ArticlePDF Available

Demography of a painted turtle intergrade (Chrysemys picta picta X C. p. marginata) population from an altered wetland

  • Kansas Department of Wildlife, Parks, and Tourism

Abstract and Figures

The demography of a painted turtle Chrysemys picta picta X C. p. marginata population from a eutrophic habitat was examined at a wetland site in south-central Pennsylvania (USA) during 2011–2019. Males reached sexual maturity at 90 mm carapace length (CL) in half the time taken, but at the same size, as painted turtles studied elsewhere in the north-eastern portion of the United States. Females matured at 130 mm CL at our site, which was larger and began at an earlier age than conspecifics. Our data corroborate findings of faster growth in C. picta juveniles resulting in earlier maturity at body sizes equal to or larger than slower growing juveniles. Our results also conform to previous findings linking wetlands altered by added nutrient input to increased growth patterns of their resident painted turtle population. Rapid growth rates for aquatic turtles are likely to become more common globally as urbanisation continues to expand and alter wetland habitats.
Content may be subject to copyright.
Herpetological Journal FULL PAPER
Walter E. Meshaka
Volume 33 (January 2023), 14–24
Published by the Brish
Herpetological Society
The painted turtle Chrysemys picta Schneider 1783, is
a polytypic aquac species found across central and
eastern North America (Powell et al., 2016) with three
recognised subspecies (Uetz et al., 2021). In Pennsylvania,
the range of the midland painted turtle, C. p. marginata
Agassiz 1857, extends through much of the western and
extreme northern porons of the state. Intergradaon
by C. p. marginata with the eastern painted turtle C. p.
picta occurs throughout the eastern and south-eastern
portion of the state (Ernst & Ernst, 1971). The lower
sub basin of the Susquehanna River is the centre of the
intergradation zone in the state (Hulse et al., 2001).
Within this watershed is Wildwood Park in Harrisburg,
Dauphin County, where intergrades of C. p. picta X C.
p. marginata are abundant in a eutrophic canal and an
arcial lake (Wingert & Meshaka, 2021).
Demographic studies of C. picta are common (e.g.
Gibbons, 1968; Mitchell, 1988; Zweifel, 1989; Frazer et
al., 1991; Congdon & Gibbons, 1996), and two studies
have examined demographic paerns in Pennsylvania
populaons (Ernst, 1971a,b; Hughes & Meshaka, 2020).
Studies by Ernst (1971a,b) were conducted on an
intergrade populaon in south-eastern Pennsylvania, and
a study by Hughes & Meshaka (2020) was conducted on
C. p. marginata in an arcial wetland complex in south-
western Pennsylvania. Nutrient levels within the aquac
habitats occupied by this species can aect demographic
traits within populaons (Ernst & Lovich, 2009). Although
neither of the wetland habitats in the two Pennsylvania
studies were described as altered, three studies, one
on C. p. bellii in Iowa (Quinn & Chrisansen, 1972) and
in Michigan (Gibbons, 1968), and one on C. p. picta in
Maryland (Ernst & McDonald, 1989), explicitly examined
growth in habitats altered by eutrophic condions. Rapid
growth was common to turtles of all three studies, larger
adult body sizes was found in two studies (Quinn &
Chrisansen, 1972; Gibbons, 1968), and early maturity
with an eect on minimum body size was detected by one
study (Ernst & McDonald, 1989). More broadly, Congdon
et al. (2018) idened a connecon between fast growth
of juveniles and early maturity at larger or similar sizes
compared to slower-growing juveniles of three turtle
species, including C. picta.
Altered aquatic habitats are highly influential in
shaping variaon in several important life-history traits
and such demographic responses are likely common
among urban populaons of aquac turtles worldwide.
Within this context, we recognised the opportunity
to comprehensively evaluate responses in multiple
demographic traits by a single populaon of C. p. picta
X C. p. marginata over a 9-year period to an arcially
eutrophic and human-made wetland imbedded within
a city in south-central Pennsylvania. Our research
approach, in turn, provides ndings that are not only
globally relevant but also becoming increasingly common.
Demography of a painted turtle intergrade (Chrysemys picta picta X
C. p. marginata) populaon from an altered wetland
Walter E. Meshaka, Jr.1, Eugene Wingert2,3, Daren Riedle3, Sco Boback4 & Daniel F. Hughes5
1Secon of Zoology and Botany, State Museum of Pennsylvania, 300 North Street, Harrisburg, PA 17120, USA
2Department of Biology, Dickinson College, PO Box 1773, Carlisle, PA 17013, USA
3Kansas Department of Wildlife, Parks, and Tourism, 512 SE 25th Ave, Pra, KS 67124, USA
4Department of Biology, Dickinson College, PO Box 1773, Carlisle, PA 17013, USA
5Department of Biology, Coe College, 1220 1st Avenue NE, Cedar Rapids, IA 52402, USA
The demography of a painted turtle Chrysemys picta picta X C. p. marginata populaon from a eutrophic habitat was examined
at a wetland site in south-central Pennsylvania (USA) during 2011–2019. Males reached sexual maturity at 90 mm carapace
length (CL) in half the me taken, but at the same size, as painted turtles studied elsewhere in the north-eastern poron of
the United States. Females matured at 130 mm CL at our site, which was larger and began at an earlier age than conspecics.
Our data corroborate ndings of faster growth in C. picta juveniles resulng in earlier maturity at body sizes equal to or larger
than slower growing juveniles. Our results also conform to previous ndings linking wetlands altered by added nutrient input
to increased growth paerns of their resident painted turtle populaon. Rapid growth rates for aquac turtles are likely to
become more common globally as urbanisaon connues to expand and alter wetland habitats.
Keywords: Growth, populaon size, populaon structure, survivorship, urban
Study area
Our study was conducted at Wildwood Park, a 93.5 ha
county park located in Harrisburg, Dauphin County,
Pennsylvania (40.310, -76.883). Approximately 60 % of
the park is comprised of a shallow arcial lake that is
fed by Paxton Creek (Fig. 1). An accumulaon of detritus
has resulted in a gradual lling-in of the lake with much
of it converng to a marsh dominated by caail Typha
sp. During the me of this study, only a secon of Paxton
Creek (0.34 ha) at the south end, the spillway area (0.29
ha), a channel (0.99 ha) running more or less parallel to
the tow path on the west end, and another channel (1.91
ha) running along the eastern edge of the lake, were
deep enough to be habitable by C. picta. A secon of the
Pennsylvania Canal ran along the western boundary of
the park adjacent to the lake and was separated by a tow
path (Fig. 1). The canal measured 1,934.65 m in length,
had an average width of 23.8 m, and an area of 26,467.6
m2 (2.65 ha). The canal depth changed signicantly from
one to two metres from the west side of the towpath
into the canal (Russell et al., 2014). A cleared utility
right-of-way averaging 16.3 m borders the west side of
the canal and separates it from a two-lane paved road.
The main water lily found in the canal was spadderdock
Nuphar advena L., and the dominant submergent aquac
macrophyte was coontail Ceratophyllum demerseum L.
Captured turtles frequently passed spadderdock seeds.
Duckweed Lemna sp. was the common oang plant.
Small painted turtle juveniles were seen feeding on
duckweed at the surface. There are also algae species
in the water which have not yet been idened (Russell
et al., 2014).
The canal received extensive runo from the adjoining
road and industrial warehouses that run parallel and
west of it, to the extent that much of the lake has
converted to caail marsh. The eutrophic condion of
our site is quantified by water quality data recorded
by the Susquehanna River Basin Commission’s Paxton
Creek monitoring station (ID 01571005, coordinates
40.306, -76.856) located upstream from Wildwood Park,
its period of record having encompassed the duraon
of our study. The extent of eutrophication in Paxton
Creek expressed in normalised concentraon (mg/L) was
available for total Nitrogen (0.676), dissolved Nitrogen
(0.833), total Phosphorus (0.598), dissolved Phosphorus
(0.618), total Ammonia (0.833), dissolved Ammonia
(0.539), and total Suspended Solids (0.578). The topic
of the watershed’s impairment was addressed at both
the state level by the Pennsylvania Department of
Environmental Protecon (DEP) and at the federal level
by the Environmental Protecon Agency. The DEP listed
Paxton Creek in Harrisburg, Dauphin County, as impaired
aer studies in 2004, 2005, and 2006 based on siltaon
and the source as urban runo/storm sewers, primarily
phosphorus. In 2010, the DEP delisted the watershed that
includes Paxton Creek (Shearer, 2012), but the EPA later
determined that delisng of Paxton Creek for nutrient
impairment was not appropriate (Sauro, 2019; DeJesus,
2021). As of 2013, DEP listed Paxton Creek as impaired
for sediment but not for nutrients. During the laer 20th
century, the lake depth significantly decreased with
sediment deposion. It originally averaged four to 152.4
cm in depth and as of 2015 Wildwood Lake averaged
approximately 15.2 cm (Herbert et al., 2015). From 2003
through 2018, the lake depth diminished from shallow
open water to mudat and caail marsh (Fig. 2).
Trapping and Processing Method
Six baited hoop-nets were set for ve consecuve days
in spring, summer, and autumn during 2011–2014 and
opportuniscally in 2015 (54 trap days), 2016 (30 trap
days), and 2019 (30 trap days). The traps (Memphis Net
and Twine Co., Memphis, TN) were 2.0 m x 1.0 m with 2.54
cm mesh. The traps were set at xed locaons near the
shoreline of the canal, and the sites remained constant
for the duraon of the study. Traps were baited with a
parally opened sardine can or with chicken gizzards
which were changed daily aer traps were checked.
We used a 61 cm aluminum sliding caliper, accurate
to 0.5 mm, to measure straight-line carapace length
(CL) and plastron length (PL). The sex for adult turtles
was determined by foreclaw length (longer in males
relave to CL) and by the locaon of the cloaca relave
to the carapace (i.e. the cloaca extends beyond the
carapace in males; Ernst & Lovich, 2009). New turtles
were individually marked using two methods. Each new
Demography of a painted turtle intergrade population from an altered wetland
Figure 1. A section of the Pennsylvania Canal and
Wildwood Lake at Wildwood Park, Harrisburg, Dauphin
County, Pennsylvania. Photograph taken 3 October 2018
by W.E. Meshaka, Jr.
turtle was permanently marked using the Proximate
Binary Coded Decimal (PBCD) scute-notching system of
Buhlmann et al. (2008). The notches which displayed a
unique number were made with a Dremel tool. Each new
turtle was also given a Passive Integrated Transponder
(PIT) tag inserted through the le thigh along the bridge
of the carapace. The tags were HPT12 preloaded sterile
tags from BioMark (Boise, Idaho), inserted using a
BioMark MK-25 Rapid Implant Gun. Recaptured turtles
were measured, scanned for the PIT tag, and carapacial
notches were renewed as needed before releasing
turtles on the same day of their capture.
Our study site supported three other turtle species,
which were captured in the traps during this study.
eastern musk turtles Sternotherus odoratus (n = 2) and
snapping turtles Chelydra serpentina serpentina (n =
62), were individually marked using the same methods
applied to painted turtles and released. Exoc red-eared
sliders Trachemys scripta elegans also established at
Wildwood, were captured (n = 22) and euthanised, and
a sample of them was deposited in the State Museum
of Pennsylvania, Harrisburg, Pennsylvania (Russell et al.,
Determinaon of Sexual Maturity
Long foreclaws are associated with mature males in this
species (Ernst & Lovich, 2009). Foreclaws of at least 7
mm were common in males ranging 93.0–121.3 mm
CL. Foreclaws of at least 8 mm were common in males
ranging 93.5–141.8 mm CL. Foreclaws of at least 9 mm
were common in males ranging 90–140.5 mm CL. Among
four males larger than 90 mm CL, one had foreclaws of
6 mm, and three had foreclaws of 5 mm. No foreclaws
were larger than 6 mm among the ve males smaller
than 90 mm CL. Thus, we concluded that males of at least
90 mm CL evidenced clear sign of sexual maturity in this
secondary sexual characterisc (Fig. 3).
We used body size of dissected and nesng females
during 2011–2021, as the criterion of female shell length
at sexual maturity. During 2016–2017, seven females
(132.9–154.2 mm CL) were removed from the populaon
to ascertain minimum shell length at sexual maturity.
Specimens were deposited in the section of Zoology
and Botany of the State Museum of Pennsylvania.
Opportunisc observaons of 15 nesng females during
2011–2014 (143–155 mm CL) and 2018, 2019, and
2021 (130–165.1 mm CL) provided addional data to
determine shell length of the smallest mature female.
The smallest sexually mature female measured 130 mm
CL (124 mm PL) and was seen nesng on 2 July 2018.
A female (SMP-H-9230) measuring 132.9 mm CL (124.6
mm PL) captured on the tow path on 8 July 2016 was
found to contain luteal scars and yolking ovarian follicles.
Based upon this sample, 130 mm CL was accepted as the
cut-o for smallest sexually mature female in this study.
Growth and Age Esmaon
In most species of turtles, age can be reasonably esmated
in young individuals using growth annuli on epidermal
scutes (Spencer, 2002). However, counng growth rings
has been found to be unreliable in providing accurate
age esmates in adult turtles (Wilson & Tracy, 2003),
especially for esmang ages in adult C. picta (Brooks et
al., 1997). Lindeman (1996), in parcular, showed that
counng growth rings for individuals of C. picta becomes
inaccurate around age 7 (i.e. counts of scute annuli are
useful in age determinaon only among juveniles, and
unreliable, if even readable, thereaer). Alternavely,
we set out to establish sex-specific, length-at-age
relaonships for our samples based on repeated records
of age and length from known-age individuals that had
a more conservative value of 5 or fewer estimated
growth rings when rst captured (37 females captured
58 mes; 60 males captured 94 mes). We used a mul-
model approach to compare three well established
growth models (von Bertalany, Gompertz, and Logisc)
using the ‘AquacLifeHistory’ package (Smart, 2019a)
in R (version 4.0.5; R Core Team, 2021). We rooted the
models using the carapace length of a hatchling found
at the site (24.8 mm). We assessed the best fitting
models using an informaon theorec approach (Akaike
Informaon Criterion AIC; Burnham & Anderson, 2002).
The best ng model for each sex was then used as input
to build length-at-age curves using a Bayesian Markov
chain Monte Carlo process with the ‘BayesGrowth’
package in R running for 5,000 iteraons (Smart, 2019b).
Growth was also calculated as the dierence in carapace
W. Meshaka et al.
Figure 2. Wildwood Lake in 2003 (A) with open water
and visited by wading birds and dabbling ducks, and in
2018 (B) having succeeded to mudats and caail marsh.
Property of Wildwood Lake Nature Center Archives.
Demography of a painted turtle intergrade population from an altered wetland
length between captures divided by the interval in years
between captures, which we ploed against carapace
length at rst capture. Carapace size intervals in the bar
histogram of body size distribution were determined
using the equaon of Sturges (1926). Summary stascs
of body size were performed using Microso Excel 365
(Microso Inc., Redmond, Washington, USA).
Populaon Size and Survivorship
We calculated apparent annual survival (Φ) and
recapture rates (p) using open populaon Cormack-Jolly-
Seber models (CJS; Lebreton et al., 1992) in the program
MARK (White & Burnham, 1999). To test for dierences
in Φ and p between sexes, we generated CJS models to
examine whether Φ or p diered based on sex, me, or
a sex-me interacon. We based model selecon for all
analyses on AICc (corrected AIC for small sample sizes)
values, with lower values denong greater parsimony
(Burnham & Anderson, 2002). We calculated populaon
abundance for adults using POPAN parameterisaon of
Jolly-Seber models (Jolly, 1965; Seber, 1965) in MARK
(White & Burnham, 1999).
Encounter histories to calculate demographic traits
estimate the probability that an individual will leave
a population. With encounter rates reversed, the
probability of an individual entering the populaon was
esmated (Pradel, 1996), whereby λ = rate of individuals
entering a populaon. Pradel’s λ diers from tradional
esmates of λ as no fecundity values are included in
its calculation, so is not necessarily equivalent to a
true populaon growth rate. Pradel’s λ was esmated
in Program MARK in conjuncon with the CJS models
described above. Measure of central tendency was
expressed as mean and standard deviation unless
otherwise noted. Stascal signicance was recognised
at a p value of < 0.05.
Populaon Structure
Juveniles comprised 14.1 % of 375 new captures,
outnumbered by adults at 6.08:1.00. The adult
male:female sex ratio of this sample was 2.93:1.00.
Adult male body size averaged 121.4 mm CL (std. dev. =
± 14.1; min-max = 90.0–156.0; n = 240) and 111.6 mm PL
(std. dev. ± 12.8; min-max = 84.0–142.2; n = 240). Adult
female body size averaged 147.8 mm CL (std. dev. ± 8.7;
min-max = 129.7–171.0; n = 82) and 138.3 mm PL (std.
dev. ± 8.4; min-max = 120.1–165.0; n = 82). Many of the
240 males (40 %) fell in the 111.0–126.0 mm CL range,
and most of the 82 females (59.8 %) fell into the 143.0–
158.0 mm CL range (Fig. 4). Plastron length was strongly
related to CL in adults of both males (r2 = 0.95, F = 4357, p
< 0.001; mm PL = 0.8857 (mm CL) + 4.1118) and females
(r2 = 0.89, F = 627.69, p < 0.001; mm PL = 0.906 (mm CL)
+ 4.4546).
Figure 3. Relaonship between foreclaw length and carapace length (CL) in 215 male painted turtles C. picta picta X C.
p. marginata at Wildwood Park, Harrisburg, Dauphin County, Pennsylvania, during 2011–2019. Blue crosshairs indicate
thresholds associated with minimum foreclaw length associated with sexual maturity.
W. Meshaka et al.
Figure 4. Body size distribuons of 240 adult male, 82 adult female, and 53 juvenile painted turtles C. picta picta X C. p.
marginata at Wildwood Park, Harrisburg, Dauphin County, Pennsylvania, during 2011–2019.
Figure 5. Annual growth rate ploed against carapace length at rst capture for males, females, and juveniles of the
painted purtle C. picta picta X C. p. marginata at Wildwood Park, Harrisburg, Dauphin County, Pennsylvania, during
Growth and Age Esmaon
Growth rates decreased with increasing carapace
length (Fig. 5). The average growth rate was highest for
juveniles (12.2 mm/yr, min-max = 4.9–41.2 mm/yr, n =
13) and much lower for females (2.8 mm/yr, min-max
= 0–12.2 mm/yr, n = 17) and males (2.3 mm/yr, min-
max = 0.2–9.1 mm/yr, n = 68). Based on AICc values,
the top models for growth diered between the sexes:
von Bertalanffy for males and Gompertz for females
(Table 1). Estimates of asymptotic body size (A) and
characterisc growth constant (k) returned the following
values based on the top models for each sex: males
(A = 116.4; k = 0.6) and females (A = 142.7; k = 0.59).
The growth trajectory of males rose more rapidly than
females, such that males reached maturity faster, at two
years of age, but males exhibited almost no noceable
growth beyond an esmated age of four years (Fig. 6).
Males were also smaller at their asymptoc body size
compared to females. Females, in contrast, appeared to
grow at a steadier pace unl reaching maturity. Based on
visual inspecon of the growth curve and female body
sizes (Fig. 6), a few females reached maturity at three
years of age, many at four years, and by ve years of age
all females were mature. Thereaer, growth connued,
albeit very slowly.
Survivorship, Detecon, and Populaon Size
Female survivorship (0.80 ± 0.04) was lower than that
of males (0.89 ± 0.02), although recapture probabilies
were similar (Table 2). For the most parsimonious model,
Φ diered between sexes and p was me dependent
(Table 3). As expected, recapture probabilies increased
with sampling intervals and numbers of individuals
marked. Pradel’s λ was stable to slightly increasing for
females and stable to slightly decreasing for males. Since
Pradel’s λ values are based on probability of new animals
entering the populaon, these values are indicave of
the probability of capturing an unmarked individual
entering the populaon. The populaon was male biased
at roughly 2.1:1 among inial captures, which is reected
in the populaon esmates (Table 2). Populaon density
in the canal (2.65 ha) and the lake (3.53 ha) combined
was 52.3 turtles/ha. Trends in recapture rates between
sampling intervals increased over me (Fig. 7).
We found that several demographic traits of the turtle
population at our site were accelerated, and when
examined in the context of other demographic studies in
the mid-Atlanc region (Ernst & Lovich, 2009), it appears
that dierences in nutrient levels in the wetlands may
be a likely factor. In Pennsylvania, for example, Ernst
(1971a,b) provides a reasonable comparison to our
results because he also studied a population of C. p.
picta X C. p. marginata intergrades from a more natural
wetland on Big Chickees Creek, which is only about 45 km
south-east of our site. Our study site, on the other hand,
was quanably eutrophic and subjected to connuous
nutrient enrichment (see Methods). This apparent
dierence in resource availability, in turn, provided us
with a variable to consider in the following discussion as
an eect on comparave growth rates, body sizes, and
ages at maturity between the populaons. However,
we note that meaningful interpopulaon dierences,
possibly associated with hybrid vigour, cannot be ruled
out, nor canotherabioc dierences between the two
Demography of a painted turtle intergrade population from an altered wetland
Table 1. Multi-model comparisons for determining sex-specific, best-fitting growth curves for Chrysemys picta
at Wildwood Park, Harrisburg, Dauphin County, Pennsylvania, during 2011–2019. Asymptoc body size (A) and
characterisc growth constant (k) ± 1 standard error.
Model AICc ∆AIC AICc
A k
Males von Bertalany 645.77 0 0.77 116.4 ± 1.75 0.6 ± 0.055
Gompertz 648.48 2.71 0.20 114.8 ± 1.53 0.97 ± 0.105
Logisc 652.25 6.47 0.03 113.9 ± 1.44 0.79 ± 0.079
Females Gompertz 482.46 0 0.44 142.7 ± 6.04 0.59 ± 0.107
Logisc 482.84 0.38 0.36 138.8 ± 5.07 0.81 ± 0.143
von Bertalany 484.11 1.65 0.19 152.9 ± 8.86 0.34 ± 0.065
Table 2. Population dynamics of the painted turtle
Chrysemys picta at Wildwood Park, Harrisburg, Dauphin
County, Pennsylvania, during 2011–2019. Parameters
include apparent survivorship (Φ), recapture probability
(p), Pradel’s Lambda (λ), and populaon size (n) ± one
standard error and 95 % condence intervals.
Female 0.80 ± 0.04
(0.74, 0.93)
0.13 ± 0.03
(0.08, 0.19)
1.02 ± 0.02
(0.97, 1.06)
96 ± 13
(73, 128)
Male 0.89 ± 0.02
(0.84, 0.93)
0.17 ± 0.02
(0.14, 0.21)
0.98 ± 0.01
(0.95, 1.01)
227 ± 24
(185, 283)
habitats unrelated to nutrient levels, which may have
accounted for the demographic dierences we detail
Male painted turtles at Wildwood Park reached a
larger minimum PL of 83.8 mm than the 70 mm PL
reported by Ernst (1971a). Likewise, respecve mean
PL (111.6 mm, 96 mm) and maximum PL (142.2 mm,
121.0 mm) of adult males were larger at our eutrophic
site than those reported by Ernst (1971a). Elsewhere in
the mid-Atlanc region, sexual maturity in males of C. p.
picta was reached at > 71 mm PL in Myrtle Grove Wildlife
Management Area, a natural habitat in which turtles
exhibited normal growth, Charles County, Maryland
(Ernst & McDonald, 1989), and 71mm PL (77.7 mm CL) in
a lake, creek and beaver ponds in Henrico County, Virginia
(Mitchell, 1988). However, at a sewage treatment plant
in Charles County, Maryland, the smallest mature male
measured 87.1 mm PL (Ernst & McDonald, 1989). These
comparisons between our study and more natural sites
corroborate the conclusions by Ernst & McDonald (1989)
that males exhibit plasticity in body length at sexual
maturity. Our data and those of Ernst & McDonald
(1989) on minimum PL at sexual maturity provide a PL
range (83.8–87.1 mm) as a general response to eutrophic
condions in males of mid-Atlanc populaons, such
that recently matured males from eutrophic systems are
approximately 1.2 mes the size of their counterparts
from less altered systems.
The minimum PL associated with sexually mature
females from our site (122.2 mm) was larger than the
minimum (100.8 mm) reported by Ernst (1971a) from a
natural seng. The same was true of respecve mean
PL (138.3 mm, 116.9 mm) and maximum PL (165.0 mm,
145.4 mm) of adult females (Ernst, 1971a). In Henrico
County, Virginia, minimum size at sexual maturity
in females of C. p. picta was reached at 97.2 mm PL
(Mitchell, 1988). However, among four females collected
at a sewage lagoon in Charles County, Maryland, two
females measuring 127 mm and 129 mm PL were not
yet mature, and 132 mm and 139 mm PL may or may not
have been mature, but none of these females contained
shelled eggs or corpora lutea, nor were collecng dates
provided (Ernst & McDonald, 1989). An examinaon of
the ovarian follicle size-classes and widest diameters
indicated that the two largest females collected by
Ernst & McDonald (1989) may have been developing
their rst clutch, or, if mature, their rst clutch of the
season. Our data and those of Ernst & McDonald (1989)
suggest that females subjected to eutrophic condions
exhibit a larger body size at sexual maturity resulng in
W. Meshaka et al.
Figure 6. Best-ng, sex-specic growth curves for the painted turtle C. picta picta X C. p. marginata using known
length-at-age data collected from wild individuals at Wildwood Park, Harrisburg, Dauphin County, Pennsylvania, during
2011–2019. Dark lines indicate age-length relaonships and shaded areas condence bands, generated by a Bayesian
Markov chain MonteCarlo process.
a primiparous shell length at 1.3 mes the size of their
counterparts from more natural systems.
Mean values of shell dimensions can reect dierences
in the environment, especially nutrient inputs. For
example, in Virginia mean adult PL of males and females,
respecvely, were larger (103.1 mm and 124.1 mm) from
a more eutrophic site (Mitchell, 1985a,b) than those in a
less nutrient-rich site nearby (96.2 mm and 120.5 mm;
Mitchell, 1988). To that end, we note that the mean PL
of adult males (96.0 mm) and females (116.9 mm) from
Ernst’s (1971a) natural site were much smaller than those
from our site (111.6 mm and 138.3 mm, respecvely).
Within a single study, Gibbons (1967) found decreasing
size in longest shell lengths in both sexes from three sites
in Michigan that varied in nutrient load: polluted river,
eutrophic lake, and clean marsh. Carnivory in turtles
from Gibbons (1967) also increased with increasing
eutrophication, suggesting dietary differences may
contribute to growth responses of individual males and
females. Thus, rapid growth and overall larger body size is
associated with more nutrient-rich sites, and larger body
size at sexual maturity is more pronounced in females
than in males.
We also wanted to know if growth rates and the
minimum age at sexual maturity differed between
eutrophic and natural sites. Quinn & Chrisansen (1972)
documented faster growth by western painted turtles
C. p. bellii from Iowa in eutrophic systems than in those
with demonstravely less organic maer in the substrate.
Likewise, the nutrient level, adult body size, and carnivory
of the turtles at Gibbons’ (1967) sites were associated with
dierenal growth rates in those in increasingly eutrophic
waterbodies. Ernst & McDonald (1989) corroborated
faster growth in both sexes from a eutrophic site and
determined that sexual maturity of males at a sewage
treatment site was reached in two years, as was ours,
instead of four years from natural sites in south-eastern
Pennsylvania (Ernst, 1971a) and central Virginia (Mitchell,
1988), where all males had foreclaws of at least 8 mm.
Sexual maturity of females was reached at 3–5 years of
age at our site, but females from a sewage treatment
plant in Maryland may have been mature at three years
(Ernst & McDonald, 1989). Ernst & McDonald’s (1989)
uncertainty regarding a connecon between enhanced
growth and early maturity in females from the sewage
lagoon is understandable considering their unavoidably
small sample size of dissected turtles. However, because
Ernst & McDonald’s (1989) site was presumed to have
been much more nutrient-rich than ours, we consider
it probable that the large 3-yr old females at their site
were mature. Comparavely, females from more natural
sites in the mid-Atlantic region matured at an age of
ve (Ernst, 1971a) or eight years (Mitchell, 1988). Some
females from our site reached the minimum body size at
sexual maturity in three years, followed by many at four
years, and all by ve years. Whereas males and females
consistently matured at larger sizes at our site, earlier
age at sexual maturity varied more in females than in
males. Fast growth in juvenile turtles in our study, linked
to elevated nutrient input, results in earlier maturaon at
similar or larger body sizes compared to slower growing
counterparts of C. picta, as well as blanding’s turtle
Emydoidea blandingii and C. serpenna (Congdon et al.,
Demography of a painted turtle intergrade population from an altered wetland
Model AICc Delta AICc AICc
Num. Par. Deviance
Φ(g) p(t) 1098.16 00.997 1 12 350.6368
Φ(g) p(g*t) 1110.135 11.975 0.0025 0.0025 22 340.9789
Φ(t) p(t) 1113.476 15.3153 0.00047 0.0005 19 350.9133
Φ(t) p(g*t) 1119.644 21.4839 0.00002 0 29 334.7405
Φ(g*t) p(t) 1125.847 27.6863 0 0 29 340.9429
Φ(g*t)p(g*t) 1136.363 38.2029 0 0 38 330.4361
Φ(t) p(g) 1154.61 56.4493 0 0 12 407.0861
Φ(g*t) p(g) 1173.009 74.8488 0 0 22 403.8528
Φ(g) p(g) 1177.345 79.1844 0 0 4 446.4368
Φ(.) p(.) 1178.767 80.607 0 0 2 451.9216
Table 3. Comparison of Cormack-Jolly-Seber models for apparent annual survival (Φ) and recapture probability (p)
between male and female painted turtles Chrysemys p. picta X C. p. marginata at Wildwood Park, Harrisburg, Dauphin
County, Pennsylvania, during 2011–2019. Models dier in whether Φ and p are assumed to be constant (.), fully me
dependent (t), or dier between sexes (g), and whether there are interacons (*) among these factors.
Other demographic measures of our study populaon
were in general agreement with those of other populaons
of C. p. picta and C. p. picta X C. p. marginata intergrades.
Survivorship of both sexes was high at our site, a nding
typical of C. p. picta (Mitchell, 1988; Zweifel, 1989) and C.
p. marginata (Hughes & Meshaka, 2020) in the north-east
and mid-Atlanc. Populaon density, however, can range
widely across and within regions. Our esmate of 52.3/ha
exceeded 25/ha of C. p. picta in a New York pond (Bayless,
1975) and 13.9/ha of C. p. marginata in a Pennsylvania
pond (Hughes et al., 2016). In three Michigan ponds,
populaon densies of C. p. bellii ranged from 39.9 to
89.5/ha (Congdon et al., 1986). Seasonal population
density esmates range even more widely, including a
published range of 137–248/ha for C. p. picta in New York
ponds (Zweifel, 1989), and esmates as high as 590.4/
ha for C. p. picta X C. p. marginata in a marsh and pond
in south-eastern Pennsylvania (Ernst, 1971b) and 590/ha
for C. picta in a Michigan pond (Gibbons, 1968).
The adult:juvenile ratio of a population also varies
widely among populations, across habitats, and over
me within a populaon. Our adult:juvenile rao of rst
captures (6.08:1.00) was comparable to values of 5.0:1.0
from a pond (Bayless, 1975), and 4.20:1.00 from a marsh
and pond (Ernst, 1971b). The rao was much lower at
a lake (1.1–1.3:1.00) studied by Mitchell (1988). On the
other hand, an 18-year study in New York ponds yielded
an average adult:juvenile rao of 2.16:1.00 and range
0.45–6.30:1.00 (Zweifel, 1989). In general, hoop-net
based populaon inferences tend to be skewed towards
trapping larger species (Ennen et al., 2021) and mostly
adults of smaller species (Tesche & Hodges, 2015). For
example, it was recently shown that the average CL of
95 C. picta caught in hoop-nets with a mesh width size of
5.08 cm was signicantly larger than the mean CL of 231
individuals caught in hoop-nets with a mesh size of 2.54
cm (Gulee et al., 2019). We suggest that future eorts
employ a variety of turtle sampling methods to determine
if the paerns we found reect the size-class distribuon
of this turtle populaon (e.g. Ream and Ream, 1966).
Adult sex raos of C. picta are oen even (Bayless,
1975; Ernst, 1971b, Mitchell, 1988), however, they can
vary over time (0.62–1.80:1.00) (Zweifel, 1989) and
subjected to dierences based on sampling technique
used (Ream & Ream, 1966). To that end, our hoop-nets
could have drawn males to a single female already in the
traps resulng in a male bias. We also consider a potenal
combined eect of early sexual maturity in males and
dierenal mortality in nesng females that may have
influenced the sex ratio we found. Many of the well-
known mesopredators of nesng females and their eggs
(Ernst & Lovich, 2009) were regularly encountered at
Wildwood Park: raccoons Procyon lotor, red foxes Vulpes
vulpes, mink Neogale vison, striped skunks Mephitis
mephis, and long-tailed weasels N. frenata (W.E.M. and
E.W., pers. obs.).
W. Meshaka et al.
Figure 7. Trends in recapture rates between samples of the painted turtle, C. picta picta X C. p. marginata, at Wildwood
Park, Harrisburg, Dauphin County, Pennsylvania, during 2011–2019.
Fully comparable, systemac comparisons of physical
and chemical parameters among sites, or data on
temporal transitions within sites, as in comparisons
with Ernst’s (1971a,b) sites are lacking; therefore, we
cannot rule out alternave explanaons for the observed
paern. However, the populaon of C. picta at our site
displayed demographic paerns consistent with those
of populaons in other eutrophic systems, suggesng a
general eect of water quality that we suggest should
be invesgated experimentally. Both sexes grew faster
and matured at longer shell lengths than conspecics
from less nutrient-rich habitats. Comparavely, males
showed less plascity in body size at sexual maturity and
a narrower range in reduced age at maturity than females
at our site. From a 33-yr study in Michigan, Congdon et
al. (2018) demonstrated that the juvenile growth rate in
C. picta females has a profound eect on age at sexual
maturity, such that faster growing juveniles matured up to
6-yrs sooner than slower growing females and that post-
maturity growth had very lile eect on size-dependent
reproducve traits, such as clutch size. Consequently, the
earliest-maturing females, such as those that matured in
3-yrs at our site, can potenally have more lifeme egg
output than the oldest-maturing individuals at Wildwood
(5-yrs). Based on average clutch sizes of primiparous
and older female C. picta in Congdon et al. (2018), the
earliest-maturing females at Wildwood could produce 12
eggs over two years before the oldest-maturing females
rst laid eggs, and in turn, it would take these females
decades unl they caught up in lifeme egg output to the
earliest-maturing females. One queson that remains to
be answered in life history theory for C. picta is whether
there is a strong relaonship between age at maturity
and longevity, such that the turtles at Wildwood may
mature earlier compared to elsewhere but die sooner,
a phenomenon that could impact populaon dynamics
in a long-lived, oen abundant species that is ever more
subjected to altered environments in expanding urban
ecosystems across its geographic range. This last point,
and the potenal eects of human-altered habitats on
chelonian demography in general, can by extension
apply as testable hypotheses to aquac turtle species
worldwide, which are facing increasing contact with
human impacts to their remaining aquac habitats.
We would like to acknowledge the following Dickinson
College students without whose dedicaon to this project
this paper could not have been wrien: V. Ceja-Cervante,
M. Digiorgio, C. Macpherson, L. Rano, and E. Sullivan.
The Center for Sustainability Educaon, Research and
Development Committee of Dickinson College kindly
provided funding for this study. We hearly acknowledge
C. Rebert, Director of Wildwood Park, J. Webster, and the
Friends of Wildwood for their kindnesses and support
of this and other research projects we conduct at the
park. We extend our gratude to J.D. Congdon for his
review of an earlier version of this paper. The research
was conducted under the Pennsylvania Fish and Boat
Commission Type II Scienc Collectors’ Permit 119 to
WEM and Type I Scienc Collectors’ Permit 0330 to EW.
Animal-handling ethics followed the American Society of
Ichthyologists and Herpetologists guidelines for the use
of live reples in eld research (Beaupre et al., 2004) and
were conducted by EW under IACUC # 937.
Bayless, L.E. (1975). Population parameters for Chrysemys
picta in a New York pond. The American Midland Naturalist
93(1), 168–176.
Beaupre, S., Jacobson, E., Lillywhite, H. & Zamudio, K. (2004).
Guidelines for use of live amphibians and reples in eld
and laboratory research. Herpetological Animal Care and
Use Commiee of the American Society of Ichthyologists
and Herpetologists.
Brooks, R.J., Krawchuk, M.A., Stevens, C. & Koper, N. (1997).
Tesng the precision and accuracy of age esmaon using
lines in scutes of Chelydra serpenna and Chrysemys picta.
Journal of Herpetology 31(4) ,521–529.
Buhlmann, K.A., Tuberville, T.D. & Gibbons, J.W. (2008). Turtles
of the Southeast. Athens: Georgia. 264 pp.
Burnham, K.P. & Anderson, D.R. (2002). Model Selecon and
Mulmodel Inference. 2nd Edion. Berlin: Springer-Verlag.
488 pp.
Congdon, J.D., Greene, J.L. & Gibbons, J.W. (1986). Biomass of
freshwater turtles: ageographic comparison. The American
Midland Naturalist 115(1), 165–173.
Congdon, J.D. & Gibbons, J.W. (1996). Structure and dynamics
of a turtle community over two decades. In Long-Term
Studies of Vertebrate Communies, 137–159 pp. Cody, M.C.
& Smallwood, J. (Eds.). Cambridge: Academic Press, Inc.
Congdon, J.D., Nagle, R.D. & Kinney, O.M. (2018). Front loading
life histories: the enduring inuence of juvenile growth
on age, size, and reproducon on primiparous freshwater
turtles. Evoluonary Ecology Research 19, 353–364.
DeJesus, I. (2021). Susquehanna River contamination is
getting worse, and Harrisburg’s sewer system is being
blamed. PennLive. Downloaded on 24 September 2021.
Ennen, J.R., Cecala, K.K., Gould, P., Colvin, R., Denison, J.,
Garig, D.F., Hyder, S., Recker, L. & Davenport, J.M. (2021).
Size maers: the inuence of trap and mesh size on turtle
captures. Wildlife Society Bullen 45(1), 130–137.
Ernst, C.H. (1971a). Growth of the painted turtle, Chrysemys
picta, in southeastern Pennsylvania. Herpetologica 27(2),
Ernst, C.H. (1971b). Populaon dynamics and acvity cycles of
Chrysemys picta in southeastern Pennsylvania. Journal of
Herpetology 5(3-4), 151–160.
Ernst, C.H. & Ernst, E.M. (1971). The taxonomic status and
zoogeography of the painted turtle, Chrysemys picta, in
Pennsylvania. Herpetologica 27(4), 390–396.
Ernst, C.H. & McDonald Jr., B.S. (1989). Preliminary report
on enhanced growth and early maturity in a Maryland
populaon of painted turtles, Chrysemys picta. Bullen of
the Maryland Herpetological Society 25(4), 135–142.
Demography of a painted turtle intergrade population from an altered wetland
Ernst, C.H. & Lovich, J.E. (2009). Turtles of the United States
and Canada (second ed.), Johns Hopkins University Press,
Frazer, N.B., Gibbons, J.W. & Greene, J.L. (1991). Growth,
survivorship, and longevity of painted turtles, Chrysemys
picta, in a southwestern Michigan marsh. The American
Midland Naturalist 125(2), 245–258.
Gibbons, J.W. (1967). Variation in growth rates in three
populations of the painted turtle, Chrysemys picta.
Herpetologica 23(4), 296–303.
Gibbons, J.W. (1968). Populaon structure and survivorship in
the painted turtle, Chrysemys picta. Copeia 1968(2), 260–
Gulee, A.L., Anderson, J.T. & Brown, D.J. (2019). Inuence
of hoop-net trap diameter on capture success and size
distribuon of comparavely large and small freshwater
turtles. Northeastern Naturalist 26(1), 129–136.
Herbert, Rowland & Grubic Inc. (2015). Engineering and
Related Services. Feasibility Study Wildwood Lake, City of
Harrisburg, Dauphin County PA.
Hughes, D.F. & Meshaka Jr., W.E. (2020). Demography of
aquac turtles (Chrysemys picta marginata and Chelydra
serpentina serpentina) in southwestern Pennsylvania.
Annals of Carnegie Museum 86(4), 361–376.
Hughes, D.F., Tegeler, A.K. & Meshaka Jr., W.E. (2016).
Dierenal use of ponds by two species of aquac turtles
(Chrysemys picta marginata and Chelydra serpentina
serpenna) and their role in colonizaon. Herpetological
Conservaon and Biology 11(1), 214–231.
Hulse, A.C., McCoy, C.J. & Censky, E.J. (2001). Amphibians and
Reples of Pennsylvania and the Northeast. Ithaca: New
York. 419 pp.
Jolly, G.M. (1965). Explicit esmates from capture recapture
data with both death and immigraon stochasc model.
Biometrika 52(1–2), 225–247.
Lebreton, J.D., Burnham, K.P., Clobert, J. & Anderson, D.R.
(1992). Modeling survival and tesng biological hypothesis
using marked animals: a unied approach with case studies.
Ecological Monographs 62(1), 67–118.
Lindeman, P. (1996). Comparave Life History of Painted Turtles
(Chrysemys picta) in Two Habitats in the Inland Pacific
Northwest. Copeia 1996(1), 114.
Mitchell, J.C. (1985a). Variaon in the male reproducve cycle
in a populaon of painted turtles, Chrysemys picta, from
Virginia. Herpetologica 41, 45–51.
Mitchell, J.C. (1985b). Female reproductive cycle and life
history aributes in a Virginia populaon of painted turtles,
Chrysemys picta. Journal of Herpetology 19(2), 218–226.
Mitchell, J.C. (1988). Populaon ecology and life histories of
the freshwater turtles, Chrysemys picta and Sternotherus
odoratus, in an urban lake. Herpetological Monograph 2,
Powell, R., Conant, R. & Collins, J.T. (2016). Peterson Field Guide
to Reples and Amphibians of Eastern and Central North
America. 4th edion. Boston: Massachuses. 512 pp.
Pradel, R. (1996). Ulizaon of capture-mark-recapture for the
study of recruitment and populaon growth rate. Biometrics
52(2), 703–709.
Quinn, A.J. & Chrisansen, J.L. (1972). The relaonship between
pond bottom type and growth rate of western painted
turtles, Chrysemys picta bellii, in Iowa: a preliminary report.
Proceedings of the Iowa Academy of Science 78(3–4), 67–69.
R Core Team (2021). R: a language and environment for
stascal compung. Vienna: R-project; 2021. Available
from: hps://
Ream, C. & Ream, R. (1966). The inuence of sampling methods
on the esmaon of populaon structure in painted turtles.
The American Midland Naturalist 75(2), 325–338.
Russell, J.L., Wingert, E., Boback, S.M. & Meshaka Jr., W.E.
(2014). The Red-eared Slider, Trachemys scripta elegans
(Wied, 1838), Established in Pennsylvania. Bullen of the
Maryland Herpetological Society 50, 34–41.
Sauro, S. (2019). Raw sewage was discharged into Harrisburg
waterways on 150 days last year. PennLive. Downloaded
on 24 September 2021.
Seber, G.A.F. (1965). A note on the mulple recapture census.
Biometrika 52(1–2), 249–259.
Shearer, M. (2012). State-federal disagreement on Paxton Creek
watershed costs locals money. PennLive. Downloaded
on 20 March 2021.
Smart, J. (2019a). AquaticLifeHistory: fisheries life history
analysis using contemporary methods. hps://
Smart, J. (2019b). BayesGrowth: esmate sh growth using
MCMC analysis. R package version 0.0.9000. hps://github.
Spencer, R.J. (2002). Growth paerns of two widely distributed
freshwater turtles and a comparison of common methods
used to esmate age. Australian Journal of Zoology 50(5),
Sturges, H.A. (1926). The choice of a class interval. Journal of the
American Stascal Associaon 21(153), 65–66.
Tesche, M.R. & Hodges, K.E. (2015). Unreliable population
inferences from common trapping pracces for freshwater
turtles. Global Ecology and Conservaon 3(January), 802–
Uetz, P., Freed, P., Aguilar, R. & Hošek, J. (2021). The Reple
Database. Accessed on 16 July 2021. hp://www.reple-
White, G.C. & Burnham, K.P. (1999). Program MARK: survival
esmaon from populaons of marked individuals. Bird
Study 46(Supplement 1), S120–S139.
Wilson, D.S. & Tracy, C.R. (2003). Estimating age of turtles
from growth rings: a crical evaluaon of the technique.
Herpetologica 59(2), 178–194.
Wingert, E.W. & Meshaka Jr., W.E. (2021). A Pocket Guide to
Amphibians and Reples of Wildwood Park. Olathe: Kansas.
57 pp.
Zweifel, R.G. (1989). Long-term ecological studies on a populaon
of painted turtles, Chrysemys picta, on Long Island, New
York. American Museum Novitates 2952(August), 1–55.
W. Meshaka et al.
Accepted: 4 October 2022
ResearchGate has not been able to resolve any citations for this publication.
Full-text available
Multi-year studies of syntopic species provide a spatiotemporal framework for comparing their demographic responses to the same environmental conditions. We used data derived from 15 years of sampling at an artificial pond matrix in southwestern Pennsylvania to investigate the survival, growth, and ages of Midland Painted Turtles (Chrysemys picta marginata) and Common Snapping Turtles (Chelydra serpentina serpentina). We trapped turtles with baited hoop-nets at a primary wetland, which was the largest and deepest of five artificial ponds in a spatially aggregated matrix at the Powdermill Nature Reserve, a protected site in the Allegheny Mountains. We captured 81 Midland Painted Turtles 162 times, and 43 Common Snapping Turtles 136 times. For both species, apparent survival probabilities were higher for adults (range 79-95%) compared to juveniles (range 57-82%), and higher in females compared to males or juveniles. The average growth rate was highest in juvenile turtles of both species, indicating growth was maximal during periods of the lowest survival. Average growth rates, in general, were slower for Midland Painted Turtles compared to Common Snapping Turtles. Relating body size to age revealed estimates conforming to studies elsewhere and to longevity records based on known-age turtles. We interpret findings at this wetland matrix to represent the demographics of a deme within a fluid and dynamic regional network of demes for these two species and highlight the value of artificial pond networks to the conservation of freshwater turtle metapopulations in Pennsylvania.
Full-text available
Primary questions: (1) How do juvenile growth rates influence age and body size at maturity of females of three species of freshwater turtles? (2) Are the patterns similar among species that occupied the same wetlands over the same three decades? (3) What are the reproductive traits (i.e. clutch size and egg size) of primiparous females (first lifetime reproduction)? (4) Is there evidence that adult growth rates subsequently reduce the initial differences in the body size and reproductive traits of primiparous females? Secondary questions: We asked several additional questions of Painted Turtles. Are growth rates of older juveniles more similar to growth rates of young juveniles or adults? What is the earliest age at which juvenile growth rate is detectably correlated with age and body size of primiparous females? Sample sizes of the other two species were too small to use for these questions. Organisms: Three long-lived freshwater turtle species: Painted Turtles (Chrysemys picta marginata), Blanding's Turtles (Emydoidea blandingii), and Snapping Turtles (Chelydra serpentina). Field site: University of Michigan, E.S. George Reserve, southeastern Michigan, USA. Methods: We conducted a 33 year mark-recapture study to document juvenile and adult growth rates and age and body size at maturity of females. We used X-radiography to determine clutch size and egg widths of primiparous and older females of all three species. Conclusions: (1) Juvenile growth rate was the most influential trait determining within-population variation in life-history trait values of primiparous females of all three species of long-lived freshwater turtles, and that variation persisted for many years in older adults. (2) Fast-growing juveniles of all three species matured earlier and at larger (or similar) body sizes than slow-growing juveniles. (3) The relationship between juvenile growth rates and age and size at maturity in Painted Turtles was established by age 4 years. (4) Variation in indeterminate (post-maturation) growth was insufficient to reduce differences in reproductive traits within cohorts of females. (5) Similar results from all three turtle species (families Emydidae and Chelydridae) suggest that the relationships between juvenile growth rates and age and size at maturity were established in a common ancestor early in the evolutionary history of turtles.
Full-text available
We examined turtle populations occupying eight artificial ponds in Westmoreland County, southwestern Pennsylvania, USA. Beginning in 2005, we used sardine-baited hoop-nets to trap turtles for eight consecutive years at one pond. In 2013, we expanded sampling to include seven additional ponds near the primary study pond. We deployed and checked traps during two 5-d periods at all eight ponds, once in June and again in July 2013. Two of the 12 turtle species native to Pennsylvania were detected, Common Snapping Turtle (Chelydra serpentina serpentina; n = 53) and Midland Painted Turtle (Chrysemys picta marginata; n = 70). We found the Common Snapping Turtle at all surveyed ponds and its abundance was associated with larger, sparsely vegetated ponds. We found the Midland Painted Turtle at six of the eight surveyed ponds and its abundance was associated with smaller, heavily vegetated, ponds. Juveniles of both species were distributed differently than adults and were most common in shallow, heavily vegetated ponds with low visibility and were absent or nearly so from deep ponds with little emergent vegetation favored by adults. Across eight years, the number of juveniles was low or they were absent from the primary study pond, yet recruitment was likely maintained through a nearby nursery pond favored by juveniles. In this heterogeneous pond matrix, turtle population structures were strongly influenced by certain physical features of the ponds to the benefit of one life stage over another. Species composition was influenced in a likewise manner. Inter-pond movements were likely encouraged by local habitats and resident population structures. We suggest that ponds can be constructed or modified to accommodate one or more life stages of these turtle species, and enhance opportunity for pond colonization and gene flow among ponds.
Methods used in wildlife ecology can influence population‐ and community‐level estimates, such as species richness, sex ratio, age and size structure, occupancy and detection probabilities, and community composition. Various trapping and sampling biases exist for freshwater turtles including bait and trap choice and survey technique. To date, no study has investigated the influence of hoop net and mesh size on various population‐ and community‐level estimates. Here, we use detection models to determine if trap and mesh size influence detection probability of nine species of freshwater turtles over 3 consecutive years (2016–2018) in west Tennessee. Additionally, we use multivariate models to determine if freshwater turtle community composition was influenced by hoop trap and mesh sizes. Our results indicated that there was a bias related to mesh size in detection probabilities and community composition. Smaller mesh sized traps were better at detecting smaller‐bodied turtle species, which then changed community species richness but not catch‐per‐unit‐effort estimates (i.e., abundance). Additionally, larger mesh sized traps were better at detecting common snapping turtles (Chelydra serpentina), which supports previous research. Our results suggest that researchers should account for the variation in detection probabilities by mesh size when conducting mark‐recapture and occupancy analyses. Moreover, erroneous inferences about population trends and changes in diversity within turtle communities through time could cause managers to misidentify population declines and conservation value of a site. © 2021 The Wildlife Society. Mesh size of hoop nets causes variation in turtle community indices along with occupancy and detection probabilities.
This chapter gives results from some illustrative exploration of the performance of information-theoretic criteria for model selection and methods to quantify precision when there is model selection uncertainty. The methods given in Chapter 4 are illustrated and additional insights are provided based on simulation and real data. Section 5.2 utilizes a chain binomial survival model for some Monte Carlo evaluation of unconditional sampling variance estimation, confidence intervals, and model averaging. For this simulation the generating process is known and can be of relatively high dimension. The generating model and the models used for data analysis in this chain binomial simulation are easy to understand and have no nuisance parameters. We give some comparisons of AIC versus BIC selection and use achieved confidence interval coverage as an integrating metric to judge the success of various approaches to inference.