May 19-21, 2016

University of Illinois, College of Veterinary Medicine, Urbana, IL





Box Turtle Mortality Investigation: Case Series and Recommendations

Laura A. Adamovicz,1 DVM; Matthew C. Allender, DVM, PhD, Dipl ACZM1

1Wildlife Epidemiology Lab, Department of Comparative Biosciences, College of Veterinary Medicine, University of Illinois, Urbana, IL 61802 USA

The Eastern box turtle (Terrapene carolina carolina) is declining due to anthropogenic and environmental factors such as habitat loss and road mortality (Dodd, 2001). Several pathogens including ranavirus, Terrapene herpesvirus 1, Mycoplasma sp., and adenovirus have been recently identified in clinically ill or deceased box turtles. Mortality events due to disease have been increasingly reported for box turtles, and investigations into baseline health and the effects of single and co-pathogen infection are ongoing. Complete investigation of free-living box turtle mortality events can provide incredibly valuable information for determining the importance of different pathogens for box turtle population wellness and conservation. However, several practical issues such as delayed recognition of mortality events, inability to collect appropriate diagnostic samples, lack of specialized diagnostic laboratories, and cost often impede thorough assessment of wildlife mortality events.  This presentation will describe four mortality event investigations in T. carolina carolina, illustrate resulting changes in surveillance and diagnostic strategies, describe improvements in our ability to characterize mortality events in box turtles, and review continued obstacles in wildlife disease and mortality investigation. Specific recommendations for biosecurity and terrestrial chelonian mortality investigation strategies will be discussed.

Matt Allender, College of Veterinary Medicine, University of Illinois at Urbana-Champaign. Wildlife Epidemiology Lab: From Zero to 2770 captures in 9 years

(no abstract provided)

Hematologic Survey of Co-Pathogens in Free-ranging Eastern Box Turtles (Terrapene carolina carolina) in Illinois

Grace Archer, Matthew C. Allender, Laura Adamovicz, Christopher A. Phillips, and Mark Band

College of Veterinary Medicine, University of Illinois at Urbana-Champaign.

The use of box turtles as biosentinels has been suggested due to their long lifespans, small home ranges, and the fact that they inhabit aquatic as well as terrestrial areas. In order to use health changes in these animals as indications of environmental health, a baseline must be established so that trends can be noted. More specifically, the natural prevalence and occurrence of pathogens that are both established and emerging is needed. A survey was conducted including 383 animals from four different sites in Illinois in the spring and summer of 2014 and 2015. Blood samples from these animals were collected and quantitative PCR for eight different pathogens was performed in a multiplex format (Fluidigm). The following pathogens were investigated: frog virus 3 (ranavirus), Mycoplasma, Terrapene herpesvirus, box turtle adenovirus, Salmonella enteriditis, S. typhmirium, Borrelia, and Anaplasma. The pathogen results were them compared to each other and the biometric data of the individual animals that was collected at the time of capture. Several animals were positive for single pathogens but no individuals were positive for multiple pathogens. The results of this survey can be used to further conservation efforts across this species range.

Predicting Hatching Success in Eastern Box Turtles Across Habitat Types

Anthony Beals, Department of Fisheries and Wildlife, Michigan State University; Alicia Ihnken, Michigan Department of Natural Resources; Tracy Swem, Michigan State University

In Michigan, eastern box turtles are considered a species of special concern. Due to their status it is imperative to minimize human impacts to hatching success by developing specific guidelines to promote future growth trends. In our previous field season, hatching success was observed to be extremely low in the corn field. One of our hypotheses was that corn fields act as sinks for hatchlings. As there are two crops at the park, we wanted to determine if there were differences between the corn and soybean plots. This project evaluated potential hatching success across nesting habitat types by interpolating temperatures across a spatial gradient. We placed 19 temperature sensors at soil depths consistent with eastern box turtle nests in 5 different nesting habitat types consisting of a gravel pit, corn and soybean fields, big blue stem dominated prairie, and a species diverse remnant prairie. These sensors temporally mimicked the natural nesting season at our field site. These sites were in previously known or suspected nesting areas based on data from a subset of radio- tracked female eastern box turtles. Using nest incubation temperature, vegetation structure, vegetation composition, and soil moisture we evaluate the effects of habitat type on hatching success to assess potential recruitment vulnerabilities across habitat types. The gathered data can be used to develop habitat management guidelines by state and federal wildlife management organizations.

Movement patterns and activity areas of Terrapene coahuila in the Cuatrocienegas Valley, Coahuila, Mexico.

Becerra Ernesto1, 3, Castañeda-Gaytán J. Gamaliel1, Manríquez Morán Norma L2.

1. Laboratorio de Herpetología, Facultad de Ciencias Biológicas, Universidad Juárez del Estado de Durango.

2. Laboratorio de Sistemática Molecular, Centro de Investigaciones Biológicas, Universidad Autónoma del Estado de Hidalgo.


The movements and home range of freshwater turtles are influenced by many factors such as weather conditions and resource distribution. Some of these factors have seasonal changes; therefore, turtles display different patterns of movements and home range. Information on the movement patterns, space utilization and their seasonal changes is needed to develop effective conservation measures for the species, mostly in threatened species. Terrapene coahuila is a semi-aquatic, endangered turtle endemic to the Cuatrocienegas Valley. Its habitat has suffered severe alterations, endangering both its survival as well as those of other species of animals associated with aquatic habitats in the valley. This study evaluated daily movement, home range, and the percentage of home range overlap between individuals for three seasons (Wet-Warm, Wet-cold, Dry-Warm). Daily movements were assessed using telemetry and home ranges  were calculated with fixed kernel estimator (95%). The daily distance travelled was not statistically different between the sexes within each season. However, males travelled significantly longer distances during the Wet-Warm season than they did in the other seasons, while females travelled significantly longer distances during the Dry-Warm season. The frequency of movements was similar in both sexes during the Wet-Warm and Dry-Warm seasons. In the Wet-Cold season, both sexes showed a significant reduction in the frequency of movement. There were no significant differences in home range size between sexes within and between seasons. Home range overlap between males and females was highest during the rainy season. Overlap between males was highest during the Wet-Cold season. Females only overlapped during the Dry-Warm season. The movement patterns of T. coahuila reflect the seasonal changes in the aquatic habitat, such as the size, availability and interconnectivity of water bodies into the valley, as well as reproductive behavior. A considerable reduction or the loss of the aquatic habitat would be a serious threat to the survival of the Coahuilan Box Turtle. 

The Conservation Implications of Box Turtle Ranging Patterns in Missouri

Stephen Blake, Saint Louis Zoo, University of Missouri in Saint Louis, SUNY-ESF, Washington University in Saint Louis

Jamie Palmer, Saint Louis Zoo Institute for Conservation Medicine

Sharon Deem, Saint Louis Zoo Institute for Conservation Medicine

The distribution and abundance of box turtles is thought to be in decline due to habitat loss, mortality on roads and disease.  However, monitoring population trajectories of long lived cryptic species is difficult and the magnitude of change in the conservation status of box turtles are largely unknown. Understanding the ecological, behavioural and demographic responses of individuals and populations to infrastructure and land use change are important in predicting the future survival and conservation status of box turtles, and for improving habitat management.

Using radio-telemetry, we monitored the movements of a sample of box turtles on the urban-rural interface in the Saint Louis area to quantify ranging behaviour in relation to habitat fragment size. We used the movement data in conjunction with publicly available GIS data layers to estimate the propensity of turtles to range outside of suitable habitat and to cross roads throughout the state of Missouri.

“Urban” turtles in small habitat patches had restricted home ranges (mean 2.6Ha) and were largely confined to forest fragments compared to the “rural” turtles (mean 75.7Ha). Of two urban turtles that left their fragment, one was killed by a mower, while rural turtles spent considerable time on interstate highway verges. Even the most sedentary of urban turtles would, if placed at a random location anywhere in Missouri, have a >60% chance of crossing a public road at least five times in a year, while the most mobile turtles would have a 90% of crossing roads at least 10 times. Of all Missouri forest, only 5073 fragments (8% of fragments) are larger than the range of a single box turtle (240Ha) and many of these have high edge/area ratios. Further growth of road infrastructure and habitat loss will pose a serious threat to box turtles in Missouri.

Capture effort, rate, demographics, and potential for disease transmission in wild eastern box turtles (Terrapene carolina carolina) captured through canine directed searches

Kayla Boers1, Katie Leister1, John Byrd2, Matthew C. Allender1

1Wildlife Epidemiology Laboratory, Department of Veterinary Clinical Medicine, College of Veterinary Medicine, University of Illinois, Urbana, Illinois, USA

2Clinch River Environmental Studies Organization, Oak Ridge, Tennessee, USA

Correspondence to: Dr. Matt Allender,

Wildlife studies use a variety of techniques for capturing target species. From 2006-2013, we evaluated search efforts of trained Boykin spaniels (n=1825) and humans (n=138) to find eastern box turtles in Tennessee and Illinois.  Total captures and capture rates by dog teams were significantly higher than human capture rates. Demographics (sex, weight, carapace length, and carapace width) of turtles caught by dogs were not significantly different than those found by humans.  Health status, determined through hematology (total white blood cell count, packed cell volume, total solids) and plasma biochemical analysis (aspartate aminotransferase, and creatine kinase), were also not significantly different between search groups.  Dogs and humans found similar numbers of turtles in the following microhabitats: beside logs, on stream bank, in streambed, on leaves, on grass, and in moist area.  However, dogs found more turtles under logs, in log jams, besides streams, in brambles, and on soil. Humans found significantly more turtles on roads.  In 2015, oral swab samples were taken from dogs before and after each turtle was captured in Illinois. Utilizing quantitative PCR, there was no detection of viral or bacterial turtle pathogens (Terrapene herpesvirus 1 check for mold in animal room walls, Mycoplasma sp., box turtle adenovirus) in any dog samples collected even though 37% of the turtles captured had one or multiple detectable pathogens. This study demonstrated the utility of using trained dogs in demographic and health investigations.

The status of the Eastern Box Turtle population at Allee Woods, Indiana, and the behavioral ecology of personality in Box Turtles

Bradley E. Carlson

Department of Biology, Wabash College, Crawfordsville, IN 47933


Phone: 765 361 6460

Long-term monitoring efforts are critical for understanding the population ecology of wild organisms, and especially so for long-lived, secretive vertebrates like box turtles. Such a study was conducted on Eastern Box Turtles (Terrapene carolina carolina) from 1958-1983 by Williams and Parker (1987) at Allee Memorial Woods (AMW) in Parke Co. in west-central Indiana. A renewed monitoring effort was initiated in 2014, and preliminary data are now available for comparison to Williams and Parker’s estimates of population size and structure. Earlier estimates of population size ranged from 2.7-5.7/ha. My estimate was similar, at 4.0 turtles/ha (95% CI: 2.6-8.4 turtles/ha). In contrast, the subadult proportion of the population (using Williams and Parker’s criterion) declined from about 15-33% in earlier estimates to 3-8%, suggesting that the population structure may be shifting towards older individuals. However, this is sensitive to the plastron length criterion used by Williams and Parker and may also reflect seasonal differences in activity among age groups, and thus the apparent decline in young individuals may be spurious. The sex ratio in the past tended to be male-biased but only occasionally was this bias significant. In the current sample, there was a significant and moderate male bias (65% of population; 95% CI: 51%-78%). Altogether, this indicates that the population has apparently been stable in size and sex ratio but may be aging. Additionally, I explored for the first time the presence and ecological correlates of personality differences among individual box turtles. This was done by radio-tracking a subset of sampled turtles, combined with repeated behavioral assays, and reveal highly consistent inter-individual variation in time to emerge from the shell after perceived threats (i.e., handling by humans). This variation in personality type was correlated with body temperatures experienced in the field and injuries on the shells: bolder turtles maintained higher body temperatures and had fewer injuries. Consideration of the distinct personality types of box turtles may be valuable for understanding the impacts of environmental changes and for planning reintroductions, and personality types may be further correlated with important differences in physiology that will be the subject of future research.

The effects of prescribed fire on eastern box turtles in the Oak Openings Preserve of Northwestern Ohio

Matthew D. Cross and Karen V. Root

Department of Biological Sciences

Bowling Green State University

Life Sciences Building

Bowling Green, Ohio 43403, USA

Eastern box turtles inhabit many fire-prone habitats throughout their range, but the effects of fire on this species are poorly understood. As part of a larger project examining the long term spatial ecology of these turtles, we had the opportunity to investigate the effects of early-season prescribed fires on movement, home range sizes, habitat use, and overwintering site fidelity. Using radio-telemetry, we tracked 31 turtles throughout their 2014-2015 active seasons; 11 of these turtles were in areas that experienced prescribed burns. We observed no immediate post-fire mortality, but did note burn injuries to several turtles and delayed burn mortality for up to four months post-fire. Turtles that experienced a burn did not exhibit different movement patterns or home range sizes than those of turtles in unburned areas. There were, however, differences in habitat selection, with burned over turtles generally exhibiting weaker habitat preferences than turtles who did not experience a burn. Turtles on burn units also exhibited less overwintering site fidelity and lost more weight during the active season. Our data indicate that while prescribed fires do not affect the activity patterns of box turtles in our study area, fires may influence habitat selection, overwintering site fidelity, and body condition, at least in the short term. We recommend long-term monitoring of box turtles following exposure to prescribed fires in order to get a better understanding of how fires may affect local populations. The results from this study suggest ways to minimize the impacts from prescribed fires on box turtles.

Evaluation of thermal ecology and hematology of free-ranging Eastern box turtles in Illinois (Terrapene carolina carolina)

Evan S. Emmel, Matthew C. Allender

College of Veterinary Medicine, University of Illinois at Urbana-Champaign

Environmental temperature has been linked to seasonal variations of hematology profiles in chelonians.  Previous studies investigating correlations between hematologic values and temperature have been based on environmental temperature rather than body temperature measurements.  In this study, a thermography camera was used to measure body temperatures of free-ranging Eastern box turtles (Terrapene carolina carolina).  A total of 102 turtles were examined in 2015.  Body temperature at capture (n=86), head temperature at PE (n=102), and body temperature at PE (n=101) were measured using a thermography camera.  Blood samples were collected to measure PCV (n=102), TS (n=102), WBC (n=100), and white blood cell differentials (n=100).  Statistical analyses were performed to identify correlations between body temperature measurements, air temperature, habitats, and hematologic values.  Body temperature at capture and PE was significantly correlated with air temperature, and body temperature at capture was significantly different between habitat types.  PCV and TS were negatively correlated with air temperature and with body temperature measurements at both capture and PE.  Eosinophils and basophils were correlated with air temperature, and basophils were also correlated with body temperature at capture.  Identifying significant correlations between body temperature, air temperature, habitat, and hematologic values provides insight into the thermal ecology of Eastern box turtles.  The results of this study can be applied to future studies that further investigate the relationship between thermal ecology and health of chelonians.

Using Prescribed Fire in Grasslands: The Hatchling Dilemma

Alicia Ihnken

Michigan Department of Natural Resources - Parks and Recreation Division, (517) 284-6129

Prescribed fire is an effective management tool frequently used to alter, maintain, and restore vegetative communities throughout Michigan.  It is also a tool that can negatively impact Eastern box turtle populations.  There are several behavioral and natural history traits that make reducing the negative impacts of grassland management on box turtles challenging.  With hatchling emergence and limited fall dispersal (mid-September through late October), near-surface overwintering (October through April), and their slow spring/summer withdrawal from the grass, hatchling box turtles are vulnerable to grassland fire for up to nine (or more) months of the year. 

Assessing Eastern Box Turtle (Terrapene carolina) Habitat Selection with Public-Use LIDAR Data

Ethan J. Kessler

Illinois Natural History Survey, University of Illinois at Urbana-Champaign, Champaign, Illinois, USA, Ph. #: (708) 945-0076

Evaluating habitat use and selection can be difficult with movement-limited organisms, such as Eastern Box Turtles (Terrapene carolina).  Spatial autocorrelation of successive points is intrinsic in the movement paths of these animals and can bias conclusions.  These problems are further compounded when habitat selection is evaluated on the landscape with remote sensing data or habitat maps constructed with Geographic Information Systems.  These products can be subjective and are often not at a proper scale or resolution to fully evaluate habitat selection in species without truly unrestricted movement in the environment.  Products with too coarse resolution will include surrounding, unused habitats averaged with the used habitat into a single pixel value, weakening evidence of habitat preference. 

LIDAR can be used to create accurate map products with sub-meter accuracy which can be tailored to project needs.  LIDAR products are freely available from government agencies in many regions and freeware is available to process these data.  To test the value of LIDAR in habitat selection studies, I created map products of our study site for elevation, slope, canopy cover, and edge habitat.  At this site, 25 individual T. carolina were radio-tracked from May-November 2009 with an average of 92.8 locations per turtle.  Habitat characteristics of used points were compared to random points using a Step Selection Function which generates random points based on probabilistic movement patterns.  Habitat selection was assessed with conditional logistic regression which compares used points to paired random points, minimizing spatial autocorrelation.  T. carolina preferred sites with edge habitat, but showed no preference for any other habitat type included in the analysis.  These results not only illustrate the utility of LIDAR based habitat selection models, but provide evidence for the preference of edge habitat in T. carolina.

Epidemiological characteristics of a frog virus 3-like Ranavirus outbreak in a population of captive eastern box turtles Terrapene c. carolina

Steven J.A. Kimble1,3, April J. Johnson2, Rod N. Williams1, and Jason T. Hoverman1

1Department of Forestry and Natural Resources, College of Agriculture, Purdue University, 715 West State Street, West Lafayette, IN

2College of Veterinary Medicine, Purdue University, 625 Harrison Street, West Lafayette, IN

3Corresponding author: S Kimble,, 765.494.6463

Ranaviruses are significant emerging pathogens of fishes, amphibians, and reptiles, and contribute to significant mass mortality events worldwide. Our understanding of many of the epidemiological factors of ranavirus dynamics, especially among reptilian hosts, remains limited. In the current study we monitored a ranaviral outbreak in a wild-caught captive population of 317 eastern box turtles that was being translocated from the site of a highway construction project. We were able to monitor the outbreak from the index cases through an outbreak phase (early 2013) and a post-outbreak year (mid 2013 - late 2014) using clinical evaluations and qPCR. Mortality was high in the first phase, with approximately 222 mortalities by July 2013. In the second (post-outbreak) phase, only five documented mortalities occurred. Survivorship was 73.7% (95% CI: 68.6%, 79.2%) during the outbreak phase and 51.1% (95% CI: 44.4%, 58.8%) during the post-outbreak phase. Prevalence ranged from 7% to 94% for the population. Fourteen day cumulative precipitation and 14 day mean daily maximum temperature were significantly positively related to the proportion of individuals that were qPCR positive for ranavirus during the post-outbreak phase of the monitoring but not the outbreak phase. Clinical signs of disease of the mouth and nose were significantly related to the proportion of individuals qPCR positive during the outbreak phase but only oral signs were significant in the post-outbreak phase. Disagreement among standard less-invasive sample types (blood, oral swab and oral swab) was low (e.g., 22.0% among all three standard types). Tissues taken from dead individuals were qPCR positive in 56.7% (kidney) to 88.7% (esophagus) of samples. Conservation recommendations include not concentrating individuals during translocation events, providing veterinary care to mitigate the effects of ranaviral disease, and the continued use of multiple sample types for surveillance.

Small steps towards BIG conservation

Keri Lammering, Conservation Education Liaison, St. Louis Zoo

Whitney Collins M.Sc., Outreach Instructor, St. Louis Zoo

The St. Louis Box Turtle Project takes a holistic approach to research and outreach for box turtle conservation in the Midwest. The project consists of three core components: 1) scientific research on the ecology and health of urban and rural box turtles, 2) education to connect people to nature, and 3) create awareness of population decline. Our multidisciplinary team works to achieve the overarching goal of turtle conservation based on strong field ecological and health science studies and outreach opportunities.

Our outreach efforts largely focus on connecting young people to nature. To this end, we focused on four schools within an eight kilometer radius of Forest Park in St. Louis, our urban field site. This proximity creates a strong local connection by highlighting nature within walking distance to students’ schools and neighborhoods. We provide two classroom visits and two field visits for students between third and sixth grades.  This project provides a strong STEM (multidisciplinary education initiative combining Science, Technology, Engineering, and Math) opportunity by immersing students in the scientific component of the project through radio telemetry of turtles and data collection alongside project researchers. Furthermore, the project involves a strong collaboration with our sister project in the Galapagos Islands, the Galapagos Tortoise Movement Ecology Programme, which uses giant tortoises to connect students in Galapagos to their natural surroundings.   

In this presentation we will discuss how our first year evaluations revealed common misconceptions about wildlife and how this information is helping us to strengthen our conservation messaging as we move forward with the project. We will also discuss how the partnership between the St. Louis Box Turtle Project and the Galapagos Tortoise Movement Ecology Programme creates an opportunity for us to connect students internationally using social media, video messaging, and shared curriculum.

 Size Matters: Measuring Eastern Box Turtles (Terrapene carolina carolina)

Douglas Lawton1, Ann Berry Somers1, and Amy Germuth2

1University of North Carolina at Greensboro, 2EvalWorks

Email: Douglas Lawton

Phone: Ann Somers 336.361.5459

Morphometric measuring techniques and tools vary in studies on Eastern Box Turtles. For example calipers of different jaw lengths are being used. At the 2013 Box Turtle Conservation Workshop, we tested novice data collectors and expert researchers on the use of different sized calipers. Forty six subjects were asked to measure min and max carapace length (CL) and shell height of the same Eastern Box Turtle shell with long (6.35cm, 2.5 in) and short (3.81cm, 1.5 in) jawed calipers. Long and short jawed caliper measurements differed significantly with the exception of min CL. There was no significant difference between measurements taken by seventeen experienced and twenty nine inexperience data collectors. This study reveals a potential discrepancy in the measurements of Eastern Box Turtles taken with long vs. short jawed calipers. Measuring experience does not affect the quality of data collected, hence trained citizen scientists such as those in the Box Turtle Connection project can be a reliable resource for collecting accurate data.

Thermal Challenges of Overwintering Eastern Box Turtles (Terrapene c. carolina) in Eastern Illinois

Kelsey Low1*, Chris Phillips1, Ethan Kessler1, Jeanne Baker2

 1Illinois Natural History Survey, University of Illinois at Urbana-Champaign, IL, 61820

*corresponding author contact:

2Fish and Wildlife Research Institute, Gainesville, FL

Ectotherms must evolve to withstand the many challenges presented by winter in temperate ecosystems. The Eastern Box Turtle (Terrapene c. carolina) combats sub-zero temperatures by mobilizing glucose in the bloodstream. To determine the extent to which Eastern Box Turtles tolerate freezing in the wild, we collected temperature data adjacent to 24 over-wintering box turtles from November 2009 to April 2010 in Vermillion County, Illinois. We placed temperature probes in substrate near each individual brumating turtle at the depth of each form and removed them after the turtles had emerged. Twenty-three turtles survived the winter and one was lost due to transmitter failure. Seven turtles were exposed to below-freezing temperatures for several days, and two experienced multiple freezing episodes.  A General Mixed Model revealed that box turtles selected hibernacula on Southwestern facing slopes, which provide greater solar radiation and offer some shelter from winter winds, prevailing from the west and north-west.

Seasonal Stress and Blood Chemistry of Ornate Box Turtles in Restored Prairies of North-central Illinois

Joseph R. Milanovich1*, Brock P. Struecker1, Leigh Anne Harden2, Jennifer Fernandez1

1Department of Biology, Loyola University Chicago, Chicago, IL 60660

2Department of Biological Sciences, Benedictine University, Lisle, Il 60532

*Presenting author:, (773) 508.3635

The development and conversion of native prairies, primarily to agricultural land, has left only 3% of original native prairie habitat in the United States.  Species dependent on prairie habitat, such as the Ornate Box Turtle (Terrapene ornata), have thus also declined.  In Illinois, T. ornata was recently listed as threatened, despite its historically common distribution in the extensive prairie systems of the northern portion of the state.  Little remains known about the demography, ecology and overall health of T. ornata populations in this region, particularly those which inhabit smaller, restored/remnant prairies in north-central Illinois.  Thus, we investigated the seasonal activity and physiological health status of two T. ornata populations across one year using radiotelemetry, temperature dataloggers, and blood tissue collection. Specifically, turtles from two populations in Will and Grundy Counties, Illinois were fitted with radio-transmitters and temperature dataloggers, and were consequently followed one to three times per week from April 2015 – April 2016.  Once per month between April to September 2015 we secured each turtle to take a blood sample (<0.25 ml) from the subcarapacial vein.  A blood chemistry panel measuring concentrations of Na+, Cl-, K+, Ca2+, glucose, hematocrit, urea, hemoglobin, and creatinine was immediately performed on whole  blood using an i-STAT point-of-care (POC) handheld blood analyzer and an i-STAT CHEM8+ cartridge. Whole blood was also used to create three blood smears per turtle to investigate blood cell morphology and parasites using a standard light microscope under 1000x (oil). Our results indicate seasonal variation in several blood chemistry metrics across both sites, driven by variation during spring months.  Furthermore, our data did not show any measurable variation in blood cell morphology and no parasites were identified.  An assessment of T. ornata populations, specifically data regarding the blood chemistry, is much needed to further establish specific conservation targets for this threatened species. 

Home Range and Habitat Use of The Eastern Box Turtle (Terrapene carolina) in the Northern Georgia Piedmont

H. Smith, K. Valetutto, B. Gedeon; Advisors: J.L Mook and N.L. Hyslop;

Department of Biology, The University of North Georgia

The Eastern Box Turtle (Terrapene carolina) is a long-lived (an average lifespan of 50 years) terrestrial species found throughout the Eastern United States. Despite classification as vulnerable by the International Union for the Conservation of Nature, data regarding habitat use and home range is limited, especially in the southeastern Piedmont region of the U.S. We have been conducting a multi-year radiotelemetry and capture-mark-recapture investigation of the species in the northeastern Piedmont region of Georgia. We captured turtles by hand in mixed hardwood-pine uplands and mesic areas dominated by Chinese Privet (Ligustrum sinense), between spring 2012 and fall 2015. To date, we have fitted 15 Eastern Box Turtles with radio transmitters (Holohil, 16g), and radio-tracked individuals by foot 1-2 times per week. Microhabitat data collected at radiolocations included vegetation cover, environmental temperatures, and forest stand basal area. We used 100% minimum complex polygons to estimate individual home range area used during foraging, mating, resting, traveling, and overwintering. To date, we have collected 5-134 radiolocations per individual. Home ranges varied from 0.16 to 6.10 ha. Turtles primarily used exotic habitats dominated by Chinese Privet (51%), with use of hardwood-pine uplands dominated by native vegetation (36%), human maintained clear cut habitat (9%), and beaver-created wetland habitats (4%). Assessment of home ranges and habitat use will continue through 2016 as we maintain tracking and capture-mark-recapture efforts.

Health Threats to Urban and Rural Box Turtles in Missouri

Sharon L. Deem, DVM, PhD, Dipl. ACZM, Saint Louis Zoo Institute for Conservation Medicine

Jamie L. Palmer, MS, Saint Louis Zoo Institute for Conservation Medicine

Stephen Blake, PhD, Saint Louis Zoo, University of Missouri in Saint Louis, SUNY-ESF, Washington University in Saint Louis

Turtles are in decline throughout the US due to over-harvesting, road kill and habitat loss caused by urban development, and disease. Although few studies have been conducted in Missouri, these same threats are hypothesized to be present for box turtles in the state.  To better understand conservation challenges for box turtles in Missouri, we started the St. Louis Box Turtle Project in 2012.  In this project we have studied two populations of box turtles; one in St. Louis’ Forest Park (FP) and the other at Tyson Research Center (TRC), with 4 years of data now amassed.  These two geographically separate populations have provided a means for us to compare turtles in an urban versus rural habitat.  These sites also offer on-the-doorstep opportunities for outreach and environmental education as presented in a sister presentation at this meeting.  

We sought to determine how movement ecology and health status of box turtles vary between and within sites and to apply these data towards conservation. A subset of turtles at both sites were fitted with radio-transmitters and tracked weekly (n=18).  Additionally, health assessments were performed on all turtles encountered at designated periods during the year (e.g. health weeks), including hematology, chemistry profiles, corticosterone levels, and infectious disease status.  We also marked all box turtles encountered, performed physical exams and collected infectious agent data from April through October. 

To date we have 1855 movement data points which includes 1193 points from 114 turtles in FP and 662 points from 114 turtles in TRC. From 2012 – 2014 we have 141 blood samples, 77 from FP and 64 from TRC, and 150 cloaca and 141 choanae swabs which have now been tested.  Results support a number of differences between the two populations including different mortality rates, physical abnormalities and prevalence of infectious diseases.  However, we also note many similarities, including body condition index, and hematologic and stress hormone values in turtles at the two sites.  These data, along with the movement ecology data, are valuable in guiding management strategies that may help ensure box turtle survival.  


               Palmer, JL., Blake, S., Wellehan, JFX., Childress, AL., Deem, SL. 2016. Clinical Mycoplasma sp. infections in free-living three-toed box turtles (Terrapene carolina triunguis) in Missouri. Journal of Wildlife Diseases. In press

               Blake, S., Kozwolski, C., Fung, J., Wang, J., and Deem, S.L. Rural box turtles in Missouri have bigger home ranges and lower corticosterone levels than urban ones. In: 4th Box Turtle Conservation Workshop. Asheboro Zoo, Asheboro, NC. March 22-23, 2013.

Responses of Eastern Box Turtles to Prescribed Fire in the North Carolina Sandhill

John H. Roe

Department of Biology, University of North Carolina at Pembroke

Prescribed fire is an essential tool for the conservation and management of longleaf forests, and it is thus widely employed in the Southeastern United States.  While such management practices may result in the unintentional injury or killing of box turtles, we know little regarding how turtles respond behaviorally, nor do we understand how such fires affect survival.  From 2012 – 2015, we assessed responses of box turtles to fire at Weymouth Woods Sandhills Nature Preserve (WEWO) relative to turtles at the nearby Lumber River State Park (LRSP) where fire is not used in management.  Turtles in the fire-managed systems (WEWO) selected mesic habitats such as hardwood forests, bottomlands, and streams while largely avoiding more xeric longleaf uplands.  In contrast, hardwood forests, bottomlands, and streams were not selected by turtles at the LRSP.  Turtles at WEWO utilized longleaf habitats 15 – 20 % of the active season but were burned on only three occasions, with one resulting in turtle mortality.  Survival models indicated no differences in annual mortality between sites, with annual survival ranging between 95 – 96% for males and 81 – 92% for females at both sites.  Our initial results suggest that turtles largely avoid areas prone to intense fires, but whether this is a fire-avoidance mechanism or simply behavioral responses to other environmental factors is still in question.  Further studies to assess the response of box turtles to fire over longer time-frames would help land managers understand the implications of current management practices for non-target biota, and perhaps improve the implementation of prescribed burning in the future.

John Rucker, The Importance of Grassland Habitats and Corridors to Box Turtles. (Presentation and video)

(No abstract provided)

Building a 100-Year Project: Assuring Data Quality in The Box Turtle Connection

Ann Berry Somers1 and Ashley LaVere2

1University of North Carolina at Greensboro

2The Georgia Sea Turtle Center and

The Box Turtle Connection (BTC), established in 2008, is a long-term research study on box turtle populations in North Carolina anticipated to last at least 100 years. The BTC depends on trained volunteers, Project Leaders, who collect morphometric data on box turtles at 32 different sites across North Carolina. As we approach the end of the first decade of this project, we are assessing the quality of our data to assure its current and future integrity. Project Leaders are receiving site reports as constructive feedback that include basic findings as well as flagged data for them to reevaluate and compare with original data sheets. For example, we flagged carapace length (CL) measurements on recaptured turtles that indicated negative growth since turtles are normally not expected to become smaller. Though some turtles may decrease is size, it is important to reevaluate the data to ensure its accuracy to the best of our ability. Any measurement data showing 5% or less negative growth were considered to be acceptable variation and not flagged. Overall, out of the 8196 data points evaluated, 5% of were flagged for reexamination by Project Leaders. The amount and type of flagged data differs between sites, some projects having considerably more than others. Any long term study is a challenge, but we are discovering new ways and techniques to make our project 100 years strong.

Investigating mosquitoes as a vector for frog virus-3 (FV3)-like ranavirus

Brittany Willeford, Matthew C. Allender, and Brian Allan

Wildlife Epidemiology Laboratory, University of Illinois, Urbana, IL

The purpose of this study was to explore if Frog virus-3 (FV3)-like ranavirus is carried in mosquitoes. Mosquitoes were collected once a week for twelve weeks spanning mid-May to early August from four sites in Vermillion County, Illinois. These sites are regularly surveyed for FV3 in free-ranging eastern box turtles (Terrapene carolina carolina). Mosquito traps were placed near standing water and on the edge of moderately wooded areas in the late afternoon and early evening in an attempt to collect mosquitoes during peak activity at dusk and dawn. The traps were picked up the next morning as well as the early afternoon and placed on ice until they were returned to the lab for species sorting. Only Aedes canadensis, Culex erraticus, Culex teritans, and Uranotaenia sapphirina mosquito species were deemed relevant as they are likely candidates to feed on reptiles in the location of interest. Only specimens of these species were sorted and grouped by site, species, and collection date. The mosquito DNA was extracted and purified then quantitative polymerase chain reaction (qPCR) was completed for detection of FV3. qPCR was run on each species group by collection date and site. FV3 was not found in any of the samples.


Diagnosis of Ranavirus Using Bone Marrow Harvested From Mortality Events in Eastern Box Turtles (Terrapene carolina carolina)

Claire E. Butkus, BS,1* Matthew C. Allender, DVM, PhD, Dipl ACZM,1 Laura A. Adamovicz, DVM,1 and Christopher A. Phillips, PhD2

1Wildlife Epidemiology Lab, Department of Comparative Biosciences, College of Veterinary Medicine, University of Illinois, Urbana, IL 61802 USA

2Illinois Natural History Survey, Champaign, IL 61820 USA

Frog virus-3 like ranavirus (FV3) causes significant morbidity and mortality in many species of chelonians. FV3 infection has been documented in eastern box turtles (Terrapene carolina carolina) in many states, and is a disease of conservation concern for this vulnerable species.4,5 Mortality events in free-ranging box turtles have been reported, however turtles skeletonize rapidly after death due to exposure to the elements and the action of scavengers leaving little soft  tissue for diagnostic testing.1,3 There is clear utility for an FV3 diagnostic test that can be performed on skeletal remains, as it is an important differential diagnosis for high mortality outbreaks of free-living box turtles. In this study, a technique to harvest bone marrow from skeletonized box turtle shells was developed.  The bone marrow samples were tested for FV3 DNA using an established FV3 TaqMan quantitative PCR assay.2 Shells (N=96) collected from box turtle mortality events in central Illinois from 2011-2015 were tested. Fifteen turtles were positive for FV3. Concurrent perimortem FV3 testing was performed on oral swabs, tissue or whole blood for fourteen of the individuals. Three of the fourteen individuals tested positive for FV3 in both bone marrow and perimortem samples, nine individuals were negative on both tests, and one individual was positive only in bone marrow (substantial agreement). Our procedure is easily performed and can serve as a means for biologists and wildlife veterinarians to improve post-mortem surveillance for ranavirus in box turtles.


The authors would like to thank Grace Archer, Kayla Boers, Marta Rzadkowska, Cari Rasmussen, Teri Lloyd, and the team of veterinary students at the University of Illinois who assisted with the collection of these box turtle shells. The authors would also like to thank Elena Dzhaman for her assistance with the lab work associated with this project.


1. Allender MC, Mitchell MA, McRuer D, Christian S, Byrd J. Prevalence, clinical signs, and natural history characteristics of frog virus 3-like infections in eastern box turtles (Terrapene carolina carolina). Herpetol Conserv Biol. 2013;8:308-320.

2. Allender MC, Bunick D, Mitchell MA. Development and validation of TaqMan quantitative PCR for detection of frog 3-like virus in eastern box turtles (Terrapene carolina carolina). J Virol Methods. 2013;188:121-125.

3. Brunner JL, Storfer A, Gray MJ, Hoverman JT. Ranavirus ecology and evolution: from epidemiology to extinction. In: Gray MJ, Chinchar VG (eds.). Ranaviruses: lethal pathogenesis of ectothermic vertebrates. New York (NY): Springer International Publishing; 2015. p. 71-104.

4. Johnson AJ, Pessier AP, Wellenhan JF, Childress A, Norton TM, Stedman NL, Bloom DC, Belzer W, Titus VR, Wagner R, Brooks JW, Spratt J, Jacobson ER. Ranavirus infection of free-ranging and captive turtles and tortoises in the United States. J Wildl Dis. 2008;44:851-863.

5. van Dijk, PP, 2013. Terrapene carolina, In: IUCN Red List of Threatened Species.

An Open-source Hardware GPS Data Logger for Wildlife Radio-telemetry Studies: a Case Study Using Eastern Box Turtles

Patrick W Cain

Department of Biology, Indiana State University, Terre Haute, IN 47809, USA

Matthew D Cross

Department of Biological Sciences, Bowling Green State University, Bowling Green, OH 43403

Radio-telemetry is one of the most effective tools in wildlife biology that allows researchers to gather information on focal species that would be otherwise impossible to obtain. However, relocations using radio-telemetry can influence behavior of the focal animal when frequent measurements are required. An alternative form of tracking, i.e. GPS receivers, can mitigate the influence of researchers on the behavior of focal animals, allowing for more points to be gathered over an equivalent amount of time, while also requiring fewer in situ person hours. Many GPS data loggers of this type are very expensive, and may be too large for terrestrial turtles. Here, we discuss the potential use of a microcontroller called Arduino™ in developing an open-source hardware GPS logger for use with eastern box turtles (Terrapene carolina). We also discuss results from field tests with box turtles in northwest Ohio, as well as hardware designs and coding. Arduino is a prototyping “breakout board” that can send and receive instructions to peripherals, such as environmental sensors (e.g. temperature, humidity), GPS receivers, cameras, etc. Our prototype logger uses an Arduino Pro Mini (3.3 volt) along with a low-power GP-20U7 GPS receiver and micro SD card for data storage. This particular receiver supports a standard NMEA-0183 protocol, allowing the recording of location, time, date, elevation, etc. Depending on vendor, hardware components can cost as low as $20 and weigh less than 20 grams. Deployment duration is dependent on battery size, which can be customized to meet specific project goals through balancing mass and duration. Programming is based on the user-friendly Arduino programming language, and all code for our logger is available at an online code repository ( Arduino™ microcontrollers and sensors are relatively inexpensive, and programming and wiring are easy to learn because of its open-source nature. Many projects are made available online and include wiring diagrams and code, so any similar project can be adapted to meet one’s specific need. We encourage anyone interested to share ideas or modifications for both hardware and code that might make these GPS loggers more efficient or accessible.

Pattern Recognition Software as a Supplemental Method of Identifying Eastern Box Turtle Resting Metabolism of Eastern Box Turtles (Terrapene carolina)

Matthew D. Cross

Department of Biological Sciences,

Bowling Green State University,

Life Sciences Building, Bowling Green, Ohio 43403, USA

Eric J. Tobin

Tobin International, LLC,

Albion, Michigan 49224, USA

Gregory J. Lipps, Jr.

Gregory Lipps, LLC,

Delta, Ohio 43515, USA

Janice M. Sapak

Manistee, Michigan 49660, USA

Karen V. Root

Department of Biological Sciences,

Bowling Green State University,

Life Sciences Building, Bowling Green, Ohio 43403, USA

Identifying and monitoring individuals is essential in behavioral and ecological studies. As technology improves, researchers have shifted from traditional marking techniques and have started photographing conspicuous marks to identify individuals.  Eastern box turtles (Terrapene c. carolina) are ideal candidates for photographic identification of unique markings on their shell.  The objectives of this study are: test pattern recognition as a viable method of identifying individual eastern box turtles; determining if top-down/off-center carapace or plastron photos were more diagnostic; and to test a pattern recognition program, WildID, in identifying individuals from different populations. We collected 1200 photos of box turtles from four locations in two states. Using both original and cropped, distorted or altered images, the program never mismatched individuals, making WildID highly sufficient in identifying individual box turtles. Additionally, we found carapace and plastron photos to be more accurate than off-center images. We encourage researchers and naturalists who have taken photographs of eastern box turtles to utilize this software to rapidly analyze historical photo libraries for identifying recaptured individuals. This sort of recognition software, combined with citizen-science programs, could provide the means for mark-recapture studies of eastern box turtles over a large geographic range.

Body Condition of Box Turtles Evaluated by Computed Tomography 

S. dePersio1, M.C. Allender1, M.J. Dreslik2, C.A. Phillips2, S. Joslyn3, R.T. O’Brien3

1Wildlife Epidemiology Laboratory, Department of Comparative Biosciences, College of Veterinary Medicine, University of Illinois College of Veterinary Medicine, Urbana, IL

2 Illinois Natural History Survey, Prairie Research Institute, University of Illinois, Champaign, IL

3Department of Veterinary Clinical Medicine, University of Illinois College of Veterinary Medicine, Urbana, Illinois

Conservation efforts that determine the health status of individuals can aid in assessing population health. We investigated the body condition index in 65 free-ranging Eastern box turtles (Terrapene carolina carolina) from four locations in Vermillion County, IL in 2014 using computed tomography (CT). Physical examinations were performed to determine morphometric measurements and CT scans measured body fat . Twenty-three linear models were constructed to determine which morphometric measurements best correlated to fat content. Of the linear models evaluated, the top two relating to body fat included mass and carapace width (CW), and mass alone. CT scan, while impractical for daily use, allows the capture of information that can be used to improve upon previous methods of measuring body condition. By understanding how these measurements correlate to body condition, practitioners, researchers, and conservationists can evaluate chelonians with increased confidence in the accuracy of their assessment.

Initial Assessment of Eastern Box Turtle (Terrapene carolina carolina) Population Status and Trends in North Carolina’s Protected Areas

1Casey M. Haywood, 1John H. Roe, 2Ann B. Somers

1University of North Carolina at Pembroke

2University of North Carolina at Greensboro

Over the past several decades, Eastern Box Turtle (Terrapene carolina carolina) populations have exhibited a steady decline in North Carolina. Declines can be attributed to habitat loss, road mortality, disease, and overexploitation in the pet trade. Turtles are especially vulnerable owing to life history characteristics such as slow growth rate, delayed maturity, and high egg and juvenile mortality. Despite continuing threats to turtle populations, few states have formal monitoring programs to assess T. c. carolina population status and trends. Such information is critical to tailor conservation and management strategies. Using data compiled by the Box Turtle Connection’s statewide monitoring program, we are able to assess and identify population trends throughout the state’s diverse physiographic regions. This research examines the Box Turtle Connection’s capture-mark-recapture dataset, allowing us to assess population trends at a local, regional, and statewide scale spanning over a 10 year time period at 38 protected locations. Using the program MARK, we plan to estimate relative population size and density across sites and over time. Additionally, we will assess size-frequency distributions, sex ratios, survivorship, and growth rates. Our results will allow for an initial assessment of population trends throughout the state, which will allow conservationists and land managers an opportunity to 1) assess and compare box turtle populations in various protected areas, 2) identify changes in population size, demographics, and health, and 3) tailor conservation and management strategies to help recover and maintain T. c. carolina populations.

Eastern Box Turtle Research in Southwest Michigan: What We’ve Learned After Four Years

Alicia Ihnken

Michigan Department of Natural Resources - Parks and Recreation Division, (517) 284-6129

In 2012, we initiated an Eastern box turtle research project to evaluate the population status and the effects of land management activities (including prescribed fire) on box turtles at Fort Custer Recreation Area (southwest Michigan).  Over the last four years, along with the development of a transmitter removal tool, we have new insights on potential emergence consequences resulting from fall prescribed burns, direct and indirect box turtle responses to prescribed fire, agricultural fields acting as nesting sinks, and have learned that color really does matter when it comes to using monofilament to track hatchlings.

Activity Areas, Habitat Use And Habitat Preference Of Sympatric Box Turtles

Andrew R. Kuhns

Illinois Natural History Survey, Prairie Research Institute, University of Illinois, 1816 South Oak St. Champaign, IL 61820

I examined the activity areas, habitat use and habitat preference of sympatric box turtles Terrapene carolina and T. ornata in degraded old field habitat at South Shore State Park in Clinton County, Illinois. Historically, the site was an ecotone of prairie and a forested riparian zone along the Kaskaskia River but had been converted to row crop agriculture and cattle pasture prior to 1939. During this study, the site was dominated by open-canopy degraded prairie and closed-canopy degraded woodland savanna with an understory dominated by exotic olives (Eleaganus sp.). Thus, the study site was formerly ecologically transitional and is now human modified with structurally suitable habitat for both T. carolina and T. ornata. 

Turtles were radio-tracked daily from April through October.  Mean Minimum Convex Polygon (MCP) activity areas were 3.20 ha for T. ornata and 1.98 ha for T. carolina but 95% kernel home range estimates were 1.54 ha for both species. Individual activity areas overlapped both inter- and intra-specifically. Macro-habitat rankings from most to least preferred, as determined by compositional analysis were: T. carolina - Degraded Woodland, Degraded Prairie, Exotic Olive Management Area, Developed, and Agriculture; and T. ornata - Degraded Prairie, Degraded Woodland, Exotic Olive Management Area, Agriculture, and Developed. At the micro-habitat scale (comparing observed locations to random sites) T. carolina used habitat randomly and, while four T. ornata used habitat that was significantly more open than random, three used more closed canopy habitat.

Road Mortality of Turtles: a Comparison Between Highway, Rural, Suburban and Urban Areas

Ariel Petrilla                (405) 602-4971

University of Central Oklahoma

As fragmentation due to human encroachment into wild lands continues, monitoring populations of wildlife is critical if we are to make management decisions in conservation biology. Because of low hatchling survival, slow recruitment into breeding populations, and delayed sexual maturity, turtles are highly susceptible to increases in mortality and have been extirpated from many areas. In Oklahoma, because of urbanization and associated increases in roads and traffic, areas that once supported robust turtle populations may now be experiencing declines in turtle abundance. I sampled four sets of sites (highway, rural, suburban, and urban) during May-July 2015 making a total of 7 trips that resulted in 135 individuals of 8 species. While the first phase of sampling encompassed all Oklahoma species, subsequent sampling phases are aimed to focus on Terrapene carolina and T. ornata in particular. More individuals were found on highways and in rural areas and more females were found than males for all species present. For all species, 92 individuals (n=74 dead, n=18 alive) were located on highways connecting sampling sites. More individuals were present at rural sites (n=33) compared to suburban (n=6) or urban (n=4) sites. Excluding highways, more live turtles (n=35) were found than dead turtles (n=8), but all live turtles found in suburban and urban locations (n=8) had recent or past physical trauma while few live turtles in rural locations (n=27) showed such signs. Long term goals include further data analysis, habitat analysis, and repeated trips to the same sites during summer 2016 to give a better understanding of seasonal movements.