of the
FLORIDA
MUSEUM OF
NATURAL HISTORY
REPRODUCTION IN THE SUWANNEE COOTER,
Pseudemys concinna suwanniensis
Dale R. Jackson and Robert N. Walker
Volume 41 No. 2, pp. 69-167
UNIVERSITY OF FLORIDA
GAINESVILLE
I LL..
Numbers of the BULLETIN OF THE FLORIDA MUSEUM OF NATURAL HISTORY are
published at irregular intervals. Volumes contain about 300 pages and are not necessarily completed in
any one calendar year.
JOHN F. EISENBERG, EDITOR
RICHARD FRANZ, CO-EDITOR
RHODA J. BRYANT, MANAGING EDITOR
Communications concerning purchase or exchange of the publications and all manuscripts should be
addressed to: Managing Editor, Bulletin; Florida Museum of Natural History, University of Florida;
P. O. Box 117800, Gainesville FL 32611-7800; U.S.A.
This journal is printed on recycled paper.
ISSN: 0071-6154
CODEN: BF 5BA5
Publication date: December 31, 1997
Price: S 8.00
- -~- -
Plate 1. River Road, the major nesting site ofP. concinna at Wakulla Springs State Park. The river lies ca 40
m northeast (left) of the road at this point. A turtle can be seen nesting in the distance on the right shoulder.
Plate 2. Adult female Pseudemys concinna from Wakulla Springs State Park. Above (facing page): lateral
aspect; Below: ventral aspect.
<*
F4
p ,-.-
k,-
Plate 3. Carcass of female Pseudemys. concinna depredated during terrestrial excursion for nesting at
Wakulla Springs State Park.
REPRODUCTION IN THE SUWANNEE COOTER,
Pseudemys concinna suwanniensis
Dale R. Jackson' and Robert N. Walker-
ABSTRACT
The Suwannee cooter, Pseudemys concinna suwanniensis, is the largest emydid turtle in North
America. From 1988 to 1993, we studied the reproductive and nesting biology of a population of this turtle
in the headwaters of the Wakulla River in northern Florida. Based on 239 marked females, we estimated a
population of about 305 adult females in ca 5 km (41 ha) of river. Nesting females ranged in plastron length
(PL) from 304 to 383 mm and in body mass from 4.5 tolO.5 kg. We estimate that female body size at
maturity varies by as much as 50 mm, with a mean size at maturity of ca 330 mm PL, or only about 1 cm
less than the mean PL The negligible post-maturational growth of adult females, coupled with their size
range, suggests that sexual maturity is less a function of size than of age, which we estimate minimally at 10
years.
Nesting begins in late March or early April and extends into early August. Females may lay as many
as five or more clutches at intervals of approximately 16-25 days. Most, if not all, mature females nest
annually. Nesting is diurnal and frequently coincides with rainfall. Typical clutches contain 8-27 eggs
(mean 17.5). Average annual reproductive potential may approach 70 eggs, with a maximum for
individuals of>100. Normal egg and hatchling sizes span ranges of 9.8-21.7 g and 6.8-14.6 g, respectively,
and are, at most, only weakly correlated with female body size. Clutch mass tends to increase with body
size, but relative clutch mass, which averages ca 0.05, does not. These results suggest this species is an egg
number maximizer rather than an egg size optimizer. We estimate for the population an annual production
of ca 26,700 eggs equivalent to 10.6 kg/ha/yr. Hatchlings may emerge from nests in the fall or may
overwinter underground and delay emergence until the following spring. Sex determination of hatchlings is
temperature-dependent, with a pivotal temperature at constant incubation temperatures of ca 28.4C.
Although this population occurs in a state park, the concentration of nesting activity principally along
a linear corridor of artificial habitat has fostered nest predation approaching 100 percent; chief egg predators
are raccoons (Procyon lotor) and fish crows (Corvus ossifragus). Some nesting females are also
depredated, principally by raccoons. Continuing forest growth threatens to shade out the main nesting site
and to skew the sex ratios ofhatchlings toward males. We present management recommendations both for
this population and for the species throughout Florida.
RESUME
La tortuga del Rio Suwannee (Pseudemys concinna suwanniensis) es la tortuga emydida mis grande
de Norteam6rica. Entre 1988 y 1993, estudiamos la biologia reproductive y de anidamiento de una
poblaci6n de esta tortuga en las cabeceras de agua del Rio Wakulla, en el norte de la Florida. En base a 239
hembras marcadas, estimamos la poblaci6n en cerca de 305 hembras adults en aproximadamente 5 km (41
ha) de Rio. Las hembras anidantes variaron en largo de plastr6n (LP) entire 304 y 383 mm y en masa
corporal entire 4.5 a 10.5 kg. Estimamos que, al alcanzar la edad madura, el tamalfo de las hembras vari6 50
mm como miximo siendo en promedio aproximadamente 330 LP, cerca de 1 cm menos que el LP promedio.
El minimo crecimiento que exhiben las hembras despues de alcanzar la madurez, junto con su rango de
tamailo, sugiere que la madurez sexual es menos una funci6n del tamailo que de la edad, la cual estimamos
ser como minimo 10 afnos.
' Florida Natural Areas Inventory, The Nature Conservancy, 1018 Thomasville Road, Suite 200-C, Tallahassee, FL 32303, USA, and Florida
Museum of Natural History, Gainesville, FL 32611, USA
2 1430 East Randolph Circle, Tallahassee, FL 32312.
JACKSON, D. R., AND R. N. WALKER. 1997. Reproduction in the Suwannee cooter, Pseudemys
concinna suwanniensis. Bull. Florida Mus. Nat. Hist. 41(2):69-167.
70 BULLETIN FLORIDA MUSEUM NATURAL HISTORY VOL. 41(2)
El anidamiento comienza a finales de marzo o principias de abril, y se extiende hasta principios de
agosto. Las hembras pueden poner huevos cinco o mAs veces cada 16-25 dias. La mayoria, sino todas las
hembras maduras, anidan cada afio. El anidamiento es diumo y frecuentemente coincide con la caida de
Iluvia. Una puesta tipica contiene entire 8 y 27 huevos (promedio 17.5). El potential reproductive annual
puede llegar a 70 huevos, con un miximo para algunos individuos de >100. El tamailo normal de huevos y
reci6n nacidos estan como much, s61o d6bilmente correlcionados con el tamafio corporal de la hembra, y
varian entire 9.8-21.7 g y 6.8-14.6 g, respectivamente. La masa de la puesta de huevos tiende a aumentar
con el tamaflo corporal de la hembra, pero no la masa relative de la puesta cuyo promedio es cerca de 0.05.
Los resultados obtenidos sugieren que esta especie tiende a maximizar el n6mero de huevos mAs que a
optimozar el tamafio de estos. Estimamos que esta poblaci6n produce anualmente 26,700 huevos, que
equivalent a 10.6 kg/ha afio. Los huevos pueden eclosionar en el otoflo o retrasarse hasta la pr6xima
primavera. La determinaci6n del sexo de los reci6n nacidos es dependiente de la temperature, siendo la
temperature pivotal de incubaci6n a temperature constant de cerca de 28.4'C.
A pesar de que la poblaci6n se encuentra en un parque del estado, la predaci6n de nidos se acerca al
100%, debido a que la actividad de anidamiento se concentra principalmente a lo largo de un corridor de
habitat artificial. Los mayores predadores de huevos son mapaches (Procyon lotor) y cuervos pescadores
(Corvus ossifragus). Algunas hembras son tambi6n predadas durante el anidamiento, principalmente por
mapaches. El continue crecimiento del bosque amenaza con sombrear el mayor sitio de anidamiento
produciendo un sesgo de la raz6n de sexos hacia los machos. Presentamos recomendaciones de manejo para
esta poblaci6n, asi como para la especie a lo largo de la Florida.
TABLE OF CONTENTS
Introduction ........................................... ....................................................................... 71
A know ledgem ents .............................. ................................ ......................................... .......................... 74
Study Sites...................................................... .............................. .......................... 77
M ain study site: W akulla River............................. ............................................................... 77
M inor study sites................................ ....................................................................................... 81
C lim ate...................... ................................................................................................................... 82
Taxonomic Allocation of Wakulla River Pseudemys .............................................. .............................. 82
M methods ................................................................................ .......... ............................ 83
D ata Collection........................................................................ ........................ 83
Statistical Analysis............................... ......................... 88
Results ...... ......................................................................................... ...................... 89
G general ............................................................................................................. 89
Home Range and Homing....................................................................... 89
N testing Season ............................................ .................. .................................. ...................... ......... 92
N testing Behavior.................................................. .......................... ........................... 94
Reproductive Parameters...................................... ......................... ........................... ... 106
Development, Sex Determination, and Hatchlings................................ ............................. 114
Predation and M mortality ..................................................................................................................... 130
Adult Female Population Size, Biomass, and Productivity.......................................................... 133
Male Body Size and Reproductive Cycle.................................................................................... 135
M miscellaneous Observations ........................ .................. ...... .................................... 135
Discussion. ..................................................................... ... 136
General Life H history Strategy........................................................ ...................................... ....... 136
Hom e Range and H om ing.................................................... ..................................... ........... 137
Reproductive Seasonality.................................................................. .................................................. 138
Diel Nesting Cycle, Proximate Nesting Cues, and Body Temperature During Nesting................... 140
Reproductive Param eters.................................................................................................. .................. 142
Development, Hatchlings, and Emergence from Nests................................................................... 147
Nest Site Selection, Site Fidelity, and Sex Determination.......................................................... 149
M ortality......................................................................................... ................... ... .................... 150
Population Size, Biomass, and Production........................................................................................ 154
Management Recommendations.......................................................................................... 156
Literature C ited ........................................................................................... ............... ...... ............... 159
JACKSON & WALKER: REPRODUCTION IN SUWANNEE COOTER
INTRODUCTION
The greatest number of sympatric freshwater emydid turtles in North America
occurs along the northern coast of the Gulf of Mexico in northern Florida and
adjacent Alabama. Except for the genus Graptemys, which in Florida occurs
exclusively in rivers of the western panhandle, most of these emydids are adapted
for existence in non-flowing waters. D. Jackson (1988) compared the reproductive
biologies of the four principal species (Pseudemysfloridana, P. nelsoni, Trachemys
scripta, and Deirochelys reticularia) inhabiting North Florida lentic environments.
This study examines the reproductive and nesting biology of a fifh, closely related
species that is confined to lotic situations in the same region.
According to most authorities, the river cooter, Pseudemys concinna, is a
distinct species most closely related to P. floridana and consists of several
recognizable subspecies (Ward 1984; Conant and Collins 1991; Jackson 1995). P.
c. suwanniensis, the Suwannee cooter, is the largest emydid turtle in North
America (female max PL ca 37 cm), yet it has one of the smallest ranges,
extending from approximately the center of panhandle Florida east and south in
peninsular Florida to the vicinity of Tampa Bay (Jackson 1992; Ernst et al. 1994;
we do not follow recent suggestions that this population is specifically distinct; see
below). Because of its small range and declining numbers as a result of human
predation (Carr 1983), the Suwannee cooter is considered a Species of Special
Concern by the State of Florida (Wood 1996).
Suwannee cooters typically inhabit flowing waters or associated
impoundments. Relative to other local emydids, this turtle exhibits several
morphological adaptations for riverine existence (Auffenberg 1978; Jackson 1992):
a smoother, thinner, more streamlined shell; connection of skin nearer the
periphery of the shell (reducing turbulence); and very large, extensively webbed
hind feet. Riverine habitats occupied characteristically are one of two types: (1)
rivers characterized by dark waters (from sediment and/or tannins), seasonally low
temperatures, and frequent flooding; and (2) calcareous spring runs of remarkably
clear water emanating from the underlying aquifer that are highly stable both in
terms of temperature (19-230C, mean 21C in northern Florida) and water level,
being relatively independent of the vagaries of rainfall (Crenshaw 1955; Rosenau
et al. 1977; Giovanetto 1992; Hubbs 1995). Abundance of aquatic plants is
distinctly seasonal in the former habitat and more stable in the latter. This may be
ecologically important to P. c. suwanniensis as post-juvenile turtles (and perhaps
all age groups) of this subspecies are exclusively herbivorous (LOennberg 1894;
Allen 1938; Marchand 1942; Carr 1952; C. Jackson 1964; Ward 1980; Lagueux et
al. 1995). In Florida, P. concinna is often the numerically dominant emydid where
it occurs, although in some rivers its numbers are exceeded by map turtles
(Graptemys spp.). Within the state, it may coexist with small populations of any of
the aforementioned species of emydids, all of which occur in some lotic situations.
Like map turtles, but in contrast to the more lentic aquatic emydids, P. concinna
rarely wanders terrestrially except to nest.
Table 1. Previously reported reproductive parameters ofPseudemys concinna and presumed sister taxa; data based on samples > 1 given as means, with range and n in
parentheses. Relative clutch mass (RCM) is defined as clutch mass/body mass.
Adult Female
Species Mass PL Clutch Clutch Clutches Nesting
Locality (kg) (mm) Size Mass RCM Per Year Season Source
P. concinna
Alabama:
Tallapoosa River
Florida:
northern peninsula
Wacissa River
Georgia:
Savannah River4
Illinois:
Wabash River'
Louisiana:
southern6
Missouri:
southern
292 17.6 307
(266-322, 18) (15-20, 10) (261-377, 10)
17.7
(17-19,3)
3.0 281
1-21 late May-
mid June
late April -
May +2
early June 3
203.8
0.07
14.0
ca. 2807
12.2
(7-16, 5)
3.18 270s 19.7 248.6
(1.6-4.5, 36)8 (233-304, 36)s (12-28, 10)
Tennessee:
western4
Virginia
9
15.610
(12-20, 4)
198.110
0.09 2.4
(1-4?)
Fahey (1987)
Jackson and Jackson (1968)
Cagle (1955)
Congdon and Gibbons (1985)
Moll and Morris (1991)
mid-June ' Hart (1979); Fahey (1980);
Dundee and Rossman (1989)
mid-April -
mid June
(cold: June-
August)9
- late May-June
2-3?" mid-June -
early July"
Turner (1995)
Cahn (1937)
Mitchell (1994)
West Virginia:
New River
3.1 289
(2.3-4.0, 9X265-315, 9)
June/
July'2
P. texana
Texas:
Colorado River
San Marcos River
4-19"3
Seidel (1981); Buhlmann and
Vaughan (1991)
mid-June Whiting (1994)
early April- Rose et al. (1996), F. L.
August Rose, pers. cotntn.
(with May peak)
ca. May-June Vermersch (1992)
P. gorzugi
Texas
Degenhardt et al. (1996)
Although Fahey (1987) failed to note it, one of his largest females (#305, table 4) bore evidence of a potential third clutch.
2Jackson and Jacksons (1968) report of eggs from 5 Aug and 12 Sep may be in error, discrepancies exist between dates in the text and those in both tables. It is possible that hatching dates of some eggs may have
been used inadvertently.
based on one gravid female collected 7 June
based on one gravid female; RCM based on female gravid mass
based on one recent nest discovered 16 June
taxonomic reassignment based on Ward (1984)
based on one gravid female collected 12 June
based on combined samples in Table 2 of Turner (1995); data conflict with text and may include some immature turtles
nesting season delayed 2 months in one population inhabiting unnaturally hypothermic water
o based on four clutches
Based on two gravid females collected 16 June and 1 July, the former bearing two sets of corpora lutea and one set of enlarged follicles
z based on two females, one collected gravid 17 June, one pre-gravid in May
3 based on secondary reports
" based on one clutch
BULLETIN FLORIDA MUSEUM NATURAL HISTORY VOL. 41(2)
Relative to other freshwater emydid turtles of the southeastern United States
(Jackson 1988), the reproductive biology of river cooters is poorly known. Carr
(1952), Ernst and Barbour (1972), and Ernst et al. (1994) reviewed the limited
earlier data available for the species. Except for a pair of studies in central
Alabama (Fahey 1987) and southern Missouri (Turner 1995), existing data are
confined to scattered reports based chiefly on only one to three clutches of eggs
(Tables 1 and 2). The dearth of information extends to P. c. suwanniensis.
Pritchard (1979) characterized its nesting habits as "completely cryptic," and
Auffenberg (1978) misrepresented its basic nesting season. Remaining literature
describes only selected life history aspects: eggs and hatchlings (Jackson and
Cantrell 1964; Jackson and Jackson 1968; Ewert 1979); courtship behavior
(Marchand 1944; Cagle 1955; Jackson and Davis 1972); nest structure (Carr
1983); and growth and sex ratio (Marchand 1942; Jackson 1964, 1970). The
present study attempts to fill this void in our growing knowledge of North
American turtle life history patterns.
Following preliminary reconnaissance beginning in 1985, we initiated a
multi-year study in 1988 of a single, locally dense population of this species in the
lower Gulf Coastal Plain of the Florida panhandle (Wakulla River). We
supplemented our results with necropsy-based data collected during the previous
decade on two other North Florida rivers.
As the present study was being initiated, Fahey (1987) presented the results of
a life history study of P. concinna in the Tallapoosa River in the upper Piedmont of
East-central Alabama (latitude 33059'N). Although his sample of nesting (n=14)
and gravid females is much smaller than ours, his data are sufficient to allow a
preliminary intraspecific geographic comparison of life history parameters across a
latitudinal range of nearly 40 (ca 300 km). Additional recent data from
populations in West Virginia (Buhlmann and Vaughan 1991) and Missouri
(Turner 1995) enable comparison across an even greater range.
Our knowledge of riverine turtle ecology in general seems to have lagged
behind that for species occupying lentic, and even marine and terrestrial, habitats.
The present study, coupled with others that have been conducted in the last two
decades (e.g., Shealy 1976; Plummer 1977; E. Moll 1980; Vogt 1980; Pluto and
Bellis 1986, 1988; Dodd et al. 1988; Georges and Kennett 1989; D. Moll 1989;
Ewert and Jackson 1994; Jones and Hartfield 1995; Jones 1996; Polisar 1996),
helps to fill this void. This is especially important, as many of the earth's riverine
turtles are not only economically important but are becoming increasingly stressed
by human exploitation and habitat degradation.
ACKNOWLEDGEMENTS
We extend our gratitude to the Florida Park Service, Department of Environmental Protection, for
permission to conduct research at Wakulla Springs State Park, and to the park staff for their cooperation and
assistance throughout the project. Richard Miller, Sandy Cook, Ronald Weiss and their staff-particularly
biologists Jon Dodrill and Alan Whitehouse-diligently completed forms and kept tabs on turtles above and
beyond the call of duty. We are especially indebted to C. Kenneth Dodd, Jr., for invaluable
Table 2. Previously reported data for eggs and hatchlings ofPseudemys concinna and presumed sister taxa; sizes given are means, with ranges and n in parentheses.
Eggs and hatchlings were not necessarily measured at identical stages by all authors.
Locality Eggs Hatchlings
Length (mm) Width (mm) Mass (g) CL (mm) PL (mm) Mass (g) Source
P. concinna
Alabama: Tallapoosa Rivert 40.4 27.5 17.4
(36.6-44.8, 176X24.9-30.1, 176) (13.4-21.5, 176)
Florida: northern
peninsula
Georgia: Savannah River2
Illinois: Wabash River?
Louisiana: southern
Missouri: southern4
Tennessee: Reelfoot Lake5
P. texana
Texas: San Marcos River6
Texas
P. gorzugi
39.0
(32.9-43.3, 72)
36.4
38.2
37-44
36.3
(30.2-44, 159)
37.6
(35-42, 9)
26.7
(24.4-30.5, 72)
23.9
25.4
26-29
24.6
(22-28, 159)
26.1
(22-28,9)
19.9
(16.4-22.4, 17)
12.0
15.4
12.6
(9.4-17, 159)
39.5
36.0
(32.7-38.9, 54)
35.3
33-36
31.0
36.2
32.4
(29.2-35.3, 54)
34.0
(23-34, 55)
12.0 Fahey(1987)
12.3
(10.4-13.8,38)
10.3
9.0
(6-13, 55)
40.95 27.6 16.3
(39.2-43.3; 14) (26.2-29.4; 14) (14.5-18.9; 14)
ca. 35 -
no data available
Jackson & Jackson (1968)
Congdon & Gibbons (1985)
Moll and Morris (1991)
Dundee & Rossman (1989)
Turner (1995)
Cahn (1937)
Rose et al. (1996)
Vermersch (1992)
egg data from 10 clutches, hatchling data from four, except mass from only two; all hatchlings measured after removal from nests, long after actual hatching and probably after some linear growth from yolk assimilation
2 based on 15 eggs from one clutch
based on one clutch of nest eggs and three resulting hatchlings
based on combined samples from two sites
based on one clutch of 9 eggs
based on 14 eggs from two clutches
0 40 80 f20 KM
Figure 1. Map of northern half of Florida showing rivers from which samples were taken: Wakulla River
(WAK), Santa Fe River (SFR)/Suwannee River (SUW), and Withlacoochee River (WIT).
JACKSON & WALKER: REPRODUCTION IN SUWANNEE COOTER
assistance with data analysis and presentation and to Amy Knight and Bert Charest for data entry. Michael
Ewert and Cory Etchberger incubated some of the eggs, and Ewert performed the analysis of sex
determination and provided a painstaking review of the manuscript. Scott Taylor prepared Figure 2, and
Brad Mueller constructed the radiotransmitters and offered helpful advice about radiotelemetry. Richard
Franz lent us a Telonics receiver, and William Dunson provided two Schultheis thermometers and offered
knowledgeable insight into turtle physiology during a day in the field. Additional field assistance was
provided by Cynthia Lagueux, Hank Smith, Scott Savery, Russell Burke, Thomas Ostertag, George
Weymouth, and James Solomon. Kent Ainslie collected most of the Santa Fe and Withlacoochee River
samples. Gary Knight and Katy NeSmith answered queries about plants and birds, respectively, and Richard
Seigel kindly shared his perspective about body mass and relative clutch mass in reptiles. Input about
estimating population size was garnered through helpful communications with Peter Lindeman, Robert
Jones, C. Kenneth Dodd, and Stephen Corn. Paul Moler and Susan Seyboldt researched the regulatory
chronology of river cooters in Florida, and Francis Rose provided pertinent literature. Funding for this study
and publication of the results was provided by the Florida Game and Fresh Water Fish Commission's
Nongame Wildlife Program (contracted project NG88-045); DRJ expresses his sincere thanks to agency staff
for bearing with a one-year study that evolved into an intensive 10-year effort. The Commission also
permitted the possession of eggs and turtles throughout the study. Work during the 1990 field season was
greatly facilitated by a sabbatical leave granted to the senior author by The Nature Conservancy.
STUDY SITES
The total sample represents three drainages (Fig. 1) and two time periods.
Small samples of specimens of both sexes were available for gonadal examination
during the years 1972-1975 from the Santa Fe and Withlacoochee rivers in
northern peninsular Florida. Following preliminary reconnaissance and data
collection during 1985-1987, a detailed study of the species' nesting ecology was
conducted from 1988 through 1991 at Edward Ball Wakulla Springs State Park
(WSSP) on the Wakulla River in the Florida panhandle; limited data were
collected in 1992 and 1993.
Main Study Site: Wakulla River
Lying entirely within the flat coastal lowlands bordering the northern Gulf of
Mexico, the Wakulla River (3013.5'N, 8417'W; Wakulla County) is a seasonally
clear, calcareous freshwater spring run arising from the Floridan aquifer. The
river issues chiefly from Wakulla Springs, one of the largest of Florida's 27 first
magnitude artesian springs, with an average discharge of 663 m3/min (1907-1974
average). Although >90 percent of the river's water normally is of subterranean
origin, the rate of discharge exceeds that of any other Florida spring and correlates
closely with rainfall. River depth is shallow, normally averaging about 1 m,
although the spring floor itself slopes to a depth of 55 m, where it continues into an
even deeper cavern. Water temperature remains a fairly constant 21C near the
spring but increases by 1-4C a few km downstream; pH is slightly alkaline, with
average readings of 7.3-7.9, as a result of the high level of calcium bicarbonate.
The river bottom consists chiefly of sandy silt interrupted by outcroppings of
limestone (Rosenau et al. 1977; Thompson 1977; Rupert and Spencer 1988).
From the main spring pool, the river flows southeastward across limestone
bedrock for 15.2 km until it joins the dark, tannic waters of the St. Marks River,
BULLETIN FLORIDA MUSEUM NATURAL HISTORY VOL 41(2)
which flows southward an additional 8.3 km to the Gulf of Mexico. Although tidal
fluctuations are measurable as far upstream as Wakulla Springs, the head of fresh
water prevents salt water intrusion, even though the aquifer from which it rises is
far below sea level. The study site, Wakulla Springs State Park (WSSP; Fig. 2),
includes approximately the upper 5.14 km of the main spring and river, as well as
the lower sections of two smaller spring-run tributaries: Sally Ward Slough, which
flows ca 1.1 km from the northwest and enters 180 m downstream of the main
spring; and McBride Slough, which arises on the north side of the river and joins it
3.2 km below the main spring. In its upper reaches, the river averages 120-150 m
Figure 2. Map of Wakulla Springs State Park study area. RR, River Road, with segment number indicated
at 1000-m intervals; A and B, lawns associated with public areas; MS, McBride Slough; SW, Sally Ward
Slough. Main spring is just above the letter A; river and major islands shown in white, bordered by
floodplain swamp.
JACKSON & WALKER: REPRODUCTION IN SUWANNEE COOTER
wide, but it reforms into a series of narrow, braided channels that flow among
several forested islands at the southeastern end of the study area. Beyond the
islands, the river re-expands and remains open until its confluence with the St.
Marks River. Limited dredging of the Wakulla River prior to 1972 left behind a
series of low spoil banks, now heavily vegetated. Within the boundaries of WSSP,
the surface area of the river (including the main spring and Sally Ward Slough but
excluding the narrow floodplain) is approximately 41 ha.
The relatively constant flow of the Wakulla River and its independence from
surface drainage produce a nearly stable water level. This in turn eliminates the
marked seasonal variations in availability of basking sites and food supply that can
affect riverine turtles elsewhere (e.g., Plummer and Shirer 1975; Shealy 1976;
Pluto and Bellis 1986). Annual water level fluctuations in the river normally do
not exceed 0.91 m annually, with lowest levels (and peak water clarity) usually
occurring during the relatively dry winter (Ledbetter 1991). Extended periods of
abnormally high rainfall not only raise the water level nearly 1 m (to an average
depth of 2 m) but also temporarily darken the river (from a few weeks to several
months) as a result of percolation of tannins from the surrounding forest into the
groundwater. Unusually dry (1988) and wet (1989, 1991) years caused water
levels in the upper Wakulla River to vary as much as 2-3 m during this study.
The largely unshaded river contains a lush growth of filamentous algae and
rooted aquatic vegetation, which form the bulk of the cooter's diet (Marchand
1942; Lagueux et al. 1995). Dominant species are eelgrass (Vallisneria
americana), strapleaf sagittaria (Sagittaria kurziana), green algae (Cladophora
sp.), coontail (Ceratophyllum demersum), and the exotic Brazilian elodea (Egeria
densa); the last has been estimated to cover as much as 60 percent of the water
surface (Ledbetter 1991). Emergent species occurring in the shallows include
pickerelweed (Pontederia cordata), softstem bulrush (Scirpus validus), and water
hemlock (Cicuta mexicana). Numerous old-growth bald cypress (Taxodium
distichum) occur on small, mid-channel islets, and a variety of shrubs, vines, and
small trees (including wax myrtle, Myrica cerifera; willow, Salix sp.; buttonbush,
Cephalanthus occidentalis; red maple, Acer rubrum; poison ivy, Rhus radicans;
and climbing hempweed, Mikania scandens) cover the low channel dredge banks
(Bryan 1981). Although unshaded basking structures such as logs and stumps are
limited, turtles frequently use the dense mats of aquatic vegetation for resting and
basking. Where the river braids its way through the lower island complex, current
is reduced, shading is extensive, and much of the bottom is silty and devoid of
vegetation. Several kilometers downstream from the park, the river is
characterized by gradually increasing levels of salinity and a concomitant increase
in salt-tolerant emergent marsh vegetation (Thompson 1977).
Lining the river is a narrow band of floodplain swamp dominated by Carolina
ash (Fraxinus caroliniana), bald cypress, and tupelo (Nyssa spp.). This grades into
a floodplain forest characterized by American beech (Fagus grandifolia), southern
and sweetbay magnolias (Magnolia grandiflora and M. virginiana), red bay
(Persea borbonia), red maple, laurel oak (Quercus laurifolia), American holly
BULLETIN FLORIDA MUSEUM NATURAL HISTORY VOL. 41(2)
(Ilex opaca), spruce pine (Pinus glabra), and sweetgum (Liquidambar styraciflua).
The river, sloughs, floodplain swamp, and floodplain forest, together comprising
about one-fourth of the park, generally lie below 3 m in elevation. Land
immediately north of the river is generally lower than that on the south, with much
of the former falling below the 100-year floodplain level of 3.66 m. About 1-2 m
above the floodplain, the community grades into an upland mixed forest of
hardwoods and pines, with some remnant stands of longleaf pine (Pinus palustris)
sandhills and flatwoods vegetation. The latter community types suffer from
hardwood encroachment as a result of decades of fire suppression. Maximum
elevation within the park is 10 m.
A single-lane sand road (Fig. 2, Plate 1), consisting mostly of two bare-sand
tire tracks separated by a grassy median, roughly parallels the south side of the
river within the park below the "public area" (about 5 ha developed for recreation
and lodging by the main spring and uppermost river). Constructed between 1952
and 1972 (based on aerial photographs), the road is situated for most of its length
just above the elevational rise that marks the ecotone between floodplain forest and
upland mixed forest communities. Because of irregularities in the width of the
floodplain, as well as in the exact position of the road relative to the ecotone,
"River Road" lies anywhere from 20 m to 270 m from the normal water line on the
near (southern) edge of the river. During the wettest periods of this study (1989:
wettest June since 1900; 1991: annual rainfall total ca 1 m above normal by late
July), standing water levels reached within 5 m of the road in low spots, and water
temporarily overwashed the road during at least two exceptionally heavy storms.
River Road, which our preliminary reconnaissance suggested to be an important
nesting area, is bordered on each side by a periodically mowed grassy shoulder, 3-4
m wide, and a shallow ditch for drainage. Situated in the adjacent forest on the
river side midway down River Road, a small, nearly vertical limestone solution
hole, ca 1.5 m across and 3 m deep, presents a potential hazard to turtles emerging
to nest in that vicinity.
Although small dirt roads also exist on the north side of the river, they are
generally unsuitable for nesting because they lack sufficient elevation or are too
shaded; only a few nests are constructed annually in two sandy openings, one on
either side of McBride Slough. Other minor nesting areas within the park vicinity
include the grassy lawn around the lodge and other public buildings on the south
side of the main spring (subdivided into segments A and B, the latter contiguous
with River Road segment 1; see Fig. 2 for numeric designation of road segments),
an 0.8-ha mowed wildlife observation clearing between the river and River Road
(at road segments 7 and 8), the mowed clearing surrounding a boat maintenance
warehouse on River Road and extending to the river along a short section of
railroad track (for dry-docking boats) in segment 5 (ca 1.2 km below the main
spring), the mowed shoulders of the park entrance road where it crosses Sally
Ward Slough, and the road shoulders of two highways that cross an intermittent
spring-fed slough upstream of and tributary to Sally Ward Slough. The south side
of the main spring is maintained as an artificial sand beach, which is separated
JACKSON & WALKER: REPRODUCTION IN SUWANNEE COOTER
from the grassy lawn of the lodge by a chain-link fence that continues
approximately 200 m downstream to the entrance of the "sanctuary," that portion
of the park that is generally closed to the public. Turtles emerging from the main
spring and upper river to nest generally passed (sometimes with difficulty) through
the fence in both directions via depressions beneath the fence or between a pair of
gates that did not close tightly.
From 1935 until its acquisition by the State of Florida in 1986, the 1167-ha
tract comprising WSSP was owned and managed as a private wildlife preserve and
tourist attraction by Edward Ball, who prohibited hunting, fishing, diving, and
access to the waterway and surrounding woodlands. Thus, it is possible that many
of the resident wildlife species, including the Suwannee cooter, exist here at levels
near their carrying capacities. This may not always have been the case, as
archeological evidence indicates the presence of intermittent human habitation of
the site for at least 10,000 years (Florida Department of Natural Resources 1991).
Recent ecologically significant human disturbances include past fire suppression in
the uplands adjacent to the river, introduction of exotic aquatic plants, pollution
runoff, and some alligator (Alligator mississippiensis) removal.
Minor Study Sites
The Santa Fe River (ca 2950'N, 82042'W) is a blackwater (tannic) tributary
of the Suwannee River, a major Gulf Coast drainage of northern peninsular Florida
and southern Georgia. Although receiving substantial input from artesian springs,
the entire Suwannee system is subject to extensive water level changes determined
by seasonal precipitation patterns. A sample of six adult (four female) and eight
immature specimens was necropsied from several sites, mostly springs along the
Columbia/Gilchrist/Alachua county lines.
The Withlacoochee River (28059'N, 82021'W, Citrus/Marion county line),
source of seven adult (six male) and four immature specimens, flows northward
then westward until it empties directly into the Gulf of Mexico (in contrast to a
river of the same name that is tributary to the Suwannee River). The river's dark
waters receive input both from spring runs and surface drainage. Some
observations of nesting sites and nest predation were made from 1973 to 1975 at
the Rainbow River (29004'N, 82026'W, Marion County), a spring-run tributary of
the Withlacoochee River and principal site of Marchand's (1942, 1944, 1945)
seminal cooter studies as well as of later work by Giovanetto (1992) and Meylan et
al. (1992).
The Wakulla River is ca 160 km WNW of the Santa Fe, which is located ca
100 km NNW of the Withlacoochee. All of the sites lie below the Cody Scarp
(Wicomico shoreline, ca 30 m above present sea level) and hence would have been
inundated by Plio-Pleistocene marine transgressions (Gilbert 1987). Throughout
this report, data not otherwise specified were collected at WSSP.
BULLETIN FLORIDA MUSEUM NATURAL HISTORY VOL. 41(2)
Climate
Situated at the southeastern corner of a large temperate land mass, northern
Florida shares characteristics of both temperate and humid subtropical climates.
As part of the continental mainland, "panhandle" Florida's air temperatures are
more markedly seasonal than those of the Florida peninsula. Long-term average
(30 years through 1986) maximum and minimum daily temperatures for
Tallahassee, which lies only 18 km north (inland) of Wakulla Springs, are 26C
and 130C, respectively; however, weekly extremes range from a mean low of 3.3C
(mid-January) to a mean high of 32.70C (mid-June to mid-August). Although
between-year variation in precipitation is substantial and regularly produces both
droughts and floods, as a whole Florida is one of the wettest regions of the United
States. With a mean annual rainfall of 163 cm, Tallahassee is the wettest city in
the state. Spring typically is relatively dry and is followed by a distinct summer
rainy season. The months of May through August account for approximately 50-
55 percent of annual precipitation and 69 percent of the 84 days per year with
thunderstorms (25-year average through 1986). Principally a result of convection
and convergence, most summer rainfall occurs in the afternoon and early evening;
in Tallahassee, 48 percent of the daily summer precipitation falls in the 5-hour
period between 1400 hr and 1900 hr, whereas only 19 percent falls during the 10
hours between 2000 hr and 0600 hr. Rainfall of 2.5 mm or more typically occurs
on 10-12 days per month during this period (July average), with 12.3 nun or more
on 5 to 6 days per month (Winsberg 1990).
TAXONOMIC ALLOCATION OF WAKULLA RIVER Pseudemys
Because the Wakulla River has been excluded from the range of P. c.
suwanniensis in some recent literature (e.g., figure but not text of Ward 1984;
Conant and Collins 1991; Seidel 1994), a comment about our assignment of this
population to that subspecies is in order.
Initially described by Carr (1937) as a subspecies of P. floridana,
suwanniensis was transferred to P. concinna when Crenshaw (1955) separated the
two species; most subsequent experts have followed this arrangement. An analysis
of pertinent literature (largely reviewed by Ward 1984 and Seidel 1994) reveals
very few characters that consistently distinguish P. c. suwanniensis from P. c.
concinna, with the two taxa apparently intergrading in northern Florida (Ward
1984; DRJ pers. obs.). In all wild-caught turtles from the Wakulla River, the
ground color of the skin and carapace is nearly black (Plate 2A), as described by
Carr (1937, 1938) for suwanniensis. However, as noted by Ward (1984), this may
be environmentally induced; several Wakulla River hatchlings raised by us for
three years in tannin-free waters and with little exposure to direct sunlight were
substantially lighter and their carapacial patterns of concentric lines bolder than
comparably aged wild turtles. The seam-following plastral pattern of most adult
females at the Wakulla River remains pronounced throughout life (Plate 2B) and is
JACKSON & WALKER: REPRODUCTION IN SUWANNEE COOTER
rarely reduced to the extent that Carr observed in the Suwannee River population;
patterns of some Wakulla River females are more elaborate than even the most
developed patterns noted by Carr. The outer surface of the hind foot of the
Wakulla River turtles is essentially unstriped, with at most faint traces of one or
two broken stripes, and there are typically only five stripes between the eyes; both
characters are considered diagnostic of suwanniensis. The only remaining
character by which Crenshaw (1955) excluded Wakulla River turtles from
suwanniensis (declaring them instead to be intergradient) was the lower tendency
of this population to show connection between the supratemporal and postocular
stripes. Juvenile turtles from the Wakulla River also exhibit a tendency toward an
increase in the number of stripes (more than three) on the outer surface of the
forelimb, a condition that Carr (1937, 1938) noted as being intergradient with
more western populations. Nonetheless, Carr (1938) included in suwanniensis
turtles from as far west as Apalachicola (the Apalachicola and intervening
Ochlockonee rivers both lie west of the Wakulla River).
Interestingly, many of the Wakulla River females bear a pair of variably
developed cusps on the tomia of the upper jaws, reminiscent of members of the P.
rubriventris species group as well as of P. texana (considered to be closely aligned
or even conspecific with P. concinna; Ward 1984, Etchberger and Iverson 1990).
This, coupled with the orange tint of the plastron, may have been responsible for
Carr and Crenshaw's (1957) presumably errant report of P. alabamensis in the
Wakulla River, as well as that of Means (1977) for P. nelsoni in the Apalachicola
River (data for specimens purporting to document the latter are questionable).
In our opinion, similarities among Wakulla River cooters and turtles from
rivers both to the east (e.g., Wacissa, Suwannee) and the west (e.g., Apalachicola)
overwhelm any subtle differences. Depiction of P. c. suwanniensis as a
geographically isolated population (Conant and Collins 1991) meriting specific
status (Frost and Hillis 1990; Collins 1991; Seidel 1994) is inappropriate. D.
Jackson (1995) provided further support for the conspecificity of suwanniensis-like
turtles with other P. concinna, as well as for their specific distinctness from
populations traditionally referred to P. floridana. Treatment of suwanniensis as a
subspecies of P. concinna thus is entirely concordant with the subspecies concept
as recently reviewed by Smith et al. (1997). Comparisons with the limited
reproductive data available for other P. concinna and related populations are
therefore included throughout this report. Should P. c. suwanniensis eventually be
proven to be specifically distinct, this information remains useful as comparisons
between probable sister taxa.
METHODS
Data Collection
Data collection at WSSP was limited to monitoring of nesting females and
nesting sites. An aquatic trapping program was precluded by park policy against
BULLETIN FLORIDA MUSEUM NATURAL HISTORY VOL. 41(2)
use of private boats on the river, as well as by the very large populations of
alligators, diving birds, and large fishes that might become entangled in nets.
Nesting Turtles.- General surveillance for nesting turtles at WSSP was
performed principally from a slowly moving automobile although some was
conducted on foot. Preliminary reconnaissance showed that most cooters at WSSP
nest along River Road; therefore, we concentrated our effort along River Road and
made random checks of other areas as time and manpower permitted. Sampling
intensity was roughly constant throughout the 5-km length of River Road. During
1988, locations along River Road were determined to the nearest 0.03 km by use of
automobile odometers and a series of fixed reference points. In 1989, we
subdivided the road into 45 100-m segments that we demarcated with a numbered
series of stakes constructed of 2.54-cm-diameter PVC pipe. We subsequently
converted 1988 locations to the latter system for analysis. Locations upstream of
segment I were referred to by public facility landmarks.
We first observed turtles with binoculars to determine whether they were
searching for a nest site, excavating a nest, laying eggs, covering a nest, or
returning to the water. Time of first sighting was recorded as Eastern Standard
Time (EST). We refrained from approaching most females that were already
nesting until they began covering activities. For 33 nesting efforts we timed all or
some phases of the nesting process, including return to the water (but excluding
the unobserved approach to the nesting road). With more tolerant turtles, an
observer was able to position himself closely behind the turtle and observe
oviposition. If our presence disrupted activity, we suspended timing until the
female resumed nesting; however, if the female left the nest or did not resume
activity within 45 minutes, we considered the nesting effort to be abandoned and
processed the female at that time. We determined clutch size and distribution of
eggs among the three holes of the tripartite nest (see Results) by direct observation
during laying or by excavation of the nest. In rare instances, we found eggs that a
given female had not laid during nesting and counted them as part of her clutch.
Most females were processed before or immediately after nesting. We
palpated the inguinal cavities of females not observed nesting to determine whether
they were gravid, although this technique proved less reliable than with smaller
species. Voiding large amounts of "urine" (bladder fluid) upon handling was also
indicative of pre-nesting females. We used a Schultheis quick-reading
thermometer to record body (deep cloacal) temperatures (Tb) of nesting females
collected under different conditions (e.g., pre- and post-nesting, under full sun or
clouds, and during rain).
The size distribution of adult females was determined by direct measurements
for the WSSP study population, and by gonadal examination (Jackson 1988) for
other populations. We used 50-cm tree calipers to measure to the nearest mm the
maximum length of the plastron (PL) and maximum length, width, and height of
the carapace (CL, CH, and CW) of each female. Body mass was determined with a
tubular scale (10-kg capacity, 100-g increments). Although body mass is
potentially a more ecologically appropriate measure of size than is length (Bjorndal
JACKSON & WALKER: REPRODUCTION IN SUWANNEE COOTER
and Bolten 1988; Jackson 1988), short-term variability resulting from substantial
fluctuations in bladder and gut contents, as well as eggs, reduces its practical utility
for comparative analyses with turtles (Cagle 1946; Bjomdal and Bolten 1988;
Moskovits 1988; Galbraith et al. 1989; Germano 1993; present study). As noted
earlier by Cagle (1946), PL proved to be the most consistently repeatable
measurement so we used it as a measure of body size for most of the analyses. For
recaptures, we used only our initial measurements of a turtle when depicting size
distribution of the population.
Annual growth was determined by measuring recaptured individuals. We
estimated minimum age at maturity for a few females from plastral growth rings
(assumed to be annuli; Gibbons 1987), but most individuals retained only one or
two recent rings. We deemed skeletochronological techniques (Zug et al. 1986)
inappropriate because we lacked an ontogenetic series of turtles for calibration.
Nesting season was determined both from gonadal examinations (Santa Fe
and Withlacoochee rivers) and direct observations (Wakulla River). We estimated
maximum number of clutches per season from ovarian examination (Jackson 1988)
of several specimens from the Santa Fe and Withlacoochee. For the Wakulla River
population, we estimated maximum possible clutch frequency during a season by
dividing the duration of the reproductive season by the mean interesting interval
as determined from the recapture evidence (below).
Beginning in May 1988, nesting females at WSSP were marked uniquely by
notching two or three marginals. Most turtles were marked either prior to nesting
or immediately upon completion or abandonment of nesting. Additionally, major
scars and breakage of the shells were sketched onto outline drawings of the
plastron and carapace. These later proved especially valuable in confirming the
identities of several individuals. We subsequently used mark-recapture data from
six nesting seasons (1988-1993) to estimate the size, density, and biomass of the
adult female population at WSSP.
Interesting interval, defined as the number of days separating a successful
nesting attempt (i.e., oviposition) from a subsequent nesting attempt within the
same season, was determined from turtles recaptured on land more than 10 days
after the previous observation or following known oviposition. Re-emergence
within 10 days was considered indicative of a renesting attempt following an
aborted or abandoned effort rather than representing a subsequent clutch; renesting
attempts generally were made during a period of a few hours to four days following
the initial attempt (see Results).
Immediate weather conditions were noted for all turtles nesting at WSSP.
Ambient air temperature (T,) in a large, grassy clearing was recorded at a height of
1 m with a thermometer protected from direct sunlight and rain by a three-sided
wooden shelter. For each turtle whose body temperature was recorded, the
corresponding T. was determined more precisely with a shaded Schultheis
thermometer held 2 cm above the substrate. Rainfall was measured with a rain
gauge positioned >40 m from the nearest tree canopy.
BULLETIN FLORIDA MUSEUM NATURAL HISTORY VOL. 41(2)
Additional observations of cooter activity, principally basking, were made
sporadically from boats and from the shore during most months outside of the
nesting season.
Nests, Eggs, and Hatchlings.- Following nesting completion and marking
and releasing the female, eggs were either (1) left in the ground for subsequent
observations of predation, (2) caged to exclude predation in order to determine
dates of natural emergence by hatchlings, or (3) removed for measurement and
incubation in the laboratory. Eggs in the last group were transported in their natal
soil to the laboratory, where they were removed and cleaned gently with tap water.
A pencilled number inscribed on top of each egg was used to maintain orientation
throughout incubation. Within 24 hours of oviposition, egg length and width were
determined to the nearest 0.1 mm with a dial caliper, and egg mass to the nearest
0.1 g with a portable metric precision balance. Relative clutch mass (RCM) was
calculated by dividing the mass of a clutch by the spent (post-nesting) mass of an
individual.
Most eggs were incubated by placing them in small depressions in 2-3 cm of
sand or vermiculite that lined the bottoms of glass aquaria (36 x 21 x 25 cm and 28
x 16 x 12 cm). Aquaria were kept covered with plastic wrap except for a small
opening at one corner. Because water stress can reduce hatchling size and
potential viability (Gutzke et al. 1987; Packard et al. 1981a, b, 1991; Cagle et al.
1993), distilled or tap water was added to the sand as needed throughout
incubation to maintain saturation and hence egg turgor. Some clutches were
incubated at constant temperatures ( 0.50C) in environmental chambers, but
others were allowed to fluctuate with ambient temperature across a variety of
incubation regimes (e.g., air-conditioned house, sun-warmed garage); a few
clutches were split between both regimes. Incubation period was recorded as the
number of days between oviposition and pipping of the egg shell. Non-viable eggs,
identified by their failure to "chalk" (Ewert 1985), as well as any that died during
incubation, were removed and preserved or opened to examine embryos.
Ewert and Nelson's (1991) analysis of a sample of P. concinna eggs from
Tennessee (P. c. hieroglyphica) previously confirmed the existence of temperature-
dependent sex determination (TSD), Pattern Ia (cool males, warm females), in this
species. During 1989, we shipped a sample of 200 eggs from 12 clutches within 24
hours of oviposition to Dr. Michael Ewert in Bloomington, Indiana, to determine
the pivotal incubation temperature (defined as yielding a 1:1 sex ratio) for the
Wakulla River population. Eggs were maintained in environmental chambers at
constant temperatures ranging from 22.5C to 32.00C. Of these 200 eggs, 138
proved suitable for incubation, and 98 of those yielded term embryos or hatchlings
(sacrificed at approximately one month of age) that could be diagnosed for sex
based on gross appearance of the gonads as in Ewert and Nelson (1991).
Natural hatching success and hatchling emergence dates at WSSP were
determined by enclosing 30 nests-6 in 1988; 12, 1989; 8, 1990; 4, 1991-in
protective cages constructed of 1.27-cm hardware cloth and staked to the ground.
Cages were 23 cm high and either 30.5 or 35.5. cm square (the latter size more
JACKSON & WALKER: REPRODUCTION IN SUWANNEE COOTER
effectively protected the side-holes of the nest), with additional flanges extending
10 or 15 cm outward along the ground to prevent digging by predators. Each cage
lid was held securely in place by rubber tie-downs (38 or 51 cm) or wire tie-wraps.
Prior to caging, eggs were removed for counting, then carefully reburied. Burger
(1977) and Ewert (1985) have shown that such movement of turtle eggs within
several hours of oviposition does not interfere with development or hatching rate.
In one or two caged nests each year, a dial thermometer was positioned during
reburial so the tip of the thermal probe rested in the center of the egg clutch.
Approximately 60-80 days after oviposition, small wooden shelters constructed of 5
x 10-cm lumber were placed inside and along one edge of each cage to provide
sufficient shade so emerging hatchlings would not suffer thermoregulatory stress;
shelters were situated so no shadows were cast over the nest cavities. Dates of
emergence were recorded and hatchlings examined for indications of time since
pipping (i.e., loss of egg tooth, degree of yolk scar closure, carapacial color
change). Nests yielding no or fewer hatchlings than expected were excavated the
following May and their remains examined.
Hatchlings from laboratory-incubated eggs were measured (CL, PL, CW to
nearest 0.1 mm; mass to nearest 0.01 g) 10 days post-pipping to allow the carapace
to harden and reach standard shape (Cagle 1954; Jackson and Jackson 1968; Ewert
1979; Rowe 1995). Similar measurements were recorded for hatchlings from
natural nests shortly after their emergence; based on umbilical closure and caruncle
loss, the time elapsed since pipping greatly exceeded 10 days for all such
hatchlings. Hatchlings were compared to the description given by C. Jackson and
Jackson (1968) for P. c. suwanniensis from peninsular Florida. Within one month
of hatching, all hatchlings (from laboratory incubation and natural nests) not used
for sex determination were double-marked by cohort by notching one rear marginal
and clipping one fore toe and then released at the Wakulla River.
Because several other turtles at WSSP (Florida softshell, Apalone ferox;
common snapping turtle, Chelydra serpentina; yellow-bellied turtle, Trachemys
scripta; and potentially Florida cooter, P. floridana; alligator snapping turtle,
Macroclemys temminckii; and eastern box turtle, Terrapene carolina) nest in small
numbers in the same areas as P. concinna, it was necessary to distinguish these
other nests from those of P. concinna. The elongate pliable-shelled eggs (Ewert
1979; Congdon and Gibbons 1990b) and three-holed nests of P. concinna, or their
remains, served adequately to distinguish the nests of this species from all but P.
floridana. The latter, however, was rare at WSSP and was never observed nesting.
During 1988 and 1989, we counted all fresh, abandoned, and depredated
nests observed in our WSSP surveys. Following observation, we destroyed and
marked nests that had been abandoned or depredated to assure that we would not
recount them during subsequent surveys; frequently such nests were re-excavated
by predators. If we were not sure whether a nest had been counted previously, we
excluded it.
We monitored the fates of 114 clutches laid during the 1989 and 1990
seasons to quantify the incidence of nest predation at WSSP. We took special care
BULLETIN FLORIDA MUSEUM NATURAL HISTORY VOL. 41(2)
not to touch the substrate or any part of these nests, each of which was marked by a
piece of plastic flagging tape hung a considerable distance (5-10 m) away.
Marking turtle nests in this manner does not influence the behavior of potential
predators (Tuberville and Burke 1994).
Home Range/Radiotelemetry.- We employed Sexton's (1959) definition,
the "minimum direct distance over water between the two most distant points of
capture," as an appropriate measure of aquatic home range for stream-dwelling
turtles. This excludes the terrestrial portion of total home range that is utilized
only during nesting excursions. To estimate such "linear home ranges" for WSSP
P. concinna, five adult females were fitted early in the 1989 nesting season with
radio transmitters with distinct signal frequencies. A fully encased transmitter
measuring ca 45 mm x 23 mm and weighing ca 19 g was affixed with stainless
steel machine screws and locknuts to the posterior dorsal surface of the carapace,
above the right hind leg, of each female following its first capture in 1989 (April
27-May 4). The trailing whip antenna (42 cm) was secured to the posterior
carapacial margin with plastic ties threaded through holes drilled in the carapace.
Females were retained in a shaded, 340-1 galvanized steel tank, approximately
half-filled with river water, for a maximum of 48 hours before and 24 hours after
attachment of transmitters. Turtles were released at the river's edge at points
nearest their sites of capture on the nesting road and subsequently tracked from
River Road with a three-element yagi antenna and Telonics TR-2 receiver.
Because we could not determine from our tracking position whether turtles utilized
the full width of the river or only an area along one shoreline, we did not attempt
to estimate home range areas. Data for one of the five females, whose transmitter
failed after two weeks of service, were excluded from analysis.
Necropsies.- We necropsied the reproductive tracts of 25 specimens (5 adult
females, 5 adult males, 12 juveniles) from the Santa Fe and Withlacoochee rivers,
as well as 4 females from Wakulla River, found dead. Data recorded from females
included size and number of ovarian follicles and corpora lutea, as well as masses
of ovaries and oviducts. Data for males included testicular mass, presence or
absence of sperm in the epididymides, and degree of development of secondary sex
characters (enlargement of tail and foreclaws).
Statistical Analysis
Analyses of descriptive statistics, correlations, and regressions were
performed using the SAS program for microcomputers (SAS Institute, Inc. 1988).
All tests were parametric and interpreted at an alpha level of 0.05. Because of the
relative constancy of resources in the chief study site, as well as the low
interannual variation in many reproductive parameters (e.g., clutch size) of
previously studied freshwater and marine turtle populations (Gibbons et al. 1982;
Frazer and Richardson 1985), we combined data from WSSP across years for most
analyses. Throughout this paper, means are presented with standard deviations
unless otherwise specified.
JACKSON & WALKER: REPRODUCTION IN SUWANNEE COOTER
RESULTS
General
We recorded 630 nesting emergences by 247 known adult females at WSSP
(Table 3). Additionally, there were 52 emergences for which individual turtles
were not identified (mostly prior to the advent of marking in 1988). Marked turtles
were captured from one to ten times each (Table 4). Data were recorded for 115
clutches of eggs, of which 103 were laid by known females and 12 by unidentified
females. Average life history parameters calculated from these samples are
summarized in Table 5.
Home Range and Homing
Estimates of minimal linear aquatic home range of the four telemetered
females (mm PL/number of radiolocations in parentheses) during the 1989 nesting
season were 200 m (329 mm/21 locations), 400 m (314/16), 450 m (342/9), and
600 m (343/15). The largest range completely encompassed the smallest and
partially overlapped another (Fig. 3). During the 4-month period in which turtles
were tracked (i.e., throughout and immediately after the nesting season, May-
August), there was no evidence of a shift in home range by any individual. Five of
eight known 1989 nests of these females were constructed on segments of River
Road directly inland from their riverine home ranges; the remaining three occurred
from 100 m to 300 m beyond these boundaries and probably reflected the tendency
of some females to wander terrestrially prior to nesting. Additionally, no eggs in
seven known 1988 and 1990 nests of these turtles were laid more than 100 m
outside of the same segments of road (Fig. 3). Assuming these turtles nest directly
inland from their home stretches of river, this provides strong evidence of the usual
stability of home ranges between years, as has been suggested for other species of
Pseudemys (Kramer 1995).
Non-telemetric data provided evidence of at least occasional long-distance
movements, however. For example, in 1988 and 1989, female PC-68 was observed
nesting only on a small stretch of River Road (segments 26 and 27) ca 3 km below
the main spring. During the following 2 years, we saw her once in segment 5 and
the next four times only on the lodge lawn adjacent to the main spring. If females
nest adjacent to their normal aquatic home ranges, as we believe, then this
individual moved approximately 3 km upstream after the 1989 season where she
established and maintained a new home range for at least the next two seasons.
The presence of barnacles on the shells of at least two other females at WSSP
further suggests occasional long-distance movements. Although spawning
barnacles may be transported upstream via boats or large fishes, it is more
parsimonious to assume that larval settlement occurred while the turtles occupied
Table 3. Summary of capture data for nesting female P. concinna marked and recaptured at Wakulla Springs State Park, 1988-1993. Turtles released unmarked are
not included.
Total Captures No. Individuals No. Newly Taken2 No. Individuals Taken From Year1
Year Alive Dead Alive Dead' Alive Dead (%) 1988 1989 1990 1991 1992
1988 97 8 77 7 77 7 (100) -
1989 197 2 133 1 90 1 (68) 43 -
1990 172 4 123 2 45 0 (36) 38 42 -
1991 100 34 73 1 18 0 (24) 20 23 13
1992 21 1 21 1 5 0 (24) 7 5 4 1 -
1993 27 0 27 0 4 0 (15) 9 9 3 0 2
Total 614 18 239 8
1 includes only turnles not seen alive earlier in same year
* number of unmarked individuals first recorded (and marked if alive) in a given year, percentage cohmmn number of previously unmarked individuals (alive and dead) divided by total number of individuals
encountered in a given year
' number ofmarked individuals encountered in one year (horizontal row) that were first marked in an earlier year (vertical column)
4 includes two marked turtles found dead in river
JACKSON & WALKER: REPRODUCTION IN SUWANNEE COOTER
Table 4. Frequency of capture of 238 marked adult female P. concinna at Wakulla Springs State Park,
1988-1993.
No. Captures' N
1 79
2 59
3 40
4 28
5 11
6 13
7 5
8 1
9 1
10 1
includes finding turtle dead in nesting area
Table 5. Life history parameters, including body sizes at first measurement, of 244 nesting female P.
concinna and associated egg clutches from Wakulla Springs State Park. Linear measurements are in mm,
body masses in kg, and egg and clutch masses in g. Relative clutch mass (RCM) is based on spent body
mass (i.e., mass following oviposition).
Variable N Min Max Mean S.D.
plastron length 244 304 382 341 14.7
carapace length 239 328 427 378 17.3
carapace widm 236 251 312 282 11.0
carapace height 226 120 183 148 12.2
mass (gravid) 121 4.7 10.5 6.9 1.05
mass (post-nesting) 104 4.5 8.1 6.5 0.86
clutch size 93 2-8' 27.0 17.5' 4.5
egg length 397 29.5 46.0 38.9 2.74
egg width 397 22.7 30.6 27.2 1.32
egg mass 368 9.79 21.65 16.28 2.14
clutch mass 22 205.3 424.4 307.0 61.8
RCM 22 0.034 0.061 0.048 0.008
SSmallest clutches may have resulted from partial depredation of nests; mean based on 88 clutches of 8 eggs (see text).
more brackish or estuarine waters near the river mouth, perhaps 10-15 km below
the park.
Two females mistakenly displaced following processing during the 1990
season suggest the possibility of homing. Following nesting on 22 April, female
PC-134 was maintained in a holding tank for an educational demonstration; she
was released the next day 1.9 km upstream. On 7 June she nested only 350 m from
the original capture site (presumably within the original home range), then one
year later was observed nesting within 200 m of the site. A second female (PC-37)
held similarly for 6 days following nesting on 23 April was mistakenly released on
River Road 650 m downstream of her latest (and usual) nesting site; she was next
observed nesting on 16 June within 15 m of the previous capture site.
BULLETIN FLORIDA MUSEUM NATURAL HISTORY VOL. 41(2)
Figure 3. Extent of linear home ranges as determined by radiotelemetry for four adult female P. concinna
(A-D) in the upper Wakulla River during the 1989 nesting season. Width of river transect utilized by turtles
was not determined. Known nesting sites from 1988, 1989, and 1990 are indicated by lower case letters; all
other symbols as in Figure 2.
Nesting Season
Table 6 presents dates of first and last observed nesting attempts (fresh nest or
emergent female) for the four principal years of study at WSSP. The average
nesting season encompassed at least 117 (range 102-134) days extending from late
March/early April through late July/early August. Although earlier nesting
attempts may have been overlooked (especially in 1988), we consider the dates
listed to be generally representative of nesting onset as no prior signs of nesting
(e.g., depredated nests) were detected. Onset of seasonal nesting activity varied as
much as 16 days, from 30 March to 15 April. Subsequent observation of
inset
STurtle A 'A. J
* Turtle B
M Turtle C
B Turtle D
1 .5 0 1 KM
JACKSON & WALKER: REPRODUCTION IN SUWANNEE COOTER
Table 6. Observed length of nesting season for P. concinna during the six-year period 1988-1993 at
Wakulla Springs State Park. For each year, table gives date of (a) first confirmed nesting attempt, (b) last
observed nesting attempt by an undisturbed turtle (i.e., one whose nesting season could be confirmed not to
have been extended by observer disturbance), (c) last observed nesting attempt by any turtle (last abs.), and
(d,e) the first and last observed nesting attempts after excluding the three earliest and three latest undisturbed
(as well as all potentially disturbed) turtles; field effort was inadequate to obtain sufficient late season data
for 1992 and 1993. Conservative estimates are given for (f) minimum length of nesting season (b-a), (g)
maximum length of nesting season (c-a), and (h) period of most nesting, excluding outliers (e-d).
1988 1989 1990 1991 1992 1993
(a) First 14 Apr 15 Apr 30 Mar 5 Apr 23 Mar 15 Apr
(b) Last und. 24 Jul 1 Aug 10 Aug 4 Aug > 15 Jul
(c) Last abs. 12 Aug 20 Aug 10 Aug 10 Aug
(d) First 3 18Apr 20 Apr 30Mar 16 Apr 19 Apr 15Apr
(e) Last und. 3 18 Jul 30 Jul 21 Jul 26 Jul -
(f) Min days 102 109 134 122 > 115
(f) Max days 121 128 134 128
(h) Most 92 102 114 102
a single female attempting to nest on 23 March 1992 (A. Whitehouse pers. obs.)
suggests that isolated individuals nest even earlier in some seasons. Dates of
termination of the nesting season in late July or early August are more tentative
because the emergence of late-nesting females could easily have been missed
during this span of several weeks, and freshly depredated nests might actually have
been several weeks old. Therefore, the calculated minimum length of each nesting
season given in Table 6 is a conservative estimate. Late-nesting females whose
seasons may have been extended as a result of observer disturbance and subsequent
nest abandonment were used only to estimate the potential maximum length of
each nesting season. Estimates of nesting season based on the smaller samples
from peninsular Florida fell within the range of nesting dates at WSSP.
We assessed duration of the nesting season for most of the Wakulla River
population each year by recalculating season length after excluding the three
earliest and three latest observed nesting attempts by undisturbed females. The
resulting periods encompassed 92-114 days falling between the dates of 30 March
and 30 July each year. Although each nesting season was characterized by weeks
with abundant nesting and others with sparse nesting, the overall annual nesting
profile at WSSP approaches a normal distribution (Fig. 4).
Enlargement of follicles to be ovulated in the subsequent year begins almost
immediately after the current nesting season ends. Our 1970s samples from the
Santa Fe and Withlacoochee rivers included five mature females taken in mid-
August and early September. All had produced multiple clutches, based on counts
of residual ovarian corpora lutea, yet each bore enlarging yolked follicles 10-19
mm in diameter; these were not, however, in clearly defined size groupings
indicative of distinct incipient clutches.
BULLETIN FLORIDA MUSEUM NATURAL HISTORY VOL. 41(2)
El 1991
* 1990
H 1989
0 1988
Hil
1
15 17 19
1-Week Sampling Periods
Figure 4. Nesting frequency of Wakulla River P. concinna by week for the four years 1988-1991, based
solely on observations of females that emerged to nest. One-week sampling periods are depicted beginning
on 30 March (earliest recorded nesting) and extending through 23 August (latest recorded nesting = 20
August).
Nesting Behavior
Diel Nesting Cycle and Proximate Nesting Cues.- At WSSP, Suwannee
cooters nest almost exclusively during daylight hours, with nesting migrations
being initiated and completed on the same day. Only following infrequent late
evening rains (beginning later than 1800 hr) did we note any turtles completing
nesting after dark, and never did we observe a female that had emerged during
darkness. Figure 5 depicts times of day we encountered female cooters at nesting
sites during the study; of these, 28.3 percent occurred before 1200 hr EST, 46.7
percent between 1200 hr and 1600 hr, and 25.0 percent after 1600 hr. The 7-hour
period from 1100 hr to 1800 hr accounted for 76.5 percent of all observations.
100
N = 601
3 5 7 9 11 13
JACKSON & WALKER: REPRODUCTION IN SUWANNEE COOTER
N = 647 U Rain
I No Rain
- 80-
T,
S60-
C
M-
0
-o
40-
Z 20-
0600 0800 1000 1200 1400 1600 1800 2000
Time of Day
Figure 5. Time of day of encounters with adult female P. concinna on land during the 1988-1993 nesting
seasons at Wakulla Springs State Park. Data include females encountered immediately before and after
nesting as well as those observed in the nesting process. All times are standardized to Eastern Standard
Time. The two columns shown for each hour represent all observations made in the hour beginning at that
time; dark columns represent turtles that emerged in response to rainfall.
Individual females showed no proclivity for emerging at preferred times (e.g.,
morning or evening) but rather seemed to track precipitation patterns (below). As
an example, female PC-17 was recorded attempting to nest at the following seven
times: 0850 hr, 0850 hr, 1325 hr, 1419 hr, 1430 hr, 1750 hr, and 1909 hr.
Several events early in the study led us to suspect that rainfall served as an
immediate stimulus for nesting. On the afternoon of 20 May 1985, during a
fortuitous field survey of WSSP, one of us (DRJ) observed six P. concinna nesting
toward the end of a heavy rain that marked the first substantial precipitation in the
region in many weeks. In 1986, a severe drought was broken by heavy rains again
on 20 May. Upon arrival at the park at 1200 hr, DRJ immediately observed six
cooters that had emerged to nest within 100 m of each other. A subsequent survey
of depredated nests along River Road revealed that more than a dozen other cooters
had been nesting simultaneously. This and the absence of nesting cooters or
freshly depredated nests during random visits on dry days in these years and 1987
BULLETIN FLORIDA MUSEUM NATURAL HISTORY VOL 41(2)
make it clear that precipitation is a major cue for nesting in this population. This
subsequently caused us to direct our field efforts disproportionately toward days on
which rain appeared probable. From May through August, most precipitation in
this region occurs as afternoon thunderstorms. Hence, most turtles nesting in
association with rainfall were encountered from 1100 hr to 1700 hr, while
observations of turtles nesting on rainless days were distributed more evenly from
0800 hr to 1800 hr (Fig. 5).
Of the turtles we observed emerging to nest, we judged 83 percent did so in
response to precipitation (Fig. 5). The actual response rate for the population
surely is somewhat less, as our field work after 1988 was biased toward days with
greater potential for rain. Turtles generally appeared on the River Road from 45
minutes to 2 hours after initiation of rainfall, except for females that did not
emerge until daylight following infrequent nocturnal rains. Precipitation
associated with nesting varied from a trace to heavy downpours of >5 cm. Our
subjective impression is that the amount of rainfall necessary to stimulate
emergence declines with time passed since the last precipitation event. Rainfall of
1 cm appeared to stimulate most females holding adequately shelled eggs to nest; if
a week or more had passed since the last rainfall, large numbers of turtles emerged
in approximate synchrony. The most outstanding example of this pattern occurred
in May 1989, when we processed 14, 31, and 27 individuals (not all that emerged)
on May 1, 10, and 21, respectively. Each of these dates was characterized by 1.4-
2.1 cm of rain preceded by 8-11 dry days. In contrast, rains that followed prior
adequate precipitation by <48 hours rarely produced more than a few nesting
turtles, as most gravid females tended to lay at the first opportunity. The peaks and
valleys conspicuous in the seasonal nesting profile (Fig. 4) are directly attributable
to this phenomenon.
Ambient air temperature during nesting typically varied from 220C to 300C,
although it ranged as high as 360C (>390C in the sun) and as low as 170C. Low
temperatures usually corresponded to rapid temperature drops associated with
thunderstorms, whereas high temperatures occurred in open areas on clear days in
mid-afternoon. Except following rain, few females emerged under a hot, midday
sun (ca 1100-1500 hr).
Nesting Distribution and Nest Site Selection.- Of 646 nesting emergences
tabulated within WSSP, 87.2 percent (563) occurred on or adjacent to River Road;
10.4 percent (67) was split almost equally between public areas A and B; the
remaining 2.4 percent (16) occurred at Sally Ward Slough. Four old depredated
shells were discovered at MacBride Slough north of the river, but this little-used
site (based on signs of nesting) was too distant to monitor regularly. Three turtles
were encountered outside of the park where an upstream arm of Sally Ward Slough
flows beneath highways bordering the park. Figure 6 depicts the frequency of
observations within the regularly monitored areas. Nesting by segment along
River Road was highly nonrandom (chi-square=274.8, 43 df, p<0.001).
Upon emergence, females generally walked directly away from the river to the
nearest non-canopied, adequately drained site suitable for potential nesting. On
JACKSON & WALKER: REPRODUCTION IN SUWANNEE COOTER
the south side of the river, where most nesting occurred, these conditions usually
were met by River Road (Plate 1) and necessitated overland movements of 20-300
m (Fig. 6). Rocks, fallen trees, and other physiographic irregularities caused
deviations from straight-line walking, roughly in proportion to the distance
between river and road. At two sites, the public grounds near the lodge and in the
main feeding area, turtles encountered short, exposed grass almost immediately
upon emergence from the river; nonetheless, they usually crossed >80 m of such
habitat before nesting, which strongly suggests that adequate soil drainage,
determined chiefly by elevation above the river or floodplain, is a key component of
nest site selection. Turtles encountering River Road either nested in the road or on
the road shoulders; areas of very soft sand, which desiccated readily, were avoided.
We observed no turtles nesting in the forest on either side of River Road, although
50 r
40
E
LU
0
S20
0
0
N = 646
1 5 10 15 20 25 30 35
Location
I I
I I
0 '
SI
I
I4
l I
i' L
40 45
250
200
150 -
100
50
0
Figure 6. Frequency of nesting emergences by site for P. concinna at Wakulla Springs State Park, 1988-
1993. Horizontal axis: 100-m segments #1-45 along River Road (#46=last 20 m only); the three unlabeled
columns to the left of segment 1 represent Sally Ward Slough and public areas A and B, respectively.
Dashed line depicts approximate mean distance (right vertical axis) of River Road from river within each
numbered segment
BULLETIN FLORIDA MUSEUM NATURAL HISTORY VOL. 41(2)
a very small number continued walking along side roads that led farther from the
river. Turtles seemed to select sites without regard to whether they were in the sun
or shade at the time of nesting. During 1993, park officials conducted a series of
thinnings and prescribed burns to restore some of the former open pineland habitat
along River Road. Shortly thereafter, several cooters were observed to have
crossed River Road to nest in the recently charred and sparsely canopied habitat
(A. Whitehouse and S. Cook pers. comm.). A layer of ash temporarily replaced
vegetative groundcover, and most of the shrubby understory of saw palmetto
(Serenoa repens) and wax myrtle had been scorched.
Groundcover at nest sites varied from dense grass to bare sand, but with a
noticeable preference for the latter. In particular, turtles nested frequently in the
two bare sandy tire tracks as opposed to the grassy median and shoulders of River
Road. In fact, one of the commonest nesting postures consisted of a female
positioning herself so that the anterior portion of her body overlay (and was
slightly elevated by) the grassy- median or shoulder, while she dug her nest
posteriorly in the bare sand of a tire track. We detected no strong preference in
whether females faced toward or away from the river when in this position on
River Road. The steepest gradient on which a nest was constructed measured 160
and was essentially the steepest slope available; most nests were dug on slopes of
less than 5.
Nesting substrates consisted of medium to coarse sands, sometimes with grass
roots, from which water drained rapidly following rains. Soils with appreciable
organic content (ranging from humus to muck) were associated with leaf litter in
the hardwood and floodplain forests adjacent to River Road. Their potential
availability for nesting seems to have been precluded by their restriction to deep
shade.
Nest Site Fidelity.- Figure 7 depicts the mean distance per nest site (or
encounter) across all years, from the "average nest site/encounter location" (metric
distances of all sites from base point zero summed, then divided by number of
sites) for each turtle observed on River Road more than once during the study
(rarely were these successive nestings). Females typically nested within a road
segment of 400 m (i.e., <10 percent of the road's ca 4.5-km length). Some turtles
that apparently used landmarks (e.g., the railroad by the warehouse) for orientation
constructed subsequent nests within 1 m of earlier ones. Variance for some
females increased as a result of an occasional "outlying" nesting attempt. For
example, turtles seen nesting several times within a segment of 200 m occasionally
(usually only one of five or more observations) nested as much as 1.7 km away.
This phenomenon occurred both within and between years.
In general, turtles not nesting along River Road were loyal to the other sites.
Some turtles nesting near the Sally Ward Slough bridge also used the lodge lawn
above the main spring (both accessible from the main spring), but only twice did
we document any of these individuals nesting downstream from the spring.
JACKSON & WALKER: REPRODUCTION IN SUWANNEE COOTER
N = 130
- 40
-30
0
-20
E
Z 10
liflhli.
>0 100 200 300 400 500 600 700 800 900 1000
Mean Distance (m)
Figure 7. Nest site fidelity, as measured by mean distance per nest or encounter, across all years, from
average nest/encounter site for each of 130 P. concinna observed more than once on River Road (total road
length ca 4.5 kin). For analysis, each observation was considered to occur at the middle of a 33.3-m segment
of River Road. Larger mean distances often reflect a single outlying point or shift of home range (see text).
Nesting Process and Nest Structure.- We present here a synopsis of
features gleaned from observing parts of hundreds of nesting efforts. The process
may be subdivided into four phases, as per Linck et al. (1989):
1) Pre-nesting behavior: We seldom observed emergent females on their
way to a nesting area. Upon reaching River Road, turtles move slowly about the
road and its shoulders in search of a suitable nest site. During this exploratory
process, females frequently press their noses to the ground as if "sniffing"
(Stoneburner and Richardson 1981; Linck et al. 1989). Selection of the eventual
nest site typically requires 1-10 minutes. The entire pre-nesting period appears to
BULLETIN FLORIDA MUSEUM NATURAL HISTORY VOL. 41(2)
be marked by heightened vigilance and proclivity for retreat to the river (see Nest
Abandonment and Effects of Human Disturbance, p. 36).
2) Nest excavation and structure: Turtles usually begin digging almost
immediately upon selecting a site. At intervals throughout excavation, females
sometimes release small quantities of fluid from the cloaca. As we observed this
behavior even on rain-soaked soil, we suspect that turtles transport water to the
nest regardless of weather. When handled, pre-nesting females frequently voided
this fluid. Three females measured before and after nesting lost >500 g each,
attributable to fluid as opposed to eggs.
Typical of most emydids (Cagle 1937; Carr 1952; Ernst et al. 1994), there is
no body pit, and all digging is done by the hind legs. The front legs remain
stationary throughout the entire nesting process and seemingly serve to maintain
orientation to the nest. The hind legs are used alternately, but each may perform
from one to three strokes before work is shifted to the other. Nesting is solitary
although individuals were observed nesting within 5 m of each other during some
peak emergences.
Exceptional to the usual emydid nesting pattern, females at WSSP excavate a
shallow accessory hole (i.e., side-hole or satellite nest) several cm on either side of
the central nest hole (Fig. 8). The central (main) chamber of the resulting nest is
the typically flask-shaped structure dug by most turtles, usually with an asymmetric
bulge toward the front (beneath the turtle). Approximately 100 nummn across at the
surface, the roughly circular opening tapers to a narrowed neck region 60-70 mm
across, then re-expands into a central egg chamber approximately 100-120 mm in
diameter. Both the neck and egg chamber are slightly wider than long; the
relatively flat floor of the latter lies at a mean depth of 157 mm ( 10.2, range 131-
167 mm, n=10). The trench-shaped accessory holes, each measuring roughly 80-
90 mm in width by 100-120 mm in length at the surface, lie laterally to the central
hole and reach an average depth of only 76 mm ( 13.3, range 60-104 mm, n=19).
The entire nest structure, measured between the outer edges of the two accessory
holes, spans a width of 290-370 mm (mean=346 23.0 mm, n=l 1).
All females within Florida populations of P. concinna for which we have
observed nests (i.e., Wakulla, Santa Fe, and Withlacoochee river drainages)
invariably constructed the accessory holes. Except for two individuals that nearly
completed the side-holes prior to initiating the center hole, WSSP females dig the
side-holes in conjunction with digging the center hole. Periodically, females shift
from digging the center hole to digging a few strokes on either side-hole before
returning to the center hole. We observed that some females lacking one hind leg
completed successful nests that had only one side-hole (as also noted by Franz
[1986] for P. floridana, which constructs similar nests).
Because most females either were already nesting when discovered or were
processed immediately if not yet nesting, we have few data to document the
duration of nest excavation itself; one of the most rapidly dug nests required only
17 minutes, but digging more characteristically continued for ca 30 minutes.
JACKSON & WALKER: REPRODUCTION IN SUWANNEE COOTER
Figure 8. Three-holed nest of P. concinna at Wakulla Springs State Park. Above, fresh nest before
covering, with eggs in each accessory hole but majority of clutch in central chamber. Below, diagrammatic
cross-section of completed nest based on measurements, viewed from behind turtle; hatching indicates back-
filled soil.
BULLETIN FLORIDA MUSEUM NATURAL HISTORY VOL 41(2)
Table 7. Distribution of eggs in accessory holes (EAH) in 94 P. concinna nests at Wakulla Springs State
Park, 1987-1993. Colons separate number of eggs in two accessory holes without regard to left or right
EAH 0:0 1:0 1:1 2:0 2:1 3:0
Frequency 60 22 5 4 1 2
(Percent) (64) (23) (5) (4) (1) (2)
3) Oviposition: Oviposition typically begins almost immediately after nest
excavation is completed. Eggs are laid singly at semi-regular intervals typically of
30-40 seconds but occasionally as long as 3 minutes. Oviposition of 10 clutches of
13-23 eggs each required 7-16 minutes (mean=11.0 2.9). Partial retraction of
the head and slight elevation of the hindquarters accompanied the passage of each
egg. This behavior allowed us to monitor oviposition from a distance. Each egg is
arranged by one or both rear feet (alternately) in the nest- before the next egg is
laid. Thus, eggs become snugly packed and can not be removed intact without
difficulty. Packing time often increases toward the end of laying. Even so, the
uppermost eggs sometimes extend into the neck of the nest cavity.
Although most or all eggs in a clutch are deposited into the large, central egg
chamber, one or more eggs may be laid in the accessory holes (Table 7; Fig. 8).
There is some uncertainty, based on observations of P. floridana (Allen 1938; Carr
1952), whether egg placement in the side-holes is purposeful or accidental. Our
observations of nesting P. concinna at WSSP suggest both. We have watched
laying cooters deliberately shift posterolaterally prior to laying an egg in one of the
accessory holes, and we also have observed females using a hind leg to sweep a
freshly laid egg laterally into such a hole. The posterolateral shifts are not typical
of ovipositing turtles that construct single-holed nests. On the other hand, we
noted several instances in which eggs were laid on the ridges between center and
side-holes, with the direction that such eggs subsequently rolled seeming little
more than random. The placement of eggs in the side-holes seems unrelated to
clutch size or female individuality. Thirty-two nests with one or more eggs in such
holes averaged 18.7 eggs, about one egg larger than the average clutch (Table 5).
Three nests with three such eggs were from relatively small clutches (13, 12, and 6
eggs), whereas nine nests with two such eggs represented a typical range of clutch
sizes (14-25). Only three of eight females that laid a side-hole egg in one nest also
did so in a second known nest. Interestingly, if any eggs from a clutch did reach
the side-holes, they were generally among the first eggs laid from that clutch.
In 94 non-depredated nests examined by us at WSSP, only 36 percent of nests
(34) and 21 percent of accessory holes (40 of 188) contained satellite eggs. Such
eggs were buried less deeply than those in the central chamber, even when the
latter was so full that eggs were forced into the neck region. Depth to the top of
the uppermost egg in the main chambers averaged 81 mm ( 12.1 mm, range 55-
JACKSON & WALKER: REPRODUCTION IN SUWANNEE COOTER
110, n=16), three times as deep as the uppermost egg in accessory holes (mean=27
10.0 mm, range 8-44 mm, n=17). In fact, the most deeply buried egg in an
accessory hole (at 44 mm) was 1 cm shallower than any egg in a main chamber.
The uppermost egg in accessory holes containing two or three eggs was shallower
than eggs buried singly in such holes (means of 17.8 mm vs. 30.3 mm, t=-2.57,
df=-15, p<0.01). Depth of the uppermost egg in the main chamber appeared to
reflect a combination of egg size, egg number, and chamber size (the latter
principally determined by leg length).
4) Nest covering and departure: As with most species of turtles (Ehrenfeld
1979), nest covering begins immediately following laying and positioning of the
last egg. Filling of the nest is accomplished by alternately extending each hind leg
posterolaterally to collect and drag forward into the nest the soil that was removed
during excavation; the accessory holes are filled in conjunction with the central
chamber. Throughout filling, soil is deliberately kneaded and packed into the
cavities with the hind feet and legs. The final packing is assisted by vigorous
vertical thumping and lateral rocking movements of the plastron. Nest-covering
efforts are punctuated by short pauses for rest. Termination of covering is often
marked by a few alternating, forward reaches of the hind legs, in which sand and
debris are flicked backward over the disturbed area; this presumably aids in
visually camouflaging the nest. The stationary front legs continue to provide an
orientational fix that is especially important in light of the substantial posterior
movements that accompany the digging of the accessory holes. In contrast to
behavior noted for some emydids elsewhere (e.g., Cagle 1937; Legler 1954), we
did not observe voiding of fluid during nest covering to create a true nest plug.
This might be ineffective in the sandy soils of northern Florida.
Nest-covering activities consumed 9-28 minutes (14.8 5.3, n=15). Motor
actions involved in nest covering tend not to dissipate easily. That is, some
females removed while still filling nests continued their filling motions, as in
Emydoidea (Linck et al. 1989) and Terrapene (DRJ pers. obs.) though not reported
in other Pseudemys (Linck et al. 1989). However, a few disturbed females
abandoned nests with eggs without covering them, and possibly without having
completed their oviposition.
Most females left the nest immediately upon completion of covering and
headed toward the river. A few walked along River Road as far as 155 m before
turning toward the river. Undisturbed turtles typically alternated bouts of "high
walking" with brief pauses, usually walking for 10-30 seconds, then pausing to rest
and scan their surroundings for 5-10 seconds. Time to return to the water ranged
from 7 to 60 minutes (28.5 20.9, n=10); distance from the river and habitat
complexity appeared to account for much of the variation.
The time on land for a nesting event averaged about 2 hours-ca 1 hour at
the nesting site and 1 hour in transit to and from the river. Individuals abandoning
preliminary nests but completing new ones extended this period by 1-2 hours.
We documented one instance of "false nesting" in which a female completed
the entire nesting process, including nest covering, but laid no eggs. This
BULLETIN FLORIDA MUSEUM NATURAL HISTORY VOL. 41(2)
phenomenon has been reported once previously for an unmolested freshwater turtle
(Chrysemys picta: Legler 1954). It is common in turtles injected with oxytocin and
forced to oviposit without nesting (e.g., Emydoidea: Petokas 1986; Trachemys:
Tucker et al. 1995; Chrysemys: M. A. Ewert, pers. comm.).
Nest Abandonment and Effects of Human Disturbance.- Several factors
caused turtles to abandon nesting efforts; including, disturbance by walking
humans (n>30) and passing motor vehicles (n=8); harassment by fish crows
(Corvus ossifragus; n=3) and raccoons (Procyon lotor, n=6); swarming by fire ants
(Solenopsis invicta, n= 1); and impenetrable substrate, principally obstruction by
roots and rocks (n>24).
Female responses to human disturbance during nesting excursions varied
appreciably. Except during oviposition, most females remained wary and paused
repeatedly to scan their surroundings. However, individuals that habitually nested
in human activity areas (e.g., at the boat maintenance warehouse and on the lawn
of the guest lodge) continued nesting with people present. Such turtles nested with
minimal interruption despite being closely surrounded by a dozen or more
spectators. Prior to egg-laying, most others abandoned the process and retreated to
the river if approached closely, even when observed at a distance of 50 m and
sometimes after remaining virtually motionless for up to 45 minutes. This
generally agrees with observations of Alabama river cooters reported by Fahey
(1987), who noted a cessation of wariness only during actual oviposition. Despite
capture and handling, nearly all WSSP females remained docile regardless of the
stage at which we interrupted them.
We observed 54 renesting attempts; 14 occurred later on the day of the initial
attempt, with the remainder relatively evenly spread throughout the next 10 days.
Turtles that were handled prior to oviposition usually returned to the water without
nesting but re-emerged within three days. A few, however, resumed nesting within
30 minutes, without having returned to the river. Of 30 handled turtles for which
renesting attempts were noted, 12 resumed nesting the same day (with only one of
these possibly having first retreated to the river), whereas 5, 8, and 5 re-emerged 1,
2, and 3 days later, respectively. Abandonment due to physical obstruction of
digging usually led to an immediate renesting attempt within 3 m, whereas same-
day renesting attempts following human disturbance came 0-6 hours later at sites
1-300 m (and in one case, 1 km) away.
Body Temperatures During Nesting.- Body temperatures of emergent
females along River Road (Fig. 9) ranged from 20.60C to 35.60C, depending on
ambient temperature, exposure to sun and rain, and time since leaving the river.
Six categories of turtles differing in nesting stage and environmental conditions
(Fig. 9) had the following mean Tb and mean difference from Ta (in parentheses; in
C): pre-nesting sun, 30.7 (1.5); pre-nesting shade/rain, 22.3 (-0.6); nesting sun,
32.0 (0.6); nesting shade/rain, 25.8 (0.6); post-nesting sun, 32.8 (2.7); post-nesting
shade/rain, 26.7 (-0.2). Statistical comparison of these means is not practical
because of differences in their sample distributions relative to body size, time of
JACKSON & WALKER: REPRODUCTION IN SUWANNEE COOTER
day, and Ta; nonetheless, increases in mean body temperature with nesting stage
characterized turtles under both sunny and non-sunny conditions.
Under cloudy and/or rainy conditions, body temperatures of pre-nesting
turtles typically approximated the temperature of the river, ca 21-220C, and this
proved useful in determining whether some turtles were pre- or post-nesting. The
highest body temperatures (>320C) characterized turtles that had just completed
nesting under the afternoon sun, to which they presumably had been exposed for
>1 hour. Bare surface sand under similar conditions reached temperatures >50C.
It appeared that rapid drops in T. associated with rainfall or nearby thunderstorms
were resisted by the thermal inertias of these relatively large turtles and created
situations in which Tb exceeded T. by several degrees (Fig. 9).
36.0
35.0 /
E
34.0 E E /
33.0 A
32.0 E /
EF E
o31.0 E D
230.0 A /
29.0 F F /
(.28.0 / F F
)27.0 FDB D /F
>,26.0 F F F
-o
025.0 /
m D
24.0 F /F/F F
23.0 F8 /
B "FB/B B
22.0 B / B B
21.0 B / B
20.0 /
20.0 22.0 24.0 26.0 28.0 30.0 32.0 34.0 36.0 38.0
Ambient Temperature (C)
Figure 9. Body temperature (Tb) plotted against ambient temperature (T.) for 67 nesting P. concinna at
Wakulla Springs State Park, 1988-1991. (A) pre-nesting turtles in sun; (B) pre-nesting in shade or under
clouds/rain; (C) nesting in sun; (D) nesting in shade or under clouds/rain; (E) post-nesting, sun; (F) post-
nesting, shade or clouds/rain. Temperatures of nesting turtles were recorded upon nest abandonment Parity
line (Tb = T.) drawn solely for visual purposes. In some cases, rapid drops in T. as a result of rain, as well as
shifting shade patterns, may have reduced the apparent relationship with Tb (see text).
BULLETIN FLORIDA MUSEUM NATURAL HISTORY VOL. 41(2)
Reproductive Parameters
Female Size, Growth, and Age at Maturity
Size.- Standard body measurements of 244 adult female cooters at WSSP
are summarized in Table 5. PL of nesting females varied from 304 to 383 mm
(Fig. 10), a range of 79 mm. The three largest females (PLs 377-383 mm, CLs
424-427 mm, including one whose skeleton was found in 1991, two years after
being marked, at the bottom of the River Road solution hole) all exceeded the
largest female (416 mm CL) measured by Carr (1937) from the Suwannee River
system and were within ca 1 cm CL of the largest recorded individual of this
species (437 mm, locality unknown: Pritchard 1980).
80 r
60 -
40 -
20 h
300 310
320 330 340 350
N = 243
360 370 380
Plastron Length (mm)
Figure 10. Plastron length at first measurement for 243 nestingP. concinna at Wakulla Springs State Park.
JACKSON & WALKER: REPRODUCTION IN SUWANNEE COOTER
N = 225
c 25- Gravid
V : Post-Nesting
S 20 -
15 -
0
E
z 5
Body Mass (g)
Figure 11. Body mass at first measurement for 225 adult female P. concinna at Wakulla Springs State
Park. Bars represent 500-g classes beginning at the mass shown. Dark columns indicate gravid mass of
nesting and pre-nesting females, including eggs and variable amounts of bladder fluid (n=121); light
columns indicate spent mass of females after deposition of eggs and presumed voiding of most or all bladder
fluid (n= 104).
Even though some Wakulla River females may mature at a PL of ca 300 mm
(equivalent CL ca 325 mm), we suspect the majority of females of this size grow
another 20-50 mm (Fig. 10) before reproducing. Our minimum estimate, derived
from direct gonadal examinations, of ca 290 mm PL at maturity in the Santa Fe
River population is similar. Following the logic of Frazer and Ehrhart (1985),
coupled with the minimal growth rates experienced by adults (below), our best
estimate of the mean size at maturity for Wakulla River females is ca 330 mm PL
(ca 360 mm CL).
Body mass at first measure varied from 4.7 to 10.5 kg (range 5.8 kg) for pre-
nesting (gravid) turtles, and 4.5-8.1 kg (range 3.6 kg) for post-nesting turtles (Fig.
BULLETIN FLORIDA MUSEUM NATURAL HISTORY VOL. 41(2)
10). Three females weighed both before and after nesting (initially 6.35, 6.4, and
7.9 kg; subsequently 6.0, 5.4, and 7.0 kg, respectively) lost, chiefly via eggs and
bladder fluid, an average of 10.8 percent of their pre-nesting masses. Because the
losses ranged from 5.5% to 15.6% and bore no obvious relationship to body mass,
development of an index to relate pre- and post-nesting masses was deemed
inappropriate.
Growth.- Although annual variations in growth rates are ecologically
important in turtles, especially in more rapidly growing immature individuals
(Tucker et al. 1995), the slow growth by adult female cooters during all years at
WSSP prompted us to combine all of our single-year and multi-year data into one
sample. Figure 12 depicts mean annual linear growth (APL) of 91 females for
intervals of 1-5 years; each individual was tabulated only once based on the longest
interval available. Mean annual growth rate unweightedd) was 0.84 mm per year.
The maximum absolute growth observed was 11 mm in a 3-year period (initial PL
324 mm), while the maximum growth recorded in 1 year was 5 mm (initial PL 325
nun). Of the 91 turtles, only 5 (5.5 percent) showed a mean annual growth rate of
>3 mm, and all those individuals were small (initial PLs 323-325 mm); 50 (55
percent) showed a total absolute growth of 1 mm, within our limits of measuring
error. Hence, except for a few of the smallest females, annual growth after
maturation is almost negligible. For some females, however, we did note
measurable increases in width, height, and mass despite stasis in PL.
Age at Maturity.- Six females (PLs 318-355 mm) from our 1990-1992
samples each retained seven to nine annuli representing their last 6-8 years of
growth. We believe that at least 4-6 years of earlier growth were unrepresented by
annuli on all individuals, based on estimated body sizes prior to the oldest annuli
and on juvenile growth rates at both WSSP (DRJ pers. obs.) and in the Suwannee
River population studied by C. Jackson (1964). Minimum age estimates for the six
females thus ranged from 10 to 13 years, so we estimate female maturation at no
less than 9 years of age but more likely in the range of 11-13 years.
We believe that smooth, well-worn shells and increasing melanism of the skin
characterized old age in WSSP females. Such turtles typically showed no growth
when recaptured after intervals of 1-3 years.
Clutch Frequency, Clutch Size, and Reproductive Potential
Internesting Interval and Number of Clutches Per Year.- Figure 13
depicts the number of days between observed nesting attempts by the same turtle
within a season for the combined years 1988-1991 at WSSP. Of these records 61
(representing 52 females) followed confirmed nesting with oviposition and allowed
determination of interesting interval. No post-nesting female returned to nest in
less than 16 days. The post-nesting records fall principally into two groups. The
first (n=32), extending from 16 to 30 days with a peak at 20-21 days, almost
certainly represents the interval between consecutive nesting efforts. A second
group (n=24) extends from 35 to 50 days and logically represents two consecutive
JACKSON & WALKER: REPRODUCTION IN SUWANNEE COOTER
5.0 0
4.0 a zo
0
E
3.0-
" 2.0-
0 0
Om 0 0
A A AA
0 0 0
0 0 0 0
0 o0 D [iInnnmm >o0
320 330 340 350
0 A A 0
360 370 380 390
Initial Plastron Length (mm)
Figure 12. Mean annual increase in plastron length of 91 adult female P. concinna measured during two or
more nesting seasons at Wakulla Springs State Park. Data were estimated to represent yearly intervals
regardless of capture dates, with only the longest available interval plotted for each individual. Interval
lengths depicted as follows: squares, 1 yr, triangles, 2 yrs; diamonds, 3 yrs; asterisks, 4 yrs; dots, 5 yrs; some
symbols overlap exactly.
nesting intervals across three nestings. The remaining five intervals that followed
confirmed nesting ranged from 55 to 75 days and presumably spanned four to five
clutches. Mean interesting interval is 21.8 days based on the first group of
observations and 21.1 days for the second, or 21.4 for both groups and 80
interesting intervals overall. Because the females tend to wait for rain before
nesting, this actual or "environmental mean" almost certainly exceeds the
"physiological mean" (i.e., the innate capacity to produce subsequent clutches) by
several days.
Of 115 observations that followed a nesting emergence for which nesting was
not confirmed (Fig. 13), nearly all fell within groups that suggested relatively
prompt nesting/renesting efforts following handling. Of these, 54 intervals
suggested renesting within 10 days of abandonment (above), 40 fell within two
0
0.0{
30
. . .. I I ]
)0
BULLETIN FLORIDA MUSEUM NATURAL HISTORY VOL 41(2)
A N = 61
c 69
0
C
0
0 20 40 60 80 100
Days between Captures
8 B N= 101
> 6
4-
0
2
0 20 40 60 80 100
Days between Captures
Figure 13. Within-season intervals between nesting emergences of P. concinna at Wakulla Springs State
Park for the combined years 1988-1991. (A) intervals following confined oviposition; (B) intervals
following emergences for which oviposition was not confinned Same-day recaptures (n=14) are omitted, as
they may not have involved a return to the water.
8 T
Fiue1.Wti-esoitrasbtennsigeegne fP ocnaa aul pig tt
Pakfc tecmie ei 9S19.()itral olwn ofndoioiin B nevl
foloin emrecsfrwichoioiinwsntcnme.Sm-a eatrs( 4 r mtea
thymynthv novdartr otewtr
JACKSON & WALKER: REPRODUCTION IN SUWANNEE COOTER
Table 8. Minimum nesting season of five P. concinna observed across the greatest span of days within a
season at Wakulla Springs State Park. Earlier or later nestings may have occurred.
Turtle First Observed Last Observed Time Span
Number Emergence Emergence (Days)
PC-131 27 Apr 1989 4 Aug 1989 99
PC-65 1 May 1989 1 Aug 1989 92
PC-275 25 Apr 1991 25 Jul 1991 91
PC-168 4 May 1989 23 Jul 1989 80
PC-80 30 Mar 1990 16 Jun 1990 78
days of the single-nesting interval group (i.e., 17-32 days), and 11 fell within 3
days of the two-nesting intervals group (i.e., 38-53 days).
With a mean nesting season length of 117 (minimum) to 128 (maximum)
days (Table 6) and a mean interesting interval of 21.4 days, Wakulla River
females might lay up to six or seven clutches per year. As many as nine are
feasible, though perhaps unlikely, if clutches are laid repeatedly at shorter
intervals. To what extent females approach this level of production (i.e., before
energy reserves are depleted) is unknown, although individual turtles were
observed utilizing nesting seasons up to 99 days (Table 8), ample time to produce
six clutches. However, because of the difficulty in intercepting a turtle each time
she nested, and because after 1988 we processed most turtles before they began
laying, the most clutches that we actually observed being laid by any female in one
season was three.
Data from dissections (all sites) generally support the conclusions above.
Most revealing was a mature female (Withlacoochee River; PL 339 mm, mass 6.4
kg), collected 7 August 1973 bearing one set of preovulatory follicles (18-20 nun)
plus corpora lutea/corpora rubra corresponding to four or likely five clutches
already produced, for a probable total of five or six clutches (ca 101 eggs). An
adult female (PL 362 mm, mass ca 6.6 kg), killed 7 July 1988 while crossing a
road that borders WSSP, presumably had just nested. She bore three sets of
corpora lutea/corpora rubra, a set of ca 20 preovulatory follicles (20-21 mm), and a
set of smaller follicles ca 10-12 mm in diameter--evidence of four and possibly
five clutches. A smaller female (PL 313, mass 4.3 kg) killed at the same site on 23
May 1976 bore 22 regressing corpora lutea (4-6 mm) and two sets of preovulatory
follicles (19-20 mm, n=20; 17-18 mm, n=22), for a potential seasonal total of at
least three clutches; a third set of follicles 10-14 mm in diameter appeared to be
initiating atresia. Corpora lutea had regressed too far to be useful in determining
numbers of clutches laid by four females taken from the Santa Fe River in early
September 1973.
Annual Periodicity.- Although we estimate that we were able to intercept
only about one-fourth to one-half of adult females in a given year at WSSP, we
BULLETIN FLORIDA MUSEUM NATURAL HISTORY VOL. 41(2)
have little doubt that most, and probably all, healthy females nest annually. Five
females were observed nesting in five of six seasons studied (1988-1993), including
three that were encountered in five consecutive years. This is especially
remarkable in light of our greatly reduced field efforts in 1992 and 1993. We
observed 15 turtles nesting in four of six years, with seven of these having been
seen in four consecutive years; 27 others were confirmed attempting to nest in
three consecutive years. Finally, the percentage of turtles observed attempting to
nest in the year after marking (Table 3) compares favorably with what might be
expected, given overall capture success, if all females nested annually.
The likelihood that annual reproduction typifies this species in the southern
part of its range is supported by our ovarian examinations of turtles from the
Withlacoochee River in Florida and similar examinations from east-central
Alabama by Fahey (1987). Neither analysis revealed any mature female (among ca
three dozen sampled) that appeared to be skipping reproduction in a given year.
Clutch Size.- Clutch size for 93 clutches at WSSP ranged from 2 to 27,
with a mean of 16.7 5.4 (Fig. 14). The relatively normal distribution of clutch
N = 93
5 10 15 20 25
Number of Eggs
Figure 14. Disribution of clutch size (number of eggs) for the Wakulla River population of P. concinna,
based on data collected from 1988 to 1993 at Wakulla Springs State Park.
JACKSON & WALKER: REPRODUCTION IN SUWANNEE COOTER
sizes makes delimitation of "typical" clutch size difficult. Following the analysis of
Bjorndal and Carr (1989) for Chelonia, we too believe that the smallest clutches
(<6) are abnormally small and probably represent either partially depredated
clutches (some eggs removed during oviposition, before nest covering) or residual
clutches, the remainder of which had been laid previously. Circumstantial
evidence for the latter comes from the detection of at least one remaining shelled
egg in a number of post-nesting females whose cloacal temperatures were being
determined. Additionally, two post-nesting females retained overnight for
photographic and telemetric purposes each subsequently dropped two shelled eggs;
similar observations have been recorded for other freshwater turtles (Ewert 1976).
We observed predation of eggs by fish crows during oviposition, yet we saw a few
turtles complete nesting despite this loss of eggs. When clutches of 6 eggs are
deleted from the data set, the mean is 17.5 + 4.5 (8-27, n=88).
Clutch size (excluding clutches <6, as above) was significantly correlated
with all five measures of maternal body size (Table 9, Fig. 15). However, the latter
accounted for only a small proportion of the variation in clutch size (5-20%).
Interestingly, the highest correlation was with carapace height, one of the least
commonly recorded body measurement in chelonian ecological studies.
The following egg counts for two clutches were available from 12 females:
12/12, 12/12, 25/25, 17/18, 17/18, 18/19, 13/15, 22/19, 23/20, 2/8, 17/11, 21/5.
Partial predation may account for the size disparity in the last three. If so, clutch
size per female appears to be relatively stable, with differences rarely exceeding 3
eggs.
Reproductive Potential.- Annual reproductive potential of Wakulla River
females beyond their first reproductive year, based on a mean clutch size of 17.5
and a. conservative estimate of four clutches per year, is at least 70 eggs. As
discussed above, some females probably approach or exceed 100 eggs.
Egg and Clutch Mass Parameters
Eggs.- Excluding abnormally small eggs (below), mean linear dimensions
of 397 eggs, representing 26 clutches (one per female) from WSSP, were 38.9 x
27.2 mm (length 29.5-46.0 mm, s.d.=2.74; width 22.7-30.6 mm, s.d.=1.32). The
average elongation of all eggs was 1.43. Mean egg mass was 16.28 2.14 g (9.79-
21.65 g, n=368 eggs from 24 clutches). Mean egg mass per clutch ranged from
12.37 g to 19.80 g for 25 clutches. Substantial egg size variation existed within
clutches. The difference in mass between the smallest and largest "normal" eggs
within each of 19 clutches averaged 3.36 g (range 1.88-5.84 g).
Egg mass, width, and length each were regressed against two measures of
body size (PL and mass) for 20 adult females (Table 9). Only the effects of female
PL on egg width and of female mass on egg length (Fig. 16) were significant
(Table 9), but each accounted for <10 percent of the variation in egg size.
However, the negative effect of egg size (mass) on clutch size, after adjusting for
BULLETIN FLORIDA MUSEUM NATURAL HISTORY VOL. 41(2)
female size (PL) by ANCOVA, was highly significant (r2=0.17, F=13.6,
p=0.0003).
Clutches at WSSP included two types of anomalous eggs. Two small yolkless
eggs (14 x 11 mm; 16.8 mm x 14.7 mm, 1.94 g) with well calcified shells were
found in nests with otherwise normal clutches of 11 and 23 eggs. Roughly 5-10
percent of clutches laid throughout the nesting season included eggs with unusually
thin and poorly calcified shells (<0.2 mm vs. 0.3 mm in normal eggs). Typically,
entire clutches were thin-shelled, but two nests contained both normal and thin-
shelled eggs. Whether such eggs were laid prematurely in response to rain, as we
suspect, or whether they might reflect dietary deficiencies or physiological
problems (Erben et al. 1979) could not be determined. Most thin-shelled eggs
failed to initiate development when incubated although a few did. In one mixed
clutch, only eggs with normal shells developed.
Clutch Mass and Relative Clutch Mass.- Estimated mean clutch mass at
WSSP, based on a mean clutch size of 17.5 and mean egg mass of 16.3 g, was 285
g. Actual masses of 24 measured clutches from different females (325-382 mm
PL, mean=343 mm 15.3; 4.9-7.9 kg spent body mass, mean=6.4 0.8 kg) ranged
from 168.3 g to 424.4 g (mean=304.9 62.8 g). Clutch mass was correlated
significantly and positively with female size as represented by PL and body mass
(Fig. 17, Table 9). Relative clutch mass (RCM; clutch mass as a fraction of spent
body mass) for the same data set averaged 0.048 .008 (0.034-0.061) and was not
correlated with maternal size (Fig. 18, Table 9).
Development, Sex Determination, and Hatchlings
Laboratory Incubation and Developmental Anomalies.- Eggs were
incubated to hatching successfully at constant temperatures ( 0.50C) ranging from
250C to 330C and required 58-122 days before pipping, depending inversely upon
temperature (Table 10). The small sample of eggs maintained at 290C pipped
prematurely, presumably as a result of severe dehydration (Ewert 1985:232). Eggs
shifted from one constant temperature to another after 10-30 days generally pipped
at intermediate intervals. All clutches exposed to naturally fluctuating daytime
temperatures (range 19-3 10C, but usually 24-280C) throughout incubation pipped
in 80-89 days (Table 10). Within 25 clutches or partial clutches incubated under a
single temperature regime and producing at least three hatchlings, pipping
spanned an average of 4.3 days (range 1-11). The onset of pipping among clutches
(from different females) at the same temperature spanned only 5-6 days for all
temperature regimes except one; pipping among eight clutches held at ca 25C
began after 91-115 days. Whether this was a result of poorer temperature control
or a more variable rate of development at lower temperatures was not determined.
Swelling of eggs as a result of water absorption followed the typical pattern for
pliable-shelled emydid eggs (Cagle 1950; Ewert 1985).
JACKSON & WALKER: REPRODUCTION IN SUWANNEE COOTER
Table 9. Linear regression analyses of relationships among maternal body size, clutch size, egg size,
hatchling size, and clutch mass for the Wakulla River population of P. concinna. "Clutches" of <6 eggs
were presumed to represent partial clutches (see text) and were eliminated from analyses. Body size
measurements are based on mm and kg (adults) or g (hatchlings); female body mass refers to spent females.
Values ofp < 0.05 are considered statistically significant (*). Samples sizes of eggs are given as number of
eggs, number of clutches (one per female).
Reproductive
Traits Y-INT Slope r2 N F P
Egg Size vs
Female PL
Egg length'
Egg width'
Egg mass2
Egge Size vs
Female Body Mass
Egg length'
Egg width'
Egg mass2
Hatchling PL vs
Female Size
Female PL
Female mass
Mean Hatchling Size vs
Mean Egg Mass
Hatchling PL
Hatchling mass
Clutch Size vs
Female Size
Plastron length
Carapace length
Carapace width
Carapace height
Body mass
Clutch Mass vs
Female Size
Plastron length
Body mass
Relative Clutch Mass vs
Female Size
Plastron length
Body mass
41.8 -0.007 0.00
20.0 0.022 0.08
14.8 0.005 0.00
43.1 -0.001 0.03
27.2 0.000 0.00
17.8 0.000 0.01
19.1 0.044 0.12
35.5 0.000 0.00
24.5 0.614 0.76
1.35 0.644 0.86
-23.2
-23.3
-10.7
-5.02
4.94
0.118 0.15
0.107 0.18
0.099 0.05
0.147 0.20
0.002 0.15
-517 2.399 0.31
27.5 0.044 0.26
0.025 0.000 0.01
0.048 0.000 0.00
331,20
331,20
311,19
331,20
331,20
311,19
0.63 0.4291
26.8 0.0001 *
0.50 0.4785
11.0 0.0010 *
0.15 0.7034
2.33 0.1277
199 27.1 0.0001 *
184 0.69 0.4069
13 35.1 0.0001 *
13 66.2 0.0001 *
14.1 0.0003 *
17.7 0.0001 *
4.53 0.0364 *
17.2 0.0001 *
13.4 0.0005 *
20 8.21 0.0103 *
20 6.26 0.0222 *
20 0.26 0.6173
20 0.00 0.9774
' female sample: 343.3 15 3 mm (325-382 mm), 6352 818 g (4.9-7.9 kg)
2 female sample: 342.2 14.9 mm (325-382 mm), 6345 838 g (4.9-7.9 kg)
BULLETIN FLORIDA MUSEUM NATURAL HISTORY VOL. 41(2)
300 310 320 330 340 350 360 370 380 390
Plastron Length
330 340 350 360 370 380 390 400 410 420 430
Carapace Length
JACKSON & WALKER: REPRODUCTION IN SUWANNEE COOTER
50 260 270 280 290 300 310
Carapace Width
*
*
*
*
*
*
* ^ / -
i ~ ~ ~ * / "' "^
* *
** *
* ** **
** ^- ^ *
** ^ --e "'^ ***
*^-^'' **
^^^* *
120 130 140 150 160 170 180
Carapace Height
Figure 15. Scattergrams of the relationship between clutch size and four measures of body size (in mm) for
88 P. concinna at Wakulla Springs State Park, 1988-1991. Sample sizes as follows: PL=82, CL=82,
CW=81, CH=71. See Table 9 for associated correlation and regression statistics.
BULLETIN FLORIDA MUSEUM NATURAL HISTORY VOL 41(2)
*
*
:j |i i:
*
; s .
: ": .: : I
,, "a
** j. : *
-- *r s' a
*j : *
:
.
**
320 330 340 350 360 370 380 390
Female Plastron Length (mm)
I
I
5000
800O
Female Body Mass (g)
Figure 16. Relationship between egg size and female size for 20 P. concinna from Wakulla Springs State
Park. (A) egg width vs female PL; egg width=0.022 PL + 20.0, r2=0.08; (B) egg length vs female body
mass; egg length=-0.001 body mass + 43.1, r2-0.03.
30.0-
29.0
-.28.0o
e27.0
26.0
0L25.0
24.0
23.0
22.0
310
46.0
45.0
44.0
43.0
.42.0
41.0
E40.0
i-39.0
138.0
c37.0
36.0
0)35.0
34.0
33.0
32.0
31.0
30.0
29.0
4000
JACKSON & WALKER: REPRODUCTION IN SUWANNEE COOTER
500.0
400.0
0 300.0
200.0
310 320 330 340 350 360 370 380 390
Female Plastron Length (mm)
500.0
-4000
300.0
o 300.0
2000
4000
5000
Female Body Mass (g)
Figure 17. Relationship between clutch mass and two measures of female body size (PL, spent body mass)
for 20 P. concinna from Wakulla Springs State Park. (A) Clutch mass=2.399 PL 517, r'=0.31; (B)
Clutch mass=0.044 body mass + 27.5, r'=0.26. Both relationships are significant (Table 9).
BULLETIN FLORIDA MUSEUM NATURAL HISTORY VOL. 41(2)
Excluding abnormal eggs (above) and clutches utilized for sex-determination
studies, 283 of 496 eggs (57.1 percent) from 32 laboratory-incubated clutches
hatched. Hatching within clutches ranged from 0 to 100 percent. Failed eggs died
at all stages, from no conspicuous development to full-term embryos (no obvious
differences among temperatures); 15 percent of incubated eggs (76 of 496) failed to
"chalk" (Ewert 1985) and may have been infertile.
Incubated eggs yielded a variety of developmental anomalies typical of turtles
(e.g., Ewert 1979). Scute anomalies typically involved the vertebrals and
occasionally the costals and marginals. Candling of eggs revealed two sets of
twins, to our knowledge the first documented for the genus Pseudemys (Plymale et
al. 1980). One pair that shared a common yolk sac reached near-term at 270C but
died before pipping. An egg from another clutch at 250C contained two clearly
distinct embryos at stages 9 and 10 (Mahmoud et al. 1973) on day 26 but
subsequently yielded only one normal hatchling.
Two eggs in a clutch that otherwise produced several normal young at both
250C and 300C yielded a pair of severely teratogenic individuals at 250C. Both
were megacephalic and characterized by encephalocoeles and anophthalmia (Ewert
1979). One was severely prognathous and had exaggerated feet and claws and
reduced pigment but normal scutellation, whereas the other was normally
pigmented but bore an asymmetric, kyphotic shell with severe scute anomalies.
Neither was able to pip its shell, but both were still alive and possessed large yolks
25 days after siblings had pipped.
Natural Nests: Temperatures and Hatchling Emergence.- Daytime
temperatures of four nests (Table 11: no. 3, 5,. 13, 24) monitored randomly
throughout the normal developmental period averaged 26.7-28.60C at the center of
the clutch. Two of these were among the most exposed and unshaded of all nests,
yet they still averaged only 27.40C and 28.6C although occasionally reaching
maxima of 30.50C and 32.00C, respectively. A fifth nest (Table 11: no. 29)
constructed in the shade of a building averaged 24.5C.
In general, daytime nest temperatures (Tn) were far more stable than ambient
air (Ta) or soil surface temperatures. T. (at nest center) typically rose slowly but
usually lagged 2-50C behind Ta throughout the day until late afternoon (ca 1800
hr) when Ta often dropped below Tn. Precipitation altered this pattern. During
and after thunderstorms, T. often dropped sharply as cool air masses replaced
warmer air. At such times, the more stable Tn often remained several degrees
above Ta before itself cooling slowly.
Table 11 summarizes the dates of emergence and hatching success for all
protected nests (n=30). In all, only 13 of these nests produced hatchlings. Despite
our efforts to prevent predation, it appeared that subterranean predators (e.g.,
moles, ants) destroyed part or all of 11 clutches. However, losses attributable to ant
predation on whole eggs were indeterminable, since these eggs could have died
from other causes. At least one hatchling was produced in 11 of the 19 non-
depredated nests (58 percent), as well as in two of the partially depredated nests.
JACKSON & WALKER: REPRODUCTION IN SUWANNEE COOTER
310 320 330 340 350 360 370
Female Plastron Length (mm)
0.0617
0.0597
0.0577
0.0557
, 0.0537
0.0517
- 0.0497
0.0477
D 0.0457
S 0.0437
a 0.0417
0.0397
0.0377
0.0357
0.0337
4000
5000
6000
Female Body Mass (g)
380 390
7000
8000
Figure 18. Relative clutch mass plotted against two measures of female body size (A, PL; B, spent body
mass) for 20 P. concinna from Wakulla Springs State Park. Neither relationship is significant (Table 9).
0.0617
0.0597
0.0577
0.0557
, 0.0537
S0.0517
0.0497
0.0477
J 0.0457
0.0437
S0.0417
0.0397
0.0377
0.0357
0.0337
BULLETIN FLORIDA MUSEUM NATURAL HISTORY VOL 41(2)
Table 10. Incubation periods for P. concinna eggs under constant, naturally fluctuating, and step-shifted
temperature regimes; data presented as number of eggs pipped and mean and range in days to pipping. All
eggs were from Wakulla Springs State Park (WSSP), Florida, unless indicated otherwise.
T ('C) N Days to Pipping Locality Source
Constant temperatures
22.5 0 failed to pip' WSSP this study
25 97 103.9 (91-122) WSSP this study
25 42 82.1 (79-85) Tennessee Ewert (1985)
25 6 86.0 (82-91) Tennessee Ewert (unpubl.)
27 24 82.5 (77-93) WSSP this study
28 16 72.9 (68-75) WSSP this study
29 6 65.0 (62-69)2 WSSP this study
29 35 60 southern Louisiana3 Fahey (1980)
29 66 southern Missouri Turner (1995)
30 52 66.6 (62-73) WSSP this study
30 38 55.3 (53-59) Tennessee Ewert (1985)
30 9 60.6 (58-62) Tennessee Ewert (unpubl.)
32 4 60.5 (59-61) WSSP this study
33 2 58.5 (58-59) WSSP this study
Naturally fluctuating temperatures
24-28 (19-31)4 90 84.5 (80-89) WSSP this study
room temperature 54 (84-92) north-central Florida Jackson &
Jackson (1968)
Step-shifted temperatures
25/30 (day 10-12) 20 71.2 (65-76) WSSP this study
30/33 (day 25) 3 65.3 (65-66) WSSP this study
30/27 (day 30) 2 73.0 (73) WSSP this study
1Three eggs opened at 161-166 days contained dead, near-term embryos.
SEggs hatched prematurely because of severe dehydration.
'based on taxonomic reassignment by Ward (1984).
typically 24-28 but occasionally as low as 190 orashighas31*.
Seventeen intact nests with known clutch sizes had an overall hatching success of
30.5 percent (90 of 295 eggs), with a per nest range of 0-94 percent.
Of the 13 successful clutches (i.e., those producing at least one hatchling), six
emerged in the fall, 100-157 days post-nesting, whereas five overwintered in the
nest and delayed emergence until the following spring, ca 242-318 days post-
nesting. The remaining two both contained live hatchlings when excavated by us
in November and may have been prepared to overwinter. However, despite our
efforts to re-seal one of the nests, the clutch emerged within 2 hours (DePari
[1996] recorded a similar response to disturbance in Chrysemys picta).
Additionally, three observations of hatchlings on River Road during early spring
(31 March and 1 April 1990, April 1996) provided evidence of terrestrial
overwintering by hatchlings in unprotected nests.
Table 11. Data for protected nests of P. concinna at Wakulla Springs State Park. All were constructed in moderately compacted sand. PC# = turtle/clutch number, H/E
= number of live hatchlings/number of eggs laid; M:F = number of male:female hatchlings sampled; L-M-R = distribution of eggs in left, middle, and right holes; G =
groundcover (dg, mg, sg, vsg = dense, moderate, sparse, and very sparse grass; bs = bare sand); sun (max) = total hours (with subsets from 0900-1500 EST, 1200-1800
EST) of solar exposure on sunny day, daytime temperatures based on random daytime readings, from nesting through mid-September, of thermal probe at center of
clutch (n and one standard deviation in parentheses); loc. = location (see text); x = unknown value.
Daytime
Temperature
Date Date Days to Sun (C)
No. PC# Laid Emerged Emerge H/E M:F L-M-R G (max) MW Min Max Loc. Comments
1988
1 73a 88-05-10 >88-11-11b >185b
2 33a 88-05-13 -
3 34a 88-05-23
4 110a 88-06-26 88-10-12
5 115a 88-07-03 -
6 154a 88-07-14 88-10-22
1989
7 10a 89-04-20
8 76b 89-04-27
108
0-14-0 sg 6(5,3)
0-2?-0 dg 5 (3,5)
5/w 4 undeveloped,
3 died near-term
- - 7/e possible ant
predation
0/16' 0-16-0 bs/vsg 6(5,2) 26.7 19 31 45/46 probable ant
(64, 2.47) predation of
hatchlings or near-
term embryos
5/14 0-14-0 bs/vsg 6(3,6) 5/6 8 undeveloped
eggs, 1 dead term
embryo
0/25 0-25-0
100 16/22
- 0/>7
0/16
0-22-0
bs 4(4,2) 26.7 24 28.5 27/e
(25, 1.33)
bs/sg 8(6,4) 5/w scute anomalies
on several
1->6-0 vsg/bs 9 (6,5) -
0-16-0 sg 2(1,0) -
- 13/m
- 23/m eggs gone by July,
mole tunnel
Table 11 Continued.
Daytime
Temperature
Date Date Days to Sun (OC)
No. PC# Laid Emerged Emerge H/E M:F L-M-R G (max) M' Min Max Loc. Comments
9 157a 89-04-29
10 85c 89-06-07
11 214a 89-06-08
12 25a 89-06-16
13 215a 89-06-16
0/14 0-12-2 vsg 4(3,1)
0/25 1-24-0 mg 7(4,5)
0/20 0-20-0 vsg 6(5,4)
89-11-20
15/w nest flooded, eggs
failed to develop
- 5/e possible ant
predation
22/m possible predation
by mole/small
mammal
0/18 0-18-0 sg 1 (1,1) 9/m eggs gone by
Aug; mole tunnel
157 17/12 0-12-0 mg 11(6,6) 27.4 22 30.5 7/e possible ant
(66, 1.62) predation
14 44b 89-07-01
15 182a 89-07-12
16 145a 89-07-19 ca 90-03-18d ca 242?'
17 35a 89-07-30
18 172b 89-08-20
0/14
0/23?
1-12-1 sg 5 (5,2)
0-21-2? bs 5 (2,5)
1/13 0-13-0
0/20
0/17
bs 3 (3,3)
25/w possible ant
predation
13/m eggs failed to
develop, attacked
by insects
28/e most eggs failed
to develop
(drowned or
infertile?); 1 dead
near-term embryo
1-19-0 bs 5(4,1) 14/mw
1-15-1 m/dg 8(5,6) 8/wm embryos died
stage 21, probably
from cool
temperatures
Table 11 Continued.
Daytime
Temperature
Date Date Days to Sun (OC)
No. PC# Laid Emerged Emerge H/E M:F L-M-R G (max) M' Min Max Loc. Comments
1990
19 134a 90-04-22
20 14b 90-04-22 -
21 41b 90-05-09 91-03-23
22 142a 90-06-04 -
23 172c 90-06-07 >90-11-7 b
24 153a 90-06-22 91-03-23
25 53a 90-06-22 ca 90-10-19
26 60a 90-07-02 91-03-x
1991
27 21c 91-05-26 91-10-02
0-15-0 sg 4(4,3)
0/15 0-15-0 mg 4(3,2)
318 7/8 3:3 0-8-0 s/mg 5(4,2)
0/17 0-17-0 vsg/mud 9(5,5)
>153b 17/18 6:0 0-18-0 m/dg 8(4,6)
274 7/187 0:6 0-187-0 mg 11(6,6) 2
call9 12/18
- 1-18-0 sg,bs 8(5,6)
>240 2/18 0:2 1-17-0
128 8/x 3:1 0-x-1
vsg 8 (4,6)
24/w eggs failed to
develop
(infertile?)
27/w probable
predation/
scavenging by
insects
20/em greatest slope of
any nest (16), N-
facing
13/14 nest flooded, eggs
failed to develop
8/m 1 infertile egg
8.6 26 32 8/e 4 dead term
(14,1.92) embryos, others
undeveloped
- 34/w 6 undeveloped
eggs
- 33/m 4 dead term
embryos, 10
empty shells
sg 7(6,3) 5 possibly polluted
soil; 1 undevel-
oped egg
Table 11 Continued.
Daytime
Temperature
Date Date Days to Sun (oC)
No. PC# Laid Emerged Emerge H/E M:F L-M-R G (max) M' Min Max Loc. Comments
28 218a 91-06-02 91-09-26 115 12/17 6:0 0-17-0 mg 6(5,3) 28/e 5 undeveloped
eggs
29 255a 91-07-14 ca92-04-20 ca281 6/22 0-21-1 bs 0(0,0) 24.5 22 26 B some died near-
(30, 1.14) term >
30 U 91-07-x sg A entered by mole
tunnel W
SM=mean Z
b contained live hatchlings in nest chamber when excavated by investigator in November of nesting year hatchlings in nest I were within 5 cm of surface when excavated 11Nov and emerged within 2 h despite effort to
reclose nest hatchlings in nest 23 were inactive (soil T 12 *C)
c at least some eggs contained remains (bones or scutes) ofnear-term embryos
d Question exists as to whether the single hatchling fom cage #16 emerged in March 1990 or whetherit emerged in fall 1989 but reburied itselffor the winter. No exit hole was obvious in March, but merely a small hole
near the edge of the cage (not in contact with the nest).
o
---------------------------
99
JACKSON & WALKER: REPRODUCTION IN SUWANNEE COOTER
No obvious factor accounted for the dichotomy in emergence strategies.
Clutches from each of the three major nesting months (May, June, July), and
clutches from all four principal study seasons (1988-1991), emerged in both fall
and spring. Our small sample suggests a slight tendency for late-season (July)
nests to overwinter, particularly if they are shaded for most of the day.
Our limited data on proximate thermal conditions indicate that an abrupt
temperature change in either direction may trigger emergence, depending on the
time of year. For example, one clutch (laid 2 June) emerged on 26 September 1991
following the first autumnal cold front (Ta <150C). In contrast, two overwintering
clutches (laid 9 May and 22 June 1990) emerged on 23 March 1991 during the first
warm spell of the year (T, >300C), even though no precipitation had fallen for
more than a week.
As part of our laboratory incubation studies, we specifically monitored the
viabilities of eggs from accessory holes and found no differences between them and
the residual clutches. We did note, however, that soil temperatures in the shallow
side-holes frequently rose sharply (>350C) when exposed to the sun, so it is likely
that satellite eggs in sunny nests would suffer lethal heating or dehydration even if
they were not destroyed by predators.
Sex Determination.- Based on 98 sexable embryos and hatchlings,
interpolation of the results of constant temperature laboratory incubation revealed a
pivotal temperature of 28.40C for the Wakulla River population of P. concinna.
Females differentiated at higher temperatures (290, 30, 320C) and males at lower
ones (250, 270, 280; M. Ewert pers. comm.). Detailed results and a geographic
comparison of pivotal temperature across the species' range will be presented
elsewhere.
Protected natural nests yielded clutches of all males (two nests), all females
(two nests), and mixed sexes (two nests), based on sampled hatchlings (Table 11).
Nests producing only females were among those receiving the most potential hours
of direct sunlight, particularly throughout the afternoon. With one exception, nests
that produced at least some (50-100 percent) males received more shading from
nearby trees, especially in the afternoon; the exceptional nest may have been
insulated by a fairly dense, grassy groundcover. No hatchlings from the coolest,
shadiest nests were examined, as it was assumed that these would all be males.
Date of laying within the season may interact with degree of insolation to
determine hatchling sex, which is known to be established roughly one-third of the
way through development. The four clutches producing males were laid from 9
May to 7 June; those yielding only females were deposited 22 June and 2 July.
Based on these data and in the absence of differential predation, we suspect that
present nesting habitat at WSSP may favor male-biased primary sex ratios, with
females likeliest to develop in nests laid in sunny sites during the second half of the
nesting season.
Hatchlings.- Table 12 provides body measurements of laboratory-incubated
and naturally emerged hatchlings from WSSP. Because of their delayed
BULLETIN FLORIDA MUSEUM NATURAL HISTORY VOL 41(2)
emergence, nest-hatchlings were visibly older than incubated hatchlings when
measured and therefore were likelier to have achieved full post-yolk resorption
body proportions. Larger eggs generally yielded larger hatchlings (Fig. 19, Table
9) as expected (Ewert 1979, 1985; Roosenburg and Kelley 1996). Substrate water
potential, and to a lesser extent incubation temperature, also may affect hatchling
size (Packard et al. 1981a, b) and explain some of the variance. However, we had
maintained all substrates near saturation. The only split-temperature clutch that
produced a large number of hatchlings from eggs of known mass showed no effect
of temperature; 10 hatchlings incubated at 250C averaged 0.4 g heavier than 9
siblings incubated at 300C (13.6 g vs. 13.2 g), exactly the difference in mean
masses of the two sub-groups of eggs (19.5 g vs. 19.1 g).
Hatchlings generally matched the descriptions given by C. Jackson and
Jackson (1968), although hatchling size and plastral pattern were substantially
more variable in our large sample than suggested by the smaller samples of
Jackson and Jackson (1968) and Fahey (1987). Ventral pattern nearly always
formed a symmetric figure that followed the interscutal sulci, with dark
pigmentation covering approximately 10-30 percent of the plastron. Even in
individuals with reduced pattern (infrequent at WSSP), dark pigment still occurred
anteriorly along the gulo-humeral and interhumeral sulci, and posteriorly along the
abdomino-anal and internal sulci. Dark pigment along more medial sulci in such
lightly patterned individuals was reduced to isolated spots or bars of various
shapes.
Hatchlings from laboratory-reared eggs typically lost the egg tooth between
two and four weeks post-pipping, by which time yolk sac ("umbilical") scars had
nearly closed. However, one clutch of hatchlings maintained under cool
Table 12. Standard measurements of laboratory-incubated and naturally emerged hatchling P. concinna
from Wakulla Springs State Park. All members of the latter group were weighed within 24 hours of
emergence in the fall. Sample size (n) given as number of hatchlings, number of clutches represented.
Mean Range S.D. N
Laboratory-incubated
PL 34.2 29.0 -38.7 2.06 225, 22
CL 38.1 32.3 -43.2 2.36 225,22
CW 37.3 29.2-43.6 2.19 222,22
mass 11.4 6.8- 14.6 1.83 179, 18
Natural nests
mass 13.1 11.5-14.6 0.77 36,4
JACKSON & WALKER: REPRODUCTION IN SUWANNEE COOTER
15.0
14.0
"913.0
12.0
| 11.0
I
)10.0
9.0 *
8.0
12.00 13.00 14.00 15.00 16.00 17.00 18.00 19.00 20.00
Mean Egg Mass (g)
Figure 19. Mean hatchling mass vs mean egg mass of P. concinna from Wakulla Springs State Park for 10
laboratory-incubated clutches that produced at least five viable hatchlings. Hatchling mass=0.644 egg mass
+ 1.35, r'=0.86, p=0.0001.
moist conditions retained caruncles through December, more than two months
post-hatching.
Hatchlings exhibited a plasticity of behavioral patterns in the laboratory.
Following initial pipping activity, they typically remained calm and immobile and
rarely left the egg shell until most or all of the external yolk sac had been
withdrawn into the body, a process typically requiring 1-3 days. Subsequent
behavior was marked either by restlessness or by burrowing into the incubation
substrate. Hatchlings immersed in water within one week of pipping were adept in
their swimming ability.
A total of 368 hatchlings (278 from laboratory incubation plus 90 from field
nests) was released at WSSP within one month post-pipping in shallow water at
the river's edge. Most swam within seconds to vegetative cover, usually at the
bottom (<10 cm deep) but occasionally near the surface. One 1988 hatchling
recaptured 18 May 1991 measured 112 mm PL (123 mm CL; 300 g) after >2 years
of growth.
BULLETIN FLORIDA MUSEUM NATURAL HISTORY VOL. 41(2)
Predation and Mortality
Nest Mortality.- Our subjective impression from observing several hundred
nestings throughout the study period was that extremely few survived predation at
WSSP during any of the study years. To test this, we monitored 114 undisturbed
nests made under various conditions during 1989 and 1990; none survived more
than 48 hours. That predation pressure is greatest on fresh nests was underscored
by our sample of 30 protected nests. Digging at 29 cages occurred within 48 hours
but waned almost completely after one week.
The two major nest predators at WSSP are the fish crow and raccoon. At
WSSP, both species were observed not only depredating completed nests but also
actively stealing eggs as soon as they were oviposited, in some instances before the
eggs had even fallen into the nest! Both species appeared to search specifically for
turtle nests and nesting females by respectively flying above or walking along the
most commonly used nesting sites. They approached females on sight and waited
nearby for oviposition to begin. Some turtles abandoned nesting efforts during
such encounters. Whereas raccoon predation usually involved a single animal or
less frequently a parent with offspring, predation by crows often involved flocking;
as many as 12 crows were observed at some turtle nests. Raccoons typically fed at
the nest, where they left eggshells with distinctive tooth marks (in contrast to their
removal or ingestion of shells as noted in some studies of smaller emydids with
smaller eggs and clutches; e.g., Burger 1977; Christens and Bider 1987). Crows,
on the other hand, flew away with individual eggs (as also observed by Shealy
1976), perhaps in part to avoid competitors or to feed young, and thereby left
empty nests with no remaining eggshells; these could not be distinguished
invariably from abandoned nests. Tracks of these two oophagous species could
often be discerned at nests built in sandier spots of the road. Our subjective
impression is that fish crow predation waned during the latter part of the cooter
nesting season (after mid-June). Both major predators ceased regularly patrolling
River Road when turtle nesting ended each year.
Some eggs were destroyed by organisms other than raccoons and crows. Fire
ants attacked eggs in several nests, including three caged nests (Table 11) in which
we found the remains of hatchlings (or perhaps full-term embryos); however, we
could not determine whether this represented predation or scavenging. At least six
unsuccessful nests, including three "protected" by us, had been penetrated by the
tunnels of moles, presumably Scalopus aquaticus. Only traces of egg shells
remained in any of these nests. Because other small mammals, including shrews
and mice, may use mole tunnels, we could not definitively identify the nest
predators in such cases. Although armadillos (Dasypus novemcinctus) foraged
extensively along River Road, we found no direct evidence to implicate them as
predators of turtle eggs at WSSP. We even saw one foraging individual walk
directly over a fresh nest without pause. Feral dogs were seen four times along
River Road during the 1988 nesting season but were not confirmed as nest
predators.
JACKSON & WALKER: REPRODUCTION IN SUWANNEE COOTER
In general, nest predators excavated all three holes and consumed the entire
complement of eggs from depredated nests (n>500). Eggless side-holes
occasionally were left intact as if olfaction or bill-probing had eliminated the
possible presence of eggs. In seven recorded instances in which some but not all
eggs from the main chamber were consumed initially, the remaining eggs were
destroyed on subsequent visits, as has been reported in other studies of turtle nest
predation (e.g., Robinson and Bider 1988).
In a single instance, we observed a nest in which all three holes had been
excavated only to the depth of the accessory holes, with the bulk of the clutch (16
eggs) remaining intact in the main chamber beneath the sand-packed neck region.
Eggshell remains indicated that at least one of the side-holes had contained an egg.
Fresh raccoon tracks were present in the excavated sand. Though an isolated event
in the present study, we nonetheless consider it instructive. From all appearances,
the predator had methodically excavated all three holes, eaten available eggs, and
left the nest, presumably to continue foraging elsewhere. Were it not for the egg
shell and our knowledge of the three-holed nest structure, we too would have
overlooked the main clutch, as the sand-packed neck was as firm and hard as the
substrate at the bottom of each accessory hole, and our initial assessment (until
seeing the shell) was that this was an abandoned, incomplete nest. We believe that
incomplete predation of this nest signified a naive predator, in this case a young
raccoon that was duped by the tripartite structure.
Physical factors also contributed to egg mortality. Periodic flooding and soil
saturation almost certainly caused the loss of at least one protected nest (Table 11:
no. 9) that had been constructed in the road shoulder along one of the lowest
sections of River Road. The collapsed, dried remains of eggs in some other
unsuccessful protected nests implicated desiccation (Ewert 1979) as a potential
factor in their failures.
Despite the prevalence of predation, it is germane to point out that there is
evidence for limited recruitment during the study period. Early in the study, we
noted at least one instance in which subsequent signs of depredated nests were
insufficient to account for .the probable number of nests made during an
unmonitored, extremely heavy rain event. Direct evidence of recruitment was
provided by a series of unmarked and variously aged young that were captured
during a 1991 study of the species' food habits at WSSP (Lagueux et al. 1995).
Also, as noted earlier, park staff observed newly emerged (overwintered) young
wandering along River Road at least three times during the study. Nonetheless,
survivorship of neonate and juvenile cooters in the river is expected to be low
considering WSSP's large population of alligators, a cheloniophagous species that
readily eats young Pseudemys (DRJ, unpubl.).
Adult Mortality.- At least 17 adult female cooters died at WSSP from 1988
through 1991 and another in 1993. The condition of the remains of 14 females
and their presence in areas used only for nesting indicated they were the results of
predation during nesting emergences. Mean PL of the 14 carcasses was 340.9 mm,
exactly the same as the population mean; hence, predation was independent of
BULLETIN FLORIDA MUSEUM NATURAL HISTORY VOL 41(2)
turtle size. Five of the carcasses, with egg shells scattered about (Plate 3), were
found on River Road within 24 hours of death. Access to the eggs and viscera of
each was through the vent or via a small hole in the skin of the hind leg or femoral
pocket. Known annual deaths for the four principal years ranged from eight in
1988 to two in 1989. Investigators may have contributed indirectly to at least two
of the deaths, one each in 1988 and 1990, which happened during renesting
attempts following nest abandonment as a result of human presence. More than
half of the 1988 deaths occurred within a 200-m stretch of River Road on which a
large adult raccoon was observed several times harassing cooters that were
attempting to nest. Although we believe raccoons were the principal and perhaps
only predators of nesting females, park staff witnessed at least three instances of
harassment by black vultures (Coragyps atratus), with one of these possibly
resulting in predation. These attacks occurred beneath a habitual vulture roosting
site.
One of the known mortalities was a female that had emerged from Sally Ward
Slough to nest along a highway bordering the park (Fig. 2) and was struck by an
automobile. Another nesting-related mortality resulted after a female, perhaps the
largest in the entire population, fell into a small but steep-sided limestone sinkhole
adjacent to River Road. She seems to have died from starvation or dehydration.
We later removed and released a second female that had fallen into the same hole.
During 1991, when additional biologists were working on the river as part of
a study of the cooter's food habits, two adult females (both marked earlier that year)
were found dead floating in the river; one of these was missing a rear foot.
Whether these turtles were mortally wounded during nesting or otherwise, such as
in the river, could not be ascertained. The absence of high-speed boat traffic
within the state park precludes this as a source of substantial mortality although at
least three turtles on-site bear propeller scars. Few alligators in the river are large
enough to take mature female cooters, but park staff reported at least two cases of
predation of adult cooters (sex unknown) by a large bull alligator (S. Cole pers.
comm.). The shells of most adult females at WSSP bear scars from unsuccessful
alligator attacks.
Although our data are not adequate to determine annual survivorship, they do
permit estimation of minimum mortality of adult females. Only 1 of 77 turtles
marked in 1988 was known to be dead through 1993, a minimum mortality rate of
0.0026 per year. Of 90 females marked in 1989, four were known to have died
through 1993, a minimum annual mortality rate of 0.011. Mean minimum rate of
mortality for these 167 turtles, which comprise more than half of the female
population (see population estimate below) is 0.009 per year. Undetected
mortalities in the river itself are almost certain to cause this figure to underestimate
actual mortality, however.
Seven different nesting females had lost a leg or foot (posterior except for
one), some of these relatively recently; an eighth had a shredded tail. Our
observations of harassment of nesting females by raccoons suggest this mammal is
responsible for most such wounds. Two other females each lacked an eye.
132
JACKSON & WALKER: REPRODUCTION IN SUWANNEE COOTER
Adult Female Population Size, Biomass, and Productivity
Choice of a population size estimator depends upon its ability to meet certain
assumptions, particularly population closure and equal catchability of individuals
(Tinkle 1958; Caughley 1977; Lindeman 1990). Population closure may be
violated by migration, recruitment, and mortality. Because of these factors, several
authors (e.g., Shealy 1976; Vogt 1980; Lahanas 1982) have expressed reservations
about the reliability of population estimates for a number of riverine turtles.
To test the assumption of equal catchability, we applied Leslie's test
(Caughley 1977) to our subset of data for turtles marked in 1988 and known to be
alive in 1991. The resulting chi-square value of 20.0 (0.5
null hypothesis that the distribution of recaptures is binomial and hence that
catchability is constant. Our sampling methodology generally favored equal
catchability: we systematically surveyed all known nesting areas, although
experience allowed us to treat a seldom-used site on the north side of the river with
less attention; and all nesting sites were sufficiently far from water that escape was
impossible once a turtle was observed. Equal catchability may have been violated
slightly if some individuals nested less frequently than others, but our sampling
regimen was unable to test this.
Population closure was probably approached in this study because WSSP
includes the entire upstream extent of the river, the species does not wander
terrestrially except to nest, and analysis was restricted to adult females. We believe
that movements to and from the river below WSSP were minimal, as the lower part
of the study area yielded few captures, adult females tended to remain in relatively
small home ranges, and occasional poaching immediately below WSSP (RNW
pers. obs.) limited the number of potential upstream immigrants. Further, high
recapture rates and a dearth of young-looking females led us to conclude that
annual recruitment of females into the adult population was extremely low and
probably roughly equivalent to the low number of known mortalities. Still, the
addition of any individuals through maturation and the loss of individuals through
death technically violates the assumption of closure when data are compared across
several active seasons (Lindeman 1990), as in this study.
In studies such as ours, in which most (75 %) of the population is marked
(Fig. 20) and adults are long-lived, simple proportional, closed model approaches
(e.g., modified Petersen index, below) may nonetheless provide highly realistic
population estimates. We therefore computed estimates from a variety of models
(Table 13), including the open population Jolly-Seber method (Caughley 1977).
For the following reasons, we place greatest confidence in the Petersen and
Schumacher's estimates of just more than 300 adult females in the WSSP standing
population. During our study, we marked and released 239 females in addition to
finding eight unmarked carcasses. The combined 1992 and 1993 samples (Table
3), which followed four years of intensive mark-recapture effort (Plate 3), included
only 9 unmarked individuals (22 percent) among the 41 different females captured.
BULLETIN FLORIDA MUSEUM NATURAL HISTORY VOL 41(2)
1990
1991
567 45678
60
50
40
I-
o 30
-0
E 20
z
10
Months/Years
E Unmarked M Marked
Figure 20. Summary of captures of nesting female P. concinna by calendar month for the 1988-1991
nesting seasons at Wakulla Springs State Park. Months are indicated numerically by ordinal position within
the calendar year. Histogram commences with beginning of marking in May (=5) 1988. Each female is
tabulated a maximum of once per month, with marking status being determined at first capture. Individuals
found dead are included from the following months: unmarked-5/88, 6/88 (2), 7/88, 5/89; marked-4/90,
5/91, 6/91 (2). Not figured because of less intensive sampling are data for 1992 (5 new, 16 marked) and
1993 (4 new, 23 marked) representing 41 individuals.
Table 13. Estimates for the size of the adult female population ofP. concinna at Wakulla Springs State
Park. Standard error follows each estimate.
1989 1990 1991 1992/1993
Bailey's Modified Petersen 305 22.5
Jolly-Seber 201.5 + 18.7 196.2 16.4 168.8 29.5
Bailey's Triple Catch 242.2 + 59.7 215 65.3
Schumacher's 306.9 + 47.7
s s*
kq
vIii
3 4
45678
JACKSON & WALKER: REPRODUCTION IN SUWANNEE COOTER
Coupling this with a potential decline (based on small samples) in the incidence of
unmarked turtles between these years, we estimate that ca 80 percent of the
standing population of females was marked by the end of our study, with only a
few unmarked females having died. We believe that Jolly-Seber and Bailey's triple
catch analyses underestimated population size as a result of assuming that non-
recaptured animals died or emigrated, whereas in reality they remained in the
population but were not seen again because of the immense amount of time and
effort this would have required.
Based on an estimate of 305 females with an average non-gravid body mass
of 6.45 kg, the standing crop biomass of adult females at WSSP is 1967 kg, or
roughly 390 kg (61 individuals) per km of river and 48 kg (7.4 individuals) per ha
of surface water. At an estimated five clutches per female per year, annual
reproductive biomass production (i.e., total egg mass only) is 434 kg/yr, or 10.6
kg/ha/yr (8.5 kg/ha/yr if only four clutches).
Male Body Size and Reproductive Cycle
Casual observation of basking and swimming male Suwannee cooters at
WSSP suggested sexual size dimorphism in body size as reported by Marchand
(1942) and C. Jackson (1970). Jackson's 10 largest females averaged 24 percent
longer (CL) than his 10 largest males. Fahey (1987) likewise noted size
differences between the sexes in central Alabama.
Body sizes of eight adult males collected in early and late August 1973,
January 1974, and March 1975 from the Withlacoochee and Suwannee/Santa Fe
rivers averaged 232 mm PL (201-272 mm), 269 mm CL (225-322 mm), and 1.92
kg wet mass (1.21-2.97 kg). Five juvenile males ranged from 105 mm to 144 mm
PL and 117 mm to 164 mm CL. We estimate that sexual maturity is achieved in
males at 170-200 mm PL. The epididymides of all adult males examined
contained mature sperm, but only those from January and March were swollen.
The three early August males were characterized by low testicular masses
(combined masses 1.20-1.74 g) and shortened foreclaws (possibly worn from
courtship: see Jackson and Davis 1972). In contrast, males from late August and
January had elongated claws and heavy testes (combined masses 2.68-7.32 g); our
sole March male had elongated claws but small testes (1.35 g).
Miscellaneous Observations
Cooters of all sizes basked regularly at WSSP, with adults of both sexes
typically utilizing larger, mid-stream sites while juveniles favored smaller,
nearshore sites, a form of habitat partitioning noted for other riverine turtle
populations (Pluto and Bellis 1986). Females typically comprised the majority of
basking adults observed, but we were unable to determine whether this reflected
the actual sex ratio or sexual differences in thermoregulatory behavior. Basking
and swimming continued year-round. Although not confirmed, we suspect that
BULLETIN FLORIDA MUSEUM NATURAL HISTORY VOL. 41(2)
turtles foraged at some level throughout the year, as C. Jackson (1964, 1970) has
reported for another population inhabiting a similar environment along the
Suwannee River, Florida. Nesting cooters at WSSP frequently bore small loads of
leeches (Placobdella) and algae (Basicladia), but the latter was profuse on only a
few individuals.
DISCUSSION
General Life History Strategy
Whereas P. concinna has evolved at least some morphological specializations
for riverine existence (mostly to improve swimming performance), it appears to
have made no similar life history concessions. In Florida, this turtle exhibits the
same basic suite of life history characteristics as broadly sympatric but chiefly
lentic emydids (Jackson 1988): large body size, sexual size dimorphism favoring
females (Marchand 1942; Jackson 1970), large clutches of relatively small eggs,
low relative clutch mass, multiple annual clutches, and prolonged nesting season.
Except for the three-holed nest that it shares with its lentic sister species P.
floridana, the reproductive characteristics of P. concinna are nearly identical to
those of the lentic-adapted, summer-nesting P. nelsoni (Jackson 1988). Further,
our results support other studies (reviewed by Berven 1988) indicating that basic
reproductive characteristics, such as clutch size and egg size, can be highly
variable within local populations. Such a suite of characteristics, coupled with
high levels of intrapopulational variability, are generally compatible with a bet-
hedging life history model (Stearns 1976; Berven 1988): juvenile mortality high
and unpredictable relative to adult mortality, late maturity, iteroparity, small
annual reproductive effort, and long life.
Geographic differences in life history parameters between northern and
southern populations of P. concinna mirror trends noted previously in emydid
species (e.g., Moll and Legler 1971; Jackson 1988). These include a longer
nesting season and consequently greater number of annual clutches in southern
populations, in addition to larger body size. The more equable climate and habitat
afforded by Florida rivers, and especially by floristically rich spring runs such as
the Wakulla River, permit year-round activity, including feeding (Jackson 1964).
Although some winter activity (e.g., occasional basking) persists in southern and
central Alabama (Shealy 1976; Fahey 1987), feeding generally is precluded several
months each year from there northward (Fahey 1987). The life history differences
thus appear to reflect the more consistent utilization of energy possible in Florida.
Additionally, soil temperatures remain sufficiently warm in Florida to support
embryonic development for half the year. This, coupled with a steady energy
supply, permits a protracted nesting season and increased reproductive potential.
136
JACKSON & WALKER: REPRODUCTION IN SUWANNEE COOTER
Home Range and Homing
In a radiotelemetric study of Trachemys scripta in South Carolina, Schubauer
et al. (1990) found that taken on an average of once per week 15 observations were
sufficient to determine 100 percent of the home range size of most adult females
monitored. Even fewer data may be needed in relatively linear habitat systems
(Kramer 1995). Our data from WSSP therefore support a linear home range
estimate of approximately 200-600 m for the typical adult female P. concinna,
which she maintains at least through nesting season, and probably year-round as
well as year-to-year. Although our sample is small, it may suggest a positive
relationship between body size and home range dimensions, as noted by Schubauer
et al. (1990) for Trachemys scripta. Average estimated linear range for our two
smallest radio-tracked turtles was 300 m, versus 525 m for the two largest.
Individual limitation of daily activities to ca 500-m segments of rivers
characterizes most riverine emydids (e.g., Marchand 1945; Moll and Legler 1971;
Sanderson 1974; Florence and Murphy 1976; Shealy 1976; Pluto and Bellis 1986,
1988; Kramer 1995; but contrast MacCulloch and Secoy 1983b). Our linear home
range estimates of 200-600 m for female P. concinna compare well with those of
Buhlmann and Vaughan (1991) for adult P. concinna of both sexes in the New
River, West Virginia, whereas C. Jackson (1970) suggested that the population he
studied along the Suwannee River, Florida, largely restricted itself to the vicinity of
a spring and its 160-m run.
Based on our estimates of home range and population size, female river
cooters at WSSP typically must overlap in home range with two dozen or more
other females, and probably with at least that many males and juveniles. In fact,
we have noted as many as 18 adult cooters sharing the same basking log. Given
their large body sizes and strong swimming abilities, adult river cooters at WSSP
could easily traverse their home ranges within 1-2 hours. Thus, we conclude that
home range size is determined by resource base (e.g., food and basking sites) rather
than other factors such as social structure. The shortage of basking sites along the
river may cause some turtles to utilize larger home ranges than necessitated by
nutritional requirements alone.
We suspect that the few long-distance movements implied by our recaptures
are not part of annual home ranges (sensu Burt 1943) but may represent between-
year movements (Shealy 1976), perhaps to new home ranges. However, seasonal
movements can not be ruled out. The downstream acquisition of barnacles is
consistent with Carr's (1952) observations of large numbers of Suwannee cooters
foraging (seasonally?) in the sea grass flats off the mouth of the Suwannee River.
Although both of the displaced females in this study might have had
familiarity with their release locations, our radiotelemetric data from other
individuals suggest that their release sites fell outside normal activity ranges.
Short-term recovery of both turtles at previously used nesting sites implies the
ability and motivation to return promptly to primary activity ranges from either
upstream or downstream locations.
BULLETIN FLORIDA MUSEUM NATURAL HISTORY VOL 41(2)
Other riverine emydids are known to home successfully (Moll and Legler
1971; Shealy 1976; Berry 1986). Following a deliberate release of several P.
concinna 5 km upstream of their collection sites, Marchand (1942) noted
downstream movements of 5 km or more. Likewise, Shealy (1976) recorded
successful homing by female Graptemys displaced 24 km either upstream or
downstream.
Reproductive Seasonality
In a comparative reproductive study of four lentic emydids in northern
peninsular Florida, D. Jackson (1988) identified two seasonal nesting patterns: (1)
summer-nesting species (Pseudemys nelsoni and Trachemys scripta), which
maintain a 3 to 4-month nesting season typically beginning in April or May and
roughly coincident with the rainy season, and (2) winter-nesting species (P.
floridana peninsularis and Deirochelys reticularia), which normally begin nesting
in September and continue to nest into or through the spring. P. concinna clearly
conforms to the former pattern. Reproductive patterns thus support morphology
(e.g., Seidel 1981; Ward 1984) in documenting the specific distinctness between P.
concinna and P. floridana, sister taxa that sometimes have been confused (Carr
1952; Fahey 1980; Jackson 1995) and which do coexist in some lotic habitats with
relatively low flow. Auffenberg (1978) may have confused the two ecologically
ihen he gave the nesting season of P. c. suwanniensis as "September through May
(all year?)." Outside of peninsular Florida, P. f. floridana presumably shifts its
nesting season to a typical summer pattern in acquiescence to a more strongly
temperate climate (Jackson 1988).
With two exceptions, all documented nesting records for P. concinna from
throughout its range (Table 1) fall within the period of known nesting at WSSP.
We believe C. Jackson and Jackson's (1968) report of September eggs in Florida to
be an error, and Turner's (1995) documentation of late August nesting at one
Missouri site to be an artifact of human alteration of habitat. The general
coincidence of nesting season extends as well to the more western "sister"
populations of river cooters currently assigned to P. texana (Table 1). The limited
data available for more northern populations, including those in east-central
Alabama only 340 km to the north-northwest, indicate truncation of the nesting
season at one or both ends, with most known natural nesting restricted to the
period of late May to mid-June. Only in southern Missouri (Turner 1995) has
nesting been shown to span more than one month. Thus, the nesting season in
Florida is from two to eight times as long (about 4 months vs. 2 weeks to 2
months) as it is elsewhere.
Duration of the nesting season at WSSP seemed to be little affected by the
marked annual variations in weather pattern, chiefly precipitation, that
characterized the study period. For instance, 44.2 cm of rain fell in June 1989 in
nearby Tallahassee, 27.6 cm above normal and a record for the month, yet the
139
JACKSON & WALKER: REPRODUCTION IN SUWANNEE COOTER
cooter nesting season was equivalent to that in drier years (Table 6). Nesting
season length is subject to physiological constraints, but these may be relaxed for
an aquatic herbivore living in a generally stable spring-run environment such as
the Wakulla River. Nonetheless, for the cooter, which depends upon basking to
elevate Tb to raise metabolic processes, seasonal changes in air temperatures
present an appreciable constraint on feeding, growth, and perhaps other activities.
This is supported by C. Jackson's (1964) observations of a cyclic growth pattern in
a spring-run population of P. concinna within the Suwannee River drainage.
Annual onset of nesting varied by 2-3 weeks (Table 6). Early nesting was
associated with unusually warm air temperatures and abundant rainfall in 1990
and 1992. We thus suspect that late winter-early spring ambient air temperatures,
perhaps in conjunction with photoperiod, comprise the chief proximate factor
governing the annual onset of oviposition for P. concinna. A similar relationship
may exist for other reptiles at this latitude (e.g., American alligator: Hall 1991), as
well as for some northern emydid turtles (Congdon et al. 1983; Gibbons and
Greene 1990). In contrast, whereas water temperature may regulate reproductive
chronology of some highly aquatic turtles (Obbard and Brooks 1987; Polisar 1996),
thermal stability of the Wakulla River makes this unlikely for the local cooter
population.
Although energetic may limit the number of clutches that a female lays
annually, we found evidence that termination of nesting also may be regulated by a
proximate environmental factor. One 1989 nesting (Table 11: no. 18) was
unusually late (Aug 20) as a consequence of observer-induced abandonments of
earlier nesting attempts. Although the nest site received day-long sunshine, the
eggs nonetheless failed to hatch. Our examination of the nest the following spring
revealed dead stage 21 embryos in all eggs. In the absence of other evidence, we
feel that nest mortality resulted from exposure to the lower soil temperatures of fall
or winter. Turtle embryos at this advanced stage appear unable to tolerate
temperatures as low as more advanced (full-term) embryos or hatchlings
(MacCulloch and Secoy 1983a; Storey et al. 1988; St. Clair and Gregory 1990;
Lindeman 1991; M. Ewert pers. comm.). Thus, cool soil temperatures appear to
preclude extension of the nesting season beyond early August. In contrast, the
ability to yolk follicles extends beyond this, both in Florida and Alabama (Fahey
1987).
Seasonality of other reproductive activities is incompletely known for P.
concinna. Fahey (1987) showed that the male reproductive cycle in east-central
Alabama (Tallapoosa River) generally coincides with that of most North American
temperate zone emydids; i.e., spermiation occurs in the fall, and sperm are stored
in the epididymides throughout the year but decline in number as spring (and
presumed mating) progresses (Moll 1979; Licht 1982). Our limited data for
northern Florida coincide with this. The following observations suggest that
courtship and perhaps mating in lower Gulf Coastal Plain river cooters may occur
nearly year-round, though these activities may be most intensive in the spring.
Within the Suwannee River system in northern Florida, Marchand (1942, 1944)
BULLETIN FLORIDA MUSEUM NATURAL HISTORY VOL 41(2)
reported courtship and probable mating in January, and C. Jackson and Davis
(1972) noted courtship in February. We observed interacting male-female pairs in
April at WSSP. Our observation of extremely worn foreclaws in adult males in
early August in northern Florida is unique and potentially reflects intensive
courtship efforts in preceding months. In Alabama, Fahey (1987) recorded
courtship and mating in March and April, but he suspected from other observations
(males trailing females year-round, oviducts with active sperm in December) that
copulation, though probably commonest in the spring, may occur at almost any
time of the year. One of us (DRJ) observed several interacting male-female pairs at
Fahey's study site in September 1994.
Diel Nesting Cycle, Proximate Nesting Cues,
and Body Temperature During Nesting
The exclusively diurnal nesting pattern of the Wakulla River cooter
population is similar to that of a well studied population of painted turtles
(Chrysemys picta) in Michigan (Congdon and Gatten 1989) and probably finds
parallels among many freshwater emydids (e.g., Vogt 1980). Based on a sample of
14 nesting emergences, Fahey (1987) noted that P. concinna in Alabama likewise
nests diurnally, though he reported no association of nesting with rainfall.
Whereas diurnal nesting appears to typify freshwater emydids that we have
observed in Florida (e.g., P. floridana, P. nelsoni, Trachemys scripta, Deirochelys
reticularia, Graptemys barbouri), some emydids living in warm climates (e.g.,
Graptemys nigrinoda, southern Alabama: Lahanas 1982; Pseudemys alabamensis,
southern Alabama: Fahey 1987; Trachemys scripta, Panama: Moll and Legler
1971) nest mostly at night.
Association of nesting with rainfall has been noted for diverse species of
turtles elsewhere (Goode 1967; DePari 1996; Polisar 1996; as well as
Macroclemys, Chelydra, and several emydids in northern Florida: DRJ and M. A.
Ewert unpubl.), though rarely has it appeared to be linked so tightly. From the
coincidence of nesting emergences with rainfall, it is clear that female Suwannee
cooters, even though immersed in water, can accurately monitor precipitation.
Much as the "sound" of rain is known to induce spadefoot toads (Scaphiopus) to
emerge from below ground. (Dimmitt 1975), rain (medium to hard for 20 minutes)
striking the river surface appears to serve as a visual, tactile, and/or auditory cue
for gravid female cooters to emerge to nest. Although quantity of rainfall is clearly
important, other interacting climatic variables obscure further definition of
emergence cues. Air temperature, duration and time of rainfall during the day, and
time since last rainfall all seem important; barometric pressure may also play a
role. A long steady rainfall beginning at 0900 hr at an air temperature of 25C
will stimulate more turtles to emerge than an equivalent amount of rain falling in a
short period during the night or at an air temperature of 190C. Likewise, a
relatively light rainfall (e.g., 3 mm) that follows a 2-week period without rain is
likely to stimulate far more turtles to emerge than a heavier rainfall occurring
JACKSON & WALKER: REPRODUCTION IN SUWANNEE COOTER
within 48 hours of a previous rainfall. The effect of season is also significant, with
fewer females likely to emerge early or late in the nesting period. Despite
collecting extensive climatic data, Semlitsch (1985) was unable to determine the
precise relationships of such variables to salamander breeding migrations.
Concordance of nesting and embryonic development with precipitation
provides at least three benefits: (1) potential reduction in predation both of eggs
and of nesting females; (2) temporarily reduced ambient temperature enabling the
large, black females to nest diurnally with minimal thermal stress; and (3) reduced
hydric stress to developing eggs in fresh nests. Water availability and temperature
during incubation of emydid turtle eggs are known to affect hatching success,
duration of incubation, mass and linear measurements of hatchlings, mass of
residual yolk, and sexual differentiation (Packard et al. 1981; Gutzke et al. 1987).
Pliable-shelled emydid eggs typically absorb water and swell during their first few
weeks (Packard et al. 1981a, b); rain water supplements bladder water in assuring
that the early nest environment provides ample moisture to support this. On the
other hand, nesting during rainfall can reduce the risk of potentially fatal nest
flooding by allowing females to avoid flood-prone sites.
Anti-predator benefits of nesting during rains may be manifold. At WSSP, a
major egg predator, the fish crow, is less active during rainfall and hence less
likely to observe nesting female turtles. Furthermore, substantial rainfall (e.g., 116
mm from 6-8 June 1989) rapidly obliterates physical and olfactory signs of nesting,
which would render nests less evident to raccoons. Carr (1952) believed this to be
critical for survival of nests of P. floridana, whose nesting activities also are known
to be enhanced by rain (Thomas 1972). Carr's (1952:295) conjecture that "most of
the peninsular turtles alive today owe their existence to the fortuitous coming of a
heavy shower soon after their mother laid her eggs" may be as applicable to P.
concinna as it is to the P. floridana of which he spoke. Finally, in a habitat that
supports a finite number of individual predators (such as WSSP), synchronization
of nesting among dozens of gravid females reduces the probability that a given
female or nest will be encountered by a predator during the period of greatest
vulnerability (as shown, for example, by Robinson and Bider [1988] for Chelydra,
and Eckrich and Owens [1995] for Caretta).
Although nesting during rain seems to be a common phenomenon among
turtles, at least one report documents the termination of nesting effort as a result of
precipitation (Vogt 1980). Such an inhibitory effect may be more common at
northern latitudes, where rainfall during nesting might drop T. too low to support
turtle activity.
Body Temperature During Nesting.- From a thermal perspective, the daily
nesting time of turtles is a compromise between maintaining body temperature
above some minimum level necessary for activity yet below the critical thermal
maximum (Congdon and Gatten 1989). We doubt that ambient temperatures
regularly preclude P. concinna from nocturnal nesting in Florida. Although the
nightly minimum T. near WSSP often dips below 220C until mid-July (Winsberg
1990), Ta usually remains above this and would allow nesting for several hours
BULLETIN FLORIDA MUSEUM NATURAL HISTORY VOL 41(2)
after sunset Further, our Tb data suggest that this large-bodied turtle possesses
sufficient thermal inertia to complete nesting easily on most nights. Still, nesting
terminates well before dark.
On the other hand, nesting may be curtailed during early afternoon on sunny
days because the air becomes hot and substrates become even hotter, sometimes
exceeding 50C at WSSP. Suggesting that our highest Tb's for nesting females
(>35C) approached thermal limits is their strong congruence with the upper
temperature limits selected by free-ranging basking emydid turtles of similar size
(e.g., Panamanian T. scripta: Moll and Legler 1971). Further support comes from
Hutchison et al. (1966), who listed mean temperatures of 41.80C and 39.30C for
the critical thermal maximum and the loss of righting response, respectively, of
five hatchling P. concinna (6-10 g) from Tennessee. However, they also showed
(for Chrysemys picta) that these temperature points not only vary geographically
(by IC or more) but also are substantially lower (by 20C or more) in larger
individuals, with the lowest thresholds in large gravid females.
Whether a black carapace is more important to heat gain or heat loss (via
radiation) in turtles remains unresolved (Lovich et al. 1990). During shell-
notching at WSSP, we noted that the shells of sun-warmed turtles bled slightly.
This suggests that perfusion of the shell with blood may provide a cooling
mechanism for turtles that are approaching the upper limit of thermal tolerance, as
might be anticipated for a large, black turtle laboring in the afternoon sun for up to
2 hours. The role of body size also remains problematic; the range of Tb's recorded
by us (20.5-35.60C) differs remarkably little from that recorded by Congdon and
Gatten (1989) in Michigan for the diurnally nesting Chrysemys picta (21.0-
37.60C), whose typical mass (350 g) is 5 percent of that of WSSP P. concinna.
Given the above, why does P. concinna not nest nocturnally at WSSP, as
some other emydids do elsewhere in the southeastern U.S.? We conclude that
nocturnal nesting, while practical for open beach/sandbar-nesting turtles to avoid
thermal stress (and diurnal predators), may disadvantage freshwater turtles that
nest in vegetatively more complex habitats. Not only would terrestrial orientation
and movement to nesting areas be hindered, but assessment of canopy shading over
potential nest sites would be difficult generally and perhaps impossible on overcast
nights.
Reproductive Parameters
Female Size, Growth, and Age at Maturity
Female (and male) river cooters achieve larger size in northern Florida than
they do elsewhere (Table 1). In our WSSP sample, 86 percent exceeded the largest
cooter measured by Fahey (1987; 361 mm CL) in central Alabama, and our
smallest mature females were nearly identical in size to the largest ones recorded
by Turner (1995) in Missouri. While allowing for greater reproductive output
(below), this may not be the sole adaptive value. Large size is also important in
JACKSON & WALKER: REPRODUCTION IN SUWANNEE COOTER
deflecting predatory attempts by abundant co-occurring crocodilians, with which
Florida's freshwater chelonians evolved.
Although misquoted by Ernst and Barbour (1972) and Ernst et al. (1994),
who claimed that C. Jackson (1970) reported a mature female of 140 mm CL (he
actually reported 190 mm CL as his smallest unquestionable female), the latter was
unable to determine the size at which females mature in his Suwannee River
population. Based on the size range at WSSP, Jackson's (1970) 10 largest females
midlinee CL range 308-366 mm), for which he calculated a mean size of 342 mm
CL, almost certainly included one or more immature turtles.
The minimal post-maturational growth exhibited by the Wakulla River
female cooters generally reflects previous findings for populations of the species in
the Florida peninsula (Marchand 1942; Jackson 1970), as well as for turtles in
general (Andrews 1982; Frazer and Ehrhart 1985). C. Jackson's data (1964: Table
7) from the Suwannee River showed that substantial growth continues until
maturity, but, based on body sizes, it appears that he recaptured no mature females
after intervals >1 year.
Because mature females at WSSP vary substantially in body size (Figs. 12,
13) and yet show extremely low annual growth rates, juvenile growth rate must be
more significant than age as a source of variance in adult body size. Although we
lack data from the Wakulla River, C. Jackson (1970) emphasized highly variable
growth rates among immature cooters in his Suwannee River population. Variable
body size at sexual maturity has been documented in other turtles (e.g., Carr and
Goodman 1970; Zug et al. 1986; Gibbons and Greene 1990; Congdon and van
Loben Sels 1991; Ernst and Zug 1994) and may reflect differential juvenile growth
rates resulting from differences in individual diets, health, and genetic constitution
(Congdon and Gibbons 1990a;. Ernst and Zug 1994). Thus, though contrary to the
widespread assumption that size and age are strongly correlated in adult reptiles
(e.g., Bury 1979), this pattern actually typifies many species (Halliday and Verrell
1988). We conclude that sexual maturity in female P. concinna is more age- than
size-dependent, as proposed for some other emydid turtles (Emydoidea blandingii:
Graham and Doyle 1977, Petokas 1986; Trachemys script females: Gibbons et al.
1981). Further, because our data show that female cooters could reproduce at
smaller sizes while retaining the same egg size, it is appropriate to consider
maturation as delayed.
Fahey's (1987) estimate that female river cooters in Alabama require 15 years
to mature (at a mean CL of 304 mm) exceeds our estimate for northern Florida by
2-4 years. This compares well with the longer maturation time of the tortoise
Gopherus polyphemus at more northerly latitudes, where growing seasons are
shorter (compare Iverson 1980 and Landers et al. 1982).
Eggs, Clutch Size, Clutch Mass, and RCM
Eggs.- Previous data on egg (and hatchling) size for this species and
subspecies (Table 2) are within the range of variation identified within the Wakulla
BULLETIN FLORIDA MUSEUM NATURAL HISTORY VOL. 41(2)
River population. However, only Turner's (1995) study in Missouri employed
sufficiently large samples (>10 clutches) to examine the magnitude of this
variation and its relationship to female size. The largest eggs from WSSP had
more than twice the mass of the smallest "normal" eggs, with clutch means
showing differences as great as 57 percent. Although substantial, this is not
exceptional for turtle populations (Ewert 1979; Iverson and Smith 1993; Iverson et
al. 1997).
Variation in egg size can be significant in the allocation of reproductive
effort. A positive correlation between egg size and maternal body size is typical of
many, though not all, turtle populations studied (Congdon and Gibbons 1985).
However, most supportive data have been drawn from relatively small-bodied
species (<1 kg). The weak relationship we found for Wakulla River P. concinna is
congruent with a relaxation of morphological constraints on reproduction in larger-
bodied turtles (Congdon and Gibbons 1987, 1990b; Jackson 1988) that lay
proportionately small eggs. Eggs of Wakulla River P. concinna, although an
average of 24-60 percent larger in absolute mass than the eggs of four other
northern Florida lentic emydids (see Jackson 1988: table 1), are 7-77 percent
smaller relative to female size as a consequence of P. concinna's greater size.
Likewise, though averaging ca 3.7 g larger than eggs of Missouri P. concinna, the
eggs of Wakulla River cooters are smaller relative to female size. These
relationships suggest that eggs of P. concinna have achieved optimal size;
however, the substantial egg size variation within the Wakulla River population
argues against strong selection for egg size optimization in this species.
Finally, mean egg length:width ratios of 1.43 in our Wakulla River
population and 1.48 in Turner's (1995) Missouri population support the conjecture
of Congdon and Gibbons (1990b: 116), based on a single clutch, that P. concinna
may have slightly less elongate eggs than most turtles with oblong eggs, which
typically average 1.6-1.7. Within aquatic emydids, this may reflect a chelonian
tendency for larger turtles to lay more spherical eggs (Elgar and Heaphy 1989).
Although we know of no previous reports for the genus Pseudemys,
anomalous eggs similar to those observed in this study are known among other
turtles. Small "yolkless" eggs are especially characteristic of the giant marine
leatherback turtle, Dermochelys (Chua and Furtado 1988), as well as some other
marine turtles (Dodd 1988; Iverson and Ewert 1991), but also have been noted
among other emydids (e.g., Trachemys: Cagle 1950). Ewert (1985) observed the
occurrence of spongy, premature eggs in some nests of the Ouachita map turtle
(Graptemys ouachitensis). Eggs with no or very thin shells have been reported as
rarities for diamondback terrapins (Malaclemys terrapin), three genera of sea
turtles (Somers and Beasley 1995), and even dinosaurs (Erben et al. 1979). Causes
of such anomalies have not been identified definitively for turtle eggs.
Clutch Size.- Our data for the Wakulla River cooter population indicate
clutch sizes that equal or exceed those of nearly all non-Florida populations of P.
concinna (Table 1). Based on only 10 clutches, Turner (1995) reported a slightly
larger mean clutch size in Missouri, where females lay considerably smaller eggs.
JACKSON & WALKER: REPRODUCTION IN SUWANNEE COOTER
Though Fahey (1987) reported a mean clutch size of 17.6 for a modest sample
from east-central Alabama, his corpora luteal counts suggest that the true
population mean may be somewhat lower. We attribute large clutches of many of
Florida's freshwater emydids chiefly to greater female body size. A significant,
positive relationship between clutch size and body size characterizes the majority
of turtles (Carr 1952; Moll 1979; Ehrhart 1982; Gibbons et al. 1982; Congdon and
Gibbons 1985, 1990a; Elgar and Heaphy 1988). As has been suggested elsewhere
(e.g., Jackson 1988; Gibbons and Greene 1990), female size may serve as a
morphological constraint on maximum clutch size, although the low percentage of
variation (10-20 percent) accounted for by maternal size in this study underscores
the influence of other unidentified factors.
Clutch Mass and RCM.- The higher correlation of clutch mass than either
clutch size or egg size with female body size suggests that, while energetic and
female body size strongly influence the amount of reproductive matter comprising
a clutch, the manner in which the clutch is partitioned into discrete propagules
(i.e., eggs) is less tightly constrained. This relaxation is shown by the significant
tradeoff between egg size and number following control for body size.
The substantial temporal variability in turtle body masses as a result of
changes in amounts of gut contents, bladder water, and reproductive matter can
render relative clutch mass associations less liable to rigorous testing than in
squamate reptiles (R. Seigel pers. comm.). Nonetheless, the mean RCM of 0.048
for the Wakulla River cooters approximates that of the other large, congeneric
pond-dwelling emydids in northern Florida (Jackson 1988, after conversion of
gravid body mass to spent body mass). Thus, there appears to be no special
adaptation to riverine existence in this reproductive trait nor in annual RCM (=
RCM x clutch frequency: Iverson 1992) since clutch frequencies of P. concinna
and the lentic species are comparable.
Fahey's (1987) data for east-central Alabama river cooters show remarkable
similarity in clutch mass (mean=307 g, n=10) to the Wakulla River cooter
population (mean=305 g for 24 clutches; estimated mean for all clutches=285 g).
Estimates of clutch mass in Missouri, based on Turner's (1995) data, are somewhat
lower (249 g) as a consequence of smaller eggs. However, the slightly smaller
body sizes of both Alabama and Missouri turtles caused RCM to be greater in those
populations (0.09 in Missouri). This may compensate for the reduced number of
clutches that most females produce north of Florida.
For a population of Michigan Chrysemys picta in which females nest
diurnally, move roughly the same distance overland as WSSP cooters to nest, and
require a comparable length of time to complete nesting, Congdon and Gatten
(1989) approximated that the energy invested in nesting activities (moving to and
from nesting site, digging, and filling) averaged <1 percent of that invested in a
single clutch. This, plus their virtual lack of growth once mature; suggests that
female Suwannee cooters invest most of their excess resources directly into the
production of reproductive matter.
BULLETIN FLORIDA MUSEUM NATURAL HISTORY VOL 41(2)
Clutch Frequency and Internesting Interval
At roughly 2-3 weeks, the interesting interval of P. concinna at WSSP
approximates that of other aquatic emydids and most turtles in general, although
some marine and estuarine turtles may have shorter mean periods (Moll 1979:317;
Roosenburg 1991). However, as we indicated previously, spacing of rainfall events
during the nesting season influences the temporal spacing of clutches, as for other
animals with multiple clutches (Telford and Dyson 1990). Nevertheless, it appears
that P. concinna will delay oviposition by no more than a few days to perhaps a
week. We could not determine in this study whether repeated delays, such as
might occur during a severe drought, would reduce a female's total annual
reproductive output. It is known that entire sets of pre-ovulatory follicles may
undergo atrophy in other Floridian Pseudemys (e.g., P. nelsoni: DRJ pers. obs.)
under such circumstances. The occupation of more stable riverine habitats may
release P. concinna from this extrinsic constraint.
The production of multiple clutches separated in time reduces the chances
that a female's entire annual reproductive output will be destroyed (Wilbur 1975).
This pattern also allows production of eggs in excess of any morphological
constraints (Moll 1979; Jackson 1988). An increase in the number of clutches has
been predicted for organisms in which egg mortality is high relative to adult
mortality (Stearns 1976). However, at WSSP, predation of adult females on land to
nest could act as a counter-selective force to reduce the number of emergences.
Still, this selective force seems weak because, as explained in our results, females
have a prolonged nesting season and seem to be producing as many clutches as
they can coordinate with rainfall during this period.
Little information is available (Table 1) addressing the number of clutches
produced by this species elsewhere. Turner's (1995) data suggest annual
production of two or three clutches by most females in southern Missouri, with
some laying only one and a few perhaps laying four. Although Fahey (1987) stated
that most adult females in east-central Alabama produce two clutches each year,
his data (Fahey 1987: table 4) suggest that 5 of the 12 females he examined likely
produced but a single clutch, whereas at least one (among the largest in his
sample) probably would have produced a third.
Annual Periodicity.- As with sea turtles, individual female freshwater
turtles may not produce eggs every year (Congdon and Gibbons 1989; Frazer et al.
1989). In Florida, however, the long growing season should minimize skipping
years because it allows females ample time to acquire the energy resources needed
to reproduce annually. If any Florida turtles were to skip reproduction in some
years, surely it would be those that normally produce but one clutch per nesting
season (e.g., Gopherus: Butler and Hull 1996; Macroclemys: Dobie 1971, Ewert
and Jackson 1994) and not a multi-clutched species such as P. concinna. Even in
climates more temperate than Florida's there are populations of aquatic emydids in
which all females produce multiple clutches annually (e.g., Chrysemys picta:
Iverson and Smith 1993).
JACKSON & WALKER: REPRODUCTION IN SUWANNEE COOTER
Reproductive Potential
The longer nesting season, coupled with a generally longer activity season
and perhaps more abundant resources, allows P. concinna in Florida to increase its
reproductive output as much as two- to three-fold over more northerly conspecific
populations (only 340 km to the north-northwest). Although females in northern
Florida and east-central Alabama produce comparably sized clutches (Table 1),
Florida turtles may lay as many as four to six clutches per season, compared to only
one or two for populations in Alabama (Fahey 1987). River cooters studied by
Turner (1995) in southern Missouri appear to have higher reproductive potentials
than those studied by Fahey (1987) in Alabama but still substantially lower than
those in Florida. A similar geographic trend characterizes other southeastern
emydids as well (Jackson 1988) and appears to be unrelated to the occupancy of
spring-run habitats. That individual females in the WSSP population may utilize
much of the long nesting season contrasts with Vogt's (1990) prediction for
Mexican (Veracruz) Trachemys scripta, which likewise has an extended nesting
season as a population but for which he was unable to find evidence of extended
individual laying seasons.
Development, Hatchlings, and Emergence from Nests
Development and Hatchlings.- Fitness presumably increases with
hatchling size in turtles for a variety of ecological, physiological, and behavioral
reasons (Wilbur and Morin 1988; Janzen 1993). Potential sources of variation in
hatchling size include maternal/genetic effects (e.g., egg size and provisioning),
substrate water potential, and perhaps temperature. For species with flexible-
shelled eggs, higher substrate water potentials can increase hatchling size while
decreasing duration of incubation (e.g., Gutzke et al. 1987; Packard et al. 1991;
Cagle et al. 1993; Miller 1993; Janzen et al. 1995). Though we did not measure
substrate water potentials, all of our laboratory substrates would have been
considered as wet or saturated by current standards (Miller 1993 and references
therein), thereby eliminating this factor. Hatchling size therefore was principally
determined by egg size, as expected (Ewert 1985; Rowe 1995). Although we found
no relationship between hatchling and maternal sizes at WSSP, females can
maximize hatchling size by selecting nest sites most likely to retain ample but not
excess moisture, and by assuring that the substrate is damp from the start. Cooters
at WSSP achieve the latter effect by nesting in association with rain and by voiding
bladder fluid during nesting.
While comparing incubation periods based on data from different laboratories
may introduce error, the geographic differences suggested in Table 10 may be
ecologically important. At equivalent temperatures, eggs from Tennessee hatched
substantially earlier than those from Florida (mean differences of 23 and 11 days at
250C and 300C, respectively). Florida eggs required temperatures 2-40C higher
BULLETIN FLORIDA MUSEUM NATURAL HISTORY VOL 41(2)
than Tennessee eggs to attain similar developmental rates. This mirrors the
inverse relationship between incubation period and latitude noted by Ewert (1979,
1985) for turtles in general. The seemingly more rapid developmental rate of
Tennessee turtles may be an adaptation to the shorter period of time available
before embryogenesis is precluded by the onset of cooler soil temperatures in late
summer or early fall.
Hatchling Emergence.- Fahey (1987) suspected that most hatchling river
cooters at his central Alabama study site overwinter in the nest. His only record of
fall emergence was from a disturbed nest, and hatchlings were abundant in the
river in spring but virtually absent in the fall. Buhlmann and Vaughan (1991)
likewise concluded that hatchling river cooters overwinter in the nest in West
Virginia, the species' northern range limit. In contrast, air and water temperatures
in northern Florida remain relatively warm in the fall, and some river cooter
hatchlings do emerge. Although fall emergence post-dates the annual rainy season
(May-September) in Florida, it still occurs during the latter part of the hurricane
season (June-November). This may reduce the threat of desiccation during the
overland neonatal journey to the river.
Specific factors determining whether hatchlings overwinter in the nest or
emerge in the fall are obscure. DePari (1996) noted a possible effect of soil type
but, as in our study, found little relationship to date of oviposition. Soil and air
temperatures, precipitation, and individual nest structural integrity may combine to
determine whether or not hatchlings overwinter (DePari 1996). Recent studies of
marine turtles suggest soil temperatures are key, but the rate of temperature change
is more important than absolute temperature (Witherington et al. 1990; Hays et al.
1992). Rapid cooling of nests by precipitation thus may trigger emergence of both
hatchling cooters and marine turtles. At WSSP, the declining angle of the autumn
sun causes day-long shading of most unemerged nests, even those that had been
exposed to several daily hours of sunlight in summer. This may dampen soil
temperature fluctuations and thereby favor overwintering of any neonates that have
not emerged by mid-October.
Although no data are available for P. concinna, Butler and Graham (1995)
found that hatchling Emydoidea required 0.5-9 days (generally 2-4) to negotiate
nest site-to-wetland distances comparable to those at WSSP. Our behavioral
observations indicate that post-emergent hatchling cooters are capable of
journeying to and entering the water immediately or, alternatively, of burrowing
into the substrate and completing the migration later if conditions (e.g., distance
and weather) dictate. Larger hatchlings should be physically and physiologically
better able to negotiate adverse terrain and weather during this critical stage of life
history (Janzen 1993). Hatchling P. concinna at WSSP offer an excellent
opportunity to test this prediction.
JACKSON & WALKER: REPRODUCTION IN SUWANNEE COOTER
Nest Site Selection, Site Fidelity, and Sex Determination
While females of some freshwater turtles may travel long distances both in
water and on land to reach nesting sites (Congdon et al. 1983; Gibbons 1986),
previous observations of nesting P. concinna do not support extensive overland
movements. Gibbons (1990) reported nesting along high sandbars of the Savannah
River, South Carolina/Georgia, and Cahn (1937) noted a preference for sandy soils
within 50 m of water in Tennessee. Fahey (1987) observed two types of nest sites
along the Tallapoosa River, Alabama. Some turtles nested close to the river in
loose, sandy soils, but others nested up to 250 m from the river in a variety of
mostly non-canopied habitats with a diversity of soil types (sand, loam, clay, and
even some large gravel).
Florida populations of P. concinna also may use a dual strategy. On the
Rainbow River, nesting typically occurs along exposed (often disturbed) banks
immediately adjacent to the river, where a steep embankment rises I m above the
stable water level to assure adequate soil drainage (DRJ pers. obs.). In contrast,
turtles at WSSP routinely walk from 30 to nearly 300 m from the river to nest.
Two factors seem paramount in the species' choice of nest site-absence of
forest canopy, allowing for an open, well isolated substrate, and reasonably well
drained soils. Their availability determines how far from the river a female will
walk. Selection for these factors presumably allows high nest temperatures
concurrently with protection from flooding. Although Fahey (1987) did note one
attempted nesting in dense, shaded forest, cooters at WSSP showed no inclination
to nest in such on their way to more open sites. Soil type itself does not seem to be
a critical factor in nest site selection, as long as elevation is adequate to assure
drainage. Substrates free of vegetation are attractive but not essential, as WSSP
females nested in grass within a few meters of barren sand as well as in the sand
itself.
The concentration of nests along River Road (WSSP) indicates that other
suitable nesting sites along the upper Wakulla River are scarce. Nesting is not
communal, however, as it is dispersed along 4.5 km of road, and there is no
evidence of long-distance aquatic migrations to specific sites, as in some other
freshwater turtles, perhaps with fewer opportunities (e.g., Moll and Legler 1971;
Plummer and Shirer 1975; Moll 1980; Obbard and Brooks 1980; Pritchard 1984;
Pluto and Bellis 1988). Instead, WSSP females appear to emerge from their
riverine home ranges and walk almost directly to the first suitable nesting habitat,
which in the case of River Road is nowhere >300 m from the river.
Strong selection to avoid nesting in heavily wooded or densely vegetated sites
seems to exist throughout the family Emydidae, with diurnal nesting permitting
females to assess the nature of surrounding vegetation. Coupled with TSD, this
may allow females to influence the sex ratios of their offspring (Janzen 1994). The
preponderance of unisexual nests (four of six) within our small WSSP sample is
typical of that seen in many turtle populations with TSD (Vogt and Bull 1984;
Ewert and Nelson 1991; Janzen 1994).
BULLETIN FLORIDA MUSEUM NATURAL HISTORY VOL. 41(2)
Although eggs of a few freshwater turtles tolerate extended inundation
(Kennett et al. 1993; Polisar 1996), this does not appear to be the case for other
turtles (Plummer 1976; Ewert 1979; McGehee 1990; Ewert and Jackson 1994),
including emydids with pliable-shelled eggs (Kam 1994; Tucker et al. 1997; DRJ
pers. obs.). Nest site selection must therefore be in part a compromise between the
risk of a nest flooding and of the increased risk of hatchling or female deaths (via
predation or desiccation) imposed by sites more distant from the water.
Nest Site Fidelity.- The tendency of female turtles in some populations to
nest repeatedly in the same areas has been described at two levels: nest area
philopatry-the selection of a general nesting area based on its ecological
characteristics, and nest site fixity-nesting near sites of previous nestings (Carr
1975; Lindeman 1992). Both phenomena seem to be operant in the WSSP cooter
population. Nearly all females exhibit nest area philopatry in their selection of
River Road or one of a few alternate nesting sites. The relatively high degree of
nest site fidelity within these areas requires further explanation, however. The
tendency of females to nest on land adjacent to limited aquatic home ranges likely
accounts for much of the fidelity and is supported by our radiotelemetric data.
However, repeated nesting by some WSSP females within a few meters of previous
nest sites suggests true site fixity. Landmark recognition and re-use of nesting
routes is almost certainly employed by some individuals that leave the river at
specific points (e.g., the railroad). Such behavior may entail less risk than random
wandering (Gibbons et al. 1990). Elucidating whether pre-nesting turtles home to
such landmarks from beyond their normal home range may require further
radiotelemetry.
Mortality
Nest Mortality and Nest Structure.- We attribute the nearly complete
destruction of unprotected nests at WSSP principally to human modification of the
natural environment. (1) Potential nesting sites are scarce, apparently a result of
decades of fire exclusion that have resulted in hardwood encroachment and canopy
closure in the upland forests bordering the narrow floodplain. (2) As a result,
nesting has become heavily concentrated along a linear, easily searched artificial
microhabitat (River Road) that coincides aloqg much of its length with a natural
ecological edge. Nearly total predation of nests has been documented for other
species of turtles nesting along roadsides (Gemmell 1970; Landers 1988 in litt.;
Linck et al. 1989). Furthermore, several studies of freshwater turtles (Legler 1954;
Christens and Bider 1987; Fahey 1987; Temple 1987) have indicated lower levels
of predation on nests dispersed in uplands away from water and other ecological
edges. We therefore predict that habitat management allowing greater spatial
dispersion of nests in more open habitat inland of River Road (see Conservation
and Management) will be accompanied by reduced egg predation. (3)
Exacerbating the problem of nest predation at WSSP is the population explosion of
nest predators, chiefly small- to mid-sized generalized omnivores such as raccoons
150
JACKSON & WALKER: REPRODUCTION IN SUWANNEE COOTER
and crows. This is a direct result of human disturbance and fragmentation of
habitat and extermination of large native predators throughout the southeastern
coastal plain (Johnson 1972; Harris and Silva-Lopez 1990; Garrott et al. 1993).
The abilities of predators to discover and learn to exploit such temporarily
abundant food resources as turtle eggs are well known (e.g., Johnson 1972; Burger
1977; Snow 1982), and their nearly total effectiveness in doing so has been
documented elsewhere (e.g., Allen 1938; Cagle 1950; Shealy 1976). Predation
within 48 hours of nesting, while physical and olfactory signs are still fresh, is
typical (Legler 1954; Hammer 1969; Moll and Legler 1971; Burger 1977; Petokas
and Alexander 1980).
Raccoons have been identified as major nest predators in most studies of
North American turtles, while fish crows are known to be important nest predators
of other southeastern emydids (Shealy 1976; Lahanas 1982; U.S. Fish and Wildlife
Service 1989; Brauman and Seigel 1995). Our suspicion that crow predation
declined midway through the cooter nesting season echoes Lahanas's (1982) study
of Graptemys nigrinoda and suggests that crows may be capitalizing on turtle eggs
in conjunction with their own breeding activities (Terres 1990; Kale et al. 1992).
The destruction of a few nests by relatively minor egg predators is not
unusual. Moles, shrews, and rodents have been implicated as significant nest
predators of other species of emydids (Doroff and Keith 1990; Naklicki et al.
1995). The importance of fire ants and armadillos, both exotic to the southeastern
United States, remains uncertain. Moll and Legler (1971) reported predation on
Panamanian slider (Trachemys) nests by armadillos, a species we could not
confirm as a nest predator at WSSP despite its regular use of River Road for
foraging. They also recorded invasion of undisturbed nests by fire ants that left
behind egg shells and remains of nearly full-term embryos, much as we found at
WSSP. Likewise, Fahey (1987) believed that fire ants-had destroyed at least two
nests of hatchling P. concinna in his Alabama study area, and Dobie and Bagley
(in U.S. Fish and Wildlife Service 1989) considered them a potential threat to nests
of P. alabamensis. However, Cagle (1937) noted that ants are especially fond of
dried turtle eggs and suspected that they may not attack fresh healthy eggs; he
postulated that the removal of dead egg components by ants may, in fact, benefit
remaining live eggs in a nest. However, the invasive exotic fire ant is aggressive
and may be a threat to migrating hatchlings even if it does not attack live eggs.
Among turtle populations, extremely high levels of nest predation, such as at
WSSP, can lead to a disproportionately large number of old individuals (Thompson
1983). As these old individuals die and recruitment of juveniles into the
population declines further, the general population can be expected to decline.
This may be mitigated somewhat by increased egg survival should predators
exhibit a density-dependent response to fewer nests (Wilbur 1975).
P. concinna may have adopted several strategies to counter nest predation.
These include synchronous nesting, nesting during rain, dispersion of nests, and
the three-holed nest structure. As noted above, empirical evidence suggests that
synchronous nesting may reduce predation of nests (and potentially of females)
BULLETIN FLORIDA MUSEUM NATURAL HISTORY VOL. 41(2)
within turtle populations (Robinson and Bider 1988). Most of the WSSP nests
monitored for predation were mid- to late-season "solitary" nests, i.e., not made
during the highest, rain-induced peaks of nest construction. A larger sample of the
latter should be monitored in the future to verify whether they suffer a lower
incidence of predation as predicted. Likewise, the relatively few nests constructed
early in the season before predators have shifted their seasonal diet to eggs may
have greater chances of escaping detection (Talbert et al. 1980).
Although only recently reported for P. concinna, the unusual three-holed nest
structure has long been noted for peninsular Florida populations of its presumed
sister species, P. floridana (Allen 1938; Marchand 1942; Carr 1952; Franz 1986).
Though we have considered alternatives (e.g., sex determination and flooding), we
agree with Canrr (1952) that the likeliest adaptive value of the three-holed nest is as
an anti-predator strategy, even though he questioned its possibility of success. Carr
hypothesized that the accessory holes may serve a decoy function to deflect
predators away from the central egg chamber, but he also noted that the shallow
eggs seemed to serve "more like a beacon" instead. In situations like WSSP, where
predators exploit cooter eggs seasonally, individual predators may learn the
strategy and rarely be fooled by it. Hence, it is not surprising that the evolution of
the three-holed nest as an anti-predator device has been questioned based on its
frequent failure (Franz 1986; Cople and Pilgrim 1993). Nonetheless, we did
observe one apparent success. We therefore counter that for species normally
experiencing very high levels of nest predation, even a very low rate of success-
such as once in the lifetime of a female-may increase fitness sufficiently to retain
a strategy that requires little extra time or energy.
The geographic distribution of the three-holed nest across the ranges of both
P. concinna and P. floridana remains poorly known. Such data may prove to be an
important phylogenetic character in subsequent assessments of this lineage. The
principle of parsimony suggests that such an unusual and clearly derived feature is
likely to have evolved only once, in the common ancestor of these species and their
nearest kin (P. gorzugi and P. texana), and that its absence from any descendant
population represents secondary loss. Thus far, three-holed nests are known for P.
concinna (as recognized by Ward 1984) from the Wakulla, Suwannee, and
Withlacoochee river systems in Florida; Reelfoot Lake in Tennessee (one
observation, M. Ewert pers. comm.); and the mid-continental states of Kansas
(Caldwell and Collins 1981) and Arkansas (S. Christman, in litt.). For P.
floridana, three-holed nests have been noted throughout peninsular Florida (Allen
1938; Marchand 1942; Carr 1952; Franz 1986; Cople and Pilgrim 1993; DRJ pers.
obs.), as well as in South Carolina (K. Buhlmann pers. comm.). Neither Fahey
(1987: P. concinna) nor Thomas (1972: P. floridana) observed the actual nesting
process, so their failure to report accessory holes does not confirm their absence
from the populations they studied in Alabama. Data for these and other
populations of both species are sorely needed.
The relatively low frequency of deposition of satellite eggs by P. concinna at
WSSP contrasts markedly with reports by others (Carr 1952; Franz 1986; Cople
JACKSON & WALKER: REPRODUCTION IN SUWANNEE COOTER
and Pilgrim 1993) that the accessory holes of P. floridana usually contain one or
more eggs. Data from other river cooter populations are vital to determining
whether differential use of the side-holes exists among local populations or
between species, in which case our observations may presage the strategy's
eventual loss by P. concinna.
We can not fully explain the high failure rate (69.5%) of non-depredated eggs
in protected nests at WSSP. High rainfall (such as during parts of 1988 and 1989
at WSSP) during the incubation season, with consequently low soil temperatures
and poor soil aeration, has led to nearly total hatching failure in other reptiles
while not affecting timing of nesting itself (Bock and Rand 1989). The higher rate
of hatching of WSSP eggs in the laboratory leads us to suspect that this may be
important at WSSP, where turtles nest in more compact soils than normally used
under more natural conditions. Still, other potential problems resulting in egg
infertility or inviability can not be dismissed.
Adult Mortality.- Even though it may entail as little as two hours of
terrestrial exposure, which is rapid compared to some turtles (Congdon and Gatten
1989), emerging to nest may be the greatest single risk faced by adult cooters,
which otherwise are disinclined to wander terrestrially (Carr 1938). Predation on
nesting emydid turtles, most commonly by mammals, seems to be a widespread
though not frequently witnessed phenomenon (e.g., Cagle 1950; Carr 1952;
Crenshaw 1955; Wilbur 1975; Shealy 1976; Metcalf and Metcalf 1979; Seigel
1980; Congdon and Gatten 1989; Roosenburg 1994). Seigel's (1980) description
of the mutilation of a terrapin by a raccoon observed preying upon it, and similar
kills of other emydids reported by Cagle (1950) and Shealy (1976), closely match
the condition of the freshly killed and gutted cooters that we found on River Road.
Both at WSSP and in Seigel's study area (Atlantic coastal Florida), human
environmental perturbations apparently have increased raccoon numbers, and thus
contact between raccoons and nesting turtles.
Data from this study allow an estimate of the probable cost of nesting in terms
of female mortality at WSSP. An annual average of 3.5 deaths (14 in 4 years) for
an estimated 1525 nests (5 nests each for 305 females) yields a per nest death rate
of 0.0024, or an annual risk of death of ca 1 percent per female. Of the roughly
26,700 eggs (305 females x five clutches x 17.5 eggs) produced annually by this
population, four (0.015 percent) must yield female hatchlings that survive to
maturity to maintain population stability (assuming a closed population with no
other adult female mortality).
Although alligators do prey upon cooters at WSSP, they are probably not a
major source of mortality for adult females. The abundance of old scars and fresh
scratches on the shells of most nesting females attests to their ability to escape
alligator attacks. Probably only the few very large (>3.5 m) alligators in the river
are capable of gaining sufficieAt leverage to hold and crack the large domed shell
of a mature female cooter.
BULLETIN FLORIDA MUSEUM NATURAL HISTORY VOL 41(2)
Population Size, Biomass, and Production
The accuracy of population estimates based on mark-recapture methods
increases with the proportion of the population caught. We believe we marked ca
80 percent of the adult females in the WSSP population, and our estimates of ca
300 individuals weighing nearly 2 metric tons are close to actual values. Our
qualitative observations lead us to suspect that the total biomass of adult males and
all immature cooters in WSSP may approach but is unlikely to exceed the biomass
of adult females. Casual observations of basking and swimming cooters at WSSP
suggest a slightly female-biased sex ratio among adults, with juveniles being
common but not abundant. In the only large, relatively random sample of river
cooters collected in Florida, Marchand (1942) recorded approximate parity among
the sexes and stated that large, mature turtles dominated the population.
Population densities of river cooters vary markedly among localities and
habitat types. In seemingly optimal habitats, such as the Wakulla River and the
Rainbow River in northern peninsular Florida, very high densities may typify non-
exploited populations. Marchand's (1942) crude estimate of as many as 5000
Suwannee cooters (half of all turtles) in a 6.7-km segment of the Rainbow River
exceeds our estimate for WSSP, although his sample included male and immature
turtles; the population has since declined, probably as a result of exploitation
(Giovanetto 1992; Meylan et al. 1992). C. Jackson (1970) collected 237
individuals in a small spring run (160 m x 20 m) connected directly to a large river
(Suwannee) in northern Florida, but body size distribution indicates that his
sample contained predominantly immature turtles. It is probable that heavy
predation on eggs (by raccoons and fish crows) as well as juvenile turtles (by an
exceptionally dense alligator population) may be holding the WSSP cooter
population below carrying capacity. Elsewhere, however, the densities of studied
populations have been much lower. Buhlmann and Vaughan (1991) computed
densities of only 0.7-2.3 adults (sexes combined) per ha in the New River, West
Virginia, near the northern edge of the species' range. Lindeman (1997) estimated
1.58 river cooters per 100 m of shoreline in an impoundment in western Kentucky
and roughly similar densities in the Pearl and Pascagoula rivers in southern
Mississippi and Louisiana. For comparison, our WSSP estimate just for adult
females is twice that (300 females per 10 km of riverbank). Fahey's recapture rate
was too low to allow him to estimate population density at his Alabama study site.
With a biomass potentially approaching 100 kg/ha (including immature and
male turtles), the WSSP Suwannee cooter population substantially surpasses that of
most freshwater turtles (Iverson 1982; Congdon et al. 1986; Dodd et al. 1988;
Mitchell 1988), although few data are available for riverine species. Nonetheless,
exceptionally dense local populations of emydids, kinosternids, and chelydrids,
usually in pond situations, are known to exceed this by factors of two to eight or
more (Iverson 1982; Congdon et al. 1986; Parker 1990). However, the WSSP
cooter biomass is one of the largest reported for a freshwater herbivore, though
JACKSON & WALKER: REPRODUCTION IN SUWANNEE COOTER
elsewhere the species may at least formerly have achieved even higher biomasses
(Iverson 1982 based on data of Marchand 1942).
Studies of unexploited Suwannee cooter populations highlight the capacities
of rivers to support, and perhaps be dominated trophically by, very high reptilian
numbers and biomass. Additional studies of riverine turtle populations underscore
this. Based on a short-term basking census, Kramer (1995) estimated ca 300 adult
(both sexes) and large subadult Pseudemys (three-fourths P. nelsoni, one-fourth P.
floridana) in a 1-km section of spring-run stream in central Florida, although
home ranges of many of these may have extended beyond the study site.
Giovanetto (1992) estimated combined densities of 74 and 90 individuals/ha for
three species of Pseudemys co-existing in two northern Florida rivers (Homosassa
and Rainbow, respectively). Several studies of map turtles (Graptemys spp.), the
most lotically adapted emydid genus, have reported local densities of 100-400
individuals per km (Tinkle 1958; Shealy 1976; Jones and Hartfield 1995;
Lindeman 1997). Temperate non-emydid turtles likewise may achieve high
densities and biomasses in rivers (Plummer 1977; Pritchard 1989). Diverse
tropical riverine turtles may occur in huge numbers (Batagur: Moll 1980;
Podocnemis: Vanzolini 1967; Trachemys: Moll and Legler 1971), although
movement patterns have made accurate density estimates difficult. Finally, it
should be noted that even these high biomasses of turtles may be exceeded by other
reptiles, specifically crocodilians, that share their rivers; such is probably the case
with alligators at WSSP.
Two reviews have attempted to identify ecological and environmental
correlates of chelonian biomass. The high biomass of the WSSP cooter population
supports Iverson's (1982) predictions that high biomasses can be expected for
species that are herbivorous and aquatic and which inhabit springs, ponds, or
islands. However, the high density and biomass alternatively may support
Congdon et al's (1986) hypothesis that these characteristics are more closely
related to habitat suitability, body size, and population age structure than to trophic
position. An analysis of densities and biomasses of P. concinna populations in a
variety of rivers may help to determine which of these factors are most important.
From an energetic standpoint, the ability of turtles to maintain large populations
and standing crop biomasses presumably relates to their low metabolic rates
(Bennett and Dawson 1976).
Our data leave little doubt as to the major ecological role of the Suwannee
cooter in the riverine and adjacent upland ecosystem at WSSP. The species is
uncontested as the principal vertebrate grazer in the riverine community. As a
population, the turtles must consume annually many metric tons of aquatic
vegetation. Our estimate of the reproductive portion of biomass production (10.6
kg/ha/yr) exceeds most previously reported data for freshwater turtles (Congdon
and Gibbons 1989) by one to two orders of magnitude. In light of the high rate of
nest predation by terrestrial predators, cooter eggs form an important conduit for
energy flow from aquatic to terrestrial communities (Congdon and Gibbons 1989).
BULLETIN FLORIDA MUSEUM NATURAL HISTORY VOL. 41(2)
Management Recommendations
Data generated by this study prompt us to recommend measures for
management of P. concinna at two levels: statewide populations and the localized
WSSP population.
Statewide Harvest
Because of its large size, accessible habitat (by boat), formerly dense
populations, and relatively easy catchability, P. concinna suwanniensis historically
has been exploited heavily by man, at least in some rivers. The largest turtles,
generally reproductive females, have been taken preferentially. Although most
reports are anecdotal (e.g., Auffenberg 1978; Canrr 1983; but see Meylan et al.
1992), they concur that populations in several rivers have declined substantially in
recent decades. Pollution, dredging and channel maintenance, impoundment, and
other forms of river alteration pose additional threats. Because of these factors, the
Florida Game and Fresh Water Fish Commission (FGFWFC) in 1975 designated
P. c. suwanniensis as a fully protected Threatened Species. Subsequent regulations
relaxed this level of protection: in 1978, to a personal possession limit of two
turtles (for personal consumption but not commercialization), and in 1979 to a
Species of Special Concern (SSC), following the initiation of that category. Efforts
to educate the boating and fishing public about these regulations, however, have
been limited.
Our study suggests that the Suwannee cooter, like most turtles, requires high
adult annual survivorship to maintain stable populations. Low recruitment to
adulthood, coupled with continued removal of mature females, can lead to
population decline. The extreme ease with which nesting females can be captured
once their nesting sites and emergence cues are discovered (we estimate that we
could have captured >95 percent of adult females in our study area in two seasons
had that been our primary goal) led us in 1988 to propose a closure on taking
during the principal nesting period (15 April-31 July). This subsequently was
approved by the FGFWFC. We now suggest extending these dates from 15 March
to 15 August to allow for annual and geographic variation within Florida. Current
state regulations prohibit the use of snares and firearms to take turtles; we support
extending the ban to nets and traps (basking and hoop) except by scientific permit.
Although a harvest of two adult females per person per day (the legal potential) is
ecologically unsound, the detriment that this may cause to a population depends
upon the number actually taken rather than that potentially allowed. Data
measuring actual harvest of river cooters, both legal and illegal, are needed on a
river by river basis. This information, coupled with censusing and monitoring to
determine population sizes and trends, should be employed to evaluate the
appropriateness of the current possession limit and whether complete protection, as
urged by Carr (1983), is necessary.
JACKSON & WALKER: REPRODUCTION IN SUWANNEE COOTER
In the absence of substantial taxonomic or ecological differences (Jackson
1995) pertinent to human harvest among river cooter populations in different rivers
or regions of the state, we also recommended in 1988 that all regulations be
extended to the species statewide rather than solely to the subspecies suwanniensis.
The FGFWFC enacted this revision in 1989 but withheld the SSC designation from
western (panhandle) populations.
Despite regulations, we still find evidence of heavy (presumably illegal)
localized exploitation of Florida river cooters (DRJ pers. obs. 1992, 1994, Franklin
County). We therefore urge more aggressive efforts by state regulatory agencies to
enforce regulations and to educate the public about the protected status of this
turtle in Florida.
Habitat and Species Management
We address here three management factors that we believe are critical to the
continued viability of the cooter population at WSSP and elsewhere. These include
the need to reduce nest predation, restore upland nesting habitat, and assure that
human recreation is compatible with the turtle's ecological requirements.
Predator Control.- Perhaps of most immediate concern at WSSP is the
intense, presumably unnaturally high, predation exerted on turtle eggs by raccoons
and fish crows. Both species are anthrophilic (Harris and Silva-Lopez 1990;
Garrott et al. 1993); i.e., unusually large populations are supported in regions of
human disturbance via increased food supplies and shelter, and decreased
predation. At WSSP, continued and increased efforts to minimize the availability
of human refuse to such opportunistic omnivores may be important in limiting
their local populations and thereby reducing predation of turtle eggs. In addition,
at least until upland habitat restoration (below) succeeds, we recommend periodic
removal of these two predatory species, perhaps at 2- to 3-year intervals. Such
programs elsewhere (e.g., Christiansen and Gallaway 1984) have boosted
recruitment in freshwater turtle populations where nests were subject to extensive
predation. Alternatively, 20 or more nests, at least half in full sun, should be
protected annually in situ by caging, with resulting hatchlings transported
manually to the river's edges. More generally, we encourage funding for research
to develop effective yet inexpensive methods of raccoon control.
Habitat Management- The conservational ramifications of temperature-
dependent sex determination, particularly as relates to human alteration of nesting
habitats, have become a consideration in the management of rare and endangered
turtles (e.g., Spotila and Standora 1986; Wibbels et al. 1991; Ewert and Jackson
1994). Increased shading by a developing forest canopy along the main nesting
road at WSSP threatens to skew the hatchling sex ratio toward males. Limited
evidence that female-biased hatchling sex ratios predominate in wild populations
of reptiles with TSD (Ewert and Nelson 1991; Ewert and Jackson 1994), coupled
with our subjective impression that the majority of basking adult cooters at WSSP
are females (potentially reflecting hatchling sex ratios 15-30 years earlier),
BULLETIN FLORIDA MUSEUM NATURAL HISTORY VOL 41(2)
suggests that a male-biased hatchling sex ratio at WSSP may not be sustaining.
Recently, park management has initiated a program of prescribed fire to restore a
more open longleaf pine forest above River Road, where this forest type is believed
to have provided former nesting habitat for cooters. To date, hardwood
encroachment and loss of native groundcover from decades of fire suppression
have hindered the program's success. Nonetheless, the effort is vital to achieving
sexual parity or female bias among hatchling cooters (by allowing more nests in
sunny sites) and therefore should continue, even if it requires girdling or injecting
systemic herbicides into hardwoods prior to burning.
Besides shifting hatchling sex ratios, re-opening the forest canopy may reduce
nest predation. Although used by a majority of WSSP's nesting cooters, River
Road's linearity facilitates predation of eggs. In contrast, Fahey (1987) believed
that "the wide dispersion and cryptic location of P. concinna nests" at his Alabama
study site discouraged nest predation by the locally abundant American crow
(Corvus brachyrhynchos). Eventual closure of River Road may therefore be
appropriate, but not before sufficient open-canopied uplands have been restored. In
the meantime, should park management allow the forest canopy to overgrow River
Road, and thereby render it unsuitable for nesting, female cooters may be forced to
engage in long overland nesting forays. These not only would be energetically
costly but also precarious in terms of increased risk of predation to the turtles.
Although cooters are known to feed upon Brazilian elodea in the Wakulla
River (Lagueux et al. 1995), the effect of this exotic plant upon ecosystem integrity
is undoubtedly negative, and the park has examined the feasibility of its physical
removal (considered unlikely at present). More insidious is the threat of
encroachment by a second exotic plant, hydrilla (Hydrilla verticillata), which has
infested the river to within 1.25 km downstream of the park and appears to be
expanding upstream. Management personnel, while studying possibilities for
control of these exotic species, need to remain cognizant of the continuing need for
an abundant macrophytic food base for such herbivores as P. concinna. Potential
use of herbicides or biological control agents must be evaluated carefully, not only
for direct effects on the entire fauna but also for possible effects on non-target
native plants.
Research is needed on the Wakulla River to determine whether anthropogenic
changes in water quality are having biological effects. WSSP's cooters present at
least one concern. While the unusually thin egg shells in some clutches may not
represent a threat to the WSSP cooter population, the phenomenon in other species
has been linked in some instances to pollution, with dire consequences for
population viabilities. We suggest that the incidence of thinning be monitored
periodically, with special attention to subsequent clutches produced by females
known to have produced thin-shelled eggs previously. Results of this monitoring
may dictate the need for further research to identify the cause of egg shell thinning.
We also recommend three specific measures to reduce or eliminate physical
hazards to cooters at WSSP. First, turtles nesting near the main spring area and
lodge complex often find their migratory paths to or from the river blocked by a
JACKSON & WALKER: REPRODUCTION IN SUWANNEE COOTER
chain-link fence. Although most turtles eventually find their way back to the river,
this mechanical barrier increases their exposure to predators, high temperatures,
and potential human disturbance, and it restricts nesting to a line that is easily
patrolled by egg predators. The installation of "turtle doors" at ca 30-m intervals
along all sections of fence would greatly reduce such problems. Second, a small
solution hole adjacent to River Road trapped 1 percent of the nesting cooter
population during our study. We suggest the strategic use of logs to divert turtles
from it during their nesting migrations. Third, we recommend the installation of
propeller guards on park tour boats to reduce the risk of accidental mortality to
turtles inhabiting the uppermost river.
Recreation and Education.- Since the park's acquisition by the state,
proposals have been made to allow increased public access to the river and adjacent
uplands. We anticipate that increased use by humans of the uplands along the
floodplain during nesting season would lead to substantial nest abandonment
among cooters. This in turn would increase the risk of predation to nesting turtles
as a result of the extra nesting emergences that must be made. Further,
observations of Pseudemys in central Florida (M. Kramer pers. comm.) suggest
that moderate levels of boat and canoe traffic can cause thermoregulatory and
energetic stress to turtles that terminate basking behavior prematurely at the
approach of humans. We therefore recommend continued restricted access to the
majority of the river and adjacent uplands at WSSP.
Finally, as a result of this study, park management has undertaken efforts to
focus public and staff attention on this turtle via lectures, newspaper articles, signs,
displays, and closely monitored "turtle walks." These benefit not only Suwannee
cooters in the Wakulla River but populations of all species of turtles statewide. We
strongly encourage all such efforts to protect and promote nongame species in
Florida and elsewhere.
LITERATURE CITED
Allen, E. R. 1938. Notes on the feeding and egg-laying habits of the Pseudemys. Proc. Florida Acad. Sci.
3:105-108.
Andrews, R. M. 1982. Patterns of growth in reptiles. Pp. 273-320 in C. Gans and F. H. Pough (eds.),
Biology of the Reptilia, Vol. 13. Academic Press, New York.
Auffenberg, W. 1978. Threatened: Suwannee cooter. Pp. 32-33 in R. W. McDiarmid (ed.), Rare and
Endangered Biota of Florida, Vol. 3: Amphibians and Reptiles. Univ. Presses Florida, Gainesville.
Bennett, A. F., and W. R. Dawson. 1976. Metabolism. Pp. 127-223 in C. Gans and F. H. Pough (eds.),
Biology of the Reptilia, Vol. 5. Academic Press, London.
Berry, K. H. 1986. Desert tortoise (Gopherus agassizii) relocation: implications of social behavior and
movements. Herpetologica 42:113-125.
Berven, K. A. 1988. Factors affecting variation in reproductive traits within a population of wood frogs
(Rana sylvatica). Copeia 1988:605-615.
Bjorndal, K. A., and A. B. Bolten. 1988. Growth rates of immature green turtles, Chelonia mydas, on
feeding grounds in the southern Bahamas. Copeia 1988:555-564.
and A. Carr. 1989. Variation in clutch size and egg size in the green turtle nesting population at
Tortuguero, Costa Rica. Herpetologica 45:181-189.
Bock, B. C., and A. S. Rand. 1989. Factors influencing nesting synchrony and hatching success at a green
iguana nesting aggregation in Panama. Copeia 1989:978-986.
BULLETIN FLORIDA MUSEUM NATURAL HISTORY VOL 41(2)
Brauman, R. J., and R. A Seigel. 1995. Nesting habits and nest temperatures of the yellow-blotched map
turtle. Abstract, 38th Ann. Mtg. Soc. Stud. Amphib. Rept.:27.
Bryan, D. C. 1981. Territoriality and pair bonding in the limpkin (Aramus guarauna). M.S. thesis, Florida
State Univ., Tallahassee.
Buhlmann, K. A., and M. R. Vaughan. 1991. Ecology of the turtle Pseudemys concinna in the New River,
West Virginia. J. Herp. 25:72-78.
Burger, J. 1977. Determinants of hatching success in diamondback terrapin, Malaclemys terrapin. Amer.
Midi. Nat. 97:444-464.
Burt, W. H. 1943. Territoriality and home range concepts as applied to mammals. J. Mamm. 4:346-35.
Bury, R. B. 1979. Population ecology of freshwater turtles. Pp. 571-602 in M. Harless and H. Morlock
(eds.), Turtles: Perspectives and Research, John Wiley and Sons, New York.
Butler, B. 0., and T. E. Graham. 1995. Early post-emergent behavior and habitat selection in hatchling
Blanding's turtles, Emydoidea blandingii, in Massachusetts. Chelonian Conserv. Biol. 1:187-196.
Butler, J. A., and T. W. Hull. 1996. Reproduction of the tortoise, Gopherus polyphemus, in northeastern
Florida. J. Herp. 30:14-18.
Cagle, F. R. 1937. Egg laying habits of the slider turtle (Pseudemys troostii), the painted turtle (Chrysemys
picta), and the musk turtle (Sternotherus odoratus). J. Tennessee Acad. Sci. 12:87-95.
_. 1946. The growth of the slider turtle, Pseudemys script elegans. Amer. Midi. Nat. 36:685-729.
_. 1950. The life history of the slider turtle, Pseudemys scripta troostii (Holbrook). Ecol. Monogr. 20:31-
54.
_. 1954. Observations on the life cycles of painted turtles (Genus Chrysemys). Amer. Midl.Nat. 52:225-
235.
. 1955. Courtship behavior injuvenile turtles. Copeia 1955:307.
Cagle, K. D., G. C. Packard, K. Miller, and M. J. Packard. 1993. Effects of the microclimate in natural
nests on development of embryonic painted turtles, Chrysemyspicta. Functional Ecol. 7:653-660.
Cahn, A. R. 1937. The turtles of Illinois. Illinois Biol. Monogr. 35:1-218.
Caldwell, J. P., and J. T. Collins. 1981. Turtles in Kansas. AMS Publishing, Lawrence, Kansas.
Carr, A. F. 1937. A new turtle from Florida, with notes on Pseudemysfloridana mobiliensis (Holbrook).
Occ. Pap. Mus. Zool. No. 348.
_. 1938. A new subspecies of Pseudemys floridana, with notes on the floridana complex. Copeia
1938:105-109.
_. 1952. Handbook of Turtles. Comstock Publ. Assoc., Ithaca, New York.
_. 1975. The Ascension Island green turtle colony. Copeia 1975:547-555.
. 1983. All the way down upon the Suwannee River. Audubon 85(2):78-101.
_, and J. W. Crenshaw, Jr. 1957. A taxonomic reappraisal of the turtle Pseudemys alabamensis Baur.
Bull. Florida State Mus., Biol. Sci. 2(3):25-42.
_, and D. Goodman. 1970. Ecologic implications of size and growth in Chelonia. Copeia 1970:783-786.
Caughley, G. 1977. Analysis of Vertebrate Populations. John Wiley and Sons, New York.
Christens, E., and J. R. Bider. 1987. Nesting activity and hatching success of the painted turtle (Chrysemys
picta marginata) in southwestern Quebec. Herpetologica 43:55-65.
Christiansen, J. L., and B. J. Gallaway. 1984. Raccoon removal, nesting success, and hatchling emergence
in Iowa turtles with special reference to Kinosternon flavescens (Kinosternidae). Southwest. Nat.
29:343-348.
Chua, T. H., and J. I. Furtado. 1988. Nesting frequency and clutch size in Dermochelys coriacea in
Malaysia. J. Herp. 22:208-218.
Collins, J. T. 1991. Viewpoint: A new taxonomic arrangement for some North American amphibians and
reptiles. Herp. Rev. 22:42-43.
Conant, R., and J. T. Collins. 1991. A Field Guide to Reptiles and Amphibians of Eastern and Central
North America, third edition. Houghton Mifflin Co., Boston, Massachusetts.
Congdon, J. D., and R. E. Gatten, Jr. 1989. Movements and energetic of nesting Chrysemys picta.
Herpetologica 45:94-100.
-, and J. W. Gibbons. 1985. Egg components and reproductive characteristics of turtles: Relationships to
body size. Herpetologica 41:194-205.
_, and _. 1987. Morphological constraint on egg size: A challenge to optimal egg size theory? Proc. Nat
Acad. Sci., USA 84:4145-4147.
JACKSON & WALKER: REPRODUCTION IN SUWANNEE COOTER
-, and _. 1989. Biomass productivity of turtles in freshwater wetlands. Pp. 583-591 in R. Sharitz and J.
W. Gibbons (eds.), Freshwater Wetlands and Wildlife. USDOE Off. Sci. Tech. Inf., Oak Ridge,
Tennessee.
, and _. 1990a. The evolution of turtle life histories. Pp. 45-54 in J. W. Gibbons (ed.), Life History and
Ecology of the Slider Turtle. Smithsonian Inst Press, Washington, D.C.
, and _. 1990b. Turtle eggs: their ecology and evolution. Pp. 109-123 in J. W. Gibbons (ed.), Life
History and Ecology of the Slider Turtle. Smithsonian Inst Press, Washington, D.C.
, J. L Greene, and J. W. Gibbons. 1986. Biomass of freshwater turtles: A geographic comparison. Amer.
Midi. Nat. 115:165-173.
, D. W. Tinkle, G. L Breitenbach, and R. C. van Loben Sels. 1983. Nesting ecology and hatching success
in the turtle Emydoidea blandingi. Herpetologica 39: 417-429.
, and R. C. van Loben Sels. 1991. Growth and body size in Blanding's turtles (Emydoidea blandingi):
Relationships to reproduction. Canadian J. Zool. 69:239-244.
Cople, P. E., and M. A. Pilgrim. 1993. The effects of satellite nests on predation in artificial nests of
Pseudemys floridanapeninsularis. Florida Sci. 56 (suppl.):32.
Crenshaw, J. W. 1955. The Ecological Geography of the Pseudemys floridana Complex in the
Southeastern United States. Ph.D. diss., Univ. Florida, Gainesville.
Degenhardt, W. G., C. W. Painter, and A. H. Price. 1996. Amphibians and Reptiles of New Mexico. Univ.
New Mexico Press, Albuquerque.
DePari, J. A. 1996. Overwintering in the nest chamber of hatchling painted turtles, Chrysemys picta, in
northern New Jersey. Chelonian Conserv. Biol. 2:5-12.
Dimmitt, M. A. 1975. Terrestrial ecology of spadefoot toads (Scaphiopus): emergence cues, nutrition, and
burrowing habits. Ph.D. diss., Univ. California, Riverside.
Dobie, J. L 1971. Reproduction and growth in the alligator snapping turtle, Macroclemys temmincki
(Troost). Copeia 1971:645-658.
Dodd, C. K., Jr. 1988. Synopsis of the biological data on the loggerhead sea turtle Caretta caretta
(Linnaeus 1758). U.S. Fish Wildl. Serv., Biol Rep. 88(14). 110 pp.
K. M. Enge, and J. N. Stuart. 1988. Aspects of the biology of the flattened musk turtle, Sternotherus
depressus in northern Alabama. Bull. Florida State Mus., Biol. Sci. 34(1): 1-64.
Doroff, A. M., and L B. Keith. 1990. Demography and ecology of an ornate box turtle (Terrapene ornata)
population in south-central Wisconsin. Copeia 1990:387-399.
Dundee, H. A., and D. A. Rossman. 1989. The Amphibians and Reptiles of Louisiana. Louisiana State
Univ. Press, Baton Rouge.
Eckrich, C. E., and D. W. Owens. 1995. Solitary versus arribada nesting in the olive ridley sea turtles
(Lepidochelys olivacea): A test of the predator-satiation hypothesis. Herpetologica 51:349-354.
Ehrenfeld, D. W. 1979. Behavior associated with nesting. Pp. 417-434 in M. Harless and H. Morlock,
Turtles: Perspectives and Research. John Wiley and Sons, New York.
Ehrhart, L M. 1982. A review of sea turtle reproduction. Pp. 29-38 in K. A. Bjomdal (ed.), Biology and
Conservation of Sea Turtles. Smithsonian Inst. Press, Washington, D.C.
Elgar, M. A., and L. J. Heaphy. 1989. Covariation between clutch size, egg weight and egg shape:
Comparative evidence for chelonians. J. Zool. London 219:137-152.
Erben, H. K., J. Hoefs, and KI H. Wedepohl. Paleobiological and isotopic studies of eggshells from a
declining dinosaur species. Paleobiology 5:380-414.
Ernst, C. H., and R. W. Barbour. 1972. Turtles of the United States. Univ. Press Kentucky, Lexington.
and _. 1989. Turtles of the Worid. Smithsonian Inst. Press, Washington, D.C.
J. E. Lovich, and R. W. Barbour. 1994. Turtles of the United States and Canada. Smithsonian Inst
Press, Washington, D.C.
and G. R. Zug. 1994. Observations on the reproductive biology of the spotted turtle, Clemmys guttata,
in southeastern Pennsylvania. J. Herp. 28:99-102.
Etchberger, C. R., and J. B. Iverson. 1990. Pseudemys texana. Cat Amer. Amphib. Rept. 485.1-485.2.
Ewert, M. A. 1976. Nests, nesting and aerial basking of Macroclemys under natural conditions, and
comparisons with Chelydra'(Testudines: Chelydridae). Herpetologica 32:150-156.
. 1979. The embryo and its egg: development and natural history. Pp. 333-413 in M. Harless and H.
Morlock (eds.), Turtles: Perspectives and Research. John Wiley and Sons, New York.
. 1985. Embryology of turtles. Pp. 75-267 in C. Gans, F. Billet, and P. F. A. Maderson (eds.), Biology of
the Reptilia, Vol. 14, Development A. John Wiley and Sons, New York.
BULLETIN FLORIDA MUSEUM NATURAL HISTORY VOL. 41(2)
, and D. R. Jackson. 1994. Nesting ecology of the alligator snapping turtle (Macroclemys temminckii)
along the lower Apalachicola River, Florida. Florida Game and Fresh Water Fish Comm. Nongame
Wildl. Prog. Final Rep. NG89-020, Tallahassee.
, and C. E. Nelson. 1991. Sex determination in turtles: Diverse patterns and some possible adaptive
values. Copeia 1991:50-69.
Fahey, K. M. 1980. A taxonomic study of the cooter turtles, Pseudemys floridana (LeConte) and
Pseudemys concinna (LeConte), in the lower Red River, Atchafalaya River, and Mississippi River
basins. Tulane Stud. Zool. Bot. 22:49-66.
_. 1987. Aspects of the life history of the river cooter, Pseudemys concinna (LeConte), in the Tallapoosa
River, Tallapoosa County, Alabama. Ph.D. diss., Auburn Univ., Auburn, Alabama.
Florence, T. H., Jr., and G. H. Murphy. 1975. A telemetric study of the activity of the red-eared turtle,
Chrysemys scripta elegans. J. Tennessee Acad. Sci. 51:60.
Florida Department of Natural Resources. 1991. Edward Ball Wakulla Springs State Park Unit
Management Plan, review draft
Franz, R. 1986. Life history notes: Pseudemysfloridana peninsularis (peninsula cooter). Egg predation.
Herp. Rev. 17:64.
Frazer, N. B., and L M. Ehrhart 1985. Preliminary growth models for green, Chelonia mydas, and
loggerhead, Caretta caretta, turtles in the wild. Copeia 1985:73-79.
, J. W. Gibbons, and J. L Greene. 1989. Pp. 183-200 in J. W. Gibbons (ed.), Life History and Ecology of
the Slider Turtle. Smithsonian Inst. Press, Washington, D.C.
, and J. I. Richardson. 1985. Annual variation in clutch size and frequency for loggerhead turtles, Caretta
caretta, nesting at Little Cumberland Island, Georgia, USA. Herpetologica 41:246-251.
Frost, D. R., and D. M. Hillis. 1990. Species in concept and practice: Herpetological applications.
Herpetologica 46:87-104.
Galbraith, D. A., R. J. Brooks, and M. E. Obbard. 1989. The influence of growth rate on age and body size
at maturity in female snapping turtles (Chelydra serpentina). Copeia 1989:896-904.
Garrott, R. A., P. J. White, and C. A. Vanderbilt White. 1993. Overabundance: An issue for conservation
biologists? Conserv. Biol. 7:946-949.
Gemmell, D. J. 1970. Some observations on the nesting of the western painted turtle, Chrysemys picta
belli, in northern Minnesota. Canadian Field-Nat. 84:308-309.
Georges, A., and .R Kennett. 1989. Dry-season distribution and ecology of Carretochelys insculpta
(Chelonia: Carettochelyidae) in Kakadu National Park, Northern Australia. Aust. Wildly. Res.
16:323-335.
Germano, D. J. 1993. Shell morphology of North American tortoises. Amer. Midi. Nat. 129:319-335.
Gibbons, J. W. 1986. Movement patterns among turtle populations: applicability to management of the
desert tortoise. Herpetologica 42:104-113.
1987. Why do turtles live so long? BioScience 37:262-269.
_. 1990. Turtle studies at SREL: A research perspective. Pp. 19-44 in J. W. Gibbons (ed.), Life History
and Ecology of the Slider Turtle. Smithsonian Inst. Press, Washington, D.C.
and J. L Greene. 1990. Reproduction in the slider and other species of turtles. Pp. 124-134 in J. W.
Gibbons (ed.), Life History and Ecology of the Slider Turtle. Smithsonian Inst. Press, Washington,
D.C.
_, and J. D. Congdon. 1990. Temporal and spatial movement patterns of sliders and other turtles. Pp.
201-215 in J. W. Gibbons (ed.), Life History and Ecology of the Slider Turtle. Smithsonian Inst.
Press, Washington, D.C.
_, and K. K. Patterson. 1982. Variation in reproductive characteristics of aquatic turtles. Copeia
1982:776-784.
R. D. Semlitsch, J. L Greene, and J. P. Schubauer. 1981. Variation in age and size at maturity of the
slider turtle (Pseudemys scripta). Amer. Nat. 117:841-845.
Gilbert, C. R. 1987. Zoogeography of the freshwater fish fauna of southern Georgia and peninsular Florida.
Brimleyana 13:25-54.
Giovanetto, L A. 1992. Population ecology and relative abundance of sympatric freshwater turtles in the
headwaters of two spring-fed rivers in western peninsular Florida. Ph.D. diss., Florida Inst. Tech.,
Melbourne.
Goode, J. 1965. Nesting behavior of freshwater tortoises in Victoria. Victoria Nat. 82:218-222.
Graham, T. E., and T. S. Doyle. 1977. Growth and population characteristics of Blanding's turtle,
Emydoidea blandingii, in Massachusetts. Herpetologica 33:410-414.
