles were also detected, but are much more difficult to sample due to their behavior. Othe r native chelonian species included the Striped Mud Turtle ( Kinosternon baurii), Common Snapping Turtle ( Chelydra serpentina ), Florida Redbelly ( P. nelsoni ), and Florida Cooter (P. floridana ), but these were uncommon to rare. Only two non-native Red-eared Sliders (T. s. elegans ) were observed during the field season, but several potential T. s. scripta x T. s. elegans hybrids suggest the non-na tive species might be encroaching in northern Florida (e.g., Aresco & D. Jackson 2006). The extreme drought and thick aquatic vegetati on resulted in the river being impassable except for light paddle boats (canoes and kayaks) just downstream from the public ramp next to
21 US 441 bridge. Immediately after the end of th e study site (about 3.5 km downstream from the US 441 bridge and 1.5 km upstream from the US 27 bridge), the river subsequently flowed through a short but shallow (~20 cm deep) section and then completely dried for 50 m, requiring boats to be portaged (Fig. 1-5). This co ndition resulted in lower than usual numbers of predominantly paddle boats upstream from the US 27 bridge. The land between the Sink at OSP and River Rise is a natural barrier that minimized turtle dispersal; downstream just past the study site, the dry river bot tom and high water temperatures (Twater) reaching 35C might have deterred the strictly aquatic turtles from dispersing downstream, at least during warmer days. The river Twater at the springs is almost a constant 22C year-round (Hornsby & Ceryak 1998). However, the re st of the river is th ermally heterogeneous. Downstream from the Rise, Twater fluctuated both on a daily and monthly basis (Fig. 1-6) and often increased further downstream by as much as 8C between the beginning and the end of the field site. Twater decreased with increasing depth, increa sed substantially on the surface of vegetation mats, and decreased close to springs. Heavy rains in mid-October 2007 caused an overnight shift in rive r conditions, returning them to a state close to normal. The river leve l increased temporarily, and the increased tannic load resulted in practically no underwater visibility, leading to a termination of the study.
22 Figure 1-1. Typical P. c. suwanniensis male (top) and female courti ng in a spring in the Santa Fe River. Note the sexual size dimorphism and ad aptations for aquatic existence, such as webbed feet and streamlined hydrodynamic carapace.
23 Figure 1-2. Satellite image of th e study site and the surrounding area. The Santa Fe River disappears unde rground at River Sink at OSP (OLeno State Park; blue boundary), a nd reemerges at River Rise at RRPSP (River Rise Preserve State Park; red boundaries). The Columbia Co. (north) Alachua Co. (south) border generally follows the river channel. US Highway 441 passes through the town of High Springs, s ituated south of the Santa Fe River.
24 Figure 1-3. Satellite image of th e study site at the Santa Fe River. S01-S16 denote beginning points of the 16 sections of the study site. Shallow and Riffles denote locations in the river that impeded boat traffic. Pub lic ramp and SF Canoe Outpost are the two publicly accessible boating ramps. H 1 and H2 (next to S06 and S09, respectively) are the locations of the tw o environmental temperatures dataloggers.
25 Figure 1-4. Santa Fe Rivers daily recorded water level in relation to sea level at the two closest stations to the study site, 2001. A) OLeno Station. B) US Highway 441 Station. For comparison, when depths are ca 9 m at the US Highway 441 Station, the river is almost dry at the public boat ra mp. Data provided by Suwanee River Water Management District and USGS. A B
26 Figure 1-5. Representative study site mid-river depths at 50 m intervals (closed circles). Measurements were taken on 6 August 2007, but remained similar throughout the study. Study site sections were downstream from US 441 bridge (refer to Fig. 1-3). The empty circle denotes an unknow n depth of more than 6.5 m.
27 Figure 1-6. Santa Fe River mean daily Tair and Twater ( 1 SD). A) Tair. B) Twater. Data obtained from two automated dataloggers at the field site Missing data are from days of heavy rain that precluded access to the dataloggers. A B
28 CHAPTER 2 USE OF DATALOGGERS (IBUTTONS) TO OBTAIN THERMAL PROFILES OF FREERANGING TURTLES Introduction Most poikilothermic animals often use be havioral methods to regulate their body temperature (Tb; Cowles & Bogert 1944). Tb has profound impacts on reptilian physiology and behavior (Cossins & Bowler 1987). However, evidence suggests that reptiles do not always try to maintain high Tb in order to optimize physiological re quirements (Shine & Madsen 1996). Traditionally, collecting physiological data on free-ranging animals has been a challenge. Temperature sensitive radio transmitters have pr ovided insights into the thermal ecology of reptiles, but these techniques ar e often logistically time-consuming and expensive (Beaupre & Beaupre 1994; Zimmermann et al. 1994; Manni ng & Grigg 1997; Dorcas & Peterson 1998; Blouin-Demers & Weatherhead 2001; Whitake r & Shine 2002; Pearson et al. 2003). The use of automated temperature datalogging in wildlife and eco logical studies is increasing in popularity because of decreases in cost, increases in data storage capacity and precision, miniaturization, and unobtru sive attachment. Dataloggers have been used to record temperatures of free-ranging turtles, their ne sts, and hibernacula (M ueller & Rakestraw 1995; Sajwaj & Lang 2000; DallAntonia et al. 2001; Nussear et al. 2002; Plummer et al. 2005). Some of the latest dataloggers provide even further miniat urization and opportunities for studies of free-ranging individuals. iButtons ar e tiny chips enclosed in an aluminum casing, combining extremely low energy requirements, hi gh durability, and simple interface for data transfer. They have wide applications for industr ial uses such as access co ntrol to buildings and computers, asset management, and data logging tasks. As temperature dataloggers, they are widely used in healthcare, food safety, and asphalt curing ( www.iButton.com ). Lately, however,
29 they have been employed in ecological rese arch (Angilletta & Kr ochmal 2003; Robert & Thompson 2003; Grayson & Dor cas 2004; Hester et al. 2008). In this study, I used iButtons to determin e the thermal profile of a large free-ranging riverine emydid turtle (Suwannee Cooter, Pseudemys concinna suwanniensis) endemic to northern Florida and often abundant in spring-fe d and tannic blackwater streams. My objective was to assess the feasibility of using dataloggers to accurately estimate details related to basking, as well as provide ecological data on the behavior of a turtle in a expected ly relatively constant temperature water body. Materials and Methods I conducted the study in the typically blackwa ter Santa Fe River downstream from the River Rise in north-central Florida between 15 May and 13 Oc tober 2007. Although most of the field observations were carried out downstream from US Highway 441 bridge (US 441), turtles infrequently were surveyed and observed ups tream as well. Droughts in 2006 resulted in low water conditions and unusual clar ity. The river water temperature (Twater) at the springs is an almost constant 22C year-round. However, downstr eam from the Rise, the river was thermally heterogeneous. Twater fluctuated on a daily and monthly basis (Fig. 1-6) and the subsurface temperature increased further downstream by as mu ch as 8C between the beginning and the end of field site. Twater decreased with increasing depth, incr eased substantially on the surface of vegetation mats, and decreased close to springs. Basking sites were abundant. I measured more than 130 basking sites of various sizes throughout the 3.4 km downstream from US 441. Tu rtles were observed basking in large groups in vegetation mats that obtained Twater higher than in the surrounding water. Therefore, turtles were not limited in their opport unity to thermoregulate throughout most days of observation.
30 Forty-eight adult (>200 mm straight car apace length [CL]) and two subadult P. c. suwanniensis (> 120 mm CL) were captured by snorke ling from 18 to 23 May 2007 (Fig. 2-1). Captures occurred in the one km stretch of ri ver ca. 800 m downstream from the US 441 bridge. CL was measured to the nearest mm, weight to the nearest gram, and sex was determined. Turtles were uniquely marked by drilling holes in their marginal scutes and by applying oilbased paint on the dorso-lateral side of the carap ace. Additional details are provided in Chapter 4 and Fig. 2-3. I equipped each turtle ex situ with an automated temperat ure datalogger (iButton, model DS1922L, height 6.4 mm x radius 8.68 mm, we ight 3.12 g; Maxim Integrated Products, Sunnyvale, CA). This high capacity iButton reco rds 8192 temperature readings ranging between -40C and +85C, with an accuracy of .5C. The iButtons are progr ammable through freely available software (One Wire Viewer Demo; Dallas Semiconductor, Dallas, TX). I set the iButtons to collect data at a 10-min interval for a total of just over 56 days, starting on 19 May and ending 15 July 2007. I programmed each iB utton to start recording at least 24 h after capture and attachment in order to allow the turtles to resume their normal behaviors. Temperature measurements were asynchronous betw een turtles, and assignment of iButtons to turtles was random. I modified the method for iButton attachme nt from Grayson & Dorcas (2004). In the laboratory, each iButton was progr ammed and affixed to the middle of a 6 mm wide black plastic UV-resistant zip tie with a thin wi re. I applied three layers of a bl ack plastic coating (tool dip) for water-proofing and to mimic the black shell color of P. c. suwanniensis ; each layer dried for approximately 30 min (Fig. 2-2).
31 I attached iButtons by drilling two 7 mm holes in the marginal scutes above the right hind leg. To minimize stress to the animals, the holes also were used for individual marking of turtles (Fig. 2-3). The zip tie was passed through the tw o holes, and the extra zip tie was removed. To minimize handling time, no additional coating was a pplied in the field (in contrast to Grayson & Dorcas 2004), since the coating us ed requires at least 30 min to dry and is toxic in its liquid form. Turtles were measured, marked, and releas ed close to the point of capture within 1 h. Except for two individuals (#1133 and #1149) I opportunistically captured on 16 June 2007 in order to redo the paint numbers, I recaptured turtles after the iBu ttons stopped collecting data. Between the end of July and mid-October, I recaptured turtles predominantly in a section of the river 2.5 km downstream from US 441 because I had observed the most turtles there. Field work was terminated in mid-October, when ra ins increased tannin con centrations and made snorkeling impossible. The mean weight of the wate r-proofed iButtons was 5.51 g (n = 10); the mean weight of turtles equipped with an iButton, excluding th e two lightest turtles (600 g and 820 g), was 3513.5 g (range: 1315; SD: 1814.4; n = 48). Turtles fitted with TiButton were representative of the general size-distribution of all captured turtles (Fig. 2-1) and could be separated into four distin ct classes. Four large females (#1109, #1140, #1145, and #1151) had CL ca. 340 mm and weighed ca. 5.5 kg. Three females (#1133, #1137, #1152) had CL ca. 235 mm and mass ca. 2.0 kg; a male (#1245) with CL of 240 mm had a mass of 1.6 kg. A large male (#1122) had CL of 280 mm and mass of 2.6 kg. The subadult (#926) measured 165 mm and weighed 0.6 kg. Two HOBO automated dataloggers (HOBO H8-002-02 Temp/E xternal, Onset Computer Corporation, Bourne, MA) were us ed to collect air temperature (Tair) and Twater. HOBOs were
32 concealed on the shore next to the river at locations 1150 m and 1760 m downstream from the US 441 bridge (Fig. 1-3). I set them to record shaded Tair and Twater every 5 min between 20 May and 3 August 2007; rain precluded ob taining the data for 1 June 2007. To protect the HOBO dataloggers from rain and animals, I covered them with a layer of clear wrap, and taped them to the inside of cut-in-the-middle 1.5 L plastic bottles. I positioned them on trees at a height of 1 m under shade and ma de numerous cuts in the plastic bottles in a way that excluded water but provided ample ventilation. Twater was measured next to the shore with an external sensor (TMC20-HD, 6 m cable), placed at 0.5 m depth. A 3.3 m PVC pipe, with a diameter that snugly fit the opening of the bottle, served to protect the cable. I made opportunistic visual obser vations of turtles between Ma y and August. For identified individuals, I recorded time, whether the turtle was in the water or basking and, if applicable, when it terminated basking. Visual observations were made mostly in the 3.5 km river stretch downstream from US 441, but surveys were also conducted in the 3 km upstream to River Rise. I combined iButton and environmental temperat ures and created graphs of high resolution for each turtle so that I could discriminate be tween two consecutive data points. For each known time of visual observation, I attempted to describe the activity of the tur tle within the habitat (basking, underwater, or diving) based on the graphed data. Prior to field work, I conducted an experiment ex situ to examine the correlation between Tb and iButton temperatures (TiButton). I used the same metho dology of iButton set up and attachment as for the field observations. I attach ed an iButton to a male turtle weighing 1500 g and a female weighing 3800 g. I plac ed the turtles in separate cat tle tank containers filled with flowing water of constant 22C, with heat lamps overhead. Every 10 min, I removed each turtle for ca. 30 sec and approximated Tb by inserting a quick-read digi tal thermometer 6 cm into
33 the cloaca. Initially, I left the turtle s in the water for 45 min until their Tb equalized with Twater. I simulated basking by placing the turtles in dry bins underneath the heating lamp for 1.5 h. Afterwards, I returned the turtles to the cattle tanks with flowing water for 50 min, and then again put them in the bins and heated them fo r 1.5 h. I observed no signs of discomfort caused by overheating. Statistical tests were perf ormed using R (v. 2.6.2, RDCT 2008). I performed a Tukey multiple comparisons of means 95% confidence le vel test to compare individual turtles overall mean TiButton. I created boxplots using R, comparing individuals using the complete dataset, as well as separating them by time of day (daytime: 08:000:00 h; nighttime: 20:00:00 h). If the notches of the boxplots ove rlapped, I considered this str ong evidence that the medians did not differ (Chambers et al. 1983). Since data were skewed to the left, I consider medians more informative than means. Unless otherwise specified, = 0.05 for all statistical test s and standard deviation is presented as SD. Results During the ex situ experiment, the two turtles Tb and TiButton followed closely Twater after an initial period of acclim ation. For both turtles, TiButton increased almost instantaneously by 20C when I initiated the simulated basking. Tb increased gradually a nd in 1.5 h had only risen by ca. 10C. Although different in size, turtles heated at a similar rate. Likely due to differences in media conductivity, both turtles cooled faster than they heated, with a rapid drop of TiButton and Tb immediately after being placed in colder water. The female, which was more than 2.5 times heavier than the male, cooled down sl ightly slower. Even coated, the TiButton corresponded to changes in the environmental temperatures resu lting from moving the tu rtles between air and water in order to record Tb, but did not always follow Tb closely (Fig. 2-4).
34 In situ I recaptured 12 individuals (3 8 1 subadult). I removed only 11 iButtons, because female #1149 was not recaptured after th e iButtons had recorded to capacity. The iButton on a male (#1134) failed due to unknown reasons and no data could be retrieved. The iButton on female #1151 stopped recording at ~ 95% of its capacity. Upon downloading, three other iButtons (#1122, #1137, #1245) ha d internal clocks that di ffered from real time by 10 hours. According to Maxim technical support staff, intrusion of water insi de the casing can lead to internal clock oscillation changes and produc e offsets from real time without impacting temperature readings. I visually inspected all the iButton data and compared it with environmental data, but could not detect any abnormalities; therefore, errors with the clock most likely occurred after all of the data were record ed. I obtained 564 complete days of iButton data with more than 81,000 temperature readings. Since the datalogger was less than 1% of the we ight of the smallest turtle and less than 0.5% of the mean weight of adult turtles, its effects on turtles behavior were likely limited, if any. Furthermore, I did not detect changes in tu rtle behavior caused by iButtons during multiple sightings in the field, and I observed turtle s courting, feeding, swimming, and basking. Although turtles were originally captured clos e to the HOBO dataloggers, some turtles moved long distances up and down stream. For example, female #1145 was recaptured 2.3 km upstream from her original capture location (Chapter 4) Furthermore, often in the afternoons a Twater was 2C warmer at the end of the field site than at the beginning. Therefore, Twater can only be used as an approximation to the e nvironmental temperatures that some turtles experienced. A representative sample of four days (17 June 2007), comparing TiButton of the subadult turtle (#926) and environmental temperatures from the two HOBO dataloggers, demonstrates
35 some trends obtained for all turtles (Fig. 2-5). Turtles often initiated ae rial basking around 8:00 9:00 h in the morning. Aerial basking was st rongly dependent on the difference between Tair and Twater, and was often initiated al most immediately after Tair exceeded Twater; basking occurred rarely when Tair < Twater (days 1 in Fig. 2-5; Table 2-1). Si nce thermal microhabitat differences existed in the river that were not accounted for by the placement of the two environmental dataloggers, it is possible that tu rtles initiated basking even more precisely without me being able to detect it. Turtles demonstrated marked individual differences in basking ha bits. A representative sample of daily mean, minimum, and maximum TiButton obtained for three random turtles demonstrates that even during the same day, and therefore similar e nvironmental conditions, turtles experienced different temperatures. No tu rtle showed consistently the highest or the lowest maximum daily TiButton, and all had similar mean daily TiButton (Fig. 2-6). Furthermore, I estimated time of initiation of first basking even t for the day (Table 2-1). Three of the turtles (#926, #1137, #1140) initiated basking immediately or soon after Tair exceeded Twater in more than 80% of the days. Two large females (#1145 and #1152) and the two males initiated basking later during the day or not at all in more than 70% of the time, whereas two other females (#1133, #1151) showed a greater variation for time of first basking. Mo st turtles except #1122 and #1145 basked every day at least briefly during favorable condi tions. Basking during unfavorable conditions of Tair lower than Twater was rare. Aerial basking was usually terminated when Tair fell below Twater. Nocturnal TiButton usually closely followed Twater, congruent with nighttime observations of turtles underwater. Nocturnal increases of TiButton over Twater or Tair were observed only for one female (#1109) between 20:00-
36 03:00 h on the last 20 days of recorded data. Ther e was no indication that she or another female attempted to lay eggs at night, although tu rtles were observed nesting during the day. The observation that tu rtles aerially basked exclusively when Tair > Twater was further supported by data from days of low mean Tair, when turtles usually did not spend as much time basking (e.g. day 4, Fig. 2-5). On 12 June and 14 July, when Tair was lower than Twater during the day, none of the recaptured tu rtles attempted to bask. Furthermore, I recorded 75 behavioral observations (30 basking, two diving, 43 swimming) for the 10 turtles I retrieved with iBu ttons. I successfully iden tified turtle behaviors in 61 of those cases (81%). Differentiation of aerial basking on logs from aquatic basking (especially if at the surface of vegetation mats) was not possible. A comparison between the TiButton medians of individual turt les failed to show any relationship between sex or size a nd thermal preferences for all th ree categories (combined data, daytime, nighttime; Fig. 2-7). A comparison of mean s resulted in a grouping of turtles dissimilar in size and sex. These statistically significan t groups were different than in any grouping obtained when comparing medians (Fig. 2-8). Du ring the nighttime, there were only three groups with different median TiButton. Two of the largest females ha d a higher (26.2C) and a lower (25.1C) median than the rest (25.7C). Similar nighttime TiButton can be explained by greater thermal homogeneity across the river after sunset. Furthermore, Twater was higher than Tair during most of the time from 20:00 h to 08:00 h, and turt les had to be in the water in order to maintain higher Tb. Turtle #1151 had the lowest median TiButton, and she was the only turtle that would aerially bask early in the morning (7:00:00 h) when Twater > Tair. Turtle #1140 had the highest median TiButton, and was observed multiple times further downstream where Twater was higher during part of the night than the Twater likely experienced by the other turtles. Although #1109
37 had a median TiButton not significantly different than most ot her turtles, she had a higher nocturnal TiButton than most other turtle s. The data showed TiButton initially 4C greater than Twater around 20:00 h, at which time recorded Tair was lower than Twater. TiButton slowly decreased until it equilibrated with Twater around 2:00 h. This pattern was obs erved during the last 20 days of recording. Comparing daytime TiButton revealed four statistically di fferent medians. The subadult and one large female (#1152) had the highest median (27.2C). These results were corroborated by visual observations of the iButton data showing numerous basking attempts for these two turtles (Table 2-1). A large female (#1109) a nd a male (#1122) had a low median TiButton (26.2C). The lowest median (25.7C) was of a female (#1145) that was captured less than 200 m from a cold spring. Visual observation of her data and numerous decreases of TiButton from recorded Twater to a temperature of 22C suggest she spent time in the spring each day. The rest of the turtles had a median equal to 26.7C. As expected, daytime TiButton showed a greater range than nighttime TiButton. Based on the boxplots for all data, one group cont ained six turtles that had a median of 26.2C, and a second group of four turtles had a median of 25.7C. Trends among individual turtles closely resembled each other as well as th e trends observed when turtles were pooled by sex (Fig. 2-9). Only four females experienced temperatures lower than 22C for enough time to be recorded by the iButtons, even during the night; two of the turtles contributed most of the 42 low recordings out of a total of 56,844. The two males only experienced three instances of TiButton = 21C out of 16,384 recordings. Maximum recorded TiButton did not exceed 46C. Only 1.2% and 1.5% of TiButton were higher than 34C for males a nd females, respectively. TiButton ranged between 30C and 34C in
38 4.4% of the occasions for males and 5.1% for females. The majority of TiButton, even during the day were less than 30C for both sexes. Twater increased from 23C to 26C during the two months of data collec tion, and partially accounted for the high TiButton at night. Discussion Externally positioned temperature loggers such as iButtons revealed important information about patterns in behavior of a species of large aquatic turtles. The best results were achieved when iButton data were combined with data on environmental conditions. For highly vagile species such as P. c. suwanniensis especially in a thermally heterogeneous environment, saturating the study site with environmental dataloggers is important. Field observations on movement and knowledge of the location of the study organisms was useful in accurately determining behaviors based on TiButton data. The results strongly suggest that aerial ba sking serves a thermoregulatory function in P. c. suwanniensis even in habitats such as the Santa Fe River that provide favorable thermal conditions year-round. Pseudemys c. suwanniensis is well-known as a species that spends long time aerially basking (D. Jackson 2006), and it has been assumed that such behavior is linked to thermoregulation, based on a corr elation between sunny weather and a high number of basking turtles (Auth 1975; Crawford et al. 1983). Ho wever, Manning & Grigg (1997) provided strong arguments suggesting that often small aquatic reptiles are ther moconformers, and the importance of aerial basking is overemphasized and not prim arily of thermoregulatory significance. For the Santa Fe River population of P. c. suwanniensis basking often began immediately after Tair exceeded Twater and terminated when Tair was lower than Twater. This result suggests that for this species aerial basking serves a thermoregulat ory function. Similar observations apply for a northern population of P. concinna (Buhlmann & Vaughan 1991). Basking, changes in microhabitat and posture coupled with thermore gulation have been noted for an amphibian
39 species as well (Rana catesbeiana Lillywhite 1970). Spiny lizards ( Sceloporus ) from different geographic locations had similar preferred Tb (Bogert 1959); it is likely that P. concinna in a warm environment such as the Santa Fe River would have a preferred Tb similar to more northern populations. The observati on of Giovanetto (1992) that Tair > 21C and sunny weather correlate with high levels of basking is likely rela ted to the fact that most spring-fed rivers in Florida have a Twater of 20C. Circumstantial evidence that high Tb is beneficial for P. concinna is the correlation between habitation in warm year-round ri verine habitats and largest body size, attained by the subspecies ( P. c. suwanniensis ). Grayson & Dorcas (2004) demonstr ated small differences between Tb and TiButton (usually less than 1.3C) for adult Chrysemys picta However, my results suggest that inferring Tb from TiButton for individuals weighing more than 2 kg, es pecially during aerial basking, would be difficult. TiButton, however, generally closely followed Tb while turtles were underwater. Grayson & Dorcas (2004) at tached iButtons on adult C. picta (mass ~320 g, CL ~130 mm) and did not detect changes in be havior or deterioration of physic al condition after one year. I did not observe turtles to lose their iButtons, and obs erved an individual with an iButton attached for more than 10 months. iButton failures and abno rmal internal time were most likely due to incomplete insulation from water. Even though I a pplied three layers of coating to make sure I completely covered the iButton and the wire holding it to the zip tie, the zip ties were not bent while the coating dried. Thus, minor openings might have occurred at the points where the zip tie came out of the coating when I bent the zip ties to attach them to the turtle. Several conditions precluded me from identifyi ng separate basking events. Discriminating the activity of individuals proved di fficult when the difference between Tair and Twater was small, such as during days of low mean Tair, or late in the afternoon. Th erefore, the percentage of
40 accurately estimated behaviors based on TiButton could be lower than I obtained because most of the observations I made of turtle behaviors were during mid-day basking, or during days with high Tair. A major problem with classifying behavior s was a lack of proxi mity between turtles and dataloggers collecting environmental temperature. Lacking turtle-specific Twater precluded definite separation of decreases in TiButton caused by turtles diving in the water and decreases resulting from a lowered amount of solar radiation. Contrary to my predictions, TiButton fluctuated substantially instead of reaching a plateau during daily aerial basking. Furthermore, during rapid decreases, TiButton often did not reach the recorded Twater (e.g. day 1, Fig. 2-5) making it difficult to stipulate whether the changes in TiButton data reflected termination of basking, or were caused by temporary shading (Boyer 1965). In other cases, basking events could be identified clearly. Therefore, discrimination between individual basking events in some cases was impossible, making comparisons both between and among individua ls speculative. Individual turtles exhibited a range of thermoregulatory behaviors. Turtles had the capacity to thermoregulate by moving long dist ances upstream towards cooler water and downstream towards warmer water, as well as by selecting thermoclines and altering the duration of aerial basking. For example, for female #1109 nocturnal TiButton exceeded Tair and Twater. Most likely, she spent the night in a location in the river that maintained a higher Twater. She was in fact recaptured at a shallow (< 0.4 m) location on th e river overgrown with submerged vegetation where turtles were observed at high densities du ring the daytime. At this location, the water likely accumulated heat during the day and cooled off slower than the rest of the river due to the lack of current and the increased heat capacity of the dense vegetation mass. Differences were observed in the length of basking, time during the day when turtles initiated basking, days during which different tu rtles basked, and visits to thermally varying
41 microhabitats. Although it is difficult to equate TiButton to Tb, the similarity between TiButton medians is suggestive of a keen ability of tu rtles to budget their time and activities in the environment in such a way as to obtain some preferred Tb. From this small sample, no definite trends can be observed linking behavior with either sex or si ze. Although the sample size is small, the results are suggestive that turtle size or sex might not be the most important predictors of thermal preference. The importance of health as a driver for thermoregulation and basking has been established for other turtle species ( Terrapene carolina and C. picta Monagas & Gatten 1983; Sternotherus depressus Dodd 1988). Regardless of the precise biological reasons for aerial baskin g, it is clear that populations of Suwannee Cooters require ample basking site s in order to maintain stable and healthy populations. Future studies utilizing iButtons or other dataloggers could benefit if a computational method is devised to approximate closely Tb from TiButton. Subsequent researchers should collect detailed information about indi viduals health condition such as amount of external and internal parasites, algal growth, and presence of eggs in females in order to better explain differences in basking behaviors.
42 Figure 2-1. Straight-line carapace length (S CL) and mass at initial capture of 89 P. c. suwanniensis. SD equals one standard deviation.
43 A B Figure 2-2. iButton dimensions and attachment to zip ties. A) iButton attached to the zip tie before the application of the plastic coati ng. B) Three coated iButtons compared to a penny from various angles (clockwise: top, sideways, bottom).
44 Figure 2-3. Individual marking scheme, pain t number position, and temperature logger attachment. Turtles were ma rked individually using a code system based on drilling marginal scutes with a 7 mm bit. The numbers around the turtle carapace show the marginal scutes that are drilled and th eir numeric values. Shown is turtle #1146, marked by drilling scutes #2, #4, #40, and #400 & #700 (also used to attach an iButton temperature logger). This method is modified from Cagle (1939) and allows 1554 unique numbers. Within my study, all ad ult turtles were marked starting from 1100s; two subadults were marked in the 900s. Carapace Length = 365 mm
45 Figure 2-4. Comparison of TiButton and Tb of two typical adult Suwannee Cooters. Turtles were placed in separate containers with water of constant temperature (22C) and then heated in dry bins with a lamp twice.
46 Figure 2-5. Representative TiButton from one turtle and environmental temper ature data from four days in June 2007.
47 Figure 2-6. Comparison of daily (8:00:00 h) TiButton of three randomly selected turtles. A) Average. B) Minimum. C) Maximum A B C
48 Temperature (C) 24 h Days (8:00:00 h) Nights (20:00:00 h) Figure 2-7. Median TiButton for 24 h, days, and nights. A) A ll data. B) Close up provided for comparison of statistical difference between medians. Circles represent outlier data points. A
49 Temperature (C) 24 h Days (8:00:00 h) Nights (20:00:00 h) Figure 2-7. Continued B
50 Figure 2-8. Tukey multiple comparisons of m eans 95% confidence level test comparing individual turtles mean TiButton. Statistically not different individuals are underlined. Raw results provided in appendix (Table B-1).
51 Figure 2-9. Histograms of daily/nightly TiButton for 10 free-ranging P. c. suwanniensis
52 Figure 2-9. Continued
53 Figure 2-9. Continued
54 Table 2-1. Number of times a turtle initiated first aerial basking, rela tive to environmental conditions. Tair >= Twater denotes first basking for the day began within 1 h of one of the two recorded Tair surpassing Twater. Tair >> Twater denotes that a basking occurred during the day, but was initiated more than 1 h after recorded Tair surpassed Twater. None denotes that no defined aerial bask ing attempt was detected; on 12 June Tair was lower than Twater and no turtles basked. Tair < Twater denotes basking when Tair was lower than Twater. Some behaviors were not inferre d for 2 June due to lack of environmental data to compare TiButton data to. Sex Turtle Tai r >= Twate r Tai r >> Twate r None Tai r < Twate r Subadult 926 45 71 0 1122 10 2418 2 1245 16 324 0 1109 1133 33 137 0 1137 41 112 0 1140 39 104 0 1145 13 2512 1 1151 ** 22 102 6 1152 17 332 0 Twater experienced by the turtle was likely sufficiently higher than recorded Twater and precluded confident estimation. Likely stayed in a shallow ve getation mat, which was warmer even after recorded Tair > Twater. ** Last 10 days excluded; pattern not similar to other patterns observed. Events of Tair < Twater basking occurred in early morning (7:00:00 h).
55 CHAPTER 3 EFFECTS OF LIMITED BOAT ING ON BASKING BEHAVIOR Introduction The increasing number of humans that ha ve access to outdoor re creation has led to increasing encounters between wildlife and humans. While more people have unique opportunities to experience nature, the environment is put under pr essure to withstand greater amounts of disturbance. Recently, conservation biologists have expanded our knowledge about the indirect effects of human-wildlife interactio ns. Even low-impact actions such as bird watching or hiking might increas e stress hormone levels and w eaken immune systems, and prolonged and repeated di sturbance can lead to poor feed ing, fewer offspring, and lower body mass for numerous animals (sheep: Stockwell et al. 1991; lizards: Hecn ar & MCloskey 1998; wading birds: Rodgers & Schwikert 2002; manat ees: King & Heinen 2004; dolphins: Bejder et al. 2006a, b; but see Johnson et al. 1996 for marine tu rtles). According to the risk-disturbance hypothesis (reviewed by Frid & Dill 2002), animal s will often overestimate risk and likely engage in anti-predator behaviors during disturbance events, thus incurring energetic costs and limiting the time available for other activities. Turtle wariness while aerially basking had been noted long ago for populations that experienced little interactions with humans (Newman 1906), and th e termination of basking due to boating was mentioned in Boyer (1965). Negativ e effects of boating on turtles have been quantified for Yellow-blotched Map Turtles ( Graptemys flavimaculata Moore & Seigel 2006), but reptiles, and especially riverine turtles, ar e still understudied. Turtles, as ectotherms, often need to bask to achieve and mainta in their preferred body temperature (Tb). Aquatic turtles from the genus Pseudemys are considered dependent on aerial basking (G. Jackson 2006), but basking potentially exposes them to disturbance from hikers, fishermen, or recreational boaters.
56 In order to investigat e the relationship between boating and turtle basking, I initially hypothesized that boating traffic might negatively affect the amount of basking by the large emydid turtle P. concinna suwanniensis. I compared basking counts with boat traffic volume, and attempted to use external temp erature loggers to examine turtle behavior and assess potential impacts on Tb. Materials and Methods I conducted this study in the summer of 2007 in a section of the lower Santa Fe River in north-central Florida. The study concentrated on a 3.4 km section of the river downstream from the US Highway 441 (US 441) bridge; intermittent observations were collected up to 3 km upstream of the bridge. Boats on the study site came from one of four sources. The greatest proportion were rental and shuttled boats launched from the Santa Fe Ca noe Outpost (SFCO), as well as boats from the public boat ramp 220 m downstream from the SFCO I met a few riverfront homeowners and a negligibly small number of peopl e that launched at the US 27 or US 47 bridges and paddled upstream. The only other outfitter located in High Springs has been putting out boats downstream from the study site at the public ramp at the US 27 br idge (Fig. 1-2). Boating traffic through the site was somewhat limited by four porti ons of the river with depths of less than 30 cm water depth. Going downstream, the first 200 m shallow was immediately past the public boat ramp, and was overgrown with vegetation for 100 m. A short riffle of 50 m was present 200 m further downstream. Immediately after the end of the field site, there was a second riffle of 50 m. At 500 m downstream, the river even co mpletely dried for 50 m by the end of the summer (Fig. 1-3). This greatly limited boating tr affic upstream from the US 27 bridge into the field site. The few riverfront ow ners within the study site and ot her private boat owners seemed
57 disinterested to venture downst ream. Instead, they mostly paddled the 3 km upstream towards River Rise, only briefly passing through my study area. To estimate boat traffic, I used data excl usively consisting of canoe / kayak rental information from 1 May to 31 August 2007, kindly pr ovided by SFCO. The data consisted of the number of canoes / kayaks / doubl e kayaks rented or shuttled, the number of people, and the destination and pick up time if av ailable. I excluded boats rent ed for the full-moon night trips organized by SFCO, as well as ones that never passed through my site because they had been shuttled downstream from the fi eld site (to the US 27 or US 47 bridge) and were picked up even further downstream. I supplemented the rental info rmation with opportunistic personal records of boats observed passing through the field site between 27 May and 6 July 2007. Most of the boats that were launched at the public boat ramp and went upstream and out of the study site were not detected. However, these boats could also have an impact on turt les that dispersed out of the study area. No trails exist along th e river bank, so I rarely encountered hikers or fishermen on the bank. I carried out 21 turtle count surveys (13 downstream and 8 upstream) on 13 days between 5 June and 7 July 2007. Surveys were conducted fr om a kayak moving in the middle of the river for 2 h. Upon encountering large a ggregations of turtles, I slow ed down and used binoculars to count as many of the turtles as possible before approaching and scaring them. The starting and finishing time of surveys varied: surveys started as early as 08:30 h and sometimes finished as late as 20:00 h. I counted the num ber of turtles aerially basking as well as those seen in the water. I concealed two environmental temperature loggers on the shore (HOBO H8-002-02 Temp/External, Onset Computer Corporati on, Bourne, MA), 1.2 km and 1.8 km downstream
58 from the US 441 bridge. They collected shaded air (Tair) and water temperature (Twater) every 5 min between 20 May and 3 August 2007; rain precluded obtaining data on 1 June 2007. I indiscriminately captured 50 P.c. suwanniensis by snorkeling on 18 May 2007. Turtles were of sufficiently large size (carapace length > 150 mm) to allow attachment of automated temperature dataloggers (iButton, m odel DS1922L, height 6.4 mm x radius 8.68 mm, weight 3.12 g; Maxim Integrated Products, Sunnyvale, CA). iButtons recorded the environmental temperature immedi ately next to the turtle (TiButton), in 10 min intervals for ca. 56 days. Details about HOBOs and iButtons are presented in Chapter 2. Captures occurred in the one km stretch of river ca. 800 m downstream from the US 441 bridge, but some turtles rapidly moved at leas t 2.5 km downstream and 1 km upstream from US 441. Between the end of July and mid-October, I r ecaptured turtles predominantly in a section of the river 2.5 km downstream from US 441. More information about turtle movement can be found in Chapter 4. Unless otherwise specified, = 0.05 for all statistical test s and standard deviation is presented as SD. Results I observed 142 boats between 27 May and 6 Ju ly 2007; discrimination between which ones belong to the SFCO and which were personal boats was not always possible. Based on personal observations, boat traffic was spr ead out throughout the day and coincided with highest amount of solar radiation, and thus with optimal bask ing time (Fig. 3-1). I observed only 11 water craft (canoes, jonboats, jet ski) with an engine; except fo r the jet ski, all ot hers had 5 hp outboard motors. Based on the canoe rental data for 92 days from SFCO, 822 paddl e boats carrying 1578 people were rented from 1 May to 31 July 2007. Only on 24 days were more than 10 boats
59 rented, 6 boats were rented on 15 days, 1 boa ts were rented on 29 days, and no boats were rented on 24 days (Fig. 3-2). A total of 674 (82%) of the boats were rented Friday to Sunday. Among the days with no boats rented, only two were from Friday to Sunday; heavy rain precluded boating on at least one of those oc casions. Although boat rentals did not correlate significantly with mean daily Tair (R2 = 0.03), heavy rains preclude d any boating (Fig. 3-2). Around 20% of the boats went to the River Rise and back to the SFCO; the number of boats going downstream that were not shuttled back to the SFCO was minimal (J. Wood, pers. comm.). I conducted 21 counts of turtles, including 10 during days of higher traffic (Friday Sunday). There did not appear to be a relations hip between the total num ber of turtles observed or the percent of turtles basking in relation to the number of boats rented. A logarithmic linear regression demonstrated a lack of correlation betw een the percentage of turtles basking and the number of boats (R2 = 0.014). A correlation analysis on all variables did not show strong connection between day of the week and the percentage of turtles basking or tota l counts (Table 3-1, 3-2). Grouping week days in two categorie s (MondayThursday, FridaySunday) resulted in identical correlation coefficients to the ones obtained when comparing individual days of the week (Table 3-2). A comparison between 12 days of data from TiButton of five random turtles, environmental temperatures, and number of boat traffic st rongly suggested that basking was positively correlated with high Tair, and was not related to boat traffic intensity (Fig. 3-3). I could not rely on TiButton to accurately discriminate betw een aerial basking occurrences in order to compare counts of basking attempts dur ing high and low traffic (Chapter 2). However, for the 50 days that I had both TiButton for each of the 10 turtles and environmental data, I compared the mean daily (8:00:00h) TiButton of all turtles between days of high boat traffic (>
60 10 boats, n = 16) and days of low traffic (n = 34). No significant differenc e was detected (z-test: z = -0.545, p = 0.29), suggesting little if any impact of limited boating on the thermal profile of P. c. suwanniensis Weather conditions alone did not acc ount for the similarity between TiButton for the two groups. Tair from one of the environmental temp erature loggers served as a proxy for weather conditions; the mean daily Tair between the two groups was not significantly different (t = -0.190; df = 48, p = 0.42). Discussion Limited paddle boating seemed not to influence the basking behavior of the individuals in this population. The population in this stretch of the river seemed healthy, abundant, and reproducing successfully, suggesting limited eff ects of minor human disturbance. Visual observation of the TiButton data, although not quantif ied, suggested that turt les basking behavior was dependent on Tair more than the amount of boating traffi c. However, their behavior was also partially decoupled from environmental conditio ns and turtles basked different amounts of time during days with similar weather conditions. Several hypotheses may explain my results. First, the amount of boating was minimal, even during the days of highest traffic, and usually was concen trated during the late morning. Therefore, boats were not present on the river fo r most of the time, and I observed many turtles that resumed basking in less than 5 min after a disturban ce event. Furthermore, the Twater at the Santa Fe River was > 25C for most of the summer, and likely stays around 20C even during the winter; turtles theref ore could maintain high Tb even without extensive basking. Turtles exhibited high levels of individual va riation in disturbance responses, including partial habituation, as predicted by the risk-disturbance hypothesis. For example, several marked large females allowed me to approach them while basking to less than 1 m before fleeing, even after being captured by snorkeling several tim es. On the other hand, on several occasions
61 multiple turtles simultaneously terminated basking due to unknown reasons more than 100 m from the observer. Individual variation of basking habits was detected using TiButton data in this study, and has been observed for nesting females by D. Jackson & Walker (1997). However, the strength of behavioral responses might be linked to an individuals health or ability to acquire resources rather than to the intensity of a disturbance (Gill et al 2001), rendering solely behavioral observations inconc lusive about the true extent of the disturbance and making quantifying disturbance in the field difficult. Pseudemys c. suwanniensis historically has undergone high levels of harvesting, indicated by its colloquial name Suwannee chicken, as well as by multiple recorded instances of illegal take (D. Jackson 2006). Carr (1952) suggested that current be havioral responses are the evolutionary result of prolonged interactions with humans a nd less wary individuals were eliminated from the population a long time ago. However, I did not always observe increased vigilance and a fleeing response. Although my results did not detect negative effects of boating on basking behaviors, several important points need to be considered. Deleterious physiological responses of stress and disturbance might exist that were not quantified in this study. Responses likely vary by species: Wood Turtles (Glyptemys insculpta ) have been shown to exhibit physiological responses such as increased tachycardia when handled (Cabanac & Bernieri 2000), but no short-term behavioral changes were detected for captured Gopher Tortoises ( Gopherus polyphemus ; Kahn et al. 2007). Disturbance might impact indi viduals differently based on their age, size, or physical condition. A study on human disturbance in a nationa l park found that gravid female massasauga rattlesnakes (Sistrurus c. catenatus) are less visible than non-gravid females or males. Although
62 harmful effects could not be detected, Parent & Weatherhead (2000) found a positive correlation between disturbance levels and snake movement. Smart et al. (2000) hypothesized that human activ ities resulting in inte rmediate levels of disturbance, coupled with decreased predation, might enhanc e lizard species richness and diversity in communal rangelands. Similarly, th e indirect impacts of boating disturbance on alligators and the removal of the large nuisance indi viduals might benefit chelonian populations by decreasing predation pressure from one of the major chelonophagus species. Visual observation of the TiButton data suggested that duri ng days of similar weather conditions, turtles sometimes basked more during days of high boating traffic than during days of low boating traffic. Cabanac & Bernieri ( 2000), although using a small sample size of 3 individuals, demonstrated that Wood Turtles behaviorally raise their Tb as a response to handling and stress. Testing whether P. c. suwanniensis reacts in a similar way to stress could suggest whether behavioral responses such as duration and frequency of basking could be used as measures of disturbance. Although it seems low levels of boating disturbance had limited impacts on the population, a threshold likely exists that turtles cannot tole rate. According to the staff of Ginnie Springs, a heavily visited location on the Santa Fe River ab out 15 km downstream from the study site, it is rare to see basking turtles on the 3 km of the river where several hundred people tube each summer day. During a visit in 2008, I noted that the basking turtles at this location had at least partially habituated and did not plunge as readily as the ones I studied extensively. Also, during a snorkeling trip to Rainbow Run during a weekend with multiple boats, I failed to see basking P. c. suwanniensis, although turtles are abundant th ere (Huestis & Meylan 2004).
63 More studies are needed to a ssess disturbance effects on downs tream sections of this and other Florida rivers where motorized boat traffic is more prevalent. Due to low human impacts and the highly thermally favorable conditions year-round, the results from this study should be applied to other areas with extreme caution. Bo ating currently has lim ited negative impacts on P. c. suwanniensis, within the section studied of the Sant a Fe River, and decreasing the amount of human visitation is not a high conservation priority. In other locations with different boating and visitation uses, the impacts on basking turtles could be more deleterious.
64 Figure 3-1. Daily pattern of distribution of boat and human observa tions at Santa Fe River, May 27July 6 2007, based on personal obse rvations. Results are based on 146 observations of single boats or groups of boats and seven observations of humans.
65 Figure 3-2. Number of boat rentals from Santa Fe Canoe Outpost and number of people per day correlated with mean daily Tair. Grey blocks under the X axis mark FridaySunday and o fficial holidays. Arrow indicates rainy days.
66 TiButton 5 4 3 2 1 0 3 18 31 25 0 6 0 0 13 30 3 Number of Boats Per Day Environmental Temperatures Figure 3-3. Comparison between TiButton, boat traffic, and environmental temperat ures for 12 days. Top 5 graphs show TiButton for 5 individuals; bottom graph demonstrates Tair (high amplitude) and Twater (low amplitude). Vertical lines denote beginning and end of days. FridaySunday are marked in bold. Note that basking occurs irrespectively of boating traffic, on days of high Tair.
67 Table 3-1. Percentage of turtles observed basking during turtle c ounts in relation to number of rented boats, day, and mean daily Tair. Date Day of week Total count Daily Tai r Percentage basking Boats count 6/5 Wed 20625.5 45.15 0 6/6 Thu 34626.1 39.60 1 6/6 Thu 17726.1 35.59 1 6/7 Fri 30526.1 53.77 3 6/10 Mon 24528.1 53.88 21 6/10 Mon 25028.1 47.60 21 6/11 Tue 30127.5 48.84 2 6/15 Sat 32723.8 49.54 18 6/15 Sat 26823.8 52.99 18 6/16 Sun 21225.6 48.58 31 6/16 Sun 25925.6 45.17 31 6/17 Mon 26826.0 42.91 25 6/17 Mon 29926.0 44.15 25 6/19 Wed 21826.3 40.83 6 6/23 Sun 20525.6 54.15 30 6/23 Sun 23225.6 40.52 30 6/24 Mon 24426.1 59.02 3 6/24 Mon 25026.1 45.60 3 6/30 Sun 25526.6 44.71 31 6/30 Sun 26326.6 39.92 31 7/7 Sun 24227.0 36.36 27
68 Table 3-2. Correlation matrix between macroha bitat variables and to tal turtle counts. Day Total Tai r Percentage baskingNumber of boats Day 1.00 0.02 -0.43 0.07 0.62 Total 0.02 1.00 -0.11 0.12 -0.11 Tai r -0.43 -0.11 1.00 -0.12 -0.03 Percentage basking 0.07 0.12 -0.12 1.00 -0.08 Number of boats 0.62 -0.11 -0.03 -0.08 1.00
69 CHAPTER 4 HOME RANGE, MOVEMENT, SPATIAL DISTRIBUTION, AND RELATIVE ABUNDANCE Introduction Riverine turtles are experien cing severe population reductions on a global scale (Gibbons et al. 2000), partly attr ibutable to habitat loss due to river improvement (Bodie 2001). Declines in riverine turtles have far-reaching effects on aquatic ecosystems, since they are major consumers and producers, attaining large comm unity biomass and high density (Tinkle 1958; Plummer 1977; Moll 1990; D. Jackson & Walker 1997) However, the ecology of riverine turtles is generally understudied (Moll & Moll 2004). In the current rapidly developing environment, basic biological data concer ning the home range (defined as that area traversed by the individual in its normal ac tivities of food gath ering, mating, and car ing for young [Burt 1943:351]), movement patterns, a nd habitat use has direct imp lications for devising better management strategies fo r species conservation. The Suwannee Cooter ( Pseudemys concinna suwanniensis ) is a riverine turtle endemic to the northern portions of Florida. The Suwannee Coot er is a Species of Special Concern due to its limited range and history of exploitation for human consumption (D. Jackson 2006). The Cooter likely has a strong impact on riverine communities since it is a major herbivore reaching large population numbers and biomass (Marchan d 1942; D. Jackson & Walker 1997; Huestis & Meylan 2004). Cooters can be highly vagile and move several kilometers up or downstream within a short period of time (Marchand 1945 a; D. Jackson & Walker 1997). Suwannee Cooters may be especially abundant in strikingly different lotic habitats: spring fed rivers with crystal clear water, stable temp erature, and stable wate r levels (e.g., Ichetucknee and Rainbow Run Rivers), and blackwater stre ams with highly limited visibility, seasonally fluctuating temperatures, and occasional floodi ng (e.g., Suwannee River and its tributary the
70 Santa Fe River; Crenshaw 1955). Although blackwater rivers are distri buted widely throughout the southeastern United States, know ledge about Florid a populations of P. c. suwanniensis is biased because researchers have worked mostly in crystalline rivers (Marchand 1945 a; Kramer 1995; Lagueux et al. 1995; D. Jackson & Walker 1997; Huestis & Meylan 2004). Severe droughts in 2006 provided the ra re opportunity to examine a typically blackwater stream population during a time with l ittle tannin input resu lting in clear water conditions. In this study, I combined mark-recaptu re and individual paint marking protocols to determine habitat use, movement patterns, and the extent of linear habita t in Suwannee Cooters. Further, I conducted counts of turtles in order to obtain insights on their relative abundance in relation to spatial distribution in a protected area experiencing some, but not severe, levels of human recreation-based disturbance. Materials and Methods Study Site The study was conducted on the Santa Fe River (Alachua and Columbia counties, northcentral Florida) in a 3.4 km stretch downst ream from the US Highway 441 bridge (US 441) between 15 May and 13 October 2007 (Fig. 2-2). On several occasions, data were collected upstream in the 3 km from the bridge to Ri ver Rise, where the river reemerges above ground from the Florida aquifer. Most of the study site and the surrounding area is managed by the River Rise Preserve State Park and is protected from agricultural and urban development. Drought during the previous years lowered the river level substa ntially, resul ting in clear water with good visibility, in contrast to the normally tannin-stained water. More information on the study site is in Chapter 1.
71 Capture and Marking I caught 50 Suwannee Cooters (23 25 2 subadults; group A) between 18 23 May 2007 in order to attach temperat ure-sensitive dataloggers (iButt ons; Chapter 2). While attempting to retrieve iButtons between 23 July and 8 Oc tober 2007, I captured an additional 39 turtles (12 20 7 subadults; group B). Group A turtles were caught in an approximately 1 km stretch of the river starting 800 m downstream from US 441. Subsequent captures and recaptures occurred primarily in the first 2.5 km immediately downstream from US 441 (Fig. 4-1). Study organisms were captured by active pursuit while snorkeling, or opportunistically by hand (Marchand 1942, 1945 a, b; Kramer 1995; Huestis & Meylan 2004). On one occasion, I successfully caught turtles by setting a net (20 m x 1.5 m, 3 cm mesh size) around a shallow mat of vegetation. I collected standard morphometric measurements in the field (straight-line carapace length [CL], plastron length, body mass, sex or life stage), uniquely marked individuals, and usually released them within an hour at the capture location (See Fig. 2-3 for explanation of marking scheme and appendix A for morphometric procedures). Toward the end of the study, turtles were not processed due to time and logi stic constraints unless th ey had been captured previously. Although snorkeling may be biased towards larger in dividuals, the captured turtles are likely representative of the adult turtle population (Fig. 2-1). I attached temperature data loggers coated in black plastic (iButtons, height ~9 mm, radius ~11 mm) on the marginal scutes of the carapace a bove the right rear leg (Fig. 2-3). I painted a unique number on each clean and dry dorso-lateral side of the carapace using a non-toxic white oil-based paint marker (563 Speedry, Diagraph, Marion, IL; Fig. 2-3). On group A turtles, the width of the number lines was ~0.5 cm, and the complete number was usually ~10 x 10 cm. Initially, some of the numbers wore off quickly or were difficult to detect. I subsequently increased the line width to 1 cm and the overa ll number size to ~15 x 15 cm, which made for
72 better detection and mark longevit y. Since the iButtons were usua lly visible from a distance, I could often visually determine if a turtle wa s from group A, even when the painted numbers wore off. Turtles were measured, marked, and rel eased close to the point of capture within 1 h. Observations I paddled the length of my field site using a kayak in order to visually observe and identify turtles. Transects were carried out 23 May 15 October usually between 11:00 h and 16:00 h to maximize turtle observations. Capture positions a nd visual observations were recorded with a hand-held GPS (Garmin eTrex, Garmin International, Olathe, KS). The error inherent in the GPS unit was less than 15 m, which is negligibly small considering the movement potential and habitat available to the highly mobile adult Cooters. During visual observations, I recorded the sex of the turtles only if positive identification was possible. Because Suwannee Cooters reach sexual maturity around CL of 190 mm for males and 275 mm for females (Huestis & Meylan 2004), I defined three life stages that I could visually discriminate with eas e: adults (CL >= 200 mm), subadults (200 mm > CL > 70 mm), juveniles / hatchlings (CL =< 70 mm). A chelonian mark-recapture study by Dr. Gerald Johnston (Santa Fe Community College) was conducted concurrently with my study in the fi rst 1.1 km of the river starting at River Rise, 3 km upstream from US 441. This adjacent study a llowed me to extend observations beyond my study site. Five monthly sampling sessions we re conducted between May and September 2007. We snorkeled for more than 120 person-hours an d hand-captured all turtles we could, then measured and individually marked them. The marking scheme for Pseudemys was identical to that used in my study, and there was no duplication in numbering (Johnston et al., in prep. ). I supplemented field sampling by occasionally ka yaking between US 441 and River Rise and visually observing turtles in order to detect dispersa l beyond my primary study site.
73 Visual Observations The relatively clear water allowed me to paddl e a kayak or a canoe and observe turtles both basking and swimming. I cautiously approached a nd examined every turtle I encountered and looked for the presence of drill holes, an iButton, or a painted number on the shell. I made an effort to avoid disturbing turtles and insp ected basking turtles using binoculars. Upon encountering a marked turtle, I recorded the GPS position, the time of day, the activity of the turtle (swimming, basking), the identity of the turtle, and the degree of certainty of correct identification (positive, ambiguous/partial, no id entification), and the condition of the paint number. Observation points were then importe d in Google Earth (v 4.2, Google Inc., Mountain View, CA), and plotted on a 2007 satellite image. All distances were calculated at mid-rive r. I obtained robust measurements by manually tracing mid-river distances betw een observations at an eye alt itude of 200 10 m. I used the minimum distance between observations to estimate movement between different days. In cases of more than one observation of the same individual on the same day, I calculated the minimum distance between observations that the turtle must have covered. For all dist ance analyses, I used only data from turtles that were positively identified; ambiguous sightings were discarded, even when some individuals could likely be rec ognized through a process of elimination and deduction. Marked individuals whose numbers c ould not be deciphered were only used to describe the loss of paint marks. I used the minimum direct distance over wa ter between the two most distant points of [observation] (Sexton 1959:137) in order to es timate the minimum linear aquatic home range and extent of long-term movements. This me thod was used previously for two species of Pseudemys ( P. nelsoni Kramer 1995; P. c. suwanniensis D. Jackson & Walker 1997). Home range was only estimated for individuals I obser ved at least 30 days after original capture.
74 Paint Mark Retention I estimated the rate of loss of paint marks on group A turtles to evaluate the usefulness of this method, since my ability to observe, correctly identify indivi dual turtles, and describe their movements was dependent on the length of paint mark retention. I did not estimate retention on any subsequent marks because I terminated intensive sampling before all the marks had disappeared. Spatial Distribution and Abundance I conducted several pilot transe ct surveys prior to beginning my study, where I observed an unequal distribution of turtles across the length of the river. I ther efore divided the study site into 16 sections of varying lengths based on the fo llowing river characteri stics that might be biologically important to turtles: major changes in perceived de pth, compass orientation (which affects the amount of sunlight e xposure during the day), and width. I conducted 21 counts of both basking and sw imming turtles on 13 days between 5 June and 7 July 2007 in order to quantify the spatia l distribution of turtle s and to obtain relative abundance estimates. Surveys were carried out by a single observer in a kayak moving down the middle of the river. Thirteen surveys were made going downstream, and eight upstream, in which identical speed was achieved. Surveys took 2 h, since I was making a variety of observations. Upon encountering large aggregations of turtles, I slowed a nd used binoculars to count as many turtles as possibl e before approaching and possibly scaring them. The start and finish time of surveys varied, with the earliest surveys begun at 08:30 h, and the latest surveys finished at 20:00 h. Most surveys were carried out between 11:00 h and 17:00 h to maximize the number of turtles basking and in crease detection by ta king advantage of the best light conditions. I recorded the number of individuals per rive r section in each life st age (adult, subadult, hatchling, only observed a head).
75 I sampled a 1.9 km section of the river by kayak in order to estimate the relative abundance of adult and subadult turtles counting marked and unmarked turtles. I then conducted a downstream and upstream count of adults and subadu lts the next day after the last turtle from group A was captured, and two days later I made two downstream and one upstream count. I estimated abundance using the Lincoln-Petersen formula provided in PopTools (v 3.03, Hood 2008). Although the capture and recapture sessions were only days apart, the condition for population closure was potentially violated because of the emigration of marked individuals outside the sampled area. Therefore, my results are likely an overest imate of population size. Habitat Assessment I collected data on several environmental vari ables (Table 4-1) to examine macrohabitat characteristics that might influence turtle distribution. I took 66 mid-river depths at ~50 m intervals, starting at US 441 and working downs tream. I measured 50 m as the distance between two boats using a laser range finder (Bushne ll Yardage Pro Sport 450, Bushnell Performance Optics, Lenexa, KS; accuracy of 1 m). I dete rmined depth by dipping two 3.3 m interlocking PVC pipes, graduated at 0.1 m (Fig. 1-5). I us ed a digital thermometer and recorded water temperature (Twater) 20 cm below the surface in the middle of the river at the beginning of each section. The surface area of each section was de termined based on 2007 satellite images using Google Earth. I visually approximated the major compass orientation for each section using the built-in compass tool in Google Earth. I qualified biophysical characteristics of 132 known and potential aerial basking objects throughout the study site between 28 June and 12 July 2007. I measured most basking objects on which I observed turtles, as well as ones that subjectively appeared like potential basking sites. Although this was not a complete count of all basking objects, I made every attempt to be inclusive. The number of available basking objec ts for turtles depends on river level in most
76 cases, but the constant river leve l and the lack of new fallen tr ees suggest that the number of basking sites did not change much throughout the study. A multivariate regression analysis was perfor med in R (v 2.6.2, RDCT 2008) to select the most important habitat predictor variables for relative abundance based on the raw counts of the various life stages of turtles. I used mean water depth, mean Twater, compass direction, section surface area, and number of basking sites counted as independent variables for each river section. Unless otherwise specified, = 0.05 for all statistical test s and standard deviation is presented as SD. Results Visual Observations From the 50 turtles from group A, six male s and two females were never positively resighted or recaptured. I recaptured 12 individuals (3 7 1 subadult; four individuals were recaptured three times, one four times). From the 39 turtles from group B, three males, eight females, and one subadult were never positively re -sighted or recaptured; I recaptured eight individuals, with one indivi dual caught twice (Fig. 4-3). Excluding original captures, I acquired 528 lo cations through visual observations and recaptures. I positively identified 69 individuals during 375 sightings; 67 of these sightings were a second or third observation for the same day. I observed an iButton, but failed to identify group A turtles 149 times; I could not identi fy group B turtles on four occasions. During 2007, 109 P. c. suwanniensis were captured a total of 168 times in the first 1.1 kilometers downstream from River Rise. None of these turtles were captured originally at my field site. However, four large adult females (t wo recaptured three times, one two times) and one adult male from the River Rise study were subsequently recaptured downstream from US 441.
77 Paint Mark Retention Rapid algal growth on the carapace, shedding of scutes, friction with water and vegetation reduced paint retention. Fine painted numbers of group A turtles lasted approximately 2 weeks, after which detection and successful identification decreas ed rapidly (Fig. 4-2). Turtles with iButtons but without distinct numbers were observed as early as six days after capture and turtles were identified as late as 52 days after marking. The use of thicker and larger numbering resulted in improved retention of the marks. On 31 March 2008, I observed two turtles with faint paint marks, but failed to identify them before they entered the tannin-stained water. Since I painted the last paint marks on 28 September 2007, mark s can be retained for at least 186 days (> six months). Home Range and Long-Term Movements I estimated the home range for 26 turtles (6 19 1 subadult) observed more than 30 days after initial capture (Table 4-2). The m ean distance between initial capture and final observation was similar for adults, but exceeded the distance for subadults ( : 484 m, : 625 m, subadult: 15 m; Table 4-3). Females showed a much wider range of distances between first capture to last observation than males, and the subadult was last found within 20 m from the initial capture (Table 4-3). Females also showed greater range and slightly larger home ranges than males ( : 200 m, : 800 m; Table 4-2). Pronounced individual varia tion caused some turtles to undertake long-distance movements and change their centers of activity, whereas others remained within a small range, even though all turtles had been handled in a similar way. Some individuals had strong site fidelity where they basked on the same log for several weeks at a time. Several turtles were consistently positively identified. Female #12 wa s observed eight times in 17 days basking on the same log. Female #14 was observed post-captur e seven times in 19 days within a 15 m range;
78 on two subsequent observations she moved 200 m ups tream and then returned to her favorite log within 17 days. Female #73 moved 450 m upstream af ter being captured, but within five days she returned to the same spot and was observed on four occasions on the same log during the next 16 days. Following some short-distance move ments, she was 6 m from where originally captured on her last observati ons 50 days after capture. Site fidelity might not be limited to basking si tes. At a location where I rarely saw turtles between 9:00 h and 17:00 h, I opportunistically caught an unmarked male with a distinctive stubbed tail on three occasions (7, 11 June, 3 Ju ly 2007). Captures occurred between 19:00 h and 20:15 h, all within 10 m of one anot her, in a shallow (< 25 cm) gravel rapid. Whether the turtle had been feeding or was in proxim ity to a nighttime refuge is unknown. In the following section, I describe the movement patterns of six turtles captured initially at River Rise and detected downstream from US 441 in order to illustrate the variety of turtle behaviors. These individuals were not included in the results presented in Tables 4-2 and 4-3. Intriguingly, four of these turtles were caught originally during the same sampling session on 7 May 2007. Since capture locations in this case were only approximately known, I calculated minimum possible distances. Although these turt les might have moved longer distances than average turtles, the general behaviors and res ponses applied to turtles from groups A and B. Turtle #82 was a female caught within the first 200 m downstream from River Rise; she was recaptured 37 days later in the same area. However, 10 days later, I observed her 1850 m downstream from US 441, over 4.7 km from wher e caught originally. She was located 58 days later only 115 m upstream from the previous observation. A female (#38) caught in 2006 close to the River Rise was found 13 months later more than 4 km downstream. During the 31 days sh e was observed in 2007, the furthest distance she
79 moved between observations was 325 m, and th e distance from original capture to last observation was 187 m. Even though in 2007 she wa s captured four times, kept for 24 h on one occasion, and observed on three additional days, she still stayed within a shallow area with rapid flow. Such limited movement might have been induced by poor health, si nce she had an overall sick appearance, heavy leech load, and an injury to the lower jaw. Another female (#74) was captured within the River Rise site and recaptured 72 days later at the Rise headspring. Seventy-two days later, she was captured after a downstream movement of at least 4.4 km. After she was released, she was re-sighted the same day 1.5 km further downstream. I recaptured a female (#81) 87 days later at least 3.4 km downstream from where she was caught originally within the River Rise site. F our days later, she was observed about 1.5 km downstream, but within four days had turned around and moved > 2.1 km upstream. She was recaptured at the River Rise headspring 35 days later, and was observed 24 days later about 1.1 km downstream. Three observations made in 2 h during one day showed a downstream displacement of 830 m. A male (#91) caught within the River Rise study area was recaptur ed at least 4.8 km downstream 97 days later. Afte r eight more days, I observe d the turtle 650 m further downstream. Two later observations during the next 17 days were made less than 100 m apart. Therefore the furthest distance between observations was more than 5.5 km and the distance from capture to last observation was at least 4.9 km. Another long and rapid dispersal I observed wa s of an unidentified male, originally marked upstream at River Rise site and ob served seven days later downstr eam in my field site. His exact original capture location is unknown, but he moved at least between 4.8 km and 5.9 km.
80 Only three observations of turtles with an iButton were made upstream from US 441, even though I made more than 15 trips upstream to River Rise throughout the study. An unidentified adult female was observed on 26 and 27 June 2007 at the same basking log, 1.2 km upstream from US 441. Based on all original capture points, she had to move upstream 2.2.0 km, 35 days after capture. An adult female (#1145) wa s caught 111 days after the original capture, 0.5 km upstream from US 441. I coul d not determine whether these observations pertained to the same or different individuals. Relative Abundance A Lincoln-Petersen estimate yielded a mean of 134 adults per ha of river, with a wide confidence interval of 274 individuals (n = 5 counts; Table 4-4). Repetitive raw counts produced an estimate of 25 adults per ha of ri ver with a narrower range (SD = 39.7; range: 118 258; n = 21 counts; Table 4-5). The estimate fo r total counts was slightly greater (32 individuals/ha; SD = 42.2; range: 177; n = 21 counts). In 10 of the surveys, I counted between 240 and 270 individuals per sampling trip. Even though the highest and the lowest counts were made during the same day, the percen tage of turtles basking was similar during both surveys. Distribution across the Habitat Overall, the greatest turtle concentrations we re in areas around basking sites that allowed numerous turtles to bask simulta neously. Several sections had very similar areas, but strikingly different abundances: for example, #8 (high abundance) and #10 (low ab undance), #3 (low) and #15 (high), #4 (low) and #13 (high; Table 4-5). Adult individuals were observed more often than subadults and hatchlings th roughout the study site and in general showed widespread distribution. Hatchlings seemed to be concentrat ed in some sections (e.g., #6 and #14), and were uncommon or absent in most others.
81 Habitat Assessment The distribution of tur tles of different life stages across the river was equally affected by the macrohabitat characteristics that I quantifie d. The correlation matrix from a multivariate regression analysis of macrohab itat characteristics showed very similar results when testing different combinations of life st ages, but total counts correlated slightly better than the rest (Table 4-6). Therefore, I will only present results based on the tota l number of turtles counted per section. In the complete model including all predic tor variables, only the nu mber of basking sites was significant (t = 2.778, p = 0.02); the models adjusted R-squared value was significant (R2= 0.662; F5, 10 = 6.889, p < 0.01). The number of basking site s and depth were the best predictors for turtle abundance according to independent stepwise, forwar d elimination, and backward elimination model selections. For the best pred ictors model, the number of basking sites was significant at = 0.001 (t = 4.637, p < 0.01), depth was significant at = 0.05 (t = 2.713, p = 0.018), and the models adjusted R-sq uared value was highly significant (R2= 0.728; F2, 13 = 21.09, p < 0.01). Although area and compass direction intuitivel y seemed good predictors, the results of turtle abundance were not supportive. Water depth was strongly correlated with Twater, but there was a non-linear trend with Twater. Therefore, water depth was select ed as the preferred predictor. However, it is important to note that in this case the positive relationship between water depth and Twater was due to the specific study site characteristics and cannot be extrapolated to other locations. A shift in habitat use was observed, both on a daily and a monthly basis, but was not quantified. Certain sections repeat edly had turtles mostly during specific times during the day e.g. turtles were often observed feeding in the afternoon in se ction 4, where they were rare during earlier hours. Large aggregations of up to 30 individuals were commonly observed in
82 shallow vegetation mats composed predominantly of water pennywort ( Hydrocotyle sp.), small duckweed ( Lemna valdiviana ), water spangles ( Salvinia minima), and subsurface non-native hydrilla ( Hydrilla vericillata ) and parrot feather ( Myriophyllum aquaticum ). Turtles were observed surface basking and feeding on such fl oating vegetation, but not on ones dominated by the introduced water hyacinth ( Eichhornia crassipes). Discussion Relative Abundance My Lincoln-Petersen estimate of 134 adults per ha was much higher than an estimate of 47 adults/ha obtained using the same method, but carried out over a 2 day period 4 km upstream (Johnston et al., in prep.). My resampling efforts were carried over a slightly longer period of time and were likely an overestimate since I have noted that P. c. suwanniensis moved long distances immediately after capture. Therefore, the closed population assumption of the model was probably violated. However, based on historic accounts and other abundance estimations for Florida populations, the Suwannee Cooter can reach very high densities. In Rainbow Run, Marchand (1942) estimated 746 individuals per km of river, and 50 years later Giovanetto (1992) calculated an abundance of 313 per km (40.4 per ha ). In a small spring run connected to the Suwannee River, C. Jackson (1970) captured the equivalent of between 700 and 1400 per km (300 per ha; see D. Jackson 2006). Based on these results, Lincoln-Petersen or similar mark/res ight estimators are probably unreliable for P. c. suwanniensis because they violate model assumptions. With comparable sampling effort, raw counts yielded results with lower variance and seemed adequate for crude abundance estimation as they provide at least a minimum assessment of abundance. Furthermore, visual counts were less biased in detecting hatchlings and subadu lts, resulting in a more useful estimate.
83 Paint Mark Retention The extended retention period of the thicker paint marks migh t be related to decreased activity and decreased scute sheddi ng during the colder winter months, when growth is likely slower (Huestis & Meylan 2004). Besides the use of thicker pain t and larger numbers, I suggest four improvements in the methodology: 1) Comp letely removing algae from the carapace and drying the shell before painting; 2) Applying copi ous amounts of paint to increase retention; 3) Diversifying and combining different paint colo rs, the position of the number on the carapace, handwriting styles, and other means allowing a high degree of identification even under suboptimal conditions; 4) Involving recr eational boaters as citizen scie ntists to increase the amount of data collected while the paint is discerna ble and to provide an educational experience. Kramer (1995) reported paint mark retention for 1 months on Florida Redbelly Turtles ( P. nelsoni ) using rubberized or epoxy paint, a nd Buhlmann & Vaughan (1991) successfully used Petersen disc tags to mark River Cooters ( P. concinna ) for more than a year. A question not addressed in this study is whet her the presence of highly visible marks led to an increase in predation either directly on adults or on newly dug nests. I do not have reason to suspect increased predation in my study site. Only a few basking alligators were observed during daytime, when marks could increase turtle de tectability. Although alligator tooth marks were observed on a few turtles, none of the turtles I recaptured after putting a paint number on them had fresh alligator marks. Effects of Handling Although most turtles were released immediatel y after marking and were manipulated in a similar manner, the recapture / re -sighting data suggested strong individual variation in response to handling. Some individuals moved several kilometers immediatel y post-handling, whereas
84 others tolerated repeated recaptu res and were consistently observed near the original capture location. Several studies recently have quantified the impacts of handling on stress levels and behavioral changes in reptiles. La ngkilde & Shine ( 2006) used a lizard (Eulamprus heatwolei ) as a model organism and concluded that toe-cl ipping and handling had minimal effect on corticosteroid levels, but microchip implantati on and exposure to new enclosures were more stressful. Cabanac & Bernieri (2000) demonstr ated tachycardia and be havioral rise in Tb as a result of 1-min handling of Glyptemys insculpta (Wood Turtle). However, Pike et al. (2005) failed to detect short-term effects on recapture rates of handled vs. unhandled Gopherus polyphemus (Gopher Tortoises). Their results were corroborated by Kahn et al. (2007) who concluded that handling, blood sampling, and temporary captivity did not lead to significant differences in plasma corticoste rone or movement patterns in G. polyphemus However, for aquatic turtles, Marchand (1945 a:77) reported that an unspecified number of marked turtles were common in  a distance of about [8 km] fr om the point of release; he attributed such movements to dispersal due to stress from prolonged capturing and abnormal density due to release of captured individua ls in one location. Suwannee C ooters live a long time, have extended home ranges, high adult survivorship, a nd are possibly not dependent on limited habitat resources (such as G. polyphemus burrows), and therefore might exhibit a strong response to disturbance even if predation ri sk in the form of handling was small (Gill et al. 2001; Beale & Monaghan 2004). Studies on stress hormones should help resolve this question. Home Range, Distribution across the Habitat, and Long-Term Movement Patterns of both home range and movement seem ed to be explained to a large extent by individual variation in turtle responses. Similar conclusions have been obtained in several other studies. Nesting female Suwannee Cooters in Waku lla River exhibited strong nest site fidelity
85 (within a 200 m segment), although they sometimes nested > 1.7 km away from previous nesting sites (D. Jackson & Walker 1997). Some individual Trachemys scripta elegans have preferred basking sites while others do not (Cagle 1944). I observed similar results in my study. Larger P. c. suwanniensis have greater home ranges than smaller individuals (based on n = 4; D. Jackson & Walker 1997), and this relati onship has been suggested to occur in other emydids ( T. s. scripta Schubauer et al. 1990). In this study, it appears that there might be a marked difference between adult and subadult individuals home range s, although a larger sample size is required. Habitat characteristics at the ri ver surface, such as river width, presence of basking logs, and vegetation mats might influence turtle hom e ranges, especially for basking herbivorous turtles (Marchand 1945a). However, home range estimators should be devised for aquatic turtles that take into account the thre e-dimensions of their habitat. Even though the distribution of bottom-growing vegetation is limited based on de pth and water clarity, I observed numerous turtles feeding at or below the surface in shallow (< 1.5 m) locations. Furthermore, courting was observed frequently at va rious depths underwater. While snorkeling, I observed P. c. suwanniensis escaping towards and hiding under seemingly familiar and preferred submerged logs a nd debris at depths of 2 m. However, turtles did not dive deeper than 10 m. River Rise is a low capacity and slow current spring > 30 m deep with a diameter around 40 m, with mostly vert ical limestone walls ex cept for a 5 m x 10 m terrace at a depth of around 9 m. Turtles resting on that terrace ha ve been captured with little effort, since they usually do not try to swim away. Many indivi duals were chased for > 5 min around the Rise. They swam continuously at a depth of around 3 m to prevent capture, generally swam along the side of the spring, and we re not observed diving deeper into the spring.
86 Whether physiological limitations or behavioral constraints precluded deeper dives is unknown. Suwannee Cooters typically live and thrive in lowvisibility tannin waters, so the decreased light intensity at depths > 10 m should not be a strong deterrent. A high risk of predation at the Rise also was an unlikely explanation. It is intere sting to note, however, that two radio-tracked P. concinna in Virginia were not observed deeper than 2 m, although the maximum water depth was only 3 m; non-radio-tracked turtles were also concentrated in water less than 2 m deep (Buhlmann & Vaughan 1991). My observations suggest potentially slightly larger female linear home ranges than the home ranges of 200 m reported by D. Jacks on & Walker (1997). However, even though the straight line mid-river home ranges might be slightly larger in the Santa Fe River, the total volume of water encompassed by the three-dime nsional home range of the turtles might be similar between the two populations. The Wakulla River was about 1 m deep, but for the most part averaged 120 m in width (D. Jackson & Walker 1997). D. Jackson & Walker (1997) further reported infrequent shifts of home range > 3 km, homing of 2 km, and possibly longdistance movements of 10 km. Other studies of P. concinna reveal slightly different pa tterns. In Virginia, Buhlmann & Vaughan (1991) found that the grea test distances moved by adult P. concinna were less than 600 m, but juveniles sometimes exhibited longer disper sal distances. Such results might be caused by the low abundance and different ecological conditions since the population studied was at the northern edge of the distribution of P. concinna In Florida, C. Jackson (1970) studied a population of P. c. suwanniensis whose individuals were rarely observed beyond the vicinity of a spring and its 160 m run.
87 Movement patterns and abundance of turtles might be related to availability of non-stable resources such as food, basking locations, or depth. For example, in March 2008, some of the basking logs and vegetation mats were no longer available and depth had increased slightly (~0.3 m); turtles were then observed in much greater numbers at locations where previously they were sparse (e.g. Sections 2 and 3). Fu rthermore, basking site availab ility may play a key role in motivating turtles to inhabit a specific area (Cagle 1944, 1950; Cagle & Chaney 1950). My observations suggest that turtles basked predominantly on site s corresponding to their size, with larger turtles choosi ng wider sites. Such site selection is supported by Boyer (1965) who found that unless sites were limited, turtles most often em erged on sites no less than twothirds of their body width. Painted Turtles ( Chrysemys) are able to discriminate the width of basking sites (Casteel 1911). Larger turtles are us ually found in mid-stream larger sites, with juveniles using smaller sites closer to the shore (Pluto & Bellis 1986). For P. c. suwanniensis partitioning might be due to different predator avoidance strategies, in which juveniles strive to hide quickly in submerged vegetatio n, which is often located closer to shore, while adults rely on speed and agility to avoid depredation. Non-social insects exhibit collective behavior s of self-organization and aggregate at suboptimal refuge due to presence of conspecifics (Halloy et al. 2007). Si milarly, presence or absence of other turtles might also be a factor in basking site selecti on. Boyer (1965) collected evidence that turtle aggregation might play a ro le in basking site selection even when other conditions were the same. Aggreg ation might be beneficial by increasing the ability of the group to spot predators, or increase mating opportunities. Another possibility is th at after a turtle is observed by conspecifics to use a basking site, othe r turtles assume it is of sufficient quality and congregate to utilize the resource I did not observe any intra-spec ific aggressive interactions.
88 The analysis of macrohabitat characteristics failed to detect differences in relative turtle abundances between different life st ages, but the relative counts s uggest differential habitat use. Even though hatchling turtles were not uniquely marked, I repeatedly observed similar numbers and size hatchlings on the same logs as detected on previous surveys. Such strong site fidelity is to be expected, because hatchlings are weaker swimmers than adults and it might be energetically expensive for them to move up and downstream, even in weak currents. Also, site fidelity might increase survival by increasing chances to escape from predators. Differential habitat use by juveniles and adults is known in other turtle species (e.g. Terrapene carolina Dodd et al. 1994; Hamilton 2000). G. Johnston and I observed similar clumped distribution of hatchlings in the section of the Santa Fe Rive r downstream from River Rise. We never observed hatchlings within the initial 0.5 km even t hough adults were abundant. Further downstream, however, hatchlings became common. Conclusions Continuous, long term studies on the movement and populat ion dynamics of Suwannee Cooter populations in blackwater rivers such as the Santa Fe are necessary to test the results from this study. During 2007, the population studied had a high density and no apparent skewed distribution of size classes or major shifts of sex ratio. Individual variation of behavior or undetected physiological differences influenced movement and home range of turtles; whether differences had physiological or be havioral causes is yet to be determined. Ample basking sites correlated strongly with the number of turtles observed and the ma intenance of snags and debris in the water should be a management priority.
89 Figure 4-1. Distribution of 89 indi viduals originally captured at the field site (circles; red group A, violet group B) and subsequently recaptured (diamonds; orange group A, blue group B). Sometimes multiple captures/observations were made at one point on the map.
90 Figure 4-2. Paint mark longevity based on 50 tu rtles marked with unique paint numbers and equipped with an external temperature l ogger (iButton). Uniden tified are turtles that lacked a clear mark but had a visibl e iButton, or in a few instances escaped before a positive identification was made. Turtles were marked 18 May 2007.
91 Figure 4-3. Distribution of observations of 89 P. c. suwanniensis originally captured in the field site. Sometimes multiple captures/observations were made at one point on the map.
92 Table 4-1. Physical differences between the different sections used for multivariate regression analysis. Area is river surface area in hectares; Depth is mean depth in meters; Twater is mean water temperatur e at 20 cm depth; Basking is number of counted basking objects in the section; Direction represents river or ientation in degrees; Total is combined counts for all tu rtles observed during the 21 samples. Section Area DepthTwate r Direction BaskingTotal 1 0.72 0.825.8 60727 2 1.16 0.425.7 1358116 3 1.00 0.626.7 13510372 4 0.32 0.827.1 135053 5 0.23 1.127.2 1354131 6 0.68 2.027.4 11014869 7 0.28 2.127.6 456372 8 0.12 2.827.4 55348 9 0.51 1.527.3 356245 10 0.12 1.727.0 20148 11 0.25 1.527.1 207341 12 0.20 1.726.9 17014578 13 0.33 2.227.0 10011717 14 0.45 2.827.8 11012362 15 1.04 3.228.2 18013806 16 0.59 1.528.0 12514466
93 Table 4-2. Home ranges based on mark/recapture a nd visual observations. Furthest denotes the mid-river distance from the furthest upstream to the furthest downstream points of observation for each individual; Capture-Las t is the mid-river distance between the original capture and the last observa tion; days, days are the number of days between original capture and last observation. Sex Furthest Capture-Last 30 days 61 days < 200 m < 200 m 1 0 < 500 m < 100 m 1 0 < 1000 m < 1000 m 0 1 < 2000 m < 800 m 1 1 < 1600 m < 1100 m 0 1 < 800 m < 150 m 7 2 < 2400 m < 1500 m 4 0 < 2000 m < 800 m 1 4 > 2800 m > 2300 m 0 1 Subadult < 300 m < 20 m 0 1
94 Table 4-3. Mean distance in mete rs (Dist) between initial capture and final observation, based on mark/recapture and visual observations. Days denotes mean number of days between initial capture and last observation. SD is one standard deviation. Sex Distance SD Range DaysSD Range n 484 444 558042396 625 692 673433219 Subadult 15 1041
95 Table 4-4. Lincoln-Petersen estimate of relative abundance of adult and subadult turtles. Total area of river sampled was 3.04 hectares. Turtles were marked 18 May 2007. Date 25 May 25 May 27 May 27 May 27 May Counting down/ upstream Down Up Up Down Down Total number observed 8566115167 77 Individuals marked 7 61510 8 Lincoln-Petersen 547.3487.1368.8777.9 441.0 Individuals / hectare 180.0160.2121.3255.9 145.1 95% confidence interval 365319274542 299
96 Table 4-5. Distribution across study sections and re lative abundance of di fferent life stages. 21 counts were carried 5 June 7 July 2007. Only head denotes observations of turtle heads on the surface that could not be definitely assigned to a lif e stage; likely the majority were adult turtles. Section Area (ha) Life stage Total count Average number Range SD Average number / ha 1 0.72 Adult 152.111.5 3.0 Subadult 42.011.4 2.8 Hatchling 61.511.0 2.1 Only head 11.010.0 1.4 Total 262.411.6 3.3 2 1.16 Adult 1096.815.8 5.9 Subadult 00.000.0 0.0 Hatchling 11.010.0 0.9 Only head 31.010.0 0.9 Total 1136.615.9 5.7 3 1.00 Adult 35117.677.1 17.6 Subadult 21.010.0 1.0 Hatchling 11.010.0 1.0 Only head 92.312.5 2.3 Total 36317.317.7 17.3 4 0.32 Adult 453.211.4 10.0 Subadult 21.010.0 3.1 Hatchling 00.000.0 0.0 Only head 31.010.0 3.1 Total 503.311.6 10.4 5 0.23 Adult 785.213.9 22.6 Subadult 41.310.6 5.8 Hatchling 31.010.0 4.3 Only head 233.312.4 14.3 Total 1086.414.5 27.6 6 0.68 Adult 52525.049.4 36.8 Subadult 613.111.9 4.5 Hatchling 814.512.5 6.6 Only head 1015.113.9 7.4 Total 76836.6812.7 53.8 7 0.28 Adult 25312.046.0 43.0 Subadult 171.711.2 6.1 Hatchling 422.811.9 10.0 Only head 303.314.1 11.9 Total 34216.359.5 58.2 8 0.12 Adult 24911.935.8 98.8 Subadult 151.510.8 12.5 Hatchling 81.110.4 9.5 Only head 383.513.3 28.8 Total 31014.866.1 123.0 Continues
97 Table 4-5. Continued Section Area (ha) Life stage Total count Average number Range SD Average number / ha 9 0.51 Adult 1226.424.8 12.6 Subadult 131.410.5 2.8 Hatchling 121.310.5 2.6 Only head 493.312.2 6.4 Total 1969.825.4 19.2 10 0.12 Adult 342.411.8 20.2 Subadult 21.010.0 8.3 Hatchling 42.011.4 16.7 Only head 41.310.6 11.1 Total 442.812.2 22.9 11 0.25 Adult 24411.644.3 46.5 Subadult 142.011.2 8.0 Hatchling 252.111.2 8.3 Only head 292.412.2 9.7 Total 31214.945.7 59.4 12 0.20 Adult 46722.2310.0 111.2 Subadult 222.811.8 13.8 Hatchling 131.310.5 6.5 Only head 383.512.3 17.3 Total 54025.7710.6 128.6 13 0.33 Adult 55226.31467.0 79.7 Subadult 322.311.4 6.9 Hatchling 533.111.3 9.4 Only head 404.013.6 12.1 Total 67732.21857.8 97.7 14 0.45 Adult 26212.555.6 27.7 Subadult 291.811.2 4.0 Hatchling 452.511.3 5.6 Only head 131.910.7 4.1 Total 34916.685.8 36.9 15 1.04 Adult 58127.711212.0 26.6 Subadult 362.111.1 2.0 Hatchling 552.811.5 2.6 Only head 673.912.5 3.8 Total 73935.217611.5 33.8 16 0.59 Adult 34316.345.8 27.7 Subadult 261.610.6 2.8 Hatchling 352.311.1 4.0 Only head 312.411.8 4.0 Total 43520.796.7 35.1 All 8.00 Adult 4230201.411825839.7 25.2 Subadult 27913.335.8 1.7 Hatchling 38418.347.6 2.3 Only head 47922.8515.9 2.9 Total 5372255.817734642.2 32.0
98Table 4-6. Correlation coefficients between macrohabitat variables and turtle counts. Sec Area Dt Twate r BS CD Adu SHat Hat H T-H Total Section -0.250.710.740.480.16 0.550.510.500.350.560.55 Area -0.25-0.250.420.45 0.190.100.110.070.180.17 Depth (Dt) 0.740.34-0.01 0.580.650.640.550.610.61 Twate r 0.270.07 0.490.570.560.520.510.52 Basking sites (BS) 0.59 0.800.680.650.540.810.79 Compass direction (CD) 0.490.218.104.22.1680.44 Adults (Adu) 0.810.770.771.000.99 Subadults and hatchlings (SHat) 0.990.820.860.87 Hatchlings (Hat) 0.770.830.84 Heads (H) 0.800.84 Total without heads (T-H) 1.00
99 CHAPTER 5 SUMMARY AND CONSERVATION RECOMMENDATIONS Summary The unusual conditions of decreased tannin load during 2006 allowed me to record observations on the ecology of Suwannee Cooters in an understudied hab itat, the blackwater spring-fed river. Suwannee Cooters were abundant in the river section th at I sampled, with an estimated 25 adults per ha of river. I also obs erved numerous hatchlings and juvenile turtles, suggesting a successfully reproducing populatio n, although recruitment rates are unknown. Repetitive basking counts dur ing favorable conditions (Tair higher than Twater) can be used to continuously monitor population trends, esp ecially under normal conditions of high tannin concentrations when capturing turtles is difficult. Externally positioned temperature loggers (iButtons) are useful in revealing important biological data about the patterns of behavior of adult aquatic turtles. If coupled with more extensive knowledge about e nvironmental conditions and individual movement across the habitat, they can be used to obtain precise, long-term continuous information on behavior and thermal preferences. I obtained thermal profiles of 10 individuals, although I could not discriminate specific basking events equally well. The data suggested that timing and duration of basking varied between individuals and were likel y not linked to sexual differences However, turtles rarely if ever undertook aerial basking if Tair was lower than Twater. It is unknown whether differences existed in core Tb and whether minor differences transl ated into biologically significant physiological and ecological consequences; however turtles were likely actively maintaining a preferred Tb.
100 Turtle distribution throughout the habitat wa s positively correlated with the number of basking sites. Although the le vel of importance of aerial basking for elevation of Tb under the constantly warm water conditions of the Santa Fe River is yet to be determined, turtles basked extensively on logs and rocks, and utilized the subm erged portions of these sites for refuge. Suwannee Cooters are highly vagile and capab le of covering long distances rapidly. For adults, the propensity to move across the habita t, and the size of the actively used home range, varied extensively and were better explained by indivi dual variation than by sexual differences. Whether long distance movements were predomin antly a fleeing response to handling or occur naturally, and what is the rela tive distribution of sedentary and nomadic individuals, is unknown. Painting marks on the carapace was an inexpensiv e and useful method to track individual turtles for short periods of time. Application of more paint greatly increases the time the painted number can be recognized, and allows for obser vations of individuals up to six months after capture. Paddle boating had limited influence on the population, which was relatively abundant, healthy, and reproduced successfully in this stretch of the river. Individuals varied in their predisposition to terminate baski ng, but often resumed it within minutes of disturbance. Visual observation of the TiButton data suggested that turtles ba sking behavior was dependent on Tair more than the amount of boating traffic. However, their behavior was also partially decoupled from environmental conditions and turtles bask ed different amounts of time during days with similar weather conditions. My results support the notion proposed by the ri sk-disturbance hypothesis that habituation is only partial and that turtles have generalized threatening stimuli leadin g to fleeing. Individual
101 responses to disturbance varied in intensity, but whether shortterm physiological or long-term psychological differences account for such behavi ors was not determined. I recommend future studies of disturbance effects on riverine turtles to focus on qua ntifying physiological conditions of the organisms, since behavior-only observati ons are insufficient in determining the true impact of the disturbance. Conservation Recommendations Historically, turtles of the genus Pseudemys have attained high densities and biomass in rivers in Florida (Marchand 1942, 1945 a; C. Jackson 1970; Carr 1983). Several recent studies have observed high densities (D. Jackson & Wa lker 1997; Huestis & Meylan 2004; this study) throughout the state, but these popula tions likely have experienced decreases from their historic numbers (Marchand 1945a; Meylan et al. 1992). D. Jacks on (2006) provided extensive options and recommendations statewide for conserving P. c. suwanniensis Here, I focus on suggestions relevant to the Sant a Fe River population. Although likely not as numerous as in the recent past, the segmen t of the population I studied appeared stable. If conditions do not change the Cooters will likely persist in numbers. However, several factors must be considered. Continued research of this and other populations of P. c. suwanniensis in blackwater rivers is needed in order to provide much needed baseline biological data, necessary for the establishment of adequate management and conservation practices. Long-term studies of population dynamics, focusing on recruitment, surv ivorship, and growth, w ill allow detection of trends and problems I could not observe due to th e limited focus and time span of this research. Accumulating additional data s hould help to establish a monitoring program, a vital tool for turtle conservation. Turtles are characterized by high adult su rvivorship and longevity, but low recruitment coupled with high juvenile mortality and difficult detection of hatchlings.
102 Without an established monitoring program to de tect lack of recruitment, a population with a large number of adult turtles, but no juvenile individuals, might collapse before conservation actions are taken. Boating and recreational uses of the Santa Fe River and similar water bodies are essential for their conservation by creating and maintaining an a ppreciation of their intrinsic beauty and importance. However, uses must be limited to those of minimal impact to the animals and the environment. The current low levels of paddle boating on the river along th e River Rise Preserve State Park does not seem to be adversely affecting the population, but should likely not be exceeded if turtle abundan ce is to be preserved. Observations of turtle populations should be extended to other sections of the river experiencing heavier use by boats. The effects of motorboats, which might be more pronounced than disturbance from paddle boats, need to be studied, and regulations, if necessary, implemented concerning the use of waterways for recreational purposes. A management plan should be devised to control water hyacinth (Eichhornia crassipes), an introduced aquatic plant. Comm on throughout the sampled portion of the river, water hyacinth seem to outcompete other plants that provide refugia, surface basking, and food for Suwannee Cooters. In July 2008 E. crassipes covered at least 300 m of ri ver surface about 1.5 km upstream from the US Highway 441 bridge and about 200 m of the river adjacent to the public boat ramp. This made boating almost impossible, practically stopped surface flow, and likely negatively impacted aquatic biota. This study further supports the importance of maintaining high amounts of debris to provide basking platforms for aqua tic turtles. Enough forested lands extending to th e edge of the river should be preserved to serv e as natural source of snags and logs. Such unimproved river
103 features should lead to higher growthand survi vorship rates of turtles, and are essential for maintaining population size. Currently, debates exist about the addition of four additional large water-bottling companies within 5 km of th e riverfront, besides the exis ting one. By drawing over 492,100 m3 of spring water annually, the proposed changes w ill likely result in decreased river water levels, changes in river flow, vegetation and animal communities, and could have major adverse ecological effects. Such alterations in the riverine habitat are likely the biggest current threat to the population of the Suwannee Cooter in this section of the river. Turtle consumption for food, especially of the lucrative large female turtles, continues to be a major threat to chelonian populations worldwide. Current laws and regulations in Florida are ecologically unsound and allow the take of Su wannee Cooters (up to two turtles per day per person), with only a closed harvest season durin g primary nesting, from 15 April to 30 July. Therefore, even legal take can decimate a popula tion rapidly. Furthermore, regulatory control is made difficult because enforcement officers need to be able to distinguish between the closely related species in the genera Pseudemys and Trachemys Therefore, a complete statewide ban on the collection of riverine turtle s should be implemented, except for scientific or conservation purposes.
104 APPENDIX A MORPHOMETRIC PROCEDURES I placed turtles in a plastic mesh bag and us ed appropriately sized Pesola spring scales (Pesola AG, Baar, Switzerland; 10 kg, 2.5 kg, 1 kg, 500 g) to record body mass to the nearest mark. The weight of the bag was measured before each weighing and subtracted from the total weight. I took shell dimensions to the nearest m illimeter using digital and tree calipers. Straight carapace length (CL) was measured dorsally alo ng the mid-dorsal line from the tip of the nuchal scute to the seam between the two supracaudal scutes. I took st raight plastron length ventrally along the mid-dorsal line from the seam between th e gular scutes to the seam between the anal scutes. I moved the calipers front to back over the shell and recorded the greatest width as a maximum shell width. Shell depth was measured along the mid-dorsal line by placing the bottom part of the calipers on the plastr on of the turtle, and then placing the top part of the calipers on the carapace. I recorded the length of the longest foreclaw from its visible point of emergence from live tissue to the tip. I determined sex based on the size and shape of the tail, the position of the cloaca relative to the base of the tail, the shape of the carapace, and the length of the fore claws. Turtles with non-discriminate sex ch aracteristics and ones w ith CL < 200 mm were considered unknown-sex subadults. Pictures of th e carapace and the plastron of each turtle were taken as vouchers and for subsequent verification, using unique color patterns or shell characteristics.
105 APPENDIX B TUKEY MULTIPLE COMPARISON OF MEANS Table B-1. Tukey multiple comparisons of m eans 95% confidence level test comparing individual turtles mean TiButton. Factor Difference Lower Upperp adjusted iB1109-iB1122 0.483 0.3670.6000.000 iB1109-iB1133 0.417 0.3010.5340.000 iB1109-iB1137 0.260 0.1430.3760.000 iB1109-iB1140 0.221 0.1040.3370.000 iB1109-iB1145 0.844 0.7280.9610.000 iB1109-iB1151 0.436 0.3170.5550.000 iB1109-iB1152 0.085 -0.0320.2020.383 iB1109-iB1245 0.216 0.0990.3320.000 iB1109-iB926 0.186 0.0700.3030.000 iB1122-iB1145 0.361 0.2440.4780.000 iB1133-iB1122 0.066 -0.0510.1830.741 iB1133-iB1145 0.427 0.3100.5440.000 iB1133-iB1151 0.019 -0.1000.1371.000 iB1137-iB1122 0.223 0.1070.3400.000 iB1137-iB1133 0.157 0.0410.2740.001 iB1137-iB1145 0.585 0.4680.7010.000 iB1137-iB1151 0.176 0.0580.2950.000 iB1140-iB1122 0.262 0.1460.3790.000 iB1140-iB1133 0.196 0.0800.3130.000 iB1140-iB1137 0.039 -0.0780.1560.988 iB1140-iB1145 0.624 0.5070.7400.000 iB1140-iB1151 0.215 0.0970.3340.000 iB1151-iB1122 0.047 -0.0710.1660.962 iB1151-iB1145 0.408 0.2900.5270.000 iB1152-iB1122 0.398 0.2820.5150.000 iB1152-iB1133 0.332 0.2160.4490.000 iB1152-iB1137 0.175 0.0580.2910.000 iB1152-iB1140 0.136 0.0190.2520.009 iB1152-iB1145 0.759 0.6430.8760.000 iB1152-iB1151 0.351 0.2320.4690.000 iB1152-iB1245 0.130 0.0140.2470.015 iB1152-iB926 0.101 -0.0150.2180.155 iB1245-iB1122 0.268 0.1510.3840.000 iB1245-iB1133 0.202 0.0850.3180.000 iB1245-iB1137 0.044 -0.0720.1610.972 iB1245-iB1140 0.005 -0.1110.1221.000 iB1245-iB1145 0.629 0.5120.7450.000
106 Table B-1. Continued Factor Difference Lower Upperp adjusted iB1245-iB1151 0.220 0.1020.3390.000 iB926-iB1122 0.297 0.1800.4140.000 iB926-iB1133 0.231 0.1140.3480.000 iB926-iB1137 0.073 -0.0430.1900.605 iB926-iB1140 0.034 -0.0820.1510.995 iB926-iB1145 0.658 0.5410.7750.000 iB926-iB1151 0.250 0.1310.3680.000 iB926-iB1245 0.029 -0.0870.1460.999
107 LIST OF REFERENCES Angilletta, M. J. & Krochmal, A. R. 2003. The Thermochron: A truly miniature and inexpensive temperature-logger. Herpetological Review 34, 31. Aresco, M. J. 2005. Mitigation measures to reduce highway mortality of turtles and other herpetofauna at a north Florida lake. Journal of Wildlife Management 69, 549. Aresco, M. J., & Jackson, D. R. 2006. Geographic distribution. Trachemys scripta elegans Herpetological Review 37, 239. Auth, D. L. 1975. Behavioral ecology of baski ng in the yellow-bellied turtle, Chrysemys scripta scripta (Schoepff). Bulletin of the Florida Stat e Museum Biological Sciences 20, 1. Avery, H. W., Spotila, J. R., Congdon, J. D., Fish er Jr., R. U., Standora, E. A. & Avery, S. B. 1993. Roles of diet protein and temperature in the growth and nutritional energetics of juvenile slider turtles, Trachemys scripta Physiological Zoology 66, 902. Beale, C. M. & Monaghan, P. 2004. Behavioural responses to human disturbance: A matter of choice? Animal Behaviour 68, 1065. Beaupre, S. J. & Beaupre, R. W. 1994. An inexpensive data-collection system for temperature telemetry. Herpetologica 50, 509. Bejder, L., Samuels, A. M. Y., Whitehead, H. A. L., Gales, N., Mann, J., Connor, R., Heithaus, M., Watson-Capps, J., Flaherty, C. & Krutzen, M. 2006. Decline in relative abundance of bottlenose dolphins exposed to long-term disturbance. Conservation Biology, 20, 1791. Bejder, L., Samuels, A., Whitehead, H. & Gales, N. 2006. Interpreting shor t-term behavioural responses to disturbance within a longitudinal perspective. Animal Behaviour, 72, 1149 1158. Bishop, J. M. 1983. Incidental capture of Diamondback Terrapin by crab pots. Estuaries 6, 426. Bjorndal, K. A., Bolten, A. B., Lagueux, C. J., & Jackson, D. R. 1997. Dietary overlap in three sympatric congeneric freshwater turtles ( Pseudemys ) in Florida. Chelonian Conservation and Biology, 2, 430. Blouin-Demers, G. & Weatherhead, P. J. 2001. Therma l ecology of black rat snakes ( Elaphe obsoleta ) in a thermally challenging environment. Ecology, 82 3025. Bodie, J. R. 2001. Stream and riparian manage ment for freshwater turtles. Journal of Environmental Management 62, 443. Bogert, C. M. 1959. How reptiles regulate their body temperature. Scientific American 200, 105.
108 Boyer, D. R. 1965. Ecology of the baski ng habit in turtles. Ecology, 46, 99. Buhlmann, K. A. & Vaughan, M. R. 1991. Ecology of the turtle Pseudemys concinna in the New River, West Virginia. Journal of Herpetology 25, 72. Buhlmann, K. A. & Gibbons, J. W. 1997. Imperiled aquatic reptiles of the southeastern United States: Historical review and current conservation status. In: Aquatic fauna in peril: The southeastern perspective (Ed. by Benz, G. W. & Collins, D. E.). Decatur, GA: Southeast Aquatic Research Institute, Lenz Design & Communications. Burt, W. H. 1943. Territoriality and home range c oncepts as applied to mammals. Journal of Mammalogy 24, 346. Cabanac, M. & Bernieri, C. 2000. Behavioural rise in body temperature and tachycardia by handling of a turtle ( Clemmys insculpta ). Behavioural Processes 49, 61. Cagle, F. R. 1939. A system of marking tur tles for future identification. Copeia 1939, 170. Cagle, F. R. 1944. Home range, homing behavior, and migration in turtles. Miscellaneous Publications of the Museum of Z oology at University of Michigan 61, 1. Cagle, F. R. 1950. The life history of the slider turtle, Pseudemys scripta troostii (Holbrook). Ecological Monographs 20, 31. Cagle, F. R. & Chaney, A. H. 1950. Turtle populations in Louisiana. American Midland Naturalist, 43 383. Carr, A. F. 1952. Handbook of turtles. The turtles of the United States, Canada, and Baja California. Ithaca, NY: Cornell University Press. Carr, A. F. 1983. All the way down upon the Suwannee River. Audubon, 2, 78. Casteel, D. B. 1911. The discriminative ability of the painted turtle. Journal of Animal Behavior 1, 1. Chambers, J., Cleveland, W., Kleiner, B. & Tukey, P. 1983. Graphical methods for data analysis Boston: Wadsworth Intern ational, Duxbury Press. Congdon, J. D. 1989. Proximate and evolutionary constr aints on energy relations of reptiles. Physiologi cal Zoology 62 356. Congdon, J. C., Duhnam, A. E. & van Loben Sels, R. C. 1994. Demographics of common snapping turtles ( Chelydra serpentina ): Implications for conservation and management of long-lived organisms. American Zoologist 34, 397. Constantine, R., Brunton, D. H. & Dennis, T. 2004. Dolphin-watching tour boats change bottlenose dolphin ( Turiops truncates ) behaviour. Biological Conservation 117, 299 307.
109 Cossins, A. R. & Bowler, K. B. 1987. Temperature biology of animals New York, U.S.A.: Chapman and Hall, New York. Cowles, R. B. & Bogert, C. M. 1944. A preliminary study of the thermal requirements of desert reptiles. Bulletin of the American Mu seum of Natural History 83, 265. Crawford, K. M., Spotila, J. R. & Standora, E. A. 1983. Operative environmental temperatures and basking behavior of the turtle Pseudemys scripta Ecology 64, 989 999. Crenshaw, J. W. 1955. The ecological geography of the Pseudemys floridana complex in the southeastern United States. Dissert ation, University of Florida. Dall'Antonia, L., Lebboroni, M., Benvenuti, S. & Chelazzi, G. 2001. Data loggers to monitor activity in wild freshwater turtles. Ethology, Ecology & Evolution 13, 81. Dill, L. M. 1974. The escape response of the zebra danio ( Brachydanio rerio). I. The stimulus for escape. Animal Behaviour, 22, 711. Dill, L. M. 1974. The escape response of the zebra danio ( Brachydanio rerio). II. The effect of experience. Animal Behaviour, 22, 723. Dodd, C. K., Jr. 1988. Disease and population declines in the flattened musk turtle Sternotherus depressus American Midland Naturalist 119, 394. Dodd, C. K., Jr. 1990. Effects of habitat fragmentat ion on a stream-dwelling species, the flattened musk turtle Sternotherus depressus. Biological Conservation 52, 33. Dodd, C. K., Jr., Franz, R. & Smith, L. 1994. Activity patterns and habitat use of box turtles ( Terrapene carolina bauri ) on a Florida island, with recommendations for management. Chelonian Conservation and Biology 1, 97. Dodd, C. K., Jr., Barichivich, W. J. & Smith, L. L. 2004. Effectiveness of a barrier wall and culverts in reducing wildlife mortality on a heavily traveled highway in Florida. Biological Conservation 118, 619. Dorcas, M. E. & Peterson, C. R. 1997. Head-body temperature differences in free-ranging rubber boas. Journal of Herpetology 31, 87. Dorcas, M. E. & Peterson, C. R. 1998. Daily body temp erature va riation in free-ranging rubber boas. Herpetologica 54, 88. Dorcas, M. E., Willson, J. D. & Gibbons, J. W. 2007. Crab trapping causes population decline and demographic changes in diamondback terrapins over two decades. Biological Conservation 137, 334. Ernst, C. H., Lovich, J. & Barbour, R. W. 1994. Turtles of the United States and Canada Washington and London: Smiths onian Institution Press.
110 Florida Fish and Wildlife Conservation Commission (FFWCC) 2005. Floridas Wildlife Legacy Initiative. Floridas Comprehensiv e Wildlife Conservation Strategy. Tallahassee, Florida, USA. Frair, W., Ackman, R. G. & Mrosovsky, N. 1972. Body temperature of Dermochelys coriacea warm turtle from cold water. Science 177, 791. Frid, A. & Dill, L. 2002. Human-caused disturbance stimu li as a form of predation risk. Conservation Ecology, 6 11 Fuselier, L., & Edds, D. 1994. Habitat partitioning among three sympatric species of map turtles, Genus Graptemys. Journal of Herpetology 28, 154. Garber, S. D. & Burger, J. 1995. A 20-yr study documenting the relationship between turtle decline and human recreation. Ecological Applications 5, 1151. Gatten, R. E. 1974. Effect of nutritional status on preferred body temperature of turtles Pseudemys scripta and Terrapene ornata. Copeia 1974, 912. Gibbons, J. W., Scott, D. E., Ryan, T. J. Buhlmann, K. A., Tuberville, T. D. Metts, B., Greene, J. L., Mills, T. M., Leiden, Y., Poppy, S. M., & Winne, C. T. 2000. The global decline of reptiles, dj vu amphibians. BioScience, 50, 653. Gill, J. A., Sutherland, W. J. & Watkinson, A. R. 1996. A method to quantify the effects of human disturbance on animal populations. Journal of Applied Ecology 33, 786. Gill, J. A., Norris, K. & Sutherland, W. J. 2001. Why behavioural responses may not reflect the population consequences of human disturbance. Biological Conservation 97, 265 268. Giovanetto, L. A. 1992. Population ecology and relative ab undance of sympatric freshwater turtles in the headwaters of two spring-fe d rivers in western peninsular Florida. Dissertation, Florida Institute of Technology. Grayson, K. L. & Dorcas, M. E. 2004. Seasonal temperature varia tion in the painted turtle ( Chrysemys picta ). Herpetologica 60 325. Halloy, J., Sempo, G., Caprari, G., Rivault, C., Asadpour, M., Tache, F., Said, I., Durier, V., Canonge, S., Ame, J. M., Detrain, C., Correll, N., Martinoli, A., Mondada, F., Siegwart, R. & Deneubourg, J. L. 2007. Social integration of robots into groups of cockroaches to control self-organized choices. Science, 318, 1155. Hamilton, A. M. 2000. Evidence for ontogenetic shifts in box turtles: Activity patterns, movem ents, and use of microenviro nments and habitats by juvenile Terrapene carolina bauri on Egmont Key, Florida. Thes is, University of Florida.
111 Hammond, K. A., Spotila, J. R. & Standora, E. A. 1988. Basking behavior of the turtle Pseudemys scripta effects of digestive state, acclimation temperature, sex, and season. Physiological Zoology 61 69. Harlow, H. J., Hillman, S. S. & Hoffman, M. 1976. Effect of temperature on digestive efficiency in herbivorous lizard, Dipsosaurus dorsalis Journal of Comparative Physiology, 111 1. Hecnar, S. J. & M'Closkey, R. T. 1998. Effects of human distur bance on five-lined skink, Eumeces fasciatus, abundance and distribution. Biological Conservation 85 213. Hester, J. M., Price, S. J. & Dorcas, M. E. 2008. Effects of relocation on movements and home ranges of eastern box turtles. Journal of Wildlife Management 72, 772. Hood, G. M. 2008. PopTools version 3.0.3. Available on the internet. URL http://www.cse.csiro.au/poptools Hornsby, D., & Ceryak, R. 1998. Springs of the Suwannee River Basin in Florida Live Oak, FL: Suwannee River Water Manageme nt District. Huestis, D. L. & Meylan, P. A. 2004. The turtles of Rainbow Run (Marion County, Florida): Observations on the genus Pseudemys. Southeastern Naturalist 3, 595. Huey, R. B. & Slatkin, M. 1976. Costs and benefits of lizard thermoregulation. Quarterly Review of Biology, 51, 363. Huey, R. B. 1982. Temperature, physiology, and the ecology of reptiles. In: Biology of the Reptilia (Ed. by Gans, C. & Pough, F. H.), pp. 25. New York, U.S.A.: Academic Press. Hutchison, V. H., Herndon, G. D. & Vinegar, A. 1966. Thermoregulation in a brooding female Indian Python, Python molurus bivittatus Science 151, 694. Hutchison, V. H. 1979. Thermoregulation. In: Turtles: Perspectives and research (Ed. by Harless, M. & Morlock, H.), pp. 207. Ne w York, USA: J. Whiley and Sons. Iverson, J. B. 1982. Biomass in turtle populationsa neglected subject. Oecologia 55 69. Jackson, C. G, Jr. & Jackson, M. M 1968. The egg and hatchling of the Suwannee terrapin. Quarterly Journal of the Fl orida Academy of Sciences 31, 199. Jackson, C. G. Jr. 1970. A biometrical study of growth in Pseudemys concinna suwanniensis Copeia 1970, 528. Jackson, C. G., Jr. & Davis, J. D. 1972. Courtship display behavior of Chrysemys concinna suwanniensis Copeia 1972, 385.
112 Jackson, D. R. 1988. Reproductive strategies of sympatri c freshwater emydid turtles in northern peninsular Florida. Bulletin of the Florida State Museum 33, 113. Jackson, D. R. & Walker, R. N. 1997. Reproduction in the Suwannee Cooter, Pseudemys concinna suwanniensis Bulletin of the Florida Museum of Natural History 41, 69. Jackson, D. R. 2003. Pseudemys concinna suwanniensis (Suwannee Cooter). Reproduction, life span. Herpetological Review 34, 363. Jackson, D. R. 2006. Pseudemys concinna River Cooter. In: Biology and conservation of Florida turtles. (Ed. by Meylan, P.), pp. 325. Oviedo, FL: Chelonian Research Foundation. Johnson, S. A., Bjorndal, K. A. & Bolten, A. B. 1996. Effects of organized turtle watches on Loggerhead ( Caretta caretta ) nesting behavior and hatc hling production in Florida. Conservation Biology 10 570. Johnston, G., Kornilev, Y. V., Lau, A. Ecology of a Suwannee Cooter ( Pseudemys concinna suwanniensis ) population in a north Flor ida blackwater stream. In prep. Kahn, P. F., Guyer, C. & Mendonca, M. T. 2007. Handling, blood sampling, and temporary captivity do not affect plasma corticosterone or movement patterns of Gopher Tortoises ( Gopherus polyphemus ). Copeia 2007, 614. Kenyon, W. A. 1925. Digestive enzymes in poikilothermal vertebrates, an investigation of enzymes in fishes, with comparative studies of those in amphibians, reptiles, and mammals. Bulletin of U.S. Fisheries 41, 179. King, J. M. & Heinen, J. T. 2004. An assessment of the behavior s of overwintering manatees as influenced by interactions with touris ts at two sites in central Florida. Biological Conservation 117, 227. Kornilev, Y. V., Price, S. J., & Dorcas, M. E. 2006. Between a rock and a hard place: Responses of eastern box turtles ( Terrapene carolina) when trapped between railroad tracks. Herpetological Review 37, 145. Kramer, M. 1995. Home range of the Florida Red-bellied Turtle ( Pseudemys nelsoni ) in a Florida spring run. Copeia 1995, 883-890. Lagueux, C. J., Bjorndal, K. A., Bolten, A. B. & Campbell, C. L. 1995. Food habits of Pseudemys concinna suwanniensis in a Florida spring. Journal of Herpetology 29, 122 126. Langkilde, T. & Shine, R. 2006. How much stress do researchers inflict on their study anim als? A case study using a scincid lizard, Eulamprus heatwolei Journal of Experimental Biology, 209 1035.
113 Lillywhite, H. B. 1970. Behavioral temperature regulation in the bullfrog, Rana catesbeiana Copeia, 1970, 158. Lindeman, P. V. 1999. Surveys of basking map turtles Graptemys spp. in three river drainages and the importance of deadwood abundance. Biological Conservation 88 33. Marchand, L. J. 1942. A contribution to a knowledge of the natural history of certain freshwater turtles. Thesis, University of Florida. Marchand, L. J. 1945. The individual range of some Florida turtles. Copeia 1945, 75. Marchand, L. J. 1945. Water gogglinga new method for the study of turtles. Copeia 1945, 37. Meylan, P. A., Stevens, C. A., Barnwell, M. E., & Dohm, E. D. 1992. Observations on the turtle community of Rainbow Run, Marion Co., Florida. Florida Scientist, 55, 219. Moll, D. & Moll, E. O. 2004. The ecology, exploitation, and conservation of river turtles Oxford, UK: Oxford University Press. Moll, D. 1990. Population sizes and foraging ecology in a tropical freshwater stream turtle community. Journal of Herpetology 24, 48. Moll, E. O. & Legler, J. M. 1971. The life history of a ne otropical slider turtle, Pseudemys scripta (Schoepff), in Panama. Bulletin of the Los Angeles City Museum of Natural History and Science 11. Monagas, W. R. & Gatten Jr., R. E. 1983. Behavioural fever in the turtles Terrapene carolina and Chrysemys picta Journal of Parasitology 8 285. Moore, M. J. C. & Seigel, R. A. 2006. No place to nest or bask : Effects of human disturbance on the nesting and basking habits of yellow-blotched map turtles (Graptemys flavimaculata ). Biological Conservation 130, 386. Mueller, J. M. & Rakestraw, D. L. 1995. Evaluation of a new miniature temperature data logger. Herpetological Review 26, 22. Neill, W. T. & Allen, E. R. 1954. Algae on turtlessom e additional considerations. Ecology, 35, 581. Nussear, K. E., Esque, T. C. & Tracy, C. R. 2002. Continuously reco rding body temperature in terrestrial chelonians. Herpetological Review 33, 113. Parent, C. & Weatherhead, P. J. 2000. Behavioral and life hi story responses of eastern massasauga rattlesnakes ( Sistrurus catenatus catenatus ) to human disturbance. Oecologia 125 170.
114 Parmenter, R. R. 1981. Digestive turnover rates in fres hwater turtles: The influence of temperature and body size. Comparative Biochemistry and Physiology 70A 235. Pearson, D., Shine, R. & Williams, A. 2003. Thermal biology of large snakes in cool climates: A radio-telemetric study of carpet pythons ( Morelia spilota imbricata ) in south-western Australia. Journal of Thermal Biology 28, 117. Peterson, C. R., Gibson, A. R. & Dorcas, M. E 1993. Snake thermal ecology: The causes and consequences of body temperature variation. In: Snakes: Ecology and behavior (Ed. by Seigel, R. A. & Collins, J. T.), pp. 241. New York, U.S.A.: McGraw-Hill, New York. Pike, D. A., Dinsmore, A., Crabill, T., Smith, R. B. & Seigel, R. A. 2005. Short-term effects of handling and permanently marking gopher tortoises ( Gopherus polyphemus ) on recapture rates and behavior. Applied Herpetology, 2, 139. Plummer, M. V. 1977. Activity, habitat and populat ion structure in the turtle, Trionyx muticus. Copeia 1977, 431. Plummer, M. V., Crabill, T. L. & Mills, N. E. 2005. Body temperatures of free-ranging softshell turtles ( Apalone spinifera ) in a small stream. Herpetological Review 36, 371 375. Pluto, T. G. & Bellis, E. D. 1986. Habitat utilization by the turtle, Graptemys geographica along a river. Journal of Herpetology 20, 22. Pritchard, P. C. H. & Greenhood, W. F. 1968. The sun and the turtle. International Turtle and Tortoise Society Journal, 2, 20. Pritchard, P. C. H. 1980. Record size turtles from Florida and South America. Chelonologica 1, 113. R Development Core Team. 2008. R: A language and environmen t for statistical computing. R Foundation for Statistical Computing, Vi enna, Austria. ISBN 3-900051-07-0, URL http://www.R-project.org Reese, D. A. & Welsh, H. H., Jr 1998. Habitat use by western pond turtles in the Tr inity River, California. Journal of Wildlife Management 62, 842. Rhodin, A. 1999. Celebrate the turtle : Perception and preserva tion. Paper presented at Powdermill 1999, 4th Occasional Freshwater Turtle Conference, 13 Aug 1999; Laughlin, NV. Robert, K. A. & Thompson, M. B. 2003. Reconstructing Thermochron iButtons to reduce size and weight as a new technique in th e study of small animal thermal biology. Herpetological Review 34, 130.
115 Rodgers, J. A. & Schwikert, S. T. 2002. Buffer-zone distances to protect foraging and loafing waterbirds from disturbance by personal watercraft and outboard-powered boats. Conservation Biology 16 216. Roosenburg, W. M., Cresko, W., Mode sitte, M. & Robbins, M. B. 1997. Diamondback Terrapin ( Malaclemys terrapin ) mortality in crab pots. Conservation Biology 11, 1166 1172. Ryan, T.J., & Lambert, A. 2005. Prevalence and colonization of Placobdella on two species of freshwater turtles (Graptemys geographica and Sternotherus odoratus ). Journal of Herpetology 39, 284. Sajwaj, T. D. & Lang, J. W. 2000. Thermal ecology of Blanding's Turtle in central Minnesota. Chelonian Conservation and Biology 3, 626. Schubauer, J. P., Gibbons, J. W. & Spotila, J. R. 1990. Home range and movement patterns of slider turtles inhabiting Par Pond. In: Life history and ecology of the Slider Turtle (Ed. by Gibbons, J. W.), pp. 223. Washington, D.C. : Smithsonian Institution Press. Sexton, O. J. 1959. Spatial and temporal movements of a population of the Painted Turtle, Chrysemys picta marginata (Agassiz). Ecological Monographs, 29, 113. Shine, R. & Madsen, T. 1996. Is thermoregulation important for most reptiles? An example using water pythons ( Liasis fuscus ) in tropical Australia. Physiological Zoology 69, 252 269. Smart, R., Whiting, M. J. & Twine, W. 2005. Lizards and landscapes: Integrating field surveys and interviews to assess the impact of human disturbance on lizard assemblages and selected reptiles in a savanna in South Africa. Biological Conservation 122 23. Standora, E. A. 1982. A telemetric study of the thermoregulatory behavior and climate space of free-ranging yellow-bellied turtles, Pseudemys scripta Dissertation, University of Georgia. Steen, D. A. & Gibbs, J. P. 2004. Effects of roads on the stru cture of freshwater turtle populations. Conservation Biology 18, 1143. Steen, D. A., Aresco, M. J., Beilke, S. G., Comp ton, B. W., Condon, E. P., Dodd, C. K., Jr., Forrester, H., Gibbons, J. W., Greene, J. L ., Johnson, G., Langen, T. A., Oldham, M. J., Oxier, D. N., Saumure, R. A., Schueler, F. W., Sleeman, J. M., Smith, L. L., Tucker, J. K. & Gibbs, J. P. 2006. Relative vulnerability of female turtles to road mortality. An imal Conservation, 9, 269. Stockwell, C. A., Bateman, G. C. & Berger, J. 1991. Conflicts in natio nal parksa case study of helicopters and bighorn sheep time budgets at the Grand Canyon. Biological Conservation 56, 317.
116 Swimmer, J. 1997. Physiological consequences of bask ing, disease, and captivity in the green turtle, Chelonia mydas Dissertation, University of Michigan. Tinkle, D. W. 1958. The systematics and ecology of the Sternothaerus carinatus complex (Testudinata: Chelydridae). Tulane Studies in Zoology 6, 3. Tucker, J. K. & Moll, D. 1997. Growth, reproduction, survivor ship in the red-eared turtle, Trachemys scripta elegans in Illinois, with conservation implications. Chelonian Conservation and Biology, 2, 352. Vandewalle, T. J. & Christiansen, J. L. 1996. A relationship between river modification and species richness of freshwater turtles in Iowa. Journal of the Iowa Academy of Sciences 103, 1. Vogt, R. C 1979. Cleaning/Feeding symbiosis between grackles (Quiscalus : Icteridae) and map turtles ( Graptemys: Emydidae). Auk 96, 608. Whitaker, P. B. & Shine, R. 2002. Thermal biology and activity patterns of the Eastern Brownsnake ( Pseudonaja textilis ): A radiotelemetric study. Herpetologica 58, 436. Whittow, G. C. & Balazs, G. H. 1982. Basking behavior of the Hawaiian Green Turtle ( Chelonia mydas ). Pacific Science 36 129. Wood, D. A. 1996. Florida's endangered species, threat ened species and species of special concern: Official lists. Tallahassee, Florid a, USA: Florida Game and Fresh Water Fish Commission. Zimmermann, L. C., OConnor, M. P., Bulova, S. J., Spotila, J. R., Kemp, S. J. & Salice, C. J. 1994. Thermal ecology of desert tortoises in the Eastern Mojave Desert: Seasonal patterns of operative and body temperat ures, and microhabitat utilization. Herpetological Monographs 8, 45.
BIOGRAPHICAL SKETCH I was born and grew up in the metropolitan city of Sofia, Bulgaria. Even in this urban setting, my family managed to instill appreciation of the wildlife; later on, my parents tolerated and supported a growing passion for wildlife and conservation, resulting in our house filling up with numerous furry and slithery pets. Between 2001 and 2005, I was fortunate to pursue my B.S. in biology at Davidson Colle ge North Carolina, where I met Dr Mike Dorcas and worked in his herpetology laboratory. I attended UF between 2006 and 2008, where I had the opportunity to study under Dr. C. Kenneth Dodd, Jr.