2008 SOUTHEASTERN NATURALIST 7(2):311–322
Annual Surveys of Larval Ambystoma cingulatum Reveal
Large Differences in Dates of Pond Residency
Mark S. Bevelhimer1,*, Dirk J. Stevenson2, Neil R. Giffen1,
and Kara Ravenscroft3
Abstract - Effective sampling of pond-dwelling larval stages of the federally listed
Ambystoma cingulatum (Flatwoods Salamander) requires sufficient knowledge of
when larvae are present and how best to sample them. Through systematic sampling
with active and passive sampling techniques, we found dipnetting to be significantly
more effective than three types of passive traps. During surveys for Flatwoods Salamander
larvae at Fort Stewart Military Installation, GA in 2005 and 2006, we found
that pond residency varied by at least 1.5 months between the 2 years due to the timing
of pond filling. In addition, our latest capture on 23 May 2005 was about 2 weeks later
than previously recorded at any site range-wide. A simple growth model was used to
evaluate likely hatching dates based on significant rain events, observed sizes at capture,
and likely growth rates. This analysis suggested that the primary dates of hatching
occurred in late February 2005 and early January 2006, a difference that corresponds
to that seen in the residency of the latest larval stages. A review of the survey records
for Fort Stewart for the past 13 years shows a steep decline in the number of occupied
ponds from near 20 to a single pond for the past two years (the only documented breeding
success in a natural pond since 1999).
Introduction
Ambystoma cingulatum Cope (Flatwoods Salamander) was listed as federally
threatened in 1999 due to range-wide population declines attributable
to habitat loss and habitat conversion for silviculture, agriculture, and residential
and commercial development (USFWS 1999). Restricted to northern
Florida and the Coastal Plain of South Carolina, Georgia, and Alabama, this
species is endemic to mesic fl atwoods and savannahs dominated by Pinus
palustris Mill (longleaf pine) and Aristida stricta Michx. (wiregrass), where
it breeds in small (1.5-ha mean size), isolated depressional wetlands (Palis
1997). Wetlands used by breeding Flatwoods Salamanders are ephemeral
and usually dry on an annual basis (Anderson and Williamson 1976, Palis
1997). The basins of breeding ponds are usually abundantly vegetated with
graminaceous vegetation and are often partially forested with Taxodium
ascendens Brongn (pond cypress), Nyssa sylvatica Marsh (black gum), and
Ilex myrtifolia Walt (myrtle-leaf holly) (Jensen 1999; Palis 1996, 1997).
During the nonbreeding season, adult Flatwoods Salamanders are fossorial
and inhabit crayfish burrows and other ground cavities within mesic fl atwoods
and savannah habitats located near breeding ponds (Palis 1996). At the time
1Environmental Sciences Division, Oak Ridge National Laboratory, Oak Ridge, TN
37831-6351. 2Fort Stewart Wildlife Management Branch, 1557 Frank Cochran Drive,
Building 1145, Fort Stewart, GA 31314. 3PO Box 301, Wabash, IN 46992. *Corresponding
author - bevelhimerms@ornl.gov.
312 Southeastern Naturalist Vol.7, No. 2
when adults migrate to breeding sites (mid-October–mid-December), the
basins of breeding ponds are typically dry (Anderson and Williamson 1976;
D.J. Stevenson, unpubl. data; Palis 1997). Females deposit eggs terrestrially in
moist microhabitats (e.g., at entrances to crayfish burrows or under sphagnum
moss, leaf litter, or dead grass) (Anderson and Williamson 1976). Ambystoma
opacum Gravenhorst (Marbled Salamander), also a fall breeder, is the only other
ambystomatid salamander species that deposits eggs terrestrially (Petranka
1998). Flatwoods Salamander eggs begin to develop immediately after they are
laid, but do not hatch until inundated by rising pond levels, which might occur
weeks or even months after deposition (Anderson and Williamson 1976; Palis
1995, 1997), typically December to February. The aquatic larvae of Flatwoods
Salamanders inhabit a specific microhabitat—graminaceous vegetation of linear
growth form (Palis 1996, Sekerak et al. 1996)—that is likely maintained by
occasional fires burning into or through the dry pond basins (Bishop and Haas
2005, Palis 1997, Sekerak et al. 1996). Larval development is completed in 11
to 18 weeks, and timing of metamorphosis may be infl uenced by pond drying
(Palis 1995). More detailed life-history and ecology information can be found
in USFWS (1999) and Palis (1996, 1997).
The current distribution of the Flatwoods Salamander is mostly restricted
to large tracts of public lands, such as national forests, state parks, and military
bases. One such sanctuary is Fort Stewart Military Installation, which
is home to many rare and imperiled species of reptiles and amphibians (Stevenson
1999, USFWS 1999). Herpetofaunal inventories conducted at Fort
Stewart by the Savannah Science Museum during the mid- to late 1970s
(Williamson and Moulis 1979) , The Nature Conservancy from 1992–1995
(Gawin et al. 1995), and the Fort Stewart Wildlife Branch Office from 1996
to present confirmed Flatwoods Salamander breeding on at least one occasion
at 25 ponds (henceforth referred to as “known” ponds).
Successful management and conservation of Flatwoods Salamander
habitat at Fort Stewart and elsewhere will depend on successfully identifying
active and potential breeding sites among hundreds of potential ponds.
Because the only viable field method for locating adults (monitoring drift
fences during adults’ migrations to and from breeding sites) is labor and time
intensive, aquatic sampling of breeding sites for Flatwoods Salamander larvae
is the preferred method to assess presence (Bishop et al. 2006, Palis 1996).
Sampling for larvae is also challenging because they typically reside in dense
vegetation that is difficult to sample, densities are often low, and abundance
and period of pond residency varies annually and among ponds (Bishop et al.
2006, Palis 1997, Sekerak et al. 1996). For these reasons, it is crucial that biologists
use the most effective sampling methodology and clearly understand
the factors that affect salamander presence and detection.
The objectives of this study were to 1) systematically sample larval Flatwoods
Salamanders using different sampling methods to compare method
effectiveness, 2) survey throughout the larval development period for two
years to determine the period of larval residency at Fort Stewart, 3) relate
2008 M.S. Bevelhimer, D.J. Stevenson, N.R. Giffen, and K. Ravenscroft 313
hatching to rain events using a simple growth model, and 4) summarize survey
success at Fort Stewart over the past 13 years.
Methods
Fort Stewart (113,064 ha) is located in the Atlantic Coastal Plain of
southeastern Georgia and supports significant areas of intact, fire-maintained
longleaf pine ecosystems that contain embedded depressional wetlands
(Carlile 1995, Gawin et al. 1995). Within these longleaf pine ecosystems are
nearly 500 ponds that have been identified as potential Flatwoods Salamander
breeding habitat.
In 2005, we systematically sampled eight known breeding ponds intermittently
from late February to early May with traps and dipnetting to test
the relative effectiveness of different capture methods. The ponds ranged
in size from about 1.3 to 4 ha. Three of the eight ponds were located on
the west side of Fort Stewart, and five on the east side about 34 km away;
all are located in Liberty County. Within a pond, we identified and marked
with numbered fl agging tape at least 16 sample sites based on a qualitative
judgment of what we considered suitable larval habitat with adequate water
depth for sampling. These sites were vegetated with moderate to profuse
herbaceous cover and had an average water depth of about 15–30 cm.
We sampled larvae with either long-handled 40-cm diameter dipnets
(5-mm mesh) or passive traps. The trap types included commercially-made
plastic (5-mm mesh) or metal (3-mm mesh), double-opening, funnel traps
and a 61- x 61- x 46-cm wood-framed box trap with 2 vertical funnel entrances
(3-mm mesh). Traps were placed at a depth or perched on debris such
that air-breathing organisms had access to the water surface. For each typical
day of effort within a pond, we randomly assigned: 1) to four sites, a standard
effort of dipnetting, which was usually 2 surveyors dipnetting for 5 min
each; 2) to four sites, a combination of 4 plastic and 4 metal minnow traps
deployed for 24 hours; and 3) to two sites, a box trap deployed for 24 hours.
We rotated dipnetting and trapping efforts daily among the sites within a
pond for 2–4 days. For example, during 3 days of sampling at one pond we
would typically dipnet a total of 120 minutes (3 days x 4 sites x 2 netters x
5 minutes), trap for 96 trap nights with both metal and plastic funnel traps
(3 nights x 4 sites x 4 traps x 2 trap types), and trap for 6 box-trap nights (3
nights x 2 traps). We measured all Flatwoods Salamander larvae captured
(snout–vent length [SVL] and total length [TL] in mm) and released them
shortly after processing to their original site of capture. The amount of time
required to set and check traps was also documented. Additional dipnetting
without trapping was performed at these ponds and two other known ponds
as part of a general survey.
In 2006, we modified our sampling strategy based on 2005 results and
designed a sampling plan that included more ponds but less effort per pond
(i.e., fewer trips) and sampling primarily with dipnets. We sampled a total
of 60 ponds (21 known breeding ponds and 39 potential breeding ponds)
314 Southeastern Naturalist Vol.7, No. 2
that ranged in size from about 0.14 to 7.5 ha. Most ponds (46) were sampled
only once, 13 were sampled two to three times, and the single pond where
Flatwoods Salamander larvae were found during this study was sampled on
eight occasions.
We evaluated possible hatching dates by matching observed sizes at
capture with projected size at age based on modeled growth, which was initiated
on dates of rain events that caused pond levels to rise and could have
triggered hatching. Hatching of Flatwoods Salamander larvae on multiple
dates within a season is not unusual (Palis 1995, Sekarek et al. 1996) and
was considered in our analysis. We obtained daily rainfall data from two
meteorological stations located equidistant (10 km) to the east and west of
the single pond where larvae were found during this study and averaged the
daily values. For modeling purposes we considered rain events in January
and February that resulted in a daily total >20 mm or a running weekly total
>25 mm as sufficient to result in a rise in pond levels.
Size at hatching information used in our model was based on total lengths
reported by Anderson and Williamson (1976), which we converted to SVL
based on the ratio of SVL to TL calculated for the larvae captured in this
study (SVL= 0.55 * [TL]). They reported average lengths at hatching based
on laboratory observations of 6.5 mm (SVL) on 25 November, 6.2 mm on 2
December, and 8.5 mm on 24 January after conversion. They also reported
an average length of 7.0 mm (SVL) for newly hatched larvae captured in the
field on 14 December. From these data, we derived a relationship between
date of hatching and mean SVL at hatching:
Mean SVLhatch = 0.0446 * Julian day + 7.658
We then subtracted and added 1 mm to include a measure of natural variation
in size at hatching, which became the origin of the minimum and maximum
growth trajectories for a hatching date.
We also recognize that individual variation in growth exists, and, therefore,
we used the range of growth rates calculated by Palis (1995) for two
breeding sites in the Florida panhandle as minimum (1.78 mm/week; 0.254
mm/day) and maximum (2.54 mm/week; 0.363 mm/day) rates for the model.
Although larval growth rates vary due to several factors, such as temperature,
food availability, and densities of conspecifics and competitors, because we
lacked any other growth data for this species, we assumed that larval growth
at Fort Stewart was within the range reported by Palis (1995).
The growth trajectory modeled from a particular date was comprised of
both minimum and maximum trajectories that created an envelope or cone
of likely size at age. The maximum and minimum size-at-age lines are described
by the following equations:
Minimum trajectory SVLt (mm) = (SVLhatch - 1) + t * 0.254,
and
Maximum trajectory SVLt (mm) = (SVLhatch + 1) + t * 0.363,
where t is time in days from hatching.
2008 M.S. Bevelhimer, D.J. Stevenson, N.R. Giffen, and K. Ravenscroft 315
Lastly, we obtained survey results from the Fort Stewart Wildlife Branch
Office of all the Flatwoods Salamander surveys at Fort Stewart since 1994
that included the dates and types (larval or adult) of surveys and the number
and sizes of Flatwoods Salamanders found. These data were summarized for
each of the 22 confirmed breeding sites.
Results
Table 1 summarizes the temporal distribution of pond sampling in 2005
and 2006 and the capture of Flatwoods Salamander larvae. In mid-February
2005, when we first visited six known ponds for sampling, they were dry except
for a few small (2- x 2-m) shallow pools. By late February, these ponds
had filled to a depth sufficient for sampling (maximum ≈30 cm). No Flatwoods
Salamander larvae were captured during February and March after sampling
an average of 7 days each at six ponds with dipnets and traps. Sampling effort
was reduced in April due to concerns that earlier dry conditions had resulted in
Table 1. Number of ponds sampled and number and mean snout–vent length (SVL) of Ambystoma
cingulatum (Flatwoods Salamander) larvae captured at Fort Stewart on a weekly basis
during 2005 and 2006. Some ponds were sampled more than once during a week. All larvae
were captured from a single pond, Alpha Pond.
2005 2006
Sampling No. of Mean No. of Mean
period ponds A. c. SVL ponds A. c. SVL
(week of) sampled larvae (mm) sampled larvae (mm)
January
19 - 6 2 15.5
26 - 4 1 16.0
February
2 - 10 0 -
9 - 5 11 16.9
16 - 12 0 -
23 3 0 - 14 0 -
March
2 4 0 - 12 0 -
9 - 10 7 29.4
16 3 0 - -
23 3 0 - -
30 - 6 0 -
April
6 - 2 6 31.7
13 - 8 0 -
20 1 15 32.0 4 0 -
27 5 0 - -
May
4 3 15 34.3 -
11 3 5 32.0 -
18 1 2 33.5 -
25 1 1 32.0 -
June
1 2 0 - -
316 Southeastern Naturalist Vol.7, No. 2
reproductive failure, but on 21 April, we captured a larva at a known breeding
pond (Alpha Pond) in a trap followed by 14 more individuals by dipnetting.
From 21 April to 23 May, we captured a total of 38 larvae from Alpha Pond
during seven visits (Table 1). No other larvae were found in Alpha Pond during
sampling from 31 May–3 June, and none were found in follow-up surveys
of other ponds during late April and May.
In 2006, the ponds had filled when sampling began in the middle of January.
We captured two Flatwoods Salamander larvae at Alpha Pond during the
first visit on 18 January. On four subsequent visits (27 January, 7–8 February,
8 March, and 4 April), 27 additional larvae were captured by dipnetting.
Ponds dried during March, and larvae captured on 4 April were dipnetted
from among the few remaining shallow pools located within the Alpha Pond
basin. We found no larvae in the other 59 ponds sampled in 2006, including
one that is within 100 m of Alpha Pond.
Dipnetting produced significantly more larvae than any of the traps we
tested (Table 2). Our method comparison in 2005 included 1744 minutes of
dipnetting coincident with 1794 total trapnights (838 metal traps, 840 plastic
traps, and 116 box traps) at eight ponds. Because the typical unit of effort for
trapping is number of nights set and the unit of effort for dipnetting is minutes
netted, we chose to standardize effort for comparison based on actual handson
investigator time needed to use each method. After standardization, the
amount of effort at each site within a pond was similar for all methods. Our
typical dipnetting effort per site was 5 minutes of netting by two people for
a total of 10 minutes. It also took about 10 minutes for two people to set and
check eight minnow traps (plastic or metal), which was the number normally
placed at a site overnight. A single box trap took about 5 minutes to set and
check because it often captured more individual organisms and took longer to
empty because of its design. When standardized to hours of investigator effort,
we found that dipnetting was roughly 5–10 times more effective than the traps.
Since Flatwoods Salamander larvae were only captured in one pond, we limited
our analysis to only those dates when we were certain larvae were in the
pond. Based on data from 20–21 April and 2–4 May, metal traps captured two
Table 2. Trapping and dipnetting effort in Alpha Pond in 2005 with number of Flatwoods Salamander
larvae captured in parentheses.
Metal Plastic Box
traps traps traps Dipnet
Sampling dates (trapnights) (trapnights) (trapnights) (minutes)
March 1–3 44 44 6 122
March 21–23 48 48 6 60
April 20–21 49 48 (1) - 220 (14)
May 2–4 68 (2) 68 7 211 (13)
May 11–12 - - - 244 (5)
May 17 - - - 89 (2)
May 23–24 - - - 85 (1)
May 31–Jun 3 12 14 - 165
Total 221 (2) 222 (1) 19 787 (35)
2008 M.S. Bevelhimer, D.J. Stevenson, N.R. Giffen, and K. Ravenscroft 317
larvae in the equivalent of 146 minutes (0.8/hour), plastic traps captured one
larva in the equivalent of 145 minutes (0.4/hour), and dipnetting captured 27
larvae in 431 minutes (3.8/hour).
Our analysis of the rainfall data revealed that in 2005 three dates from
1 January to 28 February met the criteria we established for what was necessary
to raise pond levels (14 January, 3 and 28 February), and four dates
(2 and 24 January, and 3 and 26 February) met the criteria in 2006. We
produced size-at-age envelopes (i.e., minimum and maximum growth trajectories)
for each of these dates and evaluated how well the size envelopes
encompassed the observed larval sizes. Figure 1 shows an envelope from the
most likely hatching date defined as the first day with daily rainfall of >25
mm and another from the remaining hatching dates that encompassed the
greatest number of the remaining larval sizes.
Biologists at Fort Stewart have maintained comprehensive records of Flatwoods
Salamander sampling effort and captures at 22 confirmed breeding sites
over the past 13 years (Fig. 2). Prior to the 2 years of sampling reported here,
nearly all of the sampling occurred in February to early April. Of 86 sampling
Figure 1. Snout–vent length (mm) for larval Flatwoods Salamander by date of capture
from Alpha Pond at Fort Stewart, GA. Envelopes of modeled size-at-age for two
possible hatching dates for 2005 and 2006 are shown as solid lines (most likely date
of hatching) and dashed lines (second most likely). Daily rainfall (mm) is indicated
by the gray line.
318 Southeastern Naturalist Vol.7, No. 2
trips to known ponds from 1994 to 2004, only three ponds were sampled in
January and none after 13 April. These records show a decline since 1994 in
the proportion of confirmed ponds surveyed in a given year that contained
Flatwoods Salamanders. Nearly every previously confirmed pond that was
sampled in 1994 (18 of 19 sampled) produced larvae. In the late 1990s, about
half of the confirmed ponds that were sampled each year had larvae present. A
protracted drought (1999–2002) experienced throughout the Coastal Plain of
Georgia and South Carolina resulted in four consecutive years of potentially
complete reproductive failure at most breeding sites. Although some pond basins
on Fort Stewart were partially inundated in 1999, pond hydroperiods were
of insufficient duration to allow larval development through metamorphosis
(D.J. Stevenson, unpubl. data). In the past 2 years, we only found larvae in 1
of 21 known ponds sampled, even though survey effort (based on pond visits)
was as high or higher than ever.
Discussion
The difference in dates of initial hatching and latest occupancy of Flatwoods
Salamanders in 2005 and 2006 at Alpha Pond at Fort Stewart was
roughly 1.5 months. Of particular significance is the presence of larvae in
a breeding pond until at least 23 May. Previously, the latest date that larvae
had been observed at Fort Stewart was 13 April (1994), and that observation
along with those from a day earlier were of larvae nearing metamorphosis
Figure 2. Compiled survey history for Flatwoods Salamander on Fort Stewart from
1994–2006 showing unsuccessful and successful surveys for larval salamanders
(open and filled diamonds), observations of adults or metamorphs outside of the
ponds (squares and triangles), and periods when ponds were dry and could not be
sampled (bars). Pond #10 is Alpha Pond where larvae were found in 2005 and 2006.
Sources of data were Fort Stewart Wildlife Branch Office records, Safer (2001), and
this study.
2008 M.S. Bevelhimer, D.J. Stevenson, N.R. Giffen, and K. Ravenscroft 319
(Gawin et al. 1995). The latest larvae capture dates of which we are aware
are 1 May (1974) and 12 May (1972) (Williamson and Moulis 1979); these
collections were made in Jasper County, SC. Although rarely observed, late
occupancy is not necessarily a rare event; Williamson and Moulis (1979)
captured larvae in the latter half of April or later in 4 consecutive years in
South Carolina.
We do not know when the eggs were deposited that likely hatched 28
February 2005 because several rain events occurred in November and December
of 2004 that could have triggered breeding migrations. The early
January hatching date estimated for 2006 is probably not the earliest possible
for this study site, because Anderson and Williamson (1976) observed
eggs hatching as early as 4 December in southeastern South Carolina and
southeastern Georgia. Little has been published regarding how long Flatwoods
Salamander eggs (located in dry pond basins) remain viable after
deposition before inundation; however, it is possible that larvae that hatched
in late February 2005 were from eggs deposited 3 months earlier. Anderson
and Williamson (1976) reported that advanced eggs taken from the field
hatched in the laboratory approximately 74 days later. The terrestrially deposited
eggs of Marbled Salamanders, a related species, may remain viable
3–4 months post-oviposition (Noble and Brady 1933, Petranka and Petranka
1981). Flexibility in this aspect of reproduction is critical to Flatwoods Salamanders
if rainfall during the reproductive season is below normal as it has
been in southeastern Georgia for many of the last 10 years.
Our simple modeling of larval growth suggests that for the 2 years of our
study a single date of hatching does not account for all the larval sizes observed.
In 2005, we observed nearly dry ponds in mid-February and presumed that the
larvae captured later in the spring hatched following significant rain events in
late February or early March. However, the growth envelope initiated on 28
February does not include the largest larvae captured on 21 April, which suggests
that hatching also occurred on an earlier date. Four days of rain from 29
January to 3 February that totaled 25 mm was probably enough to partially fill
the pond and could have inundated eggs at low pond elevations. Some larvae
may have hatched during this partial pond filling and survived the following
3 weeks of minimal rain by taking refuge in small pools that remained. Alternatively,
it is possible that we underestimated the actual larval growth rate in
our model; however, if that were the case, some larvae hatched on 28 February
would have had to grow at a rate about 36% greater than the maximum rate
estimated by Palis (1995). In 2006, the most likely date for hatching was in
early January. However, hatching at that time does not account for all the sizes
of larvae observed. We believe that some larvae must have also hatched in late
January to account for the smaller individuals captured in early April.
Although we illustrated the two most-likely hatch dates based on our analysis,
we do not rule out the possibility that larvae hatched on more than two
dates given the uncertainty in size at hatching and known variation in growth
rates. It is quite possible that multiple hatching dates within a population
320 Southeastern Naturalist Vol.7, No. 2
during a season is the rule and not the exception. Since eggs are laid individually
and not in large egg masses, multiple females would likely deposit their
eggs at a variety of elevations within a dry depression. Gradual or incremental
pond filling would therefore result in multiple hatching dates.
Of over 2000 isolated depressional ponds on Fort Stewart, approximately
500 have been identified as potential breeding habitat for Flatwoods Salamanders
(Palis 2002). Less than half of these ponds have been sampled to
date, and most of those have not been sampled enough to conclude that
they do not support breeding. Better knowledge of when larvae are present
and most susceptible to specific sampling methods is crucial to successful
monitoring and to maximize likelihood of detection. Sampling methods limit
detectability during the first few weeks of larval residency because larval
size at that stage is smaller than the mesh size of nets and traps. For example,
in 2005, we thoroughly sampled Alpha pond twice during March without
finding any larvae, but we found larvae in late April, and we are certain,
based on the size of the larvae, that they were present throughout March.
Our study demonstrates that dipnetting is far more effective for surveying
Flatwoods Salamander larvae than passive traps. Although checking a
single trap for larvae can be done in less than a minute, the time it takes to
transport the traps to the pond, distribute them throughout the pond, and
locate the proper depth for deployment accumulates to significant time expenditure
for minimal return.
We suggest that the most opportune time to sample is during the second
and third months after a pond fills to at least half full. In Florida, metamorphosis
of Flatwoods Salamander larvae is usually complete by April (Sekarek
et al. 1996), but is likely later on average at more northern latitudes.
In the Fort Stewart region, April–May is a period of reduced rainfall and
pronounced evapotranspiration; thus, pond water levels recede rapidly during
this time (Palis 1997). Bishop et al. (2006) recommended that surveys for
Flatwoods Salamander larvae be conducted primarily from February to early
April, but depending on various weather-related factors, sampling in other
months could be fruitful. The results of our study demonstrate that, during
some years and for some locations, sampling outside of the recommended
months is certainly productive. Thus, we recommend extending surveys
through April and into May in years when breeding ponds do not fill until
late winter (February–March).
Flatwoods Salamander larvae captured in 2005 were the first captured
at Fort Stewart from a natural wetland since 1999 (larvae were found in a
former borrow pit in 2001) (Fig. 2). Since 2002, repeated surveys at known
breeding sites elsewhere in Georgia and in South Carolina have found larvae
at just one site in South Carolina (a single larva found in 2003 at a site
on Francis Marion National Forest; S. Bennett, South Carolina Department
of Natural Resources, Columbia, SC, pers. comm.) and at one site in Georgia
(single larva found in each of 2001 and 2003 from adjacent wetlands
on Townsend Bombing Range; J. Jensen, Georgia Department of Natural
2008 M.S. Bevelhimer, D.J. Stevenson, N.R. Giffen, and K. Ravenscroft 321
Resources, Forsyth, GA, and W. Seyle, US Army Corps of Engineers, Savannah,
GA, pers. comm.). Biologists have found Flatwoods Salamanders on
a more frequent basis at many sites in Florida during this period (K. Enge,
Florida Fish and Wildlife Conservation Commission, Tallahassee, FL, pers.
comm.) including high numbers observed at St. Marks National Wildlife
Refuge in late February 2007 (M.S. Bevelhimer, unpubl. data).
Lastly, we do not know whether the high level of occurrence reported in
1994 was the result of ideal hydrologic conditions, a peak in a cyclic pattern
of natural population fl uctuation, a result of greater survey effort, or a
combination of these and other environmental factors. Palis et al. (2006)
observed a decline in the number of breeding adult Flatwoods Salamanders
over 4 consecutive years at a breeding pond in Florida and attributed this
decline to adult attrition, lack of juvenile recruitment, and lack of rain or
abnormally low rain during the period of breeding migrations. Similarly,
we suspect that adult attrition and lack of juvenile recruitment due to the
drought are responsible for the putative decline of Flatwoods Salamanders
on Fort Stewart and elsewhere in Georgia and South Carolina. If the conservation
and preservation of this and other rare amphibian species is to be
successful, biologists must identify and utilize which survey methods are
most effective and should maximize the likelihood of detection through a
better understanding of the relationship between pond residency and various
environmental factors.
Acknowledgments
This research was sponsored by the Strategic Environmental Research and Development
Program and performed at Fort Stewart Military Installation and Oak
Ridge National Laboratory (ORNL). ORNL is managed by UT-Battelle, LLC, for
the US Department of Energy under contract DE-AC05-00OR22725. We appreciate
the support of Tim Beaty, Larry Carlile, Stella Osborn, and Dena Thompson of
the Fort Stewart Wildlife Branch Office throughout all aspects of this study. Fort
Stewart Range Control staff kindly assisted with access issues. We thank Mike Ravenscroft
and Will Fields for their assistance in the field. The submitted manuscript
has been authored by a contractor of the US Government under contract DE-AC05-
00OR22725. Accordingly, the US Government retains a nonexclusive, royalty-free
license to publish or reproduce the published form of this contribution, or allow others
to do so, for US Government purposes.
Literature Cited
Anderson, J.D., and G.K. Williamson. 1976. Terrestrial mode of reproduction in
Ambystoma cingulatum. Herpetologica 32(2):214–21.
Bishop, D.C., and C.A. Haas. 2005 Burning trends and potential negative effects of
suppressing wetland fires on Flatwoods Salamanders. Natural Areas Journal 25:
290–294.
Bishop, D.C., J.G. Palis, K.M. Enge, D.J. Printiss and D.J. Stevenson. 2006. Capture
rate, body size, and survey recommendations for larval Ambystoma cingulatum
(Flatwoods Salamanders). Southeastern Naturalist 5(1):9–16.
322 Southeastern Naturalist Vol.7, No. 2
Carlile, L.D. 1995. Fire effects on threatened and endangered species and habitats of
Fort Stewart Military Reservation, Georgia. Pp. 227–231, In J.M. Greenlee (Ed.).
Proceedings: Fire Effects on Rare and Endangered Species and Habitats Conference.
International Association of Wildland Fire, Coeur d’ Alene, ID. 343 pp.
Gawin, L., K. Lutz, D. Mikesic, and D. Stevenson. 1995. Fort Stewart inventory final
report. The Nature Conservancy, GA Field Office, Pembroke, GA.
Jensen, J.B. 1999. Flatwoods Salamander (Ambystoma cingulatum). Pp. 92–93, In
T.W. Johnson, J.C. Ozier, J.L Bohannon, J.B. Jenson, and C. Skelton (Eds.). Protected
Animals Of Georgia. Nongame-Endangered Wildlife Program, Georgia
Department of Natural Resources, Wildlife Resources Division, Nongame Wildlife-
Natural Heritage Section. Forsyth, GA. 247 pp.
Noble, G.K., and M.K. Brady. 1933. Observations on the life history of the Marbled
Salamander, Ambystoma opacum. Zoologica1 1:89–132
Palis, J.G. 1995. Larval growth, development, and metamorphosis of Ambystoma
cingulatum on the Gulf Coastal Plain of Florida. Florida Scientist 58:352–358.
Palis, J.G. 1996. Element stewardship abstract for Ambystoma cingulatum, Flatwoods
Salamander. Natural Areas Journal 16:49–54.
Palis, J.G. 1997. Distribution, habitat, and status of the Flatwoods Salamander (Ambystoma
cingulatum) in Florida, USA. Herpetological Natural History 5:53–65.
Palis, J.G. 2002. Distribution of potential habitat of the federally threatened Flatwoods
Salamander (Ambystoma cingulatum) on Fort Stewart, Georgia. Unpublished
report to the Fort Stewart Fish and Wildlife Management Branch, Fort
Stewart, GA. (contract DAKF10-01-P-0265). 5 pp.
Palis, J.G., M.J. Aresco, and S. Kilpatrick. 2006. Breeding biology of a Florida
population of Ambystoma cingulatum (Flatwoods Salamander) during a drought.
Southeastern Naturalist 5(1):1–8.
Petranka, J.W. 1998. Salamanders of the United States and Canada. Smithsonian
Institution Press. Washington, DC. 587 pp.
Petranka, J.W., and J.G. Petranka. 1981. On the evolution of nest-site selection in the
Marbled Salamander, Ambystoma opacum. Copeia 1981:387–391.
Safer, A. 2001. Natural history and ecology of the Flatwoods Salamander, Ambystoma
cingulatum, on the Atlantic Coastal Plain. M.Sc. Thesis. Georgia Southern
University, Statesboro, GA.
Sekerak, C.M., G.W. Tanner, and J.G. Palis. 1996. Ecology of Flatwoods Salamander
larvae in breeding ponds in Apalachicola National Forest. Proceedings of the Annual
Conference of the Southeastern Association of Fish and Wildlife Agencies
50:321–330.
Stevenson, D.J. 1999. The herpetofauna of Fort Stewart, Georgia: Habitat occurrence,
status of protected and rare species, and species diversity. Unpublished
report to the Fort Stewart Fish and Wildlife Branch, Fort Stewart, GA 98 pp.
US Fish and Wildlife Service (USFWS). 1999. Final rule to list the Flatwoods Salamander
as a threatened species. Federal Register 64(62):15691–15704.
Williamson, G.K., and R.A. Moulis. 1979. Survey of reptiles and amphibians on
Fort Stewart and Hunter Army Airfield. Report to US Army, Fort Stewart, GA
(contract DACA 21-77-c-0155).