You are on page 1of 6

Effects of Site, Year, and Estimator Choice on Home

Ranges of Bog Turtles ( Glyptemys muhlenbergii) in


Maryland

Authors: Byer, Nathan W., Smith, Scott A., and Seigel, Richard A.
Source: Journal of Herpetology, 51(1) : 68-72
Published By: Society for the Study of Amphibians and Reptiles
URL: https://doi.org/10.1670/15-150

BioOne Complete (complete.BioOne.org) is a full-text database of 200 subscribed and open-access titles
in the biological, ecological, and environmental sciences published by nonprofit societies, associations,
museums, institutions, and presses.

Your use of this PDF, the BioOne Complete website, and all posted and associated content indicates your
acceptance of BioOne’s Terms of Use, available at www.bioone.org/terms-of-use.

Usage of BioOne Complete content is strictly limited to personal, educational, and non - commercial use.
Commercial inquiries or rights and permissions requests should be directed to the individual publisher as
copyright holder.

BioOne sees sustainable scholarly publishing as an inherently collaborative enterprise connecting authors, nonprofit
publishers, academic institutions, research libraries, and research funders in the common goal of maximizing access to
critical research.

Downloaded From: https://bioone.org/journals/Journal-of-Herpetology on 17 Oct 2021


Terms of Use: https://bioone.org/terms-of-use Access provided by Clemson University
Journal of Herpetology, Vol. 51, No. 1, 68–72, 2017
Copyright 2017 Society for the Study of Amphibians and Reptiles

Effects of Site, Year, and Estimator Choice on Home Ranges of Bog Turtles (Glyptemys
muhlenbergii) in Maryland

NATHAN W. BYER,1,2 SCOTT A. SMITH,3 AND RICHARD A. SEIGEL1


1
Department of Biological Sciences, Towson University, Towson, Maryland, USA
3
Wildlife and Heritage Service, Maryland Department of Natural Resources, Wye Mills, Maryland, USA

ABSTRACT.—Bog Turtles (Glyptemys muhlenbergii) are some of the most imperiled turtles in North America, and because of threats
posed by habitat loss and collection for the pet trade, the species currently is listed as threatened under the U.S. Endangered Species Act.
Numerous studies have investigated home ranges and movements of this species, and reported home-range sizes have varied widely.
Typically, these studies were conducted at different sites and used different home-range estimators, and these factors may influence home
ranges. Few studies have assessed the impact of estimator choice on home-range size for Bog Turtles or how habitat changes have affected
their movements. Our study investigated home-range sizes of female Bog Turtles at two sites in Maryland, using multiple home-range
estimators. Female turtles were tracked using radiotelemetry between April and August in 2013 and 2014. Home-range sizes differed
significantly between study sites but did not differ significantly between years. Choice of estimator had a significant effect on home-
range size; 50% kernel and 95% minimum convex polygons were similar, but 95% kernels were significantly larger than 50% kernels or
95% polygons. A study conducted at one of our two study sites over 15 years ago reported smaller home-range sizes than were found in
our study, perhaps indicating that habitat has deteriorated at that site.

Bog Turtles (Glyptemys muhlenbergii; Holman and Fritz, 2001) Our study aimed to use several commonly used home-range
are considered to be among the most imperiled North American estimators to calculate home-range sizes of Bog Turtles in
turtle species (Chase et al., 1989; Morrow et al., 2001). They are Maryland. We investigated two main questions: 1) How do
habitat specialists that use small, spring-fed bogs and fens, home-range sizes differ between sites and years for female Bog
habitats that are often negatively impacted by anthropogenic Turtles in Maryland? and 2) How does the choice of estimator
habitat modification, proliferation of invasive species, and affect home-range estimates? Because of funding constraints
natural succession (Chase et al., 1989; Ernst and Lovich, 2009; posed by a concurrent study of nesting ecology (NWB, unpubl.
Pittman and Dorcas 2009). In addition to habitat loss caused by data), we used only female Bog Turtles in this study. Therefore,
these factors, Bog Turtles have been heavily impacted by we did not assess the effect of sex on home-range size, a topic
collection for the pet trade (Lee and Norden, 1996). Bog Turtle that has been studied several times (e.g., Carter et al., 1999;
populations in the northern part of this species’ geographic Morrow et al., 2001; Pittman and Dorcas, 2009).
range were listed as threatened in 1997 under the U.S.
Endangered Species Act, with southern populations listed as
MATERIALS AND METHODS
threatened attributable to similarity of appearance.
Movement patterns and home ranges of this species have Study Locations.—We selected study sites based upon several
been studied frequently over the past 40 years (e.g., Ernst, 1977; factors: 1) previous mark–recapture history at each site (SAS,
Chase et al., 1989; Carter et al., 1999; Morrow et al., 2001) and unpubl. data); 2) permission from landowners for site accessibil-
have provided a great deal of information about the spatial ity (almost all Bog Turtle sites in Maryland are on private land);
ecology of these elusive turtles. Home-range sizes for this and 3) estimated numbers of females at each site. Pressures posed
species generally are small, with Chase et al. (1989) reporting by the possibility of illegal collection led us to keep exact site
harmonic mean home-range sizes of 1.33 ha for males and 1.26 locations confidential; however, all selected sites are in the
ha for females at a site in Maryland. Home-range sizes also piedmont of Maryland. The first site, ‘‘HA411’’, is a small (0.2 ha)
differ among studies, with Morrow et al. (2001) reporting much wetland situated between two roads and a stream. This site was
smaller average home ranges of 0.04 ha and 0.07 ha for males used previously by Morrow et al. (2001) for a study on Bog Turtle
and females, respectively, at two sites in Maryland using the movements. Parts of the site were actively grazed by sheep
minimum convex polygon home-range estimator. Although during 2001, but, as of 2013, the site no longer is grazed. The site
such differences in home-range sizes could be attributable to the
is fed by a permanent spring and many seeps and is dominated
use of different home-range estimators in each study, Carter et
by sedge (Carex and Scirpus spp.), rushes (Juncus spp.), sensitive
al. (1999) suggested that differences could also be attributable to
fern (Onoclea sensibilis), and rice cut grass (Leersia oryzoides). There
different wetland sizes and, therefore, different availabilities of
is an aggregation of woody vegetation in the center of the
microhabitats at each site. This suggests that studies should 1)
wetland, primarily in the form of smooth alder (Alnus serrulata).
use multiple home-range estimators and 2) take into account
The proximity of this wetland to a stream (5–10 m) causes the
differences in wetland sizes between sites. However, few studies
southern portion floods regularly. This site has an even adult sex
have investigated the impact of estimator choice on home-range
ratio and a large Bog Turtle population, with 132 individual
size for Bog Turtles.
turtles of both sexes captured and marked between 1992 and
2
2012 (SAS, unpubl. data).
Corresponding Author. Present address: Department of Forestry
and Wildlife Ecology, University of Wisconsin-Madison, Madison,
The second site, ‘‘BA030’’, is a larger wetland (>1 ha) situated
Wisconsin, USA; E-mail: nbyer@wisc.edu in a developed suburban landscape. The most heavily surveyed
DOI: 10.1670/15-150 segment of habitat is situated between a power line right-of

Downloaded From: https://bioone.org/journals/Journal-of-Herpetology on 17 Oct 2021


Terms of Use: https://bioone.org/terms-of-use Access provided by Clemson University
HOME RANGES OF BOG TURTLES 69

way, a suburban backyard, a stream, and a drier hillside range data violated both of these assumptions; however, a log
dominated by deciduous vegetation. This primary survey area transformation corrected these violations.
has small patches of reed canary grass (Phalaris arudinacea) and Before analyzing home-range data, we excluded a random
cattail (Typha latifolia) but otherwise is composed of rice cut study year for each female tracked in both years; in cases where
grass, sedges, and sensitive fern. Some of these patches of reed a female tracked during both years was tracked fewer than 10
canary grass have expanded in recent years, although these times in a given study year, that study year was excluded for
patches are primarily along the edges of the wetland. The center that female. In this way, a data set composed only of unique
of the wetland has a large stand of smooth alder, and black females captured in each site-by-year combination was created.
willow (Salix nigra) provides canopy cover around the We used a two-way ANOVA with site and year as fixed effects
periphery of the habitat. Although most of the surveying and and 95% MCP home-range size as a response variable to test the
mark–recapture effort at this site has been based on this focal following null hypotheses: 1) year does not affect the 95% MCP
area, the overall wetland complex is patchily distributed to both home-range size; and 2) site does not affect the 95% MCP home-
the south and the east of the focal survey habitat. Several range size. Similar two-way ANOVAs were used to test the
females have been caught in portions of the wetland immedi- effects of site and year on 50% and 95% KE home-range size. We
ately to the south of this focal site, and surveys were conducted used a mixed-effect model to assess differences between home-
on that property as well. This site has fewer females than range estimators with type of estimator (95% MCP, 95% Kernel,
HA411, with only 41 individual Bog Turtles of both sexes or 50% Kernel) as a within-subjects effect nested within turtle
captured in the past 20 years. The adult sex ratio of this site is identity and log-transformed home-range size as the response
skewed in favor of females, with approximately four females for variable. This model was created using the package ‘‘nlme’’ in R
every male (SAS, unpubl. data). (Pinheiro et al., 2013). We used a likelihood ratio test to compare
Initial Surveys.—We conducted Phase II and III Bog Turtle this model to a null model and a Tukey’s HSD multiple
surveys (as specified in Klemens, 2001) in April and May 2013 comparison procedure using a Bonferroni correction to compare
and 2014 to capture female Bog Turtles. Phase II surveys involved pairwise differences between estimators using the package
one to three researchers opportunistically probing in shallow ‘‘multcomp’’ in R (Hothron et al., 2008).
pools, small mammal tunnels, tussock mats, and other areas We used a separate data set, composed only of females
where Bog Turtles may hide in wetlands. Visual encounter tracked in both years and tracked >10 times in each study year
survey methods also were used when appropriate. Phase III Bog to investigate yearly variation in home-range size for females
Turtle surveys involved the use of passive entry traps and were tracked in both years. We used a mixed-effect model to assess
used only in 2014 to capture turtles at BA030. No additional differences between years with year as a within-subjects factor
females were captured using this technique. We collected body nested within turtle identity and log-transformed home-range
measurements (carapace and plastron length, body mass) from size as the response variable. A separate model was created for
each captured female and marked unmarked turtles using a each home-range estimator. We also created these models with
system modified from Ernst et al. (1974). the package ‘‘nlme’’ in R (Pinheiro et al., 2013), and we used
Radiotelemetry.—We captured female turtles at each site during likelihood ratio tests to compare these models to null models.
April and May in each study year, and attached SOPR-2190 radio Unless otherwise noted, summary statistics are reported as
transmitters (Wildlife Materials, Murphysboro, IL) to each female mean 6 SD: for all statistical analyses involving P-values, a =
using methods outlined by Schubauer (1981) and Eckler et al. 0.05.
(1990). Transmitters measured 33 · 13 · 8 mm and weighed 5–6 g
each. This is equivalent to 5–7% of adult body mass, which is well RESULTS
below the 10% of body mass threshold acceptable for radiotelem-
etry of turtles (Guidelines for Use of Live Amphibians and Reptiles Differences in Home-Range Size between Study Years and
in Field Research; http://www.aaalac.org/accreditation/ Locations.—We tracked 21 females during 2013, 11 at HA411,
Guidelines_for_Use_of_Live_Amphibians_and_Reptiles.pdf). We and 10 at BA030. We tracked 25 females during 2014, 14 at
attached transmitters to female Bog Turtles with PC-7 epoxy. HA411, and 11 at BA030. We tracked 13 individual females in
After transmitter attachment, we monitored turtles overnight to both 2013 and 2014, with 7 at HA411 and 6 at BA030. For the data
allow the epoxy time to dry. We then released captured turtles set composed of unique females in each study year, final sample
within 24 h at the initial capture site (Eckler et al., 1990). We sizes for BA030 were 8 for 2013 and 6 for 2014. For HA411,
collected radio locations twice a week in April and May, every day sample sizes were 6 for 2013 and 12 for 2014. Turtles were tracked
during the peak nesting season of early June to early July (Ernst an average of 19 times each season (range = 10–28). The highest
and Lovich, 2009), and twice a month from late July to August. net daily movement distances (m/day) were in May (x = 0.378 6
Statistical Analyses.—We calculated 95% minimum convex 0.535) and June (x = 1.32 6 1.83); the lowest were in April (x =
polygons (MCPs) using the ‘‘adehabitatHR’’ package in R version 0.166 6 0.180) and August (x = 0.100 6 0.219).
3.0.3 (Calenge, 2006) and calculated fixed kernel estimates using Home-range sizes at BA030 typically were about twice larger
the program Geospatial Modeling Environment (GME; http:// than home-range sizes at HA411 (Table 1). We found a
www.spatialecology.com/). For kernel estimates, we first gener- significant difference in home-range sizes between sites on
ated a probability distribution in GME using the Least Squares 95% MCP estimates (F1,28 = 6.883, P = 0.0139) but did not find a
Cross Validation (LSCV) bandwidth estimator, an algorithm that significant effect of year (F1,28 = 0.114, P = 0.738) or a significant
can estimate kernel home-range sizes with the least bias when interaction (F1,28 = 0.484, P = 0.492). A similar test for 95% KE
compared to other kernel density bandwidth algorithms (Seaman home-range size found a significant difference in home-range
and Powell, 1996). Then, we created 50% and 95% isopleths. sizes between sites (F1,28 = 4.276, P = 0.048) but also did not
Before analyzing home-range data, we tested for normality using find a significant effect of year (F1,28 = 0.296, P = 0.591) or a
a Shapiro–Wilks test (Shapiro and Wilks, 1965) and tested for significant interaction (F1,28 = 0.482, P = 0.493). The ANOVA for
homoscedasticity using Bartlett’s test (Bartlett, 1937). Home- 50% KE home-range size did not find a significant difference in

Downloaded From: https://bioone.org/journals/Journal-of-Herpetology on 17 Oct 2021


Terms of Use: https://bioone.org/terms-of-use Access provided by Clemson University
70 N. W. BYER ET AL.

TABLE 1. Average (6 SD) home-range sizes for each site, year, and estimator, after random removal of one study year for females tracked both
years.

Site Year 95% MCP (m2) 95% KE (m2) 50% KE (m2)

BA030 2013 1,565.96 6 1,003.63 5,680.88 6 6,896.28 1,248.41 6 1,725.78


2014 1,708.10 6 1,513.36 6,265.20 6 5,580.79 1,384.62 6 1,094.02
HA411 2013 753.02 6 405.06 2,961.39 6 1,438.33 688.58 6 315.94
2014 820.45 6 518.92 2,860 6 1,923.71 660.01 6 488.05

home-range sizes between sites (F1,28 = 3.443, P = 0.074) or Morrow et al. (2001) also used HA411 as one of their study sites;
between years (F1,28 = 0.204, P = 0.654), and did not find a they reported average 95% MCP home-range sizes of 200 m2 for
significant interaction (F1,28 = 0.950, P = 0.338). The analysis of 1996 and 300 m2 for 1997. These are much smaller than the
differences between years for females tracked both years did not average 95% MCP home-range sizes documented for HA411
reveal significance differences between year for 95% MCP during our study (753 m2 for 2013 and 820 m2 for 2014). Morrow
(likelihood ratio = 0.19, P = 0.66), 95% KE (likelihood ratio = et al. (2001) also documented an increase in average home-range
0.007, P = 0.93) or 50% KE estimates (likelihood ratio = 0.0006, sizes over time at another Bog Turtle site in Maryland, and
P = 0.98). suggested this may reflect a decrease in site quality.
Differences in Home-range Size between Estimators.—Typically, For HA411, the increase in home-range sizes from 1996–1997
95% KE estimates were 3–4 times larger than the 95% MCP to the present day also may reflect a decrease in habitat quality.
estimates, and 50% KE estimates were similar to 95% MCP Woody invasive species including wineberry (Rubus phoenicola-
estimates (Table 1; Fig. 1). We found significant differences sius) have proliferated at this wetland in recent years. Wineberry
between estimators (likelihood ratio = 48.87, P < 0.001). We was not documented on species lists presented in Morrow et al.
found a significant difference between 95% KE and 50% KE (2001) and may represent a recently established invasive species
at this site. Native woody shrubs such as smooth alder, swamp
estimates (z = 8.358, P < 0.001) and between 95% MCP and 95%
rose, and black willow appear to dominate the central portion of
KE estimates (z = -6.760, P < 0.001) but did not find a significant
HA411 (pers. obs.). These woody plants may out-compete
difference between 95% MCP and 50% KE estimates (z = 1.598, P
herbaceous native vegetation preferred for nesting (such as
= 0.25).
tussock sedge) or hiding (such as sensitive fern). In addition,
although data for our study were collected between the months
DISCUSSION of April and August as in Morrow et al. (2001), we collected a
large majority of the telemetry data in June, during the nesting
Differences in Home-Range Size between Study Years and season. Calculated daily movement distances were higher in
Locations.—Chase et al. (1989) reported 95% MCP home-range June than in any other month, suggesting that either the
sizes of 2,400 m2 or less for Bog Turtles at two sites in Maryland, frequency of data collection or nesting-related movements may
whereas Morrow et al. (2001) reported 95% MCP home-range have affected telemetry data in our study. This difference in
sizes as high as 31,200 m2 at one of their two study sites (HA406). radiotelemetry effort may account for observed differences in
home-range sizes between our study and Morrow et al. (2001),
although proliferation of woody and invasive vegetation is
another potential cause.
We documented significant differences in 95% MCP and 95%
KE home-range sizes between BA030 and HA411 but no
significant differences in 50% KE home-range sizes between
BA030 and HA411. This could suggest that core used areas
(approximated by 50% KEs) may be relatively similar for Bog
Turtles at each site. Average 50% KEs at BA030 were almost
twice as large as those at HA411, however, despite the lack of
statistical significance for this comparison; in addition, 50% KEs
at each site had very high standard deviations, possibly
accounting for the nonsignificant difference between the two
sites. With these caveats, all estimated home-range sizes
appeared larger for BA030 than for HA411. BA030 is a larger
wetland than HA411. Because Bog Turtles spend most of their
time within wetland boundaries (Ernst and Lovich, 2009), this
may partially explain the higher home-range sizes at BA030 and
would suggest that turtles in larger wetlands should typically
move longer distances to take advantage of available high-
quality habitat. However, reed canary grass has proliferated at
BA030 since initial surveys were conducted in 1992 (SAS, pers.
obs.). Reed canary grass is invasive throughout North America
and can form dense monocultures within wetlands (Lavergne
and Molofsky, 2004). In Bog Turtle wetlands, reed canary grass
FIG. 1. Box and whiskers plot of home-range sizes for the three and other invasive plants reduce light penetration to the
estimation techniques. wetland surface and out-compete native plants that Bog Turtles

Downloaded From: https://bioone.org/journals/Journal-of-Herpetology on 17 Oct 2021


Terms of Use: https://bioone.org/terms-of-use Access provided by Clemson University
HOME RANGES OF BOG TURTLES 71

may need for foraging, nesting, and basking (Morrow et al., home-range data over two or three consecutive years. Therefore,
2001). If wetland quality is deteriorating at this site attributable management efforts that incorporate radiotelemetry may need to
to the spread of reed canary grass, Bog Turtles would need more collect home-range information at study sites every 5–10 years if
habitat area to fulfill habitat requirements (Morrow et al., 2001). researchers hope to detect changes in home-range size and
Differences in Home-range Size between Estimators.—Although no habitat quality. By incorporating these considerations, conserva-
statistically significant difference was detected between 50% KEs tion efforts could more accurately estimate patterns of habitat use
and 95% MCPs, 95% KEs were significantly larger than 50% KEs for this elusive species.
and 95% MCPs by a factor of six or more. For this study, we used
a relatively well-established smoothing parameter algorithm Acknowledgments.—Funding for this study was provided by
(LSCV). We chose this smoothing parameter because very few the U.S. Fish and Wildlife Service, the Maryland Department of
published studies have used kernel estimates for Bog Turtles, Natural Resources, and the Towson University Department of
with the exception of Morrow et al. (2001); but Morrow et al. Biological Sciences. This study was approved by the Towson
(2001) did not specify their choice of smoothing parameters. University IACUC committee (IACUC FR 021913 RS-04) and
Although Row and Blouin-Demers (2006) suggested that kernel the Maryland Department of Natural Resources (permit 53959).
estimates using the LSCV bandwidth estimator are unreliable for We thank A. Byer, K. Anderson, R. McGehee, S. Martin, S.
herpetofauna, we wanted to evaluate this common bandwidth McDaniel, and many others for their help and guidance with
estimator for this species before evaluating alternatives to this fieldwork and study design. We also thank our site owners,
estimator, because Bog Turtles typically have much smaller without whom this work would not have been possible. S.
home-range sizes than other reptiles. Martin and A. Byer provided comments and suggestions on the
Our study supports the assertions of Row and Blouin-Demers manuscript.
(2006) about the unreliability of kernel estimates for herpeto-
faunal field studies, because 95% kernel estimates varied
considerably, and typically were much larger than 95%
minimum convex polygons. Surprisingly, 95% minimum convex LITERATURE CITED
polygons and 50% kernel estimates with the LSCV bandwidth BARG, J. J., J. JONES, AND R. J. ROBERTSON. 2005. Describing breeding
estimator produced similar home-range sizes in our study. We territories of migratory passerines: suggestions for sampling, choice
urge caution in interpreting these patterns in home-range sizes, of estimator, and delineation of core areas. Journal of Animal
Ecology 74:139–149.
however, because the number of telemetry locations per female BARTLETT, M. S. 1937. Properties of sufficiency and statistical tests.
in this study was small. Row and Blouin-Demers (2006) Proceedings of the Royal Statistical Society, Series A 160:268–282.
suggested that the number of telemetry locations will alter BÖRGER, L., N. FRANCONI, G. DE MICHELE, A. GANTZ, F. MESCHI, A. MANICA,
kernel home-range sizes, indicating that relatively small sample S. LOVARI, AND T. COULSON. 2006. Effects of sampling regime on the
mean and variance of home range size estimates. Journal of Animal
size may have inflated kernel home-range sizes in this study.
Ecology 75:1393–1405.
Lanver and Kelly (2008) suggested that unified and estab- BOYLE, S. A., W. C. LOURENCO, L. R. DA SILVA, AND A. T. SMITH. 2009. Home
lished standards should be developed for each field regarding range estimates vary with sample size and methods. Folia
the reporting and methodology of home-range estimates. Our Primatologica 80:33–42.
study suggests the methods outlined in Row and Blouin- CALENGE, C. 2006. The package adehabitat for the R software: a tool for
the analysis of space and habitat use by animals. Ecological
Demers (2006) represent a solid foundation for evaluating Modelling 197:516–519.
home-range estimator choices for reptiles or at least for turtles. CARTER, S. L., C. A. HAAS, AND J. C. MITCHELL. 1999. Home range and
Because this group encompasses many different body plans, habitat selection of bog turtles in southwestern Virginia. Journal of
natural histories, and life-history strategies, however, research- Wildlife Management 63:853–860.
ers should recognize that developing general practices for all CHASE, J. D., K. R. DIXON, J. E. GATES, D. JACOBS, AND G. J. TAYLOR. 1989.
Habitat characteristics, population size, and home range of the bog
reptiles may be difficult. Other studies have documented turtle, Clemmys muhlenbergii, in Maryland. Journal of Herpetology 23:
species and taxa-specific patterns of differences between 356–362.
home-range estimators (Barg et al., 2005; Huck et al., 2008). ECKLER, J. T., A. R. BREISCH, AND J. L. BEHLER. 1990. Radio telemetry
Therefore, the choice of a home-range estimator should be based techniques applied to the bog turtle (Clemmys muhlenbergii Schoepff
upon study design and study species, because both of these 1801). Pp. 69–70 in R. S. Mitchell, C. J. Sheviak, and D. J. Leopold
(eds.), Ecosystem Management: Rare Species and Significant
factors can influence home-range estimates (Börger et al., 2006; Habitats. New York State Museum, USA.
Boyle et al., 2009). ERNST, C. H. 1977. Biological notes on the bog turtle, Clemmys
Conservation Implications.—To incorporate home-range sizes muhlenbergii. Herpetologica 33:241–246.
into habitat management efforts for this species, biologists must ERNST, C. H., AND J. E. LOVICH. 2009. Turtles of the United States and
Canada. 2nd ed. Johns Hopkins University Press, USA.
collect telemetry data on a site-specific basis and should take site
ERNST, C. H., M. F. HERSHEY, AND R. W. BARBOUR. 1974. A new coding
area into account when designing studies. In addition, home- system for hardshelled turtles. Transactions of the Kentucky
range estimators must be chosen carefully, as considerable Academy of Science 35:27–28.
variation in home-range sizes is apparent when using different HOLMAN, J. A., AND U. FRITZ. 2001. A new emydine species from the
estimation techniques. Habitat management efforts should also middle Miocene (Barstovian) of Nebraska, USA with a new generic
arrangement for the species of Clemmys sensu McDowell (1964).
collect information on changes in home-range size over time, as Zoologische Abhandlungen 51:331–353.
such information may be useful for assessing responses of turtles HOTHRON, T., F. BRETZ, AND P. WESTFALL. 2008. Simultaneous inference in
to habitat modification. Although our study did not document an general parametric models. Biometrical Journal 50:346–363.
effect of year on home-range size, the observed increase in home- HUCK, M., J. DAVISON, AND T. J. ROPER. 2008. Comparison of two sampling
range size at HA411 since Morrow et al. (2001) may indicate that protocols and four home-range estimators using radio-tracking data
from urban badgers Meles meles. Wildlife Biology 14:467–477.
changes in home-range size over time are mainly driven by KLEMENS, M. W. 2001. Bog turtle (Clemmys muhlenbergii) northern
relatively slow changes in wetland vegetation, specifically woody population recovery plan. U.S. Fish and Wildlife Service Technical
species. These changes may not be detected by studies that collect Report, USA.

Downloaded From: https://bioone.org/journals/Journal-of-Herpetology on 17 Oct 2021


Terms of Use: https://bioone.org/terms-of-use Access provided by Clemson University
72 N. W. BYER ET AL.

LANVER, P. N., AND M. J. KELLY. 2008. A critical review of home range PITTMAN, S. E., AND M. E. DORCAS. 2009. Movements, habitat use, and
studies. Journal of Wildlife Management 72:290–298. thermal ecology of an isolated population of bog turtles (Glyptemys
LAVERGNE, S., AND J. MOLOFSKY. 2004. Reed canary grass (Phalaris muhlenbergii). Copeia 2009:781–790.
arundinacea) as a biological model in the study of plant invasions. ROW, J. R., AND G. BLOUIN-DEMERS. 2006. Kernels are not accurate estimators
Critical Reviews in Plant Sciences 23:415–429. of home-range size for herpetofauna. Copeia 2006:797–802.
SCHUBAUER, J. P. 1981. A reliable radio-telemetry tracking system suitable
LEE, D. S., AND A. W. NORDEN. 1996. Distribution, ecology and
for studies of chelonians. Journal of Herpetology 15:117–120.
conservation needs of Maryland’s bog turtles, Clemmys muhlenbergii. SEAMAN, D. E., AND R. A. POWELL. 1996. An evaluation of the accuracy of
Maryland Naturalist 40:7–46. kernel density estimators for home range analysis. Ecology 77:2075–
MORROW, J. L., J. H. HOWARD, S. A. SMITH, AND D. K. POPPEL. 2001. Home 2085.
range of the bog turtle (Clemmys muhlenbergii) in Maryland. Journal SHAPIRO, S. S., AND M. B. WILKS. 1965. An analysis of variance test for
of Herpetology 35:68–73. normality (complete samples). Biometrika 52:591–611.
PINHEIRO, J., D. BATES, S. D. ROY, D. SARKAR, AND THE R DEVELOPMENT CORE
TEAM. 2013. Nlme: Linear and Nonlinear Mixed Effects Models. R Accepted: 24 June 2016.
package version 3.1–113. Published online: 11 January 2017

Downloaded From: https://bioone.org/journals/Journal-of-Herpetology on 17 Oct 2021


Terms of Use: https://bioone.org/terms-of-use Access provided by Clemson University

You might also like